Title: The World of the Huns
Subtitle: Studies in Their History and Culture
Date: 1973
Notes: Edited by Max Knight. Note: the text currently contains many errors where other language letters haven't been converted correctly from the source PDF.
ISBN: 978–0520015968

    [Front Matter]


      [Title Page]


      List of Illustrations


      Editor’s Note

      Fragments from the Author’s Preface

      Author’s Acknowledgments (Fragments)

    I. The Literary Evidence



      Ammianus Marcellinus

      Cassiodorus, Jordanes

    II. History

      From the Don to the Danube

        Ermanaric’s Kingdom

      The Huns at the Danube

        Visigoths and Huns Cooperate

        The Huns Threaten Pannonia

        Hunnic Pressure on the Lower Danube

        Hunnic Horsemen Ride to Gaul

      The Invasion of Asia

        The Sources

        The Course of the War



        New Raids into Thrace

        The Huns Help Aelius and Lose Pannonia

      Octar and Ruga


        The Huns Threaten the West

        The War in the Balkans

      Attila’s Kingdom

      The Huns In Italy

        The First Phase of the War

        In the Po Valley


      Collapse and Aftermath

        Revocatio Pannoniarum

        The Nedao River

      The First Gotho-Hunnic War


      The Second Gotho-Hunnic War (463/4-466)

      The End

    III. Economy


      Hunnic Agriculture?


      Income in Gold




    IV. Society



    V. Warfare

      General Characteristics



        Horse Marks




        Coexistence of Various Types

        The Skill Required


        Sasanian Bows


        The Sword of Altlussheim


      The Lasso


        Body Armor



      Huns in the Roman Army

    VI. Religion

      The Huns and Christianity

      Seers and Shamans

      Divine Kingship?


      The Sacred Sword

      Masks and Amulets

    VII. Art

      Gold Diadems













      Personal Ornaments

        Gold Plaques on Garments



    VIII. Race

      The Hsiung-nu

      Europoids in East Asia

    IX. Language

      Speculations about the Language of the Huns



      Germanized and Germanic Names








      Iranian Names






        Xxoga^ and Dm)VT}<;



      Turkish Names









      Names of Undetermined Origin

































      Tribal Names




        Xa.fi toot


    X. Early Huns in Eastern Europe

    XI. Appendixes

      1. The Chronicle of 452

      2. Armenian Sources

      3. Figures in Olympiodorus

      4. The Alleged Loss of Pannonia Prima in 395

        I now come to the frail fortunes of human life, and my soul shudders to recount the downfall of our age. For twenty years and more the blood of Romans has every day been shed between Constantinople and the Julian Alps. Scythia, Thrace, Macedonia, Thessaly, Dardania, Dacia, Epirus, Dalmatia and all the provinces of Pannonia have been sacked, pillaged, and plundered by the Goths and Sarmatians, Quadi and Alans, Huns and Vandals and Marcomanni.

      5. Religious Motifs in Hunnic Art?

    XII. Background: The Roman Empire at the Time of the Hunnic Invasions


        I. Abbreviations

        II. Classical and Medieval Register of Cited Names and Titles

        III. Sources


    [Back Cover]

[Front Matter]



Studies in Their History and Culture


Edited by Max Knight

Few persons know more about the Huns than their reputation as savage horsemen who flourished at the beginning of the Middle Ages and the name of one of their leaders, Attila. They appeared in Europe from “somewhere in the East,” terrorized the later Roman Empire and the Germanic tribes, caused the greatest upheaval that the Mediterranean world had ever seen—the Great Migrations — and vanished. Illiterate, they left no written records; such literary evidence of them as exists is secondary, scattered in the writings of contemporary and later reporters, fragmentary, biased, and unreliable. Their sole tangible relics are huge cauldrons and graves, some of which contain armor, equestrian gear, and ornaments.

Who were the Huns? How did they live? Professor Maenchen-Helfen dedicated much of his life to seeking answers to these questions. With pertinacity, passion, scepticism, and unsurpassed scholarship he pieced together evidence from remote sources in Asia, Russia, and Europe; categorized and interpreted it; and lived the absorbing detective story presented in this volume. He spent many years and extensive resources in exploring the mystery of the Huns and in exploding popular myths about them. He investigated the century-old hypothesis that the Huns originated in the obscure borderlands of China, whence in the course of several generations they migrated westward as far as Central Europe. In his quest for information about them. Professor Maenchen-Helfcn sifted written sources in the original classical, Slavic, and Asian languages. He also devoted years of travel to visiting archeological digs from Hungary to Afghanistan, from Nepal to Mongolia, and from the Caucasus to the Great Wall. His subject fascinated him, and this book conveys something of this fascination.

Intrigued by each report of the discovery of a new grave somewhere, the author could not bring himself to complete the work. At his death in 1969 his study was filled from floor to ceiling with binders, folders, and photographic negatives. Some chapters were in final form; many were not. The editor had to choose among various drafts, to decipher and organize, and to enlist the assistance of the author’s scholar-friends in filling gaps in the material.

The work, intended primarily for scholars, lacks an introduction. However, Paul J. Alexander, Professor of History and Comparative Literature at the University of California, Berkeley, has provided a historical background designed for the general interested reader. The book is richly illustrated with rare photographs collected by the author, and includes a specially prepared map of the Roman Empire at the time of the Huns.

“Professor Maenchen’s knowledge of Central Asian peoples, including the Hsiung-nu and Huns, was absolutely unparalleled; his learning in several disciplines, including philology (with language knowledge ranging as widely as Greek and Chinese), was unique. His study includes the latest Soviet discoveries. Books of comparable scope have been attempted, but none of them has been buttressed by a comparable erudition; none has been infused with comparable understanding of the panorama of Eurasia; none has achieved in comparable degree the placing of the actors on this vast stage and their correct perspectivesand proportions.”

[Title Page]



Studies in Their History and Culture



University of California Press / Berkeley / Los Angeles / London / 19 75


University of California Press

Berkeley and Los Angeles, California

University of California Press, Ltd.

London, England

Copyright © 1973, by

The Regents of the University of California

Library of Congress Catalog Card Number: 79–94985

International Standard Book Number : 520-01596-7

Designed by James Mennick

Printed in the United States of America

List of Illustrations


1 A horse with a “hooked” head and bushy tail represented on 205 a bronze plaque from the Ordos region. From Egami 1948, pl. 4.

2 Grave stela from Theodosia in the Crimea with the representation of the deceased mounted on a horse marked with a Sarmatian 212 tamga, first to third centuries a.d. From Solomonik 1957, fig. 1.

3 Two-wheeled cart represented on a bronze plaque from the Wu-216 huan cemetery at Hsi-ch’a-kou. From Sun Shou-tao 1960, fig. 17.

4 Bronze plaque from Sui-yiian with the representation of a man 217 holding a sword with a ring handle before a cart drawn by three horses. From Rostovtsev 1929, pl. XI, 56.

5A Miniature painting from the Radziwil manuscript showing the wa-218 gons of the Kumans. From Pletneva 1958, fig. 25.

5B Miniature painting from the Radziwil manuscript showing human 218 heads in tents mounted on carts. From Pletneva 1958, fig. 26.

6 Ceramic toy from Kerch showing a wagon of Late Sarmatian type. 219 From Narysy starodav’noi istorii Ukrains’koi RSR 1957, 237.

7 Detail of a Sasanian-type silver plate from a private collection. 229 Detail from Ghirshman 1962, fig. 314.

8 Detail of a Sasanian silver plate from Sari, Archaeological Museum, 230 Teheran. Detail from Ghirshman 1962, pl. 248.

9 Silver plate from Kulagysh in the Hermitage Museum, Leningrad. 232 From SPA, pl. 217. “

10A Scabbard tip of a sword from Altlussheim near Mainz. From J. 233 Werner 1956, pl. 58:4.

10B Detail of the sword from Altlussheim near Mainz. From J. Werner 234 1956, pl. 38 A.

11 Stone relief from Palmyra, datable to the third century a.d. 235 Ghirshman 1962, pl. 91.

12 Agate sword guard from Chersonese, third century a.d. From 237 Khersonesskii sbornik, 1927, fig. 21.

12A Bronze pendant said to have been found in a grave at Barnaul, 243 Altai region, showing a man in scale armor and conical hat with

an hour-glass-shaped quiver, datable to the fourth century a.d.

From Aspelin 1877, no. 327.

12B Two horsemen in scale armor shown in gold pendants from western 244 Siberia. From Kondakov and Tolstoi, 3, fig. 49.

12C The representation of a Sarmatian member of the Roxolani tribe in 246 a detail of the marble relief from Trajan’s Column, in the Forum of Trajan, Rome. Datable to the second decade of the second century a.d. Photos courtesy Deutsches archaologisches Institut, Rome.

13 Mask-like human heads stamped on gold sheet from a Hunnic burial 281 at Pokrovsk-Voskhod. From Sinitsyn 1936, fig. 4.

14 Mask-like human heads stamped on silver sheet on a bronze phalera 281 from kurgan 17, Pokrovsk. From Minaeya 1927, pl. 2:11.

15 The representation of the head of a Scythian in clay from Transcau-283 casia. Photo courtesy State Historical Museum, Moscow.

16 Bronze mountings from a wooden casket from Intercisa on the Da-284 nube. From Paulovics, AE, 1940.

17 Flat bronze amulet in the shape of an ithyphallic human figure of 286 Sarmatian type. (Source not indicated in the manuscript. — Ed.)

18 Sandstone pillar in the shape of a human head from kurgan at Tri Brata near Elista in the Kalmuk steppe. (Height 1 m.) From Sinitsyn 1956b, fig. 11.

19 Chalk eidola from an Alanic grave at Baital Chapkan in Cherkessia, fifth century a.d. From Minaeva 1956, fig. 12.

20 Chalk figure from a Late Sarmatian grave in Foc§ani, Rumania. (Height ca. 12 cm.) From Morintz 1959, fig. 7.

21 Stone slab at Zadzrost’, near Ternopol’, former eastern Galicia, marked with a Sarmatian tamga. (Height 5.5 m.) From Drachuk, SA 2, 1967, fig. 1.

22 Fragment of a gold plaque from Kargaly, Uzun-Agach, near Alma Ata, Kazakhstan. (About 35 cm. long.) Photo courtesy Akademiia Nauk Kazakhskol SSR.

23 Hunnic diadem of gold sheet, originally mounted on a bronze plaque, decorated with garnets and red glass, from Csorna, western Hungary. (Originally about 29 cm. long, 4 cm. wide.) From Archao-logische Funde in Ungarn, 291.

24A-C Hunnic diadem of gold sheet over bronze plaques decorated 301 with green glass and flat almandines, from Kerch. Photos courtesy Rheinisches Museum, Bildarchiv, Cologne.

25 Hunnic diadem of thin bronze sheet over bronze plaques set with convex glass from Shipovo, west of Uralsk, northwestern Kazakhstan.

From J. Werner 1956, pl. 6:8.

26 Hunnic diadem of gold sheet over bronze plaques set with convex 302 almandines from Dehler on the Berezovka, near Pokrovsk, lower Volga region. From Ebert, RV 13, “Siidrussland,” pl. RV 41 :a.

27 Hunnic diadem of gold sheet over bronze plaques (now lost) set with 303 convex almandines, from Tiligul, in the Romisch-Germanisches Zentralmuseum, Mainz. From J. Werner 1956, pl. 29:8.

28 Bronze circlet covered with gold sheet and decorated with conical 304 “bells” suspended on bronze hooks, from Kara Agach, south of Akmolinsk, central Kazakhstan. (Circumference 49 cm., width ca. 4 cm.) From J. Werner 1956, pl. 31:2.

29A Terminal of a gold torque in the shape of a dragon, decorated with 304 granulation and cloisonnd garnets, amber, and mother-of-pearl. From Kara-Agach, south of Akmolinsk, central Kazakhstan. From IAK 16, 1905, p, 34, fig. 2.

29B Gold earrings from Kara-Agach, central Kazakhstan. From IAK 305 16, 1905, fig, 3:a-b.

30 Silver earring decorated with almandines and garnets from kurgan 36, SW group, near Pokrovsk. From Sinitsyn 1936, fig. 10.

31 Fold earring from Kalagya, Caucasian Albania. From Trever 1959, 305 167, fig. 18.

32 Fragment of a bronze lug of a cauldron from BeneSov, near Opava 307 (Troppau), Czechoslovakia. (Height 29 cm., width 22 cm., thickness 1 cm.) From Altschlesien 9, 1940, pl. 14.

33 Hunnic bronze cauldron Jedrzychowice (Hockricht), Upper 308 Silesia, Poland. (Height 55 cm.) From J. Werner 1956, pl. 27:10.

34 Hunnic bronze cauldron found at the foot of a burial mound at Tortel, 309 Hungary. (Height 89 cm., diam. 50 cm.) From Archaologische Funde in Ungarn, 293.

35 Hunnic bronze cauldron found in a peat bog at Kurdcsibrdk, in the 310 Kapos River valley, Hungary. (Height 52 cm., diam. 33 cm., thickness of wall O.8 cm., weight 16 kg.) From Fettich 1940, pl. 11.

36 Hunnic bronze cauldron from B^ntapuszta, near V^rpalota, Hungary. 311 From Takats, AOH, 1959, fig, t.

37 Fragment of a bronze cauldron from Dunaujv^ros (Intercisa), 311 Hungary. From Alfoldi 1932, fig. 6.

38 Hunnic bronze cauldron from a lake, Desa, Oltenia region, Rumania. 312 (Height 54.1 cm., diam. 29.6 cm.) From Nestor and Nicolaescu-Plopsor 1937, pls. 3a-3b.

39 Fragment of a bronze lug from a lake, HotSrani, Oltenia region, 313 Rumania. (Height 16.2 cm., width 19.7 cm.) From Nestor and Nicolaescu-Plopsor 1937, pl. 39:1.

40 Fragment of a bronze lug probably from western Oltenia, Rumania. 313 (Height 8.4 cm.) From Nestor and Nicolaescu-Plopsor 1937, pl. 39:2.

41 Fragment of a bronze lug found near the eastern shore of Lake Motistea, from Bosneagu, Rumania. (Height 18 cm.) From Mitrea 1961, 1–2?

42 Fragments of a lug and walls of a bronze cauldron from Celei, 314 Muntenia, Rumania. From Takdts 1955, fig. 13:a-d.

43 Hunnic bronze cauldron from Shestachi, Moldavian SSR. From 315 Polevoi, Istoriia Moldavskoi SSR, pl. 53.

44 Bronze cauldron from Solikamsk, Perm region, USSR. (Height 9 cm.) 316 From Alfoldi 1932, fig. 5.

45 Bronze cauldron found in thesand near the Osoka brook, Ul’yanovsk 317 region, USSR. (Height 53.2 cm., diam. 31.2 cm., weight 17.7kg.) From Polivanova, Trudy VII AS 1, 39, pl. 1.

46 Bronze cauldron from Verkhnii Konets, Komi ASSR. From Hampel, 318 Ethnologische Mittheilungen aus Ungarn 1897, 14, fig. 1.

Bronze cauldron from Ivanovka, gubernie Ekaterinoslav, USSR. From Fettich 1953, pl. 36:4.

Bronze cauldron found near Lake Teletskoe, in the High Altai, now in the State Historical Museum, Moscow. (Height 27 cm.,diam. 25–27 cm.) Photo courtesy State Historical Museum, Moscow. Fragment of a bronze lug from Narindzhan-baba, Kara-Kalpak ASSR. From Tolstov 1948, fig. 74a.

Fragment of a bronze lug, allegedly found “on the Catalaunian battlefield.” (Height 12 cm., width 18 cm.) From Takdts 1955, fig. l:a-b.

Bronze cauldron from Borovoe, northern Kazakhstan. From Bernshtam 1951a, fig. 12.

The representation of a cauldron in a detail of a rock picture from Pis annaya Gora in the Minusinsk area. From Appelgren-Kivalo, fig. 85. Representation of cauldrons in a rock picture from Bol’shaya Boyarskaya pisanitsa, Minusinsk area. From Ddvlet, SA 3, 1965, fig. 6. Bronze cauldron of a type associated with Hsiung-nu graves at Noin Ula and the Kiran River. From Umehara 1960, p. 37. Ceramic vessel from the Gold Bell Tomb at Kyongju, Korea, showing the manner in which cauldrons were transported by nomads. From Government General Museum of Chosen 1933, Museum Exhibits Illustrated V.

Clay copy of a Hunnic cauldron of the Verkhnii Konets type (see above, fig. 46), from the “Big House,” Altyn-Asar, Kazakhstan. (Height 40 .cm.) From Levina 1966, fig. 7:37–38.

Chinese mirror of the Han period found in burial 19, on the Torgun River, lower Volga region. From Ebert, R V, «Siidrussland,»pl. 40: c:b. A Sarmatian bronze disc in the shape of a pendant-mirror, of a type found in the steppes betweenVolga and lower Danube, from the first century b.c. to the fourth century a.d. From Sinitsyn 1960, fig. 18:1.

Bronze mirror of a type similar to that shown on fig. 58, but provided with a tang that was presumably fitted into a handle. From Gushchina, SA, 2, 1962, fig. 2:5.

Bronze pendant-mirror from the cemetery at Susly, former German Volga Republic. From Rau, Hugelgraber, 9, fig. la.

Bronze pendant-mirror from the cemetery at Susly, former German Volga Republic. From Rykov 1925, 68.

Bronze pendant-mirror from Alt-Weimar, kurgan D12. From Rau, Ausgrabungen, 30, fig. 22b.

Bronze pendant-mirror from kurgan 40 in Berezhnovka, lower Eruslan, left tributary of the Volga. From Khazanov 1963, fig. 4:9. Bronze pendant-mirror from kurgan 23, in the “Tri Brata” cemetery, near Elista, Kalmuk ASSR. From Khazanov 1963, fig. 4:8. Bronze pendant-mirror from the lower Volga region. From Khazanov 1963, fig. 4:6.

Bronze pendant-mirror from a catacomb burial at Alkhaste, northwestern Caucasus. From Vinogradov 1963, fig. 27.

An imitation of a Chinese TLV mirror from Lou-Ian. From Umehara, O bei, 39, fig. 7.

68 Small bronze mirror with simplified decoration from Lo-yang. 347 From Lo-yang ching 1959, 80.

69 Small bronze mirror with simplified decoration from Lo-yang. 347 From Lo-yang ching 1959, 82.

70 Bronze mirror from Mozhary, Volgograd region, now in the Hermitage Museum, Leningrad, datable to about a.d. 200. (Diam. 7.4 cm.) From Umehara 1938, 55.

71 Bronze mirror from Kosino in Slovakia. From Eisner, Slovensko 349 v praveku 1933, fig. 2:7.

72 Bronze pendant-mirrors from the Dnieper and Volga regions. From 350 Solomonik 1959, fig. 6.

73 Sarmatian imitation of a Chinese mirror (cf. the example from 351 Lo-yang, above, fig. 69), from Norka, lower Volga region. From Berkhin 1961, fig. 2:2.

74 Small bronze plaque showing a horseman with prominent cheekbones 370 and full beard, from Troitskovavsk in Transbaikalia. From Petri, Dalekoe proshloe Pribaikal’ia 1928, fig. 39.

75 Bronze plaque from the Ordos region, showing a man of Europoid 371 stock with wide open eyes and moustache. British Museum. Photo G. Azarpay.


Few scholars would care to risk their reputation in taking on the monumental task of straightening out misconceptions about the Huns, and incidentally about the many peoples related to them, allied with them, or confused with them. At the foundation there are philological problems of mindboggling proportions in languages ranging from Greek to Chinese; above that, an easy but solidly professional familiarity with primary sources for the history of both Eastern and Western civilizations in many periods is required; finally, a balanced imagination and a prudent sense of proportion are needed to cope with the improbabilities, contradictions, and prejudices prevailing in this field of study. The late Professor Otto Maenchen-Helfen worked on this immense field of research for many years, and at his death in 1969 left an unfinished manuscript. This is the source of the present book.

Maenchen-Helfen differed from other historians of Eurasia in his unique competence in philology, archaeology, and the history of art. The range of his interests is apparent from a glance at his publications, extending in subject from “Das Marchen von der Schwanenjungfrau in Japan” to “Le Cicogne di Aquileia,” and from “Manichaeans in Siberia” to “Germanic and Hunnic Names of Iranian Origin.” He did not need to guess the identities of tribes, populations, or cities. He knew the primary texts, whether in Greek or Russian or Persian or Chinese. This linguistic ability is particularly necessary in the study of the Huns and their nomadic cognates, since the name “Hun” has been applied to many peoples of different ethnic character, including Ostrogoths, Magyars, and Seljuks. Even ancient nomadic people north of China, the Hsiung-nu, not related to any of these, were called “Hun” by their Sogdian neighbors. Maenchen-Helfen knew the Chinese sources that tell of the Hsiung-nu, and thus could evaluate the relationship of these sources to European sources of Hunnic history.

His exceptional philological competence also enabled him to treat as human beings the men whose lives underlie the dusty textual fragments that allude to them, and to describe their economy, social stratifications, modes of transportation and warfare, religions, folklore, and art. He could create a reliable account of the precursors of the Turks and Mongols, free of the usual Western prejudice and linguistic limitations.

Another special competence was his expertise in the history of Asian art, a subject that he taught for many years. He was familiar with the newest archaeological discoveries and knew how to correlate them with the available but often obscure philological evidence.

To define distinctive traits in the art of a people as elusive as the Huns requires familiarity with the disjointed array of archaeological materials from the Eurasian steppe and the ability to separate materials about the Huns from a comparable array of materials from neighboring civilizations. To cite only one example of his success in coping with such thorny problems, Maenchen-Helfen’s description of technical and stylistic consistencies among metal articles from Hunnic tombs in widely separated localities dispels the myth of supposed Hunnic ignorance of metal-working skills.

Archaeological evidence also plays a critical role in the determination of the origin of the Huns and their geographical distribution in ancient and early medieval times, as well as the extent of Hunnic penetration into eastern Europe and their point of entry into the Hungarian plain. Maenchen-Helfen saw clearly how to interpret the data from graves and garbage heaps to yield hypotheses about the movements of peoples. “He believed in the spade, but his tool was the pen,” he once said about another scholar — a characterization that perfectly fits Maenchen-Helfen himself. Burial practices of the Huns and their associates indicate that Hunnic weapons generally originated in the east and were transmitted westward, while the distribution of loop mirrors found in association with artificially deformed skulls — a Hunnic practice — gives proof of Hunnic penetration into Hungary from the northeast. (An unpublished find of a sword of the Altlussheim type recently discovered at Barnaul in the Altai region, east Kazakhstan SSR, now in the Hermitage Museum, is a forceful argument in favor of Maenchen-Helfen’s assumption about the eastern connections of this weapon. See A. Urmanskii, “Sovremennik groznogo Attily,” A Itai 4 [23], Barnaul 1962, pp. TS- 93.) His findings define and bring to life the civilization of one of the most shadowy peoples of early medieval times.

Maenchen-Helfen’s account opens in medias res, with a tribute to that admirable Roman historian Ammianus Marcellinus, whose view of the Hunnic incursions was, despite his prejudices, in some respects clearer than that of Western historians. Abrupt as this beginning may seem, the author perhaps intended the final version of his book to begin with such a striking evaluation of a basic text. In so doing, he underlined the necessity for sharp and well-reasoned criticism of the sources of the history of the Huns. From the beginning these people were denigrated and “demonized” (to use his own term) by European chroniclers and dismissed as avatars of the eternal but faceless barbarian hordes from the east, against whom vigilance was always necessary, but whose precise identity was of little importance. The bulk of the book discusses the history and civilization of the “Huns proper,” those so familiar — and yet so unfamiliar — to Europeans. (Here we use the term “civilization” purposefully, since reports of this folk have tended to treat them as mere barbaric destroying agents — “vandals” spilling blood across the remnants of the declining Roman Empire. Maenchen-Helfen saw them with a clearer vision.)

The style is characteristically dense with realia. Maenchen-Helfen had no need to indulge in generalizations (read “unfounded guesses”). But he was not absorbed in details to the exclusion of a panoramic view. He saw, and presents to us here, the epic character of the great drama that took place on the Eurasian stage early in our era, the clash of armies and the interaction of civilizations. The book is a standard treatise not likely to be superseded in the predictable future.

Guitty Azarpay

Peter A. Boodberg

Edward H. Schafer

Editor’s Note

In early January 1969 Professor Otto Maenchen-Helfen brought a beautifully typed manuscript from the Central Stenographic Bureau of the university to the University of California Press. It seemed to represent the final result of his monumental study of the Huns, to which he had devoted many years of research and travel. A few days later, on January 29, he died. In the memorial speeches at the Faculty Club in Berkeley, several friends mentioned that he had truly completed his lifework, and that his manuscript was ready to go to press.

The impression that the delivered manuscript pages constituted the complete manuscript turned out to be erroneous. Mr. Maenchen had brought only the first of presumably two batches of manuscript. The chapters representing that second batch were not in final form at the time of his death, the bibliography was missing, footnotes were indicated but the sources not stated, an introduction and a complete preface were lacking, the illustrations were scattered in boxes and desk drawers and not identified. There was no table of contents, and the chapters were not numbered; although some groupings of chapters are suggested in the extant part of the author’s preface, it was not clear in what order he intended to arrange his work.

On Mrs. Maenchen’s suggestion I searched the author’s study and eventually found a tentative draft of a contents page. It was of unknown age, and contained revisions and emendations that required interpretation. On the basis of this precious page, the “Rosetta Stone of the manuscript,” the work was organized.

Several chapters mentioned in this page were not in final form. But three-ring folders in the author’s study, neatly filed on shelves, bore the names of most missing chapter headings. The contents of these folders were in various stages of completion. Those that appeared to be more or less finished except for final editing were incorporated into the manuscript; also sections which, although not representing complete chapters but apparently in final form, were included and placed where they seemed to fit mostly logically. In several instances, different drafts of the same subject were found, and it was necessary to decide which was the most recent one. Occasionally, also, only carbon copies of apparently finished sections were in the folders.

Errors in judgment in these editorial and compiling activities cannot be ruled out, but wherever doubts existed about the preferred version or the placement of a fragment the material was excluded. Many notes, isolated pages, and drafts (frequently written by hand, with various kinds of emendations) remain in the author’s study, including undoubtedly valuable research results.

In retyping the parts of the manuscript that existed only in draft form with many emendations and hand-written corrections, every effort was made not to introduce errors, such as misspellings of foreign words, especially in the notes and bibliography. For errors that undoubtedly slipped in nevertheless, the author is not responsible.

Although the work addresses itself to specialists, it is of interest to a broader range of educated readers who cannot, however, be expected to be familiar with some of the events, persons, institutions, and sources the author takes for granted. For these readers Professor Paul Alexander has provided an introduction; in deference to the author it was placed as “background” at the end of the book, but it may usefully be read first, as a preparation for the text.

The editorial preparation of the manuscript required the help of an unusually large number of persons, reflecting the wide range of the author’s competence. The Russian references were checked by the author’s friend, the late Professor Peter A. Boodberg, who delivered the corrected pages just a few days before his death in the summer of 1972. The Chinese references were checked or supplied by Professor Edward H. Schafer, also a friend of the author. The Latin and Greek passages were translated by Professor J. K. Anderson and Dr. Emmy Sachs; Mr. Anderson also faithfully filled lacunae in the footnotes and unscrambled mixups resulting from duplicated or omitted footnote numbers. Professors Talat Tekin and Hamid Algar checked and interpreted Turkish references. Professor Joachim Werner of Munich counseled on the Altlussheim sword. Questions about Gothic, Iranian, Hungarian, Japanese, and Ukrainian references or about historical (ancient and medieval) and many other aspects of the text that needed interpretation were answered by a long list of scholars contributing their services to the cause.

Miss Guitty Azarpay (to whom the author used to refer fondly as his favorite student) selected and painstakingly identified the illustrations. She also verified references with angelic patience.

The formidable task of compiling a bibliography on the basis of an incomplete set of cards and of the text itself was performed by Mrs. Jane Fontenrose Cajina. The author’s working cards, assembled over many years, were not yet typed in uniform style, many entries were missing, and many lacked essential information. For Russian transcriptions in the bibliography and bibliographical footnotes (but not in the text), the Library of Congress system was used.

The map was drawn by Mrs. Virginia Herrick under the supervision of Professor J. K. Anderson. The index was prepared by Mrs. Gladys Castor.

The editor is indebted to all these many competent and sympathetic helpers; clearly, without their devotion the conversion of the Maenchen papers into the present volume would not have been possible.

Max Knight

Fragments from the Author’s Preface

[Among the author’s papers were several fragments, partly written in pencil, bearing the notation “for the preface,” and evidently intended to be worked into a final draft. He may have wished to say more; all we found is presented below.—Ed.]

The author of the present volume, in his early seventies, may make use of the privilege, usually granted to men in the prime of their senility, to say a few words about himself, in this case the sources of his interest in the Huns. All my life I have been fascinated by the problems of the frontier. As a boy I dug Roman copper coins along the remnants of the earthen walls that, as late as the seventeenth century, protected Vienna, my native town, from the East. Two blocks from the house in which I was born there still stood in my youth a house above whose gate a Turkish stone cannon ball from the siege of 1529 was immured. My grandfather spent a year in jail for fighting in 1848 with the revolutionaries against the Croatian mercenaries of the Habsburgs. My doctoral dissertation dealt with the “barbarian” elements in Han lore. In 1929 I lived for months in the tents of Turkishspeaking nomads in northwestern Mongolia, where the clash between “higher civilization,” represented by Tibetan Lamaism, and the “primitive” beliefs of the Turks was strikingly visible. In Kashmir, at Harwan, I marveled at the artificially deformed skulls on the stamped tiles of Kushan times, those skulls that had impressed me so much when I first saw them in the museum in Vienna and that I had measured as a student. In Nepal I had another chance to see the merging of different civilizations in a borderland. I spent many days in the museum at Minusinsk in southern Siberia studying the “Scythian” bronze plaques and cauldrons. In Kabul I stood in awe before the inscription from Surkh Kotal: it brought back to me the problems of the barbarians at China’s border about which I had written a good deal in previous years. Attila and his avatars have been haunting me as far back as I can recall.

In the history of the Western world the eighty years of Hun power were an episode. The Fathers assembled in council at Chalcedon showed a sublime indifference to the barbarian horsemen who, only a hundred miles away, were ravaging Thrace. They were right. A few years later, the head of Attila’s son was carried in triumphal procession through the main street of Constantinople.

Some authors have felt that they had to justify their studies of the Huns by speculating on their role in the transition from late antiquity to the Middle Ages. Without the Huns, it has been maintained, Gaul, Spain, and Africa would not, or not so soon, have fallen to the Germans. The mere existence of the Huns in eastern Central Europe is said to have retarded the feudalization of Byzantium. This may or may not be true. But if a historical phenomenon were worth our attention only if it shaped what came after it, the Mayans and Aztecs, the Vandals in Africa, the Burgundians, the Albigenses, and the crusaders’ kingdoms in Greece and Syria would have to be wiped off the table of Clio. It is doubtful that Attila “made history.” The Huns “perished like the Avars” — “sginuli kak obry,” as the old Russian chroniclers used to say when they wrote about a people that had disappeared forever.

It seems strange, therefore, that the Huns, even after fifteen hundred years, can stir up so much emotion. Pious souls still shudder when they think of Attila, the Scourge of God; and in their daydreams German university professors trot behind Hegel’s Weltgeist zu Pferde. They can be passed over. But some Turks and Hungarians are still singing loud paeans in praise of their great ancestor, pacifier of the world, and Gandhi all in one. The most passionate Hun fighters, however, are the Soviet historians. They curse the Huns as if they had ridden, looting and killing, through the Ukraine only the other day; some scholars in Kiev cannot get over the brutal destruction of the “first flowering of Slavic civilization.”

The same fierce hatred burned in Ammianus Marcellinus. He and the other writers of the fourth and fifth centuries depicted the Huns as the savage monsters which we still see today. Hatred and fear distorted the picture of the Huns from the moment they appeared on the lower Danube. Unless this tendentiousness is fully understood — and it rarely is — the literary evidence is bound to be misread. The present study begins, therefore, with its reexamination.

The following chapters, dealing with the political history of the Huns, are not a narrative. The story of Attila’s raids into Gaul and Italy need not be told once more; it can be found in any standard history of the declining Roman Empire, knowledge of which, at least in its outlines, is here taken for granted. However, many problems were not even touched on and many mistakes were made by Bury, Seeck, and Stein. This statement does not reflect on the stature of these eminent scholars, for the Huns were on the periphery of their interests. But such deficiencies are true also for books which give the Huns more room, and even for monographs. The first forty or fifty years of Hun history are treated in a cursory manner. The sources are certainly scanty though not as scanty as one might believe; for the invasion of Asia in 395, for instance, the Syriac sources flow copiously. Some of the questions that the reign of Attila poses will forever remain unanswered. Others, however, are answered by the sources, provided one looks, as I have, for sources outside the literature that has been the stock of Hunnic studies since Gibbon and Le Nain de Tillemont. The discussions of chronology may at times tax the patience of the reader, but that cannot be helped. Eunapius, who in his Historical Notes also wrote about the Huns, once asked what bearing on the true subject of history inheres in the knowledge that the battle of Salamis was won by the Hellenes at the rising of the Dog Star. Eunapius has his disciples in our days also, and perhaps more of them than ever. One can only hope that we will be spared a historian who does not care whether Pearl Harbor came before or after the invasion of Normandy because “in a higher sense” it does not matter.

The second part of the present book consists if monographs on the economy, society, warfare, art, and religion of the Huns. What distinguishes these studies from previous treatments is the extensive use of archaeological material. In his Attila and the Huns Thompson refuses to take cognizance of it, and the little to which Altheim refers in Geschichte dec Hunnen he knows at second hand. The material, scattered through Russian, Ukrainian, Rumanian, Hungarian, Chinese, Japanese, and latterly also Mongolian publications, is enormous. In recent years archaeological research has been progressing at such speed that I had to modify my views repeatedly while I was working on these studies. Werner’s monumental book on the archaeology of Attila’s empire, published in 1956, is already obsolete in some parts. I expect, and hope, that the same will be true of my own studies ten years from now.

Although aware of the dangers in looking for parallels between the Huns and former and later nomads of the Eurasian steppes, I confess that my views are to a certain, I hope not undue, degree influenced by my experiences with the Tuvans in northwestern Mongolia, among whom I spent the summer of 1929. They are, or were at that time, the most primitive Turkish-speaking people at the borders of the Gobi.

I possibly will be criticized for paying too little attention to what Robert Gobi calls the Iranian Huns: Kidara, White Huns, Hepthalites, and Hunas. In discussing the name “Hun” I could not help speculating on their names. But this was as far as I dared go. The literature on these tribes or peoples is enormous. They stand in the center of Altheim’s Geschichte der Hunnen, although he practically ignores the numismatic and Chinese evidence, on which Enoki has been working for so many years. Gobi’s Dokumente zur Geschichte der iranischen Hunnen in Baktrien und Indien is the most thorough study of their coins and seals and, on this basis, of their political history. And yet, there remain problems to whose solution I could not make a meaningful contribution. I have neither the linguistic nor the paleographic knowledge to judge the correctness of the various, often entirely different, readings of the coin legends. But even if someday scholars wrestling with this recalcitrant material do come to an agreement, the result will be relatively modest. The Huna Mihirakula and Toramana will remain mere names. No settlement, no grave, not so much as a dagger or a piece of metal exists that could be ascribed to them or any other Iranian Huns. Until the scanty and contradictory descriptions of their life can be substantially supplemented by finds, the student of the Attilanic Huns will thankfully take cognizance of what the students of the so-called Iranian Huns can offer him; but there is little he can use for his research. A recently discovered wall painting in Afrosiab, the ancient Samarkand, seems to show the first light in the darkness. The future of the Hephthalite studies lies in the hands of the Soviet and, it is hoped, the Chinese archaeologists. ’Ev ftvOcp yaq aAijOeia.

I am aware that some chapters are not easy reading. For example, the one on the Huns after Attila’s death draws attention to events seemingly not worth knowing, to men who were mere shadows; it jumps from Germanic sagas to ecclesiastical troubles in Alexandria, from the Iranian names of obscure chieftains to an earthquake in Hungary, from priests of Isis in Nubia to Middle Street in Constantinople. I will not apologize. Some readers surely will find the putting together of the scattered pieces as fascinating as I did, and I frivolously confess to an artistic hedonism which to me is not the least stimulus for my preoccupation with the Dark Ages. On a higher level, to pacify those who, with a bad conscience, justify what they are doing — Historical Research with capital letters — may I point out that I fail to see why the history of, say, Baja California is more respectable than, say, that of the Huns in the Balkans in the 460’s. Sub specie aeternitatis, both dwindle into nothingness.

Anatole France, in his Opinions of Jerome Coignard, once told the wonderful story of the young Persian prince Zemire, who ordered his scholars to write the history of mankind, so that he would make fewer errors as a monarch enlightened by past experience. After twenty years, the wise men appeared before the prince, king by then, followed by a caravan composed of twelve camels each bearing 500 volumes. The king asked them for a shorter version, and they returned after another twenty years with three camel loads, and, when again rejected by the king, after ten more years with a single elephant load. After yet five further years a scholar appeared with a single big book carried by a donkey. The king was on his death bed and sighed, “I shall die without knowing the history of mankind. Abridge, abridge I” “Sire,” replied the scholar, “I will sum it up for you in three words: They were born, they suffered, they died I”

In his way, the king, who did not want to hear it all, was right. But as long as men, stupidly perhaps, want to know “how it was,” there may be a place for studies like the present one. Dixi et salvaui animam meam ....

0. M.-H.

Author’s Acknowledgments (Fragments)

[The author left some pencil jottings of names on several slips of paper under headings indicating that he wished to acknowledge them in the preface. Some are not legible, others lack initials. They are consolidated here, initials added when known, and the spelling of unidentified names as close as the handwriting permitted. Within the various countries, the order is random; the list of country names includes France, Rumania, Taiwan, and Korea, but the names of the scholars whom the author undoubtedly intended to acknowledge under these headings are lacking. The fragments include acknowledgments of help received from the East Asiatic Library and the Interlibrary Borrowing Service of the University of California. The notes must have been written at various times throughout the years, and it is obvious that the list is not complete.—Ed.]

In Austria: R. Gobi, Hancar.

In England: Sir Ellis Minns, E. G. Pulleyblank.

In Germany: J. Werner, K. Jettmar, Tauslin.

In Hungary: Z. Takats, L. Ligeti, D. Czallany, Gy. Moravcik, K. Csegledy, E. Liptak.

In Italy: M. Bussagli, P. Daffina, L. Petech.

In Japan: Namio Egami, Enoki, Ushida.

In Soviet Union: K. V. Golenko, V. V. Ginsburg, E. Lubo-Lesnichenko, C. Trever, A. Mantsevich, M. P. Griaznov, L. P. Kyzlasov, I. A. Zadneprovsky, I. Kozhomberdiev, M. Saratov, A. P. Okladnikov, S. S. Sorokin, B. A. Litvinsky, I. V. Sinitsyn, Gumilev, Belenitsky, Stavisky.

In Sweden: B. Karlgren.

In Switzerland: K. Gerhart, I. Hubschmid.

In United States: P. Boodberg, E. Schafer, R. Henning, A. Alfoldi, R. N. Frye, E. Kantorowicz, L. Olschki, K. H. Menges, N. Poppe, I. Sevcenko.

I. The Literary Evidence

The chapter on the Huns written by the Roman historian Ammianus Marcellinus (330–400 a.d.) is an invaluable document.{1} Coming from the pen of “the greatest literary genius which the world has seen between Tacitus and Dante,”[1] it is also a stylistic masterpiece. Ammianus’ superiority over the other writers of his time who could not help mentioning the Huns becomes evident from their statements about the first appearance of the savage hordes in the northern Balkan provinces. They tell us in a few scanty words that the Goths were driven from their sites by the Huns; some add the story of a doe which led the Huns across the Cimmerian Bosporus. And this is all. They did not care to explore the causes of the catastrophe of Adrianople, that terrible afternoon of August 9, 378, when the Goths annihilated two-thirds of the Roman army, else they would have found that “the seed and origin of all the ruin and various disasters”[2] were the events that had taken place in the transdanubian barbaricum years before the Goths were admitted to the empire. They did not even try to learn who the Huns were and how they lived and fought.

It is instructive to compare the just quoted words of Ammianus with the following passage by the historian-theologian Paulus Orosius (fl. 415 a.d), St. Augustine’s disciple:

In the thirteenth year of the reign of Valens, that is, in the short interval of time that followed the wrecking of the churches by Valens and the slaughtering of the saints throughout the East, that root of our miseries simultaneously sent up a very great number of shoots. The race of the Huns, long shut off by inaccessible mountains, broke out in a sudden rage against the Goths and drove them in widespread panic from their old homes.”[3]

If the Arian heresy of Valens was the root of all evils and the attack of the Huns on the Goths only a shoot, then it was clearly a waste of time and effort to occupy oneself with the Huns. There was even the danger that by looking too closely at gesia diaboli per Hunnos one might lose sight of the devil himself. Orosius pays attention only to supernatural agents, God or the demons. Unconcerned about the antecedents of a happening or its consequences unless they could be used for theological lessons, Orosius, and with him all the Christian authors in the West, showed no interest in the Huns. Ammianus called the battle of Adrianople another Cannae.[4] He never doubted, even when all seemed lost, that every Hannibal would find his Scipio, convinced that the empire would last to the end of the world:[5] “To these I set no boundary in space or time; unlimited power I have given them.” (His ego nee metas reriim nex tempora pono: imperium sine fine dedi.[6]) Among the Christians, Rufinus was the only one who could say that the defeat of Adrianople was “the beginning of the evil for the Roman Empire, then and from then on.”[7] The others saw in it only the triumph of orthodoxy, indulging in lurid descriptions of the way in which the accursed heretic Valens perished. Orosius adduced the death of the unfortunate emperor as proof for the oneness of God.


Possibly the lack of interest in the Huns had still another reason: the Huns were demonized early. When in 364 Hilary of Poitiers predicted the coming of the Antichrist within one generation,[8] he repeated what during the two years of Julian’s reign many must have thought. But since then Christ had conquered, and only an obdurate fanatic like Hilary could see in the emperor’s refusal to unseat an Arian bishop the sign of the approaching end of the world. Even those who still adhered to the chiliasm of the pre-Constantine church, and took the highly respected Divinae institutiones of Lactantius as their guide to the future, did not expect to hear themselves the sound of Gabriel’s trumpet. “The fall and ruin of the world will soon take place, but it seems that nothing of the kind is to be feared as long as the city of Rome stands intact.”[9]

The change set in early in 387. Italy had not been invaded by barbarians since Emperor Aurelian’s time (270–275). Now it suddenly was threatened by an “impure and cruel enemy.” Panic spread through the cities; fortifications were hastily improvised.[10] Ambrose, who shortly before had lost his brother Saturus, found consolation in the thought that he was “taken away that he might not fall in the hands of the barbarians.... that he might not see the ruin of the whole earth, the end of the world, the burial of relatives, the death of fellow-citizens.” It was the time which the prophets had foreseen, “when they felicitated the dead and lamented the living” (gratulabantur mortuis et vivos plangent).[11] After Adrianople, Ambrose felt that “the end of the world is coming upon us.” War, pestilence, famine everywhere. The final period of the world’s history was drawing to its close. “We are in the wane of the age.”[12]

In the last decade of the fourth century, an eschatological wave swept over the West from Africa to Gaul. The Antichrist already was born, soon he would come to the throne of the empire.[13] Three more generations, and the millennium would be ushered in, but not before untold numbers would have perished in the horrors which preceded it; the hour of judgment drew nearer, the signs pointing to it became clearer every day.[14]

Gog and Magog (Ezekiel 38:1–39:20) were storming down from the north. The initial letters suggested to some people, said Augustine, who himself rejected such equations, identification with the Getae (Goths) and Massagetae.[15] Ambrose took the Goths to be Gog.[16] The African bishop Quodvultdeus could not make up his mind whether he should identify Magog with the Moors or the Massagetae.[17] Why the Massagetae? There were no Massagetae in the fifth century. But considering that Themistius, Claudian, and later Procopius called the Huns Massagetae,[18] it seems probable that those who identified Magog with the Massagetae thought of the Huns. In the Talmud, where the Goths are Gog,[19] Magog is “the country of the kanths” (Sogdian kanf), that is, the kingdom of the white Huns.[20]

Jerome did not share the chiliastic fears and expectations of his contemporaries. In reshaping Victorinus of Poetovio’s Commentary on the Revelation he substituted for the last part, full of chiliastic ideas, sections from Tyconius.[21] But when in 395 the Huns broke into the eastern provinces, he, too, feared that “the Roman world was falling,”[22] and the end of Rome meant the end of the world.[23] Four years later, still under the impression of the catastrophe, he saw in the Huns the savage peoples kept behind the Caucasus by the iron gates of Alexander.[24] The ferae gentes were Gog and Magog of the Alexander legend. Flavius Josephus (37/8100 a.d.), the first to speak of Alexander’s gates,[25] equated the Scythians and Magog.[26] Jerome, who followed him,[27] identified Herodotus’ Scythians with the Huns,[28] in this oblique way equating the Huns and Magog. Orosius did the same; his “inaccessible mountains” behind which the Huns had been shut off were those where Alexander had built the wall to hold back

Gog and Magog. In the sixth century, Andreas of Caesarea in Cappadocia still held the view that Gog and Magog were those Scythians in the north “called Hunnica by us” aneo xaXovpev Ovvvixa.[29] If even the sober Jerome was inclined, for a time, to see in the Huns the companions of the apocalyptic horsemen, one can easily imagine how the superstitious masses felt.[30]

After 400, the chiliastic fears were somewhat abated.[31] But behind the Huns the devil still was lurking. The curious story in Jordanes[32] about their origin almost certainly is patterned on the Christian legend of the fallen angels:[33] The unclean spirits “bestowed their embraces on the sorceresses and begot this savage race.” The Huns were not a people like other peoples. These fiendish ogres,[34] roaming over the desolate plains beyond the borders of the Christian oecumene, from which they set out time and again to bring death and destruction to the faithful, were the offspring of daemonia immunda. Even after the fall of Attila’s kingdom, the peoples who were believed to have descended from the Huns were in alliance with the devil. They enveloped their enemies in darkness vnd ttvag payelag.[35] The Avars, whom Gregory of Tours called Chuni, “skilled in magic tricks, they made them, that is, the Franks, see illusionary images and defeated them thoroughly” (magicis artibus instructi, diversas fantasias eis, i.e., Francis ostendunt et eos valde superant).[36]

To be sure, this demonization of the Huns alone would not have prevented the Latin historians and ecclesiastic writers from exploring the past of the Huns and describing them as Ammianus did. But the smell of sulphur and the heat of the hellish flames that enveloped the Huns were not conducive to historical research.


How did the Eastern writers see the Huns? One should expect the Greek historians to have preserved at least some of the ethnographic curiosity of Herodotus and Strabo. But what we have is disappointing.

Instead of facts they serve us with equations. The Latin chroniclers of the fifth century, in calling the Huns by their proper name, were less guided by the intention to be precise than forced to be factual by their ignorance of literature. They knew next to nothing about the Scythians, Cimmerians, and Massagetae, whose names the Greek authors constantly interchanged with that of the Huns. However, even at a time when there still existed a Latin literature worthy of its illustrious past, the Latin writers, both prosiasts and poets, shunned the circumlocutions and equations in which the Greeks indulged. Ausonius rarely missed an opportunity to show how well read he was, yet he refrained from replacing the real names of the barbarians with whom Gratian fought by those he knew from Livy and Ovid.[37] Ambrose, too, avoided the use of archaic or learned words. The Huns, not the Massagetae, attacked the Alans, who threw themselves upon the Goths, not the Scythians.[38] In Ambrose, the former consularis, Roman soberness and aversion to speculation were as much alive as in Ausonius, the rhetor from Bordeaux. A comparison of Pacatus’ panegyric on Theodosius with the orations of Themistius is revealing: The Gaul called the Huns by their name;[39] the Greek called them Massagetae.[40]

As in the West, many writers in the East lacked interest in the invaders. They looked on them as “bandits and deserters,”[41] or they called them Scythians, a name which in the fourth and fifth centuries had long lost its specific meaning. It was widely applied to all northern barbarians, whether they were nomads or peasants, spoke Germanic, Iranian, or any other tongue. Nevertheless, in the vocabulary of the educated the word retained, however attenuated, some of its original significance. The associations it called forth were bound to shape the way in which the barbarians were seen. That makes it at times difficult to decide whom an author means. Are Priscus’ “Royal Scythians” the dominating tribe as in Herodotus, or are they the members of the royal clan, or simply noblemen ?

It is not enough to say that the phrase is merely one of the several instances of Priscus’ literary debt to Herodotus. It certainly is. But it would be strange if the man who used this and other expressions of the great historian would not, here and there, have succumbed to the temptation to see the Huns as the ancients had seen the Scythians.

The Greek historians equated the Huns and the Cimmerians, Scythians, and other peoples of old not just to display their knowledge of the classics or to embellish their accounts,[42] but first of all because they were convinced that there were no peoples which the wise men of the past had not known. And this, in turn, was not so much narrow-minded traditionalism—it was that, too—as, to use a psychological term, a defense mechanism. Synesius of Cyrene (ca. 370—412), in his “Address on Kingship,” explained why there could not be new barbarians:

Now it was not by walling off their own house that the former rulers prevented the barbarians either of Asia or Europe from entering it. Rather by their own acts did they admonish these men to wall off their own by crossing the Euphrates in pursuit of the Parthians, and the Danube in pursuit of the Goths and Massagetae. But now these nations spread terror amongst us, crossing over in their turn, assuming other names, and some of them falsifying by art even their countenances, so that another race new and foreign may appear to have sprung from the soil.[43]

This is carrying the thesis of the identity of the old and new barbarians to absurdity. But it is, after all, what so many Roman generals said so many times on the eve of a battle: our fathers conquered them, we shall conquer them again. The ever recurring oi TtaXai serves the same purpose. It deprives the unknown attacker of his most frightening feature: he is known and, therefore, needs not be feared.

In the equation of the Huns and the peoples of former times both motives, the emotionally conditioned reductio ad notum and the intention .of the learned historian to show his erudition, play their role, whereby the former, I believe, is more often in the service of the latter than is usually assumed. With which of the known peoples an author identified the Huns depended on his information, the circumstances under which he wrote, and the alleged or real similarity between the known and the barely known. The result was invariably the same. All speculations about the origin of the Huns ended in an equation.

Philostorgius, in his Ecclesiastical History written between 425 and 433, “recognized” in them the Neuri.[44] A well-read man, he may have come across a now lost description of the Neuri which reminded him of what he had heard of the Huns. One could think that Philostorgius, less critical than Herodotus, believed the werewolf stories told about the Neuri.[45] Synesius[46] and Jerome[47] were probably not the only ones to compare the Huns with wolves. It was not beyond Philostorgius to identify the “wolfish” Huns with the werewolves of Scythia. But the most likely explanation of his belief is the location of the Neuri: They were the northernmost people, the Huns came from the extreme North—ergo the Huns were the Neuri. To say that they lived along the Rhipaean Mountains, as Philostorgius did, was merely another way of placing them as far north as possible; since the legendary Aristeas[48] the Rhipaean Mountains were regarded as the region of the eternal snow, the home of the icy Boreas.

Procopius’ identification of the Huns with the Cimmerians[49] is neither better nor worse than his assertion that the Goths, Vandals, and Gepids were in former times called Sauromatae.[50] As a rule Procopius, like Themistius and Claudian,[51] equated the Huns and the Massagetae.[52] The later Byzantine writers repeated monotonously the formula: the former x, the present y.

There is finally the historian Eunapius of Sardes (ca. 345—420). The following fragment from him shows (in Vasiliev’s opinion) what a conscientious historian Eunapius was:

Although no one has told anything plainly of whence the Huns came and by which way they invaded the whole of Europe and drove out the Scythian people, at the beginning of my work, after collecting the accounts of ancient writers, I have told the facts as seemed to me reliable; I have considered the accounts from the point of view of their exactness, so that my writing should not depend merely on probable statement and my work should not deviate from the truth. We do not resemble those who from their childhood live in a small and poor house, and late in time, by a stroke of good fortune, acquire vast and magnificent buildings, and none the less by custom love the old things and take care of them.... But we rather resemble those who first using one medicine for the treatment of their body, in the hope of help, and then through their experience finding a better medicine, turn and incline towards the latter, not in order to neutralize the effect of the first one by the second but in order to introduce the truth into erroneous judgment, and, so to speak, to destroy and enfeeble the light of a lamp by a ray of the sun. In like manner we will add the more correct evidence to the aforesaid, considering it possible to keep the former material as an historical point of view, and using and adding the latter material for the establishment of the truth.[53]

All this talk about medicines and buildings, the pompous announcement of what he is going to write on the Huns, is empty. Eunapius’ description of the Huns is preserved in Zosimus.[54] It shows what a windbag the allegedly conscientious historian was. One half of it Eunapius cribbed from Ammianus Marcellinus;[55] the other half, where he “collected the accounts of the ancient writers,” is a preposterous hodgepodge. Eunapius calls the Huns “a people formerly unknown,”[56] only to suggest in the next line their identity with Herodotus’ Royal Scythians. As an alternative he referred to the “snub-nosed and weak people who, as Herodotus says, dwell near the Ister [Danube].” What he had in mind was Herodotus V, 9, 56, but he changed the horses of the Sigynnae, “snubnosed and incapable of carrying men,” into “snub-nosed and weak people” (at/zov; xai ativvarovs d’vdpa; cpEQEiv into xai aaOsvsa^ avOga)-


Ammianus Marcellinus

Seen against this background of indifference, superstition, and arbitrary equations, Ammianus’ description of the Huns cannot be praised too highly. But it is not eine ganz realistische Sittenschilderung, as Rostovtsev called it.[58] For its proprer evaluation one has to take into account the circumstances under which it was written, Ammianus’ sources of information, and his admiration for the styli veteres.

He most probably finished his work in the winter 392/3,[59] that is, at a time when the danger of a war between the two partes of the empire was steadily mounting. In August 392, the powerful general Arbogast proclaimed Eugenius emperor of the West. For some time Theodosius apparently was undecided what to do; he may have thought it advisable to come to an agreement with the usurper who was “superior in every point of military equipment.”[60] But when he nominated not Eugenius but one of his generals to hold the consulship with him, and on January 23, 393, proclaimed his son Honorius as Augustus, it became clear that he would go to war against Eugenius as he had against Maximus in 388. There can be little doubt that the sympathies of Ammianus, the admirer of Julian, lay from the beginning not with the fanatic Christian Theodosius but with the learned pagan Eugenius.[61] Ammianus must have looked with horror at Theodosius’ army, which was Roman in name only. Although it cannot be proved that the emperor owed his victory over Maximus to his dare-devil Hun cavalry,[62] they certainly played a decisive role in the campaign. Theodosius’ horsemen were “cafried through the air by Pegasi”;[63] they did not ride, they flew.[64] No other troops but the Hun auxiliaries could have covered the sixty miles from Emona to Aquileia in one day.[65] Ammianus had all reasons to fear that in the apparently inevitable war a large contingent of the Eastern army would again consist of Huns. It did.[66]

Ammianus hated all barbarians, even those who distinguished themselves in the service of Rome:[67] He called the Gallic soldiers, who so gallantly fought the Persians at Amida, dentatae bestiae[68] he concluded his work with an encomium for Julius, magister militiae trans Taurum, who, on learning of the Gothic victory at Adrianople, had all Goths in his territory massacied. But the Huns were the worst. Both Claudian[69] and Jordanes[70] echoed Ammianus when they called the Huns “the most infamous offspring of the north,” “fiercer than ferocity itself.” Even the headhunting Alans were “in their manner of life and their habits less savage” than the Huns.[71] Through long intercourse with the Romans, some Germans had acquired a modicum of civilization. But the Huns were still primeval savages.

Besides, Ammianus’ account is colored by the bias of his informants. He went to Rome sometime before 378 where, except for a short while in 383, he spent the rest of his life. The possibility that he met there some Hun or other cannot be entirely ruled out,[72] but it is inconceivable that a Hun who at best understood a few Latin orders could have told Ammianus how his people lived and how they fought the Goths. The account of the war in South Russia and Rumania is based largely on reports which Ammianus received from Goths. Munderich, who had fought against the Huns, later dux limitis per Arabias,[73] may have been one of his informants. One could almost say that Ammianus wrote his account from a Gothic point of view. For example, he described Ermanaric as a most warlike king, dreaded by the neighboring nations because of many and varied deeds of valor;[74] fortiter is a praise which Ammianus did not easily bestow on a barbarian. Alatheus and Saphrax were “experienced leaders known for their courage.”[75] Ammianus names no less than eleven leaders of the Goths,[76] but not one of the Huns. They were a faceless mass, terrible and subhuman.

Ammianus’ description is distorted by hatred and fear. Thompson, who believes almost every word of it, accordingly places the Huns of the later half of the fourth century in the “lower stage of pastoralism.”[77] They lived, he says, in conditions of desperate hardship, moving incessantly from pasture to pasture, utterly absorbed by the day-long task of looking after the herds. Their iron swords must have been obtained by barter or capture, “for nomads do not work metal.” Thompson asserts that even after eighty years of contact with the Romans the productive power of the Huns was so small that they could not make tables, chairs, and couches. “The productive methods available to the Huns were primitive beyond what is now easy to imagine.” To this almost unimaginable primitive economy corresponds an equally primitive social structure, a society without classes, without a hereditary aristocracy; the Huns were amorphous bands of marauders. Even the Soviet scholars, who still hate the Huns as the murderers of their Slavic ancestors, reject the notion that the economy and society were in any way primitive.[78]

Had the Huns been unable to forge their swords and cast their arrowheads, they never could have crossed the Don. The idea that the Hun horsemen fought their way to the walls of Constantinople and to the Marne with bartered and captured swords is absurd. Hun warfare presupposed a far-reaching division of labor in peacetime. Ammianus emphasizes so strongly the absence of any buildings in the country of the Huns that the reader must think they slept the year round under the open sky; only in passing does Ammianus mentions their tents and wagons. Many may have been able to make tents, but only a few could have been cartwrights.

The passage which, more than any other, shows that Ammianus’ description must not be accepted as it stands is the following, often quoted and commented on: Aguntur autem nulla severitate regali; sed tumultuario primatum ductu conlenti, perrumpunt quidquid incident.[79] In Rolfe’s translation, “They are subject to no royal constraint, but they are content with the disorderly government of their important men, and led by them they force their way through every obstacle.” It is not very important that this statement is at variance with Cassiodorus-Jordanes’ account of the war between the Goths and Balamber, king of the Huus, who later married Vadamerca, the granddaughter of the Gothic ruler Vinitharius;[80] whoever Balamber was, Cassiodorus would not have admitted that a Gothic princess could have become the wife of a man who was not some sort of a king. More important is the discrepancy between Ammianus’ statement and what he himself tells about the deeds of the Huns. Altough the cultural level of Ermanaric’s Ostrogoths and the cohesion of his kingdom must not be overrated, its sudden collapse under the onslaught of the Huns would be inexplicable if the latter were nothing but an anarchic mass of howling savages. Thompson calls the Huns mere marauders and plunderers. In a way, he is right. But to plunder on the scale the Huns did was impossible without a military organization, commanders who planned a campaign and coordinated the attacking forces, men who gave orders and men who obeyed them. Altheim defines tumultarius ductus as eine aus dein. Augenblick erwachsene, improvisierte Fuhrung,[81] which renders Ammianus’ words better than Rolfe’s “disorderly government.” However, the warfare of the Huns reveals at no time anything that could be called improvised leadership.[82]

For some time the misunderstanding of the Hunnic offensive tactics— sudden, feigned flight and renewed attack—was, perhaps, inevitable.[83] But Ammianus wrote the last books fourteen years after Adrianople. He must by then have known or, at least, suspected that the early reports on the Huns’ improvised leadership were not true. Yet he stuck to them, for those biped beasts had only “the form of men.”[84] He maintained that their missiles were provided with sharp bone points.[85] He may not have been entirely wrong. But the tanged Hun arrowheads of which we know are all made of iron. Ammianus made the exception the rule.

In describing the Huns, Ammianus used too many phrases from earlier authors. Because the Huns were northern barbarians like the Scythians of old and because the styli veteres wrote so well about the earlier barbarians, Ammianus, the Greek from Antioch, thought it best to paraphrase them. One of the authors he imitated was the historian Trogus Pompeius, a contemporary of the emperor Augustus. Ammianus wrote: “None of them ever ploughs or touches a colter. Without permanent seats, without a home, without fixed laws or rites, thye all roam about, always like fugitives... restless roving over mountains and through woods. They cover themselves with clothes sewed together from the skins of forest rodents.” (Nemo apud eos aval nec stivam aliquando contingit. Omnes sine sedibus fixis, absque lore vet lege aut ritu stabili dispalantur, semper fugientium similes... vagi monies peragrantes el silvas. Indumentis operiuntur ex pellibus silvestrium murum consarcinatis.)[86] This clearly is patterned on Trogus’ description of the Scythians: “They do not till the fields. They have no home, no roof, no abode... used to range through uncultivated solitudes. They use the skins of wild animals and rodents.” (Neque enim agrum exercent. Neque domus Ulis ulla aut tectum aut sedes est... per incultas solitudines errare solitis. Pellibus ferinis ac murinis utuntur.)[87]

It could be objected that such correspondences are not so very remarkable because the way of life of the nomads throughout the Eurasian steppes was, after all, more or less the same. But this cannot be said about other statements of Ammianus which he took from earlier sources. “From their horses,” he wrote, “by day and night every one of that nation buys and sells, eats and drinks, and bowed over the narrow neck of the animal relaxes in a sleep so deep as to be accompanied by many dreams.”[88] His admiration of Trogus here got the better of him. He had read the following description of the Parthians: “All the time they let themselves be carried by their horses. In that way they fight wars, participate in banquets, attend public and private business. On their backs they move, stand still, carry on trade, and converse.” (Equis omni tempore vectantur; Ulis bella, Ulis convivia, Ulis publica et privata officia obeunt; super illos ire, consistere, mercari, colloqui.)[89] Ammianus took Trogus too literally; he rendered “all the time” (omni tempore) by “day and night” (pernox et perdiu) and had, therefore, to keep the Huns on horseback even in their sleep.

Ammianus’ description of the eating habits of the Huns is another example of his tendency to embroider what he read in old books. The Huns, he says, “are so hardy in their form of life that they have no need of fire nor of savory food, but eat the roots of wild plants and the halfraw flesh of any kind of animals whatever, which they put between their thighs and the backs of their horses, and thus warm it a little.”[90] This is a curious mixture of good observation and a traditional topos. That the Huns ate the roots of wild plants is quite credible; many northern barbarians did. Ammianus’ description of the way the Huns warmed raw meat while on horseback has been rejected as a misunderstanding of a widespread nomad custom; the Huns are supposed to have used raw meat for preventing and healing the horses’ wounds caused by the pressure of the saddle.[91] However, at the end of the fourteenth century, the Bavarian soldier, good Hans Schiltberger, who certainly had never heard of Ammianus Marcellinus, reported that the Tatars of the Golden Horde, when they were on a fast journey, “took some meat and cut it into thin slices and put it into a linen cloth and put it under the saddle and rode on it.... When they felt hungry, they took it out and ate it.”[92]

The phrase “Their mode of living is so rough that they eat half-raw meat” (ita uictu sunt asperi, ut semicruda carne uescantuf) is taken from the geographer Pomponius Mela (fl. 40 a.d.), who described the Germans as “Their mode of living is so rough and crude that they even eat raw meat” (uictu ita asperi incultique ut cruda etiam carne uescantur)[93] The Cimbri, too, were said to eat raw meat.[94] Syroyadtsy, a Russian word for the Tatars, possibly means “people who eat raw [meat],” syroedtsy[95] Like so many northern peoples, the Huns may, indeed, have eaten raw meat. Ammianus, however, goes one step further; he maintains that the Huns did not cook their food at all, which is disproved by the big copper cauldrons for cooking meat, one of the leitmotifs of Hunnic civilization. But Ammianus felt he had to force the Huns into the cliche of the lowest of the barbarians.[96]

All this is not meant to dismiss Ammianus’ account as untrustworthy. It contains a wealth of material which is repeatedly confirmed as good and reliable by other literary testimony and by the archaeological evidence. We learn from Ammianus how the Huns looked and how they dressed. He describes their horses, weapons, tactics, and wagons as accurately as any other writer did.

Cassiodorus, Jordanes

In his Hunnophobia, Ammianus was equaled by Cassiodorus (487–583), of whose lost Gothic History much has been preserved in Jordanes’ The Origin and Deeds of the Getae, commonly called Getica. But Cassiodorus had to explain why the Huns could make themselves the lords of his heroes, the Ostrogoths, and rule over them for three generations. His Huns have a wicked greatness. They are greedy and brutal, but they are a courageous people. Attila was a cruel and voluptuous monster, but he did nothing cowardly; he was like a lion.[97] According to Ammianus,[98] the Huns “furrowed the cheeks of children with the iron from their very birth,” words which Cassiodorus copied. But whereas Ammianus continued “in order that the growth of hair, when it appears at the proper time, may be checked by the wrinkled scars,” Cassiodorus wrote “so that before they receive the nourishment of milk they must learn to endure wounds.”[99]

In his account of the early history of the Goths, Jordanes followed Cassiodorus, though not always verbatim. For the proper evaluation of the Gothic tradition about the struggle against the Huns in South Russia, one has to keep in mind that it has come down to us in an expurgated and “civilized” form. In Ostrogothic Italy the memory of the great wars fought side by side with the Huns and against them must still have been alive. Cassiodorus’ sources were songs, cantus maiorum, cantiones, carmina prisca, and stories, some of them told “almost in the way historical events are told” (pene storico ritii). The pene must not be taken seriously. Cassiodorus wrote his Gothic History “to restore to the Amal line the splendor that truly belonged to it.” He wrote for an educated Roman public whose taste would have taken offense at the crude, cruel, and bloody aspects of early Germanic poetry. A comparison of the Getica with Paul the Deacon’s History of the the Langobards shows to what extent Cassiodorus purged the tradition of his Gothic lords all of barbaric features.

But this is not all. The Origo gentis Langobardorum, written about 670, one of Paul’s sources, is full of pagan lore. More than two hundred years after the conversion of the Danes to Christianity, the old gods, scantily disguised as ancient kings, still were wandering through the pages of Saxo Grammaticus. In the 530’s when Cassiodorus wrote his history, there still were alive men whose fathers, if not they themselves in their youth, had sacrificed to the old gods. The Gothic “Heldenlieder” were certainly as pagan as those of the Danes and Langobards. The original breaks through in a single passage in the Getica, taken over from Cassiodorus: “And because of the great victories the Goths had won in this region, they thereafter called their leaders, by whose good fortune they seemed to have conquered, not mere men but demigods, that is, ansis.”[100] Even here Cassiodorus euhemerized the tradition. Everywhere else the pagan elements are radically discarded. The genealogy of the Amalungs, which in the carmina and fabulae was almost certainly full of gods, goddesses, murder, and homicide, reads like a legal document.

Where, as in the account of the war between the Ostrogoths and the Huns, Cassiodorus-Jordanes and Ammianus differ, Ammianus’ version is, without the slightest doubt, the correct one. We cannot even be sure that Cassiodorus’ quotations from Priscus are always exact. However, because so much of our information on the Huns is based on these quotations, they have to be taken as they are. Occasionally (as, for instance, in the story of Attila and the sacred sword, or in the description of Attila’s palace) Cassiodorus renders the Priscus text better than the excerpts which the scribes made for Constantine Porphyrogenitus in the tenth century. And for this alone we must be grateful to the stammering, confused, and barely literate Jordanes. But to elevate him to the ranks of the great historians, as some years ago Giunta tried, is a hopeless undertaking.[101]

II. History

From the Don to the Danube

The first chapters of Ammianus Marcellinus’ last book contain the only extant coherent account of the events in South Russia before 376. From his Gothic informants Ammianus learned that the Huns “made their violent way amid the rapine and slaughter of the neighboring peoples as far as the Halani.”[102] Who those peoples were obviously no one could tell him, and the monumenta vetera supplied no information about them; they were among those “obscure peoples whose names and customs are unknown.”[103] Ammianus’ actual information begins with the Hun attack on the Alans: The Huns overran “the territories of those Halani (bordering on the Greuthungi [Ostrogoths]) to whom usage has given the surname Tanaitae [Don people].”

This passage has been variously interpreted.[104] How far to the east and west did the “Don people” live? In one passage Ammianus locates all Alans—and that would include the Tanaitae—“in the measureless wastes of Scythia to the east of the river,”[105] only to say a few lines later that the Alans are divided between the two parts of the earth, Europe and Asia,[106] which are separated by the Don.[107] The Greuthungi-Ostrogoths[108] were the western neighbors of the Tanaitae, but in another passage Ammianus puts the Sauromatae, not the Greuthungi, between the Don and the Danube,[109] and in still another one (following Ptolemy, Geography V, 9, 1) also east of the Don.[110]

Ammianus suffered from a sort of literary atavism, garbling the new reports with the old.[111] The chapter on the Alans in book XXXI includes a lengthy dissertation on the peoples whom the Alans “by repeated victories incorporated under their own national name.”[112] Ammianus promises to straighten out the confused opinions of the geographers and to present the truth. Actually, he offers the queerest hodgepodge of quotations from Herodotus, Pliny, and Mela,[113] naming the Geloni, Agathyrsi, Melanchlaeni, Anthropophagi, Amazons, and Seres, as if all these peoples were still living in his time.

The Huns clashed with Alanic tribes in the Don area. This is all we can retain from Ammianus’ account. If Ammianus had used the term Tanailae as Ptolemy used it,[114] the “Don people” would have lived in European Sarmatia. But a river never formed a frontier between seminomadic herdsmen, and certainly not the “quietly flowing” Don. The archaeological evidence is unequivocal: In the fourth century Sarmatians grazed their flocks both east of the Don as far as the Volga and beyond it, and west of the river to the plains of Rumania. Exactly where the Huns attacked cannot be determined; like the later invaders, their main force probably operated on the lower course of the river.

Ammianus’ account of an alliance between a group, or groups, of Alans and the Huns cannot be doubted. In the 370’s and 380’s Huns and Alans are so often named together, that some kind of cooperation of the two peoples would have to be assumed even without Ammianus’ explicit statement:

The Huns killed and plundered them [i.e., the Tanaitae] and joined the survivors to themselves in a treaty of alliance, [reliquos sibi concordandi fide pacta iunxerunt]; then in company with them they made more boldly a sudden inroad into the extensive and rich cantons of Ermenrichus.[115]

For a long time [diu],[116] the king of the Greuthungi “did his best to maintain a firm and continued stand, but as rumor gave wide currency to and exaggerated the horror of the impending danger,” he killed himself. His successor Vithimiris

resisted the Halani for a time [aliquantisper], relying on other Huns, whom he paid to take his side. But after many defeats which he sustained, he was overcome by force of arms and died in battle. In the name of his little son, Viderichus, the management of affairs was undertaken by Alatheus and Saphrax, experienced generals known for their courage; but since the stress of circumstances compelled them to abandon confidence in resistance, they cautiously retreated until they came to the river Danastius.[117]

Ammianus’ account has been rejected by the Croatian scholar L. Hauptmann, who thought that either Ammianus made a bad blunder or that the text was corrupt.[118] Not Hanis aliis fretus Vithimir must have resisted the Alans but *Halanis aliis fretus, the Huns. Hauptmann referred to Jordanes, in whose account the only enemies of the Ostrogoths are, indeed, the Huns. But Jordanes’ compilation is tendentious from beginning to end. He not only retained the transfiguration of the early history of the Goths as he found it in Cassiodorus; he also changed what he read in Ammianus in favor of the Alans.[119] They were, wrote Ammianus, Hunis per omnia suppares (XXXI, 2, 21); Jordanes, Getica 126, changed this into pugna pares. According to Ammianus, the Alans were, in comparison with the Huns, victu mitiores et culhr, Jordanes replaced mitiores by dissimiles, and cultu by humanitate. He read in Ammianus that the Alans attacked the Ostrogoths after Ermanaric’s death. But this did not fit the picture of the noble Alans, so he left it out.

The fights between the Alans and Goths also are attested by Bishop Ambrose of Milan (374–397 a.d.). In the Expositio evangelii secundum Lucam, probably written at the end of 378,[120] he summarized the events which led to the disaster of Adrianople: “The Huns threw themselves upon the Alans, the Alans upon the Goths, and the Goths upon the Taifali and Sarmatae; the Goths, exiled from their own country, made us exiles in Illyricum, and the end is not yet.”[121]

The information Ambrose received in Milan was not quite correct. Ermanaric’s kingdom collapsed under the onslaught of the Huns. But the testimony of both Ammianus and Ambrose leaves no doubt that at one time in the apparently long struggle the main enemies of the Goths were, indeed, Alans. Were they only those Alans who had made an alliance with the Huns? This is possible. But the following strange story told by Jordanes might preserve a dim memory of an uprising of Alanic groups within the Ostrogothic kingdom:

Now although Hermanaric, king of the Goths, was the conqueror of many tribes, as we have said above, yet while he was deliberating on this invasion of the Huns, the treacherous tribe of the Rosomoni [Rosomonorum gens infida], who at that time were among those who owed him homage, took this chance to catch him unawares. For when the king had given orders that a certain woman of the tribe I mentioned, Sunilda by name, should be bound to wild horses and torn apart by driving them in opposite directions (for he was roused to fury by her husband’s treachery to him), her brothers Sarus and Ammius came to avenge their sister’s death and plunged a sword in Hermanaric’s side. Enfeebled by this blow, he dragged out a miserable existence in bodily weakness. Balamber, king of the Huns, took advantage of his ill health to move an army into the land of the Ostrogoths.[122]

Whereas Sunilda is unquestionably a Germanic name, the derivation of Sarus from Gothic sarwa, “weapon, armor,” and of Ammius from Gothic *hama, “to arm,”[123] is unconvincing. There is no satisfactory etymology of Rosomoni.[124] Sarus occurs later as the name of a Goth,[125] but this does not necessarily make Sanis of the Rosomoni a Goth. The name can be compared with Sarosius or Saroes,[126] who in about 500 was king of the Alans in the Caucasus. Sarakos in an inscription from Tanais (early third century a.d.) probably is derived from the Sarmatian word that corresponds to Avestic sara-, Ossetic sar-, “head”;[127] Sarus could mean “caput, captain.” Saphrax (Safrax) and Lagarimanus, prominent leaders of the Goths, had Iranian names;[128] they might have been Alans. Although it cannot be proved that the Rosomoni were rebellious Alans, the discessus of an Alanic gens at a time when Alans attacked the Ostrogoths seems more likely than the treachery of Gothic noblemen.

It was almost certainly the concordia with large groups of Alans which enabled the Huns to move against Ermanaric. Ammianus does not say what the terms of the alliance were. When one considers that those Alans who in 418 subjugated themselves to the palrocinium of the Vandal king, retained their tribal organization until the end of the Vandal kingdom, it may be assumed that the Hunno-Alanic alliance guaranteed the Iranian partner a considerable degree of independence and a large share in the loot. It was certainly not the first time that other tribes joined the Huns, nor was it the last. In some cases the alliance seems to have resulted in a real symbiosis, in others the tribes united temporarily for raids and looting expeditions. The Hunno-Alanic alliance lasted three decades.

Ammianus’ account on the Alanic attacks on the Goths is borne out by Ambrose, but there seems to be no other authority to confirm what it says about the Huns who sided with Vithimir. Why should Huns, even if they were paid by the Gothic king, fight for him at a time when his situation was so obviously hopeless? If they stayed with those hordes which not even the great Ermanaric could withstand, they could expect to loot at their hearts’ desire; shortly afterward, the Huns who broke into the land of the Visigoths were quickly so loaded down with booty that they had to break off the attack.[129] Were the Huns siding with the Goths a

part of the people who had crossed the Don? Or did Huns live west of the river, tribes which found themselves as threatened as the Goths and decided, when Vithimir appealed for their help, to make common cause with the Germans against the invaders?

A passage in the Getica of Jordanes, going back to the fifth-century historian Priscus, gives the answer. “Like a whirlwind of nations the Huns swept across the Alpidzuri, Alcildzuri, Itimari, Tuncarsi, and Boisci who bordered that part of Scythia.”[130] As we shall see, the first two names stand for one, the Turkish name *Alp-il-cur, which cannot be separated from the Hunnic names ending in -fur. The other names will occupy us later. In the present context this one name, *Alpilcur, suffices to prove the existence of Turkish-speaking nomads [131] on or near the northeastern shore of the Black Sea before the Huns came. In the 430’s the same peoples, listed in the same order and now under Hun domination, had their pastures along the Danube.[132] Whether they migrated or were settled there by their Hun lords is of minor importance. What matters is that their alliance withstood all the vicissitudes of those stormy decades. Because in both passages the *Alpilcur are named first, they apparently were the leading tribe. Overrun by the Huns near the Maeotis, they were, sixty years later, still bitterly opposed to their masters; they made a treaty with the Romans. In a later chapter I shall come back to those “Huns before the Huns.”

Ermanaric’s Kingdom

It is often assumed that Attila ruled over all the peoples once under the king of the Ostrogoths, Ermanaric. Archaeologists perhaps would have hesitated to attribute graves in the forests of central Russia to the nomadic Huns had they not believed that at one time Ermanaric’s Goths had ruled there. The assertion of the West Roman ambassadors at Attila’s court that the Hun king was the lord over the islands in the ocean would not have been so widely accepted were it not for Jordanes’ statement that the Aesti on the Baltic coast were Ermanaric’s subjects. Lack of criticism and chauvinistic bias either enlarged the Gothic realm out of all proportions or practically denied its existence.[133]

Jordanes’ description of it is almost a hymn.[134] “Some of our ancestors,” he wrote, “have justly compared Hermanaric to Alexander the Great.” Obviously it was Jordanes’ source Cassiodorus, not an illiterate Goth, who made this comparison and called Ermanaric “the ruler of all nations of Scythia and Germania.” Jordanes listed thirteen peoples which the Amalung ruler Ermanaric conquered in the north: Golthescytha, Thiudos, Inaunxis, Vasinabroncae, Merens, Mordens, Imniscaris, Rogas, Tadzans, Athaul, Navego, Bubegenes, Coldas. The uncertain readings and the queer forms of these names make them an ideal hunting ground for name chasers. Tomaschek took Athaul for the name of a Hunnic tribe, Turkish *ataghul, “archer.”[135] Mullenhoff thought that scytha in Golthescytha was Latinized chad, the designation of Finnish tribes in the early Russian chronicles.[136] Marquart took golthe for another form of Scoloti, connected it with thiudos, dismissed scytha as a gloss, and arrived thus at “the Scolotic peoples.”[137] He and Grienberger had no doubts that thiudos was Gothic, meaning “peoples,” but Grienberger suspected in golthe Latin gothice, connected scytha and thiudos, and translated “in Gothic, the Scythian peoples.”[138] To discuss these and equally fanciful etymologies would be a waste of time. The Mordens[139] are the Mordvins and the Merens the Mari.[140] Whether Ermanaric actually domuerat them is more doubtful. The ethnic names may merely reflect the extent of the geographical knowledge of Jordanes or his sources.

Ermanaric is also said to have subdued the Aesti on the Baltic coast “by his wisdom and might,” which probably means no more than that there existed some trade relations between the Goths and the tribes in the amber countries, as they possibly existed in Hunnic times[141] and under the great Ostrogothic king Theoderic (Theodoric).[142]

After the conquest of the northern peoples, Ermanaric “reduced to his sway” the Heruli near the Azov Sea, which is quite credible. Since the middle of the third century a tribe of the East Hermanic Heruli had dwelt on the shores of the Mueotis.[143]

Finally, Ermanaric attacked and subjugated the Venethi. Translated from the hymn into prose: “from time to time the Goths made raids into Slavic territory in the northwest.” In the confused account of the years following Ermanaric’s death, Jordanes speaks of a war between a section of the Ostrogoths and the Antes, led by King Boz.[144] After the victory over the Antes, the Goths were attacked by the Huns and defeated on the river Erac.[145]

The boundaries of the Ostrogothic “empire” cannot be defined because it had none. Around a more or less compactly settled Gothic area lay the sites of various tribes. Some of them may have paid regular tribute; others only bartered their goods, presumably mostly furs, for what the Goths got either from the Bosporan kingdom or the Danube provinces; still others occasionally may have joined the Ostrogoths in looting expeditions. The rapid collapse of Ermanaric’s kingdom clearly indicates its lack of coherence.

To analyze once more Ammianus’ account of the war between the Huns and the Visigoths, the southern neighbors of the Ostrogoths, is not our task. This has been done by all the historians of the Migration Period, most competently and succinctly, in my opinion, by Putsch.[146] The Visigoths under Athanaric expected the attack of the Huns on the right bank of the Dniester but could not hold it; they retreated behind the Sereth. The larger part of the people decided to seek a new home in the empire; Athanaric and his followers marched through Oltenia into Caucalandis locus. According to Putsch, Cuuculund wus the mountainous purt of the Bunut, between the rivers Muros, Theiss, und Dunube.[147] The objections to his thesis[148] are based on doubtful equations of the name Cauca; they disregard the events in the late 370’s, which definitely point to Visigoths in the eastern Banat. I, therefore, accept Patsch’s location.[149]

From about 376 on, the Huns were the rulers of a large area in South Russia. They stood at the lower Danube. The picture that can be drawn from Ammianus is not wrong but onesided. He says nothing about the fate of the Bosporan kingdom, the life of the peoples whom the Huns overran, their economy, their social institutions, their interrelations. It would be unfair to blame Ammianus. He wrote a history of the Roman Empire, not one of the barbarians. Fortunately the cultures of the peoples west of the Don can, at least in their outlines, be reconstructed, mainly with the help of the archaeological material.

The Huns at the Danube

In the summer of 376, tens of thousands of Visigoths were encamped on the northern bank of the lower Danube around Durostorum (modern Silistra), anxiously waiting for permission to cross the river and settle in Thrace. They were the greater part of the proud nation which only a few years before had forced the Romans to deal with their leader Athanaric as an equal of the king of kings. Now, defeated by the Huns (see preceding section) and starving, they were deadly scared lest their enemies fall upon them again before they were admitted to a refuge in the empire.

Permission came in the fall. The Visigoths, shortly followed by Ostrogoths,[150] Taifali,[151] and other transdanubian barbarians,[152] crossed the Danube. The following struggle between the Visigoths[153] and the East Romans, which for years raged throughout Thrace, at times engulfing large tracts of Macedonia, has been thoroughly studied. This is understandable and legitimate. The Germanic invaders developed into great nations; in France and Spain they shaped the fate of the Western world. Except for the few years of Attila’s reign, the Huns loomed on and beyond the periphery of the oecumene. Their history in the last decades of the fourth century seems to be bare of all interest. Even those scholars who made the Huns the special object of their studies paid no attention to it.[154]

It is true that our information about the Huns in that period is scanty, although not much scantier than for others. But this should be only a challenge to make the most of the few data. To extract from the annals, commentaries on the Bible, homilies, edicts, and poems the few passages dealing with the Huns and to determine what happened, when, and where, requires an inordinately large apparatus. But that cannot be helped if we want to learn how the Huns moved into central Europe.

Visigoths and Huns Cooperate

After the sanguinary battle Ad Salices in the northern Dobrogea between Visigoths and imperial troops (see Chapter XII) in the summer of 377, the Romans retreated behind the Haemus (Balkans). Their losses were not quite as heavy as those of the Visigoths. But even with the reinforcements being sent to him, the Roman commander could not risk another battle. The Visigoths were still far superior in numbers. Their strength was, however, at the same time their weakness. They were not an army, they were a whole people: women, children, sick people, old people, four or five times outnumbered the warriors. “Everything that could serve as food throughout the lands of Scythia and Moesia had been used up. All the necessities of life had been taken to the strong places, none of which the enemy even attempted to besiege because of their complete ignorance of these and other operations of the kind.”[155]

The Romans hastily fortified the mountain passes. The Goths found themselves “crowded between the Hister [Danube] and the waste places.” Their situation was rapidly getting desperate. Roman troops would easily have broken through the aggeres celsi or high ramparts of the Goths, obviously mere stockades: to the Goths they proved unconquerable. “Driven alike by ferocity and hunger,” they attacked time and again, only to be driven back. Hemmed in by the sea to their left, the mountains to the right and in front of them, in their back the Danube, the Goths could not hold out much longer. “Compelled by dire necessity they gained an alliance with some of the Huns and Halani by holding out the hope of immense booty.” As soon as the Roman commander heard of this, he evacuated his positions and retreated to the Thracian plain.

Ammianus Marcellinus gives a picturesque account of the events following. But instead of telling his readers what actually happened, he describes at great length and with gruesome details the horrors of the barbarian invasion. We hear much about the misery of women and freeborn men driven along by cracking whips, but we do not learn why the Romans retreated. The Huns had as little experience in storming even improvised fortifications as the Visigoths. In the mountains their horsemen were as good as lost. The few who might have sneaked behind the Roman lines could be cut down easily. The Goths did not need more men; they had enough. Besides, Ammianus himself stresses that the number of the Huns and Alans was small, Hunnorum et Halanorum aliquos. Why, then, did the blockade break down? Looking at the map, Seeck found the answer: the Huns most probably crossed the Danube far to the west. Riding down the Morava valley to Naissus (modern Nis, Yugoslavia) and turning east, they threatened the rear of the Romans.[156] Saturninus, the Roman commander, had no choice. He left the passes. The Goths were saved.

A strategic move on such a scale required more than an agreement between the Visigoths and “some” Huns. It presupposed on the part of the Huns the capacity of throwing hundreds of horsemen into action. What the status of their leaders was we do not know. But whether they were “kings,” or phylarchoi (tribal chieftains), or hetmans, whether their men followed them out of loyalty, or to gain military laurels, or simply in order to make, in the shortest time, as much booty as possible, is irrelevant compared with the fact that these horsemen could be assembled, that their leaders did come to an agreement with the Visigoths, that the Huns were kept together over hundreds of miles. The very first account of a Hun raid into the Balkan provinces refutes the view that for half a century after the invasion of South Russia Hun society consisted of a large number of tiny independent groups. But the problems of Hun society will occupy us in another context.

It sometimes has been maintained that the Huns fought at Adrianople (see Chapter XII) side by side with the Goths.[157] Adrianople (378 a.d.) was a Gothic victory. “The Roman legions were massacred by the Goths” (Romanae legiones usque ad internicionem caesae sunt a Gothis), wrote Jerome one year after the catastrophe, and none of those who made use of his chronicle had in this respect anything to add from other sources. Ammianus’ account of the battl eis far from being as precise as one would expect from an author of his military experience and grasp for essentials. Yet so much is certain: the decision fell with the arrival of the Ostrogoths. Fritigern’s Visigoths could not withstand the fierce attack of the Roman cavalry. Driven back to their wagons, hard pressed by the advancing legions, they were rescued by Alatheus’ and Safrax’s Ostrogothic horsemen. The Visigothic leader avoided giving battle as long as he could, partly because he still hoped to come to an understanding with the emperor, but mostly because he did not dare to fight alone. The Romans had their Saracen horses; Fritigern needed desperately the Ostrogothic cavalry. Had they not rushed in just in time, the Visigoths in all probability would have been defeated, if not annihilated. The sudden Ostrogoth attack threw the Romans into confusion, then into panic, and what followed was a massacre.

Adrianople, one of the decisive battles of history, was won by equitatus Gothorum. It is true that there were a few men of other tribes with them, but these were not Huns. Ammianus speaks specifically of Halanorum manus.[158] Had the account been written by Jordanes, we might suspect that he did not want to give the Huns credit for a Gothic victory. Ammianus had no reason to prefer the Alans to the Goths. In his narrative the Huns reappear after the battle. When the Goths set up their camp at Perinthus at the Sea of Marmara, they were Hunis Halanisque permixti.[159] The Huns had stayed away from the fight. Their descendants, the “Massagetae” in the Roman army in Africa, did the same more than once. They waited to see who would win. The Huns were out for looting, and had no desire to spill their blood pour le rot des Goths.

In the following two years our sources repeatedly name Huns, Goths, and Alans together,[160] but whether the Huns looted and burned down the villages of the unfortunate population of Thrace alone or as the allies of the Goths is not known. Some contemporary authors saw in the Huns the worst villains. They were “more fierce than any kind of destruction” (omni pernicie atrociores’).[161] Orosius names the Huns and Alans before the Goths.[162]

After 380, neither Huns nor Alans are mentioned among the barbarians in the Balkan provinces.[163] Goths served in the imperial armies by the thousands. The Roman commanders Botherich, Eriulf, Fravitta, Gainas, and Rumorid were Goths. But we do not hear of Hun contingents or Hun officers. The Huns returned beyond the Danube.

Although the Huns did not fight at Adrianople, indirectly they might have decided the outcome of the battle. The following chronological and geographical deliberations seem to lead away from the Huns. But without them the events in the barbaricum (the territories beyond the Roman frontiers) cannot be reconstructed.

The Huns Threaten Pannonia

In the beginning of June 378, Gratian’s army, which was supposed to join as quickly as possible the Eastern Romans hard-pressed by the Visigoths, finally set out for Thrace. The young emperor’s frivolous wish to present himself to Valens as the victor over mighty barbarians in the West delayed the march for at least a month.[164] But now Gratian hurried. He led his troops in long marches, porrectis itineribus, from Felix Arbor on Lake Constance to Lauriacum, the present Lorch in Upper Austria. There the army, which had marched 300 milia,[165] rested for a short time.[166] Gratian himself “sent on ahead by land all his baggage and packs, and descending the Danube... came to Bononia [in Pannonia superior; now Banostor] and entered Sirmium [in Pannonia inferior; now Sremska Mitrovica]. Having been delayed there four days, he went on over the same river to Castra Martis,[167] although attacked by intermittent fevers. In that region the Halani unexpectedly fell upon him, and he lost a few of his followers.”[168] It was the first encounter with the enemy.

Gratian would not have dared to sail down the Danube with only “a band of light-armed troops,” unless he could have been sure that the Quadi, Jazygi, and Sarmatae on the left bank of the river would keep the peace. They still suffered from the defeats which three years before Valentinian had inflicted on them. The Quadi were forced to provide recruits for the Roman army and the alliance with the Sarmatae Argaragantes in the Banat had been renewed. To prevent the recurrence of surprise attacks like those which in 374 and 375 carried the barbarians deep into Roman territory, the frontier fortifications were greatly strengthened.[169] Pannonian soldiers could be detailed for service in Britain.[170] In the spring of 378, Gratian’s general Frigeridus with his Pannonian and transalpine auxiliaries joined the forces in Thrace.[171] Gratian had nothing to fear from the peoples east of the Danube. But only a few months later Valeria, the easternmost province of Pannonia, was overrun by Goths, Huns, and Alans.

Assuming that Gratian traveled as fast as Emperor Julian (a.d. 360363) who, in the summer of 361, in exceptionally good weather, sailed with three thousand men from “the place where the river is navigable” to Sirmium in eleven days,[172] Gratian could have arrived in Bononia at the end of June or early in July. He probably was in Martis Castra not later than the middle of July. Whether he could have joined Valens before August 9 the day of the fateful battle, is a moot question. The letter he sent to Valens shows that he was determined to throw his cavalry into the struggle as fast as he could.[173] Yet a passage in Zosimus’ New History, composed in the sixth century, seems to indicate that Gratian suddenly stopped, turned around, and rode back to Sirmium.

Victor, commander of the horse, one of the few high officers to survive the massacre at Adrianople, fought his way, with some of his horsemen, “through Macedonia and Thessaly to Moesia and Paiones to inform Gratian, who was there, of what had happened.”[174] Paiones stands here for the province of Pannonia secunda.[175] If Victor was indeed the first to report to Gratian the death of Valens and if he met him in Pannonia secunda, Gratian must have returned to Sirmium not, as is generally assumed, because he realized that after the annihilation of the Eastern army he alone was too weak to continue the fight with the Goths but before he learned about the catastrophe. Zosimus is not a very reliable author,[176] his to av/j,- fiav may only mean “all the details.” The most important single news, that of Valens’ death, Gratian may have received while he was still marching east.[177] If, however, he should have returned before, there could have been only one reason: his troops, although needed in Thrace, must have been needed even more urgently in Pannonia. Valens fought the Goths; Gratian had to fight the peoples driven into Pannonia by the Huns, and the Huns themselves.

Gratian had asked Bishop Ambrose, first in letters, later at their meeting in Sirmium,[178] to write for him a treatise on the orthodox faith. Ambrose composed the first two books De fide “hastily and summarily, and in rough rather than exact form.”[179] He wrote them after he learned of the heretic Valens’ death,[180] which did not particularly grieve him. He hailed the young orthodox emperor Gratian as “the ruler of the whole world” who would conquer the Goths.[181] In the midst of theological arguments and scriptural proofs for the consubstantiality of Father, Son, and Holy Ghost, there is a passage that calls for close attention: “Have we not heard,” wrote Ambrose, “from all along the border, from Thrace and through Dacia ripensis, Moesia, and all of Valeria of the Pannonias [omnemque Valeriam Pannoniarum], a mingled tumult of blasphemers [sc. Arians] preaching and barbarians invading?”[182] Ambrose left out Pannonia secunda, where evidently Gratian’s main force stood. That he stressed the invasion of Valeria, of all Valeria, is all the more significant.

In De fide the Goths were still the only enemy. Ambrose soon received more exact, and more alarming news. “The Huns,” he wrote now, “threw themselves upon the Alans, the Alans upon the Goths, and the Goths upon the Taifali and Sarmatians; the Goths, exiled from their own country, made us exiles in Illyricum, and the end is not yet.”[183] The blurred picture the Romans had of the happenings beyond the Danube became clearer: Athanaric’s Visigoths, who had not joined Fritigern, threw themselves upon the Taifali in Oltenia and then upon the Sarmatians in Caucaland.[184] Throughout the barbaricum, “as far as the Marcomanni and Quadi,”[185] the peoples began to stir. We have no information about the resistance which the Sarmatians in Caucaland, the Banat, put up against the Goths. It must have been stubborn; the Argaragantes were known to be brave and resourceful.[186] But it was overcome, and an apparently large group of Sarmatians was forced to cross the Danube into Valeria. In December 378, the retired general Theodosius, hastily called from Spain, defeated the invaders.[187] Although the account of the battle by the church historian

Theodoret is heavily embroidered, it is substantially true. One passage even sheds light on the composition of the invading hordes. “Many of the barbarians,” wrote Theodoret, “were slain by their own countrymen.” Evidently the Sarmatae Limigantes, the “slaves” of the Argaragantes,[188] turned against their lords and killed them with the weapons they were supposed to use against the Romans.

Theodosius’ victory may have slightly eased the pressure on one sector of the front. But it was a mere episode in the gigantic struggle. In January 379, when Gratian proclaimed Theodosius emperor, the situation was almost hopeless. “The cities are devastated, myriads of people are killed, the earth is soaked with blood, and a foreign people [Aad; aAAo’yAcoo’o’o?] is running through the land as if it were theirs.”[189] Gratian could no longer from his headquarters in Sirmium direct the operations on a front that reached from western Hungary to the Black Sea. Eastern Illyricum, comprising the dioceses of Dacia and Macedonia, was added to the praetorian prefecture Oriens, to be governed by Theodosius as Valens’ successor.[190] The division of Illyricum into an eastern and a western portion was necessitated by purely military reasons. Gratian took over the fight against the invaders of Pannonia.

The ecclesiastical historians Socrates and Sozomen speak vaguely about the tribes from the banks of the Hister, or just barbarians.[191] The Roman orator Symmachus (ca. 340–402), too, refers to the victories of the two emperors without saying who the enemies were.[192] The poets are, fortunately, more specific. From Pacatus and Ausonius we learn that the peoples who had driven the Sarmatians against and west of the Danube were now attacking the limes themselves and piercing them at many points. Theodosius was still in Spain when the Goths, Huns, and Alans broke into Valeria. “Whatever the Goth wastes, the Huns plunders, the Alan carries off, Arcadius will later wish [to recapture]” (Quidquid atterit Gothus, quidquid rapit Chunus, quidquid aufert Halanus, id olim desiderabit Arcadius).[193] “Alas, I have lost the Pannonias” (Perdidi infortunata Pannonias) laments the res publica, imploring Theodosius to come to her rescue. Pacatus was exaggerating. Pannonia was not yet lost, but it was under heavy attack. At the end of 378, Ausonius, friend and teacher of Gratian, consul for 379, received in Trier good news:

All foes now vanquished (where the mixed Frankish and Suebian hordes vie in submission, seeking to serve in our Roman armies; and where the wandering bands of Huns had made alliance with the Sarmatians; and where the Getae with their Alan friends used to attack the Danube—for victory borne on swift wings me the news of this), lo now the Emperor comes to grace my dignity, and with his favor crowns the distinction which he would fain have shared.[194]

It perhaps would be wrong to attach too great importance to the differentiation between Sauromatae and Alani and the alleged alliance between the barbarians, though the Sarmatians, attacked by the Goths, actually might have turned to the Huns for help. The victories cannot have been as decisive as they looked from far-away Trier—for the war went on.

Gratian stayed in Sirmium throughout February and the first half of March. On April 5, he was in Tricciana,[195] the present Sagvar, a town on the road from Sopiana to Arrabona, about 10 miles south of the northwestern shore of Lake Balaton.[196] What he did in northern Pannonia we again learn from a few passages in Ausonius. The Gallic rhetor may have somewhat exaggerated the emperor’s exploits but he did not invent them, as the outcome of the fighting shows. In the thanksgiving for his consulship, addressed to Gratian at Trier at the end of 379,[197] Ausonius extols the young ruler for having “pacified in a single year the Danubian and the Rhenish frontiers.”[198] He hails him as Sarmaticus “because he has conquered and forgiven [vincendo et ignoscendo] that people.”[199] In an epigram Ausonius praises Gratian, who “midst arms and Huns ferocious and Sauromatae dangerous in stealth, whatever rest he had from hours of war, in camp he lavished it all on the Clarian muses.”[200] In a nightmare Ausonius saw himself as a disarmed Alan prisoner of war dragged through the streets.[201]

By the middle of June the situation had so much improved that Gratian could hand over the command to one of his generals and leave for Italy.[202] Besides, the new uprising of the Alamanni in the West required his presence on the Rhine.

Hunnic Pressure on the Lower Danube

We need not follow the struggle between the Visigoths and Theodosius’ armies. If there were still Huns among the barbarians, they were at the most a few stragglers who had been separated from their hordes, or broken men. But the Hunnic danger was by no means over. In the winter 381/2, Sciri and Carpodacians, “mixed with Huns,” crossed the Danube,

to be driven back after a few skirmishes.[203] The episode seems to be significant only insofar as it shows that the Huns were unable to prevent more active tribes north of the Danube from acting on their own. Yet Theodosius could not have failed to realize that the terrible horsemen who had made themselves masters, though as yet not absolute masters, of the teeming mass of barbarians in “Scythia” might someday prove to be a greater danger to his pars than the Goths. He made peace with the Visigoths in the fall of 382.

Weakened by epidemics,[204] their bands thinned out by desertions, deadly tired of incessantly moving from place to place, the Visigoths were more than willing to come to some agreement with the emperor. They wanted land to settle and, if they could get them, subsidies. Theodosius wanted soldiers. The peace treaty gave the Goths large tracts in Moesia inferior and eastern Dacia ripensis;[205] it gave the emperor troops to guard the Danube from Oescus (on the Danube near the confluence with the river Golem Iskr) to Durostorum. Themistius’ New Year’s address of January 1, 383, must not be taken literally. After his experiences with the barbarians Theodosius could not have expected that, like the Celts in Galatia,[206] the Goths would become good and law-abiding Roman citizens. But he certainly hoped they would serve him as a defense.[207] A year before, Athanaric’s retainers were settled on the right bank of the river “to prevent any incursions being made against the Romans.”[208] Zosimus, like Themistius, did not name the potential enemy. Eunapius of Sardes was explicit. The emperor, he wrote, gave the Goths cattle and land, expecting them to form “an unconquerable bulwark against the inroads of the Huns.”[209] As federates the Visigoths were bound to serve whenever and wherever they were called, but their main and permanent assignment was to defend themselves. By fighting for their new home, they fought for Rome. As long as they held the watch on the Danube, the northern Balkan provinces, except easternmost Scythia minor, seemed to be safe. For a few, all too few years, the Roman population in the ravaged towns and villages enjoyed a modicum of peace. In 384 or 385 a barbarian horde crossed the frozen Danube near its mouth and took Halmyris.[210] But this was outside the Gothic territory. Shortly afterwards Hunnic hordes raided Scythia.[211]

In 386, again to the east and west of the Gothic watch on the Danube, barbarians struck, in some parts deeply, into Roman lands. An edict of July 29, 386, gives a strange picture of the situation in the Balkans: “Because the procurators of the mines within Macedonia, Dacia mediterranea, Moesia, and Dardania,[212] who are customarily appointed from the decurions and who exact the usual tax collections, have removed themselves from this compulsory public service by pretending fear of the enemy [simulate hoslili metu], they shall be dragged back to the fulfillment of their duties.”[213] The procurators were certainly willing to use any excuse for shirking their most unpleasant duties, but they could not invent an enemy if there was none. The sequence in which the four provinces are named leaves no doubt that it was the Morava-Vardar Valley in which the enemy operated; that they could spread fear as far as Macedonia shows that the raiders were swift-riding horsemen. They may not have been many; still they were strong enough to overrun the Roman troops, probably by-passing fortified places, and returning unmolested with their booty from where they came. There was no other enemy then and there which could make such raids into the western Balkans but the transdanubian Huns.

The invaders in the East were Germans. In the summer of 386, Greuthungi, led by Odotheus, and their allies appeared on the left bank of the lower Danube and asked Promotus, master of the soldiers in Thrace, for permission to cross the river; they wanted land for settlement. When their request was rejected, they tried to force their way into the empire. Promotus inflicted a crushing defeat on them.[214]

Zosimus, following two sources, tells the same event twice. He gives a detailed account of the stratagem by which Promotus deceived the barbarians; the poet Claudian indulges in a gory description of the slaughter of the Greuthungi. But neither of these two authors, shows any interest in the antecedents of the short war: it was just another outbreak of the well-known “insanity” of the savages. Though unlikely, it is not impossible that Zosimus’ sources contained more about the Greuthungi and the reasons why they trekked south. For it was a trek, the migration of a very large group of peoples in search of a new home. Zosimus stresses that they had their wives and children with them. How many they were we are not told. Claudian certainly exaggerates the number of boats manned by the flower of barbarian youth and sunk by the Romans. But even if their number was not three thousand, as he wrote, but only one thousand, with no more than three or four men in each, we would arrive at a figure of close to ten thousand arms-bearing men. A German army could number a quarter or a fifth of the population. However, even if the Greuthungi, together with all the tribes and fractions of tribes which joined them,[215] numbered not fifty but thirty or twenty thousand (both of Zosimus’ sources call them “an immense horde”), the fact that such a great mass was able to defy their Hun lords and break through to the Danube is most significant.

In 381, five years before, a few Huns had joined the Sciri and Carpodacians on a quick looting expedition. This time it was a whole people, led by an Ostrogothic prince,[216] that threw off the Hunnic yoke. No wonder that Cassiodorus-Jordanes ignores the trek of the Greuthungi: The other Ostrogoths, those who followed the Amalungs, Cassiodorus records, did not dare to rise against the Huns. Unfortunately, we know nothing about the circumstances under which the Greuthungi were able to escape the Huns. There might have been dissension among their masters; perhaps those Huns who ruled over the Greuthungi were engaged in a looting expedition in the north. But the fact remains that many thousand of the “human cattle” broke through the Hunnic fences. Hun power in the plains north of the lower Danube was still not firmly established.

Hunnic Horsemen Ride to Gaul

The situation at the borders of Pannonia and in the plain east of the Danube remained fluid also. Only a small part of the Sarmatians made peace with the Romans. The war with the others lasted throughout 383.[217] Whether the victory that Valentinian’s troops[218] won over the elusive enemy in the spring of 384 was as decisive as it looked to the spectators in the Colosseum[219] in Rome is rather doubtful. The continuous attempts of the Sarmatians to cross to the right bank of the Danube have their parallel in the migration of Odotheus’ Greuthungi; they, too, seemed to have tried to shake off the Huns and find new pastures. Of the Huns themselves we get only a glimpse.

In the spring of 384, Hunnic horsemen rode through Noricum and Raetia towards Gaul, allies of the legitimate ruler, barbarians thrown against barbarians, called forth from their tents in the East as they were to be called so often afterward. The only source for the first appearance of the Huns in western Europe is a short passage in a letter of Bishop Ambrose to Valentinian II.[220] It is not easy to date. Ambrose alludes to events of which we know little or nothing. Yet in view of the absence of any other information about the Huns in those years, even the smallest bit of information is of value.

On his return from Trier to Milan in December 383,[221] Ambrose met in southern Gaul the troops of the usurper Maximus. They were on the march to occupy the passes over the Maritime Alps and the blocks along the Riviera. In Italy Ambrose saw the imperial army on its way in the opposite direction with the same destination. In the four months that had elapsed since Gratian was murdered, Maximus had made himself the undisputed master of Gaul; he could have invaded Italy anytime, and would not have hesitated could he have been sure that he had to fight there only the troops of Gratian’s little brother Valentinian or, rather, of Bauto, his Frankish generalissimo.

Bauto was an experienced and resourceful soldier but his troops were few and, except for the Gothic mercenaries, not reliable. On the one side stood Maximus, a most orthodox man; on the other, the Arian empressdowager Justina—the boy Valentinian did not count—and the pagan Bauto. When four years later Maximus marched into Italy, he met practically no resistance. Bauto’s army would have fought better in 383 and 384, before Justina began to “persecute” the orthodox majority of her subjects, but it almost certainly would have been defeated had Maximus decided to march. It was only the fear of Theodosius, ruler of the East, that held Maximus back. It was only the hope for help from the East that kept Bauto up. Maximus knew that an attack on Italy meant war with Theodosius. Bauto displayed all his forces along the western frontier; their task was to hold out as well as they could until Theodosius’ armies joined the battle.

Maximus did strike, but not at Italy. He instigated the Juthungi to reassume their raids into Raetia.[222] Still suffering from their defeats in 378 and 379, kept in check by the greatly strengthened garrisons along the limes Ralic’is,[223] the Juthungi did not move until the summer of 383. At that time, when a terrible famine hit a vast part of the Western empire, and particularly Italy,[224] “the second Raetia learned the danger of her own fertility. For being used to security from her own poverty, she drew an enemy on herself by her abundance.”[225] The invaders were the Juthungi. Gratian was about to march against them when the greater danger in the West forced him to leave the defense of the province to the troops stationed there and throw the mobile army into Gaul to stop Maximus.[226]

In the first month of 384, the Juthungi were preparing a new attack. It is unlikely that Maximus concluded a formal alliance with the barbarians; all they needed was the consent, perhaps even only the tacit consent, of Maximus to the invasion of Raetia. If they pressed the attack, if they crossed the Alpine passes, Bauto was lost. Maximus could just walk into Italy, not as aggressor but as savior of the Roman world from the barbarians.

It was then that Bauto turned to the Huns and Alans.[227] From Ambrose’s letter we learn nothing about the strength of the Hunnic and Alanic cavalry, the men who led them, the battles they fought. He speaks only in passing about their triumphs. It seems that they crushed the Juthungi in one great sweep. Their task was fulfilled. The Juthungian danger was removed. The Huns could return to their country.

But they did not return. They kept riding west, “approaching Gaul” (appropinquantes Galliae). When the news reached Milan, Bauto must have been horrified. Athough Theodosius had decided to defend Italy, he was anything but willing to assist Bauto in an attack on Maximus. If the Huns, Bauto’s allies, broke into Gaul, Maximus must take this as an open declaration of war. They had to be stopped, and they were. Bauto purchased the retreat of the federates with gold.[228] Again we are not told how much he paid them, but it may be assumed that they were richly compensated for the loss of booty they could have expected to make in Gaul. The Huns turned and rode home.[229]

In the history of the late Roman Empire all this would not deserve more than two lines; but for the study of the Huns the episode of 384 is of considerable importance. We can draw from it the following conclusions:

In one passage Ambrose names the Huns first, the Alans second, and in another one only the Huns, so the Huns were apparently not only the stronger but also the dominating group.

The Huns to whom Bauto turned for help cannot have lived deep in the barbaricum, far to the east. If their sites were not already west of the Danube, which is possible, they must have lived along or very close to the left bank of the river. As early as 384, large tracts of the Hungarian plain were held by the Huns and their Alanic allies.

The ductus of the Hun primates was not tumultuarius. As in 378, they made an agreement with a non-Hunnic power; they assembled the horsemen, this time many more than in 378; they led them hundreds of miles through unknown lands. It would be absurd to suppose that Bauto’s emissaries paid each Hun so and so many solidi. The gold was received by the Hun leaders. How they distributed it among their followers we do not know. But that they could keep their promise to ride back, although the temptation for a good number of the barbarians to take the money and continue looting must have been great, proves that the horsemen were firmly in their hands. These leaders, whatever their position, were men of authority.

Our information about the Huns, both west and east of the Carpathians, after 386 is even scantier than what we could extract from the very few sources so far. All we have are brief allusions in poetical works.

When in the summer of 387 Maximus offered to send a body of troops[230] from Gaul to Italy to assist Valentinian against the barbarians who were threatening Pannonia,[231] the situation along the middle Danube must have been very serious. Only the danger that the frontier defense might collapse completely and the barbarians pour into Italy itself could compel Valentinian, who had all the reasons to mistrust the unexpected readiness of his brother’s murderer to help him, to accept the offer. Within a few weeks the “auxiliary” troops were, indeed, followed by Maximus’ whole army, and Valentinian had to flee to Constantinople.

Zosimus, the only source for these events, wrote what his public expected from him. He did not say who the enemies were, where they attacked, and what the outcome of the fighting was. His readers were interested only incidentally in history; they wanted to hear court gossip and malicious anti-Christian anecdotes. Neither did the pious crowd which filled the cathedral in Milan care who were the savages against whom the soldiers of their emperor or, for that matter, those of the other one in Gaul were fighting. In his sermons at Whitsuntide 387, Ambrose called them simply barbarus hosiis.[232] Fortunately, Pacatus is, though in a roundabout way, very explicit.

As is known, his Panegyric on Theodosius is the main source for the campaign against Maximus in 388. The army that the emperor assembled consisted almost wholly of barbarians. Theodosius made careful diplomatic and military preparations; the peace with Persia was renewed,[233] the Saracens were appeased.[234] Theodosius “accepted the barbarian peoples who vowed to lend him their help as fellow combatants.”[235] In concluding alliances with them, he not only removed the threat to the frontiers, he also increased the strength of his forces sufficiently to avoid the need to draft Roman citizens.

The barbarian horsemen fought magnificently. This was to be expected. But what surprised all who knew their barbarians was the exemplary discipline they held. “The army”—it is Christ who addresses the emperor[236]—“gathered from many unsubdued nations, I bade to keep faith, tranquillity, and concord as if of one nation.” Pacatus has nothing but praise for the allies:

0 memorable thing: There marched under Roman leaders and banners as Romans those who before had been our enemies, following the signs against which they had stood, and as soldiers filled the cities of Pannonia which they had emptied with fiendish devastation. Goths and Huns and Alans answered the roll call, changed guards, and rarely feared to be reprimanded. There was no tumult, no confusion, no looting in the usual barbarian way.[237]

In another passage Pacatus refers to the allies as barbarians who came “from the threatening Caucasus and the iced Taurus and the Danube which hardens the gigantic bodies.” The last ones are evidently the Goths. Causasus and Taurus are not the mountains from which the Huns and Alans descended to join Theodosius but their original homes “somewhere in the east.”[238]

Theodosius marched from Thessalonica up the Vardar and Morava valleys to Singidunum (modern Belgrade) and from there westward along the Sava to Siscia (modern Sisak, Yugovlavia), where he inflicted the first defeat on Maximus’ troops. The second battle took place near Poetovio (modern Ptuj, Yugoslavia). The road from Singidunum via Siscia to Poetovio leads through Pannonia secunda and Savia. The towns which the Goths, Huns, and Alans raided before 388 were in those two provinces.

It is most unlikely that Valeria had been immune to their inroads. In 387, the barbarians must have penetrated deep into Pannonia prima. Ambrose would not have spoken about a few marauders at the Danube; they would not have prompted Valentinian to accept Maximus’ help.

Pacatus’ testimony bears out the conclusions drawn here from Ambrose’s letter: Eastern Hungary was Hun land. It certainly was not one great pasture for the herds and flocks of the Huns alone; there were also Alans and Goths, allied or subject to the Huns, Jazygian Sarmatians, Germanic tribes, and the aboriginal Illyric population. But the Huns were the lords.

If in 388 Huns fought for the Romans, four years later Hunnic horsemen ravaged again the unfortunate Balkan provinces. From Claudian’s In Rufinum and his Panegyric on Stilicho’s Consulship, we learn that Huns crossed the Danube and joined the German enemies of the Romans. Claudian’s poems, the one a vitriolic invective, the other a hyperbolic eulogy, are not exactly reliable sources for the dark period that followed Theodosius’ victory over Maximus. Still, Claudian is a paragon of exactitude compared with Zosimus, whose anecdotic account permits the reconstruction of the events of those years barely in their broadest outlines.

A good number of barbarians, apparently mainly Visigoths, deserted the imperial standards on the eve of the campaign in 388 and turned robbers. For almost four years they terrorized Macedonia, pillaging farms, investing highways, swiftly rushing out from their hiding places in the swamps and forests and as swiftly disappearing “like ghosts.”[239] Their ranks, swelled by more deserters after the end of the war in Italy, grew into large and well-organized bands, like the Vargi and Scamarae half a century later. In the summer of 391, the situation became so desperate that Theodosius granted civilians the right of using arms against the brigands,[240] a bold measure when one considers how easily the miners and other proletarians could have joined the bands as they had joined the Goths in 378.

In the fall the emperor himself took the field. Already the first encounters proved that the local forces were insufficient; after a severe defeat in which he almost lost his life, Theodosius called in reinforcements from the army in Thrace. The result was that large hordes of transdanubian barbarians broke through the limes and poured deep into the plain north of the Haemus (Balkans). What until then was a punitive expedition, though on a great scale, became a horrible war.[241] Jerome was not sure that in the end the Goths might not conquer.[242] John Chrysostom’s letter to a young widow gives an idea of the magnitude of the catastrophe that befell Thrace. He consoled her by pointing out how much more miserable women like the empress were. Theodosius’ wife

is ready to die of fear, and spends her time more miserably than criminals condemned to death because her husband ever since he assumed the crown up to the present day has been constantly engaged in warfare and fighting.... For that which has never taken place has now come to pass; the barbarians leaving their own country have overrun an infinite space of our territory, and that many times over, and having set fire to the land, and captured the towns, they are not minded to return home again, but after the manner of men who are keeping holiday rather than making war, they laugh us all to scorn. It is said that one of their kings declared that he was amazed at the impudence of our soldiers, who although slaughtered more easily than sheep still expect to conquer, and are not willing to quit their own country, for he said that he himself was satiated with the work of cutting them to pieces.[243]

Theodosius returned to Constantinople in 391, “so depressed at what he and his army had suffered from the barbarians in the marshes that he decided to renounce wars and battle, committing the management of those affairs to Promotus.”[244] The experienced general had no better luck. Whether the enemy was actually as strong as Claudian indicates is not known. He never gives numbers in his poems; instead he heaps names upon names. In the invective against Rufinus Claudian lists Getae, Sarmatae, Daci, Massagetae, Alani, and Geloni,[245] in the Panegyric on Stilicho, written three years later, Visi, Bastarnae, Alani, Huns, Geloni,

Getae, and Sarmatae.[246] Promotus was killed in an encounter with the Bastarnae. Stilicho, his successor, is said to have scattered the Visigoths, and overthrown the Bastarnae;[247] he would have annihilated the barbarian hordes, penned in the limits of a small valley, “had not a traitor [Rufinus] by a perfidious trick abused the emperor’s ear and caused him to withhold his hand; hence the sheathing of the sword, the raising of the siege, and the granting of treaties to the prisoners.”[248]

Rufinus acted as did Stilicho three years later and again in 402 when he made a compact with the Visigoth king Alaric and allowed him to withdraw. What Claudian said in praise of Stilicho, he could have said about Rufinus: “Concern for thee, 0 Rome, constrained us to offer a way to escape to the beleaguered foe lest, with the fear of death before their eyes, their rage should grow the more terrible for being confined.”[249] The “prisoners” with whom Rufinus, clearly with the consent of Theodosius if not at the emperor’s direct instructions, concluded alliances were Goths and Huns.[250] What the conditions of the foedera were, Claudian does not say. But many of the Huns did not ride back to their tents across the Danube; they stayed, as we shall see, in Thrace.

In the summer of 394, Theodosius again led an army against an usurper in the West, Eugenius. It was at least as strong as the one with which he had taken the field in 388. “The fortunes of Rome stood at a razor’s edge.”[251] It was not, as six years earlier, a war between the legitimate ruler and an usurper; it was a war between Christ and Jupiter, the monks of the Thebais and Etruscan augurs, the God-loving East and the idolworshippers of the West. Eugenius fought for the gods, and the gods fought for him. His soldiers carried on their standards the picture of Hercules Invictus[252] and on the height of the Julian Alps stood golden statues of

Jupiter,[253] ready to throw their thunderbolts at the Galilaeans should they dare to approach the sacred soil of Italy. In Rome, Nicomachus Flavianus, the leader of the turbulent pagan revival, read the coming victory of Eugenius in the entrails of the sacrificed bulls;[254] in Constantinople, Theodosius waited anxiously for an answer from the prophetic hermit John of Lycopolis as to whether he or the godless tyrant would win the war.[255] He prayed and fasted. “He was prepared for war not so much with the aid of arms and missiles as of fasts and prayers” (Praeparatus ad bellum non tamen armorum talorumque quam ieiuniorum orationumque subsidiis), said Rufinus,[256] and all Christian authors are agreed that it was the power of God which granted Theodosius the glorious victory over the pagans. Ambrose compared him with Moses, Joshua, Samuel, and David.[257] Yet when the emperor finally went to war he did not carry a sling; he marched at the head of a huge army.

Theodosius busied himself through the winter of 393/4 with elaborate military preparations.[258] His recruiting officers in the East enlisted Armenians, Caucasian mountaineers, and Arabs. The Visigothic allies were ordered to furnish as many troops as they could. Even if those did not number more than twenty thousand, as Jordanes asserts,[259] they must have formed a large contingent.[260] Alans came, led by Saul,[261] whom we shall meet soon again. And then came, to strengthen God’s warriors, “many of the Huns of Thrace with their phylarchoi.”[262]

The chronicler John of Antioch is the only one to mention the Huns. It is understandable that the church historians passed them over in silence; they were not interested in the composition of the auxiliaries.[263] That Jordanes spoke only of the Goths is in no way remarkable. But the absence of the Huns from the long list of peoples in Claudian requires an explanation.

The poet names Arabs, Armenians, Orientals from the Euphrates, Halys and Orontes, Colchi, Iberians, Medes from the Caspian Sea, Parthians from the Niphates, and even Sacae and Indians.[264] He mentions the Goths and, by a circumscription,[265] the Alans. But the Huns do not exist for him, although he must have known that they fought for Theodosius. He may barely allude to them by listing the Geloni among the auxiliaries.[266]

One could think that by ignoring the Huns Claudian expresses his abhorrence of those lowest of the barbarians, his reluctance to give them any credit for the victory of the good cause. But I believe the close relationship between the Huns and the hated Rufinus was the real, or at least, the stronger motive. It is true that Claudian depicts Rufinus in the blackest colors as the devoted friend of the Goths. But when Stilicho, at Rufinus’ orders, had to give up the command of the Eastern troops, these were not, as one would expect, afraid that now the Goths would be their masters. They feared, rather, that Rufinus would make them “the slaves of the foul Hun or the restless Alan.”[267] This is strange. The only explanation of which I could think would be Rufinus’ decision to rely on the Huns and Alans to counterbalance the power of the Goths. It would have been not the most pleasant, but certainly the most efficient means. A few years later the anti-Gothic faction in Constantinople played, indeed, with the idea of allying itself with the Huns against the Goths, the wolf against the lion.[268] I suspect that Rufinus had the same intention. It cannot be a coincidence that in the autumn of 395 he had a Hunnic, not a Gothic, bodyguard; only after they were cut down to the last man could General Gainas’ soldiers kill him.[269]

That he gave them land in Thrace points also to a most unusual and close relationship between Rufinus and the Huns. This is the only time that Huns were admitted into the empire. All other alliances with the Huns were concluded with tribes or tribal coalitions in the barbaricum. The Huns in Thrace must have numbered several thousand, for it is most unlikely that the Hun warriors, made Roman federates, should have been willing to live without their wives and children, herds of cattle, flocks of sheep, and their carts, which they obviously did not take with them when they broke into Thrace. They must have sent for them.

John of Antioch’s explicit statement that the Huns lived under phylarchoi allows us also to draw some conclusions as to their political organization. In the usage of the Byzantine writers the term (pv2.aQ%o<; is not sharply defined; it is interchangeable with yyep,(ov, yyovpevo:;, aQxcov, and even /taodev?. Phylarchos means the leader of any larger group; the phyle can be a tribe, comprising a number of clans, a multitude of tribes, or a whole people. If the Huns in Thrace had a king, a ruler over the phylarchoi, John could not have failed to say so. Their phylarchoi were almost certainly tribal leaders. But from this it does not necessarily follow that the Huns beyond the Danube were likewise divided into independent tribes without a common leader. It is conceivable that those Huns who allied themselves with the Romans did not want to submit to a ruler over them. In any case, there evidently was no Hun ruler in the 390’s strong enough to enforce his will on all tribes, to prevent Hun groups from waging their own wars and making their own peace. Those in the Hungarian plain pillaged Pannonia, those in Rumania Thrace; they concluded alliances and broke them at their, not a king’s, pleasure. This did not exclude the possibility of concerted action of groups of Huns on a large scale. Such was the great raid into Asia in 395.

The Invasion of Asia

In the summer of 395, large hordes of Huns crossed the Don near its mouth, turned southeast, and broke through the Caucasus into Persia and the Roman provinces to the south and southwest of Armenia.

One group devasted the country south and west of the Anti-Taurus. When they crossed the Euphrates, the Romans attacked and destroyed them. Another group, led by Basich and Kursich, rode down the valleys of the Tigris and Euphrates as far as Ctesiphon. On the report that a Persian army was on the march against them, they turned back but were overtaken. One band was cut down; the other, leaving their prisoners behind, fled through Azerbaijan and returned over the Caspian Gates to the steppes. A third group ravaged eastern Asia Minor and Syria.

In the following year the East was trembling with fear that the Huns, this time as the allies of the Persians, would come back. But the danger passed, possibly because the Romans came to an agreement with the Persians. When in 397 a few Hun hordes broke once more into Roman Armenia, they were easily driven back.

The cause of the invasion in 395 is said to have been a famine in the country of the Huns. Indeed, they drove away as many herds of cattle as they could. But first of all they made thousands of prisoners. The raid became a gigantic slave hunt.

These are, in broad outlines, the events. Instead of referring to the texts in footnotes, which themselves would require more notes, I shall discuss the various topics and problems one by one, incorporating the material that ordinarily would go into annotations.

The Sources

The sources flow so copiously that there is no need to make use of works of doubtful value as, for example, The Life of Peter the Iberian.[270] Except Theodoret (see below), the Greek and Latin sources[271] are adduced by all standard works, but most of the information contained in Syriac literature has been disregarded. I refer to the legend of Euphemia and the Goth,[272] a mamre (poem) of Cyrillonas (fl. ca. 400),[273] John of Ephesus ca. 507–586) ;[274] and the Liber Chalifarum[275] In various respects they complement the Western sources. Some texts have been misunderstood and misinterpreted with the result that Hunnic history has been strangely distorted. Two examples will suffice.

The Arian historian Philostorgius (368 to after 433) begins his fairly detailed description of the Hunnic invasion of Asia in 395 with a brief summary of the earlier history of the people: “They first conquered and laid waste a large part of Scythia, then crossed the frozen Danube and, swarming over Thrace, devastated the whole of Europe.”[276] These lines have been quoted as referring to a Hunnic invasion of Thrace in the same year.[277] Actually, Philostorgius telescoped three or more decades, from the Hunnic victory over the Goths to the repeated incursions into the Balkan provinces. The poet Claudian, too, is supposed to have described in the Invective against Rufinas an invasion of Europe by the Huns in 395. But the barbarians who devastated “all that tract of land lying between the stormy Euxine and the Adriatic” were Goths, Geticae cavernae.[278] Of the church historians, neither Socrates nor Sozomen[279] mentions a Hunnic invasion of Thrace or any other province of the Balkans in 395.[280] The Eastern sources, though mainly concerned with the events in the Orient, know nothing of Hun raids into Thrace, not to speak of the “devastation of the whole of Europe.”

Another often misunderstood passage occurs in Priscus’ account of the East Roman embassy to Attila’s court, Excerpta de legationibus Romanorum ad gentes (cited as EL), 46. In a conversation between the envoys from Rome and Constantinople, the West Roman Romulus spoke about Attila’s ambitious plans:

He desires to go against the Persians to expand his territory to even greater size. One of us asked what route he could take against the Persians. Romulus answered that the land of the Medes was separated by no great distance from Scythia and that the Huns were not ignorant of this route. Long ago they had come upon it when a famine was in their country and the Romans had not opposed them on account of the war they were engaged in at that time. Basich and Kursich, who later came to Rome to make an alliance, men of the Royal Scythians and rulers of a vast horde, advanced into the land of the Medes. Those who went across say that they traversed a desert country, crossed a swamp which Romulus thought was the Maeotis, spent fifteen days crossing mountains, and so descended into Media. A Persian host came on them as they were plundering and overrunning the land and, being on higher ground than they, filled the air with missiles, so that, encompassed by danger, the Huns had to retreat and retire across the mountains with little loot, for the greatest part was seized by the Medes. Being watchful for the pursuit of the enemy, they took another road, and, having marched ... days[281] from the flame which rises from the stone under the sea, they arrived home.

The scribes, who made the excerpts, shortened the text, as they, incidentally, also shortened the immediately following story of the discovery of Ares’ sword, much better preserved in the Getica. It is unlikely that Romulus merely said that the Romans did not oppose the Huns “because of the war they were engaged in at that time.” He must have been more specific. And why should the Romans have opposed the Huns if their goal was Media? Evidently, Romulus spoke also about the Hun incursions into Roman territory, but the scribes omitted everything that had no immediate bearing on the invasion of Persian lands.

A comparison between Priscus and the Liber Chalifarum shows that both sources deal with the same invasion.

Priscus: “When the Persians counterattacked, the Huns retreated. The greater part of their loot was seized by the Medes.”

Liber Chalifarum: “When the Huns learned that the Persians advanced against them, they turned to flight. The Persians chased them and took away all their loot.”

Priscus is also in agreement with Jerome:

Priscus: “The Romans did not oppose them on account of war they were engaged in at that time.”

Jerome, speaking of the Hun invasion in 395: “At that time the Roman army was away and held up by a civil war in Italy.”

The war was the struggle between Stilicho and Rufinus in 395 (see Chapter XII). The Huns broke into Asia while the greater part of the Eastern army stood in Italy or was on the march to Illyricum; it did not return to Constantinople and Asia Minor until the end of November.

It is hard to understand how in spite of their preciseness the texts could have been so often and so strangely misunderstood. Bury identified the Huns with the Sabirs,[282] Demougeot with the Hephthalites.[283] Thompson dates the invasion of the Priscus account to 415–420;[284] Gordon, at least recognizing that the war in which the Romans were engaged had to be dated, decided on the one in the years 423–425.[285] That the leaders of the Huns who came to Rome to conclude an alliance were the same who rode to the Tigris proves that their sites were in Europe. The Hunnish federates of the Romans were not Huns in Dagestan or the Kuban region; Aetius’ friends lived on the Danube.

Basich and Kursich may have come to Rome in 404 or 407. Emperor Honorius was in Rome from February to July 404; two years later Stilicho defeated Radagaisus with the help of Hunnic auxiliaries. Except for the month of February, Honorius was again in Rome throughout 407, where he stayed until May, 408.[286] In 409, Huns served in the Roman army.

The Chronicle of Edessa gives the most exact date: “In the year 706, the month tammuz (July 395), the Huns reached Osroene in northern Mesopotamia.”[287] They waged a veritable Blitzkrieg, so they cannot have crossed the Caucasus much earlier. The years in the Syriac sources vary slightly,[288] but the texts agree in the main. “In the days of the emperors Honorius and Arcadius, the sons of Theodosius the Great, all Syria was delivered into their [i.e., the Huns’] hands by the treachery of the prefect Rufinus and the supineness of the general Addai.”[289] “But the Romans killed Rufinus, the hyparch of the emperor, while he was sitting at the feet of the emperor, for his tyranny was the cause of the coming of the Huns.”[290]

They [i.e., the Huns] took many captives and laid waste the country, and they came as far as Edessa. And Addai, the military governor [stratelates] at that time, did not give permission to the federates to go out against them because of treason in their midst.[291]

The rumor that Rufinus let the Huns into the empire was as current in the East as it was in the West. Rufinus was killed on November 27, 395. Addai (Addaeus), comes et magister utriusque militiae per orientem, is last named in an edict issued to him on October 3, 395.[292]

In 396, a new Hun invasion seemed to be imminent. “After a little while the Goths came again to Edessa with a certain general who had been sent by the emperor to his place to keep it from the enemies, the Persians, I mean, and the Huns, who had agreed to make war on this country.”[293] Claudian, too, alluded to a threatening war with the Persians,[294] but did not mention the Huns as their allies. We learn more about the feelings of the Syrians from the moving mamre of Cyrillonas:

Every day unrest, every day new reports of misfortunes, every day new blows, nothing but fights. The East has been carried into captivity, and nobody lives in the destroyed cities. The West is being punished, and in its cities live people who do not know Thee. Dead are the merchants, widowed the women, the sacrifices have ceased... the North is threatened and full of fight. If Thou, 0 Lord, doest not intervene, I will be destroyed again. If the Huns will conquer me, oh Lord, why have I taken refuge with the holy martyrs? If their swords kill my sons, why did I embrace Thine exalted cross? If Thou willst render to them my cities, where will be the glory of Thine holy church? Not a year has passed since they came and devastated me and took my children prisoners, and, lo, now they are threatening again to humiliate our land. The South is also being punished by the cruel hordes, the South full of miracles, Thine conception, birth and crucifixion, still fragrant from Thine footsteps, in whose river Thou wert baptized, in whose siloe Thou hast cured, in whose jars was Thine precious wine, and in whose laps Thine disciples lay at the table.[295]

There was no other invasion of Syria in 397 as Claudian, against his better knowledge, asserted.[296] He simply transferred the events of 395 to 397, equating the hated eunuch Eutropius with the equally hated Rufinus. No Greek or Syrian writer knows of a second coming of the Huns. Eutropius fought some barbarian hordes, among whom there may have been Huns, in the Caucasus.[297]

The Course of the War

If Claudian is to be believed, the Huns crossed the Caucasus over the Caspia claustra,[298] the Darial Pass; he adds: inopino tramite, “a pass where they were not expected,”[299] because the northern barbarians came, as a rule, over the pass of Darband.[300] It is difficult to determine how far the Huns penetrated into Asia Minor, Syria, and western Persia.

Socrates, Sozomen, and some Syriac sources describe the theater of the war in general terms: Armenia and other provinces of the East; Syria and Cappadocia; all Syria. In his commentary on Ezekiel SSdO-^,[301] probably written before 435,[302] Theodoret wants to prove that Gog and Magog, whom he identifies with the Scythian peoples, live not far from Palestine. He reminds his readers that “in our times the whole Orient was occupied by them.” The Scythians are the Huns, as in Jerome. They made war on the Phrygians, Galatians, Iberians, and Ethiopians. The first three names stand for Ooyagpa, Fop^Q, and OofteX in the Septuaginta as interpreted by Josephus.[303]

Philostorgius is more specific: The Huns broke through Greater Armenia into Melitene, reached from there Euphratesia, riding as far as Coelesyria.[304] Claudian speaks of Cappadocia, Mount Argos, the Halys River, Cilicia, Syria, and the Orontes. Jerome names the cities on the Halys, Cydnus, Orontes, and Euphrates.[305] The Huns came as far as Antioch and Edessa.[306]

Two Syriac sources give more details. There are, first, the excerpts from the Ecclesiastical History of John of Ephesus:

In the same year the Huns invaded the country of the Romans and devastated all regions of Syria along the Cahja mountains, namely Arzon, Mipherqet, Amid, Hanzit, and Arsamisat.[307] When they had crossed the Euphrates, the bridge was cut off and the troops of the Romans gathered from various sides against them and annihilated them, and no one of the Huns escaped.

“Syria” here means Mesopotamia; the cities named are on and to the north of the upper Tigris. The author continues to describe how the Huns, by cutting the aqueduct, forced the people who had taken refuge in the fortress of Zijat to surrender; most of them were massacred, the rest led away into captivity.

The Liber Chalifarum gives the following account:

In this year the cursed people of the Huns came into the land of the Romans and ran through Sophene, Armenia, Mesopotamia, Syria, and Cappadocia as far as Galatia. They took many prisoners and withdrew to their country. But they descended to the banks of the Euphrates and Tigris in the territory of the Persians and came as far as the royal city of the Persians. They did no damage there but devastated many districts on the Euphrates and Tigris, killed many people and led many into captivity. But when they learned that the Persians advanced against them, they turned to flight. The Persians chased them and killed a band. They took away all their plunder and liberated eighteen thousand prisoners.

In the history of the Huns the invasion of Asia was an episode, though an important one. Three things can be learned from it. First, it shows what great distances the Huns were able to cover in one campaign, something often overlooked in the historical interpretation of isolated Hunnic finds. Second, the Huns carried many young people, “the youth of Syria,”[308] into captivity. Although this could have been surmised, the explicit testimony of the texts is definitely welcome. Third, there are a few lines in Theodoret which, as the whole text, have been ignored by all students of the Huns. According to Theodoret, many people in the regions overrun by the Huns joined them. Some were forced; we may assume that they had to do slave labor, collecting fuel, attending to the more unpleasant jobs in the households of the upper-class Huns, and so forth. But others ran over to the Huns and fought voluntarily in their ranks. Theodoret did not paraphrase Ezekiel, nor did he interpret the words of the prophet; Ezekiel did not say that the Israelites would join the armies of Gog and Magog.

Theodoret’s source is unknown. He was a small child when the Huns came dangerously close to Antioch, his birthplace.[309] What he says about the flight to the Huns he may have heard from older people. At any rate, it is most remarkable. I shall come back to it in another context.


After the shadowy Balamber,[310] Uldin is the first Hun mentioned by name. The literary evidence contains enough material for a picture, if not of the man, of his deeds. We know when and where he led his Huns into battle, and we even get a glimpse of the happenings in Hunnia.

In 400, Uldin was the ruler of the Huns in Muntenia, Rumania east of the Olt River. When Gainas, the rebellious former magister militum praesentalis, and his Gothic followers fled across the borders (see Chapter XII), Uldin “did not think it safe to allow a barbarian with an army of his own to take up dwellings across the Danube.” He collected his forces and attacked the Goths. The short but sanguinary campaign ended with a Hunnic victory. Gainas was killed.[311] Because only eleven days later[312] his head was displayed in Constantinople,[313] the last fight probably took place near Novae, the place at the Danube nearest to the capital, connected with it by a first-rate road.[314]

Gainas wanted to join his countrymen; he fled “to his native land” (sIq rd oi’xeia).[315] It follows that in Muntenia Goths lived under Hun rule. We do not know how far to the east and north Uldin’s realm extended. In the west his power reached to the banks of the Danube in Hungary, which is evident from the alliance he concluded with the West Roman generalissimo Stilicho in 406.[316]

At the end of 405,[317] Italy, barely recovering from the first Gothic war, again was invaded by Goths. Under their king Radagaisus the barbarians descended on Venetia and Lombardy,[318] overran Tuscany, and were nearing Rome when they were finally stopped. The regular Roman army was too weak to stem the Germanic flood. Stilicho turned to Uldin the Hun, and Sarus the Goth, for help. Near Faesulae the Hun auxiliaries encircled a large part of Radagaisus’ hordes;[319] he tried to escape but was captured and executed (April 406). The survivors were sold as slaves.[320] What happened to those Goths who had not been with Radagaisus is not known. Some seem to have been enrolled in Stilicho’s army,[321] others may have fought their way back to their transdanubian homes. The Gothic nation was “forever” extinguished. At least this was to be read on the triumphal arch erected in 406,[322] just four years before Alaric took Rome.

It has often been assumed that the Gothic invasion was a repetition of the events in the 370’s. The Goths of Radagaisus are supposed to have fled from the Huns who themselves were pushed westward by other nomadic groups which, in turn, were set in motion by an upheaval in the Far East. It is the well-known billiard ball theory, the primum movens being hidden, “in the vast plains of Eurasia.” Nothing in our authorities indicates that behind Radagaisus stood another barbarian leader whose people were pushed by still another one, and so on.[323] All we know is that the Goths came from the countries across the Danube.

If they actually were fleeing, it was not a headlong flight. Although the figures in Orosius and Zosimus are grossly exaggerated,[324] we may believe that Radagaisus led, indeed, a large army into Italy.[325] The Gothic warriors were not raiders; they were the armed part of a people on the trek to a new home. From the fact—if it is a fact—that Radagaisus was a pagan,[326] some scholars have concluded that his hordes were Ostrogoths, because by 400 all Visigoths are supposed to have been good Christians. But the Visigoth Fravittas, consul in 401, East Roman general, was a staunch pagan, and among the Visigoths beyond the Roman border there must have been many thousands not yet baptized.[327] Besides, a little-noticed entry in the Chronicle of 452 proves that there were Arian Christians among the Goths of Radagaisus.[328] Putsch might well have been right in assuming that a good part of them came from Caucaland.[329]

There is no reason to assume that Stilicho’s Hunnic auxiliaries came from far away, or, specifically, from the Dobrogea.[330] Huns had camped in Hungary since 378. They are, as we saw, well attested there in the middle 380’s. They certainly did not voluntarily give up the land, and no enemy was strong enough to drive them out. Stilicho concluded an alliance with the Huns in Hungary. Uldin was king of the Huns to the west and to the east of the Carpathian Mountains, in the Alfold as well as in Muntenia.

He was not the ruler of all Hun tribes; not even Attila at the height of his power was. But Uldin could throw his horsemen into Italy and Thrace. In the winter of 404/5 Uldin broke into the Balkan provinces. We read in Sozomen:

About this time the dissensions by which the church was agitated were accompanied, as is frequently the case, by disturbances and commotions in the state. The Huns crossed the Ister and devastated Thrace. The robbers in Isauria, gathered in great strength, ravaged the towns and villages between Caria and Phoenicia.[331]

When Sozomen interrupts his narrative of the synods, elections of bishops, and the fights between the various cliques at the metropolitan sees to deal with secular events, he treats them, with rare exceptions, only as they have a bearing on the never-ending struggle between orthodoxy and heresy. The dates of the ecclesiastical history are given as precisely as possible; political events take place “about the same time.” Still, I think Uldin’s first invasion of Thrace can be dated fairly well.

The “dissensions” were the fights of the patriarch of Alexandria Theophilus (384–412) against John Chrysostom. Chapters 20 to 24 of Book VIII cover the period from the autumn of 403 to November 404.[332] In chapter 26 Sozomen gives the translation of the letters which in the fall of 404 Pope Innocent sent to John.[333] In chapter 27 he mentions the death of Empress Eudoxia (October 6, 404), the death of Arsacius (at the end of 405),[334] and the ordination of Allicus, his successor (late in 405, or in 406).[335] Therefore, the invasion of Thrace falls somewhere between 404 and 405. I believe it can be dated even more precisely. From John Chrysostom’s letters we know that the Isaurians broke out of the valleys of Mount Taurus in the summer of 404, probably in June.[336] They were soundly defeated.[337]

In the following year they repeated their raids, this time extending their ravages over nearly the whole of Asia Minor.[338] In 404, the Isaurians were unable to take walled towns,[339] so the conquests of both towns and villages, of which Sozomen speaks, must fall in the year 405. The transdanubian barbarians used to cross the river in winter, when the fleet was immobilized and they could recross while it was still frozen. All these considerations lead to the winter of 404/5 as the most probable date of the Hun invasion of Thrace.

Sozomen is the only early writer to mention it. The account of Nicephorus Callistus (1256–1311) is a paraphrase, but one with a notable exception: He gives the name of the Hun leader—Uldin.[340] Nicephorus’ main source was probably a compilation of the tenth century, based on Philostorgius, Socrates, Sozomen, Theodoret, and Evagrius.[341] Which of these authors named Uldin cannot be determined. It may have been Philostorgius, of whose works we have only excerpts; it may have been Sozomen himself, because it is unlikely that the Sozomen text, as we have it, is word for word identical with the original. The possibility that Nicephorus himself supplied the name Uldin may be ruled out. He was too dependent on his sources to alter them; the best he could do was to dress up what others had written before him. Whatever Nicephorus’ ultimate authority, there was one in which Uldin was named as leader of the Huns 404–405.

Sozomen mentions the invasion only in passing. It may have been a quick raid, or the Huns may have been looting the unfortunate provinces for weeks or months. Still, it was in importance far surpassed by the one which, a few years later, carried Uldin’s horsemen deep into Thrace.

In the summer of 408, the Huns crossed the Danube.[342] As usual, well informed about the situation in the Balkans, they chose the right time to attack. In the spring of 408, Stilicho abandoned his plan to throw Alaric’s Visigoths into Illyricum. Shortly afterward they were on the march to Italy.

With the danger of a Gothic invasion over, the greater part of the East Roman troops was moved to the Persian frontier where hostilities were expected to break out any day.[343] The government in Constantinople was well aware that the transdanubian Huns might take advantage of the weakening of the Balkan army to make inroads into the border provinces. In April, 408, Herculius, praetorian prefect of Illyricum, was instructed “to compel all persons, regardless of any privilege, to provide for the construction of walls as well as for the purpose and transport of supplies in kind for the needs of Illyricum.”[344] If the Huns should by-pass the strong places along the limes, they could, for awhile, plunder the helpless villages, but eventually they would be caught between the unconquered towns in the interior and the troops holding out in the fortifications along the frontier, and forced back into the barbaricum. What the Romans could not expect was that the Huns would take the strategically important fortress Castra Martis in Dacia ripensis by treachery.[345] Whether other fortified places fell into the hands of the Huns is not known but is possible.

Our main source for Uldin’s second invasion is again Sozomen’s Ecclesiastical History. The other one, Jerome’s Commentary on Isaiah, has been ignored by all students of the Huns. Commenting on 7:20–21, Jerome wrote:

But now a large part of the Roman world resembles the Judaea of old. This, we believe, cannot have happened without God’s will. He does by no means avenge contempt of him by Assyrians and Chaldaeans, rather by savage tribes whose face and language is terrifying, who display womanly and deeply cut faces, and who pierce the backs of bearded men as they flee.

(Ac nunc magna pars Romani orbis quondam ludaeae similis est; quod absque ira Dei factum non putamus, qui nequaquam contemptum sui per Assyrios ulciscitur, et Chaldaeos: sed per feras gentes, et quondam nobis incognitas, quorum et vultus et sermo terribilis est, et femineas incisasque facies praeferentes virorum, et bene barbatorum fugientia terga confodiunt.)[346]

This was written in June or July 408.[347] That Jerome’s ferae gentes were the Huns is evident from their description: They were formerly unknown and they cut their faces because they wanted to look like women rather than men with beards. As I have shown elsewhere,[348] Jerome followed Ammianus’s description of the Huns. What matters here is the date of the passage in the commentary and in particular the phrase [ferae gentes] bene barbatorum fugientia terga confodiunt. If Jerome in faraway Jerusalem, as early as the summer of 408, received reports about the defeats of the Roman troops by the Huns, the losses must have been unusually heavy.

Even through Sozomen’s edifying account one senses how serious the situation must have been. With his few troops the Roman commander in Thrace could not drive the Huns back. He made peace propositions to Uldin, who replied by pointing to the rising sun and declaring that it would be easy for him, if he so desired, to subjugate every region of the earth enlightened by that luminary. But while Uldin was uttering such menaces and ordering as large a tribute as he pleased, and that on condition peace could be established with the Romans or the war would continue, God gave proof of his favor toward the present reign; for shortly afterward Uldin’s own people and captains, (olxetoi xai Ao/ayot) were discussing the Roman form of government, the philanthropy of the emperor, and the promptitude and liberality in rewarding the best men. Together with their troops, they seceded to the Romans, whose camps they joined. Finding himself thus abandoned, Uldin escaped with difficulty to the opposite bank of the river. Many of his troops were lost, and among others the whole of the barbarian tribe called the Sciri. This tribe had been strong in numbers before falling into this misfortune. Some of them were killed, and others were taken prisoners and conveyed in chains to Constantinople. The authorities were of the opinion that, if allowed to remain together, they might revolt. Some of them were, therefore, sold at a low price, while others were given away as slaves for presents on the condition that they should never be permitted into Constantinople or anywhere in Europe, but be separated by the sea from the places familiar to them. Of these a number were left unsold, and they were ordered to settle in different places. Sozomen had seen many in Bithynia, near Mount Olympus, living apart from one another and cultivating the hills and valleys of that region.[349]

Sozomen does not say when the war came to an end, but from an edict of March 23, 409, it can be concluded that by that time the Huns had recrossed the Danube.[350]

Sozomen’s account must not be taken literally, of course. The Sciri did not vanish from history.[351] But Uldin’s boast sounds genuine, and Sozomen doubtless correctly reports the content of Uldin’s demands. This is the first time our sources say something about the object of a Hun invasion. Uldin was not merely set on TrpatdEvetv,[352] “plundering,” and taking prisoners who could be sold as slaves. He did not demand the cession of Roman territory either. There were no pastures large enough for all the Huns under Uldin. If, however, some groups stayed in the empire, like those around Oescus, they would have been separated from the other tribes, and this was counter to Uldin’s interests. He rather demanded that the Romans pay him tribute, dactfiov, probably a fixed annual sum.

The Huns were mounted, the Sciri evidently mostly foot soldiers. The edict of April 12, 409,[353] provided only for the settlement of the Sciri. The Hun prisoners were either killed or drafted into the ranks of the auxiliaries. Who Uldin’s “own people” were is not quite clear; the word may mean nothing more specific than the people who usually stayed with him. The members of Belisarius’ olnia, of which in the sixth century Procopius speaks so often, were not necessarily his kinsmen. Paulus, for instance, who for a time was in charge of the obcla, was a Cilician;[354] Ataulf took over a man from Sarus’ olxla.[355] The word lochagos is not well defined either. That some lochagoi went over to the Romans together with their troops seems to indicate that a close bond existed between them and their followers. Ammianus and Orosius speak of the cunei of the Huns. Although cuneus, as used by them, is a tactical unit, the word may still have preserved some of the meaning it had in Tacitus: “Their squadrons or battalions, instead of being formed by chance or by a fortuitous gathering, are composed of families and clans.” (Non casus nec fortuita conglobatio turnam aut cuneum facit sed familiae et propinqui tales.)[356]

The cohesion of Uldin’s kingdom has been overrated,[357] but it should not be underrated either. Uldin was not the leader “of a mere fraction”[358] but of many tribes able to operate from the Rumanian plains to the Hungarian puszta. And yet, although the incipient royal power gradually was strengthened, it was by no means stabilized. How it weakened in Uldin’s last years becomes clear when we return to the West.

Shortly before Uldin’s Huns broke into the Balkan provinces, the Visigoths began the long trek which a century later ended in Spain. There is no need to recapitulate in detail the events preceding it; they have been thoroughly discussed by Santo Mazzarino in his masterful Stilicone. For our purposes a brief outline will suffice.

After the battle of Verona in the summer of 402,[359] Alaric led his hosts back to the Balkans. In the following three years he strictly kept his treaty with Stilicho. From the “barbarous region bordering on Dalmatia and Pannonia”[360] assigned to them, the Goths made occasional raids into eastern Illyricum,[361] but they were careful not to provoke a conflict with the West, partly because they had not yet recovered from their defeats, partly (and perhaps mainly) because they hoped to come to a closer and better agreement with Stilicho. In 405, he concluded, indeed, a foedus with Alaric, an alliance for the conquest of eastern Illyricum.[362] The Gothic king was promised the position of magister militum per Illyricum. He moved into Epirus where he stayed for three more years. First the invasion of Radagaisus, then the rebellion of Constantine in Britain forced Stilicho to postpone the Illyrian expedition, and finally the plan was dropped altogether.

Early in 408, Alaric turned against the West. By May[363] he had reached Noricum. Whether he encamped near Virunum, the present Maria Saal near Klagenfurt, or at Celeia[364] cannot be determined. What matters is that he passed through Emona (modern Ljubljana, Yugoslavia). The way from Epirus to Emona leads through Pannonia secunda and Savia.[365]

Of all the students of the Huns only Alfoldi realized that their inactivity in the eventful years 408–410 calls for an explanation.[366] As we shall see presently, they did not keep so quiet. But it is true that in 408 Alaric could march westward as if there were no Huns, those Huns who, as Alfoldi rightly stresses, were otherwise so eager to fish in troubled waters. Alfoldi assumes that they did not join the Goths because they were allied with the Romans. According to him, Stilicho settled them in 406 as federates in the province Valeria, the same year in which, presumably, young Aetius went as hostage to the Huns.[367]

If this assumption were correct, the Huns should have done more than stay away from the fight. They should have fought the Goths, attacking

them in the right flank while Alaric’s people, slowly traveling in their wagons, were on the move to Emona. But the Huns made neither common cause with the Goths nor did they fulfill their supposed obligations as allies of the Romans.

The reason for their inactivity is, in my opinion, much simpler. The Huns did not fight in Pannonia secunda and Savia because they fought under Uldin in Illyricum and Thrace. By dating Uldin’s invasion to 409 instead of 408,[368] Alfoldi had to find an explanation for something that does not need one.

This is not to say that the Huns had followed Uldin to the last horseman. There were Huns in the West Roman army under Stilicho, and Ravenna also had a Hun garrison after the execution of the great ductor in August 408.[369] Besides, many Huns must have stayed at home in order to prevent an uprising of their subjects while the “mobile” army was engaged in fighting south of the Danube. This was, I believe, an additional reason why they did not interfere in the war between Alaric and the Romans.

The Huns became active in the West only after Uldin’s hordes had returned to their sites beyond the Danube. How the defeat he had suffered undermined his authority can be deduced from two passages in Zosimus, who copied them from Olympiodorus.

In the summer of 409, Honorius is said to have called ten thousand Huns to his assistance.[370] Most historians accept this figure as if it had come from an official document.[371] Actually, it is one of those exaggerations in which Olympiodorus indulged.[372] What did those ten thousand Huns achieve? Nothing. At the end of the year, Alaric stood again at the gates of Rome. In 410, he marched to Ariminium, into Aemilia, to Liguria, back to Ariminium. In August, he took Rome. We hear nothing about the gigantic Hun army. Evidently it was a small contingent, probably not more than a few hundred horsemen. Still, the fact that some Huns joined the Roman army while others fought against it indicates a weakening of the royal authority.

In the later part of 409, Visigoths in upper Pannonia—a part of Alaric’s troops who for some reason had not marched with him all the way—rode into Italy. They were joined by Huns.[373] Their number may have been small. Yet they, too, acted on their own.

Still others, perhaps those who were still obeying Uldin, were engaged in fighting the Romans in Pannonia. In the summer or fall of 409,[374] Honorius “entrusted Generidus with the command of the forces in Dalmatia; he was already general of the troops in Upper Pannonia, Noricum, and Raetia, as far as the Alps.”[375] This passage has been variously interpreted. Swoboda dismisses it as invention; there were, he maintains, no troops in Upper Pannonia after 395.[376] Alfoldi thinks it supports his assumption that at that time Valeria already was ceded to the Huns.[377] Lot went a step further; from the fact that neither Valeria nor Pannonia secunda was under the command of Generidus, he concluded that both provinces were no longer held by the Romans.[378]

None of these assertions and assumptions is warranted by literary or archaeological evidence, direct or circumstantial. Even at the height of Attila’s power a part of Pannonia prima was held by the Romans, and there was, in all probability, never a formal “cession” of Valeria.

Generidus held no well-defined title or rank; he was “one of those commanders of the field forces who were appointed during the reign of Honorius to meet the emergencies of the time.”[379] From his position in the provinces named—Egger called it a Generalkommando[380]—it does not follow that there were no Roman troops in the provinces not named. True, we have no information about Roman forces in Valeria, but if it were not for the Vita s. Severini, we would have none about the garrisons in Noricum either.

In his pagan bias Zosimus probably exaggerated the achievements of his coreligionist Generidus, who is said to have drilled his troops, seen to it that the soldiers got their rations, and spent among them what he received from the treasury. “In this way he was terrible to the adjacent barbarians and gave security to the provinces which he was chosen to protect.”[381]

There were no “adjacent barbarians” of importance but the Huns. The difference between 408 and 409 is striking. In 408, the Huns in the West did not move. In 409, the troops from Raetia to Dalmatia were put under the command of one man to repulse them.

The picture which emerges from the sources and their admittedly conjectural interpretation is blurred. Yet it seems that we can discern four groups of Huns in the early 400’s. First, Uldin and his followers who, returning from the campaigns in Illyricum and Thrace, fought the troops of Generidus; second, Huns who in 408 formed a part of the Roman army in Italy; third, the Huns who joined it in 409; .fourth, a group that rode with Athaulf’s Visigoths against the Romans. The overall picture derived from the few bits of information is one of disintegration of the power of “the first king of the Huns,” as Olympiodorus would have called Uldin.

In his time falls the dissolution of the Hunno-Alanic alliance. Until 338, Huns and Alans are constantly named together, the Huns mostly, though not always, in the first place. But in 394, only the transdanubian Alans, led by Saul,[382] joined Emperor Theodosius;[383] of the Huns only those in Thrace marched under the imperial dragons. Alans, but no Huns, served Stilicho in 398 and, still under Saul, in 402.[384] In 406, however, Stilicho’s barbarian auxiliaries consisted of Huns and Goths; his bodyguard was formed by Huns.[385] Huns, but no Alans, served in the Roman army in 409.[386]

After 406, Western writers knew of Alans only in Gaul, Spain, and Africa. No author of the fifth century mentions Alans as allies of the Huns.[387] Jordanes knew of Sarmatians, not of Alans in Pannonia. The few Alans who after the fall of Attila’s kingdom settled in Scythia minor and lower Moesia[388] evidently moved there from the Wallachian Plain. All this cannot be a coincidence, and we know, indeed, the reason: The Alans moved from their old sites to Gaul; together with the Vandals they crossed the Rhine on the last day of 406.[389]

Why the Alans broke their alliance with the Huns is not known. There is a hint in Orosius that the relationship between the two peoples was already tense after 402. “I say nothing,” he writes, “of the many internecine conflicts between the barbarians themselves, when two cunei of the Goths, and then the Alans and Huns, destroyed one another in mutual slaughter.”[390] This passage has been strangely misunderstood. Most authors thought that Orosius referred to wars between Huns and Alans in their sites somewhere in the East.[391] But Orosius, who became jubilant whenever he could report how many barbarians in this or that battle were killed, most certainly would not have deplored the mutual slaughter of Rome’s enemies. Orosius’ taceo in VII, 37, refers to events unfortunate to the Romans: the escape of the defeated Alaric, and the “unhappy doings at Pollentia.” The cunei of the Huns and Alans were Roman auxiliaries, and Orosius deplores that Stilicho could not prevent those savage clashes in his own army.[392] As Gothic troops also fought each other, national antagonism between Huns and Alans, if it existed at all, may have been only a contributing factor.

According to Procopius, the Vandals left Hungary because “they were pressed by hunger”;[393] probably the people had outgrown the facilities for producing food.[394] The same may have been true for the Alans. Clashes with the Huns and the unwillingness to be forever the junior partners in an alliance which profited mainly the Huns may have been additional reasons for the Alans to seek new homes.

The Hunnic noblemen, Attila’s relatives and retainers, have either Turkish or Germanic names. There evidently were few, if any, Alans among the leading group. As no people ever emigrated to the last man, some Alans presumably stayed in Hungary after 406, but they played a minor role. Most of their tribal and clan leaders had left.


No period in the political history of the Huns is darker than the 41 O’s and 420’s. The loss of Olympiodorus’ History written in the second quarter of the fifth century, is, to quote Thompson, “a disaster for our knowledge of the nomads.”[395] It is true that Olympiodorus lacked the capacity to present the obviously rich material at his disposal in a coherent narrative;[396] at times he was gullible;[397] his figures are fantastic.[398] Yet of all the writers of the fifth century only he and Priscus traveled to the country of the Huns. What would we give to have his account of the negotiations with King Charaton instead of the few lines to which Photius reduced it I I put the name of the Hun king at the head of this section more in conformity with the titles of the other sections than to indicate its content. All we have for the two dark decades are a few isolated facts. In some cases it is, paradoxically, the very absence of information about the Huns that sheds some dim light on the events.

A fragment of Olympiodorus runs as follows:

Donatus and the Huns, and the skillfulness of their kings in shooting with the bow. The author relates that he himself was sent on a mission to them and Donatus, and gives a tragic account of his wanderings and perils by the sea. How Donatus, being deceived by an oath, was unlawfully put to death. How Charaton, the first of the kings, being incensed by the murder, was appeased by presents from the emperor.[399]

The date is the end of 412 or the beginning of 413, after Sarus’ death, referred to in the preceding fragment, and before Jovinian appointed his son Sebastian Caesar, reported in the following one. Altheim’s assertion that Charaton and Uldin reigned together as late as 414[400] has no textual support, but he rightly rejected the assumption that Donatus was a Hunnic king.[401]

From these few lines of the Olympiodorus fragment several unwarranted conclusions have been drawn. Charaton is supposed to have been

Donatus’ successor.[402] The text contains nothing of that sort. Assuming that Olympiodorus was sent to the Huns by the East Roman government, most historians place the center of Hun power somewhere near the shores of the Black Sea. This is certainly incorrect. As Thompson noticed,[403] Olympiodorus’ History deals exclusively with the Western empire. Haedicke assumed that Olympiodorus was in the civil service of the government of Ravenna.[404] Olympiodorus’ knowledge of Latin, his use of Latin words, the Latin forms of barbarian names, leave, indeed, no reasonable doubt that Haedicke was right. Olympiodorus, sent to the Huns by Honorius, crossed not the Euxine but the Adriatic Sea.[405] The Huns he visited lived in Hungary. How long Charaton “reigned” is as unknown as the number of tribes who acknowledged his hegemony. If the Huns to the north of the lower Danube should have belonged to the confederacy headed by Charaton, a possibility which cannot be excluded, they certainly did not feel themselves bound by any agreement which he made with Honorius. “Their” Romans were those of the East.

New Raids into Thrace

At the same time that Honorius sent gifts to Charaton, the Huns in Muntenia began to stir again. Moesia inferior and Scythia were most exposed to barbarian inroads. The praetorian prefect Anthemius did what he could to strengthen the border defenses, in particular the Danube fleet.[406] In 413, the walls of Constantinople were rebuilt and enlarged.[407] In spite of the repeated orders which restricted trade with the barbarians, enterprising traders were still finding ways and means to buy from them and, more important, to sell them forbidden goods. The decree of September 18, 420, differs from similar previous ones in one respect: it prohibits the export of merces inlicitae in ships.[408] Could it have been aimed at the trade with the Huns on the shores and in the hinterland of the Black Sea? To answer this question we have to make a short digression.

In the third quarter of the third century, the Goths were the terror of Asia. They sailed from the Black Sea ports as far as Ionia; the Heruli took Lemnos and Skyros, the Borani pillaged Pityus and Trapezunt.[409] But the Huns never took to the sea.

The Goths were no sailors either. When, in the last years of Ostrogothic power in Italy, King Totila (541–552) decided to build a fleet to deny the Byzantines the hitherto undisputed command of the Italian waters, he could not find enough Goths to man the ships. The sea battle of Senigallia, in which the Romans sunk or captured thirty-six of the forty-seven enemy vessels, marked the end of the Gothic fleet.[410] To- tila’s ships were built by Romans. The boats which in the third century carried the barbarians across the Euxine were built in Panticapaeum and were sailed by Bosporan crews.[411] Unable to navigate the vessels themselves, the Goths and Borani forced the Bosporans to supply them with convoys for their expeditions to Pontus, Paphlagonia, and Bithynia. Why the Gothic naval actions ceased after 276, we do not know. But it is probably not a coincidence that Dacia was abandoned at about the same time. With the emigration of a large section of the Goths to the former Roman province, the tribes to the east of them could expand westward. As greedy as they were for the riches of the Roman cities, they wanted and needed, first of all, land to settle. This was not true for the Huns. Why, then, did they not turn into pirates like the Goths before and the Slavs after them? They tried, but they failed.

In 419, Asclepiades, bishop of Chersonese, petitioned the emperor to free from punishment “those persons who have betrayed to the barbarians the art of building ships, that was hitherto unknown to them.” The petitions was granted. “But,” concludes the edict, “we decree that capital punishment shall be inflicted both upon these men and any others if they should perpetrate anything similar in the future.”[412]

Chersonese was the only place on the west coast of the Crimea still under Roman rule. The barbarians nearby were Goths and Huns. It is extremely unlikely that the Crimean Goths in their mountain homes should have wanted to build ships. This leaves the Huns. They probably needed ships both for piratical raids and for trade. They could not get them, and the government in Constantinople saw to it that no Roman ships sailed to Euxine Hunnia. If the Huns wanted merces inlicitae, they had to pillage the border provinces, which they did.

On March 3, 422, Theodosius II issued the following edict, which has not found the attention it deserves of students of the Huns:

Our most loyal soldiers returning from battle or setting out for war shall have for themselves the ground floor rooms of each tower of the New Wall of the sacred city. Landholders shall not be offended on the ground that the order which had been issued about public buildings has been violated. For even private homeowners customarily furnish one third of their space for this purpose.[413]

Nine years before, the landholders on whose properties the wall was built had been granted immunity from the law of compulsory quartering.[414] The upper part of the towers was set apart for military purposes; the lower part, however, could be used by the landlords without restrictions. When one considers what a heavy and hated burden the compulsory quartering of soldiers was, and how carefully the government refrained from extending it beyond the minimum just compatible with military necessities,[415] it becomes evident how tense the situation in and around Constantinople in the spring of 422 must have been to enforce the abolition of a regulation which was “to be observed in perpetuity.” Translated from legal into military terms, the edict says that the garrison of the capital is to be held in constant readiness against an enemy nearby. A terse entry in the Latin chronicle of Marcellinus Comes, s.a. 422, furnishes the commentary: “The Huns devastate Thrace.”

Nowhere in the history of the Huns is the one-sidedness of our sources more manifest. Hun bands skirmished with Roman soldiers almost at the gates of Constantinople. Yet no word about it appears in the detailed ecclesiastical histories, no allusion in the vast theological literature of the time. Theophanes registered that on September 7, 422, in Alexandria the praefeclus Augustalis Callistus was killed by his slaves,[416] a fate he probably deserved. But neither Theophanes nor any other writer thought it worthwhile to mention the peasants killed in Thrace, to speak about the people thrown out of their homes in the towers, the drudgery of the soldiers. Unlike the “illustrious persons” and bishops, they were expendable.

The Huns Help Aelius and Lose Pannonia

We have no information about Hun raids in the West in the 420’s.

Among the troops which, in 424, Castinus, commander in chief of the usurper John (see Chapter XII), sent against Boniface in Africa were also Huns.[417] The date is of some importance. Because the expeditionary force left immediately after Castinus had gone over to John,[418] these Huns must have formed a part of the regular army. There was not enough time to turn to federates beyond the border; the Huns must have been stationed in Italy. This, in turn, points to friendly relations between the Western empire and at least some Huns at the time that Aetius was still holding the modest position of cura palatii. Sanoeces, one of the three duces in Africa,[419] might have been a Hun.

A year later, in 425, Aetius marched with a huge Hun army[420] into Italy, to help John in the war with the East Romans.

John sent Aetius with a great sum of gold to the Huns, a people known to him since the time when he was their hostage and attached to him by a close friendship; he added the instructions that as soon as the enemy, that is, the army of the Eastern empire, entered Italy, Aetius should fall upon them from the rear while he himself would engage them at the front.[421]

The Huns came too late; three days before their arrival John had been executed. But Aetius, either unaware of what had happened or unwilling to believe the news, engaged the Eastern forces in a battle in which many were slain on both sides. The short campaign ended with the reconciliation of Aetius and Empress Mother Galla Placidia. The Huns received a sum of gold, returned hostages, exchanged oaths, and rode back to their country.[422]

Aetius, who probably spoke their language, was the best man John could find for his negotiations with the Huns. Of course, they sent their horsemen to Italy not out of friendship with Aetius but because they were paid “a great sum of gold.” They received more for breaking off the fight, and it is almost certain that they were promised regular annual tributes. Had Aetius stayed in Ravenna, the alliance with the Huns might have lasted for years. But he was sent to Gaul, and, for reasons we cannot guess, in 427 the Romans attacked and conquered the Huns in Pannonia.

Under the year 427, the sixth-century chronicler Marcellinus Comes has the short entry, “The provinces of Pannonia, which for fifty years were being held by the Huns, were retaken by the Romans” (Pannoniae quae per quinquaginta annos ab Hunnis retinebantur, a Romanis receptae sunt.)

These two lines have been discussed by generations of historians;[423] they were dismissed as nonsense[424] and were made the basis for far-reaching conclusions; they were interpreted, and reinterpreted, translated and retranslated to fit all possible theories about the fate of the former Roman provinces in the Danube basin.

It has been maintained that Marcellinus’ Romani must have been the Eastern Romans.[425] It is true that in the preface to his chronicle Marcellinus wrote that, in continuing Jerome’s work, “I write of the Eastern empire only” (orientate tantum secutus imperium). On the whole he did. But before the entry s.a. 427 Marcellinus dealt with purely Western affairs no less than thirteen times.[426] Whether it was the Eastern or the Western Romans who took back Pannonia, Marcellinus could in either case use only one word, namely Romani.[427] Until 476, the two partes formed the one Roman Empire, Romanum imperium of the Romanus populus or Romana gens.[428]

If, taken by itself, the passage in Marcellinus permits an “Eastern” as well as a “Western” interpretation, the parallel in Jordanes, Getica 166, leaves no doubt about its meaning (see Chapter XII). Under the consulship of Hierius and Ardabures, we read there, “Almost fifty years after the invasion of Pannonia the Huns were expelled by Romans and Goths.” (Huni post pene quinquaginta annorum invasam Pannoniam a Romanis et Gothis expulsi sunt.) Until recently it generally was assumed that Cassiodorus simply copied Marcellinus. That he smuggled the Goths into the text was in no way remarkable; he did that more than once.[429]

The other differences between the Getica and Marcellinus were regarded as too minor to deserve attention. Ensslin made these differences between the Romana, Getica, and Marcellinus the object of an admirable study.[430] He proved that Cassiodorus and Jordanes as well as Marcellinus drew heavily on the lost Historia Romana of Symmachus (f 525), great-grandson of the famous orator of the same name. It is practially certain that the two passages go back to it.[431] Bringing in the Goths, Cassiodorus had to change the colorless receptae—the Huns did not give up Pannonia, they were driven out. But he retained pene quinquagin ta of the original.[432]

For the rest, the two passages need hardly a commentary. Pannoniae means the same as Pannonia.[433] Retinebantur is, perhaps a little more emphatic than tenebantur.[434] A Romanis receptae means, of course, “were taken back, regained, recovered by the Romans.” I mention this only because Lizerand translated regues des Romains[435] As he understood the entry in Marcellinus, d une possession de fait succtde en 427, pour les Huns, une possession de droit, which is clearly incorrect.

The archaeological evidence does not bear out Symmachus. Nowhere in Pannonia prima or in Valeria exists a fortification, a military camp, or even a simple building that could be dated in the 420’s. Yet Symmachus could not have simply invented the reconquista. He probably exaggerated the successes of the Romans. Perhaps they merely reoccupied a number of fortified places. It is likely that they drove back some Hun bands which had ventured too closely to Noricum. Possibly Roman horsemen dashed deep into long abandoned tracts; here and there they may even have reached the Danube. In any case, the Western Romans did go to war against the Huns, with whom only two years before they had concluded an alliance, and defeated them.

Guldenpenning rejected the “Western” interpretation of the passages under discussion on the ground that Placidia’s government was so fully occupied in Gaul and elsewhere that it could not, at the same time, undertake an offensive against the Huns.[436] In a way, this is true. But now, because we know that in 427 Pannonia was, if not reconquered, at least partially made Roman again, Giildenpenning’s argument must be turned around. The Romans, indeed, could not attack the Huns unless the latter were so weakened that even the limited forces along the “frontier” sufficed for a local offensive. The Romans had not much strength; the Huns must have had even less.

As so often in these studies, we are dealing with such scanty evidence that it might seem best to register the various fragments of information and leave it at that. The gaps are too wide, not to speak of the chronological and geographical uncertainties, to seek any trend or development. Still, seen as a whole and against the background of the events of Uldin’s time, these dark decades seem to reveal at least two crises in the “body politic” of the Huns.

About 410, the Hun hordes acted as if there existed no ties, or only the loosest ones, to bind them together. It may, and it may not, be a coincidence that, shortly before, the Alans broke their alliance with the Huns. If, as we may assume, the mightiest Hunnic tribes were those which had forced the Alans to join them, the secession of the Alans must have sapped their strength.

Only a few years later, the Hun kings again acknowledged the leadership of one man. Charaton may have been only primus inter pares. However, even if he was not more, he probably owed his position not so much to his personal qualities, though they may have been of some importance, as to the preeminence of those Huns who followed him. The crisis was over. In 425, the tribal confederacy was again so well organized that the Huns could send several thousand horsemen to Italy, evidently more than what a single tribe was able to raise. The Huns whose help Aetius sought and got must have been under the leadership of a group in a position to coordinate the efforts of a number of tribes, perhaps even to enforce its will on others.

But then again, for reasons unknown, the confederacy lost much of its cohesion. Even if the successes of the Western Romans in Pannonia were relatively modest, the fact that the Huns west of the Danube had to give up a part of what they had been holding indicates the inability of the Huns as a people to rally their forces for a common cause. When the Huns in Pannonia were attacked, they must have called on their countrymen in the East for help. They received none. Nor do we hear that in the following five years the Huns beyond the Danube made an effort to reconquer the lost territory.

Octar and Ruga

Although the sources for the history of the Huns in the 430’s flow comparatively copiously, it is not easy to reconstruct even the main events. In 432, Ruga was king of the Huns. This seems to be the only certain date. In what year he became king, over what territory he ruled, to what extent he expanded it, what wars he fought and when, who after 430 his coregent was (if he had one)—these are questions to which the most divergent answers have been given.[437] Under such circumstances the smallest bit of information has to be carefuly scrutinized. We begin with a passage in the Getica:

For this Attila was the son of Mundzucus, whose brothers were Octar and Ruas, who were supposed to have been kings before Attila, although not altogether of the same [territories] as he. After their death, he succeded to the Hunnic kingdom together with his brother Bleda. (Is namque Attila patre genitus Mundzuco, cuius fuere germani Octar et Roas, qui ante Attilam regnum tenuisse narrantur, quamvis non omnino cunctorum quorum ipse. Post quorum obitum cum Bleda germano Hunnorum successit in regno.)[438]

Jordanes, or rather Cassiodorus, telescoped his source;[439] Octar died about 430, Ruga a few years later. But apart from this mistake, the statement is so precise that one can only wonder how it could have been misinterpreted. Yet both Bury and Thompson made Mundzuc the coregent of Octar and Ruga.[440] Jordanes’ style is sloppy, but had he meant to say that the three brothers ruled the Huns, he would have written Mundzuco, qui cum germanis Octar et Roa regnum tenuisse narratur. No author mentions Mundzuc as king of the Huns. From Priscus we know that there was a fourth brother, Oebarsius, who was still alive in 448.[441] He did not share the rulership with Octar and Ruga either. Only these two were kings. Before discussing their alleged double kingship, I have to deal with Socrates’ account of Octar’s fight with the Burgundians.

In the first half of the tenth century the Magyars raided western Europe from the North Sea to the Mediterranean. Between 900 and 913, they devastated Silesia, Thuringia, Franconia, and Bavaria. In 912, they crossed the Rhine. In 915 they took Bremen. They ravaged Lorraine twice, in 917, and again in 919, when they turned south and raided northern Italy. In 924, they appeared in southern France; Verdun fell to them in 926. Magyar horsemen camped before Lyon in 937. In 951, they rode as far as Calabria.[442] Summoning all the forces of the empire, Otto I finally defeated them decisively in the battle on the Lechfeld in 955.

The Germanic neighbors of the Huns were split into tribes, none of them even approximately as strong as the weakest of the German principalities of the tenth century. Incapable of any concerted action for any length of time, divided by mutual mistrust, periodically at war with each other, they were incomparably less able to defend themselves against the Huns than five hundred years later the dukes of Bavaria or Thuringia against the Magyars. Even without the not-too-exact literary evidence we would have to assume that the Huns made raids into the territories of the Germanic tribes to the west as they raided the Balkan provinces to the south.

There exist, indeed, two accounts of such predatory expeditions. The first comes from Socrates. How he received the information is not known, except that Uptaros, the name of the Hun king in Socrates, Jordanes’ Octar, points to informants who spoke Latin. Socrates wrote:

There is a nation of barbarians dwelling beyond the Rhine, called Burgundians. They lead a peaceful life. Being almost all carpenters, they support themselves by their earnings from this craft. The Huns, by making continuous eruptions on this people, devastated their country, and often destroyed great numbers of them. In this perplexity, the Burgundians resolved to have recourse not to any human being, but to commit themselves to the protection of some god; and having seriously considered that the God of the Romans defended those who feared him, they all with common consent embraced the faith of Christ. Going therefore to one of the cities of Gaul, they requested the bishop to grant them Christian baptism; who ordering them to fast seven days, and having meanwhile instructed them in the principles of faith, on the eighth day baptized and dismissed them. Accordingly becoming confident thenceforth, they marched against the tyrants;[443] nor were they disappointed in their hope. For the king of the Huns, Uptaros by name, having burst asunder in the night from surfeit, the Burgundians attacked that people then without a leader; and although few in numbers and their opponents many, they obtained a victory; for the Burgundians were but 3,000 men, and destroyed no less than 10,000 of the enemy. From that time on this nation became zealously attached to the Christian religion.[444]

Socrates’ account of this Hun raid about 430[445] has been dismissed as devoid of any historical value.[446] No other author knows of a struggle between Huns and transrhenanian Burgundians. The traditional miracle motifs can be discounted. But there still remain such absurdities as the existence of a Germanic tribe of peaceful carpenters, their conversion within a week, and the victory of three thousand artisans over ten thousand of the most formidable warriors of the century. Besides, it has been asserted that the story is at variance with all we know about the history of the Burgundians. They crossed the Rhine shortly after 406. In 411, they helped Jovinus to the throne. In 413, they obtained partem Galliae propinquam Rheno.[447] Aetius’ Hun auxiliaries slew King Gun- dahar, his whole family, and twenty thousand of the Burgundians.[448] If any Burgundians stayed behind on the right bank of the Rhine, they cannot have numbered more than a few hundred. These are strong arguments. And yet, Socrates’ story contains a historical kernel.

In the Panegyric on Avitus, Sidonius lists Burgundians among the peoples who followed Attila on his march to Gaul.[449] His catalogue of ethnic names has been denounced as untrustworthy,[450] and it must be admitted that it is a strange hodgepodge of names of real peoples and of those who had long ceased to exist or lived only in poetry.[451] Sidonius wrote the panegyric five years after the Hun war in 451.[452] Everyone in Gaul knew that Attila had neither Geloni nor Bellonoti among his troops, but no one would have objected to Sidonius naming them. It was different with the Burgundians. Avitus himself had fought them in Belgica prima.[453] In 443, after the catastrophic defeat by Aetius’ Huns, they were settled in Sapaudia.[454] Eight years later they fought under Aetius and Avitus against Attila’s Huns.[455] How could Sidonius, in an address delivered before Avitus and in the presence of the praetorian prefect,[456] have said that Attila had Burgundians among his hosts if he had none? The names of the Germanic tribes in his list shows how accurate Sidonius’ list was (apart from the poetic names.) The Rugi, Sciri, and Gepidae marched indeed with the Huns to Gaul. No poet before Sidonius mentioned the Toringi,[457] no other source mentions them as having taken part in the war. All this makes it practically certain that transrhenanian Burgundians did join Attila.

Socrates’ account of the conversion of the Burgundians to the orthodox faith is confirmed, though not in the details, by Orosius, according to whom the Burgundians “have by divine providence recently become Christians of the Catholic faith” (providentia Dei Christiani omnes modo facti catholica fide).[458] This statement, too, has been called a pious invention.[459] However, the thorough analysis of Orosius’ text by Coville[460] leaves no doubt that the Burgundians before they became Arians, probably under Visigothic influence, had been Catholics.[461]

Werner considers it possible, even probable, that the Burgundians east of the Rhine were for some time the subjects of the Huns.[462] But the arguments he adduces indicate rather a symbiosis of Alans and Burgundians in Sapaudia. It seems best to take Socrates’ story as it stands: The Huns raided the Main region as centuries later the Magyars raided Lorraine.

Seeck thought that Octar-Uptaros and Roas-Ruga might have been Uldin’s sons.[463] Perhaps they were. They as well might have been relatives of Charaton. It is equally possible that their family came from a tribe which until then had played a minor role in the confederacy. We simply have no information about the forebears of Octar and Ruga, nor do we know how they acquired their positions of authority. If it were not for Jordanes, we would not even know that for some years they jointly ruled the Huns.

On the basis of the short passage in the Getica and some vague analogies it has been suggested that Hun kingship was a Doppelkonigtum. If this means that two kings jointly ruled a common territory, the suggestion should be dismissed because it is at variance with the texts.

Jordanes is quite explicit: “Bleda ruled over a large section of the Huns.” (Bleda magnae parti regnabat Hunnorum.) After his death Attila “united the entire people under his rule” (universum sibi populum adunavit). The chronicler Prosper of Aquitaine says the same: “Attila, king of the Huns, killed Bleda, his brother and colleague in the royal office, and forced his peoples to obey him.” (Attila rex Hunnorum Bledam fratrem et consor- tem in regno suo perimit eiusque populos sibi parere compellit.)[464] The sources do not indicate different functions for the two kings, for example, the one being the religious, the other the secular leader of his people. Against the thesis of dual kingship as an institution speaks also the fact that after Octar’s death in 430 no one succeeded him; his brother became the sole ruler, like Attila after he had murdered Bleda. Dual kingship is supposedly characteristic for large groups of the Eurasian nomads. I doubt it. The Goths were not Turks or Mongols, but in the fourth century they had at one time two kings.[465] Among the Alamanni Chnodomarius and Serapio were potestate excelsiores ante alios reges.[466]

The distinction which Prosper made between Bleda’s and Attila’s peoples clearly points to a geographical division. That the “dual kingship” was indeed nothing but just that follows from two seemingly contradictory entries in the Gallic chronicles. According to the Chronicle of 452, Bleda succeeded Rugila;[467] the chronicler of 511 made Attila Rugila’s successor.[468] Considering that the Chronicle of 452 reflects in more than one passage an Eastern source,[469] the contradiction becomes a plain statement: Bleda ruled over the tribes in the east, Attila over those in the west. The same division seems to have existed with their predecessors. Octar had nothing to do with the East Romans, whose only enemy was Ruga.

It would be risky to make a rule of what very well may have been caused by unique circumstances. After Attila’s death his many sons “were clamoring that the nations should be divided among them equally.”[470] That we hear later of only two kings, Dengizich and Ernach, does not exclude the possibility that there were more before. Attila had his coregent killed, and Dengizich and Ernach may have killed their coregents too. However, if the Huns, or rather their “eminent men,” should have decided to have again two kings, they allotted to each a definite territory. Dengizich and Ernach, though at times cooperating, ruled each over his own lands. The possibility that such a geographical division was rooted in cosmological or religious ideas cannot be ruled out, but there is no evidence for it. Perhaps it was dictated by purely practical reasons: Only exceptionally able men could hold all the tribes together. The “dual kingship” may have been the result of the coalescence of two groups of tribes which to a certain extent continued to preserve their identity. Finally, it is even possible that the Huns divided their territories into two parts to deal with the two partes of the Roman Empire.

Compared with Octar, Ruga is a more substantial figure. We do not know how he, after his brother’s death, became sole ruler of the Huns; he was ruler at the latest in 432, when Aetius turned to him for help.

After the loss of his office, Aetius lived on his estate. When there some of his enemies by an unexpected attack attempted to seize him, he fled to Rome, and from there to Dalmatia. By way of Pannonia [per Pannonias],[471] he reached the Huns. Through their friendship and help he obtained peace with the rulers and was reinstated in his old office.[472]

At that time “Ruga was ruler of the gens Chunorum”[473]

The terseness of the few entries in the chronicles is a temptation to read more into them than they can yield. Whether the mere threat to march into Italy with a Hun army sufficed to make the Empress Placi- dia accept Aetius’ terms, or whether he actually crossed the Julian Alps at the head of Hun horsemen is not known; still, most historians have decided for the latter view.[474] Prosper’s per Pannonias has been taken as proof that in 432 the Romans were masters of all land west of the Danube,[475] although the two words only indicate that Ruga’s residence was east of the river.[476] Still, these are at least interpretations of the sources. But the thesis that the cession of a large part of Pannonia was the price Aetius had to pay for the help of the Huns is not warranted by any text. Yet by now it has become almost an article of faith.

The alleged and actual cessions of Roman territory to the Huns take a prominent place in nearly all studies on the barbarians. I could have discussed Alfbldi’s view on the fate of the province of Valeria when I dealt with the events in 408, but it seemed preferable to approach the problem of the cessions as a whole and from a wider angle.

How should one imagine the abandonment of Valeria, which had no natural borders in the west and south? Neither to the Huns nor to any other barbarians on the frontiers of the empire did the delineations of the Roman provinces have any meaning. No Hun horseman would have stopped at the sight of a border mark—or turned around because only Valeria had been ceded to his chieftain. The lines that on the maps in the offices at Rome and Ravenna divided Valeria from Pannonia prima and secunda could not prevent a single Hun from driving his herds and flocks across them. Most students of the Huns are students of the later Roman Empire and cannot help thinking in Roman administrative terms. If the barbarians knew the borders of the provinces, they paid no attention to them. After the migration of the Ostrogoths to the Balkans, the Gepids held not only Sirmium but also the adjacent regions of Moesia prima.[477] By the treaty of 510, Pannonia secunda was divided: The far greater part of the province became Ostrogothic; only the territory of Bassiana remained Roman.[478] In 528, Justinian ceded to the Heruli a territory which did not coincide with any of the old administrative units; it comprised tracts both on the right and left bank of the Sava.[479] Isidor of Seville gave the best definition of such a “cession.” He did not enumerate the provinces which fell to the Vandals in 435; the barbarians, he wrote, received partem Africae quam possederunt.[480] When the Romans “ceded” land to the barbarians, they merely—with very few exception—recognized a de facto situation which the barbarians wanted legalized, to be acknowledged as federates (which meant the paying of tribute money) or to regulate trade relations.

The archaeological evidence cannot reveal the exact time when a province, or a part of a province, or a province and some regions of the adjacent one was abandoned. Again to take Valeria as an example: on archaeological grounds the evacuation of Aquincum (modern Budapest) has been dated in the last decades of the fourth century,[481] whereas the much smaller Intercisa is supposed to have stayed Roman beyond the beginning of the fifth century.[482] There is no proof for such dates and there can be none. The finds from Intercisa have been thoroughly studied. Many clay vessels are dated in the fourth century. But not even the most meticulous analysis of their shapes and decor can establish the decade, not to speak of the year, in which they were made. Was it 379? Perhaps. Or 390? Possible. 410? This, too, cannot be excluded. It is impossible to set a deadline after which these plain jugs or dishes could not have been made. The ethnic attribution of the finds with marked barbarian features is equally uncertain. After Klara Sz. Poczy assigned one type of vessel to the Visigoths, another to “an Ostrogothic, respectively Hunnic- Alanic people,” whatever that means, she admits at the end of her study that it is hardly possible to differentiate between the pottery of the Goths, Huns, and Sarmatians. However, even if the exact dates when these vessels were made and the nationality of the makers could be established, we still would not know when and under what circumstances the garrisons were withdrawn. The dwellings in and around the squalid camps were occupied by barbarians, who probably included some Huns. Where these free Huns or, nominally, subjects of the emperor? We do not know. In Fenekpuszta in Pannonia prima, south of Balcum on Lake Balaton, Romans lived side by side with half-Sarmatized Germans.[483] Small Roman settlements were holding out here and there. The Huns apparently found it to their advantage to spare them because they needed the artisans. Had they wanted, they almost certainly could have overrun Vindobona (modern Vienna) any time; they left the poor people there to live in peace.[484] It is quite probable that a strip of wasteland separated Hun land from Romania in the west, as it did for a number of years in the south. But wherever the borders were, if one can speak of borders, they were not those of the former provinces.

The assumption that Aetius ceded a part of Pannonia to Ruga rests on a misinterpretation of a passage in Priscus. In his account of the East Roman embassy to Attila, Priscus calls Orestes a Roman who “lived in the land of the Paeonians on the river Sava, which according to the treaty of Aetius, general of the Western Romans, belonged to the barbarian” (tpxfit tt]v nQOQ t(i) Xdq>norapa) Haiovcov yc’joav r qj 0 a q 0 a q w xaxa rat; ’Aeti’ov OTQaTTjyov T(~)V sGjIEoudv ‘Painauuv avvOTpca$ vnaxovovaav).[485] The view that Aetius ceded Pannonian territory to Ruga is so firmly established that to my knowledge no one paid attention to the spaced words. Priscus could have written rep ‘Povq or rot; ftaefiaQois. But he wrote to flaofldoq). “The barbarian” occurs in the fragment in three more passages: (1) Attila ordered that neither Bigila nor the other East Romans must buy horses or anything else except the most necessary food.— “This was a shrewd plan of the barbarian”;[486] (2) the West Romans sent an embassy “to the barbarian”;[487] (3) “the barbarian” named the men whom he would accept as negotiators.[488] It follows that the barbarian to whom Aetius ceded the land along the Sava was Attila, not Ruga. Note also that not a province but a territory designated by the river was ceded.

Disregarding the precise statement of Priscus, historians of the Huns arbitrarily dated the “cession” of Pannonia prima[489] in 425,[490] or 433.[491] Another passage in the same fragment shows that not even in 448, when Attila was at the height of his power, was the whole province Hun land. Constantiolus was a man “from the land of the Paionians that was ruled by Attila” (ex rrjc; JJaidvcov %d)Qa<; Ttjs vtio ’ArT’qXa TaTTopdvrjq).[492] The specification makes sense only if a part of Pannonia was not under Attila’s rule.

It is conceivable that Aetius paid for Ruga’s help in land, which can only mean that he officially consented to the Huns’ keeping, to vary Isidor’s words, partem Pannoniae quam possederunt. It is equally possible that he paid for it in cash; he may have concluded an alliance with Ruga and promised to pay him subsidies. Or he may have done all that at the same time. But these are mere surmises. We should turn now to Ruga’s relationships with the East Romans.

Our information comes from Socrates, Theodoret, Priscus, and the Chronicle of 452. The excerpt from Priscus[493] seems to be somewhat shortened, and in the second half there is a gap of a few words; still it is by far the most important and most reliable source for the history of these dark years.

When some tribes on the Danube fled into Roman territory and offered their services to Theodosius, Ruga demanded through his envoy Esla that these and all other fugitives be surrendered to him; a refusal he would regard as a breach of the peace. Shortly afterwards Ruga died, succeeded by Attila and Bleda. The new treaty they concluded at Margus (near the modern village of Dubravica east of Belgrade) with Plintha, the Roman plenipotentiary who was accompanied by the quaestor Epigenes, was entirely to the advantage of the Huns. It provided for the surrender of all fugitives from the Huns and of those Roman prisoners of the Huns who had returned to the empire without paying ransom; the latter had to be sent back, unless 8 solidi were paid for each of them. The Romans undertook not to form an alliance with a barbarian people with whom the Huns went to war. At the fairs Huns and Romans should have the same rights and the same security. The annual tribute was raised from 350 to 700 pounds of gold. Among the fugitives surrendered by the

Romans were Mamas and Atakam, two young men of royal descent; they were handed over to the Huns at Carso[494] and crucified.

This Priscus fragment is in various respects most instructive. We learn from it of a previous war which the Huns had won. A tribute of 350 pounds of gold, 25,200 solidi, is not a very large sum, but the fact that the Romans paid it to Ruga indicates his eminent position. He must have been more than “the first of the kings,” as Charaton was. He, and not the “kings” or phylarchoi, received the money. How he distributed the gold among the tribal leaders and other members of the Hun aristocracy cannot be ascertained. However, he obviously was able to enforce his decision: those who were dissatisfied could not rebel; they fled to the Romans. Ruga had his diplomats; Esla, says Priscus, was experienced in negotiating with the imperial government. It seems that Ruga played also, though indirectly, a role in the domestic struggle at the court, in Constantinople. Plintha urged him to negotiate with him and not with any other Roman, which makes sense only under the assumption that the ex-consul used his connections with the Huns as a weapon against his rivals, as Aetius did in Ravenna. Still, the power of the king was not yet unlimited. The Huns fought together, but, at the same time, each one fought for himself. The prisoners a Hun made were his, not Ruga’s. Under Attila only men as prominent as Onegesius could keep their own prisoners;[495] all others were Attila’s property. How far eastward Ruga’s power reached cannot be decided. That he did not rule from Hungary to the Volga, as some scholars thought, follows from the treaty of Margus. The Romans could form alliances only with peoples who lived not far from their frontiers.

The chronology of Ruga’s last years is not easy to establish. The Chronicle of 452 lists his death under 434: “Aetius is restored to favor. Rugila, king of the Huns, with whom peace was made, dies. Rleda succeeds him.” (Aelius in graliam receplus. Rugila, rex Chunorum, cum quo pax firmala, moritur, cui Bleda succedit.)[496] It is well known how unreliable the chronology of the Gallic chronicle is.[497] If it were our only authority, we could date Ruga’s death as early as 431 or as late as 437. Seeck[498] thought that the date in the chronicle was confirmed by the edifying story about the ignominious death of the Hun king in Socrates.

The church historian relates[499] that Emperor Theodosius II (408–450), being informed that the barbarians were making preparations to ravage the Roman provinces,

committed the management of the matter to God, and, continuing in earnest prayer, he speedily obtained what he sought. For the chief of the barbarians, whose name was Rugas, was struck dead by a thunderbolt. Then a plague followed which destroyed most of the men who were under him, and if this was not sufficient, fire came down from heaven, and consumed many of the survivors. On this occasion Pro- clus the bishop preached a sermon in the church in which he applied a prophecy of Ezekiel[500] to the deliverance effected by God in the late emergency, and was in consequence much admired.

Proclus succeeded Maximian as bishop of Constantinople in April 434. However, Socrates does not say that Proclus preached the sermon in the capital. The story forms part of a panegyric on Theodosius, who evidently was as devout and meek before 434. Ruga’s death could have happened at a time when Proclus was still bishop of Cyzicus. And it did in the source from which Socrates drew. “It is because of this [i.e., Theodosius’] meekness that God subdued his enemies without martial conflicts, as the capture of the usurper John [in 425] and the subsequent discomfiture of the barbarians[501] clearly demonstrate.” Ruga’s hordes were those whom John had called to his assistance against the Romans; they attacked “after the death of the usurper.”[502]

The date in Socrates, not long after 425, is not only at variance with that in the Gallic chronicle; it is also irreconcilable with the one given by Theodoret, who tells exactly the same story both in his Ecclesiastical History[503] and the commentary on Psalm 22:14–15.[504] God helped Theodosius against the Huns because the emperor had proved his devotion to the true religion by issuing a law that ordered the complete destruction of all pagan temples. The victory over Ruga was “the abundant harvest that followed these good seeds.” The edict was issued on November 14, 435,[505] so Ruga would have been killed after that date. That this was, indeed, Theodoret’s information is confirmed, if confirmation is needed, by the other victory which God granted Theodosius as a reward for his pious zeal. He smote the Persians[506] in 441.[507]

We have, thus, three dates for Ruga’s death: shortly after 425, 434, and after November 435. Properly, none is correct, for Ruga died, as we know from Priscus, not in a campaign in Thrace[508] but in his own land. Still, Socrates’ and Theodoret’s accounts cannot be dismissed as valueless. The Romans did wage war against Ruga. The legend reflected and distorted its first phase. It has a close parallel in the homily of Isaac of Antioch. As the Huns in 447 had to retreat for a short time, incidentally also because of a plague, only to attack again and conquer, Ruga’s hordes, too, apparently suffered a temporary reverse.

At the time of the negotiations which led to the treaty of Margus, the Huns were still holding Roman prisoners, so the peace seems to have been concluded not very long before.[509] This is also indicated by the Gallic chronicle: “Ruga, with whom peace was made, dies” (Ruga, cum quo pax firmata, moritur). Who made peace with Ruga? Not the West Romans, as it is usually assumed;[510] they had not been at war with the Huns. But the East Romans were. Furthermore, when we consider that the Gallic chronicle draws more than once on Eastern sources,[511] it is practically certain that the peace referred to is the one that brought the fighting in Thrace to an end.

The date in Socrates is unacceptable; the one in the Gallic chronicle uncertain. Theodoret’s “after the end of 435” is in agreement with Priscus. Epigenes, Plintha’s companion on the embassy to Ruga’s successors, on November 15, 438, was still magister memoriae. Because Priscus describes him as quaestor, the embassy falls after that date.[512] Thompson thinks that Priscus made a slip, but his only argument is the date of Ruga’s death which he, arbitrarily as we may say now, places in 434. Plintha’s role in the negotiations with Ruga and, then, with Bleda and Attila, furnishes another argument for a late date. He was, says Priscus, magister militum. Anatolius, who in 447 concluded the peace treaty with Attila, was magister militum praesentalis. Plintha’s position at the court, his apparently strained relationship with other high dignitaries, his interference in diplomatic affairs, all this leaves little doubt that he, too, had the rank of m. m. praesen- talis. In 434, Saturninus, who was to take the place of Dorotheus, bishop of Marcianopolis, deposed by Maximian, came to the town cum magni- ficentissimo et gloriosissimo magistro militae Plintha.[513] At that time Plintha was still magister militum per Thracias. His promotion falls, thus, after 434.

To summarize, Ruga’s war with the East Romans, his death, and the beginning of the reign of Bleda and Attila are to be dated in the second half of the 430’s.


In the Bazaar of Heracleides, the ex-patriarch of Constantinople, Nes- torius (428–431), since 436 exiled to Oasis in Egypt, with deadly monotony turned to the injustice done to him at the Council of Ephesus (June 431) and to the unspeakable evils that came from it. In its rabies theo- logica the book surpasses even the writings of the patriarch Cyril of Alexandria, Nestorius’ enemy. Only occasionally does its author cast a quick glance at the world outside the conclaves where the enemies of God plotted his downfall. Yet this narrow-minded fanatic understood the causes of the rapid ascendancy of the Huns better than most of his contemporaries. Toward the end of the Bazaar, speaking of, or rather alluding to, the wars with the Huns in the last decade of the reign of Theodosius the Younger, Nestorius writes:

The people of the Scythians were great and many, and formerly were divided into people and into kingdoms and were treated as robbers. They used not to do much wrong except through rapidity and through speed. Yet later they were established in a kingdom, they grew very strong, so that they surpassed in their greatness all the forces of the Romans.[514]

Though this is an oversimplification, basically Nestorius was right. Until the end of the 430’s the Huns were a great nuisance, much worse than the Saracens or the Isaurians, but they were not a danger. Their inroads carried them at times deep into the Balkan provinces, but they were always either driven out or bought off.

At the end of the 440’s, the barbarians were, in Nestorius’ words, “the masters, and the Romans, slaves.”[515] This, too, is an exaggeration, but not even the most abject flatterers of the Christ-loving Theodosius could have denied that within a few years the bands of “robbers” had grown into a military power of the first rank. They would have rejected Nes- torius’ explanation of the change, namely, that it was “the transgression against the true faith of God impassible” which made it possible for the Huns to unite under one ruler, and in their way, they would have been right. The Huns did not become mightier because the Romans grew weaker. The Eastern army was as strong in 447 as it was in 437, the fortifications along the limes were as well garrisoned, if not better, and there is no reason to assume that the Roman troops were led by incompetent generals. Besides, the great Hun victories fell in a time when the Eastern empire was at peace with Persia. The explanation of the radical change in the relative strength of the Huns and the Romans must be sought not in Romania but in Hunnia.

It has become the fashion to deny Attila practically any merit for the short-lived greatness of his people. He was, we are told, neither a military genius nor a diplomat of exceptional ability, but a bungler who would not have made such awful blunders had he had a professor of history as advisor. The purpose of the following pages is not to prove that Attila was another Alexander; if as a result of a new study of the years 441–447 the personality of Attila turns out to have been a decisive factor, I am far from maintaining that it was the only one. But before speculating about primary, secondary, and tertiary factors, about the direct and remote causes of this or that event, the events themselves must be established. The standard histories give, I believe, an erroneous picture of the Hun wars in the 440’s and a distorted one of the relationship of the Huns with the West. A number of sources exist which have been either ignored or treated too cavalierly. None of them, taken by itself, is very revealing. Only by combining them all, and paying attention to the details, may we hope to reconstruct the happenings in this decisive decade of Hun history.

The Huns Threaten the West

Merobaudes’ Second Panegyric on Aelius[516] is a mediocre poem (see Chapter XII). More than half of it is lost; many verses in the only extant manuscript are mutilated and can be restored with no more than a varying degree of probability. Like the other poems of Merobaudes, the panegyric takes a very modest place in late Latin literature. But its value as a historical document cannot be overrated. It sheds light on the relationships between the Huns and the Western empire in a period about which we know next to nothing from other sources.[517]

It is not germane to my purpose to discuss the panegyric in all its historical aspects, but the parts dealing with the Huns can be set into the proper context only after the date has been established at which the poem was recited. Aetius was consul in 432, 437, and 446. Mommsen,[518] Seeck,[519] and Levison[520] assumed that Merobaudes addressed the consul of 437; Vollmer,[521] Bury,[522] Stein,[523] and Thompson[524] pleaded for 446. The obscure style of Merobaudes and his tendency to use circumlocutions instead of naming persons and places makes the interpretation often difficult. But taken as a whole the poem presents a clear picture of the events preceding the third consulship of the great ductor.

The year begins in peace (vv. 30–41). The clarions are silent, the arms at rest, Bellona has put down her helmet, her hair wreathed with olive, Mars stands by inactively while Aetius dons the consular toga. “The weapons and the chariot of the god are silent, and his idle steeds lay bare the pastures hidden under the Riphaean rime.”[525] In Claudian, whom Merobaudes closely followed, Mars’ steeds disport themselves in the pastures of the Eridanus.[526] The difference is significant. To Merobaudes the home of Mars is the far north. Aetius enters his consulship “with the northern regions subdued.”[527]

But it is hard-won peace. To secure it, Aetius had to fight many wars.[528] In the prooemium the poet rapidly surveys the achievements of his hero. He does not follow a chronological order. Starting in the north, on the Danube, he proceeds westward to the Rhine and the tractus Armoricanus (modern Brittany), turns south to Gallia Narbonensis, and ends in Africa.

After a reference to Aetius’ deeds at the Danube, which I shall discuss later, Merobaudes speaks of the Franks:

The Rhine has added an alliance serving the wintry world. The river is satisfied being bent by western chains and rejoices to see the Tiber [Rome] grow on its other bank.

(Addidit hiberni famulantia foedera Rhenus

orbis et Hesperiis flecti contentus habenis

gaudet ab alterna Thybrin sibi crescere ripa [vv. 5–7.])

These stilted verses are an example of Merobaudes’ style. He means to say that Aetius forced the peoples on the Rhine to conclude an alliance with Rome: the territory east of the river has been made Roman again.[529]

Taken by itself, the passage could refer to the late 420’s, the early 430’s, or the 440’s. Mommsen thought that Merobaudes alluded to Aetius’ victory over the Franks in 428.[530] In 432, Aetius conquered the Franks again.[531] In a context which points to a time not long before 439, Jor- danes speaks, somewhat vaguely, about the “crushing defeats” which Aetius inflicted on the proud Suevi and the barbarous Franks.[532] About 440, Cologne and a number of other cities in the Rhineland were a gain in the power of the Franks. A few years later—the exact date is unknown—they withdrew, to attack anew in 455. Because in 451 they fought against Attila as the allies of the Romans, the foedus between them and the empire, which means Aetius, must have been renewed after 440. Stein[533] was inclined to think that Merobaudes alluded to this last alliance. In other words, verses 5–7 are compatible with either date suggested for the panegyric. But the following verses 8–15, point unmistakably to 446.

The rosy picture which Merobaudes draws of the tractus Armoricanus, where the former Bacaudae (see Chapter XII), now law-abiding peasants, are peacefully tilling the long-neglected fields, was not true at any time in the first half of the fifth century. But even if all allowances are made for the exaggerations in which the panegyrist was expected to indulge,[534] Merobaudes could not have written those verses in 436. The Bacaudae, “an inexperienced and disorderly band of rustics” (agrestium hominum imperita el confusa manus);[535] were no match for a regular army,[536] but to the motley hosts of federates, which were thrown against them by Aetius, they offered the toughest resistance. It took the Romans a long time to put down the uprising which began in 435.[537] Tibatto, the leader of the Bacaudae, was still fighting in 436 and 437. When Aetius entered his second consulship, Litorius and his Huns were fully occupied hunting down the elusive bands which, driven into the woods,[538] inaccessible to the horsemen, broke out again and again. It was only after most of their leaders were either killed or captured that the commotio Bacaudarum “came to a rest.”[539]

Not only was the tractus Armoricanus not pacified in 437, war was also raging in southern Gaul. Narbonne, under siege by the Goths for months,[540] was at the point of surrender when Litorius relieved the city early in 437.[541] Merobaudes aptly describes Gallia Narbonensis, stressing the importance of the province as a link between Italy and Spain. Aetius drove the bandits out, the roads were open again, the people had returned to their towns.[542] Later in the poem (vv. 144–186), where he deals with the “warlike deeds” (Ttgdfet? xara noAepov) of his hero, Merobaudes draws a remarkable picture of the war in Gaul. The Goths were no longer the primitive savages whom Caesar had fought. They had learned the art of war, bravely holding out in fortified places, a people noble in deeds, if not noble in mind. In 439, the war ended with an alliance between the Goths and the Romans.[543]

Verses 24–29 refer to still later events. At a time when the Romans were still holding Carthage, Merobaudes could not have called Geiseric insessor Libyae, he could not have said that the Vandal king had torn down the throne of the Elissaean kingdom and that Nordic hordes filled the Tyrian towns.[544] Carthage fell on October 19, 439. Geiseric’s eagerness to arrange a betrothal between one of his sons and a Roman princess, to which Merobaudes alludes in verses 27–29, likewise presupposes that the war had come to an end. The peace concluded in 442[545] is the latest datable event named in the prooemium.

Now we can return to the beginning of the poem. The first verse is lost. It must have been a short praise of Aetius who

... Daniwii cum pace redit Tanainque furore

exuit et nigro candentes aethere terras

Marte suo caruisse iubet; dedit otia ferro

Caucasus et saevi condemnant proelia reges.

... comes back with peace on the Danube [or: from the banks of the Danube] and strips the Tanais of its furor, and orders the countries glistening under the black sky to be without their Mars. The Caucasus lets rest the iron, and the savage kings condemn the battles.

Mommsen’s conjecture[546] that the four verses refer to Aetius’ stay with the Huns after 409 (followed by Vollmer,[547] Bugiani,[548] and Thompson[549]) cannot be true. The avfyait; of the virtues of his hero was certainly the duty of the r/ie/or.[550] But there were limits. Not even the most servile sycophant could have said that the boy Aetius came back with peace on the Danube. He was sent to the Huns as hostage, to guarantee the observance of a treaty. He did not give orders, he received them.

But it is not only the content of the verses which forbids taking them as an allusion to Aetius’ youth, it is also the context in which they stand. Merobaudes puts the conclusion of the peace with the barbarians in the north at the head of the list of Aetius’ achievements. All achievements— the reconquest of the left bank of the Rhine, the pacification of the Are- morica, the victory over the Goths, Geiseric’s attempt at a rapprochement with the court at Ravenna—fall between 437 and 446. Aetius’ dealings with “Caucasus” and “Tanais” must be dated in the same period.

Verses 50–97 refer again to the barbarians in the far north. A nefarious goddess complains that she is held in contempt everywhere. “We are beaten back from the waves and not admitted on land.” Unwilling to bear this any longer, she is determined to call forth the distant peoples from the extreme north. Breaking the alliances of the kingdoms, regnorum foedera, she will plunge the world into misery. She drives to the Rhipaean Mountains where Enyo dwells. The goddess of war is depressed because peace has reigned for such a long time. The diva nocens exhorts Enyo to take heart and instigate the Scythian hordes of the Tanais to make war on the Romans.

These verses reflect Claudian’s influence in thought as well as in words.[551] But it is the content that interests us. The speeches of the diva nocens cannot be the prelude to a description of the Gothic war as Vollmer suggested. Verses 52–53 indicate the date with all the precision one can expect from Merobaudes. “We are driven from the sea and are not allowed to rule on land” (Depellimur undis nec terris regnare licet). The only people to fight the Romans at sea were the Vandals. Prosper, Marcellinus Comes, and the Chronicon Paschale (see Chapter XII) record their piratical expeditions in 437, 438, and 439.[552] After the conquest of Carthage, “they created a fleet of light cruisers and attacked the empire by sea, as no other Teutonic people had done or was to do in the Mediterranean.”[553] In 440, the Vandals landed in Sicily and ravaged Bruttium. It was only after 442, when Geiseric tried to get on better terms with the Romans, that the furies “were beaten back from the waves.”[554] In a decade of the fiercest onslaughts on the empire, a period of three or four years of peace must have been regarded as a rather long one, pax annosa. The threat

of war, Enyo’s appeal to the barbarians in the north, must be dated between 443 and 446.

The “savage Scythian hordes” were the Huns. By the middle of the fifth century no other people was strong enough to threaten Italy. Besides, Merobaudes characterizes the enemy so clearly that there can be no doubt whom he meant. The sites of the barbarians were near the Rhipaean Mountains, on the Tanais (the river Don),[555] and on the Phasis (the river Rion, east of the Black Sea): “Trembling Tiber will be attacked by his friend, the Phasis” (Phasiacoque pavens innabitur hospite Thybris (v. 56). The meaning of the bizarre metaphor is obvious: The people from the Phasis will break into Italy.

The Riphaean Mountains could be connected with any people in the north. The Tanais in poetic language is the river of the north, kat’ exochen, as the Nile is that of the south.[556] Alaric’s Goths were a people from the Tanais and the Hister.[557] Sidonius called even Geiseric a rebel from the Tanais.[558] But no Germanic tribe has ever been associated with the Phasis and the Caucasus. These were the regions from which the Huns came. The Hun auxiliaries in Theodosius’ army poured forth from “the threatening Caucasus and the wild Taurus” (minax Caucasus et rigens Taurus).[559] In 395, the Huns broke into Asia “from the distant crags of the Caucasus.”[560] They came from the land beyond the cold Phasis.[561] The Huns were not just barbarians far in the north. They were “ a people from the farthest boundaries of Scythia, beyond the icy Don” (genus extremos Scythiae ver- gentis in ortusj trans gelidum Tanain.)[562] And Merobaudes also so calls them: “tribes living in the farthest north” (summo gentes aquilone repostas [v. 55]). Caucasus, Tanais, Phasis, the extreme north—this is the country of the Huns, and only of the Huns.

The study of Merobaudes’ panegyric leaves no doubt that at one time between 437 and 446 the relationship between the Western empire and the Huns was extremely tense. The phrase “Caucasus granted leisure to the sword” (dedit otia ferro Caucasus) points to actual, though perhaps only limited war. Our interpretation is supported by the inscription on a (now lost) tombstone:[563]

Here the glory of Italy is buried, the hero Constantius, who was the shield of his country, its walls and weapons.

Invincible in war, a lover of true peace, though pierced with wounds, he was victorious everywhere.

He subdued the race that crossed the middle of the sea, and likewise the land refused to give aid to the vanquished.

He was sober, mighty in battle, chaste, a powerful commander, first in judgment, first in war.

He was as much burning in love and devotion to the Romans as he was bringing terror to the Pannonian tribes.

In war he sought honors for himself and his sons, to the nobles he gave as gifts the cut-off heads.

In the midst of his sons the father lies stabbed; the grievous mother does not know whom to lament, overwhelmed by her sorrow.

Worse is the misfortune of Rome, robbed of so great a senator; she has lost her ornament, she has lost her arms.

The saddened armies are standing still, after their great commander has been taken away, with whom Rome was powerful, without whom she is lying prostrate.

This tumulus, o great leader, has been erected for you by your wife, who lies here, reunited with you.

(Hie decus Italiae tegitur Constantinus heros qui patriae tegmen, murus ac arma fuit.

Invictus bello, non fictae pads amator, confixus plagis, victor ubique fuit.

Hie mare per medium gentem compressit euntem, et victis pariter terra negavit opem.

Sobrius armipotens castus moderamine pollens primus in ingenio, primus in arma fuit.

Romanis blando quantum flagravit amore, tantum Pannoniis gentibus horror erat.

Iste sibi et natis bello marcavit honores, munera principibus colla secata dedit.

Natorum medio fixus pater: anxia mater quern plangat nescit, stat stupefacta dolens.

Peius Roma gemit tanto spoliata senatu, perdidit ornatum, perdidit arma simul.

Tristes slant acies magno ductore remote, cum quo Roma potens, quo sine pressa iacet.

Hunc tumulum, dux magne, tuum tibi condidit uxor, quae tecum rursus consociata iacet.)

Constantius, a man of modest origin, distinguished himself in the service of Rome. He fought a barbarian people at sea and on land; in an engagement with the Pannonian peoples he was killed.

Who was this Constantius, and when did he live? The name is extremely common; there must have been dozens of senators called Constantius. Mommsen[564] surmised that the verses glorify the emperor Constantius Chlorus (305–306). But they were written at a time when Pannonia, or at least the larger part of it, was no longer a Roman province.[565] That the sea-going people were the Vandals has been recognized by Seeck,[566] Sundwall,[567] and Fiebiger.[568] However, these scholars could not fit the deeds of Constantius into the history of the 430’s or the 440’s. I think we can. The only time that the West Romans fought a barbarian seagoing nation, first at sea and then on land, was between 437 and 440. Tan- tum Pannoniis gentibus horror erat points to fighting in and around Pannonia, to a commander of troops on the frontier, now repulsing raiding bands, now making inroads in the territory of the enemy, to constant clashes along the border: munera principibus colla secala dedit.

There exist two more documents which reflect the threat of war with a formidable enemy between the second and third consulship of Aetius. By the novella issued at Ravenna on July 14, 444,[569] a large group of officials lost with one stroke privileges they had enjoyed for more than thirty years. Not only had they been exempt from the duty of supplying recruits from among their tenants; they did not even have to make the money payments which most landowners made instead of furnishing the men.[570] The new law provided that the illustres, who were inactive, pay in money for three recruits each, the price of one recruit being assessed at 30 solidi; that the counts of the consistory and those of the first order, the tribunes and notaries and ex-provincial governors pay for one recruit each, and inactive tribunes, counts of the second and third class, and other clarissimi for one-third of a recruit. The government, aware what a storm of indignation would sweep through the middle and lower ranks of the bureaucracy, hastened to assure them that the decree was issued only for the present time. But the government had no choice: because of “the necessity of imminent expenses,” the resources of the treasury did not suffice.

If Valentinian’s ministers expected that the new tax would alleviate the frightful financial stress in some degree, in aliqua parte, they soon realized that more radical measures had to be taken. Whether, as the emperor said, the merchants and in particular the landowners were really unable to pay more taxes may be doubted. The other way out, a cut in the military expenses, was impossible. “Nothing is for the afflicted condition of the state as necessary as a numerous army.” In the autumn of 444, the government devised a new tax, the silignaticum, a payment of 1 siliqua per solidus, that is a twenty-fourth, on all sales.[571] The government was barely able to feed and clothe the veteran army, and yet it issued the strictest orders to recruit more and still more soldiers. These were “difficult times”; an army as strong as possible was “the foundation of full security for all.”[572]

The preparations for the war with the Huns—there is, as we now may confidently say, no other explanation of the laws—fall in the second half of 444. Aetius negotiated with the saevi reges, Bleda and Attila. If Bleda’s death could be exactly dated, it would give the terminus ante quern for the renewal of the treaty between Huns and Romans. Our authorities give different dates. According to Prosper,[573] Attila put his brother to death in 444, possibly, as this is the last entry under this year, in the autumn or winter. Marcellinus Comes dates the murder early in 445,[574] the Chronicle of 452—notoriously inaccurate—in 446.[575] Theophanes, Anno Mundi 5943, is in his chronology hopelessly confused; Bleda was most certainly not killed in 441 as Theophanes seems to indicate.

That the tension was over in 445 can be concluded from the biography of the Greek renegade whom Priscus met at Attila’s court. Made prisoner in Viminacium (now Kostolatz, Yugoslavia) in 441, he fought under One- gesius first against the Romans, and then against the Acatiri, with such bravery that his lord made him a free man. He took a Hun wife who bore him children.[576] He told Priscus his story in 449. Therefore, his marriage falls in 446 at the latest. Because it is unlikely that the Roman prisoner immediately was put on a horse and sent against his countrymen, the campaign in which he fought was evidently the one in 442 or, more probably, 443. It preceded the war with the Acatiri, which, therefore, is to be dated in 443 at the earliest. Priscus says explicitly that Kuridach, the pro-Hunnic king of the Acatiri, appealed to Attila for help against the pro-Roman leaders of the people.[577] Therefore, the war falls after the death of Bleda. Attila led a large army against the Acatiri; he conquered them only after many battles. A hundred years later Jordanes still called them gens for- tissima.[578] To fight the Romans and the Acatiri at the same time was beyond the power of the Huns. All this leads to 445 as the only year in which, all circumstances considered, the war with the Acatiri should be dated. And this, in turn, narrows the period in which “peace on the Danube” was concluded to the winter 444/5 or the following spring.[579] If follows, furthermore, that shortly afterward, that is, in 445, Attila murdered his brother Bleda.

Our information about the following years comes from three sources. Two of them have been ignored by students of the Huns, the third one has been misinterpreted. There is, first, the letter of Cassiodorus in which he describes his grandfather’s[580] meeting with Attila:

With Carpilio, the son of Aetius, he was sent on no vain embassy to Attila. He looked undaunted at the man before whom the Empire quailed. Calm in his conscious strength, he despised all those terrible wrathful faces that scowled around him. He did not hesitate to meet the full force of the invectives of the madman who fancied himself about to grasp the Empire of the world. He found the king insolent; he left him pacified; and so ably did he argue down all his slanderous pretexts for dispute that though the Hun’s interest was to quarrel with the richest Empire in the world, he nevertheless condescended to seek its favor. The firmness of the orator roused the fainting courage of his countrymen, and men felt that Rome could not be pronounced defenseless while she was armed with such ambassadors. Thus did he bring back the peace which men had despaired of, and as earnestly as they had prayed for his success, so thankfully did they welcome his return.[581]

The grandfather of Cassiodorus dealt not with the saevi reges but with Attila alone. The characterization of the king as a man “who, driven by some fury, seems to strive for the domination of the world” (qui furore nescio quo raptatus mundi dominatum videbatur expetere) leaves no doubt that he had made himself the sole ruler of the Huns. The embassy must be dated after 445.[582] It would be of interest to know what Attila’s calum- niosae allegationes were. Perhaps he was complaining, as he did so often in his dealing with the East, that the Romans did not hand over all Hun fugitives. Or Aetius may not have paid the tribute as regularly as the king demanded. He may have tried to win to his side Germans over whom Attila claimed suzerainty. But all these are guesses. What we learn from the Variae is that the Huns renewed their threats to attack the West and that Aetius’ ambassadors barely succeeded in preventing the savages from breaking into Italy or Gaul. It was, of course, not Cassiodorus’ superior diplomatic skill that made Attila change his mind. Roman rhetorics never prevailed with Attila unless they were accompanied by the sound of Roman solidi.

The second source which sheds some light on the events in the second half of the 440’s is a short passage in the work of Anonymus Valesianus, which contains, among other things, an account of King Theodoric the Ostrogoth (493–526): Orestes, the father of the last Western emperor Romulus Augustulus, joined with Attila at the time the king came to Italy, and was made his secretary.[583] In 449, Orestes had already a responsible position; he accompanied Edecon on his mission to Constantinople.[584] Considering Attila’s mistrust of his Roman secretaries—he had one of them crucified[585]—it must have taken some time before he took Orestes in his confidence. But it is obviously impossible to date Attila’s stay in Italy on such a shaky basis. Much more significant is the fact that the Hun king did go to Italy. In 449, in a tense situation, Attila notified the East Romans that he was willing to meet their ambassadors in Serdica (modern Sofia), provided they were men of the highest rank.[586] It was not Attila’s custom to make pleasure trips to enemy country. We may assume that he met Aetius, or his plenipotentiaries, on Italian soil, probably not far from the frontier, because decisions of great importance had to be made.

The third source from which information can be drawn about the relations of the Hun king with Aetius is the short passage in Priscus discussed in the previous chapter. The Roman ductor ceded a large tract of Pannonia to Attila.

There can be no longer any reasonable doubt that Attila’s journey to Italy, Cassiodorus’ negotiations with him, and the cession of the land along the Sava belong together. It may have been on this occasion that Attila was nominated magister militum, naturally with the salary due him.[587]

Attila was appeased, but he did not become Aetius’ friend, as nearly all modern authors maintain.[588] That Aetius sent him secretaries and gifts is of little importance. In 484, Eudoxius, leader of the Bacaudae, fled to the Huns.[589] Had he been extradited to the Romans, as Aetius undoubtedly requested, the chronicler who reported the flight would not have failed to say so. He did not. Eudoxius was certainly not the only rebel to whom Attila granted asylum. That the Huns did not raid Noricum and Raetia, as they raided the Balkan provinces, had nothing to do with their allegedly friendly feelings for Aetius; there was little to loot there. Ail treaties the Huns concluded with the East bound the government in Constantinople to pay them tribute. They doubtless demanded, and received, gold, and ever more gold, from the West as well. Aetius was no more Attila’s friend than he was the friend of the other 2.^aTag%oi from Africa to the Danube. The Hunnic invasion of Gaul in 451 was merely the continuation of politics by other means, if the word politics can be applied to systematic extortion.

The War in the Balkans

In the early summer of 440, the government in Ravenna learned that a large Vandalic fleet had left Carthage. Whether it was headed for Spain, Sardinia, Sicily, Egypt,[590] or even Rome or Constantinople, no one knew.[591] The treacherous capture of Carthage by Geiseric, in the year before, was a blow not only to the Western Romans. In possession of the best harbor west of Alexandria with its shipyards and experienced shipbuilders, Geiseric could be expected in a short time to have a fleet able to carry the Vandalic pirates anywhere in the Mediterranean. The walls of Rome were hastily repaired,[592] the shore and harbors of Constantinople were fortified.[593] In a proclamation to the Roman people, the emperor Valentinian III assured them that the army of “the most invincible Theodosius” soon would approach to take part in the fight against the Vandals.[594]

Geiseric landed in Sicily. The Vandals took Lilybaeum on the west coast of the island, pillaged the helpless towns and villages, persecuted the Catholic clergy, and even crossed the Strait of Messina.[595] Late in 440 or early in 441,[596] Geiseric broke off the campaign and sailed back to Carthage. The Eastern army under Areobindus as commander in chief, which was supposed to drive out the Vandals, arrived in Sicily after the evacuation of the island by the enemy.[597] Behaving not much better than the Vandals, the preponderantly Germanic troops[598] soon became “more of a burden to Sicily than a help to Africa.”[599]

The Sicilian expedition was a failure. For one thing, it came too late. Valentinian’s ministers may have been overly optimistic when they announced its coming as early as June 440. The difficulties and risks of such an enterprise were greater than the hard-pressed West was willing to concede. To assemble the transports, to provide the necessary supplies, to move the troops to the ports of embarkation—all this needed time.[600] Yet this alone does not quite account for the delay. The East could not come to the rescue of the West because it was itself threatened on two fronts, in the Balkans and in Armenia.[601]

About the short conflict with the Persians little is known.[602] They attacked the region of Theodosiopolis and Satala.[603] It seems that the Romans stayed entirely on the defense, eager to come to a quick agreement with the enemy. Theodoret’s miracle stories[604] can be dismissed. But his source correctly connected the events in the East with those in the West:

At a time when the Romans were occupied against other enemies, the Persians violated the existing treaties and invaded the neighboring provinces, while the emperor, who had relied on the peace which had been concluded, had sent his generals and his troops to embark in other wars. Anatolius, magister mililum per orienlem, consented to all demands of the raging tyrant.[605]

By June, 441, the war in Armenia was over.[606]

But there was still another war raging in the western provinces. The Huns had broken into Illyricum. From Priscus[607] we learn that at the time of the annual fair, held at one of the phrouria north of the Danube, the Huns suddenly attacked the Romans and cut down many of them.

When the government in Constantinople protested the breach of the treaty which provided that the fairs should be held with equal rights and with no danger to either side, the Huns maintained that they had only avenged grave injustices done to them. The bishop of Margus, they said, had crossed the river and robbed the royal tombs[608] of their treasures. Besides, contrary to the stipulations of the treaty, the Romans again had sheltered many Hun fugitives. Although the Romans denied these charges, the Huns were undoubtedly right.[609]

Crossing the Danube, the Huns took the important city of Viminacium in Moesia superior. The bishop of Margus, afraid that the Romans, to appease the barbarians, would give him up, treacherously handed the city over to the enemy, “and the power of the barbarians increased to an even greater extent.”

For the following events our main source is Marcellinus Comes. Under 441, he has two entries dealing with the Huns. The first one is a telling example of the way in which Marcellinus thoughtlessly shortened what he found in his sources. “The Persians, Saracens, Tzanni, Isaurians, and. Huns came forth from their countries and ravaged the lands of the Romans. Anatolius and Aspar were sent against them and made peace for one year.”[610]

Who was sent against whom? With which of the enemies was the armistice concluded? Not with the Persians, for the peace treaty which Anatolius signed was not limited to one year; in fact, there was no war between Rome and Persia for more than sixty years, from 441 to 502. The wild Tzanni and Saracens, not to speak of the Isaurian robbers, were not parties with which the imperial government concluded treaties. This leaves the Huns. Anatolius was in the east, commander in chief of the troops in the Orient since 438[611] at the latest. He held the same position in 441, 442[612] and still early in 443.[613] The truce with the Huns was arranged by Aspar, comes, magister militum, and ex-consul.[614]

That Areobindus, not Aspar, was made the commander of the army which finally was sent to Sicily illustrates the hesitations and doubts with which the expedition was undertaken. Aspar knew Africa. He had fought the Vandals in 431; he was in Carthage in 434.[615] He was the most distinguished general of the east. But he stayed in Illyricum, evidently because the situation, in spite of the truce, was too precarious to be handled by anyone else. In addition to his troubles with the Huns, Aspar was confronted with savage rivalries among his generals, which further reduced the fighting power of his army. “John, magister militum, a Vandal by race, was killed in Thrace by the treachery of Arnegisclus.”[616]

Emperor Theodosius II began negotiations with Geiseric. The army in Sicily would possibly soon be needed at another front. It could not be brought back once it was engaged in fighting in Africa. Marcellinus Comes has as last entry under 441 the lines: “The kings of the Huns broke with many of their warriors into Illyricum; they lay waste Naissus, Sin- gidunum, and other cities, and many towns in Illyricum.” In 442, “the brothers Bleda and Attila, the kings of many peoples, ravaged Illyricum and Thrace.”[617]

Ignoring the campaign in 441, Prosper has under 442: “Because the Huns ravaged Thrace and Illyricum with wild devastation, the army, which stayed in Sicily, returned to defend the eastern provinces.”[618] In 442, Theodosius made peace with the Vandals.[619]

The first phase of the war can be reconstructed at least in its outlines, but its second phase is highly controversial. Since the publication of the sixth, posthumous volume of the great Tillemont’s Histoire des empereurs, more than three hundred years have elapsed. Gibbon, Wietersheim, Giilden- penning, Kulakovskii, Bury, Seeck, Stein, and Thompson struggled with the chronological problems of the Hun wars in the 440’s. Yet not a single date seems to be definitely established. When did the Huns take Philip- popolis and Arcadiopolis ? In 441–442, as Thompson assumes, or in 447, as Tillemont and Seeck maintained? When was the peace concluded of which Priscus speaks in fragment 5? Bury, Stein, and Thompson dated it in 443, Gibbon insisted on 446, Wietersheim and Kulakovskii favored 447, and Tillemont thought that the war did not end before 448. It would seem that the available evidence admits practically any date.

The crux is the long entry under a.m. 5942 in Theophanes’ Chrono- graphia. Its importance for the events in the 440’s is obvious. That it cannot be accepted as it stands is, or should be, equally obvious. However, some historians were, and still are, making use of the passage as if it were written by Clio herself.[620] As a matter of fact, the long entry is a galimatias unusual even for Theophanes, who wrote in the ninth century. In a.m. 5942 the following events are said to have occurred:

1. Emperor Theodosius II, recognizing that he had been deceived by Chrysaphius, exiled the eunuch to an island.

If this were true,[621] it would lead to the first months of 450.

2. Empress Eudocia withdrew from the court and went to Jerusalem.

This was in 443[622] or 444.[623]

3. On Theodosius’ orders Presbyter Severus and Deacon John were executed.

This happened in 444.[624]

4. Pulcheria had Bishop Flavian’s remains brought back to Constantinople and laid in the Church of the Apostles.

The translation took place in November, 450.[625]

5. Pulcheria converted a Jewish synagogue into a church, Geoxonoq T(OV XaXHOTCQaTElCDV.

This might be true,[626] but it should be noted that in another passage[627] Theophanes gives Justin II credit for the pious deed.

6. “While the army was in Sicily, waiting for the arrival of Geiseric’s ambassadors and the orders of the emperor, Attila, Mundius’ son, the Scythian, overthrew Bdella, his older brother, made himself the sole ruler of the kingdom of the Scythians, who are also called Huns, and overran Thrace. Thereupon Theodosius made peace with Geiseric and recalled the army from Sicily. He sent Aspar with the forces under him, Areobindus, and Argagisclus against Attila who already had taken Ratiaria, Naissus, Philippopolis, Arcadiopolis, Constantia, and many other towns, making many prisoners and amassing an enormous booty. In a succession of battles the Roman generals suffered heavy defeats, and Attila reached the sea, both the Pontus and Propontis, at Gallipolis and Sestus. He took every town and fortress except Adrianople and Heraclea, even the fortress Athyras. Theodosius saw himself forced to send ambassadors to Attila and grant him 6,000 pounds of gold for the retreat as well as an annual tribute of 1,000 pounds.”

7. Theodosius II died (July 2, 450).

8. Pulcheria married Martian, who was proclaimed emperor (August 24, 450).

The end of Theophanes’ account of the war agrees, more or less, with the beginning of Priscus, fragment 5:[628] After the battle in the Chersonesus the Romans, through the ambassador Anatolius, concluded peace with the Huns. The fugitives were to be handed over, the arrears of tribute, 6,000 pounds of gold, to be paid at once. The annual tribute was fixed at 2,100 pounds of gold.

Theophanes squeezed within twelve months events which lay as much as eight years apart. The war with the Huns broke out (a) while the greater part of the army stood in Sicily, thus in 441–442; (b) after Bleda’s death, thus 444 at the earliest; the war is (c) placed in a.m. 5942, which began on March 25, 450, and the forty-second year of Theodosius II, which was conventionally reckoned from September 1, 449, on. If (a) is right, (b) and (c) are wrong, and vice versa.

It could be argued that, since we know from other sources that the Huns did invade the Balkan provinces at the time of the Sicilian expedition, Theophanes had this first Hun war in mind, brought in Bleda’s death by mistake, and in this way got mixed up in his chronology. This is, indeed, the opinion of most students of the late Roman Empire. They assume that the war which Theophanes mentions is the one that broke out in 441 and ended in 442 or 443. Consequently they date the events described in the Priscus fragment 5 in the same years.

The few dissenting interpretations have been practically ignored. Tille- mont, who dated the war of a.m. 5942 in 447,[629] is almost forgotten. Ku- lakovskii held the same view, [630] but his excellent work, written in Russian, published in Kiev, remains unknown to Western scholars. It is true that Seeck came to the same conclusions.[631] However, like Tillemont and Ku- lakovskii, he merely opposed his chronology to the generally accepted one without stating his reasons.

The following considerations are not meant to establish the actual sequence of events for its own sake. Sub specie aeternitatis they are trivial. But the historian, a loyal citizen of the civitas terrena, cannot help going into details if he wants to determine Attila’s place in the history of the Huns and the Roman Empire. •

1. When did the Romans pay Attila 6,000 pounds of gold? The strain on the imperial treasury must have been heavy. Priscus may have exaggerated the hardships that befell the Romans, yet it is quite credible that many had to sell their furniture and the jewelry of their wives to raise the money the inexorable tax collectors demanded from them. A few are said to have committed suicide in their desperation.[632] Whether the tax load could have been more justly distributed need not be discussed. After paying 6,000 pounds at once and being forced to pay, year after year, 2,100 pounds of tribute, the government could not very well reform the tax system.

If the war that put such a heavy burden on the unfortunate East Romans was the one which ended in 442 or 443, one should think that the taxes in 444 were exceptionally high. The last thing one should expect would be a tax reduction. But the taxes were reduced in 444. “The exaction of delinquent taxes for the past time is remitted for the landed estates... and in the future no such tax assessment shall be feared.”[633] This edict was issued in Constantinople on November 29, 444. It alone would be sufficient proof that the great war, which ended with the financial catastrophe, took place after the issuance of the edict.

2. In the late spring of 443, Theodosius made a journey through some provinces of Asia Minor. He stayed some time in Heraclea in Bithynia;[634] the emperor had a predilection for that province which, in tribute to his uncle, he renamed Honoria.[635] Turning south, he leisurely traveled to Caria. At the end of May, he was in Aphrodisias.[636] On August 27, he returned from the expeditio Asiana to Constantinople.[637] In the spring of 443, the war must have been over. Theodosius hardly could have left the capital while the fighting was still going on. But if he did, he would have crossed over to Chalcedon, as, for example, Leo did after the great fire of 465,[638] and stayed there. In the dedication of his Church History to Theodosius, Sozomen was flattering the emperor, but he could not have written about the journey the way he did[639] had Theodosius been on the flight from the Huns. Furthermore, there is good evidence that the war practically ended in 442. On January 11, 443, the Thermae Achilleae, rd dryioaiov Xovxqov 6 ’AxiAAstit;, were solemnly opened.[640] The people of Constantinople were certainly pleasure-loving, but it is hard to imagine that they should have been in the mood to celebrate the opening of a new bath at a time when the Huns stood at the gates.

3. St. Hypatius, abbot of the monastery of Drys, a suburb of Chalcedon, died in June, 446.[641] Seven months later began the earthquakes which tumbled a large part of the great land wall of Constantinople. And then came the Huns. Callinicus, the biographer of Hypatius, was a conscientious chronicler. He not only recorded the many miracles his hero worked; he also kept a sharp eye on all secular events which affected his brethren. Callinicus would not have passed over a war in which the enemy came close to Constantinople. Indeed, he did not. But the only war of which he knew was the one in 447.

4. Evagrius mentions only “the famous war of Attila” in 447.[642]

5. Jordanes must have read in his sources that Bleda and Attila devastated Illyricum and Thrace in 441 and 442. But he mentions this first war neither in the Romana nor the Getica. As for Callinicus and Evagrius, for Jordanes there existed only one Hun war, the great war in 447.

6. Had the battle on the Chersonesus marked the end of the fighting in 442 or 443, the Romans would have had to negotiate the peace conditions with Attila and Bleda. In fragment 2, dealing with the first phase of the war in 441, Priscus speaks of the kings of the Huns. In fragment 5, which is supposed to conclude an account of the events in 443, Bleda’s name does not occur. Anatolius has to deal with Attila, and only with him. Attila is the king of the Huns,[643] the Hun army is his army.[644]

7. During the first war Anatolius was not in Thrace but in Antioch, the headquarters of the magister militum per orientem. When he concluded peace with Attila he was magister militum praesen tails.

All these data establish the date of the war in the Priscus fragment 5 beyond any reasonable doubt. It took place in 447.

We may now summarize: In 441, the Huns broke into the western Balkan provinces. After a short campaign, during which they took Vi- minacium, they agreed to a truce. In 442, the attacks were resumed. The Romans, led by Aspar,[645] suffered one defeat after another. After the fall of Margus, the key to the Morava Valley, the Huns pushed south and took Naissus.[646] Even if we did not know from Marcellinus Comes that Singidunum was lost in that year, we would have to assume that the defense system along the Danube and the Sava broke down. The road Sirmium-Singidunum-Margus-Viminacium-Naissus was for all practical purposes and, especially, for military purposes the only one that connected Pannonia secunda and Moesia superior with Thrace. With the fall of Naissus the fate of Singidunum was sealed. Everything west of Singidunum now was bound to fall to the Huns. They took Sirmium.[647] They broke into Thrace. Then something must have happened to the Hun armies. They may have been hit by epidemics as later in 447 and again in 452. There may have been an uprising in their rear that forced them to break off the campaign and turn against the rebels. Perhaps some of the peoples such as the “Sorosgi,” with whom Attila and Bleda had waged wars before, used their chance and attacked the Hun heartland while the main strength of their enemy was engaged elsewhere.

In preceding paragraphs, I adduced some arguments for the end of the war before the beginning of 443. The law of August 21, 442,[648] suggests that at least in most of the provinces it was over even earlier. The reference to the advocates, who resumed their practice in the Illyrian prefecture, presupposes that large parts of it were again firmly under Roman control In the fall of 443, the Danube flotilla was being strengthened, the camps along the river were repaired, the garrisons along the limes brought up to full strength.[649] In the same year or, perhaps, in 444, the Romans stopped the payments to the Huns.

In 447, Attila calculated the arrears of tribute at 6,000 pounds of gold.[650] This was apparently a lump sum, but it must have roughly corresponded to the actual arrears. In the treaty of Margus the annual tribute was fixed at 700 pounds.[651] From 447 on, the Huns received 2,100 pounds per annum, clearly a much higher sum than that agreed on in 442 or 443. Assuming that the latter was double the tribute of the treaty of Margus, say 1,400 pounds, the Romans must have refused to pay the Huns anything as early as 444 or perhaps even earlier, in 443. In any case, whatever the tribute was, it was not paid to the kings of the Huns for a number of years. After one or two payments, the government in Constantinople felt strong enough to repudiate its obligations, and the Huns did nothing. It was in those years that they tried to blackmail the West. The East was too strong for them.

From whatever angle we look at the war of 441–442, the picture is the same. All direct and indirect sources are in agreement. Favored by the absence of the Roman army from the western frontier, the Huns were able to inflict heavy defeats on the Romans. To get rid of the savages, Theodosius paid them off. Once they were back, Theodosius tore up the peace treaty. The Huns had proved to be a formidable enemy, but they were not yet a great power.

The contrast between the war in 441–442 and its results, and the war in 447 is so striking that it calls for an explanation. It cannot be a concatenation of coincidences, a mysterious weakening of the power of the Eastern pars. Between the two wars falls the ascendancy of Attila. Except for three fragments, and a few lines in the Gallic chronicle of 452, Priscus’ account of the great war is lost. So is the chronicle of Eustathius of Epiphaneia who, in the main, followed Priscus.[652] Of the Western authors,

Quintus Aurelius Memmius Symmachus seems to have been the only one to write about the war in 447 but his work is also lost.[653] Prosper does not mention it. Under these circumstances the course of events can be reconstructed only in the broadest outlines.

The Priscus fragment 3 deals with the beginning of the war[654]: Attila, the king of the Huns, assembled his own army and sent letters to Theodosius, demanding the fugitives and the tribute which, under the pretext of that war, had not been paid. About the future tribute envoys should be sent to him. If the Romans should delay or prepare for war, not even he would be able to hold back the hordes. The advisors of the emperor read the letter and declared that the fugitives must not be surrendered; it would be better, together with them to wait for the outbreak of the war. However, envoys should be sent to settle the controversies. When Attila was notified about the decisions of the Romans, he got angry, devastated Roman territory, took some fortresses, and attacked the large and populous city of Ratiaria.[655]

Many historians date the events told in this fragment in 442.[656] This is certainly not correct. Attila is the sole ruler of the Huns, 6 raw Ovvvcdv Paat2.Ev<;. He sends letters to the emperor, he is ready to receive the Roman envoys, he demands the tribute money. There are no more “kings of the Huns.” Bleda is dead. We are, at the earliest, in 445.

The phrase ov8s avzov eti eOeAovto. to ExvOlxov sqtE^Eiv nXfjOot; has often been misunderstood. Thompson circumscribed it by “he would no longer hold back the Huns.”[657] Actually, Attila warned the Romans that, unless his demands were granted, it would not be even in his power to prevent the Scythian mass from breaking loose.[658] The Romans did not pay the tribute Tigo^doet roods tov noXepov. What war? Not even Attila, with all his arrogance, could expect that Theodosius would send him the “subsidies,” as if he were still an “ally,” adhering to the stipulations of the foedus, while he was actually waging war with the Romans. Attila did not fight. He assembled his own army, tov obtslov argatov. The stress is on oIkeTov. There must have been other Huns, not those of Attila, who already were fighting the Romans while he was still negotiating with them. Attila declined any responsibility for “that war.” But he let the Romans know that the EhvOlhov nXfjOoQ leaned toward those who already were raiding and looting Roman territory.

We can, I think, discern three groups of Huns. There was Attila with his army; there was “the Scythian mass,” impatient, dissatisfied with their king, ready to go to war unless they got all the gold they thought they were entitled to; and there were Huns already waging war on the Romans.

This, and only this, is the context into which another Priscus fragment can be fitted. Theodosius sent the ex-consul Senator to Attila. But Senator, “although he had the name of an envoy,- did not dare to go to the Huns by land; instead, he sailed up the Pontus to Odessus (modern Varna), where also the general Theodulos, sent there, stayed.” In the Excerpta de legationibus Romanorum ad gentes, this fragment follows the one on the treaty of Margus and precedes the one that deals with the embassy of 449. Senator was consul in 436. But it does not follow that our fragment can refer to any time between 436 and 449.[659] Again it must be noted that Senator was sent to Attila, which narrows the date to the years 445–449. The men who negotiated with Attila in 447 were Anatolius and Theodulos, the latter as commander of the military forces in Thrace. Senator’s voyage falls, therefore, in 445 or, more probably, 446.

Our fragment has either been ignored or misinterpreted by most students of the Huns. Thompson[660] thinks he can discover in it Priscus’ contempt for the cowardly Senator. There is nothing of that sort in the text. The key to an understanding are the words “although he (Senator) had the name of an ambassador.” They can mean only that he could not assume that the people in the area he had to pass through on his way to Attila would respect his status. And these could be no others than those Huns who, in defiance of Attila, were waging their own war with the Romans. Senator obviously returned to Constantinople without having achieved his purpose. Had he met Attila, we would read about the encounter in Priscus.

Marcellinus Comes has four entries under 447: In a tremendous war, greater than the first one, Attila ground almost the whole of Europe into the dust; the walls of Constantinople collapsed in an earthquake and were rebuilt in three months; Attila came as far as the Thermopylae; Arnegisclus, after bravely fighting and killing many enemies, fell in a battle against Attila near the river Utus in Dacia ripensis. The last battle occurs also in Jordanes, who adds that Arnegisclus was magister militum Mysiae, set out from Marcianople, and went on fighting even after his horse was killed beneath him.[661] The fall of Marcianople and the death of the brave general occurs also in the Chronicon Paschale.[662] All three references to Arnegisclus clearly go back to the same source. But where did Marcellinus read that Attila went to war before the earthquake ? And could he actually have meant to say that the war began between January 1 and 27? Obviously not. It would seem that what is now the first entry under 447 was originally the last one under 446. From the Gallic chronicle of 452 nothing about the course of the war can be learned. Yet there are three sources which, in combination, throw some light on the sequence of the events.

We start best with the great earthquake. In 439, under the direction of Cyrus, praefectus urbis, the Anthemian wall, which protected the city only against attacks on land, from the west, was extended along the Golden Horn and the Sea of Marmora.[663] A part of it collapsed on Sunday, January 27, 447, in the second hour after midnight.[664] The whole district between the porticus Troadensis, near the Golden Gate, and the Tetrapylon, where now the Sahzade Mosque stands, was in ruins. When the morning came, ten thousand walked barefoot, the emperor at their head, to the campus of the Hebdomon where the patriarch held a special service.

Whether by that time the Huns already had opened hostilities or not, it was important that the walls were rebuilt as quickly as possible. They were. Flavius Constantinus, praefectus praetorio orientis,[665] mobilized the circus parties. He assigned to the Blues the tract from the Blachernae to the Porta Myriandri, and to the Greens the tract from there to the Sea of Marmora.[666] He had the moats cleared of rubble, “joined wall to wall,”[667] and built new towers and new gates. At the end of March the land wall stood as before, “even Pallas could not have built it quicker and better.”[668]

In the Bazaar of Heracleides the ex-patriarch Nestorius could not pass over the earthquake, for it proved once more what happened to those “who denied that God the word was immortal and impassible.” He wrote:

God shook the earth with earthquakes, the like of which there was none that remembered.... In Constantinople, the imperial city, the the towers of the wall collapsed and left the wall isolated. [This was at a time] when the barbarian again was stirred up against them, massacring and swarming over all the land of the Romans and overturning everything. And they had no means of escape nor refuge but were stricken with fear and had no hope. And he had closed them in and made them insufficient in everything they were doing for their salvation; and, because they understood not their former salvation, he had sent this man whom he had taken from pasturing sheep, who had protested against the privy purposes of the heart of the emperor. And already he had been stirred up by God, and he commanded to make a cross; and as though he, that is, the emperor, believed him not, he made it of wood with his own hands and sent it against the barbarians. But he planted another cross also within the palace and another in the forum of Constantinople in the midst of the city that it might be seen of every man, so that even the barbarians, when they saw it, fled and were discomfited. And the emperor himself, who was ready to flee, gained confidence to remain, and the nerves of the city which was enfeebled grew firm and all things happened thus.... The barbarians fled in discomfiture, while none was pursuing them, and the emperor was mightily heartened to engage in thought of his Empire. [But the barbarians returned, and this time the Romans became] the slaves of the barbarians and were subjected into slavery by the confession of written documents. The barbarians were masters and the Romans slaves. Thus the supremacy had changed over to the barbarians.[669]

The text is not very clear, possibly because of the awkwardness of the Syriac translator. Still one gets the impression that in his exile Nestorius had received some rather detailed information about the war in Thrace. What he wrote about the flight of the Huns is, indeed, confirmed by the Homily on the Royal City by Isaac of Antioch, another of those documents which have been ignored by the students of the Huns:

Again offer up praise to the power which delivered thee from the sword, again give thanks to the cross that it may again fence in the breaches. He [i.e., God] did not wear away the strength in war, thou did not see the faces of the pursuer—by means of sickness he conquered the tyrant who was threatening to come and take thee away captive. Against the stone of sickness they stumbled and the steeds fell and their riders,— and the camp which was prepared for thy destruction was silenced.... With the feeble rod of sickness he smote mighty men and laid them low, and fierceness could not stand before the feebleness which struck at it. With a mean and weak staff he bound for thee the warlike forces, the swift ones sought their feet but sickness weighed them down. The horse came to nought, the horsemen came to nought, and the arms and the assault came to nought.... Through sickness he laid low the Huns who threatened thee.... By his fiat will he caused the sword to cease.... The Hun desired thy property and from desire he changed to wrath—his desire was transformed into anger and it roused him to war and sword. The greedy one mingled desire with wrath and dared to come against the city—for this is the character of plunderers that from desire they come to quarrel. The Hun in the midst of the field heard about thy majesty and envied thee, and thy riches kindled in him the desire to come for the plundering of thy treasures. He called and gathered together the beasts of the field, the host of the desert that he might bring the land into captivity. He hung the sword from his right hand and he had laid his hand on the bow and tested it with the arrow which he sent forth through it. But the sinners drew the bow and put their arrows on the string—and preparation had perfected itself and the host was on the point of coming quickly—then sickness blew through it and hurled the host into wilderness.... He whose heart was strong for battle waxed feeble through sickness. He who was skillful in shooting with the bow, sickness of the bowels overthrew him—the riders of the steeds slumbered and slept and the cruel army was silenced. The assembled army in which the Hun had boasted fell suddenly. Lo the tumult of the battles has died away.... The war with the foreigners has come to an end.[670]

C. Moss, the translator of the Homily, dates it to 441: “as it is obviously impossible that the author could have recorded the events of 447 without mentioning the great earthquake.” This is not a convincing argument. An earthquake has no place in a homily on the royal, rich, flourishing city. Besides, in 441 and 442, the Huns were not even close to Constantinople.

Two writers, independently of each other, describe the flight of the Huns: Nestorius conjuring up a God-inspired shepherd and the Syrian preacher, though with much flourish, ascribing it more soberly to “a sickness of the bowels.” This has, incidentally, a parallel in the siege of Constantinople by the Arabs in 717: they, too, were hit by an epidemic, “and an innumerable number of them died.”[671] The two pious writers with all possible exaggerations preserved to us a phase, or, at least, an episode in the war of 447. In April or May, after the walls had been rebuilt (that it was after the earthquake we learn from Nestorius), a group of Huns advanced to the Bosporus, and the walls collapsed. In 452, five years later, the Huns broke off the campaign in Italy when illness “hit them from heaven.” In 447, they were more fortunate. The main army, under Attila as we may assume, apparently was not infected by the pestilence.

Callinicus certainly knew about the retreat of the advance group, but it paled into nothingness in comparison with the terrible fate that befell the poor people in Thrace:

The barbarian people of the Huns, the ones in Thrace, became so strong that they captured more than a hundred cities and almost brought Constantinople into danger, and most men fled from it. Even the monks wanted to run away to Jerusalem. There was so much killing and blood-letting that no one could number the dead. They pillaged the churches and monasteries, and slew the monks and virgins. And they devastated the blessed Alexander and carried away the treasures and heirlooms, something that had never happened before, for although the Huns had often come close to the blessed Alexander, none of them dared to come near the martyr. They so devastated Thrace that it will never rise again and be as it was before.[672]

The “blessed Alexander” was the church of the martyr in Drizipera, the present Karishtiran, on the road from Heraclea (Perinthus) to Ar- cadianopolis.[673] From Priscus we learn that the Roman army was defeated in the Chersonesus, from Theophanes that the Huns reached the sea at Callipolis and Sestus, and that Athyras was occupied.

Theodosius begged for terms. Anatolius, who negotiated with Attila, was not in a position to reject anything the Hun king demanded. The arrears in tribute had to be paid at once; they amounted to 6,000 pounds of gold. The annual tribute was set at 2,100 pounds.[674] This was harsh enough. But most dangerous for the future was the evacuation of a large territory south of the Danube, a belt “five days’ journey wide,” from Pannonia to Novae (modern Sistova).[675] Most towns within the march, and many to the south and east of it, had been laid waste. Naissus was deserted; when Priscus saw it in 449, the ground adjacent to the bank of the river was still covered with the bones of the men slain in the war; there were only a few people in the hostels.[676] Serdica was destroyed. But slowly, hesitatingly, timidly, the people who had fled would come back. In the march Attila would not even admit a Roman shepherd. He demanded again and again the strictest fulfillment of the treaty provisions.[677] Only the peasants, like peasants everywhere and at all times, tenaciously clung to their land. They fled when the Huns came, taking with them what they could carry, driving the cattle into the woods, and then filtered back when the storm had blown over. The emperor was as unable to drive the peasants out as the Huns were.

And yet, even though the march was not as devoid of any population as Attila wanted it to be, it served its purpose—it left the Romans defenseless. The Huns did not want, or need, the march for their herds and flocks; it was little suited to their extensive cattle and sheep breeding. Attila may have liked to go hunting there,[678] but there were other hunting grounds in his kingdom. The Huns aimed at one thing: at pushing the Romans back from the Danube, thereby removing the main obstacle that could prevent them from breaking into the empire. The Danube limes was not impenetrable. In the winter the riverboats were immobilized; the mostly barbarian garrisons in the forts were not entirely reliable, and even if they were, they could be overpowered. But it cost the Huns much blood to break through the frontier defenses. Despite its weaknesses, the defenses of the Balkan provinces along the Danube had been incomparably stronger than what the Romans now could hope to build up south of the new march. They were at Attila’s mercy.

The war was over in the fall of 447.[679] It began, if my reading of the sources is right, with an incoordinated attack of Hun hosts; when it ended with the greatest victory the Huns ever won, Attila was the ruler of a great power. Our texts tell us nothing about the apportionment of authority within the “Royal Scythians” after Bleda’s death. When the big war broke out, Attila’s authority, though great, was still not quite firmly established. The victory was his victory. From 447, Attila, king, commander in chief, supreme judge,[680] was unconditionally obeyed.

Attila’s Kingdom

To determine the expansion of Hunnic power in the middle of the fifth century is a thankless task. A sober approach is bound to hurt feelings of pride and clash with long-cherished myths. Although no one in Hungary really believes any longer in the great Attila of the medieval chroniclers, his image has not lost its hold over the imagination. To be sure, the peasants, bearers of the national tradition, always named their boys Istvan and Lajos, but in Budapest and Debrecen there still live not a few Attilas.[681] In the Germanic countries, Attila, milte and terrible at the same time, became at an early time a figure of superhuman greatness.[682] Even historians cannot free themselves from the idea that the Hunnic king was a forerunner of the great Mongol captains. Grousset subtitled his L’Empire des steppes “Attila, Gengiz-Khan, Tamerlane.” Attila’s kingdom, he wrote, englobait et entrainait toils les Barbares sarmates, alains, ostrogoths, gepides, etc., repandus entre I’Oural et le Rhin. Mommsen thought that the islands in the ocean over which Attila was said to rule were the British Isles,[683] Thompson thinks of Bornholm in the Baltic Sea,[684] Werner turns Bashkiria, 1,500 miles to the east of Attila’s residence, into his province.[685]

The slight heuristic value of comparing Attila’s kingdom with the great Inner Asiatic Mongol empires is, I am afraid, outweighed by the temptation to look for analogies where there are none. The Hun, whatever his ambition may have been, was not regna or mundi, but lord over a fairly well-defined territory. It was not much larger than the one held in the middle of the first century b.c. by the Dacian king Burebista, who in ten years expanded his rule form the mouth of the Danube to Slovakia and subdued the greater part of the Balkan peninsula. Burebista’s meteoric rise and the sudden collapse of his power were not due to the supposed latent possibilities and liabilities of nomadic societies. The Dacians were not mounted archers. Or, to take another example, to compare Attila with the Gothic condottiere Theoderic Strabo (“the Squinter”) may to some sound sacrilegious. But with all the differences in magnitude the two have much in common. For a few years in the later half of the fifth century, the Squinter was the terror of the East Romans. He forced them to appoiot him magister militum, of course with the salary that went with it. He defeated one Roman army after the other. In 473, Emperor Leo pledged to pay him 2,000 pounds of gold yearly,[686] only 100 pounds less than what Attila received as annual tribute at the height of his power. Theoderic Strabo was not another Attila, but Attila was not another Genghiz Khan either. After the murder of Bleda, Attila was the sole ruler of the Huns, his “own people,” tov a^evegov EOvooq,[687] and lord of the Goths and Gepids, a mighty warrior, for a few years more than a nuisance to the Romans, though at no time a real danger.

Those romantic souls who still see in Attila Hegel’s Weltgeist zu Pferde, should read the acts of the Council of Chalcedon. Among the voluminous documents there are a few letters with casual allusions to fights between Roman troops and Huns somewhere in Thrace. In the very detailed protocols of the meetings, the Huns are not mentioned. It is true the bishops were passionately involved in their dogmatic quarrels. Still, one could not understand their utter disregard for the deadly danger only a hundred miles away, threatening Christendom with extinction, had it really been so deadly.

In the West, Prosper has not one word about the invasion of Gaul in 451. He may have had a personal reason. In his hostility to Aetius, Prosper may not have wanted to give him credit for the victory. But he could not have passed over the invasion in silence unless he, and not only he, took it for just another of the constant barbarian raids into the empire, an episode as later the Magyar raids were episodes. As in the eighth and ninth centuries no one thought for a moment that the Magyars could make themselves masters of Europe, so the idea would have been absurd to the Romans that Attila could take Constantinople and hold it.

In the west, south of the Danube, Noricum remained a Roman province. In 449, the East Roman ambassadors met Promotus, governor of Noricum, at Attila’s court.[688]

North of the Danube the Langobards successfully defended their independence from the Huns. With the help of the story of Agelmund, Lamissio, and the Vulgares, the disputes between the two peoples can be reconstructed in broad outline. The story is preserved in Paul the Deacon’s Historia Langobardorum, who took it from the Origo Gentis Lango- bardorum, written about the middle of the seventh century. Not in spite of, but because of its gaps and inconsistencies,[689] the Origo is a historical document of the first order. To the living tradition of the Langobards it stands incomparably closer than Jordanes’-Cassiodorus’ History of the Goths to the Gothic cantus maiorum. The story runs as follows:[690]

The Langobards are said to have possessed for some years Anthaib and Banthaib, and in like manner, Vurgundaib. There they made Agelmund their king. He led them over a river, defended by Amazons. After passing it, the Langobards, when they came to the lands beyond, sojourned there for some time. Meanwhile, since they suspected nothing hostile, confidence prepared for them a disaster of no mean sort. At night, when all were resting, relaxed by negligence, the Vulgares, rushing upon them, slew many, wounded more, and so raged through their camp that they killed Agelmund, the king himself, and carried in captivity his only daughter.

Nevertheless, the Langobards, having recovered their strength after these disasters, made Lamissio their king. And he turned his arms against the Vulgares. And presently, when the first battle began, the Langobards, turning their back to the enemy, fled to their camp. Then King Lamissio urged them to defend themselves.... by arms. Inflamed by the urging of their chief, they rushed upon the foe, fought fiercely, and overthrew the adversaries with great slaughter.

Lamissio was followed by Lethu, Hildeoc, and Gudeoc, at whose time Odovacar defeated the Rugi. “Then [under Gudeoc] the Langobards, having moved out of their territory, came to Rugiland and because it was fertile in soil they remained in it a number of years.”[691]

After his victory over the Rugi in the winter 487/8, Odovacar broke their last resistance in 488. Rugiland is Lower Austria, north of the Danube, west of Korneuburg. It is the first identifiable geographical name in the Historia Langobardorum, and 488 the first identifiable date. Everything before seems to be lost in impenetrable fog. Any interpretation seems to be as good as any other.

Kemp Malone dates the war between the Langobards and the Vul- gares in the later half of the second century and places the story in the Baltic. He arrives at this astounding result by taking Vulgares for the Latinized form of Langobardic *Wulg(w)aras—wulg, “she-wolf”, and a Germanic plural suffix.[692] It would be difficult to find a more fanciful etymology, thought up in complete disregard of the text.

Convinced that the Langobards lived in Silesia before they moved to Rugiland, some scholars located the battle at the Oder.[693] Klebel is more specific. According to him, the Langobards defeated the Vulgares in the region of Glogau or still farther to the east.[694] He thinks the Vulgares are the Bulgars of South Russia; he even derives their name from that of the Volga.[695]

The question is not the etymology of Vulgares, but what the ethnic name meant in Paul’s writings. In the Historia the Vulgares are (1) the enemies of the Langobards; (2) a people living among the Langobards in Pannonia, later in Italy;[696] (3) the followers of dux Alzeco, who left his country and joined the Langobards in the reign of Grimoald (662–671); settlers in former Samnium;[697] (4) the Vulgarians at the lower Danube.[698] The Bulgars of (3) and (4) are obviously not the Vulgares of our story. The Pannonian Bulgars (2), probably a tribe, or tribes, who stayed in Hungary after the collapse of Attila’s kingdom, appear under this name only in the 480’s, too late for the story.

As unreliable as the Origo and Paul are when they give the names of the stations of the Langobardic migration,[699] in listing the kings, they follow a tradition in which, like in that of the Goths and Burgundians, the names of the rulers and their succession are well preserved. Lamissio reigned forty years. How long his successor Lethu reigned is not known.[700] Allowing him a reign of only one and a half years, the shortest reign of a Langobardic king known from reliable sources, and assuming that Gudeoc led his people into Rugiland in the first year of his reign, the war with Vulgares would fall in the year 446. The average reign of the Langobardic rulers was nine years. Giving Hildeoc nine years, the victory would fall in the year 439. The computations are admittedly anything but conclusive. Still, both point to the first half of the fifth century. The powerful enemy of the Langobards must have been the Huns. This was conjectured long ago, and should never have been doubted. But why did Paul call the Huns Vulgares ? Because had he spoken of the Huns, his readers might have thought he meant the Avars. In the Historia Langobardorum the Hunni are always the Avars, “who were first called Huns, but afterward from the name of their own king: Avars” (qui primum Hunni, posted de regis proprii nomine Avares appellati sunt).[701] Gregory of Tours, too, called the Avars Huns, and so did a century later the Langobard who wrote the Origo. In Byzantine historiography of the sixth, seventh, and eighth centuries, the use of Ofivvoi for M/tagot is common.[702]

Until recently it would have been impossible to determine where the Langobards fought the Huns. Thanks to Werner’s thorough study of the archaeological evidence,[703] we know by now that southern Moravia was held by the Langobards before they settled in Rugiland. Twenty-four findspots testify to their prolonged stay in this area.

[It is possible that this section is incomplete. —Ed.]

The Huns In Italy

The generally known sources for the Hunnic invasion of Gaul in 451 have been so thoroughly studied that their reexamination is unlikely to yield new relevant results.[704] But there are still a few which have been ignored. We learn, for instance, from the letters of Pope Leo (440–461) and only from them that in the early summer of 451 the West Romans expected Attila to march into Italy.

In this letter of April 23 to Emperor Marcian (450–457), Leo expressed the conviction that by the concord of the two rulers of the empire “the errors of the heretics and the hostility of the barbarians” would be overcome.[705] This could have been said anytime; it is an ever-recurring commonplace. But when in May the emperor decided to convoke an ecumenical council in the East, in Nicaea, the Pope implored him to postpone it; because of the threat of war, the bishops of the most important provinces would be unable to leave their churches.[706] And when Marcian insisted on the convocation, Leo sent Paschasinus, bishop of Lilybaeum, as his delegate; Sicily was “the most secure province.”[707]

Why Attila did not march into Italy but into Gaul is not known. He certainly did not undertake the campaign against the Visigoths because he was bribed by Geiseric, their enemy, as Jordanes asserts.[708] The idea that agents of the Vandal king, carrying bags of gold, sneaked through the empire, from North Africa to Hungary, is grotesque. The sixth-century chronicler Malalas’ worthless account is still being given some credit. Malalas mixed everything up. He called Attila a Gepid, confused Theoderic and Alaric, and shifted the decisive battle from Gaul to the Danube. Attila is said to have sent ambassadors to Rome and to Constantinople who ordered the two emperors to make their palaces ready for him.[709] Gibbon, followed by Thompson,[710] thought he could recognize in the order “the original and genuine style of Attila.” It is rather the style of the most stupid of the Byzantine chroniclers. I disregard the often told melodramatic story of the vicious Princess Honoria, her clandestine engagement to Attila, and what follows from it. It has all the earmarks of Byzantine court gossip.

After the battle at the locus Mauriacus in the first week of July 451,[711] Attila retreated to Hungary. About the situation in Gaul we again find some information in Pope Leo’s letters. Leo was eager to communicate with the transalpine bishops, but it was only in January 452 that their spokesman, Ingenuus Ebredurensis, came to Rome.[712] Evidently the violent rivalry between the Visigothic princes after the death of King Theoderic made all travel impossible.

The acts of the Council of Chalcedon throw a little light on the Hunnic raids into the Balkans in 451. Emperor Marcian issued a summons to meet in Nicaea on September 1, 451. At that time he hoped to be able to be there, “unless some urgent affairs of state should detain him in the field.”[713] He apparently expected trouble on the Danube frontier. It actually broke out in the summer. In August, Marcian asked the bishops assembled at Nicaea to pray for victory over the (unnamed) enemy.[714] He was then in Thrace;[715] fighting was still going on in parts of Illyricum. As no bishop from Moesia prima and Dacia ripensis attended the council when it was finally assembled in Chalcedon,[716] it may be assumed that the Huns were again ravaging the two unfortunate provinces. Scythia, too, was threatened: Alexander, bishop of Tomis, stayed with his flock.[717]

The archaeological evidence is of little help for the reconstruction of the campaign in Gaul. According to Gesta Trevirorum, Attila took Trier. This seems to be borne out by recent excavations: The Eucherius church was destroyed in the early 450’s.[718] A few years ago a fragment of a Hunnic cauldron allegedly was found in northern France, allegedly near Troyes.[719] It gave new impetus to the search for the battlefield near the locus Mauriacus, a favorite hobby of local historians and retired colonels.[720]

All these additions do not change the over-all picture of the events in 451. Attila’s campaign in Italy, however, calls for a reexamination. Nearly all modern historians, from Mommsen to Thompson, took it, first of all, as an occasion to prove what an incompetent statesman and general Aetius was.[721] Some pious souls still regard the war in Italy as a duel between a blundering Roman leader and a bloodthirsty savage that ended happily with the intervention of Pope Leo as pontifex ex machina.[722] Aetius, whom his contemporaries called “the last Roman,” is the bile noire of the moderns. It is not my intention to rehabilitate him. But I want to show, among other things, that Aetius did not make the dilettantic mistakes of which he has been accused.

The First Phase of the War

The losses of the Huns in 451 must have been very heavy. The mere fact that Attila began to move his army early in the summer shows how much time he required to recuperate from the disaster of the year before. He must have been aware of the dangers of a campaign in the hot season. Why did he attack at all? Attila was doubtless “furious about the unexpected defeat he had suffered in Gaul,” to quote the Chronicle of 452,[723] the only source to establish some sort of connection between the two wars. He certainly hated Aetius. But why did he not wait another year to take his revenge? His relationship with the East Romans could not have been worse. When he marched into Gaul, in 451, ostensibly to fight the Visigoths, Emperor Marcian did not move. But Attila could not count on the neutrality of the eastern part of the empire if he invaded Italy. He may have hoped to crush Aetius’ army before the East came to the help of the West. He may have thought that once his horsemen swarmed over the Po Valley Aetius would sue for peace. Perhaps he expected Valentinian to sacrifice Aetius in order to save his throne. We know nothing about the political situation in Italy; it may have been such that Attila had reason to expect a quick collapse of the enemy. But it also may have been the pressure of his own hordes, intent on looting and more looting, that forced the king prematurely to undertake another predatory war. There were times when the Huns in Upper Italy moved very slowly because their carts were loaded with so much loot.

When did the Huns cross the Julian Alps from present Yugoslavia into Italy ? The chronicles do not give the date, and only a few years later, even the sequence of the major events was forgotten. Hydatius, otherwise so well informed, thought that Attila marched from Gaul straight into Italy;[724] according to the Chronicle of 511, the Huns took Aquileia (in the

northeastern corners of the Adriatic Sea) on their retreat from Gaul to Pannonia.[725] Only Priscus seems to give a hint as to the date of Aquileia:

The siege of Aquileia was long and fierce, but of no avail, for the bravest of the soldiers of the Romans withstood him [sc. Attila] from within. At last his army was discontented and eager to withdraw. Attila chanced to be walking around the walls, considering whether to break camp or delay longer, and noticed that the white birds, namely the storks, who build their nests in the gables of houses, were bearing their young from the city and, contrary to their custom, were carrying them into the country. Being a very shrewd observer of events [sagacissimus inquisitor], he understood this and said to his soldiers: “You see the birds foresee the future. They are leaving the city sure to perish and are forsaking strongholds doomed to fall by reason of imminent peril. Do not think this a meaningless or uncertain sign; fear, arising from the things they foresee, has changed their custom.” Why say more? He inflamed the hearts of the soldiers to attack Aquileia again.[726]

If Priscus’ story should contain a kernel of truth, the fall of Aquileia would have to be dated at the end of August or the beginning of September. According to Pliny, the storks leave Italy after the Vulcanalia, August 23.[727] The siege is said to have lasted three months.[728] The Huns would, thus, have crossed the Julian Alps in May or June. But it is more than doubtful that the story of Attila and the storks permits such an interpretation. It rather seems to throw light on the superstitious awe with which his subjects, especially the Germans, looked up at the king.

The movements of birds were considered ominous by Greeks, Romans, and Germans. Like many heroes of Germanic tradition, Hermenegisclus, king of the Varni, understood the language of the birds.[729] The western Germans regarded the raven and the stork as prophetic birds.[730] One could, therefore, conjecture that Germans told Priscus the story; they may have spoken about Attila as in later times Swedes and Norwegians spoke about the dreaded Finnish and Lappish sorcerer. Priscus himself was possibly not above the superstition of the Greeks before and after him, to whom the “Scythians” were great sorcerers. The Hyperborean magician in Lucian’s Philopseudes[731] brings up supernatural beings, calls corpses back to life, makes Hecate appear, and pulls down the moon. In Empress Eudocia’s Discourse with the Martyr Cyprian (Adyoi eh; paQTVQa Kvnqtavdv), Cyprian relates how the Scythians taught him the language of the birds, how he learned to understand the sounds of boards and stones, the creaking of doors and hinges, and the talk of the dead in their graves.[732] Geiseric, another “Scythian,” and, like Attila, “ a very sagacious man,” interpreted correctly the flight of the eagle over the sleeping Marcian.[733] Thus, it is quite possible that the story was told by people disposed to believe it, but it is itself of more remote origin.

The legend of Attila and the storks of Aquileia is, in fact, a variant of a story which occurs in chapter 122 of the Chin shu, the biography of Lii Kuang who reconquered Turkistan for Fu Chien of the Former Chin. In February 384, he besieged Ch’iu-tz’u (Kucha). “He once more advanced to attack the city. In the night he dreamed that a golden image flew over and beyond the city walls. Kuang said: ‘This means the Buddha and the gods are deserting them. The Hu will surely perish.’”[734]

Folklorists presumably will be able to adduce other versions of this story, perhaps connecting them more closely with the proverbial rats which leave the sinking ship, a story widespread in the West. But stories like the ones told about Attila and Lii Kuang are unknown in Europe.[735] It must have been the Huns who brought them from the East.

Leaving the storks of Aquileia, we turn to the letters of Pope Leo; they provide a safer ground for dating the beginning of the Hunnic invasion. On May 22, 452, Leo wrote long letters to Marcian, Pulcheria, Anatolius, and Julian, bishop of Kios, in which he explained why he could not approve of the disciplinary canons passed by the Council of Chalcedon.[736] There is not one word in them that would indicate that Italy had become a theater of war. The same is true for the letter which the Pope sent to Theodor, bishop of Forum Julii,[737] on June 11. The decretal in which he defined the conditions that should govern the granting of absolution in the administration of penance could have been composed in any year of his pontificate.[738] It is unimaginable that the man who dictated it should have passed over the fate of the cities and towns in northern Italy had they been already under attack by the Huns. Aquileia had not yet fallen; probably it was not even besieged.

And then there is the Novella Valentiniana 36 of June 29, 452, on the duties of the swine, cattle, sheep, and goat collectors, a subject obviously of no interest to the students of the Huns, and, therefore, ignored by them. But the Novella is the only document of 452 which contains an allusion to the war. In the introduction the emperor praises Aetius, who even “among his warlike troubles and the blare of trumpets” finds time to think of the meat provision of the sacred city. The object of Aetius’ bellicae curae at the end of June could be no other but the Hunnic invasion. Attila’s hordes descended into the plains in the early summer of 452.[739]

If Prosper were to be believed, the invasion came to Aetius as a complete surprise. He wrote:

After Attila had made up for the losses suffered in Gaul, he intended to attack Italy through Pannonia. Our general had not taken any provisions as he had done in the first war, so that not even the defenses of the Alps, where the enemy could have been stopped, were put to use. He thought the only thing he could hope for was to leave Italy together with the emperor. But this seemed so shameful and dangerous that the sense of honor conquered the fear.[740]

It is amazing that all modern historians believed Prosper. “The news of Attila’s arrival in Italy,” says Thompson, “must have struck the patrician with the violence of a thunderstroke.”[741] Nothing could be further from the truth.

The passes over the Julian Alps, to begin with, can in no way be compared with the Gotthard or even the Brenner Pass. In the Historia Lango- bardorum, Paul the Deacon described the approaches to the peninsula:

Italy is encompassed by the waves of the Tyrrhenian and Adriatic seas, yet from the west and north it is so shut in by the range of the Alps that there is no entrance to it except through narrow passes and over lofty summits of mountains. Yet from the eastern side by which it is joined to Pannonia it has an approach which lies open more broadly and is quite level.[742]

Second, the limes on the Karst[743] consisted of light fortifications, or better roadblocks and watchtowers, garrisoned by forces which were unable to withstand a determined attack.[744] At the best they could delay the enemy; they could not stop them.

Third, and this is most important, Aetius acted exactly as other generals before and after him acted in the same situation. In the course of the fifth century Italy was invaded six times.[745] With the possible, though improbable exception of Radagaisus’ hosts, each time the enemy descended into the plain from the east, each time he crossed the passes of the Julian Alps without having to overcome any resistance. This is true for Alaric in 401 and again in 408, the East Roman army under Aspar in 425, Attila in 452, and Theoderic’s Ostrogoths in 489. Neither Stilicho, nor the usurper John, nor Odovacar defended Italy in the passes. It could be objected that they had not enough troops for both the passes, and, if those were broken through, the plain. But in 388, the rebel Maximus had a very strong army, and yet he made no attempt to stop Emperor Theodosius; the emperor “crossed the empty Alps” (vacuas transmisit AIpes).[746] In 394, the Alps again “lay open” to Theodosius’ army.[747] The struggle for Italy began in the valley of the Isonzo or before the walls of Aquileia.

Attila could not by-pass the strong fortress; its garrison it seems, strengthened in anticipation of the siege.[748] Only after Attila used siege engines,[749] obviously built by Roman deserters or prisoners, were the walls of Aquileia breached and the city stormed. It was thoroughly plundered; those who could not flee in time were massacred or carried away into captivity.[750] The devastation was certainly cruel, but Jordanes’ assertion that no trace of the former great city was left to be seen is one of those exaggerations of the Hunnic outrages which, enormous as they were, were later magnified out of all proportion. By the middle of the sixth century, Aquileia had long been rebuilt. It is true that the fortification had not been restored. The Ostrogoth Theoderic in 489 and the Byzantine general Narses in 552 by-passed the city. But only a few years after the devastation by the Huns, Aquileia was again the seat of a bishop.[751] The Christian community was strengthened as more and more fugitives returned.[752] In the sixth century, a basilica with a splendid mosaic floor was built.[753] The metropolitan of Aquileia was in rank equal to the metropolitans of Milan and Ravenna.[754]

In the Po Valley

Jordanes, probably following Priscus, named three cities which fell to the Huns—Aquileia, Mediolanum (present Milan), and Ticinum (present Pavia)—speaking also vaguely about “the remaining towns of the Veneti” and “almost the whole of Italy.”[755] Paul the Deacon copied Jordanes,[756] but as a native of Cividale he naturally was more interested in the towns of upper Italy than Jordanes. He first enumerated three places near Aquileia: Concordia, Altinum, and Patavium, and then “all the cities of Venetia,” namely, hoc est, Vicetia, Verona, Brixia, and Pergamum.[757] It is quite possible that the Huns indeed took all those places, but Paul’s hoc est makes the list somewhat suspicious.[758]

The Huns crossed the Po and devastated the province of Aemilia. to the south of the river. After his experiences before Aquileia, Attila could not hope to take the incomparably stronger fortress Ravenna. Like Alaric half a century earlier, he could have marched on Rome, where the emperor stayed.[759] However, the Huns did not cross the Apennines. Whether they tried and were repulsed, or were so heavily engaged by Aetius in the plain that Attila could not spare troops for an attack, is not known. Possibly the ranks of the Huns were already thinned out by sickness. Perhaps many horsemen had hastily been ordered back to Hungary. Our only source of information is a passage in Hydatius: “Auxiliaries were sent by the emperor Marcian, and under the commandership of Aetius they [i.e. the Huns] were slain. Likewise they were subdued in their own seats, partly by plagues from heaven, partly by Marcian’s army.” (Missis per Marcianum principem Aetio duce caeduntur auxiliis pariterque in sedibus suis et caelestibus plagis et per Marciani subiuguntur exercitum.)[760]

This often has been misunderstood. It has been maintained[761] that the dux Aetius was an East Roman general, commander of the troops which attacked the Huns in sedibus suis, by coincidence bearing the same name as the West Roman generalissimo. But Hydatius makes a clear distinction between the auxilia sent by Marcian to Aetius and Marcian’s exercitus. There is no reason whatever to postulate the existence of two generals, both named Aetius, both fighting the Huns.[762] The passage sheds some light on the preparations Aetius must have made when the first information about Attila’s plans reached him. Because the East Roman auxilia could not march through Pannonia, they must have sailed from some eastern ports to Italy, probably Ravenna. To move the troops to the ports of embarkation, to assemble the ships, to provide food for the soldiers and fodder for the horses in the landing port, all this required considerable time. An expedition, if it were to be of real help to the West, could not be improvised. Aetius and the government in Constantinople must have worked out a plan of coordinated action against the Huns, should Attila invade Italy.

According to Paul the Deacon, the Huns took Mediolanum and Ti- cinum, but the two cities were not plundered nor the citizens massacred. As a sermon given after the Huns evacuated Milan shows,[763] Paul was misinformed. Many houses and churches were destroyed.[764] The basilica of St. Ambrose[765] was set on fire and collapsed.[766] The Huns killed not a few clerics and laymen.[767] Still many survived, not because the Huns were so mild, but because the Milanese ran away faster than the Huns could pursue them: heavily loaded with booty, the Hunnic carts were too slow, ut velocissimi equiles tarda atque onere gravata suo trepidanttum plaustra fugerunt.[768] It is quite possible, I believe, that Aetius abandoned the cities in northern Italy to slow down the savage hordes. Their loss may have been the price paid to save Rome. Attila took up residence in the imperial palace.[769] His horsemen looted and killed to their hearts’ desire.

The Huns did not stay long in Milan;[770] they evacuated the city and retreated east. Food and fodder for the horses must have been hard to find among the charred ruins; besides, illness hit the Huns “from heaven.”[771] This we learn only from Hydatius. Prosper certainly knew of the epidemic which raged among the Huns but, according to him, the merit of having rescued Italy from the savages belonged exclusively to the Holy Father.

Northern Italy became to the Huns what fifty years earlier it had been to Alaric’s Visigoths, regio funesta;[772] a land of death, where “pestilence raged, brought by foul food and aggravated by the season’s heat.”[773] The situation of Attila’s army must have been almost the same as in 447 when he stood before Constantinople. Perhaps it can be best compared with the fate of the Frankish invaders in 540, who “were unable to obtain any provisions except cattle and the waters of the Po. Most of them were attacked by diarrhea and dysentery, which they were quite unable to shake off because of the lack of proper food. Indeed they say that one third of the Frankish army perished in this way.”[774] The same fate befell another Frankish army in 553; it was almost wiped out.[775] It seems that the Goths, many of whom fought under Attila, were particularly susceptible to epidemic diseases.[776]


In a few weeks, a month at the most, the Huns, under the threefold attack of Aetius’ troops, Marcian’s army, and sickness, would have been forced to ride back to Hungary. At that very moment the Romans decided to open negotiations with Attila. In Prosper’s words:

In all the deliberations of the emperor, the senate, and the Roman people nothing better was found than to send an embassy to the terrible king and ask for peace. Relying on the help of God, who, he knew, never failed in works of piety, the most blessed Pope Leo undertook these negotiations together with the ex-consul Avienus and the exprefect Trygetius. Nor did it turn out otherwise than faith had expected. The king received the whole delegation courteously, and he was so flattered by the presence of the highest priest that he ordered his men to stop the hostilities and, promising peace, returned beyond the Danube.[777]

How should this passage be interpreted? Hydatius, a faithful son of the Church, knows nothing about such an embassy. It need not be proved that in the middle of the fifth century the “Roman people” had nothing to decide, and it is more than doubtful that Attila was so overwhelmed by the saintliness of the pontiff or, as he probably called him, the chief shaman of the Romans, that he meekly made peace. And yet, Prosper did not invent the meeting. Jordanes, too, knew about it and even named the place where Leo met Attila.[778] What, then, was Leo’s mission? He himself never mentioned it, not in his writings, nor letters, nor sermons. Legends embellished at an early time the encounter of the Pope who became a saint and the Hun king who became the “scourge of God.”[779] Paul the Deacon knew of a venerable old man, who, standing at Leo’s side, threatened the king with a drawn sword.[780] The man is the Christian counterpart of Achilles and Athena Promachos who protected Athens from Alaric.[781] Paul’s warrior looks like a combination of Mars and Saint Peter, who “in all emergencies was close” to the Pope.[782] It is hopeless to search these stories for their historical content.

A fortunate coincidence has preserved a letter which the oriental bishops sent to Pope Symmachus in 512 or 513.[783] From it we learn that Leo negotiated with Attila about the release of the captives in the hands of the Huns, not only of the Christians, but also, “if that can be believed,” of Jews and pagans.[784] How successful Leo was we do not know, but it is quite believable, even probable, that Attila released the more prominent prisoners, naturally for a substantial ransom.[785] The others, for whom no one was willing to pay, were dragged off to Hunnia. No chronicler was interested in their fate. What happened to them we shall learn presently.

Attila’s campaign was worse than a failure. He could not force the Romans to conclude another treaty with him, to pay tribute again, or to reappoint him magister militum. The hated Aetius remained the factual ruler of the Western empire. The loot may have been considerable but it was bought at too high a price, too many Hunnic horsemen lay dead in the towns and fields of Italy. A year later Attila’s kingdom collapsed.[786]


In March 458, Nicetas, bishop of Aquileia, asked Pope Leo for help and advice. His letter has not been preserved but its content can be reconstructed from Leo’s answer.[787] In the beginning the Pope speaks of the wounds inflicted by the attacks of the enemy. Through the disasters of war and the grievous inroads of the enemy, families were broken up, the men were carried off in captivity and their wives remained forsaken. Now, through the Lord’s help, things have turned to the better. Some of those who were thought to have perished have returned. Leo decided that the women who had remarried should go back to their former husbands. He let the bishop know what should be done with those who, while captives, were, by hunger and terror, compelled to eat sacrificial food, and those who were baptized by heretics. Leo concluded the letter with the request that its contents should be brought to the knowledge of the bishop’s brethren and fellow bishops of the province.

Half a year later, Neon, bishop of Ravenna, had more questions to ask. Some of the returned prisoners craved the healing water of baptism; they went into captivity when they could have no knowledge of anything, and in the ignorance of infancy they could not remember whether they had been baptized or not. Should they be baptized, which could mean that perhaps they would be baptized twice?[788]

The theological and moral problems are of no interest to the present studies. What matters is the fact that men and children who a few years earlier had been dragged into captivity returned to the dioceses of Aquileia. The only enemy who had been there were the Huns. The circumstances under which the Huns let the prisoners go back will be discussed in a later chapter. When one considers how many must have died and how many still came back, one can imagine how large the number must have been of those for whom no one paid ransom. It should be noted that the Pope said nothing about women who came back. They apparently stayed in the harems of the Hunnic nobles and went with them, after the breakdown of the kingdom, to the northern Balkan provinces. The heretics who baptized the children must have been Arian Goths and, possibly, Gepids.[789]

Collapse and Aftermath

Attila spent the last months of his life preparing a campaign against the East. In the early fall of 452, his ambassadors threatened Emperor Marcian that their lord would “devastate the provinces because that which had been promised him by Theodosius was not paid; the fate of his enemies will be worse than usual.”[790]

The Priscus text which Jordanes followed is partly preserved in fragment EL 9.[791] Attila sent the ambassadors after he returned from Italy and before the events referred to in fragment EL 11. The peace which Maximus concluded with the Blemmyes and Nobades provided that the barbarians were again admitted access to the Isis temple on the island of Philae. As a priest of the Blemmyes (a tribe living in the modern Sudan) dedicated an inscription to Osiris and Isis on Philae on December 19, 452,[792] the peace cannot have been concluded later than in October or November.[793] The emperor rejected the demands of the Hun even more firmly than in 450. At that time, Attila had been at the height of his power. Since then his glory had faded, the myth of his invincibility had been exploded, his armies defeated, his resources greatly diminished. Although the battle of the locus Mauriacus was in the strict sense undecided, for the Huns, a battle which cost them thousands of horsemen and in which they neither took prisoners nor could rob the dead was a lost battle. The invasion of Italy ended in failure. It now had been years since Attila had received tribute from either half of the empire. Although we have no evidence of unrest among his German subjects or among the ruling group of the Huns, we may safely assume that the former were more heavily exploited than ever and the latter grew increasingly dissatisfied with the king who failed to provide them with booty and gold. Nevertheless, Attila remained a formidable adversary. Marcian was “disquieted about his fierce foe.”[794] Fortunately, Attila died in the beginning of 453.

The contemporary Prosper, whom Cassiodorus copied, the sixth-century chronicler Victor Tonnenensis, and the Gallic chronicle of 511[795] agree on the year. According to Hypatius, Attila died shortly after his retreat from Italy;[796] “454” in Marcellinus Comes[797] is certainly wrong. The circumstances of Attila’s death were soon embroidered with all kinds of inventions.[798]

Except for a not very enlightening entry in Marcellinus Comes and a few insignificant lines in the Vita s. Severini,[799] Jordanes is the only source for the events after Attila’s death. In the Romana he mentioned it in passing. For the Getica he drew on Cassiodorus and, either directly or through Cassiodorus, on Priscus, occasionally looking at a map;[800] in the main, he relied on oral tradition. The results are meager. The history of the transdanubian barbaricum after Attila’s death can be reconstructed only in the broadest outlines:

Attila’s sons were “clamoring that the gentes [sc. the Hunnic gentes] should be divided among them equally and that warlike kings with their populi should be apportioned to them like a family estate [instar familiae].” A coalition of Germanic tribes, led by Ardaric, king of the Gepids, revolted.[801] After a succession of battles they defeated the Huns in Pannonia at the Nedao River. Among the alleged thirty thousand slain Huns was Ellac, Attila’s eldest son.

The Goths did not fight on the Nedao.[802] Some Goths may have joined the rebels; others probably remained loyal to the Huns. Goths trekked with Huns as late as 468. The great mass of the people remained neutral. On this point there is general agreement. There remain two questions: When was the battle at the Nedao fought, and where is the Nedao?

Revocatio Pannoniarum

According to Marcellinus Comes, the fight between the savage peoples was still raging in 453. Before the summer of 455, the Huns were defeated. This follows from Sidonius Apollinaris’ Panegyric on Emperor Avitus. In the autumn of 455, Avitus “recovered the lost Pannonias after so many generations by a mere march.”[803] These two verses puzzled historians for more than two hundred years. Je ne sgai pas quelle verite il peut y avoir dans ce que dit Saint Sidoine, qu’Avite avoit reuni lesPannoniens, wrote Tillemont, and it would seem that we are not much wiser.

There is, first, the question of the date. Avitus was proclaimed emperor in Arles on July 9, 455; on September 21, he entered Italy.[804] On January 11, 456, he was in Rome where Sidonius delivered the panegyric.[805] When was he in Pannonia ? Not between July and September, as Stevens thought.[806] Even if Avitus left Gaul immediately after he was raised to the throne, which is improbable, he could not have journeyed north, turned east, crossed Raetia and Noricum, received the submission of the barbarians in Pannonia, and still have been in Italy by September. He could not have gone later, either. To cross Italy from her western border—Avitus came from Gaul—to the Julian Alps and to proceed to Pannonia, to stay there, even only a few days, and to return in winter to Italy in order to be in Rome at the end of the year required incomparably more time than Avitus had in the last quarter of 455. Not Avitus, but one of his officers was in Pannonia. What he achieved there was, in the tradition of the paatXixdt; Xdyoq, attributed to the emperor.[807]

Sidonius lavished the most extraordinary praises on Avitus. Like Claudian on such occasions before him, he set the whole divine machinery in motion to present to the assembled senators the new emperor as the savior of the world. Jupiter tells the gods and goddesses, even the fauns and satyrs, the exploits of the hero. Well informed by the poet who happened to be the emperor’s son-in-law, the thunderer not only describes minutely all that Avitus had done for the beloved Gaul, he also invites the listeners to go back with him to the years when the future Augustus, still a boy, emulating Hercules, killed a she-wolf with a stone. He leaves out nothing; he spends forty hexameters on a duel between Avitus and a Hun. The panegyric contains 603 lines. But only at its very end, in one and a half verses, does Sidonius allude to what one should think was the most glorious deed of Avitus—the recovery of the Pannonian provinces, “whose march alone sufficed to recover the Pannonian provinces” (Cuius solum amissas post saecula multa / Pannonias revocavit iter). The discrepancy between Sidonius’ words and the importance of the event is striking. The Western empire was losing one province after another to the barbarians. Only seven months before, the Vandals had looted Rome so thoroughly that we still speak of Vandalism. At such a time the recovery of the two long-lost Pannonias should have been hailed by Sidonius as the beginning of a new era. He should have compared Avitus to Alexander, Scipio, the divine Julius, and all the other great captains of the past. But instead of singing a paean, Sidonius whispers, as if not to forget, that the Pannonias were Roman again.

Seeck assumed that Avitus went to Pannonia to reconquer the country from the Huns and to wait there, closer to Constantinople, for recognition by Marcian before he presented himself to the senators in Rome,[808] to which Stein rightly objected that for Avitus to interfere in the Danube provinces would have been about the worst way of winning Marcian’s favor; he thought Avitus’ action actually forfeited his recognition by the Eastern court.[809]

The revocatio Pannoniarum is not mentioned in the chronicles. It is true, they are terse, but not so terse that they could have ignored such a momentous event. Neither the Western nor the Eastern sources contain as much as an allusion to the alleged reconquest of the provinces. However, the Fasti Vindobonenses priores have under 455 a curious entry: “and Sabaria has been destroyed by an earthquake seven days before the Ides of September on a Friday” (et eversa est Sabaria a terrae motu VII idus Septembr. die Veneris).[810]

As long as Pannonia prima was under Roman rule, Sabaria, the present Steinamanger (Szombathely), was the most important town of the province. The last indirect reference to it occurs in the Notitia Dignitatum of the early fifth century where (occ. VII, 82) the lanciarii Sabarienses are listed. Like the other Romans in Pannonia, the people of Sabaria must have lived a wretched life in the first half of the fifth century, but somehow they held out, possibly because they arrived at an agreement with Germanic settlers in the neighborhood who, as well as the Huns, needed the craftsmen of the town. When the power of the Huns broke down, there were still Romans in Sabaria. The entry in the chronicle, exact down to the day of the week when the earthquake struck, presupposes the resumption of relations with the West Romans. Pannonia did not become Roman again—otherwise Sidonius would have spoken in a different vein—but it was again, however loosely, in the orbit of the empire.

Sidonius contrasted iter to bellis. Avitus’ officers did not fight in or for Pannonia, neither against the Huns nor any other barbarians. Avitus needed all the troops he had, his own and, hopefully, what the Visigoths might send him for the war against Geiseric’s Vandals. Avitus will restore Libya to Rome, exclaims Sidonius, immediately before he comes to the recovery of Pannonia. The only conceivable reason for Avitus’ officers going to Pannonia must have been the same which Marcian had two years later, when he sent his officers to the Danube countries to recruit soldiers for the war against the Vandals. Avitus could not expect to succeed as long as Pannonia was under the Huns. And this, together with the report on the earthquake in Sabaria, presupposes that by that time the power of the Huns had collapsed. The battle at the Nedao was fought in the summer of 455 at the latest.

A passage by Sidonius in Panegyric on Anthemius points to a still earlier date. In 454, the future emperor, then a comes, “traversed the banks of the Danube and the whole length of the wide border, exhorting, arranging, examining, equipping.”[811] The verses reflect the situation after the battle at the Nedao, when the defeated Huns, some of their Alanic allies, and splinters of the Germanic tribes which had fought on the side of the Huns crossed the Danube at a number of places and settled there. The battle was fought in 454.

The Nedao River

The name Nedao occurs only in Jordanes. This does not necessarily mean that the Nedao was an insignificant streamlet. The names of three more rivers in Pannonia[812] and two in Dacia[813] occur likewise only in Jordanes; he is the only author to call the Bug, certainly a major river, by the enigmatic name Vagasola.[814] On the other hand, it is conceivable that tradition preserved the name of a brook as small as the Katzbach where Blucher defeated Macdonald in the fall of 1813.

None of the various attempts to locate the Nedao has succeeded. It cannot be the Neutra, a left tributary of the Danube,[815] because the Neutra is not in Pannonia. Nato, mentioned once in Marcellinus Comes,[816] sounds vaguely similar to Nedao, but has nothing to do with it;[817] Nato was a fortress near Horreum Margi, a town not in Pannonia but in Moesia superior.[818] Netabio, according to the Anonymus of Ravenna a civitas in Pannonia,[819] was possibly named after the river;[820] unfortunately, Netabio cannot be located. Nedao might be the ablative of *Nedaus< *Nedavus,[821] a name that sounds Celtic. If it were, it would point to southern Pannonia, the old country of the Sordisci and Taurini; Neviodunum, Mursa, Sopiana, Taurunum, Cornacum, and Singidunum, all in southern Pannonia, were originally Celtic settlements.[822] Nedao might be Illyric.[823] The form of the name is of no help in locating the battle.

Nor can any conclusion be drawn from the singular Pannonia in Getica 260.[824] Jordanes uses both, Pannonia and Pannoniae, with a slight preference for the former.[825] Pannonia may be the name of the whole territory between Vindomina (Vindobona) and Sirmium;[826] or of one of the two provinces, Pannonia prima and secunda;[827] or of both. It may simply mean the former Roman land east of the Alps and north of Dalmatia.[828]

Linguistics and philology, thus, lead nowhere. There is, however, a passage in Jordanes’ Romana which points to the region where the Nedao must be sought.

At the end of the Romana, Jordanes hastily added the latest news, the victory of the Langobards over the Gepids in 552.[829] More than sixty thousand fell on both sides. Such a battle, he added, has not been heard of “in those places in our times since the days of Attila, except that which had taken place before this battle under the magister militum Calluc, likewise with the Gepids, or the combat of Mundo with the Goths” (in nostris temporibus a diebus Attilae in Ulis locis, praeter ilia quae ante hanc contin- gerat sub Calluce mag. mil. idem cum Gepidis aut certe Mundonis cum Gothis).

In 536, Justinian’s general Mundo fought the Ostrogoths in “Dalmatia,”[830] which was “not far distant from the borders of Pannonia.”[831] Three years later, in 539, the Gepids invaded the Roman provinces south of the Danube, in particular Dacia ripensis.[832] They set out from their newly won territory in Pannonia secunda. Calluc drove them back, but the Gepids rallied and the Romans suffered a crushing defeat.[833] The war in 551 began with an attack of the Langobards on the Gepids, with Sirmium as their aim.

The two campaigns were fought in “Dalmatia” and between the Morava and the Sava. Jordanes compares the great battles “in these regions” with the one fought there in the days of Attila. He must have thought of the battle on the Nedao; there was no other great battle in or near Pannonia in the middle of the fifth century. The Nedao was a river in southern Pannonia, probably a tributary of the Sava.

For the events after 454, Jordanes is again our main, though fortunately not our only source. There exists rich material in contemporary documents: poems, papal letters, ecclesiastical histories, the correspondence between Emperor Leo and the bishops of the Eastern empire, and of course, though indirectly, the Priscus fragments. Together with the Getica they permit a fairly accurate reconstruction of the last period of Hunnic history.

Jordanes writes: “When Ellac was slain, his remaining brothers were put to flight near the shore of the Sea of Pontus where we have said the Goths first [prius] settled.”[834] The Gepids occupied Dacia, the territory of the Huns; the other nations, formerly subjects of Attila, received from Emperor Marcian the abodes allotted to them to dwell in. Then Jordanes turns to the Goths: “Now when the Goths saw the Gepids defending for themselves the territory of the Huns, and the people of the Huns dwelling again in their ancient abodes [suis antiquis sedibus], they preferred to ask for lands from the Roman Empire, rather than invade the lands of others with dangers to themselves. So they received Pannonia.”

Macartney,[835] followed by Thompson,[836] thought that “the ancient abodes” were those of the Huns and located them in the Danube-Theiss Basin for two reasons. First, he said, Dacia was Transylvania; second, the Goths were in Pannonia. This leaves a geographical gap, for Dacia never extended west of the Theiss. The Huns had lived there before, and now, after a short flight to the east, they came back.

However, Jordanes’ Dacia of the Gepids was not identical with Roman Dacia. In Getica 33 he stated expressly that the Theiss flowed through, discurrit, the land of the Gepids. It is true that a good deal of what Jordanes reports about the happenings in the following years makes no sense if the Huns should have lived in the Pontic region. In this respect, Macartney is right, but it does not entitle him to distort Jordanes’ text. What Jordanes maintains is impossible, but what he means is clear; the Huns fled to the Pontus littoral, the ancient seat of the Goths. As we shall see presently, a large number of the Huns stayed for a long time in the northwestern Balkans.

Alfoldi[837] and Schmidt[838] placed the Goths in South Russia before they “received” Pannonia. But how did they, not only the warriors, but the whole people, with their wagons, flocks, and herds, migrate from there to Hungary? Did they ask the Gepids for a transit visum through Transylvania? They were, Alfbldi thought, ferreted out, aufgescheucht, by the Huns. But the Huns moved where the Goths first had settled, to the ancient seats of the people and not the ones held at the time of the breakdown of Attila’s kingdom. These misunderstandings stem from the disregard for the other previously mentioned sources. Before we turn to them we have to listen again to Jordanes:

The Sauromatae, whom we call Sarmatians, and the Cemandri and certain of the Huns dwelt in Castra Martis, a city given to them in the region of Illyricum.

By Sarmatae Jordanes obviously means those Sarmatian tribes which, like the Jazyges, had been in Hungary before the Huns..Jordanes continues:

Of this race [ex quo genere] was Blivila, duke of Pentapolis, and his brother, Froila, and also Bessa, a patrician in our time.

Blivila and Froila are Germanic names,[839] Bessa was “a Goth by birth, one of those who had dwelt in Thrace from the old and had not followed Theoderic when he led the Gothic nation thence into Italy.”[840]

The Sciri, moreover, and the Sadagarii and certain of the Alani with their leader, Candac by name, received Scythia minor and Moesia inferior.

On the Sadagarii, see p. 441. Braun emended certi Alanorum[841] into *ceteri Alanorum, which would make the Sadagarii Alans, but certi Alanorum has a close parallel in the preceding quidam ex Hunnis:

The Rugi, however, and some other races asked that they might inhabit Bizye and Arcadiopolis.

Bizye is the present Vize, Arcadiopolis the present Luleburgaz, 50 miles northwest of Constantinople.

Hernac, the younger son of Attila, with his followers, chose a home in the most distant part of Scythia minor. Emnetzur and Vltzindur, kinsmen of his, seized [potiti sun/] Oescus and Vtus and Almus in Dacia on the bank of the Danube, and many of the Huns, then swarming everywhere, betook themselves into Romania; descendants of them are to this day called Sacromontisi and Fossatisii.

In the Getica, potiri means “to seize by force” (cf. Getica 108, 138, 145, 250, 264, 288). Vtus, at the mouth of the river Vtus (Vit),[842] Oescus, near the present Gigen, at the mouth of the Isker,[843] and Almus, the present Lorn,[844] were in Dacia ripensis.

In the fifth and sixth centuries, fossatum[845] meant “military camp.” The fossatisii in the East correspond to the castriciani and Castellani in the West.[846] Procopius lists four fossata: one in Moesia, the one of Longinus in the country of the Tzanni, the fossatum of Germanus in Armenia, and GesHa-fossatum in Haeminontus.[847] Gesila[848] is Gothic *Gaisila;[849] the camp was obviously garrisoned by Goths. Fossatisii, Latin with a Greek ending, points to Moesia, where the two languages met.[850] The Hunnic Fossatisii were probably those of the camp in Moesia. The Sacromontisi may have received their name from the “holy mountain” in Thrace.[851]

Although the greater part of the Huns preserved their tribal organizations, many were leaderless, broken men who had no choice but to surrender to the Romans. Jordanes contrasts Hernac, Emnetzur, and Ultzindur and their follow’ers to the Huns who, “swarming everywhere, rushed into Romania.” He also makes a distinction between the Hunnic leaders and the Alanic and Germanic fugitives—the Huns seized land, the others received it. Incidentally, the list of the Germanic refugees, Sciri, Rugi, and the Goths at Castra Martis, is rather instructive; had they fought against the Huns, they would not have fled across the Danube. Jordanes was so vague about the origin of Blivila, Froila, and Bessa because had he openly said they were Goths, he would have admitted that Goths did fight under Attila’s sons.

After 455, there existed two Hunnic pockets within the empire: one under Hernac, in the Dobrogea, and the other one in Dacia ripensis. Of the former one we hear nothing before the second half of the 460’s. The western Huns, however, having overcome the first shock, soon became active again. This we learn from Jordanes and the Nordic Hervararsaga.

The First Gotho-Hunnic War

The lines 119–122 of the Old English poem Widsith[852] allude briefly to a war between Huns and Goths:

Wulfhere I sought and Wyrmhere: there full oft war was not slack, what time the Hraede with sharp swords must defend their ancient seat from the people of /Etla by the Wistlawood.[853]

The Hraede are the Ostrogoths, 2Etla is Attila, Wistlawood is Vistula Wood or the wood of the Vistula people.

Another, much later version of the same tradition is preserved in the “Lay of Angantyr,” the oldest part of the Icelandic Hervararsaga. Heusler and Ranisch gave the “Lay” the first place in their edition of Eddica minora.[854] Some of its stanzas have such an archaic ring that Heusler and Genzmer dated the original (from which the “Lay” derives and of which it still contains so much) in the middle of the first millennium. Once it was imbedded in the Hervararsaga[855] the Icelandic redactors tried to fit the “Lay” into the framework of the saga; a number of verses were dissolved into prose. But even in its diluted form the “Lay”[856] stands closer to the heroic epic of the Migration Period than any other Germanic poem.

Heithrek, king of the Goths, had two sons, Angantyr and, from a Hunnic wife, Hloth. Hloth was brought up by his maternal grandfather, Humli, king of the Huns. After Heitrek’s death, Hloth claimed an equal share of the inheritance:

The half will I have / of what Heithrek owned of awl and of edge, / of all the treasure, of cow and of calf, / of quern harsh-grinding, of thrall and of bond-maid, / and those born of them, the mighty forest I which is Myrkwith height, the hallowed grave / which in Gotland stands, the shining stone / which a stodum Danpar stands, half of the war-weeds I which Heithrek owned, of land and lieges / and of lustrous arm-rings.

Angantyr was willing to compromise, but his counsel Gizur, leader of the Grythings, objected that too much was offered to a bond-woman’s son. Enraged, Hloth returned to the Huns. When spring came, King Humli and Hloth drew together so great a host that there was dearth of fighting men in Hunland. They rode through Myrkwith. As they came out of the forest, they saw a castle. There ruled Hervor, Angantyr’s and Hloth’s sister, and with her, Ormar, her foster father. In the ensuing battle Hervor was killed. Ormar escaped and made report to Angantyr; “From the South am I come I to say these tidings: / burned is the far-famed/ forest Myrkwith I all Gotland drenched/ with the gore of the fallen.” Angantyr sent Gizur as herald to challenge the Huns to battle. The place should be

at dylgiu and in the Dun-heath

and all the lassar mountains,

where the Goths so often had won victory. The battle lasted eight days. At last the Huns were forced to give way; Angantyr slew both Hloth and Humli. The Huns took to flight, and the Goths slew so many that the rivers were dammed up and overflowed their banks, and the valleys were filled with dead men and horses.

The Widsith and the “Lay of Angantyr” refer to the same struggle: (1) the Goths fight the Huns: (2) they defend their ancient seat, “the hallowed grave which in Gotland stands”; (3) there they had often won victories; (4) the Goths defeat the Huns; (5) Wyrmhere is Ormar.

One should think that these data, in combination with the personal and place names, would make it comparatively easy to determine when and where the battle was fought, provided, of course, that the “Lay” is not pure fiction. The vast literature shows the opposite. For awhile, Heinzel’s interpretation was widely accepted—the kernel of the story was supposed to be the victory of the allied Visigoths and Romans over the Huns in Gaul in 451.[857] After this view was abandoned, the battlefield was located near the Waldai heights in Russia, in southern Silesia, somewhere in the Ukraine, and near the Marchfeld in Lower Austria. The dates suggested ranged from the first to the middle of the fifth century. There understandingly arose the question whether the problem could be solved at all. Was it not an equation with too many unknowns? The majority of the Germanic scholars seem now inclined to regard the Battle of the Huns nicht als die dichierische Formung eines geschichtlichen Ereignisses, sondern einer geschichllichen Zustandigkeit[858] whatever that means. But it was not so much the intricacy of the problem and the ambiguity of the poetic language that seemed to defy all attempts to date and place the battle as the wild guesses of the philologists. If the Dun of the Dun-heath is identified with the Don, the lassar Mountains with Jasaniky (the stretch of hilly country which forms the broad gap between the Sudeten and the Carpathians), Gizur with the Vandal Geiseric, and Heithrek with the Gepid Ardaric, the Battle of the Huns becomes a geographic and historical monstrosity.[859]

Johannson[860] proved that the names Harwada[861] and Grafa in the Her- vararsaga have nothing to do with the original “Lay.” Arnheimar, “river home,” is not a real place name; anyway, it cannot be placed. Dylgia, v. 11, Dilgia and Dyngia, means “struggle, enmity.”[862] Myrkwidr, “murky wood,” can be as little found on a map[863] as der schivarze ivald in Stefan George’s “Waffengefahrten”: er zog mich heut aus manchen fesseln. Im schwarzen wald wo unheil haust war ich verstrickt in tiefen nesseln. There remain Danpar, Dun, and lassar. In spite of Heinzel’s doubts,[864] partly repeated by Schramm,[865] Danpar is certainly the Dnieper. Duna is the Danube.[866] lassar will be discussed later. Except for Grytingalidi, “leader of the Grytings=Greutingu,” the personal names are obscure.

It seems, the only reasonable approach to the problem of the Battle of the Huns is to look for a historical event that fits the geographical setting of the “Lay.” When and where did the Ostrogoths win a decisive victory over the Huns? Not in South Russia. There in the fourth century they were attacked, defeated, and, except those who succeeded in fleeing, remained loyal to their Hunnic lords to the very end. Nevertheless, Baesecke[867] and Altheim[868] indulge in wild speculations about the lassar Mountains, which they connect with the Ossetes in the Caucasus. Baesecke, again followed by Altheim, brings together Dylgia and Kossa dolgjana, near Mariupol in the Ukraine. Kosa dolgaya—this is the correct form—is good Russian and means “a long narrow tongue of land.”[869] The sandy Kosa dolgaya on the southeastern shore of the Azov Sea[870] is nowhere wider than 500 meters, as fit for a battle between horsemen as the top of the Matterhorn. Schramm’s assumption that sometime before 375 the Goths clashed with nomads whom later tradition turned into Huns is sheer arbitrariness.[871] Malone, disregarding all other place names, makes the Vistula in the Widsith the basis of a peculiar hypothesis[872]—after overrunning the southern part of Ermanaric’s Ostrogothic kingdom, the Huns are supposed to have tried to conquer the Ostrogoths in the Vistula Valley; the often renewed struggle has a happy ending. Needless to say, the Vistula woods were impenetrable to the Hunnic horsemen.

Jungandreas thinks the poet localized the battle on the Vistula because in Old English poetry the seats of the Goths were traditionally in the northeast.[873] Schramm assumes that the Vistula took the place of the Dnieper because no one in England in the eighth century had ever heard of the river in South Russia.[874] Linderski gives what I think to be the best explanation of the alleged mistake:[875] King Alfred’s Wislelond is taken from ancient maps; both in the Divisio orbis terrarum and the Dimen- suratio provinciarum, “Dacia finitur ab occidente flumine Vistula.” Wislelond lies east of Moravia and west of Dacia. The Goths did not fight the Huns in Silesia but somewhere in the Carpathian Basin. It must be left to the philologists to decide how Vistula got into the Widsith or what it meant.

The Battle of the Huns reflects the wars which the Goths waged against the Huns after the collapse of Attila’s kingdom.[876] We are well informed about them by Jordanes:

Let us now return to the tribe with which we started, namely the Ostrogoths, who were dwelling in Pannonia under their king, Valamir, and his brothers Thiudimer and Vidimer. Although their territories were separate, yet their plans were one [consilia tamen unita]. For Valamir dwelt between the rivers Scarniunga and Aqua nigra, Thiudimer near Lake Pelso, and Vidimer between them both. Now it happened that the sons of Attila, regarding the Goths as deserters from their rule, came against them as though they were seeking fugitive slaves [velut fugacia mancipia requirentes] and attacked Valamir alone, when his brothers knew nothing of it. He sustained their attack, though he had but few with him, and after harassing them a long time, so utterly overwhelmed them that scarcely a portion of the enemy remained. The remnant turned in flight and sought the parts of Scythia which border on the stream of the river Danaber, which the Huns call in their own tongue the Var. Whereupon he sent a messenger of good tidings to his brother Thiudimer, and on the very day the messenger arrived he found even greater joy in the house of Thiudimer. For on that day Theoderic was born.[877]

The passage poses a number of difficult problems. Where, for instance, are the two rivers between which Valamir dwelt? Neither Scarniunga nor Aqua nigra is mentioned elsewhere. Alfbldi identified Aqua nigra with Karasica, a tributary of the Drava, assuming that Karasica goes back to Karasu, in Turkish, “black water.”[878] This has been rejected by Moor.[879] There are so many Black Waters between Vienna and Belgrad that if the name of the river were the only thing to go by in localizing Vala- mir’s territory, it might have been anywhere. It is true that Aqua nigra cannot be the Raab, as has been so long assumed and is now assumed again, but in northwestern Hungary alone there are the Schwarza, the Schwarzbach, the Schirnitzbach, and the Csornopatak, a now obsolete, originally Slavic name of the upper Herpenyb, a tributary of the Raab.[880] Other Black Waters can be found all over Pannonia secunda. Lake Pelso is Lake Balaton (in German, Plattensee). If Vidimer lived between Thiu- dimer at Lake Balaton, followed, evidently in the south, by Valamir, then Valamir must have lived near the Drava. Alfoldi, though wrong in detail, was basically right.[881]

The Goths lived in Pannonia, but they did not occupy the two Pannonias from border to border. This follows not only from the previously quoted account—the Huns attacked Valamir “when his brothers knew nothing of it,” which precludes a compact Gothic settlement—but also from the account of the second Hunnic attack: a part of inner Pannonia was held by the Sadagis. Two passages in the Vita s. Severini show that easternmost Noricum Mediterranense was Gothic.[882] The small Ostrogothic fibulae found in Slovakia north of the Danube[883] point to Ostrogoths, who did not follow the Amali princes.

When and under what circumstances the Goths settled in Pannonia either is not known. That they did not move there after Attila’s death is by now almost generally agreed. Although Jordanes may have maintained that the Goths “received” Pannonia from Marcian merely in order to stress the bonds between them and the Eastern Romans, he probably was right. Avitus did try to make an agreement with them, but it evidently came to nothing. Whereas his finances were in such a bad state that he was forced to melt down the bronze statues of Rome and sell the metal in order to pay the soldiers, the rich East could afford to pay the Goths subsidies, not as much as to the Gepids, but enough to keep them quiet for a few years.

When did the Huns attack the Goths? Ensslin, who first dated the war in the winter of 456/7,[884] later took this overexact date back.[885] It would seem that the date of Theoderic’s birth determines also the date of the Gothic victory. Unfortunately, the historians cannot agree on the year. It might be 454, 455, and even 456.[886] Besides, the connection between the Gothic victory and Theoderic’s birth could very well be due to the wish to let the heroic life of the great king start with a propitious event, auguring his later greatness. In any case, the battle on the Nedao cannot have been fought many years before the Huns attacked the Goths. The war is probably to be dated about 455.

Where did the Huns come from and where did they flee? The objections to the alleged trek of the Goths from the Pontus to Pannonia are equally valid for the march of a Hun army from the Black Sea to Hungary. How came the Huns to ride clean through the intervening nations, asks Macartney rightly.[887] He spoils his case by tampering with the text in the Getica. The Huns, he maintains, not only came from their center between the Danube and the Theiss, they also returned there after their defeat by the Goths, to the Danube, not the Dnieper.

The text in Jordanes translated previously runs as follows: pars ostium... in fuga versa eas partes Scythiae peteret, quas Danabri amnis fluente prae- termeant, quam lingua sua Hunni Var appellant.

The variae lectiones of the name of the river are bewildering. For danabri in H, PVO have danubri, and XYZ danapri; danubii occurs only in the codex Ambrosianus, which teems with misspellings; danubri is obviously a cross between danabri and danubii. The scribes were not sure what to write. This is not so surprising. The names of the two rivers, Danubius and Danaper, sounded so similar that they easily could be confused, and actually often were. Jordanes himself wrote in Getica 54 Hister= Danubius where he should have written Danaper. Tanais and Danubius similarly were mixed up. In his account of Decius’ campaign against the Goths in I, 23, Zosimus wrote three times Tanais instead of Danube= Hister. The seven mouths of the Danube are many times named in Greek and Latin literature, but Horace, Troades 8–9, has the seven mouths of the Tanais. In his commentary on Horace, Pseudo-Acro (fifth century?) stated explicitly that “a river of Scythia is called Tanais, which is the same as Danube” (Tanais flumen Scythiae dicitur, qui et Danubius est).

Still, Jordanes must have written Danabri, not Danubii. The river is in Scythia, not in Pannonia. Var, is, indeed, the Dnieper. The whole relative clause from quam to appellant cannot be a later addition either.

When the Huns attacked the Goths in Pannonia a second time, they again could not and did not ride all the way from the Danaber-Dnieper in the southern Ukraine to Hungary. There is only one explanation of the passage on the flight of the Huns after the first attack: Jordanes pro- leptically located the Huns of the later 450’s where they were at his time. He displays the same cavalier treatment of geography in the account of the second Gothic war with the Huns. But before we deal with it, we have to look closer at the Huns in the northeastern Balkans. It was from there that they rode against Valamir.

Late in the year 457,[888] Emperor Leo sent a sacra to all metropolitan and many other bishops asking for their opinions on the validity of the consecration of Timothy Aelurus as bishop of Alexandria and on the point of upholding the Council of Chalcedon.[889] The list of the provinces to which the letter was sent as well as the answers to it[890] permit some conclusions about the situation in the Balkan peninsula in the first year of Leo’s reign.

All provinces of the Thracian diocese were again firmly under Roman rule. Whereas none of the bishops of Moesia inferior had been present either at the Robber Synod of 449 or at the Council at Chalcedon in 451, now not only Marcianopolis, Nicopolis, and Odyssus, but also Novae, Abrittus, Appiaria, and Durostorum on the Danube[891] could freely communicate with Constantinople. Even the bishop of Tomis in the “Scythian region” received and answered the circular letter.[892]

It was different in eastern Illyricum. The bishops of Dyrrhachium, Scampa, Lychnidus, Bullis, Apollonia, and Aulona in Epirus nova assured the emperor of their unshakable orthodoxy.[893] A similar letter was sent from Dardania.[894] The answer of Zosimus, metropolitan of Dacia medi- terranea, has not been preserved, but the fact that Leo wrote to him[895] show’s that Serdica, which eight years ago was in ruins, had to some extent regained its former importance.[896] But the emperor did not send letters to the metropolitan bishops in Dacia ripensis, Moesia superior, and Prae- valitana.[897] Evidently there were no bishops in those provinces to whom he could write.

One could presume that communications with Ratiaria and Vimina- cium, or such an important place as Naissus, were interrupted because those bishoprics were in a war zone. But 457 had been a year of peace for the Balkan provinces. The “war-loving nation,”[898] “rebels,”[899] elusive bands “now so utterly crushed that not even their name could be found anymore,”[900] were probably Lazic marauders.[901] The letters from the bishops in Moesia inferior, Dardania, Epirus nova, and the “Scythian region” mention no military operations in their regions.[902] If anything bigger than occasional clashes with latrunculi had taken place, the bishops would have been likely to hint at it.

In 449, eight years before, there still existed Christian hostels in Naissus.[903] Shortly afterward the Huns evacuated the strip of territory south of the Danube they had occupied in 447.[904] One should, therefore, expect that since then the Roman population had come back, and with them the clergy. But if they did, they had fled again. They fled from the Huns who, after the collapse of their kingdom, seized not only the three places in Dacia ripensis which Jordanes names, but, to repeat his words, “swarming everywhere betook themselves into Romania.”[905] There is no other explanation of the breakdown of all ecclesiastical life in the northwestern Balkans.[906] If small Christian communities still existed, which is unlikely though not impossible, they were cut off from the churches farther east. They were not in Romania but in Hunnia.

This does not necessarily mean that the whole, rather large territory was held by the Huns only. Between the Timok and the Arter lived Sarmatians, Cemandri, and some Huns. As late as the end of the 460’s, Goths lived side by side with Huns. But the political power lay nevertheless with the Huns, the same who about 455 had tried to reconquer Pannonia.

By 456 at the latest, the government in Constantinople must have realized that it lacked the power to reconquer the Hunnic territories south of the Danube. It kept the peace, or, rather, the silent truce, and so did the Huns. In 457, Hunnia was still inaccessible to the Romans. The change came in 458.


The “very considerable” army which, early in 458,[907] the Western emperor Majorian (451–461) collected in preparation for the campaign against the Vandals consisted almost entirely of barbarians. In the panegyric which Sidonius Apollinaris addressed to the emperor in Lyon at the end of the year,[908] he named the tribes which followed the imperial standards:

Thou dost carry off to the war the frozen army of the seven-mouthed Danube. All the multitude that the sluggish quarter of the north doth produce in the Sithonian region beneath the Parrhasian bear ... Bas- tarna, Suebus, Pannonius, Neurus, Chunus, Geta, Dacus, Halanus, Bellonotus, Rugus, Burgundio, Vesus, Alites, Bisalta, Ostrogoths, Procrustes, Sarmata, Moschus ... the whole Caucasus and the Tanaitic drinker of the Scythian water.[909]

This is one more of those lists of names in which Sidonius liked to indulge. Most names he borrowed from earlier poets,[910] others were obsolete,[911] adduced to impress the listener with his erudition. Among the names retained is Chunus, as the following verses show:

Now thou wert moving thy camp and around thee thronged thousands under diverse standards. Only one race denied thee obedience, a race who had lately, in a mood even more savage than their wont, withdrawn their untamed host from the Danube because they had lost their lords in warfare, and Tuldila stirred in that unruly multitude a mad lust for fight which they must needs pay dear.[912]

The verses refer to the Huns.[913] They had lost Ellac and other domini; the battle on the Nedao was fought only a few years before, nuper; they had withdrawn from the Danube, and their sites had been occupied by their former Germanic subjects.

Although Sidonius did not say where Tuldila’s Huns lived, it is clear that Majorian could not have recruited them in the Pontic littoral; Tuldila could not have come from the Dnieper. His Huns must have lived close to the borders of the Western empire. Priscus says, indeed, that Majorian “brought the peoples near his domains to his side, some by arms, some by words.”[914] There were no Huns at the borders of Noricum; Pannonia was held by the Goths, the greater part of the regions east of the Danube by the Gepids. We are, thus, led to the same areas from which the Huns had moved against the Ostrogoths, that is, Moesia superior and Dacia ripensis. Whether Majorian had won the Huns “by words” or “by arms” we do not know.

Two pieces of information, so far not used by the students of the Huns, throw more light on the situation in 458. The year before, Moesia superior, Dacia ripensis, and Dacia mediterranea had been inaccessible to the messengers sent from Constantinople to the bishops in the Balkan provinces. But in the summer of 458, the body of Sancta Anastasia was transferred from Sirmium to Constantinople and buried ev xolq Aopvlvov EpfioAoiQ.[915] The routes from the capital to Pannonia secunda were open again. This presupposes the pacification of the northwestern Balkan provinces, the establishment of a modus vivendi with the barbarians there. The East gained from the peace the body of a martyr, the West auxiliaries and many women and young people who had been given up for lost. In the spring of 458 the first prisoners of war, carried off by the Huns in 452, came back to Aquileia. They were released by the Huns, the same, as we now may say with confidence, who joined Majorian’s army and let the East Romans through their land to Sirmium.

The Second Gotho-Hunnic War (463/4-466)

In the years following the pacification of the northwestern Balkans (before 458), the Huns were at peace with the East Romans but, as we learn from Jordanes, and only from him, they once again, as in around 455, attacked the Goths in Pannonia.

Now after firm peace was established between Goths and Romans, the Goths found that what they received from the emperor was not sufficient for them. Furthermore, they were eager to display their wonted valor, and so began to plunder the neighboring peoples around them, first attacking the Sadagis, who held the interior of Pannonia. When Dintzic, king of the Huns, a son of Attila, learned this, he gathered to him the few who still seemed to have remained under his sway, namely, the Ultzinzures, the Angisciri, the Bittugures, and the Bardores. Coming to Bassiana, a city of Pannonia, he beleaguered it and began to plunder its territory. When the Goths learned this, they abandoned the expedition they had planned against the Sadagis and turned upon the Huns and drove them so ingloriously from their own land that those who remained have been in dread of the arms of the Goths from that time down to the present day.[916]

There follows the description of the war between the Goths and the Sciri.

What Jordanes’ source was is difficult to decide, if it can be decided at all. Mommsen suggested that Jordanes followed Priscus.[917] The endings of the Hunnic tribal names point, indeed, to a Greek author, but why should Jordanes have changed Priscus’ Dengizich into Dintzic? Priscus certainly did not praise “the wonted valor” of the Goths either. This sounds more like Cassiodorus. If Jordanes followed Cassiodorus, the strange sentence at the end of his account becomes understandable. In 551, the year he wrote the Getica, the Goths had been in Italy for more than seventy years and therefore could not be dreaded by the Huns “to the present day.” By the middle of the sixth century there were no Huns even near Totila’s kingdom. For a moment one might think of a passage in the letter which in 476 Apollinaris Sidonius wrote to his friend Lampridius, full of the most extravagant eulogies for Euric, king of the Visigoths in Gaul.[918] To him as the arbiter mundi came ambassadors from everywhere, even from Persia and the Ostrogoths who, with Euric’s help, pressed hard on the Huns.[919]

From 475 on, Theoderic had his headquarters in Novae in Moesia secunda.[920] Sidonius’ letter confirms our thesis about the prolonged stay of Huns in the Balkans, but neither Cassiodorus nor Jordanes could refer to the 470’s as to “the present day.” The phrase makes sense if the Huns were the Bulgars of 505, when Pitzia and his Goths defeated Sabinian’s army, which consisted of ten thousand Bulgarian horsemen. Cassiodorus, writing his Gothic history in the 520’s or early 530’s, and Ennodius (t 521) repeatedly calls the Bulgarians “Huns.” It seems that Jordanes copied Cassiodorus without changing the text itself, a text which, with slight changes, was based on Priscus.

When did the Huns attack? The first two sentences in Jordanes’ report give the answer.[921] In 459, the Goths, led by Valamir, took Dyrrhachium (Durazzo).[922] The Bomans were under Anthemius, the future emperor.[923] In 461, Valamir concluded a foedus with the Romans and received a yearly subsidy of 300 pounds of gold.[924] It is unlikely that the Goths broke the treaty after only one year.

The war between the Huns and the Goths preceded the war between the Goths and the Sciri. Priscus dealt with its beginnings in EL 17. He wrote about the visit of Gobazes, king of the Lazi. In the preceding excerpt he described Gobazes’ visit to Constantinople after the big fire. Emperor Leo, who had fled from the burning city,[925] met him in Chalcedon. The second Gotho-Hunnic war, therefore, falls between 463/4 and 466.

The Huns came from the south. The first fortified place which stood in their way was Bassiana, the respublica coloniae Bassianorum of the local inscriptions, between Sirmium (now Mitrovica) and Singidunum (now Belgrade). The tribal names lead likewise to the south of the Danube as the region from where the Huns marched against the Goths. The Ult- zinzures lived between Utus (now Vit) and Almus (now Lorn), both in Dacia ripensis, and the Bittugures joined the Ostrogoths on their trek from Moesia secunda to Italy in 488. The theater of both Gotho-Hunnic wars was Pannonia secunda. Of course, the Huns did not respect the borders of the former Roman provinces; the fighting certainly spread to the province of Savia, between the rivers Drava and Sava.

And now we come back to the “Lay of Angantyr.” Was southern Pannonia the ancient land of the Goths (as mentioned in Widsilh), where the former kings lay in the hallowed grave? Were there mountains the name of which sounded like Jassar? Cassiodorus, inscriptions, and ancient geographers give the answer.

Theoderic, who in 504 had conquered Sirmium from the Gepids, sent there Colossaeus, vir illustris and comes. In the letter of appointment which Cassiodorus wrote and found so good that he included it in his Variae,[926] we read:

You are sent with the dignity of the illustrious belt to Pannonia se- cunda, the former seat of the Goths [quondam sedem Gothorum]. Protect the province committed to you with arms, so that she can gladly receive her old defenders [antiquos defensores], as she used gladly to obey our fathers [quae se nostris parentibus feliciter paruisse cognovit].

Sirmium, says Ennodius, was in ancient times the border of Italy where seniores domini kept guard against the barbarians.[927]

The name las is well attested in the ancient land of the Goths.[928] North of the mountainous country between the Sava and the Drava, which might be the Myrkwidr of the “Lay,” lived the Iasi of Pliny. Iasi served in the Roman army. Aquae lasae, the present Varazdinske Toplice, was a flourishing city as late as the fourth century. Constantine ordered a bath destroyed by fire to be rebuilt there. At the time the Goths moved into Pannonia the name las must still have been quite alive. It cannot be separated from the lassar Mountains.

In the Germanic tradition the two wars were merged. That so much of the actual events and place names like the Danube heath and the lassar Mountains has been preserved in it is truly remarkable.[929]

The End

In 465 or, more probably, 466, Dengizich and his brother Hernach sent ambassadors to Constantinople. They wanted to make peace, provided that a market place be established at the Danube where “according to the ancient custom” Romans and Huns could exchange “what they needed.” The emperor rejected their demands.[930]

After his last attempt to reconquer at least some land in Pannonia had failed, Dengizich had crossed into Wallachia. The Bittugures and apparently other tribes had left him and stayed south of the Danube. Under what circumstances Hernach gave up the Dobrogea is not known. In any case, the remnants of Attila’s Huns were no longer anywhere in the border provinces; otherwise their demand for a market place on the Danube would not have made sense.[931]

When Hernach, engaged in disputes in his own country presumably with the Saraguri,[932] refused to join his brother, Dengizich moved his own hordes closer to the Danube, threatening to break into Thrace unless the emperor granted him and his people land and subsidies. Scorning the offers of Anagastes, “to whom the defense of the river was entrusted,” to negotiate with him, Dengizich sent his envoys directly to the emperor. Leo “answered that he would readily do all these things if they would be obedient to him, for he rejoiced in men who came into alliance with him from his enemies.”[933] At this point Prisons’ text breaks off.

Gordon thinks that Constantinople’s willingness to come to terms, contradicting her earlier attitude, may be accounted for by the necessity of protecting her northern frontiers in preparation for the approaching expedition against the Vandals in Africa.[934] He may be right. That the negotiations with the Huns eventually broke down had, I believe, another reason. The situation was, on a minor scale, a repetition of 376, though with the essential difference that the Huns, unlike the Goths, needed wide pastures for their flocks and herds, not land for the plough. To accommodate them in the empire would have necessitated the expulsion of the peasants from a large territory, including many of those Goths in the Thracian dioceses on whose support Aspar, for many years the nearly all-mighty major domus, brother-in-law of the Gothic leader Theoderic the Squinter, depended.

Dengizich crossed the frozen Danube. He evidently expected that the Huns still south of the river would join him. Some probably did. But large groups of barbarians acted on their own, using the chance to back their demands with arms:

Anagastes, Basiliscus, Ostryis, and other generals penned up and blockaded the Goths in a hollow place. The Scythians, hard pressed by starvation and lack of necessities, sent an embassy to the Romans. They said they were ready to surrender, if only they were given land. The Romans answered that they would forward their requests to the emperor. But the barbarians said that they must come to an agreement right away; they were starving and could no longer wait. The Roman generals took counsel and promised to supply food until the decision of the emperor came, provided the Scythians would split themselves into just as many groups as the Roman army was divided into. In this way the Roman generals could better care for them. The Scythians accepted the terms brought by their ambassadors and drew their forces up in as many sections as the Roman army. Chelchal, a man of Hunnic race, the lieutenant general of those in charge of Aspar’s forces, came to the barbarian horde allotted to them. He summoned the prominent Goths [logades], who were more numerous than the others, and began a speech to the following effect: The emperor would give land, not for their own enjoyment but to the Huns among them. For these men did not care for tilling the soil and, like wolves, attacked and plundered the provisions of the Goths. They themselves, the Goths, were treated like slaves and forced to feed the Huns, although there never had been concluded a treaty between the two peoples, and the Goths had been pledged by their ancestors to escape from an alliance with the Huns. Thus, the Goths thought lightly of their ancestors’ oaths and the loss of their own property. He, Chelchal, was a Hun and proud of it, but he was saying these things to the Goths from a desire of justice, so that they should know what must be done.

The Goths were greatly disturbed by this and, thinking that Chelchal had said these things with good will toward them, attacked the Huns in their midst and killed them. Then, as if at a signal, a mighty battle rose between the races. When Aspar[935] learned of this, he and the commanders of the other camps drew up their troops and killed the barbarians they came upon. When the Scythians preceived the intent of the trick and the treachery, they gathered together and turned against the Romans. Aspar’s men anticipated them and killed the barbarian horde allotted to them to the last man. But the fight was not without[936] danger for the other generals, as the barbarians fought courageously. Those who survived broke through the Roman formations and escaped the blockade.[937]

The date is 467. Basiliscus, brother of Empress Verina, was still magister militum per Thraciam?[938] in the spring of 468 he was commander in chief of the African expedition. It throws an interesting light on the Byzantine armies of the late fifth century that Basiliscus was the only Greek among the commanders. Anagastes, *Anagasts, the son of Arni- gisclus, and Ostrys were, as their names indicate, Goths.[939] Chelchal kept his Hunnish name; evidently he was not yet baptized. As his high rank shows, he had served a long time in the Roman army. Chelchal may have been one of those Huns who deserted to the Romans in Attila’s time or joined them after 455.

The fact that Aspar sent a large contingent of his buccellarii against the barbarians shows their strength. Although the Romans had some successes,[940] the war dragged on for two more years. In 468, the greater part of the army was sent to Africa.

The end came only in 469. Marcellinus Comes has the short entry “The head of Dinzic, son of Attila, king of the Huns, was brought to Constantinople.”[941] The Chronicon Paschale gives more details: “Dinzirichus, Attila’s son, was killed by Anagastes, general in Thrace. His head was brought to Constantinople, carried in procession through the Middle Street, and fixed on a pole at the Wooden Circus. The whole city turned out to look at it.”[942]

The few Huns south of the Danube who did follow the Ostrogoths, like the Bittugur, gradually lost their ethnic identity or joined the Bulgarian raiders.

III. Economy

The written sources contain little about the economy of the Huns before they made contact with the Roman world. It certainly changed in the eight or nine decades we can follow their history, though the change has been exaggerated. By the middle of the fifth century the great majority of the people lived almost the same nomadic life—with animal husbandry as the economic mainstay and hunting and fishing as subsidiary occupations—as their ancestors had lived.

Whether we realize it or not, when we speak of nomads, often Father Abraham, archetype of the Beduin sheikhs, comes to mind, pitching his tent one week here and the other there, constantly on the move from pasture to pasture. This is the way Ammianus described the Alans: “When they come to a place rich in grass they feed like wild beasts. As soon as the fodder is used up,” they move to another place. It is the Chinese stereotype: “they follow water and grass.” The mobility of the herders always has struck farmers—Greeks, Indians, and Chinese—as incomprehensible, uncanny, and inhuman. The archaeological evidence refutes Ammianus.

In the steppe and the wooded grassland from western Kazakhstan to the Carpathian Mountains many hundred kurgans have been excavated and thousands of graves of all Sarmatian periods opened. So far no traces of settlements have been found. One could think that the buildings, if the Sarmatians had any, were of perishable material, but then at least fireplaces and garbage pits should have been preserved. They were not. And yet two facts are incompatible with the idea of the restlessly wandering Sarmatians. First, the large grave fields. Sinitsyn was impressed by the many kurgans in the Sarmatian cemeteries on the Kolyshlei River; in some the burial mounds numbered fifty and more.[943] But they were in the wooded steppe, close to the forests, where the normal mobility of the sheep, horse, and cattle breeders was possibly restricted by natural obstacles. However, large kurgan grave fields also are known from the treeless steppe. In Berezhnovka II, on the left bank of the Volga, about two hundred kurgans were counted, and how many have been plowed over can no longer be determined. Sarmatians buried their dead there from the sixth century b.c. to the third or fourth century a.d. Of the burials excavated, 38 were Sauromatian, 29 Early, 18 Middle, and 17 Late Sarmatian.[944] In the two kurgan groups at Bykovo, in the same region, 20 burials were of the Sauromatian and 60 of the Early and Middle Sarmatian periods.[945] At Kalinovka, Shilov excavated 62 kurgans with 253 burials of which 5 were Sauromatian, 64 Early, 60 Middle, and 31 Late Sarmatian.[946] Those were not princes’ graves, not sacred burial grounds as in the High Altai. Many graves contained very few goods or none at all. It was the same in the West. In the lower valley of the Molochnaya River there is one Sarmatian kurgan after another. Of the 369 burials excavated in 1950 and 1951, 54 were Sarmatian.[947] This points, as Vyazmitina rightly stressed,[948] to a semisedentary life.

Then there is the Sarmatian pottery. True nomads like the Beduins or the Mongols have leather and wooden not clay vessels. From the earliest to the latest period, the Sarmatians used clay pots, bowls, and dishes. This proves, as Arzyutov emphasized[949] (though other archaeologists did not see it), that the Sarmatians were shepherds. Even if all wheel-made clay vessels found in the graves were imported—a rather unlikely assumption— there are the many handmade flat-bottomed pots. People who frequently move from one place to another have, as a rule, round-bottomed vessels which can be put in the soft ground or carried on cords or in a net. The Sarmatians had vessels of both types, obviously used for different purposes. Still, that they did make vessels fit for a longer stay speaks, like the large cemeteries, for prolonged stays in one place.

The wanderings of the ancient, medieval, and modern nomads of central and eastern Asia may at times and depending on geographical factors have been very long,[950] but as a rule they always repeated themselves: from the same winter quarters to the same summer pastures, and back. There was a certain latitude in the choice of the summer pastures, but the winter quarters remained the same. “True” nomadism of the Beduin type was a rare exception; in central Asia only the Kazakhs and Turkmens of the Aral-Caspian steppes and half-deserts are or until recently were constantly moving from pasture to pasture.[951]

The seminomadism of a Hunnic tribe is attested by Jordanes: In the summer the Altziagiri put up their camps in the steppe near Cherson in the Crimea where their cattle found food pasturage, and in the winter they moved above the Pontic Sea,[952] presumably to Sivash, the lacus putidus, where the luscious reed provided good fodder for the animals.[953] Jordanes’ statement, valuable as it is, must not be taken literally. There were and are no nomads who live exclusively on horned cattle. Compared with horses and sheep, cattle always and anywhere played a secondary role.

According to Ammianus, the Huns had all kinds of domesticated animals.[954] Whereas we are comparatively well informed about their horses, we hear very little about their cattle. In the version of the sacred-sword legend which Jordanes took from Priscus, we read of a herdsman and the heifer which stepped on the sword,[955] and Priscus mentions an ox which Attila sent to the Roman ambassador.[956] In the economy of the Eurasian nomads, goats take a small place. The skins of the haedus with which the Huns “protected their hairy legs”[957] were perhaps the skins of the ibex, a motif occurring quite frequently in the art of the Scythians and their relatives.[958]

No Greek or Roman author mentions sheep, without which the Huns could not have lived. The meat they boiled in the big cauldrons was mutton. Sheep provided milk and cheese. The tents were made of sheepskin or felt,[959] which was made out of sheep’s wool. Like the shoes of the Sarmatians, those of the Huns were made of sheep’s leather.[960] The curved caps of the Huns[961] were doubtless made of felt. Jerome corrected Ammianus’ slightly vulgar term galerus.[962] He called the Huns’ dress a tiara, which he describes as “a round cap, as we see it depicted in Odysseus, as if a ball were divided in the middle and one of the parts placed on the head. This the Greek and our people call ridgav, some call it galerus” (Rotundum pilleolum quale pictum in Ulixe conspicimus, quasi sphaera media sit divisa, et pars altera ponatur in capite. Hoc Graeci et nostri xiagav, nonnulli galerum vacant). But Ammianus’ galerus incurvus was almost certainly not a round but a curved cap, pointed like the Phrygian cap, a type known from the Black Sea to the borders of China.

The Huns, maintains Thompson, could not weave because they had no time for it. How strange 1 The Sarmatians seem to have had plenty of leisure, for in their graves many hundred spin whorls have been found, made of stone, alabaster, and cut from the bottom of clay vessels. Burials on the Torgun and the right Iio via yielded twilled wool fabrics. Like the Sarmatians, the Huns spun the wool of their sheep. They also made linen. Ammianus speaks of their linen dress; the canopies under which rows of girls met Attila when he entered his residence were of white linen, and in Queen Ereka’s house linen cloth was embroidered. Did the Huns import linen ? This is unlikely, for the Goths in south Russia also wore linen clothes, and pieces of linen were found in Late Sarmatian graves in the lower Volga region.


In the economy of the Huns in the Hungarian plain, camels were of little or no importance. Had Priscus seen any he could hardly have failed to mention them. On their retreat from Persia in 395, the Huns may have driven a few camels with them.[963] But on the Danube the beast could not have been more than an exotic curiosity. Farther to the east, however, in Rumania and particularly the Ukraine, the Huns, like the Sarmatians before them, may well have kept two-humped Bactrian camels.[964] In the last centuries before and the first centuries after the beginning of our era, the camel, long domesticated, served the barbarians from the Great Wall to the Crimea as pack and riding animal.

In an instructive article, Schafer marshalled the literary evidence for the presence of the camel among the Hsiung-nu, T’u-yii-hun, and T’o-pa, in Shan-shan, Kucha, Karashahr, and K’ang-chii (Sogdiana).[965] The archaeological evidence is no less eloquent. A plate, representing two camels, was found in the Hsiung-nu cemetery at Hsi-ch’a-kou in the province Liao-ning;[966] camel bones were found in the Hsiung-nu settlement on the Ivolga near Ulan-Ude;[967] a bronze plaque with a camel rider[968] and another one with two standing camels[969] come from the Minusinsk area.[970] Among the rock pictures on the Pisannaya gora at Sulek in the old Kirgiz country are fighting camels.[971] A gold plaque in the Siberian collection of Peter the Great shows a tiger attacking a camel.[972]

Some objects showing representations of camels, found in Sarmatian territories, were of Western provenance. An open-work plaque from the Manych River showing a camel[973] can be dated to the first half of the second century b.c.; like the Greek cantharus in the same grave, the piece probably was imported. The same could be true for a finger ring with two kneeling camels found in a kurgan at Bolshaya Dmitrievka in the province Saratov.[974] In form, technique, and style, the ring is related to one showing a human head from Ust’-Labinskaya, datable to the first century a.d.[975] and another one showing goats of about the same date from stanitsa Tifliskaya.[976] But there also exist representations of camels which are of Sarmatian provenance. One is a bronze plaque showing two fighting camels from Pyatimary on the Ilek River[977] and another with a camel in low relief, a stray find from Aktyubinsk, farther to the east;[978] both are of the Sauromatian period (VII-IV b.c.). A buckle from Veselyi, east of Rostov on the Don, showing a lying camel,[979] is Early Sarmatian (IV-II b.c.).

Camel bones were found in a settlement on the Yurgamysh River near Chelyabinsk,[980] at Zolotaya Balka on the lower Dnieper (not later than the second century a.d.),[981] and, in the fourth century, graves in the Necropolis at Panticapaeum.[982] There is, furthermore, the bashlyk made of camel hair in a burial near Phanagoria, datable to the third century a.d.[983] It is unlikely that such a simple hood should have been imported; it was made where it was found, in the Bosporan kingdom.

We may, therefore, assume that the Hunnic tribes in the Black Sea region, the conquerors and successors of the Sarmatians, had camels. Their herds were apparently small; no Byzantine writer mentions camels among the Pontic Huns.[984]

Hunnic Agriculture?

Our sources are unanimous in denying the Huns any knowledge of agriculture. “No one among them plows a field or touches a plow handle,” wrote Ammianus. According to Claudian: “The chase supplies their food; bread they will not eat.” Asterius of Amasea described the Huns at the Black Sea as a people “who have not learned to grow wheat and other grains”; they have no grapevines and do not till the soil. The Huns “despised” agriculture, said Chelchal, himself a Hun.[985] The same was said about the Alans. They, too “cared nothing for using the plowshare.”[986] Literally taken, Ammianus was right. Neither the Huns nor the Alans, nor any other Sarmatians, plowed their fields. Nowhere between the Volga and the middle Danube has a plowshare been found that could be connected with the Huns or Alans. As late as 1925, when quite a number of kurgans had been excavated, Rykov could say that in the Sarmatian finds in the Volga region neither corn-grinders nor sickles occurred.[987] This is no longer true.

In 1936, Sinitsyn found in the mound over a Late Sarmatian grave at Tsagan-El’sin near Elista in the Kalmuk steppes the two parts of a primitive implement for crushing seeds of a cereal plant: a long, narrow, shallowly concave bedstone and a round grinding stone.[988] In Middle Sarmatian graves, millet had been found before and also occasionally charred wheat in the remnants of the funeral feast, but such finds were merely registered, and that was all.[989] The Sarmatian corn-grinder did not fit the picture of the shepherds who supposedly knew nothing of agriculture. Like P. D. Stepanov, author of a study on the history of agriculture in the lower Volga region,[990] Sinitsyn thought that the Volga Sarmatians got grain from the Kuban and Azov Sea areas;[991] they ground it—this could no longer be doubted—but they did not grow it. Why they imported it was not clear. Obviously only small amounts could be carried over such distances, so only two explanations were possible: either grain was used in religious ceremonies, or the chieftains cherished grain as a delicacy. Neither was exactly convincing. When later the fragments of another corn-grinder were found in a Middle Sarmatian grave at Berezhnovka,[992] they were not even recognized as such.

The find of an iron sickle in a Late Sarmatian grave at Kalinovka on the left bank of the Volga north of Volgograd[993] proves definitely that in the first centuries a.d. the Sarmatians did grow grain.

The sickle, 16 cm. long, its point broken, lay at the feet of a man in the niche of a narrow rectangular pit; other finds were a wire fibula with a piece of cloth still on it; an iron buckle; bone strips from a bow; and bone arrowheads. The southwest orientation points to the early stage in the Late Sarmatian Period.[994]

Agricultural implements are rarely found in graves. It is rather surprising that any were found in Sarmatian burials at all. The many hundred graves of the Gepidic peasant population in Hungary yielded one sickle.[995]

Whether the corn-grinders in some kurgans at Novo-Filippovka in the valley of the Molochnaya River, between the Dnieper rapids and the Azov Sea, are Middle or Late Sarmatian cannot be determined. They are only once mentioned in passing.[996] The graves in the cemetery are preponderantly Middle Sarmatian, but some seem to be as late as the third century a.d.[997]

In the 1920’s, Rau found in a Middle Sarmatian grave on the Torgun River an iron implement which he called a lapped axe.[998] He did not comment on it, and for years nothing like it was found until Shilov opened a grave in kurgan 8 at Kalinovka, also of the Middle Sarmatian period. There lay what used to be called a celt.[999] It was an adz so well preserved that it was possible to determine its function: the weak socket, not even closed around, and in particular the bluntness of the edge leave no doubt that the material on which the adz was wrought was rough and loose, thus, earth.[1000] On the walls of grave pits, the traces of narrow adzes, 3 centimeters, 4 centimeters, at the most 5 centimeters wide, are frequently visible.[1001] Such adzes were used for digging pits as early as the fifth century b.c. It seems rather unlikely that the Sarmatians used the adzes for this purpose only.[1002] These tools cannot have been lying around to be picked up when someone died. They must have been used for digging much more regularly. In other words, they were hoes, tools for tilling the soil in which seeds of cereal grasses were planted. The remnants of soft food found in pots were, as a rule, porridge of millet, Panicum miliaceum,[1003] the fastest growing cereal grass, just right for shepherds.[1004] Indeed, according to Pliny and Aelian,[1005] millet was the food of the Sarmatians.

Future excavations undoubtedly will prove that in wide areas of Middle Asia agriculture played a greater role in the economy of the nomads and seminomads than we still are prepared to admit, certainly subordinated to sheep and cattle raising and yet of considerable importance. Kadyrbaev found corn-grinders in graves of nomads in central Kazakhstan, some to be dated as early as the fifth century b.c.,[1006] Litvinsky in kurgans in the Kara-Mazar Mountains in Tadjikistan, datable to the second and third centuries a.d.[1007] The long acquaintance with agriculture, though primitive and limited, made it comparatively easy for Sarmatians to give up their nomadic way of life. To give a few examples, in the small fortified settlements near the present Ivanovka and Tarsunov on the Kerch peninsula, Sarmatian soldiers of the Bosporan kingdom cultivated their fields like the limitanei in the West.[1008] The Sarmatians of the settlement Kobyakovo at the mouth of the Don had become farmers.[1009] In 442, to King Goar’s Alans “land in farther Gaul was assigned by the patrician Aetius to be divided with the inhabitants. The Alans subdued those who resisted by force of arms, and, ejecting the owners, took possession of the land by force.”[1010] This is, so far as I know, the only case of resistance against barbarian hos- pites. Under the hospitalitas system the barbarians received a third of the land.[1011] But apparently this was not enough for the Alans. They needed more land; they had come with their wives and children, tents and carts.[1012] Though they could not have with them large herds and flocks, they probably wanted to live like their fathers in Hungary and their ancestors in South Russia. They wanted pastures, not just fields. Aetius made a mistake. But that he could think the Alans would be satisfied with what he gave them shows that he expected the Alans to cultivate the land.

As we now return to the writers who denied the Huns any knowledge of agriculture, we shall, perhaps, be less inclined to accept their statements. Claudian’s characterization of the Huns as mere hunters is so much nonsense. Ammianus transferred to the Huns what Trogus had said about the Scythians. Nevertheless, he presumably was right for his time. In times of war and migration, the Huns lived on their sheep and cattle. Once they had made themselves masters of a peasant population, like the settled Sarmatians and Germanic tribes in Hungary, they found it simpler and more pleasant to rob their subjects than to work themselves. Only the poorest Huns may have been forced to supplement their meat, milk, and cheese diet with self-grown grain. But that was probably different in the past.

Finds in Kunya Uaz in Khwarezm and on the upper Ob indicate that in former times the Huns tilled the soil. The racially mixed population of Kunya Uaz, Europoids with a Mongolian admixture, people who practiced cranial deformation, cannot be separated from the Huns. They had sickles.[1013] It could be argued that by the third and fourth centuries the Hunnoids in Kunya Uaz had been assimilated with the earlier local population; their sickles could have been taken over from the Khwarezmian peasants. But the people on the upper Ob (likewise Europoids with a Mongoloid admixture, likewise practicing cranial deformation, and at that of the same circular type as in Kunya Uaz) met hunters and fishers when they moved there in the second or third century. And yet, as Nerazik noticed, their sickles resembled closely those of the Kunya Uaz people.[1014] If the Hunnoids on the Ob and east of the Aral Sea cut stalks of grain with iron sickles, the conclusion that components of the great Hun horde, and not only the Alans, did the same in the past seems inevitable.


“The Huns,“says Ammianus, “are never protected by any building, but they avoid these like tombs, which are set apart from everyday use. For not even a hut thatched with reed can be found among them. But roaming at large amid the mountains and woods, they learn from the cradle[1015] to endure cold, hunger, and thirst. When away from home [peregre], they never enter a house unless compelled by extreme necessity; for they think they are not safe when staying under a roof.”[1016]

It would seem that the Huns had read Seneca, who praised the happy age when men spent their lives under the branches of the trees, dwelling according to nature in which it was a joy to live, fearing neither for the dwelling itself nor for its safety.[1017] Actually, Ammianus transferred again on the Huns the primitive traits of the “Scythians,” the “noble savages,” so dear to the Stoic philosophers, only using them as evidence of the beastliness of the hated barbarians. In his time the northern peoples’ fear of houses had become a topos. He speaks of the Alamanni who avoided cities “as if they were tombs, surrounded by nets.”[1018] The Goths are said to have thought that people living in cities lived not like men but birds in a cage.[1019]

Gainas fled Constantinople which looked to him like a crowded and sumptuous tomb.[1020]

In South Russia, the Huns had no permanent dwellings but they certainly had shelters, tents of felt and sheepskin,[1021] materials which probably most of them were still using after they had settled in the Hungarian plains. Priscus once mentions Attila’s tent.[1022] It probably was similar to the large tent of the Sarmatized Bosporan depicted on the wall of the catacomb of Anthesterius.[1023] On the painting, the interior of the tent is blue, evidently representing a woolen carpet like the one in Queen Ereka’s house.[1024] Incidentally, as everyone who has lived in Mongolian felt yurts knows, they are quite comfortable, spacious, well aired, and easily kept clean. Living as a prisoner in the Chinese capital, Hsieh-li, kaghan of the Turks, refused to move into a house and put up his tent.[1025] The crown prince Li Ch’eng-ch’ien preferred a Turkish tent to the palace,[1026] but he was a noted and crazy Turkophile. Attila lay in state in a silk tent.[1027] The one he used when he was not in one of his residences and some tents of high-ranking Huns may have been made of the same materials.

By the middle of the fifth century, the Hun nobles had houses in the villages which they owned,[1028] better built than the modest huts of the native population,[1029] probably similar to the wooden buildings in the king’s residences, only on a smaller scale. The walls of the latter were made of well-planed planks and panels. Attila’s “palace” consisted of a single square or rectangular room, furnished with seats and a bed or couch, xMvq, screened off at one end of the room by tapestries. Thompson rightly pointed out that the “palace,” the other one-room houses, and the two stockades around the camp were not built by Huns but by either Romans or Goths.[1030]

Clemmensen adduced good arguments for the Germanic, which in our case means Gothic, technique of the wood construction.[1031] In the second half of the fourth century there were Christian churches, monasteries, and convents in Gothia,[1032] evidently wooden buildings.[1033] Since the discovery of the Gothic long houses in the Chernyakhov settlements, the existence of Gothic wooden architecture need no longer be proved. The Sarmatian Jazyges had no houses in their old sites in South Russia, but after more than two centuries of close contact with the Germanic Quadi, they lived in thatched huts. Attila and his retainers most probably had their houses built in Gothic fashion by Gothic carpenters.[1034]

Income in Gold

In the 440’s, the East Romans paid the Huns about 13,000 pounds of gold, more than 900,000 solidi. This was, from whatever angle one may look at it, a great sum. In particular, the payment of 6,000 pounds of gold in 447 must have been a heavy blow to the imperial treasury. But did it really spell the complete financial ruin of the prosperous East, as Mommsen thought?[1035] For a proper evaluation of the “subsidies” paid to the Hun “federates,” a brief survey of comparable public and private expenditures in the fifth and sixth centuries may be helpful.

In 408, Alaric blackmailed the West Romans to pay him 4,000 pounds of gold;[1036] in the same year, he blockaded Rome, and the senate bought him off with 5,000 pounds of gold, 30,000 pounds of silver, and other gifts in kind.[1037] These figures, coming from Olympiodorus, may not be entirely trustworthy. But there is no reason to doubt Malchus’ statement that in 473 Theoderic Strabo, leader of the Gothic federates, got an annual payment of 2,000 pounds of gold.[1038] The sums offered or actually paid to the Goths varied considerably according to the circumstances. The subsidy paid to Valamir was only 300 pounds of gold per annum.[1039] In 479, his nephew Theoderic, the later great king, was offered an annual subsidy of 10,000 solidi, that is about 140 pounds of gold, but immediate payment of 1,000 pounds of gold and 40,000 pounds of silver.[1040] In 570, Emperor Tiberius offered the Lombards 3,000 pounds of gold if they would stop their raids in Italy.[1041] This was the same year in which Bayan, the caganus of the Avars, was paid his annual subsidy of 80,000 solidi, or more than 1,000 pounds of gold.[1042]

In 532, Emperor Justinian concluded the “endless peace” with Chosroes; one of its conditions was the payment of twenty annual contributions to the maintenance of the fortifications in the Caucasus for which the Romans were in arrears, amounting to 11,000 pounds of gold.[1043] In 540, the Persians received again 5,000 pounds; in 545, 2,000 pounds; in 551, 2,600 pounds; and in 561, 3,000 pounds.[1044] From 484 to 492, Zeno paid the gangs of robbers in the Isaurian highland a yearly subsidy of 1,400 pounds of gold.[1045]

In order to put the tribute paid to the Huns in the proper perspective, it should not only be compared with the payments to the “allies.” Measured by the expenditures made by high-ranking people on worthy and sometimes not so worthy causes, it was not so exorbitant. To give a few examples: Empress Eudocia contributed 200 pounds of gold to the restoration of the public baths in Antioch;[1046] Empress Eudoxia gave the same sum for the building of a church in Gaza.[1047] When Paul, ex-consul of 498, was in financial trouble, Emperor Anastasius helped him out with 2,000 pounds of gold.[1048] In 514, Anastasius ransomed Hypatius from Vitalian for 5,000 pounds of gold.[1049] In 526 and 527, Emperor Justin sent 4,500 pounds of gold to Antioch, which had been heavily damaged by an earthquake.[1050] To celebrate his consulship in 521, Emperor Justinian spent 4,000 pounds of gold on the games and for distribution among the populace;[1051] in 532, he gave 4,000 pounds of gold for the building of Saint Sophia.[1052] The sums spent in the vicious ecclesiastical fights were enormous. In the 430’s, Bishop Cyril of Alexandria bribed court officials with more than 2,000 pounds of gold.[1053] Between 444 and 450, Nomus, magister officiorum, consul in 445, and patricius, extorted from Anastasius and Paul, Cyril’s nephews, 1,400 pounds of gold.[1054]

In the fifth century, the revenue of the Eastern empire has been estimated as being on the average 270,000 pounds of gold a year, of which approximately 45,000 were spent on the army.[1055] The 6,000 pounds of gold paid to Attila in 447 were a little more than 2.2 percent of the money the treasury received in a year, and the highest annual tribute was about 4.7 percent of what the army required. Still, had this gone on for a number of years, it would have been a great, though still not an unbearable strain. But Attila was paid the tribute only in 448, 449, and, possibly, in 450. In the following three years he was at war with both the East and West and consequently received nothing.

A passage in John Lydus, which escaped Mommsen, shows how far from the alleged bankruptcy the East was. When Leo followed Marcian to the throne in 457, he found in the treasury more than 100,000 pounds of gold, “which Attila, the enemy of the world, had wanted to take.”[1056] Of all the Byzantine emperors after Marcian, only Anastasius left a larger reserve at his death.[1057]

The tribute was not the only so-to-speak legitimate source of the gold income of the Huns. Before, and for some time while, they received annual subsidies, the Hun leaders were paid in gold for the auxiliaries they lent the Romans. Aetius, in particular, must have paid large sums for the contingents of horsemen he obtained in the 430’s. Whether by the middle of the 440’s Attila blackmailed the Western Romans into sending him gold for keeping the peace is not certain, but in 449 he drew a salary as Master of the Soldiers, which, as Priscus said, was a pretext for concealing the tribute.[1058]

The Huns probably insisted that part of the tribute should be handed over to them in ingots. They must have known as well as the Romans that many clipped, debased, and counterfeit solidi were in circulation. In 366, the taxgatherers were ordered to reduce the solidi “to a firm and solid mass of gold”;[1059] a year later, the edict was repeated: “Whenever solidi must be paid to the account of the sacred largesses, the actual solidi shall not be delivered, because adulterated coins are often substituted for such solidi. The solidi shall be reduced to a mass.... Whenever a definite sum of solidi is due under a title of any kind, and a mass of gold is transmitted, a pound of gold shall be credited for seventy-two solidi.”[1060] As the sixteen ingots found in 1887 at Krasna in Transylvania[1061] show, the Visigoths were likewise on their guard against such attempts at deception. The Huns hardly put more trust in the honesty of the Romans. Besides, not all solidi were of the same weight, though the deviation from the standard was, as a rule, insignificant. It is, therefore, all the more remarkable that just in a barbarian hoard from Kirileny in the Moldavian SSR, hidden about 400 a.d., there was a solidus which, instead of the standard 4.54 grams, weighed only 3.90 grams.[1062] The barbarian had been cheated. As the Huns had no mints,[1063] they obviously demanded only that amount of gold in ingots which they intended to use for ornaments; for commercial transactions at the fairs, and otherwise, they needed coins.

The Persian kings often lifted the siege of a city as soon as the beleaguered raised the money demanded from them. In 540, Edessa, for example, paid Chosroes 200 pounds of gold and four years later 500 pounds.[1064] There is no evidence that Attila or the kings before him made a town an offer to save it at a price. They obviously thought it more profitable to storm a place at the cost of a few hundred men, mostly expendable foot soldiers, to loot it, and to carry away the captives to be sold or ransomed.

After their victory at Adrianople, the Goths offered so many ten thousands of captives for sale; the Huns, temporarily allied with the Goths, certainly had their share in the lucrative business. St. Ambrose did what he could to ransom the Christian prisoners. In De officiis he wrote:

The highest kind of liberality is to redeem captives, to save them from the hand of the enemies, to snatch men from death, and most of all, to restore children to their parents, parents to their children, and to give back a citizen to his country. This was recognized when Thrace and Illyria were so terribly devastated. How many captives were then for sale all over the world? Could one put them all together, their number would have surpassed that of a whole province.... It is then a special quality of liberality to redeem captives, especially from barbarian enemies, who are moved by no spark of human feeling to show mercy except so far as avarice has preserved it with a view of redemption.... I once brought odium on myself because I broke up the sacred vessels to redeem captives, a fact that could displease the Arians. Who can be so hard, cruel, ironhearted, as to be displeased because a man is redeemed from death, or a woman from barbarian impurities, things that are worse than death, boys and girls and infants from the pollution of idols, where through fear of death they were defiled?[1065]

In 395, the Huns took thousands of prisoners in the Asiatic provinces and the Caucasus. Far away from their homes, these unfortunates were not ransomed and most of them were sold at the slave markets on the Danube. Although the tribute was paid to the Hun kings, the prisoners were sold by the men who took them, who apparently received also the ransom for Roman soldiers who fell into their hands, 8 solidi a head before 435 and 12 thereafter. How much gold flowed into Hunnia in this way is difficult to say; it seems to have been rather considerable. The ransom for civilian captives could be quite high. When Attila wanted to show his generosity, he asked only 500 solidi for the widow of a wealthy citizen.[1066] The ransom for Bigilas was 50 pounds of gold, that is, 3,600 solidi,[1067] but this was a special case.

How much gold unminted and in coins the Huns brought back from their raids and looting expeditions cannot even be guessed. After their victory over the Ostrogoths they did not press the attack on the Visigoths “because they were loaded down with booty,”[1068] certainly not cooking pots and wooden benches but gold, silver, and precious weapons. The same happened in upper Italy in 452.

In addition to the tribute, the Romans had to send “gifts” to the Huns. This was, in itself, nothing unusual. Even if the treaties between the Romans and the barbarian rulers provided only the payment of a certain sum, it was customary to give the latter presents,[1069] among them objects of precious metal. The Huns did not expect gifts, they demanded them. When in 450 the Roman ambassadors whom Attila refused even to see would not hand over the gifts they had brought with them, the king threatened to kill them.[1070]

On their departure from Constantinople, foreign envoys were given presents. It was an act of courtesy for distinguished guests. The sums involved could be huge. Procopius estimated the total lavished by Justinian on a Persian ambassador at 1,000 pounds of gold.[1071] Attila made a lucrative business out of this custom. Under the flimsiest of pretexts he would send embassy after embassy to the imperial court. To keep the savage in good humor, they all were given rich presents for which, on their return, they had to give account to the king.[1072]

Another, probably very considerable, source of income in gold was the sale of horses to the Romans. Besides slaves and, possibly, furs, there was not much else the Huns could offer the Roman traders. A passage in Vegetius’ Mulomedicina shows that at times the export of horses from Hunnia was a flourishing business. It probably shrank in the later 440’s, after the Huns in two sanguine wars lost not only many men but many horses.

A little-noticed passage in Priscus indicates that in Hunnia gold coins were, though probably only to a modest extent, in circulation as a medium of exchange. In 449, Attila forbade the Roman envoys “to buy any Roman prisoner or barbarian slave or horses or anything else except things necessary for food until the disputes between the Romans and the Huns had been resolved.”[1073] The king had a good reason for this prohibition; he wanted to catch Bigilas with the 50 pounds of gold to be paid to Edecon for killing his lord. When later Bigilas was led before Attila and asked why he was bringing so much gold, he was unable to explain away the 3,600 solidi he was carrying with him.[1074] The passage shows that not only at the frontier but also deep in Hunnia, slaves, horses, and food could be bought and sold for Roman gold coins. Whether in Attila’s time the Huns used solidi as currency only in their contacts with the Romans or also among themselves we do not know. The latter possibility cannot be ruled out entirely.


The long, costly, and indecisive war which Emperor Valens waged with the Visigoths ended in 369 with a treaty that reduced to a minimum the formerly fairly close contacts between the empire and the barbarians across the Danube. The Romans stopped paying the annual subsidies to which the Goths had been entitled as long as they were federates. The one-sided exchange of “gifts” between the emperor and his “friends” came to an end. Before the war, Romans and Goths had been bartering all along the river, and many officers of the frontier army were merchants and slave dealers rather than soldiers.[1075] From 369 on, the trade between Romania and Gothia, which was now as independent as Persia, was restricted to two market places on the left bank of the Danube.[1076] Even there, to judge from analogies, traders were permitted to bring their wares and transact business only at certain times of the year.

The imperial government saw strictly to it that the commercial relations between its subjects and the free barbarians were kept within the narrowest limits. There were only two market places for trading with the Quadi and Marcomanni.[1077] To control the trade with the Jazygi, a burgus “Commercium” was built near Gran in 371 ;[1078] the other burgi were obviously too engaged throughout the year to keep watch over the restless barbarians and to prevent “the furtive crossings of pillagers” (clande- stinos latrunculorum transitus). A law of 368 forbade the export of wine and oil to the barbaricum.[1079] A few years later, merchants who paid in gold for slaves or other goods were threatened with death.[1080] The same punishment threatened those who sold weapons[1081] and materials for making weapons to any and all barbarians.[1082] Whether the trade treaties with the Persians stipulated what goods could be exported is not known, but we may be sure that the Romans did not sell arms to the King of Kings. On the Persian frontier, too, trading was restricted to a few places. In 409, “lest foreigners might find out secrets, which would be improper,” the Romans permitted trade with the Persians at three places only: Nisibis, Artaxata, and Callinicum.[1083]

What Priscus, our only authority on trade relations between the Eastern Romans and the Huns, has to say on the subject fits this picture. The fairs were held at fixed dates, once a year,[1084] probably in late spring or early summer.[1085] As long as the frontier ran along the Danube, the market was there, presumably on the northern bank. After 447, it was shifted to Naissus (Nis).[1086] When Dengizich and Hernac, Attila’s sons, asked for peace, they requested, among other things, that the market on the Danube be reopened “as in former times.”[1087] There apparently was only one place where Romans and Huns met for barter.

It does not follow that the trade with the Huns was negligible. In addition to the legal trade, Roman goods probably were smuggled into Hunnia, and Hunnic horses and slaves into Romania. Still, the volume of both legal and illegal trade was apparently modest. Thompson’s assertion that the whole bourgeoisie of the Eastern empire was vitally interested in maintaining and expanding its commercial relations with the Huns[1088] has no basis in either the literary or archaeological evidence. Undoubtedly, some people did good business. If at fairs within the empire profits of 50 percent could be made,[1089] the trade with the barbarians was certainly even more lucrative, in particular because the traders had no pangs of conscience about cheating the Huns. Saint Ambrose thought it not a sin to lend barbarians money at usurious interest: “On him whom you cannot easily conquer in war, you can quickly take vengeance with the hundredth. From him exact usury whom it would not be a crime to kill. Therefore, where there is the right of war, there is also the right of usury.”[1090]


Like the barbarians on China’s borders who valued silk more than any other product of their neighbor and enemy,[1091] the barbarians in the West esteemed Roman silk very highly. In 408, Alaric demanded and received from the city of Rome four thousand silk tunics.[1092] His successor, King Ataulf, gave fifty young men clad in silk as wedding present to Galla Pla- cidia.[1093] In the shiploads of clothes which the Eastern Romans for many years sent to the Visigoths[1094] there were doubtless many silk tunics.

The Huns obtained silk in various ways. First, they brought it home from their raids. Like the Goths in Italy, the Huns, while they were still in Roman territory, bought silk from Roman dealers. Unde pellito serica vestimenta’l asked Maximus of Turin.[1095] Second, the Huns bought silk at the fairs; in the preceding centuries silk reached the barbarians in the steppe via the cities on the Euxine; silk was found near Kerch in the Crimea,[1096] in a Late Sarmatian grave at Marienthal (now Sovetskoe) on the Big Karman River in the former German Volga Republic[1097] and in a grave at Shipovo.[1098] Finally, the emperor sent silk as gifts to the Hun nobles and Attila, as he later sent silk clothes to the Avar caganus.[1099] Attila lay in state in a silk tent.[1100] Edecon and Orestes may have looked strange in the Roman silk garments, but they evidently liked them.[1101]


If Asterius of Amasea is to be believed, the Huns on the Black Sea did not drink wine,[1102] probably not because they did not like it but because they could not get it. It was very different in Hungary. From Priscus we learn that at Attila’s court wine was drunk in great quantities. One- gesius’ wife offered Attila a goblet of wine. At the great banquet, before food was served, Attila toasted all the prominent guests, including the Roman ambassadors, with wine, and they in turn, toasted the king. After the first course again wine was drunk, and after the second, and when the Romans left, late at night, the Huns kept on drinking. At the dinner in Adamis’ house each guest was given a beaker of wine from the others, and he had to reciprocate. As neither the Huns nor their subjects, with the possible exception of the few Romans, knew how to grow grapes and make wine, it is evident that wine was imported to Hunnia in great quantities. In the sixth century, the Massagetae—Huns in the Byzantine army—were the most intemperate drinkers,[1103] even worse than the Goths.[1104]

IV. Society

In no area of Hunnic studies is the discrepancy between the few facts and the theories built on them as striking as in the study of Hun society. The temptation to force the Huns into the favorite socioeconomic category of the student seems to be irresistible. Later Byzantine authors often transcribe the titles of the barbarians; they speak of the %aydvog of the Avars, the fioiXac; of the Bulgars, and the xovdovvot; of the Khazars. Priscus used only Greek words for the ranks and titles of the Huns. What word he rendered by PaatXevq is not known. But some modern authors call Attila “kagan” as if they had been with Priscus at his court and, knowing Hunnish, understood how his subjects addressed the king. Others, lumping together all Eurasian nomads and seminomads, from the Scythians to the Kazakhs of the nineteenth century, construct what they call the nomadic society, throwing around supposedly technical terms like il and ordu. The worst sinner in this respect was T. Peisker, who still has his followers.[1105] Thompson views the Huns as a howling mass of half-naked savages. In his tendency to push not only the Huns but also their allies way down the ladder of evolution, Thompson even mistranslates the texts. He refers to Sozomen IX, 5: “The ecclesiastical historian saw numbers of them [Sciri] scattered over the foothills and spurs of Mount Olympus in Bithynia, presumably acting as shepherds on Imperial estates.”[1106] Actually, Sozomen saw them tilling the soil, yetogyowra?. Soviet historians find for the Huns a place in the unilinear evolution of social functions drawn up by Lewis Morgan and more or less faithfully followed by Engels. The Huns are said to have been in the last stage of “barbarism,” when “gentile society” developed into “military democracy,” which Engels characterized as follows:

The military commander, the council, and the popular assembly formed the organs of military democracy, military because war and the organization of war were now the regular functions of life of the people. The wealth of their neighbors excited the greed of the peoples, who began to regard acquisition of wealth as one of the main purposes in life. They were barbarians: plunder appeared to them easier and even more honorable than productive work. War, once waged simply to avenge aggression or as a means of enlarging territory that had become inadequate, was now waged for the sake of plunder alone, and became a regular profession.... The growth of slavery had already begun to brand working for a living as slavish and more ignominious than engaging in plunder.[1107]

According to Engels, the Greeks in the heroic age were typical representatives of military democracy. The Soviet historians, untiringly repeating that the Huns had reached the same stage,[1108] of course do not even try to prove it. Attila and Agamemnon shared the initial vowel in their names, but this is about all. If all peoples who under military leaders robbed their neighbors lived in a military democracy, Assyrians, cattle-raising Zulus, agricultural Aztecs, and the Viking pirates would belong together. After many attempts to define military democracy more precisely, it eventually has become an empty phrase. It is, we are now told, a type of political superstructure which does not reflect the processes going on in the economic base.[1109]

The only Soviet student of the Huns who took Engels seriously was A. N. Bernshtam. Because the society that follows another is supposed to represent a higher stage in the development of mankind, the young military democracy of the Huns must in its time have played a progressive role. Bernshtam gave the concept of progress an original twist. He did not maintain that the Huns themselves were more developed than the peoples they conquered. Their contribution to progress was rather an indirect one: they helped to break down the “slave-holding” societies, including the Roman Empire, thereby clearing the way for more progressive feudalism. This was the thesis which Bernshtam presented in his Ocherki po istorii gunnov (History of the Huns in Outline).

He was furiously attacked.[1110] It is true, Bernshtam committed some bad blunders;[1111] instead of referring to the sources, he often quoted from hopelessly obsolete compilations. But his main sin was to put the Huns on the same level as the young barbarian peoples, the Slavic and—though this was merely whispered aside—the Germanic tribes. Like their successors, the Avars, Pechenegs, and Mongols, the Huns were the arch enemies of the peace-loving nations of eastern Europe. Bernshtam’s book was taken out of circulation.

The obligation to stay within the Marxian framework leads to strange results. The Hungarian historian and philologist Harmatta published a number of stimulating articles on Hun society, with long quotations from the original Sanskrit, Akkadian, Pehlevi, and Sogdian.[1112] After carefully weighing the pros and contras he came to the conclusion that Hun society became a state in 445 a.d., give or take one or two years. Yet the recalcitrant Huns refused to fit into one of the stages permitted by Engels. Hun society, Harmatta admitted, “had no definite character of its own.”[1113]

The meaning of the terms for the social institutions of the Huns has to be established by the context. Aoyddeq, say the dictionaries, means “picked men.” Is this the meaning of the word in Priscus? Because students of the Huns read the early Byzantine texts as if they had been written by Thucydides, their works contain a number of misunderstandings. In the following I shall deal only with Thompson’s and Harmatta’s views of Hunnic society. They are the only authors who gave the subject some thought.

Priscus, the only writer to speak about the logades of the Huns, calls five of them by name:

1. Onegesius, “who held power second only to Attila among the Scythians.” EL 1342.

2. Scottas, Onegesius’ brother. He boasted that he could “speak or act on equal terms with his brother before Attila.” EL 12718>23.

3. Edecon, a famous warrior of Hun descent. EL 1246_7.

4. Berichus, lord of many villages. EL 1471O.U.

5. Orestes, a Roman from Pannonia, Attila’s secretary. EL 12522.

The word occurs also in eight additional passages:

6. Edecon and Orestes and Scottas “and the other logades.” EL 12522.

7. “The logades of the Scythians, after Attila, took the captives from the well-to-do because they sold for the most money.” EL 13532-1362.

8. Onegesius took council “with the logades.” EL 14530.

9. The Roman ambassadors went to the house of Adamis “with some of the logades of the people.” EL 1469_10.

10. Attila ordered “all the logades around him” to show friendship to Maximinus. EL 14726.

11. Chrysaphius inquired whether admission to Attila’s presence was easy for Edecon, who answered that he was an intimate friend of Attila and entrusted with his bodyguard, “along with the logades chosen for this [duty].” EL 58020–25.

12. Chelchal summoned the logades of the Goths. EL 5892o-21.

13. Kunchas, king of the Kidaritae, wishing to punish Peirozes for his falsehood, “pretended to have a war with his neighbors and to need men, not soldiers, suited for battle, for he had an infinite number of these, but men who would prosecute the wars as generals for him.” Peirozes sent to him three hundred logades. EL 15420-21.

In Thompson’s opinion the logades were the hinge on which the whole administration of the Hunnic empire turned. He identified them with Attila’s eztiT^beioi and the olneloi xal Xo%ayol of Uldin. They were supposed to have ruled over specific portions of the empire, kept order among the subject nations, and collected tribute and foodstuff from them. During a campaign they commanded not only specific squadrons of the Huns assigned to them, but also contingents of subject warriors provided by the districts they possessed. Thompson did not translate the word as if logades were a technical term; he even speaks of the time when the logades were “instituted.”

Harmatta first stressed the fact that the logades mentioned by Priscus had not Hunnish but Germanic and Greek names; Attila was supposed to have liquidated the old tribal organization and to have ruled with the help of the logades. Later Harmatta rejected Thompson’s equation of the logades with eTtrnydetot, which, indeed, means nothing more specific than “friends,” and also of Uldin’s olneloi xal Xo%ayol, his “kinsmen and officers.” He dropped the dependency of the logades on Attila. Now they were supposed to have been the ruling class, comparable to the va- zurgan ad azadan, “the great and noble” of Sassanian Persia, or the bdg- lar of Turkish society in the sixth century. For the rest, Harmatta agreed with Thompson. His logades likewise ruled over their territories, collected taxes, and so forth.

Altheim takes the logades for “a new closed estate.” They got their name, he says, because they were, literally, picked by Attila,[1114] who employed them in his campaigns, on diplomatic missions, and for collecting taxes.

These scholars read too much into Priscus. He says nothing about the collection of taxes. He mentions that Berichus was lord over many villages, but it does not follow that all logades, even Attila’s secretary, were large landowners. The wretched Goths who in the late 460’s roved through the northern Balkans had their logades. As they had no land— they asked the Romans for land—their logades could not own large estates; nor did they have a king to “select” the logades. Fascinated by the word which they cannot find in the writings of Priscus’ time, Thompson, Har- matta, and Altheim turn it into the designation of a well-defined social group.

Actually, since the third century, logos means just “prominent, outstanding, distinguished.” In IIeqI ETttdeixTixcbv (ch. HI, Menander), Rhetor speaks about the dvdgsi; Aoyads^ of Athens; they were not picked by anyone, owned neither land nor horses, but were full of wisdom and virtue, ooeplaq xat aQETfj<; Tgoylpovi;.[1115] Basil, the older brother of Gregory of Nyssa, was Aoya? avyg xal dvopaoxoQ xaxa (ptkoaoqilav.[1116] In his refutation of Julian’s treatise Against the Galileans, Cyril of Alexandria, Priscus’ contemporary, praises the logades of the .Greeks, again neither landowners nor military leaders; the logades are Plato and Plutarch.[1117] In the eighth century, Theophanes, probably quoting an earlier work, wrote of the logades in Antioch who followed Nestorius’ doctrine.[1118] I learn from Professor I. Sevcenko that in the Russian chronicles logades is rendered by luchshie lyudi, “the best people.” In his translation of Anna Comnena’s Alexiad, B. Leib translates logades by “elite.”[1119] This is also the meaning in modern Greek: ol xov eOvovg XoyddeG are “the elite of the nation.”[1120]

There is no evidence that these prominent people of the Huns had anything in common except prominence. Had Priscus written Latin, he probably would have called them optimates. As used by Ammianus Mar- cellinus, optimates comes very close to logades. Ammianus was well informed about the ranks among the Alamanni against whom Julian fought. They were led by Chnodomarius and Serapio, potestate excelsiores ante alios reges; then came five kings, potestate proximi, then regales, a long train of optimates, and only then came the commoners (XVI, 12, 25–26). The Sarmatians in Hungary had regales, subreguli, and optimates (XVII, 12, 9, 12). It would seem that Ammianus’ optimates stood just one step above the common people. But speaking about the optimates of the Armenians (XXVII, 12, 2) and Goths (XXXI, 4, 1), Ammianus evidently not only had the lower nobility in mind but all men who had something to say. Emperor Valens refused to talk to the people whom Fritigern sent to him as envoys because they were of low rank; he demanded that the Goths send optimates, prominent men (XXXI, 12, 13). Hortarius, one of the two Alamannic primates whom Valentinianus appointed to commands in his army, was in treacherous contact with King Macrianus and the barbarian optimates (XXIX, 4, 7); here again optimates means simply prominent people or, as we may now say, logades. To the Hunnic logades correspond the Hogvcpalot of the Goths.[1121]

Priscus was well aware that not all the prominent people at Attila’s court had the same rank. He noticed that Onegesius sat at the right of the king, “the more honorable side,” and others, like Berichus, at the left. Bigilas told him that the Hun Edecon was far superior to the Roman Orestes. But Priscus was not much interested in the finer differences among the prominent men. The Roman ambassadors had, first and last, to do with Attila, and besides him no one really counted. Only what Priscus says about the Akatir[1122] gives us some information about the structure of Hun society.

The Akatir, a Hunnic people, eOvoq, were divided into tribes and clans under numerous rulers, tioAAcov xcrrd (pv).a xal yevg dgyovrcov. Kuri- dachus was the highest in power, 7iqeo.[1123]

[There is a break in the manuscript here.—Fd.]

This is one of the rare cases in early Byzantine literature in which the context permits one to determine the meaning of the terms for the subdivisions of barbarian peoples and their leaders. The people, eOvoq, consists of tribes, (pvl.a, and clans, ysvg. Kuridachus is (1) a Paai2.Evq of the Akatir ; (2) an ag^cov; (3) as a leader of a (pvkov, a (pv2.ag%oi;.

Thompson maintains that Olympiodorus distinguished carefully between the military commander of a confederacy of barbarian tribes and the military leader of an individual tribe, calling the former <pvXag%oq and the latter q^.[1124] He is mistaken. ‘P-fy- is, of course, Latin rex, whatever its original relationship to Celtic rigs may be.[1125] The Latin writers of the fourth and fifth centuries make no distinction between the reges beyond the frontiers. Ammianus Marcellinus calls the ruler of the Burgundians (XXVIII, 5, 10,13, 14), Quadi (XVII, 12,21), and wild Moorish tribes (XXIX, 5, 51) reges as well as the seven rulers of the Alamanni (XVI, 12, 25) and the great Shapur, partner of the Stars, brother of the Sun and Moon (XVII, 5, 3). The word, in the same meaning, was taken over by the Greeks. In a letter, written in 404, John Chrysostom speaks of the of the Goths in the Crimea.[1126] At about the same time Olympiodorus, who was fond of Latin words, talks about the first among the QfjyeQ of the Huns; they were, in their way, great lords, and the translation “king” seems quite in place. But Olympiodorus calls also the condottiere Sarus, dux of a small group of Goths,[1127] a p^.[1128] In Malalas, Brennus is rhex of the Gauls, Odovacar of the barbarians, the Vandal rulers are rheges of Africa or the Africans, those of the Ostrogoths rheges of Italy; Styrax and Glones are rheges of the Huns, and Boa is a rhegissa.[1129] One can, if one so wishes, translate rhex by “king,” but it seems preferable to transcribe the word. In any case, rhex in Olympiodorus is no more the military leader of a tribe than the Gothic ruler who asked for a new bishop for his people.

PaaiAev<; is another term with two meanings. In official documents as, for example, in diplomatic notes, it was used exclusively for the Roman emperor.[1130] The West Romans were afraid that some day Attila would insist on being addressed basileus instead of magister mi liturn, his (strictly nominal) title.[1131] It would be of interest to know how the East Romans addressed the king. They could have used such neutral terms as hegemon or hegoumenos of the Huns. After Bleda’s death monarchos would have been an appropriate title; Priscus called the Persian king monarchos,[1132] and Menander did not feel he had to explain why he wrote about the monarchos of the Langobards.[1133] Attila could have been addressed as xardexcov x&v Ovvvcov, as Theoderic Strabo was xardp/cov xoov rdxOoov, the Squinter was also even avtoy.odtWQ,[1134] a title usually reserved for the imperator.[1135]

Harmatta thought that Attila was the first barbarian ruler whom the East Romans granted the title basileus because his

social standing, power and absolute rule was similar to the position held by the Roman emperor; only once the term basileus became current in connection with Attila, i.e., with a barbarian ruler, the earlier sharp distinction between the Byzantine monarch and the barbarian kings became gradually obliterated in linguistic usage. This explains in Har- matta’s opinion why Priscus applied the term basileus on one occasion to the king of the Franks and on another to the chieftains of the Acat- ziri.[1136]

This sounds quite plausible, but it is not true. Harmatta overlooked the way authors, when they wrote history, spoke about barbarian kings. Had Eunapius been sent as ambassador to a Visigothic leader, he most certainly would not have addressed him as basileus. But writing about Athanaric, he did not hesitate to call him basileus.[1137] The ruler of the Cha- mavi, Julian’s enemies, was a basileus.[1138]* Eunapius wrote many years before Attila. And Priscus himself was rather generous in bestowing the title basileus on Hunnic as well as on other barbarian rulers. Attila was not the first king of the Huns. Ruga was also fiaaiteviov.[1139] The PaaiXela[1140] of the Huns devolved on Attila and Bleda;[1141] they are the fiaaiteu; of the Huns.[1142] Priscus speaks not only of the kings of the Franks and Akatir but also of the kingdom, /taaiAeia, of the Lazi in the Caucasus.[1143] In the writings of the sixth century we read about the kings of the Auxumitae[1144] and Iberians.[1145] Here again these kings were not acknowledged as such by the East Romans, but this was of no concern to historians.

A third ambiguous term, whose meaning Thompson and Harmatta defined much too narrowly, is gw2ap%o;. According to Thompson, in Olympiodorus phylarchos means the military leader of a confederacy of tribes. The word occurs in Olympiodorus five times. Alarich and Valia are phylarchoi of the Visigoths, Gunthiarius is the phylarchos of the Burgundians, and the Blemmyes have phylarchoi and prophetae, the priests of Isis.[1146] There is no reason whatever to assume that in the fifth century the Visigoths and Burgundians were confederacies of tribes.[1147] And the Blemmyes did not consist of confederacies which in their turn consisted of so and so many tribes; their phylarchois were clearly tribal leaders. Even further from the truth is Harmatta’s definition. “The word phylarchos,” he says, “denotes an official title given by the East Roman or Byzantine emperors to the leaders of the allied barbaric peoples, at least since the end of the fourth century.” But this is not enough. He continues: “These barbarian chieftains were given Roman auxilia, money, provisions, Roman advisers, and Roman dignities—in a word everything was done to stabilize their authority and power against the other members of their tribe.”[1148] Harmatta is mistaken; he refers to Olympiodorus. It is sufficient to read the first fragment on Alaric. Harmatta quotes only “Alarich, the phylarchus of the Goths.” But in the following lines Olympiodorus narrates how this “ally” of the Romans takes Rome, sacks it ruthlessly, and carries Galla Placidia, the sister of the emperor, into captivity.


Above the common people, qara budun, as they are called in the Orkhon inscriptions, stood the noble families. Both Attila and his father were “well-born.”[1149] In 449, when Priscus met Attila, the king’s beard was sprinkled with gray;[1150] he cannot have been born later than about 400, his father about 370, or even earlier, which proves the existence of a hereditary aristocracy long before the Huns broke into the Ukraine.[1151] How large it was, we have no means to determine. Priscus mentions noblemen only two times more. Berich, a prominent man, lord over many villages, was “well-born.”[1152] Somewhat more revealing is a passage preserved in Suidas: Bleda gave Zerco “from among the well-born women a wife who had been one of the attendants of the queen but who, on account of some misdemeanor, was no longer in her service.”[1153] The daughters of noblemen were, thus, also “well-born” but could be married to commoners, in this case the feeble-minded court jester, who was not even a Hun.

Provided that what Ennodius said about the Bulgars, whom he equated with the Huns, can be transferred to the latter, the distance between noblemen and commoners was apparently not great. In his Panegyric on Theoderic, Ennodius describes those Bulgars whom the Ostrogoths fought in Pannonia in 486 as a nation in which the man who killed the most enemies had the highest rank; their leaders were not born to nobility but became noblemen on the battle field.[1154] This was, of course, seen through Gothic eyes. Theoderic was a scion of the half-divine Amalungs, the Goths had their great noble families, and the relative social mobility of the Bulgars must have struck them as sheer savagery.


Some of the captives whom the Huns led away from the Balkan provinces and Italy were ransomed by their relatives and friends. Others served under their masters in the Hun armies until they were able to buy their freedom with their share in the booty.[1155] But most captives were sold to Roman slave dealers either at the annual fairs or, before these were held regularly, wherever the Huns had close contact with the Romans, even while the invaders were still in Roman territory. In 408, the Romans bought so many captives from Alaric’s Visigoths that a law had to be issued to specify the conditions under which these unfortunates could regain their freedom.[1156] According to a homily of Maximus of Turin, the barbarians sold country lads to Roman slave dealers not from distant districts but from villages near Turin.[1157] It may be assumed that the same sordid transactions took place during the Hunnic raids south of the Danube.

The Huns sold most of their captives not merely because they “burned with an infinite thirst for gold.”[1158] They themselves had little use for them. In the economy of pastoral nomads only a small number of slaves can be usefully employed; besides, it is difficult to prevent their escape. This was true for the Mongols in Chinghiz Khan’s times as well as for the Kazakhs of the eighteenth and nineteenth centuries.[1159] It was also true for the Huns. They had domestic slaves. Priscus saw two slaves who had killed their masters and were caught;[1160] the plural indicates that they were the only slaves of their \nasters. At Attila’s court Priscus recognized the captives from their ragged clothes and the squalor of their heads.[1161] Presumably they were held for eventual ransom. Once they became members of a Hunnic household they apparently were well treated. The captives from Aquileia took part in religious ceremonies; those kept by Christian subjects of the Huns were baptized.

V. Warfare

General Characteristics

In the seventy years between the first clash of the marauders with Roman frontier troops and the battle at the locus Mauriacus, the warfare of the Huns remained essentially the same. Attila’s horsemen were still the same mounted archers who in the 380’s had ridden down the Vardar Valley and followed the standards of Theodosius. Their tactics were determined by the weapons they carried, and as these did not change, the Huns fought at Metz and Orleans as they had fought at Pollentia. It is true that in Attila’s army there were men who could build and serve siege engines,[1162] clearly not Huns but Roman prisoners or deserters. Unlike Alaric, who boasted that Thrace forged him spears, swords, and helmets,[1163] Attila had no Roman fabricae work for him. But at least some Huns, like the Goths in 376, must have “plundered the dead bodies and armed themselves in Roman equipment,”[1164] and others may have fought with Persian weapons. But all this has little significance. Had Priscus in the 470’s described the weapons and tactics of the Huns, he would have written more or less as Ammianus Marcellinus wrote in 392:

When provoked they sometimes fight *singly but they enter the battle in tactical formation,[1165] while their medley of voices makes a savage noise. And as they are lightly equipped for swift motion, and unexpected in action, they purposely divide suddenly in scattered bands[1166] and attack, rushing about in disorder here and there, dealing terrific slaughter; and because of their extraordinary rapidity of movement, they cannot be discerned when they break into a rampart or pillage an enemy’s camp. And on this account you would not hesitate to call them the most terrible of all warriors, because they fight from a distance with missiles having sharp bone, instead of their usual points, joined to the shafts with wonderful skill; then they gallop over the intervening spaces[1167] and fight hand to hand with swords regardless of their own lives; and while the enemy are guarding against wounds from sword-thrusts, they throw strips of cloth plaited into nooses over their opponents and so entangle them that they fetter their limbs and take from them the power of riding or walking.

The Goths from whom Ammianus gathered his information were even after so many years still deafened by the wild howls[1168] of the Huns and dazed by the incredible speed of their attacks. About the social and political structure of the Huns the Goths knew next to nothing. They could not fail to notice that the Huns formed cunei but whether these consisted of the members of one clan or tribe,[1169] or were formed ad hoc, they could not tell Ammianus. From a passage in Procopius it appears that in the initial phase of a battle hereditary privileges played some role with the later Huns.[1170] The same may well have been true for their predecessors; the fur probably handed down their rank for generations.[1171] Strangely, Ammianus did not mention the feigned flight, a stratagem of the Huns as of all steppe warriors.[1172] Still, incomplete as his description is, it shows that the tactics of the Huns was not markedly different from those of the other mounted bowmen of northern Eurasia. The volleys of arrows with which the enemy was showered were followed by hand-to-hand fighting.

I pass over the “war crimes” of which the Huns were so often accused. In an apocalypsis of the seventh century, a Syriac cleric let his fancy run a little too wild: the Huns (he probably meant the Hephthalites) roast pregnant women, cut out the fetus, put it in a dish, pour water over it, and dip their weapons into the brew; they eat the flesh of children and drink the blood of women.[1173] Most Germans of the Folkwandering period behaved in no way more humanely than the Huns. In 406, the Germanic invaders of Gaul killed the hermits, burned the priests alive, raped the nuns, devastated the vineyards, and cut down the olive trees.[1174]


The Huns “are almost glued to their horses, which are hardy, it is true, but ugly, and sometimes they sit on them woman-fashion, and thus perform their ordinary tasks. And when deliberations are called for about weighty matters, they all consult for a common object in that fashion” (Ammianus, XXXI, 2, 6).

The Huns, indeed, carried on their negotiations with the Roman diplomats on horseback.[1175] The Sarmatians in South Russia and the Lazi in the Caucasus often rode side saddle also.[1176]

The characterization of the Hun horses as deformes is too vague to draw conclusions from it.[1177] To a Roman most steppe horses must have looked as misshapen as the horses of the Scythians, with their short legs and big heads,[1178] or those of the Sigynnae, shaggy and snub-nosed, allegedly too small to ride upon.[1179]

The only author to give a good description of the Hun horse is Vegetius. For a long time, he complains in the prologue to the second book of his Mulomedicina, veterinary medicine has been steadily declining. Horse doctors are so poorly paid that no one devotes himself any longer to a proper study of veterinary medicine. Of late, however, following the example set by the Huns and other barbarians, people have altogether ceased to consult veterinarians. They leave the horses on the pasture the year round and give them no care whatever, not realizing what incalculable harm they thereby do themselves. These people overlook that the horses of the barbarians are quite different from Roman horses. Hardy creatures, accustomed to cold and frost, the horses of the barbarians need neither stables nor medical care. The Roman horse is of a much more delicate constitution; unless it has good shelter and a warm stable, it will catch one illness after another.[1180] Although Vegetius stresses the superiority of the Roman horse, its intelligence, docility, and noble character, he concedes that the Hun horse has its good points. Like the Persian, Epirotic, and Sicilian horses it lives long.[1181] In the classification of various breeds according to their fitness for war, Vegetius gives the Hun horse the first place because of its patience, perseverance, and its capacity to endure cold and hunger.[1182] As his description shows, Vegetius, who probably kept a few Hun horses himself,[1183] had ample opportunity to observe them. They have, he says, great hooked heads, protruding eyes, narrow nostrils, broad jaws, strong and stiff necks, manes hanging below the knees, overlarge ribs, curved backs, bushy tails, cannon bones of great strength, small pasterns, wide-spreading hooves, hollow loins; their bodies are angular, with no fat on the rump or the muscles of the back, their stature inclining to length rather than to height, the belly drawn, the bones huge. The very thinness of these horses is pleasing, and there is beauty even in their ugliness. Vegetius adds that they are quiet and sensible and bear wounds well.[1184]

Although this description, in spite of its preciseness, does not permit a determination of the type of the Hun horse, it clearly precludes the Prze- walsky horse, which has an upright mane and a turniplike tail with short hair, only the end part being longhaired.[1185] A bronze plaque from the Or- dos region (fig. 1) shows a warrior with a pointed cap and a small bow on a horse with a “hooked” head and a long bushy tail.[1186] The man might be a Hsiung-nu. Another bronze from the river Yar in the former gu- bernie Tomsk, looks very much like the Ordos horse.[1187] It seems, however, that the Hsiung-nu had horses of various breeds,[1188] among them one with an upright mane,[1189] the opposite of the Hun horse with the mane “hanging below the knees.” The “typical” Hun horse may have been not much different from some of the Hsiung-nu and the Scythian horses.

Fig. 1. A horse with a “hooked” head and bushy tail represented on a bronze plaque from the Ordos region. From Egami 1948, pl. 4.

The Huns were superior horsemen. Sidonius compared them with centaurs: “Scarce had the infant learnt to stand without his mother’s aid when a horse takes him on his back. You would think that the limbs of man and horse were born together, so firmly does the rider always stick to the horse; any other folk is carried on horseback, this folk lives there.”[1190] The horsemanship of the Huns and Alans was unsurpassed.[1191]

As the Huns had no spurs, they had to urge the horses to a faster pace by using whips, handles of which were found in many graves.[1192] So far no stirrups have been found which could be assigned to the Huns. If the Huns had them, they must have been of perishable material, wood or leather. A potent argument against the assumption that the Huns had stirrups is the fact that the Germanic horsemen rode without them for centuries after the fall of Attila’s kingdom. Unlike the composite bow, leather or wooden stirrups could have been easily copied. But the specific factor that gave the Hun archers an advantage even over the best troops in the Roman armies may have been the stirrup. Laszlo rightly stresses the stability which stirrups give to the mounted bowmen.[1193]

“The soldiers of Rome,” wrote Jerome in the summer of 396, “conquerors and lords of the world, now are conquered by those, tremble and shrink in fear at the sight of those who cannot walk on foot and think themselves as good as dead if once they reach the ground.”[1194] Jerome’s odd description of the Huns was not based on observation; he never had seen a Hun. Like Eunapius[1195] who too maintained that the Huns could not “stand firmly on the ground,” Jerome copied Ammianus, who wrote: “Their shoes are formed upon no last, and so prevent their walking with free steps. For this reason they are not adapted to battles on foot.”[1196]

Ammianus’ explanation of the peculiar gait of the Hun horsemen when they dismounted and walked is naive. All equestrian nomads who spend a great part of their lives on horseback walk clumsily.[1197] And yet the Hun shoes must have struck Ammianus’ Gothic informants as strange, very different from their own. Apparently these shoes were fitted to the specific needs of the horsemen. So were those of the Magyars in the tenth century. Their soles were soft and pliable, so that the shoes could be slipped into the nearly round W’ooden and iron stirrups and be held firmly on them.[1198] The stirrups from the Korean tombs of the fifth and sixth centuries were likewise round. Some of them were of iron,[1199] but the most sumptuous ones, overlaid with gold, were made of wood. The gold shoes in these tombs are evidently replicas of leather shoes.[1200] Had Ammianus seen them, he probably would have called them formulis nullis aptati.

The problem of the origin of the metal stirrup is still unsolved.[1201] If the stray finds of miniature metal stirrups in the Minusinsk area could indeed be dated to the Syr or Uibat period (first three centuries a.d.) or even to the Syr period (first and second centuries a.d.),[1202] they would be the earliest stirrups so far known, but their date is controversial.[1203] No riders with stirrups are found in the numerous representations of northern barbarians in Chinese art of the Han period; the horsemen on the gold plaques in the collection of Peter I are not using stirrups either.


After his defeat at the locus Mauriacus, Attila “shut himself and his companions within the barriers of the camp, which he fortified with wagons. But it was said that the king remained supremely brave even in this extremity and had heaped up a funeral pyre of horse saddles \equinis sellis construxisse pyram] so that if the enemy should attack him, he was determined to cast himself into the flames, that none might have the joy of wounding him.”[1204]

This passage in the Getica has often been adduced[1205] as proof that the Huns had wooden saddles. But shabracks could have burned as well. The history of the saddle of the Eurasian nomads is anything but clear. In the third and fifth kurgans at Pazyryk and at Shibe in the High Altai rather primitive saddles were found. They consisted of two big leather pillows, stuffed with deer hair and covered with felt; small pillows at the front and back of the big ones were stiffened and strengthened with narrow wooden frames, the forerunners of the wooden saddle bows.[1206] To judge from the representations,[1207] the Scythian saddles were like those of the Altaians. The same is true for the saddles of two bronze horsemen from western Siberia, probably of the same date as the Pazyryk kurgans, and on an often reproduced later golden belt buckle in the Siberian Treasure of Peter the Great.[1208]

On the other hand, what looks like the wooden front bow of a miniature saddle from the Uibat chaatas in the Minusinsk region,[1209] datable to the beginning of our era, was possibly part of a true wooden saddle.[1210] The fragments of a saddle of about the same date from the Karakol River, not far from Shibe, might also come from a saddle with a tree between the two bows.[1211] In Kenkol, Bernshtam found a curved piece of wood which could be the bow of a saddle.[1212]

The Chinese of the Han period had wooden saddles. Although most representations of the horsemen do not permit one to decide whether they rode on saddles, the riderless horses on some reliefs from Shantung undoubtedly were saddled; the front and back bows and the saddle tree are clearly delineated.[1213] It has become almost a dogma to derive everything in the equipment of the cavalry of the Chinese from their barbarian neighbors. One should, therefore, expect that the Hsiung-nu and other nomads on the Chinese frontiers not only had saddles, but that they had them before the Chinese. The archaeological evidence does not bear out such an assumption. The barbarian hunters on the stamped tiles from Loyang have only saddle pads.[1214] In the Hsiung-nu graves at Noin Ula, wooden frames of the front and back pillows were found; they show that the Hsiung-nu at the beginning of our era had shabracks like the people in Pazyryk.[1215] Whether later the Hsiung-nu rode on wooden saddles we do not know. The Koreans of the fifth century did. A number of front and back bridges exist, made of gilt bronze and even of iron.[1216]

The literary evidence, one short passage in Jordanes, is ambiguous and the little we know from the earlier finds in the eastern steppes sheds only a dim light on the saddles of the Huns. However, the gold, silver, and bronze mountings of saddle bows in nomadic graves of the fourth and fifth centuries leave no doubt that the Hun saddle did consist of a wooden tree, with a straight vertical bow in front and a somewhat larger inclined bow in the back. Such mountings were found in the Hun heartland in Hungary, where they were unknown before the coming of the Huns, and in the steppes as far east as the Volga; silver sheet mountings from the front bow of a wooden saddle were found in the grave of a Germanic warrior at Blucina near Brno in Moravia; one find was made in Borovoe in Kazakhstan.[1217] Ten out of the thirteen mountings were decorated with a scale pattern, impressively showing that they belong to the same group. The Hun saddles were presumably similar to the wooden saddle from Bo- rodaevka (formerly Boaro), Marks, Saratov, on the right bank of the Big Karaman River.[1218] It lay in a grave of a man with an artificially deformed head. The burial rite (a horse skull and four feet cut off above the hoofs) and the furniture are characteristic for the post-Hunnic burials in the Volga steppes, preserving many elements of late Sarmatian civilization, datable between the fifth and the seventh centuries.[1219] The saddle from Borodaevka is similar to some Sasanian saddles.[1220]

Horse Marks

Where horses, owned by several families, clans, or tribes, graze over a common pasture, they are marked, either by cuts in the ears or by burning the hindquarters or shoulders with a hot iron. The former, more primitive, method is attested as early as the fourth century b.c.; all horses in the first and fifth kurgans at Pazyryk in the High Altai were earmarked.[1221] Until recently the Kirghiz on the Manyshlak peninsula in the east side of the Caspian Sea used to cut triangles in the ears of their sheep and to notch the ears of their horses.[1222]

Marco Polo wrote about the Mongols:

The land is so secure that each lord or the other men who have animals in plenty, have them marked with their seal stamped on the hair, that is, the horses and the mares and camels and oxen and cows and other large beasts; then he lets them go safely to graze anywhere over the plains and over the mountains without a watchman; and if on Lheir return they are mixed the one with the other, each man who finds them recognizes the owner’s mark and immediately takes pains to inquire for him and quickly gives back his own to him whose mark is found. And in this way each finds his own animals.[1223]

Two horses of the K’itan on a painting by Hu Kuei have on their hindquarters tamgas[1224] I use this Turkish word for “seal, property mark,” because it was borrowed not only by the Mongols, but also by the Tadzhiks, Persians, and even the Russians,[1225] although the Persians branded their horses before they had any contact with Turks, and the Russians had their own, Slavic word for brandmark (pyatnd). It may have been the technically superior form or application of the branding iron of the Turks that superseded the earlier methods of branding, and therefore the older word. The Tang hui yao, chapter 72, contains a list of the tamgas of thirty-seven, mostly Turkish tribes;[1226] the foreign, again preponderantly Turkish, horses in the great pastoral inspectorates in T’ang China were in addition branded on many parts of the body, to show ownership, age, type, quality, and condition.[1227]

Turkish tribes marked their horses before[1228] and after[1229] the T’ang period. In Persia the brandmarks can be traced to the third century. A graffito in Dura-Europos shows an early Sasanian tamga.[1230] The horse of Peroz at Taq-i-Bustan has a mark on the right hindquarter,[1231] and the steeds of the Sasanian kings on a fabric in the Horyuji at Nara are branded on their flanks, but the Persian tamga has been changed into the Chinese character chi, “auspicious.”[1232]

Although the Sasanian tamga brings us closer to the Huns in time and space, the horse of a hunter on an often-reproduced mosaic from Borj Djedid, Carthage, now in the British Museum,[1233] leads to a milieu intimately associated with the Huns. The man, to judge by his dress, could be a Roman, a Vandal, or an Alan.[1234] The strange crosslike tamga on his horse has been taken for Roman.[1235] But as Janichen noticed, it has a striking resemblance to the tamga in a rock picture on the upper Yenisei,[1236] datable to the middle of the first millennium a.d. It is an Asiatic tamga; the hunter must be an Alan.

Sarmatian tamgas of the second and third centuries are well attested. There is, first, a grave stela from Theodosia in the Crimea, of a type known from many places in the Bosporan kingdom. The stela has been set up by the religious society to which the deceased Atta, son of Tryphon, belonged. In spite of his Greek name, the man is dressed like a Sarmatian horseman; he carries a Sarmatian dagger with a ring handle and his horse is marked with one of those Sarmatian signs which occur on reliefs, mirrors, cauldrons, buckles, jewelry, and coins from the Bosporan kingdom and adjacent areas in the first three centuries a.d. (fig. 2).[1237] There is, second, the fragment of a stela found at the khutor Malaya Kozyrka, north of Olbia, representing a hunting scene. The horse is marked on the flank; another Sarmatian tamga is between the front and hind legs.[1238] There is, third, the clay figure of an ox, a toy, from Glinishche near Kerch-Panticapaeum, with a Sarmatian tamga branded on the shoulder.[1239]

Fig. 2. Grave stela from Theodosia in the Crimea with the representation of the deceased mounted on a horse marked with a Sarmatian tamga, first to third centuries a.d. From Solomonik 1957, fig. 1.

The Mongols, the K’i-tan, the Turks before and after the T’ang period, and the Kirghiz north of the Sayan Mountains branded their horses like the Sarmatians in the second and third centuries and the Alans in the fifth. The Huns had large herds of horses. In the campaign of 451, Attila’s army is said to have numbered five hundred thousand men,[1240] though actually it cannot have had more than a fifth of this figure, and probably even less. A good part of the army consisted of Germans, many of whom were foot soldiers. Still, counting the reserve horses and the draught horses, Attila must have had fifty or sixty thousand horses when he set out for Gaul. The long frontiers of the loosely knit kingdom had to be guarded while the mobile army was away, and a considerable force stayed at home to keep the conquered peoples in subjection. To these war horses the mares and foals have to be added. The Huns must have had some means of identifying the owners of their horses. It is, I believe, practically certain that they branded their horses with tamgas similar to those of the Sarmatians.[1241]


All frozen horses in the kurgans at Pazyryk were castrated:[1242] The princes who were buried there rode only geldings. The same was true more than two millennia later: No well-off Altaian rode a stallion or a mare.[1243] In the 1860’s their herds consisted of 20 to 60 horses: 1 stallion, 8 to 25 mares, 5 to 15 one-year-old colts, 4 to 14 two- and three-year-old colts, and 5 to 10 geldings;[1244] the stallion colts were castrated in their second year.[1245] In the herds of the Kirghiz the relation between stallions and mares was one to nine.[1246]

“The knowledge of castration,” says Lattimore, “is essential to the technique of steppe pastoralism. Otherwise the unnecessary large number of male animals, fighting each other and attempting to lead away bands of females, would make it impossible to keep stock in large, tractable herds on unfenced pasture.”[1247] The Scythians and Sarmatians in South Russia castrated their horses “to make them easy to manage; for although the horses are small, they are exceedingly quick and hard to manage.”[1248]

In spite of the absence of any literary evidence, there can be no doubt that the Huns, too, rode mostly geldings.[1249]


The literary evidence for the wagons of the Huns is scanty: a few lines in Ammianus, a sentence in Priscus, and a subordinate clause in the Getica.

According to Ammianus (XXX, 2, 10), “no one in their country ever plows a field or touches a plow-handle. They are all without fixed abode, without hearth, or law, or settled mode of life, and keep roaming from place to place, like fugitives, accompanied by the wagons [cum carpentis} in which they live.” This is a paraphrase of the description of the Scythians in Trogus Pompeius.[1250] Ammianus uses almost the same phrases when he speaks about the incessant wandering of the Alans (XXXI, 2, 18). In their wagons the Hunnic women cohabit with their husbands, bear children, and rear them to the age of puberty; in the wagons of the Alans the males have intercourse with the women, and in the wagons their babes are born and reared. If Ammianus were to be believed, the Hunnic women even wove their garments in the wagons. But he cannot be believed. He turned into the ordinary way of Hunnic life what his informants told him about a Hun horde on the move. Besides, he followed the Greeks who were so impressed by the wagons of the Scythians that they took the vehicles, mostly used for moving the tents, for the homes of the nomads. To the Greeks the Scythians were and remained “wagon dwellers” (ap,a£6ftioi} and “men who carried their own houses with them” (g^egeoixoi),[1251] epithets endlessly repeated and occasionally embroidered by Latin authors.[1252]

Priscus mentions the wagons, a/za|at, of the Huns, on which rafts or pontoons, a^edlat, for use in marshy places were carried.[1253]

The third reference to the Hun wagons occurs in Jordanes’ account of the battle at the locus Mauriacus. In the evening of the first day Attila retreated and “shut himself and his companions within the barriers of the camp, which he had fortified with wagons [plaustris vallatum]” (Getica 210).

Although Priscus says nothing about the number of the rafts put on a wagon, their size, and the material of which they were made—it could have been wood, wickerwork, or hides—the wagons were probably heavy four-wheeled vehicles. The Huns could not have encumbered their swiftmoving cavalry with such carts. Those in Attila’s army must have been light, probably two-wheeled wains. In the fourth and fifth centuries camps with a defensive barrier of wagons were nothing specifically Hunnic. “All the barbarians,” wrote Vegetius, “arrange their carts around them in a circle and then pass their nights secure from surprise.”[1254] Like other Germans of their time and before it,[1255] the Goths formed carragines[1256] with great skill.

Although there is no archaeological evidence for the wagons of the Huns, we can form an approximate picture of them from the finds in the graves of other northern barbarians, who put the dismantled or broken funeral cart, or parts of it, into the pits or catacombs. Fragments of such carts were found in Scythian kurgans from the sixth to the third century b.c. in the Kuban, Taman, Dnieper, and Poltava groups,[1257] in four of the five burial mounds at Pazyryk in the High Altai,[1258] in Sarmatian graves from the fourth to the first century b.c.,[1259] and in the Hsiung-nu graves at Noin Ula.” Some Sarmatian carts were light vehicles with two or four wheels. The wheels in kurgan 12, grave 9, at Politotdel’skoe on the lower Volga measured 1.2 meters in diameter and had at least twenty spokes.[1260] The wheels of the impressive four-wheeled cart in the fifth kurgan at Pazyryk, each with thirty-four spokes, had a diameter of about 1.5 meters; there was a raised seat for the driver, and a superstructure covered with black felt, decorated with stuffed felt swans.[1261] The absence of metal parts indicates that the big cart was of local provenance, though possibly made in imitation of Chinese wagons. The Kao-chii tribes, the later Uigurs, had wagons with very high wheels; the Chinese named the people after them: kao chii means “high chariots.” The various names under which the Kao-chii were known before, ti-li, t’e-le, t’it-te, and ting-ling,[1262] are possibly variants of a Turkish word for wheel.[1263]

Fig. 3. Two-wheeled cart represented on a bronze plaque from the Wu- huan cemetery at Hsi-ch’a-kou. From Sun Shou-tao 1960, fig. 17.

There exist a few representations of carts of the eastern barbarians: a two-wheeled cart on a bronze plaque from the Wu-huan cemetery at Hsi-ch’a-kou (fig. 3),[1264] another one on a Chinese incense burner of late Chou or early Han date in the Freer Gallery.[1265] A bronze plaque from Sui- yiian shows a man in a long coat and wide trousers, holding a sword with a ring handle, in front of a car drawn by three horses (fig. 4).[1266] The two heads on the cart are not the cut-off heads of enemies[1267] but are meant to represent people in a small tent. The miniatures in the Radziwil manuscript show the wagons of the Kumans with the same heads in the tents mounted on the vehicles (figs. 5A and 5B).[1268]

Fig. 4. Bronze plaque from Sui-yiian with the representation of a man holding a sword with a ring handle before a cart drawn by three horses. From Rostovtsev 1929, pl. XI, 56.
Fig. 5A. Miniature painting from the Radziwil manuscript showing the wagons of the Kumans. From Pletneva 1958, fig. 25.
Fig. 5B. Miniature painting from the Radziwil manuscript showing human heads in tents mounted on carts. From Pletneva 1958, fig. 26.

Even without the specific statements in the cited passages it would have to be assumed that the Huns had wagons. They broke off the pursuit of the Goths because they were “loaded down with booty.”[1269] The Huns must have had wagons like those of Alaric’s Visigoths in Italy which were loaded with precious stuff, such as mixing bowls from Argos and lifelike statues from Corinth.[1270] On their migration to the Don and from the Don to the Danube, the Huns probably transported their old people, women, and children in wagons.[1271] Toy wagons found in Kerch show what the wagons of the later Sarmatians looked like. Some of them have pyramidal towers, doubtless movable tents;[1272] others are heavy four-wheeled vehicles (fig. 6).[1273]

Fig. 6. Ceramic toy from Kerch showing a wagon of Late Sarmatian type. From Narysy starodav’noi istorii Ukrains’koi RSR 1957, 237.

The wagons of the Huns must have been similar to the toy wagons from Panticapaeum.

Horses played a prominent role in the economy of the Huns. Although our authorities do not mention that the Huns ate horse meat—perhaps because this went without saying[1274]— they certainly did, like the Scythians,[1275] Sarmatians,[1276] and all other steppe peoples. The meat was boiled in large cauldrons[1277] and fished out with iron hooks. The Scythians were IziTurjpolyol and yaXawrorpayoi’, the Alans “lived on an abundance of milk.[1278] There can be no doubt that the Huns, too, drank mare’s milk and made kumys and cheese.[1279]

Claudian and Sidonius at times named the Geloni where we would expect the Huns. In addition to the reasons adduced in another context, Claudian and Sidonius may have thought of some epithets of the Geloni, like sagittiferi[1280] or volucres,[1281] which also fit the new barbarians whose hated name could, therefore, be exchanged for one almost consecrated by the great poets of the past. Sidonius may have had a verse of Virgil[1282] in mind when he associated the equimulgae Geloni with the Sygambri and Alans of his time.[1283] Like the Massagetae, the Geloni were said to have mixed milk and horse blood.[1284] Perhaps by substituting Geloni for Huns, the poets[1285] indicated that the Huns, too, drank the blood of their horses. When Ennodius ascribed this custom to the Bulgars,[1286] he could have followed a topos. But neither Marco Polo[1287] nor Hans Schiltberger[1288] thought of Virgil when they described how the Mongols and the Tatars of the Golden Horde bled their horses and boiled and ate the blood when they had nothing else to eat.

Bows and Arrows

“A wondrous thing,” wrote Jordanes, “took place in connection with Attila’s death. For in a dream some god stood at the side of Marcian, emperor of the East, while he was disquieted about his fierce foe, and showed him the bow of Attila broken in the same night, as if to intimate that this race owed much to that weapon [quasi quod gens ipsa eo telo mul- tum praesumat].”[1289]

The bow was the weapon of the Huns. In Ammianus’ description of their armament, bow and arrow take the first place.[1290] Olympiodorus praised the skill of the Hunnic leaders in shooting with the bow.[1291] Aetius, who got his military education with the Huns, was “a very practiced horseman and skillful archer.”[1292] Shapely bows and arrows, said Sidonius Apollinaris, were the delight of the Huns; they were the best archers.[1293] He found no higher praise for Avitus’ bowmanship than by saying that he even surpassed the Huns.[1294] In the battle on the Nedao the Huns fought with bows and arrows.[1295]

A century later, after the East Romans had taken over so many of the weapons and tactics of the barbarians, they were “expert horsemen, and able without difficulty to direct their bows to either side while riding at full speed, and to shoot at opponents whether in pursuit or in flight.”[1296] And yet Belisars’ Massagets,[1297] that is, Huns, were still the best bowmen. Even dismounted and running at great speed, they “knew how to shoot with the greatest accuracy.”[1298]

Although Ammianus had the highest respect for the Hunnic bow, he was not well informed about it. The Huns could, he said, easily be called the fiercest of all warriors, because they fight from a distance with missiles having sharp bone points instead of the ordinary points, joined to shafts with wonderful skill. Why the bone points should have turned the Huns into such superior archers is by no means clear. From Ammianus’ assertion that the Scythian and Parthian bows are the only ones that have a straight rounded grip[1299] it would follow that the Hunnic bow was bent in a continuous curve, which is contradicted by the archaeological evidence. As so often, a single find tells us more than all the written sources.

As early as 1932, when only a few finds were known, Alfoldi and Werner were able to reconstruct the Hunnic bow.[1300] Their results are by now generally accepted, but in the past thirty years the material has grown immensely.[1301] New problems have arisen. Unexpected finds reopen questions which seemed to be answered definitely. In a way, the history of Eurasia sep- tentrionalis antiqua runs parallel to the history of the Hunnic bow.

It is a reflexed composite bow, 140–160 centimeters in length. Its wooden core is backed by sinews and bellied with horn. What distinguishes it from other composite bows are the seven bone plaques which stiffen the ears and the handle, a pair on each ear and three on the handle, two on its sides and one on its top. The string is permanently made fast to the end of the bow, which is stiffened for the greatest length; the nock is square, or almost square; in the finds it shows little evidence of rubbing. The nock in the ear of the shorter, more flexible arm is round; the string is looped into it when the bow is strung. In the finds it is much worn. This bow was spread from the British Isles to northern China. The earliest known bone strips come from graves of the fourth or third centuries b.c. The Russians used such bows as late as the twelfth century.[1302]

Before attacking the specific problems which the Hunnic bow poses, some preliminary remarks and general considerations seem to be in order. The lack of a generally agreed on terminology in the study of the bow sometimes results in an annoying confusion.[1303] I will use the following terms:

Self bow: the plain wooden bow in one piece.

Reflexed bow: a bow which, when unstrung, reverses its curve.

Compound bow: a bow built up by uniting two or more staves of similar material.

Reinforced bow: a bow with a layer of longitudinally disposed sinew applied to the back.[1304]

Composite bow: a bow whose stave embodies a laminated construction involving more than one type of material, such as wood, sinew, and horn; as a rule the wooden core is backed by sinew and bellied by horn.[1305]

Handle or grip: the space occupied by the hand in holding the bow.

Arms: the regions between the handle and the tip.

Nocks: the depressions or notches on the ear which serve to keep the string from slipping.

Ear: the part of the arm with the nock.

Back: the side of the bow away from the string; the concave side when the bow is strung.

Belly: the side of the bow next to the string; the convex side when the bow is strung.

Bracing: setting the string tight on the bow.

Length: the distance from tip to tip before the bow is strung.

Span: the distance from tip to tip when the bow is strung.

There are bows which do not fit these definitions. The English longbow, for example, is a self bow but also a variety of the compound bow. “In making a yew bow, the wood that is used is that which is nearest the outside of the log, consisting of practically all the light-colored sapwood immediately under the bark and only as much of the darker heartwood as may be needed. This combination of sap and heartwood in yew provides the two properties required, for the sapwood is resistant to stretch and therefore suitable for the back, and the heartwood resists compression and is therefore perfect for the belly.”[1306]


Representations of bows in paintings, reliefs, metalwork, and on coins are in general of limited value for determining their anatomy. Doublecurved bows are not necessarily composite bows. Those in Attic Geometric art, for instance, their inward curve reaching almost to the string when they were strung, were self bows, made entirely of wood.[1307] Brown’s in some respects very valuable study[1308] is in others misleading because he deals almost exclusively with representations. After having established the shape of a bone-stiffened bow at Yrzi in a necropolis on the Euphrates about 40 kilometers southeast of Dura-Europos, Brown looked for more bows like it. As could be expected, he found them nearly everywhere. The bows of the royal guard on the tile reliefs in Susa and on Chinese vases of the Han period looked to him like “Yrzi” bows, hence they were “Yrzi” bows. But Achaemenid findspots, in particular the arsenal at Persepolis where they should have been found by the hundreds, yielded not a single bone strip,[1309] and the same is true for the Han graves. Emeneau collected an impressive number of representations in early Indian art;[1310] so did Au- boyer.[1311] But whereas Auboyer, wisely in my opinion, merely classified them according to their curves,[1312] Emeneau drew from the monuments conclusions as to the structure of the bows. In some cases they may be right, but there is no archaeological evidence to bear them out.

Had the bow which Stein found in the Tibetan T’ang fortress at Ma- zardagh[1313] occurred in a wall painting, it easily could have been taken for a bone-stiffened bow. The gently curved ears with their notches look like those of bone-stiffened bows from the Chinese borderland. Khazanov included it in his list.[1314] Evidently he did not read the text. In the dry desert bone strips would have been splendidly preserved. Stein found none. The ears, made of tamarisk wood, had no traces of glue on them. Bone plaques on the ears were sometimes painted or wrapped up in colored strings,[1315] in which cases it is impossible to recognize them in paintings.

There are, however, representations of bows which can be of help in determining some of the anatomy of the bow. The strongly curled ends of the Scythian bow preclude the application of bone strips on the ears. There is, furthermore, a type of Sasanian bow with very long ears. They must have been stiffened with bone plates, otherwise they could not have resisted the strain when the bow was drawn; ears of plain wood have been broken in pieces. Conversely, such very long bone strips in graves prove that the bow on which they were applied was of this Sasanian type.

Coexistence of Various Types

The bow of the Huns discussed on the following pages was a war bow, as presumably most bows in burials were; they lay in the graves of warriors. About the hunting bows of the Huns we have no information, but that they were different from the war bow is practically certain. The Huns could not have hunted ducks or foxes with their precious composite bows. War bows and hunting bows were often as different as rifles and shotguns. On a stela in already strongly Sarmatized Panticapaeum, a young man is drawing a long C-bow[1316]; behind him stand his groom and his horse, which is neither bridled nor saddled; it is a peaceful scene.[1317] The bow is a hunting bow. In the battle scenes in the Stassov catacomb in Panticapaeum[1318] and the stelae with the likeness of the dead as warrior, the bows are short and double curved, of the Scythian type. Unless one keeps such differences in mind, one can easily draw wrong conclusions from one-sided evidence. On stamped tiles from Old Loyang, probably made from stamps designed in the third century b.c.,[1319] occur hunters, Hsiung-nu or people closely related to them, chasing deer. The bows in these pictures tell us little about the war bows of the nomads.

The Skill Required

Composite war bows technically as perfect as those of the Huns could only be made by professional bowyers. They must have had workshops like those in the Roman fort at Carleon[1320] and Parthian Merv.[1321] The making of even such a simple bow as the English longbow required a good deal of craftmanship. It had to be tapered correctly, with patience and care, from the middle toward each end to bring it to an even curve when full drawn; all knots and irregularities in the grain had to be carefully watched and “raised” or followed skillfully to eliminate weak spots.[1322] For a detailed description, the chapter on “Making the Bow” in Pope’s classical Hunting with Bow and Arrow[1323] should be read. “While the actual work of making bows,” he wrote, “takes about eight days, it requires months to get one adjusted so that it is good.” Turkish manuals on archery contain the names of outstanding bowyers, and there exist long lists of Japanese bowyers, who wrote their names and the date on their bows. Elmer, one of the greatest experts on archery, wrote: “I know of only three men of our race who had been successful in making one or more composite bows, though none of them has produced a weapon which could vie in quality with the best products of the ancient Orient. All started with the slogan ‘A white man can do anything a brown man can,’ but none has seen his boast fulfilled.”[1324] Luschan estimated that the time required for making a good Turkish bow, including the intervals of drying and seasoning between operations, was from five to ten years.[1325] These were, of course, particularly well made bows, mostly used for flight shooting. But the ordinary bows also required a high degree of skill and thorough familiarity with all details. In 1929 old men in the Barlyq-Alash-Aksu region in western Tuva told me that in their youth, in the seventies and eighties, there were only two men in their khoshuns who could make bows. To find the appropriate materials, to cut the wood, horn, and bone into the right shape, to mold the sinews for the back, to determine the best proportions between the weak and rigid parts of the bow, all this and much more presupposed long training. The idea that each Hunnic archer could make his own bow could have been conceived only by cabinet scholars who never held a composite bow in their hands.

Such bows were not easily replaced; once they were broken, they could not be repaired. This explains the character of the finds. Mere lists of findspots give a distorted picture. One has to go through the reports carefully to realize what the bows meant for their owners. Whenever a report is sufficiently detailed, it invariably turns out that the set of bone plaques is incomplete: one, two, or three instead of four from the ears, or only one plaque instead of three on the handle. The only complete set known to me, all the nine plaques of the bow of the latest Sarmatians found by Sinitsyn at Avilov’s Farm, comes from a damaged bow.[1326] Marmots dislocated the skull of the dead man; they damaged the quiver of birch bark. The bow lay in situ, and yet not a single bone plaque was intact. The more interesting cases are those in which the plaques do not belong together.

Werner mentions the plaques in the rich grave at Blufcina in Moravia, where a Germanic nobleman was buried shortly after the collapse of Attila’s kingdom; Khazanov refers to Werner, and Tihelka in his report gives them a few lines;[1327] fortunately he also brings drawings, which allow a closer study. The plaques do not fit one bow. Two fragments of endpieces are of different width and have notches on the same side; they cannot have formed a pair. Two long strips cannot have come from one pair either; one is almost straight, the other markedly curved. The grave was not disturbed. The strange ensemble admits only one explanation: They are the broken part of two or, perhaps, three bows.

A find from Ak-Tobe in the Tashkent oasis throws more light on the reluctance to put an intact bow into the grave. In a burial dated to the end of the fourth century, the excavators found what they took to be a bow in situ. But the two long bone strips come from two bows. One has a round, the other one a triangluar notch; they are differently curved.[1328] The people buried the dead warrior with a sham bow.

The difficulty of making a bow like that of the Huns is indirectly proved by the inability of the Germanic tribes to produce one.[1329] The Gepids for many years had lived under and together with the Huns in Hungary. They buried their dead, even after their conversion to Christianity, with weapons. The graves contain swords, daggers, armor, helmets, umbones, arrow points, but not one bone strip.[1330] Though the Goths had archers,[1331] they never learned to shoot from horseback.[1332] “Practically all the Romans and their allies, the Huns,” Procopius wrote, “were good mounted bowmen, but not one among the Goths had any practice in this branch. Their bowmen entered battle on foot and under the cover of heavy-armed men.”[1333] The Goths in Italy were excellent riders[1334] but unable to emulate the Huns because they had no bows like the Huns.[1335]

As Alfoldi recognized first,[1336] the Hunnic bow had limbs of uneven length. It would not be worthwhile to mention it again if it were not for the insistence of some scholars on the inferiority of such a bow. I need not enumerate the peoples who had bows of this type. It will suffice to point to the Japanese. It is, to say the least, unlikely that they, who made the best swords in the world, should have been unable to make limbs of the same length.


Thanks to McLeod’s careful analysis of the Greek and Latin sources, the range of the ancient composite bows has been definitely established: bowmen were quite accurate up to 50 to 60 meters, their effective range extended at least 160 to 175 meters, but not as far as 350 to 450 meters.[1337]

According to a Moroccan archery manual of about 1500, “archers throughout the world agree that the limits beyond which no ...”

[The manuscript breaks off here in mid-sentence.—Ed.]

The bone strips found in and near Roman camps from Scotland to Vindonissa and Egypt[1338] show that Oriental Sagittarii[1339] used bows stiffened and reinforced like those of the Huns.[1340] When one considers how strong Parthian influence on the armament of the Palmyreans[1341] and other Syrians was and that bone lamellae began to appear only at the end of the first century b.c.,[1342] the Parthian provenance of the bows of the Eastern archers seems highly probable. Possibly some bows, or rather fragments of bone strips, found along the limes[1343] were actually Parthian; among the archers whom Severus Alexander sent from the Orient to Germania were Parthian deserters.[1344] The archers who left bone strips in a late building in Carnuntum[1345] unfortunately cannot be identified. An ear piece was nailed to the wooden core[1346] as in the camp of Bar Hill,[1347] where a cohort of archers from Emesa in Phoenicia was stationed;[1348] this could indicate that the troops in Carnuntum were Orientals, but this unusual way of attaching the strip occurs also in Avar and Hsiung-nu graves.[1349] Alfoldi is inclined to take the archers of Carnuntum for Huns.[1350] The fragment of a Hunnic cauldron in Aquincum (Budapest) seems to support his suggestion, but near the cauldron fragment lay Oriental officers’ helmets.

Sasanian Bows

Sasanian bows are known only from reproductions. The most common type, to be seen on numerous silver plates, has the long ears sharply set off the arms, exactly as on the wall paintings from Dura-Europos (figs. 7 and 8). The Sasanians took it over from the Parthians. Assuming that the handle is about 15 to 16 centimeters (the hand’s width with one or two centimeters on each side), the length of the hunting bow, measured along the curve, varies from 70 to 110 or 115 centimeters. The latter figure is possibly an exaggeration: The great size of the bow corresponds to the superhuman height of the royal hunter. If the king is not on horseback but stands in a boat, as on the reliefs at Taq-i-Bustan,[1351] his bow is also extremely long.

Fig. 7. Detail of a Sasanian-type silver plate from a private collection. Detail from Ghirshman 1962, fig. 314.
Fig. 8. Detail of a Sasanian silver plate from Sari, Archaeological Museum, Teheran. Detail from Ghirshman 1962, pl. 248.

On some plates the notch in the ear is clearly visible; on others the string just touches the bow, an indication of the craftsman’s carelessness. As is known, not a few plates are copies of older originals, and not very exact ones. Occasionally the silversmith, who may never have held a bow in his hand, made even stranger blunders. On a plate from Kulagysh[1352] the two heroes[1353] carry bows with the strings fastened to loops on the belly. In most cases it cannot be decided whether the notch was cut in the wood or in the bone strip. There are, however, silver plates on which the bow has strings tied around the ear (fig. 7).[1354] This would be superfluous had the nock been cut in the wood but makes sense as a means to hold the bone strips and the wood between them firmly together. From the fact that some bone strips from Carnuntum are roughened on the surface, Werner concluded that they were wrapped around with strings.[1355] As in similar cases, the strings were probably colored. Incidentally, this shows that the archer held the bow with the longer end up, the one to which the bowstring was permanently made fast.

The long ears are additional proof that the bows were stiffened with bone or horn. Unless the ears were encased in bone strips, they could not possibly be so rigid.[1356] Sidonius Apollinaris had such bows in mind when, in the Panegyric on Emperor Anthemius, he wrote, “In boyhood it was his sport to handle eagerly arrows that had been seized from the foe, and on captive bows to force the resisting strings on to the curving horns.”[1357] Procopius, young Anthemius’ father, fought against the Persians in 422.[1358]

If the silver plates could be dated more exactly, it should be possible to follow the development of the Sasanian bow, or, better, bows, for it is unlikely that they were all of the same type, from Egypt to Afghanistan. The plate from Kulagysh is of Sogdian origin.[1359] So is probably the often reproduced plate with the lion hunter, whose stirrups point to post-Sa- sanian times (fig. 9).[1360] The Sogdians fought with weapons identical with or very similar to those of the Sasanian Persians. From the war bows on Sogdian plates we may conclude that the Sasanian war bows were the same, though possibly of slightly different size, as the bone-stiffened hunting bows.

There were others. On the plate from Akinovo the warriors defending the fortress carry M bows,[1361] the same as the bow depicted on a vase from Merv, datable to the fifth century.[1362] In a battle scene on a weave from Arsinoe in Egypt of about 600 a.d., both foot soldiers and horse-archers carry bows with strongly curled ends,[1363] very similar to the Scythian and Scythian type bows of the Parthians.

The term Sasanian bow is, strictly, a misnomer, for the same type, the bow with the long ears set off at an angle, occurred also outside Sasanian Persia and before the Sasanids. Yet the term is so commonly used and so convenient that I will retain it, with the understanding that the “Sa- sanian” bow is not exclusively Sasanian but designates only a specific type. To deal with all “Sasanian” bows, from India[1364] to southern Siberia[1365] and Chinese Turkestan,[1366] would lead us too far away from the Huns. Why, for instance, Virudhaka on a relief from the Silla kingdom in Korea holds a “Sasanian” bow[1367] is a question for historians of Far Eastern art to answer.

Fig. 9. Silver plate from Kulagysh in the Hermitage Museum, Leningrad. From SPA, pl. 217.


[The section on swords is missing in the manuscript, except for the following fragments.—Ed.]

The Sword of Altlussheim

The scabbard tip of the much-discussed sword from Altlussheim near Mainz{2} (fig. 10A)[1368] (fig. 10A)[1369] was, as Werner proved, originally an attachable sword guard, comparable to the guards of the Chinese swords of the Han period. Because of its shape and material, Werner takes it to be of Sasanian or Hephthalite provenance (fig. 10B).[1370]

Fig. 10A. Scabbard tip of a sword from Altlussheim near Mainz. From J. Werner 1956, pl. 58:4.

He refers to the representation of a sword on a relief from Palmyra (fig. Il);[1371] the lower edge of its guard has the same obtuse angle as that of the piece from Altlussheim. As so much in the armament of the Pal- myreans, the sword is supposed to be either of Persian provenance or made in imitation of a Persian sword. The guard, so incongruously fixed to the shape of the sword from the Rhine, is cut from a piece of lapis lazuli. This semiprecious stone is said to be mined only in the Badakhshan Mountains in Afghanistan, an area which until the middle of the fifth century was a part of the Sasanian empire, later lost to the Hephthalites.

[[o-j-otto-john-maenchen-helfen-the-world-of-the-hun-15.jpg][Fig. 10B. Detail of the sword from Altlussheim. From J. Werner 1956, pl. 38 A.

Fig. 11. Stone relief from Palmyra, datable to the third century a.d. Ghirshman 1962, pl. 91.

Werner’s argumentation is ingenious but inconclusive, for several reasons. First, the Palmyrean sword, if it should be of Persian origin, would go back not to a Sasanian but a Parthian prototype. Maqqai, on whose triclinium the sword is represented, died in 229, only one year after the collapse of the Parthian kingdom. Second, no sword guard of that type is known to have existed in Sasanian or, for that matter, in Parthian Persia. Third, if we apply the ratio between the guard and the neck or arm of the warrior in the Palmyrean relief to life proportions, the guard must have been, at least, 30 centimeters wide, that is, three times the width of the Altlussheim guard. It is almost certain that a guard of the size of the Palmyrean one could not have been detachable but must have formed a part of the sword, cast or forged together with the blade and the handle. Fourth, there is no indication of the saddle between the shoulders, which the lapis lazuli guard, though rather battered, shows quite clearly. Fifth, it is true that the source of lapis lazuli has long been the Kokcha Valley of Badakhshan, but it was not always the only one. Darius I got lapis lazuli for the building of the apadana at Susa from Sogdiana,[1372] where the stone was still mined in Marco Polo’s days.[1373] Besides, lapis lazuli was worked by craftsmen from Egypt to China. The Chinese imported it via Kashgar and Khotan as early as the second century b.c.;[1374] they may have obtained the se-se[1375] either directly from Afghanistan or from Persia, where lapis lazuli was widely used by the Parthians.[1376]

Although the piece of lapis lazuli from which the guard was cut may have come from an outlying province of the Sasanian kingdom, the guard itself shows Chinese workmanship. Such sword guards, with the characteristic saddle between the shoulders, cast of bronze,[1377] carved out of jade, made of glass, or cast together with the handle and the blade, often inlaid with turquoise,[1378] are among the most common objects found in Han tombs. Agate was another material used for the decoration of swords. The Chinese cut pommels of bluish and reddish agate[1379] even before the Han period.[1380] A sword guard of agate, of exactly the shape and size of the guard from the Bhine, was found in vault 1013 at Chersonese (fig. 12).[1381] Its date is the same as that of the little rabbit of rock crystal found together with it: the third century a.d. As Chinese scabbard slides of jade and chalcedony have come to light from the Volga to Panticapaeum, and as far north as Perm,[1382] the find of a Chinese sword guard of agate in Chersonese is in no way surprising. The piece from the Crimea is as Chinese as the sword guard from Altlussheim.

Fig. 12. Agate sword guard from Chersonese, third century a.d. From Khersonesskii sbornik, 1927, fig. 21.

[The following two paragraphs were found loose; since they discuss swords they are inserted here.—Ed.]

Although Ostrogothic swords are not preserved or depicted or decribed, we know that they were heavy cutting weapons. In the battle Ad Salices in 376, the Ostrogothic cavalry “with mighty strength slashed at the heads and backs” of the fleeing Romans.[1383] Even more instructive is John of Antioch (fr. 214a): Theoderic dealt Odovacar “a blow with his sword upon the collar-bone. The weapon pierced his body down to the hip. It is said that Theoderic exclaimed ‘ In truth, the wretch has no bones.’”[1384]

There is no direct proof that in South Russia the Goths had swords like those from Altlussheim, Pouan, and from the grave of the Frankish king Childeric. These sumptuous weapons, glittering from gold and al- mandins, were made in Pontic workshops.[1385] Two similar swords were, indeed, found at Taman and at Dmitrievka.[1386] When one considers how fond of luxurious gold jewelry the Gothic nobles were—there must have been many hoards like that from Pietroasa, though perhaps not quite so rich—it is probable that at least some Gothic swords were as richly decorated as the just-mentioned weapons from South Russia. It is, I believe, not too bold an assumption that they were not markedly different from them.


The long and heavy lances of the South Russian Sarmatians are well known from wall paintings and reliefs of the first and second centuries a.d. The artists at times exaggerated their length; in the frescoes of the tomb of Anthesterius in Kerch[1387] they are represented to be 15 to 20 feet long.[1388] Still, the lance on Tryphon’s dedication from Tanais[1389] must have been nearly 10 feet long; the galloping horseman is holding it with two hands. The Roxolani did the same, as we know from Tacitus, who, however, was not impressed by what he thought to be a clumsy weapon.[1390] Other Romans thought differently. “Stretching out over the horse’s head and shoulders,” we read in Valerius Flaccus,[1391] “the fir-wood shaft, firmly resting on their knees, casts a long shadow upon the enemy’s field and forces its way with all the might of both warrior and steed.” In the second century, Roman horsemen, heavily or light armored,[1392] carrying long lances, xovrovg, attacked “in the manner of the Alans and Sauromatians.”[1393] The hastae longiores of the transdanubian Sarmatians[1394] were probably javelins, whereas the conti of the Alans and Sarmatians, mentioned by Claudian,[1395] were still the same thrust lances which the tribes in the East had carried since the sixth century b.c. They are attested for the Sauromatian,[1396] Early,[1397] and Middle Sarmatian periods, particularly for the latter (lance heads from stanitsa Kazanskaya and stanitsa Ust’-Labinskaya in the Kuban region,[1398] Tarki in Dagestan,[1399] Kalinovka[1400] and Lyapichev[1401] in the province Volgograd). The lance head in a woman’s grave at Tri Brata near Elista in the Kalmuk steppe[1402] is probably to be dated to the first century a.d., and the one from kurgan 28/2 at Kalinovka, 22 centimeters long,[1403] cannot be much later. The lance head from the river burial at Pokrovsk- Voskhod[1404] shows that the half-Alanized Huns on the Volga were armed like the Sarmatians in the preceding centuries.

It is a priori almost certain that the heavily armored Hunnic cavalry, like the Alanic and Roman cataphracts, carried long thrust lances. Avitus and the Hun wore the same equipment: the thorax and the lance. Among Narses’ horsemen were Huns beyond the Danube; their weapons were aaQiaaai.[1405]

In one of the graves at Hobersdorf in Lower Austria a 28-centimeter long lance head was found.[1406] Werner and Mitscha-Marheim[1407] date the graves to the first half of the fifth century which, I believe, is too early: in any case, the people buried there were Huns or closely related to them. The unsightly lance head from Pecs-t)szbg[1408] was probably the weapon of a Hun who rode in the king’s “household” (comitatus, druzhina).

The Lasso

“While the enemy are guarding against wounds from the sword-thrusts, the Huns throw strips of cloth plaited into nooses over their opponents and so entangle them that they fetter their limbs and take from them the power of riding or walking.”[1409] Ammianus’ statement is confirmed by Sozomen:[1410] A Hun “raised up his right hand in order to throw a rope [/3go/ov] over Theotimus, bishop of Tomis, for the Hun intended to drag the bishop away to his own country; but in the attempt, the Hun’s hand remained extended in the air, and the barbarian was not released from the terrible bonds until his companions implored Theotimus to intercede with God in his behalf.”[1411]

The Goths, the only Germans to use the lasso,[1412] took it over either from the Huns or the Alans. The Alans almost caught King Tiridates with their throwing ropes;[1413] in the fourth century, the lasso was their typical weapon.[1414] The lasso was used throughout such a wide area[1415] that it cannot be assigned to a specific cultural circle. It was known to the Scythians[1416] and Sarmatians,[1417] the Sargatians, a people “of Persian extraction and language,[1418] the Thatae, Sirachi, Phicores, and laxamatae, peoples between Bosporus and the Don,[1419] the Parthians,[1420] and the Persians in Sasanian times.[1421] In India the art of casting the lasso, pasa, was one of the martial arts studied by princes.[1422] It is the weapon of the Hindu gods.[1423] In the fourth century, the Kuai Hu, west of Kucha, used rawhide lariats which, whipping their horses, they threw at men.[1424]


Body Armor

To those historians who deny the Huns the capability of forging their swords,[1425] the mere question of whether they made their own armor must sound strange. Besides, in his description of the Huns Ammianus Marcellinus says nothing about armor. No iron or bone lamellae, no scales, no plates from splint armor[1426] have been found in Hun graves, or in association with what are doubtlessly Hunnic objects. The chain mail from Fedorovka in the former district Buzuluk, province Chelyabinsk,[1427] and Pokrovsk-Voskhod[1428] are suspect of being of Persian origin.[1429]

However, to wear armor, and especially metal armor, was everywhere and at all times the privilege of a few. Wegen des muhsamen, zeitraubenden und grosse Fertigkeit voraussetzenden Arbeitsgangs sind Kettenhemden zu alien Zeiten grosse Kostbarkeiten gewesen[1430] Medvedev adduces telling testimonies for the esteem in which chain armor was held in late medieval Russia.[1431] The much plainer scale armor also was apparently handed down from father to son and grandson rather than buried with the dead. A picture of Sarmatian civilization in Hungary, drawn from the finds, would not include scale armor. Yet we know from Ammianus that the cuirasses of the transdanubian Sarmatians were made of smooth and polished pieces of horn, fastened like scales to linen or leather shirts.[1432]

There is good, though indirect archaeological evidence that the Hun nobles, and perhaps not they alone, long before their first engagement with armored Roman troops wore some covering to protect their bodies in battle.

Recent finds enlarge considerably the material on which Thordeman and Arwidsson[1433] based their admirable studies on the history of armor.

I realize that the following survey has many gaps and that in a few years it will be obsolete. Yet for our purpose I hope it will suffice.

Bone lamellae are known from a much wider area and from much older sites than had been assumed before. They were found in graves of the Glazkovo period (eighteenth to thirteenth century b.c.) at Ust’-Igla on the Lena River in Cis-Baikalia and at Perevoznaya near Krasnoyarsk.[1434] Far to the west, on the lower Ob, bone lamellae and a technically and artistically marvelous breastplate of whalebone, found in settlements at Ust’-Polui near Salekhard, are datable between the fifth and third centuries b.c.[1435] Bone lamellae of the late Ananino period (fourth to third century b.c.) are known from Bol’shoi Skorodum[1436] and Konets-Gor[1437] in the Kama Basin.

Bone and horn armor is not necessarily inferior to, nor always earlier than, metal armor. To judge by the other grave goods, the bronze scales in the cemetery on the Morkvashka near Kazan[1438] are about two centuries earlier than the earliest bone scales found there. Pausanias greatly admired the Sarmatian corselets. The Sarmatians, he wrote,[1439]

collect the hoofs of their mares, clean them, and split them till they resemble the scales of a dragon. Anybody who has not seen a dragon has at least seen a green fir cone. Well, the fabric which they made out of the hoofs may not be inaptly likened to the clefts on a fir cone. In these pieces they bore holes, and having stitched them together with the sinews of horses and oxen, they use them as corselets, which are inferior to Greek breast-plates neither in elegance nor strength, for they are both sword-proof and arrow-proof.

The cuirasses of the horsemen in the Hellenistic armies were sometimes made of horn,[1440] and if the author of the Sylloge Tacticorum does not copy earlier authors but describes the armament of the Byzantine army of his time, the clibania of the horsemen were, as late as the tenth century, either of iron or of horn.[1441]

In the last centuries b.c. and the first centuries a.d., armor of one type or the other was widely used in northern Eurasia. I pass over the well-known and often discussed representations of armored warriors in Parthian[1442] and Gandharan art,[1443] but I would like to draw attention to two little- noticed metal figures from the Altai region and western Siberia.

A bronze pendant (fig. 12A),[1444] said to have been found in a grave at Barnaul in the Altai region, shows a man wearing a scale armor and a conical helmet; his quiver has the hour-glass shape that occurs from China to the Caspian Sea. To draw any conclusion from the style of the pendant—provided the drawing is correct—would be risky. The earliest hour-glass quiver is datable to the fourth century a.d.[1445]

Fig. 12A. Bronze pendant said to have been found in a grave at Barnaul, Altai region, showing a man in scale armor and conical hat with an hour-glass-shaped quiver, datable to the fourth century a.d. From Aspelin 1877, no. 327.

Then there are the horsemen on two gold pendants from western Siberia (fig. 12B).[1446] They also wear scale armor. Their similarity to the rider on the famous wall hanging in Pazyryk (fourth century b.c.)[1447] indicates an early date. The short jacket, the boots, the horse trappings are the same here and there. The square tuft in the mane, clearly discernible on one plaque, corresponds to the crenelation of the mane in Pazyryk.[1448]

Fig. 12B. Two horsemen in scale armor shown in gold pendants from western Siberia. From Kondakov and Tolstoi, 3, fig. 49.

The most common and probably the earliest type of armor in the steppes was scale armor. In spite of the strong influence the civilization of Urartu exerted on Scythian metal work, the Scythians did not take over Urartian lamellar armor.[1449] Throughout the centuries that we can follow their history they wore scale armor.[1450] The scales, sometimes gilded,[1451] were of bone, bronze, or iron;[1452] occasionally the two metals were combined.[1453] As the finds from Kobylovka near Atkarsk west of Saratov, from Tonku- shorovka (formerly Marienthal), and from the province Astrakhan show,[1454] the Sarmatians of the Sauromatian period also had bronze scale and lamellar armor. Bronze lamellae of the Early Sarmatian period are rare.[1455] Bronze, iron, and bone scales were found in Middle Sarmatian graves in the Trans-Ural steppes,[1456] on the lower Volga (Kalinovka,[1457] Pogromnoe,[1458] Usatovo on the Eruslan[1459]), in the Kuban Valley,[1460] and in the southern Ukraine.[1461] Scales in a grave at Vor’bi[1462] indicate Sarmatian influence on the P’yany-bor civilization. Sarmatians and Sarmatized Bosporans,[1463] horse and man covered with corselets of scale armor, are depicted on the wall paintings at Panticapaeum-Kerch.[1464]

For the Late Sarmatian period (II-IV a.d.), we have the adduced testimony of Pausanias (about 175 a.d.). The Sarmatians of Emperor Galerius’ bodyguard on the arch of Thessalonica[1465] wear the same scale armor as the galloping horseman on a relief from Tanais,[1466] datable to the third century (see also the figure of a member of the Roxolani tribe, fig. 12C), or the Bosporan kings Cotys II and Sauromates II on their coins.[1467] There is, finally, a stone relief from Chester in the Grosvenor Museum.[1468] It shows a Sarmatian, a cloaked horseman, with a tall conical helmet, holding with both hands a dragon standard or pennon. The surface tooling on man and horse is much worn, but what remains suggests that both were shown clad in scale armor.

Fig. 12C. The representation of a Sarmatian member of the Roxolani tribe in a detail of the marble relief from Trajan’s Column, in the Forum of Trajan, Rome. Datable to the second decade of the second century a.d. Photos courtesy Deutsches archaologisches Institut, Rome.

For the tribes on the eastern end of Eurasia we have in the main to rely on Chinese sources. We learn from them that the primitive, perhaps Tungus, Su-shen in Manchuria had leather and bone armor,[1469] and we read about the armor of the Fu-yu[1470] and Jo-chiang.[1471] In the armies of the Hsiung-nu rode “cuirassed horsemen.”[1472] Their armor is called chia which, according to Laufer,[1473] in Ssu-ma Ch’ien means “hide armor.” But in 1956 Dorzhsuren found in one of the Hsiung-nu graves at Noin Ula an iron scale, with the fabric, to which it had been fastened, still on it.[1474] Iron scales occur in Tuva in graves of the Shurmak period (second century b.c. to first century a.d.).[1475] Like the Chinese of the Han period,[1476] the Hsiung- nu probably also had bronze and leather scale armor.

Finds of metal lamellar armor in the steppes are rare. Those from Kutr- Tas, province Kustanai,[1477] and Tomilovka on the Tobol River[1478] are probably of Persian provenance. Gryaznov thinks an oblong iron plaque with perforations around the edges, found at Blizhnie Elbany north of Barnaul, is the lamella of an armor.[1479]

At the beginning of our era, or, perhaps, even earlier, chain mail began to take its place next to scale armor among the Sarmatians in the Kuban Basin.[1480] In the first century a.d., Valerius Flaccus[1481] described the Sarmatian catafractarii: “Their armor is bristling with flexible chains and their horses have the same protective cover” (riget his molli lorica catena, id quoque tegimen equis). The chain mail in Karabudakhkent in Dagestan is certainly of Sarmatian provenance.[1482] The same is probably true for the chain mail found in the basin of the Kama and its tributaries: Vichmar’,[1483] Atamonovy kosti,[1484] Gainy,[1485] and Pystain.[1486]

The figure incised on a sheep astragal, found in Kobadian in Tadjikistan (third to second century b.c.),[1487] seems to represent a warrior in a long coat, with what might be a helmet on his head; D’yakonov takes the crisscross lines on the coat for chain mail, but they could be just quilts. The same might be true for another figure of a warrior, incised on a bone, from the cemetery at Kuyu-Mazar near Tashkent (second to first century b.c.).[1488] In the fourth century a.d., the armor of the Kuai Hu, west of Kucha, was like “linked chain, impenetrable to bow and arrow.”[1489]

In view of the literary and archaeological evidence for the spread of body armor in the first centuries a.d. from the Ukraine to Manchuria, it is a priori unlikely that the warlike Hun tribes fought without protection of some sort of armor. In addition, we have the testimony of Greek and Latin sources.

The Huns, Alans, and Goths in the army which Theodosius led against Maximus in 388 were not Roman soldiers but free barbarians, enlisted for one campaign. They were not outfitted with weapons manufactured in the armorum fabricae; they brought their own equipment with them. It included heavy iron cuirasses.[1490]

Fifty years later, some Huns under the command of Litorius in Gaul wore the same armor. They, too, were not milites but auxiliatores and socii.[1491] Sidonius Apollinaris describes a duel between Avitus and a Hun in Litorius’ contingent that reads like one taken from a medieval romance:

When the first bout, the second, the third have been fought, lo I the upraised spear comes and pierces the man of blood; his breast was transfixed and his corselet twice split, giving way even where it covered the back [post et confinia dorsi cedit transfosso ruptus bis pectore thorax]; and as the blood came throbbing through the two gaps, the separate wounds took away the life that each of them might claim.[1492]

The thorax was clearly not a mere breastplate but a piece of armor protecting the body on all sides, not a leather corselet but a metal shirt. It may have been of the same type as the one worn by the Hun Bochas, one of Belisarius’ bodyguards:

He came to be surrounded by twelve of the enemy, who carried spears. And they all struck him at once with their spears. But his thorax withstood the other blows, which therefore did not hurt him much; but one of the Goths succeeded in hitting him from behind, at a place where his body was uncovered, above the right armpit, right close to the shoulder, and smote the youth, though not with a mortal blow.[1493]

Pacatus in the fourth, Sidonius in the fifth, and Procopius in the sixth century testify that the Huns were “men with iron cuirasses” (ctWoeg aideQw TeOojQaxtapxvoi).[1494] There are three more sources which, to my knowledge, have not been utilized. The first is a homily on St. Phocas by Asterius of Amasea. The saint was venerated throughout the world. Even “the most ferocious Scythians who lived on the other side of the Euxine, near the Maeotis and Tanais, and as far as the Phasis River,” were deeply devoted to him. One of their rulers “took off his crown, sparkling with gold and jewels, and put off his war cuirass of precious metal (for the armor of the barbarians is ostentatious and sumptuous)” and sent them to St. Phocas’ church in Sinope.[1495] The homily[1496] was written about 400,[1497] at a time when the greater part of the territory described by Asterius was under Hun domination. The ao/<w xat /ZaatAevs[1498] was most probably a Hun.

The other source is Merobaudes’ panegyric on Aetius’ third consulship. In verses 79–83,[1499] the poet describes the equipment and weapons of the Huns:

fulgentes i]n tela ruunt: gravis ardeat auro
a][1500] uratae circumdent tela pharetrae,
aurea cri]spatis insidat lamna lupatis:
incendant] gemmas chalybes ferroque micantes

*cassidis[1501] a]uratis facibus lux induat enses.

Belts, quivers, horse bits, helmets, and the armor, studded with precious stones, were gilded. The hexameters cannot be dismissed as a mere imitation of what Merobaudes read in Claudian and Statius. They were obviously patterned on In Rufinum II, 352–377, and other passages dealing with the sumptuous equipment of the Roman elite cavalry. It is, furthermore, true that nearly all the golden or gilt weapons of the Huns have their Roman counterparts.[1502] Some may actually have been of Roman provenance;[1503] a few Huns may have worn gilt Persian armor.[1504] But all this does not detract from Merobaudes’ description of the Hunnic arms. The poet could not have drawn such a picture had it not corresponded to reality. His public, and first of all Aetius, knew the Huns. By calling felt caps helmets and leather jackets armor, Merobaudes would have made himself ridiculous. Some Huns did wear costly armor.

There is, third, a short passage from Priscus, preserved in Suidas: Zercon, the Moorish jester, accompanied Bleda in his campaigns in full armor.[1505]

Six authors, independently of one another, speak of the body armor of the Huns. This by no means proves that all Huns wore armor. Most of them were, as Ammianus said, lightly equipped, and remained so until the end of Attila’s kingdom. But many Hun nobles were heavily armored, and their number was apparently growing as the Huns acquired riches from booty and tribute.


The Huns in the army of Theodosius must have worn metal helmets. With their bodies protected by iron armor, they could not have fought bareheaded or worn soft leather or felt caps. As we know from Merobaudes, in the 440’s the helmets of the Hun nobles were gilt. What such cassides looked like can be learned from Sidonius Apollinaris. In the Panegyric on Avitus (253–255), he describes how the heads of the Hun boys were flattened:

The nostrils, while soft, are blunted by an encircling band, to prevent the two passages [i.e., the nose] from growing outward between the cheekbones, that thus they make room for the helmets [utgaleis cedant]; for these children are born for battles, and a mother’s love disfigures them, because the area of the cheeks stretches and expands when the nose does not interfere.

From these verses, stilted but clear in their meaning, Arendt rightly concluded that the Hunnic helmets had nosepieces.[1506]

It is understandable that helmets do not occur in Hunnic graves. Like armor, helmets were so costly that they were handed down from generation to generation—the Spangenhelm from Gammertingen was more than a hundred years old when it finally was deposited in the grave.[1507] Perhaps the widespread belief that the dead were proof against attack also played a role.

The Hunnic helmet was possibly a Spangenhelm. If, as Post maintains, the Spangenhelme with a copper framework are to be strictly separated from those with an iron framework,[1508] the Hunnic helmets cannot belong to the former group in which nosepieces occur in rudimentary form or not at all. They may, however, have resembled the iron Spangenhelm from Der-el-Medineh in Egypt,[1509] which Werner dates to the fifth century; he takes it for the helmet of a Roman officer.[1510] The Sarmatian helmets on the Galerius arch in Thessalonica are provided with nasals, though whether they are Spangenhelme cannot be determined.[1511]

Furthermore, there is the curious helmet, laced with leather thongs, from Kerch, uncovered by Kulakovskii in a catacomb.[1512] It was found together with lamellae from mail shirts, a lance head, twenty arrowheads, pieces of gold-embroidered fabrics, golden plaques from a belt, and a coin of Emperor Leo (457–473). Because the coin was pierced, the find has been dated to the sixth century.[1513] But the tomb was plundered at an early time, and the coin may have been lost, as Kulakovskii thought, by the tomb robbers. Grancsay dates the helmet to the fifth century and takes it for Avaric,[1514] although in the fifth century the Avars were not even near the Crimea. If the man was buried in a catacomb built for him, he could have been a Hun. Such catacombs were not built after the fifth century. But the grave may be a secondary burial. Non liquet.

Lately it has become fashionable to give the Sasanians credit for every advance in military technique in late imperial times. The nasal of the Roman helmet is supposed to be of Eastern origin, but whether the Sasanians were the givers is at least doubtful. The few preserved Sasanian helmets have no nosepieces. The face of the horseman on the rock sculpture at Taq-i-Bustan is covered, except for the eyes, with a defense of mail, suspended from the rim of the helmet.[1515] Around a Sasanian helmet in the Metropolitan Museum of Art in New York there are several perforations to which a similar defense of mail must have been attached.[1516]

As for the mentioned conical helmets of the Sarmatians on the Galerius arch, there existed more of this kind in the East. Among them are the iron conical helmets found on the Vangai River between Tobolsk and Omsk in western Siberia. One is gilt, the other has gilt inlays of dragon and griffinlike figures.[1517] In the hoard were also two Chinese mirrors of the Han period[1518] and a silver disk with the representation of Artemis (?), probably made in Bactria in the first half of the second century b.c.[1519] It is evident that the helmets were not made where they were found, but neither their construction nor their decoration gives an indication where they came from. The incised figures of horsemen on the plaques found together with the helmets[1520] show that the tribes on the lower Ob wore conical helmets with nosepieces about the beginning of our era.

It is possible that the Hun nobles wore helmets of various forms and constructions, Spangenhelme, helmets like those of the Sarmatians on the Galerius arch, helmets of the Vangai type, and still others.[1521]


If the Hunnic shield had an umbo, the hollow boss of iron or bronze covering the aperture of the Roman and common Germanic shield, one should have been found in a Hunnic grave. Its absence indicates that the shield of the Huns, like that of the Scythians,[1522] the Persian infantry,[1523] and some of the Roman troops,[1524] was of wickerwork, possibly covered with leather.[1525]

Very little about the shields of the Sarmatians is known. According to Strabo, the Roxolani had shields of wickerwork.[1526] The copper umbo in kurgan 10 of the Maeoto-Sarmatian cemetery at stanitsa Elizavetovskaya[1527] seems to be of Greek provenance. Two umbones in the Sarmatian cemetery Malaaeshti in the Moldavian SSR[1528] might have come from the Bosporan kingdom.[1529] There remains the umbo in the Gepidic cemetery at Kiszombor in Hungary; it is in the shape of a high cone, different from the usual low or hemispherical umbo.[1530] In Csongrad a similar umbo was found in a grave together with a trihedral arrowhead[1531] which could be Hunnic or Sarmatian; Parducz takes the grave for Vandalic.[1532] Another umbo of this type lay next to a sitting skeleton in a grave at Nyiregyhaza,[1533] possibly attributable to a Sarmatian. There is, as we see, very slight evidence that the western Sarmatians had wooden shields with umbones, and none that the Sarmatians in the East had them.

A passage in Sozomen tells us something about the shields of the Huns at the end of the fourth century. While talking to Theotimus, bishop of Tomis, a Hun leaned on his shield, “as was his custom when parleying with his enemies.”[1534] As the Hun was standing, his shield must have been at least as big as some of the Scythian oblong shields on the Kul Oba vase,[1535] or those of the Sarmatized Bosporan foot soldiers on the wall paintings in Panticapaeum.[1536] If we apply the ratio determined from the terra cotta figures of Sarmatians from Kerch, with their long shields,[1537] to life proportions, the Hun, assuming that he was 5 feet, 5 or 6 inches, carried a shield 2 and possibly 3 feet long. Such a large shield was not suitable for use on horseback. Narses’ cavalry, which must have included many Huns, were armed with javelins, bows, long lances, and small shields, pelted.[1538]

Huns in the Roman Army

The Notitia Dignitatum lists Frankish, Alamanic, Gothic, Vandalic, Herulic, Marcomannic, Quadic, and Alanic alae, vexillationes, cohortes, cunei, and auxilia[1539] but no Hunnic units. Those Huns who went over to the Romans were apparently distributed among the numeri barbari or the Theodosiani, Arcadiani, and Honoriani equites, Sagittarii, and ar- migeri. Only under exceptional circumstances were Huns kept together. Such a formation was, I believe, the Unnigardae.

Of the units which in the first decade of the fifth century served in Libya Pentapolis, only the Balagritae were Africans; the cavalry consisted mostly of Thracians, the infantry of Dalmatians and Marcomanni.[1540] The best troops were the Unnigardae. Synesius, bishop of Ptolemais, praised them as the savior of his beloved city. It was true, they sometimes got out of hand, “like young hounds,” but their leader “would take them by the throat and call them in, even before they sated themselves with their charge and their wild-beast slaughter.”[1541] The Unnigardae[1542] were a small corps of horsemen, excellent in lightning attacks and dashing raids, at their best as scouts and vanguards. “They are in need of a rearguard and an army drawn up in order of battle.” From Synesius’ letter to his friend Anysius in Constantinople,[1543] we learn that the Unnigardae formed an independent troop, receiving their relays of horses, equipment, and pay directly from the emperor. Their status, tactics, and ferocity are comparable to those “Massagetae” who in the sixth century fought under Belisarius in Africa and Italy.

Unni is undoubtedly the ethnic name.[1544] It must be left to Germanic scholars to decide whether gardae could reflect the Latin pronunciation of the Germanic word that gave Old Italian guarda and French garde[1545]

Another Hunnic formation was possibly stationed in Britain. One of the commanders per lineam valli, Hadrian’s wall, was the praefectus alae Sabinianae, Hunno [1546] Could Hunnum be “the fort of the Huns”?

The ala Sabiniana, in Britain since the second century,[1547] was of course not a Hunnic unit. But the Notitia Dignitatum, in the form it has come down to us, is a patchwork composed of “returns” of various times. The name Hunno might have been substituted for an older one. It would not be the only case where the compilers brought an obsolete “return” wholly or partly up to date.

Under Stilicho the defenders of the Wall were mostly native federates;[1548] one or another of the more important forts was possibly held by auxiliaries. But it seems unlikely that Stilicho, who as we know did have Hun auxiliaries in his armies, would have used them for the defense of a half-abandoned province. If Huns were actually stationed at Hunnum, they could have been there in the last years of Gratian, who had Huns in his service.

There was a fort Ovvvcov near Oescus on the right bank of the Danube.[1549] The place was named after the Hun garrison, as Baaregvat^[1550] after the Basternae and Sagp,aOd>v in the Haemimons[1551] after Sarmatians. However, Hunnum might be a Celtic word.[1552]

The Unnigardae and the Huns in Britain—provided they were there —served far away from their homes. But there were also Huns in the garrisons of the Roman camps along the Danube in Pannonia and the Balkan provinces, both before and in Attila’s time. Whether the bone strips of composite bows found in Carnuntum[1553] point to Huns or Alans cannot be decided. The fragments of bronze cauldrons in Intercisa[1554] and Sucidava[1555] leave no doubt that at one time or another Huns lived there. Werner thinks that broken cauldrons and nomadic mirrors were left there by Huns and other barbarians who settled in the abandoned camps, where they also could find metal.[1556] The find circumstances are not in favor of such an interpretation.

The fragment from Intercisa was found in a burned-down late Roman building; in another room lay the fragments of fifteen to twenty iron helmets. It is a priori unlikely that Huns, who disliked so intensely to live in houses, would have settled in the ruins of a Roman camp. It is even more improbable that the Huns, who were so desperately short of metal, should have overlooked the helmets in the other room. In Sucidava four fragments of Hunnic cauldrons were found in the layer of ashes which covered the whole area of the castellum. Two of the fragments lay near hearths, evidently to be melted for the fabrication of bronze objects. The milites riparenses were extremely poor. Neither gold nor silver coins were found. No wonder that every bit of bronze was to be used. The cauldrons were certainly not owned by Huns who settled in the former camp. There was nothing left of it. The barracks were burned down; of the amphorae in which the Danube flotilla brought oil even when both banks of the river were held by the barbarians only fragments were found; there were no human bones anywhere in the ashes. It is practically certain that the camp was hastily evacuated,[1557] and it is quite possible that the garrison itself put the fortress on fire.[1558] How did the Hun cauldrons get there? They could not have been booty; neither in the first nor the second Hun war in the 440’s did the Romans defeat the Huns even in a single encounter. It is hard to imagine that at the annual fair on the Danube a Roman bought a Hunnic cauldron. There is only one explanation of the presence of the Hunnic vessels in Sucidava: They must have belonged to Huns serving in the Roman army.[1559]

The archaeological evidence is supplemented by a few lines in Priscus. In his negotiations with the East Roman envoys in 449, Attila “would not allow his own servants to go to war against him, even though they were unable to help those who turned over to them the protection of their native land, for, said he, what city or what fortress he set out to capture would be saved by these refugees ?”[1560] Sucidava was one of those fortresses. There certainly were more of them along the lower Danube.

Like the Unnigardae, Aetius’ Hunnic auxiliaries were excellent fighters, but their lack of discipline made them often more a terror to the provinces they were supposed to defend than to the enemy. Again and again they broke loose and “with raid and fire and sword and savagery and pillage destroyed all things nearby.”[1561] In Gaul the Romans had to keep garrisons in the cities to protect them from their own auxiliaries.[1562] Years later the atrocities committed by the Huns were still vividly remembered. In his biography of St. Martin, Paulinus of Perigueux wrote, “Seized by sudden fear Gallia admits the Huns as auxiliaries. One can barely suffer as allies those who behave more cruel than the enemy, and in their savagery throw off the foedus. ”[1563]

VI. Religion

The huns, wrote Ammianus Marcellinus, were a people without religion. Like unreasoning beasts, they were utterly ignorant of the difference between right and wrong, deceitful and ambiguous in speech, faithless and unreliable in truce, niillius religionis vel superstitionis re- vereniia aliquando districti (XXXI, 2, 11).

There was hardly a barbarian people that did not lack the virtues in which the Romans excelled. The Parthians held their promises only as long as it was to their advantage.[1564] The Heruli were not bound by any convention.[1565] The Moors, like the Huns, “did not care for oaths,” and for the same reason: “Among them was neither fear of God nor respect of men.”[1566] The Avars, successors of the Huns, were “the most faithless of all nomads.”[1567] The list could be continued.

Ammianus’ statement about the irreligion of the Huns was not based on firsthand knowledge; it was the conclusion he drew from their behavior or, to be more exact, from what people who had unpleasant experiences with the Huns told him. Actually, he did learn about a religious custom of the savages from his informants, though he did not recognize it as such.

When they once put their neck into a faded tunic, it is not taken off or changed until by long wear and tear it has been reduced to rags and fallen from them bit by bit.[1568]

Ammianus cannot be blamed for taking the aversion of the Huns to washing their clothes for just another mark of their beastliness. Ibn Fadlan, a keen observer and ever ready to ask questions, noticed the same unclean habit among the Oguz without suspecting that it might have religious significance.[1569] The object of the Turkish and Mongol[1570] custom was to avoid offense to the water spirits.[1571] It probably was the same with the Huns, and it presumably corresponded happily with their natural inclinations. Priscus noticed as remarkable that Attila’s dress was clean.[1572] The “Massagetae”- Huns were as dirty as the Sclaveni.[1573]

The Huns and Christianity

By the middle of the fourth century, the Roman and Romanized population of Pannonia was preponderantly Christian. Arianism was fairly strongly entrenched; the bishops of Mursa and Sirmium staunchly upheld the heretic tradition. In the 380’s and 390’s it took all the zeal of St. Ambrose, efficiently supported by the secular arm, to bring the Danube provinces back into the orthodox fold.[1574] In Attila’s time Pannonia, both the part which had been ceded to him and the ill-defined no man’s land east of Noricum, was apparently solidly Catholic. In Pannonia secunda the Christian community of Sirmium survived the Huns,[1575] Ostrogoths, and Gepids.[1576] Sopianae in Valeria, from where Christianity had been carried north and west,[1577] withstood all storms of the migration period.[1578] In Pannonia prima urban life had almost ceased when the Huns came, but there too small Christian communities seem to have held out.[1579]

Cut off from the churches in Romania, the Catholics in Pannonia offered no political problems to the Huns. The big landowners had fled, and the small people who stayed were utterly unable to organize any resistance against their lords. The danger that the Catholics might act as a fifth column for the Romans, at times so acute in Persia and a permanent threat to the Vandals in Africa, did not exist in Hunnia. Attila could

afford to be tolerant: He allowed his Catholic subjects to pray and fast as long as they meekly worked for him.

Most of the Christian Germans under Hun rule were Arians.[1580] One may doubt whether Attila knew the difference between the Arian heresy and the orthodox creed. It is hard to imagine the Hun king listening to a discussion about the consubstantiality of the Father and the Son. But he must have been aware that his Germanic followers and subjects were not of the same religion as the emperors in Ravenna and Constantinople. The mere fact that the Arian clergy under the Huns were not persecuted for their faith as they were in the empire, both in the Western and Eastern part, ensured their loyalty to the Hun kings.

The Huns had Christian slaves. What Prosper said about the ways by which the gospel reached the pagans beyond the borders may also have been true, to a modest extent, for the Huns:

Some sons of the church, made prisoners by the enemy, changed their masters into servants of the gospel, and by teaching them the faith they became the superiors of their own wartime lords. Again, some foreign pagans, while serving in the Roman armies, were able to learn the faith in our country, when in their own land they could not have known it; they returned to their homes instructed in the Christian religion.[1581]

One or another Hun mercenary in the Roman army may have been baptized. A particularly zealous slave may have converted his master,[1582] or, more likely, his master’s wife. But it is improbable that men like One- gesius, Attila’s prime minister, should have renounced the faith of their fathers because their bath attendants read the Bible to them.

Had their kingdom not so suddenly collapsed, the Huns would sooner or later have embraced Arian Christianity. The Arian Goths were much closer to them than the Romans. Compared with wretched Catholics in the dying towns of Pannonia, not to speak of the prisoners of war, the

Gothic chieftains were almost the equals of the Hun nobles. But two generations of slowly growing symbiosis of the upper strata of Hun and Germanic society were too short to bring the Huns over to the religion of the Goths.

Salvian, writing about 440,[1583] classified the Huns among the heathen nations:

I shall discuss the pagans first, since theirs is the older delusion: among these, the nation of the Saxons is savage, the Franks treacherous, the Gepids ruthless, the Huns lewd—so we see that the life of all the barbarians is full of vice.... Can you say that their vices imply the same guilt as ours, that the lewdness of the Huns is as sinful as ours, the treachery of the Franks as reprehensible as that of the Christians, the greed of the Alans as much to be condemned as that of a believer? If a Hun or Gepid is deceitful, what wonder is it in one who is utterly ignorant of the guilt involved in falsehood? Can it be said of the Huns: See what sort of men these are who are called Christians?[1584]

The only Huns Salvian knew were those who served under the Romans in Gaul. But the Hun kings obviously did not draft only pagans into the auxiliary corps they lent to their Roman “friends.” The Huns, as a people, were as pagan in the middle of the fifth century as they had been when they crossed the Don.

Salvian’s statement is at variance with what Jerome and Orosius say. It seems, furthermore, contradicted by an often quoted passage in Theodoret about the successes of the priests whom John Chrysostom allegedly sent to the Huns. Niceta of Remesiana also is said to have carried his missionary activities beyond the Danube into the country of the Huns. A closer examination of the evidence reveals that it is either untrustworthy or has been misunderstood.

In 399, Jerome called the Huns “wild beasts.”[1585] But when shortly after, in his letter to Laeta, he described Christ’s triumph over the demons, he wrote: “From India, from Persia, and from Ethiopia we welcomed crowds of monks every hour. The Armenians have laid aside their quivers, the Huns are learning the psalter, the frosts of Scythia are warmed by the fire of the faith.”[1586] At about the same time Jerome explained the psalter to two Goths, Sunnia and Fretela.[1587] He may have lumped together the Huns and the Goths, both Scythian peoples, but it is more probable that he simply invented the psalm-singing Huns as he invented the crowds of monks from India.[1588]

The Huns, wrote Orosius in 418, filled the churches of the West and the East.[1589] This is the statement of a theologian. The early Christian belief in the imminent end of the world implied the certainty that the gospel was being preached to all nations.[1590] What Tertullian and the apologists of the third century said about the spread of Christianity to the Scythians, Parthians, and Indians were merely the conclusions they drew from the scriptures. Had they known the Huns, they would have included them in the number of baptized barbarians. If the lists of converted peoples in Tertullian and Arnobius were the products of exegesis, those of the post-Nicaean fathers were pure rhetoric. Poets and theologians indulged in exotic names. The Scythians, Massagetae, Sauromatae, Tibareni, Hyrcanians, Caspians, Geloni, Moors, Indians, Ethiopians, Persians, Bactrians, Cimbri, even the Seres were now Christians.[1591] Orosius was the pupil of St. Augustine, who rejoiced at the fact that “what as yet is closed to those who fight with iron is not closed to him who fights with the wood [cross].”[1592] In a way Orosius was right: The Huns did fill the churches,[1593] but only to ransack them. In the East, in Thrace, they killed the monks, raped the nuns, and put fire to the churches; first, of course, they carried the sacred vessels away. They did the same in the West, in Gaul.[1594]

About the relationship between the Huns and the Christian priests in the border provinces only Sozomen has something to tell us:

The church of Tomis, and indeed all the churches of Scythia [t.c., Scythia minor], were at this time under the guidance of Theotimus, a Scythian. He had been brought up in the practice of philosophy, and his virtues had so won the admiration of the barbarian Huns, who dwelt on the banks of the Hister,[1595] that they called him the god of the Romans, for they had experience of divine deeds wrought by him. It is said that one day, when traveling toward the country of the barbarians, he perceived some of them advancing toward Tomis. His attendants burst forth into lamentations, and gave themselves up for lost; but he merely descended from horseback, and prayed. The consequence was that the barbarians passed without seeing him, his attendants, or the horses from which they had dismounted.[1596]

The passage, which refers to the last years of Theodosius I,[1597] not only throws a sharp light on the inefficiency of the frontier defense, it also shows that Theotimus could not have had as much success with the Huns as has been claimed.[1598] A missionary who has to make himself invisible when he meets those he is supposed to convert will not baptize many. Indeed, if Theotimus or his successors or any other bishop anywhere in the Eastern empire had won more than a few Huns to the faith, the Byzantine church historians are not likely to have failed to report their successes.[1599]

Some scholars adduced the beautiful poem in which Paulinus of Nola praises the zeal of his friend Niceta of Remesiana[1600] for spreading the gospel among the Scythae, Getae, and Daci[1601] as another proof of the conversion of Hunnic tribes north of the Danube.[1602] But Scythae, Getae, and Daci are only archaic names of the Bessi and other tribes in the mountain glens of Haemus and Rhodope.[1603]

There remains Theodoret. John Chrysostom, bishop of Constantinople, says Theodoret in his Church History,[1604] was informed that some nomadic Scythians, who pitched their tents along the banks of the Hister, thirsted for salvation but had no one to bring it to them. John sought men willing to imitate the labors of the apostles and sent them to these people. Theodoret himself saw the letter which John wrote to Leontius, bishop of An- cyra in Galatia, in which he informed Leontius about the conversion of the Scythians, asking him to send them men capable of guiding them.

The “nomadic Scythians” are supposed to be Huns.[1605] It is true that in another passage Theodoret calls the Huns nomadic Scythians.[1606] But it does not follow that all nomadic Scythians were Huns. Asterius of Amasea, John Chrysostom’s contemporary, wrote about the nomadic Scythians on the Cimmerian Bosporus and near the Rhine.[1607] In enumerating the nations to which the fame of Symeon Stylites spread, Theodoret named, besides the Persians, Indians, and Ethiopians, also the nomadic Scythians,[1608] obviously a collective term without sharp ethnic or linguistic definition. In the Vita Athonitae, the Magyars were called nomadic Scythians.[1609] In one of his orations[1610] on John Chrysostom, Theodoret spoke once more about the Scythians whom the sainted bishop had converted. There he called them wagon dwellers, another of the stereotyped attributes of the Scythians carried over from one author to the other. In the Church History Theodoret wrote about the guidance which the Scythians still needed. In the oration they are already exemplary Christians: “The barbarian, dismounting from his horse, has learnt to bend his knees, and he, who was not moved by the tears of the prisoners, has learnt to cry over his sins.” Theodoret was not satisfied with one nation converted by his hero. He let him bring the gospel also to the Persians, and they too worship Christ.

Still, Theodoret’s account cannot be dismissed as fiction. He embroidered the little he knew, but he had at least one reliable piece of information: John’s letter to Leontius. Theodoret could not have invented it. Its content is too strange and its recipient too unfit for such an assumption. John wrote the letter in Constantinople, thus between February 398 and June 404, probably nearer the last date, for in the first years of his office he was fully occupied with reforms in the capital itself. Leontius played a leading role in the intrigues which resulted in the downfall of John. The hatred of the crafty scoundrel almost cost John his life on the journey to Cucusus.[1611] It must, therefore, have been a very special reason that induced the bishop of Constantinople to ask the bishop of Galatia to send priests to the barbarians on the Danube. Why were people of this province in Asia Minor so much better equipped for such work than the priests of John’s own diocesis?

There is, I believe, only one answer to this question. John must have thought that in the whole Eastern empire Galatians were the only ones who could preach the gospel to the “nomadic Scythians” in their own language. The missionaries whom John sent to the Goths were “talking the same language as those” (6/zo’yAcorroi exeivoi?).[1612]

Besides Greek, not so few Galatians spoke their Celtic language as late as the end of the fourth century. Jerome recognized the close relationship of Galatian to the Celtic dialect he had heard spoken around Trier.[1613] There was only one people on the Danube that spoke “Galatian,” namely the Bastarnae. With the exception of Strabo, who had some doubts, all Greek authors regarded them as Celts. Plutarch spoke of the “Galatae on the Hister, who are also called Bastarnae.”[1614]

Only a few years before John wrote to Leontius, Bastarnae, Goths, and Alans crossed the Danube and ravaged Thrace.[1615] After 400, the Bastarnae are not mentioned, but BaatEQvai;, the name of a fortress built by Justinian II in Moesia on the Danube,[1616] shows that they lingered on in the northern Balkans, preserving their ethnic identity as late as the sixth century.[1617]

Later authors were puzzled by the “Celts” converted by John Chrysostom; they identified them with the Arian Goths,[1618] although, of course, no author of the fifth century would have mixed up Goths and Celts. At the frontiers of the Eastern empire no other Celts existed except the “Galatian”-speaking “nomadic Scythians,” the Bastarnae.[1619]

The Apocritus of Macarius Magnes contains a list of the peoples to whom the gospel has not yet been preached: the seven races of the Indians who live in the desert in the southeast, the Ethiopians who are called Macrobians, the Maurusians, “and those who dwell beyond the great northern river Ister which shuts off the country of the Scythians, where twelve tribes of nomad barbarians live, of whose savage state Herodotus tells us, and their evil customs derived from their ancestors.”[1620]

Seers and Shamans

Litorius, one of Aetius’ generals, was supposedly the last Roman general to perform the ancient pagan rites before battle.[1621] In 439, under the walls of Toulouse, his army was destroyed by the Visigoths, he himself wounded, taken prisoner, and put to death. The Romans, maintained Prosper, were defeated because Litorius refused to listen to the advice of his officers; instead, “he trusted the responses of the haruspices and the monitions of the demons.”[1622] Can Prosper be believed?

It is, perhaps, not particularly significant that in his chronicle Hy- datius said nothing about the soothsaying. He barely mentioned the war in 439;[1623] besides, he may have thought it would cast a doubtful light on Aetius, whom he held in high esteem, if one of the ductor’s most trusted lieutenants was a pagan. Salvian’s silence is more important. He lived in Gaul; he must have known Litorius. Ever ready to accuse his countrymen of all possible sins, Salvian would not have passed over Litorius’ “crime,” had the unfortunate commander committed it. According to Salvian, the Romans lost the war because, unlike the Goths, they did not put their hope in God but relied on their Hunnic auxiliaries.[1624]

That Litorius should have had professional Roman haruspices in his army is unlikely for another reason. In times of stress the Christian government was forced to tolerate the stubborn paganism of an exceptionally able general as, for instance, in 409 when an antipagan law was temporarily revoked in favor of Generidus. But since then one edict after another had been issued which threatened with capital punishment those who dared to indulge in the “insanity” of consulting haruspices. It is inconceivable that as late as 438 a Roman general could have the entrails of victims inspected before engaging the enemy.

And yet there must be a kernel of truth in Prosper’s accusation. When we consider that the troops under Litorius’ command were Huns,[1625] the explanation becomes clear at once. Not Litorius but his Huns wanted to know the outcome of the battle they were about to enter. The harus- picatio was performed not by Roman but Hunnic diviners. Litorius’ Huns did before Toulouse what twelve years later Attila, “a man who sought counsel of omens in all warfare,” did on the eve of the battle at the locus Mauriacus: He “decided to inquire into the future through haruspices.”[1626] In the ninth century the Bulgars, before a battle, “used to practice enchantments and jests and charms and certain auguries” (exercere incan- tationcs ct ioca cl carmina cl nonnulla auguria).”[1627]

The Hunnic diviners are also mentioned by Priscus. At the banquet at Attila’s court he noticed that the king pinched Ernach’s cheeks and looked on him with serene eyes. Priscus was surprised that Attila should take small account of his other sons but give attention to this one. He learned from a Latin-speaking Hun that the seers [of pavTscq] had prophesied to Attila that his genos would fail but would be restored by this son.[1628] We may assume that the seers were Prosper’s and Jordanes’ haruspices.

The Hunnic diviners may, at the same time, have been shamans. The shamans of the Turkish tribes in the Altai, informed by their guiding spirits, occasionally foretell the future;[1629] they are also more experienced in the interpretation of naturally occuring omina, but they have no monopoly on what Bawden calls “involuntary divination.”[1630] The same is true for the Buryat shamans.[1631] In the middle period the Mongols distinguished between seers and shamans.[1632] But as so often in these studies, we have to resist the temptation to conclude from the customs and practices of later steppe peoples what those of the Huns were or may have been. That the Huns had shamans is certain. Kam in the names Atakam and Eskam is qam, the common Turkish word for shaman.[1633] To judge from the two names of high-ranking Huns, the shamans seem to have belonged to the upper stratum of Hun society. Malalas’ ieqelq were possibly shamans.

The Hunnic method of deliberate foreknowing was scapulimancy.[1634] Attila’s haruspices “examined the entrails of cattle and certain streaks in the bones that had been scraped.”[1635] From Eisenberger’s excellent monograph[1636] we have learned to distinguish between two forms of this method of prognostication. In the “Asiatic” form the bones, mostly the shoulder bones of sheep, after having been carefully scraped clean, are exposed to fire: The fissures caused by the heat are then “read.” In the “European,” supposedly more primitive form, the bones are “read” as they are.

Because Jordanes does not state whether the bones were scorched or not, Eisenberger does not dare to decide whether the Huns practiced the “Asiatic” or “European” form of scapulimancy. The latter he traced back to a Stone Age hunter culture. He may be right. In any case, the “European” scapulimancy is attested only many centuries after Attila[1637]. No ancient writer knows about it. It was unknown to the Alans.[1638] Nothing in the earlier Sarmatian graves indicates that scapulimancy was ever practiced. Only in the Sarmatoid cemeteries at Vrevskii, south-west of Tashkent,[1639] and Lavyandak, near Bukhara,[1640] both of them datable to the last centuries b.c., were shoulder blades of sheep, one of them scorched, found. If they had been used for divination, as Voronets and Obel’chenko think, they would point to an Eastern, non-Iranian element.

The Huns could not have borrowed scapulimancy from their neighbors and subjects in Hungary and the western steppes. In China it had been practiced since pre-Shang times.[1641] The Turkish word for divination, yrq< yryq, means originally “fissure, cracks”; the Mongolian ttilge, “portent,” goes back to tiile, ttili, “to burn.”[1642] There can be no reasonable doubt that the scapulimancy of the Huns was of Eastern origin.

Divine Kingship?

At the dinner which the East Roman ambassador gave to Edecon and his entourage, the Huns lauded Attila and the Romans the emperor. Bigilas, the typical meddlesome Levantine dragoman, “remarked that it was not fair to compare a man and a god, meaning Attila by the man and Theodosius by the god. The Huns grew excited and hot at this remark.”

This passage in Priscus’ report[1643] has been adduced as proof that the Huns regarded Attila as a god. But such an interpretation does not take into account the Roman meaning of “god” when applied to the emperor. As dominus totius mundi he was “God on earth” (deus in terra), not really god; deus was in the fifth century understood as quasi or tamquam deus “For when the emperor has accepted the name Augustus, sincere devotion must be offered to him as if he were God incarnate and present” (Nam imperator cum Augusti nomen accepit, tamquam praesenti et corporali Deo fidelis est praestanda devotio), wrote Vegetius.[1644] Or to quote the sixthcentury Agapetus: “Though an emperor in body is like all others, in power of office he is like God.”[1645] Pacatus could call the good Christian Theodosius a god,[1646] and as late as the eleventh century the Byzantine emperor was “God on earth” (0ed; sjttycto;).[1647] The fugitive Athanaric, overwhelmed by the sight of Constantinople, conceded that the emperor was “truly God on earth,” deus terrenus.[1648] To acknowledge him as such meant acceptance of his claim to be the lord of the world, domitor omnium gentium barbarorum. It was this implication that aroused the ire of the Huns at Bigilas’ remark.

The Roman ambassadors were not allowed to pitch their tents on higher ground than that on which Attila’s tent stood.[1649] G. Staunton’s account of the first English embassy to Ch’ien Lung’s court offers an instructive parallel:

When a splendid chariot intended as a present to the Emperor was unpacked and put together, nothing could be more admired, but it was necessary to give instructions for taking off the box; for when the mandarins found out that so elevated a seat was destined for the coachman who was to drive the horses, they expressed their utmost astonishment that it should be proposed to place any man in a situation above the Emperor. So easily is the delicacy of this people shocked at whatever related to the person of their exalted sovereign.[1650]

Ch’ien Lung was “the son of Heaven,” but he was not a god. Neither was Attila. In his relationship with the Huns, Attila in no way behaved like a divine being. There was none of the elaborate ceremony which stressed the distance between the god-like basileus and his subjects, not to speak of the abyss that separated the Sasanian king of kings from ordinary mortals.[1651] Attila wore neither a diadem nor a crown; his dress was plain; his sword, the clasps of his shoes, and the bridle of his horse were not, like those of the Hunnic nobles, adorned with gold and gems. He drank from a wooden goblet and ate from a wooden plate.[1652] With only his bodyguard standing by, Attila, in front of his house, listened to the disputes of his Huns and arbitrated their quarrels.[1653] The most Attila claimed for himself was that he was well-born.[1654] In the dirge sung at his funeral the dead king was praised as a great conqueror, not worshipped as a god.

On the other hand, Kuridach, king of the Hunnic Acatziri, refused to come to Attila’s court because, he said, it was difficult to face a god. “If it be impossible to look upon the orb of the sun, how could one behold the greatest of the gods [piytaxov xu>v 0£c5v] without injury?”[1655] To be sure, Kuridach may have used this language merely because he feared a trap and hoped, by flattering the terrible king, to save himself. Still, this does not seem sufficient to explain his hyperbolic comparison of Attila with the sun.

In the late Roman Empire the ruler was often compared and even equated with the sun. The inscription on an equestrian statue of Theodosius I reads:

You lept up from the East, another light-bearing Sun, 0 Theodosius, for mortals in the midst of heaven, 0 gentle-hearted one, with the ocean at your feet and the boundless earth.[1656]

But the emperor was a mild, not a fierce and blinding sun. Kuridach’s simile reminds one rather of Indian expressions. The great bowman Bhishma looked “like the all-consuming sun himself, incapable of being looked at like the sun when in his course he reaches the meridian and scorches everything underneath.”[1657] However, there is nothing else that could connect the Acatziri with India.

The titles and epithets of the Hsiung-nu kings and the rulers of the Orkhon Turks offer no parallels to Kuridach’s word either. Ch’eng li ku fu, the title of the Hsiung-nu king as given in the Han shu,[1658] has been explained as tayri qut, “heavenly majesty.” Ch’eng li — d’vng Iji is undoubtedly tayri, “heaven, god.” Pan Ku states expressly that this is the meaning of the word in the Hsiung-nu language. Ku fu, he says, means “son.”[1659] Shiratori’s etymology of ku t’u, which he took for a Tungus word for “son,” may be unconvincing, but it is consistent with the text. F. W. K. Muller,[1660] followed by A. von Gabain,[1661] rejected Pan Ku’s translation; convinced that the Hsiung-nu spoke a Turkish language, he maintained that ku t’u could be only Turkish qut. Muller was certainly wrong. Why should the Chinese have transcribed qut by two characters and not, as they did in T’ang times, by ku<kust? If, however, the older form of qut was bisyllabic, it was probably *qaivut.[1662] By adopting the Chinese title fieri tzu, the shan-yii proclaimed himself the equal of the emperor.[1663] There leads no way from “son of heaven” to Kuridach’s “greatest of the gods.”

In the Orkhon inscription the kagan is given his power by Tayri; he fulfills the mandate of Tayri. But he is not Tayri himself, and he is never compared with the sun.[1664] In the following centuries the epithet tayri became quite common. In the eighth century a ruler of the eastern Turks,[1665] in the ninth century an Uigur king[1666] called themselves tayri qayan. The %atun is tanri quniuy, “the divine princess”;[1667] tayrim, “my god,” means “princess.”[1668] The Uigur king is tayri qan or tayri ilig, “divine king,”[1669] but he can also call himself tayri without any additions.[1670] In the confessions of the lay sister Utrat are named tai%an %an, kiimsa, %atun tayrim, misan, %an, (ai§i wang bag,“ and the other tayri”[1671] In all these titles the meaning of tayri fluctuates between “god” and “majesty,” exactly like in bayan in Middle Persian.[1672] However, Buddha is “the tayri of the tayri,”[1673] and Mani “the greatest tayri,“[1674] which corresponds exactly to Kuridach’s neyiatoz tmv Oeojv. It seems that the Uigurs and also other Turks borrowed, though in an attenuated from, the concept of divine kingship from the Persians.[1675]

Kuridach’s words have a decidedly Persian ring. The deification of the Persian monarch began under the first Darius and persisted throughont the Parthian and Sasanian periods. Shapur I was 0id? and “of divine descent” (ex yivovQ Oecov).[1676] Bahram II was a god.[1677] Chosroes called himself “the divinity who takes his form from the gods” (0eio; J; ex Oemv %a- QaxTTjQl^ETai).[1678] The Sasanian king, with his head crowned with rays, appears in the guise of the sun, radiato capite solis in figura.[1679] When taking his seat on the royal throne the ruler had his face veiled. According to the New Persian court ceremonial, it was imperative when entering before the shah to cover one’s face with one’s hands, exclaiming at the same time, “misuzam, I am burning up !”[1680]

Although to his Huns Attila was most certainly not a divine being,[1681] the Acatziri, particularly after they had been forced to acknowledge him as their supreme lord, looked up to Attila as, in the same time, the Persians looked up to their king.


When Attila died, the Huns “as it is the custom of that race, cut off a part of their hair and disfigured their faces horribly with deep wounds, so that the gallant warrior should be mourned not with the lamentations and tears of women, but with the blood of men.”[1682] Sidonius had the Huns in mind when he wrote about the peoples “to whom wailing means selfwounding and tearing the cheeks with iron and gouging the red traces of scars on the threatening face.”[1683] A line in Kalidasa’s Raghuvamsa alludes to the same custom among the Huna on the Oxus: “The exploits of Raghu, whose valor expressed itself among the husbands of the Huna women, became manifest in the scarlet color of their cheeks.”[1684]

Slashing or scratching the face as an expression of mourning was so widespread[1685] that only a few parallels to the Hunnic custom need to be adduced: The Kutrigur cut their cheeks with daggers;[1686] the Turks cut off their hair and slashed their ears and cheeks;[1687] so did the Magyars[1688] and Slavs;[1689] on a wall painting in Pandzhikent, patterned on Parinirvana scenes,[1690] the mourners are shown cutting their cheeks with knives.[1691] Until quite recently the custom was observed by Serbs and Albanians[1692] and in some regions in Tadjikistan.[1693] The lines on a gold mask found at the Shami Pass in the Chu Valley, datable to the fourth or fifth century, might represent scars.[1694]

The Getica account of Attila’s obsequies, going back to Priscus,[1695] reads:

His body was placed in the midst of a plain and laid in state in a silken tent as a sight for men’s admiration. The best horsemen of the entire tribe of the Huns rode around in circles, after the manner of circus games, in the place to which he had been brought and told of his deeds in a funeral dirge in the following manner:

Here lies Attila, the great king of the Huns,

the son of Mundzucus,

the ruler of the most courageous tribes;

enjoying such power as had been unheard of before him, he possessed the Scythian and Germanic kingdoms alone and also terrorized both empires of the Roman world after conquering their cities, and placated by their entreaties

that the rest might not be laid open to plunder he accepted an annual tribute.

After he had achieved all this with great success

he died, not of an enemy’s wound, not betrayed by friends, in the midst of his unscathed people, happy and gay, without any feeling of pain.

Who therefore would think that this was death which nobody considers to demand revenge? (Praecipuus Hunnorum rex Attila, patre genitus Mundzuco, fortissimarum gentium dominus,

qui inaudita ante se potentia solus Scythica et Germanica regna possedit nec non utraque Romani orbis imperia captis civitatibus terruit, et ne praedae reliqua subderentur, placatus praecibus annuum vectigal accepit: cumque haec omnia proventu felicitatis egerit, non vulnere hostium, non fraude suorum, sed genie incolumi inter gaudia laetus sine sensu do loris occubuit.

quis ergo hunc exitum pulet, quern nullus aestimat vindicandwn?)[1696]

When they had mourned him with such lamentations, a strava, as they call it, was celebrated over his tomb with great reveling. They connected opposites and showed them, mixing grief over the dead with joy.[1697] Then in the secrecy of night they buried the body in the earth. They bound his coffins, the first with gold, the second with silver, and the third with the strength of iron, showing by such means that these three things suited the mightiest of kings: iron because he subdued the nations, gold and silver because he received the honors of both empires. They also added the arms of foemen won in the fight, trappings of rare worth, sparkling with various gems, and ornaments of all sorts whereby princely state is maintained. And that so great riches might be kept from human curiosity, they slew those appointed to the work—a dreadful pay for their labor; and thus sudden death was the lot of those who buried him as well as of him who was buried.

Priscus’ source is not known. If, as I am inclined to assume, strava is a Slavic word, his informant may have been an escaped prisoner of war, or one of those “Huns” who, after 453, took service in the East Roman army. Priscus may have heard the dirge from a Goth to whom it was translated from Hunnish and who rendered it into Greek. The song was translated at least once, probably twice, and possibly three times before Cassiodorus or Jordanes gave it in the present form. The “reconstructions” of the supposed Gothic text[1698] from a version so far removed from the original are as fanciful as attempts to discover in it the Weltanschauung of the ancient Turks.

Mommsen praised the beauty of the song.[1699] In the desert of Jordanes’ prose it is certainly an oasis, but a small one. The Huns’ boast that their king extorted so much money from the Romans might be genuine: It sounds, mutatis mutandis, like an epitaph for an American gangster of the prohibition era. The rest are banalities. Not a Hun poet but Cassiodorus-Jordanes called Attila king of kings, rex omnium regum-[1700] one thinks, of course, of the title of the Persian kings, but as early as the first century b.c. Pharnaces, ruler of the Bosporus, called himself PaaiXevg fiaaiMcov.[1701] Attila was “the lord of all Huns, and of the tribes of nearly all Scythia, he was the sole ruler in the world” (Hunnorum omnium dominus et paene totius Scythiae gentium solus in mundo regnatofy[1702] however, Ermanaric, a century before him, was also the ruler of “all the nations of Scythia and Germania.”[1703] Non fraude suorum almost sounds as if it were taken from Ammianus XXV, 3, 20, where Emperor Julian on his deathbed thanks the godhead that he does not die clandestinis insidiis. Subdere for subicere[1704] and the spelling Mundzuco incidate that Jordanes made some changes in Cassiodorus’ text. All in all, the song throws a very dim light on the poetry of the Huns.

The authenticity of Priscus’ account of the funeral rites cannot be doubted, although in the shortened version of the Getica various themes seem to be telescoped. It is hard to imagine how the horsemen could ride around the tent and sing at the same time. In modum circensium cursibus may in the original have referred to horse races which so often and among the most different peoples are connected with burials.[1705] Here and there a feature seems to have been misunderstood or misinterpreted. It is a little strange that Attila’s coffin should have been covered with gold, silver, and iron like the walls of the inner sanctum of the Serapaeum in Alexandria.[1706] To kill the laborers who buried the king was an inefficient means to prevent the robbing of the tomb for thousands must have known of it. Besides, who killed the killers? The slaughter was probably a sacrificial act, comparable to the killing of prisoners after the death of Silzibulos.[1707]

The Hunnic rites must have reminded Priscus of similar ones he knew from his Homer.[1708] The Thracians, reported Herodotus (V, 8), “lay out the dead for three days, then, after killing all kinds of victims and first making lamentations, they feast; after that they make away with the body either by fire or else by burial in the earth, and when they have built a barrow they set on foot all kinds of contests.” Priscus may have heard of the drinking bouts and horse races with which the Othtrysae honored the dead.[1709] The association of burial and games is known from Greece to the Nicobar Islands, and from the Bedouins in the Sinai Peninsula to the Bashkirs on the Volga.[1710]

There is nothing in the Hunnic rites to which analogies could not be found throughout Eurasia. The assertion that the Huns buried their dead like the Goths[1711] is as unfounded as the opposite statement that no part of the rites at Attila’s funeral can be claimed as Germanic. The mixture of grief with joy is well attested for both Germans[1712] and non-Germans. The hunt for parallels is futile, and the assumption that the Huns buried their kings after Gothic, Sarmatian, Slavic, or any other but Hunnic fashion is untenable. Attila’s shade, says Jordanes, was honored “by his tribe” (a sua gente), “as is the custom of that tribe” (ut gentis illius mos est). The strava, which the Huns celebrated over his tomb with great revelry, was a Hunnic custom.

The Sacred Sword

The Huns are said to have worshipped a sacred sword. At Attila’s court the East Roman ambassadors were told the following story:

When a certain shepherd beheld one heifer of his flock limping and could find no cause of this wound, he anxiously followed the trail of blood and at length came to a sword it had unwittingly trampled while nibbling the grass. He dug it up and took it straight to Attila. The king rejoiced at this gift, and, being ambitious, thought he had been appointed ruler of the whole world, and that through the sword of Mars supremacy in all wars was assured to him.[1713]

Jordanes read this story in Cassiodorus, whose source was Priscus.[1714] Cassiodorus shortened the passage but not as much as we have it now in the Constantinian excerpts. There it is compressed into a pale “discovered through the agency of an ox,” but the sword itself is more exactly described than in the Getica. It was “sacred and honored among the Scythian kings, dedicated to the overseer of wars. It had vanished in ancient times.”[1715]

All this sounds like the combination of a folktale, transferred on Attila, and Herodotus IV, 62: “The Scythians worship Ares in the form of an acinaces [a scimitar-Ed.], set up on a platform of bundles of brushwood.”[1716] Herodotus’ statement, with slight variations, has often been repeated. It occurs in Eudoxius of Cnidos,[1717] Apollodorus,[1718] Mela,[1719] Lucian,[1720] Solinus,[1721] and, cited from secondhand sources, in the writings of Christian apologists.[1722]

Occasionally newer tribes took the place of the Scythians. Hicesius ascribed the worship of the sacred sword to the Sauromatae,[1723] Dionysius to the Maeotians,[1724] and Ammianus, in a passage in which he is not above the suspicion of having followed the styli veteres, to the Alans (they “fix a naked sword in the ground and reverently worship it as Mars, the presiding deity of those lands over which they range”).[1725] If, however, Ammianus should actually have referred to the Alans of his time, it could be argued that the Huns had taken over an old Iranian cult.

On the other hand, the Hsiung-nu of the Han period likewise worshipped a sword:[1726] The ching-lu was both a sword, tao, and a god, shen, to whom prisoners of war were sacrificed in the same way as to the Scythian Ares- acinaces.[1727] Besides, at least three more “Altaic” peoples held the sword so sacred that they swore by it. The Avar kagan took an oath after the manner of his people on his drawn sword,[1728] the Bulgars swore on their swords,[1729] and Suleiman the Great, undoubtedly following an old Turkish custom, took an oath on his sword.[1730]

But there were more, neither Iranian nor Altaic, peoples for whom the worship of the sword is attested. The Quadi, “drawing their swords, which they venerate as gods, swore that they would remain loyal.”[1731] The Franks swore by their swords.[1732] The warriors in ancient India worshipped their swords.[1733]

In spite of the literary overtone, we may believe Priscus: Like so many peoples, from Mongolia to Gaul, the Huns worshipped the god of war in the form of a sword. The origin of the cult cannot be determined.

Masks and Amulets

Like the Germans and the Celts in the West,[1734] the nomads of the eastern steppes took a fancy to the frontal representations of heads, an old and
widespread motif in the higher civilizations to the south in direct or indirect contact with the barbarians.[1735] Masklike heads occur on horse trappings, stamped silver and bronze sheets, in Hunnic burials as well as in those which in one way or another indicate a Hunnic milieu; in Szentes- Nagyhegy[1736] and Pecs-Uszog[1737] in Hungary, Novo-Grigor’evka in the southern Ukraine,[1738] and Pokrovsk-Voskhod (fig. 13)[1739] and Pokrovsk kurgan 17 (fig. 14)[1740] and 18[1741] on the lower Volga.

Fig. 13. Mask-like human heads stamped on gold sheet from a Hunnic burial at Pokrovsk-Voskhod. From Sinitsyn 1936, fig. 4.
Fig. 14. Mask-like human heads stamped on silver sheet on a bronze pha- lera from kurgan 17, Pokrovsk. From Minaeva 1927, pl. 2:11.

The wooden and leather masks in the kurgans at Pazyryk in the High Altai, datable to the fourth century b.c., still reveal their foreign prototypes; some are derived from the head of Bes,[1742] others betray in the palmette on top their origin in Greek art.[1743] The gradual simplification of the heads, their progressive coarsening, whether they were Sileni, Negroes, Gorgoneia, the heads of Hercules or Dionysus,[1744] led independently both in the East and West to similar results.[1745] On early Celtic masks the hair is done in vertical strokes covering the forehead down to the eyebrows[1746] as it is on the masks from Intercisa, Szentes-Nagyhegy, and Pokrovsk-Voskhod.[1747] The masks may have been carriers of apotropaic powers, could (pars pro toto) have stood for god or demons, or may have been merely decorative. The juxtaposition and superimposition of the masks probably had no meaning: They seem to result from the technique of stamping thin metal sheets.[1748]

Some of the Hunnic or probably Hunnic masks are of Iranian origin. The Huns, said Ammianus Marcellinus, looked like eunuchs. He exaggerated, as usual. But their thin beards also struck the observant Priscus.[1749] The masks from Pecs-Uszbg, Pokrovsk 17 (fig. 14), and Pokrovsk-Voskhod (fig. 13), with their luxuriant beards, cannot represent Huns, or their gods.

Fig. 15. The representation of the head of a Scythian in clay from Transcaucasia. Photo courtesy State Historical Museum, Moscow.

Although most masks of the barbarians are mechanical and increasingly debased replicas of motifs unintelligible to their makers, occasionally one finds new and unexpected features in them, apparently attempts to make them more like the people who used them. The just-mentioned masks, with their shaved upper lip and the fan-shaped beard, render a fashion which at one time was current among Eurasian nomads. The head of a “Scythian” in the Historical Museum in Moscow (fig. 15), found in Transcaucasia,[1750] makes it probable that originally it was the fashion of Iranian tribes.[1751]

To Iranians point also the curious bronze mountings on a wooden casket from Intercisa on the Danube, south of Aquincum-Budapest (fig. 16).[1752] They sometimes have been claimed for the Huns. Radnoti, on the other hand, comparing the heads, or masks, of the figures with those on Germanic buckles, takes the mountings for Germanic; he dates them tentatively to the middle of the fifth century.[1753] Now it is true that Germans settled, at one time or another, near Intercisa; however, the masks of the Ostrogothic buckles,[1754] which, by the way, belong to the last third of the fifth century,[1755] are quite different from those on the Intercisa casket. Apart from the higher relief, the mountings are technically identical with the many late Roman pieces found both on the Danube and the Rhine.[1756] The man who made them must have been a Roman or a barbarian using Roman techniques. But this is of minor importance, for the figures on the mountings are as un-Roman as possible. The heads are distantly related to the mask-heads from Hunnic burials, but the impressive mustache occurs on none of them. It has some resemblance to the mustache of Turkish stone figures in Southern Siberia and Mongolia,[1757] or on Sasanian silver phalerae.[1758] But neither the Turkish nor the Persian heads have the luxuriant beards of the Intercisa figures, a feature which also rules out the Huns. This leaves only one possibility: The figures must be Sarmatian.

Fig. 16. Bronze mountings from a wooden casket from Intercisa on the Danube. From Paulovics, AT?, 1940.

Their meaning is obscure. Still, perhaps we may make a guess. The two standing figures cannot be the images of mortal women. The finds in Sarmatian graves, from the earliest to the latest, prove that the Sarmatian women did not bare their breasts like the Intercisa figures; they wore shirts leaving only the neck free. Besides, the lozenges on the figures emphazise the genital region in the strangest way. In his Aletheia, Claudius Marius Victor of Marseille, who died about 425,[1759] says that the Alans worshipped their ancestors.[1760] I am inclined to assume that the women on the casket from Intercisa represent Sarmatian maires.

To Sarmatians lead likewise flat bronze amulets, angular figures of men and women, which Kruglikova took for Hunnic.[1761] The women with marked breasts and the ithyphallic men (fig. 17) were evidently meant to avert evil. Such amulets were found in the Crimea (Chersonese, Panti- capaeum, Tyritace), on the Kuban (stanitsa Pashkovskaya, cemetery 3) and in the northern Caucasus (Kumul’ta, Kamunta, Aibazovskoe).[1762] The grave of a child at the Syuyur Tash on the Azov Sea, in which such an amulet was found, contained a fabric with a quotation from the New Testament and, written on it, the date: 602 of the Bosporan era=305 a.d.[1763] It is, therefore, pre-Hunnic.[1764]

Fig. 17. Flat bronze amulet in the shape of an ithyphallic human figure of Sarmatian type. (Source not indicated in the manuscript.—Ed.)

Although from the masks little, if anything, can be learned about the religion of the Huns, some of them point to apparently early contacts between Huns and Iranian tribes, presumably Sarmatians.


Between 452 and 458, some of the Roman prisoners of war in the country of the Huns “by hunger and terror” were forced to eat sacrificial food. Who forced them? The Goths. It is true that Athanaric, iudex of the Visigoths, ordered that the people who were suspect of being Christian worship a wooden figure[1765] and make sacrifices to it.[1766] But that was in the 370’s. By the middle of the fifth century, most Visigoths and Ostrogoths were Christians, not always very devout ones, but definitely no longer fanatical pagans. The date of the conversion of the Gepids is controversial. Thompson thinks it improbable that they were baptized when they were still under Hunnic rule; he even suspects that their “most savage rites,” of which Salvian wrote in the early 440’s, may have been human sacrifices.[1767] Schmidt’s assumption that the Gepids embraced Christianity under King Ardarich[1768] has no textual support. A Gepidic nobleman who died about 480 wore a finger ring with a cross on it but was buried with pagan rites. As late as 580, when the Langobards had long been Arian Christians, it could happen that forty Italian peasants who refused sacrificial meat were slaughtered.[1769] However, after the battle on the Nedao, the Gepids, Christians and pagans, and the Huns no longer lived together.

This leaves the Alans and Huns. We seem to learn something about Hunnic sacrifices from a short passage in the Getica which probably goes back to Priscus: When the Huns first entered Scythia, they sacrificed to victory, litauere victoriae, as many as they captured.[1770] This is the only time the Huns were accused of having sacrificed their prisoners. In Attila’s time, and also before him, those captives who could not be sold or who were not ransomed were kept as domestic slaves. Priscus apparently transferred a Germanic custom, of which he knew from literature, to the Huns.[1771] But the Huns may have sacrificed animals to their gods. Did they worship gods in human or animal form ?

Throughout northern Eurasia, from Lapland to Korea, the figures of the shamanistic pantheon, in particular the shaman’s “helpers,” were represented in various ways: painted on drums; cut out of felt; cast in bronze and iron and attached to the shaman’s coat; carved out of wood and put up in the tent or glued to the drum.[1772] The shamanistic Huns, too, may have had eidola (I avoid the missionary term “idols”). There is, indeed, both literary and archaeological, though circumstantial, evidence of their existence.

According to Malalas, Gordas, prince of the Huns near Bosporus in the Crimea, was baptized in Constantinople in the first year of Justinian’s reign, 527–528. After his return to his country, he ordered the dydfytava, made of gold and electrum, to be melted down; the metal was exchanged for Byzantine money in Bosporus. Incensed at the sacrilege, the priests, in connivance with Muageris, Gordas’ brother, put the prince to death.[1773]

There is no reason to doubt Malalas’ account. Besides, the statement that the figures were of gold and electrum, while the cliche would call for gold and silver, speaks in favor of the story. It does, of course, not prove that the Attilanic Huns, too, had figures of their gods made of precious metals. But the possibility cannot be ruled out, certainly not because of the low level of Hun metal work.[1774] The impressive bronze horseman from Issyk in Kazakhstan, datable to the fifth or fourth century b.c.,[1775] shows the skill of metalworkers in the early nomadic societies of Eurasia. The Hsiung-nu had their “metal men,”[1776] and the silver figures at the court of the Turk Silzibulos greatly impressed the Byzantine ambassador.[1777] The common Hunnic eidola—provided that they did exist—were probably much more like those of the Sarmatians, about which we are fairly well informed.

The earliest one is of sandstone, about one meter high, a pillar rectangular in cross section, except the upper part, which is rounded to represent the head (fig. 18); it was found in kurgan 16 at Tri Brata near Elista[1778] in the Kalmuk steppe.[1779] The arrowheads date the grave to the fifth century b.c.[1780] Smirnov lists a similar stone figure from Berdinskaya Gora near Orenburg and two from the trans-Volga steppes which, however, stand closer to the well-known kamennye baby, “stone women.” Two more eidola from the lower Don, stone slabs showing human figures in silhouette, may be somewhat later.[1781]

Fig. 18. Sandstone pillar in the shape of a human head from kurgan 16 at Tri Brata near Elista in the Kalmuk steppe. (Height 1 m.) From Sinitsyn 1956b, fig. 11.

Smirnov assumed that these Sarmatian figures were put up on, or near, burial mounds as representations of local gods or deified ancestors. Their similarity to the silhouette stone slabs from the Bosporan kingdom, dating from Hellenistic to Roman times, speaks for the latter interpretation; the Bosporan figures, some of them with the name of the dead written on them,[1782] are doubtless tombstones.

From the Early Sarmatian period two chalk eidola are known,[1783] both about 13 centimeters high, too small to be erected on a kurgan or on the ground. The one from Bliznetsy, west of Ak-Bulak in the province Orenburg, is a human figure in the round, so crude that not even the sex can be determined; the other one from Zaplavnoe between Volgograd and Elista, one or two centuries earlier, is a slab with the merest indication of the head.

In the Middle Sarmatian period, eidola were made over a wide territory. In the grave of a young woman in kurgan 5/3, in the burial ground at Bykovo on the Volga, oblast Volgograd, a crude chalk figure was found, 8 centimeters high, head, shoulders, and legs barely indicated .[1784] Four even cruder eidola from the Kuban area probably are to be dated to the early first century a.d.; one was found at Krasnodar, three at Elizavetskaya stanitsa.[1785]

In a sacrificial pit at Neapolis near Simferopol in the Crimea lay unburnt clay figures: the head and neck of a ram, the fragment of a human torso, and two coarsely modeled heads;[1786] the building near the pit was destroyed about 200 a.d., the beginning of the Late Sarmatian period when numerous elements of Sarmatian civilization began to appear in the late Scythian civilization of Neapolis. Of about the same time is the clay figure of a seated woman with a hollowed head, 7 centimeters high, found in the town site Zolotaya Balka on the lower Dnieper.[1787] Clay figures of the Late Sarmatian period were found in small rural settlements on the periphery of the Bosporan kingdom: The terracottas from Semenovka represent women;[1788] a female torso and a head were excavated at Mysovka,[1789] and another head at Tasunovo.[1790] A limestone figure, 9.5 centimeters high, 3.5. centimeters across the shoulders, comes from a kurgan at Perezdnaya in the uezd Bakhmut, gubernie Ekaterinoslav. It represents a woman with what looks like a vessel in her hands, the body apparently bare, the the head covered. Veselovsky took it for pre-Mycenean; Gorodtsov dated it rightly to the second or third century.[1791]

Two chalk eidola have come to light from Alanic graves of the fifth century a.d. at Baital Chapkan in Cherkessia.[1792] One is round in cross section, modeled on one side only, the shoulders being indicated by round projections (fig. 19); the other eidolon is merely a cone, somewhat wider in the upper part.

This list is incomplete. Many Sarmatian eidola mentioned in excavation reports are neither properly described nor properly illustrated. A few examples follow: a piece of wood with a human head in a kurgan at Susly in the former German Volga Republic;[1793] two stone “stelae” in a cemetery at Zemetnoe near Bakhchisarai in the Crimea;[1794] wooden statues, 56 inches high, in a barrow in the former okrug Sal’sk, southeast of Rostov;[1795] an anthropomorphic copper figure in a kurgan between Kapustin and Po- gromnoe at the border of the oblasts Astrakhan and Volgograd.[1796]

Fig. 19. Chalk eidola from an Alanic grave at Baital Chapkan in Cher- kessia, fifth century a.d. From Minaeva 1956, fig. 12.

Some of the small terracotta, lead, and copper figures in Sarmatian graves in the Kuban area, excavated by Veselovsky, but never published,[1797] may have been dolls. A small bronze figure in a Late Sarmatian grave at Ust’-Kamenka, district Apostolovo, oblast Dnepropetrovsk,[1798] might also be a doll; its leather belt, with a bow at the back, is well preserved; the absence of a loop indicates that the statuette was not carried around the neck as an amulet. The silver figure of a mustachioed man in a short coat found in a grave in the cemetery at Novo-Turbasly near Ufa,[1799] datable to the fourth or fifth century, had a loop at the back.

Minaeva compared the Alanic eidola from Cherkessia with the pieces of chalk in Late Sarmatian graves which for a long time have claimed the attention of Soviet archaeologists. Rykov[1800] and Ran[1801] attributed to them ritual significance without attempting to define it; Grakov[1802] and K. F. Smirnov[1803] think the white chalk symbolizes purity: the pieces of chalk meant to purify the corpse. This is an attractive suggestion which may be valid in some cases but does not account for all. In the Early Sarmatian cemeteries at Berezhnovka and Molchanovka, no pieces of chalk were found, but many of realgar. The same is true for the Don region.[1804] The orange-red realgar cannot very well stand for purity. From most excavation reports, one gets the impression that the lumps of clay were just thrown into the grave pit. However, there are exceptions. In Susly, kurgan 35, in the grave of a woman with a deformed skull, the chalk lay in a small, round vessel with a hole in its side.[1805] In the Late Sarmatian graves at Ust’-Labinskaya the pieces were carefully placed next to clay vessels; one was in a bowl and five were in pitchers, intentionally kept away from the corpses they were allegedly to purify.[1806] It seems that it was rather the shape of the chalk pieces than their color that counted. Many seem to be merely irregularly shaped cones and pyramids, but others had been worked over. The piece in kurgan 8/3 in Susly looks like the cocoon of a silkworm.[1807] In the Late Sarmatian grave of a woman, in Focsani in Rumania, lay a rather remarkable “piece of chalk” (fig. 20).[1808] Almost 12 centimeters high, it represents a human being: the round line of the chin separates the head from the body; eyebrows, pupils, nose, and mouth are crudely but unmistakingly rendered.

So far, no sandstone or chalk eidola have been found in Hungary. In view of the very small number of Alanic graves in the Danube basin, this is not surprising. A curious find proves the identity of the religion of the Alans in Hunnic Hungary and Cherkessia. At Fiizesbonyban, a cone-shaped cavity, lined with polished clay, contained a horse skull.[1809] There was no cemetery nearby; nothing similar is known from Hungary. But in Cherkessia, in Baital Chapkan and Atsiyukh, three such small “graves” with only the skull and the fore- and hindlegs of a horse have been found, again unconnected with other burials.[1810] If the Alans in Cher- kessia put eidola in their graves, those in Hungary almost certainly did the same.

Fig. 20. Chalk figure from a Late Sarmatian grave in Focgani, Rumania. (Height ca. 12 cm.) From Morintz 1959, fig. 7.

The Alans in Hungary stayed as pagan until the end of the Hunnic kingdom as those who in the beginning of the fifth century moved to Gaul. About 440, Salvian of Marseilles spoke about the greedy pagan Alans.[1811] In the sixth century a few Alans in Gaul were Christians. We hear of St. Goar from Aquitania whose parents, Georgius and Valeria, had already been baptized;[1812] they apparently had left their compatriots and moved into a Roman milieu which, however, did not prevent them from giving their son the pagan Alanic name, Goar. In the second half of the sixth century, Venantius Fortunatus named the Alans among the peoples who worshipped the Virgin, but the list (Ethiopians, Thracians, Arabs, Dacians, Alans, Persians, and Brittons)[1813] is patterned on an old cliche and without value. In an inscription in Spain, St. Martin is praised for converting the Alans;[1814] there, too, they are among the same exotic peoples as in Ve- nantius Fortunatus. In any case, by the middle of the fifth century, the Alans in Gaul were still pagans. Their king, Goachar (Goar), rex ferocissi- mus, was idolorum minister.[1815] If this is not a conventional phrase, Goachar’s eidola were probably not different in shape from those in the Sarmatian graves in the East, though possibly bigger.

In his admirable study of the Sauromatian cult objects, K. F. Smirnov assumes that the small chalk eidola in the burials were replicas of large stone statues like the one in kurgan 16 at Tri Brata.[1816] He lists more of its kind, unfortunately mostly undatable. Still, one needs only to compare the piece of chalk from Foc§ani with the stone figure from Tri Brata to see that the main, if not the only, difference between them is their size. The same is true for a stone figure found at khutor Karnaukhova near ancient Sarkel on the lower Don[1817] and a small clay statue, a pyramid with a round head from Znamenka south of Nikopol on the lower Dnieper.[1818] Both are Sarmatian. Had the eidola which Muageris melted down been of small size, he would not have received more than a few solidi when he exchanged the metal for Byzantine money. This speaks for the assumption that, in analogy with the Sarmatian custom, the Huns in the Crimea, and and not only there, also had small eidola. This seems to be borne out by two eidola from Altyn Asar in ancient Khwarezm.[1819] They are of unburnt clay, the one 8 centimeters high and the other 4 centimeters high. The upper strata of the lower horizon in the “Big House” are datable to the third or fourth century.[1820] The eidola belong to the same Hunnoid civilization as the bone lamellae and the clay cauldrons from Altyn-asar. The extremely crudely modeled eyes, nose, and mouth are barely indicated by dots and strokes. The small clay cauldrons from Altyn-asar are, as we saw, replicas of bigger copper cauldrons. Therefore, we may conjecture that the eidola from Altyn-asar stand likewise for bigger ones worshipped by the Hunnoid population in Khwarezm in the third or fourth century.

In her analysis of the pottery from Altyn Asar, Levina found numerous parallels to the Late Sarmatian civilization on the lower Volga and to the west of the river, but neither she nor Tolstov noticed that one eidolon has a typically Sarmatian tamga cut into the clay. Exactly the same tamga is carved on the side of a stone slab at Zadzrost’ near Ternopol’ in former eastern Galicia (fig. 21 ).[1821] On the front are more tamgas, likewise typically Sarmatian. The slab is no less than 5.5 meters high, and below 1.21, above 1 meter wide. How it got into the northwestern Ukraine, where Sarmatians never lived, is obscure. Some Polish archaeologists took it for a Gothic monument, others saw in it a Turkish kamennaya baba with Runic letters; Drachuk, who discussed it most recently, regards it as a symbol of Sarmatian power. Actually, it is an eidolon, the biggest known so far: the upper part, carefully cut and set off the carelessly cut lower part, represents the head and the neck of the figure. It is in large size what the clay eidolon from Bykovo is in a small size. Similar stone slabs, also with tamgas on them, are known from the Crimea.[1822] I do not dare to decide whether the eidola from Altyn Asar were those of Huns under Sarmatian influence or of Sarmatian under Hunnish influence. Because of the Hunnish cauldron and the bone lamellae, the former seems more likely.

Fig. 21. Stone slab at Zadzrost’, near Ternopol’, former eastern Galicia, marked with a Sarmatian tamga. (Height 5.5 m.) From Drachuk, SA 2, 1967, fig. 1.

The metal, stone, clay, and wooden anthropomorphic sculptures in ancient northern Eurasia must be left to scholars who have access to all museums in the Soviet Union, not just to those in Leningrad and Moscow. A first and promising attempt was made by Davidovich and Litvinskii.[1823] The material presented in the foregoing makes it probable that the Attilanic Huns and their Alanic allies worshipped, next to the sacred sword, also eidola in human form.

VII. Art

Gold Diadems

About 400 a.d., a “leader and king of those most savage Scythians who hold the other side of the Euxine Sea, living on the Maeotis and the Tanais as well as the Bosporus and as far as the Phasis River,” is said to have sent “his crown, covered with gold and set with stones,” to the church of St. Phocas in Sinope.[1824] The “Scythians” were Hunnic tribes, among them, on the Phasis, the Onogur,[1825] and, probably, their Alanic allies. As- terius actually may have seen the crown; it is remarkable in any case that he spoke of a crown covered with, not made of gold, atEyavov... xqvgm TtegdapTio/iEvov. A number of such sumptuous headgears, usually, though not quite correctly, called diadems[1826] have been known for some time. In his BeUrcige Werner discussed them in a special chapter.[1827] Recently three more and the fragment of a fourth, possibly a fifth one, have come to light, and a report on a sixth, now lost, has been published. Their study can now start from a fairly wide base; besides, the circumstances under which the diadems were found are now better known, which, as will be seen, is of some importance for their interpretation.

Before going into details, the two fragments of a gold plaque (fig. 22)[1828] from Kargaly in the district Uzun-Agach, not far from Alma Ata in Kazakhstan, which Bernshtam published,[1829] must be eliminated from the discussion. First, because they were not a part of a diadem. More than 35 centimeters long, straight, not curved, they could not have been worn around the head. Second, their decor has nothing to do with the Hunnic or Alanic diadems. Bernshtam, followed by Werner, admitted strong Chinese influence in the a jour relief but insisted that it still reflected the shamanistic reliefs of its barbarian owners. He thought he could recognize in the diadem a renaissance of Scythian art which had led a subterranean existence in a conservative shamanistic milieu, to come suddenly to the fore around the beginning of our era. Actually the design is purely Chinese. The horse standing on a column is a variety of the quadruped, its feet gathered together on a pole, known not only from Scythian graves in South Russia but also from Perm, Kazakhstan, the Altai, southern Siberia, and the Ordos region.[1830] In China the motif occurs as early as the Chou period.[1831] The winged horse is likewise a well-known Chinese motif which had a great appeal to the northern barbarians; the gold plaque from Noin Ula has often been reproduced;[1832] gilt bronze plaques with winged horses were recently found in Inner Mongolia.[1833] The long-haired genii, hsien jen, have hundreds of parallels on Han stone reliefs, metal work, tiles, lacquers, vases, and textiles. They represent no more the shamanistic gods of the T’ien-shan nomads than the Nereids on a Greek cylix found in South Russia represent the goddesses of the Scythians.[1834]

Fig. 22. Fragment of a gold plaque from Kargaly, Uzun-Agach, near Alma Ata, Kazakhstan. (About 35 cm long.) Photo courtesy Akademiia Nauk Kazakhskoi SSR.

I first list the diadems which were known to Werner.

1. Csorna in western Hungary (fig. 23).[1835] Found on the skull of a north-orientated skeleton. A gold sheet, broken in several pieces, 26.5[1836] (originally about 29) centimeters long, 4 centimeters wide; the edges had been bent around a bronze plaque which has disappeared. Traces of copper oxide on the skull indicate that the diadem was worn without stuffing or leather lining. Garnets and red glass in cloisons.

Fig. 23. Hunnic diadem of gold sheet, originally mounted on a bronze plaque, decorated with garnets and red glass, from Csorna, western Hungary. (Originally about 29 cm long, 4 cm wide.) From Archdologische Funde in Ungarn, 291.

2. Kerch (figs. 24A,B, C).[1837] Said to have been found on the Mithridates Mountain in a grave next to the skeleton of a man with an artificially deformed skull.[1838] Gold sheet over bronze plaques. Except the two big round cells and the lozenge one on the top ornament, which enclose green glass pieces, the 257 cloisons contain flat almandines.[1839]

3. Shipovo, west of Uralsk, northwestern Kazakhstan (fig. 25).[1840] Found on the forehead of a north-orientated skeleton in a wide rectangular pit under a kurgan; 25.2 centimeters long, 3.6 centimeters wide.[1841] Thin bronze sheets over bronze plaques, set with convex glass. The bronze plaques were originally lined with leather and, on it, thin silk; on the latter, small lozenges of gilt leather. The absence of weapons and a clay spin whorl indicate that the dead was a woman. Except for a crescent-shaped golden earring, the other metal objects in the grave were of bronze: buckles, a gold-covered necklace of twisted wire, and another earring. The bronze mirror with a long handle, preserved only in a fragment, is typical of the Middle Sarmatian period (I b.c.-I a.d).[1842]

Figs. 24A-C. Hunnic diadem of gold sheet over bronze plaques decorated with green glass and flat almandines, from Kerch. Photos courtesy Rhei- nisches Museum, Bildarchiv, Cologne.

Fig. 25. Hunnic diadem of thin bronze sheet over bronze plaques set with convex glass from Shipovo, west of Uralsk, northwestern Kazakhstan. From J. Werner 1956, pl. 6:8.

4. Dehler on the Berezovka near Pokrovsk, lower Volga region (fig. 26).[1843] The diadem was on the skull of the skeleton. Bronze plaques covered with gold sheets, which are set with convex almandines. Of the other grave goods, only big amber beads and a mirror were preserved. The mirror is of a type which in the Caucasus occurs from the fifth century a.d. on;[1844] in the West it made its appearance about the same time.[1845] The diadem was probably made about 400 a.d. or a little later.

Fig. 26. Hunnic diadem of gold sheet over bronze plaques set with convex almandines from Dehler on the Berezovka, near Pokrovsk, lower Volga region. From Ebert, RV 13, “Siidrussland,” pl. RV 41 :a.

5. Tiligul (fig. 27).[1846] Formerly in the Diergardt collection, now in the Romisch-Germanisches Zentralmuseum, Mainz. Similar to the diadem from Dehler but technically inferior. The bronze plaques are lost. On the front part, convex almandines; flat triangular and rectangular ones on the side parts. Allegedly none was found in the same grave “at Tiligul,”[1847] which, however, is not the name of a place but a river between the Prut and Dniester and the liman (lagoon) at its mouth.

Fig. 27. Hunnic diadem of gold sheet over bronze plaques (now lost) set with convex almandines, from Tiligul, in the Rdmisch-Germanisches Zentralmuseum, Mainz. From J. Werner 1956, pl. 29:8.

6. Kara-Agach, south of Akmolinsk in central Kazakhstan (fig. 28).[1848] Found near the skull of a skeleton in a stone cist under a kurgan. The bronze circlet, 4 centimeters wide, 49 centimeters in circumference, is covered with a sheet of very pale gold, decorated with stamped triangles in imitation of granulation; fifteen conical “bells” (without clappers) hang from bronze hooks. Among the other finds, there were two gold dragons (fig. 29A),[1849] the ends of a torque, richly decorated with garnets, amber, and mother-of-pearl in cloisons and, in between, triangles in granulation. Skalon published a very similar dragon found in a cemetery at Stavropol, together with many ornaments typical of Sarmato-Alanic graves of the fourth and fifth centuries in the North Caucasus.[1850] The combination of garnets and mother-of-pearl occurs in Con^esti[1851] at the beginning of the fifth century. Emphazising the similarity, in many details amounting to identity, of the dragons from Stavropol and Kara-Agach, Skalon rightly assumes that they were made in the same workshops, probably in Bosporus.[1852] This is also true for a type of earring represented in Kara Agach (fig. 29B).[1853] Such earrings were worn over a very wide area. Their simplest, though not necessarily the original form, without the rings of granules around the inlays, occurs as early as the second or third century; one, with inlays of glass, was found in a rich grave at Usatovo in the lower Volga region.[1854] The earring from a grave at Kotovo (Mozhary), district Kamyshin, province Volgograd, of the same period or perhaps a little later, has the cloisons ringed with granulation.[1855] A coarse version in silver comes from a kurgan at Pokrovsk (fig. 30).[1856] Two such golden earrings with an attached small tube and clusters of granules at the end were in a jug grave in Kalagya in Caucasian Albania (fig. 31).[1857] Similar earrings are among the treasures in the Siberian collection of Peter the Great; they have gold balls and pyramids of granules attached.[1858] Clasps of approximately the shape of the earrings from Kara-Agach, set with semiprecious stones clearly representing a face occur in Sarmatian graves in Checheno- Ingushetia in the northern Caucasus.[1859] All these pieces of jewelry are by their form and technique so closely related that they must have been made by highly skilled goldsmiths who transmitted their craft from generation to generation in one and the same place.

Fig. 28. Bronze circlet covered with gold sheet and decorated with conical “bells” suspended on bronze hooks, from Kara-Agach, south of Akmolinsk, central Kazakhstan. (Circumference 49 cm, width ca. 4 cm.) From J. Werner 1956, pl. 31:2.
Fig. 29A. Terminal of a gold torque in the shape of a dragon, decorated with granulation and cloisonnd garnets, amber, and mother-of-pearl. From Kara-Agach, south of Akmolinsk, central Kazakhstan. From IAK 16, 1905, p. 34, fig. 2.
Fig. 29B. Gold earrings from Kara-Agach, central Kazakhstan. From IAK 16, 1905, fig. 3:a-b.
Fig. 30. Silver earring decorated with almandines and garnets from kurgan 36, SW group, near Pokrovsk. From Sinitsyn 1936, fig. 10.
Fig. 31. Gold earring from Kalagya, Caucasian Albania. From Trever 1959, 167, fig. 18.

The people who buried their dead in Kara-Agach in Kazakhstan were as unable to make the earrings and dragons as they were to make the glass beaker which ... [The manuscript breaks off here in mid-sentence.—Ed.]


The Hunnic cauldrons have long claimed the attention of the archaeologists. In 1896, Reinecke separated a small group of cylindrical or bellshaped bronze vessels, which until then had been classified as Scythian, from the hemispherical cauldrons of South Russia.[1860] His assumption that they go back to Western prototypes, shared by Posta[1861] and Ebert,[1862] proved to be wrong, but he assigned them the right date. Because the bell-shaped cauldron from Jedrzychowice (formerly Hbckricht) was found together with jewelry of the Folkwandering period, Reinecke dated it and, consequently, all similar cauldrons to the first centuries of our era. In 1913, Zoltan Takats (Takacs) published the first of a long series of articles[1863] in which he argued for the Hunnic provenance of the cauldrons. Although Takats at times indulged in wild speculations, in the main he was right, and his views prevailed: both the distribution of the vessels and the context in which they were found leave no doubt that they were cast by Huns for Huns.

Since Werner’s discussion of the cauldrons in 1956,[1864] so many more were found and so much new evidence on the cauldrons of the nomads in Central Asia and the Far East has accumulated that the problems which the Hunnic cauldrons pose call for a reexamination.


In the following list the numerous misspellings and distortions of the names of the findspots have been silently corrected. I did not aim at bibliographical completeness; to refer to the often poor illustrations in old Hungarian publications would serve no useful purpose; those in Japanese works[1865] are taken from Western books and articles.

Fig. 32. Fragment of a bronze lug of a cauldron from BeneSov, near Opava (Troppau), Czechoslovakia. (Height 29 cm, width 22 cm, thickness 1 cm.) From Altschlesien 9, 1940, pl. 14.

1. BeneSov (Bennisch) near Opava (Troppau). Fragment of a lug 29 centimeters high, 22 centimeters wide, up to 1 centimeter thick. Said to be found in a peat bog or on an old road running through a forest; the absence of patina typical of bronze objects found in bogs speaks for the latter. On the outside the cauldron had been exposed to strong fire. Fig. 32.

V. Karger, “Neues zu den Fund- und Erwerbsumstanden des Bronze- kessels von Bennisch-Raase, Bezirk Troppau,” Altschlesien 9, 1940, 112114, pl. 14 (our figure 32). G. Raschke, “Zum Bronzekessel von Raase- Bennisch,” Altschlesien 9, 1940, 114–119. Fettich 1953, 144, n. 47, took the vessel for a poor local imitation; he was certainly wrong.


2. Jedrzychowice (Hockricht), district Oawa, Upper Silesia. Height, 55 centimeters. Fig. 33.

E. Krause, “Der Fund von Hbckricht, Kreis Ohlau,” Schlesiens Vor- zeit in Bild und Schrift, N.F. 3, 1904, 47, fig. 12; Alfoldi 1932, pl. 19:9; Werner 1956, pl. 27:10.

Fig. 33. Hunnic bronze cauldron from Jedrzychowice (Hockricht), Upper Silesia, Poland. (Height 55 cm.) From J. Werner 1956, pl. 27:10.

3. Tortel, county Pest. Height, 89 centimeters, diameter, 50 centimeters. Found at the foot of a burial mound. Fig. 34.

Alfoldi 1932, pl. 18: 2; Fettich 1940, pl. 10, and 1953, pl. 36:1; Archaolo- gische Funde in Ungarn, 293.

Fig. 34. Hunnic bronze cauldron found at the foot of a burial mound at Tortel, Hungary. (Height 89 cm, diam. 50 cm.) From Archaologische Funde in Ungarn, 293.

4. Kurdcsibrak, between Hogyesz and Regoly in the valley of the Kapos River, county Tolna. Height, 52 centimeters; diameter, 33 centimeters; thickness of the wall, 0.8 centimeters; weight, 16 kilograms. Found in a peat bog. Fig. 35.

Fettich 1931, 523; Alfoldi 1932, pl. 18:1; Fettich 1940, pl. 11; and 1953, pl. 36:2.

Fig. 35. Hunnic bronze cauldron found in a peat bog at Kurdcsibrdk, in the Kapos River valley, Hungary. (Height 52 cm, diam. 33 cm, thickness of wall 0.8 cm, weight 16 kg.) From Fettich 1940, pl. 11.

5. Bantapuszta near Varpalota, county Veszprdm. Said to have been found in a marsh. Dimensions not given. Fig. 36.

I understand that the cauldron is bigger than the one from Kurdcsibrak.

Z. Takats, “Neuentdeckte Denkmaler der Hunnen in Ungarn,” Acta Orientalia (Budapest) 9, 1959, 86, fig. 1.

6. Dunaujvaros[1866] (Intercisa), county Feher. Fragment of a wall, found in a late Roman building; fragments of iron helmets were also found.[1867] Fig. 37. Fettich 1931, 524; Alfoldi 1932, 33, fig. 6.


7. Desa, district Calafat, reg. Craiova, Oltenia. Height, 54.1 centimeters; diameter, 29.6 centimeters; maximal height of the lugs, 11.4 centimeters; height of the stand, 9.8 centimeters. Fished out from a lake between Ciu- perceni and Ghidiciu. Fig. 38.

Nestor and Nicolaescu-Plop§or 1937, 178, pl. 3a and 6; Fettich 1953, pl. 36:3, Takats 1955, fig. 10; Werner 1956, 58, n. 10, pl. 28:3.

Fig. 36. Hunnic bronze cauldron from B&ntapuszta, near Vdrpalota, Hungary. From Takdts, AOH, 1959, fig. 1.
Fig. 37. Fragment of a bronze cauldron from Dunaiijvdros (Intercisa), Hungary. From Alfoldi 1932, fig. 6.
Fig. 38. Hunnic bronze cauldron from a lake, Desa, Oltenia region, Rumania. (Height 54.1 cm, diam. 29.6 cm.) From Nestor and Nicolaescu- Plop§or 1937, pls. 3a-3b.

8. Hotarani, district Vinju Mare (formerly Meheninfi), Craiova, Oltenia. Fragment of a lug, 16.2 centimeters high, 19.7 centimeters wide. Found in the mud of a lake. Fig. 39.

Nestor and Nicolaescu-Plopsor 1937, 178–179, pl. 39:1; Werner 1956, 58, n.8, pl. 28:1.

9. Probably from western Oltenia. Fragment of a lug, 84 centimeters high. Fig. 40.

Nestor and Nicolaescu-Plopsor 1937, 179–180, pl. 39:2; Takats 1955, fig. 12; Werner 1956, 58, no. 11.

10. Bo§neagu, community Doroban|u, district Calarasi, reg. Bucuresti, Muntenia. Two fragments of lugs. The bigger one is 18 centimeters high, 12.7 centimeters wide, 1.3 centimeters thick. Found in 1958, 1.5 meters under the ground, at the border of the inundation area of the Danube, near the eastern shore of Lake Moti§tea. Fig. 41.

Fig. 39. Fragment of a bronze lug from a lake, Hotarani, Oltenia region, Rumania. (Height 16.2 cm, width 19.7 cm.) From Nestor and Nicol&escu- Plopgor 1937, pl. 39:1.
Fig. 40. Fragment of a bronze lug probably from western Oltenia, Rumania. (Height 8.4 cm.) From Nestor and Nicolaescu-Plopgor 1937, pl. 39:2.

Nestor 1960, 703; B. Mitrea and N. Anghelescu, “Fragmente de Cazan Hunic descoperite in sud-estul Munteniei,” SCIV 11, 1960; Mitrea 1961, 549–558, figs. 1, 2 (our figure 41), 3, 4.

11. Celei (Sucidava), district Corabia, reg. Bucure§ti, Muntenia. Four fragments of walls and a Jug. Found in a layer of ashes in the Roman cas- tellum. Fig. 42.

Fig. 41. Fragment of a bronze lug found near the eastern shore of Lake Motigtea, from Bogneagu, Rumania. (Height 18 cm.) From Mitrea 1961, figs. 1–2.

D. Tudor, Dacia 7–8, 1937–1940, 375, fig. 10c, and 11–12, 1945–1947, 189, fig. 35:1, 2, 7; Takats 1955, 166, fig. 13:a-d; Tudor 1548, 161–162; Werner 1956, 58, n.8, pl. 64:18–21.

Fig. 42. Fragments of a lug and walls of a bronze cauldron from Celei, Muntenia, Rumania. From Takats 1955, fig. 13:a-d.

12. Shestachi, district Rezina, Moldavian SSR. Fig. 43.

L. L. Polevoi, Isioriia Moldavskoi SSR, 53; G. A. Nudel’man, SA 4, 1967, 306–308.

Fig. 43. Hunnic bronze cauldron from Shestachi, Moldavian SSR. From Polevoi, Istoriia Moldavskoi SSR, pl. 53.

13. District Solikamsk, obi. Perm. Height 9 centimers. Fig. 44.

Alfoldi 1932, 32, fig. 5 (after a sketch by Fettich); Fettich 1940, pl. 13:3 and 1953, pl. 26:11; Werner 1956, 58, n. 2. A poor photograph in SA 10, 1948, 201, fig. 15:5.

Fig. 44. Bronze cauldron from Solikamsk, Perm region, USSR. (Height 9 cm.) From Alfol di 1932, fig. 5.

14. Osoka, district Sengilei, obi. Ul’yanovsk (formerly Simbirsk).[1868] Height, 53.2 centimeters; diameter, 31.2 centimeters; weight, 17.7 kilograms. Found in sand near the brook Osoka. Fig. 45.

V. Polivanova, “Zametka o proiskhozhdenii mednago sosuda iz Sen- gileevskago uezda, Simbirskoi gub.,” Trudy VII AS (Yaroslavl) 1, 39, pl. 1; Werner 1956, pl. 27:11 (most of the other reproductions are poor drawings).

15. VerkhniI Konets, region of Syktyvkar, Komi ASSR (formerly Ust’sysol’sk). Fig. 46.

J. Hampel, “Skythische Denkmaler aus Ungarn,” Ethnologische Mii- theilungen aus Ungarn 1897, 14, fig. 1 after a drawing by Prince Paul Putyatin, repeated by all later authors.

16. Ivanovka, gubernie Ekaterinoslav.[1869] Fig. 47.

Fettich 1940, pl. 8:10 (photo taken by A. Salmony in the museum in Novocherkassk), and 1953, pl. 36:4; a drawing in side view in Takats 1955, 166, fig. 15.

17. Found near Lake Teletskoe in the High Altai. Height, 27 centimeters; diameter, 25–27 centimeters. Aspelin, who first published the cauldron, gave as its findspot Teletskoe,[1870] which later authors changed to Biisk; but Biisk, 100 miles northwest of the lake, was only the place where the cauldron was given to Grand Duke Vladimir Aleksandrovich, who donated it to the Historical Museum in Moscow. Fig. 48.[1871]

Fig. 45. Bronze cauldron found in the sand near the Osoka brook, Ul’ya- novsk region, USSR. (Height 53.2 cm, diam. 31.2 cm, weight 17.7 kg.) From Polivanova, Trudy VII AS 1, 39, pl. 1.
Fig. 46. Bronze cauldron from Verkhnii Konets, Komi ASSR. From Hampel, Ethnologische Mittheilungen aus Ungarn 1897, 14, fig. 1.

18. Narindzhan-baba, district Turtkul, Kara-Kalpak ASSR. Fragment of a lug.[1872] Fig. 49.

S. P. Tolstov, Dreunyi Khorezm, 130, fig. 74a.

19. Allegedly found on the “Catalaunian battlefield.” Fragment of a lug 12 centimeters high, 18 centimeters wide. Fig. 50.

Takats 1955, 143, figs, la, b. E. Salin, Academic des inscriptions et belles-lettres. Comptes rendues des seances de I’annee 1967, 389, fig. 2.[1873]

The cauldrons, from the plainest to the most ornate, have four features in common: Their cylindrical or bell-shaped bodies are supported on a stand in the shape of a truncated cone which is slightly curved inward; their rectangular lugs project vertically from the rim; they are cast; with the exception of one or two, they are technically inferior vessels.

Fig. 47. Bronze cauldron from Ivanovka, gubernie Ekaterinoslav, USSR. From Fettich 1953, pl. 36:4.

The cauldron from Tortel was cast in four,[1874] those from Jedrzy chowice, Kurdcsibrak, and Osoka in two molds, which is probably also true for the other vessels. Body and stand were cast separately, hooked, and soldered together. The stand, which broke off easily, is often missing.

The Huns were not good at casting the comparatively large vessels. The traces of the joints of the mold sections were rarely removed, the horizontal ribs running around the upper part of the body almost never meet where they should. Not even on the poorest Chinese ritual bronze would a dot like the one in the triangle of the cauldron from Teletskoe (fig. 48) have been left; apparently the casters had no tools to file it off.

Fig. 48. Bronze cauldron found near Lake Teletskoe, in the High Altai, now in the State Historical Museum, Moscow. (Height 27 cm, diam. 25–27 cm.) Photo courtesy State Historical Museum, Moscow.
Fig. 49. Fragment of a bronze lug from Narindzhan-baba, Kara-Kalpak ASSR. From Tolstov 1948, fig. 74a.

It is regrettable that only one fragment from Bosneagu and another one from Sucidava have been analyzed;[1875] the results might be of historical importance. As the chemical and spectrographical analysis of twenty cauldrons from the Semirech’e shows, the copper corresponds to the local copper ore, which makes it practically certain that the cauldrons were cast where they were found.[1876] The metal of the Eurasian “bronze” cauldrons is actually copper, mixed with various impurities. The metal of the Scythian cauldron from Karagodeuakhsh is almost pure (99 percent) copper.[1877] The alloy—if it can be called alloy— of the Semirech’e pieces consists of 95.4 to 99 percent copper. The two fragments from Rumania do not come from bronze but from copper cauldrons. The material of one is 75 percent copper, 25 percent red oxide of copper (ruby red, cuprite, Cu2O), and a negligible amount of lead; that of the other one is 71 percent copper, 25 percent red oxide of copper, and 4 percent lead. The “bronze” of the cauldron from Desa is described as “reddish”; the material of the one from Benesov is “bronze with a strong content of copper.” According to Po- livanova, the metal of the Osoka cauldron is pure copper. In the cauldron from Jedrzychowice, “the ingredients are so unevenly mixed that in some places the copper appears almost pure; in others, tin is preponderant.” The distribution of the metals in the alloy in the lug from Sucidava is “extremely irregular.”

Fig. 50. Fragment of a bronze lug, allegedly found “on the Catalaunian battlefield.” (Height 12 cm, width 18 cm.) From Takats 1955, fig. l:a-b.

How the Huns got the copper is not known.[1878] Its poor quality seems to indicate that the smiths themselves heated and reduced the ore with charcoal or wood with the help of blast air in some form of furnace. Occasionally they may have pillaged graves. Had they melted Roman bronze vessels and recast the metal, the results would have been much better. The cauldrons are in every respect barbaric.

Yet with all their flaws and imperfections, the Hunnic cauldrons decisively refute the views of those historians who, like Thompson, deny the Huns the capacity of working metal. The Sarmatian cauldrons were cast by professional metalworkers;[1879] so were those of the Huns.


Like Alaric’s Visigoths who drank from Greek mixing bowls,[1880] if they did not cook in them, the Huns probably used all kinds of iron, bronze, copper, and silver vessels. Peoples on a trek and nomads cannot afford to insist on stylistic uniformity. Three of the four cauldrons in a hoard northeast of Minusinsk are of the common South Siberian type, but the fourth one is closely related to vessels best known from the Semirech’e.[1881] In the hoard from Istyak in Kazakhstan, cauldrons with three legs occur side by side with cauldrons on conical stands.[1882] The Hsiung-nu also had bronze vessels of various shapes.[1883] Some they carried back from their raids into China or bartered for horses, but those cast for themselves also differed in shape and size; the one found in Noin Ula by the Kozlov expedition[1884] and the high bronze vessels which Dorzhsuren excavated in 1954[1885] have only the decoration in raised lines in common. The Germanic and Alanic chieftains of the fifth century likewise had metal vessels of various origin; I need only to refer to the silver jugs from Congesti and Apahida. At Jedrzychowice a Hunnic cauldron was found together with a Roman bronze bowl. Looking around at a banquet in Attila’s palace, a guest would have seen sacred Christian vessels like those which the bishop of Margus handed over to the Huns,[1886] profane ones brought to Hungary from everywhere between the Loire and the Dardanelles, and Hunnic cauldrons.

Werner claims a footed bronze bowl from Munstermaifeld in the Eifel[1887] for the Huns. In his opinion,[1888] it is similar to a bronze cauldron from Bri- getio-Oszony in Hungary[1889] and another one from Borovoe in northern Kazakhstan (fig. 51 ).[1890] Because he takes the other finds from Borovoe for Hunnic, he thinks that the bowl from the Eiffel must be Hunnic too. Actually, the three pieces belong to three different types.

Fig. 51. Bronze cauldron from Borovoe, northern Kazakhstan. From Bernshtam 1951a, fig. 12.

The vessel from Brigetio is probably of late Scythian origin; in any case, the figures on its surface[1891] set it widely apart from the two others. The finds from Borovoe play a prominent part in the speculations about the Huns in Central Asia. Werner thinks that they indicate the expansion of Attila’s empire deep into Kazakhstan; only the other allegedly Hunnic findspot in Kara-Agach lies still farther east. To Bernshtam the finds are of even greater importance. They are supposed to prove the polychrome style of jewelry to be the product of the “creative” meeting of a local Central Asiatic culture and the political rise of the Huns. What bourgeois “fal- sificators” call Gothic art is actually the art of the Huns carried by them as far as Hungary.[1892]

Borovoe[1893] in the district Suchinsk, Kokchetav, lies in an archaeologi- cally little known region. The grave has some unique features, for example, a granite slab on top of it 4.5 meters long, 1.5 meters wide, 0.7 meters thick, weighing 4,000 kilograms. Underneath there were two more slabs, each 0.12 meters thick and a layer of rubble and pebble in which the cauldron was found. Still deeper in the ground was the pit. Of the skeleton, only the skull was “more or less” preserved. It would be of interest to know what the original position of the skeleton was. Was it extended or flexed, lying in a niche or a catacomb at the end of a dromos ? The very heavy stone slabs prove that the grave was not that of a Hun. Among neither the graves which Werner assigns to the Huns nor those which Bernshtam regards as Hunnic occurs anything similar to the construction of the grave in Borovoe.

The tomb furniture was a strange hodgepodge. The arrowheads were of three types, trihedral, three-flanged, and rhombic in cross section. Side by side with technically superb jewelry occur such primitive things as small blue-dyed bone beads, a copper buckle, and bronze wire earrings. As Werner noticed, aP-shaped sword mount is similar to one from the Taman Peninsula.[1894] In the same direction, the Bosporan workshops, point also the gold objects with their combination of triangular clusters of granulation and cloisons filled with red stone. It is infinitely more probable that the pear-shaped cloison within a border of grains from Borovoe[1895] comes from an East Roman workshop than that the almost identical one from Cyprus[1896] was made by a Hun.

Some of the things found in Borovoe occur also in Hunnic finds. But this does not make the cauldron Hunnic. An almost identical one was found near Tashkent.[1897]

There is little resemblance between the footed Munstermaifeld bowl and the cauldron from Borovoe. The former is an elegant vessel with two plain round handles, the latter a crude piece with four scalloped handles. The Munstermaifeld bowl contained the charred bones of a very young individual,[1898] a form of burial foreign to the Huns. It may not be a coincidence that the bowl was found in a field next to which there were many traces of a Roman villa.[1899] In the fourth century, Sarmatians were settled in the Moselle region.[1900]


Compared with the big Scythian cauldrons as, for example, the one from Chertomlyk, which is 3 feet high,[1901] or even those from Kazakhstan and Kirgizia, some of which could hold 140 liters,[1902] the Hunnic cauldrons were, as a rule, of moderate size. They were cooking vessels. The solid conical stand was not quite as effective for the maximum utilization of fuel, always scarce in the steppes, as the tripod or the perforated stand,[1903] but it helped. Like the Scythians and Sarmatians, the Huns used the cauldron for boiling meat; it was lifted out with a hook similar to those found in Verkhne-Kolyshlei and Khar’kovka. (Such hooks are still used by the Kazakhs and the Abkhaz in the Caucasus.)[1904]

The usual assumption that nearly all Eurasian cauldrons were sacral vessels has rightly been doubted by Werner and Spasskaya. It is true that in the larger ones food for more than one person was prepared, but this does not prove that the meal was always sacrificial. The rock pictures from the Pisannaya Gora in the Minusinsk area (fig. 52)[1905] have frequently been interpreted as reproductions of religious ceremonies. Such big cauldrons, it was thought, cannot have been ordinary cooking vessels. However, their size in the drawings only betrays the artist’s ineptitude. The ladle to scoop out the broth which the man to the left is holding is of the same gigantic proportions as the hook in the hand of the man to the right. There were cauldrons bigger than a man. On the Bol’shaya Boyarskaya pisanitsa, in the same region, twenty-one buildings and sixteen cauldrons are depicted. Evidently, such a small settlement could not have had so many sacrificial vessels; besides, they are of moderate size (fig. 53).[1906]

Fig. 52. The representation of a cauldron in a detail of a rock picture from Pisannaya Gora in the Minusinsk area. From Appelgren-Kivalo, fig. 85.
Fig. 53. Representation of cauldrons in a rock picture from Bol’shaya Boyarskaya pisanitsa, Minusinsk area. From Devlet, SA 3, 1965, fig. 6.

Another argument in favor of the sacral character of the cauldrons from southern Siberia, Kazakhstan, and Kirgizia is the circumstances under which they were found. None of the numerous cauldrons from the Minusinsk area—I saw dozens in the museum in Minusinsk—and only two of thirty-three found in Kazakhstan and Kirgizia come from graves.[1907] As they were not buried with the dead, they supposedly were not owned by one person but by a larger group and, therefore, clearly not used for preparing everyday meals. The findspots are probably the places where the sacrifices were performed.

With regard to the Hunnic cauldrons, we are confronted with a similar situation. Of the eighteen finds, only the cauldron from Jedrzy chowice was allegedly found in a grave.

Alfoldi and Werner agree that in Jedrzychowice a Hunnic nobleman was buried. The objects found were (1) the cauldron; (2) a Roman bronze bowl; (3) two iron buckles; (4) a gold buckle, its rectangular plate decorated with red stones in cloisons; (5) two gold strap ends; (6) six pieces of thin gold sheets with rectangular and triangular red stones encased in cloisons; (7) a gold chain.[1908] Goetze recognized that the gold sheets were originally parts of a diadem which had been cut up to decorate a leather belt and a buckle. Straps and buckles are known, though not exclusively, from Hunnic graves. The cauldron is undoubtedly Hunnic. Jedrzychowice is supposed to be a Hunnic grave.

This, in my opinion, is open to doubt. In his article on the find, E. Krause reprinted the original report in the catalogue of 1838:[1909] A peasant, ploughing a flat potato field, hit with his ploughshare the handle of the cauldron; the vessel lay [1910]</em> in fine white sand and was filled with sand and dirt; in the same depth, about 2 feet to the west, was the bronze bowl; north of the cauldron was a strip of white sand, 12 to 16 inches wide, 5 to 6 feet long, and in it a dark brown band, about a hand’s breadth and barely 1 inch high, in which lay scattered traces of bones, small wooden sticks of various shapes, mostly with silver mountings, the gold sheets, and the shoe buckle. At the end of the strip was a 3-inch square of dark brown dirt, and in it lay the gold chain.

As far as I know, only Takats paid attention to this description.[1911] He thought that the white sand was the bed of a small creek, but he drew no conclusions from this strange choice of a site for a grave. He only insisted that the tiefernste sacrificial vessel had nothing to do with the flimsy gold sheets from East Roman workshops.[1912]

There is something else peculiar about the alleged grave. I asked Professor Paul Leser of the Hartford Seminary Foundation, the leading authority on early ploughs, what depth the plough of a Silesian peasant in 1830 could have reached. I quote from the letter he kindly sent me on August 28, 1964:

It would be quite impossible, in my opinion, that any plough used in Upper Silesia in the 1830’s would have reached a depth of 3 feet. The average plough there at that time dug a depth of 4–10 inches (1025 cm.) The deepest ploughing plough available in Central Europe in the first half of the nineteenth century scarcely ever ploughed as deep as 15 inches.

But this is not all that puts this “grave” in a peculiar light. Werner noticed that the gold platings of the strap ends and the gold sheets were fastened to the leather in such a sloppy way, with one or two tiny rivets, that the belt and the straps could never actually have been used. The sheets must have been cut out from the diadem and fastened to the belt of the dead man; the gold platings were specially made for shoes to be put into the grave. Werner’s observations are correct. What follows from them ? Did the horseman carry with him the gold platings for his shoes and straps and the diadem, whole or already cut to pieces, for his belt in case he died far away from home? Was he accompanied by a goldsmith who made the gold rivets on the spot? Or did the survivors send a man from Silesia to Hungary to fetch the gold things for the burial? One explanation is more farfetched than the other. They are all to be rejected, for, unless the report on the finds is utterly unreliable, the “grave” contained no skeleton. At least the skull should have been preserved. The few bones probably were in the cauldron and fell out when it was overturned.

From whatever angle one looks at this curious ensemble in the bed of a creek, this “grave” without a pit, less than a foot under the ground, it remains puzzling. One could think of a hoard, consisting of objects of various provenance, partly loot (the cauldron, the bowl, the gold chain), partly from a pillaged grave, but the bronze buckles were hardly objects worthy to be hoarded.

On the other hand, the parallel with the find from Osoka is striking. There, too, the cauldron was found in the sand near a creek. This could be a coincidence, if not a third find, still farther to the east, would not make it probable that the two cauldrons were intentionally deposited where they were found. A cauldron with round handles, the surface decorated with raised lines in the same pattern as on one from Noin Ula and many from the Ordos region, was found in the bed of the Kiran River in northern Mongolia.[1913] Although the connection of the cauldron from J^dryzchowice with the other objects in the find remains obscure, it lay, like the other cauldrons, in or near running water. To the same, though somewhat looser connection with water, point, as Nestor and Takats noted,[1914] our numbers 4 (bog), 5 (marsh), 7 (lake), and 10 (near a lake), to which we now can add number 17 (near a lake).

Such a location was not limited to the Huns and Hsiung-nu. The Hsiung- nu were never in the Cis-Baikal forests, yet a large cauldron was found on the bank of the river Kutullaki in the former district Kiren, gubernie Irkutsk, and a similar, only smaller one, on the island Shchukin in the Angara River, about 13 kilometers north of Irkutsk.[1915] Three-legged cauldrons, typical for the Semirech’e long before the Huns, came to light on the shores of the Issyk-kul.[1916] In the Minusinsk area cauldrons were found on the left and right bank of the Yenisei and on the bank of the river Shush.[1917]

In his discussion of the cauldrons from Kazakhstan and Kirgizia, Spasskaya offers an attractive explanation of their location.[1918] She thinks that the nomads performed some rites on watercourses in the spring, stored the vessels near the water when they moved to the higher summer pastures, and used them again when they came down in the fall. This assumption seems to be supported by the association of cauldrons, sometimes more than one, with other bronze objects in sacrificial rites. If thereby the cauldron itself should have acquired a sacred character, one would understand a find like the one from Bo§neagu, where a lug was buried 1. 5 meters under ground; it must not be profaned. One may conjecture that the particularly sacred part of the more sumptuous Hunnic cauldrons was the handles with the “mushrooms”; following other considerations, Werner arrived at similar conclusions.[1919] Although there were no high summer pastures in the Hungarian and Rumanian plains, in depositing cauldrons near creeks, lakes, or marshes the Huns might have preserved an old custom under changed circumstances. In any case, the location of a number of cauldrons near water strongly points to their use in some ceremonies.[1920] On the other hand, there is no reason why other cauldrons could not have been just everyday cooking vessels like those found in Late Sarmatian graves.


This or that feature of our cauldrons occasionally appears in pre- Hunnic times, which is in no way surprising. Their function bound all cauldrons togethers; they all must have a round body and handles. A Sauromatian cauldron from the Orenburg area, for example, has an almost cylindrical body, but its handles are round with a knob on the top.[1921] On the whole, however, the differences between the Scythian, Sarmatian, Semirech’e, and Far Eastern cauldrons are sharply marked.[1922] Werner derived the “mushrooms” of the Hunnic cauldrons from the three knobs on the handles of late Sarmatian cauldrons.” Taken by itself, this seems to be quite plausible. But the Sarmatian egg-shaped cauldrons had no stand; and by the beginning of the third century they went out of existence.[1923] The round, low, flat-bottomed imported Roman kettles mainly known from the lower Volga region,[1924] have nothing to do with ours.

The Huns did not create their cauldrons out of nothing. Their affinity for those of the first centuries a.d. from northern China, Mongolia, and the Ordos region long has been recognized by Japanese[1925] and Western scholars.[1926] It is true that the cauldrons from the Hsiung-nu graves at Noin Ula and the Kiran River (fig. 54)[1927] have their almost hemispherical bodies decorated with raised lines in wide waving curves which have no parallels in the Hunnic cauldrons. The handles, round or rectangular with a scalloped upper rim, do not occur in our cauldrons either. However, plain rectangular handles, comparable with those on the cauldrons from Jedrzychowice and Osoka, are also known from Ordos cauldrons.[1928] On a footless vessel with an elongated body, found in 1950 in a rich Hsiung- nu grave of the later Han period in Inner Mongolia near Erh-lan-hu-kou, one handle was round and the other rectangular.[1929] It seems that the rectangular handle with a scalloped upper rim is merely a variant of the plain rectangular handle. There exist, indeed, a number of handles with scallops so shallow that the upper rim looks almost straight.[1930]

Fig. 54. Bronze cauldron of a type associated with Hsiung-nu graves at Noin Ula and the Kiran River. From Umehara 1960, p. 37.

Takats was, I believe, right in comparing the scalloped rim from Noin Ula with the rim of the cauldron from Lake Teletskoe (fig. 48).[1931] If one imagines the rounded triangles of the cauldron from the Altai put on stalks, they would come close to the “mushrooms.”[1932] It is true that the cauldrons from the borderlands of northern China are somewhat smaller than the Hunnic cauldrons; they are, as a rule, squatter, the handles are mostly round, and the stands nearly always perforated. However, there exist also Ordos cauldrons with elongated bodies and solid stands.[1933]

The Hunnic cauldrons cannot be derived from the Scythian and Sarmatian ones, not to speak of the three-legged cauldrons from the Semirech’e. If they are not the direct descendants of the Ordos cauldrons, they certainly are their cousins. Some, probably many, Ordos cauldrons were cast by and for Hsiung-nu.[1934] But not all, as not all small Ordos bronzes (all those knives, daggers, belt buckles, discs, pendants, horse-frontlets, and so forth) were of Hsiung-nu origin.[1935] Ordos cauldrons were bound together with small Ordos ornaments.[1936] But in Inner Mongolia Ordos cauldrons were found in graves which the Chinese excavators, probably rightly, date to the Northern Wei period (424–534 a.d.).[1937]

The farther to the west the cauldrons were cast, the more they differ from their prototypes. It looks as if they fell under the influence of basically related but more richty decorated bronze and copper vessels. On the cauldrons from Lake Teletskoe (no. 16) and Solikamsk (no. 15), two raised lines or ribbons, starting below the handles, run down the body, sharply curving outward at the lower ends. This pattern, foreign to the Ordos bronzes, occurs on Scythian cauldrons from the Don region as early as the fourth century b.c.[1938] and on the Kuban as late as the first century a.d.[1939] The same two Hunnic cauldrons have, along the rim, square compartments formed by raised lines, with one or two lines between the opposite corners. This pattern, likewise foreign to the Ordos, seems to be related to the one on the cauldron from Chertomlyk.[1940]

The origin of the “pendants” on the cauldrons from Tortel, Desa, Shest- achi, Osoka, and Verkhnil Konets is obscure. Takats noted that similar “pendants” occur on Chinese pots from the Neolithic period, and derived the former from the latter.[1941] In view of the two or more millennia which separate the clay from the copper “pendants,” such a connection is out of the question. But the parallel may give a hint to the origin of the Hunnic “pendants”: They might be replicas of cords or fringes. In the early art of the barbarians at China’s border as well as in China proper the rendition of cords on bronzes was quite common. The cords probably served also a practical purpose. The Korean vessel from the Gold Bell Tomb at Kyongju in Korea (fig. 55)[1942] shows how the nomads transported the cauldrons over long distances.

To arrange all Hunnic cauldrons in a typological series does not seem possible. The upper edge and the sides of the handle of the Ivanovka cauldron (fig. 47) are curved as on the fragment from Bosneagu (fig. 41), which seems to indicate that the two vessels were cast at approximately the same time. Fragment no. 19 (fig. 50) shares with the cauldron from Shestachi (fig. 43) the circles on the “mushrooms.” It seems reasonable to assume that the cauldrons with the plain handles are earlier than those with the “mushrooms.” But the Huns may very well have cast plain and elaborately decorated cauldrons, perhaps for different purposes, at the same time. If someday it should be possible to date the cauldrons more exactly, they still would give the context in which they were found only a terminus post quern, for they were used for generations. Many were repaired. “A man’s life span is fifty years, a cauldron can be used for a hundred,” says a Kazakh proverb.[1943]

Fig. 55. Ceramic vessel from the Gold Bell Tomb at Kyongju, Korea, showing the manner in which cauldrons were transported by nomads. From Government General Museum of Chosen 1933, Museum Exhibits Illustrated V.

Scattered from the borders of China to eastern Europe, the cauldrons of course can not indicate the way over which they spread to the West. They are absent from Tuva and the Minusinsk area,[1944] have so far not turned up in Kazakhstan, but are known from Khwarezm.

Fig. 56. Clay copy of a Hunnic cauldron of the Verkhnil Konets type (see above, fig. 46), from the “Big House,” Altyn Asar, Kazakhstan. (Height 40 cm.) From Levina 1966, fig. 7:37–38.

A number of clay vessels (fig. 56),[1945] about 40 centimeters high, found in the upper horizon of the “Big House” at Altyn Asar (Dzheti Asar 3), are copies of Hunnic cauldrons of the type Verkhnil Konets. They not only show the seams where the sections of the mold met; the handles with their parallel lines are the same here and there; the rings of the pendants on the upper zone of the body of the copper vessel appear as dots on the pottery copy. A fragment of a lug with a “mushroom” comes from approximately the same region. Tolstov dates the upper horizon from the third to the seventh,[1946] Levina from the beginning of the fourth to the seventh or eighth centuries.[1947]

In the upper strata of the lower horizon, thus not much earlier than the clay cauldrons, and in the kurgans near Dzheti Asar lay bone strips from composite bows. Some of the persons buried in the kurgans were Europeoids with a Mongoloid admixture; some had deformed heads. The Hunnic (or Hunnoid) population in the delta of the Syr Darya had cauldrons of the Verkhnii Konets type.

The imitations of Hunnic metal cauldrons in Khwarezm are just as closely connected with other elements of Hunnic civilization as the cauldrons with the “mushroom” handles in Hungary and Rumania, areas ruled by the Huns in the fifth century. Hunnic soldiers in Sucidava broke their cauldrons in the 440’s. If, as the decoration on the fragment from Intercisa indicates, the cauldron had “mushroom” handles, the type must have existed at the end of the fourth century, when the camp on the Danube was still Roman. There remain three more cauldrons with such handles: Ivanovka, Benesov, and Narindzhan-baba. The vessel from Ivanovka has “no passport,” as the Russians would say. How it got into the museum in Rostov is not known. The fragment from Narindzhan- baba is possibly to be connected with the finds from Altyn Asar. But what about Benesov? It has been argued that the last owner of the cauldron as well as the man who carried a cauldron to Jedrzychowice in Silesia were Huns, subjects of Attila or one of his predecessors. Alfoldi, Werner, and Sulimirski[1948] are convinced that Benesov and Jedrzychowice were Hunnic camps. By the same reasoning the Huns should have had garrisons in Osoka, Solikamsk, and Verkhnii Konets. Werner evidently feels that would expand the Hunnic “empire” too far and consequently speaks somewhat vaguely about the Fundmilieu ostlicher Reiterkrieger.[1949]

A glance at the map is sufficient to exclude the possibility that Huns or any other “Eastern mounted warriors” could push even close to Solikamsk or Verkhnii Konets, across the forests and swamps into the land of the Komi (Zyryans). Verkhnii Konets is at the latitude of Helsinki. The cauldron from Solikamsk is no more proof of the presence of Huns in the northern parts of the oblast’ Perm than the Roman,[1950] Sasanian,[1951] and Byzantine[1952] bronze and silver vessels prove the existence of foreign troops in northeastern Russia; the primitive hunters on the Vyshegd were not the subjects of the basileus in Constantinople or the king of kings in Ctesiphon. They never had heard of Attila. As the Sasanian and Byzantine luxury vessels and coins [1953] testify to fur trade, over many middlemen, between the Permian lands and the higher civilizations in the south,[1954] the Hunnic cauldrons probably point to similar relations between the northern tribes and the ancestors of the Huns. I say ancestors because a considerable time must have passed before the cauldrons from Lake Teletskoe, Solikamsk, Osoka, and Verkhnii Konets changed into the vessels of the fourth and fifth centuries.

It is to be assumed that future excavations will close the many gaps between the Kerulen River and the Danube. Still, even now there can be no doubt that the Hunnic cauldrons originated on China’s northern and northwestern borders. The crude, often truly barbaric copper cauldrons link the Huns with the area of the Hsiung-nu confederacy.


Objects of Central Asiatic origin have been found at various places in eastern Europe: Bactrian silver phalerae of the second century b.c. at Novouzensk in the oblast’ Kuibyshev;[1955] a Bactrian tetradrachm in Chersonese;[1956] Kushan coins in the Volga region[1957] and in Kiev.[1958] They are oddities, though not quite as odd as the Shang bronze fished up at Anzio, the Late Chou bronzes unearthed at Rome and Canterbury, and the Chinese coins of the third century b.c. dug up in southern France.[1959] Chinese objects found in eastern Europe belong to a different category. They were actually used by the barbarians. The jade scabbard slides in Sarmatian graves, for instance, came from China; the nomads had no access to the gemstone, and the dragons carved on some slides are unmistakably Chinese. The Sarmatians fitted them on their scabbards in the same w’ay they used their wooden slides.[1960] Pieces of Chinese silk from dresses were found in a Late Sarmatian grave at Marienthal (now Sovetskoe), on the Big Karman River in the former German Volga Republic,[1961] and in a grave at Shipovo.[1962] The Han mirror in kurgan E 26, burial 19, on the Torgun River in the lower Volga region (fig. 57)[1963] may have been cherished for its magical power, but it was also a toilet implement.

Fig. 57. Chinese mirror of the Han period found in burial 19, on the Torgun River, lower Volga region. From Ebert, RV, « Siidrussland,» pl. 40: c:b.

As there existed no direct trade relations between the Chinese and the Sarmatians on the Volga, the Chinese objects reached the east European steppes via Central Asia. The striking similarity of Sarmatian gold and clay vessels with animal handles to a Chinese ritual bronze with a tiger handle in the British Museum finds its explanation in the origin of the motif in Central Asia, possibly Fergana, from where it spread both east and west.[1964] The westward spread of Chinese mirrors through Central Asia and their gradual transformation can be fairly well traced.[1965]

The earliest Chinese mirrors found outside of China are two Huai mirrors in the Hermitage. The one from Tomsk[1966] is identical with a mirror in the Lagrelius collection which Karlgren dates to the fifth century b.c.;[1967] the other one, from the sixth kurgan in Pazyryk in the High Altai,[1968] is about a century later.[1969] In the past forty years no more Huai mirrors have turned up in southern and western Siberia, and it is unlikely that many more will be found in the future. Han mirrors, however, have come and are constantly coming to light in northern Eurasia, from Outer Mongolia to the Ob River and the lower Volga.

Some of those found near the frontier as, for instance, in the tombs of the Hsiung-nu princes at Noin Ula, were probably gifts of the emperors; others testify to trade relations with China. In the barbaricum Chinese mirrors were bartered from tribe to tribe. Even fragments were highly appreciated. The edges of a broken Han mirror in the Izykh chaatas in the Minusinsk[1970] region were smooth, not sharp as they would have been had the mirror, as so often, been intentionally broken before it was put in the grave. This proves that the fragments had been held in many hands.

To draw up a list of the Han mirrors found in the barbaricum must be left to scholars who have access to the museums in Inner and Outer Mongolia and the Soviet Union. Only a small fraction has been published; many more are merely listed as “ancient Chinese mirrors.” Still, even the little that is known is impressive.

As was to be expected, Han mirrors were found in the barbarian graves beyond the northern and northeastern frontiers of China: in the Hsiung- nu graves at Noin Ula,[1971] Il’mova Pad’,[1972] and Burdun,[1973] and in the large, presumably Wu-huan, cemetery at Lo-shan-hsiang in Manchuria.[1974] Of the numerous stray finds in the Minusinsk area only a few have been published. One is of the i t’i tzu (“quaint script”) type, five are “TLV” mirrors.[1975] In the Kenkol cemetery in the Talas Valley a chang i tzu sun[1976] and a “hundred nipples” mirror[1977] were found. With the exception of a TLV mirror in Kairagach,[1978] the mirrors from Fergana were chang i tzu sun mirrors: three from Tura-tash,[1979] one from Kara Bulak,[1980] one and a fragment from the kurgans in the Isfara Valley.[1981] Of the same type was a fragment from the northwestern part of the oblast’ Leninabad[1982] and a mirror from Vrevskii southwest of Tashkent.[1983]

Like most of the objects in the Istyatsk hoard on the Vangai River between Tobolsk and Omsk in western Siberia, the chang i tzu sun mirror[1984] must have been brought from the south. The same is true for a Han mirror in a kurgan near Tobolsk.[1985] In a kurgan at Zarevshchina in the former gubernie Astrakhan, a “four S spirals” mirror was found together with a Turkish stirrup.[1986] How the Turkish nomads got the mirror can only be guessed. They were not averse to occasional grave robbings; a Sarmatian kurgan at Politotdel’skoe in the lower Volga region for example, was ransacked in the time of the Golden Horde.[1987] Mirrors were often used for a long time before they accompanied the dead to the other world. To give just one example, in a kurgan at Naindi sume on the Tola River, about 120 kilometers southwest of Ulan Bator, a Han mirror was found together with a piece of Chinese silk which the Sasanian pattern dates to the sixth or seventh century.[1988]

From Middle Sarmatian graves in the lower Volga region two Han mirrors are known: a fragment of what seems to be a “four nipples” mirror from a diagonal burial at Berezhnovka II, kurgan 3,[1989] and the abovementioned mirror from the Torgun.[1990] The westernmost Han mirror with a long inscription,[1991] unfortunately without a date, comes from the Kuban region in the Caucasus.[1992]

The list, incomplete as it is, shows how popular Han mirrors were among the peoples and tribes west of the Great Wall. The absence of mirrors of the Six Dynasties period finds its explanation in the breakdown of Chinese power in the western regions; Chinese mirrors reappeared in Central Asia only in the T’ang period.

When the supply from the big state factories dried up in the latter half of the second century, but occasionally also before, the barbarians tried to cast their own mirrors in the shape and with the designs of the admired Chinese bronze disks. In many cases the so-called imitation mirrors, the ho sei kyo of the Japanese archaeologists,[1993] can be easily recognized, though not all coarsened versions of the standard types are necessarily imitations. There exist a large number of small mirrors of the later Han period[1994] and the Three Kingdoms of such poor casting and such crude decor that they were rarely collected and, except in recent publications, hardly ever illustrated. Being very cheap, they must have been eagerly sought by the barbarians. I listed a mirror from Kenkol as a chang i izu sun mirror, although it has small circles between the leaves of the quatri- foil instead of the four characters chang i tzu sun. It could be argued that the barbarians, having no use for them, transformed the characters into ornaments. But identical mirrors are known from undoubtedly Chinese graves.[1995]

The imitations of Han mirrors vary greatly in quality. In Japan they are, as a rule, well cast; their decoration, deviating from the original sometimes

slightly, sometimes drastically, often foreshadows the future breakthrough of the native genius. In Korea the earlier imitations can barely be distinguished from Chinese mirrors, but they soon became cruder and thicker; their decoration has less and less in common with the prototypes. The least changed imitations come from the oasis cities in Hsin-chiang; only the simplification of the decoration gives them away. Whereas these three groups, in particular the Japanese one, have often been studied, very little about the imitation mirrors found farther to the west is known. Possibly some of the above-listed mirrors look genuinely Chinese only in the inadequate reproductions. Sometimes the excavator did not recognize the imitation. A mirror from the kurgan cemetery Kok-el in Tuva[1996] is, as the odd decoration shows, most probably a ho sei kyo.

In the second and third centuries Chinese mirror decorations were transferred to Sarmatian so-called pendant-mirrors, widely spread through the steppes between the Volga and the lower Danube from the first century b.c. to the fourth century a.d. The small bronze disks, occasionally silvered on the smooth side, sometimes with a high content of tin, were worn around the neck on a cord which ran through a perforated square or rectangle on the edge (fig. 58) ;[1997] some mirrors have instead of a rectangle a short flat tang (fig. 59),[1998] to be fitted into a wooden, bone, or horn handle. The designs in raised lines are the same in both variants.

Fig. 58. A Sarmatian bronze disc in the shape of a pendant-mirror, of a type found in the steppes between Volga and lower Danube, from the first century b.c. to the fourth century a.d. From Sinitsyn 1960, fig. 18:1.

The origin of the pendant-mirror is controversial. Rau thought it reached the steppes from the Caucasus,[1999] but it appeared in both areas at about the same time.[2000] Khazanov traces the pendant-mirrors back to Siberia; however, the mirrors to which he refers[2001] have handles in the shape of animals.[2002] The earliest undecorated pendant-mirrors were found in Wu- sun graves of the third or second century b.c.[2003] But we are less interested in the origin of the pendant-mirrors than in their decorations, particularly in a group from the lower Volga region and the northwestern Caucasus.

Fig. 59. Bronze mirror of a type similar to that shown on fig. 58, but provided with a tang that was presumably fitted into a handle. From Gushchina, SA, 2, 1962, fig. 2:5.

The findspots are

1, 2. The cemetery at Susly in the former German Volga Republic (figs. 60,[2004] 61).[2005]

3. Alt-Weimar (now Staraya Ivantsovka), kurgan D 12 (fig. 62).[2006]

4. Kurgan 40 in Berezhnovka in the lower Eruslan, a left tributary of the Volga (fig. 63).[2007]

5. Kurgan 23 in the cemetery “Tri Brata” near Elista in the autonomous Kalmuk SSR (fig. 64) ,[2008]

6. Lower Volga region (fig. 65).[2009]

7. A catacomb burial at Alkhaste in Checheno-Ingushetia in the northeastern Caucasus (fig. 66).[2010]

Fig. 60. Bronze pendant-mirror from the cemetery at Susly, former German Volga Republic. From Rau, Hiigelgraber, 9, fig. la.
Fig. 61. Bronze pendant-mirror from the cemetery at Susly, former German Volga Republic. From Rykov 1925, 68.
Fig. 62. Bronze pendant-mirror from Alt-Weimar, kurgan D12. From Rau, Ausgrabungen, 30, fig. 22b.
Fig. 63. Bronze pendant-mirror from kurgan 40 in Berezhnovka, lower Eruslan, left tributary of the Volga. From Khazanov 1963, fig. 4:9.
Fig. 64. Bronze pendant-mirror from kurgan 23, in the “Tri Brata” cemetery, near Elista, Kalmuk ASSR. From Khazanov 1963, fig. 4:8.
Fig. 65. Bronze pendant-mirror from the lower Volga region. From Khazanov 1963, fig. 4:6.
Fig. 66. Bronze pendant-mirror from a catacomb burial at Alkhaste, northwestern Caucasus. From Vinogradov 1963, fig. 27.

It is, first of all, the border, a band filled with radiating lines, which sets this group of mirrors apart. Rau derived the motif from the Caucasus, where it occurs on an antimony medallion of the Early Iron Age, and traced it back to Mycenean times. Such a simple motif can originate everywhere and at all times, and it might be a mere coincidence that it is found on small Sarmatian bronze mirrors and small Chinese bronze mirrors of about the same date. Such bands, encircling the central field, are known from Chinese mirrors before, in, and after the Han period. It is, however, remarkable that there exist imitation mirrors whose whole decor consists of two bands, one with radial lines and the other with dog-tooth ornaments.[2011] The mirror from Berezhnovka has two concentric bands. This, too, would not be particularly remarkable if the strokes in the outer one were not slanting, something alien to all other Sarmatian mirrors. Not the technique of casting a mirror, nor its shape, nor any other conceivable reason accounts for the combination of a square in the center field and a striated band around it as on these Sarmatian mirrors. The chances that this identity with the squares and the same borders on hundreds of Chinese TLV mirrors is still coincidental are very small. They become zero when we see the Sarmatian craftsman put a small knob in the center of the square. It has no function. It has no aesthetic value. It is the imitation of the perforated knob on the Han mirrors.

TLV mirrors were imitated in Japan, Korea, and the western regions of China. On a mirror from Lou-Ian (fig. 67),[2012] only the cross stroke of the T is left; L and V have disappeared. On some Japanese imitation mirrors the TLV’s have been entirely discarded. The Japanese craftsmen to whom inscriptions on the Chinese mirrors meant nothing changed them into fancy lines, but kept the dog-tooth, zig-zag, and radial lines of the border. The Sarmatian coarsened the Chinese patterns much more radically but still not beyond recognition. It would be unfair to place one of the Sarmatian mirrors next to a fine Chinese TLV mirror. They should rather be compared with the small Chinese mirrors in which the decoration has also been radically simplified as, for example, two mirrors recently found at Lo-yang, both lost the central square and the L’s and V’s (fig. 68, 69).[2013]

Fig. 67. An imitation of a Chinese TLV mirror from Lou-Ian. From Umehara, 6 bei, 39, fig. 7.

In one of them the T’s consist only of strokes, and the hundred and more radial lines of the border on good TLV mirrors are so widely spaced that they approximate those on the Sarmatian mirrors.

Fig. 68. Small bronze mirror with simplified decoration from Lo-yang. From Lo-yang ching 1959, 80.
Fig. 69. Small bronze mirror with simplified decoration from Lo-yang. From Lo-yang ching 1959, 82.

The Sarmatians not only transferred Chinese designs to the pendantmirrors, they also cast mirrors in direct imitation of bronze mirrors of the Han period. Werner was the first to recognize the importance of what he called the ostliche Nomadenspiegel for the study of the Huns.[2014] They are disks of whitish bronze with a loop or perforated knob on the back for attaching the cord which served to hold them. The decoration consists of various patterns in raised lines. With few exceptions and in contrast to the manifold and often gracious ornaments on the pendant-mirrors, the decor is monotonous: two or more concentric circles, divided by lines radiating from the center, occasionally with dots in the compartments thus formed; in later mirrors a zig-zag line runs between the circles. Rau called the group “Sibero-Chinese,”[2015] Khazanov lists it as Group X of the Sarmatian mirrors.[2016] For brevity’s sake, I shall call them loopmirrors.

Werner distinguishes four types of decoration. The earliest one is supposedly found on four mirrors from Mozhary, Susly, Atkarsk, and Tanais.[2017] Actually the group “Mozhary” consists only of two mirrors. The one from Atkarsk belongs to another type, and the mirror from Susly does not exist. Neither Rykov, who excavated the cemetery, nor Pater Beratz, who dug three kurgans in it, knew of such a mirror. The drawing in Merpert’s article[2018] is the Mozhary mirror; Merpert mixed up his notes. In 1963, Khazanov published more mirrors of the Mozhary type, but the best and most important of the group remains the mirror from Mozhary (diameter, 7.4 centimeters), often reproduced[2019] and dated between the first and fourth centuries (fig. 70).[2020] It was found by peasants who did a little grave robbing in a kurgan on the mountain Mozhary near the settlement Kotova in the district Kamyshin, gubernie Saratov, later oblast’ Stalingrad, now Volgograd. Only after I. I. Berkhin published the whole find[2021]—it is in the Hermitage—could it properly be evaluated. One may disagree with Berkhin on minor points, but the date at which he arrived after a thorough study cannot longer be in doubt: it is the beginning of the third century.[2022] As the mirror was evidently used for a considerable time, it cannot have been cast much later than about 200 a.d.

Fig. 70. Bronze mirror from Mozhary, Volgograd region, now in the Hermitage Museum, Leningrad, datable to about a.d. 200. (Diam. 7.4 cm.) From Umehara 1938, 55.

Werner rejected the possibility that the patterns on the loop-mirrors could have anything to do with the “artistic ornaments” of the Chinese mirrors; Khazanov would not exclude it entirely. As Rau before them, the two archaeologists think that the patterns on the loop-mirrors had been taken over from the pendant-mirrors. In some cases this might be true; as a whole, however, the two groups have very little in common. No loop-mirror has a tamga or a swastika or any of the ingenious combinations of patterns of the pendant-mirrors.

Berkhin and Solomonik tried to interpret the design on the Mozhary mirror. Berkhin took the “trees” on the square for a possible reflection of the cult of the Tree of Life, which is not exactly convincing.[2023] Solomonik spoke of “birds’ claws,”[2024] by the quotation marks indicating that this is meant to be a purely descriptive term. She referred to a mirror from Krasnodar and another one from Kosino in Slovakia (fig. 71 ).[2025] In the Krasnodar mirror she saw a combination of a swastika with “birds’ claws”; the same on pendant-mirrors from the Dnieper and the Volga (fig. 72, line II, last on the right). The similarity between the “birds’ claws” on the mirror from Kosino and the “trees” on the Mozhary mirror cannot be denied. However, the differences between the Slovakian and the Kuban mirrors on the one side and the mirror from Mozhary on the other outweigh by far the similarities. The first two mirrors have neither a rim with radial lines nor a square in the center field.[2026]

Fig. 71. Bronze mirror from Kosino in Slovakia. From Eisner, Slovensko v praveku 1933, fig. 2:7.

Fig. 72. Bronze pendant-mirrors from the Dnieper and Volga regions. From Solomonik 1959, fig. 6.

The design on the Mozhary mirror remains a riddle to Western archaeologists. Japanese archaeologists riddled it long ago. As early as 1925, Umehara wrote, “Anyone who looks at the design must certainly conclude that it is an extremely crude imitation of the popular TLV mirror.”[2027] Mizuno and Egami and a few years later Egami again listed the Mozhary mirror among the imitation mirrors of the West. The Sarmatian craftsman possibly transformed the lines on mirrors like the one from Lo-yang (fig. 69) into “birds’ claws,” but for the rest he copied the Chinese design as well as he could.

Even more simplified imitation mirrors of this type, without the “tree,” have been found at Blumenfeld and Khar’kovka in the lower Volga region.[2028] In mirrors from Norka (fig. 73)[2029] and Kalinovka[2030] only the rim with the radial lines is left.

Fig. 73. Sarmatian imitation of a Chinese mirror (cf. the example from Lo-yang, above, fig. 69), from Norka, lower Volga region. From Berkhin 1961, fig. 2:2.

The influence of Chinese mirror designs on non-Chinese mirrors before and after the period in which we are interested would deserve a special study. The scallops on a mirror with a long side handle from Tyukova near Tobolsk,[2031] for example, are doubtlessly copied from Western Han mirrors, Karlgren’s type k.[2032] Even designs on other metal objects occasionally betray their origin from Chinese mirrors. A stamped bronze plaque in the grave of later nomads at Akkermen[2033] looks almost like a “hundred nipples” mirror.

Although the present studies are not concerned with the origin of the Scythian and Sarmatian loop-mirror, I may remark that in my opinion they ultimately go back to Chinese mirrors. The earliest Chinese loopmirrors precede those of the Scythians by at least half a millennium.[2034] It is hardly a coincidence that the earliest datable Sarmatian loop-mirror, the Mozhary mirror, is an imitation mirror. Unadorned loop-mirrors, which might be the forerunners of the Sarmatian loop-mirrors, have been unearthed in Central Asia. Werner refers to one excavated in the T’ien Shan by A. N. Bernshtam, who dated the grave between the fourth and third century b.c.[2035] Others were found in the upper Irtysh Valley,[2036] the Chu Valley in Kirgizstan,[2037] and in western Siberia.[2038] Dated in the sixth or between the sixth and fourth century, they may very well be later, though not much. But there exist also plain pre-Han Chinese mirrors, but they are little known; collectors, interested in beautiful decorations and inscriptions, paid no attention to them. In recent years such mirrors were unearthed in Ch’ang-sha,[2039] Hsi-an,[2040] and Ch’eng-tu;[2041] those from Hu-nan have tentatively been dated between the seventh and fourth century. Their relationship with the plain mirrors of the Western barbarians needs further investigation.

The preceding survey has shown once more how strong the influence of the civilizations of Central Asia, themselves in contact with China, was on the Sarmatians. Mirrors of the Mozhary type were fairly common in the lower Volga region in the Late Sarmatian period. They represent the earliest phase in the development of the loop-mirrors of the Sarmatians, forerunners of the type which Werner after the easternmost and westernmost findplace, Chmi in the Caucasus and Brigetio on the Danube, calls the Chmi-Brigetio type; its decor consists of a circle in the center and radial lines between it and the rim. The type Berezovka-Carnuntum is typologically more developed, but on the whole contemporaneous with Chmi-Brigetio. Still later, but again not much later, is the type Karpovka- St. Sulpice.

On the basis of the rich evidence, collected from often rather remote publications, Werner shows that the loop-mirrors of all three types spread from the east westward as the Huns did. None of the mirrors is, in his opinion, earlier than about 400 a.d. Their bearers were supposedly the Huns, from whom their Germanic subjects took over the mirrors.

More recent finds are incompatible with Werner’s thesis. Chmi-Brigetio mirrors occur in Sarmatian graves as early as the third century. One was found in a grave at the stanitsa Vorozhenskaya in the Kuban area which by its furniture, among other things an amphora, must be dated to the third century.[2042] The Sarmatians in the necropoles of Phanagoria and Tanais were buried together with their Chmi-Brigetio and plain loop[2043] mirrors before the 370’s. The Huns wiped out the Chernyakhov civilization, so the Chmi-Brigetio mirror in the Chernyakhov cemetery at Vorokh- tanskaya Ol’shanka southwest of Kiev[2044] was cast before the Hun storm. A “flat,” that is, unadorned mirror of 7.5 centimeters’ diameter, which originally had a loop on the back, was found in a building at Toprakkala in Khwarezm,[2045] datable to the middle of the third century at the latest.[2046] Werner is right: The loop mirrors came to the West together with the Huns. But they were not Hunnic mirrors. They were the mirrors of Sarmatians, who had them long before the Huns. We can be even more specific: They were the mirrors of eastern Sarmatians, those whom the Huns forced to join them east of the Don and those with whom they made an alliance on the Don.

These small bronze mirrors permit an answer to a question rarely asked. Where did the Huns cross the Carpathian Mountains into Hungary? Some hordes may have ridden through the passes over the southern Carpathians into Transylvania and from there into the Hungarian plain, but this would have been difficult for horsemen accompanied, as they most probably were, by their wagons. In the course of her history Hungary was repeatedly invaded from the northeast, through the valley of the upper Theiss: Ko- lomyya-Yablonsky (Tatar) Pass—► Sighet (Sziget)—► Khust (Huszt). The Huns and their Alanic allies took this route.

Studying the distribution of the Chmi-Brigetio and the Berezovka- Carnuntum mirrors south and west of the Carpathians, Ilona Kovrig noted that it almost coincided with the distribution of artificially deformed skulls.[2047] Loop-mirrors were found in several graves with deformed skulls. One group of mirrors, associated with silver fibulae, is rather dense in the Upper Theiss Valley, another one stretches north of the Danube from the bend at Waitzen to Vienna. Compared with these two groups, the number of loop-mirrors in the Danube Valley and the great Hungarian plain is insignificant. From this distribution Kovrig drew the conclusion that the greater part of the ethnic groups which brought the mirrors to Hungary came—probably in several waves—through the passes of the northeastern Carpathians. This argument is strengthened by the absence of pendant-mirrors in Hungary. They are, on the other hand, characteristic of the Late Sarmatian graves in Rumania where loop-mirrors do not occur. Had the Huns and their allies and subjects come from the southeast, Sarmatians in the Rumanian plains, forced or voluntarily, would have joined them. But not a single grave of the fifth century in Hungary contains a pendant-mirror.

Now we can take a step further. In the third century the loop-mirrors were Sarmatian. In Hungary they still were almost absent from the Hunnic heartland east of the Danube. In other words, even at a time when the Huns and Alans lived closely together, the Huns did not take over the Sarmatian mirrors. This, of course, does not mean that only Sarmatians had them. It is unlikely that all Hunnic women, out of national pride, refused to look into a loop-mirror. The one found in Strazhe in Slovakia[2048] comes from a grave in which a racially mixed Europoid-Mongoloid individual was buried. Many fifth-century graves with loop-mirrors were Germanic. If Goths and Gepids in Hungary followed the Sarmatian custom, the Huns could not reject it forever.[2049] Still, chances are that the graves with loop-mirrors are not Hunnic. We have gained another criterion for separating Hunnic and non-Hunnic finds.

The observation that many loop-mirrors were intentionally broken when they were put in the grave led archaeologists to all kinds of speculations.[2050] The plethora of ethnographical parallels and the lack of at least relatively constant association of the custom with other features in the archaeological material account for the futility of such often ingenious and erudite essays.

Personal Ornaments

Gold Plaques on Garments

In Attila’s time and long before it, the custom of sewing small stamped gold plaques on garments was widespread in the barbaricum. To trace it back to its origin is not our task, nor need we investigate who the givers and takers in each case were. The Vandals in Slovakia[2051] may have adopted the fashion from the Jazygi[2052] and carried it to Africa,[2053] although some plaques found there could have been Alanic. The Kushans who had rosettes and ringlets sewn on their coats possibly imitated the Parthians;[2054] the small plaques on the coats of nobles in Parthian costume in Hatra[2055] are doubtless of gold, and in Sirkap small gold rosettes were found in a Parthian stratum.[2056] But both Kushans and Parthians independently may have followed an older Central Asian fashion. In a grave at Kyzyl-kyr in the Bukhara oasis, datable between the third and second centuries b.c., ninety small hemispherical gold plaques lay on the chest of a woman.[2057]

The custom is well attested for the Sarmatians as early as the Sauroma- tian period,[2058] for the Scythians (see for example, the plaques in the Chastye kurgan),[2059] and in Khwarezm.[2060] In the Middle Sarmatian period, garments were decorated with gold plaques from the Volga (Kalinovka,[2061] Berezovka)[2062] to the Ukraine (Svatova Luchka and Selimovka).[2063] The garment of a woman in a Late Sarmatian grave at Wiesenmiiller (Lugovoe) on the Eruslan was richly decorated with gold and silver plaques.[2064] Such plaques are also known from the possibly Hunnic burials at Shipovo[2065] and Novo-Grigor’evka.[2066]

In the unquestionably Hunnic find from Szeged-Nagyszeksos occur twenty-six electron plaques and seventeen fragments.[2067] They are square; a beaded frame encloses four triangular faces meeting at a point; the corners are pierced. The Hunnic plaques are identical with those from Pusztabakod and from Carthage.[2068] As the similarity, at times amounting to identity, of the gold plaques in the fourth and fifth centuries from one end of the barbaricum to the other proves, they were the products of Roman workshops using the same technique and the same patterns. Among the plaques on the dress of the Germanic or Alanic lady from Airan in Normandy[2069] are some exactly like plaques from Novo-Grigor’evka;[2070] others have counterparts in Papkezsi in Hungary[2071] and Panticapaeum.[2072] In other words, the Huns of the fifth century followed an “international” fashion.[2073]


“In the house of Queen Ereka, maidservants, sitting on the floor in front of her, were embroidering with color fine white linen to be placed as ornaments on the barbarian clothes.”[2074] Spherical, cylindrical, and flat embroidery glass beads are known from most Middle and Late Sarmatian women’s graves, even the poorest ones. They were all imported. They were sewn on the shoes, the lower part of the trousers, the sleeves, the collar of the tunic.[2075] In grave Fl 6 at the khutor Schulz (now sovkhoz Krasnyi Oktyabr’) on the Torgun, almost seven hundred, mostly green and blue ones, lay near the feet of the woman;[2076] they were in the same position in grave 3 in the second- or third-century cemetery Bel’bek II in the Crimea.[2077] In some cases the shoe soles could have been embroidered as in Pazyryk;[2078] the woman sat crosslegged on the floor.

Huns and Sarmatians shared their love for multicolored articles of dress with many northern barbarians. On the silk cloths in the Hsiung-nu graves at Derestui, beads of carnelian, jasper, gilded glass, limestone, and paste were sewn.[2079] In Noin Ula, very small perforated pyrite crystals were found, originally fastened to cloth or leather.[2080]


Like their Sarmatian sisters, the Hunnic women wore necklaces and bracelets, perhaps also anklets, of beads of all sorts of material: coral, carnelian, mother-of-pearl, quartz, pyrite, lapis lazuli, Egyptian paste, amber, lignite, but also stone and clay.[2081] Only the latter were homemade; the others came from all parts of the Roman Empire, Persia,[2082] Khwarezm,[2083] India,[2084] and also the barbaricum itself. By the fifth century a good part of the amber, worked into beads or used for inlays, came from the banks of the Dnieper and other places in the Ukraine.[2085] Lignite[2086] seems to have been imported from the Caucasus where lignite beads are quite common. Bracelets and necklaces formed by amber, glass, and semiprecious stones were worn throughout the northern steppes, as far east as Tuva and Outer Mongolia.[2087]

VIII. Race

The following investigation is largely based on paleoanthropological evidence.[2088] To the reader who has been exposed to so much that was merely a reasonable guess, exact measurements must come as a relief. The date of a battle may be controversial, but the naso-malar angle and simotic height of a skull are never in doubt. And yet the many hundred pages and the tens of thousands of figures with which the paleoanthropologists overwhelm us are of little value for historical studies unless they are supplemented by literary and archaeological evidence. Even if, for instance, the number of skulls from the thirteenth-century graves between the Kerulen and the Volga were twenty times greater than it is now, they would be useless in retracing the campaigns of Genghiz Khan, Batu, and Subotai. Mongoloid skulls of the paleo-Siberian type in the Avar graves in Hungary prove that one group of the multiracial hordes came from northeastern Asia, but they cannot tell us when these Mongoloid Avars left their pastures and over which routes they reached the middle Danube. These are limitations which are almost self-evident, but the historian faces other difficulties which he is well advised to recognize in order not to set unrealistic hopes in paleoanthropological studies.

Paleoanthropology is a relatively new science, and its terminology is still fluid. At times this can be rather bewildering. To give examples which refer directly to our problems, Nemeskeri regards the “Ural-Altaic” or “Sub-Uralic” type as Mongoloid;[2089] other anthropologists assign it to an intermediate position between Mongoloids and Europoids (or Europeids

or Europids; there is not even general agreement on what the adjective should be). Debets’ distinction between the paleo-Siberian and Baikal type[2090] is ignored by others. “South Siberian” and “Turanian” mean the same, but there is no equivalent to the “Tungid” type of the Hungarian anthropologists in Soviet taxonomy, although its “short-faced” Mongoloid type seems to be the same; Debets’ suggestion to call it the Katanga type has not been generally accepted.[2091]

In the present studies mainly the paleoanthropological material from the Soviet Union will be discussed, so I adhere to the terminology used in Osnovy Antropologii by Roginskii and Levin, and Ethnic Origins of the Peoples of Northeastern Asia by Levin.

“Great race” designates the three basic racial divisions of mankind, the Negroid, Europoid, and Mongoloid; “race,” the large subdivisions within the great races. Thus the Mongoloid great race comprises, among others, the North Asiatic, Arctic, and Far Eastern (Sinid) races. Within the races “types” are distinguished, for example, within the North Asiatic race, the Baikal and Central Asiatic types.[2092]

The paleoanthropological findings permit only a partial reconstruction of the physical appearance of the people. They remain silent about so much one would like to know; the color of the skin, eyes, and hair; the shape of the lips and eyelids; the patterning of the subcutaneous fat, to mention some of the characteristics by which, without measuring the skull, we can tell between, say, a Russian from Vologda and a Madrileno.

For reasons I do not quite understand the Soviet paleoanthropologists are exclusively, or almost exclusively, interested in skulls. This is all the more regrettable as stature is often of considerable importance for the racial diagnosis. To give an example, the burials in the kurgan cemetery at Shipovo take a prominent place in Hunnic studies. The furniture in kurgans 2 and 3 has been minutely described by Minaeva.[2093] Maslovski carefully measured the skull from kurgan 3.[2094] But only Rykov gave the length of the skeletons. The woman in kurgan 2 was 176 centimeters, the man in kurgan 3 was 170 centimeters tall; the man in kurgan 2 had the imposing height of 185 centimeters.[2095] These people could not be Huns, who were exigui forma, of small stature, as Jordanes said.[2096]

In the evaluation of the paleoanthropological evidence one must, furthermore, not lose sight of the fact that the reconstruction of the racial history of the Eurasian steppes rests on a narrow base. According to the Han shu, the Wu-sun numbered 630,000,[2097] which is of course too exact; who could have counted them? Still, the figure probably was in the neighborhood of half a million. In the five centuries we can follow the history of the people, there lived several million Wu-sun. But to date not even two hundred of their skulls have been found. In 71 b.c., the Wu-sun took 39,000 Hsiung- nu prisoners.[2098] Where are their skulls? About 150 b.c., the Chinese princess Hsi-chiin for political reasons had to marry a Wu-sun king.[2099] She came to his tents with several hundred servants and eunuchs.[2100] It was sheer luck that in the Wu-sun graves at least one Chinese skull was found.

Finally, it must not be overlooked that the graves can very rarely be dated as exactly as the historian would wish. The skull in kurgan 12 at Kurgak in the Alai Valley is artificially deformed.[2101] Bernshtam dated the grave to the third century b.c., which puzzled Ginzburg, for cranial deformation was supposed to make its appearance with the coming of the Huns in the first century b.c. So he called this premature occurrence an echo, otgolosok, of the connections of the Kurgak people with the Huns,[2102] although the echo does not precede the sound. Later Bernshtam changed his mind and dated the kurgan to the beginning of our era.[2103] Perhaps he was right this time, perhaps not. I do not want to be misunderstood. The paleoanthropological contributions to the study of the Huns cannot be overrated, but the uncertainties inherent in them must not be overlooked either. They can be somewhat reduced if the written sources come to our help. We now turn to them.

There exist four descriptions of the appearance of the Huns. The first and earliest one, written by Ammianus Marcellinus[2104] in the winter of 392/3, was paraphrased by Jerome[2105] and Claudian.[2106] The second was the Gaulish writer Sidonius Apollinaris; although some of his expressions were taken over from Claudian, his description of the Huns is based on autopsy.[2107] The third picture of the people was drawn by Jordanes,[2108] who must have seen Huns in the East Roman army. His portrait of Attila,[2109] however, goes, through Cassiodorus, back to Priscus, our fourth source. As the king “showed the evidence of his origin,” we may take what Priscus said of him to be racial characteristics of the Huns.[2110]

Ammianus’ description begins with a strange misunderstanding: “Since the cheeks of the children are deeply furrowed with the steel from their very birth, in order that the growth of hair, when it appears at the proper time may be checked by the wrinkled scars, they grow old without beards and without beauty, like eunuchs.” This was repeated by Claudian and Sidonius and reinterpreted by Cassiodorus. Ammianus’ explanation of the thin beards of the Huns is wrong. Like so many other people, the Huns “inflicted wounds on their live flesh as a sign of grief when their kinsmen were dying.”

Ammianus not only misinterpreted the Hunnic custom; his description of the Huns as beardless is at variance with Priscus. Ammianus may have seen an occasional Hunnic mercenary; in the main he had to rely on his Gothic informers. Priscus, in contrast, was personally acquainted with Attila, his sons, his uncles, and many Hunnic dignitaries. Attila, Priscus wrote, had a thin beard, rarus barba. To a Roman of the fifth century, a time when the beard was valued as a sign of manhood, indicium virilitatis, as Jerome said,[2111] the beards of the Huns may have looked sparse. But Attila did not look like a eunuch. His thin beard was not necessarily a racial characteristic, a Mongoloid feature as has been maintained, any more than the sparse beard of Mynheer Pepperkorn in Thomas Mann’s Magic Mountain. The definitely Europoid Scythians were often depicted with thin beards.[2112] Besides, Ammianus speaks of the hairy legs, hirsuta crura, of the Huns.

That in the eyes of the Romans and Germans the Huns were an ugly crowd[2113] does not mean much, and when Ammianus compares them to “the stumps, rough-hewn images, that are used in putting sides to bridges,” he evidently wants to emphasize the coarse features of the Huns. Only reluctantly he has also two good words for the hated savages: They have compact, strong limbs and, like Ammianus’ beloved emperor Julian,[2114] strong necks.

The wide shoulders and the broad chest, scapulis latis (Jordanes), lato pectore (Priscus), insignes umeri, pectora vasta (Sidonius) are for the racial diagnosis as irrelevant as the narrow waist, succincta sub ilibus alvus (Sidonius). The great sitting height might be of more importance: “The figure of the foot soldiers is of medium height, but it is elongated if you look at the horsemen. Thus they often are considered tall when they are sitting” (Forma quidem pe- diti media est, procera sed extat, si cernas equites; sic longi seape putantur, si se- deant, Sidonius). Like the Huns, the Bashkirs, with their considerable Mongoloid admixture, are long-bodied, well muscled, and robust, with wide shoulders.[2115] But the Belgian Flemings and Walloons also are described as “moderately thick-set in bodily build; their shoulders are broad, and their relative sitting height great.”[2116]

Jordanes stressed the small stature, exigui forma, and the swarthy complexion of the Huns, species pavenda nigridinis; Priscus described Attila as swarthy, teter color, and of short stature, forma brevis. Althias, commander of the Hunnic auxiliaries in Belisarius’ army, was “lean and not tall of body.”[2117] Asterius of Amasea called the Huns nimble and slender.[2118] But Emperor Arcadius was also of short stature and dark complexion.[2119] Ammianus called the Persians subnigri;[2120] Emperor Valens was nigri colorist so was the Egyptian philosopher Pamprepius,[2121] whom Hodgkin[2122] took for a Negro.[2123] Whereas their height and the color of their skin did not markedly set the Huns apart from many Romans, the difference between them and their Germanic and Alanic white-skinned and tall subjects and allies must have been striking. The Alans were a tall, blond people.[2124] In the Middle and Late Sarmatian graves in the Volga region lay men as tall as 182, 185, 187, and 189 centimeters.[2125]

Only the statements about the heads and the physiognomy of the Huns are really revealing. The heads were round and shapeless (informis offa, Jordanes), “a round mass rises into a narrow head” (consurgit in artum [or arcum] massa rotunda caput, Sidonius); the eyes small and deepset: “tiny eyes, perforations rather than lights” (minutis oculis, havens magis puncta quam lumina, Jordanes), “their sight is there in two hollows beneath the forehead; while the eyes are not visible, the light that enters the dome of the skull can hardly reach the receding eyeballs” (geminis sub fronte cavernis visus adest, oculis absentibus acta cerebri in cameram vix ad refugos pervenit orbes, Sidonius). The nose was flat; this follows from Sidonius’ description of the way the skulls of the children were deformed, and Jordanes, quoting Priscus, says expressly that Attila had a flat nose, semo nasu.

The weakly accentuated profile, together with the small eyes, point to a Mongoloid strain in the Huns. How strong it was cannot be determined from the few words in our sources. The more pronounced racial features in a mixed population always attract the most attention. Movses Das- xuran^i ignored the Europoids among the Khazars and described the whole people as “an ugly, broad-faced, eyelashless mob”.[2126] The women in the Kiptchak horde, wrote William of Rubruk, were exceedingly fat “and the smaller their noses, the fairer they were esteemed”;[2127] he was so impressed by the flat Mongol faces that he had no eyes for the non-Mongols who constituted the majority of the population.[2128] One must also not forget that Ammianus and Jordanes hated the Huns with such an intensity that, however the savages may have looked, they had to be depicted as subhuman monsters. A comparison between Ammianus’ and Jordanes’ descriptions of the Huns and what Western chroniclers wrote about the Magyars is instructive. To the Germans and Italians the Magyars were “a monstrous nation, a horrid tribe, a tribe more cruel than any wild beast” (mostrifera natio, horrenda gens, gens omni belua crudeliof). Crossing Hungary on his voyage to the Holy Land, Otto of Freising admired God’s patience in giving so beautiful a country not to human beings but such monsters.[2129] But Gardizi, a disinterested observer, called the Magyars handsome and pleasant-looking.[2130]

Ammianus and Jordanes may be forgiven, but what excuse have modern authors who ascribe to the Huns swollen lips, beady eyes, and bandy legs?[2131][2132] Eickstedt’s mistranslation of the Latin texts is fantastic; Attila had aus- einanderstehende Zahne, which pretends to be the translation of canis aspersus, “sprinkled with gray,” said of his beard.[2133]

The descriptions give a somewhat distorted picture of the Huns. What is known about other steppe peoples of northern Eurasia in the first millennium a.d. makes it unlikely that the Huns were as Mongoloid as, say, the Yakut or Tunguz of our times. Many Huns were halfbreeds. Balamber married a Gothic princess,[2134] Attila’s last wife had the Germanic name Ildico,[2135] the Gepid Mundo was of Attilanic descent.[2136] Though we do not hear of Alano-Hunnic marriages, the Mongoloid strain in the Alans of Sa- paudia shows that such marriages were fairly common. The leader of Stilicho’s Alanic auxiliaries was a small man;[2137] among his ancestors were probably Huns.

Most large cemeteries of the post-Hunnic centuries in the steppes reveal a mixture of races. The Gepidic cemetery at Kiszombor shows ein Rassenkonglomerat, das sich ausden Elementender nordischen, mediterranen, osteuropiden, turamiden, mongoliden und palaoasiatischen Rasse zusammen- setzt.[2138] In their Scandinavian home the Gepids may not have been pure Nordics, but there were no Mongoloids among them; in Hungary they mixed with the Huns. In the Avar cemeteries; next to Europoids, at least four Mongoloid types are represented: Sinid, Baikal, Tungid, Yenisei.[2139] In the cemetery at Kyukyal’dy in the valley of Kzyl-Alai, datable to the sixth and seventh centuries, Mongoloids with both wide and narrow faces were buried side by side with Europoids of the Andronovo and protoMediterranean type with varying degrees of Mongoloid admixture, testifying to the complex composition of some groups in the western Turkish kaganate.[2140]

The paleoanthropological evidence indicates that the Huns were likewise racially mixed. In 1939, when Bartucz published his fundamental study on the races in Hungary, he did not know “of a single skull which could, beyond any doubt, be regarded as Hunnic.”[2141] This is still true.

The chances that someday a tombstone will be found with the inscription Hie iacet ... genere Hunus and a well-preserved skeleton beneath it are slim. Yet the situation is not as bad as it looks. The following list of non- Europoid skulls in graves of the Hunnic period is probably not complete,[2142] but it suffices for our purposes:

Vienna-Simmering: Skull of a mature man. “Alles deutet darauf hin, dass wir einen Mongo len oder Mongo liden vor uns haben”[2143]

Strazhe I near Piest’any, Slovakia: woman. E(uropoid)+M(ongoloid).[2144]

Besenov V, district Surany, Slovakia: man. E-|-M.[2145]

Adony, Hungary: One artificially deformed skull of a child which “seems to belong to the Europid type.” Of the twenty-one skulls not deformed, “ten are dolichocranic, six mesocranic, and four brachycranic. In one case it was impossible to determine the index. As to the distribution of varieties, the Europoid type is represented by the Northern, the Mediterranean, and the East-Europid varieties. In the case of four skulls, we have to do with the so-called dolimorphic Ural-Altaic or Sub-Uralic varieties of the Mongolid type. The skulls belonging to this type are characterized by a long and moderately wide cranium cerebrale (mesocrany); by a low cranium viscerale,by amoderately vaulted forehead, and pronounced browridges.”[2146]

Gy dr, Szechenyi Square, Hungary: Twenty-three skulls from a cemetery in and outside of a Roman camp. One artificially deformed skull of a child. “The skulls belong to the Europid and Mongolid types, represented by six skulls each. No clear assignation to types was possible in the rest of the cases. The Mongolid varieties show a predominance of Tungid characteristics. Special importance attaches to the skulls found in graves nos. 9 and 21: these skulls belong to the dolichocranic Mongolid type. The closest parallel is the classical type found in the Avar cemetery at Mosonszentjanos.”[2147]

Dulceanca, rayon Ro§iori in Muntenia, Rumania: Deformed skull of a man of about fifty years. E-|-M.[2148]

Although none of these finds can be dated exactly, they cannot be earlier than the last quarter of the fourth century. No Mongoloids lived between Vienna and Dulceanca before the coming of the Huns. On the other hand, the locations and the grave goods preclude the possibility of dating the skulls later than the fifth century. They are those of Huns or people who came with the Huns.

The descriptions and racial diagnoses which have been quoted verbatim require some comment. There are first the skulls that show both Europoid and Mongoloid features. Some anthropologists refuse to go beyond the statement that in a given skull characteristics of the two major races can be discerned. The artificial, and in particular the circular, deformation affects nearly all cranial indices to such a degree that it is often impossible to determine even the major races.[2149] If, in addition, a deformed skull shows, or seems to show, features of both major races, the diagnosis of the types becomes an extremely difficult task. Most Soviet anthropologists are content with classifying such skulls as Europoid-Mongoloid.

In the list of skulls of the Hunnic period, I did not include the deformed skulls from Szekszard, Mohacs, Gyongyosapati, and Szirmabesenyo. Nemeskeri thought he could detect Mongoloid features in them. Werner accepted his diagnoses and drew from them far-reaching conclusions.[2150] But the diagnoses seem to be wrong. Liptak measured the skulls again, and his results were quite different from those of Nemeskeri. According to Liptak, none of the skulls shows any Mongoloid admixture.[2151] Of those from Strazhe and Besenov which Vliek took for Mongoloid,[2152] Liptak accepted only two as E-|-M.[2153]

The historian finds himself in a quandary. Whose judgment should he believe? The Soviet anthropologists whom I asked were inclined to take Liptak’s side. Fortunately, the situation is not hopeless. For even after the elimination of the controversial skulls, there remain a number of Mongoloid and E-]-M skulls datable to the Hunnic period. To be sure, the possibility that one or another of the supposedly Mongoloid skulls may turn out to be E+M or even Europoid cannot be ruled out. However, it is unlikely that all diagnoses were wrong. Nemeskeri could not have been mistaken when he found the closest parallels between two skulls from Gyor and the Avar skulls from Mosonszentjanos. It is by now generally agreed that the latter are of the Baikal type.[2154]

The material from Hungary, Slovakia, and Rumania is by far too small to determine the numerical relationship of the various races in the Hunnish hordes. Besides, most of the skulls come from the graves of poor people. The prominent Huns, or, to be more cautious, some of them, cremated their dead. Some E-f-M skulls might also be Alanic. There were individuals of the South Siberian type among the Sarmatians at Kalinovka in the Volga region. The skulls in the graves at Saint Prex, canton Vaud, with their considerable Mongoloid admixture, were in all probability the skulls of Alans or descendants of Alans. Such a halfbreed was also the man in whose grave at Vienna-Simmering objects were found[2155] that could be Hunnic. The man himself was 180 centimeters tall,[2156] thus clearly not a Hun.

The Hsiung-nu

Until the 1940’s, the identity of the European Huns with the Hsiung- nu on China’s borders was rarely questioned. As no one doubted that the Hsiung-nu were Mongoloids, the Huns must have been Mongoloids too. Are there paleoanthropological finds to reconstruct the routes over which they migrated into eastern Europe?

The answer given by A. N. Bernshtam in 1926 was for a while widely accepted: In the last century b.c., Hsiung-nu were supposed to have moved to eastern Middle Asia and from there spread westward. Bernshtam’s thesis centered on a catacomb in the cemetery on the Kenkol River in the Upper Talas Valley. Bernshtam excavated kurgan 10. “In the catacomb,” he wrote, “lay two Mongoloid skeletons with deformed skulls; the skeletons in the dromos were Europoids, apparently slaves from the local population of the Pamiro-Fergana race.”[2157]

Bernshtam was an excellent and indefatigable excavator who went on digging when he hardly could walk any more; he died from cancer at the age of forty-six. Bernshtam was also a courageous man. He defended the views of the eminent but often mad linguist N. Marr at a time when so many Soviet scholars who had praised Marr to heaven were kicking the dead lion after Stalin had branded him an anti-Marxist. But Bernshtam wrote in too great haste, reconstructing whole periods of world history on the narrowest foundations. His interpretation of the Kenkol finds is a telling example. The two Mongoloids became in no time Turkishspeaking Hsiung-nu, and the Europoids in the dromos Wu-sun slaves. Because the Mongoloids were buried in catacombs, all catacomb burials in Middle Asia were declared Hsiung-nu burials. The shepherds from Kenkol were the missing link between the Hsiung-nu in Mongolia and the Huns in Hungary.

Zhirov doubted Bernshtam’s interpretation as early as 1940.[2158] But it won, as I said, wide acceptance both in the Soviet Union and in the West. By now it is practically abandoned. A closer study of the Chinese annals led S. S. Sorokin[2159] and N. Negmatov[2160] to doubt that the Mongoloids in the Talas Valley had anything to do with the Hsiung-nu of Chih-chih’s shortlived robber state as Bernshtam thought. The date of the finds suggested by Bernshtam became questionable. Gryaznov proved that the “slaves” in the dromos belong to a secondary burial.[2161] Finally, the alleged difference between the “lords” and the “slaves” turned out to be nonexistent. Debets measured the horizontal profiles of the couple in the catacomb and the two men in the dromos.[2162] They are as follows:

Naso-malar Zygo-maxillary Dacryal Simotic
angle angle height height
Catacomb, man 141 129 13.3 4.4.
Catacomb, woman 133 132 13.9 4.0
Dromos, man 140 139 12.3 2.8
Dromos, man 140 132 11.1 3.2

The angle of nasal prominence of the skulls in the catacombs is 26, of those in the dromos, 26 and 25. In other words, there are no real differences between the “lords” and the “slaves” in the degree of the horizontal profile of the face. The ones are not more Mongoloid than the others. All four skulls are Europoids with some Mongoloid admixture.

Debets’ almost indignant refutation of Bernshtam’s thesis of course does not solve the problem of the Kenkol finds. Where did the Mongoloid admixture come from ? The wider question still remained whether the Mongoloids in the graves in Hungary had anything to do with the Mongoloid Hsiung-nu.

The number of Hsiung-nu skulls is still small but large enough to draw from them rather important conclusions. Debets measured sixteen from the kurgans in the Selenga Valley near Ust’-Kiakhta, between 1897 and 1903 excavated by the Polish anthropologist Talko-Hryncevics (in Russian transcription, Tal’ko-Gryntsevich), and a female cranium from Noin Ula, found by the Kozlov expedition in 1925.[2163] The skull of a man, found by the Hungaro-Mongolian expedition in Noin Ula in 1961, has been measured and described by T. Toth.[2164] He found in it the features of the Baikal (paleoSiberian) type: dolichocephalic, low skull, high and orthognathous face, very slight horizontal profile, that is, a very flat face and a broad, flat nose, sloping forehead, strong browridges. The other skull from Noin Ula is of the same type; so are the skulls from the Selenga Valley, although among them one has somewhat attenuated Mongoloid features (as, possibly, the whole series).[2165] The skull from the Ivolginskoe gorodishche which Gokhman studied is likewise of the Baikal type.[2166]

The earliest Baikal skull was excavated in 1952 in a cave near the Shilka River; Okladnikov dates it to the Glazkovo period (about 1700–1300 b.c.), though it might be later.[2167] The skulls from the slab graves in Transbaikalia of the beginning of the Iron Age (fourth to second century b.c.) are of greater importance to us. They are the low-faced skulls of the pre- Hsiung-nu population of the area.[2168] When the Hsiung-nu came, the lowfaced skulls gave way to the high-faced ones of the Hsiung-nu. In the early and the beginning of the later Han period a great part of the Hsiung- nu confederacy, perhaps we may say its nucleus, consisted of Mongoloids of the Baikal type. This does not make all Mongoloids of the Baikal type into Hsiung-nu. Nor does it prove that all members of the confederacy were of the Baikal type. Besides, what was true for the last two centuries b.c. and the beginning of our era was not necessarily true for the third and fourth centuries. We turn to the written sources and the archaeological monuments.

Europoids in East Asia

A stone horse at a tomb in the valley of the Wei River in Shensi is trampling a barbarian under its hoof.[2169] The tomb has been identified as that of Ho Ch’ii-ping, who died in 117 b.c., the great general famous for his victories over the Hsiung-nu. Although the exact date of the sculpture is not quite certain,[2170] it is doubtless of the Han period.[2171] The general buried under the earth mound was perhaps not Ho Ch’ii-ping, but he must have been an outstanding man, and the enemy was definitely a Hsiung-nu. He has a flat face and prominent cheekbones, but a luxuriant beard which is quite un-Mongoloid.[2172] In this respect he closely resembles the horseman on a small bronze plaque found by P. S. Mikhno near Troitskovavsk in Transbaikalia (fig. 74).[2173] A bronze in the British Museum, from the Ordos region, which was for a long time held by the Hsiung-nu, represents a Europoid; note the thick moustache and the wide open eyes (fig. 75).

Fig. 74. Small bronze plaque showing a horseman with prominent cheekbones and full beard, from Troitskovavsk in Transbaikalia. From Petri, Dalekoe proshloe Pribaikal’ia 1928, fig. 39.

The Mongoloid elements in the Hsiung-nu were considerably strengthened by the many Chinese renegades[2174] and prisoners of war. Of the Hsiung-nu’s Ch’iang, Ta Hu, and Ting-ling slaves in the third century,[2175] the Ch’iang were almost certainly Mongoloids. But from their raids into the oasis cities of Hsin-chiang,[2176] the Hsiung-nu must have brought back quite a number of Europoids. A double burial in the desert region north of Min- feng hsien is instructive. The polychrome silk, jackets, trousers, stockings, and shoes are the same as in Noin Ula. But on a fabric a man is represented whose features are distinctly Europoid. The couple in the grave was also Europoid.[2177]

Fig. 75. Bronze plaque from the Ordos region, showing a man of Euro- poid stock with wide open eyes and moustache. British Museum. Photo G. Azarpay.

We are not concerned here with the first appearance of the Europoids on the borders of China. Two references will suffice to indicate the problem. Karlgren pointed out that the bronze figure of a kneeling man from one of the Chin Ts’un graves, datable between about 450 and 230 b.c., does not represent a Mongoloid;[2178] I would rather say that the flat face is Mongoloid, but the wide open eyes are Europoid. The hunter on an often reproduced gold plaque in the Siberian collection of Peter the Great[2179] is undoubtedly Europoid. The plaque has been dated between the third and first century b.c., if not earlier.[2180]

As the account of the massacre of the Hsiung-nu Chieh in Chao in 349 a.d. shows, the great majority of that people were Europoids. When Jan Min made himself lord of Chao in northern Honan, which until then had been ruled by the Chieh, he ordered the extermination of all Chieh. In and around Yeh more than two hundred thousand were slain. The Chieh soldiers were recognized by their high noses and full beards.[2181]

Uchida Gimpu[2182] and I,[2183] independently of each other, adduced this characterization of the Chieh as proof of the existence of a Europoid group among the Hsiung-nu in the fourth century.[2184] This was rejected by Tsunoda Bumie, who maintained that the Chieh were not of Hsiung-nu origin,[2185] and again by S. G. Klyashtornyi with reference to Yao Wei-yuan, who tried to prove that the Chieh were originally Yiieh-chih.[2186] Taking one step farther, Pulleyblank declared the Chieh to be Tokharians.[2187]

It is entirely possible that the Chieh were ethnically different from other Hsiung-nu; but this does not change the fact that they were, one of the nineteen tribes of the Hsiung-nu. When they joined the Hsiung- nu confederacy is not known. At any rate, by the middle of the fourth century there were Europoids among the Hsiung-nu.

Liu Yuan, the Hsiung-nu conqueror of Lo-yang in 311, was 184 centimeters tall; there were red strains in his long beard.[2188] The Hsiung-nu Ho-lien Po-po, founder of the short-lived Hsia dynasty, a contemporary of Attila, was 195 centimeters.[2189] Some T’u-yu-hun princes were also very tall.[2190] The Mu-jung T’u-yu-hun were a branch of the Hsien-pei. An anecdote in the Shih-shuo hsin-yii, compiled by Liu Yi-ch’ing in the first half of the fifth century, shows that the Hsien-pei, who are supposed to have spoken a Mongolian language, were racially anything but Mongoloid. When in 324 Emperor Ming, whose mother, nee Hsiin, came from the Hsien-pei kingdom of Yen, heard about the rebellion of Wang Tun, he rode into the camp of the rebels to find out their strength. He rode in full gallop through the camp. His puzzled enemies thought he was a Hsien-pei because of his yellow beard.[2191] One would like to know from what tribe Sakanoke no Tamuramaro’s “Chinese” ancestor Achi no Omi came; he had a reddish face and a yellow beard.[2192]

The T’ang period falls outside the framework of the present studies. I mention only in passing the Europoid “Tokharians,” depicted with their red hair and green eyes on the wall paintings in northern Hsin-chiang. K. I. Petrov thinks the Chinese misinterpreted the ethnic name (which, according to him, means “the red ones,” after the red color of the earth), and ascribed to the people red hair ![2193] The barbarian horsemen from Yu-chou in a poem by Li Po, probably Turks, had green eyes. Even later the Chinese knew of Mongol Huang t’ou Shih-wei, “Shih-wei with the yellow heads,” and Gengiz Khan and his descendants had blond or reddish hair and deep-blue eyes.[2194]

One could think that the Europoid Hsiung-nu were originally members of subjugated tribes, prisoners of war, or slaves. Some probably were. But Chin-jih-ti, 191 centimeters tall, a contemporary of Ho Ch’ii-ping, was crown prince of the Hsiu-t’u, a royal branch of the Hsiung-nu.[2195] After the conquest of present Tuva by the Hsiung-nu in the second century b.c., the population, which had been racially mixed with a preponderance of Europoid features,[2196] became not less but more Europoid.[2197]

Yen Shih-ku’s often quoted descriptions of the Wu-sun, neighbors and hereditary enemies of the Hsiung-nu, seems to prove that at one time the Wu-sun were preponderantly Europoid: “Of all the Jung of the western lands the Wu-sun look the most peculiar. Those of the present Hu who have cerulean eyes and red beards and look like Mi monkeys[2198] are their descendants.”[2199] Yen Shih-ku (579–645) evidently relied on an earlier source. But is the earlier source reliable?

Already at a time when only a small number of skulls from the territory held by the Wu-sun were known, they were recognized as Europoid.[2200] Debets admitted a slight Mongoloid admixture. The Wu-sun were not as purely Europoid as the preceding Saka, who looked like Afghans or North Indians, but “physiologically the Wu-sun resembled the present day clanless Uzbeks or Fergana Tadjiks, that is, the Europeoid features were still decidedly prominent.”[2201] As the material accumulated, local differences turned out to be more prominent than it was first thought. The development also did not go in the same direction. As late as the third century some Wu-sun were almost purely Europoid, whereas others were of the South Siberian type, that is, with a marked Mongoloid admixture.[2202] Still, there was nothing in the material that would have confirmed Yen Shih-ku’s statement until the young Kazakh anthropologist, 0. Ismagulov, published the results of his studies. Of eighty-seven skulls from graves in the Se- mirech’e, six, datable around the beginning of our era, were either of the European type or close to it.[2203] These Wu-sun did not resemble Uzbeks or Tadjiks; they were people with “cerulean eyes and red beards.”

The paleoanthropological work in Hsin-chiang has barely begun. It is, therefore, all the more remarkable that some of the skulls collected by the Sino-Swedish Expedition in 1928 and 1934 and studied by C. H. Hjortsjd and A. Walander point to Europoids of the northern type in the ancient population. Of the three skulls from Miran, datable between the last century b.c. and the third century a.d., one is probably Chinese, one probably Tibetan with a strong Nordic admixture, one preponderantly Nordic, possibly with some Indoid or Mongoloid features. In the third century Miran was a Tibetan fortress, so the Mongoloids were possibly soldiers of the garrison. The presence of Indoid features could be expected; the men on the third-century wall painting are Indians, the inscriptions are in Karosfhi.[2204] But the Nordic features come as a surprise. A skull from Charchan, unfortunately undatable, is predominantly Nordic, with Indoid and Mongoloid admixture. One of the earlier crania from the Lopnor region, presumably datable to the first three centuries a.d., is Mongoloid with some Nordic features. From the mass cemetery in the same region, which only approximately can be dated after 200 a.d., comes the skull of a Mongoloid with some Nordic features and another one which is Indoid with Nordic and weak Mongoloid admixture.[2205] Around the beginning of our era, Europoids of the Nordic type lived, thus, both in the Semirech’e and Hsin-chiang.

IX. Language

Speculations about the Language of the Huns

The Germans in Attila’s kingdom apparently did not use the script which Wulfila had invented to translate the Bible into Gothic; they scratched their runes on swords, lance heads, brooches, and buckles as their ancestors had done. The Huns, “barbarous even in the eyes of the barbarian peoples around them,”[2206] had no script. Attila’s scribes were not Huns but Romans: the Gaul Constantius,[2207] an Italian by the same name,[2208] the Pannonian Orestes,[2209] and Rusticus from Upper Moesia.[2210] In the middle of the sixth century Procopius described the Huns west of the Maeotis as “absolutely unacquainted with writing and unskilled in it to the present day. They have neither writing masters nor do the children among them toil over the letters at all as they grow up.”[2211]

All we know of the language of the Huns are names. Our sources do not give the meaning of any of them. These names have been studied for more than a century and a half.[2212] Some were assigned to this, others to that group of languages, from Slavic to proto-Chuvash and Old Khvar- telian.[2213] The task of the historian with some linguistic training or the philologist with a knowledge of history cannot consist of singling out this name or that and comparing it with what he happens to know. It should consist, rather, of studying the entire material in all its complexity. This has been done only once. Vambery listed not merely the names he thought he could explain but all he could find.[2214] His list is incomplete, and many of his etymologies strike us as fantastic. Yet methodologically Vambery was on the right track.

Although the present studies deal with the Attilanic Huns (to use the perhaps not quite correct but convenient term coined by B. von Arnim[2215]), the lists on the following pages also include names of other Huns. It has often been maintained, and I have said so myself, that the Byzantines spoke of Huns as loosely as they spoke of Scythians. This is true for later writers, but in the fifth and sixth centuries Byzantine authors definitely distinguished the Huns from other northern barbarians.

Priscus, who was interested in foreign languages, set Hunnish apart from other languages spoken at Attila’s court. During his stay with the Huns, and perhaps also before, he learned enough Hunnish and Gothic to be able to distinguish between them at least by their sound. He described how Zerco, the Moorish jester, threw the guests at the king’s banquet “into fits of unquenchable laughter by his appearance, his dress, his voice, and the promiscuous jumble of words, Latin mixed with Hunnish and Gothic.”[2216] By calling Edecon a Hun,[2217] Priscus implied that the man’s tongue was Hunnish.[2218]

Although Procopius’ definition of an ethnic group would not satisfy modern anthropologists, it is not as vague as it is sometimes presented. He wrote:

There were many Gothic peoples in earlier times, just as also at the present, but the greatest and most important of all are the Goths, Vandals, Visigoths, and Gepids. All these, while they are distinguished from one another by their names, do not differ in anything at all. For they all have white bodies and fair hair, and are tall and handsome to look upon, and they use the same laws and practice a common religion. For they are all of the Arian faith, and have one language called Gothic; and, as it seems to me, they all came originally from one tribe, and were distinguished later by the names of those who led each group.[2219]

Procopius applied to the Huns two of the four criteria of what constitutes a people in his view. Like the Goths, the Ovvvtxa eOvt] were characterized by their racial type—they were ugly and their bodies were dark; and by their manner of life—they were nomads.[2220] That Procopius passed over their religion is understandable: unlike the antagonism between Arianism and orthodoxy, it played no role in the relationship with the Romans. Nor had Procopius any reason to pay attention to the language of the Huns. As Belisarius’ consiliarius he had the opportunity to pick up some Gothic and possibly Vandalic; these were the languages of great kings and warriors. But it was not worthwhile to learn the gibberish which the uncouth Massagetic bodyguards spoke. To Procopius’ ear it must have sounded, to use a Chinese simile, like “the croaking of a shrike.” Yet he spoke of Hunnic peoples as he spoke of Gothic peoples. If the latter had one language, the same must be true for the former. In one instance we are explicitly told that the Kutrigur and Utigur, called Huns by Procopius,[2221] Agathias,[2222] and Menander,[2223] were of the same stock, dressed in the same way, and had the same language.[2224] “Same” does not necessarily mean identical. Vandalic was certainly close to Gothic but not the same. There may have been marked dialectical differences in the speech of the various Hunnic peoples and tribes, yet they apparently understood one another.[2225]

A little-noticed passage in John of Antioch sheds more light on the early Byzantine concept of the ethnic name “Hun.” In 513 Hypatius, the nephew of Emperor Anastasius, was made prisoner by Vitalian’s Hunnic federates. Polychronius and Martyrius “whose office it was to deal with the envoys of the Huns” (rd? twv Ovvvodv JtQeaPelat; £mr£TQa/j,£vvoi), were sent to the Huns with 1,100 pounds of gold to ransom Hypatius.[2226] This shows that among the interpretes diversarum gentium[2227] under the magister officiorum some were in charge of dealing with the envoys of the Huns. Not with this or that tribe, but with the Huns who evidently spoke one language.

The present investigation could not have been undertaken without Gyula Moravcsik’s invaluable Byzantinoturcica. They lead to the sources. Only by a careful study of the literary context in which the names appear can we hope to bring the problem of the Hunnish language closer to its solution. It is of little help to know the alleged Byzantine rules for transcribing foreign names. They change from author to author and from century to century. Before the twelfth century could render both foreign b and v. Sozomen has JBag5^ordv^;[2228]=Bar-Daisan, and Blxxwo,[2229] Priscus ’.4@(5a/?ov@to£[2230]=Ardabures in the Latin sources, and JBa2dju£go;[2231]=Va- lamer. Mzt for initial b appears for the first time in the twelfth century;[2232] the traditional transcription BovXyagot was retained much longer. Only by lumping all transcriptions together, from the earliest to the latest, and regardless of the language of the author, ranging from classical pure Greek to vulgar colloquial, can one say that a stands for a, o, u, e, a, i, and i, in Turkish names.[2233] What matters is the specific idiom of the writer, his dependence on earlier works, the manuscript tradition, and a number of other factors, to be discussed presently, which account for the form of a name in a text.


They were Tatos and Chales and Sesthlabos and Satzas (for I must give the names of the highest-born of these, although the elegant appearance of my history is spoiled by them).[2234]

It is a priori certain that the phonetic system of the Hunnish language, whatever it may have been, was different from that of Greek and Latin. Even if an author wanted to render a Hunnish name faithfully, the mere fact that he had to use the letters of his own alphabet forced him to distort it. A few names may have passed the process of transcription relatively unscathed; others must have suffered badly. What name is hidden behind Mdct/zi?? The name of Queen Erekan’s steward occurs only in Priscus, and in the dative case at that: ’^Ida/zet.[2235] -ig is not the Greek ending tacked on to the name. Priscus may not have been a good Christian but he must have heard of the protoplast. If the Hun’s name had been Adam, Priscus would have written ’Ada/j,. The Greek, having no letters for supradental s and palatal s’, transcribed these consonants by sigma. ’Adapts could be Adamis or Adamis (s, s). But because in the transcription of Germanic names the ending ip is sometimes rendered by -t;, ’Adapts could also be Adamip[2236]

Foreign names were not only adapted to Greek and Latin phonetics but also to the morphology of the writer’s language. The Byzantines often treated names ending in -an or -in as if they were in the accusative. If we had only the forms OvAdys and OvAdts,[2237] it would be impossible to determine whether the name of the Hun king was Uldis or Uldin. Fortunately Orosius mentions it in the nominative: It was Uldin.[2238] In some transcriptions the Greek and Latin endings can be relatively easily distinguished, but in others it is impossible to decide where the barbarian name ends. Procopius admired Belisarius so much that he even described the horse of his hero. “Its body was dark grey, except that the face from the head to the nostrils was of the purest white. Such a horse is in Greek called (pantos, the barbarians call it /Id^av.”[2239] Was it Balas, or Balan, or Bal? Balas is a Germanic word, OHG balas, equus maculosus, English blaze, German Bless.[2240] The word can be recognized because it occurs in a group of well-known languages. But what if the meaning of a name is as unknown as the language? The Hunnic names in the Latin and Greek sources can be reconstructed within limits, but these limits are rather wide. [v]Ha).aQ could be the transcription of Esl, Esla, Eslas, Esl, Esla, ESlas, Eslas, Eslas, Eslan, and ESlan. Was the stress on the first or the second syllable? Was the §—if it was an s—palatal or supradental? We do not know.

Besides the orthography of the writer and the possibility of morphological change, three more factors must be considered when we try to “retranscribe” Hunnish names. It is, first, not certain that all the names in our sources are those by which the Huns called themselves. Before the East Romans had any contact with the Huns, they heard about them from the Goths. They must have heard many names as they were pronounced by Goths and other non-Huns. Octar, the name of Attila’s paternal uncle, is a good example of the modification which a Hun name underwent in the course of transmission from Hunnish through Latin into Greek. Jordanes has Octar,[2241] Socrates OvmaQoi[2242]. These forms have a parallel in Accila and Optila. Eastern writers call the Ostrogoth, who killed Valentinian III, Accila or Occila; Marcellinus Comes, Jordanes, and John of Antioch call him Optila.[2243] The transition from -cl- to — pt- is characteristic of Balkan Latin.[2244] It was probably there that Octar became Optar — Uptar.

The second factor to keep in mind is the tendency of late Roman and Byzantine writers to alter foreign names until they sounded like Latin or Greek ones. In this way Bagrat became Pankratios.[2245] The name of the Langobard Droctulft appears in his Latin epitaph as Drocton.[2246] At times names were translated: Ammianus Marcellinus mentions an Iberian prince by the strange name of Ultra;[2247] the prince’s name was Piran; so Ammianus made it into TiEQav and then translated it into Latin.[2248]

The third reason for treating transcribed Hunnish names with utmost caution lies in the circumstances under which they have come down to us. Proper names are particularly liable to corruption in the manuscript tradition. The Procopius manuscripts have Ovq(}i(}evto<; for Urbs Vetus and OvQ^ioalta for Urbs Salia.[2249] It seems unlikely that Procopius is responsible for such forms.[2250] Most of the Priscus fragments are in the collection of excerpts made by Constantinus Porphyrogenitus in the tenth century.

All existing codices, none older than 1500, are copied from the one burned in the fire which destroyed the greater part of the library in the Escorial in 1671. Six Hun names in Priscus are hapax legomena: Adamis, Basich, Eskam, Mamas, Kursich, and Oebarsius. The last one appears in all manuscripts as wTjftaQaiov.[2251] In a Priscus fragment dealing with the siege of Naissus by the Huns, preserved in a single manuscript of the tenth century, the city is said to be situated etil 4dvovfta.[2252] Naissus was not an obscure village but an important town, the junction of several roads. Priscus could not have called the river “Danube.” davovfia is evidently a scribal error. But what was the name of the river? As a rule, if a name occurs in a single passage in the writings of a single author in a single manuscript, it has to be taken as it is. But identical forms in all codices are not necessarily the correct ones. If Persian were as unknown as Hunnish, [s]AQ-caftlbri<; in Theophanes Simocatta III, 18, 9, could never have been recognized as a clerical mistake for * ’Aoyafiid-rjQ = Arghabad.[2253]

Different transcriptions of the same name are of help, though not always. The name of the commander of the troops in Thrace in 447 appears in Priscus as ’Ooiuy/crzXo?,[2254] in Theophanes as ’AyaQaxiaXoQ,[2255] and in the Chronicon Paschale as ’Avdoyioxo^.[2256] Which of these forms is the correct one? None, for they are all distorted from Arnigisclus,[2257] Arnegisclus,[2258] and ’Agv^yiaxXoQ,[2259] Germanic ♦Arnegisl.[2260]


Many languages were spoken in Attila’s kingdom. His “Scythian” subjects were “swept together from many nations.”[2261] They spoke, wrote Priscus, “besides their own barbarian tongues, either Hunnish, or Gothic, or, as many have dealings with the Western Romans, Latin; but not one of them easily speaks Greek, except captives from the Thracian or Illyrian frontier regions.”[2262] We must be prepared to meet among the names borne by Huns Germanic, Latin, and (as a result of the long and close contact with the Alans) also Iranian names. Attempts to force all Hunnic names into one linguistic group are a priori doomed to failure.

“Let no one,” warned Jordanes, “who is ignorant cavil at the fact that the tribes of men use many names, the Sarmatians from the Germans and the Goths frequently from the Huns.”[2263] Tutizar was a Goth[2264] and Ragnaris a Hun,[2265] but Tutizar is not a Gothic name and Ragnaris is Germanic.[2266] The Byzantine generals who in 493 fought against the Isaurians were Apsikal, a Goth, and Sigizan and Zolban, commanders of the Hun auxiliaries.[2267] Apsikal is not a Gothic but a Hunnic name; Sigizan might be Germanic.[2268] Mundius, a man of Attilanic descent,[2269] had a son by the name of Mauritius;[2270] his grandson Theudimundus bore a Germanic name.[2271] Pa- tricius, Ardabur, and Herminiricus were not a Roman, an Alan, and a German as the names would indicate, but brothers, the sons of Aspar and his Gothic wife.[2272] There are many such cases in the fifth and sixth centuries. Sometimes a man is known under two names, belonging to two different tongues.[2273] Or he has a name compounded of elements of two languages.[2274] There are instances of what seem to be double names; actually one is the personal name, the other a title.[2275] Among the Hun names, some might well be designations of rank.[2276] It is, I believe, generally agreed that the titles of the steppe peoples do not reflect the nationality of their bearers.[2277] A kan, kagan, or bagatur may be a Mongol, a Turk, a Bulgar; he may be practically anything.

The names of the Danube Bulgars offer an illustration of the pitfalls into which scholars are likely to stumble when they approach the complex problems of the migration period with their eyes fixed on etymologies. In spite of the labor spent on the explanation of Bulgarian names since the thirties of the past century, there is hardly one whose etymology has been definitely established. The name Bulgar itself is an example.[2278] What does it mean ? Are the Bulgars “the Mixed ones” or “the Rebels ?” Pelliot was inclined to the latter interpretation but thought it possible that bulgar meant les trouveurs.[2279] The Turkish etymology was challenged by Detschev; he assumed that Bulgar was the name given to the descendants of the Attilanic Huns by the Gepids and Ostrogoths and took it for Germanic, meaning homo pugnax.[2280] Still another non-Turkish etymology has been suggested by Keramopoulos.[2281] He takes Bulgarii to be burgaroi, Roman mercenaries garrisoned in the burgi along the limes. Without accepting this etymology, I would like to point out that in the second half of the sixth century a group of Huns who had found refuge in the empire were known as fossatisU.[2282] Fossatum is the military camp.

In addition to the objective difficulties, subjective ones bedevil some scholars. Turkologists are likely to find Turks everywhere; Germanic scholars discover Germans in unlikely places. Convinced that all protoBulgarians spoke Turkish, Nemeth offered an attractive Turkish etymology of Asparuch; other Turkologists explained the name in a different, perhaps less convincing way.[2283] Nowit has turned out that Asparuch is an Iranian name.[2284] Validi Togan, a scholar of profound erudition but sometimes biased by pan-Turkism, derived shogun, Sino-Japanese for chiang chiin, “general,” from the Qarluq title sagun.[2285] Pro-Germanic bias led Schbnfeld to maintain, in disregard of all chronology, that the Moors took over Vandalic names.[2286]

In view of the difficulties concerning Ilie study of Hun names—the inexactness inherent in transcriptions, the morphological changes which many names must have undergone, the ever present possibility that the names were Gothicized, the wide margin of error in the manuscript tradition—in view of all these one cannot help marveling at the boldness with which the problem of the Hunnish language has been and still is being attacked.[2287]

Germanized and Germanic Names


The name[2288] seems to offers neither phonetic nor semantic difficulties. Attila is formed from Gothic or Gepidic atta, “father,” by means of the diminutive suffix -ila. It has often been compared to batyushka, the di- minituve of batya, “father,” as the Russian peasants used to call the tsar. In 1962 the Ozbek poet Kamil Nughman Yasin addressed Nikita Khrushchev as “the dear father of the Ozbek people.”[2289]

Attila is not a rare name. Venantius Fortunatus mentions a regulus aulae domesticus by that name.[2290] /Etla, bishop of Dorchester,[2291] was certainly not named after the Hun king.[2292] 2Etla seems to be concealed in some English place names (Attleford, Attlefield, Attleborough, Attlebridge).[2293] Attila occurred as a monk’s name in Switzerland as late as the twelfth century.[2294]

Some scholars, impressed by the similarity of Attila to Atil, the Turkish name of the Volga, equated the two names without caring for their phonetic and semantic relationship.[2295] Rasonyi was slightly troubled by the final -a in Attila, but he thought that he could dispose of it by going back to what he took to be the earliest from. He regarded -a? in Priscus’ Mwifjlaf as the Greek ending and -a in Kezai’s Ethela as the old Magyar diminutive. In this way he arrived at Atil = Atil, Volga or perhaps just “big water.”[2296] However, the thesis that Kezai, who dedicated his Gesta Hungarorum to Ladislaus IV (1272–1290), preserved genuine Magyar traditions about the Huns has long been refuted. Eighty years ago Hodgkin wrote: “The Hungarian traditions no more fully illustrate the history of Attila than the Book of Mormon illustrates the history of the Jews.”[2297] Rasonyi’s explanation of the name in Priscus is unconvincing. As Latin Attila shows, the name ends in -a, not in -Z; compare ’Avattat; = Ansila, OvvDmq = Hunila, TtorZAa; = Totila, 0W.q)l2.a<; = Vulfila, and so forth.

Pritsak[2298] offered an etymology of both the name of the king and that of the river. In his opinion Atil, Adil, and so forth, meant the same as Attila. He argues as follows:

1. In theByzantine sources the name of the Volga appears as ’XrrZAav[(acc)], TiX, ’AcrT-rfo, and ’Arr/A.

2. These forms show that the Altaic name of the Volga is compounded of two words: aq and rtA, tt]X, teA. The second word could have the enlarged form tiX a.

3. There are two rivers called Tai; one flows into Lake Balkhash and the other one is in the region of the Syr-Darya.

4. Common Turkish a/a changed in Chuvash into i/i’ in very early times,

5. Chuvash *as, preserved only in suffixed forms, means “great, big.”

6. In Hunnish, which developed into Bulgar-Chuvash, *as-tll, *as-til-a must have meant grosse Wassermenge, grosser Fluss, grosses Meer.

7. On analogy with Cingis qa’an and dalai-in qa’an, “oceanic = universal ruler,” the Uigur title koi bilga qan, which is said to mean “the qan whose mind is like a lake,” and Dalai lama, “oceanic = universal religious lord,” Attila, *attila < *as-t’ila means “oceanic > all embracing > universal (ruler).”

This is an ingenious but for many reasons unacceptable etymology. To begin with the arguments based on Chuvash words and forms, according to Benzing (the leading authority on Chuvash), Turkish a/a changed to Chuvash t/t not before the eleventh or twelfth century.[2299] Even if there existed a Chuvash word *as, “big, great, large,” how can we know that in the language of the Huns in the fifth century the same word existed with the same meaning? [At this point, one or two manuscript pages are missing.—Ed.]


Attila’s older brother. The Greek sources have BkgdaQ and BldbaQ, the Latin Bleda.[2300] The Arian bishop whom Marcian sent as his ambassador to Geiseric,[2301] and one of Totila’s generals[2302] had the same name. It is generally agreed that Bleda is Germanic, the short form of a name like OHG Bladardus, Blaigildus, Blatgisus.[2303] Bleda of Marcellinus Comes (s.a. 442) appears in Bede’s Chronicle in the strange form Blaedla.[2304] The English scribes “corrected” the name; they knew it as Blaedla from oral tradition where the name was adapted to 2Etla.[2305]


One of Attila’s counselors,[2306] by birth a Hun.[2307] Edekon is Grecized *Edika;[2308] the hypocoristic form applied to a person whose true name began with Ed-, such as Edivulf.[2309]


Killed in the battle at the locus Mauriacus. The Gallic chronicle of 511 calls him cognatus Altilae.[2310] Laudaricus is Germanic *Laudareiks.[2311]


Attila’s prime minister.[2312] Onegesius is evidently not Greek[2313] but the Grecized form of a barbarian name. Hodgkin[2314] boldly Hunnicized it into Onegesh. *Oneges seems to be Hunigis,[2315] as a spatharius of Theoderic the Great was called.[2316] -gis appears in Greek transcriptions as yig and y^,[2317] huni- is rendered by ovvi- and o’vco-.[2318] Hun- in East Germanic

names is most probably the same as hun in OHG, OE, and ON names, namely either ON hiinn, “cub of a bear, young man,” or proto-Germanic hun, “high.”[2319] Hunila, a Gothic bishop of about 400,[2320] was born and named before the Huns crossed the Don.

I think Thompson is right in identifying Onegesius with Hunigasius, Attila’s interpreter and spokesman in the Vita s. Lupi.[2321] Rasonyi, taking -sios for the Greek ending, suggests a Turkish etymology: oneki, “twelve.”[2322] However, among the hundreds of transcriptions of foreign names listed by Moravcsik there is not one ending in -esios. Oneki would have been transcribed *Onekios. Onegesius is spelled like [2323]Ovxjaipo<;, ’OvTpHXQaxTjs, and so forth.


Leader of the Ostrogoths in the last campaign against the East Romans in 552–554.[2324] He was not opocpv/.oq with them but a Hun from the BIxxoqbqP* Ragnaris is a Germanic name.[2325]


The Eastern sources call Attila’s uncle ‘Povyaq/Povvaq, and ‘Pcoi’Aa?,[2326] the Western Ruga,[2327] Roas,[2328] and Rugila.[2329] These forms lead to Ruga> Rua and, with the suffix -ila, to Rugila > Ruila. Compare Rugemirus, Rugolf, and similar names.[2330] The connection with Turkish uruq, favored by Markwart,[2331] is phonetically unsound.

With the possible exception of Laudaricus and Ragnaris, these names were not the true names of the Hun princes and lords. What we have are Hunnic names in Germanic dress, modified to fit the Gothic tongue, or popular Gothic etymologies, or both. Mikkola thought Attila might go back to Turkish atliy, “famous”;[2332] Poucha finds in it Tokharian alar, “hero.”[2333] The first etymology is too farfetched to be taken seriously, the second is nonsense.

Iranian Names


“Massaget,” doryphorus in the Byzantine army about 540.[2334] -manos is Iranian -mani- or -manah-, which is also transcribed manus, manes, and menes.[2335] No satisfactory etymology has been offered for the first element.

A Hun chieftain in the Caucasus about 500.[2336] “Having arms with power,” Old Iranian *ama-bazuka.[2337]


Together with Sinnion, commander of six hundred Massaget auxiliaries, all mounted archers, in Belisarius’ army in 533.[2338] Balas, transcribed BdXai;, OtiaAat;, Bkaarji;, and BkaaoQ, is a common Persian name.[2339]


Leader of the preponderantly Hunnic hordes which in the winter 465/6 devastated Dacia ripensis and mediterranea. When one considers that poets often slightly changed foreign names to fit them in the meter— Valerius Flaccus, Argonautica VI, 96, has Batarna instead of Bastarna; in Dionysius, Periegesis 302, Zag/iatat became Xaparat—it seems quite probable that Hormidac is Hormizdak, a common Middle Persian name in Sasanian times.


“Massaget,” bodyguard of Belisarius.[2340] According to Abaev, Ossetic xorz-aman, “(having) good intentions.”[2341]


“Massaget,” bodyguard of Belisarius.[2342] Abaev takes it to be Ossetic xors-amond, “(having) good luck.”[2343]

Xxoga^ and Dm)VT}<;

The only source for the war between the Sabir and the Caucasian Huns, led by Styrax and Glones, is the Chronography of Malalas, preserved in a single manuscript, the codex Baroccianus,[2344] which bristles with corrupt readings.[2345] Some of them can be emended with the help of quotations in later works. Theophanes, in particular, often has the correct forms, confirmed by the Slavic translation of Malalas and, though to a very moderate degree, by John of Nikiu. In the codex Baroccianus the names of the two Huns are Tvgay^ and rM/i. Theophanes has Xxvga^ and rkd)vr]Q-, the Slavic translation, Sturaks and Eglon; John of Nikiu, Asterd and ’Agldn6s.[2346] These forms show that the original Malalas text had Styrax and Glones.

Glones is the Grecized form of a Persian name. The general ridivyq, commander of the garrison of Amida in 503, was “a Persian man.”[2347] Z’lcov- afys was the mobadhan mobadh who “refuted” the Mazdakites in the great religious discussion which marked the beginning of the end of the heresy.[2348] Although les formes iraniennes des noms de Glonazes and Boazanes [bishop of the Persian Christians] ne se distinguent pas avec certitude,”[2349] there can be no doubt that the name of the highest Zoroastrian priest was Persian.

As Professor W. B. Henning informed me, Glones may be compared with Golon-Mihran, a Persian commander in Armenia mentioned by Sebeos; there is a variant in other Armenian sources—Wlon-Mihran. Henning took Wlon-Golon-rXcbv for a late form of Vrthraghna (Varhran, Bahram, and so forth.)

Styrax is a common Greek name.[2350] Malalas altered the barbarian name of the Hun into one which was familiar to him and sounded better to his ear. Styrax is, I believe, the same as Xvvqclxoq in an inscription from Gorgippia, a transcription of *sturak, which V. Miller connected with Orgor stur-, “big.”[2351]


Leader of the Kutrigur Huns about 550–560.[2352] Justi compared the name with Zafiagyoi; in two inscriptions from Tanais, assuming that -an was the patronymic -ana, — an.[2353] Zabergan is a Persian name. In the inscription of Shapur I, 261, a.d., it occurs as Pahlavi zplk’n, Parthian zbrkn, and Greek Zaflotyav.[2354] Although Zafleoyav, the general who in 586 defended the fortress Chlomaron against the Romans,[2355] might have been the commander of barbarian auxiliaries and, therefore, a barbarian himself, Za — ^sQydvTji;, a minister of Chosroes I,[2356] was certainly a Persian.[2357]


“Massaget” in the Byzantine army about 549.[2358] The etymology has been found by Professor Henning.[2359] The second half is the Persian divinity Tlr.[2360] Zar-tir is a twin brother of Zar-mihr, a name of the same period. Zag-wfe stands to Zarmihr in the same relationship as TTjQiddvrjq to MtOQiddTT]q

Turkish Names

In the Turkish “runic” inscriptions occur many names with the apposition cur (or cor),[2361] for example,

Al ci cur kuc bars;[2362] Qan cur;[2363] Tadiqin cur;[2364] K61 cur of the Tardus;[2365] Unagan cur;[2366]Yigan cur;[2367] Isbara tamyan cur;[2368] Sabra tamyan cur;[2369] ... t cur;[2370] Bag cur.[2371]

It has long been recognized that dur is a title or rank;[2372] its meaning, however, has not been ascertained so far. Though all the men called cur were members of the aristocracy, their status was not the same. The dur who represented the Kirghiz qaghan at Koi Tegin’s obsequies,[2373] and Isbara bilga koi (i) cur of the monument at Ikhe-khushotu[2374] were high dignitaries; Bogii cur, to judge by the simple slab used for his epitaph,[2375] held a modest position. The various dur named in Arabic sources[2376]—all Turks as it seems—were great lords, but whether dur designated a rank in the military or administrative organization, was hereditary or not, higher or lower than bag or tarqan, is anything but clear. The same is true for the chari and chara = dur in the Khotanese documents.[2377] Cor (hjor) in the Tibetan names Drugu cor, ’Bug chur, and Khri-skugs-hjor in the old Shan-shan kingdom and western Kansu[2378] are Turkish cur,[2379] but what it means is not known.

The Chinese sources are of no help either. In the dynastic annals a considerable number of dur among Turkish-speaking groups are named. As in the inscriptions, c/i’o[2380] (=cur) is often added to another title: for instance, in A ch’o,[2381] Mo ch’o,[2382] P’ei-lo ch’o,[2383] or Shih-chien ch’o.[2384] It frequently occurs in the names of qaghans and other persons of high rank,[2385] sometimes preceded and followed by more titles, as in the monstrous Hsieh to teng-li ku ch’o mi-shih ho chii-lu ying yi chien li pi-ch’ieh k’o-han = El tbbar tagri qut cur toym’is alp chii-lu ying yi chien li bilga qayan.[2386] But none of the chroniclers stated exactly what cur meant.[2387]

The closer one studies the titles of the steppe peoples in the Chinese annals, the more perplexing are the constant contradictions. They are only partly due to misunderstandings on the part of the recorders, although the Chinese, bewildered by the complexities of social and political systems so unlike their own, must often have been tempted to twist titles and ranks until they somehow fitted their ideas of a state, be it ever so barbaric. The nomadic societies, especially those nearer to China and therefore more exposed to her influence, were not unchangeable entities. As testified by the numerous Chinese titles in the Turkish inscriptions, the barbarians saw themselves forced to take over a number of institutions from the hated and admired empire. This meant more than the addition of a set of Chinese titles; it meant a marked change in the political structure. The old titles themselves, as far back as they can be traced, were by no means uniform. Some of them seem to be rooted in the shamanistic oligarchy of an early period, becoming unstable as the functions to which they belonged were withering away; others were closely connected with the ascendancy of the qaghanate. If the pictures the Chinese drew of a given nomadic society differ from one another, at times in the same chapter of the annals, the cause has to be sought primarily in the continuous, now slow, now accelerated shift of importance and power from one group to another. Confronted with reports which contradicted one another because they referred to different periods—not necessarily far apart—the chroniclers often saw no way out but to tuck together what they found in their material and leave it to the reader to make sense out of it. One of the titles which must have puzzled the Chinese was far.

About 635, Sha-po-lo tieh-li-shih qayan divided the western Turks in ten tribes. The five Tu-lu tribes, forming the left division, were under the five “great far” the Nu-shih-pi tribes of the right division under the five “great ch’i-chin.”[2388] The titles of the chiefs were as follows:

Tu-lu Nu-shih-pi
Lii cur[2389] (tribe Ch’u-mu-k’un) Ch’iieh ch’i-chin[2390] (A-hsi-chieh)
Ch’iieh[2391] cur (Hu-lu-wu) Ch’iieh ch’i-chin (Ko-shu)
Tun[2392] cur (She-she-t’i) T’un sha-po[2393] (Pa-sai-kan)
Ho-lo-shih cur (Tu-ch’i-shih) Ni-shu ch’i-chin (A-hsi-chieh)
Ch’u-pan cur[2394] (Shu-ni-shih) Ch’u-pan ch’i-chin (Ku-shu)

The “great far” obviously have the same rank as the “great ch’i-chin” But we have lists of high dignitaries of the western Turks in which the ranks are quite differently arranged: yi-chin, ch’ii-li far, yen-hung-ta, hsieh-li-fa, t’u-t’un, ch’i-chin,[2395] The ch’u-lu far is also the second in a list of high dignitaries of the Turks in T’ang Shu,[2396] but he again heads the list of the officials of the Northern Turks.[2397] Both lists end with ch’i-chin.

It seems that Sha-po-lo promoted the ch’i-chin from a lower rank to that of the dur. The whole system was an innovation, and not a stable one. According to it, there should be no dur in the right division. But the two ch’iieh far whom Mi-she, leader of the Tu-lii, killed in 659 were Nu-shih-pi chieftains.[2398] The Kirghiz seemed not to have been divided into a left and right division. Yet they had their kiiliig far’s, as, for example, Kiiliig dur Baina Sarjun, who was buried by the Barluk River in Tuva.[2399]

One gets the impression that dur was a rather general term, whose specific meaning was determined by the preceding adjective: the great dur, the minor dur, the wise dur, the loyal dur and so forth. Still, the far of the western and northern Turks were all men of considerable importance.[2400] This was not so with the Uyghurs in the eighth century.

The Mahrnamag[2401] lists eleven Manichaean auditores whose names end in dur. None of them was a high official. The princes are called tegin. The “rulers” have either Chinese titles[2402] or are addressed as tirak and il iigdsi. Then follow officials with the title iigd. Of the following “lords” of towns only two are far. The other cur are a physician, a scribe, and various lower officials. The last one is the long list is kill Cur.

The names of the Uyghur far are as follows:[2403]

kwrtr twr = kortla, “beautiful,” far.

bgr’k twr = bagrak, “princely,” far.

yddwy twr = yduq, “holy,” far.

lywl’ng xwm’r twr. Liu-lang is evidently Chinese. Benveniste takes xwm’r to be Buddhist Sogdian ywm’r, *humar, “consolation, encouragement.”[2404]

xr’kwl I” twr. Whether this is one name or two is not clear, xr’kwl — qara qul.[2405] L” might be Chinese.

‘wn twr. Perhaps on, “ten.”[2406]

by’mnwrz twr. A Sogdian name.

twnk whmn twr. Another Sogdian name.

sp’r xr’ twr = isbara qara far.

’Ip cwr = alp, “hero,” far.

qwyl twr = koi far.

In the Mahrnamag, far is not the designation of a function. If it was an inherited title, it amounted at best to a honorific adjunct to a name. We know too little about Uyghur society to determine the causes of this devaluation of far. Life at the court of the Manichaean qaghan was not the same as in the steppe. The change, the disintegration of the old order which made far an empty title, was possibly the result of the strong impact of Sogdian civilization. Together with the new religion, new arts and crafts, new techniques, a new division of labor came into the life of the herdsmen. The Mahrnamag mirrors an urban civilization. Those Uyghurs who returned to their more primitive life after the collapse of their kingdom kept far as a title as, for example, Na hsie ch’o t’e-le = Nahid (“Venus”) cur tagin.[2407] The latest datable Uyghur name of the type x- fur is Inal cur; [2408] it occurs in an inscription of the tenth century.[2409]

The meaning of fur, like that of any other title, was bound to change in time. A closer study of the titles of the Turks and non-Turks in the post-T’ang period may reveal more instances of the restricted or modified use of fur. But it is doubtful whether much more can be learned from Chinese sources. They certainly cannot tell us what the archaic meaning of fur was.

Pelliot was inclined to assume that fur was an Avar word; he even thought it might ultimately be of Indo-European origin.[2410] But no such Avar word exists. I know of no word in the vocabulary of the Hsiung- nu, T’o-pa, or any supposedly Altaic people that might be regarded as an older form of, or related to, cur.[2411] We know practically nothing of the Indo-European languages spoken at the borders of China in early times. Yet there are some documents which lead further back and are more revealing than those discussed so far.

The date of the Turkish inscriptions from the Talas Valley and the shores of the Issyk-kul is the sixth and seventh centuries.[2412] They were the epitaphs of warriors who stood culturally much lower than the Turks in the Orkhon region, not to speak of the Uyghurs. The letters do not have the more or less standardized forms they have on the Orkhon, and the lines are so irregularly arranged that it is often difficult to read them. We may hope to learn from the Talas inscriptions, if not the original, at least the more primitive meaning of dur.

There is, first, an inscription found by Kallaur in the district Aulie Ata. It has been translated three times,[2413] and although a few words are still obscure, the content is clear: A man named dur says farewell to his thirty oylan, his loyal men, and to the pleasures and blessings of the world; he leaves behind his widow and oylan cur.

There is, second, a much longer inscription from the same region with a similar content, known since the 1890’s, but translated only in 1926 by Nemeth. He had to use a squeeze published by Heikel, the same text which Malov translated some years later.[2414] In the fall of 1961 the inscribed stone was rediscovered in situ, photographed and edited by Ch. Dzhuma- gulov.[2415] It turned out thatHeikel’s squeeze was imperfect; both Nemeth’s and Malov’s translations are therefore obsolete. Dzhumagulov’s new translation probably is not final either; the sequence of the lines is still not quite certain and some letters are unreadable. Nevertheless, further study will not change what matters to us: A man bearing the “heroic” name

Qara Cur leaves behind his loyal (or close) friends, the thirty oylan, and his son Qara Cur.

Thirty oylan occur in a third, recently found inscription. Again a man is separated from them. His name is Agus, he is su Cur.[2416] The phrase otuz oylan occurs once more in a fourth, newly found, very mutilated inscription.[2417]

In the Yenisei inscriptions oylan means “boy, son, warrior”; in those on the Orkhon, “son of someone, hidalgo, prince.”[2418] Malov thinks the thirty oylan were the sons of the deceased and their comrades,[2419] which obviously cannot be true for all four inscriptions. But why then the recurrent thirty? When one considers that the armies of nearly all Turkic peoples were divided into units of tens and multiples of tens, it seems much more likely that the thirty oylan were a military unit. It could be a coincidence that a document from the Tun-huang, written in runes, mentions thirty “men of rank and distinction” under the command of a higher officer.[2420] But the men in another inscription who, led by a nobleman, rode nine times around the tomb of their lord, likewise numbered thirty.[2421]

In the inscriptions the thirty oylan are under a dur, whose son is also a dur. With the western Turks under Isbara qaghan the title and rank of “eminent dur” were handed down from father to son. The same must have been true for the more primitive tribes in the Semirech’e.

The Talas inscriptions permit, I believe, only one interpretation of dur: It must mean “commander, leader, captain.” Compared to the great TarduS koi dur, the Cur and Qara dur of our inscriptions were minor figures. They had thirty men under their command; the Tardus officer must have led thousands. But both he and they were “commanders, captains.” Our interpretation is also borne out by the rank of Agus in the third Talas inscription. He was su Cur, “Cur of the troops.” This corresponds to sii bad, “captain of the troops,” in the Tonukuk inscription and in the Vienna manuscript of the Qutadyu Bilig.[2422]

KovaQT^tT^ovo (for short T^ovq), one of the eight tribes, y eweal, or units of the military administration, 0^/zara, of the Pechenegs in the tenth century,[2423] were *kiiarfi fur, “the fur with the pigeon-blue horse-tail flag.”[2424] The Pecheneg fur had nothing to do with the fire of the hearth or the drum, they were neither shamans nor judges, but horsemen and leaders of horsemen. In the language of the Pechenegs fur must have meant “commander, leader.” Among the Kirghiz the word has preserved its military connotations to the present day. It is true, it does not amount to much but it shares this fate with many feudal-military terms. As John Smith, Esq. < scularius no longer bears a shield, so the Kirghiz foro no longer rides into battle at the head of his oylans. In everyday language foro means “boy, lad” in the household of a nobleman.[2425] In the epic, however, foro is still “the warrior, companion in arms, one of the troop [druzhennik].”[2426] Now we can turn to the Huns.

In his account of the war in Lazica in 556,[2427] Agathias[2428] mentions among the Byzantine officers of barbarian origin a Hun by the name of ’Efydy- yeiQOQ’, he was lochagos, commander of a lochos, a regiment. Agathias also mentions the name and the nationality of Elmingeiros’ superior: He was the taxiarchos Dabragezas of the people of the Antes. In order to overcome the difficulties of transmitting orders, a formidable task in mercenary armies of as many different nationalities as the armies of Justinian and his successors, barbarians of the same regions were kept together in the same units. Dabragezas[2429] must have come from those Antes who, according to Procopius, together with Huns and Sclaveni, “lived across the Danube or not far from it.”[2430] Elmingeiros was probably from the same region. The battle in which he distinguished himself took place in the spring of 556.

In the summer of the same year Justin, commander of the army in Phasis, sent one of the taxiarchoi, a Hun by the name of Efyiivtoi$Q, with two thousand horsemen to occupy the fortress Rhodopolis.[2431] In the index of his edition of Agathias, Niebuhr listed Elminzur with the note, fortasse idem cum praecedente, i.e., Elmingeiro.”[2432] Stein identified Elmingeir and Elminzur.[2433] It would be a strange coincidence indeed if in the same army and in the same months, there had been two Hun officers bearing names as similar as Elmingeir and Elminzur.

It is not necessary to know the exact foreign sounds represented by the Greek letters,[2434] nor what the names mean, to recognize that the first is compounded of elmin and geir, the second of elmin and zur. If Elmingeir and Elminzur were actually two names of the same man, the change from -geir to -zur could correspond to his promotion from lochagos to taxiarchos, or, to use the Latin terms, from tribunus to dux[2435] This would support our assumption that cur means “captain, leader.”

There are three more Hunnish names ending in -zur:

1. After the collapse of Attila’s kingdom, his kinsmen Emnetzur and Ultzindur occupied Oescus, Utum, and Almus on the right bank of the Danube.[2436] On the analogy with Elminzur, Emnetzur must be Emne-tzur.

2. Another name of this type is Ultzinzures, OvXtIv^ovqoi.[2437] Together with other Hunnic tribes they followed Dengizich in the second war with the Goths.

3. Prisons’ ’ApiXtovgoi, Urtpagoi, Tovvcjovqeq, and Botanoi appear in Jordanes as Alpidzuri, Alcildzuri, Itimari, Tuncarsi, and Boisci.[2438] The explanation of the difference between Priscus and Jordanes was found by Krasheninnikov:[2439] The archetype of the Jordanes manuscripts had alpidzuros, with the emendation alcildzuros written over it, which leads to alpildzuros. Only this form is compatible with the name in Priscus which, therefore, must be emended to read AAIJIAZOYPOI>AMIA- ZOYPOI.

In the Chinese annals, the titles of tribal leaders are sometimes used for the tribes themselves. In Han times the Chinese spoke of the Sai wang, the “Saka kings,” under the T’ang of the Hu-lu-wu chiieh, She-she-t’i tun, and Shu-ni-shi ch’u-pan.[2440] This was not a misunderstanding on the part of the Chinese, as some scholars thought.[2441] To the Tibetans the kingdom of the second dynasty of the northern Turks was known as Bug-dor = Mo ch’o.[2442] Did they make the same mistake as the Chinese? Should we assume that Constantine Porphyrogenitus was also misinformed when he spoke of the ★kiiarfifur? And before him Priscus about the alpildzuri, or, as we now may say, the alpilfur‘l This is most unlikely. Even today Kirghiz tribes, subtribes, and clans exist which call themselves foro and x-foro: Qara-coro (tribe), Coro, Zol-coro (subtribes), Boro-doro, Ono-coro (clans). The Kazakh have the clans Zhan-cura, Bai-cura, and Qara-cura.[2443]

In an epitaph from Uibat in Tuva the deceased glories that he exerted himself for the people il cur.[2444] Whatever the origin of cur may be, in the inscription from Uibat il fur is as Turkish as il qan and il basi. Hunnish *Alpilfur cannot be anything else but alp-il-cur, “hero-people-dur.”[2445]

The thesis that the Huns spoke a Turkish language has a long history behind it. Its earlier phase is no longer of interest. The later is still with us. Taking the identity of the Huns and the Hsiung-nu for granted, some scholars have no doubt and need no proof that the Huns spoke the same language as the “eastern” Huns, which they take to be Turkish. By the same reasoning the Norman conquerors of England should have spoken Old Norse.

That the Huns included Turkish-speaking tribes can be regarded as established only if a number of personal and tribal names of the Huns are undoubtedly as Turkish as orfevre is French, goldsmith English, and Goldschmied German. One such name is *alp-il-fur.

The formal analysis of Turkish-sounding Hunnic names requires utmost caution. If English were as unknown as the language of the Huns, one could conjecture that fe- in female is a prefix and -diet in maledict a suffix to the root male.

-gir, like fur, occurs in both a Hunnic personal name (Elmingir) and the name of a tribe of the Pontic Huns, named twice in Jordanes, Getica 37. In the first passage, page 63n, all codices, except the inferior ones of the secundus ordo, have altziagiri or altziagri. In the next line, page 6312, the forms are: primus ordo altziagiri (H), ultziagiri, uultziagiri, autziagi- ri; secundos ordo alugiagiri, aulziagri; tertius ordo ultziagri, altziagri (Y).

Mommsen put Altziagiri in both passages in his text. Closs in his edition of the Getica, page 29, preferred Ultziagiri. He was right, in my opinion. In the second passage the name began with u. Three codices still have it; au obviously was u with superscribed a; the forms in H and Y were adapted to altziagiri in the first passage. We have, thus, altziagiri and ultziagiri. Although Altziagiri has no parallel in Hun tribal names, Ultziagiri can be compared with Ultzinzures, OvXtivCovqoi. When we think of the personal name Uldin, and in particular of Elming(e)ir and Elminzur, the conclusion that * OYATIAFIP is but a slightly blundered *0YATimP, *Ultingir, seems inescapable. Gir, like fur, must be a rank or title. It seems to occur in xvQiy’qQ, a Bulgarian genos,[2446] and Yazghyr and Uragir, two Oghuz tribes named by Kashghari.[2447]

Five Hun names end in : ’Aynx, Baal%, Beqi^oq, AeyyiCix> and Kovq<jI%. Standard pronunciation treated % as aspirant in Byzantine Greek until the ninth century.[2448] In the Greek transcription of Germanic names x corresponds to c in Latin forms. The same is true for Hunnic names. AEyyt&x and Atv£l% appear in Latin sources as Denzic and Dintzic. Priscus wrote ’Hovax, Jordanes Hernac. There is no evidence that in fifth-century Greek transcription of foreign names % can reflect g or y.[2449] Therefore, etymologies based on the equation = ig, ty or a% = ag,ay are inadmissible.

The name of an Utigur prince about 550–560 occurs in two forms. Agathias and Menander call him ZavdiXxoq; in Procopius his name is EavdlX.[2450] Sandilchos is Sandilk, Sandil-k.

KovqoIx is[2451] the name of a Hun leader in 395. It could be Kurs-ik or Kur-sik. Kovqq, the name of a barbarian officer in the Byzantine army about 578,[2452] seems to indicate that Kursik is Kurs-ik.

Tov 2.6 lx was the qaghan of the western Turks about 580,[2453] Tuldila a Hun leader in Majorian’s army in 45S.[2454] -ila is evidently the same as -ila in Attila and Rugila, namely the Germanic diminutive suffix. It corresponds to the Turkish diminutive suffix +°q, +°k.[2455] Tuldiq would be in Turkish what Tuldila is in Germanic: “little Tuld.” This tuld can be compared with Ultinzur, Uldin, and l/Zdach, names which seem to be compounded of uld or ult and in/ach = *in/aq.

To maintain that all Hun and Turkish names ending in are diminutives would probably be wrong, but some of them apparently are. Take, for instance, Baot/.[2456] Basich and Kursich are named together. If Kursich is Kurs-ich, *Kurs-iq, then Basich is probably Bas-ich, *Bas-iq, which can hardly be anything else but basiq, “little captain.”

It is almost generally agreed that Aeyyi^lx contains Turkish dayiz. Dengizich cannot be Dengir-siq[2457] because if it were, Priscus would have written AEyyiQalx,[2458] nor can it be Dengis-siy (see above). AEyyt&x is a perfectly normal transcription of *dayiz-iq, “little lake.”

Another formant in Hun names is +1. ’AynxdA, the name of a barbarian exarch,[2459] stands in the same relation to ’Xyu#[2460] as to XavdiL. It evidently is Apsik-al.

The number of Hun names which are certainly or most probably Turkish is small. But in view of the wild speculations and irresponsible etymologies still being expounded, to lay a narrow but firm basis for studying all the names seems preferable to dreamily wandering through dictionaries. Some of the names in the following list have been etymologized before; instead of repeating the arguments brought forward, in particular the many parallels, I refer to Moravcsik, BT 2, where the literature is carefully listed.


Leader of Hun auxiliaries in the Byzantine army about 530.[2461] Alt’i, “six.” In his study of names formed by numerals, Rasonyi (1961, 55–58) listed the Kazakh patronymic Altyev and a large number of personal and clan names having alti as the first element: Altybai, Altyortak, Altyate, and so forth. Compare also Alty bars (Sauvaget 38).


A Hun of noble birth, about 433.[2462] The name could be compared with Iranian ’Xpraxd/za?.[2463] In an Iranian dialect spoken in South Russia the change from -rt- to -t- can be followed in the inscriptions: ’/iTa/cuo?[2464] cannot be separated from ’Xpraxat^?.[2465] Some names beginning with ata are Iranian, for example, ’^ra/zdCe? and ’/IrTa/zdCa;[2466] (*maza, “greatness”)[2467] or ’Xraxoda?[2468] There exist dozens of Iranian names ending in kam, “wish,” from Maaxaprjt;[2469] to Xudkam and Sadkam.[2470] However, Eskam, another name ending in kam, has no similarity to any Iranian name and a most plausible Turkish etymology. Therefore, I accept Vambery’s etymology: ata, “father,” and qam, “shaman.”[2471] Similar Turkish names, for example, Atabag, are[2472] fairly common.[2473]


Hun leader about 395. Basich is probably BaSiq.

Lord of many villages,[2474] Berik, “strong.”[2475] The king under whom the Goths are said to have left Scandinavia had a similar name: Berig, Berg, Berigh, Berich, Berice, Berige; see Getica 2594. Although the Goths took over Hunnic names, they certainly did not rename one of their half- mythical rulers. Berig is probably *Bairika, the hypocoristic form of a name beginning with Bere-, like Beremod.[2476]

A son of Attila. Dagiziq, “little lake.”[2477] Dengizich, as Priscus heard the name pronounced at Attila’s court,[2478]is the only authentic form. Denzic,[2479] Dintzic,[2480] apparently renders the Germanic pronunciation Denilsik, with the frequent dropping of g. AivCIqi%o<; is assimilated to names like rev^eQi^oQ.[2481]

The fact that layiz, ddyiz is not attested before the eleventh century is of little importance.[2482] It occurs in all Turkish languages; besides, there is no language known from which the Turks could have borrowed the word. Mongol Tangiz is a Turkish loanword.


Attila’s oldest son.[2483] The scribes who made the excerpts from Priscus left the name out. It should be in EL 13036 and 18328. Jordanes’ Ellac presupposes in Priscus; compare “Hqvclx = Hernac. Ellac seems

to be alik (ilik), “ruler, king.”[2484] To be sure, in Priscus’ transcriptions of Germanic personal and Latin place names alpha always renders a, never i.[2485] But a in the second syllable occurs also in Armenian, alphilaq>alp ilig.[2486] Apparently Ellac was not the name but the title of the prince who was governor of the Acatziri. Latin and Greek authors often mistook foreign titles for names.[2487]


*Elmingir. Tunguz elmin, “young horse,” also the name of a Manchu tribe,[2488] is probably a coincidental homophone; it would be the only Tungus word in the language of the Huns. El seems to be el, al, il,[2489] “realm”; -min- can be compared with -min in Bumin, Chinese T’u-men and Ch’i-men.[2490]


* Elminfur, see p. 401.


*Emnecur, see p. 402.


Priscus mentions Attila’s wife, Ellac’s mother, in two passages. In the first, EL 13922, all codices have xqexo.-, in the second, EL 1467, M and P have rjoExa B and E rjQtxav, C has rjOExav. The copyists repeatedly dropped v at the end of personal names, but they never added it where it did not belong.[2491] The name ended in -av. To choose between tcQEtcav and TjQenav would be impossible were it not for the Germanic names of Attila’s wife: Herche, Helche, Hrekja, and Erka.[2492] They prove that Priscus wrote rjQsnav. Bang’s etymology is convincing: rjQsnav is *ari(y)-qan, “the pure princess.”[2493] Aruvkhan (aruv, “pure”) is a Qaraqalpak girl’s name.[2494]-[2495]


Eskam’s daughter was one of Attila’s many wives.[2496] Eskam is most probably qam, as, “friend, companion”, and qam, “shaman.”[2497] The non-Tokharain name Yarkam in a Tokharian document[2498] might be a hybrid name with the same meaning (Persian yar, “friend”).


A Sabir, about 555.[2499] Probably *Ilig-dr.[2500]


A Sabir, about 555.[2501] When one thinks of the many Turkish names with qul, “majesty,” it seams very likely that the name was qut-il-ti or qut-elci.


The name of Attila’s father occurs as Movvdlov%oq in Priscus, Mundzuco[2502] in Jordanes, and Movvdiov[een] in Theophanes.[2503] The last one is so corrupt that it can be disregarded.[2504] Cassiodorus undoubtedly wrote *Mundiucus, which Jordanes changed to Mundzucus as he changed Scandia to Scandza[2505] and Burgundiones to Burgunzones.[2506] In vulgar Latin d before i and e, followed by a vowel, became t/z.[2507] Jordanes pronounced Mundiucus as Mundzucus, and consequently wrote Mundzucus. But this does not necessarily prove that the Hunnic name was *Mundiuk. If Priscus should have heard a Pannonian Roman or a Latin-speaking Goth say “Mundzuk,” he still could have written Movvdlov%oq on the assumption that his informant mispronounced the name in the same way he said dzaconus for diacon us.[2508]

Nemeth and Rasonyi[2509] take Movvdiovxot;, Mundzucus, to be the transcription of Turkish munjuq, buniuq, Perle, Glasperle, Kiigelchen oder Perlen, die man am Haise des Pferdes befestigt (Radlov). “Pearl” would indeed be an appropriate name for a prince.[2510] I prefer Vambery’s etymology which took the name to mean Fahne, eigentlich Fahnenknauf, Koralle, die apfelartige Rundung, in welcher der Rossschweif, die primitive Fahne des Tiirkenvolkes befestigt ivurde, und nach welcher das game militarische Abzeichen spater den Namen erhielt.[2511]

In his review of Moravcsik’s Byzantinoturcica, Ligeti doubted the correctness of Nemeth’s etymology.[2512] At my request to state his reasons, he was so kind to write to me: L’expose des raisons de ma reserve vis-a-vis de cette etymologic depasserait les cadres de cette lettre. Je me contenterai de vous indiquer qu’il m’est impossible de concilier cette etymologie avec ce que nous savons de I’histoire des langues turques. Ainsi, le j est caracteristique des langues oghouz, en face du c offert par les autres langues turques. En meme temps I’initiale m caracterise les langues offrant un c, en face de I’ini- tiale b qu’on attend dans les langues oghouz.[2513]

To these objections of the eminent Hungarian scholar one could perhaps answer that to a Greek, in whose language J and c did not occur, the two must have sounded very much alike. More important is the known fact that b interchanges with m within a number of Turkish languages: ban in Osmanli in the eastern and man in the western Crimea,[2514] mindi and bindi in Nogai; borti in the southern and morii in the northern group of Altai Turkish.[2515] One cannot even say that the Oghuz languages have the initial b, for although Osmanli, its Rumelian dialects, and Azerbaijan Turkish have it, the East Anatolian dialects have m.[2516] Except the Auslaut, in the Osmanli dialect of Kars our word has the allegedly impossible form mungux.[2517]

“Flag” as title or rank of the flagbearer occurs in many languages. Ensign, for instance, is both the insignium and the one who bears it: “hee is call’s aunchient Pistoll,” Henry V (aunchient, corrupt for ensign). It is the same in the East. Tuy,[2518] “standard with a horse or yak tail,” occurs by itself or with a suffix in early Turkish and Uyghur names: Tuy A§uq, Tuyluy, “he who was the tuy” Tuyic, “Tuy bearer.”[2519] Munjiiq probably means the same. Qizil MonSuq, the name of a Mongol commander in Afghanistan about 1223[2520] means “Red Flag” rather than “Red Pearl.”

In the eighth century the leaders of the ten arrows (tribes) of the Tiirgas bore the standards.[2521] The cauda equi was the signum militare of the protoBulgars.[2522] It may have been that of the Huns, too.

The Germanic etymology of Mundzucus[2523] is to be rejected. It is not only phonetically unsound. About 370, when Mundzucus was born, no Hun could have been given a Germanic name.[2524]

2av<5tA, Xavdifyos

Ruler of the Utigur, about 555.[2525] Sandil cannot be separated from the Mamluk name Sandal, “boat.”[2526]


Commander of Hun auxiliaries in the Byzantine army, 491 a.d.[2527] Zol- bon is “the star of the shepherd,” the planet Venus, colban, colbon, solbon, and so forth.[2528] Colpan is a Mamluk name.[2529]

Names of Undetermined Origin

The following names, taken by themselves, might be Germanic, Iranian, Turkish or even Latin, or they defy any attempt to connect them with any known language group.


Steward in Queen Erekan’s household. See p. 380.


“Massaget,” cavalry commander in Belisarius’ army, first in the Persian, then in the African campaign.[2530] Without stating his reasons, Justi listed the name as Iranian, but left it out in the enumeration of names ending in -an or -gan.[2531] Aiydv might be Turkish a’i-%an, “prince moon,” as one of the six sons of Oyuz-/an was called.[2532] Compare Ai-bak,[2533] Ai- tekin,[2534] Ai-tas, and Ai-kun.[2535] Incidentally, the Manichaean terms ai tayg and kun ai tiiygri in the Chuastanift[2536] and other Manichaean writings have nothing to do with these Turkish names. Mas’udi’s Aiyan in Gilgit were probably Tibetans; see Markwart 1938, 101, 110.


Magister militum per Illyricum in 538.[2537] Malalas callshim “the Hun.” Not even the correct form of the name can be established,[2538] so speculations about its etymology are futile.[2539]


Ruler of the Utigur about 57G.[2540] Anagai has been equated with A- na-kuai,[2541] the name of the Juan-juan ruler whom the Turks defeated in 552.[2542] Could Anagai be the Turkish name of a bird? According to E. Frankie (1948, 54), “the suffix -qaj, -kaj, -yaj, -gaj, embraces the function of forming designations for bird and the like.” She adduces Osmanli darayai, “black bird,” duryaj, “lark,” and similar names of birds. Duryaj, Turyaj, and Tory a] are both Turkish and Mongol names.[2543] One is also reminded of Mongol names like Piano Carpini’s Eldegai, or Taqau, Ta/ai.[2544]


Hun doryphorus who distinguished himself in the defense of Edessa in 544.[2545]


The Massagetae Simmas and Askan were commanders of a corps of six hundred horsemen in Belisarius’ army in the Persian war about 530.[2546] Justi regarded Askan as an Iranian name.[2547] It might be Turkish *as-qan, “the qan of the As (Az),” although the leader of such a small troop would hardly have been called qan. Besides, it is anything but clear who the As or Az were.[2548]


Rex Hunnorum about 370.[2549] Nomen nemo nisi imperitus pro germanico vendet, said Miillenhoff more than eighty years ago.[2550] The name of the king who is said to have married a Gothic princess[2551] was apparently assimilated to Gothic Valamer. It was Balimber.


Leader of the Sabir, husband of Boarex, about 520.[2552] *Balaq is possibly malaq, “calf.”[2553]


Queen of the Sabir. The bewildering variety of the readings[2554] makes any attempt to etymologize the name a hopeless task. Sinor sees in Germanic reiks,[2555] which for historical and geographical reasons is unacceptable.


Massaget, one of Belisarius’ doryphori in the Gothic war about 536.[2556] I do not know why Justi (1895, 72) listed the name as Iranian; perhaps he thought of Beuca, mentioned in Getica 277 as king of the Sarmatians in southern Pannonia about 470. Bochas could be *Bochan, Bdi^avog.^ Either by itself or with some addition, buqa (buyd), “bull,” occurs as a name since very early times among nearly all Turks.[2557]


Attila’s favorite son.[2558] Ernak is supposed to be Turkish er, dr, ir, “man,” with the suffix -nak, -nik. Professor 0. Pritsak informs me that -nak, -nik as diminutive suffix occurs only in the Altai dialects and in Tuva. He regards -nik as a combination of -n and -k, suffixes which are sometimes used to express not a diminution but an augmentation: ar-an means “he-man, hero.” In his opinion Ernak could be dr-an-ak >ar-nak, “great hero.” Ernak has often been identified with npHHKT> in the Bulgarian Princes’ List.

On the other hand, it is noteworthy that the Armenian Arnak lived at the same time as Ernak (see Justi 1895, 27). Compare also ’AQvaxTji; in an inscription of the second century from Tanais (Vasmer 1923, 33, Zgusta 1955, § 543).


A Hun of high rank, first in Rugila’s, then in Attila’s service.[2559] Harmatta (1951, 145) suggested a Germanic etymology; he thought the name might be aisila >esla and connected it with *ais, “to be respectful, to honor.” But the name might be Turkish: as, es, “comrade,” + -Za.[2560]


Hunnic ruler near the Maeotis.[2561] Bowd in Malalas is almost certainly misspelled. The Turkish etymologies listed by Moravcsik are not particularly convincing.


Doryphorus of Valerian in 538.[2562] Although the best codices have yovftovXyovbov, Comparetti and Moravcsik prefer the reading ftovXyov- dov. There can be little doubt that the longer form is the correct one. To some scribes the accumulation of the barbaric syllables, with their u-u-u-u, in addition preceded by another word ending in u, Ba2.EQiavov proved too much. They decapitated the monster. Gubul occurs as a Jazygian name in a Hungarian document of the fourteenth century.[2563]


Doryphorus in the Byzantine army about 545.[2564] Chalazar brings Tu- tizar to mind.


“The first of the kings of the Huns,” about 412.[2565] Olympiodorus, the only author to mention the name, took great liberties with foreign names. His BeAAeqISoi;,[2566] possibly taken from a Latin source, seems to be a capricious rendering of *Valari/>. Instead of ’^Act^i/o?,[2567] Olympiodorus wrote ’[2568] as if to indicate that the man was aAAo/ev???. -on in Charaton may be the Greek ending. If we had only Mooxaywv, MovQxayayv, and MovxQayayv,[2569] it would be impossible to decide whether -on belongs to the name of the Bulgarian ruler. As the inscriptions with OfiovQxay[2570] show, it does not. -on might also stand for -a. Note that Olympiodorus, like all Greek authors, wrote XceM%(ov for Stilika.[2571] As so often in the endings of foreign names, -on could be -o. Finally, -ton may stand for -tom. Nearly all Greek writers had a marked aversion to -m at the end of a word. Propocius wrote ttivxov (I, 22, 4), novxrjv (De aedif. VI, 6, 16), TtdxEv (De aedif. VI, 3, 11), and qetixov (III, 1, 6). In other words, the name transcribed XaodxMV may have ended in -tom, -ton, -to, -ta, and -t.

Vambery (1882, 45) took Charaton for Turkish qara ton, “black mantle.” This is phonetically sound. But can we be sure that ton was a Turkish word as early as the fifth century? Uigur ton is borrowed from Khotanese or a related dialect: thauna, later thaum, thau, “piece of cloth, silk.”[2572]

If -ton in Charaton were the Iranian word, chara- might be the same as in the Parthian name Xaqdanri^ “having a dark (hara, xara) horse (aspa)”.[2573] Charaton, furthermore, is reminiscent of Sardonius, *Sardon, the name of a Scythian, that is, Rhoxolanic leader whom Trajan defeated,[2574] and the Ossetian Nart name Syrdon.[2575]

If Charaton should actually mean “black mantle,” peAay/Acuvo?, it could be the name of the clan or tribe to which the man belonged. There is the Kirghiz tribe Bozton, “Gray Coats,” and the Kirghiz clans “White Coats,” “Yellow Caps,” and “High Caps” have analogous names.[2576] However, it must be stressed that both the Turkish and Iranian etymologies presuppose that the name ended in -ton.


Hun general in the East Roman army, about 467.[2577] If Chelchal were Chel-chal, one could think of Chalazar. If -al were the formans -al, one could think of Chelch, KoA/, an Ogur tribe.[2578] *Kolk might be koliil, kolok, “(pack) animal,” Kirghiz kiiliik, “race horse.”[2579] But this threatens to degenerate into the well-known play with assonances.


Leader of the Kutrigur, about 550.[2580]


See page 437.


A “royal Scythian,” who fled to the Romans.[2581] Hammer-Purgstall and Vambery compared the name with Mamai, emir of the Golden Horde.[2582] But Mamas, bishop of Anaea,<