Excavations at Aksum
By Stuart Munro-Hay
This final chapter is based on a talk given by the author to the Society of Antiquaries of London in October 1987, designed more or less to coincide with the publication of the report on the excavations undertaken for the BIEA by Dr. Neville Chittick at Aksum from 1972-1974. Chittick’s untimely death in 1984 prevented him from writing a fuller account than his Preliminary Report of 1974, but this task was undertaken by the present author and has now been published (Munro-Hay 1989). Since the last major book on Aksumite archaeology appeared before the First World War (Littmann 1913), new studies based on archaeological excavation are long overdue. In addition, the architectural, numismatic, chronological and general cultural information revealed by Chittick’s excavations has radically changed the impression gained of Aksum and its civilization through previously published material, and it is evidently useful to recapitulate some of the main points here. Some of this material has been mentioned in previous chapters but is here described altogether within the context of the two main archaeological campaigns of 1973 and 1974.
The BIEA excavations which Chittick directed were on a large scale, and there was a great deal of information to sift through. The result is that we have not only a much clearer picture of many facets of Aksumite life, but also valuable indications towards a chronology, one of the perennial problems in Aksumite studies. As usual, more information produces more problems; we cannot claim to have more than begun to solve them, but a good deal of progress has been made, and the general schema of Aksumite history presented by this book has greatly benefited from Chittick’s work. The excavations explored a large number of sites in and around present-day Aksum. The archaeology of these is fully described in Munro-Hay 1989. The location of the most important sites was as follows. The easternmost sites excavated were those flanking or near to the superstructure covering the so-called Tombs of Kaleb and Gabra Masqal. Next to the west were a number of stele sites, called Geza `Agmai (GA), `Enda Yesus (EY), and Ghele Emmi (GE). In the eastern central part of the town the site DA revealed the Tomb of the Brick Arches, and many trenches were laid out around the Stele Park (ST) to investigate and to try to date the stelae.

Among these trenches many tombs were discovered, including Shaft Tombs labeled A-C, and the Mausoleum and East Tomb near the largest stele. Certain other large tombs were also cleared; the Nefas Mawcha (NM) south of the great terrace wall above which the three largest stelae stood, and, to the west of the main stelae group, the Brick Vaulted Structure and the Tomb of the False Door (the THA and THC trenches). In the ancient residential center of the city, two sites revealed what were almost certainly small parts of large mansions; they were labeled IW and ES. Finally, apart from a number of relatively unimportant exploratory trenches around the Stele Park (designated HAW, PW, WC, and ML), some trenches were laid out in a stelae field west of the modern town, opposite the Dungur villa excavated by Anfray (1972); these were called GT after the traditional identification of the area with the legendary queen Gudit.
An important, and previously unknown, the feature found at Aksum, among the ST trenches, was a series of buried stone-revetted platforms, the earliest of which seems to have been constructed in the first century AD, according to radiocarbon readings. The latest seems to have been erected or rather expanded, in the fourth century. Behind their facades, the platforms were filled with freshly-quarried stones with almost no admixture of earth, and topped, relatively carefully, with layers of white and red soils. These, and considerable deposits elsewhere, at such sites as GA, GE, and HAW, lay in levels which yielded no coins and thus seem to precede the first issues in c270-90AD. But in some of the GA and ST, early levels were glass fragments, including types such as the mosaic or millefiori glass generally dated to between the first century BC and AD 100. The Periplus and other accounts mention the glass of several types among the items imported into Aksum from the Roman empire, and these finds not only confirm the Periplus’ report but help to date these platforms as the earliest yet known features at Aksum.
There are further indications that there was a considerable period of occupation at the site before the `Classical Aksumite’ period of the third and fourth centuries. Some stelae, found standing upright in pits which had been dug into the earlier platforms, had been completely buried by subsequent deposits. All were of a rough undressed type which preceded the later carefully shaped and sometimes elaborately carved examples, some of which could be dated to the later third and the fourth century AD by accompanying material. The stelae at Aksum have never been properly dated. During his excavations, Chittick was surprised to find that, according to his estimate, `the coins indicate that the deposits on which the stelae were erected accumulated in the Christian Aksumite period’. This certainly seemed a little unlikely, in view of the mounting evidence as the excavations progressed that the stelae were closely connected to tombs and were probably memorials to the deceased Aksumite kings; but, as usual, the key to his dating was the coinage, which has since been radically re-dated (Munro-Hay 1978 et seq.).
The particular coin type which led Chittick to assume that he had found stelae of Christian date was an issue attributed by earlier numismatists (Anzani 1926) to the sixth-century king Kaleb or his immediate successors. But the type can now be re-dated to a considerable time before Kaleb, on the basis of overstriking on coins of king MHDYS, probably a close successor of Ezana. Further, the study of the stratigraphy in the trenches concerned has resulted in a different interpretation of the sequence of events and no longer supports the idea that stelae were a late phenomenon at Aksum. The French archaeologist Henri de Contenson, working at Aksum in the 1950s, found that fragments of the broken summit of the largest stele of all (no. 1) lay below an occupation level containing coins of the late fourth-century king Ouazebas (de Contenson 1959: 29). The stele, according to the coin evidence, fell most probably in the late fourth or early fifth century; that is, after the official conversion of King Ezana in c330AD, but at a period sufficiently close to this event to make it likely that certain burial traditions such as the erection of stelae had not yet lapsed.
Other structures in the Stele Park area are also dated to the later fourth century and confirm that the cemetery was in use after the advent of Christianity, but probably for decades rather than centuries. The stelae seem to have all been associated with tombs, but as yet the direct pairing of certain tombs with certain stelae remains difficult. Almost every trench opened in the central (ST) area of Aksum yielded either a fallen stele, broken fragments, buried upright stelae, or shafts leading to tombs and tunnels, and it is certainly premature to assume that we know the layout of the necropolis. Two of the tombs were of quite unforeseen dimensions and sophisticated architecture. One, possibly associated with the largest stele, was dubbed the `Mausoleum’, and consisted of a 15 × 15 meter complex of rooms off a central passage. Included in its construction were dry stone walling, a brick arch, three shafts, dressed or rough granite roof-blocks, and a magnificent granite doorway in typical Aksumite style. The second, called by the Tigrinya name `Nefas Mawcha’, a name meaning something like `the place where the winds go out’, consisted of two outer corridors roofed with dressed granite slabs, built around a central room which was covered with a single slab measuring some 17m × 7m × 1½m.
the Tomb of the False Door
By a curious chance this tomb, roofed by the second largest stone known to have been employed in Aksumite construction work, was severely damaged when the largest stone of all, the greatly carved stele, crashed down and struck the tomb’s north-west corner. This upset the complicated balance of roofing blocks and the entire tomb subsided. However, enough has remained intact for the excavators to be able to propose a restoration of its original design. Though the stelae may not have been objected to for religious reasons (it has even been suggested that some of them bore crosses at the top, where nail-holes indicate some applied decoration; van Beek 1967), the collapse of the largest one, and possibly of the second largest too, may have been sufficient reason for the Aksumites to turn to a simpler but essentially similar memorial, the house-tomb.
The most accomplished monument of this type at Aksum, the Tomb of the False Door, was a surprising discovery. It is entirely made of dressed granite blocks, in the form of a house-superstructure with a magnificently carved granite door over a tomb chamber and a surrounding corridor, reached by a separate staircase from a paved courtyard. It was dated by Chittick to the pre-Christian period since it was overlain by deposits containing glass attributed to the third century. However, it was later found that a stratum running beneath the stones of its courtyard abutted against an earlier stone wall, part of a building called the Brick Vaulted Structure. This latter, though incompletely excavated, appears to have consisted of a series of burnt-brick vaults, with horseshoe-shaped arches and granite relieving lintels, closely resembling the architecture of yet another tomb, the Tomb of the Brick Arches, so-called from its three horseshoe-shaped arches. This latter tomb contained material of probably mid-fourth century date. It, therefore, seems that the Tomb of the False Door is later than the arched structures, and probably of late fourth or early fifth-century date. Very likely it was the next stage in the development of the necropolis architecture, since the fall of the great stele would probably have discouraged further such attempts, and the house-tomb type is a logical successor.
A very close stylistic link between tomb and stelae is provided by the doorway and lintel of the tomb, carved in exactly the same manner as the doors depicted on the two largest, and latest, stelae. The early material found over the tomb, which was one reason for Chittick assumption that it was of pre-Christian Aksumite date, appears to have been washed down from the higher slopes of the Beta Giyorgis hill which dominates the necropolis; it included a large number of stone scrapers also found in quantity on the top of the hill. Two other house-tombs, the double tomb building locally attributed to the sixth-century kings Kaleb and Gabra Masqal, and another found at the Eritrean site of Matara, are comparable. The picture we have of the town is not all taken from the necropolis. At the same time, the excavations cleared a number of domestic structures, particularly at two sites which were designated IW and ES. These revealed rough stone-built walls, strengthened by two techniques, wooden interlacing or the use of granite corner blocks. In addition, the walls were arranged in a series of recesses so that there were no long stretches of wall, and each wall rose in rebated steps, each lined with slate.
This is, as we have noted (Ch. 5: 4), typical Aksumite `mansion’ architecture. Only a few rooms were cleared in each place, but it was evident from the finds that these dwellings were the houses of prosperous Aksumites in possession of a high standard of living. Objects found include the fragments of many polished breccia bowls, glassware, and elegant pottery, metalwork, coins, and other items. Temples or churches were not found by the BIEA expedition, nor were examples of the more humble dwellings which were probably built with perishable materials such as wood and mud plaster with thatched roofs. The tombs, though only one was completely cleared, yielded rich grave-goods. The cleared one, marked by a rough stone stele, was not found in the main necropolis, but in the Gudit Stele Field west of the town (GT II). It appears to date from the mid-third century and was merely a small chamber cut into the earth, with no built elements at all. It contained particularly fine pottery, two sets of glasses (stem goblets and beakers), and a large number of iron tools such as tweezers, saws, knives, and a sickle. The stele marking the grave, and the pottery, are `Aksumite’ elements; the glassware and tools could as well be from a Roman site as an African one.
Tombs in the main necropolis were evidently much richer, and the excavation of only one room in the 4th century Tomb of the Brick Arches revealed piles of grave-goods (mentioned above, Ch. 12), including glassware, pottery vessels in a multitude of shapes, some painted and decorated, all sorts of metalwork, including glass-inlaid bronze plaques, fittings for what was probably a wooden chest, gold fragments, a silver amulet case, a bronze belt-buckle inlaid with silver and enamel crosses, iron knives with bone or ivory handles, and even leather and wood. In a number of small inner loculi, constructed by dividing the interior of the simply cut tomb by built stone walls, stone coffins could be seen behind a partly broken blocking wall. This tomb was also a surprise from the architectural point of view since it was the first of those excavated which revealed the burnt brick horseshoe arches later found in even more elaborate styles in the Brick Vaulted Building. These must be among the earliest horse-shoe arches known, and were quite unexpected elements in Aksumite architecture, not being repeated in the later rock-cut churches of Tigray and Lasta. Other tombs consisted of carefully cut shafts leading beneath the stelae into vast roughly cut chambers (Shaft Tombs A, B, and C), or into long winding corridors (the Tunnel Complex), which may have belonged to tombs, or perhaps more likely were robber tunnels. Though pottery, cut stone, skulls and so forth could be seen lying in the rooms and corridors, little clearing could be done in the time available. Much more work is necessary in the tombs found by the British Institute expedition, but political events have precluded a return as yet.
Nothing significant was found in the tombs or buildings at Aksum which can be certainly attributed to a later date than the sixth or early seventh century AD. The archaeological record shows that the large residences were occupied or built around by squatters, even, apparently, in the time of the last coin-issuing kings, then gradually covered by material brought down by run-off from the deforested hills. The excavations thus confirm the theory suggested above that by about 630AD the town had been abandoned as a capital, although it continued on a much-reduced scale as a religious center and occasional coronation place until the present.