cross and other motifs. Particularly common on the red-brown bowls were little incised or
stamped crosses or palmettes (see for example Paribeni 1907; fig. 60). Other incised
decoration was also common, lines and panels predominating, but including many
different styles of cross and even sometimes roughly incised inscriptions.
It may be suggested that some of the shapes, mainly hole- mouth bowls and round-bodied
flasks or jars, owe their origin to the `African' side of Aksum, while the ledge-rimmed or
ring-based bowls owe more to the pottery styles of the Roman world. Such a range of
influences affecting Aksumite pottery is to be expected considering Aksum's particular
position, but should not be over-emphasised; the overall style of decoration seems
certainly to have been a local inspiration. Differences in the pottery found at such sites as
Adulis, Matara, and Aksum, typifying three different regions of the country subject to
different influences and developments, are not well understood, but for the time being a
western and eastern `provincial' style seem to be recognised (Anfray 1973i; Michels, in
Kobishchanov 1979: 26).
Some tomb vessels, largely in the red wares, seem to have been created with some very
specialised use in mind. Bowls containing modelled images of two oxen yoked together,
`foot-washer' bowls, with, in the centre, a kind of platform (occasionally two), sometimes
with a runnel for water, conceivably for painting henna on to the feet, stem bowls, bird-
shaped vessels, tripod jars, and strainer jars are among these (Chittick 1974: fig. 19k, pl.
XIIIb, pl. XIIa, pl. XIIc, pl. XIIb and fig. 19c, and pl. Xiib illustrates all these types).
Another unusual pottery type is the human-headed ointment (?) jar. The necks of these
round-bodied jars (Chittick 1974: pls. XIIb-c, XIVb; Leclant 1959: pl. III-IV bis; de
Contenson 1959: pls. XV-XVII), accessible from the top through a narrow opening,
depict women with a hair-style in which the hair turns out sharply at about chin level.
There are several differences in detail, some with zigzags representing the strands of
(plaited?) hair, others with a sort of cap on the top of the head. Earrings are sometimes
shown, and on the whole one gains the general impression of a coiffure some what like
that of the women of Tigray even today. Plaited hair, lying close to the head at the top,
but allowed at the base to frizz out freely, is a style which can still be seen in preparation
by the hairdressers during market days in Aksum nowadays. On one example from the
Tomb of the Brick Arches at Aksum, part of the globular body of the pot survived, and
bore an arm painted in yellow paint on one side. Interestingly, pots of the same general
type, though with different hairstyles, were collected relatively recently in the Azande
country of south-western Sudan by Sir Harold MacMichael (Horniman Museum,
Illustration 59. Profile view of a painted head from a jar from the Tomb of the Brick
Arches at Aksum.
Illustration 60. Only one of the human-headed jars found in the Tomb of the Brick
Arches was still almost intact, and traces of one painted arm could still be seen.
Painted decoration on pottery might consist of crosses in various colours, plant motifs, or
panels filled in in different ways. Red, black, and white were popular colours, and there
was a purplish paint in use in later Aksumite times (Chittick 1974: pl. XIVc). The
ordinary shapes included, very frequently, globular-bodied jars with more or less long
and thin necks, sometimes very close to the typical coffee-pot used even now in Ethiopia
and the Sudan, where their round bases rest on rings of plaited straw or other fibres.
Bowls, either round or ring-based, were also very much used, together with beakers, and
many types of handled jars, cauldrons, storage vessels and the like.
Imported amphorae are also not uncommon, and were employed for various purposes
after their original contents had disappeared. Some were cleverly used to form a sort of
water supply-pipe to a baptismal pool near a basilica at Matara (Anfray 1974: 757-8, and
pl. IV, 2); others served as coffins for the burial of babies at both Adulis and Matara
(Anfray 1974: pl. II, 1; Paribeni 1907: 452, 480); and a third use was as furnaces or
ovens. At Adulis seve ral examples of the latter were found, and also other types of vessel
employed as ovens or for industrial purposes; Paribeni suggested that some examples
from Adulis were employed in liquefying tar and in gold-working. His excavations at
Adulis revealed what was apparently a gold-workshop, with amphorae and ashes in
association with a collection of gold rods, earrings, and two elaborate bejewelled crosses
with chains; in a different place a stone mould for making jewellery was found (Paribeni
1907: 453, 483-6 and figs. 20-21, 461, fig. 7). Many of the amphorae may have come
from Alexa ndria (Paribeni 1907: 455). Some bore incised or painted identification marks
at shoulder level, and they were sometimes sealed with terracotta discs, plastered over,
with an identification mark stamped or painted on top (Paribeni 1907: 456, 520, 522, 524,
figs. 4, 39, 41, 43, 59).
One very specialised imported vessel found at Adulis was a flask stamped with a design
showing the Egyptian St. Menas between two kneeling camels; such vessels are supposed
to have held water from a spring near the saint's tomb in Egypt (Paribeni 1907: 538, fig.
54), and this one may have been brought to Adulis by a pilgrim.
Among imported pottery types were a number of lamps. While at Aksum lamps tended to
be of a very simple open type, probably local, Adulis produced a wider variety including
some closed lamps with moulded decoration which certainly came from the Roman
empire, probably Egypt (Paribeni 1907: 499, fig. 28; 518, fig. 38). Others, perhaps local,
consisted of small double or single spouted jars (Paribeni 1907: 460-1, fig. 5; 522-3, fig.
42; 526, fig. 45).
Most of the pottery shapes were evidently designed for eating, drinking, storage and
cooking, but the more unusual ones perhaps served for special purposes like personal
hygiene, cosmetics, or ceremonial occasions. Some may have been for ritual use, or made
as specifically funerary goods to serve the dead in some way. This might account for the
apparently long use of certain types; they could have been fossilised designs essential to
some funerary purpose, though not in general outlasting the fourth/fifth-century change to
One spherical pot from Adulis was closed completely except for a rectangular opening
around which were impressed four crosses; it evidently served as a money box, since