LATE ROMAN SOCIETY AND ECONOMYwhich suited a different kind of social and economic life. Public buildings
were not repaired, or fell out of use, their stone being used as spolia for pur-
poses including fortification. Houses were subdivided; shops or workshops
encroached on colonnaded streets. In the Balkans, urban sites began to give
way to fortified hill settlements, while on the eastern frontier sites described
by Procopius in his Buildings, such as Resafa/Sergiopolis near the Euphrates,
were characterized by their military provision and fortifications and their
large churches and episcopal buildings rather than by public streets or grand
houses.^58 But it is difficult to generalize: the large urban site of Scythopolis in
Palestine for example continued to flourish into the Umayyad period until it
was hit by a major earthquake in 749, and the large number of ‘dead cities’ of
northern Syria were able to maintain a prosperous lifestyle not based on state
investment or the presence of large landowning magnates. Similar changes
can be detected in different parts of the empire, as continues to be shown
by a huge mass of recent and current work, but at very different rates and
for different local reasons. They cannot be explained in simple or universally
applicable economic terms, but rest on much deeper changes in society, and
especially on the changing role of elites and degree of maintenance of the late
Roman administrative structure.
Long-distance trade and exchangeOver the last few decades the study of the vast amount of pottery evidence,
and in particular the diffusion of amphorae, the containers of the late Ro-
man world, has transformed our understanding of specialized production and
exchange. The pioneering work of John Hayes in the early 1970s provided a
typology of late Roman pottery which enabled archaeologists to date exca-
vated material far more securely and to log stratigraphic evidence; these fi nd-
ings were also extended into the study of late Roman amphorae, particularly
by Italian archaeologists. The evidence from ancient shipwrecks sometimes
provides valuable dating material of this kind.^59 With the growth of Constan-
tinople and the diversion of Egyptian grain to the eastern capital, an eastern
axis, Carthage/Constantinople, became important, and long-distance ex-
change continued into the fi fth century, without serious break after the Van-
dal conquest of North Africa in 439;^60 this eastern axis continued in existence
throughout the period, as the importance of Constantinople reached its peak,
until the Persian and then the Arab invasions cut the connections between
Constantinople and Egypt.
Many questions suggest themselves as a result of these conclusions, of
which the following are only the most obvious. For example, whether these
fi ndings actually represent trading links (the evidence of pottery will not tell
us the why of transmarine exchange, only the how). What if anything can they
tell us about the impact of the post-Roman kingdoms in the west on the Medi-
terranean economy? Finally, to what extent does this evidence tally with that
of urban change to suggest that a signifi cant weakening of the Mediterranean