80 THE ROMAN EMPIRE
labour, or in other other ways, was self- sustaining and resulted in structural
change. The inherited framework of economic life remained largely intact.
The problem of documenting economic growth and gauging its
signifi cance is particularly acute in the area of trade. An ingenious attempt
by Hopkins to demonstrate an expansion of trade (see p. 50) revolves
around four propositions: fi rst, the imposition of money taxes in the
provinces by the Roman government greatly increased the volume of trade
in the empire; secondly, levels of consumption were considerably higher in
the Roman than pre-Roman period, at least in the West, ‘illustrated but not
I think proven’ from artefact fi nds; thirdly, a greater incidence of shipwrecks
shows more seaborne trade than ever before; fourthly, there was a growth
of the supply of money such as to fi nance greater interregional trade. The
arguments cumulatively make an impact, although singly they remain
unsubstantiated, as is disarmingly admitted, or (as is the case with the second
and fourth arguments), if anything, they establish that trade expanded in the
late Republic rather than the early empire.^16
An alternative approach, which we favour, is to investigate the possibility
of changes in the infrastructure of trade, in technology and commercial
institutions, such as to point to increased activity in the trading sector. Here
there was little signifi cant development in the period of the Principate. For
example, in the central matter of ship construction, the late Republic and
late empire are the main periods of innovation.^17 First, Mediterranean
shipbuilders were already in the fi rst century BC constructing large ships in
the range of 250–400 tonnes for the bulk transport of food and other
commodities. Secondly, throughout our period, and indeed until about AD
400, they appear to have built merchant ships outer- shell-fi rst, a laborious
and expensive method, instead of building them up from the internal frame,
the skeleton method. This is particularly signifi cant in view of the fact that
skeleton construction was known and used in the Celtic provinces.
The several changes in the containers by which primary products were
carried, wine in particular, have long intrigued archaeologists. A
developmental model, according to which each change represented an
improvement in the ratio of contents to container and can therefore be seen
as a technical advance with commercial implications, is over- optimistic. The
weight, solidity and capacity of amphorae were matters of more than
academic interest to Romans (cf. Pliny, HN 35.161).^18 But one cannot be
certain that technical and commercial considerations lay behind the
abandonment over three or four decades at the end of the fi rst century BC of
one Graeco-Italian amphora (or clay jar) as a main carrier of wine (the so-
called Dressel 1) in favour of another, lighter amphora, modelled on the
amphorae of Cos (the so- called Dressel 2–4). Again, although the introduction
of ships that transported wine in massive jars ( dolia ) anchored in the ship-
centre, known from the Augustan period, may be seen as a technological
innovation, it was one that was not persisted with after what seems to have
been an experimental phase. Finally, it is too early to say whether the wooden