(Fig. 4.1). The lower estuaries of these river valleys, which are often the
areas that have attracted the greatest interest of historians, have under-
gone substantial changes since antiquity. Strabo commented on the fact
that in his day the River Nestos frequently changed its bed andflooded
the surrounding countryside (Geogr.7 fr. 443 (44)).^7 The area of the
Thermaic Gulf has been particularly affected, but the coastal zone
between the Chalkidic peninsula and the Straits has also been altered,
with the shoreline receding progressively south and eastwards through-
out the course of thefirst millennium bc. These natural processes
coincided with various forms of intensive land management, including
the partial draining of lakes and waterlogged terrain, activities that have
been dramatically speeded up during the twentieth century.
Modern urban development has obviously played a key part in
changes to the external appearance of the landscape, most notably in
and around Istanbul, Thessaloniki and the western parts of Chalkidike.
The construction of major motorway schemes is having a profound
impact on our knowledge of past landscape use in the region as a
whole, but is also changing traditional ways of accumulating environ-
mental information using survey techniques, by making these areas
much less amenable to pedestrian scholarly access (the Egnatia Hodos
across lower Macedonia and Aegean Thrace to Istanbul; the A1/E80
between Sofia and Edirne for Istanbul; the new‘Trakya’and‘Maritsa’
motorway routes in east-central Bulgaria; new gas plants in the same
region; the Haliakmon Dam project in western Macedonia; new railway
lines along the Pierian coast of Macedonia, and between Plovdiv and
Svilengrad). Development often brings with it money for the investi-
gation of past landscapes, which can be beneficial for our understanding
of past societies. However, the pace and pattern of development rarely
matches the needs of historical research. This means that the scope of
our knowledge of the past is uneven and remains intractably unbalanced
in favour of lowland, coastal zones, river valleys, and plains, at the
expense of upland and marginal areas.
Any attempt to create a balanced portrait of past economies has to
allow for such factors. Information tends to be highly abbreviated and
fragmented in urban locations, more extensive but selective in extra-
urban areas. Inevitably, regional analyses must work with samples that
may not be fully representative and evidence that undoubtedly fails to
(^7) Perissoratis and Konispoliatis 2003, on sea-level changes; Pavlopoulos et al. 2011 on
uplift in the north-east Aegean and subsidence in the Thermaic Gulf; Chiverrell and
Archibald 2009 for analysis of thefluvial geomorphology of one major river, the middle
Hebros (Maritsa) during thefirst millenniumbc.
Thelongue duréein the north Aegean 133