settlement has been found in a number of separate locations, both on the
acropolis of Nebet Tepe, and in the lower city at its foot. A configuration
of multiple concentrations, which together amounted to a single collect-
ive, or a group of closely related local communities, would explain the
persistent nature of this pattern and the apparently spontaneous emer-
gence of a politically well-defined civic community when the earliest
Hellenistic documents with which it can be associated become archaeo-
logically visible.^59 The evidence that has accumulated makes clear that
the centre of exchange focusing on the hilltop of Nebet Tepe was indeed
an urban agglomeration, not just a loose collocation of residential dis-
tricts. The dispersed nature of the sites identified does not conform to
expectations of urban organization; but there is plentiful evidence of
Early Iron Age andfifth to fourth centurybcactivity in and around the
centre of Plovdiv. We need to rethink our ideas about urban planning in
contexts where there was far less pressure on space, as compared with
central and southern Greece. In the north Aegean area the challenge was
the breadth not the lack of space.
Comparative research on urban planning can help us understand the
implications of this type of spatial evidence. Students of urban planning
have long recognized the operation of‘Zipf’s law’, namely the phenom-
enon that in human societies settlements fall into a natural hierarchy
based on size, with a small number of genuinely large cities at the top of a
progressively expanding logarithmic curve. This pattern of site sizes has
recently been explored in the context of Roman imperial cities.^60
I suggested more than a decade ago that we should expect to see a similar
pattern in pre-Roman settlement distributions in Macedonia, Thessaly
and Thrace.^61 The comparative sizes of settlements are hard to evaluate
in practice whilst so little is known about their overall spatial organiza-
tion, whilst the‘dispersed’nature of some sites makes this task even more
challenging. However, the well-founded principles behind Zipf’s law
offer a reliable way of thinking about sites in general. The majority of
sites identified through currentfield investigation will have been small
rural establishments, whether these were individual farms, hamlets,
mining villages, or agricultural processing units (such as foresters’or
charcoal burners’huts; and hunting lodges). Some will have been larger
agglomerations, particularly those at road and river junctions, where
natural routes crossed; and a few will have been much larger. The
(^59) Philippopolis:Inventory, no. 655; see esp. Koleva et al. 2000, 103–21; Koleva and
Katsarov 2009. 60
See the contributions to Bowman and Wilson et al. 2012, esp. A. Marzano.
(^61) Archibald 2000, 218–19.
Herdsmen with golden leaves—narratives and spaces 69