Demography

So far I have avoided the question of how to date the isolated farmsteads. Although they in general can be considered a characteristic feature of the classical Greek landscape, clear differences have been noticed between different regions as to when they start to occur and when their numbers peak. The peak is usually accompanied by an increase of off-site scatter in the landscape, possibly the result of intensive manuring of the fields. Even those who believe that the isolated farmsteads were not settled permanently agree that the parallel occurrence of dispersed small sites alongside nucleated settlements indicates an intensification ofland use and thus also a population climax. The fact that many of the small, dispersed settlements are located in marginal lands provides further support for such an interpretation.

By studying the occurrences of farmsteads at different times across ancient Greece we can thus establish divergent regional demographical trajectories (Bintliff 1997). Thus, small isolated farmsteads occur already during the archaic period in Attika, Lakonia, Boiotia, parts of Argolido-Korinthia, as well as on Melos and Keos. However, the number of these settlements peaks during the classical to early Hellenistic period in a wide arc of south-eastern Greece, not only throughout all of the regions mentioned above, but also in Arkadia, Euboia and on other Aegean islands (with the exception of Melos, which shows a decrease) and possibly also on the islands of Leukas and Kephallenia. In Messenia the first farmsteads do not appear until after 369. In a similar way the peak appears in the early Hellenistic period in Aitolia and Epeiros, and possibly even later in the Hellenistic or early Roman period in Achaia and on Crete. In most of the rest of Greece there is a marked decrease in the number of rural sites during the late Hellenistic and early Roman period.

Landscape archaeology’s main contribution to ancient demography is the fact that it is the only way in which we can detect changes in population size across time. Landscape archaeology can also be of some help when estimating the size of population at a given point in time, but it cannot give us any information about nativity and mortality rates, the distribution of population among sex and age groups, life tables or the average life expectancy, all basic concepts in historical demography. A close study of the epigraphical and skeletal data from cemeteries can help us to approximate some of these rates, but not all, thus invalidating any attempt at a full demographic analysis (Sbonias 1999a).

Lately it has also become fashionable to use the data collected by intensive surveys to estimate the size of population in a region at a given point in time. First, the total

Settled area (including all types of settlements) at a certain point in time has to be established. Then, the probable population density per hectare is estimated with the help of information from the residential quarters of urban settlements that have been excavated. On the basis of these two factors the number of people living in the region can finally be calculated. However, it must be pointed out that this method is full of pitfalls and should be used with care. When it has been applied (in Southern Argolid, Lakonia, Boiotia and on Keos), it has always been possible to compare the results with some sort of written demographic evidence. This may give some confirmation of the estimated settlement density, but cannot tell us whether the density was the same 100 years later, or, for that matter, in a totally different region.

Analogous with the method described above, it is often argued that it is possible to calculate the population of a city on the basis of the area inside its walls. Consequently it would also be possible to estimate the difference in population size between poleis for which we know the size of the walled area. The Copenhagen Polis Centre has collected all available data of this kind and presented it in a most useful article (Hansen 2004). For a total of233 poleis of archaic or classical date there are sufficient remains of the circuit wall to allow an assessment of the area enclosed by the walls. Twenty-three of these poleis have a walled area exceeding 150 ha. In Table 4.1 they are arranged according to size.

Hansen assumes that all 23 of these poleis had a population larger than 10,000. This may very well be the case, but does Table 4.1 necessarily give the correct internal order of the poleis in terms of population? Was the population of Athens, for instance, really of the same size as that of Sybaris and Taras, and only slightly larger than that of Maroneia? It is also clear that some regions are over-represented in the table. It includes a total of eight poleis from Magna Graecia and three from Arkadia, but none, or, very few, e. g., from the Black Sea region, the Aegean islands and the Ionian colonies in Asia Minor. Important poleis surprisingly excluded from the table include Kos (112 ha), Mytilene (140 ha), Miletos (130 ha), Samos (103 ha), Eretria (81.5 ha) and Thasos (70 ha), just to mention a few. These examples clearly show the deficiencies of the method and the danger of using a standard figure for the relationship between population and space. The fact that certain regions are overrepresented whereas others are totally absent in Table 4.1 could indicate that the population densities varied between the different regions for some economic or social reason. Thus, the method may perhaps be considered reliable only when comparing poleis of similar general character within one and the same region.

Despite these methodological caveats, the method brought forward by landscape archaeology to estimate the size of the population at a given point of time should not be fully rejected. Such data may confirm or refute figures quoted in written sources, and in those cases where no written information exists they constitute the only available ‘guesstimate’. Furthermore, they can also give us some idea of how the population of a certain region was distributed over the landscape. In this case most of the intensive surveys seem to end up with similar results, i. e., that the majority of the population lived in the urban centres and that the rural population generally only constituted a small part of the total population (Hansen 2004 suggests a figure between 10 per cent and 33 per cent). The proportion of the rural population may perhaps be larger in the colonies - one thinks of the 400 or more farmsteads recorded outside the walls of Chersonesos on Crimea (Pedrka 1973) or the 870 farmsteads

Table 4.1 The 23 Greek poleis of archaic and classical date with an area enclosed by walls larger than 150 ha (according to Hansen 2004).

Outside the walls of Metapontion (Carter 1990). Then again, much depends on whether the farmsteads were occupied permanently and on the number of residents assumed to live on each farm.