Gardening, foraging and herding: Neolithic land use and social territories in Southern Italy. (Research).

Robb, John ; VanHove, Doortje

Neolithic land use: Economy or culture?

Most discussions of Neolithic land use and settlement start from one of two premises. Traditionally, the Neolithic has been defined economically by the use of agriculture and stock raising. Neolithic land use is essentially seen as reflecting the needs of farmers for farmland and herding land (cf. Higgs 1975), and a persistent explanation for the spread of agriculture has been the constant need for new farmland under demographic pressure on resources (Ammerman & Cavalli-Sforza 1984). But recent critiques have argued that Neolithic societies cannot be viewed as passive reflections of their economies (Thomas 1991; Whittle 1996). One line of critique is based on the continuing use of non-domestic resources in Neolithic societies, most evidently in Britain but also elsewhere in Europe. But even when a Neolithic society made extensive use of domesticates, its use of space may have been based upon other factors such as cultural or social preferences for mobility (Whittle 1996) or the use of space as a symbolic resource (Hodder 1992; Tilley 1994).

The goal of this article is to explore relations between culture and land use. Culture and social relations can be envisioned as affecting land use in several ways. On a symbolic level, people's experience of their landscape is culturally conditioned, a line of argument well-explored in recent phenomenological analyses (following especially Barrett 1994; Thomas 1991; Tilley 1994) which we do not pursue further here. However, culture and land use must also be theorised from other perspectives. Central to any full theory of practice is the question of human-environment interaction. Put most simply, humans do not simply respond to environmental stimuli, but neither can they act freely without external constraint. Rather the relationship is inextricably reciprocal. The 'objective conditions of existence' (Bourdieu 1977) provide a context within which humans make decisions influenced by perception, symbolism, social relations, and many other factors. These human decisions have unintended consequences which alter extra-cultural conditions and hence the context of social action in the future. A full agency-oriented exploration of either phenomenological or environmental aspects of cultural land use is beyond current GIS capabilities, but GIS can profitably be used to explore specific aspects of the question.

Here, we focus on one link of the inextricably reciprocal relationship between human agency and resources mentioned above: how cultural choices affected land use and the resulting human landscape. Neolithic Italian people knew how to farm, herd, hunt and gather. All of these were potentially viable activities in most environments. How Neolithic people decided to allocate their work between these activities must thus have involved cultural choice as well as environmental practicalities. We use GIS to trace out the implications of this cultural choice systematically. What would the resulting landscape have looked like? What areas would have been used for farming, herding and foraging, and where would they have been located? What would the implications have been for site spacing and for potential conflicts over resources? Does this provide any support for economic or cultural theories of the Neolithic spread? Ultimately, this may provide material for symbolic interpretation as well. Given the mosaic of areas used for different purposes, what possibilities would the resulting landscape have afforded for symbolisms such as 'wild' zones? Can we infer anything about patterns of use and the associated values put on sectors of the landscape?

To anticipate our conclusions, we argue that even in overwhelmingly agricultural Neolithic communities, the greatest amount of space was used for non-agricultural purposes, and space between settlements would have been a valuable economic and social resource. Theoretically, this implies that Neolithic land use cannot be reduced to the economic needs of agriculture, but must have involved both the use of wild resources and considerable cultural and social choice.

Land use in the Southern Italian Neolithic

There is a surprising lack of discussion on Neolithic Italian land use. A series of site catchment analyses (Barker 1975, 1981; Jarman & Webley 1975; Sargent 1983) have tended to view site location as a correlate of the needs of an agricultural economy, though Barker also notably emphasized access to wild resources, particularly in highland sites. Both Italian (e.g. Cassano et al. 1987: 25; Tine 1983: 185) and Anglophone (e.g. Ammerman 1985: 95) works have tended to focus upon the need for arable land. Use of space for other purposes has rarely been considered a serious factor on site location and spacing.

The Southern Italian Neolithic is generally known in the Anglophone world as a landscape of agricultural villages, a picture which partly reflects the long tradition of research on the villaggi trincerati of the Tavoliere. This picture is certainly justified by substantial village sites known in Apulia, Basilicata, and Sicily. However, there was also substantial variation in both settlement and economy. Inhabited sites included non-ditched habitation sites, scattered huts which apparently did not form villages (e.g. Ammerman 1985; Maniscalco 1997, Albore Livadie et al. 1987-88), caves, and ritual sites both in and out of caves (Whitehouse 1992). This variation in settlement is paralleled by variation in economy. Palaeobotanical samples often contain gathered plants (Castelletti et al. 1987; Hopf 1991), most strikingly at Uzzo Cave (Costantini 1989). At present, however, the use of wild resources is best traced through the faunal record. It is common for faunal remains from village sites to contain over 95% domesticated species, with minimal contribution from hunted game (Castelletti et al. 1987; Sorrentino 1983, 1996; Tagliacozzo 1992: 90-92; Villari 1995; Wilkens 1996). But an increasing number of sites demonstrate more extensive use of wild animals (Tagliacozzo 1992: 92). Some are caves, for example Grotto Pacelli, where fox, hare and turtle were especially common (Striccoli 1988 cf. Sorrentino 1984) or Uzzo Cave (Tagliacozzo 1992). Others are cult sites (e.g. Santa Barbara, Castelletti et al. 1987) or settlements (e.g. Mulino Sant'Antonio, Albarella 1987-88). In addition, many Neolithic pigs archaeologically identified as domesticated may in fact have been wild (e.g. Rowley-Conwy 1997). As more non-village sites are excavated, the overall pattern emerging is similar to that for Central Italy (Barker 1975, 1981; Castelletti et al. 1987), where the proportion of foraged resources varied according to setting. Intriguingly, faunal assemblages from Southern Italian villages seem more rigidly domesticated than those from non-village sites even when the two kinds of site occur in close proximity, suggesting that some wild resources may have been consumed off-site.

It should also be remembered that even when wild resources were unimportant in total quantity, they may have been necessary for many purposes. Wild resources can be used for expediently gathered snack foods, tools and ornaments, antler, fur, feathers (Albarella 1997), gums and glues, flavourings, salt, dyes, lithic raw materials, building materials, and firewood (cf. Delano Smith 1983: 23). Wild resources would have been used to buffer lean times of year and as an emergency resource. Wild resources were also sometimes symbolically important; for instance, cave paintings depict hunting (Graziosi 1974, 1980; Whitehouse 1992), and Capreolus skulls were ceremonially deposited in the Ipogeo Manfredi cult site at Santa Barbara (Geniola 1987).

This evidence suggests that it is worth exploring the question to which other uses beyond agriculture, and beyond subsistence needs in general, may have structured Neolithic landscape use.

The study area

The specific focus of this study is southern Aspromonte in Calabria, at the tip of the Italian peninsula (Figure 1), where the Bova Marina Archaeological Project has been excavating and surveying since 1997. The landscape rises from sea level to above 1700 meters 20 kilometres inland, and is dominated by rugged hills and steep valleys. Level land is confined to small coastal plains, narrow river valleys, and a few confined montane plateaus.

[FIGURE 1 OMITTED]

GIS study areas can be defined arbitrarily, through practical considerations of data availability and preparation, or with reference to a theoretical problem. Here we choose the latter approach. The study area was devised to encompass the range of Neolithic settlement, from the coast to the highest known settlement at Bova Castello (ca. 800 meters above sea level). The east-west dimensions were chosen to create a representative 'slice' of terrain encompassing two major river valleys (San Pasquale and Siderone) and the ridges between them. The overall scale of the area was chosen to represent a human landscape; given the small scale of the Neolithic communities known, it seems highly likely that this area encompasses all the land areas directly exploited by residents of these sites. The source map utilised (see below) restricted the potential territory of Bova Castello to the north. However, this will not have altered calculations of the land requirements, and the slight shift in the distribution of utilised land around this site does not alter our conclusions significantly given how infrequently site territories overlapped. An enhanced study with a territory approximately six times as large is now in progress.

Four Neolithic sites are known within the study area, the territories of Bova Marina and Bova Superiore: San Pasquale, Umbro and Penitenzeria, and Bova Castello. Several other sites may be Neolithic but cannot be dated securely from surface remains. All four known sites are associated with the Stentinello culture, which represents the local Early and Middle Neolithic and dates to the sixth and fifth millennia BC; Umbro and Penitenzeria also have a Late Neolithic Diana culture occupation (Robb 1998; Robb 2001).

The highest site is Bova Castello, located under the medieval castle atop the rocky peak of Bova Superiore at 800 meters above sea level. This is in a very varied landscape with a moderately sloping hillside just below cliffs, which could have been used for habitation and cultivation as well as for other purposes. The site is known from survey and has never been excavated. Umbro and Penitenzeria are located about 150 meters apart on the margins of a small plateau at 400 meters altitude. Umbro is a very small rock shelter, which may be a special function site rather than a habitation (Robb 1998). Penitenzeria is an open air site which may have been a primary habitation, although excavation here is just beginning (Robb 2001). The sites are so close together that, if contemporary, they are very likely to have been used by the same community; hence for purposes of a regional GIS model they are considered together as a single social unit. San Pasquale, an unexcavated site, is located on a clay terrace just inland from the coast.

Although only Umbro has yet been excavated extensively, none of the four sites display characteristics of the large Neolithic villages known elsewhere in Italy, such as dense artefact scatters and concentrations of daub. However, all three are located near springs and the catchment for each site would have included arable land and wooded slopes and ridges; San Pasquale would also have been near coastal marshes and river valley resources. Economic evidence is available only for Umbro, where wheat and barley remains have been recovered (Ciaraldi 1999) and the faunal remains contain many domesticated animals, particularly sheep and goats (U. Albarella, pers. comm. 2000). Residents of Umbro used clay, stone, sand, marine shell and probably many other archaeologically less visible local resources.

While a few Neolithic villages elsewhere in Italy are thought to have contained up to 250 people (e.g. Passo di Corvo, Tine 1983), it is clear from site sizes and house numbers that both in Southern Italy in general and Aspromonte in particular most settlements were substantially smaller. Here we modelled each site as if it were occupied by 50 people. While notional, this figure provides a point of reference to base calculations upon. If anything, it is probably an over-estimate for these relatively small sites; if so, areas of utilised land would have been even smaller than those reconstructed below, which only reinforces the implications of the simulation.

Archaeologists commonly focus on sites, but ancient people lived in landscapes: areas whose ancient use is archaeologically invisible may have been culturally or economically very important. What would the landscape used by Neolithic people have looked like with different social choices of economic regime?

GIS reconstruction of land use: methods

In this study, GIS modelling is used as an exploratory tool rather than to recreate Neolithic land use exactly. Because many important parameters can only be guessed at, and others would have varied greatly from region to region and from year to year, it makes more sense to start with reasonable estimates of model parameters and observe the effect of varying them than to try to create a single best model of Neolithic economy. GIS provides a well-established, flexible tool for generating hypothetical scenarios and providing alternative perspectives of human behaviour (Verhagen et al. 1995; Harris & Lock 1995).

As a baseline for land use, a 1:25000 scale topographic map from the Istituto Geografico Militare was digitised. The resulting digital data served as a spatial database of which maps of landscape slope, altitude (DEM) and hydrology were derived. This was supplemented with geological data from the Carta Geologica d'Italia. These were created and combined in GRASS 4.2 to create a map of local environmental zones. Eight zones were defined through map algebra calculation of environmental characteristics: these included beach, river channel, river valley, clay terrace, sedimentary plateau, sedimentary hill slopes, metamorphic plateau and metamorphic hill slopes (see Figure 2). These reproduce modern environmental niches available in the region. The modern Calabrian landscape has been severely affected by deforestation, erosion and deposition. However, since the base map does not take into account present-day vegetation or soil cover, it can serve as a reasonable proxy for the Neolithic natural setting pending more detailed investigation of landscape change. Each environmental niche was then ranked for its utility for three kinds of ancient economic exploitation--pastoralism, foraging and agriculture; this ranking was for use in preferentially choosing land for specific production purposes. For this purpose, quantitative estimates of yield were not needed, and a simple qualitative relative ranking into more and less productive land was sufficient. For instance, clay terraces, sedimentary plateaux and metamorphic plateaux were modelled as productive farmland, with the other niches less productive; steep slopes were modelled as relatively more useful for foraging and pastoralism than for agriculture. Similarly, river valleys were considered a rich resource for foraging purposes, as they afford a greater diversity of herbs and water-loving plants, as well as the accompanying fauna.

[FIGURE 2 OMITTED]

We then incorporated the three Neolithic sites currently known in the study region, San Pasquale, Umbro/Penitenzeria and Bova Castello. For each one we posited a population of 50 people as discussed above.

The next step was to estimate the resources a group of 50 people would have required annually to survive. A family of 4-6 people living primarily on grain would have required about 1000 kg per year (Gregg 1988), and Neolithic grain yields are thought to have been between 500 and 1000 kg/ha (Barker 1985; Gregg 1988; Jarman & Webley 1975; Halstead 1981). Allowing some overproduction to insure against a bad season (Halstead 1989), a group of 50 would have required perhaps 10-15 ha of active grain plots with an additional amount in fallow fields and gardens of other crops. Other resources are less straightforward to estimate, and estimating overall land use also depends on the mixture of resources being combined. We have followed Gregg (1988), who provides the most detailed resource by resource estimate of land use needs for European Neolithic populations. Based on her figures, and adjusting for group size, 50 people subsisting entirely on crops would require 41 ha; the same group would require 564 ha to live entirely on gathered plants, 2130 ha to live entirely on herds, and 12 300 ha to live on hunted game. These are not viable economic regimes in most situations but these estimates can then be used to calculate land needs for groups living on various mixtures of economic resources. These figures were used to estimate the land needs and sustainable population densities of Neolithic communities according to 20 different models of subsistence (Figure 3, Table 1).

[FIGURE 3 OMITTED]

In modelling where utilised lands would be located, we also included accessibility. An accessibility map was created, using a simple friction algorithm based on slope (friction = log (slope + 1) +.5) to represent the time and effort of travelling outwards from each site by foot over a landscape dominated by steep hills and valleys. Again, the goal was not to reproduce Neolithic reality exactly but to model it adequately; in our on-the-ground observations, cliffs and watercourses do not create large-scale barriers in this area, and simple distance and slope account for most of the effort involved in moving around the territory.

Finally, to reconstruct the overall map of land use, we combined the modelled economic needs of each site with the environmental niche map and the accessibility map. It was assumed that each settlement would satisfy its needs by using the most accessible available land of the appropriate quality for each type of use. In calculating the total amount of land used, we use the maximum land used for a specific purpose rather than the total of land used for all purposes, since land can be used for multiple purposes (for instance, fallow fields can be used for herding or hunting).

The GIS model was run for 20 different mixes of resources to calculate land, potential population densities, and potential population figures for a group of 50 people (Figure 3; Table 1). Population density here is understood as the maximum possible within given economic choices rather than as actual prehistoric density, as are the population estimates. Ideal site spacing is based on assuming circular territories; this was probably not the case prehistorically, but the figure can give a notional measure of site spacing. These models cover the range of possibilities from pure foragers (models 1-3) to pure pastoralists (model 12) and pure farmers (model 7). Models 6, 8 and 17 are based on Gregg's actual estimates for resource mixes in LBK villages, while the others show resource mixes which may fit sites within the great range of variability known archaeologically for prehistoric sites in Southern Italy.

Resource land use maps for forager-farmers (model 6), farmers (model 8) and pastoralists (model 17) are shown in Figures 4-6. It should be emphasised that these are only samples of a wide range of resource choice. Our intention is to explore the effects of variation rather than to determine a single best-fitting model for Neolithic economies. Indeed, we doubt it is even possible to determine one best model for Italian Neolithic economies, if only because they varied greatly.

[FIGURES 4-6 OMITTED]

Social choice and the importance of non-agricultural resources in land use

This reconstruction sheds light on one linkage of the environment-agency dialectic: the effects of social choices on utilised landscapes. Inasmuch as land use over extended periods would have modified the local environment, such choices created the context for future actions, as well as defining areas of human frequentation available as symbolic resources.

The most important point of our reconstruction of Neolithic land use is that arable land was probably never the limiting resource on Neolithic settlement (cf. Jarman & Webley 1975). All models (including Model 7, 100% dependence on crops) showed adequate land for cultivation within a reasonable distance from the known sites. For sites in hilly zones such as around Umbro and Bova Castello, farmland was located in scattered small plots. However, in small-scale horticulture, small scattered plots would not have lowered productivity, and they may have been both ecologically less risky and more appropriate to Neolithic social organization. This lack of pressure on arable land was probably true for the Italian Neolithic in general. If arable land was adequately plentiful even in an extremely hilly landscape such as the Aspromonte slopes, it is even less likely to have been scarce in more open Neolithic landscapes such as the Tavoliere and the Matera regions. Moreover, since a relatively small proportion of a group's exploitation territory was gardened at any point, soil exhaustion seems unlikely. This conclusion, supported by the high site densities found in areas such as the Tavoliere (Cassano & Manfredini 1983; Tine 1983), is probably true at virtually all population densities likely in prehistory.

Returning to our theoretical premises, we need to free Neolithic land use from agricultural determinism. Clearly social choice was involved. If even mountainous zones could support fully agricultural or pastoral economies, we cannot necessarily ascribe the high levels of foraging in some central Apennines Neolithic groups (Barker 1975, 1981) to the need to exploit a specific ecological niche. Rather, other factors may have dictated economic choices. While physical factors such as soil character may have been important, Neolithic groups may have had cultural preferences for economic activities such as hunting, gathering or herding. It is tempting to relate such choices to historic moments in the development of the Neolithic, such as the 500-year delay in the spread of the Neolithic from Puglia northwards and inland (Skeates 1994).

This implies that we need to recognise and theorise the importance of 'wild' areas and resources in early farming societies. Counter-intuitively, even when farming was the mainstay of the economy, wild resources were probably the limiting factor on Neolithic settlement, with hunting the primary determinant on site spacing. As long as about 3% or more of resources came from game, foraging would have been the most extensive land use within a Neolithic economy. Most Neolithic Italian faunal assemblages contain at least this amount of game in terms of bone fragments, and the relative contribution to the diet would probably have been higher considering the meat weight of relatively large game such as deer and wild pig. Many other gathered resources such as honey, salt, shellfish, nuts, and stone of various types occur patchily too, and an extensive territory would have been needed to ensure of access to them. If Neolithic people symbolically marked their territories to differentiate them from those of other groups, it would have been primarily pasture and forest rather than agricultural land they were staking claims to. Negotiation of access to potentially overlapping areas would have been for forest and pasture. Since such small communities would not have been demographically or culturally self-sufficient, overlaps in areas used would often have been negotiated within the context of a common regional identity.

Space has social uses as well. In the Italian Neolithic landscape, sites were probably associated with concentric zones of space, with an area around each settlement generally understood as its territory for activities such as garden plots and short-range pastoralism, and a larger area understood as belonging generally to a network of related sites such as Pasquale, Umbro and Bova Castello. Right of access to land and use of its products would likely have been mushy, with less permeable boundaries close to settlements and more open access further away and where territories used by different settlements overlapped. Spaces between villages would have been an important social resource. Uncultivated space between sites would have afforded an ecological resource for foraging in case of crop failure. Among the uses ethnographically attested for the 'bush' are hunting, foraging, quarrying, trading, fighting, burial, public ceremonies, ritual seclusion, activities requiring privacy such as sex, and dangerous or objectionable work such as potting (e.g. Barley 1994). Carrying out some of these, such as trading, fighting, or political negotiation, can be much easier, or in fact only possible, in areas not intimately connected with a particular group. Other uses may be enhanced by the 'wild' symbolic connotations of interstitial areas, which may have had gender connotations as well (Robb & Morter 1998; Whitehouse 1992).

Regional population dynamics and the Neolithic spread

Relations between population density, economic intensification, and the use of space have often been discussed for the Mesolithic/Neolithic transition. The models explored here confirm generalisations that foraging populations such as Ice Age hunters (models 1-2) and Holocene foragers (models 2-3) would have lived in widely dispersed groups at low population densities. What has been less appreciated is the amount of variation possible within broadly agricultural and pastoral societies. The initial Neolithic at Uzzo Cave, for instance, may have involved a land use similar to our model 5, which allows a relatively low maximum population density of 1.55 people/[km.sup.2]. The same may be true for areas such as the highland Abruzzo. At the other end of the scale, the Tavoliere Neolithic villages, with a greater reliance on agriculture and far less hunting, may have resembled model 10, allowing a maximum population density of up to 8.38 persons/[km.sup.2] and a site spacing dictated primarily by pastoralism rather than foraging. These reconstructions suggest the population of Lipari was never more than marginal in terms of demographic self-sufficiency. Neolithic Malta may never have had more than a few thousand inhabitants and animal products may have been in shorter supply there than in less circumscribed environments (Table 1). Conversely, they may hint at links between a relatively high population density on the Tavoliere, the presence of large, nucleated villages, a high dependency on agriculture, and elaborate local pottery traditions which may represent small, bounded communication networks.

Similarly, it has often been proposed that Italian Final Neolithic and Copper Age societies relied on pastoralism to a greater extent than earlier Neolithic groups. If so, this would require more space and involve living at lower overall population densities than comparable agricultural populations (for instance, compare model 10 and model 14). In this context, the apparent late Neolithic 'abandonment' of the Tavoliere (Tint 1983) may reflect a shift to both a lower population density and a settlement pattern around the margins of the plain incorporating slope areas exploitable for herding. This would presumably be the result of a social choice (c.f. Robb 1999); simple demographic pressure would favour intensification of grain agriculture rather than extensification of herding.

Finally, demographic pressure on potential resources was not responsible for the spread of the Neolithic (e.g. Ammerman & Cavalli-Sforza 1984). Neolithic economies throughout Europe incorporated hunted and foraged resources (Barker 1985; Whittle 1996). Given a baseline economy with at least some foraged resources, spacing between settlements would have been determined by non-agricultural land uses, and each group's territory would have included much non-intensively used land. When adjacent territories overflow into each other, it is virtually always the wild resource zones which overlap. While local crop fluctuations might have caused famines, the initial effect of population pressure per se on agricultural economies would not have been starvation; it would have been restriction of the use of wild resources. At the low population densities prevalent in prehistory, and even with low-technology farming regimes, higher total productivity was virtually always available within an existing territory simply by changing the mixture of resources used, and particularly by planting more and foraging less. Thus, acceptable levels of landscape occupation may have been determined not by subsistence needs nor by a theoretical ecological carrying capacity, but by a feeling of 'crowding' reflecting the need for distance from neighbours, game for social consumption, and uninhabited spaces for a wide variety of social activities. As this implies, the decision of farmers to colonise new territories rather than intensify their economic regime, like the decision of foragers to adopt domestic plants and animals, would have been a social and cultural decision, not a simple response to need.

Table 1 GIS models of Neolithic land use with varying resource mixes.

 % % % %
 reliance reliance reliance reliance
 on on on gathered on hunted
Model crops herds plants game

1 Forager 1 0.00 0.00 20.00 80.00
2 Forager 2 0.00 0.00 50.00 50.00
3 Forager 3 0.00 0.00 80.00 20.00
4 Forager-farmer 1 15.00 15.00 35.00 35.00
5 Forager-farmer 2 40.00 10.00 25.00 25.00
6 Forager-farmer 3 * 66.60 3.30 6.40 23.70
7 Farmer 1 100.00 0.00 0.00 0.00
8 Farmer 2 * 62.50 22.90 2.80 11.80
9 Farmer 3 65.00 25.00 5.00 5.00
10 Farmer 4 68.00 28.00 3.00 1.00
11 Farmer 5 70.00 30.00 0.00 0.00
12 Pastoralist 1 0.00 100.00 0.00 0.00
13 Pastoralist 2 50.00 50.00 0.00 0.00
14 Pastoralist 3 48.00 48.00 3.00 1.00
15 Pastoralist 4 30.00 70.00 0.00 0.00
16 Pastoralist 5 28.00 68.00 3.00 1.00
17 Pastoralist 6 * 25.80 71.00 3.20 0.00
18 Pastoralist 7 0 50 25 25
19 Pastoralist 8 15 70 7.5 7.5
20 Mixed 33.5 33.5 16.5 16.5

 Total Max.
 land Ideal pop.
 needs site density
Model ([km.sup.2]) spacing ([km.sup.2])

1 Forager 1 99.53 11.26 0.50
2 Forager 2 64.32 9.05 0.78
3 Forager 3 29.11 6.09 1.72
4 Forager-farmer 1 45.02 7.57 1.11
5 Forager-farmer 2 32.16 6.40 1.55
6 Forager-farmer 3 * 29.51 6.13 1.69
7 Farmer 1 0.41 0.72 121.95
8 Farmer 2 * 14.67 4.32 3.41
9 Farmer 3 6.43 2.86 7.77
10 Farmer 4 5.96 2.76 8.38
11 Farmer 5 6.39 2.85 7.82
12 Pastoralist 1 21.30 5.21 2.35
13 Pastoralist 2 10.65 3.68 4.69
14 Pastoralist 3 10.22 3.61 4.89
15 Pastoralist 4 14.91 4.36 3.35
16 Pastoralist 5 14.48 4.29 3.45
17 Pastoralist 6 * 15.12 4.39 3.31
18 Pastoralist 7 32.16 6.40 1.55
19 Pastoralist 8 14.91 4.36 3.35
20 Mixed 21.23 5.20 2.36

 Max. Max. pop.
 pop. Bova study
 Lipari area
Model (38 [km.sup.2]) (132 [km.sup.2])

1 Forager 1 19 66
2 Forager 2 30 103
3 Forager 3 65 227
4 Forager-farmer 1 42 147
5 Forager-farmer 2 59 205
6 Forager-farmer 3 * 64 224
7 Farmer 1 4634 16098
8 Farmer 2 * 129 450
9 Farmer 3 295 1026
10 Farmer 4 319 1107
11 Farmer 5 297 1033
12 Pastoralist 1 89 310
13 Pastoralist 2 178 620
14 Pastoralist 3 186 646
15 Pastoralist 4 127 443
16 Pastoralist 5 131 456
17 Pastoralist 6 * 126 436
18 Pastoralist 7 59 205
19 Pastoralist 8 127 443
20 Mixed 90 311

 Max. Max.
 pop. pop.
 Malta Tavoliere
Model (316 [km.sup.2]) (4500 [km.sup.2])

1 Forager 1 159 2261
2 Forager 2 246 3498
3 Forager 3 543 7729
4 Forager-farmer 1 351 4997
5 Forager-farmer 2 491 6996
6 Forager-farmer 3 * 535 7624
7 Farmer 1 38537 548780
8 Farmer 2 * 1077 15335
9 Farmer 3 2456 34981
10 Farmer 4 2649 37726
11 Farmer 5 2473 35211
12 Pastoralist 1 742 10563
13 Pastoralist 2 1484 21127
14 Pastoralist 3 1545 22007
15 Pastoralist 4 1060 15091
16 Pastoralist 5 1091 15534
17 Pastoralist 6 * 1045 14878
18 Pastoralist 7 491 6996
19 Pastoralist 8 1060 15091
20 Mixed 744 10600

 Max.
 pop.
 Italy
Model (301225 [km.sup.2])

1 Forager 1 151327
2 Forager 2 234161
3 Forager 3 517355
4 Forager-farmer 1 334516
5 Forager-farmer 2 468322
6 Forager-farmer 3 * 510344
7 Farmer 1 36734756
8 Farmer 2 * 1026536
9 Farmer 3 2341612
10 Farmer 4 2525360
11 Farmer 5 2357003
12 Pastoralist 1 707101
13 Pastoralist 2 1414202
14 Pastoralist 3 1473127
15 Pastoralist 4 1010144
16 Pastoralist 5 1039854
17 Pastoralist 6 * 995917
18 Pastoralist 7 468322
19 Pastoralist 8 1010144
20 Mixed 709579

Max.: Maximum; Pop.: Population.

Received: 12 December 2000; revised: 4 January 2001; accepted: 17 March 2003

Acknowledgements

We are grateful to David Wheatley, Lee Hazelwood and Rob Hosfield for discussion of the GIS model and to two anonymous reviewers whose comments improved the manuscript.

John Robb (1) & Doortje Van Hove (2)

(1) Department of Archaeology, University of Cambridge, Downing Street, Cambridge CB2 3DZ UK (Email: jer39@cam.ac.uk)

(2) Department of Archaeology, University of Southampton, Highfield, Southampton S01 71BF UK (Email: dvh198@soton.ac.uk)

References

ALBARELLA, U. 1987-88. La fauna, Rivista di Scienze Preistoriche 41: 84-98.

--1997. Crane and vulture at an Italian Bronze Age site, International Journal of Osteoarchaeology 7: 346-349.

ALBORE LIVADIE, C., F. FEDELE, U. ALBARELLA, F. DE MATTEIS, E. ESPOSITO & R. FEDERICO. 1987-88. Ricerche sull'insediamento tardo-neolitico di Mulino Sant'Antonio (Avella), Rivista di Scienze Preistoriche 41:65-103.

AMMERMAN, A. (ed). 1985. The Acconia Survey: Neolithic settlement and the obsidian trade. London: Institute of Archaeology.

AMMERMAN, A. & L. CAVALLI-SFORZA. 1984. The Neolithic transition and the genetics of populations in Europe. Princeton: Princeton University Press.

BARKER, G. 1975. Prehistoric territories and economies in Central Italy, in E. Higgs (ed.), Palaeoeconomy: 111-175. Cambridge: Cambridge University Press.

--1981. Landscape and Society: Prehistoric Central Italy. New York: Academic.

--1985. Prehistoric Farming in Europe. Cambridge: Cambridge University Press.

BARLEY, N. 1994. Smashing pots: feats of clay from Africa. London: British Museum.

BARRETT, J. 1994. Fragments from Antiquity. Oxford: Blackwell.

BOURDIEU, P. 1977. Outline of a Theory of Practice. Cambridge: Cambridge University Press.

CASSANO, S. & A. MANFREDINI. 1983. Studi sul Neolitico del Tavoliere della Puglia. Oxford: British Archaeological Reports.

CASSANO, S., A. CAZZELLA, A. MANFREDINI, & M. MOSCOLONI (ed). 1987. Coppa Nevigata e il suo territorio: testimonianze archeologiche dal VII al II millennio a.C. Roma: Edizioni Quasar.

CASTELLETI, L., L. COSTANTINI & C. TOZZI. 1987. Considerazioni sull'economia e l'ambiente durante il neolitico in Italia, Atti, Riunione Scientifica dell'I.I.P.P. 26:37-55.

CIARALDI, M. 1999. Paleobotanical remains, in J. Robb (ed.) Bova Marina Archaeological Project: Survey and Excavations at Umbro, Preliminary Report, 1999 Season: 38-40. Southampton: Department of Archaeology, University of Southampton.

COSTANTINI, L. 1989. Plant exploitation at Grotta dell'Uzzo, Sicily: new evidence for the transition from Mesolithic to Neolithic subsistence in Southern Europe, in D. Harris & G. Hillman (eds.), Foraging and farming: the evolution of plant exploitation: 197-206. London: Unwin Hyman.

DELANO SMITH, C. 1983. L'ambiente, in S. Tine (ed.), Passo di Corvo e la civilta neolitica del Tavoliere: 11-22. Genova: Sagep.

GENIOLA, A. 1987. La cultura di Serra d'Alto nella Puglia centrale, Atti, Riunione Scientifica dell'I.I.P.P. 26:771-782.

GRAZIOSI, P. 1974. L'arte preistorica in Italia. Firenze: Sansoni.

--1980. Le pitture preistoriche di Porto Badisco. Firenze: Martelli.

GREGG, S. 1988. Foragers and farmers: population interaction and agricultural expansion in prehistoric Europe. Chicago: University of Chicago Press.

HALSTEAD, P. 1981. Counting sheep in Neolithic and Bronze Age Greece, in I. Hodder, G. Isaac, & N. Hammond (eds.), Pattern of the past: studies in honour of David Clarke: 307-339. Cambridge: Cambridge University Press.

--1989. The economy has a normal surplus: economic stability and social change among early farming communities of Thessaly, Greece, in P. Halstead & J. O'Shea (eds.), Bad Year Economics: 68-80. Cambridge: Cambridge University Press.

HARRIS T. M. & G.R. LOCK. 1995. Towards an evaluation of GIS in European archaeology: the past, present and future of theory and applications, in G. Lock and Z. Stancic (eds.) Archaeology and Geographical Information Systems: A European Perspective, 349-365. London: Taylor and Francis.

HIGGS, E. (ed). 1975. Palaeoeconomy. Cambridge: Cambridge University Press.

HODDER, I. 1992. The domestication of Europe, in I. Hodder (ed.), Theory and practice in archaeology: 241-253. London: Routledge.

HOPF, M. 1991. South and Southwest Europe, in W. Van Zeist, K. Wasylikowa, & K. E. Behre (eds.), Progress in Old World Palaeoethnobotany: 241-277. Rotterdam: Balkema.

JARMAN, M. & D. WEBLEY. 1975. Settlement and land use in Capitanata, Italy, in E. Higgs (ed.), Palaeoeconomy: 177-231. Cambridge: Cambridge University Press.

MANISCALCO, L. 1997. L'insediamento preistorico presso le Satinelle di San Marco, in S. Tusa (ed.), Prima Sicilia: 193-197. Palermo: Ediprint.

ROBB, J. 1998. Bova Marina Archaeological Project: Survey and Excavations at Umbro. Preliminary Report, 1998 Season. Southampton: Department of Archaeology, University of Southampton.

--2001. Bova Marina Archaeological Project: Survey and Excavations. Preliminary Report, 2001 Season. Cambridge: Department of Archaeology, University of Cambridge.

--1999. Great persons and big men in the Italian Neolithic, in R. H. Tykot, J. Morter, & J. E. Robb (eds.), Social Dynamics of the prehistoric Central Mediterranean: 111-122. London: Accordia Research Centre.

ROBB, J. E. & J. MORTER. 1998. Space, gender and architecture in the southern Italian Neolithic, in R. Whitehouse (ed.), Gender and Italian archaeology: challenging the stereotypes: 83-94. London Accordia Research Centre.

ROWLEY-CONWY, P. 1997. The animal bones from Arene Candide: final report, in R. Maggi (ed.), Arene Candide: a functional and environmental assessment of the Holocene sequence. 153-277. Rome: Il Calamo.

SARGENT, A. 1983. Exploitation territory and economy in the Tavoliere of Apulia, in S. Cassano & A. Manfredini (eds.), Studi sul Neolitico del Tavoliere della Puglia: 223-236. Oxford: British Archaeological Reports.

SKEATES, R. 1994. Towards an absolute chronology for the Neolithic in central Italy, in R. Skeates & R. Whitehouse (eds.), Radiocarbon dating and Italian prehistory: 61-72. London: Accordia Research Center.

SORRENTINO, C. 1983. La fauna, in S. Tine (ed.), Passo di Corvo e la civilta neolitica del Tavoliere: 149-158. Genova: Sagep.

--1984. Lo studio della fauna di Tirlecchia, Rivista di Scienze Preistoriche 39: 73-84.

--1996. Favella, in V. Tine (ed.), Formi e tempi della neolitizazzione in Italia meridionale e in Sicilia: 520-525. Cosenza: Rubbettino.

STRICCOLI, R. 1988. Le culture preistoriche di Grotta Pacelli (Castellana Grotte, Bart). Brindisi: Schena Editore.

TAGLIACOZZO, A. 1992. I mammiferi dei giacimenti pre- e protostorici italiani: un inquadramento paleontologico e archeozoologico, in A. Guidi & M. Piperno (eds.), Italia preistorica: 68-102. Rome: Laterza.

THOMAS, J. 1991. Rethinking the Neolithic. Cambridge: Cambridge University Press.

TILLEY, C. 1994. A phenomenonology of landscape: places, paths and monuments. Oxford: Berg.

TINE, S. 1983. Passo di Corvo e la civilta neolitica del Tavoliere. Genova: Sagep.

VERHAGEN, P., J. MCGLADE, S. GILI & R. RISCH. 1995. Some criteria for modelling socio-economic activities in the Bronze Age of South East Spain, in G. Lock and Z. Stancic (eds.) Archaeology and Geographical Information Systems: A European Perspective, 187-209. London: Taylor and Francis.

VILLARI, P. 1995. Le faune della tarda preistoria nella Sicilia orientale. Siracusa: Ente Fauna Siciliana.

WHITEHOUSE, R. 1992. Underground religion: cult and culture in prehistoric Italy. London: Accordia Research Center.

WHITTLE, A. 1996. Europe in the Neolithic: the creation of new worlds. Cambridge: Cambridge University Press.

WILKENS, B. 1996. Latronico 3, Santo Stefano, Avezzano, Ripa Tetta, in V. Tine (ed.), Formi e tempi della neolitizazzione in Italia meridionale e in Sicilia: 507-510. Cosenza: Rubbettino.