A characteristic feature of human subsistence as the last glaciation ended was the turn towards new food sources, in a `broad spectrum' transformation. Australia took an unusual course, and the trajectory in its arid zone is especially striking. What were the broad spectrum diets in arid Australia? Why did they arise so late? Did they arise late?
Following the peak of the last glaciation, human diets in most parts of the world changed dramatically. Wild plant and animal foods previously available but not much exploited suddenly became important, sometimes dominant, elements of local diets. Because of the diversity of new foods involved, this change is sometimes called the `broad spectrum revolution' (Flannery 1969). Many connect its onset with other changes in human behaviour, including the appearance of new technologies, larger and more stable settlements, higher regional population densities, increased reliance on food storage, and, in some instances, the development of agriculture and the emergence of complex social organization.
Experts disagree about the causes of this dietary `revolution'. Some see it as a product of terminal Pleistocene climatic changes that made these resources more widely available, and more attractive to potential consumers. Others, noting high processing costs and low nutrient return rates, attribute their adoption to the elimination of previously preferred prey, either by climatic change, human population growth and consequent over-predation, or a combination of the two. Still others take the change in diet to reflect the emergence of social institutions that require more intensive exploitation of all resources, particularly those that are most abundant, regardless of cost and independent of climatic change.
Australia provides an interesting and potentially important example of the problem. Judging from the recent literature, it apparently witnesses the broad spectrum change in diet continent-wide, in combination with the appearance of new technologies, changes in settlement pattern at several scales, and a sharp increase in refuse deposition rates read by many to mark a general increase in population density. But these changes are late: on present evidence no earlier than the mid Holocene, well after the first evidence for similar changes on other continents, and unconnected with any major climatic changes of the kind some see as provoking its onset elsewhere. This difference makes Australia a critical test for any general explanation of this phenomenon, particularly those that appeal to climate as the cause.
Here we discuss the origin of the broad spectrum revolution in Australia, with special reference to the arid zone. We focus on this region because of its size, the availability of an ethnographic record in which the use of broad spectrum resources figures prominently, and because their exploitation may be understandable in terms of the same few constraints region-wide.
We begin with an overview of regional environments and continue with summaries of pertinent ethnographic, paleoenvironmental, and archaeological data. We then elaborate general arguments about the origins of the broad spectrum revolution, and evaluate each in light of the arid Australian data.
The arid zone
The arid zone can be defined as that large part of the continent (c. 5,000,000 sq. km, 70% of its present land surface) where evaporation equals or exceeds precipitation (Figure 1). Analysts have long agreed that its occupation by humans is tightly constrained by the distribution of food and water, both in turn largely determined by climate and local topography.
By definition, the arid zone is hot and dry (Gentilli 1971). Summers are very hot (mean January highs >35[degrees]C); winters are moderate (mean July lows ~5-10[degrees]C. Rainfall is low and highly variable. The annual average is less than 500 mm everywhere, less than 125 mm in the driest parts. Inter-annual variation ranges from 30-60% and correlates inversely with the annual mean. In the northern part of the zone, most rains fall in the summer; along the southern margin, they occur primarily in autumn and winter. Elsewhere, seasonal patterning is absent.
Land-forms can be divided into six broad categories (Figure 2, after Mabbutt 1971; see also Beard 1981; 1984). Two are relatively favourable for human settlement.
Uplands (18% of the arid zone) are large areas of locally high relief. They include the MacDonnell and Musgrave Ranges of the Central Desert, the Pilbara District to the west, and the southern margins of the Kimberley to the northwest. None are high enough to affect rainfall, but all concentrate run-off in stream channels and near-surface aquifers, providing the best and most accessible water sources available over most of the region. The combination of high relief and a diverse substrate also contributes to relatively high biotic diversity.
Riverine flood-plains (4% of the arid zone), encountered along the margins of most uplands, are best developed along the eastern edge of the desert, where perennial and substantial seasonal streams (e.g. Darling River, Cooper Creek) that rise in the Eastern Highlands transect the zone. Most contain surface waters in at least some seasons; all store readily accessible sub-surface reserves of various sizes and degrees of persistence, some quite large and effectively permanent. Biotic diversity is also comparatively high in these areas. They are unique among arid Australian habitats in offering a wide range of aquatic resources, including fish, shellfish, waterfowl and hydrophytic plants (e.g. Allen 1974; Kimber 1984).
The remaining four land-form categories, stony, sand, and shield deserts and clay pans (collectively, 78% of the arid zone, shown as `open desert' in Figure 2), are less attractive. All are marked by uncoordinated drainage patterns: generally, well-defined stream channels are either absent or small and very short, running from local high points to near-by basins. Surface water sources may be widespread immediately after rain but disappear quickly thereafter. Most sub-surface catchments are small and seldom retain water for more than a few weeks. Biotic diversity is also limited.
Ethnographic patterns of land and resource
At the beginning of European contact in the mid 19th century, Aboriginal populations throughout the arid zone were small and widely dispersed. Estimates range from a high of one individual per 13 square kilometres among the Arrernte (Strehlow 1965) to a low of one per 200 square kilometres among the Pintupi (Long 1971). Density seems to have varied directly with the availability and abundance of food and, especially, water: highest in upland areas and along riverine flood-plains; lowest in stony or sandy deserts, particularly those with low, highly variable rainfall regimes.
Local diets everywhere were broad and broadly similar. The list of prey taken included birds, marsupials, reptiles, larvae and a wide range of fruit, tubers, and seeds (for details, see Cane 1987; 1989; Cleland 1966; Cleland & Johnston 1933; Cleland & Tindale 1959; Gould 1969; Latz 1982; Meggitt 1957; O'Connell & Hawkes 1981; O'Connell et al. 1983; Peterson 1977; 1978; Silberbauer 1971; Sweeney 1947; Thomson 1992; Tindale 1972; 1977; Veth & Walsh 1988; Walsh 1987).
Post-encounter returns available from these resources varied widely (Cane 1987; 1989; O'Connell & Hawkes 1981; O'Connell et al. 1983). Among plants, the best yields were derived from fruits and tubers, some of which (e.g. Solanum, Ipomoea) produced average returns of several thousand calories per hour spent collecting and processing. Despite their high caloric content, herb, tree and grass seeds (e.g. Acacia, Bracheria, Chenopodium, Eragrostis, Panicum) gave much lower returns - in the 300-1100 kcal/hr range - largely as a function of heavy processing requirements.(1)
These seeds are classic broad spectrum resources. Comparison of the more extensive resource lists shows they were very important in ethnographic diets throughout the arid zone. Their widespread use indicates that, at least at contact, diet breadth was essentially constant across the entire region, regardless of differences in local resource diversity or productivity.
Seeds were probably most critical during dry seasons, when local groups were closely tied to permanent or near-permanent water sources and entirely dependent on foods available within a day's round-trip walk. As such sites were often occupied for periods of up to several months, higher-ranked prey are likely to have been depleted, necessitating heavy reliance on seeds. Along with the availability of water, abundance of seeds may have been a key determinant of variation in local human population density.
It is sometimes suggested that seeds were also critical to the support of large groups (numbering in the hundreds) that gathered in favourable seasons for periods of up to several weeks in connection with male initiation and other ritual activities (e.g. Hamilton 1980, Spencer & Gillen 1912:259, Tindale 1972:244-5). Evidence cited in support of this proposition is entirely anecdotal.
Past climatic and environmental change
Data on past environmental conditions within the arid zone are limited. As a result, comprehensive inferences on this topic must proceed through an extended chain of argument beginning with data on past land-forms, pollen, and lake-levels elsewhere in Australia. Implications of these data are then developed as models of general atmospheric circulation, which in turn support inferences about past climates in the arid zone, some of which can be tested with evidence from the zone itself.
Sandy Harrison & John Dodson (1993; also Harrison 1993) provide the basis for an argument of this type (see Bowler 1976; Bowler et al. 1976; Chappell 1991; Chappell & Grindrod 1983; Harrison et al. 1984; Kershaw, this volume; Wasson 1984; and Webster & Streten 1978 for alternatives). Their model assumes the existence of a generalized global temperature curve marked by:
* very low temperatures at the peak of the last glaciation (18,000-20,000 b.p.),
* a subsequent warming trend accelerating after 12,000 b.p. through a peak at about 5000-7000 b.p., and
* a return to somewhat cooler, essentially modern conditions thereafter. It also recognizes that regional rainfall patterns are determined largely by trends in atmospheric circulation. The present limits of the Australian arid zone are set primarily by the breadth and seasonal movement of the subtropical high-pressure belt (Gentilli 1971). In summer the belt moves south, allowing monsoonal fronts on its northern margin to penetrate the arid zone. In winter the belt shifts north, permitting rain-bearing westerlies to cross the southern desert margin.
Drawing on fossil pollen and lake-level data from New Guinea and southern Australia, Harrison & Dodson (1993) argue that the subtropical high was broader than at present from about 25,000 b.p. through the peak of the last glaciation. Its northern margin may have been at or just north of its current location, its southern margin at least 8[degrees]S of its present position. Winter westerlies were forced well south of their current track, leaving the southern margin of the continent much drier than it is at present. Summer monsoons may have been weakened, not only by any northerly movement of the subtropical high, but also by lower sea-levels and the consequent exposure of the broad Arafura Plain (Nix & Kalma 1972). After 12,000 b.p., the subtropical high evidently began to contract, progressively enhancing westerly-related moisture conditions in Tasmania and southeast Australia as its southern margin moved north. At the same time, rising sea-levels and the flooding of the Arafura Plain led to increased monsoonal activity across the newly reconfigured northern edge of the continent. By 6000 b.p., the high-pressure belt was narrower than at present, with its southern margin somewhat north of its current limit. After 6000 b.p., the process was reversed, with the subtropical high assuming essentially modern form by 3000 b.p.
By this argument during 12,000-25,000 b.p. (especially 16,000-20,000 b.p.), conditions throughout the arid zone were much colder and drier than they are at present. Local pollen, lake-level and dune-formation data are consistent with this proposition (Bowler 1976; Bowler & Wasson 1984; Bowler et al. 1986; Chen 1989; Chen et al. 1990; Jacobson et al. 1988; Jennings 1975; Martin 1973; Patton et al. 1993; Ross et al. 1992; Singh a Luly 1991; Wasson 1984; 1989). Before 25,000 b.p., conditions were significantly wetter but not much cooler (<3[degrees]C) than at present. Lake-levels were higher- surface waters generally more widespread. By 20,000 b.p., average temperatures had evidently fallen 5-8[degrees]C below modern averages, seasonal temperature contrasts were greater, rainfall was down > 50%, wind speeds had increased sharply, and evaporation rates were significantly higher. Vegetation (particularly tree cover) was much reduced, dune systems more active, and the arid zone itself much larger than it is today (Figure 1).
These conditions had important implications for local human populations (Smith 1989a: 100-102). Plant and animal resources were probably less diverse than they are now, tropical flora less widespread, and overall productivity lower. Water sources were fewer and less reliable, making large areas inaccessible. Reliable waters may still have been found in the uplands, particularly where near-subsurface aquifers with long recharge times ([10.sup.2]-[10.sup.4] years) remained available (e.g. Jacobson et al. 1988; 1989).
The principle exception to this general picture may have been on the southeastern margin of the zone, where several lines of evidence indicate high lake levels during the period 14,000-17,000 b.p. (Balme 1995; Harrison 1993; Hope 1993; Prentice et al. 1992; Ross et al. 1992). Harrison & Dodson (1993: 282) suggest that these were supported by rains associated with `troughs between the travelling anticyclones, just as is observed today when the tropical high-pressure belt is unusually far south'. Whatever the cause, their presence may have created especially favourable habitats, not only around the lakes themselves but along the rivers that fed them. They were apparently dry after 14,000 b.p.
After 16,000 b.p., conditions throughout much of the arid zone improved. Plant cover increased, dunes became stabilized, tropical flora spread south, overall biotic productivity was enhanced, water sources were more numerous, more widely available and more reliable, and the zone itself contracted in size. Opportunities for occupation were probably most favourable between 5000-7000 b.p. when lakes along the southern margin of the zone and elsewhere were full. After that time, conditions become somewhat cooler and drier, approximating those known historically.
Archaeological research in the arid zone
Archaeological research in the arid zone has until recently been quite limited (Ross et al. 1992). The first serious work was initiated in the mid 1960s by Richard Gould (1977; 1978). His excavations at Puntutjarpa, in the Western Desert, and Intirtekwerle (James Range East Rockshelter), south of Alice Springs, revealed long occupation sequences marked by relatively unchanging tool-kits whose composition generally matched the lithic component of those used throughout the zone ethnographically. These results led him to propose that the pattern of local Aboriginal land and resource use, including heavy reliance on tree and grass seeds, had been characteristic of the area from the time it was first inhabited, an event he dated as early as 10,000 b.p. Most commentators initially agreed, arguing that it was the development of seed-grinding technology, then dated to 15,000 b.p. elsewhere in Australia, that permitted the arid zone to be occupied in the first place (e.g. Bowdler 1977; Horton 1981; O'Connell & Hawkes 1981).
These propositions were challenged by several studies reported in the early to mid 1980s. Three points were particularly important:
* Excavations at several sites showed that some parts of the arid zone were occupied as early 26,000 b.p. (Brown 1987; Lampert & Hughes 1988; Maynard 1980; Smith 1987; Wasson 1983; also Wright 1971a).
* These and other excavations, in combination with systematic archaeological survey, indicated that different parts of the zone may have had very different occupational histories (see Ross et al. 1992 for a general summary; also Balme 1995; Davidson et al. 1993; Hiscock 1988; Hope 1993; Lampert & Hughes 1988; Morse 1988; O'Connor et al. 1993; Smith 1988; 1989a; 1993; Smith & Clarke 1993; Smith et al. 1991; Veth 1989; 1993; 1994; Veth et al. 1990; Williams 1988). Some areas were evidently used prior to onset of Last Glacial Maximum aridity, others were first occupied after 15,000 b.p., still others only after 5000 b.p., and some even more recently. All sequences display fluctuations in occupation intensity; most (but not all) show an increase in use sometime after the mid Holocene.
* Detailed reconsideration of the evidence for the antiquity of seed-grinding technology indicated that tools like those known ethnographically appeared quite late in the archaeological record, no earlier than 5000 b.p. anywhere on the continent, possibly as late as 1500-2000 b.p. in the arid zone (Smith 1985; 1986; 1988; 1989b). While there are earlier examples of grinding tools, from both the arid zone and elsewhere dating before the Last Glacial Maximum, their role in seed processing remains unclear (e.g. Furby et al. 1993; Gorecki et al. in prep.; Hope 1993; cf. Allen 1990; Balme 1991; Smith 1985 1986).
Analysts attempting to account for these data fall into two groups: those who see arid zone prehistory as largely the product of environment and demography, and those who emphasize the importance of independent changes in technology and social organization. Michael Smith (1988; 1989a; 1993; Smith et al. 1991), a proponent of the former view, argues that the region was widely but thinly occupied as early as 30,000 b.p. Increased aridity after 25,000 b.p. forced abandonment of all areas except the larger uplands. Wherever local groups persisted, they were very small and highly mobile. Diet breadth increased but not enough to include seeds. When conditions improved after 15,000., populations grew and dispersed throughout the zone, reaching their maximum size and extent by 6000 b.p. Cooler and drier conditions after that time reduced the availability of key subsistence resources and restricted the number and distribution of surface waters. Relatively high population densities everywhere prevented the general increase in foraging ranges that might have been anticipated as a result. Instead, local groups were forced to broaden their diets by adopting the use of tree and grass seeds that had been available before but not exploited because of their high processing costs.
The best-developed argument for the importance of social and technological change is presented by Peter Veth (1987; 1989; 1993; 1994; Veth et al. 1990). He divides the arid zone into three habitat types: refuges, corridors and barrier deserts. Like Smith, he argues that the first two (including all upland areas plus certain intervening, relatively well-watered lowlands) were occupied at least intermittently (refuges perhaps permanently) from a time prior to the Last Glacial Maximum. He also reckons that variation in their use was directly related to past changes in rainfall. Unlike Smith, he thinks `barrier' deserts (mainly dunefield areas of the Great Sandy, Gibson, and Simpson Deserts) present a very different set of problems, primarily as a function of their very limited food and water resources. He contends they were not inhabited until after 5000 b.p., when key innovations in technology (seed-grinding tools, adzes suitable for working desert hardwoods, the ability to exploit desert aquifers by means of deep wells) and social organization (extended social networks, long-distance trade) were adopted. He asserts that these innovations developed in response not to immediate climatic stimuli but to `a combination of demographic pressure, technological shifts and changes in social organization' (Veth 1989: 83) that had `roots ... within the refuges of the glacial maximum' but did not emerge as `basic elements of the ethnographic desert economy' until the mid Holocene (Veth 1989: 90).
Anne Ross and colleagues (1992) take this argument a step further, suggesting that mid-Holocene developments in the arid zone were in fact part of a continent-wide pattern of sharp population growth and intensified use of broad spectrum resources. They think this pattern had little to do with climatic or environmental change but instead appeared as the result of social forces that required increased production of foodstuffs for reasons unrelated to basic subsistence requirements.
As indicated above, general explanations for the adoption of broad spectrum diets fall into three categories. Two appeal to changes in resource availability; the third to social forces. Having reviewed the evidence of dietary change from arid Australia, we are now in a position to evaluate these arguments.
The first explanation is based on the proposition that terminal Pleistocene climatic change made certain resources more widely available, and hence more attractive to human consumers (e.g. Binford 1968; Henry 1989; McCorriston & Hole 1991). Once adopted, they are said to have provided the basis for fundamental changes in human behaviour, including sedentary life-styles, increased population growth rates, related changes in technology and social organization, and (in some cases) the development of agriculture.
The initial proposition can be challenged on economic grounds. There are good reasons to expect that the decision to take a particular resource is determined in part by the returns available from pursuing other options (Stephens & Krebs 1986: 17-24). While many classic broad spectrum foods are seasonally abundant and nutrient-rich, they are also relatively expensive to collect and process, making them comparatively unattractive candidates for inclusion in the diet, regardless of their own abundance.
This argument is not often pursued in Australia (but see Tindale 1959), perhaps because its inherent limitations are generally recognized there. Whatever the reason, data from the arid zone provide a useful illustration of its shortcomings. As indicated above, post-encounter returns on tree and grass seeds commonly exploited there fall in the 300-1100 kcal/hr range. This means individual collectors relying solely on these resources must have devoted 2-7 hours/day to collecting and processing alone, independent of travel and search, simply to satisfy their own daily energy requirements. If they were responsible for 50% of the total caloric intake for a family of five, and if they drew that entirely from seeds yielding an average post-encounter return of 700 kcal/hr, then they must have spent 7 hours/day every day processing food, again independent of travel, search, or any other activity.
Assuming that maximizing the rate of nutrient capture is a key determinant of resource choice, it is difficult to see why seeds yielding such low returns would be added to the diet unless other, more attractive resources were unavailable. The fact that seeds were typically dropped from arid zone Aboriginal diets when European foods became available in quantity, while fruits and tubers yielding post-encounter returns up to an order of magnitude higher continued to be exploited, is consistent with this line of argument (Cane 1987; O'Connell & Hawkes 1981).
If broad spectrum resources world-wide are generally expensive to collect and process relative to other foods, then, contrary to common argument, changes in their own abundance, however great, are unlikely to account for their adoption, not only in arid Australia, but anywhere.
This is the converse of the argument from abundance. It turns explicitly on the proposition that many broad spectrum resources are expensive and will only be taken when other, higher-ranked prey are unavailable (e.g. Broughton 1994; Cohen 1977; Hawkes & O'Connell 1992; Layton et al. 1991; Moore & Hillman 1992; O'Connell & Hawkes 1981; Wright 1994). We favour this argument in principle because of its simplicity and potential generality, and because it recognizes the importance of resource cost.
Smith's (1988; 1989a; 1993) model of prehistoric arid zone resource use is generally consistent with this hypothesis. Its key points are that seeds were low ranked, entering the diet only after a long period of human population growth (marked by increased rates of artefact discard, larger numbers of sites, and more widespread occupation) that reduced the availability of preferred foods. The slight trend toward cooler temperatures and increased aridity after 6000 b.p. is said to have accelerated the process.
Despite its theoretical appeal and apparent fit with available data, there are two important problems with Smith's argument. First, the actual evidence for mid-Holocene population growth in the arid zone is limited and equivocal. Apparent increases in site numbers and deposition rates may prove illusory once larger samples become available, and, in any case, are subject to alternate interpretations. At this stage, and despite the arguments for similar phenomena elsewhere on the continent (e.g. Beaton 1991), a pattern of very slow, long-term net population growth from initial occupation through the early Holocene, followed by sharp acceleration in the mid Holocene seems unlikely. In general, we expect just the opposite-rapid initial growth followed by a slow-down (perhaps even a decline in population density) coincident with the depletion of high-ranked prey and the subsequent shift to more expensive resources (Hawkes & O'Connell 1992; see also Porch & Allen, this volume). On this logic, human populations in the arid zone should have reached the point at which seeds became the economically optimal choice within a few millennia after initial occupation, probably before the Last Glacial Maximum, and certainly well before the mid Holocene.
Even if regional populations did remain small until late in the sequence, a second problem emerges. The most significant changes in continental Australian environments over the past 50,000 years coincide with the peak of the last glaciation. Throughout the arid zone, the availability of food and water were sharply reduced. Some areas were entirely abandoned by humans (O'Connor et al. 1993); others may have been continuously occupied, though by very much smaller groups (Smith 1989a; Veth 1993). If these changes in settlement pattern were driven in part by food shortages, and if seeds were available, then they should have been exploited in continuously inhabited areas from the time climate first began to become more arid (c. 25,000 b.p.). The available archaeological evidence does not support this proposition.
The absence of evidence might be explained in several ways:
* The arid zone was in fact completely abandoned during this period. This possibility has been proposed by Peter Hiscock (1988), who argues that evidence for continuous occupation at proposed refuge sites, e.g. Mount Newman (Brown 1987) and Puritjarra (Smith 1987), is equivocal due to stratigraphic mixing and inadequate chronological control. Smith (pers. comm.) maintains that more recent work at Puritjarra resolves these problems, and that essentially continuous occupation through the Glacial Maximum is indeed indicated. Since pertinent data remain unpublished, it is impossible to assess this proposition independently.
* As Smith suggests, Last Glacial Maximum populations in refuge areas were so low and so highly mobile that seeds never became the optimal choice. This seems unlikely. Climatic reconstructions strongly suggest that sources of food and water were at least as limited and unreliable at that time as in any part of the arid zone known historically. Seasonal temperature contrasts were evidently much greater. In ethnographic times, seeds were probably most important in the dry season. They should have been even more important during the Glacial Maximum if higher-ranked resources were indeed less abundant, as every environmental indicator suggests they were.
* Seeds were part of Last Glacial Maximum diets but the archaeological sample is too small to show it (Gorecki et al., in prep.). Despite the importance of seeds in ethnographic economies, recent archaeological evidence of their use is surprisingly rare: seed-grinding tools make up only ~0.4% of total assemblage content in arid zone sites of late Holocene age (e.g. Gould 1978: 108-11; Smith 1988: 99-132; Veth 1993: 51-8). Archaeological samples of Last Glacial Maximum age are few in number and small in size, perhaps too small to support the conclusion that seed-grinding tools are unrepresented. Data
(1) With the assistance of Melissa Heck, we recalculated the range of return rates reported by Cane (1987: 402-4, 430). Resting and eating times were eliminated from the equation. While Cane presents a range of return rates of 246-810 kcal/hr for grasses, our calculations show the upper end of the range closer to 1100 kcal/hr (1106 kcal/hr for Panicum australiense). The median return rate from all available data on arid Australian tree and grass seeds remains at or below 700 kcal/hr. from Puritjarra (Smith 1988: 99-132) provide the basis for a test. Located near a permanent spring in the Central Desert Ranges, this site should have been used in dry seasons if the area were occupied during the Glacial Maximum; and, if it were, seeds should have been part of the diet of its occupants. Although grinding tools make up the expected ~0.4% of the late Holocene assemblage at this site, they are absent from the glacial-age deposits. Nevertheless, Chi-square analysis indicates their absence could be a product of sample size. In other words, the available archaeological data are not sufficient to rule out the possibility that seeds were exploited during the Last Glacial Maximum.
* Seeds were token during the Last Glacial Maximum but the evidence is unrecognized or unrecognizable. Smith's (1985; 1986) proposed criterion for prehistoric seed exploitation is the presence of seed-grinding tools similar to those used ethnographically. As indicated above, these have so far been found only in deposits of mid-Holocene age or younger. If seeds were used prior to that time, they may have been processed with different implements. Amorphous, grinding tools are occasionally reported from Pleistocene and early Holocene deposits in various parts of the continent, including sites within or adjacent to the arid zone (Gorecki et al. in prep.; Smith 1985). So far, it has been difficult to determine whether they were used as seed-grinders. The best candidates may, be those from ~30,000-year-old deposits at Cuddie Springs, in western New South Wales (Furby et al. 1993). These artefacts resemble the shallow grinding slabs commonly encountered at many sites in arid western North America, except that their grinding surfaces are smaller. It is also possible that wooden tools were used in seed processing (a common pattern elsewhere in the world), although the evidence for severely reduced tree cover over much of the arid zone makes this seem unlikely. Appeal to a very different technology for seed processing in Pleistocene times would also require an explanation for the late change to stone seedgrinders.
On present evidence, the question of seed use during the Last Glacial Maximum cannot be resolved. If the arid zone were occupied at that time, then, on the logic of Smith's own argument, seeds must have been part of the diet. In fact, they might well have been used even earlier. If further research shows that they were, then the depletion, hypothesis is strongly supported. If not, that hypothesis might still be supported if it were shown that the arid zone was abandoned during the Last Glacial Maximum, then reoccupied (after c. 16,000 b.p.) by populations that subsequently adopted seed use when local densities reached some critical threshold. But if this does not occur until the mid Holocene, then proponents of this argument need to tell why the build-up took so long.
The third general explanation appeals to social process as the catalyst for the broad spectrum revolution. It is grounded in the proposition that various forms of exchange between individuals and groups foster increased demand for the production of critical resources, including food. This in turn favours technological and organizational innovations that promote production and encourage the development of new or improved resources. Broader exchange networks are created and sustained, economic stability is enhanced; population growth follows. In some instances, the larger volume of resources in circulation creates opportunities for social stratification based on differential control of production. Agriculture and complex societies are among the results (e.g. Bender 1978; Hayden 1981).
This line of argument has been very attractive to Australianists (e.g. Beaton 1977; 1982; 1983; Bowdler 1981; Lourandos 1983; Ross 1985; Williams 1987), primarily because so many important shifts in diet, technology, and settlement pattern throughout the continent seem to date to the mid Holocene, a period of relatively modest environmental change, at least in terrestrial habitats (cf. Beaton 1985). Proponents begin by noting the ethnographic importance of large-scale regional exchange networks continent-wide, contending that such networks facilitated various forms of social and economic interaction, including the exchange of mates and other key resources. Some hold that these networks enabled individuals and families to move widely across the landscape, beyond their normal foraging ranges, either to escape local food shortages or take advantage of especially abundant harvests elsewhere. Others observe that they provided a framework that enabled ambitious individuals to pursue positions of prestige and influence. All recognize that such networks were marked by ritual convocations that drew many participants together over great distances. Crucially in this context, all agree that these convocations were fuelled by locally abundant foods, many of which (e.g. tree and grass seeds, bunya nuts, cycads) are definitive broad spectrum resources. The development of these convocations and the social networks that underlie them required and encouraged the use of these resources. As these systems developed, so the exploitation of resources essential to their maintenance intensified.
There is less agreement about the factors that provoked the appearance of the networks themselves. Some attribute it to the passage of a critical demographic threshold (e.g. Beaton 1983; see also Bender 1985; Hayden 1981; 1990. Others (e.g. Lourandos 1983: 92), explicitly deny any link with population pressure, instead characterizing network development as a natural evolutionary process. Still others (e.g. Bowdler 1981) see it as the product of one or more key innovations in food-processing technology, a proposition that softens the emphasis on social factors. From this latter perspective, the temporal coincidence between initial evidence of these innovations and the first appearance of microliths and the dog tall about 5000 b.p. on a continental scale) is especially significant, implying an exotic (i.e. non-Australian) origin for both the technology and the social networks its presence facilitates.
Veth's (1987; 1989; 1993; Veth et al. 1990) scenario for mid-Holocene changes in arid Australian land use entails a similar line of argument. Specifically, he contends that the mid-Holocene occupation of barrier deserts, was facilitated by: * The development of extended social networks that enabled local groups to offset the uncertainty inherent in use of very low rainfall habitats by allowing them to move from areas of temporary drought to others favoured by recent rainfall. * Intensified exploitation of herb, tree, and grass seeds (particularly those common to the hummock grassland associations typical of barrier deserts) that supported the ceremonies associated with these networks and the occupation of resource poor habitats. * The invention of hafted adzes said to be essential to the manufacture of hardwood bowls used in seed winnowing and hence important to the intensified exploitation of seeds. * Innovations in well construction that facilitated access to the deep aquifers that are often the only sources of water available in barrier deserts.
As indicated above, Veth says that some of these innovations, notably the social networks and seed grinding technology (see also Gorecki et al. in prep.), originated in arid zone refuge areas during the last Glacial Maximum, but did not develop or coalesce in a way that enabled the occupation of barrier deserts until the mid Holocene. Ross et al. (1992) tie Veth's argument to the continent-wide pattern of mid-Holocene intensification, linking changes in desert land use to the operation of 'social forces, elsewhere on the continent, but Veth himself (pers. comm.) sees this connection as unnecessary.
Taken collectively, 'intensification' arguments for the use of broad spectrum resources have two important shortcomings. First, appeal to this line of argument effectively precludes the possibility of accounting for the shift to broad spectrum diets in global terms. Seeds and other high-cost resources entered human diets in many parts of the world at various points in the late Pleistocene and Holocene, and while in some cases their adoption may have been connected with the development of regional exchange networks like those known in ethnographic Australia, such links are far from ubiquitous.
Second, and ultimately more important, none of the arguments that identify social processes as the explanation for mid-Holocene changes in Australian diet and technology tells how and why those social processes come into play when and where they do. Most simply rely on the ethnographic observation that ritual convocations involving representatives of many local groups were frequently supported by the exploitation of abundant, high-cost foods. The correlation is projected back to the earliest date for use of the resources in question, the ritual link assumed, and the social processes behind the ritual given explanatory primacy. In its present form, this line of argument is uncompelling because the social processes themselves are neither explained nor directly monitored archaeologically. Though they may be correlated with population growth, the causal link between demography and social process remains unexplored. The argument for a mid-Holocene population peak, is also subject to the same objections outlined above under the depletion hypothesis.
Veth's version of this argument has its own special problems, most of which have been identified elsewhere (Smith 1993; cf. veth 1994). The key criticisms are these: * Extended social networks are typical of ethnographically known hunter-gatherers everywhere in the world and thus not necessarily tied to the development of complex Australian-style regional ex@ change systems in the manner suggested. Like Smith (1993), we anticipate their presence from the time the arid zone was first occupied. * Exploitation of herb, tree and grass seeds is certainly important @probably essential) to the occupation of barrier deserts, but need not be linked with social intensification. Nor is it clear, contra Veth (e.g. 1989: 90), that seed resources common to hummock grasslands represent any special problems in handling not already encountered among similar (sometimes precisely the same) taxa in other habitats. * Efficient use of seeds anywhere in the arid zone may well necessitate the use of winnowing trays, but, as Smith (1993) points out, these need not be made of hardwoods of the kind generally worked with adzes (e.g. Acacia, Eucalyptus). A simple bark sheet will suffice, as will a bowl made of the soft wood of the bean tree (Erythrina vespertillo) (O'Connell et al. 1983: figure 3; Thomas 1989. figure 15.2). Neither of these requires the use of a specialized adze in its manufacture. * Though Veth maintains that the excavation of deep wells requires special skills, he provides no details on what these might be. Smith contends that whatever expertise was needed probably developed early, coincident with initial use of the arid zone. Our own limited experience leads us to side with Smith, though we observe that, as currently stated, the case is weak either way.
We recognize that social considerations may shape, sometimes largely determine, resource exploitation (e.g. Hawkes et al. 1991), but the argument that they do in any particular case must be made, preferably in archaeologically testable terms, not simply asserted on the basis of an ethnographic correlation. More precisely, one must show in terms of general theory how and why social processes are likely to affect resource choice, how and why these social processes come into play when and where they do - and by extension, why they are not implicated in other, non-Australian contexts. None of the intensification, arguments offered to account for the apparent mid-Holocene expansion in diet breadth anywhere in Australia meets these criteria.
Our discussion of current explanations for the broad spectrum revolution is grounded in the proposition that the revolution, itself is a world-wide phenomenon, and, as such, is likely to be the product of very general processes. Whether this proposition is valid depends in part on how that revolution, is defined. If we take it to mean the regular use of resources that require heavy investments in processing, such as tree and grass seeds, then its broadly coincident onset in many parts of the world seems consistent with the notion that the same general processes are involved in every instance. While its explanation may ultimately prove more complicated, it is best to eliminate this simplest of possibilities first. Arid Australia provides a good venue in which to do so because, on current evidence, the 'revolution' begins there later than anywhere else, making it an obvious challenge to the most general hypotheses, those that appeal to terminal Pleistocene climatic change as a catalyst.
Of the three arguments for the adoption of broad spectrum diets commonly discussed in the world literature, one - climate-driven in creases in the abundance of broad spectrum resources themselves - can be eliminated. Though past changes in the abundance of tree and grass seeds in the arid zone are still poorly controlled, they were unlikely to have had any effect on the initial use of these resources, simply because their collecting and processing costs are so high relative to their nutritional value. The same objection applies to similar arguments for the adoption of broad spectrum resources elsewhere.
The converse of the abundance, argument - that broad spectrum resources were included in local diets whenever higher ranked prey were unavailable or rarely encountered - fares better in some parts of the world than in arid Australia, where these resources were apparently adopted so late, well after any period of major climatic change. The logic of the argument suggests seeds should have been used much earlier, at least as early as the onset of Last Glacial Maximum aridity, perhaps even coincident with initial occupation of the zone. We think it would be difficult for any hunter - gatherer population to have persisted there without regular access to these resources. The archaeological data needed to test this proposition are unavailable. If it is refuted, and the mid-Holocene date for initial use of seeds confirmed, then the depletion of higher-ranked prey may still have been the cause, although the slow rate of regional population growth implied by this timing would itself require explanation.
Social process arguments remain plausible to many, but require a better theoretical warrant and clearer archaeological test. If they are shown to be the cause of a mid-Holocene broad spectrum revolution in Australia, and depletion arguments correspondingly refuted, then the proposition that a single process underlies this revolution world-wide should be rejected.
Acknowledgements. This work was supported by the Department of Anthropology, University of Utah. J. Allen invited us to write the paper; H. Allen, J. Balme, J. Beaton, S. Cane, I. Davidson, R. Fullagar, D. Grayson, K. Hawkes, M. Heck, P Hiscock, D. Metcalfe, M. Smith, P. Veth and P. White provided access to key data, unpublished manuscripts, and comments on previous drafts; J. Graves and L. Hunsaker helped prepare the manuscript.
Allen, H. 1972. Where the crow flies backwards: man and land in the Darling Basin. Unpublished Ph. D dissertation, Department of Prehistory, RSPacS, Australian National University, Canberra. 1974. The Bagundji of the Darling Basin: cereal gathers in an uncertain environment, World Archaeology 5:309-22. 1990. Environmental history in southwestern New South Wales during the late Pleistocene, in C. Gamble & O. Soffer (ed.), The world at 18000 BP 2: Low latitudes: 297-321. London: Unwin Hyman. Balme, J. 1991. The antiquity of grinding stones in semi-arid western New South Wales. Australian Archaeology 32:3-9. 1995. 30,000 years of fishery in western New South Wales, Archaeology in Oceania (in press). Beard, J.S. 1981. Vegetation of central Australia, in J. Jessop (ed.), Flora of central Australia: 21-26. Sydney: Reed. 1984. The vegetation of the Australian arid zone, in H.G. Cogger & E.E. Cameron (ed.), Arid Australia: 113-17. Sydney: Australian: Museum. Beaton, J.M. 1977. Dangerous harvest. Unpublished Ph.D dissertation, Department of Prehistory, RSPacS, Australian National University, Canberra. 1982. Fire and water: aspects of Australian Aboriginal management of cyacads, Archaeology in Oceania 17:59-67. 1983. Comment: does intensification account for changes in the Australian Holocene archaeological record?, Archaeology in Oceania 18:94-7. 1985. Evidence for a coastal occupation time lag at Princess Charlotte Bay (North Queensland) and implications for coastal colonization and population growth theories for Aboriginal Australia, Archaeology in Oceania 20: 1-20. 1991. Colonizing continents: some problems from Australia and the Americas, in T. Dillehay & D. Meltzer (ed.), The first Americans, search and research: 209-30. Boca Raton (FL): CRC Press. Bender, B. 1978. Gatherer-hunter to farmer: a social perspective, World Archaeology 10(2):204-22. Emergent tribal formations in the American midcontinent, American Antiquity 50(1):52-62. Binford, L. 1968, Post-pleistocene adaptations. in I,. R. Binford & S. Binford (ed.), New perspectives in archaeology: 313-41. Chicago (IL): Aldine. Bowrdler. S. 1977. The coastal colonization of Australia, in J. Allen, J. Golson & R. Jones (ed.), Sunda and Sahul: prehistoric studies in Southeast Asia, Melanesia and Australia: 205-46. London: Academic Press. 1981. Hunters in the highlands: aboriginal adaptations in the eastern Australian uplands, Archaeology in Oceania 16:99-111. Bowler, J.M. 1976. Aridity in Australia: age, origins, and expression in aeolian landforms and sediments. Earth-Science Reviews 12:279-310. Bowler, J.M., G.S. Hope, J.N. Jennings, G. Singh &, D. Walker. 1976. Late Quaternary climates of Australia and New Guinea, Quaternary Research 6:359-94. Bowler. J.M., Huang QI, Chen Kezao, M.J. Head & Yuan Baoyin. 1986. Radiocarbon dating of playa-lake hydrologic changes: examples from northwestern China and central Australia, Paleogeography, Paleoclimatology, Poleoecology 54: 241-60. Bowler. J.M. & R.J. Wasson. 1984. Glacial age environments of inland Australia, in Vogel (ed.): 183-208. Broughton, J. 1994. Late Holocene resource intensification in the Sacramento Valley, California: the vertebrate evidence, Journal of Archaeological Science 21:501-14. Brown. S. 1987. Towards a prehistory of the Hamersley Plateau, Northwest Australia. Canberra: Australian National University Press. Occasional Papers in Prehistory 6 Cane. S. 1987. Australian Aborginal subsistence in the Western Desert, Human Ecology 15(4):391-435. 1989. Australian Aboriginal seed grinding and its archaeological record: a case study from the Western Desert, in Harris & Hillman (ed.): 99-119. Chappell, J. 1991. Late Quaternary environmental changes in eastern and central Australia, and their climatic interpretation, Quaternary Science Reviews 10:377-90. Chappell, J.M.A. & A. Grindrod (ed.). 1983. CLIMANZ: a symposium of results and discussions concerned with late Quaternary climatic history of Australia, New Zealand and surrounding seas. Department of Biogeography and Geomorphology. RSPacS, Australian National University, Canberra. Chen, X.Y. 1989. Lake Amadeus, Central Australia: modern processes and evolution. Unpublished Ph.D dissertation, Australian National University, Canberra. Chen, X.Y., J.R. Prescott & J.T. Hutton. 1990. Thermo-luminesence dating on gypseous dunes of Lake Amadeus, central Australia, Australian Journal of Earth Sciences 37:92-103. Cleland, J.B. 1966. The ecology of the Aboriginal in south and central Australia, in B.C. Cotton (ed.), Aboriginal man in south and central Australia: 111-58 . Adelaide: Government Printer. Cleland, J. B. & T.H. Johnston. 1933. The ecology of the Aborigines of central Australia: botanical notes, Transactions of the Royal Society of South Australia 57:113-24 Cleland, J. B. S, N. Tindale. 1959. The native names and uses of plants at Haast Bluff, Central Australia, Transactions of the Royal Society of South Australia 82:123-40. Cohen. M. 1977. The food crisis in prehistory. New Haven (CT): Yale University Press. Davidson, I., S.A. Suttons &, S.J. Gale. 1993. The human occupation of Cuckadoo 1 Rockshelter, northwest central Queensland, in Smith et al. (ed.):164-72. Flannery, K.V. 1969. The origins and ecological effects of early domestication in Iran and the Near East, in P.J. Ucko & G.W. Dimbleby (ed.), The domestication and exploitation of plants and animals: 73-100. London: Duckworth. Furby, J.H., R. Fullegar, J.R. Dodson & I. Prosser. 1993. The Cuddie Springs bone bed revisited, 1991, in Smith et al. (ed.):204-10. Gentilli, J. (ed.). 1971. Climates of Australia and New Zealand. Amsterdam: Elsevier. World Survey of Climatology 13. Gorecki, P., M. Grant, S. O'Connor A P. Veth, 1994. The morphology, function and antiquity of grinding implements in northern Australia. Unpublished manuscript. Gould, R.A. 1969. Subsistence behaviour among the Western Desert Aborigines of Australia, Oceania 34(4):253-74. 1977. Puntutjarpa Rockshelter and the Australian Desert Culture. Anthropological Papers of the American Museum of Natural History 54. 1978. James Range East Rockshelter, Northern Territory, Australia: a summary of the 1973 and 1974 investigations, Asian Perspectives 21(1):85-125. Hamilton, A. 1980. Dual social systems: technology, labour and women's secret rites in the Eastern Western Desert of Australia, Oceania 51:4-19. Harris D.R., & G.C. Hillman(ed.). 1989. Foraging and farming: the evolution of plant exploitation. London: Unwin Hyman. Harrison, S.P. 1993. Late Quaternary lake-level changes and climates of Australia, Quaternary Science Reviews 12:211-31. Harrison, S.P. & J. Dodson. 1993. Climates of Australia and New Guinea since 18,000 BP, in H.E. Wright, Jr. et al. (ed.), Global climates since the Last Glacial Maximum: 265-93. Minneapolis (MN): University of Minnesota Press. Harrison, S.P., S.E. Metcalfe, F.A. Street-Perrott, A.R. Pittock, C.N. Roberts & M.J. Salingero. 1984. A climatic model of the last glacial/interglacial transition based on paleotemperature and paleohydrological evidence, in Vogel (ed.): 21-34. Hawkes, K. &, J.F. O'Connell. 1992. On optimal foraging models and subsistence transitions, Current Anthropology 33(1):63-6. Hawkes, K., J.F. O'Connell a N. Blurton Jones. 1991. Hunting income patterns among the Hadza: big game, common goods, foraging goals, and the evolution of the human diet, Philosophical Transactions of the Royal Society, London B 334;243-51. Hayden, B. 1981. Research and development in the Stone Age: technological transitions among hunter/gatherers, Current Anthropology 22:519-48. 1990. Nimrods, piscators, pluckers and planters: the emergence of food production, Journal of Anthropological Archaeology 9:31-69. Henry, D. 1989. From foraging to agriculture: the Levant at the end of the last Ice Age. Philadelphia (PA) University of Pennsylvania Press. Hiscock. P. 1988. Prehistoric settlement patterns and artefact manufacture at Lawn Hill, Nortbwest Queensland. Unpublished Ph.D dissertation, Department of Anthropology, University of Queensland. Hope, J. 1993. Pleistocene archaeological sites in the central Murray-Darling Basin, in Smith et al. (ed.):183-96. Horton, D.R. 1981. Water and woodland: the peopling of Australia, Australian Institute of Aboriginal Studies Newsletter 16:21-7. Jacodson, G., A.V. Arakel & C. Vijian. 1988. The central Australian groundwater discharge zone: evolution of associated calcrete and gypcrete deposits, Australian Journal of Earth Sciences 35:549-66. Jacobson, G., G.E. Calf, J. Jankowski & P.S. McDonals. 1989. Groundwater chemistry and palaeocharge in the Amadeus Basin, central Australia, Journal of Hydrology 109:237-66. Jennings, J.N. 1975. Desert dunes and estuarine fill in the Fitzroy Estuary, northwestern Australia, Cateno 2:215-62. Juell, K.E. 1990. Groundstone utilization, in R.G. Elston H, E.E. Budy (ed.0, The archaeology of James Creek Shelter, 247-55. Salt Lake City (UT): University of Utah. Anthropological Papers 115. Kimber, R.G. 1984. Resource use and management in central Australia. Australian Aboriginal Studies 2: 12-23. Lampert, R.J. & P.J. Hughes. 1987. The Flinders Ranges: a Pleistocene outpost in the arid zone?, Records of the South Australia Museum 20: 29-34. 1988. Early human occupation of the Flinders Ranges, Records of the South Australian Museum 22(2):139-68. Latz, P.K. 1982. Bushfires and bushtucker: Aborigines and plants in central Australia, Unpublished MA thesis, Department of Prehistory, University of New England, Armidale. Layton, R., R. Foley A E. Williams. 1991. The transition between hunting and gathering and the specialized husbandry of resources, Current Anthropology 32:255-74. Long, J.P.M. 1971. Arid region Aborigines: the Pintubi, in Mulvaney &, Golson (ed.): 262-70. Lourandos, H. 1983. Intensification: a late Pleistocene-Holocene archaeological sequence from southwest Victoria, Archaeology in Oceania 18:81-94. 1985. Intensification and Australian prehistory, in T.D. Price & J.A. Brown (ed.), Prehistoric hunter-gatherers: the emergence of cultural complexity: 385-426. New York (NY) Academic Press. McCorriston, J. S, F. Hole. 1991. The ecology of seasonal stress and the origins of agriculture in the Near East, American Anthropologist 93(1):46-69. Mabbutt, J.A. 1971. The Australian arid zone as a prehistoric environment, in Mulvaney &, Golson (ed.):66-79. Martin, H.A. 1973. Palynology and historical ecology of some cave excavations in the Australian Nullarbor, Australian Journal of Botany 21: 283-16. Maynard, L. 1980. A Pleistocene date from an occupation deposit in the Pilbara region, Western Australia, Australian Archaeology 10: 3-8. Meggitt, M.J. 1957. Notes on the vegetable foods of the Walbiri of central Australia, Oceania 28:143-5. Moore, A.M.T. A G.C. Hillman. 1992. The Pleistocene to Holocene transition: the impact of the Younger Dryas, American Antiquity 57:482-94. Morse, K. 1988. Mandu Mandu Creek rockshelter: Pleistocene human coastal occupation of North West Cape, Western Australia, Archaeology in Oceania 23:81-8. Mulvery, D.J. & I Golson (ed.). 1971. Aboriginal man and environment in Australia. Canberra: Australian National University Press. Nix, H.A. A J.D. Kalma. 1972. Climate as a dominant control in the biogeography of northern Australia and New Guinea, in D. Walker (ed.), Bridge and barrier: the natural and cultural history of Torres Strait: 69-92 Canberra: Australian National University. O'Connell, J.F. S, K. Hawkes. 1981. Alyawara plant use and optimal foraging theory, in B. Winterhalder a E. Smith (ed.), Hunter-gatherer foraging stategies: ethnographic and archaeological analyses:99-125. Chicago (IL): University of Chicago Press. O'Connell, J.F., P.K. Latz S, P. Barnett. 1983. Traditional and modern plant use among the Alyawara of central Australia, Economic Botany 37:80-109. O'Connor, S., P. Veth &, N. Hubbard. 1993. Changing interpretations of postglacial human subsistence and demography in Sahul, in Smith et al. (ed.): 95-105. Patton, P.C., G. Pickup & D.M. Price. 1993. Holocene paleofloods of the Ross River, central Australia, Quaternary Research 40:201-12. Peterson, N. 1977. Aboriginal uses of Australian Solonaceae, in J.G. Hawkes, R.N. Lester a A.D. Skedling (ed.), The biology and taxonomy of the Solonaceae: 171-88. London: Linnean Society. Symposium Series 7. 1978. The traditional pattern of subsistence to 1974, in B.S. Hetzel &, H.J. Frith (ed.), The nutrition of Aborigines: 25-35. Melbournm: CSTRO. Prentice, I. C., J. Guiot & S.P. Harrison. 1992. Meditettanean vegetation. lake levels add paleoclimate at the Last Glacial Maximum. Nature 360: 658-60. Ross, A. 1985. Archaeological evidence for population change in the middle to late Holocene in southeastern Australia, Archaeology in Oceania 20:81-9. Ross, A., T. Donnelly & R. Wasson. 1992. The peopling of the arid zone: human-environment interactions, in J. Dodson (ed.), The naive lands: prehistory and environmental change in Australia and the southwest Pacific: 76-114. Melbourne: Longman Cheshire. Silberbauer, G. 1971. Ecology of the Ernabella Aboriginal community, Anthropological Forum 3:21-36. Singh, G. S, J. Luly. 1991. Changes in vegetation and seasonal climate since the last full glacial at Lake Frome, South Australia, Paleogeography, Polaeoclimatology, Palaeoecology 84: 75-86. Smith, M.A. 1985. A morphological comparison of central Australian seedgrinding implements and Australian Pleistocene-age grindstones, The Beagle, 23-38. 1986. The antiquity of seedgrinding in arid Australia, Archaeology in Oceania 21:29-39. 1987. Pleistocene occupation in arid Central Australia. Nature 328:710-11. 1988. The pattern and timing of prehistoric settlement in central Australia. Unpublished Ph.d dissertation, Department of Archaeology and Paleoanthropology, University of New England, Armidale. 1989a. The case for a resident human population in the Central Australian Ranges during full glacial aridity, Archaeology in Oceania 24:93-105. 1889b. Seed gathering in inland Australia: current evidence from seed-grinders on the antiquity of the ethnohistorical pattern of exploitation, in Harris & Hillman (ed.): 305-17. 1993. Biogeography, human ecology and prehistory in the Sandridge Deserts, Australian Archaeology 37:35-49. Smith, M.A. &, P. Clarke. 1993. Radiocarbon dates for prehistoric occupation of the Simpson desert, Records of the South Australian Museum 26:121-27. Smith, M.A., M. Spriggs & B. Fankhauset (ed.). 1993. Sahul in review: Pleistocene archaeology in Australia, New Guinea, and island Melanesia. Canberra: Australian National University Press. Smith, M.A., E. Williams &, R. J. Wasson. 1991. The archaeology of the JSN site: some implications for the dynamics of human occupation in the Strzelecki Desert during the late Pleistocene, Records of the South Australian Museum 25:175-92. Spencer, B. & F.J. Gillen. 1912. Across Australia. London: MacMillan. Stephens, D.W. & J.R. Krebs. (1986). Foraging theory. Princeton [NJ]: Princeton University Press. Strehlow, T.G.H. 1965. Culture, social structure and environment in Aboriginal Central Australia, in R.M. Berndt A C.M. Berndt (ed.), Aboriginal man in Australia: essays in honour of emeritus professor A.P. Elkin: 121-45. Sydney: Angus and Robertson. Sweeney, G. 1947. Food supplies of a desert tribe, Oceania 17: 289-99. Thomas, D.H. 1989. Archaeology. @nd edition. Fort Worth (TX) Holt, Rinehart and Winston. Thomson, L. 1992. Australia,s subtropical dry zone Acacia species with human food potential, in A.P.N. House & C.E. Harwood (ed.), Australian dry zone acacias for human food: 3-36. Canberra: Australian Tree Seed Centre. Tindale, N.B. 1959. Ecology of primitive aboriginal man in Australia, in A. Keast, R.L. Crocker &, C.S. Christian (ed.), Biogeography and ecology in Australia: 36-51. The Hague: W. Junk. 1972. The Pitjandjara, in M.G. Bicchieri (ed.), Hunter gatherers today: 217-68. New York (NY): Holt, Rinehart and Winston. 1977. Adaptive significance of the Panara or grass seed culture, of Australia. in R.V.S. Wright (ed.), Stone tools as cultural markers: 345-49. Canberra: Australian Institute of Aboriginal Studies. Veth, P.M. 1987. Martu prehistory: variation in arid zone adaptations, Australian Archaeilogy 25: 102-11. 1989. Islands in the interior: a model for the colonization of Australia's arid zone, Archaeology in Oceania 24:81-92. 1993. Islands in the interior: the dynamics of prehistoric adaptations within the arid zone of Australia. Ann Arbor (MI) International Monographs in Prehistory. 1994. Marginal returns and fringe benefits: characterizing the prehistory of the lowland deserts of Australia (a reply to Smith). Unpublished manuscript. Veth, P., G. Hamm &, R.J. Lampert. 1990. The archaeological significance of the lower Cooper Creek, Records of the South Australian Museum 24:43-66. Veth, P. &, F.J. Walsh. 1988. The concept of `staple' plant foods in the Western Desert region of Western Australia, Austrulian Aboriginal Studies 2:19-25. Vogel, J.C. (ed.). 1984. Late Cainozoic palaeoclimates of the Southern Hemisphere. Rotterdam: A.A. Balkema. Walsh, F.J. 1987. The influence of the spatial and temporal distribution of plant food resources on traditional Martujarra subsistence strategies, Australian Archaeology 25:88-01. Wasson, R.J. 1983. The Cainozoic history of the Strzelecki and Simpson dunefields (Australia), and the origin of the desert dunes, Zeitschrift fur Geomorphologie, N.F. Supplement 45:85-115. 1984. Late Quaternary palaeoenvironments in the desert dunefields of Australia, in Vogel (ed.419-32. 1989. Desert dune building, dust raising and palaeoclimate in the Southern hemisphere during the last 280,000 years, in T. Donnelly SR. Wasson (ed.), Late Quaternary climatic history of Australasia: 1978. Canberra: CSIRO. Webster, P.J. & N.A. Streten. 1978. Late Quaternary ice age climates of tropical Australasia: interpretations and reconstructions, Quaternary Research 10:279-309. Williams, E. 1987. Complex hunter-gatherers: a view from Australia, Antiquity 61:310-21. 1988. The archaeology of the Cooper Basin: report on fieldwork, Records of the South Australian Museum 22(1) 53-62. Wright, K. 1994. Ground-stone tools and hunter-gatherer subsistence in Southwest Asia: implications for the transition to farming, American Antiquity 52(2): 238-63. Wright, R.V.S. (ed.). 1971. Archaeology of the Gallus site, Koonalda Cave. Canberra: Australian Institute of Aboriginal Studies.
(1) Distribution of seed-grinding tools at Puritjarra: Level I (Holocene-age deposits) ground-stone artefacts=15, other artefacts=4445; Level II (Pleistocene-age deposits) = O, 593, respectively. Chi square - 1.07; not significant @ p=0.05. (2) Other arid zone data sets of Pleistocene age that might be expected to include seed grinding tools are those from various locations along major drainages in western New South Wales (e.g. Allen 1972) and from Lawn Hill, north Queensland (Hiscock 1988). When first described, the former were said to contain seed-grinders, but both the age and function of the objects in question are now disputed (Allen 1990; Balme 1991; Hope 1993; Smith 1985; 1986; 1989b) and can and can probably be established only through further fieldwork and functional analysis. No grinding tools have been identified in the Pleistocene-age samples from Lawn Hill, but because the assemblage is small (total artefacts from Louie and Colless Creek sites = 1197), their absence could be attributed to sample error. (Assuming ground stone tools were deposited in the same proportions as at arid zone sites of late Holocene age. Chi square 3.24, not significant d p=0.05). (3) The grinding surface on the most complete fragment reported by Furby et al. (1993) has an estimated diameter of cm. Similar surfaces on grinding slabs from the Great Basin are generally in excess of 15 cm in diameter (e.g. Juell 1990: table 85).…