Zubrow, E.B.W. Up on the Prehistoric Farm. Review of Rindos, David. The Origins of Agriculture: An Evolutionary Perspective. Orlando, Florida: Academic Press, 1984. xvii + 324 pp. including bibliography and index. $32.50 cloth. [In Reviews in Anthropology 13(3) 1986: 210-222.]

My ten-year-old son thinks that book reports are dull, boring and as he says "for the birds." He has thought so since he did his first one in third grade. When I told him that I was writing a grown-up book report, preempting the personal computer and preventing him from playing "Wizardry" or "Chivalry," he was sufficiently outraged to comment, "Couldn't you write about gnomes and trolls and other important things?" Those would be the only topics that could make such a task as interesting as Mr. Kaiser's fifth grade creative writing assignments. I hope that my son's tastes will change as he grows older, because David Rindos's Origins of Agriculture is not about trolls or gnomes, but for some of us it surely is as fascinating.

The origin of agriculture has been a subject which has attracted archaeologists, botanists, and various members of the historical and agricultural sciences for the best part of a century. Their interest has been neither casual nor precipitous, for the transition from hunting and gathering to agriculture was one of the most profound and far-reaching in human history. The absolute magnitude and the relative recency of this change may be sketched. Of the approximately 80 billion people who have lived since the emergence of our species, Homo sapiens, over 90 percent have lived as hunters and gatherers. For more than 99 percent of the duration of human occupation (broadly defined since the time of Homo habilis), people made no use of domesticated plants and animals. Agriculture appears for the first time in archaeological contexts about ten millennia ago. Only 6 percent of this more broadly defined human population ever living has been agricultural, and of these only 3 percent have had the benefit of industrial agriculture. Put somewhat more baldly, most people, most of the time have had no conception of agriculture, its benefits, and its costs.

There has been a tide in the flow of scientific discoveries concerning early agriculture. It has risen and now is ebbing. At best the advance was one of "fits and starts.'' The locus of discovery moved from the Middle East, through Europe, to both continents of the New World, touching Africa and Southeastern Asia in passing. The disciplines of plant genetics, chemistry, geology. paleobotany, ecology, and systems theory have all made theoretical and substantive contributions. New ideas are being developed still. However, as in many fields, the developments have been surpassed by the phenomenal growth in literature. When I recently did a bibliographic search of Cambridge University's libraries on the origin of domestication and agriculture I was taken aback by the discovery of over a 1000 sources published since 1978. There are articles and monographs in all of the expected places as well as in the unexpected. I will emphasize the English speaking or, more accurately, the English publishing world. The archaeological journals such as American Antiquity (Dean et al. 1985; Farnsworth 1985), Antiquity (Higgs and Jarmon 1969; Guilaine 1979), and World Archaeology (Bender 1978; Dennell 1979) are full of articles on the origin or the implications of the discovery of agriculture. Similar articles addressed to a wider audience are to be found in the American Anthropologist (Wilson 1981), Anthropology (Foster 1985) and Current Anthropology (Rindos 1980; Peters and O'Brien 1981), as well as in such reviews of the field as the Annual Review of Anthropology (Flannery 1973) and Advances in Archaeological Method and theory (Dunnell 1980; Green 1980). The more factual discoveries are reported frequently in Science (Flannery 1965; Harlan 1971. Cavalli-Sforza et al. 1982) or Nature (Hansen and Renfrew 1978). Biological journals Such as Economic Botany (Galinat 1974), American Journal of Botany (Iltis and Doebley 1980), Annual Review of Ecology and Systematics (Heiser 1979), and Agro-Ecosystems (Gliessman et al. 1981) often publish works regarding the origin, the locus. or the dissemination of agriculture. Less accessible to the average student might be the Texas Agricultural Experimental Bulletin, which as early as 1939 included an article by Mangelsdorf and Reeves on the "Origin of Indian Corn and Its Relatives," the Israeli Journal of Earth Science, Progressive Geography (Grigg 1976), or the Cold Spring Harbor Symposia on Quantitative Biology (Braidwood and Reed 1957). Internationally, one finds important works in all the major scientific languages, including German, Russian, Spanish, and French as well as Hebrew and Arabic. There are major studies in Chinese and Japanese as well. For the latter, Koyama Shuzo (1976) not only provides an extensive Japanese bibliography but also has a succinct summary of the relevant data in English. ''Postglacial Foraging and Early Farming Communities in Japan and Korea: A West European Perspective" by Peter Rowley-Conway (1984) provides a more recent review, but is limited to English literature.

Rindos's Origins attempts to explain in biological and cultural terms how and why agriculture began. It is based upon a theoretical concept of "coevolution" and is both processual and behavioral. He delineates three types of behavioral processes which endogenously transformed one to another. These are "incidental-opportunistic," "specialized-obligate," and "agricultural-selective." In order to understand the importance of this book, one must know what has gone before, what is known about the origins of agriculture, and in which directions research has developed. I will divide the contributions into two broad categories -- substantive discoveries, and theoretical explanations. In the former, I will briefly comment upon the earliest evidence for agriculture, regional analyses, diffusionary studies, different forms of agriculture, generalizations, and recent problems. In the latter, I touch upon the four traditional models, discuss the relationship to MalthusianBoserupian controversy, examine the relationship to stable and unstable growth, and point out the importance of new studies in foraging theory. It should be recognized that this substance/theory dichotomy is a heuristic division, for the borderlands are fuzzy. Empirical works blend into theoretical explanations and vice versa.

Substantively, the archaeological "Holy Grail'' has been to discover the earliest agriculture. One genre of literature is the chronicles reporting the various crusading attempts to find the earliest domesticated plants. Kathleen Kenyon (1957, 1979) at Jericho and Robert Braidwood (1975) at Ksar Akil and Jarmo excavated intensively for the earliest domesticated plants. They made remarkable discoveries. Others surveyed extensively in search of early agricultural remains in cave shelters in Mexico and Peru (MacNeish 1964, 1967). Hole et al. (1969) used a similar, survey strategy adapted to examining the Del Luran plain. This search for early cultigens began in the 1950s and often relied upon the National Science Foundation for support. It grew throughout the 1960s and 1970s, but appears to be diminishing in the 1980s. I do not know why this change is occurring, but it is part of the ebbing tide. One reason might concern what has actually been discovered. Once having found the earliest domesticated form of corn, wheat, or any other cultigen, has one found the earliest agriculture? Whether early cultigens and the origin of agriculture are equivalent has not been a research priority, and neither money nor effort has been expended trying to find the answer. As will be seen, Rindos's book begins to answer this question and the answer appears to be negative.

The regional substantive studies tracing the origin of agriculture in different geographic and environmental zones have been less appealing to the general public than finding the actual earliest agricultural remains worldwide. Yet, they have been at least as academically relevant and empirically important. Generally, they have been concerned with the alternative processes of diffusion and independent invention. Typical of research in this area is the British Academy's major project on the early history of agriculture. begun by Higgs and published in three volumes. The third volume, Early European Agriculture edited by M. R. Jarmon ( 1982), was the culmination of ten years' work and could be considered the exemplar of the type. I shall briefly recapitulate some of that project's conclusions, as they are exemplary of the results of this type of regional research, and they are so different in style and substance from Rindos.

Rindos's conclusions concern the relationships and relative importance of such processes as population pressure and feeding pressure. Regional studies emphasize topographic and environmental variables. From a European perspective the development of agricultural subsistence economies was the result of three primary factors -- the size of population. the availability of resources. and the technology of exploitation. Research ranking these factors makes it possible to discern which topographic areas exhibit high potential for early agricultural development. One may divide Europe topographically into the highlands, lowlands, and coastal areas. The highlands are shown to be impossible for early agriculture, the coastal areas' potential overemphasized by earlier scholars, and the lowlands subdivided into areas of high and low arability for further analysis. In the areas of high arable potential there is rapid growth of agriculture and such cultures and sites as "Proto-Sesklo," "Star cevo," "Koros," and "Criv" figure dramatically. This rise of agriculture is synchronous with, if not causal to, the shifting patterns in settlement location. The increasing spatial restriction of neolithic occupation is significant when compared to the preceding Paleolithic and Mesolithic periods.

The role of two variables should not be underestimated. First. the developing relationship between crops and domestic animals needs to be taken into account. Second, there is a critical lack of drinking water on the otherwise rich loess lands. This level of specificity is greater than that used by Rindos, who tries to be global in his relationships rather than regional.

Diffusionary studies have been important on both sides of the Atlantic as well as in Southeast Asia and Australia. There has been considerable interest in the post-origin spread of agriculture and its implications for agricultural evolution and development. The issues have been more clearly crystallized in Europe, where the spread of the earliest agriculture has been the center of intellectual controversy for more than fifty years. In the New World and Asia where the chronological and diffusionary patterns have not been as well developed the problems have masked the underlying processes. Ammerman and Cavalli-Sforza's classic study in Man (1971) followed the existing view of prehistoric agriculture. They showed that the northern spread of agriculture through the great river valleys of Europe took place at a rate of approximately one kilometer a year. The accepted doctrine was that the earliest agriculture was very generalized with either a "slash and burn" or ''long fallow" field rotation pattern. These systems supposedly substituted land extensivity for intensivity. They are more appropriate for colonization than the presumably later, more labor-oriented systems. Concomitantly, the population movement associated with the spread of agriculture had been thought to be "wavelike." This aggregate wave may be composed of numerous individual random walks of the type often modeled by Markov chains or Brownian motion models.

According to such scholars as Jarmon (1982), the evidence does not substantiate the above characterization of these earliest agricultural systems and has to be inverted. Namely, the earliest European systems were short fallow. They were specialized in the sense that specific and precise environmental conditions were recognized and sought. The result was a patchy rather than wavelike settlement pattern. In short, their evidence pointed to the neolithic "invasion'' as being not an invasion but a type of "neolithic opportunism.'' This was believed to be a process of selective colonization and expansion under population pressure which still continues today What was true for Europe was also thought to be true for the other continental traditions. Of course, each was thought to have its own particular developmental sequence.

Ammerman and Cavalli-Sforza published in 1984 their major study, The Neolithic Transition and the Genetics of Populations in Europe, which culminated more than a decade of research. In this volume they recalculated the rate of expansion of agriculture through northern Europe. Both they and Rindos use a demic model. Ammerman and Cavalli-Sforza use it as a model of expansion which posits the spread of farming correlating with local population growth which results in expansion of neolithic populations in short steps. They deemphasize innovation or stimulus diffusion and restate their belief in the 'wave of advance" model originally proposed by R. A. Fisher (1937). This model predicts the movement of the "wave front" at a constant radial rate based upon population growth and local migratory activity. Ammerman and Cavalli-Sforza tested in Calabria, Italy and elsewhere for rates of migration and growth and found that the observed rate of advance is compatible with the expectation based upon the wave of advance model. A simulation based upon the Bandkeramic settlement pattern showed that it took only small amounts of migratory behavior to create the wave of advance. Their map of the present-day distribution of 39 independent genes shows that the neolithic transition is the backbone of the geographic distribution of genes across Europe. The spread of agriculture is a good predictor of the location of modern human genes across Europe. The correlation coefficient between the values of the first principal component of the genetic distribution and the dates of arrival of early agriculture turns out to be very high i.e., r = .89. In other words there is clearly a functional relationship between the two maps. Subdividing the genetic groups in a variety of ways does not change the results significantly. I must say that I find the evidence both persuasive and significant.

Archaeologists have extended their substantive work by examining different social systems associated with different forms and periods of agriculture. Not only are the now traditional skills of the paleobotanists and prehistoric archaeologists used, but the more rarely seen skills of the archaeological linguist and translator are making significant contributions as well. Typical of the genre is a slight volume edited by Trevor Rowley (1981), The Origins of Open Field Agriculture. It contains an article entitled ''Archaeology and Origins of Open Field Agriculture'' (Taylor 1981) in which the advantages and problems of applying prehistoric archaeological techniques to historical agriculture are explicitly addressed. Similarly, Maria de J. Ellis explores new territory in "Agriculture and the State in Ancient Mesopotamia" (1976). This study was one of a number of textual analyses whose purpose was to explore the social nature of early agriculture. She concluded that there were two forms of agricultural production in which "ownership of land and crop" by the state is contrasted to the "usufruct right to use." This type of study brings a new social consciousness to the archaeology of agricultural studies. It raises issues which must be scrutinized by archaeologists working in preliterate areas and times. A similar literate theme was analyzed in Roman Palestine by scholars at Bar llan University using rabbinic sources. The pattern of land ownership changed but not the units of farming. In short, this trend of increasing interest in historic and classical studies of agriculture is part of the trend of increasing interest in the social rather than material aspects of the origin of agriculture.

As early as 1977 Charles A. Reed in the Origins of Agriculture summarized what was known in twenty-one conclusions, which can be compressed into the following eleven propositions:

1. Cultivated crops were being used for food prior to the cultivation process and, thus, sufficient botanical knowledge preexisted agriculture for the neolithic transition to occur.

2. The postglacial periods produced agriculture in several areas on several continents. The Near Eastern and Central Chinese agricultural centers, as defined by Harlan (1971, 1975), are characterized by limited domesticated forms, places, and times; and they contrast with the two agricultural non-centers in sub-Saharan Africa and South East Asia with variable forms, cultigens, and times of ascendancy. The New World centers and non-centers of Mesoamerica and Peru do not conform to the model as well as their Old World counterparts.

3. The domestication of animals was not the result of biological changes per se. but of social attitudes towards animals.

4. There exists no documented evidence for diffusion from a unicentral origin of agriculture, nor for home gardening preceding agriculture

5. Ethnographic evidence points to hunters and gatherers expending less per capita energy per unit of time than do agriculturalists for a successful adaptation. Their low-density populations are maintained by widely spaced births, frequent infanticide, and intensified death rates.

6. People did not seek agriculture as much as they were forced into it by population pressure which increased due to rising sea levels and decreased coastal land.

7. Sedentism is dependent upon size and scale of available resources and not upon the development of agriculture, as is demonstrated by settlements in the Middle East and Peru. It is correlated with increased population growth and decreased age for first pregnancy.

8. Emphasis was on the storage of food.

9. Agricultural innovation and migration reduced population pressure.

10. The shift to agriculture was not sudden, b instead consisted of societies relying on small amounts of cultivation within a mixed economy which gradually shifted to greater dependence on agriculture. Agricultural innovation and migration reduced population pressure.

11. Only marginal populations can afford to drop agriculture, for having once adapted to it, they are trapped by rising expectations and needs of population growth.

None of these propositions has been contradicted since their original statement.

According to the paleobotanist Jane Renfrew (1985) there have been no critical discoveries within the last five years that have amended our understanding of the development of agriculture. However, the changes are coming. First, and foremost, she believes that the origin of agriculture is more independent of location than was previously thought. Numerous centers of development are expected and are being found. The reality is crop diversity in many localized environments and temporal uniformity. Multiple synchronic inventions with many failures for every success were the standard. Second, the principles of agriculture (land and seed preparation, crop and seed storage, planting, harvesting, etc.) as well as agricultural techniques (digging stick and hoe design, the use of the plow, calendrical and weather information, etc.) closely followed the trade routes for obsidian. Third, the widely accepted view that agriculture developed in areas where the modern descendants grow is now under severe criticism. For example at Franchthi (Hansen and Renfrew 1978), a cave in the Argolid, there is a full sequence from Upper Paleolithic to Neolithic agriculture. According to Jane Renfrew and Jule Hansen, cereals with pulses were collected at a level dated 11,000 B.C. This site lies outside the modern distribution of barley and outside the distribution of ancestral species. It is also earlier.

Theoretically, there have been four major models and a small number of variants prior to Rindos's Origins. They are Childe's (1929) "propinquity" hypothesis, Braidwood's (1957) ''settling in" hypothesis, Binford's ( 1968) "population model," and Flannery's ( 1968) ''systems model." Each has had a developmental role in the study of how agriculture originated. Childe's model emphasized climates Braidwood's knowledge, Binford's population, and Flannery's interaction as the exogenous causes. Today. it is felt that none of these models is adequate. The realization that hunters and gatherers do not live at a minimum subsistence level and that their per capita labor for subsistence is relatively low in comparison to early agriculturalists has Created problems. If hunters and gatherers can sustain themselves with a minimal effort and if they have a fair amount of leisure time, what was the driving force to create the transition to early agriculture? Given relatively higher labor costs, relatively higher rates of failure, and relatively lower production, there is no clear economic incentive for agriculture. Suggested explanations have included a sudden rise in the expected standard of living, greater standardization of crops, and more regular times for harvesting. One can raise the argument to a higher level of theoretical generality. The origin of agriculture is an important case. It allows one to evaluate two competitive theories of development. For the Boserupian optimists it is the "piece de resistance" in two ways. First, they see agriculture as the ultimate technological response to increasing population, and second they see the increasing intensification of agricultural labor as a mechanistic response to decreasing output per capita. For the Malthusian pessimists, it is another example of the "dismal" and "utterly dismal'' laws in operation (Boulding 1959). Namely, agriculture was an economic innovation which reduced the pressure on resources for a while until population growth once again reasserted the pressure. Every agricultural innovation simply delays the ultimate return to the balance between resources and population at the subsistence level. The only long-term effect is that each innovation allows more people to exist at minimal subsistence (Boulding 1959) . Mark Cohen fully explored the implications of both these views in The Food Crisis in Prehistory ( 1977), although it should be renoted in this context that the European experience does not corroborate Boserup's expectation that the development of agriculture has been one of increasing shortening of fallow periods (Jarmon 1982).

The importance of this view was recently reiterated and explored in a lecture entitled ''Long-Term Change in the Economies of Societies" by W. H. Day (1985). Prior to commenting on the long-term change in the economies of societies, he pointed out that the field of economics has been dominated by explanations and questions of short-term economic change, while long-term change has ''gone wanting" for lack of both data and interest. Conversely, most archaeologists have been limited to considering long-term economic change. It is the ''bread and butter'' of their discipline. They have the data without the benefit of an economic theory to explain it. The classical economists such as Malthus and Smith were the exception. They had an interest in long-term change and conceived of the economy as converging eventually toward a stable state. Therefore, their focus was on stable, homeostatic, and equilibrating systems. These economies are characterized by increasing efficiency, increasing resources, and balanced growth. If there is anything which economic history has shown, it is that economies do not converge towards stability. Rather, there are large changes with massive redistribution of resources, as in the case of the neolithic transition. The aggregate effects are the result of a mosaic and result in a non-equilibrating growth similar to a life cycle. Ultimately, modern economic or archaeological theorists must place their faith on endogenous changes rather than exogeneous ones. As we will see, Origins is an example of this faith.

From the anthropological position, this viewpoint is the one which has divided archaeologists from their social anthropological brethren. Since the days of Gordon Childe (1929) archaeologists have viewed the development of agriculture as a revolution and a transition. It was an evolutionarily non-stable, temporally fleeting, and at best momentary system. This perspective contrasts with that of the social anthropologists who have viewed peasant agriculture as a conservative and stable system whose success is secure unless perturbed externally. This occurs most often by more advanced industrialized societies. However, if one looks at the development of agriculture evolutionarily and in terms of economics, it is a classic example of non-stable growth -- a discontinuous step process, not a continuous process.

Our knowledge of the variety of agricultural systems and their predecessors has increased vastly in the last decade. Since the days of Boserup's (1965) classic examination of shifting cultivation which pointed out the progressive shortening of fallow periods and her developmental theory of intensification, there has been a large number of studies -- Ucko and Dimbleby's The Domestication and Exploitation of Plants and Animals (1969), Struever's Prehistoric Agriculture (1971), Grigg's (1974) The Agricultural Systems of the World: An Evolutionary Approach, and Ruthenberg's Farming Systems in the Tropics (1980), to name only four. Similarly, there has been a series of analogous works for hunter and gatherers and their relationship to the development of agricultural systems. Winterhalder and Smith (1981) explicitly are concerned with harvesting nonagricultural crops in their discussions of optimal foraging strategies in ''Hunter-Gatherer Foraging Strategies: Ethnographic and Archaeological Analysis." How closely related this is to agriculture is directly raised in Prehistoric Foraging in a Temperate Forest (1981) by Arthur Keene, who asks whether optimal foraging leads to agriculture and whether these are equivalent adaptive strategies.

As a hunting and gathering group develops agriculture and pastoralism, their interaction with other hunting and gathering groups continues. These intergroup relationships figure heavily in The Archaeology of Frontiers and Boundaries, edited by Green and Perlman (1985). In particular the relationships of foragers, pastoralists, and subsistence farmers as defined through group size, mobility, cultural boundaries, and ecological frontiers are emphasized. More ecologically, this same question has been taken on in a mathematical form in the discussion of "humans as a robber species" by Tivy and O'Hare in Human Impact on the Ecosystem: Conceptual Frameworks in Geography (1982).

The question of whether agriculture was a successful adaptation originally has been questioned. There have been studies which indicated that the pa capita resources per energy expended may have dropped, as did the size and life expectancy of the agriculturalists relative to the hunter-gatherers. The conception of the "hungry and grim life of the hunter and gatherer" evolving into a richer and more ''pleasant life for the peasant'' has been replaced by the "happy hunter and gatherer" and the more "pessimistic peasant." These viewpoints have resulted from studies such as those of Yellen (1977), Lee (1972), Howell (1979), and Chagnon (1983). Even my own most recent work on simulating kinship from life tables shows that if one uses mortuary life tables, there is a drop in size of the kin set (total available relatives) at the beginning of the Neolithic (Zubrow 1985).

David Rindos's book The Origins of Agriculture: An Evolutionary Perspective is an important study, for it is simultaneously a description of personal research, a synthesis of other people's ideas, and an exploration of a school of thought. As one reads this book one has the distinct impression that one is privileged to accompany a subtle and well-read mind as it considers and analyzes some of the more profound questions of evolution and human subsistence. How has our species adapted to our environment? How have plants and people coevolved and who is ''doing it faster"? Who are the domesticators and who are the domesticated -- the plants or we? This is a dense book, by which I mean it is not a superficial study of one or more ideas. Rather it is so full of insights that I found each page to be gratifying, and it probably deserves to be read more than the twice-over coverage that I have presently given it. My own copy has a Jackson Pollock look, with sections of every page underlined in red. blue, and black ink, and with rapidly scribbled notes in the margins.

First, I wish to comment on the book as an entirety It is a compendium of scientific explanation and analysis, not of biological or historical data. Theory and methodology are its strengths. It is divided into six chapters: "Agriculture, Evolution, and Paradigms"; "Darwinism and Culture"; "The Naturalness of the Human-Plant Relationship"; "The Evolution of Domestication"; "Feeding Behavior and Change in Diet"; and "Instability, Cultural Fecundity, and Dispersals."

I agree with Dunnell's foreword. The difference between Rindos's sophisticated biological and ecological concepts of evolution and agriculture and the usual anthropological concepts is substantial. He avoids the traps of environmental determinism and cultural selectionism. The strategy is evolution, and the substance is biological ecology. The problem is the explanation of behavior, and the document is the archaeological record. Rindos believes that domestication is a coevolutionary process between plants and people. Neither domesticated the other, rather both codomesticated each other. Thus, domestication cannot be understood as a discovery, or an innovation. The process is "mediated by morphological and autoecological adaptations in the plant and by behavioral changes in man." He believes that there are three major and different modes of domesticatory behavior -- incidental, specialized, and agricultural.

Incidental domestication describes the relationship between a nonagricultural society and some of the plants on which it feeds. Rindos claims that these are not specialized behaviors. Rather, people (as well as other incidentally domesticating animals) harvest the reproductive or vegetative propagules of the plant and effectively disperse it. The animals act as an opportunistic but not solely necessary agent for plant dispersal. The result is that there are two basic types of incidental domesticates -- wild plants whose distribution is partially determined by human feeding behavior or, conversely, truly domesticated plants whose distribution and dispersal is non-humanly or incidentally determined. It is important to recognize that, unlike other forms of domesticatory behavior, this relationship resides in the nonagricultural environment. The amount of niche space of the plant is environmentally, not culturally, predetermined. Thus, neither the size of the crop nor its yield is under human control. The relationship is fundamentally independent of the subsistence strategy used by the society. The yield obtainable from the environment places a limit on the carrying capacity and on human populations. Rindos distinguishes three evolutionary processes as acting on incidental domestication. First, the spotty spatial distribution of incidental domesticates reduces the speed with which human selection can affect the evolution of the domesticate. Second, the plant is adapted to a nonagricultural environment, and thus many of its adaptive attributes will be lost as the specialized, culturally created environments of agroecology are brought into being. Finally, there are likely to be coevolutionary interactions with nonhuman agents. Taken together this means that the rate of change caused by incidental domesticates is likely to be quite low.

The second process refers to specialized domesticatory behavior. It is both an intensification of incidental domesticates and a rise of new behavioral phenomena. Humans are no longer the opportunistic, but are now the obligate agents of plants. Rindos believes that their behavior is specifically directed towards enhancing the success of the plants. Survival itself becomes dependent upon domesticates and vice versa. The plants rely upon humans for the majority of their dispersal. With specialized domestication not only do people modify their agricultural environment, they emphasize storing and planting behavior. The firing of the grasslands, the cutting of the forests, and the disturbance of the soil modify the local environment, increasing the size of the potentially agricultural habitat. Thus, both populations are freed from the iron rule of intrinsic environmental parameters. The probability of environmental modification is directly related to the amount of time of settlement. Storing creates reservoirs of plants and changes their dispersal systems. It makes the agriculturalist less bound by temporal constraints . Planting of stored seeds or tubers is done on demand of the planter and is not dependent on the exact time that the plant releases its propagative parts. Rindos believes that planting consists of two quite different types of behavior for humans; one is conservative, the other progressive. Replacement planting involves maintenance of the plant species in a preexisting niche. Agricultural planting is the colonization of the plant into new areas created by human disturbance. The latter favors a growing agroecology in which some but not all domesticates are located. Rindos claims that no longer do people directly affect the plant's evolution. Rather it is affected by increasing mediative forces of the agroecology.

Agricultural domestication is the third behavioral process which Rindos postulates. The evolution of domesticates becomes the establishment and refinement of systems of production and takes place solely within the agroecology. Control of the environment is relatively complete using harvesting, seed selection, storage, weeding, and tillage. The life cycle of the plant is under the proverbial thumb of human control. While the relationships of the populations are increasingly obligate during specialized domestication, they have culminated in being completely obligate in this ongoing process. Perhaps the most important tendency of agricultural domestication is the development of systems of selection which create intentionally greater productivity. Rindos has developed two versions of a model to explain how this coevolution from incidental, to specialized, to agricultural domestication takes place. It has three major variablesthe resources comprising diets, their changing abundance, and the absolute and relative amounts of food consumption. The first version is derived mathematically, and Rindos calls it his general model. Underlying his model is the ability to differentiate wild and agricultural resources as well as intentional and non- intentional diet selection. The model operates so that the behavioral interactions of incidental domestication increase the total potential yield from the environment. However, it depresses the rate of growth of the domesticates themselves. This is somewhat counter-intuitional. As specialized domestication is reached, the model reflects an important shift in feeding strategy in that humans reach into the trough in proportion to the perceived abundance of the resource. Thus, increasing diet breadth and yield brings about the transition to agricultural domestication. Full domestication brings rapid population growth due to increases in yield. With population pressure, wild resources are subject to high levels of consumption. The highest yield from domesticates is obtained from minimal reliance upon them and results in a slow rate of evolution. By minimizing the relative contributions of domesticates to the diet one maximizes highest yield and contributes to the evolution of the agroecology .

One could make several specific comments about the mathematics of the general model, but they would actually be minor caveats and technically difficult to explain. I will relinquish in order to make only a few more catholic comments. First, the mathematics are well explained and quite elegant. Second, many of his equations work because he divides the entire resource universe into two sectors -- domesticates and wild resources. His critical equation is ut = dt / (dt + wt) where u is the relative abundance of domesticates at a time t, d is the amount of domesticates at time t, and w is the amount of wild resources at time t. If he used a tripartite or greater classification, the mathematical model would become vastly more complex. Third, the model created a growth curve for domesticates which will be very familiar to demographers. It is essentially a logistic curve (equation 7, figure 5.3) Manipulation shows the effect of an early origin of agriculture on the ultimate development of the system (figure 5.5, etc.) .

I personally find his second version, the graphic model, more interesting. My preference is based on my concurrence that cultural and personal values are very important in subsistence decision making. Rindos models the society's valuation of given resources. This model, along with Cavalli-Sforza and Feldman s Cultural Transmission and Evolution: A Quantitative Approach (1981), is one of the few formal attempts to consider cultural preferences as an operating mechanism in evolution. He is able to show theoretically the importance of differential valuation of resources. Namely, although domesticates increase the total potential yield available to people, the conservative valuation of resources produces an actual diet which is not dominated by domesticates. It is in many ways reminiscent of the models developed by theoretical economists to relate supply, demand, and utility using isopreference curves. Thus, because the rate of evolution for domesticates is a function of the feeding pressure placed upon them, one can expect, and does get, a stable equilibrium between people and early agriculture.

This book is insightful and meaningful. It should and will be on every archaeologist's reading list. There were two aspects of reading this book which I found bothersome. First, I kept wishing that I had written it because it explains so many issues that I think are important. Second, I kept realizing that I could not have. I do not know which is worse.

In short, this book is excellent. If you do not read it you will be missing a classic. The Origins of Agriculture may not be On the Origin of Species (Darwin 1859), but it is in the same tradition of excellence .

Ezra B. W. Zubrow is Associate Professor of Anthropology in the Department of Anthropology, State University of New York, Buffalo. He has published widely on prehistoric ecology and the demography of small populations. He has been working on the size of the prehistoric family in conjunction with the Cambridge Group for the Study of Population and Social Structure. University of Cambridge. He has done fieldwork in North America and Europe.