8 The Monte Alban State: A Diachronic Perspective on an Ancient Core and its Periphery Gary M. Feinman and Linda M. Nicholas In a review of contemporary archaeology, Trigger (1984:286) wrote that: ...what is important is the growing realization that societies are not closed systems with respect to their neighbors any more than with respect to their environment and that the develop- ment of a culture or society may be constrained or influenced by the broader social network of which it is a part. There is also increasing recognition that the rules governing these processes are themselves worthy of scientific investigation. The challenge is to extend a systemic analysis to incorpo- rate what used to be called diffusion. More recently, Schortman and Urban (1987:- 54) noted that: "(t)he long-dormant debate concerning the relation between external contacts and local social change has been reopened." Both of these statements call attention to the growing theoretical rumina- tions in archaeology generated by an increas- ing dissatisfaction with purely endogenous or local models of societal change. These models were swept into vogue two decades ago with a new perspective now referred to as "new ar- chaeology" (e.g., Binford and Binford, 1968). At the same time, few archaeologists would welcome a return to the trait-based diffusion- ism that was associated with an earlier era (see Willey and Sabloff, 1980; Trigger, 1984, for general discussions of theoretical devel- opments in archaeology). In grappling with the organization of a prehispanic Mesoamerican social system at a scale larger than the site or a tightly de- fined region, it is not our intent to add to or sort through the burgeoning jargon that already includes cluster interactions, bound- aries and frontiers, world-systems, and peer-- polity interactions. Nor is it our aim to ignore the importance of local environmental conditions for the elucidation of long-term social change. Rather, our principal objec- tive is to contribute to and expand our under- standing of how ancient Mesoamerica was orga- nized and interconnected. To do this, we examine the long-term and changing interrela- tionship between the prehispanic inhabitants of the Ejutla Valley and those immediately to the north in the larger Valley of Oaxaca (Figure 8.1), both located in the Southern Highlands of Mesoamerica. Soon after the beginning of sedentary agricultural settlement in the Valley of Oaxaca around 1500 B.C., the region was the focus for political and demo- graphic centers that were larger than those found in neighboring regions (including the Ejutla Valley). In this paper, we investigate the dynamic interactions between these centers (first San Jos‚ Mogote and later Monte Alb n) and their surrounding hinterlands (both near and far). Figure 8.1. Map of central and southern Mexico, with areas mentioned in the text. In recent years, macro-regional investiga- tions in Mesoamerica have focused primarily on two "text-aided" contexts, the structure of the Late Postclassic Aztec tribute domain (e.g., Smith 1986, 1987; Berdan, 1987), and the relations between the multiple Maya poli- ties that inhabited the eastern lowlands during the Classic period (e.g., Marcus, 1976a; Mathews, 1985; Freidel, 1986; Culbert, 1988) (see Table 8.1). Here, in our long temporal focus on the Southern Highlands, we must depend largely, although not exclusively, on the sketchy outline that can be discerned from the archaeological record (e.g., Marcus, 1976b, 1983; Marcus and Flannery, 1983; Pad- dock, 1983a). Table 8.1. Chronological Sequence in Valleys of Ejutla and Oaxaca. Ejutla Valley Valley of Oaxaca Mesoamer- ica 1500 1300 Monte Alban V Monte Alban V Late Postclassic 1100 900 Monte Alban IV Early Postclassic 700 Monte Alban IIIB/IV Monte Alban IIIB Late Classic 500 Monte Alban IIIA Monte Alban IIIA Early Classic 300 AD 100Monte Alban IIMonte Alban II Terminal Formative BC 100 Monte Alban Late I Monte Alban Late I 300 Late Formative Monte Alban Early IMonte Alban Early I 500 Rosario 700 Rosario Middle Formative Guadalupe 900 San Jose 1100 Early Formative Early Formative 1300 Tierras Largas In previous studies of precapitalist macro- -regional systems, a central concern has been the spatial division of labor (Blanton and Feinman, 1984; McGuire, 1986; Chase-Dunn and Hall, Chapter 1; Schneider, Chapter 2). Thus, in this analysis, a more specific aim is to examine the spatial arrangement of archaeolog- ically discernible economic (craft) special- izations in prehispanic Ejutla and Oaxaca, as well as to evaluate both endogenous and macro- -regional factors that might account for their distribution. Although these investigations do not yet provide definitive answers, the intent is to give us a better perspective from which to address and assess a series of fundamental issues. Was agricultural tribute the key force behind political expansion? Was the division of labor across the Ejutla-Oaxaca study region uniform, implying only a thin veil of political/elite integration above intraregional self-sufficiency? Can the distribution of craft specialists be accounted for by local agricultural or resource-based considerations? Was macroregional economic interdependence more developed, which by inference would justify at least guarded experimentation with and modification of the fundamental concepts and framework (see Ragin and Chirot, 1984; Abu-Lughod, 1989; Chase-Dunn and Hall, Chapter 1) advanced by Wallerstein (1974). EMPIRICAL BACKGROUND Before preceding to a discussion of our empir- ical foundation, it is important to place the Valley of Oaxaca and Monte Alb n in a broader Mesoamerican context. The Valley of Oaxaca has long been recognized as a key region of prehispanic political and demographic impor- tance (Palerm and Wolf, 1957). Soon after the advent of sedentary village life, San Jos‚ Mogote rose to prominence as one of the larg- est and architecturally most elaborate centers in Mexico's highlands. Yet, this settlement was neither as monumental as several contempo- raneous Gulf Coast lowland communities, nor did it control areas outside the Valley of Oaxaca. By 500 B.C., Monte Alb n, a hilltop community located at the hub of the Valley of Oaxaca, was established (see Blanton et al., 1981, for a discussion of the rise of this early center). In size, Monte Alb n rapidly eclipsed San Jos‚ Mogote, and the later site is generally considered to be one of the earliest cities in Mesoamerica. Yet, even at its apogee (after 200 B.C.), Monte Alb n appears never to have conquered or controlled areas outside the bounds of the contemporary state of Oaxaca. For most (if not all) of its history, Monte Alb n was neither the largest nor the most architecturally monumental site in Mesoamerica (For general discussions of Oaxacan prehistory see Blanton et al., 1981; Flannery and Marcus, 1983). In general, prehispanic Mesoamerica was a world composed of multiple, competing cores and shifting peripheries (see ChaseDunn and Hall, Chapter 1). In this analysis, we rely primarily on the findings of the regional archaeological sur- veys undertaken by the Valley of Oaxaca (Blan- ton, 1978; Blanton et al., 1982; Kowalewski et al., 1989) and Ejutla Valley (Feinman, 1985; Feinman and Nicholas, 1988) Settlement Pattern Projects. During the last two decades, these large-scale projects have systematically mapped, recorded, and dated archaeological remains over a contiguous 2672 km2 area (Figure 8.2). Although this large study region cer- tainly does not represent an entire macro-- regional system, it does allow for the exami- nation of an area larger than the physiograph- ically defined Valley of Oaxaca (Welte, 1973). Comparable pedestrian survey procedures were employed over the entire study region. This methodology (see Feinman et al., 1985; Kowalewski et al., 1989:24-38), which was borrowed (with slight modifications) from the archaeological surveys of the highland Basin of Mexico (Sanders, 1965; Parsons, 1971; Blanton, 1972; Sanders, Parsons, and Santley, 1979), entails the systematic coverage of every field, knoll, ridge, arroyo, and street by crews of three to five people walking 50 to 100 m apart (depending on terrain and the visibility of surface artifacts). Site dimensions, environmental variables, earthen or rubble mounds (the remnants of prehispanic platforms and buildings), pottery, chipped stone, ground stone, artifactual indications of craft activities, defensive walls, and all other important or unusual features were recorded. Where possible these features were mapped directly on 1:5000 aerial photographs that each crew carried into the field. Over time, all archaeological remains were recorded on aerial photographs of the region studied. These field procedures were chosen because they yield information on a large corpus of sites across a broadly defined region at an affordable expense in money and time. They provide, at least for the highlands of Meso- america, the most systematic means available for producing an inventory of the sizes and distributions of archaeological sites at a spatial scale adequate to study long-term societal change. Figure 8.2. Map of the Valleys of Oaxaca and Ejutla, with major sites mentioned in the text. Population estimates for each occupation were determined as a function of site area, following procedures utilized in previous highland Mesoamerican survey research (Sand- ers, 1965). Except in those cases where residential features were visible or where surface artifact densities were extraordinari- ly light or heavy, settlements were calculated as having 10-25 people per hectare of occupied area (see Kowalewski et al., 1989:35, for a fuller discussion). These demographic figures are expressed most appropriately and conserva- tively as ranges; however, for ease and clari- ty of presentation (as well as our sanity), we manipulate and refer to mean population values in most of our analyses. Although we recog- nize the numerous possibilities for error in these estimates, we argue that these data are the closest approximation to a diachronic census of prehispanic settlement that archae- ologists can presently achieve--providing relative patterns of population change at broad spatial and temporal scales. To examine the role of agricultural produc- tion and human-land relations in the regional division of labor, data on agricultural strat- egies and the spatial variability of agrarian resources in ancient Oaxaca was necessary. For this information we rely principally on Anne Kirkby's (1973) landmark, diachronic analysis of the use of land and water resourc- es in the past and present Valley of Oaxaca, Mexico. Kirkby's observations of land quality and productivity, which were supplemented by our own field-by-field observations, were compared with archaeological settlement infor- mation for the sequence of prehispanic phases (Table 8.1). Following Kirkby (1973:124-126) and Kowalewski (1980, 1982; see also Feinman and Nicholas, 1987a; Nicholas, 1989), these analyses take into account the increasing productivity of maize during the prehispanic era. We assume that even the region's earli- est farmers had the knowledge and the tools to implement the basic water control techniques that can be used locally (Kowalewski, 1982:- 150). All available archaeological and ethno- historic evidence suggests that the irrigation and drainage techniques utilized prehispanic- ally were relatively simple (Kirkby, 1973; Lees, 1973; Flannery, 1983), and most of these methods were employed early in the Formative period (Flannery et al., 1967; Drennan and Flannery, 1983). Because we do not know the specific cropping practices and rotations employed prehispanically on each field, we have followed previous investigators (Kirkby, 1973:124-126; Kowalewski, 1982:149-150) in adopting agricultural productivity estimates based entirely on maize (using maize yields as a proxy for total agricultural production). Pre-Conquest maize consumption was presumed (Kowalewski, 1982:158) to correspond to known ethnographic ranges (160-290 kg per person per annum). In our examination of economic specializa- tions, settlement patterns, and human-land relationships, we examine three analytical scales smaller than the entire survey block. To compare the valley's central hub with a southern edge, we contrast the Valley of Oaxaca Settlement Pattern Project survey area with the Ejutla region (Figure 8.2). For other analyses, we compare seven sub-regions, Ejutla and six contiguous segments of the valley that are internally similar in environ- ment and demographic history (Figure 8.3). For a finer scale of investigation, we have broken the study region into 229 grid squares, 4 km on a side (see Figure 8.4 below). The grid square size roughly corresponds to the amount of land in easy walking distance of sites situated in each square (see Chisholm, 1968). Use of the grid greatly facilitates spatial comparisons and cross-phase analyses. The remainder of the paper is divided into two sections. First, we approach the question of the spatial division of labor through an examination of the archaeological evidence for craft activities. In so doing, we evaluate for Oaxaca-Ejutla several extant models that endeavor to explain the distribution of non-- agricultural production. This discussion of craft activities leads us to a consideration of prehispanic agricultural production and its spatial arrangement. Second, we review long-- term changes in the relationship between the Ejutla and Oaxaca regions. Through the inte- gration of these empirical analyses, we gain insight into the macro-regional structure of the ancient Mesoamerican world as seen through the perspective of the Southern Highlands of Mexico. THE SPATIAL DISTRIBUTION OF ECONOMIC SPECIAL- IZATION Numerous models have been proposed for the emergence and distribution of craft special- ization. Generally, these models emphasize local factors, which for prehispanic highland Mesoamerica seem an appropriate starting point given the limitations of transportation tech- nology. In archaeology, a frequently used model (e.g., Arnold, 1975:192, 1980:147) implies that unpredictable or inadequate agricultural resources, particularly "popula- tion pressures," are likely to promote non-- agricultural production. This expectation is exemplified by Howard (1981:7), who noted that: "specialization tends to develop as a necessary adaptation to population pressure and poor agricultural land." An archaeological examination of this proposition is difficult if one wants to extend the test beyond individual specialists and sites. Yet, such scalar expansion clearly is requisite for an adequate evaluation. To gain a broad spatial perspective, dependence on surface remains is necessary, and, as Spence (1983:434) so eloquently recognized, this entails "the imprecision inevitable when working on a regional scale." Furthermore, as Paddock (1983b:433) rightfully has warned, many prehispanic Mesoamerican craft special- izations may be invisible to the archaeolo- gist, particularly when reliant on surface remains. Figure 8.3. Sub-regional divisions of the combined Oaxaca-Ejutla survey block. Nevertheless, in the combined Oaxaca-Ejutla survey block, we have found 268 occurrences of unusual surface residues of shell, spindle whorls (cloth production), ceramics, obsidian, other chipped stone and ground stone (see Figure 8.4). We are not foolish enough to assert that these sites represent the entire corpus of production locations for these materials; in fact our sample probably is skewed toward larger-scale specialization. Nor do we believe that every identified loca- tion was necessarily a locus of non-agricultu- ral production. We also doubt whether even the finest-scale archaeological analysis would be able to distinguish convincingly and re- peatedly between seasonal, half-time, three-- quarter-time, and full-time specialists. Even workshops may not operate year-round. Who can determine from archaeological or even archival sources what craftsmen did in their off-time or how many hours they worked? Despite the obvious limitations, these 268 locations represent the best regionalscale record for these five economic specializations that we have (or are likely to have in the nearfuture) for the study region. To examine the proposition advanced by Howard (1981) and others, we worked at the level of the grid square. The 268 locations were located in 94 squares, just under half of the 201 squares that were inhabited at some time in the prehispanic era. At that scale, a previous study (Feinman, 1986) has shown that the distribution of Monte Alb n I ceramic production sites in the Valley of Oaxaca was not spatially coterminous with areas in which the estimated population would have exceeded the available agricultural resources under average rainfall conditions (see also Kowalew- ski and Finsten, 1983:420). Yet, unlike pottery, the other craft specializations are more difficult to date, particularly at sites with more than one occupational episode. Consequently, to sidestep the problem of chronology, we devised a less demanding, atemporal test of association. For the 201 grid squares, we isolated those squares that had a "dependent population" of at least five people in any one temporal phase (Figure 8.4). By "dependent population," we simply mean a population larger than the number of people that could have been sustained in that partic- ular square by its available land and labor resources. For example, grid square 1206, which includes the site of Monte Alb n, could not have provided sufficient maize to feed its occupants during much of the sequence, so it would be included among those squares with "dependent population." By structuring our analysis independent of time, we not only lessened the interpretive demands on the data, but we increased the likelihood of an associa- tion, since squares in which population depen- dence occurred during a different phase than economic specialization would still yield a positive association. Yet, a significant association was not found (x2=.96, df=1, not signif. at .05) as only roughly half of the squares with "dependent populations" also had indications of economic specialization. Figure 8.4. Spatial relationship between craft specialization and dependent population in the Valleys of Oaxaca and Ejutla. Using this grid-based approach, we also found that squares with the least productive agricultural land did not have greater evi- dence of economic specialization. Here, we limited the analysis to the 129 squares in which 1400 hectares or more had been surveyed (fully surveyed squares have 1600 ha), and then identified the 50 squares with the lowest agricultural potentials. This analysis ex- cluded partially surveyed squares (primarily located in the mountainous zones at the edge of the surveyed area) because a dispropor- tionate number of them would rank among the least productive squares. Most of these edge squares also have little occupation and no evidence of economic specialization. The inclusion of the edge squares in the sample would create an artificially strong negative association between those squares with the least productive land and the presence of economic specialization. Nevertheless, even with the edge squares eliminated, we found no significant association between the loci of economic specialization and poor quality land (x2=3.49, df=1, not signif. at .05). In fact, the opposite tendency was observed. Although the relationship was not statistically signif- icant, economic specialization tended to occur in those squares with better agricultural resources. Evidence for craft specialization was present in only 24 of the 50 least produc- tive squares, yet more than 64% of the remain- ing squares (79) had such evidence. Although recent archaeological models of craft production have tended to focus on population pressure or agricultural marginali- ty, a more traditional perspective (e.g., Childe, 1950) viewed subsistence surplus as the trigger for craft specialization. To assess this factor, we first focused on the 50 grid squares in the study block with the most productive agrarian resources. All but four- teen of these squares were spatially cotermi- nous with evidence for economic specializa- tion, hence the association was statistically significant (x2=16.97, df=1, signif. at .01). Yet, clearly this factor cannot explain the relative abundance of economic specialization in the eastern arm of the valley (see Figure 8.4) where agricultural conditions were gener- ally less favorable. A stronger relationship (x2=30.55, df=1, signif. at .01) was found between high population and economic special- ization. Here we defined "high population squares" as those ranked in the top 15% (by population) during any prehispanic phase. Although this relationship probably has been strengthened by temporal imprecision, the findings do conform with two sets of theoreti- cal expectations. First, specialists, partic- ularly those producing at a relatively large scale, would tend to situate where there is relatively high demand (e.g., Feinman, Kowa- lewski, and Blanton, 1984:301). Second, certain specialists may have been "attached" (see Brumfiel and Earle, 1987:5) to particular sponsors, either social elite or civic-- ceremonial institutions, concentrated at the major population centers. Although the spatial arrangement of occupa- tional specialization was partially accounted for by local socio-political and demographic factors, several aspects of the distribution require additional discussion and considera- tion. For example, evidence for economic specialization was abundant in the eastern or Tlacolula arm, while it is underrepresented at the central core of the valley as well as in distant Ejutla. Tlacolula's agricultural marginality may relate to this apparent densi- ty of specialists; however, we already saw that population pressure could not account for the distribution of specialists at the region- al scale. Furthermore, the northern arm or Etla had the second-highest concentration of occupational specialization, and it is one of the region's most fertile agricultural areas. Significantly, Tlacolula was densely inhabited in Monte Alb n IV and V, and much of the evidence for specialization may pertain to those phases (Finsten, 1983; Kowalewski et al., 1989:348-363). The relative abundance of raw chert sources in both Tlacolula and Etla (Parry, 1987) also may help account for the prevalence of chipped-stone production loca- tions in both of those areas. Yet, pottery and obsidian production locations were abun- dant in those areas respectively (Figure 8.5), and good clay sources can be found throughout the Valleys of Oaxaca and Ejutla, while no obsidian sources are known in the Central Valleys of Oaxaca. Specialized stone tool and ceramic produc- tion locations were scarce in Ejutla (Figure 8.5), and this relative dearth may be account- ed for by the generally lower prehispanic population densities noted in this region as compared to the Valley of Oaxaca (Feinman and Nicholas, 1987b, 1988). In Ejutla, these basically utilitarian goods were more likely to have been made by individual households or at a smaller scale of manufacture. Curiously, the production of higher status goods, like shell and cloth, were more abundant than expected in Ejutla. Shell was recorded at only 20 archaeological locations (out of 2700) in the Valley of Oaxaca; whereas in Ejutla, shell was found at 21 (of 423) sites. Rela- tive to the number of sites in each region, surface shell was between six and seven times more prevalent in Ejutla than Oaxaca. Of the three shell production sites noted in the combined survey area, one was found in Ejutla, while the other two were at San Jos‚ Mogote and Monte Alb n. A fourth shell working area was recorded by Brockington (1973) and Markman (1981) just south of the Ejutla region in Miahuatl n (Figure 8.1). Figure 8.5. Distribution of craft activities and selected exotic items in the Oaxaca-Ejutla survey block. Utilitarian craft activities are represented by the open symbols. Spindle whorls were rare items in both Ejutla and the Valley of Oaxaca, yet their greater prevalence in Ejutla relative to Oaxaca was even more marked than the differen- tial in the presence of shell. Twice as many spindle whorls were found on the ground sur- face in Ejutla as in Oaxaca (11 for Ejutla and six for Oaxaca, including Monte Alb n), even though Ejutla is only one-fourth the areal size (Feinman and Nicholas, 1987b). Based on the three important dimensions of hole diameter, total diameter, and weight, nine of the 11 Ejutla spindle whorls would fit neatly into Mary Parsons' (1972) smallest, or Type III, category of Basin of Mexico spindle whorls, which she associates with the spinning of cotton. The other two spindle whorls are similar in size to her larger Type I whorls and may have been used for spinning maguey. According to Late Postclassic ethnohistoric accounts, cotton did enter the Valley of Oaxaca from lowland areas to the south (Ball and Brockington, 1978), probably through Miahuatl n and Ejutla, and at least some highland Oaxacan towns were importing raw cotton (Ball and Brockington, 1978). Some cotton also may have been grown in Ejutla, which is slightly lower in elevation than the Valley of Oaxaca. Spinning, whether of cotton or maguey, apparently was a more prevalent activity among the prehispanic residents of Ejutla than Oaxaca, and the finished product (as with shell ornaments) may have been traded north into the larger valley. It may be tempting to see the cloth and shell working in Ejutla as simply due to the region's relative proximity to coastal prod- ucts, yet Oaxaca's eastern arm, through Mitla, provides almost comparable access to shell and cotton. Another exotic material, obsidian (a volcanic stone that was highly desired for its cutting capabilities), could have entered the Central Valleys of Oaxaca from any direction through either the eastern, northern, or southern arms. Yet, it seems to have been worked most frequently in Tlacolula (the eastern arm), not Ejutla (Feinman and Nicho- las, 1987b). Consequently, the regional and sub-regional division of labor seems neither entirely accountable to resource proximity nor local environmental conditions. Even demo- graphic factors cannot explain the sparse evidence for craft specialization in the central part of the valley, which often was settled very densely in the prehispanic past. A simulation of potential agricultural production provides an interesting alternative perspective on this issue (Feinman and Nicho- las, 1987a; Nicholas, 1989; Feinman, 1989). For the purposes of this analysis, we define potential production as the quantity of maize that could have been produced at a particular time, if populations farmed only terrain (starting with the best land) within the grid square in which they were located. For each grid, a maximum surplus can then be estimated by subtracting that population's required maize consumption from its potential produc- tion. Although in all phases a surplus could have been produced in the majority of squares (or the inhabitants at least could have fed themselves), a few would have had dependent populations. We also modeled whether these grid square imbalances could have been evened out within larger sub-regional units. For example, following the foundation of Monte Alb n, the site could not have been supported by the land available in the grid square in which it was located. One might then ask, could the site be fed by the immediately surrounding population? By our figuring, even the potential sur- pluses produced during average rainfall years in the squares of the Central sub-region (which includes Monte Alb n) did not compen- sate for the deficit incurred by the popula- tion of Monte Alb n. From Monte Alb n Early I through IV, the Central area had to import maize (see Nicholas, 1989:Figure 14.7). As we saw in the earlier grid square analysis, we again do not see a strong positive relation- ship between population resource imbalance and craft specialization (at the sub-regional scale). Whereas the population of the Central sub-region seems to have faced a somewhat recurrent maize deficit, there is very little evidence for prehispanic craft activities in that part of the region. Rather, our alterna- tive argument follows Blanton's (1985) more general discussion regarding the spatial structure of prehispanic highland Mesoamerican political-economies. We argue that the occupants of the Central area, with the excep- tion of the inhabitants of Monte Alb n, were not involved in craft activities because they were encouraged or coerced to emphasize agri- cultural production to help feed the non-prod- ucers at the urban center. This hypothesis makes sense given the high transportation costs for grain (e.g., Lightfoot, 1979, Dren- nan, 1984a, 1984b). In earlier works (Nicholas et al., 1986; Feinman and Nicholas, 1987a), we illustrated that in Oaxaca the potential to produce large local (or grid square) food surpluses general- ly was centered around major population cen- ters. This pattern is significant and would seem to relate to the concentration of agri- cultural labor around non-food producers. If we examine sub-regional surplus from a slight- ly different angle (one that eliminates from consideration those grid squares with a popu- lation-maize imbalance), the gross agricultur- al potential of the rural population of the Central area is illustrated further. The cumulative surpluses of agricultural produc- tion at the gridsquare level (if those sur- pluses simply were summed rather than shifted over to feed deficit squares in the sub-- region) were not evenly distributed across the study region. Figure 8.6. Surplus production in Monte Alban Late I. The map displays the additional population that could be supported by the gross maize surplus of each sub-region (prior to shifting any produce to squares with food deficits). For example, in Monte Alb n Late I (as in other phases not illustrated here), the poten- tial surplus of rural producers in the Central area was relatively high (Figure 8.6), despite the sub-region's comparatively small areal size and average environmental potential. These food surpluses, which most likely were consumed by the inhabitants of Monte Alb n, clearly were a consequence of the relative abundance of labor in this central portion of Oaxaca. If we summarize the spatial arrange- ment of economic specialization in Oaxaca, what we seem to find is less emphasis on craft specialization at the valley core, where food imports often were required. Yet, the poten- tial to produce sizeable local surpluses was present as long as labor was directed toward agriculture. In contrast, at the edge of the Central Valleys of Oaxaca in Ejutla, easily transportable, laborintensive, high status goods were produced more frequently than expected. The craft work of bulkier, heavier, perhaps less costly items (ceramics, obsidian, chipped and ground stone) tended to be concen- trated inbetween. Although, an interpretation of this pattern is highly speculative, it does conform to Blanton's (1985:400-402) scenario in which highland Mesoamerican states encour- aged intensive food production near their cores and craft manufacture in the political margins (see also Brumfiel, 1976). A sub- -regional division of labor in which status- related goods served in part to interdigitate moredistant regions is suggested (Blanton and Feinman, 1984; Schneider, Chapter 2). The preceding interpretation has glossed over significant temporal variation. For example, a sizeable number of the economic specialization locations in Tlacolula may pertain to the Postclassic period, a time when that sub-region was also a demographic core. The close spatial association between major Postclassic Tlacolula centers, like Mitla, and relatively high rural densities of craftwork may point to another significant organization- al difference between Oaxaca during the Post- classic period and earlier (500 B.C.-A.D. 700) when the region was dominated by Monte Alb n (see Kowalewski et al., 1983; Kowalewski and Finsten, 1983; Marcus, 1989, for discussions of organizational differences between the Classic and Postclassic periods in Oaxaca). THE VALLEY OF OAXACA AND THE EJUTLA VALLEY: A DIACHRONIC PERSPECTIVE We have argued that the character of occupa- tional specialization was different in Ejutla than in the Valley of Oaxaca. In so doing, we have suggested that Ejutla was part of a larger socioeconomic system. In this section, we investigate the long-term interrelationship between these two adjoining valleys. Was the smaller Ejutla Valley simply a microcosm of the larger region, or was its history influ- enced by its changing ties to the region to the north? How was this relationship struc- tured over time? Was the Ejutla Valley simply an additional source of agricultural tribute for Monte Alb n? In this discussion of the relationship between the Ejutla region and the larger valley to the north, we borrow a conceptual distinction made previously by Strassoldo (1980). He distinguishes "frontiers" as open, sparselysettled, almost "virgin" areas of potential growth from "peripheries," which are dependent, more-closed domains that are dis- tant, yet linked, to more developed cores. The Ejutla region was settled late and rather sparsely compared to the Valley of Oaxaca (Table 8.2). While sedentary settle- ments were present in the Valley of Oaxaca during the Tierras Largas phase, the earliest ceramics in the Ejutla region resemble valley ceramics of the subsequent San Jos‚ phase. Such early pottery has been found at only three very small sites in Ejutla, all near the R¡o Atoyac (Feinman and Nicholas, 1988). In comparison, San Jos‚ Mogote in the Valley of Oaxaca extended over 79 ha and included civic- -ceremonial structures that were built during the San Jos‚ phase (Flannery and Marcus, 1976; Kowalewski et al., 1989). The first village settlements in the Ejutla survey region were small pioneering communi- ties that extended down from Oaxaca along the course of the R¡o Atoyac. At this time, these occupations may have formed a true frontier for the Central Valleys of Oaxaca, as no Early Formative settlements have been discovered in the Miahuatl n Valley directly to the south (Markman, 1981). In the subsequent Rosario phase, the Ejutla region remained a sparsely inhabited frontier, occupied by only four small communities. The Rosario-Early I transition in the Valley of Oaxaca was characterized by the emergence of Monte Alb n, as well as the foundation of a series of smaller centers with nonresidential architecture. Dozens of Early I settlements larger than 2 ha were located in Oaxaca, with at least several of these posi- tioned in each sub-region (Kowalewski et al., 1989). As in the Valley of Oaxaca, the popu- lation of Ejutla also increased between the Rosario and Early I phases. Most of the Ejutla settlements still were small farming hamlets located along the Atoyac and its tributaries (Feinman and Nicholas, 1988). Yet in contrast to Oaxaca, none of the Ejutla settlements were larger than 2 ha, and no public architecture could be linked defini- tively with Ejutla's Early I settlements. The first settlements also were recorded in the Miahuatl n Valley during Monte Alb n I (Mark- man 1981). Consequently, at this time when the early center of Monte Alb n was founded on a hilltop at the center of the larger valley to the north, both Ejutla and Miahuatl n remained sparsely occupied. Even the two valley sub-- regions farthest from Monte Alb n, the south- ern Valle Grande and eastern Tlacolula, were settled more than twice as densely as the Ejutla region (Figure 8.7). Since eastern Tlacolula has less fine bottomland and is generally more agriculturally marginal than the Ejutla region, the demographic sparsity of Ejutla relative to Oaxaca seems at least partially a consequence of its spatial posi- tion. The Ejutla region continued to be a lightly settled frontier lacking any archaeological indication of the emergent hierarchical insti- tutions so evident in Oaxaca at this time. The absence of large or civic-ceremonially important Early I centers in both Ejutla and along the southern edge of the Valley of Oaxaca survey region (Kowalewski et al., 1989:103) leads us to suggest that most interactions may have been handled reciprocal- ly by individuals at small, relatively autono- mous communities along this southern frontier. The most rapid episode of prehispanic demographic increase in the Ejutla region occurred between Monte Alb n Early I and Late I (Table 8.2). The number of Late I settle- ments in Ejutla increased three-fold, and these occupations were increasingly differen- tiated in size and architectural complexity. The population density was roughly similar to what it had been earlier in the Valley of Oaxaca during Monte Alb n Early I. Yet based on surface assessments, no Late I settlement in Ejutla was comparable in size or architec- tural complexity to Late I Monte Alb n or for that matter, to Early I Monte Alb n, pre-Monte Alb n San Jos‚ Mogote, or even the larger Late I secondary centers in the Valley of Oaxaca. Figure 8.7. Monte Alban Early I population density by sub-region. In Late I, the Ejutla study area was not dominated by one or two centers as was the Valley of Oaxaca (Kowalewski et al., 1989:113- -152). While almost 20% of the Late I Ejutla sites had estimated mean populations greater than 100, none had more than 350 people. Mounds are associated with 16 Late I compo- nents; however, most of these sites had no more than four structures, and most of the mounds were very small. Later, larger occupa- tions also were present at the four sites where Late I ceramics were associated with more monumental or more numerous structures, and the larger, more substantial construction almost certainly pertains to these later phases. At four sites where Late I was the sole ceramic phase associated with the struc- tures, the single mounds or plaza groups were low and very small. Pending excavations at several of these sites, we suspect that Late I architectural construction generally was internally similar and simple in plan. The settlements with civic-ceremonial buildings were welldispersed along the region's rivers and tributaries, suggesting that Ejutla was not dominated by one principal settlement. In Monte Alb n II, the number of small hamlets occupied in Ejutla remained roughly constant, yet most of the small Late I Ejutla centers diminished considerably in size or were abandoned entirely. Concurrently, three strategically positioned Late I centers in- creased in extent and most probably in archi- tectural elaboration. The two smaller and southernmost of these sites grew to their maximum sizes in Monte Alb n II. Both were associated with 13 comparatively large struc- tures. Although this construction cannot be placed securely in time, these settlements clearly were much larger in Monte Alb n II than they had been earlier. Figure 8.8. Distribution of Monte Alban Late I cremas in Ejutla. The third Monte Alb n II center, placed in the middle of the R¡o Ejutla drainage and positioned underneath the contemporary dist- rito head town of Ejutla (Figure 8.2), grew to roughly twice the size of any prior Ejutla settlement or other contemporaneous community. During two summers of houselot-by-houselot survey, nine very large structures were re- corded and measured, and these have a total volume of approximately 80,000 m3, more than four times the estimated volume of the con- structions at the other two large sites. Several of these mounds were built up more than 12 m, and observations of mound fill indicated a Monte Alb n II construction date. Earthen platforms of this scale were unprece- dented in Ejutla prior to this date. Other factors also point to a significant Monte Alb n II transition in the Ejutla region and a change in the region's interconnection with Oaxaca. The number of small low-lying hamlets in northern Ejutla and the southern Valle Grande decreased, indicating a drop-off in the kinds of open, horizontal communica- tions that are expected along a more open frontier. Instead, for the first time, sever- al Ejutla sites were positioned in defendable locations. However, unlike later defendably situated localities that tended to face out- side the valley, the phase II sites were inward looking, positioned over the most direct route between Monte Alb n and the Ejutla site. A shift in the nature of inter- actions between Oaxaca and Ejutla also is suggested by the changing distribution in the latter region of Monte Alb n I and II cream paste pottery. These distinctive painted crema serving bowls, which were produced (and were recorded) most abundantly in the central and northern parts of the Valley of Oaxaca (Feinman, 1980), were very rare in Ejutla in phase I contexts. The few definitive Monte Alb n I cremas found were distributed rather randomly in terms of site size and location (Figure 8.8). Yet, the temporally specific and highly decorated Monte Alb n II cremas were found more frequently and are particular- ly abundant at the three aforementioned cen- ters (Figure 8.9). Hence, interactions be- tween the two regions may have been handled more directly through elite individuals living at the major centers. Suggestively, prior studies by Spencer (1982) and Redmond (1983) in the Cuicatl n Ca¤ada, a canyon area north of Oaxaca where tropical fruits are grown, document a contemporary episode of local conquest that they relate to Monte Alb n expansionism. Marcus' (1980) analysis of the glyphic record on Building 'J,' an arrowhead-- shaped structure on the Main Plaza at Monte Alb n, led her to a similar hypothesis, that the hilltop center exerted force against other external domains in the state of Oaxaca during Monte Alb n II. Figure 8.9. Distribution of Monte Alban II cremas in Ejutla. Given the apparent transition of Ejutla from a sparsely settled frontier to a nearp- eriphery of the Valley of Oaxaca, it is worth returning to the sub-regional modeling of agricultural surplus potential that we dis- cussed above. Through Monte Alb n Early I, the small Ejutla population precluded the production of any significant maize surplus. Yet, by Late I, a rather large maize surplus could have been produced, equal to or greater than the surplus potential of three valley sub-regions, western Tlacolula, eastern Tlaco- lula, and the southern Valle Grande (Figure 8.6). Whether the demand for tributary agrar- ian surplus was a factor in the southern expansion of the Monte Alb n-centered polity remains unknown; however, our figures suggest that if bulk agricultural surplus was their goal, Monte Alb n's incorporation of Ejutla may not have been immediately successful. In Monte Alb n II, with the concentration and decline of the sub-regional population, Ejut- la's potential for production of maize surplus declined by almost 50% (Figure 8.10). Figure 8.10. Surplus production in Monte Alban II. The map displays the additional population that could be supported by the gross maize surplus of each sub-region (prior to shifting any surplus to squares with food deficits). Ejutla's potential surplus was roughly compa- rable to that of eastern Tlacolula, yet the latter sub-region was closer to the valley core, lessening necessary transportation costs. Although some bulk maize could have been sent to Monte Alb n from Ejutla in Monte Alb n II, the quantities had to have been small. More likely, Ejutla was incorporated as a link to exotic raw materials and craft goods and to solidify defensive and communica- tion networks. In addition, certain tropical plants that could not have been produced in the Valley of Oaxaca may have been grown with greater success in Ejutla's slightly lower elevations. During Monte Alb n IIIA, the great emphasis on new settlement in the southern arm of the Valley of Oaxaca (Kowalewski et al., 1989:201- -250) apparently extended into the Ejutla region, where there was a proliferation of many small hamlets, especially in northern Ejutla and in the central part of the study area along the R¡o Ejutla drainage. Perhaps this greater exploitation of the southern reaches of the Central Valleys of Oaxaca was interrelated with Monte Alb n's loss, between Monte Alb n II and IIIA, of more distant northern peripheries, such as Cuicatl n (Spen- cer 1982). The extension of the demographic patterns observed in the southern Valle Grande into Ejutla during Monte Alb n IIIA suggests that the two regions remained interconnected, with perhaps a basic continuation of the core/periphery relationship that developed in Monte Alb n II. In IIIA, much of the Valley of Oaxaca population was concentrated in several large sites, such as Monte Alb n, and Jalieza in the Valle Grande, and this pattern also was ob- served in Ejutla with the establishment of a large center near San Joaqu¡n, in central Ejutla (Feinman and Nicholas, 1987b, 1988). This site had almost four times the population of the next largest contemporary Ejutla set- tlement. While the old Monte Alb n II center, situated under modern Ejutla, increased in extent, it was completely overshadowed in size by the new center 5 km to the northwest (San Joaqu¡n). In Ejutla, the number of sites with nonresidential architecture also increased, and their distribution became more spatially widespread. The ring of hilltop defendable sites that flanked the eastern and southern arms of the Valley of Oaxaca (Elam, 1989:389) extended well into Ejutla. Thus, the incorpo- ration of Ejutla into the Valley of Oaxaca polity appears to have been more complete by Monte Alb n IIIA (Feinman and Nicholas, 1987- b). Although several factors point to Ejutla's continuation as an important communication and military link, as well as a source for exotic craft goods in IIIA (Feinman and Nicholas, 1987b), the region could have contributed more than twice the maize surplus of the prior phase. Yet, potential surplus maize was still far less in Ejutla than in any valley sub-- region with the exception of eastern Tlacolula (Figure 8.11). For Monte Alb n, grain tribute from Ejutla may have been more of an adminis- trative consideration than it was earlier, yet we doubt that the three to four day round-trip movement costs and the relatively limited potential returns would have made such bulk tribute the central rationale for continued incorporation. Figure 8.11. Surplus production in Monte Alban IIIA. The map displays the additional population that could be supported by the gross maize surplus of each sub-region (prior to shifting any surplus to squares with food deficits). After IIIA, Monte Alb n's hegemony over the Ejutla region began to wane. By Monte Alb n IV, both the Valley of Oaxaca and the Ejutla region were politically fragmented (Kowalewski et al., 1989:251-305). The population of Ejutla declined, and no Ejutla settlement was markedly dominant in size or architectural complexity. Ejutla's population was clustered into several similarly sized settlement con- centrations that were separated by sparsely settled zones. We found a relative abundance of imitation fine gray and fine orange pottery at the easternmost Ejutla settlement clusters. These wares, rare in the Central Valleys of Oaxaca, are thought to signal affiliation with lowland areas to the south and east. Their relative abundance in Ejutla suggests that the region may have developed its own, more direct ties with Mesoamerican populations situated outside the Oaxacan highlands. During Monte Alb n V, the relative politi- cal fragmentation of the Central Valleys of Oaxaca was maintained (Flannery and Marcus, 1983: Chapter 8; Kowalewski et al., 1989:307- -366). The Ejutla region contained settlement clusters, perhaps representing semiautonomous petty states, that were roughly comparable in size to smaller settlement concentrations in Oaxaca. In both Oaxaca and Ejutla, these settlement clusters often were separated by unoccupied or sparsely settled shatter zones. The presence of these zones indicates that the entire region was not well integrated politi- cally in this last prehispanic phase, a point borne out in ethnohistoric records (Flannery and Marcus, 1983: Chapter 8; Marcus, 1989). The relative autonomy of the Ejutla population also is suggested by changes in the propor- tions of obsidian derived from different sources. Ejutla's procurement network for this desirable stone material appears to have been more independent in the Late Postclassic period than it had been during the Early Classic (Feinman and Nicholas, 1987b). During the Postclassic period, the popula- tion densities in Ejutla remained below those in the valley (Feinman and Nicholas, 1987b). Yet, they were closest to the demographic densities observed for the Etla sub-region, an area markedly superior in agricultural poten- tial. However, with Valley of Oaxaca popula- tion concentrations now densest in Tlacolula, Etla also was spatially wellremoved from the areal core (as was Ejutla), suggesting a basis for their relative demographic (as well as architectural) marginality. In sum, in the Late Postclassic, the extent of IIIA politi- cal incorporation of Ejutla by Oaxaca was not repeated, and the degree of economic autonomy of the Ejutla region appears to have been somewhat greater. SUMMARY AND CONCLUSION This paper has approached the macro-regional organization of the prehispanic Southern Highlands of Mexico from two analytical direc- tions. The first examined the spatial distri- bution of economic specialization, while the second focused on the diachronic relationship between the Ejutla Valley and the Valley of Oaxaca. Perhaps we should have listened to the old saying that "to do two things at once is to do neither." Yet, the dual examination has afforded us the opportunity to draw to- gether a series of related observations and inferences that may not have emerged from either investigation alone. We have seen that large-scale differences in economic specialization existed in the Central Valleys of Oaxaca. This diversity and its spatial arrangement could not be explained by either local resource distributions or agricultural conditions alone. Yet these economic divisions may have contributed to the integration and interdependence of the larger socioeconomic system. Clearly, by implica- tion, the notion of relatively uniform, self-- sufficient households bound together by a thin political or economic veneer seems much too simple, as does the traditional redistributive model that envisions that the major intercom- munity linkages were merely environmentally induced. The complexity of ancient Oaxaca's regional and macro-regional connections is further implied by the diachronic transitions and transformations that occurred in and between sub-regional cores and margins. Even Ejutla, always demographically marginal to the valley core, underwent major organizational shifts. Episodes of political fragmentation and cen- tralization in Ejutla were roughly concurrent with similar cycles of change in the valley. Yet, at the same time, demographic processes were certainly not coterminous in the two regions. Luxury goods and craft items apparently had as important a role in the interconnection of the Southern Highland communities and polities as they did in the Central Mexican Aztec world (e.g., Blanton and Feinman, 1984). It seems doubtful that the peripheralization of Ejutla was spurred by Monte Alb n's demand for bulk agricultural surplus alone. The food needs of the ancient Oaxacan capital could have been supported from much closer at hand and at much reduced transport costs. Access to shell, cloth, certain varieties of obsidian, and a range of lowland products not presently visi- ble in the archaeological record, as well as to labor to work those goods, and defensive considerations appear more likely factors behind valley expansionism. In ancient Mesoamerica, the production, exchange, and consumption of ritually impor- tant and luxury items should not be divorced entirely from the consideration of more basic necessities. In the pre-Columbian world, where extractive technology was simple and transportation costs high, labor was a criti- cal variable in food production. Yet, symbol- ically imbued items and non-local products may have played a critical role (through ritual and patron-client transactions) in attracting and integrating the labor necessary to support major centers like Monte Alb n. While we would not want to extend these arguments blanketly to other continents and eras, where beasts of burden and wheeled vehicles made grain transport a more efficient proposition, it is our view that systemically significant (as opposed to epiphenomenal) long-distance relations in ancient Mesoamerica were rarely accountable to food and fuel considerations alone. Although the inhabitants of prehispanic Oaxaca relied on local natural resources, their regional and macro-regional organiza- tions were not simple reflections or conse- quences of such "natural" factors. Rather, ancient Oaxaca had a much more complex politi- cal economy, distinguished by civic-ceremonial cores and peripheries, shifting economic linkages at various scales, uneven demographic change, and a spatial division of labor. As in Colonial and more recent times, inequality, interdependence, and economic differentiation were integral (yet changing) aspects of Oaxa- ca's archaeological past. In sum, our find- ings, though tentative, lead us to support Kohl (1987:5) when he argues that "a position that altogether rejects any correspondences between capitalist and precapitalist or west- ern and non-western societies often tends to distort our vision of the present and idealize that of the past." ACKNOWLEDGMENTS We are grateful for the National Science Foundation support given to both the Valley of Oaxaca Settlement Pattern Project (GS-28547, GS-388030, BNS19640 to Richard E. Blanton; BNS-7914124 to Stephen A. Kowalewski) and the Ejutla Valley Settlement Pattern Project (BNS-84-06229, BNS-85-42668 to Gary M. Fein- man). The permission and assistance of the Instituto Nacional de Antropolog¡a e Historia and the Centro Regional de Oaxaca are recog- nized with great appreciation. Joaqu¡n Garc¡a B rcena and Manuel Esparza have been particu- larly supportive over the years. We also would like to thank the Rota project and Apple Computer for granting us the equipment on which most of our figures were prepared. The initial draft of this paper was pre- sented at the 46th International Congress of Americanists in Amsterdam, Holland, in July, 1988. We would like to thank Peter Druijven and Jan Hardeman for the invitation to par- ticipate in their stimulating session. Ste- phen A. Kowalewski, Scott Cook, Richard E. Blanton, and Joseph Whitecotton read earlier versions of this paper, and we thank them for their useful and provocative comments. Chris- topher Chase-Dunn and Thomas D. Hall offered insightful suggestions that we incorporated into the published manuscript. The final draft of this paper was prepared while the senior author was a resident scholar at the School of American Research. We thank Douglas W. Schwartz and the SAR staff for providing such a wonderful atmosphere for productive research. REFERENCES Abu-Lughod, Janet 1989. Before European Hege- mony: The World System A.D. 1250-1350. New York: Oxford University Press. Arnold, Dean E. 1975. "Ceramic ecology of the Ayacucho Basin, Peru: implications for prehistory." Current Anthropology 16:183-205. ______ 1980. "Localized exchange: an ethno- archaeological perspective." Pp. 147-156 in R.E. Fry (ed.) Models and Methods in Regional Exchange. Society for American Archaeolo- gy, Papers 1. Ball, Hugh G. and Donald L. Brockington 1978. "Trade and travel in prehispanic Oaxaca." Pp. 107-114 in T.A. Lee, Jr. and C. Navarette (eds.) Mesoamerican Communication Routes and Cultural Contacts. New World Archaeolog- ical Foundation, Papers 40. Berdan, Frances, F. 1987. "Cotton in Aztec Mexico: production, distribution and uses." Mexican Studies/Estudios Mexicanos 3:235-262. Binford, Lewis R. and Sally R. Binford (eds.) 1968. New Perspectives in Archeology. Chicago: Aldine. Blanton, Richard E. 1972. Prehispanic Settle- ment Patterns of the Ixtapalapa Peninsula Region, Mexico. University Park: Pennsylvania State University, Department of Anthropolo- gy, Occasional Papers 6. ______ 1978. Monte Alb n: Settlement Patterns at the Ancient Zapotec Capital. New York: Academic Press. ______ 1985. "A comparison of early market systems." Pp. 399-416 in S. Plattner (ed.) Markets and Marketing. Monographs in Economic Anthropology 4. Lanham, MD: University Press of America. Blanton, Richard E. and Gary M. Feinman 1984. "The Mesoamerican world-system: a compara- tive perspective." American Anthropologist 86:673-82. Blanton, Richard E., Stephen A. Kowalewski, Gary M. Feinman, and Jill Appel 1981. Ancient Mesoamerica: A Comparison of Change in Three Regions. Cambridge: Cambridge Uni- versity Press. ______ 1982. Monte Alb n's Hinterland, Part 1: The Prehispanic Settlement Patterns of the Central and Southern Parts of the Valley of Oaxaca, Mexico. Ann Arbor: University of Michigan, Museum of Anthropology, Memoir 15. Brockington, Donald L. 1973. Archaeological Investigations at Miahuatl n, Oaxaca. Nashville: Vanderbilt University, Publications in Anthropology 7. Brumfiel, Elizabeth M. 1976. Specialization and Exchange at the Late Postclassic (Aztec) Community of Huexotla, Mexico. Ph.D. Dissertation, Anthropology, University of Michigan. Brumfiel, Elizabeth M. and Timothy K. Earle 1987. "Specialization, exchange, and complex societies: an introduction." Pp. 1-9 in E.M. Brumfiel and T.K. Earle (eds.) Specializa- tion, Exchange, and Complex Societies. Cam- bridge: Cambridge University Press. Childe, V. Gordon 1950. "The urban revolu- tion." The Town Planning Review 21:3-17. Chisholm, Michael 1968. Rural Settlement and Land Use: An Essay in Location. 2nd ed. London: Hutchinson University Library. Culbert, T. Patrick 1988. "Political history and the decipherment of Maya glyphs." Antiq- uity 62:135-152. Drennan, Robert D. 1984a. "Long-distance transport costs in pre-hispanic Meso- america." American Anthropologist 86:105-112. ______ 1984b. "Long-distance movement of goods in the Mesoamerican Formative and Clas- sic." American Antiquity 49:27-43. Drennan, Robert D. and Kent V. Flannery 1983. "The growth of site hierarchies in the Valley of Oaxaca: part II." Pp. 65-71 in K.V. Flannery and J. Marcus (eds.) The Cloud People: Divergent Evolution of the Zapo- tec and Mixtec Civilizations. New York: Aca- demic Press. Elam, J. Michael 1989. "Defensible and forti- fied sites." Pp. 385-407 in S.A. Kowal- ewski, G.M. Feinman, L. Finsten, R.E. Blanton, and L.M. Nicholas Monte Alb n's Hinterland, Part II: The Prehispanic Settlement Patterns of Tlacolula, Etla, and Ocotl n, the Valley of Oaxaca, Mexico. Ann Arbor: University of Michigan, Museum of Anthropology, Memoirs 23. Feinman, Gary M. 1980. The Relationship be- tween Administrative Organiza- tion and Ceramic Production in the Valley of Oaxaca, Mexico. Ph.D. Dissertation, Anthropology, City University of New York. ______ 1985. "Investigations in a near-- periphery: regional settlement pattern survey in the Ejutla Valley, Oaxaca, Mexico." Mexicon 7:60-68. ______ 1986. "The emergence of specialized ceramic production in Formative Oaxaca." Pp. 347-373 in B.L. Isaac (ed.) Economic Aspects of Prehispanic Highland Mexico. Research in Economic Anthropology. Supplement 2. ______ 1989. "Demography, surplus, and in- equality: early political formations in highland Mesoamerica." Forthcoming in T. Earle (ed.) Chiefdoms and their Evolutionary Significance. Feinman, Gary M., Stephen A. Kowalewski, and Richard E. Blanton 1984. "Modelling ceramic production and organizational change in the pre-hispanic Valley of Oaxaca, Mexico." Pp. 297-333 in S.E. van der Leeuw and A.C. Pritchard (eds.) The Many Dimensions of Pottery. Amsterdam: University of Am- sterdam. Feinman, Gary M., Stephen A. Kowalewski, Laura Finsten, Richard E. Blanton, and Linda M. Nicholas 1985. "Long-term demographic change: a perspective from the Valley of Oaxaca, Mexico." Journal of Field Archaeology 12:333-362. Feinman, Gary M. and Linda M. Nicholas 1987a. "Labor, surplus, and production: a region- al analysis of Formative Oaxacan socio-- economic change." Pp. 27-50 in S. Gaines (ed.) Coasts, Plains and Deserts: Essays in Honor of Reynold J. Rupp‚. Tempe: Arizo- na State University, Anthropological Research Papers 38. ______ 1987b. "Prehispanic interregional interaction in southern Mexico: the Valley of Oaxaca and the Ejutla Valley." Forthcoming in E.M. Schort man and P.A. Urban (eds.) Resources, Power, and Interregional Interac- tion. ______ 1988. "The prehispanic settlement history of the Ejutla Valley, Mexico: a preliminary perspective." Mexicon 10:5-13. Finsten, Laura 1983. The Classic-Postclassic Transition in the Valley of Oaxaca, Mexi- co: A Regional Analysis of the Process of Political Decentralisation in a Prehistoric Complex Society. Ph.D. Dissertation, Anthro- pology, Purdue University. Flannery, Kent V. 1983. "Precolumbian farming in the Valleys of Oaxaca, Nochixtl n, Tehuac n, and Cuicatl n: a comparative study." Pp. 323-339 in K.V. Flannery and J. Marcus (eds.) The Cloud People: Divergent Evolu- tion of the Zapotec and Mixtec Civilizations. New York: Academic Press. Flannery, Kent V., Anne V.T. Kirkby, Michael J. Kirkby, and Aubrey W. Williams, Jr. 1967. "Farming systems and political growth in ancient Oaxaca." Science 158:445-453. Flannery, Kent V. and Joyce Marcus 1976. "Evolution of the public building in Forma- tive Oaxaca." Pp. 205-221 in C.E. Cleland (ed.) Culture Change and Continuity: Essays in Honor of James Bennett Griffin. New York: Academic Press. Flannery, Kent V. and Joyce Marcus (eds.) 1983. The Cloud People: Divergent Evolution of the Zapotec and Mixtec Civilizations. New York: Academic Press. Freidel, David A. 1986. "Maya warfare: an example of peer polity interaction." Pp. 93-108 in C. Renfrew and J.F. Cherry (eds.) Peer Polity Interaction and Socio-Political Change. Cambridge: Cambridge University Press. Howard, Hilary 1981. "In the wake of distribu- tion: towards an integrated approach to ceramic studies in prehistoric Britain." Pp.1-30 in H. Howard and E.L. Morris (eds.) Production and Distribution: A Ceramic Viewpoint. British Archaeological Reports, International Series 120. Kirkby, Anne V.T. 1973. The Use of Land and Water Resources in the Past and Present Valley of Oaxaca, Mexico. Ann Arbor: Universi- ty of Michigan, Museum of Anthropology, Memoirs 5. Kohl, Philip L. 1987. "The use and abuse of world systems theory: the case of the pristine West Asian state." Pp. 1-35 in M.B. Schiffer (ed.) Advances in Archaeological Method and Theory, Volume 11. San Diego: Academic Press. Kowalewski, Stephen A. 1980. "Population-- resource balances in period I of Oaxaca, Mexico." American Antiquity 45:151-165. ______ 1982. "Population and agricultural potential: Early I through V." Pp. 149-- 180 in R. Blanton, S. Kowalewski, G. Feinman, and J. Appel Monte Alb n's Hinterland, Part I: The Prehispanic Settlement Patterns of the Central and Southern Parts of the Valley of Oaxaca, Mexico. Ann Arbor: University of Michigan, Museum of Anthropology, Memoirs 15: . Kowalewski, Stephen A., Richard E. Blanton, Gary M. Feinman, and Laura Finsten 1983. "Boundaries, scale, and internal organiza- tion." Journal of Anthropological Archaeolo- gy 2:32-56. Kowalewski, Stephen A., Gary M. Feinman, Laura Finsten, Richard E. Blanton, and Linda M. Nicholas 1989. Monte Alb n's Hinterland, Part II: The Prehispanic Settlement Patterns of Tlacolula, Etla, and Ocotl n, the Valley of Oaxaca, Mexico. Ann Arbor: University of Michigan, Museum of Anthropology, Memoirs 23. Kowalewski, Stephen A. and Laura Finsten 1983. "The economic systems of ancient Oaxaca: a regional perspective." Current Anthropology 24:413-441. Lees, Susan H. 1973. Sociopolitical Aspects of Canal Irrigation in the Valley of Oaxaca. Ann Arbor: University of Michigan, Museum of Anthropology, Memoirs 6. Lightfoot, Kent G. 1979. "Food redistribution among prehistoric Pueblo groups." Kiva 44:319-339. Marcus, Joyce 1976a. Emblem and State in the Classic Maya Lowlands: An Epigraphic Approach to Territorial Organization. Washing- ton, D.C.: Dumbarton Oaks Research Library and Collections. ______ 1976b. "The iconography of militarism at Monte Alb n and neighbor ing sites in the Valley of Oaxaca." Pp. 123-139 in H.B. Nichol- son (ed.) Origins of Religious Art and Iconography in Pre-Classic Mesoamerica. Los Angeles: UCLA Latin American Center. ______ 1980. "Zapotec writing." Scientific American 242:50-64. ______ 1983. "The conquest slabs of Building J, Monte Alb n." Pp. 106-108 in K.V. Flannery and J. Marcus (eds.) The Cloud People: Diver- gent Evolution of the Zapotec and Mixtec Civilizations. New York: Academic Press. ______ 1989. "From centralized systems to city-states: possible models for the Epi- classic." Pp. 201-208 in R.A. Diehl and J.C. Berlo (eds.) Mesoamer ica after the Decline of Teotihuacan A.D. 700-900. Washington, D.C.: Dumbarton Oaks Research Library and Collec- tions. Marcus, Joyce and Kent V. Flannery 1983. "The Postclassic balkanization of Oaxaca." Pp. 217-226 in K.V. Flannery and J. Marcus (eds.) The Cloud People: Divergent Evolution of the Zapotec and Mixtec Civilizations. New York: Academic Press. Markman, Charles W. 1981. Prehispanic Settle- ment Dynamics in Central Oaxaca, Mexico: A View from the Miahuatl n Valley. Nashville: Vanderbilt University, Publications in Anthropology 26. Mathews, Peter 1985. "Maya early Classic monuments and inscriptions." Pp. 5-54 in G.R. Willey and P. Mathews (eds.) A Consider- ation of the Early Classic Period in the Maya Lowlands. Albany: Institute for Mesoamerican Studies, State University of New York at Albany, No. 10. McGuire, Randall H. 1986. "Economies and modes of production in the prehistoric south- western periphery." Pp. 243-269 in F.J. Math- ien and R.H. McGuire (eds.) Ripples in the Chichimec Sea: New Considerations of South- western-Mesoamerican Interactions. Carbondale: Southern Illinois University Press. Nicholas, Linda M. 1989. "Land use in prehis- panic Oaxaca." Pp. 449-505 in S.A. Kowal- ewski, G.M. Feinman, L. Finsten, R.E. Blanton, and L.M. Nicholas Monte Alb n's Hinterland, Part II: The Prehispanic Settlement Pat- terns of Tlacolula, Etla, and Ocotl n, the Valley of Oaxaca, Mexico. Ann Arbor: University of Michigan, Museum of Anthropolo- gy, Memoirs 23. Nicholas, Linda M., Gary M. Feinman, Stephen A. Kowalewski, Richard E. Blanton, and Laura Finsten 1986. "Prehispanic colonization of the Valley of Oaxaca, Mexico." Human Ecology 14:131-162. Paddock, John 1983a. Lord 5 Flower's Family. Nashville: Vanderbilt Univer sity, Publica- tions in Anthropology 29. ______ 1983b. "Comment on 'The economic systems of ancient Oaxaca: a regional per- spective,' by Stephen A. Kowalewski and Laura Finsten." Current Anthropology 24:433. Palerm, Angel and Eric R. Wolf 1957. "Ecologi- cal Potential and Cultural Development in Mesoamerica." Pan American Union Social Sci- ence Monograph 3:1-37. Parry, William J. 1987. Chipped Stone Tools in Formative Oaxaca, Mexico: Their Procure- ment, Production and Use. Ann Arbor: Universi- ty of Michigan, Museum of Anthropology, Memoirs 20. Parsons, Jeffrey R. 1971. Prehistoric Settle- ment Patterns in the Texcoco Region, Mexi- co. Ann Arbor: University of Michigan, Museum of Anthropology, Memoirs 3. Parsons, Mary Hrones 1972. "Spindle whorls from the Teotihuac n Valley, Mexico." Pp. 45-79 in M.W. Spence, J.R. Parsons, and M.H. Parsons Miscellaneous Studies in Mexican Prehistory. Ann Arbor: University of Michi- gan, Museum of Anthropology, Anthropological Papers 45. Ragin, Charles and Daniel Chirot 1984. "The world system of Immanuel Wallerstein: sociology and politics as history." Pp. 276-- 312 in T. Skopol (ed.) Vision and Method in Historical Sociology. Cambridge: Cambridge University Press. Redmond, Elsa M. 1983. A Fuego y Sangre: Early Zapotec Imperialism in the Cuicatl n Ca¤- ada, Oaxaca. Ann Arbor: University of Michi- gan, Museum of Anthropology, Memoirs 16. Sanders, William T. 1965. "The cultural ecolo- gy of the Teotihuacan Valley." University Park: Department of Sociology and Anthropolo- gy, Pennsyl vania State University. Sanders, William T. and Deborah L. Nichols 1988. "Ecological theory and cultural evolu- tion in the Valley of Oaxaca." Current Anthro- pology 29:33-80. Sanders, William T., Jeffrey R. Parsons, and Robert S. Santley 1979. The Basin of Mexico: Ecological Processes in the Evolution of a Civilization. New York: Academic Press. Schortman, Edward M. and Patricia A. Urban 1987. "Modeling interregional interaction in prehistory." Pp. 37-95 in M.B. Schiffer (ed.) Advances in Archaeological Method and Theory, Volume 11. San Diego: Academic Press. Smith, Michael E. 1986. "The role of social stratification in the Aztec empire: a view from the provinces." American Anthropologist 88:70-91. ______ 1987. "Archaeology and the Aztec econo- my: the social scientific use of archaeo- logical data." Social Science History 11:237-- 259. Spence, Michael W. 1983. "Comment on 'The economic systems of ancient Oaxaca: a re- gional perspective,' by Stephen A. Kowalewski and Laura Finsten." Current Anthropology 24:433-34. Spencer, Charles S. 1982. The Cuicatl n Ca¤ada and Monte Alb n. New York: Academic Press. Strassoldo, Raimondo 1980. "Centre-periphery and system-boundary: culturological per- spectives." Pp. 27-61 in J. Gottmann (ed.) Centre and Periphery: Spatial Variation in Politics. London: Sage Publications. Trigger, Bruce G. 1984. "Archaeology at the crossroads: what's new?" Annual Review of Anthropology 13:275-300. Wallerstein, Immanuel 1974. The Modern World-- System I. New York: Academic Press. Welte, Cecil R. 1973. "Ready reference release No. 2." Oaxaca. Willey, Gordon R. and Jeremy A. Sabloff 1980. A History of American Archaeology. San Francisco: W.H. Freeman.