14 Herds, Fish, and Fowl in the Domestic and Ritual Economy of Formative Chiripa KATHERINE M. MOORE, INTRODUCTION THIS CHAPrER PRESENTS THE results of a pilot study of Early Chiripa phase bone remains from the 1996 excavations at the Santiago sector of Chiripa and describes the ongoing research program for animal bones of the Taraco Archaeo- logical Project. Analysis of animal bone material from the 1992 season and of other components excavated during the 1996 season is in progress. Of the three authors of this work, Katherine Moore is responsible for the work on the large mammal remains, Susan deFrance for the fish remains, and David Steadman for the remains of birds, small mammals, and herps. Previous work on animal bones from the site has been undertaken by Jonathan Kent (1982) and Darwin Horn (Erickson and Horn n.d.) using material from the Monticulo complex. We approach our study with several goals. We wish to contribute to a better understanding of the daily activities and economic and ecological constraints upon the prehistoric populations of Chiripa. We hope to identify the most important DAVID STEADMAN, AND SUSAN DEFRANCE animal foods and estimate their relative dietary contribution. Taking these goals together, we hope to describe the cultural ecology of the southern Titicaca lakeshore and offer insights into the ritual system that is thought to have emerged there. MATERIALS The total bone assemblage from the 1992 and 1996 excavations amounts to several hundred thousand specimens. The material discussed here is a small sample selected from the total. The sample selected for this pilot study was identified based on an examination of Early Chiripa ceramic samples from the Santiago sector of the site (primarily B 16). This surface includes several pits, at least one burial, several hearths or dumps from hearths, and fill around a wall of a structure. We have begun our analysis of the 1996 remains by examining the bone recovered from this well-dated and relatively well-understood component. The collection includes screened samples from 36 loci; a smaller selection of 10 flotation samples from the same units was also analyzed (tables 16 and 106 Ear/N Settlement at Chiripa, Bolivia 17). In total, a sample 'of about 1500 mammal and bird bones was studied along with about 20,000 fish bones. THE BONE ASSEMBLAGE As DIETARY REMAINS One of the most important goals of our work is to estimate the relative importance of dietary components at Chiripa. In previous work at Chiripa estimates of the importance of large mammals in relation to fish ranged from 2 to 12% fish (Kent 1982, 1989, Erickson and Horn n.d.). In deriving dietary information from bone re- mains, problems arise when standard 1/4 inch mesh screens fail to recover the bulk of a probable important dietary component, small fish. Finer scale recovery techniques were used together with this standard screening program to improve the recovery of fish and small bird and mammal remains. Another issue is the difficulty in assigning taxa identification to large mammal bone. In the Andes, camelids (guanaco, llama, alpaca, and vicufia) are usually impossible to separate in fragmentary assemblages, and their fragmented bones are also difficult to separate from the bone of the common deer species in the region. Thus, there is a large category of unidentifiable "large mammals" whose remains clearly represented an important portion of the animal biomass obtained by prehistoric inhabitants. In this sample, 89% of fragments and 50% of the bone weight fell into an unidentifiable "large mammal" category. For the Santiago Early Chiripa phase dietary analysis, bones identified to the level of camelid were lumped with large mammal bone to give a general estimate of the importance of that ecological niche. Birds and small mammals were also lumped together, and fish remains were considered as one group. The first step in our dietary reconstruction was to determine from the screened samples that meat from large mammals had made up about 93% of the diet (table 16). About half of the large mammal was securely attributed to camelids and the rest was assumed to be camelid (no post- cranial remains of deer appeared in the sample). The remainder of the animal economy was made up of birds and small mammals at 2.5% and fish at 3.6%. The flotation samples from Chiripa revealed that the tiny bones of lake fish are much more common than the 1/4 inch screened samples would indicate. To improve the accuracy of our dietary and economic estimates, we worked backwards from the proportions of fish bones in carefully measured flotation samples to estimate how much fish bone had been lost from the screened samples. Proportions by weight of large mammal bone to fish bone from the flotation samples were applied to the amount of large mammal in the screened samples, producing an estimate of fish bone that had passed through the screen into the backdirt. When these missing fish bones were factored back into our dietary reconstruction, we were able to estimate that fish bone was more abundant by weight than large mammal bones for many of our test samples. To dramatize the difference made by a full accounting of the fish, we calculated the biomass represented by the various dietary components (table 18). These estimates were made on bone weight determinations, following the techniques of Reitz et al. (1987). In a further phase of this study, we hope to refine these biomass estimates by using fish total length estimates rather than fish bone weights. Using only the sample of fish bone from the 1/4 inch screened samples, estimates for each locus for fish ranged from 2 to 30% of biomass or total meat weight. With the full sample of fish bone added to calculations of total biomass, the importance of fish in Early Chiripa fill and pits ranges from 2 to 94% of biomass. The average fish biomass becomes 67%. This finding parallels those from other sites where controlled recovery has been added to traditional excavation tech- niques. The proportions revealed by fine-screen- ing force us once again to consider how important aquatic resources (marine and fresh water) have been in provisioning early sedentary societies. The importance of camelid meat in the daily diet is swamped by the use of fish as a staple. This is in line with expectations based on modeling ecologi- cal abundance and examining ethnographic accounts of how rarely Titicaca Basin herders slaughter their llamas (see Nachtigall 1966, for example; Horn 1984). We have identified small amounts of deer antler but suspect that it was obtained as a raw material for tools and not necessarily as part of a carcass. Herds, Fish, and Fowl 107 TABLE 16 Santiago Early Chiripa Phase Animal Bone (screened sample, bone wt.) Locus CAMELID CAMELID CAMELID LARGE CARNI- SMALL FISH TOTAL WT. LARGE SMALL INDETERM. MAMMAL VORE MAMMAL, (g) FORM FORM INDETERM. BIRD 400.1 49.5 66.7 9.7 66.5 37.3 33.7 5.9 10.1 122.7 20.0 45.3 26.7 7.8 2.7 904.7 28.1 8.6 46.2 1.1 1.1 33.8 38.5 38.3 42.5 22.5 15.4 2.9 3.7 25.6 2.4 38.0 40.6 60.1 157.1 22.6 40.4 28.5 6.8 23.4 17.7 3.3 2.3 20.3 5.0 55.4 10.4 31.0 61.0 30.2 10.0 186.6 6.1 29.6 25.0 48.0 179.1 1.9 64.6 24.9 29.8 11.8 5.7 42.0 46.1 10.3 53.1 2.0 38.3 2.1 46.6 183.9 12.2 49.3 2.5 61.6 103.6 0.5 57.5 24.1 13.5 35.0 2.7 43.2 1.8 3.7 0.9 60.7 5.7 11.1 21.0 0.6 4.5 243.1 538.9 1,746.0 THE CAMELIDS The camelid remains at Chiripa are part of a remarkable archaeological and anthropological tradition. Effigies of llamas and alpacas in precious metals have been recovered from pre- sumed offering sites around Lake Titicaca, camelid bones have themselves been recovered with associated underwater offerings in the lake, and Aymara folk practice makes use of dried fetal alpacas and llamas in several rituals. Unfortu- nately it is difficult even to reconstruct the basic economic patterns of hunting, herding, and cara- 3.4 4.8 0.3 3.1 27.2 0.7 0.7 0.1 3.7 3.6 31.7 0.2 3.4 1.4 0.2 2.3 7.1 1.0 1.3 5.5 0.2 0.6 1.1 0.5 77.2 4.9 2.2 2.4 0.8 0.7 0.6 0.8 15.5 15.0 0.2 2.4 1.2 0.4 0.6 2.0 0.4 0.5 1.2 0.2 563.6 10.8 37.3 152.0 39.4 2.9 30.7 41.2 312.0 110.6 161.3 3.3 23.0 75.9 171.5 264.7 38.7 295.0 67.9 67.8 105.9 75.3 41.6 355.8 81.5 214.7 86.2 24.5 58.9 49.2 5.5 1.1 63.5 39.6 0.6 5.1 3,678.5 1.2 0.5 0.7 0.7 0.4 0.4 0.6 0.9 1.1 8.6 100.5 2.2 0.2 2.0 0.7 0.6 136.5 van transport from the fragmented bones from most sites, including Chiripa. The interpretation of camelid remains can be highly ambiguous since the skeletons of the four living species are difficult to separate objectively. The most common practice is to report the number of bones whose measurements fall above and below a series of decision rules or "cut-off points" originally provided by Wing (1972). This provides an estimate of the proportion of animals in the large size class (guanaco and llama) and the smaller size class (alpaca and vicunia), but it does not reveal whether the animals are wild or domesticated. 1198 1199 1235 1236 1238 1240 1244 1247 1248 1275 1276 1279 1281 1282 1285 1294 1301 1302 1305 1307 1316 1317 1319 1320 1322 1323 1324 1405 1406 1407 1408 1480 1483 1484 1496 1499 Total 108 Earlv Settlement at Chilripa, Bolivia TABLE 17 Archaeological Context of Early Chiripa Fauna (1/4" screen only): Density, Burning, and Taphonomic Indicators. Domesticated or wild status has sometimes been inferred based on geographic range or date. The dental differences between vicunia and the other forms are also sometimes reported (see Wheeler 1984, for example). Ideally, we would like to be able to provide estimates of the proportions of the wild and domesticated species of camelids and to estimate the age and sex structures of the original herds. For the Formative periods in Bolivia, studying the appearance of domestic herds of llamas and alpacas is a key research goal. Using remains from the 1974-75 excavations at Chiripa, Kent (1982) used a discriminate function analysis to assign camelid species to measured bones. He found that 80% of the camelids were of the large form (that is guanaco or llama) and the rest the small form (alpaca with one specimen assigned to vicuiia). In strata dating from about 1400 B.C., Kent identified domesticated llama. Determinations of large and small camelid reported here come from the Wing decision rules for all the elements that have been studied (table Locus DENSITY DENSITY BURNED FRAGMENT FRAGMENT FRAGMENT FRAGMENT TAPHONOMIC L. MAM. FISH BONE % WT. GM. WT. GM. WT. GM. WT. GM. HISTORY (gm/I) (gm/I) OF WT. (L. MAM.) UNBURNED BURNED CALCINED 1198 8.050 1.287 14.2% 6.0 11.5 2.4 1.9 1199 0.037 0.163 0.0% 0.4 0.4 1235 0.062 0.004 4.0% 1.7 1.8 0.7 1236 0.314 0.006 8.1% 2.1 2.3 1.2 erosion, carnivore damage 1238 0.158 0.003 0.0% 7.6 7.6 1240 0.048 0.0% 1.0 1.0 1244 0.225 0.005 9.2% 1.3 1.4 0.7 0.8 1247 0.238 0.004 25.7% 1.8 1.6 2.6 1248 0.640 0.002 15.4% 2.2 2.5 1.4 erosion, rodent and carnivore damage 1275 1.144 0.194 50.3% 1.7 2.5 1.8 0.3 1276 1.146 0.150 26.4% 3.2 3.8 2.2 1279 0.165 0.0% 3.3 3.3 1281 0.161 0.001 18.6% 1.6 1.7 1.4 1282 0.539 0.018 3.3% 1.3 1.3 0.8 1285 2.111 0.015 15.2% 3.2 3.3 2.9 1294 1.651 0.003 41.7% 1.7 1.8 1.6 1.4 carnivore damage 1301 0.397 0.007 26.9% 1.5 1.5 1.6 1302 0.866 0.006 19.6% 3.1 3.4 2.2 carnivore damage, digestion 1305 2.217 0.013 14.9% 1.1 1.1 1.4 0.9 1307 1.108 59.8% 4.4 3.0 6.6 carnivore damage 1316 0.767 0.004 30.9% 1.4 1.4 1.6 carnivore damage 1317 0.525 0.009 7.5% 1.2 1.5 0.3 1319 0.212 0.001 3.2% 1.6 1.6 1.3 1320 1.269 13.8% 3.7 3.7 3.5 2.6 1322 2.038 24.3% 3.3 3.6 2.5 1323 1.065 0.002 2.6% 2.7 2.8 1.4 1324 0.273 0.002 11.4% 1.9 1.9 0.9 1405 0.201 0.003 23.2% 2.2 2.1 2.8 1406 0.225 0.009 6.7% 2.6 2.6 1.9 carnivore damage 1407 0.810 10.9% 1.7 1.7 1.3 root etching 1408 0.079 0.0% 1.8 1.8 1480 0.008 0.001 0.0% 0.9 0.9 1483 0.319 0.010 9.6% 1.3 1.4 0.6 1484 0.378 0.007 18.5% 2.1 2.4 1.4 1496 0.030 0.0% 0.6 0.6 1499 0.225 0.030 35.6% 0.9 0.7 1.6 Averages 0.825 0.070 15.3% 2.2 2.4 1.8 1.3 Herds, Fish, and Fowl 109 TABCE 18 Calculation of relative dietary importance of fish. 19). In addition, Moore is using here decision rules for carpal and tarsal bone (wrist and ankle) measurements based on previous work with col- lections in the Laboratorio de Paleoetnozoologia of the Facultad de Medicina Veternaria, Universidad N.M. de San Marcos in Lima in 1981 (see Malaga et al. 1976 for background on this sample). These auxiliary decision rules allow a significantly larger sample of measured bones to be studied, since many of the complete bones from Chiripa are not included in the Wing decision rule set. The main weakness with using carpals and tarsals for size determination is that it is impos- sible to securely exclude immature animals from consideration since these bones do not have multiple centers of ossification. The fusion of multiple centers of ossification at epiphysial lines is typically used to determine a bone's maturity (though it is known that these mature elements do increase in size with time after fusion has taken place). Note that such measurement schemes on mature bones provide size estimates for adults only, and provide no comparable determinations of size class for juvenile animals. In the Early Chiripa phase Santiago sample, 65.2% of measured camelid bones were from large camelids (amounting to 80% by weight), a propor- tion lower but similar to that found by Kent (1982). This is concordant with data for presumed early herding communities from southern Peru reported by Miller (1979) and Wing (1972, 1977). Farther north in Peru, proportions tend to shift to a predominance of smaller forms (alpacas and/or vicunias). At least one individual in the Early Chiripa sample was larger in most dimensions (width but not length of joints in the leg) than any comparative specimen in the collections that underlie these studies. A few loose teeth confirm the presence of the vicunia in the deposit, but dental remains are rare overall. One archaeological measure that has been associated with herding domesticated camelids in the Andes is a high proportion of very young animals, presumably lost during the difficult wet season birthing period. Neither Kent (1982) nor I have identified this pattern at Chiripa, contrasting with the strong pattern in components of contem- porary Peruvian sites such as Panaulauca (Moore 1989) and Telarmachay (Wheeler 1984, 1985). The majority of camelid remains in both studies represent animals who were older juveniles and adults. Of the 132 bones for which fusion status could be determined, 10.6% were from fetal or neonate animals, 38.6% were from unfused elements, and the remainder (50.7%) came from bones that were in the process of fusing or fused. This thumbnail sketch of the assemblage cannot be translated into estimates of the ages of the animals Locus no. Fish bone, Fish bone, Fish bone, Fish bone, Total fish bone Fish, % wt. Fish Fish of Flot wt. gm, 1/4" wt. gm, wt. gm, 1/4" lost in 1/4" wt. gm; in total Biomass Biomass, % Sample fraction < 1/4" screen screen, re- screened and screened (meat kg) of total meat from flot fraction sample constructed and "missing" sample from flot wt. gm 1238 0.1 4.6 0.76 35.0 35.7 48.5% 0.81 54.0% 1244 0.01 5.5 0.68 374.0 374.7 92.7% 5.47 90.9% 1276 0 15 15.0 11.6% 0.40 17.7% 1282 0.5 5.6 2.37 26.5 28.9 29.2% 0.69 36.3% 1302 0.1 7 2.04 142.8 144.8 33.6% 2.53 37.2% 1316 0.01 5.3 0.51 270.3 270.8 73.1% 4.20 71.8% 1404 0.8 5.9 6.7 27.3%* 0.21 26.9%* 1405 0.3 7 0.36 8.4 8.8 26.7% 0.26 36.1% 1478 0.5 3.2 3.7 21 .9%* 0.13 39.8%* 1499 0.2 15.8 0.6 47.4 48.0 91.4% 1.04 91.0% *No screened samples (100% of locus floated). Calculations based on flot samples only. 110 Early Settlenment at Chiripa, Bolivia TABLE 19 Early Chiripa Sample, Santiago: Camelid and Large Mammal Summary. since each center of ossification or bony growth has its own fusion timetable. The major impres- sion drawn from these data is the ready availability of prime age animals. A slightly greater level of detail on camelid age structure is available for a few elements for which an adequate sample has been studied. For the first phalanx or toe bone, I of 14 specimens appeared to be from an animal less than three months, 2 from animals from 3 to 18 months, and 11 from animals older than 18 months. Little information can be gathered at this time on how old the adult animals might be. Once the bone of the skeleton fuses completely at about 3 years for camelids, the process of aging can no longer be tracked by this method. (Relative aging techniques for older animals rely on tooth wear observations, which we will apply to any dental remains that may be recovered.) This pattern of adult animals could be consistent either with prime-oriented hunting or with a conservative (and fortunate in the face of usual disasters) strategy of harvesting domesti- cated animals as they reach their maximum weight. Abundant spinning and weaving equip- ment from the site indicate the importance of textile production, at least in the later periods, which we would expect to be associated with a relatively mature population of animals. No dental remains indicating advanced age (with deeply worn molars) have been observed in the sample so far. We are alert to the possibility of sacrificial slaughter but have no evidence from this sample. A faint seasonal signal can be derived from determinations of camelid age at death. Occupa- tion during the wet season and early dry season may be inferred by the presence of the youngest age category. Occupation during the later dry and early wet season cannot be established using these data. Other seasonal data may later be available from camelid dental eruption data, from incre- mental structures in bones from various animals, and from bird migration and breeding patterns. We know from other parts of the assemblage (the birds and small mammals, for example) that hunting was a persistent part of food gathering. Hunting and fishing smaller animals may have used simpler technologies and involved very different social organization than that used while hunting large game. Specifically, women and children could have collected eggs, netted fish, and set snares and traps within a short distance from the settlement along the shoreline. Identification Total % of Ct. Camelid Size Classes and Wt. Camelid Large Form Ct. 43 3.0% Large form ct. 65.2% Wt. (g) 904.7 26.3% wt. 78.8% Camelid Small Form Ct. 23 1.6% Small form ct. 34.8% Wt. (g) 243.1 7.1% wt. 21.2% Camelid Form Indeterminate Ct. 86 5.9% Wt. (g) 538.9 15.7% Camelid Subtotals Ct. 152 10.5% Wt. (g) 1686.7 49.0% Carnivore Ct. 2 0.1% Wt. (g) 8.6 0.2% Large Mammal Indeterminate Ct. 1296 89.1% Wt. (g) 1746 50.7% Total Early Chiripa Large Mammal Ct. 1450 Wt. (g) 3441.6 Herds, Fish, anid Fowl III FISH As noted above, simple analysis of bulk remains indicates that fish were the major source of animal food for the site's occupants. Scales and bones of fish are common through the entire sequence. They typically represent very small species, probably less than 10 cm in length. The analysis of individual specimens of the fish bone has just begun. The modem fish fauna of Lake Titicaca is dominated by 23 endemic species of killifish in the genus Orestias, with various species being specialized inhabitants of the shallow-, mid-, and deep-waters (Parenti 1984). The great major- ity of fish bone appears to have come from several species of Orestias, though Trichomycterus (small, burrowing catfishes, Trichomycteridae) has also been identified. Fishing in Lake Titicaca is done with dip nets and seines today and was likely done this way in the past. Many bone tools from the site appear to relate to the tasks of net making, including finely made net gauges. We plan to take the research on the fish bones in several directions. First, more precise identifications of species, or at the very least, determinations of the sizes of fish taken may allow us to identify specific niches or feeding conditions exploited by Chiripa residents. Second, the analysis of the density of fish remains and fish skeletal completeness will provide a picture of the daily routine of food preparation, consumption, and discard. Preliminary data on fish bone density are noted in table 17, indicating a pattern of uneven density of these remains. This level of detail is important since our concerns surround the operation of household economies within this (probably) ritual local context. BIRDS In contrast to the very low diversity of the fish and mammal remains, the bird fauna of Lake Titicaca is rich, including local species drawn to the lake margins and many migratory species (table 20). David Steadman, studying the 1992 materials, has identified an even more diverse archaeological bird fauna than is known from the area today, producing a list of 23 taxa from all components taken together. (The bird and herp sample discussed here includes later materials as well as Early Chiripa materials.) Aquatic birds are more important than terrestrial birds at Chiripa, emphasizing the economic importance of the lakeshore edge. These water birds include coots, grebes, and ducks, as well as larger birds such as the flamingo and the Andean goose. Species characteristic of both shallow and deep water are well represented. Many of these could have been taken with nets, snares, or bolas. The land birds include tinamous, doves, and flickers which were common prey for hunters throughout the Andes, and several passerines. In addition to bird bone, bird egg shell is a common find in flotation samples, indicating that lake edge nests were robbed for food. Two bird specialists have visited the region around the site and noted the bird species and their habitats. Five of the birds represented among the bones from the 1992 excavations were not seen within 2 km of Chiripa during June-July 1996 field season (Kent et al. n.d.). These include a grebe, cormorant, goose, duck, and owl. As identifica- tion of the bird bones proceeds to more refined levels, this list of past birds is sure to grow. OTHER SMALL FAUNA Finds of small rodents, reptiles, and amphib- ians are rare in this assemblage. The many hundreds of smaller mammal bones are primarily from rodents. These include caviomorphs, espe- cially guinea pigs (two species of which poten- tially occur in the region; as well as cricetid rodents, such as the small mouse Akodon and a medium-sized mouse cf. Phyllotis). The small rodents include probable prey such as guinea pig- sized animals as well as owl prey such as mice. Ethnohistorically, larger rodents would have been taken in their burrows with sticks or snares. The excavators identified a number of complete rodent skeletons that have tentatively been identified as offerings or other ritual deposits (such entire or partial skeletons were recovered from loci 1194, 1276, 1292, 1487, and 1489). We cannot evaluate the possibility of the domestic guinea pig at this time. The amphibian bones consist of what seems to be one species of small to medium-sized toad, perhaps in the widespread Andean genus Atelopus. The meat of these toads may be edible although their skin is toxic. The only reptiles recorded are a 112 Earbly Settlement at Chiripa, Bolivia small species of lizard, perhaps Tropiduros sp., and a small colubrid snake, both of which are rarely represented in the bone at the site. ANIMAL BONE IN THE ARCHAEOLOGICAL CONTEXT Our preliminary analysis of camelid and large mammal body parts suggests that the sample does not represent a full range of butchering waste. Weights of bone elements recovered were compared to weights of those bones from a known modern skeleton (data from Mengoni-Gonalons 1991) to examine relative skeletal completeness of the archaeological sample (table 21). Using bone weight rather than count allows us to make use of fragmented and unidentifiable bone which may still be identifiable to body part. Bones of the neck, pelvis, and lower limbs (radius, carpals, metapodials, and toes) are over-represented compared to their abundance in a whole camelid carcass; and the bones of the rib cage (in particular the meaty brisket portion carried by the sternum), back and upper limbs are under-represented. Two interpretations are possible: first, that the missing bones were discarded in another portion of the site and were not recovered; and second, that the missing bones were so fragmented by food pro- cessing and gnawing that their remains were identified as "unidentifiable" even though they were in fact recovered in this sample. These two possibilities are not mutually exclusive. Relative to the value of the meat and fat that would be associated with those skeletal parts, both high value bones (ribs and sternum) and lower value (axis and atlas) are under-represented (utility indices from Mengoni-Gonalons 1991). The majority of the high-utility parts are under- represented, suggesting that high fragmentation in cooking and consuming has destroyed them. It is difficult to account for the almost total non- representation of some very dense and resistant elements such as the distal humerus and the astragalus, however, without invoking some agent of differential transport (the part being left behind during dismemberment or carried off during distribution). Larger and more representative samples from other parts of the site should shed some light on this pattern. Small percussion flakes and abundant long bone scrap do indicate that processing of bones for marrow, a typical house- hold activity, took place within the areas sampled. Marks interpreted as cut marks were re- corded based on visual inspection with the aid of a 10 power loupe. The surfaces of the bone were substantially intact, so this estimate of cut mark frequencies should be reliable (table 17). Nineteen cut marks were recorded, and comparisons with the rest of the assemblage reveal that 14% of the cervical vertebrae (n=3 bones with sets of cut marks recorded), 2% of the ribs (n=2), and 18% of the first phalanges were cut (n=3). Single inci- dences of cuts were observed on a thoracic verte- brae, a proximal radius, and a radial carpal. These were interpreted as marks left by skinning (radial carpal and phalanges) and dismemberment (verte- brae and radius). Seven cut marks were observed on the exterior of unidentifiable long bone shaft fragments; these cuts were likely to have been made while scraping long bones before cracking to remove marrow. In addition to these cut marks, remains of percussion fracture were also recorded (n=4). These sharp, hacked bits of bone debitage are thought to have been produced when bones are struck with enough force to crack open the medul- lary cavity of a long bone. In addition, many cut bones were removed from the animal bone assemblage because they were bone implements or the waste from produc- ing them. Most of the cut marks noted on bird bones (not part of the sample reported on here) were for producing bone beads. The same simple stone tool technology seemed to have been used to produce marks for butchery and for tool manufac- ture (with the marks on the bone tools being much deeper, repeated, and more carefully patterned). No indication of the use of metal tools on bone has been observed in this sample. While cut marks are the classic indication of ancient butchery, they are only a small part of the behavior revealed by fragmented bone remains. Bones may be transported from the sample, as inferred above. In addition, bones change size and shape with continued fragmentation during their use-life at the site. Bone fragmentation in a particular deposit may reveal a complex of impor- tant cultural steps (cooking, fat rendering, site maintenance), as well as non-cultural ones (tram- pling, gnawing, burning, weathering) (Moore 1997, Stiner et al. 1995). The uniformly small fish Herds, Fish, and Fowl 113 TABLE 20 List of species identified taken from Steadman's report and augmented. Calculation of representation of body parts. Family Species Common Name Habitat Fish Cyprinodontidae Trichomycteridae Amphibia Atelopidae Reptiles cf. Tropiduridae cf. Colubridae Birds Tinimidae Nothoprota sp. 2 (small) Podicipidae Rollandia sp. (small) Phalacrocoracidae Ardeidae Phoenicopteridae Anatidae Accipitridae Rallidae Fulica sp. 1 Fulica sp. 2 Charadriidae Laridae Columbidae Psittacidae Strigidae cf. Ciccaba sp. * Picidae Passeriformes Oscine spp. Mammals Dasypodidae Caviidae Cricitidae cf. Phyllotis Canidae Camelidae Lama vicugna Cervidae Orestias spp. Trichomycterus sp. Atelopus sp. cf. Tropidurus sp. Species 1 Nothoprota sp. 1 (large) Tinamou Podiceps sp. (large) * Grebes Rollandia cf. micropterum Phalacrocorax cf. brasilianus* Nycticorax nycticorax Phoenicopterus sp. Cloephaga Oxyura sp. * Anas sp. cf. Buteo Rallus sp. Coot Coot Vanellus sp. Larus serranus cf. Metriopelia cf. Bolborynchus Bubo virginianus cf. Colaptes Suboscine spp. Oscine songbirds Species 1 (small) Species 1 Akodon sp. Leaf-eared mouse Canis sp. Lama glama Vicuna cf. Hippocamelus Killifish Burrowing catfish Lake shallows Toads Lava lizards Harmless snakes Tinamou Quail-like, ground feeder Grebes Grebe Cormorant Black-crowned Night Heron Flamingo Andean Goose Ruddy duck Teal/pintail etc. Hawk Rail Nest along lake edge Typically, shallow water Flightless/ lakeshore & reeds Diving waterbird Reed beds Shallow water High lakes and marshes Lakes, including open water Lakes, ponds, streams Raptor Shorebirds Plovers, lapwings Open ground, marsh, lakeshore Andean Gull Breed in reedbeds Black winged dove Groundfeeding dove Rufous parakeet/Andean parakeet Horned owl Nocturnal raptor Flicker Songbirds Groundfeeding insectivore Armadillo Burrowing insectivore Guinea pig and related forms Many habitats including rock piles and reedbeds Small mouse Unid. canid Llama/Guanaco/Alpaca Taruca deer Puna grazers Moist puna grazers Grazer/browser *Bird species not recorded within 2 km of Chiripa during survey June-July 1996 by A. Kent and T.A. Weber 114 Early Settlement at Chiripa, Bolivia N- -- F- C-j rllCli 4 -4 1- I,) 'I- 10 'I OCtN 00 \, 4i 66ci ti 0 o * a c 00 00 0 ON 0 - o O _ --: N x ci4 - N I_ "tt ON - im CA - r- -t - ) M - ' N - 00- _ N IC i 0 o '(N 0 C m 0 0 0 0 - CtN 0o0 0 o i i- _r eq 00 0 "t O 7t 0 ON - crN 'O N 6 o0 - - N kNo0 r-N ,6 ci ci tc N ci C.0 00 00 00 ooi Lf N ON, o6 r 4 o cN u( (0N N - (?N m cr W) M r - M ?c ,t .c Tt c c oo ,-t cr 006 ~16 _ -0 00 t ON 0 00 C 'N m 0 ON 00 ci ?0 ' ( Ci Cic 00 e e-~ ~ - (O V) V) Ot " * oo - NO N _ _ N N m - -- r- O O - -- 4 -: tr wi tri 00 ON c i ON1 C f f a) C/- lq a) ~Z ~ a ) 0 o -z o) - t oo s. t N (N- W) - O N c C ci c N c i C N " - - c i ___ CZ~~~~~~~~~a & CZ ~ ~ W u ~ ~ ~ ~ C, a) a * 0 0 o1 * a- . E '. 0 r. - co aE) I Z a II2 C4 wii U , o 0t C t * a) aX - a) * a) B-u C Q o u a; Q c "0 al) "0 "0 t S c) 0 0 c- E *a_ a) H u: 'a E CZ S E E C) C0 0 0 06 (N ci a) C5 ON ON 9 0 * * a) EN ON. 9 0h 00 a) "0 S C.) * ci M O IN _a) a) a) 0u O O C C) 0Y au V: . a) 0 0 r.. 0 a) S 00 4- a. 0 a CZ 0 Herds, Fish, and Fowl 115 bone are excluded from consideration in fragment size analysis. Typically, the largest bone frag- ments are the least-modified, unburned, large bone scrap, those bones which are primary refuse from the dismemberment and preparation of the basic units of the carcass. Fragments in this sample ranged up to 0.5 kg for two camelid cervical (neck) vertebrae from Locus 1198, a pit with high overall density. Unburned fragments averaged 2.4 gm/fragment and burned bones averaged 1.8 gm/ fragment. This size difference is typical and is attributed to the friability of burned bones and the greater likelihood that smaller fragments will end up being burned. Even smaller fragments, which might reflect intense use of within-bone nutrients such as marrow, and/or intense site use and maintenance, occur with low density in the samples that have been examined so far. It may be that the examination of the flotation samples will reveal more of these fragments, but at this time, it appears that the intense domestic use and discard of large mammals is not represented in this portion of the site. On the other hand, intense use of fish remains is clearly attested to. Burned bones were separated into two categories: blackened and calcined (bluish, greenish, or whitened). While burning is some- times attributed to direct heat treatment in cook- ing, observations and experiment have shown that cooking rarely produces more than a faint singeing of the tips of bones. Such so-called "broiling" marks are not common in this sample. Fish bones at Santiago are rarely burned, suggesting that they were prepared by boiling or steaming. Calcined bone, the result of burning bone at the highest temperatures, is also rare in the Early Chiripa Santiago samples, but it does appear, often in deposits that had been identified as trash by excavators. Blackened burned bones are common, ranging from a few percent of the total in most units to almost 60% in Locus 1307, identified as fill around a rock wall. We feel that this blacken- ing comes mostly from contact with heat apart from cooking, perhaps in site maintenance or during use of some substantial heat feature. Units with no burned bone seem to be either those heavy in fish bones (loci 1199 and 1408) or from depos- its where accumulation of bone overall was very light (loci 1238, 1240, and 1496). Rodent and carnivore gnawing were noted on bones in several units. It is not possible to identify the specific agents involved. Elkin and Mondini (1996) demonstrated that both humans and foxes are capable of producing scratch marks on bone that would traditionally have been attrib- uted to dogs. Until the canids from the site can be more securely identified it is difficult to offer further interpretation. Neither carnivore- nor rodent-gnawing is inconsistent with continuous occupation of the site by humans, but such gnaw- ing probably did result from debris being left in open trash areas. Taken together, fragmentation, burning, and other modification can help support functional or contextual determinations made in the field; in other cases they may suggest a contrasting inter- pretation. The assemblages of bone from loci 1236, 1294, and l248 seem to have been formed by a variety of discard and modification events while the bone was exposed on the surface. Bones from these units were weathered and rounded, marked by carnivore and rodent gnawing, and had often been burned. Other loci had large fragments but much less burning and other modification (for example loci 1238 and 1240, identified as eroded gravel matrix) suggesting that those bones had had less chance to be worked over by humans and other agents before final burial. Locus 1305, which had been tentatively identified as an ephemeral burned area, was probably a dump of ash and debris from a hearth or heat feature at another location, as little of the bone was burned. The bone assemblage from this locus contained a distinctively high proportion of small, calcined, whitened bone fragments that result from high temperatures. Locus 1 199, another ash filled feature, had no burned bone at all, suggesting that the ash had been mixed with fresh bone after the burning event in the original hearth. While the assemblage of species is quite uniform across the deposit for the most part, occasional units have distinctive assemblages, in particular those heavy in fish bones. Such units as loci 1480 and 1499, which were extremely high in fish bone but low in burned fragments and larger fragments from mammal bone, may represent trash from households, or areas of the site, or meals, or even seasons, in which the staple fish was the only animal food. 116 Ear-/v Settlement at Chiripa, Bolivia ANIMALS IN THE CONTEXT OF THE CHIRIPA ENVIRONMENT These initial results indicate the potential complexity of analyzing the cultural ecology of the Titicaca lakeshore. This lacustrine zone has been a highly productive one for the native fauna and for the human occupants as well. The first level we can analyze is the intensive human exploitation of easy-to-collect fish, birds, eggs, and rodents along the shallow lake edge and totora reed beds. Second, we should take into account the human disruption of those resources with continued occupation and the alterafion of lakeshore terrain with fields, terraces, and drainage ditches. Third, we can speculate on the impact of camelid herds on local settlements and environ- ments, as camelids provided meat, fiber, and dung but put pressure on the supply of lake vegetation, field stubble, and natural pasture above the site. As our sample of later period remains increases, we hope to contribute to evaluating these three aspects of the Chiripa local economy through time.