Summarizing the results presented in Chapters 6 and 7, a number of changes in obsidian procurement at the Chivay source correspond to changes recognized regionally in south-central Andean development. While the activities at the Chivay source area did not exactly meet the expectations of the project that were implied by patterns of regional consumption data. The differences between the expectations and the results of the Upper Colca project are informative, and these differences will be highlighted in the text. The differences appear to be largely the result of a historical bias in research, in that there has been a predominant focus on the more socially complex, later episodes in Andean prehistory. For example, obsidian has been found in many ceremonial Tiwanaku contexts and in the Tiwanaku core area such that some investigators in the heartland area argue that the Chivay source was state-controlled (Giesso 2003). Yet, in numerous studies in the Colca region, and this study of the Chivay source in particular, archaeologists have yet to find definitively Tiwanaku artifacts in the Colca area. Contrasting with this lack of evidence from the Tiwanaku period, consider the evidence from earlier periods: the Terminal Archaic and Early Formative. While relatively few sites have been excavated that date to the Archaic / Formative transition in the south-central Andean highlands, the evidence from the Chivay source shows that obsidian production intensified during that time. Indeed, a review of the little evidence available from Terminal Archaic and Early Formative sites in the adjacent consumption zone indicates a spike in obsidian consumption during that time.
In other words, archaeologists in the south-central highlands have historically focused on late prehispanic complex polities, and today the legacy of data from excavations aimed at regional centers has biased the overall picture of obsidian consumption towards the later Prehispanic period. While there were major social and economic changes underway in the third millennium BC, the findings here reveal that obsidian was one of the earlier materials to show a distinct increase in production and circulation in the south-central Andes.
During the lengthy Forager period there were a number of gradual changes that may be observed in obsidian procurement and circulation. Obsidian appears to have been procured directly from the Maymeja area since at least the Middle Archaic. Evidence from the dating of a glacial moraine, and from obsidian consumption at the distant site of Asana, suggests that the Maymeja area itself was glaciated until approximately 9000 cal BCE.
The Ob1 material (obsidian free of heterogeneities) appears to have been used preferentially for tool production over the Ob2 material (obsidian with heterogeneities) throughout the Forager period, as was true in later times. During the Late Archaic, obsidian use for projectile point production declined and points were chiefly made from fine-grained volcanic stone, such as the andesite material common in Block 2.
Establishing the degree of embeddedness of obsidian procurement during particular periods of the Archaic Forager time frame is not a straightforward task. If evidence of advanced reduction stages is taken as a sign of more embedded procurement, then the evidence for advanced reduction comes in several forms.
(1) There was a greater proportion of diagnostic points identified for series 1-4 points than series 5 points in Block 1.
(2) Most of the diagnostic points identified in the Block 1 area are incompletely flaked and were broken longitudinally as would occur during production.
(3) The bases of latitudinally snapped, non-obsidian projectile points were encountered that appeared to have been discarded in the Maymeja area. The implication is that as the hafts were reused for newly-made obsidian projectile points, the broken bases were discarded from hafts that were being retained.
Embedded procurement may have taken the form of blank production with later advanced reduction occurring in adjacent blocks, such as Block 2 that contains stronger evidence of Archaic residential occupation.
Elsewhere in the study region more distinct site size differences are apparent during the Archaic Foragers period. Operating within the confines of surface survey information, sites were differentiated into residential bases and small logistical sites based primarily on lithic artifact density and diversity, shelter opportunities, and other site features. Archaic occupation in the high puna area of Block 2 appears to have been more long-term. Dense scatters, and variability in lithic materials were encountered, along with identifiable series 1-4 projectile points at a various late stages of manufacture and use. The sites with an abundance of flakes of fine-grained volcanic stone all have a Late Archaic component, and given the predominance of fine-grained volcanic material in the production of Late Archaic point types, the presence of andesite and rhyolite flakes at sites in Block 2 appears to correlate most strongly to Late Archaic occupation. Archaic Forager sites with no pastoral component rarely had obsidian flakes, and obsidian use appears to have been relatively diminished. This lack of obsidian flaking debris in Forager sites is further supporting evidence for embedded procurement and source-area tool production during the Archaic Forager times.
Numerous projectile points diagnostic of the Archaic Forager period were found in Blocks 4 and 5, but the surface of these sites was dominated by evidence from later reoccupation and therefore it is difficult to isolate these lithic concentrations as belonging to the non-pastoral or pastoral period. Rock shelters in Block 5, like Mollepunku and Kakapunku, appear to have had high rates of reoccupation throughout prehistory.
The Block 3 area around Callalli has the lightest evidence of Archaic Forager period occupation. Chert production sites on river banks, and some small, isolated sites in sheltered areas, provide the bulk of the evidence in this region. Perhaps the paucity of the Archaic Forager component in Block 3 reflects the better hunting opportunities elsewhere, such as on the high puna, and better vegetation and gathering opportunities in the lower altitude Colca valley downstream. It appears that the Callalli area was relatively lightly occupied during the Archaic and perhaps the more specialized ecology of the puna above and the lush valley below were more conducive to Archaic Forager subsistence. The sites identified here appear to have been the result of short-term stays in the course of travel through the river valley.
Obsidian procurement in the Early Agropastoralist period took place in the context of sweeping changes in economic and political circumstances in the south-central Andes. A marked increase in interregional exchange was complemented by greater sedentism and possible circumscription, an increased reliance on food production, and early evidence of social ranking in the form of differences in supra-household architecture and grave goods. The social and economic integration of altiplano communities through networks of regular camelid caravans has been proposed as an important preliminary stage in early economic coalescence in south-central Andean highland communities prior to the emergence of regional polities of the Late Formative (Browman 1981;Dillehay and Nuñez 1988). While an increased reliance on camelid pastoralism is evident from regional settlement patterns and faunal assemblages, the date of the onset of the system of regional caravan networks is difficult to establish.
The existing evidence from diagnostic ceramic distributions suggest that regional integration through regular caravan traffic occurred when particular styles became more widely distributed, such as Middle Formative Qaluyu pottery, or earlier with Early/Middle Formative fiber-tempered pottery in the southern Titicaca Basin. However, the regional styles of projectile points changed dramatically in a time period that is earlier still: the Terminal Archaic (3300 cal BCE). While changes in projectile point styles are likely connected to technological modification, such as the use of bow and arrow and differences in hafting, the widespread adoption of the series 5 projectile points throughout the central and south-central Andes from the Junin puna to northern Chile suggests that interregional contact was pervasive.
Evidence showing that quarrying and workshop activity at the Chivay source began during the latter part of the Terminal Archaic, suggests that the regional economy was already becoming relatively integrated and responsive to non-local demand. First, the evidence from this period shows that production targeted larger nodules that served to signal surplus and abundance, and would retain greater use-life with distance from the source. Second, the quarrying activity appeared to focus on obtaining the Ob1 obsidian nodules of clear or grey material. Finally, the date inferred for the construction of a swept path out of the Chivay source quarry area appears to coincide with production at the source; although the specific dating of the construction of the road is tenuous. Changes in the regional demand, the responsiveness of procurement, and the ability to transport obsidian, are part of a larger suite of changes during the Terminal Archaic and Early Formative, and this production system will be discussed below in the context of regional integration during that time period.
The Early Agropastoralist period in Block 2 shows evidence of growing pastoral herds and the construction of a variety of well-established animal control structures. These Block 2 facilities, many of which are abandoned today, are notable in that a variety of ceramics and lithic artifact styles are scattered on the surface of these sites. These pastoral structures may be interpreted as corrals, or some other animal control structure but with additional ritual functions. For example, at the site of Pausa, three large ovals varying in size from 20m in diameter to 50m in diameter were examined. These structures are delimited by solidly placed large rocks between 30-100cm across and the ovals have several curious attributes: (1) the ovals, regardless of size, all have 26 to 28 large rocks, (2) niches along the east side of the largest oval suggest that this structure was more than merely a corral, (3) the quantity of obsidian and ceramics at the site of Pausa far exceeds similar configurations elsewhere in Block 2, suggesting a particularly important function for this site. The 2003 testing at two of the Pausa structures revealed what appears to have been a domestic hearth dating to the Late Formative, around AD200, and the construction of a secondary structure and platform at the end of the Late Formative around AD400. The evidence from surface ceramics distributions suggest that these Late Formative dates are linked to a number of those corral mounds. Advanced-stage obsidian reduction is found in abundance at these mound structures, which implies that the people pasturing the expanding herds of camelids in Block 2 were involved in obsidian tool manufacture, but the presence of exclusively small, advanced stage flakes (never greater than 4 cm) implies that these people were not responsible for the initial quarrying of large nodules since the evidence shows that they never discarded large flakes or cores in Block 2.
In the upper Colca drainage in Block 3, the Early Agropastoralist evidence is less defined than was the occupation in the Block 2 puna area. The quality of pasture is known to be considerably better, particularly for alpacas, above 4200 masl. Several sherds of non-local ceramics with similarities to Middle Formative (Qaluyu) and Late Formative (Pukara) in the Lake Titicaca Basin were encountered in the course of work in this area which suggests that the Callalli area was a major cross-roads and thoroughfare, as it is today. Obsidian was found primarily in the form of finished bifacial tools, but some obsidian production was encountered in Terminal Archaic contexts at the rock shelter of Quelkata, and at the open air site of Taukamayo. While the test unit at Taukamayo produced Tiwanaku period dates (circa AD 650), many of the ceramics that were observed in the landslide debris appear to date to the Formative.
Evidence from regional consumption patterns indicate that while Chivay obsidian was in wide circulation within the Tiwanaku economic sphere, the regional distribution and quantity consumed declined during the subsequent LIP and Late Horizon. As one of the research questions motivating this study at the Chivay source, evidence was specifically sought to explain the economic circumstances around obsidian production in Late Prehispanic contexts. New evidence will be reviewed that was acquired from fieldwork in 2003, and later in this chapter contrasting models of production will be explored in more detail.
No direct evidence of Tiwanaku was encountered at the Chivay source or in the immediate vicinity during fieldwork in 2003. A single obsidian preform was encountered that appears to have been an early stage of a type 4E (Tiwanaku stemmed) projectile point in Block 2, but the identification is not definitive. Titicaca Basin materials were encountered deriving from particular time periods in the form of non-local decorated ceramics, such as sherds of possible Qaluyu and Pukara (Formative) styles, and Colla and Chucuito (LIP and Late Horizon) ceramic styles, but no Tiwanaku pottery was encountered any where in the Upper Colca research area. In Block 3, a calibrated radiocarbon date of A.D. 650 - A.D. 780 (Figure 7-3) from Taukamayo places the occupation in the Tiwanaku period, but no Tiwanaku or Wari evidence was found at the site. Rather, pottery akin to the local Chiquero material was encountered that is diagnostic to the Formative in the main Colca valley (Wernke, 2003). The asymmetrical relationship between Chivay obsidian consumption in the Tiwanaku economic sphere, and the lack of Tiwanaku diagnostic materials in the Chivay source area will be discussed in more detail below.
Local Middle Horizon patterns are of particular interest because regional distributions of these ceramics may shed light on the nature of the frontier relationship between Tiwanaku and Wari that appears to have been occurring in the upper Colca valley during the latter half of the first millennium AD. Ceramics in the local Middle Horizon style (Wernke 2003: 466-478) were encountered in Block 2, but these sherds were confined to the northern half of Block 2. Middle Horizon pottery was found scattered throughout Block 3. A single local Middle Horizon sherd was found mid-way between the Chivay source and the town of Chivay, and one sherd was found in Block 6 upstream of Tuti. In short, the Colca Middle Horizon type defined by Wernke was encountered throughout the Upper Colca study area although, as discussed in Chapter 6, there is a distinct zone in Block 2 south of which the local MH pottery is not found.
Ceramics of the local Collagua style dominate the decorated ceramics found throughout the Upper Colca survey area. Pastoral facilities are rich in decorated ceramics resembling the main Colca Valley styles described by Wernke and others and underscoring the close links between Collagua polity in the Colca valley, and their expanding herding sector on the adjacent puna. Sherds of the Colla type, the Lake Titicaca Basin LIP style, were found in the northern part of Block 2 and also along the trail leading from Block 2 towards Block 1 and the Colca valley. Relatively close economic and cultural links are expected between LIP groups in the Colca and the Titicaca Basin as these groups shared a number of traits including the Aymara language, chulpamortuary architecture, and the defensive pukarahilltop settlements. Yet, interestingly, no Colla sherds were encountered in Block 3, the upper Colca valley, although the ceramics and architectural evidence shows that the Collagua presence was relatively strong in this area. This may be explained in terms of ecological complementarity, as one might expect the Colla to have been actively trading with the lower Colca valley agriculturalists, but the upper Colca valley is at the same elevation as the Titicaca Basin and thus it had relatively little to complement the goods available in the Titicaca Basin.
The LIP occupation of Block 3 appears to have been focused around agricultural plots and pastoralism in the adjacent high puna. Evidence of abandoned agricultural plots was noted both upstream and downstream of Callalli (3,900 masl) with furrowed areas that were likely planted in frost resistant crops like quinoa or tubers.
During the Late Horizon the reorganization that took place under Inka rule is evident in settlement pattern changes and ceramics distributions. At the Chivay obsidian source, the highest quantity of diagnostic ceramics were those in Late Horizon styles. These ceramics may be derived from the mortuary structures that were encountered in the Maymeja source area, or the LH ceramics may be a reflection of the increased investment in water control projects and expanded herding. A possible extension to the Huarancante canal capturing water from Quebrada de los Molinos would have had its origins precisely at the Maymeja workshop at the Chivay source.
Elsewhere in the survey area the Inka period introduced a number of changes including distinctive differences in long-distance interaction. In Block 2, the Late Horizon component appears to have been consolidated around a few larger sites. The presence of Titicaca Basin pottery is notably lower in the Late Horizon in Block 2, and Cusco Inka styles are introduced in that area. Fourteen sherds of Colla style pottery (Titicaca Basin LIP) were found in Block 2, and these were primarily scattered in the northern half of the block. By the subsequent Late Horizon only two Chucuito style sherds (Titicaca Basin LH) were found in Block 2. The most common decorated pottery was the local LH Collagua-Inka and the Collagua-3 style. In Block 3, settlement became concentrated at agricultural plots along natural river terraces near the confluence of the Río Colca and Río Llapa in an area known as Callalli Antiguo. A concentration of structures with non-local Inka pottery along the principal road system on the south side of the Colca suggests that this site may have had administrative functions in the Late Horizon settlement pattern.
This review of the major findings from research in 2003 shows that, in general, it was regional forces that dominated the changes that were observed in obsidian procurement through time. Raw material for use in the local economy was relatively steady, with local cherts dominating Block 3 assemblages in most time periods, and local fine-grained volcanics prevalent in Block 2 during the Late Archaic. However, changes on the regional scale in long-distance interaction and in demand for obsidian beginning in the Terminal Archaic introduced distinct intensification in procurement of obsidian. The later changes that occurred in the Early Pastoralists and Late Prehispanic times reflected the growing intensification in both agricultural and pastoral economies in the Colca. In contrast, the procurement and circulation of Chivay obsidian is relatively sustained and is not subject to this expanded production in later times.