A03-1000 "Kakapunku" [A03-1000 - A03-1013]

The site of Kakapunku [A03-1000] (a.k.a. C'oponeta) is located on the south side of the Río Llapa on a north facing bluff just upstream of the confluence of the Río Llapa and the Río Pulpera. This site consists of three rock shelters spaced about 15m apart along the base of a 10m high irregular cliff band that lies approximately 30m up the hill and south of the high-water line on the Río Llapa. There is a colluvial ramp below the rock shelters and on a terrace below that ramp lie number of lithic loci and some ceramics. The rock shelters are generally north facing so that they are oriented to the mid-day sun and the shelters are warm and well-lit.

ArchID	SiteID	FileType	Description	Area (m2)
1000	1000	Site_a	"Kakapunku"	3400.0
1004	1000	Lithic_a	High Dens, ~70 Obs	1771.3
1010	1000	Lithic_a	High Dens, ~70 Obs	76.7
1013	1000	Lithic_a	Medium Dens, ~70 Obs	148.6

Table 6-35. Loci in Kakapunku [A03-1000].

ArchID	Description	Height	Depth	Width	Width Entrance
A03-1001	Mortuary, looted. MNI=7	1.6	3.2	5.2	1.9
A03-1002	Domestic, warm rock shelter	4	3.5	6	5.1
A03-1003	Mortuary, looted. MNI=14	1.6	~8	1.4	1

Table 6-36. Rock shelters at Kakapunku [A3-1000], dimensions in meters.

Figure 6-26. Site of Kakapunku [A03-1000].

A03-1001 -East rock shelter at Kakapunku

This is a looted mortuary context. The rock shelter is moderately difficult to access on a cliff face, and the entrance appears to have been formerly walled off. The original rock shelter height is difficult to determine because cist tombs in the floor of rock shelter have been excavated by looters. A number of complete crania and long-bones were present. The crania and long bones were examined in the field by Mirza del Castillo, a team member who is a physical anthropologist. From her in-field visual assessment she cautiously suggests that a number of the crania had female metric proportions, and that a number of the crania appeared to have hypoplasia of the bone marrow on the posterior edge of the parietal bones which is a possible indicator of anemia.

Figure 6-27. The entrance of A03-1001 with looted cist tomb visible inside the rock shelter.

A03-1002 - Center rock shelter at Kakapunku

This rock shelter is relatively shallow but other characteristics made it appear to be to be an inviting residential location for a small group of foragers or pastoralists. The shelter is dry inside the dripline, north facing, flat interior, defensible, a short distance from water, and immediately south of a terrace for open-air activities where loci a03-1004 through 1013 are located.

A03-1003 - West rock shelter at Kakapunku

This rock shelter is a lava tube sloping upwards into the cliff face. It appears to contain many disturbed interments. Undisturbed human remains perhaps exist here but they were not visible on the surface.

Discussion

Kakapunku is another case of a multicomponent site where discerning the Archaic component is difficult. Five projectile points were identified belonging to different Archaic Periods, and combined with the ceramics (described later), there is a some affiliation with virtually all time periods at this site.

ArchID	Index	Material Type	Form	Type	Temporal
1001	1	Obsidian	Proj Point	3a	Early-Middle Archaic
1002	1	Obsidian	Proj Point	2b	Latter part of Middle Archaic
1004	1	Obsidian	Biface	n/a
1004	2	Obsidian	Biface	n/a
1004	3	Obsidian	Proj Point Broken	5	Late
1005	1	Volcanics	Core	n/a
1007	1	Obsidian	Proj Point	3b	Middle Archaic
1008	1	Volcanics	Proj Point	4d	Late Archaic
1009	1	Obsidian	Preform	5a	Late
1010	1	Obsidian	Proj Point Broken	5d	Late

Table 6-37. Selected Lithics from Kakapunku [A03-1000].

On the terrace below the rock shelter [A03-1002] the largest concentration of artifacts were noted. Samples collected from the medium and high density lithic loci on the terrace are approximately 50% obsidian materials with the non-obsidian primarily consisting of aphanitic volcanic stone. The northern high density locus, A03-1010, is eroding down the cutbank caused by erosion and it is possible that this terrace was significantly larger but has been truncated by fluvial erosion of the terrace. In sum, Kakapunku appears to have served as a regular residential area with three relatively small rock shelters.

A03-985 "Quelkata"

This rock shelter lies just below the confluence of the Río Llapa and Río Pulpera along a sweeping bend in the river channel where the Río Llapa first approaches the Castillo de Callalli formation. This rock shelter is well situated and probably served as a residential base through most of prehistory, however surface collections by Jose-Antonio Chávez Chávez (1978) for his Bachelor's thesis revealed primarily preceramic occupation levels with late, Series 5 projectile points. By the time Chavez was able to systematically investigate the materials in the rock shelter it had been largely destroyed in 1976 by the road construction crews associated with the Majes Project (Chávez 1978: 3). The rock shelter was dynamited because it is unfortunately located precisely at a point of the Río Llapa where the project planned to build a bridge on the road to the Condoroma reservoir. What remains today of the rock shelter is a concavity in the tuff with a shelter dry zone measuring approximately 150m wide inside the dripline. Chávez systematically collected and analyzed projectile points from this rock shelter but, curiously, the artifactual evidence available shows an occupation sequence going only as far back as the Terminal Archaic.

Figure 6-28. Cueva de Quelkata was dynamited by Majes Project road crews in 1976. Two people are visible for scale inside the shelter to the left of center.

Figure 6-29. Quelkata in 1975 redrawn from Chavez (1978: 20). Red dotted line shows estimated border of the portion of the rock shelter that remains after the dynamiting for road widening.

Description

The rock shelter was significantly larger in the recent past. Chavez (1978: 20-21) estimates that the dry portion of the rock shelter was 7m deep, 50m wide, and 2.65m high at the mouth. The sheltered area amounted to a 365 m²area with the bulk of cultural debris concentrated in the south portion of the rock shelter. Chávez estimated that the stratigraphy included up to 3m of cultural material. During the 2003 season a great abundance of cultural material was not observed in the slopewash between the highway and the river, as one might expect if the highway were constructed on such quantities of cultural fill, but the area has been much disturbed in recent years.

Figure 6-30. Quickbird satellite image of Quelkata area in 2005. The shelter is under the tuff outcrops immediately to the left of the bridge. Data courtesy of Google / Digital Globe.

Quelkata is located at a major cross-roads in the local transportation network and it has probably been a significant landmark among people in the Upper Colca since the earliest human occupation of the region. The rock shelter mouth opens east-north-east and deposits still evident in the back of the rock shelter are approximately 5 vertical meters above the river and 2.5 meters above the modern highway. As the rock shelter is on a bend in the river the shelter appears to have been scoured out of the outcrop of welded tuff by the waters of the Río Llapa until the river channel was sufficiently incised that it not longer entered the rock shelter proper. Recently Noble (2003) collected a K-Ar sample from precisely this location and the resulting age of 20.7±0.6 million years indicates that the formation dates to the Early Miocene.

Geographical attributes of this location can be summarized as follows.

(1) Quelkata is a large, sheltered site close perennial to water source that probably provides year-round fishing opportunities.

(2) The east-north-east aspect results in warming by the morning sun.

(3) It is located at major intersection in local travel routes which suggests that many were familiar with the site, but also exposes the occupants to anyone traveling in this portion of the valley.

(4) Summits of the "Castillo de Callalli" tuff outcrops immediately above are defensible and appear to have been fortified during the LIP.

Figure 6-31. A 65cm column of cultural deposits still exists at the back of Quelkata at a height of approximately 1.8m above datum (at head height between the gravels and the tuff in photo).

Preservation of Quelkata

Road crews associated with the Majes Project dynamited the rock shelter in 1976 but this was not the first important disturbance in this high traffic area. Significant impacts on preservation at Quelkata are the following

(1) Although the rock shelter lies approximately 5 meters above the dry season water level of the Río Llapa, the rock shelter has probably been scoured repeatedly by high water in past millennia. Quelkata is located on the outer edge of river bend precisely at a channelized portion of the river where 1000 year or even 100 year floods potentially entered the rock shelter and disturbed the contents.

(2) As noted by Chávez (1978: 17) a road has long existed in this area due to the fact that it is a crossroads location and, as a channelized portion of the river, it probably has had bridges prior to the modern bridge construction. With the mining history in the Cailloma area, a road was likely constructed to connect this ore-rich province with the Sumbay railroad sometime in the early 20^thcentury. The dynamiting by the Majes crew was intended to widen the road at the bridge and the intersection, and it significantly impacted the rock shelter. However, the downslope colluvial outwash from the rock shelter, which are frequently deposits of great archaeological interest from a rock shelter, had long been disturbed by road construction and by fluvial erosion prior to the destruction by the Majes crew.

(3) The construction crews dynamited the site in 1976 and largely destroyed the southern portion of the rock shelter, the portion that Chávez (1978) describes as having the deepest deposits. There are still significant deposits at the back of the rock shelter, however. As part of the archaeological reconnaissance of the Llapa valley materials systematically surface collected from the rock shelter as will be described below.

(4) The intersection and bridge at Quelkata was a principal junction in the Arequipa highway network until 2002 when the highway connecting Arequipa with Juliaca was paved. Until the asphalting of the Arequipa-Juliaca highway virtually all of the highway traffic between Arequipa and Cusco passed directly in front of this rock shelter where one of the few restaurants and gas stations along the route was located. Thus, a relatively high number of travelers probably visited the rock shelter and its surroundings while waiting for transportation. Fortunately, the dynamiting of the shelter created a sheer 2 m wall so that a brief segment of 4^thclass rock climbing is required to enter the rock shelter, discouraging casual visitors.

There are still cultural deposits in a broad but thin and shallow section at the back of Quelkata. While much of the strata along the back of the rock shelter are disturbed sand and river gravels, a cultural level approximately 60cm deep is found across a swath approximately 40m wide at the back of the rock shelter.

Figure 6-32. Profile of extant deposits at Quelkata shown in terms of relative collection units.

Proveniencing

As GPS cannot function in rock shelters, a collection strategy was devised using a datum inside the rock shelter. First, the shelter area was mapped expediently using tape and compass from a semi-permanent datum established on large rock on the north end of the occupation area near the road cut. Artifacts were surface collected from Quelkata in four horizontal collection units that were given ArchID# 986, 987, 988, and 989 to conform to the horizontal collection strategy throughout the survey. The deposits in the back of the rock shelter occurred on natural shelves and the vertical provenience was retained by collecting these shelves from the top (level 1) to bottom (level 5). The lowest level, the shelter floor, appeared to contain a mixture of materials that had cascaded down from above.

Site Collection

In the 2003 fieldwork season 22 flakes and retouched flakes from the rock shelter were collected. These consisted primarily of chert, chalcedony, and andesite. Of the 22 flakes 10 (45%) were made from obsidian, but these obsidian artifacts amount to only 7% of the flaked stone artifacts by weight. A number of bones and goat horns were among the faunal materials collected from the upper levels of the rock shelter, above the rock ledge. It is possible that the horns were placed in this location to dry out.

Chávez Collection

In his Bachelor's thesis José-Antonio Chávez (1978: 52-70) describes and analyzes surface collections that he conducted shortly after the dynamiting of the cave. Chávez includes photographs and drawings of projectile points that appear to be primarily of Type 5B and 5D forms. Chávez reports that he collected 36 projectile points and that the largest group was the concave base form (n = 17) that are probably Type 5B or Type 5D, diagnostic to the Terminal Archaic through the Late Horizon. Of these concave base forms 15 of them were made from obsidian (88%) and the remaining two were chert. Pressure flaking was noted on 16 of the points (Table 6-38).

	Obsidian	Non-obsidian	Totals
Unmodified	434	150	584
Modified	121	57	178
Total	555	207	762
Percent	72.8%	27.1%	100%
P.Pt Base: Convex	7	1	8
P.Pt Base: Flat	8	0	8
P.Pt Base: Concave	15	2	17
Probable Concave Base	12	0	12
Cores	2	3	5
Flakes - cortical	4	12	16
Flakes - noncortical	141	93	234
"Blades"	57	40	97
"Choppers"		5	5

Table 6-38. Counts of Material types by Artifact Form from 1977 Quelkata surface collections (Chávez 1978: 50-72).

A relatively large proportion of the points are non-diagnostic using Klink and Aldenderfer's typology. Three points described as "convex base" appear to be type 3D, a style which persists throughout the Archaic. Chávez notes that a large percentage of the artifacts were made of obsidian, although the identification of the Chivay source, 20 km to the south-west and 1000 meters in elevation above Quelkata, would not occur for another 15 years. Chávez attributes the relatively large percentage of obsidian to the many very small flakes of obsidian that were included in his sample. Had his analysis included weight of artifacts, the relative presence of obsidian by weight would have been considerably lower than that of other material types. It is apparent that obsidian was being transported to Quelkata and worked at the rock shelter from one of the Chivay obsidian source exposures in the study region.

Another notable pattern in these data is the lack of non-obsidian projectile points. This pattern is consistent with the predominantly base-notched (and Series 5) nature of this assemblage. However, results from the 2003 fieldwork show that this is anomalous in the region, even for the Terminal Archaic and onwards.

Discussion

Several of the projectile points that were noted on the top level of the rock shelter (in unit A03-988-level I) during a preliminary visit to the shelter in 2001 were gone from the rock shelter by the fall of 2003. The points required more careful study, but unfortunately this did not occur, and it was hoped that the points would have been retrieved during the course of the 2003 fieldwork.

It is notable that, despite the apparent importance of Quelkata in the regional settlement pattern, there were relatively few Archaic projectile points found in this rock shelter. Based on the evidence from Chávez and the 2003 examination of the rock shelter, the only definitive projectile point styles encountered at Quelkata are from the Terminal Archaic and onward. One parsimonious explanation for the lack of earlier projectile point styles is that the location of this rock shelter, only 5 m above the dry season water line, was insufficient to prevent particularly intense flooding during the Terminal Archaic which removed evidence from earlier time periods. Given data encountered at the site, this description of the Quelkata rock shelter perhaps belongs in the Early Agropastoralist discussion rather than the Archaic Foragers section of this dissertation. However, due to the centrality of this rock shelter in the settlement pattern, the shelter was likely a prominent settlement throughout the prehispanic period and thus it was included here.

A02-3 "Mollepunco"

The rock shelter of Mollepunco (aka Mollepunku, Cueva de Pirita) is one of the principal tourist destinations in the Upper Colca area (Cardona Rosas 2002: 143-147;Hostnig 2003: 52;Linares Málaga 1992: 287, 290, 291, 298) and it probably receives visitation by about ten groups of tourists per week during the high season. Formerly known as "Pirita" the shelter was renamed "Mollepunco" by Eloy Lináres Malaga in 1984 (1990;1992). This dry rock shelter at the base of a partly welded tuff outcrop consists of two principal cavities and with flat floors and sloping ceilings, and a dry outdoor activity area. The principal cavity is 7m deep, 4m wide, and 5m tall. The shelter is approximately 100 vertical meters above the Pulpera valley floor and it is adjacent to a small perennial stream. The shelter features pictographs and petroglyphs of camelid, bird, and humanoid figures that Augusto Cardona (2002: 143-147) argues are significant because they show the transition from pictographs to petroglyph techniques, and the art is believed to have been executed by herders and is therefore no older than 3000 BCE. The art has been interpreted as dating to relatively late periods because plump camelid figures are often shown in herds, and one solitary camelid has what appears to be a rope around its neck. The shelter was excavated by Linares (1992) and despite the gate blocking the entrance of one sector, it appears that the shelter was significantly looted. An open burial chamber probably dating to the LIP is evident high in the southern end of the rock shelter.

Figure6-33. Cueva de Mollepunco (A02-3).

Figure 6-34. Petroglyph of camelid on wall of Mollepunco.

Lithic debris is predominantly chert but a significant portion of the flaked material is obsidian. Bone and LIP ceramics are also evident at the site, perhaps resulting, in part, from the looted burial. Given the heavily visited and previously collected nature of this site surface the 2003 Upper Colca project field crew did not attempt a systematic recording of surface artifacts.

A03-1074 "Pokallacta"

This is a small, sheltered site that has evidence of occupations in both the Archaic Foragers period and subsequent times. The site is immediately to the west of a steep tuff outcrop approximately 100m high. The site is in the shadow much of the day, but it is a sheltered location that is protected and yet offers large views of the river valley.

The rock shelters are fronted by a flat area that today is a corral. This flat activity area is mapped as two Structure Loci though they are essentially one terrace. The two loci delimit the artifact spill zone for each respective rock shelter and each structure locus was assigned the same ArchID number as the rock shelter above it.

ArchID	SiteID	FileType	Description	Area (m2)
1074	1074	Site_a	"Pokallacta"	4,342.7
1075	1074	Struct_a	Southwest shelter	63
1076	1074	Struct_a	Northeast shelter	45

Table 6-39. Loci in Pokallacta [A03-1074].

ArchID	Description	Height	Depth	Width	Height At Back
1075	Residential, pastoral	2.5	9	10	1.2
1076	Residential	2.2	6	5.3	0.9

Table 6-40. Rock shelters at Pokallacta [A3-1074], dimensions in meters.

ArchID	Index	Material Type	Form	Type	Temporal
1074	1	Obsidian	Proj Point Broken	5d	TA - LH
1074	2	Obsidian	Biface Broken	n/a
1074	3	Obsidian	Proj Point Broken	5a	Term Archaic
1074	4	Volcanics	Biface Broken	n/a
1075	1	Chalcedony	Proj Point	5	TA - LH
1076	1	Obsidian	Proj Point	5b	TA - MH
1078	1	Obsidian	Proj Point	2a	Early - Middle Archaic
1080	1	Obsidian	Proj Point	3d	Archaic

Table 6-41. Selected Lithics from Pokallacta [A03-1074].

The bifacially flaked items are largely obsidian but notably the bulk of the grab sample of unmodified flakes was of chert and chalcedony. This reflects the location of the site on the north side of the Río Llapa, and is consistent with a raw material use pattern observed elsewhere on the survey.

A03-1075

This rock shelter has a very thick dung cap and it is difficult to see the ground. The slanting roof and open west side provides a large but only partly sheltered area. It appears that a protective wall was constructed on the west side in the past.

A03-1076

This rock shelter is the more protective residential shelter of the two, offering a view of the river valley and the surrounding terraces.