(1) Larger nodules could be acquired through quarrying.

Energy was evidently expended in digging to extract obsidian at the quarry pit in Maymeja and the motivation behind this effort an important question. Looking at the general patterns over the entire study area is instructive because it reveals some general patterns that were counter to expectations of this research project. If the original nodules from the Block 1 area were of a larger size it may account for the quarrying activity observed in Maymeja.

	Points and Tools			Cores			Simple Flakes			Total
Block	No.	mLn	sLn	No.	mLn	sLn	No.	mLn	sLn
1	4	38.2	10.0	110	48.5	8.7	88	38.7	14.9	202.0
2	3	27.0	4.7	8	38.9	7.7	26	23.8	10.9	37.0
3	3	27.8	2.4	9	41.9	8.0	14	26.3	11.1	26.0
4	-	-	-	33	41.6	7.5	34	34.9	12.2	67.0
5	1	28.0	-	11	36.9	11.8	82	22.0	9.9	94.0
Total	11	31.8	8.2	171	45.8	9.5	244	30.2	14.5	426.0

Table 6-6. Lengths of complete obsidian artifacts with > 30% cortex by survey block surface collections, showing means and standard deviations.

In Table 6-6 the length of cortical obsidian artifacts from surface collections from throughout the study region with a minimum threshold of 30% cortex are shown. It should be noted that tools and flakes must have 30% covering of dorsalcortex to qualify for this table, whereas cores need only have 30% cortex anywhere on the exterior surface to be included here.

The first indicator of differential activity in Block 1 is the sheer number of cortical cores in these surface collection data. Table 6-6 reveals that the mean size of cortical artifacts in all three technical classes is notably longer in Block 1, and that the second largest mean lengths are from Block 4, adjacent to the Maymeja area. The artifacts from Block 2 are surprisingly small considering that the Chivay source is only one day's travel from this location. The fact these cortical artifacts are small suggests that they did not have access to large starting nodules. Table 6-6 also shows that the cortical artifacts from Block 3 were slightly larger than one would expect considering that Blocks 5 and 2 are equidistant, if not closer, to the Maymeja area of the obsidian source than is Block 3.

(2) Obsidian from the quarry pit had more transparent coloration.

In the course of this research it was observed that there is variability in the transparency and color of culturally-modified obsidian throughout the study area and the spatial distribution is related to the appearance of natural obsidian in geological contexts. Chivay obsidian is renowned for its clarity and it is possible that material from one survey block had a greater frequency of transparency than did material from surrounding survey blocks. If clarity was a desirable characteristic of obsidian artifacts then the evidence may show a greater focus on production in areas where transparent obsidian was common. Much of the obsidian contained dark banding as well, and this banding was more visible with clear obsidian than when the obsidian matrix was a darker coloration.

Figure 6-2. Proportion of obsidian for four colors (shades) of glass, by count.

	Black		Clear		Grey		Other
Block	No.	%	No.	%	No.	%	No.	%
1	13	4.0	245	74.9	66	20.2	3.0	0.9
2	1	0.6	111	66.9	40	24.1	14.0	8.4
3	3	2.7	71	63.4	33	29.5	5.0	4.5
4	12	8.2	79	54.1	55	37.7		0.0
5	5	2.4	141	66.5	61	28.8	5.0	2.4
6		0.0	7	33.3	14	66.7		0.0
Total	34	3.5	654	66.5	269	27.3	27.0	2.7

Table 6-7. Obsidian artifact color (shade) by survey block surface collections. Includes obsidian with bands and without bands.

As shown in Table 6-7, Block 1 indeed has a higher fraction of clear obsidian than other blocks in the survey. The high incidence of clear obsidian at Block 2 (67%) further suggests that whoever was quarrying and reducing obsidian at the A03-126 workshop in Block 1 was also associated with the settlements in Block 2, as there is no naturally occurring obsidian in Block 2. As Block 2 is on the direct transport route towards the Lake Titicaca Basin, this evidence suggests that some of the clear obsidian from Block 1 was being consumed in Block 2 en route to the larger consumption zone of the south-central Andean highland region and the Titicaca Basin where Chivay obsidian was purportedly prized for its clarity. It should be cautioned that the distinction between grey and clear obsidian appears to be correlated with thickness. That is, a "clear" artifact is more likely to be considered "grey" if it is thicker because it appears to be less transparent as a result of thickness.

(3) Block 1 had more homogeneous Ob1-type obsidian.

A further line of inquiry relates to the question of presence of heterogeneities in the obsidian. The Block 1 area with both the quarry pit, and the greatest abundance of large nodules, did not entirely consist of Ob1 homogenous obsidian. Investigating all the artifactual obsidian collected from surface contexts in the course of this project by survey block, a number of general patterns emerge.

Figure 6-3. Proportion of obsidian material as Ob1 and Ob2 (heterogeneities), by count.

	Homogeneous: Ob1			Heterogeneous: Ob2
Block	No.	%	m%Cortex	No.	%	m % Cortex	Total
1	355	93.2	36.6	26	6.8	35.4	381
2	329	90.6	6.1	34	9.4	15.9	363
3	96	64.9	15.4	52	35.1	10.2	148
4	140	73.7	23.0	50	26.3	31.2	190
5	172	73.2	30.8	63	26.8	22.5	235
6	17	77.3	1.2	5	22.7	2.0	22
Total	1109	82.8	22.6	230	17.2	21.7	1339

Table 6-8. Obsidian artifact material type by Survey Block surface collections.

Table 6-8 reveals that a small percentage of Ob2 material was actually found to have been used in Blocks 1 and 2, although the representation of Ob2 material was considerably higher in the other blocks in the survey. On the whole, Ob2 cores are decorticated to the same extent as Ob1 cores, although further upon exploration in Table 6-8) reveals that in Block 2 there appears to be a distinct preference for Ob1 obsidian as cores of this material have only 6% cortex while those of Ob2 have nearly 16% cortex. This is a pattern that might be expected, as the Ob2 obsidian has flaws that both negatively affect knapping quality and affect the visual appearance. However, the pattern is reversed in Blocks 3 and 5 where the Ob2 obsidian is decorticated to a greater extent than Ob1 material. There is a link between the color of the obsidian and the material quality because 45% of the grey obsidian is the Ob2 material, whereas for the other shades of obsidian the Ob2 ratios are smaller (10%-25%).

It was noted that in the Maymeja area of Block 1, unmodified obsidian nodules on the surface were often Ob2 material because they had small gas bubbles in them. Accordingly, it appears that some fraction of the obsidian that was knapped in the Maymeja area was made from this Ob2 surface material because it represented 6.8% of the collections from Block 1. The distribution of Ob1 and Ob2 material at the Chivay source will be further explored below prior to examining the results of the survey work in detail.

Projectile points made from Ob2 obsidian

Eleven obsidian projectile points (4%) were made from Ob2 obsidian; a surprisingly high number under the operating assumption that fracture and visual quality of the material were important characteristics in bifacial tool production. Briefly exploring these eleven Ob2 projectile points may shed light on the characteristics that guided material selection in prehistory.

Ob2 materials form a much higher percentage (15%) of the obsidian flake surface collection than do Ob2 bifacial tools (4%) of the obsidian tool collection, which suggests that Ob2 material was being knapped but apparently not bifacially retouched.

The projectile points made from Ob2 material tend to have small or low-density heterogeneities that do not appear to greatly affect knapping quality, although visually the pieces appear mottled. These points were found in the south-eastern part of the study area in the San Bartolomé area (including one from a Late Formative excavated context), and in the reconnaissance blocks 4 and 5.

ArchID	Block	Period	PPt Type	Weight (g)	Length (mm)	Retouch Index
953.1	2	M. Archaic	2c	4.1	38.28	0.9375
820.1	2		3b	4.1	45.82	1
918.1	5		2c	6.9	48.73	0.96875
818.1	2	Late(2) - T. Archaic	4f	2.9	31.13	1
231.10	4	T. Archaic - Late Horizon	5	5.3	37.62	0.84375
994.1	2		5d	0.5	21	1
1014.3	2		5	2.5	25.52	0.9375
1026.9	2		5	1.9	Broken	1
1038.3	2		5	11.3	19.9
2061.3	2		5d	1.2	Broken

Table 6-9. Projectile Points made from obsidian containing heterogeneities (Ob2).

Period	Ob1	Ob2	Percent with Heterogeneities	Total
Middle Archaic	18	3	14.3	21
Late Archaic	4	1	20	5
T. Archaic - Late Horizon	221	7	2	227
Total	243	11	3.9	253

Table 6-10. Ratio of Obsidian Projectile Points with heterogeneities.

Due to low cell counts, conducting a chi-squared test required aggregating the counts from the Middle and Late Archaic Periods. A chi-squared test on the aggregated table (Table 6-10) showed that the difference between projectile points from Group 1: Middle and Late Archaicand Group 2: the Terminal Archaic through the Late Horizonwith respect to the use of obsidian with heterogeneities is very significant (c²= 9.976, .005 > p> .001). It appears that Ob2 was very significantly less used for point production in the later time period.

Note that the material used for projectile point production in Block 2 was at times the cloudy Ob2, and it is likely that this reflects, in part, the availability of this material on the southern and eastern flanks of Cerro Hornillo. However, the vast majority of the Ob2 obsidian flakes are actually found in Block 3, at the site of Taukamayo.

Obsidian source material with and without heterogeneities

A number of the lag gravel deposits encountered in Blocks 4 and 5 of the survey are Ob2 material. Accordingly, obsidian artifacts from these blocks are higher in heterogeneities, indicating that there was a utility for this type of obsidian despite the imperfect matrix of the material. Investigating the distribution of Ob1 and Ob2 material across all obsidian artifacts (primarily flakes) shows that the Ob2 make up approximately one half of the obsidian artifacts even in Block 3 some distance from the Maymeja zone where Ob1 was observed in situ.

The mean size of Ob1 flakes is notably smaller, which suggests that more advanced reduction was occurring on the Ob1 material. There may be some size bias occurring with observations of heterogeneities because small flakes struck from Ob2 nodule will often appear relatively homogenous and clear if few bubbles or particles are included in the glass in that portion of the flake.

	Homogeneous (Ob1)			Heterogeneous (Ob2)			Total count
Block	No.	m Length (mm)	m Weight (g)	No.	m Length (mm)	m Weight (g)
1	315	40.6	18.8	24	40.5	25.4	339
2	240	25.7	3.6	21	33.8	10.5	261
3	62	30.1	12.4	32	30.1	6.2	94
4	104	35.1	18.3	38	36.8	20.7	142
5	134	23.2	6.2	43	25.5	6.5	177
6	12	25.2	3.1	3	31.7	6.0	15
Total	867	30.0	10.4	161	33.1	12.6	1028

Table 6-11. Obsidian: mean sizes of complete Ob1 and Ob2 artifacts, by Survey Block.

These data, show patterns in terms of the mean length and weight differences between Ob1 and Ob2 artifacts. In all blocks, Ob1 artifacts are on average lighter than their Ob2 counterparts except for in Block 3. Furthermore, in most blocks the mean lengths of Ob1 and Ob2 material are very close but as the weights are different and therefore width or thickness must vary between Ob1 and Ob2 material. Further investigation of the metric data shows that, indeed, Ob1 artifacts have narrower and thinner medial measures, on average, than do Ob2 artifacts except for in Block 3 where Ob2 materials are thinner.

It appears that throughout the study region, Ob1 materials were preferentially knapped into artifacts that were narrower and thinner, but not necessarily shorter, than the Ob2 materials except for in Block 3. Ob2 material was much more common in Block 3, as will be discussed below, and it appears to have been used for more immediate butchering needs rather than for production of bifacial tools, a pattern that is consistent with the later date of the Callalli occupation. There is also a possibility of size bias where smaller Ob2 flakes are classified as Ob1 because no heterogeneities were evident in that particular small flake.