An examination of obsidian by color per excavation level, and color by rotations as a measure of investment but discard, reveals that grey and black obsidian artifacts had the greatest number of rotations while clear and clear-banded obsidian, while making up over 50% of the obsidian present, was seldom rotated more than twice. A number of researchers have observed that Chivay obsidian is unusually transparent. Yet, this evidence indicates that grey and black obsidian was predominant in certain levels at the workshop and these artifacts were rotated more extensively.

Table 7-15. Attributes of obsidian artifacts from Q02-2u3.

In the course of the analysis, Alex Mackay noted the presence of Kombewa flakes (Schick and Clark 2003;Tixier, et al. 1980), also as known as "Janus flakes", in analysis. This a flake-as-core technique where

(1) A relatively large flake, Flake A, is struck from a core.

(2) Flake A then becomes Core A when a flake is struck from the ventral surface of Flake A removing the bulb of percussion from step 1 as a thinning technique and producing Flake B.

(3) Flake A is a Kombewa core, and the Flake B that results from step 2 is a Kombewa flake that has bulbs of percussion on both sides.

Table 7-16. Proportion of Kombewa flakes by level in Q02-2u3.

The Kombewa core is a flake that is thinned by this process, and the Kombewa flake that is produced is small and triangular. Mackay observed that a number of series 5 projectile points had vestiges of a bulb of percussion on one of the proximal tangs on one side, and a bulb of percussion in the middle of the piece on the other side, which suggests that it was produced on a Kombewa flake. The only other location where Kombewa flakes were encountered in significant quantities was on the surface of the site of Collpa [A03-910] in Block 5 midway between the Chivay source and Callalli. Seven Kombewa flakes (6 of obsidian, one of chert) were encountered on the surface at Collpa. A number of series 5 points were also encountered at Collpa.