Maize Project Progress Report (Part 1) - eScholarship
Melanie Wright · 1994
www.escholarship.org/uc/item/7d96x2cz Maize Project Progress Report #1 October 13, 1994 I Archaeological Maize We are still in the process of pulling maize from the Tiwanaku samples. So far approximately 207 capsules have been pulled, of which 94 were kernels and 113 were cupules (45% and 55% respectively). Of the capsules pulled, however, only 100 were found to be measureable (ie 48%). The cupules are disproportionately more measurable than the kernels: 70% of the measurable capsules were cupules, 30% kernels. In the unmeasurable sample, 40% were cupules and 60% were kernels. Or to look at it another way, 62% of the cupules pulled were measurable, where only 32% of the kernels pulled were measurable. This may pose an important difference in interpretation, because of the possibility (and in my opinion liklihood) that whether a kernel survives fairly intact (ie is measurable) is directly related to the variety of corn it represents : I think it is quite possible that the smaller kerneled, harder varieties (like Confite Punefi.o and Altiplano) may survive to measured better than the larger, lower valley floury types like Huilcaparu or Checchi. This is something that we can address experimentally when we burn our modern comparatives. There are approximately 104 capsules left to pull : 42 of kernels and 62 of cu pules (40% and 60% respectively). If the previous trend keeps up, this would represent 38 capsules of measureable cupules and 13 capsules of measureable kernels. Each capsule most commonly has one or two measurable item contained in it, but in quite rare cases a few particularly dense samples may contain up to 10-15 measurable items per capsule. So our total sample size from Tiwanaku mismo is likely to be around 100-200 cupules and 50-75 kernels, which is a little low. There are a couple possible ways to address this problem. First, I could redefine what I mean by measurable to include more broken kernel and cob fragments, either by estimation of the bits missing, or simply measurement of fewer variables. The latter would cause problems in the stat analysis down the road, because many analytical procedures won't accept absent data. The other possible route is to include in the sample maize from the valley sites near-ish to Tiwanaku which date to Tiwanaku times. I have not scanned Juan and Jim's book to see which/how many sites that might be, but it is sure to increase our sample size at least somewhat. And I feel fairly comfortable that the people at Kk'arafia, for example, weren't living that differently from their counterparts in other valley sites. This is problematic, however.