ABSTRACT
The genetics that distinguish the "domesticated" form of maize from its wild progenitor teosinte are now believed to involve relatively few genes, with one of those genes being teosinte glume architecture which is thought to affect silica deposition. Cluster analysis of morphological trait presence or trait absence data of populations of 100 phytoliths is successful in differentiating among known varieties and has been useful in classifying archaeological phytolith samples into known extant varieties. This chapter examines the underlying patterns of variation that result in what appears to be a successful clustering methodology, as a first step in ultimately discovering which morphological traits are the best indicators of genetic variation in maize rondel phytoliths. This chapter assumes all of the ethnographic samples are from highland Bolivia and vary randomly among samples rather than being clustered by ethnographic variety.
