This article explores the chemical-element composition of a wide range of local domestic wares (late 6th century b.c. to early 4th century a.d.). Using instrumental neutron activation analysis and two statistical approaches, we examine whether function or date explains variations in the chemical groups we found and if they match their archaeological labels. The initial choice of ceramic pastes (cooking fabric or buff ware) relates to function, but within those main categories, vessel function did not affect grouping. Fabrics did vary with time. Such changes do not occur at the same time in the cooking and buff wares, nor do they correlate well with the political history of the city.