Two typical archaeological artifact distribution problems are discussed with respect to the most effective statistical analysis of multivariate data: (a) Upper Palaeolithic assemblages of stone tools in terms of concentration -- context correlation, covering 46 tool categories for 50 sites. Computer programmes Average-link Cluster Analysis and Principal Components Analysis are referenced. (b) Copper and Bronze tools in terms of their chemically analyzed trace impurity contents extracted mainly from the data of Junghans (who carried out the original grouping at Stuttgart). The Junghan's grouping has been heavily criticized by Waterbolk and Butler (at Groningen), but note is made here that the original treatment was univariate. Average-link Cluster Analysis is employed on 100 metal objects and encouraging clarification of the basic data is presented. -- AATA