Mathematical approaches assessing similarity in terms of culture, geography, and zoological components were applied to nine online collections of New Guinean necklaces. When mapped in multidimensional space for peoples, no strong clustering of collections was found, and for provinces two collections formed a distinct cluster from the rest. In zoological space there was no clustering, but one collection occupied a distinctly separate space. A highly significant ( p < .001) effect of collection on the zoological species richness was found. There were significant differences ( p < .05) in zoological entropy between several collections, and a degree of uncertainty or surprise in the zoological composition of the necklace collections. The processes behind such patterns are likely complex, and may reflect issues of funding, unconscious bias, and colonial or missionary histories. The methods explored provide diagnostic tools useful for testing the underlying structures and bias of collections.