A Mathematical Model of the Effects of Screen Size on Zooarchaeological Relative Abundance Measures

Zooarchaeologists have long recognized that the use of different sizes of screen in excavation can result in varying degrees of bias in faunal samples. This paper presents a graphical, mathematical representation of the ways in which screen size choice can affect measures of taxonomic relative abundance. Data from the site of Mogollon Village in southwestern New Mexico show that the rank order of samples along a measure of relative abundance like the Artiodactyl Index can vary with screen size. This occurs because specimens of a given taxon can respond to changes in screen size in different ways in different provenience units, a fact which also severely limits the utility of ‘‘correction factor’’ methods for dealing with screen size bias. The model presented here and the Mogollon Village fauna have important implications for studies based on temporal or spatial trends in the abundances of large-bodied taxa relative to small-bodied taxa, especially in cases where large mesh sizes or no screens at all were used in excavation. ? 1999 Academic Press