Classification bias in discriminant function analyses used to evaluate putatively different taxa

Discriminant function analysis has been used to assess morphological distin ctiveness of putatively different taxa. We used randomizations of previousl y published morphological data for 2 subspecies of the coyote, Canis latran s frustror and C. l. thamnos, to quantify a recognized but previously unexa mined bias in discriminant-function analyses that use resubstitution classi fication. This bias results in overestimates of intertaxon distinctiveness and is exacerbated when sample sizes are small. An alternative classificati on technique, jackknife sampling, is relatively unbiased.