.We used six restriction endonucleases (D&I, HinfI, HpaII, MM, TagI, and Zmo37), each with a unique four-base recognition site, to digest 26 mitochondrial genomes from six Platycercus species in an investigation of the systematics of this speciose and variable Australian genus. The resulting fragments were end labeled and separated by means of polyacrylamide-gel electrophoresis. The number of substitutions per nucleotide site between each of the genomes was estimated on the basis of the distribution of fragments between samples by means of the method of Nei and Li. These pairwise distances were used to construct evolutionary trees either by assuming an evolutionary clock or by allowing the branch lengths to vary independently. The presence or absence of each restriction fragment was also scored and used as characters in Wagner parsimony analysis. In general, the phylogenies produced from mitochondrial genome diversity mirror the current classification of the genus, except for P. icterotis, which is not clustered with either of the superspecies of Platycercus and, as such, should probably be placed in a third group. To confirm that the family of parrots to which Platycercus belongs has a conservative nuclear genome, Nei’s genetic distance between and within Platycercus and other related species was estimated by means of cellogel electrophoresis of liver isozymes. Of the 26 Platycercus mitochondrial genomes studied, 23 different haplotypes were identified according to the limits of resolution of the technique. The mean interspecific mitochondrial genome diversity was 0.0457, and the intraspecific genome diversity varied between 0.003 1 (for P. caledonicus) and 0.0 142 (for P. eximius). These diversity values indicate either that the rate of rosella genome evolution is 2-3 orders of magnitude below the rate reported for primates or that the effective sizes of the rosella populations are 2-3 orders of magnitude below the sizes estimated on the basis of ecological methods.