James L. Boone
Museum of Natural History, University of Georgia, Athens, GA 30602

Co-workers: Dr. Joshua Laerm and Dr. Michael H. Smith

My revisionary studies concerning the subspecific taxonomy of the cotton mouse (Peromyscus gossypinus) continue, and I am very interested in discussing these finding and ideas with other interested researchers. In a paper recently submitted for review to the Journal of Mammalogy, I have proposed synonymizing the Federally endangered Key Largo cotton mouse, presently referred to P. g. allapaticola, with the more widespread mainland subspecies, P. g. palmarius. In this study, I examined cotton mice, white-footed mice (P. leucopus), and deer mouse (P. maniculatus) to assess the taxonomic validity and affinities of the mice on Key Largo, Florida, presently referred to P. g. allapaticola. I performed genetic analysis on 40 loci in 658 mice from 35 populations (6 island, 29 mainland) and morphometric analysis on 18 characters on 1167 mice from 36 populations. Every population was genetically and/or morphologically different from every other population for at least one character. However, when I examined all populations simultaneously, the differences between the Key Largo and other populations were not significant. While each population was statistically distinct, none was unusually distinct. Therefore, the Key Largo population does not warrant recognition as a separate subspecies and should be referred to the widespread mainland subspecies, P. g. palmarius.

These results are quite similar to those I found for the mouse population on Cumberland Island, Georgia, formerly referred to P. g. anastasae (Boone et al., 1993). As above, this island population appeared to be morphologically distinct, but in fact was not any more distinct than would be expected among randomly chosen populations. This is a species where we must be particularly careful to ensure that subspecies of P. gossypinus are not simply monuments to "the persistence and patience of the systematist" (Lidicker, 1962: 168).

Since Boone et al. (1993) was submitted for publication, I have greatly expanded the data sets (allozymes and morphology), and I am tempted to revise my earlier position concerning all the Georgia and north Florida islands (Amelia Island, FL north to Skidaway Island, GA). Populations on these islands do appear as a distinct cluster in both genetic and morphometric spaces (Fig. 1). However, I find it difficult to conclude that they should be a single subspecies, but I find it equally difficult to conclude that they should each be a different subspecies.

We have also been investigating a possible case of reticulate evolution on Cumberland Island, Georgia where it appears that a subspecies level of distinction may have been achieved by hybridization between P. gossypinus and P. leucopus, and by "hybridization" between P. g. gossypinus and P. g. megacephalus. The inadvertent movement of animals as a result of trade between island residents and people around Baton Rouge, Louisiana, and trade with "Northerners" can account for the distribution of alleles at two loci (Transferrin and GPI). Of course, common ancestry can also account for these observations. We have found a few hybrid P. gossypinus X P. leucopus individuals from areas where these species are sympatric, so hybridization on Cumberland Island is a possibility.

Literature Cited

Boone, J. L., J. Laerm, M. H. Smith. 1993. Taxonomic status of the Anastasia Island Cotton Mouse (Peromyscus gossypinus anastasae). J. Mamm. 74:363-375. Abstract

Lidicker, W. Z., Jr. 1962. The nature of subspecies boundaries in a desert rodent and its implications for subspecies taxonomy. Syst. Zool., 11:160-171.

Figures

Figure 1. Phenogram of three Peromyscus species produced from UPGMA cluster analysis (17 morphometric characters, correlation matrix extraction, varimax rotation) of population centroids.