I constructed a best estimate phylogeny based on congruence of multiple data sources. In recent years molecular data has been used both to construct phylogenies of taxonomic groups and to aid in the delimitation of new species. I generated and analyzed sequence data for forest spiny pocket mice (Genus Heteromys) for the mitochondrial gene cyt b (1143 bp) and two nuclear gene segments MYH2 (252 bp) and EN2 (189 bp). I used maximum parsimony and Bayesian analyses to infer relationships among species and to provide a framework for using a species delimitation method to investigate the possibility of multiple species within the widespread Heteromys desmarestianus. I found several well-supported lineages within the H.desmarestianus complex, including H. goldmani and H. oresterus. Incorporating karyotype and allozyme data from earlier studies, I found sufficient supporting evidence to justify maintaining H. goldmani and H. oresterus as species as well as identifying four lineages as candidate species. I present a revised taxonomic arrangement within the genus; the subgenus Heteromys should be divided into three species groups: anomalus (H. anomalus and H. australis), gaumeri (H. gaumeri), and desmarestianus (H. desmarestianus, H. goldmani, H. oresterus, and the four candidate species).

This study aims to determine species-level relationships within the genus Heteromys, as well as generic-level relationships among members of the subfamily Heteromyinae using a phylogenetic framework. Molecular sequence data were generated from two mitochondrial genes (cytochrome b and cytochrome oxidase I) and three nuclear gene segments (β-fibrinogen, engrailed protein II, and myosin heavy chain II), and analyzed under maximum parsimony, maximum likelihood, and Bayesian optimality criteria to infer relationships. Chapter 1 focuses on the phylogenetic and taxonomic implications for Heteromys from the analyses of sequence data. Phylogenies also provided a framework for delimiting species boundaries within the wide-ranging Heteromys desmarestianus complex using the Wiens and Penkrot method. Several well-supported clades within this complex were recovered, including H. goldmani, H. nubicolens, and H. oresterus, as well as five groups identified as candidate species. Heteromys oasicus was not found to be genetically diagnosable from H. anomalus, and was relegated to subspecific status. I present a revised taxonomy as follows: the monotypic subgenus Xylomys is maintained (H. nelsoni); the subgenus Heteromys is divided into three species groups – anomalus (H. anomalus [including H. oasicus], H. australis, and H. teleus), desmarestianus (H. desmarestianus, H. goldmani, H. nubicolens, H. oresterus, and the five candidate species), and gaumeri (H. gaumeri). Chapter 2 describes phylogenetic inferences made from analyses of heteromyine taxa, genera Heteromys and Liomys. Many studies have recovered Liomys as paraphyletic relative to Heteromys, and the goal of this chapter was to address this taxonomic problem. The Liomys pictus species group (L. irroratus, L. pictus, and L. spectabilis) was recovered as sister to Heteromys rather than to the L. salvini group (L. adspersus and L. salvini). I recommend a revised taxonomy for the subfamily as follows: the genus Heteromys is retained as delineated in Chapter 1; the genus Liomys is reduced in scope to include only L. irroratus, L. pictus, and L. spectabilis; the subgeneric name Schaeferia is elevated to generic rank and includes S. adspersus and S. salvini. This classification better reflects the phyletic diversity within the subfamily Heteromyinae, and requires fewer name changes; thus providing nomenclatural stability.

Heteromys pictus-spectabilis is a species complex within the subfamily Heteromyinae (Family: Heteromyidae) that is distributed along the western and southern Mexican coast and surrounding environments. Currently, the species complex is accepted as being 2 separate species (H. pictus and H. spectabilis), but this also renders H. pictus paraphyletic. Therefore, the species complex requires re-evaluation in order to resolve the paraphyly. Mitochondrial DNA sequences from a previously existing ~720 specimen database compiled by Victoria Vance were used in conjunction with new nuclear DNA sequences sequenced for the purpose of this study to generate multiple phylogenetic trees via the software programs RAxML, BEAST, and MrBayes to evaluate how different haplotype networks were related to each other. Using these molecular datasets in consideration with Kimura two-parameter values, time calibrations via BEAST, and the relative geographic locations of the haplotype networks, the results strongly indicate this species complex is composed of multiple cryptic species and potentially multiple genera. This was a preliminary exploration into this species complex however, and future research will be required to verify these findings.