WHAT DOES SKELETAL MORPHOLOGY REALLY TELL US? VARIATION IN PINE MARTEN LIMBS BY BIOME AND MOLECULAR PHYLOGENY

Living species are usually diagnosed by molecular criteria. Fossil species are diagnosed based on skeletal morphology, because DNA is usually too degraded for sequencing. Skeletal morphology, however, reflects habitat as well as relatedness. How good a proxy is bone morphology for relatedness?

I tested this question in the North American pine marten, Martes americana and Martes caurina, by performing a 3D geometric morphometric analysis of limb morphology and by sequencing the 12S, 16S, and d-loop genes from the same specimens measured. These specimens were collected 2000-2013 and span three biomes in the U.S.: temperate coniferous forest (TCF); temperate broadleaf mixed forest (TBM); and boreal forest (BF). A PERMANOVA of PC scores found significant morphological differences in the limb bones between specimens from all three biomes. A mixed-model Bayesian phylogeny of the 12S, 16S, and d-loop genes from these same specimens supported two distinct clades: TBM/BF and TCF. The TBM/BF clade contained well-supported nodes with specimens from both biomes.

Phylogenetic and morphological separation of TBM/BF and TCF specimens supports the presence of two species, M. americana and M. caurina, as has been previously proposed from the cytochrome b gene and cranial morphology. Reintroduction of martens from BF to TBM may explain the lack of phylogenetic separation between TBM and BF specimens despite significant morphological differences. This likely resulted in haplotypes from both biomes present in TBM, but suggests that limb shape of reintroduced populations could readily adapt to differences in habitat. This indicates interspecific phylogenetic relationships are congruent with morphological variation, while intraspecific morphology may correlate more with habitat. Skeletal morphology, therefore, is appropriate for diagnosing fossil species, but likely not subspecies.