Remarkable species diversity in Malagasy mouse lemurs (primates, Microcebus)

Abstract

Phylogenetic analysis of mtDNA sequence data confirms the observation that species diversity in the world's smallest living primate (genus Microcebus) has been greatly underestimated. The description of three species new to science, and the resurrection of two others from synonymy, has been justified on morphological grounds and is supported by evidence of reproductive isolation in sympatry. This taxonomic revision doubles the number of recognized mouse lemur species. The molecular data and phylogenetic analyses presented here verify the revision and add a historical framework for understanding mouse lemur species diversity. Phylogenetic analysis revises established hypotheses of ecogeographic constraint for the maintenance of species boundaries in these endemic Malagasy primates. Mouse lemur clades also show conspicuous patterns of regional endemism, thereby emphasizing the threat of local deforestation to Madagascar's unique biodiversity.

Figure 1

Results from discriminant function analysis of 34 cranial, dental, and external morphometric characters. Body mass was not considered in this analysis. Detailed character descriptions are given in ref. . Functions 1 and 2 (A) show conspicuous discrimination of M. berthae from other species. Functions 2 and 3 (B) show discrimination of all species. Combined, the first through third discriminant functions describe 94.5% of the variance in the data set; 55.9%, 31.7%, and 6.9% for the first, second, and third discriminant functions, respectively. Dashed lines are drawn around species clusters for purposes of illustration; they do not convey statistical information.

Figure 2

Phylogeny derived from sequence alignment of 2,404 bp of combined mtDNA sequences from the control region homologous with the hypervariable region 1 region in humans, COII and cytochrome b. Clades are color-coded to emphasize species diversity. Individuals are identified by unique laboratory extraction number (Yoder Lab Extraction; YLE) and by locality. Distance tree was generated in paup* 4.0b4a (PPC) (53) by using HKY85 correction model (54) and weighted least squares (power = 2) algorithm. A total of 1,000 replicates of the random addition option were executed without branch swapping. TBR branch swapping then was performed on best tree (hit 399 of 1,000 trials). A single tree of score 0.83 (%SD = 3.27) resulted from the search and is shown with midpoint rooting. Location of midpoint root was confirmed by multiple outgroup rootings. Clade resolution and hierarchy is congruent with strict consensus of 12 trees derived from maximum parsimony analysis in which 100,000 random additions were performed without branch swapping, followed by TBR branch swapping of the 10 best trees. Numbers on branches indicate statistical support from 100 bootstrap replicates with one random addition per replicate. Propithecus, Varecia, and Eulemur were used to root the bootstrap tree. Circled numbers highlight bootstrap support for two primary clades. Asterisks beside species designations indicate species that have been reported to occur in sympatry with another mouse lemur type. See ref. for details.

Figure 3

Illustration of Microcebus mtDNA branch lengths (shaded boxes) compared with those of Eulemur species (open boxes). Two species highlighted for Eulemur (E. fulvus and E. rubriventer) are morphologically distinct and occur in noninterbreeding sympatry. The following individuals (identified by YLE number and by field number) were chosen as representative haplotypes: M. ravelobensis, YLE66 (RMR58); M. sambiranensis, YLE72 (RMR40); “M. rufus1,” YLE138 (SA M102); M. myoxinus, YLE62 (RMR30); M. berthae, YLE148 (JUG 72); M. tavaratra, YLE110 (RMR76); “M. rufus2,” YLE190 (SMG 8774); M. murinus (west), YLE74 (RMR47); M. murinus (east), YLE199 (JUG A-8A1D); M. griseorufus, YLE116 (RMR67). Tree was generated in paup* 4.0b4a (PPC) by using the maximum-likelihood optimality criterion. Settings corresponded to the HKY85 model with rate heterogeneity. -Ln likelihood, 12982.79; estimated ti/tv, 4.60 (kappa, 9.60); estimated gamma shape parameter, 0.26. Circled numbers indicate quartet puzzling values from 1,000 puzzling steps of HKY85 model with sites assumed to evolve at single rate. Likelihood ratio test conducted in puzzle 4.0.1 (55) indicates that a molecular clock cannot be rejected (critical significance level = 14.96%).

Figure 4

Combined mtDNA haplotype phylogeny (from Fig. 2) superimposed on mouse lemur collecting localities. Figure shows segregation of haplotypes into northern and southern clades with 85% and 100% bootstrap support, respectively. The placement of the root was confirmed by multiple rooting techniques (outgroup and midpoint), a variety of outgroup taxon samples, and three different optimality criteria (maximum parsimony, distance, and maximum likelihood). Note that Kirindy/CFPF contains individuals from both northern and southern clades. Map modified with permission from ref. . Ecotype designations are general; for site-specific information, refer to Table 1.