11Center for Conservation and Research (CCR), Henry Doorly Zoo, Omaha, NE, USA
22Anthropology Department, University of Antananarivo, Antananarivo, Madagascar
33Conservation International, Arlington, VA, USA
*Authors'addresses:Edward E. Louis, Jr., Shannon E. Engberg, Susie M. McGuire, Marilyn J. McCormick, Richard Randriamampionona, Carolyn A. Bailey, Runhua Lei, Center for Conservation and Research (CCR), Henry Doorly Zoo, 3701 South 10th St., Omaha, NE 68107-2200, USA. E-mail of first author: <edlo@omahazoo.com>.
**Jean Freddy Ranaivoarisoa, Anthropology Department, University of Antananarivo, Antananarivo, Madagascar.
***Russell A. Mittermeier, Conservation International, 2011 Crystal Drive, Suite 500, Arlington, VA 22202, USA.
It is not available for individual sale.
Molecular genetic sequence variation of northern and northwestern mouse lemurs (Microcebus) was examined during a phylogenetic analysis of mitochondrial DNA (mtDNA) sequence data (c. 3,000 bp) for the entire genus. Phylogenetic inference of the mitochondrial DNA sequence data was generated from 132 individuals, representing 15 species of mouse lemurs. The database distinguished the 15 described Microcebus species and also provided diagnostic evidence for two further species. A comparison of the data for two mouse lemur species described from Nosy Be confirmed the existence of just one for this island population. The localities of the newly identified species are within the distributions previously recognized for Microcebus sambiranensis and Microcebus tavaratra. Formal descriptions, drawn from molecular genetic data, are presented for the two newly named species: one from Antafondro Classified Forest and the other from Montagne d'Ambre National Park. We revise the Inter-River-System hypothesis concerning the biogeographic patterns of the distributions of the northern and northwestern mouse lemurs according to our findings concerning the two species described here.
Introduction
Due to its unique species biodiversity and to the continued pressure from human encroachment, Madagascar is among the highest conservation priorities worldwide (Myerset al. 2000). With 40% of the forest cover lost between the 1950s and 2000, rapid and comprehensive surveys of the remaining forest are essential (Harperet al. 2007). Dufils (2003) estimated that 90% of Madagascar's biodiversity is found exclusively in forest or woodland tracts, making these research efforts more urgent still. Recent molecular genetic and morphological studies of lemurs, particularly the mouse lemurs (Microcebus) and sportive lemurs (Lepilemur), have led to a great increase in the number of recognized species (Andriantompohavanaet al. 2006;Craulet al. 2007;Kappeleret al. 2005;Louiset al. 2006a,2006b;Olivieriet al. 2007;Radespielet al. 2008). Even with these taxonomic revisions and the consequent realignments of the distributions of the species, regular re-evaluations are needed to monitor the conservation status of each taxon (Louis Jr.et al. 2006b).
All lemurs are currently protected under the Convention on International Trade in Endangered Species (CITES). Fortyone lemurs (43% of the 96 species and subspecies listed) were categorized on the2008 IUCN Red List of Threatened Species as threatened (IUCN 2008). The status of a further 43 lemurs (45%) were, however, too poorly known to be assessed and were classified as Data Deficient. Distributed throughout the island, lemurs are particularly susceptible to extinction from stochastic and deterministic factors due to their relatively small and fragmented geographic ranges (Jernvall and Wright 1998). Mouse lemurs are adaptable, being found in secondary or otherwise degraded forest tracts, even along roads. They live in small social units, being solitary or forming small family groups (Guschanskiet al. 2007), and are limited in their capacity to disperse because they are nocturnal and small (30–80 g) and have small home ranges of 0.3–1.5 ha (Schwab 2000;Weidtet al. 2004;Louis Jr.et al. 2006a).
Until recently, the northern and northwestern mouse lemurs were represented by the northern mouse lemur (Microcebus tavaratra) found at Ankarana National Park, and the Sambirano mouse lemur (Microcebus sambiranensis) found at Manongarivo Special Reserve (Rasoloarisonet al. 2000). Based on phylogenetic inference of mitochondrial DNA (mtDNA) sequence data, Andriantompohavanaet al. (2006) presented evidence for a new species of mouse lemur they namedM. mamiratra in northwestern Madagascar at Nosy Be Island, and also indicated the probability of another, which they referred to asMicrocebus sp.nova #5, at Antafondro Classified Forest. Olivieriet al. (2007) presented a biogeographic model for the northern mouse lemurs, and described three new species, including one,M. lokobensis, from Lokobe Special Reserve on Nosy Be Island and Manehoka on the mainland of Madagascar.
Three biogeographic models have been proposed for the distribution patterns of mouse lemurs, based on different relative contributions of factors that include large rivers (>50 m wide at 20 km inland), retreat dispersion watersheds, and topographical barriers such as mountains (Martin 1995;Wilméet al. 2006;Craulet al. 2007;Olivieriet al. 2007). Olivieriet al. (2007) and Craulet al. (2008) presented biogeographic models which defined “centers of endemism” based on the isolation effects of paired rivers, or Inter-River-Systems (IRS;Fig. 1). During the course of a number of biogeographic reviews of northern and northwestern Madagascar, the number of Inter-River-Systems has increased from four (Martin 1995), to five (Wilméet al. 2006) to nine (Craulet al. 2008).
In this paper, we present a comparative phylogenetic analysis of the northern and northwestern mouse lemurs. With comprehensive sampling in this region (novel samples and sites, along with accessioned published sequences), we re-evaluate the biogeographic partitions, define the relationship betweenMicrocebus mamiratra andMicrocebuslokobensis described independently from the island of Nosy Be, and provide descriptions of two mouse lemurs that we consider to be distinct species; one from Antafondro Classified Forest and the other from Montagne d'Ambre National Park.
Methods
Sample collection
All lemurs in this molecular study were free-ranging, wild-caught, adults (Fig. 1;Table 1;Appendix I(a)). All mouse lemurs were hand-caught and subsequently immobilized using 1.0–3.0 mg of Telazol® (Fort Dodge). Two 2.0-mm biopsies and 0.01–0.05 cc of whole blood were collected and stored in room temperature tissue preservative (Longmireet al. 1992). The lemurs designated as outgroups were immobilized with a CO2 projection rifle or blowgun with 10mg/kg of Telazol® (Fort Dodge;Appendix I(a)), and four 2.0-mm biopsies and 1.0 ml/kg of blood were collected and stored in room temperature tissue preservative (Longmireet al. 1992). Genomic DNA was extracted from a 2.0-mm ear punch using a phenolchloroform extraction (Sambrooket al. 1989). All measurements were taken on sedated animals as described in Andriantompohavanaet al. (2006). We measured the weight (±0.1 g), head crown (total length from the tip of the nose [soft tissue of the nose is not included] to the occipital crown ±0.1 cm), body length (total length of body from the occipital condyle to the base of the tail ±0.1 cm), tail length (total length from the base of the tail to the end of the last caudal vertebra ±0.1 cm), ear length (total length from the tip of the ear to the base ±0.1 mm), ear width (total width across the widest portion of the pinna ±0.1 mm), and muzzle length (total length from the tip of the nose [soft tissue of the nose is not included] to the medial corner of the eye ±0.1 mm). For presentation purposes we provide the weight, head crown, body length, and tail length following the guidelines of Smith and Jungers (1997). (SeeTable 1. Appendices I(a–b).)
Data generation
To compare our data with previously published molecular studies, we analyzed the following regions of the mitochondrial DNA (mtDNA): D-loop or control region (D-loop;Bakeret al. 1993;Wyneret al. 1999); and a fragment of the cytochrome oxidase subunit III gene (COIII); NADH-dehydrogenase subunits 3, 4L, and 4 (ND3, ND4L, and ND4); as well as the tRNAGly, tRNAArArg, tRNAHis, tRNASer, and partial tRNALeu genes (PAST;Pastoriniet al. 2000;Louis Jr.et al. 2006a). Using 50 ng of genomic DNA, the D-loop (487–531 base pairs (bp)) and the PAST fragments (2367 bp) were amplified by the polymerase chain reaction (PCR) using the following conditions: 94°C for 30 s, a primer-specific annealing temperature for 1 min, and 72°C for 5 min for 35 cycles. Since all potential sites or populations of mouse lemurs have not been collected, accessioned sequences were used to compare and augment the datasets to evaluate the current taxonomic knowledge of the genus Microcebus (Andriantompohavanaet al. 2006;Yoderet al. 2000;Louis Jr.et al. 2006a;Olivieriet al. 2007; seeTable 1;Appendix III(a)). The species described by Radespielet al. (2008) were not included in these analyses since sequence fragments could not be compared at this time. To evaluate the two described species of Nosy Be,Microcebus mamiratra andM. lokobensis, representative sequences for the D-loop were added to the data file (Appendix III(a)).
PCR products were confirmed visually on a 1.2% agarose gel, and purified using QIAquick PCR purification kit (Qiagen, Valencia, CA). Using the BigDye terminator cycle sequencing ready reaction kit by Applied Biosystems, the sequence was generated with a 7% Polyacrylamide gel by an ABI 3100 automated sequencer (Applied Biosystems, Inc; Foster City, CA). The sequence fragments were aligned to generate a consensus sequence using Sequencher (Gene Corp; Ann Arbor, MI), and the consensus sequences were aligned using ClustalX (Thompsonet al. 1997). The consensus sequences were submitted to GenBank and Accession Numbers are listed inTable 1 (seeAppendix I(a)). The sequence alignments for the data sets are available from the first author upon request.
Figure 1.
Distribution map of the mouse lemur (genusMicrocebus) samples of northern and northwestern Madagascar. Each sample site is color-coded to a specificMicrocebus species. The Inter-River-System (IRS) data is based on Olivieriet al. (2007).
Phylogenetic analysis
To examine the genetic diversity of the mouse lemurs of the northern region of Madagascar, maximum-parsimony (MP), maximum likelihood (ML), and neighbor-joining (NJ) analyses were implemented for the D-loop and PAST, and combined (D-loop//PAST) sequence data with PAUP software (Swofford 2001). The trees described in this paper are all consensus trees except for the bootstrap analysis (all trees were presented as phylograms for presentation purposes only). Bootstrap analyses were accomplished with 1000, 3000, and 4000 pseudoreplicates with the D-loop; PAST; and D-loop/PAST combined sequence files, respectively, with 10 random addition heuristic searches per replicate option selected. Only nodes with greater than 50% support were reported. The D-loop NJ tree was generated using the Tamura-Nei model (Tamura and Nei 1993). The stepwise addition option was selected for MP analyses, and corrections for nucleotide sequence data suggested by Kimura (1980) were used with the NJ analyses. Gaps were considered as a fifth character in MP analyses, whereas gaps were treated as missing data in the NJ analyses. The ML trees were estimated via the best-fit model selected by the hierarchical likelihood ratio test (hLRT) in ModelTest3.5 (Posada and Crandall 1998). The best-fit model selected by the hLRT criteria was the TrN+I+G model [(0.2750 0.0996 0.2552), Nst=6, Rmat=(1.0000 13.5199 1.0000 1.0000 8.4486), Gamma= 1.0731, Pinvar=0.4333]. In addition to characterbased phylogenetic analysis of DNA sequences, PAUP software (Swofford 2001) was also used to calculate uncorrected pairwise distances (‘p’) and Kimura distance measures for D-loop and PAST fragments.
Table 1.
Samples (27 total) from free-ranging mouse lemurs (Microcebus) used in this study. MtDNA sequence data for each mouse lemur sample are available from GenBank under the listed accession numbers. The TK number is the catalogue of the paratype DNA sample stored at the Museum of Texas Tech University, Lubbock, Texas. Global Positioning System (GPS) shows the site where the animal was immobilized. The samples not listed in this manuscript are available in Louiset al. (2006a) and Andriantompohavanaet al. (2006).
Bayesian inference analyses were conducted using MrBayes 3.0b4 (Huelsenbeck and Ronquist 2001;Ronquist and Huelsenbeck 2003). The model of evolution was selected by using MrModeltest 2.2, a modified version of Modeltest 3.6 (Nylander 2004;Posada and Crandall 1998). A Markov Chain Monte Carlo (MCMC) run with four simultaneous chains and 1,000,000 generations was performed. Every hundredth generation, the tree with the best likelihood score was saved, resulting in 4000 trees. The 4000 trees were condensed in a majority rule consensus tree using PAUP Version 4.0b10 software (Swofford 2001). Branch supports were assigned as posterior probabilities on the consensus tree. The pattern of sequence evolution was estimated by conducting a minimum spanning network generated with the program NETWORK Version 4.11 (Bandeltet al. 1999;Forsteret al. 2001;Gonzaleset al. 1998) and Arlequin, Version 2.0 (Schneider et al. 2000).
As described in Andriantompohavanaet al. (2006), Davis and Nixon (1992), Wyneret al. (1999), Mayoret al. (2004), and Louis Jr.et al. (2006a,2006b), we used MacClade 3.01 (Maddison and Maddison 1992) and MEGA version 2.0 (Kumaret al. 1993) in a diagnostic search to designate Evolutionary Significant Units (ESU) for the Microcebus species using a Population Aggregate Analysis (PAA) of the D-loop (487–531 bp) and PAST (2367 bp) sequence data. In this paper, the current Microcebus taxonomy for northern and northwestern Madagascar was examined according to the Phylogenetic Species Concept (PSC)sensu (Wheeler and Platnicket al. 2000; Louis Jr.et al. 2006;Mayoret al. 2004). With the sequential addition of each individual without an a priori species designation, a PAA distinguishes attributes or apomorphic characters according to the smallest definable unit (Andriantompohavanaet al. 2006;Davis and Nixon 1992;Mayoret al. 2004;Louis Jr.et al. 2006a,2006b;Ravaoarimananaet al. 2004).
Results
Mitochondrial DNA sequence data were completed for two fragments, D-loop and PAST (approximately 3,000 bp), for 121 individuals, representing all 15 recognized species of mouse lemurs from a total of 32 sites (Figs. 1–4, Appendices II(a–e)). Based on the phylogenetic inferences of the NJ, MP, and ML analyses of three sequence alignments (D-loop, PAST, and combined), the 15 Microcebus species were represented in 15 well-supported terminal clades (Figs. 2–4; the newly described species by Radespielet al. (2008) were not included in these analyses since sequence fragments could not be correlated). All three phylogenetic methods corroborate the monophyly ofM. griseorufus andM. murinus and the monophyly ofM. bongolavensis,M. danfossi, andM. ravelobensis as presented in Radespielet al. (2008). Additionally, the sister relationship betweenM. myoxinus,M. berthae,M. lehilahytsara, andM. rufus exists with all three methods for the D-loop sequence fragment, but cannot be confirmed for the PAST or D-loop/PAST concatenated due to the unavailability of samples sets forM. bongolavensis andM. danfossi. The mouse lemur samples from the island of Nosy Be, comprising Microcebus sp. nova #4 from Louis Jr.et al. (2006a),M. mamiratra from Andriantompohavanaet al. (2006), andM. lokobensis from the IRS VI in Olivieriet al. (2007; Lokobe Special Reserve on Nosy Be and Manehoka from mainland Madagascar) were found to form a single terminal clade (Figs. 1 and2). The minimum spanning network for theMicrocebus D-loop haplotypes reveal a similar evolutionary pattern as the three phylogenetic methods (Fig. 5). Interestingly, Microcebus jollyae, an east coast reddish morph, is aligned intermediately between theM. griseorufus andM. murinus group, west coast gray forms, and theM. mittermeieri andM. simmonsi, east central coast reddish morphs. The samples from Nosy Be, representing the two described species,M. mamiratra andM. lokobensis, along with the samples from Manehoka (mainland Madagascar), clustered together as one well-supported terminal clade. Furthermore, all three phylogenetic methods support two distinct subpopulations,Microcebus sp.nova #5 (Antafondro) andMicrocebus sp. nova #6 (Montagne d'Ambre;Figs. 1–4; Appendices II(a–e)).
A review of the morphometric data for 13 described species of mouse lemurs are presented inTable 2 (detailed morphological measurements of the novel individual mouse lemurs are available inAppendix I(b)). No extensive quantitative analyses were conducted on the morphometric data. Inherent inconsistencies found or produced within morphologic data sets prevent a statistically reliable conclusion. Numerous factors such as small sample sets, independent data sets, multiple data collectors, the variance between live, sedated individuals versus processed museum vouchers, along with seasonal and age differences of individual mouse lemurs, currently restrict any comprehensive analysis of the genusMicrocebus. With that said, this morphometric information is provided as supplemental data, only complementing the partitioning of unique biodiversity (Table 2).
The results from the population aggregate analysis of the D-loop and PAST sequence data are presented inTables 3 and4, respectively (Appendices III(b–e)). Multiple diagnostic characters distinguish each established Microcebus species, along withMicrocebus sp.nova #5 andMicrocebus sp.nova #6 at Antafondro and Montagne d'Ambre, respectively (Tables 3 and4; Appendices III(b–c)). Microcebus sp; nova #5 had seven diagnostic sites, whereasMicrocebus sp. nova #6 had nine. The complete uncorrected ‘p’ distance and the Kimura two-parameter distance measures are presented inTables 5a and5b. The absolute pairwise distances generated between undefined terminal clades and described mouse lemur species corresponds to the observed interspecific values found between described species (Andriantompohavanaet al. 2006;Louis Jr.et al. 2006a;Olivieriet al. 2007). Although the absolute pairwise distance betweenM. mamiratra andMicrocebus sp. nova #5 is the smallest percentage between the terminal clades, the geographic distance between sampling sites is also reduced (Appendix II(h)). Values ranged mostly from 10% to 15% with the lowest percentage found betweenMicrocebus sp. nova #5 andM. mamiratra (4.9% and 2.5%, D-loop and PAST, respectively) and the highest percentage was found betweenM. ravelobensis andM. jollyae (24.3% and 10.7%, D-loop and PAST, respectively).
Figure 2.
Neighbor-joining phylogram derived from the D-loop DNA sequence data from 82Microcebus individuals with 18 out-group taxa. Species designated according to the distribution in the current literature (Andriantompohavanaet al. 2006;Louis Jr.et al. 2006a;Mittermeieret al. 2006;Olivieriet al. 2007). Values above branches indicate number of changes between nodes. Values within circles indicate support of bootstrap pseudoreplicates.
Figure 3.
Fifty percent majority-rule consensus phylogenetic tree from the Bayesian analysis derived from the PAST sequence data from 97 haplotypes from 121Microcebus individuals with 18 out-group taxa reconstructed using the computer program package MrBayes. Branches without posterior probability values are supported by less than 50% of the sampled trees.
Figure 4.
Maximum-likelihood phylogram derived from concatenated D-loop and PAST sequence data from 107Microcebus haplotypes with 18 out-group taxa. The phylogram is presented with branch lengths proportional to the number of changes (values specified on the branches). We obtained the maximum likelihood phylogram (-ln likelihood=4921.54) from the D-loop and PAST concatenated alignment (K=7) and γ shape parameter of 1.07.
Figure 5.
Minimum spanning network ofMicrocebus D-loop haplotypes calculated using Arlequin 2.0 and Network 4.11. Identification numbers denote unique haplotypes. The minimum number of mutational steps separating matriarchal lines is indicated above the branches. Nucleotide substitutions are indicated by dashes. The number of nucleotide differences (more than two) in their connecting lines of the network is indicated by the number at each connecting link. Missing intermediates are indicated by conical pink circles. The size of circles approximates the number of individuals with matching haplotypes corresponding to information inAppendix III(d) (circles without any number represent one individual).
Table 2.
Morphometric data collected from sedatedMicrocebus individuals. (Individual morphological data available online; seeAppendix I). Morphological data taken from immobilized animals.
Table 3.
Summary of Population Aggregate Analysis (PAA) D-Loop diagnostic sites for the genusMicrocebus. Refer toAppendix III(b).
Table 4.
Summary of Population Aggregate Analysis (PAA) Pastorini fragment diagnostic sites for the genusMicrocebus. Refer toAppendix III(c).
Table 5a.
Genetic distance matrix for D-loop sequence data for the genusMicrocebus. 1.M. tavaratra; 2.M. ravelobensis; 3.M. sp. nova #5; 4.M. sambiranensis; 5.M. sp. nova #6; 6.M. mamiratra; 7.M. berthae; 8.M. murinus; 9.M. rufus; 10.M. simmonsi; 11.M. mittermeieri; 12.M.jollyae; 13.M. lehilahytsara; 14.M. griseorufus; and 15.M. myoxinus. Genetic distance based on absolute differences is displayed above the diagonal, and genetic distance as a percentage is displayed below the diagonal.
Table 5b.
Genetic distance matrix for PAST fragment sequence data for the genusMicrocebus. 1.M. tavaratra; 2.M. ravelobensis; 3.M. sp. nova #5; 4.M. sambiranensis; 5.M. sp. nova #6; 6.M. mamiratra; 7 .M. berthae; 8.M. murinus; 9.M. rufus; 10.M. simmonsi; 11.M. mittermeieri; 12.M.jollyae; 13.M. lehilahytsara; and 14.M. griseorufus. Genetic distance based on absolute differences is displayed above the diagonal, and genetic distance as a percentage is displayed below the diagonal.
Discussion
The persistent and rapid loss of habitat and the resulting fragmentation of panmictic populations have compelled wildlife and conservation agencies to define management decisions according to existing guidelines and data with the ultimate goal of prioritizing species and/or sites (Wilméet al. 2006;Kremenet al. 2008). Many studies have shown that molecular genetics technology offers a reliable and rapid method of identifying unique and cryptic biodiversity (Louis Jr.et al. 2006a;Olivieriet al. 2007;Radespielet al. 2008). With this in mind, we present another revision of the genusMicrocebus, concentrating on the biogeographic distribution of the mouse lemurs in northern and northwestern Madagascar. Through the analyses of accessioned and novel sample sets, we found that each described mouse lemur clusters in distinct and well-supported terminal clades.
Since Radespielet al. (2008) demonstrated the same result with an alternative data set, a singular terminal clade for both described mouse lemur species from the island of Nosy Be, we have established thatMicrocebus mamiratra has precedence overM. lokobensisAndriantompohavanaet al. 2006, which should consequently be regarded as a junior synonym. Furthermore, the distribution ofM. mamiratra not only extends throughout the island of Nosy Be, but also exists on mainland Madagascar, occupying IRS VI (Olivieriet al. 2007; seeFig. 1).
In addition to the well-supported terminal clades of the 15 acknowledged mouse lemur species, the data revealed a distinct clade for the mouse lemur initially proposed in Andriantompohavanaet al. (2006) at Antafondro Classified Forest, and also showed a remarkable cryptic diversity from Montagne d'Ambre National Park (Figs. 2–4). Three main criteria provide support for the definition of the two new species indicated, as follows: molecular genetic parameters, geographic and topographic barriers, and relative partitions between species.
By providing the initial criterion for the justification of species-level status for the two undefined mouse lemur taxa, molecular genetic data and inference offers the first line of argument. According to the Phylogenetic Species Concept (PSC)sensu Wheeler and Platnick (2000;Groves 2001;Louis Jr.et al. 2006a), diagnostic characters or attributes define Evolutionary Significant Units (ESUs). Several authors suggest that ESUs are equivalent to species and reflect species barriers (Cracraft 1983). Given this criterion, the two undefined species had multiple molecular diagnostic sites (Tables 3 and4). The constant addition of samples to the PAA data set will continue to test the distinction and diagnostic ability of these characters; and, therefore, the ongoing status of each species.
The second line of argument is as follows. The two undefined mouse lemur taxa,Microcebus sp.nova #5 and Microcebus sp.nova #6, have distributions defined by geographic and topographic barriers. Following the initial proposal in Andriantompohavanaet al. (2006),Microcebus sp.nova #5 is bounded by the Andranomalaza River to the northwest, the Sambirano River to the northeast, and the Maevarano River to the south (Fig. 1). Although the Andranomalaza River does not meet the large river criterion (>50 m wide, 20 km inland), geographic barriers in combination with the small size of mouse lemurs and limited dispersal ability essentially could drive allopatric speciation (Wilméet al. 2006). With the distribution ofM. mamiratra extended to mainland Madagascar (directly east of the island of Nosy Be in IRS VI), the topographic presence of Tsaratanana, one of the three mountains in Madagascar with an altitude above 2,000 m, could create a significant geographic barrier toMicrocebus sp.nova #5 just north of the Sambirano River.Microcebus sp.nova #6 is found in the montane rainforest of Montagne d'Ambre National Park, north of the Irodo River. Ankarana National Park and Analamerana Special Reserve establish the southern boundary to this undefined mouse lemur's range. As a limestone plateau and tsingy formation intermixed with dry deciduous forest, Ankarana and Analamerana could be acting as a significant barrier to dispersal (Fig. 1). Additionally, the Bobakindro River courses along the northern margin of Analamerana Special Reserve. Again, the Irodo and Bobakindro Rivers do not meet the criterion of a major river barrier, however the topographic features and habitat differences offer strong support for the uniqueness of this undefined species.
Third, each undefined mouse lemur is found paired geographically (smallest geographic distance) with a defined species that is also segregated by an Inter-River-System but is not its genetically most proximal sister taxon (Fig. 1). All three phylogenetic analyses, along with the spanning network, demonstrated the phylogenetic proximity between Microcebus sp.nova #5 andM. mamiratra, on the one hand, andMicrocebus sp.nova #6 andM. sambiranensis, on the other. WithMicrocebus sp. nova #5 at Antafondro, the distribution ofM. sambiranensis would be limited to the Manongarivo Special Reserve, north of the Andranomalaza River and south of the Sambirano River, placing its range in between the distribution ofM. mamiratra and the undefined species. Similarly, the distribution ofM. tavaratra in Ankarana, Andra-fiamena, Analamerana, and Andavakoera andM. mamiratra in Manehoka provide a significant species barrier betweenM. sambiranensis and its genetically closest sister taxa Microcebus sp.nova #6.
Species Descriptions
Microcebus margotmarshae new species
FormerlyMicrocebus sp.nova #5; initially proposed in Andriantompohavanaet al. (2006). SeeFig. 6,Appendix II(f).
Holotype. TAFO6.1; adult female captured in Antafondro Classified Forest on 21 May 2006. Material: Total genomic DNA (50 ng/µl) for TAF06.1 (Bar Code 145314), adult female. Total genomic DNA materials are stored and curated at the Museum of Texas Tech University, Lubbock, Texas, USA. Two 2.0-mm biopsies from ear pinna tissue are stored at Henry Doorly Zoo, Omaha, Nebraska, USA. A microchip pit tag was placed subcutaneously between scapulas and recorded as 4722607B5D. TAF06.1 was collected by Francois Randrianasolo, Richard Rakotonomenjanahary, Jean Amié Andriamihaja, and Rambinintsoa Andriantompohavana on 21 May 2006.
Paratypes. TAF06.2 (Bar Code 145315), adult female and ANT5.1 (Bar Code 145313), adult male; captured in Antafondro Classified Forest. Total genomic DNA (50 ng/µl) TAF06.2 (Bar Code 145315), adult female; and ANT5.1 (Bar Code 145313), adult male; are stored and curated at the Museum of Texas Tech University, Lubbock, Texas, USA. Two 2.0-mm biopsies from ear pinna tissues are stored at Henry Doorly Zoo, Omaha, Nebraska, USA. Individual measurements, e-voucher photos, and collection data are given inAppendix I(b) and are available at the Museum of Texas Tech University, Lubbock, Texas, USA. Francois Randrianasolo, Richard Rakotonomenjanahary, Jean Amié Andriamihaja, and Rambinintsoa Andriantompohavana collected TAF06.2 and ANK5.1 on 21 May 2006 and 4 October 2005, respectively.
Type Locality. Madagascar: Province de Antsiranana, Antafondro Classified Forest Special Reserve (approximately 14°02′44.5″S, 48°13′23.4″E, 134 m above sea level).
Measurements of holotype. Recorded in the field catalog on 21 May 2006: weight: 49.0 g; head crown: 3.2 cm; body length: 8.4 cm; tail length: 14.3 cm; muzzle length: 9.5 mm; ear length: 15.4 mm; and ear width: 8.7 mm.
Description.Microcebus margotmarshae is a small mouse lemur (41.0 g). The dorsal and tail pelage is predominantly reddish-orange with gray undertones, (Fig. 6;Appendix II(g)). The ventral fur is white to cream. The head is largely bright reddish-orange. The ears are small. The muzzle and the area surrounding the eyes are light brown, and there is a small, bright white spot on the nose ridge between the eyes.
Diagnosis. In the D-loop and PAST sequence fragments,M. margotmarshae differs from its closest relatives,M. tavaratra,M. sambiranensis,M. mamiratra andM. arnholdi, by both genetic and geographic distance by 12.3% ±1.6% (73 informative sites), 9.5%±1.4% (42 informative sites), 4.9%±1.0% (23 informative sites) and 9.5%±1.3% (57 informative sites); 9.7%±0.7% (227 informative sites), 5.1%±0.5% (132 informative sites), 2.5%±0.3% (54 informative sites) and 5.1%±0.5% (134 informative sites), respectively. Even thoughM. margotmarshae is a rufous-type mouse lemur asM. mamiratra (genetically the closest related),M. margotmarshae (41.0 gm) is significantly smaller thanM. mamiratra (60.8 gm).
Distribution.Microcebus margotmarshae is known from the Antafondro Classified Forest, south of the Andranomalaza River and north of the Maevarano River, Madagascar.
Comparisons and remarks. Andriantompohavanaet al. (2006) proposed that the mouse lemurs from Antafondro Classified Forest should be considered a separate species (Microcebus sp.nova #5), based on the PAST sequence fragment from one individual that was included in the analyses (Table 3 and4;Appendix III(b–c)). Of the recognized mouse lemurs that are in the adjacent regions of Madagascar,Microcebus margotmarshae (41.0 gm) is approximately the same size asM. sambiranensis (44.0 gm), but smaller thanM. mamiratra (60.8 gm),M. tavaratra (52.3 gm), andM. ravelobensis (65.9 gm). Additional samples from the entire region south of the Andranomalaza River and north of the Maevarano River are needed to define the distribution ofM. margotmarshae. Olivieriet al. (2007) presented the course of the Maevarano River in an east to west direction, when, in fact, this river travels in more of a northwest to southeast direction, increasing as such the size of IRS V (Fig. 1). Samples should be collected from mouse lemurs from Tsaratanana Special Reserve. It is possible that mouse lemurs can be found at high altitudes there.
Etymology. Microcebus margotmarshae is named in honor of the late Margot Marsh, who during her lifetime contributed very generously to primate conservation initiatives in many different countries, including the publication of the first edition of the field guide Lemurs of Madagascar in 1994 (Mittermeieret al. 1994). The Margot Marsh Biodiversity Foundation was created after her death in 1995, thus continuing support for efforts that help safeguard the future of threatened primates.
Vernacular names. Margot Marsh's mouse lemur or Antafondro mouse lemur.
Microcebus arnholdi new species
FormerlyMicrocebus sp.nova #6 (Fig. 7,Appendix 11(g)).
Holotype. AMB5.39; adult female; collected on 27 November 2005, captured at Montagne d'Ambre National Park. Material: Total genomic DNA (50 ng/µl) for AMB5.39 (Bar Code 145310), adult female stored and curated at the Museum of Texas Tech University, Lubbock, Texas, USA. Two 2.0-mm biopsies from ear pinna, and 0.07 cc of whole blood tissues stored at Henry Doorly Zoo, Omaha, Nebraska, USA. A microchip pit tag was placed subcutaneously between the scapulas and recorded as 4657027B18. AMB5.39 was collected by Richard Randriamampionona, Richard Rakotonomenjanahary, Jean Amié Andriamihaja, Fidelis Razafimananjato Tsirivaliniaina, John R. Zaonarivelo, and Edward Louis Jr. on 27 November 2005.
Paratypes. AMB5.40 (Bar Code 145311), adult female; and AMB5.43 (Bar Code 145312), adult female; captured at Montagne d'Ambre National Park. Material: Total genomic DNA (50 ng/µl) for each are stored and curated at the Museum of Texas Tech University, Lubbock, Texas, USA. Two 2.0-mm biopsies from ear pinna, and 0.07 cc of whole blood tissues stored at Henry Doorly Zoo, Omaha, Nebraska, USA. Individual measurements, e-voucher photos, and collection data are given inAppendix I(b) and are available at the Museum of Texas Tech University, Lubbock, Texas, USA. Richard Randriamampionona, Richard Rakotonomenjanahary, Jean Amié Photo by Edward E. Louis Jr. Andriamihaja, Fidelis Razafimananjato Tsirivaliniaina, John R. Zaonarivelo, and Edward Louis Jr. collected AMB5.40 and AMB5.43 on 28 November 2005.
Type Locality. Madagascar: Province de Antsiranana, Montagne d'Ambre National Park and Montagne d'Ambre Special Reserve (approximately 12°31′28.1″S; 049°10′22.8″, 990 m above sea level).
Measurements of holotype. AMB5.39; adult female. Recorded in the field catalog on 21 November 2005. Weight 71.0 grams; head crown 3.3 cm; body length 8.1 cm; tail length 12.9 cm; muzzle length 9.4 mm; ear length 17.8 mm; and ear width 10.1 mm.
Description.Microcebus arnholdi is a medium-sized mouse lemur (49.7 gm). The overall dorsal pelage is a mixture of dark brown, red and gray (Fig. 7;Appendix II(g)). There is a dark brown midline dorsal stripe that runs down to the base of the tail. The tail is dark brown near the tip. The ventral fur is white to cream, with gray undertones. The head is predominately red, with dark brown on the muzzle and surrounding the eyes and with a white nose ridge that stops at the distal end of the muzzle. The ear length ofM. arnholdi (17.5 ±0.4 mm) is smaller thanM. tavaratra (21.7±0.7 mm).
Diagnosis. In the D-loop and PAST sequence fragments,M. arnholdi differs from its closest relatives,M. tavaratra,M. sambiranensis,M. mamiratra andM. margotmarshae, in both genetic and geographic distance, by 12.6% ±1.5% (91 informative sites), 6.9% ±1.1% (53 informative sites), 9.6%±1.3% (60 informative sites) and 9.5%±1.3% (57 informative sites); 9.2%±0.7% (233 informative sites), 3.5%±0.4% (113 informative sites), 5.0%±0.5% (151 informative sites) and 5.1%±0.5% (134 informative sites, respectively). Of the recognized mouse lemurs that are in the adjacent regions of Madagascar, Microcebus arnholdi (49.7 gm) is smaller thanM. tavaratra (52.3 gm), andMicrocebus mamiratra (60.8 gm). Even thoughM. arnholdi is a rufous type mouse lemur asM. mamiratra, the pelage ofM. arnholdi is more grayish brown.
Distribution.Microcebus arnholdi is known from the Montagne d'Ambre National Park and Special Reserve, northwest of the Irodo River, Madagascar.
Comparisons and remarks.Microcebus arnholdi can be found in montane rainforest, whereasM. tavaratra occupies the dry deciduous forest in the Ankarana and Analamerana IRS VII;Fig. 8). As shown inFigure 8,M. arnholdi is a new species in a new Inter-River-System (IRS) VIII; the tenth IRS in northern and northwest Madagascar.Figure 8 also illustrates the need for comprehensive sampling in this intensely researched region of Madagascar, a detailed distribution map of the species sampled that correlates to the existing forest tracts, and accurate mapping of the course of all river systems. The distributions of other genera in the region should be overlaid to provide us with a better understanding the biogeography of lemurs in general. Lastly, molecular genetic data should be generated for the all lemur holotypes, and included in the phylogenetic inferences and diagnostic evaluations of lemur taxonomy. Of the recognized mouse lemurs that are in the adjacent regions of Madagascar,Microcebus arnholdi (49.7 gm) is slightly larger thanM. sambiranensis (48.0 gm), but smaller thanM. tavaratra (52.3 gm),Microcebus mamiratra (60.8 gm), andM. ravelobensis (65.9 gm).
Etymology. The namearnholdi honors Henry Arnhold of New York, who has supported conservation efforts throughout the developing world, with a particular focus on linking the well-being of the people with the protection of their environment. Conservation International's Healthy Communities Initiative and Conservation Stewards' Program has come into existence because of Mr. Arnhold's commitment to linking the well-being of people with the protection of critically important biodiversity hotspots. Madagascar has been among the places that have benefited substantially from the support that Henry Arnhold has provided. By naming this species after him, we recognize his great commitment and express the appreciation of the conservation community for all that he has done to further the cause for biodiversity conservation in Madagascar and around the world.
Vernacular names. Arnhold's mouse lemur or Montagne d'Ambre mouse lemur.
Figure 6.
Microcebus margotmarshae, Margot Marsh's or Antafondro mouse lemur, at Antafondro Classified Forest (Maromiandra). Photo by Rambinintsoa Andriantompohavana.
Figure 7.
Microcebus arnholdi, Arnhold's or Montagne d'Ambre mouse lemur, at Montagne d'Ambre National Park and Classified Forest.
Figure 8.
Distribution of the mouse lemurs (genusMicrocebus) of northern and northwestern Madagascar. Color-coded circles represent the samples (sites) that were included in the analyses (these samples include accessioned GenBank sequences and the colors are species specific (see Appendices II(h) and III(a)). The map was modified from an image provided by Conservation International, Arlington, VA (Harperet al. 2007), and incorporates the Landsat Enhanced Thematic Mapper Plus (ETM+) data from 1999–2001, predominantly from 2000. Colored triangles represent accessioned samples not used in this study, but the color of the triangle is representative of a specificMicrocebus species.
Note
As discussed in Andriantompohavanaet al. (2006,2007), Louis Jret al. (2006a, 2006b), and Thalmann and Geissmann (2005), the use of whole vouchers as the designated holotype for a new species is not a prerequisite for describing an undefined species. Opportunistic collection, however, can later supplement morphological, and/or molecular data in combination with curated blood and/or tissue samples. Total genomic DNA for the holotypes and paratypes of the newly describedMicrocebusmargotmarshae andMicrocebus arnholdi, along with e-vouchers and field data, are currently curated at the Museum of Texas Tech University, Lubbock, Texas, USA, under the following catalogue numbers: TK145310; TK145311; TK145312; and TK145313; TK145314; TK145315, respectively;Appendix I(a).
Acknowledgments
This project would not have been possible without the support of the Association Nationale pour la Gestion des Aires Protégées (ANGAP), the Ministère des Eaux et Forêts of Madagascar, the staff, guides, and drivers of the Institute for Conservation of Tropical Environments, Madagascar (ICTE-MICET), as well as, Parc Botanique et Zoologique de Tsimbazaza, and US Fish and Wildlife Service. This manuscript was supported in part by a grant from the Ahmanson Foundation. We also thank the reviewers for their comments, suggestions, and helpful critiques of this manuscript. We acknowledge the field efforts of Rambinintsoa Andriantompohavana and John Zaonarivelo, and the generosity of Bill and Berniece Grewcock for their long-term support and commitment. Furthermore, we would like to acknowledge that this research would not be possible without the incredible support by the Theodore F. and Claire M. Hubbard Family Foundation, and the James Family. Thanks to Lisa Kimmel for creating the various web page and manuscript documents, and John Musinsky and Marc Steininger of the Center for Applied Biodiversity Science, Conservation International, for the use of the Landsat image of Madagascar.
Literature Cited
Appendices
The following Appendices are available online at the indicated website addresses and can be downloaded as pdf documents.
Appendix I
Appendix 1(a).Microcebus table of individual samples and corresponding information for each sample (bar code number, site, original species designation, current species designation, GenBank accession numbers of sequence data).
Appendix 1(b). Field notes forMicrocebus margotmarshae (formerlyMicrocebus sp.nova #5) andMicrocebusarnholdi (formerlyMicrocebus sp.nova #6)
Appendix II
Appendix 11(a). Maximum parsimony phylogram derived from the D-loop sequence data from 82Microcebus individuals with 18 out-group taxa (one of 3886 most parsimonious trees). Values above branches indicate number of changes between nodes. Values within the circles along the branches indicate support of bootstrap pseudoreplicates. Length=2052; CI = 0.4235; RI = 0.8460; RC = 0.3583; HI = 0.5765.
Appendix 11(b). Part A. Neighbor-joining phylogram derived the D-loop DNA sequence data from the 121Microcebus individuals with 18 out-group taxa. Values above branches indicate number of changes between nodes. Values within circles indicate support of bootstrap pseudoreplicates. Solid black circle indicates the branch that connects in-group taxa to the out-group taxa (displayed on next page (Part B)).
Appendix 11(b). Part B. Neighbor-joining phylogram derived the D-loop DNA sequence data from the 121Microcebus individuals with 18 out-group taxa. Values above branches indicate number of changes between nodes. Values within circles indicate support of bootstrap pseudoreplicates. Solid black circle indicates the branch that connects to the in-group taxa (displayed on previous page (Part A)).
Appendix 11(c). Maximum parsimony phylogram derived from the D-loop sequence data from 77Microcebus haplotypes with 18 out-group taxa (one of 364 most parsimonious trees). Values above branches indicate number of changes between nodes. Values within the circles along the branches indicate support of bootstrap pseudoreplicates. Length=2138; CI = 0.4574; RI = 0.8578; RC = 0.3924; HI = 0.5426.
Appendix 11(d). Fifty percent majority-rule consensus phylogenetic tree from the Bayesian analysis derived from the D-loop sequence data from 77 Microcebus individuals with 18 out-group taxa reconstructed using the program MrBayes. Branches without posterior probability values (PP) are supported by less than 50% of the sampled trees.
Appendix 11(e). Maximum parsimony phylogram derived from the D-loop and PAST sequence data from 89 Microcebus individuals with 18 out-group taxa (one of 4112 most parsimonious trees). Values above branches indicate number of changes between nodes. Values within the circles along the branches indicate support of bootstrap pseudoreplicates. Length=6539; CI = 0.4271; RI = 0.8755; RC = 0.3739; HI = 0.5729.
Appendix 11(f).Microcebus margotmarshae, Margot Marsh's or Antafondro mouse lemur at Antafondro Classified Forest (Maromiandra). Photo by Rambinintsoa Andriantompohavana.
Appendix 11(g).Microcebus arnholdi, Arnhold's or Montagne d'Ambre mouse lemur at Montagne d'Ambre National Park and Classified Forest. Photo by Edward E. Louis Jr.
Appendix 11(h). Distribution map of the mouse lemurs of Madagascar. Designated sites and species are based on molecular genetic data. The species legend corresponds to the color coded to the sites.
Appendix III
Appendix 111(a). Summary of the acronyms and GenBank accessioned sequences used in this study.
Appendix 111(b). Table 1A. Diagnostic nucleotide sites from the D-loop Pairwise Aggregate Analysis (PAA) of genus Microcebus.Appendix III(c). Table 1B. Diagnostic nucleotide sites from the PAST Pairwise Aggregate Analysis (PAA) of genus Microcebus.
Appendix III(d). Haplotypes forMicrocebus D-loop SequencesAppendix III(e). Haplotypes for Microcebus PAST Sequences.
