Historical biogeography of tamarins, genusSaguinus: the molecular phylogenetic evidence

Historical Biogeography of Tamarins, Genus Saguinus:The Molecular Phylogenetic Evidence

SUSAN JACOBS CROPP,1,2* ALLAN LARSON,1AND JAMES M. CHEVERUD2

1Department of Biology, Washington University, St. Louis, Missouri 631302Department of Anatomy and Neurobiology, Washington University Schoolof Medicine, St. Louis, Missouri 63110

KEY WORDS dispersal; phylogeny; mtDNA

ABSTRACT Hypotheses of the historical biogeography of tamarins (genusSaguinus) based on variation in coat colors and body size are tested usingphylogenetic relationships inferred from mitochondrial DNA (mtDNA) se-quence data. Samples from all 12 species of Saguinus and several subspeciesare included in the analysis. Approximately 1,200 bases of mtDNA sequencefrom the cytochrome b and D-loop regions are reported for the tamarins andseveral outgroup taxa. Parsimony analysis of the mtDNA sequence datareveals Saguinus to be a monophyletic taxon composed of two major clades:one, the Small-bodied clade, contains S. nigricollis, S. tripartitus, and S.fuscicollis, and the other, the Large-bodied clade, contains the other ninespecies. The phylogenetic relationships among tamarins inferred from themtDNA sequence data reject previous hypotheses for the historical biogeogra-phy of tamarins and suggest different dispersal routes for this group of NewWorld monkeys. The molecular data suggest that tamarins dispersed acrossSouth America in two major waves from an origin somewhere south of theAmazon. One wave moved in a westerly direction, whereas the other moved ina northeastern direction toward the Amazon delta and then west along thenorthern portion of the continent into northern Colombia and Panama. Am JPhys Anthropol 108:65–89, 1999. r 1999 Wiley-Liss, Inc.

Of the South American primate genera,Saguinus is one of the largest in numbersof species and geographical distribution(Hershkovitz, 1977). Tamarins are thoughtto be most closely related to the marmosets(Callithrix and Cebuella), the lion tamarins(Leontopithecus), and Goeldi’s monkey (Cal-limico) (Hershkovitz, 1977; Mittermeier andCoimbra-Filho, 1981; Rylands et al., 1993;Jacobs et al., 1995). The phylogenetic rela-tionships among the five aforementionedgenera are not well resolved, and muchdebate exists concerning their proper taxo-nomic affiliations. Hershkovitz (1977) andothers (Mittermeier and Coimbra-Fihlo,1981; Ford, 1986; Rylands et al., 1993) placeSaguinus, Callithrix, Cebuella, and Leon-topithecus in the family Callitrichidae

and Callimico in the Family Callimiconidae.Some (Rosenberger, 1981a; Schneider et al.,1993; Harada et al., 1995), however, wouldplace all five genera in the subfamily Callit-richinae of the family Cebidae. Regardless ofthe higher-level taxonomy, most would agreethat tamarins, marmosets, and Callimicoform a monophyletic group.

There are twelve recognized species oftamarins, ranging from the Amazon basinnorthward into Panama (Hershkovitz, 1977;Mittermeier and Coimbra-Fihlo, 1981; Thor-ington, 1988; Moore and Cheverud, 1992;

Grant sponsor: National Science Foundation; Grant number:9411169.

*Correspondence to: Susan Cropp, University of Texas, Hu-man Genetics Center, 6901 Bertner Ave., S250, Houston, TX77030. E-mail: [email protected]

Received 22 October 1997; accepted 22 September 1998.

AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 108:65–89 (1999)

r 1999 WILEY-LISS, INC.

Rylands, 1993). Tamarins are typically foundin edge and secondary forest habitats. Mostof the species are allopatric, with the mostnotable exception being S. fuscicollis, whichis sympatric with other tamarin speciesthroughout much of its range (Fig. 1). Themost common boundaries between taxa arerivers (Peres et al., 1996), and Ayres andClutton-Brock (1992) report that rivers widerthan about 0.3 km are sufficient barriers totamarin dispersal.

In his classic volume on New World mon-keys, Hershkovitz (1977) proposed the dis-persal routes for Saguinus shown in Figure2A. The ancestral tamarin was postulated tohave arisen in the central Amazon basin inthe general vicinity of the current distribu-tion of S. fuscicollis and subsequently tohave spread northeast across the Amazon togive rise to S. midas and S. bicolor. Anotherwave of dispersal north along the Andes wasproposed to have resulted in the extantspecies in northern Colombia and Panama.This hypothesis is based largely on theprinciple of metachromism and the currentdistribution of Saguinus. Metachromism isa theory of mammalian coat color evolutionthat Hershkovitz developed, in part, fromhis studies on geographic variation in pelageamong tamarins (Hershkovitz, 1968, 1977).According to the theory, the primitive statefor mammalian coat color is agouti and canevolve in an orthogenetic fashion througheither of two color pathways, both leadingfinally and irreversibly to white. The phe-nomenon is proposed to be geographic andphylogenetic (Hershkovitz, 1968, 1977).However, because of extensive parallelismsamong species, coat color is thought not to bea reliable phylogenetic character for distin-guishing evolutionary relationships amongspecies (Shedd and Macedonia, 1991; Jacobset al., 1995). A hypothetical phylogeny con-structed from Hershkovitz’s (1977) descrip-tion of species relationships based on thetheory of metachromism and biogeographyis shown in Figure 3A. He applied the theoryof metachromism to subspecies relation-ships as well. The largest of the subspeciesgroups is S. fuscicollis; phylogenetic relation-ships predicted by Hershkovitz’s (1977)

analyses of these subspecies are shown inFigure 3B.

An alternative to Hershkovitz’s theory ofdispersal has been proposed by Ferrari(1993a), based primarily on the hypothesisthat callitrichids are phyletic dwarfs. Tradi-tionally, callitrichids had been consideredprimitive due to their small size (and associ-ated ecological characteristics) compared toother New World monkey families (Hershko-vitz, 1977). However, according to the phy-letic dwarfism hypothesis, the precursor tothe callitrichid lineage was larger in bodysize than any of the extant species, whichdisplay a secondary reduction in body size(Ford, 1980; Leutenegger, 1980). Consider-ing body-size differences and related eco-logical variables to be the most importantfactors in the evolutionary history of callitri-chids (Ferrari, 1993b), Ferrari (1993a) for-mulated a theory of the phylogenetic rela-tionships and dispersal of tamarins based onthese factors.

Ferrari’s (1993a) proposed dispersal oftamarins across the Amazon basin is de-picted in Figure 2B. The ancestral tamarin,or prototype, is hypothesized to have beenmidas-like in size (average male size 5 533 g[Fleagle and Mittermeier, 1980]) and isthought to have arisen in the general vicin-ity of the present S. midas distribution.From the lower Amazon, tamarins dispersedsouthwest, giving rise to the moustachedtamarin group (S. mystax, S. labiatus, S.imperator). The final radiation of Saguinuswould have given rise to the saddlebacktamarins (S. fuscicollis). The origin of thespecies in northern Colombia and Panama(S. oedipus, S. leucopus, and S. geoffroyi) isnot discussed. A hypothetical phylogenybased on Ferrari’s (1993a) discussion of spe-cies relationships is shown in Figure 4A.

Although both of these dispersal theoriesare based, in part, on some hypothesis of thephylogenetic relationships among tamarins,neither of these studies utilized a phyloge-netic data analysis. The purpose of thisstudy is to test each of the above dispersalhypotheses for tamarins using phylogeneticevidence generated from mitochondrial DNAsequence data. The molecular phylogenetic

66 S.J. CROPP ET AL.

Fig. 1. Geographic distribution of Saguinus (with the exception of S. fuscicollis and S. tripartitus) A:and geographic distribution of S. fuscicollis subspecies and S. tripartitus B: based on Hershkovitz (1977)and others (for a review see Rylands et al., 1993).

67HISTORICAL BIOGEOGRAPHY OF TAMARINS

Fig. 2. Hypothesized historical dispersal routes for Saguinus based on Hershkovitz (1977) A: andFerrari (1993a) B.


relationships of portions of the genus havebeen reported (Jacobs et al., 1995). Thepresent study extends the previous one toinclude all 12 recognized species of Sagui-nus.

METHODS AND MATERIALSDNA extraction, amplification,

and sequencing

Samples from ten species and subspeciesof Saguinus, in addition to those repre-sented in Jacobs et al. (1995), were obtainedfrom a variety of sources (Table 1). GenomicDNA was extracted from the museum skinspecimens (2–3 mm2 of skin) using the tis-sue protocol of the QIAamp Tissue Kit (Qia-gen Inc., Chatsworth, CA). This same proto-col was used to extract genomic DNA froman immortal cell line of S. leucopus.

Three segments of mitochondrial DNA(mtDNA) in the cytochrome b and D-loop

regions were amplified and sequenced forthis analysis (Fig. 5). These regions havebeen shown previously to be phylogeneti-cally informative for Saguinus (Jacobs et al.,1995). Because of the degraded nature ofDNAextracted from museum skins, tamarin-specific primers were designed for this studyto amplify smaller portions of the regions ofinterest. Primer sequences are listed in Table2; all were used for both polymerase chainreaction (PCR) amplification and sequenc-ing. Double-stranded PCR amplificationswere run for 30 cycles under the followingconditions: 92–94°C denaturation (35 sec),50–55°C annealing (35 sec), and 70–72°Cextension (2.5 min for the first cycle with anaddition 4 sec for each ensuing cycle).

Because only a limited amount of DNAcould be extracted from museum skins, 25 µlamplifications were first performed for thesesamples. DNA extracted from museum skins

Fig. 3. A: Hypothetical phylogeny for Saguinus basedon species relationships described by Hershkovitz (1977)(redrawn from Jacobs et al., 1995). B: Hypotheticalrelationships among the S. fuscicollis subspecies basedon Hershkovitz (1977) (redrawn from Cheverud andMoore, 1990).

Fig. 4. A: The hypothetical phylogenetic relation-ships among Saguinus based on Ferrari’s (1993) descrip-tion of species relationships. B: The most parsimoniousbifurcating tree (see Results) consistent with the po-lytomy shown in A.


has been found to contain degradation prod-ucts and other contaminants that have aninhibitory effect on PCR amplifications. Toenhance the reactions, we added 10 µg/µlbovine serum albumin (BSA) (Promega,Madison, WI) to the initial 25 µl reactions,

and the amount of TAQ DNA polymerase(Gibco, Inc., Grand Island, NY) in the initialamplifications was increased to twice thenormal amount (Paabo et al., 1988). Theproduct was then purified on a low-meltagarose gel and reamplified in 100 µl reac-

TABLE 1. Sources for Saguinus samples used for this study1

Species Tissue type Source Specimen ID

S. bicolor bicolor Museum skin R. Thorington 2053, 2074, 2075, 2076Extracted DNA H. Schneider 1070

S. bicolor martinsi Extracted DNA H. Schneider 814S. bicolor ochraceus Museum skin AMNH 94093, 94100S. fuscicollis lagonotus Frozen muscle WU none

Museum skin FMNH 60349S. fuscicollis fuscus Museum skin FMNH 123377, 123378S. inustus Museum skin AMNH 78596, 79415, 79418S. leucopus Immortal cell line D. Evans EBU

Museum skin USNM 216697, 216698S. midas niger Museum skin AMNH 10159, 70166S. mystax pluto Museum skin FMNH 134479S. tripartitus Museum skin FMNH 57620

Museum skin AMNH 72098, 1659341 Sample sources for all species other than those listed here can be found in Jacobs et al. (1995). AMNH, American Museum of NaturalHistory, New York; FMNH, Field Museum of Natural History, Chicago; USNM, National Museum of Natural History, Washington, DC;WRPRC, Wisconsin Regional Primate Research Center, Madison; WU, Washington University, Department of Anthropology, St. Louis.

Fig. 5. Regions of mtDNA sequenced for phylogenetic analysis. The attachment sites for the primerslisted in Table 2 are indicated by labeled arrows. Right arrows indicate light-strand primers. and leftarrows indicate heavy-strand primers. F, phenylalanine transfer RNA gene (tRNA); P, proline tRNA; T,threonine tRNA.


tions to obtain sufficient product for thesequencing reactions. Negative controls wererun for all sets of PCR reactions to ensurethe reactions were not contaminated withextraneous DNA. Any reactions containingproduct in the negative controls were dis-carded. Amplification products from the 100µl reactions were purified on a 3.5% poly-acrylamide gel in preparation for the se-quencing reactions.

Double-stranded sequencing reactionswere performed using either Sequenase ver-sion 2.0 (U.S. Biochemical, Cleveland, OH)(Hillis et al., 1996) or using a fmol cyclesequencing kit (Promega Corp., Madison,WI). Radiolabelled 32P or 35S (NEN, Boston,MA) was incorporated into the sequencingreactions, which were electrophoresed on a6% polyacrylamide Long Ranger gel (FMCBioProducts, Rockland, ME). The regionssequenced included approximately 1,200 nu-cleotide bases, which includes more of theD-loop region than was previously reported(Jacobs et al., 1995).

Phylogenetic analysis

Homologous sites of the mtDNA sequenceswere aligned manually and analyzed withPhylogenetic Analysis Using Parsimony(PAUP), version 3.1.1 (Swofford, 1993). Us-ing Callimico, Callithrix, Cebuella, and Le-ontopithecus as the outgroup reference taxa,we used the heuristic search option (randomaddition of taxa, 20 replications) to find themost parsimonious phylogenetic reconstruc-

tion for Saguinus. More alignable sequencedata were obtained for Saguinus than forany of the outgroup taxa. Once the mostparsimonious reconstruction was obtainedfor the tamarins using the outgroup taxa,the phylogenetic relationships among eachof the two major clades within the genuswere analyzed separately, each using theother clade as the outgroup. The purpose ofthe additional analyses was to take advan-tage of the additional data obtained for theSaguinus species and subspecies.

A branch-and-bound analysis was thenperformed on the data using the length ofthe most parsimonious reconstruction as anupper bound. The data were also subjectedto a bootstrap analysis (Felsenstein, 1985b)using the heuristic search option with arandom-addition sequence (three random-addition replications per bootstrap replica-tion) for 200 bootstrap replications. Theaforementioned analyses were performed onthe entire data set including the other callit-richid genera and on the individual cladeswithin Saguinus. Alternative hypotheses forthe entire Saguinus phylogeny and for theS. fuscicollis subspecies were compared us-ing a nonparametric Wilcoxon signed-rankstest (Templeton, 1983a,b; Felsenstein,1985a). Branch lengths for the most parsimo-nious tree were estimated using theACCTRAN and DELTRAN options of PAUP(Swofford, 1993). ACCTRAN resolves hom-plasies preferentially as reversals, whereasDELTRAN favors parallelisms. Homoplasyis quantified using the consistency indexand the homoplasy index.

Geographic distributions

The current geographic distributions forthe 12 tamarin species were obtained mainlyfrom Hershkovitz (1977). Information frommore recent censuses was incorporated aswell (for a review see Rylands et al., 1993).

RESULTS

Sequence data spanning approximately1,200 bases in the cytochrome b and D-loopregions of the mtDNA were obtained for

TABLE 2. Primers used for PCR amplificationand sequencing of the mtDNA1

Primer Sequence (58 to 38)

282 (H)2 AAGGCTAGGACCAAACCT283 (L)2 TACACTGGTCTTGTAAACC464 (L)2 TGAATTGGAGGACAACCAGTSCJ1 (H) GAGCGAGAATACTAGTAGAAGSCJ2 (L) ACCCTTCAGAGAATAAACTTASCJ3 (L) GTTAGTCATTCAGGGGATASCJ4 (L) GCACTAATTACATAACCAASCJ5 (H) TTGGTTATGTAATTAGTGCH15149 (H)2 TGACTGTGGCACCTCAGAATGATATT-

TGGCCTCAL14996 (L)2 AGCCCCATCCAACATCTCTGCTTGAT-

GAAA1 An (L) indicates a light-strand primer, and an (H) indicates aheavy-strand primer. Attachment sites for these primers areshown in Fig. 5.2 For references for these primers, see Jacobs et al., 1995; allothers were designed for this study.


Saguinus as well as representatives fromCallimico, Callithrix, Cebuella, and Leonto-pithecus. The aligned sequence data can befound in the Appendix. For many of themuseum skin specimens, the DNA was de-graded, and only small fragments could beamplified and sequenced. Pairwise dis-tances between the taxa (adjusted for miss-ing data) are listed in Table 3.

An exact search of the aligned sequencedata using Callimico as an outgroup yieldeda single most parsimonious tree (Fig. 6).This tree is 1,325 steps long, with a consis-tency index of 0.543 and a homoplasy indexof 0.457. The branch lengths were calculatedwith both the ACCTRAN and DELTRANoptions of PAUP. Heuristic searches of thealigned sequence data using other callitri-chid outgroups resulted in less resolutionbut were consistent with the tree in Figure 6.

As demonstrated before (Jacobs et al.,1995), Saguinus is a monophyletic taxoncomposed of two major clades: one contain-ing S. nigricollis, all the S. fuscicollis subspe-cies, and S. tripartitus, and the other cladecontaining all remaining species. All speciesin this second clade have a 77 base deletionin the middle of the D-loop region comparedto the nigricollis/fuscicollis/tripartitusclade. For simplicity, the clade containing S.nigricollis, S. tripartitus, and the S. fuscicol-lis subspecies will hereafter be known as theSmall-bodied clade and the other major cladewill be termed the Large-bodied clade, inreference to the relative body sizes of thespecies in the two clades.

The addition of S. tripartitus and S. fusci-collis fuscus to the analysis has revealed S.fuscicollis to be a paraphyletic taxon; S. f.fuscus is most closely related to S. nigricol-lis, and S. tripartitus is the sister taxon of S.f. lagonotus. However, with the exception ofS. fuscicollis, all other species containingmultiple subspecies are monophyletic. Sagui-nus mystax appears to form a monophyleticgroup with S. labiatus and S. imperator, arelationship that was not revealed with theprevious, smaller data set (Jacobs et al.,1995). Saguinus leucopus is the sister taxonof S. oedipus and S. geoffroyi, which is notsurprising given the close proximity of the

geographic ranges of these three species.That clade is the sister group to the cladecontaining S. midas and S. bicolor. Sagui-nus inustus, a species about which very littleis known, appears to be the sister taxon tothe labiatus/imperator/mystax clade.

The 50% majority-rule bootstrap analysisyielded the topology shown in Figure 7.Many of the nodes in this tree appear wellresolved, as indicated by bootstrap valuesgreater than 70. The bootstrap analysissupports Saguinus as a monophyletic taxoncontaining two major clades. Although therelationships within the Small-bodied cladeare not all well resolved, S. fuscicollis stillappears to be paraphyletic. The bootstrapanalysis supports the conclusion that S.mystax, S. midas, and S. bicolor each formsmonophyletic groups of populations. Thestructure of the clade containing S. geoffroyi,S. oedipus, S. leucopus, S. midas, and S.bicolor also is well supported. The deeper-level relationships within the second majorclade lack resolution in this analysis.

Further phylogenetic analysis of the indi-vidual tamarin clades failed to yield moreresolution. As can be seen in Figure 8A, thesingle most parsimonious reconstructionfrom an exact search of the sequence datafor the Small-bodied clade using S. m. mystaxas an outgroup has virtually the same struc-ture as in Figure 6. The only differences arethe relationships of S. tripartitus and S. f.lagonotus to the remainder of the clade.Although the 50% majority-rule bootstrapconsensus for the single-clade analysis (Fig.8B) reveals the same ambiguity in the basalportions of the clade as seen in the previousanalysis (Fig. 7), there is great support forthe grouping of S. f. fuscus with S. nigricol-lis. An exact search of the sequence data forthe Large-bodied clade (using S. f. melanoleu-cus and S. nigricollis as outgroups) yieldedthe same topology as before (Fig. 9A), and abootstrap analysis of the same data (Fig. 9B)again failed to resolve the deeper relation-ships within the clade.

The molecular tree (Fig. 6) was comparedto the hypothetical tree based on Ferrari(1993a) (Fig. 4A) using a Wilcoxon signed-ranks test (Templeton 1983a,b; Felsenstein,


TABLE 3. Pairwise comparisons of mtDNA sequences among the taxa used in the study1

Taxon 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27

1. Callimico sp. 0.149 0.148 0.157 0.151 0.163 0.151 0.152 0.157 0.170 0.139 0.154 0.130 0.146 0.157 0.152 0.144 0.114 0.170 0.162 0.178 0.177 0.184 0.184 0.174 0.205 0.1172. Cebuella sp. 0.174 0.109 0.182 0.188 0.174 0.178 0.182 0.193 0.177 0.190 0.164 0.183 0.190 0.179 0.168 0.151 0.194 0.194 0.190 0.183 0.211 0.196 0.190 0.245 0.2343. Leontopithecus sp. 0.167 0.177 0.168 0.174 0.174 0.170 0.172 0.168 0.184 0.150 0.174 0.190 0.178 0.170 0.161 0.199 0.197 0.204 0.179 0.215 0.192 0.182 0.206 0.1474. Callithrix sp. 0.185 0.189 0.172 0.179 0.187 0.189 0.166 0.201 0.150 0.182 0.204 0.200 0.164 0.167 0.215 0.215 0.218 0.182 0.219 0.198 0.204 0.225 0.2485. S. f. fuscicollis 0.052 0.046 0.070 0.066 0.065 0.053 0.090 0.057 0.086 0.183 0.167 0.151 0.134 0.188 0.184 0.206 0.169 0.200 0.204 0.191 0.185 0.2346. S. f. nigrifons 0.052 0.075 0.064 0.074 0.058 0.083 0.064 0.086 0.181 0.171 0.151 0.149 0.201 0.195 0.207 0.171 0.200 0.204 0.197 0.202 0.2217. S. f. illigeri 0.025 0.049 0.061 0.056 0.046 0.051 0.053 0.140 0.128 0.134 0.135 0.160 0.166 0.173 0.173 0.203 0.178 0.169 0.204 0.1648. S. f. leucogenys 0.072 0.075 0.057 0.063 0.038 0.078 0.175 0.153 0.133 0.128 0.199 0.189 0.212 0.160 0.179 0.208 0.193 0.202 0.2519. S. f. melanoleucus 0.050 0.060 0.091 0.056 0.098 0.179 0.167 0.153 0.138 0.198 0.192 0.206 0.166 0.193 0.201 0.185 0.187 0.242

10. S. f. weddelli 0.061 0.092 0.057 0.101 0.186 0.167 0.150 0.136 0.203 0.194 0.210 0.173 0.197 0.204 0.193 0.190 0.22211. S. f. fuscus 0.052 0.045 0.024 0.144 0.123 0.135 0.125 0.172 0.164 0.152 0.167 0.208 0.170 0.170 0.197 0.16812. S. f. lagonotus 0.035 0.094 0.178 0.158 0.155 0.160 0.203 0.194 0.206 0.177 0.195 0.198 0.186 0.208 0.23213. S. tripartitus 0.051 0.138 0.122 0.127 0.126 0.158 0.157 0.142 0.150 0.153 0.157 0.154 0.181 0.19414. S. nigricollis 0.181 0.161 0.136 0.121 0.204 0.195 0.230 0.175 0.195 0.221 0.206 0.181 0.24815. S. labiatus 0.088 0.089 0.102 0.138 0.148 0.155 0.147 0.165 0.152 0.153 0.173 0.14416. S. imperator 0.096 0.116 0.132 0.132 0.144 0.143 0.163 0.139 0.131 0.159 0.10517. S. mystax mystax 0.021 0.137 0.140 0.137 0.154 0.167 0.169 0.164 0.175 0.09118. S. mystax pluto 0.134 0.129 0.139 0.151 0.113 0.157 0.158 0.137 0.11219. S. geoffroyi 0.049 0.105 0.142 0.165 0.141 0.137 0.169 0.17620. S. oedipus 0.107 0.129 0.172 0.142 0.132 0.171 0.17421. S. leucopus 0.109 0.100 0.122 0.127 0.116 0.13522. S. midas midas 0.034 0.095 0.091 0.110 0.18623. S. midas niger 0.121 0.110 0.121 0.24324. S. bicolor bicolor 0.041 0.034 0.15025. S. bicolor martinsi 0.017 0.14326. S. bicolor ochraceus 0.21127. S. inustus

1 Distances were calculated using PAUP version 3.1.1 and have been adjusted for missing data. S. 5 Saguinus; f. 5 fuscicollis.

Fig. 6. Single most parsimonious tree resulting from a heuristic search (random additions option, tenreplications) of the sequence data using Callimico as the outgroup reference taxon. The numbers abovethe branches are the ACCTRAN/DELTRAN branch lengths. The total number of changes on the tree is1325. The consistency index is 0.543, and the homoplasy index is 0.457.


1985a). Before comparing the two trees, werearranged the polytomy in Figure 4A toproduce the most parsimonious bifurcatingclade consistent with the polytomy (Fig. 4B).

Ferrari (1993a) made no mention of therelationships among any of the subspeciesgroups, so only S. f. fuscicollis was retainedin both trees as a representative of the S.

Fig. 7. The 50% majority-rule bootstrap consensus of the mtDNA sequence data using Callimico as theoutgroup. Bootstrap values obtained from 200 replications (three random additions per replication) areshown above the branches.


fuscicollis subspecies. In addition, only S. b.bicolor and S. b. martinsi were used asrepresentatives of the S. bicolor subspeciesgroup to yield a bifurcating S. bicolor clade.Saguinus bicolor ochraceus was eliminatedfrom both trees in this particular analysisbecause the least amount of sequence wascollected for this subspecies. The hypotheti-cal phylogeny based on Ferrari (1993a) wassignificantly less parsimonious than the mo-lecular tree (N 5 87, T 5 560, P , 0.001,two-tailed test).

The molecular tree (Fig. 6) was comparedalso to Hershkovitz’s (1977) hypothetical

phylogeny (Fig. 3A) using the relationshipsamong the S. fuscicollis subspecies depictedin Figure 3B. Hershkovitz also made specificpredictions concerning the relationshipsamong the S. bicolor subspecies that wereused to create a bifurcating structure forthat clade in the hypothetical phylogeny.

Fig. 8. A: The single most parsimonious tree result-ing from an exact search of the mtDNA sequence datafor the Small-bodied clade using S. mystax mystax as anoutgroup (tree length 5 412 steps, C.I. 5 0.771, H.I. 50.229). B: The 50% majority-rule bootstrap consensus ofthe same data set. Bootstrap values obtained from 200replications (three random additions per replication) areshown above the branches.

Fig. 9. A: The strict consensus of the three equallymost parsimonious trees resulting from an exact searchof the mtDNA sequence data for the Large-bodied cladeof Saguinus using S. fuscicollis melanoleucus and S.nigricollis as outgroups (tree length 5 751, C.I. 5 0.679,H.I. 5 0.321). B: The 50% majority-rule bootstrapconsensus of the same data set. Bootstrap values ob-tained from 200 replications (three random additionsper replication) are shown above the branches.


Hershkovitz’s (1977) hypothetical phylog-eny was significantly less parsimonious thanthe molecular tree (N 5 83, T 5 975, P ,0.001, two-tailed test). The Small-bodiedclade of the molecular tree alone was com-pared to the hypothetical tree in Figure 3B.The phylogenetic relationships among S.nigricollis and the S. fuscicollis subspecieshypothesized by Hershkovitz (1977) weresignificantly less parsimonious than the to-pology of the Small-bodied clade of the mo-lecular tree (N 5 38, T 5 211, P 5 0.02,two-tailed test). Also, the Small-bodied clade(Fig. 8A) was rearranged to reflect a mono-phyletic S. fuscicollis clade by grouping S. f.fuscus with the other S. fuscicollis subspe-cies. This rearrangement was significantlyless parsimonious (N 5 15, T 5 24, P 5 0.04,two-tailed test).

Hershkovitz postulated that the northernbare-face tamarins (S. oedipus, S. leucopus,and S. geoffroyi) arose from ancestral S.inustus. A topology grouping S. inustus basalto the clade of the three northern bare-facetamarins was compared with the moleculartree. This rearrangement of the Large-bodied clade was not significantly less parsi-monious than the molecular tree (N 5 14,T 5 30, P . 0.1, two-tailed test).

DISCUSSION

Tamarin phylogeny

The molecular phylogenetic relationshipsamong some of the tamarin species havebeen reported previously (Jacobs et al., 1995),but the present analysis includes all 12species and nearly all of the subspecies. Thepresent analysis is congruent with the find-ings of Jacobs et al. (1995). Saguinus is amonophyletic taxon composed of two majorclades: the Small-bodied clade containing S.nigricollis, S. tripartitus, and S. fuscicollisand the Large-bodied clade, which containsthe remainder of the species. These resultsagree with Garber’s (1992) division of Sagui-nus into two major clades (S. fuscicollis andS. nigricollis in one clade and all others in asecond clade) based on feeding adaptationsrelated to body size. Whether Garber (1992)considers S. tripartitus as a subspecies of S.

fuscicollis and therefore a member of thesame clade is not explicitly stated.

Three more species are now included inthe Large-bodied clade than in the Jacobs etal. (1995) study: S. leucopus, S. bicolor, andS. inustus. Hershkovitz (1977) proposed thatS. leucopus was most closely related to S.oedipus and S. geoffroyi. His hypothesis issupported by the molecular data (Fig. 6) andis not surprising given the geographic prox-imity these three species and their geo-graphic isolation. Cladistic and morphomet-ric analyses of craniofacial features alsogroup these three species (Natori and Hani-hara, 1988; Moore and Cheverud, 1992),although Moore and Cheverud (1992) foundS. oedipus to be more similar morphologi-cally to S. leucopus than to S. geoffroyi.

Based on the sequence data, the S. bicolorsubspecies form a monophyletic group andare the sister taxon to S. midas. Hershkovitz(1977) hypothesized this same relationshipbased on coat color and the adjacent geo-graphic ranges of these taxa north of theAmazon (Fig. 1A). The sister-group relation-ship of S. bicolor and S. midas also agreeswith cladistic analyses of morphological fea-tures (Natori and Hanihara, 1988; Natori,1988) and with ecological evidence (Ferrari,1993a).

In the most parsimonious phylogeneticreconstruction (Fig. 6), S. inustus is a sistertaxon to a group containing S. mystax, S.labiatus, and S. imperator. Jacobs et al.(1995) found a monophyletic group com-posed of S. labiatus, S. imperator, and S.mystax, but a bootstrap analysis of the rela-tionships among this clade lacked sufficientresolution to make any firm conclusions.The grouping of these three species togetherwith S. inustus is not well supported bybootstrap analysis (Fig. 7). The failure ofadditional sequence data for S. labiatus, S.imperator, and S. mystax to resolve therelationships among these three species sug-gests that a rapid radiation may have oc-curred amongst these species that would notbe recoverable using parsimony analysis.The lack of resolution for the phylogeneticposition of S. inustus could be attributed tothe same phenomenon or could be due to therelatively small amount of sequence data


generated for S. inustus (Appendix) com-pared to the other species in the analysis.Hershkovitz (1977) proposed that S. inustuswas most closely related to the northernbare-face tamarins, S. oedipus, S. leucopus,and S. geoffroyi, and rearranging the phylo-genetic tree to reflect this relationship is notsignificantly less parsimonious than the mo-lecular tree in Figure 6. Ferrari (1993a)grouped S. inustus with S. bicolor and S.midas, although this decision was basedpartially on the lack of evidence linking S.inustus to any other tamarins. Cladisticanalysis of craniodental features lacked reso-lution in the relationship of S. inustus to theother tamarins (Natori, 1988). The lack ofconsensus among data sets indicates theneed for more information about this little-known species.

The phylogenetic relationships amongmembers of the Small-bodied clade, al-though not all completely well resolved, domake sense geographically (Fig. 1). Themolecular data indicate that S. f. fuscicollisand S. f. nigrifons are sister taxa, and theranges of these two subspecies are separatedonly by the Rio Javari (Hershkovitz, 1977).The ranges of S. f. melanoleucus and S. f.weddelli are geographically close as well,which agrees with the grouping of these twoas sister taxa. Cheverud and Moore (1990)also found that the facial morphology of S. f.melanoleucus and S. f. weddelli was quitesimilar. Together, all four subspecies form awell supported clade that ranges from thewestern bank of the Rio Madeira/Rio Ma-more to the eastern side of the Rio Ucayali.These results are supported by a clusteranalysis of morphological distances thatgrouped S. f. nigrifons and S. f. weddelli(Moore and Cheverud, 1992), although theother two species in the clade, S. f. melano-leucus and S. f. fuscicollis, were not includedin the morphological analysis. Had S. f.avilapiresi been included in the presentstudy, it seems likely that it would havebeen related to this clade, given that thegeographic range of this subspecies is withinthe bounds of the clade.

Although the support for S. f. illigeri andS. f. leucogenys as sister taxa is relatively

weak (Fig. 7), their geographic ranges doadjoin on opposite banks of the Rio Hual-laga. However, a systematic study of cranialmorphology failed to yield a close relation-ship between the two subspecies (Moore andCheverud, 1992). The lack of resolution inthe more basal portions of the S. fuscicollisgroup could be due to lack of sufficientvariability in the regions of mtDNA thatwere sequenced. However, the D-loop regionof the mtDNA is the most variable in mam-mals (Avise et al., 1987) and should yield thegreatest number of informative sites. Analternative explanation is that a rapid ini-tial radiation of the S. fuscicollis groupoccurred.

In the Jacobs et al. (1995) study, S. fusci-collis was a monophyletic species. However,the inclusion of S. f. fuscus and S. tripartitusin the present study shows S. fuscicollis tobe a paraphyletic species. As can be seen inFigure 6, S. f. fuscus groups with S. nigricol-lis rather than the other S. fuscicollis subspe-cies, a relationship that is well supported bythe bootstrap analysis (Fig. 7). This relation-ship is again revealed by the single mostparsimonious tree and the bootstrap consen-sus resulting from the single-clade analysisas shown in Figure 8. In addition, groupingS. f. fuscus with the other S. fuscicollissubspecies is significantly less parsimoniousthan the molecular tree shown in Figure 8A.Based on the degree of divergence in facialmorphology of S. f. fuscus from the other S.fuscicollis subspecies, Moore and Cheverud(1992) have suggested that this subspeciesmight deserve species status. The presentresults support that suggestion, and a changein the taxonomic status of S. f. fuscus to S.fuscus would maintain the monophyly of S.fuscicollis. Saguinus tripartitus also dis-rupts the monophyly of S. fuscicollis becausethe most parsimonious phylogenetic recon-struction links it with S. f. lagonotus (Fig. 6).However, this relationship is not well sup-ported by the bootstrap analysis (Fig. 7), noris it well supported in the single cladeanalysis (Fig. 8). Hershkovitz’s (1977) origi-nal taxonomy included S. tripartitus as asubspecies of S. fuscicollis. Subsequent stud-ies of the geographic range of this taxon


have shown that this group should havespecies status (Thorington, 1988; Albuja,1994), and this revised classification for S.tripartitus is generally accepted. Because ofthe ambiguity of the phylogenetic relation-ship of this species to the remainder of the S.fuscicollis clade, S. fuscicollis could still beconsidered tentatively monophyletic pro-vided the taxonomy of S. f. fuscus is revised.

Historical biogeography

Ferrari’s (1993a) dispersal scheme for tam-arins was derived from a phylogenetic hy-pothesis based on ecological factors relatedto body size. However, the phylogenetic rela-tionships upon which his dispersal schemeis based are significantly less parsimoniousthan the phylogeny inferred from the mtDNAsequence data. Phyletic dwarfism was origi-nally proposed to account for several appar-ently derived morphological features sharedby extant callitrichids. These features in-clude small body size, loss of the third molar,reproductive twinning, and claws instead ofnails (Ford, 1980; Leutenegger, 1980). Thetheory did not espouse body size as a phylo-genetic character. In the absence of goodfossil evidence, dwarfing was hypothesizedas a single rapid event that occurred in theancestral callitrichid lineage. In fact, a testof the phyletic dwarfism hypothesis basedon dental metric data failed to support cer-tain morphological predictions of the hypoth-esis (Plavcan and Gomez, 1993a,b). There-fore, the dispersal routes for Saguinusderived from Ferrari’s (1993a) phylogenetichypothesis are unlikely.

The dispersal scheme of Hershkovitz(1977) for Saguinus also seems improbable,since his phylogenetic hypothesis based onthe theory of metachromism also was signifi-cantly less parsimonious than the molecularphylogeny. Jacobs et al. (1995) obtainedresults similar to the ones presented herewith a partial tamarin phylogeny. Hershko-vitz (1968, 1977) proposed the theory ofmetachromism to be a phylogenetic phenom-enon. However, others have cautionedagainst using coat color as a phylogeneticcharacter due to the likelihood of parallel-isms occurring among lineages (Shedd and

Macedonia, 1991; Jacobs et al., 1995). Ja-cobs et al. (1995) demonstrated coat color tobe a poor character for phylogenetic analy-sis.

Given the phylogenetic relationshipsamong tamarins inferred from the mtDNAsequence data, a plausible version of thehistorical dispersal routes for tamarins isshown in Figure 10. The most basal divisionamong the ancestral tamarins is a bifurca-tion between the Small-bodied clade and theLarge-bodied clade. If we take the region ofhighest diversity as the most likely area oforigin for the tamarins, the ancestral Sagui-nus would have arisen somewhere south ofthe Amazon and west of the Rio Madeira.This scenario is in agreement with Hershko-vitz (1977). The split between the ancestorsof the two clades could have occurred by avicariance event or by migration. A subse-quent dispersal of S. fuscicollis in a south-easterly direction would have brought themback into sympatry with the southern spe-cies of the Large-bodied clade. Two waves ofdispersal north across the Amazon wouldaccount for the distributions of S. inustus tothe west and S. bicolor and S. midas to theeast. A monophyletic grouping of S. bicolor,S. midas, and the northern bareface tama-rins indicates that the ancestors of the lattergroup arrived in northern Colombia andPanama from the northern regions of SouthAmerica rather than the southern originsproposed by Hershkovitz (1977). This dis-persal hypothesis reflects the most parsimo-nious phylogenetic relationships amongSaguinus, as shown in Figure 6.

The phylogenetic relationships among S.fuscicollis suggest that this group radiatedin a southeasterly direction instead of thesouthwestern direction suggested by Hersh-kovitz (1977) (Fig. 10B). Peres et al. (1996)have shown that hybridization (and hencedispersal) occurs at the headwaters of theRio Jurua, which separates the geographicranges of S. f. fuscicollis and S. f. melanoleu-cus. Therefore, the dispersal route for S.fuscicollis depicted in Figure 10B is conceiv-able, given that such a dispersal could have


Fig. 10. Proposed historical dispersal for Saguinus A: and the S. fuscicollis B: subspecies based on themolecular phylogenetic evidence.


occurred around the headwaters of the vari-ous rivers separating the subspecies.

Dating the dispersal events that led to thegeographic distribution of extant tamarinsis difficult. The paucity of primate fossilremains from the Neotropics allows for onlythe roughest estimate for the divergencetime of Saguinus from the other callitrich-ids. The oldest known South American pri-mate fossil, Branisella, was recovered fromlate Oligocene deposits in Bolivia (Hoffstet-ter, 1969; Rosenberger, 1981b; Wolff, 1984).These remains place early primates in SouthAmerica approximately 25 million years ago(Ma). More recent specimens that appear tobe more closely related to extant callitrichi-nes than Branisella have been unearthedfrom the La Victoria Formation in Colombia.Remains of a giant tamarin, Lagonimicoconclucatus, have been dated to the middleMiocene, approximately 13.5 Ma. The origi-nal fossil, described by Kay (1994), appearsto be a sister taxon to living callitrichids(callitrichines) rather than a direct ancestor.From the fossil evidence, therefore, we canestimate the age of the genus Saguinus aspossibly 25 Ma but more likely closer to 13.5Ma. Either way, divergence of Saguinusfrom the other callitrichids is clearly pre-Pleistocene.

Of course, the best way to test the dis-persal hypothesis set forth in this study is todemonstrate that the pattern also holds truefor other organisms similar to Saguinus insite of origin and physiogeographic featuresperceived as barriers to dispersal. Saimiri(squirrel monkeys) is another primate genuswith a geographic distribution similar tothat of tamarins. Squirrel monkeys are foundin secondary and riverine forest throughoutthe Amazon basin, with a small disjunctpopulation of Saimiri oerstedi located inCosta Rica (Hershkovitz, 1984). These pri-mates are larger in body size than tamarins,but their geographic ranges, too, are limitedby riverine barriers, although to a lesserextent than tamarins (Ayres and Clutton-Brock, 1992; S. Boinski, personal communi-cation). Preliminary data collected for theCosta Rican Saimiri species and two speciesof South American squirrel monkeys indi-

cate a similar geographic pattern of phyloge-netic relationships as Saguinus (Lepp et al.,1997). However, more extensive geographicsampling from Saimiri is needed before anydefinitive conclusions can be drawn.

CONCLUSIONS

Within Saguinus, there are two majorclades: the Small-bodied clade composed ofS. fuscicollis, S. tripartitus, and S. nigricol-lis and the Large-bodied clade that containsthe remainder of the species. A previousreview of the taxonomy of S. fuscicollissubspecies elevated S. f. tripartitus to spe-cies status (Thorington, 1988), yet under thecurrent taxonomy S. fuscicollis is still aparaphyletic taxon. Revision of the taxo-nomic status of S. f. fuscus from subspeciesto species, S. fuscus, would maintain mono-phyly of S. fuscicollis. Such a revision issupported by morphological evidence (Mooreand Cheverud, 1992) as well as the molecu-lar evidence.

The molecular evidence suggests a differ-ent historical dispersal for tamarins thanpreviously hypothesized (Hershkovitz, 1977;Ferrari, 1993a). The phylogenetic relation-ships among Saguinus suggest that tama-rins dispersed in two major waves from anorigin south of the Amazon and west of theRio Madeira: one in a western directionleading to the distributions of the small-bodied tamarin clade and one in a northeast-ern and then western direction resulting inthe distributions of S. bicolor, S. midas, andthe northern bareface tamarins.

ACKNOWLEDGMENTS

We thank the following individuals andinstitutions for their generosity in providingsamples for the molecular analysis: RichardThorington, Horacio Schneider, James Pat-ton, Dave Evans, the National Museum ofNatural History, the Field Museum of Natu-ral History, and the American Museum ofNatural History. Nathan Lepp provided tech-nical assistance. Jonathan Losos, CharlesHildebolt, Robert Sussman, and P. Mick


APPENDIX: Alignment of the sequence data used in the parsimony analysis 82S

.J.C

RO

PP

ET

AL

.

(continued)

83H

IST

OR

ICA

LB

IOG

EO

GR

AP

HY

OF

TA

MA

RIN

S

Appendix (continued) 84S

.J.C

RO

PP

ET

AL

.

(continued)

85H

IST

OR

ICA

LB

IOG

EO

GR

AP

HY

OF

TA

MA

RIN

S

Appendix (continued) 86S

.J.C

RO

PP

ET

AL

.

87H

IST

OR

ICA

LB

IOG

EO

GR

AP

HY

OF

TA

MA

RIN

S

Richardson provided comments on the manu-script. This work was supported by a disser-tation improvement grant (NSF 9411169)for S.J.C.

LITERATURE CITED

Albuja L. 1994. Nuevos registros de Saguinus triparti-tus en la Amazonia equatoriana. Neotrop Primates2:8–10.

Avise JC, Arnold J, Ball RM, Bermingham E, Lamb T,Neigel JE, Reeb CA, Saunders NC. 1987. Intraspecificphylogeography: the mitochondrial DNA bridge be-tween population genetics and systematics. Annu RevEcol Syst 18:489–522.

Ayres JM, Clutton-Brock TH. 1992. River boundariesand species range size in Amazonian primates. AmNat 140:531–537.

Cheverud JM, Moore AJ. 1990. Subspecific morphologi-cal variation in the saddle-back tamarin (Saguinusfuscicollis). Am J Primatol 21:1–15.

Felsenstein J. 1985a. Confidence limits on phylogenieswith a molecular clock. Syst Zool 34:152–161.

Felsenstein J. 1985b. Confidence limits on phylogenies:an approach using the bootstrap. Evolution 39:783–791.

Ferrari SF. 1993a. The adaptive radiation of Amazoniancallitrichids (Primates, Platyrrhini). Evolucion Bio-logica 7:81–103.

Ferrari SF. 1993b. Ecological differentiation in theCallitrichidae. In: Rylands AB, editor. Marmosetsand tamarins: systematics, ecology, and behaviour.New York: Oxford University Press. p 314–328.

Fleagle JG, Mittermeier RA. 1980. Locomotor behavior,body size, and comparative ecology of seven Surinammonkeys. Am J Phys Anthropol 52:301–314.

Ford SM. 1980. Callitrichids as phyletic dwarfs, and theplace of the Callitrichidae in the Platyrrhini. Pri-mates 21:31–43.

Ford SM. 1986. Systematics of the New World monkeys.In: Swindler DR, Erwin J, editors. Comparative pri-mate biology, vol. 1: systematics, evolution, andanatomy. New York: Alan R. Liss. p 73–135.

Garber PA. 1992. Vertical clinging, small body size, andthe evolution of feeding adaptations in the Callitrichi-nae. Am J Phys Anthropol 88:469–482.

Harada ML, Schneider H, Schneider MPC, Sampaio I,Czelusniak J, Goodman M. 1995. DNA evidence on thephylogenetic systematics of New World monkeys:support for the sister-grouping of Cebus and Saimirifrom two unlinked nuclear genes. Mol PhylogenetEvol 4:331–349.

Hershkovitz P. 1968. Metachromism or the principle ofevolutionary change in mammalian tegumentary col-ors. Evolution 22:556–575.

Hershkovitz P. 1977. Living New World monkeys(Platyrrhini), vol. 1. Chicago: University of ChicagoPress.

Hershkovitz P. 1984. Taxonomy of squirrel monkeysgenus Saimiri (Cebidae, Platyrrhini): a preliminaryreport with description of a hitherto unnamed form.Am J Primatol 7:155–210.

Hillis DM, Mable BK, Larson A, Davis SK, Zimmer EA.1996. Nucleic acids IV: Sequencing and cloning.In Hillis DM, Moritz C, editors. Molecular systemat-ics, 2nd ed. Sunderland: Sinauer. p 321–381.

Hoffstetter R. 1969. Un primate de l’Oligocene inferieursud-americain Branisella boliviana gen. et sp. nov. CR Acad Sci, Paris, D 269:434–437.

Jacobs SC, Larson A, Cheverud JM. 1995. Phylogeneticrelationships and orthogenetic evolution of coat coloramong tamarins (genus Saguinus). Syst Biol 44:519–532.

Kay RF. 1994. ‘‘Giant’’ tamarin from the Miocene ofColombia. Am J Phys Anthropol 95:333–353.

Lepp NT, Jacobs SC, Larson A. 1997. Phylogeneticrelationships and geographic dispersal in two generaof New World primates. Am J Phys Anthropol Suppl24:154 (abstract).

Leutenegger W. 1980. Monogamy in callitrichids: aconsequence of phyletic dwarfism. Int J Primatol1:95–98.

Mittermeier RA, Coimbra-Filho AF. 1981. Systematics:species and subspecies. In: Coimbra-Filho AF, Mitter-meier RA, editors. Ecology and behavior of Neotropi-cal primates, vol. 1. Rio de Janeiro: Academia Bra-sileira de Ciencias. p 29–109.

Moore AJ, Cheverud JM. 1992. Systematics of theSaguinus oedipus group of the bare-face tamarins:evidence from facial morphology. Am J Phys Anthro-pol 89:73–84.

Natori M. 1988. A cladistic analysis of interspecificrelationships of Saguinus. Primates 29:263–276.

Natori M, Hanihara T. 1988. An analysis of interspecificrelationships of Saguinus based on cranial measure-ments. Primates 29:255–262.

Paabo S, Gifford JA, Wilson AC. 1988. MitochondrialDNA sequences from a 7000-year old brain. NucleicAcids Res 16:9775–9787.

Peres CA, Patton JL, da Silva MNF. 1996. Riverinebarriers and gene flow in Amazonian saddle-backtamarins. Folia Primatol (Basel) 67:113–124.

Plavcan JM, Gomez AM. 1993a. Dental scaling in theCallitrichinae. Int J Primatol 14:177–192.

Plavcan JM, Gomez AM. 1993b. Relative tooth size anddwarfing in the callitrichines. J Hum Evol 25:241–245.

Rosenberger AL. 1981a. Systematics: the highertaxa. In: Coimbra-Filho AF, Mittermeier RA, editors.Ecology and behavior of neotropical primates, vol. 1.Rio de Janeiro: Academia Brasileira de Ciencias, p9–27.

Rosenberger AL. 1981b. A mandible of Branisella bolivi-ana (Platyrrhini, Primates) from the Oligocene ofSouth America. Int J Primatol 2:1–7.

Rylands AB. 1993. The bare-face tamarins, Saguinusoedipus oedipus and Saguinus oedipus geoffroyi: sub-species or species? Neotrop Primates 1:4–5.

Rylands AB, Coimbra-Filho AF, Mittermeier RA. 1993.Systematics, geographic distribution, and some noteson the conservation status of the Callitrichidae. In:Rylands AB, editor. Marmoset and tamarins: system-atics, behavior, and ecology. New York: Oxford Univer-sity Press. p 11–77.

Schneider H, Schneider MPC, Sampaio I, Harada ML,Stanhope M, Czelusniak J, Goodman M. 1993. Molecu-lar phylogeny of the New World monkeys (Platyrrhini,Primates). Mol Phylogenet Evol 2:225–242.


Shedd DH, Macedonia JM. 1991. Metachromism andits phylogenetic implications for the genus Eulemur(Prosimii: Lemuridae). Folia Primatol (Basel) 57:221–231.

Swofford DL. 1993. PAUP: Phylogenetic analysis usingparsimony, version 3.1.1. Champaign: Illinois NaturalHistory Survey.

Templeton AR. 1983a. Convergent evolution and nonpar-ametric inferences from restriction data and DNAsequences. In: Weir BS editor. Statistical analysis of

DNA sequence data. New York: Marcel Dekker. p151–179.

Templeton AR. 1983b. Phylogenetic inference from re-striction endonuclease cleavage site maps with particu-lar reference to the evolution of humans and the apes.Evolution 37:221–244.

Thorington RW. 1988. Taxonomic status of Saguinustripartitus. Am J Primatol 15:367–371.

Wolff RG. 1984. New specimens of the primate Bra-nisella boliviana from the early Oligocene of Salla,Bolivia. J Vert Paleontol 4:570–574.


Documents

Historical biogeography of tamarins, genusSaguinus: the molecular phylogenetic evidence