AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 124:285–296 (2004) Atelinae Phylogenetic Relationships: The Trichotomy Revived? A.C. Collins* Department of Anthropology, University of Wisconsin, Madison, Wisconsin 53706, and Department of Conservation Biology, Daniel F. and Ada L. Rice Conservation Biology and Research Center, Brookﬁeld Zoological Institution, Brookﬁeld, Illinois 60513 KEY WORDS taxonomy; cladistics; systematics; molecular; Atelini; Cebidae; Platyrhinne ABSTRACT This research examines phylogenetic relationships between members of the Atelinae subfamily (Alouatta, Ateles, Brachyteles, and Lagothrix), based on analysis of three genetic regions. Two loci, cytochrome c oxidase subunit II (COII) and the hypervariable I portion of the control region, are part of the mitochondrial genome. The other is a single-copy nuclear gene, Aldolase A Intron V. Analysis of these genetic regions provides support for tribe Alouattini containing the Alouatta species, while tribe Atelini contains the other three genera. However, these three genetic regions produce conﬂicting results for relationships among tribe Atelini members. Previous genetic studies supported grouping Brachyteles with Lagothrix, leaving Ateles in a separate subclade. The present data sets vary based on the genetic region analyzed and method of analysis suggesting all possible cla- distic relationships. These results are more consistent with investigations of morphology and behavior among these primates. The primary cause of discrepancy between this study and previous genetic studies is postulated to reside in increased sampling in the present study of genetic variation among members of the Atelinae, speciﬁcally Ateles. The present study utilized samples of Ateles from all postulated species for this genetically variable primate, while previous studies used only one or two species of Ateles. This paper demonstrates that shifting relationships are produced when different species of Ateles are used to reconstruct phylogenies. This research concludes that a trichotomy should still be supported between members of tribe Atelini until further analyses, which include additional Atelinae haplotypes are conducted. Am J Phys Anthropol 124:285–296, 2004. © 2004 Wiley-Liss, Inc. Most authors agree that the genera Alouatta (howler monkeys), Ateles (spider monkeys), Brachyteles (woolly spider monkeys, or muriqui), and Lagothrix (woolly monkeys) are monophyletic descendants of a single common ancestor, constituting the subfamily Atelinae (Dunlap et al., 1985; Ford, 1986; Rosenberger and Strier, 1989; Kay, 1990; Paterson, 1991; Strier, 1992; Ford and Davis, 1992; Schneider et al., 1993; Harada et al., 1995; Schneider and Rosenberger, 1996; Canavez et al., 1999; von Dornum and Ruvolo, 1999; Meireles et al., 1995, 1999). Early taxonomists grouped the members of the Atelinae together primarily on the basis of their unique prehensile tail (Gray, 1870). Comparison of these genera shows that they all share a number of morphological traits, including: 1) large body size (Brachyteles is the largest Platyrrhine), 2) a prehensile tail with ventral grasping surfaces, and 3) frequent suspensory behaviors (Rosenberger and Strier, 1989; Ford and Davis, 1992). In most other aspects they display a wide array of morphological and behavioral traits (Rosenberger and Strier, 1989; Strier, 1992). Members of the Atelinae are the most geographically widespread subfamily among all New World primates, ranging from southern Mexico to the Uruguayan-Brazilian border (Rosenberger and Strier, 1989; Rowe, 1996). See Figure 1. Morphological studies arranged the three genera of tribe Atelini (Ateles, Brachyteles, and Lagothrix) in every possible taxonomic relationship (Dunlap et al., 1985; Ford, 1986; Rosenberger and Strier, 1989; Kay, 1990; Ford and Davis, 1992; Strier, 1992), suggesting a trichotomous relationship between the three genera. Cytogenetic research of tribe Atelini does not completely resolve this issue. Chromosomal evolution (Dutrillaux, 1979) suggests Ateles and Lagothrix are sister taxa; however, comparison of chromosome number and implied amount of rearrangement would support Brachyteles and Lagothrix as more closely related genera (Koiffman and © 2004 WILEY-LISS, INC. Grant sponsor: Chicago Zoological Society; Grant sponsor: Institute of Museum and Library Services; Grant number: 1C-40190-94; Grant sponsor: Sigma Xi. *Correspondence to: A.C. Collins, Department of Anthropology, 125A Sabin Hall, 3413 N. Downer Ave., University of Wisconsin, Milwaukee, WI 53211. E-mail: email@example.com Received 10 September 2001; accepted 16 June 2003. DOI 10.1002/ajpa.10358 Published online 19 November 2003 in Wiley InterScience (www. interscience.wiley.com). 286 A.C. COLLINS Fig. 1. Geographic distribution of tribe Atelini members in southern and eastern portions of the Amazon Basin and Atlantic coastal forests of South America. Both species of Lagothrix (shaded line) and disjunct populations of Brachyteles are indicated. Four species of Ateles (patterned line) are included. Approximate geographic locations of samples utilized in this study are indicated with numerals that correspond to those found in Table 1. Species identities are based on taxonomy of Groves (1989), Rowe (1996), Pope (1998), and Collins and Dubach (2000, 2001). Saldanna, 1974; Pequignot et al., 1985; Rumpler and Strasbourg, 1992). A steadily growing number of genetic studies support a phylogeny in which Brachyteles and Lagothrix are sister taxa, with Ateles as the outgroup (Schneider et al., 1993; Harada et al., 1995; Schneider and Rosenberger, 1996; Barroso et al., 1997; Porter et al., 1997; Canavez et al., 1999; von Dornum and Ruvolo, 1999; Meireles et al., 1999; Schneider, 2000). These genetic studies are based on sequence information from a variety of nuclear genes as part of large-scale studies to determine higher-level relationships among all Platyrrhine primates. Figure 2 summarizes the relationships supported, the morphological characteristics or genetic regions investigated, and the taxonomy suggested by previous studies. The present study investigated phylogenetic relationships among the Atelinae, based on DNA sequence variation for a mitochondrial, protein-coding region, cytochrome c oxidase subunit II (COII), as well as the hypervariable I area of the mitochondrial control region. In addition, a single-copy nuclear gene, Aldolase A intron V, was sequenced. The higher mutation rate and associated increase in genetic substitutions in these loci, compared to more conserved nuclear genes utilized in previous Platyrrhine-wide studies, were hypothesized to produce a more discriminate measure of variation among the haplotypes. Atelinae relationships were investigated using only one or two haplotypes from each genus in the previous genetic studies, since their primary focus was Platyrrhine-wide relationships (Schneider et al., 1993; Harada et al., 1995; Schneider and Rosenberger, 1996; Barroso et al., 1997; Porter et al., 1997; Canavez et al., 1999; von Dornum and Ruvolo, 1999; Meireles et al., 1999; Schneider, 2000). However, Alouatta has 6 or 7 species (Groves, 1989), Ateles has 3– 6 species (Kellogg and Goldman, 1944; Groves, 1989; Froehlich et al., 1991; Collins and Dubach, 2000), Lagothrix has 2 species (Groves, 1989), and Brachyteles is comprised of 1 or possibly 2 species (Groves, 1989; Pope, 1998). The results of the genetic studies contrast with the trichotomous relationship supported by sampling multiple members of each genus for morphological traits. I propose that signiﬁcant intrageneric variation among Atelinae genera, especially in Ateles (Collins and Dubach, 2000), prohibits the use of the limited sampling approach for phylogenetic studies of this subfamily. Ruvolo (1997) suggested that the very close, previously trichotomous, genetic relationships among the Hominoidea can be adequately discriminated using single-haplotype representatives from each genus. But genetic distances between genera are closer for the Hominoidea, their last common ancestor lived more recently, and their geographic distributions are more limited than members of the Atelinae. All of these factors are ATELINAE PHYLOGENETIC TRICHOTOMY 287 Alouatta spp., Alouatta seniculus, Alouatta caraya, and Lagothrix spp. were obtained from GenBank (Adkins and Honeycutt, 1994; Figueiredo et al., 1998; Ascunce et al., 2001). Table 1 and Figure 1 detail the geographic locations and sources of samples from the various Atelinae genera and species used in this study. DNA isolation and nucleotide sequencing Fig. 2. Various different cladistic arrangements among tribe Atelini genera in a wide variety of previous and present morphological and genetic studies. Source of analysis is indicated in parentheses after each author, in boxes below various clades. hypothesized to reduce genetic variation in the Hominoidea as compared to the Atelinae. To increase inclusion of genetic variation sampled in the Atelinae, the present study included haplotypes of Ateles from each of the six species suggested by Groves (1989) in the most expansive taxonomy for this genus. Alouatta, consistently the basal taxon in all previous genetic studies and most morphological studies, was represented in the present research by one species, except for COII, where additional sequence information from GenBank allowed expanded analysis. Both Brachyteles and Lagothrix are represented by samples from the single most widely distributed species. This study omitted only the minimally distributed Lagothrix ﬂavicauda among traditionally recognized tribe Atelini species. This research is an extension of phylogenetic studies of Ateles phylogenetic relationships conducted by Collins and Dubach (2000, 2001). MATERIALS AND METHODS Materials Samples of DNA and/or tissue were obtained directly from wild specimens for most haplotypes. Additional whole-blood samples were obtained from various zoological and research institutions throughout North America. Only specimens with wild origins from known geographic locations, conﬁrmed by karyotype analysis, were utilized (Table 1). Subspecies names were used as labels to identify haplotypes, but all haplotypes were analyzed and grouped irrespective of any previous taxonomic identities. Additional DNA sequences for COII from Total genomic DNA was isolated from blood and tissue samples, using standard phenol/chloroform extraction techniques with ethanol precipitation (Sambrook et al., 1989). DNA was ampliﬁed using the polymerase chain reaction (PCR). Conserved primers located in the lysine and asparagine tRNA were used to amplify the COII gene. The control region was ampliﬁed using conserved primers located in the proline tRNA and a conserved location near the origin of replication (Kocher et al., 1989). Primers reported by Hillis et al. (1996) were used for ampliﬁcation of the Aldolase A intron. All double-stranded DNA ampliﬁcation products were directly sequenced on both model 377 and 377A automated DNA sequencers (ABI). All samples were sequenced in both directions, and random samples were repeated to ensure sequence accuracy. For speciﬁc details of PCR conditions and sequencing of DNA, refer to Collins and Dubach (2000, 2001). DNA sequences obtained for all three genetic regions used in this study were deposited in GenBank under the following sets of accession numbers: control region, AF213940 –AF213966; COII, AF216225AF216253; and Aldolase A, AF242395–AF242410. See Table 1 for exact sequences utilized. DNA sequence alignment and phylogenetic analysis Atelinae sequences were aligned using the MacDNASIS version 3.2 (Hitachi Software Engineering, America Ltd.) contig option to align sequences from opposite orientations, and then were compiled by inspection. Gaps were introduced in order to align haplotypes, as dictated by a Higgins-Sharp multiple alignment program. Sequences were analyzed using Phylogeny Analysis Using Parsimony (PAUP), version 4.0b1 (Swofford, 1998). Trees were produced using both unweighted and weighted ratios of transversions to transitions. Exhaustive searches were performed to identify the best-ﬁt tree in most cases. Bootstrap analysis was carried out using heuristic searches, with random-sequence addition to assess support for various branches in the phylogram, based on 5,000 replications. For analysis of various genetic regions, midpoint rooting was initially performed to assess that Alouatta was the most basal member among Atelinae haplotypes. Alouatta was then utilized as an outgroup for further analyses of relationships among members of tribe Atelini and bootstrap conﬁdence assessment. 288 A.C. COLLINS TABLE 1. Geographic location and specific identification of Atelinae haplotypes1 Subspecies population Ateles paniscus Ateles belzebuth marginatus Ateles belzebuth chamek Ateles belzebuth chamek Ateles hybridus Ateles geoffroyi (fusciceps) robustus Ateles geoffroyi yucatanensis Alouatta palliata Alouatta spp. Alouatta seniculus Alouatta caraya Brachyteles arachnoides hypoxanthus Lagothrix lagotricha Lagothrix spp. GenBank ID AF213961 AF216248 AF216247 AF213941 AF216226 AF242400 AF213945 AF216240 AF242403 AF213944 AF216239 AF242401 AF213948 AF216227 AF242398 AF213959 AF216245 AF242397 AF213951 AF216230 AF242396 AF213964 AF216252 AF242409 L22774 AF054300 AF396459 AF213966 AF216253 AF242410 AF213965 AF216251 AF242408 L22779 Map location Geographic area 1 Rio Trombetus, Para State, Brazil 2 Tocantins River, Para State, Brazil 3 Rio Jiparana, Rondonia State, Brazil 4 Peru 5 Magdalena River Valley, Colombia 6 Darien Peninsula, Panama 7 Yucatan Peninsula, Mexico Central America From Adkins and Honeycutt (1994) From Figueiredo et al. (1998) From Ascunce et al. (2001) Fazenda Esmerelda, Sao Paulo State, Brazil Unknown From Adkins and Honeycutt (1994) 1 Provides information on species/subspecies identity of various haplotypes utilized in this study. GenBank accession numbers are provided for each haplotype. Accession numbers are for COII, control region, and Aldolase A regions from top to bottom. Identical haplotypes were not sequenced for some animals for Aldolase A. Geographic information indicates broad area each sample came from in the wild, or references published source of sequence information. Map locations correlate to numbers in Figure 1 on source of Ateles haplotypes. Maximum likelihood analysis was performed using the likelihood option of PAUP, version 4.0b1 (Swofford, 1998), with heuristic searches and addition of samples as presented in the data set. The transition transversion ratio was estimated from the data. Data sets were randomly rearranged manually and reanalyzed, to insure that order of addition was not affecting the resultant phylogeny. Sequences were also analyzed using the neighborjoining optimization in PAUP, version 4.0b1 (Swofford, 1998), to produce a distance-based phylogenetic tree. Bootstrap analysis was performed with random sequence addition on 10,000 replications of the data set. The computer phylogeny program MEGA, version 1.02 (Kumar et al., 1993), was also used to compute the standard error test of neighbor-joining tree branch lengths from zero H0:BL ⫽ 0. The conﬁdence probabilities (CP) reported are equivalent to 1 ⫺ ␣, where ␣ is the signiﬁcance level of type I error. Thus, CP ⱖ 0.95 corresponds to ␣ ⱕ 0.05 and is considered signiﬁcant. When bootstrapping favored an arrangement seen in the neighbor-joining tree and the conﬁdence probability favored another relationship, no conﬁdence probability is reported for that node. Congruence between various data sets was examined with the incongruence length difference (ILD) or partition-homogeneity test (Farris et al., 1995), as implemented in PAUP (Swofford, 1998). Congruence between all tribe Atelini data sets (past and present) was also implemented in the same manner, but with only one species of each genus, as occurs in many of the data sets. The test was implemented on 1,000 partition sample replicates with a heuristic search and random addition of haplotypes. Topological congruence between possible gene trees was evaluated using the two-tailed test of Templeton (1983) and the winning sites test (Prager and Wilson, 1988), as implemented in PAUP. The computer program MacClade, version 3.05 (Maddison and Maddison, 1992), was used to analyze base composition frequency, set character weights, determine nucleotide coding positions for protein coding regions, calculate transition/transversion ratios, and trace character-state evolution. RESULTS The Ateles haplotypes used in this study consistently produced the taxonomy reported by Collins and Dubach (2000), and were collapsed into those ATELINAE PHYLOGENETIC TRICHOTOMY four species: A. paniscus, A. belzebuth (including A. b. chamek and A. b. marginatus), A. hybridus (formerly A. b. hybridus), and A. geoffroyi (including A. g. yucatanensis and A. f. robustus) for graphic presentation. COII gene The analysis of the COII gene included seven Ateles haplotypes, two haplotypes of Lagothrix, four Alouatta haplotypes from three different species, and one haplotype of Brachyteles (See Table 1). The data set has 143 phylogenetically informative characters, with a ratio of 5.1 transitions per transversion (Maddison and Maddison, 1992). Sequences from Alouatta and Lagothrix obtained from GenBank are shorter than those in this study and include fewer characters, which were treated as missing data. A branch-and-bound analysis (PAUP) of unweighted data produced six equally parsimonious trees, where four trees supported a Brachyteles-Ateles clade and the other two trees actually supported an Ateles-Lagothrix clade. Either clade was supported by an identical number of synapomorphies. The Brachyteles-Lagothrix clade had six fewer synapomorphies than the alternative clades. Bootstrap support for the Brachyteles-Ateles clade was 51% in the unweighted data set, using all available DNA sequences. Weighted analysis with the observed 5:1 ratio and experimental ratios as high as 20:1 supported the Ateles-Lagothrix clade. The 20:1 weighting produced two trees that varied only among the Ateles haplotypes, with better bootstrap support for the Ateles-Lagothrix clade of 62% (Fig. 3). Neighbor-joining analysis favored the Brachyteles-Ateles clade with bootstrap support of 70% (Fig. 3). A maximum likelihood analysis of this data set also favored a clade containing Brachyteles and Ateles, with an ln likelihood of ⫺2,554.6. Assigning coding positions to the weighted data set and analyzing each position separately produced support for a Brachyteles-Ateles clade for position 1, using both parsimony and neighbor-joining analysis. Position 2 does not favor relationships between any members of tribe Atelini. The third position reﬂects the complete data set relationships varying depending on the analysis method. Analysis of only the transversions in the Atelinae COII data set is unable to differentiate among the members of tribe Atelini. Control region The portion of the control region sequenced ranged from 497–522 base pairs among Ateles haplotypes. This variance was due to a 25-base pair deletion in A. paniscus, A. b. marginatus, and some A. b. chamek haplotypes. The complete aligned data set was 541 base pairs long, with appropriate insertions made. The control region sequence contained 124 phylogenetically informative characters and a 289 transition to transversion ratio of 2.56 (MacClade, version 3.05). Parsimony analysis of the control region sequences produced six equally parsimonious trees. All six trees related Brachyteles more closely to Lagothrix, contradictory to the COII gene tree, with weak bootstrap support of 56% (Fig. 3). Weighting the data set as high as 20:1 produced only one tree identical to one of the six unweighted trees, but with no bootstrap support for the Brachyteles-Lagothrix clade. Due to the contradictory topology of the two-phylograms, an analysis which eliminated the more frequent and possibly homoplasic transitions was undertaken. Analysis of transversions only produced 21 trees and resulted in a Brachyteles-Lagothrix clade in 11 of the 21 trees without any bootstrap support. Neighbor-joining analysis produced a tree with a very short branch length (less than 0.5% of the total length) leading to a Brachyteles-Lagothrix clade. Bootstrap support was similar to the parsimony tree (Fig. 3). To try and resolve possible variable rates of change among sites in the control region, a maximum likelihood search was preformed. The resultant topology related Brachyteles to Lagothrix, with an ln likelihood of ⫺2,448.4. Since the mitochondrial genome is inherited as a single unit, the relationships should reﬂect the same heritage. The fact that they do not is likely due to increased homoplasy in the more variable control region sequences and the possibility of nonequal rates of variation among sites. However, a plot of transitions vs. genetic distance for all pairwise comparisons in this data set had a high correlation coefﬁcient (R2 ⫽ 0.941) Aldolase A Intron V Aldolase A Intron V is a variable nuclear region that had only ﬁve parsimony informative characters in the 466 bp sequenced. An exhaustive search (PAUP) using midpoint rooting united all members of tribe Atelini with Alouatta as the most divergent genus. Parsimony analysis produced only one tree which grouped Brachyteles and Ateles, leaving Lagothrix as the sister taxa to this clade, with bootstrap support of 63% (Fig. 4). No informative transversions occurred in the data set, so no weighted analysis was carried out. Distance-based analyses supported grouping Brachyteles with Ateles in a clade joined basally by Lagothrix. Bootstrap analysis supported this relationship 68% of the time (Fig. 4). Maximum likelihood analysis of the Aldolase data set produced a tree that also related Brachyteles more closely to Ateles, with an ln likelihood value of ⫺821.7. Complete mitochondrial and nuclear DNA analysis To search for clarity of relationships among tribe Atelini members, all nuclear and mitochondrial data from the present study were subjected to the parti- 290 A.C. COLLINS Fig. 3. Trees A and C. Tree A is maximum parsimony (PAUP) COII tree from weighted data set, with tree length of 493 steps. Discovered using a branch-and-bound search with furthest addition sequence; consistency index ⫽ 0.7333. Tree C is maximum parsimony control region tree from weighted data set, with tree length of 547 steps. Discovered using an exhaustive search; consistency index ⫽ 0.7459. Bootstrap values are percentages out of 5,000 bootstrap replications supporting a given grouping. Values indicated along each branch correspond to following node. Values below diagonals indicate number of synapomorphies supporting clade. Trees B and D. Tree B is distance-based tree from neighbor-joining analysis of COII sequences. Tree D is distance-based tree from neighbor-joining analysis of control region sequences. Both trees were derived from a distance matrix constructed using Kimura two-parameter distance method (Kimura, 1980). Frequency of grouping among clades is based on 10,000 bootstrap replications. Conﬁdence probabilities testing signiﬁcance level of corresponding branch lengths from zero were assessed using MEGA (version 1.02). Bootstrap values are indicated along each branch. Below diagonals are conﬁdence probabilities. tion homogeneity (ILD) test (PAUP) to discern if the data from these varied genetic regions could be combined for analysis. The data were partitioned into their respective data sets, and invariant characters were set to zero weight. This test indicated that data in all three genetic regions were congruent (P ⫽ 0.994). The combined data set includes 1,715 bp and 207 phylogenetically informative characters, with a ratio of 3.03 transitions per transversion (MacClade, version 3.05). An exhaustive search (PAUP) with either midpoint rooting, or Alouatta as the outgroup to tribe Atelini, produced two equally parsimonious trees, differing only within the Ateles genus. Brachyteles is joined with Ateles in both of the most parsimonious trees to form a clade with Lagothrix as the sister taxa, but without any bootstrap support. An exhaustive search (PAUP) of the data set with transversions weighted from 5:1 to as high as 20:1 over transitions produced the same two trees, which related Brachyteles to Ateles and provided weak bootstrap support of 51% (Fig. 4). Neighbor-joining analysis of the weighted, complete data set produced a tree which related Brachyteles and Lagothrix as a clade, with Ateles as the sister taxa. This relationship received no bootstrap support, however (Fig. 4). Analysis of this data set using maximum likelihood analysis produced a tree that supported Brachyteles as more closely related to Ateles, with an associated ln likelihood of ⫺7,929.98. The three data sets produced trees with different relationships among the Atelini depending on region and method of analysis, and are incongruent upon visual inspection. The signiﬁcance of this topological incongruence was assessed using the test of Templeton (1983) and the winning sites test (Prager and Wilson, 1988), using the topology of each independent data set as a constraint for the combined mitochondrial and nuclear data. The test of Templeton (1983) did not detect any signiﬁcant incongruence between any comparisons (P ⫽ 0.9130 for comparison between the two mitochondrial trees, P ⫽ 0.2513 for the Aldolase tree compared to the COII tree). The winning sites test also did not discover any signiﬁcant incongruence (P ⫽ 1.0 and 0.4545 for the above comparisons). ATELINAE PHYLOGENETIC TRICHOTOMY 291 Fig. 4. Trees A and C. Tree A presents maximum parsimony (PAUP) Aldolase A Intron V phylogeny from data set, with tree length of 30 steps. Discovered using an exhaustive search; consistency index ⫽ 1.00. Tree C presents maximum parsimony combined mitochondrial and nuclear DNA phylogeny from weighted data set, with tree length of 1,128 steps. Discovered using an exhaustive search; consistency index ⫽ 0.7624. Bootstrap values are percentages of 5,000 bootstrap replications supporting a given grouping. Values indicated along each branch correspond to following node. Values below diagonals indicate number of synapomorphies supporting clade. Trees B and D. Tree B is distance-based tree from neighbor-joining analysis of Aldolase A Intron V sequences. Tree D is distance-based tree from neighbor-joining analysis of combined mitochondrial and nuclear DNA sequences. Both trees were derived from a distance matrix constructed using Kimura two-parameter distance method (Kimura, 1980). Frequency of grouping among clades is based on 10,000 bootstrap replications. Conﬁdence probabilities testing signiﬁcance level of corresponding branch lengths from zero were assessed using MEGA (version 1.02). Bootstrap values are indicated along each branch. Below diagonals are conﬁdence probabilities. Reexamination of previous genetic studies An attempt was made to differentiate between the alternate and poorly supported phylogenies produced by the different genetic regions and analysis methods in the present study. This involved further examination of Atelinae relationships by combining sequence data from Atelinae haplotypes from previous Platyrrhine-wide nuclear phylogenetic studies with the data from this research project. The ⑀-globin sequences reported by Schneider et al. (1993) for the Atelinae were downloaded from GenBank (accession numbers L25358, L25366, L25367, L25369, and L25370) and aligned according to that study. Additional sequence information from other available studies was downloaded and analyzed in the same manner as the present study: 1) ␥-globin (Meireles et al., 1998) (accession numbers AF030092, AF030093, AF030094, AF030095, AF030097, and AF030098); 2) interstitial retinolbinding protein (IRBP) (Harada et al., 1985) (accession numbers U18602, U18603, U18605, and U18614); 3) ␤2 microglobulin (Canavez et al., 1999) (accession numbers AF032047, AF032050, AF032053, AF032086, AF032048, AF032049, AF032051, AF032052, AF032054, AF032055, AF032087, and AF032088); and 4) G6PD (von Dornum and Ruvolo, 1999) (accession numbers AF289989, AF289985, AF028502, AF028501, AF028513, AF028498, AF028478, AF028477, AF028489, and AF028473). Sequence data for the von Willebrand Factor Gene Intron II for all Atelinae (Vieria et al., 1997; Chaves et al., 1999) were not available on GenBank. Analysis of only the Atelinae haplotypes for ⑀-globin, ␥-globin , ␤2 microglobulin, and G6PD (Schneider et al., 1993; Meireles et al., 1998; Canavez et al., 1999; von Dornum and Ruvolo, 1999) all supported a Brachyteles-Lagothrix clade, with bootstrap support ranging from ⬍50 –100%. The G6PD sequences produced no bootstrap support using the weighting scheme of von Dornum and Ruvolo (1999). The weighted IRBP data equally supported either a 292 A.C. COLLINS TABLE 2. Comparative support among possible subclades in tribe Atelini1 Clade B-L B-A A-L Length Templeton P 1,300 1,327 1,331 Best 0.0090 0.0023 N 107 103 z Winning sites P ME score ⫺Ln likelihood ⫺2.610 ⫺3.05 Best 0.0120 0.0031 0.08613 0.08737 0.08868 28,303.69 28,368.43 28,375.32 1 Three possible subclades within tribe Atelini for combined present and previous data sets. There is no weighting of data. For parsimony comparisons, favored subclade is labeled as “best,” and P values for test of Templeton (1983) along with N and z values are presented, followed by P value for winning sites test (Prager and Wilson, 1988). Signiﬁcant values of P ⬍ 0.05 are in bold. For distance-based comparisons, minimum evolution (ME) score is presented. For likelihood comparisons, ⫺Ln likelihoods of each tree are also presented. All data were produced using the tree-scores option of PAUP. Brachyteles-Ateles or a Brachyteles-Lagothrix clade, using parsimony analysis for only Atelinae samples. This is contradictory to the full data-set results of Harada et al. (1995). Neighbor-joining results for the reduced IRBP data set do favor the BrachytelesLagothrix clade as supported by Harada et al. (1995), and provide 77% bootstrap support for the clade. The inclusion of additional haplotypes from other genera in the Platyrrhine-wide studies (Schneider et al., 1993; Harada et al., 1995; Schneider and Rosenberger, 1996; Barroso et al., 1997; Porter et al., 1997; Canavez et al., 1999; von Dornum and Ruvolo, 1999; Meireles et al., 1999; Schneider, 2000) acted to modify Atelini relationships in the full studies. In a ﬁnal attempt to differentiate among tribe Atelini members, all nuclear and mitochondrial data from the present study and additional sequences from all other studies were subjected to the partition homogeneity (ILD) test (PAUP) to discern if the data from these varied genetic regions could be combined for analysis. The data were partitioned into their respective data sets, and invariant characters were set to zero weight. The test indicated that data in all eight genetic regions were congruent (P ⫽ 0.75). Analysis of this complete data set contained only one member of each Atelinae genus due to the limitations of some of the studies. The total combined data set included 15,177 base pairs of DNA for eight genetic regions. Parsimony analysis of the unweighted data set produced a tree (1,300 steps; CI excluding uninformative characters ⫽ 0.653) that related Brachyteles most closely to Lagothrix, with Ateles as the sister taxa. Neighbor-joining analysis produced the same tree, and both had bootstrap support ranging from 98 –100%. The alternate Brachyteles-Ateles tree required an additional 27 steps to render. The various data sets produced trees that sometimes differed in relationships among tribe Atelini genera, and are thus topologically incongruent upon visual inspection. The signiﬁcance of this incongruence was assessed using the test of Templeton (1983) and the winning sites test (Prager and Wilson, 1988). The complete data set had only single representatives from each taxon; thus, any phylogram had only four possible OTUs. There were only three possible arrangements of tribe Atelini members in phylograms. The test of Templeton (1983) scored the Brachyteles-Lagothrix tree as the “best” tree and indicated signiﬁcant incongruence when comparing to the other possible trees, as did the winning sites test (Prager and Wilson, 1988). The results of these comparisons are presented in Table 2. This seeming conﬁrmation of subtribe Brachytelina is tempered by the fact that investigations of ⑀-globin (Schneider et al., 1993) and ␥-globin (Meireles et al., 1999) used only two species of Ateles, while IRBP, ␤2 microglobulin, and G6PD (Harada et al., 1995; von Dornum and Ruvolo, 1999; Canavez et al., 1999) utilized only one species of Ateles in their analyses. Only one Lagothrix and Brachyteles with no more than two Alouatta haplotypes were used in these studies as well. This sampling approach does not adequately measure genetic diversity among these widely dispersed genera, as demonstrated below. To examine the effects of using limited Ateles haplotypes to construct phylogenies among Atelinae genera, the complete mitochondrial and nuclear data set from the present study was reanalyzed with the restriction of including only one or two haplotypes of spider monkeys. The reduced set of Ateles haplotypes clearly does not represent the wide degree of genetic variation present among the various populations of spider monkeys seen in the complete data set. Using only one Ateles haplotype, 2 of the 6 haplotypes produced relationships, using parsimony analysis, that favored a Brachyteles-Ateles clade, with the other 4 favoring a Brachyteles-Lagothrix clade (Table 3). Neighbor-joining analyses of these data sets produced only trees favoring a BrachytelesLagothrix clade (Table 3). Analysis of the data set including only the two species of Ateles used by Schneider et al. (1993) produced two equally parsimonious trees, indicating equal support for either a Brachyteles-Ateles clade or a Brachyteles-Lagothrix clade. A neighbor-joining analysis of this data set favored a Brachyteles-Lagothrix clade, but this received no bootstrap support (Table 3). Examination of all possible combinations of only two Ateles species using parsimony produced relationships of Brachyteles to both Ateles and to Lagothrix (Table 3). Of the 15 possible combinations, 8 phylograms favored a Brachyteles-Lagothrix clade, and 5 trees favored a Brachyteles-Ateles clade, with 293 ATELINAE PHYLOGENETIC TRICHOTOMY TABLE 3. Effect of different Ateles species on phylogenies1 Ateles species included Parsimony relationship Bootstrap support Neighbor-joining relationship Bootstrap support A. paniscus5 A. hybridus A. belzebuth chamek A. belzebuth marginatus2 A. geoffroyi yucatanensis A. geoffroyi robustus6 A. g. robustus and A. g. yucatanensis A. paniscus and A. b. chamek A. b. chamek and A. b. marginatus A. paniscus and A. g. yucatanensis4 A. g. yucatanensis and A. hybridus A. b. chamek and A. hybridus A. paniscus and A. b. marginatus A. paniscus and A. hybridus A. g. robustus and A. hybridus A. b. chamek and A. g. robustus A. b. chamek and A. g. yucatanensis A. g. robustus and A. b. marginatus A. b. marginatus and A. hybridus A. b. marginatus and A. g. yucatanensis3 A. paniscus and A. g. robustus All Ateles specimens Brachy-Ateles Brachy-Ateles Brachy-Lagothrix Brachy-Lagothrix Brachy-Lagothrix Brachy-Lagothrix Brachy-Lagothrix Brachy-Lagothrix Brachy-Lagothrix Brachy-Ateles Brachy-Ateles Brachy-Lagothrix Brachy-Ateles Equal support Brachy-Ateles Brachy-Lagothrix Brachy-Lagothrix Brachy-Lagothrix Brachy-Lagothrix Equal support Brachy-Ateles Brachy-Ateles 51% ⬍50% ⬍50% 70% ⬍50% 64% ⬍50% ⬍50% 62% 63% ⬍50% 53% ⬍50% Brachy-Lagothrix Brachy-Lagothrix Brachy-Lagothrix Brachy-Lagothrix Brachy-Lagothrix Brachy-Lagothrix Brachy-Lagothrix Brachy-Lagothrix Brachy-Lagothrix Brachy-Ateles Brachy-Ateles Brachy-Lagothrix Brachy-Lagothrix Brachy-Ateles Brachy-Lagothrix Brachy-Lagothrix Brachy-Lagothrix Brachy-Lagothrix Brachy-Lagothrix Brachy-Lagothrix Brachy-Lagothrix Brachy-Lagothrix 51% 52% 60% 70% ⬍50% ⬍50% ⬍50% 52% 57% ⬍50% ⬍50% ⬍50% 55% 51% ⬍50% 56% 51% 57% ⬍50% ⬍50% ⬍50% ⬍50% ⬍50% ⬍50% ⬍50% 60% 58% ⬍50% ⬍51% 1 Details shifting support for two most likely clades among tribe Atelini, based on number and composition of Ateles species included, following most inclusive taxonomy (Groves, 1989) in an exhaustive parsimony search (PAUP) and with neighbor-joining analysis (PAUP). Includes all possible trees with inclusion of only one or two Ateles haplotypes. Bootstrap support is based on 1,000 replications. 2 One species utilized by Harada et al. (1995). 3 Two species utilized by Schneider et al. (1993). 4 Two species utilized by Meireles et al. (1999). 5 One species used by Canavez et al. (1999). 6 One species used by von Dornum and Ruvolo (1999). At bottom, information with inclusion of all Ateles specimens is repeated for comparative purposes. none of these receiving ⬎75% bootstrap support. Two haplotype combinations produced identical support for the relationship of Brachyteles to either Ateles or Lagothrix (Table 3). Distance-based analyses using only two species of Ateles produced trees that were sometimes contradictory to the parsimony trees (Table 3). Analysis of the 15 haplotype combinations produced 3 trees relating Brachyteles and Ateles, while 12 trees related Brachyteles to Lagothrix with variable, relatively low levels of bootstrap support (Table 3). These examples demonstrate how relationships can change, based on the number and composition of Ateles haplotypes included. Results and comparisons for all these combinations are presented in Table 3. DISCUSSION Results based on the new data presented in this study demonstrate the high similarity between Atelini genera. All of the genetic regions investigated support different relationships among tribe Atelini genera, which vary by region, weighting scheme, and analysis method. The control region sequences, which support a Brachyteles-Lagothrix clade, may possibly be too homoplasic for discrimination between separate, yet closely related genera. Even combined analysis of all three loci supports different relationships, depending on the method of analysis. These ﬁndings demonstrate the very close relationships between these genera. They further demonstrate the associated difﬁculty in differentiating the actual phylogenetic relationship, even with multiple haplotypes and species/subspecies of Ateles, and Alouatta for COII, in the data sets. Some readers might suggest that these results demonstrate the overall poor resolving power of the loci chosen in this study. While this could be the case, it is more likely to be indicative of the very close genetic relationships among all three genera and the increased need for inclusion of multiple samples from different species/subspecies or genetically distinct populations in the determination of relationships in tribe Atelini. In reanalysis using only Atelinae DNA sequences from previous studies (Schneider et al., 1993; Harada et al. 1995; Canavez et al., 1999; Meireles et al., 1999; von Dornum and Ruvolo, 1999), there was a total of 44 phylogenetically informative characters when all data sets were reduced to just one member of each genus. This is compared to 70 phylogenetically informative characters for the present data set with only one member of each genus. This suggests that the resolving power of this data set is at least as good as the combination of previous data sets or better. The low bootstrap support for clades and the variable relationship of Brachyteles in these trees are indicative of the extreme similarity between these genera, and not the lack of phylogenetically informative sites in the sequences chosen. Even though the test of Templeton (1983) and the winning sites test (Prager and Wilson, 1988) found signiﬁcant support for the Brachyteles-Lagothrix 294 A.C. COLLINS tree in the combined data set, these same tests detected no signiﬁcant differences between the possible trees for analysis of the present data set alone. The inclusion of additional Atelinae haplotypes in this study likely contributed to the lack of signiﬁcant differences between possible trees in the present data set. Ateles is a genetically variable genus with significant intra- and interspeciﬁc variation (Collins and Dubach, 2000). When only limited samples of such a genetically and geographically diverse taxon are included, it is possible to omit the genetic variation needed to discern taxonomic relationships. Limited sampling of the Atelinae genera would seem likely to affect the outcome of phylogenetic reconstructions. The effects of sampling only a portion of the genetic variation in Ateles were demonstrated in Table 3, with comparisons of phylogenies with limited Ateles haplotypes included. Support switched back and forth in the limited haplotype analysis of the present data, depending on species/populations included, thus suggesting that similar results might occur if the previous data sets could be examined in a similar fashion. Since this data set was not originally gathered to address Atelinae phylogenetics, the COII data set was the only one where multiple samples of other non-Ateles genera were available for phylogeny construction via the use of sequences deposited in GenBank. Analysis of this locus goes further in demonstrating the impact of multiple vs. single samples of all genera on Atelini relationships. Inclusion of all available Atelinae haplotypes, using the weighted data set, produced 62% bootstrap support for a Lagothrix-Ateles clade, but inclusion of only the more complete Alouatta haplotype determined in this study produced a tree with only a Brachyteles-Ateles clade, with bootstrap support of 56%. In earlier versions of this research, the Alouatta caraya haplotype was not available, and all trees analyzed supported a Brachyteles-Ateles clade. It was only with inclusion of this new sample and reweighting of newly informative transversions that the data supported the Ateles-Lagothrix clade. These results clearly indicate the importance of including not only additional haplotypes of Ateles in future analyses, but also additional haplotypes of the geographically widespread and genetically variable members of Alouatta, and likely Lagothrix and Brachyteles too. The possibility that several of the gene trees do not represent the actual species tree is feasible, especially given the limited number of samples utilized for each genus. A great deal of support for a Brachyteles-Lagothrix clade comes from studies of the ⑀-globin, ␥-globin, and ␤2 microglobulin genes (Schneider et al., 1993; Schneider and Rosenberger, 1996; Porter et al., 1997; Canavez et al., 1999; Meireles et al., 1999). ⑀-globin and ␥-globin are part of the ␤-globin gene cluster, where ⑀-globin is embryonically expressed and ␥-globin is fetally expressed in anthropoids (Meireles et al., 1995; Chiu et al., 1996). The mode of expression and organization of these genes is unique among the Atelinae (Meireles et al., 1995; Chiu et al., 1996). Among Atelinae, the ␥1 portion has been inactivated due to a 1.8-kb deletion. Gene conversions between ␥1 and ␥2 are common among many Platyrrhine species (Meireles et al., 1995), but Meireles et al. (1999) found only two conversions between loci in Atelinae. They eliminated these stretches from their analyses, but it is interesting to note that one of the conversions occurred in only one of the two spider monkeys they sampled. If they had sampled more species/subspecies of Ateles, it seems possible that more conversion events might have been discovered. Imagine a scenario where only one species of Ateles had been sampled and this conversion would have been missed, and the data would have then been incorporated into phylogenetic reconstructions. This suggests that other gene conversions in other regions may have occurred in additional populations of Lagothrix and Brachyteles, which might be acting to confound the results and skew the relationships between the limited haplotypes chosen. Additionally, it is clear that ⑀-globin and ␥-globin are closely linked loci, and the ␤2 microglobulin gene is possibly linked. Thus it is not surprising that they produce similar phylogenies. As can be seen in examination of the various distance-based cladograms, the shared branch length between all of the Atelini is very short. The results of this study and the morphological analysis of Ford (1986) suggest a very close relationship among the three genera of tribe Atelini, a suggestion that was also proposed by Schneider and Rosenberger (1996). CONCLUSIONS This research provides additional support for the similarity of members of tribe Atelini, providing validity for the cladistic arrangement of two tribes, Alouattini and Atelini, among the Atelinae subfamily. The relationship of Brachyteles to either Ateles or Lagothrix is demonstrated to be unresolved due to the lack of inclusion of genetically variable samples among Atelini genera in previous genetic studies and possibly this study as well. Based on the evidence from this study on the importance of the inclusion of multiple Atelinae species, the assertions of Harada et al. (1995), Schneider and Rosenberger (1996), Canavez et al. (1999), Meireles et al. (1999), and von Dornum and Ruvolo (1999) that Brachyteles and Lagothrix should constitute a subtribe Brachytelina need to be reviewed. It appears that, while this may be the actual relationship, the evidence presently available is not clear enough to support any speciﬁc Atelini subtribes. Future studies should also attempt to obtain additional specimens of the different Alouatta species, various Lagothrix species and subspecies, and samples from both disjunct populations of Brachyteles. A large amount of sequence data exists to address the relationships among tribe Atelini, but until this ATELINAE PHYLOGENETIC TRICHOTOMY amount is obtained on a greater number of populations and species within the Atelinae, no deﬁnitive relationships among tribe Atelini can be determined. ACKNOWLEDGMENTS I thank the Conservation Biology Department of Brookﬁeld Zoo (Chicago, IL) for providing research support at the Daniel F. and Ada L. Rice Conservation Biology and Research Center’s Genetics Laboratory for this project. Grant support was awarded to J.M. Dubach and the Brookﬁeld Zoo by SEACON (Chicago Zoological Society Conservation Research Grant) and the Institute of Museum and Library Services (1C-40190-94). Funding was also provided by a Sigma Xi Grant in Aid of Research to A.C.C. I thank J.M. Dubach, Trudy Turner, and Karen Strier for comments about this manuscript at varying stages of its progress. The comments of two anonymous reviewers also served to strengthen and focus this paper. Invaluable samples of spider monkeys were provided by the following individuals and institutions: Horacio Schneider and the Universidade Federal do Para in Belem, Brazil, who provided most of the wild-caught samples; Morris Goodman at Wayne State University; Theresa Pope at Duke University; Guy Hoelzer at Utah State University; Robin Brockett, who helped obtain the Belizean samples; the Lubee Foundation; Bruce Patterson at the Field Museum of Natural History in Chicago; La Merida Zoo in Mexico; Club Auto Safari in Guatemala; the National Zoo in Washington, DC; Scott Carter at the Detroit Zoo; and the Charles Paddock Zoo (Atascadero, CA). LITERATURE CITED Adkins RM, Honeycutt RL. 1994. Evolution of the primate cytochrome C oxidase subunit II gene. J Mol Evol 38:215–231. Ascunce MS, Hasson E, Mudry MD. 2002. Description of the cytochrome c oxidase subunit II gene in some genera of New World monkeys (Primates, Platyrrhini). Genetica 114:253–267. Barroso CML, Schneider H, Schneider MPC, Sampaio I, Harada ML, Czelusniak J, Goodman M. 1997. Update on the phylogenetic systematics of New World monkeys: further DNA evidence for placing the pygmy marmoset (Cebuella) within the genus Callithrix. Int J Primatol 18:651– 674. Canavez FC, Moreira MAM, Ladasky JJ, Pissinatti A, Parham P, Seuanez HN. 1999. Molecular phylogeny of New World primates (Platyrrhini) based on ␤2-microglobulin DNA sequences. Mol Phylogenet Evol 12:74 – 82. Chaves R, Sampaio I, Schneider MPC, Schneider H, Page SL, Goodman M. 1999. The place of Callimico goeldii in the Callitrichine phylogenetic tree: evidence from von Willebrand factor gene intron II sequences. Mol Phylogenet Evol 13:392– 404. Chiu C, Schneider H, Schneider MPC, Sampaio I, Meireles C, Slightom JL, Gumucio DL, Goodman M. 1996. Reduction of two functional ␥-globin genes to one: an evolutionary trend in New World monkeys (infraorder Platyrrhini). Proc Natl Acad Sci USA 93:6510 – 6515. Collins AC, Dubach J. 2000. Phylogenetic relationships among populations of spider monkeys (Ateles): based on analysis of mitochondrial DNA variation. Int J Primatol 21:381– 420. Collins AC, Dubach J. 2001. Nuclear DNA variation in spider monkeys (Ateles). Mol Phylogenet Evol 19:67–75. Dunlap SS, Thorington RW Jr, Aziz MA. 1985. Forelimb anatomy of New World monkeys: myology and the interpretation of 295 primitive anthropoid models. Am J Phys Anthropol 68:499 – 517. Dutrillaux B. 1979. Chromosomal evolution in primates: tentative phylogeny from Microcebus murinus (prosimians) to man. Ann Genet 48:251–314. Farris JS, Kallersjo M, Kluge AG, Bult C. 1995. Testing signiﬁcance of incongruence. Cladistics 10:315–319. Figueiredo WB, Carvalho-Filho NM, Schneider H, Sampaio I. 1998. Mitochondrial DNA sequences and the taxonomic status of Alouatta seniculus populations in northeastern Amazonian. Neotrop Primates 6:73–77. Ford SM. 1986. Systematics of the New World monkeys. In: Swindler DR, Erwin J, editors. Comparative primate biology, volume I: systematics, evolution and anatomy. New York: Alan R. Liss. p 73–135. Ford SM, Davis LC. 1992. Systematics and body size: implications for feeding adaptations in New World monkeys. Am J Phys Anthropol 88:415– 468. Froehlich JW, Supriantna J, Froehlich PH. 1991. Morphometric analyses of Ateles: systematic and biogeographic implications. Am J Primatol 25:1–22. Gray JE. 1870. Catalogue of monkeys, lemurs, and fruit-eating bats in the collection of the British Museum. London: British Museum Press. Groves CP. 1989. A theory of human and primate evolution. Oxford, UK: Clarendon Press. Harada ML, Schneider H, Schneider MPC, Sampaio I, Czelusniak J, Goodman M. 1995. DNA evidence of the phylogenetic systematics of New World monkeys: support for the sistergrouping of Cebus and Saimiri from two unlinked nuclear genes. Mol Phylogenet Evol 4:331–349. Hillis DM, Moritz C, Mable BK. 1996. Molecular systematics. 2nd ed. Sunderland, MA: Sinauer Associates, Inc. Kay RF. 1990. The phyletic relationships of extant and fossil Pitheciinae (Platyrrhini, Anthropoidea). J Hum Evol 19:175– 208. Kellogg R, Goldman EA. 1944. Review on the spider monkeys. Proc US Mus Nat Hist 96:1– 45. Kimura M. 1980. A simple model for estimating evolutionary rates of base substitutions through comparative studies of nucleotide sequences. J Mol Evol 16:111–120. Kocher TD, Thomas WK, Meyer A, Edwards SV, Paabo S, Villablanca FX, Wilson AC. 1989. Dynamics of mitochondrial DNA evolution in animals. Ampliﬁcation and sequencing with conserved primers. Proceedings of National Academy of Sciences 86:6196 – 6200. Koiffmann CP, Saldanha JH. 1974. Cytogenetics of Brazilian monkeys. J Hum Evol 3:275–282. Kumar S, Tamura K, Nei M. 1993. MEGA: molecular evolutionary genetics analysis. Version 1.02. University Park, PA: Pennsylvania State University. Maddison WP, Maddison DR. 1992. MacClade: analysis of phylogeny and character evolution. Sunderland, MA: Sinauer Associates. Meireles C, Schneider MPC, Sampaio MI, Schneider H, Slightom JL, Chiu C, Neiswanger K, Gumucio DL, Czelusniak J, Goodman M. 1995. Fate of a redundant ␥-globin gene in the Atelid clade of New World monkeys: implications concerning fetal globin gene expression. Proc Natl Acad Sci USA 92:2607–2611. Meireles C, Czelusniak J, Schneider MPC, Chiu CH, Slightom JL, Gumucio DL, Muniz JAPC, Brigido MC, Ferreira HS, Goodman M. 1999. Molecular phylogeny of Atelinae New World monkeys (Platyrrhini, Atelinae) based on ␥-globin gene sequences: evidence that Brachyteles is the sister group of Lagothrix. Mol Phylogenet Evol 12:10 –30. Paterson JD. 1991. Systematics and taxonomy. In: Stephens ME, Paterson JD, editors. The order Primates: an introduction. Dubuque, IA: Kendall/Hunt Publishing. p 36 – 48. Pequignot EV, Koiffmann CP, Dutrillaux B. 1985. Chromosomal phylogeny of Lagothrix, Brachyteles, and Cacajao. Cytogenet Cell Genet 39:99 –104. Pope TR. 1998. Genetic variation in remnant populations of the woolly spider monkey (Brachyteles arachnoides). Int J Primatol 19:95–109. 296 A.C. COLLINS Porter CA, Page SL, Czelusniak J, Schneider H, Schneider MP, Sampaio I, Goodman M. 1997. Phylogeny and evolution of selected primates as determined by sequences of the ⑀-globin locus and 5⬘ ﬂanking regions. Int J Primatol 18:261–295. Prager EM, Wilson AC. 1988. Ancient origin of lactalbumin from lysozyme: analysis of DNA and amino acid sequences. J Mol Evol 27:326 –335. Rosenberger AL, Strier KB. 1989. Adaptive radiation of the Atelinae primates. J Hum Evol 18:717–750. Rowe N. 1996. The pictorial guide to the living primates. East Hampton, NY: Pogonias Press. Rumpler Y, Strasbourg SC. 1992. Importance of comparative cytogenetics for primate systematics and phylogeny: limits and alternatives. Karger Gazette 54:8 –9. Ruvolo M. 1997. Molecular phylogeny of the hominoids: inferences from multiple independent DNA sequence data sets. Mol Biol Evol 14:248 –265. Sambrook J, Fritsch ET, Maniatis T. 1989. Molecular cloning: a laboratory manual. Volumes 1–3. Cold Spring Harbor, NY: Cold Spring Harbor Publishers. Schneider H. 2000. The current status of the new world monkey phylogeny. Ann Acad Brasil Cie 72:165–172. Schneider H, Rosenberger AL. 1996. Molecules, morphology, and platyrrhine systematics. In: Norconck MA, Rosenberger AL, Garber PA, editors. Adaptive radiations of neotropical primates. New York: Plenum Press. p 3–18. Schneider H, Schneider MPC, Sampaio I, Harada ML, Stanhope M, Czelusniak J, Goodman M. 1993. Molecular phylogeny of the New World monkeys. Mol Phylogenet Evol 2:225–242. Strier KB. 1992. Atelinae adaptations: behavioral strategies and ecological constraints. Am J Phys Anthropol 88:515–524. Swofford DL. 1998. PAUP: phylogenetic analysis using parsimony. Version 4.0b1. Washington, DC: Smithsonian Institution. Templeton AR. 1983. Convergent evolution and non-parametric inferences from restriction fragment and DNA sequence data. In: Weir B, editor. Statistical analysis of DNA sequence data. New York: Marcel Dekker. p 151–179. Vieira R, Sampaio I, Schneider H, Goodman M. 1997. Relacoes intergenericas na subfamilia Atelinae (Platyrrhini, Primates) atraves do estudo do intron 11 do gene do fator de von Willebrand. VIII Congresso Brasileiro de Primatologiz e V Reunido Latino Americano de Primatologiz. Paraiba: Universidade Federal da Paraiba, Sociedade Brasileira de Primatologia. p 101. von Dornum M, Ruvolo M. 1999. Phylogenetic relationships of the New World monkeys (Primates, Platyrrhini) based on nuclear G6PD DNA sequences. Mol Phylogenet Evol 11:459 – 476.