Journal of Human Evolution 59 (2010) 218e222

Contents lists available at ScienceDirect

Journal of Human Evolution

journal homepage: www.elsevier.com/locate/jhevol

News and Views Stem taxa, homoplasy, long lineages, and the phylogenetic position of Dolichocebus

Richard F. Kay a,*, John G. Fleagle b a Department of Evolutionary Anthropology, Duke University, Durham, NC 27708, USA b Department of Anatomical Sciences, Health Sciences Center, Stony Brook University, Stony Brook, New York 11794-8081, USA

article info identifying them as being members of a clade. In our analysis, a suite of derived anatomical features that are generally lacking in Article history: the Patagonian taxa unites all living, or crown, platyrrhines. Among Received 16 November 2009 other things, compared with the Patagonian taxa, crown platyr- Accepted 12 February 2010 rhines have a shorter rostrum, a more transversely arched palate,

Keywords: more convergent orbits, reorganization of the arrangement of the Platyrrhini bones at pterion, larger, more transversely positioned premolar Phylogenetics metaconids, loss of molar hypoconulids and of a postentoconid Anthropoidea sulcus, and reduction of molar roots. It is the absence of the above- Molecular clocks listed derived features of crown platyrrhines in the Patagonian taxa that suggests each of them is a stem platyrrhine, and outside of the modern radiation. The Patagonian taxa may belong to a separate stem clade or they may be a paraphyletic group, one of which may prove to be the sister to the most recent common In his reply to our recent paper (Kay et al., 2008) entitled “The ancestor of living platyrrhines. Either way, Dolichocebus, Trem- anatomy of Dolichocebus gaimanensis, a stem platyrrhine acebus and the other less-well known Patagonian early from Argentina,” Rosenberger (2010: p.1) addresses a wide range of monkeys are stem platyrrhines. topics including the anatomy of the fossils, molecular dates for platyrrhine evolution, phylogenetic systematics, parallelism, and fl more broadly “the nature of phylogenetic, functional, and adaptive Re ections on homoplasy evidence in historical evolutionary reconstruction.” These are very large and complex topics, worthy of debate. Nevertheless, we will As Rosenberger notes, parallelism or homoplasy is present in all restrict our comments in this rebuttal to a few simple observations. evolutionary reconstructions of phylogeny. However, He “ Rosenberger rejects our conclusions that many of the fossil plat- (2010: p.4) chides us for failing to acknowledge that not all ” yrrhines from the early Miocene of Argentina are stem platyrrhines. homoplasy is created equal. While this may be true, most scholars, fi He (2010: p.1) claims that our “conclusion rests on a weakly sup- including us, agree that homoplasy can only be identi ed in the ’ ported cladistic artifact, a purported monophyletic group of context of a phylogenetic reconstruction. Rosenberger s reply southern platyrrhines.” Even more clearly, Rosenberger (2010: p.2) mirrors his 2002 essay (Rosenberger, 2002) emphasizing the states “A plethora of issues cast doubt on the reliability of Kay et al.’s problems posed by homoplasy for the type of phylogenetic analyses (2008) ‘stem group’ hypothesis.” He claims that the “central tenant” used in our paper. He repeats several widely perceived miscon- of our stem group hypothesis is that early Miocene Dolichocebus, ceptions about homoplasy as a phenomenon. One is that the pres- Tremacebus, Carlocebus, and Soriacebus form a clade. But our ence of homoplasy in an analysis is some sort of analytical artifact hypothesis does not rest on this ‘central tenant.’ Rosenberger is rather than an evolutionary reality. Another is that the presence of “ ” either misrepresenting our position or completely misunder- homoplasy is the result of using the wrong kinds of characters. A fl standing of the concept of stem taxa (see Williams and Kay, 1995; third is that the level of homoplasy in an analytical result re ects the fi Williams et al., 2010). con dence one can have in the resulting tree. All studies of homo- Contrary to what Rosenberger claims, our identification of the plasy in phylogeny reject these views (e.g., Sanderson and Patagonian fossils, including Dolichocebus, as stem platyrrhines Donoghue, 1989, 1996; Sanchez-Villagra and Williams, 1998; outside the clade of living platyrrhine genera is not dependent on Lockwood and Fleagle, 1999). It is often suggested that homoplasy is more common in some types of characters than others and therefore analyses using such * Corresponding author. characters are likely to generate high levels of homoplasy and E-mail address: [email protected] (R.F. Kay). unreliable results (e.g., Collard and Wood, 2000, 2001). Certainly it

0047-2484/$ e see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.jhevol.2010.03.002 R.F. Kay, J.G. Fleagle / Journal of Human Evolution 59 (2010) 218e222 219 is true that, in any given analysis, different types of characters may between Dolichocebus and Saimiri are not even observable because show different levels of homoplasy (e.g., Ross et al., 1998). However, the relevant areas are broken away in the fossil. In his reply, these seem to be idiosyncratic results from individual analyses of Rosenberger continues to dwell on the phylogenetic significance of particular taxa and cannot be generalized from one analysis or one the midsagittal profile of the frontal bone (glabellar region) and the group of taxa to another (Lockwood and Fleagle, 1999). All broad shape and size of its underlying frontal cortex and olfactory bulbs. comparative studies of homoplasy have found no evidence that any The midsagittal part of the frontal bone is not preserved in Doli- particular type of data (e.g., molecular, osteological, dental, chocebus. Rosenberger agrees, but states that Dolichocebus’ glabella behavioral) consistently shows higher levels of homoplasy than must have been ‘domed’ as it is in Saimiri and Cebus because the another (e.g., Sanchez-Villagra and Williams, 1998; Lockwood and natural endocast of the frontal lobe of the neocortex is expanded in Fleagle, 1999). There are no general rules. Thus, features that may Dolichocebus. Specifically, he (2010: p.3) claims that Dolichocebus is be extremely labile in one clade may be essentially fixed in another. similar to cebines (Saimiri and Cebus): “the prominent olfactory For example, fusion of the mandibular symphysis has occurred bulbs and high arc of the frontal lobedthe underlying morphology several times in strepsirrhines and perhaps only once in crown of import, as stateddare well preserved and clearly show the anthropoids. The entepicondylar foramen of the humerus was lost anterior braincase had a cebine-like, domed shape rather than the many times in platyrrhines but only a few times in catarrhines. flat frontal region which is primitive among NWM.” We find three While homoplasy unquestionably makes the reconstruction of problems with this argument. First, the profile of the frontal cortex phylogenetic relationships more difficult, and character evolution of Dolichocebus is not domed as it is in Cebus and Saimiri. Second, more “messy” than one might prefer, it is a real evolutionary the olfactory bulbs of the fossil differ markedly from those of Sai- phenomenon that is present at all levels of biological analysis from miri and Cebus. And third, comparison of this part of the braincase amino-acid sequences to aspects of adult bony morphology and in extant platyrrhines demonstrates that the glabellar region of the behavior. Certainly there are methodological ways of generating frontal cannot be inferred from the endocranial profile of the “spurious” homoplasy in an analysis (e.g., through aspects of char- frontal cortex. acter definition) and ways of artificially minimizing the amount of With respect to the first two points, Rosenberger correctly states homoplasy in a phylogenetic tree (e.g., by arbitrarily eliminating that Saimiri and Cebus both have a high arc to the frontal bone, the characters with high homoplasy levels in all trees). However, ‘domed’ condition. Undoubtedly this is a correlative of the enlarged homoplasy is not primarily a result of methodological mistakes, but neocortex in these two species compared to most other NW rather the result of a variety of well-established evolutionary monkeys (Stephan et al., 1981; Fig. 1). The enlarged cebine frontal phenomena, including the nature of the genetic code, functional and neocortex expands dorsally over the olfactory fossae leaving the developmental constraints, allometry, and especially parallel or fossae ventral to the frontal lobes, not projecting rostral to it (Fig. 1A convergent adaptation (Lockwood and Fleagle, 1999). and D). In contrast, in , the frontal lobes are smaller, and While recognizing the existence of homoplasy, Rosenberger the midsagittal profile of the frontal lobes is sloping, not ‘domed’ takes the view that certain character complexes are unlikely to (Fig. 1C and F). The olfactory bulbs of Callicebus are no larger than appear in parallel in New World monkeys (NWM) because they those of cebines (bulb volume data from Baron et al. [1983]), but reflect a fundamental niche partitioning that was entrained very they project rostrally in front of the frontal lobes. However, early in platyrrhine evolution and, once set in motion, has contrary to Rosenberger’s claims, Dolichocebus more closely continued to the present day. In his view it is to those characters resembles Callicebus than cebines: the midsagittal profile of its that we should turn a priori in placing extinct platyrrhines into frontal endocast is sloping and the olfactory fossae of Dolichocebus a phylogeny. However, as noted above, most reviews of homoplasy projects far rostral to the neocortex (Fig. 1E). In short, in having indicate that “reliable” characters are only evident after an analysis a sloping frontal neocortex and rostrally projecting olfactory fossae, and cannot be identified in advance. Dolichocebus displays the primitive platyrrhine condition, like Beyond the theoretical issues raised by Rosenberger’s paper Callicebus, not the derived Saimiri-Cebus morphology. about how homoplasy is recognized and treated in a phylogenetic With respect to the third point, even if the profile of the natural analysis, we object to many of the specifics of his argument. As brain endocast of Dolichocebus did resemble that of Saimiri and discussed below, many of the features Rosenberger cites linking Cebus in being nearly vertical (which it does not) the condition of Dolichocebus with extant squirrel monkeys through cranial glabella still cannot reliably be inferred to be cebine-like. As illus- anatomy are simply not preserved in the fossil, so it is meaningless trated in Fig. 1,inSaimiri, the inner and outer tables of the braincase to discuss whether they evolved multiple times or not. Other conform to one another throughout this region but in Callicebus, characters are present in Dolichocebus and Saimiri but should not be the inner and outer tables are widely separated by an intervening accorded canonical status to which they are ascribed because they frontal sinus. demonstrably have evolved or been lost multiple times in platyr- Thus, to the extent that the frontal region is preserved in Doli- rhine evolution. chocebus, the structure is more similar to Callicebus than to Saimiri or Cebus. Disputed aspects of Dolichocebus anatomy Rosenberger persists in claiming that the masticatory apparatus of Dolichocebus is ‘gracile,’ but we remain unconvinced. A recent To coherently debate the anatomical evidence that might, or study by Perry et al. (2006, in press) shows that Dolichocebus, like might not, indicate that Dolichocebus is a sister taxon to Saimiri, Tremacebus and Homunculus, has larger tooth roots than compa- it is essential to agree about the anatomical facts. Kay et al. (2008) rable-sized living platyrrhines. We have nothing further to add to dispute a number of Rosenberger’s anatomical claims (Rosenberger, our observations about the masticatory apparatus of Dolichocebus 1979; Rosenberger et al., 1990), as summarized in Table 1. Some of other than to repeat that much of the anatomy is not preserved (see the claimed resemblances are not resemblances at all. For example, Fig. 11 in Kay et al., 2008). The claim that the interorbital region is published cranial measurements of extant platyrrhines (Kay et al., very narrow is difficult to sustain given the state of preservation of 2008: 348, Table 6,) do not support Rosenberger’s claim that Doli- the interorbital region (Kay et al., 2008: Figs. 7 and 12). In any event, chocebus exhibits dolichocephaly (to the extent that this can even having a narrow interorbital region should not be taken as char- be measured on its distorted skull) or even that dolichocephaly is acteristic of a ‘lightly built’ masticatory apparatus, because hard- characteristic of Saimiri or Cebus. Other proposed resemblances object feeders like Cebus and the pitheciines Cacajao and Chiropotes 220 R.F. Kay, J.G. Fleagle / Journal of Human Evolution 59 (2010) 218e222

Table 1 Anatomical characteristics of Dolichocebus reportedly shared either with Saimiri alone or cebines (Cebus and Saimiri). We comment on the actual anatomical condition in the right column. Seven of the ten features originally described in Dolichocebus are not preserved.

Character Condition in Dolichocebus Dolichocebus condition Dolichocebus condition Our anatomical observations as reported by Rosenberger suggested as a suggested as a (Kay et al., 2008; and this response) (1979, 2010); Rosenberger synapomorphy of Cebus and synapomorphy with et al. (1990) Saimiri (Rosenberger, Saimiri (Rosenberger, 1979, 2010) 1979, 2010) Braincase shape Dolichocephalic braincase X Living cebines are not dolichocephalic (Kay et al., 2008: 348, Table 6). The only notably dolichocephalic taxon is Leontopithecus Interorbital region Narrow interorbital pillar X Not preserved in fossil (Kay et al., 2008: Figs. 7 and 12) Nasal bones Short and narrow nasal bones X Not preserved in fossil (Kay et al., 2008: p. 347) Pyramidal process Lightly built X Not preserved in fossil (Kay et al., 2008: 335) Glabella Convex glabella X Not preserved in fossil (see text and Fig. 1) Frontal profile Probably vaulted frontal X Frontal lobes are not vaulted (see text and Fig. 1) Sylvian sulcus Sharply delineated sylvian sulcus X All platyrrhines have a well-developed sylvan fissure.1 Interorbital process of frontal Elongate descending interorbital X Not preserved in fossil (Kay et al., 2008: process of the frontal Fig. 7A and B) Interorbital fenestration Fenestrated interorbital septum X Not preserved in fossil (Kay et al., 2008: Figs. 7A, B, 11) Frontal lobe Large frontal lobe X Midsagittal profile of the frontal lobe is unexpanded, (see text).

Fig. 1. Midsagittal sections through skulls of Saimiri sciureus (A, D), Dolichocebus (B, E), and Callicebus moloch (C, F). The endocranial surface is depicted in red. Profiles of the endocranial surface of the frontal cortex and olfactory fossae are labeled. In Saimiri, the endocranial surface conforming to the frontal cortex has a vertical profile and the cranial vault is domed. The olfactory fossae do not project rostrally beneath the frontal lobe. Callicebus has a sloping endocranial frontal profile and its olfactory fossae project rostrally in front of the frontal cortex. The inner and outer tables of the frontal bone more or less coincide in Saimiri, whereas in Callicebus, a large air-filled sinus, identified by the letter S, separates the inner and outer tables. R.F. Kay, J.G. Fleagle / Journal of Human Evolution 59 (2010) 218e222 221

Fig. 2. A. Cladogram of extant platyrrhine phylogeny based on molecular sequence and Alu data. References from Kay et al. (2008), with the addition of Hodgson et al. (2009), Wildman et al. (2009), Schrago (2007), and Opazo et al. (2006). Numbers on cladogram indicate a taxon or taxa of extinct platyrrhines that establish the earliest occurrence of the extant clades. Except as noted, dates of occurrence are based upon paleomagnetic and radiometric studies reported by Flynn et al. (1997) and Madden et al. (1997) with the geomagnetics recalibrated from Gradstein et al. (2004).1.Miocallicebus (13.2e13.9 Ma; Takai et al., 2001). 2. Proteropithecia (<15.7 Ma; Kay et al., 1998), Nuciruptor (12.4e12.8 Ma; Meldrum and Kay, 1997), and (13.0e13.6 Ma). 3. Neosaimiri spp. (12.0e13.2 Ma). 4. “” (13.0e13.2 Ma). 5. (13.4e13.6 Ma), (13.0e13.2), (13.4e13.6 Ma). 6. spp. (12.8e13.6 Ma). 7. Acrecebus (6e9 Ma; Kay and Cozzuol, 2006). 8. Solimoea (6e9 Ma; Kay and Cozzuol, 2006). B. Simplified arrangement of the extant families of Platyrrhini. The crown group is all species living and extinct that can trace their ancestry to the most recent common ancestor of living platyrrhines. Any extinct taxon, such as Dolichocebus, that split off before the last common ancestor of the platyrrhine living families is a stem platyrrhine. have among the narrowest interorbital regions among living plat- models produce somewhat different results, such as multigene yrrhines (Figs. IV-115, 116 in Hershkovitz, 1977). Bayesian methods vs. local molecular clock approaches. For example, with the same molecular data set and the same under- lying phylogeny Opazo et al. (2006) date the Saimiri-Cebus split at Split times and molecular clocks 16.15 Ma using a multigene Bayesian method, but 17.3 Ma using a local molecular clock method. Third, and less appreciated, these As he did in 1979, Rosenberger states explicitly the hypothesis methods are extremely sensitive to the assumed underlying that Dolichocebus is the sister taxon to Saimiri and more broadly phylogeny. Again examining the Opazo et al. study, when the within the clade of Saimiri and Cebus (Rosenberger, 1979)2. For his underlying phylogeny is assumed to be (, Pith- hypothesis to be correct, the Saimiri lineage (e.g., after the split eciidae), Opazo et al. find a Saimiri-Cebus split time that is from 8% between Saimiri and Cebus) must already exist before 20.5 Ma to 24% more ancient than when they assume a phylogeny of Pith- (based on our best estimate of its geologic age; Kay et al., 2008). eciidae (Atelidae, Cebidae). In short, knowing the underlying Rosenberger claims that this split time is consistent with recent phylogeny is critical to assessing the reliability of the result. molecular clock estimates for the branch time between Saimiri and Fortunately, molecular phylogenetic studies using molecular Cebus. In fact the results of molecular clock studies are far less sequence data and Alus have now reached consensus about the precise than Rosenberger claims but largely reject a >20.5 Ma split underlying phylogenetic branching pattern for living platyrrhines time. Over and above the requirement that molecular clocks run on (Fig. 2). Three families are recognized with splitting first time, clock calibration is limited by the following: First, the cali- from the last common ancestor with a subsequent split between bration of the clock requires fossil evidence, which itself often is Atelidae and Cebidae, i.e., P (A, C); Fig. 2B. Filtering those studies to disputed although recent methodological advances, which allow use the P (A, C) arrangement, we get the following: 14.3 Ma multiple calibration dates and ranges of dates have ameliorated this (Bayesian credibility interval 12.6 to 17.5 Ma; Hodgson et al., 2009), problem somewhat (Yoder and Yang, 2000). Second, different 16.125 (no confidence interval [CI] reported), and 17.63 Ma (Bayesian credibility interval 15.8 to 19.43 Ma; Opazo et al., 2006). In short, multiple estimates of the split time between Saimiri and 1 Rosenberger (1979) cites Falk (pers. comm.) that the sylvian fissure is deeply grooved as in Saimiri endocasts, but admits that its taxonomic distribution is Cebus that use the most corroborated family-level phylogeny yield unknown. Hershkovitz (1977) illustrates a sylvan fissure in all extant platyrrhines. It dates that are too recent for Dolichocebus to be the sister group of is less deeply fissured in some of the smaller taxa. See also photographs of plat- Saimiri even when the Bayesian credibility interval is included. yrrhine brains at Comparative Mammalian Brain Collection (http://brainmuseum. While we caution against overreliance on molecular clocks, org). ’ 2 This is a departure from his more cautious statement of 1991 that Dolichocebus Rosenberger s (2010) claim that these clocks support a Long might alternatively be sister to a Saimiri-Cebus clade. Rosenberger et al. (1990: 226) Lineage Hypothesis is dubious at best. allow that Saimiri and Cebus may be sister taxa to the exclusion of Dolichocebus because living cebines share a “squared anterior jaw, highly dimorphic canines and Long, medium, and short lineages broad anterior premolars.” To this list might be added a number of other synapo- morphies that Saimiri and Cebus share to the exclusion of Dolichocebus, which has “ ” the stem platyrrhine condition, that is, expanded brain size and foramen magnum In his history of the Long Lineage Hypothesis Rosenberger positioned anteriorly. notes its corroboration by numerous fossil platyrrhines from 222 R.F. Kay, J.G. Fleagle / Journal of Human Evolution 59 (2010) 218e222 younger time periods, specifically Proteropithecia,aw15.7 Ma Lockwood, C.A., Fleagle, J.G., 1999. The recognition and evaluation of homoplasy in e pitheciine from northern Patagonia (Kay et al., 1998, 1999), the primate and human evolution. Yrbk. Phys. Anthropol. 42, 189 232. Madden, R.H., Guerrero, J., Kay, R.F., Flynn, J.J., Swisher III, C.C., Walton, A.H., 1997. many taxa from the Colombian locality of (Fleagle et al., The stage and Laventan age: new chronostratigraphic and geochro- 1997), and an isolated late Miocene locality in western (Kay nologic units for the Miocene of South America. In: Kay, R.F., Madden, R.H., and Cozzuol, 2006; see Fig. 2 for dates). We want to emphasize Cifelli, R.L., Flynn, J.J. (Eds.), Vertebrate Paleontology in the Neotropics. Smith- sonian Institution Press, Washington, D.C., pp. 499e519. that we agree with Rosenberger that many modern platyrrhine Meldrum, D.J., Kay, R.F., 1997. rubricae, a new pitheciin seed predator clades were clearly present in the Amazon/Orinoco basin by from the Miocene of . Am. J. Phys. Anthropol. 102, 407e427. 13.5 Ma (Fig. 2). Our disagreement is whether the extant clades can Opazo, J.C., Wildman, D.E., Prychitko, T., Johnson, R.M., Goodman, M., 2006. fi Phylogenetic relationships and divergence times among New World monkeys be identi ed in the earlier deposits of southernmost Argentina, viz: (Platyrrhini, ). Mol. Phylogenet. Evol. 40, 274e280. Dolichocebus and other Southern Patagonian taxa. As new fossil Perry, J.M.G., Kay, R.F., Colbert, M.W., 2006. Tooth root size and chewing muscle material becomes available for Homunculus, Chilicebus, and other leverage in Homunculus, a Miocene Primate from Patagonia. J. Vert. Paleontol. 26, 110A. early fossil primates in South America, we look forward to further Perry, J.M.G., Kay, R.F., Vizcaino, S.F., Bargo, M.S., Tooth root size, chewing muscle investigations and debates about the mode and tempo of platyr- leverage, and the biology of Homunculus patagonicus (Primates), from the rhine evolution. Miocene of Patagonia. Ameghiniana (in press). Rosenberger, A.L., 1979. Cranial anatomy and implications of Dolichocebus, a late Oligocene ceboid primate. Science 279, 416e418. References Rosenberger, A.L., 2002. Platyrrhine paleontology and systematics: the paradigm shifts. In: Hartwig, W.C. (Ed.), The Primate Fossil Record. Cambridge University Baron, G., Frahm, H.D., Bhatnagar, K.P., Stephan, H., 1983. Comparison of brain Press, Cambridge, pp. 151e159. structure volumes in Insectivora and Primates. III. Main olfactory bulb (MOB). J. Rosenberger, A.L., 2010. Platyrrhines, PAUP, parallelism, and the Long Lineage Hirnforsch. 24, 551e558. Hypothesis: a reply to Kay et al. (2008). J. Hum. Evol. 59 (2), 214e217. Collard, M., Wood, B.A., 2000. How reliable are human phylogenetic hypotheses? Rosenberger, A.L., Setoguchi, T., Shigehara, N., 1990. The fossil record of callitrichine Proc. Natl. Acad. Sci. USA 97, 5003e5006. primates. J. Hum. Evol. 19, 209e236. Collard, M., Wood, B.A., 2001. Homoplasy and the early hominid masticatory Ross, C., Williams, B.A., Kay, R.F., 1998. Phylogenetic analysis of anthropoid rela- system: inferences from analyses of extant hominoids and papionins. J. Hum. tionships. J. Hum. Evol. 35, 221e306. Evol. 41, 167e194. Sanchez-Villagra, M., Williams, B.A., 1998. Levels of homoplasy in the evolution of Fleagle, J.G., Kay, R.F., Anthony, M.R.L., 1997. Fossil New World monkeys. In: Kay, R.F., the mammalian skeleton. J. . Evol. 5, 113e126. Madden, R.H., Cifelli, R.L., Flynn, J.J. (Eds.), Mammalian Evolution in the Sanderson, M.J., Donoghue, M.J., 1989. Patterns of variation in levels of of Homo- Neotropics. Smithsonian Institution Press, Washington, D.C., pp. 473e495. plasy. Evolution 43, 1781e1895. Flynn, J.J., Guerrero, J., Swisher III, C.C., 1997. Geochronology of the Honda Group. In: Sanderson, M.J., Donoghue, M.J., 1996. The relationship between homoplasy and Kay, R.F., Madden, R.H., Cifelli, R.L., Flynn, J.J. (Eds.), Vertebrate Paleontology in confidence in a phylogenetic tree. In: Sanderson, M.J., Hufford, L. (Eds.), the Neotropics. Smithsonian Institution Press, Washington, D.C., pp. 44e60. Homoplasy: The Recurrence of Similarity in Evolution. Academic Press, San Gradstein, F.M., Ogg, J.G., Smith, A., 2004. A Geological Time Scale. Cambridge Diego, pp. 67e89. University Press, Cambridge, UK. Schrago, C.G., 2007. On the time scale of New World primate diversification. Am. J. Hershkovitz, P., 1977. Living New World Monkeys (Platyrrhini) with an Introduction Phys. Anthropol. 132, 344e354. to Primates, vol. 1. University of Chicago Press, Chicago. Stephan, H., Frahm, H.D., Baron, G., 1981. New and revised data on volumes of brain Hodgson, J.A., Sterner, K.N., Matthews, L.J., Burrell, A.S., Jani, R.A., Raaum, R.L., structures in insectivores and primates. Folia Primatol. 35, 1e29. Stewart, C.B., Disotell, T.R., 2009. Successive radiations, not stasis, in the South Takai, M., Anaya, F., Suzuki, H., Shigehara, N., Setoguchi, T., 2001. A new platyrrhine American primate fauna. Proc. Natl. Acad. Sci. USA 106, 5534e5539. from the middle Miocene of La Venta, Colombia, and the phyletic position of Kay, R.F., Cozzuol, M.A., 2006. New platyrrhine monkeys from the Solimões Callicebinae. Anthropol. Sci. (Japan) 109, 289e307. Formation (late Miocene, Acre State, Brazil). J. Hum. Evol. 50, 673e686. Wildman, D.E., Jameson, N.M., Opazo, J.C., Yi, S.V., 2009. A fully resolved level Kay, R.F., Fleage, J.G., Mitchell, T.R.T., Colbert, M.W., Bown, T.M., Powers, D.W., 2008. phylogeny of neotropical primates (Platyrrhini). Mol. Phylogenet. Evol. 53, The anatomy of Dolichocebus gaimanensis, a primitive platyrrhine monkey from 694e702. Argentina. J. Hum. Evol. 54, 323e382. Williams, B.A., Kay, R.F., 1995. The taxon Anthropoidea and the crown clade concept. Kay, R.F., Johnson, D.J., Meldrum, D.J., 1998. A new pitheciin primate from the Evol. Anthropol. 3, 188e190. middle Miocene of Argentina. Am. J. Primatol. 45, 317e336. Williams, B.A., Kay, R.F., Kirk, E.C., 2010. New perspectives on anthropoid origins. Kay, R.F., Johnson, D.J., Meldrum, D.J., 1999. Proteropithecia, new name for Propi- Proc. Natl. Acad. Sci. 107, 4797e4804. thecia Kay, Johnson and Meldrum, 1998 non Vojnits 1985. American Journal of Yoder, A.D., Yang, Z., 2000. Estimation of primate speciation dates using local Primatology 47, 347. molecular clocks. Mol. Biol. Evol. 17, 1081e1090. Author's personal copy

Journal of Human Evolution 59 (2010) 214e217

Contents lists available at ScienceDirect

Journal of Human Evolution

journal homepage: www.elsevier.com/locate/jhevol

News and Views Platyrrhines, PAUP, parallelism, and the Long Lineage Hypothesis: A reply to Kay et al. (2008)

Alfred L. Rosenberger

Department of Anthropology and Archaeology, Brooklyn College, CUNY, 2900 Bedford Ave., Brooklyn, NY 11210, USA City University of New York, Graduate Center, NY, USA New York Consortium in Evolutionary Primatology (NYCEP), NY, USA Department of Mammalogy, American Museum of Natural History, NY, USA

article info modern platyrrhine genera, Saimiri, Alouatta, and later Aotus could have arisen within or earlier than the approximately 11e20 Ma Article history: time bracket while a fourth, Cebus, the living sister-taxon of Saimiri, Received 1 October 2008 was inferred to be equally as old; 2) generic distinctions between Accepted 23 September 2009 the La Venta fossils Stirtonia and Alouatta and the modern Neo-

Keywords: saimiri and Saimiri, respectively, were questionable; 3) within- Fossils lineage morphological continuity between these sets was evidence Cladistics for stasis; and, 4) modern NWM differentiation has deeper New World monkeys temporal roots than comparable catarrhine splits. Two Patagonian Primates fossils, Tremacebus and Dolichocebus, at about 20 Ma, were identi- Systematics fied as the oldest affiliates of the modern taxa, related to Aotus and Saimiri, respectively; the broader affinities of Dolichocebus as a cebine (hence the Cebus-Saimiri clade) were also explicitly dis- cussed (e.g., Rosenberger et al., 1990). Kay et al. (2008) argued Kay et al. (2008) presented a parsimony (PAUP) analysis of platyr- instead that Tremacebus and Dolichocebus were part of a four-genus, “ ” rhines involving a craniodental database of 268 characters, several southern stem platyrrhine group, proving that NWM evolution “ ” fi targeted Patagonian fossils, Tarsius, and 29 anthropoid genera. Their unfolded in a more layered fashion. A fth Argentine genus, core conclusion challenges a central idea about platyrrhine evolu- Homunculus, was held to be part of this group but it was not tion, but it is flawed by interlocking problems concerning research analyzed (see below), while the older Bolivian fossil, Branisella, fell design, data quality, and methodology. This conclusion rests on outside it. a weakly supported cladistic artifact, a purported monophyletic Although Kay et al. (2008) make no mention that the LLH has group of southern platyrrhines involving four poorly known, been corroborated independently by distinctly different types of unevenly preserved, barely comparable sets of fossils that emerged data and varied methods of analysis since it was proposed, by default, juxtaposed against a grouping of essentially modern numerous molecular studies (e.g., references in Opazo et al., 2006) fi platyrrhine genera, 94% of which are living forms represented by have con rmed the majority of the underlying cladistic hypoth- effectively complete datasets. To rationalize these results, Kay et al. eses (e.g., Schneider and Rosenberger, 1997; Schneider et al., 2001; (2008) oddly misconstrue the nature of phylogenetic, functional, Rosenberger, 2002) and uniformly support the early differentia- and adaptive evidence in historical evolutionary reconstruction, tion model. The median ages (in millions of years) of the diver- ’ using false analogies. And, their argument ignores the body of gence of critical genera and subclades calculated from Schrago s evidence supporting the evolutionary model they seek to refute, the (2007) summary of six projects gives the following: Aotus, 17.6; Long Lineage Hypothesis (LLH). Callicebus, 15.4; Alouatta, 13.7; Callimico, 9.7; Cebus and Saimiri, The LLH proposes that modern New World monkeys (NWM) are 16.5; cebines vs. callitrichines, 19.9; cebids vs. atelids, 20.1. A characterized by a relatively large number of long lived genera and separate study (Opazo et al., 2006), for reference, yields even older subclades (e.g., Rosenberger, 1979, 1992, 2002; Delson and divergence dates: Aotus, 22; Callicebus, 19.3; Alouatta, 16.8; Calli- Rosenberger, 1984; Rosenberger et al., 2009). Predicated on an mico, 12.1; Cebus vs. Saimiri, 19.5; cebines vs. callitrichines, 22.8; ecophylogenetic study of all platyrrhines known up until 1979 and cebids vs. atelids, 24.4. Another recent molecular study (Hodgson “ a major reorganization of NWM systematics, its central hypotheses et al., 2009) inveighed against the LLH, saying The MRCA [most sought to establish: 1) direct fossil evidence that three of the 16 recent common ancestor] of the living platyrrhines is estimated to have lived 19.5 Ma (95% credibility interval 16.8e23.4 Ma).” and specifying that Dolichocebus and Tremacebus thus cannot be E-mail address: [email protected]

0047-2484/$ e see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.jhevol.2010.03.003 Author's personal copy

A.L. Rosenberger / Journal of Human Evolution 59 (2010) 214e217 215 related closely to modern genera. But the estimated MRCA date is primitive and/or prone to analogy. Finally, none of the four char- only 0.5 million years prior to the 20 Ma geologically estimated acters defining the “stem group” are observable in one of its four dates for the fossils, which falls well within the wide margins of genera, Tremacebus, which is known only by a basically toothless their confidence limits, and their 14.3 Ma bottom boundary for skull. differentiation of all the major platyrrhine lineages actually A variety of researchers working on disparate primate groups, conforms with the LLH model. living and extinctdplesiadapiforms, platyrrhines, catar- The pre-Pleistocene fossil record, now comprising over twice as rhinesdhave shown that the manner in which a PAUP-based study many genera as in 1979, consistently fits LLH predictions. Three is implemented has profound consequences, and may often lead to examples: 1) The case for fossil-modern congeners was first unreliable results (e.g., Collard and Wood, 2000; Young, 2005; strengthened in 1987 by discovery of a new large-eyed species at Sargis, 2007; Silcox, 2007; Matthews and Rosenberger, 2008). In La Venta, Colombia, Aotus dindensis (Setoguchi and Rosenberger, this case, there is no way of objectively quantifying the trustwor- 1987), an attribution corroborated in 2001 by new material, also thiness of the “stem group” hypothesis, specifically, within the allocated to A. dindensis (Takai et al., 2001); 2) An implicit context of the supertree. We can only assess overall tree quality by prediction of within-lineage transformational continuity has been metrics like the Consistency Index (CI), which is 0.343 (Kay et al., confirmed by a detailed micro-CT study (Kay et al., 2004), which 2008). Technically, this means the average agreement of each shows again that Tremacebus shares derived cranial features with individual character state with the tree as a whole is 34%, that is, Aotus pertaining to nocturnality (Fleagle and Rosenberger, 1983), two-thirds of the information is spurious. Biologically, it is impos- but in more primitive states than in the living owl monkey (e.g., sible to determine where in the tree the “good characters” reside, smaller orbits and larger olfactory lobes, in both cases falling but they are likely to be concentrated among the essentially between the diurnal Callicebus and Aotus in their proportions); 3) complete datasets of the modern taxa (plus Proteropithecia, well The recently discovered Killikaike (Tejedor et al., 2006), at 16.4 Ma, represented by a morphology that is strikingly similar to modern presents a complete undistorted face that is diagnostically cebi- pitheciins). So, some or all of the 34% “positive” results apportioned nedvaulted anterior braincase, large forebrain, narrow inter- to the “stem group” is likely due to methodological artifact, error, orbitum, closely appressed medial orbital wallsdall derived and chance. The CI is not an empirically proven metric, and there is cebine features securely extending the subfamily’s time depth 3e5 no agreement as to the biological importance of a CI value of millions of years beyond Neosaimiri. Indeed, since 1979, 17 new whatever magnitude (unlike Coefficient of Variation). But calcu- fossil genera have been named (see Hartwig, 2002). Six fossils lated on a 0.0e1.0 scale, a 0.343 would rate the score of this study at from mainland South America have been aligned with modern the bottom third of its own internal quality control measure. Does generic lineages and/or low-level clades without evident debate: this number represent an adequate level of scientific comfort, Caipora, Acrecebus, Nuciruptor, Patasola, Miocallicebus, Proter- credibility, and predictability? opithecia. The oldest among these is Proteropithecia,a16Ma The reliability of parsimony-based cladograms built from pitheciin from Patagonia (Kay et al., 1998), that is, within the morphological information also has a chequered history with temporospatial range of the Kay et al. (2008) southern “stem regard to modern platyrrhines, and this should be taken into group,” thus a strong indicator of historical/faunal continuity account when evaluating questionable outcomes regarding data- between the south and modern Amazonia in the north. The rela- poor fossils. A case in point is Aotus, a genus whose cladistic linkage tionships of three others, Killikaike, Soriacebus, and Paralouatta, has appeared in fundamentally different positions with nearly have elicited different opinions (e.g., MacPhee and Horovitz, 2002; every such study that has been conducted. For example, in Kay et al., 2008), but I regard them as securely cebine, pitheciin, a comprehensive craniodental and postcranial study, Ford (1986) and alouattin, respectively. Others, more poorly known, remain attached Aotus to Callicebus but linked both with Saimiri (or Sai- difficult to interpret and I ignore them here, taking a conservative miri and Cebus) as the first- or second-branching basal lineage. Kay stance. (1990), in a large study of over 100 dental characters, placed Aotus A plethora of issues cast doubt on the reliability of Kay et al.’s in a triple-junction polytomy which included atelines and calli- (2008) “stem group” hypothesis. Its central tenet, monophyly of trichines, nested deep within the NWM cladogram. Kay (1994),in Dolichocebus, Tremacebus, Carlocebus, and Soriacebus, is weakly a nine-genus study focusing on the La Venta fossil Lagonimico, supported. It is based on a proportionately small number of concluded that Aotus was the sister-taxon of Callicebus and Pithecia. features, including states that would appear rather trivial for higher Horovitz (1999), in another comprehensive study, found Aotus to be phylogeny studies. Of the final 183 dental features used, only four the sister-taxon of Cebidae sensu Rosenberger when based on synapomorphies support this group. Taken verbatim, they are: 1) morphology alone. And, Kay et al. (2008:347), when performing distolingually oriented P4 protocristid; 2) mesiodistally reduced P4 their maximum-parsimony analysis of “.the cranial and dental 4 talonid; 3) basally inflated P4; and, 4) strongly developed P matrix unconstrained by a molecular “backbone”.” found that hypocones. How much phylogenetic power should we attribute to Callicebus and Aotus were sister-groups, linked first with Cebus and this series? How much developmental and taxonomic indepen- then with a monophyletic group comprised of pitheciins and dence might they have? How confident should we be that these atelines. characters properly reflect phylogeny in this case? PAUP is programmed to produce (read “force”) a result for each These traits represent one tooth set, P4. Three are descriptive taxon in the matrix. There are no fail-safe measures akin to prob- attributes of one unit from that set, that is, they particularize P4 ability statistics, or even empirically derived rules of thumb, to shape. The first (#1) is found elsewhere only in Tarsius, Apidium, indicate when a specific result falls below a threshold of being and Simonsius, where their radically different crown morphologies realistic; or, what constitutes an adequate sampling of traits needed raise questions about character delineation and coding. Characters to characterize an individual taxon or large assemblage of taxa. #2 and #4 are complex homology and polarity problems involving Even when fossils are notoriously depleted of data PAUP finds multi-state characters coded as four and three states, respectively; a solution, but is it artifact or phylogenetic? Of 286 characters used none of the four “stem platyrrhine” genera mutually share any one by Kay et al. (2008: Appendix 4), the total proportion of missing of the #2 states. Character #3 is widely distributed within NWM data for the fossils is large: Dolichocebus, 20%; Soriacebus, 40%; and in the out-groups, occurring in five living platyrrhines plus Carlocebus, 54%; Tremacebus, 73%. By contrast, the average Proteropithecia and among four Fayum genera, which suggests it is percentage of missing data in toto for the living NWM genera is Author's personal copy

216 A.L. Rosenberger / Journal of Human Evolution 59 (2010) 214e217

6.6%. The breakdown for fossils is dramatic for cranial (Dolichoce- opposition to the “robust” pattern of atelids) is ignored because the bus, 46%; Soriacebus, 99%; Carlocebus, 99%; Tremacebus, 40%) and zygomatic arches are gone. But the gracile pattern is a combination dental data as well (Dolichocebus, 13%; Soriacebus, 13%; Carlocebus, of features involving a flat glenoid fossae, small postglenoid 33%; Tremacebus, 89%). Actually, these figures present an incom- process, shallow temporal fossa, and unremarkable temporal lines. plete picture of the nature of the evidence since Kay et al. (2008) This is a combination eminently observable in Dolichocebus and did not eliminate redundancies from the input matrix or from the found elsewhere only in cebines and callitrichines (Kay et al., 2008). results. For example, of the six derived features said to link the Kay et al. (2008) concur with Rosenberger and colleagues on these “stem group” with crown platyrrhines (p. 356), the mesial termi- descriptions, except for the temporal fossae, claiming that the nation point of the cristid obliqua on M1 and M2 is counted as two latter’s size is measured against the zygomatic arch, which is characters, and the root structure of P3 and P4 is counted as another missing, but ignoring that the fossa can also be assessed as two. Surely these traits are better treated as a pair of serial a matter of relative postorbital constriction. Kay et al. (2008:348) homologs (two traits) rather than four independent characters. explain away these features as commonly occurring or body-size Still, the overarching data quality issues are these: Dolichocebus and relateddwhich is precisely our point: they overlap with cebids, Tremacebus are essentially edentulous skulls (a few referred teeth a monophyletic, relatively small-bodied group. Overall, Kay are attributed to the former) with a limited range of well preserved et al.’s (2008) re-description and clarification of Dolichocebus comparable parts or characters; Soriacebus and Carlocebus are morphology utilizing modern technological capabilities is valuable, known from teeth, almost all representing lower jaws, without any but their insistence on treating each observable anatomical trait or cranial material; Carlocebus lacks anterior teeth, which are often term as a zero-sum data game obscures the reality that character cladistically diagnostic among NWM, and it preserves precious states evolve inter-connectedly as patterns. This undercuts the little postcanine occlusal morphology. Given these limitations, it validity of observation-based morphological inference as a neces- should come as no surprise that PAUP anchored the whole tree by sary paleontological research tool. the multiplying effect of the modern taxa with their rich informa- In conclusion, Kay et al. (2008) do not provide compelling tion content, segregating them from the depauperate fossils by the evidence for a monophyletic southern group of “stem platyrrhines” sheer weight of accessible crown-group synapomorphies that radiated separately from the modern forms and their other substantiating its many internal nodes. fossil relatives. That specific result is based on only four dental These skewed results partly arise from the fact that the taxo- character statesdand none go without question. All four appear to nomic context of the study is grossly overspecified. Kay et al. be correlated, non-independent traits of a single tooth set, which (2008:329) curiously exclude the iconic Homunculus from the denigrates their statistical power, and none of them are preserved analysis “.because some or all of the pertinent material is unde- on the best specimens under study, the edentulous (type specimen) scribed,” while criticizing Tejedor et al. (2006), on the very same skulls of Dolichocebus and Tremacebus. Nor do Kay et al. (2008) offer page, for not making explicit comparisons of the Patagonian even a single comparative anatomical datum or argu- Killikaike blakei with the “.abundant [my italics] comparative mentdindependent of the treedthat explains the counterintuitive material of Homunculus patagonicus.” This self imposed embargo sequelae of their hypothesis (see Rosenberger, 2002): for example, ostensibly relates to new material under study (Kay et al., 2005) but that parallelism is a better explanation than a transformational does not adequately explain why Homunculus could not be repre- vector behind the minor adaptive differences separating the non- sented in the matrix by the two other well prepared and published diurnal Tremacebus and Aotus; or, that parallelism is a better partial, undistorted skulls (e.g., Fleagle and Rosenberger, 1983; explanation than ecophylogenetic affinity for the extensive, derived Tauber, 1991; Tejedor and Rosenberger, 2008) and the long seed-harvesting and -feeding similarities shared by Soriacebus and known dental and mandibular remains, which include many pitheciins in the incisors, canines, and anterior premolars (and specimens in better condition than the dentitions of Carlocebus. likely jaws); or, that ecological parallelism writ large of the Pata- Two additional southern cone fossils were also excluded, Killikaike gonian primate fauna is a better explanation than their primitive- and the older Chilecebus (Flynn et al., 1995). The latter, at ness and older age relative to the anatomically more modern and a minimum, provides more occlusal morphology than any Pata- younger La Venta community. Failing to acknowledge that not all gonian fossil NWM other than Killikaike. homoplasy is created equal, Kay et al. (2008) instead present Kay et al. (2008) similarly seek to dismiss or reject cranial a specious plausibility argument by asserting that the parallelisms information supporting an alternative view of Dolichocebus. The emerging from their study are as likely to occur as other hypoth- most complete statement in this regard, albeit with redundancies esized cases of parallelism among NWM, such as the independent and flaws, is from Rosenberger et al. (1990), which proposes evolution of prehensile tails in Cebus and atelines, which now features specifically linking the genus to Saimiri and, more broadly, appears to be a universally accepted explanation. Left unstated is to cebines. While the most dramatic character, the interorbital that an extensive analysis stands behind that hypothesis fenestra, Kay et al. (2008) fairly hold in abeyance because its status (Rosenberger, 1983), testing the pros and cons of parallelism by seems ambiguous, other morphological observations or inferences morphologically and functionally deconstructing components of are negated without due consideration to established norms of the prehensile tail as a system, from bone to brain to allometric paleontological research and comparative anatomy. The presence scaling to the biological roles of tail use as informed by fieldwork, in Dolichocebus of a vaulted frontal region is dismissed because and placing this assessment in what was then a new phylogenetic glabella (a poorly chosen term on the part of the Rosenberger et al. context, which has also been confirmed independently. [1990] description) is broken away, but the prominent olfactory bulbs and high arc of the frontal lobedthe underlying morphology Acknowledgements of import, as stateddare well preserved and clearly show the anterior braincase had a cebine-like, domed shape rather than the Several colleagues and reviewers provided very helpful flat frontal region which is primitive among NWM. The narrowness comments on the manuscript, for which I am grateful. The research of the interorbital pillar is likewise discounted although the cir- was partially supported by grants from PSC-CUNY and the Tow cumorbital, intra-orbital, and forebrain morphology can only be Travel Research Fund, Brooklyn College. I thank the authorities at interpreted as being cebine-like (i.e., narrow). The inferred light- the American Museum of Natural History and the United States ness of the masticatory apparatus (describing a cebid complex, in National Museum for access to collections, as always. Author's personal copy

A.L. Rosenberger / Journal of Human Evolution 59 (2010) 214e217 217

References Rosenberger, A.L., 1983. Tail of tails: parallelism and prehensility. Am. J. Phys. Anthropol. 60, 103e107. Rosenberger, A.L., 1992. Evolution of feeding niches in New World monkeys. Am. J. Collard, M.C., Wood, B., 2000. How reliable are human phylogenetic hypotheses? Phys. Anthropol. 88 (4), 525e562. Proc. Natl. Acad. Sci. USA 97, 5003e5006. Rosenberger, A.L., 2002. Platyrrhine paleontology and systematics: the paradigm Delson, E., Rosenberger, A.L., 1984. Are there any anthropoid primate “living shifts. In: Hartwig, W.C. (Ed.), The Primate Fossil Record. Cambridge University fossils”? In: Eldredge, N., Stanley, S. (Eds.), Casebook on Living Fossils. Springer- Press, Cambridge, pp. 151e160. Verlag, New York, pp. 50e61. Rosenberger, A.L., Setoguchi, T., Shigehara, N., 1990. The fossil record of callitrichine Fleagle, J.G., Rosenberger, A.L.,1983. Cranial morphology of the earliest anthropoids. In: primates. J. Hum. Evol. 19, 209e236. Sakka, M. (Ed.), Morphologie, Evolutive, Morphogenese du Crane et Anthro- Rosenberger, A.L., Tejedor, M.F., Cooke, S., Pekar, S., 2009. Platyrrhine ecophyloge- pogenese. Centre National de la Recerche Scientifique, Paris, pp. 141e153. netics, past and present. In: Garber, P., Estrada, A., Bicca-Marques, J.C., Flynn, J.J., Wyss, A.R., Charrier, R., Swisher, C.C., 1995. An early Miocene anthropoid Heymann, E.W., Strier, K.B. (Eds.), South American Primates: Comparative skull from the Chilean Andes. Nature 373, 603e607. Perspectives in the Study of Behavior, Ecology and Conservation. Springer, New Ford, S.M., 1986. Systematics of the New World monkeys. In: Swindler, D. (Ed.), York, pp. 69e113. Comparatve Primate Biology. Systematics, Evolution and Anatomy, vol. 1. Alan Sargis, E.J., 2007. The postcranial morphology of Ptilocercus lowii (Scandentia, Liss, New York, pp. 73e135. Tupaiidae) and its implications for primate supraordinal relationships. In: Hartwig, W.C., 2002. The Primate Fossil Record. Cambridge University Press, Ravosa, M.J., Dagosto, M. (Eds.), Primate Origins: Adaptation and Evolution. Cambridge. Springer, New York, pp. 51e82. Hodgson, J.A., Sterner, K.N., Matthews, L.J., Burrell, A., Jani, R.A., Raaum, R.L., Schneider, H., Rosenberger, A.L., 1997. Molecules, morphology and platyrrhine Stewart, C.-B., Distotell, T.R., 2009. Successive radiations, not stasis, in the South systematics. In: Norconk, M.A., Rosenberger, A.L., Garber, P.A. (Eds.), Adaptive American primate fauna. Proc. Natl. Acad. Sci. USA 106, 5534e5539. Radiations of Neotropical Primates. Plenum Press, New York, pp. 3e19. Horovitz, I., 1999. A phylogenetic study of living and fossil platyrrhines. Am. Mus. Schneider, H., Canavez, F.C., Sampaio, I., Moreira, M.A.M., Tagliaro, C.H., Seuanez, H. Novit. 3269, 1e40. N., 2001. Can molecular data place each Neotropical monkey in its own branch? Kay, R.F., 1990. The phyletic relationships of extant and fossil (Platyr- Chromosoma 109, 515e523. rhini, Anthropoidea). J. Hum. Evol. 19, 175e208. Schrago, C.G., 2007. On the time scale of new world primate diversification. Am. J. Kay, R.F., 1994. “Giant” from the Miocene of Colombia. Am. J. Phys. Phys. Anthropol. 132 (3), 344e354. Anthropol. 95, 333e353. Setoguchi, T., Rosenberger, A.L., 1987. A fossil owl monkey from La Venta, Colombia. Kay, R.F., Campbell, V., Rossie, J.B., Colbert, M.W., Rowe, T.B., 2004. Olfactory fossa of Nature 326, 692e694. Tremacebus harringtoni (Platyrrhini, early Miocene, Sacanana, Argentina): Silcox, M.T., 2007. Primate , plesiadapiforms, and approaches to primate implications for activity pattern. Anat. Rec. 281A, 1157e1172. origins. In: Ravosa, M.J., Dagosto, M. (Eds.), Primate Origins: Adaptation and Kay, R.F., Fleagle, J.G., Mitchell, T.R.T., Colbert, M., Bown, T., Powers, D.W., 2008. The Evolution. Springer, New York, pp. 143e178. anatomy of Dolichocebus gaimaneensis, a stem platyrrhine monkey from Takai, M., Anaya, F., Suzuki, H., Shigehara, N., Setoguchi, T., 2001. A new plat- Argentina. J. Hum. Evol. 54 (3), 323e382. yrrhine from the Middle Miocene of La Venta, Colombia, and the phyletic Kay, R.F., Johnson, D., Meldrum, D.J., 1998. A new pitheciin primate from the middle position of Callicebinae. Anthropological Sci.: J. Anthropol. Soc. Nippon 109 Miocene of Argentina. Am. J. Primatol. 45, 317e336. (4), 289e307. Kay, R.F., Vizcaino, S.F., Tauber, A.A., Bargo, M.S., Williams, B.A., Luna, C., Colbert, M. Tauber, A.A., 1991. Homunculus patagonicus Ameghino 1891 (Primates, Ceboidea), W., 2005. Three newly discovered skulls of Homunculus patagonicus support its Mioceno temprano de la costa atlántica austral, Provincia de Santa Cruz, position as a stem platyrrhine and establish its diurnal arboreal folivorous República Argentina. Acad. Nac. de Cienc. 82, 1e32. habits. Am. J. Phys. Anthropol. 40 (Suppl.), 127. Tejedor, M.F., Rosenberger, A.L., 2008. A neotype for Homunculus patagonicus MacPhee, R.D.E., Horovitz, I., 2002. Extinct Quaternary platyrrhines of the Greater Ameghino, 1891, and a New Interpretation of the Taxon. PaleoAnthropology Antilles and Brazil. In: Hartwig, W.C. (Ed.), The Primate Fossil Record. 2008, 68e82. Cambridge University Press, Cambridge, pp. 189e200. Tejedor, M.F., Tauber, A.A., Rosenberger, A.L., Swisher III, C.C., Palacios, M.E., 2006. Matthews, L.J., Rosenberger, A.L., 2008. Taxonomic status of the Yellow tailed New primate genus from the Miocene of Argentina. Proc. Natl. Acad. Sci. USA Woolly Monkey, Lagothrix flavicauda e an object lesson in the implementation 103 (14), 5437e5441. of PAUP*. Am. J. Phys. Anthropol. 137, 245e255. Young, N.M., 2005. Estimating hominoid phylogeny from morphological data: Opazo, J.C., Wildman, D.E., Prychitko, T., Johnson, R.M., Goodman, M., 2006. character choice, phylogenetic signal and postcranial data. In: Lieberman, D.E., Phylogenetic relationships and divergence times among New World monkeys Smith, R.J., Kelley, J. (Eds.), Interpreting the Past: Essays on Human, Primate and (Platyrrhini, Primates). Mol. Phylogenet. Evol. 40 (1), 274e280. Mammal Evolution in Honor of David Pilbeam. Brill Academic Publishers, Inc., Rosenberger, A.L., 1979. Cranial anatomy and implications of Dolichocebus, a late Boston, pp. 19e31. Oligocene ceboid primate. Nature 279 (5712), 416e418.