Hominin homoiology: An assessment of the impact of phenotypic plasticity on phylogenetic analyses of humans and their fossil relatives

https://doi.org/10.1016/j.jhevol.2006.11.018Get rights and content

Abstract

Homoiologies are phylogenetically misleading morphological similarities that are due to nongenetic factors. It has been claimed that homoiologies are common in the hominin skull, especially in regions affected by masticatory strain, and that their prevalence is one reason why reconstructing hominin phylogenetic relationships is difficult. To evaluate this “homoiology hypothesis,” we performed analyses on a group of extant primates for which a robust molecular phylogeny is available—the hominoids. We compiled a data set from measurements that developmental considerations and experimental evidence suggest differ in their likelihood of exhibiting masticatory-strain-induced phenotypic plasticity. We then used the coefficient of variation and t-tests to evaluate the phenotypic plasticity of the measurements. We predicted that, if the hypothesis is correct, the measurements of skeletal features that do not remodel and therefore are unaffected by phenotypic plasticity should be less variable than the measurements of skeletal features that remodel and are subject to low-to-moderate strains, and that the latter should be less variable than the measurements of skeletal features that remodel and are subject to moderate-to-high strains. Subsequently, we performed phylogenetic analyses on character state data derived from the measurements and compared the resulting phylogenetic hypotheses to the consensus molecular phylogeny for the hominoids. We predicted that, if the hypothesis is correct, agreement between the phylogenies should be best for the non-phenotypically-plastic characters, intermediate for the low-to-moderate-strain characters, and worst for the moderate-to-high-strain characters. The results of the coefficient of variation/t-test analyses were consistent with the predictions of the hypothesis to the extent that the moderate-to-high-strain measurements exhibited significantly more variability than the non-phenotypically-plastic and low-to-moderate-strain measurements. In contrast, the results of the phylogenetic analyses were not those predicted. The phylogeny derived from the moderate-to-high-strain characters matched the molecular phylogeny better than those obtained using the non-phenotypically-plastic and low-to-moderate-strain characters. Thus, our study supports the suggestion that mechanical loading results in phenotypic plasticity in the hominin skull, but it does not support the notion that homoiologies have a significant negative impact on hominin phylogenetics.

Introduction

Knowledge of hominin phylogeny is necessary for the successful reconstruction of human evolutionary history. Without a reliable phylogeny, little confidence can be placed in hypotheses of ancestry or in hypotheses regarding the number and nature of adaptive changes in human evolution (Eldredge and Tattersall, 1975). A reliable phylogeny is also necessary to test evolutionary scenarios that link events in human evolution with changes in the environment and with wider patterns of faunal evolution (Eldredge and Tattersall, 1975). Unfortunately, the phylogenetic relationships of the species whose remains compose the hominin fossil record are currently far from certain. Despite a relatively rich, well-dated fossil record and many methodological improvements (e.g., Chamberlain and Wood, 1987, Skelton and McHenry, 1992, Lieberman et al., 1996, Strait et al., 1997, Strait and Grine, 1999), cladistic analyses have so far been unable to estimate the phylogenetic relationships of several fossil hominin species with a reasonable level of confidence (Corruccini, 1994, Lieberman et al., 1996, Wood and Collard, 1999, Strait and Grine, 2004).

Our inability to reliably reconstruct these relationships has frequently been attributed to taxonomic uncertainties, to the use of incorrect characters, and/or to the way in which the cladistic methodology has been implemented (e.g., Chamberlain and Wood, 1987, Skelton and McHenry, 1992, Lieberman, 1995, Lieberman, 1999, Strait et al., 1997, Skelton and McHenry, 1998, Strait and Grine, 1998, Lovejoy et al., 1999, Lovejoy et al., 2000, McCollum, 1999, McCollum and Sharpe, 2001). In recent years, however, attention has focused on the confounding effects of homoplasies (e.g., Wood and Chamberlain, 1986, Skelton and McHenry, 1992, McHenry, 1994, McHenry, 1996, Lieberman, 1997, Lieberman, 1999, Lieberman, 2000, Lieberman et al., 1996, Lockwood and Fleagle, 1999, Collard and Wood, 2000, Collard and Wood, 2001). Homoplasies are resemblances between taxa that result from processes other than descent from a common ancestor and which suggest relationships that are inconsistent with the best estimate of the phylogeny for the taxa (Willey, 1911, Simpson, 1961, Hennig, 1966, Cain, 1982, Patterson, 1982, Sober, 1988, Sanderson and Hufford, 1996, Lockwood and Fleagle, 1999).

Homoplasies are a problem for phylogenetic systematists because they can be mistaken for shared derived similarities (synapomorphies), which are the main evidence for phylogeny. When a character state data matrix contains a small number of homoplasies in relation to the number of synapomorphies, it is possible to obtain an unambiguous estimate of phylogeny using parsimony analysis, which favors the hypothesis of relationship requiring the least number of changes to account for the distribution of character states among a group of taxa (Quicke, 1993, Kitching et al., 1998, Schuh, 1999). However, in phylogenetic studies of the hominins, the ratio of putative homoplasies to inferred synapomorphies has generally been around 1:2 (e.g., Skelton et al., 1986, Wood and Chamberlain, 1986, Chamberlain and Wood, 1987, Wood and Chamberlain, 1987, Wood, 1991, Skelton and McHenry, 1992, Lieberman et al., 1996, Strait et al., 1997). In these circumstances, parsimony analysis tends to yield several equally plausible phylogenies (Lieberman et al., 1996). For instance, Skelton et al.'s (1986) most parsimonious cladogram, in which Homo habilis and Paranthropus formed a clade to the exclusion of Australopithecus africanus, was supported by only one more character than the next most parsimonious cladogram, which linked Paranthropus with A. africanus to the exclusion of H. habilis. Similarly, although the cladograms favored by Wood (1991) and Strait et al. (1997) suggest that Homo is monophyletic, these cladograms are only slightly shorter than ones in which Homo is paraphyletic (Wood and Collard, 1999). The ambiguity that homoplasies introduce into hominin phylogenetic studies is further illustrated by the work of Strait and Grine (2004). Their bootstrap analyses returned insignificant levels of support for many fossil hominin phylogenetic relationships, and they failed to support the widely accepted relationships among the extant hominoids at the 70% level that is commonly used to classify clades as statistically significant in biological applications of the phylogenetic bootstrap (Hillis and Bull, 1993). The presence of numerous homoplasies among the character state data employed in these and other studies means that little confidence can be placed in published hominin phylogenies (Corruccini, 1994, Lieberman et al., 1996, Wood and Collard, 1999). Thus, developing a better understanding of the distribution and causes of homoplasy among humans and their closest fossil relatives represents a major challenge for hominin paleontology (Lieberman, 1995, Lieberman, 1997, Lieberman, 1999, Lockwood and Fleagle, 1999, Collard and Wood, 2000, Collard and Wood, 2001).

It is worth noting that hominin phylogenetic analyses are not unique in being confounded by extensive homoplasy. Several recent studies have shown that hard-tissue homoplasies occur in large numbers among many primate groups. For example, Hartman (1988) found that extant hominoid molar morphology was misleading regarding phylogenetic relationships due to diet-related convergence between humans and orangutans. Likewise, Harrison (1993) concluded that his attempts to resolve the relationships among closely related fossil primates, such as the early Miocene catarrhines of East Africa and the Eurasian pliopithecids, had been largely unsuccessful as a result of homoplasy. Most recently, Collard and Wood (2000) demonstrated that the crania of extant hominoids and papionins exhibit levels of homoplasy so high that phylogenetic analyses of qualitative and quantitative characters return strongly supported estimates of phylogeny that differ greatly from the groups' consensus molecular phylogenies, which are widely considered to be accurate. As such, there is a pressing need to understand not only hominin homoplasy, but also homoplasy among nonhuman primates (Lieberman, 1995, Lieberman, 1997, Lieberman, 1999, Lockwood and Fleagle, 1999, Collard and Wood, 2000, Collard and Wood, 2001).

It has been suggested that many hominin cranial homoplasies are likely to be homoiologies (Lieberman, 1995, Lieberman, 1997, Lieberman, 1999, Lieberman, 2000, Lieberman et al., 1996, Collard and Wood, 2000, Collard and Wood, 2001, Gibbs et al., 2000, Gibbs et al., 2002). Homoiologies are a phylogenetic consequence of phenotypic plasticity, the expression by a genotype of different phenotypes in response to different environmental conditions (Reidl, 1978, Lieberman, 1995, Lieberman, 1997, Lieberman, 1999, Lieberman, 2000, Lieberman et al., 1996). That is, homoiologies are resemblances among a group of taxa that suggest relationships that conflict with the best estimate of phylogeny for the taxa, and which result primarily from epigenetic responses to internal and external stimuli. The “homoiology hypothesis” derives from studies on the effects of mechanical loading on bone, which suggest that a large, possibly predominant, proportion of variation in bone shape and size is a function of interactions between regions of the skeleton and their mechanical environments (Currey, 1984, Lanyon and Rubin, 1985, Frost, 1986, Frost, 1998, Herring, 1993, Lieberman, 1995, Lieberman, 1996, Lieberman, 1997, Lieberman, 1999, Lieberman, 2000, Lieberman and Crompton, 1998, Martin et al., 1998). Comparative studies and controlled experiments on vertebrate models, including modern humans, have shown that mechanical loading during growth substantially affects both cortical bone growth in diaphyses and trabecular bone growth in epiphyses (Currey, 1984, Lanyon and Rubin, 1985, Frost, 1986, Lieberman and Crompton, 1998, Martin et al., 1998). These effects may be systemic or local. For example, Lieberman (1996) found that pigs and armadillos exercised during growth had markedly thicker cortical bone than did individuals that were not exercised. This difference occurred not only in the limbs but also in the cranial vault, where strains are too low to induce osteogenic activity. In addition, studies of disuse (e.g., from denervation, bed-rest, and gravity-free environments) also indicate that bone resorbs—often at rapid rates—in many regions of the skeleton when subjected to lower than normal strain magnitudes or frequencies (Martin et al., 1998).

A typical example of these effects on a character that is used frequently in hominin paleontology is variation in the bicondylar angle of the femur. Mechanical loads from locomotion during growth influence this trait in many ways, as indicated by the absence of the angle in newborn infants, its increase with age until maturity, and by its absence in individuals immobilized during childhood (Tardieu, 1995). However, it should be noted that, while mechanical loading has been shown to influence many skeletal characters, the applicability of some rather extreme experimental studies (e.g., osteotomies) to natural variation is questionable (Bertram and Swartz, 1991), and for most characters it has been difficult to quantify the relative proportion of variation explained by genetic versus environmental effects.

Here we report a study in which the homoiology hypothesis was evaluated by determining whether or not predictions about the distribution and phylogenetic utility of phenotypically plastic traits were supported in analyses of extant primates. We focus on the phenotypic-plasticity-inducing effects of the strains associated with mastication, which experimental work suggests are sufficiently high in some regions of the skull to influence aspects of primate cranial and mandibular shape (Hylander, 1988). The first prediction tested was that skeletal features that do not remodel and therefore are unaffected by phenotypic plasticity should be less variable than skeletal features that remodel and are subject to low-to-moderate levels of strain, and that the latter should be less variable than skeletal features that remodel and are subject to moderate-to-high levels of strain (Wood and Lieberman, 2001). The second prediction tested in the study was that skeletal features that do not remodel and therefore are unaffected by phenotypic plasticity should be more reliable for phylogenetic reconstruction than skeletal features that remodel and are subject to low-to-moderate strain, and that the latter should be more reliable for phylogenetic reconstruction than skeletal features that remodel and are subject to moderate-to-high strain.

To test the first prediction, we employed the coefficient of variation and the t-test. To test the second, we adopted an approach that has been used to investigate the phylogenetic utility of hominoid molar morphology (Hartman, 1988), hominoid, papionin, and galagid cranial and dental morphology (Collard and Wood, 2000, Collard and Wood, 2001, Masters and Brothers, 2002, Strait and Grine, 2004, and hominoid soft-tissue features (Gibbs et al., 2000, Gibbs et al., 2002). We analyzed craniodental data for the hominoids using cladistic methods and compared the resulting phylogenetic hypotheses with the group's consensus molecular phylogeny (Fig. 1), which is widely considered to be reliable (Ruvolo, 1997, Gagneux and Varki, 2001, Page and Goodman, 2001). Incongruence between the morphological and molecular phylogenies was taken to indicate the presence of a relatively large number of homoplasies in the morphological data set, whereas congruence was assumed to indicate the presence of relatively few homoplasies. This analytical approach is controversial because it assumes that the molecular data are more reliable for phylogenetic reconstruction than the morphological data, and some researchers do not accept that certain data sets are more reliable than others for phylogenetic reconstruction (e.g., Kluge and Wolf, 1993, Kluge, 1998, O'Leary, 1999). However, we believe there are several reasons why it is reasonable to use the hominoid consensus molecular cladogram to evaluate the homoplasy content of different hominoid hard-tissue data sets. First, in phylogenetics, morphology can never be more than a proxy for genetic data because it is genes that are passed between generations, not morphological characters. Second, it is well documented that many “good,” reproductively isolated species are genetically distinct, but dentally and osteologically indistinguishable (Tattersall, 1986, Tattersall, 1992, Aiello et al., 2000). Since speciation events create phylogenetic relationships, there is thus an a priori expectation that skeletal characters will be less useful for phylogenetic reconstruction than genetic characters. Third, molecular phylogenetic techniques have been successfully tested on laboratory taxa of known phylogeny, whereas comparable analyses of morphological data have not been successful (Fitch and Atchley, 1987, Atchley and Fitch, 1991, Hillis et al., 1992). Lastly, and most importantly, multiple independent genes support the consensus molecular phylogeny for the hominoids (Ruvolo, 1997, Gagneux and Varki, 2001, Page and Goodman, 2001). Congruence among multiple independent lines of evidence is the best support possible for a phylogenetic hypothesis.

Section snippets

Materials and methods

Primate skeletal morphology is conventionally translated into character states for cladistic analysis in two main ways. The first breaks the phenotype up into character states qualitatively. Thus, a prominence is described as “strong,” “reduced,” or “absent,” a contour as “arched” or “less-arched,” and a feature as “developed” or “not developed.” This approach has been used in most hominin cladistic analyses (e.g., Eldredge and Tattersall, 1975, Delson et al., 1977, Skelton et al., 1986,

Results

Two sets of analyses were carried out to evaluate the hypothesis that homoiology is a significant form of homoplasy among hominin taxa. In the first, the CV and t-tests were used to test the prediction that the moderate-to-high-strain measurements should be more variable than the low-to-moderate-strain measurements, and that the latter should be more variable than the non-phenotypically-plastic measurements. Table 2 summarizes the average CV for each group of measurements by taxon, along with

Discussion

It has been suggested recently that homoiologies are a common form of homoplasy in the hominin skull, especially in those regions affected by mastication-related strain, and that their prevalence has contributed to our failure to date to obtain a reliable estimate of hominin phylogeny (Lieberman, 1995, Lieberman, 1997, Lieberman, 1999, Lieberman, 2000, Lieberman et al., 1996, Collard and Wood, 2000, Collard and Wood, 2001, Gibbs et al., 2000). In order to evaluate this hypothesis, we compiled a

Conclusions

The study described here was undertaken to assess the validity of the hypothesis that homoiology is an important source of homoplasy among fossil hominins. Two analyses were carried out using data from the extant hominoid primates and the group's consensus molecular phylogeny. The first sought to determine whether skeletal features that remodel and are subject to moderate-to-high masticatory strains are significantly more variable (i.e., phenotypically plastic) than skeletal features that

Acknowledgements

The study reported in this paper was carried out in collaboration with Dan Lieberman. We are grateful to Charles Lockwood and John Fleagle for inviting us to participate in the 1999 symposium. For their help with various aspects of this project, we thank Alan Bilsborough, Briggs Buchanan, Nicole Collard, Todd Disotell, Mike Lague, Stephen Lycett, Chris Paul, David Pilbeam, Todd Rae, and Jamie Tehrani. The manuscript was improved by insightful comments from three reviewers: David Begun, Callum

References (142)

  • P. Gagneux et al.

    Genetic differences between humans and great apes

    Mol. Phylogenet. Evol.

    (2001)
  • E.E. Harris

    Molecular systematics of the Old World monkey tribe Papionini: analysis of the total available evidence

    J. Hum. Evol.

    (2000)
  • S.E. Hartman

    A cladistic analysis of hominoid molars

    J. Hum. Evol.

    (1988)
  • J. Hawks et al.

    An Australasian test of the recent African origin theory using the WLH-50 calvarium

    J. Hum. Evol.

    (2000)
  • W.L. Hylander

    In vivo bone strain as an indicator of masticatory bite force in Macaca fascicularis

    Arch. Oral Biol.

    (1986)
  • W.L. Hylander et al.

    Jaw movements and patterns of mandibular bone strain during mastication in the monkey Macaca fascicularis

    Arch. Oral Biol.

    (1986)
  • A.G. Kluge

    Total evidence or taxonomic congruence: cladistics or consensus classification

    Cladistics

    (1998)
  • A.G. Kluge et al.

    Cladistics: what's in a word?

    Cladistics

    (1993)
  • A. Kramer et al.

    Out of Africa and into the Levant: replacement or admixture in western Asia

    Quatern. Int.

    (2001)
  • D.E. Lieberman et al.

    Homoplasy and early Homo: an analysis of the evolutionary relationships of H. habilis sensu stricto and H. rudolfensis

    J. Hum. Evol.

    (1996)
  • C.A. Lockwood

    Adaptation and functional integration in primate phylogenetics

    J. Hum. Evol.

    (2007)
  • M. O'Leary

    Parsimony analysis of total evidence from extinct and extant taxa and the cetacean-artiodactyl question (Mammalia, Ungulata)

    Cladistics

    (1999)
  • S.L. Page et al.

    Catarrhine phylogeny: non-coding DNA evidence for a diphyletic origin of mangabeys and for a human-chimpanzee clade

    Mol. Phylogenet. Evol.

    (2001)
  • R.R. Ackermann et al.

    Discerning evolutionary processes in patterns of tamarin (genus Saguinus) craniofacial variation

    Am. J. Phys. Anthropol.

    (2002)
  • J.C.M. Ahern

    Underestimating intraspecific variation: the problem with excluding Sts 19 from Australopithecus africanus

    Am. J. Phys. Anthropol.

    (1998)
  • L.C. Aiello et al.

    Assessing exact randomization methods for determining the taxonomic significance of variability in the hominin fossil record

    S. Afr. J. Sci.

    (2000)
  • B. Asfaw et al.

    Remains of Homo erectus from Bouri, Middle Awash, Ethiopia

    Nature

    (2002)
  • W.R. Atchley et al.

    Gene trees and the origin of inbred strains of mice

    Science

    (1991)
  • J.E.A. Bertram et al.

    The ‘law of bone transformation’: a case of crying Wolff?

    Biol. Rev. Camb. Philos. Soc.

    (1991)
  • C.L. Brace et al.

    Old World sources of the first New World human inhabitants: a comparative craniofacial view

    Proc. Natl. Acad. Sci. U.S.A.

    (2001)
  • J. Braga

    Cranial discrete variation in the great apes: new prospects in palaeoprimatology

  • A.J. Cain

    On homology and convergence

  • D. Cameron et al.

    The phylogenetic significance of the middle Pleistocene Narmada hominin cranium from central India

    Int. J. Osteoarcheol.

    (2004)
  • Caparros, M., 1997. Homo sapiens archaïques: Un ou plusieurs taxons (espèces)? Analyse cladistique et analyse...
  • J.M. Cheverud

    Phenotypic, genetic and environmental morphological integration in the cranium

    Evolution

    (1982)
  • M. Collard et al.

    How reliable are human phylogenetic hypotheses?

    Proc. Natl. Acad. Sci. U.S.A.

    (2000)
  • M. Collard et al.

    Ontogeny and homoplasy in the papionin face

    Evol. Dev.

    (2001)
  • G. Conroy

    Otavipithecus: or how to build a better hominid—not

    J. Hum. Evol.

    (1994)
  • R.S. Corruccini

    How certain are hominoid phylogenies? The role of confidence intervals in cladistics

  • P. Cranston et al.

    Cladistics and computers: a chironomid conundrum?

    Cladistics

    (1988)
  • M. Crisp et al.

    Cladistics and legume systematics, with an analysis of the Bossiaecae, Brongniartieae and Mirbelieae

  • J.D. Currey

    The Mechanical Adaptations of Bones

    (1984)
  • T.R. Disotell

    Generic level relationships of the Papionini (Cercopithecoidea)

    Am. J. Phys. Anthropol.

    (1994)
  • T.R. Disotell

    The phylogeny of Old World monkeys

    Evol. Anthropol.

    (1996)
  • T.R. Disotell et al.

    Mitochondrial DNA phylogeny of the Old World monkey tribe Papionini

    Mol. Biol. Evol.

    (1992)
  • N. Eldredge et al.

    Evolutionary models, phylogenetic reconstruction and another look at hominid phylogeny

  • J. Felsenstein

    Phylogenies and quantitative characters

    Annu. Rev. Ecol. Syst.

    (1988)
  • W.M. Fitch et al.

    Divergence in inbred strains of mice: a comparison of three different types of data

  • J.G. Fleagle et al.

    Skeletal and dental morphology supports diphyletic origin of baboons and mandrills

    Proc. Natl. Acad. Sci. U.S.A.

    (1999)
  • H.M. Frost

    Intermediary Organization of the Skeleton

    (1986)
  • Cited by (47)

    • Biosocial complexity and the skull

      2023, Behaviour in our Bones: How Human Behaviour Influences Skeletal Morphology
    • Using the shape of the basicranial portion of the temporal bone to distinguish between relatively closely-related human populations

      2019, Journal of Archaeological Science: Reports
      Citation Excerpt :

      For example, the facial skeleton has been found to be considerably reflective of climatic variables (Roseman, 2004; Hubbe et al., 2009), especially in areas linked to the nasal aperture, which varies dramatically according to humidity, latitude, and temperature (Carey and Steegmann, 1981; Franciscus and Long, 1991). Diet, through mastication of various foods, has also been found to influence cranial shape by exerting mechanical strain on the cranium (Collard and Wood, 2007; von Cramon-Taubadel, 2009). However, the shape of the cranial vault, including the temporal bone, has been found to be largely unaffected by such environmental factors, generally correlating with population history (Olson, 1981; Wood and Lieberman, 2001; von Cramon-Taubadel, 2009, 2014).

    View all citing articles on Scopus

    Lycett and Collard (2005) was written as a follow-up to this paper, but it appeared first due to publication delays associated with the collection of papers presented in this special issue. To maintain consistency with the paper by Lycett and Collard (2005), only minor changes have been made to the present article during revision.

    1

    Tel.: +1 202 994 6075; fax: +1 202 994 6097.

    View full text