IDEA ANDPERSPECTIVEPhylogenetic niche conservatism, phylogenetic signaland the relationship between phylogeneticrelatedness and ecological similarity among speciesJonathan B. Losos*Museum of ComparativeZoology and Department ofOrganismic and EvolutionaryBiology, Harvard University,26 Oxford Street, Cambridge,MA 02138, USA*Correspondence: E-mail:[email protected] are increasingly adopting an evolutionary perspective, and in recent years, theidea that closely related species are ecologically similar has become widespread. In thisregard, phylogenetic signal must be distinguished from phylogenetic niche conservatism.Phylogenetic niche conservatism results when closely related species are moreecologically similar that would be expected based on their phylogenetic relationships;its occurrence suggests that some process is constraining divergence among closelyrelated species. In contrast, phylogenetic signal refers to the situation in which ecologicalsimilarity between species is related to phylogenetic relatedness; this is the expectedoutcome of Brownian motion divergence and thus is necessary, but not sufficient,evidence for the existence of phylogenetic niche conservatism. Although many workersconsider phylogenetic niche conservatism to be common, a review of case studiesindicates that ecological and phylogenetic similarities often are not related. Consequently,ecologists should not assume that phylogenetic niche conservatism exists, but rathershould empirically examine the extent to which it occurs.KeywordsPhylogenetic effect, phylogenetic niche conservatism, phylogenetic signal.Ecology Letters (2008) 11: 995–1007INTRODUCTION…species of the same genus have usually, though byno means invariably, some similarity in habits andconstitution…(Darwin 1859, p. 76)In the last decade, ecologists increasingly have becomeinterested in the extent to which Darwin was correc t thatclosely related species tend to be ecologically similar. Thisfocus stems from a numb er of related interests. At thebroadest level, ecologists are interested in the evolutionarytapestry of life: how has ecological diversification pro-duced the biological diversity we see today? In a similarvein, community ecologists now frequently compare thedegree of phylogenetic relatedness of community mem-bers to that of species source pools to draw inferencesabout the processes structuring communities. In addition,the application of phylogenetic comparative methods tostatistical anal yses of ecological data is predicated on theassumption that closely related species tend to beecologically similar.The relationship between phylogenetic relatedness andecological similarity among species has been investigatedusing two approaches. The fir st is to quantify phylogeneticsignal1, which is the Ôtendency for related species to resembleeach other more than they resemble species drawn at randomfrom the [phylogenetic] treeÕ (Blomberg & Garland 2002,p. 905). Phylogenetic signal would result if characters evolvein a Brownian moti on-like manner, in which the am ount ofchange in any given interval is generally small and random indirection; such a pattern of evolution could ensue either fromgenetic drift or from natural selection that randomlyfluctuated through tim e in direction and magnitude. Conse-quently, in this scenario, a relationship would be expected1Also termed Ôphylogenetic effectÕ (Derrickson & Ricklefs 1988). ÔPhylo-genetic inertiaÕ is sometimes also used to mean the same thing, but this termhas a long and twisted history in which it has taken a variety of differentmeanings to different workers (reviewed in Blomberg & Garland 2002).Ecology Letters, (2008) 11: 995–1007 doi: 10.1111/j.1461-0248.2008.01229.x 2008 Blackwell Publishing Ltd/CNRSbetween the degree of phylogenetic relatedness, quantified asthe time since divergence between pairs of species, and thedegree of phenotypic similarity between them; the less theamount of time since two species shared a common ancestor(i.e. the more closely related the two species), the less theexpected phenotypic difference between them (Blomberg &Garland 2002).The second approach to understanding the relationshipbetween ecological and phylogenetic similarities revolvesaround the idea of Ôphylogenetic niche conservatismÕ (PNC).In contrast to explanations for phylogenetic signal, theexistence of PNC suggests that some factor is causing closelyrelated species to be more similar ecolog ically than would beexpected by simple Brownian motion descent with modifi-cation2(for discussion of methods to quantify phylogeneticsignal and their relevance to PNC, see Freckleton et al. 2002;Blomberg et al. 2003; Revell et al. in press).Phylogenetic signal and PNC have received increasingattention by ecologists3. In a 2002 Annual Review of Ecology andSystematics, Webb et al. (2002, p. 487) stated that Ôphylogeneticanalysis reveals that many (possibly the majority of) lineagesstudied show evidence for conse rvatism of dominantecological characterÕ. Three years later in another AnnualReview of Ecology and Systematics article entitl ed, ÔNicheconservatism: integrating evolution, ecology, and conserv a-tion biologyÕ, Wiens & Graham (2005, pp. 521–522) statedÔSpecies will always inhabit environments that bear somesimilarity to those of their close relatives…Th us, to someextent, niches will always be conserved. Yet few sister speciesmay share identical niches; so niches may never be perfectlyconserved eitherÕ.Many recent studies, however, have taken the moreextreme view that phylogenetic signal in ecolog ical charactersis widespread and pervasive, and thus constitute the a prioriexpectation when ecological variation within a clade isexamined in a phylogenetic context (e.g. Duncan & Williams2002; Anderson et al. 2004; Peres-Neto 2004; Mouillot et al.2006; Swenson et al. 2006; Johnson & Stinchcombe 2007).Nonetheless, I believe that the universality of ecologicalphylogenetic signal and PNC has been overstated. The goa l ofthis paper is to clarify the distinction between phylogeneticsignal and PNC and to examine evidence for the claims thatthey are ubiquitous. To do so, I must first explicitly definePNC and discuss how it can be quantified. I will then concludeby assessing the implications of this review for understandingthe evolution of ecological patterns and processes.WHAT IS PNC?Harvey & Pagel (1991) did not actually define PNC, but whatthey wrote suggested that it