The Proposal To Split Anolis Into Eight Genera: Time To Discuss

ResearchBlogging.orgWe’ve had a week now to let the proposed reclassification of Anolis sink it, so it’s time to start discussing it. A revolutionary new view of the scientific review process suggests that in the future, all papers will be published open access online (as this one is–thanks Nicholson et al.), the journal in which it appears (if any) will not matter, and peer review and evaluation will be conducted post-publication on internet discussion sites. Realistic? Who knows, but why not give it a try?

The paper by Nicholson et al. is undoubtedly the most important paper on anoles to be published in the last several years. Not only does it propose to split Anolis into eight genera, but it also presents provocative findings about the ecological evolution of anoles (including throwing out the ecomorph concept), anole biogeography, and the dating of evolutionary events in anole history.

Anole Annals’ goal is to be the meeting place for discussion of all things Anolis, so let’s take this post-publication review and commentary idea out for a spin. Anole Annals invites members of the anole community to post their thoughts on any aspect of the Nicholson et al. paper. We hope to get a conversation going on the merits of splitting the genus, as well as the other issues raised in the paper. In fact, this has already begun, as evidenced by the comments by Mssrs. Crother, Hillis and Duellman, among others.

To get the ball rolling, here’s a short précis of the paper:

1. Phylogenetic analysis based on previously published data of all sorts (genetic, morphological, karyological), with a smidgeon of new molecular data, reveals a phylogeny with eight strongly supported clades in a Bayesian analysis. These clades are recognized as distinct genera.

2. The ecomorph concept does not apply to mainland anoles because species similar in habitat use are not similar in morphology. Hence the term “ecomode” is coined for species similar in habitat use. Phylogenetic analysis of ecomode evolution on the phylogeny suggests that the crown-giant ecomode is ancestral for Anolis. The ecomorph concept is argued to not work for Greater Antillean anoles and should be discarded.

3. Biogeography is reconstructed on the phylogeny. Using the phylogeny, the authors argue that the eight clades differentiated about the time that the proto-antillean islands were passing between what is now North and South America. The Norops clade differentiated on several of these blocks (both island and mainland), explaining why Norops is nested within Caribbean non-Norops taxa without requiring the island-to-mainland colonization of Norops proposed by a number of previous papers.

4. Molecular clock dating reveals that anoles are surprisingly ancient, originating in South America approximately 130 million years ago.

Nicholson, K. E., B. I. Crother, C. Guyer, J. M. Savage (2012). It is time for a new classification of anoles (Squamata: Dactyloidae) Zootaxa, 3477, 1-108

About Jonathan Losos

Professor and Curator of Herpetology at the Museum of Comparative Zoology at Harvard University. I've spent my entire professional career studying anoles and have discovered that the more I learn about anoles, the more I realize I don't know.

34 thoughts on “The Proposal To Split Anolis Into Eight Genera: Time To Discuss

  1. For starters, I’d like to see Nicolson et al. weigh in on why they chose to recognize only 8 genera instead of more of the 22 subclades they identified…

    1. It’s too bad there aren’t any trees with meaningful branch lengths in the paper. If you look at the Anolis genome paper (Alfoldi et al., 2011) and look at the tree in the supplemental data ( ) it’s easier to see why eight, and why these 8 are unlikely to be mitochondrial DNA artifacts (the genome paper uses lots of loci). The key to knowing why eight, however lies in the fact that Norops has to be one of the genera — Assuming Norops is a genus, there are 8 clades with a black dot that are about the same age as Norops.
      I have mixed feeling about splitting the genus up (possibly pissing off everyone with strong feelings either way). I would like anyone working on anoles to know these eight groups, and to be familiar with the 22 subclades as well — but how to best achieve better knowledge of the phylogeny of anoles is not straightforward. Using subgeneric or clade names is fine, if they get used and get used often. If only taxonomists and serious tree-making anole workers use the names for these clades, then the phylogenetic information hasn’t been conveyed. Splitting up the genus (if recognized) forces everyone to use more phylogenetically precise language, but it clearly isn’t a popular choice.

      Other thoughts I have:
      -The biggest discrepancy on the tree between mtDNA and nuclear is A. argenteolis – and the choice to not give it a new name is likely anticipatory of nuclear support for the placement of this clade in a different place.
      -It seems like a lot of mtDNA data was not used that could have been. For example, A. bonairensis has complete ND2, except the first 50 bases – that data was not used – in the combined analysis it comes out with high support with occultus — which doesn’t make sense.
      -Reconstructing ancestral states for ecomorphs or ecomodes is not a great idea for two reasons:
      1. Demonstrated convergence (or parallelism) should make anyone suspicious of conclusions based on assumptions that have been demonstrated to have been violated.
      2. Having played around with these reconstructions – I would guess that a careful look at possible ancestral states will lead toward highly equivocal conclusions about ancestral states that will be strongly influenced by things like sampling bias in mainland anoles and whether crown giant (or any of the other ecomodes) are in fact equivalent between groups

      1. I don’t think the number 8 has that much to do with the clades being of roughly equal age. If you recognize Norops as a genus, then the minimum number of mutually exclusive clades (regardless of age) on their molecular tree (Fig. 4) that must be ranked as genera to encompass all of the remaining species of anoles is 7. Convention dictates that genera must be mutually exclusive.

      2. Regarding Todd’s point on sampling. I’ve been working on a version of this alignment for a few months and I find 214 Anolis species with at least one vouchered, full-length ND2 sequence on GenBank.

        1. Have these sequences just been posted to GenBank or would they likely have been available to Nicholson et al. when they were working to expand previously published datasets of ND2 sequences?

          1. I can’t say exactly. Most of sequences missing from Nicholson et al. are from Castañeda and de Queiroz (2011, MP&E 61:784-800) which was published 21 July 2011. Nicholson et al. was accepted 17 May 2012 but Zootaxa doesn’t report when it was originally submitted. On a related note, I don’t see the new sequences generated by Nicholson et al. on GenBank (medemi, townsendi, gorgonae, and princeps).

  2. I think it makes great sense to divide an old, diverse clade like Dactyloidae into multiple genera, and it sounds like the results of Nicholson et al. are largely congruent with the phylogeny of Alfoldi et al.
    Me asusta, pero me gusta!

    1. The conclusion that it makes sense to divide this clade into multiple “genera” rests in part on the recognition of the “family” Dactyloidae (which is identical in composition to the “genus” Anolis sensu lato). But it was unnecessary to divide the traditional Iguanidae into multiple “families” in the first place (which ultimately lead to the recognition of Dactyloidae), given that it is monophyletic. And if Anolis is one of many “genera” of iguanids, then it makes less sense to “divide” it.

      1. Three related thoughts: 1) Another thing that may contribute to the idea that it makes sense to “split” Anolis is that fact that the name Dactyloidae has (rank-based) nomenclatural precedence over Anolidae; therefore if Anolis is not “split” then the “family” Dactyloidae would not contain a “genus” Dactyloa. 2) I said above that Dactyloidae is identical in composition to Anolis sensu lato. In reality, we don’t know much about the composition of Dactyloidae except that Anolis carolinensis is included and Polychrus marmoratus is excluded, because the paper that proposed recognizing Dactyloidae included only those two species and not any from other “genera” previously assigned to the anoloids/Polychrotidae/Polychrotinae (we don’t really even know if Dactyloa is included, though it’s probably safe to assume that Anolis punctatus [the type species of Dactyloa] is more closely related to Anolis carolinensis than to Polychrus marmoratus). Third, support for the relationships of Anolis and Polychrus to other taxa (i.e., the justification for placing the former in a separate “family”) is weak.

        1. I think points one and three are good ones to make. As far as point two goes, I think that it’s safe to understand that Anolis sensu lato comprises the family Dactyloidae.

  3. I haven’t yet had time to extensively review (i.e., actually read) this very long article; but in my opinion revising a stable and widely accepted taxonomy shouldn’t be done lightly and probably not without a “phylogenetic” justification (i.e., to resolve issues of para- or polyphyly). Since this does not seem to be the case here (specifically, Anolis sensu lato is monophyletic), won’t renaming all these species serve primarily to add a lot of unnecessary confusion to the literature?

    1. There absolutely no doubt that Nicholson et al.’s proposed revisions – which involve redefining the meaning of Anolis, changing the binomials for the majority of anole species, and altering the spelling of species epithets for ~10% of anole species – would cause massive confusion in the literature. Whether this type of wholesale change in anole nomenclature is necessary is a subjective question about which systematists might reasonably disagree. Count me among those who view this change as unnecessary, and much more likely to complicate anole research than it is to facilitate future work on the group. We can use the same type of informal system of classification (which eschews formal changes to generic epithets and the resulting need to change binomials) that anole biologists have favored for decades to capture the same type of phylogenetic information contained in Nicholson et al.’s formal classification. The names applied to genera no longer have the significance they did prior to the widespread use of phylogenetic trees for studies across a range of disciplines in ecology and evolutionary biology. As a result, proposals to formally redefine or change generic epithets that have been as widely used as Anolis has for more than a half century seem destined to fail, particularly when the name(s) we’ve been using refer to monophyletic groups and are perfectly compatible with the ICZN and other documents that govern nomenclatural decisions.

  4. I am also unfond of the proposal to eliminate the term “ecomorphs” in favour of “ecomode”. Ecomode appears to be synonymous with microhabitat, so why even have a different word for this? Furthermore, having a term just to be more inclusive of mainland anoles that inhabit the same microhabitat as Greater Antillean ecomorphs but do not share the same characteristic morphological traits, completely shadows the repeated adaptive evolution that occurred on the Greater Antilles. Using the term ecomode is “a hindrance to scientific communication regarding evolutionary events and directions of future research”; the very reason why Nicholson et al. suggest splitting up Anolis!

    1. It is also worth noting that much of the discussion in the paper about the evolution of habitat specialists (a term that probably has historical priority over ecomode) relies entirely on parsimony to reconstruct ancestral character states. Other sections of the paper do not rely so heavily on a single technique with known shortcomings. In this particular example, we already know that parsimony reconstructions of habitat specialist types – ecomorphs – differ from ML, and that it is incredibly important to quantify uncertainty in ancestral states. I wonder what the ML reconstruction of ecomode looks like for this clade? I suspect that, if one were to quantify statistical uncertainty, then the ancestral “ecomode” of these lizards would be ambiguous.

    2. I agree with you, Julienne. This is a complete “baby with the bathwater” situation. Obviously, there are severe shortcomings to their methodologies, as Luke stated. Beyond that, I am just plain uncomfortable with the degree of rigor applied to correctly assigning the taxa to their arbitrary categories. For example, to the best of my ability I have assigned members of the cybotoid group to the “Trunk-ground” or to “saxicolous” ecomode. This comes from dedicating my PhD to studying the ecology of this group, so this doesn’t come out of nowhere. Anolis armouri, A. shrevei, A. marcanoi, A. longitibialis, and A. strahmi are arguably saxicolous “ecomodes”, while A. cybotes, A. whitemani, and A. breslini are more “trunk-grounds” – based on the phylogeny (A. armouri and A. shrevei are nested within A. cybotes), this would through ambiguity into the reconstruction of the ecomode for the cybotoid node. How much can I trust their assignations without any tables or ecological data?

      1. In my opinion, the issue with this entire “ecomode” discussion begins well before the analytical methods the authors employ. It is based on false assumptions and an inaccurate dissection of the classic anole literature.

        First, it is unclear which definition of “ecomorph” these authors are using prior to trying discredit it. As discussed by JBL (, some authors have equated “ecomorph” with “habitat specialist.” Nicholson et al. appear to use the broader definition of ecomorph as the backbone of their argument, which I do not find compelling in this context. This is not the original intent of the term or how it has primarily been applied within the Anolis literature historically. I will not belabor this point here. Read the JBL post for more details on this.

        But sticking with the historical context of the Nicholson argument for just another moment, it also seems as if the authors attempt to credit Losos 1992 with an inaccurate conception of ecomorph concept based primarily on observations of Jamaican species. The foundational works of Anolis ecomorphology by Williams, Rand, Schoener, Gorman, and others are not acknowledged, nor the vast majority of literature that has been published over the last ten years. I find this discussion grossly misleading, focusing only on a handful of studies that seem to support their desired narrative.

        Moving on, implied in the Nicholson et al. discussion is the notion that exceptions to the ecomorph concept automatically disqualifies this as a valid framework. To my knowledge, placing every species into a defined ecomorph category has never been an objective of community (until now apparently). The “ecomorph” designation has been used to highlight the remarkable convergence among distantly related lineages. There are recognized “unique” or “non-ecomorph” species commonly acknowledged in the literature. While quantitative analysis of mainland species has lagged somewhat behind efforts in the Caribbean increased efforts to quantify patterns of mainland ecomorphology are underway (, I suggest that further comparisons between patterns of mainland and Caribbean ecomorphology wait until comparable ecological and morphological datasets are available. As previously stated, the addition of new, less well-defined terminology only serves to confuse efforts that already underway.

  5. I’ve been trained with systematists, and created tools to help systematists, but I am not a practicing systematist. With that caveat in mind, it’s my impression that this is a valid revision, regardless of the wisdom of it [please note that, if it’s not valid, the rest of this post is pointless]. That is, as of its publication, the group formerly known as Anolis is now eight genera according to how the ICZN works. Replies to the previous AA post on this topic argued for “refraining from adopting” the new classification. That does happen in practice, of course, but it’s a violation of the Code. I’m on the consumer side of taxonomy nowadays, often aggregating trees/data from various sources, and one big problem we run into is taxonomic name resolution. This will always be a problem, because names often do have to change due to new discoveries, and that’s fine. But having one set of folks adopt a different taxonomy from a different set of folks leads to competing simultaneous taxonomies, which is exactly the problem modern taxonomy is supposed to solve. If this is an unwise revision (the repeated use of “basal” applied to extant species is one worrisome feature), it would be great to have this fixed, rather than just ignored by some but not all, leading to later confusion.

    1. Unfortunately, I can’t pretend to understand ICZN – but the implication of your post is that anyone can arbitrarily and unilaterally revise existing, stable, widely accepted, and phylogenetic consistent taxonomy, without any consultation in the scientific community, so long as their revision follows ICZN. This is quite remarkable, if you think about it!

    2. Brian’s remarks about the necessity of applying Nicholson et al.’s formal taxonomic revision under the ICZN are completely false. The ICZN does not regulate taxonomy, only nomenclature. Indeed, one key principle of the ICZN says, ‘‘The Code refrains from infringing upon taxonomic judgment, which must not be made subject to regulation or restraint.’’ It would be a violation of the ICZN to apply other generic epithets to the taxa that Nicholson et al. have diagnosed. It would also be a violation of ICZN to use Anolis if this generic epithet did not have priority over other generic epithets applied to anoles. However, it is in no way a violation of the ICZN or any other rules that govern use of taxon names to simply continue applying Anolis to all anoles.

      1. Rich is absolutely right about the ICZN not regulating taxonomy. A common misconception among nonsystematic biologists.

        1. Lest someone coming to this discussion later think there’s a controversy here about what is required, I just want to note that I was wrong. I thank Rich for the correction; he has a longer post about it here.

    1. I actually wouldn’t go so far as to say that no valid purpose is served by splitting Anolis into multiple genera. I can see how this splitting could increase our appreciation for the diversity of the group in some contexts. For example, it might be more obvious that a community of anoles on Hispaniola is phylogenetically diverse if we said it included Anolis, Ctenonotus and Xiphosurus. My point is that any such gain in our appreciation for anole diversity that might be achieved by formally recognizing new genera is marginal (particularly given the availability of phylogenetic trees and informal classifications that recognized series and species groups rather than genera) and does not justify the profound confusion that redefining and changing widely used taxonomic terms would produce for many years to come (to say nothing of the effort that would be required to re-organize and re-label hundreds of thousands of specimens and records in museums and biodiversity databases).

        1. I fully agree with Rich here in that recognizing multiple genera among anoline lizards may appear to help in appreciating the diversity of communities. However, in practice, it is completely unnecessary. The development of software such as Phylocom (Webb et al., 2002) renders the need to recognize classifications above the generic level unnecessary and does away with the historic concept of presenting community structure on the grounds of genus to species ratios. Modern phylogenetic methodologies have specifically aimed to improve accuracy in analyses related to community structure and diversity by getting away from the ambiguity of classification above the generic level. The fact of the matter is that tip-taxa are the unit of importance in phylogenetic comparative analyses and only the topology matters. Generic recognitions have their importance, but they do not benefit research from a comparative phylogenetic perspective.

          Regarding the issues surrounding the concept of “ecomodes,” Nicholson et al. appear to have completely missed the point. The recognition of the classic “ecomorphs,” while perhaps lacking in methodological rigor, exist due to strong recognition of ecological and morphological correlation. In some sense it’s a no-brainer that species’ morphology would reflect their ecology. This is just as likely the case among mainland anoles as it is among Greater Antillean (GA) anoles. The fact that mainland anoles don’t fit neatly within the GA ecomorphs only suggests that mainland anoles are faced with different ecological challenges and more work is needed to define mainland ecomorphs. Nicholson et al. misinterpret the work done on mainland ecomorphology, namely by Irschick et al., 1997 and Schaad & Poe, 2010. Nicholson et al. state “attempts to force mainland anoles into ecomorphological categories based on data from island anoles seems pointless….” No effort has been made to force mainland anoles into GA ecomorphs. Rather Irschick et al. and Schaad & Poe (among others) simply point out the fact that mainland anoles have experienced different evolutionary trajectories compared to island anoles. A point that would seem rather obvious based on past island adaptation research, but that needed to be made in order to set the stage for mainland research on anole ecomorphology. Julienne’s point (see above) is well made in that “ecomodes” seem to be akin to microhabitat. The point is not just to stick anoles in categories, but to define evolutionarily meaningful categories, such as is the case with the GA ecomorphs. Rather than solving a biological issue in mainland ecomorphology, the “ecomode” concept obscures the unique evolution of mainland ecomorphology and may hinder further research on the topic.

  6. Like Julienne, I take umbrage to the statement that “the single genus concept can be a hindrance to scientific communication regarding evolutionary events and directions of future research.” Obviously, given the tremendous amount of work conducted in anoles so far (from ecology to evolution to neurobiology to development), Anolis sensu lato hasn’t hindered research in anoles. There’s no reason to expect that it would in the future.
    Moreover, I would argue the contrary – that dividing Anolis sensu lato would only serve to obfuscate and confuse the design and scope of future studies. Dividing anoles into smaller groups will invite novices in the field to focus on subclades within the islands or the mainland rather than incorporate all the information we have about anoles into large comparative studies. As silly as may seem, a search of the literature would find that studies focus most often on congeners when studying adaptive radiation (as per Schluter 2000) and character displacement and interspecific competition and sexual selection and species recognition etc. etc. etc. So breaking up Anolis would only serve to reduce the scope of the studies that are undertaken in the future.
    Or, if the anole community remains immune to this congener effect, then we’d be bravely comparing across the 8 genera and writing papers that have to justify why we compared Anolis sensu stricto to Audantia to Xiphosurus when they’re all different genera. All the while, we’ll be using an operational definition of Anolis that fits Anolis sensu lato. If we’re operating scientifically with Anolis sensu lato, then we should keep Anolis sensu lato.

  7. While I completely agree that splitting Anolis sensu lato into several genera is uncalled for and will have severe and unnecessary downstream consequences, I think we should be careful not to throw the baby out with the bathwater in the opposite direction either. In particular, we anole biologists ought to make a serious effort to draw attention to the ecologically and morphologically distinct clades that this and previous studies have established. Studies that find general patterns across Anolis sensu lato can only be strengthened by emphasizing the diversity that this group encompasses. On the other hand, understanding patterns that differ across these diverse clades can only lead to a better understanding of the forces generating diversity in a group of what are, in many ways, very similar organisms. It sounds like a win-win situation…

  8. Couple of things I notice in this and I have seen similar arguments in the Rana debate as well. First of all is people often refer to stability. Stability is a nomenclatural concept and has nothing to do with the binomial combination used for taxa. There is no requirement in nomenclature for the combination to be stable, and that would be beyond the scope of the ICZN as they do not deal with taxonomy, and the binomen is taxonomy. Stability is for preserving names when older available names are found that have not been used for example. That is the oldest name for the Matamata is Testudo terrestris, which was dealt with by the ICZN in 1963. I find it strange people always argue about stability, ignoring its irrelevance to the issue. I specialise in nomenclature and most of the species described over 100 years ago are no longer in their original genera, so what you really mean by stability is the names you got used to in the last 30 odd years and don’t wish to update your databases on. Irrespective of what name they originally had.

    As for confusion, I would argue it is not as bad as you would think. I was involved in, rather responsible for, uncovering a major nomenclatural error in Australian turtles. It in the end required the changing of a name from one species to another and the sinking of a well used name and resurrection of an unused name. You think that can be confusing, well most people understood it without too much problem. Even though the older literature of course has the names the other way round, even my early papers do. People got used to it. They did so very quickly so the confusion argument, having been in the middle of one issue, is just a fear not an event. People get it, without too much problem. That is what synonymies are for.

    The argument of don’t change the taxonomy is, to me, anti science, I mean if we come up with new arrangements that have a founding in good science why not advance the taxonomy? Or should it be held up because people like an old name. Enormous genera that cannot be diagnosed serve little to no purpose. The genus is more than a monophyletic group, yes monophyly is the objective evidence for the genus to stand, but there are other factors. The genus should be diagnosible and informative. It should give more information than it does by the evidence of its monophyly alone. I accept this may not necessarily mean you use genera, subgenera are an option depending on the circumstances. Mayr once said that genera were defined as a group of species that can be distinguished by a decided gap to other groups of species.

    The other thing I notice in several threads on this is some sort of expectation that all genera are equal. There is no requirement for them to be equal, whatever is meant by that. The genus is a monophyletic group of species that are different from another monolphyletic group of species which are its sister taxon. However all the members within it are more similar to each other than any is to the other genus. This is based on multiple lines of evidence, not just molecular or morphology. It will and should also include zoogeography and other features unique to that group. You have to be able to key it out. That is good old taxonomy that includes developing keys.

    Just a few of my thoughts based on comments I have seen. I do not know if the genus in question should be split. I have not looked. I am commenting on the discussion. However if a good study has been made with multiple lines of evidence proposing the split then any good taxonomist will refute it, or accept it. Taxonomy is not really that complicated.

Leave a Reply

Your email address will not be published. Required fields are marked *

Optionally add an image (JPEG only)