The Battle Over Anole Classification Continues

For those of you who have been living under a rock for the last two years, here’s the short story. In 2012, Nicholson et al. published a monograph in Zootaxa on anole systematics and evolution that proposed dividing Anolis into eight genera. This paper has been much discussed and criticized in these pages (this might be a good place to start). In turn, Poe replied in Zootaxa with a critique widely considered to be overly aggressive in tone. Now, Nicholson et al. have responded to Poe in a nicely written rebuttal that for the most part clearly draws the lines on the areas of disagreement.

This Zootaxa paper is behind a pay-wall, so I thought it might be worthwhile to provide a summary of the main points of the paper in today’s post. In the next two posts, I will discuss what I see as the major issues moving forward.

I’ll summarize Nicholson et al.’s paper following their sections:


The authors introduce the paper by citing Poe’s critique and stating: “We acknowledge that science benefits from vigorous, intellectual debate, but would have preferred his commentary to be more constructive, objective, and scientifically accurate. We therefore present this rebuttal to explain how Poe erred in characterizing our work, and missed the opportunity to present an alternative comprehensive taxonomy to replace the one against which he argues so strenuously. In this contribution we explain, and correct, Poe’s errors and misrepresentations, and argue that our taxonomy is likely to be adopted because it 1) eliminates the obvious problem that will arise if the family Dactyloidae contains only a single large genus (i.e., that a single genus obscures the evolution and diversity within the group and misrepresents or cloaks it), 2) conforms with the long historical trend of dissecting large, cumbersome groups into smaller sub-units, 3) is consistent with all recent phylogenetic studies for anoles in membership within clades we recognize as genera, and 4) aids in associating these lizards with the ancient land masses that shaped their history.”

I would only comment on this introduction by saying that criticizing Poe for not providing an alternative taxonomy seems off the mark. Poe made very clear what he thinks the preferred taxonomy is—one in which a single genus Anolis is recognized.

Monophyly and Anole Taxonomy

The paper begins by taking on the criticism that some of their eight genera are not monophyletic. They point out that for any group which has been the subject of multiple studies, there always will be some disagreements about clade membership. They review how they handled these disagreements: “We did not always follow one particular analysis or dataset (i.e., only follow the molecular data or only the Bayesian analysis) because, as systematists, we are all aware that there are always shortcomings in both the data and the analyses, especially when considering large, cumbersome groups. We integrated the available information to make these predictions, and these explanations are included in the systematic section for each group. Morphological and molecular data often disagree, and investigators are left to interpret those results.”

Character Diagnoses for Clades

This section rebuts Poe on a rather technical point about whether it is sufficient to cite diagnostic characters for a clade when those characters are identified from only one of many possible equally likely trees. Although an interesting debate, it is of minor significance compared to the bigger issues under discussion.

Recognition of Monophyletic Groups across Studies

This is the most important, and most original, part of the paper. The authors state: “Eight major clades are recovered in all studies that have broadly sampled anole taxa (Alföldi et al. 2011; Jackman et al. 1999; Nicholson et al. 2005; Poe 2004), including Pyron et al’s (2013) monumental reassessment of the Squamata. We classified these clades as separate genera because clade membership is so remarkably consistent among analyses, as is membership in 21 of 22 subgroups that we recognized within these genera (Figures 1–5). There are 12 species out of the 240 included in our combined molecular and morphological analysis that are unstable with respect to generic designation in our molecular-only tree, or other recent molecular-only phylogenies.”

Upon further examination of the 12 problematic species, the authors conclude that only “… 5 species that we placed in the genera Anolis (argenteolus, cyanopleurus, lucius, and spectrum) and Chamaelinorops (christophei) …may potentially, eventually, warrant different generic assignments than those we recommended.” Moreover, in a Note Added in Proof at the very end of the paper, the authors state “The recent analysis of several large datasets leads us now to recommend placing the species christophei into our genus Xiphosurus rather than in Chamaelinorops as we suggested in our 2012 paper.” This point refers to the recent papers by Pyron and Burbrink and Gamble et al., recently discussed in our pages (see comments). The authors conclude on this point: “Poe notes that node support for some of the eight major clades of anoles is weak, and concludes that more data are required to justify them. We continue to argue that the consistent recovery of eight dominant clades of anoles in multiple independent studies is sufficient justification for recognizing eight genera. In our view, the pattern is clear; the accumulation of additional data is going to recover these same eight genera.”

The most novel and important part of this paper is the figures 1-5, which show that, for the most part, the same eight clades have been detected time and time again. Here, for example, is the phylogeny from Jackman et al. (1999):

comparison to Jackman et al

And here’s a comparison to Poe’s own work:

comparison to poe

comparison to poe legend

The authors then respond to Poe’s argument that 194 anole species have never been included in any phylogenetic analysis, so how can they be assigned to a clade? They note that most of these species have been put in the Norops and Dactyloa clades and they argue that it is unlikely that those assignments are incorrect.

Nicholson et al. then argue that it is the trend in modern systematics to subdivide large intercontinental genera into smaller clades, and they basically tell Poe and company to get over it. They argue that other proposals, such as using subgenera or the use of Phylocode-style indented and nested lists, have not caught on.

The authors then argue, as they have from the outset, that it’s important to divide large genera into smaller units to facilitate the identification of evolutionary patterns, stating “When new monophyletic structure is revealed in groups for which such structure was previously unrecognizable, taxonomy should change to incorporate that new information. This process does reveal constructs inherent to the natural world and, therefore, forces us to change the way we train future generations of biologists, design future comparative analyses, and interpret new data.”

The paper concludes by answering Poe’s question of why eight instead of some other number of monophyletic groupings. The answer to me was surprising because I thought it had already been given above when the authors stated that all previous studies had identified the same eight groups. But instead, the paper concludes this way:

“Our approach was made clear first in Guyer and Savage (1986). Starting with that paper and continuing with every phylogeny of anoles published since, the alpha section of Williams’ (1976) influential taxonomy has been demonstrated to be paraphyletic. In our opinion, responsible taxonomy requires revision of the alpha section into the seven monophyletic groups now known to be needed to replace it. Starting with Savage (1973), we have made clear our conclusion that the beta section of Williams (1976) deserves generic status (Norops); we note that no published phylogeny since then causes us to question that choice. Therefore, the seven additional genera that we propose as replacements for the alpha section represent the minimum number of genera needed to eliminate the problem of the previous taxonomy. This change is a much more important decision than Poe implies. We have noted how misinformative Williams’ (1976) taxonomy has (2012) been to those interested in evolutionary ecology (e.g. van Berkum 1986). We note that at national meetings, and in print, Poe has been remarkably unwilling to relinquish the concept of alpha anoles (see Poe 2004). We suspect his devotion to Williams (1976) prevents him from seeing value in our work in the same way that Williams (1989) was blinded in his criticism of Guyer and Savage (1986; see Guyer and Savage 1992). We are confident that those reading Nicholson et al. carefully will realize that we did consider recognizing additional genera and presented reasons why we did not do so. We expect that future revisions will continue what we have begun and welcome that process because we recognize such revision will further increase the information content of anole classification within the Linnean system. Based on what is happening to other taxa all around anoles, we think it is inevitable that Williams’ (1976a, b) concept of Anolis, as modified and promoted by Losos (2009), will, of necessity, be broken into the smaller units that we proposed (see Vences et al. 2013 for support of this conclusion). We remain confident that, once this bridge has been crossed, those who have been so reluctant to accept this change will wonder why they fought so hard to avoid it. As Greene (2001) notes, “[r]ather than hindering biology, increasingly accurate and phylogenetically based taxonomy promotes the study and appreciation of life’s diversity.” We couldn’t agree more.”


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.

17 thoughts on “The Battle Over Anole Classification Continues

  1. Nice article. Healthy debate.

    However, I wish the debate shifted to the taxonomy based on a multilocus species tree instead of the old ND2 tree.

    1. Frank, as you know, the original paper reviewed all published anole papers, and was not based on just the ND2 and tRNA’s etc., but included your paper with Alex, and the data produced in the Green Anole genome paper which was multilocus. Perhaps we could have talked about that more in this rebuttal, but we felt compelled to address several issues in the Poe critique that needed correcting. I think much more discussion needs to happen regarding taxonomic rules, consensus, etc. and perhaps this paper will do that.

  2. I wonder why subgenera are not an acceptable compromise in this case. Nicholson et al. (2014: 113, 2nd paragraph) say “To retain older generic concepts, some have attempted to create taxonomic schemes involving subgenera (e.g., Pauly et al. 2009) or nested indented names (Losos 2009). We find no history of the acceptance of such schemes in the herpetological literature.” Obviously, this is not quite true – there are many authors who do use subgenera, and even emphasize this in the titles of their papers, e.g. Wallach et al. 2009. There are many other examples from the recent literature in the Reptile Database.


    Pauly, G.B., Hillis, D.M. & Cannatella, D.C. (2009) Taxonomic freedom and the role of official lists of species names. Herpetologica, 65, 115–128.

    Wallach, V.; Wüster, W. & Broadley, D.G. (2009) In praise of subgenera: taxonomic status of cobras of the genus Naja Laurenti (Serpentes: Elapidae). Zootaxa 2236: 26–36

    1. So Peter we were referring to wide acceptance, and not everyone seems to be doing that. Personally I was a big fan of what the Phylocode hoped to accomplish. However, it seems to have stalled out and I’m unclear as to why. I’ve asked Kevin de Queiroz about it, and last I heard they were still trying to move forward with their system and a big set of papers applying the code to a variety of taxa, but I haven’t seen that come out yet. That said, I think there’s a great need to have a widespread discussion on classification and taxonomy and find consensus on best practices. There are limitations to the Linnaean system, and I’m not sure yet if I’m completely on board with the Phylocode, but I’m looking forward to seeing more on it.

    2. Nobody doubts that there are limitations to the Linnaean system but subgenera are a pragmatic solution to add information. The Reptile Database has >400 subgenus combinations that were used over the past decade and this is only the tip of the iceberg as we only scan a subset of the herpetological literature. Also, subgenera are more popular with some taxa, e.g. Varanus, chameleons, wall lizards (Lacertidae), and less so with others. My impression is that subgeneric names are only used when authors want to stress the difference between clades. In many if not most contexts they are not even worth mentioning.
      Subgenera are a fine solution with those clades that are hard to distinguish morphologically, and that is certainly true for many anoles. It doesn’t make much sense to me to distinguish genera by single scales or other minute characters, but, hey, I am openly biased, given that my background is in developmental biology: I am sure that most such characters are determined by a few genes or even SNPs.

      1. In my opinion, recognizing the genera proposed by Nicholson et al. as subgenera rather than genera would provide all of the benefits and none of the problems associated with a generic level revision. As Peter points out, this idea has been suggested before, not only by authors of other recent herpetological revisions but also in comments here on Anole Annals and elsewhere. It makes sense and seems to represent a very reasonable compromise between those who would rather leave things as they are and those who would prefer a generic revision.

  3. I’ve read the debate over the splitting of Anolis with great interest. This is the first time I have felt compelled to comment on this blog.

    Respectfully, I wish to contest the characterization of the Poe critique as overly aggressive. Personally, I feel it is one of the most well-written rebuttals I’ve ever read. Humor is a powerful way to make a point. As a good friend recently told me: Didn’t Jonathan Swift teach us that? Perhaps, in this case, some have misconstrued Poe’s application of humor as unnecessary condescension or aggression. I’d argue, instead, that the pop culture references make his paper (and by extension, the entire debate) more accessible to a broader audience. The following is an anecdote. With sample size n=1, no less. But I asked my mom (a non-scientist) to read Poe’s critique. She felt that the popular style of his writing allowed her to grasp his arguments much more clearly. Isn’t this what we, as scientists, need to do a better job of? Don’t we need to be strategic communicators? To speak not only amongst ourselves, but also to the non-scientist users of taxonomy?

    I am curious what the Anole Annals community thinks about this. Was Poe’s tone and prose overly aggressive? Or an example of a writing style we all might do well to emulate? Please comment. I yield the floor.

    1. The Editors of Zootaxa should be ashamed of themselves, and apologise, for allowing the publication of such childish personal attacks.

    2. I think that tone and prose might have better befitted another medium than “Zootaxa” so yes I am wondering about the editorial decision. But then an article like Poe’s might bring some public attention to the journal – that usually publishes articles like this one (random example):

      Zootaxa 3811 (2): 151–184 (3 Jun. 2014)
      Insects found in birds’ nests from Argentina: Coryphistera alaudina Burmeister, 1860 (Aves: Furnariidae), their inquiline birds and mammals, new hosts for Psammolestes coreodes Bergroth, 1911 and Triatoma platensis Neiva, 1913 (Hemiptera: Reduviidae: Triatominae))

Leave a Reply

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

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>

Optionally add an image (JPEG only)