The PhyloCode and the Names of Anole Clades

I’m posting these remarks at the request of Anole Annals founder Jonathan Losos in light of his suggestion that a proponent of the PhyloCode explain how this system works (with reference to anoles).  As one of the developers of the PhyloCode, as well as a systematic biologist who studies anoles, I guess I’m the logical person to do this.  These issues relate to the recent proposal to “split” Anolis into multiple “genera” following the rules of the Zoological Code (ICZN) in that the PhyloCode (ICPN) describes an alternative system for applying taxon names according to which the very idea of “splitting a genus” has no meaning (hence my use of quotation marks).  The reason is that unlike the Zoological Code, which is based on artificial ranks (e.g., genus, family), the PhyloCode is based on statements about phylogenetic relationships, which means that the PhyloCode ties names directly to clades (monophyletic groups), rather than tying them indirectly and loosely to clades through the intermediary of ranks, as in the case of the Zoological Code.  Clades are evolutionary groups about which scientists can make inferences (regarding properties such as composition, diagnostic characters, and age of origin); they are not things that scientists can “lump” or “split.”  In any case, some of the advantages of the PhyloCode are that names maintain more stable associations with clades, many unnecessary and disruptive name changes that occur under rank-based nomenclature can be avoided, clades can be named one at a time as the evidence permits (rather than requiring large-scale revisions to the taxonomy, many components of which may lack an adequate evidentiary basis), and much more information about phylogenetic relationships can be conveyed (because the system is not artificially constrained by ranks).  In the rest of this post, I’ll illustrate these points using examples involving anoles.

The Fundamental Difference

The fundamental difference between the Zoological Code and the PhyloCode concerns the way in which names are defined in the two systems.  Under the Zoological Code, the name Anolis is effectively defined as follows:  Anolis := [is defined as] the taxon ranked as a genus that contains the species carolinensis.  Now it turns out that no one has defined the name Anolis using the PhyloCode approach, which requires names to be defined explicitly.  The following examples are just two possible ways in which that name could have been defined prior to the proposal to “split” the “genus”:  Anolis := the least inclusive clade containing bimaculatus, lineatus, carolinensis, punctatus, and auratus (some of the species originally included by Daudin) or Anolis := the clade originating in the first ancestor of carolinensis that had adhesive toe pads synapomorphic with those in carolinensis (one of the diagnostic characters originally cited by Daudin).  Note that the PhyloCode style definitions tie the name directly to a clade, while that of the Zoological Code only ties the name to a taxon, which might or might not be a clade, and even if it is a clade, the tie is only indirect through the clade being ranked as a genus.  I also want to point out that PhyloCode methods for applying names are tree-based in that they require phylogenetic trees for determining the limits of the clades to which the names apply.  Although rank-based methods can be applied in the context of trees, they are not inherently tree-based in that first, their implementation doesn’t require trees (taxa can be “erected” however the taxonomist chooses), and second, the names are more strongly tied to artificial ranks (in this case the “genus”) than they are to any of the monophyletic groups (clades) implied by a tree.

Associations between Names and Clades

As a consequence of the indirect (and thus weaker) tie between names and clades under the Zoological Code, names governed by that code do not have stable associations with clades.  This should be obvious from the fact that the name Anolis is associated with a relatively large clade of ca. 385 (currently recognized extant) species according to the current widely accepted taxonomy, but that name is to be associated with a relatively small clade of ca. 44 species according to the proposed “split.”  By contrast, under the PhyloCode, names have more stable associations with clades.  Thus, if we were to adopt either of the phylogenetic definitions of the name Anolis described in the previous section, that name would apply to the same large clade of ca. 385 species under both the phylogeny of Poe (2004: Figs. 1–4), who treated the entire clade as a “genus,” and that of Nicholson et al. (2012: Fig. 4), who propose to “split” the “genus.”  The reason is that the name is defined as referring to a particular clade independent of arbitrary rank assignments (note that the phylogenetic definitions make no references to ranks).  In addition, any changes concerning hypothesized species composition under the PhyloCode can result only from revised phylogenetic inferences (i.e., new scientific results); they cannot result from artificial and non-scientific decisions to change ranks (whether a particular clade is a “genus” is not a scientific hypothesis).  Thus, if we were to adopt either of the phylogenetic definitions of the name Anolis described in the previous section, the phylogenies of both Poe and Nicholson et al. lead unambiguously to the conclusion that Anolis includes the species formerly referred to the “genera” Chamaeleolis, Chamaelinorops, and Phenacosaurus.  But this does not mean that those names must be “synonymized” with Anolis, as they would be under the rank-based Zoological Code.  Instead, the name Chamaeleolis can continue to be applied to the clade of giant twig anoles including Anolis chamaeleonides and it close relatives (rather than adopting the new and cumbersome name “Xiphosurus chamaeleonides species group” of Nicholson et al.).  Similarly, the name Chamaelinorops can continue to be applied to the clade of anoles with certain distinctive vertebral modifications that is currently considered to include only the single extant species Anolis barbouri (rather than applying that name to a larger clade including 8 other species that do not possess those vertebral modifications and were not previously included in Chamaelinorops, as Nicholson et al. were obligated to do by the rank-based Zoological Code when they chose to rank that clade as a “genus”).

Unnecessary and Disruptive Name Changes

Because names under the PhyloCode are tied to clades rather than to ranks, this system avoids unnecessary and disruptive name changes of the sort that result from artificial and non-scientific decisions to change ranks under the Zoological Code.  As we saw in the previous section, the name Anolis as defined using either of the example phylogenetic definitions (which reflect the way in which the name has traditionally been used) applies to the large clade of ca. 385 species in the context of the phylogeny of Nicholson et al. (2012: Fig. 4).  That phylogeny cannot therefore be used to justify changing the reference of the name Anolis to a smaller clade composed of only ca. 44 species.  Consequently, there is no need to change the names (binominal combinations) of the other 385 – 44 = 341 species, which (as other posts to the Anole Annals have argued) would be highly disruptive.  The only reason that we even consider haphazardly changing a name from one clade to another is that we take the rank-based approach for granted.  That approach inappropriately places greater emphasis on the associations of names (e.g., Anolis) with particular ranks (e.g., the “genus”) than with particular clades (e.g., the one composed of ca. 385 currently recognized species with which the name Anolis was previously associated).

Evidence for Named Clades

Another advantage of the PhyloCode is that it permits clades to be named one at a time as the evidence permits.  Under the Zoological Code, names are tied to ranks, and taxonomic convention dictates that all species must be assigned to a taxon at the rank of “genus.”  This situation often forces taxonomists to name putative clades that are not justified by the evidence.  For example, under the traditional approach, if one is going to recognize Norops as a “genus“ separate from Anolis, then all of the other species formerly referred to the “genus” Anolis have to be assigned to mutually exclusive taxa at the rank of “genus.”  Thus, Nicholson et al. propose recognizing seven other mutually exclusive putative clades on their molecular tree (their Fig. 4) as “genera,” which is the minimum number that is sufficient to cover all of the other species.  The problem is that several of these putative clades are contradicted by the results of their own combined analysis (their Fig. 5), and others are only weakly supported.  By contrast, because the PhyloCode operates independently of taxonomic ranks, it is not bound by taxonomic conventions associated with (artificial) ranks, thus allowing clades to be named one at a time as the evidence permits.  Interestingly, this has already been going on in anole taxonomy, where several anole biologists have adopted the general PhyloCode approach (even if they didn’t always adopt formal phylogenetic definitions).  For example, Jackman et al. (1999) noted that the names Chamaeleolis, Chamaelinorops, and Phenacosaurus could be applied to subclades of Anolis.  Ironically, Nicholson (2002), who is also the first author of the 2012 publication proposing to “split” Anolis, treated Norops as the name of a clade within Anolis (rather than as a separate “genus”).  Similarly, Brandley and de Queiroz (2004) applied the name Ctenonotus to a clade within Anolis including distichus, bimaculatus, cristatellus, and their relatives but excluding cybotes and its relatives.  By contrast, Guyer and Savage (1987) had placed all of these taxa in a “genus” of the same name even though their results did not support its monophyly (another example of the Zoological Code encouraging taxonomists to name groups that are not justified by the available evidence).  Finally, Castañeda and de Queiroz (in press) apply the name Dactyloa to another clade within Anolis (in this case using an explicit phylogenetic definition).  The point is that in all of these cases names were applied to individual clades supported by evidence in the studies in question.  Because those names were applied independent of ranks, the authors were not compelled to name putative clades for which they did not have adequate support to satisfy the convention that all of the other anole species would be assigned to taxa at the same “rank.”

Conveying Phylogenetic Information

As Jonathan pointed out, recognizing one “genus” versus eight “genera” of anoles each conveys phylogenetic information not conveyed by the other.  More importantly, neither conveys anything close to all of the phylogenetic information in the anole tree, which has ca. 384 potentially nameable clades composed of more than one currently recognized species.  Nor does either convey the information corresponding to the subset of those clades that are well supported or otherwise useful to name (e.g., because of distinctive apomorphies).  Another advantage of the PhyloCode is that it is not constrained by ranks and therefore is not limited to a choice between representing information in one versus eight “genera” (or any other number of mutually exclusive clades).  Clades of any size or age can be named wherever warranted by the available evidence without the artificial constraint that every species in some larger clade must be assigned to a taxon at “the same level” (in this case, the “genus”) and thus without disrupting the existing taxonomy (e.g., with large numbers of new binominal combinations).  Thus, even if all ca. 385 species are considered to belong to a clade named Anolis, some of them can also belong to clades named Chamaeleolis, Chamaelinorops, Ctenonotus, Dactyloa, Norops, Phenacosaurus, and others.  Moreover, rather than being limited to use of a “genus” name or a binomen, one can use as many or as few names as are necessary to convey the relevant phylogenetic information.  Thus, in some contexts, the name Anolis/inderenae may be adequate (advocates of the PhyloCode often use slashes to indicate nested relationships), in others, Dactyloa/inderenae would be appropriate (without implying that inderenae is not part of Anolis), and if necessary, one could use Anolis/Dactyloa/Phenacosaurus/inderenae.  In short, the PhyloCode approach provides simple and intuitive methods for conveying much more phylogenetic information than can names at any single “rank” and without the artificial constraints that result in many names being used (in the context of the rank-based Zoological Code) to convey dubious or weakly supported phylogenetic information.

Concluding Remarks

It’s unfortunate that the nomenclatural aspects of taxonomy have lagged so far behind the rest of biology that we’re still arguing about such archaic notions as “splitting genera.”  Tree-based methods are now standard in all other areas of evolutionary and comparative biology, and they’ve been available in biological nomenclature for over 20 years (even if they weren’t at first put into a formal code) (de Queiroz and Gauthier, 1990), yet taxonomists still employ outdated rank-based methods for applying names.  As a consequence, those taxonomists end up revising taxonomies in ways that result in disruptive and unnecessary name changes, the formal recognition of dubious or poorly supported taxa, and the misconception that by doing so they have somehow improved the taxonomy.  We can only hope that taxonomists will eventually abandon those outdated methods in favor of ones that promote the stable associations of names with clades, the ability to name clades one or a few at a time when justified by the available evidence, and the true and largely incremental improvement of taxonomies with minimal fanfare resulting from studies that emphasize phylogenies rather than taxonomies.  When this happens, we will finally stop wasting our time arguing about artificial notions such as the numbers of “genera” or “families” that ought to be recognized and focus our attention on real phylogenetic problems and the evolutionary inferences that depend on them.  Let’s hope that the case of anole taxonomy will help to move things in that direction.

References

Brandley, M. C., & K. de Queiroz (2004). Phylogeny, ecomorphological evolution, and historical biogeography of the Anolis cristatellus series. Herpetological Monographs 18:90–126.

Cantino, P. D.,  & K. de Queiroz (2010). International Code of phylogenetic nomenclature [PhyloCode].  Verson 4c.  International Society for Phylogenetic Nomenclature.  http://www.ohio.edu/phylocode/

Castañeda, M. del R., & K. de Queiroz (in press).  Phylogeny of the Dactyloa clade of Anolis lizards, insights from combining morphological and molecular data.  Bulletin of the Museum of Comparative Zoology.

de Queiroz, K., & J. Gauthier (1990). Phylogeny as a central principle in taxonomy:  Phylogenetic definitions of taxon names. Systematic Zoology 39:307–322.

Guyer, C., & J. M. Savage (1987). Cladistic relationships among anoles (Sauria: Iguanidae). Systematic Zoology 35:509–531.

Jackman, T. R., A. Larson, K. de Queiroz, & J. B. Losos (1999).  Phylogenetic relationships and tempo of early diversification in Anolis lizards.  Systematic Biology 48:254–285.

Poe, S. (2004). Phylogeny of anoles. Herpetological Monographs 18:37–89.

Nicholson, K. E. (2002). Phylogenetic analysis and a test of the current infrageneric classification of Norops (beta Anolis).  Herpetological Monographs 16:93–120.

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.

• Author
• Recent Posts

Kevin de Queiroz

Research Zoologist (Herpetologist) at the National Museum of Natural History, Smithsonian Institution.

Latest posts by Kevin de Queiroz (see all)

• Happy Thanksgiving - November 22, 2012
• The PhyloCode and the Names of Anole Clades - October 4, 2012
• Anole Poetry - July 11, 2012