Last week, I wrote a post on how the new classification for anoles proposed by Nicholson et al. 2012 might affect long term taxonomic stability for this group. That post generated some discussion, including, most recently, commentary by Kirsten Nicholson herself, explaining some of the reasons her group decided it was time to split up the anoles. Because that conversation is quickly becoming buried in the depths of Anole Annals, I wanted to continue it here on the main page. Also, I wanted to write some more to expand on some of the thoughts that prompted me to post last week, and I hope folks will continue to weigh in (note, the rest of this post will make a lot more sense if you first read my previous post, and the ensuing commentary). I’d like to make several general points:
1 – I think we can all agree that taxonomies should reflect accurate phylogenetic knowledge (I don’t think anyone here would disagree with this – let’s take it as a given that Linnean taxa should not be paraphyletic or polyphyletic).
2 – Beyond accuracy, I’m pretty sure there aren’t any rules governing the type or level of phylogenetic information that “should” be included in a binomial classification. Whether binomial names should reflect deep phylogenetic knowledge or shallow phylogenetic knowledge is a matter of opinion – I’d propose it’s completely subjective.
3 – The amount of phylogenetic information you can communicate in a binomial classification is trivial. It’s hard enough to represent phylogenetic structure across the depths of the entire Linnean hierarchy, and there’s almost no such information in Genus+species epithets. The goal of communicating finely resolved phylogenetic information probably should not be our main concern when dealing with genus names (so long as they are not phylogenetically inaccurate).
4 – Given that no one’s to say who’s right about what the appropriate phylogenetic scale of a genus is, and that Latin binomials are ineffective at communicating much of anything about phylogenetic information anyway, issues of stability are comparatively very important. It’s no small thing to propose a change for 88.6% (n = 343) of the scientific names of a group of species studied by thousands of people.
Since I think we can all agree that Linnean names should be applied solely to monophyletic groups, I’ll start with my second point, which is that there’s no “right amount” of phylogenetic knowledge that should be expressed in binomial names. Kirsten suggested we might all agree that “our classifications should mirror our phylogenetic knowledge.” I certainly agree with this statement in general, but I suspect I might disagree on some of the details. What sort of classification, exactly, would mirror our knowledge best? Should we assign genera to the smallest phylogenetic units about which we can be reasonably certain of monophyly, and continue to split them up into additional genera as we work towards resolving the entire bifurcating history of anoles? If so, we’ll probably eventually end up with a taxonomy that’s as finely parsed as that of the non-avian dinosaurs, where the genus:species ratio is 1:2 (I’m not even kidding – check it out…)! At this point, we’d have all sorts of cool binomials, like Nicholsonolis annectans and Mahlerolis gorgonae, but the genera would be functionally equivalent to species names (as they are in dinosaurs). This sort of reasoning (taxonomy should reflect ever-improving phylogenetic knowledge) is implied in the very title of the Nicholson paper, which seems to suggest that periodically, when the progress of systematics has advanced enough, “it is time” to reclassify everything (I think this contradicts the founding purpose of Linnean classification, but that’s another point). Anyhow, if this isn’t what it means for a taxonomy to mirror phylogenetic knowledge, then what exactly does that mean? Why 8 genera, and not 60, 16, or 4?
My main point here is that it’s a matter of opinion what kind of phylogenetic knowledge should be in a Latin name. One person might think that a genus should apply to the MRCA and all descendants of any two species similar enough to be confused by an experienced herpetologist (e.g., Anolis fraseri and A. biporcatus; see Williams 1966 for details). Another person might maintain that a genus should have 20 species max, no exceptions. Both are entirely matters of opinion, and such opinions abound when it comes to systematics. But since there are no official guidelines on the matter, I don’t think that such opinions can suffice to justify a disruptive taxonomic change.
I next wanted to criticize the logic of amending genus names to reflect systematic developments. The reasons are that (a) there’s very little phylogenetic information in Latin binomials, and (b) any change in the names of genera will result in a gain of some phylogenetic information (shallower information) at the expense of other phylogenetic information (the deeper stuff).
Linnean binomials contain next to no phylogenetic information. When we look at a list of scientific names, all we know is that congeners are more closely related to one another than they are to members of other genera, and other than that, they don’t tell us anything about phylogeny. To illustrate this, I created “binomial phylogenetic trees” for the Iguanidae (or Iguania, depending on who you follow..). I included all species in “Iguania” from the Reptile Database. Here’s what the traditional classification looks like, with Anolis highlighted in red:
Binomial “phylogenetic tree” of iguanian genera, following the traditional classification.
What we can see is a perfect basal polytomy giving rise to 47 lineages that vary somewhat in their species diversities. Within the genera, we of course have more polytomies. Of the tiny amount of information we have here, one thing we actually CAN tell is that a few groups like Liolaemus, Sceloporus, Stenocercus, and Anolis are pretty darned diverse, while others are relatively species poor. So this binomial classification is not totally useless – it correctly identifies most of the large radiations of similar-looking relatives – but it’s not much more informative than that.
Now let’s look at our binomial phylogeny from the Nicholson et al. 2012 classification:
Binomial “phylogenetic tree” of iguanian genera, following the Nicholson et al. 2012 classification.
The first thing we notice is that the genera in this topology are a bit more even in species numbers. We’re a bit closer to having no information about relative diversities at all (which would be the case if all genera had the same number of species). Still, it looks like Norops is a pretty spectacular group. It’s not as diverse as Liolaemus, but it’s pretty close… However, if we weren’t herpetologists, we’d probably be surprised to learn that some of the Norops are nearly indistinguishable from some of the Ctenonotus and Audantia (you’d want a loupe and a very good scientific key). Note that in this phylogeny the anoles aren’t all together. The reason, of course, is that there’s no information in this binomial classification to tell us that anoles are more closely related to one another than any of them are to, say, Phrynosoma. We’ve lost that information and gained information about the distinctiveness of 8 anole subclades (personally I’d argue that they’re pretty similar to one another when compared to most iguanian genera, but that’s just my opinion).
Anyway, I think these trees show that it’s a fallacy to think that one valid classification is more phylogenetically informative than another. Any lumper is going to remove tipward phylogenetic information from a binomial classification, probably to the dismay of those partial to that sort of information. Any splitter is going to sacrifice deeper phylogenetic information from a binomial classification, probably to the horror of people who think that information is pretty useful. But I think the more important message is that binomial names are extraordinarily thin on phylogenetic information under any circumstances. Given their utility as handles for communicating about species (across time, not just in the present), I don’t think we should upend stable taxonomies to better mirror what some people (but not others) consider important phylogenetic knowledge.
When a small number of workers propose a massive, unnecessary taxonomic change on the basis of a subjective preference for a certain type of phylogenetic information, I think it’s reasonable for others to challenge that change for the sake of stability. Stability is what makes taxonomies work, and I think that to the extent possible, it should be protected unless taxonomies inaccurately represent phylogenetic relationships.
KIRSTEN E. NICHOLSON, BRIAN I. CROTHER, CRAIG GUYER & JAY M. SAVAGE (2012). It is time for a new classification of anoles (Squamata: Dactyloidae) Zootaxa, 3477, 1-108
ERNEST E. WILLIAMS (1966). South American anoles: Anolis biporcatus and Anolis fraseri (Sauria, Iguanidae) compared. Breviora, 239, 1–14.
- The “Rediscovery” Of Anolis Proboscis, And The Evolution Of A Viral Internet News Story - October 11, 2013
- Research To Suffer As Chicago’s Field Museum Of Natural History Redefines Its Mission - December 30, 2012
- What’s In A Name?, Part II - September 27, 2012
Skip Lazell
Bravo Luke: I totally agree! Furthermore, when “Norops” was fist named, decades ago, I checked skeletons in MCZ and found several nominal “Norops” whose caudal vertebrae were wholly equivocal in form. Genera are best when absolutely distinct on some obvious morphological character. All Anolis sensu lato look like Anolis to me.
Barnaby
Having multiple genera helps the average person who doesn’t have the Anolis phylyogeny memorized to appreciate how diversity within the clade is partitioned. While it may seem inconvenient as far as immediate concerns regarding “taxonomic stability”, I think that years down the road folks will appreciate the change.
Luke Mahler
Hi Barnaby,
I appreciate that point, but I only think it makes sense if our average person is also endowed with the knowledge that the proposed anole genera form a clade. Otherwise, they won’t have any idea that there’s a big diverse clade of anoles that needs partitioning. This information (that the proposed genera form a clade) isn’t available in the scientific names of species, which are really the only pieces of taxonomic information that we can be assured will always be presented (both to the public as well as to other scientists).
The biggest problem, I think, is that you can’t communicate information about more than a single level of phylogenetic partitioning in a binomial nomenclature – this is one of the things I was hoping to illustrate with the trees above. If we carve up a genus, the new binomials will only reflect that certain groups are different, not that there’s structure within a previously recognized group (since all evidence of that previous group is erased). This is why I think we should think twice about changing binomials on the premise of somehow making them “more informative” about current phylogenetic understanding.
Of course, we can communicate much more about phylogeny by using a variety of clade or rank names (although I think these tend to receive limited use outside of systematics). You might argue that the above logic is oversimplified because systematists aren’t really setting out to reflect phylogenetic knowledge in simple binomials* – these reclassifications tend to be broader than that, and occur in the context of potentially informative new family or subfamily or subgenus (or clade) names.
I would reply that since we clearly need to invoke other ranks or clades to really express phylogenetic information, why not proceed in a manner that doesn’t disrupt species names which, as handles, are infinitely more widely used than other taxonomic names? For example, instead of elevating Anolis to family status, and then renaming nearly all of the species, why not leave Anolis as a genus, and subdivide it by assigning series and species groups? (note that the new paper assigns species groups, but not series) The exact same information can be communicated this way: both classifications require several ranks to know anything useful about phylogeny, and both will tell you that Anolis is composed of a bunch of interesting subclades. The only difference is that the former strategy changes most of the scientific names of anoles, and the latter approach leaves them be. This suggestion is not novel – Burnell and Hedges (1990) took exactly this approach when they “reclassified” the anoles more than two decades ago (Kirsten suggested in an earlier comment that Hedges had attempted to change the term Anolis, but this is not technically true – Burnell and Hedges proposed an infrageneric revision of names for anole clades, but did not propose any changes to Latin names).
I’m not sure people will appreciate these changes later. I’m currently collaborating with some folks on a plant diversification study that has us poring over old literature for basic trait data. I’m a herpetologist and know embarrassingly little about plant taxonomy. One thing I most definitely do not appreciate though is the fact that all of these plants (which are new to me in the first place) have had their scientific names changed at least once, and sometimes over half a dozen times during the last century.
*note, however, that this is a common defense for renaming binomials.
KRISTI L. BURNELL & S. BLAIR HEDGES (1990) Relationships of West Indian Anolis (Sauria: Iguanidae): an approach using slow-evolving protein loci. Caribbean Journal of Science, 26(1-2), 7–30.
Barnaby
Also, where was the outcry when Caribbean Mabuya were (needlessly) split into many genera (https://www.anoleannals.org/2012/05/11/a-shiny-new-example-of-caribbean-lizard-diversity/)?
Luke Mahler
Good point. That situation was similar in that the new taxonomy didn’t rectify any problems of paraphyly or polyphyly in the old taxonomy (unless I’m mistaken about Mabuya phylogenetics, which is possible). I think the lack of outcry was probably because the change didn’t affect anoles! There are other potential factors too: by the time Hedges and Conn split the Caribbean skinks, the genus Mabuya had already been significantly reduced in diversity by previous splittings to just 26 species (for what it’s worth, these splittings were controversial when proposed – see the intro of Miralles and Carranza for a brief review). That change, while of the same nature, undoubtedly impacted far fewer workers. Also, Hedges and Conn revised the genus in the context of a massive discovery of new unrecognized diversity, much of which is already extinct, and the taxonomic revision was a relatively minor focus of the paper – it may have simply been overlooked by a lot of folks (although Jonathan clearly mentioned it in his post).
I think this is an apt comparison, although the scientific communities and the taxonomic histories of Anolis and Caribbean Mabuya are pretty different. But I for one was perfectly happy back before 2002 – before the group was split up at all. Back then, Mabuya was the only genus of skinks with a circumtropical distribution (now, of course, there are none). Those skinks are incredible dispersers, and the name and distribution did a decent job reflecting that.
Hedges, S. B., & Conn, C. E. (2012). A new skink fauna from Caribbean islands (Squamata, Mabuyidae, Mabuyinae). Zootaxa, 3288, 1–244.
Miralles, A., & Carranza, S. (2010). Systematics and biogeography of the Neotropical genus Mabuya, with special emphasis on the Amazonian skink Mabuya nigropunctata (Reptilia, Scincidae). Molecular Phylogenetics and Evolution, 54(3), 857–869.
Marc Tollis
Luke, this is a terrific post. Robust in analysis, yet not peer reviewed while open to criticism on the web by other scientists. A good example of the utility of the internet a la Science 2.0. When are you going to start your own blog? :-0
Luke Mahler
Thanks Marc! I’m flattered. Really glad you’re interested.
cybokat
Something that has always bugged me – that a genus is what you make it.
Skip Lazell
More great points, Luke. I am opposed to the splitting of Mabuya too. Subgenera, that folks who care can put in parentheses after the generic name, work fine and those of us who do not want to use them do not have to.
Todd Jackman
A few points need to be considered, that I think are exaggerated by the figures. I can’t imagine that anyone who learns about anoles will not learn about the 8 clades and not realize (or be taught) that they are monophyletic – a shift in names will necessitate an increase in the use of the word anole, rather than Anolis.
Second, in regards to why 8, I again go back to the branch lengths I had mentioned earlier (in my first post on the topic – look at http://todd.jackman.villanova.edu/AnolisGenomeSupplement2011.pdf ). Starting with the Norops clade, there are eight clades with black dots (well supported). Between the 8 groups, only 3 of 7 of the between group nodes are well supported – 5 clades are involved in a rapid radiation. This is significant because it means you can gain phylogenetic information, and the unknown nature of the between group relationships reflects an aspect of reality. I would really favor splitting the genus if all 8 formed a hard polytomy because you would increase phylogenetic information 8-fold, without losing any between group relationships (all would be unknown).
I am assuming that the Alfoldi et al. relationships are robust and will be reflected by future additions of data.
I am not always in favor of splitting everything up, I think my opinion is that Bufo should have stayed Bufo, for example.
I think that what will happen in this case is a hodgepodge of some workers using the names and others ignoring them – but even that isn’t as hopelessly confusing as it might first seem.
The problem with using series and species groups instead of splitting Anolis, is that they may not be used widely outside of taxonomy and systematics — If Anolis is not split up, then I would like to see some ideas of how series and species groups could be used more widely.
Luke Mahler
Thanks for the comments Todd.
As for whether people in the future would learn that these 8 genera are anoles, and were once called Anolis, I’m not so sure… A lot of people would be taught, but a lot of others might come across the group without any training at all. Harvesting and re-use of data is only becoming more prevalent with each passing year, and there are now countless papers that study big-picture questions using data on organisms that the authors know very little about (think, for example, of all the synthetic work that goes on in think tanks like NESCent and NCEAS). These are the folks that aren’t going to know one genus of lizards from another, just like I don’t know anything about the various legumes that used to belong to the same genus (see above comment). This can matter if there are points in history where species stop being called one thing, and start being called another. Ten years from now, a naive meta-ecologist might be excused for thinking little is known about habitat use for Deiroptyx. Sure, people might start using “anole” more if “Anolis” no longer applies, but the problem is that any future search for “anole” will miss 93% of all “Anolis articles written between 1899 and 2012. For a group with such a rich history of research, this would be a real shame.
As for the 8 groups, what happens if ten years from now we can go one step further, and we have irrefutable evidence for 16 monophyletic groups (2 for each of the 8, let's say). Should we have another round of revision to "gain" more phylogenetic information? (I still maintain that we gain shallow information this way, but that our names lose deep phylogenetic information.) I get the feeling that this type of taxonomic turnover is considered normal or even healthy across much of the systematics community (e.g., the view that if you labor long and hard over the systematics of a group, your reward is that you get to overhaul the nomenclature). I think this is pathological. I'd like to think that some of the papers I write are about animals whose names will still be in use 100 years from now – most of the animals certainly will be!
Anyway, I'm not sure what the solution is for getting non-systematists to use the infrageneric names for Anolis – it’ll probably be an uphill battle. But whether or not they use phylogenetically informative rank names, non-systematist anole researchers have done a fantastic job explicitly incorporating anole phylogenetic information into their work (I’d wager that this is more true for anoles than for nearly any other group). It’s hard to beat that for the influence of systematists on other branches of science!
Martha Munoz
Luke, I heartily agree with the statement that non-systematist anole researchers incorporate phylogenetic information. Many times during talks have I heard Norops, Dactyloa, etc., used in a talk to discuss diversity within subgeneric clades. Most of us in the community are familiar and expound on the various clades within Anolis. We don’t need new genera to communicate phylogenetic information. We already do it!
Thomas Sanger
I also agree. But it seems like our arguments have now reached the point where they are just being restated and rehashed. Most contributors seem to agree that nomenclature is best applied in a manner that conveys phylogenetic information. There are obvious limitations to using Linnean ranks for this purpose, as Luke has so eloquently discussed. Many AA readers also seem to agree that reclassifying “Anolis” will have far reaching consequences and are not in favor of it.
Using a reclassification based strictly on genera just isn’t going to accomplish the stated goals or satisfy the community at large. In my mind our time could be better spent discussing a method that will shed these incompatibilities. Perhaps this flies in the face of the Linnean tradition but it seems like our discussions should move towards developing a method to officially name clades (hmmm, does such a thing already exist?). The Nicholson et al. scheme could then be applied to this, conveying the phylogenetic information everyone is searching for (assuming that the clades are, in fact, monophyletic and the names valid). Rather than continuing a debate that seems riddled with a mix of new ideas and old traditions I think that a pragmatic compromise is possible which will actually accomplish the needs of modern taxonomy. Then again, perhaps this is just the optimistic opinion of a naive observer. I would also accept that.