The Ants of Fiji, and the relative caution of modern taxonomic practice

Eli Sarnat and Evan Economo have a beautiful new paper out on the Ants of Fiji. It’s open-access, too:

Plate 105 from Sarnat & Economo 2012, just one of many clean, clear illustrations.

This study is not the first to cover the myrmecofauna of the Fijian islands. Worth reading, for contrast, is William M. Mann’s 1921 classic paper on the same topic:

http://gap.entclub.org/taxonomists/Mann/1921.pdf

In particular, notice that the 1921 paper is full of new species descriptions, while the newer monograph refrains from describing a single new ant. It’s tempting to think the earlier work cleared most of the descriptions out of the way, accounting for the difference. But chronology is not it at all. Sarnat & Economo include a stack of undescribed ants (see the Poecilomyrmex, for example), so they had ample opportunity to follow Mann’s lead.

Instead, this modern taxonomic caution has become the norm. It’s a cultural change in the intervening 90 years as taxonomists adopted the Darwinian synthesis. Biologists as a group are more focused on underlying evolutionary processes, rather than simply describing observed diversity.

Increasingly, taxonomists leave descriptions of new species to more detailed studies of particular lineages on a global scale, often in conjunction with a phylogeny. Thus, species are described in global monographs focused on particular genera or species groups. We see new species in revisions of the Ants of Genus X, rather than in papers on the Ants of Region Y.


source: Sarnat, E.M. & Economo, E.P. (2012) Ants of Fiji. University of California Publications in Entomology, 132, 1-398. [pdf]

20 thoughts on “The Ants of Fiji, and the relative caution of modern taxonomic practice”

  1. This makes one wonder what is better: new species that are eventually synonymized or species codes that run the risk of being single-study codes. This is especially problematic with articles that don’t show how to identify those species that have only codes, when the specimens don’t have appropriate vouchers, or when it takes years for someone to actually look at them. Those information on the species essentially becomes orphaned and can’t be filed in their proper species files.

  2. Describing species is low status, often tedious work* and requires expertise not necessarily held by those who pursue “underlying evolutionary process”. When there is no reward and often an actual penalty for doing ‘simple descriptive’ work, then there is little reason to expect people to go out of their way to describe biological diversity. Who needs names for a no doubt ephemeral phylogeny anyway? Codes for terminal taxa seem both appropriate and sufficient for something no one will read in 10 years.

    *Well, unless you are describing a dinosaur or hominid fossil – then creativity and high status journals go hand-in-hand.

  3. Henry Hespenheide

    The problem with not naming species known to be undescribed is that anyone who wants to do, for example, an ecological study can only give meaningless “morphospecies” codes that cannot be connected to a scientific name when the species are eventually described. Sarnat & Economo have essetially made their own paper irrelevant because no one can evaluate their ability to determine species. They have, in effect, failed to complete their own study.

  4. Marc "Teleutotje" Van der Stappen

    Sarnat has published already revisions about a few groups of the region. These were included as appendixes in his thesis (608 pages!). This work is a reworked version of his thesis and I think the unnamed species will be described in other articles to come…

  5. Great discussion regarding the modern angst over species descriptions. Evan and I deliberated at length on whether to describe new species in this faunistic revision. Taxonomy, perhaps a bit like politics, is the art of the possible. In the end we decided it was better to publish the work we had completed than to sit on the manuscript — for who knows how long — and hope to find some spare time work on species descriptions. The risk of delay is that despite best intentions, life moves on and new projects with new deadlines pile up and “The Ants of Fiji” starts looking more and more like a retirement project.

    Our solution to Henry’s insightful ‘morphocode dilemma’ was to designate a voucher specimen for each undescribed/undetermined presumptive species. These morphocodes can be tracked for the most recent taxonomy changes. Each voucher specimen was given a unique specimen identifier label on its pin (we use CASENT#’s), and that identifier was presented in the monograph checklist. Photographs of each undescribed/undeteremined species are presented in the monograph and available on Antweb, and each is also included in the key to species of its genus. So even though they don’t have proper scientific names, we made it as easy as we could for folks to identify every single species known from Fiji.

    As it happens, after our postdoc hiatus from Fiji ants, Evan and I now have the funding and the time to pick up where our monograph left off. We are preparing new descriptions of Pristomyrmex, Proceratium and Leptogenys, and are working with the very talented Marek Boroweic on a Fijian Cerapachys revision. Because each of our published morphocodes was assigned a voucher specimen with a CASENT#, future students of Fijian ants will be able to track the current name of that species. For example, CASENT0171144 is listed as the voucher specimen for Pristomyrmex sp. FJ02 in the checklist. If you plug that number into Antweb next year, you will find that it has been described as a new species and has a proper scientific name.

    Our description strategy at this point is to begin by picking off the three taxonomically low-hanging fruit: Pristomyrmex sp. FJ02, Proceratium sp. FJ01 and Leptogenys sp. FJ01. Arguably these are ones we could have described in the monograph without too much extra effort. These are species that (1) belong to genera which have been revised in the relatively recent past, and (2) clearly differ from any previously described species. Because we laid the groundwork in the monograph, they have all been imaged, diagnosed and keyed. Now we just need some time at the scope recording the measurements and coming up with good names.

    The next class of future descriptions are species that clearly belong to endemic Fijian radiations. These include Cerapachys (in prep.), Strumigenys, Camponotus, Poecilomyrma and a few others. (We’ve already done Pheidole and Lordomyrma). Though there are clearly undescribed species among these groups, the interspecific boundaries are often ambiguous because of the archipelago system they are evolving in. Ideally we’d have some molecular evidence to help us out, but regardless these groups will take some time to revise if we aim for names that will endure. At the rate I work, it would probably take me a few years to get these all described.

    The last class of future descriptions are species that belong to taxonomically difficult groups that have not been recently revised (at least in Melanesia). These are the Hypoponera, Ponera, Vollenhovia, Nylanderia and Tapinoma of Fiji. A responsible description of these will require studying all the material (including types) available from at least the Pacific, and possibly the world. It is unlikely that Evan or I will take on a Pacific-wide or global generic revision of any of these, and thus their future description is in the hands of souls braver than we. But when such a hero taxonomist takes up the cause, the hardship will be somewhat ameliorated by the morphocode system and the associated images, text and data provided by the Fiji Ant monograph.

    The fundamental rule of taxonomy is to leave whatever group you work on in better shape than when you found it. Perhaps the taxonomic approach taken for Ants of Fiji was a bit too conservative. Although we didn’t describe any new species, we did make a bunch of synonymies (no more Fijian trinomials!), assign lectotypes, and take care of a few other nomenclatural housekeeping chores.

    The advent of unique specimen identifiers, digital images and websites like Antweb are really changing the rules about what responsible taxonomy looks like. In other words, these ain’t your grandfather’s morphocodes.

  6. I intended to reply to earlier comments with a defense of Eli’s work, as he is not the sort to pursue trendy topics at the expense of taxonomy. But since Eli himself has stopped in with a thoughtful explanation I’ll leave it at that.

    I was trying to convey in the OP that species descriptions in this Fiji survey were preferentially being published in taxon-centered papers (as here) rather than in the faunal list, and that this prioritization of taxon-centered work over region-centered work is part of a shifting taxonomic culture. I did not mean to imply that the descriptions would not be done, or that the authors had little interest in taxonomy.

    In any case, I’m greatly enjoying this discussion.

  7. Nevertheless, there is a worrying trend globally nowadays to not bother naming new taxa, but just voucher a specimen and give it a code name. I can see this leading to future chaos, because the realities don’t quite fit the rationale behind doing it this way … scientific names are for (reliable) communication about taxa – no name, no communication …

    1. This is an argument over whether type I or type II error is preferable, and as such I’m not sure there is a solution. Do we have the authors provide names that will likely prove to be incorrect on further study? Or morphospecies that may never be addressed again?

      1. When a species name is proposed, a hypothesis is also proposed (i.e. that a new species [whatever a species might be] can be recognised by a set of diagnostic characters). No name = no hypothesis. Operational taxonomic units have no official standing and are not formal hypotheses. There are many published ecological studies using OTUs that are essentially untestable because the operational units are too obscure. I can’t see why phylogenetic hypotheses based on vague endpoints would be any better. Sarnat & Economo’s paper may be an exception – they seem to have done everything except propose a formal name and diagnosis for their OTUs – but I think it would have been better science to include the formal names. Some may be wrong, but that is not a bad thing if you believe that science advances through falsification – but maybe I’m being too pre-normative in my understanding of science.

  8. Hi Alex, Do you think you could maybe do a post on what is involved with describing a new ant species? I Doubt I’m the only student entomologist that vists your blog so it would be cool to see some content aimed at the next generation of myrmecologists 🙂

    Keep up the good work!

  9. Notwithstanding the issues discussed above, this is a both gorgeous and very useful piece of work.
    Call me wierd, but I just want to sit down in some comfortable place and spend some prolonged, quality time with it, even though the likelihood of my ever getting to Fiji and making on the ground use of it is diminishing with age!

    1. Hear, hear!

      I particularly appreciated the species accounts. In an ideal world, I think we would all have such nice elevational/habitat distribution data to graph!

  10. @Dave: You make a very good point regarding scientific names as scientific hypotheses that can be accepted or rejected as additional evidence accrues. However, I do not agree that it would have been better science to assign valid names to the morphospecies we used in the paper. I believe to do so would have effectively stopped future research of their taxonomy for at least another generation.

    For example, in our current taxonomic research on the Fijian Cerapachys Marek and I are preparing to sink the following synonymize 2 valid species and 3 morphospecies with C. vitiensis. If we had gone ahead and described those 3 morphospecies in the monograph, I can guarantee that Marek, Evan and I would not be revisiting Cerapachys with the more rigorous research their taxonomy deserves, and instead the world would be left with a sub-par hypotheses for Fijian Cerapachys diversity for a generation or two to come.
    Similarly, I doubt that the 7 Poecilomyrma taxa that Alex highlights in this post will stand up to rigorous taxonomic scrutiny when we get around to those. But had we assigned them valid names in the monograph, it would be case closed until someone else decides to revise the Fijian ant fauna. The 90 years separating Mann’s revision from ours suggests that it might be awhile.

    So perhaps the difference is that we are treating valid scientific names as rigorously tested hypotheses, and we are treating our morphospecies taxa as preliminary hypotheses.

    @Alex: I really like how you juxtapose Mann’s 1921 treatise in which he describes something like 60 new species with our 2012 revision in which we describe…uh…0* new species. It is a stimulating conversation starter. Here’s my off-the-cuff top five reasons for the discrepancy.

    #1. The asterisk after the zero is because I did actually describe 10 or so Fijian species of Lordomyrma and Pheidole — but they were published prior to the monograph.

    #2. Mann didn’t do those pesky morphometric measurements.

    #3. Mann didn’t worry about databasing the hundreds or thousands of specimens he collected (we databased >10,000 specimens which are all available to the public on Antweb).

    #4. Mann only did a handful of crude line drawings to illustrate his ~60 new species (we did 3 specimen photographs per caste per species for 186 species).

    #5. Mann described many infraspecific taxa (sub-species and varieties), often using one or two sentences for the full description.

    Mann’s work was a masterpiece for the time. But as Alex notes, it was a different time. Today’s standards for species descriptions in the ant community are much stricter than they were 100 years ago. Nowadays authors tend to be more concerned about images and illustrations, specimen databasing and the inclusion of molecular evidence — and we are less prone to describe infraspecific taxa.

    I think an important point being raised in this discussion is to what extent the tail is wagging the dog with respect to the technology used for modern species descriptions. All that specimen databasing, molecular analysis, imaging, website development, etc. etc. takes enormous amounts of time. Barry Bolton — arguably the most prolific and respected ant taxonomist of our time — dispenses with these arguably peripheral tasks and is thus free to devote his prodigious efforts to the descriptions themselves.

    1. Well, you two are the authors and in the best position to judge what needed names and what should wait. I was responding more to Alex’s general comment than to your paper per se (it does seem to be admirable). I’m no fan of minimalist descriptions attached to little or no biological information, but you do seem to have an unusually rich trove of data attached to your OTUs. I hope you do find the time and support to work them out.

      Your tail wagging the dog comment may be apt. I suppose my primary concern is that databasing and molecular phylogenies (of OTUs) become the goals, not understanding and documenting biological diversity. Managers and funding bodies tend to love new technologies – they see them as being modern and progressive – but have no interest ‘obsolete’ things like basic taxonomy.

      There are also practical concerns too. So suppose one of your OTUs shows up on the “Buy Live Ants” site that Alex posted on this morning? Or in a quarantine intercept or a backyard in Brisbane? Or someone is worried that their habitat is being destroyed and that the ants need protection? There are lots of reasons for having available names. You’ve already done Lordomyrma and Pheidole and I hope you get the chance to do the rest.

  11. Great job! This is a remarkable, very useful, and also beautiful faunal analysis. The best I’ve seen so far. Eli, wow!!!!!!!!!!

    Concerning the use of morphospecies, I fully agree with Eli’s approach. His “morphospecies codes” might not be valid names, but they are still very useful. They have been imaged, data-based, have unique specimen identifiers, and are available for the public. This is a great foundation for someone who wants to do taxonomy on them.

    I also think that it is a bit over the top to criticize the authors for not describing new species of Camponotus, Ponera, Hypoponera, or Tapinoma. This is not easily done and requires a broader scope than just a regional one.

  12. I must admit that I generally hate the use of operational taxonomic units in publication and agree with Dave: no name means no hypothesis, and it’s generally better to propose a hypothesis that is later proven false then not propose anything (e.g., describe the species and later synonymize them as needed). However, I’m extremely impressed with Eli’s thoroughness, and actually agree with his reasons for not describing all of the morphotypes as new species in the current publication.

  13. Eli and Evan deserve to be congratulated on a super piece of work. The talk of hypotheses baffles me as, in my comprehension, a hypothesis is put forward as a possible explanation and then is tested by a scientific method. But then I am a relatively old man. I fully applaud the presentation of OTU’s supported by clear illustrations and notes. A future worker should have no difficulty in matching specimens. As a cautionary tale, however. I did exactly the same, with accurate drawings and measurements for the some 180 species I encountered in Nigeria in 1974-76. Despite publication and circulation to, at least some of the right places, my work has never been fairly recognised, despite including specimens later reported as holotypes. So, beware of “mini itch”!

  14. Marc "Teleutotje" Van der Stappen

    I think the publication is superb and I also think their action not to give names is good. It is clear from their explanation that they didn’t name them because they still need to/are studying groups they don’t know very well. So, naming “new” species now can create many new synonyms not needed. So wait for the revisions where everything clears up!

    Very good work!

Leave a Reply