Wednesday, March 18, 2009

Detonating a T-bomb

A few weeks ago, I wrote a post in which I heartily praised the contributions of amateur/avocational workers to museums and museum collections. So of course, in short order, along came an example of why amateur workers might not be such a good thing, which landed in my email in-box with the electronic equivalent of a thud.

There is, in the far-off land of Oz, a man by the name of Raymond Hoser. Yes, that really is his name - for those of you unfamiliar with the McKenzie brothers, a "hoser" is Canadian slang for something that interferes with or “messes up” something else. Some snake taxonomists would argue that Ray Hoser's name is an apt one. He has carved out a formidable reputation as a professional snake handler - an important job in a country that has more than its fair share of toxic reptiles - and as a crusader against government corruption. But it's his activities as an avocational taxonomist that get his professional peers hissing like a pit full of agitated mambas (this will be the only gratuitous snake simile that I use in this post).

To a non-herpetological taxonomist like me (i.e. one who doesn't have to live with the end results) Hoser's papers are hugely enjoyable. Take, for example, his recent paper in the Australian Journal of Herpetology (owner and editor, R. Hoser) on the taxonomy of rattlesnakes. What taxonomist has not secretly wished to dismiss all those pages of tedious descriptive text with a simple statement that "This paper does not by any means seek to.... provide elaborate descriptions of taxa beyond that deemed necessary to formally resolve the taxonomy and nomenclature of this group of snakes" or better still "detail has been kept to a minimum"? Bored with compiling that bibliography? "In terms of references cited, these have been kept to a bare minimum." And there are few taxonomists out there who could close a paper on the systematics of snakes by expressing their opinion that "children discouraged from interacting with wildlife, including rattlesnakes are more likely to turn to harmful alternatives like drugs, violence and the like."

I'm not a herpetologist, so I'm not going to try and comment on the validity of Hoser's work; if you Google his name you'll find that there are plenty of herpetologists out there that are more than happy to do this. You'll also find that a lot of the commentary is negative. Certainly his buccaneering prose style does not suggest the sort of attention to detail that one would normally expect of a professional taxonomist. However, the current round of Hoser-related controversy led me back to an earlier piece by Brendan Borrell, which appeared in Nature in 2007 (vol 446, pp.253-255). This article touched on Hoser's work while considering whether or not "amateur" taxonomists are "bad for science." One of the most interesting points raised was the question of whether the International Commission on Zoological Nomenclature (ICZN), which sets the ground rules for animal taxonomy, should take a more active role in such matters. At the moment, the ICZN can only adjudicate on issues of nomenclature - i.e. is the name itself valid? As you can imagine, this leads to some fast-paced, cutting-edge scientific debate:

"The purpose of this application, under Article 23.9.3 of the Code, is to conserve the universally accepted name Coprinisphaera Sauer, 1955 for an ichnogenus attributed to dung-beetle brood balls (ichnofamily COPRINISPHAERIDAE). This name is threatened by a rarely used senior subjective synonym Fontanai Roselli, 1939. It is proposed that the name Coprinisphaera is conserved by suppression of Fontanai."

Take that, Fontanai, you threatening senior synonym for dung beetle balls whose beetle is as yet unknown. OK, I'll admit that it's easy to laugh at the work of the Commission, but it does serve a very valuable purpose in policing the massively Byzantine world of naming. Understandably, the last thing the ICZN wants is to be dragged into policing the science that underpins these names. And rightly so, because this is the sort of thing that the taxonomic community should be policing itself. The problem is that, faced with a "rogue" taxonomist who works outside conventional academic structures and has his own journal, complete with ISSN number, the sanctions available to the community are limited. It's possible that efforts currently underway to more tightly define ICZN's requirements for publication will restrict the ability of avocational workers to self-publish, but thus far these efforts are limited to electronic publications.

Here, however, I'm going to take a step back and argue that some "bad" taxonomy may actually be "good" for science.

Now obviously there are some limits to this. But setting aside sea dragons from Utah, Cadborosaurus willsi Bousfield & LeBlond, and Nessiteras rhombopteryx Scott & Rines there may be instances where a less than perfect taxonomic or systematic account can stimulate research, or blow open long-held dogma and open up new lines of investigation. I call these papers "T-bombs."

I'll give you an example from my own area of research, the systematics of phalangerid marsupials (no, no, please don't stop reading here... I promise it will be both relevant and interesting). For those of you who have never had the pleasure of meeting a phalangerid, they are small to medium-sized possums of questionable temperament that occur in Australia, New Guinea, and the surrounding islands. In the mid 1980s, two contrasting studies of the taxonomy of this group were published. The first, by Jim Menzies and John Pernetta (1986), was a model of taxonomic rectitude. Over a period of many years, the authors measured hundreds, if not thousands, of possum skulls. They conducted rigorous statistical analyses of the results, refined the description of existing species and, with considerable conservatism, erected a very limited number of new taxa, mostly at the subspecific level. It was massively thorough and almost unreadable.

By contrast, the 1987 paper by Tim Flannery, Mike Archer, and Gerry Maynes was a wild carnival ride through phalangerid systematics. The authors were based in Australia, while most of the Type specimens of the taxa they were describing were in European institutions. Menzies and Pernetta, faced with the same problem (they were based in Papua New Guinea), took years to complete their study, collecting data on their occasional trips overseas. Flannery et al just ignored the types and worked from the species available in their collection in Sydney, which in one case consisted of a single specimen, or worked from published descriptions. The characters they used were often hard to independently verify (to this day, more than 20 years after I first read their paper, I still don't know what "ventral rim of the orbit overhangs the cheek tooth row" actually looks like) and in some cases were age-related features that were not present in adults - a problem when limited specimen availability means that you have to compare adults with juveniles. They erected, or rather resurrected, 3 new genera, and combined two existing genera for reasons that were never explained. There were no quantitative analyses and in contrast to Menzies and Pernetta, whose study was published in the Journal of Zoology (Series B, 1: 551-618), Flannery et al's paper was in a set of conference proceedings (Archer, M ed. 1987. Possums & Opossums, studies in evolution. Sydney, Surrey Beatty & Sons & the Royal Society of New South Wales) that was very hard to get hold of outside of Australia.

So which do I think was the better paper? Flannery et al, by a mile. Menzies and Pernetta's account was thorough, but hidden underneath that mountain of statistics were some pretty untenable assumptions, starting from their belief that G.H.H. Tate's suprageneric classification of phalangerids, dating back to the 1940s, was basically correct. They omitted any consideration of possums outside of Tate's original "orientalis" group because they decided that the "problem" was definition of species inside this group. Flannery et al looked at all phalangerids, concluding that some of Menzies and Pernetta's Indonesian and New Guinea species were actually more closely related to Australian brushtail possums than they were to the other Indo-New Guinea species. And because Menzies and Pernetta wanted the biggest sample size possible they only measured dimensions that were found in all the specimens they studied - which meant that they had, in effect, done a study of the systematics of possum palates rather than possums.

A cursory read through Menzies and Pernetta's paper made you think that there were no questions left to be answered; you might as well go and have a beer, or several beers - which you would need to recover from the experience of reading it. By contrast, when you started reading Flannery et al, you sat up and started asking questions; going into museum collections and pulling out phalangerid skulls to try and check some of these characters; getting annoyed when you couldn't find them; then finding other characters that supported some of their contentions but not others; then going back to Menzies and Pernetta and beginning to see some of the flaws in that study. All the things, in fact, that I ended up doing as a graduate student. Twenty years on, we still don't really know how many species of phalangerid there are, and we keep finding new ones.

And this, I think, is the most important point. At some level, all taxonomists believe that they are searching for an ultimate truth - the one true tree, the one correct and all-inclusive list of species. The problem is that taxonomy is an artificial construct; it's our attempt to apply discrete systematic descriptors to a fluid and continuously changing natural world. In reality, there are a multiplicity of species and a forest of possible trees. They shift, kaleidoscope-like, according to your choice of characters, or your choice of statistical methodology; the tree-generating algorithms you use, or what species you include. They change when you find a new fossil, or when you discover a previously unrecognized species hidden in a museum collection, or when you look at an old character in a new way, or when you find new types of character. The rules applied to the practice of taxonomy are a necessary evil to stop all of this spiralling into chaos, but it's important that they don't become a straitjacket. And that's why it's a good thing if a T-bomb goes off every now and again.


  1. The crucial difference you are missing is that Flannery et al. DID examine specimens, DID provide data and DID advance knowledge. All Hoser does is throw names around for taxa previously defined by other people's hard work.

  2. Well, well, well, it's 2013 and all Hoser's work has been validated by a little bit of DNA.
    Maybe he knew a little more than you all suspected!