[Taxacom] Costello on species again: new paper

[Doug Yanega]

 From the perspective of an entomologist, the paper in question does not 
at all reflect the patterns or processes underlying insect taxonomy. It 
would be extremely interesting to see the study parameters re-evaluated 
when limited *exclusively* to arthropods. Never before in history have 
we been collecting such astonishingly large volumes of specimens as we 
are at present (thanks to the development of malaise traps, mercury 
vapor lamps, pan traps, and the like), and the backlog of unprocessed 
arthropod specimens in collections around the world is growing almost 
exponentially as a result. I was witness to a student thesis project in 
Brazil where 6 malaise traps were run continuously for 4 years, and 
generated well over one million specimens; of those, a few thousand (at 
most) were removed and processed, and the rest were neglected, had the 
alcohol evaporate, and were thrown into the trash. You could fill a book 
with other, similar stories - and for all of that, there are freezers 
and shelves in virtually every museum collection, all just waiting for 
someone to sort through them.

As for description, the typical arthropod revisionary monograph will at 
least double the number of known species, unless it deals with 
butterflies, tarantulas, or other commercially-traded taxa (various 
small lineages of beetles and moths, etc.). There are very few insect 
taxa for which we are not presumably still scrambling around at maybe 
10-20% of the extant fauna described, and for my colleagues who work 
with arachnids (including mites) the percentages are undoubtedly much 
worse. Given that well over 90% of the world's extant biodiversity is 
comprised by arthropods (and yes, I know, the argument behind that is 
somewhat circular, since it focuses on undescribed taxa), it seems 
incredibly misleading to write a paper about trends in biodiversity 
research that does not focus specifically on the taxa that comprise 
nearly all biodiversity. The recent paper refers to two insect taxa, 
Chalcidoidea and Ichneumonoidea, which are exceptional in that they 
contain species used in biological control; as such, there are far more 
papers for these taxa that describe small numbers of new species than 
there are monographic revisions, often from authors who are not career 
taxonomists. The sorts of papers involved in describing a new taxon as 
part of biocontrol research will artificially inflate two parameters: 
number of (co-)authors, and the diversity of authors, and deflate 
numbers on species described per unit effort. Those are some of the most 
important parameters used to establish the trends referred to in the 
paper, because the statistics are very sensitive to data points that 
involve fewer species described per paper, and authors whose entire 
"taxonomic output" consists of co-authorship on a tiny handful of 
papers. As such, groups used in biocontrol constitute one of the worst 
possible data sets for attempting to refute the general conclusions 
(which was ostensibly why they were included in the analysis in this 
paper - because people had complained that insects violated the trend).

In effect, the study is confounding patterns in how taxonomy is being 
funded and published with patterns in taxon description. The way the 
statistics are generated, for example, would apparently draw different 
conclusions if author X studies for 10 years and publishes 100 new 
species in a single paper in 2012, as opposed to publishing 10 papers a 
year over that span, each describing one new species. Add to that 
changes in career paths over time (past researchers could "get away" 
with sitting and writing species descriptions full-time, and never spend 
a single hour writing grant proposals, sitting on committees, or 
mentoring students), changes in peer review itself over time, and 
changes in what constitutes character evidence (i.e., use of molecules), 
and it gets a lot more muddled. Consider, for example, that some recent 
papers on Icheumonoidea talk about dozens upon dozens of new species, 
but ID them only by code names linked to gene sequences, and don't give 
formal scientific names. I'm sure those new species weren't included in 
the analysis... and I'm sure that we can expect more papers like that in 
the future. There are TOO MANY trends for some simplisitic metric of 
taxonomic output per unit effort to have ANY real explanatory power. 
Claiming that we have already described the majority of arthropods, 
based on this metric, calculated the way it was, is unbelievable.

Sincerely,

-- 
Doug Yanega      Dept. of Entomology       Entomology Research Museum
Univ. of California, Riverside, CA 92521-0314     skype: dyanega
phone: (951) 827-4315 (disclaimer: opinions are mine, not UCR's)
              http://cache.ucr.edu/~heraty/yanega.html
   "There are some enterprises in which a careful disorderliness
         is the true method" - Herman Melville, Moby Dick, Chap. 82