Providing estimates of biodiversity and pointing to the desperate need to improve our efforts in describing and naming a substantial percentage of extant species before these become extinct is definitely worthwhile. But this should be based on an adequate factual and conceptual background. Here are my major reasons for concern in respect to the analysis offered by Moura et al. (PLos Biology 9(8): e1001127).
1. Moura et al. are aware of the consequences on global estimates of biodiversity of the lack of a consensus “definition of species”. The problem, however, is not simply one of “different levels of differentiation” required by specialists of different groups to recognize and diagnose a species (not to “define” it, as incorrectly stated in the paper). Different taxonomic traditions have very often to do with differences among the different groups e.g. in geographic patterns of differentiation or in the amount or nature of genetic exchange through sex, not to mention differences in the way they are sampled in the field or studied under the microscope or through molecular analyses.
2. More worrying is however the implausibility of an equivalence between taxa formally classified at the same rank in current classifications. Is there any reason other than tradition, and the associated nomenclature, for defending the idea that beetles (Coleoptera) and bats (Chiroptera) are in any way “equivalent”, given that both of them are classified as orders? Similarly, e.g., for the weevils (the beetle family Coleoptera) and the willows (the plant family Salicaceae), as for any pair of taxa who are not mutually sister groups in the tree of life. As a consequence, what does any of the numbers used in this analysis actually mean?
3. The Catalogue of Life (CoL), the main global data base on which calculations have been based, is a dynamic repository of information which has been growing over the last ten years but is still far from covering all valid species named to date. When accessed by Moura et al., the CoL registered 1,244,360 species; the currently available version (available at http://www.catalogueoflif...) lists 1,368,009 species; a target of 1,750,000 species - as described at the time - was expected when the project started ten years ago; and as estimated total of 1,922,710 as described by 2009 is given in the document SOS – State of Observed Species. 2010, issued by the International Institute for Species Exploration at the Arizona State University (available at http://species.asu.edu/fi...).
4. An yearly average rate of ca. 6200 new eukaryotic species described over the last 20 years is unacceptably low, even taking into account the burden of synonyms. Detailed estimates for recent years are given by the cited SOS document as well as by Chapman A.D. 2009. Numbers of living species in Australia and the world. Canberra: Australian Government, Department of the Environment, Water, Heritage and the Arts (available at http://www.environment.go... ). Figures are as follows: for 2006, 16969 new species; for 2007, 18516; for 2008, 18225. Truly, Moura et al. “only used data to 2006 to prevent artificial flattening of accumulation curves due to recent discoveries and descriptions not yet entered into databases”, but the high figures I report here for the years 2006 to 2008 are hardly exceptional for the last two decades. I cannot provide detailed and taxonomically global figures, but the fact is certainly informative that, for the animals only, the “Zoological Record” recorded 11132 species described as new in the now remote year 1983 (my own data mining; cf. A. Minelli. 1993. Biological Systematics. London: Chapman & Hall).
5. Moura et al.’s estimate of the effort possibly required to describe “all” existing species ignores the principle of decreasing return. That is, the more species in a group are known, the less predictable and the more time-consuming becomes, on the average, the discovery of the remaining species. This is well known to taxonomists who have worked, within one and the same taxonomic group, with samples from a poorly known fauna or flora and samples from areas with a long tradition of studies on that group. In the first case, literally everything may be new to science and ready for description and naming; in the second case, only a long screening through ‘boring’ series of specimens of already named species may eventually lead to the discovery of an undescribed species.

What about the global estimates eventually obtained by the authors? Despite the conceptual and factual flaws listed above, their estimates are not necessarily much off the mark. Interestingly, the predicted grand total of 8 750 000 species is remarkably close to the figure obtained in 1783 by E.A.W. Zimmermann, who devoted 46 pages of the third volume of his Geographische Geschichte des Menschen und der allgemein verbreiteten vierfüßigen Thiere (Leipzig: Weygand) to “The size of the Animal Kingdom.” In Zimmermann’s times the living organisms were referred to two kingdoms only. For the plants, Zimmermann’s estimate was 175 000 species, for the animals, 7 000 000. Zimmermann certainly underestimated insect biodiversity and likely overstimated the diversity of marine life. His analysis, anyway, is still worth attention, in particular is calculations for plant-eating insects, whose number he estimated as five times as large as the estimated number of plant species: a remarkable ecological reasoning, and a remarkably sensibile estimate, for a scientist who published on this topic 199 years before the famous (and still basically useful) estimates of T.L. Erwin (1982) Tropical forests: their richness in Coleoptera and other arthropod species. Coleopterists Bulletin 36: 74-75.