Are we improving?

Fossil quality and naming dinosaurs

A major concern for dinosaur experts, and indeed for people who study any group of organisms living or extinct is whether current species lists are accurate or not, and whether estimated error rates of 50% or more in some cases still apply. In other words, are systematists any better at their job now than they were in the past? Should we assume that currently published work is just as likely to be error-ridden as more ancient work?

The type specimen of Yinlong downsi, a plant-eating dinosaur from the Jurassic of China, named by Xu Xing and colleagues in 2006, an example of an excellent type specimen.

In the case of dinosaurs, a new species is named every two weeks. This is the highest rate of naming species ever, and there is a convern that these new species might be illusory. It could be that palaeontologists are producing poor-quality work, perhaps fuelled in part by excessive interest from museums and the media worldwide. There is a risk that undue pressure from funding agencies, or even from scientific journals, might lead palaeontologists to name new species when they do not feel a new name is warranted: it is well known, for example, that the press prefers a story about a ‘new species of dinosaur’, rather than another example of a previously named form. On the other hand, it could be that the new species being reported every few weeks are generally valid, and that there really is a great deal yet to discover about these giant fossil creatures.

To study this question, the type specimen (original, ‘name-bearing’ specimen) of each of the 1401 species of dinosaurs named from 1824 to 2004 inclusive was documented, using standard references, such as Weishampel et al. (2004) and original papers. The type materials were coded as:

  1. Isolated tooth or bone
  2. Collection of isolated teeth and bones from one site
  3. Complete or partial (> 50%) skull
  4. Complete or partial (> 50%) skeleton
  5. Several complete or partial skeletons

The first two categories were designated as ‘poor’, and the remaining three as ‘good’, and the ‘quality ratio’ is simply the proportion of poor to good type specimens used in any time interval.

Quality of type materials of new dinosaur species, plotted by decades; the measure of quality is the ratio of incomplete materials (isolated teeth or bones; collections of individual elements) to complete materials (complete skull(s) or skeleton(s)). The best-fitting line (y = 0.0811 – 0.098) shows highly significant correlation (Spearman’s rho = 0.788; p < 0.0001).

The plot of dinosaur quality against time (above; from Benton 2008a) shows that the quality of type materials of new dinosaurian species has improved steadily since the first dinosaur was named in 1824. The measure of quality of type material (ratio of good to poor specimens) remained below 1.0 until 1960, and after this crossover point more new species of dinosaurs were based on complete skulls or complete skeletons than on less complete materials. In this context, ‘good’ type material consists of a complete or partial skull, or a complete or partial skeleton, where ‘partial’ means that at least 50 per cent of the skull or skeleton is preserved. ‘Poor’ type material consists of anything from a single tooth to a collection of 10 or 15 isolated elements from different regions of the skull and skeleton.

What does all this mean?

It looks as if the current huge crop of new dinosaurs is probably reliable. There is no evidence that palaeontologists are rushing new dinosaur names into print at an excessive rate; indeed this study has shown that taxonomic practice has measurably improved through time, and apparently continues to improve.

The historical rate of error in dinosaur species names, more than 50 per cent, may not apply to the species named more recently because the quality of type materials has improved. Indeed, evidence of past invalidation tends to support this. Of the 726 currently invalid dinosaurian species, 582 (80.2%) were based on isolated teeth and bones, whereas only 247 of the 675 currently valid dinosaurian species (36.6%) were based on such limited materials. As the use of incomplete specimens declines, past performance suggests that dinosaurian systematists ought to be establishing a higher propor tion of valid species now than they did in the past.

There is likely to be a limit to the number of new dinosaurs to be named at some point in the near future, however. By using various approaches, Dodson (1990) estimated that there had been perhaps 1000 genera of dinosaurs (which scales to about 1200 species), a figure revised upward to 1850 genera (and perhaps 2220 species) by Wang & Dodson (2007). These estimates suggest that we have identified approximately one-quarter of known dinosaur species (the 1990 estimate) or approximately one-third (the 2007 estimate).

Further evidence that a limit will be reached at some point is that the number of valid new dinosaur species is correlated with the number of new sedimentary basins from which dinosaurs have been recovered (Benton 2008b). The burst of new discoveries from various geological formations in China, Mongolia and Argentina in particular have fuelled much of the recent rise; there will surely be fewer and fewer such unexplored new basins as time goes on.

It could be claimed that fossils are less well understood than extant taxa. In particular, fossil species are morphospecies, and genetic and breeding tests cannot be carried out: this is not perhaps such a problem as might be thought at first, however, because most systematics of living species is also based on the morphospecies model. Further, it is not clear that taxonomic practice has been uniformly worse (or better) among palaeontologists than among systematists of living taxa. The equivalent synonymy rates for extant and fossil groups suggest that there is not much difference in practice. Among fossil groups, dinosaurs, and the North American mammals presented by Alroy (2002), have been subject to heavy scrutiny. Indeed, there have now been several cycles of revision of the species established in Victorian times, and this intense scrutiny continues. If other, less well studied, fossil or living groups are subjected to the same level of revision, invalidity rates of 50 per cent or more might be found to be more the norm than the figures of 20 to 22 per cent suggested for modern biodiversity studies.

The apparent improvement in the quality of type materials, for dinosaurs at least, suggests that such scrutiny may help the community of active systematists to define better practice in identifying new species. If such improvements are occurring, then this must be factored into the statistical techniques that have been used to estimate ultimate synonymy rates based on historical levels of synonymy (Solow et al. 1995; Alroy 2002).

Read more about the study


  • Alroy, J. 2002. How many named species are valid? Proceedings of the National Academy of Sciences USA 99, 3706-3711.
  • Benton, M. J. 2008a. Fossil quality and naming dinosaurs. Biology Letters 4, 729-732. pdf.
  • Benton, M. J. 2008b. How to find a dinosaur, and the role of synonymy in biodiversity studies. Paleobiology 34, 516-533. pdf.
  • Dodson, P. 1990. Counting dinosaurs, how many kinds were there? Proceedings of the National Academy of Sciences USA 87, 7608-7612.
  • Solow, A. R., Mound, L. A. & Gaston, K. J. 1995. Estimating the rate of synonymy. Systematic Biology 44, 93-96.
  • Wang, S. C. & Dodson, P. 2006. Estimating the diversity of dinosaurs. Proceedings of the National Academy of Sciences USA 103, 13601Ð13605.
  • Weishampel, D. B., P. Dodson, and H. Osmólska, eds. 2004. Dinosauria, 2nd ed. University of California Press, Berkeley.