The recent article by Fastovsky et al. (2004) examines the Mesozoic record of dinosaur diversity. For this study they utilized recently published dinosaur distribution data complied by Weishampel et al. (2004). A major concern of Fastovsky et al. (2004) was the dinosaur record during the closing 10 m.y. of the Cretaceous. It is this aspect of their paper on which I am commenting.

The authors conclude: “The database does not support the claim that dinosaur richness was decreasing toward extinction during the ~10 m.y. preceding the Cretaceous-Tertiary boundary” (p. 880). Note that the authors qualified this statement with the phrase “decreasing toward extinction.” This is a subjective opinion separate of the more objective issue of whether there was such a decline. When read carefully it is quite clear that their own data show a decline. For example note their absolute richness for this interval in Table 5 with a decline of 51 genera from late Campanian to 41 genera in the late Maastrichtian. Note a similar pattern for the left hand bars in Figure 5.

The authors do not accept these results but rather introduce what they call “total generic richness.” The authors are not explicit as to how “total generic richness” was obtained, but they do note: “Multiple entries of a genus in the database result from multiple geographic records based upon first-order political subdivisions (states, departments…)” (p. 877). If this is in fact what they did, parsing the data according to political subdivisions can be a useful although crude way of examining relative degrees of spatial distribution (e.g., endemicity, cosmopolitanality), but such groupings are not relevant to questions of taxonomic diversity through time.

Next, the authors introduce a rarefaction analysis, but fail to provide a description of which of the many forms of this technique they used. Tipper (1979) reviewed the problems with a naive application of rarefaction to diversity data, listing four critical assumptions implicit in its use in paleoecological studies. As these problems are not discussed by the authors, they apparently are unaware of the pitfalls of rarefaction as used in paleoecology. This is clear when they apply the unwarranted assumption that the “number of genera expected in a sample” can be estimated “if the [expected] sample had the same richness structure as the smallest-sized sample” (p. 877). This is an erroneous assumption given that they are trying to test for total generic richness through the Mesozoic for notoriously rare dinosaur taxa. In fact, rarefaction estimates are sensitive to both the relative diversity and the evenness of the comparison faunas. Determining which aspect(s) of these factors is/are similar or different between populations using the rarefaction results is impossible, especially when broad discrepancies in the protocols used to obtain samples may exist.

As the authors show, their results are largely based upon the much better North America record. I examined the same Weishampel et al. (2004) data set noted by the authors dealing with North America. I utilized genera in the data set that were identified without qualification (e.g., cf., aff., ?, etc.), apparently similar to that done by the authors. I only used localities for which unambiguous age ranges were provided. By use of these data I am not implying that I agree with all taxonomic and age assignments, but rather wish to show that these data clearly show a decline between the Campanian and Maastrichtian in North America.

In Table 1, I have divided the sampling into the normally recognized time subdivisions (early and late) of the Campanian and Maastrichtian (e.g., Gradstein et al., 1995). Localities identified as “middle” Campanian and Maastrichtian were included with late Campanian and late Maastrichtian, respectively. Although the Campanian is decidedly longer than the Maastrichtian, Fastovsky et al. (2004) show that this is not a factor in differences in dinosaur diversity. The very low early Campanian diversity is almost certainly an artifact, as probably is the early Maastrichtian as well. This interpretation is supported because five genera (Avisaurus, Leptoceratops, Pachycephalosaurus, Pentaceratop, Troodon) are reported from the late Campanian and late Maastrichtian but not the intervening early Maastrichtian. The much better sampled late Campanian and late Maastrichtian intervals show a clear decline from 48 to 32 genera—a 33% drop. As Table 1 shows, this is the case even though there are four more localities and 27 more repeated generic samplings for the late Maastrichtian compared to the late Campanian of North America. Interestingly, if one were to follow the logic of Fastovsky et al. (2004), these data would suggest that the late Campanian is more poorly sampled than the late Maastrichtian, opposite of what they are arguing.

Another method that does not have the potential problems of using unequal intervals of time and is more ecologically meaningful compares the taxonomically richest dinosaurian faunas from the Campanian and Maastrichtian. For the Campanian, this is from the Dinosaur Park Formation, Alberta, Canada, which has 31 genera, while for the Maastrichtian, this is from the Lance Formation, Wyoming, United States, which has 20 genera (Weishampel et al., 2004). This is a 35% decline.

These data speak for themselves. Whichever method is employed, there is clearly a decline of over 30% for dinosaur genera in North America during the ~10 m.y. preceding the Cretaceous-Tertiary boundary. I invite readers to examine these data for themselves.

1GSA Data Repository item 2005064, Table DR1, generic counts of dinosaurs through the Campanian and Maastrichtian of North America, is available online at, or on request from or Documents Secretary, GSA, P.O. Box 9140, Boulder, CO 80301-9140, USA.
Attribution: You must attribute the work in the manner specified by the author or licensor ( but no in any way that suggests that they endorse you or your use of the work).Noncommercial ‒ you may not use this work for commercial purpose.No Derivative works ‒ You may not alter, transform, or build upon this work.Sharing ‒ Individual scientists are hereby granted permission, without fees or further requests to GSA, to use a single figure, a single table, and/or a brief paragraph of text in other subsequent works and to make unlimited photo copies of items in this journal for noncommercial use in classrooms to further education and science.