Technical Comments

Response to Comment on “Climate, Critters, and Cetaceans: Cenozoic Drivers of the Evolution of Modern Whales”

See allHide authors and affiliations

Science  08 Oct 2010:
Vol. 330, Issue 6001, pp. 178
DOI: 10.1126/science.1190377


Pyenson et al. raise concerns about the correlation we identified between diatom and cetacean diversity through time. Although the issues raised are of investigative interest, they do not invalidate our conclusions. We agree that localized studies combined with global data sets will further our understanding of the factors that have helped shape cetacean evolution.

We showed, based on a comprehensive diversity data set, that cetacean paleodiversity may be explained by diatom diversity and variations in climate (1). Pyenson et al. (2) comment on various ways in which a global data set, especially one stretching back millions of years, can be confounded by a variety of localized patterns and geologic biases. Contrary to their claims, we neither drew any firm conclusions about the role and onset of the Antarctic Circumpolar Current nor considered data on consumer body size or abundance in our study. While appropriately pointing out a complex link between primary productivity and consumer diversity (3, 4), Pyenson et al. oversimplify the issue by ignoring other potentially important variables. Our preferred model always includes climate change as a predictor, suggesting that primary productivity works in conjunction with changes in global temperature. Given that the diversity of extant deep-water cetaceans peaks at temperatures well above those found in regions like the Southern Ocean [e.g., (5)], this might explain why cold regions with high primary productivity sustain relatively fewer species of cetaceans. It is also worth noting again that our list of proposed predictors is by no means comprehensive and that other variables may eventually turn out to play an important role in cetacean evolution.

Although the relationship between species richness and productivity is still under debate, various analyses have indeed found them to be interdependent (6), with a linear relationship observed at larger (regional) scales (7). Furthermore, as suggested by several previous studies (810), marine food web efficiency may be at least partly dependent on the size of the primary producers. Diatoms are a relatively large type of marine phytoplankton (11), and some abundant forms exhibit a colonial habit (8). Both of these traits may facilitate the shortening not only of the Southern Ocean food web with its relatively close trophic linkages but also of any food web of which diatoms form a major part, irrespective of the feeding adaptations of its cetacean constituents.

Pyenson et al. (2) reasonably draw attention to a potential geographic bias in the cetacean diversity data, which is biased toward well-sampled areas like Europe, North America, Japan, and New Zealand. Contrary to their claims, however, our data do not seem grossly biased by localized oversampling (Fig. 1). Furthermore, we note that most modern cetacean genera are widely distributed [e.g., (12)]. Hypothetically, if we did not know anything about the number of extant genera living today and restricted our search to the well-sampled areas mentioned above, we would still be likely to find all extant mysticete genera and about 77% of all extant odontocete genera identified to date [based on species distributions as shown in (12)]. Thus, if the past distribution of cetaceans was anything like the present one, even geographically restricted sampling should provide us with an adequate idea of cetacean paleodiversity. Finally, time-averaging is a virtually unavoidable consequence of temporal binning and may to a degree be beneficial in the context of our study, because it may provide a more complete picture of regional diversity owing to the inclusion of rare taxa often absent from short-term assessments of life assemblages (13, 14). Furthermore, if applied to both predictors and the dependent variable, time-averaging should only become a problem when time bins greatly vary in duration—a bias accounted for by including stage duration in all our models.

Fig. 1

Number of taxon-rich collections (A) and the effects of localized oversampling on cetacean diversity (B). Sampled-in-bin neocete diversity data were downloaded from the Paleobiology Database (17) and the number of genera counted for each collection (A). Diversity curves were then constructed using the full diversity data set (B, black curve), and with collections containing ≥10 (dark gray) and >5 (light gray) genera excluded. As shown in (B), apart from a relatively large drop in diversity around 2 million years ago (which, however, did not impact the overall pattern of the data), exclusion of genus-rich collections does not seem to distort our estimate of cetacean paleodiversity.

Overall, we think that our data represent a reliable overall assessment of neocete diversity through time, as earlier concluded by Uhen and Pyenson in their original assessment of cetacean diversity and subsequent studies (15, 16). Given this, and the interplay of temperature and productivity suggested by our results, we conclude that the issues raised by Pyenson et al. (2) neither preclude nor necessarily contradict our findings and hence do not affect the validity of our original interpretation. We agree with Pyenson et al. that more localized studies will greatly further our understanding of any potential link between productivity and consumer diversity, but also call attention to the second point they made, which is that such studies should be used together with global-scale studies looking at the bigger picture.


View Abstract

Navigate This Article