Speciation rates and duration in primates

In the most recent Journal of Evolutionary Biology, Darren Curnoe and colleagues have a paper titled "Timing and tempo of primate evolution". Here's the abstract:

Published molecular clocks for primates are used to estimate typical divergence times for phylogroups (1.6 Ma), species (3.3 Ma), sister species (2.7 Ma), genera (8.9 Ma) and sister genera (8.6 Ma). Significant median differences exist between major groups (infraorders and superfamilies) for various divergence times. These data are employed to estimate typical maximum duration of speciation. Typical primate values (1.1 Ma) suggest this process to be faster than is characteristic of many vertebrates. However, after considering divergence times for hybridizing congeneric and confamilial primates, this value is likely only to estimate the commencement of prezygotic isolating mechanisms, rather than the completion of reproductive isolation. Thus, speciation typically takes around 1.0 Ma to more than 4.0 Ma to occur, depending on whether prezygotic or post-zygotic isolating mechanisms are emphasized. Typical primate genus age is around 5.3 Ma, but we note differences among major groups. In light of these estimates, the classification of humans and chimpanzees is reconsidered using a molecular yardstick approach. Three taxonomic frameworks may flow from molecular analyses, all of them having major implications for understanding the evolution of humans and chimpanzees (Curnoe et al. 2006:59).

You can only go so far with statistics on divergence times in primates (or any other group). Literally any speciation might be unique. The origin of a species might involve very rapid evolution of intersterility with sister species, because of a rapid chromosome change or some other incompatibility.

But of course, the point of statistics is to show what the range of variation of real events actually is. Fossil hominids might or might not look like other primates, but to argue that they necessarily are unlike other primates requires evidence that we don't have.

Here's Table 4 from the paper, which provides a list of known cases of hybridization between species within a genus, and between genera:

</tr> </table>

Curnoe et al. (2006) conclude that the median speciation time represents the initiation of prezygotic isolation mechanisms, and that complete reproductive isolation might better be approximated by the median genus divergence time -- 5.3 million years.

Finally, in consideration of the "Homo trodlodytes" taxonomic mini-problem (i.e., some people think the human-chimpanzee common ancestor is too recent to place humans and chimpanzees in different genera, in comparison to other primate genera), they conclude:

An important controversy in primate taxonomy at present is the generic classification of humans and chimpanzees (Goodman et al., 1998; Castresana, 2001; Curnoe & Thorne, 2003). To allow us to consider this question, we have estimated median primate genus ages for samples excluding human-chimpanzee divergence times (Table 5). Overall, typical primate genus ages are unaltered, as are estimates for the hominoids. The estimated age of the last common ancestor (LCA) of humans and chimpanzees derived from the fossil record is presently older than 7.0 Ma (Hailie-Selassie, 2001; Pickford & Senut, 2001; WoldeGabriel et al., 2001; Brunet et al., 2002). Molecular estimates place the divergence time at 7.5 Ma or earlier (data in Appendix 1). Both estimates are older than typical primate, catarrhine and hominoid divergence times presented here, and also older than the genus yardstick of Goodman et al. (1998).
Two of us have previously supported the view that humans and chimpanzees should be classified as congeners on the basis of their estimated divergence time (Curnoe & Thorne, 2003; also see Goodman et al., 1998; Watson et al., 2001). However, we now believe this view requires reconsideration. Two other taxonomic frameworks may flow from the molecular yardstick approach. The second and third hypotheses would see humans and chimpanzees retained in separate genera (Pan and Homo), with the LCA of both classified in a separate and as yet undiscovered genus or classified in a genus that would also comprise the first 2.0-4.0 Ma of both the human and chimpanzee lineages.

An interesting suggestion, since the latter case would place the human-chimpanzee LCA in Australopithecus.

References:

Curnoe D, Thorne A, Coate JA. 2006. Timing and tempo of primate speciation. J Evol Biol 19:59-65. Full text (subscription)

GeneraTIS or TIG (Ma)
Congeners
Eulemur 10.43
Saguinus 5.72
Ateles 3.59
Colobus 2.9
Mandrillus 2.7
Papio 1.8
Macaca 5.0
Hylobates 1.31
Pan 2.55
Median</td>2.9
Confamilials
Papio-Theropithecus 4.0
Papio-Macaca 12.05
Cercopithecus-Erythrocebus 8.0
Cercocebus-Mandrillus5.0
Hylobates-Symphalangus6.8
Median6.8