The effect of data provenance on estimates of gestation length in African and Asian colobines.

Abstract

It seems to be commonly accepted that gestation length within the subfamily Colobinae lasts several weeks longer in the African tribe (Colobini) than in the Asian tribe (Presbytini) even though closely related taxa of similar body mass should have similar life histories. Suspecting problems with data provenance to cause the difference, we revisited the published records expecting similar gestation lengths in both tribes if based on vetted, accurate data. We compiled published gestation length data for Colobini and Presbytini, labeling them as "unspecified" (n=16) if the primary reference could not be located, methods were not described, and/or conceptions, the beginning of gestation, were determined based on sporadic observations of mating. If conceptions were determined based on changing hormone levels or patterns of daily mating records, we labeled the data as "accurate" (n=12). We analyzed the ln transformed data in a phylogenetic framework in relation to adult female body mass. In the unspecified dataset, gestation length in the two tribes overlapped extensively and did not differ significantly. However, in the accurate dataset, gestation length was significantly shorter in Colobini (not longer, as previously assumed). Data provenance had a strong impact on the comparison, reversing the relationship in gestation length in the two sister tribes. It remains to be determined why gestation lengths differ, whether, relative to the other primates, Colobini have a shortened gestation or Presbytini a lengthened gestation, and whether similar differences exist in other closely related taxa. Addressing these questions will require additional, broader, comparative analyses.