Molecular systematics and paleobiogeography of the South American sigmodontine rodents [published erratum appears in Mol Biol Evol 1998 Feb;15(2):224]

The murid rodent subfamily Sigmodontinae contains 79 genera which aredistributed throughout the New World. The time of arrival of the firstsigmodontines in South America and the estimated divergence time(s) of thedifferent lineages of South American sigmodontines have been controversialdue to the lack of a good fossil record and the immense number of extantspecies. The "early-arrival hypothesis" states that the sigmodontines musthave arrived in South America no later than the early Miocene, at least 20MYA, in order to account for their vast present-day diversity, whereas the"late-arrival hypothesis" includes the sigmodontines as part of thePlio-Pleistocene Great American Interchange, which occurred approximately3.5 MYA. The phylogenetic relationships among 33 of these genera werereconstructed using mitochondrial DNA (mtDNA) sequence data from the ND3,ND4L, arginine tRNA, and ND4 genes, which we show to be evolving at thesame rate. A molecular clock was calibrated for these genes using publishedfossil dates, and the genetic distances were estimated from the DNAsequences in this study. The molecular clock was used to estimate the datesof the South American sigmodontine origin and the main sigmodontineradiation in order to evaluate the "early-" and "late-arrival" scenarios.We estimate the time of the sigmodontine invasion of South America asbetween approximately 5 and 9 MYA, supporting neither of the scenarios butsuggesting two possible models in which the invading lineage was either (1)ancestral to the oryzomyines, akodonts, and phyllotines or (2) ancestral tothe akodonts and phyllotines and accompanied by the oryzomyines. Thesigmodontine invasion of South America provides an example of the advantageafforded to a lineage by the fortuitous invasion of a previouslyunexploited habitat, in this case an entire continent.