A life-history approach to the late Pleistocene megafaunal extinction.

A major criticism of the "overkill" theory for the late Pleistocene extinction in the Americas has been the seeming implausibility of a relatively small number of humans selectively killing off millions of large-bodied mammals. Critics argue that early Paleoindian hunters had to be extremely selective to have produced the highly size-biased extinction pattern characteristic of this event. Here, we derive a probabilistic extinction model that predicts the extinction risk of mammals at any body mass without invoking selective human harvest. The new model systematically analyzes the variability in life-history characteristics, such as the instantaneous mortality rate, age of first reproduction, and the maximum net reproductive rate. It captures the body size-biased extinction pattern in the late Pleistocene and precisely predicts the percentage of unexpectedly persisting large mammals and extinct small ones. A test with a global late Quaternary mammal database well supports the model. The model also emphasizes that quantitatively analyzing patterns of variability in ecological factors can shed light on diverse behaviors and patterns in nature. From a macro-scale conservation perspective, our model can be modified to predict the fate of biota under the pressures from both climate change and human impacts.