406 



p. S. MARTIN 



pericnd); by open habitat, i.e., savanna, grassland, tundra, and 

 desert, and by limited range, either insular or on such ecological 

 islands as Alpine meadows surrounded by forest. Conversely, sur- 

 vival was enhanced by small size, forest habitat, and a large range 

 for the species in question. 



O 



UJ 

 N 



CO 



NORWAY RAT 



PRONGHORN 



MAMMOTH 



FOREST 



SAVANNA GRASSLAND DESERT 



BIOCHORE 



Fig. 6. A Late Pleistocene extinction model for New World mammals. 

 This generalized model is intended to indicate the interaction of variables 

 that appear to have been important in determining probability of ex- 

 tinction for any particular species or population. Points beneath the sur- 

 face of the solid lie within the region of high extinction probability; those 

 above the surface lie in the region of high probability of survival. 



The primary question in late Pleistocene extinction revolves 

 around the herbivores (trophic level A2 in Lindeman's system). 

 According to ecological and evolutionary theory it is axiomatic 

 that a reduction in the number of species operating at one level will 

 require reduction at higher trophic levels. Our present task is to 



