Recent large-scale experimental population studies have allowed us to probe the dynamics of survival in the "oldest old" of a small number of biological species. The results of these studies add strong support to the validity of a multiphasic survival/mortality model for population survival as has been previously proposed by a number of investigators. In this paper we briefly review some of the problems with the Gompertz survival model, and the issue of locating the region in which the mortality rate might change. We derive some rigorous formulae for bounding the mortality rate change regions and compare the predictions to the available experimental data. We demonstrate that the mortality cut point parameter appears to be directly related to topological properties of the species survival curve, in particular, the inflection time. We conclude by addressing some of the pitfalls of using a mortality cut point model and propose an alternative formulation for a mortality model involving multiphasic mortality dynamics and incorporating a nonspecified upperbound on species lifespan.