A simple (deterministic) population dynamics model is used to examine the l
evel of fishing mortality at which a population is expected to be driven to
extinction (F-crash). Values for F-crash are determined for six marine fis
h species (Cape hake, blue grenadier, blue warehou, school shark, gummy sha
rk, and orange roughy) subject to commercial harvest. The sensitivity of th
e value of F-crash to changing the selectivity pattern of the fishery and t
he relationship between the size of the reproductive component of the popul
ation and subsequent births to allow for depensatory effects is examined. F
-crash is greatest for highly productive species and when the fishery does
not target immature animals. The ratio of F-crash to the fishing mortality
at which maximum sustainable yield is achieved, F-crash is a decreasing fun
ction of the productivity of the population. The possibility of depensation
has little impact on F-MSY but can substantially reduce the ratio F-crash/
F-MSY. A series of stochastic simulations is conducted to assess the probab
ility of detecting that the extent of fishing exceeds F-crash using the cur
rent IUCN A criterion. The results of these simulations are also used to de
termine the probability that the IUCN A criteria will be triggered when fis
hing takes place at F-MSY. The results indicate that there are substantial
probabilities of incorrectly identifying species being harvested at F-MSY a
s being threatened during the 'fishing down' phase and also of not identify
ing species actually at risk of extinction if fishing mortality is not redu
ced.