Is FST obsolete?

Since the introduction of allozyme methods in the mid 1960s it has been a standard practice to report Wright's measure of population subdivision, F-ST, for surveys of genetic variation. Its widespread use has provided us with a sense of what values can be expected in particular situations and how they can be interpreted. With some theoretical justification, F-ST has also been used to estimate rates of gene flow. However there are conditions under which F-ST is inappropriate for gene flow estimation and can lead to incorrect or even absurd conclusions. These pitfalls have prompted critics to suggest that F-ST has failed to deliver what its proponents have promised and should be abandoned. A further challenge has been the development of new methods that offer even greater promise. Thus it is reasonable to ask if perhaps it is time to retire F-ST and turn to new and more powerful methods for the inference of gene flow from genetic markers. Here I will argue that although gene flow should be estimated by more powerful approaches whenever practical, F-ST remains a useful measure of the average effects of gene flow and will continue to be used for comparative purposes.