Is F ST obsolete?

Since the introduction of allozyme methods inthe mid 1960s it has been a standard practiceto report Wright's measure of populationsubdivision, F _ST, for surveys ofgenetic variation. Its widespread use hasprovided us with a sense of what values can beexpected in particular situations and how theycan be interpreted. With some theoreticaljustification, F _ST has also beenused to estimate rates of gene flow. Howeverthere are conditions under which F _STis inappropriate for gene flow estimation andcan lead to incorrect or even absurdconclusions. These pitfalls have promptedcritics to suggest that F _ST hasfailed to deliver what its proponents havepromised and should be abandoned. A furtherchallenge has been the development of newmethods that offer even greater promise. Thusit is reasonable to ask if perhaps it is timeto retire F _ST and turn to new andmore powerful methods for the inference of geneflow from genetic markers. Here I will arguethat although gene flow should be estimated bymore powerful approaches whenever practical,F _ST remains a useful measure of theaverage effects of gene flow and will continueto be used for comparative purposes.