SELECTION OF VARYING DIRECTION

Some ideas in population biology developed and promoted by JBS Haldane in the 1950's are recalled. At the time his influence was out of proportion to the volume of his work because of the stimulating effect he had on others.

A possibility discussed by Haldane and Jayakar, that selection varying in direction with time would maintain a polymorphism, is examined in the context of limited populations subject to sampling error. It is concluded that if the selection is random it will not be effective if the population is isolated. In a network of interconnected populations, however, independent fluctuations in selection with time may generate genetic heterogeneity in space which could tend to maintain polymorphism.