Loss of electrophoretic variation in serially bottlenecked populations.

Experimental bottleneck lines were generated from a single outbred population of the housefly using one, four, or 16 pairs of flies. These lines were followed through one, three, or five successive founder-flush cycles. After each flush 30 individuals from each line were assayed for electrophoretic variation at four initially polymorphic loci. Electrophoretic variation largely followed neutral expectation, the major exception being that the differentiation among 16-pair lines was greater than expected. All bottleneck sizes showed significant losses in genetic variation compared to the control, particularly for losses in numbers of alleles. These findings contrast theoretical predictions in that few lines and loci would be required to detect historical bottlenecks.