The genetic basis of natural variation. X. Recurrence of cve polygenes.

A L L E L F B with simple phenotypic correlates, notably electrophoretic mobility variants of enzymes, have now been shown in many cases (e.g., LEWONTIN and HUBBY 1966; PRAKASH and LEWONTIN 1968; O’BRIEN and MACINTYRE 1969) to participate in recurrent polymorphism: the occurrence of the same two or more alleles in each of many, widely separated natural populations. Similarly, recurrent polymorphism at the gross phenotypic level is well known (FORD 1965), and it can sometimes be related to polymorphism at a single gene locus. On the other hand, less is known about the great middle ground of natural variation, where phenotypic variation is associated most directly with combinations of alleles rather than with individual alleles. Here, phenotypic variation assumes many common forms, but the genetic basis of each variant may not be constant at all. Some traits, like size and fertility, are sensitive to allelic variation at a vast number of loci; but others, including crossvein defects in Drosophila, seem to monitor a workably small number of genes (MILKMAN 1960a,b). POStenor crossvein defects are rare but ubiquitous in D. melanogaster: they have been seen in the progeny of every natural population in which they have been sought (MILKMAN 1964). Other Drosophila species are similar in this regard ( TIMOF~EFF-RESSOVSKY 1934; DUBININ 1948; PREVOSTI 1951 ) . The crossvein defects observed in the progeny of wild L). melanogaster have a polygenic basis, as demonstrated by the analysis of true-breeding crossveinless strains derived from them (MILKMAN 1962, 1964). Furthermore, three true-breeding strains, each derived from a wild inseminated female caught in a fruit market, all turned out to contain the same cve genes (MILKMAN 196513). There is, however, abundant evidence for the existence of numerous seis of cue polygenes (MILKMAN 1962, 1964, 1965b). Do these sets coexist? What is the duration and area over which a given set can be extracted? The present paper describes the comparative study of 12 true-breeding crossveinless strains, each selected from a progeny of a single female in one of numerous samples taken over a 5-month period from two wild populations. Evidence was sought for a repeated occurrence of cue alleles at the same loci.