The genetic basis of natural variation. 8. Synthesis of cve polygenic combinations from laboratory strains of Drosophila melanogaster.

RARE gene combinations resulting in posterior crossvein defects can be produced by inbreeding Drosophila melanogaster and other Drosophila species collected in nature (MILKMAN 1960c, 1964; GORDON, SPURWAY and STREET 1939; PREVOSTI 1951, 1952; DUBININ 1948). Are these combinations of genes peculiar to natural populations, or can they be produced by recombination and inbreeding in any heterogeneous stock of flies? It appears from the results to be presented in this paper that cue genes are not as a group peculiar to natural populations; that the allelic differences between two different laboratory strains can serve as a basis for the assembly of new cue combinations. This finding is important in two senses. In the broad sense, we are interested in knowing how many genes are potential cue genes; we wish to know the degree of specificity between genes and phenotype here because it is likely to indicate a general state of affairs. In the narrower sense, the answer will influence the experimental strategy employed in this theatre of investigation. There is a spectrum of possibilities with respect to the genetic basis of certain deviant phenotypes which are rare but unbiquitous in nature and whose frequency can be increased by inbreeding and selection. It may be worthwhile to consider the two extremes of this spectrum as a background for the present report. LERNER (1954) suggested that these phenodeviants (his term) stem from certain homozygous combinations. We might further suppose that although the gene combinations producing a given phenodeviant may be relatively few, practically all genes are potential participants in one effective combination or another. Thus while it is possible to have highly inbred wild-type strains like Oregon-R, the inbreeding of any heterogeneous population is likely to produce some phenodeviants which can lead via selection to true-breeding lines with high penetrance and expressivity. This experimental result would certainly appear to be general for cue on the basis of previous work with natural populations and the work to be presented here. The viewpoint is supported by two major lines of evidence. First. it is easy to produce true-breeding cve strains from the progenies of about 10% of individual wild inseminated females; and second, the cue gene combinations as assembled are often antagonistic, so that a gene which interferes with posterior crossvein formation in one genetic context favors it in another (MILK-