A major effect locus (Lc-Lc^+) exists in linkage group I concerned with the number of locules in the fruit. Since varieties are known having a wide range of mean locule number, a single locus with two alleles is not a complete account of the genetic basis of locule number. Two possibilities exist: (1) a series of multiple alleles exists at the Lc locus, or (2) differences in the modifier background exist, i.e., other loci are concerned with the trait. An experiment was designed to test the possibility of a multiple allele series. Evidence was at hand indicative of a single gene segregation for locule number between the varieties Devon (modal value, three) and Pan American (modal value, about five). A third variety, Goldball (two-loculed), was considered for testing the possibility of a third allele determining the third phenotype. Gold Ball, Pan American, the hybrid, and the backcross to Pan American were studied. A one to one segregation pattern would be expected from the backcross generation and none was observed. Segregation was noted considerably beyond the recurrent parent for mean locule number. Thus, no evidence was found for the existence of a multiple allele series.
An additional item of interest, determined from other crosses, indicates that the two-loculed phenotype of the variety Gold Ball has a different genetic base from that of other material (e.g., L. pimpinellifolium). Hybrids involving Gold Ball crossed with parents with higher locule number show an intermediate phenotype, rather than the usual dominance of the two-loculed parent.