1. Further studies on A^b.

The factor for shrunken endosperm (sh₂) is reported by E.B. Mains to be in the third linkage group about one‑quarter of one unit from a₁. Because of its favorable location it is being used as a marker gene in the current studies on the compound nature of A^b.

From the cross A^b Sh et/a sh Et x a sh twenty‑four crossovers in the a‑et segment were obtained and survived testing. Seven of these strands carried A^b and all seven were A^b sh Et. Seventeen carried a and of these fifteen were a Sh et and two were a bh Et. These data indicate that sh₂ is distal to A₁, that is, between A₁ and Et.

In past experiments the occurrence of a large majority of A^d (pale) derivatives from A^b/a plants has been associated with crossing over in the a-et segment, the isolated pales being linked with the et allele carried on the a chromosome of the tested plants. It was argued that these crossovers occur within the A^b "locus" the A^d derivative representing a proximal component of A^b. Since sh₂ is distal to A and so close to it, its use in these experiments should allow a localization of this crossover event.

From A^b Sh/a plants 27 A^d cases have been analyzed. Twenty‑three of these were A^d sh and the remaining 4 were A^d Sh; only 1 out of 400 A^d cases are expected to be linked with sh if there is no crossover basis for their origin. Hence this crossover event is localized within a segment less than one unit in length.

The four A^d Sh cases from A^b Sh/a sh plants suggest that A^d may arise without crossing over at the A^b locus. This is confirmed from studies of plants carrying A^b on one chromosome and a deficiency on the homologue and designated A^b/a‑XI the deficient piece is now known to include Sh₂ also). The occurrence of A^d here could not result from crossing over involving homologues. Among 32,000 tested gametes from such plants one A^d case has been found. Analysis has ruled out illegitimate bases for its origin. It could have resulted following mutation of the residual component of A^b to a null form but it is not possible to rule out the alternative exploration of sister chromatid exchange which in this case would have be be unequal.