HE stability of the ring chromosomes in Drosophila melanogaster raises T certain questions about the nature of chromosome structure and reduplication, since it seems possible that such chromosomes would be subject to loss by interlocking of daughter chromatids after reduplication. In a set of experiments designed to determine the extent to which newly derived ring chromosomes might be lost at the meiotic divisions, it has been found that the generalization that disjunction of chromatids at the second meiotic division is random does not appear to be completely valid when the chromatids are structurally different. The evidence for such non-randomness is presented below.
EXPERIMENTS INVOLVING ATTACHED X CHROMOSOMESThe type of genetic constitution which regularly manufactures ring chromosomes is that used by SIDOROV, SOKOLOV and TROFIMOV (1935, 1936) and by STURTEVANT and BEADLE (1936) in their demonstration of single crossing over within heterozygous inversions. The essential features of their analyses of the results of crossing over in this type of tetrad are incorporated into figure 1. At the reduction division the X-chromatids separate from the Ychromatids ; the figure shows only the second division segregation for the X-chromatids. The tetrads with no, one, or two exchanges are represented by the symbols EO, E1 or EQ, respectively. The two exchange tetrads (Ez) may involve 2, 3, or 4 strands and are designated as Ez-2s, EQ-3s and E2-4s. It is to be noted that there are two different genetic consequences from the single exchange tetrads (El), and two from each of the three types of two exchange tetrads (Ez-Zs, Ez-3.S and E2-4~). These different possibilities are distinguished by the letters a and b following the tetrad type. Three exchange tetrads are not considered here since, as will be shown below, they are relatively rare and can contribute little to the analysis.From figure 1 it is clear that 50 percent is the maximum frequency with which rings may be expected, regardless of the distribution of tetrads of the different ranks since the no-exchange tetrads give rise only to attached X's and the two-exchange tetrads to 50 percent more attached, X's than rings in the viable X-chromosome-bearing gametes.Neither the work of SIDOROV, SOKOLOV and TROFIMOV nor that of STURTE-VANT and BEADLE suggests any deficiency of the ring class. In the first case,