Consequences of crossing over in oogonial cells.

N TESTCROSSES it is commonly supposed that each recombination I offspring owes its phenotype to chiasmata in a single primary oocyte which gives rise to no other offspring. This assumption should be questioned in view of the demonstration that crossing over may be induced by X rays in oogonial cells of Drosophila melanogaster ( WHITTINGHILL 1938). Irradiation is known to increase the percentage of recombinations in females in some manner, and in males the same agent is able to induce crossovers which originate largely or exclusively in spermatogonial cells. It is therefore appropriate to consider the various possible effects of crossing over in oogonial cells upon the whole picture of recombinations as found in testcrosses. Any oogonial crossing over would make the basic assumption of a linear relationship between chiasma frequencies and recombination partially incorrect.’ Hence conclusions regarding the quantitative effects of various external and interchromosomal agents upon the process of crossing over may be subject to large errors usually unsuspected. Large changes in the frequency of certain classes of progeny will result from one crossing over in an early gonial cell, and smaller effects will follow an exchange in a later cell division. This lack of a constant proportionality between crossover offspring and inferred chiasmata, so clearly demonstrated by this author in males (1937) and revealed in c3G females (1938), may be reasonably applied also to ordinary heterozygous females. This variable relationship may be visualized in figure 1 by imagining that the pyramids shown there may be of various sizes representing the different numbers of cells formed by mitosis after any gonial exchange and before the initiation of meiosis in the descendant cells. The methods of detecting oogonial crossing over through the smoke screen of more numerous meiotic exchanges will vary depending on whether the oogonial exchanges have occurred early or late. The essential ideas of this paper have been incorporated into the diagrams and tabular material of figure 1. It shows the kinds of effects produced in a testcross by a single gonial crossing over depending on the alternate ways in which the four strands of the tetrad may be distributed to daughter cells. STERN (1936) in an extensive investigation of segregation in somatic tissues has shown that anaphase separations of the two kinds figured may be expected following crossing over during mitosis. Those of type (2) would produce segre-