Salivary chromosome bands and the frequency of crossing over in Drosophila melanogaster.

W I T H the discovery of linkage and crossing over, geneticists acquired a technique for estimating the physical distances between genes on the chromosome. From the first, however, the accuracy of linkage maps has been viewed with reservation. Too many extraneous factors, including temperature (PLOUGH 1917) and maternal age (BRIDGES 1927), were shown to modify linkage values. Furthermore, both crossover suppressors and enhancers were identified. Nevertheless, until independent procedures were developed, recombination values, of necessity, remained the only kind of measure that could be applied to the estimation of gene distances. The detailed cytogenetic analysis of rearranged chromosomes later provided a new kind of map, the chromosome or cytological map (DOBZHANSKY 1929.1930a, 1930b, 1932a, 1932b). At once, a serious discrepancy was noted between the old linkage maps and even the first crude cytological maps: although the linear order of genes was the same in both, the relative distances between them were quite different, especially in the neighborhood of the centromere. Later, with the use of salivary gland chromosomes, cytological maps were greatly refined (PAINTER 1934; BRIDGES 1935, 1938). Genes could be localized in particular bands; gene intervals could be measured in microns. BRIDGES (1937) noted the poor correlation between gene distances as seen in the salivary chromosome and in linkage maps of the right end of chromosome 2R. As a result of his studies, he proposed “coefficients of crossing over” to reflect localized variations in crossing over per unit of chromosome length. Recently, RUDKIN (1965) produced another type of map. By measuring the UV absorption of salivary chromosome bands, he determined the relative DNA content of successive intervals along the X chromosome, each interval being delimited by genes whose cytological locations were accurately known. He showed that the distribution of induced mutations among the intervals correlates closely with the proportion of DNA in each. Further. when the distribution of induced chromosome breaks (data from KAUFMANN 1946) is plotted against DNA content, the correlation is even better than that for induced mutations (LEFEVRE 1969). This clearly implies that the amount of DNA available for breakage and mutation in a given region of the gametic X chromosome is in direct proportion