420 ANISTLTAL REPORT SMITHSONIAN INSTITUTION, 1961 



the kind. Such differences of opinion — which were described in 

 textbooks, taught in universities, and expounded in theses — created a 

 verbal and literary jungle which had to be cleared before any general 

 theory could be discussed. 



Finally, there was one specific objection which no one seemed to 

 make. It was that genes were inferred to exist from breeding experi- 

 ments with "characters"; or more precisely, from studying the in- 

 heritance of differences of character when different parents were 

 crossed. Yet the whole of heredity, of the genotype, was supposed to 

 be made up of genes added together. There was a concealed gap be- 

 tween the analytical or differential gene and the integral genotype. 

 In looking at the chromosomes one could see that they added up to 

 make the nucleus. But their differences, their variations, were visibly 

 of many kinds and degrees. The gap was revealed. How could it be 

 bridged? Not, as it seemed to me, by pretending that it did not exist. 

 It was necessary to work out a system of understanding life in terms 

 of chromosomes, a system independent of experimental breeding, a 

 system which would stand on its own feet. In this view I was 

 strengthened by one man's opinion. "Cytology," Karl Belar said to 

 me in 1928, "should not be the ancilla of genetics." That was just 

 what I thought. 



CROSSING-OVER AND THE CHIASMA 



Belar had shown that mitosis and the chromosomes themselves had 

 a universal character, a character which must underlie the uniformity 

 of development of plants and animals and protista [6]. Here was a 

 great and necessaiy step forward. In genetics, as in geology a cen- 

 tury earlier, uniformity was bound up with evolution. Cytologists 

 and geneticists too, so far as they took the chromosomes seriously, 

 therefore liked to think that the same such universal character was 

 true also of meiosis and underlay some uniformity in heredity. But 

 Drosoj)hila itself, with crossing-over at meiosis in the female but not 

 in the male, faced us with the gravest objection to this view. It was 

 possible to evade the issue for the time being. It was possible to 

 begin with the simplest material offering experimental tests by purely 

 chromosome criteria. For this purpose polyploid plants with large 

 chromosomes, tulips and hyacinths, were admirably fitted. They re- 

 vealed several unexpected principles. The important ones, in the 

 present connection, concern the chiasma [7]. 



At a certam stage in the beginning or "prophase" of meiosis, like 

 chromosomes come together as threads side-by-side in pairs. The 

 association is limited to likes: it is chemically specific; and it was, I 

 found, limited to pairs even when there are three or four of a kind. 

 At a later stage the paired chromosomes reproduce, forming double 



