412 GENETICS OF SOMATIC CELLS 



or semidominant resistance to harmful agents; selection by starvation (based on the 

 fact that certain cells with two different nutritional requirements survived longer under 

 conditions of starvation than cells with only one requirement) ; and selection of cis 

 from trans heterozygotes, in cases where a strain heteroallelic in trans for a nutritional 

 requirement did not grow on a medium lacking the relevant growth factor, whereas 

 the corresponding cis heterozygote resulting from s.c.o. did, and could therefore be 

 selected. 



Another important requirement for the genetic analysis by s.c.o. was that the 

 markers used for selection of the crossover homozygotes had to be located as distal 

 from the centromere as possible, preferably quite near the tip of the chromosome. 

 Among the homozygotes for such a distal marker, produced by s.c.o., there were some 

 that were also homozygous at the next proximal locus (that is, nearer to the centromere), 

 some at the next two, some at the next three, and so on. By expressing the proportion 

 of homozygotes for any one, two, three, or more segments nearer to the centromere 

 than the selected marker, as a fraction of all homozygotes obtained for the selected 

 markers, it was possible to calculate the proportional incidence of exchanges in each 

 of the corresponding intervals between the centromere and the first marker, the centro- 

 mere and the second marker, and so on up to the selected marker. The centromere 

 itself could be located in chromosomes marked on both arms, by identifying the segment 

 where homozygosis for one of two adjacent markers was always accompanied by homo- 

 zygosis for a series of other markers in one direction, and homozygosis for the other 

 marker was always linked with homozygosis for a series of markers in the other direction. 



Essentially, then, the procedure of genetic analysis by s.c.o. was to select, from a 

 heterozygous diploid strain, segregants that were homozygous or hemizygous for 

 certain distal markers and to analyze subsequently these segregants for their residual 

 genotypes. 



In molds, the occurrence of another important process — haploidization — compli- 

 cates the s.c.o. analysis. Between 10 and 50 per cent of all phenotypically recessive 

 patches selected from heterozygous diploid colonies turned out to be haploid. Somatic 

 crossing over and haploidization result from two quite different processes that occur in 

 different nuclei and do not coincide more often than expected by chance. Haploid 

 segregants can be separated from diploid, however, by measuring the diameter of the 

 conidia or by other methods. 1011 Haploidization turned out to be a very convenient 

 tool for genetic analysis by itself. It yields all possible recombinations between chromo- 

 somes, but practically no recombination between linked markers. As a result, it 

 locates any previously unlocated marker on its appropriate linkage group at once. 



Genetic mapping by s.c.o. proved to be perfectly feasible in the studies on Aspergillus. 

 Since the absolute incidence of s.c.o. was small, difficult to measure, and variable, these 

 maps were not based on recombination fractions as in meiotic mapping, however. 

 The distances had to be expressed in different units for each chromosomal arm. This 

 was done by determining the total number of segregants for the distal selected marker 

 and expressing the incidence of crossing over in each of the intervals between the 



