176 PHYSIOLOGICAL GENETICS 



depending upon the time in development in which the somatic 

 /Nutation takes place This phenomenon is rather frequently 

 found in plants, especially in regard to colors but also affecting 

 other morphological characters. Demerec (19356) counts 63 

 such characters in plants, especially studied by Emerson, Baur, 

 Demerec, and Imai (see Demerec 19356). In animals, only a 

 few cases are known, most of them reported by Demerec (see 

 Demerec 19356) in Drosophila virilis. Here a body color is 

 involved, and the miniature wing. (We mention the important 

 point — see page 212 — that the miniature wing is a complete wing 

 which did not finish the pupal phase by growth of the individual 

 cells, i.e., by a very late process in development.) In all these 

 cases, the existence of rather small mosaic spots shows that the 

 mutant gene acted within the cells in which it was situated. 



There is a fourth type of evidence of the same order. Stern 

 (1936) studied mosaic spots that were produced by somatic 

 crossing over which occurred in small groups of cells. In this 

 case, also, very small spots were involved, which showed genie 

 action within individual cells of such an organ system as the skin. 

 Demerec (1934a) used a somewhat different method. He 

 also bred mosaic flies in which the mosaic spots were the results of 

 somatic segregation. These mosaic spots had been marked by an 

 ingenious method with small deficiencies (lack of a piece of the 

 chromosome at a definite locus); and as they arose by somatic 

 segregation, there were always adjacent twin spots with and 

 without the deficiency. Those with the deficiency will be 

 missing if the deficiency is lethal for the cells that carry it. The 

 size of the spot, of course, measures the time at which somatic 

 segregation occurred. Of 33 deficiencies in the X-chromosome 

 thus tested, all but one were lethal to the cells. This, of course, 

 shows only a general effect and does not contain informa- 

 tion regarding the time of action of a special gene-controlled 

 process. 



These facts certainly show that many gene-controlled reactions 

 may be confined to a single cell, and it is possible that this type 

 of process is typical for some gene-controlled processes that take 

 place late in development and are localized in groups of identical 

 cells, without a pattern within this group. But even in these 

 cases a certain centrifugal action away from the cells containing 

 the mutant genes has been observed. 



