MANNER OF PRODUCTION OF MUTATIONS 493 



Convenient for this purpose were the very small, phenotypieally recog- 

 nizable deletions which were know^n to occur in extraordinary abundance 

 in and near the heterochromatic regions of chromosomes (known as scute 

 8, scute SI, and Bar M2) in which by a structural change the loci for 

 certain visible "markers" had been placed near heterochromatin (Muller, 

 Prokofyeva-Belgovskaya, and Raffel, 1937; Belgovsky and Muller, 1937; 

 Belgovsky, 1938). The tests, carried out by Belgovsky in collaboration 

 Avith Muller (Belgovsky, 1939; Muller, 1938) and later confirmed (Muller, 

 1940; Muller and Valencia, unpublished data), showed clearly that the 

 frecjuency of these minute deletions is directly proportional to the dose. 

 At low doses then their frequency must be unusually high compared with 

 that of gross rearrangements. 



More recently Panshin et al. (1946) have reported that, in another 

 structurally changed chromosome, the phenotypic changes observed by 

 them in the regions bordering on and including heterochromatin showed 

 a frequency nearly proportional to the square of the dose, much like gross 

 structural changes. Their material, however, was different from that of 

 the previously cited studies in that the chromosome which they used, 

 known as white-mottled 4, allowed larger deletions to survive than were 

 viable in the earlier experiments. This was because the white-mottled-4 

 chromosome was not closely hemmed in, on the euchromatic side of the 

 region studied, by either the free end of the chromosome (as in scute 8 

 and scute SI) or by a gene whose loss was lethal to the heterozygote (as 

 in Bar M2). In addition, a larger number of phenotypieally observable 

 changes produced in this chromosome may be position effects of gross 

 rearrangements. If in their data attention is confined to changes of a 

 kind especially likely to be minute (represented here by the abundant 

 cases of viable whites, which these investigators took to be gene muta- 

 tions, but which there is good reason to classify as being in large majority 

 composed of minute deletions when this material is used), then it becomes 

 evident that here, too, the minute deletions are nearly proportional to the 

 first power of the dose. 



The interstitial fragments, the deletion of which was recognized through 

 phenotypic effects in the experiments on Drosophila just cited, were so 

 small that they could not have been seen as fragments by cytological 

 observation of chromosomes in mitotic or near-mitotic stages, although 

 in salivary-gland chromosomes their absence would sometimes have been 

 detectable. Pieces large enough for cytological detection in mitotic 

 stages should, however, be capable of formation by the clusters in the 

 tails of electron tracks formed by ordinary X and y rays since these tails 

 attain a length of somewhat over ^-^ /jl. Evidence of the formation of 

 such pieces from the chromosomes of irradiated Tradescantia microspores 

 is to be found in the cytological observations of Rick (1940). These 

 showed that chromosome fragments of about this size, of interstitial deri- 



