498 RADIATION BIOLOGY 



ated ^\'ith detectable chromosome changes, by Muller, Valencia, and 

 Hannah (see Muller, 1950). However, since, as previously noted, so 

 many more ring than rod chromosomes which undergo breakage are lost, 

 by twisted restitution and perhaps also by straightening of the ring, the 

 rings recovered in the next generation should not include nearly as high a 

 frequency of restitutions as do the rods. Therefore if the changes classi- 

 fied as gene mutations were to any marked degree associated with resti- 

 tuted breaks, a considerably lower frecjuency of both lethal and visible 

 mutations would be expected among the surviving offspring in the experi- 

 ments with ring chromosomes than in those with rods. The fact that 

 the results are in opposition to this was therefore originally regarded as 

 direct evidence that the great majority of gene mutations arise inde- 

 pendently of breaks. 



This line of reasoning is, however, rendered less secure by a consider- 

 ation of the parallel finding by Muller, J. I. Valencia, and Hannah (cited 

 by Muller, 1950) that the irradiated ring chromosomes which were recov- 

 ered in the next generation not only failed to show a lower frequency of 

 lethal and of visible mutations in general than do the rods but also failed 

 to show a lower frequency of small deletions. Since the deletions cer- 

 tainly result from breaks and since they, nevertheless, show an undimin- 

 ished frequency in the rings, then the fact that the changes classified as 

 gene mutations show an undiminished frequency in rings cannot be taken 

 as direct evidence against their being associated with breaks. 



These results on deletions might, on the one hand, be interpreted as 

 meaning that excised interstitial fragments are less likely to become inter- 

 calated into rings than into rods when the major parts rejoin, owing per- 

 haps to the rings being subject to a mechanical stress which for a time 

 (before they become greatly extended) tends to straighten them out. 

 Deletions in recovered rings would thereby be favored by one influence 

 and hindered by another and so might appear to have been unaffected in 

 frequency. In that case, gene mutations, if associated with one break, 

 would not be subject to the influence which favors the finding of dele- 

 tions, but only to that which hinders it. Hence they would occur with 

 diminished frequency in recovered rings. The fact that they do not 

 occur with diminished frequency would then indicate that they are sel- 

 dom associated with breaks. 



It is possible, alternatively, to interpret the results with deletions as 

 meaning that breaks occur far more frequently than had been suspected. 

 On this view it would be supposed that the losses of irradiated rings, 

 despite their high frequency, were so greatly outnumbered by the cases 

 of nontwisted rejoining of the rings that these losses failed to diminish 

 the frequency of deletions among recovered rings to a perceptible extent. 

 On this alternative, the apparently undiminished frequency of gene muta- 

 tions in the recovered rings would be reconcilable with the idea that they 

 too had been associated with breaks. 



