MANNER OF PRODUCTION OF MUTATIONS 517 



remain affected for some time as a result of irradiation itself. This possi- 

 bility is indeed suggested by the effects of this kind which have been 

 found for nonionizing radiation and for mustard gas (to be referred to 

 later) . 



8. SPATIAL LIMITATION AS EVIDENCED BY THE EFFECTS 

 OF ION DISTRIBUTION 



Finer scale evidence concerning the degree of spatial localization of the 

 mutagenic process than that given on pp. 507-511 is furnished by studies 

 of the effects of differences in specific ionization. For example, as already 

 noted on pp. 488-490, the finding that there is a higher observable fre- 

 quency of double breakage, involving both of two adjoining sister chroma- 

 tids of Tradescantia pollen tubes at adjoining points, when 4.1 A X rays 

 are used than when the radiation is harder or softer (Catcheside and Lea, 

 1943), can only be interpreted reasonably (regardless of the explanation of 

 the other matters at issue in the same experiments) as being a result of the 

 known fact that with 4.1-A radiation a maximum proportion of the ions 

 is distributed in clusters long enough (about 14 m) to span the two adja- 

 cent chromatids. This conclusion, in turn, implies that the breaks are 

 located very close to the actual sites of ionization, of the order of less 

 than ^■'i IX distant. In support of this interpretation is the further 

 observation that neutrons, according to the data of Thoday (1942) as 

 worked over by Lea and Catcheside (1942), are even more efficient in 

 causing these isochromatid breaks. For this result corresponds with 

 the fact that the ion clusters resulting from neutron irradiation are of 

 comparable density to those from 4.1-A X rays but, in relation to the 

 genetic structures to be affected, are indefinitely long and are therefore 

 still more likely to be in striking range of two chromatids at once. 



Conversely, a rays, although also relatively long, can be shown to have 

 their ionizations crowded to an unnecessary degree on this view (see Lea, 

 1946), and therefore they would be expected to be less efficient than neu- 

 trons in producing breaks. Data of Kotval and Gray (1947; also cited 

 by Lea, 1946) indicate that they are, in fact, less efficient. Li the reckon- 

 ing of this efficiency, due allowance had to be made for the complication, 

 discussed on pp. 490-492, that fewer of the broken ends produced by a 

 rays join again (whatever the reason for this may be), and that more of 

 the breaks originally produced therefore persist until they can be 

 detected. 



Prior to and paving the way for these observations and interpretations 

 had been work of Giles (1940, 1943) on the effects of neutrons on the 

 chromosomes of Tradescantia microspores. In fact it was this work 

 which provided the first clear evidence of the fine-scale spatial limitation 

 of ionizations or activations in causing breaks of chromatin. In this 



