INTERMEDIATE STEPS OF MUTAGENESIS 319 



must work through the production of somewhat different substances 

 from those involved in the longer-range effect produced by the shorter 

 waves. 



A somewhat similar division of effects, perhaps at a different level, 

 has been found in studies of the "reactivation" or undoing of potential 

 radiation effects by means of visible and long-wave ultraviolet light [see, 

 for instance, Novick and Szilard (61), Watson (77)]. It is found that 

 there is a fixed proportion of the ultraviolet or x-ray effect (varying with 

 the type of radiation) which is not "reactivable." Similarly, in studying 

 the influence of withdrawal of oxygen [Giles and Riley (20)] or of freezing 

 [Faberg6 (17)] upon the induction of genetic effects, certain apparently 

 irreducible minima of effects are obtained. These seem to represent 

 changes induced by a rather different mechanism from that involved in 

 those effects which are responsive to changes in the condition in question. 

 They may, of course, involve bound oxygen, yet even in that case it is 

 unlikely that the reaction formulas are identical. Again, in the analysis 

 of radiation effects on phage, Watson finds evidence of several distinct 

 types of changes, the production of each of which is differently influ- 

 enced by various conditions attendant upon or following the irradiation. 

 The whole problem therefore is, as might be expected, a multiple one. 

 And it is becoming increasingly evident that there are various different 

 pathways leading from ionization or activation to mutation, some shorter 

 and straighter, others longer and more branched and devious. 



In the present discussion, attention has been concentrated chiefly on 

 problems of gene mutation, not because structural changes in chromo- 

 somes are unimportant, or because they are unrelated to gene mutations, 

 but simply because the topic would then become too diffuse. It is, how- 

 ever, necessary for us to orient the two subjects to one another in re- 

 spect to two particulars. The first of these concerns itself with the fact 

 that there are good grounds for regarding the distinction between the 

 two categories as valid, even though in many individual cases it is im- 

 possible to discriminate between them. Reasons for this have been 

 given elsewhere [Muller (45, 47, 52)] and need not be repeated here. 

 Second, we must clear the ground of the hypothesis [tentatively sug- 

 gested but abandoned by the present writer many years ago and re- 

 cently urged by Lea and Catcheside (34), and by Herskowitz (22)] that 

 "gene mutations" are often or usually accompanied by a chromosome 

 breakage near by, which may have been foflowed by either restitution 

 or recombination. Here again the reader must be referred to other 

 papers [for example, Valencia and Muller (75), Muller (52)] for the evi- 

 dence that gene mutations unaccompanied by rearrangement are not 

 restitution phenomena, and that the effects resembling gene mutations 



