S. H. REVELL 



But more is known about exchanges than this. Although they rise as 

 about the square of the X-ray dose, and although this is taken as ruling out 

 a ' contact ' mechanism, nevertheless several different kinds of evidence all 

 indicate that two chromosome points must be both close together in space 

 and must both be affected within a very short time if they are to exchange. 

 This is conventionally interpreted as meaning that restitution usually occurs 

 quickly after breakage. Therefore only breaks induced very close together 

 (the l[x and 3^ minutes of Lea^) have much chance of reuniting because 

 restitution otherwise supervenes and makes them unavailable to one another. 

 But after all, the original ' contact first ' problem was essentially a biological 

 problem, not of how many ionization columns are required to cause an 

 exchange, but whether or not its initiation is dependent on there being first 

 a functional proximity of some kind. It was only afterwards suggested 

 that a study of dose-interchange relations might give an answer to the 

 question. 



Now in the case of the present work, it seemed to the author that the 

 occurrence of the homologous exchanges raised the possibility that there 

 might be an association involved which had a biological capacity for ex- 

 change, as distinct from there being just a spatial proximity which favoured 

 exchange as the outcome of an essentially random breakage and reunion. 



In considering the possibility of an exchange process for interchanges, the 

 question at once arises of what the single breaks might otherwise be if it 

 turned out that they were not the components of the interchanges. Now 

 the diflficulty of scoring certain changes in Vicia, which has already been . 

 mentioned, was concerned with the single breaks. Isochromatid breaks were 

 easily observed but it was difficult to have confidence in the chromatid break 

 scores because, although some of the latter were clear enough, they seemed 

 to grade imperceptibly into a whole complex of other anomalous changes 

 which, as a group, could not reasonably be described as chromatid breaks 

 at all. Part of this difficulty was concerned with numerous small vuistained 

 regions. These occur in many places. For instance, in m/^rchanges they 

 seem to mark the points where the breaks and reimions might be supposed to 

 have occurred {Figure la), but other apparently identical unstained regions 

 may also be observed in single chromosomes which have not interchanged, 

 and these might perhaps be conventionally interpreted as detected resti- 

 tutions ( Morrison 1"). This, however, did not appear to be a completely 

 convincing explanation because, at least in Vicia, these regions so often occur 

 in chromosomes with unequal chromatids and, frequently, associated minutes 

 as well. Altogether, it seemed difficult to justify the arbitrary separation of 

 the typical chromatid breaks [e.g. Figure 2c) from all these others {Figure lb). 

 In one type out of this group of anomalies it seems clear that part of one 

 chromatid has been intercalated into the other. The simplest way in 

 which this could occur is by chromatid exchange in a small loop {Figure Ic, 

 type 1). Such m^rachanges are well known in Tradescantia and Trillium, 

 there being four possible types depending on which pair of chromatids 

 exchange {Figure Ic, types 1-4 ; see also Figure 17 of Catcheside, Lea 

 and Thodayi, and Figure 10 of Darlington and La Cour^^). It was 

 at first difficult to accept this interpretation for Vicia because the three 

 other m/rrtchanges seemed to be missing {Figure Ic, types 2-4) each of 



245 



