S. H. REVELL 



undergo rearrangement only lose their paired relation during prophase. If 

 this were correct then the two fragments from an ' isochromatid break ' might 

 still remain close together at metaphase because the association between 

 them had only just lapsed (compare with Figure 17b of Catcheside et al^). 



In addition to these three items of supporting evidence, it should also be 

 pointed out that the new interpretation gives a satisfactory explanation of 

 the original difficulty encountered in scoring chromatid breaks. All the 

 aberrations which previously were difficult to distinguish would now be 

 assumed to be equivalent : all are either complete or incomplete small 

 m/rochanges. But only one out of the four types — that which gives an 

 'isochromatid break' (type 4) — would be observed and discriminated effici- 

 ently. The others, owing to their small size, would frequently be missed or 

 else confused with one another, and it would moreover be difficult in some 

 cases to decide whether they were incomplete (that is, whether or not they 

 constituted ' chromatid breaks ') . The hypothesis also accounts for the occur- 

 rence of chromatid and isochromatid breaks in the same nucleus. 



A large proportion of the whole class of 'chromatid ' aberrations as defined 

 in the orthodox theory — both chromatid and isochromatid breaks, chromatid 

 zn/^rchanges and ?Vz/rachanges — may thus be interpreted in terms of one unit 

 of change : a chromatid exchange. But it will be noted that there is still 

 an important group of m/^rchanges which remains unaccounted for ; namely 

 those which are conventionally interpreted as involving reunion of one or 

 more isochromatid breaks. It is not obvious how this type of {nterch.a.nge 

 could arise by chromatid exchange. Nevertheless, certain facts suggest that 

 they might be formed in essentially the same way. 



The most frequent of this last group of aberrations is the chromatid- 

 isochromatid /«/^rchange ( = triradial) which, according to the orthodox 

 interpretation, results from reunion between one isochromatid fragment and 

 a single chromatid break, the other isochromatid fragment commonly 

 undergoing sister union. It is often the case (a) that this latter fragment 

 remains close to its related configuration at metaphase (see paragraph (3) 

 and Figure 2g) ; and (b) that the short achromatic region is eccentrically 

 placed as it is in sister unions of non-/«/frchanged chromosomes (see para- 

 graph (/) and Figure 2g). These two facts strongly suggest that the sister 

 union event, whatever it is, which occurs in a triradial is the same as that 

 which results in ordinary isochromatid fragments. 



Now it is occasionally evident that a sister union event can occur close 

 to a chromatid interchange {Figure 2e). In view of this, the items of evidence 

 (a) and (b) above suggest that a triradial could consist of two chromatid 

 exchanges very close together — one interchange and one intrachange (Figure 

 2f) : the chromatid relations would be obscured during contraction, the 

 fragment thus separating from the main configin-ation by the time of meta- 

 phase {Figure 2g). Similar interpretations in terms of two or more exchanges 

 near together may be given of other less common types of 'interchange'. 



Thus it is possible, at least qualitatively, to interpret all recognized ' chro- 

 matid ' aberrations in terms of one or more chromatid exchanges. It is 

 evident that since incomplete m/erchanges, failed ' sister unions ' and chro- 

 matid ' breaks ' are all assumed to be incomplete exchanges their numbers 

 should be positively correlated. There is, of course, some evidence that 



249 



