S. H. REVELL 



which should be occurring with a frequency equal to the duplication- 

 deletion (type 1). 



However, the resolution of this problem could depend on one assumption : 

 that if the loops of these /«/rachanges were always small enough in Vicia 

 then, as the whole configuration shortened to metaphase, the original 

 chromatid relations would be lost. Type 2 would then yield the bent 

 chromosome with minute juxtaposed ; type 3 would probably be difficult to 

 see at all since it is merely a small inversion ; and type 4 would appear as 

 an isochromatid break with sister union {Figure Id). Carried a stage 

 further, this interpretation could account for a proportion of chromatid 

 breaks : if, like interchanges, these m/rachanges may also be incomplete — 

 that is, with one of the two pairs of chromatid ends not joined up in the 

 exchanged relation {Figure 2a ; see Catcheside et al^), then types 1 to 3 could 

 yield ' chromatid breaks ' (compare Figures 2b and 2c) and type 4 could give 

 failed 'sister union'. 



This interpretation is supported by three observations. 

 (7) If the aberrations in question are really collapsed m/rachanges — that 

 is to say, chromatid exchanges like the interc\\a.ngc?, — then some of the small 

 unstained regions already mentioned which were evidently associated with 

 these aberrations may now be rationally explained (no matter what they 

 actually are) as marking the points of exchange, just as they are seen to do 

 in the case of m^^rchanges {Figure la). For example, if 'sister unions' are 

 really collapsed mifrachanges then the points of chromatid exchange must be 

 to one side of the ' end ' of each fragment. It is frequently possible to confirm 

 this by observation {Figure Id, type 4 photomicrograph). Such achromatic 

 points may also be observed on the other presumed ?«/rachanges {e.g. Figure 

 lb). Previous to the development of this hypothesis, the author had inter- 

 preted many of these points as chromatid breaks. 



(2) In contrast to Trillium and Tradescantia, no clear intra-arm mimchanges 

 are observed at metaphase in Vicia. However, such /«/rachanges are demon- 

 strable at prophase although they are always relatively much smaller than 

 those in the former species {Figure 2d). Therefore these m/rachanges must 

 reach metaphase in a disguised form, and it seems that the most rational 

 explanation available is that they are represented by the aberrations already 

 described* {Figure Id). 



{3) As was observed by Lewitsky and Araratian^^^ the related parts of 

 affected chromosomes are frequently closely juxtaposed at metaphase. Thus 

 the fragments from an isochromatid break are often very close together and 

 minutes are usually close to the chromosomes from which they are presum- 

 ably deleted {Figure Id). (A similar phenomenon can be observed in the 

 case of ' triradials ' and certain other types of interchange — see below.) This 

 observation is not readily accounted for by the orthodox hypothesis since 

 breakage and reunion are assumed to have occurred at the time of treat- 

 ment, there being no subsequent relation between the separated chromosome 

 parts. It can, however, be plausibly explained by the hypothesis now sug- 

 gested, which postulates that the chromatids of the minute loops which 



* The same explanation could be advanced for these aberrations in Trillium and Trades- 

 cantia, although in these species the largest zViirachanges retain their form even at metaphase 

 and are therefore recognized as such. 



247 



