564 APPENDIX I 



not proved amenable to observation by the customary methods (Lucas 

 and Stark, 1931). 



In the saUvary glands it has been possible with ultra-violet light to 

 discriminate between chromomeres that were not distinguishable in 

 visible light, but whose separate existence was inferred from the 

 occurrence of X-ray breakages (Ellenhorn, Prokofieva and Muller, 

 1935)- Similarly, the use of X-rays for changing genes and for breaking 

 the chromosome thread has superseded direct observation as a means 

 of testing such delicate questions as the time of division of these 

 structures. Direct X-ray photography has not yet been brought to 

 bear on the chromosomes and the spindle [cf. Astbury, 1933). 



Evidently visible light can no longer be regarded as the final instru- 

 ment of observation for chromosomes. Our senses have been sur- 

 passed by intermediate mechanisms and in this way the old objection 

 to cytology, tot capita, tot sensus, together with its stale controversies, 

 will be put on one side. 



(ii) The observations may be compared with those of related 

 organisms at the same stage of development (after similar or different 

 treatment). In this way a study of the chromosomes at mitosis and 

 meiosis in material which has satisfied the first kind of test enables us 

 to conclude that such structures as constrictions and chiasmata are 

 present in other material less suitable for study on account of the 

 concomitant conditions found, e.g., an inturned bend in the chromosome 

 at the constriction or a change of plane in a diplotene loop or opening 

 of the chromatids in anaphase at the chiasma. The same criterion has 

 been used to show that multiple ring formation at meiosis is the result of 

 normal development of structurally dissimilar chromosomes, in hybrids. 



(iii) The observations may be compared with those of the same 

 material at the stages immediately preceding and following. This 

 method is essential in the interpretation of meiosis in regard to the 

 pairing of the chromosomes, the movements of chiasmata, the separation 

 at anaphase, the frequency of association, and so forth, and, in examin- 

 ing these, quantitative methods are often necessary. Developmental 

 comparison has shown the invalidity of various theories of pairing and 

 crossing-over. Developmental comparison is equally essential for 

 many minor inferences. For example, nucleolar material surviving 

 to metaphase at meiosis in many forms may be mistaken for free small 

 chromosomes produced by fragmentation unless different stages are 

 compared critically. 



It is not yet possible to show the whole course of meiosis in all 

 organisms on account of two major difficulties : (i) The occurrence of 

 stages when the chromosomes are collapsed to one side of the nucleus ; 

 this is usually at zygotene and diplotene, probably for the reason that 

 the chromosomes at these stages are slenderer than at the stages 



