502 CELL MECHANICS 



length of the hght used. Experiments with infra-red Hght (Hruby, 

 1935) show that incorrect illumination or the use of light of a 

 wave-length having a certain proportion to the diameter of the 

 cylinder observed, may produce this result. This explanation 

 seems to be all the more inevitable since the material and the stage 

 are the very ones in which the chromosomes were tested by X-ray 

 treatments and shown to be single. 



The question now arises as to how the chromosomes " divide." 

 The knowledge that primary attraction exists only between pairs 

 of chromomeres at once suggests an analogy with the property of 

 chromomeres reproducing only to give pairs. In reproduction a new 

 chromomere or gene identical with the parental chromomere or 

 gene is laid down beside it. This, at least, is the description necessary 

 if we look upon the structures as having an organisation that is 

 characteristic and unchanging like that of a molecule which retains 

 characteristic and unchanging chemical properties. We may say, 

 therefore, that the new particle that is laid down is attracted by the 

 old one from the substrate. And the possibility of this attraction 

 taking effect must depend on two general conditions : the degree 

 to which the attractions of the particles are already satisfied, and 

 the concentration of suitable materials in the substrate. A test 

 of this view is provided by the behaviour of unpaired threads in 

 pachytene nuclei as compared with that of paired threads. They 

 divide earlier both in triploids and in organisms with incomplete 

 pachytene pairing (D. 1935 h). Evidently, therefore, when the 

 primary attraction is unsatisfied the concentration of substrate 

 necessary for reproduction is reached earlier and presumably at a 

 lower level than when this attraction is satisfied. Further evidence 

 of a relationship between attraction and reproduction is shown by 

 the salivary gland nuclei. An unlimited reproduction is associated 

 with an unlimited attraction. 



The question that next arises is as to whether the chromosome 

 thread divides into two equivalent daughter-threads or as we have 

 supposed gives a daughter-thread distinct in its properties from a 

 parent thread. This question is answered by the observations of 

 ring chromosomes (Ch. III). Such chromosomes divide so that 

 either two free rings or two interlocked rings are formed. In the 



