56o CELL MECHANICS 



"fragmentation" and "fusion" in the past, can arise only by 

 gain and loss of centromeres. And since, so far as we know, all 

 such changes apart from simple loss or gain of chromosomes depend 

 on structural changes occurring at random in the chromosomes, 

 we have to consider how they could arise without fatal inconvenience 

 to the organism. It seems that this would be achieved most probably 

 by unequal interchange. Thus two rod-shaped chromosomes with 

 sub-terminal centromeres interchanging would give a single V-shaped 

 chromosome and a very short fragment which might then be lost, 

 and the chromosome number reduced. On the other hand if the 

 fragment were retained along with the unchanged types the chromo- 

 some number would be increased. The first would depend on the 

 region near the centromere being inert ; the frequent occurrence 

 of inert supernumerary fragments shows that this is often the case. 

 In Fritillaria, where such fragments are formed, the expected 

 concomitant changes in number and form of the chromosomes 

 are also found in comparing different species (D., 1935 g). It seems, 

 therefore, that the direction of change of chromosome number must 

 be conditioned by the genetic activity or inertness of the proximal 

 parts of the chromosomes. Evolutionary stability of basic chromo- 

 some numbers means activity of genes near the centromere and 

 the ends. Instability means their inertness. 



(d) The Molecular Basis of Cell Mechanics. In considering the 

 internal and external mechanics of the chromosomes we find a 

 wide range of movements which are co-ordinated in time and 

 therefore necessarily related in cause. In order to understand their 

 causation we must pursue the study to lower levels of integration, 

 to colloidal and molecular dimensions. Having inferred movements 

 from changes in structure, we must infer structure from differences 

 in movement. 



Let us take first the internal changes in the chromosomes. They 

 require, as we have seen, changes in a molecular spiral. Long 

 molecules of a chain structure must be twisted spirally. Such 

 changes of shape are characteristic of certain protein molecules 

 and may be determined in various ways, the most obvious being 

 by a change in the surface electrical conditions, which will necessarily 

 change the internal spatial equilibrium of the molecule {cf. Koltzoff, 



