Structural Differentiation of the Nucleus 123 



coils. The torsion hypotheses of coiling and crossing over have also 

 been checked by analyses of the relationship of successive chiasmata 

 (Huskins and Newcombe, 1941). These reveal a more complex 

 relationship than previously envisaged, and experiments with Neuro- 

 spora (Lindegren and Lindegren, 1937 and unpublished) seem to 

 substantiate the results with fourth level, genetic data. 



To sum up: The structure of the nucleus is highly involved and 

 changes in it are being correlated with function. Studies of the 

 nucleus are proceeding on four different levels which differ con- 

 siderably in the degree of accuracy or objectivity attainable. First- 

 level analyses of chromosomes have led to hypotheses on their mole- 

 cular pattern. Ultraviolet and Feulgen studies of the nucleus have 

 established the nucleic acid cycle during mitosis and interkinesis. 

 The conclusions from these regarding the time of reproduction of 

 the gene rest, however, partly on second-level or microscopic obser- 

 vations of the time of chromosome or chromonema reproduction, 

 regarding which there are sharp differences of opinion. They are, 

 therefore, much less definitely established than the first-level data 

 on which they may appear to be based. Microchemical studies on 

 genetically controlled materials involving heterochromatic regions 

 (combined first-, fourth-, and third-level studies) indicate: (a) that 

 heterochromatic regions are specialized for the production of nucleic 

 acid, (b) that they are concerned in the formation of the nucleolus, 

 and (c) that the position of genes relative to these regions affects 

 the expression of the characteristics they determine. The functions 

 of nucleoli of different types remain undetermined. 



Second-level microscopic studies are now very frequently com- 

 bined with fourth-level, genetic analyses. Where the correlation is 

 experimental, very solid progress has been made. One school of 

 second-level cytology uses genetic data extensively for the deductive 

 formulation of cytological hypotheses (Darlington, 1935a and b) . 

 Very great advances have been made by this technique, but the 

 dangers inherent in the method are apparently not realized by all 

 those following it, nor is the admittedly tentative nature of some of 

 the conclusions drawn by this school realized by many workers on 

 other levels who attempt correlations of them with their own data. 



Third-level studies by the methods of experimental cytology- 

 have not been considered to any extent in the above brief review, 

 since they are considered by another contributor to this symposium. 

 Only limited aspects of the problems of mitosis have been consid- 

 ered since this process obviously depends upon reactions of the cell 



