122 The Structure of Protoplasm 



have mostly been interpreted in accordance with the opinion the 

 investigator had previously formed from direct microscopic obser- 

 vation. The opinion is now gaining ground that X-rays are too crude 

 a tool for settling this problem. Fourth level data show that there 

 are four genetically effective strands in a meiotic bivalent so far as 

 crossing over is concerned, but this says nothing of the structure 

 within each of these. 



In our McGill laboratory we have, during the last five years, 

 attempted to get more definite evidence on some of these problems 

 by starting with statistical analyses of extensive data from the largest 

 and most easily observed spiral structures. For this purpose Trillium 

 is exceptionally favorable material. 



Our first significant observation was that the major coil changes 

 direction at numerous points in Trillium. By analysis of normal 

 material and comparison with pollen mother-cells that had different 

 degrees of lack of association between homologous chromosomes 

 and sister chromatids through temperature treatments, we have 

 been able to establish (Huskins and Wilson, 1938) that reversals 

 occur with random frequency at the kinetochore and at chiasmata, 

 and that elsewhere they may occur from unknown causes with a 

 frequency proportional to the length of chromosome involved. It 

 was next discovered (Wilson and Huskins, 1939) that the chromo- 

 nema more than doubles in length during formation of the major 

 coil — the nature of the elongation is not yet determined. These facts 

 seem decisively to rule out any hypothesis which involves an internal 

 torsion due to molecular pattern and directed as a unit. They are 

 fitted for the present by the simple assumption that a thread of a 

 resilient consistency will take up a spiral form if it expands in length 

 within a confined space— the sheath or pellicle. Almost surely this 

 concept will be found inadequate to explain some of the other spirals 

 now being studied. 



The relational coil has generally been assumed to result from 

 torsion causing the chromonemata to twist around each other. Our 

 current data (Sparrow, Huskins, and Wilson, 1941) show that the 

 microspore chromatid relational coiling results directly from the 

 partial straightening out of the gyres of the major coil of meiosis 

 and is related to the plane assumed by the tertiary split in the major 

 coil. This cannot, of course, be the direct cause of the relational 

 coiling of homologous chromosomes in the prophase of meiosis — 

 which requires much more rigid analysis than it has yet received. 

 We are proceeding at present with analyses of standard somatic 



