376 W. J. C. Lawrence Cytologia2, 



dispersed to the periphery of the spindle, without regard to the posi- 

 tion of the equator. Should secondary association occur among uni- 

 valents, then we might expect a random distribution and orientation of 

 the pairs of univalents. This has been observed by Mofpett in the 

 secondary triploid Pyrus minima f2n=51). In this species about 13 

 univalents are commonly seen at metaphase of the first division, but 

 they are frequently associated in pairs, as many as 4 pairs being found 

 distributed at random with the remainder of the univalents round the 

 periphery of the spindle. On the other hand we cannot preclude the 

 possibility of univalents associating in such a manner that their attach- 

 ments would be diametrically opposite, i.e. their close association would 

 simulate true bivalent association, except for the absence of chiasmata. 

 As a result of such an association it is possible that pairs of univalents 

 would align themselves on the equator and pass to opposite poles. 



So long ago as 1905 Wilson observed and described an association 

 of this kind in the Hemiptera. In the material he studied the small 

 m-chromosome did not pair at prophase. ' ' As the spindle forms the 



two micro-chromosomes approach each other, and in the stage just 



preceding the metaphase finally conjugate to form the small bivalent 

 chromosome. Without fusing, the two halves are then immediately 

 separated, the division always taking place more rapidly than in the 

 case of the larger chromosomes " ^' (my italics). 



It is possible that the paired univalents in Nicotiana purpurea 

 haploid behave in a similar manner on the occasions when they are 

 apparently aligned on the equator. Chipman and Goodspeed however 

 did not observe the distribution of these ' paired ' univalents. 



In conclusion, it should be noted that the deduction of a polyploid 

 nature for a number of the species mentioned in this paper is not 

 without genetical support. The complex results obtained in breeding 

 experiments on such plants as Cotton (Harland 1929-30, Kearney 1930 

 etc.), Geum (Marsden-Jones 1930), Brassica (Peace 1926, Malinowski 

 1921), Viola (Clausen 1927, 1929), Cardamine (Correns 1912), Solanum 

 tuberosum (Salaman 1910, 1930) (Collins 1924), Lythrum Salicaria 

 (Barlow 1923, East 1927), Pyrv^ (Crane and Lawrence 1931, Welling- 

 ton 1924) etc. are all indications of polyploidy. For in the allo-poly- 

 ploid, inheritance will be in the direction of disomy but modified by the 

 action of duplicate factors (Peace 1926, Shull 1929 ; Lawrence 1931 ; 

 MiJNTZiNG 1929 ; Mangelsdorf 1930) Chao 1928, and by factors which, 



1) Of. notes on disjunction on p. 363 of this paper. 



