374 W. J. C. Lawrence Cytologia2, 



mata and four types with a single terminal chiasma. Without secon- 

 dary association no evidence whatever exists for the duplication of 

 chromosome types, in this and other species. 



In the Pomoideae secondary association has enabled Darlington 

 and MOFFETT (1930) and Mopfett (1931) to demonstrate that the hap- 

 loid chromosome number of 17 is a secondary basic number derived from 

 an ancestral form with 7 pairs of chromosomes. Similarly in Rubus 

 (Crane and Darlington 1927) and Prunus (Darlington 1928) in which 

 diploid and polyploid forms occur, the numerical relation of the dif- 

 ferent chromosome-numbered forms is in agreement with the second- 

 ary association observed. In Hidalgoa, which has chromosomes of 

 varying sizes, chromosomes of the same size are secondarily associated 

 at metaphase in groups of 2 and 3. This genus is being studied fur- 

 ther. At present there is no evidence that multivalent association 

 occurs, but we may infer that Hidalgoa is a secondary polyploid with 

 some chromosome types represented 6 and others 4 times. Mention 

 has been made of Kuwada's drawings of metaphase in Oryza sativa in 

 which 4 or 5 groups of 2 bivalents each are figured. It is interesting 

 to note that Chao (1928) and others report the occurrence of 5 duplicate 

 factors in this species. While it is possible that duplication of parts of 

 chromosomes has given rise to some duplicate factors it seems probable 

 in this and other cases (p. 376) that they may be attributed to chromo- 

 some duplication. If this is true then the genetics of Oryza not only 

 confirm the evidence of secondary association but also indicate this 

 species to be a secondary polyploid derived from a form with 7 pairs of 

 chromosomes, in common with the majoriry of the Gramineae. The 

 evidence suggests that Zea similarly may have arisen by duplication 

 of a portion of the chromosome complement. The genetics of Oryza 

 should prove of great interest in view of the /fact that this genus is 

 amenable to genetic analysis, which is difficult in Pyrus. 



As previously stated, no final conclusions can usually be based 

 upon the evidence of secondary association inferred from the literature. 

 But in certain cases this evidence is so significant that there is little 

 doubt as to the reality of the apparent assQciations depicted. 



Thus the evidence from secondary association in Betula is con- 

 firmed by the quadrivalents found in Corylus, and indicates that the 

 basic number in the Betulaceae is 7. Similarly we may infer that the 

 secondary association found in the following genera indicates that they 

 are secondary polyploids, \i2.:—Salix 19; Gossypium 13; Xanthium 

 18; Ficus 13; Vitis 19. It is highly probable that natural orders 



