420 Morpho genetic Factors 



genes were operative in the former character than in the latter. There 

 was by no means free recombination of length and width in F 2 , however, 

 as independent assortment would require. The combinations that did 

 occur were only a narrow segment of those theoretically expected. When 

 length was plotted against width in F 2 , these were confined to a narrow 

 segment of the total, running from combinations rather like one parent 

 to those like the other through others resembling the Fj. He suggests that 

 factors hindering free recombination might be gametic elimination, 

 zygotic elimination, pleiotropism, and linkage. He believes that all of 

 these may here be operative and suggests that all quantitative characters 

 of an organism may be tightly linked, surely a radical conclusion. An 

 extreme instance of the hypothesis that shape is the result of genes 

 determining dimensions has been proposed by Frets (1947), who postu- 

 lates that in the inheritance of seed shape in beans there are a series of 

 independent genes for length, breadth, and thickness, respectively. 



H. H. Smith (1950) studied a cross much like that made by Anderson 

 and comes to the conclusion that there is a developmental restriction to 

 free recombination but that this is due to a "correlated growth pattern." 

 In simpler words, there are genes that control shape directly rather than 

 through individual dimensions. 



Evidence for the existence of such genes has been presented by Sin- 

 nott (1935), chiefly from studies of the inheritance of fruit shape in the 

 Cucurbitaceae. This evidence is of several types, as follows: 



1. If a race with flattened, disk-shaped fruits is crossed with a spheri- 

 cal one, the F! shows complete dominance of the disk shape and in the 

 F 2 there is sharp segregation into % disk and y± sphere. In another case, 

 two different types of spheres, when crossed, show evidence of the ac- 

 tion of complementary genes, for the F x is disk-shaped and in the F 2 

 there is dihybrid segregation into % 6 disk, % 6 sphere, and y 16 elongate. 

 Other shape differences can be analyzed in equally simple mendelian 

 terms, though more genes are usually involved. 



2. In the disk-sphere cross, F 2 segregation for shape index (ratio of 

 length to width) is sharp but those for length and for breadth are much 

 less so, suggesting that the segregating genes control shape directly and 

 not through dimensions. 



3. In one disk-sphere cross, the fruits of the disk parent were con- 

 siderably larger than those of the sphere. The size of the F x was close to 

 the geometric mean between the two parental sizes, and the means of the 

 segregating F 2 disks and spheres were essentially similar to each other 

 and close to that of the F\. This can be explained by assuming that 

 size is determined by a series of genes but that the gene for shape is 

 independent of these and molds into a particular form the material 

 made available by the genes for size. 



