The Theory of Polymery 349 



Now that we have considered several theoretical cases of dif- 

 ferent degrees of complexity, let us study an actual example. In 

 1916, Dr. E. M. East published some results of crosses between 

 two plants of the genus Nicotiana. Among other characters that 

 he studied was length of the corolla tube. This character is good 

 for genetic studies because it is little affected by environmental 

 differences. The flowers of N. Langsdorffii averaged about 21 

 mm in length whereas those of N. alata averaged about 82 mm. 

 As can be seen from Table 19, Langsdorffii shows very little 

 variability, undoubtedly because of the almost universal self- 

 fertilization of that species. The other species is somewhat more 

 variable (Fig. 98). Theoretically, the Fi should average about 

 51.5 mm, but actually the average corolla length of the Fi plants 

 was only 41 mm. This discrepancy between theory and obser- 

 vation has not been explained. As is to be expected from the 

 theory of polymeric genes, however, the variability in the Fi is 

 low. The. average length in the F2 is somewhat in excess of 38 

 mm. Again, this average is considerably less than expected, 

 but the average of the F2 is practically the same as of the Fi, 

 an observation well in accord with the expectation according 

 to hypothesis. The F2 shows far greater variability than the Fi 

 and in this respect agrees with the theory. Two of the F2 plants 

 are about as small as some of the Langsdorffii strain, but no 

 plants were as tall as the shortest plants of N. alata. If a large 

 number of genes were contributing to height, the 581 Fo plants 

 would probably not constitute a population large enough to en- 

 sure the recovery of parental types. 



A study of various F3 families from this cross offers interesting 

 support for the explanation of corolla size based upon polymeric 

 genes. Various F2 plants, when selfed, should produce F3 popu- 

 lations which differ markedly in their average length of corolla. 

 That this is true can readily be seen from Table 19. For ex- 

 ample, family 7 averages only 21 mm whereas 3 averages 39 mm 

 and 4 averages 54 mm. This splitting up of the F3 into dif- 

 ferent lines with different average corolla lengths is one of the 

 features demanded by the theory of polymeric genes. Another 

 requirement of the hypothesis is that various F3 families should 

 show different amounts of variability, whether or not they av- 

 erage the same in size. We showed previously that this is to 

 be expected because some plants are heterozygous for a number 



