V. CONCLUSIONS. 



The results obtained in these investigations may be briefly 

 summed up as follows : — 



1. In any statistical investigation on the mode of inheri- 

 tance, the uniformity of the environment in which each set of 

 cultures is grown is exceedingly important. Comparisons 

 should not be drawn between cultures unless they are grown 

 close to one another with full precautions as to uniformity in 

 environment. By careful attention to cultural details it is 

 possible to reduce greatly the effect of environmental fluctua- 

 tions. The importance of using in such investigations only 

 normal, well groAvn plants cannot be over-estimated. 



2. Parthenogenesis in N. tabacum, under the conditions 

 obtaining in hybridization work at Pusa, is negligible. 



3. In aU characters except height, the Fi generation is 

 intermediate between the parents. In the case of the height, 

 different results were obtained in different crosses. This 

 may be due to added vigour in the hybrid plants. It is suggested 

 that the differences in the increase produced by this msiy depend 

 on differences in the number of dissimilar factors in the 

 parents. 



4. In aU cases the limits of variation in the Fo generation 

 have been as great as those of both parents combined or have 

 exceeded these in both directions. In some cases, where the 

 parents and the Fj generation were all alike, the variation in the 

 Fg was very great. This can readily be explained by the 

 hypothesis that most of the factors possessed by the parents are 

 different. 



5. Selected variates of the Fo generation gave cultures 

 which differed in their range of variation from one another, and 



