■'•^'-^'5] Goodspcfd: InJtcrilancc in Nicotiana Ilijhrids 229 



J). 17;^j but that this situation might be dependent upon various con- 

 ditions external to the experiment and that the Mendelian explanation, 

 while obviously applicable, was of somewhat doubtful practical value. 

 It was suggested that the actual significance of this apparent increase 

 in variability in F. might lie in an increased fluctuating variability 

 due to environmental causes. A somewhat detailed study of the 

 parents under stril^ingly different conditions of culture showed that 

 ordinary greenhouse treatment resulted in a small but distinct increase 

 in the length of corolla tube as compared with that of sister plants 

 in the field. The most important factor considered in connection with 

 the effort to show that the increased variability in Fo was due to non- 

 inherited effects was the influence of the age of the plant, etc., in 

 determining the size of the flowers borne on a plant. A discussion 

 of these points is given in a paper to follow (Goodspeed and Clausen, 

 1915). In the earliest report, dealing with flower-size in F^ of these 

 hybrids, emphasis was laid upon the fact that all the fully opened 

 flowers were taken on each day of measurement from the various 

 plants and measured. Thus, for the F^ previously reported upon 

 and contained in the foregoing tables, the "age of plant" and the 

 "age of floAver" factor (Goodspeed and Clausen), which otherwise 

 might have accounted for the degree of variation in F,, were both 

 eliminated. The importance of attempting to do away with the de- 

 crease in flower-size brought about by the presence of maturing seed 

 capsules on the plant was imperfectly appreciated up to this past 

 season. Thus, without attempting to indicate directly the situation, 

 as has been done for other species of Nicotiana in another paper 

 {loc. cit.), it may simply be said that a certain proportion of the 

 increase in variability of the F^ population of the cross I X II 1912 

 (table 2) and the III X II 1^12 (table 5) was due to non-inherited 

 eft'ects. The distributions of the parent and Fj populations in 1912 

 and 1913 are much more nearly representative and accurate, since it 

 was possible to measure approximately all the flowers produced. 



Leaving out of account such situations as the striking difference 

 between reciprocal F, hybrids (table 5, II X HI, 1912, and III X H, 

 1012) as due to the small number of plants involved, the foregoing 

 tables seem to indicate the following facts : 



The F, generation of the .V. acuminata flower-size hybrids shows 

 .1 range of variation as great as and often greater than the parent 

 plants in corresponding years. The mean flower-size may be approxi- 

 mately the mean of the parent flower-sizes or it may not. 



