FACTORS INFLUENCING FLOWER SIZE IN NICOTIANA 37 1 
per plant which possess only a little less than the normal length of 
tube but a very small diameter of corolla limb. East further states 
(p. 182) with reference to the correlation between spread and length 
in an F2 population, that "perfect coupling of certain factors together 
with independent combination of others may be possible; partial 
coupling arising from a peculiar gametic distribution may be equally 
possible." Our results appear to favor East's assumption of perfect 
coupling of certain factors with independent combination of others, 
and to indicate that the perfect coupling of factors may be simply 
due to the fact that these factors are common to length and spread 
of corolla. 
The sum of the conclusions above drawn makes it seem possible 
to state definitely that flower size, although very evidently not so 
markedly modified by environmental conditions as height of plant, 
leaf size, etc., still is not a stable character complex and is subject to 
marked modifications under the stress of both internal and external 
conditions attending development. The demonstration of these facts 
makes necessary in the future a more detailed description of technique 
and general method of flower size investigations than is ordinarily 
included in the published reports. On general principles it would 
seem rather necessary in a report on flower size inheritance to mention 
something of the technique of taking the measurements and at least 
a statement as to the number considered necessary to determine a 
fair mean for the various individuals of the populations reported upon. 
The facts brought out in the preceding pages serve to show how great 
an influence the number of measurements taken, the period of measure- 
ment, the age of the plants, and the age of the flowers, etc., will have 
upon a flower size distribution. 
V. Summary 
In connection with the taking of some 25,000 measurements on the 
length and spread of corolla limb of flowers on plants of N. Tahacum 
var. fnacrophylla and three hybrids between N. Tahacum varieties 
and N. sylvestris the following facts were observed. 
I. When the plants first come into flower the spread and length 
of corolla is greater than the spread and length of flowers produced on 
the same plants later in the growing and flowering season. This 
difference in size is more striking in the case of corolla spread than in 
the case of corolla length. Within a month or six weeks the average 
