GENETICS; GOODSPEED AND CLAUSEN 
337 
too old to be measured and all developing seed capsules were removed 
twice a week. In a month the spread of corolla for Plant 14 decreased 
5.78 mm. and for Plant 15 only 0.75 mm. At the end of two months 
the decrease amounted to 5.40 mm. for Plant 14 and 2.29 mm. for Plant 
15. The length behaved similarly in this case, in Plant 14 there was a 
decrease of 4.55 mm. and in Plant 15 of 0.25 mm.; and at the end of 
two months, a decrease of 4.00 mm. in Plant 14 and 0,79 mm. in Plant 
15. 
During the period which elapses from the time a flower is fully opened 
to the time at which pollen is shed, there is a considerable increase in 
corolla spread and associated with it little or no increase in corolla length. 
On ten plants of F1H38, the increase in spread amounted on an average 
to 4.53 =t 0.23 mm. and for length there was an increase of 1.62 =t= 0.22 
mm. Similarly the spread of corolla for F1H137 averaged 3.18 =t 0.17 
mm. greater for flowers after shedding pollen, but in this case the length 
was 0.47 =1= 0.11 mm. smaller, a slight discrepancy undoubtedly due to 
the indirect method of comparison made, which still further emphasizes 
the increase in spread of corolla during this period. This increase in 
spread has, also, been repeatedly confirmed on individual marked flowers. 
That there is a differential distribution of flowers on tobacco plants 
according to size at any given time is shown by the comparative meas- 
urements of flowers borne among developing seed capsules on the ter- 
minal inflorescence of a plant and those borne on laterals of the same 
plants. In the case of plants of N. Tabacum var. macro phylla the flowers 
borne on laterals were found to have an average spread greater by 
2.56 =«= 0.16 mm. than those of the terminal inflorescence, and in the 
case of length, 1.24 0.11 mm. greater. In the hybrids studied the 
difference is rather more marked, so much so that there is a distinctly 
noticeable difference in size between these two classes of flowers. 
Similarly we have found that numerous other factors have a like 
influence on flower size, some relatively great and others less marked. 
For instance cuttings of F1H44 growing in the greenhouse produced 
flowers 3.95 mm. smaller in spread and 1.42 mm. greater in length than 
those on the field plants from which the cuttings were taken. Pure 
line populations of N. sylvestris grown in a shaded situation on rich 
moist garden soil produced distinctly larger flowers than plants growing 
on higher unfertilized soil and not shaded, and pot experiments likewise 
showed that flower size could be distinctly influenced by applications 
of sodium nitrate, and in a direction parallel to that of the influence on 
vegetative characters. 
The conclusion seems irresistible that flower-size in Nicotiana is not 
