348 Gates. 
Table 6. 
O. rubrinervis, 2.14. 53 buds. 
fo) | I 2 3 | 4 | 5 6 | 7 
Sept. 28. | x x x.g | Na ex | 
| x.g | XXX xxXX 
x x xX x xX 
| x x xX x X 
We | IX: BRR: 
(2—3) (85) | x x 
October 1t. | | x (4—5) ents 
| | | (4—5) | 
| | | RX: | x cx | 
| | x | 
(2—3) . (4-5) ex xx 
| (35) 
| x(4—S) | 
(2—3) |= 
| (3—5)- 4 
|_x(4—5) | | 
I | 4 (3) 9 (3) (1) 16 | 16 | 
| I 5.5 12 | I 17.5 16 | 
| | [1 
| I | 5.5 12 34.5 | 
of accuracy which is quite impossible of attainment with these color 
characters, — the element of judgment comes in, so that the results 
of course give general indications rather than mathematical data for 
the plotting of curves. But a more rigorous mathematical treatment 
would probably serve no useful purpose, even if it were practicable. 
Classes o—7 are perhaps reasonably nearly equidistant from each other, 
but the gap between classes 7 and 8 is probably nearly as wide as 
the whole distance from class o to 7. 
Tables 2 to 7 register the results for the buds of several different 
individuals of O. rubrinervis, showing that there is a single mode for 
the extent of the red color-pattern in each individual but that the 
mode may fall in different classes in different individuals. These 
tables will also serve to show the complexity of the variability, as 
indicated by the formulae for individual buds. 
B. Other color variations. 
In a culture of Oenotkeras from near Liverpool, England, the 
offspring of one individual were of a single type as I grew them at 
