III. / nheritaiice in SJiirley Poppij 
395 
The object of the first series of experiments, like those which have followed it, 
was to ascertain the intensity of heredity in a free plant population, and to com- 
pare such heredity with that of similar populations in other types of life. If we 
have to deal with evolution in actual living forms in which isolation of individuals 
is not the rule, it is the law of inheritance in populutions with which we are 
concerned. We desire to know the statistics of sub-classes, and the proportions in 
which those sub-classes reproduce themselves. A statistical investigation of this 
kind cannot be opposed to any ultimate analysis of the I'esult of crossing indi- 
viduals; it can only indeed serve as a criterion of the truth of any hypothesis 
based upon individual crossings. At the same time it avoids the dangers which 
arise from neglecting what we may term the " law of small numbers." It is too 
often forgotten that an imposing display of numbers, obtained by dealing with 
the numerous offspring of a number of crossings which can be counted on the 
fingers, must on any hypothesis of parental gametic constancy be given the weight 
of the number of crossings and not of the number of offspring, when the results 
are applied to the race as a whole. 
In our earlier experiments we were not certain to what extent Shirley Poppies 
grown in masses were cross or self-fertilised *. This doubt left a double interpre- 
tation open when we came to deal with the resemblance of siblings. Our results 
in the previous investigation were in general accordance with those for man, 
provided we assumed that the bulk of the plants were cross-fertilised. 
Another point of considerable importance was the question of whether any 
really well-defined colour sub-groups could be asserted to exist in the case of the 
Shirley Poppy. In order to determine this point water-colour sketches of the 
petals of a long series of the poppies wei-e kindly made for us by Miss Eva Biggs 
immediately after gathering. It was soon seen that in distribution of colour, 
amount of colour, and shade of pigmentation there was practical continuity. We 
found it impossible to draw a rigid line between scarlets, reds, orange-reds, terra- 
cottas, rose-pinks, pinks, and pink-whites, although such terminology served fairly 
well for rough classification. There appeared always to be transitional individuals, 
although such individuals might be numerically few. The ocular demonstration 
of this was difficult even in Miss Biggs' paintings, because the colour of the poppy 
is very difficult to match in water-colours, and being matched is very evanescent. 
We believe that Miss Biggs' careful work fully achieves what is possible in this 
direction. But the task becomes harder still when the colour printer has to match 
the artist's tints. If honest, he frankly tells you that he can approximately catch 
the artist's shades, but that he can only do so by using non-durable printer's inks. 
The alternative before the Editors was therefore to produce something like the 
artist's work which will not be permanent, or something unlike it which will be 
permanent. The latter seems to fail entirely in its object, for while it might show 
continuity, it would give colours which we should be told were not even approxi- 
mately those of the Shirley. We have therefore determined on the second course, 
* Loc. cit., p. 62. 
50—2 
