III. Inheritance in Shirleif Poppy 
409 
(5) T]ie Injiuence of Variahility of like Organs in the Plant on Coefficients of 
Heredity. Homotyposis. 
We have already seen that if all the flowers of a plant were identical, or if the 
first flower were as markedly differentiated from other flowers as one digit in a 
man is from a second (i.e. if there were a very marked correlation between order of 
flowering and character), the problem of inheritance in plants would be much 
simpler. But the variation of like organs in a plant is very considerable, and in 
many cases the degree of variation is not highly correlated with the po.sition or 
order of production. Now if there were no such correlation, it has been shown 
elsewhere that the extent of the variation of the individual plant — its horaotyposis 
— would be properly measured by the coefficient of correlation of like organs. 
Hence our selection for record of apical flowers will mean that we get closer to the 
true values of the inheritance coefficients than by using all or any flowers on the 
plant, provided 1st there be marked differentiation of the apical flower from other 
flowers (i.e. if the first flowers of two related plants are much more alike than the first 
flower of one and any other flower of the second), or again 2nd there be no variation 
of the character observed in the plant at all. For example, colour and margin in 
the broad classification of our records do not exhibit much variation in the plant. 
Number of stigmata, size of local patch and wrinkling, do exhibit much variation, 
and further this variation is by no means perfectly correlated with position or 
order of budding. In order therefore to allow for the neglected homotyposis, we 
ought to have obtained measurements of (i) the correlation between the order of 
flowering and the characters observed, and (ii) the homotyposis of these characters. 
The first would have enabled us to measure the differentiation of the first or apical 
flower from other flowers, and by the aid of the second result we could have made 
allowance for the neglected homotyposis. 
But to record some half-dozen characters on all the flowers of each plant would 
have multiplied the work of recording tenfold, and the task seemed at present 
beyond the volunteer labour, which was all that was available. Accordingly we 
must simply state that our inheritance coefficients are too low because we have at 
present no adequate measure of the differentiation of the apical flowers, nor of the 
homotyposis of most of the characters dealt with. But it is satisfactory to note 
that those characters in which mere inspection has shown least differentiation 
with position and most variation in the individual, such as number of stigmata, 
size of basal patch and wrinkling, are precisely those in which the coefficients 
of inheritance come out lowest. We hope that investigations into character 
differentiations with position and into homotyposis will be started this season. 
Meanwhile the results in sections (3) to (5) will, we hope, suffice to indicate to 
the reader that in our opinion the study of inheritance in plants is not a simple, 
straightforward matter. We have not in Shirleys at any rate found the definite 
and exclusive categories which the Mendelians appear to encounter at every turn. 
We have found our characters widely modified by environment and season ; we 
have found selection acting on the seed sown, and both type and variability 
Biometrika iv 
