J. A. Harris 
53 
between the number of pods on a plant and the fertility of the pods was so a priori 
evident that its demonstration on the bases of actual material was superfluous. 
The determination of the degree of interdependence of somatic characters and 
fertility in plants presents some special difficulties as compared with the working 
out of the same problem on zoological material. The chief obstacle is the fact 
that in the higher plants the reproductive organs are not so specialized nor so 
clearly differentiated from the other organs as in the higher animals. It is, there- 
fore, difficult to obtain " somatic " characters which are not directly connected up 
with the reproductive organs either (1) as a means of support, (2) as a conducting 
system for plastic materials, or (3) as a direct source of food supply. 
This difficulty has been incurred in the attempts which have been made to 
measure the relationship between the vegetative and reproductive organs. In the 
case of the correlation between diameter of the stem and fruit length in Oenothera 
studied by de Vries* it must be remembered that the stem not only served to 
support the fruits but transported the material of which they were built up 
and bore the leaves which are largely concerned in its manufacture. In the study 
of the relationship between length of flowering stalk and number of flowers in 
Nothoscordium and Alliunvf where a correlation of about '500 was found, it is 
certainly quite impossible to regard the length of the stem as a strict " somatic " 
character and the number of flowers as an independent " fertility " character j, so 
intimate is the structural relationship of the two. The same is true for the inter- 
dependence between the number of leaves and flowers per inflorescence in Spiraea 
Vanhouttei to be published later. In Sanguinaria§, where a correlation between 
length of peduncle and number of ovules produced and seeds developing per fruit 
has been demonstrated, there is perhaps less reason to think of a necessary 
structural interdependence of the characters. 
But when one takes the relationship between the number of fruits or flowers 
per inflorescence and the number of ovules or seeds maturing per fruit there is no 
patent morphological reason why there should be a correlation between number of 
fruits per inflorescence and number of ovules per fruit, although the biologist 
might say that the competition of a number of fruits above the average for plastic 
material would result in a negative correlation for a number of fruits and number 
of seeds per fruit. 
Very low correlations have been found for both number of flowers and number 
of fruits per inflorescence and number of seeds per fruit in Gelastrus\\ and for 
number of ovaries per inflorescence and number of ovules per ovary in Gercis and 
in a large series of unpublished observations on Staphylea — now nearly ready for 
the press. 
* Die Mutationstheorie, Vol. r. p. 113, 1901. 
t Harris, J. Arthur : Ann. Rep. Mo. Bot. Gard. Vol. xx. pp. 105—115, 1909. 
J Although for convenience I have occasionally referred to them as such. 
§ Harris, J. Arthur : Biometrika, Vol. vn. pp. 314—317, 1910. 
|| Harris, J. Arthur : Ann. Rep. Mo. Bot. Gard. Vol. xx. p. 122, 1909. 
