J. A. Harris 
329 
IV. Recapitulation. 
The present paper embodies the results of a quantitative study of the in- 
florescence and fruit of Sanguinaria with special reference to morphological and 
physiological problems. The results presented are chiefly of value as descriptive 
quantitative data. The constants will gain iti significance as others from organs of 
distinct morphological type are laid beside them for comparison. The following 
points may be noted. 
1. The variation and correlation constants for two collections of Sanguinaria 
made in the same habitat for two successive years show very close agreement for 
some characters and significant divergence for others. The significant differences 
do not necessarily indicate that the populations in the two years are really different 
for the material was not collected in a manner to secure the same proportional 
sampling of subhabitats in the two seasons, and these may differ enough among 
themselves to introduce a slight heterogeneity into the material. These results 
furnish another illustration of the difficulty of establishing " place constants " for 
an organism easily affected by environment. On the whole, the two series may be 
regarded as very satisfactorily similar as material for a morphological and physio- 
logical investigation. 
2. The length of the peduncle and the length of the fruit are to some extent 
interdependent, the magnitude of the relationship being measured by a coefficient 
of correlation of about "SoO. Number of ovules formed, number of seeds developing 
and number of aborted ovules per fruit are also correlated with peduncle length. 
For both ovules and seeds developing the correlation is about "300 — '350, the 
intensity of the relationship being slightly but not significantly higher for length 
of fruiting stalk and number of seeds developing than for length of fruiting stalk 
and number of ovules formed. 
Since our correlation coefficient is purely descriptive, we cannot say that the 
length of the peduncle has any influence upon the length of the fruit or upon its 
fertility. All that we can do at present is to measure the degree of interdependence 
of these pairs of characters. Possibly neither has any real influence upon the 
development of the other and their measurable interdependence is due merely to 
the age or physiological vigor of the plants which produced them. Interdependence 
of their magnitude would then be merely the result of their mutual dependence 
upon other factors. Perhaps questions of this sort can be solved only by appropriate 
experiments, but we sorely need quantitative data of the kind presented here for 
numerous species to guide us in the intelligent planning of experiments. 
3. It is not surprising to find that length of fruit and number of ovules formed 
and length of fruit and number of seeds developing are interdependent, but it is 
important to measure the degree of coi-relation between them in such a way that 
comparison with other forms will be possible. As indicated by our material the 
correlations for length and ovules or seeds developing lie between '500 and '700. 
