314 Morphology of the Fruit of Sanguinaria Canadensis 
^ = •03149, and ^/E^ = 4^-279. It seems rather questionable, therefore, whether 
our regression line can be regarded as strictly linear within the limits of the 
probable errors of random sampling. 
The equation of the regression straight line is 
2/ = 26-5l7 + -772.T. 
I have plotted the means of arrays and find that they are scattered with some 
irregularity on either side of the theoretical line, but considering the nature of our 
material the fit is perhaps as good as could be expected. There is no suggestion 
in the graph that any other curve would give better results. 
The biological significance of this constant showing the interdependence 
between length of flowering stalk and length of fruit seems to me to be chiefly 
descriptive. We cannot say that the length of the stalk determines to some 
extent the length of the fruit, or vice versa. In short, we have not "explained" 
anything, as we biologists are so wont to do. But we have measured a relation- 
ship and recorded it in a form that may be directly compared with other similar 
measurements when they shall have been made, and when the series so described 
are enough, comparison will permit of pushing problems further. 
3. The Relationshij) between the Length of the Peduncle and the 
Fertility of the Fruit. 
We have little definite knowledge of the relationship between the physical 
development of an organism and its fertility. In man Pearson * has brought 
forward evidence to show that tall women are more fertile than short ones. I am 
not aware of further published evidence for animals, though Professor Pearson 
informs me that data for other mammals are in existence and will shortly be 
published. 
In plant material it has been shown in an earlier paper f that there is a 
correlation of about "5 or over for length of flowering stalk and number of flowers 
per umbel in N othoscordium and Allium. In the case of GercisX and Gelastrus% 
however, there seems to be no correlation between the size of the inflorescence 
as measured by the number of flowers or fruits produced and the fertility of the 
ovaries. Before we can go very far into the problem of the relationship between 
the degree of development of the vegetative and reproductive characters we must 
have together a sufficient number of series of reduced data to enable us to make 
extensive comparisons. 
In the table of correlation constants I have given the relationship between 
the length of the peduncle and the fertility characters of " first " and " second " 
* Pearson, K., Proc. Roy. Soc. Vol. lix. pp. 301—305, 189G. 
t Harris, J. Arthur, Ann. Eept. Bio. Bot. Gard. Vol. xx. 1909. 
X Harris, J. Arthur, Biometrika. In press. 
§ Harris, J. Arthur, Ann. Rept. Mo. Bot. Gard. Vol. xx. 1909. 
