220 
Miscellanea 
if we suppose that proximity of organs does not indicate that they have any function to perform 
in common which would tend to render their correlation organic rather than homotypic*, 
it would seem quite reasonable to think that organs separated in their embryological develop- 
ment by only relatively few cell divisions would be more similar than those separated by an 
immensely large number. Concretely, the flowers of an inflorescence of Adoxa have been built 
up from a common vegetative point by only a few cell divisions as compared with the in- 
conceivably large number which have intervened between the laying down of the leaves upon 
the outermost branches of a beech tree. In the laying down of the leaves of a forest tree there 
is also opportunity for considei'able differences in the environments of the individual vegetative 
points. The actual correlation constants, calculated from the symmetrical homotypic tables 
IV— VI, are given in Table B. 
The conspicuous thing about these constants is that they are not higher, but rather lower, 
than those usually found for vegetable homotypes. Comparing the results summarized in 
Pearson's Table XXXII t, we note, however, that Ash, Wild Ivy and one series of Holly 
give values of about the same general order of magnitude as tliese. In a study of the correlation 
between the two leaves of the flowering stalk of Poclophi/llum I have shown J that the relation- 
ship in two series of material is about -45, although here too I should have expected higher 
values. Aside from any bearing which the problem of homotyposis may have on heredity, 
I think it may throw some light upon a number of morphogenetic problems. One of these 
is the question of a possible relationship between the homotypic constants for organs which are 
widely separated ontogenetically and those which are closely associated ontogenetically § . 
One further point may be noted. In any critical study of interindividual correlation in the 
Moschatel the possibility of the individuals being related vegetatively must be taken into 
account. Pearson found that in Malva. rotundifolia the homotypic coefficient was considerably 
lower than that found in most forms, and concluded that this was due to the jilants having for 
the most part spread from one clump by stolons. 
TABLE A. 
Variation and Orgamc Correlation Constants for Adoxa. 
Material 
Means 
Standard 
Deviation 
Coefficient 
of 
Variation 
Coefficient 
of 
Correlation 
Kent, Flowers 
„ Divisions ... 
4 -042 -f -007 
4-819 ±-006 
-489-1- -005 
-440 ± -005 
12-09 
9-14 
- -135 ±-014 
Surrey, Flowers ... 
„ Divisions... 
4-190-1- -024 
4 -803 ±-020 
-604+ -017 
•500 ±-014 
14-423 
10-41 
- -214 ± -038 
Essex, Flowers 
„ Divisions ... 
3- 994+ -007 
4 - 862 ±-009 
-278+ -005 
-357 ± -007 
6- 95 
7- 35 
+ -lG3±-026 
Flowers = Number of lateral flowers per inflorescence ; Divisions = Number of divisions 
of corolla. Sheppard's Correction not used. 
* See in this connectiou Pearson's very clear statements of the differences between organic and 
homotypic correlation in Biomctrika, Vol. i. p. 341, 1902. 
t Pliil. Trans. 1901, Vol. 197, A, p. 356. 
t Bot. Gaz. Vol. 47, pp. 438—444, 1909. 
§ Possibly the statistical difficulties introduced by the "small samples" necessarily used in studies 
of interindividual correlation in such forms as Podophijlliim and Adoxa may obscure the real biological 
relationships. See Pearson, Biometrika, Vol. vi. p. 403, 1909. 
