■180 



THE AMERICAN NATURALIST 



[Vol. XL VI 



Table VII, fl^V) =763,048.0000, r=.1861. 



While primarily illustrations of method, these results, 

 if they are substantiated by further work, seem to me of 

 considerable biological interest. They show not only that 

 individuals of H. Syriacus differ in the radial asymmetry 

 and in the locular composition of their fruits, but that 

 when an individual bears fruits above the average 

 asymmetry, it also produces fruits above the average in 

 number of "odd" locules. Apparently, this cross corre- 

 lation is as high as either of the direct correlations. 



Two biological interpretations are possible, {a) The 

 production of radially symmetrical ovaries and those 

 with a high number of odd locules depends upon the same 

 morphogenetic tendencies of the primordia, 13 which give 

 rise to the fruit, (b) There is in Hibiscus an intra-indi- 

 vidual selective elimination similar to that demonstrated 

 in Staphylea, 14 the intensity of which differs from indi- 

 vidual to individual in such a way as to bring about 

 (statistical) correlation for characters originally uncor- 

 related. 



The discussion of these points falls outside the scope 

 of the present note where the data serve merely as a 

 random illustration of a very rapid method of carrying 

 out the routine of a widely applicable statistical process. 

 Cold Spring Harbor, 

 April 25, 1912 



necessarily associated (cf. BiometriM, Vol 7, pp. 491-493, 1910). In 

 Staphylea, correlations of r = .22 to r = .33 have been noted. Table HI 

 above gives r = .527 for asymmetry and locular composition of the same 



Probably in all these relationships regression is not linear, and the corre- 

 lations must be interpreted with caution. 



" Biometrilca, Vol. 7, pp. 452-504, 1910; Science, N. S., Vol. 32, pp. 519- 

 528, 1910; Zeitschr. f. Ind. Abst. u. Vererbungsl., Vol. 5,'pP- 273-288, 1911; 

 Pop. ScL Mo., Vol. 78, pp. 534-537, 1911. 



