APR. 1912.-J H. NAKANO.— VARIATION IN THE SEEDS AND PULP. 



87 



De Vries'^ mentions many examples of the variations of 

 one-sided curves and states that such variations generally occur 

 in the cases of " Halbrasse."-^ Whether our unshumikan is a 

 hj'brid or not, has not yet been decided, but it seemes to be a 

 very likely conjecture, becauce the degeneration of pollen is very 

 often noticed in hybrids. This opinion has already been stated 

 by Mr. I. Osawa. 



As regards the meaning of " halbe Galton-Kurve," it may 

 be quite right, as Johannsen mentions'^^ to explain that it is no 

 more than a special case of skew curves. As will be seen from 

 the following pages, the curves of the variation of seeds in 

 Kojimikan and Kishumikan are almost symmetrical, and the 

 curve of Fukuremikan is nearly' one-sided, while that of Uushu- 

 mikati is quite half a curve. Thus it is clear, that the skewness 

 reached its maximum in Unshumikan. 



As the data relating to Unshumikan are insufficient, the cor- 

 relation-coefficient calculated will be also a defective one, which 

 shows 0.309±0.092. 



Finally the correlation tables in Kojimikan and Kishumikan 

 will be shown as follows : — 



Kojimikan 



r=0.414±0.032, 



X 



:0.414 



1) Ueber halbe Galton-Curven als Zeiclien discontinuirlicher Variation Ber. d. 

 deutsch. bot. Geselsch. 1894. p. 197-207. 



2) Mutationstheorie, Bd. I. p. 428-435. 



3) 1. c. p. 192. 



