Mai:. 1012.1 H. .NAKAXO.— rAELlTlOX IN THE SEEDS AND PULP. 



69 



of the four mandarins in the number of their seeds and vesicles. 

 Besides these I shall give my attention to the correlation bet- 

 ween the number and weight of the fruit, and between the 

 number of seeds and vesicles. 



The collection of the fruit took place only in the neighbour- 

 hood of Tok\'o, because my present profession prevented me 

 from extending my observations to the various provinces. 



A. habitat is situated on a loamy hill in the northern part 

 of the province Simosa, 30 miles east from TokA'O. 



B. habitat is also situated on a loamy hill in northern 

 Simosa, about 7.5 miles east from A. 



C. habitat is situated on the alluvial land near the western 

 end of the swamp " Kasumigaura," in the province of " Hida- 

 chi " about 15 miles north from A. 



All the trees examined were not pruned, being in almost all 

 cases in their natural condition. 



The counting of vesicles was carefully carried out by 

 separating them one by one. Two united vesicles were counted 



as one 



1) 



The variation of vesicles. 



On 3rd Nov. 1909 I collected the fruits from a Fukuremi- 

 kan tree of about 4.8 m high in A. habitat. The fruits were 

 divided into two groups, one from the branches, higher than 

 2.4 m, the other from those lower than 2.4 m. The collecting 

 was made without selection and it took place from the same 

 side of the upper and lower branches. The following data were 

 obtained : 



From the lower part. 



1) As to the phenomena of the dedoublement, we may consider the separation or 

 the fusion ^ f two elements. (Comp. Gobel : Organographie 1898 p. 710). Which 

 case is correct has not yet been ascertained in onr vesicles. 



