igis] YASUI—DIOSPYROS KAKI 363 



were observed, they were dioecious, and bore only perfect flowers. 

 One possible exception has been found recently." In D. Kaki, on 

 the other hand, there are many such exceptions, as follows. 



1. In some varieties only the pistillate flowers occur in good 

 condition, while all the staminate and some of the pistillate flowers 

 are imperfectly developed. 



2. In other varieties (gosho, egosho, yamagaki, jenjimaru) 

 both kinds of flowers are always well developed. 



3. A tree in Mr. Yabe's garden in Tokyo (no. 1 of my 

 material) has the same type of fruit as in some of the varieties in 

 group 1, but has pistillate, staminate, and perfect flowers, and 

 produces also two types of fruit from the two kinds of flowers. 



4. No tree has been discovered by the writer on which only 

 staminate flowers occur. 



5. Some varieties have pistillate flowers only, when the trees are 

 young, but as they become older or weaker (on account of poor 

 nutrition or transplanting) they begin to produce staminate flowers 

 also (hyakume-gaki) . 



6. Some trees show an alternation of fruiting and non-fruiting 

 years, which means that during one year they produce many pistil- 

 late flowers and many fruits, while during another year they pro- 

 duce a few pistillate flowers and abundant staminate flowers. 

 Under the latter conditions they bear very little fruit, or no fruit 

 at all. 



It may be added, in reference to the fourth condition referred 

 to, that in Japan the staminate trees are being continually destroyed 

 unconsciously, because the better varieties are produced always by 

 grafting, and therefore many seedlings of staminate plants are 

 destroyed, and pistillate plants take their place. The foregoing 

 data gave me the suggestion that D. Kaki is naturally a monoecious 

 plant (fig. 1), and that it is in process of losing this character, pro- 

 ducing staminate flowers under cultivation. 



I have selected five different varieties from one hundred or more 

 for this study, as follows. 



No. 1. — I am not certain of the name of this variety, but it 

 resembles "yemon," which is an astringent, flat type. The differ- 

 ence is that yemon has no staminate flowers at the Okitsu 



