240 'I'lIK INIIKRITANCK OK TIIK FI .( AVI'.RIKC; TIMK IN PKAS AND RICE 
as u'c had expected. 
The frei]uency distribution of representative values calculatetl by the 
above mentioned fornuihi in tlie G- P- offsprings, is show n in the same table. 
(The designation C nicans a family of lOO plants t,n-o\vn from the stock seeds) 
Here, we see much irrej^ularity of variation. The offsprings of those plants 
which nowered early in the forcing house (I. XV. XI. XVI.) flowered dis- 
tinctly earlier than those of the late-flowered plants (XIV. VI. VIII.). In 
the family C, the variation range is quite wide and there are two distinct 
classes with a larger frequency (class 50 and55). 
Raising in 1911. 
From the plants grown in the previous year, we selected the following 
individuals : — 
2 individuals from M- P. i 
>» >> 
»> >» 
3 
>• >> 
>> »» 
j> >> 
20 
In selecting these individuals from each family, we took care to choose 
those which flowered on different dates and which made normal growth. We 
sowed 30 seeds from each selected individual on May 6. After the sowing, 
extraordinarily dry weather followed, and during the 3 days from the 15th 
to the 17th of the month, only a partial sprouting occurred, and afterwards 
the sprouting stopped almost entirely. So, we were obliged to water arti- 
ficially on the ungerminated parts. From the 26th of the month they were 
watered every day. As a consequence, sprouting began to take place from 
the 31st of the month and continued until the 4th of June. 
Here, again, we can not use simply the number of days from sowing to 
flowering as a representative value of the flowering time. The mode of 
» 
}> 
6 
)> 
>> 
12 
>» 
G P 
I 
>> 
>> 
XII 
)) 
)> 
VI 
ft 
VIII 
)> 
)) 
c 
