BIOMETRIC DATA IN CRINUM LONGIFOLIUM. “i 
counted the number of seeds in a random sample of 1,000 
fruits, (b) weighed a series of 333 seeds individually, (c) 
determined the number of flowers formed and number of 
fruits developing as well as the number of seeds per fruit 
in a series of individual inflorescences. 
Collections were again made in 1906. In 309 inflo- 
rescences, the number of seeds per fruit was determined in 
all fruits possible.2 A lot of 2,000 seeds was weighed. 
III. CONSIDERATION OF PROBLEMS. 
1. Variation in Number of Flowers and Number of 
Fruits per Inflorescence. 
Data for number of flowers formed and number of fruits 
developing per inflorescence are given in the totals of the 
correlation tables for 1905, Table VII, and for 1906, Table 
VIII. The constants are given in Table I. 
TABLE I. 
1905 1906 Difference 
Flowers 
Mean......... 10.668+ .126 | 11.518+4 .101 850+ .161 
S. Dw ee 2.594+ .089 2.626+ .071 032+ .113 
C.V.........] 24.8164 .883 | 22.795+ .650 1.520 + 1.096 
Fruits 
Mean: .-....-.- 5.689 + .093 7.065+ .086 1.38764 .127 
Sh] Oe ee | 1.917+ .066 2.242+ .061 .325+ .090 
(Ca Aa 33.699 +1.282 | 31.731 41.291 1.968 + 1.819 
2It is almost impossible to collect inflorescences at a time to 
secure the well-developed pods without including some which have lost 
part of their seeds. There is considerable variation in the time of 
flowering and a few inflorescences are still in flower when the bulk 
of the materials must be collected. The later-flowering inflorescences 
probably produce fewer or smaller fruits, but I have no means of 
making a quantitative estimate of the difference. The material col- 
lected both years was taken almost, if not quite, entirely from the 
maximum period. Our material is, therefore, to some extent selected. 
Thus while the statistical investigation of this form is of very great 
interest, it is surrounded by many difficulties; the results must be 
interpreted with caution. Some of the difficulties met are discussed 
in their proper context. 
