BIOMETRIC DATA IN CRINUM LONGIFOLIUM. 93 
From the material collected in the spring of 1906, it is 
possible to form correlation tables for the number of seeds 
per fruit in the same inflorescence. 
A difficulty at once arises in regard to the selection of 
fruits to represent the inflorescences. The complete failure 
of some flowers to produce seeds may, perhaps, affect the 
number which develop in other ovaries of that inflorescence. 
It is difficult to decide in advance whether all pedicels shall 
be chosen quite at random and the sterile ones inserted in 
the series, or whether all pedicels bearing no fruits shall 
be excluded. 
From the 1906 lot of material, it is possible to extract 
two series. The first comprises 249 inflorescences bearing 
five fruits each. It would have been slightly better, perhaps, 
if all the fruits from these inflorescences had been used, 
but the labor of dealing with the correlations is excessive. 
The second series comprises 280 inflorescences each yield- 
ing eight pedicels with or without fruits.1® In both of these 
series the pedicels or fruits were chosen quite at chance 
by a system of numbering the recorded data backwards and 
forwards. 
The tabling of the data presented a problem of the 
greatest difficulty. As compared with many symmetrical 
intra-class correlation surfaces which have been formed, 
the number of entries is small, being only 5 x 4 « 249 = 
4,980 for the table containing fertile fruits only and 8 & 7 X 
280 — 15,680 for the surface comprising both sterile and 
fertile rays. But the range of variation in number of seeds 
is so great—from 0 to 70—that the tables are exceedingly 
large, comprising 4,900 and 5,041 compartments respec- 
tively. It is not safe to reduce the surface of the table by 
grouping, because of the skewness of distribution. 
18 Sterile fruits were counted in with those in which the fruit failed 
to develop, since the persistence of the ovary, when no seeds develop, 
cannot be regarded as forming a separate category as far as our 
present considerations are concerned. Our pedicels are classified accord- 
ing to the number of seeds which they produce, hence “no fruit” and 
“sterile fruit”? both fall in the zero class. 
