BIOMETRIC DATA IN CRINUM LONGIFOLIUM. 91 
above the average tends to cut down the number of seeds 
in these fruits, but much confirmatory evidence must be 
obtained before great stress can be laid on such a weak 
correlation. 
The only directly comparable series are one for Celastrus™* 
and several for Staphylea. For Celastrus there is no trust- 
worthy indication of an interdependence between the num- 
ber of flowers formed per inflorescence and the number of 
seeds maturing per fruit, or between the number of fruits 
matured per inflorescence and the number of seeds matured 
per fruit. . 
For Staphylea the results show the most slender relation- 
ships between numbers of fruits per inflorescence and the 
fertility of the fruits. Data for both ovules and seeds are 
available. For twenty individual shrubs of 1906, the mean 
relationship for number of fruits per inflorescence and 
number of ovules per locule is only r= -+ .0192 + .0185 
—a value of no significance. For fruits and seeds it is 
7 = — .0399 + .0080, indicating that in Staphylea the con- 
dition may be similar to that in Crinum. For the whole 
populations the results are :— 
For Ovules. For Seeds. 
1906, r= + .0391 + .0086 7 = — .0474 + .0086 
1908, r = + .0633 + .0061 r= — .0494 + .0061 
1909, r= — .0539 + .0085 r= -+ .0626 + .0085 
Taken as a whole, the data for Staphylea indicate a very 
slight negative correlation between the number of fruits 
matured and the capacity of these fruits for developing 
their ovules into seeds. 
13 Waldron (Am. Nat. 44:48-56. 1910.) concludes that the seeds 
are lighter in large heads of oats than in small ones. Harris (Torreya 
11 : 165-169. 1911.) has shown that in Staphylea and Cladrastis, the 
weight of the seed is in some measure dependent upon the number pro- 
duced in the fruit. As noted above, other data are in press, and still 
more are being reduced. 
14 Harris, J. Arthur. Ann. Rept. Mo. Bot. Gard. 20 : 116-122. 1909. 
15 Harris, J. Arthur. Beih. Bot. Centralbl. I. 28:2-10. 1912. 
