84 MISSOURI BOTANICAL GARDEN. 
of measurement. Only fully matured seeds should be 
weighed, but one fruit of an inflorescence may be quite 
mature, while others are apparently still in a developing con- 
dition. 
The first difficulty demands only time and patience. The 
second may be fairly well met by choosing only fruits with 
well-ripened walls, though even here some of the smaller 
seeds seem immature. 
In 1905, 333 seeds, taken quite at random, were weighed 
and seriated in units of half a gram range. The method 
of collection of another sample of 2,000 in 1906 is described 
under Section 7. Table VI gives the data. 
TABLE VI. 
Class 1905 | 1906 Class 1905 | 1906 
1=0 —0.50 22 219 || 10 = 4.51 — 5.00 | 10 39 
2 = 0.51 — 1.00 31 396 || 11 = 5.01 — 5.50 5 27 
3 = 1.01 — 1.50 57 352 || 12 = 5.51 — 6.00 6 26 
4 = 1.51 — 2.00 33 | 290 || 13 = 6.01 — 6.50 6 29 
5 = 2.01 — 2.50 45 | 212 || 14 = 6.51 — 7.00 5 13 
6 = 2.51 — 3.00 40 147 || 15 = 7.01 — 7.50 2 5 
7 = 3.01 — 3.50 20 106 || 16 = 7.51 — 8.00 1 6 
8 = 3.51 — 4.00 26 76 || 17 = 8.01 — 8.50 2 1 
9 = 4.01 — 4.50 22 55 || 18 = 8.51— 9.00] .... 1 
The distribution of weight of seed, like that of the num- 
ber of seeds per fruit, cannot by any stretch of the imagina- 
tion be regarded as conforming to the “normal,” Gaussian, 
or “‘Quetelet’s’” curve, which some biologists seem to regard 
as the biologically normal type. The skewness is very con- 
spicuous. The great majority of the seeds remain very 
small—those attaining a large size are rare. To what extent 
this may be due to a competition of the numerous ovules 
for a supply of food material, which is inadequate to bring 
many seeds up to the large size which we think of as 
characteristic of the bulbiform seeds of the Amaryllidaceae, 
can only be determined by further investigation along lines 
indicated later. © 
