80 MISSOURI BOTANICAL GARDEN. 
The series (c) and (d) bear to each other the relation of 
random sub-samples. Some of the fruits in the two series 
are identical. The two are inserted merely to show the 
reader how considerable irregularities may arise through 
errors of random sampling. Series (a) and (b) are both 
samples from the same population, but taken in a somewhat 
different manner. Diagram 1, in which the heights of the 
several vertical lines represent the frequencies for the 57 
classes of fruits, gives the 1905 random sample. The sta- 
tistical constants are given in Table IV. 
TABLE IV. 
A Standard Coefficient 
Series Mean Seeds Deviation of Variation 
ile See ae 10.338 +.221 © 10.342 +.156 99.55 
(aa); OG hee oe 9.242 + .246 10.095 +.174 109.23 
COR. BOO. pe eead ss 15.471 +.225 13.344+.180 86.25 
(i FO. 15.283 + .239 12.941 4.169 84.67 
Two characteristics of these distributions are remarkable. 
The range of variation is exceedingly wide. The mode lies 
far to the lower end of the range, the most frequent cases 
being fruits with but a single seed or, at most, three seeds. 
