8 BULLETIN 754, U. S. DEPARTMENT OF AGRICULTURE. 
unchallenged as very good approximations of actual observation to 
theory. 
The percentages of waxy seeds shown by the individual ears for 
the first-generation hybrids and those previously reported (7) have 
been plotted in figure 1, which shows that the deviation from 25 per 
cent is not the result of a few aberrant ears. It readily can be seen 
that while the mode and mean are considerably below the expected 
25 per cent, the curve very closely approximates the normal prob- 
ability curve. 
Although the graph (fig. 1) bears a striking resemblance to the 
normal probability curve, there is reason for believing that the indi- 
O15 76 /7° A 19 20 2/ 22 23 24 25 26 27 28 29 30 3/ 32 33 
Fic. 1.—Diagram showing the percentage of waxy seeds on 141 ears obtained from the 
second generation of the hybrids Dh 234 and Dh 237 and the ears previously reported: 
The dotted line shows the probability curve for this population of.141 ears, the mean 
and standard deviations being the same. 
viduals do not form a homogenous group deviating from a common 
mean. 
As a means of determining whether the varying percentages of the 
individual ears are chance deviation from the mean of the entire 
group, a method has been proposed by Yule (16) by which a theo- 
retical standard deviation is calculated, based on the harmonic mean 
of the observed individual ears. 
The “ goodness of fit” of the observed standard deviation to the 
calculated standard deviation is measured by the probable error of 
the former. 
