8 o 6 
Journal of Agricultural Research 
Vol. XXI, No. II 
mentioned cause may be made to disappear by arranging the abscissas in 
accordance with dates of counting. 
For all countings except June 24 the median is smaller than the mean, 
and in these cases the skewness is presumably positive. Positive skew¬ 
ness may perhaps usually be presumed in cases of this kind. A certain 
number of plants progress slowly, while the values of the remainder are 
scattered over a wide range, but all in advance of the earlier stages. This 
is indicated by a study of the extreme ranges. The differences between 
the minimum values are much less than the corresponding differences 
between the maximum values. 
More interest centers about the discussion of the mean number of 
culms taken seriatim. To facilitate discussion these means have been 
plotted and are shown in figure i as the solid line. The abscissas in the 
figure for the observations are placed in accordance with the dates of 
counting, which are usually five days apart. The plotted observations 
show, on the whole, a very smooth series. From an inspection of the 
graph there is evidently something of a spurt in culm formation around 
July 1 and also a marked falling off around August 5. There were no 
obvious weather changes to account for these comparatively slight modi¬ 
fications from what one might consider the normal course of events. 
