Aug. 3, 1920 Daily Development of Kernels of Hannchen Barley 399 



to visualize the changes that take place from day to day, the nature 

 of the individual spikes can be seen only in this table. The averages 

 do not give a correct indication of the condition of a single spike. The 

 variations between kernels are reduced by averages, and the difficulty 

 of securing such averages is not apparent in the absence of the complete 

 data. It is readily seen in the table that in instances where the spike 

 is short it is often a question which kernel should be considered the 

 third or the fourth. Such decisions affect the averages, and they must 

 be made by some arbitrary method, since the actual number of sterile 

 nodes at the base can not be used successfully as a basis. The curve of 

 the kernels of a single spike usually is better than the average of two 

 spikes unless the two spikes have the same number of kernels. The 

 extremes of the curves are especially liable to distortion in averages. 

 Of course, the curve of the growth by days is much improved by the use 

 of averages. The same process that reduces the fluctuation in the 

 record of a spike increases the gap between samples. 



Table I. — Normal growth of Hannchen barley in 12-hour periods from flowering to 

 maturity, at Aberdeen, Idaho, in igij 



JULY 16 



» The letter S indicates a sterile spikelet. 



t> Unfertilized. 



