No. 517] HEAVY AND LIGHT SEED GRAIN 55 



tions it would be somewhat reduced, though this is only 

 problematical. 



Length of head 



Since the above was written it has been found that the 

 small oat heads bear a somewhat larger percentage of 

 single grains than the large heads. If this factor be con- 

 sidered, the negative correlations would be somewhat de- 

 creased but probably not materially. 



COERELATION DATA OF WHEAT 



In looking over the literature some unreduced winter 

 wheat data were found in Bulletin 78 of the Bureau of 

 Plant Industry by Dr. T. L. Lyon. A portion of this 

 data was thrown on to correlation tables and is given 

 herewith. 



Fig. 4 shows the correlation existing between the aver- 

 age weight of kernel as subject and the average number 

 of grains per head as relative. Both the weight and num- 

 ber are averages for the entire plant. The figures are 

 thus not quite comparable to those given for the oat 

 plants, but are doubtless nearly so. Data for only 354 

 individual plants are given, but this number of plants 

 represents a considerably larger number of culms. The 

 correlation is negative and is a little over 11 per cent. 

 The regression of the average number of grains relative 

 to the average weight of grains is much diminished and a 

 selection of seed of either very high or of very low weight 

 would not indicate that the seed was from plants located 

 any appreciable distance from the mean, as far as average 

 number of grains is concerned. 



Fig. 5 shows the correlation existing in winter wheat 

 between the average weight of kernel as subject and the 

 average height of culm as relative. As in the previous 

 case, data for only 354 individuals are given. Instead of 

 a negative correlation, as in all previous cases, we have 



