Aug. 26, 1918 
Variation and Correlation in Wheat 
361 
in the correlation coefficients of the same variety for plants grown in 
the two ways. In any variety, when considerable differences occurred 
in the coefficients as obtained by the two methods, the lower was secured 
by using the plants grown in hills. From this it is concluded that differ¬ 
ences due to spacing may amount to more than varietal differences. 
Variability of yield, number of kernels, number of spikelets, and 
breaking strength of straw decreased with crowding, while for height 
the least variability occurred in hills. 
Hutchinson (1913), in a statistical study of oat plants grown from 
individually weighed seeds planted at definite distances apart, found 
medium high positive correlations between weight of seed planted and 
each of the following characters: Yield of kernels, total weight of plant, 
number of kernels harvested, height before second leaf, height at 4, 6, 
and 10 weeks, at heading, and at harvest. No correlation was found 
between weight of seed planted and average weight of seed harvested. 
In two trials medium to high positive correlation was found between 
yield of grain per plant and average weight of kernels harvested, while 
in another trial no correlation was found between these two characters. 
Atkinson ( 1912 ), using culms as the basis of determinations in a 
statistical study of eight spring-wheat populations grown under field 
conditions, obtained correlations as follows: Between yield and average 
weight of kernel a range from 0.508±0.022 to 0.837 ±0.009, between 
average weight of kernel and length of culm a range of from 0.098 ± 0.030 
to 0.523 ±0.022, and between yield and length of culms a range of from 
o.2i7±o.032 to o.863±o.oo8. 
Meyers (19x2), investigating the effect of soil fertility on variations 
and correlations in wheat, found that variability was decreased by 
increase in fertility, and that all correlations were greatest on the poorer 
soil. 
Roberts (1912) } working with three pure lines of wheat, found that 
variability was reduced in favorable growing seasons and concludes 
that seasonal and soil factors are probably sufficient to overcome heredi¬ 
tary distinctions of yield in good seasons. 
Waldron (19x0), in a statistical study of oats grown under field con¬ 
ditions, found in correlating average weight of kernels with number of 
kernels per plant a negative correlation of — 0.595 ±0.013. Between 
average weight of seed and length of head and average weight of seed 
and length of culm correlation coefficients of — 0.511±0.005 and 
— 0.404 ±0.017, respectively, were found. These correlations indicate 
that the large kernels are borne by the short plants having short heads 
and producing a small number of kernels per head. 
Love (1912), on the other hand, shows positive correlation between 
height of plant and yield, between height and average weight of kernels, 
and between yield and average weight of kernels. 
