The reliability of nursery tests as shown by correlated yields from nursery rows and field plats

Varieties of common spring wheat, durum wheat, oats, barley, and flax were grown for periods of 3 and 4 years in both regular variety test plats and nursery rows. Yield data obtained from these two sets of plats were compared from the standpoints of degrees of correlation between them and on the basis of rank agreement. Rather significant correlations were found between the yields of regular variety test and nursery plats. While the values of the coefficients correlation found for most seasons were perhaps not high enough to justify an outright recommendation for the entire substitution of the triplicated nursery row method of testing in favor of the employment of the regular variety test plats of 1/50 to 1/60 of an acre, they show that very accurate yield data can be obtained by the use of nursery plats. The 4-year average yields of seven varieties of common spring wheat from variety test and nursery plats showed a correlation of r = .9520 +/- .0241, and an identical ranking in the two sets of plats. Similar correlations of the yields of 10 varieties of oats gave a value of r = .9065 +/- .0380 with good rank agreement. The values of the correlation coefficients for four separate seasons ranged in common spring wheat from .5692 +/- .1219 to .8140 +/- .0608, with an average value of .7205 +/- .0465; in oats from .3202 +/- .1563 to .7122 +/- .0888, with an average of .5620 +/- .0613; and in barley from .2943 +/- .1647 to .7333 +/- .0865, with an average of .5429 +/- .0512. Three years yield data of durum wheat showed correlations ranging from .4590 +/- .2013 to .7007 +/- .1298 with an average value of .5645 +/- .1003. The yields between the flax variety test plats and nursery showed as light negative correlation of -.0265 +/- .2247 in 1930 when there was a difference of 18 days between the planting dates of the nursery and the variety test plats. In 1931, the value of r = .7042 +/- .1133 resulted from the correlation of the yields of the two sets of flax plats. No doubt the relationships between the yields of the regular variety test and nursery plats would have been even greater than indicated by the figures cited above had it been possible to have had the two set of plats adjacent to each other in all seasons instead as in most years of the comparison at some distance and not infrequently under somewhat different soil conditions. Nevertheless the results indicate that a great deal of significance can be attached to yield data from properly conducted nursery tests.