PLANT-BREEDING METHODS 431 
yields of the ultimate units, p, of the same combination plot, C,, the measure 
of heterogeneity required. If S indicates summation the formula, is 
_ LUS(Co4) — 8(@*)I/min(n — 1) = 3" 
py. ae ose 
where n is constant throughout the m combination plots. 
Surface and Pearl have devised a method of correcting for soil 
heterogeneity which when tested by Harris’ method was found to give 
in a row 20 inches long. (After Montgomery.) 
in all cases a very marked reduction in the amount of heterogeneity when 
the corrected figures were used, and when tested experimentally, it 
seems that this method leads to results which more nearly represent the 
truth than do the uncorrected yields. However, this method in its 
present form is adapted for use only when the plots are arranged in a 
particular way which is not always practicable with certain crops or 
on certain areas. It is probable therefore that the older method of 
check plots or rows together with replicate planting will continue to be 
used. Following are some of the conclusions reached by Montgomery 
regarding the reduction of experimental error: 
(a) Systematic repetition constantly reduces error as the number of repeti- 
tions increases, but with 16-foot row plots 10 to 20 repetitions must be made, 
depending on the degree of accuracy required. 
Digitized by Microsoft® 
