Vol: 30 Page: 42

Authors:
Wiesław Pilarczyk
Kamila Tomaszyk

Title:
 The influence of proposed changes in trial methodics on assessment of sugar beet varieties

Language: Polish

Keywords:
sugar beet
series of trials
trial methodology

Summary:
The problem of number of measurements for quality characters in sugar beet series of trials is reviewed. In trial methodology of variety testing in sugar beet trials in Poland, the values of quality characters (sugar content and content of Na, N and K) are measured in independent samples taken from every plot. So there are four measurements available for every variety in each trial, because as a rule four replicate experiments are in use. Lately the idea of reducing the number of measurements for quality characters appeared. In order to investigate the consequence of reduction of the number of measurements for quality characters in series of experiments on variety assessment, the results of sugar beet trial series conducted in 1998 in Poland were used. All the data was elaborated by two methods. In the first one all available data was used, and classical method of analyzing series of experiments with application of mixed model was used. In the second method the same approach was kept, but only part of data was used. Namely, from among four replicates in every single trial. one was randomly chosen, and measurements of quality characters from this replicate were used in calculation of biological and recoverable sugar yield for all replicates. The mean values given by these two methods for the same varieties were then compared. The ranks of varieties given by two methods being compared. The values of that coefficient varied from 0,64 in the weakest correlation case to 0,97 when the correlation was the highest. In the series of trials this coefficient was equal to 0,97 for biological sugar yield to 0,96 for recoverable sugar yield. The differences in variety assessment can be decreased if a bulk sample from all replicates is used instead of a sample taken from a chosen single replicate.