Reproducibility of grain hardness measurements in wheat

In line with the quality classification used by the processing industry, it would be useful for wheat varieties and commercial wheat lots in Hungary to be grouped on the basis of kernel hardness. This will require a rapid, accurate system of measurement. In the present experiments the kernel hardness of 18 wheat varieties was examined in ten replications with two instruments to determine the reproducibility of the measurements and the correlation between physical measurements and those made using the NIR technique. The analyses were carried out with a Perten SKCS 4100 instrument on whole kernels and with an Inframatic 8611 NIR instrument on wholemeal and flour. The analysis of separate kernels showed that the kernel hardness of soft-grained varieties was less uniform than that of hard-grained genotypes. Hard- and soft-grained varieties could be separated clearly with both instruments. Even small differences (3-5 Hardness Index values) between varieties of the same kernel type could be demonstrated significantly with the SKCS 4100 instrument. The NIR analysis was less accurate, the confidence interval of the measurements being +/-0.321-5.024 within each variety for wholemeal and +/-0.451-3.57 for flour. The correlation between the kernel hardness values obtained with the two instruments was very close (0.96-0.99) and significant at the P=0.1 level. For breeding and genetic studies, the use of the more accurate SKCS 4100 is preferable, as the reproducibility is also better. The Inframatic 8611, or other instruments working on the same principle, can be recommended for the commercial classification of wheat lots. The analysis should be carried out on wholemeal, since sample preparation is easier in this case.