The impact on a breeding programme of incorrect estimates of FD can be seen in Fig. 3where there was a large spread of FDmic data for high FDfin values. Regression analysis of the top 20% for FDfin versus FDmic gave an R2 of only 0.12 as a result of the poor relationship between the two FD calculations. If a breeder were to select the top 20% for FDfin, only 48% of the highest FDfin lines would be selected from the population if FDmic was used because of the influence of low maturity ratio on micronaire.