Estimating Kernel Weight in Maize
In an experiment involving 20 maize (Zea mays L.) cultivars and five sample sizes (100,500,750 and 1000 kernels) replicated four times, it was found that sample weight increased linearly (r2 = 0.81) as sample size increased. This suggests that mean weight/kernel did not differ significantly among sample sizes. In similar trials involving 21 S1 lines developed from two maize cultivars (FARZ 23 and FARZ 27) but utilising sample sizes of 25, 50, 75, 100 and 125, mean weight/kernel showed a curvilinear response to increased sample size with the point of inflection at approximately the sample size of 100 kernels. This suggests that mean weight/kernel was variable for sample sizes less than 100 kernels. It seems, therefore, that kernel weight in maize can be estimated from small sample sizes (e.g. 100 and 125) kernels in order to save time and expenses when evaluating a large number of lines in a breeding programme.