Abstract
AbstractResearchers often sample experimental units to measure certain characteristics, for example, percent protein in grain or percent sucrose in plant roots. Thus, in planning such experiments, one of the first questions encountered is how many samples and replications should be used.A procedure is suggested for using the non‐central F distribution to determine the number of replications and samples needed per experimental unit. The method requires the specification of the treatment effects to be detected, the probability of type I and type II errors, and an estimate of population variance. Sugar beet (Beta vulgaris L.) trials, employing sampling for sucrose content of roots, are used to illustrate the procedure in completely randomized, randomized complete blocks, latin square, and split‐plot designs. It is shown that fewer samples and/or replications are required when orthogonal comparisons can be defined.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
