gms | German Medical Science

New crop varieties are usually evaluated for their performance in a target population of environments (TPE). This evaluation requires conducting randomized field trials at several environments sampled from the TPE. Such trials are called multi-environment trials (MET).

If the TPE is large and can be suitably stratified along geographical borders or agro-ecological zonations, it may be advantageous to subdivide the TPE into sub-regions. If the same set of genotypes is tested at a number of locations in each of the sub-regions, a linear mixed model may be fitted with random genotype-within-sub-region effects that allows estimating a genotype's average performance in each sub-region using best linear unbiased prediction.

The design of MET for a sub-divided TPE involves two decisions: (1) The total number of environments at which to conduct the trials and (2) the allocation of this total number of environments to the different sub-regions. The present work is devoted to the second decision: We propose an analytical approach for computation of optimal allocation numbers per sub-region with respect to the prediction of the genotype effects.

The authors declare that they have no competing interests.

The authors declare that an ethics committee vote is not required.