Two-stage selection strategy in plant breeding

Abstract

The efficiency of two-stage selection depends upon the correlation of performance between stages. This paper gives a general treatment of two-stage selection on a frequency basis (selected/not selected), taking the costs of trials into account. A table that permits the calculation of the relevant 2×2 tables from knowledge of r and selection rates is presented. Given that the breeder is prepared to make certain initial subjective judgements/choices, optima, based on numbers of superior selections or least costs usually emerge. If the correlation is not well known it will usually be better to guess low than to guess high. Optimal solutions (which may or may not sharp) may imply the adjustment of total numbers of entries, either upwards or downwards. Examples from rubber and sugarcane breeding are discussed. © 1985 The Genetical Society of Great Britain.