INTRAPOPULATION RECURRENT SELECTION STRATEGIES IN PLANT BREEDING
The conduction of each intrapopulation recurrent selection (IRS) cycle typically requires a few years. There are several factors, with and without the control of plant breeders, that affect the success obtained with the IRS. For it to be more efficient, such factors must be judiciously assessed. In this light, this study aimed to compare a large number of variables involved in the efficiency of the IRS. Selection gain (SG) estimates with the IRS were obtained. Considering different variables under the control of the breeders such as the following: two types of progenies, half-sib (HS) and full-sib (FS); three experiment evaluation conduction strategies with multiple tree plots (MTP), single tree plots (STP), and clonal progeny test (CPT), and other variables. Variables without the control of plant breeders were also considered: mean allele frequencies of the populations (p); level of dominance (d/a); broad-sense heritability at the individual level (h²). Over 3000 SG estimates were obtained. When the CPT was used, the SG was always higher than that obtained with STP or MTP; the SG when employing HS was always higher than with FS. However, the employment of HS, especially when using clones for the recombination, has the serious restriction of not allowing higher selection intensities and maintaining the same number of progenies in the different IRS calculations. In this context, the employment of FS is more advisable since it allows applying a more significant selection intensity and enables more homogeneous recombination.