Comparison of Mass Selection and S1 Testing Recurrent-Selection Methods in Sorghum

Mass selection and S1 testing are often used for the improvement of sorghum [Sorghum bicolor (L.) Moench] populations. Both recurrent selection methods have a place in applied breeding programs and should be used where they are likely to be most advantageous. S1 families of a random-mating sorghum population mass selected three cycles (C3) for individual-panicle weight [IAP1R(M)C3] were compared with S1 families from a population (C3) advanced from the same base (CO) population by using replicated yield tests in each cycle [ IAP4R(S1)C3]. Our objectives were to compare the performance in C3 of families developed by using each selection strategy and to assess the genetic variability among S1 families. Sixty S1 families from each population were grown in five Iowa environments. Mean grain yields of S1 families derived from IAP1R(M)C3 and IAP4R(S1)C3 were not significantly different (beyond P\u3e0.05). Estimates of genetic variance among families (ail from IAP4R(S1)C3 were significantly greater for seeds/panicle and days to midbloom, but δ2g was not significantly different from IAP1R(M)C3 for grain yield, panicles/plant, and plant height. Greater δ2g for 100-seed weight was observed among the S1 families from IAP1R(M)C3. Except for 100-seed weight, heritability estimates for agronomic characters were not significantly different for the two populations. Considerations of mass selection and S1 testing should be made in relation to gains per season and use of project resources. On a preseason basis, mass selection seems preferable because S1 testing requires three seasons per cycle versus one for mass selection. If off-season nurseries are used for recombination of testing units, gain per year from S1 testing would be similar or superior to that from mass selection