Performance of S1 Alfalfa Lines from Original and Improved Populations

Abstract

Inbreeding has many uses in plant breeding. Improvement of autotetraploid alfalfa (Medicago sativa L.) by selection during inbreeding was evaluated by measuring inbreeding depression on herbage yield in 60 S1 lines from original and 120 S1 lines from improved populations. Original and improved populations were produced by synthesizing both before and after selection, respectively, in ‘Vernal’, ‘Saranac’, ‘Iroquois’, ‘Honeoye’, and an experimental containing Vernal and Saranac. Selection for self‐fertility and vigor was practiced during four generations of inbreeding. Natural and directional selection during inbreeding reduced 536 S1 lines to 152 in the S2, to 75 in S3, and to 16 in the S4 generation. The 16 selected S4 inbreds were used to make single crosses, and then two sets of double crosses (DCA and DCB) that were designated improved populations. The original materials were synthesized from open‐pollinated seed of 452 of the original 536 plants (W40) and by using remnant seed of 41 lines identified as superior during the inbreeding and selection process (W41). A higher proportion of plants in the improved populations produced sufficient SI seed for field testing (49%), than did plants in the original populations (27%). The S1 lines from the improved populations showed less inbreeding depression for herbage yield (19% below S0), while those from the originalshowed more inbreeding depression (28% below S0). Higher percentages S1 entries were included in the highest three herbage yield classes from improved populations (19.7%) than from original populations (12.4%). Thus, inbreeding depression was reduced in the improved populations. Selection during inbreeding to produce the 16 S4 parents mof the improved population likely decreased the frequency of deleterious alleles and accumulated favorable alleles.