Population improvement in lucerne (Medicago sativa L.): components of inbreeding depression are different in original and improved populations

Inbreeding depression, the lowered fitness of inbred individuals compared w ith their non-inbred counterparts, is an important concept in lucerne impro vement; but is poorly understood. Two-allele autotetraploid populations are suitable for studying inbreeding depression, especially when the populatio n improvement strategy involves inbreeding, because they are derived from c hromosome-doubling of hybrid diploid plants. They have a maximum of 2 allel es and a single allelic interaction per locus. Inbreeding depression was co mpared in original 2-allele autotetraploid populations and populations that had undergone inbreeding and selection. The original and improved (selecte d) populations were produced by intercrossing 2 single-cross lines from the original and improved 2-allele autotetraploid populations respectively. He rbage yield of the S-1 and intercrossed generations derived from these popu lations was evaluated in field trials at Arlington, Wisconsin, USA, and use d to estimate inbreeding depression. Herbage yield of the S1 and intercross ed generations derived from the improved population were significantly (P<0 .01) higher, by 13.3 and 24%, respectively, than those derived from the ori ginal population. Selection during inbreeding probably decreased the freque ncy of deleterious alleles and accumulated favourable alleles. Inbreeding d epression values were higher in the improved compared with the original pop ulation. Genetic load of deleterious alleles may account for much of the in breeding depression observed in the original population, whereas, in the im proved population,loss of heterozygosity or non-additive gene interactions: between favourable alleles on linked chromosome segments may account for th e substantial inbreeding depression. Therefore, in a population improvement program, the causes of inbreeding depression seem to be more important tha n their estimated value.