USDA-ARS and Agronomy Department

Throckmorton Hall, Kansas State University, Manhattan, KS 66506

This study was undertaken to explore the agronomic and genomic stability of alfalfa (Medicago sativa L) plants over an extended period of time. About 100 ramets of each of four genotypes of 'Riley' alfalfa were grown in a greenhouse for two years, with the top growth removed on a six-week interval. In the fall of the second year, the ramets were transplanted to a field plot in a 4 X 4 Latin square design with 24 ramets of each genotype forming a block. Vegetative yield of individual ramets were recorded a total of six times over the next three years. With each ramet grown for one year considered a plant-year of growth, each genotype was represented in nearly 500 plant-years of growth.

The relative yield of each genotype, considered over the entire Latin square, was quite consistent. Genotype 1 was the top-yielding genotype in five of the six harvests, was significantly greater than the next genotype in four of the six harvests, and yielded significantly more than the other genotypes in the combined analysis of the six harvests. The coefficient of variation generally increased over the six harvests, ranging from 35.2 in the first harvest to 47.7 in the sixth.

Variation among ramets within genotypes generally was consistent across harvests, with the largest and smallest ramets differing by a factor of about 2; the largest factor observed was nearly 4. Statistical analysis indicated the variance within genotypes was 51% of the variance between genotypes. Only about 5% of the variance was due to genotypes, less than 1% was due to rows of the Latin square, 2% was due to columns and 40% was due to date of harvest.

At the genomic level, PCR was used to test for fragment length polymorphisms within introns of nuclear, and within chloroplast, genes. No variation has yet been detected in nuclear genes. However, one plant was found which had undergone a deletion in the intron of a chloroplast gene. Sequence analysis revealed that the deletion consisted of the near-perfect excision of one repeat unit of two tandemly-repeated, 146bp units, suggesting replication slippage. This result suggests that alfalfa chloroplast introns may undergo DNA loss once in about 500 years of plant growth.