Selection for DNA markers that are linked to loci that control forage yield, quality, and water-use efficiency offers the potential to improve alfalfa production efficiency. The importance of complementary gene action in tetraploid alfalfa and the inability of diploid alfalfa to predict performance of the tetraploid suggest that linkage analysis, and identification of quantitative trait loci (QTL), should be conducted at the tetraploid level. Our goal is to develop a tetraploid alfalfa population that will be useful for both linkage mapping, and for identifying DNA markers associated with QTLs which influence forage yield, quality, and reduced water consumption. The most efficient approach for developing this type of population is to cross genetically diverse parents whose hybrid progeny demonstrate significant heterosis for important traits. Hybrid populations that demonstrate heterosis for such traits should also demonstrate linkage disequilibrium for these traits in subsequent generations. This disequilibrium will be manifested as genetic variation which will be critical for identifying marker-QTL associations. To identify useful parents for linkage mapping, our first objective was to characterize DNA polymorphisms among various alfalfa germplasms that represent much of the genetic diversity in cultivated alfalfa in the U.S. Bulked DNA samples from 30 genotypes of each parental germplasm were scored on an ABI 377 Prism DNA sequencer for amplified restriction fragment length polymorphisms (AFLPs). These data are being used to estimate genetic distances among the germplasms using NTSYS. To identify useful parents for future QTL mapping research, our second objective is to characterize heterotic responses among the germplasms for: forage yield, carbon isotope discrimination, canopy temperature, forage maturity, and forage digestibility. Selected germplasms were crossed in a half-diallel mating design. The hybrid and parental populations have been evaluated in the field for the above traits. Preliminary data indicate significant (P<0.01) midparent heterosis for forage yield for several synthetic hybrids. Our third objective is to correlate marker-based genetic distances with field performance of the hybrid populations to determine if molecular genetic distance estimates can predict general and specific combining abilities and heterosis in alfalfa. Based on heterotic response data, and AFLP diversity among the parental germplasms, we will select two germplasms to initiate linkage analysis. A single noninbred genotype from each of the two selected germplasms will be crossed to generate F1 progeny. Our fourth objective is use segregating AFLPs in this F1 population to construct a tetraploid linkage map. To circumvent many of the complexities associated with polyploid linkage analysis we will monitor segregation of AFLPs which behave as single dose polymorphisms (SDPs). An SDP is defined as a fragment that is present in only one parent and segregates 1:1 (presence: absence) in the F1 progeny. Chi square analyses will distinguish between SDPs, which will segregate 1:1, and multiple dose markers which will segregate > 3:1 in the F1 progeny. Linkage relationships of SDP markers will be determined based on two-point and multipoint analysis using JoinMap version 2.0. Ultimately, we will anchor this tetraploid map to a diploid map previously developed by Montana State University and the University of Wisconsin. The AFLP markers used for tetraploid linkage analysis will also be useful in future research to identify QTLs in tetraploid alfalfa.