The variability and genetic structure of Alfalfa mosaic virus (AMV) in Spain was evaluated through the molecular characterization of 60 isolates collected from different hosts and different geographic areas. Analysis of nucleotide sequences in four coding regions--P1, P2, movement protein (MP), and coat protein (CP)--revealed a low genetic diversity and different restrictions to variation operating on each coding region. Phylogenetic analysis of Spanish isolates along with previously reported AMV sequences showed consistent clustering into types I and II for P1 and types I, IIA, and IIB for MP and CP regions. No clustering was observed for the P2 region. According to restriction fragment length polymorphism analysis, the Spanish AMV population consisted of seven haplotypes, including two haplotypes generated by reassortment and one involving recombination. The most frequent haplotypes (types for P1, MP, and CP regions, respectively) were I-I-I (37%), II-IIB-IIB (30%), and one of the reassortants, II-I-I (17%). Distribution of haplotypes was not uniform, indicating that AMV population was structured according to the geographic origin of isolates. Our results suggest that agroecological factors are involved in the maintenance of AMV genetic types, including the reassortant one, and in their geographic distribution.