Four sources of Frankia endophyte were tested with four seed sources of Alnus spp. host plants in all possible combinations. Alnus glutinosa (L.) Gaertn. and Alnus rubra Bong. plants grew better with a spore (-) strain from The Netherlands than with a spore (+) strain from The Netherlands or with either of two strains from the United States;Two Frankia strains, CpN1 and ArI3, were compared as pure culture and crushed nodule derived inocula on two seed sources of A. glutinosa and two seed sources of A. rubra host plants. The A. glutinosa plants treated with pure culture inocula nodulated more quickly than those treated with crushed nodule inocula, bur the crushed nodule inocula induced greater total numbers of nodules. The A. rubra plants nodulated normally when treated with the pure culture inocula but failed to nodulate when treated with the crushed nodule inocula;Eight Frankia symbionts and two A. glutinosa seed sources were collected through cooperation with other researchers and a germplasm collecting expedition in Europe. The 16 host-endophyte combinations were tested for nodulation and symbiotic performance. Patterns of nodule development were observed through clear plastic Growth Pouches('(REGTM)). The spore (+) strains initiated more rapid nodulation and a larger total number of nodules. Well-nodulated seedlings from each host-endophyte combination were transplanted to Leonard jars to measure plant growth responses. Plant height was measured at approximately two-week intervals for three months. After five months of total growth, the acetylene reduction assay was used to estimate nitrogen fixation rates. There were significant differences among treatments in fixation per plant, but no significant differences among treatments in fixation per gram of nodules. Plants were then separated into roots, stems, leaves, and nodules, and dry weight measurements were recorded. A 13-fold difference in dry weight between the poorest and the best host-symbiont combination was observed. Significant intra-species differences were observed among A. glutinosa derived symbionts. In all cases plants treated with the spore (-) inocula grew larger than those treated with spore (+) inocula. Early nodulation or initial nodule number were poor predictors of subsequent plant growth. Interactions between host genotype and endophyte strain were small compared with the main effects of host and inoculum. This indicated that genetic improvement programs for either associate can probably be conducted without resorting to extremely large cross-inoculation experiments.