The relationship between Stewart's disease and subsequent kernel infection was evaluated for 77 naturally infected maize genotypes. Seed parents were visually rated for the percentage of leaf area killed by E. stewartii and percentage of infected kernels were determined by ELISA tests of individual kernels or bulk samples. Maximum likelihood statistical equations were used to estimate percentages of infected kernels in samples tested by bulk ELISA. No infected kernels were found in 59 of 77 genotypes, including all genotypes with less than 25% diseased leaf tissue. Only three genotypes had more than 5% infected kernels and those came from plants with more than 50% diseased leaf tissue. Seed transmission rates were determined by assays of more than 77,000 plants over four years of greenhouse and field trials. Sixteen seed lots used in the grow-out tests were harvested from plants inoculated, by the pinprick method, with the rifampicin and nalidixic acid resistant strain, Rif-9A or the wild type strain, SS104. Four seed lots were obtained from naturally infected plants. Kernels infection percentages for each seed lot were determined by agar plating, for the Rif-9A seed lots, or by individual kernel ELISA for the SS104-infected and naturally infected seed lots. Seed transmission assays were made by stem printing, which consisted of pressing cut stem cross-sections onto agar media. Twenty-eight cases of seed transmission were detected among 42,000 plants evaluated from the inoculated seed lots. All 28 positives came from seed lots containing >35% infected kernels. One positive plant was detected among 35,000 plants evaluated from the naturally infected seed lots. Calculations of seed transmission rates were based solely on the percentage of plants estimated to have originated from infected kernels. The transmission rate from the inoculated seed lots was 0.14% with a 95% upper confidence limit of 0.2%. The transmission rate from the naturally infected seed lots was significantly lower, at 0.022% with a 95% upper confidence limit of 0.1%. The risk of seed transmission of E. stewartii is extremely low and may be essentially non-existent from resistant cultivars or from seed lots with a low incidence of infected kernels.