Components of partial disease resistance (PDR) to fusarium head blight (FHB), detected in a seed-germination assay, were compared with whole-plant FHB resistance of 30 USA soft red winter wheat entries in the 2002 Uniform Southern FHB Nursery. Highly significant (P < 0·001) differences between cultivars in the in vitro seed-germination assay inoculated with Microdochium majus were correlated to FHB disease incidence (r = −0·41; P < 0·05), severity (r = −0·47; P < 0·01), FHB index (r = −0·46; P < 0·01), damaged kernels (r = −0·52; P < 0·01), grain deoxynivalenol (DON) concentration (r = −0·40; P < 0·05) and incidence/severity/kernel-damage index (ISK) (r = −0·45; P < 0·01) caused by Fusarium graminearum. Multiple linear regression analysis explained a greater percentage of variation in FHB resistance using the seed-germination assay and the previously reported detached-leaf assay PDR components as explanatory factors. Shorter incubation periods, longer latent periods, shorter lesion lengths in the detached-leaf assay and higher germination rates in the seed-germination assay were related to greater FHB resistance across all disease variables, collectively explaining 62% of variation for incidence, 49% for severity, 56% for F. graminearum-damaged kernels (FDK), 39% for DON and 59% for ISK index. Incubation period was most strongly related to disease incidence and the early stages of infection, while resistance detected in the seed germination assay and latent period were more strongly related to FHB disease severity. Resistance detected using the seed-germination assay was notable as it related to greater decline in the level of FDK and a smaller reduction in DON than would have been expected from the reduction in FHB disease assessed by visual symptoms.