A selection of wheat samples from the 1988 New Zealand harvest were analysed for sprout damage by visual assessment of grain, a highly standardized bread baking test, the falling number test and an α-amylase assay. The relationships between these methods were analysed and reasons for anomalous results discussed. The bake test defines six sprout-damage categories (S′1 to S′6) inclusive of sound wheat (S′1). Flour falling number and α-amylase assay clearly distinguished sound samples from S′3 to S′6 samples that would cause problems in a commercial bakery. Meal falling number was less reliable at distinguishing sprout-damage categories. The slightly sprout-damaged S′2 samples were considered to require a bake test and falling number or ′-amylase assay to estimate potential problems to bakeries. The visual method was an unreliable indicator of sprout-damage as assessed by the bake test. There was a good correlation between α-amylase activity and flour falling numbers above 62. Over the whole range of sprout-damage samples, α-amylase activity was about 1·5 times higher in the milling fraction containing aleurone and scutellum tissue than in the corresponding flour fraction. The former milling fraction had increased α-amylase activity when stored for 8 months, whereas the flour fraction had reduced activity when stored.