The suitability and reproducibility of the inhibition mode of the radioallergosorbent test (RAST) as a method for the assay of allergen extracts was examined in a collaborative study involving seven laboratories. Participants assayed five coded extracts of cocksfoot grass (Dactylis glomerata) pollen, including three batches from one manufacturer, both by a common method and by their local methods of RAST inhibition. For the versions of RAST inhibition employed in this study the estimates of relative potencies of the extracts were in agreement. Further, when laboratories used their local methods, the inter-laboratory variation in potency estimates was similar to that found when laboratories used the common method. It was concluded that RAST inhibition has a useful role in the assessment of the potencies of allergen extracts, and of the consistency of serial batches of extracts from individual manufacturers.