The reproducibility of the radicallororbent (RAST) inhibition and direct binding assays with mite allergen were investigated in the presence of heterogeneous extracts and non-mite-sensitive atopic sera. Both contain components similar to potential contaminants which would occur in the assay of mite allergen and dust allergen extracts. The standardized inhibition and direct binding assays employed had a day to day (n= 4) coefficient of variation [(s.d. ± 100)/mean] of 15% and 24% respectively. The inhibition assay for mite allergen was reproducible in the presence of protein concentrations of added plant, fungal, arthropod and animal extracts in excess of the protein concentrations that occur under the operational mite assay conditions. The mite inhibition assay was also reproducible in the presence of non-mite allergen extracts, with and without additional sera containing IgE specific for the non-mite allergens. The binding of a constant quantity of mite allergen to the activated solid phase in the direct binding assay was reproducible in the presence of added bovine serum albumin, and of a fungal or arthropod extract, representing the heterogeneous components of an allergen extract at the concentrations of total protein known to occur in the direct binding assay of mite extracts.