The influence of various filter types and extraction conditions on the quantitation of airborne endotoxin with the Limulus amebocyte lysate test was studied by using airborne dusts sampled in a potato processing plant. Samples were collected with an apparatus designed to provide parallel samples. Data from the parallel-sampling experiment were statistically evaluated by using analysis of variance. In addition, the influence of storage conditions on the detectable endotoxin concentration was investigated by using commercially available lipopolysaccharides (LPS) and endotoxin-containing house dust extracts. The endotoxin extraction efficiency of 0.05% Tween 20 in pyrogen-free water was seven times higher than that of pyrogen-free water only. Two-times-greater amounts of endotoxin were extracted from glass fiber, Teflon, and polycarbonate filters than from cellulose ester filters. The temperature and shaking intensity during extraction were not related to the extraction efficiency. Repeated freeze (-20 degrees C)-and-thaw cycles with commercial LPS reconstituted in pyrogen-free water had a dramatic effect on the detectable endotoxin level. A 25% loss in endotoxin activity per freeze-thaw cycle was observed. Storage of LPS samples for a period of 1 year at 7 degrees C had no effect on the endotoxin level. House dust extracts showed a decrease of about 20% in the endotoxin level after they had been frozen and thawed for a second time. The use of different container materials (borosilicate glass, "soft" glass, and polypropylene) did not result in different endotoxin levels. This study indicates that the assessment of endotoxin exposure may differ considerably between groups when different sampling, extraction, and storage procedures are employed.