To characterize the seasonal variability of endotoxin levels, we measured endotoxin in dust from the bed, bedroom floor, and kitchen floor in 20 homes, and in air from the bedroom in 15 of the homes. All homes were located in the greater Boston, Massachusetts, area and were sampled each month from April 1995 to June 1996. Outdoor air was collected at two locations. We found greater within-home than between-home variance for bedroom floor, kitchen floor, and airborne endotoxin. However, the reverse was true for bed dust endotoxin. Thus, studies using single measurements of dust endotoxin are most likely to reliably distinguish between homes if bed dust is sampled. Dust endotoxin levels were not significantly associated with airborne endotoxin. Airborne endotoxin was significantly (p = 0. 04) and positively associated with absolute humidity in a mixed-effect model adjusting for a random home effect and fixed effect of sampling month and home characteristics. This finding implies that indoor humidity may be an important factor controlling endotoxin exposure. We found a significant (p < 0.05) seasonal effect in kitchen floor dust (spring > fall) and bedroom airborne endotoxin (spring > winter), but not in the other indoor samples. We found significant seasonal pattern in outdoor airborne endotoxin (summer > winter).