Measurements of biocontaminants in settled house dust once a year are commonly used to assess long-term exposure. To examine stability over time and seasonal variation, we measured concentrations of mite and cat allergens, endotoxin and mold spores in living room floor dust in 745 German homes collected twice a year in two different seasons. The study population consisted of adults and children living in five different areas in Germany. All dust samples were collected in a standardized manner from the living room floor and taken during the years 1995 to 1998. The median interval between the two dust samplings was approximately 7 months. Mite and cat allergens were measured in settled house dust by monoclonal antibodies, endotoxin by the limulus amebocyte lysate method, and total spore counts by cultural methods. Crude Pearson's correlation coefficients between log-transformed concentrations in the first and second dust samples ranged between 0.65 and 0.75 for allergens, 0.59 for endotoxin and only 0.06 for total spore counts. The strongest and most consistent seasonal effects were observed for fungi with highest levels in July-September. Cat allergen concentrations were found consistently to be increased in January-March. Mite allergens did not show a strong and consistent seasonal pattern. We conclude that repeated measurements of mite and cat allergens and endotoxin in settled house dust improve the estimate for annual mean concentrations. However, even a single observation of these biocontaminants may be a good proxy for a 1-year exposure since repeated measures were highly correlated. However, repeated measurements of fungi levels were only weakly correlated and thus repeated observations for assessment of annual means of total spore counts are needed.