A year-long investigation of air change rates in an occupied house was undertaken to establish the effects of temperature, wind velocity, use of exhaust fans, and window-opening behavior. Air change rates were calculated by periodically injecting a tracer gas (SF(6)) into the return air duct and measuring the concentration in 10 indoor locations sequentially every minute by a gas chromatograph equipped with an electron capture detector. Temperatures were also measured outdoors and in the 10 indoor locations. Relative humidity (RH) was measured outdoors and in five indoor locations every 5 min. Wind speed and direction in the horizontal plane were measured using a portable meteorological station mounted on the rooftop. Use of the thermostat-controlled attic fan was recorded automatically. Indoor temperatures increased from 21 degrees C in winter to 27 degrees C in summer. Indoor RH increased from 20% to 70% in the same time period. Windows were open only a few percent of the time in winter but more than half the time in summer. About 4600 hour-long average air change rates were calculated from the measured tracer gas decay rates. The mean (SD) rate was 0.65 (0.56) h(-1). Tracer gas decay rates in different rooms were very similar, ranging only from 0.62 to 0.67 h(-1), suggesting that conditions were well mixed throughout the year. The strongest influence on air change rates was opening windows, which could increase the rate to as much as 2 h(-1) for extended periods, and up to 3 h(-1) for short periods of a few hours. The use of the attic fan also increased air change rates by amounts up to 1 h(-1). Use of the furnace fan had no effect on air change rates. Although a clear effect of indoor-outdoor temperature difference could be discerned, its magnitude was relatively small, with a very large temperature difference of 30 degrees C (54 degrees F) accounting for an increase in the air change rate of about 0.6 h(-1). Wind speed and direction were found to have very little influence on air change rates at this house.