Human exposures to ultrafine particles (UFP) are poorly characterized given the potential associated health risks. Residences are important sites of exposure. To characterize residential exposures to UFP in some circumstances and to investigate governing factors, seven single-family houses in California were studied during 2007-2009. During multiday periods, time-resolved particle number concentrations were monitored indoors and outdoors and information was acquired concerning occupancy, source-related activities, and building operation. On average, occupants were home for 70% of their time. The geometric mean time-average residential exposure concentration for 21 study subjects was 14,500 particles per cm(3) (GSD = 1.8; arithmetic mean ± standard deviation = 17,000 ± 10,300 particles per cm(3)). The average contribution to residential exposures from indoor episodic sources was 150% of the contribution from particles of outdoor origin. Unvented natural-gas pilot lights contributed up to 19% to exposure for the two households where present. Episodic indoor source activities, most notably cooking, caused the highest peak exposures and most of the variation in exposure among houses. Owing to the importance of indoor sources and variations in the infiltration factor, residential exposure to UFP cannot be characterized by ambient measurements alone. PRACTICAL IMPLICATIONS: Indoor and outdoor sources each contribute to residential ultrafine particle (UFP) concentrations and exposures. Under the conditions investigated, peak exposure concentrations indoors were associated with cooking, using candles, or the use of a furnace. Active particle removal systems can mitigate exposure by reducing the persistence of particles indoors. Eliminating the use of unvented gas pilot lights on cooking appliances could also be beneficial. The study results indicate that characterization of human exposure to UFP, an air pollutant of emerging public health concern, cannot be accomplished without a good understanding of conditions inside residences.