The idea ofprecooling and demand limiting is to precool buildings at night or in the morning during off-peak hours, storing cooling in the building thermal mass and thereby reducing cooling loads during the peak periods. Savings are achieved by reducing on-peak energy and demand charges. The potential for utilizing building thermal mass for load shifting and peak demand reduction has been demonstrated in a number of simulation, laboratory, and field studies. In a preliminary case study in a government office building in California in summer 2003, it was found that a simple demand-limiting strategy reduced the chiller power by 80%-100% (1-2.3 W/ft2, 11-25 W/m 2 ) from 2p.m. to 5p.m. without causing any thermal comfort complaints. This paper describes a follow-up study in 2004 in which tests were performed in two office buildings over a wider range of conditions. A Web-based comfort survey instrument was developed and used in the field tests to assess thermal sensation, comfort, and perceived productivity ratings in these two buildings. The results of the comfort survey indicate that occupant comfort was maintained in the precooling tests as long as the room temperatures were within the range of 70°F-76°F (21.1°C-25.6°C). Nighttime precooling was found to have varying effects on the magnitude ofthe peak the following day, with a number of factors affecting its effectiveness. It was found to be important to manage the afternoon load shedding by ramping the zone temperature setpoints rather than stepping them up. This is particularly important on hot days or in buildings with smaller time constants, where electrical power could rebound and exceed the peak demand under normal operation.