This paper studies the urban heat island in Quebec City and its suburbs on one summer night. The region under study covers an area of 20 km by 20 km which streches from lake Saint-Charles on the north to Levis on the south and from the western suburb of Cap-Rouge to Beauport on the east. The total population of the region is about 400 000.On August 22, 1979 from 22 : 00 to 23 : 30 EDT, 109 measurements of wet-and dry-bulb temperatures were taken at preselected points. On that day, a large high pressure system gave clear skies to central Quebec while Quebec City Airport recorded a maximum temperature of 23 "C. The average wind measured at Duberger meteorological tower between 6 and 121 m above ground was 1.8 m SC' at 22: CO EDT and 0.9 m s-i at 23:30 EDT. Additional data were obtained from three thermographs installed at strategic points and five regular climatological stations, which were used to estimate cooling rates.During the period of measurement, the cooling rate averaged over the eight reference points Was about 0.7 'C h-'. However, the maximum cooling rate, which occured earlier, ranged from 2.2 'C h-i at Duberger to 4.4 'C h-' at Courville-de-Poissy.No relation seems to exist, in our case, between the maximum cooling rate and the heat island spatial structure. Comparison of these results with the ones given by Oke et al. (1972) shows that the eight stations behaved more like rural stations than urban ones.The temperature gradient measured between the center of old Quebec City and the suburbs averaged 6 -C while it was 9 'C for the coldest spot. Clo units were used to characterize the microclimates prevailing on the region for that night; 1 clo would be sufficient for comfort downtown while 1.5 clo is needed in the coldest spot for a slowly walking person. Large open areas like parks or the Lava1 University campus had a definite effect on temperature.The humidity was calculated for each point using standard psychrometric tables; no significant difference could be found over the area expect at points near or above the St. Laurent river where the average dew point was 1.5 "C higher.Each point was classilied into 6 categories according to its predominant land-use as reported by the observers. A multiple regression between temperature on the one hand and altitude and land-use on the other hand was tried; it showed that temperature was significantly correlated with both. Such an equation could be used by town-planners to locate temperature gradients in future urban development over the area; land uses could be planned to enhance or reduce these gradients.This study confirms the idea that urban climate is a mosaic of interacting micro-climates. More measurements, including possibly airborne infra-red thermal imagery, will bc taken during winter in order to assess the winter urban climate of Quebec City.