Measurements of formaldehyde were made during the summer of 1996 at a high-elevation site in Quebec as part of the North American Research Strategy on Tropospheric Ozone-Canada East (NARSTO-CE) measurement program. Gas phase mixing ratios were determined continuously by removing formaldehyde from the air in a glass coil scrubber, and producing a fluorescent dimer through the Hantzsch reaction. Average mixing ratios of formaldehyde were 1.3 and 0.8 ppbv for dry and wet periods, respectively. Highest values of HCHO were observed July 1--2 with a maximum mixing ratio of 4.6 ppbv. Fog water samples were also collected and analyzed for HCHO on five afternoon periods. Comparison of HCHO in the gas and aqueous phases shows reasonable agreement with Henry's law equilibrium. For dry periods July 1--12, relationships were examined between formaldehyde and other chemical species also measured at the site. Data were segregated based on the ratio of NOx to NOy and on the level of anthropogenic hydrocarbons present in the air mass. For the majority of the data, formaldehyde increased with both ozone and products of NOx oxidation (NOz) and was inversely related to the NOx/NOy ratio. During the high HCHO episode July 1--2, HCHO was correlated with neither ozone nor NOz illustrating the different chemistry at the site on these days. A chemical box model was used to examine sources of HCHO July 1--4. The model suggests that biogenic hydrocarbons contribute on average 53% of the locally produced formaldehyde, the remainder resulting from the oxidation of methane (19%), anthropogenic VOCs (16%), acetaldehyde (7%), and organic peroxides (3%). The model cannot account for the July 1--2 formaldehyde mixing ratios from the chemistry measured at the site. This implies that an additional HCHO source not included in the model was responsible for the high levels on those days. ¿ 2001 American Geophysical Union