Use of soil water and groundwater by open-grown Acer saccharum Marsh. (sugar maple) tree canopies and forests was estimated by measuring transpiration (E) rates using porometry, sap flow methods, and the Bowen ratio method. The Bowen ratio and sap flow methods showed the best agreement; porometer measurements scaled to whole canopies always underestimated E by 15-50%. Trees of different sizes showed very different rates of E. I hypothesized that these differences were due to the differential access of large and small trees to groundwater and soil water, respectively. Transpirational flux was partitioned between soil water and groundwater by tracing the water sources based on their hydrogen stable isotopic composition (deltaD). Soil water deltaD varied between -41 and -16 per thousand seasonally (May to September), whereas groundwater deltaD was -79 +/- 5 per thousand during the entire growing season. Daily transpiration rates of large (9-14 m tall) trees were significantly higher than those of small (3-5 m tall) trees (2.46-6.99 +/- 1.02-2.50 versus 0.69-1.80 +/- 0.39-0.67 mm day(-1)). Small trees also showed greater variation in E during the growing season than large trees. In addition, compared to the large trees, small trees demonstrated greater sensitivity to environmental factors that influence E, such as soil water deficits and increased evaporative demand. Over the entire growing season, large trees and forest stands composed of trees > 10 m tall transpired only groundwater. The high rates of water loss from large trees and older forests were likely a result of the influence of an enhanced "pool" of transpirational water in the upper soil layers caused by hydraulic lift (see Dawson 1993b). The hydraulically lifted water reservoir enabled large trees to use more potential transpirational water during daylight hours than small trees, leading to a greater total water flux. In contrast, small trees and forest stands composed of younger trees almost exclusively used soil water, except during two dry periods when their transpirational water was composed of between 7 and 17% groundwater. Thus groundwater discharge from sugar maple trees and forest stands of different sizes (ages) differs significantly, and large trees and older forest stands have a greater impact on the hydrologic balance of groundwater than small trees and younger forest stands. However, mixed stands (small and large trees) may have a greater overall impact on the regional hydrologic balance than old stands, because trees in mixed stands draw on both soil water and groundwater reservoirs and thus can substantially increase total water discharge on scales from tens to hundreds of hectares.