during spring and summer in ontario, moose are commonly observed at sites known as moose aquatic feeding areas. feeding on aquatic vegetation is thought to be an important source of sodium for moose at this time of year. the effects of different timber harvesting systems on the use of aquatic feeding areas by moose was studied in the great lakes - st. lawrence and boreal transition forests of central ontario. during june to september 2002,1 compared the use of aquatic feeding sites by moose among selection cutting in the algonquin park forest management unit (fmu), uniform shelterwood cutting in the french-sevem fmu, and clearcutting in the spanish fmu. at >50 sites within each harvesting system i studied the relationships between moose use and age of forest stands adjacent to aquatic feeding areas, proximity of timber harvest, and amount of shoreline affected. the locations of potential study sites in the three fmus were initially identified using gis data (cut history and reserve widths), moose aquatic feeding area survey data, and air photos. sites were assessed for moose use by recording the characteristics of trails, tracks, pellet-groups, and browsing. physiographic and vegetative attributes of the aquatic and terrestrial landscape were also measured. overall, moose use of aquatic feeding areas was greatest in areas harvested by selection cutting, followed by shelterwood cutting, and clear-cutting, respectively. the reserve width and time since last cut influenced the use of aquatic feeding areas by moose in all three silvicultural systems. within areas harvested by selection cutting, moose use was greatest adjacent to old cuts (>20 years) and large reserve widths (>120m). the shelterwood areas showed more moose use of sites adjacent to recent cuts (<5 years) with >120m reserves. the clear-cut areas showed more moose use adjacent to cuts >10 years of age with >120m reserves. the results of stepwise multiple regressions, indicated that habitat characteristics other than forest age and reserve width were also important for moose when selecting a site. the length of aquatic vegetation along the shore and midpoint basal area were important habitat variables within the selection cut system. endpoint basal area was the only habitat variable important for moose use within the shelterwood system and there was no multiple regression model predicted in the clear-cut system. subsequent correlation analyses indicated that the length of aquatic vegetation along the shore and reserve width were the only two variables related to moose use within all three silvicultural systems. moose demonstrated both random and non-random patterns of use within reserves in all three systems. random use was identified by an interconnection of moose trails within reserves that were not used repeatedly, indicating that sites were used less frequently. non-random use was identified by a trail system heavily used within the reserve, indicating that trails were used repeatedly. because aquatic plants are an important source of nutrients for moose in spring and summer, forest management practices must ensure proper protection of these sites. this study shows that the time since last cut and the type of silvicultural system being used must be considered when applying a reserve around aquatic feeding areas, because the quality of the habitat within the adjacent reserve is important for moose using these sites. although moose used aquatic feeding areas adjacent to narrow reserves (<60 m), the results of this study show that sites adjacent to 120-m reserves, as recommended in the timber management guidelines for the provision o f moose habitat, were used the most and have the greatest potential of meeting the life history requisites of moose in all three silvicultural systems.