I evaluated methods of preventing desert mule deer from drowning in the concrete-lined Mohawk Canal, southwest Arizona, and monitored the movements of deer using this canal. A 15 km study section of the Mohawk Canal where most previous drownings occurred was checked 478 times from June 1982 through September 1985 and 5,307 deer-canal interactions (DCI) were recorded. Ninety-eight percent of the DCI were recorded from April through September. Deer fell into this canal ≥ 279 times: 116 escaped via steps, 79 via ramps, and 50 escaped unaided. Only 5 deer drowned in sections of the Mohawk Canal with escape structures; 7 deer and 2 bighorn sheep drowned in sections without escape structures. Deer approached the canal to drink, not to cross. Maintaining depths to water of ≤ 30 cm will reduce the number of deer falling into the canal. The Mohawk Canal escape structures are adequately designed and spaced to prevent most summer mortalities. Deer use of 2 water catchments build to provide alternate water sources for deer drinking from the Mohawk Canal increased significantly each year. Each time a deer drank from these catchments was one less opportunity for a deer to fall into the canal. I monitored desert mule deer movements in a xeric region of the Sonoran Desert from October 1982 through November 1984 to determine the influence water availability and rainfall patterns have on deer movements. Ten of 15 radio-collared deer monitored for >1 year migrated to areas with permanent water in April or May and left those areas soon after summer rains started. Deer did not migrate to areas receiving the most summer rainfall. Home range sizes are larger (annual x = 145.2 km², range = 47.0 - 566.6 km²) than any previously reported for mule deer.