Soil water is the most limiting factor influencing crop production in the semiarid prairies. The effects of fallow frequency and nitrogen (N) fertilization on soil water conservation were quantified for a 40-yr (1967–2006) field experiment conducted on a medium textured Orthic Brown Chernozem (aridic haploboroll) in semiarid southwestern Saskatchewan, in which soil water contents were measured each year in early spring (generally a week prior to seeding), shortly after harvest, and again just prior to freeze-up in the fall. The three treatments examined were continuous spring wheat (Triticum aestivum L.) (Cont W) and fallow-wheat (F-W), each receiving recommended rates of N and P fertilizer, and Cont W receiving only P. On average, 36% of the precipitation received during the fall and winter months for Cont W (N + P) was conserved in the soil. In the summer fallow system [F-W (N + P)] a greater proportion (42%) of the precipitation was conserved during the first fall and winter. Despite the fact that cumulative precipitation from spring to late fall during the fallow period averaged 243 mm, compared with 152 mm received during the previous fall and winter period, the amount (31 mm) and proportion (13%) of precipitation conserved was considerably less than that during the first overwinter period. These differences were attributed mainly to much higher summer evaporative losses. During the second overwinter period, only 6% of the precipitation received was conserved in the F-W (N + P) system compared with 44% in the first overwinter period. This poor conservation during the second winter was thought to be related to increased snow blowoff due to smaller amounts of standing crop residues, and to the freezing of a wet and bare soil surface, restricting water entry during snowmelt or spring thaw events. Physical based soil-crop models, which must be first improved for overwinter simulations, should be tested on the current data set to further interpret the observations and hypotheses. Compared with the 36% of fall and winter precipitation conserved in Cont W (N + P), inadequate N fertility [Cont W (+ P)] resulted in only 27% of the precipitation being conserved during this period. At harvest, F-W (N + P) and Cont W (N + P) had similar amounts of water in the soil, but Cont W (+ P) had significantly (P < 0.05) more because of reduced water use. However, by the following spring soil water recharge being proportional to crop residues produced resulted in F-W (N + P), Cont W (N + P) and Cont W (+ P) having 252, 209 and 204 mm/1.2 m soil, respectively. Equations were developed that will allow estimation of water conserved as a function of precipitation received between harvest and seeding for F-W (N + P) and Cont W (N + P) (R2 = 0.52*** in each case). Trends in grain yield were fairly closely correlated with growing season precipitation and potential evapotranspiration [squared semipartial correlation coefficients for Cont W (N + P) 0.32 and 0.17, respectively, and for F-W (N + P) 0.35 and 0.12]; soil water conservation played a minor role in determining final grain yields (squared semipartial correlation coefficient < 0.08). Key words: Soil water, spring wheat, precipitation deficit, yield trends, water conserved trends