This study was conducted to determine if deep banding of P fertilizer was as efficient as placing P in a band with the seed. Phosphorus fertilizer (MAP) was deep banded at rates of 0, 10, 20, and 30 kg of P ha-1 which was compared with equal rates applied with the seed, and with one-half seed placed, one-half deep banded. The P rates were applied onto the same plots each year in a cropping sequence of canola (Brassica napus L.), four crops of wheat (Triticum aestivum L.) and then canola. Plots were located on a Melfort silty clay soil (Orthic Black Chernozem). In combination with MAP treatments, ammonium nitrate was applied by deep banding and by broadcast-incorporation to bring the total N application rate to 75 kg ha-1. Control treatments of 0-0 and 11-20 kg of N-P, respectively, were also included. In the first year of the experiment, when soil moisture conditions were dry, seed placement of the P fertilizer, on average, resulted in a significantly higher grain yield (1.01 t ha-1) of canola than deep banding (0.88 t ha-1). In the last year of the experiment, canola grain yields for the two placements were not significantly different (1.87 vs. 1.83 t ha-1). Over the 4 yr that wheat was grown there was no significant difference in grain yield between seed placed P and deep banding (2.97 vs. 2.95 t ha-1). Seed placement of P resulted in a higher grain yield (3.05 t ha-1) with N broadcast than with N deep banded (2.90 t ha-1), but with half-seed plus half-deep-banded P the yield was higher with N deep banded (3.08 t ha-1) than with broadcast-incorporated (2.87 t ha-1). The application of N and P over the 6 yr increased the available P in the soil from 7.5 (control) to 12.9-mu-g of P g-1 soil (N P treatments with 75 kg N ha-1). In general the application of P fertilizer either by deep banding or placement with the seed of the crop gave similar yields with the exception that at the low rates of P, seed placement was better for canola in the first year of the crop sequence.