Wheat was grown in rotation with three different crops, namely wheat, chickpea and lentil, and with a fallow, in three consecutive seasons beginning in 1992 in NW Syria. Two rates of N fertilizer (0 and 30 kg N ha−1) were superimposed on these four rotations, giving eight treatments which were replicated three times in each season. 15N-labelled fertilizer was applied to microplots within the fertilized plots at sowing when unlabelled fertilizer was broadcast on the rest of the plot.

Yields of grain and dry matter were generally greatest when wheat followed a fallow, and least in the continuous wheat rotation; this was significant in 1993 and 1994. Applications of N fertilizer had no effect on productivity in 1992, but in the other two seasons grain yields were increased by 550 kg ha−1, on average. Depending on the season, between 8 and 26% of the 15N-labelled fertilizer was recovered in the shoot dry matter, while between 18 and 54% of the fertilizer remained as N in the soil at harvest, mostly in the 0–20 cm soil layer. More than half the fertilizer in the crop at harvest had been taken up by the end of March, although by March the plants were only c. 10% of their mass at harvest. Conversely, <35% of the soil-derived N in the crop at harvest had generally been taken up by March. This temporal difference in the pools of N utilized by the crop was attributed to the drying of the soil surface layers where most of the N fertilizer remained. Approximately 50% of the 15N-labelled fertilizer was unaccounted for in the crop and 0–40 cm soil layer at harvest.

The yield benefit of growing wheat in rotation with a fallow or either grain legume rather than continuously, ranged from nothing to the equivalent of a fertilizer application of at least 30 kg N ha−1 to continuous wheat, depending upon the season and the previous crop. Only in the season where the residual effect of the rotation on wheat yields was greatest (1993) did the preceding grain legume crop or fallow appear to contribute between 10 and 20 kg N ha−1 to the wheat crop. Otherwise they contributed no N at harvest. While a fallow may increase the availability of water to the succeeding wheat crop, the benefit from the preceding grain legumes may lie in their effect on the soil physical structure or on the increased availability of other limiting nutrients rather than additional N.