Theoretically, growth of stressed plants is maximised when all resources are equally limiting. The concept of co-limitation could be used to integrate key factors affected by crop rotation. This paper tested the hypothesis that the effect of crop rotation on the yield of wheat is partially mediated by changes in the degree of co-limitation between nitrogen and water.

Four rotations were established on a sodic, supracalcic, red chromosol in a Mediterranean-type environment of southern Australia. Rotations included wheat grown after (a) faba bean harvested for grain, (b) faba bean incorporated as green manure, (c) ryegrass pasture, or (d) medic pasture; barley was grown after wheat in all cases. The response of wheat to the rotations during 3 growing seasons was analysed in terms of nitrogen and water co-limitation, and the response of barley was taken as a measure of the persistence of rotation effects.

Daily scalars quantifying water and nitrogen stress effects on tissue expansion were calculated with a crop simulation model. These scalars were integrated in a series of seasonal indices to quantify the intensity of water (S_W ) and nitrogen stress (S_N ), the aggregated intensity of water and nitrogen stress (S_WN ), the degree of water and nitrogen co-limitation (C_WN ), and the integrated effect of stress and co-limitation (SC_WN 25 C_WN/S_WN ). The expectation is that grain yield should be inversely proportional to stress intensity and directly proportional to degree of co-limitation, thus proportional to SC_WN .

Combination of rotations and seasons generated a wide variation in the amount of water and inorganic nitrogen in the 1-m soil profile at the time of wheat sowing. Plant-available water ranged from 33 to 107 mm, and inorganic nitrogen from 47 to 253 kg N/ha. Larger amounts of nitrogen were found after green-manured faba bean, and smaller after grass pasture. There was a consistent effect of rotation on wheat yield and grain protein content, which persisted in subsequent barley crops. Measured grain yield of wheat crops ranged from 2.5 to 4.8 t/ha. It was unrelated to water or nitrogen stresses taken individually, inversely related to the aggregated stress index S_WN , and directly related to the C_WN index of co-limitation. The combination of stress and co-limitation in a single index SC_WN accounted for 65% of the variation in measured crop yield. This is a substantial improvement with respect to the stress effect quantified with S_WN , which accounted for 43% of yield variation. It is concluded that rotation effects mediated by changes in the relative availability of water and nitrogen can be partially accounted for by degree of resource co-limitation.