Intercropping, growing two or more crop species in the same field, is a practice that can contribute to ecological intensification of agriculture. Intercropping has been shown to increase productivity in low phosphorus (P) soils compared to sole crops. The ecological mechanisms underlying these benefits mainly include complementarity and facilitation with respect to resource acquisition. The contribution of these mechanisms to the yield benefit of intercropping probably depends on crop species traits, soil nutrient availability and agronomic practices. This thesis aims to contribute to the design of intercropping systems for improved P acquisition and yield gain by testing the resource partitioning hypothesis (dissimilarity in P acquisition traits among plant species leads to enhanced P uptake by mixtures of crop species compared to sole crops), and by quantifying the absolute yield gain of intercrops and the effect of agronomic practices. It presents results from empirical studies and meta-analyses.It appeared that the conditions under which the P resource partitioning hypothesis can be tested are limited. The tested crop species had inconsistent abilities to access the sparingly soluble Ca-bound P, phytate P, P-coated Fe(hydr)oxide, and competitive inequality between them largely determined the interaction. In a pot experiment with low P soils containing a mixture of organic and inorganic P sources, complementarity and facilitation did not result in increased P uptake by species mixtures, because P uptake was also affected by the competitive equality of species in the mixture. In a field experiment on a low P soil, complementarity and facilitation with respect to P uptake occurred in millet/chickpea relay strip intercropping, but they were not the main drivers for overyielding. In the subsequent meta-analysis on intercropping at the field level, the absolute yield gain of intercropping was mainly attributed to a positive complementarity effect, an outcome of any mechanism reducing competition. This positive complementarity effect was related to complementarity in using resources at different times caused by a relay sequence in the sowing and harvesting of intercropped species. The temporal complementarity in using resources plays an important role in competitive relaxation and hence overyielding. On a global level, there was a set of coordinated management factors rather than a single factor that drove the yield gain of intercropping, resulting in two contrasting syndromes of production in intercropping. The first syndrome was a high input -high output strategy that is maize-based with species arranged in relay strips and with high fertilizer input. The second syndrome was a low inputlow output strategy that is legume-based and arranged in simultaneous full mixtures or alternate row intercrops with low fertilizer input. Both intercropping strategies saved 16-29% land and 19-36% fertilizer compared to monocultures grown under the same management as the intercrop.To conclude, there was no evidence for the resourc...