Aims: The phosphorus (P) resource partitioning hypothesis assumes that dissimilarity in P acquisition traits among plant species leads to enhanced P uptake by crop combinations compared with their sole crops. We developed and implemented a test for this hypothesis. Methods: Two pot experiments were conducted with quartz sands. In Experiment 1, the ability of the crop species to acquire P from sparingly soluble sources (Ca phosphate (CaP), phytate (PhyP) and P-coated Fe (hydr)oxide (FeP)) was tested. In accordance with the species performances in Experiment 1, combinations of millet/chickpea and cabbage/faba bean (which have dissimilar P acquisition traits) and wheat/maize (which have similar traits) were selected for Experiment 2. The biomass production and P uptake were compared between the sole crops and species combinations as well as between the single and mixed P sources. Results: A dissimilarity in P acquisition traits enhanced P uptake by millet/chickpea on CaP/PhyP (as expected) but not by cabbage/faba bean on FeP/PhyP. Despite their similar P acquisition traits, we found enhanced P uptake by wheat/maize on CaP/PhyP. Conclusions: Because of complicating factors such as unstable P acquisition traits and competitive inequality between species, the conditions under which the P resource partitioning hypothesis can be tested are limited. This challenge complicates designing for complementarity in soil P pools by intercrops.