Summary Evidence is accumulating that belowground crop residues contribute more efficiently to the build-up and maintenance of native soil organic carbon (SOC) than aboveground plant parts. We corroborated previous testing of the hypothesis that root-C would preferentially accumulate in soil microaggregates where it is physically protected against microbially mediated decomposition. In three European field trials with C3 to C4 crop transitions, we compared the content of maize-C (Zea mays L.) in soil from rotations with grain maize (MG) or silage maize (MS), i.e. with incorporation of roots and shoots or roots only. After decades of maize cultivation, SOC content did not differ within three out of four MS?MG pairs, although obviously larger amounts of shoot biomass were added to soil in the case of MG. We found that relative contribution of roots was on average 3.5 times more than shoots to the build-up of SOC per equivalent mass of residue C added to soils. Preferential occlusion of root-C as silt-sized intra-microaggregate particulate organic matter (iPOM) was not observed. There were much larger effects from shoot incorporation on maize-C in the >53-?m fraction and free silt and clay. Storage of root-C as sand-sized iPOM was not quantified here, but first estimates suggested that physical entrapment at this level could only partly explain the longevity of root-C in soil. We reconfirm the relative stability of root-C in soil, but do not conclude that this stems from preferential physical entrapment over shoot-C. Future work should investigate the cause of preferential root-C association with the clay-sized fraction and if this occurs before or after microbial processing.