Soil food webs are mainly based on three primary Carbon (C) sources: root exudates, litter, and recalcitrant Soil Organic Matter (SOM). These C sources differ in their availability and accessibility to soil organisms, which could lead to different pathways in soil food webs. The presence of three C isotopes (12C, 13C and 14C) offers a unique opportunity to trace all three C sources in one study. In a microcosm experiment we investigated the effect of food web complexity on the utilization of these three carbon sources, which originated from i) rhizodeposition of living plant – corn (C4 signature), ii) litter of Lolium perenne (C3 signature, but labeled with 14C) and iii) SOM originated from C3 vegetation. Four levels of food web complexity were investigated. The simplest food web included autochthonous microorganisms only. The most complex food web consisted of autochthonous microorganisms, nematodes, collembolan, predatory mites, endogeic and anecic earthworms. Corn growing on soil developed under C3 vegetation and the application of 14C labeled ryegrass shoots as a litter layer allowed tracing of all three C sources in soil, CO2 efflux and different groups of organisms. The presence of living plants and litter had a major influence on C pathways, much more than food web complexity. Growing plants or litter added to the soil surface increased total CO2 efflux by nearly 3 times. Compared to the control with SOM as the sole C source, six and three times higher microbial biomass (fumigation – extraction) was measured in the living plant and litter treatments respectively. Anecic earthworms, Lumbricus terrestris, in the most complex food web enhanced CO2 efflux and plant growth. Living plants promote the growth of microbial biomass near the soil surface. Both the presence of litter and of anecic earthworms had positive effects on above ground shoot dry weight. Neither nematodes, nor microarthropods, nor endogeic earthworms had effects on CO2 efflux, on microbial biomass, or on shoot biomass. Without plants, 71% of the added 14C labelled litter were mineralized to CO2 after 30 days, whereas in the presence of living plants only 33% of the added litter were mineralized to CO2. Rhizodeposition promoted incorporation of 14C from litter into microbial biomass by six times (5% and 30% of 14C input without and with plants, respectively) and increased incorporation of litter C (14C) into SOM. The significant decrease of d13C value of microbial biomass, nematodes (for 1.8‰ after 30 days) and predatory mytes (for 1.6‰) in the presence of living corn plants showed that rhizodeposits as a C source were passed through rhizosphere microorganisms to the predator level (at least to the third trophic level). The changes of d13C value decreased with each next trophic level, showing slower turnover rates of predators compared to primary consumers. The changes of d13C value of SOM and of Lumbricus terrestris after 30 days were not significant. We conclude that besides anecic earthworms, living plants were the main modifiers of the food web and strongly affected the C flows 1) directly: by additional C source and 2) indirectly: by modification of the existing C flows within the food web including decomposition of the litter already present and CO2 efflux from the soil. The promising approach of coupling 13C natural abundance with 14C pulse labelling showed many advantages for simultaneous tracing of three C sources through the food web from primary producers up to predator levels.