In most terrestrial ecosystems ants (Formicidae) as eusocial insects and spiders (Araneida) as solitary trappers and hunters are key predators. To study the role of predation by these generalist predators in a dry grassland, we manipulated densities of ants and spiders (natural and low density) in a two-factorial field experiment using fenced plots. The experiment revealed strong intraguild interactions between ants and spiders. Higher densities of ants negatively affected the abundance and biomass of web-building spiders. The density of Linyphiidae was threefold higher in plots without ant colonies. The abundance of Formica cunicularia workers was significantly higher in spider-removal plots. Also, population size of springtails (Collembola) was negatively affected by the presence of wandering spiders. Ants reduced the density of Lepidoptera larvae. In contrast, the abundance of coccids (Ortheziidae) was positively correlated with densities of ants. To gain a better understanding of the position of spiders, ants and other dominant invertebrate groups in the studied food web and important trophic links, we used a stable isotope analysis (¹⁵N and ¹³C). Adult wandering spiders were more enriched in ¹⁵N relative to ¹⁴N than juveniles, indicating a shift to predatory prey groups. Juvenile wandering and web-building spiders showed δ¹⁵N ratios just one trophic level above those of Collembola, and they had similar δ¹³C values, indicating that Collembola are an important prey group for ground living spiders. The effects of spiders demonstrated in the field experiment support this result. We conclude that the food resource of spiders in our study system is largely based on the detrital food web and that their effects on herbivores are weak. The effects of ants are not clear-cut and include predation as well as mutualism with herbivores. Within this diverse predator guild, intraguild interactions are important structuring forces.