Wood ants are known to affect the structure of invertebrate communities in
tree canopies but it is not clear whether wood ants can affect the structur
e of soil animal communities. To isolate the direct effects of wood ants on
soil fauna, I manipulated wood ant density in a 6-week mesocosm experiment
by placing mesocosms (O 14 cm, height 13.5 cm) permeable to ants either on
or outside the ant trails. The mesocosms contained litter, humus, and a di
verse soil fauna typical of coniferous forest. I controlled earthworm speci
es composition and biomasses by inoculating 470 mg fresh mass of epigeic lu
mbricid Dendrobaena octaedra (Sav.) or Dendrodrilus rubidus (Sav.) to the m
esocosms.
A 15-fold difference in ant density affected both the taxonomic and the tro
phic structure of the soil invertebrate community. The biomass of epigeic a
rthropod predators, Arachnida and Opiliones, was reduced in the high ant de
nsity treatment, whereas the biomass of mesofauna predators increased in th
e presence of ants. Although total earthworm biomass was not affected by th
e wood ants. reproductive output of epigeic earthworm Dendrodrilus rubidus
increased under high wood ant density, while reproduction of Dendrobaena oc
taedra was unchanged. The overall effects of wood ants on soil fauna was re
latively weak: despite the 15-fold difference in wood ant density between t
he low and the high density treatments, only some taxa were affected. Conse
quently, the direct effects of wood ants on soil animal food web is likely
to be small. (C) 1999 Elsevier Science Ltd. All rights resented.