In a Quercetum petraeae-cerris forest in northeastern Hungary, we examined effects of litter input alterations on the quantity and quality soil carbon stocks and soil CO 2 emissions. Treatments at the SĂkfĆkĂșt DIRT (Detritus Input and Removal Treatments) experimental site include adding (by doubling) of either leaf litter (DL) or wood (DW) (including branches, twigs, bark), and removing all aboveground litter (NL), all root inputs by trenching (NR), or removing all litter inputs (NI). Within 4 years we saw a significant decrease in soil carbon (C) concentrations in the upper 15 cm for root exclusion plots. Decreases in C for the litter exclusion treatments appeared later, and were smaller than declines in root exclusion plots, highlighting the role of root detritus in the formation of soil organic matter in this forest. By year 8 of the experiment, surface soil C concentrations were lower than Control plots by 32% in NI, 23% in NR and 19% in NL. Increases in soil C in litter addition treatments were less than C losses from litter exclusion treatments, with surface C increasing by 12% in DL and 6% in DW. Detritus additions and removals had significant effects on soil microclimate, with decreases in seasonal variations in soil temperature (between summer and winter) in Double Litter plots but enhanced seasonal variation in detritus exclusion plots. Carbon dioxide (CO 2 ) emissions were most influenced by detritus input quantity and soil organic matter concentration when soils were warm and moist. Clearly changes in detritus inputs from altered forest productivity, as well as altered litter impacts on soil microclimate, must be included in models of soil carbon fluxes and pools with expected future changes in climate.