We used sugar maple litter double-labeled with 13C and 15N to quantify fluxes of carbon (C) and nitrogen (N) between litter and soil in a northern hardwood forest and the retention of litter C and N in soil. Two cohorts of litter were compared, one in which the label was preferentially incorporated into non-structural tissue and the other structural tissue. Loss of 13C from this litter generally followed dry mass and total C loss whereas loss of 15N (20–30% in 1 year) was accompanied by large increases of total N content of this decaying litter (26–32%). Enrichment of 13C and 15N was detected in soil down to 10–15 cm depth. After 6 months of decay (November–May) 36–43% of the 13C released from the litter was recovered in the soil, with no differences between the structural and non-structural labeled litter. By October the percentage recovery of litter 13C in soil was much lower (16%). The C released from litter and remaining in soil organic matter (SOM) after 1 year represented over 30 g C m−2 y−1 of SOM accumulation. Recovery of litter 15N in soil was much higher than for C (over 90%) and in May 15N was mostly in organic horizons whereas by October it was mostly in 0–10 cm mineral soil. A small proportion of this N was recovered as inorganic N (2–6%). Recovery of 15N in microbial biomass was higher in May (13–15%) than in October (about 5%). The C:N ratio of the SOM and microbial biomass derived from the labeled litter was much higher for the structural than the non-structural litter and for the forest floor than mineral SOM, illustrating the interactive role of substrates and microbial activity in regulating the C:N stoichiometry of forest SOM formation. These results for a forest ecosystem long exposed to chronically high atmospheric N deposition (ca. 10 kg N ha−1 y−1) suggest possible mechanisms of N retention in soil: increased organic N leaching from fresh litter and reduced fungal transport of N from soil to decaying litter may promote N stabilization in mineral SOM even at a relatively low C:N ratio.