In temperate forests, soils are the main sink for atmospheric N deposition. The main processes proposed for N retention are microbial and abiotic immobilization in soil organic matter. The relative importance of these processes as well as the kind of resulting chemical compounds are not totally understood. We carried out a laboratory incubation of Hg-sterilized and non-sterilized organic and organo-mineral soil horizons, labelled with either 15 NO 3 À or 15 NH 4 þ . The labelled samples were incubated for 1 hour, 1 day, or 6 days, then subjected to K 2 SO 4 extraction and analysed with 15 N CPMAS NMR spectroscopy. N immobilization was already effective in all samples and treatments after 1 hour. The corresponding NMR spectra showed that part of the immobilized 15 N was already incorporated into an amide structure. In the sterilized soils labelled with 15 NH 4 þ , the tracer was rapidly and largely immobilized by an unknown process related to the presence of Hg. In the sterilized soils labelled with 15 NO 3 À , between one-third and one-half of the added tracer was immobilized during the first hour and only 10% more over the 6 days. These results suggest that the sterilization was incomplete at first, allowing relatively great microbial immobilization during the first hour. By contrast, over a longer time, NO 3 À immobilization was significantly reduced to a level corresponding to an abiotic process as Hg sterilization became more effective. Even if the low signal-to-noise ratio precluded quantitative 15 N NMR measurements, we showed that the amide-peptide signal, considered as a biotic signature, was dominant in all cases. RésuméDans les foreˆts tempe´re´es, les sols constituent le puits principal vis-a`-vis des apports atmosphe´riques azote´s. Deux processus sont a`l'origine de la re´tention de l'azote dans le sol, l'immobilisation microbienne et l'immobilisation abiotique sur la matie`re organique. L'importance relative de ces deux processus ainsi que les formes chimiques qui en re´sultent ne sont pas e´lucide´es. Nous avons mene´au laboratoire une expe´rience d'incubation qui porte sur des horizons organiques, organo-mine´raux, ste´rilise´s ou non a`Hg et marque´s ou non a`1 5 NO 3 À ou 15 NH 4 þ . Suite aù ne incubation d'une heure, d'un jour et de 6 jours, des extractions chimiques a`K 2 SO 4 et de la N CPMAS spectroscopie de Re´sonance Magne´tique Nucle´aire ont e´te´re´alise´es sur les e´chantillons marque´s 15 NO 3 À , 15 NH 4 þ . Apre`s une heure, on note une immobilisation de l'azote dans tous les e´chantillons de sol, quelque soit le traitement effectue´. Les spectres RMN correspondant indiquent qu'une partie du 15 N immobilise´est sous forme d'amides. Dans les e´chantillons ste´rilise´s marque´s a`1 5 NH 4 þ , le traceur est rapidement et massivement immobilise´par un processus non identifie´et lie´a`la pre´sence de Hg. Dans les sols ste´rilise´s marque´s a`1 5 NO 3 À , entre un tiers et la moitie´de la quantite´du traceur sont immobilise´s au cours de la premie`re heure. Au cours des 6 jours suivants...

In temperate forest ecosystems, soil acts as a major sink for atmospheric N deposition. A (15)N labeling experiment in a hardwood forest on calcareous fluvisol was performed to study the processes involved. Low amounts of ammonium ((15)NH(4)(+)) or nitrate ((15)NO(3)(-)) were added to small plots. Soil samples were taken after periods ranging from 1 h to 1 yr. After 1 d, the litter layer retained approximately 28% of the (15)NH(4)(+) tracer and 19% of (15)NO(3)(-). The major fraction of deposited N went through the litter layer to reach the soil within the first hours following the tracer application. During the first day, a decrease in extractable (15)N in the soil was observed ((15)NH(4)(+): 50 to 5%; (15)NO(3)(-): 60 to 12%). During the same time, the amount of microbial (15)N remained almost constant and the (15)N immobilized in the soil (i.e., total (15)N recovered in the bulk soil minus extractable (15)N minus microbial (15)N) also decreased. Such results can therefore be understood as a net loss of (15)N from the soil. Such N loss is probably explained by NO(3)(-) leaching, which is enhanced by the well-developed soil structure. We presume that the N immobilization mainly occurs as an incorporation of deposited N into the soil organic matter. One year after the (15)N addition, recovery rates were similar and approximately three-quarters of the deposited N was recovered in the soil. We conclude that the processes relevant for the fate of atmospherically deposited N take place rapidly and that N recycling within the microbes-plants-soil organic matter (SOM) system prevents further losses in the long term.

Nitrogen (N) from atmospheric deposition has been shown to be mainly retained in the organic soil layers of temperate forest ecosystems, but the mechanisms and the physico‐chemical fractions involved are still poorly defined. We performed a hot‐acid hydrolysis on 15N‐labelled soil samples collected 1 week, 3 months and 1 year following a single in situ application of either 15NO3− or 15NH4+ in two montane forest ecosystems in Switzerland: Grandvillard (beech forest on a calcareous, well‐drained soil, 650 m above sea level) and Alptal (spruce forest on hydromorphic soil, 1200 m above sea level). After 15NH4+ application, recovery rates in the soil were smaller in Alptal than in Grandvillard through a large rate of absorption by mosses. At both sites, the organic soil layers retained most of the tracers at all three sampling times between 1 week and 1 year. In Grandvillard, the hydrolysable fraction (hydrolysable N : total N) of 15N was on average 79% and thus similar to the hydrolysable fraction of native N. This similarity is probably because of the rapid incorporation of N into organic molecules, followed by stabilization of the recalcitrant N pool through organo‐mineral bonds with soil minerals. In Alptal, the 15N hydrolysable fraction was greater than that of native N, particularly after 15NH4+ application (15N, 84%; native N, 72%). At both sites, 15N and the fraction of hydrolysable native N remained constant between 1 week and 1 year. This shows that both the recalcitrant and the hydrolysable pools are stable in the mid‐ to long‐term. We present arguments indicating that biological recycling through microbes and plants contributes to the stability of the hydrolysable N fraction.