-The behaviour of nitrogenous fractions in the rumen of sheep fed grasses at different vegetation stages and preserved by different methods was examined in this study. The measurements were made on four sheep fed forages. In Experiment 1, we used fresh cocksfoot cut at three stages of vegetation in the first growth stage (heading, first flower and full flower), two silages (one made with formic acid and the other without a preservative) and a first flower hay. In Experiment 2, we used fresh perennial ryegrass cut in the first growth stage (end of heading) and in the second growth stage at 7 week regrowth. A silage made with formic acid, two wrapped big bales harvested at 42% and 58% dry matter and a hay cut at the first growth stage (end of heading) were compared. The effective degradability of crude protein (DegN) estimated using the in situ method was lower (P < 0.05) with the latter vegetation stage for cocksfoot: 0.693 for the heading, 0.667 for the first flower, 0.597 for the full flower. For cocksfoot, the DegN of the silages was higher (P < 0.05) for silage without additive (0.770) than for silage with formic acid (0.705) and higher than that of the fresh forage at first flower (0.667). The DegN of hay was markedly lower (0.537, P < 0.05) than that of the original fresh forage. The DegN of silage of perennial ryegrass (0.760) and wrapped big bales harvested at 42% DM (0.739) were higher (P < 0.05) than that of fresh forage harvested at the end of heading (0.705). The DegN of wrapped big bales of perennial ryegrass harvested at 58% DM (0.667) and hay (0.536) were lower than those of the other forages (P < 0.05). Whatever the forage studied, the concentration of total nitrogen (tN), ammonia nitrogen (NH 3 -N) and non-ammonia nitrogen (NAN) were high 1 or 2 h after feeding and diminished rapidly up to 7 h after feeding. Some of the solubilised nitrogen remained as proteins 1 to 2 h after feeding for fresh forage harvested at various growth stages, but no protein was found in the rumen fluid after sheep were fed the silage (with or without preservative), wrapped big bales or hay. The proportion of dietary NAN flow (relative to ingested nitrogen) escaping rumen degradation was 5.8% to 10.1% for perennial ryegrass and 10.9% to 15% for cocksfoot. Quel que soit le fourrage étudié, les teneurs en azote total, en azote ammoniacal, et en azote non ammoniacal dans le rumen sont maximales 1 ou 2 h après le repas et diminuent jusqu'à 7 h après le repas. Une partie des matières azotées solubilisées reste sous forme de protéines 1 ou 2 h après le repas pour les fourrages verts tandis qu'on ne retrouve pas de protéines solubilisées dans le rumen des moutons recevant des fourrages conservés (ensilages, balles rondes, foin). Le flux d'azote alimentaire non ammoniacal (exprimé par rapport à l'azote ingéré) qui échappe à la dégradation dans le rumen et transite avec le liquide ruminal représente de 5,8 % à 10,1 % pour le ray-grass anglais et de 10,9 % à 15 % pour le dactyle.dactyle / ray-grass anglais / mode de conse...