Adult sows fed a constant amount of a basal diet received purified cellulose either orally at levels of 0 and 475 g/animal . d (Experiment 1) or intracecally at levels of 0, 285, 570 and 855 g/animal . d (Experiment 2). Each experiment consisted of subsequent periods of faeces and urine collection with the animals re-allocated to the treatments each time. With that, a total of 36 observations on each parameter was achieved. The faecal samples were analyzed for the contents of organic matter, cell wall carbohydrates and various nitrogen fractions such as bacterial N and undigested dietary N. Furthermore, N balance, urinary allantoin excretion and plasma urea concentrations were determined. In a preliminary study, the effects of freeze-drying and of shaking of the faecal samples as suspensions with water (in order to release bacteria from fibre) on content and composition of faecal nitrogen had turned out to be reproducible. Cellulose significantly enhanced faecal nitrogen loss whereas N retention was not affected due to the counteraction of urinary N loss. Plasma urea concentration reflected the situation with urinary N. The proportion of undigested dietary N and of water-soluble protein in total faecal N was somewhat increased by cellulose at cost of the bacterial N proportion which accounted for about 72 % of total N on average. Urinary allantoin did not respond to the higher bacterial activity in the hindgut in the presence of supplementary cellulose. Cellulose significantly decreased the apparent N digestibility by on average about 3 percentage units per 100 g of supplementary cellulose. True N digestibility was also reduced by cellulose but did not go below 95 %. The supplementary cellulose was fermented in the hindgut at similar rates of on average about 60 % regardless of the route of administration. The almost 100 g of native cellulose incorporated in the basal diets were lignified by about 20 %, and that is why they were fermented at a rate of only about 30 %. The rate of fermentation was only slightly decreasing with increasing amounts of supplementary cellulose, and a daily quantity of 564 g (11 g/W0.75) cellulose was fermented on average if the highest level of cellulose was provided. This was within a range exclusively reported for easily-fermentable carbohydrates but was achieved in the case of cellulose only at a consistently higher level of supply. The true efficiency of bacterial protein synthesis was 5.2 g bacterial protein/100 g supplementary cellulose on average. The apparent efficiency was 60 % higher averaging 8.4 g bacterial protein/100 g further apparently fermented organic matter.