Tannins were added to experimental diets at levels of 1 and 2 g/kg DM (hydrolysable tannins; Castanea sativa wood extract) and saponins at 2 and 30 mg/kg DM (sarsaponin; Yucca schidigera extract). These levels were far below thresholds expected to be adverse in ruminants. Effects were measured in lambs by comparison with unsupplemented control diets calculated to be either deficient (10%) or adequate in protein. The diets consisted of hay, concentrate (1:1) and extra wheat starch with increasing body weight. Ruminal pH, VFA concentration, protozoa count and apparent digestibilities of organic matter and fibre did not differ among treatments. The low tannin dose significantly decreased bacteria count compared to the high saponin dose. Saponin supplementation and the high tannin dose showed some potential to reduce ruminal ammonia concentration. This was associated with weak trends towards lower urine N excretion (only tannins) and ammonia emission from manure. Methane release was increased by the low tannin dose compared to the unsupplemented control. Diet effects on heat production were not systematic. In conclusion, the extracts rich in tannins or saponins gave only slight indications for either increased body nitrogen retention or reduced nitrogen emission. However, effects might have been larger with more pronounced dietary protein deficit.