Abstract:The effects of various purified carbohydrates, given as supplements to a basal diet of grass silage, on ruminal fermentation and on the urinary excretion of purine derivatives (PD) were investigated using six sheep in a 6 x 6 Latin square experiment with period lengths of 14 days. The six experimental treatments were (i) the basal diet of 4 kg day-' of grass silage, supplemented with (ii) sucrose, (iii) lactose, (iv) xylose, (v) wheat starch, and (vi) fructose. All supplements were given at 200 g day-'. The basal diet and the supplements were given in two equal meals each day. Relative to the basal diet, xylose, starch and fructose all reduced (P < 0.05) the mean ruminal pH; for sucrose, the depression was non-significant (P > 0.05) whilst lactose produced virtually no depression of pH. All supplements reduced (at least P < 0.05) the ruminal concentrations of ammonia compared with the basal diet but, excepting xylose, all the sugar supplements produced a greater (P < 0.05) reduction than did the starch treatment; for xylose, this effect just failed to reach statistical significance (0.10 > P > 0.05). Starch did not alter the volatile fatty acid (VFA) pattern from the basal diet but the sugars produced very marked changes in the molar proportions of VFA. All sugars reduced the proportions of acetic acid (for xylose, 0.10 > P > 0.05, for the others, at least P < 0.05) but compensation was in terms of increased proportions either of propionic acid (xylose and fructose) or of butyric acid (sucrose and lactose). All four sugars increased (P < 0.05) the urinary output of PD over that seen with the basal treatment but the much smaller increase with the starch supplement was statistically non-significant (P > 0.05). The calculated supply of microbial nitrogen to the small intestine was 10.2, 14.8, 14.3, 13.1, 11.9 and 13.7 g day-' for the basal, sucrose, lactose, xylose, starch and fructose treatments respectively; the value for sucrose was significantly (P < 0.05) greater than that for starch but differences between sugars were statistically non-significant (P > 0.05). It is concluded that sugars, particularly sucrose, are clearly superior to starch as an energy source for the microbial fixation of nitrogen in the rumen.