When feeds rich in starch are used as supplements to straw, both the digestibilitiy of fibre and straw intake may be reduced. The objectives of the study were to determine the effects of starch per se on straw intake and digestion, and to compare starch with pure cellulose of varying degradabilities. Four complete diets based on barley straw, molasses, fishmeal and urea, and containing starch 0(control), 100, 200, 300 g.kg^-1 were either fermentated in vitro or given to four cannulated sheep. The Control and 300 g starch kg^-1 diet were compared with Courcel (highly degradable cellulose) and Solkabloc (low degradable cewllulose), when the purse celluloses replaced straw of the basal diet at 300 g.kg^-1 and the above measurements were made including rumen dilution rates. Supplementation of straw with starch in different levels as complete diets linearly increased cumulative gas production in vitro and intake and digestion in vivo of the diets with little or no depression of straw digestion in sacco or fibre digestion in vivo. But the feeding of 300 g starch by the rumen cannula in a single meal in the morning reduced straw digestion in sacco. However, it was compensated by the restoration or increase of cellulolytic activity in the afternoon and no effect of feeding systems on straw intake and fibre digestion in vivo was found. Starch did not reduce and may have increased straw intake (not significantly) up to 300 g.kg^-1. Both starch and Courcel (at 300 g.kg^-1) increased the rumen outflow rates of solids and liquid. Courcel as a highly degradable cell wall polysaccharide also increased straw intake without showing any positive effect on straw digestion in sacco but degradation was still higher than that with starch. Solkafloc, possibly owing to its physical form and different feeding system, gave an atypical result on the rumen environment and failed to show any positive effect on straw intake and digestion in sacco. The cumulative gas production from the fermentation in vitro of the straw diets containing pure carbohydrates was highly correlated with its intake and digestion in vivo and may be a good predictor of intake and digestion of a mixed diet.