Forty days after mating, 16 pregnant sows were randomly allocated to two diets containing either sugar-beet pulp (SBP) or wheat straw (WS) each offered at two levels (1·0 or 1·5 × maintenance energy requirement (M), where M = 460 kJ digestible energy (DE) per kg0·75 per day). Diets were iso-nitrogenous, and contained similar levels of DE together with 260 g non-starch polysaccharides (NSP) per kg dry matter (DM) largely derived from either the SBP or the WS. A 7-day NSP and DM balance was made between day 60 and 100 of gestation. During lactation sows had ad libitum access to a conventional sow diet which contained neither SBP or WS.
Outputs of both fresh and dry faeces from sows given diet WS were significantly greater than the outputs from sows given the corresponding SBP diets (P <0·05), and outputs were significantly greater at the higher feeding level for both diets (P < 0·05). There was a tendency for higher urinary outputs from sows given diet WS, but these were variable and not significantly different from those from sows given diet SBP. The apparent digestibility and retention of DM was significantly greater for the SBP diet than for diet WS but were not significantly affected by feeding level. The apparent digestibilities of the total NSP and each of the individual constituent monomers were significantly higher for diet SBP than for diet WS (P < 0·001), but were independent of feeding level (P> 0·05). Xylose was the most poorly digested NSP constituent from both diets. Apparent digestibility coefficients for the remaining NSP constituents in diet SBP were > 0·8 and were > 0·52 in diet WS, with the uronic acids and mannose being the most highly digested NSP fractions from diets SBP and WS respectively.