A first group of three horses was given diet 1 (D1) allowing 1180 g per 100 kg body weight (BW) of a pelleted food rich in fibre (P1) and 556 g per 100 kg BW of straw during a 20-day period to allow for adaptation. A second group of four horses were given diet 2 (D2) allowing 1180 g per 100 kg BW of a pelleted food rich in cereals (P2) and 1000 g per 100 kg BW of meadow hay during the same period. Digesta was collected from the stomach, duodenum, jejunum, ileum, caecum, right ventral colon, left ventral colon, left dorsal colon, right dorsal colon, and small colon, and faeces were collected under general anaesthesia 2·5 h after the ingestion of the morning pelleted meal. The concentration of total anaerobic, cellulolytic and lactic acid-utilizing bacteria, lactobacilli and streptococci were determined in all these segments except for the duodenum, left ventral colon, right dorsal colon and small colon. D -/ Llactic acid, volatile fatty acids and pH were measured in all anatomic segments of the digestive tract (from stomach to small colon). The caecal concentration of total anaerobic bacteria was the lowest (7 ⋅ 9 5 10 7 colony-forming units (c. f. u. ) per ml), whereas that of the stomach was the highest (1·4 5 10 9 c. f. u. per ml) ( P < 0 ⋅ 001). Cellulolytic bacteria did not exceed 3·0 5 10 2 c. f. u. per ml in the ante-caecal segments whereas in the hindgut the average concentration was 5·3 × 10 5 c. f. u. per ml ( P < 0 ⋅ 001). Likewise, VFA concentrations were also greater in the large intestine (on average, 96·3 mmol/l v. 8·8 mmol/l in the ante-caecal segments) ( P < 0 ⋅ 001), confirming the limited extent of fibre degradation in these ante-caecal segments. Lactobacilli, streptococci and lactate-utilizing bacteria colonized all the digestive tract; the stomach and the small intestine tended to host the greatest numbers of these bacteria, which suggests a high interference of micro-organisms with the digestion of readily fermentable carbohydrates. Compared with the other ante-caecal segments, the stomach ecosystem seemed the most affected by the composition of the last pelleted meal ingested : the concentrations of lactobacilli and lactate-utilizing bacteria were higher ( P < 0 ⋅ 05) with P2. The lower concentration of D -/ L -lactate with P2 ( P < 0 ⋅ 05) was concomitant with a greater proportion of propionate ( P < 0 ⋅ 05), probably related to a greater fermentation of lactate. In the large intestine of horses given D2, cellulolytic bacteria tended to be lower, whereas VFA concentrations were higher ( P < 0 ⋅ 05). The lower [NDF/starch] ratio of D2 was probably less propitious for the proliferation of cellulolytic bacteria but was compensated by the higher cellulose intake brought by the hay.
