Lactating cows were chosen for high or low susceptibility to bloat from within herds bred for high (HS) and low susceptibility (LS) for about three generations. For 4 experimental weeks they were allocated to a latin-square change-over design consisting of four feed treatments. The treatments were one of ad libitum hay and three levels of a lucerne diet -15 kg or 30 kg fresh material per day or ad libitum. Eight cows per herd were used, two per herd being allotted each week to each feed treatment. The objective was to compare HS and LS cows for food intake, empty liveweight, bloat score, heart girth, rib girths before and after feeding, saliva proteins, and three blood components. HS cows were 41 kg (12.1070) lighter than LS cows and 7.5 cm (4.5070) smaller in heart girth. Pre-feeding rib girths were not significantly different. However, adjusting the pre-feeding rib girths for heart girth by linear regression, the rib girths were respectively 193.7 and 187.8 em, a difference of 3.1 % (P< 0.06). There was no bloat among cows on the hay diets. The HS cows could bloat on 15 kg of lucerne, so that high intakes were not required, whereas the LS cows consumed up to 30 kg of lucerne with little effect. Mean bloat scores on lucerne (0 -4 scale) averaged 1.03 for HS and 0.11 for LS cows. Differences in post-feeding 1 Deceased Received 22 July 1987; accepted 3 March 1988 rib girths were affected by differences in intake and in bloat; for a unit increase in rib girth in HS cows as a result of food intake there was an additional increase of 0.77 as a result of bloat. The ratio of rib girth increase to dry matter (DM) intake was significantly higher in the HS cows (5.64 cm/kg) than the LS cows (3.20 cm/kg). The correlations of bloat score with heart girth, pre-feeding rib girth, and post-feeding rib girth were -0.32,0.19, and 0.41, and the best linear combination of the first two girths had a correlation of 0.63 with bloat score. The HS cows produced significantly lower daily yields of milk and milkfat. Saliva volumes were obtained by front and rear sampling bits from cows after feeding. There was a significant difference between herds in the ratio of rear to front saliva volumes. Twelve saliva band concentrations were assayed and results from front saliva in HS cows were significantly higher than for LS cows on lucerne diets, and vice versa on hay. There were no significant differences between herds in sodium (Na) or potassium (K) concentration from rear saliva and packed cell volume, nor in protein and albumin concentrations in plasma. However, in one of the two sampling schedules the ratio of Na to K was significantly lower in HS than LS cows.