Serum calcium concentrations were determined in 164 cases of milk fever (mostly Jersey cows) attended in the course of practice in a cold, wet winter rainfall area where the cows calved at pasture. The mean calcium concentration of the cows receiving no calcium therapy subcutaneously before sampling (1.0 mmol/litre) was significantly (P < 0.01) less than the mean of 1.4 mmol/litre for cows receiving 300 ml of 33.3 or 40 per cent calcium borogluconate solution, which was significantly (P < 0.01) less than the mean of 2.1 mmol/litre for those receiving 600 ml subcutaneously. Neither the rectal temperature of the cows nor the average time interval between the injection of the two bottles and blood sampling had any effect on the serum calcium of the latter cows, whereas the rectal temperature but not the time interval had a significant (P < 0.01) effect on the serum calcium of the cows receiving 300 ml; the lower their temperature, the lower the calcium concentration. The mean (+/- sd) interval between treatment by the herdsman and veterinary attention for the cows receiving 300 ml (4.8 +/- 4.9 hours was significantly (P < 0.01) shorter than for the cows receiving 600 ml (12.0 +/- 9.4 hours). At the time of sampling, 48 per cent of the cows receiving 600 ml of calcium borogluconate solution had a serum calcium concentration below 2.0 mmol/litre.