Abstract
The biases inherent in the use of the growth curves of the litter and of milking devices to estimate milk yield in the rat are discussed. A method is described in which the pups are separated from their dams for a period of hours and are then allowed to suck for about 60 minutes. The increase in litter weight during suckling is used as an estimate of milk secretion during the period of separation. Various components of this method were examined and it was shown that (i) the rate of milk secretion in the rat is constant for at least 16 h; (ii) the administration of oxytocin at suckling does not influence the yield at that suckling, but does increase the rate of milk removal; (iii) gland evacuation is completed within 50 to 60 minutes of suckling; (iv) the emptying of the mammary gland prior to the secretion interval increases the sensitivity of the estimate of milk yield; (v) the yield after a standard period of separation is a function of the level of hunger of the litter prior to suckling; (vi) cross-fostering and frequent randomized reallocation of pups between treatment groups eliminate biases inherent in earlier methods. The problems associated with the technique are discussed and a laboratory routine is outlined. The use of the technique in cross-over designs is advocated.