In the course of an investigation on the yield and composition of sows' milk, we have experienced difficulties in obtaining representative samples of milk for analysis.

The technique adopted was that described by Braude et al. (1) in which 5 units of pitocin were injected into the ear vein of the sow and the milk withdrawn by hand. Since the yield of these sows was being measured over a 24 hr. period each week, it was possible to compare the quantities obtained, following injection, with the average amount suckled by the litters during natural sucklings. The comparison was never particularly accurate as we found that two operators, each milking one teat, were all that could be conveniently accommodated on each side of the sow. This meant that only four glands could be milked properly. We found, too, that variations in teat size and mammary gland size affected the speed of withdrawal and the final amount obtained.