THE use of control and experimental groups of animals for the study of in vivo effects of drugs and foreign chemicals is a standard practice. Considerable care is given to several accepted precautions such as the establishment of identical conditions for both groups and the choice of litter mates of the same sex. If a drug is to be given parenterally to the experimental group each corresponding control animal must receive an injection of equal volume and ionic composition given by the same route. Less attention is usually paid to the segregation of the animals during the experiment. A common method is for several animals to be kept in a number of separate cages. The first animal taken from each cage may be used for control purposes and the remainder given different doses of the drug. Alternatively either one cage of animals may be designated the control group and other cages used for a particular drug dose or the control and experimental animals may be distributed randomly throughout the cages. In all these procedures it is implicit that the order in which the animals are used should not affect the biological system being measured. We have obtained some results which suggest that this assumption is not justified.