(1)The single injection-indicator dilution method of Crone was adapted for use in the rat. A bolus containing a non-diffusible reference tracer and the test tracer was injected into the internal carotid artery. Simultaneously venous blood was sampled from the confluence of sinuses. It proved necessary to take special steps to avoid extracerebral contamination of the samples. Therefore the extracerebral branches of the carotid tree on both sides were ligated, and the venous sampling was done without suction. Results obtained in animals prepared surgically in different ways were compared. (2) 24Na+ was used to assess the degree of extracerebral contamination, the transcapillary loss of 24Na+ being a sensitive measure since the permeability of cerebral capillaries for sodium is about 4-5% that of muscle. (3) The transcapillary fractional loss, extraction E, was determined in this preparation with minimal extracerebral contamination. ENA+ was 2.9%, Eglucose was 12.5% at normal blood glucose levels whilst the net glucose extraction (calculated from arterio-venous difference) was about 10%. (4) It is concluded that extracerebral contamination is an important pitfall in indicator dilution studies on the brain. Use of 24Na+ as an indicator of extracerebral contamination is therefore recommendable.