We collected interstitial fluid for measurement of colloid osmotic pressure (COPi) of normally hydrated and dehydrated rats with two methods: (1) saline-soaked nylon-wicks implanted in vivo for 1 h; and (2) the crossover method, i.e. titration with wicks pre-loaded with serial dilutions of rat plasma implanted post mortem for 15-20 min. All wicks were implanted in subcutis on the back. Dehydration was induced by injection of 40 mg furosemide followed by 24 h water deprivation. The saline-soaked wick method gave an average control COPi of 10.8 mmHg, 2-3 mmHg lower than the corresponding pressure obtained with the crossover method. This discrepancy between the methods increased to 10-11 mmHg in dehydrated animals, indicating that the saline-soaked wick method underestimates COPi even more during dehydration than during control conditions. The mean COPi of 22.8 mmHg (SD = 3.7, n = 9) obtained in dehydrated rats using the crossover method corresponds well with what is to be expected from the reduction in interstitial fluid volume. The present study suggests that the frequently used implantation period for saline-soaked wicks of 60 min should be prolonged to 90-120 min. It further shows that the rise in COPi is about 3 times more important than the fall in interstitial fluid hydrostatic pressure in restricting mobilization of interstitial fluid during dehydration.