The wick method for sampling of interstitial fluid from subcutis was applied in fluid balance studies in young pigs. Colloid osmotic pressure was measured in serum (COPs) and interstitial fluid (COPi) using a membrane colloid osmometer. Our aims were to determine the 'true' COPi, and to find the optimal duration of wick implantation. In series I (n = 6) a 'crossover' experiment was performed using wicks soaked in different priming solutions (non-diluted and diluted serum protein solutions or isotonic salt solution). Circulatory arrest was induced just before wick insertion in order to eliminate the vascular part of the acute inflammation. In series II (n = 6) wicks were removed in sequence after 60, 90, 120 and 180 min sampling time in anaesthetized pigs in vivo. COPs, COPi and haematocrit (HCT) together with haemoglobin (Hgb), serum albumin and total protein concentrations were determined in the same animals. In series I average COPs and COPi were 13.7 (1.4) and 7.2 (1.4) mmHg respectively (SD). In series II the optimal wick implantation times were estimated to be 60-90 min for wicks soaked in diluted protein solution, and 90-120 min for dry and saline-soaked wicks. COPs averaged 13.0 (0.7) mmHg, HCT 30.0 (1.6)%, Hgb 8.3 (0.9) g/dl, s-albumin 22.7 (0.6) g/l and s-protein 47.3 (2.3) g/l. Compared to commonly reported reference values, we found surprisingly low values for most of the measured variables. This may be related to the fact that we used immature pigs. An analysis of the validity of the wick method based on our own results and published reports is presented. We conclude that sampling of interstitial fluid with subcutaneous wicks is easy to perform in young pigs. However, the COP-values measured in wick fluid have to be carefully evaluated especially when sampling is performed in vivo.