Several experimental approaches have been used in the past for the cannulation of the thoracic duct in mice. Most, however, are characterized by a modest reproducibility and a low lymph yield. Here, we describe a cannulation technique modified with respect to the anesthesia, the use of a silicone cannula and a simple and efficient intraabdominal fixation of the drain. Surgery averaged 45 min with an intra- and postoperative mortality rate of zero. Postoperatively, mice were given access to an exercise wheel allowing increased mobility and consequently a good lymph flow, thus maintaining the function of the cannula. The mice yielded a mean of 29.3 ml/24 h (range 8-40 ml) thoracic duct lymph, which contained a mean of 2.2 x 10(6) lymphocytes/ml during the first 24 h, decreasing to 0.1 x 10(6) lymphocytes/ml on the 2nd day after cannulation. Patency of the cannulae was 100% after 3 days. Interestingly, we have detected strain dependent differences in the anatomy of the thoracic duct in the mouse and these need to be considered when cannulation procedures are attempted.