An international tritium model validation experiment was held at Chalk River, Canada, during June 1987. The Princeton Plasma Physics Laboratory (PPPL) Differential Atmospheric Tritium Sampler (DATS) was one of the many types of tritium samplers used for this experiment. Besides the modeling data that were produced from this experiment, we learned how well our tritium samplers performed when a known tritium quantity was released. The DATS were set up at 50, 100, 200, and 400 meters downwind from the release point. Data were collected during the release period and for the next 24 hours. While the units worked very well in the field, valued operational experience was gained in the recovery of the tritium from the silica gel. Because of delays in the analysis of the collected samples, it became difficult to recover the HTO fraction quantitatively. Indications are that molecular sieve is more suitable for samples which are not going to be processed immediately. This paper reports on the field set up, the measurement results, and operational experience in the use of the DATS.