The effect of surface cooling on intra-articular temperature was examined in dogs' knees. Four treatment protocols were examined: local ice compress application for 5, 15 and 30 min and local ice bath immersion for 15 min. Intraarticular temperatures were recorded using a needle microprobe inserted into the knee and continuous temperature recordings were made before, during and after the treatment. Rectal temperatures were also simultaneously recorded. Intra-articular temperatures rapidly dropped during icing. Five minutes of ice compress application resulted in a 2.2 ± 1.2°C intra-articular temperature drop with no change in rectal temperature. After 15 min of ice compress application, the joint temperature fell 4.1 ± 1.3°C with no change in rectal temperature. Thirty minutes of local ice compress application dropped the knee temperature 6.5 ± 4.0° C with a slight 0.5 ± 0.3° C drop in rectal temperature. Fifteen minutes of ice water immersion caused a much greater drop in intra-articular temperatures (20.2 ± 8.4° C) than could be achieved with ice compress. Rectal temperatures fell slightly during immersion (1.6 ± 0.3°C). After the removal of any type of cryotherapy, intra-articular temperatures continued to drop for several minutes and then a prolonged rewarming period commenced. The mean time required to return to baseline intra-articular temperature varied from 22-60 min, depending on the type and duration of cryotherapy. We conclude that brief periods of topical cold application to a dog's knee can induce significant and long lasting depression of intra-articular temperatures and that this is a local effect not dependent on core temperature cooling.