The objectives of the investigation were to identify the muscle spindle endings which respond to cooling of the relaxed muscle and to study their response to stretch. 2. The discharge of single afferents from 162 de-efferented muscle spindles in the relaxed medial gastrocnemius muscle of the anaesthetized cat was studied in vivo during cooling of the muscle from 37 to 24 degrees C. Temperature measurements were made at the inner surface of the muscle, while cooling (never below 15 degrees C) was applied at the skin over the muscle. 3. The endings were classified as primary or secondary endings on the basis of their conduction velocity, the dividing line being set at 70 m/sec. A response to cooling was obtained only from endings with afferents conducting at velocities of 20-70 m/sec. These fifty-six endings (CR) represented 65% of the secondary endings studied; the remaining secondary endings (NCR) and the primary endings showed no activity during cooling of the relaxed muscle. 4. During maintained stretches of 4-12 mm, activity of the NCR and primary endings decreased when the muscle was cooled. Cooling affected the CR endings in the same way, but only if the muscle was stretched 6 mm or more. During a smaller maintained muscle stretch, cooling caused an increase in CR activity, superimposed on the response to stretch. 5. The response to a 10 mm stretch at velocities of 10-70 mm/sec was studied in twenty-six CR, eleven NCR and twenty-one primary endings. 6. The dynamic responses of CR endings were intermediate between those of the primary endings and NCR endings. For any velocity of stretch the mean dynamic index of the CR endings was significantly greater than that of the NCR endings but significantly less than that of the primary endings. 7. The mean static responses of the CR and primary endings, measured 0-5 sec after the end of ramp stretch, were the same and significantly greater than that of the NCR endings. 8. The results indicate that cooling of the relaxed mammalian muscle may be used to differentiate between primary endings and about two-thirds of the secondary endings. The remaining secondary endings can be recognized by their small dynamic and static response to stretch.
