1. The receptive fields of 48 specific cold units, located in the hairy and glaborous skin of fore- and hindlimbs of rhesus monkeys, were mapped and scale drawings made. Forty-five percent had single spotlike receptive fields of 1-2 mm diameter or were up to 2 mm wide and 3-5 mm long, and formed straight, crescent, or right-angle shapes. The remaining 55% had two to five discrete spots or small areas innervated by a single fiber. Area or number of receptive fields per unit did not vary significantly along the limb axis. 2. Of the 48 specific cold units, 15 were held sufficiently long to record their dynamic responses to intensity-rate series of temperature changes from several adapting temperatures (ATs) between 20 and 45 degrees C. Six cold units showed a paradoxical increase in the steady-state response at the 45 degrees C compared to that at the 40 degrees C AT. Of the 15 cold units, 7 discharged in bursts at the 35 degrees C or lower ATs. The proportion of short intervals (intraburst intervals) increased as the adapting temperature decreased. 3. Both the latency and the temperature change at the onset of the dynamic response to cooling remained relatively constant for ATs up to and including 35 degrees C. The response latencies were approximately 200-400 ms, while the temperature changes at response onset were -0.02 to -0.06 degrees C. Both increased sharply at the 40 degrees C and still more at the 45 degrees C ATs. 4. The rate of increase in the frequency of the dynamic response to cooling by the fast rate increased to maximum at the 30 degrees C AT and then decreased at the higher ATs. For the slow rate the highest rate of increase in frequency occurred at the 20 degrees C AT. The fast rate of cooling always induced a faster rate of increase in frequency than slow rate of cooling. 5. Four indices of response magnitude were used in the analysis of the dynamic responses. These were peak frequency, cumulative impulses in the first 4 s following stimulus onset, average frequency during stimulations plus 3 s, and the total impulses during stimulation. The first three indices gave similar representations of the dynamic responses to the rate and intensity of cooling and the effect of the AT. The slow rate of cooling invariably yielded a smaller index of response magnitude than the fast rate. The difference became more pronounced as the intensity of cooling increased. The stimulus intensity-response magnitude functions were nonlinear. This nonlinearity was more pronounced at the low than at the high ATs. As the AT was increased, the response indices approached linearity at the 35 and 40 degrees C ATs. The greatest sensitivity of cold units to cooling was from the 35 degrees C AT. The fourth index, total impulses during stimulation minus steady state, gave a different picture. This index of response magnitude was linearly related to stimulus intensity for both rates of cooling from all adapting temperatures except for the slow rate of cooling from the 45 degrees C AT...