1. 1.

    The circadian clock that controls CAP frequency was stopped at or near its lowest phase point by long duration cold pulses of 6 °C. On return to normal recording temperature (15 °C), the rhythm was always reinitiated from this phase point.

  2. 2.

    Following long cold pulses, there was often a transient peak of CAP activity lasting 2-6 h. It is thought that this was an effect of rise in temperature after prolonged cooling and not an effect on the clock itself.

  3. 3.

    Twelve h cold pulses, spanning the rhythm peak, caused phase delays. 9 °C pulses caused small delays (e.g. 1.7 h) while large phase delays (e.g. 6.7 h) followed pulses of 5 °C. Some pulses at an intermediate temperature (8.5 °C) caused abnormal post-pulse cycles lasting several days, and resulting in very large phase delays (10–14 n).

  4. 4.

    The abnormal CAP frequency curves following 12 h cold pulses of 8.5 °C spanning the rhythm peak are interpreted as rhythm splits. It is postulated that part of the population of coupled oscillators comprising the circadian clock was slightly delayed by the cold pulse, while the other part was driven further towards the “stopped” state, thus producing a large phase angle difference between the two subpopulations. These drew one another back into phase during several cycles to reform a normal circadian rhythm.

  5. 5.

    It is hypothesized that the circadian oscillations of the two subpopulations did not sum to produce the observed CAP frequency curve; rather the level of CAP output was controlled by whichever subpopulation was discharging at the higher frequency.


Laboratory of Sensory Sciences, University of Hawaii at Manoa, 1993 East-West Road, Honolulu, Hawaii 96822, U.S.A.

This content is only available via PDF.