Temperature compensation of the circadian oscillation in drosophila pseudoobscura and its entrainment by temperature cycles.

-Detailed analysis of the effects of temperature on the circadian rhythm of eclosion in Drosophila pseudoobscura adults yielded results similar to those in similar studies of other poikilotherms : although the period, T, of the free-running oscillation varies little with different constant temperatures, it can be phase shifted by changes (steps and pulses) and entrained by temperature cycles. Temperature steps-up generate only phase advances ( + A$) ; steps-down generate only phase delays ( -A+). Pulses, however, give both + A+ and A$. The size of A$ generated depends on the point in the oscillation’s cycle exposed to the temperature signal. An empirical relationship is found between temperature steps and pulses : to first approximation, the +A+ or A+ generated by 12 hr pulses may be computed from the A# generated by the steps comprising that pulse. This finding justifies the assumption that the A+ normally effected by the step is accomplished rapidly (within a few hours). An empirical response curve model for (square wave) temperature cycle entrainment of the eclosion rhythm is developed: from the A# response curves for single 12 hr pulses (high or low), the phase angle difference, Y, between 24 hr recurrent pulses (square-wave forcing oscillation) and the eclosion rhythm (driven circadian oscillation) is correctly predicted. From the response curves for a temperature step-up ( T 1/ T 2) and a temperature step-down (T2/T1) the A# response curves for single pulses (T1/T2/T1) of different duration were computed. These computed response curves differ in their zero phase shift points; that is, exposure of certain points in the oscillation’s cycle to a temperature step-up will generate a + A+ equal to the A+5 generated by the step-down which comes n hr later, n being the duration of the pulse. Assuming the validity of the empirical response curve model for temperature entrainment, the variation in zero phase shift points of the computed pulse response curves predicts a parallel variation of Y between the eclosion peak and 24 hr recurrent pulses of different durations (different thermoperiods). Such an experiment thus tests both the validity of deriving pulse response curves from of * The experiments reported here were performed under contracts from the Office of Naval Research (Nonr-1838(28)) and the National Aeronautics and Space Administration (Nas-223) to C. S. P. They were also supported by grants from the National Science Foundation to W. F. Z. (GB-5819) and T. P. (GK-1630).