Forward and reverse genetic approaches, together with transcriptome-scale gene expression analyses support the working model of the Arabidopsis circadian clock as a pair of interlocking negative feedback loops. Two Myb-domain transcription factors, CIRCADIAN CLOCK ASSOCIATED 1 (CCA1) and LATE and ELONGATED HYPOCOTYL (LHY), play central roles in each loop. CCA1 and LHY target three PSEUDO RESPONSE REGULATOR (PRR) genes with opposite effects. In one loop, CCA1 and LHY repress expression of TIMING OF CAB EXPRESSION 1 (TOC1), also known as PRR1. TOC1/PRR1 is a positive regulator of CCA1 and LHY, completing the first negative feedback loop. In the second loop, CCA1 and LHY induce the expression of PRR7 and PRR9. PRR7 and PRR9 are involved in entrainment of the oscillator by both light and temperature, although the details of resetting are incompletely known. Mutational analysis indicates that the temperature sensing mechanism that provides input to the clock is distinct from that used to respond to cold. Although temperature steps or pulses provide potent entraining cues, it is well established that the period of the clock is more or less constant across a broad range of temperatures (temperature compensation), and this clock property will also be discussed. It has long been assumed that the circadian clock influences the fitness of an organism and we are testing this conjecture. PRR7 lies in an interval identified by QTL mapping as contributing to period length. Considerable genetic variation at the PRR7 locus exists among natural accessions. Molecular analysis of this sequence diversity indicates that this locus is evolving in a nonneutral fashion. Moreover, the geographical distribution of these alleles does not appear to be random. We are investigating whether PRR7 sequence variation is associated with the differences in circadian function observed among natural accessions. This work is supported by the NSF.