Photochemistry and Photobiology, 1997, 66(5): 549-561 zyxwvut Invited Review Twilight Times: Light and the Circadian System Till Roenneberg*’ and Russell G. Foster2 ’Institute for Medical Psychology, Ludwig-Maximilian University, Munich, Germany and 2Departmentof Biology, Imperial College, London, UK Received 21 May 1997; accepted 31 July 1997 zyxwvutsrqp Whose twilights were more clear, than our mid-day. John Donne (1571-1631) zyxwvut Of the progress of the soul THE CIRCADIAN CLOCK zyxwvutsr The spatial and temporal features of the environment have provided the abiotic selection pressures that have shaped the evolution of life on earth. We are accustomed to accept that the spatial world offers specialized niches. However, we rarely consider that the temporal structure of our planet of- fers similar opportunities. Most organisms have evolved spe- cializations that allow them to exploit their environment in terms of both space and time, and this demands that indi- viduals have an endogenous representation of these environ- mental features. We know surprisingly little about the en- dogenous representation of space, but in recent years an un- derstanding of how organisms build a representation of their temporal world has begun to emerge. Organisms do not merely respond to their environment, they also have the capacity to adjust physiology and behav- ior in anticipation of changing environmental conditions. In a competitive world, “being prepared” offers a great selec- tive advantage. It takes considerable time to bring about the complex realignments of physiological systems that permit an optimal expression of different behavioral states (e.g. ac- tivity and rest or exploitation of temporally restricted re- sources). By fine-tuning physiology in advance of the chang- ing conditions, an organism can be ready to exploit the changed conditions to its best advantage. At the heart of the biological machinery that “creates” a day within us is a biological or Circadian clock. In mammals, for example, this resides within a small paired nucleus in the brain located above the crossing of the optic nerves, the suprachiasmatic nuclei or zyxwvutsrq SCN.? The neuronal activity of this nucleus continues to oscillate with a 24 h rhythmicity *To whom correspondence should be addresed at: Institute for Med- ical Psychology, Chronobiology Group, Ludwig-Maximilian Uni- versity, Goethestr. 31, 80336 Munich, Germany. Fax: 49 89 5996- 615; e-mail: zyxwvutsrqp till@imp.med.uni-meunchen.de jAhhreviations: CT, circadian time; PRC, phase response curve; RGC, retinal ganglion cells; RHT, retino-hypothalamic tract; SCN, suprachiasmatic nuclei. Q 1997 American Society for Photobiology 003 I -8655/97 $5.00+0.00 even if it is isolated from the rest of the brain (1). Long before the primary circadian clock of mammals was identi- fied in the SCN, the endogenous nature of circadian rhythms was recognized, first in plants (2-5) and then in many dif- ferent animal species (6). Identification has been based upon the critical observation that when organisms are kept in iso- lation, void of any temporal cues, their daily routine contin- ues unabated, although the endogenous daily period zy (7) may deviate from the external 24 h cycle (T), hence the term circadian (about 1 day). Such drifting rhythms are often termed “free-running’’ rhythms, and depending on the or- ganism and on the nature of the constant conditions (e.g. constant light or constant darkness), the endogenous period of the free-running rhythm can range from about 19 to 28 h. The circadian period not only depends on the quality of environmental conditions, as will be discussed below, but is also under tight genetic control as shown by classic as well as by molecular genetics (7). Single genes or gene complex- es have been isolated in the fruit fly Drosophila (8), the fungus Neurospora (9,10), in Chlamydomonas (11) and in the cyanobacterium Synechococcus (12), as well as in the hamster (13) and the mouse (14), that profoundly influence the length of the circadian period and are likely to be mo- lecular components of the clock itself. The current models for molecular pacemakers will be briefly described in a sep- arate section. Considerable research has shown that the mechanisms of the circadian pacemaker (the endogenous rhythm generator) are part of the biochemistry of single cells (15-19). This has been known for decades because the me- tabolism of single cell organisms, such as Euglena (20,21), Chlamydomonas (22), Acetabularia (23), Gonyaulax (24), Pyrocystis zyxwvu (25), Tetrahymena (26) and Paramecium (27) is also controlled by a circadian system (for an extensive re- view of circadian rhythms in unicellular organisms see Ed- munds (28)). In addition to the free-running nature of circadian rhythms, two other universal properties have been identified: (a) The period of circadian rhythms is not greatly affected by changes in environmental temperature. The rates of most biochemical reactions approximately double with a 10°C rise in temperature (Qlo 2 2). By contrast, circadian rhythms show temperature compensation so that period length changes very little over larger temperature fluctuations (29). As a result, those circadian rhythms studied have a Ql0 that 549