DAILY RHYTHMS 483 



period of an overt biological rhythm as the average time interval (in 

 hours) between successive events in the periodic sequence. No bio- 

 logical rhythms are strictly periodic in the mathematical sense, and in 

 practice one takes some arbitrary and well-defined statistic of the 

 rhythm as a measure of the periodically repeated events. For example, 

 in the case of Drosophila we take the median of the daily distribution 

 of eclosion. For hamsters, mice, and many other organisms the time of 

 onset of running-wheel activity is taken as the statistic which we use 

 to discuss the formal properties of the rhythm. It is evident that the 

 diverse assayed periodic events may occur at different clock hours in 

 an alternating cycle of dark and fight. For example, the median of the 

 Drosophila eclosion peak occurs soon after dawn (Fig. 2), whereas the 

 time of onset of running-wheel activity in a nocturnal animal may oc- 

 cur just after sunset (Fig. 3). Thus, the characterization of the phase 

 of an overt rhythm with respect to an external entraining periodicity 

 is arbitrary. For reasons given later on, we shall characterize the phase 

 of an overt rhythm as the time interval from "dawn" to the arbitrary 

 statistic defining the rhythm. For example, in a typical light-dark cycle 

 of 12 hr of light and 12 hr of dark, the phase of the Drosophila rhythm 

 would be about Wi hr (Fig. 2), whereas the phase of the Peromyscus 

 rhythm would be about 13 hr (Fig. 3). We extend the definition to the 

 overt rhythm of the organism in constant conditions (such as DD) in 

 the following way. When the rhythm is in a steady state, "dawn" 

 comes 11/2 hr before the median of the eclosion peak in Drosophila or 

 13 hr^ before the time of onset of running-wheel activity in Pero- 

 myscus. Thus, when we discuss the phase of the rhythms of different 

 organisms in constant conditions we have a common basis for com- 

 parison. In a steady-state condition, such as may happen in continuous 

 fight or darkness or in alternating cycles of light and dark, the actual 

 intervals between successive events may not deviate much from the 

 average interval. When the organism is subjected to nonrepeated dis- 

 turbances (such as a single shock of light) the intervals between suc- 

 cessive events may be systematically shortened or lengthened until a 



2 Strictly speaking this should be 13/24 times the natural period in constant 

 conditions. ^ 



