DAILY RHYTHMS 481 



Again there are difficulties here which stem from the fact that diverse 

 physical systems can be devised (or evolved) to behave in a formally 

 similar way. In the comparison of living clock systems for instance, we 

 can take neither temperature independence nor entrainability by light 

 as evidence of a common mechanism; selection must have demanded 

 both these features as functional prerequisites of clock systems and 

 will, of course, have been indifferent to the particular physical 

 mechanism by which the functional end is met. If we are to test the 

 proposition of a common mechanism and use only formal properties, 

 these must be of such a nature that selection can reasonably be dis- 

 missed as the agent responsible for their universal association with 

 cellular clock systems; in short, the properties must lack adaptive 

 value. 



In the following sections of this paper we turn to a group of physio- 

 logical effects that we have recently found to be widespread in living 

 clocks. They have the two merits of ( 1 ) demanding a more explicit 

 and restrictive form of the oscillator model than has previously been 

 necessary and (2) of being apparently devoid of adaptive meaning. 



PHASE SHIFTS WITH SINGLE SIGNALS 



Definitions and Qualifications 



It is necessary to establish provisional working definitions for the 

 terms period, phase, and transients used in discussing the oscillator 

 model. Figure 1 in an earlier paper (Pittendrigh and Bruce, 1957) 

 summarizes the hypothesis of a fundamental endogenous self-sustain- 

 ing oscillation (ESSO) underlying and controlling overt rhythms 

 (OR). The type of experimental conditions which produce overt 

 forced (or field) rhythms and overt persistent rhythms are summarized 

 in the figure. Examples of overt forced rhythms and overt persistent 

 rhythms are shown in Fig. 2 and 3, illustrating eclosion in Drosophila 

 and running-wheel activity of small mammals. 



We shall not attempt to do more than make operational definitions 

 for the terms period, phase, and transients relative to overt rhythms. 

 Ultimately we should like to relate these formal characteristics to 

 equivalent ones for the underlying oscillatory clock mechanism, but at 

 the present time these relationships must be inferential. We define the 



