PERIODICITY IN HUMAN BEINGS AND MICE 837 



whether or not the dominant synchronizer varied in prominence 

 and/or kind among the subjects as well, the same investigation by 

 Lewis and Lobban brought to the fore a most important point, namely 

 a dissociation among different daily rhythms on a given unusual 

 routine. Thus the temperature rhythm adapted almost immediately 

 to the experimental routines in most subjects, while for excretory 

 rhythms the reverse held true. Moreover, rhythms in the excretion of 

 water, chloride, and potassium also revealed dissociations, that in 

 potassium excretion being usually out of phase with those in water 

 and chloride. In human subjects living on different routines, a similar 

 dissociation also has been recorded for the rhythms in serum iron and 

 blood eosinophils (Halberg and Howard, 1958). It seems likely, 

 particularly under unusual circumstances, that at the same time and 

 in the same individual, different synchronizers can gain dominance 

 over the phase control of different rhythms. 



But is the nervous system involved in more than phase control of 

 rhythms? Can it account for the free-running circadian periods as 

 well, as we record them presumably in the absence of environmental 

 synchronizers? We must accept the circadian period of a rhythm as 

 endogenous rather than exogenous, if we find that it happens to be 

 sHghtly yet significantly different from 24 hours, if it is maintained 

 as such for prolonged periods and if we can assume that no environ- 

 mental factor with similar frequency existed. As long as circadian 

 periods in the mammal describe only body activity (Figs. 9 and 10), 

 we might think of a neural oscillator possessing the corresponding 

 natural period. The detection of circadian periods in rectal tempera- 

 ture behavior may lead us to believe that such an oscillator may be 

 located at the site of temperature control. But in the mammal we 

 now have indications for the operation of circadian periods in blood 

 eosinophils (Fig. 8) and epidermal mitoses as well, which may serve 

 to suggest that we are dealing with spontaneous cellular rhythms, 

 comparable to those encountered in unicellulars (Bruce and Pit- 

 tendrigh, 1957; Hastings and Sweeney, 1957a, b; Hastings et al., 1956; 

 Pittendrigh, 1954; Pittendrigh and Bruce, 1957). But obviously, 

 free-running periods in a given variable must not necessarily result 

 from an "oscillator" located specifically in the underlying function; if 

 this were true, the list of prospective cellular (or other) oscillators 



