808 



PERIODIC FUNCTIONS IN MAMMALS 



1.4 



1.2 



a. 



I 1.0 



V 



•o 



•^ 0.8 



i s 



o 



a 



£ 



Q. 

 Q. 



0.6 



0.4 



0.2 



— • Shom-operation 

 --- ° Bilateral optic 

 enucleation 



Estimate of period (hours) 



Fig. 3. Slight but clear shortening of period (t) of temperature rhythm 

 during the first month after blinding revealed by periodograms on data 

 from individual mice. Note that such circadian periods vary from animal 

 to animal. 



temperatures, measured at 4-hr intervals (Ar = 4 hr), over a period T 

 of several days, served for the computation of the curves in Figs. 2-5. 

 The maximum at about 24 hr (but not always at exactly 24 hr) stands 

 out clearly, unmasked from the interference of shorter-term variations 

 in rectal temperature such as are associated with the spurts of activity 

 and feeding of rodents ( cf . Fig. 1 ) . 



Next, we shall examine the periodogram as applied to the study of 

 temperature rhythm in human beings. Rectal temperatures were meas- 

 ured at M= 1.5 hr for T = 97.5 hr, on several human subjects. 

 Periodograms computed for the group mean of rectal temperature are 

 shown in Fig. 6. All the data obtained on the group of subjects served 

 for computing the curve drawn as a continuous line; only one-half of 

 the data was used for the computation of the curve shown as a dotted 

 line, every second group mean of the time series being omitted. It is 

 readily apparent from Fig. 6 that the periodogram technique yielded 



