PERIODICITY IN HUMAN BEINGS AND MICE 



809 



a. 



6 



07 



0.6 



0.5 



-r 0.4 



a. 



e 



o 



o 



E 



o. 



Q- 

 < 



0.: 



0.2 



0.1 



Sham - operotion 



^ 



3rd month after bilateral 

 optic enucleation 





J^ 



J_ 



± 



22 



23 24 



T 



Estimate of period (hours) 



25 



Fig. 4. During the third month after operation, differences in the 

 circadian temperature periods of individual bhnded mice bring about a 

 decrease in ampUtude of the mean for the Winded group. 



results which describe the daily periodicity studied in terms of both its 

 period and its amplitude. Moreover, the periodogram based upon Ar 

 = 3 hr is roughly comparable to that based upon ^t = 1.5 hr. Conse- 

 quently, in subsequent studies on the subjects, measurements were 

 made at 3-hr intervals, rather than at intervals of 1.5 hr. This reduc- 

 tion by one-half of the number of measurements certainly eased the 

 workload of the investigators. But, possibly, a more important conse- 

 quence of the choice of Ar = 3.0 over ^t = 1.5 is the probable reduc- 

 tion of interference (by measurement) with the phenomenon studied 

 (Halberg, in press; Koehler et al., 1956). 



Other illustrative periodograms are shown in Fig. 7. We check such 

 results invariably by the empirical interpretation of plots of the meas- 

 urements, which served for their computation, but this is not always 

 easy, when T is short and t is close to 24 hr, and we also need a meas- 

 ure of variability of the estimates of t and C, a measure not yielded in 

 the Schuster periodogram. By using the extended Cramer-Rao in- 



