396 ANNUAL REPORT SMITHSONIAN INSTITUTION, 19 64 



and so the coimts are measured actually at points C and D on the curve. 

 The difference in counting rates between the two scalers measures the 

 change in frequency due to the gravitational effect. 



Because of the 74-foot distance between the source and detectors, 

 the counting rates are rather low, and it takes many hours to acquire 

 sufficient counts to give good statistics. In the course of the experi- 

 ment it was noticed that there was a slow drift in the relative counting 

 rates of the two scalers — a drift large enough to wipe out the effect 

 being sought. The drift was somewhat periodic — with a period of 

 about 2 days. 



After a good deal of soul-searching and examination into possible 

 causes for this drift, it was finally realized that this error was a result 

 of the small temperature difference between the source and absorber. 

 It requires a difference of only 1° C. to produce a frequency shift as 

 large as the one being looked for. 



The temperature correction turns out to be directly related to the 

 relativistic Doppler shift mentioned previously. The atoms of the 

 radioactive source are vibrating in their crystal lattice with a mean- 

 square velocity proportional to the temperature, and the relativistic 

 Doppler shift depends upon just this velocity. Utilizing this as a 

 basis. Pound calculated the amount of correction to apply in order to 

 eliminate the temperature effect. 



When this correction was used, it was found that the remaining 

 frequency shift agreed very well with that calculated from the theory 

 of relativity, using the 74-foot difference in height between the source 

 and absorber. 



In this w^ay the gravitational "red shift" was verified. 



THE CLOCK PARADOX 



Following publication of this experiment, it was noticed by C. W. 

 Sherwin, of the University of Illinois, that the clock paradox plays 

 a role in this situation (Sherwin, 1960) . 



We recall that the relativistic Doppler shift is associated with the 

 time dilatation of the moving source. In the present experiment, the 

 source of radiation is an atomic nucleus which is vibrating back and 

 forth in a crystal lattice with a mean-square velocity proportional to 

 the temperature of the material. The emitting nucleus actually goes 

 back and forth many times during the time that the wave packet is 

 being emitted. 



This, we see, is very much like the clock paradox situation described 

 at the beginning of this article. Instead of a spaceship going away 

 and coming back, we have a radioactive nucleus going away and com- 

 ing back many times. The radiation passing between the emitter and 

 absorber is a means of continually comparing the clocks located on the 



