OF CLOCK RATES UPON LONGITUDE WORK. 135 



If the phase of variation has a constant epoch, the maxima would 

 occur at the same instant for all stations, and the exchange of signals 

 would exhibit no changes. 



Without regard to the true difference in longitude, an exchange 

 of clock signals at intervals of six hours through one day, between 

 two stations 90° apart, may furnish a test of the phase of the vari- 

 ation, if it is in clock rates. The only necessary conditions are two 

 clocks with well-established uniform daily rates and the requisite 

 precision in recording the signals. 



It is hoped, and even hopefully expected, that a transatlantic 

 record of the wireless signals from Bordeaux may be employed for 

 this purpose. 



Our current series of observations should have a weight of ten 

 per night, as compared with the earlier work tabulated above, since 

 there are about eighty stars observed in each period of six hours, 

 while most of the earlier work included but eight stars in periods of 

 four hours. 



Whatever the result, when we close the cycle of observations next 

 September the weight will be assumed equal to that of all the pre- 

 ceding observations. This policy is justified by the rigorous atten- 

 tion that is being paid to the necessary details of program and of 

 reduction. An interchange of stars or the loss of a single observation 

 will always mask, if not mar, the effect of the small correction we 

 are sifting out. Nevertheless there has been, personally, a convincing 

 effect in the weight of evidence of the old work, from which these 

 variations emerge as a by-product — to close with an industrial figure 

 of speech. 



Lick Observatory, 



Mt. Hamilton, Calif. 



