316 REPORT— 1893. 



motion is on an average very nearly proportional to either the quantity of 

 sunshine or the maximum oscillation of the temperature during the day. 

 Though one might expect that in a series of observations of more than 

 half a year's length, grouped in such a regular way as here, meteorological 

 effects would be sufficiently eliminated, this is really not the case. In 

 comparing the meteorological data with the results mentioned above, it 

 was found that the first class of periodical changes can be explained to a 

 great extent by meteorological changes. 



The best way to avoid these disturbing influences would be to place 

 the horizontal pendulum in the bottom of a mine, where the effects of 

 meteorological changes are reduced to a minimum whilst, as far as we 

 may judge, the lunar effects must remain the same as on the surface of 

 the earth. 



Since every periodical change in the coefficients a, h can be considered 

 as the combined effect of changes of the first and second kind, it is im- 

 possible to decide in the cases of the two stations at Potsdam and Puerto 

 Orotava, where the effects of temperature appear to be especially strong, 

 which part is due to the first, and which to the second, cause. We must 

 content ourselves with the following result : The comparison of the ob- 

 served temperatures with the harmonic constants of the daily oscillation 

 shows that the former play an important part in producing periodical 

 variations in the latter. But, on the other hand, there is some evidence 

 of the existence of a small lunar wave in both places, amounting to 0"'01, 

 or probably more. I do not hesitate to express the opinion that if it had 

 been possible to continue the observations through a longer interval of 

 time, and thus to eliminate the effects of temperature, a more positive 

 result would have been obtained. 



In a former investigation,^ in which I employed the nsual method of 

 tide reduction without having regard to the motion of the zero-point and 

 the variation of the daily period, numerical values of a supposed semi- 

 diurnal wave caused by the moon have been computed. These values 

 agreed pretty well with the formula representing the ordinary gravita- 

 tional deflections of the plumb-line, for whilst we flnd the semi-diurnal 

 wave to be 



+ 0"-0142 cos (2i-247°-5) and -l-0"-0128cos (2i-228°-5) 



for Potsdam and Puerto Orotava respectively, the formula of deflection 

 are : — 



+ 0"-0099 cos (2^-270°) and +0"-0142 cos (2<-270°). 



Until further researches have been made, it seems advisable not to 

 place too much confidence in these numbers, though it is not impossible, 

 from the way in which they were computed, that they represent a near 

 approach to the real values. 



Whilst we now see that the observations obtained at Potsdam and at 

 Orotava were insufficient to demonstrate the effects of the moon on the 

 plumb-line, the third series of observations at Wilhelmshaven, as might be 

 expected after the preliminary investigation mentioned above, led to a 

 positive result of some interest. It will be shown later on that at Wil- 

 helmshaven, though the range of the daily oscillation is extremely 

 variable, there does not exist such a clear relation between it and the 



• See Bos Horizontal^endel, &c. pp. 87-104. 



