ON EARTH TEEMOBS. 



315 



forming gi-oups, certain enlarging factors being applied, as is done in th© 

 rednotion of tides. 



When these results had been obtained it immediately appeared that 

 all the constants down to the smallest terms presented more or less con- 

 siderable periodic changes. The following table contains two examples, 

 one showing the coefficients of the first harmonic term for Potsdam * 

 (a, cos ^ + 61 sin i), the second those of the second term for Wilhelms- 

 haven - (a2 cos 2^ + 62 sin 2t). 



Each group from which these numbers were computed contained 

 about twenty days. It is evident that these two periodical changes pre- 

 sent an entirely diiferent character. In the first case a^ and h^ vary with 

 each other,^ and the changes are proportional to the mean values of these 

 constants. Putting w, cos (Mi — t) in place of a^ cosi + &i sini, there is 

 no considerable change in M,. In the second case the changes of a^ and 

 62 are independent of each other, but of about the same range, and the 

 phase is subject to considerable fluctuations. It is immediately seen that 

 the variations in a<^, h^ are due to the superposition of a lunar wave, which, 

 moves on as C changes, whilst the variations in a^, h^ may be very nearly 

 represented by a periodical change of mi alone. 



Similar observations may be made in comparing any of the twenty-four 

 lists of coefficients printed in the ' Astr. Nachr.' There are several cases in 

 which the influence of a lunar term is easily recognised, and others in 

 ■which it is easy to see that both causes act together to produce periodical 

 changes. 



When treating of the daily period it will be shown that the range of 



' Expressed in units of 0"0020. = Expressed in units of 0"-0028. 



' Tliis is not quite so evident in the values of «i on account of their smallness- 

 and irregularity. 



