1 82 Mr J. D. FOKBES on the Horary Oscillations 



dify a and , we may obtain a different expression, which shall 

 very closely represent the observations. We shall find for the 

 new equation 



z =. 3.07 cos 3 .36 as a near approximation. 



The first of the above integrals then becomes 



cos? 6 .36) dn, 



which, taken between the limits 6 = and 6 =. \ *, gives for the 

 mean oscillation, in relation to the quadrant of latitude, 



m ' m - 

 _ .36 = 0.94. 



The other integral above given becoming 



/(3.07 cos 4 .36 cos <) dt, 



there results, within the same limits, 



Q m.m. 



3.07.-^*- .86 = 1.45 

 lo 



for the mean value with regard to the surface of a hemisphere. 



These numbers correspond to homogeneous columns of air 

 at the ordinary pressure and temperature 10^ and 16 metres in 

 height respectively. In the investigation of any supposed con- 

 nection with temperature, these mean results will be of some 

 value. 



18. I am satisfied for the present with having pointed out a 

 formula representing very closely the observed amount of oscil- 

 lation at the level of the sea, as depending upon latitude. For 

 any successful generalization upon the influence of height, we 

 must wait for vastly more extended data than we already pos- 

 sess ; and the same remark is applicable, though to a less extent, 

 to the influence of the seasons. That temperature has an im- 

 portant connection with the geographical distribution of this 

 phenomenon, I have no doubt. It appears probable, that, 

 among places under the same latitude, and having different 



