470 EDITOR'S NOTES. 



8> 36. The equatorial rate 86263 S -60 consequently has a 

 probable error of + O s *36. 



If attention be now directed to the " differences " at 

 the ten stations between the parallels of 40 and 60 

 (which may be regarded as the representatives of the 

 middle latitudes in this investigation), we do not find, 

 either in the individual amounts, or in the systematic 

 occurrence of either + or signs, the indications which 

 we should expect to find, if there were a notable discre- 

 pancy peculiar to those latitudes in the general form of 

 the ellipticity of the hemisphere. On the contrary, 

 the signs of opposite character are interspersed in a 

 manner which indicates that they are of an accidental 

 rather than of a systematic nature ; and in regard to 

 amount, the sum of the actual discrepancies at the ten 

 stations (giving their proper value to the signs) is 8 *57; 

 whilst the (e probable discrepancy" of 10 stations (com- 

 puted from the discrepancies at the 22 stations,) is, as 

 above stated, O s< 55. 



If we combine the equatorial vibration 86263*60 

 successively and separately with the vibrations observed 

 at each of the 10 middle stations, and take the arith- 

 metical mean of the 10 results, we obtain from these 

 stations, so combined with the equatorial rate, '0051828 

 as the increase of gravity from the equator to the pole, 

 differing only -0000057 from the results of the 22 



stations, and corresponding to the ellipticity L_ j 



which, viewed as a partial result derived from the 

 middle latitudes, may be regarded as sensibly the same 

 as g 8*3. 4> already stated as the general result from the 

 equatorial and polar stations. 



Having thus reasserted, and more fully justified, the 



