of the Barometer near Edinburgh. 175 



for the Equator ; and I believe there is no series in existence, con- 

 tinued for several days together, which gives such a mean result. 

 The observations of DUPERREY, at Payta, which, by HUMBOLDT'S 

 table, was reckoned at so extreme an amount as 3.4, has dwindled, 

 by the analysis of M. BOUVARD, to 2.66, which is, in all probability, 

 a very correct result ; and my own opinion is, that future obser- 

 vations will limit the equatorial oscillation to 3 millimetres, or 

 perhaps even lower, except in anomalous situations. Now, the 

 result of the table we are discussing, generally gives an oscillation 

 nearer 4 than 3 millimetres ; but what is remarkable, the larger 

 numbers occur only where the amount of reduction is considerable. 

 Thus, confining ourselves to observations not at extraordinary 

 height? above the level of the sea (that is, under 2000 metres), we 

 do not find a single observation under latitude 45, affording an 

 equatorial oscillation amounting to 4 millimetres ; whilst between 

 45 and 55, in less than half the number of localities, we have no 

 less than five results above 4 millimetres. This alone indicates 

 something wrong. Had M. BOUVARD limited his passion for ge- 

 neralization to the law of latitude only, he would easily have given 

 a nearer approximation ; this was what I had proposed to under- 

 take before I had the good fortune to meet with his paper, and 

 what I afterwards executed with the aid of his data. 



13. As M. BOUVARD has presented us in his table with a con- 

 siderable number of unpublished results, which I was unwilling 

 to lose, I have trusted to the skill and fidelity of the author in 

 deducing the numbers which he has given as the observed amount 

 in his table. From observations continued but a few days, and at 

 irregular hours, it requires some labour and judgment to obtain 

 the truest results. It is therefore almost entirely in equatorial 

 results that M. BOUVARD differs in his numbers from the quan- 

 tities already assigned by such observers as have published their 

 registers, and from a very careful and extended comparison which 

 I have made with the originally published synopses of observations 



