VOL. LXIX.] PHILOSOPHICAL TRANSACTIONS. 575 



gave 4 feet less than the geometrical measure for the former of the depths, and 

 2 feet more for the latter. 



General Observations, — So long as in barometrical measurements we shall 

 consider as given only the difFereiices in the weight and heat of the air at the 

 places of observation, we shall be subject to errors ; because there are many 

 other causes of modification in the air : and all the exactness to which we can 

 pretend, will be to determine a formula which ])reserves a mean among the pos- 

 sible variations. This is, says Mr. De Luc, what I have proposed to do in my 

 own formula, and it seems to answer this end. In the different trials which have 

 been made of it, it has sometimes given the heights too great, at other times 

 too little, without distinction of climate. Thus, for example, the experiments 

 at Spitzbergen by Lord Mulgrave, and at the Pike of Teneriffe by Mr. De 

 Borda, one of the French academicians, gave the heights too great ; those of 

 Col. Roy and Sir Greorge Shuckburgh, made in mean latitudes, and partly in the 

 places where I myself had observed, gave them too little. 



Tiiese differences do not seem to depend on the climate, and indeed I have 

 frequently observed them to happen in the same places. Thus, for example, my 

 observation on the Glaciere de Buet, cited by Sir George Shuckburgh, gives the 

 height of that mountain a little less than the geometrical measure ; but Mr. 

 De Saussure having repeated the barometrical observation, it agreed with that 

 measure by the same formula ; and Mr. Marc Pictet, by a 3d observation, 

 found the height a little too great. In these 3 observations, the corresponding 

 point was Geneva, distant about 10 or 12 leagues. At that distance there are 

 doubtless some causes of variations which are irremediable ; since the formula 

 supposes that the observations are made in the same column of air. It is there- 

 fore only in the cases in which that supposition approaches near the truth that 

 we can hope to perfect the rule. But this can only be by introducing new con- 

 ditions into it ; that is, other modifications of the air, of which we have not as 

 yet taken any account. 



In meditating on the causes of the diversity of results in experiments, it has 

 always appeared to me, that the differences of the effects of heat on the air, 

 according to the different states of it, was the principal ; that is, that the air not 

 being always of the same nature, heat, that grand cause, whose effects we ought 

 principally to determine, does not always act equally. Besides the particular 

 experiments which prove it, we can attribute to these differences only those of 

 the results of the researches of some philosophers concerning the dilatations of 

 the air by heat, applied to various physical uses. 



In a paper lately read at the r. s., on the subject of refractions, I analysed and 

 compared different formulse of this kind, given by philosophers on whom we 

 can depend. The result of that examination was that, supposing the volume of 



