206 " I'HILOSOPHICAL TliANSACTIONS. [aNxVO 1777. 



D.'s were also. Within what limits my barometrical errors are to be found, is 

 not difficult to determine from what has been before premised. That the scale 

 of Mr. De Luc's barometer was less accurate than mine, is, I think, without a 

 doubt ; and indeed he never attempted a division less than -Jl. of a French line, 

 or about -r-oV<r of an inch English : and yet when I consider the number of his 

 observations, and the unexampled diligence and care with which he made them, 

 I am obliged to attribute the difference of our results to some other cause than 

 that of inaccuracy. If then future experience should demonstrate, that the 

 density of the atmosphere with a given heat is invariable, or nearly so ; while 

 the pressure of a whole column of it continues the same, we may perhaps search 

 for the cause of our disagreement from hence, viz. the barometers of Mr. 

 De Luc were not sufficiently near each other in an horizontal direction : mine 

 were separated from 2 to 3 miles ; and his, I believe, at double or triple that 

 distance. It may be suspected, I am well aware, that the syphon construction 

 of Mr. De Luc's barometer might occasion this difference : let us see whether 

 this be the case. Mr. De Saussure (whose instrument was of Mr. De Luc's 

 construction, and made as I understood^ under his inspection) observed at the 

 top of the Mole, or at least nearly on the same level with my barometer, as 

 follows: after relating and calculating the observations by Mr. Saussure's obser- 

 vations, it turns out that the mean of 2 sets gives a difference of about 10 

 inches only in deducing the height of the mountain, a quantity wholly to be 

 neglected. Finally, the mean of Mr. De Saussure's observations gives the 

 defect of Mr. De Luc's rules 21.g in 1000. The construction of the barometer 

 had therefore no influence on this difference. But further, while Mr. De Saus- 

 sure observed the height of the barometer on the Mole, Mr. De Luc, the 

 brother, made a corresponding observation with a similar instrument at Geneva j 

 the- result of which, computed after Mr. De Luc's manner, gives, for the sum- 

 mit of the Mole above the lake of Geneva 4814 English feet ; whereas, by a 

 mean of Sir Geo. S.'s trigonometrical operations, this height was 4883 feet. 



This last observation serves at least to show, that the error I am contending 

 for is on the defective side, though it gives the quantity of it somewhat less, but 

 by no means deserves that confidence which the other comparisons do ; for, 

 besides that this single observation may be concluded less decisive, the trigo- 

 nometrical measurement is also less accurate Ironi the distance ; and lastly, to 

 suppose the state of the atmosphere precisely the same with respect to weight in 

 two places 20 miles asunder, is, I am afraid, a postulatum too hazardous to 

 grant. I therefore say, that all these observations confirm the same truth, that 

 the atmosphere is lighter than Mr. De Luc presumed it. What had already 

 been done may seem sufficient for the establishment of this iiict ; for I have 

 always held, that a few observations, well made and faithfully related, do more 



