VOL. LXVIII.] PHILOSOPHICAL TRANSACTIONS. 407 



Thus, if the mean height of the barometer were 27 inches, and the tempera- 

 ture 52", the difference of the results would be 4 feet in a thousand; liow far, 

 therefore, this is of moment is left to the judgment of the observer. I conclude 

 lastly, sir, with a comparison of your actual observations made in Great Britain 

 computed after my tables, deduced from a series of observations made in Savoy. 

 I have indeed only collected \6 of your observations; but as I have chosen such 

 as presented themselves as most proper, either on account of their heights or 

 temperatures, I imagine these will be very satisfactory. From these it seems, 

 that the error of my tables, from a mean of all these observations, amounts to 

 -j- ^ „ Vo -o? of yours, to + -totto-o; but it must be remarked, that the standard 

 temperature, from which I compute, is 31°.24 of Fahrenheit, whereas in your 

 computations it is assumed at 32°.0; this difference of 0°.76 is equal to , „'/„ „ in 

 the correction for the expansion of the air: if then we were to set out from the 

 same zero, viz. 32°, which I have proposed in my memoir, the error of my 

 tables, according to your observations, would become only , „ g „ ^ , that of yours 

 remaining tVoVo-' I would by no means from hence conclude, that any prefer- 

 ence is to be given to the former, but would say, that in most practical observa- 

 tions, in these countries at least, it is indifferent which method is used. These 

 same comparisons also afford us another piece of information, viz. that under 

 similar conditions the density of the atmosphere is the same, whether under the 

 parallel of 46 or 56 degrees of latitude. Till, therefore, more accurate obser- 

 vations than those of Mr. Bouguer can be obtained in the neighbourhood of the 

 equator, 1 should be extremely cautious how I admitted a latitudinal equation; 

 nor do I think the single observation, related in Lord Mulgrave's Voyage towards 

 the North Pole, of sufficient authority itself to establish such a theory on. 



I shall now beg leave to conclude with what I flatter myself will not here ap- 

 pear improper, a new rule for reducing the observations, and which I hope will 

 be found particularly commodious, as it requires no logarithms, nor any other 

 than the following short table, which may be engraven on the scale of a thermo 

 meter, and therefore, always accompanying the instrument, will serve for com- 

 puting the observations on the spot, if the height should not exceed 4000 or 

 5000 feet, which will frequently be found very satisfactory. 



The adjoined table gives the value of -^ of an 



Ther 

 32 

 35 

 40 

 4-5 

 50 

 55 

 60 

 65 

 70 

 75 



85.86ft."l -« A c 

 87.49 



6 .- s' <J c 



S ^ ~ S a inch on the barometer in feet in the atmosphere. 





re a 







O _ T3 



<U 0) "^ j3 



when the quicksilver stands at 30 inches, for 

 S every 5 degrees of temperature from 32° to 80°; 



88. 5i 



89.60 



90.66 ,p o * ^ ^ _ 



91.72 >'''SS.S2o£ and for any other height of the barometer it 



a J o I S g I will be in the inverse ratio of that height to 30 



92.77 

 93.82 

 94.88 

 95.93 

 80 J 96.99 



-; -a 



"2 S S " II |. inches. Thus, let a be the mean height of the 

 £ "o e •£ .S 1^ i two barometers in inches ; « the difference of 



