Mr. Meikle's Reply to Mr. Ivory. 319 



(448=+.) [(A)-- 1] 



where no inconsistency can be detected, and which, for aught 

 that yet appears to the contrary, may hold good throughout 

 the whole range within which air maintains the same form 

 and constitution.^ This latter formula is equally free from 

 the glaring case of tinder kindling under the boiling point, 

 from a limited rise of temperature, and from an indefinite 

 descent below the impassable limit of 448°. 



It is curious that Mr. Ivory's answer should conclude with 

 his old complaint against M. Poisson's integrations. For if 

 we will persist in assuming inconsistent hypotheses, there is 

 no need for wondering at strange and incoherent results. 

 The whole mystery arises from their assuming the common 

 scale of the air-thermometer to be a true scale of temperature, 

 which is utterly inconsistent with what they also lay down — 

 that the specific heat of air under a constant pressure has an 

 invariable ratio to its specific heat under a constant volume. 

 Such mathematicians ought to know (for it is upwards of two 

 years since I laid it before the public), that, if the invariable 

 ratio just mentioned be made a fundamental principle, the 

 necessary and unavoidable consequence is, that, when the 

 variations of the quantify of heat in air are uniform, those 

 of its volume, under a constant pressure, form a geometrical 

 progression. I have more recently touched on this point, in 

 an article which was written and sent off before seeing Mr. 

 Ivory's answer. From it, the reader will be enabled to estimate 

 what confidence is to be put in those *' easy deductions from 

 the usual theory of the thermometer," on which' Mr. Ivory is 

 incessantly harping*. 



I wish it to be distinctly understood, that, in discussing this 

 subject, I do not endeavour so much to establish a particular 

 theory, as to point out some of the consequences which are 

 unavoidable, when we proceed on certain data ; and I only 

 insist on these consequences within ihe range throughout 

 which the data are supposed to hold good. 



Henry Meikle. 



two * "^^^^ article will appear in our next Number.— -Ed. 



Z 2 



