from metallic fubftances, as I have found by experiment, while 

 rain is falling in fummer, or dew forming in an evening; for if 

 a vefTel of tinned iron be rubbed over with a wet fpunge, 

 and then fufpended with its mouth dow^nwards, its inner furface 

 will foon become dry again, though rain be falling or dew forming 

 at the time. 



As thij evaporation cannot be entirely prevented by any 

 method of conflruding the gage, if the quantity of water loft 

 this way could be determined, this imperfedion would be 

 correded; and I fhould think that this quantity might be 

 determined by two contiguous gages. For, let A and a = 

 the areas of the apertures of the two gages ; B and b = the 

 curve furfaces of the funnels ; S, s = the quantities of water 

 colledcd by them in a given time in grains ; X and x the 

 quantities loft by evaporation in the fame time ; then S -f- X 

 and s -\- X, being the quantities received by the gages, we have 



a S— A s + a X 



A : a : : S + X : s + x, and x = ; butfincethe 



A 



quantities evaporated in the fame time are as thofe furfaces, 



T, , V J b X _ a S— A s + a X , 



B : b : : X : x, and x = -— - — ; hence X = 



D A 



5JL5iI^:^, and the quantity correded = S + X =, 

 Ab— B a ' H ^ 



A b ^ i 5_A s j^^j. ^-^^ cones muft not be fimilar, for in that 



A b— B a 

 cafe both the numerator and denominator would be = o , and 

 confequently nothing could be determined. Indeed, if it was 

 worth the while, it might even be determined in this cafe, by 

 taking the fluxion of the quantity. 



K k 2 The . 



