' QUANTITY OF HORlZONtAL REFRACTION. 47 



The definite refleding furface which he fuppofes to take Objeftion to ths 

 place between two flrata of air of different denfity, is by no ^""^^^ °"^** 

 means confiftent v/ith that continued afcent of rarefied air 

 which he himfelf admits ; and the explanation founded on this 

 hypothefis will not apply to other cafes, which may all be 

 fatisfaftorily accounted for, upon the fuppofition of a gradual 

 change of denfity, and fucceffive curvature of the rays of light 

 by refraflion. 



I have fince learned that the fame fubjed has alfo been ably The fubjedl well 



treated by Mr. Woltman, in Gilbert's Annalen der P//j/^i ; ^""^^^^^ ^J' Wolt- 



■' ' •^•' man. 



but I have to regret that his differtation, as well as that of 



Gruber, in the fame Annals, were written in a language that 

 was unknown to me, and that I could not avail myfelf of the 

 affiftance that I might otherwife have received from their 

 refearches. 



When I forrperly engaged in this inquiry, being impreffed Obfervations 

 with the advantajje to be derived from it to nautical aftronomy, °^^'' ^^^ Surface 

 ^ f.i ■ .• • ., T f.i ,1 • of the Thames, 



on account of the variations in the dip ot the apparent horizon, 



from which all obfervations of altitude at fea mull neceffarily 

 be taken, I fuggefted the expediency of a feries of obfervations, 

 to be made by a perfon attentive to thofe changes of tempera- 

 ture or moifture of the atmofphere, on which he might find the 

 deprefllon of his horizon principally to depend. I had at that 

 time no expedation that I could myfelf purfue this fubje6t 

 farther to any ufeful purpofe, having little profpedl of refidiug 

 for a fufficient length of time in view of the fea, and feeing no 

 other method by which the fame end might be accomplifhed. 

 I have, however, fince that time, found means to fatisfy myfelf^ 

 by obfervations over the furface of the Thames, that although 

 the quantity of refra6lion varies in general with any change of 

 the thermometer or hygrometer, yet the law of thefe variations 

 is not altogether fo fimple as I had hoped it might be found. 



I fliall, on the prefent occafion, firft relate the fa6ts on which Narrat'iTc. 

 this opinion is founded, and which are in themfelves fufficiently 

 remarkable, on account of the unexpeded quantity of refradtion 

 obfervable over a ftiort extent of water ; I iliall, in the next 

 place, (liew that the exact determination of the concurrent 

 changes of the atmofphere are of lefs value, and their irregu- 

 larities of lefs confequence, than I had conceived, as there is a 

 very eaiy method whereby the quantity of dip at fea may be at 

 any time corredtly meafured ; and therefore the end which I 

 fought by indired means, may be at ence diredly attained. 



The 



