HIMALAYA MOUNTAINS. 257 



However this allowance of an eighth part of the contained arc £1111 

 exceeds the mean of terreftrial refraction: as appears from the trials 

 conducted by General Roy and Colonels Williams and Modge,* 

 and efpeeialjy from thofe of the lad mentioned obferver. They found 

 terreilrial refraction fubjeel: to great variation, amounting to no lefs than 

 ■id of the contained arcin* fqme inftances, and fo f mall as ^th of the 

 intercepted arc, and even lefs> or abfolutely o, in others* But, in the 

 numerous obfervations of thofe gentlemen, the extreme in (lances are 

 few; and the range of variablenefs is commonly within narrower 

 limits, from fth to vith, being on a mean either T vth or —th part. The 

 trials, moft to be depended on, being thofe which were conduced by 

 means of correfpondent and contemporary pbfervations, give a mean 

 of _? T th.. It appears alfo, that the refraction is leafl variable where the 

 ray panes through the air at a confiderable diftance from the furface of 

 the earth, for the greatefl part of its course: which is eminently the cafe 

 in the infiance under confederation; and efpecially in fome which will 

 be fubfequently noticed, where the altitude of the mountains was ta- 

 ken from elevated fpots: and, in all. the ray mud pafs for a great par£ 

 of its courfe through a ftratum of the atmofphere of much lefs denfity 

 than in the experiments of General Roy and Colonel Mudge, to which 

 reference has been made. 



It follows from thefe confiderations, that the mean terreftrial refraction 

 fhould not be taken at more than T Vth of the arc contained between the 

 object and ftation. This allowance agrees with that which Delambre 

 directs to be made: it exceeds what was found by Legendre, (viz. 

 T \.th) j and it approaches v«ry near to Maskelyne's eltimate of T vth 



* Phil. Trans, vol. 8o» 85, and 87. 



R3 



