VOL. XVI,] PHILOSOPHICAL TRANSACTIONS. 303 



is much about as 1 to QO, or as the excess of the secant of about 8-i^ degrees to 

 radius : for if E be the eye of the observer, S a place where the sun sets at the 

 end of twiHght in E, and the angle ECS, or TCA be found 18 degrees, the 

 excess of the secant of its half ECH, would be the height of the air, viz. GH : 

 but the sun's ray ASH, and the visual ray EH, do each of them suffer a re- 

 fraction of about 32 or 33 minutes, by which being bent inwards from H to- 

 wards G, the height of the air need not be so great as if they proceeded straight 

 on ; and having from the angle ECS taken the double refraction of the hori- 

 zontal ray, the half of the remainder will be Si degrees, whose secant being 10 

 HI, it follows that as lOOOO to ill, so is the semidiameter of the earth, sup- 

 posed 4000 miles, to 44,4 miles ; which will be the height of the whole air, if 

 the places E, S, whose visible portions of the atmosphere, ERZH and SHKB, 

 just touch each other, be 18 degrees asunder. 



At this height the air is expanded into above 300 times the space it occupies 

 below ; and we have seen the experiment of condensing it into the 6oth part of 

 the same space ; so that it should seem that the air is a substance capable of 

 being compressed into the 1 80000 part of the space it would naturally take 

 up, when free from pressure. Now what texture or composition of parts shall 

 be capable of this great expansion and contraction, seems a very hard question ; 

 and which is scarce sufficiently accounted for, by the comparing it to wool, 

 cotton, and the like springy bodies. 



Hitherto I have only considered the air and atmosphere as one unaltered 

 body, as having constantly at the earth's surface the 800th part of the weight 

 of water, and being capable of rarefaction and condensation in infinitum ; 

 neither of which is rigidly true: for the weight of the whole atmosphere is 

 various, being counterpoised sometimes by 28^ inches of mercury, and at other 

 times by no less than 30-i- ; so that the under parts being pressed by about a 15th 

 part less weight, the specific gravity of the air on that account will sometime* 

 be a 15th part lighter than another. Besides, heat and cold very considerably 

 dilate and contract the air, and consequently alter its gravity; to which add 

 the mixture of effluvia or streams rising from almost all bodies, which, assimi- 

 lating into the form of air, are kept suspended therein, as salts dissolved in 

 liquors, or metals in corroding menstruums ; which bodies being much heavier 

 than air, their particles, by their admixture, must needs increase the weight of 

 that air, with which ^hey lie incorporated, after the same manner as melted 

 salts augment the specific gravity of water. The other consideration is, that 

 the rarefaction and condensation of the air is not precisely according to the pro- 

 portion here laid down ; for though experiment very nearly agrees with it, as 

 may be seen in the 58th chap, of Mr. Hook's Micrographia, yet are the con- 

 densations not possible beyond certain degrees ; for air being compressed inta 



