( 549 ) 



Fii'st (»r all WO may diM-ixc iVoiu /> llic pressure w liieli llie reiiiaiiidor 

 <>r tlie (ixyu'eii would liaxc exerled in Hie \oliiiiie O.DDDT'.I I "., and 

 al a liMnpei-aliire of ll)\(S| ('. In ihis calcnJalioii die law of liovi.K 

 may i)e a|>|)lie(l w idi sufficient accuracy. We dien lind : 0. 022 cm. 

 The |>i'essiiiv \\lii(di llie ((iiandly of ().\y,i>'en Iraiisferred lo die mixinre 

 would (\xefl ill die \(.liimo 0.1)1)1)711 T,, al U)°.<SJ (\ is dicii re[)i-cseiile(l 

 \vidi siiflicient accuracy liy : 



40.396 - 0.022 = 40.374 cm. 



The |)ressiii-es of Ihe oxyudi and the carhoii dioxide may l>e reduced 

 lo the same lem|)crature 1 1)'^'.<S2 ( \, l>y deri\iii.u' coeflicieiits cd" pressure 

 \arialion from the series üi\es in ('oium. N". 71, or Ity means of 

 the \alue for oxyueii 0.003()74 found directly l»y .loi.i.v, and ihe 

 value 0.003711 \\ liich follows from ('iiapimis' data (Traw el Mém. 

 <\\i Hiireaii inlenialional des Poids et Mesures. t. ii p. 124) for tli(> 

 "reaT' coefficient of pressure xai'iation of cai-hon dioxide at 20^ (^ 

 and au initial pressure of 1 atmosphere ^). 



We liiid for \\\o oxyueJi : 40.375, for Ihe carlion dioxiih' : (iH. 1)59 cm. 

 If /y,. I" are the jiressiire and the Noliime of ihe lirst i;as, p^, 1"^ 

 of the second, it may he easily found that the niiniher of molecules 

 of the first ^as is proportional to: 



r^'^i 



that of the second lo : 



/\ K 



1 I ^^^ 



where, at least with reiiard to Ihe denominators, />, and j)^ are 

 expressed in atmospheres. We lind from the data of (*omni. N". 71, 

 These Proc. Ill -liuie 29, 1901 p. 130 and 132: 

 for oxygen : 



.lj,„ = 1.074237 



Ba.^ = — 0.74278 X 10 '. 



i) Wo shall not consider lierc tlir viU'ialion f)!' the coefficient of pre.s.siire-variation 

 Avitli the initial piessiiie: liiis variation \V(»iil(l he for a (lillerence of ."")() cm. in 

 llie initial pressure accortling to lliiArruis, 1 c., O.OOOO;! I, whicli with a reduction 

 of tlic pre-ssure for 1'^ (litreiciice of lein|ieiatiiie would give an error of only 



— —- . Ixeeping in view llie t'aci that such large (lilt'er(!nce.s in the temperature 

 30000 



llo not occur, this uiav he neglected, 



30 



