( 583 ) 



smaller, iiaiiiciv oils, Ihc raiio of llic xaliics ol' {h,i - hj^ would also 

 lia\'e hccii round ('(|iial lo 1,7.'). WC iiiav llicrcloi-c safch a^siiiiic 

 lliat {!>,, h)" is foiiiid lo he acciii-alclv /iroitoji/ona/ lo Hie l('iii|i(M-a- 

 liii'c wilhiii so larii'c an intcrxal of Icinpci-alnrc as llial Itdwccn (P 

 and 200', in conscMiucncc of w Incli llic (|naiitilv a ninsi he (jiu/i' 

 'm<l('ih'ii(l('nl of llic l(Mn|KM-aliir('. 



Il is not asloiiisliiiiLi' thai a is indc[)(Mi<leiit of the l('in|»('raliii'(> ; 

 \\\v conli-arv would ratlici- seejii lo he roniai'kahlc. iiein*;' induced 

 to make this coiilrarv sn|)|>()siti()ii lor llio holtor ajire(Mïioiil of the 



(|uaiitilA" loi* ^VA, \\\\\\ the e.\|KM-inienls, Prof. v. i). Waals ') 



innnedialelv poiiiled out its astoiiishinu- charaelei-. 



We shall just draw attention to the following eojisetinenee of the 

 fact, that h,i — />,, is projxu'tional to \^T. 



If we put : 



then e(juation (1) mav he Avritten as follows: 



v—b yT " l/y2' yT 



With small value of />,y — h and great value of r, \\(' get anproxi- 

 mateh' : 



V l/y7'' 



therefore 



yT 



r being' in this ease approximatelv e([ual lo , we get: 



P 



I 1 — ^^ 

 ' 2R ' 



or 



h — h,, — y' p. 



i.e. the \alue of h depends oidy on /y and Jio more on /• or 7', the 



value of h,i being neai'lv constant, 'idie values of />, ealeulale<l for 



tin' same pressures, ha\'e therefore the same (lilference whether the 

 tempei-ature be 0^ or 200 . For we have: 



l>,-l>, — y' {/>,—/>,). 

 We found this fact affirmed in the abo\-e tables-). 1^'or th(^ pnrpose 



1) 1. c. Ill, p. Oi(i. 



-) I pointed tlii.s out nli-cady before in u paper in llie .[ir/tirt's l)///ir („Sur 

 rintliieiice des correclion.';, etc." (•2) VII, 8me partie, p. -H't — 27.) I tested there the 

 /'-values I'm- liydroiicu to an einpiiieu! luiuiula n\' KAMi-iinxüH Onxks. 



