( 559 ) 



Jïivoii ill § 3, wliih^ llic \oliimr ol' llic- iiaiTow caiiilliii'v ahox'o 

 l)('l\\(MMi llie iiiai'k and llic place w lici-c il is sc^alcd was accoiinlcd 

 tor, a lalilo \vas dorixed. This iiexv lahic i;i\('s lur each division (^ \\\c 

 deiisilv (/a <>t Ihe hvdrogeii Avhen Ihe ineivui'v readies thai dixision 

 at a leinperalure of 20° (\, expressed in terms of Ihe iiorinal deiisit\ 

 (O (\, 1 alm. 45 N.L.) as nnilv. Kor llie \alnes of (^ helween 

 25 and 50 die table increases bv 0.5; lieiiee also tbr lli(> hij^-hcr 

 values of tlie pressure (as far as (^ = 40) the erior made b\ 

 interpolating linearly is less than 1 atni. 



To derix'e from these densities the corresponding' pressures we niiist 

 i<now the isothermal of hydrogen at 20"^ (/. Measiireiiicnis of this 

 ha\e been made by Schalkwijk; they do not, lioxve\er, exceed the 

 density 54. An extrapolation from these observations for the densities 

 wanted is not allowed with a view to the mean error of his determination 

 of the (' of the series of KamkrlinCtH ( )nnks (cf. SciiALKWi.ik's Thesis 

 for the doctorate, p. 115). Obserxations at higher densities luue been 

 made by Amaoat (fi-om 100 to oOOO alms.) The isothermal for hydrogen 

 at 20" V. derixed from Amaoat's data : 



IH'A ,, — i. 07252 + 0.0007 J 94 ,1a + 0.000000672 d a ' 

 (cf. Schalkwi.jk's Thesis for the doctorate p. i2J) does nol agree, 

 howexer, xvitli that from Schalkw i.ik's observations. From these 

 for instance, foUoxxs at (/_i := 55 : /y/" j.^,, ^ J .1 1 194 (cf. I.e. p. 124), 

 hence y> = 61.157, xvliereas frcnn the isothermal gixeii bv A:\ia(iat 

 we deri^■e at ƒ/ = 61.157 : r/j = 54. <S97, Schai.kwi.ik's observations 

 have been made very carefnlly es[)ecially nis determination of the 

 normal volume (see C-omni. N". 70, V, These Proc. 1\ lune 29, 1901, p. 

 107). If with this we conipai'e the ^vay in xvhich Amac^at has determined 

 his normal xolnme (Ann. de ('himie el de IMiysiqiie, t. 39, 1893, p. 83) 

 it seems not entirely xxdthont reason if, xvhile xvaiting for moi-e 

 accnrate <leterniinations of the isothermals of hydi-ogen at higher 

 densities, we make those of Amagat agree xvith those of Schalkwijk 



by multiplying all the volumes of tlu^ former by the factoi- — '--— 



55 



So we obtain : 



pvA,, — 1.0705 -f- 0.000717 r/.t -|- 0.00000067 dX' ■ 



To test this xxc compai-e the value of /y/\,j,„ at <l ,\ := 1, \'va. -. 1.0712 

 resnlting from it, with the xaliie of />/•.(.,„ xvhich folloxvs from 

 Ihe value of Ihe coefficient of expansion according to (^happuis: 



ri, = 0.0036606 '), jvi/: 1.0732. The differejice is , so that the 



1) See Schalkwijk, Thesis tor lliu doctorate ji. 110. 



