(511 ) 



(leiice of the distance sphei'cs, 2'"' a^ a i-oal cliaii.uv causcvl bv 

 (•()iii|)rossl()ii ^). 



At first 1 thouiiiit that 1 had lo make use of tlie foi'iniila, derived 

 bv Prof. VAX DKR Waai,s oji llie second sii[)posilion willi the aid 

 of llie tiieory of cyclic motion : when testin<^' this fornuda to the 

 Ijydrogen isolhenns of Amaoat. \'an Laai{ '), found it to harnioni/.e 

 well with lli(^ ol>ser\atioiis. lu tlie cah'ulation we are, however, 

 cojifronted with the dilïiculty tliat for tlie accurate detei'niination 

 of the constants A,, and h^, a pi-(diininary accurate kiiowled,u-e of 

 the a is re([nired. It ]iot having- been ascei-lained to which of the 

 two causes the \ariability of the /ms due, and it being iniju-obable, liiat 

 the first mentioned cause can be left out of consideration, I ha\e used 

 tlie formula deri\ed on the lirst suj)[)ositio]i for the variability of the /> : 



/, — /,,^ ]\—a - -f ^?-.; — r , + • • ■ • 



.a 



Of the eleven correction terms which will occur for s[»herical 

 molecules'''), only tlie tirst tw(t ha\e been calculated. 1 have conh'ned 

 myself here lo three terms. \\\ a comparison \vith tlie values of 

 y> and r observed i)y Amai^at at 15, '7 ('. we shall lia\e to deter- 

 mine the values of a, l>,,, «, ,i and y w liich agree closest with the 

 ol)servations between J 00 and oOOO ats. In tn'der to avoid when 

 applying the method of least st(uares, the elaborate calculation of 

 five n(U-nial e(piations with twiMity coefticitMils, 1 ha\'e determined 

 the most pi'oliablc \abies of d and A,, with the aid of assumed 



values of <?, ^i and y. For this purp()se I put ^e ^ ^, which value was 



b 



found according to two dilferent methods by Holt/mann ') and by \'ax 

 UKK Waai.s .lull. ') : further ji=:0,0i)5(S, Avhicli \alue has been calculated 

 by VAN r.AAK ') and a(lo[)te(l '') by IJoltzmann: (piite arbitrarily 1 put 

 y =r: 0,01 and I assumed as approximated \alues of r/ and />,y a := 0,002J-' 

 /v,/ = 0,0020 '). So if L(t and L It.j are the dillerences between the 



1) See Van dkr Waals, Tiie.se Proe. V. June 27 1903, p. 1;>3. 



, , n , BoLTZ.MANN Festsclll'if't p. 305. 



2) These Proe. V. March 28, 1903 p. 573. 



3) Vax Laak, Evaluation de la deuxième coi'iection sur la grandeur b. Arch. 

 Teyler, série 11, t. VI 1899 ]>. iS. 



■^) Gastheorie II p. I'yl. 



■') These Proe. V. February i28 1903 p. 4-87. 



'^) These Proe. I March 25, 1899, p. 398. Adopted namely lor the calculation 

 of Ills second correction term, which has the value 



,_3,^ 957 



' " Y ~ 89(jo' 



Here has the value 0,0958 calculated by van Laap. so /?' = 0,03(;9. (Febr. 24, 1904). 

 ') In Cont. I r/ = 0,0037, /> — 0.0026 IS derived h'om the observations of Kkgx.-vult. 

 When nuiiliplied willi 0,7(i with change oi' the unity of pressure, the aljove values 

 are obtained. 



