630 



ing pressure in order to' test wliellier leiiipernture t'orres|i(»iide(1 lo 

 pressure, the only stirring possible being made hy tlie moving np 

 and down of I lie carrier '). The magnetisations observed pointed to 

 much lower temperatures than corresponded to tlie aclua! pressures 

 and thus lo a large tem|)eralure lag. Tlierefoi-e these points have 

 been left out of consideration. 



^ 3. Discussion. For Gdll 0,02024 ') has been accepted as Cnrie- 

 conslant and with this value '5^^^ and o,,,,, have been calculated 

 according to formula (2). Half the real molecular weight has been 

 used in calculating o,,,^ from (J^, as (he atoms of Gd ai'e assumed 

 to have rotational freedom. This is usually done for salts containing 

 more than one metal atom in the molecule'); moreover, if (he 

 whole molecular weight had been taken, (j,„^ would have become 1^2 

 larger, a„ |/2 smaller and thus (i : (J«, again I 2 larger and one would 

 have found values larger than 1, as for o : «» the value 0.84 has 

 been attained (Cf. table IV^). 

 We find: 



^;,„<^ =z 434,2 iO" (38.65 WEiss-magnetons). 

 öo, =116,25. 

 For the Curie constant of Gd I we found ') 



C' = 0,02149 

 a,„«, =447,4 . 10' (39.82 WEiss-magnetons) 

 flee =119,79. 

 From the tables 1 and II o: o^ and a have been calculated for 

 Gd I and Gdll, with its own particular Curie constant for each 

 substance. The results have been collected in tables III and IV. The 

 values placed in stpiare brackets are a priori less reliable, mostly 

 because during or immediately after (he measurement the gadolinium 

 siilphale appeared to be not sufficiently below the liquid helium 

 level'). The differences between the observed values of (J : öqc, and 



') I.e. § 3«. 



') Gf. the preceding communication § 4, where on account of a later somewhat 

 modified calculation 0,02030 has been given. The dilïerence is of no importance. 



>) P. Weiss, Arch. d. So. phys. et nat. (4) 31 (1911) 



B. Gabrkea, J. de Cliim. Phys. 6 (1918) p. 442, especially p. 462. 



*) Cf. the preceding communication § 4, where the difference between both 

 results has been discussed. 



5) At the points marked with an asterisk the helium level was certainly below 

 tlie spring Fo (cf. the preceding communication § 3(0. Though a general tendency 

 to higher values of !r : o-^ (cf. the diagram) must be acknowledged to exist, there 

 is no systematic diflference between the points with and without asterisk. 



