EQUATION AND THE NATURE OE COHESION. 39 



! 2. Snip linr dioxide. S0 2 . 



{a). a = 2.03 BT c F e /im — PlF c 2 .P c = 77.65; T r = 4.30 (Car- 

 doso and Bell). r/,. = 0.52 (Cailletet and Mathias). 

 r t . = 123.07; #=0.513 by Cardoso, 1912. M = 64. 

 The computations have been made with the density 0.52. 

 I did not find Cardoso's value until some time after- 

 wards a = 7.48 X 10 12 



(/,). a = 6.5 P,.V? a = 7.75 X 10 12 



(c). a = fjc'M 2 Y I, L85 X 10 7 /3 d^. ft' = 85.5 . a = 7.56 X I0' 2 



(</). a = C'BT C V C . C' = 1.707. . . . .- a — 7.50 X 10 2 



(e). a = 3 sv 2 ' 3 MN l > 3 T? % \{d — D) (T c — T)' m - 

 s= 33.29 dynes at — 25° (Grumach by 

 capillary waves); sv 2!3 = 407.05- ergs, d 

 at — 25° = 1.4968 Lange); 2) is neg- 

 lected. T c — T=\S2 a = 7.84Xl0 12 



The mean of all these values is 7S2 X M 12 , a computed from the 

 weight and valences in the usual way with 8 valences in the mole- 

 cule, 4 to sulphur and 2 to each oxygen atom, is 7,53 X10 12 which 

 is in excellent agreement with the found values. The value computed 

 by van Laar is 6.74 which is as usual too low, being in fact some 

 12 % too low. With the usual formula of a = 27 7', 2 /64 X 273 2 P,. 

 we compute 6.63. It is clear that the computation from the number 

 of valences and molecular weight gives a correct value and the 

 method of van Laar does not. 



1 3 . Methyl formate. C 2 H 4 2 . 



(a), a = 2.03 BT e F~ e / 1.08 — P..F, 2 . T c = 48 7; P c = 45.030 

 mm.Bg.\ d c = .3489; V c = 228.95. a = 11.66 X 10' 2 



(Ô). a = 6.5 P c r r 2 a = 11.54 X »0 12 



( C ). a = MY !, 1 85 X 10 7 /3r/,- /i . /x'.= 1 Ï9.S6 



M = 60 a = 12.14 X 10 12 



(d). a = C'ET r r„. C' = 1.706 a = 11.28 X 10' 2 



(e). a = 3 MN ild T c C/d . C= 2.00 (Morgan and 



Schwartz) d = 1.887. N 1 ' 3 as before. a = I1.10X10 12 



The mean of these results is 11.54. Van Laar gives 11.91. The 

 calculation from van der Waals' on the assumption that b and a 

 are constant (« = 0.2371) gives 11.67. The correct value is evi- 

 dently about 11.50. The calculation from the molecular weight ami 

 valence number with 16 valences, i.e. 8 C valences, 4 H valences 

 and 4 O valences, gives 11.45 X 10 12 in perfect agreement. Van 

 Laar's value is here about 4 ° I too high. 



