EQUATION AND THE NATURE OF COHESION. 53 



39, and it may be higher, in spite of the lower value obtained 

 by the second formula. It may be possible that there are some 

 extra valences, and in support of this it may be said that octane 

 behaves chemically somewhat as if it were an unsaturated compound. 

 On the other hand it has also occurred to me that possibly the 

 valences are not strictly additive. Here we have at least 50 valence 

 electrons in the molecule. Now it may be that several valences 

 together might produce a little greater effect, electro-magnetic, or 

 cohesive, than an equal number taken separately. However that 

 will have to be determined by further work. The only indication 

 of this possibility is that there is apparently a slight tendency for 

 the computation from the molecular weight and valences to be a 

 little low as the number of the valences in the molecule increases. 

 But whether this is a real tendency or only apparent due to the 

 empirical latent heat formulae giving less good results in the more 

 complex substances, or to slight decomposition as the critical tem- 

 perature gets high, cannot at present be said. We have then as a 

 general result in octane to record that the determination of a by 

 the formula Q.5P C V C 2 gives a result in agreement with that calcu- 

 lated from the molecular weight and number of valences, but that 

 the other methods yield results, one of which is 15°/ above the 

 required value. Van Laar's value for a 1/2 is .276, which calculated 

 in the units I have used is 33.42 X 10 12 . This, as will be seen, 

 is very much lower than any of the values found. 



40. Diisobutyl. 6' 8 N l8 . 



(a), a = 2.03 B2 C F C /I.03 — P,.V?. T c = 549.8; d c = .2366; 

 P c = 18,660 mm.Eg. ; V, = 481.8.. a = 37.65 X 10 1 ' 2 



{b). a = 6.5P c r c 2 a = 37.54 X 1<> 12 



(c). a = fi!M*l$dW. M= 1 14; p' = 87.72 . . a = 41.56 X 10 12 



(d). a = C'RT C V.. C' = 1.813 a = 39.95 X 10 12 



(e). a = sv m 3 MN^ Z Tjd (T r — T). At 107.4° 

 5= 11.82 dynes (Schtff). sv 2i:i = 383.3 

 ergs. d= .6166; d = Sd c = .9014; 



T r — T= 169.4 a = 39.95 X I0' 2 



The mean value of all is 39.33 X *0 12 . The required value com- 

 puted from the molecular weight and the number of valences, 

 there being 50 valences in the molecule, is 37.54 X 10 12 . This is 

 exactly the value calculated from the second of the above formulae 

 but is between 4 and 5 °/ lower than the mean. The agreement 

 between the found values and that computed from the weight and 



