Affinity and Dissected [Structural) Formulae, 409 



When t is above 100, the last term s 2 (100 — t) becomes 

 negative. 



15. By means of this formula the values of 2SF ]00 for various 

 substances given in the following Table (pp. 410, 411) are calcu- 

 lated. In several instances the data for the calculation of the 

 value of H 2 are only approximately known; and in some the 

 value of H 2 is only estimated as a probable guess. Inasmuch, 

 however, as the experimental errors in the determination of the 

 values of H l are considerable, and often as large as the value of 

 H 2 , a small error in the estimation of the latter does not much 

 affect the general result. 



The initials affixed to some of the numbers stand for : — 



R, Regnault. 



K, Kopp. 

 FS, Favre and Silbermann. 

 JT, Julius Thomsen. 



D, Dulong. 



A, Andrews. 



16. On examination of the affinity-values given in the 

 Table, it is evident that as any homologous series is ascended 

 there is usually an increase in the value of 2SF ]00 ; and so general 

 is this rule, that it seems extremely probable that the few excep- 

 tions met with are only apparent exceptions due to experimental 

 errors in the determination of the values of Hj &c. used in the 

 calculations. Thus the value for ethyl butyrate (147*7) is consi- 

 derably too high, 110*0 being probably near the true value 

 (§ 39); whilst that for ethyl formate (92*0) is probably also 

 rather too high, as a priori the true value may be expected to lie 

 between 68*0 and 91*5, the numbers found for methyl formate 

 and ethyl acetate respectively; probably 86*0 is near the true 

 value (§ 39). Similarly the value for amylic ether (83*5) is 

 probably a little too high, as the substance used by Favre and 

 Silbermann appeared to contain a little amylic alcohol, which 

 would diminish the value found for H x , and hence increase the 

 value of 2SF 100 . On the other hand, the value for stearic acid 

 (154*0) appears to be considerably too low, suggesting the pro- 

 bability of the presence of a little stearine, &c. in the acid used. 

 The values for amyl acetate and methyl, ethyl, and amyl valerates, 

 being deduced from one determination of R 1 only in each case, 

 can only be regarded as approximations. Of these, the values of 

 amyl acetate and methyl valerate are probably the most incorrect, 

 the true values being probably near 114*4 and 1 10*0 respectively 

 (§ 39) : when these corrections are made, and if 105*0 be taken 

 as the value for methyl butyrate instead of 115*1, all the excep- 

 tions disappear. 



