and the Surface- Tension of Solutions. 485 



and FJ2 due to metal and acid respectively, with the values 

 2-7 for Mg, 1-7 for Zn, and 4"1 for S0 4 '. Whether this 

 assumption is true has still to be decided ; it is made here 

 only to see to what results it leads, and as it gives the ratios 

 PM 2 (M 2 /^ 2 )—«H-(M 2 %)* with values 18*7 and 23'4, not far 

 from the 19 characteristic of the types MgCl 2 and Na 2 S0 4 , 

 we have a certain amount of encouragement for believing our 

 assumption to be true, but our values of the ratio are not 

 reliable enough to make us sure. As regards 



c(p,V)M 2 (M 2 /p 2 )-^(M 2 2 Z 2 )i, 



the smallness of c causes it to be unreliable. If we made 

 the guess that as for the types NaCl and MgCJ 2 the ratio is 

 8*5, and for the type Na 2 S0 4 it is 4'4, so for the type MgS0 4 

 it is 2*2, we should not involve ourselves in an error greater 

 than the experimental ones, but the final decision of the point 

 must rest with special experiments. It is obvious that we 

 can hardly draw any reliable inferences from experiments on 

 the surface-tension of dilute solutions of compounds of the 

 type ZnS0 4 unless they are of a very high degree of accuracy, 

 because c being small and p 2 2 small, the value of ci/p 2 — a 1 /p^ 2 

 is small. 



When we come to treat of solutions of organic compounds 

 we have to separate them into two classes, the exceptional and 

 the normal. Traube's experiments (Ber. cL deutsch. Chem. 

 Ges. xvii.; Ann. der Chem. cclxv.) show that mixtures of water 

 with the alcohols and fatty acids are exceptional, the most 

 pronouncedly exceptional case being that of water and amyl 

 alcohol, for 2*5 per cent, of amyl alcohol in water reduces the 

 surface-tension to 2'8, that for the pure alcohol being 2'4 and 

 for water 7*3. The probable cause for the exceptional be- 

 haviour of these mixtures is a difference of composition in the 

 surface-layers from that in the body of the liquid. Certain 

 other solutions of organic compounds appear to behave nor- 

 mally as regards surface-tension, and for them Traube's data 

 (Journ. f. prak. Chem. cxxxix.) are given in the following 

 table. The values of (M 2 2 / 2 ) * have not been directly deter- 

 mined for any of these substances, but they can be estimated 

 according to the values given for the different atoms and 

 radicals on pp. 4 and 7 of " Further Studies," although on 

 account of the exceptional nature of the alcohols and fatty 

 acids it has not been possible hitherto to get a satisfactory 

 value for the characteristic OH group, but if provisionally we 

 take 1*3 as the increase of (M 2 2 / 2 )% when H is replaced by 

 OH, then we can obtain values of (M 2 2 / 2 ) 5 in the following 

 manner. For oxalic acid (COOH) 2 we start from 2 H 6 with 



