t!0Lt.0lt) CtlEMlS'fllY. 200 



by ahother ion of tlie same sign by washing with the appropriate salt.' 

 In each case an equivalent quantity of the non-precipitating ion is set 

 free, and in the case of the stronger acids and bases may be estimated in 

 the residual liquid by titration. These phenomena are not confined to 

 inorganic colloids, but are very marked in many organic colloid precipita' 

 tions, and apparently the same laws hold good in many pseudo-solutions, 

 such as that of clay or of gum mastic in water, which are generally dis- 

 tinguished as mere mechanical suspensions. The phenomenon of the 

 retention of a portion of the precipitating ion with the precipitate is well 

 known in many ionic reactions, and is usually termed ' co-precipitation ' 

 or ' adsorption.' It is in fact common to all bodies of largely extended 

 surfaces and to jellies, though the question whether in the last case the 

 action is surface, or extends throughout the substance, must be discussed 

 later. Even with solids, this is still, to a certain extent, an open ques- 

 tion. Davis ^ has shown that in the absorption of iodine by carbon, an 

 equilibrium is established with the surface action in a few hours, but that 

 a further absorption goes on for weeks or months, which he considers due 

 to solid solution. The quantity of absorbed or adsorbed substance varies 

 with the concentration of the solution with which it is in equilibrium, 

 relatively more being taken from weak than from stronger solutions. 

 The laws by which it is governed show close relations to those of solu- 

 tion-equilibria, and the phenomenon might perhaps be not improperly 

 described as surface-solution. If a body be brought in contact with two 

 immiscible solvents, it distributes itself between them in a ratio which, 

 so long as its molecular condition is the same in both solvents, is, in 

 accordance with Henry's law, a constant fraction dependent on its rela- 

 tive solution-pressure in the two solvents ; or if C„ and C(, be its concentra- 

 tion in the solvents a and b and ft a constant, C^,=/3Ci,. If, however, its 

 molecular weight in the solvent b is n times as great as in a, the equation 



takes the form C„=/3C(,". Written in a general form, where x is the 

 weight of the absorbed substance and ??i that of the absorbent, while C(, 



X -* 



is the concentration of the solution, the equation becomes — =Ca=/3C4 '',' 

 where /3 and p are constants to be determined experimentally, and repre- 

 sents very approximately most adsorption-equilibria ; but it is evident 

 that p cannot represent the molecular weight in the ordinary sense, since 

 it is seldom a whole number, and usually indicates greater complexity in 

 the solvent than in the adsorbent, which is impossible, when, as in the 

 case of most electrolytes in very dilute solutions, the dissolved body is 



' Lindcr and Picton, Jour. Chem. Sue, 189.3, C7, C3. 

 " Trans. Ok. Soc, 1907,91, 1666-1683. 



' The exponential character of a curve of experimental results is easily tested 

 by plotting the logs, either natural or common. If the logs of the equation 



C„ = fiCt' be taken, we obtain log C„^log /3 + ^log C,, x -V Hence, if we plot the 



experimental logarithmic concentrations of the solution as abscisstB and those of 

 absorbed substance as ordinates, we shall, if the reaction is rigidly exponential, obtain 



a straight line, of which the slope (tangent) is ~, and which cuts the vertical line, 



passing through the origin log C't, = (or Ci,= ]) at log /8. It must not be forgotten 

 in plotting logarithms that in fractional numbers the indices are -, while the 

 mantissajare always +. Exponential curves rise rapidly at first, but afterwards 

 much more slowly and approximate to a straight line. 



1908. p 



