THE UNION OF OXYGEN WITH HEMOGLOBIN 121 



the adsorbed substance, is now impossible, since it appears proved by Langmuir's 

 work that there is no difficulty in accounting both qualitatively and quantitatively 

 for many cases of adsorption, by means of the already very familiar forces which 

 cause combination between metals and oxygen to form oxides, or the forces which 

 bring about solution. 



Obviously adsorption cannot be defined as "that which occurs at the surface of 

 a colloid"; since colloids themselves are not yet a well-defined class of substances, 

 and indeed the best studied cases of adsorption are at plane interfaces, not at the 

 surfaces of colloids. 



With the definition of adsorption proposed, a process would be excluded if, as with 

 haemoglobin and oxygen, combination occurs only at some defined locality on the 

 surface. Similarly, the ordinary reactions of organic chemistry wiU be excluded, as 

 they should be, since the substances taken up go to definite atomic groupings in the 

 molecule. The combination of oxygen with haemoglobin is seen to belong to the same 

 class as most organic reactions. 



It remains to examine whether the definition is practically applicable to known 

 cases of adsorption, as well as theoretically justified; and whether, in the case of 

 oxygen and haemoglobin, the arguments originally put forward in support of the 

 adsorption process are cogent enough to override the definition. 



All cases of adsorption, from a gaseous phase, or from solution, on plane, or nearly 

 plane, interfaces, are obviously compatible with the definition, since the common 

 method of calculating the amount of adsorption assumes uniformity of distribution 

 on the surface, and the results are generally expressed per sq. cm. of interface. 



In the cases of adsorption on colloidal surfaces, when the extent of surface is usually 

 not known, and the adsorption is expressed per gram of adsorbent, the definition is 

 probably ^.Iso applicable. Mecklenburg (Zeitsch. f. physikal. Chemie, vol. Lxxxiii. 

 p. 622) described experiments showing in several cases that the adsorption on different 

 specimens of the same adsorbent, prepared, however, under different conditions, 

 varied in a precisely similar way with concentration for each adsorbent, but the total 

 amount adsorbed per gram was propori;ional to a factor in each case, this factor being 

 presumably proportional to the area of the adsorbent. 



In proposing the theory that the oxygen in oxyhaemoglobin is held by adsorption, 

 Wo. Ostwald (KoU. Zeitsch. vol. n. pp. 264, 294) based the argument on two supposed 

 facts: first, that no definite saturation point of oxygen with haemoglobin could be 

 found, a fact now shown to be incorrect; and, second, that the amount of oxygen 

 taken up at different pressures could be fairly accurately represented, under certain 

 conditions, by the so-called "adsorption isotherm," y = kc^ (y = amount taken up, 

 c = concentration of oxygen). 



The mere fact that the variation of the amount taken up fits the "adsorption 

 isotherm" does not seem now to be a sufficient ground for classing a process as 

 adsorption. The "isotherm" has, until quite recently, been an empirical fact without 

 theoretical explanation; and not only does it contain two independent arbitrary 

 constants, which makes the fitting of a set of experimental data easier than would be 

 the case otherwise; but also it is, at the best, usually only accurate at low con- 

 centrations, divergencies being found at higher concentrations. 



A more accurate equation relating amount adsorbed to concentration has been 

 deduced recently (Henry, Phil. Mag. vol. xliv. p. 689, 1922) on the assumptions of 



