54 



NA TURE 



[May 17, 1906 



of osmotic pressures in plants and animals, chemo- 

 taxis, the theory of ionisation and its application to 

 the germicidal action of disinfectants, the permea- 

 bility of membranes and the influence of this on secre- 

 tion, the velocity of reactions, catalysis, colloidal solu- 

 tions, and the bearing of physical chemistry on serum 

 therapy, in which connection the work of Ehrlich, 

 Arrheriius, and Madsen is briefly reviewed. Alto- 

 gether this book supplies a decided want, and can be 

 thoroughly recommended. 



LETTERS TO THE EDITOR. 

 [The Editor does not hold himself responsible for opinions 

 expressed by his correspondents. Neither can he undertake 

 to return, or to correspond with the writers of, rejected 

 manuscripts intended for this or any other part of Nature. 

 No notice is taken of anonymous communications.] 



Osmotic Pressure. 



'■ In the issue of N.4TURE for May 3 (p. 19) appeared an 

 abstract of a recent paper by Prof. Kahlcnberg on 

 "Osmosis and Osmotic Pressure." In Prof. Kahlenberg's 

 paper, and also in the abstract, it is claimed that his ex- 

 periments invalidate van 't Hoff's theory of osmotic pres- 

 sure, by which the concordance between the pressure of 

 gases and the osmotic pressure of dilute solutions was 

 established. As the basis of that theory seems sometimes 

 to be misunderstood, may I be allowed to recall the prin- 

 ciples on which it is founded? 



In a paper published in the Zeitschrift fiir physikalische 

 Chcmie for 1887, van 't Hoff showed that, from the well- 

 known experimental relation between the solubility of a gas 

 and the pressure, it followed by a simple application of the 

 second law of thermodynamics that the osmotic pressure 

 of a dilute solution must possess the same value as the 

 ordinary pressure of a gas at the same concentration. The 

 solution must be so dilute that the dissolved systems, each 

 made up of a particle of solute as nucleus, and the portion 

 of solvent which it influences, are beyond each others' 

 spheres of action. The proof has been put in a modified 

 form by Lord Rayleigh (Nature, 1897), and Prof. Larmor 

 has obtained the same result by using the fundamental 

 conceptions of the molecular theory as a basis, instead of 

 the experimental solubility relations of a gas (Phil. Trans., 

 A, 1S97). In all these proofs no assumption is made as to 

 the nature of osmotic pressure. It may be due to molecular 

 impacts or to chemical affinity, or to some other undis- 

 covered cause. The strength (and weakness) of a thermo- 

 dynamic proof lies in this very independence of assumptions 

 as to the mechanism by which the effects are produced. 

 Prof. Kahlenberg and his followers seem to consider that 

 the thermodynamic theory of solutions stands or falls with 

 the hypothesis that the pressure is due to molecular bom- 

 bardment. 



If the conditions assumed in the proofs are realised, the 

 whole authority of thermodynamics goes to support the 

 result. The importance of experiments on osmotic pressure, 

 such as those of Prof. Pfeffer, Lord Berkeley and Mr. 

 Hartley, and Prof. Kahlenberg, lies in the question how 

 far the assumptions made in the thermodynamic proofs can 

 be realised experimentally. This is a much humbler r6]c 

 than that assigned to the experiments by Prof. Kahlen- 

 berg, who claims that the application of gas laws to solu- 

 tions is based on the few observations of Pfeffer and others 

 by which those laws have been verified directly. Never- 

 theless, the experiments are of great interest. The gas 

 value for the osmotic pressures measured by Pfeffer shows 

 that the conditions laid down in the thermodynamic theory 

 are realised in practice : (i) that for sugar solutions in water 

 an approximately perfect semi-permeable membrane has 

 been obtained ; (2) that no selective action such as could be 

 produced by a Maxwellian dremon is in operation ; (3) that 

 the molecules of cane sugar in solution are the simple mole- 

 cules indicated by the chemical formula, though they may 

 or mav not be combined with solvent molecules ; (4) that 

 a solution which is dilute in the thermodynamic sense can 



be realised at possible concentrations ; (5) that a theory 

 deduced for volatile solutes may be extended to other cases. 

 When other solutions and different membranes are 

 employed, one or more of these conditions may fail, and 

 the theoretical value be beyond the reach of e.xperimental 

 attainment. Prof. Kahlenberg remarks that because a 

 semi-permeable membrane does not exist, a theory which 

 postulates one cannot be maintained. We might construct 

 a parallel statement by saying that because a frictionless 

 piston is not practically obtainable, in Carnot's engine and 

 the science of reversible thermodynamics physicists and 

 engineers have imagined a vain thing. 



But I may point out that at least two perfect semi- 

 permeable surfaces are probably known : (i) when a solu- 

 tion freezes to give the solid of the pure solvent, the solute 

 is compressed into a smaller volume of liquid solution ; the 

 surface of the growing crystals is semi-permeable. 

 (2) When a volatile solvent evaporates from the solution 

 of a non-volatile solute, the free surface of the liquid is 

 again a semi-permeable membrane. From these two factj 

 follows the validity of the thermodynamic relations between 

 the osmotic pressure on the one side and the freezing point 

 and vapour pressure on the other. This is important, for 

 it enables us to use measurements of freezing points or 

 vapour pressures when it is not possible to realise the 

 experimental conditions necessary for a satisfactory deter- 

 mination of the true osmotic pressure. 



Osmotic pressure is a thermodynamic conception. The 

 pressures observed in practice may or may not represent 

 the same thing. We may define osmotic pressure as the 

 excess of hydrostatic pressure it is necessary to exert on 

 a .solution in order that it may be in equilibrium with the 

 solvent through a perfect semi-permeable membrane. 

 With this definition we may use the conception of osmotic 

 pressure as a basis for a Carnot's cycle and a thermo- 

 dynamic theory of solutions. Prof. Kahlenberg writes that 

 opponents of van 't Hoff's idea have generally held that 

 the so-called osmotic pressure is an ordinary hydrostatic 

 pressure, brought about by the entrance of liquid into the 

 osmotic cell. It is delightful to find one point at least in 

 which the supporters of van 't Hoff, and van 't Hoff him- 

 self, are in complete agreement with his opponents. 



In the abstract of -Prof. Kahlenberg's paper which 

 appeared in Nature we are warned that, among the general 

 ruin of physical theories which is to follow his experi- 

 ments, the hypothesis of ionic dissociation is involved. I 

 confess that the warning leaves me unmoved. The idea 

 that the ions of electrolytic solutions are dissociated from 

 each other during their movement (though possibly or 

 probably combined with the solvent) is required by the 

 electrical phenomena. The abnormally great osmotic 

 pressures of certain electrolytes dissolved in water indi- 

 cate some kind of dissociation, but cannot tell us whether 

 or not that dissociation takes place so as to give rise to 

 electrified systems. In simple salts such as potassium 

 chloride, which we know by their electrical properties to be 

 electrically dissociated, it is difficult to see how a second 

 kind of simultaneous dissociation could occur. But that 

 non-electrical separation is sometimes found is indicated 

 by some older experiments of Prof. Kahlenberg himself, 

 who found that solutions of diphenylamine in methvl 

 cyanide show abnormally low molecular weights, but are 

 non-conductors of electricity. The theory of ionic dis- 

 sociation rests upon electrical evidence, and by such evidence 

 it must be tried. W. C. D. Whetham. 



Trinity College, Cambridge, May 12. 



Considerable importance seems to be attached to a 

 recent paper by Prof. Kahlenberg on " Osmosis and 

 Osmotic Pressures " (Jour. Phys. Chem., vol. x.), as is 

 evidenced by a separate summary published in Nature 

 (May 3, p. 19). In these circumstances it may not be out 

 of place to point out that the conclusions Prof. Kahlen- 

 berg deduces are not warranted. 



On p. 142 he says " indirect measurements of osmotic 

 pressures . . . from vapour tensions . . . involves the 

 assumption that the gas laws hold for solutions." This is 

 contrary to fact. We have shown experimentally (see 

 vol. Ixxvii. Proc. Roy. Soc.) that aqueous solutions of cane 

 sugar give the same osmotic pressure whether observed 



NO. 1907, VOL. 74] 



