Oct. 5 1876] 



NA TURE 



513 



to be practically impermeable to gas by either of the two modes 

 of passage just described, is readily penetrated by the agency of 

 the molecular or diffusive movement. The times of passage 

 through a graphite plate into a vacuum have no relation to the 

 capillary transpiration times of the gases, but they show a close 

 relation to the square roots of the densities of the respective 

 gases, and agree with the theoretical times of diffusion usually 

 ascribed to the same gases. 



These latter results were obtained by the graphite diffusio- 

 meter of which a sketch is given (Fig. i). It stood over mercury, 

 and was raised or lowered by an arrangement introduced by 

 Prof. Bunsen. 



Mr. Graham subsequently employed the barometrical diffusio- 

 meter shown in Fig. 2. It consists of a tube in which a Torri- 

 cellian vacuum could be produced. The upper end was closed 

 by the porous septum, and a slow stream of the gas under exa- 

 mination was allowed to pass over the plate through the india- 

 rubber hood by which it was covered. 



I might mention that the very exact and illustrious experi- 

 menter, Prof. Bunsen, was led to doubt the accuracy of Graham's 

 law of ihe diffusion of gases, but he employed plugs of plaster 

 of Paris which impaired the results by introducing the pheno- 

 menon of transpiration ; and probably also as Mr. Graham 

 observed to me, by an actual retention of hydrogen in the pores 

 of the plaster It is interesting from our point of view, because 

 it shows that the simple apparatus employed by Mr. Graham 

 really gave the only trustworthy results. 



The results of the later experiments led him to prove that 

 mixed gases might be separated from each other by diffusion. 

 Stems of tobacco-pipes were employed, arranged inside a glass 

 tube, which could be rendered vacuous, the mixed gases being 

 passed through the tobacco-pipe. For example, when this ex- 

 plosive mixture of 66 per cent, of hydrogen, and 33 per cent, of 

 oxygen is passed through this tube (Fig. 3) a mixture is obtained 

 containing only 93 per cent, hydrogen, and is therefore non- 

 explosive. With air it was found possible to concentrate the 

 oxygen by 3 5 per cent. 



With the apparatus now before us (Fig. 4) Mr. Graham subse- 

 quently worked on liquid transpiration in relation to chemical 

 composition. He started from the discovery of M. Poiseuille, that 

 a definite hydrate of one equivalent of alcohol with six equivalents 

 of water is more retarded than alcohol, containing either a greater 

 or a smaller proportion of water. The rate of transpiration de- 

 pending upon chemical composition and affording an indication 

 of it, it thus appeared probable that a new physical property might 

 become available for the determination of the chemical constitu- 

 tion of substances, and the experiments appeared to establish 

 " the existence of a relation between the transpirability of liquids 

 and their chemical composition. It is a relation analogous in 

 character to that subsisting between the boiling point and com- 

 position so well defined by Hermann Kopp." ^ The apparatus 

 consists of a strong glass jar closed at the top by a biass plate 

 into which a condensmg syringe is screwed. This plate also had 

 a tube screwed into it, and into the tube the glass bulb with a long 

 capillary tube was fixed. The fluid under examination was 

 placed in the bulb, which communicated freely with the interior 

 of the jar, containing compressed a-r. 



To revert to the chronological order. His next paper in De- 

 cember, 1849, formed the Bakerian lecture of the Royal Society. 

 It was on the Diffusion of liquids, and the only apparatus em- 

 ployed was very similar to that adopted in his earliest paper on 

 the diffusion of gases ; it consisted of a bo'.tle and glass jar (Fig. 

 5), the fluid under examination being placed in the bottle, which 

 was immersed in the water with which the jar was filled. With 

 this simple apparatus he found that when two liquids of different 

 densities, and capable of mixing, are placed in contact, diffusion 

 takes place between them much in the same manner as between 

 gases, except that the rate cf diffusion, which varies with the 

 nature of the liquids, the temperature and the degree of concen- 

 tration is slower. Common salt when placed in the inner vessel 

 will diffuse twice as rapidly as sulphate of magnesia, and this 

 salt will diffuse twice as rapidly as gum arable. Subsequently 

 Mr. Graham modified thedisposiiionof the apparatus and simply 

 introduced the salt to be diffused by means of a pipette to the 

 bottom of a jar filled with water. These experiments led to the 

 very remarkable and important discovery that different com- 

 pounds might be separated from each other by diffusion, and 

 this was not all, for it was proved that a partial decomposition 

 of chemical compounds was effected by diff jsion. Thus ordinary 

 alum was partially decomposed into sulphate of potassium and 

 ' Phil. Trans., 1861, p. 373. 



sulphate of aluminium, which is less diffusible than the first- 

 named salt. Mr. Graham considered this research to be very 

 important, and he remarks, "in liquid diffusion we appear to 

 deal no longer with chemical equivalents or Daltonian atoms, 

 but with masses even more simply related to each other in 

 weight." We may suppose that the chemical atoms "can group 

 together in weights which appear to have a simple relation to 

 each other. It is this new class of molecules which appear to 

 play a part in solubility and liquid diffusion, and not the atoms 

 of chemical combination." 



Continuing the investigation he described in a paper of singular 

 beauty, his well-known experiments on the varying rates of liquid 

 diffusion of various soluble substances, which led him to divide 

 them into crystalloids and colloids, the former having a rapid 

 diffusion rate, the latter being marked by low diffusibility. He 

 placed the substance under experiment in a tambourine of parch- 

 ment paper (Fig. 6) which was floated on the surface of a com- 

 paratively large volume of water, the highly diffusive crystalloid 

 passed through the membrane, the colloid remained behind, for 

 "the diffusion of a crystalloid appears to proceed even through 

 a firm jelly with little or no abatement of velocity." 



I have here the very interesting series of colloids prepared by 

 Mr. Graham, and of these perhaps the most interesting is the 

 soluble silicic acid. If silicate of soda is poured into diluted 

 hydrochloric acid, the acid being maintained in large excess, a 

 solution of silicic aciJ is obtained. But this solution also con- 

 tains, in addition to the silicic acid, chloride of sodium, from 

 which it may be freed by the action of dialysis, and by this 

 means a solution, which is not in the leist viscous, is obtained, 

 containing 14 per cent of silicic acid. The coagulation of the 

 silicic acid is effected, however, by the addition of a solution 

 containing the iTToirth part of any alkaline or earthy carbonate. 

 Mr. Graham therefore described this gelatinous state as the 

 "pectous," as distinguished from the " peptous " or dissolved 

 form. 



By a similar process Mr. Graham obtained specimens of 

 soluble alumina, peroxide of iron, chromic oxide, and stannic 

 acid, all of which have their pectous and peptous states. And 

 he showed that in most cases alcohol, sulphuric acid, and 

 glycerine can replace part of the water of these colloids. I 

 cannot describe these interesting substances now, nor can I do 

 more than remind you of the use of dialysis iu medico-legal 

 inquiries. I must content myself with summing up a few of 

 Mr. Graham's conclusions with reference to crystalloids and col- 

 loids. Although chemically inert, in the ordinary sense, colloids 

 possess a compensating activity of their own, arising out of their 

 physical properties. While the rigidity of the crystalline struc- 

 ture shuts out external impressions, the softness of the gelatinous 

 colloid partakes of fluidity, and enables the colloid to become 

 a medium for liquid diffusion like water itself. Another and 

 eminently characteristic quality of colloids is their mutability, as 

 fluid colloids often pass from the fluid to the pectous or gela- 

 tinous condition under the slightest influences. The colloid is, 

 in fact, the dynamic state of matter, the crystalloid being the 

 statical condition. The colloid possesses energy, and it may be 

 looked upon as the primary source of the force appearing in the 

 phenomena of vitality." 



The next instruments to be considered are those with which Mr. 

 Graham studied osmotic force. When a solution of a salt, or a 

 liquid, is separated by a membrane or porous diaphragm from a 

 mass of water, a flow of liquid takes place from one side of the 

 septum to the other. This action was discovered by Dutrochet, 

 and is known as osmose. Dutrochet and Mr. Graham both 

 used a narrow glass tube, having a funnel-shaped expansion at 

 the bottom, covered at that end by a piece of bladder (Fig. 7). 

 Mr. Graham also used porous earthenware and albuminated 

 calico. 



In some cases the flow of liquid into the bulb is sufficiently 

 powerful to sustain a column of water many inches high in the 

 glass tube. Dutrochet inferred from his experiments that the 

 velocity of the osmotic current is proportional to the quantity of 

 salt or other substance originally contained in the solutior^. 

 He attributed the action of the septum to capillarity, but 

 Mr. Graham ultimately considered that the water movement in 

 osmose is "an affair of hydration and dehydration of the sub- 

 stance of the membrane or other colloid septum," and that the 

 diffusion of a saline solution only acts by affecting the hydration 

 of the septum. The outer surface of the membrane being in 

 contact with pure water, tends to hydrate itself in a higher 

 degree than the inner surface does, the latter surface being sup- 

 posed to be in contact with a saline solution. When the full 



