'94 



NA TURE 



[June 20, 1901 



going on, and to show you that they have the exact shape which 

 the theory requires. 



I have here a rectangular glass cell divided horizontally by a 

 ihin plate of celluloid having a circular hole punched through 

 it. The lower half of the cell is filled with a solution of gelatine 

 containing a little barium chloride, and the upper half with a 

 solution of sodium sulphate. 



The relative strengths of the solutions are so adjusted that 

 the two salts, diffusing in opposite directions, shall meet some- 

 where in the gelatine where a precipitate of barium sulphate 

 is thrown down at the surfaces of contact of the two opposing 

 streams of diffusion. The result is that we get a slowly growing, 

 spheroidal mass of precipitate, starting from the aperture and 

 resembling in shape the head of an inverted mushroom. 



If we arrange for the diftusion of the sodium sulphate to be 

 intermittent, or, better still, if we alternate the diffusion of a 

 sulphate with that of a chromate, we get well-marked zoiiings 

 in the precipitate forming the spheroid, zonings which corre- 

 spond to the successive forms which the spheroid has assumed 

 during growth, and which, therefore, must have been zones of 

 equal density of the difftising substances. We can study the 

 forms which these assume in relation to the aperture by subse- 

 quently cutting sections through the gelatine, but by a little 

 arrangement we can make the apparatus cut its own sections as 

 the diffusion goes on. 



This is done by making the aperture in the diaphragm soni- 



series of wavy lines, which become more and more horizontal 

 as the distance gets more remote. 



Could they be rendered visible these are also the forms which 

 we should expect the lines of equal density of a substance to take 

 when it is difi'using through a series of small apertures. I am 



circular instead of circular, and bringing its 

 straight edge close up to the side of the glass 

 vessel. 



I will now throw on the screen some photo- 

 graphs of vertical sections of spheroids of difiu- 

 sion of this kind. (.See Fig. 3 and Fig. 4). 



On comparing the lines of equal density around 

 the aperture with the diagrams on the wall you 

 will at once see that their shape is exactly that required by 

 theory — they describe a series of ellipses having their common 

 foci in the edges of the aperture through which the diffusion is 

 taking place. 



The actual stream lines of the diffusing substance are not 

 visible, but as these must necessarily be normal to the curves of 

 equal density they can only 

 be represented by a series 

 of hyperbolas, also having 

 their foci in the edges of 

 the aperture. 



" The electrostatic analogy 

 which has served us so well 

 in determining the form of 

 the zones of equal density 

 around single apertures 

 may also be used for pre- 

 dicting their distribution 

 around a series of apertures 

 in a diaphragm. 



If we regard the indivi- 

 dual holes in a multiper- 



forate diaphragm as so "^' 



many minute discs, all 



electrified to a common potential, the lines of equi-potential 

 and the lines of force should take a form something like that 

 represented in the diagram, Fig. 5, the lines of equi-potential 

 forming complete ellipses in the immediate neighbourhood of 

 the electrified discs, but gradually intersecting and forming a 



NO. 165 1, VOL. 64] 



able to give you a verification of this by throwing on the screen 

 photographs showing the result of inleimitient diffusion through 

 a series of such apertures (Figs. 6 and 7). The lines of equal 

 density are marked out by the alternate bands of sulphate and 

 chromate of barium as they 

 w ere in the last experiment. 

 From the shape of these 

 lines of equal density it is 

 possible to determine the 

 form of the stream lines of 

 tlie diffusing substance and 

 tci show that the tendency of 

 11 multi-perforate septum of 

 this kind is to locally in- 

 crease the gradient of density 

 in its neighbourhood and so 

 to accelerate the flow through 

 the small apertures. We get, 

 in fact, a complete and satis- 

 factory explanation of the 

 small amount of obstruction 

 which such a diaphragm 

 produces when put in the 

 way of a diffusive flow of 

 gas or liquid. 



Intermittent diffusion such 



as I have described may be 



^ '^- *■ used to illustrate in a variety 



of ways the distribution of 



electric potential around electrified bodies which are within the 



sphere of each other's action. 



It is generally a difficult and laborious task to work out the 

 distribution of the surfaces of equi-potential around electrified 

 bodies which are near enough to influence each other. By this 

 systein of intenniitent diffu>ion we may sometimes make Nature 



work out the problem for us. Here, for instance (see Fig. 8), is 

 a figure copied from Clark Maxwell's "Electricity and Mag- 

 netism," representing the form which is assumed by equi-potential 

 surfaces around two points charged with quantities of electricity 

 of the same kind in the tratio of 4 to i. If the analogy is 



