January 30, 1896] 



NATURE 



297 



Helmholtz made the great discovery that, by virtue of 

 their vorticity, vortex rings floating in a perfect fluid are 

 unable to destroy or create one another ; although these 

 vortices may distort each other, becoming drawn out into 

 thin threads or rolled into spherical balls, one cannot 

 destroy another. This discovery it was that afforded a 

 basis for those speculations of Lord Kelvin which would 

 identify atoms with vortex rings moving in a perfect 

 fluid ; the indestructibility of atoms finds a parallel in 

 the permanency of vortex rings, and the two have many 

 properties in common. As, however, our knowledge of 

 vortices has increased, so obstacles to the acceptance of 

 the atomic vortex hypothesis have arisen. Thus the 

 energy and the inertia of vortex rings increase together 

 whilst their rate of motion decreases, so that on raising 

 ihe temperature of a gas composed of vortex atoms, and 

 therefore increasing the rate of motion of its particles, 

 it would seem that, in some mysterious way, more energy 

 leaves the gas than enters it. Similarly, unless the weight 

 of a body alters appreciably as its temperature changes, 

 it is not easy to see how the simple vortex theory of matter 

 can be true ; the difficulty of determining weights at 

 different temperatures of course stands in the way of an 

 experimental examination of this point. Many modifica- 

 tions of the vortex theory have been proposed, but the 

 only statement that can be made with certainty is that 

 the space between the atoms, whatever their nature may 

 be, must be filled with some complicated structure, the 

 postulation of which is essential for the explanation of 

 electro-magnetic actions. It is therefore impossible to 

 believe that atoms are simply thin vortices floating in an 

 otherwise motionless and structureless medium. 



A curious analogy is noticeable between the stability of 

 vortex systems and chemical valency. A system of two 

 vortex rings, both rotating in the same direction, assumes 

 a state of fairly stable equilibrium in which the two rotate 

 round one another, whilst a system of three vortex rings 

 is stable in a state in which the vortices are situate at the 

 apices of a triangle. Similarly, a condition of stable 

 equilibrium is possible for systems of four, five, or six 

 rings ; a system of seven vortex rings, however, is un- 

 stable, and vortex systems generally become unstable 

 when composed of more than six rings. The curious 

 analogy between this result and the fact that the atom of 

 no chemical element requires to combine with more than 

 six monovalent atoms, should be kept in view in default 

 of a sounder dynamical conception respecting the 

 limitation of chemical bonds. 



The atomic vortex theory again meets with difficulties 

 in connection with homologous series of organic com- 

 pounds and with the atomic weights ; the atomic weight 

 of mercury is 200, that of hydrogen being unity, and it 

 can be shown that the volume occupied by the inercury 

 atom should be some 2800 times that occupied by the 

 atom of hydrogen, a result hardly reconcilable with the 

 known properties of these elements. Valency also pre- 

 sents obstacles to the theory ; thus nitrogen and carbon 

 should be respectively mono- and di-valent unless the 

 vortex rings are doubled on to themselves, and even' then 

 the doubling indicates the existence of two allotropic 

 modifications of carbon, a right- and a left-handed form, 

 for which no evidence exists. The vortex theory of 

 atoms and the experimental facts regarding atoms are 

 thus sadly at variance, and much still remains to be done 

 in clearing up the questions at issue. 



The theory of semipermeable membranes, which is of 

 such importance in certain branches of physical chemistry, 

 is as yet in a very unsatisfactory state. The absolute 

 disregard of any possible heating effects occurring during 

 osmosis may lead to serious errors, corresponding to those 

 which crept into the theory of galvanic cells by neglect of 

 the thermal effects which arise when electrical currents 

 enter or leave a liquid ; possible causes of error, such as 

 these, should be well borne in mind until the theory and 



NO. 1370, VOL. 53] 



practice of semipermeable cells are in better agreement 

 than at present. These semipermeable membranes are 

 frequently regarded as being only some kind of molecular 

 sieves, although they are really much more analogous to 

 Graham's second class of membranes, which only allow 

 the passage of gases soluble in the membrane itself ; the 

 laws governing the two kinds of membranes are quite 

 dissiinilar. It is not easy to sharply distinguish between 

 physical and chemical permeability when molecular 

 magnitudes are dealt with, and one molecule may pass 

 amongst others not so much by reason of possessing the 

 right size as the right shape. There seems some hope of 

 extending our methods of " chemical filtration " by means 

 of sets of properly constituted diaphragms, each of which 

 is penetrable by certain classes of molecular groups. 



The application of thermodynamics to chemical m- 

 vestigations is full of pitfalls ; the law of conservation of 

 energy has been often misapplied, and it is not sufficiently 

 realised that the second law of thermodynamics is not 

 strictly applicable to irreversible chemical changes, such 

 as explosions, &c. 



The tendency to regard chemical forces as electrical 

 ones is not altogether justifiable ; too many instances of 

 irreversible chemical changes exist to permit a parallel 

 between chemical actions and simple reversible 

 electrolysis. Chemical actions are of a far more complex 

 nature than simple electrolysis, and that other than 

 purely electrical forces are operative in solution is 

 indicated by Helmholtz's investigations of electrical 

 diffusion through fine tubes. No statical theory of solid 

 or liquid media which supposes the action of none but 

 electrical forces is possiljle, for such media would be 

 essentially unstable ; as far, then, as solids and liquids can 

 be conceived as statical systems, the postulation of other 

 than purely electrical forces is imperative. The success 

 which has attended the accepted theories of crystal 

 structure and of the asymmetric carbon atom, makes it 

 pretty safe to conclude that many properties of molecules 

 are deducible from purely statical theories of structure. 



The enormous increase of knowledge which has 

 attended the assumption that a substance in liquid 

 solution behaves in some important respects like the 

 same substance in a pure gaseous state, has led to the 

 grave error of supposing that the physical conditions of 

 molecules of a substance when gaseous and when dis- 

 solved are similar. A dissolved molecule is always 

 within the spheres of action of countless neighbours ; its 

 path is of the order of one-hundredth of its diameter, 

 and it receives, perhaps, 10^* blows per second, so that 

 its vibrations are comparable with those of radiant heat ; 

 in the gaseous state, however, the molecule has a free 

 path thousands of times its diameter in length. The 

 dynamical conditions of gaseous and dissolved molecules 

 are thus absolutely dissimilar. Although it is curious 

 that the osmotic pressure of a dissolved substance should 

 be even roughly identical with the vapour pressure of 

 the same quantity of the substance as a gas under 

 similar conditions of volume and temperature, it is wholly 

 erroneous to attribute this coincidence to a similarity 

 between the dynamical states in the two cases. Osmotic 

 pressure is more nearly related to Laplace's internal 

 pressure in a liquid, which depends on intramolecular 

 forces, than to a gaseous pressure which is practically 

 independent of the forces operating between the mole- 

 cules. Considerations respecting the capillarity and 

 vapour pressure of solutions and solvents shows that 

 some very close connection exists between osmotic 

 pressure and capillarity, and afford a trustworthy method 

 of applying thermodynamics to the calculation of osmotic 

 pressure. 



It is almost impossible to explain dynamically the 

 assumption that free electrically charged ions wander 

 about in a liquid in a condition at all rightly described 

 as one of dissociation. The term " dissociation " should 



