September 6, 1900] 



NATURE 



455 



concerned, considered by themselves ; this, however, he would 

 allow to be satisfied indirectly, if the effects could be traced 

 across the intervening rether by stress, so that the tractions on 

 the two sides of each ideal interface are equal an! opposite.' 

 As above argued, this view appears to exclude ab initio all 

 atomic theories of the general type of vortex atoms, in which 

 the energy of the atom is distributed throughout the medium 

 instead of being concentrated in a nucleus ; and this remark 

 seems to go to the root of the question. On the other hand, 

 the position here asserted is that recent dynamical developments 

 have permitted the extension ot the principle of Action to 

 systems involving permanent motions, whether obvious or latent, 

 as put of their constitution ; that on this wider basis the atom 

 may itself involve a state of steady disturbance extending 

 through the medium, instead of being only a local structure act- 

 ing by push and pull. The possibilities of dynamical explana- 

 tion are thus enlarged. The most definite type of model yet 

 imagined of the physical interaction of atoms through the ;i^ther 

 is, perhaps, that which takes the a-ther to be a rotational ly 

 elastic medium after the manner of MacCullagh and Rankine, 

 and makes the ultimate atom include the nucleus of a permanent 

 rotational strain-configuration, which as a whole may be called 

 an electron. The question how far this is a legitimate and 

 effective model stands by itself, apart from the dynamics which 

 it illustrates ; like all representations it can only cover a limited 

 ground. For instance, it cannot claim to include the internal 

 structure of the nucleus of an atom or even of an electron ; for 

 purposes of physical theory that problem can be pat aside, it 

 may even be treated as inscrutable. All that is needed is a 

 postulate of free mobility of this nucleus through the tether. 

 This is definitely hypothetical, but it is not an unreasonable 

 postulate because a rotational i\;ther has the properties of a per- 

 tect fluid medium except wliere differentially rotational motions 

 are concerned, and so would not react on the motion of any 

 structure moving through it except after the manner of an ap- 

 parent change of inertia. It thus seems possible to hold that 

 such a model forms an allowable representation of the dynamical 

 activity of the a;thcr, as distinguished from the complete con- 

 stitution of the material nuclei between which that medium 

 establishes connection. 



At any rate, models of this nature have certainly been most 

 helpful in Maxwell's hands towards the effective intuitive grasp 

 of a scheme of relations as a whole, which might have proved 

 too complex for abstract unravelment in detail. When a physical 

 model of concealed dynamical processes has served this kind of 

 purpose, when its content has been explored and estimated, and 

 has become familiar through the introduction of new terms and 

 ideas, then the ladder by which we have ascended may be kicked 

 away, and the scheme of relations which the model embodied 

 can stand forth in severely abstract form. Indeed, many of the 

 most fruitful branches of abstract mathematical analysis itself 

 have owed their start in this way to concrete physical concep- 

 tions. This gradual transition into abstract statement of physical 

 relations in fact amounts to retaining the essentials of our work- 

 ing models while eliminating the accidental elements involved 

 in them ; elements of the latter kind must always be present 

 because otherwise the model would be identical with the thing 

 which it represents, whereas we cannot expect to mentally grasp 

 all aspects of the content of even theisimplest phenomena. Vet 

 the abstract standpoint is always attained through the concrete ; 

 and for purposes of instruction such models, jiroperly guarded, 

 do not perhaps ever lose their value : they are just as legitimate 

 aids as geometrical diagrams, and they have the same kind of 

 limitations. In Maxwell's words, "for the sake of persons of 

 these different types scientific truth should be presented in dif- 

 ferent forms, and should be regarded as equally scientific 

 whether it appear in the robust form and the vivid colouring of 

 a physical illustration, or in the tenuity and paleness of a sym- 

 bolical expression." The other side of the picture, the neces- 

 sary incompleteness of even our legitimate images and modes of 

 representation, comes out in the despairing opinion of Young 

 ("Chromatics," 1817), at a time when his faith in the un- 

 dulatory theory of light had been eclipsed by Malus's discovery 

 of the phenomena of polarisation by reflection, that this diffi- 

 culty "will probably long remain, to mortify the vanity of an 



J Cf. also Hertz on the electro-mapiietic equations, § 12, IVied. Ann.. 

 1890. The problem of merely repl.icing a system of forces hy a statical 

 stress is widely indeterminate, and therefore by itself unreal ; the actual 

 question is whether any such representation cin be co-ordinated with existinR 

 dynamics. 



NO. 16 10, VOL. 62] 



ambitious philosophy, completely unresolved by anythe:)ry": 

 not many years afterwards the mystery was solved by Fresnel. 



This process of removing the intellectual .scaffolding by which 

 our knowledge is reached, and preserving only the final formul;? 

 which express the correlations of the directly observable things, 

 may moreover readily be pushed too far. It asserts the concep- 

 tion that the universe is like an enclosed clock that is wound up 

 to go, and that accordingly we can observe that it is going, and 

 can see some of its more superficial movements, but not much 

 of them ; that thus, by patient observation and use of analogy, 

 we can compile, in merely tabular form, information as to the 

 manner in which it works and is likely to go on working, at any 

 rate for some time to come : but that any attempt to probe the 

 underlying connection is illusory or illegitimite. As a theo- 

 retical precept this is admirable. It minimises the danger of 

 our ignoring or forgetting the limitations of human faculty, 

 which can only utilise the imperfect representations that the 

 external world impresses on our senses. On the other hand such 

 a remainder has rarely been required by the master minds of 

 modern science, from Descartes and Newton onwards, what- 

 ever their theories may have been. Its danger as a dogma lies 

 in its application. Who is to decide, without risk of error, 

 what is essential fact and what is intellectual scaffolding? To 

 which class does the atomic theory of maf.er belong ? That is, 

 indeed, one of the intangible things which it is suggested may 

 be thrown overboard, in .sorting out and cUsiifying our scientific 

 possessions Is the mental idea or image, which suggests, and 

 alone can suggest, the experiment that adds to our concrete 

 knowledge, less real than the bare phenomenal uniformity which 

 it has revealed ? Is it not, perhaps, more real in that the uni- 

 formities might not have been there in the absence of the mini 

 to perceive them ? 



No time is now left for review of the methods of molecular 

 dynamics. Here our knowledge is entirely confined to steady 

 states of the molecular system : it is purely statical. In ordinary 

 statics and the dynamics of undisturbed steady notions, the 

 form of the energy function is the sufficient basis of the whole 

 subject. This method is extended to thermodynamics by 

 making use of the mechanically available energy of Rankine 

 and Kelvin, which is a function of the bodily configuration and 

 chemical constitution and temperature of the system, whose 

 value cannot under any circumstances spontaneously increase, 

 while it will diminish in any operation which is not reversible. 

 In the statics of systems in equilibrium or in steady motion, this 

 method of energy is a particular case of the method of Action ; 

 but in its extension to thermal statics it is made to include 

 chemical as well as configurational changes, and a new point 

 appears to arise. Whether we do or do not take it to be pos- 

 sible to trace the application of the principle of Action through- 

 out the process of chemical combination of two molecules, we 

 certainly here postulate that the static case of that principle, 

 which applies to steady systems, can be extended across chemical 

 combinations. The question is suggested whether extension 

 would also be valid to transformations which involve vital pro- 

 cesses. This seems to be still considered an open question by 

 the best authorities. If it be decided in the negative a distinc- 

 tion is involved between vital and merely chemical processes. 



It is now taken as established that vital activity cannot create 

 energy, at any rate in the long run, which is all that can from 

 the nature of the case be tested. It seems not unreasonable 

 to follow the analogy of chemical actions, and assert that it 

 cannot in the long run increase the mechanical availability of 

 energy — that is, considering the organism as an apparatus for 

 transforming energy without being itself in the long run 

 changed. But we cannot establish a Carnot cycle for a portion 

 of an organism, nor can we do so for a limited period of lime ; 

 there might be creation of availability accompanied by changes 

 in the organism itself, but compensated by destruction and the 

 inverse changes a long time afterwards. This amounts to as- 

 serting that where, as in a vital system or even in a simple mole- 

 cular combination, we are unable to trace or even assert complete 

 dynamical sequence, exact thermodynamic statements should 

 be mainly confined to the activity of the existing organism as a 

 whole ; it may transform inorganic material without change of 

 energy and without gain of availability, although any such 

 statements would be inappropriate and unmeaning as regards 

 the details of the processes that take place inside the organism 

 itself. 



In any case it would appear that there is small chance of 

 reducing these questions to direct dynamics ; we should rather 



