39Q 



NA TURE 



[September 16, 1922 



it is true, is contradictory, hut, such as it is, it seems 

 to show a balance against the existence of the displace- 

 ment. The extreme difficulty and complexity of the 

 experimental work must, nevertheless, be borne in 

 mind. Perhaps it is scarcely possible, in the present 

 state of our knowledge and experimental equipment, 

 to obtain a definite solution of the problem. The 

 third test — concerning the deflection of light in a 

 gravitational field — accordingly becomes of very con- 

 siderable importance, and to many minds constitutes 

 the deciding factor in their judgment of the theory. 



Consequently, the chief item in the programme 

 of the Royal Astronomical Society's expedition to 

 Christmas Island, on the occasion of the total solar 

 eclipse of September 21, will be the investigation of 

 this particular problem. It will be remembered that 

 the original test, on May 29, 1919, was considered by 

 the observers and a large number of others to give 

 conclusive evidence in favour of the relativity theory : 

 it was this result, in fact, that directed general atten- 

 tion to the theory, and made Einstein, for a brief 

 spell, a noteworthy figure in public esteem. The 

 interpretation of the observations, however, has been 

 subjected to various criticisms. A refracting atmo- 

 sphere of the sun has been proposed. Attempts have 

 been made to explain the effect as a result of terrestrial 

 atmospheric refraction arising from a temperature 

 gradient across the boundary of the moon's shadow- 

 cone. Still more serious is the evidence of the mutual 

 displacement of adjacent photographic images : the 

 question arises whether the observed positions of the 

 star images might not be, to some extent, dependent 

 on the intensity of the coronal light. It must be 

 admitted that the criticisms have been well met. 

 Nevertheless, so fundamentally important a matter 

 can scarcely be regarded as finally settled by a single 

 set of observations, and the repetition which is about 

 to take place is anything but a superfluous confirmation 

 of previous knowledge. 



It will be opportune at this time to recall the nature 

 of the problem, and see wherein lies the difference 

 between the traditional and the relativity conceptions 

 which makes the prospective test possible. From the 

 time of Newton until quite recently, gravitation has 

 been looked upon as an essential property of matter — 

 as characteristic as the property of inertia. Whenever 

 we find matter showing the unmistakable effects of 

 inertia, we find also evidence of gravitational influence. 

 The universality of these twin phenomena has so 

 impressed physicists that they have come to look upon 

 them as the fundamental properties of matter. Matter 

 is, by definition, that which has inertia and exerts 

 gravitational attraction. Nevertheless, it is well to 

 point out that inertia and gravitation are not the 

 properties by which matter is generally recognised. 

 With regard to most of the matter in the universe, 

 there is no evidence that they exist. We announce the 

 presence of matter when we see it : in other words, 

 matter everywhere has the power of emitting or 

 absorbing light — or, more generally, radiation. ' We 

 see light, and we deduce a star ; the light fluctuates, 

 and we deduce absorbing matter. If, according to the 

 true scientific method, we establish our fundamental 

 conceptions on the groundwork of pure observation, 

 we must place the power to radiate and absorb light 



NO. 2759, VOL. I 10] 



at least as deep down in the nature of matter as the 

 inertial-gravitational property. The recognition of 

 either property is universally accepted as evidence of 

 the existence of matter. 



The difference with which we are concerned between 

 the traditional and the relativistic conceptions may 

 be expressed in this way : that whereas the older view 

 gives no a priori indication of a relation between the 

 two fundamental material influences, gravitation' and 

 light, it is an essential condition of the relativity theory 

 that such a relation exists. A large and valuable 

 system of thought has been built up — mainly during 

 the last hundred years — in which radiation and gravita- 

 tion are completely independent. Radiation submits 

 to analysis and invites correlation with other physical 

 phenomena ; gravitation stands inaccessibly apart. 

 The complex organism of electromagnetism, embracing 

 as it does radiation, the aether, electricity, magnetism, 

 the atom — even inertia (for radiation possesses inertia) 

 — seems capable of assimilating the whole of physics — 

 except gravitation. Matter appears to be the source 

 of two streams of phenomena, one summarised in 

 electromagnetism and the other in gravitation, and 

 between them there is a great gulf fixed. The complete- 

 ness of the duality lies, of course, only in the con- 

 ceptions. Experimental evidence of a bridge across 

 the gulf might have arisen at any time. Standing now 

 on the bridge, it seems a little strange that it was not 

 sought before. Inertia, in submitting to the electro- 

 magnetic scheme, might carry with it the gravitational 

 property with which, in material bodies, it is always 

 associated, and the electromagnetic inertia of light 

 might be accompanied by a proportionate power to 

 exert and respond to gravitational influence. There 

 is no reason, according to pre-relativity physics, why 

 it should, but neither is there any reason why it should 

 not. The impotence of the electromagnetic theory 

 even to suggest the more probable of the alternatives 

 is its main defect. 



There is no such ambiguity in the utterances of the 

 relativity theory. Here gravitation — as a physical 

 existence giving rise to a gravitational field — is ignored : 

 the field alone is considered. The seat of the pheno- 

 menon is not sought in the secret nature of matter ; it 

 is sought in the space surrounding matter, and is, in 

 fact, regarded as a property of that space — or, rather, 

 space-time. The justification for this view is found 

 in the facts, first, that the evidence for the existence 

 of gravitation is the observed acceleration of one body 

 in the neighbourhood of another ; and, second, that the 

 acceleration produced by one body in another is in- 

 dependent of every property of the latter except its 

 position relative to the former. Now a phenomenon 

 manifesting itself as an acceleration (involving the 

 dimensions of space and time only) and producing 

 effects depending only on position, can be submitted 

 to a geometrical treatment, provided that the dimension 

 of time is added to those of space. Instead of speaking 

 of the curved paths of bodies in a homogeneous space- 

 time, we can speak of the straight paths of bodies in a 

 heterogeneous space-time. The same phenomenon 

 is indicated by both statements. Expressed in this 

 way, it seems as though there could be no difference 

 between the two views, except that one might be more 

 convenient than the other. Considered from a physical 



