PEE 
TRANSACTIONS OF SECTION A. 621 
ends—that does not appear to me to lessen the present value of the hypothesis. 
We can recognise to the full how well it enables us to group together large masses 
of facts which, without it, would be scattered apart, how it serves to give 
working explanations, and continually enables investigators to think out new 
questions for research. We can recognise that it is the symbolical form in which 
much actual knowledge is cast. We might almost as well quarrel with the use of 
the letters of the alphabet, inasmuch as they are not the sounds themselves, but 
mere arbitrary symbols of the sounds. 
In this country there is no need for any defence of the use of the molecular 
hypothesis. But abroad the movement from the position in which hypothesis is 
confounded with observed truth has carried many through the position of equili- 
brium equally far on the other side, and a party has been formed which totally 
abstains from molecules as a protest against immoderate indulgence in their use. 
Time will show whether these protesters can do without any hypothesis, whether 
they can build without scaffolding or ladders. I fear that it is only an attempt to 
build from balloons. 
But the protest will have value if it will put us on our guard against using 
molecules and the ether everywhere and everywhen. There is, I think, some 
danger that we may get so accustomed to picturing everything in terms of these 
hypotheses that we may come to suppose that we have no firm basis for the facts 
of observation until we have given a molecular account of them, that a molecular 
basis is a firmer foundation than direct experience. ; 
Let me illustrate this kind of danger. The phenomena of capillarity can, for 
the most part, be explained on the assumption of a liquid surface tension. But if 
the subject is treated merely from this point of view it stands alone—it is a por- 
tion of the building of science hanging in the air. The molecular hypothesis then 
comes in to give some explanation of the surface tension, gives, as it were, a sup- 
orting understructure connecting capillarity with other classes of phenomena. 
ut here, I think, the hypothesis should stop, and such phenomena as can be 
explained by the surface tension should be so explained without reference to mole- 
cules. They should not be brought in again till the surface-tension explanation 
fails. It is necessary to bear in mind what part is scaffolding, and what is the 
building itself, already firm and complete. 
Or, as another illustration, take the Second Law of Thermodynamics. I 
suspect that it is sometimes supposed that a molecular theory from which the 
Second Law could be deduced would be a better basis for it than the direct experi- 
ence on which it was founded by Clausius and Kelvin, or that the mere imagining 
of a Maxwell’s sorting demon has already disproved the universality of the law; 
whereas he is a mere hypothesis grafted on a hypothesis, and nothing correspond- 
ing to his action has yet been found. 
There is more serious danger of confusion of hypothesis with fact in the use of 
the ether: more risk of failure to see what is accomplished by its aid. In giving 
an account of light, for instance, the right course, it appears to me, is to describe 
the phenomena and lay down the laws under which they are grouped, leaving it 
an open question what it is that waves, until the phenomena oblige us to introduce 
something more than matter, until we see what properties we must assign to the 
ether, properties not possessed by matter, in order that it may be competent to 
afford the explanations we seek. We should then realise more clearly that it is 
the constitution of matter which we have imagined, the hypothesis of discrete 
particles, which obliges us to assume an intervening medium to carry on the 
disturbance from particle to particle. But the vortex-atom hypothesis and Dr. 
Larmor’s strain-atom hypothesis both seem to indicate that we are moving in the 
direction of the abolition of the distinction between matter and ether, that we 
shall come to regard the luminiferous medium, not as an attenuated substance here 
and there encumbered with detached blocks—the molecules of matter—but as 
something which in certain places exhibits modifications which we term matter. 
Or starting rather from matter, we may come to think of matter as no longer 
consisting of separated granules, but as a continuum with properties grouped round 
the centres, which we regard as atoms or molecules. 
