APPENDIX A XLIII 
It is the characteristic of genuine scientific hypothesis not to 
depart far from the facts nor to neglect them. It consists merely in 
adopting another point of view. 
Thus Newton showed that the earth’s attraction extends to the 
moon by comparing the curvature of the moon’s orbit, or the distance 
by which the orbit deviates from a straight line, in a given time, with 
the distance through which a body near the earth’s surface falls in the 
same time. Between the moon sweeping through its monthly orbit, 
and an apple falling from a tree, there is little outward similarity; but, 
taking another view point, the eye of genius detected the hidden 
analogy. 
An example of a theory too far away from the fact is the elastic- 
solid theory of light. Light not being transmitted instantaneously, 
but taking several minutes to reach us from the sun, it is necessary to 
postulate the existence of some medium in which it subsists during its 
passage to us, and which fills all space. The phenomena of light show 
that it is propagated in this medium like a wave motion in an elastic 
solid. The induction should have stopped there. It was a wrong 
inference that, because the medium behaved in certain ways like an 
elastic solid, it was therefore in all respects an elastic solid. Con- 
sidering it as such, the circumstances of the vibration of light made it 
logically necessary to hold that the ether which pervades all space 
and penetrates most if not all substances was not merely highly elastic 
but absolutely solid. Taking this point of view, such a supposition 
raises mechanical difficulties, insuperable to our minds, for how could 
the planets move without hindrance through a medium immensely 
more rigid than steel? 
The intellectual value of sound theory is well illustrated by 
Poincaré who compares Newton’s discovery with Kepler’s. Kepler 
discovered that the paths of the planets are ellipses. Newton’s law 
of gravitation gives the reason for this; but it does more: it also 
accounts for all the effects of the perturbations, through which 
the orbits, as is shown by more accurate observations than those 
which Kepler used, are only approximately ellipses. Instead they are 
exceedingly complicated curves and are continually changing their 
forms. If, then, Newton’s generalization had never been made, the 
evidence of the reign of law which astronomy affords would have been 
lacking: the progress of observation would have created a belief in 
chaos. 
In the foregoing I have spoken of the observations as the first step, 
and the framing of the theory as the second. The order is not essential. 
In my illustration the city may first be viewed from the tower, if the 
enc uirer is careful afterwards to verify his conclusions nearer at hand. 
