TRANSACTIONS OF SECTION A. 5?! 



length X in some or in nil directions. P probably acts somewhat like a Hertzian 

 vibrator ; but whatever be its modus operandi it is an agent which makes a disturbance 

 in the jether and sets up what we may call turmoil in its immediate neighbour- 

 hood. This turmoil is of a special kind, its action on the tether beyond adding 

 wave after wave to an undulation of regular waves, which advances outwards. 



It is this undulation of regular waves beyond the region of turmoil that is the 

 light radiated from P. 



The aether is competent to propagate these waves forward without external aid 

 and by reason of forces developed within itself when strained ; but the turmoil in 

 Ihe vicinity of P requires that forces supplied byP shall co-operate with the forces 

 developed in the medium to keep it going. If P ceased to maintain it, the 

 turmoil would quickly disappear after expending whatever energy had been stored 

 up in it in adding a few additional waves to the inner fringe of the great undula- 

 tion travelling outwards. 



Let us draw round /a tiny sphere with radius p, which we may call sphere p, 

 just sufficiently large to include the region of turmoil. In the case of light, one or 

 two wave lengths is a sufficient radius for this sphere, since beyond that short 

 distance the events in the aether do not differ sensibly from regular wave-motion. 



P, which emits the light, is a portion of the non-sether. It is a 'source' 

 through which energy is transferred from the non-aether to the aether. By reason 

 of its presence the aether is not a ' self-contained system ' of the kind which is 

 necessary to justify an application to it of the principle of reversal. But we can 

 bring about this requisite state of isolation by supposing that P, after having 

 emitted light for a definite time, say for one minute, not only ceases to emit light, 

 but ceases to exist. This total suppression of P cuts off the communication 

 between the aether and the non-aether, and thenceforward the aether is a self- 

 contained system in which we may investigate ihe further progress of events by 

 employing the principle of reversal. It will be convenient to divide time into 

 equal intervals — say into minutes — and the definite supposition we shall make is 

 that P emits light of wave length X from the epoch t = till the epoch ^ = one 

 minute, and that at the close of this period all the contents of the sphere p, including 

 P and the disturbed aether near it, are suddenly annihilated, and quiescent aether 

 put in their place. 



By the end of the first minute, when these events are supposed to take place, 

 the undulation beyond sphere p has extended to a distance from/, which is about 

 forty-seven times the distance from the earth to the moon. After those events 

 take place, the undulation continues to advance outwards ; and we may now 

 employ upon it the principle of reversal, with the advantage of being at liberty 

 to confine the reversal to the reversal of motions in the aether. This provides us 

 with the means of investigating events after the first minute. 



We may also include the events of the first minute by introducing two 

 reversals ; since by this contrivance we can succeed in reproducing under the new 

 conditions, i.e., within a. self-contained aether, precisely the same undulation as 

 existed during the first minute while P was emitting light. To this end let us 

 imagine the undulation to continue its outward journey for any convenient period 

 —say for two minutes after the annihilation of the contents of sphere p. This 

 brings us to the epoch i = three minutes. At this instant let reversal of all 

 motions in the aether take place. The outtiowing waves then retrace their steps, 

 so that after the reversal the undulation becomes light converging towards the 

 focus/. When the time f = eight minutes arrives the undulation has not only con- 

 verged upon /, but after passing that focus it has become an undulation of 

 divergent spherical waves, each part of the undulation when passing the focus 

 having crossed to the opposite side of /. At the instant t = eight minutes let a 

 second reversal of all motions in the aether take place. The light which, im- 

 mediately before this second reversal, was diverging from / again becomes 

 convergent, and within the period from ^ = ten minutes to ^ = eleven minutes 

 each spherical wave for the second time passes the focus and becomes divergent, 

 and each of these divergent waves now finds itself under such circumstances that 

 so soon as it gets beyond little sphere p it becomes for all future time an exact 



