PROGRESSIVE MOTION OF LIGHT. ]21 



The next important step in the progress of optical science was the discovery 

 of the progressive motion of light, and tlie determination of its velocity. Though 

 every theory which had ever been suggested to account for the phenomena of 

 light presumed that there must be a progress from the luminous origin, and 

 therefore that time must be an element in the solution of every optical problem, 

 still so nearly instantaneous are all the cfiects produced at the; distanccis to 

 which our ordinary observation extends, as apparently to render hopeless any 

 ]>hui for experimentally determining the velocity. This circumstance rendered 

 the efforts made by the celebrated Galileo, and by the academicians of Florence, 

 to settle the question, completely nugatory. 'J'he method of proceeding adopted 

 by Galileo was to place himself upon an eminence opposite to an assistant 

 observer something more than a mile distant; both being provided AVith lanterns 

 Avhich could be darkened by a slide. The lights being arranged, Galileo dark- 

 ened his lantern ; and the assistant, immediately on noticing its disappearance, 

 darkened his also. Apparently both Avere extinguished at tlie same instant. 

 The Florentine academicians repeated the experiment, increasing the distance 

 between the stations, but the result Avas the same. The problem remained 

 imsolved ; but its solution came at last, Avhen demanded by the exigencies of a 

 higher branch of science. 



In 1675 Roemer, an astronomer of Copenhagen, in his obsej'A'ations upon the 

 eclipses of the first satellite of Jupit(;r, became perplexed by irregularities for 

 Avhich lie could conceive no means of accounting. It Avas suggested by Dominic 

 Cassini that these difficulties might perhaps be removed by supjjosing that the time 

 occupied by light in passing through the vast distance betAveen Jupiter and our 

 planet may be large enough to be ajipreciablc; and therefon; that, as our dis- 

 tance A'aries, this time must vary also. Assuming this hypothesis to be true, 

 and that the epoch on Avhich our computations of futvu-e (;clipscs are founded is 

 the date of some eclipse actually observed Avhen the two bodies Avere occupying 

 their points of nearest approach, it will follow that if the accuracy of the deter- 

 minations is affected only by the motion of light, all subsequent eclipses, 

 observed when the distance is the same as at the epoch, will agree with the 

 prediction, and all others will be in retardation by an amount of timi; equal to 

 that which light requires to pass over ihe space by which the distance has been 

 increased. In like manner, if the epoch had been an eclipse observed in the 

 position of greatest distance betAveen the bodies, subsequent eclipses Avould be 

 in ad\"ance of the prediction ; and if the epoch had been an obserA-ation made 

 from some position intermediate betAveen the points of greatest and least distance, 

 the eclipses afterAA\ards occurring Avould be sometimes in advance and some- 

 times in retardation. The test of the correctness of the hyjjothesis Avould be a 

 careful comparison of the observed irregularities of time Avith the variations of 

 distance — a compai'ison involving no slight labor. Cassini, Avith wliom the idea 

 originated, seems to have abandoned it; but Roemer folloAved it up Avilh such 

 perseverance as at length conclusively to establish its truth. He demonstrated 

 that the time occupied by light in passing over the entire diameter of the earth's 

 orbit is 16 minutes and 26 seconds. But at that period the dimensions of the 

 earth's orbit Avere not accurately known, and this determination Avas insufficient to 

 fix the absolute value of the velocity of light. Assuming the sun's mean parallax 

 to be 8. "6, the mean diameter of the orbit must be about 190,000,000 of miles, 

 and this number divided by 986, the number of seconds in 16 minutes and 26 

 ;K!Conds, gives for the velocity in miles 192,700. 



The velocity of light has, since the time of Roemer, been ascertained, with a 

 probably near approximation to the truth, by other inch'pendent methods, and 

 the results tend to confirm the substantial correctness of his original determina- 

 tion. . The first of these methods is that Avhich rests upon the measurement of 

 the aberration of the stars, a phenomenon discovered by Bradley, afterAvards 



