140 ESSAY ON THE VELOCITY OF LIGHT. 
variation in the time comprised between seeing and hearing each blow; but 
even when we only hear the blows without seeing them, that irregularity is 
always detected by the attentive consideration of the successive sounds; it 
seems as if the sounds experienced a periodic perturbation—that they take place 
now before, now after the instants at which we ought to hear them so that they 
would follow each other regularly. An observer placed in the conditions 
which we have supposed, and ignorant of the conditions of the progressive* 
transmission of sound, and who would thus hear the sounds of the axe or ham- 
mer succeed each other now with rapidity, now with slowness, as he walked 
round his circular path, might well imagine, at first, that the blows themselves 
had this irregularity; but in thinking of the manner in which the blows are 
produced, he would have reason to think that such a periodie irregularity in 
their production was not very probable; and especially as he would remark that 
the retardation and acceleration of the successive blows coincided always with 
his going away from or his approach to the place where they originated, he 
would naturally be led to attribute the irregularity of the succession of sounds 
to his own change of position, and would thus be led to the discovery of the 
progressive transmission of sound. 
It was in exactly similar conditions that Roémer found himself when he 
examined and discussed with care the observations of the eclipses of the first 
satellite of Jupiter. This satellite, in entering the umbral cone of Jupiter, 
suddenly loses its light, and in coming out of the umbral cone it instantly 
regains it. There is, therefore, in the successive immersions and emersions of 
that satellite, a series of phenomena of light which are reproduced as regularly 
as the blows of the hammer of which we have spoken above. ‘The astronomer 
who observes these phenomena is carried by the earth each year over an orbit 
which differs little from a circle having the sun as a centre. It is true that 
Jupiter moves at the same time around the sun, so that the place where the 
eclipse of its first satellite takes place is displaced continually round that cen- 
tral body ; but the time of the revolution of the earth around the sun being 
much shorter than that of Jupiter, the distance of the observer from that 
planet and from its satellite experiences the same-alternative increase and dimi- 
nution of distance as if Jupiter were stationary. ‘The period of the returns of 
the observer to the same distance from Jupiter is alone modified; it is about 
399 days instead of only one year. If, therefore, light is not imstantaneously 
transmitted to all distances, if it requires an appreciable time to run over a 
length such as the diameter of the orbit of the earth, we ought to find the 
effects in the observations of the eclipses of which we have just spoken. If 
from observations continued through several years we deduce the mean interval 
of time which separates two consecutive eclipses, we ought to find that the 
time is greater than the interval really observed between two consecutive 
eclipses while the earth is approaching Jupiter than the interval observed 
while the earth is going away from Jupiter. If, in using the value of this 
mean interval of time which elapses between one eclipse and the following and 
setting out from an eclipse observed when the earth is at its mean distance ; 
from Jupiter, we wish to predict the return of future eclipses, it happens that | 
the instants when the eclipses are really observed are sometimes in advance of 
and sometimes in retard of the predicted epochs, aceording as the earth is 
nearer to or further from Jupiter than at the moment of departure from the 
point of mean distance. ‘The times furnished by the observations for the com- 
mencement or the ending of the different eclipses will therefore appear sub- 
mitted to a periodic perturbation agreeing with the period of variation of the 
distance of the earth from Jupiter. ‘This is precisely what Roémer found in 
discussing the numerous*observations which he had in his possession of the 
eclipses of Jupiter’s first satellite; he thus found that light employs 22 
minutes to run over the diameter of the orbit of the earth around the sun, and 
