HISTORY OF PHYSICAL ASTROXOM.

81

be no equation of tlie centre; but, as Delambre justly remarks, theremight exist the inequality of the variation even in a circular orbit, and, aslong as it was neglected, it would occasion an apparent discordancebetween the several equations of light. The arguments of Maraldi were,however, considered by some of his most eminent contemporaries to befatal to the theory of Koemer. “ It appears then,” says Fontenelle, “ thatwe must renounce, though perhaps with regret, the ingenious and se-ductive hypothesis of the successive propagation of light.How

little prevents us from falling into great errors! If Jupiter had hut onesatellite, or if the eccentricity had been less, and these two things arevery possible, we should have concluded with the utmost confidence thatlight traversed the annual orbit of the earth in 14 minutes.” *

Maraldi first established the important fact, that the inclination of thesecond satellite is variable. He was led to this discovery by observingthat the duration of eclipses was not always the same when the satellitewas at the same distance from the nodes, a fact of which he assured him-self by a careful comparison of a great number of eclipses. For example,on the 21st January, 1608, when Jupiter was in that part of his orbitwherein the eclipses are shortest, he found that the semi-duration ofthe eclipse was l h 19 m ; on the 17th September, 1715, when all thecircumstances were the same, the semi-duration was only l h 7 m 14 s .The difference amounted to ll m 46 s , and, as this quantity was too greatto be ascribed to the errors of observation, he concluded that it musthave proceeded from a change in some of the elements upon whichthe phenomenon depended. He remarked that the duration of theeclipse might he modified by three distinct causes; 1st, a variation inthe eccentricity of the satellite; 2nd, a variation in the inclination;3rd, a variation in the place of the nodes. With respect to the firstof these causes, the variation would require to be enormous, in orderthat it might occasion so great a difference between the eclipses;with respect to the second, he remarked that the duration of the eclipsevaried even when the satellite was at the same distance from the node.He concluded, therefore, that the phenomenon must be ascribed to achange in the inclination of the orbit. In 1707 he found the inclinationto be equal to 3° 33'; whence it appeai’ed to have increased nearly adegree since the publication of Cassini’s tables.

Bradley is the next astronomer whose researches contributed to throwlight upon this interesting subject. In 1710 he constructed tables of thesatellites; but they were not given to the world until 1740, when theywere published along with Halley’s tables of the planets. In these tablesthe places of the satellites were given by Bradley, in degrees and minutes ofspace; but there were appended to them ecliptic tables of the first satellitecalculated in time by Pond, his uncle. He determined the mean motionswith great accuracy, by means of a comparison between the observations ofpreceding astronomers and those made by himself at Wanstead, afterJupiter had completed four revolutions. We have seen that Cassini andMaraldi refused to admit the equation of light; Halley, in 1694, arguedmore philosophically on the subject; for he maintained the necessity ofapplying it to ail the satellites. Bradley was the first who introduced thisequation into the tables of their motions. He fixed it at the usuallyreceived value of 14 m , adding a smaller equation of 3| m to account for the

* Mem. Acad, des Sciences, Hist. p. 80.

o