HISTORY OF PHYSICAL ASTRONOMY.

947

the spots generally assign to the planet a period of rotation less than thatwhieh appears to be its most probable value, it may be inferred that the-wind constantly blows in the same direction as that in which the planetrevolves upon its axis. Herschel found that a spot, after its first appear-ance, continued to complete its successive rotations in less and less inter-vals of time, affording thereby a clear indication that its proper motionwas constantly increasing. This circumstance, he remarks, may naturallybe explained by supposing that a considerable time necessarily elapsesbefore the spot can acquire the full velocity of the wind by which it is im-pelled. It is evident, from the amount of discordance exhibited by a com-parison of the different periods of rotation derived from the movements ofthe spots, that the currents in the atmosphere of the planet must possessvery considerable velocities. Sir William Herschel found, from his observa-tions, that the spots, on one occasion, must have been transported through60° in twenty-two rotations of the planet. This would indicate that thewind on the surface of the planet was travelling at the rate of somewhatmore than 200 miles in an hour, a velocity which, however, does not ex-ceed that of occasional winds on the surface of the earth. The nearestapproach to the velocity of the wind will of course be indicated by thesmallest period of rotation. In 1692, Cassini observed a Spot near theequator of the planet, which performed a complete rotation in 9 h 50™*.Assuming (which is certain to be true) that the period of the planet’s realrotation lies between 9 h 55™ and 9 h 56™ f, this would indicate that the cur-rent in the atmosphere of the planet was travelling at the rate of about250 miles in an hour. In general the spots complete their rotations inperiods varying between 9 h 50™ and 9 h 56'", whence it may be concludedthat the currents in the atmosphere do not possess a velocity exceedingthe hundredth part of the rotatory motion of the planet.

Although the spots on Jupiter , by their transient existence and variablemotion, generally indicate that they are of atmospheric origin, there aresome of a more permanent character, whose frequent re-appearance in thesame position can only be explained by supposing them to adhere to thesurface of the planet. The most remarkable of these is the spot situateclose to the northern margin of the southern belt, by watching the appa-rent motion of which Cassini originally succeeded in determining theperiod of the planet’s rotation. From its first discovery in 1665 down to1713, it vanished and re-appeared no less than nine times. When thesouthern belt disappeared, the spot disappeared also, but the belt was fre-quently perceived without the spot. On no occasion did the spot continue

* Mem. Acad, des Sciences, tome x., p. 8. Siemens d’Astronomie, p. 406.

+ It appears, from the researches of Cassini, Schroeter, and Airy, that the period of rota-tion is included between 9 11 56 m and 9 1 ' 55 ln . Independently of the results arrived at by thetwo last-mentioned astronomers, it is evident, from observations of the permanent spotoriginally discovered by Cassini, on the occasions of its successive re-appearances between1665 and 1713, that the period of rotation is somewhat less than 9' 1 56 m . In 1672, Cassini,by means of it, made the period of rotation 9 h 55 m 50 s . In 1708, Maraldi deduced fromit a period equal to 9 !l 56'" 48 s , and from observations of it in 1713 he made the periodequal to 9 h 56 ". On the last-mentioned occasion he compared Cassini’s original observa-tions of the spot in 1665 with his own observations in 1713. Assuming the period ofrotation to be 9 h 56 m , he found that in 1713 the actual position of the spot was about 36°in advance of the computed position. This would afford a valuable means of determiningthe period of rotation, if we could be assured of the precise number of complete rotationsembraced between the observations of 1665 and 1713; but this is an uncertain datum,on account of the magnitude of the probable error of the element to be determined.