240 Royal Astronomical Society. 



improbable that so difficult an object as the satellite c could have 

 been detected ; a was then in the south following quadrant, and 

 rather near the planet, so that it was not seen at either place of ob- 

 servation. 



On the whole, therefore, it seems highly probable that the interior 

 satellites observed by Mr. Lassell and M. O. Struve are entirely 

 distinct ; that the one becomes invisible in the southern part of its 

 orbit, and the other in the northern ; and that, during the last appa- 

 rition of Uranus, it has so happened, that each satellite has been seen 

 only by its own observer. If either of them is the same as was dis- 

 covered by Sir W. Herschel, it seems most likely to be that observed 

 at Poulkova. 



The object observed at Starfield on Nov. 6, 1847, and supposed 

 to be an intermediate satellite between I. and II., was estimated to 

 be only 10'' distant from the planet. But its position was then 

 almost precisely opposite to that of I., whose distance was estimated 

 at 20". It is therefore obvious, that if this were a satellite, its orbit 

 must be interior to that of I. It could not however have been c, 

 which must at that time have been near its greatest northern elon- 

 gation, and therefore probably invisible. Moreover, its greatest 

 distance could be only about 15*4; and it would therefore be nearly 

 intermediate between a and c, and for the present it may be distin- 

 guished as b. Though it may seem premature to attempt any con- 

 clusions from a single observation of so difficult an object, yet, on 

 the other hand, the observation of Nov. 6 was peculiarly worthy of 

 reliance. The night was unusually fine. The planet was viewed 

 for two hours, during which time the supposed satellites were carried 

 along with it. The distance also of b was estimated just half that 

 of I., which was almost exactly opposite to it. And that the esti- 

 mated distance of I. (20") was nearly correct, appears from comparing 

 it with the measured distance of the same satellite (20 /, 57) on Oct. 11, 

 when it was in almost precisely the same part of its orbit. If there- 

 fore it were really a satellite, it appears probable that there are three 

 satellites revolving within the orbit of I., at apparent mean distances 

 of about 12", 15", and 18". 



Mr. Lassell having favoured me with a communication of all his 

 observations of a close satellite of Uranus, with the position-angles 

 estimated at the time, and copies of the diagrams, it appears from 

 them, that only on one occasion has such an attendant been un- 

 doubtedly seen on the southern side of the planet. This occurred 

 on Sept. 27, 1845, at 12 h 6 m Greenwich mean time. The estimated 

 angle was 1 60°, the distance was three diameters of the planet from 

 the edge of the disc, which gives 14" for the central distance. This 

 does not at all agree with any probable period of the satellite a. 

 But on calculating the period of c from this observation, compared 

 with each set obtained by M. O. Struve in 1847, employing as a 

 guide to the number of entire revolutions the approximate period 

 deduced by him, viz. 3 d 22 h 10 m , and giving a weight to the Poul- 

 kova observations proportional to the number from which each result 

 is derived, the period comes out from the mean of the whole, — 3 d 



