70 Royal Astronomical Society. 



Nov. 1. The satellite better seen than I. and II., which are very near each 

 other. 

 28. The satellite quite as visible as I. 

 Dec. 9. The images very good, but the observation not sure. The satellite 

 seen by glimpses. 

 10. The satellite as well seen as II., but fainter than I. 

 The observations of Nov. 28 show clearly enough that the object was a 

 satellite: the position angle of a fixed star would have been altered 10°. 



It may at first sight seem odd, that all the observations are made 

 on the same side of the planet. There are, however, many analogous 

 cases among the satellites, if this should always occur. The satel- 

 lites I. and II., for instance, are found to alternate in brightness. 

 There has been, however, no really fine observing night after Oc- 

 tober 8, on which the satellite could have been observed on the other 

 side. 



The observations of this inferior satellite are not sufficiently nu- 

 merous to yield an accurate orbit. It may, however, be assumed 

 pretty safely, that it has been observed near its greatest elongation, 

 i. e. that the apparent semiaxis is between 17" and 18" ; in which 

 case it must have a period somewhere about three or four days. 



Now the satellite has, approximately at least, always been observed 

 at the same part of its orbit, and hence the intervals of twenty-four, 

 twenty-seven, and twelve days must, approximately, be multiples of 

 the periodic time. The hypothesis of a period of three days, which 

 first presents itself (or more accurately, from the observations 

 of October 8 and December 10, 2 d 23 h 45 m ), M. Struve finds wholly 

 incompatible with the observation of December 9, and he prefers a 

 period of four days (more accurately 3 d 22 h 10 m ), which, admitting 

 a slight error on Nov. 1, agrees equally well with the observations. 

 This result also accords better with the presumed mass of the planet ; 

 for the first period and greatest distance, 17"*5, gives the mass 15 ^ 8 ? 

 larger than Bouvard's, while the second period assigns 26 | 6 q to the 

 mass, somewhat less than Lamont's, but yet a tolerable approxima- 

 tion to it. 



The difference between M. O. Struve's period for this satellite, 

 and that of 5 d 21 h 25 m , assigned to Sir W. Herschel's, is considered 

 by M. O. Struve to be due to an erroneous estimate of the semiaxis. 

 He has no doubt of the identity of the two objects. 



After remarking that the foregoing observations completely prove 

 the existence of Herschel's first supplementary satellite, M. Struve 

 remarks that the visibility of this satellite depends less upon the 

 aperture of the telescope employed than on the sharpness of the 

 image shown. All the telescopes, in his opinion, which will show 

 I. and II., are large enough to show I. if they are good enough. It 

 is not fainter than the other two, but, from its greater proximity 

 to Uranus, requires better defining power and freedom from stray 

 light. 



M. O. Struve has not hitherto seen any other supplementary sa- 

 tellite, but he remarks that the position of Uranus is every year 

 becoming more favourable for this inquiry ; and there can be little 

 doubt of their speedy rediscovery, so far as they actually exist. 



