GOULD REDUCTION OF D'AGELET'S OBSERVATIONS. 25 



It is manifest that those present reductions, being purely differential, can lead to no deter- 

 mination of the latitude of the place of observation. This has been used throughout us 48° 51' 5", 

 and any error in this assumption is merged with such other corrections in declination as were 

 constant for the date. 



$ 10. OBSERVATIONS. 



The crude observations printed by Lalande are given in the present memoir in their 

 reduced form, as already stated. The first column, entitled T, shows the clock-time of transit 

 over the mean of the three wires; the second, entitled "Sidereal times," gives the sidereal time 

 corresponding, or (for those dates when the clock was running at sidereal rate, but with very large 

 error) the time T, increased or diminished by a constant amount representing the approximate 

 correction. Next follow the values of the corrections m -f- At, n tg 3, q, from data presented 

 in § 8 ; their sum when applied to T giving the apparent right-ascension. 



Column 8 gives the approximate apparent declination obtained by subtracting the sup- 

 posed latitude, 48° 51' 5", from the mean of the recorded zenith-distances, and is followed by 

 the corrections for refraction and q'. The -'equatorial point," or index-error to be used, is 

 printed at the beginning of the observations for each date, and when summed with the other 

 corrections and applied to £ — <p yields the apparent declination. All these quantities have been 

 computed in duplicate, and the errors of calculation must, I think, be very few. 



The observations of sun, moon, and planets are omitted from the subsequent discussion. 

 The data given render their reduction easy; but it has seemed hardly advisable to carry the 

 computation farther at present. I was not without hopes of finding some observations of 

 Neptune, but none have been discovered. 



A very considerable number of marginal notes will be found accompanying the reduced 

 observations. These relate in most instances to assumptions of errors in the original record. 

 To decide what alterations are requisite, what warrantable, and what inadmissible, is always a 

 matter of delicac}', and owing to the peculiar circumstances of the present case, has here been 

 more difficult than usual. The existence of a large number of errors of press and pen being 

 indisputable, I have felt justified in making a liberal use of probable conjectures, so long as the 

 rule was strictly followed that no change, however slight, be made in the original, without a 

 corresponding record on the same page. This rule does not apply to manifest errors in the 

 column of the original entitled "reduction," where the readings of the division into 96 parts 

 and its subdivisions are translated into the sexagesimal notation. These reductions have been 

 made afresh, without use of the originals, which were found both imperfect and untrustworthy. 



One remarkable kind of error is that where both the recorded readings of the limb agree 

 in giving an erroneous result, precisely as though one reading had been erroneously made, and 

 the other subsequently constructed to agree with it. Fortunately three-quarters of these cases 

 are for stars which are well known, and whose names are given, and the others were fortunately 

 observed on other dates, so that the assumptions made seem fully justified. 



(Sfi) 



