66 Royal Astronomical Society. 



longest levels the construction with which I had become familiar In 

 Germany and Russia, in which the glass tube of the level is supported 

 in Ys : since that time the levels have been fairly accordant. Another 

 contrivance extensively applied to German levels, namely the cover- 

 ing by glass shades, has also been applied here. A difficulty which 

 can be surmounted only by constant care has sometimes presented 

 itself, namely, that when the dome is opened very shortly before 

 observation, the changes of readings of the upper and lower levels do 

 not exactly correspond. Lastly, when the best values of instru- 

 mental errors of every kind are applied, the accuracy of every part 

 of observation, of calculation and application of instrumental errors, 

 and of tabular calculation, is checked by the determinations of the 

 zero of azimuth, lliese determinations are sufficiently steady in any 

 one evening, or perhaps in groups of several evenings ; but they are 

 not steady from time to time, the variation amounting to three or four 

 seconds of arc. Whether this arises from a twist of the brick pier, 

 or from a twist of the piers of the transit instrument (the times being 

 obtained from the transit- clock), or from a change in the observer's 

 personal equation, I cannot tell. The substitution of improved me- 

 ridional instruments for those now in use will enable me to remove 

 one of these conjectural causes. 



" I have spoken hitherto solely with respect to the azimuthal ob- 

 servations, in which alone, from the first, I anticipated any difficulty. 

 The zenith distance observations have never given the smallest 

 trouble. 



" The accuracy of the results, as estimated by the observation of 

 stars, is somewhat less than that of the mural circle, perhaps nearly 

 in the degree which might be expected in circles whose diameter is 

 half that of the mural circle. 



" For observations with the altitude and azimuth instrument, the 

 following rule is laid down. The moon is to be observed if visible, 

 and the observer is bound to watch if necessary while the moon is 

 above the horizon, and the sun is not more than an hour above the 

 horizon. One azimuth and one altitude are to be observed, and if 

 possible, two azimuths and two altitudes in reversed positions of the 

 instrument : and if the night is fine, a low star and a high star are 

 to be observed in azimuth, both in reversed positions of the in- 

 strument, and one star in altitude, in reversed positions. Thus a 

 complete set includes ten observations. These rules have been fol- 

 lowed carefully during the thirteen lunations intervening between 

 1847, May 15, and 1848, May 30; and I am now able to give a 

 comparison of the number of days of observation of the moon with 

 the altitude and azimuth instrument and with the meridional instru- 

 ments. With the altitude and azimuth instrument no days are in- 

 cluded except both altitudes and azimuths are observed ; and with 

 the meridional instruments, no days except both right ascensions and 

 polar distances are observed." 



The number of days during the last year in which the moon has 

 been observed at Greenwich are — 



With the altitude and azimuth circle =203 

 meridional instruments ... =111 



