226 ON THE ACCURACY OF THE TABULAR LONGITUDE OF THE MOON. 



I have also selected from the same work, some observations where both limbs of the 

 moon have been observed, and should show the same tabular error. 



Now should an error of observation be made at the time of the maximum tabular error, 

 both with the same sign (and it is as liable to be made at this time as at any other,) the 

 apparent tabular error will be greater than the real error by the amount of the error of 

 observation, which I have shown often amounts to from 3" to 6". In the following list 

 there are comparatively few errors amounting to over 10" when the observations have 

 been made under favourable circumstances, and it is but reasonable to suppose that the 

 excess over 10" is due to the error of observation; and if we reject all observations made 

 while the sun is above the horizon, and the observations of November 12, 1821, and of 

 January 24, 1823, (which are doubtless erroneous,) and allow for possible error of 

 observation as here shown, there is no error that will exceed 7".5 or a half second of time. 



Damoiseau, in his Tables de la Lune, 1824, has given the errors of his tables with fifty 

 observations made at Greenwich, together with those of Burg and Burckhardt. I 

 have also computed the moon's place for the same observations, and included the errors 

 in the following list; together with those of Burg, Burckhardt, Damoiseau, and Plana for 

 comparison. The errors stated in the Tables de la Lune being modified according to the 

 alteration made in the moon's R. A. in the Reduction of the Greenwich Lunar Observa- 

 tions. 



Philadelphia, First month 3rd, 1851. 



