415 
the error in the first two cases is only casual: and also, what is 
of more importance, that the centering is very nearly accurate ; 
and though it is just possible that the large error in the mea- 
sured distance in the two first cases is due to bad graduation, it 
is clear that the fault extends to both sides of the imstrument. 
The index-error (additive) at the zero-point having remained 
constant for some months at about 10°, this amount has been 
added in all cases. Having thus shewn that the anomalies can 
only be due to a fault in the graduation on the corresponding 
opposite sides of the instrument, such as can only be accurately 
estimated by a long series of observations and comparison with 
other instruments, I will proceed to give the elements of the 
observations themselves. 
G.M.T. L.M.T. © app. alt. © true alt.) app. alt. ) true alt. 
h. m. Ss. h. m. s. SOUS ayy Go iat ane Oper tars oR a ae 
(1) 4.32.26 4.33.22 6.37.20 6.29.38 5.43.32 6.29.37 
(2) 4.39. 4 4.40. 0 5.40.20 5.31.23 6.18.13 6.59.50 
(3) 4.41.18 4.42.14 5.20.58 5.11.38 6.22.52 7. 9.40 
(4) 4.44. 3 4.44.59 4.57.24 4.47.18 6.34.50 7.21.50 
(5) 4.46.35 4.47.31 4.35.33 4.24.45 6.47.39 7.34.52 
(6) 4.53.26 454.22 3.36.45 3.23.45 7.13.40 8. 1.12 
(7) 4.55.22 4.56.18 3.20.13 3. 6.28 7.21.10 8. 8.50 
yee ne &). pele Error in time. Error in are. 
ee ees eee | ane ae tae 
(1) 101.26.50 101. 21.27 5.34 2.36 
2) 10L2 29. 8 101 .24.36 5.31 2.34 
(3) 101.30.40 101 . 26.26 3.35 1.38 
(4) 101.32.20 101.28 .30 1.59 0.56 
(5) 101.32.10 101.28 .42 4.46 2.13 
(6) 101.35.20 101.32.58 2. 8 1. 0 
(7) 101.36. 0 101.33.58 1.44 0.49 
As the G. M.T. was obtained by comparing the watch with 
a carefully rated chronometer immediately after the observa- 
tion, I am certain that it is not more than 5° to 10°, at the 
most, out. 
