Levelling Instrument by the Greenwich Mural Circle. 365 



telescope may be regarded as unexceptionable, but were those 

 of the level equally delicate they could scarcely have been 

 efficient in staff-le veiling at short distances. 



On describing one of the numerous methods made use of 

 to determine the cylindrical error of my horizon sector, I re- 

 marked that the results, which varied from 17"*5 to 27"*5*, 

 were scarcely worth transcribing, owing to the almost impos- 

 sibility of placing one horizontal wire exactly before another. 

 Unable to comprehend how the method which had been con- 

 sidered successful at Greenwich should fail so signally in the 

 above experiments, it occurred to me that the subject might 

 be best investigated by observing through what (minute) 

 range of inclination of the telescope its horizontal wire could 

 be deemed passably coincident with that of the (fixed) colli- 

 mator. My horizon sector, of which the wires are extremely 

 fine, served as collimator, the light being derived, as in the 

 Greenwich observations, from a piece of white paper set up 

 at the most favourable distance from the eye-tube. The tele- 

 scope employed (taken from my repeating circle) is fitted up 

 with wires (by Dollond), apparently of the same fineness as 

 those of the sector, placed between the two nearest lenses of 

 the eye-tube. It was mounted with a very large spirit level, 

 of which the divisions of the scale, about thirty- two to an inch, 

 were equal to l" each. Both instruments stood on a plank 

 resting on brackets driven into solid masonry; and, as an ad- 

 ditional precaution, one of the great levels of the sector served 

 to detect any slight change of inclination in the plank. Du- 

 ring the experiments, of which the results are subjoined, both 

 horizontal wires, as the vertical ones were kept coincident, 

 must have been parallel throughout their length. 



No. I.— February 7, 1835. Temp. 49° Fahr. throughout. 



Point of Scale opposite Annearance of Wir^ 



middle of Bubble. Appearance ot Wires. 



o 



115*5 — Slightly separate. 

 I* 117*5 — As one wire, but rather thicker. 

 J 120* — More like one wire, 

 j 121*5 — Blacker, and as one wire. 

 ^124*5 — As one, but rather thicker. 



126* — Almost as one wire. 



127*5 — Penumbra below. 



129*5 — Scarcely passable. 



132*0— A light line below. 



* See Phil. Mag. and Annals for 1831, N.S. vol. x. p. 347. 



