Royal Astronomical Society. 533 



was very much disturbed on finding variations, which sometimes al- 

 tered the factor as much as 0*"3 in a day, and more than P*0 in the 

 course of the year. These changes in azimuth had been remarked by 

 Professor Henderson, and were attributed to the irregular action of 

 the counterpoises, which were consequently removed. On a com- 

 parison of these errors with the indications of thermometers plunged 

 in the rock there were apparent marks of correspondence. 



There are several thermometers inserted at different depths in the 

 rock near the observatory, which had been carefully observed in the 

 year 1841*. The indications of these thermometers were projected 

 on paper, and the curves thus formed compared with a curve traced 

 according to the course of the azimuthal deviation. It was thus 

 made evident, that the curve of azimuthal deviation, though having, 

 like the other curves, an annual maximum, did not otherwise resem- 

 ble the curves belonging to the deep-seated thermometers at all ; and, 

 in fact, it came nearest the curve traced out by the thermometer at- 

 tached to the barometer and by the free thermometer exposed to the 

 outer air. Hence the cause of the deviation was not to be looked 

 for in the effect of temperature on the foundations or on the massive 

 transit piers, but on smaller parts more readily affected, such as the 

 metallic mounting. These were accordingly examined. In the 

 azimuthal Y, the construction was found to be much as usual, but 

 the artist has adopted an adjustment for the vertical Y, which seems 

 liable to suspicion. There are two vertical screws applied from be- 

 low ; one, pushing, on the north side of the middle, and the other, 

 pulling, at the south side. The Y is prevented from turning in a 

 vertical plane by jamming horizontal screws, which press a plate 

 against the north face of the Y so as to bring the whole tightly 

 against a stopping-piece, which blocks the south face. Professor 

 Smyth's present opinion is, that the effect of expansion on the two 

 screws, which are in contrary states of constraint, is to alter the ad- 

 justment ; certainly the arrangement looks unmechanical. In the 

 ordinary mode of construction, in this country at least, the elevating 

 Y is either raised by one central screw, or by two screws, one on 

 each side of the centre ; in which case a drawing-screw may be placed 

 at the centre. There is thus no tendency to twist, and the side- 

 plates which confine the Y laterally have to exert little restraining 

 force. Professor Smyth has communicated with MM. Repsold, the 

 makers of this magnificent instrument, and is awaiting their reply 

 before adopting any remedy t- 



* Some years ago, Professor J. D. Forbes had four thermometers sunk in 

 the rock with their bulbs at the depths of 24, 12, 6, 3 French feet and a 

 fifth on the surface merely covered with sand. 



t Sudden and lawless changes in azimuth forbid independent determina- 

 tions of the azimuthal deviation (which are also the best;, viz. from the 

 consecutive semidiurnal transits of circumpolar stars. The possessor of an 

 imperfectly mounted instrument must content himself with assuming the 

 fundamental places of his close circumpolar stars, and determine his azi- 

 muthal error from each of them. This will, with proper caution, be found 

 quite sufficient for objects not too near the pole, especially when the clock- 

 error stars are pretty numerous, and situated above and below the object 

 to be determined. 



