143 



which ho conceived to shew the temperature of the surface of the I'ock 

 outside the Observatory ; and was accustomed latterly to detorniino 

 the level error of the axis merely from the indications of the ther- 

 mometer. 



The author has found from Professor Henderson's own observa- 

 tions, a temperature effect on the azimuthal position of the axis four 

 times as strong as in the case of the level. 



Indeed the extent was so great as almost to vitiate the observa- 

 tions ; for the amount of error varied regularly to such an extent from 

 day to day, as to preclude the possibility of employing the only unex- 

 ceptionable method of determining the azimuth of the instrument : 

 viz. three consecutive observations of a star above and below the 

 pole. 



In order to try to ascertain how heat produced this effect, or what 

 were the parts acted on, six different thermometers were discussed. 

 Five (Professor Forbes's) had their bulbs buried in the rocks 50 feet 

 outside the Observatory, at the depths of 24, 12, G, 3, and 0- French 

 feet respectively ; the last having its bulb merely covered with sand ; 

 a sixth thermometer had been observed under the floor of the Ob- 

 servatory, and its indications came between the 0" and the 3 feet in- 

 struments. 



The deeper the thermometers, the greater was found their differ- 

 ence from the azimuthal errors of the transit ; the test consisting in 

 the pi-ominence and similarity of the daily variations, and the degree 

 of retardation of the grand annual wane of heat. 



Of the former it was difficult to get a numerical estimate, but of 

 the latter a result was obtained shewing the ratio of the quantity of 

 heat in the first half of the year, to that in the last : — 



In the case of the 3 feet Thermometer, the ratio was 1 : 1"941 

 Thermometer under floor, . 1 : 1"639 



0- feet, .... 1:1-351 



Instrumental errors of Transit, 1:1-158 



plainly shewing that the parts acted on by temperature were more 

 quickly affected by it than a thin coat of sand, and so could neither 

 be the rocky foundations of the piers, or the massive stone piers 

 themselves ; but must be, considering too the effect of the walls and 

 roof covering the instrument, something very small, and most pro- 

 bably metallic. 



The field of inquiry being thus curtailed, the brass supports of the 

 axis on the top of the stone piers were examined, and the cause was 



