It appears from these results that the place of the pole or 

 the instrument, whenever it had undergone no change, is sub- 

 ject to very small variations, which in the course of each pe- 

 riod always appeared insensible. In order to trace, however, 

 likewise these small inequalities, I have noted down for some 

 months the state of a thermometer attached to the pillars, and 

 the length of the air-bubble of the level of the altitude in the 

 last column of the observations, e.g. No. 4 bubble 63*6 = 

 7°*6; but when I found no connexion, I did not continue 

 this practice any further. The instrument got repeatedly new 

 wires, the intervals between which are stated in the observa- 

 tions. At the passage of the sun the first limb was usually 

 observed on the first four, and the following limb on the last 

 three or four wires ; in the mean time the northern or southern 

 limb was brought between the horizontal wires, and the cir- 

 cle read off. There is only time to do this if the observation by 

 the circle is made 20" or 30" before the culmination, and two 

 verniers only are read off". For this purpose it becomes ne- 

 cessary to know the deviation of the horizontal wires from 

 parallelism to the equator, and the difference of two verniers 

 from four. The former was investigated by means of ob- 

 serving a. Ursa Minoris at the two extremities, and in the 

 middle of the system of wires. By this means the correc- 

 tions to be applied to the readings of the circle to correct this 

 error, were found as follow : 



From 



