TRANSACTIONS OF THE SECTIONS. 17 
If the image of the declination line be made to coincide with it, the optic axis of the 
telescope is obviously vertical, and the reading of the circle gives the nadir. When Dr, 
R. found Kater’s collimator uncertain, he made an illuminating eye-piece, and essayed 
this method with tolerable success in 1832, but as the power was but 25, he did not 
pursue it further than was necessary to obtain a value of his latitude. About three 
years since his attention was recalled to it by Mr. Henderson, who was using it with 
great advantage; and he made the necessary alteration in the observing eye-piece 
(power 250) to avoid the necessity of changing it when illuminating. He now finds this 
method much superior to the pole star observation in expedition, notinferior in certainty, 
and so easily practicable, that he in general’ determines the index correction at the 
close of each night’s work. From the obvious fact that the angular movement of the 
image is twice that of the telescope, the precision is double that ofa star, independent 
of the fluttering of the latter. Dr. R. would also call the attention of astronomers to 
the fact, that in determining the division corrections of a circle it must be remem- 
bered that they are occasionally variable with the position of the instrument, and with 
its temperature. It will, for instance, occasionally be found that the mean of six 
microscopes will differ from that of two, or of twelve from four by unequal quantities, 
when the readings of the index differ 180 degrees; and this throws a suspicion on the 
usual mode of examining divisions. The difference is far too great to be attributed to 
the error of observation, and appears to occur in every circle of which a detailed ex- 
amination has been published. In the Armagh circle Dr. R. has found a few cases of 
the effect of temperature. The most prominent is that of the four divisions used for 
the nadir point, and it is remarkable that if this had not been attended to he would have 
found for « Lyre and « Cygni parallaxes very nearly equal to those assigned them by 
Dr. Brinkley, Bishop of Cloyne. These remarks seem to furnish an argument in favour 
of the use of moderate-sized instruments, and the improvement of engine division. 

On the Barometric Compensation of the Pendulum. By Dr. Rosinson. 
At the Manchester Meeting of the Association Professor Bessel made a communi- 
cation on the improvement of the astronomical clock, which, with other valuable mat- 
ter, contained a proposal to compensate for the changes of rate produced by the vary- 
ing density of the atmosphere. This appears in the Report of the Sectional Proceed- 
ings, and also at much greater length in No. 465 of the ‘ Astronomische Nachrichten.” 
At the time Professor Stevelly remarked that I had not merely proposed but applied 
this compensation twelve years ago *; and I should not have reverted to it, but that I 
think my method possesses certain advantages over that proposed by the illustrious 
astronomer of Kénigsberg, which entitle it to the preference in practice. It was long 
believed to be demonstrated that the rate of a pendulum was influenced by the air’s 
density only as far as it lessened the are of vibration and diminished its gravity by 
buoyancy. The researches of Kater on the length of the second pendulum are all 
Vitiated by this mistake, which was discovered by Bessel during a similar investigation, 
in which he found, by using balls of different specific gravities, that the received 
buoyancy correction is too small. As early as 1825, and without any knowledge of 
what Bessel was doing, I had ascertained the same fact by comparing the rates of my 
transit clock with the barometric indications; and Colonel Sabine gave the tinal proof 
of it by swinging the pendulum in a vacuum apparatus in the year 1829. The amount 
of it is far from inconsiderable; even with the mercurial pendulum of my transit 
clock, which weighs 21 pounds, and presents a very small surface, it is 0°36 for an 
inch change of the barometer. Now the remedy for this is obvious. If we attach a 
barometer to the pendulum, its fall transfers a cylinder of mercury from point near 
the axis of motion to a greater distance from it; the time of vibration may thus be 
made to increase by the same amount that it decreases in consequence of the dimi- 
nished density of the air. By placing the clock in vacuo, as Bessel proposes (and as 
Sir James South has actually done for several years past), the effect of resistance can 
be determined exactly, and the diameter of the tube selected, which will nearly cor- 
rect it. This is not mere speculation, for I have verified it by trial. The diameter 
to Astronomical Memoirs, vol. vy. Dependence of Clock’s rate on Barometer. 
. Cc 
