microscopes a, b, c, d, e, f, are attached. The face of the pier which carries the microscopes, fronts either

the east or the west. The construction of the instrument is this; there is a circle BC, which turns round an axis DE, (not visible in the view) that passes through the pier A. The circle BC has the telescope FG attached. This circle is graduated into degrees and minutes and other sub-divisions on its outside, containing 360 degrees in its whole circumference. Its position would be sufficiently observed for ordinary purposes if there were a pointer fixed to the pier at one part, but there are reasons (depending on the liability of the axis to be disturbed in its bearing, and on the tendency of the circle to bend under its own

weight) which make it desirable that there should be pointers at several parts of it. In the instruments at Cambridge and Greenwich, and other places, there are six pointers; they are not ordinary pointers, but microscopes, by means of which the spaces between the divisions can be sub-divided with greater accuracy than they could be by other means. Therefore you will perceive very easily, that by the use of these microscopes, viewing the circumference of this circle, it is possible to determine and register the position of this circle, (and consequently the position of the telescope which is fastened to it,) with very great accuracy indeed.

In all measures, however, we want a starting-point. What we want to ascertain with the circle is, how far the telescope is pointed above the horizon. It is therefore a very important thing to ascertain what is the reading of the circle when the telescope points horizontally. There is a contrivance used in most modern Observatories for this purpose which is worthy of attention-it is the use of observations by reflection. Suppose that a star is seen by the observer to be approaching the meridian, he places a trough of quicksilver O in such a direction that the star can be seen by reflection in the quicksilver. When the telescope is pointed towards the reflection in the quicksilver, then we know that the telescope is pointed below the horizon, just as much as it is pointed above the horizon to see the star by direct vision. This results from the optical law of reflection. For in Figure 13, if FG be the position of the telescope placed to receive the light which comes from the star in the direction SG, and if F'G' be the position of the telescope placed to receive the light which comes from the star to the quicksilver in the direction

S'O, and is thus reflected in the direction OG', then by the law of reflection, S'O and G'O make

equal angles with the surface of the quicksilver. But as the quicksilver is perfectly fluid, its surface is exactly horizontal. So that G'O and S'O make equal angles with the horizon; and therefore FG'O points as much below the horizon as OS' or FGS points above it. The observer therefore looks at the star by reflection in the quicksilver; he takes the reading of the microscopes; he then turns the instrument so as to see the star by direct vision in the telescope, and then he takes the reading of the microscopes; then by taking the mean between the reading of the circle corresponding to these two observations, it is certain that we have got the reading corresponding to the horizontal position of the telescope. That gives us a starting-point; and having got that, whenever we

observe a star in any position whatever on the meridian, inasmuch as we have got the reading of the circle when the telescope is directed to that star, and as we know the reading of the circle which correponds to the horizontal position of the telescopethen by taking the difference between these readings, we know in degrees and minutes the inclination of the telescope, or the degrees and minutes by which the star is elevated above the horizon. The method of observation which I have described is going on at an Observatory every day. It is necessary, however, to remark that (as has been already said) every star appears too high, in consequence of refraction; a correction must therefore be subtracted from the elevation thus found, in order to discover at what elevation it would have been seen, if there had been no atmosphere about us.

Now, suppose that we observe the Polar Star. This star, though very near the Pole, describes a small circle round the Pole, and therefore goes as much above the Pole at one time when it is highest, as it does below the Pole at another time when it is lowest. Therefore, by taking the angular elevation above the horizon, in degrees, minutes, and seconds, of the Polar star when at the highest point above the Pole, and applying the proper correction for refraction; and taking its angular elevation in degrees, minutes, and seconds, when at the lowest point below the Pole, and applying the proper correction to this for refraction; and taking the mean between the two elevations so corrected; we get the true angular elevation of the celestial Pole. In that manner we have got the accurate calculation of the angular elevation of that Pole in the North side, round which the heavens appear to turn.

D

Now, allow me to point out what we have obtained with regard to these celestial objects.

By the use of the transit instrument, when properly adjusted, and the clock, we observe the time of transit of a principal star, and we observe the time of transit of any other objects, smaller stars, planets, or whatever else they may be. By means of these observations, we have a difference of times of transit. We can place no reliance upon the clock, except this, that it gives us the difference of time between the passage of principal stars, and that of other objects. Suppose that our clock is so adjusted, that if we observe the time of that principal star passing the instrument to-day, and again observe the time at which it passes the instrument to-morrow, the clock describes accurately twenty-four hours. If it does not describe accurately twenty-four hours, we know how great its error is in twenty-four hours, and we can apply a proportionate part of the error to every interval of time; so that it is in every respect as serviceable as if it were accurately adjusted. Supposing, then, that our clock was adjusted in such a manner that it indicated twenty-four hours, from the time of a principal star passing to the time at which the same object passed again-this amounts to saying, that it indicates twenty-four hours in the time in which the whole heavens turn round. Assuming, then, that the planet which we have observed passes the telescope one hour after the principal star passes, then we must conclude that the heavens have turned for one hour; or have performed one twenty-fourth part of their whole revolution, before that part of the heavens in which the planet is seen passes our meridian. And this is precisely one of the co-ordinates which, as I said, serves to determine the position of