nearly a semicircle, the same process will take place at the other spring; these springs, therefore, afford the means of limiting the horizontal motion of the pan to 180 degrees.

Figs. 7. and 8. are the plan and section of a float of cast iron, 10.6 inches diameter, one inch thick, and having an opening in the middle of 5.7 inches diameter, consequently the breadth of the annulus is about two inches and a half. Into the sides of this float, at opposite points and equidistant from the two surfaces, two steel pins are screwed, of such a thickness as to pass freely into the grooves of the guides before mentioned with a very little shake. The ends of the pins are to be hemispherically rounded, and very smooth. When the float is placed in the pan, with the pins in the guides, the distance between the terminations of the pins should be such as to leave them just clear of the backs of the guides.

A section of the float with the bridge and telescope attached to it is given at fig. 9. The bridge is made of wrought iron. The length at the top is seven inches, and the perpendicular part of the bridge is of a sufficient height to enable the bridge to clear the middle part of the pan when the float is placed in it. In my vertical floating collimator the bridge is an inch and three-quarters above the float. The middle of the bridge consists of a piece of brass tube three or four inches long, destined to receive the telescope, which fits tightly into it. The width of the iron part of the bridge is half an inch, and the thickness about a quarter of an inch; but this may be varied according to circumstances which will hereafter appear. The parts where the bridge is screwed to the float have cross pieces four inches long, intended to give a firmer bearing. The heads of the screws are long, to serve as pins, upon which weights with holes in them may be placed, for the purpose of adjusting the float,

The bridge is screwed to the float in such a position that its length is at right angles to the pins inserted in the edge of the float, as before described. The intention being, that when the collimator is employed, the bridge should be in the direction of the meridian, and the guides at right angles to it. Two additional pins are fixed perpendicularly in the float near the guide pins. These are represented in the plan, fig. 7, and are intended to receive some of the weights by which the float is to be adjusted.

The telescope is achromatic, the object-glass in my collimator being about

eight inches focal length, and one inch and a quarter aperture. The objectglass is fixed in a separate piece of tube, which slides within another tube, to which, after adjustment, it may be firmly attached by two opposite screws moving in longitudinal slits.

In the focus of the object-glass of my collimator, a diaphragm is placed, carrying fine cross wires flattened. These wires, however, do not form angles so neat as could be wished, in consequence of their thickness, and of a want of perfect straightness of their edges; and I am indebted to Dr. WOLLASTON for the suggestion of a method of constructing a substitute for the cross wires, which has been applied to a vertical floating collimator made for Captain FOSTER, R.N. and which I found to answer perfectly well.

The surfaces of a plate of brass (bell metal would perhaps be preferable), about the twentieth of an inch thick, were ground parallel. The plate was then cut in half, and the surfaces cemented together. One of the sides of this double plate was then formed by grinding, into a salient angle of about 135°, the faces of this angle being slightly bevilled and carefully finished. The plates were then separated, and their opposite surfaces cemented together, the angular points and edges being made accurately to coincide. One of the sides (not that opposite to the angle) of this compound plate was then ground at right angles to the surface, and the plates again separated and carefully cleaned.

A circular piece of glass having been fitted into the diaphragm, a bit of brass with a straight edge was cemented to the glass on one side by shell lac. This was at such a distance from the middle of the diaphragm, as to allow the angular points of the brass plates to be in its centre, when their ground sides were in contact with the straight edge.

One of the brass plates was now attached by shell lac to the glass, having its ground edge in contact with the brass straight edge, the obtuse angle of the bevilled edge being next the glass, and the angular point in the middle of the diaphragm. It is now evident that if the ground edge of the other brass plate be placed in contact with the straight piece of brass, the obtuse angle of the bevil being next the glass, the two angular points will, upon sliding the brass plate along the straight edge, be brought accurately into contact, and in this position the brass plate was fixed to the glass by shell lac.

This arrangement afforded an object having well defined acute angles of

about 45°, which could be bisected with great precision. As this is a very important part of the instrument, I have been induced to describe thus minutely the manner in which it may be made. The diaphragm is placed in the telescope, with the brass plates next to the object-glass, in order that when the edges of the plates are seen with perfect distinctness, no particles of dust may be visible which may have lodged upon the surface of the glass. The angular point is brought into the centre of the tube in the usual manner, by means of opposite screws passing into the diaphragm. A piece of plane glass closes the end of the telescope, to keep out dust and to protect the diaphragm from accident. The diaphragm is represented at fig. 10*.

When observations are made with the floating collimator, it is necessary that all light should be excluded, except that which passes through its telescope. For this purpose a piece of sheet iron is formed, as represented at fig.11. The aperture is the same as that of the iron basin of the collimator, and one side externally is straight, in order to form a hinge upon which the plate B moves. This plate may be raised to a perpendicular position, and when permitted to fall, rests steadily upon the ring. In the centre of the plate B, is a circular opening, rather less in diameter than the telescope. The three projecting parts of fig. 11. are intended, when bent close to the ring, to form legs, to be attached to the mahogany support, fig. 1, the length of these legs being such as to allow the plate B, when resting on the ring, to be quite clear of the telescope of the collimator.

For the purpose of illuminating the diaphragm of the collimator, a small plane mirror is used, similar to that of a microscope. This is attached to the plate B over the opening, by a short firm pedestal, the manner of doing which is so obvious, as to need no figure. The instrument is represented in Plate XIV. The collimator being thus completed, it will be found convenient to adjust the object-glass of its telescope very nearly to the distance of its focal length from the diaphragm. For this purpose, any telescope may be employed which has been carefully adjusted to distinct vision upon a fixed star. The telescope of the collimator is to be supported in a convenient position for looking into it through its object-glass with the other telescope, and a lamp so placed as to illuminate the diaphragm. All false light being excluded from the telescope

* Since the above was written, I have found that a cross of strong spiders threads answers very well.