In the following table the expansions are marked in parts, of which each is the part of an inch : 2. Desaguliers' Improvement on ilfzischenbroek's meter.
Dr. Desaguliers made several important improvements on this instrument. In place of square rods of metal, Dr. Desaguliers used cylindrical ones, B N, Plate CCCCLXXI. No. II, Fig. 4, as they could be made more uniform, by being drawn like wires. Instead of the pinion F, he used a small roller FI, made of steel, truly tempered, but not polished, and filed on its surface, in the direction of its axis, so that it became a small wheel, with infinite number of teeth. The wheel gg, on the same axis, has no teeth, but has a groove, to receive a fine watch chain, or a horse's hair, which carries round a roller i, having also a small groove. Upon the upper end of the axis of this roller the index k i 72 is fixed. In order that the chain by which gg carries i may be properly extended, the whole dial plate, and the cock and pinion i, can be moved to or from the wheel gg by a screw fastened to the upper frame plate. In stead of the rack NL, Dr. Desaguliers substitutes a long thin plate of steel LN, about of au inch broad, and tiled roughly, so as to move the first roller H, by rubbing against it. It is spring tempered and is a little convex towards but w hen it is fastened to the rod at N, there is a spring fixed to the lower brass plate, which draws it straight and tight by the end L, in the direction NL. Instead of the cocks there are two pulleys I', P, placed horizontally, whose broad vertical grooves receive and direct the steel plate or roller LN, that is substituted for the rack. In place of the watch spring to support the bars, Dr.'Desa gulicrs used a small brass roller, four tenths of an inch in diameter, having its axis horizontal. This roller is raised tip by a screw Q so as to support any bar of metal at its end N. By means of these alterations the sticking of the teeth is prevented, and the motion of the index commences at the instant that the heat is applied to the metallic rods. See Desagulicrs' F.xperimental Philosophy, vol. i. p. 444.
Description of Mr. .Ellicot's Pyrometer The pyrometer, as constructed by Mr. Ellicott, is repre
stnted in Plate CCCCLXXI. Fig. 5, where AA is a flat plate of brass screwed down to a thick piece of mahogany. Upon the plate is screwed two pieces of brass, two of which B, B support the flat iron bar C, called the standard bar. The upper part of the third piece of brass D is a circle, about three inches in diameter, divided into 360 de grees, and within the circle is a moveable plate d, divided also into 360 degrees, and a small steel index. The bar of metal E, upon which the experiment is to be made, rests upon the standard bar C. A bar, two inches and a half long, is fixed to an axis, which turns on two pieces of brass screwed to one of the supports B; and to the end of the lever is fastened a chain or silk line, which, after being coiled round a small cylinder, to which the index on the brass circle D is fastened, passes over a pulley, and has a weight hung to the end of it. Upon the axis which car ries the lever is a pulley one-fourth of an inch in diame ter, to which a piece of watch chain is fastened. The other end of the chain is hooked to a strong spring G, bearing against one end of the metallic bar E. There is another lever H exactly similar to F, but the chain fastened to the pulley on its axis, is hooked to the standard bar. The line fastened to the end of this lever, 'after being coiled round a cylinder, to which the moveable plate is fixed, passes over a small pulley, and has a weight hung to the end of it ; or the same line, passing under a pulley, to which the weight is hung, has its other end fastened to the lever F, so that one weight will serve for both levers, as in the figure.
When the bar E lengthens by heat, it allows the weight to draw the lever F upwards, by its action on the spring G, and, by means of the silk line, the index will be, at the same time, carried forward in the circle. When the bar E contracts, the index will return back again, and the same motion will be communicated to the standard bar. An elongaton of the bar E one-twentieth of an inch cor responds to 360° or one revolution of the index, and one degree to the 7200th part of an inch.