432 ^ PHILOSOPHICAL TRANSACTIONS. "^[ANNOlGgQ. 



j4n Experiment on the Refraction of the Jir. By J. Lowthorp, A. M. 



N°257, p. 339. 

 We took a cylinder of cast-brass, abcd, fig. 3, pi. 9, and cut one end cd 

 perpendicular to the axis acx, the other end ab inclined to it at an angle of 

 about 27° 30', and therefore the perpendicular to this inclining plane, pc, and 

 the axis of the cylinder ax comprehended an angle pea of about 62° 30'. These 

 ends were ground very true on a glass-grinder's brass tool, and each of them 

 was compassed about witli a narrow ferrule of thin brass bbbb. Into the upper 

 side of the cylinder at e was soldered the brass pipe ef, and into the under side 

 at G the other brass pipe gh ; the former of these pipes being about 3 inches 

 long, and the latter 6 inches. Upon the plate ddd were fixed two other plates 

 LL, perpendicular to it, and parallel to each other. Each of these two plates 

 had an arch of a circle, equal to the circle of the cylinder, cut out of its upper 

 edge : so that when the pipe gh was let through a hole near the middle of the 

 plate ddd, the cylinder fell into the arches ; and being fastened there with solder 

 the axis ax lay parallel to the plate ddd, and about an inch and a half above it. 

 The perpendicular end of the cylinder uc was closed with an object glass of a 

 7 V-foot telescope, 00, and the other end ab, with a well polished flat glass i(^ 

 which was carefully chosen to transmit the object distinct enough, notwith- 

 standing its obliquity to the visual rays. The ferrules were well filled with 

 cement round about the edges of the glass, and they laid flat and every where 

 touched the smooth ends of the cylinder, that they might firmly support the 

 pressure of the excluded air. 



Instead of a cistern, as in the Torricellian experiment, we used the inverted 

 siphon of brass, mno, fig. 4, soldered to the plate ggg. One of the sides mn 

 stood perpendicular to the plate, and the other side no inclined to it, and was 

 supported near the upper end o by a little prop kk. 



We then placed the cylinder, as in fig. 5, upon a table which was well fas- 

 tened to a firm floor ; the pipe gh was let through a hole, and the axis laid 

 almost parallel to the sides of the table, and the plate ddd was nailed down to 

 it. The tube of the telescope sss, with the eye-glass in it, was applied to the 

 object-glass, and a hair fixed within it at x, the common focus of both glasses, 

 in the axis of the cylinder continued. Upon the floor, under the cylinder, we 

 nailed the plate ggg, with the inverted siphon upon it, and joined m to h by 

 the insertion of the glass tube x. The joints were very carefully closed with 

 cement ; and then they were covered over with pieces of a bladder, and tied 

 hard with strong thread. There was also a bladder tied below each joint at m, 

 and when it was filled with water it was tied above it at n : so that no air could 



