520 



HISTORY OF SCIENCE. 



mercury is passed back into F. The lengths of mercury in A B act as 

 so many pistons, and as the exhaustion proceeds, the air-spaces between 

 them are observed to get smaller and smaller, and at length the descent 

 of the mercury takes place with the clicking sound peculiar to all 

 liquids shaken in a vacuum. Wh en the mercury falls without enclosing 

 tne smallest bubble, if the tub e at G be clamped, the column in A B 

 will stand at the full height of the barometer. The exhaustion can 

 by this apparatus be carried beyond the millionth part of an atmo 

 ?re, and even the ^1^ th part has been attained. The most per 

 feet vacuums possible are, however, pro- 

 cured by using tubes filled with carbonic 

 acid gas, and after this has been exhausted 

 as far as possible by the Sprengel pump, 

 the residual gas is absorbed by caustic 

 potash. 



, It was in general in vacuums obtained 

 Toy use of the Sprengel pump that most of 

 Mr. Crookes's experiments were con- 

 ducted. He was led on to try one form 

 of apparatus after another. One of the 

 earliest forms had a straw carrying a pith 

 ball at each end. balanced in the tube 

 on pivots formed by passing through the 

 straw, a little above the centre of gravity 

 of the system, the pointed end of a fine 

 needle broken oifa very little shorter than 

 the interior diameter of the containing 

 tube, and pointed very finely at the broken 

 end. The needle was thus supported very 

 delicately by its points against the sides 

 of the tube, and with the least possible 

 friction. This form of experiment was 

 superseded by an instrument on the same 

 principle as the torsion balance (p. 335). 

 The straw beam was supported horizon- 

 tally by means of a single fibre of cocoon 

 silk, or of an extremely delicate filament 



ot glass, and the movement was indicated by a reflected ray of 

 light, as in Thompson's " reflecting galvanometer." The horizontal 

 beam was made to carry various substances, and the amount of the 

 deflecting force under assigned circumstances was noted in a great 

 number of cases. Finally, Mr. Crookes was led to a form of apparatus 

 which has since been the subject of many experiments, and is now- 

 well known as the " Radiometer." Its ordinary form is represented in 

 Fig. 239. A is the exhausted vessel, in which a vertical spindle cb turns 

 freely upon its point at b. The spindle carries four arms at right 



FIG. 239. THE RADIOMETER. 



