294 



HISTORY OF SCIENCE. 



wire would then be seen intersecting some other division of the vertical 

 scale, as at o. The angles D E H and F E G being of course equal, the 

 .distance D H would have the same proportion to F G that E D has to 

 E F ; and therefore D H, the amount of expansion, could easily be cal- 

 culated from the known distances F G, E F, ED. In the apparatus 



FIG. 141. 



actually employed by Lavoisier and Laplace, by the movement of the 

 end of the bar at D the optical axis of the telescope was displaced on 

 the scale FG through 744 times the distance D H. The expedient 

 here adopted consists, it will be noted, in magnifying in a known pro- 

 portion the quantity to be measured. Of course the proportion must 

 be very accurately known ; and these expedients often prove illusory 

 from the difficulties of accurately determining this proportion. Thus, 

 in the case before us one of the most essential points is to find the 

 exact length of the lever E D. This is a difficult matter, and may in- 

 troduce serious errors. A more direct mode of measuring the elonga- 

 tion of bars was practised by Ramsden. Three metallic troughs are 



fixed parallel to each other, as at A, 



: : B, c, Fig. 142. A and c contain bars 



of iron about 6 feet in length, and 

 these bars are immersed in ice at 

 32 during both phases of the experi- 

 ments. The middle trough contains 

 a bar of the substance of which the 

 expansion is to be ascertained, and 

 the ends of the three bars are pro- 

 vided with certain optical and me- 

 FlGt I42< chanical arrangements, which are 



partly displayed in Fig. 143. /and 



i are attached tubes, containing a lens which acts like the object- 

 glasses of microscopes ; while g and k contain the corresponding eye- 

 pieces, each provided with cross-wires. At e and h rings are mounted 

 containing cross-wires to serve as fixed marks to be viewed through 



A.I 



