the Chhride^ Bromide, and Iodide of Silver, 283 



perfect continuity o£ the parts is established, and the micrometer-head is 

 turned until the needle stands at zero ; the reading on the micrometer is 

 registered, and the head is then turned until the index B has passed to 

 180° of arc ; by again reading the micrometer, the relation between the 

 micrometric divisions and the divisions on the scale of C is established. 

 By this means it was found that a movement of '0035 inch in the 

 micrometer moves the index through 180° of arc. Xow since the index 

 is six inches long, in moving through half a circle its extremity passes 

 over 18-84 inches, and this motion is produced by a moAxment of -0035 

 inch ; hence the levers multiply any motion communicated to them 5382 

 times. The precise value of this index-scale having been ascertained, the 

 determination of the expansion of a body by heat becomes an easy matter, 

 provided that we are careful to secure perfect rigidity in all the fixed 

 parts of the apparatus (by no means an easy task), and are further 

 careful to prevent the conduction of heat from the short glass rods (E and 

 H) to the interior of the apparatus. The coefficient of expansion of E 

 and H must also be ascertained, and allowed for in all determinations. 

 If, on the other hand, we wish to determine the contraction of a body, 

 the index is pushed up to 180° of arc by means of the micrometer- screw, 

 and is pushed back to zero as the bar contracts by the spriug T and the 

 hehcal mainspring Z. The precise value of the return movement must 

 be, of course, ascertained by means of the micrometer-head. It is obvious 

 that in moving from 0° to 180^ of arc the spring T acts against the 

 motion of the expanding body ; while in moving from 180° to 0° it moves 

 with the motion of the contracting body. 



A bar of fine homogeneous silver was used in order to test the delicacy 

 of the apparatus. It was placed between the short glass rods (E and H), 

 and cold water was poured into the trough. The temperature was in- 

 dicated by two thermometers reading well together, placed near the 

 opposite extremities of the silver bar, and the needle was brought to zero. 

 The water was then heated, being constantly stirred to produce uni- 

 formity of temperature throughout the whole mass, until the needle had 

 been forced round to 180° of arc. The number of degrees of heat being 

 read off, showed the heat necessary to produce a certain known expansior, 

 and from this the coefiicient of expansion was easih'- ascertained. Six 

 results gave 10° C, 10°, 9°, 10°, 10°, 9°-5. The coefiicient deduced from 

 the mean of the determinations was for 1° C. ■ 



•0000193083. 



Now Lavoisier gives it as -00001910 ; Daniell -00001951 ; Eizeau (deter- 

 mined by the same method as that which he applied to the iodide, chloride, 

 and bromide of silver) gives -00001921 ; and more recently Matthiessen 

 (Phil. Trans. 1866) found it to be -000019436. 



