the Chloride, Bromide, and Iodide of Silver. 281 



116° C. it undergoes sudden and violent expansion, passing from tlie amor- 

 phous into the crystalline condition ; (d) after undergoing this expansion 

 the mass, on further cooling, undergoes slight expansion, and (e) the 

 coefficient of contraction diminishes as the temperature decreases (or 

 otherwise expressed, the coefficient of contraction augments with the 

 temperature). The following experiments were made in order to examine 

 these effects more minutely, to determine the coefficients of contraction 

 and expansion of the iodide, and to determine the coefficients of expan- 

 sion of the chloride and bromide of silver between as great extremes of 

 temperature as should be found practicable. 



1. Apparatus employed for the determination of coefficients of expansion 



or contraction. 



As mercury attacks the iodide of silver at a temperature far below its 

 boiling-point, and as the beautiful optical method of Fizeau does not 

 seem to be capable of application at high temperatures (his own experi- 

 ments in no case embracing a higher temperature than 100° C), it was 

 considered advisable to seek for some special method by means of which 

 the degree of expansion and contraction of the iodide could be accurately 

 determined. I finally decided to enormously multiply a small motion by 

 a system of levers, to cause short bars of the iodide, chloride, and 

 bromide to act on these levers while being heated between various tem- 

 peratures, and to measure the dilatation by a micrometer-screw. 



The following apparatus was demised in order to carry out this idea. 

 The box A, fig. 1, p. 282 (drawn to scale), contains the levers, shown 

 in plan and section in figs. 2 and 3 ; B is an index attached to the 

 axis X, which is in connexion with the final lever, B mo™g through 

 180° of arc over the graduated half-circle C. D is a sliding bar, one 

 end of which bears upon the first lever, while the other is in 

 contact with a short rod of glass E, which moA'es freely, but water- 

 tight, in a metal stuffing-box. The rod of glass thus enters a brass 

 trough r, which can be filled with water or melted paraffine, and which 

 can be heated by means of a Bunsen's burner. The water or paraffine is 

 allowed to run of£ at the end of an experiment by the tap Gr. H is a 

 rod of glass similar to E, which moves freely but water-tight in the 

 metal stuffing-box I. Between E and H is placed the rod of substance 

 the expansion or contraction of which we desii'e to determine. The end 

 of H external to the trough E rests against the end of the micrometer- 

 screw K, which is tipped with agate, and which moves steadily in the 

 support L firmly clamped to the iron bar M, ^^hich is screwed to the 

 base of the instrument JN". The head of the micrometer-screw, 0, is 

 graduated into 250 divisions, and is figured to 500 ; the graduatious are 

 read off against the cross arm P. A plan of the levers is shown in fig. 2, 

 where D (the end of which is seen in fig. 1) is the movable bar, sliding 

 in a socket E and bearing against the fii'st lever S, which is pivoted at T. 



