the Chloride, Bromide, and Iodide of Silver, 285 



point ; the substance was allowed to solidify, and the whole was weighed- 

 The cavity due to the contraction of the mass in solidifying was 

 then filled with mercury to the level of the mouth of the conical tube, 

 and was weighed. Then, knowing the capacity of the tube, the specific 

 gravity of mercury and of the substance under examination, the tem- 

 perature of the fusing-point, and the coefficient of expansion of the 

 platinum cone, we have all the data requisite for the determination. 

 The real difficulty is the determination of temperatures above the boiling- 

 point of mercury ; and until a trustworthy method of general application 

 has been devised, such determinations must be regarded as approxi- 

 mations. The expressions " below a red heat," " a dull red heat," &c. 

 are still common in text-books and in memoirs ; but the very definitions 

 of what is meant by a " dull red heat " vary, as also do the temperatures 

 assigned to it by different writers. 



3. Effects of Heat on the Iodide of Silver. 



Iodide of silver, carefully prepared by precipitating the nitrate by 

 means of iodide of potassium and subsequent drying*, was fused in a 

 porcelain crucible and cast in warm tubes of glass. The cylinders thus 

 cast were from | to J an inch diameter ; they were levelled at the end 

 by a fine steel saw, and were reduced to 6 inches in length. Every pre- 

 caution was adopted in order to secure homogeneity ; but the rifts, both 

 longitudinal and horizontal, produced during the sudden expansion of 

 the substance in passing from the amorphous to the crystalline condition 

 sometimes rendered the bar so weak and brittle that it had to be recast. 

 I have before pointed out that as a bar of molten iodide cools it contracts 

 regularly to its point of maximum density, then expands considerably 

 just below the temperature of maximum density, passing simultaneously 

 into its crystalline condition, and finally continues to slightly expand as 

 the temperature falls. Its behaviour in the expansion-apparatus may 

 consequently be readily inferred ; and experiment justified the inference. 



A bar of the iodide was placed between the movable glass rods (E H, 

 fig. 1), resting on glass rollers ; water was placed in the trough, and the 

 thermometers were put in their places. The index was placed up to 180° 

 of arc, and the water was heated. As the temperature rose the needle 

 gradually returned towards zero as the bar contracted ; and the amount 

 of contraction was measured and compared with the range of tempera- 

 ture. When the needle reached zero, the bar was allowed to cool, and 

 the needle now moved forward, indicating expansion. The water was 

 now replaced by a paraffine of high boiling-point, and the heating ^^'as 

 recommenced. The index (at 180° of arc) began to retreat as before^ 

 showing that the bar was contracting ; as 140° C. was approached the 



* I must express my great indebtedness to Mr. Valentin for allowing me to have a 

 quantity of the iodide, bromide, and chloride prepared at South Kensington, and tQ 

 Mr. Greenaway for preparing it. 



VOL. XXV. X 



