and Attached Water. 113 



the question, Is this a case of fusion or solution ? is to be 

 answered by the reply, It is continuous with both. 



The above results with this salt-alloy gave encouragement 

 for the examination of nitre alone with water. The thermal 

 relationship of nitre and water from —3° C, the melting- 

 point of the cryohydrate, to the boiling-point of a saturated 

 solution at one atmospheric pressure, about 115° C, has been 

 pretty fully examined by others and myself. The results are 

 shown in Table LVI. The first sixteen are from my memoir 

 IV., § 129; the next five are on the authority of Gay-Lussac; 

 the last seven, which are those to which I wish here to direct 

 particular attention, are obtained by the use of sealed tubes 

 as described in § 240. The result corresponding to T = 115° 

 is interpolated for the sake of verification from Table LVII. 

 § 253, where a different method of experimenting was under- 

 taken for a different purpose. For T = 100° C, the solu- 

 tion saturated at 114° C. was allowed to cool for several 

 hours in a tube surrounded by boiling water (16*05 grams 

 gave 11*485 grams nitre). For T = 109° C, the solution 

 saturated at 114° C. was allowed to cool for several hours in 

 a tube surrounded by a boiling saturated solution of chloride 

 of sodium (17*83 grams of the solution gave 13*2945 grams 

 of nitre). 



§ 252. It must not be overlooked that, although the tubes in 

 the last seven experiments were filled as full of the nitre as the 

 exigencies of manipulation permitted, a certain air-space was 

 unavoidably left into which the water in the tube was free to 

 evaporate. In the last one, indeed, in which water was present 

 its actual weight was only 0*114 gram. This, heated by itself 

 in the free space to 300° C., would have become dry superheated 

 steam. But the fact that it lowered the temperature of solidifi- 

 cation 20° C show r s that in the presence of nitre it is not all 

 free.' The vapour-tensions of strong solutions of nitre will 

 have to be determined. Correlation with this table or with 

 Wiillner's results is all the more necessary, because, as 

 has been shown by Sorby and others, variation in pressure is 

 not without influence on solubility as well as upon fusion, 

 per se. 



Phil. Mag. S. 5. Vol. 18. No. 111. Aug. 1884. 



