45G MR. A. E. TUTTON ON THE THERMAL DEFORMATION OF THE CRYSTALLISED 
suitable for au investigation of the thermal deformation. The double sulphates are 
unsuitable on account of the ease with which most of them lose water of crystallisa¬ 
tion when their temperature is raised, and a similar remark applies to the double 
selenates, whose investigation with respect to their morphological and physical 
properties is now proceeding. The simple selenates offer great difficulties on account 
of their excessively hygroscoj^ic nature, which is so marked in the case of caesium 
selenate, in accordance with the rapidly progressive advance in the solubility of the 
three salts wliich has l)een shown {loc. cit. p. 851) to follow the order of the atomic 
weights of the metals, as to j^lace it in the category of effective desiccating agents. 
The normal sulphate of potassium is absolutely free from this disadvantage, being one 
of the least soluble of the salts usually classed as soluble in water, 100 cub. centims. 
of this liquid at the ordinary temperature only dissolving 10 grams of the salt {loc. cit. 
p. 851 and sulphate memoir loc. cit. p. 632). Rubidium sulphate is so slightly 
hygroscopic, its solubility being only 44 per cent., as to present no difficulty on this 
ground. Caesium sulphate is decidedly hygroscopic, the solubility being so relatively 
great as 163 grams in 100 cub. centims. water. Although this characteristic is by no 
means so strong as in the analogous selenate, the solubility of caesium selenate being 
no less than 245 grams in 100 cub. centims. water, still it is sufficiently marked to 
render the use of the salt for the purjDose in question imjDossible in damp weather. 
The difficulty has, however, been successfidly overcome in the case of caesium sulphate, 
by taking advantage of the driest days of the recent remarkably dry summer, and of 
a few dry frosty ones of the early winter, together with the expedient of utilising the 
inner chamber of the air bath of the dilatometer as a desiccator, by placing a vessel 
containing oil of vitriol therein until the actual moment of commencing the obser¬ 
vations. 
In the present memoir, therefore, are presented the results of an investigation 
of the thermal deformation of the orthorhombic normal sulphates of j^otassium, 
rubidium, and cmsium. It is scarcely necessary to remark that the series of these 
particular three metals lias been chosen throughout the whole of the author’s work 
on the relations between the chemical composition of salts and the properties of their 
crystals, because of their well-established close relationship, as being in the strictest 
sense members of the same family group of the periodic system, the definitely 
established and relatively large difterences between their atomic weights, and the 
fact that they form the most strongly electro-positive series of elements. 
Pre'paration of the Crystals. 
Although the new compensation method does not require crystal blocks of greater 
thickness than 5 millims., the greatest difficulty has been experienced in obtaining 
crystals of the commonest of the three salts, potassium sulphate, of adequate thickness 
in all three of the axial directions along which measurements of expansion or con- 
