NORMAL SULPHATES OF POTASSIUM, RUBIDIUM, AND CESIUM. 
457 
traction by heat were desired. By the slow evaporation of cold saturated solutions 
over oil of vitriol in vacuo it is possible, given adequate length of time and sufficient 
amount of solution, to obtain excellent crystals of the more soluble sulphates of 
rubidium and cresium, of the requisite size to furnish blocks from 5 to 10 millims. 
thick in the three axial directions. But in the case of the sparingly soluble potas¬ 
sium salt the crystals are almost invariably small. Moreover, when by exception 
they are larger, they are either pseudo-hexagonal triplets or other twinned forms, 
useless for the purpose in view owing to the unequal expansion of the interpene¬ 
trating parts due to different axial expansion in accordance with orthorhomhic 
symmetry; or otherwise, they are individuals of an elongated prismatic nature, the 
elongation being in the direction of the axis ct and the jDrisms being too narrow for 
use along the other two axial directions, particularly that of the axis h. It was 
found exceedingly difficult to induce well-formed individual prisms, deposited from 
a large quantity of a cooling saturated solution, to grow further to the required trans¬ 
verse dimensions in cold saturated solutions over vitriol in vacuo. After attaining 
a thickness of aljout 4 millims. it almost always happened that, rather than grow 
further, fresh crystals began to be deposited. After months of fruitless labour, 
Messrs. Hopkix and Williams kindly undertook to attempt to obtain larger crystals 
by the use of very much larger quantities of solution, and eventually succeeded in 
producing seven crystals of exceptional thickness, and which, after a little further 
growth in a cold saturated solution over vitriol in vacuo, have at length yielded 
transparent blocks 6 A to 9’6 millims. thick along the direction of the axis c and 5'1 
to 5 "9 millims. along the axis h. The author desires to exjiress liis great indelited- 
ness to the firm in question for so kindly placing their resources at his dis])osal, and 
thus enabling this investigation to be completed. 
Peepakation of the Paeallel-Faced Crystal-Blocks. 
Ini207'ovements on the Cutting and Grinding Goniometer. 
The preparation of the crystal-1 docks, each provided with a pair of truly plane and 
truly parallel surfaces accurately perpendicular to the particular crystallograpliical 
axis along which the linear expansion or contraction was to be measured, was carried 
out with the aid of the author’s new cutting and grinding goniometer. The instru¬ 
ment in question is similar in principle and general appearance to that which was 
described to the Royal Society in December 1894, (‘Roy. Soc. Proc.,’ vol. 57, p. 324), 
and which is now in the National Collection in the South Kensington Museum. It 
differs from the latter instrument in including a few slight improvements which 
prolonged use has shown to be advisable for the sake of greater convenience and ease 
in manipulation. The perfected instrument affords the highest satisfaction, enabling 
the most accurately orientated and truly plane surfaces of crystals of any degree of 
VOL. CXCII.-A, 3 N 
