NORMAL SULPHATES OF POTASSIUxM, RUBIDIUxM, AND CHESIUM. 
4(J1 
milled collar, Ijy the screwing of which down the cone the three portions of the 
tubular holder are com23ressed together. The latter are padded inside with broad¬ 
cloth, within which, on rotating the collar, the crystal is firmly clamped. This holder 
is suitable for the gripping of prismatic crystals. The other two, which difier only in 
the size of their apertures, somewhat resemble miniature tuning forks, the stems being 
of the same size as those of the other holders and similarly grooved to fit the ribbed 
socket at the base of the crystal-adjusting apparatus. The two prongs are in each 
case relatively Ijroad, 1‘3 centim., are padded inside with broad-cloth and can ])e drawn 
together so as to eftect the grip by means of a milled-headed clamping screw, passing 
loosely through one i^rong and screwing firmly through the other. One of the forks 
takes crystals 5 millims. thick and the other takes crystals up to 1‘2 centim. thick. 
If the crystal is not sufficiently tabular for direct gripping by one of tliese two latter 
holders, and not sufficiently eveidy prismatic for the advantageous use of the split-tube 
grippei', it is })acked in a rectangular Ijlock of cork iield in the larger forked liolder; 
the cork can be cut with a sharp penknife so as to accommodate the most incon¬ 
veniently shaped crystal in the position required for grinding. When the jjrongs are 
screwed together as much as possible so as to tightly grip the cork setting and tlie 
contained crystal, the latter is found to be rigidly held without any danger of 
cracking, the cork lending itself to an even distribution of pressure. These grip- 
holders were devised in order to avoid the frequent cracking which large crystals 
sufi'er when warm opticians’ wax is employed to cement them to the ordinary holders. 
The importance of this point is obvious, when it is remarked that a cracked crystal is 
totally unsuitable for use in determinations of thermal expansion Ijy the interference 
method, the crack being sure to develop further during the observation and derange 
the interference bands. 
Another addition is a special crystal-adjusting apparatus, shown resting on the table 
to the right in the illustration, intended for use in preparing 60° prisms for refractive 
index determinations, in cases where it is a difficult matter, by reason of delicjuescence 
or other rapid deterioration of the substance under Investigation, to prepare tlie two 
inclined surfaces by separate settings on a crystal-holder enq>loyed Avith the ordinary 
adjusting apparatus. In order to pre])are tAvo surfaces inclined at 60° by one setting 
of the crystal, it is obviously necessary to rotate the crystal for 60° on each side of 
the })articular principal optical plane Avliicii has been adjusted vertical to the grinding 
plane, Avith the aid of the goniometrical arrangements provided on the instrument, 
and to Avhich optical j)lane the tAvo recjuired surfaces are to be symmetrical. The 
adjusting apparatus pi'ovided for ordinary purposes is similar to the second one 
described in the memoir on the first and smaller pattern grinding goniometer, intended 
for use in grinding small artificial salt crystals (‘ Phil. Trans.,’ A, 1894, p. 895), but 
of larger size, corresponding to the larger instrument. It includes tAvo cylindrical 
adjusting movements provided Avith divided silver arcs and indicators, and a divided 
horizontal circle betAveen them to enafjle tlie loAver nio\'ement to be set at any desired 
