96 G. H. KNIBBS. 
3 mm. in thickness, about 9 mm. in breadth, and from 30 to 200 
mm. in length, suitably held in position. Usually there is a con- 
tact piece symmetrically placed on each side of the specimen, both 
being held in place by means of a simple spring clamp, as shewn 
in Fig. 3 illustrating Professor Warren’s paper. The other 
knife-edge G of the prism rests in a groove in the contact piece. 
Attached to the prism is a mirror M/, which can be rotated on its 
pivots PP’ by means of a screw C not shewn, working against @ 
slight spring A, and by means of this mirror a scale 7UJ, is read, 
the reading being determined by the way in which the sight line 
of the telescope 7’ meets the plane of the mirror.1_ On the appli- 
cation of the stress to the specimen, the prism, and consequently 
the mirror attached to it, are rotated; the result being that 
the scale reading is altered. The difference of these readings is 
the datum from which may be deduced the amount which the 
knife-edge / has shifted, that is the amount of extension oF 
compression which the applied stress has produced. The scale, 
an ordinary millimetre scale with black lines on a white ground, 
is set approximately perpendicular to the sight line of the telescope, 
and can be rotated in that plane through any angle. The 
telescope is of short focus; has an objective of 28°5 mm. clear 
aperture ; its focal limits are 5030 and 868 mm., the conjugate 
focus at the greater limit being about 256 mm., and at the less 
about 348 mm.? The ocular of the telescope is an ordinary 
Ramsden or positive, of considerable magnifying power ; the cross- 
wire diaphragm is susceptible of slight rotatory adjustment, 80 
that the wires may be set horizontally or vertically. By means 
of slow motion screws the telescope can be moved either horizon- 
tally or vertically through small arcs. Other features of the 
instrument will be referred to when dealing with the question of 
its adjustments and use. 
1 See Figures 2, 3, and 4, Professor Warren’s paper. 
2 The distances are only roughly measured. The relation ;\-5 + s48 
= hs +5}% should hold good. The results are ‘00411 and -00403, the 
accordance is evidently satisfactorily. 
