THEORY OF THE REFLECTING EXTENSOMETER. 95 
although the formula given for finding the extension from the 
scale readings is, as Professor Warren indicates, but approximate, 
it is nevertheless as regards precision fairly satisfactory within 
those defined limits. 
Instruments constructed on the same principle may however, 
be made serviceable with larger extensions than those which 
Professor Warren had in view, but in such cases the theory of the 
instrument requires further development. And moreover, for 
the higher readings of the scale in the instrument exhibited, the 
approximate theory is not quite adequate, for its error, though 
absolutely, is not relatively small, and may be entirely eliminated. 
I propose therefore to consider the more rigorous theory of the 
instrument, and to illustrate it by applications to the extensometer 
referred to, which, I may here say, Professor Warren has very 
courteously placed at my disposal for the purpose in view. 
2. Description of the Hxtensometer.—The following _ brief 
description of this instrument will be necessary in order to admit 
of the discussion of its theory being properly elucidated. The 
knife-edge F, see Fig. 1, of a small steel prism FG, lozenge- 
Fig. 1. 
shaped in transverse section, is held against a specimen of material 
EF —the elongation or compression of which under stress is to be 
measured—by means of a small contact piece GHZ, about 2 or 
