248 
FISCHER. 
Paris. In order to test whether this had occurred, the re¬ 
flection of a straight rod or string from the two surfaces was 
viewed. If the thread or rod appeared continuous, except 
where it was interrupted by the space between the two pieces 
of glass, the appliance was considered ready for use. 
The use of the arrangement was likewise very simple. The 
upper parts of the contact surfaces, namely, the parts above 
the line surfaces, were placed in the opening between the 
pieces of glass, as shown in Fig. 10, and when the line sur¬ 
faces were in contact with the glass the pieces were clamped 
together with a U-shaped spring, after which they were 
mounted on an adjustable tripod under the microscope. By 
means of the tripod the plane of the two surfaces was made 
J 
1 
G 
G 
_ 
E 
Fig. 10. 
perpendicular to the axis of the microscope, and the distance 
between the two lines was then ready to be determined. A 
slight error, always operating to make the distance between 
the lines too short, would result from small inclinations of 
the line surface. 
To determine what inclination would be permissible, let d 
equal the true distance between the lines, which is approxi¬ 
mately 160(F, and a the angle of inclination. Then the 
error would be d cos a — d; or, d (cos a — 1). Substituting 
for «, 10', 20', 30', we get the following errors: 0.01, 0.03, 0.06 
microns. 
It would be an easy matter to keep the inclination below 
10', and I am satisfied such was the case in this work. 
