262 | A. BOYD. 
knife-edge is being moved. In this case, the movement of 
the knife-edge on the glass scale was added to the reading 
on the telescope scale just before the knife-edge was moved. 
Then, if there had been no deflection of the rim while the 
knife-edge was being moved, this should give the reading 
of the telescope scale after the knife-edge had been moved. 
This usually was not the case, the difference being the 
deflection of the rim during the movement, e.g. zero reading 
of telescope scale was 49; deflection took place down to 2. 
The knife-edge, which was at 4°30, was shifted, making 
the reading on the telescope scale 40. The knife-edge was 
found to be at 470 on glass scale, i.e. it was moved 40 
divisions. Then 40 + 2 should have been the reading on 
the telescope scale after the movement. But it was found 
to be 40, so that the deflection during the movement was 
two divisions. 
Mr. Barraclough and the author carried out a number 
of preliminary experiments in order to arrive at a satis- 
factory form of distance rod. Three forms of rod were 
tried (see figs. 3,4 and 5). These were all constructed of 
steel, the first and third being attached to the rim by a 
pair of chisel points, while the second was held by a 
hardened steel set screw. The distance between the chisel 
points in the case of the first rod was 0°6 inch, and in the 
case of the third 0°45 inch. Great care was taken in 
balancing the rods both in a horizontal and vertical plane; — 
the rods (figs. 4 and 5) were counterbalanced in a vertical 
plane by means of pieces of metal soldered to their upper 
faces, making them symmetrical, while the rod (fig. 3) was 
counterbalanced by means of a coil of copper wire-(equal’ 
in weight to the terminal bent portion of the rod) wound 
on the upper part of the small vertical steel bar which 
acted as a safety arrangement to prevent the rod from 
slipping off the rim. | Lh 
