4 
MR, H. M'LEOD AND LIEUTENANT G. S. CLARKE ON THE 
To prove that no great error can have crept into the calculations from any inaccuracy 
' the observations 
Gaining rate 
per day. 
seconds. 
2-463 
2 - 232 
2-282 
2-297 
2-673 
2- 476 
3- 098 
2-504 
The error of the clock rate when checked was not more, and probably less, than one 
second per day. 
The apparatus used for the measurement of the tuning-forks was made from a 
portion of an astronomical chronograph constructed by Messrs. It. and J. Beck. The 
brass drum, which is two feet long and one foot in diameter, rests on friction wheels and 
is driven by a weight and train of wheels. The original regulator, consisting of a 
conical pendulum, w T as removed and its place supplied by a fan kindly designed for us 
by our friend and colleague, Professor W. C. Unwin, B.Sc. The fan and case are 
shown full size in Plate 2, fig. 4. The diameter of the fan is 3*8 inches, and it is 
attached to a vertical steel axis resting in an agate cup. The fan rotates in the lower 
part of a circular box 5 inches in diameter and 2 inches deep, the box being divided 
into two horizontal compartments by a fixed diaphragm d d, the upper compartment 
being '75 inch and the lower l - 2 inch deep; the diaphragm is perforated by four 
segmental apertures a a which can be closed by a similarly perforated disk resting on 
the diaphragm and capable of being moved by an arm r on the top of the box. When 
the apertures in the diaphragm are opened, air is driven from the lower compartment 
into the upper, returning through the openings in the middle of the fan, and when the 
apertures are closed the air circulates with the fan in the lower half of the box. In 
the latter case the fan rotates about twice as fast as when the diaphragm is fully 
opened. The axis carries at the top a small pulley p by which it is connected to the 
drum of the chronograph by a piece of thread passing round the latter. 
Bound one end of the drum, strips of dark blue paper are wuapped, on which white 
equidistant lines parallel to the axis of the drum are ruled. The paper was graduated 
of the chronometer, the following table is given showing the dates of 
and the chronometer rate in the interval. 
1878. Jan. 16 
Gaining rate 
per day. 
seconds. 
1878. May 4 
Jan. 19 = 
► 2-91 
May 14 
Jan. 23 
> 2-527 
June 3 
Feb. 25 - 
- 2-741 
June 11 
Mar. 3 - 
2-657 
> Q.QQQ 
June 19 
Mar. 10 = 
> Z oZo 
June 22 
Mar. 12 « 
2-270 
June 27 
April 6 : 
> 3-394 
July 15 
April 22 i 
> 3-528 
Aug. 7 
May 4 j 
► 3-546 
