290 
the line crosses the meridian at an angle of 40° nearly, so that 
I scarcely anticipated any sensible tremors. 
«<J, however, have found that all the phenomena which I 
observed at Dodder bank are reproduced here. 
‘©1, Disturbance of the mercurial horizon. The importance 
of the observations made with this can scarcely be overrated. 
In the case of stars, they give the index correction of a circle, 
the verification of its divisions, and the means of eliminating 
flexion; nor are they less useful for the transit. But another 
application of them, in which the reflected image of a tele- 
scope’s Own wires is made to coincide with the direct one, is 
perhaps even more valuable, and is coming into general use. 
Both require almost absolute quiescence of the mercury; an 
agitation of it, which no other method of examination can ap- 
preciate, will make the image nebulous and confused, or even 
cause it to oscillate round its true place. Even a moderate 
breeze of wind will do this; and therefore, as the amount of 
time during which such observations can be perfectly made is so 
limited, the introduction of any extra disturbance is injurious to 
an extent much more than proportioned to its actual duration. 
Now, I find that with an outgoing train, the image is so much 
disturbed as to make its coincidence doubtful for 4” 36° on an 
average: in one instance even for 6" 17%, at which time, as- 
suming the velocity as given above, the distance must have 
been 11,130 feet. The mercury vessel is the same which was 
used in my former experiments; but the magnifying power is 
the habitual observing one of 250. The time of disturbance 
for the incoming train is seven seconds less. During half this 
time the image cannot be seen, and occasionally the waves on 
the mercury are directly visible to the eye. 
«2, In bisecting a star, or making a coincidence of the 
wire’s images, the circle is moved by a fine screw. However 
perfect its centre work may be, or however complete the adap- 
tation of counterpoises and friction wheels, an amount of fric- 
