VELOCITY AND PRESSURE INSTRUMENTS, 
363 
Should the disjunctor readings become irregular, the following points must 
be looked to, viz:— 
(1) If one of the magnets has not become too strong. 
(2) If the instrument is properly levelled; that is, if the rods hang 
vertically, and do not rest too heavily against the support. 
(3) If there be not an imperfect connexion in the circuits, including the 
battery and the screens. 
Should the mark obtained be indistinct, the chronometer must be brought 
nearer to the knife (by means of the levelling screws), care being taken that 
it still falls freely and without friction. 
If, during the experiments, one of the currents becomes broken without 
any apparent cause, try (after ascertaining that the screens are properly 
mended) whether there is contact between the plates, c[ of the disjunctor 
(Plate II., Pig. 9) and the pins, r r' } by removing the wire from the binding 
screw at one extremity of the plate, and bringing it into contact with the 
screw at the other end. If the current is thus re-established, it show r s that 
the break occurs at the point of contact of this plate with the screw, w r hich 
should be cleaned by passing a piece of paper between them. 
The only parts of the instrument that require special attention are the 
conical points of contact between the rods and the electro-magnets. 
These four points ought to be kept clean and polished, and it is as well 
never to touch them with the fingers, and to rub them frequently with a 
chamois leather. The rest of the instrument may be covered with rust and 
dirt without affecting its working, whilst a single spot of rust on one of these 
points may cause irregularity in the disjunctor readings. If by accident they 
should get rusty, very fine emery cloth must be used to clean them, care 
being taken to rub round the point so as not to alter its form. 
Prom the nature of the instrument itself, nothing can affect the chronometer 
while falling, and the rate of falling being according to a well known and 
invariable law, it is evident that there can be no constant error in the measure¬ 
ment of velocities, provided that the scale is correctly graduated, and the 
disjunctor is in proper working order. 
The accidental errors which may be committed correspond to those which 
occur when a series of disjunctor readings are taken, after the instrument 
has been properly regulated; and any one at all accustomed to using the 
instrument will see at once that the errors in determining velocities (including 
errors in reading the scale) do not exceed a few thousandths of an inch, and 
that the results are sufficiently accurate for ordinary experiments, the varia¬ 
tions being far less than those due to other causes. 
If the two currents are not broken by the disjunctor identically at the 
same instant, there will be a constant error in the readings. The disjunctor 
is not liable to get out of order, but, if required, its accuracy can at any time 
be verified as follows :— 
Determine the height of the disjunctor reading, and then invert the currents 
by removing the wires which were first at s and v to s' and v', and those at 
s' and v’ to s and v (Plate II., Pig. 9), so as to send the chronometer current 
through the side on which the registrar current first passed, and vice versa. 
Having done this, take several readings, and ascertain whether they agree 
