o 
o 
of Edinburgh, Session 1884-85. 
But the use of all these methods is very laborious, for the whole 
apparatus has to be opened for each individual reading. Hence it 
struck me that, instead of measuring the compression produced by 
a given pressure, we should try to measure the pressure required to 
produce an assigned compression. I saw that this could be at once 
effected by the simplest electric methods ; provided that glass, into 
which a fine platinum wire is fused , were capable of resisting very 
high pressures without cracking or leaking at the junctions. This, 
on trial, was found to be the case. 
We have, therefore, only to fuse a number of platinum wires, 
at intervals, into the compression tube, and very carefully calibrate 
it with a column of mercury which is brought just into contact 
with each of the wires successively. Then if thin wires, each 
resisting say about an ohm, be interposed between the pairs of 
successive platinum wires, we have a series whose resistance is 
diminished by one ohm each time the mercury, forced in by the 
the pump, comes in contact with another of the wires. Connect 
the mercury with one pole of a cell, the highest of the platinum 
wires with the other, leading the wires out between two stout 
leather washers ; interpose a galvanometer in the circuit, and the 
arrangement is complete. The observer himself works the pump, 
keeping an eye on the pressure gauge, and on the spot of light 
reflected by the mirror of the galvanometer, The moment he sees 
a change of deflection he reads the gauge. It is convenient that 
the external apparatus should be made to leak slightly ; for thus a 
series of measures may be made, in a minute or two, for the contact 
with each of the platinum wires. Then we pass to the next in 
succession. 
I have found this method perfectly successful in practice, 
enabling me to do in an afternoon (and far more certainly than 
before) as much as I formerly could manage in a month. But I 
cannot properly apply it to the compression of water at various 
temperatures (from 0° to 100° C.) until I get a new, light but 
strong, steel compression apparatus, which has been ordered for 
this special problem. 
The experiments hitherto made by this process (to satisfy myself 
that it would work) were made in my smaller compression 
apparatus ; whieh is a massive iron cylinder, and very unwieldy for 
