226 
Proceedings of the Pioyal Society 
greater for tlie smaller tension — precisely as Joule’s results would 
lead one to expect. This effect of tension seems to have escaped 
the notice of Wiedemann, but it was undoubted. 
In the case of nickel, the direction of the twist was different from 
that for iron, a conclusion which Barrett’s discovery of the contraction 
of nickel under magnetisation rendered not unexpected. No maxi- 
mum twist was reached for an intermediate current. Otherwise the 
results were the same as for iron, as, for example, in the case of the 
tensions. 
The experiments were tried with different thicknesses of Avire, 
and all gave the same conclusions. 
3. Further Note on the Maximum Density Point of Water. 
By Professor Tait. 
During my long investigations of the ‘‘ Pressure-Errors of the 
^Challenger’ Thermometers,” but more especially two years ago {Pro- 
ceedmgSj May 1881), I was led to suspect a lowering of the maximum 
density point of water by pressure. For I found that the change 
of temperature of water increased faster than in direct proportion 
to the sudden change of pressure which produced it. These ex- 
periments were, at my request, more fully carried out by Messrs 
Marshall, Smith, and Omond {Proceedings^ July 1882). Their 
result was (approximately) a lowering of the maximum density point 
by 5° C. for 1 ton-weight of pressure per square inch (roughly 
speaking, about 150 atmospheres). In a note appended to their 
paper I deduced from their experimental data a lowering of 3° ’6 C., 
and from my own a lowering of about 3° C. for the same pressure. 
In November last, when reducing the observations made in the 
“Triton” (ante, p. 45), I found that water becomes more compressible 
as its temperature is loAvered, at least down to 3° C. ; and this I 
regarded as another indirect proof of the lowering of the maximum 
density point by pressure. 
I then determined to try a direct process analogous to that of 
Hope, for the purpose of ascertaining the maximum density point 
at different pressures. The experiments presented great difficulties, 
because (for Hope’s method) the vessel containing the water must 
have a considerable cross section, and thus I could not use my 
