1875.] Physics. 123 
of water falls 0°007° C. for every additional atmosphere of pressure. To deter- 
mine the true freezing-point, take a glass tube closed at one end, 20 centimetres 
long and 2 wide, fill it almost full with sulphuric acid, and heat. Then pour 
out the acid, and rinse repeatedly with pure distilled water. The tube is then 
two-thirds filled with distilled water, which has been boiled for some time ina 
clean beaker, and a small quantity of filtered oil of turpentine (about 1 centimetre 
in depth) is poured upon the water. The tube is then carefully heated in the 
oil-bath, without allowing the temperature to rise to the boiling-point lest an 
explosion should ensue. The object of the heating is to remove any air bub- 
bles which may adhere to the side of the glass or may remain between the 
turpentine and the water. When the water has been thus exposed for a 
considerable time to a temperature very near to the boiling-point, the tube is 
taken out of the oil-bath, cooled in cold water, and then placed in a freezing 
mixture (water and nitrate of ammonia). After a few minutes the water is 
cold, and in most cases a portion of it freezes at once if a thermometer is 
inserted, and moved up and down. If this does not take place the tube must 
be returned to the freezing mixture, and cooled more strongly. The ther- 
mometer may be previously placed in an empty test-tube, which is then plunged 
in the freezing mixture. It is very important that the thermometer should be 
cooled down close to the freezing-point before being introduced into the water. 
The best thermometers when tested in this manner show a freezing-point too 
high by about or1° C. 
Evecrriciry.—In a paper read by Prof. W. F. Barrett, F.R.S.E., M.R.1.A., on 
the “ Molecular Changes that accompany the Magnetisation of Iron, Nickel, 
and Cobalt,’’ at the British Association Meeting, at Bradford, it was shown 
that certain molecular phenomena known to attend the magnetisation of iron 
were found to accompany the magnetisation of nickel and cobalt. Notably 
this was the case with the peculiar sound emitted on magnetising and demag- 
netising these metals ; with cobalt the note was clear and metallic, and louder 
than in the case of iron. The physical as well as the chemical properties of 
the three magnetic metals were, in every case, proved to be so closely similar 
that it was reasonable to expect that any cause producing a molecular change 
in the one metal would produce a similar change in the other. A series of 
preliminary experiments established the fact that a change in the dimensions 
of nickel and cobalt occurred on their magnetisation by an electric current, 
corresponding to the elongation and retraction of iron investigated by Dr. 
Joule. To pursue this enquiry further a committee was appointed, consisting 
of Prof. Balfour Stewart, and subsequently Prof. Clerk Maxwell, in conjunction 
with the author, and the results were embodied in areport read before Section A 
Fia. I. 
of the British Association at Belfast ; after the trial of various arrangements an 
instrument was devised, and has been constructed by Messrs. Yeates, of Dublin 
and London, by which it is hoped precise determinations may shortly be made 
