BY MIXTURES OF ELECTROLYTES.—BARNES. 145 
This inner freezing-tube was 28 cm. in length and 2.8 em. in 
diameter and had its lower end re-entrant. The outer tube was 
25 em.in length and 3.15 cm. in diameter. The thickness of the 
glass being about 1 mm., there was an air space of about 1.5 mm. 
between the tubes. This space was found to be quite sufficient 
to prevent the formation of ice on the wall of the tube. The 
inner tube was supported in the outer by means of two rubber 
bands, one at the top and the other at the bottom. These bands 
also prevented the walls from touching one another. The 
length of the tubes allowed the solution to be well submerged in 
the protection bath and therefore almost freed it from the 
influence of the outside temperature. 
In the determination of the freezing-point of a solution these 
tubes were surrounded by a mixture of salt water and pieces of 
ice, contained in a vessel of glass 35 em. high and 11 em. in 
diameter (called the protection bath). The cover for this vessel 
was one taken from the protection bath of a Beckmann apparatus, 
The glass of this vessel was $ em. thick and was covered with 
asbestos paper that the effect of the temperature of the room 
might be lessened. It was found necessary to keep this bath 
at a constant temperature within a fiftieth of a degree, if values 
of the freezing-point agreeing with the mean value to less than 
a thousandth of a degree were desired. This was obtained by 
keeping the bath continually stirred. Any change of tempera- 
ture was quickly recorded by means of a thermometer graduated 
to a fiftieth of a degree. The temperature of the bath could be 
raised by the addition of water at the temperature of the room, 
or lowered by the addition of pieces of ice, snow when obtain- 
able being preferable. 
The platinum stirrer for the freezing-tube was of the common 
ring shape, having wound around its ring a thin platinum wire, 
which would rub against the wall of the tube and thereby 
prevent the formation of an ice sheath. With constant stirring 
there was found no tendency for the ice to mass itself together 
and float to the surface, but it could be seen moving through the 
whole solution in tiny glistening particles. The stirrer for the 
Proc. & Trans. N. S. Inst. Sor, VoL X. TRANS. J. 
