148 ROYAL SOCIETY OF CANADA 
specimens of ice formed under different conditions, the artificial ice 
experimented upon being formed similarly to the ice mantles formed 
in the Bunsen Ice Calorimeter. 
In reviewing these experiments, it will at once be seen that there 
are three general methods of determining ice density, viz., equilibrium, 
dilatometric and weighing methods. 
In the first method, obviously, great accuracy cannot be obtained. 
Of the two experimenters using this method, Dufour obtained by 
far the most accurate results. The advantages of the method are that 
both artificial and natural ice can be experimented upon, and the deter- 
minations can be repeated upon the same specimen as often as desired. 
However, these advantages are far outweighed by the above mentioned 
disadvantages, and the method as used by Thomson and Dufour may 
be set aside as insufficiently accurate. 
The next general method to be considered is the dilatometric. 
The construction of the dilatometer for use in these determinations 
is very varied; but it will be sufficient to study the forms used by 
Kopp, Plicker and Geissler, and by Bunsen. 
The instrument used by Kopp (see fig. 1) had many 
| disadvantages, which need not be entered into here at 
| great length. 
Whenever the water in the inner tube was frozen, 
a small bubble was always observed in the ice. which 
| would, of course, tend to increase the observed volume 
CH) and alter the density. The liquid used to fill the 
| vessel was turpentine oil. The oil would tend to pene- 
7 
EC NSRE 
AA 
SESS 
| 
RUE irate the cork, and the joints between the cork and 
Hk the glass tubes, and so vitiate the results. When the 
il (à . ; ‘ : 
E | ice formed, the oil was forced up higher in the capillary 
=n 
— 
A tube and thus the interior pressure on the walls of the 
vessel was increased and its dimensions altered. The 
| order of accuracy of the instrument is shown by the 
KZ fact that, while only two sets of readings were taken, 
Fic. 1. Kopps these varied 2 parts in 900, and as these were only read 
Dilatometer to 1 part in 1100, Kopp himself evidently realized the 
limitations of the instrument. The results obtained were far below 
those of the other experimenters. 
The apparatus employed in Plücker and Geissler’s determinations 
was far more delicate and accurate. There was no cork to add 
uncertainty to the action of the instrument. Also the difficulty 
of the different internal pressure, owing to the column of mercury 
being forced up in the tube, was partially overcome by having the 
capillary tube widened from a to b, causing the difference of head 


FRERE 
TRES 
