154 ROYAL SOCIETY OF CANADA 
The following is a table of the results obtained by the method: 
which has just been described: 
TABLE II. ON THE Density OF ICE. METHOD OF WEIGHING IN WATER AT (0° C. 

Difference from 
Date Year of Formation Density Men 
March 9th 1901 “91684 ‘00023 
a oh | 1901 “91665 00004 
ie ef 1900 ‘91661 “00000 
‘ 16th 1900 91642 “00019 
x a 1899 “91650 *O0u11 
<i Ÿ 1899 “91648 ‘00013 
‘ 923rd 1900 "91678 00017 




Mean = ‘916611 + ‘000065 
In order to check the accuracy of the method, two determinations 
were made upon the same specimen (the 1899 ice), the ice being 
removed, washed, and the whole experiment repeated. The two 
results agree to 2 parts in 90,000. All the other determinations were 
made with different specimens, sizes ranging from 150 — 200 grams. 
From these experiments it appears that the density of the St. 
Lawrence river ice may be taken as 
0-91661 + -00007. 
a value agreeing very closely to Nichols’ value for old river ice, but 
considerably lower than his value for natural ice newly cut. It 
also agrees very closely with Bunsen’s result (2 in 10,000), which is 
the generally accepted value. 
Quite recently, experiments have been carried out by J. H. 
Vincent,’ on the Density and Cubical Expansion of Ice. His method 
consisted in weighing water in mercury. The water was weighed 
both as liquid at 0° C., and as solid at several temperatures below 
the freezing point. It was necessary for him to assume values for 
the density of water and mercury at 0° C. The density of ice at 
0° C. was then calculated, assuming that the densities of ice and 
mercury are linear functions of the temperature. The measure- 
ments are therefore not direct, but depend on certain pre-arranged 
assumptions. 
The mean value of the density of ice at 0° C. is found by Vincent 
to be -9160. The variations of the individual computations, which 
had to be made in order to arrive at a value of the density at 0° C., 
are of the order of 1 in 1,000. 
1 Proc. Roy. Soc., 69, 422 (1902). 


