59 
The specimen enclosing this cavity was fastened to a piece of glass, and 
this was fixed in a beaker containing water, supported so that the cavity 
was in the focus of the microscope under a low power. The temperature 
was raised very slowly, and was maintained for some minutes at each par- 
ticular degree at which it was thought desirable to measure the volume of 
the liquid ; and this was usually repeated over and over again when the 
heat was both rising and falling, so as to obtain as accurate a result as 
possible. In making the measurements with the micrometer, care was 
taken to allow for the tapering end of the cavity and the curved surface 
of the liquid. The results are given in degrees Centigrade. Though 
the expansion below 30° was very great, compared with that of any other 
known substances except liquid carbonic acid and nitrous oxide, when the 
temperature rose above 30° it was so very extraordinary that it was not 
until after having performed the experiment over and over again that Mr. 
Sorby felt confidence in the results. This will not be thought surprising 
when we state that from 31 w to 32° the apparent expansion of the liquid 
is no less than one-fourth of the bulk it occupies at 31° ; the length of 
the column increasing for that single degree from to -5_o_ inch. 
This is about 780 times as great as the expansion of water would be, and 
even 69 times as much as that of air and permanent gases. It was not 
possible to ascertain the amount of expansion above 32 c C. , because the 
cavity was quite filled at that temperature. If the expansion increase at 
the same increased rate, the liquid would soon occupy several times as 
much space ; but it seems very probable that before then it would pass 
into the state of gas. At all events it appears as if this enormous rate of 
expansion indicated a close approach to a fresh physical condition. The 
following Table gives the results of the experiments; and it has been 
found by drawing them as a curve, that their general relations indicate 
that there cannot be any serious error ; but at the same time, considering 
all the circumstances, they must only be looked upon as tolerably good 
approximations to the truth. 
Temperature. Volume, 
o 
OC 100 
174 109 
20 113 
25 122 
28 130 
