I40 THE NORTHERN MICROSCOPIST. 

the temperature far below the freezing point of water and it would 
not freeze. Still I felt persuaded that it was water, from other 
considerations, and that caused me to investigate the effect of the 
freezing of water in minute tubes, and I found that when it is in 
capillary tubes and small cavities that you can reduce the freezing 
point of the water far below its usual one ; but if the least portion 
of ice were in contact with it it would freeze immediately. But 
you may reduce the temperature to below 10°, I believe, without 
freezing it, if no ice is present. 
But that the substance was water was easily ascertained by its 
optical properties, because they are so different to the optical 
properties of ice. The above will explain, then, why it did not 
freeze in small cavities. Then one could determine the rate of 
expansion that agreed with water. 
Then there was another method. You see cavities filled with 
fluid in the quartz, and you know the diffusion of steam will burst 
the cavities if you heat the specimen. I had a tube and put the 
crystal at the bottom, then. pumped thoroughly dry air into the 
tube. Then I had a freezing mixture. After heating the crystal 
I got a deposition inside the tube of what looked like hoar frost 
and had all the character of frozen water. I now put this into 
salt and water below the freezing point, and found as soon as the 
temperature rose and got to the melting point of ice, the material 
obtained in this manner thawed. 
But the liquid is not pure water. In some cases it contained 
chloride of sodium or chloride of potassium. 
Sometimes the cavities are large crystals of these substances, 
this being especially the case in volcanic rocks, so that some of 
the salt so present then can be dissolved at the ordinary temperature. 
That was the kind of evidence on which I relied in determining 
that the liquid in the cavities was water. 
In addition to the two chlorides named, I think there is some- 
times hydrochloric acid; but that is rather doubtful, because it 
may have been set free by the action of the quartz on the chlorides. 
Of course, many cavities contain liquid carbonic acid. I first 
established that fact by proving that the rate of expansion was the 
same as that of carbonic acid, and very unlike that of anything 
else. 



