28 The N.Z. Journal of Science and Technology. [Feb. 
of a glacier below the surface therefore has a temperature below the freezing- 
temperature at the surface. The Mueller River in March—a very dry 
month—was chiefly draining the glacier of water produced by pressure 
thaw, and any surface water that reached them would be cooled to the 
temperature of the glacier’s drainage-channels. As the super-cold water 
gradually escaped from pressure in the rising pipe, and doubtless to some 
extent under the broken-down end of the glacier also, a small proportion 
of it became ice, to rectify the temperature to the new pressure conditions 
and finally to the subaerial normal ; and our observations showed that it 
takes the form of granular crystals. 
Being necessarily at the melting-temperature, the ice-crystals were 
ready under the smallest external pressure to cohere (regelate), and in con¬ 
sequence of this readiness to stick together they built up a wall where they 
were driven against each other by the movement of the water. The wall 
necessarily started from a foundation of ice on one or more sides, and thence 
was added to till a complete circuit was made. 
The deposition of crystals took place on the inside of the rising pipe 
as well as at the surface. On throwing stones into the pool anywhere but 
in the middle, smaller or larger masses, from a handful to a bucketful, 
of spongy sand-stained masses of crystals were detached, rose, and floated 
away. It may be that this process of deposition on the walls sometimes 
results in choking the passage. The deep eddy pool above mentioned 
must have been an extinct spring-pit, and we saw two other pits recently 
extinct and two older ones within a few chains of the living spring. 
The white wall of the Mueller spring was visible from the Hooker track 
opposite, but at that distance no one would suspect its peculiar nature. 
A simple quantitative statement may be made of the explanation given 
above of the production of ice in drainage-water from a glacier. The 
phenomenon results from the operation of James Thomson’s law that pressure 
lowers the freezing-point of water by 0-0074° C. (or 0-0075° as some give it) 
per atmosphere, and the law of latent heat. Taking a cubic foot of water 
as a convenient unit mass, if it were 1° C. below freezing-point in air the 
ice produced to raise the temperature to 0° C. would be one-eightieth, or 
21-6 cubic inches. As the temperature due to the pressure of one atmosphere 
is only —0-0074°, the ice produced will be so much less (21-6 X 0-0074, 
or 0-16 cubic inches), and proportionately for other pressures in atmo¬ 
spheres. Taking 100 ft. of ice as equal to three atmospheres, the ice neces¬ 
sarily formed in seepage water to rectify the temperature to 0° C. in air 
is 0-48 cubic inch per cubic foot of water per 100 ft. of ice. That is the 
minimum, however. In nature the ^drainage of the terminal ice of a valley 
glacier is likely to have a temperature lower than that required by the 
pressure of ice at the terminal, because the ice was previously, when higher 
up the valley, of a considerably greater depth, the low temperature corre¬ 
sponding to which is in part retained and carried down to the terminal. 
This fact is observable in the dryness of the ice in terminal faces. 
In reply to an inquiry as to whether he had observed such a phenomenon, 
Mr. Peter Graham, chief guide at the Hermitage, writes that the condition 
is not common, but he has several times observed it at the Mueller and 
Hooker Glaciers, and also once at the end of the Fox Glacier on the West 
Coast. Once he saw one built up 3 ft. or 4 ft. Mr. Graham remarks that 
the walled springs occur mainly in the fall of the year, and thinks that they 
are due to a fresh fall of snow on the mountains being washed down through 
the underground channels in the glacier. 
