( 889 ) 
and no ice. Although a network of ice is formed in the immediate 
neighbourhood of the wire, a great part of the water does not freeze. 
It is exactly for this reason that empty spaces are formed. 
Also in the following manner I convinced myself that above the 
wire much liquid is found. By slightly pulling the wire sideways, 
putting some cotton-blue on it and then letting the wire go back 
again, some colouring substance is carried into the block of ice in 
the immediate vicinity of the wire. If not too much has been intro- 
duced, one only sees those spaces coloured where the empty bubbles 
occurred ; the network between them is colourless, since by the 
formation of ice the colouring substance is excluded. Sometimes a 
space in which a bubble occurs is seen to shrink near the wire, so 
that the freezing process can be directly observed. To be sure the 
introduction of the colouring matter will interfere with the freezing 
of the melting-water, but very little of the colouring substance is 
already sufficient for the experiment. 
If there are empty spaces above the wire, then — contrary to 
the former supposition — a flow of water in the piece of ice must 
be expected from the sides inwardly. In fact I could state this 
inward flow in a few cases by observing the small solid particles 
that had entered with the colouring substance. (Experiments with 
fine chalk particles gave no result). Where a narrow channel existed 
in the network of ice, the particles were forced through with great 
velocity and in an inward direction. 
The path of the wire consequently consists for a great part of 
liquid with a network of ice. The liquid is enclosed in this network, 
for a little cotton-blue, put on the piece of ice where the wire had 
sunk in, does not penetrate, whereas otherwise this colouring sub- 
stance enters through the finest channels. If a piece of ice is taken 
which is not perfectly clear and if then a little of the colouring 
substance is put on it, the whole piece is in a short time coloured 
blue all through. 
The heat necessary for the melting under the wire is consequently 
only partly furnished by the solidification above it. The question arises 
whence the rest of the heat is derived. We may not assume that 
it is by conduetion through the ice. It is not impossible that radiation 
plays a part here. The foilowing experiment supports this view. If 
a strong solution of cotton-blue is made to freeze not too slowly in 
a freezing mixture, the ice which is formed is not clear and has a 
red-violet tint’), the colour of the colouring substance when dry. 
1) When the freezing is slow the ice is in this case also perfectly clear and 
colourless. 
60 
Proceedings Royal Acad. Amsterdam Vol. XI. 
