( 887 ) 
ice contained some salt-solution (dissolved from the glass), according 
to QUINCKE’s observations, but certainly less than the common arti- 
ficial commercial ice. One experiment was made with iee, formed 
from water which had been distilled directly through a metal cooler 
into a vessel of sheet copper, had next been boiled for an hour and 
then been placed at once outdoors during a frosty night. Of this 
ice a layer near the wall was clear as glass; nearer the middle it 
contained small bubbles. Hence the water had not been sufficiently 
freed from air by boiling or had taken up air again. In the experiment 
the clear part only was used. 
The experiments were made as in my former communication, with 
a steel wire of 0,4 mm. thickness, cut from the same piece from 
which I had taken the wire, used in the former measurements. This 
steel wire had been greased and kept since; it showed no perceptible 
changes, also under the microscope. 
The results of the different measurements are given in the following 
table *). 
E Dad CAE C Remarks: 
1. 2150 3.8 23 76° 0.037 Ice frozen from distilled water 
in glass tube, perfectly clear. 
2. 2150 2.2 36 80° 0.034 Ice frozen from distilled water 
in glass tube, perfectly clear. 
3. 1150 1.2 28 81° 0.027 As before; had lain outdoors 
during the morning with thaw 
and so was certainly at 0° at 
the interior. 
4. 2150 23 34 83° 0.084 As number one and two. 
5. 2150 207 32 83° 0.028 Lower part of the piece of number 
4, not quite clear, descent 
irregular. 
6. 2150 0.71 91 82° 0.027 Ice from distilled water, frozen 
in metal vessel. 
that such a plume formed each time the tube had for a moment been removed 
from the freezing bath, in order to see how far the freezing had proceeded. Thus 
it could afterwards be seen from the number of plumes how many times this had 
been done. 
1) The notations are the same as in the former communication. I avail myself 
of this opportunity to correct an error in the formulae of that communication. In 
formulae (1) and (3) the factor 2 in the numerator must be dropped. In the paper 
by Dr. Ornstein, from which the formulae had been taken, P denoted the weight 
at either side of the wire; in mine P is the total weight. The values for C are 
correctly calculated, however; the 2 was dropped in the calculation, so that the 
mistake had no further consequences. 
