1023 
against each other), they are cut through in the middle; the planes 
of section must then be ground once more and polished. 
A second circumstance which gave rise to difficulties, and some- 
times does so now, is the becoming opaque of the cones. If such 
a cone which has become opaque, is removed from the cylindre, 
it appears that innumerable planes of cleavage have arisen at right 
angles to the axis of the cone, so that it can be easily broken up 
with the hand into a great number of plane plates. In consequence 
of these cracks the at first perfectly transparent cone has become 
quite opaque. It appears that the cones hardly ever or never become 
opaque with rising pressure. It is probably the consequence of the 
compressibility of the steel cylindre. This extends with rising pres- 
sure, and so the glass cone is driven deeper and deeper into the 
eylindre. If then the pressure diminishes, the cone cannot return to 
its first position and is cracked by the immense pressure of the 
steel cylindre. In agreement with this is the fact that in experiments 
at higher temperature the cones become opaque still more frequently 
than at lower temperatures; the difference of the coefficient of expan- 
sion of glass and steel then acts in the same direction. Moreover by 
means of the brass model used for grinding the glass cones into 
the eylindre, it could be clearly demonstrated that one of the steel 
eylindres had widened by use. A cylindre made expressly of speci- 
ally hard nickel-steel yielded better results in this respect. When this 
was used, the cones were less liable to crack, though even then it 
occasionally happened. To protect the window from injury as much 
as possible it is also desirable to diminish the pressure as carefully 
as possible; a rapid increase of the pressure, on the other hand, 
rarely, if ever, gives rise to an accident. Though this cleaving of 
the cones perpendicular to the axis continues to be a drawback, 
because the preparation and adjustment of new cones always remains 
a rather lengthy work, the observations themselves are not disturbed 
by it, if only the experiments are made as much as possible with 
rising pressure, and decrease of pressure is effected with the utmost care. 
At present the apparatus cannot be used for temperatures much 
above 70°; the Cailletet cement with which the windows are fastened 
into the steel cylindres, melts at that temperature, or at least rapidly 
dissolves in the liquid which is in the pressure tube. First of all 
this renders the liquid turbid, but moreover it gives rise to leakages 
and breaking of the windows, which are now directly pressed against 
the steel. We are now trying to find means to apply the windows 
also at higher temperatures. As to the limit of pressure, we think 
we have to fix this for the present at about 5000 atm. 
