AT HIGH PHESSUEES BY OPTICAL METHODS. 
131 
at the extremely narrow fringe at the edge of the glasses, this device has proved, 
itself to work very satisfactorily in most cases. 
It has indeed been found that with the help of this little piece of apparatus the 
substances could be equally well observed at high pressures inside the pressure 
apparatus as under the ordinary polarization microscope between object-glass and 
cover-glass. Also, most crystal-optical determinations, such, for example, as those 
serving to identify the rock-forming minerals in thin sections of rocks, can be used 
during a research at high pressures with the above described apparatus. 
(2) Isothermal Melting and Crystcdlization. —^When an alteration of pressure at 
will is possible, the melting-point of a substance can, of course, be determined 
either at constant temperature by altering the pressure, or at constant pressure by 
altering the temperature. The determination at constant temperature, what one 
may conveniently call isothermal crystallization and isothermcd melting, is l)y far 
the more convenient method of the two, and has throughout been employed in this 
research for the determination of melting-points. In the case of the determination 
of transition-points the method of work at constant pressure is of special use in 
some cases, and will be further referred to in cases where it has been applied. 
But, with regard to isothermal crystallization, there are some peculiarities which 
should be briefly discussed at this point. When a crystallized substance is melted 
in the ordinary way by being heated, it melts gradually at constant temperature, 
but at the same rate as the heat is supplied to it. It is not possible to heat instantly 
a melting mass of a substance through and through ; the heat is taken up by the 
parts of the mass in contact with the heating device or vessel which is being heated, 
and spreads from those parts to the other parts of the mass. But in the case of 
isothermal melting matters are difierent in this respect, that any change of hydraulic 
pressure almost instantly takes place throughout the entire mass. As most crystals 
need a certain amount of time both to grow and to melt, the result is that, when 
working isothermally, we are able either slightly to compress the crystals in the 
“ melt” of the substance to pressures above their melting-point pressure, or to lower 
the pressure beneath the pressure of the melting-point pressure. The extent to 
which this can be done depends almost entirely upon the velocity with which a 
crystal melts or crystallizes at its melting-point, and the rate at which this velocity 
is altered with increase or decrease of pressure, and these factors are very difierent 
for different substances. The two cases, pressures above and pressures below the 
actual melting-point pressures, correspond to a superheating of crystals in the 
presence of the liquid-phase or super-cooling of the liquid-phase in the presence of 
the crystal-phase at constant pressures. Only in the case of a totally melted 
substance is a super-cooling in most cases actually possible, and it has generally 
been thought that a superheating of a crystallized body to temperatures above its 
melting temperature could not be effected. According to some recent work, however, 
it seems probable that a superheating occurs in the case of some extremely slowly 
s 2 
