134 
DE. WALTER WAHL: PHYSICO-CHEMICAL DETERMINATIONS 
supported by the ol3servation by Tammann that at temperatures near a the growth 
in some cases occurs in the form of comparatively large crystals, which point in 
different directions inside the tulje, and crystal faces are richly developed, but that 
at temperatures h growth occurs in a certain direction only, so that long crystal 
threads are produced in the tiibe.^ In the case of the unstable modification III. of 
benzophenane melting at 26° C., Tammann found that the crystal grows inside the 
tube pushing forward a crystal corner, terminated by even crystal faces.^ 
As Tammann occasionally points out, every crystal face probably possesses its own 
velocity of crystallization at a given degree of super-cooling, and it is also probable 
that the velocity of the different crystal faces changes at a different rate with the 
degree of super-cooling. The result of this is that at a certain temperature, or 
pressure corresponding to this temperature, the velocity of crystallization in one 
crystal direction so much exceeds the velocity in other directions that, in consequence, 
only these crystal faces are developed and crystal-growth structures produced. It 
may suffice in this place to point out that these considerations give the clue to the 
general explanation of the production of growth-structures and of spherulites, and 
are of interest with regard to the origin of crystal-growths and spherulites in the 
igneous rocks; a fuller discussion of these problems, however, l^eing outside the scope 
of the present communication. But the possibility of determining two distinct 
crystallization pressures, one at which the growths of the crystal only just takes 
place (at a, fig. 5), and one at which the formation of growth-structures begins (at 6, 
fig. 5), is also of practical importance when the melting-point is determined at 
constant temperature, and the problem has therefore been discussed here already 
before the data of measurements are given. These two pressures will in the following 
tables be indicated as Pj = the melting-point pressure, and as Pg = the higher pressure 
at which growth-structures are developed. 
As seen from the figures in Table I. the melting-point of CBiq is raised very 
rapidly by pressure, the pressure required to raise the melting-point 1° C. being only 
16 kg./cm.^. This ratio remains constant, within the limits of accuracy of the 
measurements, up to about 10° C. above the melting-point at 1 atmosphere. 
Unfortunately it is not possible to determine the melting-points at higher 
temperatures and pressures as a slight decomposition then begins, and as the 
decomposition products tend to lower the melting-point, higher melting-point 
pressures are recorded the longer the substance has been kept at high temperatures. 
This becomes still more marked at 120° C. and at 130° C., at which temperatures the 
melting-point jDressures in different series of measurements differ very widely from 
each other. The melting-point pressure for preparations on which even only a single 
rapidly carried out observation at 120° C. was made, was, after cooling down, at 
95°'4 C. about 200 kg./cm.^ higher than it originally had been. Under such 
circumstances it is, of course, not possible to determine if the raising of the melting- 
* G. Tajimann, ‘Zeitschr. f. Phys. Chem.,’ 23, p. 68 (1899). 
