331 
upper portion of the mass follows surfaces which curve round 
the naujaite fragments, while in the main part of the rock the 
structure follows the general inclination of the batholitic sheets 
which are indicated in Fig. 29 (p. 322), though there is an even 
transition between the two arrangements. Observations, thus, 
seem to indicate that the schistose structure of the whole 
lujavrite mass is causally connected with the sinking down of 
the upper strata. But here we meet a difficulty, for it would 
appear that a structure produced in this way must be limited 
to the border zone, whereas, as a matter of fact the schistose 
structure affects the whole mass. 
The lujavrite has crystallized after the other rocks of the 
batholite, and, therefore, probably at the lowest temperature. 
The high content of low melting point minerals, such as arf- 
vedsonite, and the behaviour of the felspars of the rock (see 
p. 160), confirm the assumption that the crystallization occurred 
at a low temperature. On the other hand the large quantity 
of secondary analcime replacing almost all the light coloured 
minerals occurs as anhedra which are larger than the average 
crystals of the primary constituents of the rock. The very 
common pneumatolytic alteration products which are found in 
all rocks of the batholite, also seem to indicate that the 
lujavrite during its crystallization must have given off great 
quantities of volatile substances. 
For these reasons the lujavritic magma is supposed to have 
been very rich in volatile substances, and to have cooled down 
to a relatively low temperature before it became finally conso- 
lidated. 
Another peculiar circumstance relating to the crystalliza- 
tion of this rock may be mentioned, namely the absence of an 
active pressure from below. When an ordinary abyssal rock 
crystallizes, it must be assumed that the decrease of volume 
caused by the crystallization, as a rule, is compensated by an 
additional supply of magma from below. But since the over- 
