Rit (Vísindafélag Íslendinga) - 01.06.1984, Blaðsíða 262

258
PÁI.L IMSLAND
Some of the evolved rocks of this stage contain plagioclase cores that might
be crystallized at depths below the crust-mantle boundary. In other cases
the evolved rocks contain in fact no phenocrysts and have been taken to the
surface in a totally liquid state.
B. PRIMARY, PRIMITIVE AND EVOLVED MAGMAS
In order to recognize the processes responsible for the compositional
diíl'erences within the rock suite, the primary magmas themselves must be
identified.
The widest possibility is are to look upon all the rocks as representing
crystallized primary magmas. From such a definition it follows that all the
minerals contained by the rock, both phenocrysts and groundmass grains,
are crystallized out of that primary magma under equilibrium conditions,
no crystals have been removed or added, and no liquid has been removed or
added. In other words, each rock represents a magma batch taken from the
place of generation to the surface without any modification or any com-
munication to its surroundings, other than giving off heat, i.e. as a chemi-
cally closed system all the way. Numerous aspects of the mineralogy of the
Jan Adayen rocks, discussed in the previous chapters, prove this wide view to
be unrealistic, before even considering other lines of argument. This is
apparently a general conclusion regarding rock suites of broad composi-
tional spectra.
The narrowest view of primary magmas is to select a single major and
trace element composition as being primary and to evolve all other magma-
and rock compositions from it by the appropriate evolutionary processes.
Whether this is realistic or not, in general, depends on the compositional
width of the rock suite and the range of the physical conditions at the site of
magma generation and magma solidification. It may be realistic in cases of
compositionally narrow rock suites, especially if they are crystallized
relatively close to the magma generation site, but it becomes more unlikely
as the compositional spectra of the rock suite widens and the site of
solidification is removed from that of the magma formation. Nevertheless, in
spite of occasionally poor fit, this approach to the internal relations of rock
suite members has frequently been used.
Between these two extremes, there are numerous possibilities and varia-
tions. Of these the following two (in their simplest form) are probably of the
greatest importance.
1) Two (or more) end member magmas exist which can be mixed in
various proportions to form a continuous rock suite.
2) Two (or more) primary magmas of diíferent compositions exist and
evolve along a common and continuous trend, but starting at different loci
on the trend, as one is more primitive than the other. The more primitive