sure, and fugacities of volatiles, especially water. It
would be useful for the petrologist if inexpensive pet-
rographic examination of rock fabric could determine
the order of crystallization and thereby provide insight
into the nature of changing intensive parameters with-
out recourse to costly, tedious, less available laboratory
equipment and techniques. Four textural criteria have
been employed during the past decades to ascertain or-
der of crystallization:
1. Comparison of phenocrystic and groundmass min-
erals in quenched magmas, usually volcanic rocks
2. Reaction rims
3. Mineral inclusions in larger grains
4. Relative euhedralism at mutual grain contacts
The latter two criteria are examined here (see also
Flood and Vernon [1988, pp. 105–116] for a more
thorough critique).
Inclusions of, for example, pyroxene poikilitically
enclosed in a larger oikocryst of plagioclase are com-
monly interpreted to mean that pyroxene crystallized
before plagioclase. But this interpretation is not neces-
sarily correct in every case. At least three alternate pos-
sibilities are:
1. Late pyroxenes might have crystallized from melt
inclusions entrapped within skeletal plagioclase af-
ter it ceased crystallizing.
2. Late pyroxenes might have crystallized from melt-
filled cracks that cut thorough the plagioclase after
it grew.
3. The inclusion-filled plagioclase might be a xeno-
cryst or a restite grain derived from metamorphic
rocks in the deep continental crust where the
magma was generated.
The order of crystallization indicated by the relative
euhedralism of grains in mutual contact can also be
ambiguous (Figure 6.26). Ophitic texture (Figure 7.15)
and hypidiomorphic-granular texture (Figure 7.16)
have mutual grain contacts that tempt one to interpret
order of growth. For ophitic texture, it is commonly as-
sumed that the euhedral plagioclases grew freely in the
melt, followed by pyroxene, which is anhedral because
of the restricted available space within the network of
earlier plagioclases. However, this is only one of at least
three different possible orders of crystallization of pla-
gioclase and pyroxene. In Figure 7.22a, a few nuclei
formed in the melt allow early growth of pyroxene, fol-
lowed by abundant nucleation of plagioclase and
growth of both phases; the final resulting ophitic tex-
ture is shown on the right. In Figure 7.22b, pyroxene
and plagioclase nucleate in the melt and grow simulta-
neously to create the final ophitic texture. In Figure
7.22c is the conventionally assumed order of crystal-
lization in which plagioclase nucleates and grows be-
fore pyroxene does. In all three instances, pyroxene nu-
cleates less abundantly than plagioclase; that can be
taken as the only valid conclusion for ophitic texture. A
and alkali-enriched boundary layer (Figure 6.16) forms
around it. Continued crystallization may create a den-
dritic to spherulitic intergrowth of alkali feldspar and
quartz, the former crystallographically continuous with
and projecting radially outward from the euhedral pla-
gioclase core. Some Ostwald ripening might coarsen the
large-surface-area intergrowth (Figure 6.21), producing
granophyric texture. Whether phaneritic graphic gran-
ite might be created by further Ostwald ripening is a
speculative possibility.
A superficially similar intergrowth in granitic rocks
consists of an intergrowth of vermicular (“wormy”)
quartz in a sodic feldspar host typically in contact with
K-rich alkali feldspar (Figure 5.18b). The origin of this
myrmekite has been ascribed to direct crystallization
or to replacement but Castle and Lindsley (1993)
believe it may originate in subsolidus unmixing of
K-bearing plagioclase in a ternary feldspar system open
to excess Si.
7.5 A WORD OF CAUTION ON
THE INTERPRETATION OF
CRYSTALLINE TEXTURES
Kinetic investigations of the dynamic behavior of melts
in the laboratory, coupled with theoretical models and,
as always, evaluated by observations of real rocks, pro-
vide insights into the origin of magmatic fabrics. How-
ever, the behavior of natural magmas is impossible to
duplicate wholly by experiments and models. A defi-
ciency in experiments is an appropriate handling of the
time scale of rock-forming processes; the slow crystal-
lization of plutonic magma bodies can never be dupli-
cated in one researcher’s professional career! In plu-
tonic rocks, and to some extent in volcanic as well, the
observed fabric is the cumulative product of a long pe-
riod of changing geologic conditions and declining ki-
netic rates that took place above and below the solidus;
the effects of overprinting processes are not easily un-
raveled. The spontaneous drive toward a state of tex-
tural equilibrium that potentially can modify grain
boundaries is largely unexplored in slowly cooled mag-
matic rocks. The kinetic history of magmatic systems as
indicated by the rock fabric is frequently ambiguous
except in the simplest of cases.
To emphasize and illustrate this sense of caution we
next briefly discuss a line of interpretive petrology that
has been a part of the discipline virtually since its in-
ception, namely, the determination of the order of crys-
tallization of minerals from the magmatic rock texture.
7.5.1 Magmatic Rock Texture and
Order of Crystallization
The sequence in which minerals crystallize with chang-
ing intensive parameters, most commonly declining T,
between the liquidus and solidus of the magma, de-
pends upon the magma composition, confining pres-
Kinetic Paths and Fabric of Magmatic Rocks
163