Textures of Igneous Rocks
Questions to be Considered in this Chapter:
1. What textures may be produced as magma cools and crystallizes to form igneous rocks?
2. What physical variables control the development of igneous textures, and how do they do so?
3. What recrystallization textures may result as high-temperature igneous minerals, once formed, cool further
toward near-surface conditions?
4. How can we work backward from the knowledge we have and use the textures we see to interpret the
developmental history of the rock exhibiting them?
P
etrography is the branch of petrology that deals with the description and classification of rocks. You should al-
ready know how to categorize and name an igneous rock, and, because the classification scheme is now largely
developed, most modern petrography involves the detailed study of rocks in thin section, using the polarizing
light (“petrographic”) microscope. Thin sections are cut from rock samples, cemented to microscope slides, and ground
down to 0.03 mm thickness so that they readily transmit light. From a purely descriptive standpoint, a good rock depic-
tion should include the mineralogy, a proper name, and a good description of the rock’s texture in hand sample and from
thin sections. But textures are much more important than mere descriptive aids. The texture of a rock is a result of vari-
ous processes that controlled the rock’s genesis and, along with mineralogy and chemical composition, provides informa-
tion that we may use to interpret the rock’s origin and history. It is thus important for us to be able to recognize and
describe the textures of a rock and to understand how they are developed. For example, interlocking texture is produced
by crystallization from a melt and can be used to infer the igneous origin of a rock. In this chapter, we will explore ig-
neous textures in more detail, seeking to discover what controls those textures so that we can use textural criteria to aid
us in understanding the crystallization history (and perhaps also some of the post-crystallization history) of a particular
rock. I have supplied a glossary of textural terms at the end of the chapter, where you will find the definitions of terms
you may encounter. In some of the references listed at the end of the chapter, you will also find excellent color photo-
graphs and line drawings (based mostly on thin sections) that further illustrate many of the textures.
The textures that you observe in an igneous rock result from a number of processes that can be grouped into two
principal categories. Primary textures occur during igneous crystallization and result from interactions between miner-
als and melt. Secondary textures are alterations that take place after the rock is completely solid. The following is a very
general discourse on how a number of the most common textures develop. I will concentrate on thin section study, but
many of the textures described can be recognized in hand specimens also.
, Textures of Igneous Rocks
1 PRIMARY TEXTURES (CRYSTAL/MELT high-energy faces, the overall energy of the system is lower
INTERACTIONS) and, hence, more stable. In more complex silicates, this ten-
dency may be superseded by preferred growth in directions
The formation and growth of crystals, either from a melt or with uninterrupted chains of strong bonds. Pyroxenes and
in a solid medium (metamorphic mineral growth), involves amphiboles thus tend to be elongated in the direction of the
three principal processes: (1) initial nucleation of the crys- Si-O-Si-O chains, and micas tend to elongate plate-like in
tal, (2) subsequent crystal growth, and (3) diffusion of chem- the directions of the silicate sheets. Defects such as screw
ical species (and heat) through the surrounding medium to dislocations may also aid the addition of new ions to a
and from the surface of a growing crystal. growing face, and impurities may inhibit growth in some
Nucleation is a critical initial step in the development directions. The surface energy on different faces of a crystal
of a crystal. Very tiny initial crystals have a high ratio of sur- may vary disproportionately with changing conditions, so
face area to volume and, thus, a large proportion of ions at the shape of a particular mineral may vary from one rock to
the surface. Surface ions have unbalanced charges because another. Discussions of crystal growth based on crystal de-
they lack the complete surrounding lattice that balances the fects, the nature of lattice-building elements, the nature of
charge of interior ions. The result is high surface energy for the crystal–melt interface, and structural coherence be-
the initial crystal and, therefore, low stability. The clustering tween the melt and growing faces may be found in Kirk-
of a few compatible ions in a cooling melt will thus tend to patrick (1975, 1981), Dowty (1980a), Lofgren (1980), and
spontaneously separate, even at the saturation temperature Cashman (1990).
when conditions are otherwise suitable for crystallization of In most situations, the composition of a growing
a particular mineral. Under such conditions, crystallization crystal differs considerably from that of the melt. Only in
would be possible, but the prerequisite nucleation isn’t. simple chemical systems, such as water–ice, is this not
Before crystallization can take place, a critically sized “em- true. In the general case, then, the growth of a mineral will
bryonic cluster” or “crystal nucleus” must form, with a suf- gradually deplete the adjacent melt in the constituents that
ficient internal volume of fully bonded ions to overcome the the mineral preferentially incorporates. For growth to pro-
surface-related instability. This typically requires some de- ceed, new material must diffuse through the melt, cross the
gree of undercooling (cooling of a melt below the true crys- depleted zone, and reach the crystal surface. In addition,
tallization temperature of a mineral) or supersaturation the formation of a crystal from a melt produces heat (the
before a sufficient number of ions to be stable can sponta- latent heat of crystallization, which is merely the opposite
neously cluster together (“homogeneous nucleation”). Alter- of the latent heat of fusion). This heat must also be able to
natively, a preexisting crystal surface may be present: either diffuse away from the crystal, or the temperature at the
a “seed crystal” of the same mineral or a different mineral growing surface may become too high for crystallization to
with a similar structure on which the new mineral can easily proceed.
nucleate and grow (“heterogeneous nucleation”). For re-
views of the kinetics of nucleation, see Dowty (1980a),
1.1 Rates of Nucleation, Growth, and
Kirkpatrick (1981), and Cashman (1990).
Diffusion
Several experimental studies have indicated that crys-
tals with simple structures tend to nucleate more easily than Because there are three main processes involved in mineral
those with more complex structures. Oxides (such as mag- development, and not just one, their relative rates have
netite or ilmenite) generally nucleate more easily (with less considerable influence on the ultimate texture of the result-
undercooling) than does olivine, followed by pyroxene, pla- ing rock. We shall see that, as with the weakest link in a
gioclase, and alkali feldspar, with progressively more com- chain, whichever rate is the slowest will be the overall rate-
plex Si-O polymerization. This may explain why oxides are determining process and exert the most control on crystal-
typically small and numerous, whereas alkali feldspars gen- lization. There is a further rate that we must also address: the
erally grow quite large, seemingly regardless of the degree cooling rate of the magma. If the cooling rate is very slow,
of undercooling. equilibrium is maintained or closely approximated. If the
Crystal growth involves the addition of ions onto ex- cooling rate is high, significant undercooling can result
isting crystals or crystal nuclei. In a simple structure with because there is seldom time for nucleation, growth, or dif-
high symmetry, faces with a high density of lattice points fusion to keep pace. The cooling rate is an important exter-
({100}, {110}) tend to form more prominent faces (the nally controlled variable that influences the rates of the
“Law of Bravais”). Different faces also grow at different other crystal-forming processes. Much of the textural infor-
rates. As a rather simplistic generalization, fast-growing mation that we observe is thus used to interpret the cooling
faces tend to be those with smaller interplanar lattice spac- rate of a rock.
ings (and higher surface energies). If the c-axis unit cell The rates of both nucleation and crystal growth are
spacing is particularly small, for example, a crystal may be strongly dependent on the extent of undercooling of the
expected to become elongated in the c-axis direction. Fast- magma. Initially, undercooling enhances both rates, but fur-
growing faces thus tend to grow themselves out of exis- ther cooling decreases kinetics and increases viscosity, thus
tence. In general, faces with low surface energy become inhibiting the rates. As illustrated in Figure 1, the maxi-
more prevalent. When low-energy faces predominate over mum growth rate is generally at a higher temperature than is
, Textures of Igneous Rocks
the maximum nucleation rate because it is easier to add an crystals. When there is a distinctly bimodal distribution in
atom with high kinetic energy onto an existing crystal lattice grain size, with one size considerably larger than the other,
than to have a chance encounter of several such atoms at the texture is called porphyritic. The larger crystals are
once to form an embryonic cluster. Further undercooling in- called phenocrysts, and the finer surrounding ones are
hibits growth because atoms have to diffuse farther to add called matrix or groundmass. A porphyritic rock is consid-
onto a few existing crystals, and it is easier for the slowed ered plutonic or volcanic on the basis of the matrix grain
atoms to nucleate in local clusters than to move far. size. If the phenocrysts are set in a glassy groundmass, the
We can use Figure 1 to understand why the rate of texture is called vitrophyric. If the phenocrysts contain nu-
cooling so profoundly affects the grain size of a rock. As it merous inclusions of another mineral that they enveloped as
relates to Figure 1, “undercooling” is the degree to which they grew, the texture is called poikilitic. The host crystal
temperature falls below the melting point (which, of course, may then be called an oikocryst.
is also the crystallization temperature when we consider The growth rate of a crystal depends upon the surface
cooling) before crystallization occurs. For example, if the energy of the faces and the diffusion rate. For a constant
cooling rate is low, only slight undercooling will be possible cooling rate, the largest crystals will usually be those with
(such as at temperature Ta in Figure 1). At this tempera- the most plentiful or fastest-diffusing components. The dif-
ture, the nucleation rate is very low, and the growth rate is fusion rate of a chemical species is faster at higher tempera-
high. Fewer crystals thus form, and they grow larger, result- ture and in material with low viscosity. Diffusion rate is thus
ing in the coarse-grained texture common among slow- low in highly polymerized viscous melts. (Such melts are
cooled plutonic rocks. Quickly cooled rocks, on the other generally silica rich and also tend to be cooler than mafic
hand, may become significantly undercooled before crystal- melts.) Small ions with low charges diffuse best, whereas
lization begins. If rocks are undercooled to Tb in Figure 1, large polymerized complexes diffuse poorly. In general, dif-
the nucleation rate exceeds the growth rate, and many small fusion in a fluid is better than in a glass, and it is better in
crystals are formed, resulting in the very fine-grained texture glass than in crystalline solids. H2O dramatically lowers the
of volcanic rocks. Very high degrees of undercooling (Tc in degree of polymerization of magma, thereby enhancing dif-
Figure 1) may result in negligible rates of nucleation and fusion. Alkalis have a similar effect, although less extreme.
growth, such that the liquid solidifies to a glass with very The very coarse grain size of many pegmatites may be at-
few or no crystals. tributed more to the high mobility of species in the H2O-rich
Two-stage cooling can create a bimodal distribution of melt from which they crystallize than to extremely slow
grain sizes. Slow cooling followed by rapid cooling is the cooling.
only plausible sequence and might occur when crystalliza- The rates of nucleation and growth vary with the sur-
tion began in a magma chamber, followed by the opening of face energy of the minerals and the faces involved, the de-
a conduit and migration of magma to the surface. Initially, gree of undercooling, and the crystal structure. These
the magma would be only slightly undercooled, and a few values can be different for different minerals, even in the
coarse crystals would form, followed by volcanism and finer same magma. Different minerals can be undercooled to dif-
fering extents because the melting point in Figure 1 is
specific to each mineral. Minerals develop sequentially in a
cooling magma as the melting point of each is progressively
Tc Tb Ta
reached. The temperature may thus be lower than the melt-
ing point of one mineral (undercooled) and higher than that
Nucleation of another. Many stable nuclei of one mineral may thus
Melting Point
form, while only a few of another may form, resulting in
many small crystals of the former and fewer, larger crystals
of the latter. The popular notion that the large crystals in a
Growth
Rate
porphyritic rock must have formed first or in a slower-cool-
ing environment is thus not universally valid. The sudden
loss of an H2O-rich fluid phase from a melt will quickly
raise the crystallization temperature and can also produce
porphyritic texture in some plutonic rocks.
When the diffusion rate is not the limiting (slowest)
rate, crystals growing free and unencumbered in a melt will
Temperature tend to be euhedral and nicely faceted. Different crystal
FIGURE 1 Idealized rates of crystal nucleation and growth faces have different atomic environments and surface ener-
as a function of temperature below the melting point. Slow gies. As discussed above, faces with low surface energy will
cooling results in only minor undercooling (Ta), so that rapid generally be most stable and manifest themselves when
growth and slow nucleation produce fewer coarse-grained
growing freely in a liquid; other factors may influence
crystals. Rapid cooling permits more undercooling (Tb), so that
slower growth and rapid nucleation produce many fine-grained growth rates in some directions, with considerable effect on
crystals. Very rapid cooling involves little if any nucleation or crystal shapes.
growth (Tc), producing a glass.
Questions to be Considered in this Chapter:
1. What textures may be produced as magma cools and crystallizes to form igneous rocks?
2. What physical variables control the development of igneous textures, and how do they do so?
3. What recrystallization textures may result as high-temperature igneous minerals, once formed, cool further
toward near-surface conditions?
4. How can we work backward from the knowledge we have and use the textures we see to interpret the
developmental history of the rock exhibiting them?
P
etrography is the branch of petrology that deals with the description and classification of rocks. You should al-
ready know how to categorize and name an igneous rock, and, because the classification scheme is now largely
developed, most modern petrography involves the detailed study of rocks in thin section, using the polarizing
light (“petrographic”) microscope. Thin sections are cut from rock samples, cemented to microscope slides, and ground
down to 0.03 mm thickness so that they readily transmit light. From a purely descriptive standpoint, a good rock depic-
tion should include the mineralogy, a proper name, and a good description of the rock’s texture in hand sample and from
thin sections. But textures are much more important than mere descriptive aids. The texture of a rock is a result of vari-
ous processes that controlled the rock’s genesis and, along with mineralogy and chemical composition, provides informa-
tion that we may use to interpret the rock’s origin and history. It is thus important for us to be able to recognize and
describe the textures of a rock and to understand how they are developed. For example, interlocking texture is produced
by crystallization from a melt and can be used to infer the igneous origin of a rock. In this chapter, we will explore ig-
neous textures in more detail, seeking to discover what controls those textures so that we can use textural criteria to aid
us in understanding the crystallization history (and perhaps also some of the post-crystallization history) of a particular
rock. I have supplied a glossary of textural terms at the end of the chapter, where you will find the definitions of terms
you may encounter. In some of the references listed at the end of the chapter, you will also find excellent color photo-
graphs and line drawings (based mostly on thin sections) that further illustrate many of the textures.
The textures that you observe in an igneous rock result from a number of processes that can be grouped into two
principal categories. Primary textures occur during igneous crystallization and result from interactions between miner-
als and melt. Secondary textures are alterations that take place after the rock is completely solid. The following is a very
general discourse on how a number of the most common textures develop. I will concentrate on thin section study, but
many of the textures described can be recognized in hand specimens also.
, Textures of Igneous Rocks
1 PRIMARY TEXTURES (CRYSTAL/MELT high-energy faces, the overall energy of the system is lower
INTERACTIONS) and, hence, more stable. In more complex silicates, this ten-
dency may be superseded by preferred growth in directions
The formation and growth of crystals, either from a melt or with uninterrupted chains of strong bonds. Pyroxenes and
in a solid medium (metamorphic mineral growth), involves amphiboles thus tend to be elongated in the direction of the
three principal processes: (1) initial nucleation of the crys- Si-O-Si-O chains, and micas tend to elongate plate-like in
tal, (2) subsequent crystal growth, and (3) diffusion of chem- the directions of the silicate sheets. Defects such as screw
ical species (and heat) through the surrounding medium to dislocations may also aid the addition of new ions to a
and from the surface of a growing crystal. growing face, and impurities may inhibit growth in some
Nucleation is a critical initial step in the development directions. The surface energy on different faces of a crystal
of a crystal. Very tiny initial crystals have a high ratio of sur- may vary disproportionately with changing conditions, so
face area to volume and, thus, a large proportion of ions at the shape of a particular mineral may vary from one rock to
the surface. Surface ions have unbalanced charges because another. Discussions of crystal growth based on crystal de-
they lack the complete surrounding lattice that balances the fects, the nature of lattice-building elements, the nature of
charge of interior ions. The result is high surface energy for the crystal–melt interface, and structural coherence be-
the initial crystal and, therefore, low stability. The clustering tween the melt and growing faces may be found in Kirk-
of a few compatible ions in a cooling melt will thus tend to patrick (1975, 1981), Dowty (1980a), Lofgren (1980), and
spontaneously separate, even at the saturation temperature Cashman (1990).
when conditions are otherwise suitable for crystallization of In most situations, the composition of a growing
a particular mineral. Under such conditions, crystallization crystal differs considerably from that of the melt. Only in
would be possible, but the prerequisite nucleation isn’t. simple chemical systems, such as water–ice, is this not
Before crystallization can take place, a critically sized “em- true. In the general case, then, the growth of a mineral will
bryonic cluster” or “crystal nucleus” must form, with a suf- gradually deplete the adjacent melt in the constituents that
ficient internal volume of fully bonded ions to overcome the the mineral preferentially incorporates. For growth to pro-
surface-related instability. This typically requires some de- ceed, new material must diffuse through the melt, cross the
gree of undercooling (cooling of a melt below the true crys- depleted zone, and reach the crystal surface. In addition,
tallization temperature of a mineral) or supersaturation the formation of a crystal from a melt produces heat (the
before a sufficient number of ions to be stable can sponta- latent heat of crystallization, which is merely the opposite
neously cluster together (“homogeneous nucleation”). Alter- of the latent heat of fusion). This heat must also be able to
natively, a preexisting crystal surface may be present: either diffuse away from the crystal, or the temperature at the
a “seed crystal” of the same mineral or a different mineral growing surface may become too high for crystallization to
with a similar structure on which the new mineral can easily proceed.
nucleate and grow (“heterogeneous nucleation”). For re-
views of the kinetics of nucleation, see Dowty (1980a),
1.1 Rates of Nucleation, Growth, and
Kirkpatrick (1981), and Cashman (1990).
Diffusion
Several experimental studies have indicated that crys-
tals with simple structures tend to nucleate more easily than Because there are three main processes involved in mineral
those with more complex structures. Oxides (such as mag- development, and not just one, their relative rates have
netite or ilmenite) generally nucleate more easily (with less considerable influence on the ultimate texture of the result-
undercooling) than does olivine, followed by pyroxene, pla- ing rock. We shall see that, as with the weakest link in a
gioclase, and alkali feldspar, with progressively more com- chain, whichever rate is the slowest will be the overall rate-
plex Si-O polymerization. This may explain why oxides are determining process and exert the most control on crystal-
typically small and numerous, whereas alkali feldspars gen- lization. There is a further rate that we must also address: the
erally grow quite large, seemingly regardless of the degree cooling rate of the magma. If the cooling rate is very slow,
of undercooling. equilibrium is maintained or closely approximated. If the
Crystal growth involves the addition of ions onto ex- cooling rate is high, significant undercooling can result
isting crystals or crystal nuclei. In a simple structure with because there is seldom time for nucleation, growth, or dif-
high symmetry, faces with a high density of lattice points fusion to keep pace. The cooling rate is an important exter-
({100}, {110}) tend to form more prominent faces (the nally controlled variable that influences the rates of the
“Law of Bravais”). Different faces also grow at different other crystal-forming processes. Much of the textural infor-
rates. As a rather simplistic generalization, fast-growing mation that we observe is thus used to interpret the cooling
faces tend to be those with smaller interplanar lattice spac- rate of a rock.
ings (and higher surface energies). If the c-axis unit cell The rates of both nucleation and crystal growth are
spacing is particularly small, for example, a crystal may be strongly dependent on the extent of undercooling of the
expected to become elongated in the c-axis direction. Fast- magma. Initially, undercooling enhances both rates, but fur-
growing faces thus tend to grow themselves out of exis- ther cooling decreases kinetics and increases viscosity, thus
tence. In general, faces with low surface energy become inhibiting the rates. As illustrated in Figure 1, the maxi-
more prevalent. When low-energy faces predominate over mum growth rate is generally at a higher temperature than is
, Textures of Igneous Rocks
the maximum nucleation rate because it is easier to add an crystals. When there is a distinctly bimodal distribution in
atom with high kinetic energy onto an existing crystal lattice grain size, with one size considerably larger than the other,
than to have a chance encounter of several such atoms at the texture is called porphyritic. The larger crystals are
once to form an embryonic cluster. Further undercooling in- called phenocrysts, and the finer surrounding ones are
hibits growth because atoms have to diffuse farther to add called matrix or groundmass. A porphyritic rock is consid-
onto a few existing crystals, and it is easier for the slowed ered plutonic or volcanic on the basis of the matrix grain
atoms to nucleate in local clusters than to move far. size. If the phenocrysts are set in a glassy groundmass, the
We can use Figure 1 to understand why the rate of texture is called vitrophyric. If the phenocrysts contain nu-
cooling so profoundly affects the grain size of a rock. As it merous inclusions of another mineral that they enveloped as
relates to Figure 1, “undercooling” is the degree to which they grew, the texture is called poikilitic. The host crystal
temperature falls below the melting point (which, of course, may then be called an oikocryst.
is also the crystallization temperature when we consider The growth rate of a crystal depends upon the surface
cooling) before crystallization occurs. For example, if the energy of the faces and the diffusion rate. For a constant
cooling rate is low, only slight undercooling will be possible cooling rate, the largest crystals will usually be those with
(such as at temperature Ta in Figure 1). At this tempera- the most plentiful or fastest-diffusing components. The dif-
ture, the nucleation rate is very low, and the growth rate is fusion rate of a chemical species is faster at higher tempera-
high. Fewer crystals thus form, and they grow larger, result- ture and in material with low viscosity. Diffusion rate is thus
ing in the coarse-grained texture common among slow- low in highly polymerized viscous melts. (Such melts are
cooled plutonic rocks. Quickly cooled rocks, on the other generally silica rich and also tend to be cooler than mafic
hand, may become significantly undercooled before crystal- melts.) Small ions with low charges diffuse best, whereas
lization begins. If rocks are undercooled to Tb in Figure 1, large polymerized complexes diffuse poorly. In general, dif-
the nucleation rate exceeds the growth rate, and many small fusion in a fluid is better than in a glass, and it is better in
crystals are formed, resulting in the very fine-grained texture glass than in crystalline solids. H2O dramatically lowers the
of volcanic rocks. Very high degrees of undercooling (Tc in degree of polymerization of magma, thereby enhancing dif-
Figure 1) may result in negligible rates of nucleation and fusion. Alkalis have a similar effect, although less extreme.
growth, such that the liquid solidifies to a glass with very The very coarse grain size of many pegmatites may be at-
few or no crystals. tributed more to the high mobility of species in the H2O-rich
Two-stage cooling can create a bimodal distribution of melt from which they crystallize than to extremely slow
grain sizes. Slow cooling followed by rapid cooling is the cooling.
only plausible sequence and might occur when crystalliza- The rates of nucleation and growth vary with the sur-
tion began in a magma chamber, followed by the opening of face energy of the minerals and the faces involved, the de-
a conduit and migration of magma to the surface. Initially, gree of undercooling, and the crystal structure. These
the magma would be only slightly undercooled, and a few values can be different for different minerals, even in the
coarse crystals would form, followed by volcanism and finer same magma. Different minerals can be undercooled to dif-
fering extents because the melting point in Figure 1 is
specific to each mineral. Minerals develop sequentially in a
cooling magma as the melting point of each is progressively
Tc Tb Ta
reached. The temperature may thus be lower than the melt-
ing point of one mineral (undercooled) and higher than that
Nucleation of another. Many stable nuclei of one mineral may thus
Melting Point
form, while only a few of another may form, resulting in
many small crystals of the former and fewer, larger crystals
of the latter. The popular notion that the large crystals in a
Growth
Rate
porphyritic rock must have formed first or in a slower-cool-
ing environment is thus not universally valid. The sudden
loss of an H2O-rich fluid phase from a melt will quickly
raise the crystallization temperature and can also produce
porphyritic texture in some plutonic rocks.
When the diffusion rate is not the limiting (slowest)
rate, crystals growing free and unencumbered in a melt will
Temperature tend to be euhedral and nicely faceted. Different crystal
FIGURE 1 Idealized rates of crystal nucleation and growth faces have different atomic environments and surface ener-
as a function of temperature below the melting point. Slow gies. As discussed above, faces with low surface energy will
cooling results in only minor undercooling (Ta), so that rapid generally be most stable and manifest themselves when
growth and slow nucleation produce fewer coarse-grained
growing freely in a liquid; other factors may influence
crystals. Rapid cooling permits more undercooling (Tb), so that
slower growth and rapid nucleation produce many fine-grained growth rates in some directions, with considerable effect on
crystals. Very rapid cooling involves little if any nucleation or crystal shapes.
growth (Tc), producing a glass.
