Crystals grow wherever atoms have an ordered place to go and enough freedom to get there. Cooling magma supplies both, and so does mineral-laden water. Less obviously, so does solid rock that never melts.
A garnet in a mica schist had no melt to grow out of and no cavity to grow into. The rock around it stayed rock the whole time. Bonds broke, atoms moved through the solid by diffusion or travelled a short distance in the thin fluid between grains, and where they arrived a new mineral assembled itself out of ingredients that had been sitting in the old one.
This route has a name in the textbooks and a branch of petrology behind it. Garnet, kyanite, staurolite and andalusite, the minerals that make a schist worth picking up, all grew this way, in rock that was hot and under pressure and never liquid at any point.
Growing inside solid rock
Metamorphism is the name for what happens when heat, pressure or chemically reactive fluids change a rock without any of it melting. The starting material is called a protolith, meaning the earlier rock the new one was made out of. A mudstone buried deep enough becomes a slate, then a phyllite, then a schist, and at each step the mineral list changes because the old minerals are no longer the stable ones at that temperature and pressure.
Perkins puts the mechanism in a single sentence. Bonds are broken and atoms migrate by solid state diffusion, or are transported short distances by intergranular fluids, to sites where new minerals crystallize and grow. Atoms genuinely do walk through solid rock, slowly, and they are also helped along by water films thin enough to be invisible and thick enough to dissolve things. Metamorphic rocks are not dry. Most metamorphic fluids are rich in water. They are simply unmelted.
The crystals this produces have a name of their own. A porphyroblast is a large crystal grown inside a fine-grained metamorphic rock, and the most striking examples are the mica-rich schists carrying andalusite, chloritoid, kyanite, garnet or staurolite. Grain size tracks the grade of metamorphism: low-temperature metamorphic rocks tend to hold small grains, and as the grade rises the grains generally get larger.
Porphyroblasts are not uniformly ragged, whatever growing inside solid rock might suggest. Open Petrology notes that they may be anhedral but are often subhedral or euhedral, meaning partly or fully faced, which is to say a crystal can push its way through a solid rock and still come out with proper faces. It rarely comes out clean, though. It grew through a fabric of mica and quartz that was there before it started, and that fabric tends to end up wrapped around the crystal and caught inside it as inclusions. A crystal that grew into an open cavity carries neither.
Melt, water, and the fluids in between
Crystals form as magma cools, because a melt holds elements that stop being soluble as the temperature drops and start bonding into minerals instead. Granite is the whole rock version, an interlocking mass of quartz, feldspar and mica that grew together underground. Crystals also precipitate out of water, as salt does from an evaporating lake and quartz along a vein. Hydrothermal growth is the hot end of the same idea, mineral-charged water moving through fractured rock and depositing its load where the temperature or the chemistry changes.
And crystals form by chemical reaction, a category that covers the metamorphic case above along with the slow work of weathering and diagenesis near the surface. Perkins sets out those three mechanisms on one page and then divides them across five sections: igneous, aqueous, hydrothermal, metamorphic, and weathering and diagenesis together. One mineral can arrive by more than one route. Quartz grows in granite and along veins, and it also turns up as the recrystallised silica of a metamorphosed sandstone. Same species every time.
Termination often gets described as a third stage of growth, and it is not a stage at all. A termination is the set of faces that closes off the end of a crystal, a shape and not a moment, and a section can be cut below one. On why crystals stop, the literature is more concrete: they grow until they run into other crystals or the supply of chemical constituents is cut off. Vanadinite makes that visible. It crystallises as small barrel-shaped hexagonal prisms scattered across a matrix, deep reddish-brown, each one finished and none of them touching its neighbours. Nothing on a piece like that ever had to compete for room. The supply simply stopped arriving while every crystal still had space around it.
Why size is a question about how many started
The folk rule says slow growth makes big crystals, which is close enough to be useful and still wrong.
Crystal size is set by the balance between two rates. Nucleation is the rate at which new crystals start; growth is the rate at which existing ones get bigger. Cashman puts it as a trade-off. For the same total crystallinity, high nucleation rates produce numerous small crystals, while high rates of growth on a limited number of sites produce fewer but larger crystals. Both rates depend on how far the system is from equilibrium, which in a magma is measured as effective undercooling.
Nelson’s petrology notes lay out three cases, and the third is the one that breaks the folk rule. At low undercooling there are few nuclei, so the crystals that do start can grow to a relatively large size and the texture comes out coarse. At large undercooling many crystals nucleate and all grow fast, but there are so many of them that they run into each other before they have time to get anywhere. At high undercooling both the growth rate and the nucleation rate are low, so few crystals form and none of them grows large, and the result is glass.
Obsidian is that third case made into an object. Some extrusive igneous rocks cool so quickly that they contain glass, and obsidian is the version most people have handled. It had almost no nuclei, and it has no crystals either. Few nuclei on its own does not buy you a large crystal; you also need something to feed it, for long enough, without the conditions changing.
The same magma can produce two rocks that look nothing alike. Cooled slowly at depth it comes out coarse, with grains large enough to pick apart by eye. Erupted and chilled at the surface it comes out fine-grained, or glassy. The chemistry can be near enough identical. The difference is entirely in how many places the atoms had to start from.
A magma-grown crystal interlocks with its neighbours along shared boundaries. A vein crystal terminates into what was once open space. A porphyroblast has the rock’s fabric wrapped round it and pieces of that fabric caught inside, and of the three that one can be checked on a hand specimen without cutting anything. The species that turn up in each of those settings, and what distinguishes them, get their own pages in the Crystalance Mineral Library. Whether the word crystal covers all of them is a separate question and a contested one: see what counts as a crystal.
Sources
- Dexter Perkins, Mineralogy, section 4.2.4, Metamorphic Minerals, for bonds broken and atoms migrating by solid state diffusion or transported short distances by intergranular fluids; and the parent page, 4.2, Forming Crystals, for the three mechanisms and the five sections they divide into.
- Open Petrology, chapter 9, Introduction to Metamorphism, for metamorphism as change with no melting involved, and for the water-rich character of metamorphic fluids.
- Open Petrology, chapter 10, Metamorphic Fabrics, for porphyroblasts in mica-rich schists, for porphyroblasts being anhedral but often subhedral or euhedral, and for grain size rising with metamorphic grade.
- Katharine Cashman, “Crystal Size Distribution (CSD) Analysis of Volcanic Samples”, Frontiers in Earth Science, 2020, for nucleation rate against growth rate and for supersaturation as effective undercooling.
- Stephen Nelson, EENS 212 Petrology, Introduction to Igneous Textures, Tulane University, for the three undercooling cases and for crystals growing until they impinge or the supply is cut off.
- USGS, What are igneous rocks?, for slow cooling giving coarse grains and quick cooling giving fine-grained or glassy texture.
