Mineral twinning explained
Twinning is recorded for 69 of the 108 mineral species we checked but called common for only 27. It is a fluorite cube interpenetrating, not a diagnostic test.
A twin is two or more crystals of one species grown together in a fixed, symmetrical relationship rather than at random. Of 108 species this site covers, 69 record twinning in the Handbook of Mineralogy, but only 27 describe it as common or better and 18 call it rare or uncommon. Twinning is a collecting feature far more often than it is an identification test.
In short
- A twin is not two crystals stuck together. It is one growth in which parts of the structure are related by a specific symmetry operation — a reflection, a rotation or an inversion — that is not already part of the crystal's own symmetry.
- Twinning is common in the reference and uncommon on the shelf. 69 of 108 species record it, but only 27 are described as commonly or very commonly twinned. For 18 the Handbook explicitly says rare or uncommon.
- That is why twinning rarely identifies anything. Finding a twin confirms what the species can do; not finding one tells you nothing at all, because most specimens of most twinnable species are not twinned.
- The named laws are the ones worth learning: Carlsbad, Baveno and Manebach on orthoclase; Dauphiné and Brazil on quartz; the spinel law on magnetite. A law names a plane or an axis in a quartz or fluorite crystal and nothing more.
- Where it does identify, it is decisive. Scolecite, mesolite and natrolite cannot be separated by hardness or density at all, and their twin laws differ — so twinning is the whole determination rather than a clue in it.
| Form | Species | Twin law as recorded | Frequency in the Handbook |
|---|---|---|---|
| Pseudohexagonal trilling | Aragonite | Repeated contact twinning on {110} | Common |
| Stellate pseudohexagonal | Cerussite | Simple or cyclic contact twins on {110} | Common |
| Cross or cogwheel | Bournonite | On {110} | Common |
| Six-pointed star | Muscovite | Composition plane {001}, twin axis [310] | Given without qualification |
| Star-shaped trilling | Arsenopyrite | On {012} | Common |
| Fiveling | Safflorite | {011} as twin plane | Given without qualification |
| Interpenetrant cube | Fluorite | On {111}, interpenetrant, flattened | Common |
| Cruciform, V-shaped, swallowtail | Gypsum | Contact on {100} | Very common |
| Carlsbad, Baveno, Manebach | Orthoclase | Simple, contact or penetration twins | Common |
| Ubiquitous cruciform | Stilbite-Ca | On {001} | Ubiquitous |
What a twin actually is
Two quartz crystals that happen to have grown against each other are an intergrowth. A twin is something stricter.
In a twin, the parts are related by a symmetry operation that the crystal's own structure does not already contain. Reflect one half across a particular plane, or rotate it about a particular axis, and it maps onto the other half. Because the operation is fixed by the structure rather than by chance, the same twin appears again and again on specimens from localities thousands of miles apart, at the same angles.
The vocabulary follows from that. The twin plane or twin axis is the operation, quoted in Miller indices — {110}, {001}, [310]. The composition plane is the surface the two halves actually meet along, which is often but not always the twin plane. A contact twin meets along a clean surface; a penetration twin has the two individuals passing through one another. Polysynthetic or lamellar twinning repeats the operation many times in thin slices. A trilling, fourling or fiveling is three, four or five individuals in a cyclic arrangement.
A twin law with a name — Carlsbad, Baveno, Manebach, Dauphiné, Brazil, spinel — is simply a particular operation common enough in a particular species to have earned one. There is no more to it than that, and the names are historical rather than systematic.
The count, and why twinning is a poor identification test
Textbooks list twinning among the diagnostic properties. We wanted to know how far that holds for the species a collection actually contains, so we counted.
Method. On 18 September 2026 we read the Twinning field on the Handbook of Mineralogy sheet for each of the 108 species this site covers, and classified the frequency word the Handbook itself uses.
69 of 108 record twinning at all. Of those 69: 27 are described as common, very common or ubiquitous; 18 as rare, very rare or uncommon; and 24 are stated with no frequency qualifier at all.
So on roughly a quarter of the species in a collection, twinning is something that happens often enough to expect. On another 18 it is explicitly a rarity — atacamite, autunite, azurite, chalcanthite, clinozoisite, cobaltite, conichalcite, corundum, hemimorphite, lepidolite, mimetite, pharmacosiderite, pyromorphite, schorl, siderite, stibnite, torbernite and uraninite.
The logic that follows is one-way and worth stating plainly. Finding a twin narrows the field. Not finding one narrows nothing, because most specimens of most twinnable species are untwinned. Twinning is therefore useful as positive evidence and worthless as negative evidence, which is the opposite of how a diagnostic property is usually used. Our note on identifying a specimen puts it where it belongs in the order.
The one place twinning is the whole determination
There is a group where twinning is not a supporting clue but the entire answer, and it is worth knowing about because it is the exception that justifies learning the laws.
Scolecite, mesolite and natrolite — the natrolite-group zeolites — cannot be separated by hardness or density. Scolecite is 5 to 5.5 and 2.25 to 2.29; natrolite is 5 to 5.5 and 2.20 to 2.26; mesolite is 5 and 2.26. All three ranges overlap, all three have perfect {110} cleavage, all three are colourless to white with a vitreous lustre.
The twin laws differ and are readable. Scolecite twins on {100} with [001] as twin axis, commonly, as contact or penetration twins. Mesolite is characteristically twinned on {010} or {100}. Natrolite twins on {110}, {011} and {031}. Our scolecite note works the group through in full.
The same pattern appears, less starkly, wherever physical properties converge: stilbite against heulandite, where stilbite's twinning on {001} is recorded as ubiquitous and heulandite's as merely present; and among the feldspars, where Carlsbad, Baveno and Manebach twins are how orthoclase is read in the field.
Twinning as a collecting feature, which is what it usually is
For most collectors most of the time, twinning matters because of what it does to a specimen rather than what it proves about one.
The forms worth looking for are the ones in the table above: aragonite's pseudohexagonal trillings, which make an orthorhombic mineral look hexagonal; cerussite's stellate and reticulated groups, among the most striking things a lead carbonate does; bournonite's cogwheels, which are unmistakable and are the reason the species has a nickname; fluorite's interpenetrant cubes; gypsum's swallowtails; and the six-pointed stars of muscovite.
A well-formed twin of a species that twins rarely is a genuinely scarce specimen, and that is worth recording on a catalogue card. A twin of a species the Handbook calls ubiquitously twinned — stilbite, for instance — is simply a normal specimen, and describing it as twinned on a label adds nothing. Our note on labelling and cataloguing covers what belongs on a card.
One practical warning. Twin boundaries are planes of weakness. A polysynthetically twinned crystal can part along the lamellae under shock in a way an untwinned one will not, and a reticulated cerussite group is among the most fragile things a cabinet holds. Our notes on handling and storage and on spotting repairs both bear on this, because twinned groups are repaired more often than most material.
We hold no catalogue and nothing here is offered for sale. Our wanted list sets out what we look for, and a sharply twinned specimen of a rarely twinned species is on it. The rest of our collecting notes cover identification and specimen care.