Gemini’s emblem is a pair of twins, and the stone at the head of its list is the one that visibly does two things at once. Agate, then citrine, tiger’s eye, aquamarine, pearl and chrysocolla.
A crystal shop asked for a Gemini stone reaches for agate, and the reason sits on the surface of the piece. Agate is banded. Two colours, or four, laid in parallel through one stone, which is about as close as a mineral comes to being two things at the same time.
That is a selection rule, and it is not a hidden one. Where a sign has an emblem somebody can picture, crystal practice matches the picture, and it does this openly. Sunstone reached Leo that way, and the fire signs collected their red and orange palettes the same way.
What the rule selects, in Gemini’s case, is not a mineral. Nothing sold as agate has an entry in the Handbook of Mineralogy or a place on the International Mineralogical Association’s list of valid species, because agate describes how quartz has grown and not what it is.
One substance, a shelf of names
The Handbook’s quartz entry disposes of the whole family in a parenthesis, defining the fine-grained to microcrystalline form as chalcedony. Everything after that is a description of pattern.
Banding makes chalcedony an agate. Add dendritic inclusions, the little black feathers that look like moss trapped in the stone, and it becomes moss agate; flatten the bands out parallel and it becomes onyx. Iron oxide turns it red, at which point the name is carnelian, or jasper if the piece is opaque. One species, hardness 7 throughout, and a shelf of different words, none of which tells a buyer anything the previous word did not.
A picture-matching rule finds this convenient, because a look is exactly what it is shopping for and chalcedony supplies more looks than most common materials. Anybody trying to sort the subject out finds it maddening.
Grape agate, which is neither
The name fails in two directions at once on a single common object. Grape agate is a trade name for botryoidal purple chalcedony, a tight cluster of small spheres with no banding on it anywhere, so in the strict sense it is not an agate at all. It is usually sold as one, and the name is what people search for. It is also material with a pronounced light-dependent colour shift. Indoors a piece looks muted, almost grey-lilac. Carried into direct sun the purple comes up so far that photographs of the same specimen taken an hour apart look like two different objects.
A cluster ordered online for this collection on the strength of a listing photograph arrived noticeably duller than the picture, and nothing about the material had been misrepresented. The photograph had been taken in strong sunlight, with the light passing through the piece instead of bouncing off it. That is a condition a buyer at home will almost never reproduce, and on material with a shift this strong the indoor appearance is the one to buy against.
The name gets the mineral wrong, says nothing about the structure, and does not predict the colour the piece will show. It has still sold every specimen it was ever attached to. That is the picture-matching rule working exactly as designed, on a stone where the stakes are a disappointing parcel rather than an argument about a tradition.
Topaz and sapphire, neither of which stuck
Cornelius Agrippa, printing his table of correspondences in 1533, gives Gemini topaz, on a row that carries the twelve months as well as the twelve signs. Morales, in a Spanish treatise of 1604 that Kunz calls an old list probably representing Arab tradition, gives sapphire. Kunz’s own zodiacal list of 1913, the one the twentieth-century crystal trade inherited, gives agate, and carries no citation of any kind in a chapter otherwise densely footnoted.
Topaz, sapphire, agate. Neither of the first two appears on any modern Gemini list, and nobody has to explain why, because no compiler was answerable to any other. Sign stones were never a single thing handed down; they are a handful of separate lists that happen to cover the same twelve slots. The Aquarius page sets out how few of the surviving attributions match a modern list, and by which test.
What is left, once the tables have cancelled each other out, is a picture, which is at least a reason a reader can check without a library.
The planetary layer is the one part of this material that is genuinely documented, and it points somewhere else. Ptolemy gives Mercury both Gemini and Virgo in the second century. Marsilio Ficino, in 1489, names the planetary stones in a single sentence of Latin, and Mercury’s is an emerald it holds jointly with Venus and the Moon. Emerald is on no modern Gemini list at all.
The other five
Citrine is quartz with a yellow colour. GIA states its origin directly, that because yellow quartz colours are rare in nature most citrine is the result of heating, which converts less valuable shades of purple amethyst to the golden shades of citrine. A rule that selects on appearance has no way to register that, and neither does a shop shelf.
Tiger’s eye is quartz again, and it earns its place on the same logic as agate. A silky band of light travels across the polished surface as the stone turns, so a piece is never showing you one thing. Chatoyancy is the word for it, and parallel fibrous inclusions running through the stone are the cause, catching light along one axis and not the others. The band sits inside the stone and moves with the viewer, so a photograph fixes it in one position and the stone will not stay there. Aquamarine is beryl again, in blue, and its presence here is harder to account for, since it is the March birthstone and Gemini contains no part of March. Pearl comes from June, which Gemini does touch, and is not a mineral at all.
Chrysocolla is the sixth and the strangest. It does have a Handbook entry, and the entry is hedged in a way none of the others is. The formula carries brackets around two of the metals, the crystal system is given as orthorhombic with a question mark after it, and the point group is recorded as not determined. It forms in the oxidised portions of copper deposits and is commonly cryptocrystalline, opaline or enamel-like. A reference work definite about several thousand species declines to be definite about this one.
Agate, carnelian and onyx all resolve to one entry in the Crystalance Mineral Library, and the entry is quartz. Filing by species cuts directly across the way these stones were chosen, which is what makes it useful here.
None of that is an argument against choosing them by the picture. Matching a stone to an emblem is a legible thing to do and a reader can audit it in a second, which is more than a table nobody can date will ever allow. The cost is that the answer stops being a mineral. Ask what crystal is for Gemini and the truthful answer names a pattern of growth, and that pattern can turn up on any piece of chalcedony in the shop.
Sources
- Handbook of Mineralogy, quartz, on the fine-grained form: handbookofmineralogy.org
- Handbook of Mineralogy, chrysocolla: handbookofmineralogy.org
- Cornelius Agrippa, Three Books of Occult Philosophy, Book II ch. XIV, in the 1651 English translation of the 1533 original: archive.org
- George Frederick Kunz, The Curious Lore of Precious Stones, 1913, ch. X: gutenberg.org
- Ptolemy, Tetrabiblos I.17, Robbins translation: penelope.uchicago.edu
- Marsilio Ficino, De vita libri tres, Book III, Lyon 1560: archive.org
- GIA, treated citrine: gia.edu
