Digital TownsCrystalsShop crystals
  1. Crystals & Minerals
  2. How they form
  3. Dendrites

Mineral Dendrites: The Fern Patterns in Stone That Are Not Fossils

By Digital Towns Crystals ยท Last reviewed October 8, 2026

Short answer: mineral dendrites are branching, fern-like growths of black or brown minerals, usually manganese oxides, that crystallize from fluids seeping along thin cracks and bedding planes in rock. They look like pressed plants because branching growth follows the same physics in minerals, snowflakes and metals, not because anything ever lived there. Analysis at Caltech found them made of minerals such as hollandite, cryptomelane, romanechite and todorokite, and, despite many textbooks, not pyrolusite.

Are dendrites fossils?

No, and the confusion is understandable. A good dendrite on a split slab of limestone looks exactly like a fern frond or a tiny tree, complete with trunk, branches and twigs. The classic slabs even come from fossil country: George Rossman's manganese dendrite page at Caltech pictures dendrites from the limestone quarry near Solnhofen, Germany, and the University of California Museum of Paleontology describes the Solnhofen Limestone as famous for fossils of exquisite detail, including fragile and soft-bodied organisms. Real fossils and mineral dendrites can sit in the same quarry, sometimes on the same slab.

The difference is material and structure. Rossman defines manganese dendrites as branching features containing manganese oxides found on rock surfaces and in cracks in rocks. Under a scanning electron microscope, he shows a cryptomelane dendrite resolving into individual crystals in a field of view only 30 micrometers wide. There is no carbon film, no cell structure, no repeated leaf shape: just mineral grains stacked into a branching pattern.

Four quick checks: dendrite or fossil?

Check Manganese dendrite Plant fossil
Where it lies On a fracture or bedding surface, often spreading from a crack edge Within a bed, with consistent orientation of leaves
Branching Irregular, every tip splits again at slightly different angles Repeated, species-specific leaf and vein pattern
Thickness A stain with no relief; scrapes off on weathered surfaces Often a carbon film or impression with slight relief
Symmetry Fans outward from a source, no midrib Midrib and paired leaflets

How does a dendrite grow?

The mechanism comes from the physics of solidification, and the pieces are well understood even where the mineral details are not.

  1. A thin film of fluid enters a crack. Groundwater carrying dissolved manganese, often with iron, barium or potassium, seeps along a hairline fracture or bedding plane. The space is extremely thin, which is why most dendrites are flat.
  2. A seed forms. Where conditions change, manganese oxide starts to precipitate at a few points, often along the edge of the crack where fluid enters. Rossman shows a specimen from the Panamint Mountains, California, with dendrites formed from numerous nucleation centers on the rock surface.
  3. The growing edge becomes unstable. As mineral precipitates, the fluid next to it is depleted. Any bump that sticks out reaches fresher fluid, grows faster and sticks out further. H. K. D. H. Bhadeshia's Cambridge notes on dendritic solidification explain the general rule: a dendrite tends to branch because the interface instability applies at all points along its growth front.
  4. Branches split again. Each new tip faces the same instability, so it branches in turn, building the tree shape. Growth stops when the fluid stops moving or the supply runs out.

Snow shows the same physics in a form anyone has seen. Kenneth Libbrecht at Caltech explains that when the branching instability applies itself over and over to a growing snow crystal, the result is called an ice dendrite, and that crystals grown in air below atmospheric pressure have fewer branches because diffusion limits growth less. The supply of material, not biology, is what draws the tree.

Which minerals make dendrites?

Mineral Group Example noted by Rossman
Hollandite Hollandite group Mn oxide Afton Canyon and the Argus Range, Mojave Desert, California
Cryptomelane Hollandite group Mn oxide Southwestern USA
Romanechite Mn oxide Siltstone near Barstow, California; South Dakota
Todorokite Mn oxide White Queen beryl mine near Pala, San Diego County
Coronadite Hollandite group Mn oxide Listed among dendrite minerals
Cinnabar Mercury sulfide Funderburk mine, Pike County, Arkansas

Rossman makes three points that matter for anyone labeling a specimen. First, dendrites are commonly but incorrectly believed to be pyrolusite (MnO2): his group never found a dendrite containing it. Second, dendrites from one locality tend to be a single mineral. Third, one can not tell the mineralogical composition of a dendrite by its physical appearance. The same black fern can be romanechite in one place and hollandite in another.

Minerals.net agrees on the identification problem in general terms, noting that the manganese oxides are all dark gray to black and may be very difficult to distinguish from one another, especially in massive form. Its pages for romanechite and psilomelane list dendritic and arborescent among their habits.

Manganese is not the only option. Native copper grows true three-dimensional dendrites, and Minerals.net's copper page names the Itauz mine at Dzezkazgan, Kazakhstan, for very intricate dendritic crystals and illustrates dendritic copper from the Ray mine in Pinal County, Arizona.

What is still unknown about mineral dendrites?

The branching physics is general, but several mineral-specific questions remain open in the sources used here.

  • Why one mineral per locality? Rossman reports the pattern but the page does not explain it. Local fluid chemistry is the obvious suspect, yet the sources do not show that it has been tested site by site.
  • What sets the charge balance? In these manganese oxides, Rossman notes, other cations sit in open channels in the structure, and the average oxidation state of manganese is typically 3.4 to 3.7. How fluid chemistry controls that mix in a given dendrite is not addressed.
  • Mixed dendrites. Rossman illustrates a mixed dendrite containing hollandite, braunite and todorokite from Badersdorf, Austria, so a single "tree" can hold several minerals. Whether they grew together or one replaced another is not stated.

None of this reopens the fossil question. It means that the black pattern on a slab should be labeled "manganese oxide dendrites" unless someone has actually analyzed it.

Dendrites, moss and plumes: what is the difference?

The agate trade uses plant names loosely, and the inclusions behind them differ.

  • Dendritic agate. Minerals.net describes it as chalcedony containing manganese oxide impurities that form tree-like patterns. These are true dendrites, trapped inside the silica.
  • Moss agate. Minerals.net's agate entry describes moss agate as chalcedony with dense inclusions of green hornblende that make the pattern resemble moss. Green, fuzzy and three-dimensional, it is an inclusion texture rather than a flat dendrite.
  • Sweetwater agate. The same page describes Wyoming agate from the Sweetwater River with star-shaped patterns of manganese oxide inclusions, and classes it as a form of moss agate.
  • Plume agate. Minerals.net's plume agate entry defines it as agate with inclusions in feather-like patterns, a different texture from a flat dendrite spreading along a crack.

Minerals.net also notes that dendritic and moss agate are not true agates because they lack banding, though they are traditionally called agates. Our agate guide covers banded agate, and the inclusions guide covers other mineral guests inside quartz.

Choosing plant-pattern agates and what to check

The pieces we stock now are the moss, tree and plume side of the family rather than flat manganese dendrites on limestone, and that distinction is worth knowing when you compare listings elsewhere.

  1. Ask what the pattern is. Black, flat and fern-like suggests manganese oxide; green and tufted suggests hornblende or chlorite moss.
  2. Look at it edge-on. Surface dendrites on a rock face are a stain a fraction of a millimeter thick and can be scuffed; dendrites sealed inside chalcedony are protected.
  3. Beware of painted or printed "dendrites." Real dendrites branch irregularly at every scale under a loupe; repeated identical fronds are a warning sign.
  4. Price follows the picture. A crisp, isolated tree in clear chalcedony is worth far more than a cloudy tangle.
  • Moss Agate: larger moss agate pieces; tilt them to see the three-dimensional green tufts that separate moss from flat dendrites.
  • Tree Agate: an inexpensive example of an agate named for branching green patterns; compare it with photos of black manganese dendrites.
  • Plume Agate (Pink): feather-shaped plumes, the third plant-like texture, best seen with light behind the stone.
  • Moss Agate Fairy Guardian .77lb: a Brazilian moss agate carving of about 349 g where inclusions run through a polished, curved surface.

Digital Towns Market

Agates with plant-like inclusions in stock

See all crystals and minerals

Frequently asked questions

What are the black tree patterns on rocks?

Usually manganese oxide dendrites: branching growths of minerals such as hollandite, cryptomelane, romanechite or todorokite that crystallized from fluid seeping along a crack or bedding plane. They are mineral deposits, not plant fossils, and are common on limestone, sandstone and siltstone surfaces.

Are dendrites made of pyrolusite?

Usually not. Many books say so, but George Rossman's group at Caltech reports never finding a dendrite containing pyrolusite. The dendrites they analyzed were other manganese oxides, with one mineral tending to dominate at any one locality.

Is moss agate the same as dendritic agate?

No. Dendritic agate contains flat, tree-like manganese oxide inclusions. Moss agate contains dense green hornblende inclusions that look like moss. Both are chalcedony without true agate banding, but the patterns and minerals are different.

Can dendrites be fake?

Yes. Patterns can be painted, printed or etched onto stone, and some dyed material imitates them. A real dendrite branches irregularly at every magnification and lies on a crack or inside the stone; repeated identical shapes or patterns sitting on a polished surface coat suggest an imitation.

How old are mineral dendrites?

They are younger than the rock they sit on, since they grew in cracks after it formed, but the sources here give no dating method for them. A dendrite on a Jurassic limestone slab is not necessarily Jurassic itself.

Sources