Identification · Morgan Hill Poppy Jasper Field Guide

Identification key

What are you holding?

Only the top left is orbicular. The test is not that the bodies are round - it is that each one has a center with fibres radiating out from it, or rings around it. A plain round spot of solid color is mottled, and mottled chert is not evidence for H1. Diagram, not a photograph: schematic, drawn to isolate the structural difference.

An orb is a structure, not a shape. Under the hand lens a true orb shows a core with fibres radiating from it, and often one or more rings - the project’s own definition, “rounded, radiating spherulitic bodies set in a chert groundmass” (GLOSSARY.md), and the 1930 first notice of this material describes exactly that: “background of brown and red jasper intermixed with white spherules of quartz” tier 2. The lower body is a plain patch. Both look like “spots” at arm’s length; only one of them is an orb.

Franciscan chert tier 1

This is the one you must be able to name, and it is the reason this page exists. USGS OFR 98-795 describes the Franciscan melange here as containing blocks of chert that are “red, white or green, radiolarian-bearing, and locally tuffaceous”, and defines unit fmc as “Red to green radiolarian chert” tier 1. Dibblee’s 1973 Mount Madonna sheet maps bodies of fc, “varicolored chert”, in the Watsonville Road corridor tier 1. Read that carefully: the sheet was georeferenced on 2026-08-22 and PJ-0016’s coordinate sits on unlabeled valley floor, with the nearest fc body 714 m west. Three bodies fall inside the entry’s ±1 km of along-channel uncertainty, so the host is not settled - the chert is nearby, not underfoot at the recorded point.

BreakCurved and glassy, like the inside of a shell. Edges sharp enough to cut.

EdgeA thin chip glows at the very edge against the sky. Body opaque.

BeddingOften in flat parallel ribbons a few centimeters thick, separated by thin partings. Also occurs massive, with no bedding at all.

HardnessWill not be scratched by a steel key. It scratches the key.

ColorRed is the common one here, but white, green and brown are all normal and all count.

Green and white chert still fire the control. The control asks whether you can see Franciscan chert, not whether you can see red rock. Counting only the red pieces would make the control easier to fail on a bar that is full of chert.

Hornitos poppy jasper tier 6

The honest answer is that you cannot separate them by looking. Russ Kaniuth, in Rock & Gem: “Hornitos poppy jasper is often mistaken for Morgan Hill poppy jasper” tier 6. Hornitos is in Mariposa County, roughly 150 km away and in a different geological setting entirely.

TellProvenance, not petrography. A piece you picked up yourself in a Llagas Creek gravel bar is not Hornitos material. A piece handed to you, bought, swapped or found in a box is not evidence of anything about location, whatever it looks like.

This is why the key asks where the rock came from before it asks anything about the rock. A dealer label documents a label. Dealer labels are the largest single source of bad locality data in mineral collecting, because they are relabeled at every sale and “Morgan Hill” is desirable enough to be applied loosely.

Manganese-stained chert tier 2

What state geologist E. F. Davis actually described in place at the manganese prospects on 3 November 1917, printed in Trask 1950: “The area west of Mr. Kelley’s house consists principally of radiolarian chert, much of which is thin-bedded and of intrusive greenstone and serpentine. At several places the chert is brightly colored as a result of the contact action of these igneous rocks, and in some places it is stained with manganese oxide. The massive chert is cut by numerous fissures filled by oxides of manganese which form stains and impregnations” tier 2.

TellSooty or metallic black films along cracks, black dendrites like tiny ferns on a joint face, or a general dull black impregnation. It comes off dark on your fingers where it is soft.

MattersThis is the material at PJ-0004, PJ-0005 and PJ-0006. Whether it is the same rock as the gem jasper is hypothesis H1, and it is not established. Finding brightly colored, manganese-stained chert with no orbs at the lot 16 district is the observation that would weaken H1.

Silicified limestone tier 1

El Toro maps as greenstone with silicified-limestone lenses, which is a different host rock from the varicolored chert at the historic deposit tier 1. USGS MF-2373 describes the associated assemblage in Marin Headlands basalt as “recrystallized limestone altered to calcsilicate mineral assemblages (epidote, pumpellyite, calcite-aragonite, and hematite)” tier 1.

TellPale grey, cream or greenish white. Dense and heavy for its size. May show pale green or yellow-green patches of calcsilicate. Often has a chalky or sugary rather than glassy break.

NotIt does not glow at a thin edge the way chert does, and it rarely gives clean conchoidal fracture.

MattersRelevant to H3. Orbicular jasper found genuinely in place on El Toro would not merely weaken the El Toro-is-an-error finding, it would break the host-rock model.

Silica-carbonate rock, and jasperised serpentine tier 1

USGS OFR 98-795, unit sc: “Silica-carbonate rock (Miocene?) -- Present locally along the Sargent and Berrocal faults, associated with hydrothermal alteration of serpentinized ultramafic rocks” tier 1. It is what serpentine turns into when silica-bearing fluids work on it, and it can be strongly red-brown and hard.

TellRusty orange to buff-brown, often porous or cavernous with a boxwork of ribs. Noticeably lighter in the hand than dense chert. Frequently has pale grey-green serpentine remnants still in it.

MattersThis is the rock H5 turns on. Bowen 1962, a CDMG geologist, puts the host at Paradise Valley in “veins and irregular masses in altered serpentine” tier 2 - not the Franciscan chert this project has been modeling. Host rock at the outcrop is the single observation that decides it, and it also moves H1, because the manganese prospects sit in serpentinite and melange.

One hill carries both candidate host rocks, so one competent look tests both models. Dibblee’s plate 4, georeferenced 2026-08-22, maps serpentine sp 225 m from PJ-0014 at the creek bend and fc varicolored chert about 500 m away on the same hill tier 1. Bowen’s model and this project’s are not geographically separable here. So the field question is not which rock is nearby - it is which rock the jasper is actually in. Record the host, not the neighborhood.

Serpentinite tier 1

Mapped at PJ-0004 and against the contact at PJ-0005 tier 1. California’s state rock, and abundant in this corridor.

TellGreen to blue-green to nearly black. Soapy or waxy to the touch. Soft: a steel key scratches it easily. Often slickensided - polished, grooved, slippery-looking surfaces. Cobbles are smooth and rounded rather than sharp.

NotNever conchoidal, never glass-hard. If a key scratches it, it is not chert or jasper.

The ordinary Llagas Creek assemblage tier 1

This is what you will actually be holding, nearly all day. CJMG 1954 on Llagas Creek gravels: “gravels are composed chiefly of rock types common to the Franciscan formation ... arkosic sandstone, chert, shale, basalt, gabbro, and quartz” tier 1.

Jasper is not in that list. That is a genuine negative and it is the strongest single piece of evidence against H4, the downstream-concentration hypothesis. It is weakened, but not removed, by Bulletin 152 p.269 stating outright that “The jasper is strikingly similar to the red chert of the Franciscan group” tier 2 - in the vocabulary of the period, chert and jasper are not reliably distinguished.

SandstoneArkosic: you can see individual sand grains with the lens, including blocky pink or white feldspar. Gritty, not glassy. Grey to buff.

ShaleSplits into flat sheets. Dull. Soft enough to mark with a fingernail. Dark grey to black.

BasaltDark grey to greenish black, very fine grained, sometimes with small round gas holes or white and green blebs filling them. Tugs a magnet slightly. Where metamorphosed it is greenstone, which is the same rock and the one H2 cares about.

GabbroDark and coarse. You can see separate black and white crystals without a lens. Heavy.

QuartzWhite, grey or clear. Glassy, greasy lustre on a break. No color banding, no orbs. Very hard.

Wetting it

Wet the rock before you judge anything. A dry creek cobble carries a dulling film of silt and a weathered rind, and both hide color, orbs and banding. Water restores the color a cut face would show, at no cost and with no tools.

DoDip it, or spit on it and rub with a thumb. Look at it wet, in the shade of your own body if the sun is direct.

WarningWet rock oversells. Everything looks better wet, including ordinary chert. Judge structure wet; judge how good a piece is when it is dry.

Conchoidal fracture

A conchoidal break is a smooth curved dish, often with faint ripples spreading from where it was struck, and it leaves an edge sharp enough to cut. It means the rock has no grain and no crystals to break along, which in this creek means silica: chert, jasper, chalcedony or quartz. Diagram, not a photograph.

WhereLook at a chipped corner or an old break, not at the water-worn outside. A rounded cobble may have no fresh surface at all, in which case skip this question rather than guessing.

NeverDo not hammer anything that is in place. Read the breaks the creek already made.

Light through a thin edge

Only the thinnest millimeter of a chipped edge will show it. A thick edge, or a water-worn cobble with no chip at all, reads dead opaque whatever the rock is - so a “no” from a thick edge is worth nothing. Chert and jasper are opaque in the body but pass light through a genuinely thin edge, giving a warm glow. Quartz and chalcedony pass much more. Basalt, gabbro, sandstone and shale pass none at all. Diagram, not a photograph.

Weak testTranslucency is a soft character and it varies within one piece. A “not sure” here is completely reasonable and the key will carry it forward as a flag rather than pretending you decided.

Hardness, with a key and a streak plate

Two cheap tests that between them split the silica family from nearly everything else in this creek.

KeyPress a steel key hard across a clean flat spot. Steel is about 5.5 on the Mohs scale. If it leaves a groove, the rock is softer than steel: serpentine, shale, weathered basalt, silica-carbonate rock. If the key skates off and leaves only a metal smear you can rub away, the rock is harder: chert, jasper, quartz.

PlateDrag a corner across unglazed porcelain. Chert and jasper leave no powder and may scratch the plate. A rusty red-brown powder is hematite. A pale green-white powder that feels slick is serpentine. Black powder is manganese oxide.

CarefulA weathered rind is always softer than the rock under it. Test a fresh surface, or test twice.

Reference plates

Real rocks from this creek

Every specimen below was picked up in the upper Llagas catchment above Chesbro on 2026-08-22 and identified in hand by physical test, not by looking. None is a dealer specimen. The exact collection point is held internally and is not published.

These plates speed recognition. They do not make determinations. Every rock here was settled by hardness, streak or a scratch comparison - characters no photograph carries. Match a picture and you will be wrong eventually; test a character and you will not.

Wet, side by side: the mistake that actually happened

Grey greywacke - NOT chert, wet — Grey greywacke - NOT chert

Red chert, crackle-veined, wet — Red chert, crackle-veined

The left piece was field-logged as "grey chert." It is not chert - a steel blade scrapes it. The discriminator is lustre. Chert goes waxy wet; greywacke stays matte. Wet every candidate before you judge it.

Wet, side by side: chert is not only red

Green Franciscan chert, wet — Green Franciscan chert

Red chert, brecciated, wet — Red chert, brecciated

Both fire the chert control. USGS OFR 98-795 defines the chert here as "red, white or green" (tier 1). A search image built only on red rock will miss half of it - and a chert control that misses chert reports a failed search as an absence.

The four tests

What am I holding?

Sorted by color, because that is what you notice first. A dashed tile is a rock the project has not photographed yet - it opens to the drawings and the text.

What is still missing. Bedded ribbon chert, basalt with pillow structure, quartz and shale have no entry of their own, and the target has no photograph. The dashed tiles are the work list, and they are shown rather than omitted so the guide cannot read as complete when it is not.

Orbicular jasper

THE TARGET

•the target

This is what you are looking for, and the project owns no photograph of it. The drawings below are the sharpest statement of the boundary that decides H1: an orb is a structure, not a shape. A plain round spot is mottling.

Not yet photographed. What follows is the authored description and, where one exists, a diagram. A drawing is not a specimen and this entry says so rather than implying otherwise.

Red chert / jasper

the large specimen

•fires

Llagas Creek, upper catchment above Chesbro, Santa Clara County.

Red chert, the large specimen, wet — wet · Rich brick red, glossy, white veinlets.

Red chert, the large specimen, dry — dry · Duller, weathering rind evident.

What settled it

Too hard to scrape with a steel nail. Examined under a hand lens: no orbs, no radiating structure, no concentric rings.

What it shows

The best-quality red chert from the visit - glossy and waxy wet, with white quartz veinlets branching through a brick-red groundmass.

Watch for

This is what NOT orbicular looks like on good material. Pale material as branching vein fill following fractures. The hand-lens check is what settled it; a photograph could not.

Red chert, brecciated

angular clasts in vein fill

•fires

Llagas Creek, upper catchment above Chesbro, Santa Clara County.

Red chert, brecciated, dry — dry · Angular blocks and pale veins at their clearest.

What settled it

A steel nail does not mark it, so harder than about 5.5; chert and jasper are 7.

What it shows

Broken and healed. Angular dark-red fragments separated by pale silica vein fill.

Watch for

This is the confusable that most resembles poppy jasper without being it - red groundmass, pale pattern, striking wet. It is definitively not orbicular: the pale material is angular vein fill between angular clasts, not rounded bodies with internal structure.

Red chert, crackle-veined

fine pale network

•fires

Llagas Creek, upper catchment above Chesbro, Santa Clara County.

Red chert, crackle-veined, dry — dry · Fine pale crackle network visible.

What settled it

A steel nail does not mark it. Used as the hardness reference that identified the green chert.

What it shows

The same rock as the brecciated plate in a different habit - a fine pale network crazing through the groundmass rather than coarse blocks.

Watch for

Two habits of one rock. A guide that shows only one teaches a narrower search image than the creek actually contains.

Hornitos poppy jasper

the one you cannot rule out by looking

•not chert

Not yet photographed. What follows is the authored description and, where one exists, a diagram. A drawing is not a specimen and this entry says so rather than implying otherwise.

Manganese-stained chert

sooty black on joints

•fires

Not yet photographed. What follows is the authored description and, where one exists, a diagram. A drawing is not a specimen and this entry says so rather than implying otherwise.

Green Franciscan chert

chert is not only red

•fires

Llagas Creek, upper catchment above Chesbro, Santa Clara County.

Green Franciscan chert, dry — dry · Sage to grey-green, matte.

What settled it

A steel nail does not mark it. Mutual scratch against confirmed chert (piece 6), run both directions: no mark either way, so equal hardness at about 7. Greenstone at about 6 would have been scratched.

What it shows

Franciscan chert here is not only red. USGS OFR 98-795 defines unit fmc as "Red to green radiolarian chert" and describes melange chert blocks as "red, white or green" (tier 1). This plate is that mapping confirmed in hand.

Watch for

Greenstone is the confusable, and it looks almost identical. It is softer - about 6 - so a confirmed chert will scratch it. Color will not separate them.

Serpentinite

soapy, soft, green

•not chert

Llagas Creek, upper catchment above Chesbro, Santa Clara County.

Serpentinite, wet — wet · Deep olive-green, waxy, white vein bright.

Serpentinite, dry — dry · Grey-green, mottled, matte.

What settled it

A steel nail scratches it and a fingernail does not, so between about 2.5 and 5.5; serpentine is 2.5-4. Slightly slippery. Fully opaque at a thin edge.

What it shows

Dark olive-green, mottled with pale celadon, cut by a blocky white vein. Three serpentinites came from roughly 25 m of this channel.

Watch for

Silica-carbonate rock is what serpentinite becomes under silica-bearing fluids - rusty, porous, ribbed, light for its size. It is the rock H5 turns on, and its parent rock is demonstrably in this creek. Add it to the search image.

Silica-carbonate rock

what serpentinite becomes

•not chert

Not yet photographed. What follows is the authored description and, where one exists, a diagram. A drawing is not a specimen and this entry says so rather than implying otherwise.

Grey greywacke

NOT chert - the mistake that happened

•not chert

Llagas Creek, upper catchment above Chesbro, Santa Clara County.

Grey greywacke - NOT chert, dry — dry · Pale grey-buff, dusty, fine-grained.

What settled it

A steel blade scrapes it, so softer than about 5.5. Chert and jasper are 7 and would not scrape.

What it shows

A fine-grained clastic sedimentary rock, most likely Franciscan greywacke or sandstone. CJMG 1954 lists "arkosic sandstone" first among Llagas Creek gravel constituents (tier 1).

Watch for

This piece was field-logged as "grey chert" and it is not chert. It is the one misidentification the first visit actually produced. The discriminator is lustre: it stays matte when wet, where chert goes waxy. Compare the two wet plates side by side - that is the whole lesson.

Silicified limestone

pale, dense, sugary

•not chert

Not yet photographed. What follows is the authored description and, where one exists, a diagram. A drawing is not a specimen and this entry says so rather than implying otherwise.

Arkosic sandstone

visible grains, salmon feldspar

•not chert

Llagas Creek, upper catchment above Chesbro, Santa Clara County.

Arkosic sandstone, wet — wet · Darker salmon-brown, still matte.

Arkosic sandstone, dry — dry · Salmon-pink, granular, pitted.

What settled it

A steel nail scratches it. Individual sand grains visible to the naked eye, with pits where grains have plucked out.

What it shows

The salmon-pink color is potassium feldspar, and feldspar-rich sandstone is arkose. First-named constituent of CJMG 1954's tier-1 gravel description, now confirmed in hand.

Watch for

This is most of what you will pick up all day. Naming it correctly and quickly is what leaves time for the chert.

Coarse igneous rock

interlocking crystals

•not chert

Llagas Creek, upper catchment above Chesbro, Santa Clara County.

Coarse igneous rock, wet — wet · Contrast between white and black sharpens.

Coarse igneous rock, dry — dry · Grain boundaries clearest.

What settled it

A steel nail does not mark it. Coarsely crystalline - two distinct minerals, white and black, with real grain boundaries.

What it shows

Diorite, plagiogranite or a leucogabbro - white plagioclase feldspar with black amphibole or pyroxene. Gabbro is on CJMG 1954's list of these gravels.

Watch for

Crystalline texture is the tell. Discrete interlocking grains mean igneous or metamorphic - never chert, which has no grain at all.

The ordinary assemblage

shale, basalt, gabbro, quartz

•not chert

Not yet photographed. What follows is the authored description and, where one exists, a diagram. A drawing is not a specimen and this entry says so rather than implying otherwise.

Poppy Jasper Field Key

What are you holding?

Franciscan chert tier 1

Hornitos poppy jasper tier 6

Manganese-stained chert tier 2

Silicified limestone tier 1

Silica-carbonate rock, and jasperised serpentine tier 1

Serpentinite tier 1

The ordinary Llagas Creek assemblage tier 1

Wetting it

Conchoidal fracture

Light through a thin edge

Hardness, with a key and a streak plate

Record it here, now

Reports held on this phone

Real rocks from this creek

Wet, side by side: the mistake that actually happened

Wet, side by side: chert is not only red

What am I holding?

Orbicular jasper

Red chert / jasper

Red chert, brecciated

Red chert, crackle-veined

Hornitos poppy jasper

Manganese-stained chert

Green Franciscan chert

Serpentinite

Silica-carbonate rock

Grey greywacke

Silicified limestone

Arkosic sandstone

Coarse igneous rock

The ordinary assemblage