A Field Guide to Fossils: What People Get Wrong First

The first thing people get wrong with a field guide to fossils is the order of operations. The instinct is to open the book and hunt for a photograph resembling the rock in your hand. Identification runs the other way: establish where and when the rock formed, then open the guide, because age and formation eliminate most candidate species before you compare a single image. The National Audubon Society Field Guide to North American Fossils covers a continent in 846 pages, according to the Internet Archive's catalog record of the 1982 Knopf edition. The Paleontological Research Institution's Devonian Atlas of Ancient Life covers over 200 Devonian species from New York, Michigan, and Wisconsin, and it names the rock formations each one comes from. If your rock is Devonian, the second guide reaches an answer faster.

I came to this from a fault desk rather than from paleontology. On nights in a network operations center, a city's dropped calls arrive as a list, and every line invites you to blame the handset. I spent one whole shift doing exactly that: a cluster of no-service reports, all filed as device faults, for eight hours. One failed rectifier at a single cell site had caused every one, and the evidence had been on my screen since midnight, because the reports shared a coverage footprint. The rock unit is the cell site. The photograph is the handset.

What do people get wrong first with a field guide to fossils?

They use the guide as a lookup table for shapes, and shape is the least reliable thing a fossil hands you. Two-shelled specimens are the classic trap: brachiopods and clams both produce them, and the difference sits in the symmetry rather than the outline. William H. Matthews III made that the backbone of his identification key in Texas Fossils: An Amateur Collector's Handbook, published as Guidebook 2 by the Bureau of Economic Geology at the University of Texas at Austin. His key sorts specimens by radial versus bilateral symmetry, and he tells the reader plainly that "to use the key the beginner should know something about symmetry."

Preservation compounds it. The same animal reaches you as a mold, a cast, or a worn fragment, and Matthews separates those states in his opening pages before showing a single species.

None of this makes the guide useless. It changes what the guide is for. A field guide confirms a hypothesis you narrowed by other means, and it performs badly as the instrument that generates one.

Why does the rock unit narrow the answer faster than the photograph?

Because occurrence data excludes, and photographs only suggest. Take one species page from the Devonian Atlas of Ancient Life. Mucrospirifer mucronatus, a brachiopod, carries a geologic range spanning the Middle and Late Devonian, twenty-six named formations and members including the Moscow, Ludlowville, and Skaneateles, and occurrences in Indiana, Michigan, New York, Ohio, Ontario, Pennsylvania, and West Virginia. Its description, from Wilson (2014), gives a size of 20 to 50 mm and wing-like alae extending as much as 2 cm beyond the shell.

Every one of those fields is a filter. If your specimen came out of a Cretaceous limestone in central Texas, that page is gone before you look at its pictures, along with every other Devonian brachiopod in the atlas.

Getting the age right requires a source that states its own version. The International Commission on Stratigraphy publishes the International Chronostratigraphic Chart with a version stamp, and the current one, v 2026/06, places the base of the Devonian at 419.62 ± 1.36 million years ago and its top at 358.86 ± 0.19 million years ago. The Ordovician runs from 486.85 ± 1.5 to 443.1 ± 0.9 million years ago on the same chart. Those numbers get revised; the June 2026 release adjusted three stage boundaries in the Triassic and Permian. A figure copied from a guide printed in the 1980s will orient you, though it is worth knowing which version you are quoting on a specimen label.

For the rock unit itself, your state geological survey is the practical source; Matthews built a simplified geologic map of Texas into his handbook for that reason.

What does the "×6" under a fossil photograph actually mean?

It is the magnification of the printed image, and skipping that line is the fastest way to misidentify something small. In Matthews's Texas handbook the plate captions run from ×¼ to ×25; a caption reading "Avonia sp., ×6" describes a specimen printed at six times life size. A bryozoan colony reproduced at ×15 fills the page and looks nothing like the gray crust you picked up.

The correction costs nothing. Measure the specimen in millimeters before you open any guide, write it down, and treat a photograph whose caption lacks a scale as an illustration rather than as evidence. Online atlases hold a structural advantage here: the Devonian Atlas states a measured size range in the species description, so the number travels with the identification instead of living in a caption you might not read.

The Digital Atlas project also maintains a Virtual Collection of photogrammetry models, and its Sketchfab account held 635 models when I checked in September 2026, all under Creative Commons Zero licensing and free to download and 3D print. For a question about shape, rotating a model beats one fixed viewpoint, and printing one at a known size is the closest thing to a reference specimen most amateurs will hold.

How do you tell a fossil from something that only looks like one?

Some rocks imitate life convincingly. Matthews grouped them under a heading his readers still need: pseudofossils, or "false fossils." Three account for most misidentifications.

Dendrites are the common one. Those branching black patterns spreading across a bedding plane look exactly like fern fronds, and they are thin incrustations of manganese dioxide. Matthews records them along the bedding planes of Cretaceous and Paleozoic limestones across much of Texas, which is to say in the rocks a Texas collector is most likely to split.

Slickensides are striations produced when rock surfaces grind past each other during fracturing, and they pass for Pennsylvanian coal plants. The handbook's test is geometric and takes two seconds: "Since slickensides are commonly at an angle to the bedding plane and plant remains lie parallel to the bedding plane, the two are usually easily distinguished." Concretions round out the set, weathering into shapes that suggest eggs, bones, or turtle shells.

I like the bedding-plane test because it reads a relationship rather than an appearance, which is what fault-desk work rewards: an alarm alone means little, an alarm timed to a power event means something.

National guide, regional atlas, or state handbook: which should you use?

These three do different jobs, and the honest comparison is granularity rather than quality.

| Property | Audubon national guide (print) | Devonian Atlas of Ancient Life (online) | Matthews, Texas Fossils (print, free digital) | | --- | --- | --- | --- | | Scope | North America north of Mexico | Devonian of New York, Michigan, Wisconsin | Common fossils of Texas | | Organized by | Form and color | Time period and region, then taxonomic group | Rock type and taxonomic group, plus a state geologic map | | Stated coverage | 846 pages, 1982 Knopf edition | Over 200 species; its index listed 246 species pages in September 2026 | 49 plates, first printed 1960 | | Locality detail | Period and broad distribution | Named formations and members, plus state occurrence | Texas formations and regions | | Scale handling | Photographic plates | Measured size ranges in the text | Caption magnifications, ×¼ to ×25 | | Cost and access | Purchased, works with no signal | Free, requires connectivity | Free full text at Project Gutenberg | | Currency | 1982 text | Under revision; atlases merging into a version 2.0 | 1960 text, nomenclature dated in places |

My preference, for a reader who collects near home, is the regional atlas as the working tool and the national guide as the shelf reference. Continental coverage forces breadth at the cost of the formation-level detail that resolves a specimen. The Digital Atlas team published the same rationale: collectors build from localities near where they live, usually within one geologic period, so a guide scoped to time and place matches how people collect. The Devonian atlas is supported by National Science Foundation Grant No. 2127427.

One caveat on the print guides. I could not confirm a per-species count for the Audubon volume from any library record, and publisher copy advertises "over 500 identification photographs," which is a marketing number rather than a catalog one. The 846-page figure appears in both the Internet Archive record and Wikidata.

Where can you legally collect, and what needs a permit?

Land status decides this, and it decides before you pick anything up.

On lands administered by the Bureau of Land Management you may collect common invertebrate and plant fossils without a permit, within limits. The regulation at 43 CFR 49.810 defines a reasonable amount as "a maximum of 25 pounds of common non-vertebrate paleontological resources per day per person," allowing a heavier slab when a specimen is embedded in rock and you use non-motorized hand tools. BLM requires the collection be for personal use and states the fossils "may not be bartered or sold." Power tools and heavy equipment are out. Vertebrate fossils sit outside all of it: BLM says it is illegal to dig, remove, collect, or make molds or casts of dinosaurs, mammals, fish, and reptiles without a research permit.

National parks operate under a flat prohibition. The National Park Service states that "collecting fossils for recreational, commercial, or educational use is prohibited in all units of the National Park System," with scientific collection permissible only under a park-approved research permit. The Service asks instead that you photograph the fossil, record its location, leave it in place, and tell a ranger.

Private land needs one thing, and Matthews italicized it in 1960: "The collector should always ask the land owner's permission before entering or collecting on private property." That has not aged.

The order I work in now

This is the sequence that fixed my own hit rate, and it deliberately puts the guide near the end.

1. Confirm the land status and collecting rules before the trip. BLM, park, forest, state, and private land carry different answers; the field office or landowner is the authority.
2. Record where the specimen sat. Coordinates, rock unit, bed, and a photograph in place. People skip this and cannot recover it later.
3. Fix the geologic age of that unit. A state survey map plus a versioned chart such as the ICS v 2026/06 gives you a bracket in millions of years.
4. Measure the specimen in millimeters. Write it on the label before any comparison.
5. Rule out the pseudofossils. Dendrites, slickensides, and concretions, using the bedding-plane relationship rather than the silhouette.
6. Choose a guide matching that period and region. A regional atlas organized by time beats a continental guide here.
7. Match occurrence data, then images. Formation and range narrow the field; photographs confirm what is left.

Steps two and three do most of the work. When I get an identification wrong now, it is because I skipped them.

Questions people ask about fossil field guides

How do I identify a fossil I found?

Start with the rock rather than the picture. Record the location, the rock unit, and its geologic age, then open a guide covering that time and place, since age and formation eliminate most candidates immediately. Measure the specimen in millimeters before comparing it to any photograph, because printed plates are frequently enlarged.

Is there a free field guide to fossils online?

Yes. The Paleontological Research Institution runs the Digital Atlas of Ancient Life, a set of free online fossil field guides organized by geologic time period and region. Its Devonian atlas covers over 200 species from New York, Michigan, and Wisconsin, and the project's Virtual Collection adds interactive 3D models you may download.

Can I keep fossils I find on public land?

On Bureau of Land Management land you may collect common invertebrate and plant fossils for personal use without a permit, up to 25 pounds per day per person under 43 CFR 49.810, using surface collection and non-powered hand tools. They may not be sold or bartered. Vertebrate fossils require a research permit.

Why can't you collect fossils in national parks?

The National Park Service prohibits collecting fossils for recreational, commercial, or educational use in every unit of the park system, allowing scientific collection only under a park-approved research permit. When you find one, the Service asks you to photograph it, record its location, leave it in place, and tell a ranger.

What looks like a fossil fern but isn't one?

Dendrites. Those branching patterns spreading across a rock surface are thin incrustations of manganese dioxide rather than plant remains. Matthews's Texas Fossils handbook classes them with slickensides and concretions as pseudofossils. Slickensides usually sit at an angle to the bedding plane, while genuine plant remains lie parallel to it.

How old are Devonian fossils?

The International Commission on Stratigraphy's chart version 2026/06 places the Devonian Period between 419.62 ± 1.36 and 358.86 ± 0.19 million years ago. A Devonian brachiopod, trilobite, or coral falls inside that window, which is why establishing the age of your rock unit eliminates so many candidate species at once.

What does "×2" mean under a fossil photograph?

It gives the magnification of the printed image. In Matthews's Texas Fossils, plate captions run from ×¼ to ×25, so a specimen labeled ×6 appears at six times its real size. Reading a plate without checking that number is the quickest route to misidentifying a small fossil.