Golden poison frog: where the poison really comes from

An adult golden poison frog averages 47 millimeters (about 1.9 inches), roughly half the width of a credit card, and it comes in mint green, yellow or orange, colors more at home on a highlighter than on a rainforest floor. The Animal Diversity Web, a University of Michigan database, says those bright colors warn predators that the skin is toxic. Adults are largely protected by it.

The odd part is where the poison comes from. The frog does not seem to cook it up itself. The main idea in the research is that poison frogs collect their alkaloids (bitter chemicals, many of them toxic) from the small animals they eat, and that frogs kept away from their wild menu tend to lose their toxicity. For this species the leading guess for the original source is a group of small beetles, but nobody has confirmed that for frogs in Colombia.

This article follows the toxin from forest floor to skin, explains why the frog is so colorful, and covers the lab work on how it survives its own chemistry. It is not a guide to handling poisonous animals. Watch this species from a distance, or in photographs.

Meet the golden poison frog

According to Mongabay, the golden poison frog lives only in southwest Colombia, where mountains and rainforest meet the mangroves of the Pacific coast. The Animal Diversity Web (ADW) puts the core of its range in the upper Río Saija drainage and describes very humid lowland forest at 100 to 200 meters above sea level, with up to 5 meters (about 16 feet) of rain a year. That page also calls the forest Amazonian, a word that does not match the Pacific-coast accounts, so we follow the Pacific descriptions.

Where the poison comes from: a diet, not a recipe

Think of the frog less as a chemist and more as a collector. According to the ADW, it mainly eats ants of the genera Brachymyrmex and Paratrechina, plus termites and beetles, and captive frogs tend to lose their toxicity without their wild prey. None of our sources says that ants carry batrachotoxin itself. The ADW points to a beetle, and the ants in the 2021 study below carried other alkaloids.

So which prey carries the batrachotoxins? In 2004, Dumbacher and colleagues reported these toxins at high levels in Choresine beetles (family Melyridae) collected in Papua New Guinea. Insects of that size, and in one case a named Choresine species, also turned up in the stomachs of toxic birds called pitohuis. The authors proposed that relatives of these beetles in Colombian rainforest could be the source for Phyllobates frogs. They did not test that, and they say fieldwork restrictions in Colombia got in the way. The ADW says feeding on a Choresine beetle may be the main source of toxicity, which is still a hypothesis, not a result.

The paper is also candid about messy results. Toxin amounts varied widely. One batch of three beetles averaged about 1.8 micrograms per beetle, while another mixed batch had none, and one single beetle showed none. The authors cite a figure of about 1 milligram of batrachotoxin in one frog skin. Dividing one number by the other gives somewhere past 500 beetles, which is only a rough scale check of ours and not a claim from the paper.

There is one more wrinkle. A frog does not store everything it eats. In a 2021 study, captive-bred green and black poison frogs (a different species) ate fire ants for four months, and the ants’ piperidine alkaloids never showed up in the frogs’ skin. Sequestering, the technical word for storing a toxin taken from food, is selective.

The toxin, and the frog’s own shield

Batrachotoxins keep voltage-gated sodium channels open, according to the Animal Diversity Web. Sodium channels are tiny gates in nerve and muscle cells that open and close to pass signals. Prop them open and the signals stop making sense, which the same page says can lead to heart failure. Mongabay adds that if the toxin enters the bloodstream it can stop the heart within minutes.

Estimates of the dose do not agree well. The ADW gives about 1,900 micrograms per frog and a lethal range for humans of 2 to 200 micrograms, a hundredfold spread. Mongabay says one frog carries enough to kill 10 people. Treat all of these as rough.

How does the frog avoid its own gates being jammed open? In 2017, Sho-Ya Wang and Ging Kuo Wang at the State University of New York reported in PNAS that they put five changes from the frog’s muscle sodium channel into rat channels. Only one, a swap from asparagine to threonine called N1584T, gave strong resistance. The limits matter: the tests used rat channels in cultured cells, not frogs, and a small toxin effect remained at high doses. The authors also say the finding is unlikely to produce an antidote.

Why wear gold on a poison budget?

Bright color that advertises danger is called aposematism. Per the ADW, the frog’s colors warn predators, and adults wear a single bright color while juveniles carry stripes along their sides. The same page lists only one known predator, a small snake named Leimadophis epinephelus, which eats juveniles.

We opened no experiment that tested predators’ reactions to this species, so the warning role rests on the database’s description, not on a field test we can point to. That is worth saying out loud, because “bright means poisonous” is easy to believe and not always true.

What this looks like in practice

Here is the chain of events as the sources describe it, with how firm each link is.

For example, a frog eats small prey on the forest floor, some of the prey’s alkaloids end up in its skin, and a predator that tastes the skin learns a lesson. Step 6 is the one link nobody has confirmed for Colombian frogs.

Common mistakes and myths

Myth: the frog makes its own toxin. Some news coverage describes the toxin as produced in the skin glands. The skin stores it, but the evidence above says the raw material arrives with the food.

Mix-up: golden poison frog and Panamanian golden frog. They share a color word and a taste for warning colors, but they are separate species with separate stories. The Panamanian golden frog lives in Panama, sits inside the toad family and carries different toxins, according to its own article here.

Mix-up: poisonous and venomous. A poisonous animal carries a toxin that harms when touched or eaten. A venomous animal injects it. This frog’s toxin sits in its skin. Another toxic amphibian, the rough-skinned newt, makes a good comparison. This is a lesson in vocabulary, not an invitation to test anything.

Mix-up: every poison frog is equally dangerous. The Animal Diversity Web describes this species’ toxins as far stronger than those of other poison dart frogs, and the fire ant study shows that even among poison frogs, uptake depends on the chemical.

Mix-up: all frog parents are the same. Here the male carries his young on his back to a pool. The ADW says eggs are carried, though a later passage on the same page mentions tadpoles, so we say “young.” The Darwin’s frog has a very different parenting trick.

Conservation status

The golden poison frog is listed as Endangered on the IUCN Red List of Threatened Species. That is the category as distributed by GBIF, the Global Biodiversity Information Facility, and we checked it on November 9, 2026. Endangered means the species was evaluated and is considered to face a very high risk of extinction in the wild.

The pressures are mostly about space. Mongabay reports that agriculture, mining and illegal coca farming are shrinking the frog’s habitat, and that illegal mining upstream is polluting rivers. The ADW names deforestation as the main threat and says populations are declining, while Mongabay says there is no hard population data since the reserve was founded. Community members told Mongabay the frog has rebounded, but they want to know for certain. Some things are going right: in September 2019 the Eperãra Siapidaarã people announced the 11,641-hectare (28,765-acre) K´õk´õi Eujã Traditional Natural Reserve, which Mongabay describes as protecting the frog’s habitat, after meetings funded by the Rainforest Trust. We have no confirmed trend beyond those statements.

Categories are reviewed and can change, and the copy GBIF distributes can trail the IUCN website by some months, which is why we print the date we checked. GBIF’s copy does not give the year of the assessment, so we do not state one.