An axolotl is a salamander that never grows up: it keeps its larval body and gills for life and breeds in that state. It can rebuild an amputated limb, complete with bone and nerve, and it is critically endangered in the wild while being one of the most common laboratory vertebrates on earth.
Two things about the axolotl are famous and both are usually explained backwards. The first is that it never turns into an adult salamander: it keeps the feathery external gills and the tail fin of a larva and becomes sexually mature anyway, a condition called neoteny. This is not a failure to develop. The metamorphosis machinery is intact and can be triggered in the laboratory with thyroid hormone, producing a land salamander that then lives a shorter life — so what evolution removed was the signal, not the ability, in a lake where staying in the water was the better bet. The second is regeneration, and the popular version treats it as spare parts. It is closer to an instruction problem. An amputated limb forms a blastema from local tissue that has reverted to a less specialised state, and what gets rebuilt is dictated by positional information still held in the stump; cut the nerve supply and no blastema forms at all, because the material is still there and the instruction is not. Underneath both is a genome about ten times the size of ours, which turns out to be mostly repetitive sequence rather than extra genes — so whatever explains the axolotl, it is not simply more instructions. And then the fact that makes the animal genuinely strange to write about: there are tens of thousands in tanks and, on the most recent surveys, a wild population in a few canals at Xochimilco that field teams struggle to find at all.
Developed record · 58% complete · reviewed 2026-08-11
What this page covers
One species, Ambystoma mexicanum, native to a single lake system in Mexico City. The laboratory animal and the wild animal are related populations that have been separate for over a century.
Often confused with: Ambystoma tigrinum, the tiger salamander — a close relative that metamorphoses normally; Andrias and Cryptobranchus, the giant salamanders, which are not axolotls; Fish, which it is not
Quick facts
What it is
A salamander that keeps its larval body and breeds anyway
Regeneration
Limbs, tail, jaw, spinal cord and parts of the heart and brain
Genome
About 32 gigabases — roughly ten times human, mostly repetitive
The metamorphosis is not missing. The signal to start it is.
Most salamanders hatch as aquatic larvae with external gills, then metamorphose — losing the gills, growing eyelids and lungs, and moving onto land. The axolotl stops halfway and stays there. It keeps the gills, keeps the tail fin, never develops functional eyelids, and becomes sexually mature in that body. The technical name is neoteny: the adult retains juvenile form.
What makes this interesting rather than merely odd is that the machinery still works. Metamorphosis in amphibians is driven by thyroid hormone, and an axolotl given thyroid hormone will metamorphose — losing its gills, developing eyelids, and walking out onto land as something that looks like a tiger salamander. So the species has not lost the ability to become an adult. It has lost the hormonal trigger that would start it, and the result is an animal that spends its whole life in the stage other salamanders pass through.
That is a reasonable bet in the environment it evolved in. A deep, permanent, food-rich lake with few aquatic predators rewards staying in the water; the land around it does not. Where a related species faces a pond that dries, metamorphosis is worth the cost. Where it does not, the animal that skips it breeds sooner and keeps the gills that let it stay.
Induced metamorphosis is not a party trick with no consequences. Metamorphosed axolotls are generally shorter-lived and, importantly for the laboratory, they regenerate considerably less well. Whatever keeps the larval body also keeps the repair.
Words used here
Neoteny
Retaining juvenile features into a sexually mature adult. The axolotl is the standard example.
Metamorphosis
The reorganisation from larva to adult. In amphibians it is driven by thyroid hormone.
A regrowing limb is told what to build by the stump it grows from
Established
Specialists would state this without hedging. Multiple independent lines of evidence agree.
Amputation in the axolotl produces a wound epidermis and then a blastema formed from local cells that partially dedifferentiate rather than becoming pluripotent, largely retaining memory of their tissue of origin. What is rebuilt is determined by positional information held in the connective tissue of the stump, so a blastema formed at the wrist produces a hand and one formed at the shoulder produces the whole limb. Blastema formation requires nerve input: a denervated limb does not form one.
Who this applies to
axolotls, with the same broad mechanism in other salamanders
Studied in
Ambystoma mexicanum, Caudata
You may have heard
“Axolotls can regrow any body part”
It regrows a great deal, and not because it keeps spare parts. The tissue is ordinary; what is unusual is that the stump can still say where it is and what should be there, and the nerve carries the permission to start. Cut the nerve and the same tissue does nothing at all. Regeneration is a set of instructions, not a reserve — which is also why it can only rebuild the pattern the animal already had.
Why we rate it this way, and what the caveats are
EstablishedHigh confidence
The nerve dependence is a direct manipulation with an unambiguous result, and lineage tracing has established the origin and fate of blastema cells.
How far it can be extended
Blastema-based limb regeneration and its nerve dependence are documented across salamanders; the molecular detail is characterised mainly in the axolotl.
Caveats
Positional information is demonstrated behaviourally and only partly identified molecularly.
Regeneration fidelity declines slowly over repeated amputations.
Laboratory animals of one species at controlled temperature; wild axolotls are almost unstudied in this respect.
Still unanswered
How is positional identity stored in adult connective tissue for the life of the animal?
Blastema formation, partial dedifferentiation, positional information and the nerve requirement.
How we know
Amputating a limb with the nerve cut, and watching nothing happen
An axolotl rebuilds an amputated limb. Is that because the tissue is capable of it, or because something tells the tissue to do it?
The two possibilities predict the same thing in an ordinary amputation, so the design removes one of them. The nerve supply to a limb is severed and the limb is then amputated, leaving every cell type that would normally contribute to a blastema present, healthy and in place. If regeneration is a property of the tissue, it should proceed. If it requires a signal the nerve carries, nothing should form.
What happened
Denervated limbs healed the wound and formed no blastema. Tissue that regenerates reliably when innervated did nothing when it was not, and the capacity returned when the nerve did.
What it shows
That regeneration is a signalling problem rather than a materials problem. Everything needed to rebuild the limb remains present in a denervated stump; what is missing is the instruction to begin. It reframes the axolotl from an animal with unusual tissue to an animal with an unusual instruction system, which is why the search moved to what nerves supply rather than to what salamander cells are made of.
What it does not show
It identifies a requirement, not a mechanism: knowing the nerve is necessary does not say what it delivers, and the factors involved were established much later and are still incomplete. It says nothing about why mammals fail, since mammalian limbs are innervated and still do not regenerate — nerve supply is necessary and clearly not sufficient. And it is a salamander result, from laboratory animals of one species.
The controls — what makes this evidence rather than a story
Amputation performed identically on innervated limbs in the same animals, so the only difference is the nerve.
Wound epidermis formation assessed separately from blastema formation, because the first happens without nerves and the second does not — collapsing them would hide the result.
Histology rather than external appearance, since a stump that has healed over and a stump that has begun regenerating look similar from outside.
Reinnervation used as a recovery test: restoring nerve supply restores blastema formation, which rules out the surgery itself as the cause.
An amputated axolotl limb does not scar. Within hours a thin layer of epidermis covers the wound without forming the collagen barrier a mammal would lay down, and cells beneath it — muscle, cartilage, connective tissue — stop being what they were and start dividing. The mass they form is the blastema, and it is the whole of the trick.
Two findings from that blastema matter more than any other. First, the cells do not become fully unspecialised. They step back rather than back to the beginning, and largely remember their tissue of origin — cartilage rebuilds cartilage. So this is not an animal keeping a reserve of embryonic stem cells. Second, what gets built is determined by positional information retained in the connective tissue of the stump: a blastema on a wrist builds a hand, and a blastema on a shoulder builds the whole arm. The stump knows where it is.
The clearest evidence that this is an instruction system rather than a material one is what happens when the nerves are cut. A denervated limb has all the same tissue available and forms no blastema at all. Regeneration is not the tissue being capable; it is the tissue being told.
Limbs, tail and jaws regenerate repeatedly, with fidelity that declines only slowly over many amputations.
Spinal cord regenerates functionally, which no mammal manages.
Substantial portions of heart and brain tissue are replaced after damage.
Scarring is largely absent, and that absence appears to be a precondition rather than a side effect.
What it cannot do is build anything the animal did not have: regeneration restores the pattern, it does not invent one.
Ten times our genome, and an ordinary number of genes
Established
Specialists would state this without hedging. Multiple independent lines of evidence agree.
The axolotl genome is approximately 32 gigabases, around ten times the human genome and among the largest animal genomes assembled. The excess is overwhelmingly repetitive sequence and expanded introns rather than additional protein-coding genes, whose number is unremarkable for a vertebrate. Salamanders as a group have the largest genomes of any tetrapods, and the expansion predates the axolotl.
Who this applies to
axolotls, within a salamander-wide pattern of very large genomes
Studied in
Ambystoma mexicanum, Caudata
You may have heard
“Axolotls have ten times more DNA than us, which is why they can regenerate”
The first half is right and the second does not follow. The extra DNA is repeated sequence and oversized introns, not extra instructions — the gene count is ordinary. That is a useful result rather than a disappointing one: it rules out the easiest explanation, and leaves regeneration to be explained by how familiar genes are used rather than by a hoard of unfamiliar ones.
Why we rate it this way, and what the caveats are
EstablishedHigh confidence
Direct sequence assembly, subsequently revised and improved without altering the conclusion about composition.
How far it can be extended
Large genomes dominated by repetitive sequence are general to salamanders; the axolotl assembly is the best characterised.
Caveats
Assembly of highly repetitive regions remains difficult and later versions revised the details.
Whether genome size relates to regeneration at all is unknown; the association is suggestive and untested.
One laboratory strain from a very small founding population.
Still unanswered
Is the salamander genome expansion adaptive, or drift that selection tolerates?
Genome size, its repetitive composition, and ordinary gene count.
The axolotl genome is about 32 billion base pairs against the human three billion, and when it was assembled in 2018 it was the largest animal genome ever sequenced. The obvious inference — ten times the instructions, hence the extraordinary abilities — is wrong. Gene count is unremarkable. The size comes from repetitive sequence and enormous introns, the non-coding stretches within genes, which in salamanders have expanded to a degree seen nowhere else in tetrapods.
This is a useful negative result. Whatever underlies regeneration, it is not a large set of extra genes that mammals lack, and explanations of that shape can be set aside. Attention moved instead to how existing genes are regulated, and to a small number of genes expressed specifically in the blastema.
There is a cost to carrying a genome that size. Every cell must copy all of it, so salamander cells are large and divide slowly, and the whole animal runs at a leisurely pace. It is a reasonable guess that the slowness and the regeneration are related, though that is a guess rather than a finding.
Words used here
Intron
A stretch within a gene that is removed before the gene is translated. In salamanders these are enormous.
Repetitive sequence
DNA consisting of copies of the same short sequence. Most of the axolotl genome, and most of what makes it large.
The abundance and the endangerment are both real, and the first does not fix the second.
Axolotls are everywhere and nowhere. They are among the most widely kept laboratory vertebrates in the world, sold in pet shops, and bred by the thousand. The wild population lives in the remains of one canal system at Xochimilco in southern Mexico City — what is left of the lake complex the Aztecs built their city among — and repeated surveys have found so few that estimates are given in animals per square kilometre rather than as totals. The IUCN lists the species as Critically Endangered.
The captive population is not a solution to that, for reasons that are worth stating precisely rather than assuming. The laboratory line descends from a small number of animals taken to Paris in 1864 and has been bred in captivity ever since; it carries introgressed genes from the tiger salamander, deliberately introduced in the twentieth century, and it has been selected for over a century and a half of life in a tank. Releasing those animals into Xochimilco would introduce a domesticated, partly hybridised population into the last habitat of the wild one.
Water quality: the canals receive treated and untreated wastewater from a city of over twenty million people.
Introduced tilapia and carp, which eat eggs and larvae and compete for food.
Habitat loss as the canal system is drained and built over.
Chytrid fungus, which has devastated amphibians globally and is present in the region.
The conservation work that appears to be helping is not captive breeding but the restoration of chinampas — the traditional raised agricultural beds — with filtration and refuges that keep introduced fish out. That reconstructs the habitat rather than restocking it, which given the state of the captive line is the more useful direction.
Words used here
Chinampa
A raised planting bed built in shallow lake water, the traditional agriculture of the Valley of Mexico. Restored chinampa channels are the main axolotl refuge work.
Introgression
Genes from one species entering another through hybridisation. The laboratory axolotl carries tiger salamander DNA.
The permanent smile is anatomy rather than mood — the mouth is shaped that way and cannot change — and the same caution applies as with dolphins: an axolotl in poor condition looks exactly like one in good condition. The signs that actually matter are the gills, which shrink and lose their filaments in poor water, and forward-curled gill stalks, which indicate stress.
Cold water: they are lake animals from a high-altitude basin and suffer above roughly 20°C.
No gravel small enough to swallow; impaction is a common cause of death in captivity.
Solitary by preference. Axolotls bite each other’s gills and limbs, which regrow, which is not a reason to let it happen.
They are illegal to keep in some jurisdictions, including several US states, generally because of the risk to native salamanders if released.
Legal
Before keeping one
Check local law first: axolotls are prohibited in a number of jurisdictions, and where they are permitted, release into the wild is almost always an offence and always ecologically harmful. Never release a captive axolotl, including in Mexico — the laboratory and pet lines are genetically distinct from the wild population and carry disease risk.
Where this applies: Legality varies by country and by state or province; several prohibit private keeping outright.
When to get help: Consult local wildlife authorities on legality, and an amphibian-experienced veterinarian on care.
The scientific interest is not cosmetic either. Axolotl regeneration is studied because it is the clearest vertebrate example of something mammals cannot do, and the hope is not to make people regrow limbs but to understand why our wounds scar instead. Scarring is what stops regeneration, and the axolotl is the animal that shows it is not inevitable.
A shipment to the Muséum national d’Histoire naturelle established the population that most laboratory axolotls descend from. The animals bred readily in captivity, which is the reason this species rather than another salamander became the model.
1865
First observation
Captive axolotls metamorphose, and the puzzle appears
Some of the Paris animals unexpectedly transformed into land salamanders, showing that the species retained a metamorphosis it normally never performs — and turning neoteny from a description into a question about what was switching it off.
1920
Landmark experiment
Thyroid hormone identified as the trigger
Feeding thyroid tissue induced metamorphosis in axolotls, establishing that the species retains the full response and lacks the hormonal signal. Neoteny became a matter of endocrine regulation rather than lost capacity.
1952
Landmark experiment
Regeneration shown to depend on nerves
Denervation experiments established that an amputated limb with an intact nerve supply forms a blastema and one without does not, moving regeneration from a question about tissue to a question about signalling.
2015
Reinterpretation
The blastema reframed as positional memory rather than stem cells
Lineage tracing showed blastema cells largely retain memory of their tissue of origin instead of becoming pluripotent, and that positional information in connective tissue determines what is rebuilt.
Changes how the 1952 result reads
The nerve requirement had shown that regeneration needs an instruction. This identified what the instruction is about — not "grow", but "grow the part that belongs here".
The genome sequenced, and the simplest explanation ruled out
A 32-gigabase assembly, ten times the human genome, showed the size to come from repetitive sequence rather than from extra genes. Whatever underlies regeneration is not a large novel gene set.
Why it matters: Mammals retain much of the same machinery and deploy it differently. Whether scarring is a constraint or a trade-off — faster closure against the risk of infection — is the question the whole field is arranged around.
What would settle it: Identifying what in the mammalian wound response blocks blastema formation, and whether suppressing it produces regeneration rather than a tumour.
How is positional information stored in adult tissue?
Why it matters: A stump behaves differently depending on where along the limb it is, decades after development finished. Something durable encodes that, and what it is molecularly is not established.
What is the enormous genome for, if anything?
Why it matters: Salamander genomes are the largest of any tetrapod and the expansion may be a consequence of weak selection against repetitive DNA rather than an adaptation. Whether it is connected to regeneration at all is unknown.
Is there a viable wild population left?
Why it matters: Recent surveys find very few animals, and the difference between "extremely rare" and "functionally gone" determines whether habitat restoration has something to restore.
What would settle it: Systematic environmental DNA surveying across the canal system, which is now underway.