The Blood Was
Made of Ants

A Texas schoolyard in the 1970s. A horned lizard — a Horny Toad, as we called it then — sitting in the hollow of a tree, entirely calm, while a line of children took turns looking at it. It showed no inclination to shoot blood from its eyes. One of us grew up along the I‑35 corridor and never saw another one in the wild again.

Hold onto three things in that sentence, because every one of them turns out to be load-bearing, and none of them is decoration.

The animal was calm, and it was right to be. The child was not the harm. And the memory is not an anecdote — it is the kind of observation that a state agency later built a monitoring programme out of, because in Texas it was the only kind available.

Lee Ann Johnson Linam, writing up ten years of that programme for Texas Parks and Wildlife, put it plainly: “many older Texans can recount personal experiences with horned lizards.” That is a sentence about a dataset.

So this zine goes the long way round. Not to the video that is currently circulating about this animal — we will get to that, and grade it — but to the papers. There is a literature on this lizard, and it is stranger and more generous than anything anybody has made up about it.

A mature Texas horned lizard eats seventy to a hundred ants a day. Roughly 69% of the diet is harvester ants of the genus Pogonomyrmex; at one Arizona site, ants were up to 99% of prey items and harvester ants were 95% of those.

Harvester ants are a bad thing to eat. They husk seeds with mandibles strong enough to bite defensively, and they carry a sting whose venom is among the most toxic insect venoms known to vertebrates. In some species the sting autotomises — it detaches and keeps working.

So the lizard did not pick an easy food. It picked a food that fights back, and then it built a body around the fight.

And it crops that food lightly. When Walter Whitford and Marilyn Bryant watched these lizards sunrise to sunset in 1979, the pattern they found was not raiding: the lizards fed mostly on ants that were away from the nest disc and away from the foraging column, and took only a few at any one place. Not a siege. A slow, distributed, careful tax.

Which matters later, when somebody claims this animal is capable of controlling an ant population. A hundred ants a day is a life, not a lever.

In 2021, four researchers filmed fifty strikes by five lizards at high speed and measured where the tongue landed as a percentage of the ant’s body length. The answer was not the middle.

Divide the ant into head (0–20%), mesosoma (21–50%) and abdomen (51–100%). 76% of hits landed on the mesosoma — the thorax, the one segment carrying neither the mandibles nor the sting. Hits on the head and on the abdomen came in significantly below chance. And this: “The lizard tongue never touched the ants at their rostral or caudal end, even if it was the closest position to strike at.”

Read that twice. Given a nearer target that was armed and a further target that was not, the lizard extended its tongue further. Fifty strikes, zero exceptions.

Then they moved ants mechanically, forwards and backwards, and the aim did not shift — so the lizard is not extrapolating from a direction of travel. It is reading the body.

The whole event takes 60.1 ms of tongue protrusion, 5 ms from first contact to retraction, and 15.74 ms coming back. In 43 of 46 captures the ant arrived head first, so the mandibles are the part that gets coated in mucus first. The lizard does not bite or chew. It embeds.

In 1989, three researchers measured how much Pogonomyrmex maricopa venom it takes to kill. The lethal dose for half a sample — the LD50 — came out at 0.119 micrograms per gram in mice. In Yarrow’s Spiny Lizard, Sceloporus jarrovii: 28. In the Texas horned lizard: 162.

That is roughly 1,360 times a mouse. And notice the middle rung — an ordinary lizard is already 235 times a mouse. The horned lizard is a further 5.8 times beyond that.

Then the experiment that settles the mechanism. They injected mice with two and a half times a mouse’s own lethal dose, plus horned lizard blood plasma. Those mice survived with no symptoms of envenomation. Mice given plasma from other lizards, or from snakes, all died.

So there is something in the plasma. Something transferable, that works in a body it did not evolve in.

And here is the part worth sitting with. The venom is still venom. Nothing about the ant got weaker, safer or more convenient. The ant remained one of the most dangerous things its size on the continent, and the lizard changed itself until it could be in the room with it. That is not a cure. That is an accommodation, and it is carried in the blood.

Horned lizards squirt blood out of the sinuses around their eyes. It is the fact everybody knows about them. What almost nobody knows is where the blood gets its power.

In 2024, Wade Sherbrooke and Bruce Kimball fractionated horned lizard plasma and ran the fractions past coyotes and mice. They isolated the active material to a molecule of between 800 and 1,600 in molecular weight. Then they compared lizards on different diets, and reported this:

“Plasma fractions from individuals of P. cornutum, a blood-squirting species, fed Pogonomyrmex had more aversive levels of the active compound(s) compared to those given a cricket diet.”

And extracts from the ants located the compound in the ant’s abdomen — where the venom is stored. It is present in blood-squirting species that eat harvester ants, and absent in species that do not squirt.

So the defence is not manufactured. It is eaten. The thing that makes a coyote spit out a horned lizard and back away is harvester ant chemistry, taken in through the mouth, carried in the blood, and fired back out through the eye.

Nothing in this animal is its own. Not the atoms — those came out of stars. Not the venom resistance, which is a factor in the plasma. And not the weapon, which belongs to the ant. The lizard is a way of holding borrowed things together.

Which means the loss runs deeper than hunger. Take the harvester ants away and you have not only taken the food. You have taken the blood.

The lizard in the schoolyard did not squirt, and nobody touched it. A line of children took turns looking into a tree hollow, and the animal sat there. For years the calm could read as luck, or as a tame animal, or as one more thing a childhood memory got slightly wrong.

It was none of those. Blood-squirting is aimed at canids. Emily Henry and colleagues, writing up their fieldwork in 2008, describe it exactly that way: a defence “employed primarily against canids.” The 2024 plasma work ran its bioassays on coyotes, and the animals reacted with what the authors call distinct revulsion. Kit foxes, coyotes and a bobcat all responded strongly. Roadrunners and grasshopper mice do not trigger it at all.

And the chemical does nothing to us. The oral receptors that make a coyote recoil appear to be absent in humans.

It is also the last rung of a ladder, and the children never got near it. Henry and colleagues set the order out: these lizards depend primarily on cryptic coloration, and “when threatened, a horned lizard’s first line of defense is to flatten itself against the substrate. If pressed further, it will run a short distance and then ‘freeze’ to reestablish its camouflage or seek cover under nearby vegetation.” Inflating the body, hissing, lunging, jabbing with the occipital horns and squirting blood are all further up. A lizard being looked at and not handled is on the first rung, and the first rung is to hold still.

Which means the stillness those children remember was not tameness, and it was not trust. It was the defence, working.

So nothing about that afternoon was a fluke. The animal stayed at the bottom of its own ladder because nothing made it climb — and the expensive, borrowed, dietary thing at the top was being kept for what it was for.

Which is worth naming, because the story we are usually told about a threatened animal and a child is a story about the child. The child was not the harm. The harm was arriving by road and by cargo ship, and it was going to take the ants.

Scott Henke and W. Scott Fair, writing a management bulletin for the Caesar Kleberg Wildlife Research Institute in 1998, report the specification. Texas horned lizards prefer sandy loam and loamy sand: more than 67% sand, under 15% silt, under 15% clay. They avoid predominantly clay soils, and they avoid soils above 2.5% moisture for bedding and nesting.

Be careful with that number in both directions — the species has been recorded on everything from pure sand to coarse gravel and desert pavement. It is a measured preference, not a wall. But it is a real preference, precisely stated, and it points at something.

Because harvester ants want the same thing. Their queens found new mounds in open ground with little vegetation. In one south Texas study, harvester ant abundance and activity were greater in burned pastures — and the lizards’ home ranges there were smaller, because the food was closer together.

So the whole arrangement rests on ground that is dry, sandy, open, and disturbed often enough to stay that way. Fire did that. Grazing did some of it. Drought did some of it.

And now follow the sand down. Sand is mostly quartz — silicon and oxygen — and both of those were cooked inside massive stars and thrown out when those stars came apart. At Mason Mountain, the wildlife management area where captive-raised hatchlings are now released, roughly two-thirds of the soils weather off granite: the Llano Uplift, a window down onto crystalline basement rock older than 1.3 billion years, some of the oldest exposed stone in North America.

Star, then granite, then sand, then ant, then lizard, then a defence fired out of an eye. Every link in that is a borrowing, and every one of them has to still be there.

Red imported fire ants, Solenopsis invicta, came off a cargo ship at Mobile, Alabama in the 1930s. They were first detected in Texas in 1953. They out-compete native harvester ants for food and for space, and where they get dense they become the dominant ant.

Now the sentence that undoes the story going around. Henke and Fair, 1998: “Texas horned lizards do not actively eat fire ants. This may be due to their inability to effectively neutralize fire ant venom; whereas, horned lizards are resistant to the venom of harvester ants.”

The observation is firm; the explanation is offered as a possibility, and we will hold it that way. And a 2019 study of introduced populations in South Carolina did find Solenopsis in the scat — 1.2%, 8.5% and 1.5% by number across three sites — but identified only to genus, and Solenopsis contains plenty of native species, so that is not a finding about fire ants.

Either way it is not a rescue. The fire ant does not replace the relationship. It deletes it.

Everything on the last four spreads still works. The tongue still finds the mesosoma in sixty milliseconds. The plasma still neutralises a venom among the most toxic known. The eye is still loaded — except that the ammunition was a diet, and the diet is going.

A perfect apparatus, intact, pointing at nothing. That is not the story of an animal that was too fragile to make it. That is the story of a support being withdrawn from an animal that was doing fine.

For decades the decline was anecdote — long-time Texans saying they had not seen one in years. In 2022 four researchers at Texas State turned museum records and iNaturalist observations into the first quantitative test of it, and the eastern edge of the range came in moving toward the range centre at 4.4 kilometres a year between 1960 and 2019.

They are careful about what that does and does not show, and so are we: the more recent iNaturalist data alone, from 2000 to 2019, showed no significant decline in any direction. Citizen-science records can measure where a species is. They cannot measure how many.

Fourteen years earlier, Texas Parks and Wildlife had already drawn the line. The pessimistic results were for the eastern third of the state, “especially in those areas lying north and east of IH‑35 and IH‑37.”

That is a road. Somebody grew up beside it, saw one horned lizard, and never saw another. The boundary of the loss was published, and it runs through a childhood.

The species was listed as threatened in Texas on 18 July 1977 — within a few years of that schoolyard. Henke and Fair spell out what the listing means: it is illegal to pick up, touch or possess one in Texas, and “Handling horned lizards is illegal even if your intentions are good.” Move one off a road and you are in violation.

And that law does not reach what those children did. Nobody picked it up. They looked, took turns, and left it in the hollow of the tree — which was lawful then and is lawful now, and is exactly why a memory like that can end up as evidence. Looking leaves the animal where it is.

The statute draws its line at the next step, and it draws it there because that is where the harm was: collection sits in the decline literature alongside the roads, the cities and the poison. Not because children were the problem. Because there is not enough animal left for taking one home to be free.

There is a population of Texas horned lizards on an urban reserve at Tinker Air Force Base in Oklahoma, and there are no harvester ants there. None. The specialist’s specialism is simply absent, and the lizards are still there.

So five researchers asked what they were eating. They baited for what was available, then examined 124 scat samples. Available: Monomorium 69%, Forelius 11%, Pheidole 10%, Crematogaster 7%, Tapinoma 2%. Eaten: Crematogaster 81%, Pheidole 12%, Formica 6%, Monomorium 1%.

Look at the two ends of that. The ant making up 69% of what was going past got eaten 1% of the time. The ant making up 7% got eaten 81% of the time. The paper’s word for this is that the lizards consumed prey “disproportionately to their availability.”

This is the fact that keeps the zine from being a tragedy, and it is sharper than the usual consolation. Deprived of the specific thing it was built around, the lizard did not become a generalist. It did not learn to make do with the average ant, the abundant ant, the ant that was everywhere. It went and found a different specific thing, and it went narrow again, hard, against availability.

The same shape shows up elsewhere. Introduced populations in South Carolina eat 94% ants and no harvester ants whatsoever, mostly Dorymyrmex, and each of the three populations rounds it out differently. In southern California, coast horned lizards shifted diet when the Argentine ant displaced their harvester ants. The authors of the South Carolina paper say it outright: “Phrynosoma species have greater dietary flexibility than is commonly thought and can persist without harvester ants.”

Specific is not rigid. The support need was real, and the animal was never the caricature the label made of it.

A video essay is going around titled Fire Ants Are Vanishing Across Texas — Scientists Finally Found What’s Hunting Them. Its claim, as described to us: Texas horned lizards are recovering in protected limestone corridors and reducing fire ant numbers through “grazing pressure” that keeps harvester ant colonies compact and defensible.

We could not read it. The transcript was not retrievable and a site republishing the script refused us, so what follows grades the claim as reported to us, not the video’s own words. Say that first, or we are doing the thing we are complaining about.

Ten papers on this animal were read for this zine. Zero contain the phrase, the mechanism, or the recovery. The chain needs three links:

One: horned lizards eat harvester ants. Documented, half a century over. Two: moderate predation makes a harvester colony better at resisting fire ants. No published evidence found. Three: that reduces fire ants across a region. No published evidence found — Texas Parks and Wildlife’s volunteer data suggested a negative harvester/fire-ant relationship in some years, and the same report says “pooled data do not reveal a clear relationship.”

Then the arrow. The literature runs one way: suppress fire ants and small vertebrates recover. Linam’s conclusion is an ask — “recovery efforts will have to address their management in order to be successful.” Reverse it and a request for work becomes a reassurance that no work is needed.

And the geography is inverted. The claim puts the recovery in limestone. The eastern Edwards Plateau — the limestone country — is precisely where the same report says volunteers “confirm some pessimism.” The encouraging reports came from the Post Oak Savannah, the coastal barrier islands and the Dallas–Fort Worth Metroplex, and are described as “small and isolated.” The place releases actually happen, Mason Mountain, is about two-thirds granite.

Here is the sting. Fire ants genuinely are being eliminated across parts of Texas. Where tawny crazy ants, Nylanderia fulva, get dense, fire ants are wiped out — work led by Edward LeBrun at the University of Texas. That is one invasive replacing another, and it is worse news, not better. The true half of the headline is the bad half. The invented half is the comforting one.

In 2014, seven researchers modelled two Texas horned lizard populations a thousand kilometres apart — an urban reserve in Oklahoma and a six-thousand-hectare wildland site in south Texas — over eight and nine years. The populations differed in almost everything. They converged on one answer.

Hatchling survival had the strongest influence on population growth at both sites. Adult survival had the smallest effect. Whatever else is true, the leverage is at the beginning of a life.

And then the honest sting, which the authors put in their own abstract: that conclusion holds “within the constraint of the need to more accurately estimate hatchling survival.” The number that matters most is the number nobody has well. Henke and Fair, sixteen years earlier: “Very little is known about the diet of hatchling and juvenile horned lizards.” The most important stage is the least known one.

Which is what the real comeback is aimed at, and it is made of people. In autumn 2021 the Texas Horned Lizard Coalition released 263 captive-raised hatchlings at Mason Mountain and Muse Wildlife Management Areas and on private land; 159 of them hatched at the Fort Worth Zoo, which also hatched its thousandth horned lizard that season. That August, biologists and graduate students found 25 lizards believed to be the offspring of hatchlings released in 2019 — to their knowledge the first time captive-reared horned lizards had survived long enough to breed in the wild.

Ask the geneticist working on it what the obstacle is and he does not say fire ants. Dean Williams at TCU, in August 2026: “snake predation is a big problem… The big effort now is to try to reduce predation to the point that the population can become established.”

No corridor did this. Zoos did, and a state agency did, and volunteers did, and it is slow, and it is working a little. That is what a real recovery looks like, and it is smaller and better than the story.