Archerfish (Toxotes spp.) water jetting

You might have heard that chimpanzees were the first animal found making and using tools in the wild. That’s simply not true. What is true is that opening the door to chimpanzees—in the 1960s—was the least threatening way to bring animals inside our exclusive tool users’ clubhouse.

Today, we look at one of the many animals that got in ahead of the apes. We follow two groups of scientists, working 250 years apart, as they uncover one of the most dramatic and skilful examples of tool manufacture that we know of. It’s a fishy story, in a good way.

The Jaculator

The Systema Naturae of Carl Linnaeus is central to the history of biology. In several editions of his book starting in 1735, Linnaeus spelled out the genus and species naming of all the known plants and animals, and even minerals. This is where we first learned our own species name, Homo sapiens, although until the tenth edition (1758) Linnaeus used a simpler designation of ‘Homo’ along with the terse Latin inscription ‘Nosce te ipsum’: know yourself.

In a short note for the 1767 edition (the 13th: he really churned them out), we find a description of a curious fish:

Habitat in India; victitat Insectis supra aquam volitantibus, in quae explodit rostro tubuloso exquisitissime guttulam, ut ea cadant in aquam et devorentur.

Loosely translated, this tells of a fish that ‘lives in India; it defeats insects flying above the water, into which it explodes a drop with its tubular beak in a most exquisite manner, so that they fall into the water and are devoured’. Linnaeus named this animal as Chaetodon rostratus, although today we know its genus as Toxotes: the archerfish.

Linnaeus lived in northern Europe, and likely never saw an archerfish in its natural, tropical habitat. He also mistakenly attributed the report to butterflyfish, which still populate the genus Chaetodon and have a more ‘tubular beak’ than the pointy-faced archer. So how did he come to know of this behaviour at all? For that, we need to turn to his contemporary, a Dutch physician named Johannes Albertus Schlosser.

Schlosser’s nationality gave him access to a trading network halfway around the globe, in what the colonisers called ‘Dutch East India’: modern Indonesia. As with most naturalists of the day, he was enmeshed in a network of letter writing that spread tales of new and unusual things back to Europe, for repetition and dissection by the learned societies of the day. Schlosser was a Fellow of the Royal Society in London, and in two short reports to the Philosophical Transactions of the Royal Society—in 1764 and 1766—he introduced the English-speakers of the world to ‘the Jaculator or shooting fish’. The drawing accompanying his 1766 report is quite accurate:

Even Schlosser was an intermediary in this game of scientific telephone: the actual fish-watcher was one Wilhelm August Hommel, governor of the hospital in Batavia, now known as Jakarta. Hommel had heard tales from local people of this spitting fish, and wasn’t sure whether to believe them. After all, it was an age where centaurs and unicorns and men with their faces in their chests hadn’t entirely been disproven. So he did exactly what researchers do now: he built a makeshift seawater aquarium out of a large cask, and had several archerfish placed inside. Once they were acclimatised (or rather ‘seemed reconciled to their confinement’), Hommel used a fly on a stick to test their abilities, as relayed in Schlosser’s 1764 letter:

It was with inexpressible delight, that he daily saw these fish exercising their skill in shooting at the fly, with an amazing velocity, and never missed the mark.

In the 1766 follow-up report, Hommel goes into more detail, giving us the first clear scientific report of archerfish natural behaviour:

When the jaculator fish intends to catch a fly or any other insect, which is seen at a distance, it approaches very slowly and cautiously, and comes as much as possible perpen­dicularly under the object: then the body being put in an oblique situation…and the mouth and eyes being near the surface of the water, the Jaculator stays a moment quite im­moveable, having its eyes diredly fixed on the insect, and then begins to shoot, without ever shewing its mouth above the surface of the water, out of which the single drop, shot at the object, seems to rise…

As this modern video shows, its a good description:

Note that Hommel stresses that he never saw the fish poke its head out of the water, even as it shot repeatedly at its prey. These fish can jump to catch food, but not as part of their shooting routine. If you ever see an image of an archerfish performing a jump-shot—including the painting at the top of this post, from the 1893 edition of Brehm’s Tierleben—you know it’s almost as fanciful as a picture of a centaur.

Of possible relevance to our tale, Schlosser and Linnaeus were in direct conversation as part of that 18th century letter-circle. If you’re trying to flatter someone, you could do worse than follow Schlosser’s example in addressing his famous elder:

Translated, he writes to: ‘Carolus Linnaeus, Golden Knight etc. etc. etc. etc. A truly noble and wise man’. Perhaps a little over the top, but it was also a time of fancy powered wigs and flowing robes, so a quick ‘Hey Carl, check out this weird fish lol’ likely wouldn’t do the job.

And so we close the circle, with Linnaeus either reading the Philosophical Transactions of the Royal Society, or hearing about it through his morning mail, and adding a latinised footnote to his latest edition.

Back to the future

Fast forward two-and-a-half centuries. The French and American Revolutions pass by, Darwin is born and dies, modern zoology and ethology (animal behaviour studies) arise, electricity emerges, cameras are invented and get really good at taking fast photos. Aquarium science moves on from large casks. I actually guest edit an issue of the Philosophical Transactions of the Royal Society. And all that time, archerfish haunt the mangroves and freshwater of south and southeast Asia, and northern Australia.

What didn’t really move on was our understanding of exactly how archerfish do their famous trick. Luckily, in the past few decades that has changed, dramatically.

Let’s start in Germany, at the Universitat Freiburg, where Stefan Schuster and his colleagues were building increasingly clever systems for testing just how these fish jaculate. It’s now 2004, and Schuster’s team have already shown that archerfish are able to calculate where a struck insect will fall, within just a few dozen milliseconds of a strike. It needs that speed to outcompete all the hungry fellow fish around them.

That year, they published another paper showing that archerfish can accurately judge just how big their prey is. That may seem like a simple task, but remember that the fish keeps its head underwater, so it’s dealing with the refraction of images as they bend through the water surface. Knowing how big their prey is tells the fish whether it’s a big thing nearby or a smaller thing further away. And that matters, because the archerfish isn’t some garden sprinkler, spraying water in the random hope of wetting a bug. A second study by the same team in 2006 showed that the fish shoot with more force (using bigger droplets) at bigger prey, effectively hitting insects with about ten times as much force as the bugs need to hold on to the leaf or other object they’re clutching.

The fish are also remarkably accurate, as Hommel/Schlosser spotted in 1764. It is not entirely true that they ‘never miss the mark’, but they do have a trick or two for making sure that they deliver the most punch for their puff. One key is their ability to use their inbuilt distance gauge to bundle the jetted stream together at the point of impact. This was proved by Peggy Gerullis and Schuster (now at University of Bayreuth) in a 2014 project, in which cameras recorded archerfish shooting at either 250 or 5000 frames per second. This figure from their paper sums it up nicely:

On the left we see a stream of ejected water strung out during its journey, coalescing into a power-drop just as it reaches the target. The same images on the right show that the fish is judging distance to manage this trick, whether it’s trying to hit something 20, 40 or 60cm away. And archerfish do this with a minimum of effort, keeping their body remarkably still and their upper jaw fixed at the water’s surface. All they need to do is change the timing of their mouth opening, using longer times for further targets.

Actually, we need a short aside here to look more closely at what’s happening inside the archerfish mouth. Apologies if you’re not ready for this:

That’s a front-on view of a sevenspot archerfish, Toxotes chatareus. If you’re searching for an animal exquisitely adapted for tool making, this is what one looks like. The jet is formed by pushing water along that groove in the upper mouth, as explained in technical detail by Gerullis and Schuster:

The true mastery of the fish therefore seems to be covarying the stability and the initial speed profile of their jets. To shoot, archerfish rapidly compress their gill covers, forcing water out from a ‘‘gun barrel’’ whose upper half is a ridge in the roof of their mouth and whose bottom half is formed by the hardened surface of their tongue.

This more recent work has helped refine those initial 18th century reports. Hommel and his barrel of fish suggested that archerfish preferred to shoot straight upwards, but that’s not exactly true. Schuster’s research showed a spitting range of 60-88 degrees relative to the water’s surface, still steep, but not directly up. Of course, this angle further complicates distance measurement, because gravity will make the water drops bend downwards slightly as they travel. The problem is exacerbated when prey is moving, but again Schuster and his team found extraordinary abilities: not only can archerfish learn to hit moving prey by matching prey speed or using a ‘leading’ shot, but other archerfish can learn the same trick without practice simply by watching another trained fish. Any human who has tried archery will know that watching others is definitely not enough to make us masters, so the fish have one up on us there.

By this point it should be clear that treating water jets as a reflex, rote action is not giving the fish nearly enough credit. Stripping leaves from a twig tool using ape hands is clever, but forming and throwing tools with just your mouth, all while allowing for refractive effects (and even ripples on the water surface distorting your view) is a next level skill. Not to mention then tracking the falling prey within milliseconds and getting to it first.

No-one is suggesting that the fish are deliberately doing calculus in their heads, but their ability to adjust for changing conditions shows that they have an inborn talent for complex trajectories. For example, a recent study by Svetlana Volotsky and researchers at Ben-Gurion University of the Negev placed wild-caught fish into tanks and set up a target (Governor Hommel would approve). They then turned on a small wind generator that blew each shot slightly off target. In response, the fish adjusted their next few shots until they were back hitting the right spot, proving that there is an ongoing learning component to their archery. There is a lot happening in that little pointy body.

The original tools

Tool use in our own lineage (the hominins) goes back at least 3.3 million years to large stone pounders in East Africa. How far back can these fish trace their own technology?

They don’t leave stone tools—as far as we know—but their unique adaptations to forming and shooting water drops have shaped their skeleton. By looking at closely related fish, we can use those bony kinks to estimate origins and track the currents of their development.

Our guide here is a study by Matthew Girard and colleagues from 2022. The team looked at the two best-known shooters, Toxotes chatareus and T. jaculatrix (recognise that species name from the 1760s?). They used hard and soft tissues, and genetics, to compare those taxa to others in the Toxotes genius, as well as to the related Protoxotes and the sister group Leptobramidae. Not only did they manage to track the emergence of the shooting apparatus, they even cleaned up a bit of taxonomic muddle along the way.

Here’s a map of their samples, which also gives a clue to where these fishy gunslingers are plying their trade:

Along with detailed anatomical discussions, the researchers tried to distinguish between two hypotheses over how the fish eject their weapon. One theory, the ‘blowpipe’, is that they rapidly change the shape and size of their mouth to shoot. The other, the ‘pressure tank’, suggests that the water builds up pressure in the fish’s mouth until it is suddenly released by a valve. The results come down plausibly on the blowpipe side, but even at these fine levels of detail the possibility of some kind of valve remains slightly open (ironically).

The analysis was more conclusive on the topic of origins. Surprisingly, the non-archerfish group Leptobramidae (or beachsalmon) also have an upper mouth groove and keeled tongue or basihyal. The shape of the groove is more curved than in archerfish, but presumably if they had the coordination and motivation, they could learn to be effective sharpshooters. But instead they feed on animals already in the water, including other fish.

One interpretation of these data is that while archerfish and beachsalmon both feed mainly on prey in the water, the archers have added an aerial attack to their arsenal. In fact, the original function of that power jet could have been to forage not upwards but down into the muddy or sandy substrate. Similar to dolphins using sponge tools to disrupt the seafloor and reveal hidden prey, archers are able to create short underwater jets that break up sediment. Work from Gerullis and Schuster’s team in 2017, led by Jana Dewenter, showed that they get much closer to the sediment than they do to their above-water prey, no more than a few centimetres (versus up to 2 metres for aerial shots). But again they are careful, changing the timing of their pulses depending on whether they are disturbing mud, fine sand or larger sand particles. This is also an example of tool manufacture and use, of course, just not as splashy as the jets we’ve been looking at.

So when did this tool use start? Toxotes split from Leptobramidae around 40 million years ago, in the Eocene epoch. If both sets of fish have been tool users since at least that time, albeit with later adoption of aerial weapons by archerfish, it puts our 3.3 million years of stone tools to shame. It would even put suspected plant tool use by the common human-chimpanzee ancestor some 6-8 million years ago firmly in the ‘newcomers’ category.

This bottom-to-top theory is speculative for now. But many fish make use of underwater jets—or their opposite, suction—as a routine part of feeding. Given their social nature, if even one animal started shooting for the stars, every other archerfish would have a chance of sharing in the spoils. It might not be long before they started taking their own shot. For a young, scrappy and hungry fish, that’s not an opportunity you’d throw away lightly.

Further viewing

If you want to judge for yourself the difference between the notes of a mid-1700s naturalist and the recordings of a high-speed, high-definition camera system, this behind the scenes footage from the BBC Planet Earth III camera team is for you. Not that having all the latest equipment means that everything goes their way, of course:

Sources: Linnaeus, C. (1767) Systema Naturae, 13th ed. || Schlosser, J.A. (1764) XIV. An account of a fish from Batavia, called jaculator: In a letter to Mr. Peter Collinson, F.R.S. from John Albert Schlosser, M.D. F.R.S. Philosophical Transactions of the Royal Society 54: 89-91. || Schlosser, J.A. (1766) XXI. Some further intelligence relating to the jaculator fish, mentioned in the Philosophical Transactions for 1764, Art. XIV. from Mr. Hommel, at Batavia, together with the description of another species, by Dr. Pallas, F.R.S. in a letter to Mr. Peter Collinson, F.R.S. from John Albert Schlosser, M.D. F.R.S. Philosophical Transactions of the Royal Society 56: 186–188. || Shih et al. (2017) Archer fish jumping prey capture: kinematics and hydrodynamics. Journal of Experimental Biology 220: 1411-1422. || Schuster, S. et al. (2004) Archer fish learn to compensate for complex optical distortions to determine the absolute size of their aerial prey. Current Biology 14: 1565-1568. || Schlegel, T. (2006) Archerfish shots are evolutionarily matched to prey adhesion. Current Biology 16: R836-R837. || Gerullis, P. & Schuster, S. (2014) Archerfish Actively Control the Hydrodynamics of Their Jets. Current Biology 24: 2156-2160. || Schuster, S. (2018) Hunting in archerfish – an ecological perspective on a remarkable combination of skills. Journal of Experimental Biology 221: jeb159723. || Schuster et al. (2006) Animal Cognition: How Archer Fish Learn to Down Rapidly Moving Targets. Current Biology 16: 378–383. || Volotsky, S. et al. (2024) The archerfish uses motor adaptation in shooting to correct for changing physical conditions. eLife 12: RP92909. || Girard, M.G. et al. (2022) Phylogenetics of Archerfishes (Toxotidae) and Evolution of the Toxotid Shooting Apparatus. Integrative Organismal Biology 4: obac013. || Dewenter et al. (2017) Archerfish use their shooting technique to produce adaptive underwater jets. Journal of Experimental Biology 220: 1019-1025. || Harrington et al. (2016) Phylogenomic analysis of carangimorph fishes reveals flatfish asymmetry arose in a blink of the evolutionary eye. BMC Evolutionary Biology 16: 224. || Schuster, S. (2024) Shooting in archerfish: The art of transferring force to distant aerial objects. Encyclopedia of Fish Physiology (2nd Ed.), p.429-435. Academic Press.

Main image credit: Brehm’s Tierleben (1893) 3rd edition, p.49 “Schutzenfisch” || Second image credit: Schlosser (1766) || Third image credit: The Linnaean Society of London, document #L2021 || Fourth image credit: Gerullis & Schuster (2014) Figure 1 || Fifth image credit: Australian Museum; https://australian.museum/learn/animals/fishes/sevenspot-archerfish-toxotes-chatareus/ || Sixth image credit: Volotsky et al. (2024) Figure 1 || Seventh image credit: Girard et al. (2022) || Eighth image credit: Schuster (2024) || First video credit: Smithsonian Channel; https://www.youtube.com/watch?v=nrITCFbuhTw || Second video credit: BBC Earth; https://www.youtube.com/watch?v=Tq5KxIc12Uo