Picture this: a small, sandy-brown bird is foraging in the Negev Desert when it spots something coiled and still in the shade of a rock – a venomous viper, waiting in ambush. The bird doesn’t fly away. Instead, it starts calling loudly and fanning its wings, and within moments, several more birds have gathered around the snake, calling and displaying together until the threat gives up and the group moves on.
This is “mobbing,” and the bird is the Arabian babbler (Argya squamiceps), a desert-dwelling species that lives in cooperative social groups. Researchers at Tel Aviv University, led by Marie Guggenberger, studied a population at Shezaf Nature Reserve in Israel’s Negev Desert that’s been color-banded and monitored since Amotz and Avishag Zahavi started the project in 1971 – now about 120 birds across 21 groups, habituated enough to humans to be observed and recorded from just a few meters away – and found something remarkable hiding inside it: the babblers aren’t just making noise. They’re running two conversations at once. One is a private alarm – a call pitched so high the snake can’t hear a note of it. The other is a performance – a visual display aimed at the snake, but close enough that flockmates catch it too.
Why Arabian Babblers Mob Snakes Instead of Fleeing
Before getting into the mechanics of the signals themselves, it’s worth understanding why a bird would ever approach a venomous predator in the first place. Mobbing is inherently risky. By moving toward a viper instead of away from it, a babbler trades safety for proximity to danger.
But that risk buys something valuable. A coordinated mob can drive a predator off, reducing the chance that anyone in the group gets bitten later. It lets birds gather firsthand information about how dangerous the threat actually is. And prior research on this same population has floated a reputational angle too – one earlier study on Arabian babbler mobbing was framed explicitly around the possibility that showing up to defend the group functions partly as self-advertisement, a display of vigor that could matter later when birds are choosing dispersal partners.
Given those stakes, it makes sense that babblers have evolved a communication system finely tuned to make mobbing as fast, coordinated, and low-cost as possible.
Two Signals, Built for Two Different Sets of Ears and Eyes
The babbler’s toolkit has two components:
- The “zwick” call – a sharp, high-frequency vocal alarm, peaking around 6,200-6,300 Hz.
- The wing-lift – a visual posture where the bird spreads its tail and fans its wings while moving around the threat.
Here’s the trick: vipers cannot hear the “zwick” call at all. Snakes have no external or middle ear, and their hearing effectively cuts off below 1,000 Hz – well under the babbler’s calling frequency. So from the snake’s point of view, the call simply doesn’t exist. It travels freely to other babblers without ever registering as a threat cue to the predator itself.
The wing-lift solves the opposite problem. Ambush vipers like the Saharan sand viper hunt primarily through vision and by detecting vibrations traveling through the ground, and a bird spreading its wings and circling nearby delivers exactly that kind of signal. This lines up with an idea called the “pursuit-deterrent hypothesis”: by visibly showing itself, the babbler is essentially telling the snake, “I’ve spotted you, and I’m alert and ready to escape” – which may reduce the snake’s incentive to strike, since a successful ambush depends on surprise.
So the call recruits friends invisibly, and the display likely warns off the predator visibly. Researchers describe this as a “two-signal, two-receiver” system: one message, split cleanly across two channels, aimed at two audiences who each mostly pick up the part of it meant for them. There’s a small twist, though – raptors, unlike snakes, can likely hear the high-pitched “zwick” call, so the acoustic channel isn’t perfectly private. It’s tuned specifically around the hearing limits of snakes, not every possible predator.
Why Showing Up With Both Signals Works Better – Probably
Signals sent together turn out to be more powerful than the sum of their parts, at least in terms of speed. When another babbler is exposed to a combined call-and-wing-lift display, it responds in about 0.61 seconds on average. To a call alone, that response takes about 0.84 seconds – nearly a quarter-second slower. Interestingly, a wing-lift by itself, with no call, produced a response nearly as fast as the combined signal – both were significantly quicker than a call alone, and not significantly different from each other.
The type of response also changes depending on what’s received. A bystander who hears just a call is likely (about 85% of the time) to simply call back. But a bystander who sees a wing-lift, whether alone or combined with a call, is far more likely (roughly 59-63% of the time) to respond with a fully combined display of its own.
One honest caveat from the researchers themselves: because a bird’s call is physically altered whenever it’s produced during a wing-lift (more on that below), it’s hard to be fully certain whether the faster response is driven by the visual signal itself, or simply by the fact that wing-lift calls sound acoustically different. The study can’t yet fully rule out that possibility – untangling it would likely require playback experiments that isolate the two channels.
When Posture Bends Sound
Perhaps the most unexpected discovery in this research is that these two signals physically interact. Holding a wing-lift posture requires sustained effort from muscles that put mechanical pressure on the upper part of the bird’s vocal tract. As a result, a “zwick” call produced during a wing-lift sounds measurably different from one produced while the bird is standing still.
Three specific changes show up:
- Pitch drops – by about 90 Hz on average, and by as much as 241 Hz near the end of a long display.
- Calls lengthen slightly – by roughly 9.2 milliseconds (about 8.5%), a shift that falls below the 10-20% threshold typically needed for birds to consciously discriminate call duration – so it likely isn’t perceived as a deliberate signal.
- Calls come faster – the gap between one call and the next shrinks by about 165 milliseconds.
Whether these shifts mean anything to a listening babbler depends on whether they cross known perceptual thresholds. Birds can typically detect pitch changes of about 1% or less, and the frequency drops measured here (1.4-3.7%) clear that bar comfortably – so other babblers likely do register the lower, “strained” pitch as meaningful. The faster call rate is even easier to detect, since the timing shift is far larger than the minimum gap a bird needs to notice a change in rhythm.
Toward the end of a mobbing bout, the pattern reverses: calls slow back down and pauses lengthen again, and each successive wing-lift a given bird performs tends to be shorter than the last – a signature, researchers suggest, of accumulating physical fatigue from repeatedly holding a demanding posture.
The Group Discount: Why More Birds Means Less Individual Effort
Not every babbler pulls the same weight in a mob, and the deciding factor is often simply how many others have shown up. As group size increases from a single bird up toward nine, individual call rates fall by around 77% – a clear, statistically robust decline. Time spent in the wing-lift posture also trends downward by a comparable amount (around 56%), though this particular pattern was more marginal statistically and not quite as certain as the drop in calling.
Two ideas from behavioral ecology help explain this pattern. Social buffering describes how being near familiar group members can calm an animal’s physiological stress response, lowering the felt urgency to keep signaling. Predation risk dilution describes a more mathematical effect: in a bigger group, any one bird’s odds of being the snake’s actual target go down, so there’s less individual payoff to working flat-out.
What’s notable is that this scaling-back of effort has nothing to do with rank. When researchers tracked which birds joined a mob first, the order was unrelated to sex, age, or social status – dominant breeders were no faster to respond than young, low-ranking helpers. Predator defense, in this species, looks less like a job assigned by hierarchy and more like a responsibility that falls to whoever happens to be closest when the alarm goes up.
Different Birds, Different Signals
Age and sex do shape how individuals use the two-signal system.
Young birds under six months old call at a noticeably higher rate than adult females (about 0.32 calls per second versus 0.19). Juveniles are also the group most likely to perform a silent wing-lift – the visual display without any accompanying call – doing so about 25% of the time. Taken together, researchers interpret this as a signature of inexperience: young birds haven’t yet learned to smoothly coordinate their two channels.
Adult females also perform silent wing-lifts fairly often – around 17% of the time – far more than adult males, who almost never go silent (about 4%). The likely explanation here isn’t inexperience but tactics: females tend to spend more time in close, direct confrontation with the snake, and going silent may let them commit their full attention to deterrence.
Silent wing-lifts occur in about 20% of displays overall, but vocal effort is at its highest when a bird is on its own. When an individual first discovers a snake, it relies on ‘zwick’ calls to recruit groupmates from a distance. As more flockmates arrive and group size grows from one to nine, individual calling rates drop by 77% – showing that while silent displays are shaped primarily by age and sex, heavy vocalization is most intense before help arrives.
Catching It All on Camera
None of this fine-grained detail would be visible to the naked eye. Telling exactly which bird in a fast-moving cluster of five to ten made a particular call, and matching it to the millisecond with a wing posture, required specialized equipment: a SoundCam 2.0 acoustic camera that visualizes sound as a pressure map, run in sync with a high-resolution Canon EOS 5D video camera used to track posture and identify birds by their individual leg-band colors, positioned about three meters from the action.
To reliably bring on mobbing behavior, researchers used realistic dummy vipers, partly buried in the sand to mimic an ambush position. Because these decoys don’t move the way a live snake would, the resulting mobbing bouts were likely somewhat shorter than they’d be with a real predator – and notably, the babblers never actually attacked the dummy, unlike their documented behavior toward other snake species, eventually just flying off instead. The resulting audio and video – over 3,000 individual calls in total – were annotated with BORIS and Adobe Premiere Pro and analyzed acoustically with Raven Pro, achieving better than 91% synchronization accuracy between audio and video streams.
The Takeaway
Strip away the technical details, and what’s left is a genuinely elegant piece of natural engineering, with a few open threads still being pulled on. A single Arabian babbler, spotting a snake, doesn’t just sound one alarm – it runs two signals that are usually coordinated, each shaped for a different kind of listener, and each capable of working somewhat on its own. The very act of performing one signal reshapes the other, in ways that make the acoustic signal itself carry more information than it otherwise would. The intensity of the response scales down as the group’s safety in numbers increases. And even the choice of whether to call at all shifts depending on who’s doing the signaling and who else is around to help.
It’s a reminder that “communication” in the animal world is rarely a single, simple broadcast. For the Arabian babbler, warning a friend and warning off an enemy are two different jobs – handled, brilliantly, in the very same breath.
Source
Study: A two-signal, two-receiver signalling system: Snake mobbing in the Arabian babbler
Authors: Marie Guggenberger, Oded Keynan, Yossi Yovel (2026)
Read the full paper: https://www.biorxiv.org/content/10.64898/2026.08.25.747098v2
Leave a Reply