Fork-tailed drongos in South Africa’s Kalahari Desert can produce up to 51 distinct mimicked alarm calls and deploy them selectively against whichever neighboring species holds the most food. Researchers tracked 64 individual drongos over 847 hours of field observation at the Kuruman River Reserve, documenting a system of vocal deception far more flexible than scientists had previously recognized. The findings raise a pointed question: if a small African passerine can run a con this sophisticated, what stops other vocal mimics, including crows and ravens, from doing the same?
How 51 fake alarms let drongos steal meals in the Kalahari
The core discovery centers on a behavior called kleptoparasitism, in which one animal steals food that another has caught or gathered. Drongos perch near mixed-species foraging groups in the Kalahari scrubland, watching meerkats and pied babblers dig up insects and larvae. When a target finds a high-value food item, the drongo sounds a false alarm call that mimics the target’s own species-specific warning. The target drops its meal and flees; the drongo swoops in and eats.
What makes this system remarkable is the switching. Targets quickly learn to ignore a repeated false alarm. A meerkat that hears the same drongo-produced warning twice in a row stops responding. To counter that habituation, drongos rotate through their repertoire of up to 51 mimicked warning calls, drawing on the alarm vocabularies of multiple species in the area. When one call loses its effect, the drongo switches to a different species’ alarm or even to its own species-specific mobbing call, resetting the deception.
This flexible variation strategy was documented in a field study published in Science, which showed that targets reduce their responses when a false alarm is repeated but keep reacting when the sequence of calls varies. In that work, researchers followed color-banded drongos and recorded their vocalizations and the behavior of nearby animals, demonstrating that the birds alter their calls as soon as a particular sound stops yielding food. The evidence supports the idea that drongos are not simply noisy opportunists; they track which signals are still effective and adjust in real time.
Crucially, the Science study did more than catalog an impressive vocal range. It linked specific call types to measurable outcomes, such as whether a meerkat abandoned its prey or whether a pied babbler resumed foraging after an alarm. By combining detailed observations with playback experiments, the authors were able to show that variation itself is the key to sustained deception: when the same false alarm was played repeatedly, responses dropped off, but when different alarm types were interleaved, the rate of food theft stayed high.
Field data, spectrogram proof, and how we know it is mimicry
The evidence for drongo deception rests on multiple layers of primary research. An earlier study published in Proceedings of the Royal Society B first proposed and documented the kleptoparasitism-by-false-alarm mechanism, establishing that drongos direct mimicked calls at foraging competitors to steal food. In that work, detailed focal follows revealed that food theft attempts were tightly associated with alarm-like vocalizations, and that successful thefts were far more likely when potential victims actually fled in response.
That initial study laid the observational groundwork for the later, larger project that quantified the 847-hour tracking effort and the 51-call repertoire. Together, the two datasets bridge a gap between anecdote and mechanism: they show not only that drongos shout and then steal, but that they do so in a way that appears tuned to their listeners’ learning and forgetting.
Separate research using spectrogram comparisons confirmed how scientists distinguish genuine mimicry from coincidental vocal similarity. By analyzing the acoustic structure of drongo vocalizations against recordings of the species they imitate, researchers demonstrated that the mimicked calls closely match the originals in frequency, timing, and tonal shape. This kind of context-dependent analysis also showed that drongos produce their mimicked calls disproportionately in alarm-related situations rather than during song or territorial displays, reinforcing the idea that the imitations are part of a functional deception system.
Those spectrograms matter because they close off a simple alternative explanation: that drongo calls just happen to sound a bit like other birds. Instead, the data show that when a drongo mimics a meerkat’s alarm, the match is so close that human observers cannot reliably tell the difference by ear, and neighboring animals behave as though a genuine warning has been given. The mimicry is precise, and it is deployed in exactly the contexts where it can move food from one beak to another.
What about crows and ravens?
Corvids, the family that includes crows, ravens, magpies, and jays, are also well-known vocal mimics. In captivity and in urban environments, they can copy human speech, mechanical noises, and other birds’ calls. Their cognitive abilities rival those of many primates, and their social lives are complex. It is tempting, therefore, to assume that if drongos can run elaborate vocal scams, corvids must be doing something similar.
Yet the current scientific record does not support that leap. While there are many reports of corvids imitating sounds, there is no equivalent field dataset that links corvid mimicry to systematic food theft in the way the drongo studies do. A recent synthesis of corvid vocal behavior catalogs which species copy which kinds of sounds and under what general circumstances, but it does not provide the kind of paired call-and-response data-who called what, who fled, who stole-that underpins the Kalahari work.
In other words, the specific headline claim that “crows can mimic over 50 alarm calls, aiming them at whichever species has the most food” currently fits drongos, not corvids. Fork-tailed drongos are the species for which researchers have documented a large, flexible alarm repertoire, demonstrated context-specific deployment, and shown that variation in calls maintains the effectiveness of deception. Whether any crow or raven uses mimicry in an equally targeted way remains an open empirical question.
Untested predictions and the gap between drongos and corvids
The drongo research nonetheless points toward clear, testable predictions for corvid biology. Species that forage in mixed-species groups and compete for clumped, high-value food patches should face strong selection for context-dependent alarm mimicry. If a crow or raven regularly feeds alongside smaller birds or mammals that respond to particular warning calls, there is an obvious potential payoff to faking those signals when a rival uncovers a prized item.
Researchers could probe this by combining detailed behavioral observations with targeted playback and food supplementation experiments. One approach would be to identify corvid populations that habitually forage near other species, then record their vocalizations and the responses of neighbors during normal feeding. After establishing a baseline, experimenters could provide temporarily rich food sources and test whether the frequency or type of mimicry shifts when competition intensifies. Crucially, they would need to document not just that mimicry occurs, but that it reliably precedes successful thefts.
Several factors complicate a direct comparison between drongos and corvids. Corvids tend to be among the dominant competitors in their foraging communities, often able to displace rivals through size, aggression, or social mobbing. Drongos, by contrast, are smaller than many of their targets and would likely lose most direct confrontations. For them, deception may be the only viable route to a stolen meal. The evolutionary pressure to develop flexible alarm mimicry may therefore be strongest in species that lack the physical means to take food by force.
The structure of the local community also matters. Drongos operate in a relatively stable assemblage of mammals and birds that share a common set of predators and, therefore, a broadly overlapping alarm vocabulary. That ecological backdrop makes it worthwhile for a single species to invest in learning dozens of different warning calls. In more fluid or species-poor systems, where neighbors change frequently or share fewer predators, the payoff to such extensive mimicry might be lower.
Finally, there is the question of cognition and constraint. Drongos show that a small passerine with a modest brain can track which calls work on which targets and swap signals when one loses its punch. Corvids clearly have the neural machinery for at least equal flexibility. The absence of evidence for drongo-style food theft in crows and ravens may reflect a lack of focused fieldwork rather than a genuine absence of the behavior. Until researchers apply the same intensive, call-by-call methods to corvids that they have brought to bear on drongos, the safest conclusion is a narrow one: in the Kalahari, it is the fork-tailed drongo-not the crow-that currently stands as the master of vocal fraud.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
