# Deception by Flexible Alarm Mimicry in an African Bird

> In the Kalahari, fork-tailed drongos act as watchmen for babblers and meerkats, but when one finds a juicy meal, a drongo may shout a false alarm and steal it. This study shows that drongos mimic their targets’ own alarm calls and switch to a different call when a trick has just failed, so their victims never get used to the lie.

Summary by Kersey Lawrence, Original Wisdom. Updated 9 November 2026. Page: https://www.originalwisdom.com/tracking-literature/2014-flower-gribble-deception-flexible-alarm-mimicry/

## Citation and access

- **Authors:** Tom P. Flower, Matthew Gribble and Amanda R. Ridley
- **Year published:** 2014
- **Journal or source:** Science
- **Full citation:** Flower, T. P., Gribble, M. and Ridley, A. R. 2014. Deception by flexible alarm mimicry in an African bird. Science 344(6183): 513–516.
- **DOI:** 10.1126/science.1249723
- **Access status:** Publisher link only
- **Where to find it:** https://doi.org/10.1126/science.1249723

## Tags

- **Species:** Birds > Fork-tailed drongo; Birds > Southern pied babbler; Mammals > Suricate
- **Region:** Africa > South Africa
- **Behaviour and ecology:** Behaviour; Communication; Feeding; Kleptoparasitism
- **Methods and technology:** Direct observation
- **Data type:** New field data
- **Observer qualification:** Not needed
- **Discipline:** Behaviour and ecology
- **Year published:** 2014

## Detailed keywords

- **Species:** Studied: fork-tailed drongo (Dicrurus adsimilis), the deceiver, and southern pied babbler (Turdoides bicolor), the target in the playback experiments. Other target species: meerkat (Suricata suricatta). Mimicked in the experiments: Cape glossy starling (Lamprotornis nitens). Mentioned only for comparison: infectious organisms that cause influenza, sleeping sickness (written in the paper as Trypanosomiasis brucei) and malaria (Plasmodium falciparum).
- **Location and habitat:** Kalahari Desert. The paper names Northern Cape Conservation (research permit) and the Kalahari Research Trust (site access), which places the study in the Northern Cape, South Africa; no reserve name is given in the main text.
- **Topics:** deception; false alarm calls; vocal mimicry; alarm call mimicry; food theft (kleptoparasitism); eavesdropping on other species’ alarm calls; habituation; frequency-dependent constraints on deception; honest and dishonest signals; tactical deception; feedback in communication; drongo sentries and target vigilance; comparison with antigenic variation in disease organisms; theory of mind.
- **Statistics or models:** generalized linear mixed models (GLMM, binomial); linear mixed models (LMM) on log response time; Tukey post-hoc tests; mean and standard error (SE). Possible inconsistencies: (1) the chi-square values reported for the binomial GLMMs (for example 2.55 with P = 0.011 in Fig. 1, 2.15 with P = 0.039 in Fig. 4A, 2.01 with P = 0.045 in Fig. 4B) would give much larger P values as chi-square tests with 1 degree of freedom; the P values fit if these are z values; (2) the Fig. 3 legend gives “P< 0.996” for a non-significant result, probably meaning P = 0.996; (3) methods details are cited to reference 7 (Thorogood and Davies 2012, about cuckoos), while the supplementary materials are reference 14, so the numbering looks shifted. The supplementary materials (methods, tables S1 to S5) are not in this PDF.
- **Data type:** New field data: 847.5 hours of focal observations on 64 colour-ringed wild drongos (688 food-theft attempts), plus two playback experiments on habituated pied babblers. Dates of data collection are not given in the main text (the research permit is dated 2008). Data are in the Dryad Digital Repository (doi 10.5061/dryad.f60j5).

## Synopsis

### 1. What they did

Fork-tailed drongos in the Kalahari spend over a quarter of their time following other animals, such as southern pied babblers and meerkats, often for several hours a day. They act as sentries, giving honest alarm calls when they see a predator approaching, and the other animals flee to cover when they hear them. But when one of those animals finds a large food item, a drongo may give a false alarm, then swoop in to steal the abandoned food. Drongos can produce many alarm calls, including mimicked alarm calls of other species. The authors asked why drongos use so many different false alarms.

They worked with 64 wild drongos that were used to observers and could be recognised by coloured leg rings. The study had four parts:

- Observations. They watched drongos for 847.5 hours and recorded 688 food-theft attempts, noting which alarm call was used on which target.

- Playback experiment 1. They gave 20 babblers (in 10 groups) a food item and played four drongo calls 20 minutes apart: a territory call (not an alarm), a drongo-specific alarm, a drongo mimic of a Cape glossy starling alarm, and a drongo mimic of a pied babbler alarm. They timed how long each babbler took to go back to foraging.

- Playback experiment 2. They played 22 babblers (in 11 groups) three alarm calls in a row, 20 minutes apart. In two treatments all three calls were the same; in the other two, the third call was switched between a drongo-specific alarm and a mimicked starling alarm. Treatments were 3 days apart.

- Repeated thefts in the wild. They recorded 151 occasions when a drongo (42 drongos in total) made repeated food-theft attempts on the same target, and whether the drongo changed its call.

### 2. What they found

Drongos’ alarm calls

- An individual drongo’s repertoire ranges from 9 to 32 calls (average 17). In all, 51 alarm call types have been recorded in false alarms: 6 are drongo-specific and 45 mimic other species.

- Food theft makes up 23% of what drongos eat, by biomass.

- Drongos that could mimic a target’s alarm call did so more often when stealing from that species than when stealing from other species (147 food-theft attempts by 10 drongos).

Experiment 1: does mimicry work better?

- Babblers took longer to go back to foraging after mimicked alarms (of babblers or starlings) than after drongo-specific alarms, and longer after any alarm than after the territory call.

- Babblers did not respond differently to mimics of their own alarm than to mimics of the starling alarm.

Experiment 2: does changing the call work better?

- When the same alarm call was played three times in a row, the babblers responded less each time.

- When the third call was changed, the babblers kept responding as strongly, and they responded longer to a changed third call than to a repeated one. This held whether the drongo-specific or the mimicked starling call came first.

In the wild

- On repeated attempts on the same target, drongos changed their alarm call type on 74% of occasions.

- They were more likely to change their call after a failed attempt.

- After a failed attempt, changing the call made the next theft more likely to succeed than repeating the same call.

### 3. What they concluded

- By varying their false alarms, drongos gain twice: they use the calls most likely to fool a particular target, and they avoid the target getting used to one repeated false alarm.

- This lets drongos get around the usual limit on deception, where a lie stops working once it is used too often compared with the honest signal. The authors compare this to disease organisms that keep changing their surface proteins to stay ahead of the immune system.

- Flexible variation may not work forever: targets might eventually ignore all the calls in a drongo’s repertoire, or recognise individual drongos and judge their reputation.

- The study shows a deceptive use for vocal mimicry, a behaviour for which few benefits have been shown, and suggests that paying attention to the target’s response can be adaptive.

- Behaviour like this, honest and dishonest alarms used tactically, has been taken as evidence of advanced mental abilities similar to theory of mind. The authors note that similar behaviour in different species need not have the same underlying mechanism.

### 6. Other technologies, skills or innovations

- Playback experiments with recorded drongo calls, including drongo mimics of other species’ alarm calls, shown as spectrograms.

- Long-term habituated study population: drongos, babblers and meerkats used to observers, with individuals recognised by colour rings.

## Cite this summary

Lawrence, K. 2026. Summary of: Deception by Flexible Alarm Mimicry in an African Bird (Flower and Gribble et al., 2014). Original Wisdom, Tracking Literature. https://www.originalwisdom.com/tracking-literature/2014-flower-gribble-deception-flexible-alarm-mimicry/