Sparrowhawk imitation or convergent alarm signal? A new hypothesis for bubbling call of cuckoos with empirical testing

Abstract

Alarm calls in bird vocalizations serve as acoustic signals announcing danger. Owing to the convergent evolution of alarm calls, some bird species can benefit from eavesdropping on certain parameters of alarm calls of other species. Vocal mimicry, displayed by many bird species, aids defense against predators and may help brood parasites during parasitism. In the coevolutionary dynamics between brood parasites, such as the common cuckoo (Cuculus canorus), and their hosts, female cuckoo vocalizations can induce hosts to leave the nest, increasing the probability of successful parasitism and reducing the risk of host attacks. Such cuckoo calls were thought to mimic those of the sparrowhawk. However, owing to their similarity to alarm calls, we propose a new hypothesis: Female cuckoos cheat their hosts by mimicking the parameters of the host alarm call. In this study, we tested this new hypothesis and the sparrowhawk mimicry hypothesis simultaneously by manipulating the syllable rate in male and female common cuckoo vocalizations and playing them in front of the host Oriental reed warbler (Acrocephalus orientalis) for examination. The results indicate that similar to a normal female cuckoo call, a female call with a reduced syllable rate prompted the hosts to leave their nests more frequently and rapidly than male cuckoo calls. Additionally, the male cuckoo calls with increased syllable rate did not prompt the host to leave their nests more frequently or quickly compared with the male cuckoo calls with a normal syllable rate. Our results further confirm that female common cuckoos mimic the vocalizations of Eurasian sparrowhawks (Accipiter nisus), reveal the function mechanisms underlying such mimicry, and support the theory of imperfect mimicry.

Communication among organisms occurs through auditory, visual, olfactory, and tactile means (Partan 2013; Rubi and Stephens 2016). Auditory-based communication relies predominantly on sound signals. Birds produce sounds through the vibration of the syrinx, and birdsong includes both songs and calls (Suthers 1990). Songs generally refer to the longer, more complex vocal signals produced by males during the breeding season, which serve to attract females or proclaim territorial boundaries. Calls encompass all vocalizations other than songs. Typically, calls are shorter in duration and simpler in structure than songs and can be produced by males and females year-round as a means of communication (Collins 2004). Calls have a wider range of applications than songs: Calls are used to attract mates and defend territory and for various other biological functions. Calls can be categorized into food calls, contact calls, flight calls, alarm calls, mobbing calls, and begging calls (Catchpole and Slater 1995).

An alarm call serves as a rapid, effective means of conveying danger signals. It can convey knowledge about the type of threat (Pereira and Macedonia 1991), the direction from which it is approaching (Evans et al. 1993), and its urgency (Templeton et al. 2005). The innate ability to recognize conspecific alarm calls is well documented (Herzog and Hopf 1984; Hollén and Radford 2009). In addition to warning conspecifics, these calls can be used by other species vulnerable to the same threats, who eavesdrop and gain information about danger (Fallow and Magrath 2010). By contrast, many studies have indicated that the recognition of heterospecific alarm calls is acquired through experience and learning (Hauser 1988; Ramakrishnan and Coss 2000; Magrath and Bennett 2012). However, some playback experiments have shown that birds respond to heterospecific alarm calls even when they have had no prior opportunity to learn them (Johnson et al. 2003; Fallow et al. 2011, 2013). These responses to unfamiliar alarm calls may have been due to the heterospecific calls sharing acoustic properties with the conspecific calls of the species tested, which is consistent with the view that alarm calls share acoustic structures that prompt unlearned responses (Rendall et al. 2009). Marler (1957) analyzed the alarm calls of various bird species and suggested that alarm calls are structurally and functionally similar, leading to the hypothesis of convergent evolution in bird alarm calls.

Convergent evolution refers to the phenomenon where species with different evolutionary origins develop similar phenotypic or behavioral traits due to adaptation to similar environments (Grant et al. 2004; Losos 2011). Examples of convergent evolution are countless. For instance, the repeated evolution of flight is a classic example: Flying insects, birds, pterosaurs, and bats all independently evolved wings or similar structures to achieve flight. Similarly, marine mammals such as whales and dolphins, after adopting an aquatic lifestyle, evolved streamlined bodies, fins, and tails similar to those of fish and show some degree of convergence in their movement patterns (Gleiss et al. 2011). In studies on convergent evolution of alarm calls in birds, Jurisevic and Sanderson (1994) studied the alarm calls of 30 Australian bird species and revealed that the alarm calls of 26 of these species consisted of rapidly repeated and species-specific single-syllable types. This finding indicated that the rapid repetition of a single-syllable type may be a shared acoustic feature driving convergent evolution in the alarm calls of birds. We assessed the audio recordings from Jiang et al. (2022); our findings showed that in areas where the common cuckoo coexists with hosts, such as the Oriental reed warbler and the vinous-throated parrotbill (Sinosuthora webbiana), the alarm calls of these hosts share the characteristic of rapid repetition of a single syllable.

The similarity of the sounds might also be attributed to vocal mimicry (Dalziell et al. 2015). Mimics can simulate signals or cues in various ecological contexts, including antipredation; promotion of parasitism; and attraction of mates, pollinators, or prey (Ruxton et al. 2004; Kilner and Langmore 2011; Schiestl and Johnson 2013). Mimicry is estimated to occur in vocalizations of approximately 15–20% of all songbirds (Marshall 1950), with 1 survey suggesting that up to 40% of European passerines exhibit this behavior (Garamszegi et al. 2007). Similarly, non-passerines exhibit this behavior: For example, burrowing owls (Athene cunicularia) produce hisses resembling the rattle of a rattlesnake (Rowe et al. 2010).

Parasitic cuckoos also benefit from vocal mimicry. They are obligate avian brood parasites that have evolved plumage patterns and colors similar to those of sympatric raptors (Davies and Welbergen 2008; Spottiswoode 2013). Cuckoos do not construct nests; instead, they lay their eggs in the nests of other species, relying on the host to rear their offspring (Davies 2000). Preventing parasites from laying eggs in host nests is an effective strategy that reduces the risk of parasitism (Feeney et al. 2012). However, the degree of harm to parasites varies depending on the host species and defensive action. For example, Oriental reed warblers mob and even kill common cuckoos during nest defense (Zhao et al. 2022). Several parasitic species employ strategies that mimic host predators to minimize the risk of injury. For example, many Cuculus cuckoos visually resemble the host predator, the Eurasian sparrowhawk (Trnka and Prokop 2012; Møller et al. 2015; York and Davies 2017). Moreover, the “bubbling call” of the adult female common cuckoo mimics the vocalizations of the Eurasian sparrowhawk, serving as a deceptive form of acoustic mimicry akin to phenotypic mimicry (York and Davies 2017). This mimicry capitalizes on the host’s recognition of sparrowhawk vocalizations and can distract the host’s attention from the clutch during parasitism when the host is out of the nest. This may increase the probability of parasitism and reduce the risk of host attack, ultimately increasing the likelihood of parasitic egg acceptance (York and Davies 2017; Wang et al. 2022).

Although studies have suggested that the common cuckoo influences its hosts by mimicking the vocalizations of a sparrowhawk, there are 2 possible explanations for the host’s response to the cuckoo’s bubbling call. This call, characterized by the rapid repetition of a single syllable, may lead hosts to mistakenly regard it as either 1) a sparrowhawk call or 2) an alarm call. Although research has hypothesized that the cuckoo’s bubbling call mimics the sparrowhawk call, we propose a new hypothesis: It may also resemble the typical features of an alarm call. These 2 hypotheses are not mutually exclusive. To reveal the potential element of alarm call mimicry in bubbling call, we used sound manipulation to reduce the rapid syllable rate of the bubbling call to match that of the normal call of a male common cuckoo (i.e., the “cu-cu” call) while maintaining the syllable unchanged. We also increased the slow syllable rate of the cu-cu call to match that of the bubbling call in a contrasting manner. A playback experiment was then performed to investigate the reaction from the hosts. We hypothesize that 1) the manipulated bubbling call would reduce the reaction intensity or change the reaction mode of the hosts compared with that of the natural bubbling call if the natural bubbling call is composed of an element of mimicry to the alarm call. If this prediction is supported, the support suggests that it no longer mimics an alarm call as well but equally could also no longer perfectly mimic a sparrowhawk call as well. If it is not supported, this goes against the alarm call hypothesis. Furthermore, we also predict that 2) elevating the syllable rate of the cu-cu call would increase the reaction intensity or change the reaction mode of the hosts compared with case for when the natural cu-cu call was used. Finding support for this would suggest that an increased syllable rate increases the “alarm-call”-like nature of the call, which may indirectly support the notion that the bubbling call has “alarm-call”-like properties. Finding no support for this would again go against the alarm call hypothesis.

Materials and Methods

Study area and species

The study was conducted in Baiyangdian National Wetland Park in Hebei Province (38°43ʹ–39°10ʹN, 115°38ʹ–116°19ʹ) and Zhalong National Nature Reserve in Heilongjiang Province (46°48ʹ–47°31ʹN, 123°51ʹ–124°37ʹ). The main vegetation in both wetlands consists of reeds (Phragmites australis) and cattails (Typha latifolia), which provide suitable nesting habitats for Oriental reed warblers. The breeding season for hosts in these regions extends from May to August (Wang et al. 2022; Trnka et al. 2023). The Oriental reed warbler is a commonly observed host of the common cuckoo in both studied areas, exhibiting high levels of coevolution (Yang et al. 2014, 2016, 2017). Oriental reed warblers in these areas defend their nests against cuckoos, and the female cuckoo call evokes nest departure in approximately half of the hosts when they are incubating (Yu et al. 2019a; Wang et al. 2022).

Sound manipulation

Natural calls of male and female common cuckoos were identified from an existing study that combined different samples of each type of sound to avoid pseudo-replication (Wang et al. 2022). The bubbling call of female cuckoos is characterized by a rapid repetition of single-syllable types, which share acoustic parameters with alarm calls and can increase the vigilance of incubating hosts. To ascertain whether the bubbling call mimics an alarm call, we altered its rapid syllable rate characterized by alarm calls, reducing it to match the natural syllable rate of the “cu-cu” call from a male cuckoo (hereafter manipulated bubbling call; Figure 1). We selected the syllable rate of male cuckoos as a reference and control not only because the cu-cu call is from the same species of cuckoos but also because the bubbling call was found to trigger a stronger reaction (i.e., leaving the nest more frequently) than the cu-cu call in hosts (Wang et al. 2022). Furthermore, the reaction of the hosts to the cu-cu call is consistent with the vocalizations of a harmless control species, the Oriental turtle doves (Streptopelia orientalis) (Wang et al. 2022). Therefore, the cu-cu call of the male common cuckoo is an optimal control and reference for syllable rate manipulation in this study. Corresponding to the manipulation of bubbling call, we increased the syllable rate of male common cuckoo calls to match the natural syllable rate of females (hereafter a manipulated cu-cu call; Figure 1), imbuing them with characteristics typical of alarm calls. Natural calls of male and female cuckoos were used as controls (hereafter a natural bubbling call and natural cu-cu call). Therefore, 4 types of playback (2 min each) were used in the playback experiment: 2 manipulated calls (manipulated bubbling and cu-cu calls) and 2 controls (natural bubbling and cu-cu calls). The natural bubbling call was a control for the manipulation of the syllable rate, and the natural cu-cu call was a control for both predator mimicry and manipulation of the syllable rate. Syllable rate manipulation and noise elimination were achieved via Raven Pro 1.6 software (Cornell Lab of Ornithology, Cornell, NY, USA). The amplitude of each sound was set to 80 dB at a distance of 1 m from the speaker and was measured via a digital sound level meter with a fast-response setting (SmartSensor Ar824, Shenzhen, China).

Figure 1.

Sound spectrogram of common cuckoo callings and their manipulation.

Playback procedure

Playback experiments were performed during the breeding season of Oriental reed warblers from May to August 2021. During the incubation period, each host nest received 4 sounds (2 manipulated calls and 2 normal calls) in randomized order, creating 72 records of host responses from 18 nests (9 nests for each population). A miniature camera (WJO3, Hisilicon, Shenzhen, China) was installed ~30 cm above each nest, and a playback device (BV370, SEE ME HERE Electronic Corporation, Shenzhen, China) was placed 1 m from the nest (Yu et al. 2019b; Wang and Yang 2020; Wang et al. 2022). The observer remained concealed in reeds ~10 m away. After the host parent returned to the nest and incubated for 2 min, the sound playback commenced and lasted for 2 min. The minimum interval between 2 consecutive playback sessions was 15 min, and the maximum interval was 30 min. The behavioral reactions of incubating hosts, including 1) staying in or leaving the nest and 2) latency to leave the nest (measured in seconds) were recorded (Wang et al. 2022). We also paid attention to any changes in the reaction mode from the hosts. Specifically, we recorded whether the hosts would leave the nests but wander around to spot the sound source of playback. The Oriental reed warblers wandered around their nests after they heard the playback of alarm calls (Wang and Yang 2020). However, we did not find such a reaction from the hosts. The only identified reaction was leaving or staying in the nests.

Statistical analyses

Generalized linear mixed models based on Markov chain Monte Carlo techniques (MCMC-GLMMs) were used to investigate the effects of treatments (playback of natural bubbling and cu-cu callings and their manipulations), populations (Baiyangdian and Zhalong), playback order (4 orders), incubation day, clutch size, and date of experiment on the behavioral reaction of the hosts (leaving or staying). The nest identities were included in the random effect analysis. In MCMC-GLMMs, a Bayesian model with a default value of prior probability was used to calculate the posterior means and 95% credible intervals (CI), with a 95% CI excluding zero indicating a significant result (Gómez-Rubio 2020). Deviance information criterion (DIC) was used to evaluate the model; thus, the final model was confirmed according to the values of DIC (Supplementary Material). Kaplan–Meier curves with Gehan–Wilcoxon test, which calculate the probabilities of events by integrating both the occurrence of the events and their latency time (Cuthill et al. 2005; Stevens et al. 2008), were used to investigate the probability of the hosts staying in nests during the playback experiment, incorporating the host reaction and its latency (Yang et al. 2023; Yang and Zhang 2024). The MCMC-GLMMs, Kaplan–Meier curves, and Gehan–Wilcoxon tests were performed using the MCMC-glmm, survival, and survminer packages in R (v. 4.2.2) for Windows (R Foundation for Statistical Computing, Vienna, Austria).

Results

The model of MCMC-GLMMs including playback treatment, population, and playback order as predictors has the smallest value of DIC (100.62) that playback treatment alone could predict the host reactions. Specifically, host responses to manipulated bubbling calls differed significantly from responses to manipulated cu-cu calls (posterior mean and 95% CI: −0.422 and −0.721 to −0.131, MCMC-GLMM; Figure 2) and natural cu-cu calls (Posterior mean and 95% CI: −0.323 and −0.614 to −0.028, MCMC-GLMM; Figure 2). Conversely, the responses to manipulated and natural bubbling calls did not show a significant difference (posterior mean and 95% CI: −0.252 and −0.552 to 0.049, MCMC-GLMM; Figure 2). Population and playback order did not predict the host reactions (Figure 2). The total rates of nest abandonment by the hosts during playback were 27.78%, 22.22%, 38.89%, and 61.11% (n = 18 for each) in response to the natural cu-cu call, manipulated cu-cu call, natural bubbling call, and manipulated bubbling call, respectively. Therefore, the leaving rate of hosts was higher in response to the bubbling call than to the cu-cu call. Additionally, Kaplan–Meier survival analyses revealed that the probability of hosts staying in nests differed among treatments (χ² = 8.1, df = 3, P = 0.04, Gehan–Wilcoxon test; Figure 3) and between syllable types if the manipulated and natural calls of either bubbling or cu-cu call were aggregated, respectively (χ² = 7.3, df = 1, P = 0.007, Gehan–Wilcoxon test; Figure 3).

Figure 2.

Results of generalized linear mixed models using Markov chain Monte Carlo techniques (MCMC-GLMMs). Points and bars refer to the posterior means and 95% CI in the Bayesian model of the MCMC-GLMM, respectively. Significant predictors were highlighted by red color.

Figure 3.

Kaplan–Meier curves illustrate the probability of the hosts staying in the nest during the playback experiment. P values were calculated using the Gehan–Wilcoxon test.

Discussion

The results indicate that neither reducing the syllable rate of the bubbling call nor increasing the syllable rate of the cu-cu call altered the host’s response to natural cuckoo vocalizations. Therefore, our 2 hypotheses were not supported. The hosts were found to leave their nests more frequently in response to the bubbling call than to the cu-cu call. The leaving behavior without wandering around the nests was a fear reflection from the hosts, which was consistent with the reaction to the natural bubbling call of hawk mimicry, according to this study and an existing study (Wang et al. 2022). Consequently, our findings do not support the hypothesis that cuckoo vocalizations mimic the convergent evolution of alarm call parameters. By contrast, our data provide evidence that further confirms that female cuckoo vocalizations mimic those of the sparrowhawk. Specifically, this result implies that the bubbling call is a highly effective mimicry of the syllable of the sparrowhawk sound regardless of the syllable rate.

Unlike signals aimed at conspecifics, such as mating signals, alarm calls benefit from similarity across sympatric species without compromising their effectiveness (Marler 1957). This similarity often stems from the retention of characteristics from a common ancestor (de Kort and ten Cate 2001) or convergent acoustic structure (Marler 1955; Fallow et al. 2011). For example, Jurisevic and Sanderson (1994) analyzed alarm calls from 30 bird species in Australia and reported that 29 species emitted broad-band alarm calls of a single-syllable type, partially with noise or clear harmonics. In a study on the recognition of alarm calls in the superb fairywren (Malurus cyaneus), Fallow et al. (2013) suggested that peak and modulation frequencies affect the recognition of unfamiliar heterospecific alarm calls by the superb fairywren. They assembled natural heterospecific calls from 3 avian families in Australia and systematically manipulated the peak and modulation frequencies of these calls. Through playback experiments, they reported that even if other acoustic parameters differed, when the peak frequency of the alarm call playback was similar to that of conspecific alarm calls, the reaction elicited in the test subjects was similar to those elicited by conspecific alarm call playback. Thus, single syllables, peak frequency, modulation frequency, and potentially other sound parameters might all be directions in the convergent evolution of alarm calls. In our study area, the common cuckoo coexists with hosts such as the Oriental reed warbler and the vinous-throated parrotbill (S. webbiana), and both the alarm calls of these hosts share the characteristic of rapid repetition of a single syllable (Jiang et al. 2022). Therefore, if the call of the common cuckoo mimics this characteristic of alarm calls, it can effectively influence both hosts within the same habitat via a single call type.

However, our results do not support the hypothesis that we proposed: The common cuckoo mimics alarm calls through a rapid syllable rate. Nonetheless, we cannot rule out the possibility that other acoustic parameters in the bubbling call, possibly convergently evolved with the alarm calls of the Oriental reed warbler, could elicit behavioral responses from the host. Parameters such as bandwidth and harmonics, which have been proven effective for mimicking alarm calls (Jurisevic and Sanderson 1994), suggest avenues for further research. This could include synthesizing artificial calls by modifying the bandwidth and harmonics of female cuckoo calls to test whether other parameters can also function as alarm calls commonly recognized by the Oriental reed warbler.

Our results indicate that the bubbling call of the common cuckoo mimics that of a sparrowhawk. The common cuckoo leverages this mimicry to exploit the host’s recognition of predatory sounds, effectively distracting the host’s attention from the clutch during parasitism to its advantage. Mimicry signals target specific perceptual biases to manipulate the host’s perception, relying on the host’s sensitivity to specific acoustic structures, even though the hosts do not perceive mimicry as similar to their signals (Dalziell et al. 2015). Although the bubbling call of female common cuckoos may not perfectly mimic that of the sparrowhawk, mimicking certain key parameters can still be effective, potentially more effective than the sound of the mimicked entity itself. Specifically, our study supports the opinion of the aforementioned perceptual biases because cuckoo hosts did not decrease but increased their reaction to the bubbling call even when its syllable rate was reduced.

Additionally, although the bubbling call of female cuckoos is renowned for mimicking sparrowhawks, its role extends beyond parasitic adaptation. Bubbling calls are crucial for intraspecific communication (Moskát and Hauber 2019; Xia et al. 2019). Moskát and Hauber (2023) manipulated the harmonics of sparrowhawk calls and demonstrated that the bubbling call mimics only the fundamental frequency, not the harmonics, of the sparrowhawk call. They also suggested that high mimicry between species can diminish the effectiveness of intraspecific acoustic communication. Therefore, the imperfect simulation of the call of sparrowhawks by the bubbling call might enhance intraspecific communication. Although this study does not support the hypothesis that the bubbling call of a female cuckoo mimics the acoustic parameters of convergently evolved alarm calls, it provides precise evidence that the bubbling call mimics that of hawks.

In summary, in this study, we proposed a new hypothesis regarding the female cuckoo call. Although evidence supporting this hypothesis was not found, our results reveal that the mimicry of the sparrowhawk call relies on the syllable type rather than the rate. Additionally, the reduced syllable rate in female cuckoo calls was found to be as effective as the normal rate. This finding supports the theory of imperfect mimicry, suggesting that as long as key acoustic parameters are mimicked, mimicry can achieve effects similar to those of the model’s sound. Therefore, although our results did not support the new hypothesis we proposed, they reveal the function mechanism of sparrowhawk mimicry in the bubbling calls (i.e., the syllable type rather than the rate) and provide evidence to support the theory of imperfect mimicry.

Supplementary Material

Supplementary material can be found at https://academic.oup.com/cz.

Authors' Contributions

CY conceived this study, designed the experiments, and analyzed the data. HW, XC, JW, and LM conducted the experiments. CY and HW wrote the draft manuscript, and CY improved the manuscript. All authors read and approved the final manuscript.

Funding

This research was funded by the Education Department of Hainan Province (no. HnjgY 2022-12) and the National Natural Science Foundation of China (no. 32260127).

Conflict of Interest

The authors declare that they have no conflicts of interest.

Ethical Statement

The experiments reported here comply with the current laws of China. Fieldwork was conducted under permission from Zhalong National Nature Reserves, and Baiyangdian National Wetland Park, P.R. China. Experimental procedures were in agreement with the Ethical Evaluation Group for Animal Behavior Study (permit no. EEGABS-006).