[dataset] Host-specific vocal similarity in fledgling Levaillant’s Cuckoos and babblers in West Africa
Description
Audio and video recordings of Levaillant’s cuckoo and host fledglings reared by Blackcap Babblers (Turdoides reinwardtii) and Brown Babblers (Turdoides plebejus) in The Gambia were collected or curated by Clive R. Barlow between 2006 and 2025. Recordings of one Arrow-marked Babbler cuckoo and two host fledglings associated were provided by Derek Engelbrecht and Isaac Kilusu, some previously published (Derek Engelbrecht/Macaulay Library, Cornell Lab: ML617040028, ML617040040, ML617040041, ML617040048; Isaac Kilusu: XC819163, www.xeno-canto.org/819163; XC819164, www.xeno-canto.org/819164). We also included a unique recording of a Levaillant’s cuckoo fledgling begging and being fed by a Kurrichane Thrush (Turdus libonyana) in South Africa (Tony Archer, XC418749, www.xeno-canto.org/418749). No recordings of Kurrichane Thrush fledgling calls were available, but we included fledgling begging calls of the closely related African Thrush (Turdus pelios) recorded by CRB in The Gambia. This study investigated the presence of vocal mimicry in Levaillant’s cuckoos. The dataset contains acoustic variables automatically extracted with Avisoft SASLab (see “Steps to reproduce”), including call duration, and the mean and standard deviation of frequency bandwidth, entropy (energy dispersion across frequencies—low in pure tones, high in broadband calls), and the first and third frequency quartiles of the energy distribution (robust indicators of low and high frequency limits). Acoustic analyses and classification methods revealed that cuckoos reared by their primary host, the Brown Babbler (Turdoides plebejus), produced contact and food-related calls closely resembling those of host fledglings. By contrast, cuckoos raised by secondary hosts, including Blackcap Babblers (T. reinwardtii), showed little or no vocal similarity.
Files
Steps to reproduce
All original audio files were annotated by the recordists to identify call types and behavioural context. A descriptive list of recordings is provided in the “List of recordings” Excel file. Vocalisations were analysed with Avisoft-SASLab Pro 5.3.2-36. For each recording, continuous annotated spectrograms were generated to identify callers, describe behavioural context, and classify vocal elements. Each element was defined by its distinctive amplitude envelope; where two successive elements formed a single vocalisation, they were analysed separately. Fledgling calls were classified as food-transfer calls, given during food delivery by adults, or begging contact calls, given in the long intervals between feedings and often accompanied by wing-quivering or crouching. Cuckoos reared by Blackcap Babblers produced two distinct types of contact calls that were categorized separately. To minimise background noise, a 0.6 kHz high-pass filter was applied. Spectrograms were generated with a 512-point FFT, FlatTop window, 100 % frame size, 50 % overlap, and a bandwidth of 162 Hz, yielding 11.6 ms temporal and 43 Hz frequency resolution. For each element, non-overlapping extraneous sounds were manually removed and amplitude thresholds adjusted to optimise signal-to-noise ratio. Nine acoustic parameters were automatically extracted for each element. As most vocalisations were harsh or multiband calls with low-frequency modulation, spectral measurements were based on the mean spectrum of the entire element as described in Avisoft-SASLab manual.