The presence of an avian co-forager reduces vigilance in a cooperative mammal
L. L. Sharpe, A. S. Joustra and M. I. Cherry
MATERIALS AND METHODS
The study was undertaken in the lowveld of north eastern South Africa, in dry, wooded savannah that was interspersed with eroded granite outcrops (koppies). The dominant tree species were velvet corkwood (Commiphora mollis), knob thorn (Acacia nigrescens) and marula (Sclerocarya birrea), above an under storey of bush willow (Combretum spp.), thorn bush (Acacia spp.) and raisin bush (Grewia spp.). Rainfall was highly seasonal with wet summers (October to March) and dry winters, when the vegetation progressively lost its leaves.
Dwarf mongooses (Helogale parvula) often forage in the company of birds, and frequently join mixed-species foraging flocks. At the study site these flocks usually included fork-tailed drongos (Dicrurus adsimilis), yellow-billed and red-billed hornbills (Tockus leucomelas and T. erythrorhynchus), black-backed puff-backs (Dryoscopus cubla), southern black tits (Parus niger), blue waxbills (Uraeginthus angolensis), green-winged pytilias (Pytilia melba) and firefinches (Lagonosticta spp.). The mongooses were accompanied by birds for approximately 50% of their foraging time (Sharpe, unpublished data).
Fork-tailed drongos, which are medium sized passerines (45-65g; Ridley et al. 2007), also accompanied the mongooses when they were alone, and both species appeared to actively seek out and maintain the association, although this has yet to be confirmed experimentally. Fork-tailed drongos and dwarf mongooses are both territorial species, and the home range of each mongoose group overlapped the territories of two or three pairs of drongos. Normally one or two drongos (but up to six; i.e. a pair and its offspring) accompanied the foraging mongooses. The commensal association of fork-tailed drongos with ground foraging species was found to be widespread in a study conducted in Botswana (Herremans & Herremans-Tonnoeyr 1997). Drongos were observed to associate with both livestock and 37 species of wildlife, capturing insects that these animals flushed. Although fork-tailed drongos are known to undertake kleptoparasitism (Ridley & Raihani 2007), this was rare, occurring only when the abundance of flying insects was low due to drought or exceptionally cold weather (Herremans & Herremans-Tonnoeyr 1997). We occasionally observed drongos attempting to steal prey from the mongooses (fluttering above an individual while uttering loud alarm calls), but the birds were rarely successful, and primarily fed on airborne grasshoppers flushed beyond reach by the mongooses.
Dwarf mongooses are vulnerable to a wide array of predators due to their small size (250-300g; Rood 1986). At the study site, the most common terrestrial predators were the African wildcat (Felis lybica), black-backed jackal (Canis mesomelas), side-striped jackal (Canis adustus), African civet (Civettictis civetta), honey badger (Mellivora capensis), serval (Felis serval), slender mongoose (Galerella sanguinea), rock monitor (Varanus exanthematicus), African rock python (Python sebae), puff adder (Bitis arietans), Mozambique spitting cobra (Naja mossambica), snouted cobra (Naja haje) and black mamba (Dendroaspis polylepis). Frequenting the area were 24 species of diurnal raptor known to prey on mammals the size of a dwarf mongoose. The most important of these aerial predators (based on the severity of the mongooses’ response to their presence) were the African hawk eagle (Hieraaetus spilogaster), brown snake eagle (Circaetus cinereus), black-breasted snake eagle (Circaetus pectoralis), steppe eagle (Aquila nipalensis), tawny eagle (Aquila rapax) and dark chanting goshawk (Melierax metabates). Although fork-tailed drongos are not susceptible to terrestrial predators, they are considered vulnerable to all but four of the mongoose’s aerial predators, plus four additional species that are bird specialists.
All mongooses in the study population (120 animals) were individually ID marked using Garnier Nutrisse blonde hair dye, applied with a long-handled paintbrush while the group was sunning.
For all data collection, we considered a drongo to be present if one or more fork-tailed drongos were located within 10m of any of the group’s members (group dispersion was normally around 15m). Ten metres was chosen because it was the distance that an observer could accurately assess for the presence of birds. However, because the drongos accompanied the mongooses to hawk the insects that they flushed (Herremans & Herremans-Tonnoeyr 1997), they always positioned themselves directly above the foraging mongooses (usually in the centre of the group) and were only ever more than 10m away (but within auditory range) when briefly transiting the area.
To document vigilance in foraging mongooses, we recorded the focal individual’s vigilance behaviour on a handheld computer, Psion organiser II, model LZ64. We paused data collection if the focal individual stopped foraging or ceased to be visible, and suspended it for 10 minutes after any predator alarm severe enough to disrupt foraging for more than 15 seconds. To ensure a homogenous sample, we selected eight adult (>12 months old), non-alpha mongooses, because youngsters and alpha individuals appeared likely to differ in their rates of vigilance (e.g. alphas undertake sentinel duty at twice the rate of subordinates; Sharpe unpublished data).
To simulate the presence of a fork-tailed drongo, we chose to use playbacks of the drongo’s non-alarm vocalisations because this species is highly vocal and readily detectable by its almost continuous calls. The species’ non-alarm vocalisations include a single ‘twank’ call, repeated every 4-5 seconds and a complex song consisting of a ‘loud jumble of strident, twanging, rasping and creaking sounds’ which also incorporates the mimicked calls of other species (Hockey et al. 2005). The function of these calls has yet to be established, but in the greater racket-tailed drongo (Dicrurus paradiseus) songs containing mimicry are more effective at attracting the bird species involved in mixed-species foraging flocks (Goodale & Kotagama 2006). We used recordings of the ‘twank’ call (figure 1a) in all experiments.
All playbacks were recorded in the field using a Marantz Professional portable recorder (model PMD670) and a Sennheiser microphone (model ME66/K6). The amplitude of calls was measured using a Velleman sound level meter (model DVM805; range 30-100 db; 1.5 db accuracy) both at the time of recording (to ascertain normal levels) and at playback. The recordings (ranging in duration from 12 secs to 2.5 mins) were edited, using Adobe Audition 2.0, to remove background sounds and then were replicated to produce a 10-min playback.
When conducting the experiments, the observer, carrying the portable Marantz recorder, followed the test individual at a distance of 2-3 m. All bird calls were played at an amplitude of 55 db (when measured two meters from the recorder) which was within their normal range (40-65 decibels). Calls were played continuously for a period of ten minutes, although both data collection and playbacks were paused if the test subject stopped foraging or ceased to be visible, or any birds came within visual or auditory range. Similarly, both playbacks and data collection were suspended for 10 minutes after a predator alarm severe enough to disrupt foraging for more than 15 seconds. Playbacks and data collection were resumed after such interruptions and continued until 10 minutes of vigilance data were obtained for each playback condition. We randomised the order in which the recordings (drongo versus control) were presented for each experiment. At least four days were allowed to elapse between experiments at the same group (mean 4.7 days).
Because some test subjects were drawn from the same group, the independence of the trials may have been compromised if group members influenced one another’s behaviour. However, due to the mongooses’ small size, the density of their habitat and their tendency to forage 2-3 meters from their nearest neighbour, there was virtually no visual contact between foraging group members (hence the need for continuous ‘contact calls’) and individuals were unable to monitor one another’s vigilance behaviour. An analysis of variance confirmed that group membership did not affect an individual’s response to the playbacks, with no difference in rate of vigilance scanning between groups for either the control or test conditions (ANOVA: control: f 1,6 = 0.041, p = 0.846; drongo: f 1,6 < 0.001, p = 0.999).
First playback experiment.
The first playback experiment sampled eight test individuals from two groups and was undertaken in September 2007, In this experiment we used a single exemplar of drongo vocalisations for all eight trials. Hence we tested the mongooses’ response to one individual drongo rather than a sample of the species, reducing the generalizablity of the findings. Theoretically, it is also possible that the mongooses were responding to some other aspect of this particular playback (e.g. an undetected background noise or recording artefact) but it seems unlikely that any such defect would cause a reduction in vigilance. This was verified by the finding that vigilance rates during drongo playbacks did not differ between the first and second experiments (see below) (1.95+/-0.21 and 1.65+/-0.17 scans per min, respectively; t-test: t14 = 1.10, p = 0.289).
To control for the potential effect of the experimental procedure, we chose to use playbacks of the ‘contact calls’ of mongooses from the test subject’s own group as the control condition. Individual dwarf mongooses give ‘contact calls’ whenever the group is foraging or travelling, to maintain contact with one another and ensure the group remains unified (Rasa 1987). These high-pitched ‘peeps’ form a constant, ubiquitous background noise to a foraging mongoose group, and as such were considered an appropriate control. Frequency of contact calls does not reflect the number of mongooses present (group composition does not alter in the short term) but appears to vary in response to such factors as group dispersion, habitat density and rate of travel. At each of the two mongoose groups used in the first experiment, we made a two-minute recording of contact calls while the animals were foraging. These recordings featured the vocalisations of most of the group’s members (collectively averaging one call every 4 seconds). Each of these two-minute recordings was replicated to create a 10-minute playback for use, as the control playback, at the group at which it was recorded.
The eight test subjects used for the first experiment were drawn from two mongoose groups (five individuals from one group and three from another). Although the trials were undertaken on different days, and the recordings were played specifically to the test subject, other group members foraging nearby would also have heard the playbacks. Thus six of the eight test subjects received prior exposure to the playbacks (i.e. the second subject tested in each group may have been exposed to 10 minutes of drongo playbacks prior to its own trial, the third animal 20 minutes, the fourth 30 minutes and the fifth animal tested could potentially have heard 40 minutes of drongo playbacks). However, this prior exposure did not appear to affect the test subjects’ response to the playbacks, and there was no relationship between extent of prior exposure and the degree to which individuals reduced their vigilance scanning in response to the drongo playback (Pearsons correlation: r7 = -0.011, p=0.979).
Second playback experiment.
The second playback experiment sampled eight individuals, from four groups, and was undertaken in Oct-Nov 2009. Because we wished to demonstrate that the mongooses were responding specifically to the drongos (rather than to the presence of any bird) we chose to use recordings of the non-alarm vocalisations of the white-bellied sunbird (Cinnyris venusta) for the control playback (figure 1b). This highly vocal species is a common resident at the study site but it does not forage with the mongooses nor join mixed-species foraging flocks.
We recorded non-alarm vocalisations from eight different drongos and eight different sunbirds (either known individuals, or birds living >3km apart) and created 16 unique exemplars for use in the eight trials (no exemplar was used more than once).
Goodale, E. & Kotagama, S. W. 2006. Vocal mimicry by a passerine bird attracts other species involved in mixed-species flock. Anim. Behav. 72, 471-477.
Hockey, P. A. R., Dean, W. R. J. & Ryan, P. G. 2005 eds. Robert’s birds of Southern Africa. 7th ed. Capetown: The Trustees of the John Voelcker Bird Fund.
Rood, J. P. 1986. Ecology and social evolution in the mongooses. In: Rubenstein, D. I. & Wrangham, R. W. eds. Ecological Aspects of Social Evolution. Princeton Univ. Press, Princeton. pp. 131-152.
Figure one. Spectrogram of (a) the ‘twank’ vocalisation of a fork-tailed drongo (Dicrurus adsimilis), as used in both the first and second playback experiments; and (b) the territorial call of a male white-bellied sunbird (Cinnyris venusta), as used in the second playback experiment as the control condition. Horizontal axes show seconds. Spectrograms prepared in Raven Lite 1.0 (brightness 55%, contrast 82%, sharpness 512).