Wild American crows gather around their dead to learn about danger
Highlights
- •
- We examined whether crows learn places and predators associated with conspecific death.
- •
- Crows took longer to approach food in areas associated with conspecific death.
- •
- Crows scolded humans previously seen near a dead crow, a hawk and a hawk with a dead crow.
- •
- A hawk with a dead crow elicited the strongest immediate antipredator behaviours.
- •
- Dead pigeons did not elicit similar antipredator responses in crows or pigeons.
While
a growing number of animals demonstrate avoidance of areas associated
with conspecific death, the extent to which wild populations may use
these experiences to learn about novel predators remains unclear. Here
we demonstrate with experiments that wild American crows, Corvus brachyrhynchos,
respond to dead conspecifics by mobbing, increasing the time to
approach food in areas associated with these events, and learning new
predators based on their proximity to dead crows and hawks. Avoidance of
either dead conspecifics or areas associated with them is not shared by
another urban bird, the rock pigeon, Columba livia. Crows
mobbed and increased the time to approach food over the next 72 h after
observing novel humans paired with a dead crow, a red-tailed hawk, Buteo jamaicensis,
or a hawk with a dead crow. The sight of a dead pigeon did not elicit
these responses. These findings suggest that, for crows, dead
conspecifics, but not dead heterospecifics, represent a salient danger
akin to the observation of a predator. On the day the stimulus was
presented, the number of trials that resulted in mobbing and avoidance
of the food was strongest when crows were presented a hawk with a dead
crow. In addition, we demonstrate that crows use the proximity of a
human to predators, to dead conspecifics and to predators with dead
conspecifics as cues to learn to recognize and subsequently scold the
associated human after only one training event, and that this
association can last 6 weeks. Together, our results support previous
findings that crows learn places associated with conspecific death, and
further demonstrate that crows can learn and remember people who appear
complicit in these events.
Keywords
- American crow;
- Columba livia;
- conditioned predator learning;
- Corvus brachyrhynchos;
- fear extinction;
- necrophobia;
- risk assessment;
- rock pigeon
Animals can reduce their risk of predation by attending to cues in the environment such as predator odours (Eichholz, Dassow, Stafford, & Weatherhead, 2012), observations of predators (Cooper, 2005) and observations of predators with prey (Conover and Perito, 1981 and Kruuk, 1976). Risk may also be communicated through conspecific and heterospecific alarms such as vocalizations (Shriner, 1998 and Templeton et al., 2005) and olfactory cues (Ferrari, Wisenden, & Chivers, 2010). In fish, these cues trigger area avoidance and increased shelter activity (Lawrence & Smith, 1989). Less is known, however, about the extent to which animals use visual remains of conspecifics as evidence of predation risk.
Humans place substantial significance on conspecific death (Tattersall, 1998), whereas few animals have been reported to show more than a passing interest. Black-billed magpies, Pica hudsonia ( Miller & Bringham, 1998), western scrub-jays, Aphelocoma californica ( Iglesias, McElreath, & Patricelli, 2012), chimpanzees, Pan troglodytes ( Stewart, Piel, & O'Malley, 2012), African elephants, Loxodonta africana ( Douglas-Hamilton, Bhalla, Wittemyer, & Vollrath, 2006), and bottlenose dolphins, Tursiops aduncus ( Dudzinski et al., 2003),
are among those that congregate around or touch and groom dead
conspecifics. The evolutionary basis for these behaviours in mammals
remains unclear (McComb, Baker, & Moss, 2006).
Emerging evidence suggests that, for some birds, these interactions are
used to assess danger and trigger antipredator behaviours.
Wild common ravens, Corvus corax, and American crows, Corvus brachyrhynchos, mob in response to distress call playbacks paired with a dead conspecific and avoid areas where they are present ( Avery et al., 2008 and Peterson and Colwell, 2014). Western scrub-jays also mob in response to dead conspecifics, and they do so in the absence of artificial distress calls (Iglesias et al., 2012).
Furthermore, following carcass removal, scrub-jays show reduced feeding
activity in the area for 24 h. A similar effect was seen when
scrub-jays were presented an upright-mounted great-horned owl, Bubo virginianus,
suggesting that dead conspecifics are used as indirect evidence of
predators. These behaviours were not observed in response to jay-like,
novel objects or upright-mounted scrub-jays. A subsequent study showed
that mobbing and area avoidance are also extended to sympatric and
allopatric jay-sized heterospecifics ( Iglesias, Stetkevitch, & Patricelli, 2014).
Together, these studies suggest that dead conspecifics, and certain
heterospecifics, can elicit learning and avoidance of places associated
with death. What remains unclear, however, is whether the presence of
dead conspecifics is an effective trigger to induce conditional learning
of a novel predator in wild populations.
Through
classical conditioning, naïve animals can learn about novel predators
through exposure to an unfamiliar predator in association with
conspecific alarm cues, such as odours or vocalizations, or by watching
responses of knowledgeable individuals (Griffin, Blumstein, & Evans, 2000). These stimuli subsequently prompt antipredator behaviours such as mobbing (Curio, Ernst, & Vieth, 1978) and site avoidance, even at the cost of avoiding high-quality or abundant food (Lima & Dill, 1990).
Fear can be extinguished, however, through repeated exposure to the
conditioned stimulus without reinforcing its predictive value of the
unconditioned stimulus (Myers & Davis, 2007). While brain-imaging studies suggest that captive American crows learn to recognize people associated with dead crows (Cross et al., 2013),
it remains untested whether wild animals can use dead conspecifics to
infer novel predators in the absence of alarm call playbacks.
Understanding this potential, and its vulnerability to extinction, could
inform management for both the reintroduction of naïve individuals, and
as a means to create more effective ‘scarecrows’.
Here,
we add to previous studies by testing whether, as in scrub-jays, wild
crows reduce feeding activity after only a brief exposure to a predator,
to a dead conspecific or to a dead, similarly sized heterospecifc.
Furthermore, we expand by asking whether crows' interest in dead
conspecifics facilitates learning of novel, threatening people and
whether this knowledge is resistant to extinction. Lastly, we also
determine whether another urban bird, the rock pigeon, Columba livia,
uses dead conspecifics to assess risk. To test danger learning, we
conducted three experiments on wild crows. In experiment 1, we examined
(1) whether the sight of a dead conspecific is sufficient to elicit
alarm calling and recruitment, or whether the presence of an
unconditioned predator is also necessary, (2) whether crows learn areas
associated with these dangers and subsequently avoid them, (3) whether
crows use dead conspecifics to identify novel predators and, if so, how
this process compares to conditioned learning when novel predators are
paired with unconditioned stimuli (hawks), and (4) whether fear
extinction can be achieved with a minimum of three additional exposures.
For experiment 2, we determined whether a dead conspecific is a more
salient source of dangerous information than a similarly sized, dead
heterospecific. In experiment 3, we compared responses of rock pigeons
and crows to dead conspecifics.
