M. Broom et Gd. Ruxton, Modelling responses in vigilance rates to arrivals to and departures from a group of foragers, IMA J MATH, 15(4), 1998, pp. 387-400
Citations number
18
Categorie Soggetti
Multidisciplinary
Journal title
IMA JOURNAL OF MATHEMATICS APPLIED IN MEDICINE AND BIOLOGY
Feeding birds must balance food gathering against predation risk. Group for
aging is generally considered as a strategy for optimizing this trade-off.
Previous modelling studies assumed that group sizes were static and that al
l group members are informed of an imminent predator attack if one of them
detects it. These models implicitly assumed that birds could estimate group
size, and adopt a fixed rate of anti-predator vigilance scanning. Recent e
mpirical results suggest that group sizes are generally dynamic rather than
static and the group members are often unaware of another's detection of i
mminent attack. It has also been observed that vigilance rates are not stat
ic but change after the arrival or departure of another. Here, we present a
model which allows feeding-group size (and individuals' vigilance rates) t
o vary dynamically, and investigate the implications which this has for the
optimum trade-off between foraging and avoiding predation. We find that ne
wly arrived birds should generally be the most vigilant and that vigilance
rates should decrease after the arrival of another into the feeding group b
ut increase after a departure. Vigilance rates should increase as the cost
of predator attack increases or if the reward rate from foraging decreases.
Vigilance should increase if predator attacks are more common but decrease
if predators require a longer time undetected to approach the feeding grou
p. In common with many experimental studies, we observe that vigilance rate
s decrease as the average number of birds feeding together increases. Hence
, the main conclusion of previous works (that foraging in groups is an effe
ctive strategy for balancing the conflicting pressures of foraging and avoi
ding predation) is obtained by our model, despite relaxing several previous
ly used assumptions. Finally, we discuss some of the open questions related
to group vigilance and how extensions to our modelling framework might be
used to address these.