Prey species can increase their chances of detecting predators before it is too late by visually monitoring their surroundings for signs of danger (Beauchamp, 2015a). Many features of the habitat, however, can obstruct the field of view during vigilance. Examples include stubble in a field (Murton and Isaacson, 1962), rocks on the shore (Metcalfe, 1984), and vegetation cover (Underwood, 1982). Visual obstruction can thus interfere with vigilance and increase predation risk (Lima, 1987). Observational studies with naturally-occurring visual obstacles and experimental studies, in which the field of view is altered by the researchers, have typically documented an increase in vigilance with visual obstruction (Lima, 1992; Arenz and Leger, 1997; Lima and Bednekoff, 1999; Blumstein and Daniel, 2003; Whittingham et al., 2004; Butler et al., 2005; Ebensperger and Hurtado, 2005; Fernández-Juricic et al., 2005; Devereux et al., 2006; Bednekoff and Blumstein, 2009; Iribarren and Kotler, 2012) although there are exceptions (Lima, 1987; Foster-McDonald et al., 2006; Hannon et al., 2006; Hall et al., 2013).

Visual obstruction is thought to increase vigilance by decreasing the ability to detect predators. However, other researchers have argued that visual obstruction can also affect the ability to monitor nearby companions (Elgar et al., 1984; Duncan and Jenkins, 1998; Harkin et al., 2000; Fernández-Juricic et al., 2005). For instance, visual obstacles might temporally hide companions from view and also give the impression that neighbors are further away. The perception of a smaller group with more distant neighbors would be sufficient to increase vigilance on its own. In addition, subtle signs of alert, which are important to communicate danger across the group, might also be missed more often with visual obstruction (Harkin et al., 2000). Teasing apart the contribution of these two factors, namely, the ability to monitor predators or neighbors, requires experiments. In one type of experiment, the ability to monitor neighbors can be maintained constant so that visual obstruction only prevents the detection of predators. In another type of experiment, the ability to detect predators can be maintained constant so that visual obstacles only prevent the detection of neighbors. Few such experiments have been carried out thus far for group-foraging prey species (Fernández-Juricic et al., 2005; Pomeroy et al., 2006; Beauchamp, 2015b).

Visual obstruction could also influence other aspects of vigilance. For prey species that live in groups, vigilance was first thought to be carried out independently by each group member (Pulliam, 1973). However, recent models suggest that group members pay attention to the vigilance of their neighbors and might actually copy their vigilance. One model proposed that the level of vigilance maintained by neighbors can be seen as public information about current predation risk (Sirot, 2006). Such information could be used by others to adjust their vigilance according to the prevailing perception of predation risk in the group. It has also been suggested that by synchronizing their vigilance bouts, individuals could reduce the risk of being left behind during an attack, which could be dangerous if predators preferentially target laggards (Sirot and Touzalin, 2009). Vigilance synchronization has been documented in several species of birds and mammals, suggesting that monitoring neighbors is common in prey species (Fernández-Juricic et al., 2004; Pays et al., 2007a,b, 2009, 2012; Beauchamp, 2009; Ge et al., 2011; Michelena and Deneubourg, 2011; Öst and Tierala, 2011; Butler et al., 2016; Podgórski et al., 2016).

How animals synchronize their vigilance is not well known. The above models suggest that visual cues are important although auditory cues, such as calls or sounds from neighbors, could also be used (Radford and Ridley, 2007; Pereira et al., 2012). Visual obstruction would thus be predicted to alter the ability to synchronize vigilance if visual cues are important when copying vigilance. Synchronization could also be achieved by independent responses by each group member to external cues such as noises or movement in the vegetation, in which case synchronization serves no purpose (Ruxton and Roberts, 1999; Hoppitt et al., 2007). If such external cues rather than those provided by neighbors are predominant, visual obstruction is not expected to have a strong impact on the synchronization of vigilance.

To address these issues, I performed an experiment in which the ability to monitor neighbors was prevented by visual obstruction while the ability to monitor predators was maintained for some group members but not for others. I predicted that the level of vigilance would increase with visual obstruction. If visual obstruction restricts the ability to monitor neighbors as well as predators, I predicted that the increase in vigilance with visual obstruction would be higher for individuals prevented from monitoring both neighbors and predators rather than neighbors only. If visual cues generated by the behavior of neighbors drive synchronization of vigilance, I predicted less synchronization with visual obstruction.

Obstruction of the field of view by the wood panel increased time spent vigilant [F_{(1, 13)} = 6.9, p = 0.02; Figure 2]. Time spent vigilant by the inner bird of a pair did not differ from that of the outer bird during obstructed trials [back-transformed means (95% C.I.); inner: 32.4% (19.2%, 47.1%), outer: 31.1% (18.2%, 45.8%); F_{(1, 6)} = 0.04, p = 0.85].

With respect to the synchronization of vigilance, mean latency to become vigilant following the initiation of a vigilance bout by the companion was significantly shorter than predicted by the null hypothesis of random vigilance by each bird in unobstructed trials (t = 2.5, p = 0.015; Figure 3) but not in obstructed trials (t = 1.3, p = 0.19; Figure 3).

Visual obstruction caused an increase in vigilance. In previous experiments with birds and mammals, visual obstruction also typically increased the allocation of time to vigilance as noted earlier. Why visual obstruction causes an increase in vigilance, however, is less well known. Vigilance can increase because visual obstruction reduces the ability to monitor the surroundings and/or because neighbors are more difficult to monitor. In obstructed trials, birds in the inner position could not monitor their neighbors and much of the surroundings for lurking predators. By contrast, the outer birds were only prevented from monitoring their neighbors. Nevertheless, time spent vigilant did not differ for inner and outer birds, which suggests that the main reason why vigilance increased with visual obstruction here was related to the inability to monitor neighbors. Earlier work proposed that social monitoring might be hindered by visual obstruction (Elgar et al., 1984; Duncan and Jenkins, 1998; Harkin et al., 2000; Fernández-Juricic et al., 2005), and this experiment provides experimental evidence in support of this hypothesis.

The finding that vigilance of experimental birds did not appear to respond to threats in the surrounding environment does not mean that visual monitoring of the surroundings is not important. In a natural experiment in which visual obstruction prevented visual monitoring of the surroundings but not visual monitoring of neighbors, vigilance was indeed higher with visual obstruction (Beauchamp, 2015b). Other studies with solitary foragers also suggest that visual obstruction can increase predation risk by blocking the view of the surroundings (Arenz and Leger, 1997; Bednekoff and Blumstein, 2009; Iribarren and Kotler, 2012; Wheeler and Hik, 2014). For social species, my results indicate that visual monitoring of neighbors can be an important component of the increase in vigilance with visual obstruction.

Visual obstruction decreased the synchronization of vigilance in pairs of birds feeding close to one another. Synchronization of vigilance was documented in unobstructed trials using the timing of vigilance bouts between pairs of birds in visual and auditory contact. Vigilance synchronization has been documented in several species of birds and mammals as noted earlier. Nevertheless, how such synchronization is achieved is less well known. My results indicate that visual cues are important in triggering synchronization because the level of synchronization decreased in trials with visual obstruction. Alternatively, synchronization could occur through independent responses to external events (such as noises or movement in the vegetation) (Ruxton and Roberts, 1999; Hoppitt et al., 2007). This hypothesis can be ruled out because synchronization did not persist to the same level when birds could not see one another but could still respond to the same external stimuli. Use of auditory cues from a neighbor could also induce synchronization. Work with rats suggests that lack of feeding noises from a neighbor, which can be indicative of higher vigilance, is sufficient to induce vigilance in a visually isolated companion (Pereira et al., 2012). Vocal cues such as calls can also trigger vigilance adjustments in group members that are visually isolated (Radford and Ridley, 2007). In this experiment, I rarely recorded vocalizations during trials. Other auditory cues, if any, were apparently insufficient to trigger synchronization of vigilance in obstructed trials.

Synchronization of vigilance can be used to get information from neighbors about current predation risk and also to reduce the chances of being targeted during an attack (Sirot, 2006; Sirot and Touzalin, 2009). Visual obstruction could thus decrease the amount of public information about predation risk and the risk of being targeted by predators. It has long been thought that visual obstruction can decrease the ability to detect predators. In species where public information about predation risk is important and/or where predators can target laggards, visual obstruction could also increase predation risk even if visual obstruction did not prevent the detection of predators. This could be further tested in prey species that can detect predators easily because, say, they tend to attack from below or above, but where visual obstacles prevent monitoring of neighbors.

This experiment shows that the ability to visually monitor neighbors is an important determinant of the level vigilance maintained by a prey species, and also an important factor in the synchronization of vigilance between neighbors. Future work with other species, especially in the field where animals are exposed to direct threats from predators, could determine the extent to which visual cues from neighbors influence vigilance patterns.

The author confirms being the sole contributor of this work and approved it for publication.