key: cord-0272478-u1vm9kpg
authors: Feist, Franziska; Graham, Paul
title: An Ethogram Identifies Behavioural Markers of Attention to Humans in European Herring Gulls (Larus argentatus)
date: 2022-02-23
journal: bioRxiv
DOI: 10.1101/2022.02.20.481240
sha: 6d0181c9fe124f4a454fb2e1e800ffe2b0935ca3
doc_id: 272478
cord_uid: u1vm9kpg

Herring gulls are one of the few species that thrive in anthropogenic landscapes. Their longstanding history of urbanisation and familiarity with people makes them an excellent target for studies of the effects of urban living and cognition. Previous studies highlight a connection between food-stealing behaviour, success in anthropogenic areas and an increase in attention towards humans, opening up questions about the exact extent of a gull’s knowledge of human food cues and how this manifests itself in behavioural changes. To investigate these questions further, behavioural responses to human cues in a food-related context are investigated and presented in a systematic ethogram, which identifies three distinct markers of attention. Those behaviours – head turns, approaches, and the angular body position relative to the experimenter – differ significantly between control and food conditions, showing that attention towards humans in a food-related context is upregulated and this is reflected in the gull’s behaviour. In food condition trials, head turns occur more often and gulls face towards the experimenter (≤90°) with occasional approaches that are never seen in control conditions. Acoustic and behavioural human food-like cues alone were insufficient to elicit a response similar to that seen in presence of a real food item, indicating that gulls specifically pay attention to humans in foraging contexts. The results show situation-dependent attentional modulation in gulls, highlight knowledge gaps about how Herring gulls might identify food objects and provide an ethogram description of attentive behaviours that can be used in further study of attentional modulation in gulls.

The human population is increasing, and with it comes the expansion of cities and 31 urban landscapes (Lowry, Lill and Wong, 2013). Often, this negatively affects local wildlife and 32 causes species to retreat and avoid urbanised areas (Bateman and Fleming, 2012) . Few 33 manage to survive in anthropogenic landscapes, but, interestingly, those that do can be very 34 successful and in some species, urban numbers exceed the population densities found in 35 natural environments (Bateman and Fleming, 2012) . This is true for the European Herring gull 36 Based on these recent findings, we sought to continue the investigation of human-gull 54 interactions in a food-related context. To aid future investigations, observational studies were 55 conducted with the aim to create a systematic ethogram that describes distinct attentive 56 behaviours and confirms whether urban gulls pay attention to humans in general or specifically 57 in a food-related context. 58 59

Recorded video files were analysed using the Behavioural Observation Research 94

Interactive Software (BORIS) program (Friard and Gamba, 2016). General variables such as 95 weather condition (sun, cloudy, rain), average temperature in C° and wind speed in mph during 96 the time of recording were noted, along with low and high tide time and height on survey days. 97

Later analysis showed that weather conditions on recording days only affected gull presence, 98 but not their behaviour. All gulls were given a unique subject ID and their age was recorded as 99 either juvenile, adult or NA if unsure based on recorded footage. Recordings were then re-100 watched until the three selected behaviours of interest, head turns, body orientation relative to 101 the experimenter and approaches towards them, had been logged for all subjects. These three 102 distinct behavioural responses were hypothesised to be indicative of attention towards a 103 person or situation. 104

To investigate whether the three selected behaviours were indeed markers of vigilance, 107 statistical analysis focussed on comparing them across conditions. If they were indicative of 108 an increase in attention in a food-related context, it was expected to see more head turns and 109 approaches during FC trials and gulls were hypothesised to orient themselves towards the 110 experimenter. Head turns and approaches were counted, and the weighted average angular 111 body position (0-180°) of test subjects relative to the experimenter was calculated for each 112 individual to obtain one measurement per bird per trial. Furthermore, head turns were divided 113 by the time a gull was in camera view to give "head turns per gull minute", which was used for 114 analysis. Chi-square test was conducted to confirm that there was a relationship between condition type 126 and approach occurrence. Here, the complete dataset (N=211) was used as no order effects 127 were found. 128 129

The aim of this observational study was the creation of the systematic ethogram to 131 identify markers of attention. This could then be used to investigate whether gulls specifically 132 pay more attention to humans in a food-related context, and whether acoustic and behavioural 133 human food-like cues without the presence of a real food object are enough to elicit a 134 behavioural response similar to that evoked by a food item. It was found that, compared to 135 NFC trials, attentive behaviours are upregulated during FC trials but not during PC trials which 136 mimicked acoustic and behavioural food stimuli. 137

The presence of a food stimulus increases head turn counts. After pooling data 138 from trials that were not influenced by order effects, statistical analysis of the effect of condition 139 type on head turns found that during FC trials, counts were significantly increased from NFC 140 trials (NFC=53, FC=22, p=<0.0001) but counts from PC trials did not differ significantly from 141 Approaches only occurred when the experimenter signalled food. Order effects 153 did not influence approach occurrence and, therefore, the entire data set could be used 154 (NFC=111, FC=58, PC=42). Figure 1E highlights that while no birds started approaching 155 during NFC or PC trials, 14 out of 58 gulls (24.14%) did approach the experimenter during FC 7 trials, with between one and five approach instances per trial. A Pearson's Chi square test 157 confirmed that there was an association between trial type and approach occurrence 158 (Pearson's Chi-squared test, X-squared=39.56, df=2, p-value=<0.0001, N=211). 159

Together, these results highlight that gulls pay increased attention to humans in 160 possession of food (FC), but not when they only mimic acoustic and behavioural food-like cues 161 (PC). In turn, this lead to particular behavioural changes in food trials. Head turn counts 162 increased significantly, more gulls oriented themselves towards or perpendicular to the 163 experimenter, and some started approaching, which is not seen in other conditions. In contrast, 164 the acoustic and behavioural human food-like cues given during PC trials were not sufficient 165 to elicit responses similar to those recorded in the presence of a real food item. 166 167

We have identified three behaviours: head turns, orientation towards an experimenter 169 and approaches, that are upregulated during FC trials only, which shows that they are reflective 170 of increased attention towards people in possession of food, and, with that, confirms that the 171 presented stimulus is interesting to the birds. It is suggested that all three behaviours allow 172 gulls to combine attention towards a potential food source with generally higher vigilance and 173 the possibility to escape when necessary. The fact that PC trials, during which human food-174 like cues were mimicked, seemed insufficient to elicit behavioural responses similar to those 175 seen in FC trials confirms that the recorded response is specific to a foraging context. 176 177 What purpose do these behaviours serve? 178

Having identified specific markers of attention towards humans in possession of food, 179

we must ask what purpose these behaviours serve, and how this might benefit the birds. 180

First, approaches can be interpreted as interest in the experimenter, the food item and, 181 potentially, a sign for preparations for a food-stealing attempt. We report that approaches only 182 ever occur in trials where a food item was present, which supports this idea and confirms that 183 the stimulus presented in this condition was attractive to the gulls. The approach rate of the 184 Interestingly, the importance of each type of sensory cue varies between species (Laska, Freist 250

and Krause, 2007) and is associated with their foraging ecology (Rushmore, Leonhardt and 251 Drea, 2012). To reliably identify preferred food items, folivores have been shown to require 252 both visual and olfactory cues, while generalists could identify their preferred foods using either 253 cue alone, and frugivores could identify their favourites using olfactory cues alone while visual 254 cues by themselves did not suffice (Rushmore, Leonhardt and Drea, 2012) . The importance 255 of olfactory cues in food choice is further highlighted by studies of a variety of taxa. In fruit bats, 256

foraging activity is altered based on the presence of light, but olfactory cues are used to locate 257 food items. In addition, they can quickly learn to associate the presence of accessible fruits 258 with non-fruit scents, such as that of cedar wood oil (Kshitish Acharya, Roy and Krishna, 1998). 259

Sea turtles have been shown to be able to detect airborne odours and start to display foraging 260 behaviours when such odours signal the presence of food (Pfaller et al., 2020), suggesting 261 that, for them, olfactory stimuli alone are sufficient to identify foraging opportunities. Lastly, 262 models of foraging strategies in sea birds have shown that olfactory senses confer significant 263 advantages in ocean environments (Bastos et al., 2020) , and experimental studies confirm that 264 some species make use of olfactory cues to locate food (Schreiber and Burger, 2001) . 265

It seems, therefore, that both visual and olfactory cues specifically play an important 266 role in food choice in a variety of taxa, which may indicate that gulls apply a similar strategy to 267 identify food objects as such, providing a possible explanation for why human acoustic and 268 behavioural food-like cues were insufficient to elicit a response similar to that seen when a real 269 food item is present. Further investigations similar to those conducted with other species will 270 provide insight into which type of sensory cue is necessary and sufficient for gulls to identify a highlighting increased attention towards them. As such, these behaviours can be used as 280 markers of attention towards human cues, which will, in turn, be useful in future studies of 281

Herring gull behaviour and attention. It has previously been suggested that their success in 282 urban environments may be due to their cognitive capabilities and high behavioural flexibility 283 (Bateman and Fleming, 2012; Plumer, Davison and Saarma, 2014) , which is what makes them 284 an interesting subject for future studies of cognition, urbanisation and human-wildlife conflicts. 285

Aided by the description of attentive behaviours we present here, to expand our 286 knowledge in this area one could first investigate how such a response, food-dependent 287 modulation of attention, develops in urban Herring gulls. While our more complex model did 288 not find an effect of age on behavioural responses, a more detailed investigation would be 289 useful. Here, we split gulls in "adult" or "juvenile" based on the colouration of their plumage, 290 but since Herring gulls have a juvenile period of multiple years (Mullarney, Zetterström and 291

Grant, 2010), differences between first winter juveniles and older individuals may be present. 292

Investigations of potential age differences would highlight whether this specific attentional 293 modulation is a learned skill and, if so, when it develops. 294

Similarly, we may want to ask whether a gull's response would be the same when they 295 pay attention to other animals they have identified as a potential target for kleptoparasitism, or 296 whether these behaviours are human-specific. We know that gulls specifically pick more 297 successful individuals as food-stealing targets (Busniuk, Storey and Wilson, 2020), which may 298 imply that gulls have a certain degree of knowledge about their targets that allows them to 299 modulate their approach strategy accordingly. As a result, their behavioural response to a non-300 human individual in possession of food may differ from the responses we present here. 301

Comparisons of how urban gulls attend to food cues from other species and of how non-urban 302 gulls react to human food cues will provide more detailed insight into whether urbanisation and 303 frequent contact with people may have caused specific human-centred behaviours to arise in 304 urban populations. 305

Lastly, with the ethogram presented here it will be possible to investigate social 306 contagion in Herring gull groups. Attentional cues may be transferred across a group, even 307 when direct visual cues of the item are lacking. Previous investigations of vigilance in Herring 308 gull groups have reported mixed results. Some highlight that vigilance spreads throughout a 309 group, with individuals interrupting their sleep more often to scan their environment if their 310 neighbours are more vigilant (Beauchamp, 2009) . Others suggest that groups should follow 311 the "many eyes" hypothesis, according to which individuals decrease their own vigilance when 312 surrounded by highly vigilant neighbours (Roberts, 1996). Using the above identified 313 attentional markers we can investigate how vigilance travels through a group of urban gulls, 314

and how the presence of human food cues could affect this.

We present a systematic ethogram that identifies three specific markers of attention in 318 urban Herring gulls. Head turns and approaches are upregulated during trials in which a real 319 food item was present, and gulls face towards or perpendicular to the experimenter more often, 320 highlighting context-dependent attentional modulation in gulls. Individuals seem to be more 321 attentive to people in possession of food as well as their surroundings, which could be 322 indicative of preparations for a kleptoparasitic attack. Subjects may face a trade-off between 323 ideal approach conditions for a food-stealing attempt and retaining an escape route should the 324 person pose a threat to them. In contrast, such behavioural responses are not seen in 325 response to acoustic and behavioural human food-like cues without the presence of a real food 326 item, which suggests that gulls may identify objects as potential food without relying on human 327 cues, using those only when making foraging decisions about two food items. 

Oceans of stimuli: an individual-based model to assess the role of 357 olfactory cues and local enhancement in seabirds' foraging behaviour

Big city life: carnivores in urban 360 environments

Sleeping gulls monitor the vigilance behaviour of their 363 neighbours

Larus argentatus. The IUCN Red List of Threatened 365

Sight or Scent: Lemur Sensory 431

Reliance in Detecting Food Quality Varies with Feeding Ecology

Biology of Marine Birds

Kleptoparasitism in 435 gulls Laridae at an urban and a coastal foraging environment: an assessment of 436 ecological predictors

We thank Professor Pierre Nouvellet (University of Sussex) for his 352 support with the statistical modelling of the data. 353 354