Spontaneous approaches of divers by free-ranging orcas (Orcinus orca): age- and sex-differences in exploratory behaviours and visual laterality

Running comparative studies of laterality in mammals is a way to deepen our understanding of the evolution of the brain hemisphere functions. Studies on vision highlighted a possible task-sharing between hemispheres depending on the characteristics of the observers, the nature of the observed stimulus and the context of the observation, a phenomenon that could go beyond the monitoring of conspecifics. Cetaceans are predators that adapted to an aquatic habitat and display a clear crossing of fibers to the side of the brain opposite the eye of origin. Here, we analysed the interactions between humans and cetaceans when free-ranging orcas approach divers. Our study concentrated on the spontaneous exploratory behaviours of divers by orcas depending on their age and sex, and on the possible expression of a visual laterality. The results showed a significant preference for the use of the left eye but exclusively in adult females. Adult males had a more sustained attention than adult females, marked by a higher spatial proximity to divers, slower approaches and longer look durations. Adult females, probably more cautious, explored from the distance and more furtively. Our findings support a possible link between attentional/motivational states and visual laterality in mammals.

confirmed with a beluga population in managed care 40 . Nevertheless, differences are found between social and non-social monitoring. When hunting, orcas attack their prey by placing them in their right visual field, whereas no lateral bias is found when they jump out of water and fall back on their side 41 . The right hemisphere of cetaceans would thus play a predominant role for social interactions 18,39,42 , whereas in response to non-social or non-familiar stimuli, a predominance of the left hemisphere can be found 26 .
In captivity, the left eye was preferentially used by bottlenose dolphins (Tursiops truncatus) to observe a human being, familiar or not, approaching the pool 17 . The individual left eye lateralized preferences were later confirmed by studies with captive belugas and pacific white-sided dolphins (Lagenorhynchus obliquidens) 43 . In the same experimental conditions, when the human being was replaced by an object, the degree of familiarity with the observed item was important. Indeed, Blois-Heulin et al. 22 found that, while non-familiar objects are observed preferentially with the right eye, the left eye is used to look at very familiar (previously manipulated) objects. Both eyes are used indiscriminately for intermediary items, i.e. non-familiar objects which have become to some extent familiar from previous observations but never manipulated. All in all, the mother, the human (assimilated to the one that is providing care and food) and the toys in captivity are preferentially processed by the right hemisphere, whereas prey and new objects are processed by the left hemisphere. Interestingly, the degree of familiarity with the human or object observed also influenced gaze duration in several cetacean species 21,[43][44][45] . The question thus remains open as to whether or not the observer's attention and motivational state may be a determining factor. More studies are now needed, notably with free-ranging cetaceans, and the cases of human -orca spontaneous encounters is particularly interesting because the former is a potential prey for these animals.
Orcas show numerous and complex capacities of social coordination, which involve both communication and cognition. They indeed show a great diversity of innovated and learned foraging techniques and feeding cooperation strategies [46][47][48] , individual and matrilineal acoustic signatures with dialectal cultural variations [47][48][49] and a long period of maternal dependence for the single offspring 50 . Orcas are also curious animals characterized by a highly developed neophilia in captivity 51 and frequent manipulation of various objects in the wild 52 . Recent studies have described cases of play behaviour innovations in resident orcas living in the U.S. west coast 46,53 .
In several mammal species, innovations, risk-taking and exploratory behaviours vary with the age and sex of the individual (e.g. rats and primates: [54][55][56][57] ). In captive dolphins, youngsters are generally more curious than adults 58 , and 80% of the new behaviours observed in plays come from juveniles 59 . However, Baird et al. 60 found no influence of age and sex in diving rates in orcas, but male adults would dive more often during the day than female adults. The question of exploratory behaviour regarding age, sex, as far as we know, has never been studied in free-ranging orcas. Previous studies in cetaceans (e.g. dolphins and belugas) have shown that gaze duration (and possibly swimming speed and distance of approach) plays a key role during exploration of familiar and unfamiliar humans and objects 21,[43][44][45] . Although the exploratory behaviours of cetaceans have still been poorly described in the literature, particular behaviours like head and body movements, some being associated with vocalisations (notably echolocation), have also been previously described in this context 61,62 .
How humans are perceived by free-ranging orcas is a current hot topic, notably because whale-watching and swimming-with-whales ecotouristic sites are multiplying. To this day, little research has been conducted scientifically about the nature of human -orca spontaneous encounters in the wild and in the diversity of associated behaviours. Orcas' interest is expressed by behaviours like following boats and many spontaneous interactions with humans (cetologists, fishermen, tourists), including play and even actions of supposed cooperation with humans in their fishing activities 63 . In the wild, orcas are totally free of their movements and voluntarily approach divers. By using underwater videos, made in Norway, we therefore concentrated on gazes and other behaviours in the context of spontaneous approaches of divers by orcas. We studied an area where diving is organised every year. We have analysed the influence of individual's age and sex on the behaviours displayed during these approaches and tested a possible eye-preference during visual monitoring.
In most cases (86%), the "approach consisted in a progressive movement towards the divers and more rarely (14%) in a "passing by" movement (i.e. quasi-linear trajectory, constant distance with the divers). Moreover, orcas approached preferentially from the front or below (98%) and in very rare cases from the back (2%). After having excluded the rare approaches from behind, we found that all units were relatively frequent (i.e. from the right side 39%, from the left side 29%, facing 14%, just under 10%, from deep water 8%).
However, certain types of approach were more frequently observed than others in the three age-sex classes of individuals (see all statistical intra-class comparisons in Table 1). First, concerning the social context and the minimum distance of approach, we found that all individuals approached significantly more often alone than accompanied, and more often close than far. However, while adult females approached more often close than very close, no difference was found for adult males and subadults. Second, concerning visual monitoring and swimming speed, while adult females and subadults displayed more short than long gazes, no difference was found for adult males. This has to be related to some extent to the swimming speed that was more frequently slow than fast in adult males but equally slow and fast in both adult females and subadults. Third, concerning head movements, while adult males displayed more horizontal than vertical movements, both types were equally observed in adult females and subadults. Fourth, all individuals approached significantly more often silently than accompanied by whistle calls.
The swimming speed was also significantly slower in adult males than in subadults (z = 2.29, p = 0.007), while adult females' speed did not differ (when applying Bonferroni correction) with those of adult males (z = −2.001, p = 0.045) and subadults (z = 1.35, p = 0.18) (Fig. 4) In order to better understand the context of whistling, we tested whether certain variables were more often associated than others. Indeed, we found that while there were as many approaches with whistles when the orca approached alone or not (Wilcoxon matched-paired test: N = 83, Q = 1806, p = 0.743), there were significantly more approaches with whistles when the swimming speed was slow than fast (N = 83, Q = 2478, p = 0.0001).

Discussion
This first investigation of the nature of the exploratory behaviour of free-ranging orcas when approaching human divers shows two important findings. First, adult females prefer to visually monitor humans with their left eye than with their right eye, but no eye preference was found in adult males and subadults. Second, the approach behaviours differed significantly between different individuals; adult males demonstrated stronger attention, as well as potentially greater curiosity (shorter distance, slower swimming and longer gazes) than adult females and subadults, the latter being possibly more cautious.
The visual laterality in favour of the left eye is in accordance with the previous findings on social interactions in wild and captive cetaceans 18,39 and with captive dolphins when looking at humans 21 and at familiar toys 22 . It   therefore confirms the dominance of the right hemisphere in the visual exploration of some salient and familiar items by cetaceans (see also ref. 64) and allows us to support the hypothesis that the human diver is a source of interest for the free-ranging orcas habituated to their presence (for about 18 years). This hypothesis is also supported by the fact that no aggressiveness or sudden withdrawal has been noticed.
However, not all individuals preferred a left-side gazing in our study. Several non-mutually exclusive hypotheses may explain the individual variations observed. First, it is possible that laterality is more pronounced in older individuals than younger ones, due to their experiences with those unusual encounters, which would account for the absence of laterality in subadults. Indeed, the strength of laterality in animals is frequently related to the degree of familiarity with the observed item 22 . This is also supported by the fact that more adult females are in contact with divers than adult males in this area, and thus more experienced with human encounters. Over two consecutive years, the number of approaches by adult females was 4.5 times higher than the number of approaches by adult males. Second, it is possible that the emotional state of male and female adults differed in our study, as this is another factor of variation in humans and animals for both strength and direction of laterality [65][66][67] . The fact that females generally keep a greater distance than males here and only observe humans briefly could be the sign of cautiousness in females that, while approaching, would remain more vigilant. Third, it is very likely that there is a link between the direction of the laterality and the degree of attention of the observer, as it was found earlier in birds 68 . Males, being more attentive, perhaps integrate more details, thus mobilizing their two hemispheres. The left hemisphere is known to play a key role for detailed visual exploration in various animal species [69][70][71][72][73] . Fourth, the hormonal difference between males and females may also be responsible for the sex-dependent laterality observed here. A study on testosterone-treated chicks found that an effect of testosterone was greater reliance on the frontal visual field 74 , which could be a possible explanation for the absence of lateral bias due to a stronger collaboration between the two hemispheres in male orcas. Fifth, differences in cognitive abilities between males and females could explain the sex-dependent laterality found in our study. In humans, females perform better than males in various memory tasks 75,76 . This female superiority would be explained by a more effective use of the abilities of the right hemisphere for accessing memory of patterns made up of multiple items 2 .
Also in our study, adult males approached less often than adult females; this is probably due simply to their smaller number in the overall population. The fact that males approached divers at a closer distance and for longer exploration sequences is an important point of discussion. One abovementioned explanation is that males may be more curious. Unfortunately, to our knowledge, there is no such "personality type" study done with cetaceans that could confirm this hypothesis. However, sex-specific responses to human presence have been found earlier in cetaceans. Williams et al. 77 studied the habituation to human activities by orcas in Johnson Detroits (British Columbia). They found that the animal's reaction to boats changed in a sex-specific way with time. While females tended to accelerate their swimming speed and increase the angle of their diving routes, males kept their swimming speed constant with less predictable trajectory. This said, the fact that females approached less closely in our study does not mean that they did not "observe"; cetaceans may use echolocation to complete their visual exploration 36,78 . It is possible that females prioritized long-distance echolocation over short-distance eye observation. In line with this, vertical head movements, typically associated with echolocation 61 , were particularly frequent in adult females. Scheer et al. 61 also described this echolocation behaviour directed to humans at a 20-meter distance in free-ranging pilot whales (Globicephala macrorhynchus). As Blois-Heulin et al. 22 , we suppose that echolocation, associated to a binocular vision, completes the amount of information collected on objects, facilitating the construction of its representation.
The approaches described here are signs of a certain curiosity, and possibly attempts to initiate interactions with divers. The fact that the vast majority of the approaches come from the front clearly indicates that animals behave to remain visible. Scheer et al. 61 with free-ranging pilot whales and Duroyon 62 with captive bottlenose dolphins (Tursiops truncatus) have also described behaviours like belly presentation, body rotation, horizontal head movements that were clearly associated with non-agonistic encounters and interaction initiations. To our knowledge, the function of whistling in this context has never been studied, but it is intriguing to find more calls associated with closer and longer approaches. This also shows that orcas do not seek to be silent, whereas they do when hunting mammals [79][80][81] . We must acknowledge however that we could not identify callers and more solid conclusions require further investigation.
In conclusion, the fact that visual laterality and exploratory behaviours are sex-and age-dependent has interesting perspectives in both basic (understanding of the underlying, notably psychological, mechanisms and evolutionary processes) and applied research fields. Indeed, human-cetacean encounters are increasing all around the world 82,83 and the development of ecotourism over the last 20 years has created a significant pressure on the environment of cetaceans, and this is not limited to cetaceans. Complementary studies are necessary to explore further the different hypotheses raised here, and notably to test to which extent the motivation, the temperament, the degree of attention and the level of experience of animals are determining factors of behavioural variations.

Measured variables
Behavioural unit definitions Distance (minimum distance between the orca and the diving guide reached during a given approach) The orca approaches the diver from: -Very close (less than 5 meters) -Close (between 5 and 20 meters) -Far (more than 20 meters).

Angle of approach
The orca approaches the diver from: -The front: * from the right side of the diver (angle: 30 to 120°) * from the left side of the diver (angle: 200 to 330°) * facing the diver (angle: 0°) -Below: * moving just under the diver * arriving vertically from deep water -The back.

Swimming speed
-Slow: the orca moves but the movements of its caudal fin are hardly detectable -Fast: at least one caudal fin movement per second.

Gaze laterality
The orca looks at the diver with: -its right eye -its left eye -both eyes (bilateral) For a given approach, the eye used may change during the behavioural sequence and each gaze change is counted for the laterality analysis.

Gaze duration (mono-and binocular pooled)
The orca looks at the diver for: -less than 6 seconds (Short gaze) -more than 6 seconds (Long gaze) For a given approach, we measured the total gazing duration.

Head movements
-Horizontal (the orca turns slightly its head towards the diver) -Vertical (the orca is facing the diver and moves the head up and down several times in a row).

Whistle calls
-Whistling (a high-pitched modulated frequency is heard) -Silence (no species-specific sound is heard) It is not possible to confirm whether the caller is the observed orca or a neighbouring one.

Belly presentation
The orca moves the head up presenting its belly to the humans.

Body rotation
The orca rotates its body along the horizontal axis.

Materials and Methods
Study animals, context of the study and video recordings. The presence of orcas in the Kalkfjord (Norway) is directly linked to the massive presence of specific prey, i.e. atlantic herrings (Clupea harengus). As a result, the same population of orcas comes back to this location every year 84 . Individuals are identified based on classic criterion for that species, such as the size and the shape of the dorsal fin and the saddle patch as well as bodies scars 85 . Only individuals clearly identified and observed at least twice during the study have been included in the data set, i.e. 60 adult females, 18 adult males, 11 unsexed subadults. The underwater videos used for the analyses have been made by Pierre Robert de Latour during ecotouristic dives organised in Norway on winters 2014 and 2015 (November to February). The expeditions have been set up from the base M/S SULA and the dives started from a small barge, within a narrow period of daylight (i.e. 10 am-2 pm). Orcas are first observed at a distance, and then the barge is used to approach the animals following the USEA (Undersea Soft Encounter Alliance) method developed by Pierre Robert de Latour. This method is based on the principle of a limited intrusion into the orcas' space, who can then decide to approach the barge and the divers further or not. The divers (including snorkelers), with a maximum of eight people simultaneously in the water, are gathered just behind the guide, Pierre Robert de Latour. All people are commanded to stay as a group, move slowly and silently.
Video recordings have been made with an underwater camera Sony Ilce A6000 with a Nauticam NA 6000 housing, and equipped with a 24 mm focal equivalent to 24*36 (shutter speed: 1/25 to 1/100, aperture: f2.8, ISO: automatic limited at 3200, White Balance: 4800 k). All in all, 902 video sequences (collected with an ad libitum sampling method, i.e. at various times of the diving session and with various durations) have been analysed, that is 540 minutes of film sampled (2014: 323 video sequences lasting 163 minutes in total; 2015: 579 video sequences lasting 377 minutes in total).
Our study is observational and totally non-invasive, based on video scoring. Videos come from the archive of the USEA -Orques Sans Frontieres program (permit #USEA2017-0007).
Behavioural repertoire and data collection. A detailed behavioural repertoire has been established to sample the different movements and other exploratory behaviours of orcas when approaching the divers (see photo examples in Fig. 5). Here, an approach is defined as follows 1 : an approach is systematically an animal movement oriented in the direction of the divers 2 ; an approach starts from the moment where an orca enters the camera range and ends as soon as it gets out. The behavioural repertoire, composed of 10 measured variables (most of them being divided into mutually exclusive units), is given in Table 2. For each single approach, we used a 1-0 sampling method to score all measured variables.

Statistical analyses.
In a first step, to study the possible existence of a visual laterality bias during approaches at the population level, we compared, using Wilcoxon tests for matched samples, the number of gazes from the left eye with the number of gazes from the right eye directed to the humans based on all individuals sampled in each age-sex class.
In a second step, we tested whether the different individuals differed in the way they approached humans. On the one hand, using Wilcoxon tests for matched samples, we compared, for each measured variable and each age-sex class of individuals, the number of occurrences of the different behavioural units sampled. On the other hand, we calculated for each individual sampled the percentage of expression of a given unit over all the behavioural units of each measured variable. These individual percentages were then used to compare the behavioural profiles of adult males, adult females and unsexed subadults using Kruskal-Wallis and post-hoc Mann -Whitney tests (applying Bonferroni corrections for multiple comparisons). Significant threshold was set at 0.05 (and 0.0167 in case of Bonferroni correction).