Multiple social benefits drive the formation of mixed-species groups of Australian humpback and Indo-Pacific bottlenose dolphins

Mixed-species groups are common amongst diverse taxa including fishes, birds, and mammals. Antipredator, foraging, and social benefits have been proposed as functional explanations for mixed-species group formation. Amongst delphinids, mixed-species groups are widespread, but little is known about their function. To investigate the potential benefits of delphinid mixed-species groups, we compared the number of individuals, the age composition, and the behaviour of single- and mixed-species sightings of Australian humpback (Sousa sahulensis) and Indo-Pacific bottlenose dolphins (Tursiops aduncus) observed around the North West Cape, Western Australia. We found no difference in the number of individuals or the age composition of humpback dolphins present in single- and mixed-species sightings, whereas bottlenose dolphins were present in larger numbers in single-species sightings than in mixed-species sightings due to a higher number of adults. Socialising was the initial observed behavioural state of 36.1% of mixed-species sightings, compared to only 5.1% and 10.3% of humpback and bottlenose dolphin single-species sightings, respectively. Furthermore, both species travelled and foraged less frequently when in mixed-species groups. Of 93 mixed-species groups observed during a focal follow of ≥ 10 min, 32 (34.4%) involved aggressive and/or sexual behaviours typically initiated by bottlenose dolphins towards humpback dolphins while the remaining 61 (65.6%) involved only neutral and affiliative behaviours. The results of this study suggest that the observed mixed-species groups provide multiple social benefits, particularly those pertaining to socio-sexual behaviours and the development and care of young. Numerous species of dolphin are known to form mixed-species groups. Few studies have, however, investigated the antipredator, foraging, and/or social benefits that these species may gain by grouping with other species. Investigating these benefits is key to understanding the impacts of mixed-species groups on the species and individuals involved. We evaluated the potential functions of mixed-species groups of two sympatric, inshore dolphin species — the Australian humpback dolphin and the Indo-Pacific bottlenose dolphin. We found that both species engage in socialising behaviour more frequently when in mixed-species groups and that they engage in a diverse variety of interspecific behavioural interactions. Our results not only indicate that social benefits are the key drivers of these mixed-species groups, but they also highlight the complexity of social interactions between dolphin species.


Introduction
Mixed-species groups have been observed amongst diverse taxa, most notably fishes, birds, and mammals (Stensland et al. 2003;Sridhar et al. 2009;Cords and Würsig 2014;Heymann and Hsia 2015;Goodale et al. 2017;Syme et al. 2021). Mixed-species groups form when an attraction between heterospecific individuals causes them to actively Communicated by S. Parks achieve and maintain spatiotemporal proximity (Stensland et al. 2003;Goodale et al. 2017;Syme et al. 2021). Accordingly, mixed-species groups should not be confused with chance encounters or with aggregations of animals that display shared responses to environmental stimuli because, unlike these situations, the formation of mixed-species groups is driven by evolutionary benefits (Waser 1984;Whitesides 1989;Stensland et al. 2003). These benefits typically fall within three main hypotheses -the antipredator, foraging, and social advantage hypotheses (Whitesides 1989;Stensland et al. 2003;Syme et al. 2021). These hypotheses are not mutually exclusive and, in some situations, a combination of drivers may affect the occurrence of mixed-species groups (Stensland et al. 2003;Zaeschmar et al. 2014;Syme et al. 2021).
There are diverse mechanisms by which individuals in mixed-species groups may gain evolutionary benefits. For example, antipredator benefits may be derived from the dilution effect, improved defence against predators, or increased group vigilance (Whitesides 1989;Stensland et al. 2003;Goodale et al. 2017) while foraging benefits may be obtained by mutual or non-mutual exchange of information to increase feeding opportunities or cooperative foraging (Whitesides 1989;Stensland et al. 2003;Goodale et al. 2017). Social benefits may be gained when interactions with heterospecifics increase the ability to defend territory or potential mates, increase the number of socio-sexual opportunities, or provide opportunities for individuals to practice a range of social behaviours, from play and alloparental care, to courtship and sexual behaviours, to infanticide (Spinka et al. 2001;Stensland et al. 2003;Syme et al. 2021). The social interactions that bring such benefits can, however, also result in certain costs including physical injury, displacement, and hybridisation (Stensland et al. 2003;Heymann and Buchanan-Smith 2007;Syme et al. 2021).
These different benefits, and their corresponding mechanisms, result in patterns in certain group and individual characteristics (e.g., group size, behaviour, and habitat use) according to how and why a given mixed-species groups was formed (Syme et al. 2021). Mixed-species groups that are larger in size than single-species groups can be suggestive of antipredator or foraging benefits (Heymann and Buchanan-Smith 2007;Sridhar et al. 2009). Furthermore, species that gain antipredator benefits from mixed-species groups often increase their association rate in places and at times of higher predation risk, such as where predators are abundant or when the group contains more vulnerable young individuals (Noë and Bshary 1997;Chapman and Chapman 2000;Wolters and Zuberbühler 2003;Kiszka et al. 2011). Foraging benefits are indicated by increased foraging behaviour and increased feeding rates in mixed-species groups compared to single-species groups (Heymann and Buchanan-Smith 2007;Sridhar et al. 2009). Also suggestive of foraging benefits are observations of cooperative foraging and a widening of foraging niche as the result of using novel microhabitats or food resources made available by heterospecifics (e.g., primates flushing insects from where they are hiding) when in mixed-species groups (Wolters and Zuberbühler 2003;Zaeschmar et al. 2013;Heymann and Hsia 2015). Where species form mixed-species groups to gain social benefits, interspecific social behaviour is frequent and involves recurring, and often diverse and complex, behavioural interactions from affiliative (e.g., play, grooming, and alloparental care), to aggressive (e.g., chases, forceful physical contact, and biting), and sexual behaviours (Van Lawick-Goodall 1968;Herzing and Johnson 1997;Acevedo-Gutiérrez et al. 2005;Struhsaker 2010;Syme et al. 2021). Investigating such behavioural interactions and their relationship to the ratio between species and the sex and age of the individuals involved can provide further insights into the nature and directionality of any social benefits (Herzing and Johnson 1997;Elliser and Herzing 2016).
Numerous species of delphinids have been observed in mixed-species groups, however most published reports are brief accounts and, consequently, whether they represent true mixed-species groups rather than chance encounters or aggregations is often unknown (Cords and Würsig 2014;Syme et al. 2021). Furthermore, the underlying causes of these interactions remain poorly understood, with several studies proposing the antipredator, foraging, and social advantage hypotheses as potential explanations (Stensland et al. 2003;Cords and Würsig 2014;Syme et al. 2021). These conclusions have typically been reached by analysing and comparing characteristics, such as the number of individuals, the sex and age composition, and the behaviour, of single-and mixed-species groups of the species involved. For example, Kiszka et al. (2011) suggested that spinner dolphins (Stenella longirostris) form mixed-species groups with pantropical spotted dolphins (Stenella attenuata) around the western Indian Ocean island of Mayotte to reduce predation risk when travelling in deep water on the basis that mixedspecies groups were larger than single-species groups, frequently exhibited travelling and resting behaviour, and occurred in areas associated with heightened predation risk. Observations off the California coast of common bottlenose dolphins (Tursiops truncatus) joining groups of foraging short-finned pilot whales (Globicephala macrorhynchus) led Shane (1994) to conclude that common bottlenose dolphins gain foraging benefits. Social benefits are harder to unravel but the prominence of socialising behaviour and the variety of direct interspecific interactions suggests that they may be the drivers behind several delphinid mixed-species groups including, for example, those of common bottlenose and Guiana dolphins (Sotalia guianensis) in Costa Rica (Acevedo-Gutiérrez et al. 2005; May-Collado 2010) and common bottlenose and Atlantic spotted dolphins (Stenella frontalis) in the Bahamas (Herzing and Johnson 1997;Melillo et al. 2009;Elliser and Herzing 2016).
Australian humpback (Sousa sahulensis) (hereafter "humpback dolphin") and Indo-Pacific bottlenose dolphins (Tursiops aduncus) (hereafter "bottlenose dolphin") occur in sympatry across northern Australia (Corkeron 1990;Allen et al. 2012;Brown et al. 2012;Palmer et al. 2014) and have been observed to regularly form mixed-species groups (Corkeron 1990;Brown et al. 2012;Hunt 2018). Humpback and bottlenose dolphins around the North West Cape, Western Australia, co-occur more often than would be expected by chance given their shared responses to key environmental variables (Syme et al. 2023). Furthermore, their patterns of co-occurrence do not resemble those of species aggregated around shared resources, suggesting that observed mixedspecies groups are indeed the result of an attraction between the species (Syme et al. 2023). Nevertheless, the group size, age composition, and behaviour of mixed-species groups of humpback and bottlenose dolphins are yet to be described in detail and their functional significance remains unknown.
Here, we analysed the characteristics of single-and mixedspecies groups of humpback and bottlenose dolphins around the North West Cape in order to assess the possible function of these mixed-species groups. If these species form mixedspecies groups to reduce predation risk, we would expect mixed-species groups to be larger, contain more calves, and to travel and rest more frequently than single-species groups. Alternatively, if they form mixed-species groups to improve foraging, then we would expect mixed-species groups to forage more often than single-species groups and, depending on how the foraging benefit is obtained, observations of one species following another during foraging and/or cooperative foraging between the species. Finally, if humpback and bottlenose dolphins gain social benefits from forming mixedspecies groups, then we would expect an increase in the frequency of socialising combined with numerous, and potentially diverse, direct interspecific behavioural interactions. Our research contributes to the limited, but increasing, knowledge of the complexity of delphinid mixed-species groups and provides direction for future research to investigate how these potentially diverse benefits apply to and impact the individuals and the species involved.

Study site and populations
The North West Cape is found in the Pilbara of Western Australia and is notable for its high biodiversity. To the east, lies Exmouth Gulf, a shallow, turbid embayment with sand and mud seabeds that are home to seagrass meadows, scattered coral reefs, and mangrove forests (Cassata and Collins 2008;Wilson 2013). To the north and west, lies the World Heritage Listed Ningaloo Reef which separates sandy coral lagoons from the open waters of the Indian Ocean (Cassata and Collins 2008;Wilson 2013).
The diverse habitat types around the North West Cape are important for both humpback and bottlenose dolphins. Hunt et al. (2017) estimated that the humpback dolphin population consists of 65-102 animals and has, at approximately one dolphin per km 2 , the highest recorded density for this species. Consistent with other humpback dolphin (Sousa sp.) populations, humpback dolphins around the North West Cape were found to form small groups of up to 19 individuals with an average (± SD) of 4.6 ± 3.2 (Hunt et al. 2017). The bottlenose dolphin population is larger, with Haughey et al. (2020) estimating the resident population at 141 and the super-population at 370 (including resident and transient individuals) animals across a three year study period (2013)(2014)(2015). Bottlenose dolphins are also found at a higher density, between 2.4 and 2.8 dolphins per km 2 , and form slightly larger groups of up to 30 individuals with an average (± SD) of 6.4 ± 5.2 (Haughey et al. 2020).

Data collection
Dolphin sightings were recorded during boat-based surveys conducted during the austral winter period of April to October over six years (2013-2015, 2018-2019, and 2021). Surveys were conducted from a 5.6 m research vessel that travelled at an average speed of 7 knots along two pre-determined, opposing, zigzag transect lines and one additional straight transect line. This ensured even coverage over the principal study area, which encompassed approximately 175 km 2 of shallow waters, to 30 m deep, on both sides of the Cape, from Exmouth to South Lagoon (Fig. 1). Surveys were restricted to daylight hours and good conditions (Beaufort scale ≤ 3, no rain or fog) to ensure a consistent sighting rate. Additional sightings were recorded during travel to and from the transect lines and during non-transect surveys both inside and outside the principal study area.
We use the term sighting to refer to both singletons (i.e., single animals) and groups, which were operationally defined as two or more individuals within 100 m of one another and engaged in similar behaviour (Hunt et al. 2017;Syme et al. 2022). The same definition was applied to both single-and mixed-species groups (Stensland et al. 2003;Syme et al. 2021). Additionally, following Weir et al. (2008) and Deutsch et al. (2014), we defined nursery groups as those groups with two or more calves and where calves constituted at least 25% of group members.
Upon each sighting, the dolphins were approached to within 10 -30 m and key characteristics were recorded including: the species present; the initial observed behavioural state of the majority of animals sighted (see Table 1 for behavioural state definitions); an estimate of the number of individuals; and the age composition, defined as the number of adults, juveniles, and calves (see Table 2 for age class definitions). For mixed-species sightings, the number of individuals and the age composition were recorded separately for each species. Where possible, the presumed sex of individuals was determined by either regular close association with a calf or juvenile, observation of the genital area or, for humpback dolphins, sexually dimorphic differences in the location and extent of patches of pigmentation loss and spotting (Brown et al. 2016;Hunt et al. 2019). When conditions permitted, focal follows were conducted until either the dolphins were lost from sight, weather conditions deteriorated, or a limit of one hour was reached. During focal follows, behavioural events, such as chasing, biting, forceful body contact, synchronous surfacing, play, and following, were recorded ad libitum (Altmann 1974;Mann 1999). During observations, the impact of the research vessel on dolphin behaviour was minimised by driving at a similar Fig. 1 The North West Cape, Western Australia, showing the two opposing, zigzag transect routes (blue and red) and the additional straight transect route (green) that were used to survey for Australian humpback (Sousa sahulensis) and Indo-Pacific bottlenose dolphins (Tursiops aduncus) as well as the vessel launch sites, Ningaloo Marine Park (dotted area), sanctuary zones (grey shaded areas), and reef crests (crosshatched area) Slow, non-directional movements with short dives at shallow angle, but most time spent at surface Close proximity, but few interactions Resting Very slow, non-directional movements at surface with low activity level Tight group formation Socialising Non-directional movements with dives of varying length Frequent aerial behaviours (e.g., leaping) and heads and flukes often above the water surface Close proximity, high levels of interaction including physical contact (e.g., touching, rubbing) Travelling Constant, directional movement with regular surfacing, shallow dive angles Group spacing varies Boat Interaction Interacting with boat (e.g., bow riding, wake riding) Boat Avoidance Rapid movement away from the boat, directional changes away from the boat, prolonged dives

Age Class Definition
Adult > 2 m long Juveniles 1 ∕ 2 to 2 3 the length of an adult Usually swimming alongside an adult, but sometimes swimming independently Calf ≤ 1 ∕ 2 the length of an adult Regularly swimming beside or slightly behind an adult speed to the dolphins and approaching them from the side.
Behavioural observations were limited to behaviours visible from outside the water, principally surface behaviours and occasionally underwater behaviours when conditions allowed. It was not possible to record data blind because our study involved observations of animals in the field.

Data analysis
Sightings of unknown species as well as sightings where the number of individuals, the age composition, or the initial observed behavioural state were unable to be determined were deemed incomplete and were not included in the analyses. All statistical analyses were conducted in RStudio 1.2.5 (RStudio Team 2019) and evaluated at a significance level of α = 0.05.

Comparisons of the number of individuals and age composition
For all the following comparisons of the number of individuals and the age composition, non-parametric randomisation tests were used as the observed data did not meet the assumptions of homogeneity of variance or normality. All randomisation tests utilised 10 000 iterations where, for each iteration, the data were randomly assigned to two samples and the mean difference between the samples was calculated and recorded. Significance was evaluated as the number of times that the mean difference obtained from an iteration was greater than that obtained from the observed data divided by the total number of iterations (Manly 1997). The p-values produced from the randomisation tests were adjusted using the Benjamini-Hochberg Procedure to reduce the chance of Type I error (Benjamini and Hochberg 1995).
To evaluate how mixed-species groups fit into the broader patterns of grouping dynamics, the number of individuals and the age composition of single-species sightings were compared to those of mixed-species sightings. Additionally, the number of individuals and the age composition of conspecifics within mixed-species sightings were compared to those of single-species sightings of the same species. Finally, given that the relative number of individuals of each species can influence the nature of interspecific interactions, the number of individuals and the age composition of conspecifics within mixed-species sightings were compared to each other.

Frequencies of initial observed behavioural states
To determine if the presence of heterospecifics influences the species' behaviour, the frequencies of the initial observed behavioural states of each species were compared between single-and mixed-species sightings with a chi-square contingency table. Sightings with an initial observed behavioural state of boat avoidance or boat interaction (n = 21) were excluded from this analysis.

Frequency of nursery groups
For each species, the frequency of nursery groups was compared between single-and mixed-species sightings using a chi-square contingency table.

Nature of interspecific interactions
To better understand how the relative numbers and age composition of each species relate to the nature of interspecific interactions, we evaluated the behavioural events observed during focal follows of single-and mixed-species sightings. Sightings were then categorised into those involving aggressive (e.g., tail slaps, forceful body contact, and chases) and/ or sexual (e.g., copulation and ventral-to-ventral positioning) behavioural events (hereafter "aggressive/sexual sightings") and those not involving such behavioural events (hereafter "non-aggressive/sexual sightings") (Herzing and Johnson 1997;Parra 2005;Elliser and Herzing 2016). Aggressive and sexual behavioural events were often concurrent and difficult to distinguish in the field, hence sightings involving such behaviours were combined into one category. Only sightings observed during a focal follow lasting ≥ 10 min were  Total   2013 84  62  122  32  216  45  45  21  111  2014 94  139  109  23  271  97  29  23  149  2015 70  45  86  15  146  17  12  6  35  2018 85  57  94  11  162  44  80  10  134  2019 93  85  118  30  233  69  107  27  203  2021 42  39  55  8  102  36  49  6  91  Total 468  427  584  119  1130 308  322  93  723 analysed to ensure that adequate opportunity was had to observe any aggressive or sexual behavioural events, which were recorded ad libitum (Altmann 1974;Mann 1999). For both humpback and bottlenose dolphins, the frequency of aggressive and/or sexual behaviour was compared between single-and mixed-species sightings with a chi-square contingency table. The characteristics of mixedspecies sightings were then analysed with randomisation tests, as described previously, firstly, by comparing the total number of individuals (i.e., both species combined) in aggressive/sexual and non-aggressive/sexual mixed-species sightings, then, by comparing the number of individuals and the age composition of each species between aggressive/ sexual and non-aggressive/sexual mixed-species sightings and between species.

Sightings summary
Both humpback and bottlenose dolphins were observed throughout all six field seasons, resulting in a total of 1130 dolphin sightings (Table 3, Fig. 2). Of these sightings, 427 (37.8%) were of humpback dolphins only, 584 (51.7%) were of bottlenose dolphins only, and 119 (10.5%) contained both species (i.e., mixed-species groups) (Table 3). Accordingly, 21.8% of the 546 sightings containing humpback dolphins were mixed, as were 16.9% of the 703 sightings containing bottlenose dolphins. Both species were observed across the study area, primarily in shallow inshore waters. Bottlenose dolphins were also found further offshore, particularly in Exmouth Gulf, but were observed less frequently than humpback dolphins outside the reef crest on the western side of the North West Cape (Fig. 2). These trends are in agreement with previous studies of habitat partitioning of these species (Corkeron 1990;Hanf et al. 2022). Mixed-species groups were also distributed around the Cape although they occurred primarily in inshore waters where both species were frequently present (Fig. 2).

Frequencies of initial observed behavioural states
The initial observed behavioural states differed significantly between humpback dolphin single-species sightings and mixed-species sightings (χ 2 4 = 99.6, p < 0.001) and between bottlenose dolphin single-species sightings and mixed-species sightings (χ 2 4 = 61.8, p < 0.001) (Fig. 4). Most notably, socialising was the initial observed behavioural state of only 5.1% of humpback and 10.3% of bottlenose dolphin single-species sightings but was the most common initial observed behavioural state of mixed-species sightings (36.1%) (Fig. 4). Furthermore, travelling was the second most common initial observed behavioural state of mixed-species sightings (28.6%) but was noticeably less frequent for mixed-species sightings than for humpback dolphin single-species sightings (57.8%) and, to a lesser extent, Fig. 3 Comparisons between Australian humpback (Sousa sahulensis) and Indo-Pacific bottlenose dolphins (Tursiops aduncus) with regards to (a) the number of individuals as well as the number of (b) adults, (c) juveniles, and (d) calves present in single-and mixedspecies sightings around the North West Cape, Western Australia. The box plot shows the interquartile range with the line showing the median and the whiskers 1.5 times the interquartile range while dots represent outliers and diamonds indicate the mean. Bar plots show mean values with standard error bars. Significance between key pairs is written above the brackets ▸ bottlenose dolphin single-species sightings (35.1%) (Fig. 4). Similarly, foraging was the initial observed behavioural state of 21.4% and 32.0% of humpback and bottlenose dolphin single-species sightings, respectively, but only 10.9% of mixed-species sightings (Fig. 4).

Frequency of nursery groups
Of the mixed-species sightings, 18 (15.1%) were classified as nursery groups (i.e., groups with two or more calves and where calves constituted at least 25% of group members), as were 61 (14.3%) of the humpback and 26 (4.5%) of the bottlenose dolphin single-species sightings. For bottlenose dolphins, nursery groups were more frequent amongst mixedspecies sightings than single-species sightings (χ 2 1 = 17.4, p < 0.001) however there was no such difference for humpback dolphins (χ 2 1 = 0.007, p > 0.05).

Nature of interspecific interactions
Across all mixed-species sightings, the number of individual humpback and bottlenose dolphins did not exhibit any significant difference (humpback: mean ± SD = 4.6 ± 2.8; bottlenose: mean ± SD = 4.6 ± 3.5; p = 0.872) (Fig. 3a), however this apparently equal representation was not observed after considering the nature of the behavioural events exhibited by the individuals during interspecific interactions. Of 93 mixed-species sightings observed during a focal follow of ≥ 10 min, 32 (34.4%) involved aggressive and/or sexual behavioural events (Fig. 5a). In comparison, 31 of 308 (10.1%) humpback dolphin and 47 of 275 (14.6%) bottlenose dolphin single-species sightings observed during a focal follow of ≥ 10 min involved such aggressive and/or sexual behavioural events. For both species, aggressive and/or sexual behavioural events were more frequent in mixed-species sightings than in Fig. 4 The frequency of the initial observed behavioural states of single-species sightings of Australian humpback dolphins (Sousa sahulensis), single-species sightings of Indo-Pacific bottlenose dolphins (Tursiops aduncus), and mixedspecies sightings of both species observed around the North West Cape, Western Australia single-species sightings (humpback: χ 2 1 = 30.2, p < 0.001; bottlenose: χ 2 1 = 17.1, p < 0.001). Both species were observed to perform a variety of aggressive behavioural events, including forceful body contact and tail slapping, however observed chases involved bottlenose dolphins pursuing humpback dolphins (Fig. 6a). Sexual interactions between humpback and bottlenose dolphins were characterised by high levels of physical contact (e.g., rubbing and rolling over one another), ventral-toventral positioning, mounting, and erections. Additionally, avoidance behaviours, such as positioning the ventral side towards the water surface and raising the head above the water surface (Fig. 6b), were also observed.
The remaining 61 (65.6%) mixed-species sightings, including 14 (93.3%) of 15 nursery groups observed for ≥ 10 min, did not involve any such aggressive and/or sexual behavioural events (Fig. 5b). They did, however, involve neutral behavioural events, such as individuals swimming alongside one another without any observed direct behavioural interactions, as was often displayed during travelling and resting (Fig. 6c), and affiliative behavioural events, including synchronous surfacing, non-aggressive tactile interactions, and playful behaviour often involving calves and juveniles. Calves were also observed swimming closely alongside a heterospecific adult in events that resembled interspecific alloparental care on four occasions: one involving a bottlenose dolphin calf and a humpback dolphin adult and three involving a humpback dolphin calf and a bottlenose dolphin adult (Fig. 6d).

Discussion
Mixed-species groups of delphinids are widespread, yet their dynamics and potential functions have rarely been investigated (Stensland et al. 2003;Cords and Würsig 2014;Syme et al. 2021). Australian humpback and Indo-Pacific bottlenose dolphins around the North West Cape, Western Australia, were regularly observed in mixed-species groups (10.5% of all dolphin sightings were mixed). Previous studies of mixed-species groups of small, coastal delphinids have reported similar, although generally higher, values (see Syme et al. 2021 for a review). For example, in the Bahamas, 8.9% of Atlantic spotted and common bottlenose dolphin sightings were mixed (Elliser and Herzing 2016), as were 32.4% of Guiana and common bottlenose dolphin groups in Costa Rica (Acevedo-Gutiérrez et al. 2005). The drivers of mixed-species groups of humpback and bottlenose dolphins around the North West Cape are, perhaps, not as strong as those that drive more frequently occurring mixed-species groups. Nevertheless, mixed-species groups account for a considerable proportion of sightings of both humpback and bottlenose dolphins and occur more frequently than expected by chance (Syme et al. 2023). Thus, the formation of mixedspecies groups of humpback and bottlenose dolphins around the North West Cape is presumably driven by evolutionary benefits that one or both species gain from grouping with the other. Here, by assessing the relationship between group characteristics (e.g., size and age composition) and the frequency and nature of their behavioural interactions, we show that the potential functions of mixed-species groups of sympatric humpback and bottlenose dolphins around the North West Cape are most consistent with the social advantage hypothesis and, at least in certain cases, the antipredator advantage hypothesis.

Antipredator advantage hypothesis
The larger number of individuals in mixed-species groups could lead to a reduction in predation risk, as has been hypothesised for other species (Scott and Chivers 1990;Scott and Cattanach 1998;Heymann and Buchanan-Smith 2007;Kiszka et al. 2011). Little information is available on predation risk around the North West Cape, however numerous humpback and bottlenose dolphins bear scars from shark bites (JS and GJP personal observations) and both species are known to share predators, including tiger (Galeocerdo cuvier), bull (Carcharhinus leucas), and white sharks (Carcharodon carcharias) (Heithaus et al. 2017;Parra and Jefferson 2018;Smith et al. 2018;Wang 2018). The observed mixed-species groups of humpback and bottlenose dolphins do not, however, resemble mixed-species groups that form to reduce predation risk in several aspects. Notably, humpback dolphins and, to a lesser extent, bottlenose dolphins travelled less frequently when in mixed-species groups and direct behavioural interactions between the species were frequent. This differs from spinner dolphin -pantropical spotted dolphin mixed-species groups around Mayotte (Kiszka et al. 2011) and Indian Ocean humpback dolphin (Sousa plumbea) -bottlenose dolphin groups in South Africa (Koper and Plön 2016) where travelling behaviour was more frequent in mixed-species groups and no direct interactions were observed. Furthermore, the frequent occurrence of socio-sexual interactions observed in mixed-species groups of humpback and bottlenose dolphins in this study would likely lead to decreased vigilance, further arguing against the antipredator advantage hypothesis.
Certain species have been shown to form mixed-species groups when the number of vulnerable young individuals, and therefore the predation risk, is highest (Chapman and Chapman 2000). This does not appear to be the case around the North West Cape where, for both species, the number of calves was not higher in mixed-species groups. Nevertheless, calves were present in mixed-species groups, including 18 observed nursery groups, and nursery groups were more frequent for bottlenose dolphins in mixed-species sightings than single-species sightings. Furthermore, we observed instances of humpback and bottlenose dolphins travelling or resting alongside each other without any evident direct behavioural interactions. Thus, it seems plausible that in travelling and resting groups, particularly nursery groups, mixed-species group formation could result in increased group vigilance and, therefore, reduced predation risk. Therefore, it remains a possibility that, at least in certain cases, humpback and bottlenose dolphins around the North West Cape form mixed-species groups to gain antipredator benefits. Future research could investigate whether vigilance behaviour changes when individuals are Fig. 7 Comparisons between Australian humpback (Sousa sahulensis) and Indo-Pacific bottlenose dolphins (Tursiops aduncus) with regards to (a) the number of individuals as well as the number of (b) adults, (c) juveniles, and (d) calves present in mixed-species sightings around the North West Cape, Western Australia, that involved aggressive and/or sexual behaviours (i.e., aggressive/sexual) or did not involve such behaviours (i.e., non-aggressive/sexual). The box plot shows the interquartile range with the line showing the median and the whiskers 1.5 times the interquartile range while dots represent outliers and diamonds indicate the mean. Bar plots show mean values with standard error bars. Significance between pairs is written above the brackets ▸ 43 Page 12 of 16 in mixed-species groups or whether mixed-species groups are more frequent in areas of higher predation risk (e.g., areas where large sharks are abundant). Such investigations have been conducted on primate and avian species and can provide more insight into the potential antipredator benefits of mixed-species group formation (Stensland et al. 2003;Heymann and Buchanan-Smith 2007;Sridhar et al. 2009;Goodale et al. 2017).

Foraging advantage hypothesis
Mixed-species groups where foraging benefits are gained often involve heterospecifics with similar diets as this leads to higher quality information concerning foraging opportunities (Sridhar and Guttal 2018). No dietary data from the study populations are available, however both species are generalists and analyses of stomach contents and observations of foraging from other locations indicate that there is potentially some interspecific dietary overlap (Amir et al. 2005;Parra 2006;Cagnazzi et al. 2011;Kiszka et al. 2014;Parra and Jedensjö 2014). Our observations of mixed-species groups do not, however, correspond to those expected from species that gain foraging benefits by forming mixed-species groups. Most notably, both species foraged less frequently when in mixed-species groups. This contrasts with mixed-species groups where foraging behaviour is a prominent feature such as those of false killer whales (Pseudorca crassidens) and common bottlenose dolphins off New Zealand (Zaeschmar et al. 2014) and Diana monkeys (Cercopithecus diana) and Campbell's monkeys (Cercopithecus campbelli) in the Taï Forest, Côte d'Ivoire (Wolters and Zuberbühler 2003). Furthermore, we did not observe a clear tendency of one species to follow or join the other, as is the case for common bottlenose dolphins that follow and join foraging short-finned pilot whales and Risso's dolphins (Grampus griseus), seemingly to obtain information about prey (Shane 1994;Bacon et al. 2017), or rock kestrels (Falco rupicolus) that preferentially follow "travel foraging" (i.e., searching for food while moving) chacma baboons (Papio ursinus) to catch insects flushed by the primates' movements (King and Cowlishaw 2009). Finally, the absence of observed intra-and interspecific cooperative foraging, combined with the tendency of humpback dolphins to forage in small groups (2-3 individuals) (Parra et al. 2011) argues against the possibility of interspecific cooperative foraging as has been observed, for example, between false killer whales and common bottlenose dolphins (Zaeschmar et al. 2013). Thus, it seems unlikely that foraging benefits drive the formation of mixed-species groups of humpback and bottlenose dolphins in the coastal waters of the North West Cape.

Social advantage hypothesis
The marked increase in the frequency of socialising behaviour when in mixed-species groups and the numerous, diverse behavioural interactions between the species indicate that social benefits may play a key role in the formation of mixed-species groups of humpback and bottlenose dolphins around the North West Cape. The interactions recorded between humpback and bottlenose dolphins covered a wide range of behaviours, from affiliative to aggressive. Similar variation in interspecific interactions has previously been reported between certain primate (Van Lawick-Goodall 1968;Struhsaker 2010) and dolphin species (Herzing and Johnson 1997;Acevedo-Gutiérrez et al. 2005;Parra 2005;Melillo et al. 2009;May-Collado 2010) and suggests diverse potential social benefits including the practicing of sociosexual behaviours, the provision of an ideal environment for the development of calves, and alloparental care.

Aggressive and sexual interactions
Aggressive and/or sexual interactions, which were observed in approximately a third of mixed-species groups, were recorded more frequently in mixed-species sightings than single-species sightings of either species. Previous research on agonistic interspecific interactions within mixed-species groups indicates a tendency for larger species to be dominant over smaller ones (Heymann 1990;Herzing and Johnson 1997;Psarakos et al. 2003;Parra 2005;Heymann and Buchanan-Smith 2007;Elliser and Herzing 2016). Given the similarity in body size of humpback and bottlenose dolphins (Parra and Jefferson 2018;Wang 2018), it is not apparent that either species should have a noticeable physical advantage over the other. Bottlenose dolphins had, however, a clear numerical advantage during aggressive and sexual interactions and, accordingly, typically appeared to initiate such interactions with humpback dolphins -a role that is proving to be widespread with both Indo-Pacific and common bottlenose dolphins having been reported to initiate aggressive and/or sexual interactions with heterospecifics (Ross and Wilson 1996;Herzing and Johnson 1997;Stensland et al. 2003;Wedekin et al. 2004;Acevedo-Gutiérrez et al. 2005;Barnett et al. 2009;Melillo et al. 2009;Cotter et al. 2012). During interactions not involving aggressive and/or sexual behaviours, however, bottlenose dolphins were outnumbered by humpback dolphins and neither species appeared to be dominant. This pattern is suggestive of the possibility that bottlenose dolphins tend to initiate aggressive and sexual behaviours when they outnumber humpback dolphins but not when they are outnumbered. A similar dynamic has been reported in the Bahamas where male common bottlenose dolphins are dominant over male Atlantic spotted dolphins, except when the latter are more numerous, in which case they chase the common bottlenose dolphin males away (Herzing and Johnson 1997). Alternatively, it is possible that mixed-species groups start with fairly equitable numbers and that the imbalance develops as more bottlenose dolphins are drawn to the commotion from aggressive and sexual behaviours while any humpback dolphins in the vicinity stay away. Although it is not clear as to which is the cause and which is the consequence, the nature of the interspecific interactions in mixed-species groups is apparently linked to the number of bottlenose dolphins and, by extension, to the relative numbers of each species.
The socio-sexual behaviours exhibited by bottlenose dolphins towards humpback dolphins around the North West Cape are similar to those reported elsewhere. In the Bahamas, for example, immature or low-ranking male common bottlenose dolphins engage in sexual interactions with Atlantic spotted dolphins (Melillo et al. 2009) while in Zanzibar, in the western Indian Ocean, young male bottlenose dolphins display aggressive and sexual behaviours towards female Indian Ocean humpback dolphins, possibly in order to practice social behaviours or for mating (Stensland et al. 2003). Around the North West Cape, male bottlenose dolphins may similarly use interactions with humpback dolphins to practice socio-sexual behaviours and, in doing so, increase their chances of subsequent reproduction with conspecifics. Further investigation could test such a hypothesis by analysing long-term data on the individuals involved in these interactions and their reproductive success (Stensland et al. 2003).
The presence of sexual interactions between humpback and bottlenose dolphins leads naturally to the question of whether they result in hybrid individuals. Numerous cases of hybridisation involving Sousa and/or Tursiops species have been reported (Stensland et al. 2003;Brown et al. 2014;Crossman et al. 2016), including, notably, potential hybridisation of bottlenose dolphins and Indian Ocean humpback dolphins in South Africa (Karczmarski et al. 1997;Koper and Plön 2016). Humpback dolphin -bottlenose dolphin hybrids are yet to be confirmed by genetic analysis at the North West Cape or in another location but could represent a potential cost to these interspecific social interactions. Other potential costs could include physical injury, the energetic costs of high intensity social behaviour, and displacement if aggressive interactions lead to certain individuals avoiding others.

Affiliative and neutral interactions
Around the North West Cape, two thirds of the mixed-species groups of humpback and bottlenose dolphins observed for ≥ 10 min did not involve aggressive or sexual behaviours. While it is possible that the benefits from these mixed-species groups fit within the antipredator advantage hypothesis, the nature of these interactions suggest the existence of social benefits, in particular, those pertaining to the development and care of young individuals (Whitehead and Mann 2000; Spinka et al. 2001). Humpback and bottlenose dolphin calves and juveniles were observed swimming alongside each other and engaging in apparent play behaviour (e.g., leaping, chasing, and breaching). This raises the possibility that mixed-species groups provide opportunities for young dolphins to play and, thus, to develop social and physical skills (Spinka et al. 2001), as may be the case for mixed-species groups of primates (Struhsaker 2010).
Behaviours that involved potential alloparental care between humpback and bottlenose dolphins were also observed on four occasions, although it is impossible to know whether they truly represent alloparental care as it is not known if the calf and the mother benefitted (Whitehead and Mann 2000). Similar interactions resembling interspecific alloparental care have been observed previously amongst wild delphinids (Bearzi 1996;Herzing and Johnson 1997;Stensland et al. 2003;Markowitz 2004). The potential benefits and costs of interspecific alloparenting are presumably similar to those of intraspecific alloparenting, including respite from caregiving for the mother and practice parenting for the alloparent (Mann and Smuts 1998). Whether these affiliative behaviours, such as play and alloparental care, are rare or simply rarely observed is unclear and, thus, the extent to which they influence the formation of mixed-species groups is difficult to determine.
Nevertheless, they may be particularly pertinent to mothers and their young, given the presence of such affiliative behaviours in mixed-species nursery groups. Generally speaking, nursery groups may reduce predation risk and provide social benefits, such as protection from male harassment and opportunities for the development of physical and social skills (Wells et al. 1987;Weir et al. 2008;Deutsch et al. 2014). Both species formed singleand mixed-species nursery groups and, for bottlenose dolphins, nursery groups were more frequent amongst mixedspecies sightings than single-species sightings. If maternal care places similar requirements on both humpback and bottlenose dolphin mothers, the benefits of forming nursery groups may extend to mixed-species nursery groups, as has been suggested for dusky (Lagenorhynchus obscurus) and Hector's dolphins (Cephalorhynchus hectori) in New Zealand (Markowitz 2004). For example, the need to protect young calves from adult male harassment could lead mothers of both species to avoid groups of individuals displaying aggressive or sexual behaviours (Pearson 2011;Cords and Würsig 2014), thus explaining the near absence of these behaviours in mixed-species nursery groups.

Summary
In this study we have provided evidence that humpback and bottlenose dolphins around the North West Cape form mixed-species groups primarily to gain social, and perhaps antipredator, benefits whose relevance may depend on the individuals involved and their traits (e.g., age, sex, social status, and reproductive stage). For example, we can hypothesise contrasting situations such as the following: a young male will seek opportunities to learn and practice social skills, both with con-and heterospecifics, resulting in aggressive and sexual interactions, whereas a mother with a calf will seek out other mothers and calves, be they con-or heterospecifics, to provide a safe environment for the development of her young. Future research could further unravel the complexity of these social benefits by studying these mixed-species groups on an individual, rather than a group or species, level. This would provide a more detailed understanding of what drives particular individuals to form mixed-species groups and the benefits and costs that they may experience from doing so.