Whale sharks (Rhincodon typus) are often associated with other fishes, such as remoras (echeneids), large predatory species including cobia (Rachycentron canadum) and schools of baitfishes (Carangidae spp.). For remora, the association is mostly symbiotic, with these fishes benefiting from reduced costs of movement by clinging to the shark and providing an advantage to the host by feeding on parasitic copepods (Pandarus rhincodonicus) (Dove and Robinson 2021; Norman et al. 2021). However, for baitfishes and cobia, the drivers of the association with whale sharks are less obvious. Thousands of small (10–20 cm length) carangid baitfishes can swarm around the head and body of whale sharks, but are not consumed by the whale shark (Fig. 1a). Three (non-exclusive) hypotheses suggest that there may be benefits to baitfishes of this association by: (1) sharks providing shelter from larger predators (Diamant et al. 2021; Dove and Robinson 2021; Sheaves et al. 2023), (2) the pressure wave of the whale shark reducing costs of travel (Sumikawa and Miyoshi 2022) and (3) the whale shark providing foraging opportunities (Andrades et al. 2012; Dove and Robinson 2021) for baitfishes.

Fig. 1
figure 1

a Baitfish schooling around the head of a whale shark (Rhincodon typus) at Ningaloo Reef, WA, b Predation by trevally (Carangidae spp.) on baitfishes. Images by Ollie Clarke Photography

Full size image

Here, we report on observations of adult trevally (Carangidae spp.) attacking and feeding on schools of carangid baitfishes associated with whale sharks from camera tags deployed on whale sharks and video taken by a tourism photographer. The tags were deployed in April 2009 (towed video tag, Model Y-1205, Marshall Electronics, Davis et al. 2003) and in May of 2017 and 2018 (CATS-Camera tags; Customised Animal Tracking Solutions, Moffat Beach, Australia) on whale sharks at Ningaloo Reef, Western Australia (see Meekan et al. (2015) for location map, Table 1 for details of deployments).

Table 1 Summary of video deployments of whale sharks at Ningaloo Reef, Western Australia
Full size table

The 2009 video from an animal-borne tow tag recorded a school of adult trevally (> 20 individuals) feeding on a school of > 200 baitfishes, which were mostly consumed in an attack that lasted only 19 s (Supp video 1). The shark reacted to this activity by slowing forward movement and adopting a tail-down, vertical posture. Once the attack ceased, it resumed horizontal forward movement. Feeding by the trevally resumed ~ 2 min later and the remaining baitfishes at the head of the shark were almost completely consumed within 8 s. A similar sequence of events was recorded by a photographer with the whale shark tourism industry at Ningaloo Reef (Supp video 2). In this video, a school of approximately 208 trevally attacked baitfishes associated with a whale shark for 9 s, with some prey attempting to flee toward the surface but still being consumed.

Further evidence of trevally attacking baitfish was recorded in videos from CATS tags deployed on five different whale sharks at Ningaloo Reef (Supp video 3). To determine how often these predation events occurred, we analyzed the footage from the CATS tags and calculated the time to a predation event or between successive events. A predation event was defined as baitfish moving around in a high-velocity swirl in response to rapid movement by adult trevally. In R Statistical Environment (R Core Team 2023, version 4.3) we used the time-to-event data to create a survival analysis with a Kaplan–Meier curve using the Surv and survfit functions from the package survival (Therneau and Lumley 2013). To obtain time-to-event probabilities we took the inverse value of the survival proportions and confidence intervals.

A total of 15 predation events occurred during the 23 h of video (0.65 predations per hour), and attacks by trevally had a median duration of 7 s, ranging from 2 to 45 s. Fewer trevally were involved in these attacks (typically < 12 individuals) and a smaller proportion of the baitfish schools were consumed, compared to the video obtained from the tourism industry and the towed tag. The time-to-event analysis indicated that the median interval between trevally attacks on baitfishes was 17.4 min (1044 s, probability = 0.51), with almost all attacks occurring within 37.9 min (2275 s, probability = 0.89), although these values are accompanied by large confidence intervals (Fig. 2).

Fig. 2
figure 2

Kaplan–Meier incidence plot showing that the probability of trevally (Carangidae spp.) attacking baitfish increased over the duration of a video. The blue shaded area shows the upper and lower 95% confidence interval of the prediction window

Full size image

The high frequency of attacks by trevally on baitfishes observed in the videos implies that shelter from predation may not be the primary driver of the association, although the rate of predation on baitfish schools away from whale sharks is currently unknown. Rather than reducing predation, this association may provide energetic (Sumikawa and Miyoshi 2022) and food acquisition advantages for baitfishes (Dove and Robinson 2021). Whale sharks at Ningaloo Reef (and other aggregation sites) move throughout the region to locate dense prey patches with high energetic value (Nelson and Eckert 2007; Gleiss et al. 2013; Ketchum et al. 2013; Meekan et al. 2015) where accompanying baitfish riding the bow wave gain the advantage of moving through the water column for a lower energetic cost (Sumikawa and Miyoshi 2022). Furthermore, these baitfish are capable of exploiting the same food resources located by the whale sharks, with negligible impacts on the food available to the sharks, potentially leading to a commensal relationship that is energetically beneficial to the baitfishes.