Illuminating otoliths: New insights for life history of Balistes triggerfishes

Abstract Our understanding of fish life‐history strategies is informed by key biological processes, such as growth, survival/mortality, recruitment and sexual maturation, used to characterize fish stocks (populations). Characterizing the life‐history traits of fish populations requires the application of accurate age estimation for managed species. Grey triggerfish Balistes capriscus and queen triggerfish Balistes vetula are important reef‐associated species for commercial and recreational fisheries in the Atlantic Ocean. Both species exhibit a unique reproductive strategy for large‐bodied fisheries‐targeted reef fishes in that they are nesting benthic spawners and invest substantial energy in defence and care of their benthic nests and fertilized eggs. Until recently, our understanding of the life‐history strategies of triggerfishes assumed the main method used to obtain age estimates, increments counted from thin sections of the first dorsal spine, provided an accurate characterization of population age‐based parameters. However, results from bomb radiocarbon validation studies on the two Balistes species demonstrated that spines do not provide accurate ages, but sagittal otoliths do. The main goal of the current study was to provide an updated understanding for triggerfish life‐history strategies by using otolith‐based age estimates to characterize population age structure and growth for grey triggerfish and queen triggerfish from waters of the south‐eastern U.S. Atlantic. The current study is the first to report on sex‐specific age and growth information for grey triggerfish using the Δ14C‐validated otolith‐based age estimation method and the results indicate that the previous characterization of Balistes species as exhibiting moderately rapid growth and as relatively short‐lived, based on spine‐derived age estimates, are flawed. Otolith‐based ages indicated that grey triggerfish and queen triggerfish are moderately slow‐growing and long‐lived species, attaining maximum ages of 21 and 40 years, respectively. Management efforts for triggerfishes should evaluate these new insights and incorporate the results of otolith‐based age estimation into future population monitoring efforts.


| INTRODUCTION
Our understanding of life-history strategies of fishes is informed by key biological processes such as growth, survival/mortality, recruitment and sexual maturation that are used to characterize fish stocks (populations). Accurate age estimation for fisheries species is a critical component of ensuring sustainable fisheries management (Beamish & McFarlane, 1983;Campana, 2001). This is because in the assessment of fisheries species, population age-based parameters, such as longevity, age at sexual maturity, age at sexual transition for sequential hermaphroditic species, growth rate, mortality, age-specific reproductive output and lifetime reproductive output, are important in understanding overall life-history strategies of managed species (Beverton, 1998;King & McFarlane, 2003;Lorenzen & Enberg, 2002;Winemiller, 2005).
B. capriscus and B. vetula co-occur throughout much of their ranges in the western Atlantic but appear to vary from each other in their geographic patterns of abundance. In the northern hemisphere of the Atlantic both species are strongly associated with shelf and slope habitats characterized by hard-bottom and reef-like structures (Garcia-Sais, 2010;García-Sais et al., 2007;Glasgow et al., 2021;Sedberry & Van Dolah, 1984;Shervette & Rivera Hernández, 2022b). B. capriscus occur with greatest abundance at higher latitudes of the shared range, from as far north as waters offshore of North Carolina, United States, down through much of the waters off Florida (Glasgow et al., 2021;Kellison & Sedberry, 1998;Muhling et al., 2014;Sedberry et al., 1998Sedberry et al., , 2006. B. capriscus also occurs at high abundance in the northern Gulf of Mexico (GOM) (Allman et al., 2018). B. vetula occurs at higher abundance in the lower latitudes of its range, including waters throughout the Caribbean Sea (Robertson & Van Tassell, 2019).
B. capriscus and B. vetula are gonochoristic species that exhibit a unique reproductive strategy compared to other large-bodied fisheries-targeted reef fishes. Both species are nesting benthic spawners that utilize nesting grounds associated with coral reef habitats (Shervette & Rivera Hernández, 2022b;Simmons & Szedlmayer, 2012). B. capriscus spawning occurs from late April to early September in waters off North Carolina through north Florida (Kelly-Stormer et al., 2017) and from late May to August in waters of the northern GOM (Ingram, 2001;Lee, 2019). B. vetula spawning occurs in the north Caribbean over a longer time period compared to grey triggerfish, starting as early as December and extends through August (Rivera Hernández et al., 2019).
Until recently, our understanding of the general life-history strategy of Balistes triggerfish species was based on the assumption that the method used to obtain age estimates, increments counted from thin sections of the first dorsal spine, provided an accurate characterization of population age-based parameters (Albuquerque et al., 2011;Allman et al., 2018;Burton et al., 2015;Manooch & Drennon, 1987). From spine-based age estimates, Balistes species were thought to exhibit moderately rapid growth (Aiken, 1983;Allman et al., 2018;Kelly-Stormer et al., 2017), reach sexual maturity within the first 2 years of life (Aiken, 1983;Ingram, 2001;Moore, 2001) and were relatively short-lived, attaining maximum ages of 14-15 years (Albuquerque et al., 2011;Allman et al., 2018;Burton et al., 2015;Johnson & Saloman, 1984). However, recent age estimation validation studies utilizing regional patterns of bomb radiocarbon concluded that the first dorsal spine does not provide accurate age estimates for Balistes species (Patterson et al., 2019;Shervette & Rivera Hernández, 2022a).
A study on B. vetula from the north Caribbean validated the accuracy of otolith-based age estimation and demonstrated that spine-based age estimates resulted in an erroneous characterization of population age structure, growth and longevity compared to the results from otolithbased ages (Shervette & Rivera Hernández, 2022a). Our new understanding of the life-history strategy of this species is that B. vetula is characterized by moderately slow growth and is long-lived.
For B. capriscus, results from radiocarbon validation efforts indicated that, similar to B. vetula, the first dorsal spine does not provide accurate age estimates, but otoliths do (Patterson et al., 2019). The next important step in updating the understanding of the general lifehistory strategy of triggerfish species is to use otolith-based age estimates to document population age structure and growth for regional contingents of B. capriscus and expand on B. vetula by characterizing the age and growth of this species outside of the north Caribbean region. Therefore, the overall goals of this study were two-fold. The first goal was to provide an updated understanding of the life-history strategy of B. capriscus by using otolith age estimates to describe agebased population parameters. The second goal was to document the age and growth of B. vetula from waters of the SEUS. The specific objectives were to (1) determine the timing of opaque zone formation in triggerfish otoliths, (2) describe population age structure and growth for the two Balistes species using samples from the SEUS and (3) compare size-at-age between grey triggerfish males and females.

| Study area
Samples for this study were collected from waters off the coasts of North Carolina and South Carolina, United States. This region of the Atlantic is characterized by a wide shelf, extending up to 145 km from the shore. In this region, B. capriscus has a relatively broad across-shelf distribution (Glasgow et al., 2021), occurring with moderate to high frequency in the mid-shelf (20-30 m depths), outer-shelf (30-50 m depths) and shelf-edge (50-100 m depths) zones (Muhling et al., 2014;Sedberry et al., 1998Sedberry et al., , 2006Sedberry & Van Dolah, 1984). In the mid-and outer-shelf zones, emergent hard bottom and rock outcrops provide low-profile three-dimensional structural complexity that is enhanced by gorgonian corals and sponges (Muhling et al., 2014). Artificial reefs and wrecks occur intermittently throughout the study area and provide additional structurally complex habitat for triggerfishes and other fisheries species targeted by commercial and recreational fishing (Kellison & Sedberry, 1998).
B. vetula is less common in this region and a dearth of published information exists on this species' habitat associations in SEUS waters.
Video surveys conducted in the SEUS from 2015 to 2017 documented a 5.4% frequency of occurrence for queen triggerfish compared to 45.6% for B. capriscus (Bacheler et al., 2019). In the waters of southwest Florida, a survey of coral and fish assemblages associated with Pulley Ridge (located in waters of the south-west corner of Florida) reported B. vetula occurred in rock rubble habitat at depths exceeding 60 m (Harter et al., 2008). Several studies from the Caribbean reported that adult B. vetula are associated with coral reef ecosystem habitats that occur in deeper shelf and shelf edge zones (Garcia-Sais, 2010;García-Sais et al., 2007;Shervette & Rivera Hernández, 2022b). Diver surveys in Saba Bank documented adults associated with the whole spectrum of coral reef ecosystem strata, but were most abundant in the outer reef flat zone characterized by hard bottom/pavement and a submerged inner reef flat zone with low relief pavement and scattered rubble (Debrot et al., 2020;Toller, 2007). B. vetula in the current study were caught as incidental catch in deeper shelf waters (>45 m) and at shelf edge sites where commercial fishers were targeting large groupers, snappers and tilefishes.  (Table 1).
All fish samples were kept on ice until processing occurred. Fish were measured for size (standard length: L s , fork length: L f , total length: L t ) to the nearest millimetre. Fish obtained whole were weighed (g). Whenever possible for FD samples and for all FI samples (including small, juvenile fish collected from sargassum), gonads were collected and preserved for histological processing to determine sex T A B L E 1 Summary results for Balistes triggerfish sample collection from waters of the south-eastern United States Sagittal otoliths ( Figure 1a) were carefully extracted following the methods described in Rivera Hernández and Shervette (2022), and saved for age estimation.
The Δ 14 C-validated otolith age estimation protocol for B. capriscus and B. vetula (Rivera Hernández & Shervette, 2022) was used to obtain ages for samples from both species via enumeration of otolith opaque zones (Figure 1b,c). Briefly, sagitta were read whole, submerged in water against a black background with a stereo microscope at a magnification range of 20-40Â (Supporting Information Figure S1). This age estimation protocol for triggerfish sagitta includes the use of concentrated light to more effectively illuminate the otolith via a fibre optic cable (this allows the reader to control light intensity and angle such that opaque zones appear to glow); each opaque zone present was counted ( Excel (Haddon, 2011). A lack of juvenile B. vetula in the SEUS collections necessitated the use of a fixed t 0 value (À0.585) that was previously computed for this species in a recent otolith-based age and growth study (Shervette & Rivera Hernández, 2022b).

| RESULTS
A total of 1044 grey triggerfish and 27 queen triggerfish samples were processed for otolith-based age estimates ( opaque edges peaked from April to July (Figure 2). This information was combined with peak spawning period of B. capriscus in SEUS waters to establish a birthdate of 1 June so that fractional ages could be computed and utilized in the growth models for this species. Based on these results, the following rules were applied for computing fractional ages: (1) for fish caught February-May, with opaque zones on the edge, fractional age = opaque zone count -((6 À month)/12); (2) for fish caught February-May with translucent zones on the edge, fractional age = (opaque zone count + 1) -((6 À month)/12); (3) for fish caught June-December with either otolith edge type, fractional age = opaque zone count + ((month À 6)/12).
A total of 665 B. capriscus otoliths had two independent age estimates which resulted in an APE of 4.7%, perfect agreement for age estimates occurred for 62% of the samples, 85% had otolith age estimates within 1 year and 95% within 2 years. Analysis of betweenreader agreement for B. veulta otolith age estimates (APE = 3.6%) was previously described in Shervette and Rivera Hernández (2022a). Ten of the SEUS B. veulta otoliths were included as part of the APE calculation for the 510 otoliths with independent age estimates from the two readers.
B. capriscus males ranged in size from 26 to 571 mm L f and in age from 0 to 17 years. Females ranged from 25 to 483 mm L f and from 0 to 20 years (Table 1). Mean size at age of B. capriscus males and females differed significantly, with males larger than females in each of the age groups analysed ( Figure 3 and combined was L f = 449 (1 À e À0.28(t+0.30) ) ( Figure 4 and Table 3).

| DISCUSSION
The results from the current study provide critical new insights into the life history of two ecologically and economically important triggerfish species. This study is the first to report on sex-specific age and growth information for B. capriscus using the Δ 14 C-validated  (Živkov et al., 1999) because of differences in the methods used to compute parameter estimates and differences in study design and sample collection methods. Therefore, we have limited our comparisons to past spinebased age estimation studies that included juvenile triggerfish collected from pelagic habitat or that included newly recruited small fish to benthic habitat and produced biologically comparable t 0 values and computed VBGF parameters from observed size-at-age data (Table 3).
For B. capriscus the maximum spine-based age from these studies was 14 years compared to a maximum otolith-based age of 21 years (Table 3). The otolith-based result extends B. capriscus longevity by approximately 30% in that this species can live up to 1.3 times longer than previously realized. The difference in longevity derived from spine-based and otolith-based ages is even more for B. vetula: otolithbased age estimation extended maximum age from 14 to 40 years, indicating that this species can live up to three times longer than previously realized.
In the current study, B. capriscus otolith-based age estimates from SEUS FD and FI samples yielded an asymptotic length (L ∞ ) of 449 mm L f [95% confidence interval (CI) 441-458] for all samples combined T A B L E 3 Comparative summary of life-history studies reporting on the von Bertalanffy growth function results for Balistes capriscus and B. vetula from spine-based age estimation and otolith-based age estimation that included small juvenile fish caught in pelagic habitat or newly recruited to benthic habitats Note: Parameter estimates computed in the current study (L ∞ , K, t 0 ) include 95% confidence intervals in parentheses. F, female; FD, fisheries-dependent; FI, fisheries-independent; K, growth coefficent; L ∞ , asymptotic size; L f , fork length; M, male; NC-FL, North Carolina to Florida; NC-SC, North Carolina to South Carolina; t 0 , theoretical age when size is zero; S.D., standard deviation. (Table 3), which was greater than the L ∞ estimates from spine-based ages from other SEUS studies (L ∞ = 382-400 mm L f ; Table 3).
Otolith-based ages resulted in a lower growth coefficient (K = 0.28, 95% CI 0.26-0.29; Table 3) than reported for spine-based SEUS studies (K = 0.63-0.67; Table 3). However, the current study obtained otolith-based age estimates from mainly FD samples from the SEUS and therefore may present a biased snapshot of growth (Taylor et al., 2005;Wilson et al., 2015)   also larger than what occurred in the GOM (387 mm L f ) using spinebased age estimates (Allman et al., 2018). The otolith-based K for females and males (0.29 and 0.27; Table 3) were lower from the current study than the GOM spine-based female and male VBGF results (0.52 and 0.55; Table 3). Otolith-based ages of long-line caught B. capriscus would be useful to obtain in future sampling efforts, as those larger fish may yield individuals that exceed the maximum age of 21 years found for B. capriscus in the current study, thereby potentially extending our understanding of longevity for this species.
B. vetula estimates of L ∞ and K exhibited a similar pattern of differences: otolith-based L ∞ (430-520 mm L f ) was larger than spinebased L ∞ (368 mm L f ; Table 3). B. vetula otolith-based age estimates from SEUS and U.S. Caribbean waters had similar K values (0.14 and 0.15, respectively) that were half of the K (0.34) reported from spinebased ages (Table 3). However, the sample size of SEUS B. vetula from the current study was small, so additional samples are needed to provide a more comprehensive understanding of the potential differences in growth parameters for this species between the two regions.
Three clear trends for both Balistes species occurred when comparing otolith-based VBGF estimates and longevity with those reported from spine-based studies: K was consistently lower, asymptotic length was larger and maximum age was greater. These new insights combined with recent Δ 14 C triggerfish age validation results, that otoliths provide accurate ages and the dorsal spine does not (Patterson et al., 2019;Shervette & Rivera Hernández, 2022a), indicate that our past understanding of basic life-history parameters for the two Balistes species was flawed. B. capriscus and B. vetula were previously described as moderately rapidly growing and relatively short-lived (Burton et al., 2015;Manooch & Drennon, 1987) compared to other large-bodied reef fishes that support fisheries in the western Atlantic, but based on our new understanding from otolithbased ages we now know that Balistes species are moderately slow to slow growing and relatively long-lived. Additional otolith-based age sampling of B. capriscus in the SEUS and GOM is needed before additional stock assessments are conducted to ensure the accuracy of age-based parameter estimates utilized in the stock assessment models. Spine-based ageing methods do not appear to produce accurate ages, underestimate age (e.g., Figure 1c (Table 3), and Hood and Johnson (1997) (Manooch & Drennon, 1987 Some triggerfish species may be capable of plasticity in growth that is responsive to anthropogenic and environmental factors (Shervette et al., 2021). In the Gulf of Guinea, during a period of climatic shifts in oceanic attributes, the B. capriscus population experienced a rapid increase in abundance over a relatively short time span of 1972of -1983of (Caverivière et al., 1980Gerlotto, 2017). The rapid increase in abundance was correlated with an expansion of favourable environmental conditions for triggerfish from its normal benthic habitat to an additional suitable pelagic habitat (Gerlotto, 2017). For a fish population to expand from an estimated regional biomass of <1 t in 1972 to over 1,000,000 t by 1978, an increase in fish growth rate would seem to be a potentially important contributing factor (Shervette et al., 2021) combined with increased reproductive success.
If growth in B. capriscus is relatively plastic, then it may vary temporarily and spatially in other regions of its range. Efforts to obtain otolithbased age data for B. capriscus from contingents and populations across its range will be necessary to better understand the potential growth plasticity in this species.  (Gladstone, 1994) in that they are relatively large-bodied reef-associated species that produce large amounts of eggs (Gladstone, 1994;Ingram, 2001;Kuwamura, 1997), but are benthic nesters that invest a large amount of energy in spawning territory defence and in brood care of their fertilized eggs (Fricke, 1980;Gladstone, 1994;Lobel & Johannes, 1980) to ensure that eggs successfully hatch into larvae that then move on to planktonic habitat (Kuwamura, 1997). Several studies on mating behaviour in triggerfishes have observed that the males defending nesting territories are larger than the females nesting within a territory (Fricke, 1980;Gladstone, 1994;Kuwamura, 1997;Seki et al., 2009;Simmons & Szedlmayer, 2012). A larger size for males may enhance their ability to successfully defend higher quality nesting territories from conspecifics and also enhance their success in attracting and mating with more females (Gladstone, 1994;Seki et al., 2009).
Mature females across a range of sizes utilize nests within a male territory and defend developing eggs from potential predators (Fricke, 1980;Gladstone, 1994;Kuwamura, 1997;Simmons & Szedlmayer, 2012). Female triggerfish invest a substantial amount of energy during their spawning season in nest preparation and maintenance, mating, tending to the fertilized eggs by fanning and blowing on them, and defending eggs from predators (Clark et al., 2015;Fricke, 1980;Gladstone, 1994;Kuwamura, 1997;Simmons & Szedlmayer, 2012). Females of several species do not appear to forage or exhibit reduced foraging efforts while caring for fertilized eggs compared to the effort they spend foraging outside of the nesting period (Fricke, 1980;Gladstone, 1994;Kuwamura, 1997). Females of several triggerfish species spawn multiple batches of eggs within a reproductive period (Gladstone, 1994;Kelly-Stormer et al., 2017;Kuwamura, 1997;Rivera Hernández et al., 2019;Seki et al., 2009).
The investment of energy by females into these reproductive activities for multiple broads each spawning season combined with reduced intake of food during that time may partially explain why females tend to be smaller than males since they are investing large amounts of energy in reproduction efforts and less energy during the spawning season in somatic growth. Further evidence of this substantial energy investment in reproduction and less in somatic growth for females from the two Balistes species comes from the maximum age results documented in the current study. The oldest B. capriscus in our study of known sex (20 years) was a 448 mm L f female and the oldest B. vetula (40 years) was a 466 mm L f female. Both of these females were in the upper size range for females of their species but were much smaller than the largest males from our study. King and McFarlane (2003)