The selective drivers of allometry in sharks (Chondrichthyes: Elasmobranchii)

Abstract

In addition to the selective importance of interspecific morphological variation, ontogenetic morphological variation may reflect different selective regimes to which successive developmental stages are subjected. The typical body form of carcharhiniform sharks is considered relatively conserved, yet sharks exhibit a wide range of body sizes and shapes, representing adaptations to distinct ecological niches. Previous investigations of ontogenetic shifts in shark body form have provided evidence for both isometric and allometric changes, depending on the morphological characters and species investigated. These findings have led to suggestions of a relationship between body size and allometric growth in sharks. In this study we present evidence of ontogenetic allometric shifts in two species of carcharhiniform sharks (Sphyrna lewini and Rhizoprionodon longurio) from novel measurements. Our results are generally consistent with previous suggestions of body form conservatism across shark phylogeny, yet also suggest potential selective factors underlying observed instances of ontogenetic allometric shifts, and highlight where additional studies are required. We propose the ‘allometric niche shift’ hypothesis for interspecific differences in scaling trends, suggesting that long-distance movements and ontogenetic trophic niche shifts represent key drivers of allometry in sharks.

ecomorphology, elasmobranchs, morphology, morphometrics, natural selection, phylogenetic allometry

INTRODUCTION

The central goal of ecomorphology is to understand the relationship between morphological and ecological variation amongst taxonomic units, which may be individuals, populations, species or higher taxa (Leisler & Winkler, 1985; Barr, 2018). Morphology is shaped by natural selection (Spence et al., 2013), such that the study of comparative morphology may shed light on selective pressures operating throughout the evolutionary history of a lineage. A plethora of studies have investigated the evolutionary history of specific morphological traits and their ecological correlates in marine vertebrates (e.g. Pyenson et al., 2013; Spence et al., 2013), with most investigations targeting the evolutionary significance of interspecific morphological variation. However, vertebrate morphology also varies with ontogeny, and patterns of ontogenetic variation may be related to alternative selective regimes to which different developmental stages are subjected (Eggold & Motta, 1992; Muir et al., 2013). Thus, an ecomorphological approach can conceivably improve our understanding of the selective underpinnings of morphological changes across phylogeny, as well as of growth patterns throughout the ontogeny of an individual.

Elasmobranchii is a clade consisting of over 1100 species (Weigmann, 2016). Crown group elasmobranchs have a long evolutionary history, having diverged during the Silurian (Heinicke et al., 2009), and remain ecologically important components of marine and some freshwater communities to the present day (Heithaus et al., 2010; Navia et al., 2017). The evolutionary ‘success’ of sharks has previously been partially attributed to conservatism in their body form (suggesting that all sharks possess one of a small number of generalized body forms) and locomotor function (Thomson & Simanek, 1977). Despite these notions, it is clear that sharks exhibit a wide range of body sizes and shapes, representing adaptations to a range of distinct ecological niches (Irschick et al., 2017; Sternes & Shimada, 2020). Different shark species employ alternative locomotor strategies, facilitated by interspecific variation in the shape, size and physiology of body appendages (Wilga & Lauder, 2002; Gleiss et al., 2011; Iosilevskii & Papastamatiou, 2016; Maia et al., 2017). Ontogenetic morphometry has also been studied in some sharks, providing evidence of both allometric and isometric growth, depending on the species under investigation and the morphological traits in question (Lingham-Soliar, 2005a; Reiss & Bonnan, 2010; Irschick & Hammerschlag, 2015; Fu et al., 2016; Ahnelt et al., 2020; Sternes & Higham, 2022). Some studies have suggested a relationship between ontogenetic scaling patterns and body size, with smaller-bodied taxa likely to grow isometrically, and larger species more likely to exhibit some degree of allometric growth (Irschick et al., 2017; Ahnelt et al., 2020).

Studies of ontogenetic shifts in shark body form have as yet been restricted to a relatively small proportion (< 2%) of extant shark diversity. Moreover, previous studies have generally relied on limited sample sizes (Fu et al., 2016), or data deriving exclusively from museum specimens (Reiss & Bonnan, 2010; Sternes & Higham, 2022). Further studies of ontogenetic morphological scaling in a range of shark species are, therefore, warranted, in order to reveal broader phylogenetic patterns of body form evolution among species and higher clades.

Here, we investigate ontogenetic allometry in two species of carchariniform sharks, drawing on a large, novel dataset of wild-caught individuals. The first species, the scalloped hammerhead shark, Sphyrna lewini (E.Griffith & C.H.Smith, 1834), is distributed globally in temperate and tropical waters (Compagno, 1984). Neonate and juveniles are thought to use coastal nearshore habitats and mangroves as nursery grounds (Duncan & Holland, 2006), which provide protection from larger predators and offer a greater abundance of prey that can be located and handled successfully by juveniles (Simpfendorfer & Milward, 1993). Trophic niche studies using stable isotope-based techniques have led to the categorization of S. lewini as a tertiary consumer (Duncan & Holland, 2006; Flores-Martínez et al., 2017). However, this species also undergoes an ontogenetic trophic niche shift (Estupiñán-Montaño et al., 2021). This shift is probably influenced by an ontogenetic shift in habitat use, with subadults and adults undergoing long-distance migrations in pelagic environments (Bessudo et al., 2011; Hoyos-Padilla et al., 2014). Sternes & Higham (2022) reported evidence of several allometric shifts in body shape in S. lewini, and provided potential ecomorphological explanations for these shifts. However, our dataset is considerably larger than that of Sternes & Higham (2022) and relies solely on measurements from freshly caught individuals as opposed to preserved museum specimens. Therefore, we assess the conclusions of Sternes & Higham (2022) in light of our expanded dataset and comment on potential ecomorphological explanations underlying discrepancies between our results and theirs.

We also investigated ontogenetic morphometric changes in the Pacific sharpnose shark, Rhizoprionodon longurio (Jordan & Gilbert, 1882), a species inhabiting coastal waters of the eastern tropical Pacific (Compagno, 1984). Like Sphyrna lewini, R. longurio is thought to also exhibit migratory behaviour (Kato & Carvallo, 1967; Márquez-Farias et al., 2005). Although Alatorre-Ramirez et al. (2013) have suggested that trophic level does not vary substantially with ontogenetic stage in R. longurio, a more recent study with a larger sample size reported significant differences between the trophic levels of neonates and those of larger juveniles and sexually mature adults (Trejo-Ramírez, 2017).

Both habitat use and dietary preferences may impart strong selective pressures on morphology, making these species ideal candidates for shedding light on potential correspondence between these factors and ontogenetic shifts in morphometric scaling. Moreover, there is a marked size difference between these two species, with S. lewini typically reaching substantially greater length and mass than R. longurio (Compagno, 1984). In light of the aforementioned hypothesized relationship between body size and ontogenetic scaling (Irschick et al., 2017; Ahnelt et al., 2020), the inclusion of taxa varying in maximum body size is important. Additionally, S. lewini and R. longurio are categorized by the IUCN as critically endangered and vulnerable, respectively (Rigby et al., 2019; Pollom et al., 2019); as such, further studies into their growth patterns are vital to help inform local management plans and conservation efforts.

We generated a large-scale morphometric dataset covering an unprecedented range of ontogenetic stages for these two species of carchariniform sharks. We also investigated within-cohort patterns of morphometric variation by analysing particular ontogenetic stages individually, and interpret our results in light of potential ecological drivers of allometric growth patterns that are supported by the literature. Furthermore, we combine these data with results from a previous multispecies study to investigate morphological changes in sharks across elasmobranch phylogeny. This study greatly increases the volume of available data on scaling trends in Sphyrna lewini and provides the first such data for Rhizoprionodon longurio, increasing the phylogenetic diversity of sharks for which detailed information on ontogenetic morphometry is available, and highlighting specific areas where further studies are required.

MATERIAL AND METHODS

Ethics statement

Data collection and analysis procedures in this study complied with national animal welfare laws; guidelines and policies were authorized by Mexican wildlife authorities under the permit PPF/DGOPA-024/20 provided by the Comisión Nacional de Acuacultura y Pesca (CONAPESCA). At no point did any participants condone, promote or encourage the harvesting of sharks.

Data collection

Data were collected from an artisanal fishing camp (El Saladito) in Baja California Sur, Mexico (Fig. 1) between 2 December 2020 and 3 March 2021. A total of 254 sharks were measured, including 129 S. lewini individuals (68 females and 61 males) and 125 R. longurio individuals (79 females, 42 males and four neonates for which sex was not determined). The two species were distinguished on the basis of gross morphological differences, and the sex of individuals was determined by the presence or absence of external claspers (male intromittent organs). Following the approach of Irschick & Hammerschlag (2015), 13 morphological measurements were taken from each individual (Fig. 2; Table 1), as well as an additional 14^th measurement in the case of S. lewini (EE, Fig. 2; Table 1). All measurements were taken in centimetres using a tape measure and recorded to the nearest 0.5 cm (apart from neonates, where measurements were recorded to 0.1 cm). This discrepancy was due to the fact that non-neonate individuals were to be sold for human consumption, so the duration of sampling had to be minimized to prevent the meat from spoiling. All measurements were taken immediately after the sharks were landed to minimize error that could result from post-mortem morphological changes. Total length (TL, Fig. 2; Table 1) was excluded from statistical analyses as it is considered redundant given the use of precaudal length (PL, Fig. 2; Table 1). However, TL values were still collected as they were necessary for the determination of niche shifts in S. lewini.

Table 1.

Morphometric measurements, including descriptions and their associated abbreviations

Abbreviation	Measurement	Morphological description
TL	Total length	Distance from the tip of the snout to the dorsal tip of the caudal fin
PL	Precaudal length	Distance from the tip of the snout to the precaudal pit (caudal peduncle)
LS	Lateral span	Distance spanning from the insertion point of the anterior edge of one pectoral fin to the same point on the other pectoral fin
FS	Frontal span	Span across the body from the horizontal plane of one pectoral fin to the horizontal plane of the other, crossing the midline of the body at the anterior edge of the first dorsal fin
PS	Proximal span	Span across the body from the horizontal plane of one pectoral fin to the horizontal plane of the other, crossing the midline of the body at the posterior edge of the first dorsal fin
KC	Caudal keel circumference	Total circumference at the base of the tail, as measured at the caudal keel
DH	Dorsal fin height	Vertical distance from the tip of the first dorsal fin to the base of the first dorsal fin
DW	Dorsal fin width	Distance horizontally along the body between the anterior and posterior insertion points of the first dorsal fin
DL	Dorsal fin length	Distance from the anterior insertion point of the first dorsal in to the tip of the first dorsal fin
UL	Upper caudal lobe	Distance from the dorsal insertion of the caudal fin to the dorsal tip of the caudal fin
LL	Lower caudal lobe	Distance from the ventral insertion of the caudal fin to the ventral tip of the caudal fin
CH	Caudal height	Distance from the dorsal tip of the caudal fin to the ventral tip of the caudal fin
PF	Pectoral fin length	Distance from the distal insertion point of the pectoral fin to the tip of the fully extended pectoral fin
EE	Cephalofoil diameter	Lateral diameter of dorsal surface cephalofoil measured from midpoint the left eye to the midpoint of the right eye

Figure 1.

Map of the study site. Geographical labels on the inset refer to states of Mexico: Baja California Norte (BCN), Baja California Sur (BCS), Sonora (SO), Sinaloa (SI), Chihuahua (CH), Durango (D).

Figure 2.

Diagrams of R. longurio (A) in lateral view and S. lewini (B) in lateral view, with morphological measurements taken during data collection (abbreviations correspond to those found in Table 1). Figure produced using original illustrations from Oliver Demuth, used with permission.

Niche shift calculations

To test for different growth trajectories between ontogenetic stages, the precaudal lengths at which ontogenetic dietary shifts (niche shifts) were reported in previous studies were determined for S. lewini. This was accomplished by entering the ages at which successive ontogenetic dietary shifts were reported (Estupiñán-Montaño et al., 2021) into the Von Bertalanffy growth function (Von Bertalanffy, 1938), using parameters from Anislado-Tolentino et al. (2008). These data had already been reported for R. longurio by Trejo-Ramírez (2017), and these are presented in the results section of this study for ease of comparison. Ontogenetic stages in S. lewini were categorized as Class One (younger juveniles), Class Two (older juveniles) and Class Three (adult), whereas R. longurio ontogenetic stages were characterized as Class One (neonate), Class Two (juvenile) and Class Three (adult).

Principal component analyses

To visualize morphological variation both between species and between ontogenetic stages prior to conducting statistical analyses, principal component analyses (PCAs) were carried out in R (R Core Team, 2020) using the packages factoextra (Kassambara & Mundt, 2020; R Core Team, 2020) and ggplot2 (Wickham et al., 2016; R Core Team, 2020). The first three principal components were calculated for the total dataset and for each species separately, so that morphological variation could be visualized between the two species and between ontogenetic stages. The complete dataset used for principal component analyses can be found in the Supporting Information, Table S1, and PCA plots can be found in the Supporting Information, Figures S1 and S2.

Linear regression analyses

To test for ontogenetic allometry, linear regression analyses of each measurement against PL were carried out using the R package ggplot (Wickham & Wickham, 2007; R Core Team, 2020). As per Irschick & Hammerschlag (2015), all data were log10 transformed prior to subsequent analyses. Regressions were then performed separately on the full S. lewini and R. longurio datasets, after which separate regressions were performed for S. lewini Classes One and Two (Class Three consisted only of a single data point), and R. longurio Classes Two and Three (Class One consisted of only four data points). In the case of R. longurio, a regression analysis combining Classes Two and Three was also performed, on account of the PCA results (Supporting Information, Figs S1, S2). Previous studies performed regressions only on total datasets, and hence our methodology (combined with our large sample size) provides finer-scale results that enable us to discern between scaling trends in discrete ontogenetic stages that may correspond to different selective regimes. The complete dataset used for linear regression analyses can be found in the Supporting Information (Table S2).

Phylogenetic generalized least squares regression analysis

To test for phylogenetic trends in body form, a phylogenetic generalized least squares (PGLS) analysis was carried out. First, mean trait data were collated from from Irschick et al. (2017) and combined with mean trait data for the S. lewini and R. longurio (Supporting Information, Table S3). Because Irschick et al. (2017) used only adult and subadult individuals, S. lewini Class One, and R. longurio Classes One and Two individuals were excluded. The taxa incorporated into this study from Irschick et al. (2017) were as follows: blacknose shark [Carcharhinus acronotus (Poey, 1860)], bull shark [Carcharhinus leucas (J.P.Muller & Henle, 1839)], blacktip shark [Carcharhinus limbatus (J.P.Muller & Henle, 1839)], sandbar shark [Carcharhinus plumbeus (Nardo, 1827)], lemon shark [Negaprion brevirostris (Poey, 1868], tiger shark (Galeocerdo cuvier Péron & Lesueur, 1822), nurse shark [Ginglymostoma cirratum (Bonnaterre, 1788)] and the Atlantic sharpnose shark [Rhizoprionodon terraenovae (J.Richardson, 1836)]. All data were log10 transformed, and a pruned, time-scaled phylogeny (Fig. 3) was generated in MESQUITE (Maddison & Maddison, 2021) using the topology and branch lengths of Stein et al. (2018). To determine the model of trait evolution explaining the greatest proportion of observed variation (and hence which model should be used to perform the phylogenetic correction in the PGLS analysis) the R packages phytools (Revell, 2012; R Core Team, 2020) and mvMORPH (Clavel et al., 2015; R Core Team, 2020) were used. Table 2 shows that for each measurement, a Brownian motion model of trait evolution was most appropriate. Finally, PGLS was carried out on data for each trait that had coverage for all ten species (Supporting Information, Tables S3, S4) using the R package ape (Paradis & Schliep, 2019; R Core Team, 2020), and a Brownian motion correction as per the results of the model test (Table 2). The full original and log10-transformed datasets used for PGLS analysis are presented in the Supporting Information (Tables S3, S4, respectively).

Table 2.

Trait evolution model test, comparing log-likelihood, Akaike information criterion (AIC) and corrected Akaike information criterion (AICc) values for Brownian motion (BM), early burst (EB) and Ornstein–Uhlenbeck (OU) models for each morphological measurement included in subsequent PGLS analysis

	Log-likelihood			AIC			AICc
Character	BM	EB	OU	BM	EB	OU	BM	EB	OU
PL	2.14	2.14	2.91	–0.29	1.71	0.17	1.43	5.71	4.17
LS	0.83	0.83	1.54	2.34	4.34	2.93	4.06	8.34	6.93
FS	1.16	1.16	2.14	1.68	3.68	1.72	3.39	7.68	5.72
PS	1.72	1.72	2.71	0.56	2.56	0.59	2.27	6.56	4.59
KC	1.91	1.91	2.56	0.18	2.18	0.88	1.90	6.18	4.88
DH	1.13	1.13	2.52	1.74	3.74	0.97	3.45	7.74	4.97
DW	1.08	1.08	2.99	1.84	3.84	0.03	3.55	7.84	4.03
DL	0.96	0.96	2.12	2.09	4.09	1.77	3.80	8.09	5.77
UL	2.27	2.27	3.09	–0.55	1.45	–0.17	1.17	5.45	3.83
LL	1.37	1.37	2.56	1.26	3.26	0.87	2.98	7.26	4.87
CH	1.37	1.37	2.56	1.26	3.26	0.87	2.98	7.26	4.87
PF	–0.38	–0.38	0.98	4.76	6.76	4.04	6.47	10.76	8.04

Figure 3.

Time-scaled molecular phylogeny of ten shark species produced in MESQUITE, utilising topology and branch lengths from Stein et al. (2018), with branch lengths representing evolutionary time (Myr).

RESULTS

Ontogenetic niche shifts in S. lewini and R. longurio

Niche shifts in male S. lewini were delimited at 0.0–117.8 cm (Class One, young juveniles), 117.8–171.5 cm (Class Two, older juveniles) and > 171.5 cm (Class Three, sexually mature adults; Table 3). These values differed for female S. lewini, where niche shifts were delimited at 0.0–101.9 cm (Class One, young juveniles), 101.9–151.6 cm (Class Two, older juveniles) and > 151.6 cm (Class Three, sexually mature adults; Fig. 4). Niche shifts in R. longurio were predicted to be delimited in males at 0.0–38.0 cm (Class One, neonates), 38.0–100.5 cm (Class Two, juveniles) and > 100.5 cm (Class Three, sexually mature adults; Table 3). Similarly, R. longurio female niche shifts were predicted to be delimited at 0.0–38.0 cm (Class One, neonates), 38.0–92.4 cm (Class Two, juveniles) and > 92.4 cm (Class Three, sexually mature adults; Table 3). Values for R. longurio were extracted from Trejo-Ramírez (2017). All measurements refer to total length (TL) and individuals were categorized as belonging to the relevant Class (Supporting Information, Tables S1, S2) prior to further analyses.

Table 3.

Total lengths delimiting different ontogenetic stages in each species and hence potential niche shifts

	S. lewini			R. longurio
Class	Definition	Male (cm)	Female (cm)	Definition	Male (cm)	Female (cm)
1	Young juveniles	0–117.8	0–101.9	Neonates	0–38.0	0–38.0
2	Older juveniles	101.9–171.5	101.9–151.6	Juveniles	38.0–100.5	38.0–92.4
3	Sexually mature	> 171.5	> 151.6	Sexually mature	> 100.5	> 92.4

Figure 4.

Linear regression plots for proximal span – PS (A), caudal keel circumference - KC (B), upper caudal lobe - UL (C), caudal height - CH (D) and cephalofoil diameter - EE (E) against precaudal length - PL for all S. lewini data. All data are log10 transformed, and the 95% confidence interval for the scaling coefficient is represented by the dark-grey area.

Evidence of both allometric and isometric growth in all ontogenetic stages of S. lewini and R. longurio

Sphyrna lewini (all classes)

Regression of 12 morphological measurements against PL in all S. lewini individuals recovered seven cases of isometric growth and five cases of allometric growth (Table 4). R² values varied from 0.39 (KC; Table 4) to 0.85 (EE; Table 4). The measurements PS, KC, UL, CH and EE all showed significant negative allometry (Fig. 4; Table 4).

Table 4.

Linear regression results for all S. lewini data, with significant P values in bold

Character	Coefficient	Std.error	t value	P value	Residual SE	R2	Adj. R2	F statistic
LS	1.01	0.09	0.11	0.91408	0.08	0.48	0.47	116.60
FS	0.91	0.05	–1.89	0.0608	0.04	0.73	0.73	346.10
PS	0.87	0.04	–3.05	0.00282	0.04	0.76	0.76	402.90
KC	0.78	0.09	–2.57	0.0113	0.08	0.39	0.39	81.91
DH	1.03	0.06	0.57	0.573	0.05	0.73	0.73	350.60
DW	1.12	0.07	1.68	0.0959	0.06	0.67	0.67	255.90
DL	0.92	0.05	–1.61	0.11	0.04	0.74	0.74	369.90
UL	0.80	0.05	–3.83	0.000204	0.05	0.65	0.65	234.00
LL	1.08	0.05	1.61	0.109	0.04	0.80	0.90	521.00
CH	0.80	0.05	–4.02	9.87E-05	0.04	0.66	0.66	247.70
PF	0.97	0.04	–0.92	0.358	0.03	0.84	0.84	664.00
EE	0.77	0.03	–7.78	2.17E-12	0.03	0.85	0.85	703.50

Sphyrna lewini (class-specific)

Regression of 12 morphological measurements against PL in S. lewini Class One individuals recovered five cases of isometric growth and seven cases of allometric growth (Table 5). R² values varied from 0.16 (KC; Table 5) to 0.78 (UL; Table 5). The measurements FS, PS, KC, DL, UL, PF and EE all showed significant negative allometry (Fig. 5; Table 5). Regression of 12 morphological measurements against PL in S. lewini Class Two individuals recovered five cases of isometric growth and seven cases of allometric growth (Table 6). R² values varied from 0.08 (KC and UL; Table 6) to 0.75 (EE; Table 6). The measurements LS, PS, KC, UL, LL, CH and EE all showed significant negative allometry (Fig. 6; Table 6).

Table 5.

Linear regression results for S. lewini Class One data, with significant P values in bold

Character	Coefficient	Std.error	t value	P value	Residual SE	R2	Adj. R2	F statistic
LS	0.71	0.17	–1.68	0.0981	0.08	0.20	0.18	17.46
FS	0.81	0.09	–2.07	0.042	0.04	0.54	0.53	83.10
PS	0.78	0.08	–2.69	0.00891	0.04	0.56	0.56	92.18
KC	0.63	0.17	–2.19	0.032	0.08	0.16	0.15	13.87
DH	0.85	0.08	–1.93	0.05741	0.04	0.61	0.61	113.60
DW	1.11	0.14	0.76	0.451953	0.07	0.46	0.45	61.11
DL	0.78	12.05	–3.31	0.00144	0.03	0.67	0.66	145.10
UL	0.88	0.05	–3.43	0.00118	0.03	0.78	0.78	257.00
LL	1.00	0.10	0.02	0.986	0.04	0.60	0.59	106.70
CH	0.89	0.08	–1.44	0.153	0.04	0.65	0.65	136.00
PF	0.86	0.07	–2.00	0.04928	0.03	0.66	0.66	140.80
EE	0.66	0.06	–5.50	5.56E-07	0.03	0.62	0.61	117.00

Table 6.

Linear regression results for S. lewini Class Two data, with significant P values in bold

Character	Coefficient	Std.error	t value	P value	Residual SE	R2	Adj. R2	F statistic
LS	1.66	0.23	2.84	0.00636	0.08	0.49	0.48	50.87
FS	0.86	0.12	–1.11	0.273	0.04	0.48	0.47	47.76
PS	0.79	0.10	–2.04	0.0467	0.04	0.53	0.52	59.07
KC	0.43	0.20	–2.89	0.00568	0.07	0.08	0.06	4.60
DH	1.21	0.17	1.22	0.22934	0.06	0.48	0.47	48.29
DW	1.06	0.16	0.37	0.71255	0.06	0.45	0.44	42.01
DL	1.07	0.15	0.45	0.65751	0.50	0.50	0.49	52.65
UL	0.39	0.18	–3.43	0.00118	0.06	0.08	0.06	4.67
LL	1.28	0.10	2.69	0.00963	0.04	0.75	0.74	154.90
CH	0.36	0.15	–4.38	5.87E-05	0.05	0.10	0.08	5.86
PF	0.88	0.09	–1.31	0.19631	0.03	0.65	0.64	96.69
EE	0.70	0.06	–5.31	2.32E-06	0.02	0.75	0.75	158.50

Figure 5.

Linear regression plots for frontal span - FS (A), proximal span - PS (B), caudal keel circumference - KC (C), dorsal fin height - DL (D), upper caudal lobe - UL (E), pectoral fin length - PF (F) and cephalofoil diameter - EE (G) against precaudal length - PL for S. lewini Class One data. All data are log10 transformed, and the 95% confidence interval for the scaling coefficient is represented by the dark-grey area.

Figure 6.

Linear regression plots for proximal span - PS (A), frontal span - FS (B), caudal keel circumference - KC (C), upper caudal lobe - UL (D), lower caudal lobe - LL (E) and cephalofoil diameter - EE (F) against precaudal length - PL for S. lewini Class Two data. All data are log10 transformed, and the 95% confidence interval for the scaling coefficient is represented by the dark-grey area.

Rhizoprionodon longurio (all classes)

Regression of 11 morphological measurements against PL in all R. longurio individuals recovered five cases of isometric growth and six cases of allometric growth (Table 7). R² values varied from 0.54 (KC; Table 7) to 0.96 (LS, FS, PS; Table 7). The measurements LS, FS and PS all showed significant positive allometry, whereas the measurements DW, DL and UL showed significant negative allometry (Fig. 7; Table 7).

Table 7.

Linear regression results for all R. longurio data, with significant P values in bold

Character	Coefficient	Std.error	t value	P value	Residual SE	R2	Adj. R2	F statistic
LS	1.06	0.02	2.86	0.00496	0.03	0.96	0.96	2931.00
FS	1.18	0.02	8.44	7.33E-14	0.03	0.96	0.96	3108.00
PS	1.26	0.02	10.57	<2e-16	0.03	0.96	0.96	2658.00
KC	0.95	0.08	–0.57	0.569	0.11	0.54	0.54	146.70
DH	1.05	0.03	1.50	0.136	0.05	0.89	0.89	992.80
DW	0.87	0.05	–2.63	0.00954	0.07	0.72	0.72	321.00
DL	0.94	0.02	–2.60	0.0104	0.03	0.94	0.94	1876.00
UL	0.86	0.02	–7.91	1.27E-12	0.02	0.95	0.95	2327.00
LL	0.96	0.04	–0.94	0.349	0.05	0.83	0.83	587.10
CH	0.93	0.04	–1.88	0.0629	0.05	0.85	0.85	702.20
PF	1.02	0.03	0.64	0.521	0.03	0.93	0.93	1608.00

Figure 7.

Linear regression plots for LS (A), FS (B), PS (C), DW (D), DL (E) and UL (F) against PL for all R. longurio data. All data are log10 transformed, and the 95% confidence interval for the scaling coefficient is represented by the dark-grey area. Lateral span - LS (A), frontal span - FS (B), proximal span - PS (C), dorsal fin width - DW (D), dorsal fin length - DL (E) and upper caudal lobe - UL (F) against precaudal length - PL for all R. longurio data. All data are log10 transformed, and the 95% confidence interval for the scaling coefficient is represented by the dark-grey area.

Rhizoprionodon longurio (neonates excluded)

Regression of 11 morphological measurements against PL in all R. longurio individuals from Classes Two and Three recovered ten cases of isometric growth and one case of allometric growth (Table 8). R² values varied from 0.19 (KC; Table 8) to 0.81 (FS; Table 8). The measurement UL showed significant negative allometry (Fig. 8a; Table 8).

Table 8.

Linear regression results for all R. longurio data excluding neonates, with significant P value in bold

Character	Coefficient	Std.error	t value	P value	Residual SE	R2	Adj. R2	F statistic
LS	0.93	0.04	–1.49	0.14	0.03	0.79	0.79	441.70
FS	1.01	0.05	0.21	0.831	0.03	0.79	0.79	453.60
PS	0.98	0.05	–0.42	0.678	0.03	0.76	0.76	378.10
KC	0.68	0.19	–1.70	0.0925	0.11	0.10	0.09	13.14
DH	1.12	0.08	1.53	0.128	0.04	0.63	0.63	205.50
DW	0.99	0.12	–0.06	0.9516	0.07	0.38	0.37	72.95
DL	0.90	0.05	–1.91	0.0582	0.03	0.72	0.71	300.40
UL	0.84	0.04	–3.92	0.000146	0.02	0.77	0.77	399.70
LL	1.07	0.09	0.76	0.448	0.05	0.53	0.52	132.50
CH	1.06	0.08	0.73	0.469	0.05	0.58	0.58	164.70
PF	1.11	0.06	1.84	0.0678	0.03	0.75	0.74	347.80

Figure 8.

Linear regression plots for upper caudal lobe - UL in R. longurio following exclusion of neonates (A), dorsal fin length - DL in R. longurio Class Two (B) and upper caudal lobe - UL in R. longurio Class Three (C) against precaudal length - PL. All data are log10 transformed and the 95% confidence interval for the scaling coefficient is represented by the dark-grey area.

Rhizoprionodon longurio (class-specific)

Regression of 11 morphological measurements against PL in all R. longurio individuals from Class Two recovered ten cases of isometric growth and one case of allometric growth (Table 9). R² values varied from 0.19 (KC; Table 10) to 0.81 (FS; Table 9). The measurement DL showed significant negative allometry (Fig. 8b; Table 9). Regression of 11 morphological measurements against PL in all R. longurio individuals from Class Three recovered ten cases of isometric growth and one case of allometric growth (Table 10). R² values varied from 0.02 (UL; Table 11) to 0.49 (LS; Table 10). The measurement UL showed significant negative allometry (Fig. 8c; Table 10).

Table 9.

Linear regression results for R. longurio Class Two data, with significant P value in bold

Character	Coefficient	Std.error	t value	P value	Residual SE	R2	Adj. R2	F statistic
LS	0.94	0.07	–0.82	0.4193	0.03	0.78	0.77	165.30
FS	0.92	0.06	–1.20	0.23489	0.02	0.81	0.81	205.80
PS	0.91	0.08	–1.11	0.274	0.03	0.72	0.71	122.00
KC	0.93	0.28	–0.24	0.812	0.10	0.19	0.18	11.19
DH	1.13	0.14	0.91	0.370	0.05	0.57	0.56	63.14
DW	1.17	0.20	0.84	0.40786	0.07	0.42	0.41	34.53
DL	0.83	0.08	–2.04	0.047129	0.03	0.67	0.66	96.35
UL	0.94	0.07	–0.80	0.4282	0.02	0.80	0.79	187.40
LL	1.06	0.15	0.38	0.704397	0.05	0.50	0.49	48.32
CH	1.11	0.18	0.66	0.5155	0.06	0.46	0.45	40.39
PF	1.13	0.10	1.24	0.221	0.04	0.72	0.72	124.30

Table 10.

Linear regression results for R. longurio Class Three data, with significant P value in bold

Character	Coefficient	Std.error	t value	P value	Residual SE	R2	Adj. R2	F statistic
LS	0.96	0.12	–0.32	0.75	0.03	0.49	0.48	65.74
FS	1.02	0.14	0.14	0.889	0.03	0.44	0.43	54.52
PS	0.89	0.13	–0.82	0.418	0.03	0.40	0.39	45.01
KC	0.02	0.52	–1.87	0.0651	0.11	0.00	−0.01	0.00
DH	1.02	0.19	0.10	0.92117	0.04	0.29	0.28	28.17
DW	0.90	0.30	–0.34	0.735	0.07	0.12	0.10	8.94
DL	0.84	0.14	–1.19	0.2392	0.03	0.35	0.34	36.59
UL	0.59	0.11	–3.91	0.000212	0.02	0.31	0.30	31.17
LL	1.02	0.25	0.08	0.9407	0.05	0.19	0.18	16.69
CH	1.03	0.16	0.20	0.8386	0.04	0.37	0.36	40.17
PF	0.91	0.15	–0.57	0.5723	0.03	0.34	0.33	35.43

Table 11.

PGLS results for ten shark species, obtained using a Brownian motion model of trait evolution, and incorporating data extracted from Irschick et al. (2017), with significant P value in bold

Character	Coefficient	Std. error	t value	P value	Residual SE	DF	AIC	BIC
LS	1.26	0.10	2.54	0.029	0.09	10	–23.81	–22.90
FS	1.23	0.13	1.86	0.093	0.11	10	–19.26	–18.35
PS	0.97	0.16	–0.19	0.853	0.14	10	–14.58	–13.67
KC	1.07	0.05	1.35	0.207	0.05	10	–36.36	–35.45
DH	1.02	0.19	0.09	0.930	0.16	10	–11.53	–10.62
DW	1.05	0.22	0.25	0.808	0.19	10	–8.05	–7.14
DL	1.19	0.12	1.53	0.157	0.11	10	–19.48	–18.57
UL	0.92	0.16	–0.47	0.648	0.14	10	–14.31	–13.40
LL	1.12	0.18	0.70	0.500	0.15	10	–12.61	–11.70
CH	0.99	0.15	−0.04	0.969	0.13	10	–16.21	–15.30
PF	1.29	0.18	1.59	0.143	0.16	10	–12.02	–11.11

phylogenetic generalized least squares (PGLS)

Phylogenetic generalized least squares regression against PL of mean trait values for both S. lewini and R. longurio, together with the eight taxa studied by Irschick et al. (2017), yielded statistical evidence of interspecific allometry in LS (Table 11). All other measurements did not differ significantly from isometry.

DISCUSSION

Negative caudal fin allometry in S. lewini and R. longurio

We observed that both S. lewini and R. longurio exhibit strong negative allometry in the length of the dorsal (but not ventral) caudal lobe of their caudal fins (Figs 4, 7; Tables 4, 7), suggesting that in both species the caudal fin becomes less heterocercal (more symmetrical) through ontogeny. Caudal morphology is thought to be under strong selection due to its influence on important biological functions such as ram ventilation, reproduction and interspecific interactions (Kim et al., 2013), and similar negative caudal allometry has been recorded in other species (Lingham-Soliar, 2005a; Irschick & Hammerschlag, 2015; Ahnelt et al., 2020), as well as in museum specimens of S. lewini (Sternes & Higham, 2022). There is some degree of conservatism in the function of the heterocercal shark tail (Maia et al., 2012; Maia & Wilga, 2016). Functional conservatism is supported by morphology as well as biomechanical studies, with Sternes & Shimada (2020) finding that the majority of sharks share a common caudal fin ‘design’, despite the obvious morphological specializations of some taxa. However, biomechanical investigations of a limited range of species have suggested that caudal fin flexibility and heterocercal tail angle are key variables affecting heterocercal tail function (Ferry & Lauder, 1996; Flammang et al., 2011; Maia et al., 2012). Sharks with a higher heterocercal angle (more symmetrical caudal fin) are thought to direct water more horizontally during caudal fin-propelled swimming (Irschick et al., 2017), which may be selectively beneficial for taxa undertaking long-distance migrations (Irschick et al., 2017; Ahnelt et al., 2020). Moreover, given that sharks do not possess a gas bladder (Iosilevskii & Papastamatiou, 2016), increasing symmetry of the caudal fin through ontogeny is thought to compensate for an otherwise decreasing lift/drag ratio (Ferrón et al., 2017; Ahnelt et al., 2020). This is consistent both with previous studies of ontogenetic morphometry (Lingham-Soliar, 2005a; Irschick & Hammerschlag, 2015; Fu et al., 2016; Ahnelt et al., 2020) and with our results. Adults of both S. lewini and R. longurio are known to exhibit migratory behaviour (Márquez-Farias et al., 2005; Hoyos-Padilla et al., 2014), with subadult S. lewini commonly inhabiting the pelagic realm (Hoyos-Padilla et al., 2014). Hence, negative caudal allometry in both species may represent an adaptation to long-distance migration and/or foraging on relatively large-bodied prey in the pelagic realm. These results, in the case of S. lewini, are consistent with those found by Sternes & Higham (2022), but in this case the authors recovered positive allometry of the lower caudal lobe as opposed to negative allometry of the upper caudal lobe (Sternes & Higham, 2022). Both regimes suggest a trend towards increasing symmetry of the caudal fin, but further studies should examine whether the distinction between them is of any functional significance.

Biomechanical studies also shed light on to why some species or ontogenetic stages may benefit from a relatively heterocercal or asymmetrical caudal fin. Wilga & Lauder (2002) used flow-visualization experiments to demonstrate that sharks exhibiting a typical heterocercal tail were incapable of vector thrusting through caudal fin-mediated alterations to the orientation of the vortex wake, contrasting with results reported from white sturgeons Acipenser transmontanus (J.Richardson, 1836) (Liao & Lauder, 2000). Given that sturgeons have a relatively flexible and heterocercal caudal fin in comparison with most carcharhiniform sharks (Liao & Lauder, 2000), these findings suggest that sharks with particularly heterocercal or flexible tails may be able to manoeuvre and alter body pitch more effectively than those with lower heterocercal tail angles (Irschick et al., 2017).

In light of the literature regarding ontogenetic migration (Hoyos-Padilla et al., 2014) and trophic niche shifts (Estupiñán-Montaño et al., 2021), negative caudal allometry across all measured ontogenetic stages of S. lewini (Figs 5, 6; Tables 5, 6) suggests that a relatively heterocercal tail in early life could be favoured for enhanced manoeuvrability – perhaps related to foraging for smaller, more agile prey and escaping increased predation pressure at small body size – whereas a more symmetrical tail could be favoured later in life, related to long-distance migrations and foraging on larger-bodied prey, as initially suggested by Sternes & Higham (2022). The fact that negative allometry was observed in the total dataset, as well as in both Classes One and Two, further suggests that this allometric pattern is pervasive throughout development, and is not restricted to particular ontogenetic stages. Equivalent negative caudal allometry was also reported in sexually mature R. longurio individuals (Fig. 8c; Table 10), but this was not observed in juveniles (Table 9). Whilst we did not perform linear regressions on neonates alone due to their small sample size, following omission of neonates, support for allometric trends in many measurements was lost (Figs 7, 8; Tables 7, 8). Whilst UL was not one of these measurements (Table 8), we hypothesize that if data from more neonates were obtained, this Class would exhibit significant negative allometry in UL, and that two discrete phases of caudal allometry may exist in this species. The latter of these phases is probably related to ontogenetic shifts in habitat use. Rhizoprionodon longurio reportedly undertakes annual migrations to warmer waters (Kato & Carvallo, 1967; Márquez-Farias et al., 2005). However, this phenomenon is poorly studied, and our understanding of it derives purely from fisheries’ data (Márquez-Farias et al., 2005) and the tagging of a limited number of adults (Kato & Carvallo, 1967). Given that juvenile elasmobranchs are thought to utilize nursery areas in part due to their lower relative predation pressure (Simpfendorfer & Milward, 1993), it is likely that only sexually mature adults, and potentially subadults, take part in these long-distance migrations. There is also no evidence of significant differences in trophic niche between juvenile and adult R. longurio (Alatorre-Ramirez et al., 2013; Trejo-Ramírez, 2017). We suggest that in adult R. longurio, caudal fin allometry improves hydrodynamic locomotor efficiency for long-distance swimming. If it is only adults that take part in annual migrations, this hypothesis is insufficient to explain the putative negative allometry in neonate individuals. Rather, given that neonate and juvenile R. longurio differ in trophic niche (Trejo-Ramírez, 2017), we suggest that this initial phase of allometric growth results from selection for enhanced manoeuvrability, which in turn improves foraging efficiency (and potentially predator avoidance). Thus, whilst selective pressures associated both with habitat use and trophic level appear to drive caudal allometry consistently through ontogeny in S. lewini, such consistency is absent in R. longurio, where different phases of allometry appear to be driven by different selective pressures.

Predominance of isometric pectoral fin growth in S. lewini and R. longurio

Whilst we recovered many cases of negative caudal allometry, allometry was detected in PF only in a single Class (S. lewini Class One) in which negative allometry is present (Fig. 5; Table 5). There is debate surrounding the function of pectoral fins in sharks (Wilga & Lauder, 2004). On the basis of positive angles of attack recovered in flow-tank experiments, some authors have suggested that pectoral fins act to produce lift in the anterior body, counterbalancing lift generated by the caudal fin and thus maintaining trim (Fish & Shannahan, 2000). Alternatively, several studies have highlighted a potential role in manoeuvring torques and facilitating depth changes in the water column (Wilga & Lauder, 1999, 2001). Given these important biomechanical functions, it may seem unexpected that allometry in pectoral fin length is not more widespread. However, we note that only a single measurement of pectoral fin morphology was investigated in this study (Fig. 2; Table 1), and this may be insufficient to capture the full extent of pectoral fin variation. Moreover, as this result does not match that of Sternes & Higham (2022), who recovered positive allometry in pectoral fin length, additional studies are warranted, particularly given the differences in sample size and data source between these studies. Additionally, pectoral fins are thought to play a role in copulation (Pratt & Carrier, 2001), and hence pectoral fin morphology may not relate exclusively to selection for locomotor function, but may also be influenced by sexually antagonistic coevolution (with the relative importance of these selective pressures probably varying between taxa).

Negative dorsal fin allometry in S. lewini and R. longurio

We find evidence of negative dorsal length (DL) allometry in R. longurio juveniles (Fig. 8b; Table 9), negative dorsal length allometry in young S. lewini juveniles (Fig. 5; Table 5) and negative allometry in dorsal length (DL) and width (DW) in our total R. longurio data, such that the dorsal fin becomes taller and narrower as precaudal length increases. Harris (1936) hypothesized that dorsal fins function as stabilizers in sharks. Histological analysis of the relatively tall dorsal fins of Carcharodon carcharias support this hypothesis, such that dermal fibre stiffness in the dorsal fin varies with swimming velocity (Lingham-Soliar, 2005b). However, more recent experimental studies have suggested that, whilst the dorsal fins of other species, including the spiny dogfish Squalus acanthias Linnaeus, 1758, do indeed provide a stabilizing function, those of the bamboo shark Chiloscyllium plagiosum (Bennett, 1830) contribute to thrust generation (Maia et al., 2017). Experimental studies of dorsal fin function in elasmobranchs are restricted to these two taxa. Hence it is difficult to draw any meaningful ecomorphological inferences from our regression data alone. A similar result was recovered by Sternes & Higham (2022) for S. lewini, leading these authors to suggest that increased dorsal fin height facilitates ‘side-swimming behaviour’ (Sternes & Higham, 2022), which occurs when individuals swim at a rolled angle of 90° (Royer et al., 2020). Previously, it had been suggested that only subadults and adults engage in this behaviour (Royer et al., 2020), and our finding of allometry only in S. lewini Class One (and not Class Two or the total dataset) is consistent with this idea, as it would imply that whilst there are differences in locomotor function of the dorsal fin between young juveniles and older juveniles, no significant differences are seen between older juveniles and sexually mature adults. Given that R. longurio has not been observed engaging in side-swimming behaviour, we suggest that additional factors must be of relevance in this taxon, and future studies are required to ascertain the extent to which the dorsal fin of R. longurio functions in stabilization or thrust generation.

Contrasting growth trajectories in girth measurements in S. lewini and R. longurio

Body condition is an important biological parameter that reflects energy stores of an individual, and can be estimated using multiple girth measurements (Gallagher et al., 2014). Sphyrna lewini shows several cases of negative allometry in girth measurements (Figs 4–6; Tables 4–6), while R. longurio exhibits positive allometry in several girth measurements (Fig. 7; Table 7), although these allometric trends are lost when neonates of R. longurio are excluded from regression analysis (Table 8). The positive allometry observed in R. longurio girth measurements may be explained by a significant positive allometry in liver volume, which is thought to be a common pattern amongst sharks (Gleiss et al., 2017) as they lack the gas bladders of teleosts (Iosilevskii & Papastamatiou, 2016). Thus, the hydrodynamic lift generated by a less heterocercal tail in larger R. longurio individuals may be offset by the enhanced buoyancy conferred by an enlarged liver, providing hydrostatic lift (Gleiss et al., 2017). Evidence of allometry in girth measurements was lost following the removal of neonates, providing further evidence for an initial period of allometric growth early in life (the selective drivers of which may differ from those underlying allometry between subsequent ontogenetic stages), as hypothesized above. Sphyrna lewini differs dramatically from R. longurio with regard to growth trends related to body condition, as several girth measurements show significant negative allometry regardless of Class (Figs 4–6; Tables 4–6). These results are generally consistent with those found by Sternes & Higham (2022), although in that study only lateral span exhibited negative allometry, whereas our results suggest that different girth measurements exhibit allometric growth at different ontogenetic stages (Figs 4–6; Tables 4–6). Negative allometry in these measurements may represent a depletion in energy reserves following long-distance migration, given that all sampled adults were caught in a nearshore coastal environment, rather than a deep-water pelagic environment. Similar results have been obtained in species that undertake annual migrations (Brett, 1979; Beamish et al., 1996), with the lowest energetic condition observed among individuals upon arrival to coastal environments following migration. It should be noted that in some other elasmobranch species the opposite trend has been observed (Craik, 1978; Rossouw, 1987). Sternes & Higham (2022) suggest that negative allometry in girth may represent an adaptive trend as a result of shifts in functional demand associated with drag reduction. Alternatively, given that S. lewini is critically endangered (Rigby et al., 2019), systemic over-exploitation of both this species and their preferred prey species may mean that these observations are simply symptomatic of population decline, with abundance of multiple hammerhead species having fallen dramatically in recent decades (Pérez-Jiménez et al, 2014). Further studies will be required to discern which, if any, of these hypotheses are valid.

Negative cephalofoil allometry in S. lewini

Whilst no aspect of head morphology in R. longurio was included in this study, we found strong evidence of negative allometry in the cephalofoil diameter across all classes of S. lewini (Figs 4–6; Tables 4–6), consistent with the results of Sternes & Higham (2022). The cephalofoil has been hypothesized to perform multiple hydrodynamic and sensory functions (Kajiura, 2001; Kajiura et al., 2003, 2005; McComb et al., 2009; Gaylord et al., 2020), but few of these hypotheses have been tested empirically (McComb et al., 2009). Given that young hammerheads of several species are known to target benthic prey (Clarke, 1971; Stevens & Lyle, 1989), the increased lateral search area and manoeuvrability conveyed by a wider cephalofoil (Kajiura, 2001; Kajiura & Holland, 2002) may convey a selective advantage upon neonates and young juveniles foraging on small, agile benthic prey (Sternes & Higham, 2022). Contrastingly, individuals at later ontogenetic stages inhabiting pelagic environments are unlikely to benefit as much from enhanced electroreception, given the ontogenetic trophic niche shifts observed in this species (Estupiñán-Montaño et al., 2021). It should be noted that there is evidence supporting several different hypotheses regarding cephalofoil function (McComb et al., 2009; Kajiura & Holland, 2002; Gaylord et al., 2020) and hence the evolution of cephalofoil morphology may be driven by a complex selective regime, which may not show any obvious correlation with specific ecomorphological functions. Geometric morphometric analysis of cephalofoil form has also provided evidence for the existence of multiple cephalofoil ‘designs’ in Sphyrnidae (Cavalcanti, 2004), and thus it cannot be ruled out that the relative significance of each of these functions may vary between taxa.

Limited evidence of interspecific allometry from a ten-species dataset

Despite clear evidence for ontogenetic allometry in both S. lewini and R. longurio, PGLS analysis has revealed evidence of phylogenetic allometry only in LS (Table 11). Thus, whilst ontogenetic shifts in morphometric trends certainly exist, these results are not inconsistent with the notion of phylogenetic conservatism in shark body form. One potential adaptive explanation for positive allometry in lateral span relates to metabolic demands. Larger-bodied sharks are likely to have both increased oxygen demands (Wong et al., 2021) and reduced surface area to volume ratios relative to smaller sharks (Nakamura et al., 2020), and evidence suggests that gill surface area scales with positive allometry in sharks (Wong et al., 2021) and other fishes (Scheuffelle et al., 2021). If this relationship is valid across shark diversity, then the proportionally larger gill apparatuses of larger-bodied sharks may explain a proportional increase in girth around the gills. The gills are thought to be the primary source of heat loss across the body in most fishes (Stevens, 2011), and thus in taxa without counter-current heat exchange systems (Bernal et al., 2001) that inhabit relatively cool environments, the relationship between gill area and body mass may be more complex. Whilst this positive allometry may hold some biological significance, there are limitations with our PGLS analysis that complicate robust ecomorphological interpretations. Namely, not all ontogenetic stages can be incorporated into a single regression analysis of this nature, and for any given measurement, only ten data points are compared (potentially signifying low statistical power sensuButton et al., 2013). Moreover, species included in this analysis represent only two of the seven extant order-level shark clades; thus, future studies should seek to incorporate taxa representing a greater proportion of total elasmobranch diversity.

The ‘allometric niche shift’ hypothesis

Both ontogenetic allometry (Lingham-Soliar, 2005a, b; Irschick & Hammerschlag, 2015; Fu et al., 2016; Ahnelt et al., 2020; Sternes & Higham, 2022) and isometry (Reiss & Bonnan, 2010) have been reported in sharks, with the existing literature leading some authors to suggest that large-bodied species are likely to grow allometrically, compared to relatively small-bodied taxa that are likely to exhibit isometry (Irschick & Hammerschlag, 2015; Ahnelt et al., 2020). On the basis that R. longurio does not differ significantly in body size from Squalus acanthias (Saunders et al., 1993; Corro-Espinosa et al., 2011), we suggest that this relationship should be revisited as additional studies focusing on small-bodied taxa become available. A number of studies have suggested potential links between allometry in sharks and various aspects of habitat use and ecology (Lingham-Soliar, 2005a; Ahnelt et al., 2020; Sternes & Higham, 2022), yet these concepts are yet to be formalized into specific hypotheses. We propose the ‘allometric niche shift’ hypothesis of ontogenetic allometry in sharks, whereby selective pressures related to long-distance movements (particularly between coastal and pelagic habitats) and ontogenetic trophic niche shifts may represent important drivers of ontogenetic shifts in morphology. Thus, we predict that shark taxa partaking in long-distance movements (or those showing strong trophic differentiation through ontogeny) may exhibit increased allometric trends in the growth of important morphological features, especially caudal measurements, whereas species with relatively limited movements and consistent trophic niches throughout ontogeny may be less likely to exhibit significant allometric growth.

All shark species reported to show ontogenetic morphological shifts exhibit long-distance movements (Lingham-Soliar, 2005a; Hoyos-Padilla et al., 2014; Irschick & Hammerschlag, 2015; Ahnelt et al., 2020; Sternes & Higham, 2022), possibly exposing different ontogenetic stages to different trophic and hydrodynamic environments. Moreover, where studies of the trophic biology of these species have been conducted, evidence of ontogenetic trophic niche shifts has been recovered (Hussey et al., 2012; Dicken et al., 2017). Although Squalus acanthias has been reported to show migratory behaviour in some populations, this has not been investigated outside of the northern Pacific, and in any case this behaviour appears to be less consistent than in taxa that exhibit allometric growth (McFarlane et al., 2003). Ontogenetic trophic niche shifts have also been demonstrated in this species (although again, in a single population; Koen-Alonso et al., 2002), suggesting that trophic niche shifts alone may not provide sufficient selective pressure to drive the evolution of allometric growth. Whilst the relative contribution of migratory tendencies and trophic niche shifts to selection for allometric growth may vary among taxa, we predict that allometry is most likely to occur when both trophic niche and habitat use vary ontogenetically.

Although we only consider extant elasmobranch taxa in this study, ontogenetic niche shifts are well documented in a wide range of taxa (Nakazawa, 2015), and thus there is potential for the allometric niche shift hypothesis to apply outside of sharks. Indeed, evidence for a relationship between allometric growth and ontogenetic niche shifts exists in other, non-elasmobranch taxa. For instance, such a relationship has been recovered between shell morphology and ontogenetic shifts in flow regime in the brachiopod Terebratalia transversa (Sowerby, 1846) (Tomašových et al., 2008). Nonetheless, assessing the generality of these relationships across a wide phylogenetic range of animals remains challenging due to a lack of research effort. Future studies may reveal the extent to which common relationships between ecological niche and morphometric trajectories exist across a diverse range of taxa, by combining morphometric analyses with information on habitat use and trophic ecology.

CONCLUSIONS

In this study we assembled a large, novel dataset for two carcharhiniform species, showing clear ontogenetic shifts in patterns of allometric growth. At least in the case of the hammerhead S. lewini, these shifts are broadly consistent with those found in a previous study (Sternes & Higham, 2022). We speculate on the extent to which ontogenetic shifts in selective pressures may be related to observed growth trajectories in these species. Further taxon-specific studies (including additional morphological measurements covering multiple ontogenetic stages) investigating these putative selective drivers will be required before such speculations can be confidently accepted. After all, allometric growth patterns can also arise as a result of fundamental constraints, such as shared developmental and regulatory pathways between morphological characters (Voje et al., 2014), and hence alternative explanations, which may not be mutually exclusive with ecomorphological hypotheses, should be investigated. Whilst additional ecomorphological and biomechanical studies will be required before we can hope to fully understand the selective drivers underlying evolutionary changes in elasmobranch growth trends, this study provides a valuable complement to existing reports of ontogenetic morphological shifts in sharks, and presents a novel hypothesis for interspecific and ontogenetic differences in scaling trends against which future findings can be tested.

ACKNOWLEDGMENTS

This work was funded by Ocean Blue Tree and Paul Angell Foundation through Pelagios Kakunjá, and by UKRI grant MR/S032177/1 to DJF. For the purpose of open access, the authors have applied a Creative Commons Attribution (CC BY) licence to any Author Accepted Manuscript version arising. Authors would like to thank volunteers and interns at Pelagios Kakunjá A.C. for assisting in fieldwork and data-collection procedures. Additionally, authors thank Arjun D. Tapasvi for valuable discussions regarding statistical aspects of the methodology, and Oliver Demuth for providing the original illustrations used to produce Figure 2. The authors declare no conflicts of interest regarding any aspect of this publication.

DATA AVAILABILITY

The data underlying this article are available in the article and in its online Supporting Information.

SUPPORTING INFORMATION

Additional supporting information may be found in the online version of this article on the publisher's website.

Figure S1. PCA results for the full dataset.

Figure S2. Species-specific PCA results.

Table S1. Full morphological dataset.

Table S2. Full log10-transformed morphological dataset.

Table S3. PGLS morphological dataset, including data extracted from Irschick et al. (2017).

Table S4. Log10-transformed PGLS morphological dataset, including data extracted from Irschick et al. (2017).