Maternal investment and size-specific reproductive output in carcharhinid sharks

Authors

  • Nigel E. Hussey,

    Corresponding author
    1. School of Ocean Sciences, College of Natural Sciences, Bangor University, Menai Bridge, Anglesey LL59 5AB, UK
      Correspondence author. E-mail: nigel.hussey@bangor.ac.uk
    Search for more papers by this author
  • Sabine P. Wintner,

    1. KwaZulu-Natal Sharks Board, Private Bag 2, Umhlanga Rocks 4320, South Africa
    2. Biomedical Resource Unit, University of KwaZulu-Natal, Durban 4056, South Africa
    Search for more papers by this author
  • Sheldon F. J. Dudley,

    1. KwaZulu-Natal Sharks Board, Private Bag 2, Umhlanga Rocks 4320, South Africa
    2. Biomedical Resource Unit, University of KwaZulu-Natal, Durban 4056, South Africa
    Search for more papers by this author
  • Geremy Cliff,

    1. KwaZulu-Natal Sharks Board, Private Bag 2, Umhlanga Rocks 4320, South Africa
    2. Biomedical Resource Unit, University of KwaZulu-Natal, Durban 4056, South Africa
    Search for more papers by this author
  • David T. Cocks,

    1. School of Ocean Sciences, College of Natural Sciences, Bangor University, Menai Bridge, Anglesey LL59 5AB, UK
    Search for more papers by this author
  • M. Aaron MacNeil

    1. Australian Institute of Marine Science, PMB 3 Townsville MC, Townsville, Qld 4810, Australia
    Search for more papers by this author

Correspondence author. E-mail: nigel.hussey@bangor.ac.uk

Summary

1. Life-history theory predicts that organisms will provide an optimal level of parental investment for offspring survival balanced against the effects on their own survival and future reproductive potential.

2. Optimal resource allocation models also predict an increase in reproductive output with age as expected future reproductive effort decreases. To date, maternal investment in sharks has received limited attention.

3. We found that neonatal dusky sharks (Carcharhinus obscurus) are not independent from maternal resource allocation at the point of parturition but instead are provisioned with energy reserves in the form of an enlarged liver that constitutes approximately 20% of total body mass.

4. Analysis of long-term archived data sets showed that a large proportion of this enlarged liver is utilized during the first weeks or months of life suggesting that the reported weight loss of newborn sharks signifies a natural orientation process and is not necessarily related to prey abundance and/or indicative of high mortality rates.

5. Interrogation of near-term pup mass in two carcharhinids, the dusky and spinner shark (Carcharhinus brevipinna), further revealed an increase in reproductive output with maternal size, with evidence for a moderate decline in the largest mothers.

6. For the dusky shark, there was a trade-off between increasing litter size and near-term pup mass in support of optimal offspring size theory.

7. For both the dusky and spinner shark, there was a linear increase in near-term pup mass with month, which may indicate variable parturition strategies and/or that carcharhinids are able to adjust the length of the gestation period.

8. The identification of optimal size-specific reproductive output has direct implications for improving the reproductive potential of exploited shark populations and for structuring future management strategies.

Introduction

Life-history theory predicts that organisms should provide an optimal level of parental investment or energy expenditure to increase both offspring survival and future reproductive success, balanced against components of their own maintenance, survival and future reproductive potential (Williams 1966; Clutton-Brock 1991; Roff 1992; Stearns 1992). For many organisms, including sharks, males contribute no parental care; consequently the number, size and growth of young are determined by the energy allocated by the mother, i.e. the level of maternal investment (Evans 1990).

In terms of reproductive mode, carcharhinid sharks are analogous to marine mammals in that they bear live young (viviparity). A principal difference between the reproductive strategies of these two groups, however, is the allocation of postnatal maternal resources. For most mammals, postnatal costs are incurred through the direct supply of food to the young from the mother’s reserves (Bowen et al. 2001) and through maternal–young behavioural interactions (Mann & Smuts 1999; Szabo & Duffus 2008).

In the case of sharks, evidence suggests that at the point of parturition, offspring independence is distinct and terminal, similar to most spawning fish (Wourms 1977; Jönsson, Tuomi & Järemo 1998) and marine reptiles (Broderick et al. 2003). For these later taxa, however, the independent young and emerging hatchlings are provided with a component of prenatal reproductive resources, or a maternal head-start, through the provisioning of yolk reserves (Congdon 1989; Berkeley, Chapman & Sogard 2004; Donelson, McCormick & Munday 2008). The young ultimately do not become independent from maternal resource allocation until they have exhausted this endogenous energy store. Importantly, strong positive relationships between the quantity (size) and quality of these reserves and factors such as time to exogenous feeding, growth rate and early survival have been widely reported (Wallace & Aasjord 1984; Springate & Bromage 1985; Hare & Cowen 1997; Marteinsdottir & Steinarsson 1998; Nagle, Burke & Congdon 1998; Berkeley et al. 2004; Gagliano & McCormick 2007; Donelson et al. 2008). Prenatal maternal investment in most oviparous ectothermic vertebrates, which exhibit no direct maternal care or postnatal costs, can therefore be considered to include the allocation of reproductive resources for postnatal use or the provisioning of a maternal head-start (Congdon 1989; Berkeley et al. 2004; Wallace et al. 2007).

In contrast to teleost fish, newborn carcharhinid sharks are equipped with fully functional jaws and teeth (Wourms 1977; Carrier, Pratt & Castro 2004) and have therefore been considered independent of maternal resource allocation at the point of birth. Gilmore, Dodrill & Linley (1983), Gilmore (1983) and Francis & Stevens (2000), however, noted that near-term pups of sand tiger (Carcharias taurus), longfin mako (Isurus paucus), and porbeagle sharks (Lamna nasus) possessed a larger liver than both small prenatal and neonatal sharks. For large adult sharks, it is has long been recognized that the liver acts as a storage site for lipid reserves which are used in reproduction and migration and during periods of prey limitation (Springer 1967; Hoffmayer, Parsons & Horton 2006; Hussey et al. 2009). Gilmore et al. (1983a), Gilmore (1983b) and Francis & Stevens (2000) proposed that the enlarged liver of neonatal sharks may act as an energy reserve to be utilized after birth. Consequently, if neonatal sharks are provisioned with an enlarged liver they may be reliant on a maternal head-start to enable the transition to independent foraging and to increase their likelihood of survival.

Comparative analysis of life-history traits predicts that a mother’s phenotype will impact the phenotype of her young, independent of genotype (Bernardo 1996; Mousseau & Fox 1998). Maternal phenotypic effects are typically mediated by the age (size) and experience of the female, commonly resulting in a positive correlation between maternal size and offspring size (Chambers & Leggett 1996; Coté & Festa-Bianchet 2001; Green & McCormick 2005; Skibiel, Dobson & Murie 2009) and/or the number of offspring produced (Wootton 1990; Morris 1996; Sogard, Berkeley & Fisher 2008), i.e. an increase in overall reproductive output. These effects, which are common across taxa, likely reflect available space for developing offspring and the acquired experience of larger individuals to obtain/store food for offspring development (Mattingly & McCure 1982). Optimality models further predict that reproductive effort will increase with age, as expected future reproductive success decreases (Pianka & Parker 1975; Charlesworth & Leon 1976). At the species level, heterogeneity in near-term young mass/length and litter size vs. mother mass/length relationship is common due to multiple competing functions exerting influence. Aside from maternal size, environment-induced phenotypic effects include the mother’s ability to secure high-quality habitats (prior to breeding and throughout gestation; Mappes, Koskela & Ylonen 1995; Räsänen et al. 2008) and temporal variability, over scales of months and years, in environmental conditions and resource availability within selected habitats (Sikes & Ylonen 1998; Castro et al. 2009; Wilson et al. 2009). Flexibility in a species’ reproductive strategy, such as open breeding seasons (Hutchings & Myers 1993) and flexible gestation periods (Asher 2007), will also impact on the quality and survival of young. For many organisms, physiological changes in maternal condition cued by a general decrease in body condition with age may also occur (Roff 1992). As a result of these trade-offs, a specific age or reproductive size may be the most reproductively efficient strategy in terms of breeding resource allocation and production of healthy offspring (Williams 1966; Boltnev & York 2001; Carr & Kaufman 2009).

Considering the estimated decline in many shark populations (e.g. Baum et al. 2003) and the tendency of fisheries to target larger individuals within the population (Myers & Worm 2005), understanding the allocation of maternal investment by mothers of different sizes has important implications for potential conservation and management strategies.

The aim of this study was to determine if the reported enlarged liver of near-term shark pups represents a form of maternal investment for neonatal, free swimming sharks. A carcharhinid, the dusky shark [Carcharhinus obscurus (LeSueur, 1818)], was selected as a test species. With data for two species, the dusky and the spinner shark [Carcharhinus brevipinna (Müller & Henle, 1839)], an additional aim was to examine the variability in maternal investment in near-term carcharhinid sharks to determine (i) what maternal and environmental effects might be driving this variability and (ii) if there is evidence of optimal size-specific reproductive output/investment.

Materials and methods

Study location and sampling

Data for all sharks included in this study were recorded from animals incidentally caught in beach protection nets in KwaZulu-Natal (KZN), South Africa, between 1978 and 2008. Nets were set approximately 300–500 m from beaches, parallel to the shoreline in 10–14 m of water (Dudley et al. 2005). All dead sharks retrieved from the nets in good condition were transported to the KwaZulu-Natal Sharks Board (KZNSB) main laboratory for dissection. For specific details regarding changes to net installations over time and net servicing operations refer to Cliff, Dudley & Davis (1988) and Dudley et al. (2005). On arrival at the laboratory, data on species, sex, capture date and location, as well as basic morphometric measurements including precaudal length [PCL (cm)], and total body mass (kg) were recorded. PCL was measured as the straight line distance from the tip of the snout to the precaudal notch as defined by Dudley et al. (2005). Sharks were then stored frozen (−20 °C) prior to dissection. During the main dissection, the complete stomach was removed and total stomach mass and total liver mass of combined lobes of each shark were recorded to the nearest gramme. For net-caught neonatal/juvenile free swimming sharks, the stage of umbilical scar healing or umbilical scar stage (USS) was also recorded. A qualitative 5-point USS scale was devised, where (i) the wound was completely open, (ii) wound open but muscle tissue was closing, (iii) skin was closing, (iv) a faint scar remained and (v) no scar was present. The estimated time between USS1 and USS3 is 2–4 weeks (Bass, D’Aubrey & Kistnasamy 1973; Duncan & Holland 2006). For pregnant sharks, all pups were extracted from both uteri and the number (litter size), sex, PCL (cm) and total body mass (kg) were recorded. Accepting that net-caught pregnant sharks may abort pups, the total litter size of each individual female was confirmed by counting the total number of placental scars.

Maternal investment in free swimming neonatal/juvenile sharks

To examine if an enlarged liver represents a maternal investment in free swimming neonatal dusky sharks, we isolated data on total body mass, total liver mass, PCL and stomach mass for all small sharks with reported USS data. We then extracted total body mass and PCL data for all near-term and mid-term dusky pups. Near-term pregnant dusky sharks were defined as having pups with a PCL of ≥60 cm (Dudley et al. 2005). Liver mass data for near-term pups were recorded for eight individuals from three separate mothers between 2007 and 2008. The two most commonly applied condition indices; a somatic measure, hepatosomatic index (HSI) and a morphometric measure, condition factor (CF), were then calculated for mid-term/near-term pups and neonatal/juvenile sharks with USS data using the following equations:

image(1)

and

image(2)

where MTL equals the total mass of both liver lobes combined per individual shark and MTB is the total body mass of each individual shark. Linear models were applied to the data to examine the trends in total liver mass, total body mass and condition (HSI and CF) of sharks by the defined life stages. Because the sample sizes were unbalanced, significance of pair-wise comparisons was tested using adjusted Bonferroni tests. To visually interpret changes in total stomach mass for each USS, i.e. an indication of the level of exogenous feeding, the data were plotted in boxplot format.

Maternal investment in near-term pups

To examine what maternal and environmental factors may influence investment in near-term shark pups and if there is evidence for optimal size-specific reproductive output, we accessed all mother–pup data for near-term pregnant dusky sharks, and a second carcharhinid, the spinner shark. Near-term spinner sharks were defined as having pups with a PCL of ≥50 cm (Allen & Cliff 2000). Both species utilize the same reproductive strategy, viviparity, with developing embryos receiving nourishment through a pseudo-placental sac. The population status of both species is considered to be stable in the study region (Dudley & Simpfendorfer 2006), and therefore we could rule out any influence of complicating factors, for example, compensatory density-dependent mechanisms related to overfishing (Rose et al. 2001). We examined pup mass as a response variable representative of pup fitness.

The structure of the data (multiple pups from the same litter) necessitated the use of hierarchical linear models to properly structure the statistical dependence among pups from the same mother and the reduced number of observations for mothers, relative to pups. The pup-level model was of the form:

image(3)

with pup-level variables, x, and a mother-dependent intercept, β0k and standard deviation σ2k. The mother-level model for average pup mass within mother was:

image(4)

with mother-level variables, z. We adopted an a priori model-building strategy and developed nine candidate models of increasing complexity, factoring in the above defined mother–pup variance structure and fixed effects for mother–pup and seasonal relationships (i.e. maternal and environmental phenotypic controls) that were reasonably suspected of operating during the sampling period (Table 1). This included a pup-level sex covariate, maternal covariates of mother PCL and litter size, environmental covariates of month and year, and associated interactions. For the maternal covariates of mother PCL, we formally tested both linear and quadratic functions, the latter via pair-wise interactions, as we may expect an initial increase in maternal reproductive performance with size followed by a senescent decline in later life. Of the suspected model interactions, we included litter size–mother size as previously a weak relationship between these effects has been reported for both the spinner and dusky shark (Allen & Cliff 2000; Dudley et al. 2005). The candidate model set for both dusky and spinner sharks were implemented using the nonlinear mixed effects package (nlme; Pinheiro & Bates 2000) in the statistical program R (R Development Core Team, 2009). Given our large sample sizes, Akaike’s information criterion were used to assign relative strengths of evidence to each candidate model in the set and parameter estimates were solved for using the likelihood-ratio tests provided in r.

Table 1.   Candidate model set for quantifying near-term pup mass in dusky (Carcharhinus obscurus) and spinner (Carcharhinus brevipinna) sharks
ModelEquation
  1. Variance structure provided in eqns (2 and 3) in the Methods section of the text.

  2. pMASS, mass (g) for pup i; pSEX, pup sex (male = 1/female = 0); mPCL, group-centered precaudal length (cm) for mother k; LS, litter size; MONTH, month of capture (Feb = 2, ..., Dec = 12); YEAR, median-year centred year of capture.

M0pMASSik = γ0 + β0k
M1pMASSik = γ0 + β0k + β1pSEXi
M2pMASSik = γ0 + γ1mPCLk + β0k + β1pSEXi
M3pMASSik = γ0 + γ1mPCLk + γ2LSk + β0k + β1pSEXi
M4inline image
M5inline image
M6inline image
M7inline image
M8inline image
M9inline image

Results

Maternal investment in free swimming neonatal/juvenile sharks

A marked reduction in total liver mass was found between near-term pups and USS5 dusky sharks with mean liver mass values ranging from 1072·8 ± 78·4 g (mean ± SE) to 517·3 ± 47·4 g (Fig. 1a). When considering total body mass, a declining trend was observed between near-term pups and USS2 sharks, followed by an increase between USS2 and USS5 sharks (Fig. 1a). In agreement with total liver mass, HSI decreased from 19·5 ± 0·96 (mean ± SE) to 6·4 ± 0·50, between NTP and USS5 sharks (Fig. 1b). Data for HSI/liver mass of mid-term pups were unavailable. For CF, there was a large increase in condition from mid-term pups to near-term pups, but then a noticeable reduction in condition with increasing USS. CF values ranged from 1·55 ± 0·001 for near-term pups to 1·33 ± 0·02 for USS5 sharks, an identical value to mid-term pups (1·33 ± 0·01) (Fig. 1c). Linear model results were significant for liver mass, total body mass, HSI and CF (F5, 134 = 14·16, P < 0·0001; F5, 134 = 17·93, P < 0·0001; F5, 134 = 31·24, P < 0·0001, F6, 2226 = 84·15, P < 0·0001, respectively). For total liver mass and HSI, pair-wise comparisons were highly significant indicating a consistent graduated decrease with USS (Fig. 1a, b). For CF, pair-wise comparisons found that NTP and USS1 sharks were in significantly better condition than all other stages. No difference in CF was detected between mid-term pups and USS2-5 sharks (Fig. 1c). Total stomach mass of USS1-2 sharks was negligible while an increase in total stomach mass was observed for USS3-5 sharks (Fig. 1d).

Figure 1.

 Transition in (a) actual liver mass and total shark mass, (b) hepatosomatic index (HSI), (c) condition factor (CF) and (d) total stomach mass for mid-term dusky shark (Carcharhinus obscurus) pups to free swimming sharks with a healed umbilical scar (USS5). Data are mean ± SE. For total stomach mass, boxes illustrate the interquartile range, whiskers identify the largest non-outlier observations, solid circles are outliers and open circles are the mean values. Values displayed above a given life stage indicate the life stage(s) for which pair-wise tests revealed significant differences with the given life stage. Numbers in each plot represent the sample size of sharks per prenatal and umbilical scar stage class. For total stomach mass, samples sizes are the same as for condition factor.

Maternal investment in near-term pups

A total of 228 and 136 near-term pregnant dusky and spinner sharks and 2230 and 1277 near-term dusky and spinner pups, respectively, were included in the analyses. Model fits were adequate for all candidate models except M1 (Table 2) and quantile–quantile plots showed the data to be well-described by normally distributed errors for both species. Candidate model results for the relationship between pup mass and mother-environment phenotypic effects favoured models M7 and M6 for the dusky and spinner shark respectively (Table 2). For both species there was a clear relationship between pup mass and both linear and quadratic functions of mother PCL (Table 3; Fig. 2a, b). Mean pup mass increased with mother size to an asymptote, with evidence for a decline at the largest mother lengths of both species (Fig. 2a, b). For the dusky shark, there was an effect of litter size, with mean pup mass decreasing with increasing litter size (Table 3; Fig. 2c). A model-fit regression found an overall decrease in dusky shark pup mass of 0·79 kg between litters composed of 3 and 16 pups. A similar trend was observed for spinner sharks (Fig. 2d), but the negative effect of litter size was of a much smaller magnitude within the best-fit model (Table 3). For both species, pup mass showed a strong linear increase with month (Fig. 2e, f), with model-fit regression values ranging from 4·14 kg (February) to 5·83 kg (December) and 2·01 kg (February) to 2·27 kg (August) for the dusky and spinner shark respectively (Table 3; Fig. 2e, f). The peak catch period of postpartum dusky and spinner sharks in beach protection nets was July and June respectively (Fig. 2e, f). For spinner sharks, there was evidence for heavier female pups and a trend of increasing pup mass since the late 1970s (Table 3).

Table 2.   Model selection results for top-ranked models of dusky (Carcharhinus obscurus) and spinner (Carcharhinus brevipinna) shark near-term pup mass
ModelGOFKlogLikAICΔAICwi
  1. GOF, likelihood-ratio test goodness of fit; K, number of model parameters; logLik, model log-likelihood; ΔAIC, relative AIC differences; wi, Akaike’s weights.

Dusky
 M7<0·0017−16 548·3433 114·680·000·35
 M5<0·0016−16 549·8633 113·710·170·33
 M6<0·0017−16 549·5833 115·171·620·16
 M8<0·0017−16 548·8633 117·733·040·07
 M9<0·0017−16 549·0433 118·083·400·06
 M4<0·0015−16 553·6033 119·195·700·03
 M3<0·0014−16 573·2333 156·4743·00·00
 M20·0673−16 578·1833 194·9251·70·00
 M11·002−16 580·0033 169·7655·10·00
 M01·001−16 581·2733 168·553·90·00
Spinner
 M6<0·0017−8246·3816 510·80·000·52
 M8<0·0017−8246·2916 512·61·830·21
 M9<0·0017−8246·3716 512·71·980·19
 M5<0·0016−8250·1316 516·35·520·03
 M7<0·0017−8249·4016 516·96·100·03
 M4<0·0015−8252·1516 518·37·540·02
 M3<0·0014−8255·4516 520·910·150·00
 M2<0·0013−8254·4716 520·910·190·00
 M11·002−8265·3216 538·727·890·00
 M01·001−8267·5716 541·130·390·00
Table 3.   Parameter estimate results for best-fit models of dusky (Carcharhinus obscurus) and spinner (Carcharhinus brevipinna) shark near-term pup mass
ParameterEstimateSE95% CI
  1. Parameters not overlapping zero are given in bold; abbreviations provided in Table 1.

Dusky (M7)
 Intercept4061·2131·43804, 4319
 pSEX−14·215·1−43·9, 15·4
 LS−48·118·3−84·2, −12·0
 mPCL9·34·50·59, 18·1
 mPCL2−0·970·32−1·60, −0·35
 LS × mPCL2·351·60−0·79, 5·49
 MONTH149·622·5105·2, 193·9
Spinner (M6)
 Intercept2032·7119·11800, 2266
 pSEX−16·77·8−31·9, −1·4
 LS−7·58·0−23·2, 8·3
 mPCL7·22·22·8, 11·6
 mPCL2−0·310·15−0·60, −0·02
 MONTH52·718·117·1, 88·4
 YEAR7·52·72·1, 12·9
Figure 2.

 The relationship between near-term pup mass and maternal precaudal length (a, b), litter size (c, d) and the seasonal effect of month (e, f) for dusky (Carcharhinus obscurus; model M7) and spinner sharks (Carcharhinus brevipinna; model M6) respectively. Small grey dots are raw data (a–f), large grey dots with black outline (a, b) and large black dots (c–f) are mean data (±SD) with best-fit model regressions where significant. Grey dotted lines indicate the minimum and maximum values of the best-fit regression lines. For (e, f), the grey bar indicates the peak catch period of postpartum females for both species in beach protection nets.

Discussion

To the best of our knowledge these data provide the first characterization of a maternal head-start for newborn sharks, in the form of an enlarged liver and provide evidence for both size-specific reproductive output (or investment) and environmental phenotypic control on pup fitness for two carcharhinid sharks.

The clear decline in total liver mass, HSI and CF of neonatal dusky sharks with USS provide empirical evidence that liver reserves, provisioned by the mother, are used in the first few weeks (or even months) of life. In agreement with our observed decline in CF and neonatal dusky shark mass, Duncan & Holland (2006) reported that captive neonatal scalloped hammerhead sharks (Sphyrna lewini) lost weight with umbilical scar healing stage and that the CF of free ranging animals in Kāne’ohe Bay, Hawai’i, was lower in the months following the birthing peak. Similar findings were reported in previous studies by Lowe (2002) and Bush & Holland (2002). Lowe (2002) and Duncan & Holland (2006) concluded that newborn scalloped hammerhead sharks were in a poor nutritional state and attributed this to population size and poorly developed foraging skills and/or reduced prey availability, suggesting this was indicative of high mortality rates within the nursery habitat. It is likely that the weight loss of newborn sharks is a combined result of the utilization of the provisioned maternal head-start and both density dependent mechanisms and possible reduced prey availability within the nursery [as reported for scalloped hammerhead sharks in Hawai’i; Duncan & Holland (2006)]. Certainly the rate of total body mass loss of newborn sharks will be dependent on individual foraging development and success. Possible variation in the allocation of maternal resources may also be a contributing factor to pup survival in high density nursery regions.

The limited stomach contents of early USS dusky sharks support the role of the liver in maternal investment in agreement with observations on neonatal scalloped hammerhead sharks (Bush 2003). The enlarged liver acts as a food reserve to maintain the young sharks while they orientate themselves in their environment and develop their foraging skills. This point is further supported by observations of a newborn sand tiger shark (Carcharias taurus) born in captivity that did not feed for the first 25 days (Gilmore et al. 1983). Furthermore, Hussey et al. (2009) reported a clear difference in HSI values of suspected neonatal and juvenile dusky sharks over a seasonal cycle. Current concerns over the weight loss of newborn sharks in nursery areas and associated high mortality rates (Heupel, Carlson & Simpfendorfer 2007 and references therein) require further detailed investigation. The existence of an enlarged liver in neonatal sharks and associated natural weight loss may also be a complicating factor when studying selection on life-history traits of newborn sharks (DiBattista et al. 2007).

Although the provisioning of a maternal head-start in sharks is not surprising considering documented maternal provisioning of teleost fish and marine reptiles (Berkeley et al. 2004; Gagliano & McCormick 2007; Donelson et al. 2008), the allocation of excess liver reserves appears to be a novel maternal investment strategy among viviparous marine vertebrates and may in part explain the evolutionary success of sharks. For viviparous sharks, the decoupling of postnatal care may result in considerably higher reproductive output per individual female’s lifespan, while the birthing of multiple, large, well-provisioned live young increases the likelihood of pup survival, in the absence of maternal care.

In the case of many oviparous fishes, predation and starvation during larval development frequently govern survival and subsequent recruitment (Houde 1987). Once endogenous yolk reserves have been absorbed, larvae must begin to feed themselves immediately, without time for environment acclimatization coupled with limited ability to disperse. This may lead to high mortality rates when larvae find themselves in food-deprived environments [sensu Hjort’s critical period hypothesis; Hjort (1914)]. For sharks, the mother is able to provision her young with a level of endogenous prenatal resource allocation irrespective of exogenous resource availability at parturition and these reserves can be directly utilized by the pups in conjunction with independent foraging. Neonatal sharks can therefore maximize their survival potential and mitigate the effects of the critical period hypothesis.

How the level of maternal investment in near-term sharks and rays varies between species or families adopting different reproductive strategies (i.e. viviparous, ovoviviparous) is unknown. Reported data for the oophagous white shark (Carcharodon carcharias) suggests lower levels of provisioning than observed in this carcharhinid study [NTP/newborn HSI = 16·5 (n = 3), free swimming juvenile HSI = 12·5 (n = 6); Cliff, Dudley & Jury 1996; Francis 1996]. For sandtiger sharks, which are also oophagous but exhibit uterine cannibalism, the reported enlarged liver of a newborn animal had a HSI value of 9·9% (Gilmore et al. 1983a), suggesting placental species demand a larger liver reserve than oophagous species prior to parturition. Indeed, direct maternal–young nutrient placental transfer may enhance the fitness of viviparous newborn sharks, but may also increase the potential for parent–offspring conflicts (Crespi & Semeniuk 2004), a situation further exacerbated in populations with polyandrous mating systems (Zeh & Zeh 2000). It is likely that inter-specific variation in the level of maternal investment is related to not only reproductive mode but also the life-history strategies (region of parturition, size at birth, growth rate etc.) of the species in question.

The occurrence of heavier pups in the mid-size class of mature carcharhinid sharks provides evidence for optimal size-specific reproductive output. It is expected that mothers possess phenotypic plasticity and can adjust offspring phenotype in a way that enhances offspring fitness (Mousseau & Dingle 1991) and therefore variation in pup mass relative to maternal size may be influenced by several alternative factors, such as genotypic variation in reproductive traits, litter size, paternal genetic input, population demography, timing of parturition and environmental-resource heterogeneity (Clutton-Brock 1991; Roff 1992, 2002; Jordan & Snell 2002). For the dusky shark, we found evidence for a trade-off between pup mass and increasing litter size in support of optimal offspring size theory (Smith & Fretwell 1974; Stearns 1992). Allen & Cliff (2000) and Dudley et al. (2005) previously reported a significant relationship between litter size and maternal size for both study species, with litter size increasing with maternal PCL. The inclusion of the mother size–litter size interaction within our candidate models, however, had limited effect on best-fit model selection, indicating that the observed pup mass–litter size relationship is unlikely to be regulated by maternal size. For leatherback turtles, Wallace et al. (2007) found that females maximize the number, but not necessarily the size of the young per breeding event and that mother size exerted minimal influence on this relationship. Similar to the suggestion of Wallace et al. (2007), it seems likely that physical constraint of body size in carcharhinid sharks may have limited impact on observed litter size patterns. Therefore, differential resource availability may better explain observed variation in litter size (Reznick & Yang 1993; Reznick, Callahan & Llauredo 1996) although we cannot rule out a combination of genotype–environmental interactions driving the observed pattern (Reznick, Nunney & Tessier 2000; Roff 2002). If we consider that litter size will approach an optimal value through evolutionary time (Lack 1947), it was apparent that dusky and spinner sharks preferentially selected for litter sizes of 8–12 and 6–12 pups respectively (Fig. 3).

Figure 3.

 Percentage frequency of litter sizes for the dusky inline image (Carcharhinus obscurus) and spinner shark bsl00001 (Carcharhinus brevipinna).

The consistent linear increase in pup mass with month for both species may also be a result of phenotypic plasticity. In the case of the dusky shark, near-term pregnant and postpartum females and newborn sharks with early-stage umbilical scars are caught in beach protection nets throughout the year (Dudley et al. 2005; KZNSB, unpublished data). This suggests that either mating occurs year round and/or reported sperm storage (Pratt 1993) allows fertilization plasticity enabling females to opt for variable parturition strategies to optimize offspring survival. This would entail pupping smaller offspring earlier in the year when environmental conditions are optimal and during a period of abundant prey availability in the nursery habitat (i.e. the annual sardine run, Armstrong et al. (1991)), and pupping larger animals in the warmer, summer months when the density of large predators in coastal waters is reduced and pups are conferred the advantage of the larger size at birth. Birthing later in the year may also enable mothers to increase the size of their pups by exploiting this seasonal prey base, although there has been limited evidence to date of sardines, Sardinops sagax, in the diet of juvenile or pregnant female sharks (van der Elst 1979; Dudley et al. 2005; Hussey et al. 2009). Additionally, the effect of month and the linear increase in near-term pup mass may suggest that carcharhinid sharks are able to regulate the length of the gestation period dependent on both maternal condition and resource availability. Similar observations have been reported for large terrestrial mammals (Kiltie 1982; Asher 2007; Mysterud et al. 2009).

For the spinner shark, our best-fit model result found that female near-term pups were slightly heavier than males. Sex-biased pre- and postnatal maternal investment in young is widely recognized, and is thought to be adaptive dependent on species-specific life-history strategies (Charnov 1982; Cockburn, Legge & Double 2002). For some carcharhinid sharks, including the spinner shark, sexual-size dimorphism occurs with females attaining a larger overall size than males (Cliff et al. 1988; Allen & Cliff 2000). This is in contrast to most polygynous mammals, where sexual-size dimorphism is male biased (Weckerly 1998), and selection for larger size is thought to be related to male–male competition for females (Clutton-Brock 1989), and higher variation in male reproductive success (Cockburn et al. 2002). Our result, in conjunction with the observed increase in reproductive output with size of mature female, indicate that maternal investment is more important for the reproductive value (body size) and success (survival) of female offspring (Leimar 1996; Schulte-Hostedde, Miller & Gibbs 2002; Koskela et al. 2009). However, when we consider the difference in mean body mass between sexes is minimal (35 g; difference of 1·6% of female body weight) and that individual litter ratios can be sex biased (NSB, unpublished data), it is equally likely that this is a result of a sex ratio-pup mass-mother size relationship and future models examining pup mass as an indicator of fitness, should incorporate an associated interaction. Similarly the observed increase in spinner shark pup mass since the late 1970s, which was not seen for the dusky shark, was influenced by an increase in the length (PCL) of net-caught pregnant spinner sharks over time (Fig. 4). This increase in mean pup mass therefore reflects the higher reproductive output of larger females rather than possible density-dependent compensatory mechanisms reported for exploited fish species (Rose et al. 2001).

Figure 4.

 The relationship between pregnant female size (precaudal length) and year (1978–2006) for the dusky (Carcharhinus obscurus) and spinner (Carcharhinus brevipinna) shark. Black dots are raw data with fitted linear regression lines (for the dusky: y = 0·04 Year + 183; F1,2228 = 1·44, P = 0·23; and for the spinner: y = 0·35 Year−495; F1,1274 = 79·23, P < 0·0001).

Aside from environmental phenotypic and predicted genotypic effects, this study affirms the prediction that reproductive effort will increase with age (Williams 1966), but with some evidence of a cost to the condition of individual newborn animals at larger maternal sizes due to an eventual decline in reproductive performance (Reznick et al. 2004; Tedesco, Benito & Garcia-Berthou 2008). It is well-established that age-dependent patterns of reproductive performance can influence population dynamics of teleosts (Carr & Kaufman 2009) and consequently, neglecting the age structure of a population, may overestimate viable larvae production (Trippel, Kjesbu & Solemdal 1997; Scott, Marteinsdottir & Wright 1999); a similar case may exist for viviparous carcharhinids and the survival of their pups. Considering the large knowledge gap in our understanding of the reproductive success of female sharks and the associated fitness of their near-term pups in conjunction with the limited curvature of shark stock-recruitment curves (Kinney & Simpfendorfer 2009), the implications of a peak reproductive size in commercially harvested species are substantial. Small increases in fishing mortality can have a disproportionally large effect on population viability (Walters & Martell 2004), especially at key life-history stages. If depleted shark populations are to be efficiently restored there is considerable work to do in quantifying their life histories and setting appropriate management conditions.

Acknowledgements

We express our gratitude to the KwaZulu-Natal Sharks Board laboratory staff for their vigilant dissection work and recording of data which enabled this study to be undertaken. We thank our colleagues, John R. Turner, Lewis Le Vay, I.D. McCarthy and Aaron T. Fisk and two anonymous referees for their valuable and constructive comments on earlier drafts. NEH would like to express thanks to Anna J. and Alina J. Hussey for their continued support. NEH was funded by a NERC PhD studentship (NER/S/A/2005/13426).

Ancillary

Advertisement