Correlation between δ¹³C and δ¹⁵N in flying fish (Exocoetus volitans) muscle and scales from the South China Sea

A total of 82 modern flying fish samples were collected near the Xisha Islands, the South China Sea, including 10 samples from the Xisha Sea around Yongxing Island (17°N, 112°E) and 22 samples from the Southern Hainan Sea (18°N, 110°E, the southern Hainan Island) collected by fishermen in 2014 and 2015, respectively. Additional 50 samples were collected within the Western South China Sea (9° – 19°N, 111° – 115°E), using a small dredger during an open scientific expedition cruise in 2015. These samples represent different sizes of fish, from both nearshore and open water in the South China Sea. All fish samples were weighed and measured for their length (standard length, from the snout to distal caudal vertebrae); their scales and dorsal muscles without skin were then removed with forceps and a knife after defrosting and dried at 60°C.

Muscle samples were pretreated based on the methods of Logan and Lutcavage (2008) and Inamura et al. (2012). Each 0.2 g sample was ground using a mortar and pestle and passed through a 0.90 mm mesh sieve. Samples were then placed twice into a 10 ml chloroform/methanol (1:1, v:v) solution for more than 12 h each time to extract and remove lipids. The remaining samples were then dried at 60°C and stored in tinfoil. Fish scale samples were pretreated based on the method of Estep and Vigg (1985) and Sinnatamby et al. (2007). We used a whole scale to represent the whole fish, and placed it into a clean plastic syringe. Then, 1.2 N HCl was drawn into the syringe to remove carbonates contained in the scale, which was then washed for 2 minutes using deionized water. The scales were air-dried and also stored in tinfoil.

Stable isotopic compositions δ¹³C and δ¹⁵N were determined at the State Key Laboratory of Atmospheric Boundary Layer Physics and Atmospheric Chemistry (LAPC), Institute of Atmospheric Physics, Chinese Academy of Sciences (Beijing, China). Well-treated samples and standards (urea) were fully combusted at 1000°C using a FLASH 2000 HT Elemental Analyzer (Thermo Fisher) and the gases were separated by a “purge and trap” adsorption column and then sent to IRMS (Isotope-Ratio Mass Spectrometry) MAT 253 (Thermo Fisher) for the isotope analysis. Stable isotope ratios were expressed in δ notation as the deviation from standards in parts per thousand (‰): δ13C=[(Rsample/Rstandard)−1]×1000(o/oo), $$\begin{array}{} \rm {\delta}^{13}C=[(R_{sample}/R_{standard})-1]\,\times\,1000\,(^{o}\!/_{\!oo}), \end{array} $$ where R refers to the ratio ¹³C/¹²C, the R_standard value is based on Vienna Pee Dee Belemnite (V-PDB), and δ15C=[(Rsample/Rstandard)−1]×1000(o/oo), $$\begin{array}{} \rm{\delta}^{15}C=[(R_{sample}/R_{standard})-1]\,\times\,1000\,(^{o}\!/_{\!oo}), \end{array}$$ where R represents the ratio ¹⁵N/¹⁴N, the R_standard value is based on atmospheric air nitrogen (N₂-atm). Analytical precision (the standard deviation) for δ¹³C and δ¹⁵N is better than ±0.1‰ and ±0.2‰, respectively.

Consistent with many other fish species (Perga & Gerdeaux 2003; Roussel et al. 2014), stable isotopic characteristics of flying fish scales are strongly correlated with the characteristics of muscle (Fig. 2). Therefore, the fish scale offers a non-lethal method to study the characteristics of stable isotopes of modern flying fish as in other fishes (Jardine et al. 2011). In addition, flying fish scales enable us to infer the stable isotopic ecology of ancient fish when their muscle tissues are not preserved but their scales are.

The trophic level of an organism can be determined by δ¹⁵N due to significant nitrogen isotope fractionation, which can cause an obvious increase in δ¹⁵N in organisms from higher levels of the food chain (Schoeninger & DeNiro 1984; Wada et al. 1991; Hobson 1999; Post 2002). Similarly, δ¹⁵N contained in organisms can increase as they age because they feed on different food sources with evidently different δ¹⁵N values (Minagawa & Wada 1984; Xu et al. 2007). In this study with modern flying fish, δ¹⁵N of both muscle and scales increased with mass (Fig. 3A), with flying fish muscle and scales exhibiting a δ¹⁵N change of ~3‰ and a mass change of 350 g. Thus, we ascribe the variation of δ¹⁵N of both flying fish muscle and scales to the size of fish (and thus age).

However, the relationship between muscle δ¹⁵N and fish mass is evidently different from the relationship between scale δ¹⁵N and fish mass (Fig. 3A). Given that both the slope and intercept of δ¹⁵N_muscle -mass regression are greater than those of δ¹⁵N_scale -mass regression, a larger fractionation of stable nitrogen isotope exists in muscle than in a scale, consistent with the previous studies (Blanco et al. 2009; Vašek et al. 2017). In addition, fish muscle δ¹⁵N provides more information on recent food sources, whereas scale δ¹⁵N records an average dietary composition over the lifetime (Perga & Gerdeaux 2003). Generally, an increase in one trophic level would result in enrichment of 3.4‰ in muscle tissues (DeNiro & Epstein 1981; Minagawa & Wada 1984; Post 2002). Fish scale δ¹⁵N can also be used to calculate the actual trophic level, but only after considering the nitrogen isotopic offset; otherwise, the trophic level may be underestimated by more than half. In this study, the average values of δ¹⁵N_muscle and δ¹⁵N_scale are 8.6 ± 1.0‰ (n = 82) and 6.3 ± 0.8‰ (n = 82), respectively, and their difference (δ¹⁵N_muscle − δ¹⁵N_scale = 2.3 ± 0.6‰) can be applied as the muscle and scale nitrogen isotopic offset for flying fish.

In general, muscle δ¹³C of prey animals (flying fish in our study) varies slightly, less than 1‰ during their lifetime if they do not migrate (DeNiro & Epstein 1978). Consistently, we found that the variation of δ¹³C of fish muscle is less than 1‰. The muscle and scale carbon isotopic offset of flying fish, i.e. the difference between the average values of δ¹³C_muscle and δ¹³C_scale (Δ¹³C) is −2.1 ± 0.5‰. Several factors may cause scale δ¹³C to be higher than that in muscle. For example, glycine is the predominant amino acid of scale collagen and it is enriched by 8‰ in δ¹³C compared to other amino acids (Cano-Rocabayera et al. 2015). Muscle δ¹³C changes linearly with fish mass, while fish scale δ¹³C changes nonlinearly (Fig. 3B). Consequently, the mass or the size is a factor that accounts for the change in scale δ¹³C, but may not be the predominant one. This carbon isotopic offset value, therefore, must be fully considered if δ¹³C is used to infer the food source of fish (Saito et al. 2011).

Although the average offsets of δ¹³C and δ¹⁵N between muscle and scales can be calculated, they still vary with the increase in fish mass (Fig. 4). For δ¹⁵N, there is a weak correlation between the offset and fish mass, and the nitrogen isotopic offset between muscle and scales would increase from 2.17‰ to 2.47‰ during the growth from birth to a weight of 300 g (Fig. 4). Obviously, the variance of about 0.3‰ is small, so the average muscle and scale nitrogen isotopic offset of 2.3 ± 0.6‰ can apply to each flying fish, regardless their weight.

For δ¹³C, there is a strong negative correlation between its offset and fish mass (Fig. 4). For flying fish from birth to a weight of 300 g, the carbon isotopic offset between muscle and scales would change from −1.60‰ to −2.50‰. As flying fish age, the difference between their muscle and scale δ¹³C would slightly increase, which we believe is attributed to fish scales recording the dietary information over their lifetime. Even so, the average carbon isotopic offset between flying muscle −2.1 ± 0.5‰ is still useful for estimating the actual diet in paleoecological studies when only fish scales are preserved in the fossil record.

δ¹⁵N and δ¹³C in fish muscle as well as in scales yielded a weak correlation (Fig. 5), suggesting that the fish size, the primary factor affecting δ¹⁵N, can impact δ¹³C to some extent, but is not the most important factor.

In fact, δ¹³C depends strongly on the geographic location. For example, δ¹³C in organisms may increase with altitude in a mountainous area (Hobson et al. 2003), and dissolved inorganic carbon varies with the seawater depth (Sisma-Ventura et al. 2016). These variations are irrelevant to our study because flying fish live only in the surface layer of warm ocean waters. However, there are a number of geographical factors that may apply to this study: δ¹³C of eupelagic plankton is much lower than that of coastal species (Kaehler et al. 2000; Xu et al. 2014), and δ¹³C of organisms increases from polar areas to equatorial regions as a result of the latitude effect (Takai et al. 2000; Vaughn et al. 2010). Therefore, we attribute the variation of fish δ¹⁵N to the change of fish mass and size, but the variation of fish δ¹³C to the alteration of their geographic location.