%20-%20https://sketch.com%20--%3e%3ctitle%3e3ECBBF50-FD08-412A-943D-2CFC5E025A16%3c/title%3e%3cdesc%3eCreated%20with%20sketchtool.%3c/desc%3e%3cg%20id='Stylesheet'%20stroke='none'%20stroke-width='1'%20fill='none'%20fill-rule='evenodd'%3e%3cg%20transform='translate(-926.000000,%20-668.000000)'%20id='BUTTONS'%3e%3cg%20transform='translate(100.000000,%20597.000000)'%3e%3cg%20id='BTN-/-RIGHTS'%20transform='translate(816.000000,%2065.000000)'%3e%3cg%20id='ICON-/-Rights'%20transform='translate(10.000000,%206.000000)'%3e%3cpath%20d='M13.5752841,0%20L1.34943182,0%20C0.59375,0%200,0.586666667%200,1.33333333%20L0,22.6666667%20C0,23.4133333%200.620738636,24%201.34943182,24%20L17.6505682,24%20C18.40625,24%2019,23.3866667%2019,22.6666667%20L19,5.36%20L13.5752841,0%20Z%20M13.5752841,1.89333333%20L17.0838068,5.36%20L13.5752841,5.36%20L13.5752841,1.89333333%20Z%20M1.34943182,22.6666667%20L1.34943182,1.33333333%20L12.1988636,1.33333333%20L12.1988636,5.36%20C12.1988636,6.10666667%2012.7926136,6.69333333%2013.5482955,6.69333333%20L17.6235795,6.69333333%20L17.6505682,22.6666667%20L1.34943182,22.6666667%20Z'%20id='Shape'%20fill='%23595959'%20fill-rule='nonzero'%3e%3c/path%3e%3ccircle%20id='Oval'%20stroke='%23595959'%20stroke-width='1.5'%20fill='%23595959'%20cx='9.5'%20cy='13.5355339'%20r='5.5'%3e%3c/circle%3e%3cpolyline%20id='Path-2'%20stroke='%23FFFFFF'%20stroke-width='1.5'%20transform='translate(9.500000,%2013.035534)%20rotate(45.000000)%20translate(-9.500000,%20-13.035534)%20'%20points='8%2016.0355339%2011%2016.0355339%2011%2010.0355339'%3e%3c/polyline%3e%3c/g%3e%3c/g%3e%3c/g%3e%3c/g%3e%3c/g%3e%3c/svg%3e)
Introduction
Fisheries remain important sources of food for human populations, with a total of 81.5 million tons of marine fish captured in 2014 (FAO, 2016). Based on the FAO's analysis of assessed commercial fish stocks, the proportion of fish stocks within biologically sustainable levels decreased from 90% in 1974 to 68.6% in 2013 (FAO, 2016). The accurate identification of stocks has significant implications for fisheries management and conservation biology (Kritzer & Liu, Reference Kritzer, Liu, Cadrin, Kerr and Mariani2014), and non-consideration of stock structure may lead to decreases in stock abundances and genetic diversity, as well as changes in biological attributes (Begg et al., Reference Begg, Friedland and Pearce1999). However, describing a stock is complex and numerous studies have reached the conclusion that stock units may be identified with a high degree of confidence by integrating information from several methods and combining different spatial and temporal scales, thus making it possible to overcome many of the drawbacks of single-technique approaches (Cadrin et al., Reference Cadrin, Kerr, Mariani, Cadrin, Kerr and Mariani2014; Izzo et al., Reference Izzo, Ward, Ivey, Suthers and Stewart2017).
Mediterranean fisheries are multispecies and generally lack large monospecific stocks compared to the open ocean (Lleonart & Maynou, Reference Lleonart and Maynou2003). Indeed, more than 90% of the fish stock has never been properly assessed (Pauly et al., Reference Pauly, Hilborn and Branch2013). The scarcity of catch and stock assessment data are more pronounced in the eastern and southern parts of the Mediterranean (Tsikliras et al., Reference Tsikliras, Dinouli and Tsalkou2013), and are sometimes influenced by social and political factors. Because of this, several Mediterranean areas have been characterized as being poor in fisheries data (Pilling et al., Reference Pilling, Apostolaki, Failler, Floros, Large, Payne, Cotter and Potter2008).
Gurnards are present in the Mediterranean Sea, comprising eight species (Richards & Jones, Reference Richards and Jones2002), which represent a commercially important by-catch in demersal fisheries. The population trends of these fish have not yet been quantified in the Mediterranean (Nunoo et al., Reference Nunoo, Russell, Bannermann and Poss2015) and, due to the alarming decline of traditional exploited fish stocks (e.g. Hake, red mulled, Anchovy (Vasilakopoulos et al., Reference Vasilakopoulos, Maravelias and Tserpes2014)), the exploitation of gurnard species is gradually increasing (FAO, 2007). The longfin gurnard Chelidonichthys obscurus (Walbaum, 1792) is one of the three major species landed in Tunisia, together with the streaked gurnard, Chelidonichthys lastoviza (Bonnaterre, 1788) and the tub gurnard Chelidonichthys lucerna (Linnaeus, 1758) (Boudaya et al., Reference Boudaya, Neifar, Taktak, Ghorbel and Bouain2007). This triglid species is relatively common in the Mediterranean Sea (Sanches et al., Reference Sanches, Blanco and Gancedo2002) and is a critically endangered species on the Turkish Mediterranean coast (Turan et al., Reference Turan, Ergüden and Gürlek2016).
Methods for delineating stocks are mainly based on data extracted from artificial and natural tags, leading to considerable advances in recent years. Developments in electronic tags, image analysis, chemical methods, parasitology and molecular biology have spurred profound changes in a large number of stock identification approaches (Cadrin et al., Reference Cadrin, Kerr, Mariani, Cadrin, Kerr and Mariani2014). Furthermore, numerous studies clearly demonstrate the merits of adopting a more holistic approach, combining information on life history traits, genetics, behaviour and biological tags to assist fish stock identification (Catalano et al., Reference Catalano, Whittington, Donnellan and Gillanders2014). Taken together, these techniques make it possible to overcome many of the limitations of single-technique approaches and provide reliable evidence for population structure (Lleonart & Maynou, Reference Lleonart and Maynou2003).
In this study, we examine the population structure of the longfin gurnard off eastern Tunisia using two approaches: fish morphometry and parasite assemblages. These two stock identification methods were chosen rather than other methods owing to their relative ease of application and low cost. Morphometrics can be used to discriminate differences among species by analysing variance in body shape (Cadrin et al., Reference Cadrin, Friedland and Waldman2005). Parasites can also be used to discriminate stock structure. Short-lived parasites yield information on short-term movement, while long-lived parasites yield information on long-term movement (Lester & MacKenzie, Reference Lester and MacKenzie2009).
We integrate the results obtained from the application of the two methodologies and then use these findings to address the appropriate spatial scale for stock assessment. The results of this study could provide useful information for more effective fisheries management.
Materials and methods
Host sampling
A total of 134 specimens of C. obscurus (table 1) were sampled randomly from the landings of bottom trawlers, operating in two fishing areas off the eastern coast of Tunisia – the Gulf of Gabès and the Gulf of Hammamet (fig. 1). Samples were collected in spring 2017 at depths ranging from 80 m to 150 m, using trawls equipped with a cod-end bag liner of 40 mm stretched mesh size. Fish were either kept fresh or deep frozen in plastic bags at −18°C until examination.
Map showing locations used to sample Chelidonichthys obscurus off Tunisia.
Descriptive data of Chelidonichthys obscurus in Tunisian waters, with sample size (N), range, mean and standard deviation (SD) of standard length (SL).
GH, Gulf of Hammamet; GG, Gulf of Gabès.
Morphometric study
A total of 14 morphometric characters were measured in this study (fig. 2), including: total length, standard length, fork length, head length, body height, first dorsal fin length, second dorsal fin length, anal fin length, pelvic fin length, eye diameter, pectoral fin length, length of the first pectoral fin ray, length of the second pectoral fin ray and length of the third pectoral fin ray. The measurements were repeated by three readers using digital callipers, accurate to the nearest 0.1 mm. Between-reader differences were minimized by calibrating repeated measurements performed on the same set of fish by each reader.
Morphometric variables measured on the body of Chelidonichthys obscurus. TL, total length; SL, standard length; FL, fork length; HL, head length; H, body height; D1, first dorsal fin length; D2, second dorsal fin length; AFL, anal fin length; EFL, pelvic fin length; ED, eye diameter; PFL, pectoral fin length; LR1, length of first pectoral fin ray; LR2, length of second pectoral fin ray; LR3, length of third pectoral fin ray.
Parasite study
To detect ectoparasites, the body of the fish, as well as the fins, oral cavity, nasal pits, eyes and gills, were thoroughly examined under a stereomicroscope with incident light. To study the endoparasites, internal organs (heart, liver, spleen, gall bladder and gonads) were separated and examined. The stomach, pyloric caeca and intestine were separated and opened longitudinally. Their contents were placed in vials containing saline solution (Cribb & Bray, Reference Cribb and Bray2010) and vigorously shaken. Endoparasites were removed when the contents settled. Prevalence and mean abundance were determined for each parasite species (Bush et al., Reference Bush, Lafferty, Lotz and Shostak1997). Only parasites species with prevalence >5% were included in the statistical analysis (Bush et al., Reference Bush, Lafferty, Lotz and Shostak1997).
To assess the potential role of diet, stomach contents were preserved in 7% buffered neutral formaldehyde. Food items were identified and counted. The importance of the different prey types was calculated using the frequency of occurrence (F), as described by Hyslop (Reference Hyslop1980).
Data and statistical analysis
To remove the size effect from the variation of morphometric characters due to growth and allow a comparative analysis, all morphometric characters were adjusted by adapting an allometric method based on the formula of Elliott et al. (Reference Elliott, Haskard and Koslow1995) using Madj = M (Ls/L0)b, where M is the measurement of the observed character, Madj the adjusted measurement of the character, L0 the standard length of the fish, Ls the overall mean of the standard length for all fish from all samples in each analysis, with b being estimated for each character as the slope of the regression of log M on log L0 using all fish in a given group. The results obtained from the allometric method were confirmed by testing the significance of the correlation between transformed variables and standard length.
Non-parametric analyses were performed to evaluate the differences in characters at the population level. Chi-square analyses and a posteriori multiple comparisons of proportions were used to test for significant differences in prevalence of parasites between zones. Kruskall–Wallis and a posteriori Dunn's tests or Mann–Whitney tests were used to analyse the effects between zones on the morphometric characters and the abundance of each parasite species (Zar, Reference Zar1984). Discriminant analyses, based on Mahalanobis distances, were used to find the differences between zones and identify the morphometric characters and parasite species responsible for these differences. Indeed, for each group of our sample, we can determine the position of the average representative point of all variables in the multivariate space defined by the variables in the model. These points are termed centroids of groups or centres of gravity. In each case, it is possible to calculate the Mahalanobis distances (respective observations) to each centre of gravity. Therefore, each case is classified within the group that is closest – that is, with the shortest Mahalanobis distance (Brodgar, 2000). This procedure was applied using the software package SPSS 18.
A non-metric multidimensional scaling method (nMDS) was applied to visually assess the differences in morphometric parameters and parasites between the studied zones. Similarity matrices were constructed based on Bray–Curtis' similarity. This method assigns a non-dimensional location to each group and calculates the distances between groups. A non-dimensional plot is produced based on these distances, which reflects the similarities or dissimilarities among groups, so that similar groups are plotted close to each other and dissimilar groups further apart. These analyses were performed using PRIMER®-v6.
Results
Geographical variation using morphometric characters
Table 1 shows descriptive data for the sex ratio, mean length and standard deviation and length range of sampled specimens from the two studied zones. No sexual dimorphism is observed for any of the morphometric characters (P > 0.05). There is no significant correlation between any of the log-transformed morphometric variables and standard length (P > 0.05), indicating that the size effect is well taken into account.
A comparison of morphometric variables between localities allows us to identify three out of 13 morphometric characters as the most effective feature for differentiating samples (P < 0.05; table 2).
Comparisons of morphometric characters of Chelidonichthys obscurus from the Gulf of Hammamet and the Gulf of Gabès, Tunisia.
Mean values are given with standard deviation (±). N, number of examined fish; TL, total length; FL, fork length; HL, head length; H1, body height; D1, first dorsal fin length; D2, second dorsal fin length; AFL, anal fin length; EFL, pelvic fin length; ED, eye diameter; PFL, pectoral fin length; LR1, length of the first pectoral fin ray; LR2, length of the second pectoral fin ray; LR3, length of the third pectoral fin ray.
* Level of significance with P < 0.05.
** Level of significance with P < 0.01.
Using multivariate discriminant analysis to separate C. obscurus between the studied zones, we find that the first discriminant function explains 100% of the variance (eigenvalue = 1.321). There is a significant overall effect of locality (Wilks' lambda = 0.431, F39.63, P < 0.01). Fish data points are distributed mainly along the first axis. Dimensionality tests reveal that the two zones are significantly separated in both dimensions (X 2 = 101.89, d.f. = 4, P < 0.01). Fish are classified correctly according to the two component communities, with an accuracy of 87.2% (table 3). The scores for individual fish exhibit a clear discrimination between two groups, one including fish from the Gulf of Hammamet and the other including fish from the Gulf of Gabès. The Mahalanobis distance does not reflect any significant difference. The importance of each morphometric character with respect to discrimination between groups is evaluated as the contribution of each variable to the sum of Mahalanobis distances. Using this approach, we find that the body height, the pectoral fin length and first dorsal fin length represent the highest contribution to the differences in the composition of morphometric characters between the studied zones. Most of the patterns revealed by non-parametric discriminant analysis are validated by nMDS of the three morphometric variables identified as the most significant characters for differentiating samples (fig. 3).
nMDS ordination of Bray–Curtis similarities computed on three significant measured variables of the two groups.
Discriminant analysis classification showing numbers and percentages of fish classified in groups from each studied zone.
Geographical variation using parasite markers
Four short-lived parasite markers were identified, including two polyopisthocotylean monogeneans, Plectanocotyle major Boudaya, Neifar & Euzet, 2006 and Triglicola obscurum (Euzet & Suriano, 1974) obtained from the gills, as well as a monopisthocotylean, Paradiplactanotrema sp. from the oesophagus and a digenean Synaptobothrium caudiporum (Rudolphi, 1819) from the digestive tract. We also collected two long-lived parasite markers, larvae of the nematode Hysterothylacium aduncum (Rudolphi, 1802) and Trypanorhyncha larvae recovered in the digestive tract. The isopod Gnathia sp. was also found in the gills with a prevalence <5%, but is not included in the statistical analysis.
The mean abundance of P. major shows a significant difference between the studied zones. This parasite is more abundant in specimens from the Gulf of Gabès. Furthermore, P. major is more prevalent in fish from this latter locality. Paradiplactanotrema sp. and S. caudiporum are more prevalent and abundant in specimens from the Gulf of Hammamet (table 4).
Taxonomic composition, prevalence and mean abundance of helminth parasites of Chelidonichthys obscurus from two studied zones off the Tunisian coast.
GG, Gulf of Gabès; GH, Gulf of Hammamet.
* Level of significance with P < 0.05.
** Level of significance with P < 0.01.
A multivariate discriminant analysis, used for the separation of C. obscurus between the studied zones, shows that the first discriminant function explains 100% of the variance (eigenvalue = 0.406). There is a significant overall effect of locality (Wilks' lambda = 0.711, F15.01, P < 0.01). Fish data points are distributed mainly along the first axis. Dimensionality tests reveal that the two zones are significantly separated in both dimensions (X 2 = 37.98, d.f. = 3, P < 0.01). Fish are classified correctly according to the two component communities with an accuracy of 72.2% (table 5). Scores of individual fish display a clear discrimination between two groups, one including fish from the Gulf of Gabès and one including fish from the Gulf of Hammamet. The importance of each parasite species with respect to discrimination between groups is evaluated as the contribution of each variable to the total sum of Mahalanobis distances; this approach shows that T. obscurum is the most influential monogenean species in determining the position of samples belonging to the first group (Gulf of Gabès). Specimens from the second group (Gulf of Hammamet) are correlated positively with Paradiplactanotrema sp. and S. caudiporum. Most of the patterns observed in the discriminant analysis were validated by nMDS analysis, suggesting the presence of two main groups (fig. 4).
nMDS ordination of Bray–Curtis similarities computed on four parasites affecting the two fish population groups.
Results of discriminant analysis showing numbers and percentages of fish classified in each studied zone.
Discussion
Geographical comparison of morphometric characters and discriminant analyses allow the separation of two stocks of the longfin gurnard C. obscurus off eastern Tunisia: one stock in the Gulf of Hammamet and the other stock in the Gulf of Gabès. After allometric adjustment, morphometric data of C. obscurus show differences between the studied zones that appear solely related to body shape variation. These results are corroborated by morphometric discriminant analysis, with a relatively high (=87.2%) mean percentage of specimens correctly assigned to original populations. Discriminant analyses show that the most effective morphometric trait between the two studied zones is body height and length of the pectoral and the first dorsal fin, with significantly higher values in the Gulf of Gabès (P < 0.001). These differences in body form may reflect differential microhabitat use (Murta et al., Reference Murta, Pinto and Abaunza2008), which could be attributed to variations in prey availability (e.g. Marcil et al., Reference Marcil, Swain and Hutchings2006) and fish mobility (e.g. Swain & Holtby, Reference Swain and Holtby1989). To test this hypothesis, we investigated the stomach contents of our samples. This approach reveals that individuals from the Gulf of Gabès, with a larger number of prey (X 2 = 5926, d.f. = 1, P < 0.001), seem to be more voracious than individuals from the Gulf of Hammamet. Furthermore, the frequency of occurrence of prey in stomach contents is higher for all preys in specimens from the Gulf of Gabès compared to the Gulf of Hammamet (table 6). Based on the hypothesis of Colloca et al. (Reference Colloca, Ardizzone and Gravina1994), who proposes that the diet of gurnards is partly a reflection of prey availability, it appears that, in the Gulf of Gabès, the preferential prey of C. obscurus are more available than in the Gulf of Hammamet. Moreover, Swain & Holtby (Reference Swain and Holtby1989) concluded that differences in body form are adaptive, providing greater prolonged swimming performance. Thus, in the Gulf of Gabès, C. obscurus would appear to be more active in searching areas, where more preys occur, and have a more robust body that would be better adapted for high swimming performance. Swimming behaviour of fish can also vary according to the physicochemical characteristics of the water (Costa & Cataudella, Reference Costa and Cataudella2007) and hydrodynamic constraints. Indeed, the Gulf of Gabès is subject to one of the strongest tidal currents in the Mediterranean, which leads to considerable sea-level oscillations (Sammari et al., Reference Sammari, Koutitonsky and Moussa2006). These morphological divergences between populations could also be related to genetic differentiation or interactions between environmental and genetic factors (Favaloro & Mazzola, Reference Favaloro and Mazzola2006).
Frequency of occurrence of main prey categories of Chelidonichthys obscurus from two studied zones off the Tunisian coast.
GG, Gulf of Gabès; GH, Gulf of Hammamet.
** Level of significance with P ≤ 0.01.
In our study, a morphometric approach provides valuable information for the identification of C. obscurus stocks and the description of their spatial distributions. Several studies have demonstrated the ability of the morphometric method to discriminate marine fish stocks and determine their spatial distributions (e.g. Geladakis et al., Reference Geladakis, Nikolioudakis, Koumoundouros and Somarakis2017; Afzal Khan & Nazir, Reference Afzal Khan and Nazir2019). Nevertheless, numerous studies recommend the integration of information from multiple approaches to draw holistic conclusions about stocks that could then be used in fisheries management. Accordingly, we make use of parasites as an additional biological marker for the discrimination of C. obscurus stocks on the eastern coast of Tunisia. Parasites have been widely used as biological tags to provide information on the stock discreteness of their fish hosts in many parts of the world ocean (Cantatore & Timi, Reference Cantatore and Timi2015; Weston et al., Reference Weston, Reed, Hendricks, Winker and Van Der Lingen2015). However, the selection of suitable parasites has been discussed by several authors and a number of specific guidelines have been presented in the literature highlighting the requirements for a parasite species to be considered as a biological tag candidate (see MacKenzie & Abaunza, Reference MacKenzie and Abaunza1998, Reference MacKenzie, Abaunza, Cadrin, Kerr and Mariani2013; Mackenzie & Hemmingsen, Reference Mackenzie and Hemmingsen2015). These criteria encompass: (1) easy detection and identification of the parasite; (2) significantly different levels of infection in the host at different geographical locations; (3) a relatively long life span of the parasite within the host; (4) no significant impact of the parasite on the behaviour or survival of the host; (5) a parasite with a single-host life cycle; and (6) the presence of the parasite should be temporally stable. It should be noted that these guidelines are simply recommendations rather than established rules. Thus, compromises are necessary because a single parasite species rarely satisfies all of these criteria (MacKenzie & Abaunza, Reference MacKenzie and Abaunza1998). For instance, in contradiction to criterion (5) described above, many studies have demonstrated that anisakids and trypanorhynchs represent the best markers since these parasites require at least three host species to complete their life cycle (MacKenzie & Abaunza, Reference MacKenzie and Abaunza1998; Lester & MacKenzie, Reference Lester and MacKenzie2009; Marcogliese & Jacobson, Reference Marcogliese and Jacobson2015). There is also a general consensus that long residence time in the fish host is the most important criterion for an effective parasite marker (Lester & MacKenzie, Reference Lester and MacKenzie2009). Furthermore, long-lived parasites are commonly recommended for studies on stock identification (criterion (3)). However, researchers should be prudent when using certain long-lived markers. Due to the cumulative patterns of parasites as fish grow, these long-lived markers impose some restrictions on the methodology (Poulin, Reference Poulin2004). Consequently, ontogenetic differences in parasite burdens can be incorrectly attributed to a locality effect, so fish size must be taken into account as a potential confounding variable in the interpretation of spatial patterns and stock structure (Cantatore & Timi, Reference Cantatore and Timi2015).
In this study, long-lived parasites are not discriminant. Besides, no significant spatial differences are observed in the prevalence and abundance of nematode larvae H. aduncum and Trypanorhyncha. On the contrary, the short-life-span parasites P. major and Paradiplactanotrema sp., and the digenean S. caudiporum, appear to be more discriminative. This hypothesis is supported by the morphometric study. According to Lester & MacKenzie (Reference Lester and MacKenzie2009) and MacKenzie & Abaunza (Reference MacKenzie and Abaunza1998), short-lived parasites (e.g. ectoparasites and many adult endoparasites) are likely to be of little value for stock discrimination. Based on these results, two hypotheses can be put forward. The first hypothesis is that digenean and monagenean parasites might be used in some cases as a source of information on the discreteness of fish host stocks. These parasites satisfy the majority of the criteria required for an effective biological tag, including being easy to detect and identify, showing significantly different levels of infection within the study area. Moreover, such parasites appear to have no significant impact on the behaviour or mortality of the host, even though we could not explicitly test these effects. The only major disadvantage of using monogenean or digenean parasites as biological tags is their short resident times in their hosts. However, for many parasites whose life cycle is unknown, as is the case of the parasites identified in this study, the residence time in the host should be investigated. Furthermore, the generation time of monogeneans is highly variable (Cribb et al., Reference Cribb, Chisholm and Bray2002). Several monogenean species exhibit short life cycles (Tubbs et al., Reference Tubbs, Poortenaar, Sewell and Diggles2005; Lackenby et al., Reference Lackenby, Chambers, Ernst and Whittington2007). On the other hand, a prolonged life span (more than a year) is observed in some monogeneans (e.g. Diplozoon paradoxum von Nordmann, 1832 and D. minutus (Pallas, 1770) (Bykhowskii, Reference Bykhowskii1957, Bovet, Reference Bovet1967)).
The second hypothesis involves considering these parasites as non-discriminant. In this case, future research is required using complementary techniques, such as otolith microchemistry and genetics. Such approaches may improve our understanding of the global stock structure of the longfin gurnard to suitably inform regional fisheries management.
Acknowledgments
This study was conducted in the Marine Biodiversity and Environment Laboratory (LR18ES30) at the University of Sfax, Tunisia. This research was partially supported by the 17/TM/21 project. We would like to thank Dr. W. Boussellaa for her assistance in collecting parasites. The authors acknowledge Pr. J.C. Dauvin for his precious help and Dr. M. Carpenter for the English revision of the manuscript. They also thank reviewers for their comments and suggestions.
Financial support
None.
Conflicts of interest
None.
