INTRODUCTION

Marine fishery generates an average of 38.5 million tons of bycatch per year worldwide, a number that represents more than 40 % of total catches (^{Davies etal., 2009}). In the South Atlantic and the Gulf of Mexico, the bycatch rates of shrimp fishery reach 64 % and Brazil accounts for about 60 % of all marine catches (^{Davies et al., 2009}, ^{Keledjian et al., 2014}).

The adverse ecological impacts of fishing on marine ecosystems are widely recognized, resulting in a reduction of catches, especially in bottom trawling fishery (^{Alverson et al., 1994}; ^{Kelleher, 2005}; ^{Davies et al., 2009}; ^{Keledjian et al., 2014}).

In the state of Santa Catarina, southern Brazil, the largest national producer of fish and crustaceans from marine fisheries (^{IBGE, 2019}), the situation is equally worrisome. In the municipality of Penha, our study site, the proportion between fish biomass and seabob shrimps (Xiphopenaeus kroyeri Heller, 1862) varies from 0.3:1 to 4.7:1 (^{Barrilli et al., 2021b}).

In this area, the sciaenid Paralonchurus brasiliensis (Steindachner, 1875), a non-targeted fish of low commercial value that is sympatric to X. kroyeri, represents the highest bycatch in terms of landed biomass (^{Branco et al., 2005}; ^{Bail and White, 2007}). Paralonchurus brasiliensis is an epi-benthic species that inhabits coastal regions of sandy-muddy bottoms, up to 100m deep, and has been reported all along the Western Atlantic from Panama to Argentina, including all the Brazilian shoreline (^{Menezes and Figueiredo, 1980}).

Paralonchurus brasiliensis was identified in the stomach contents of top predators in coastal ecosystems, such as large fish (Trichiurids, Sharks and Rays) (^{Barbini and Lucifora, 2015}), marine turtles (^{Awabdi et al., 2013}), marine birds (gannets and petrels) (^{Colabuono and Vooren, 2007}) and marine cetaceans such as Tursiops truncatus Montagu, 1821 (^{Milmann et al., 2016}).

Previous studies of the stomach contents of P. brasiliensis revealed that its prey consisted mainly of crustaceans, echinoderms and polychaetes, being considered as a key species in marine food webs (^{Soares and Vazzoler, 2001}; ^{Branco et al., 2005}; ^{Araújo et al., 2014}). Besides, P. brasiliensis has been shown to be a relevant bioindicator in studies on bioaccumulation and trophic transfer of trace elements along the marine food chains (^{Kehrig et al., 2013}).

In view of this, a concern arises especially with the impact of trawling fishery on non-target species and its reflex on the trophic chains in coastal ecosystems. As feeding is determinant in the abundance and structure of the fish community (^{Piet et al., 1998}), detailed studies of their natural diet contribute to understanding their trophic ecology and existing functional groups (^{Criales-Hernández et al., 2006}).

Thus, through the study of the feeding habits of a species, it is possible to define its functional role in the food chains (^{Barreiros and Santos, 1998}), besides allowing inferences about the availability and accessibility of organisms to food resources (^{Wootton, 1990}).

According to ^{Pillay (1952)}, the best method to assessing the food habits of a species is the analysis of the stomach contents. This procedure is common in fish trophic ecology studies. For this, ^{Berg (1979)} suggests the use of a large number of samples because, although tropical fish exhibit great trophic plasticity, their diet may vary with season, latitude and ontogeny.

Several methods of analysis of the stomach contents of fish were developed to better understand the food spectrum, pointing out the importance of each item and the probable interrelationships associated with environmental characteristics. The most used traditional methods for the analysis of fish diets are numerical, volumetric, gravimetric, as well as subjective estimation techniques such as the point system, used in analyzes of the stomach volume (^{Hynes, 1950}; ^{Hyslop, 1980}). In addition to these methods, it is common to use the relative importance index (IRI) to infer the importance of prey (^{Pinkas et al., 1971}) and, more recently, stable isotopes of carbon and nitrogen (^{Buchheister and Latour, 2010}). However, regardless of the chosen method, the most appropriate is the one that allows a better comparison and interpretation of the results (^{Berg, 1979}).

Despite the ecological importance of P. brasiliensis found in the shrimp fishery area (X. kroyeri), studies on the feeding habits of this species are still not well documented (Robert et al., 2017). Thus, in this study we analyze the feeding habits and trophic ecology of P. brasiliensis in the South Atlantic of Brazil as a contribution to understanding its biology.

MATERIALS AND METHODS

Study area

The study was conducted in the municipality of Penha (26° 46' 10'' S; 48° 38' 45' W), Santa Catarina, Brazil (Fig. 1).

Figure 1 Trawling areas used for the Xiphopenaeus kroyeri shrimp in Penha, South Brazil. Isolines show isobaths of 10, 20 and 30 m and the cross-sectional area where our sampling was obtained. 

Diet description

The composition of the diet was described according to the methodology used by several authors (e.g.^{Branco et al., 2005}; ^{Machado et al., 2014}), applying a quantitative method, the Points Method (PM), a qualitative method, the Frequency of Occurrence (FO) used by ^{Hynes (1950)}, ^{Berg (1979)}, ^{Williams (1981)}, ^{Wear and Haddon (1987}) and ^{Haefner Jr (1990)} and a combined method, the Alimentary Index (IAi) of ^{Kawakami and Vazzoler (1980)}.

For PM, five degrees of repletion were adopted, and the respective percentage points were attributed: empty (0.00 %), partially empty (<0.25 %), medium (<0.50 %), partially full (<0.75 %) and full (1.00 %), according to the volume of food items of the stomach contents of each specimen.

The percentage points totaled for each food item "i" (% PM) were expressed according to ^{Berg (1979)} and ^{Williams (1981)}:

where, PM = Points Method; aij = number of points of the food item "i" in the stomachs examined; A = total number of points for all items.

The relative FO of each food item "i" (% FO) was obtained through the equation:

where, FO = Frequency of Occurrence; ni = number of stomachs containing item "i"; N = total number of analyzed stomachs.

The importance of each food item in the diet, the Alimentary Index (IAi) was calculated according to the equation proposed by ^{Kawakami and Vazzoler (1980)}:

where, IAi = Alimentary Index; i = 1, 2, 3 ... number of food item; FO = frequency of occurrence of the item "i" in %; PM = total of the average points of item "I".

The calculated variables (% FO, % PM and % IAi) quantify the importance of each item in the diet of the species under study, considering the volume occupied in the stomach and the amount of different food items (^{Castro et al., 2015}). With this method, we estimate the food preference of the species Paralonchurus brasiliensis.

For the analysis of the general composition of the food items, the groups with IAi less than 0.01 % were grouped as Others, in order to facilitate the interpretation of the results and provide a better understanding of the most representative food items in the diet of P. brasiliensis.

RESULTS

Temperature

The bottom water temperature fluctuated between depths and seasons (Fig. 2a-b). The lowest recorded value was 16 °C, in the 30 m, in the winter of 2013 and the highest of 28 °C, at 10 m, in the summer of 2014. No significant differences were found between different isobaths (F2-33 = 0.3253; p> 0.05), but significant differences exist between the seasons (F3-8 = 74.3370; p<0.05).

Figure 2 Average variations of the temperature and salinity of the bottom water per season of the year (a) and in the isobaths (b), in Penha, South Brazil. 

Stomach repletion

Of the 1219 specimens evaluated, 412 (33.80 %) were females, 256 (21.00 %) males and 551 (45.20 %) were sexually indeterminate (Table 1). The length amplitude varied between 2.7 and 23.3 cm, with an average of 12.3 ± 3.5 cm and the weight between 0.17 and 116.65 g, with an average of 19.32 ± 18.64 g.

Table 1 Frequency of occurrence of empty and with content stomachs in females, males and juveniles with undetermined sex of P. brasiliensis in Penha, South Brazil. 

Macroscopic analysis revealed that among the 1219 stomachs of the dissected specimens, 1167 (96.73 %) contained identifiable prey and/or food remains (Table 1).

Considering the degree of repletion of all the stomachs of the original sample (Fig. 3), 419 (34.37 ± 2.33 %) were full, 314 (25.76 ± 3.28 %) partially full, 307 (25.18 ± 5.70 %) medium and 127 (10.42 ± 2.33 %) partially empty. Only 52 stomachs analyzed were empty (4.27 ± 0.88 %).

Figure 3 Stomach repletion degree from a sample of Paralonchurus brasiliensis from Penha, South Brazil. Legend: F (full), PF (partially full), M (medium), PE (partially empty) and E (empty). 

Sample representativeness

The accumulated curve of the food items of P. brasiliensis showed that the occurrence of all expected items, also pointed out by the Jackknife 2 estimator, was reached with 465 stomachs examined (p<0.05), confirming that the number of dissected stomachs was sufficient to describe the diet of the species (Fig. 4).

Figure 4 Cumulative trophic diversity curves of Paralonchurus brasiliensis from Penha, South Brazil. 

Diet description

Table 2 shows the relative importance of food items and the food preference of P. brasiliensis. It is observed that the food items can be divided into two large groups, the first consisting of the preys that are prevailing in this diet and the second, of preys less expressive in their diet. Table 2 also shows that this species is an opportunistic predator, with a narrower trophic niche and specialized, mainly, in the ingestion of polychaetes, crustaceans, and ophiuroids.

Table 2 Diet composition of Paralonchurus brasiliensis (n = 1167 stomachs with contents), from Penha, South Brazil. Values of relative frequency of occurrence (% FO), mean points (% PM) and feeding index (% IAi) for each food item are presented. Items marked with + means values <0.1 %. 

These macroinvertebrates formed the dominant prey group, both in the frequency of relative occurrence of food items (36.1 %, 34.9 % and 11.2 %), in the average of the points (50.55 %, 33.12 % and 9.27 %), and in the percentages of the alimentary index (58.43 %, 37.01 % and 3.32 %) (Table 2).

The second group, the least representative in the general diet composition of P. brasiliensis, involved Nematoda, Osteichthyes, sand, organic matter and those denominated as other (Algae, Cnidaria, Nemertinea, Sipuncula, Mollusca and Scales). Together they contributed with the lowest fractions of FO (17.80 %), PM (7.07 %) and IAi (1.24 %), as shown in Table 2.

The ANOSIM statistical test showed that there were no significant differences in the diet of P. brasiliensis among isobaths (R = -0.07, p > 0.05), and seasons (R = -0.12, p ≥ 0.05).

DISCUSSION

Studies involving the analysis of stomach contents of fish are fundamental instruments to determine their eating habits and their niche in the food chains (^{Barreiros and Santos, 1998}; ^{Machado et al., 2014}). In addition, food availability and utilization indicate patterns of distribution, migration, reproduction of fish, and ecosystem changes (^{Amaral and Migotto, 1980}; ^{Zavala-Camim, 1996}; ^{Soares and Vazzoler, 2001}; ^{Araújo et al., 2014}).

The methodology used in the analysis of the feeding habits of P. brasiliensis in the South Atlantic of Brazil was adequate, as in several studies involving natural fish feeding (^{Branco et al., 2005}; ^{Machado et al., 2014}; ^{Castro et al., 2015}), although the recognition of prey was restricted to large groups. However, as suggested by ^{Branco and Verani (1997)}, the most important is to determine the food spectrum of the species and the relation among them.

The fluctuations recorded for temperature and salinity of the bottom water are typical of coastal zones (^{Matsuura, 1986}). They may be related to rainfall and to the contributions of the Itajaí Açu River (which is distant about 20 km to the South of the study area), the Coastal Water mass, low salinities (<34 %), high temperatures (19 to 28 °C), and the seasonality of the South Atlantic Central Waters between spring and summer, which favors the increased of primary coastal productivity (^{Soares-Gomes and Figueiredo, 2002}).

P. brasiliensis may occupy this area due to the combination of favorable abiotic factors and the diversity of macrobentos as food, as suggested by ^{Souza et al. (2008}). Also, areas closer to the beach offer mechanisms that provide large amounts of nutrients, since they are regions of high primary productivity, that guarantee resources at lower and higher trophic levels (^{Araújo et al., 2014}).

The food items accumulation curve, constructed in the present study, reached asymptote confirming the adequacy of the number of stomachs dissected to describe the diet of P. brasiliensis. In this sense, ^{Zavala-Camin (1996)} points out that when carrying out fish diet studies, the size of the sample should not be small enough to lose relevant information but should not be large enough to promote a waste of resources. For ^{Teixeira and Gurgel (2002)}, small samples reduce the information required whereas exaggerated samples can mean a waste of work and loss of time.

The analysis of the degree of repletion and the stomach contents of P. brasiliensis revealed a diversified and constant feeding. The high incidence of full or nearly full stomachs with no significant differences between juveniles and adults in the diet of this species was described by ^{Soares and Vazzoler (2001)}, as well as ^{Branco et al. (2005)}, who did not find significant differences when comparing the diet between males and females.

The feeding habits of P. brasiliensis verified in the present study coincide with previous descriptions for the Brazilian South region (^{Vazzoler, 1975}; ^{Branco et al., 2005}; ^{Robert et al., 2007}). The results showed that P. brasiliensis presents a broad trophic spectrum, consuming more frequently organisms of the group Polychaeta, Crustacea and Ophiuroidea, closely related to the sediment. In general, the members of the Sciaenidae family present a wide food diversity, invertebrates being the fundamental dietary items of some marine fish species (^{Amaral and Migotto, 1980}; ^{Branco et al., 2005}).

Despite a wide food spectrum, our study pointed out a strong trend towards the predation of polychaetes by P. brasiliensis. This pattern of diet is consistent with the records of several authors (^{Amaral and Migotto, 1980}; ^{Braga, 1990}; ^{Soares and Vazzoler, 2001}). According to ^{Robert et al. (2007)}, juveniles and adults of P. brasiliensis share the same food items. Therefore, it is possible to infer that there are no significant variations in the composition of food items during the ontogenetic development of this species.

The use of polychaetes as a more important item may be a consequence of a greater availability of these prey in the habitats of P. brasiliensis and may be related to the type of sediment in this area, where there is a predominance of silt-clay. According to some studies, areas with this type of sediment contribute to the retention of organic matter and, therefore, to the abundance of organisms that feed on it, such as polychaetes (^{De Almeida et al. 2012}, ^{Araújo et al., 2014}).

One fact that caught our attention is that the decapod crustacean X. kroyeri, target species of the South Atlantic trawling fishery, is not among the more important food items in the diet of P. brasiliensis. This can be related to the fact that these shrimps are often buried in the sediment (^{Simoons et al., 2013}), becoming less available to benthic and benthopelagic consumers (^{Wakabara et al., 1993}). In addition, from the anatomical point of view, P. brasiliensis has a morphologically small, inferior and almost horizontal mouth (^{Fischer et al., 2011}), that could hinder the ingestion of this type of prey.

As in our studies, ^{Braga et al. (1985)} identified Ophiuroidea among food items that constitute the diet of P. brasiliensis throughout the year, along with polychaetes, and crustaceans, and to a lesser extent fish, and scales. Several studies point to the ophiuroids among the most exploited resources in the diet of this species and of other demersal Sciaenidae (^{Vazzoler, 1975}; ^{Soares and Vazzoler, 2001}; ^{Branco et al., 2005}; ^{Robert et al., 2007}; ^{Araújo et al., 2014}).

Sand ingestion may be considered accidental and reinforces the species' benthic habit (^{Barrilli et al., 2021a}). However, it is not possible to discard the hypothesis that sand ingestion is voluntary and continuous when considering the diversity of macrobenths found in marine sediments such as Cnidaria, Nematoda, Polychaeta, Crustacea and Ophiuroidea, among others (^{Branco and Verani, 1997}; ^{Araújo et al., 2014}). Additionally, the occurrence of sediment particles in the stomach contents suggests that the species actively captures prey that relates to areas of the bottom.

The presence ofintact nematodes in the stomach contents of P. brasiliensis recorded in our studies suggests a parasitic relationship, as demonstrated by the results of ^{Turra et al. (2012)}, who described nematodes in the stomachs of around 50 % of the intestines of Sciaenidae Menticirrhus americanus (Gill, 1861). Other researchers have also reported the presence of this parasite in the P. brasiliensis digestive tube (^{Pinto et al., 1992}; ^{Luque et al., 2003}). However, ^{Araujo et al. (2014)} considered Nematoda as part of the natural diet of eight species of Neotropical Actinopterygii, including Sciaenidae P. brasiliensis, Isopisthus parvipinnis (Cuvier, 1830), and Stellifer brasiliensis (Schultz, 1945).

The relationships between organic matter, sand and other items (Algae, Cnidaria, Nemertinea, Sipuncula, Mollusca, Ostheichthyes and scales), recorded in our study were also described as stomach contents of several marine species by different authors (^{Branco et al., 2005}; ^{Araújo et al., 2014}). For these researchers, both organic matter and sand were included as food items in the function of their considerable occurrence and volume values. However, the other items were not considered because they were treated as rare occurrences or accidental ingestion.

On the other hand, this may be indicative of the opportunistic character of P. brasiliensis, due to its great trophic plasticity, especially when there is little availability of the main food source, as indicated by ^{Gerking (1994)}. According to ^{Araujo et al. (2014)}, most tropical fish are indeed trophic opportunists with a broad food spectrum, and this allows them to adapt rapidly to the availability of other resources. These authors also affirm that this high degree of eurifagia can be associated with the high faunistic diversity on one side and the smaller biomass of each species on the other.

The frequency of occurrence and volume of certain food items in the stomach contents of P. brasiliensis suggests their availability in the environment and how these food resources are being exploited by the species, following the Optimal Foraging Theory (^{MacArthur and Pianka, 1966}). Taking into account that P. brasiliensis captures a large number of small preys, this habit must compensate for the energetic gain in search, capture, and manipulation of these foods, as described by ^{Bozza and Hahn (2010)}. As foraging is dependent on prey density, this exercises an in controlling subtidal organisms' populations (^{Wellenreuther and Connell, 2002}).

The response of the ecosystem to bottom trawling depends on whether the abundance of benthos responds to top-down or bottom-up trophic control (^{van Denderen et al., 2013}). These researchers have demonstrated that there may be a positive effect of trawling, with increased fish abundance, when their prey (benthos) are also resistant to trawls in controlled systems from the bottom up. An intermediate disturbance can contribute to the maintenance of species, as it removes those that are dominant and therefore allows a better balance between them (^{Connell, 1978}).

The results obtained in this study allow us to conclude that P. brasiliensis is a demersal-benthic species with a broad trophic spectrum, predatory, carnivorous, opportunistic, and predominantly invertivorous. With this paper we intend to provide information that contributes to the understanding and discussion of the bycatch of the trawling fishery of X. kroyeri shrimp in the trophic chains of coastal ecosystems, using the case of P. brasiliensis as a study model.

CONCLUSIONS

The population of P. brasiliensis showed an intense foraging activity and a diversified feeding habit, with higher consumption of polychaetes, crustaceans, and ophiuroides. The presence of sand in the content suggests accidental ingestion since this item can adhere to the surface of the other items. The species might be characterized as a demersal-benthic species, predator, opportunistic, of the broad trophic spectrum and having an important ecological role in the maintenance of marine trophic web in coastal in the South Atlantic of Brazil. We emphasize that studies on ecological aspects are necessary, as they assist in knowing the role of species in environments, conservation issues, and fisheries management.