1. Introduction
Fish and seafood are highly valued and play an important role in human diet because of their content of high quality protein and essential amino acids, polyunsaturated fatty acids and micronutrients [1]. In 2017, according to FAO estimates, fish provided at least 50% of the total animal protein intake of humans in several West African countries including Togo [2]. In these countries, artisanal catches could represent the bulk (60%) of the domestic fisheries catches [3]. Unfortunately, catches are not expected to maintain their current growth rate in the next decade. The region is expected to become more dependent on imports to satisfy demand, reaching 41% of consumption in 2026 [4]. In Togo, in the last decade, the production of most species by artisanal fisheries was stable, or fluctuating around the average without a real tendency to increase [5]. More worrying, total catches of many valued food fishes (silver catfish, blackchin tilapia…) have decreased because of overfishing, drought and watershed degradation [3] [6]. It was therefore, strongly recommended that the necessary efforts be made to achieve a rational management of the stock and to find ways of improving the level of productivity of coastal lagoons and estuaries [5].
The study of the food and feeding habits of fish is a subject of continuous research, because it plays an integral part in the development of a successful fisheries management program on fish capture and culture [7] [8]. Indeed, feeding behavior based on the analysis of stomach contents is widely used in fish ecology as an important tool for understanding the role of fishes in aquatic ecosystems since they indicate relationships based on feeding resources and indirectly indicate community energy flux [9]. From a practical standpoint, information on quantity and quality of food consumed and the feeding behavior pattern is needed to understand the predicted changes that might result from any natural or anthropogenic intervention. In addition, knowledge of the feeding ecology of commercial as well as non-commercial fish species is essential for implementing a multispecies approach to lagoon and estuary fisheries management [10].
The bagrid catfish species, notably Chrysichthys nigrodigitatus Lacépède, 1803 and C. auratus Geoffroy Saint-Hilaire, 1809 are important members of fresh and brackish water fish highly sought after for their flavor and chemical composition [11]. They have been reported from practically all the river and lagoon systems of Africa within latitudes 25˚N and 25˚S and from Tanzania in the east to Senegal in the west [12] [13]. Bagrid catfish species have great commercial and economic value in West African countries [14] [15] and ensure food security and the livelihoods of local populations. Studies on the fauna in inshore habitats revealed that Chrysichthys spp could represent, at certain periods of the year, the highest biomass of the littoral ichthyofauna, accounting for 17% to 43.8% of total catches [5] [16] [17]. The most recent works on feeding ecology come mostly from the lagoons and rivers of Ghana [18] and Nigeria [15] [19] [20] [21]. In Togo, studies on the bioecology of fish, which have mostly been limited to the description of fisheries, are very rare and old [14]. Accordingly, very little information is available on diet of brackish and freshwater fishes in the country. The present study aims to characterize the diet of C. nigrodigitatus from the hydrosystem Lake Togo-Lagoon of Aného through a qualitative and quantitative description of the stomach contents and its temporal and ontogenetic variation.
2. Materials and Methods
2.1. Study Area
The Lake Togo-Lagoon of Aného complex is part of a lagoon system set in south-eastern Togo located between latitudes 6˚14'38'' and 6˚17'37'' North and longitudes 1˚23'33'' and 1˚37'38'' East. It stretches from the villages of Dékpo and Sévatonou in the north-west to of Aného in the south-east. It covers 64 km2 and includes the Lake Togo (46 km2), the Lagoon of Togoville which is a channel of 13 km length parallel to the coast with a width varying between 150 and 900 m, the Lake Zowla (6.55 km2) and the Lagoon of Aného in the South-East. It communicates with the ocean at Aného. In addition, the Lagoon of Aného communicates with the Mono river in the East via the Gbaga channel. The Zio and Hahorivers are the main tributaries of the hydrosystem [22]. The Lake Togo watershed enjoys a subequatorial or Guinean climate with two rainy seasons composed of a large one (mid-March to mid-July) and a small one (mid-September to mid-November) alternated by two dry seasons composed of a large one (mid-November to mid-March) and a small one (mid-September to mid-November). The most popular economic activities around the lagoon system are fishing, agriculture and livestock. Phosphorite mining takes place in this watershed with discharge of all kinds of mining waste. Figure 1 shows the location of the study area and sampling points.
2.2. Sample Collection and Determination of Stomach Contents
A total of 332 fishes were collected monthly at two landing stations over an annual cycle, from January to December 2017, in collaboration with professional fishermen of the lagoon complex [15] [23] [24]. The gear types often used by fishermen include drag net, hook and line, bottom-set gillnet and bottom-set traps. They were set the day before between 07:00 and 08:00 am; then raised and the fish collected the next day between 05:00 and 06:00 am. Samples were immediately
Figure 1. Location map of the study area showing the sampling points.
packaged in a cooler containing ice and transported to the Laboratory of Animal Ecology and Ecotoxicology of the University of Lomé. After identification according to Lévêque et al. [25] and Stiassny et al. [26], each specimen was measured to the nearest millimeter (total length), then weighed (total weight) using a KERN type balance with an accuracy of 0.01 g. The fishes were then dissected using stainless dissection materials. The entire gastrointestinal tract was removed and stored in 5% formalin to stop decomposition and consolidate the preys until their contents analysis [15] [23] [24] [27].
The digestive tracts were opened longitudinally and their contents were emptied into petri dishes. The decomposed preys were counted based on the number of heads, eyes, legs etc. Recognizable preys were identified to the lowest possible taxon under a stereomicroscope OPTIKA LAB-20 with a magnification range of 7 to 45 times. For identification of food items, the standard literature on the systematics of aquatic invertebrates such as Grassé et al. [28], Durand and Lévêque [29] and Bouchard [30] was referred to.
2.3. Sex Ratio
Fish sexes were determined by macroscopic observation of the gonads after dissection. The sex ratio (SR) was calculated using the Equation (1) [15] [31] [32]:
(1)
2.4. Data Analysis of C. nigrodigitatus Stomach Contents
Numerous food indexes have been described and used to analyze and quantify the importance of different prey items in the diets of aquatic species [7] [33]. Some of them were used in the present study:
1) The vacuity index (VI) was calculated using the Equation (2):
(2)
where Nes is the number of empty stomachs and Nts is the total number of stomachs examined.
2) Frequency of occurrence (%Fo). It was expressed according to the Equation (3):
(3)
where, Ns is the number of stomach containing a given prey and Nt is the number of non-empty stomach. This index was interpreted according to the scale of Pillay [34] as modified by Gning [35] as follows: Fo > 30%, the prey consumed is considered preferential prey, 10% < Fo < 30%: the prey is qualified to be secondary and Fo < 10%: the prey can be considered to be accidental.
3) The numerical abundance index (%N). It was calculated using the Equation (4)
(4)
where Ni is the number of a given prey in the stomach and Nt is the total number of preys ingested.
4) The gravimetric abundance index (%W). The following equation allowed its calculation
(5)
with Wi is the weight of a given prey in the stomach and Wt is the total weight of preys ingested.
5) The index of relative importance (IRI) was used to assess the relative importance of a prey category in the diet by minimizing bias caused by each food index [36]. It was calculated as follow:
(6)
where, %N is the numerical abundance index, %W is the gravimetric abundance index and %Fo is the frequency of occurrence [23] [37].
2.5. Statistical Analysis
The Chi-square test (χ2) was used to compare the observed sex-ratios to the theoretical sex-ratio of 1:1 [15] [38]. The importance of the food spectrum was determined using the Shannon-Wiener diversity index (H). A value of H = 0.5 indicates a very low diversity:
(7)
where, Pi is the numerical abundance of the prey “I” [15] [39]. The specific richness of ingested prey was determined using the Margalef index (d). This value ranges between 1 et 4.5:
(8)
where S is the number of species and N is the number of individual preys [15] [40] [41]. The degrees of similarity of diets or dietary overlap between sexes, size classes and states of sexual maturity were determined using the Schoener index (SI):
(9)
where, Pxi = the numerical abundance of the prey “i” in the diet of fish group x and y [42]. This index varies from 0 to 1. The diets are considered to be significantly similar when SI value is superior or equal to 0.6 (SI ≥ 0.6) [43] [44] [45].
The ontogenetic variation in diet was highlighted by the Cluster Analysis (CA) of size classes on the basis of numerical abundance of the prey items. The temporal variation of the diet was assessed by principal component analysis (PCA) on the basis of the numerical abundance of the prey items [23] [35]. The CA and PCA were performed using STATISTICA 6.1 software.
3. Results
3.1. Biometric Parameters of C. nigrodigitatus
3.1.1. Sexratio and Proportions of Sexes
The overall result of the monthly sex ratio shows that of the 332 fishes examined 195 (58.73%) were males and 137 (41.27%) were females giving a sex ratio (M/F) of 1.42 or 1:0.70. A chi-square analysis of the result shows that this sex ratio is significantly different from the theoretical one (1:1). Thus, the number of males was significantly higher than females in the population of C. nigrodigitatus examined (χ2 = 10.13; p = 0.0014).
3.1.2. Variation of Morphometric Parameters
The largest C. nigrodigitatus caught during the study was a ripe male. It measured 53.50 cm and weighed 1.660 kg. The total lengths, all sexes combined varied between 7.95 and 53.50 cm (average size 25.10 ± 7.66 cm) while the total weights ranged from 13.29 to 1660.50 g (average size 171.73 ± 178.41 g). A perusal of the result indicates similar morphometric parameters in both sexes with total length and weight averages respectively 24.25 ± 8.01 cm and 167.69 ± 207.17 g for males and 26.32 ± 7 cm and 177.48 ± 127.23 g for females (Table 1).
Table 1. Morphometric parameters of C. nigrodigitatus from Lake Togo-Lagoon of Aného.
Figure 2. Frequency of lengths (a) and weights (b) in C. nigrodigitatus individuals.
However, large and small individuals are poorly represented in the samples. The population exhibited a unimodal distribution skewed towards middle sizes, with fishes measuring 15 to 35 cm and weighing between 10 to 130 g dominating the catch the two sexes (Figure 2). These fish are composed of juveniles, pre-adults (young individuals) and adults.
3.2. Diet Composition
3.2.1. Vacuity Index
Of the 332 stomachs examined, 99 were empty corresponding to an overall vacuity index (VI) of 29.82%. The vacuity index varied according to sex with 34.36% for males and 23.36% for females.
The lowest vacuity indexes were observed in January (20.59%), February (16.33%), March (22.22%) and August (21.74%) (Figure 3). These periods correspond to the dry season in accordance with the climatic calendar of the coastal zone of Togo.
In general, for both sexes, the vacuity index values increase with the size of the fish (Figure 4). The lowest values were recorded in young individuals belonging to 7 - 14 cm size class (24%) while the highest was observed in adults measuring between 49 and 56 cm (75%).
3.2.2. Overall Composition of the Food Items
Analysis of the contents of 233 full stomachs yielded a food spectrum comprised of 23 types of preys that can be divided into 7 major groups: fish, molluscs,
Figure 3. Temporal variation of the vacuity indexes.
Figure 4. Variation of the vacuity index according to fish sizes.
crustaceans, insects, annelids, vegetables and sediments. Qualitatively, Table 2 indicates that C. nigrodigitatus of the lagoon complex has a wide range of prey. Indeed, the frequency of occurrence (%Fo) indicated that, shrimps (Farfantepenaeus notialis) with a frequency of occurrence of 49.36% and unidentified fish (33.91%) were the most preferred food items consumed by the silver catfish. While the secondary prey consists of clams (Galactea paradoxa) (29.61%), mud (22.32%), fish (Ethmalosa fimbriata) (20.6 %), the decapod crustacean (Callinectes amnicola) (12.88%), the gastropod mollusc (Pachymelania fusca) (12.45%) and vegetable debris (16.74%). The other preys were accidental in the diet with frequencies of occurrence less than 10%.
Based on the numerical abundance (%N), the food spectrum was numerically dominated by juvenile and adult clams (44%) and shrimps (13.62%). They are followed by the unidentified fish (6.94%), E. fimbriata (6.24%), vegetable debris (5.28%) and gastropod mollusk P. fusca (4.06%). In term of gravimetric abundance (% W), the clam G. Paradoxa (40.17%), the shrimp F. notialis (26%) and the fish E. fimbriata (10.27%) were the most abundant prey. The decapod crustacean
Table 2. Prey found in C. nigrodigitatus stomachs from Lake Togo-Lagoon of Aného.
Note: %N: = numerical abundance, %W = gravimetric abundance, %Fo = frequency of occurrence, %IRI: index of relative importance.
C. amnicola, the unidentified fish, the gastropod mollusk P. fuscatus and the bivalve mollusk M. perna followed by respective gravimetric abundances (%W) of 6.57%, 5.95%, 3.99% and 2.11%.
The relative importance (IRI) of the different food items in the stomachs of the silver catfish based on numerical abundance (%N), gravimetric abundance (%W) and frequency of occurrence (%Fo) indexes is given in Table 2. The results indicated that clams G. paradoxa (40.49%) and shrimps F. notialis (35.85%) which together make up 76.34% of the IRI are the most important and preferred preys of C. nigrodigitatus. Unidentified fish (8.08%), E. fimbriata (6.24%), the decapod crustacean C. amnicola (2.15%), the gastropod mollusk P. fusca (1.83%), vegetable debris (1.68%) and juvenile clams (1.10%) represented the secondary prey of the species in the lagoon complex with a total of 21.02% of the IRI. The others prey with a total IRI of 2.64% are in the accidental prey category. The Shannon-Wiener diversity index (H) calculated on prey showed that the overall food spectrum of C. nigrodigitatus is very diverse (H = 3.34). In addition, the Margalef index (d) indicated that this spectrum has a high specific richness (d = 5.36).
3.3. Variation of Diet According to Sex
The collected fish are sorted by sex, based on the numerical abundance (%N) of their prey. Figure 5 indicates that the most abundant preys are clams (G. paradoxa) and penaeid shrimps (F. notialis) with respectively %N = 33.58% and 16.10% in females and %N = 34.75% and 11.99% in males. Data shows that males consumed many more clams (% N = 12.57%) than females (% N = 5.07%). Furthermore, juvenile gastropods are very poorly represented in males and absent in females. However, the Schoener dietary overlap index (C) calculated for male and female preys yielded a significant result (C = 0.85 > 0.6) indicating similarity between the diets of both sexes.
Figure 5. Variation of diet according to sex in C. nigrodigitatus.
3.4. Ontogenetic Variation of Diet
3.4.1. Variation According to Fish Size
According to the dendrogram obtained from the cluster analysis (Figure 6) two main clusters are identified: one composed of four isolated size classes (a, b, c and d) which are respectively 7 - 14 cm, 35.1 - 42 cm, 42.1 - 49 cm and 49.1 - 56 cm and the other (Group A) grouping three size classes (14.1 - 21 cm, 21.1 - 28 cm and 28.1 - 35 cm). This suggests that there is a significant ontogenetic change in the diet of the silver catfish in the hydrosystem Lake Togo-Lagoon of Aného. The Schoener indexes calculated between the different size classes (Table 3) shows that this change in the diet composition of the species according to size classes occurs gradually. Indeed, there is no significant similarity (C < 0.6) between the isolated classes a, b, c and d which represent the smallest sizes (a, b) and the largest sizes (c, d) of C. nigrodigitatus in the lagoon complex. However, the intermediate size classes belonging to group A are significantly similar to each other (C = 0.64 to 0.79). Furthermore, the difference between the largest class of the smallest sizes (14.1 - 21 cm) and the smallest class of intermediate sizes of group A (21.1 - 28 cm) is small (C = 0.59), suggesting that the first two
Figure 6. Dendrogram of cluster analysis of the size classes of C. nigrodigitatus according to the numerical abundances of the prey (size in cm).
Table 3. Schoener similarity indexes between size classes of C. nigrodigitatus.
classes of group A are closer to the smaller class sizes.
Figure 7 shows that the diet of fishes of size 7 - 17 cm is largely dominated by juveniles of G. paradoxa (58.97%) followed by fry, chironomid larvae and unidentified fish with 11.28%, 9.23% and 7.69% respectively. This diet was different from that of the next class (14.1 - 21 cm) where the food items are largely dominated by juvenile (23.61%) and adult (17.53%) clam G. paradoxa. They were followed by vegetable debris (9.66%), unidentified fish (8.77%) E. fimbriata (5.90%), M. perna (5.72%) and F. notialis (5.37%). Then, the data shows a gradual decrease of juvenile clams G. paradoxa from 58.97% for sizes 7 - 14 cm to 23.61% for sizes 14.1 - 21 cm. The diet changed from juvenile clam-dominated to adult clam-dominated. The latter appears in the diet along with E. fimbriata, C. hippos, T. zillii, P. fusca., juvenile gastropods, F. notialis, C. amnicola and vegetable debris. In addition, the numerical proportions of fry have significantly decreased from sizes 7 - 14 cm (11.28%) to sizes 14.1 - 21 cm (0.54%).
The food spectra of the size classes of group A are similar in that they contain almost the same preys. Indeed, G. paradoxa and F. notialis are the most abundant preys of the three size classes with respective proportions of 37.66% and 19.48% for sizes 21.1 - 28 cm, 42.65 % and 18.11% for the sizes 28.1 - 35 cm and 63.20% and 11.52% for the size class 35.1 - 42 cm. However, a gradual disappearance of some preys such as fry, amphipods, ostracods, chaoborid larvae, oligochaete annelids and filamentous algae were noted. Therefore, this group can constitute a transition between individuals of medium sizes and those of large sizes. On the other hand, the diet of individuals of sizes 42.1 - 49 cm is
Note: Preys codes are defined in Table 2.
Figure 7. Numerical abundances of prey for each size class of C. nigrodigitatus.
dominated by fish in particular E. fimbriata (35.48%), crabs C. amnicola (29.03%) and shrimps F. notialis while those of 49.1 - 56 cm consisted only of crustaceans, with a notorious dominance of shrimp F. notialis (77.78%). In addition, the data shows a decreasing trend in predation intensity as C. nigrodigitatus grew. It can therefore be concluded that up to 14 cm, C. nigrodigitatus, has a diet based mainly on juvenile clams (G. paradoxa). Then, the diet changes with the integration of a wide variety of preys when the catfish measure between 14.1 and 42 cm and before stabilizing around fish (E. fimbriata), shrimps (F. notialis) and crabs (C. amnicola).
3.4.2. Variation According to Sexual Maturity
Individuals of C. nigrodigitatus have been grouped into two categories namely immature and mature on the basis of their average size at first sexual maturity which is approximately 21 cm. Figure 8 shows that immature individuals have a diet dominated by juvenile clams (32.76%), diverse fish species (16.86%) and adult clams (13%). While, adult clams (43.38%), shrimps (17.82%) and various species of fish (15.44%) were the most abundant prey in the dietary spectrum of mature individuals. A comparative analysis of the two food spectra indicates that the observed difference is due to the disappearance of juvenile clams and the appearance of shrimps. This difference was confirmed by the Schoener dietary overlap index which revealed a significant difference between the diets of immature and mature individuals (C = 0.47 < 0.6).
The Shannon-Wiener diversity index (H) showed that immature individuals have a more diverse food spectrum (H = 3.41) than mature individuals (H = 2.86). However, the Margalef index (d) showed a slight increase in the specific richness of prey ingested by mature individuals (d = 5.64) compared to those consumed by immature individuals (d = 5.21).
Note: Preys codes are defined in Table 2.
Figure 8. Variation of diet according to the sexual maturity of the fish.
3.4.3. Temporal Variation in Diet
Temporal variation in diet was carried out by principal component analysis applied to the numerical abundances of prey. Results in Table 4 indicated that the
first two factor axes (Fact 1 and Fact 2) explain 74.35% of the total variance, suggesting that the factorial plan Fact 1 × Fact 2 can restore most of the information contained in the data.
The projection of months in the factorial plane (Fact 1 × Fact 2) shown in Figure9(a) indicates a slight temporal variation in the food items of C. nigrodigitatus. It is evident from this Figurethat the twelve months of the year form four groups: 1) March to August 2) November, December and February, 3) September and October, and 4) January. The projection of the preys in the same plane is presented in Figure9(b). This figure, compared to that of the months indicates that the months of the first group are characterized by a high abundance of shrimps in the food spectrum. The second group is mostly dominated by clams. The unidentified fish were abundant during September and October while juvenile clams dominated the food spectrum in January.
The detailed presentation of the monthly food spectrum of the species (Figure 10) shows that this variation in diet is mainly due to temporal variation in the abundance of shrimps in the ecosystem. Indeed, the first appearance of shrimps (F. notialis) in the diet of C. nigrodigitatus was noted in December at the beginning of the dry season. The predation on shrimps (F. notialis) increased gradually to a maximum reached in May before decreasing until it is canceled from
Table 4. Eigenvalues, total variances explained and cumulative variances.
Note: The prey codes in the legend are defined in Table 2.
Note: Preys codes are defined in Table 2.
Figure 9. Projection of months (a) and preys (b) in the factorial plane Fact 1 × Fact 2.
Note: Preys codes in the legend are defined in Table 2.
Figure 10. Monthly numerical abundances of different prey.
September to November. This variation seems to be inversely proportional to those of clams and fishes.
4. Discussion
Sex ratio and size structure constitute the basic information for assessing reproductive potential and estimating the stock size of populations [46]. Analysis of the sex ratio of C. nigrodigitatus in Lake Togo-Lagoon of Aného showed that the number of males was significantly higher than that of females. It can be explained by the fact that the fishing gear was not installed near the breeding grounds [19]. These results are similar to those obtained for the same species in some aquatic ecosystems in Nigeria [19] [31] [47]. However, they differ from those reported by Vanderpuye [17] in Volta Lake. These apparently contradictory results could be attributed to partial segregation of mature forms through habitat preferences and migration or behavioral differences between sexes rendering one sex more easily caught than the other [46]. According to Offem et al. [19], males migrate more frequently from breeding grounds to feeding areas located in shallow parts of the water body where they become vulnerable to be captured. Also, they suggest that females could take more shelter for incubation and protection of their offspring. Likewise, some samples obtained by Vanderpuye [17] suggest that females of Chrysichthys spp aggregate during certain periods of the year. Nevertheless, sex ratio divergence might also be explained by food availability and changes in environmental conditions. Nikolsky [48] observed that when food is limited, males predominate, with the situation reversing in regions where food is abundant.
The length frequency distribution of silver catfish collected in Epe Lagoon in Nigeria showed a triple mode suggesting that the species were made of three age groups during the study period [31]. On the contrary, in the present study, there was no evidence of more than one mode in the size structure which might suggest spawning periods. Similar results were obtained by Vanderpuye [17] in Volta Lake (Ghana). These findings conform to the assertion of year-round recruitment and breeding which is common in tropical species because of the relatively stable and elevated water temperatures in the tropics [49] [50]. The maximum total lengths and total weights recorded during this study (Lt = 53.50 cm and Wt = 1660.50 g) are similar to those observed by Idodo-Umeh [51] in the Ase River in Nigeria (Lt = 57.5 cm, Wt = 1500 g). However, these values are much higher than those recorded in C. nigrodigitatus from Epe Lagoon in Nigeria (24.30 cm and 178.87 g) by Lawal et al. [31] and from the Aiba Reservoir in Nigeria (25.6 cm and 288.7 g) by Atobatele and Ugwumba [15]. The variations of size (length and weight) in fishes may be due to a number of factors including season, habitat, genetic and environmental factors [37] [52]. In addition, these variations can be affected by gonad maturity, sex, diet and stomach fullness, health and preservation techniques [53].
The vacuity indexes obtained during this study appear to be relatively high (16.33% to 45.45%) and variable. The obtained vacuity indexes could be attributed to regurgitation during capture in fixed gillnets [54] and/or to excessively long fishing periods during which digestion continues [55]. The monthly variations in the vacuity indexes recorded in the present study (Figure 3) indicate the existence of a seasonal rhythm in the feeding activity of C. nigrodigitatus in Lake Togo-Lagoon of Aného [7] [24]. The periods of intense feeding activity probably correspond to those of the availability of preferred prey in the environment [56]. Thus, the rhythm of feeding activity in fishes is conditioned by temporal variations in the availability of food in the environment [57] and environmental factors such as the transparency of the water [58]. The increase of the vacuity index with size of the species (Figure 4) could be explained by the fact that young individuals appear to be more agile than adults in finding and capturing active prey [56].
Analysis of the stomach contents revealed that C. nigrodigitatus in Lake Togo-Lagoon of Aného complex fed on a wide range of food items notably bivalves and gastropods (mainly G. paradoxa and P. fusca), decapod crustaceans (F. notialis and C. amnicola) and various fish species. In addition, vegetables, amphipods, ostracods, insect larvae, annelids and mud were present but in very small quantities. These results in agreement with the findings of Lalèyè et al. [59] in the Lake Nokoué-Porto Novo Lagoon complex in Benin and those of Lawal et al. [31] in the EpeLagoon in Nigeria. However, C. nigrodigitatus has been reported to feed mainly ongastropod mollusc and ostracod crustaceans in Lekki Lagoon in Nigeria [60]. On the other hand, Oronsaye and Nakpodia [61] and Esenowo et al. [21] found that the diets of the silver catfish respectively from the Ethiope and Nwaniba Rivers (Nigeria) were dominated by detritus, plant matters, insects and fish remains. Atobatele and Ugwumba [15] reported that in the Aiba reservoir, the diet of C. nigrodigitatus is dominated by crustaceans (copepods, ostracods) and various species of insects. Overall, it emerges from these studies that C. nigrodigitatus has a very eclectic diet based mainly on benthic food resources, molluscs, aquatic larvae of insects, shrimps and crabs but also copepods, ostracods, filamentous algae and small fishes. These food categories can be found in the diet either in greater or lesser numerical importance depending on the biotopes and the hydrological season. Based on these results, C. nigrodigitatus is an omnivorous species with a carnivorous tendency. This omnivorous character of the species was reported by other studies in tropical aquatic ecosystems [19] [20] [31] [59] [61]. Furthermore, according to the classification established by Lauzanne [62], the species can be considered a secondary consumer consuming mainly benthic invertebrates, zooplankton and zooperiphyton. Thus, although C. nigrodigitatus has a morphology suitable for feeding on the bottom of water bodies, a wide variety of prey has been found in its stomach. This suggests the ability of this species to move to different aquatic habitats to capture different kinds of prey. The plasticity of their diets gives the silver catfish the power to adapt to various biotopes, to very different geographic and climatic conditions. In the different ecological conditions in which they may live, they will find food that suits them [19] [62] [63]. The presence of mud and sand in the stomachs of these fish could be due to accidental ingestion along with other foods [64].
The dietary spectra of the male and female are found to be similar in the present study according to the Schoener Overlap Index. This lack of variation in utilization of resources between sexes indicates that resource sharing or potential competition might exist between males and females. This finding defers from the results of Lauzanne [62] in which the males of C. nigrodigitatus fed more on planktonic crustaceans while the females fed more on benthic insect larvae in Aiba reservoir (Nigeria). According to these authors, this suggests a strategy for reducing intraspecific competition. The sex based differences in food consumption by mature individuals could be related to gender asymmetry in the energy invested in the development of primary and secondary sexual characters [65].
The variation in diet composition according to fish sizes was observed in the present study. This ontogenetic variation has also been reported by many studies in diet of C. nigrodigitatus and other fish species [15] [24] [45] [56] [59] [66] [67]. These changes in diet spectra may be an adaptation to reduce intraspecific competition between individuals belonging to different size classes [68]. However, this can be due to the opportunistic nature of the species in the environment or its ability to search for the preferred foods [45]. It is generally known that fishes preferentially consume the most abundant preys in the environment [62]. Nikolsky [48] had suggested that variation in the composition of food with age and size is a substantial adaptation towards increasing the range of food supply of their population by enabling the species to assimilate a variety of foods. Diet changes are often linked to anatomical changes in fish, allowing them to have a preference for large preys [24]. These large preys provide them with maximum energy for growth and reproductive functions [69].
The temporal variations observed in the diet are mostly dependent on the presence or absence of penaeidshrimps (F. notialis) in the hydrosystem. In fact, shrimps are completely absent from the diet of the catfish during September, October and November. This is due to the fact that these months correspond to a high water period in the hydrosystem via freshwater discharge leading to a considerable decrease of water salinity. Indeed, average salinity values in the hydrosystem dropped quickly from its maximum value of 14.8 g/l in March to 2.19 g/l in October [70]. This decrease in salinity triggers the return of shrimps to the ocean leading to their absence in the food spectrum of C. nigrodigitatus. Shrimpsre-appear in small quantities in December, corresponding to the start of the decrease in water level and the intrusion of marine waters into the lagoon. It is noted that in the absence of shrimps, the species feeds mainly on fish and bivalve molluscs (Figure 10). Some previous studies on silver catfish food habits also demonstrated a shift in diet depending on prey availability. Dada and Araoye [71] reported that plant materials were high in the months that coincide with rainy season. Choaborus and chironomid insects commonly found in the stomachs of the species also coincide with reproduction period of insect, especially dipterans. Similar observations were made by Atobatele and Ugwumba [15]. Thus, the present study may conclude that the occurrence of different types of food items in stomach and gut contents of C. nigrodigitatus in different months depend on their availability rather than selection by the catfish. The species is a non-migratory fish and remains in a specific habitat throughout its life and has to adapt the food available in the habitat during all seasons of the year.
5. Conclusion
The present study gives the first information on the food and feeding habits of C. nigrodigitatus in the Lake Togo-Lagoon of Aného complex. The findings confirmed that C. nigrodigitatus is an eclectic omnivore consuming mainly benthic invertebrates. Its diet varies depending on the size of individuals, their sexual maturity and time (months). However, no significant differences were observed between the sexes in terms of food preference. The dominance of Malacostraca and Mollusca in the diet is likely to be one of the more important considerations for future management plans. In this study, freshwater clams (juveniles and adults) constitute a considerable portion of the diet (44%), suggesting that the recent declining trend in G. paradoxa populations in the Mono River, in the Lake Togo and Lagoon of Anéhoas a result of habitat alteration and overfishing could have a negative effect on the catfish fisheries. Future studies should therefore focus on the macro-invertebrate fauna and its spatial and temporal distributions in the hydrosystem.
Acknowledgements
This study was co-funded by the International Foundation for Science (IFS) in Sweden and the organization for the prohibition of chemical weapons (OPCW) in Netherland (IFS Scholarship: I-2-A-6056-1). We also wish to express our gratitude to the Laboratory of Management, Treatment and Valorization of Waste (Laboratoire Gestion, Traitement et Valorisation des Déchets) of the University of Lomé (Togo).