Evaluation of Hybridization in Two African Catfishes (Siluriformes, Clariidae): Exotic (Clarias gariepinus Burchell, 1822) and Native (Clarias jaensis Boulenger, 1909) Species under Controlled Hatchery Conditions in Cameroon

Laboratory of Applied Ichthyology and Hydrobiology, School of Wood, Water and Natural Resources (SWWNR), Faculty of Agronomy and Agricultural Sciences (FAAS), !e University of Dschang, P.O.Box 786, Ebolowa Antenna, Cameroon Department of Aquaculture, Institute of Fisheries and Aquatic Sciences at Yabassi (IFAS), !e University of Douala, P.O.Box 2701, Douala, Cameroon Department of Processing, Quality Control of Aquatic Products, Institute of Fisheries and Aquatic Sciences at Yabassi (IFAS), !e University of Douala, P.O.Box 2701, Douala, Cameroon


Introduction
Hybridization is the mating of genetically differentiated individuals or groups and may involve crossbreeding within a species (also known as line crossing or strain crossing) or crosses between separate species [1]. is breeding technique is used by aquaculturists in the hope of producing aquatic organisms with specific desirable traits or general improvement in performance [1].
ere are three types of hybridization in aquaculture: intraspecific, interspecific, and intergeneric hybridizations.
Interspecific hybridization has been practiced for a long time in a variety of fish species to increase growth performance, produce sterile animals, improve flesh quality, increase disease resistance, improve environmental tolerance, and manipulate sex ratios, as well as various other traits (overall improvement, hybrid polyploidization, experimental hybridization, and unplanned or accidental hybridization) to make fish more profitable to raise [1,2]. e African catfish Clarias gariepinus is an exotic species in Cameroon, where it is established in many rivers [3,4]. Introduced for aquaculture purposes, as early as the 1970s, in association or in monoculture, Clarias gariepinus was considered as a fish of the future for breeding due to its hardiness, its omnivorous diet, its rapid growth, and its much appreciated flesh [5]. is species was accidentally introduced into the Mbô Floodplain by escaping from the fish ponds located upstream of the University of Dschang [4]. In this plain, she cohabits with her sister Clarias jaensis, an indigenous catfish of Cameroon, and was the most abundant alien species [6]. e abundance of exotic species leads to the rarefaction or disappearance of native species because of the competition they engage in ecological niches [7]. Despite the consequences of "introductions" on native species [8], the Mbo Floodplain has become over the years an "attic" of catfish juveniles for the stocking of fish ponds in Cameroon since the introduction of this species; to the point where its endogenous sister C. jaensis would be in the process of disappearing in the catches. Furthermore, several studies have reported low rates of survival and growth in captivity in Clarias jaensis [9][10][11]. In addition to captivity, many failures have been recorded in the artificial reproduction of Clarias jaensis [12,13]. e first successful artificial breeding of this species in captivity reported very low hatching rates (<20%) [13,14], while male C. gariepinus × female C. jaensis hybridization appears to give a better reproductive performance and a good larval survival rate under the same conditions (Tiogué, unpublished). According to Cacot and Lazard [7], hybridization is a threat to the genetic heritage of native species because the hybrids produced are not sterile. If this hybridization is possible in the natural environment between the two species, it may eventually lead to the total disappearance of the two purebred Clariidae, especially if their hybrids are interbreeding [7]. e question that arises is whether the abundance of juveniles recorded in the Mbô Floodplain since the introduction of C. gariepinus should be due to the hybridizations between the two sister fish species? erefore, the present work aimed at evaluating the reciprocal hybridizations between C. gariepinus and C. jaensis on some reproductive parameters, survival and growth performances, and heterosis at larval and fry stages. .05 cm of mean total length)) captured by fishermen in the Menoua River at the Mbô Floodplain (5°10′-5°30′ N, 9°50′-10°10′ E) were used. All these broodstock were transported in a container of 10 liters to the fish culture station in Batié. Once in the GCI-IWA hatchery, same-sex broodstock of each species were acclimatized in a concrete tank (2 × 0.86 × 0.53 m 3 ) for one month during which they were fed twice per day with wheat bran.

Broodstock Selection.
Six females of each species having a swollen and soft belly and a protruding reddish genital papilla and whose oocytes of greenish color were easily obtained by light manual pressure of the abdomen [16] were selected and transported to the hatchery. In the hatchery, oocytes were taken from each female using a plastic cannula and their diameters were measured on millimetric paper. Six females of each species with an oocyte diameter of between 1.1 and 1.6 mm for C. gariepinus (Figure 1(a)) [16] and between 1.65 and 1.73 mm for C. jaensis (Figure 1(b)) [13] were selected and kept individually, fasting for 24 hours in 20 l plastic containers. ese containers were numbered according to the species, and each container was connected to a water renewal system at a rate of 2 l·min − 1 . All these containers were covered with a black plastic to keep the temperature inside. ey were supported by boards to prevent the escape of fish. ree males (Figures 1(a) and 1(c)) per species of weight more than 200 g were selected according to their well-developed genital papilla [16]. All males selected by species were kept in the same concrete tanks. A mercury thermometer was placed in one of the tanks for the control of the water temperature. Some external characteristics of the selected broodstock are presented in Figure 1.

Hormonal Induction.
e final oocyte maturation was induced using the ovaprim hormone at a dose of 0.5 ml of stock solution per kg of body weight. Each female received two injections at 6 h intervals, one as a preparatory injection (1/3) and a second 2/3 (effective). e administered dose was proportional to the weight of the fish. Males were not injected.

Harvesting Gametes.
Oocyte maturation and ovulation were obtained after a latency period of 18 h for C. gariepinus and 24 h for C. jaensis [13], in contrast to the defined time (13 hours) by Hogendoorn and Vismans [17] for a water temperature of 23°C. e milt was collected using the technique of partial gonadectomy described by Nguenga [18] and then diluted in physiological solution of 0.9% NaCl. Ovules obtained by manual pressure of the abdomen of females in the order of hormonal injection were collected in well-dried plastic bowls and then weighed on a "Sartorius AG (Goettingen, Germany)" balance of precision 0.1 g. Eggs from females of each species were mixed as well as the milt of males of the same species.

Fertilization and Incubation.
For each cross, the female was chosen first: 1 g of oocytes (100 to 183 oocytes) were taken from each species and introduced into a plastic bowl (three bowls per cross), for a total of twelve bowls. e milt was collected with a syringe and mixed with oocyte in bowls for fertilization between the female C. gariepinus × male C. gariepinus (Fcg × Mcg), female C. jaensis × male C. jaensis (Fcj × Mcj), female C. gariepinus × male C. jaensis (Fcg × Mcj), and female C. jaensis × male C. gariepinus (Fcj × Mcg). A fertilization solution (NaCl: 2 g·l − 1 and urea: 4 g·l − 1 water) was used to rinse the mixture for 1 minute. One bowl containing unfertilized oocytes of each cross was used as a control to evaluate the fertilization rate. e oocytes in each bowl were spread with a bird's feather. e 12 bowls labelled according to the crosses and the 4 control bowls were incubated in a concrete tank (2 m × 0.86 m × 0.53 m), equipped with a pump and a UV lamp (water height � 0.37 m and water flow � 5.18 ml·s − 1 ), and located inside the hatchery at an average temperature of 25°C. e rest of the eggs were fertilized, spread on racks, introduced into happas arranged in triplicate corresponding to each cross and fixed in a concrete tank filled with water and located outside the hatchery. Time taken by the eggs of the control bowls to become white was noted, and all other eggs were considered to be fertilized and counted to assess the fertilization rate. e incubator was covered with a black tarp to prevent nighttime cooling.
At the end of hatching (D0), the deformed larvae (2 heads, 2 tails or hard-swimming) or dead were counted and siphoned to separate them from normal larvae. e enumeration of the larvae was made by direct observation with the naked eye by daylight. e water temperature of the rearing tanks was recorded daily in the morning, at noon, and in the evening with a thermometer.

Larval
Rearing. For the total length of the larvae, about forty larvae were taken from each bowl using a Pasteur pipette and then placed on absorbent paper and their total lengths were measured quickly on millimetric paper at the end of the hatching (D0). e survivors of each bowl were counted daily and their total lengths were measured using millimeter paper until the end of vitelline resorption (D3). On the fourth day (D4), the larvae of each bowl were mixed with those of the same crosses incubated in the concrete tanks.

Larval Rearing in Happas.
Two hundred four-day-old larvae (D4) from each cross were counted and weighed on a Sartorius Competence 1 mg precision scale, and their total lengths were measured on millimetric paper. ey were then transferred into twelve happas of 0.2 cm 3 (0.5 cm × 0.5 cm × 0.8 cm) installed in a concrete tank of 250 m 3 with a water depth of 35 cm corresponding to the four previous treatments.
is tank was fertilized using pig droppings four days before the start of the test. e larvae were fed with local powdered food formulated at the station. e amount of food distributed was measured to twice a day assess the food consumption index. Every week, the larvae were captured with a plastic bowl and weighed with a 1 mg sensitivity balance. Five larvae of each happa were taken at random and weighed and their total length was measured until the end of the test (32 days of rearing (D32)). e happa was then removed from the rinsed water before reintroduction of larvae with each control fishing to avoid clogging.
where Nof � total number of fertilized eggs and Noi � total number of eggs incubated.

Hatching
Rate. e hatching rate was calculated using the following formula: where NL � total number of larvae obtained (normal, deformed, and dead).

Deformed Larvae and Survival
where Ld � number of larvae deformed at the hatchling.

Survival
Rate. e survival rate was calculated at the larval (D3) and fry (D32) stages: where Nf � number of larvae at yolk resorption and the number of larvae at the end of the experiment and Ni � number of larvae at hatching and loading.

Linear Growth
Total length (mm) Size heterogeneity (SH) (or coefficient of variation): where Xw � average weight (g) and SD � standard deviation.

Weight Growth
Mean weight (mW) in mg Mean weight gain (mWG) in mg where fmW and imW are the final and initial mean fish weights in mg. Average daily weight gain (ADWG) in mg/t: where t � duration of the experiment in days. Specific growth rate (SGR) in % day -1 : where Ln � natural logarithm. Coefficient of condition K: e formula of EI Bolock [19] was used to calculate this parameter: where mW � mean live weight in g and L � total length in mm.

Consumption Index (CI).
e consumption index was calculated using the following formula: where Qf � quantity of feed distributed (g).

2.9.
Heterosis. Heterosis was calculated as follows: where H1and H2 are the average weights or the number of survivors of the hybrids and P1 and P2 those of their parents at the end of the test.

Statistical Analyses.
Descriptive statistics such as mean, standard deviation, and percentage were used. Student's t-test was used to compare the values of the allometric coefficient b of the weight-length relationship with the isometry value b � 3. e one-way ANOVA was used to separate the means between the genetic types for the reproduction and growth parameters and the consumption index. When the differences were significant, they were separated by Duncan's multiple test at the 5% significance level. All analyses were done using SPSS software version 2013. All graphs were plotted using Excel 2013 software.

Fertilization and Hatching Rates in the Reciprocal
Hybridizations of Clarias gariepinus × Clarias jaensis in Controlled Conditions. Fertilization and hatching rates as shown in Figure 2 varied from one genetic type to another. ey were raised in the Fcj × Mcg hybrid followed by the parental Fcg × Mcg and the Fcg × Mcj hybrid which were not significantly different (p > 0.05). e parental crossing Fcj × Mcj recorded the lowest values significantly (p < 0.05). Figure 3 illustrating the deformed larvae rate in the hybridization of C. gariepinus × C. jaensis catfish shows that it varied from one genetic type to another. It was significantly lower (p < 0.05) in the hybrid Fcj × Mcg compared to other crosses that were otherwise comparable (p > 0.05). Figure 4 shows the survival rates at the larval and fry stages of the different genetic types. It appeared that at the larval stage (D3) (Figure 4(a)), the survival rate was very high (>80%) compared to the fry stage. e lowest rate (p < 0.05) was recorded by the hybrid Fcg × Mcj. It was comparable (p > 0.05) in all other crosses.

Survival Rates of Larvae at Different Stages of Development in Reciprocal Hybridizations of Clarias gariepinus × Clarias jaensis under Controlled Conditions.
In the fry stage (D32) (Figure 4(b)), the survival rates were less than 70% and the highest (p < 0.05) were recorded in the parental Fcg × Mcg and the hybrid Fcj × Mcg, followed by the hybrid Fcg × Mcj. e parental crossing Fcj × Mcj had a significantly lower survival rate (41.16%) (p < 0.05).

Linear Growth in Reciprocal Hybridizations of Clarias gariepinus × Clarias jaensis in Controlled Conditions
(1) Evolution of Total Length at Different Stages of Development in Reciprocal Hybridization of Clarias gariepinus × Clarias jaensis. e evolution of the total length of larvae of different genetic types studied is presented in Figure 5. (2) Coefficient of Variation or Size Heterogeneity. Figure 6 illustrating the evolution of the coefficient of variation (CV) of length as a function of age between the genetic groups studied shows that it has evolved into sawtooth whatever the genetic type and stage of development considered. is variation was very pronounced at 18 days of age (D18), where this coefficient was very low in the parental cross Fcg × Mcg and higher in the hybrid Fcg × Mcj. However, at 32 days of age, this coefficient of variation increased significantly in all genetic types. At this age, the hybrid Fcj × Mcg had a significantly higher coefficient of variation (p < 0.05), followed by the parental Fcg × Mcg and the hybrid Fcg × Mcj, which were otherwise comparable (p < 0.05).

(2) Characteristics of Weight Growth and Food Consumption Index (CI) Based on Genetic Type at 32 Days of Age in
Reciprocal Hybridization of C. gariepinus × C. jaensis in Happas. Weight growth characteristics and food consumption index based on genetic type to 32 days of age, as presented in Table 1, shows that the hybrid Fcj × Mcg and parental Fcg × Mcg had significantly higher growth performances (p < 0.05) and lower food consumption indices (p < 0.05) compared to other genetic types. Parental Fcj × Mcj had significantly lower growth performance and a higher feed conversion rate (p < 0.05).

Length-Weight Relationship and Growth Type of Fry in
Reciprocal Hybridizations of Clarias gariepinus × Clarias jaensis at D32. Table 2 shows the length-weight relationship and growth type of fry in reciprocal hybridizations Clarias gariepinus × Clarias jaensis at 32 days of captive breeding. It   followed that the coefficient of determination was significantly high (p < 0.05) regardless of the crossing considered. e allometric coefficient b was very low and comparable (p > 0.05) between all crosses; it was lower than the isometric value b � 3, thus giving to all crosses a negative allometric growth.

Heterosis of Growth and Survival in Reciprocal
Hybridizations of C. gariepinus × C. jaensis. Table 3 showing growth and survival heterosis values in reciprocal hybridizations of C. gariepinus × C. jaensis at different stages of development shows that negative heterosis was recorded only for survival at the larval stages (D0 to D3). e heterosis effect for growth was positive regardless of the type of growth and stage of development.

Reproductive Performances in Reciprocal Hybridizations of Clarias gariepinus × Clarias jaensis in Controlled
Conditions. Fertilization rates were comparable to those generally observed in Clariidae under controlled conditions: 80 to 95% in C. gariepinus [20], 87.1 to 95.2% in H. longifilis [21], and 79.25 to 87.57% in C. jaensis [13]. However, the parental crossing Fcj × Mcj had the lowest fertility rate (68.33%). is would be related to the species because many artificial reproduction failures had already been reported [12,13]. e hatching rates were lower in all genetic types (42.77 to 67.77%) compared to those obtained in H. longifilis (90.37 to 92.96%) by Agnèse et al. [22] and those from 68.07 to 72.82% reported in a closed loop by Tiogué et al. [23] in C. gariepinus induced by ovaprim. is should be partially due to the low oxygenation of the tanks caused by pump downtime and water renewal due to power cuts. ese rates, however, were close to 51-64 to 65.58% recorded by Tiogué et al. [23] in C. gariepinus induced by the pituitary gland in the closed circuit. However, they were higher than the results (1.10 to 19.38%) reported by Zango et al. [13] in C. jaensis induced by hCG and the pituitary gland of the same species and in the same station. is difference would probably be related to the types of hormones used. Indeed, these authors have used hCG and the pituitary gland of Clarias contrary to this study where it is ovaprim, a hypothalamic analogous hormone known to be more effective in inducing oocyte maturation than other hormones [23,24]. e high rates of deformed larvae recorded in all genetic types were significantly higher than those reported by Agnèse et al. [22] in wild and domestic strains of H. longifilis (7.5%), those of 6.44% recorded in plastic bowls in the cross of wild and domestic strains of C. gariepinus by Tiogué [25], and those reported in a closed loop (6.3%) by Tiogué et al. [23] in C. gariepinus induced by ovaprim. However, the hybrid Fcj × Mcg recorded the lowest rate of deformed larvae, thus making this crossover the most improving for this parameter.

Survival Performance in Reciprocal Hybridizations of Clarias gariepinus × Clarias jaensis in Controlled Conditions.
e survival rate of larvae for vitelline resorption was higher than 80% according to the results usually obtained in Clariidae at this stage of development.
ese results are similar to those reported by Agnèse et al. [22] from H. longifilis.
In the fry stage, the survival rates obtained were low (41.16 to 61.16%) compared to the results of 80 to 96% reported under controlled conditions in C. gariepinus and in H. longifilis fed on the living food Artemia from the first days of nursery [21]. ese results should be due to the low availability of plankton at the beginning of the test. Indeed, during loading, the water was not yet green so was poor in phytoplankton and consequently poor in zooplankton, essential food for the larvae at the beginning of rearing. In addition, the animals were fed with a powdered food made locally with local ingredients. ese results could also be explained by the effects of stress on animals during control fisheries. However, these rates were comparable to those of 42.68% reported by Tabaro [26] on larvae raised in basins in a closed circuit of water and fed with an artificial feed, of 30-50% reported by Campbell et al. [27] in a pond protected from a nylon paper fence and fertilized with chicken droppings and of 37.8-56.4% obtained from pure lines and crossbreeds of Clarias gariepinus and Heterobranchus longifilis in plastic aquaria and fed ad libitum twice daily with dried decapsulated cysts of Artemia sp obtained by Ataguba et al. [28]. Although relatively low, these survival rates remain interesting. Hybridization would have increased the survival rate in the hybrid cross Fcj × Mcg. Indeed, the positive value of the heterosis of survival reflects its superiority over pure parents.
is cross would be more suitable for improving the survival rate at this stage of development.

Food Consumption Index (CI) in Reciprocal Hybridizations of Clarias gariepinus × Clarias jaensis in Controlled
Conditions.
e index of food consumption is the efficiency of a food on fish production: the higher it is, the less the food is effective. In the fry stage, the values recorded for this parameter were higher (2.8 to 3.87%) than those of 0.45 to 1.87% obtained by Tabaro et al. [24] and Tabaro [26] in C. gariepinus reared at different stocking densities. For this author, this parameter varies according to the food used. Compared to parental hybrids, it was found that the hybrid Fcj × Mcg has consumed more food (2.91 ± 0.05%) in contrast to parental Fcj × Mcj (3.87 ± 0.09%). is is similar to the results obtained by Tonguthai et al. [29] with the artificial crossing of a female Clarias macrocephalus with a male C. gariepinus, in plastic aquaria whose hybrids were fast-growing and disease-resistant, with a high feed conversion rate.

Linear Growth in Reciprocal Hybridizations of Clarias gariepinus × Clarias jaensis in Controlled Conditions.
e hybrid Fcj × Mcg significantly increased in total length at yolk resorption and at day 32 compared to the parental crossing Fcj × Mcj. Indeed, according to Fontaine and Le Bail [30], during crossbreeding, the latter acquired high growth performances by genetic mutations that would have accelerated the morphoanatomical or morphophysiological changes affecting the growth of fish in the early stages of development. ese results are similar to those reported by Ruzzante [31] and Agnèse et al. [22] in H. longifilis. On the other hand, the positive value of growth heterosis explains   e heterogeneity of the size appeared in all the genetic groups at the end of the hatching (D0) could be explained on the one hand by the size of the oocytes of each species (1.5 mm in C. gariepinus against 1.7 mm in C. jaensis); the sexual dimorphism characteristic of Clariidae observed in the early stages of ontogeny, when males develop faster than females; and on the other hand by the cannibalism of the first larvae to hatch on the deformed larvae, and even the normal larvae 20 hours after hatching, and which consequently become more developed than the others [25].
us, the Fcj × Mcg hybrids recorded a more significant coefficient of variation comparable to that of the Fcg × Mcg parental crossbreeding. Whatever the stage of development, size heterogeneity has evolved into saw tooth between genetic types and regardless of age. Kestemont et al. [32] working with Dicentrarchus labrax and Perca fluviatilis explain this phenomenon by the fact that the final size heterogeneity is independent of the heterogeneity of the initial size.

Growth Weight in Reciprocal Hybridizations of Clarias gariepinus × Clarias jaensis in Controlled Conditions.
e final average weights obtained are lower than the final weights of 0.3 g reported by Campbell et al. [28] in a pond (10 × 10 × 1 m) protected by a nylon paper fence. ey are also lower than those of 0.466 g obtained by Tabaro et al. [24] and Tabaro [26] in larger basins (1.5 and 2.4 m 3 ) in a closed water circuit. ese low weights are due to the fact that in the present study, happas (0.5 × 0.5 × 0.7 m) with 0.5 mm mesh and a water depth of 35 cm used as nursery infrastructure had to be very narrow for free movement of fish or for good oxygenation. e plankton water poverty at the beginning of loading would also explain these results.
Moreover, by comparing the genetic types with each other, the best growth rates were observed in parental Fcg × Mcg and hybrid Fcj × Mcg crossbreeding.
is fast growing characteristic transmitted to the hybrid Fcj × Mcg was so dominant that it recorded the highest growth rate than all other genetic types.
is was explained by the heterosis that has been positive, thus showing the superiority of this heterozygous over the homozygous parents.
e K factor values for overweight characteristic of fish during rearing were low (0.79 to 0.84). ese results could be explained by the stress of the manipulations during the conduct of the test. However, they were close to those of 0.77 to 0.80 obtained by EI Bolock [19] in C. lazera fed with rice bran, blood meal, and kitchen waste and from 0.79 to 0.83% reported by Tabaro et al. [24] and Tabaro [26] in C. gariepinus reared at several densities and fed on an artificial feed. Despite the low condition factor K recorded by the hybrid Fcj × Mcg, its growth was greater than that of the other crossbreeding's. e allometric coefficient b was less than that in all crosses, thus showing negative allometric growth. In all these crosses, the growth is in favor of the length; therefore, the fish lengthen more than they gain weight. Moreover, the values of b have all been outside the range defined by the literature: between 2.5 and 4.0; or between 2.0 and 4.0 or between 2.5 and 3.5 [33]. In fact, the values of b less than 2.5 or b more than 3.5 would be caused by the large size of the samples arranged very closely [6]. e length-weight relationship was very significant in all crosses with a coefficient of determination R 2 more than 0.9 as reported in most silurids [10, 21-23, 25, 31].

Conclusion
e present study aimed at improving the production capacity of Clariidae fry in the CIG-AIW of Western Cameroon, through the evaluation of reciprocal crosses between Clarias gariepinus and Clarias jaensis under controlled conditions. Fertilization and hatching rates were better with parental (Fcg × Mcg) and hybrid (Fcj × Mcg) crosses; deformed larvae rate was significantly low in the hybrid Fcj × Mcg and the survival rate at the end of vitelline resorption was greater than 80% in all genetic types. It was significantly low in the hybrid (Fcg × Mcj) cross.
At the fry stage (at D32), the survival rate was in general less than 65%.
e highest values were recorded in the parental crosses Fcg × Mcg and hybrid Fcj × Mcg and the weakest in parental Fcj × Mcj, e hybrids Fcj × Mcg and Fcg × Mcj recorded significantly very high coefficients of variation of size compared to parental e condition factor K was low regardless of the genetic crossing considered. e length-weight relationship was significant with a coefficient of determination R 2 greater than 0.9. e growth was of allometric type in all crosses. e parental (Fcg × Mcg) and hybrid (Fcj × Mcg) crosses recorded significantly higher linear and weight growth performances and the higher food consumption index, followed by the hybrid (Fcg × Mcj) crossbreeding. e lowest rates were recorded at the parental crossbreeding' Fcj × Mcj Positive heterosis was recorded for both growth and survival In view of these results, Clarias jaensis × Clarias gariepinus hybridization is indeed possible and confirms the "hybrid threat" that occurred in the Mbô Floodplain Rivers on Clariidae species. To complete this study, a biomolecular study should be done on juvenile of Clariidae from the Mbô Floodplain.

Data Availability
e data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
e authors declare that they have no conflicts of interest. Advances in Agriculture 9