Fisheries  and  aquaculture  play  an   important   role   as sources   of   food   and   income,  and  is  thus  critical in addressing global human food and nutrition insecurity, and economic demands (FAO, 2016). In 2013, fish contributed about 17% of the animal protein to the global human population, while  the global per capita fish consumption reached 20 kg in 2014 (FAO, 2016). Additionally, fisheries and aquaculture was a source of livelihood for 56.6 million people globally in 2014 (FAO, 2016). In Uganda, fisheries and aquaculture sector significantly contributes to economic growth. The sector contributed 3% to Gross Domestic Product (GDP) in the 2015-2016 financial year. Additionally, it directly employed over 1.2 million people, generated greater than 89 million USD in export earnings, and accounted for up to 50% of animal protein food (MAAIF, 2016a). However, the global fish production from wild fish stocks has generally stagnated, with most fisheries already fully exploited or over-exploited (FAO, 2014, 2016); yet human population is increasing, thus widening the fish demand-supply gap. In Uganda, this has been accompanied by a decline in per capita fish consumption to as low as 8 kg (MAAIF, 2017) which is below the 17 kg recommended by FAO (2012). With the fast growing human population, to about 35 million people (UBOS, 2016); this per capita fish consumption will further decline due to increase in the fish demand-supply gap. Therefore, aquaculture provides one of the plausible solutions for increasing fish production, and will subsequently lessen this demand-supply gap. Moreover, Uganda has a significant potential for aquaculture development, since it has numerous water bodies that cover up to 20% of the country’s total surface area (MAAIF, 2012). It also has favourable climate, good culture species such as Oreochromis niloticus, and availability of raw materials for feed (NAFIRRI, 2012).

As such, the Government of Uganda through her Agriculture Sector Strategic Plan (ASSP) 2016-2020, recognised the need to promote and support aquaculture in order to boost fish production to at least 300,000 tonnes annually by 2020 (MAAIF, 2016b). Cognizant of the need to meet this projected fish production target, augmenting investment in cage fish culture is one of the viable approaches that Uganda could adopt. This is because cage fish culture is known to increase fish production (Gentry et al., 2017), and can therefore result into higher returns. It, however, requires adherence to best practices including, adopting an ideal fish stocking density, in order to maximize production efficiency. Stocking density is critical since it directly influences the economic viability of fish culture enterprises (Osofero et al., 2009; Baldwin, 2010).

Stocking density directly affects the growth rate and survival of fish, and subsequently the  productivity  of  the fish culture operations (North et al., 2006; Osofero et al., 2009; Pouey et al., 2011; Mensah, 2013). In tilapia culture, better growth performance and survival rate are obtained in lower stocking densities (Sorphea et al., 2010). On the other hand, a positive relationship between stocking density and yield of Nile tilapia was reported by Watanabe et al. (1990), Cruz and Ridha (1991), Alemu (2003) and Gibtan et al. (2008). These studies reported higher yields with increasing stocking density. Therefore, stocking densities that result into higher fish yields, and subsequently higher economic returns from a cage culture enterprise, are ideal for catalysing cage aquaculture development. However, the inadequacy of empirical information on stocking densities that result into high growth and survival rates accompanied with higher yield continues to constrain cage fish culture productivity in Uganda. Consequently, the farmer’s capacity to operate cage culture systems is low, resulting into low cage fish production (Mbowa et al., 2016, 2017). Therefore, this study investigated the effects of stocking densities on the growth performance and survival of Nile tilapia (O. niloticus), reared in cage facilities, to contribute to the available cage aquaculture technical information, towards increasing cage production in Uganda.

 

 

Selection of study site

The study was conducted from March to August 2016 at Butiaba fish landing site, Piida Bay (20241.72 N and 313352.59 E) on Lake Albert in Buliisa District , Uganda (Figure 1). This experimental site is located in the Lake Albert Crescent Zone (LACZ) at an altitude of 616 m above sea level, approximately 68 km from the town of Hoima. Consideration of this area for the experiment was based on the baseline site suitability survey. The site is well suited for cage culture of Nile tilapia, because it is partly sheltered with an average depth of 8 m, water flow rate of 48 cm/s, pH of 7.5, Dissolved oxygen concentration 6.88 mg/L and temperature of 27.8°C, which are all within the prescribed parameters for cage fish farming (Howerton, 2001; Queensland Water Quality Guidelines, 2009).

 

 

Experimental design

Cages of 8 m³ (2 m × 2 m × 2 m), with metallic frames and an enclosure of 10 mm mesh size Nylon butane treated netting material, with 1 mm mesh size nylon cage liners were used during this experiment. The cages were secured in blocks (three per block) and fitted with cage liners as nursery cages for the small sized fish at the time of stocking. The cages were thereafter randomly stocked with monosex male Nile tilapia fingerlings, of uniform average body weight (4.07 g). The fingerlings were stocked at densities of 200, 250, and 300 fish/m³ per cage, with each stocking density  having  2 replicates. These stocking densities were based on an aquaculture policy brief, which suggested that fish stocking densities in a cage rearing facility can be high as 300 to 600 fish/m³ (NAFIRRI, 2012). After two months of the experiment, the liners were removed from the cages because the fish had increased in size. Further, this  was meant to abate clogging that would inhibit appropriate water exchange.

A locally produced commercial floating fish feed of 35 and 30% Crude Protein (CP) levels (based on the manufacturer’s provided information) was used during  the  experiment.  The  35%  CP  feed was in crumble and 2 mm pellet form, while the 30% CP feed was in a 3 mm pellet form. These different forms of the feed were adjusted based on the average body weight of the fish, and their gape (mouth) size. The 35% CP feed was initially fed to the fingerlings for a period of 3 months, while the 30% CP feeds were administered for the remaining experimental period. The fish were feed three times per day, from 09:00 am, at an interval of four hours.

Data collection

Water quality parameters

Physico-chemical parameters (Water temperature, pH, and dissolved oxygen concentration) were measured monthly within the cages, during fish sampling in the morning hours (0900-1100 h) of the experimental period. Water temperature and Dissolved Oxygen were measured using a digital probe (YSI model, 550A), while pH was measured using a digital combo pH meter (Hanna model, H1 98129).

Fish sampling

The fish in all cages were randomly sampled at the end of every month between 08.00 – 11.00 h from March - August 2016 from cages using a scoop net (1 mm mesh size). Thirty live fish were randomly scooped out of each cage unit. Their Total length (TL) and Standard (SL) length, and individual live body weights were measured using a measuring board, and an electronic weighing scale (Model JZC-TSC-03;V5.1-2010, Minimum: 2 g, Maximum: 3000 g), respectively. The cages were also inspected and cleaned during fish sampling.

Growth analysis

Growth changes in the fish were calculated based on the following parameters; Mean Weight Gain (MWG), Condition factor, Survival Rate and Feed Conversation Ratio (FCR).

Mean weights

Monthly mean weight of fish for each stocking density was determined by averaging the individual weights of a sample of 30 individual fish in each cage. MWG was calculated as:

MWG = Wt₂ – Wt₁

Where, Wt₁ = Mean Initial weight of fish, Wt₂ = Mean Final weight of fish.

Calculation of relative condition (K_n)

A total of 1,068 O. niloticus individuals were used in the calculation of Relative Condition Factor (RCF). To compare fish condition among cages with different stocking densities, relative condition factor (the ratio of observed individual fish weight to expected weight of an individual of a given length), was calculated using the formula: Kn = W/aL^b (Le Cren, 1951); where W is observed individual fish weight, L is observed individual fish total length, “a” is the intercept of the length-weight regression and “b” is the slope of the regression line. The values of a and b were obtained from the overall length-weight relationship (W=aL^b) derived by pooling data for all the cages across stocking densities. Length and weight data were log-transformed and the  resulting  linear  regression  fitted  by Least Squares Method using weight as the dependent variable. A minimum of 30 individuals per stocking density was considered acceptable for computing the RCF.

Feed conversion ratio (FCR)

The amount of feed used to produce 1kg of the fish biomass was calculated as:

FCR =Weight of feed given (g)/Fish weight gain (g)

Survival rate

Daily fish mortalities were recorded during the experimental period. At the end of the culture period, all the cages were emptied and the number of fish in each cage determined. The percentage survival rate was calculated as;

Survival Rate (%) = [Number of fish at harvest / Total Number of fish stocked] × 100.

Statistical analyses

One-way analysis of variance was used to test for differences in means of the growth performance and relative condition factor of fish, for the different stocking densities (200, 250 and 300 fish/m³). Turkey’s post-hoc test was used for multiple comparisons, to study any difference among treatment means. All statistical analyses were performed with SPSS for Windows (version 20.0) at 0.05 level of significance.

 

 

Water quality

The mean values of key water quality parameters monitored throughout the study are shown in Table 1. Temperature and Dissolved Oxygen concentration ranged from 27.45±0.47 to 27.46±0.48 (°C) and 3.44±0.80 to 3.63±0.67 (mg/L)  respectively, while pH ranged from 7.52±0.36 to 7.67±0.32. There were no significant differences in temperature (p=0.987), Dissolved oxygen concentration (p=0.451) and pH levels (p=0.124) across the different stocking densities in all cages.

 

 

Fish growth performance

The mean final weight, mean weight gain, mean relative condition, percentage survival rates and FCR of fish in all the treatments at harvest are presented in Table 2. The mean weight of fish in all stocking densities increased with time during the experiment, with slight variations in the mean weights during the first four months of culture. From the fifth month, fish stocked in cages at 200 fish/m³ had a higher mean weight compared to the cages with 250 and 300 fish/m³ stocking densities (Figure 2). The mean final weights of fish were 150.79±85.71 g, 127.82±68.43 g and 118.73±49.29 g in stocking densities  of 200, 250 and 300 fish/m³ respectively, from the initial mean weights of 4.07 g for all stocking densities (Figure 2). There was a significant difference (P<0.05) in mean final weights of the 200 and 300 fish/m³ stocking densities though no significant disparity (P>0.05) was observed in mean final weights of the 200 and 250 fish/m³ stocking densities.

 

 

 

Length-weight relationship and relative condition factor

The size of the O. niloticus ranged from 120-262 mm TL and 38.2-402.2 g TW. The relationship between length and weight derived by pooling data across all stocking densities (Figure 3) was highly significant (P<0.001). Similarly, the length and weight relationship for fish in each stocking density, obtained by the regression of length and weight of individuals across treatments was significant (P<0.001). For individual stocking densities, negative allometry was obtained with ‘b’ values deviating from the reference b value (b=3) (Table 3). The mean relative condition of O. niloticus differed among stocking densities ranging from 1.02-1.06. The fish exhibited a slightly higher condition than average in the 200 fish/m³ cages compared to the 250 and 300 fish/m³ cages (Figure 4). However, there were no significant differences in   condition   of   fish   among   the    stocking   densities (P>0.05).

 

 

 

 

Survival

Overall, the survival rates across the different stocking densities had an inverse relationship with stocking density (Table 2). The survival rates (%) did not change significantly (P>0.05) amongst the three different stocking densities, and these ranged between 92.98 and 94.19. The mean survival rates (%) were 94.19, 94.10 and 92.98 for 200, 250 and 300 fish/m³, respectively. The highest mean survival rate was observed in the 200 fish/m³ stocking density, 250 fish/m³ stocking density and lowest in the 300 fish/m³ stocking density at the end of the culture period.

Feed conversion ratio

This study did not determine the amount of feed lost during feeding, and therefore the actual amount of feed consumed by the fish. Consequently, the study  compared  the  FCR  across the three stocking densities  based  on  the  total amount  of  feed  given  during the experiment  for  the different treatments. The FCR ranged from 3.16 to 3.68, and increased with increase in fish stocking density. A relatively low FCR was obtained in cages stocked with 200 fish/m³, followed  by  250  fish/m³ and finally a higher FCR in 300 fish/m³ of a cage (Table 2).

 

 

 

 

The effects of stocking density were evident on the growth of Nile tilapia, in terms of weight gain and subsequently the final weight of fish. The best stocking density with regard to growth performance and feed conversion efficiency was 200 fish/m³ of cage.  However, the resultant final weight of the fish stocked at 200 fish/m³ of cage was still low. This study, therefore, recommends further research, with low stocking densities, which would result into a higher final weight of the fish, and consequently higher yields from the cages.

 

The authors have not declared any conflict of interests.

 

The authors are grateful to the World Bank and Government of Uganda that funded this research under the Agricultural Technology and Agribusiness Advisory Services (ATAAS) project through the National Agriculture Research Organization (NARO), Uganda and appreciate the  National  Fisheries  Resources  Research Institute (NaFIRRI) and Department of Zoology, Entomology and Fisheries Sciences of Makerere University for providing technical support. We are also grateful to Buliisa District Local Government (Fisheries Department) and the participating fisher folk during the study.