Growth performance and economic analysis of Nile tilapia ( Oreochromis niloticus ) fed on black soldier fly larvae ( Hermetia illucens)

The depletion of fish stocks in Lake Victoria is putting severe strain on fish supplies for human consumption and aquaculture feed formulation. Due to the fish stocks high demand and limited supply, fishmeal prices are rising, increasing the cost of feed for aquaculture. As a result, up to 70% of variable aquaculture expenditures are related to feed. Alternative feed sources, such as insect meals, have been heavily advocated for in light of the limited supply of fishmeal. This study replaced fish meal (FM) with black soldier fly larvae (BSF) meal (in Oreochromis niloticus diet, with replacement levels ranging from FM 0%, BSF 2 25%, BSF 3 50%, BSF 4 75% and BSF 5 100%). The feeding trial lasted 84 days and evaluated the effect of replacing BSF on fish's growth performance. Results indicate that all the fish accepted the experimental diets well. Regression on categorical variables analyses on growth performance revealed a significant relationship between the variables for length and weight respectively. For log length F (5, 99) = 628.02, P<0.05, with an R 2 of 0.96 and log weight (5, 99) = 977.61., P<0.05 with an R 2 of 0.97. In terms of economic analysis, O. niloticus fed with increasing levels of H. illucens was more profitable at replacement levels of 50, 75 and 100%. Based on these findings replacing fishmeal with Hermetia illucens by up to 50% is more cost effective for aquaculture production and can effectively contribute to higher yields and optimal O. niloticus growth performance.


INTRODUCTION
Protein demand is expected to rise due to global population growth. Unless deliberate action is implemented, increased demand will drive up prices, exerting pressure on animal production and food security (Smárason et al., 2019). Fish is a significant protein source and provides essential macro and micro-nutrients that can reduce malnutrition in Kenya. Globally, Asia is the highest producer and consumer of fish, with a per capita consumption of 24.1 kg per annum compared to Europe's and Africa's per capita consumption of 21.6 and 9.9 kg per annum, respectively (FAO, 2020). Fish contributes, on average, 19% of animal protein intake in Africa regions (Obiero et al., 2019a). In Kenya, the per capita consumption is low at 5 kg per annum (Obiero et al., 2019b), only a quarter of the world average of 20.5 kg per capita (Michèle, 2018).
The fishing sector is vital to ensuring food security and a livelihood for many Kenyans. Nile tilapia (Oreochromis niloticus) is the most commonly farmed fish species, accounting for around 75% of total production. The second most farmed species is African catfish (Clarias gariepinus), which contributes about 21% of total aquaculture production. Other cultivated species include goldfish (Carassius auratus), rainbow trout (Oncorhynchus mykiss), koi carp (Cyprinus carpio), and common carp (Cyprinus carpio auratus) (Munguti et al., 2014).
Capture fisheries has contributed significantly to food security over the years. In the recent past, aquaculture production has consistently contributed to food security, with a steady increase of 7.5% annually since 1970 (FAO, 2020). It is estimated that aquaculture production in Kenya needs to increase to 150,000 metric tons by 2030 to satisfy the increasing demand  or reach 550,000 tonnes to bring fish consumption per capita to the African average (Obiero et al., 2019b). The sector contributed 0.7% of Kenya's GDP in 2021 (KNBS, 2022). Despite the relatively small contribution to the GDP, the fishing industry is the lifeline for the Kenyan communities living around Lake Victoria and the Coast. The sector provides direct employment opportunities to over 500,000 people and indirectly supports about two million people (KMFRI, 2017). Total fish output grew from 151300 tonnes in 2020 to 163600 tonnes in 2021. Landings of freshwater fish increased by 8.5% to 136300 tonnes. Likewise, marine fish production increased by 6.2% to 27,300 tonnes. The increase was a result of the ongoing, optimal, and sustainable use of fishing resources. Fish landings totaled KES 30.4 billion in 2021, up from KES 26.2 billion in 2020. In 2020, freshwater fish caught brought in KES 20.6 billion in income, representing 76.8% of total fish landings (KNBS, 2022).
Despite aquaculture's potential to improve food security and income amongst resource-constrained households, the high costs of ingredients, particularly protein sources, used in fish feed formulation present a significant production challenge. Fish feeds contribute to over 60% of the total operational costs in a fish farm (Adeyemi et al., 2020). Protein is the major component needed in fish feeds. However, it is fish feed's most expensive dietary component (Craig et al., 2017;Mo et al., 2018). Fish meal has been an essential component in fish diets due to its high digestibility and acceptance alongside its high protein, essential amino acids and fatty acids profiles (Miles and Chapman, 2006;Cho and Kim, 2011;Nogales-Mérida et al., 2019). However, a significant impediment to the rapid growth of aquaculture is the scarcity of fishmeal, which raises its prices and increases production costs in aquaculture (Muin et al., 2017). As a result, low-cost and readily available feed ingredients are critical to ensuring food security through aquaculture. In advancing aquaculture, full utilization of insects offers an alternative protein ingredient for fish feeds (van Huis et al., 2015). Insects are natural diets for freshwater and marine fish species, including Nile tilapia. Insects are rich in protein, essential amino acids, lipids, minerals and vitamins. They can be reared in large numbers with minimal quantities of water or feeds (FAO, 2013), with promising results from the black soldier fly (Hermetia illucens), the common house fly (Musca domestica) and the yellow mealworm (Tenebrio molitor) (Sogari et al., 2019).
There is an excellent potential for H. illucens to replace fishmeal as a major source of protein for fish such as O. niloticus, catfish and rainbow trout (Stamer et al., 2014;Talamuk, 2016;Nyakeri et al., 2017). H. illucens, which contains about 49.5% crude proteins compared to the 40.3% in fishmeal (Veldkamp and Vernooij, 2019), offers a readily available protein source for aquaculture. H. illucens is widespread and considered a non-pest species with no capacity to carry pathogens, unlike M. domestica (Joosten et al., 2020;Shishkov et al., 2019). Under optimal conditions (30°C), a growth cycle takes 15 days to gain an average larva weight of 0.25 g. The larvae grow fast and have an excellent feed conversion rate of about 1.58 to 8.90 (Broeckx et al., 2021). They consume around 25 to 500 mg of fresh matter daily and feed on various substrates, from fertilizers to food wastes and leftovers (Hardouin and Mahoux, 2003). This reduces the dry matter content dramatically by 65-75% in an open field (Diener, 2010) but generates an additional valuable product, the prepupae. Their nutritional value is dependent on the feed substrates they consume. Crude protein levels range from 28 to 48%, and lipid levels from 12 to 42% ( Barragan-Fonseca et al., 2017;Liland et al., 2017). They have an amino acid profile similar to fishmeal and can be an excellent source of lipids subject to the feed the larvae are bred on (Liland et al., 2017;Wang et al., 2017;English et al., 2021). Based on its low cost of production, biological value, availability, fast growth and nutrient utilization, the H. illucens is quite promising in lowering aquaculture costs through its use as a protein source in aquaculture feeds.
Consequently, this research examined the effect of H. illucens larva meal on the growth of O. niloticus compared to fishmeal. It evaluated the cost implication when H. illucens larva meal replaces fishmeal in fish feeds. The study results may help farmers make more informed choices on the ingredients needed to formulate cheaper balanced diets with an excellent performance on tilapia growth rate reducing production costs in fish farming.

Study location
The research was done at Jaramogi Oginga Odinga University of Science and Technology (JOOUST), located in Bondo town, Siaya County, Kenya at coordinates 0.0926° S and 34.2587° E. Feed ingredients and formulated feeds were taken to the Kenya Agricultural and Livestock Research Organisation (KALRO) for proximate analysis.

Ethical considerations
For the study, the research sought the approval of Ethical Review Committee (ERC), JOOUST Postgraduate School and National Commission for Science, Technology and Innovation (NACOSTI). The study also considered ethical guidelines put in place for use of fish in experimental designs (Jenkins et al., 2014;NRC, 2011;Sloman et al., 2019). During the study period, the welfare of the experimental fish was maintained by ensuring that the fish were not stressed and were healthy.

Diet formulation
The H. illucens larva meal was obtained from the JOOUST insect farm, while other ingredients, including soybean meal, wheat bran, wheat pollard, fishmeal (FM) and vitamin premix, were purchased locally. The diets were formulated with ingredients such as H.
illucens larva meal (BSF), soya bean meal, maize, wheat pollard, wheat bran, sunflower meal, sunflower oil, fishmeal, salt, Dicalcium phosphate (DCP) and vitamin premixes (Table 1). After formulation, the diets were also analyzed to determine the nutritional value and varying levels of crude protein. The crude protein of formulated diets ranged between 29 and 30% (Table 2).

Experimental design
Mono sex male O. niloticus fingerlings of 15±0.06 gram were obtained from Siaya at Lake Agro Limited, formerly known as the Dominion Farm. The initial 14 days were used to acclimatize fingerlings to laboratory conditions. The study was conducted for 84 days at the JOOUST laboratory. The experiment employed a completely randomized design (CRD) with five treatments; FM (Control), BSF 2 (H. illucens larva meal 25%) BSF 3 , (H. illucens larva meal 50%) BSF 4 , (H. illucens fly larva meal 75%) and BSF 5 (H. illucens larva meal 100%). The treatments had a crude protein content of 29 to 30%. Each treatment was replicated three times. The research work was undertaken in 90 L rectangular glass aquaria of dimensions 70 × 35 × 40 cm each, each aquarium containing 15 fingerlings fed at 3% of their body weight daily. The daily feed ratio was divided equally amongst all treatments on each feeding schedule

Calculation of growth parameters
Fish were sampled every two weeks by siphoning water from the aquaria. Fish were collected from each aquarium using a plastic filter basket and weighed to the nearest 0.01g using a digital weighing balance. A fish measuring board was used to measure lengths (cm). The values (length and weight) were used to calculate the mean biweekly growth. The biweekly data was utilized to calculate the specific growth rate and mean weight gain. Daily mortality rates were recorded. At the conclusion of the experiment, weight and diet data were used to calculate growth performance measures such as final mean weight, mean weight gain, specific growth rate, feed conversion ratio, survival rate, and Economic feed conversion ratio (eFCR), as stated below.
Specific growth is defined as the variance in fish weight assessed as an increase in body weight as a percentage daily over a specified time period (Ricker, 1979). It was calculated using natural logarithms of body weight and given as a percent per day growth rate.

Feed conversion ratio (FCR)
For each treatment, the feed conversion ratio was calculated by dividing the amount of feed consumed by the weight gain. For each of the treatments, the FCR was established. The formula was as follows: FCR = Total weight of feed given (g) ÷ Weight gain (g) Final fish number Survival % = × 100 Initial fish number

Evaluation of physico-chemical parameters of water in aquaria
Water quality parameters, temperature (°C), oxygen (Mg -L ) and pH were measured three times daily. The mean daily temperature, oxygen and pH were calculated and recorded. Temperature and pH were measured using an electronic pH probe and dissolved oxygen was measured using an oxygen meter (Yellow Springs, YSI 550A U.S.A). NH 3 (Mg -l and NH 4 (Mg -l ) were measured bi-weekly using color disc and color comparator box after which calculations were done.

Statistical analysis
Data for all parameters were collected and recorded in a table form and stored in Microsoft Excel spreadsheet. The data was then imported into statistical software R (version 4.0.2) and/or Statgraphics Ver16 analysis. Descriptive statistics; mean and standard error (SE) were computed for each treatment. A one-way analysis of variance (ANOVA) was utilized to determine whether the diets substantially impacted survival, specific growth rate, and feed conversion ratio (FCR). Additionally, the least significant difference (LSD) test was used as a mean separations procedure. Further statistical analysis was done as follows.

Analysis of covariance (ANCOVA) for O. niloticus growth indicators
A one-way analysis of covariance was conducted to examine whether growth performance indicators (length or weight) were different based on the treatment and the control group, while controlling for time in days(x). Analysis of covariance (ANCOVA) improves the power of traditional analysis of variance (ANOVA) by reducing error. In general, since variations in days owing to temperature and influence from other factors that cannot be regulated in the experiment, their effects should be observed in tandem with growth performance indicators (total length or weight (y)). The model was specified as follows. Assuming a linear relationship between x and y the appropriate statistical model will be: is the growth performance as the response variable (weight or length) (the jth observation on the response taken under the ith treatment); i=1, 2…, a and j= 1, 2…, n; ( − . . )is the covariate or the variable used in the control of error and adjustment of the means; μ -is the initial or overall mean of the growth performance indicator (length or weight gain); -is the effect of the ith treatment (that is, the treatment levels); β is the linear regression coefficient indicating the dependency of on ; -is the random error component.
As shown in the equation, the analysis of covariance model is a combination of linear models used in the analysis of variance and regression. This means that, like in one-way classification analysis of variance, we have treatment effects t i and a regression coefficient β like in regression analysis. To keep the parameter μ as the overall mean, the covariate is expressed as ( − ..) instead of Xij.

Regression analysis for growth performance indicators
A regression on categorical variables analysis was used to predict total length and time in days based on the treatments. Coefficient of correlation (r) was used to measure the linear relationship between the length gains over time in days in consideration of the treatment effects.
The regression model is of the form: is the growth performance indicator (total length/weight); is the independent variables (time in days based on the treatments); βo is the intercept (the point where the regression line cuts the y-axis); ′ -the slope (the change in total length for every unit change time (Days)); -is the random error

Economic analysis
Enterprise budget analysis was performed to assess the financial impact of replacing FM with BSF in Nile tilapia diets. The cost of Total cost of feed fed (g) Economic feed conversion ratio (eFCR) = Total value of wet weight gain (g)  various feed materials was calculated, as well as the volume of feed (kg) used in each treatment (based on current market prices). The purpose of the enterprise budget was to look at the aquaculture business's relative profitability when O. niloticus consumed different levels of H. illucens. Feed costs, fingerlings costs, labour and transportation costs were all variable costs; whereas fixed costs were aquarium prices. The market price of O. niloticus in grams was used to assess the profitability of the treatments assumed to be different businesses. The analysis used the actual revenue values obtained after selling the fish. The profit index (PI) was calculated, and the economic feed conversion ratio (eFCR). The analysis used Kenya's local market retail prices expressed in Kenyan shillings (KES). Profitability was calculated using the gross profit margins of various treatments as shown in the equation below:

GP = TR -TC
Where; GP = Gross Profit, TR = Total Revenue (KES), TC = Total Cost (KES). Profitability was determined by the break-even price which was calculated by dividing total costs per unit price of fish (Shati et al., 2022;Musa et al., 2021).

Evaluation of physico-chemical parameters of water in aquaria
BSF 2 registered the highest temperatures with BSF 5 having the lowest. For both NH 3 and NH 4 , FM had the highest with BSF 5 having the least. Highest concentration of oxygen and pH levels was both recorded in aquaria with BSF 4 (Table 3). There were no significant differences in the physico-chemical parameters of water of any of the aquaria. Since there were no statistical differences on water quality between the treatment groups (p > 0.05), the differences in growth performance are due to the increasing levels of H. illucens in the O. niloticus diet.  (Table 4) . Furthermore, the same applies for the length gain and the daily length gain. The growth performance though is recorded highly in BSF 2 , BSF 3 and FM respectively, with BSF 4 having the lowest.
Like weight gain, the effect of BSF 4 specific growth rate varied significantly from BSF 2 and comparably to the other sets of treatments. Food conversion ratio on the contrary has the effects of BSF 4 significantly varying from FM, BSF 2 and BSF 3 except for BSF 5 . The different treatments do not have a significantly different effect on survival of O. niloticus. The survival rates of O. niloticus are the same regardless of the feed applied (Table 4).

Evaluation of O. niloticus growth performance fed on different levels of H. illucens over time based on length
There was a steady increase in fish length after feeding on the respective test diets through the 84-day culture period. The curves display some overlaps from the 30 th day onwards with more pronunciation happening from the 70 th day (Figure 1). BSF2 showcased high growth levels with BSF5 and BSF4 registering the lowest.

Analysis of Covariance based on O. niloticus length
The treatments have a significant effect on growth performance in terms of length (F (4, 95) =14.915; p< 0.05).
Days equally have a significant effect on growth performance (F (1, 95) =2335; p< 0.05). Moreover, the interaction effect between treatment and time is significant (F (4, 95) =7.595; p< 0.05) ( Table 5). While the ANOVA shows the differences in treatments means, the results of ANCOVA show performance of treatments when controlling for time in days.
Fish fed on BSF 2 , H. illucens at 25%, recorded the highest average length followed by FM (fish meal) with BSF 5 and BSF 4 , H. illucens larva at 100 and 75% respectively, having the lowest records of average length. Least significant difference (LSD) shows that the effect of BSF2 is significantly different from the effects of BSF3, BSF4 and BSF5 but comparable to the effects of FM (Table 6).     (Days)). The total log length of fish increased 0.005 daily (Table 7).

Regression analysis for O. niloticus Length
There was a significant difference for the slopes in the model with p<0.05 level with different dietary treatment levels. A test for common slopes was conducted for the five treatments. The analysis showed that the five lines presented do possess significantly different slopes (F 1,506 = 18.53, P < 0.001), indicating distinct growth patterns ( Figure 2).

Evaluation of O. niloticus growth performance fed on different levels of H. illucens over time in based on weight
O. niloticus growth performance in terms of weight followed a similar procedure as the total length analysis. There was a steady increase in fish weight after feeding on the respective test diets through the 84-day culture  period. The curves display some overlaps from the 20 th day onwards with more clarity happening from the 70 th day (Figure 3). BSF 2 showcased high growth levels with BSF 5 and BSF 4 registering the lowest.

Analysis of covariance based on treatment effect on O. niloticus weight
The treatments have a significant effect on growth performance in terms of weight (F (4, 95) =14.027; p< 0.05). Days equally have a significant effect on growth performance (F (1, 95) =3030; p< 0.05). Moreover, the interaction effect between treatment and time is significant (F (4, 95) =7.011; p< 0.05) ( Table 8). Fish fed on BSF 2 , H. illucens at 25%, recorded the highest average weight followed by FM (fish meal) with BSF 4 and BSF 5 , H. illucens larva at 75 and 100% respectively, having the lowest records of average weight. Least significant difference (LSD) shows that the effect of BSF 2 is significantly different from the effects of BSF 4 and BSF 5 but comparable to the effects of FM and BSF 3 (Table 9).

Multiple linear regression analysis for weight and days based on treatments
The analysis employed multiple linear regression analysis to predict total weight and time in days based on the treatments (Table 10). Coefficient of correlation (r) was used to measure the linear relationship between the weight gains over time in days in consideration of the treatment effects. There was a significant relationship between the variables F (5, 99) = 977.61, P<0.005 with an R 2 of 0.98. FM is embedded in the intercept as the constant and acts as the reference category. Both time and treatments have a significant effect on the growth of O. niloticus as shown by p = 0.000 for both CONSTANT and Time (Days). BSF 4 and BSF 5 vary significantly from FM as shown by the p = 0.000. On the other hand, effects of BSF2 and BSF3 are comparable to the effect of FM (p=0.5253 and p=0.1200 respectively) ( Table 10).
The output shows the results of fitting a linear regression model to describe the relationship between weight, days and treatment. The equation of the fitted model is: Where the terms similar to Treatment =BSF 2 are indicator variables which take the value 1 if true and 0 if false. This corresponds to 5 parallel lines. When treatment = FM, the model reduces to Log Weight = 1.19352 + 0.00671937*Time (Days); When treatment=BSF 2 , the model reduces to weight = Log Weight = 1.19352 + 0.00671937*Time (Days); because the P< 0.05, there is a statistically significant relationship between the variables. The total log weight of fish increases by 0.01 g daily (Table 10). The analysis showed that the five lines presented do possess significantly different slopes (P< 0.05), indicating distinct growth patterns (Figure 4).

Profitability and cost-benefit analysis
FM had the highest total cost of production, averaging (KES.3044.14), while BSF 5 had the least cost at (KES.3043.63). In terms of total yield, BSF 2 had the highest biomass at harvest of 2.69 kg. BSF 5 had the lowest yield, with a biomass of 2.34 kg (Table 9). BSF 5 had the lowest cost of producing a piece of O. niloticus, followed by BSF 4 , BSF 3 , BSF 2 and finally FM. The break-even prices for all diets were lower than the current market price of sh.350 kg -1 . The break-even price over total cost for BSF 5 was the lowest, while FM being the highest. The eFCR were compared for all the treatments. The analysis showed that the eFCR in FM (control) was higher than in all the 5 treatments (70.02). BSF 2 and BSF 3 recorded eFCR of 63.37 and 61.76 respectively while BSF 4 had an eFCR of 64.82. BSF 5 had the lowest eFCR of 59.39.

Evaluation of physico-chemical parameters of water in aquaria
In all treatments, there were limited variations in the physico-chemical water quality parameters measures ( Table 3). The lowest mean pH was recorded as 6.9546 in BSF 3 and the maximum being 7.2276 in BSF 4 . During the study period, the highest mean temperature recorded was 24.00°C, while the minimum mean temperature was 23.24°C. The lowest oxygen levels were found in BSF 3 at 4.3129 Mg -L , while the maximum levels were found in Odhiambo et al. 1061 BSF 4 at 4.9094 Mg -L . The highest quantities of ammonia were detected in BSF 4 at 0.1213 Mg -L with all the other treatments recording almost similar range. The water quality parameters taken were optimal for fish growth performance in all treatments. The survival of fish was not significantly different at the end of the experiment (P > 0.05) in all the treatments, indicating that water quality did not affect growth.
The current findings suggest that the diets used do not affect water quality. Furthermore, the fish excreta from this diet did not affect the water quality. Changes in temperature significantly influence the growth performance of larvae and fingerlings. Other water quality parameters were ideal for the growth of O. niloticus.
The water quality parameters taken were within the required range for fish growth performance in all treatments. The current findings suggest that the diets used do not affect water quality. Other water quality parameters were ideal for the growth of O. niloticus. The temperatures and pH physico-chemical parameters in the aquaria were within required limits for O. niloticus optimal growth (Rebouças et al., 2016). The results of water quality analysis were closely related to those reported by Tippayadara et al. (2021).

Analysis of covariance based on O. niloticus length and weight
This study's primary purpose was to examine the growth of O. niloticus fed on black soldier fly larvae meal and determine if growth varied predictably compared to FM. The change in length and weight throughout fish growth was tested using ANCOVA and showed a significant effect of treatment and time and an interaction between the two. After controlling for days, there was a significant difference in the weight and length of O. niloticus fed black soldier fly larvae meal at varying levels. The results of the growth study provide clear evidence that growth varies consistently with the level of inclusion of black soldier fly larvae meal. The ANCOVA showed a significant difference in weight and length of O. niloticus fed on different levels of black soldier fly larvae meal. The post hoc test determined that the length and weight of O. niloticus fed BSF 2 were considerably higher than O. niloticus fed with the control diet (FM). Compared to the control (FM), there were no significant differences in the weight and length of O. niloticus fed BSF 2 and BSF 3 . In addition, the low weight and length observed in fish fed diets at 75 and 100% could have been caused by low feed intake by the fish, due to unfavorable flavor of the diets. When insect meal is dried, a phenomenon called the maillard reaction occurs that either improves or ruins the taste of the feed affecting palatability (Tamanna and Mahmood, 2015). A study by Fasakin et al. (2003) on catfish fed with defatted and full-fat maggot meal suggested that defatting and drying processing methods assure insect meal palatability. Further, St-Hilaire et al. (2007) suggested that diets with increased levels of black soldier fly larvae meal have chitin and hence are less digestible. This in turn causes lower FCR, low weight and smaller-sized fish in turbot, pacific white shrimp, yellow catfish, Nile tilapia and rainbow trout. The amount of chitin in BSF-containing diets appears to have a deleterious influence on feed intake, growth, and digestibility (Kroeckel et al., 2012;Cummins et al., 2017;Xiao et al., 2018;Eggink et al., 2022). However, the FCR obtained in this investigation suggests that O. niloticus may have a high efficacy in digesting chitin or tolerating specific levels. Studies by El-Sayed (1998) and Cavalheiro et al. (2006) suggest that shrimp meal, which is heavy in chitin, can be used in place of fish meal without harming red tilapia (Oreochromis spp), Oreochromis niloticus, or Oreochromis urolepis hornorum. The quality of the crude protein has an impact on feed utilization as well. Increasing H. illucens larvae meal and reducing the percentage of fishmeal influence the quality of protein in the diet. The essential amino acids in the proper ratios determine the protein quality of a fish diet (Diógenes et al., 2016;Rodrigues et al., 2020). The black soldier fly larvae meal at 75 and 100% may have had a much lower amino acid content, particularly the essential amino acids. In addition, the high ammonia quantities in the BSF 4 treatment could have contributed to the lower weight and length gain registered in diet at 75% inclusion of H. illucens compared to all the other treatments

Regression analysis based on varying levels of H. illucens on O. niloticus weight and length
According to the findings, regardless of treatment, the weight of the fish increases with time as it consumes feed (P<0.05). All experimental fish fed on various diets during the investigation had significantly varied overall increases in body weight. For The model described 96% of the variables, as shown by the correlation coefficient of 0.96. The control (FM) is embedded in the intercept, where reference is made to the control feed compared to the other diets. The length and weight of O. niloticus fed BSF 4 , and BSF 5 differed significantly from the control diet (FM), P<0.05. O. niloticus fed on FM, BSF 2 , and BSF 3 gained more weight than BSF 4 and BSF 5 . Their linear regression slopes were less steep compared to the latter (Figure 4). The low weight increase in O. niloticus fed black soldier fly larvae meal at 75% and at 100% could have been caused by a decrease in intake of the feed by the fish due to the taste compared to FM. At 75% inclusion of fishmeal in the diet negligible while at 100% it is completely omitted in the feed. Nile tilapia is naturally attracted to fishmeal in their diets as it is more digestible, palatable and tastier to them (Jackson, 2006;Ryder and Balaban 2011). Kroeckel et al. (2012) suggest that lower palatability is frequently observed in diets when fishmeal is largely replaced by other protein sources, particularly if they contain anti-nutritional components (ANFs). The low weight and length in O. niloticus diets feed at 75 and 100% could have also been attributed to the low digestibility as seen in the low FCR realised by BSF 4 and BSF 5 compared to the other diets. Weight by O. niloticus fed BSF 2 and BSF 3 could have increased due to nutrient fortification. According to Ogunji et al. (2011), there appears to be a type of nutrient fortification when fishmeal is blended with maggot meal in the diet of carps up to a particular level. Other species, such as rainbow trout, Pacific white shrimp, and juvenile barramundi, have reported a lower percentage of 20 to 50% fishmeal replacement (Stamer et al., 2014;Cummins et al., 2017;Katya et al., 2017). Ogunji et al. (2011) discovered that maggot meals have more balanced amino acid profiles than numerous other potential protein sources. According to Kamarudin et al. (2021), fish fed 100% black soldier fly larvae meal had poorer growth performance. There were no indications of nutritional insufficiency or mortality in the current study. The regression model can be used by farmers to help predict and optimize the grow out period required for fish to reach the market on time.

Evaluation of O. niloticus growth performance
The body weight gains, length gains and SGR of O. niloticus slightly decreased when FM replaced H. illucens beyond 50%. Therefore, replacement of black soldier fly larvae meal by up to 50% is more profitable for farmers and contributes towards more yields. Enhanced performance could be attributed to improve essential amino acid profiles in diets below 50% (Kamarudin et al., 2021) increased feed intake by the fish.
All treatments had a significant difference in FCR, with BSF 4 and BSF 5 being significantly different from FM, BSF 2 and BSF 3. On the other hand, O. niloticus fed BSF 2 had the highest SGR (1.53±0.00) and the best FCR (0.71±0.01). The results of the FCR and SGR concur with those reported by (Kroeckel et al., 2012). A study conducted by Barroso et al. (2014) showed that, H. illucens larva meal had the most identical amino acid profile to fishmeal across a variety of insect species, which could explain the particularly positive results in the current experiment. The inclusion of chitin in the H. illucens larva meal could also explain aspects of performance improvement at 25 and 50% inclusion levels of H. illucens. Low quantities of chitin in H. illucens based diet boost the growth performance of fish and mitigate the negative effects of fishmeal replacement (Fontes et al., 2019).
The slower growth performance of O. niloticus when fed H. illucens diet at 75% and at 100% exhibits a positive relationship between decline in growth and increased quantities of H. illucens larva meal in the diet.
Increased levels of H. illucens in fish feed have been shown to inhibit fish growth due to taurine deficiency in H. illucens based diets, an amino acid essential for fish growth and health (Magalhães et al., 2017).This indicates that supplementing limited essential amino acids and adjusting the non-essential and essential amino acids ratio of substitute meals can lead to significant improvements in fish performance at increased amounts of H. illucens larva inclusion (Cummins et al., 2017). In addition growth could have been slower at 75 and 100% H. illucens diets due to increased levels of chitin in the diets. High chitin levels, which are generally indigestible to most fishes, have been linked to poor growth performance (Kroeckel et al., 2012) because they inhibit digestibility of lipids restrict the absorption of nutrients from the intestinal tract . The study's findings concur with those reported by Muin et al. (2017) that showed significant variations in growth when H. illucens completely substituted FM at 50%.The current study results differ from those of Tippayadara et al. (2021) and  who demonstrated that up to 100% replacement of fishmeal with H. illucens larva meal is possible with no negative impact on growth performance. Although there were no adverse effects on O. niloticus fed H. illucens meal up to 100%, optimal growth and revenue is obtained by substitution level of up to 50%. Using black soldier fly larvae meal by up to 50% is more economical as farmers have optimal growth and produce at reasonably lower cost compared to fishmeal. All of the dietary treatments in the current study had survival rates exceeding 95%. Fish fed on BSF4 had the highest survival (97.40%), while fish fed on BSF 2 had the lowest (95.7%), indicating that H. illucens meal had no detrimental consequences on the fish's ability to survive. Since most mortalities were noticed a day following length and weight measurements, the mortality that was found may have resulted from stress during sampling. The current study's findings concur with those of prior studies by Tippayadara et al. (2021) and Limbu et al. (2022)]

Profitability of using H. illucens larva in O. niloticus diets
As a farming enterprise, aquaculture typically requires the use of inputs such as feeds, fingerlings, water, labor, technical skills and land. The products can be used for either subsistence or commercial fishing. The availability of suitable and cost-effective fish feed, as well as increased economies of scale, are considered necessary for aquaculture success. Using low-cost alternative feed resources as protein source which is the major cost contributor will help lower production costs. As a result, finding and implementing alternatives to fishmeal will be critical to developing a more sustainable aquaculture. In addition farmers have to increase their economies of scale to maximize profits (Arru et al., 2019). All the treatments realized a negative gross profit margin. In the current study, the economic analysis for the initial production cycle, explains the negative marginal profits or losses in all treatments. The cost of aquarium paid for sh.3000 per aquarium has a significant impact on gross profit, increasing the cost of production. This rate was too high for the current study's production level, resulting in a negative gross profit. The results are consistent with those performed by Ouma (2019) in the first trial of Clarias gariepinus fed diets of plant protein origin with different crude protein levels. Due to the high capital required to purchase aquaria for the start-up, the initial production cycle was costly. In a multi-cycle economic analysis, the current setup would be profitable. Analysis on the cost effectiveness of the various dietary treatments (Table 11) in the present study shows that the cost of feeding was reduced when fishmeal was replaced with H. illucens larva meal. The highest profit index was found in the BSF 5 and BSF 4 diets, owing to the relatively low market prices of H. illucens larva and the total lack of fishmeal in the BSF 5 although higher FCR and reduced weight gain were observed. The control diet (FM) had the lowest profitability index. Generally, profitability index decreases with increasing levels of fishmeal inclusion because it is the most expensive component in the diet (Table 11). The results obtained in the current study confirms the general view held by other authors that fishmeal is the most expensive ingredient in formulated fish diet accounting for the high cost of aquaculture feed (Nguyen et al., 2013;Galkanda-Arachchige et al.,2020;Bilgüven, 2022) Increasing fish farm profits requires a minimization in the cost of feed and making considerable effort to finding alternatives of fishmeal from both plant and animal protein sources (Hossain et al., 2002). The study revealed that the use of H. illucens larva meal in the diet of O. niloticus reduces cost of production while enhancing growth performance.
Revenue findings demonstrate that partial replacement of fishmeal with alternative insects-based proteins is more profitable in aquaculture. As much as 100% replacement of H. illucens larva meal is more cost effective, BSF 5 in O. niloticus its yields are minimal returns. Therefore, BSF 2 , replacement at 25% and at BSF 3 50% are therefore the best treatment since they realised best O. niloticus growth performance and contributes optimal profits. Therefore, farmers can get more yields and income when replace fishmeal by up to 50% in O. niloticus diets The present finding on reduction in cost with increased levels of H .illucens in O. niloticus diet are consistent with that of Wachira et al. (2021) that observed significant cost reduction of use of increased levels of H. illucens in O. niloticus diets as a protein source.

Conclusion
The current study found significant differences in the growth performance of monosex O. niloticus fed on the five treatments (P<0.05). O. niloticus fed on BSF 2 25% and BSF 3 50%, inclusion of H. illucens larva diets had the best growth performance in terms of the parameters analyzed, compared to the control diet FM. However, no adverse effects were observed when fish was fed black soldier fly larvae meal at 100%. From the present results it could be concluded that BSF-meal can substitute fish meal without severe losses in body weight gain, FCR and up to an extent of 50%. Although the growth performance of O. niloticus fed the highest level of BSF-meal (BSF 75) was poorest, neither signs of nutrient deficiencies nor of higher mortalities were observed. From an economic view, the study recommends that fish farmers should produce fish in large scale and up to marketable size for sale to optimise productivity. The low profit margins in all the treatments were realised because of the low economies of scale and because the fish did not reach the market size. Therefore, if farmers increase their production levels more profits can be realised. In addition, this study recommends substitution of fishmeal with black soldier fly larvae meal up to 50% for optimal growth and optimal profits. However, farmers could still replace fishmeal by black soldier fly larvae meal by up to 100% without impacting growth performance of O. niloticus negatively although this lowers profitability. The study recommends further investigations on digestibility studies so the concrete conclusions are made on how chitin in black soldier fly larvae meal based diets affects growth performance of fish when fed with O. niloticus