Cymbopogon citratus leaves were obtained from plots around the faculty of Agriculture and Veterinary Medicine Teaching and Research Farm (FAVM-TRF) in the University of Buea, Cameroon. The fresh leaves were air dried at room temperature in a green house and milled to a fine powder using electric blender (BL-500). The powdered leaves were packaged in air tight glass jar then stored at room temperature prior to use.

Graded amounts of the lemongrass powder were incorporated into five iso-calorific and iso nitrogenous diets used for the experiment ( Table 1 ). The amounts mixed in 1000 g of a basal diet to constitute and treatment diets were: T1 (control) = 0 g, T2 = 2.0 g, T3 = 4.0, T4 = 6.0 g, mT5 = 0.8. The basal diet was (40% crude protein), comprising maize, soybeans cake, fishmeal, groundnut cake, fish oil, corn starch, DCP, methionine, lysine and vitamin premix formulated according to Fagbenro and Adebayo [11] . The ingredients were thoroughly mixed in with a wooden pestle to obtain a homogenous mass and corn starch added as a binder. The resulting mash was then passed through a pelleting machine with 4 mm die. The pellets were air-dried and packed in air tight containers. The diets were analyzed for proximate composition, including crude protein, crude lipid, crude fiber, ash and moisture as described by AOAC [12] .

Fifteen plastic tanks of 50-litre capacity, each three quarter filled with water, were aerated continuously using an air compressor. Two hundred and twenty five farm-raised C. gariepinus juveniles (21.25 ± 0.23 g) were acclimated to laboratory conditions for 14 days at the Fisheries Laboratory in Faculty of Agriculture and Veterinary Medicine Teaching and Research Farm (FAVM-TRF), University of Buea, before being distributed randomly into the 15 tanks which served as experimental units in a completely randomized design experiment. There were five treatments and each was replicated thrice at a stocking density of 15 fish per tank. The five treatments included a control and four test diets which had graded levels of Lemon grass meal powder. The treatments were labeled T1 (control = 0 g) T2 = 2.0 g, T3 = 4.0 g, T4 = 6.0 g and T5 = 8.0 g. Fish were fed at 5% of their body weight (bw) per day in two equal meals. A four stage procedure was used for randomization of the fingerlings to the various experimental units. First the 15 experimental units were placed properly on the tables with functional aerators and aerator stones. Then hand written labels indicating the treatment and the replicate number (which had been folded) were randomly drawn from a hat and assigned to each of the 15 units. The acclimatized fingerlings were pooled into one aerated tank, from which 5 fingerlings were scooped, weighed individually and placed in each experimental units. The scooping of 5 fingerlings was done 3 times to reduce scooping bias and eventually place 15 fingerlings per experimental unit. All fish were weighed and counted fortnightly and feeding rates were adjusted accordingly. The daily ration was divided into five equal parts and fed at specific periods during the day (9.00 am, 11.30 am, 2.00 pm, 4.30 pm and 7.00 pm). However, when an electrical failure lasting more than one hour just before feeding, that portion of the daily ration was carried over to the next feeding period and the amounts for the remaining feed period of the day readjusted accordingly. The experiment lasted for 60 days.

Sixty percent of all the water in the aquaria was siphoned daily every morning from 6.00 am to 8.30 am and replaced with fresh water previously stored for 24 hrs at room temperature in drums within the laboratory. Every Sunday all the water in the experimental tanks was removed, the tanks washed thoroughly and disinfested with virunet. The water source was the spring water meant to serve the university community for agricultural, domestic and research purposes. It was usually not chlorinated. However, the main catchment tank was washed once every Monday. Preliminary measurements of water quality before start of the experiment showed that nitrite was not detectable. The water quality parameters such as dissolved oxygen, temperature and pH were observed before the first feeding ration of the day. Dissolved oxygen and temperature was measured using DO meter (DO9100) in mg/l and ˚C units respectively while pH was measured using a pen type pH meter (pH-009 111).

The following growth and feed utilization indices were calculated: weight gain (WG), specific growth rate (SGR), food conversion ratio (FCR) and survival. The formulae used were as follows:

Weight gain (WG) = final average weight (g) − initial average weight (g);

Specific growth rate, SGR (% d – 1) = 100 × (ln Wt − ln W0)/t

where Wt and W0 represent final and initial body weights of fish, respectively, and t represents the duration of the feeding trial;

FCR = dry weight of feed (g)/wet weight gain by fish (g)

Survival (%) = number of surviving fish/initial number of fish stocked × 100

*Premix: composed (mg vitamin and mineral/kg premix): vitamin A 4,800,000 IU, vitamin D₃ 800,000 IU, vitamin E4800 mg, vitamin K 800 mg, thiamine 600 mg, riboflavin 2800 mg, vitamin B₃ 4800 mg, pyridoxine 600 mg, vitamin B₁₂ 4 mg, folic acid 200 mg, cobalt 160 mg, copper 1200 mg, iron 9000 mg, iodine 480 mg, magnesium 2730 mg, manganese 28,000 mg, zinc 20,000 mg.

At the end of feeding trial haematological studies was carried out on the fishes. Blood samples of about 2.0 ml were collected from three fish randomly selected from each replicate in all the treatments. Blood was collected from the caudal peduncle with heparinized plastic syringe, fitted with hypodermic needle and preserved in dipotassium salt of ethylene-diaminetetraacetic acid (EDTA) bottles for analysis. The Haemoglobin concentration (HB), Red Blood Cell (RBC), White Blood Cell (WBC) count, Mean Corpular volume (MCV) Mean Corpular Heamoglobin (MCH), Mean Corpular Heamoglobin Concentration (MCHC) Platelets (PLT) and Lymphocides (LYM), were determined in the haematology laboratory at MAFLEKUMEN Tiko, Cameroon using 5-part differential Haematology Auto-analyzer (Mindray BC 5300 model).

The cost of feed and fish were the economic criteria under consideration in this study and were based on the current market cost of feed ingredients and market value of a kilogram of fresh fish in Buea, Cameroon at the time of the experiment. The economic evaluations in terms of Cost of feed consumed (CFC), Total investment cost (TIC), Value of fish produced (VF), Incidence Cost (IC), Profit Index (PI) and Benefit Cost Ratio (BCR) were calculated as follows:

Cost of feed consumed (CFC) = Feed cost × Feed intake

Total investment cost (TIC) = Cost of feed + cost of juveniles stocked

Value of fish produced (VF) = Cost of fish × Weight gained

Incidence Cost (IC) or Economic conversion ratio (ECR) = FCR × Cost of Feed

Profit Index (PI) = VF/TCF

Benefit Cost Ratio (BCR) = VF/TIC

Data were entered into spreadsheets using Microsoft Excel and analyzed using SPSS version 22 software package. These were then used to estimate descriptive parameters like the means, the standard errors of the mean and statistical differences between group means, as well as one-way analysis of variance (ANOVA). Shapiro-Wilk test for normality and Levene’s test for homogeneity of variances were used to test whether the data meets the assumptions of ANOVA. When the result was significant, the Turkey’s post hoc test was used for multiple comparison. Significant levels were measured at 95% confidence threshold.

The water quality parameters investigated in this study are presented in Table 2 . The pH value ranged from 7.49 ± 0.41 in T3, to 7.66 ± 0.51 in T4. Temperature ranged from 26.2˚C ± 0.24˚C to 26.41˚C ± 0.41˚C in T5, while Dissolved oxygen (DO) ranged from 3.91 ± 0.47 mg/l in T5 to 4.23 ± 0.34 mg/L in T2.The highest Ammonia (NH₃) concentration (5.67 ppm) was recorded in T4 while the highest Nitrate concentration ( ${\text{NO}}_3^{-}$) was obtained in T5.

Key: DO = Dissolved oxygen, pH = hydrogen ions concentration.

The results of proximate analysis of the experimental diet is presented in Table 3 . Crude protein content was similar amongst treatments with the highest crude protein (C. P) (41.99%) in T1 and lowest value (40.06%) in T5. T5 (5.24%) registered the highest crude fiber while lowest was registered in T1 (4.24%). The highest moisture content was recorded in T4 (11.61%) while the lowest was obtained in T2 (10.40%). The Dry matter ranged from 88.39% in T4 to 89.60% in T1.

Key: CP = Crude protein, CF = Crude fiber, EE = Ether extract, DM = Dry matter, LGM = Lemon grass meal.

The results on growth performance of C. gariepinus fed with varying inclusion of lemon grass showed significant difference in the Mean final weight (MFW) with T5 recording the highest value (128.14 ± 4.04 g) while the lowest value was recorded in the control −T1 (87.34 ± 1.34 g) ( Table 4 ). The Mean weight gain (MWG), Relative growth rate (RGR) and Specific growth rate (SGR) increase with increase significantly with increased level of LGM across the treatments. However, feed conversion ratio (FCR) ranged from 1.13 ± 0.04 to 1.26 ± 0.06 and was not significantly different (P > 0.05) amongst the treatments. The Survival rate was very high (88.89 ± 3.85 in T1 to 95.56% ± 3.85% in T5) and was not significantly different (P > 0.05) amongst treatments.

Mean Values ± Standard Deviation with different superscripts a, b, c are significantly different (P ≤ 0.05) (superscripts a, b, c indicate difference significance); Key: IW = Initial weight, MIW = Mean initial weight, FW = final weight, MFW = Mean final weight, MWG = Mean weight gain, RGR = Relative growth rate, SGR = Specific growth rate, FI = Feed intake, WG = Weight gain, FCR = Feed conversion ratio.

The haematological indices of C. gariepinus juveniles fed different levels of Lemon grass (Cymbopogon citratus) Leaf meal in this study are shown in Table 5 . The recorded values of White blood cell (WBC) from group of fish fed with 0.0% C. citratus were significantly higher (P < 0.05) than the groups treated with lemon grass inclusion. The platelets (PLT) values in this study ranged from 17.33 ± 0.58 in T1 to 21 ± 1 × 10⁹ L⁻¹ in T4 with no significant difference in T1, T2, T3 and T5. The results recorded for Lymphocytes %, Red blood cell, haemoglobin, mean cell volume, mean cell haemoglobin and mean cell haemoglobin concentration were not significantly different among the various treatments.

The results of cost benefit analysis of culture of C. gariepinus juveniles using lemon grass-based diet are presented in Table 6 . The Lowest Incidence cost (IC) was registered in T5 (1.31 ± 0.04) while T4 registered highest value (1.51 ± 0.16). The Profit index (PI) ranged from 1.47 ± 0.18 in T2 to 2.90 ± 0.38 in T5 while Benefit cost ratio (BCR) ranged from 0.41 ± 0.06 in T1 to 1.27 ± 0.06 in T5.

Mean Values ± Standard Deviation with different superscripts a, b, c are significantly different (P ≤ 0.05); HB = Haemoglobin; RBC = Red blood cells; WBC = White Blood Cells; MCV = Mean cell volume, MCH = Mean corpuscular haemoglobin, MCHC = Mean corpuscular haemoglobin concentration PLT = Platelets, LYM = Lymphocytes.

Mean Values ± Standard Deviation with different superscripts a, b, are significantly different (P ≤ 0.05); Key: FC = Feed cost; CFC = Cost of feed consumed; TIC = Total investment cost; VF = Value of fish; IC = Incidence Cost; ECR =Economic conversion ratio; PI = Profit index; BCR = Benefit cost ratio. Cost of 1 juvenile = 200, FCFA; Value of 1 kg fish = 2500 FCFA; Cost of 1 Kg Maize = 200 FCFA, 1 Kg Soybean Meal (44.6%) = 400 FCFA; 1 Kg Groundnut cake (42%) = 390 FCFA 1 kg Lemon grass leaf Powder (11.2%) = 350 FCFA; 1 Fishmeal (62.6%) = 500 FCFA; 1 kg Fish oil = 500; 1 Kg Diacalcium Phosphate = 300; 1 Kg Corn starch = 300, I Kg Methionine = 1400 FCFA; ! Kg Lysine = 800 FCFA; I Kg Premix = 2000 FCFA; 1 USD = 500 FCFA.