Effect of stocking densities and feed supplements on the growth performance of tilapia (Oreochromis spp.) raised in ponds and in the paddy field

Sen Sorphea, Torbjorn Lundh*, T R Preston** and Khieu Borin

CelAgrid (Center for Livestock and Agriculture Development)
sorphea_sen@yahoo.com
* Department of Animal Nutrition and Management, Swedish University of Agricultural Science, Uppsala, Sweeden
** UTA - TOSOLY Finca Ecológica, AA # 48, Socorro, Santander, Santander del Sur, Colombia

Abstract

Two experiments were conducted at CelAgrid, Kandal Province for a period of 80days. In experiment 1, 12 ponds each with area of 10m² were used to compare four treatments arranged in a 2 x 2 factorial with a completely randomized design with 3 replicates. The first factor was fertilizer with effluent and no effluent; the second factor was stocking density of 3 or 5 fish per m². The fish in each pond were provided with feed at 5% of the fish biomass (DM basis). The feed contained 25% of water spinach, 25% of duck weed and 50% of rice bran (DM basis). The effluent from a bio-digester was applied every 7 days, at rates equivalent to 150 kg N/ha.

Survival rate was higher in ponds fertilized with effluent and in ponds with lower fish density. There were no differences for gain in weight and length due to treatment, and no interaction between fertilizer and stocking density.  However, these data were derived from random samples of fish taken at 20 day intervals and the results were partially confounded by differences among treatments in survival. Net fish yield was higher for the higher stocking density but there was no effect of fertilization with effluent. Feed conversion (DM offered/net fish yield) did not differ among treatments but this measurement was also confounded by differences in survival as amounts of feed offered were based on initial numbers of fish and the average weight estimated from the sampling at 20 day intervals. Final fish weight and net fish yield were negatively correlated with survival rate.

In experiment 2, 12 plots in a paddy field each with an area of 209m² were used to compare 4 treatments in a 2 x 2 factorial in a completely randomized design with 3 replicates. The first factor was with or without feed supplement; the second factor was different stocking densities of 3 or 5 fish per m². In each plot of paddy there was a trench 11m wide x 1m in length x 1m deep along one side of each plot. The feed supplement was the same as in experiment 1. All paddy plots were fertilized with effluent from a bio-digester every 7 days at the rate of 150 kg N per ha.

Survival rate was not affected by supplementation but there was a tendency (P = 0.10) for it to be lower on the higher stocking rate. Both final weight of fish and the net fish yield were increased by supplementation and by stocking rate with no interaction between the treatments. The FCR (for those paddies that received feed supplementation) was not affected by stocking rate.

In conclusion it would seem that in rice-fish systems, supplementation is not an appropriate intervention, in view of the lower efficiency of use of the supplement. Thus, for the additional 43 kg of net fish yield (123-80) in experiment 2,  the amount of feed provided was on average  358 kg (7.5/209*10000), that is about 7.5 kg feed per 1 kg of net fish yield.   Measures that lead to enhancement of the natural feed supply (e g: fertilization with bio-digester effluent) would seem to be more appropriate technology.

Key words: Effluent, feed conversion, rice-fish culture, supplementation

Introduction

Cambodian people prefer their protein to come from fresh water fish that is eaten fresh, salted, smoked or made into fish sauce and paste. Tonle Sap, Tonle Mekong and Basak rivers are the main capture fisheries in Cambodia (Agriculture in Cambodia 2010).

Tilapia has become popular for farmers as it is easy to culture and there is a good demand in the market. Moreover, tilapias (Oreochromis spp.) adapt well to the local environment and local feed, and have high productivity. The fish are usually kept in a pond near to their houses, as in addition to having the fish as protein source, farmers can grow vegetables and use the water from the pond to water the vegetables. The feeds used for the fish depend on the resources available in the area. Duckweed and water spinach are available almost everywhere in the villages; while rice bran is the byproduct from the rice milling.

The combination of rice and fish can be a very profitable system, since it was observed that the fish feed on organisms, such as insects and larvae which grow and live in the rice fields. This system provides both rice and fish. Besides economic benefits, the biological benefit is also a factor. Weeding and use of chemical fertilizers and pesticides are reduced when this system is practiced.  Moreover, the movement of the fish stirs the water, which increases the oxygen level and improves the development of the roots of the rice. Rice-fish culture improves the income of farmers in the rural areas, as the system requires very little inputs and farm labor. Farmers can harvest rice or fish at the same time or harvest only the rice and keep the fish, or alternatively harvest the fish before the rice (Mackay 1995). 

Hypothesis

• In pond culture the net fish yield of tilapia will be increased by fertilizing with bio-digester effluent and by increasing the stocking density.
• When tilapia is raised in the rice paddy there will be advantages from supplementation with duckweed, water spinach and rice bran.

Objectives

Two experiments were carried out to study the effects of stocking density and fertilization/supplementation on the growth performance of Tilapia raised in ponds and in a paddy field.

Location and climate

The experiments were carried out at the Center for Livestock and Agriculture Development (CelAgrid) experimental farm, located in Prah Theat village, Rolous commune, Kandal Steung district, Kandal Province, approximately 19 km from Phnom Penh City.

In Cambodia, the rainy season is from June to October, while the dry season is from November to May. This climate provides good conditions for the animal and rice production system. Average temperature is around 25^oC, with a maximum of about 40^oC in April, while the coldest month is January, when the temperature is around 21^oC, with a maximum of about 31^oC. Average annual rainfall in Cambodia varies from 1,500 mm or less in the central plain and 1,500 to 2,500 mm in the surrounding mountains. Over most of the South West coastal region, average annual rainfall is in excess of 3,000 mm. The rainfall in the East of the Mekong River is generally between 1,800 mm and 3,000 mm, while the lower Mekong valley and basin of the Tonle Sap Lake is relatively dry, with rainfall averaging between 1,200 and 1,500 mm.

Experiment 1. Effect of stocking density and fertilization with bio-digester effluent on the growth performance of tilapia (Oreochromis spp.) fed rice bran, water spinach and duckweed in pond culture

Materials and methods

Study duration

The experiment was conducted from 12 January to 1 April 2010.

Pond preparation and management

A total of 12 ponds were prepared, each with an area of 10 m² (4 m in length x 2.5 m wide) and a depth of 1.5 m at the CelAgrid centre. The ponds were lined with plastic to avoid filtration of water, and then water was pumped in from nearby canals and pond. Lime (CaO) at 200 g/m² was applied before stocking the fish, in order to kill parasites and pathogenic organisms and also to increase water pH.

Experimental design

The experiment was conducted 80days and designed as 2 x 2 factorial arrangements: the factors were with or without application of biodigester effluent, and two stocking densities (3 and 5 fish per m²) (Table 1). A  Complete Randomized Design (CRD) was used. Each treatment was replicated 3 times. In total 640 fingerlings were bought from a commercial fish farm and randomly distributed into the ponds. Weight and length of a sample of the fish were recorded as the initial weight and length.

Experimental diets

The fish were fed a mixture of rice bran, water spinach and duckweed as the basal feed. All ponds were provided with the same feed (Table 2).

Duckweed was cultivated in CelAgrid, while water spinach was bought from a local market. In order to reduce the  moisture content, duckweed was collected in the evening and wilted for the morning feeding, while the duckweed collected in the morning was wilted and fed to the fish in the evening. Water spinach was chopped into small pieces and mixed with duckweed and rice bran before feeding. The feed was provided twice daily at 8:00h and 16:00h. The amount of feed was 5% (DM basis) of the fish body weight. The amount of feed was adjusted by the average of fish sampling multiply with the initial number of the fish at the beginning. The feed was mixed and put in a floating feeding frame to avoid the feed spreading in the pond (Photo 1).

Fertilization

Ponds were fertilized with effluent from a plastic bio-digester loaded with pig manure, at a rate of 150 kg of N/ha/year (San Thy et al 2006). The effluent from the bio-digester was pumped into containers (Photo 2). Before applying it to the fish pond, a sample was taken to determine N. The amount of effluent applied to the ponds was calculated according to the concentration of N.

Photo 1: Individual pond with floating feeding frame	Photo 2: Effluent storage

Measurements

The oxygen and pH of the water in the fish pond were measured every 5 days.  Each measurement was taken 2 times at 06:00h and 16:00h using a pH meter (pHep by HANNA) and a DO₂ meter (Model 9150), respectively. The water temperature was measured 3 times every 5 days at 06:00h, at 12:00h and at 16:00h using thermometers. A thermometer was placed permanently in each pond. Water transparency was measured at 12.00h every 2 days using a Secchi disk.

Every 20 days a sample of the fish was caught with a seine net and ten individuals chosen at random. These were weighed using an electronic scale and measured with a ruler from the mouth tip to the caudal fin. Survival rate was measured at the end of the experiment by the following equation:

X (%) = (N_t / N₀) x 100

Where: N₀ : initial number of the fish;  N_t : final number of the fish

Statistical analysis

The data were subjected to analysis of variance (ANOVA) by using the General Linear Model (GLM) of the Minitab software (version 2000 release 13.1). Sources of variation were: effluent, density, effluent * density interaction and error. Gain in weight and length were measured as the linear regression of weight (or length) on days in the experiment, using the SLOPE command in the Minitab software.

Results

Chemical composition of the feed

During periods of 20 days, water spinach and duckweed were harvested at the same place to make sure that the nutrient content was not so different. The protein content of the duckweed was higher than for the water spinach (Table 3). 

Nitrogen requirement of the pond

Effluent from the biodigester was analyzed for N, and the amount calculated based on the rate of N application of 2.88 g N per pond per week (Table 4).

Water quality

The treatments had no effect on water quality, measured by pH, temperature or dissolved oxygen levels (Table 5), all of which were within the normal range for culture of Tilapia (Swingle 1969). There was an interaction between treatments for water density (Table 6), which was more transparent (less phytoplankton) with the lower fish stocking rate when effluent was applied. In contrast, in the absence of effluent, the water was more transparent at the higher stocking rate.

Changes in weight and length, and in the ratio of weight: length

There were no differences for gain in weight and length due to treatment (Table 7), and no interaction between fertilizer and stocking density.  However, these data were derived from random samples of fish taken at 20 day intervals and, as will be discussed later, the results were partially confounded by differences among treatments in mortality/survival.

The best measure of the overall effect of the treatments is the net yield of fish expressed as the weight of all the fish at the beginning of the experiment subtracted from the final weight of the survivors (Tables 8 and 9). On this basis there was no effect of the effluent but a significantly higher yield for the greater stocking density. However, as feed offered was based on average live weights in each pond multiplied by the numbers of fish at the beginning, those with the highest density received more feed. Also in the ponds with highest mortality, the survivors received more feed per fish than in the ponds with higher survival rate.

Net yield of fish and survival rate

Survival rate was higher in ponds fertilized with effluent and in ponds with lower fish density (Tables 8 and 9). This in turn affected the net fish yield which is measured as the weight of surviving fish less the weight of fish at the beginning.

Discussion

Growth rate

Growth rates in weight and in length were calculated from measurements on a random sample of 10 fish taken at 20 day intervals. The limitation of growth rate are affected by the density of fish in the pond which was determined not only by the experimental treatments (3 or 5 fish/m²), but also by the survival rate of the fish population. This varied among treatments thus in some cases the feed available per fish was influenced by the numbers of fish surviving in the pond. When the survival rate was low the feed offer per surviving fish was increased, which would stimulate faster growth rate. The problem is that survival rates could only be measured at the end of the experiment, and it is not known at what stage of the experiment the fish died.

For this reason productivity was measured as the net fish yield expressed as weight of fish at the end of the experiment less the weight at the beginning. Using this criterion it was clear that the survival rate had a determining effect on fish productivity. As the survival rate increased the final weight per fish and the net fish yield both decreased (Figures 1 and 2). In contrast, the DM offered per pond was determined by the stocking density at the start of the experiment. The relation to survival is incidental and not causal (Figure 3). The final result was that the feed conversion, measured as DM offered/net fish yield, was not affected by the survival rate (Figure 4).

The net fish yield in this experiment (a range of 1500 to 2200 kg/ha) was twice as high as the yields (760 – 1200 kg/ha) reported by San Thy and Preston (2003). These researchers also used Tilapia but in ponds of 6m² and at a lower stocking rate of 2 fish/m². They applied biodigester effluent to the ponds at a similar rate (160 kg N/ha) but the fish received no supplementary feed. Nguyen Duy Quynh Tram et al (2007) fertilized ponds with raw pig manure or bio-digester effluent (derived from the same manure) at 240 kg N/ha over 120 days. The net fish yield of a mixture of Tilapia, Silver carp and Hybrid Catfish was 1700 kg/ha with the effluent and 2100 kg/ha with the raw manure.

Conclusions

• The net yield of fish was not affected by addition of effluent to the ponds but was higher for the greater stocking density.
• Survival rate was higher in ponds fertilized with effluent and in ponds with lower fish density.
• As the survival rate increased the final weight per fish and the net fish yield both decreased.

Experiment 2. Effect of stocking density and feed supplementation on the growth performance of tilapia (Oreochromis spp.) raised in the paddy field

Materials and methods

Study duration

The experiment was conducted in the dry season, from 23^rd January to 13^th April 2010.

Pond preparation and management

Twelve plots were prepared in a paddy field. The total area of each plot was 11m x 19 m with a trench 11m wide x 1m in length x 1m deep at one side of each plot. In total 528 fingerlings were purchased from a fish hatchery farm, Preak Phnov, near Phnom Penh City. They were raised in a nursery pond in CelAgrid for 15 days before being introduced into the plots of paddy rice, which was done 7 days after rice transplanting.

Experimental design

The experiment was carried out as a 2 x 2 factorial arrangement. The factors were: with and without supplementary feed; and stocking densities of 3 and 5 fish per m² (Table 10). A Complete Randomized Design (CRD) was used. Each treatment was replicated 3 times.

Feed processing

Fish were fed on a mixture of rice bran, water spinach and duckweed (Photo 6, 7 and 8; Table 11).

The feed was provided twice daily at 8:00h and 16:00h at an estimated rate of 5% (DM basis) of the fish body weight. The amount of feed was adjusted by the average of fish sampling multiply with the initial number of the fish at the beginning. Duckweed was cultivated in CelAgrid, while water spinach was bought from the market. In order to reduce the moisture content, duckweed was collected in the evening and wilted for the morning feeding, while duckweed collected in the morning was wilted and fed to the fish in the evening. Water spinach was chopped into small pieces and mixed with rice bran and duckweed before feeding to the fish, in the feeding frame (Photo 1). During successive periods of 20 days, water spinach and duckweed were harvested at the same place to make sure that the nutrient contained in water spinach and duckweed were not so much different.

Measurements

Fish weight and length were measured on random samples (n=10) of fish taken at 08:00h every 20 days before they were given feed and before application of effluent.  Dry matter and CP of feeds were analyzed every 20 days before the fish sampling.  Survival rate was measured at the end of the experiment. Water quality was measured following the same procedures as in Experiment 1. The fish were harvested after 80 days. Total weight, length and the number of the fish were measured.

Statistical analysis

The data were subjected to analysis of variance (ANOVA) by using the General Linear Model (GLM) of the Minitab software (version 2000 release 13.1). Sources of variation were: feed, density, feed * density interaction and error.

Results

Chemical composition of the feed and fertilization

The crude protein content in duckweed was higher than in water spinach (Table 12).

All the ponds in the paddy field were fertilized with effluent from a bio-digester at a rate of 150 kg N per hectare/year (San Thy et al 2006).  The amount of effluent was calculated on the basis of its content of N (Table 13). It was supplied directly into each paddy at intervals of 7 days.

Water quality

There were no treatment effects on water quality (Table 14).

Gain in weight and in the ratio weight: gain

Growth rate in weight was increased by supplementation (Table 15), with no effect due to stocking rate. However these results were confounded by differences in survival rate (Table 16).

Survival and net fish yield

Survival rate was not affected by supplementation but there was a tendency (P = 0.10) for it to be lower on the higher stocking rate (Table 15). Both final weight of fish and the net fish yield were increased by supplementation and by stocking rate with no interaction between the treatments. The FCR (for those paddies that received feed supplementation) was not affected by stocking rate.

Discussion

The average growth rates in experiment 1 for the supplemented fish were higher (0.70 g/day) and feed conversion better (1.61) than in the present study. It is possible that some of the feed supplement was washed into the rice field which would make it more difficult for the fish to consume it. The two systems are not comparable as the net outputs are quite different. In experiment 1 the only output was the fish whereas in experiment 2 the outputs were fish and rice. The comparisons in Table 16 are more relevant, although here also the conditions varied in terms of quantity of supplement supplied and fish densities. In general, the net fish yield in our study can be said to be within the range reported in Table 16.

As a general conclusion it would seem that in rice-fish systems, supplementation is not an appropriate intervention, in view of the lower efficiency in use of the supplement. Thus, for the additional 43 kg of net fish yield (123-80) , the amount of feed provided was on average  358 kg (7.5/209*10000), that is about 7.5 kg feed per 1 kg of net fish yield.   Measures that lead to enhancement of the natural feed supply (e g: fertilization with bio-digester effluent) would seem to be more appropriate.  

Conclusions

• Net fish yield was increased by supplementation and by the higher stocking density,

• The survival rate tended to be reduced (P=0.10) by the increased stocking density but was not affected by supplementation

• As the survival rate increased the final weight per fish and the net fish yield both decreased

• Measures that lead to enhancement of the natural feed supply (e g: fertilization with bio-digester effluent) would seem to be more appropriate than supplementation.  

Acknowledgements

The authors are grateful to the MEKARN program, financed by Sida (Swedish International Development Agency) for supporting this study.

References

Agriculture in Cambodia 2010   http://en.wikipedia.org/wiki/Agriculture_in_Cambodia#Rice_production

AOAC 1990  Official Methods of Analysis. Association of Official Analytical Chemists. 15th edition (K Helrick, editor). Arlington pp 1230

Bocek A   Water harvestinng and aquaculture for rural development. http://www.ag.auburn.edu/fish/international/introrice.htm

Mackay K T  1995  Rice-Fish culture in China. http://www.idrc.ca/en/ev-9299-201-1-DO_TOPIC.html

Nguyen Duy Quynh Tram, Le Duc Ngoan and Ogle B 2007  Effect of processing pig manure through a biodigester. http://www.mekarn.org/msc2003-05/theses05/tram_p1.pdf
 

Rasowo J and Auma E O  2006  On farm trials with rice-fish cultivation in the west Kano rice irrigation Scheme, Kenya http://myais.fsktm.um.edu.my/2514/

Rothuis A 1998  The effect of rice seeding rate on rice and fish production in direct-seeding rice-fish culture. Page 43.

San Thy, Khieu Borin, Try Vanvuth, Pheng Buntha and Preston T R 2008  Effect of water spinach and duckweed on fish growth performance in poly-culture ponds. Livestock Research for Rural Development. Volume 20, Article #16. http://www.lrrd.org/lrrd20/1/sant20016.htm

Swingle H S  1969  Methods of analysis for waters, organic matter and pond bottom soils used in fisheries research. Auburn University, Auburn, AL.

Received 10 November 2010; Accepted 24 November 2010; Published 9 December 2010

Go to top