Nutrition

Effect Of Varying Levels Of Sulfate Concentration In Saline Waters On Fish Yield

R.D. Fortes, N.R. Fortes, And I.G. Pahila

Institute of Aquaculture,

University of the Philippines in the Visayas

Iloilo, Philippines

Abstract

Milkfish (Chanos chanos) were stocked in 12 90-lifer aquaria to determine the effect of sulfate on fish yield. Saline waters from various sources with known sulfate levels, and different food sources were used as treatments in first and second runs, respectively. The levels of sulfate (first run) decreased from 1,200, 1,380, 1,700, and 1,100 mg/l to 976.53, 840.53, 772.56, and 441.07 mg/l, respectively, after 24 days; and to 477.00, 490.79, 339.78, and 325.05 mg/l after 28 days. The concentrations of sulfide and ammonia increased (0.01140.0286 mg/l and 0.0105-0.1820 mg/l, respectively). Populations of microorganisms adhering to rice bran particles were highest (450, 570, 850/ml) where sulfate and sulfide concentrations (325.05 mg/l and 0.0192 mg/l) were lowest and ammonia concentration (0.08006 mg/l) was highest.

The levels of sulfate and sulfide were not different among the treatments (range of 919.5-970.5 mg sulfate/l, and 0.02252-0.02542 mg sulfide/l) (second run). The levels of ammonia, however, were highest in the rice bran-sugar (II) and commercial-feed (IV) treatments compared to treatments that received fertilizers (I and III). The populations of microorganisms that adhere to substrates (tiles) were fewer in the rice bran-sugar and commercial-feed treatments than in the treatments that received fertilizers (I and III).

In both runs, the effect of sulfate concentration on fish yield could not yet be established due to low survival (40-67 %), poor growth attributable to poor water quality, low feeding rate, and poor food quality (first run). In the second run, fish survival was very high (100%) except one replicate each in Treatments II (88%) and III (93%). The growth of milkfish in all treatments was low (4-6 % per day), apparently caused by foul water and acidity, which could have resulted from high sulfate concentration in the water. An in-depth study using stable isotope technology is in progress to identify important feeding niches and to intensify the productivity of these niches. Success in this study could lead to a new pond management strategy.

Introduction

This paper is an initial attempt to demonstrate the effect of sulfate concentration in saline water on fish yield with the goal of evaluating the role of sulfate-sulfide reaction and the role of sulfide as a toxin that limits the availability of natural foods in brackish and saltwater earthen ponds. The role of sulfate as a cause of mineral acidity in ponds through its reaction in water has been the object of investigations since the 1960s. There has been considerable research to help alleviate the negative influence of acid sulfate soils on fish and shrimp yields (Singh 1980); however, the problem of sulfide accumulation in the sediment interstices still exists. The effect of this accumulation on crude organic matter, which is the main source of nutrition for the target animals, needs to be properly understood. In milkfish (Chanos chanos) and prawn (Penaeus monodon) farming, traditional or extensive, modified extensive, and semi-intensive methods based on animal density, added food, and other inputs are in use. These provide the necessary organic matter that could be converted into nutritious food for the fish. However, the users have been unable to differentiate these methods (Fortes et al. 1989, Fortes 1991). Tilapia (Oreochromis niloticus) raised in brackish-water ponds that received feed, either alone or in combination with chicken manure and/or fertilizer, exhibited better growth and higher production (Fortes et al. 1986), but the actual sources of growth are still uncertain. In fresh water, despite the presence of full rations of protein-enriched feed pellets, natural food still accounted for more than half of the growth of the target fish (Schroeder 1983).

The goal of this study is to determine food niches in brackish-water or marine ponds that provide target animals with their nutrition. Armed with this information, a management strategy can be designed that will improve the use of fertilizers as locally available replacements for costly imported feeds. This particular component of the project was implemented (initially using milkfish as the test organism) to pursue the following objectives:

- To determine the effect of sulfate on fish yield;

- To determine the influence of sulfate and sulfide ions on the production of natural food; and

- To determine the influence of sulfate on the population of microorganisms that adhere to the added organic or inorganic material in saline water.

Materials and Methods

Location and Facilities The site is located in Barangay Nabitasan, municipality of Leganes, Iloilo Province, Philippines. The municipality of Leganes is located N 10° 8' longitude; E 122° 5.4' latitude. The laboratory facilities included 12 glass aquaria (H= 35 cm; L= 90 cm; W= 35 cm) provided with an airlift system and a bottom filter made of 10 cm sand on perforated marine plywood (0.635 cm thick). Aeration, water delivery, and lighting systems were also installed.

First Run

Collection and Preparation of Water Water used during the first run of the experiment was collected from three points (Guimaras, SM area, and Gui-gui Creek) along Guimaras Strait approximately 7 km from the laboratory. This was necessary to ensure that pure seawater was collected. Seventy-five 60-1 plastic bags filled with seawater to a 40-1 line (total of three tons of seawater) were transported to the laboratory by means of a 4-ton motor boat. Water from each source was used separately or in combination with water from other sources, including freshwater, and of the four, each treatment was replicated three times, as follows:

Treatment	Initial Sulfate Content (ppm)	Source of Water
I	1,200	Guimaras
II	1,380	Gui-gui Creek
III	1,700	SM + Gui-gui (1:1)
IV	1,100	SM + freshwater (1:1)

 

Samples of water were also sent to the service laboratory of the Natural Science Research Institute of the University of the Philippines at Diliman, Quezon City, for the analysis for cations.

Monoammonium phosphate fertilizer was added before fish stocking to permit microbial growth. Concentrations of sulfate, sulfide, and ammonia in the water were measured before and after stocking and every week thereafter using the methods described by Strickland and Parsons (1972). Salinity, pH, temperature, and dissolved oxygen were monitored every other day using an Atago refractometer, digital pH tester, and a YSI D.O. meter (model 51B), respectively.

Fish Stocking and Feeding Sixteen milkfish fingerlings (average weights: 1.27 g, 1.26 g, 1.23 g, and 1.45 g, for Treatments I, II, III, and IV, respectively) were stocked in each of the 12 units of 90-1 aquaria (equivalent to 1 fish/6 1). The fish were fed rice bran mixed with 1% refined cane sugar given at 5% of fish biomass daily and adjusted to 10% after the first sampling. The fish were raised in these aquaria for 28 days and sampled at the midpoint and at the end.

Sulfate, Ammonia, and Microbe Populations Finely ground rice bran was mixed with refined sugar at a ratio of 100:1 (rice bran: refined sugar). Two grams of the mixture were added into a 2-1 plastic jar filled with seawater collected from different sources with known levels of sulfate concentration. The mixtures were analyzed for nitrogen after 1, 2, 12, and 24 hours; and after 7, 14, and 30 days of contact with water. Nitrogen content of the water was determined by the use of a micro-kjeldahl apparatus, which converts the nitrogen from small samples into ammonia, which is then measured volumetrically. The changes in the nitrogen content in the feed substrate during each time of exposure was taken to represent the microbial processing that could take place if rice bran were not immediately consumed by the fish and remained in the water or sediments to serve as food substrate. The water was observed for visual changes, especially the occurrence of detritus in the container. Water samples and detrital material or organic residues were taken and examined under the microscope. Organisms were counted using a hemacytometer.

Second Run

Due to the difficulty of maintaining the desired level of sulfate concentration in water, the treatments were changed. Instead of using the levels of sulfate concentrations as treatments, different sources of fish food were used as treatments, then the levels of sulfate concentrations in water were monitored. This new experimental design is consistent with the tank experiments of our collaborators in Israel and a tilapia experiment in aquaria. The new treatments with the initial weights of the fish are as follows:

	Treatment	Initial Weight (g)
I	Rice bran + cane sugar + fertilizer	1.05
II	Rice bran + cane sugar	1.17
III	Natural food + fertilizer	1.09
IV	Commercial pelleted feed	1.08

 

Each treatment had three replicates (aquaria) that were stocked with 15 milkfish fingerlings each. The rate of feeding for the three different food sources was 10% of the fish biomass given every day. Natural food in the form of a benthic mat composed of lower forms of microflora and microfauna (lab-lab) and filamentous algae were added to the aquaria. The natural food was given at the rate of 20% of fish biomass and was later increased to 40%.

A total of 16 tiles (equivalent to 16 samplings) were scattered over the bottom of each aquarium. One tile was then randomly picked every sampling and samples of the microorganisms were taken by scraping off the brownish-whitish substance adhering to the tiles. The scrapings were then weighed. The wet weight of the organisms was estimated using the following equation:

OW = WTBS—WTAS

where:

OW - weight of organisms (g)

WTBS - weight (g) of tiles before scraping

WTAS - weight (g) of tiles after scraping

The samples were then fixed in formalin and the organisms were enumerated and identified. Population counts of minute organisms were made using a hemacytometer; for larger ones, the Sedgewick rafter counting chamber and cell were used.

Results and Discussion

First Run

The initial sulfate concentrations of seawater from the different sources were 1,200 mg/l, 1,380 mg/l, 1,700 mg/l, and 1,100 mg/l for Treatments I, II, III, and IV, respectively. These were

TABLE 1 Sulfate Concentration of the Different Treatment used in Aquarium Experiment No. l and No. 2

Treatment	Sulfate (ppm)	Salinity (ppt)
I	976.98	36
II	840.52	35
III	772.56	36
IV	441.07	19

 

significantly lower (P < 0.05) than the theoretical concentration of seawater of 2,657 mg/l (Church 1975). These initial concentrations decreased to 976.98 mg/l, 840.53 mg/l, 772.56 mg/l, and 441.07 mg/l after 24 days of storage, and were the initial sulfate concentrations of the various treatments at the start of the experiment (Table 1). They decreased further to 477.00 mg/l, 490.79 mg/l, 339.778 mg/l, and 325.05 mg/l, respectively, 11 days after the fish were stocked. However, an increase was observed a few days before the experiment was terminated.

The values of different water parameters that were monitored (sulfate, sulfide, ammonia, pH, DO, temperature, and salinity) are discussed as follows and are given in Table 2.

TABLE 2 Physicochemical Properties of Water and Sediment (Milkfish Aquarium Experiment -- First Run)

	Sulfate (mg/l)			Sulfide (ml/l)
A. Treatment:	Dec. 27	Jan. 7	Jan. 10	Jan. 17	Jan. 7	Jan. 10	Jan. 17	Jan. 24
I	976.98	476.98	770.72	824.13	0.019	0.025	0.024	0.025
II	840.52	490.79	772.56	813.08	0.015	0.027	0.025	0.020
III	772.56	339.78	726.52	811.23	0.022	0.025	0.024	0.023
IV	441.07	325.05	536.83	669.43	0.013	0.024	0.018	0.021
	Ammonia (mg/l)			Dissolved Oxygen (mg/l)
B. Treatment:	Dec. 27	Jan. 7	Jan. 10	Jan. 17	Jan. 24	Jan. 7	Jan. 10	Jan. 16	Jan. 24
I	0.076	0.140	0.021	0.029	0.031	3.63	2.15	4.73	4.73
II	0.097	0.130	0.069	0.065	0.038	3.47	2.05	4.77	5.13
III	0.011	0.147	0.041	0.033	0.046	3.58	2.07	4.30	4.47
IV	0.162	0.140	0.029	0.043	0.026	3.52	2.47	4.73	5.07
		Salinity (ppt.)				pH
C. Treatment:	Jan. 7	Jan. 10	Jan. 16	Jan. 24	Jan. 7	Jan. 10	Jan. 16	Jan. 24
I	36	35	39.3	37.7	7.20	7.37	7.17	7.43
II	35.7	35	40	37.3	7.03	7.17	7.07	7.33
III	36.3	34.7	39.3	38	7.27	7.23	7.20	7.50
IV	19.3	17.7	21.3	19.7	7.43	7.53	7.17	7.47
	Temperature (C°)		Organic Matter (%)		Soil Sulfate (mg/l)
D. Treatment:	Jan. 7	Jan. 10	Jan. 16	Jan. 24	Dec. 27	Jan. 28	Dec.27	Jan.28
I	24.1	25.6	25.6	27.0	0.0	0.81	0.0	3652.5
II	24.1	25.6	25.7	27.0	0.0	0.97	0.0	3160.2
III	24.2	25.8	25.7	27.0	0.0	0.85	0.0	3983.7
IV	24.1	25.6	25.6	27.0	0.0	0.85	0.0	3337.0

 

Sulfate An analysis of variance showed significant differences (P < 0.01) in sulfate concentrations in water among the treatments. The sulfate concentration in Treatment IV (mixture of seawater and rainwater) was significantly lower (P < 0.01) than those of Treatments I, II, and III. The significant reduction in the sulfate concentration could be attributed to the anoxic condition at the bottom as detected in the decreasing dissolved oxygen level of the water (Table 2). Sulfate was found to be positively correlated with dissolved oxygen and salinity (P < 0.01). As the dissolved oxygen level

increased, sulfate concentrations subsequently increased in all treatments. Sulfate concentrations were relatively higher in water with higher salinity levels. This observation is corroborated by the fact that seawater with higher salinity levels actually contains higher sulfate ions than freshwater (Church 1975). A significant correlation (P < 0.01) was also found between the concentration of sulfate and ammonia. The presence of small amounts of ammonia in the sediment could have been contributed by the accumulated unconsumed feed (rice bran), which contained 5.7%-10.9% crude protein.

Initially, the washed and dried sand bottom in all the treatments were free of sulfate, but after harvest, significant amounts of sulfate (3,337.0 mg/l-3,983.71 mall) (Table 2) were recorded from the sand bottom. The sulfate concentration of the sediment was higher than the sulfate concentration found in the water. Statistical analysis showed that sulfate in the sediment is highly correlated (P < 0.01) with sediment organic matter. It was also possible that sulfur contained in protein (e.g., methionine) from the accumulated unconsumed food could also have contributed, to some degree, to the level of sulfate in the sediment.

Sulfide Sulfide contents of waters collected from the different treatments were significantly different (P < 0.01) among each other. Hydrogen sulfide was relatively higher during the latter part of the experiment and may have been contributed by decomposing unconsumed food and enhanced by the breakdown of the aeration system. This phenomenon was highly possible because hydrogen sulfide is formed by heterotrophic bacterial metabolism; thus, unionized hydrogen sulfide usually does not occur in well-oxygenated water (Chiu 1988). Sulfide was also found to be positively correlated with salinity, (P < 0.01), pH (P < 0.05), and temperature (P < 0.01).

Ammonia Ammonia concentrations were relatively high three days after stocking. This could be attributed partly to inadequate aeration, as evidenced by the low dissolved oxygen (Table 2). From the sixth day onward, ammonia concentrations decreased, especially after water exchange, but an increase was observed toward the end of the experiment. Again, unconsumed food was noted on the bottom of the aquaria. Correlation analysis showed that ammonia was negatively correlated with temperature (P < 0.01) and dissolved oxygen (P < 0.05), which indicated that ammonia accumulates in the water column in lower dissolved oxygen and temperature.

TABLE 3 Changes in the Total Count of Organisms in Different Treatments (Water Sources) with Rice Bran

Treatment	2 Days	7 Days	14 Days
I (Guimaras)	4,210,000	7,109,000	3,468,000
II (Gui-gui)	1,160,000	3,989,000	920,000
III (SM + Gui-gui)	6,650,000	13,638,000	655,000
IV (SM + freshwater)	5,363,000	12,495,000	1,285,000

 

Sediment Organic Matter Initially, the sand bottom had practically no organic matter. After 28 days, an appreciable amount of organic matter was recorded from the samples taken from the sand bottom (0.81%, 0.97%, 0.85%, and 0.85% in Treatments I, II, III, and IV, respectively) (Table 2). This indicated that organic matter was formed from the different inputs, particularly the rice bran, which, even when put in at a relatively low rate, was not completely utilized by the fish.

Sulfate Concentration and Microbe Populations A whitish film developed on the water and spread across the surface after six days. This film was composed mostly of round and filamentous bacteria and protozoans. The total counts of these microorganisms on the second day were 4,210,000, 1,160,000, 6,650,000, and 5,343,000 for Treatments I, II, III, and IV, respectively. These populations continued to increase until day 7 and finally decreased on day 14 (Table 3). There were indications that the populations of microorganisms were higher in the treatments with lower sulfate concentration (Table 4).

TABLE 4 Population of Organisms (org/ml x 10³) in Seawater with Different Sulfate Concentrations

		January
Treatment		17	24	30
		Number of Organisms
I	1	406	1654	224
(Guimaras)	2	120	92	248
	3	150	772	756
	Mean	225.3	839.3	409.3
II	1	4	2	212
(Gui-gui)	2	112	692	394
	3	30	32	470
	Mean	48.7	242	358.7
III	1	150	200	510
(SM + Gui-gui)	2	162	158	358
	3	276	646	394
	Mean	196	334.7	420.7
IV	1	358	108	511
(SM + fresh water)	2	800	260	1774
	3	2	1600	1344
	Mean	386.7	656	1209.7

 

Obviously, the particles of rice bran harbored microorganisms and became food substrates. An organic fraction of the food was apparently converted into an assemblage of microorganisms that could serve as fish food. Schroeder (1978) and Hobbie and Lee (1980) pointed out that the relative contribution of supplied foods and fertilizers to the growth of the fish is attributable to the sunlit pond ecosystem in which minerals and organic fractions of the food and fertilizers are converted into a complex of algae, bacteria, protozoans, and their mucopolysaccharide exudates, which can be used as food for fish growth.

Changes in the nitrogen content of water from the first hour to day 30 are shown in Table 5. A build-up of nitrogen in all treatments was observed between days 2 and 7, and then decreased tremendously between days 14 and 30. Fluctuations in the number of microorganisms followed the increase and decrease of the nitrogen content of the water. Increases in the nitrogen content could be due to the organisms that adhere to the food substrate and enrich the protein source of the substrate within the one-week period of time (observed in this run). After 14 days, however, the nitrogen content abruptly decreased, probably due to the observed decrease in the microorganisms in the water.

TABLE 5 Changes in the Nitrogen Content (mg/l) of Seawater with Varying Sulfate Concentration with Rice Bran as the Organic Residue

Exposure Time	Treatment I	Treatment II	Treatment III	Treatment IV
0 hour	0.064	0.0257	0.407	0.500
1 hour	0.311	0.504	0.568	0.611
2 hours	0.203	0.332	0.311	0.268
12 hours	0.825	0.536	0.642	0.852
24 hours	0.986	0.558	0.880	1.308
7 days	12.440	4.647	15.018	3.432
14 days	1.80	2.55	1.80	2.70
30 days	1.17	-	-	0.96

Survival and Fish Yield The mean survival of milkfish on a per treatment basis ranged from 40% to 67%, although a general decrease in the average body weight was observed in all treatments (Table 6). An inverse relationship was recorded between the average body weight (ABW) of milkfish at harvest and the levels of sulfate concentration in water. The following are the means of sulfate concentrations and the average body weights of milkfish.

 

Treatment	ABW(g)	Sulfate (mg/l)
I	1.145	477.10
II	1 095	490.79
III	1.195	339.78
IV	1.375	325.05

Although there was a slight negative effect of sulfate on the growth and yield of milkfish, other parameters also could have affected the fish. Salinity, pH, and temperature were all within tolerable limits, but the dissolved oxygen contents were all in the lower range of tolerance.

TABLE 6 Sampling Weights of Milkfish Raised in Seawater in Aquaria for 28 Days

Treatment		Weights in Grams Sampling				Survival
		Initial	1st	2nd	Final (%)
I	1	1.24	1.26	1.08	1.05	63
	2	1.27	1.14	1.08	1.03	66
	3	1.30	1.15	1.05	1.11	69
	Mean	1.27	1.18	1.07	1.06	67
II	1	1.30	1.08	1.03	1.13	50
	2	1.29	1.06	1.02	0.90	38
	3	1.19	1.04	1.03	1.06	31
	Mean	1.26	1.06	1.03	1.03	40
III	1	1.14	1.20	1.13	1.24	38
	2	1.30	1.30	1.12	1.20	73
	3	1.26	1.31	1.06	1.09	38
	Mean	1.23	1.27	1.10	1.18	50
IV	1	1.54	1.41	1.03	1.23	56
	2	1.30	1.71	1.08	1.14	49
	3	1.50	1.31	1.26	1.23	62
	Mean	1.45	1.50	1.12	1.20	56

 

Number of milkfish stocked in each aquarium -- 16.

Second Run

The different feeding treatments started on April 10, 1991. Based on our data, the fish in Treatment IV (commercial feed) were observed to have a significantly higher growth rate than the fish in other treatments. Statistical analyses were run on the water and sediment parameters to determine treatment differences, and are summarized below.

Sulfate An analysis of variance showed that the amounts of sulfate in the water of each treatment were significantly different. Treatment IV, which received commercial feed, had the highest range of sulfate concentrations (787.96-1155.99 mg/l) (Table 7). The average sulfate concentrations showed lower sulfate, over time, in the treatments that received rice bran (Treatments I and II). Fluctuation in the sulfate concentrations could be due to the water change and other factors that affect the sulfate

TABLE 7 Physiochemical Properties of Water and Sediment (Milkfish Aquarium Experiment -Second Run)

	Sulfide (ml/l)
Treatment	April 10	April 17	April 24	April 29	May 9
I	0.020	0.021	0.028	0.031	0.025
II	0.021	0.019	0.026	0.033	0.028
III	0.016	0.022	0.019	0.032	0.028
IV	0.017	0.021	0.023	0.025	0.028
	Sulfate (mg/l)
Treatment	April 10	April 17	April 24	April 29	May 9
I	1,129.70	806.36	973.56	685.96	1,001.95
II	1,177.02	835.28	1,039.80	694.90	1,011.41
III	1,137.59	835.28	1,054.00	761.15	1,035.07
IV	1,155.99	883.12	1,012.46	787.96	1,012.99
	Ammonia (mg/l)
Treatment	April 10	April 17	April 24	April 29	May 9
I	0.146	0.056	0.088	0.129	0.107
II	0.139	0.036	0.108	0.578	0.125
III	0.115	0.066	0.042	0.054	0.068
IV	0.168	0.078	0.044	0.131	0.534
	Phosphorus (mg/l)
Treatment	April24	April 29	May 9
I	1.28	1.98		1.52
II	1.19	1.54		1.44
III	1.15	1.42		1.15
IV	1.10	1.35		1.11
	Soil Sulfate (mg/l)	Soil pH	Organic Matter(%)	P(mg/l)	N (%)
Treatment	Feb. 1	May 10	May 10	Feb. 1	May 10	Ma 10	May 10
I	3,760.20	3,662.97	7.57	0.98	1.09	90.9*	0.067
II	3,347.50	3,925.87	7.78	0.69	1.3	66.54	0.064
III	3,515.77	3,584.10	7.82*	0.91	0.31	56.43	0.012 *
IV	4,109.88	3,641.97	7.77	0.90	0.78	63.14	0.053

 

* Significant based on SYSTATS

levels in water. Correlation analysis showed that sulfate in water is negatively correlated to

phosphorus and temperature (P< 0.01). Also, a slight correlation was observed with soil organic

matter (P < 0.10). However, sulfate levels in soil were not significantly different (P > 0.05) among treatments, when based on analyses made before stocking and after harvest.

Sulfide An analysis of variance showed no differences among the treatments in terms of sulfide concentrations in water. Mean concentration of sulfides ranged from 0.0225 mg/l to 0.0254 mg/l. Sulfide concentrations, in general, tended to increase in time in all of the treatments (Table 7). This could be due to the accumulation of decomposing organic material from unconsumed food on the bottom and an increase in the metabolic wastes of the fish during the latter part of the experiment. Sulfides were found to be highly correlated (P < 0.01) with ammonia, pH, soil organic matter, soil phosphorus, and soil nitrogen.

Ammonia No differences in ammonia concentrations were found among the treatments. The concentration of ammonia in water followed a fluctuating trend from one sampling period to another (Table 7), wherein ammonia was seemingly influenced by certain parameters such as dissolved oxygen, temperature, and organic or nitrogen input. An analysis for this run showed significant correlations of ammonia with dissolved oxygen (P < 0.05), soil nitrogen (P < 0.01), and sulfide (P < 0.01).

Sediment Organic Matter The amounts of organic matter in the sediment increased a*er each experimental run (Table 7), obviously because of the unconsumed food that accumulated as a white precipitate on the sand. The organic matter contents of the sediment ranged from 0.69% to 1.31% (Table 7). Treatment I was supplied with rice bran mixed with 1% cane sugar plus fertilizer, while Treatment III had no food supplement except fertilizer. It seems evident that the unconsumed food contributed significantly to the organic matter in the sediment. Correlation analysis between organic matter and the different parameters showed significant relationship (P < 0.05) with sulfide, soil and water phosphorus, and soil nitrogen. It is a fact that upon decomposition, organic matter will release sulfide and phosphorus as well as inorganic nitrogen.

Phosphorus An analysis of variance for water phosphorus showed significant differences among treatments. The highest mean phosphorus concentration in the water was found in Treatment I (1.59 mg/l P) and the lowest in Treatment IV (1.19 mg/l P), which was not significantly different from Treatment III (Table 7). Phosphorus in water was found to be correlated with sulfate, water pH, sediment organic matter, and soil phosphorus. Sulfates could affect water pH, which in turn determines the solubility of phosphorus in water. Thus, the amounts of organic matter and soil phosphorus are directly related to soluble phosphorus in water.

Sulfate and Microbe Population The total population counts for the microorganisms, mostly protozoans and bacteria, are significant. Highest population counts were obtained in Treatment IV (commercial feed) (Table 7) where the sulfate concentration of seawater was recorded as 523.94 mg/l. Total population counts of organisms were significantly different from each other (P < 0.01). Based on the mean of sulfate concentration (accumulated values in time), total counts of organisms were greater in the treatments with lower sulfate concentrations. There is an indication that the sulfate-sulfide concentrations in seawater negatively affected the total populations of the microorganisms. This is likely, because of the negative effect of acidity on the organisms resulting from high sulfate concentrations in water. The other measured parameters, ammonia (P < 0.05), salinity (P > 0.10), pH (P > 0.10), dissolved oxygen (P > 0.10), and organic matter did not affect the total population.

The density of the microorganisms in terms of the total population counts for the two sampling periods in the aquaria are shown in Table 8. Mean counts show the highest density (26,610,000) in Treatment III (natural food), followed by Treatment IV (commercial feed) and Treatment I (rice bran + refined sugar + fertilizer). In terms of population counts, the treatments were not significantly different from each other, but numerically the treatments that received fertilizers exhibited more microorganisms. Average population counts of 313,000 organisms/1 and 266,150 organisms/1 were recorded from Treatments I and III, respectively, compared to 133,400 organisms/1 and 191,250 organisms/l, respectively, in Treatments II and IV. The positive effect of fertilizer on the development of organisms is consistent with the findings of Hepher (1962) that the primary productivity of chemically fertilized ponds is about 4-5 times greater than ponds that do not receive fertilizer. In terms of the weight of microorganisms that adhered to the tiles, Treatment I produced significantly greater biomass than the rest of the treatments, possibly due to microbial organisms associated with the organic matter (rice bran). Schroeder (1978) reported that large increases in sediment-related microbial protein are regularly associated with the deposition of organic matter. An aerobic environment rich in coarse organic matter can produce large communities of bacteria and protozoans in small straw-like particles that serve as substrate for microbial growth (Schroeder 1978).

TABLE 8 Total Population Count (org./1 x 104) and Weight (g) of Microorganisms on Tiles (Milkfish Aquarium Experiment Second Run)

		April 9		April 26
Treatment/Replicate		Population Count	Weight	Population Count	Weight
Treatment I	1	6.15	0.02	9.10	0.00
	2	13.05	0.12	21.40	0.11
	3	36.25	0.09	26.35	0.07
	Mean	18.48	0.075	18.95	0.06
Treatment II	1	7.20	0.01	17.25	0.02
	2	19.25	0.02	16.40	0.02
	3	12.25	0.02	7.70	0.02
	Mean	12.90	0.017	13.78	0.02
Treatment III	1	57.35	0.01	10.00	0.01
	2	27.35	0.04	13.95	0.003
	3	14.90	0.02	36.15	0.02
	Mean	33.20	0.023	20.03	0.011
Treatment IV	1	34.95	0.012	11.70	0.01
	2	6.10	0.009	15.05	0.003
	3	23.00	0.090	23.95	0.02
	Mean	21.35	0.037	16.90	0.011

 

Fish Survival and Yield The survival of milkfish in all treatments was very high (100%) except in one replicate each of Treatments II (83%) and III (93%) (Table 9). Fish in all treatments registered

a daily growth rate of 4%, 4%, 4%, and 6% for Treatments I, II, III, and IV, respectively. Such growth rates were very low, but they provide information about the indirect effect of sulfate on the growth and yield of fish under the conditions of the experiment. The expectation that Treatment IV would have the highest yield was realized. This is mainly because this treatment used commercially formulated food given at 10% of total fish biomass. The highest yield, however, coincided with the highest level of sulfate, lowest level of sulfide, and a higher level of ammonia (Table 7).

TABLE 9 Sampling Weights of Milkfish Raised in Seawater in Aquaria for 30 Days

 

Treatment		Weights in Grams Sampling			Survival
		Initial	1st	Final	(%)
I	1	1.05	1.11	1.12	100
	2	1.01	1.20	1.00	100
	3	1.10	1.30	1.09	100
	Mean	1.05	1.20	1.07	100
II	1	1.37	1.40	1.30	100
	2	1.12	1.30	1.40	100
	3	1.02	1.10	1.00	88
	Mean	1.17	1.27	1.23	96
III	1	1.20	1.21	1.20	100
	2	0.99	1.10	1.02	93
	3	1.07	1.13	1.04	100
	Mean	1.09	1.15	1.09	98
IV	1	0.97	1.55	1.90	100
	2	1.14	1.72	2.10	100
	3	1.14	1.61	1.80	100
	Mean	1.08	1.63	1.93	100

 

Treatments: I - Rice bran + cane sugar + fertilizer

II - Rice bran + cane sugar

III - Natural food + fertilizer

IV - Commercial feed

Conclusion

A negative influence of sulfate on milkfish yield and on the production of natural food (phytoplankton and other algal forms, bacteria, and protozoa) in seawater was found in this preliminary study. The duration of the experiment was not sufficient to draw concrete conclusions; however, a trend can be gleaned from the results. This experiment needs to be replicated several times to obtain consistent results.

Several experiments addressing the same objectives and overall goals of the project are in

progress. These are being done in glass aquaria, concrete tanks, and brackish-water earthen ponds.

Identification of important feeding niches with the use of stable isotope technology is also in progress.

The results of these new studies should strengthen the findings of this study.

Acknowledgments

The authors thank the Office of Research, USAID, for funding grant number DHR-5544-G-SS

9068-00, the network meeting, and publication of this paper.

References Cited

Chiu, Y.N. 1988. Water quality management for intensive prawn ponds. pp. 79-92. in Chiu, Y.N., L.M. Santos and R.O. Juliano. (ed.). Technical considerations for the management and operation of intensive prawn farms. Univ. Phil. Aquac. Soc. Coll. Fish., Univ. Phil., Iloilo City, Philippines. 172 pp.

Church, T.M. (ed.). 1975. Marine chemistry in the coastal environment. ACS Symposium Series 18. Amer. Chem. Soc., Washington, DC, U.S.A. 710 pp.

Fortes, R.D. 1991. Utilization of selected prawn farming technologies in Western Visayas. Paper presented during the Regional Symposium on R & D Highlights, WESVARRDEC-Central Phil. Univ., May 28, Iliolo City, Philippines. 11 pp.

Fortes, R.D., V.L. Corre, and E. Pudadera. 1986. Effects of fertilizers and feeds as nutrient sources on Oreochromis niloticus production in Philippine brackishwater ponds. pp. 121-124. in Maclean, J.L., L.B. Dizon, and L.V. Hosillos. (ed.). The First Asian Fisheries Forum. Asian Fish. Soc., Manila, Philippines.

Fortes, R.D., B. Posados, R. Cajilig, L. Baylon, E. Pudadera, and I. Belleza. 1989. Technology assessment of prawn, milkfish and mussels in Western Visayas. Terminal Report.

Brackishwater Aquac. Center, Coll. Fish., Univ. Phil. Visayas, Leganes, Iloilo, Philippines. 128 pp.

Hepher, B. 1962. Primary production in fishponds and its application to fertilization experiments. Limnol. Oceanogr. 7:131-136.

Hobbie, J. and C. Lee. 1980. Microbial production of extracellular material: importance in benthic ecology. in Tenore, K. and B. Coull. (ed.). Marine Dynamics. Bell Baruch Symp. on Benthic Ecology. Carolina Press, U.S.A.

Schroeder, G. 1978. Autotrophic and heterotrophic production of microorganisms in intensely manured fishponds and related fish yields. Aquaculture 14:303-325.

Schroeder, G. 1983. Natural food web contributions to fish growth in manured ponds. J. World Mariculture Soc. 14:505-509.

Singh, V.P. 1980. Management of fishponds with acid-sulphate soils. Asian Aquacult. 3:4-6. (SEAFDEC Aquac. Dept. Tigbauan, Iloilo, Philippines).

Strickland, J.D.H. and T.R. Parsons. 1972. A practical handbookfor seawater analysis. Fish. Res. Board Can., Ottawa.

 

Feeding Value Of Fresh Perennial Leguminous Shrub Leaves To Nile Tilapia (Oreochromis Niloticus L.)

M.A.G. Castanares,¹ D.C. Litile,¹ A. Yakupitiyage,¹

P. Edwards,¹ And L.L. Lovshin²

¹Division of Agricultural and Food Engineering, Asian Institute of

Technology

Bangkok, Ihailand

²Department of Fisheries and Allied Aquaculture, Auburn University

Auburn, Alabama

Abstract

The feeding value of fresh pigeon pea (Cajanus cajan), leucaena (Leucaena leucocephala), gliricidia (Gliricidia septum), and sesbania (Sesbania grandifloraJ leaves to Nile tilapia (Oreochromis niloticus) was evaluated in an outdoor recirculating water system comprising 15 5 m³ concrete tanks for 56 days. Commercial catfish pelleted feed was used as a control diet. The protein content of all green fodder was within the limits of the required dietary protein level for Nile tilapia. Average green fodder intake of fish fed ad libitum ranged from 0.2 to 0.4% body weight/day, but fish lost weight during the experiment. Results of this experiment indicate that there is no value in feeding the leguminous fodders tested here to fish.

Introduction

Legume fodder is considered cheap nutritious feed for terrestrial farm animals (NRC 1984). Pigeon pea, gliricidia, leucaena, and sesbania are four common perennial forage legumes that are often used by Asian farmers as animal feed or as green manure for conventional crops. Although some perennial legume fodders have been found to have deleterious effect to farm animals, gliricidia, leucaena, and sesbania are frequently used as ruminant feed in the tropics (Williamson and Payne 1978). Pigeon pea, which is also a food crop for humans, is usually cut 50-75 cm from the ground for forage when the pods begin to ripen (Bogdan 1977).

The nutritive value of these fodders as nonconventional feed for livestock suggests that they are a potential source of herbivorous fish feed for small-scale fish farmers. There have been several studies, especially on the inclusion of leucaena leaf meal in the diet of finfish (Jackson et al. 1982, Ferraris et al. 1986, Wee and Wang 1987a) and Penaeus monodon (Vogt et al. 1986, Penaflorida 1989). However, there is no information on the nutritive value of the fresh leaves of these plants as the major nutritional source for fish.

The present study was, therefore, designed to evaluate the potential use of these fresh legume leaves as the main dietary source for Nile tilapia. This species is primarily a microphagous feeder (Colman and Edwards 1985), but it is also known to ingest plant leaves.

Materials and Methods

This study was conducted at the Asian Institute of Technology (AIT), Bangkok, Thailand, from March 28 to May 23, 1991. A completely randomized design (CRD) with five treatments, in triplicate, was utilized. Treatments consisted of fish fed the four different legumes and a control diet. The feeding trial was conducted in a recirculating system that consisted of 15 outdoor concrete tanks (2 x 2 x 1.2 m) and a biofilter. A flow rate of recirculating water was maintained to provide sufficient oxygen and to circulate the total volume of water through the biofilter unit at least once a day. A rotating octahedron plastic net cage was placed in each tank. The cage had a frame of painted steel and wood and the sides of the cage were covered with a black plastic net (2 cm mesh size). Each cage unit was rotated daily to expose the submerged half to the air to minimize fouling. The experimental cages were stocked with sex-reversed male Nile tilapia (Oreochromis niloticus) with an initial weight of approximately 25 g, at a density of 15 fish per cage.

Pigeon pea, gliricidia, leucaena, and sesbania were cultivated on the AIT campus. Fresh leaves of the four legumes were collected daily from the field and a known weight of fodder was presented to the fish once a day to feed to satiation. A bundle of fodder was tied to a floating PVC frame (0.5 x 0.5 m) in each cage at 0900 hr and the uneaten materials were collected 24 hr after. The dry matter content of offered and uneaten fodder from each tank was determined for two successive days, at biweekly intervals, to obtain the average daily dry matter intake by the fish. A commercial catfish pellet (Charoen Pokphand Company Ltd., Thailand) was the control diet, fed at 3% body weight per day.

Experimental fish were sampled biweekly to record their survival and batch weight, and the feed ration of commercial pellets was adjusted accordingly. All fish were counted and weighed individually at the end of the experiment. Samples of initial and final experimental fish were sacrificed and dried in an oven at 100°C for 24 hours. Green fodder, commercial catfish pellets, and fish carcasses were analyzed for moisture, crude protein, crude lipid, and ash. The crude fiber content of experimental feeds was also determined (AOAC 1975).

Results

The average approximate compositions of the four green fodder types and the commercial pellets are presented in Table 1. The mean body weights of the experimental fish during the feeding trial are shown in Figure 1. The mean final body weights of the fish fed the four legumes decreased compared to their initial mean body weights (Table 2). However, the fish fed commercial pellets grew from approximately 24 g to 93 g during the feeding trial.

FIGURE 1. Growth in average weight (gram) of Nile tilapia during the experimental period; (a) treatments with green fodder and pellet; (b) treatment with only green fodder

More than half (approximately 60%) of the experimental fish fed leucaena fodder developed eye cataracts. Gross morphological disorders were not observed in the other treatments.

TABLE 1 Average Proximate Composition of Green Fodder and Commercial Pelleted Feed

		Percent, Dry Matter Basis
Feed	Dry Matter	Crude Protein	Crude Lipid	Crude Fiber	Ash	Nitrogen Free Extract
Pigeon pea	34.4	24.2	3.6	19.4	9.0	43.1
Gliricidia	20.3	27.3	3.0	12.8	8.8	48.1
Leucaena	28.2	32.5	1.9	12.5	6.3	46.9
Sesbania	20.5	26.3	2.6	12.1	10.5	48.7
Pellet	90.6	36.0	3.2	8.0	10.5	42.4

The dry matter intake of fish fed the experimental fodders was markedly low throughout the experimental period Cl able 3). The body lipid contents of experimental fish fed fodder were extremely low compared to their initial body lipid contents and that of the fish fed the control diet (Table 4).

Discussion

Chemical analysis of the tested fodders indicated that their protein contents were within acceptable limits for Nile tilapia (Tacon 1987). Crude protein contents ranged from 24.2 to 32.5% on a dry matter basis, with pigeon pea the lowest and leucaena the highest. In contrast, the crude lipid contents were relatively low. The crude fiber contents of the fodders was relatively high, ranging from 12.1% for sesbania to 19.4% for pigeon pea. The low lipid and high fiber content of these plants may pose a problem for their use as fish feed.

There was a reduction in the body weight of fodder-fed tilapia. Low feed intake indicated the unpalatability of the tested fodders since the highest estimated total dry matter consumption among the fodders was only 5.5 g for gliricidia-fed fish in 56 days. The corresponding protein intake of fish fed gliricidia was therefore only 1.5 g, which may not be sufficient to maintain the nitrogen balance of the fish. However, fish fed commercial pellets containing 36% crude protein at 3% body weight/day grew by 75.1 g within 56 days. This resulted from 27 g of protein ingested over the same period.

It can be concluded, therefore, that fish growth was hindered by poor fodder intake by the fish. Direct feeding on these fodders could not supply sufficient nutrition for normal growth of Nile tilapia. The relatively low body lipid content of the fodder-fed fish reflects the utilization of stored body fat as an energy source for metabolism during the experimental period, which resulted in weight loss.

The observed low dry matter intake of fodder-fed fish may be attributed to the presence of potential anti-nutritional factors in leguminous fodder. However, a gross morphological disorder resulting from an anti-nutritional factor was only detected among leucaena-fed fish, i.e., eye cataracts.

TABLE 2 Growth Performance of Nile Tilapia Fed with Leaves of Four Legume Species and Commercial Catfish Pelleted Feed for 56 Days + 1 Standard Error)

	Initial Weight	Final Weight	Weight Gain	Relative Growth	Percent
Feed	(g)	(g)	(g)	Rate (mg/g/day)	Mortality
Pigeon pea	25.3	23.1	-2.2	-1.6	4.4
	±0.3	± 1.4	±0.6	±0.4
Gliricidia	24.1	21.4	-2.7	-2.0	2.2
	±0.4	± 1.4	±0.4	±0.3
Leucaena	24.4	21.8	-2.5	-1.8	4.4
	±0.1	± 1.1	±0.5	±0.4
Sesbania	24.9	21.2	-3.6	-2.7	2.2
	±0.6	±5.6	±2.7	±2.0
Pellet	24.0	93.2	69.3	51.5	2.2
	±0.3	±5.7	±2.9	± 1.4
No feed	23.8	21.6	-2.1	-1.6	0.0

 

Sallmann et al. (1959) reported that a reduction of the mitotic index in lens epithelial cells of rats fed leucaena indicated a mimosine induced eye cataract. However, inclusion of leucaena leaf meal (LLM) in the diet of Nile tilapia, with or without pretreatment by soaking, did not cause eye cataracts in previous nutritional studies, although there was growth retardation with an increased inclusion of LLM in the ration (Wee and Wang 1987a, Santiago et al. 1988). The results of the present experiment indicate that it may be necessary to pretreat fresh leucaena leaves to reduce the detrimental effect of mimosine to Nile tilapia. Pretreatment of leucaena fresh leaves by soaking at room temperature for 48 hr has been reported to reduce the mimosine content (Wee and Wang 1987b).

Recirculated water in the tanks had a brown color, probably due to the diffusion of tannins and other pigments from the leaves. Krishnamurthy et al. (1972) and Rao and Mariappan (1972) reported that tannin as low as 6.5 mg/l is lethal to freshwater fish, and that 320 mg/l of tannin was toxic to Catla catla, respectively. However, direct toxicity by water soluble tannins could not be a major influencing factor in the present study because water was recirculated through all five treatments, including the control treatment wherein fish grew at 1.3 g/day. Further research on tannin toxicity is required. It may be possible to reduce the water soluble fraction of tannin in a pretreatment by submerging the fodder in water for a longer duration.

In addition to tannins, pigeon pea fodder was possibly deleterious to fish due to the presence of other toxic factors. Pigeon pea leaves, like the pods, may also contain haemaglutinins, protease inhibitors, cyanogen, and physic acid, which depresses appetite (IDRC/ICAR 1988).

TABLE 3 Feed Utilization Indexes of Nile Tilapia Fed with Four Legume Leaves and Commercial

Catfish Pellet for 56 Days

Feed	Average Total Dry Matter (g)	Average Feeding Consumed Per Fish (% Body Weight) per Day
Pigeon pea	3.0	0.2
	± 0.3	±0.0
Gliricidia	5.5	0.4
	±0.1	±0.0
Leucaena	4.2	0.3
	±0.3	±0.0
Sesbania	2.6	0.1
	±0.3	±0.0
Pellet	75.1	3.0
	±2.0	±0.0

 

The results of the present experiment suggest that it is not possible to use the four tested leguminous fodders as direct feed for Nile tilapia. If toxic components were the major factors affecting low dry matter intake, it is highly unlikely that other more voracious macrophagous fish such as grass carp (Ctenopharyngodon idella), giant gourami (Osphronemus goramy), and silver barb (Puntius gonionotus) would be able to utilize these fodders efficiently. It is known that non-ruminants have a lower toxicity tolerance to mimosine and other toxic factors than ruminants (IDRC/ICAR 1988).

TABLE 4 Initial and Final Carcass Composition of Experimental Fish

	Percent	Percent, Dry Matter Basis
Treatment	Dry Matter	Crude Protein	Crude Lipid	Ash
Initial Fish Sample	24.1	60.8	20.9	14.2
Pigeon pea	13.6	64.4	2.2	29.3
Gliricidia	12.8	61.6	3.4	31.0
Leucaena	12.2	61.2	3.3	25.9
Sesbania	13.5	63.5	2.4	28.3
Pellet	20.9	64.2	21.3	12.6
No feeding	10.7	63.5	2.7	30.4

 

However, before generalizations are made, further research should be conducted to verify the

feeding value of these fodders for other herbivorous species. Identification of the anti-nutritional

factors and the development of a cheap pretreatment for these fodder may be necessary to improve their potential as fish feed. As an alternative, however, the relatively high nitrogen content of legume leaves suggests that they may have potential as green manure and serve as a source of nitrogen for fertilizing fish ponds.

Acknowledgments

The authors thank the Office of Research, USAID, for funding grant number 493-5600-9-44

0075-00, the network meeting, and the publication of this paper. Peter Edwards and David Little are

seconded to AIT by the Overseas Development Administration (ODA), London.

References Cited

Association of Official Analytical Chemists (AOAC). 1975. Official Methods and Analysis. 12th Edition. Washington, D.C.

Bogdan, A.V. 1977. Tropical Pasture and Fodder Plants (Grasses and Legumes). Tropical Agriculture Series. Longman Inc., New York, U.S.A.

Colman, J.A. and P. Edwards. 1985. Feeding pathways and environmental constraints in waste-fed aquaculture: balance and optimization. Paper presented at the Bellagio Symposium on Detritus and Aquaculture. August 1985. Bellagio, Italy. 42 pp.

Ferraris, R.P., M.R. Catacutan, R.L. Mabelin, and A.P. Jazul. 1986. Digestibility in milkfish, Chanos chanos (Forsskal): effects of protein source, fish size and salinity. Aquaculture 59:93-105.

International Development Research Centre/Indian Council of Agricultural Research (IDRC/ICAR). 1988. Non-conventional Feed Resources and Fibrous Agricultural Residues. Strategies for expanded utilization. Devendra, C. (ed.). Proceedings of a consultation. 21-29 March 1988. Hissar, India.

Jackson, A.J., B.S. Capper, and A.J. Matty. 1982. Evaluation of some plant proteins in complete diets for the tilapia Sarotherodon mosambicus. Aquaculture 27:97-109.

Krishnamurthy, V.S., C.A. Sastry, and R. Bhaskaran. (ed.). 1972. Treatment and disposal of tannery and slaughterhouse waste. CLRI, Madras, India.

National Research Council (NRC). 1984. Leucaena: Promising Forage and Tree Crops for the Tropics. Second Edition. National Academy Press, Washington, D.C.

Penaflorida, V.D. 1989. An evaluation of indigenous protein sources as potential component in the diet formulation for firer prawn, Penaeus monodon, using EAAI. Aquaculture 83:319-330.

Rao, A.V.S.P. and M. Mariappan. 1972. Toxicity of tannery waste and their components to fish. in Krishnamurthy, V.S., C.A. Sastry and R. Bhaskaran. (ed.). Treatment and disposal of tannery and slaughterhouse waste. CLRI, Madras, India.

Sallmann, L.V., P. Grimes, and E. Collins. 1959. Mimosine cataract. Am. J. Ophthalmol. 47:107117.

Santiago, C.B., M.B. Aldaba, M.A. Laron, and O.S. Reyes. 1988. Reproductive performance and growth of Nile tilapia (Oreochromis niloticus) broodstock fed diets containing Leucaena leucocephala leaf meal. Aquaculture 70:53-61.

Tacon, G.J. 1987. The Nutrition and Feeding of Finned Fish and Shrimp. A Training Manual. FAO, Brazil.

Vogt, G., E.T. Quinitio, and F.P. Pascual. 1986. Leucaena leucocephala leaves in formulated feed for Penaeus monodon: a concrete example of the application of histology in nutrition research. Aquaculture 59:209-234.

Wee, K.L. and S.S. Wang. 1987a. Nutritive value of leucaena leaf meal in pelleted feed for Nile tilapia. Aquaculture 62:97-108.

Wee, K.L. and S.S. Wang. 1987b. Effect of post-harvest treatment on the degradation of mimosine in Leucaena leucocephala leaves. J. Sci. Food Agric. 39: 195-201.

Williamson, G. and W.J.A. Payne. 1978. An Introduction to Animal Husbandry in the Tropics. Longman, London.