Previous Page Table of Contents Next Page

Appendix 4



Freshwater fish culture is a branch of agriculture or animal husbandry, and uses the land, water and labour of farmers. If it requires the use of wastes, manures, fertilizers or feeds these must be of kinds which are available to them.

Like other animals fish do not produce food. They can only convert a part of what they eat into growth. The rest they use up for their own needs. (Plants are different. They can make new carbohydrates, proteins, fats, etc., using only air, water, sunlight and minerals). So growing fish only increase the amount of food available for people if the fish consume things which people cannot or will not eat and convert them into fish flesh which people will eat.

Low intensity animal rearing systems rely on food that the animal searches for and finds for itself. The output will be low (generally not more than a few hundred kilograms per hectare per year). For land animals this may still be worth the farmer's effort. He lets the pigs, goats, sheep, chicken or cattle wander around and only has to build a shelter for them to use at night and some fences. The animals can collect their food over a wide area. For fish this system is generally not worthwhile because the fish have to be kept inside their ponds. They cannot wander around over a wide area (or if they do they will not return). The amount that can be grown in a small pond is very little and big ponds are expensive to construct.

Medium intensity small-scale animal rearing systems require more input by the farmer. He tries to improve the pasture land for his cattle. He feeds wastes to his animals. He may use combinations of animals on the same patch of land. He may practice complicated animal and plant rotations and spend a lot of effort manuring, composting and moving by-products around. His output per unit area is likely to be several times higher, but so must his effort and skill be.

Medium intensity large-scale animal rearing systems are comparable to the big cattle estates. There are areas in the highlands, such as big swamps, which might allow fish farming of this type, but these would require huge development projects and lots of capital. There are also lowland areas in which this might be possible. They would have the advantage of being near to large markets in the towns. They could more easily exploit the resources produced by the towns, such as sewage, slaughterhouse, brewery and mill by-products, food processing wastes, scraps from large restaurants and institution canteens. They would also benefit from higher water temperatures. In some lowland areas this kind of enterprise could have a fair chance of success.

High intensity animal husbandry is practised where cattle, pigs or poultry are raised on feeds which are prepared for them and brought to them. Trout farming is usually of this type. The animals or fish are often given complete diets, which of course have to be grown somewhere else or imported. Although the yield per unit area of the feed lot or trout tank may be almost unlimited, the actual yield per unit area of land must take into account the fields which are used in growing the soybeans, corn or wheat which is fed to the animals. From the point of view of human nutrition, it would be much better to feed these products directly to people of course. In certain areas specific high-protein wastes are used such as silkworms, shrimp processing wastes or so-called trash fish, generally for the culture of high-priced species for a luxury market.

Chinese systems of growing a mixture of fish using huge quantities of plant material and manure produce high yields but require a very great input of skill and work. They also depend on using carefully balanced mixtures of plant, plankton and detritus-eating fish which are not available in PNG at the moment.


2.1 Indonesia

Indonesia is tropical, with a climate quite similar to PNG. Most people live below 1 000 m altitude. Many thousands of tons of carp are grown in Indonesia, particularly in Java. Most growers work on a small scale. They often concentrate on only one stage of carp production. Some specialize in in breeding and produce fry of 2–3 cm length (1/4 g or less). They use small ponds or hapas for spawning and grass or natural fibres for egg collecting as described in Appendix 3, section 3. Fry are grown in ponds of 500–2 000 m2. The fry are often sold at special live fish markets to other growers who raise them to fingerling size of a few grams (5–8 cm) or a few tens of grams (8–12 cm) in ponds. These fish are then grown to the relatively small size of about 200 g, at which they are often eaten. This may be done in ponds, in rice fields during the 2 months or so during which the rice is grown standing in water, or in fields which have been flooded, between crops of something else. Another stage of further growing, up to the luxury size of ½ kg or more, is sometimes carried out in running-water ponds or cages at very high density. In the running-water ponds fish have to be fed but the cages are submerged in rapidly-flowing streams in quite heavily populated suburban or village areas. The fish grow quickly on human faeces and kitchen refuse, while the rapid flow keeps the water aerated, despite its heavy load of organic materials.

The main feed used at all stages is the fine bran, rice polishings and broken rice grains which are a by-product of the rice mills. This material amounts to 10–15 percent of the huge local production of rice. Only a little is used by poultry and cattle, and most people do not keep pigs as they are Muslims. The practice of building toilets above ponds or defaecating into streams and irrigation canals also contributes to fish production (but means that all water for drinking has to be boiled). Animal manures and some composts and green manures are also used. By breaking the rearing process into many stages of a few weeks or months each, farmers use their ponds more efficiently. They also ensure a regular income from frequent small sales and keep their risks down.

Many people prosper from fish farming (although for others it may be a hobby). Carp are often sold for consumption alive. They are among the most expensive and higly-esteemed fish, eaten by well-off people or for celebrations. Small dried salted marine fish are more commonly consumed by the less wealthy.

Ponds are generally quite small, not more than a few thousand square meters, and shallow. Often they are simply rice-fields, with their banks raised to give a water depth of 30–50 cm, receiving their water from the same irrigation system. In addition to carp, tilapias, gouramies and other cyprinids are the main fish cultivated in fresh water. Productivity figures are hard to obtain but probably do not exceed 1–2 t/ha/yr in most ponds.

2.2 China

China has the biggest cultured freshwater fish production of any country. Complex polyculture systems are often used, combining silver carp, big head, grass carp, black carp, mud carp and common carp. Bream, mullet, tilapia, snakehead and catfish are also included in various combinations in some areas. Farmers use green fodder (edible plants), composted vegetable wastes, manures, fertilizers and a range of fine feeds; bean cakes, peanut cakes, rape cakes, rice bran, wheat bran, grains, silkworm pupae and brewing residues (Zhong Lin et al., 1980). Common carp are often a minor component in the ponds; the various Chinese carps which live on vegetation, plankton and detritus have greater importance. Yields appear to reach 4–8 tons/ha/yr, but to obtain this 70–230 t/ha of green fodder and up to 70 t/ha of manure will have been applied in addition to small amounts of fine feeds. The overall relationship appears to be 20–30 tons of green fodder and manure per ton of fish produced. The growth period has been divided into as many as five stages which take place in fry ponds, small-size fingerling ponds, medium-sized fingerling ponds, large-sized fingerling ponds and fattening ponds. Stocking rates and sizes differ for each species in these ponds, and each will have to spend a different number of days at each stage so there must be almost continuous netting, sorting and transferring in addition to feeding. Such heavily fed, fertilized and stocked ponds have to be closely monitored, and farmers have learned how to predict problems according to the behaviour of the different species at various times of the day or night and how to manage the pond using a wide range of locally-available fertilizers, manures, composts, feeds and poisons. It is said that large ponds produce big fish and pond areas of 2 500–7 000 m2 are recommended for fattening, with depths up to 2 or 3 m to give more space and reduce temperature fluctuations.

2.3 Japan

Japan has hot wet summers similar to those of the tropics, but in winter temperatures can fall below zero, and some areas have a lot of snow. Carp are grown particularly in inland hilly areas in the middle of the main island where fresh sea fish used to be hard to obtain. Carp growing systems in Japan consist of irrigation ponds, small fishponds, running-water ponds, and floating net cages (Suzuki, 1976). Irrigation ponds vary in size from half to twenty hectares each. The water is generally stagnant. They usually have mechanical (electric) aerators and often have automatic pellet distributing devices. They are stocked in spring (April-May) with 40–100 g fingerlings at a rate of about one fish per square metre and fed pellets or silkworm pupae 4–5 times daily. Fish are sold in the autumn (October or November) at about 800 g average size, with survival rates of 50–85 percent and feed conversion coefficients (dry feed/wet fish) of 1.3–1.7. Average production in 1973 was 3.5 tons/ha (350 g/m2). Small fishponds generally also have stagnant water. They are from 900–3 000 m2, are stocked in April with 0.5–1 fish/m2 at average weight 80–120 g, fed intensively till autumn and harvested again when fish are about 800 g each. Survival rates around 90 percent and feed conversion about 1.7. Average harvests for ponds of this type are given as 5–8 tons/ha (500–800 g/m2), but according to Suzuki (personal communication) 10 tons/ha is considered good and yields up to 22 tons/ha are achieved. In running water ponds of 20–100 m2 and 1.5 m depth, food conversion rates are lower (2.4–2.7) and yields in the range 100–200 kg/m2. Net cages are generally supported on rafts in lakes or large irrigation ponds. They vary from 25–120 m2 with a depth in the water of about 1.5 m. Stocked with about 30 fish/m2 (average size 50–120 g) and fed as intensively as the ponds, fish harvests are 20–50 kg/m2, survivals around 90 percent and food conversion about 1.4–1.6.

The production figures from Japan, all the more remarkable when the short growing season is taken into account, are clearly the result of many factors. Pellets are said to contain 50 percent fishmeal, 39 percent wheat flour, 6 percent alfalfa meal, 3 percent yeast, 1 percent salt and 1 percent vitamins, and to have a minimum protein content of 39 percent. Pellets are the main feed overall, but are supplemented at different times of the year by dry or fresh silkworm pupae, fresh or frozen fish, boiled wheat, etc. Feeding devices automatically scatter regulated quantities of pellets at preset intervals. When this happens, the fish gather in that area in a dense mass and feed vigorously so that very little of the food even reaches the pond floor.

Aeration devices allow fish to live and grow in ponds at densities which would otherwise lead rapidly to suffocation. They are needed particularly at night in hot weather and when there is no wind. They generally consist of wheels mounted on a horizontal axis just above the water surface. Each wheel has several perforated steel vanes set across the plane of the wheel. They are rotated at high speed by en electric motor.

Genetic work is carried out at several research stations and institutions. There are many pure and applied research centres, plenty of growers cooperatives, and fish grower's magazines which are published regularly and contain articles and advertisements. Thus, information on techniques, new products, disease control and so on is quickly spread.

2.4 Israel

Israel has very hot dry summers. Rain falls only in the winters. Frost is unusual except in a few mountain areas and water temperatures rarely drop below about 8°C. Much of the country is desert and many of the water sources are slightly brackish. It is these that were originally used for carp culture, which started about 40 years ago. Later, other introduced or local fish started to be cultured in the same ponds. Now common carp, Chinese carps (grass carp, bighead, silver carp and silver carp-bighead hybrids), tilapias (various hybrids of O. aureus and O. niloticus) and mullets (Mugil cephalus, M. capito and M. aureus) are produced on quite a large scale. Yields average 3–4 tons/ha per growing season. There is also small-scale or experimental cultivation of Macrobrachium, eels, sea bream, penaeid shrimps and oysters.

The quarterly fisheries journal “Bamidgeh” gives a good account of continuous experimentation and development in practical pond polyculture. This journal is sent free to any institute in exchange for their own publications. Tables have been produced which attempt to show farmers the most effective and economic feeding rates, taking into account natural pond fertility. The most commonly-used feed-stuffs are sorghum grains and 25 percent protein pellets. Table 1 was calculated taking the daily natural protein production of ponds with standard fertilization and manuring rates as being equivalent to 15 kg of 25 percent protein pellets per hectare. (These standard rates involve the application of ammonium sulphate and superphosphate at 60 kg/ha every two weeks over the 200 day growing season and a total of about 2–4 tons/ha dry poultry manure during this time.) When using this table, regular fish sampling must be made to keep track of growth and adjust the ration. Fish are usually sampled every two weeks by feeding in one corner of a pond and then quickly closing off that corner with a small pond seine. Expected growth rates are shown in Table 2 (Marek, 1975). For temperatures of 20–24°C, quantities of feed are only 70 percent of those shown. (Whole grains other than sorghum can, of course, be used with slight adjustment for their different protein and energy content.)

Table 1 concentrates on the larger sizes of carp. An alternative approach is to adjust the protein content of the feed by mixing 25 percent protein pellets and sorghum (or some other grain) according to the fish density (Table 3) and then feed on a bodyweight percentage in accordance to fish size. For 24°C or above, the Fish Grower's Association of Israel gives the figures as shown in Table 4 (for temperatures of 20–24°, reduce by 25 percent).

Fish farms in Israel are large, generally having over 40 ha of ponds each. Some of them use slightly brackish water from (inland) springs which has otherwise only limited use for agriculture. Many ponds are also water storage reservoirs for irrigation. Some farms use a percentage of urban wastewater. Water pumping is often involved; in summer in particular water is often pumped back up from drainage channels into ponds. Most ponds, except those stocked super intensively (over about 10 tons/ha) are essentially without flow except for what is needed to make up seepage and evaporation. (Evaporation is generally taken as being around 1 cm/day. This is equivalent to a flow of about 1.2 1/sec/ha.)

Almost every fish farm is part of a larger communal farm (kibbutz) owned by the members. The regular workforce on the fish farm may only be 6–8 but they can call on extra help when needed. The fish farms are highly mechanised, with tractors and other vehicles, aerated fish transporting tanks which can hold 4 tons of live fish, large seine nets often 100 m long, small boats for spreading chemicals in the ponds, etc. The use of aeration devices has become widespread in the last decade; these are often of the Japanese water-wheel type although some vertical-shaft sewage aerators are used as well. Where ponds do not yet have an electricity supply, land-based blowers can be installed with submersed air-lift tubes or diffusers. There are several types of feeders in common use. These include tractor-powered blowers which scatter food over a large area of the pond, demand feeders activated by the fish, and feeders which scatter preset quantities at certain intervals. A single pond may easily hold ten tons of fish or more. Once netted, fish are lifted up into the transport tanks with the aid of powered lifts or fish pumps. There are sorting, packing and refrigeration facilities. The kibbutz usually has workshops and garages where the equipment can be maintained and repaired. The chief limitation on production is probably the relatively small domestic market for freshwater fish.

Table 1
 Stocking density (fish/m2)
Average fish weight (g)0.2–0.40.4–0.60.6–0.80.8–1.21.2–1.51.5–2.02.0–5.0

Top figure - grams of 25% protein pellets per fish per day
Bottom figure - grams of sorghum grains per fish per day
For 20–24°C feed only 70% of these quantities

Table 2
Average fish weight (g)Expected daily weight increase per fish (g)

Table 3
Fish densitySorghum25% protein pellets
(g/m2)(% weight)(% weight)
below 751000
above 2200100

Table 4
Average fish weight (g)Feeding rate (% bodyweight)
900–1 0002.4


How much natural food can a pond produce and how much fish growth will it support? This is a key question where fish growing without the use of much supplemental feeding is planned. The answer of course is not simple; it will depend on the natural fertility of the soil and water, the climate, the turbidity, the management and so on. However, here are some calculations which may be helpful.

Boyd (1979) gives gross photosynthesis values for fertilized catfish ponds (southern USA) of 1–3 g carbon fixed/m2/day. He quotes Hepher as giving values of 3.3–6.4 g/m2/day for Israeli ponds. Unfertilized ponds will be far less productive. Since carbohydrate has the general formula (CH2O)n' if we take the figure of 3 g/m2/day and multiply it by 2.5 this gives an estimate of dry weight production of organic matter of 75 kg/ha/day. If the average algal cell contains 25 percent of its dry matter as protein (Cho et al., 1985) this suggests a primary productivity of around 20 kg/ha/day of protein and 55 kg/ha/day of carbohydrate. These figures are intended as rough estimates only, with fats, minerals and vitamins being ignored.

Common carp cannot utilize phytoplankton directly. Therefore these 75 kg/ha/day are not available to them. In fact Marek (1975) found that in carp ponds in Israel which were being fed sorghum intended for maintenance only, the growth which actually occurred could be explained on the basis that the carps' natural food contributed about 4 kg/ha/day of protein to their intake. If we estimate that the natural food of carps contains about 50 percent protein on a dry matter basis, then the pond is only contributing about 8 kg/ha/day of food to the carps' total intake.

The ratio of these figures is interesting. Out of around 20 kg/ha/day of protein which the pond produces the carp utilize about 20 percent, while from a total organic matter production of 75 kg/ha/day they can only make use of about 11 percent. This illustrates the losses in going from one trophic level to another and in raising common carp which rely largely on an animal (invertebrate) diet. The potential benefits to be gained by including phytoplankton-feeding fish in the pond are clear.

These rough figures also give some indication of the amount of fish production which one could expect from natural production alone. From 8 kg/ha/day (dry weight) of natural food, one would not expect more than about 4 kg/ha/day of fish flesh production (wet weight). In a tropical situation where the growing season is year-round, this would be equivalent to about 1½ tons/ha/year. Quite a lot of the protein in this natural food would be burned up for energy. But when this rich 50 percent protein diet is supplemented by a protein-poor starchy energy source the carbohydrate supplement provides much of the energy needs of the fish leaving more of the protein available for growth. In ponds in Israel fertilized and manured in the standard way (Appendix 4, section 5.3), stocked with 600 g fish at about 2 000–2 500 fish/ha, and subsequently fed on average 10 kg sorghum/ha/day, a growth of about 8 kg/ha/day was achieved. This is equivalent to almost 3 tons/ha/year in an area where water temperatures allow growth all year.


Fish farming specialists talk about ponds in terms of hectares or acres. An acre is about 4 000 m2, the size of a football pitch. A hectare is 10 000 m2, i.e., 100 m by 100 m. A quarter of an acre means 1 000 m2, for example, 40 m long and 25 m wide. To get an idea of a piece of land of this size, pace it out (one long step is about a metre) and mark the corners. It is a very small pond to a fish farming expert, but quite large to a man with only a spade.

In fact the ponds built by farmers have been very much smaller than this. Most have been less than 100 m2. A pond of 100 m2 is quite easy to imagine. If it were made square it would be 10 m × 10 m (but of course it can be any shape). It is only one hundredth of a hectare, or one fortieth of an acre. This is less than the minimum recommended pond size, but it is a useful size on which to base some calculations.

How long does it take to build a pond of this size? If a pond is built on reasonably flat ground, soil is dug out of the pond floor and placed around the edges to form banks. If a 10 m × 10 m square is excavated to a depth of 40 cm, and the banks built up to a height of about 60 cm, a water depth of about 80 cm should be possible. To do this 40 m3 of earth must be dug up and moved. FAO pond engineer, Mr. J. Kovari, estimates that on big ponds (where the earth has to be moved a long way from the middle of the floor to the banks) a man can shift 1½ – 2 m3 of earth per day, but on small ponds (where the earth can be thrown out directly) 3 m3/day should be possible. (Digging small ponds at Aiyura was carried out at about this rate.) The banks may still require some additional work, drainage and supply channels have to be dug, and a water outlet constructed. So on a flat but drainable bit of land consisting of grass and soil over clay, not too far from a suitable water source, a pond of 100 m2 could be built in 15–20 man days of work. A slight slope to the ground may reduce this.

However, for the construction of barrage or diversion ponds the time taken may be very much less. These are ponds built by putting a dyke across the floor of a suitably shaped valley, the soil of which may be swampy and not particularly suitable for gardens or other types of agriculture. Many details are given by Huet (1970). If there is a stream flowing down the valley floor an adequate stream bed or channel has to be constructed, passing beside the ponds to avoid the danger of flooding. Ponds may be filled by part of the stream flow or by rainfall, seepage and runoff from surrounding slopes. Large diameter overflow pipes or protected spillways will be needed. The amount of earth which has to be shifted will be much less than with dug ponds. Also the fertility may be higher, as the topsoil remains on the pond floor instead of being piled up on the banks. Villagers' ponds at Omoara, near Aiyura, of about 200 and 350 m2 were apparently constructed in ‘one or two’ and ‘three or four’ days by one or two people. Though leaving room for improvement, crops of fish had been raised in them and carp had bred in one of them too.

How much fish can one expect to produce in a small pond? This will depend on the way the pond is looked after, the climate, the nature of the soil and water and so on. However by looking at results obtained by carp growers in different parts of the world some general ideas can be obtained. These are shown below in two columns. The first is the usually-given figure of kilograms of fish produced per hectare per year. The second shows the same production, but expressed as the amount of fish produced per hundred square meters in six months (in other words the first figure divided by 200), to make it easier to compare the production of fish with, for example, garden crops.

SystemYield range kg/ha/yearkg/100 m2/6 months
Carp only, unfed, pond not manured200 – 4001 – 2
Carp only, unfed, pond manured400 – 1 6002 – 8
Polyculture, unfed, pond manured1 000 – 3 0005 – 15
Carp only, fed low protein, pond manured500 – 2 5002½ – 12½
Polyculture, fed low protein and/or pond  
very heavily manured1 000 – 5 0005 – 25
High protein feeds, natural oxygen only2 000 – 5 00010 – 25
High protein feeds, aerators/running wateralmost unlimited

The growing season in the PNG highlands continues all year round but temperatures are cooler than in tropical lowland regions. People here may not want to eat very small fish whole as they do in parts of Asia. One cannot expect to get yields as high as the very best results of farmers or experimenters who have a great deal of experience and technical back-up.

Introducing new species will require government approval. Chinese carps are obvious candidates as escaped fish are unlikely to reproduce. (They require induction to spawn outside their native rivers.) Javanese carp (Puntius gonionotus or P. javanicus) was tried some years ago. It is a good, largely herbivorous pond fish below about 800 m, but highland temperatures are probably too low. It should be possible to prevent uncontrolled breeding of tilapia, including the species now present in PNG (O. mossambicus) and raise fish of a good size by growing males only. Marine fry, particularly mullets, ofen fit well into freshwater polyculture. There may also be good untried native freshwater fish from elsewhere in the country.

Fish culture is usually promoted for one of two reasons; either to make money or as a method of improving villagers' diets. Estimates of yield per day's labout with carp monoculture systems are shown below.

Labour involved in pond construction and maintenance of structures

APond dug on flat ground or slightly sloping
Time spent on construction/100 m216 days
Assumed life of pond8 years
Annual pond maintenance/100 m23 days
Thus: annual labour on pond itself5 days
BBarrage or diversion pond
Time spent on construction/100 m22 days
Assumed life of pond8 years
Annual pond maintenance/100 m23 days
Thus: annual labour on pond itself3¼ days

Time involved in fish management and returns

ALow level of management
(Stock once, harvest once, no feeding or manuring)2 days/yr
Estimated yield2 kg/100 m2/yr
Return on (a) dug pond290 g fish/day
(b) barrage pond
380 g fish/day
B.Middle level of management (stock twice,harvest twice)
Manure 2 kg/week, feed vegetable scraps 0.5 kg/week10 days/yr
Estimated yield8.5 kg/100 m2/yr
Return on work (a) dug pond570 g fish/day
(b) barrage pond
640 g fish/day
CHigh level of management (stock four times, harvest monthly)
Manure 5–10 kg/week, feed vegetable scraps 200 g/day16 days/yr
Estimated yield15 kg/100 m2/yr
Return on work (a) dug pond710 g fish/day
(b) barrage pond
780 g fish/day

Tinned fish costs about 1-1½ Kina/kg and plantation wages are about 1.5-2 Kina/day. For somebody who has a chance to work on a plantation therefore, this would seem a better option. He could buy more fish with the wages earned than he could grow by working the same amount of time on a small fish pond. (In fact, one should also take into account the chance of failure with ponds, and many farmers talk of theft and flooding. The risk of a total loss of stock before they can be harvested could be 25 percent or more.)

Factors which might make the picture more favourable to fish farming are as follows. First, although the calculations include estimates for the time spent on feeding the fish, this is something pleasant to do at one's leisure, not really work. Second, it was estimated by the consultant that it would take an hour to collect 4 kg of manure by wandering around the village with a bucket looking for pig dung. Anybody who keeps a pig or other domestic livestock would be able to collect manure in very much less time. Thirdly, clearly the work of managing a pond does not increase in direct proportion to the size of the pond. The work of managing a pond of 400 m2 would probably only be about double that of managing one of 100 m2, not four times. So a well-managed barrage pond of 400 m2 requiring a total of 45 days work per year and yielding 60 kg of fish per year looks more worthwhile.

No estimate has been made of the possibility of selling the carp, because of lack of information about the price. Similarly, more complex systems using ducks to fertilize the pond would require information on the expected survival, growth rates, food conversion efficiency, feed costs and market prices for ducks. The value of vegetable scraps has been ignored in this very simple analysis; in times of food shortage small sweet potato tubers would be better eaten directly than fed to fish; in times of plenty they would have a certain alternative value if fed to pigs.

Of course, there are other ways villagers try to earn money or grow more and better food. Fish farming is just one of many options. It would be very useful if an agricultural college or department of agricultural economics could compare the expected financial or nutritional returns from fish ponds with those from increased staple production, wing beans or wingbean tubers, other legumes, sugar cane, corn, green vegetables, free range chickens, sheep, coffee, etc.


One of the most urgent tasks of the Highlands Aquaculture Development Centre, Aiyura, is to test growing systems which are suitable for small farmers. This has not been done in all the years of the Centre's existence, or at least there are no records to be found. Methods which are being recommended to farmers have not apparently been tested first at the Centre itself.

Due to the short duration of the consultancy there was not enough time to rectify this situation. However, a number of ideas based on observations made when visiting farmers' ponds and on data from other countries are put forward in this report to indicate directions that the work of improving carp growing might take. They should of course be tested in practice before being promoted.

5.1 Stocking and Harvesting Policy

There is confusion as to what farmers are being advised to stock. Is it one fish/m2 as often said, one fish/2.5 yd2 (as stated in “Grow Good Carp” book 2A for management of fertilized ponds without supplementary feeding) or one fish/7–8 yd2 (book 2B for ponds with supplementary feeding, both produced by the Dept. of Information and Extension Services, 1970)?

An appropriate stocking density depends on the productivity of the pond, fish sizes, types and quantities of fertilizers and manures, types and quantities of feeds, temperatures, mortalities, growth potential of the carp being produced here and so on. All of which are unknown and likely to vary. The following scheme is intended to be self-adjusting to compensate for this uncertainty.

First of all target figures should be set based on the level of management which it is anticipated the pond will receive. The farmer must be asked whether he intends to buy chemical fertilizer and if so, how much. If he is prepared to collect animal manure, how many bucketfulls might he be able to find each week? How much waste food or scraps does his household produce, what will go to the pigs and what into the pond? How far is the pond from his house? He is also asked at what size he wants to harvest. It should be borne in mind that fish can be eaten at any size; in Indonesia carp are well appreciated at 100 – 200 g. It is generally more efficient to produce 4 fish of 250 g each than to produce 1 of 1 000 g. The farmer should be shown fish of various sizes and asked at what size they appear to him to be worth eating. This is his target size.

Based on this discussion realistic a figure is decided on for estimated productivity (see Appendix 4, section 4). For example a particular farmer may seem to be able to provide a medium level of management which may give an estimated productivity of about 500 kg/ha/yr, and he wants to eat his fish when they reach 250 g. Therefore he should be able to harvest 5 kg of fish/year from each 100 m2 of pond area, that is 20 fish/yr/100 m2. If his pond area is 300 m2 he should be able to harvest 60 fish/year, or 5 per month. This is his monthly cropping rate.

If a pond is stocked with very small fish the total weight of fish in the pond (the fish biomass) at the beginning is very low. As the fish grow, this weight increases until, just before harvesting it is high. (If a number of fish are grown in a pond from an average weight of 5 g to an average weight of 1 kg, and if all survive, the fish biomass in the pond increases over this period by a factor of 200). This is a very inefficient way to use a pond. For most of the rearing period the pond will be very much under-utilized. This is why it is usual to break the rearing period into a number of stages and stock each successive stage at a lower density. Thus 1 g fry might be produced at 100 fish/m2, these grown to 10 g at 15 fish/m2, then restocked for growing to 100 g at 2 fish/m2 and finally fattened to 1 kg size at 1 fish/3 m2 in a system using fertilization and good feeding. Or the figures might be about one fifth of these in a system based on a moderate level of manuring only.

Now obviously a farmer who has only got one pond cannot do this. But what he can do is to keep the biomass in his pond roughly constant by cropping his biggest fish regularly, making room for the others to grow. In fact, if there is a range of sizes in the pond there is another benefit. As small carp and big carp will tend to choose food items of different sizes this will give, to a certain extent, some of the advantages of polyculture. Mixed-age culture is practised in parts of China for this reason although there it is also combined with the use of more than one species.

If the aim of small ponds is to improve the nutrition of the pond owner's family, as is so often stated, then it seems illogical to tell the farmer (as the “Grow Good Carp” books do) to harvest all his fish at once after 6 months or a year and then sell or eat them all. It is better to harvest small numbers at frequent intervals.

Having decided on a target size and a monthly cropping rate, the farmer is given the appropriate number of fish, in this case 60, or 1 fish/5 m2. These should be large fingerlings of 20–50 g each. At this stage the hatchery must supply good-size fingerlings. Later on, if fish culture becomes popular in a particular group of villages, one farmer can be encouraged to specialise in rearing fry to large fingerling size for the others.

The farmer stocks and manages his pond and watches his fish. However uniform in size they are at stocking some will grow faster than others. When he thinks that the largest fish in the pond have reached the target size he should partly drain the pond until he can catch one or more of these fish (using a basket, a small net or just his hands). From this moment on he should harvest his fish at the monthly cropping rate or above it. In the case of our example above, this is 5 fish per month. It does not matter if he harvests on a weekly or a monthly basis, and it doesn't really matter if one month he harvests four fish and the next month six, but he should not fall below the monthly rate when averaged over a couple of months. Each time he should harvest the largest fish (unless he plans to keep a few as breeders), and he should harvest them whether they reach the target size or not. Now he must keep a record of two numbers. The first and most important is the success total, i.e., the total number of fish taken out which have reached the target size. The second is the failure total, i.e., the number of fish taken out which were below the target size. Periodically he should get new fingerlings for stocking in his pond, but he should only get a number of new fingerlings equal to the success number, the number of fish harvested which did reach the target size since the last restocking. The intervals between adding new fingerlings should not be more than 3–4 months. If the farmer knows of any losses or has seen any dead fish these should also be replaced.

If the farmer manages his pond as he said he would and if the estimated productivity was realistic, the results of this system should be as follows. Once the largest fish have reached the size at which he crops them (the target size) and he starts to remove them, other fish should reach target size at about the montly rate. They are harvested and recorded as part of the success total. The farmer then replaces them with fingerlings. A kind of steady state is produced in which the pond holds fish of all sizes but the number of fish in the pond remains roughly constant and so does the total fish biomass.

Let us assume now that the estimate of the pond's productivity was too high. Perhaps the farmer fails to keep to the schedule of feeding or fertilization which he planned. Perhaps his soil fertility is below average. Or maybe the assumptions on which the productivity of this type of fish culture was estimated were simply very over-optimistic. In any of these cases fish will not reach the target size at the predicted rate. The farmer should continue to crop the same number of fish each month, but some of them will be smaller than he planned. These will not be included in the success total. Therefore they will not be replaced next time the farmer receives a new batch of fingerlings. Hence the total number of fish in the pond will decline. If the stocking rate was only a little too high then this slight drop in the stocking rate may be enough; the remaining fish should grow faster so that the right number of fish reach the target size each month. However, if this does not happen and the farmer continues to harvest fish (according to the monthly cropping rate) which are smaller than the target size, he could eventually end up with no fish in the pond. Obviously this is not right. So if the number of fish has dropped by 25 percent and still fish are not reaching target size at the monthly cropping rate, then one of these numbers must be reduced. In the case of our example, if the pond now holds only 45 fish instead of 60 and the farmer finds that he cannot harvest 5 fish of about 250 g each every month, he may decide to lower his monthly cropping rate to 4 or 3 and see if the fish now reach a size at which he is satisfied to harvest them. From then on he takes out this lower number of fish, and as before only those which reach the target size are replaced. This process may have to be repeated several times until a steady state is reached. Alternatively, the farmer may decide to improve his management by putting in more manure or better and more regular feeding, to see if this helps. Or he may simply decide that he doesn't mind eating smaller fish. Then he lowers his target size and replaces all those fish which he now considers acceptable. (It doesn't matter if the farmer's target size is not exactly what he said it would be when he discussed it the first time he collected fingerlings. It also doesn't matter that he doesn't know how many grams it corresponds to. All that is required is that a steady state can be reached in which he is taking out and replacing fish at a rate which utilizes the productive capacity of his pond efficiently.)

If the pond is producing fish which reach a larger size than expected, or more of them than expected reach target size each month, then the farmer can enjoy this situation without making any change in his accounting. He can crop fish at above the monthly rate, and put this bigger number into his success total. They are then replaced when he gets new fingerlings. Or he continues to crop and replace only at the monthly rate and effectively increases his target size.

From time to time the farmer should check that the total number of fish in his pond is in agreement with his calculation. This can most easily be done when he is lowering the water level to harvest big fish. In this case he will have to lower it further to catch and count also the small fish. He had better start this work early in the morning while the weather is cool, and preferably during a rainy period when there is plenty of water available to refill the pond quickly.

All this may seem complicated. However it is a way to get around the lack of production data and away from the situation where an arbitrary stocking figure is recommended to everyone, without taking any account of different levels of fertility and management. With time and experience it should become easier. Also estimates of productivity should become closer to the truth.

In addition it would be very useful indeed if certain farmers could be selected for closer follow-up, so that their fish could actually be weighed on harvesting. If fish could be weighed on stocking, harvesting weights and dates recorded, and some records kept of the approximate amounts of fertilizer, manure or feed supplied, this would be a great step forward.

5.2 Feeding

It is better to feed small quantities often rather than large amounts infrequently (since this will just lead to waste). If fish are fed regularly at one place they soon learn to come and eat there. The grower can then get a good idea of the condition and appetite of his fish by feeding them slowly, and can stop when they seem to have had enough.

In very hot climates carp are often said to lose their appetites during the middle of the day and only to feed in the morning or evening. However at Aiyura they have often been seen eating actively at demand feeders during the afternoon.

It is not recommended that crops be grown specially for feeding to fish. It would be better for the land and effort to be used on growing food which can be eaten by people directly. However, spoiled crops, kitchen scraps and leftovers can be given to the fish; it should soon be possible to see what they will take and what they leave. The grower can learn a lot about his fish by spending some time each day beside the pond feeding them.

A good way to present granular or powdery foods to the fish and reduce wastage is by the use of a plastic bag demand feeder. This simply consists of a plastic bag hanging from a string into the water in some part of the pond where the fish are not afraid to come, i.e., where the water is reasonably deep and there are not too many people passing to and fro. The bag has a few holes of about 0.5 – 1 cm diameter cut or burned into the lower part and the food placed inside, so that it does not run out into the water automatically but can be got out little by little by the fish nudging the bag or sucking at the holes. This the carp soon learn to do. Such a demand feeder is good for such things as spoiled or leftover food from the house (cooked or raw), mill screenings, grated coconut or other oil-cakes, leftover poultry feeds and so on. A quantity which can be finished in about 24 hours is suitable and any leftovers are then washed out before re-filling, so that the feed inside the bag does not become smelly.

If the mud in the area of the pond where the fish are generally fed becomes very dirty and starts to stink, either of rotten eggs or of decayed food, this is a sign that a lot of the food being given is not being eaten. Perhaps it is unsuitable, it is being given too quickly, or the fish are simply being given too much.

In the future, if Chinese carp are introduced and become a part of local pond culture, it will be worth collecting soft grasses and plants to feed to them directly. Grass carp can chew up huge quantities of vegetation but do not digest all of what they eat. Their faeces are then eaten by other fish in the system, as well as fertilizing the water. The utilization of some plants can then be increased considerably by composting them for a month or so with about half their weight of animal manure or with 1–2 percent urea plus 1 percent quicklime. Common carp does not take grass directly as food. However a similar composting system may be used to “stretch” a limited quantity of animal manure. In this case one would wait until most of the grass had decomposed, which would take longer, 2–3 months for reasonably soft grasses.

5.3 Manuring in Unfed Ponds

(i) Experimental and theoretical background

Manuring rates in fed ponds generally range from 50–250 kg dry material/ha/week. However, there are a number of reports from different parts of the world of very high yields being obtained from fishponds entirely or mainly from the application of manures at very much higher rates.

In Israel, up to 500 kg dry material/ha/week was applied in the form of liquid cow manure or sewage, plus inorganic fertilizers. Ponds, stocked with common carp, silver carp and tilapia in the ratio 3:1:2 yielded 20–40 kg/ha/day (equivalent to about 7 tons/ha/yr) without supplementary feeding.

In the USA, Buck et al., (1976) stocked fish in small ponds receiving the manure from pigs, at a rate equivalent to 39 and 66 pigs/ha while the pigs were being grown from about 30–100 kg each. Although they do not give manure production rates, pigs of this size would be fed 3–4.3 percent bodyweight of feed (dry weight) daily and about 80 percent of the nutrients would be excreted again (Morrison, 1948). Thus these ponds were receiving 300–1 100 kg dry weight of manure/ha/week. They were stocked with nearly 4 300 silver carp (9 g each), 1 700 Israeli carp (80 g each), 430 bigheads (45 g each) and 130 grass carp (65 g each) per hectare plus 700 large (150 g) silver carp/ha and some American species in small numbers. Most fish reached 0.5–2 kg average weight in 170 days, giving a production rate of 3–4 tons/ha in less than half a year. Fish growth was better with the higher number of pigs per unit area.

Moav et al., (1977) report on 2 years' research in Israel using liquid cow manure in small (400 m2) stagnant water ponds without aeration (except during one week when oxygen levels became very low). Two different stocking densities were used, some ponds received pellets, some were fed with grain and some were unfed. Manuring was applied 6 days a week in quantities which increased with the fish biomass, from 180 to 1 260 kg dry matter/ha/ week. In the second year the low stocking rate (per hectare) was 4 800 (30 g) common carp, 520 big (460 g) silver carp plus 1 000 small (45 g) silver carp, 1 650 big (100 g) tilapia and 850 small (20 g) tilapia and 850 (200 g) grass carp, while the high stocking was almost double this. In addition to the liquid manure all ponds received 50 kg/ha each of ammonium sulphate and superphosphate every second week, and 150 kg/ha chicken manure on alternate weeks. In 126 days of growth the unfed ponds both produced biomass increases of slightly over 30 kg/ha/day (equivalent to 11.7 tons/ ha/yr). Individual growth rates at the low stocking density were also very high, with carp gaining 4 g/day (to reach 550 g each), grass and silver carp gaining over 4 g/day (reaching over 700 g each) and tilapia growing about 2–2½ g per day to reach 250 – 450 g each by the end of the trial. At the high density individual growth rates were suppressed, particularly those of common carp which only reached 260 g average in weight. (In the high density ponds periodic partial harvesting or thinning was performed several times, removing some 40 percent of the total mass before the final harvest. Most of the fish thinned out were silver carp and tilapia, but it was particularly common carp which suffered from overcrowding. Therefore this did not help much to relieve the carp growth drop in these ponds.) Fish were sampled every 2 weeks by partial seining. Carp in particular responded to doubling of the manuring rate (180 kg/ha/week was clearly insufficient).

Fish biomass on stocking at the low rate was 700 kg/ha and on harvest 4.7 tons/ha. Daily biomass gains in this pond remained practically constant at 40 kg/ha/day for the first 56 days and then started dropping, suggesting that biomass should not exceed about 3 tons/ha for optimum production, with this combination of fish. It appears that about 3.8 kg of dry material (from liquid cow manure and poultry pellets) were required to produce 1 kg of fish. Common carp produced 55 percent of the biomass growth. Total yields might have been higher if silver carp had formed a greater fraction of the fish stocked.

This is an example of an experiment which, because it was carried out carefully and well, gives a great amount of information. In particular all inputs into the pond were weighed and recorded, fish growth was sampled regularly for each size group and species, and fish of each group were counted and weighed on harvesting. Thus we can learn about individual fish growth rates and survival rates, biomass growth, optimum levels of biomass, and conversion coefficients.

In Thailand, Edwards (1985) reports on trials with tilapia (O. niloticus) in 200 m2 ponds which were integrated with 30 laying ducks each. He does not report how much the ducks were fed. Williamson and Payne suggest that an average daily consumption would be about 170 g (dry matter) of which about 70 percent would be excreted, i.e., a manuring rate of 1 250 kg/ha/week. Gross fish yields are said to have been equivalent to 5.5–14.5 tons/ha/year. However, egg production was low and failed to pay for the cost of duck feed, so the system was not economically attractive.

In their excellent book on Chinese fish culture, Zhong Lin et al., (1980) discuss rates of application of animal manure, green fodder and fine feeds in relation to fish yields. Green fodder (grasses, vegetable wastes and water plants) is initially eaten by the grass carp which can chew up its own body weight of grass per day, but a large amount is excreted only partly digested. It therefore acts partly as a feed for grass carp, partly as manure. Fine feeds included oilseed cakes, brans, sauce and wine brewing residues and silkworm pupae, but in the examples given did not exceed 2 percent of the total input. The most important fish in the polyculture ponds were grass carp, silver carp, black carp, crucian carp, tilapia (O. mossambicus), snakeheads, etc. Production rates were equivalent to 4.5 – 8 tons/ha. The highest rate was obtained with inputs of 235 tons/ha of green fodder and 8.5 tons of animal manure (no fine feeds). 6.1 tons/ha were obtained virtually on green fodder alone (160 tons/ha). The highest application of animal manure was 69 tons/ha which, with 76 tons of green fodder plus 0.12 tons/ha of fine feeds gave a fish yield of 5 tons/ha. Feed coefficients (wet feed : wet fish) are from 20:1 to 30:1, i.e., 20–30 tons of green fodder, fine feed and manure applied for each ton of fish produced. Individual coefficients are given for several grasses and aquatic weeds (35:1 – 100:1), for silkworm pupae (1.8:1), bean cake (4:1) and rice bran (6.8:1). (These conversion rates would be much poorer without the different elements of the polyculture system.) Assuming that green fodder and animal manures both average about 25 percent dry matter, then rates of application were up to 60 tons/ha/year dry weight, or about 1 200 kg/ha/week.

If similar systems are to be applied in the PNG highlands, a number of points have to be taken into consideration. There is a shortage of cow and poultry manure, and great reluctance to use pig manure in food production. (This situation may well change as agriculture intensifies.) Chinese carp are not available. They play key roles in the utilization of vegetable matter (grass carp) and plankton (silver carp and bighead). To some extent tilapias, grey mullets and milkfish may serve as partial substitutes; this will have to be looked at experimentally. In Edwards' system, based entirely on tilapia (O. niloticus) figures are not given but no doubt many of the fish were very small. Temperatures in PNG highlands are generally lower than they would be Thailand or for much of the growing season in Israel, China and USA. Most important of all, the skill and experience in pond management are lacking in PNG, both among farmers and among government staff. This is something that can only be developed with years of serious work.

High rates of manuring can only be applied to ponds in which the water is essentially stagnant, otherwise the fertility is simply washed out. A delicate balance has to be struck; too little manure and production rates will drop, too much and lack of oxygen will become a problem, particularly on warm nights without wind. Chinese farmers know how to judge the seriousness of oxygen shortage by the surfacing behaviour of the different species at different times of the night. Surfacing of bighead or silver carp is an early sign of lack of oxygen; common carp are more resistant. Surfacing around dawn is less serious than surfacing which starts already at or before midnight. If fish do not go downwards from the surface when frightened the situation is already dangerous.

Schroeder (1975) looked at the oxygen balance in heavily manured ponds, and found a complex situation. The oxygen consumed in bacterial decay of manure may be more than offset by the increase in photosynthesis during the day, but manure will contribute to the plankton's biological oxygen demand (BOD) in the pond water, which may be of importance at night. The main factors were apparently the water BOD, fish respiration and oxygen transport across the water-air surface. For still water, oxygen transport was about 0.3 g oxygen/hr/m2 per 0.2 atm oxygen excess or deficit. The oxygen consumption due to the respiration of a single fish between 20–30°C was estimated using the formula oxygen consumption (mg/hour) = (fish weight in grams)0.82. This means that, roughly speaking, the respiration of about 6 tons/ha of large fish is the limit of what gaseous diffusion alone can support when there is no wind.

(ii) Practical suggestions

Manuring rates of up to 600 kg dry matter/ha/week should be tried. This would be equivalent to about 25–30 kg wet manure/100 m2/week. Clearly it would be safer to apply such large amounts on a daily basis (i.e., 4–5 kg/day) and stop for a while and then reduce the dose if there are signs that it is excessive (fish gasping, bad-smelling black mud developing). Of course, growth rates at these high manuring levels should be compared with those with more moderate manuring.

Fish biomass should not be allowed to exceed about 1–2 tons/ha (10–20 kg/100 m2) in intensively-manured growing ponds. If higher yields are aimed for, they should be based on periodic thinning to keep the biomass at or below this level.

Previous Page Top of Page Next Page