Appendix 3

TRAINING PROGRAMME IN CARP BREEDING

1. INTRODUCTION

This training programme is intended to assist the people concerned with practical carp breeding work. It is aimed at scientific officers and technicians, and those who may in future be involved in training farmers and extension workers.

Techniques and quantities given are based on work carried out elsewhere, in particular in China, Israel, Hungary and Indonesia. They can only serve as starting points for a process of trial and error aimed at developing systems which work well under local conditions. Several alternatives are usually given to enable the user to apply those which can be realized under the existing conditions, with limitations imposed by the availability of gear, chemicals, fertilizers, etc. Trials and experience will show which methods are the best applicable.

2. BROODSTOCK

2.1 Broodstock requirements

A hatchery should have about 100 kg of female brood fish per hectare of fry rearing ponds. These fish should be over about 1 and 1/2 years of age and ¾ kg in weight, although larger fish are generally preferred. About 50–70 kg of males per hectare of fry rearing ponds should be sufficient. The average size of the males may be a little smaller. It is useful to have a range of sizes of males, from a few hundred grams up to about 70 percent of the size of the large females.

2.2 Choice of strains

At present there is very little choice available. Hatcheries have to use the carp they can get hold of, whether from rivers, other hatcheries or farmers' ponds. Most of the carp which were previously introduced into the country were described as ‘golden’ or ‘Cantonese’. All the fish seen in Aiyura are fully scaled. The majority are golden in colour and rather longbodied, although some are brown or greyish and some a little deeper in shane. A few of the larger fish have eye openings which are very small so that the fish appear almost blind or eyeless, although the size of the eyeball underneath the opening seems to be normal. (This is a characteristic of an Indonesian strain known as Sinyonya which is quite well thought of there.) Behaviourally, the fish are quick and active when frightened and consequently quite difficult to catch in a net. When calm they like swimming at the surface in groups. They learn quickly to find and use a demand feeder. There is almost no data concerning their growth potential, food conversion efficiency or tolerance of different environmental conditions.

Future importation of common carp strains, subject of course to government approval, should take highland conditions into account. Pond water is generally between 20°C and 27°C, although temperatures of over 30°C and below 15°C have been found in shallow ponds. Farmers' ponds are likely to rely on manuring and feeding starchy wastes rather than high protein feeds.

More placid strains would be easier to handle than those now present. Colour and shape preferences are not known, but dark fish should be less vulnerable to predation by birds.

There is no indication that carp at Aiyura are inbred. Where they fail to grow it is due to poor husbandry, i.e. lack of natural or supplied food.

2.3 Ripening conditions for spawners

Once fish which are intended for controlled breeding have reached sexual maturity the females must be separated from the males. Fish of over 1–1½ years of age and about 400 g which do not produce at least a little whitish milt (sperm) when the abdomen is squeezed several times toward the genital pore can be considered as females. Fish which do produce milt are clearly males. Fish which are slightly younger than this must not be allowed into the pond where the females are being ripened as they may cause wild or uncontrolled breeding to take place. It will then be hard to find ripe females when they are required. (A few females getting into the pond where the males are kept will do little harm.) Males which have grown under poor conditions or in the wild can be mature at a very small size of only a few tens of grams. The female ripening pond must be screened to prevent the entrance of such fish either from the water inlet or the water outlet.

It is usually easy to find ripe males and they recover more quickly after spawning. Therefore their ripening conditions are less critical. If there is a shortage of pond space hatchery staff can keep males in fingerling ponds, in ponds for the production of fish for consumption or in small holding ponds at high density with feeding. Females can be ripened at various densities, but their growth conditions must be good if they are to ripen quickly after spawning. Low density ripening means not more than about 1 000 kg/ha. At this stocking rate it should be sufficient to manure the pond and rely mainly on natural food developing. Some supplementary feeding of starchy foods or wastes may help, but this should not be more than about 2–3 percent of the total fish body weight per day, after correcting for the water and fibre content of feeds if necessary.

Medium density means up to 1 500 or 2 000 kg/ha. Under these conditions the pond should be manured and a supplementary feed containing about 20 percent protein should be given at a daily rate of about 2–3 percent of the fish body weight. Feeds should be given either by hand twice daily or using a demand feeder (see fig. 1), and should be served in the same place of the pond so that fish can gather there to eat.

High density ripening, at over 2 000 kg/ha, requires a balanced diet containing the right fats, minerals and vitamins as well as a high level of protein (30 percent or more). It may be possible to experiment (with a small part of the broodstock) using a commercial trout-growing diet. This will be expensive, but it will save pond space if successful.

To be able to check the condition of spawners and select ripe ones, the ponds should be easy to drain or to net. This should be done while water is not too warm, i.e. in the early morning. Handling must be gentle and kept to a minimum. Knotless netting causes less skin damage than the knotted variety. If fish are transferred from pond to pond a plastic bag partly-filled with clean water is the most convenient way. Avoid sudden temperature changes of more than 2–3°C. Fish are less stressed by handling if they have not recently been fed, and it is easier to judge their condition.

Fish which have already been stressed by netting should not be left in very shallow muddy water where temperature will fluctuate greatly in the following day and night. Broodstock ponds are more conveniently managed when small, say 500–1 000 m². A pair of ponds may be used for one group of females which are then transferred from the pond which is being drained to one which has previously been filled, each time a selection is made. The pond from which they have been removed may then be dried and prepared for rearing the next batch of fry.

2.4 Selecting ripe fish for spawning

Female carp are judged to be ripe mainly from the condition of the belly. It should be large and swollen at a time when they are known not to have recently had a heavy feed. The outlines of the ovaries should be seen from above and from the side. From below the rear of the body looks rather square in cross-section. The belly should be soft, but elastic and full, to distinguish it from a possibly still swollen but empty belly of a fish which has just spawned. The anus or genital opening may be of red colour and protruding. Gentle pressure may release clear yellow fluid and one or two yellow eggs of about 1 mm in diameter.

Ripe males release sperm easily with a very gentle stroking or pressure on the belly. This should be white or yellow and creamy.

Fish selected for spawning should appear healthy and free of signs of disease or parasites. If certain characteristics are wanted from the fry, such as a particular colour, shape or temperament this may influence the choice too. Thus there has often been selection in favour of fish with a small head and broad or high body which will have a higher proportion of edible flesh and less waste. Dark coloured fish may be more difficult for birds and other two-legged predators to see; and fish which show calm, placid behaviour are easier to catch and handle. They are particuarly favoured in Indonesia where the farmers almost never use seine nets.

3. SYSTEMS OF CARP SPAWNING AND HATCHING

A wide range of systems for breeding carp are in use in different parts of the world. They vary in their yields and in their difficulty. The systems described first are simple but give little protection to the eggs and larvae. The later systems generally give higher yields per kilogram of broodstock. However they require more skill, more work and more equipment. In each case one must also consider how to look after the larvae when they are ready to start feeding (two or three days after hatching) and what treatment to apply to the pond in which the fry are to be reared. Some of these treatments should be started before the breeding activity begins, so it is important to be prepared. Examples of schedules of activity are shown in Appendix 3, section 6.

3.1 Uncontrolled breeding, without separating spawners and fry

Mature male and female carp are put into a pond, which should have some areas of grass or weeds growing inside. The fish spawn and fertilize their eggs in the grassy areas and the eggs stick to the grass. They hatch, and the fry grow in the same pond. The pond is checked and sampled for fry (with a hand net or small seine net) from time to time. When fry or fingerlings are of the size required they are collected by seining, perhaps first partly lowering the pond water level through a fine screen:

There is no control over when the fish spawn. As the fish breed when they become ripe, at different times, there will be a range of sizes in the pond. Breeders and fingerlings will eat some of the eggs and fry, and they will compete with each other for food. This cannot be prevented but it can be reduced by keeping the total weight of fish in the pond low, by fertilizing and manuring the pond (Appendix 3, section 5.2) and by feeding. Fingerlings should be harvested regularly using a net which leaves the small fry behind.

Not more than about 100 kg/ha of females and a similar total weight of males should be stocked. (The average weight of the males may be lower than that of the females, in which case there will be a larger number of them.) Select breeders according to the characteristics of colour and shape required from the fry, but try to avoid using fish which are all too closely related.

Only uncontrolled breeding has been used until now in PNG, both at government hatcheries and in some farmers' ponds. Although the number of fry produced per spawner or per hectare will not be very high it involves little work and it does not matter if a few mistakes are made in sexing broodstock. Ponds do not have to be emptied and refilled frequently which may be difficult during dry periods. If there is spare pond space and plenty of broodstock use this system in addition to more controlled methods.

3.2 Controlled Breeding

The ripening of a female carp involves development of the egg yolks and increase in size of the ovaries. The developed eggs will remain inside the ovaries until certain factors in the environment cause the female to produce hormones which lead to ovulation and spawning. Chief among these factors are the attention of a male which follows her and nudges her belly and the presence of suitable surfaces for the eggs to stick to, such as grass, leaves or fibres. Water temperature and dissolved oxygen content have to be suitable. Clear water, the absence of predators and low fish population are also helpful. If these spawning conditions do not occur, eventually the eggs will degenerate and be reabsorbed by the female.

The key for exercising control over the breeding of carp is to allow the females to ripen without allowing them to spawn until it is required. In particular, do not allow males to come into the pond where the females are being ripened. If mature males do get in, the chances are that whenever a female becomes ripe she will immediately spawn. It will not be possible to find ripe females which can be used in any of the systems of controlled breeding described below.

3.2.1 Spawning in a pond, followed by separation of spawners and eggs

In these systems ripe females and males have to be selected just before spawning is required and introduced into a spawning pond. The stocking density of fish in the spawning pond may be quite high (up to about 1 kg of breeders per 5 m² of pond).

The following conditions all help to encourage successful spawning: the spawning pond has been dried and only recently part-filled, the water level in the spawning pond is still rising when the breeders are introduced, the water in the spawning pond is slightly warmer than that in the pond from which the fish have been removed but not excessively hot (i.e. not over about 28°C), the incoming water is fairly clear and well-oxygenated, there is plenty of vegetation or collecting material suitable for the eggs to be deposited on. Fish must be handled gently, not dropped or rolled around inside a net full of mud and transferred when water temperatures are not too high in a container of good quality water.

Incubation and hatching should also take place in clear well oxygenated water. Exclusion of organisms larger than about 0.5 mm, such as insects, copepods, tadpoles and fish which might prey on or damage the eggs is beneficial. Water temperatures should remain within the range of 19–29°C and be relatively stable, not fluctuating by more than four or five degrees from day to night. Since surface water layers can be heated by the sun or cooled by heavy rain it is a good idea to arrange the materials to which the eggs are attached so that they float a few centimetres below the surface. About 2 m² of kakabans (Arenga pinnata, palm fibres held between strips of wood or bamboo), pine or casuarina branches, or about 2 kg of grass or aquatic weeds should suffice for the eggs from 1 kg female spawners. There are also artificial egg-eollecting substrates of nylon or polythylene marketed commercially, mainly for breeding aquarium fish.

The following are the three common systems in use:

1. Dubisch system - eggs on grass; spawners removed The spawning pond should have grass growing on its floor, which should be covered when the pond is full. There may also be an internal drainage ditch. In the morning after spawning eggs will be seen on the grass. The pond water level is then quickly lowered, the fish removed and the pond refilled for hatching to take place inside the spawning pond. The difficulty with this system, particularly during hot dry weather, is that some of the eggs may be out of the water for too long and die.

2. Sunda system - eggs on collection material; eggs removed The spawning pond should not have any grass or weeds growing inside. Instead egg collecting material is placed inside the pond. After spawning has occurred the egg-collecting material with the eggs attached is transferred, in the cool of the morning, to a prepared pond where hatching is to take place. The male and female fish may then be caught and examined at leisure. If some females have not yet spawned the process can be repeated for another night. Otherwise they can be returned to ponds for ripening again.

3. Cimindi system - eggs on collection material; spawners removed The small spawning pond here is often built inside the corner of a larger pond which is to be used for fry rearing. It should be free of grass and vegetation and have its own water supply. Egg collectors are introduced as in the Sunda system. After spawning the water level is lowered but the egg collectors remain immersed. The male and female fish are removed to ripening ponds and the small pond refilled for hatching to take place. Meanwhile the fry-rearing pond is fertilized, manured, treated with insecticide etc. While the eggs are hatching. As the pond in the corner has its own water supply the eggs and larvae remain in relatively clean, cool, flowing water. (The overflow from the small to the larger pond should be filtered and not too fast, i.e. not more than about 1 l/sec). Two or three days after hatching a connexion is opened between the two ponds and the larvae allowed to enter the larger pond as they spread out looking for food.

3.2.2 Spawning and hatching using tanks or hapas

Tanks may be of concrete, fibreglass, plastic, or wood. They may be indoors or outdoors. They should generally be at least 2 m² in area and have a minimum water depth of about 70 cm for spawning. They may have to be covered to prevent fish from jumping out.

Hapas are bags or cages hung in water inside a tank or pond, made of an artificial fibre which is resistant to rotting, such as nylon or polyester, with pores of ½–1 mm size (i.e. something in between mosquito netting, plankton net and curtain lining cloth or tuile). Their minimum dimensions are like those of the tanks, and if used for spawning they too should have a cover (see fig. 5).

Three systems can be used:

(i) Spawning and hatching in a tank

Broodstock and egg collecting material are introduced into the tank which was previously cleaned, dried, and freshly filled. It should have a good flow (more than 10 l/min/kg fish) of clear, well-oxygenated water. Fish density should not be more than about 1 kg/m² and egg-collecting material, as before, about 1–3 m²/kg female or 2 kg/kg female.

After spawning the brood fish are netted and removed and incubation takes place inside the tank, with a somewhat reduced water flow and screened outlet. If the tank is outdoors care should be taken to protect eggs against big changes of temperature which may occur very near to the surface.

When it is time to transfer the larvae to the fry-rearing pond they can be collected by lowering the water level in the tank gradually through a fine screen of cloth similar to that used for making hapas, and then gently netting the fish with hand-nets of the same material. Alternatively the tank contents may be drained into a basket lined with this material outside the tank where the larvae will be retained.

This is probably the method which will be found most convenient at Aiyura if suitable concrete tanks are constructed near to the ponds (see section 4.7 and figs. 2 and 3).

(ii) Spawning in a pond or tank, hatching in a hapa

Spawning is carried out in a pond or in a tank. The material with the eggs attached is then moved to a hapa, which may be inside a tank or in a pond. A hapa of 2 m² (see fig. 5) can accommodate eggs from a good spawning of about 2 kg of females. A small (screened) flow may be introduced directly into the hapa, or water can enter simply by passing through the cloth.

A day or two after hatching the kakabans or other egg-collection materials should be removed carefully from the hapa. When it is time to transfer the larvae to the fry-rearing pond they can easily be concentrated in a corner of the hapa and scooped out with a bowl into a large basin or bucket.

(iii) Spawning and hatching in a hapa

This system may be used where no tanks or small ponds are available, by arranging a hapa in any pond with good water conditions and introducing breeders and egg-collecting materials. The hapa should be covered to prevent the fish escaping. After spawning the males and females are easily caught inside the hapa and returned to their ripening ponds, while the eggs on their supporting material remain inside the hapa for hatching in the usual way. The wear and tear on hapas used this way is likely to be greater than if they are only used for hatching, and a hapa of 2 m² should not be stocked with more than 1–2 kg of females and a similar weight of males.

3.2.3 Induction, with or without stripping

The female which is fully ripe will usually spawn if placed under the right environmental conditions and stimulated by active males. This causes her pituitary gland (or hypophysis) and other glands to release the necessary hormones inside her body. Alternatively it is possible to get these hormones from somewhere else, generally the ground-up pituitaries of other carp which have been killed, and inject them. This is called induction. A certain time after receiving the right dose of hormone the female will have to spawn, whether males are there or not. She can be taken out of the water at this time and the eggs squeezed out of her. They should now run out quite freely, as they will have been released from the ovaries (ovulated). This is called stripping.

The female must be ripe before she can be induced. Quite often a small dose of hormone which is not enough to cause induction is given some time before the main dose, just to help make sure she is fully ripe. If induction is attempted with a fish which is not ready it will fail and the female may die.

(i) Induction without stripping

Ripe females (and males if necessary) are injected with hormone as described in Section ii below. They are then placed together in tanks, ponds or hapas for spawning and fertilization to take place in one of the ways described in section 3.2 above. Doses should be timed so that spawning occurs at night, as it would normally without induction. Therefore the first injection is given in the early morning and the second in the early afternoon.

(ii) Induction with stripping

Selected fish may be disinfected in 40 ppm formal in for 2–4 hours and kept unfed in hatchery tanks for 24 hours. Females are injected with fresh or preserved hypophysis according to the schedule below. They are held, without males and preferably separately, in small containers from which they cannot jump out but where they can be closely watched. When some eggs can be felt by running a finger across the walls or floor of the tank, the female is ready. She is removed from the water, quickly dried, and the eggs stripped either into a dry container (or into fertilizing solution which also inhibits the development of the stickiness). Eggs are fertilized by mixing with milt which is stripped directly when needed or stored (cool) for up to several hours. After washing, the eggs are placed in some kind of running-water incubator to hatch.

Pituitary glands may be preserved in alcohol. After being removed from donors, they are immediately dipped in absolute alcohol. This absolute alcohol is changed after about an hour and again after 24 hours. They may than be stored at room temperature in simple sealed bottles. Alternatively, acetone can be used, changed a couple of times in the first 8 hours and after 24 hours poured off and glands allowed to dry by evaporation. They may then be stored in sealed bottles in a dessicator at room temperature. Dried carp hypophysis is also commercially available.

For inducing spawning a standard dose is taken to be 3 mg dried carp pituitary per kilogram of injected female. It is divided into two portions. The first portion contains 10 percent of the material or about 0.3 mg/kg. This is administered 8 hours prior to the second or main injection which contains the remaining 2.7 kg/kg. Females are ready to strip 10–14 hours after the second injection. It is therefore convenient to give the first injection around midday or early afternoon and the second late that night. The fish should then be ready the next morning.

Males receive a half dose (i.e. about 1.5 mg/kg) at the time females receive their second injection. Dried pituitaries are usually used, homogenized but not centrifuged or filtered. It is hard to be precise with such small quantities of material. As a general rule, if in doubt be mean with the first injection and generous with the second. Although various anaesthetics can be used to pacify the females for stripping, this is not usually necessary. The fish can simply be netted from a small holding tank and a towel quickly placed over the head of the fish. Covering the eyes calms the fish and this helps the operator to hold her firmly without causing damage. Another towel is used to dry the rest of the body. During this time one finger is kept over the genital pore of the female to prevent discharge of eggs. She is then vigorously stripped into a dry plastic bowl. A similar procedure is now carried out with one or two males. Milt may be allowed to flow directly into the dry eggs or may first be collected in a dry hyperdermic syringe (without needle). Eggs and milt are then stirred together with a plastic spatula, a feather or a soft paint-brush.

When stripping and fertilizing great care must be taken that eggs do not come into contact with fresh water, as this would cause them to become sticky and form clumps. The adhesive layer is removed from the eggs as follows: eggs from one female (2–8 kg) are stripped into a dry plastic bowl. A few millilitres of clean milt from one or two males are mixed dry with the eggs, and after a minute or two a few litres of weak solution of 4 g/l NaCl + 3 g/l urea in water are added. This prolongs sperm viability. After about 4 minutes of gentle mixing and swirling, the liquid is poured off and about 5–10 litres of strong solution (4 g/l NaCl + 20 g/l urea, in water) are added. After mixing well, the eggs are left to stand in this solution for 30–90 minutes, changing the liquid for more, fresh, strong solution 2–4 times during this time, and stirring the eggs occasionally. When no tendency to clump remains, the liquid is discarded and tannin solution (1 g/l tannic acid in water) is poured onto the eggs to harden them. They are stirred at once, and then the liquid poured off within about 15 seconds. The eggs are immediately washed repeatedly with good clean, chlorine-free water for about three minutes, and then are measured into the incubators.

When eggs have finished absorbing water and swelling, about 100–300 eggs will have a volume of one millilitre. It is a good idea to count several one millilitre samples.

Commonly used incubators in Israel are conical 40–80 litre funnels made of clear plastic (PVC or acrylic) with an angle of about 36° and water inlet at the apex. The outlet is near the top and water flows down a tube into a larval tank with a screened outlet. Flow of water is carefully adjusted so that eggs are in continuous slow movement, but are not carried up to the outlet. When hatching occurs, the larvae swim upward, are carried out of the hatchery with the overflow and arrive in the larval holding tank which is usually aerated (Fig. 4). Such an incubator can hold about 300 000 eggs.

Alternative incubators consist of cones made of a porous cloth (nylon, polyester, etc.) or plankton netting (pore size around 200 microns, .2 mm). A way has to be found to attach the bottom of this cloth cone to a water supply and suspend it from the top so that it is immersed in a tank. Such hanging cloth cones were used in some Indonesian hatcheries for the eggs of the Javanese carp (Puntius javanicus) but seldom for common carp which generally spawned (without induction) onto kakabans using one of the systems described in sections 3.2.1 and 3.2.2. At the Aiyura hatchery where the pressure of the water supply is very variable this kind of cloth hatching cone was found to be better then plastic funnels as there is less danger of the eggs being washed out if water pressure increases. The apex angle should be about 36 degrees (which can be formed by sewing a 110 degree sector of a circle or slightly steeper. There should be no step or dead space at the point of attachment to the tube. A cloth cone can hold about 1 000 eggs per one litre of cone volume.

4. FRY REARING DURING THE FIRST FEW DAYS

For about three days after hatching larvae live on the reserve provided in the yolk-sac and do not feed. During this period they can easily be held at quite high density of 10–50 000 larvae/m³ in either a hapa or a tank. The hapa must be placed in good, clean and well-oxygenated water of a suitable temperature (20–28°C). It may help to introduce a small flow (5–10 l/min) of good water directly into the hapa, and aerate it. A tank should have aeration and, preferably, a water flow of 2–5 l/min/m³. This means it will have to have a screened water outlet which is covered with hapa cloth or plankton netting of mesh size about 0.5 mm. This has to have quite a large surface area (200–500 cm²) to prevent it from becoming blocked and causing the tank to overflow. These screens have to be cleaned regularly.

When the end of this stage is reached larvae have to start feeding. Their natural first food would be zooplankton organisms smaller than 250 microns, such as rotifers and perhaps crustacean nauplii. The best growth is obtained by releasing larvae at this stage into a suitably prepared fry pond. This is the most widely used method for carp fry production, and it is what is recommended here.

A lot of research has been carried out to extend the period during which carp larvae or fry can be held in tanks at high density. This work has only had limited success. It enables larvae to spend a few extra days in tanks, but they then are anyway transferred outdoors to prepared fry ponds, where their main growth takes place. (This should be clearly understood. Carp fry are produced in ponds, not in hatchery tanks.) Some of the larval feeds used in extending the period in tanks are listed, for interest only.

1. Hard-boiled egg yolk dispersed in water: one hen's egg yolk for about 100 000 3-days old larvae per day, distributed near the water surface through a fine cloth or by spraying in two or three daily feeds.

2. Whole egg dispersed in water: the egg is beaten lightly, then about 200 ml boiling water poured in quickly while stipping. Feeding rate as above.

3. Soya bean extract: 50–100 g beans soaked, ground with water and the milky liquid pressed through fine cloth, per 100 000 larvae.

4. Newly-hatched artemia nauplii: nauplii from 10–20 g good cysts are sufficient for 100 000 newly-feeding larvae, but requirements increase rapidly. Feed 2–5 times daily till larvae have pink/orange gut contents but are not full to incapacitation. Nauplii should be well separated from empty cyst cases before feeding. Alternatively, cysts may be decapsulated before hatching. Cysts are decapsulated by aeration in contact with a solution of 17 g/l NaOCl plus 25 g/l NaOH (take care to avoid contact with eyes and skin) or diluted commercial bleach solution, taking care the temperature does not rise above 40°C, until the colour changes from brown to orange. They are then washed very thoroughly and hatched in the normal way, with vigorous aeration.

5. Specially formulated larval feed such as Ewos Larvstart C10 (58 percent crude protein): 40 g per day grade 00 for 100 000 newly feeding larvae rising to 70 g grade 0 plus grade 00 after four days and 100 g grade 0 after one week, given in 4–10 daily feeds.

6. Cultured rotifers: cultivation of rotifers in tanks or small ponds and feeding them for some days to larvae in tanks or hapas is claimed to give improved larval survival. Recipes for rotifer culture in tanks or small ponds often involve heavy manuring (200–800 g/m² dry poultry manure) followed a few days later by insecticide (0.5–2 ppm). Rotifers have to be harvested using plankton netting of about 50 microns. In addition to requiring space and quite a lot of extra labour, the main problem with these cultures is that they often fail when most needed.

Dry feeds contaminate the water and make careful control of flow rates and frequent clearing necessary. Live feeds require quite a lot of daily attention. All these methods add to labour and expense, they carry a high risk of total failure, and they generally do not give growth which is nearly as good as that possible on natural plankton in a well-prepared pond.

The recommended procedure at Aiyura is as follows. Two days after hatching starts a little hard-boiled egg yolk may be given to larvae in tanks, hapas or small earth ponds. The next day larvae should be stocked in the large well-prepared fry ponds. Efforts should not be made to extend the high density rearing period beyond this time. The main effort should go into preparing and managing the fry ponds as well as possible.

5. FRY REARING IN PONDS

5.1 Preparation of fry rearing pond before filling

Some or all of the following treatments may be required before the fry rearing pond is fertilized and filled with water.

1. Remove grass or weeds growing on pond floor, as they will make fry harvesting more difficult.

2. Repair erosion damage to banks. Some soils are very unstable and banks need a lot of attention. Banks may become undermined at the water line, particularly if large carp have been living in the pond. The hollow places should be refilled with clay or with mud from the pond floor. Erosion is often particularly bad by the sides of the monk. This is a common cause of leaking and may even lead to the sudden loss of all the water from a pond.

3. Mud-plaster the pond banks. With some soils this helps to reduce leaks and erosion, as well as slowing the growth of grass and weeds.

4. Dig over or plough the pond floor. As well as serving as weed control, this may discourage some pests. It will lower the subsequent oxygen demand of the mud by helping to oxidize reduced material. It may increase fertility by the oxidization of organic nitrogenous substances to nitrates. In some soils it promotes the production of mud and lowering of porosity. In some ponds, however, staff may feel that there is already too much mud.

5. Level and smooth the pond floor using a mud hoe or some kind of roller. This will make the harvesting easier and more complete, particularly if the period of rearing is to be short. It may further aid the production of mud and pond sealing.

6. Deepen or renew drainage channels inside the pond. Most commonly a channel runs from the water inlet diagonally across the pond to the outlet. Sometimes other internal channels are needed, especially if the pond is large or the floor has very little slope.

7. Screen the water inlet to prevent the entry of unwanted fish. They may predate, compete or carry disease.

8. Clean out the last water from the drainage channels inside the pond and with it any fish, tadpoles, etc., by pushing a plug of grass and mud down the channels a few times.

9. Screen the pond outlet immediately if there is any chance of fish entering from below.

10. Dry the pond floor for several days. This promotes oxidation and helps to control many pests, parasites and diseases. It may, however, increase porosity and is difficult to achieve during wet weather.

11. Remove or kill any frogspawn, tadpoles, predatory aquatic insects or fish remaining in puddles. Cement powder or quicklime (CaO) can be sprinkled in these puddles. Rotenone is effective against gill-breathing animals. Insecticides are also effective but are dangerous to handle (see Appendix 6). In all these cases, very small water volumes are being treated so high local concentrations of the chemicals can be obtained by sprinkling small amounts of powder or a few drops of liquid into each puddle. They all lose their potency rapidly and will become very diluted when the pond is filled. Five or ten kilograms of cement or lime are enough for a nearly-dry pond of 1 000 m² and will also serve to increase water hardness and raise pH. For a similar sized pond, 100–400 ml of insecticide should be enough. The smaller quantity should be ample for killing whatever lives in the puddles, the larger quantity is likely also to have an effect on the growth of plankton in the pond if it is filled soon afterwards.

5.2 Fertilization and manuring of fry ponds

The purpose of chemical fertilizers is to provide a supply of plant nutrients to the water so that the growth of phytoplankton can proceed. The chief nutrients are soluble inorganic forms of nitrogen, phosphorus and potassium, and common agricultural fertilizers may provide one of these nutrients, or a mixture of two or three. The strength of a particular fertilizer is described by giving the percentage of its weight which consists of nitrogen (as N), phosphate (as P₂O₅) and potash (as <₂O).

Standard fertilization rates for Israel grow-out ponds are ammonium sulphate (20–21 percent N) and super-phosphate (18–20 percent P₂O₅), each at 6 g/m² applied at two-weekly intervals. Fry ponds should receive this dose once each rearing cycle in addition to 100 g/m² of dry chicken manure. This is equivalent to about 1.2 g/m² each of nitrogen (as N) and of phosphorus (as P₂O₅).

To calculate dosages using other fertilizers the concentrations of nutrients in some of the commonly-used materials are as follows:

The phosphorus provided by some fertilizers can be absorbed by pond mud and then becomes unavailable for plankton to use. To reduce this it is advisable to place phosphorus-containing fertilizers on a small platform or submerged board near the pond water inlet.

With regards to organic manures, the situation is more complicated. They provide a slow release of nutrients as they are broken down by various bacteria. The manure, or the micro-organisms growing on it, may also provide a direct food source for zooplankton.

The following table gives some idea of the percentage nutrient content of various manures (from Morrison 1948).

As can be seen poultry manure is much richer in nutrients than that of cattle or pigs. It is often said to be superior in promoting the growth of rotifers (an important first food for fry) and generally better in growing ponds too. In practice, whatever is available should be used, with a target rate of application of at least 200 g/m² of cattle or pig manure or 100 g/m² of poultry manure, applied before the start of each rearing cycle.

The use of lime in ponds is often described. Quicklime can be used to sterilize the pond mud before filling. However, it is expensive and not widely available. Agricultural limestone is used to raise water pH and improve the response of acid waters to fertilization. It should not be necessary where the pH is always over about 6 and total hardness greater than 20 mg/l CaCO₃, as at Aiyura.

Chinese fish breeders emphasize the importance of the development of Protozoa and rotifers as shown by a greenish-brown water colour before larvae are stocked. They recommend the application of soft, quick-decaying plants or of grasses (which have previously been composted with animal manure and 1% quicklime until decayed) 3–10 days before stocking, at a rate of about 100–200 g/m³ of water. Undecayed parts of plants should be removed before larvae are stocked. (Inorganic fertilizers may then be added every day after stocking. Rates suggested are ammonium sulphate ¾ g/m² and superphosphate ¾–1½ g/m²). It would be worth comparing results from different manuring systems.

5.3 Other pond treatments

(i) Mineral oil

Mineral oil is often applied to the water of first-rearing ponds, either before or after stocking with larvae. This is to reduce as far as possible the numbers of air-breathing insects which may predate on the fry. Rothbard (1982) suggests a rate of application of either edible oil or diesel fuel of 24 ml/m², given first 2–3 hours before stocking and then repeated every second day until the fish have been in the pond one week. These treatments are to be given early in the morning when there is no wind, to enable the oil to spread over the water surface.

Kerosine does not always spread when poured directly onto the water, particularly if there is a lot of dust on the water surface. For this reason, it should be applied before any dusty food is given. It may be mixed with a small amount of petrol (gasoline) to make it spread more easily or it may be mixed with sand which is then scattered into the pond.

These methods are quite effective in killing the smaller air-breathing insects, such as the notonectids or back-swimmers. However, the larger insects are generally more resistant.

(ii) Insecticides

Insecticides may also be applied to first-rearing ponds either before stocking or during the time when the larvae are in the pond. Of course, in the latter case, the dosage used and the method of spreading the chemical are more critical because of the danger of reaching concentrations which are lethal to the fry in some parts of the pond. Doses of 1–2 ppm of organophosphate formulations such as Dipterex are often applied before the larvae are stocked in order to suppress the growth of crustacea (copepods and clodocerans) in the pond zooplankton. According to Woynarovich and Horvath (1980), some of the cyclopid copepods can cause very serious harm to small larvae by riding on the larva and damaging its skin and fins with their feet. One hundred cyclops/litre can kill 90–95 percent of fish larvae within a short time. Even without this direct attack Cladocera and Copepoda are too large to be consumed at first by young larvae and they compete with and possibly prey upon the rotifers which are considered to be the most suitable early food. So treating the pond with insecticide before stocking helps rotifers to become abundant during the early stages of fry rearing. Later, after a couple of weeks rearing, crustacea may become dominant in the pond plankton, but then they cannot harm the fry and are likely to be good food for them.

As Dipterex is available in PNG it is suggested that its use be tried, but that very great care be taken in handling it, following the manufacturer's safety instructions and the notes in Appendix 6. The average water depth is estimated by sampling at various points in the pond. This is multiplied by the pond area, to find the volume of water. (For example, if the average depth is 0.5 m and the area is 2 000 m² then the pond holds about 1 000 m³ of water.) One ppm of the formulated product should then be applied. This is equal to 1 ml insecticide/m³ water volume. In the example given above 1 000 ml would be the right dose for the pond. The newly-feeding larvae can be stocked one or two days later. There is no need to spray the insecticide onto the pond. It can simply be poured into the water near the inlet and left to disperse. As experience is gained in rearing fry, with and without the use of insecticide, it should become possible to see whether its use is helpful or not.

5.4 Counting larvae into the fry pond

While higher stocking rates are possible when hatchery staff have a lot of experience in preparing and looking after fry ponds, at the beginning it is better to keep them low. Not more than about 40–50 larvae should be stocked per square metre of fry pond. Thus a pond of about 2 000 m² could receive 80 000–100 000 larvae three days after hatching. If the survival was 50 percent this pond would produce 40 000–50 000 fry three to five weeks later. The following is a simple method for stocking larvae which requires no special equipment and not too much counting.

Larvae from a hapa or tank are collected into a single large bucket or basin which should not be more than about half full of water. The basin is carried to the prepared pond where the larvae are to be stocked. This should be done at a time of day when the temperature difference between the water in the hatching tank or hapa and that in the pond is small; not more than 3–4°C. If larvae have been hatched indoors this usually means that they have to be transferred to the pond in the morning. Water from the pond is added to water in the basin until it is nearly full. The water in the basin then vigorously mixed by hand so that the larvae are evenly distributed, and 3–5 samples of fixed volume are taken while the water continues to be mixed. (An ordinary drinking glass, a tin or a cup may be used to take these samples; it should be completely and rapidly filled each time.) Each sample is transferred to a suitable container such as a plate or glass, with a volume greater than the volume of the sampling glass. The larvae in each sample are counted into the pond using a spoon. These numbers are recorded. Then the remaining volume of water in the basin is measured out into the pond using the same glass, tin or beaker as was used for taking the samples. The total number of larvae stocked is then easily calculated as the average number per sample multiplied by the total number of glassfuls in the basin (do not forget to include the number of glassfuls taken out as samples). If all the larvae are to be stocked in one pond, there is no need to keep on mixing the water in the basin during stocking. If, however, they are to be divided between two or more ponds, then the water in the basin must be continuously mixed while the required number of glassfuls is counted out.

Keep a record of the individual sample counts, number of glassfuls, and the final estimated number of larvae. This allows possible mistakes in arithmetic to be corrected later. It also gives some information concerning the accuracy of the estimate obtained, and whether the process of mixing has been sufficiently thorough while the samples were being taken.

5.5 Feeding the fry pond

The main food of the fish in the fry rearing pond are invertebrates, first from the zooplankton, and later from the mud as well. Therefore pond preparation is very important. Many growers give some larval feed such as egg yolk or soybean extract during the first days of rearing, sprinkling these feeds in liquid form or as dust around the edge of the pond in the morning and evening. This may be helpful if the larvae are to be stocked in a pond where the plankton has not had time to develop, but it should not be necessary otherwise.

Later, particularly if fry are seen to be plentiful in the pond, spread some fine prepared feed. It is usual to start feeding 5–10 days after stocking and to scatter one or more of the following materials around the edge of the pond once or twice per day:

1. Cereal by-products such as mill screenings, malt screenings or brans with a crude protein content of around 8–10 percent. These materials are usually cheap, if available from a mill or brewery, but only a fraction of the material can be taken and digested by fish. For these reasons large daily quantities are usually given, i.e. 2 g/m² after five days increasing to about 5 g/m² or more after three to four weeks. The excess decays and fertilizes the pond encouraging both wanted and unwanted organisms. Care should be taken that the bottom mud does not become excessively black and foul-smelling. Bran is sometimes cooked before use. It may be supplemented by mixing with another, richer feed.

2. Soya flour or finely ground whole soybeans. Crude protein content is 38–44 percent. This flour is not widely available, and whole beans are difficult to process unless a suitable grinder is available. The cost is also fairly high. It is probably more digestible if roasted or cooked before use. Do not use imported ‘health food’ soya flour, it is far too expensive at over 2 kina/kg.

3. Prepared animal feed concentrate. Many different formulations are commercially available. Where there is a choice, choose one with a high protein content, high level of fishmeal, finely ground texture, no antibiotics and a low price.

4. Pellets. Finely ground and then mixed with water. Where available these are probably the best feed, either alone or supplemented with rice bran. However, they require the right type of grinder and add a stage to feed preparation. An alternative, though not so good, is simply to soak pellets in water until they disintegrate. Then feed the resulting slurry.

Fry may need to eat more than 40 percent of their body weight of food each day. The pond provides a variable amount of natural food. Various feeds are given to supplement this. Overfeeding is, of course, expensive and can lead to poor water conditions, while underfeeding and reduces survival and produces fish of a wide range of sizes. The amount of feed required by the fish in a particular pond will depend on their numbers and sizes which, of course, cannot be determined until the pond is drained. However, as a rule-of-thumb, estimate the weight of fry which is expected to be harvested. About 25 percent of this weight of feed should be sufficient as a daily supplement for the last few days of rearing. To economize on materials and labour, half this quantity may be made up of cereal wastes if available. The amount of feed supplied should increase in a roughly linear fashion during the rearing period. Only experience and experiment at a particular site will show how much feed is actually required. A wide range of fry sizes in the pond at any stage suggest that food supply is insufficient. The following schedule could be used as an example or starting point.

If scattered dry, much of the fine material floats on the surface of the water. If it is very windy most will end up along the down-wind bank of the pond. Feed may be scattered this way when there is little wind. Otherwise it should be mixed with water in a bucket and the resulting slurry spread mainly around the edges of the pond. The daily ration should be divided so that fry are fed two or three times per day.

5.6 Stocking other fish during fry-rearing

It is possible to put a few broodstock into the first rearing pond after the fry have been in the pond for about ten days. This has advantages both for the broodstock which can ripen in the fry pond, and for the fry, but with certain types of soil it may lead to very high water turbidity.

Where the growth of submersed aquatic weeds or filamentous algae is a problem, the turbidity caused by the digging behaviour of larger carp or broodstock can help to suppress this growth by preventing a lot of light from reaching the bottom of the pond, especially if this is combined with increased water depth. Productivity in the pond may be somewhat reduced, and, of course, the larger carp should be taken into account when feeding the pond. This probably will not eliminate weeds once they have become established, but if the pond is well cleaned before stocking, weed growth can be prevented or reduced. The total weight of additional fish should remain low, of the order of 10 g/m² (100 kg/ha).

5.7 Fry harvesting

The length of time for which fish remain in the fry rearing pond will depend largely on what size fry are required. In countries where fish production is well developed, there is a tendency for different people to specialize in different stages (chain activity). Hatcheries can then concentrate on producing very large numbers of larvae, and keeping them in ponds for only a few weeks of rearing. Depending on rearing conditions they will then usually be sold as 1–2 cm fry(0.05–0.25 g average weight) of which there will be 800–4 000 fry per full (200 ml) drinking cup, or 2–3 cm fry (0.25–1 g average weight, i.e., 200–800 fry per cup. As the length of rearing increases, so does the spread of sizes in the pond. Numbers of fish may drop as a result of competition and cannibalism. For this reason, where possible, it is better not to continue in fry-rearing ponds for more than about six weeks. If bigger fish are required, the pond should be drained, the fish sorted by size and fingerling ponds stocked at a lower density with fish of one size range.

Small numbers of fry can be harvested without draining the pond. In Israel a fine-mesh seine net would generally be used to concentrate the fry and hold them in the pond in a bag formed of the mid-section of the net while they were scopped up using a fine hand-net or bucket with small holes. They would immediately be transferred to an aerated tank on a trailer or pick-up.

In Indonesia farmers rarely possess seine nets. They catch fish using a variety of lift nets or baskets. A square lift-net (see Figure 5) of about 1 m² is easily made and netting of various grades can be fitted according to the size of fish to be caught. Round baskets, made of tarred bamboo or rattan, are more difficult to make, but equivalent plastic utensils can be found in some PNG hardware stores.

Harvesting all the fry involves screening the water outlet, whether monk or simple pipe, and lowering the pond till most of the mud is exposed. Most of the fish collect in the pond's internal drainage channels. The fish are then caught using lift nets, baskets, handnets, sieves or small seines. If the pond outlet is just a pipe, a large basket of traditional materials or plastic can be placed over the end to serve as a screen. This may have to be covered with mosquito netting if the fry are very small.

A good harvesting technique for speed, efficiency and minimum fish damage is shown in Figure 6. In the first stage, the pond water level is lowered in the usual way, screening the inside end of the outlet pipe. When the pond floor is nearly exposed, the outlet water is allowed to flow unscreened through the pipe into a net placed below the lower end of the outlet pipe and standing in water in an external drainage canal (stage 2), so that many of the fish are carried out in the pond in the water flow and do not have to be scooped up in nets. The remainder are then gradually harvested from the downstream end of the pond's internal diagonal drainage channel, down which a small amount of water continues to flow (stages 3–5). The last fry, therefore, concentrate at the top of the drainage channel in water of good quality (stage 6). This causes less stress than concentrating them at the lower end of the internal drainage channel.

Harvesting the pond always involves some fish losses. To keep these to a minimum, the work should be carried out in the early morning. Care must be taken to minimize the time that the fish spend at high density in warm muddy water. Keep the transfer containers filled with relatively cool clean water and empty them frequently. Store the fish under the best possible conditions with a good flow. Ponds with a regularly-sloped smooth floor, free of weeds and with adequate internal drainage channels, are the easiest to harvest.

Numbers and sizes of fry harvested from each pond should be recorded. Where there is a big size range, they may be sorted first. This can be done by hand, by using netting of a suitable grade, or by using a sorting trough similar to that shown in Figure 7. Numbers and sizes can be estimated quite accurately and quickly without the use of a balance using the fact that the specific gravity of fish is almost exactly one, i.e. they have the same average density as the water in which they swim. The capacity of commonly-used cups for drinking coffee is 200 ml. If fry are caught in a sieve and quickly poured into a cup of this size so that it is packed to the brim with fish, its contents will be 200 g, to quite a high degree of accuracy. It is, therefore, only necessary to count the number of glassfuls and number of fry per glass (for each size-class) in order to know both their total number and average weight. The fish from one such cupful are simply poured into a bucket or bowl of water and then counted. If one cup, fully packed, holds 400 fish then their average weight is 0.5 g each. This gives much more information than just classing the fish as fry or fingerlings which is a very imprecise and subjective classification. It is also much more accurate than giving lengths.

6. CARP BREEDING - APPROXIMATE SCHEDULE OF ACTIVITIES

The following is a summary of activities for controlled breeding of common carp.

The following notes refer to the above schedule:

Draining the pond. If this can be done earlier and pond dried more thoroughly, so much the better; three to four days is about the minimum.

Select spawners, put together with substrate, and place the spawners together with egg-collecting substrate in a tank, pond or hapa. If the fish do not spawn on the first or second night it may be worth injecting the female and one or more of the males. In this case time the injections so that the fish will be ready to spawn during the night when they would spawn naturally, i.e. give preliminary small injection in the early morning and main dose after lunch or mid-afternoon.

Injecting for induced spawning with stripping. In this case it is more convenient to time the injections for spawning in the mid-morning, i.e. give preliminary small injection after lunch and main dose at about 10 at night.

Applying fertilizer. If this is done only after successful spawning is seen to have taken place one avoids wasting fertilizer. However, slightly better results might be obtained by fertilizing a few days earlier.

Counting the larvae into the fry-rearing pond. The timing of this has to be decided by looking at the larvae, it will depend on temperature. If the larvae are swimming actively and appear to be searching for food, with pigmented eyes, functional mouth and swim bladder, it is probably time. If a little hard-boiled egg yolk is dispersed into the water through a fine cloth, feeding larvae will soon be seen to have whitish gut contents. They are then ready to go into the pond.

Harvesting fry. Before harvesting sample fry with a small, fine seine net, life net or hand net. Are they plentiful or few, very varied in size or not? If there are expected to be more than 20–30 fry/m², it is better not to leave the pond too long before harvesting or competition for food will limit their growth and cause a very great spread of sizes. The difficulty of course is that harvesting a large pond full of very small fish will require draining the pond through a rather fine screen; this will take a long time. A possible alternative would be repeated thinning or partial harvesting with a seine net.

7. FINGERLING PRODUCTION

Farmers who are trying to grow fish for consumption in small ponds will have a much better chance of success if they are provided with fingerlings (about 10–40 g) than if they receive fry (0.25–2 g). Their mortality losses are likely to be lower, the time taken for the fish to reach a good size will be very much shorter and the efficiency of the pond use is better. In regions where fish culture is already well developed farmers can specialize in different stages of the rearing process. However, where it is just beginning, rearing and distribution of fingerlings should be the task of government centres such as Aiyura.

To obtain good growth rates in fingerling ponds stocking densities should not be too high, not more than 2–4 fry/m² in ponds which are going to be well and regularly manured, and 5–10 fry/m² in manured ponds which are fed daily on a bodyweight basis using an effective feed such as broiler crumble. Fish should be sampled regularly (once every two weeks) to keep a check on their growth rates. This can be done by giving food in a corner of the pond where fish are accustomed to being fed, then closing off this corner with a suitable fine seine net. Fish of these sizes should be increasing their weight by 50–100 percent each two weeks. If growth rates start to fall feeding has to be increased or the biomass reduced by partial harvesting.

Fingerling ponds are stocked at a lower density than fry ponds. Also, fingerling rearing is likely to take two to three months to produce suitably-sized fish, while fry rearing takes only three to six weeks. Therefore the total area of fingerling ponds has to be considerably greater than that of fry ponds. (Depending on the intensity of operation and the size fish which are to be supplied this ratio may be two to ten times.) At Aiyura as it is now, one of the ponds of about 2 000 m² (pond 2, 3 or 4) could serve as a fry pond and pond 1 (5 000 m²) as a fingerling pond.

Tilapia fry could also be raised to fingerling size for sexing together with carp in one pond as long as the pond was completely emptied every two to three months to prevent wild breeding. Mullet fry would also do no harm but would have to be handled very carefully at harvesting time.