ANNEXES

ANNEX 1
Key to larval stages of freshwater prawns (Macrobrachium rosenbergii)

THIS ANNEX PROVIDES a simplified key to the eleven larval stages of M. rosenbergii and is illustrated with some micro-photographs kindly given to the author by the late Takuji Fujimura (Annex 1, Figures 1 - 12). The most prominent characteristics are shown in Annex 1, Table 1.

ANNEX 1, TABLE 1
Selected characteristics of Macrobrachium rosenbergii larvae and postlarvae

SOURCE: DERIVED FROM ISMAEL AND NEW (2000)

ANNEX 1, Figures 1-12

Macrobrachium rosenbergii go through eleven distinct larval stages (Figures 1-11)
before metamorphosing to become postlarvae (Figure 12)

SOURCE: TAKUJI FUJIMURA

ANNEX 2
Natural beach filter for seawater

SUITABLE BEACHES can be used as natural seawater filters for hatcheries. Some hatcheries draw water from perforated pipes protected by a 150 µm nylon screen buried approximately 1 m deep in the beach. However, the screens are prone to damage and it is better to develop the beach itself as a filter. This annex describes a cheap filter probe made from plastic, which is derived from a stainless steel probe developed by a zoologist, the late George Cansdale. The following notes have been extracted from Suwannatous and New (1982). The system is easy and cheap to make and is described here in spite of some scepticism that such simple systems are effective.

1. Basic requirements and capacity

Careful thought should be given to the location of the beach filter. A permeable beach is needed, with a depth of 2-3 m under a minimum of 30 cm water. Beaches of a wide range of types may be used, including sand, gravel, broken coral, shell, etc. The bulk of the sand grains should be between 0.5 mm and 5.0 mm but a great advantage of this system is that, during the development of the filter, excess fine sand is pumped out, leaving the larger grains in and around the filter probe; thus a precise sand specification is not needed. Uniformly fine sand, especially of wind-blown origin, is unsuitable on its own, but it can be graded up by adding coarse sand or gravel under and around the unit. If most grains are above 2 mm diameter, it helps to add fine sand on the surface around the unit during development. ‘Fine sand’ is defined for the purposes of this annex as material up to 1 mm and ‘coarse sand’ from 2 mm to 5 mm, but these are not technical terms. A few stones in the beach of up to 50 mm do not prevent it being developed as a filter bed but larger stones will reduce the efficiency of the filter and should be cleared away (or a different site chosen). Sites with little or no sand are not suitable for natural beach wells. Those with soft mud cannot be used. Where the beach is rocky some people have found that excavating a large hole and filling it with sand from another site, into which the filter probe is inserted, is effective. However, this may be very difficult and costly to construct and maintain. If the hatchery site is not adjacent to a beach with a favourable structure for a beach well there are a number of choices, including choosing a better site, bringing seawater (or brine) from another location (this is essential for inland hatcheries anyway), or pumping raw seawater and treating it within the hatchery.

The equipment described in this annex can be installed in any suitable beach. When this system was developed the original probes were constructed of stainless steel, which is expensive. However, cheaper plastic pipes can be used, provided the probes are taken up and cleaned more frequently.

The capacity of the pump required to operate the filter probe and the jet probe (see section 3), and the correct pipe diameter, depends on the water requirements of the hatchery, as well as its elevation above sea level and the distance between the pump and the filter probe, and between the pump and the seawater holding tank in the hatchery. It is important to have no noticeable flow resistance at the maximum water flow required. Equipment should not be too large for the characteristics of the site and the amount of water required by the hatchery, because this will result in excessive capital costs. Conversely, buying equipment which is too small for the site is a waste of money. The choice of pipe size is discussed in detail in an FAO manual (FAO 1992b). As an example of pump capacity, a 3 HP, 1 440 RPM self-priming electric pump sucking water from a filter probe 35 m distant through a 10 cm flexible hose (adapted down to 5 cm near the pump) and delivering water through a 10 cm pipe to a hatchery above the highest high-water mark to a site 350 m distant is capable of pumping about 20 m³ /hr of seawater.

2. Construction of the filter probe

Soften the end of a 1.5 m piece of 10 cm diameter PVC pipe with heat, taper it to a point, and make sure that it is sealed. Then cut three sets of slits into it (Annex 2, Figure 1). The three sets of slits should be cut in rings. The lowest set should be 20 cm from the bottom of the pipe and the space between the three sets of rings should be 40 cm. From the upper set of slits to the top of the pipe should measure about 45 cm. Each ring of slits should be 2.5 cm long. The individual slits should be 1-2 mm wide and the spaces between them 1 cm wide. Flow can be increased by inserting more rows of slits but care must be taken not to weaken the pipe so much that it will fracture.

3. Installation and operation of the filter probe

Ideally, the filter probe should have at least 30 cm water above it at spring low tide. In very permeable sand on a level beach, a unit can sometimes be installed above low tide mark; it must be inserted as deep as possible. However, unless water has free access to it at all times, flow from probes inserted above low tide mark may be limited when the tide is out. Sea movements are daily and predictable, and a known factor in the suction head. Tidal pattern may vary widely: rise and fall may be from less than 2 m to over 15 m and the tide may recede anything from a few metres to over 500 m.

Dig a soft area in the inter-tidal zone of the beach, using a jet probe (use a pipe similar to the filter probe but with a tapered open end and no slits) connected to the outlet side of the pump that will later be used for water extraction. The closer you can get to the low-water mark the better, but you need to be able to get access to the probe for maintenance. Then fix the flexible hose to the top of the filter probe and push it into the softened sand to a depth of about 1.5 m. When it is in position, move the other end of the hose to the intake side of the pump and start extracting water, letting it flow to waste. The well will need ‘developing’, as explained below, before the water is of quality high enough to use in the hatchery. Once the well has been developed, connect the outlet side of the pump to the pipe supplying the seawater holding tank in the hatchery.

Where quality is critical, the water should be analysed and its quality monitored. In any case, the salinity of the water should be monitored during beach well site selection to make sure that it is high enough for the requirements of the hatchery (for freshwater prawn hatcheries, for example, it must always be higher than 12 ppt); in some areas the water sucked from a beach well may have been diluted to levels below this by freshwater run-off or beach springs.

4. Developing the efficiency of the filter

The area of beach around the filter probe becomes the natural ‘beach filter’. Before it operates to the best efficiency it needs development. Thorough development is the key to success and this section of the annex is most important.

When the filter probe is buried and the suction line has filled with water, this is then connected to the pump intake. Tight joints with washers are essential, for the smallest air leak delays priming and lowers efficiency. Underwater leaks may admit raw water; however, if these are very small, they soon block. The filter can be developed with a temporary pump close to the water. When the pump is fully primed, reduce its speed until it runs steadily. At first the water will be full of silt and organic matter as it cleans the bed. Depending on the site, the water will become clear in anything from a few to many minutes. Stop the pump and then re-start it; after a very short interval the water will become dirty but then it will soon clear. When this happens, stop and re-start the pump again. Releasing the partial vacuum disturbs the sand in and around the probe. This allows more fine material to be sucked out and gradually pushes back the perimeter of clean coarse sand, thus improving flow. This result is the main reason why the filter needs development so that it will work efficiently. Continue the alternate stopping and re-starting process until water no longer becomes dirty after restarting, and the pump is working to full capacity. The type of beach and the pump size determines how long this process will take. Where the beach has a lot of black organic matter, development is best spread over several days to allow this to decay aerobically, after which it is easily sucked out.

The water should now be crystal clear, free of all suspended matter and organisms down to about 1 micron (1 µm) or less. Where particularly high quality water is needed (i.e. for research work), the seawater should be pumped to waste for several hours daily for at least a week (while the biological filter is developing in the beach filter). The time needed for this varies with temperature and other factors.

Where adverse site conditions impede progress, the following procedures may be tried:

• using a jet probe, dig the beach well in the area where the filter probe is to be inserted, and let the current carry away much of the silt;
• instead of just stopping the pump momentarily during development, release the suction completely, letting the water flow right back to the filter probe before re-starting the pump;
• interchange the intake and outlet hoses and reverse the water flow so that water is pumped through the filter probe into the beach for several minutes. Somebody needs to hold the probe in position during this process, to make sure that it does not become dislodged.

Small amounts of sand may be drawn through for some days, especially when pumping is periodic, but this is sterile and can easily be settled in a primary tank or small baffle chamber.

If the filter is only used intermittently it is good practice to pump some water to waste each time it is re-started; this need be for only a few minutes if the time since it was last used is only a day but should be about an hour if the filter has not been used for a week. Local experience will show how long water needs to be pumped to waste on each occasion.

5. Maintaining the efficiency of the filter

There is a tendency for the flow of water passing through any filter to gradually decrease, as the spaces between the filter bed particles become blocked. In marine sites tidal and wave movements generally keep the surface of the beach filter clear. If blocking does occur, this will only be in the top 1-3 cm, usually only the top 1 cm. If the flow from the filter becomes reduced, and this is not due to declining pump performance or other factors, this suggests that there is surface blocking. This can be cured in several ways, including:

• stopping the pump and raking an area of about 5 m radius around the probe, working to a depth of about 5 cm, then pumping the water to waste and redeveloping the filter for as long as needed;
• skimming off about 3 cm of sand from the surface and replacing it with new sand;
• forking the area over lightly and back-washing the filter by moving the suction hose to the pump outlet and drawing water through a spare hose. Somebody needs to make sure that the probe does not become displaced during this operation;
• moving the filter probe to another area and developing a new beach filter. A firmly embedded filter probe can be quickly freed by changing over suction and delivery hoses at the pump and blowing back, after letting some air into the line.

A change in tidal pattern or a badly sited breakwater may cause a metre or more of sand to be removed from the beach, though this is unlikely near or below low tide mark. If the filter probe becomes exposed because of this type of problem, it must be re-installed and re-developed.

BERRIED FEMALES brought into the hatchery just before their eggs hatch are not normally fed. If they are, they can be given the normal grow-out diet. However, if broodstock are being maintained for long periods, it is best if you use a diet which encourages maturation. Some simple methods of supplementing grow-out feeds for this purpose are described in the section of the manual on broodstock. Annex 3, Table 1 shows two formulae for specific broodstock diets which have been shown to be effective for this species.

ANNEX 3
Maturation diets for broodstock freshwater prawns

THERE HAVE BEEN many publications on the use of Artemia as a live food since the original FAO manual on freshwater prawn farming was published in 1982. The following annex has been derived mainly from two other FAO publications (Lavens and Sorgeloos 1996; Moretti, Pedini Fernandez-Criado, Cittolin and Guidastri 1999) whose authors are hereby gratefully acknowledged. You are recommended to study these manuals for a thorough understanding of the topics in this annex.

ANNEX 3, TABLE 1
Broodstock diets for Macrobrachium rosenbergii

SOURCE: THESE TWO DIETS GAVE THE HIGHEST NUMBER OF EGGS PER G BODY WEIGHT OF FEMALE (APPROXIMATELY 1 355), THE LARGEST EGG DIAMETER (APPROXIMATELY 0.5 MM) AND THE HIGHEST HATCHING RATE (90% FOR DIET 1 AND 82% FOR DIET 2) IN A TRIAL WITH MACROBRACHIUM ROSENBERGII BY DAS, SAAD, ANG, LAW AND HARMIN (1996).

ANNEX 4
Source, hatching and enrichment of Artemia

1. Sources, quality and use of Artemia cysts for freshwater prawn larvae

Artemia cysts can be obtained in cans from commercial companies but originate in many different countries, including Brazil, China, Iran, the former Soviet Republics, and Viet Nam. The main source is still the Great Salt Lake in Utah, USA. Dry cysts (2 to 5% moisture) are very resistant to extreme temperatures (hatching viability is not affected in the temperature range -273°C to + 60°C and can even tolerate short exposure to temperatures between 60°C and 90°C. Hydrated cysts are far less tolerant. Mortalities occur below +18°C and above +40°C, and a reversible interruption of their metabolism occurs between -18°C and +4°C and between +33°C and about +40°C. Active cyst metabolism occurs between +4°C and about +33°C; hatching percentage is unaffected within this range but the nauplii hatch earlier at higher temperatures. When incubated in saline water, Artemia cysts swell up and become spherical within 1 to 2 hours (Annex 4, Figure 1).

After 12 to 20 hours of hydration, the cyst shell bursts (breaking stage) and the embryo, surrounded by the hatching membrane, becomes visible. The embryo then leaves the shell completely and hangs underneath the empty shell but may be still attached to it by the hatching membrane (umbrella stage). The hatching membrane is transparent and the development of the pre-nauplius into the instar I nauplius, which starts to move its appendages, can be viewed through it. Soon after this, the hatching membrane breaks open (hatching) and the free-swimming larva is born head first. Instar I nauplii cannot feed; thus, the older nauplii are when they are fed to freshwater prawn larvae, the more they will have used up the energy reserves with which they were born and the less nutritional value they will have for the prawn larvae. Instar II Artemia have used up 25 to 30% of their energy reserves within 24 hours after hatching (Merchie 1996). Instar II Artemia are also transparent and swim faster than instar I larvae; they are therefore less easy for prawn larvae to catch. Detailed information on the biology and ecology of Artemia is given in Van Stappen (1996).

The nutritional quality and physical size of the nauplii which hatch from Artemia cysts (referred to elsewhere in this manual as brine shrimp nauplii – BSN) vary enormously from source to source and even (in the case of nutritional quality) between individual batches from a single source. Of particular importance is the level of an essential polyunsaturated fatty acid, eicosapentaenoic acid [EPA] (20:5n-3), which depends on the composition of the primary food available to the brine shrimp in the locations where they originate. Further reading on this topic can be found in Merchie (1996). The nutritional quality of Artemia nauplii (BSN) can be improved by enrichment, as described later in this annex. For practical purposes, cysts can be categorized by the size of the first stage nauplii they produce: small (~430 µm), medium (~480 µm) and large (~520 µm) but the size is not so important for freshwater prawns as it is for some marine fish. Freshwater prawns can ingest BSN of all sizes.

The two other important Artemia factors which vary from batch to batch are the number of cysts per gram and their hatching rate. The most effective way of checking the basic quality of the cysts you buy is to measure the hatching efficiency, because this is a check not only on the percentage of cysts that hatch but is also a means of judging how much detritus (e.g. empty cyst shells, sand, salt, etc.) that the batch contains. Hatching efficiency is defined as the number of BSN hatched per gram of cysts purchased. Premium quality cysts from the Great Salt Lake should yield about 270 000 BSN per gram of cysts. Smaller cysts (e.g. the San Francisco Bay source) may provide up to 320 000 BSN per gram of cysts. However, some sources yield as low as 100 000 BSN per gram of cysts. Good quality cysts should start to hatch after 12-16 hours incubation and all should have hatched by 24 hours (Van Stappen, 1996). This measure of quality is known as the hatching rate (HR). Hatching rate curves for two samples of cysts are illustrated in Annex 4, Figure 2. A procedure for determining hatching percentage, hatching efficiency and hatching rate is given in Annex 4, Table 1; this should be used to compare different sources of cysts, so that you know whether you are getting value for money or not.

Choosing to feed prawn larvae with 24 hour old or with 48 hour old (timed from when the cyst incubation process starts) BSN depends partly on the rate of Artemia hatching (characteristic of the source and the environmental conditions you supply it with) and partly on operator preference. Some hatcheries use 24-36 hour old BSN; others use 24-36 hour old BSN at first and progress to the use of larger 48 hour or 72 hour old BSN (reared on ‘greenwater’ or rice bran) as the prawn larvae grow. However, the use of 24 hour old BSN enables you to use the same equipment on a daily basis. Keeping BSN longer before using them is more expensive as producing food to rear them needs more equipment.

Artemia cysts need treatment before the hatching process commences to ensure maximum hatching occurs and to maintain healthy conditions in larval rearing tanks. This process is described below.

2. Treatment of cysts before hatching

Artemia cysts are by nature contaminated with bacteria, fungal spores, and other micro-organisms, and may be contaminated with organic impurities. The use of BSN which arise from cysts that have not been disinfected can cause health problems in larval rearing tanks; poor water quality and larval diseases (especially those caused by Vibrio spp.) can be introduced when untreated empty shells, unhatched cysts and cyst hatching water are transferred to the larval rearing tanks. The decapsulation process also disinfects the cysts but has the additional advantages of increasing the hatching efficiency of some batches and reducing the transfer of indigestible matter to the larval rearing tanks.

DISINFECTION

Commercially disinfected cysts may be available on the market but it is safer to apply routine disinfection. Simple disinfection can be done by immersing the cysts in a hypochlorite solution (200 ppm active chlorine), according to the procedure in Annex 4, Table 2. The procedure for preparing the disinfection solution is given in Annex 4, Table 3. After disinfection, some hatcheries remove residual chlorine with sodium thiosulphate but thorough rinsing of the treated cysts is regarded as adequate in others. However, while reducing the risk of contamination, disinfection does not kill all the organisms present in the outer shells of the cysts and is not recommended in this manual. Decapsulation (see below) is a more effective means of obtaining contaminant-free cysts, as well as potentially increasing hatching efficiency.

DECAPSULATION

Decapsulation completely removes the hard shell that encapsulates the dormant Artemia embryos. This process is recommended because:

• although cyst shells can be physically removed after hatching non-decapsulated cysts (this is necessary because cyst shells are not digested and may obstruct the larval gut) this laborious process is completely unnecessary if the cysts are decapsulated;
• decapsulation improves the energy content of the BSN (which do not waste energy breaking their own way through the shell) and sometimes increases hatchability; and
• decapsulation is an efficient means of disinfection.

Decapsulation involves hydrating the cysts, removing the (then spherical) brown shells in a hypochlorite solution (500 ppm active chlorine), washing them and deactivating the remaining chlorine. Decapsulated cysts can be directly hatched into BSN, or filtered and stored in a refrigerator at 0-4°C for a few days before use, or transferred to a saturated brine solution for longer storage (up to several months). If stored, they must be protected from sunlight because the hatchability decreases when exposed to UV light.

Decapsulated cysts can be dried and fed directly (without hatching). However, this type of food is more appropriate for postlarvae because prawn larvae feed better on a moving target; this is one of the primary reasons for feeding them live BSN. Decapsulation involves the use of a source of hypochlorite, usually liquid bleach (NaOCl), and an alkaline product, usually technical grade caustic soda (sodium hydroxide, NaOH), to increase pH to above pH 10. Finally, the residual hypochlorite is neutralized with sodium thiosulphate. Details of the cyst decapsulation process are provided in Annex 4, Table 4 and the preparation of the active chlorine solution in Annex 4, Table 5. Commercial bleach varies widely in chlorine content, so it is essential to measure the chlorine content of each batch, as detailed in Annex 4, Table 6.

3. Hatching decapsulated Artemia cysts and harvesting nauplii (BSN)

Almost any type of container can be used for hatching Artemia, including rectangular and round tanks, cylindrico-conical tanks, garbage cans, modified drinking water containers (Annex 4, Figure 3), converted chemical carboys, ‘klong’ (water) pots (Annex 4, Figure 4) and other structures (Annex 4, Figure 5). However, several of those types of containers need to be siphoned to empty the BSN out after hatching. The easiest to use is a conical based circular plastic or fibreglass tank which is elevated so that the BSN can be harvested by gravity. A 1 m³ volume tank is convenient. The upper parts of the tank should not be transparent; the conical part of the tank can be transparent or translucent and have a valve at the tip of the cone for harvesting purposes. A simple tank is illustrated in Annex 4, Figure 6). Aeration should be supplied through a half-inch PVC pipe which extends to close to the conical tip of the bottom of the tank; this is to keep the cysts in vigorous suspension, as well as to keep the dissolved oxygen level above 4 ppm. The tanks should be filled with natural seawater that has been filtered and is about 33-35 ppt in salinity (well-buffered artificial seawater can also be used). The optimum hatching temperature is 25-28°C and the best hatching pH is 8.0-8.5. A solution of about 1 g of sodium bicarbonate (NaHCO3)/L can be used to achieve this pH level, if necessary. The surface of the water at the top of the tanks should be illuminated at 2 000 lux. If the available sunlight is not sufficient, two 60 watt fluorescent tubes placed just above the tank rim is sufficient to achieve this level of light intensity. The hatching tanks should be elevated, to aid the harvesting process.

A one-ton (1 m³ ) tank can be stocked with 250 g to 1 kg (0.25-1.00 g/L) of Artemia cysts. The actual amount needed depends on the source of cysts you purchase, and their hatching efficiency (see Annex 4, Table 1). No specific hatching efficiency (HE) is recommended. Your choice of cysts will be a matter of comparative cost and the results you obtain. Cheap low-quality cysts can be used but you will need more of them. The important thing is that the HE must be measured, so that you know the actual quantity of cysts you will need to obtain the number of BSN you need for your hatchery. A 1 m³ Artemia rearing tank stocked at this level should provide enough BSN to feed ten 5 m³ larval rearing tanks for one day, which are capable of producing 500 000 to 1 million freshwater prawn postlarvae per cycle. The hatching and harvesting procedure is given in Annex 4, Table 7.

Annex 4, Figure 3
Brine shrimp nauplii (BSN) can be reared in many different containers, including old drinking water bottles (Peru)

SOURCE: OSCAR ORBEGOSO MONTALVA

Annex 4, Figure 4
These ‘klong’ pots were being used to culture the larvae of Macrobrachium rosenbergii but are also sometimes used to rear Artemia (Thailand)

SOURCE: SPENCER MALECHA

Annex 4, Figure 5
These outdoor tanks are used to rear Artemia (Thailand)

SOURCE: HASSANAI KONGKEO

4. Enrichment

The nutritional quality of BSN, especially in terms of the PUFAs eicosapentaenoic acid (EPA, 20:5n-3) and docosahexaenoic acid (DHA, 22:6n-3) can be increased by enrichment. Enrichment for disease control is also possible. These processes are sometimes known as boosting or bioencapsulation and are a feature in many marine fish and shrimp hatcheries, especially when older BSN are used (Annex 4, Figure 7).

Some advantages from enriching BSN with vitamin C have been clearly demonstrated by measuring the reaction of freshwater prawn larvae to stress testing when fed starved versus enriched BSN. Lavens, Thongrod and Sorgeloos (2000) have suggested that the use of HUFA/vitamin C-enriched BSN would provide benefits to hatcheries which do not feed other HUFA- and vitamin-rich supplements (for example, included in EC diets) to larval freshwater prawns from stage V onwards. Several commercial enrichment products are now available and each supplier describes the practical enrichment process. Those Macrobrachium hatchery operators who wish to enhance the quality of their BSN should follow the suppliers’ instructions. The names and suppliers of BSN enrichment products include Super Selco and DHA Selco (INVE Aquaculture NV., B-9080 Lochristi, Belgium), Superartemia (Catvis BV., 5222 AE ‘s-Hertogenbosch, Netherlands) and SuperHUFA (Salt Creek Inc., Salt Lake City, Utah 84104, USA). If you wish to study the topic of Artemia enrichment further, it is recommended that you read the details provided by Merchie (1996).

ANNEX 4, TABLE 1
Determining hatching percentage (H%), hatching efficiency (HE) and hatching rate (HR).

ANNEX 4, TABLE 2
Disinfecting¹⁹ 1 kg of Artemia cysts

SOURCE: DERIVED FROM VAN STAPPEN (1996) AND MORETTI, PEDINI FERNANDEZ-CRIADO, CITTOLIN AND GUIDASTRI (1999)

ANNEX 4, TABLE 3
Preparing a 200 ppm active chlorine solution for use in disinfecting 1 kg of Artemia cysts

ANNEX 4, TABLE 4
Hydrating and decapsulating²⁰ 1 kg of Artemia cysts

SOURCE: DERIVED FROM VAN STAPPEN (1996) AND MORETTI, PEDINI FERNANDEZ-CRIADO, CITTOLIN AND GUIDASTRI (1999)

ANNEX 4, TABLE 5
Obtaining 0.5 g of active chlorine for use in decapsulating 1 kg of Artemia cysts

ANNEX 4, TABLE 6
Measuring the level of chlorine in commercial liquid bleach or bleaching powder

SOURCE: DERIVED FROM VAN STAPPEN (1996) AND MORETTI, PEDINI FERNANDEZ-CRIADO, CITTOLIN AND GUIDASTRI (1999)

ANNEX 4, TABLE 7
Procedure for hatching Artemia cysts and harvesting nauplii

SOURCE: DERIVED FROM VAN STAPPEN (1996) AND MORETTI, PEDINI FERNANDEZ-CRIADO, CITTOLIN AND GUIDASTRI (1999)

ANNEX 5
Production of farm-made larval feeds

MANY DIFFERENT farm-made feeds can be used in the rearing of freshwater prawn larvae, in addition to the feeding of brine shrimp nauplii (BSN). This annex describes the preparation of three versions of an egg custard diet (EC) and the use of fish flesh. The first EC diet is an egg-mussel mixture. Experience has shown that using fish flesh (larval diet No. 4), especially when carelessly prepared or overfed, can be a grave source of water pollution in larval rearing. Farm-made larval diets No. 1 and No. 3 are the most simple to prepare.

Farm-made larval diet No. 1:

Prepare as follows:

• blend 0.5 kg of shelled mussel (other molluscs can be used, but mussel seems best) in a blender;
• strain the chopped mussel through a coarse cloth and discard the connective tissue, retaining only the material which passes the strainer;
• using the whole of the mussel which has passed the strainer, add three or four whole eggs and stir thoroughly in the blender (note: it is important to use the white as well as the yolk of the egg - the white contains good quality protein - some people think using the white of the eggs causes water pollution but it does not if homogenized properly);
• steam the mixture over water (like poaching an egg) until it solidifies into a custard;
• screen to the correct size (see the main text of the manual) and feed directly; or
• you can refrigerate it for a few days for later use (however, the quality of frozen EC is not as good as fresh EC for feeding purposes).

Farm-made larval diet Nos. 2 and 3

Prepare as in larval diet No. 1 but use the ingredients shown in Annex 5, Table 1.

Farm-made larval diet No. 4

Skipjack tuna, bonito or mackerel are good types of fish to use when preparing this feed. It may also be used as an ingredient in larval diet No. 1 above, partially or totally replacing mussel. The results with mussel seem to be superior.

Prepare the fish as follows:

• fillet the fish, discarding head, bones and viscera;
• grind and liquidize the flesh, as for the mussel in larval diet No. 1, in a blender;
• force the flesh through stainless steel sieves with a strong jet of freshwater (this grades the particles and washes the flesh free of blood and excess oil). The mesh sizes should be chosen to produce particles of a size relevant to the age of the prawn larvae (see the main text of the manual);
• use the fish flesh directly; or
• form into balls of known weight for storage (it may be kept in the refrigerator for 2-3 days or frozen for longer periods; however, frozen material is less satisfactory than fresh).

ANNEX 5, TABLE 1
Ingredients for farm-made larval diets Nos. 2 and 3

NOTE: TETRACYCLINE HAS BEEN USED IN LARVAL DIET NO. 2; YOU SHOULD READ THE TEXT OF THE MANUAL CONCERNING THE DANGERS OF REGULARLY USING ANTIBIOTICS IN HATCHERIES.

SOURCE: DERIVED FROM LAVENS, THONGROD AND SORGELOOS (2000)

ANNEX 6
Stock estimation

ESTIMATING OF THE NUMBER of animals present under hatchery or pond conditions is difficult. The four critical times when it is important to assess the number (and sometimes the size) of prawns present in the system are:

• when postlarvae are harvested, to provide a record and an assessment of the production efficiency of each larval batch and tank;
• when postlarvae are transferred from the hatchery to the pond, to control stocking density and determine feeding rates;
• at intervals during the grow-out period, to check growth rate and survival; and
• when market-sized prawns are harvested, to provide a final or cumulative record (an assessment of the productivity of the pond and management system being used).

The following methods are suggested for stock estimation.

1. Stock estimation when postlarvae are harvested

The following system is suggested:

1. before the harvested PL are transferred to the PL holding tank, suspend them temporarily in a small container with a known volume of aerated water;
2. agitate the water thoroughly to evenly disperse the animals;
3. take four samples from the container in 100 ml beakers;
4. now place the bulk of the postlarvae into the holding tank (do not wait until the sample counting process is complete);
5. count every animal in each of the four 100 ml beakers (one way of doing this is to take quantities into a graduated pipette held at a 45° angle towards a lamp and to count the animals as they swim up towards the light);
6. average the number of postlarvae found in each 100 ml beaker and multiply this number by the volume of water in the container mentioned in (a) above (in ml) and divide by 100.

The following is an example of the calculation for a postlarval stock estimation:

Let us assume that the small container [see (a) above] had a volume of 25 L. You counted 80, 86, 90 and 100 postlarvae in the four beakers [see (e) above]. The total number of postlarvae in the 25 L container can be calculated as follows:

Average number of PL in 100 ml = (80 + 86 + 92 + 98) ÷ 4 = 89

Number of PL in the 25 L container = 89 x 25 x 1 000 ÷ 100 = 22 250

2. Stock estimation when postlarvae are transferred to nursery or to grow-out facilities

The following procedure is not very accurate but is sufficiently practical for use, especially where the same person always does the counting:

1. on every transfer occasion count out 100 PL individually by dipping a hand bowl into a larger bowl containing PL. Transfer them into the plastic transport bag, or the transport tank;
2. measure out further batches by visual comparison to the counted batch, and add them to the plastic bag or transport tank. You will quickly become able to estimate the number of PL in each small bowlful dipped from the larger bowl quite accurately;
3. once the PL have been counted into the bags or transport tanks in batches, usually of 1 000 or 2 000, they are not normally counted again before they are stocking into the rearing facilities.

Another version of the method described above is to weigh the first batch counted [see (a) above] and use this weight alone to measure subsequent batches. However, this may cause more stress to the animals. Note that two-month old juveniles are much easier to count than PL.

Annex 6, Figure 1
Freshwater prawns can be sampled with a cast net; the polythene sheeting is not a pond liner but is placed on the banks to prevent prawns escaping from the pond in the rainy season (India)

Annex 6, Figure 2
This sampling by cast net reveals a badly eroding pond bank (Thailand)

SOURCE: STEPHEN SAMPATH KUMAR (TOP) SOURCE: MICHAEL NEW (BOTTOM)

3. Stock estimation during the grow-out period

Once the prawns have been put into a pond, it is extremely difficult to estimate growth rate or survival. Multiple seine and cast net samples seem the only reasonable method of following the growth rate of a crop of prawns. At least this enables a comparative estimate to be made. It is important that the method of sampling on each occasion is exactly the same (the same net; the same time of day; the same areas of the pond sampled; the same method of casting or pulling the net through the pond; and preferably the same person doing the sampling).

Even though the result may be grossly inaccurate, sampling animals from the ponds with a cast net (Annex 6, Figures 1 and 2) or a seine net on a regular basis does give you a reasonable idea of how your crop is growing. It is also a good opportunity to examine the health of the animals. It does not, unfortunately, give you much more than a vague idea of the survival rate.

Annex 6, Figure 3
Measuring the length of a prawn (Brazil)

SOURCE: DEBORAH ISMAEL

4. Stock estimation when market-sized prawns are harvested

From the practical rather than the scientific point of view, there are two vitally important data which must be recorded at harvest time. One is the total drained weight of the harvest and the other is the average size of the animals harvested.

From these data the numbers of animals harvested can be estimated. Since you already have an estimate of the number of PL or juveniles stocked, you can then calculate an estimated survival rate. Together, you can use these data to assess the productivity of your pond and the efficiency of the management system you have used. Combined with the costs of production and the market value of the product, this information enables you to calculate the overall economic performance of each pond.

Although the length of the prawn [biologists usually measure them from behind the eye stalk to the tip of the tail (Annex 6, Figure 3); farmers usually measure them from the tip of the rostrum to the tip of the telson] is a more accurate form of measurement than weight, it is not so easy for the you to measure, particularly as you then have to convert the measurement to weight using a calibration curve (Annex 6, Figure 4).

The weight of the animal can easily be measured on a portable scale. It will be inaccurate because of the amount of water adhering to the animal, particularly within the gill chambers. However, especially when the same person always does the weighing, it is possible to standardize the weighing technique and to achieve reasonable comparative accuracy.

It is suggested that you individually weigh about 250 animals for every 500 kg harvested (this is equivalent to a 2% sample of the population if the average weight is 40 g). Take the sub-sample from the total harvest by dip-netting from the holding container or cage. Do not select individual animals because this will lead to a bias towards the larger animals. Use a dip-net and weigh every prawn that you obtain in your sample.

ANNEX 7
Seine nets

THERE ARE VERY MANY different types of harvesting and sampling devices used in aquaculture, including seines, gill nets, lift nets, cast nets, bag nets, traps, and barriers. The making and use of these nets and traps is described in detail in another FAO manual (FAO, 1998). The following annex deals specifically with the use of seine nets in the cull-harvesting of freshwater prawns; information from the original FAO freshwater prawn manual has been supplemented with material drawn from FAO (1998) and Valenti and New (2000).

3/8 inch (9.5 mm) polypropylene (common trade names: danaflex, nufil, ulstron) is suitable for the floater line (sometimes called the head line) and 1/2 inch (12.7 mm) polyamide (common trade names: nylon, perlon, amilan, anzalon) for the sinker (sometimes called the foot line). Nylon is soft and follows the contours of the pond, while polypropylene is light and floats, yet is stiff enough to minimize sagging. Sinker lines smaller than 1/2 inch tend to sink into the mud. Soak the ropes for 12 hours and wet-stretch and dry them to prevent twisting. Both floater and sinker lines must be 2-3 m longer than the seine itself. You will also need pulling ropes. Long seines can be handled better if each end of the floater and sinker lines is fixed to a wooden pole and the pulling rope is attached to the top and the bottom of the pole. These poles can be used to stake (moor) the seine by ramming them into the bottom of the pond.

Monofilament netting is best. Double knotted, 17 1b test netting should be used. Mesh sizes (stretched) may vary from 18-50 mm; the choice depends on the market size of prawns you wish to capture. The rostrums, claws and other appendages of prawns tend to get caught in the net, so a larger mesh size than you would use for the same size (weight) of finfish can be used for freshwater prawns. For example, a stretched mesh (Annex 7, Figure 1) size of about 40 mm will retain prawns of 45 g and above. For comparison, FAO (1998) states that a net with a stretched mesh size of this size (40 mm) retains silver carp of about 30 g, or common carp or tilapia of about 20 g. The depth of the net should be about 1.5-2.0 times the depth of the deepest water to be seined. Its length should also be at least 1.5 times the width of the pond through which it will be drawn. Monofilament of 60 lb test should be used to fix the netting to the floater and sinker lines, using a ‘double clove hitch’ at every third eye. Net ends should be reinforced with a heavy strand of nylon to prevent tearing. A seine net being used for harvesting prawns is illustrated in Annex 7, Figure 2 and a typical design for a seine net is shown in Annex 7, Figure 3.

Sufficient floats should be used to prevent the line sagging. If this occurs, some prawns will crawl over the net. In general, a float can support a weight (its ‘floatability’) equivalent to 80-90% of its volume; a single 70 mm x 40 mm egg-shaped or oval-shaped float with a hole diameter of 9 mm can support a weight of 63 g. Cylindrical plastic (PVC or polyurethane) floats, 64 mm long and 64 mm in diameter have been used on freshwater prawn seines, for example. Moulded U-shaped leads are favoured against commercial sinkers as they have a smaller cross-section. The total weight of the lead sinkers needs to be 1-1.5 times the total floatability of the floats. Leads cut to 37 mm from a 3 mm thick sheet, weighing about 60 g, hammered onto the sinkers every 28 cm are suggested for freshwater prawn seines.

Some seines used for freshwater prawns have a bag similar to that of a beach seine, except that the top is left open and the distance between the floaters is reduced to prevent prawns from escaping over the top. A bag with floor dimensions of 15 ft x 9 ft (approximately 4.6 m x 2.7 m), tapering down to 4 ft (~1.2 m), will hold 200 kg of live prawns. Many seines used for harvesting freshwater prawn grow-out ponds have no bag, but a temporary catch area is made by the seine operators drawing in the sinker line as the seine is pulled towards the bank. Where prawns are to be transferred to another pond, for example from a nursery pond to a grow-out pond, the use of a bagged seine net may lessen the amount of damage to the prawns.

ANNEX 7, FIGURE 2
Freshwater prawns (Macrobrachium rosenbergii) can be cull-harvested by seining (Hawaii)

SOURCE: TAKUJI FUJIMURA

¹¹ vitamin premix (per kg mix): 24 000 mg pyridoxine; 142 000 mg ascorbic acid; 23 700 000 IU vitamin A; 4 740 000 IU vitamin D; 24 000 mg riboflavin; 71 000 mg calcium ð. pantothenate; 142 000 mg niacin; 24 000 mg thiamine; 12 000 mg folic acid; 12 100 mg vitamin B12; 100 mg biotin.

¹² mineral premix (per kg mix): 15 270 mg copper (Cu); 100 450 mg iron (Fe); 97 500 mg manganese (Mn); 1 190 mg iodine (I); 159 180 mg zinc (Zn).

¹³ type not stated.

¹⁴ it is important to use exact quantities of cysts and water, etc., in order to ensure that valid comparisons are being made.

¹⁵ lugol solution: Dissolve 50g potassium iodide (KI) and 25g iodine (I2) in 100 ml boiling water (= solution A). Dissolve 25 g sodium acetate (CH3COONa) in 250 ml water (= solution B). When solution A cools, mix solutions A and B and store in a cool dark place.

¹⁶ the umbrella stage is when the embryo hangs beneath the empty shell of the cyst, still within the hatching membrane.

¹⁷ dissolve 40 g NaOH in 100 ml distilled water.

¹⁸ 5.25% NaOCl.

¹⁹ some of the chemicals used in disinfection and decapsulation are toxic and/or can cause burns. Wear gloves and protective eyeglasses.

²⁰ some of the chemicals used in disinfection and decapsulation are toxic and/or can cause burns. Wear gloves and protective eyeglasses. All decapsulation operations should be carried out in a well-ventilated room. Gloves and protective eyeglasses should be worn.

²¹ prepare starch solution by mixing 5 g of starch (C6H10O5)n with a little cold water and grind in a mortar; pour into 1 L of boiling distilled water, stir and settle overnight; use the clear supernatent, preserved with 1.25 salicylic acid (C7H6O3) and store in a dark bottle.