Appendix 1
LIST OF INSTITUTIONS WITH FEED TRAINING FACILITIES

South East Asian Fisheries Development Centre
Tigbauan
Iloilo
Philippines

Contact: via Dr Chen Foo Yan
Network of Aquaculture Centres in Asia
NACA
c/o UNDP
G.P.O. Box 618
Bangkok
Thailand

College of Fisheries
University of Washington
Seattle
Washington
USA

Contact: Dr John Halver

Institute of Aquaculture
University of Stirling
Stirling
Scotland
UK

Contact: Professor R. J. Roberts

Tokyo University of Fisheries
4-5-7 Konan
Minato-ku
Tokyo
Japan

Contact: Dr Takeshi Watenabe

Fisheries and Marine Service
Department of the Environment
Halifax Laboratory
P.O. Box 429
Halifax
Nova Scotia
Canada

Contact: Dr John Castell

Buhler Brothers Ltd.
Engineering Works
CH-9240 Uzwil
Switzerland

Contact: Dr Urs P. Grunder
Head, Department of Food
and Feed Technology

Bodega Marine Laboratory
P.O. Box 247
Bodega Bay
California 94923
USA

Contact: Dr Douglas Conklin

AQUACOP
Centre Oceanologique du Pacifique
BP 7004 Taravao
Tahiti
French Polynesia

Contact: Dr Gerard Cuzon

Faculty of Fisheries
University of Kagoshima
Kagoshima
Japan

Contact: Dr Akio Kanazawa

Ministry of Agriculture, Fisheries and Food
Fisheries Laboratory
Lowestoft
Suffolk
UK

Contact: Dr P. J. Bromley

Institut für Küsten-und Binnerfischerei
Bunderforchungsanstalt für Fischerei
Hamburg
Federal Republic of Germany

Contact: Dr Klaus Tiews

Appendix 2
COMPARISON OF MARINE FISH/SHRIMP VITAMIN REQUIREMENTS AND THE AMOUNTS SUPPLIED BY THE CURRENT PROJECT PREMIX¹ AND ANOTHER PREMIX NOW AVAILABLE IN MALAYSIA

¹ Currently, 2.5% of a 1/100 dilution of Roche Vitamin Concentrate 428/3, to which 5 g of Vitamin C is added, made up with wheat pollards is used (Chow, 1984)
² Recommended supplementation levels (Roche, 1985)
³ Provisional requirements for marine Percoidae (New, 1985)
⁴ Increasing the quantities of those vitamins which are deficient by decreasing the dilution of the concentrate would raise the levels of Vitamins A and D3 to excessive dietary levels. This is NOT recommended. In calculating the amount in mg/kg dry diet (zero moisture) achieved by adding 2.5% of the current vitmix to a moist diet formulation, a moisture content of 45% has been assumed. In the case of the Agrimate premix, an inclusion rate of 0.2% (manufacturers recommendation) and a dietary moisture content of 10% has been assumed.
⁵ This was increased to 908 mg during the present consultancy (= “Modified Vitmix”)
⁶ The information available on the Agrimate premix indicates Vitamin A content content in the premix to be 275 000 I/kg. There may, however, have been an error of ÷10 in the transfer of this information. If so, the Vitamin A level in the diet would become 6 100 I/kg

Appendix 3
IMPROVED VITAMIN CONCENTRATE FOR SHRIMP AND SEABASS

It is suggested that either Roche, Sin Heng Chan, or Union Farm be asked to make up the following “Improved Vitamin Concentrate” which could then be used for preparing the “Improved Vitamix” to be used in making marine finfish and shrimp feeds at Sg. Merbok. Details of the dilution necessary to prepare the “Improved Vitmix” are given in the footnote¹.

^a In calculating the amount/kg dry matter, an inclusion rate of 2.5% in a moist diet has been assumed, together with a dietary moisture content of 45%

¹ For use, 1 part of the “Improved Vitamin Concentrate” should be diluted with 6 parts of wheat pollards. The result is the “Improved Vitmix” mentioned in Appendixes 8, 10, 12, 14, 15, 16. The Vitamin Concentrate should preferably be kept refrigerated, or at least in an air-conditioned room, for not more than 6 months

Appendix 4
INGREDIENT SOURCES AND COST IN PENANG AND KEDAH (November 1985)

1. Ingredients Used in Experimental Diets:

2. Other Ingredients Locally Available:

3. Other Ingredients Nationally Available:

¹ For suppliers' addresses, see Appendix 6
² Estimated price for feed grade fish oil used for costing purposes
³ Price assumed for costing purposes

Appendix 5
INGREDIENT SOURCES AND COST IN BESUT AND JOHOR (December 1985)

¹ For suppliers' addresses, see Appendix 6

The following prices were being paid for small quantities of ingredients bought from various local sources by the Brackishwater Aquaculture Research Centre in Galong Patah, Johor (the excessive prices are caused by the small volume of purchases):

Appendix 6
ADDRESSES OF FEED INGREDIENT AND EQUIPMENT SUPPLIERS

¹ Sells shrimp feeds

² For moist feed mixing and extrusion equipment

Appendix 7
LOCALLY AVAILABLE COMMERCIAL SHRIMP FEEDS

¹ cl = crude lipid; cp = crude protein; cf = crude fibre. All analysis are as-fed
² Substantial sales, cash, and promotional discounts are available on the list prices of some manufacturers. Discounts can total as much as 17%
³ PL = post larval; ST = starter; GR = grower; FI = finisher; AD = adult; NA = not available

Appendix 8
TEST DIET FOR P5-P30 PENAEUS MONODON¹

¹ Ingredient cost (on dry matter basis; experimental diet M$ 1.8/kg; control (micro-encapsulated) diet M$ 47.6
² Method of preparation: Trash fish, mussel flesh, shrimp heads, and squid were minced before use. Whole trash fish was used. The soybean was reground locally. Dry ingredients were stored at ambient temperatures; fresh ingredients were kept frozen. Fishmeal, soybean meal, wheat flour and vitmix were weighed out and thoroughly mixed by hand. Cod liver oil and molasses were then weighed out and thoroughly mixed in. The wet ingredients (in this case trash fish, mussel, shrimp heads and squid) were then weighted and added. After thorough hand mixing, the diet was further thoroughly mixed by passing through the 8 mm and then the 5 mm dies of a table model Chinese (Sheng Yuan) meat mincer. The mixture was finally extruded through a 3 mm die plate. One third of the resultant noodles were frozen. The rest were sundried and then ground to form two fractions - a coarse crumble (from 1 up to a maximum of 4 × 2 mm) and fine crumble (from 0.25 to 1 mm). The diet was fed in three forms, therefore: moist pellet, dried coarse crumble and dried fine crumble. The two dried diets were kept at room temperature. The moist diet was stored frozen
³ Micro-encapsulated diet (Chow, 1984), modified by LKIM as shown above, but made by the technique suggested by the previous FAO consultant. The control diet was made up as necessary and kept refrigerated
⁴ “Modified vitmix” = 10 g Roche Vitamin Concentrate 428/3; 20 g Chinese Vitamin C; and 970 g wheat pollards. (The “Improved Vitmix” (Appendix 3) could be substituted for this later, at the same inclusion rate. Alternatively, use 0.1% Agrimate fish vitmix and 2.4% extra wheat pollards)
⁵ The “old” vitmix was used for the control. This consisted of 10 g Roche Vitamin Concentrate 428/3; 5 g Vitamin C; and 985 g wheat pollards

Appendix 9
1985 FEEDING TRIAL WITH EARLY POST-LARVAE OF P. MONODON

1. INTRODUCTION

Problems have been experienced at the site with the rearing of early shrimp post-larvae (up to P20). Survival rates have been poor and survivors unhealthy, when fed on the available commercial PL or Starter feeds. However, the results are reported to be better when the post-larvae are fed a modified version of the micro-encapsulated diet suggested by the previous FAO feeds consultant (Chow, 1984) for fish larvae, when fed in combination with a dry commercial feed. While the micro-encapsulated feed seems successful and is simple to make on a small scale, it is very expensive and may prove difficult to prepare and store in the quantities necessary when the hatchery is in full operation. The freeze-drying facilities suggested (Chow, 1984) as necessary to preserve large quantities of this feed are unavailable at the site.

An attempt was therefore made during this consultancy to formulate, make and test an alternative, cheaper, post-larval diet. This Appendix summarizes the experiment.

2. EXPERIMENTAL DESIGN

The experiment was conducted in eight 0.5 t plastic tanks kept outside the hatchery under shade. The tanks were aerated. Ambient salinity water was used; salinity was not measured because the same water supply was used for all tanks. Uneaten feed was removed each morning, followed by a water change. Records were kept of water changing operations and of daily observed mortalities amongst the uneaten feed; these are on file. Each tank was stocked (on 21 November = D1) with 1 500 P8 post-larval P. monodon acquired from a nearby DOF hatchery. (The animals had been received at age P4 and had been fed on the control diet alone before the experiment started.) A further 1 500 P1s were weighed in bulk on D1. On D1 all tanks were fed the micro-encapsulated control diet. The trial of the experimental diet commenced on D2 (P9). Tanks were numbered 13–20 and laid out in two groups of four. Tanks 13 and 19 were fed the control diet. The other six tanks were fed the three versions of the experimental diet. Appendix 8 gives details of the formulation and method of manufacture of these diets. The feed contained four sources of fresh marine protein. Tanks 14 and 20 received the moist form of the diet; tanks 15 and 17 received the experimental feed in the coarse (> 1 mm) form, while tanks 16 and 18 were fed the fine crumble (0.25–1 mm) form. Animals were fed three times per day, the first after the water change (10.00–11.30 h) and the others at approximately 15.00 h and 18.00 h. Records were kept of the amount of feed presented daily to each tank, but shrimp were fed to demand, not pre-determined quantities. Each tank contained a small feeding tray, suspended about 20 cm below the water surface. At first, the moist feed was pressed through a sieve during feeding but, because this caused water quality problems, from D6 onwards most moist feed was made into small balls before being put into the tank.

On D11, ten post-larvae were removed from each tank and measured for length; they were then restored to the tanks.

The experiment was terminated on D17 (7 December), by which time the animals were P24. All survivors were counted, the total biomass in each tank was weighed and ten post-larvae from each were measured for length.

3. RESULTS AND DISCUSSION

3.1 Physical Dietary Characteristics

All forms of the diet (moist, coarse crumbles and fine crumbles) had excellent water stability (> 24 h) though some soluble materials leached from the moist form. All feeds were soft in water. Water quality in the tanks fed the dry forms of the diet was as good as that of those fed the control but those fed the moist feed had to have greater water exchange. The diet contained small quantities of wheat flour and molasses but no expensive binders such as alginates. The stability of the moist form of the diet could be further improved in future batches if the noodles were sun-dried for two hours before storing in the freezer. As only a coffee grinder was available there was a considerable quantity of fines (particles < 0.25 mm) in the fine crumbles, which were sieved out before use.

3.2 Growth and Survival Rate

On D11, ten shrimps were removed from each tank for examination, after which they were returned to the tanks, live. All animals appeared healthy but those on the experimental diets appeared to be more consistant in size and better pigmented. The results of this examination are given in Table 1.

¹ C = control; W = moist form; DP = dry coarse crumble; DG = dry finecrumble

Although average total length by D11 was similar in all tanks, animal size was more consistent in those fed by the experimental feeds, particularly the fine crumble. Additionally, pigmentation was better and the animals looked plumper. For comparison, ten P15 animals from a batch received for pond stocking from the local DOF hatchery on D12 were also measured. Average length was 11.1 mm, but the size range was greater (10–13 mm). Four animals were 10 mm, 2 were 11 mm, 3 were 12 mm, and 1 was 13 mm. There was also great variation in “plumpness”. Their dietary regime was unknown but was suspected to be Artemia-based. The smaller size of the animals in the trial at P18 was probably due to the fact that from P4–8 they were only fed the control diet. Normally at the hatchery they are fed Artemia until P10.

On D17 (P24) all animals were counted and the total biomass was weighed for each replicate. A sample of 10 from each tank was measured. The results are presented in Table 2.

The survival rate in one tank (No. 17) was abnormal (5.2%). It is believed that this was due to operator error and the result, though recorded, has been excluded from the interpretation of the experiment. Survival in other tanks ranged from 30.2 to 73.4%. Overall survival for all treatments and tanks (including tank No. 17) was 48.5% from P8–P24. Survival in largescale post-larval production at Sg. Merbok is previously reported to have been around 25%. The control diet produced an average survival rate of 36.5%, while the experimental diet gave 67.5% survival when fed in moist form, 45.1% (one replicate only) when fed as a coarse crumble (“dry pellet”), and 64.9% when presented as a fine crumble (“dry ground”). The financial effect of these higher-than-normal survival rates is analysed in the conclusions of this Appendix.

Records of observed mortalities on a daily basis were also kept and are retained on file. These show that a total of 352 (out of a total animal loss of 6 180) were observed as dead during the daily removal of excess feed. Thus a further 5 828 were either cannibalized or lost during maintenance operations (probably about 600 of these were lost due to an unknown error in tank No. 17). The daily observed mortality record is interesting because it shows a marked increase in mortality rate from D12 onwards (one day after intermediate measurements on D11). Nearly 72% of all the observed mortalities occurred from D12–D17. There was no difference due to treatments but those tanks where water level was left low during sampling on D11 for the longest period had the highest mortality rate. This pattern in observed mortalities was also probably true of total mortalities although this cannot, of course, be proved.

Animal length did not increase as much as expected in any treatment from D11 (Table 1) to D17 (Table 2). It is believed that this is associated with the stress on D11, mentioned above, and perhaps with other unsatisfactory environmental factors (including the use of “old” water due to the hatchery being out of operation at the time). Despite this general performance, there are clear differences between the treatments. Animals fed the control diet averaged 10.05 mm on D17, while those fed the experimental diet achieved 10.3 mm on the moist form and 10.7 mm on both of the dry forms of the diet.

Measurements of average weight on D17 showed that animals fed the control diet averaged 2.45 mg each, while those fed the experimental diet were 3.2 mg on the moist form and 3.35 when fed either of the two dry forms of the feed.

The actual amount of feed used (presented, not necessarily eaten) is also summarized in Table 2. From this the weight of each feed used on a dry matter basis has been calculated. This shows that the coarse crumble (“dry pellet”) form of the diet seems more wasteful than the others. Much more of the experimental feeds were used than the control diet. To some extent this can be accounted for by the overfeeding of an unfamiliar diet and by differences in the number of survivors on the new diet. Using the dry weight cost of each of the two feed formulations, the total cost of feeding is also presented in Table 2. This shows that feeding costs were much less when the experimental diet was used. Converted to a larger scale basis, these results show (Table 2) that the feeding costs to rear 1 million P24 post-larvae (from P8 onwards) would be M$ 1.192 if the control, micro-encapsulated, feed were used. If the experimental diet were used, the feeding costs would be reduced to M$ 107, M$ 296, or M$ 158/million for the moist, coarse crumble (“dry pellet”) or fine crumble (“dry ground”) forms respectively. It may be possible to reduce feeding costs still further if overfeeding can be controlled.

4. CONCLUSIONS

This trial has adequately demonstrated the superiority of the new post-larval feed formulation (see Appendix 8 for formulation) over the old micro-encapsulated egg diet.

In summary, the advantages of the new diet are as follows:

1. The new diet is easier to make, from locally available materials, in large quantities than the egg diet.

2. The new diet is easier to store, especially in its dry forms

3. The new diet can be fed in moist or “dry ground” (fine crumble) form, with similar results.

4. Survival rate is 1.2 (“dry pellet”) - 1.8 (moist) times greater with the new diet than with the egg diet. This improvement in survival rate has a very great potential effect on farm profitability, a fact that has been further commented upon in section 3.3 of this report.

5. Both animal length and weight are greater when the animals are fed the new diet, particularly if fed in its dry forms. The new diet produces animals with better pigmentation, and which are more active, than those fed the egg diet. Their potential future survival in ponds is therefore greater.

6. For a 10 million post-larvae/year hatchery, it would be necessary to extrude about 500 kg of the moist diet (one batch in the equipment suggested in section 3.8 of this report) every 3 months. If dried, the final weight to be stored would be about 270 kg. If stored dry, the feed need only be made 4–6 times per year. If the micro-encapsulated diet were used, it would be necessary to make about 25 kg/week, for which refrigerated storage would be essential.

7. There would be very significant savings in using the new diet for large-scale post-larval production (in addition to the better survival rates to be achieved). For example, the raw material costs for post-larval feeding in a hatchery producing 10 million post-larvae/year (as is targeted for Sg. Merbok) would be M$ 1.580 if the fine crumble (“dry ground”) version of the new diet were used. If the micro-encapsulated egg diet were used, the raw material costs would be M$ 11.918, approximately 7.5 times more expensive.

8. The new diet is obviously successful from P9-P24. It could also be applied from P5-P8. A future experiment would determine the difference between feeding it alone to P5-P10 animals or in conjunction with Artemia. If it proves unnecessary to use Artemia to P10, this would result in further significant labour, equipment and financial savings.

One further comment is relevant. Environmental conditions for this experiment were not ideal. For example, it was necessary to use “old” aged water from a storage tank because the hatchery pumping/filtration system was closed at the time of the author's visit. The only tanks available were under shade outdoors; lighting conditions were not ideal. In addition, it was necessary to use an inadequate vitmix in the feed because no other was available on the site. Also, from the time the experimental animals were received from the DOF hatchery (P4) to first feeding of the trial diets (P9), all animals were fed only the control egg diet. No Artemia were presented. They may therefore have been stunted. It is therefore probable that even better results will be achievable with the new diet in the future. Provided that good quality ingredients are purchased, there should be no further problems with post-larval P. monodon feeds at Sg. Merbok; any poor survival rates in the future would be due to water management, water quality or disease factors, or to the general fitness of the P5 animals received into the post-larval rearing system.

¹ 1 500 were stocked per tank (total 12 000). Total survivors = 5 820 (48.5%)
² 1 replicate only; the other was discounted as a probable management error
³ Total biomass ÷ No. of animals. Average weight at start (P8) was 1.0 mg
⁴ Sample of ten animals

Appendix 10
POND GROW-OUT FEEDS FOR SHRIMP (P. MONODON)¹

Two alternative formulations for grow-out shrimp feeds are given below. Grow-out feed S2 is much more expensive than S3 because it contains mussel and squid, which are excellent ingredients from an attractant and a nutritional point of view. The raw material cost of either diet is very much less than the cost of the current Ang Hock feed and that of the more expensive version. S2 is only 47% of the discounted price for Gold Coin growers feed.

¹ These feeds should be mixed and extruded in the same way as the experimental post-larval shrimp diet (Appendix 8). The pellets should then be sun-dried. If dried to 10% moisture or less they could be stored for up to 2 months at ambient temperature. If fed in moist form they must be stored frozen or made daily

² See Appendix 15, note 4

Appendix 11
ASSESSMENT OF 1984 FEEDING TRIAL WITH SEABASS JUVENILES

1. INTRODUCTION

During the consultancy visit by Dr Kai Chow (29 February - 25 March 1984) an “artificial” semi-moist diet for seabass was formulated and made. A feeding trial was commenced during his consultancy and completed by counterpart staff after he departed. This appendix summarizes the experimental design, discusses the results and draws some conclusions.

2. EXPERIMENTAL DESIGN

Details of the experimental design are taken from the consultancy report (Chow, 1984) and discussions with counterpart staff. The experiment was conducted in eight 3 m diameter plastic pools in the hatchery. Each tank was stocked with 25 fish averaging about 38 g each. Three dietary formulations (Table 1) were tested against a control (trash fish). Each treatment was replicated. Though designed to last 3 months, the experiment was terminated after 28 days due to general disease/water quality problems in the hatchery. Animals were conditioned to the artificial feed (using feed No. 2) for 4 days prior to the trial. All feeds were presented in fixed, pre-determined quantities (Table 2) ranging from 50–73 g/day/pool for the formulated diets and 134–196 g/day/pool for the control trash fish.

3. RESULTS

The results of the experiment were summarized by counterpart staff in a report to LKIM in Kuala Lumpur but that report contained no discussion of the result, nor did it draw any conclusions from the experiment. The results are represented in Table 3.

4. DISCUSSION AND CONCLUSIONS

Although no written conclusions were given in the counterparts' report to LKIM, the verbal impression gained was that though the juvenile seabass accepted the “artificial” feeds well, the results were poor compared to those achieved on the trash fish control. The surviving replicate of the control produced an average weight gain of 41.4% in 28 days, while the best two results achieved by the trial diets were 17.6% (Diet 1) and 15.8% (Diet 2). At first sight this result discouraged further experimentation. However, there are other ways of interpreting the results and several lessons to be learnt from this trial.

First, it is clear that survival rate on the control diet was very poor (44%) compared to that on the test diets (60–96%). The result was that although the survivors on the control diet gained 41.4% in weight, the total weight of fish in the only tank left on this treatment was 340 g less at the end of the experiment than at the beginning. Total weight of fish also decreased in most of those tanks fed the test diets but, with one exception (Diet 3, replicate B), to a lesser extent. In one case (Diet 2, replicate B) there was an increase in the total weight of fish, the only tank in which this occurred in the whole experiment. Although the increase in individual weight was less than that on the control, the overall performance of fish fed Diet 2 was rather better than that of fish fed any other feed in the experiment. None of the results (including the control) can be regarded as satisfactory.

Which lessons and conclusions can be drawn from this preliminary experiment? They can be listed as follows:

1. Survival of fish fed trash fish was poor.

2. The demise of the second control replicate should have been recorded. Negative results are as valuable as positive ones and results of both successes and failures are required to draw proper conclusions from an experiment.

3. Feed was rigidly presented according to a pre-determined schedule. No attempt was made to adjust feeding rate according to demand, or to compensate for the reduced number of fish caused by mortality losses. The result was gross overfeeding which may have contributed to poor water quality and to the generally poor survival rates.

4. No record exists of whether the losses were in fact mortalities or simply “losses”. There is an important distinction between these two and absence of this information reduces the value of the experiment.

5. The result summary produced after Dr Chow left and the experiment was over, contained no comments, discussions or conclusions. If it had, the fact that Diet 2 performed rather better than the other artificial diets (taking into account average and total weight gains and the fact that one replicate exhibited the only positive feed conversion efficiency) would have emerged. This fact should have been a prelude to an examination of the possible reasons for this better performance.

6. Having drawn the conclusion that Diet 2 was superior to Diets 1 and 3 it is worth looking at the composition of the diets again (Table 1) to see why this may be so. Here it will be observed immediately that Diet 2 had a higher protein content than Diets 1 and 3, as it contained more fish and soybean meals. Thus a general conclusion could have been drawn and Diets 1 and 3 were deficient in protein content for the fish under trial. It also immediately raises a new question to be answered in a future experiment. What happens if the dietary protein level is increased above that in Diet 2? Would that still result in better performance?

7. The above observations and conclusions have been taken into account in designing the current seabass experiment although, unfortunately, it was not possible to use the same sized fish. The current experiment (Appendix 13) has been conducted with two batches of fish with an average starting weight of 0.8 g and 3.2 g respectively, instead of the 38 g animals used earlier.

8. This review of the 1984 seabass experiment has been used to illustrate experimental procedure, together with the experimental protocols for the current experiments, during tutorials with Sg. Merbok staff.

¹ Prepared as described in Table 8 (Chow, 1984)
² This diet also used for conditioning fish prior to experiment
³ See footnote in Table 1 (Chow, 1984) for composition

¹ Experiment being conducted in 3 m diameter plastic pools; 25 fish averaging 40 g in each replicate pool; 2 replicates per diet
² Day started: 20 March 1984; day first weighed 25 March 1984
³ Weights calculated on estimated daily feed intake of 2.5 percent (dry feed basis) body weight, and average feed conversion (dry basis) during period of 2.0
⁴ Feed given twice daily: first after water change at 11.00 h and second feeding at 15.00 h
⁵ Minced trash fish without vitamin supplementation
⁶ See Table 9 (Chow, 1984) for details

¹ No record of the data from the second replicate of the control exists. It is believed that all the fish died due to unknown circumstances

Appendix 12
TEST DIET FOR SEABASS FINGERLINGS¹

¹ Ingredient cost of diets (dry matter basis) was M$ 0.88/kg for the test diet and M$ 2.32 for the control diet. If whole trash fish had been used for the control, the dry matter cost would have been M$ 1.28/kg
² Method of preparation: see Appendix 8. The diet was fed in moist form, crumbled by hand during feeding. The experimental diet was kept frozen, except during the day of feeding, when it was kept at the cage site under ice
³ See Appendix 8, note 4, for composition. In future, the “Improved Vitmix” (see Appendix 3) could be substituted at the same inclusion rate, when available. Alternatively, the Agrimate fish vitmix could be used until the better one is made, at 0.1% in the moist diet. The weight should be made up with an extra 2.4% of wheat pollards

Note: This formulation (SB No. 1) is now suggested as a grow-out feed. Improved formulations for fingerling rations are given in Appendix 14

Appendix 13
1985 FEEDING TRIAL WITH FINGERLING SEABASS

1. INTRODUCTION

The rearing of small fingerling seabass is difficult. Among the problems experienced at the site (and elsewhere) is the availability of a suitable diet. At present young seabass are weaned from live foods onto Acetes spp. and from that to trash fish. Trash fish is not always available and is (on a dry matter basis) an expensive feed, especially (as is normally the case) if it is filleted before use.

An attempt was, therefore, made during this consultancy to formulate, make and test all alternative diet for feeding seabass. This feed was based on a reduced content of whole trash fish and was fed in moist extruded form. Experience gained from the analysis of an earlier experiment with seabass juveniles (Appendix 11) and from a review of others' experience with this and related species (New, 1985) was used to design the diet. The diet contained 1% squid as an attractant and had a higher protein and lipid level than that used in the 1984 experimental juvenile feed. Another difference between the diets was that trash fish has been used to supply fresh marine protein and moisture content in the current diet, whereas the earlier diet contained only dry ingredients, moistened with water before extrusion.

2. EXPERIMENTAL DESIGN

The experiment was conducted in eight nursery cages in Sg. Merbok, each measuring 1.275 m³. The cages were numbered 1–8. The cages were rafted in blocks of six but only the four corner cages of two rafts were used for the experiment, to ensure similar environmental conditions. Each of cages 1–4 was stocked with 200 1-inch fingerlings while each of cages 5–8 was stock with 100 3-inch fingerlings. Fifty fish of each size were stocked on Day 1 (20 November); on this day all the fish were fed trash fish. On day 2, feeding of the experimental feed commenced. No attempt was made to wean the fish on to the compounded diet, which was accepted immediately. Cages 1, 4, 5 and 8 were fed the control (filleted trash fish), while cages 2, 3, 6 and 7 were fed the test diet. The formulation and method of preparation of the test diet is given in Appendix 12. Fish were fed to satiation three times per day. Records were kept of the total daily amount of feed presented daily; any observed mortalities were also recorded. On Day 11, a sample of 10 fish from each cage was weighed and returned to the respective cage. The experiment was terminated on Day 21(10 December), as planned, when all survivors were counted and the total biomass in each cage was measured.

3. RESULTS AND DISCUSSION

3.1 PHYSICAL DIETARY CHARACTERISTICS

The experimental diet had good water stability (> 3 tours), which was satisfactory for feeding fish. It contained no special binders. The feed was readily accepted by the seabass fingerlings from Day 1 onwards.

3.2 GROWTH AND SURVIVAL RATE

On Day 11, a sample of 10 fish from each cage was weighed. On Day 21 the total weight of fish in each cage was measured and all animals were counted. The results of these measurements are given in Table 1.

By Day 11, the 1-inch fish had increased in weight from 87.5 to 150.0% while the 3-inch fingerlings had increased by 87.5–118.8%. No significant difference was noted between the growth rate on trash fish or the moist diet at that stage.

Table 2 summarizes the final survival and growth rates, while feed data are presented in Table 3.

Unlike the 1984 seabass trial (Appendix 11), survival and growth rates were good, resulting in positive increases in tank biomass in this trial. Individual cage survival rates after 20 days ranged from 64% to 94.5% for the 1-inch fish and 64% to 97% for the 3-inch fish. When fed filleted trash fish, survival averaged 94% for the 1-inch fish and 80.5% for the 3-inch fish. On the moist diet fish averaged 76.5% (1-inch) and 85% (3-inch) respectively.

Average growth rate was generally higher for fish fed filleted trash fish. The 1-inch fish (start weight 0.8g) averaged 3.2 g on Day 21 when fed the moist diet and 3.8 g when fed trash fish. The 3-inch fish achieved 10.2 g and 12.6g on the moist diet and trash fish respectively. These increases in individual weight represent increases of + 378% (1-inch) and + 296% (3-inch) over 20 days when fed trash fish and + 298% (1-inch) and + 220% (3-inch) when fed the moist diet. These figures compare with + 41% on trash fish and - 23.2% to + 17.6% for the test diets fed to 38 g starting weight animals in the 28 day 1984 feeding trial (Appendix 11). The average length of the fish at the end of this 1985 trial was 6.25 mm and 9.2 mm on trash fish and 6.0 mm and 8.95 mm on the moist diet for the two size groups.

Individual cage biomass increased from 285 to 640 g (starting biomass 160 g) in 20 days for the smaller fish and from 390 g to 1 055 g (starting biomass 320 g) for the larger fish. On filleted trash fish, biomass increase averaged + 349% and + 226%, while on the test diet it was + 200% and + 173% for the 1-inch and 3-inch fish respectively. In the 1984 trial most tank biomass decreased over the 28 day period, with one exception (+ 90%).

AFCR¹ averaged 0.80 and 0.84 for the smaller and larger fish respectively when fed filleted trash fish, while the moist diet gave corresponding AFCRs of 1.86 and 1.13. Because of the high unit cost of filleted trash fish, the cost of feed per unit of fish produced was less when the moist diet was used. The cost of feeding the smaller animals was M$ 1.84/kg of weight increase for trash fish and M$ 1.63/kg for the moist diet. The difference was more extreme for the larger fish, where the trash fish diet cost M$ 1.94/kg while the moist diet regime only cost M$ 1.00/kg of weight increase.

¹ AFCR = Apparent feed conversion ratio on a dry matter basis

These results, in a preliminary trial of a moist feed formulation, are very favourable. The formulation (Appendix 12) is obviously much more appropriate than those tested in the 1984 trial (Chow, 1984; Appendix 11) but, at least for seabass of the sizes used (0.8–3.2 g starting weights) was not as effective as filleted trash fish.

Nevertheless, the major achievement is a moist diet which is accepted by 1 to 3 inch fingerlings without weaning, which produces good survival (76.5–85.0%) and growth rates (fish increased by a factor of 3.2–4.0 within 20 days). This shows that the current formulation would be an acceptable substitute for trash fish when this is unavailable. The formulation can probably be improved and the current partial success can be regarded as the expected stepping stone towards the ultimate definition of a compounded fingerlings feed which is equal, or superior to trash fish for seabass.

An interesting observation which can be made from the results, as presented in Tables 2 and 3, is that the comparative overall performance of the moist diet compared to trash fish is better for the larger than for the smaller animals. The survival rate appears to be better on the moist diet for the larger animals. While the gain in cage biomass on trash fish is 1.75 times greater than on the moist diet for the smaller animals, it is only 1.31 times greater for the larger fish. While the AFCR on trash fish remains the same for both size groups the AFCR of the moist diet was much more favourable (1.13:1 compared to 1.85:1) when fed to the larger fish. Similarly feed costs per unit of weight gain were much less for the moist diet for larger fish, whereas the cost of using trash fish was more than that for the smaller fish.

These results lead to the speculation that the formulation (Appendix 12) may be suitable as a grow-out feed, in spite of the fact that it does not fully meet the requirements of young fingerlings. It is suggested that it be tried as a grow-out feed (the intention of the counterpart staff is to continue the experiment until market size using the same feed). Two alternative and improved fingerling test diets for future trials with seabass are given in Appendix 14.

4. CONCLUSIONS

The trial has demonstrated that small (0.8 or 3.2 g) seabass juveniles will immediately accept a moist feed formulation which contains fresh ingredients and that they will grow and survive well on it. Performance, however, particularly of the smaller size group, is not so good as when filleted trash fish is used. The trial indicates that the diet formulated in Appendix 12 is not completely adequate for fingerlings but that it may suffice as a grow-out feed (yet to be tested). A summary of the conclusions drawn from this experiment, together with some suggestions is given below:

1. Young fingerlings seabass will accept a moist diet without weaning provided that it includes suitable ingredients.

2. The use of the test fingerling diet produced average survival rates of 76.5–85.0%, increases in individual weight of 220–298%, and increases in cage biomass of 173–200% within a 20-day period.

3. The AFCR of the moist diet is acceptable, particularly for the larger size group (1.13:1).

4. The feeding cost per unit of fish produced is less when the moist diet is used instead of filleted trash fish.

5. Trash fish gave superior survival rates for the smaller size group and better growth rates for both groups.

6. The performance of the moist diet appears to improve as the seabass became larger.

7. Thus, while the moist diet is partially successful, and much superior to those tested in 1984, it is not good enough to recommend its routine use as a substitute for filleted trash fish. It would, however, be an acceptable short-term substitute in times of trash fish supply problems.

8. It is suggested that the test formulation be tested as a grow-out feed. Meanwhile two improved formulae for testing young juveniles are given in Appendix 14.

9. Future seabass trials, including those mentioned above, are suggested in section 4 of this report.

¹ TF = trash fish; MD = moist diet
² Average weight of 50 fish
³ Average weight of 10 fish
⁴ Total biomass ÷ No. of fish
⁵ Average of sample of 10 fish

¹ AFCR = Apparent feed conversion ration on a dry matter basis

Appendix 14
NEW TEST DIETS FOR SEABASS FINGERLINGS¹

¹ The diets are designed as a result of the 1985 trial using feed SB No. 1 (formula in Appendix 12; experimental results in Appendix 13). It is suggested that diet SB No. 1 be retained for trial as a grow-out diet
² From Soon Soon
³ See Appendix 12, note 3

Appendix 15
EXPERIMENTAL SEMI-MOIST MARINE FINFISH BROODSTOCK AND SHRIMP MATURATION FEEDS

1. SEMI-MOIST MARINE FINFISH BROODSTOCK DIET¹

The following diet is suggested for feeding to broodstock fish for the two monsths prior to normal spawning time:

¹ Diet to be made as shown in Appendix 8, note 2, and fed in moist form
² Use the “Improved Vitmix” prepared as shown in Appendix 3. Until this is available, the less suitable Agrimate premix could be used in preference to the old (current) premix. If the Agrimate premix is used (see Appendix 2 and Section 4.1 of this report), it should be included and 0.1% in the moist diet. The missing 2.4% of the diet should be replaced with more wheat pollards. If neither of these premixes is available, use the “Modified Vitmix” (see Appendix 8, note 4)

2. SEMI-MOIST SHRIMP MATURATION DIET¹

The following diet is suggested for shrimp maturation stock:

¹ Diet to be made as shown in Appendix 8, note 2. After extrusion, it should be formed into balls by hand before feeding
² Or growers feed from another reputable company, such as President Enterprises or Hanaqua. Pellets to be broken up into powder by passage through a large diameter die plate of the mincer
³ If available, replace 1% of each of these ingredients (total 4%) with blood worms
⁴ This is an expensive diet but only small quantities will be needed for maturation stock. A suitable diet is one of the keys to successful maturation

Appendix 16
SUGGESTED FORMULATION FOR DRY FINGERLING AND GROW-OUT FEEDS FOR TILAPIA REARED IN CAGES¹

The following formulae are based on locally available ingredients:

¹ Diet to be made as shown in Appendix 8, note 2, sun-dried and fed as a dry pellet
² 0.5–35 g
³ 35 g - market size
⁴ Use the “Improved Vitmix” prepared as shown in Appendix 3. Until this is available, the less suitable Agrimate premix should be used in preference to the old (current) premix. If the Agrimate premix is used (see Appendix 2 and Section 4.1 of this report), it should be included at 0.1% in the moist diet. The missing 2.4% of the diet should be replaced by a further 2.4% of wheat pollards. If neither of these premixes is available, use the “modified vitmix” (see Appendix 8, note 4)

Appendix 17
EXAMPLES OF MAJOR EQUIPMENT¹ FOR THE MANUFACTURE OF UP TO 6 t/DAY OF MOIST FEED OR 3 t/DAY OF DRY FEED

¹ Mention of specific branded products in this report does not imply that other alternatives are unavailable or unsuitable
² Supplier: Yashima Bussan (Japan) - see Appendix 6 for address

¹ These methods favoured for large batch production because of the unreliability of sunshine and its possible effect on vitamin potency

Note: Brochures on the suggested equipment are being sent to the project, addressed to Salehuddin bin Abu Bakar, c/o LKIM, Kuala Lumpur

Appendix 18
A FEW SIMPLE DO'S AND DO NOT'S FOR EXPERIMENTAL WORK

DO

1. Test only one, or at the most, two theories in each experiment.

2. Keep oil experimental parameters (e.g., environmental conditions) the same for each treatment.

3. Replicate: two, or preferably three, replicates per treatment.

4. Keep full and adequate daily records. This is essential for subsequent interpretation of the results.

5. Plan the experiment thoroughly. Write down the experimental plan and discuss it with your colleagues (critical analysis) before you start the experiment.

6. Write a report on the experimental results which includes your deductions, conclusions, and a consideration of “where do we go from here?”. Discuss the results with your colleagues.

7. Remember that apparently negative results are often as valuable as positive ones.

8. Use standardized techniques for size measurement (weighing; length determination).

9. Have the same person take all size measurements.

DO NOT

1. Try to answer too many questions in one experiment.

2. Discard any results. Try to explain any obvious discrepancies.

3. Use animals from the control to “top-up” the number of animals on other treatments.

4. Replace any mortalities during the experiment.

5. Obscure or omit apparently negative or conflicting results.

6. Forget to check that all planned jobs have been completed before you leave the experimental site each day. This includes checks on facility security (e.g., cage netting; aeration) and on record keeping.

7. Change the experimental protocol part way through an experiment. It is better to stop the experiment and analyse the results and then start a new experiment to answer the new questions which have arisen.

8. Do not entrust work on a measurement day to someone else if you are otherwise busy or away from work through sickness. It is better to prolong the experiment by another day or two if you are unable to do the work yourself (see DO No. 9).

9. Do not allow anyone to interrupt you when such interruptions may endanger the success of the experiment. In biological work, the animals come FIRST; telephone calls and meetings are less important.

Appendix 19
BASIC ANALYTICAL INFORMATION ON LOCALLY AVAILABLE INGREDIENTS AND AN EXAMPLE OF INGREDIENT SUBSTITUTION

¹ These are assumed levels. They should be replaced with actualanalytical data when these become available
² The figure in parentheses refers to true protein

The following is an example of the effect of ingredient substitution. In this example, it is assumed that you wish to substitute Peruvian fishmeal for Thai fishmeal. The Diet No. S2 (pond grow-out feed for shrimp (see Appendix 10) is used for the example.

The calculated analysis for iet S2 on a dry matter basis is 8% oil and 36% protein. These levels should be retained as closely as possible, in the modified formulation. In the author's original formulation, worked out from the data given in the table above, the moisture content of the moist form of the diet was 43.3%. Thus the oil and protein levels in the moist form (as made) formulation can be calculated by multiplying by 56.7 and dividing by 100.0. The result is a made oil level of 4.5% and a protein level of 20.4%.

To those figures, an 8% inclusion level of Thai fish meal has contributed 0.6% oil and 4.4% protein. A straight substitution of peruvian fishmeal for Thai fishmeal would result in a contribution to the total diet of 0.4% oil and 5.2% of protein. Thus the final product would be 0.2% less in oil and 0.8% more in protein than the original formulation. The way to rebalance the diet is as follows:

The peruvian fishmeal should be included at a lower rate. If included at 6.8% instead of 8%, it will contribute 4.4% of protein to the diet, the same as the Thai fishmeal. However, it will only contribute 0.3% of oil; this is 0.3% less marine lipid than before. To compensate for this, the inclusion rate of cod liver oil in the diet must be raised by 0.3% to 2.8%. There is now 1.2% less fishmeal and 0.3% more cod liver oil in the formulation, a total of 0.9% less. This can be made up by increasing the quantity of wheat flour from 5.0% to 5.9%. This will in turn have a slight effect on the oil and protein levels in the final diet but it is so small it can be ignored. So the final formulation would have 6.8% Peruvian fishmeal, 2.8% cod liver oil, and 5.9% of wheat flour, instead of 8.0% Thai fishmeal, 2.5% of cod liver oil, and 5.0% of wheat flour. The inclusion rates of all other ingredients would remain the same. The resultant analysis of the final product would be the same. The effect on the cost of the diet could be examined as follows:

It can be seen that substitution of peruvian fishmeal for Thai fish meal would increase the “as made” cost of the feed by M$ 5.25 per ton. On a dry matter basis, this would increase the feed cost by M$ 9.26 per ton, or M$ 0.009/kg. In view of the better quality of the new fishmeal, this slight cost increase would be warrented. If it gave better results when fed, the cost of feed per unit of shrimp produced would still be less than with the original diet.

It is hoped that the above simple example of ingredient substitution will serve as a reminder to project staff of how this can be done.

Appendix 20
RECORD OF CONSULTANT'S DAILY ACTIVITIES

Appendix 21
LIST OF PERSONS MET