FI:DP/CPR/88/077 Field Document 7 December 1990

PEOPLE'S REPUBLIC OF CHINA

A report prepared for the project
Fisheries Development in Qinghai Province

by

David Edwards
CTA/Coldwater fish culture specialist

This report was prepared during the course of the project identified on the title page. The conclusions and recommendations given in the report are those considered appropriate at the time of its preparation. They may be modified in the light of further knowledge gained at subsequent stages of the project.

The designations employed and the presentation of the material in this document do not imply the expression of any opinion whatsoever on the part of the United Nations or the Food and Agriculture Organization of the United Nations concerning the legal or constitutional status of any country, territory or sea area, or concerning the delimitation of frontiers.

FOOD AND AGRICULTURE ORGANIZATION OF THE UNITED NATIONS
Rome, 1990

Hyperlinks to non-FAO Internet sites do not imply any official endorsement of or responsibility for the opinions, ideas, data or products presented at these locations, or guarantee the validity of the information provided. The sole purpose of links to non-FAO sites is to indicate further information available on related topics.

This electronic document has been scanned using optical character recognition (OCR) software. FAO declines all responsibility for any discrepancies that may exist between the present document and its original printed version.

TABLE OF CONTENTS

1. TERMS OF REFERENCE

2. CONSULTANTS' VISITS

3. NATIONAL PROJECTS STAFF

4. NATIONAL FINANCIAL SUPPORT

5. TRIPARTITE REVIEW MEETING

6. FELLOWSHIPS AND STUDY TOURS

6.1 Fellowships
6.2 Study tours

7. TROUT FARM DESIGN, CONSTRUCTION AND OPERATIONS

7.1 Rainbow trout hatchery and fingerling production

7.1.1 Construction work
7.1.2 Water supplies
7.1.3 Equipment delivery and installation
7.1.4 Operating plan
7.1.5 Agreement with Local Government

7.2 On-growing units

7.2.1 Concrete raceways
7.2.2 Floating cage unit
7.2.3 Earth ponds

8. NAKED CARP HATCHERY PLANNING AND CONSTRUCTION

8.1 Data collection and hatchery planning
8.2 Hatchery construction
8.3 Operational plan for the naked carp hatchery

9. RAINBOW TROUT FEEDS

10. EQUIPMENT PROCUREMENT

11. PROGRESS AND PROBLEMS

Appendix 1 CURRICULUM OF THE FELLOWSHIP IN TROUT CULTURE

Appendix 2 PROGRAMME OF THE FRESHWATER FISHERY MANAGEMENT COURSE FOR STUDY TOUR CANDIDATES

Appendix 3 AGREEMENT BETWEEN THE BUREAU OF AQUATIC PRODUCTS, DEPARTMENT OF AGRICULTURE, QINGHAI PROVINCE GOVERNMENT, AND THE QINGHAI PROVINCE MIXED FEEDS PROCESSING FACTORY

1. TERMS OF REFERENCE

The Chief Technical Adviser/Coldwater Fish Culture Specialist consultant to project CPR/88/077 carried out his second mission from 1 May to 4 November 1990. His terms of reference for the four-year project are:

• to coordinate the activities of other consultants in collaboration with the National Project Director (NPD);

• to prepare progress reports for consideration by the tripartite review meetings and write the project terminal report;

• to be responsible for planning, building and operating the trout farms and the experimental naked carp hatchery, working in close cooperation with the NPD and consultants specialized in fish feed technology, aquaculture engineering, etc.;

• to take a major part in the in-service training of national staff.

2. CONSULTANTS' VISITS

All consultancies planned for 1990 were fielded on schedule. In addition to the CTA/Coldwater Fish Culture Specialist, four other experts were assigned to the project for periods during the year as follows:

Dr I. Dunn, Specialist in Lake Fisheries Management

This consultant is responsible for the Qinghai Lake investigation which aims to produce a scientifically-informed management strategy for the naked carp fishery in this water body, the most important capture fishery in Qinghai Province. During 1990, Dr Dunn divided his assignment into two missions, the first from 26 April to 2 July and the second from 10 August to 8 October. He plans to return to Qinghai during March-April 1991, in time to begin the season's work as soon as the winter ice cover on the lake breaks.

Mr Colin Smith, Consultant in Limnology

He worked in Qinghai from 3 May until 1 June 1990. Mr Smith's brief was to commission equipment ordered for physical, chemical and biological sampling work on Qinghai Lake, and to train counterpart staff in its use. Since some important items of equipment had not arrived at the project site in Xining by the end of Mr Smith's assignment, it may be necessary for him or another limnologist to return in 1991 to complete his work.

Dr James Muir, Aquaculture Engineering Consultant

He visited Qinghai from 17 September to 13 October 1990, replacing Dr L. Varadi, who was unavailable this year. His terms of reference comprised two main parts:

1. to advise the CTA on engineering aspects of design and construction of the experimental naked carp hatchery and the trout hatchery

2. to recommend possible solutions to the problem of excessive water fluctuations in the naked carp spawning streams around Qinghai Lake, which is reported to have adverse effects on recruitment of juveniles into the fishery.

Dr Albert Tacon, Fish Feed Specialist

Dr Tacon stayed in Qinghai from 2 October until 1 November 1990. Working closely with the CTA/Coldwater Fish Culture Consultant, he assisted in the formulation of good quality feeds for farmed trout, using locally available materials, and advised on feeding strategy for juvenile naked carp. The feeds will be required when the projects' fish farms come into full operation during 1991 and 1992.

All the above consultants have submitted reports and their findings and recommendations are therefore not repeated here.

3. NATIONAL PROJECT STAFF

Staffing of the Qinghai Lake investigation team remained little changed from 1989, but during 1990 the Lake Management Consultant (Dr Dunn) expressed serious discontent over the level of competence and motivation of the group (Field Document 4). Changes will therefore be made to the leadership of the team before the 1991 working season begins (see section 11 of this report).

Regarding fish culture, several staff members were re-assigned to other duties during 1990, with the result that the fish farming teams within the Bureau of Aquatic Products was effectively reduced to two full-time active members, Mr Zhao Yimin and Mr Wong Guojia. Mr Zhao assumed the responsibilities of team leader.

As a result of these changes, it is regrettable that only one of the four trainees sent overseas on fellowships to study trout culture early in 1990 (Mr Zhao) was still working full-time with the group by summer the same year.

During the detailed planning and construction of fish farms undertaken in 1990, this reduced level of staffing has been sufficient. However, when the fish farms begin operating in spring 1991, more personnel will be needed. The CTA is confident that adequate staff will be provided both by the Bureau of Aquatic Products and the local authorities at the various fish farm sites, but it is unfortunate that most of the staff specially trained by the project during 1989 and 1990 both in and outside China are unlikely to be among them. It will thus be necessary to hold a second short training course on trout culture in Xining during 1991 to give newcomers to the project a theoretical background in the subject. Practical skills will be taught by the CTA via continuous on-the-job training throughout the project period.

4. NATIONAL FINANCIAL SUPPORT

During 1990, the Government made available sufficient funds to finance the building of the trout and naked carp hatcheries, and to equip and operate boats and vehicles necessary for project activities. Suitable working accommodation was provided for project staff. No delays in project activities were therefore attributable to lack of Government support.

5. TRIPARTITE REVIEW MEETING

The first tripartite review meeting for the project was held in Xining, Qinghai Province, on 10 June 1990. The meeting formed part of an itinerary of visits to projects in Gansu and Qinghai Provinces undertaken by a delegation from Beijing comprising two representatives from the Ministry of Agriculture, two from CICITE, two from UNDP and two from FAO. Prior to the tripartite meeting, the group toured the trout hatchery at Nanmenxia Reservoir, the fishery base and Buha River on Qinghai Lake, and the project offices and laboratory in Xining.

A Project Performance Evaluation Report was written by the CTA and circulated to all participants in advance of the meeting. In addition, the NPD's report was distributed and read during the meeting.

Minutes of the meeting were circulated to all concerned and are not reiterated here. The meeting concluded that the implementation of project CPR/88/077 to the date of the tripartite meeting was satisfactory.

6. FELLOWSHIPS AND STUDY TOURS

6.1 FELLOWSHIPS

Overseas fellowships in rainbow trout culture were undertaken by four national project staff for three months (7 January-7 April 1990). Messrs Ying Baicai, Yang Hongzhi, Zhao Yimin and Miss Lin Li were assigned to the Institute of Aquaculture of the University of Stirling in Scotland, United Kingdom.

Much of the trainees' time was spent in practical instruction and work at the Institute's commercial hatchery and pond farm at Howietoun, and at the floating cage facility operated by the University. Their work practice included the care of trout eggs, rearing of fry and fingerlings, transport of fish to floating cages, and on-growing to market size. Techniques for prophylactic treatments against disease, fish grading, handling and harvesting were covered.

The trainees participated in formal lectures and practical classes in fish disease diagnosis and control at the University, and attended the annual Scottish fish farming conference and exhibition held in Edinburgh during February 1990. They also visited several commercial salmon and trout farms in northern and central Scotland. Tours to scientific laboratories engaged in salmonid farming research and development were undertaken. Institutions visited included the Marine Laboratories and Torry Research Institute in Aberdeen, Dunstaffnage Marine Research Laboratory in Oban, and Heriot-Watt University of Edinburgh.

A full itinerary for the fellowship is attached as Appendix 1.

Reports produced by the trainees following their return home indicated that they were satisfied with the training received, and felt it would be useful to them in their future work in China.

6.2 STUDY TOURS

A study tour in Lake Fishery Management has been organized for six project team members: Mr Qiu Benchen (NPD), Mr Wong Jilin (leader, lake study group), Mr Yang Hongzhi (deputy leader, lake study group), Mr Da Renqing, Mr Chen Yuan, and Mr Zhao Yimin (interpreter).

The group plans to leave China on 11 November 1990, and spend the first 19 days of their tour receiving formal theoretical and practical instruction at the Humberside International Fisheries Institute, Hull, United Kingdom. The programme for this course is attached as Appendix 2. Field visits to important fishery and fish processing operations in the area will also be arranged during this period.

Following the formal training course, the group will make short visits of approximately four days duration to the Netherlands and Czechoslovakia where they will observe the techniques of lake fishery management employed in those countries. From Czechoslovakia the group will divide into two. The NPD, Mr Qiu Benchen, together with interpreter Mr Zhao Yimin, will travel to Rome for about four days, where they will receive instruction in UNDP/FAO organization and procedures at FAO headquarters. The other four members of the group will return to Beijing. All participants will be back in Qinghai by the end of 1990 ready to resume project work early in 1991.

Throughout the tour, the group will be accompanied and guided by Dr Colin Cowx, a fishery management specialist from the Humberside International Fisheries Institute.

Two of the candidates, Mr Zhao Yimin and Mr Yang Hongzhi, have passed the UNDP examination for proficiency in the English language. They will act as interpreters for their colleagues.

7. TROUT FARM DESIGN, CONSTRUCTION AND OPERATION

7.1 RAINBOW TROUT HATCHERY AND FINGERLING PRODUCTION UNIT

7.1.1 Construction Work

Diagrams of the proposed layout for the rainbow trout hatchery and fingerling unit at Nanmenxia Reservoir, Huzhu Country, were included in the CTA Edwards' 1989 report (Field Document 1). Details of available water supplies were also given. During 1990, the exact site for building was moved approximately 300 m following measurements of land levels, in order to facilitate the supply of water by gravity from the springs downstream of the dam. Staff and service accommodation, originally planned under the same roof as the fish hatchery, was instead detached for construction as a separate building. Otherwise, facilities were built in accordance with the 1989 plans.

The system for obtaining official planning consent for building in China required the commissioning of a specialist design department to draw up detailed blueprints of the project. The drawings should then be approved and signed by the Provincial Government's Planning Department, before being given to the builder so that work can commence. Had this procedure been strictly adhered to, no building work would have been accomplished during 1990.

Numerous delays occurred at the design stage, partly due to deliberate stoppages resulting from disputes over money, and partly because the draftsmen were unfamiliar with fish farm construction. They therefore made many errors which necessitated repeated re-drawings of the plans. Final plans were not ready until early September, and at the end of October it was still not known whether they had yet received official approval from the Planning Department. Due to the onset of heavy frosts, the outdoor building season in Qinghai comes to an end at this time.

Fortunately such delays were anticipated, and the builder agreed to go ahead with construction work on schedule in spring 1990 before receiving the finalized official plans. It was therefore possible to complete the hatchery building, a pipeline to supply the building with water from one the springs, and staff accommodation by the end of the 1990 building season. Excavation was also done in preparation for construction of concrete raceways next year. The facilities already completed are:

Hatchery building. The 40 × 15 m hatchery building is constructed to a high standard with brick walls, concrete roofs, and a concrete floor sloping to integral drainage channels. Single-phase electricity supply to fluorescent lights and power points is in place. Imported plastic pipework and fittings are on site ready for installation inside the building as soon as the incubators and fry tanks arrive (see section 7.1.3).

Water supply pipeline. A 3 m-diameter concreate sump was constructed to collect water from one of the three main springs below the Nanmenxia Dam. This spring is 3.6 m above the level of the hatchery floor. From the sump, a 30 cm (internal diameter) concrete pipe carries water approximately 340 m to the hatchery. The pipeline runs underground, emerging inside the building where it rises through a steel pipe and gate valve to discharge into the top of a 3 × 2 × 1 m deep fibreglass head tank 2.2 m above floor level. There is thus a difference in level of about 1.4 m between the two ends of the pipeline.

It was calculated that the 30 cm pipe installed will carry about 80 litres of water per second, more than enough to supply incubators and the fibreglass fry tanks and concrete raceways inside the hatchery building, where fry will be reared to an average size of about 1 g. Excess water from the head tank will be carried via an 8-inch plastic pipe to the eight 8 m-diameter fibreglass fingerling tanks which are to be installed outside the hatchery building. However, the existing 30 cm main supply pipe is inadequate to supply the total water requirement of the outside tanks, which is estimated at more than 100 1/sec. Consequently a second water supply must be installed to these tanks before they come into operation in summer 1991. This supply could come from one of two sources:

1. If the spring from which the existing 30 cm pipeline has been constructed proves to deliver sufficient excess water (see section 7.1.2), a second pipeline will be installed from it to the outside tanks. By the time this work can be carried out, in spring 1991, the hatchery building will already be in operation, with rainbow trout eggs and fry being cultured inside it. It is therefore essential that the new, additional pipeline be installed without any interference with the water supply delivered via the existing line. To facilitate this, a stub of 40 cm internal diameter pipe has been attached to the concrete sump at the spring. The stub is plugged to prevent leakage of water through it. It is now a straightforward matter to lay a complete 40 cm pipe from the stub to the outside tanks without interrupting flow through the 30 cm hatchery line.

2. If, on the other hand, the above spring proves to deliver insufficient water to supply a second pipeline, the additional water required will be drawn from a second spring, closer to the hatchery site. The second spring originates only about 2 m above the ground level of the hatchery site. It may therefore be necessary to set the bottoms of the outdoor tanks a little below ground level. This is common practice with this type of tank, and causes no problems provided there is sufficient slope away from the site for drainage purposes, which is the case at Nanmenxia.

Staff accommodation. A single-storey brick-built block of staff and service accommodation has been constructed adjacent to the hatchery building. The accommodation comprises bedsitting rooms, recreation room, kitchen, and storage and machine rooms. In all, the 26 × 5 m block houses eight rooms. The block is supplied with three-phase electricity for machine operation. Further storage and working accommodation will be built during 1991 if needed.

Outdoor concrete raceways. Excavation work has been completed for construction of 1 mu (666 m²) of concrete raceways. Aggregate and other materials are on site ready for concreate work to begin as soon as the winter frosts abate, i.e., around May 1991. The 30 × 3 m raceways will be used mainly for trout fingerling production and retention of the broodstock necessary to meet the target of making the hatchery self-sufficient in eggs after the project ends.

It was originally planned that a further 2 mu of raceways would be built at Nanmenxia during 1991, for growing-out fingerlings to market size. Through it is still possible that this work will go ahead, the plan is in some doubt due to disputes between the Bureau of Aquatic Products and the Huzhu Country Government (see section 7.1.5).

7.1.2 Water supplies

The plan described in the CTA's 1989 report (Field Document 1) envisaged the hatchery building and outdoor fingerling tanks being supplied with water collected from three springs below Nanmenxia Dam. The springs are the result of uncontrollable seepage of water beneath the earth dam. Their water is filtered as it passes through the ground, making it at all times clear and of suitable quality for supporting rainbow trout eggs, fry and fingerlings. Historical records collected over many years by the Department of Water Conservancy, which owns the dam, had measured the normal total flow from the three springs at about 0.5 m³/sec. This would be more than enough to supply not only the hatchery and fingerling farm, but also an on-growing unit producing about 25 t of market-sized trout per year. The spring from which water has been piped to the hatchery building was alone estimated to deliver an average of 0.2 m³/sec, enough for the hatchery building and the 8 m outdoor fingerling tanks.

However, the flow rates from the springs are highly dependent on the head of water held behind the dam. During the past two summers, repair work has been carried out on the dam to stabilize it against possible structural collapse. While the repair work was being done, the reservoir behind the dam was kept empty. Consequently flows from all the springs were much reduced, although none of them stopped flowing completely. Repair work is now finished, and the dam will be allowed to fill during winter. By the time the hatchery begins operation in March 1991 the reservoir will be full. Only then can project staff be certain about the quantities of water the springs will deliver, since it is possible that the repair work itself may reduce seepage under the dam. For this reason, final decisions about water supplies to the outdoor fingerling tanks and raceways cannot be taken until early 1991.

If spring flows prove adequate, the 1 mu of concrete raceways currently under construction can be supplied with clean filtered water from these. Alternatively, a reliable supply of water can be obtained by laying a pipeline from the dam's discharge canal to the fish farm site. This water is of a more variable quality than the spring sources, sometimes becoming a little turbid during periods of heavy rain. Filtration might therefore be necessary for supplies to trout fry, but the water quality is adequate for on-growing fish.

7.1.3 Equipment delivery and installation

Completion of the hatchery building now awaits delivery of incubators, fry and fingerling tanks. These are finally being shipped from Europe, following lengthy delays in requisitioning and ordering procedures via FAO Rome (see section 10). The absence of this equipment on site has prevented project staff completing the hatchery during 1990 as required by the project work plan. However, provided delivery is made by the end of 1990, there will still be time to install the equipment before the importation of rainbow trout eggs early in 1991. No permanent delay in production of projects outputs is therefore anticipated.

7.1.4 Operating plan

One million eyed rainbow trout eggs have been ordered by FPO from a supplier in Denmark. They will be delivered by airfreight to Beijing during March 1991, promptly transferred to a connecting flight to Lanzhou, then carried by car to the Nanmenxia hatchery. The project document requires the incubation of only 700 000 eggs. The increase in number ordered reflects the rising interest of local government departments and companies in producing trout under guidance from the project. Special interest has been shown by the Longyangxia Hydro-Power Company (see sections 7.2.1 and 7.2.2). This additional interest substantially increases the expected demand for trout fingerlings in Qinghai Province during the project period.

The eggs will be hatched in 13 “California-type” incubators. Depending on water temperature, hatch is expected sometime during April 1991. The resulting fry will be reared to an average size of about 1 g in 20 indoor 2 × 2 m fibreglass fry tanks, and in specially constructed internal concrete raceways. First feeding is expected in May 1991, and it is hoped fry will reach 1 g during June. They will then be transferred to eight 8 m-diameter fibreglass tanks outside the hatchery building, for rearing to fingerling size. By summer 1991, one mu of outdoor concrete raceways will also have been constructed for fingerling production. These extra holding units are required to accommodate the increased numbers of fingerlings expected as a result of importing extra eggs and for retention of broodstock.

At an average of 20–50 g, fingerlings will be transferred by truck to the on-growing units (see section 7.2) where they will be reared to a market size averaging 250 g. According to the growth rates achieved and the weather conditions at the recipient sites, first transfers may be made any time between autumn 1991 and spring 1992.

7.1.5 Agreement with local government

Difficulties have arisen over the contract made between the Bureau of Aquatic Products and the Huzhu Country Government, owners of the Nanmenxia hatchery site, necessitating changes to the draft agreement detailed in CTA Edwards' 1989 report (Field Document 1). Though much time and effort are expended on meetings at which such agreements are negotiated, it is regrettably common that the parties involved subsequently fail to adhere to the terms of the contracts they have signed, and often hold-ups were caused to building work pending discussion of new demands. As a result, it was necessary to draft a new agreement, whereby the Bureau of Aquatic Products, not the Country, will run the hatchery throughout the project period. At the same time, local staff will be trained to enable them to take over the unit when it reverts to Country ownership after the project finishes. During the first year of the project, operating costs will be borne by a 420 000 Yuan low-interest loan from Central Government. Projected 1991 costs are shown in Table 1. In future years, most costs should be covered by income from the sale of fingerlings to the on-growing units. This revenue will be supplemented by funds from the Bureau when necessary. Any profits, however, accrue to the Country. It is unlikely that the latest agreement will hold any better than the first, and further problems can be expected in the future.

The rather one-sided distribution of benefits under this contract has made the NPD reluctant to commit project funds for building the proposed concrete raceway unit for on-growing trout to market size at Nanmenxia as originally planned (see Edwards, Field Document 1, and section 7.2.1).

7.2 ON-GROWING UNITS

7.2.1 Concrete raceways

As described and diagrammed in CTA Edwards' 1989 report (Field Document 1), it was originally planned to construct sufficient concrete raceways at the Nanmenxia Dam site to on-grow approximately 25 t of rainbow trout per year to a market size averaging 250 g. However, as mentioned above, problems between the Bureau of Aquatic Products and the Huzhu Country Government have now made this site less attractive to project management.

On-growing raceways will not be needed before spring 1992, and are scheduled for construction during summer 1991. A firm decision on whether or not to site them at Nanmenxia will be taken early in 1991. This decision will depend on the then-prevailing attitude of the Country Government and on the technical consideration of water availability. As previously discussed, project staff will not be certain about the quantity of water available from the springs downstream of Nanmenxia Dam until the reservoir has filled to its normal operating level over the coming winter.

An alternative site for construction of raceways has been offered by the Longyangxia Hydro-Power Company. The site is adjacent to the Longyangxia Reservoir, where the project's floating cages will be moored (see section 7.2.2 and Field Document 1). Following successful preliminary trials of cage culture of trout (section 7.2.2), which proved the water in Longyangxia Reservoir suitable for culture of this species, the Company is enthusiastic about expanding rainbow trout farming activities not only into more cages, but also into concrete raceways. In addition, the Company wishes to establish its own hatchery to achieve self-sufficiency in fingerling supplies at an early date.

Unfortunately no sites are available close to the reservoir to which water can be diverted by gravity. This is mainly due to the very steep-sided nature of the gorge chosen for daming of the Yellow River. Consequently all water fot the proposed hatchery and raceways would have to be pumped. Further, due to the large fluctuations in water level which result from the normal operation of the dam, the fish farm units would have to be built a considerable height above the lake surface. This would necessitate pumping water at times as high as 50 m.

The substantial amounts of electricity needed for pumping would pose no problem for the Longyangxia Power Company. As part of its agreement with Central Government, the Company has the right to use up to 4% of the total power output from the Longyangxia Dam free of charge. Only about half of this entitlement is currently used. Since the power generating capacity of the dam is 1 280 MW, with an annual output of around 6 000 GWh, the amounts of electricity needed for the proposed fish farms would be, by the Company's standards, small. In addition pumps, pipework, and abundant building materials are freely available on site, following recent completion of the major construction works on the dam.

The CTA, however, would prefer to site the project's concrete raceways at a location where a gravity supply of water is available. This is because the project's farms are partly intended as demonstration units, which will hopefully be copied at other locations in future. Though pumping large quantities of water may be feasible in the special case of the Longyangxia Power Company, it would certainly not be economically viable for most fish farmers, who are not blessed with free electricity. Therefore, if the original site chosen at Nanmenxia proves technically and administratively usable, the CTA prefers to place the project's main raceway unit there. In that case it is still probable that the Longyangxia Power Company will decide to proceed with its own land-based trout farms based on pumped water, and the CTA and other project staff will help as necessary in their design and operation. This possible additional development, with its consequent demand for extra fingerlings not foreseen in the project document, is the major reason why additional trout eggs have been ordered for 1991.

7.2.2 Floating cage unit

The site on Longyangxia Reservoir chosen for the project's cage culture facility was described in Field Document 1.

Physical and chemical monitoring of the lake indicated that water quality was probably adequate for survival of rainbow trout. Oxygen saturation levels were always above 80%, giving dissolved oxygen concentrations at maximum summer temperatures of between 6 and 7 ppm. pH was always between 8.5 and 9.0. Even in summer, no thermocline was found in this deep lake down to a depth of 30 m (the maximum length of cable available on the oxygen/temperature meter). This is probably due to mixing caused by the large amounts of river water flowing through the lake. Maximum summer temperature at the surface was 21°C.

However, useful as such scientific monitoring is, definitive answers about the suitability of any water for trout culture are most simply provided by testing a trial batch of fish at the proposed site. This was done during summer/autumn 1990.

One thousand rainbow troug averaging 100 g in weight were purchased by the Longyangxia Power Company early in July. The fish originated from a commercial trout farm close to Lanzhou in Gansu Province. The trout were stocked into a single 5 × 5 × 4 m deep cage (previously used by the Company in trials on culture of common carp), moored in a small sheltered bay on the north side of the lake, close to the town. The fish were initially fed on a diet composed of a wide variety of ingredients. These included a high proportion of corn, wheat and other uncooked cereals, and even chopped Chinese cabbage. Later the diet was changed to a more conventional composition comprising 70% fishmeal, 5% bloodmeal, 10% vegetable oil, and 15% cereal flours. The feed was produced on site with a cold pelleting machine, and delivered to the fish according to appetite partly by hand and partly by an electric automatic feeder.

Despite some initial mortalities following transport, and the dubious feed quality, surviving fish grew satisfactorily, and by October their average weight had doubled to more than 200 g.

This simple preliminary trial has therefore proved the technical feasibility of cage culture of rainbow trout in Longyangxia Reservoir, and has done much to strengthen the confidence of the Power Company in the project. The Company is now keen to go ahead with larger-scale trials during 1991 and 1992.

According to the schedule set out in the project document, it was not planned to stock trout fingerlings into the cage site until spring 1992. This was partly because it was assumed that all static fresh waters in Qinghai Province would freeze during winter. It now appears that this assumption may be wrong, at least in the case of Longyangxia Reservoir, because this water body remained ice-free throughout the winter of 1989–90. This may again be a result of the large volumes of river water continually flowing through the lake. If the reservoir remains ice-free also during the winter of 1990–91, the project plan can be brought forward, and fingerlings from the Nanmenxia hatchery be transferred to cages at Longyangxia during autumn 1991. This is the earliest that the project's own juvenile fish can be available, since the first intake of eggs to the hatchery will be in March 1991.

However, the management of the Longyangxia Power Company is anxious not to miss the opportunity of running a trial also through the summer of 1991. It is therefore planned to make six of the project's floating cages ready for spring of that year. They will be stocked with bought-in fingerlings, and harvest of market-size fish is planned for autumn.

The cages are currently being built to the specifications of the project's coldwater fish culture specialist. More robust than the Company's carp cages (which are suitable only for very sheltered locations), the new square frames are constructed of timber with corners reinforced by box-shaped structures in galvanized steel. There is a 0.5 m-wide walkway all around each cage, and above this a 1 m high wooden rail serves both for attachment of the top of the net bag and as a handrail. Flotation is provided by expanded polystyrene, sandwiched between the main structural timbers underneath the walkways. Galvanized steel eye-bolts are provided for anchorage at the corners of each cage.

The cage frames are designed to hold nets 5 × 5 × 7 m deep. Net bags and top predator nets have been ordered as part of the UNDP contribution to the project. Twenty-six cages are currently under construction, and it is planned that all will receive fingerlings from the Nanmenxia hatchery as soon as these are available in autumn 1991-spring 1992. As far as the project is concerned, target production from the cage site remains at 20–25 t of market-sized trout in 1992. However, if the 1991 summer trial is successful, the Longyangxia Power Company may decide to increase this target by building more cages and buying-in extra fingerlings at its own expense.

7.2.3 Earth ponds

At the time of writing Field Document 1, the favourite site for development as a demonstration earth pond unit was on the Guide Shijia Stream in Guide Country, about 200 km from Xining City. However, subsequent monitoring of the water source at this location has shown that whilst water quality is always good, flow rates sometimes fall to levels below 0.1 m³/sec. Such small flows (which in some years may become even lower) are insufficient to support a trout farm of a worthwhile size, and consequently plans to use this stream for project purpose were abandoned.

Alternative sites for earth ponds were thus sought elsewhere in Qinghai. Good locations for this type of unit are not common because the soft sedimentary soils covering most of the Province are unstable and do not hold water well. For the same geological reasons, most surface waters become extremely turbid after rain.

A very promising site was eventually found in a distinctly unpromising part of the Province. The second city of Qinghai, Golmud, is located 800 km west of Xining, and is surrounded by desert. Most of Golmud's 120 000 people were re-located in the city as part of a government plan to settle and develop this remote part of China. In addition to the resident population, about 50 000 migrant people (mostly Tibetan and Mongolian) pass through the city each year, and Golmud is also a garrison and supply centre for the People's army in western China. Incomes in this area are said to be above the average for the Province.

Generally, water is scarce in the vicinity of Golmud. Annual rainfall averages only 40.5 mm, and the large river shown on the map close to the city virtually disappears except during the spring thaw of snow in the mountains, when it floods. However, there is water underground, and wherever this comes (or can be brought) to the surface, agriculture thrives.

A number of springs rise close to the city, and the biggest of these, 5 km from Golmud, supplies nine ponds and a lake of total area 4 000 mu before going onto the land for irrigation. The local government uses these water bodies for culture of common and Chinese carps, but temperatures are so low (average air temperature is 4.3°C, and only on about 118 days of the year does the temperature rise above 10°C) that carps require at least four years to reach a market size approaching 1 kg. The spring supplying these ponds is said to deliver between 0.5 and 1.25 m³/sec of cold, clear water throughout the year, and local government wishes to use some of it for production of trout to supply the local market.

Following discussions between the Golmud City Government and the Bureau of Aquatic Products, the following agreement was made (subject to subsequent monitoring of the quality and quantity of spring water proving satisfactory):

• Local government will build a canal to divert water from the spring supply at a site to be chosen upstream of the existing pond system. Five earth ponds totalling 2 mu in surface area and of the average depth 1.25 m will be built to receive this water. An outlet canal will return used water from the ponds to the original stream downstream of the new fish farm. Excavation of the ponds will be done during summer 1991.

• The Bureau of Aquatic Products will give 100 000 Yuan to finance pond and canal construction.

• The Bureau will supply 50 000 rainbow trout fingerlings averaging between 20 and 50 g from the Nanmenxia hatchery during spring 1992, and assist with their transport to the earth ponds at Golmud. The fingerlings will be charged at cost. Extension advice and training in trout culture technology will be provided free of charge by the Bureau via the project.

• Local government will pay all running costs incurred to grow-on the trout throughout summer. Target production by autumn 1992 is 10 t, and it will be necessary to harvest all fish before the onset of the severe winter freeze. It is planned to sell the trout alive in the nearby city. All revenues accrue to the local government.

• Local government undertakes to buy and use the feeds recommended by the project's consultants, and to keep accurate records of fish mortalities, feed usage, and numbers and weights of trout harvested.

8. NAKED CARP HATCHERY PLANNING AND CONSTRUCTION

8.1 DATA COLLECTION AND HATCHERY PLANNING

During spring-summer 1990, preliminary trials were done to provide data on which to base the final planning of the naked carp hatchery and the work programme for artificial reproduction and rearing of fry of this species during 1991 and 1992. Further valuable historical information was obtained from Mr Qu Weiliang, Associate Professor and Director of the Department of Fish Culture, Heilongjian Fisheries Research Institute. Mr Qu has several years experience with the Qinghai Lake naked carp, and was responsible for its introduction to Mongolia. He has also conducted experiments in artificial reproduction of this species.

Previously known information about the reproduction and early life of naked carp was summarized in the 1989 report of aquaculture engineering consultant Varadi (Field Document 3). The additional information gathered from experiments in 1990 and the experience of Mr Qu is as follows:

1. The Qinghai Lake naked carp is a serial spawner, i.e., females release only a proportion of their eggs at any one time. This was confirmed by sampling fish from the Buha River, the main spawning stream for the species, by electric fishing at intervals through the run (late April to early August). At all times, apparently ripe females would give only a few hundred eggs when stripped by hand. Autopsy revealed that most eggs had not been ovulated, and were still firmly attached within the ovaries. However, a high percentage of males gave milt freely.

   At the height of the run, enormous numbers of migrating fish crowd into the Buha River. They are easily caught by electric fishing, and indeed local people catch them easily (and illegally) using a variety of dip, seine and gill nets. It was therefore concluded that there will be little difficulty in obtaining sufficient numbers of broodfish to supply the hatchery. The technical problem to be overcome is the procurement of sufficient ripe eggs from the captured females.

2. Experience in Mongolia has shown that ripe females can be induced to ovulate a greater proportion of their eggs at one time using hypophysation, i.e., injections of pituitary gland extracts of artificial hormones, as is commonly done with other species of carps which are routinely reared artificially in hatcheries. Using this technology, researchers succeeded in obtaining an average yield of about 3 500 eggs per female. Though a great improvement on the few hundreds obtainable without hypophysation, this still represents only a fraction of the total number of eggs produced in the ovaries of each female, which averages 16 000. There is thus scope for further refinement of this technique, i.e., by investigating the optimum type and dosage of hormone, the timing of injections and ovulation, preconditioning and holding environment for brooders, etc. Since this work can be more easily carried out in a large fixed hatchery, it was decided to modify the original plan for artificial breeding of naked carp as outlined in Field Document 1. Instead of building a temporary hatchery at the Buha River for broodstock holding and stripping, it is now proposed that captured broodfish will be returned immediately alive to Xining, where a well-equipped hatchery has been built (see section 8.2). There the broodfish can be held in well-prepared ponds and fed until ready for hypophysation and stripping.

3. Experiments in hatching and early fry rearing of naked carp were conducted during 1990 at the Bureau of Aquatic Product's Extension and Training building and ponds in Xining, where the project offices are located. It was found that satisfactory hatching success could be achieved by holding fertilized eggs in conventional funnel-shaped incubators provided with water gently upwelling from the bottom of the funnel. Small batches could also be hatched in glass jars with aeration but no flow of water. Hatching funnels have therefore been installed at the new naked carp hatchery (see section 8.2). However, since naked carp eggs are large (over 4 mm diameter swollen) and only slightly sticky, it is probable that they can also be incubated in “California” incubators of the type used for hatching salmonid eggs. This running water system would also more closely resemble the natural conditions in the spawning streams. Consequently three such incubators are being installed in the experimental naked carp hatchery in Xining for trials in 1991 and 1992.

4. First feeding of hatched naked carp larvae was done with finely ground boiled egg, as is normal practice for other species of carp in Qinghai Province fish farms. After several days of feeding on this, the juvenile fish were transferred to conventional carp nursing ponds. To ensure that the young fish did not become asphyxiated in the mud of the pond bottom, and to facilitate observations, the fry were held in the ponds inside small floating cages and hapas made of fine netting. Though conditions in the ponds were very different from those in the natural spawning streams, survival of the young fish was good, and growth of surviving individuals to a mean weight approaching 1 g was achieved during summer. It was concluded that conventional earth ponds are suitable for nursing naked carp fry.

8.2 HATCHERY CONSTRUCTION

Detailed plans for the naked carp hatchery were given in the report of Aquaculture Engineering Consultant Muir (Field Document 6), and are only briefly summarized here.

The hatchery is situated at the Bureau of Aquatic Products' carp fingerling production farm about 10 km outside Xining City. Outdoors, one earth pond of about 2 mu surface area has been allocated for storage of naked carp broodstock, and a further 7 mu of ponds is available for fry nursing. Stripping, fertilization, egg incubation and early fry rearing will be done inside a large glass-covered hatchery building. This existing building was previously used only for incubation of common carp eggs. Reproduction of this species finishes at this unit by the end of May, and thereafter the building is free for use in the naked carp work.

The building has been modified and re-equipped for the project's naked carp work programme as follows:

• Naked carp spawn naturally in clean, well-oxygenated water. At the hatchery site this can be provided from a 40 m deep well. Water from the well is routinely pumped to refill and top up the farm's earth ponds when necessary, the existing electric pump delivering around 70 m³/h. The glasshouse/ hatchery is supplied from the pump via a 4 cm steel pipe and valve.

Though the quality of the underground water is good, its temperature is a constant 8°C. Since this is lower than optimal for culture of naked carp, the hatchery supply will be allowed to warm up before use by standing in an existing 170 m³ rectangular concrete tank under glass inside the building. A temperature of 13°–15°C is considered optimal for naked carp reproduction, but in fact anything between 5°C and 18°C is usable (the spawning streams have this range during the breeding season).

• To maintain water temperature close to the optimum, a partial recycling system has been installed. A submersible pump lifts water from the large concrete tank via a 2-inch hose to a 15 m³ steel head tank situated on the roof of the building. From there water falls by gravity into the building to supply 13 incubator funnels, four 2 × 2 × 0.4 m deep rectangular fry tanks, and three California-type incubator troughs. The total maximum water requirement for this equipment is 15 m³/h. The pump will deliver more than this, but for simplicity the system is designed so that the pump operates continuously. Excess water flows through a 4-inch overflow pipe from the head tank back into the main concrete tank inside the building. Top-up and exchange water can be added to the concrete tank as required via the 4 cm pipeline from the well. Overflow from the storage reservoir flows out to waste via a standpipe.

• Water leaving the incubators and fry tanks is filtered by passage upwards through a 7 × 6 × 0.6 m deep filter bed filled with graded rocks, gravel and sand.

• By the end of October 1990, all the equipment was in place inside the hatchery, and pipework to and from the head tanks was installed.

The pump was wired up and operational, and the filter constructed, It remains only to make some final pipework and valve connections to incubators and fry tanks early in 1991. The naked carp hatchery will therefore be ready for use in good time for its planned start-up date of 1 June 1991.

8.3 OPERATIONAL PLAN FOR THE NAKED CARP HATCHERY

The project is fortunate to have secured the services of Mr Qu Weiliang of the Heilongjian Fisheries Research Institute (see Section 8.1) to manage the naked carp hatchery during 1991. Mr Qu is an expert in the field of artificial rearing of carps, and has as much experience as anyone in China on rearing of the Qinghai Lake naked carp. He is now semi-retired from his regular job, and the Bureau of Aquatic Products will employ him on a 3-month contract from mid-May to mid-August 1991.

The arbitrary target set for juvenile naked carp production at the Xining hatchery during 1991 is 100 000 one-summer-old fingerlings. To produce this number of survivors in autumn, it is estimated that some 1 million fertilized eggs will be needed in the spring-early summer spawning season. This will require a minimum of 300 female broodfish and a similar number of males.

Eggs will be obtained by three methods:

1. By stripping ripe fish immediately after capture by electric fishing on the Buha River, afterwards releasing the brooders back into the river.

2. By transporting captured broodfish in tanks by truck to the Xining hatchery. There they will be held in the broodfish pond until required for stripping with the aid of hypophysation.

3. In preparation for the 1992 spawning season, some brooders will be held over the winter of 1991–92 in the hatchery pond, where they will be fed a special diet and conditioned before hypophysation and stripping.

After hatching in the incubators (the relative success of the two types of incubators will be compared), half the larvae will be first-fed in the fry tanks inside the hatchery. Feeding will be on boiled egg and/or hatched Artemia nauplii. After about one week, the fed fry will be transferred to floating cages or hapas in the nursing ponds. For comparison, the other half of the hatched carp larvae will be transferred immediately, without feeding, to the nursing cages, where they must begin to eat natural plankton.

From previous experience it is known that, at a water temperature of 13°–15°C, naked carp eggs require about 6 days to hatch after fertilization. A further 3 days are required before first feeding. The hatchery incubators, fry tanks, etc., will therefore work on a production cycle of about nine days, allowing several batches of eggs and small fish to be produced during the reproductive season.

9. RAINBOW TROUT FEEDS

Fish nutrition consultant Dr A. Tacon carried out a full survey of the availability and price in Qinghai Province of ingredients suitable for inclusion in rainbow trout diets (Field Document 8). From these, good quality feeds were formulated.

Three dry, pelleted diets were redommended: a starter diet for fish from first feeding to 5 g average weight, a fingerling diet for 5–50 g fish, and a production ration for trout from 50 g up to the target market size of 250 g.

The consultant also identified the best local feed mill with the capability of making up the diets. Prices were then negotiated with this company, and finally a contract was signed committing the Bureau of Aquatic Products to buying its 1991 supplies of starter and fingerling feeds from the company. In return, the company guarantees quality and price. (The agreement and specifications for diets are shown in Appendix 3.)

Thus supplies of good quality feed sufficient for running the project's Nanmenxia trout hatchery during 1991 are assured. Project management will also strongly advise the Longyangxia Hydro-Power Company to buy production feed for its floating cage farm from the same supplier.

The Bureau of Aquatic Products is keen to begin manufacturing its own trout feeds after 1991. Though this was not foreseen in the project document, the NPD considers that this is the best way to assure supplies of high quality feeds at the lowest possible prices in the future. Quotations for grinding and pelleting mills are therefore being obtained, and the CTA will recommend the purchase of suitable machinery using money unspent in other areas of the project's equipment budget. Approval for this course of action will be sought from UNDP Beijing and FAO Rome.

On the Chinese side, the NPD has already obtained money from the Qinghai Provincial Government sufficient to finance the building of a small factory to house the new machinery. It is planned to send a member of the project team to Beijing for specialized training in fish feed formulation and manufacturing technology before the new plant becomes operational.

10. EQUIPMENT PROCUREMENT

Generally, the FPO system for ordering goods from overseas up to a value of US $ 20 000 fob worked reasonably well, and most items requisitioned in this way arrived at the project site, ready for their use to begin at the scheduled time.

Exceptions to this timely delivery were usually found to be due to delays in forwarding goods from the port of Xingang or the customs at Beijing airport to Xining. In a number of cases, consignments took well over three months to complete the rail journey from Beijing to Xining. Even more seriously, on arrival in Xining many of the packages despatched by rail were found to have been opened, and sometimes part of their contents stolen. In a few cases, entire consignments have disappeared without trace. A full account of equipment deliveries to the Qinghai Lake investigation component of the project can be found in the two reports of consultant Smith (Field Documents 5 and 9).

The most serious delay in equipment delivery, however, occurred with a purchase above the FPO price limit, which therefore had to be made through FAO Rome. Essential trout hatchery tanks, incubators and other equipment (P.O. 1841386) were first requisitioned, by letter from the CTA to Rome, dated 10 August 1989. The goods are now aboard ship for China, scheduled to dock in Xingang on 6 November, after the CTA's departure from China. Delivery of containers by rail to Xining can be expected to take a further 2–3 months. Thus the late delivery of this equipment has made it impossible to fully prepare the project's trout hatchery during 1990, as required by the project document (see section 7). Though it should still be possible to install this equipment before the arrival of rainbow trout eggs in March 1991, any further delays would threaten to set the entire throut culture programme back a full year.

11. PROGRESS AND PROBLEMS

During 1990, the fish culture work of CPR/88/077 made progress roughly in accordance with the project work plan. Delays, for example in completion of the trout hatchery caused by late delivery of equipment, should not be serious enough to damage delivery of any planned output.

Indeed, in some respects the fish culture work is ahead of schedule. For example, though it was not planned to put rainbow trout into cages in Longyangxia Reservoir until spring 1992, in fact small preliminary trials were carried out in summer-autumn 1990. The succes of this test has strengthened the confidence of the Hydro-Power Company in the project far more than any amount of words or documents. Similarly, small-scale stripping, fertilization, first feeding and fry rearing trials with naked carp were begun in 1990, one year ahead of schedule. These trials provided hard data on this little-known species, which allowed the new naked carp hatchery to be built in a more rational way than hitherto planned.

The Qinghai Lake investigation component of the project appears to have fared less well. The project work plan required full-scale physical, chemical, limnological, and especially fisheries sampling programmes to be carried out from April 1990 onwards. This has not been fully achieved. Instead, during most of the year the project team was still involved in installation and commissioning of equipment, and training in its use. Special emphasis was placed on use of the computer, and a disporportionately large amount of the team's time was spent in the office rather than in the field. Whilst it is true that some aspects of the practical investigaction were held up by the late delivery of equipment, it is equally true that much important fieldwork which could have been accomplished (e.g., intensive sampling and tagging of migration fish during the spawning run, routine measurements of trawl catches) was not approached in a determined or systematic way.

Details of the progress made and problems encountered in this part of the project are described fully in Lake Management Consultant Dunn's report (Field Document 4). The consultant's opinions about the reasons for delays in accomplishment of lake work programmes activities are also clearly stated in this report, and were forcibly presented verbally to the FAO Representative in Beijing. In particular, the consultant was highly critical of the calibre of leadership of the lake study team (though on the Chinese side only).

In response to these opinions, the NPD and CTA have decided to strengthen the team in future by assigning the Deputy NPD, Mr Ying Baicai, to the group as full-time leader. Mr Ying is a Deputy Director of the Bureau of Aquatic Products, and therefore possesses authority second only to the Bureau Director and NPD, Mr Qui Benchen. Mr Ying is also a scientist with 30 years experience in the fishery of Qinghai Lake, and is the author of a recentlywritten book on the subject. Thus his technical expertise in this field is second to none.

It must be understood that fisheries is a relatively unimportant source of food production in Qinghai, where grassland livestock are the major animal products. This was, of course, one of the justifications for giving this project to Qinghai. But for this reason there is no large pool of skilled manpower available in the fisheries of fish culture fields, such as might be found in many other Chinese provinces. Even without any change, the lake investigation team already contained most of the qualified fishery scientists available to the Bureau in Qinghai Province. With the addition of Mr Ying, during 1991 the group will contain almost all of them.

During the remaining years of the project the NPD and CTA will themselves attempt to supervise the activities of the lake study team more closely, as far as their time commitments to other work allow. As a necessary start, in 1991 and 1992 the group will be required to relocate its centre of operations from the Xining office to the fishery base on Qinghai Lake throughout the working season of approximately April-October. If the lake work is directed in a more practical and vigorous way during this period, all the outputs required are still achievable by the end of the project.

Appendix 1

CURRICULUM OF THE FELLOWSHIP IN TROUT CULTURE

hosted by the Institute of Aquaculture, University of Stirling, Scotland, UK, 7 January-7 April 1990

Week 7–12 January 1990

Appendix 2

PROGRAMME OF THE FRESHWATER FISHERY MANAGEMENT COURSE FOR STUDY TOUR CANDIDATES

to be given at Humberside International Fisheries Institute, Hull, U.K., beginning around 11 November 1990

Appendix 3

AGREEMENT BETWEEN THE BUREAU OF AQUATIC PRODUCTS, DEPARTMENT OF AGRICULTURE, QINGHAI PROVINCIAL GOVERNMENT, AND THE QINGHAI PROVINCE MIXED FEEDS PROCESSING FACTORY

1. The Bureau agrees to buy all trout feeds required for fish production as part of UNDP project CPR/88/077 during 1991 from the Company.

2. The Company agrees to make these feeds using the formulations given by the Project's consultants, and in accordance with the technology specified by these consultants (e.g., freshness of ingredients, size of particles to be ground, pellet size). Agreed specifications are attached to this agreement.

3. The price of feeds shall be: Fry diet… Y 3837, Fingerling diet… Y 3445, Production diet… Y 3027. Prices are per tonne ex factory, and will be held throughout 1991.

4. The Bureau expects to need approximately 5 t of fry diet and 25 t of fingerling diet for its hatchery production during 1991. However, should unexpected mortalities occur, or fish growth be poorer than expected, the Bureau reserved the right to buy only as much food as is needed to feed its fish stocks.

5. The Company agrees to allow a member of the Bureau staff to be present as an observer in the factory when trout feeds are being made. The Bureau employee will from time to time take random samples of ingredients and/or feeds for independent analysis. Should these prove substandard, the Bureau reserves the right to terminate this agreement.

Recommended dietary formulations for rainbow trout¹

¹ Recommended dietary formulations for rainbow trout for use within the project's fish farms

² Ingredient name is followed by reported crude protein and lipid content. All ingredients should be first ground so as to pass through a 0.25 mmsieve (for premixes and starter rations) and a 0.354/0.5 mm sieve for fingerling and production diets.

³ Starter: 0–5 g fish; Fingerling 5–50 g fish; Production 50–250 g fish. Starter feeds should be prepared as 0.5 and 1 mm crumbles, fingerling feeds as 1.5–3 mm crumbles and 3 mm pellets, and production feeds as 5–8 mm pellets. Granules or crumbles should be prepared by crushing or crumbling 3 or 4 mm pellets between rollers and then screening out the desired particle size. After sieving, the finished feed should contain not more than 10% oversized or undersized granules. “Fines”, which result from the manufacture of pellets or crumbles of an intended sizeshould be not used as part of starter feed and should be recirculated continuously so as to minimize variations in the formulations intended. In general “fines” content (defined as feed particles below 0.42 mm) should not exceed 2–3% of any finished feed

⁴ Deshydrated brewers grains containing 10% moisture,

⁵ Stabilized with antioxidant

⁶ In the event that the vitamin premix specifications cannot be produced, the trout vitamin premix produced by the Institute of Aquatic Products (Beijing) can be used. The Institute supplier's recommended dosage level for all diets is 200 g premix/200 kg dry diet; vitamin premix cost given as Yuan 20/200 kg diet or Yuan 0.1/kg diet

⁷ In the event that the mineral premix specifications cannot be produced locally, the trout mineral premix produced by the Institute of Aquatic Products (Beijing) can be used. The Institute supplier's recommended dosage level for all diets is 2 000 g premix/200 kg dry diet; mineral premix cost given as Yuan 8.8/200 kg diet or Yuan 0.044/kg diet.

⁸ Calculated composition based on the feed analysis reports of ingredients used by the Qinghai Province Mixed Feed-Processing factory, Xining

⁹ October 1990; Yuan 4.71 = US$ 1

¹⁰ Raw ingredients costs CAF Xining: fishmeal (Peruvian) Y 3.25/kg, meat and bone meal Y 0.99/kg, blood meal Y 2.48/kg, feather meal Y 2.19/kg, soybean meal Y 1.50/kg, rapeseed meal Y 0.58/kg, wheat bran Y 0.56/kg, corn meal Y 0.6/kg, dehydrated brewers grains Y 0.35/kg, fish oil Y 5.0/kg, trout vitamin premix (Beijing) Y 100/kg, trout mineral premix (Beijing) Y 4.4/kg

Recommended dietary nutrient levels for rainbow trout¹

¹ Dietary nutrient levels recommended for clear water aquaculture systems

² Starter: 0–5 g fish, Fingerling: 5–50 g fish, Production: 50–250 g fish

³ Amino acid requirement based on the essential amino acid compositions of fish

⁴ Fish: plant lipid ratio of 5–7:1

⁵ Suggested dietary vitamin levels taking into account processing, storage and leaching losses. Production diets used for broodstock should have their vitamin levels increased by 50% to those employed in the fingerling diet

⁶ Silicon or fat coated

Recommended crumble/pellet sizes for trout of different age groups

¹ When moving to a larger feed size, it is recommended that the two sizes are mixed together for 5–10 days, gradually reducing the amount of the smaller pellet until it is finished