Lake Kariba Fisheries Research Institute

Project Report Number 67

1990 Annual Report

Compiled by C. Machena

,ional Parks and Wildlife Management.

COMMITTEE OF MANAGEMENT

MINISTRY OF ENVIRONMENT AND TOURISM

Mr K.R. Mupfumira - Chief Executive Officer

DEPARTMENT OF NATIONAL PARKS AND WILDLIFE MANAGEMENT

Dr. W.K. Nduku

Mr. G. Panqeti

Mr. R. B. Martin

Mr. Nyamayaro

LAKE KARIBA FISHERIES RESEARCH INSTITUTE

Dr. C. Machena - Officer - in - Charge

Mr. N. Mukome - Executive Officer and Secretary ofCommittee of Management.

- Director and Chairman of Committee ofManagement

- Deputy Director

- Assistant Director (Research)

- Assistant Director (Administration)

TABLE OF CONTENTS PAGE

Officer-in-Charge's Reoort 4

The Zambia/Zimbabwe SADCC Lake Kariba FisheriesResearch And DeveloQment Pro:iect 12

Comparative Study Of Growth Of Limnothrissanijodon(Boulenger) in Lake Kariba 18

An Analysis Of The Effects Of Fishing Location AndGear On Kapenta Catches On Lake Kariba 19

Hydro-acoustic Surveys In Lake Kariba 20

The Pre-recruitment Ecology of The FreshwaterSardine Limnothrissamíodon (Boulenger) In LakeKariba 22

Report On Short Course In Zooplankton QuantitativeSampling Methods held at the Freshwater BiologyLaboratory Windermere From 19 to 30 November 1990. .. 25

Report On Training: Post-Graduate Training -Humberside Polytechnic, U.K 32

Postharvest Fish Technology In Lake Kariba.Zimbabwe 33

Assessment Of The Abundance Of Inshore Fish StocksAnd Evaluating The Effects Of Fishing Pressure OnThe Biology Of Commercially Important Species AndEcological Studies On Synodontis zambezensís.......35

11. Publications 41

3

1. OFFICER - IN - CHARGES'S REPORT

Institute Finances

Like in the last financial year, (1989/90). the financialposition of the Institute is such that several activitiescould be accomplished without much constraints. The 1989/90fund was $768 000.00. Within this financial year the fundstands at $795 000.00

Staffing and staff training.

Junior staff

A number of vacant posts were filled in June 1990 asbellow:-

4 for general hands and2 for office orderlies

The typist/stenographer D. Mwaita swopped positions with F.Gapara. F. Gapara was based at the Mushandike NaturalResources Col lege.

The position of the two vacant posts for Scouts (1/II)remain unresolved. The Personnel section in National Parkscontends that Gahamadze and Chinoda transferred with thoseposts to Lake Kyle Research Station and the Bulawayo officerespectively. This Cannot be the case and the ChiefEcologist (Aquatic) has been asked to verify this.

Mr Farai Gapara attended a computer training course in DOS,and Word processing. Mr. Nyaude completed a correspondencecourse in book-keeping.

Senior staff

The post of the technician Mr. Nyamhanza has not beenresolved. As the Institute's fleet of vehicles as well asthe number of vessels will expand under the Zambia/ZimbabweSADCC Fisheries Project it is imperative that a competentjourneyman be appointed. Problems have been encountered intaking vehicles to local garages. The service is not alwaysgiven when required and furthermore as there are only 2garages in town these tend to be monopolistic and chargeexorbitantly. It would be appreciated if the Committee ofManagement took steps to resolve the problem.

4

Mr Power Chitaunhike the E.O. left the Institute for theMinistry of Environment and Tourism. His place was takenby Mr. Naboth Mukome who has also since been able to attenda computer course in DOS and Spread sheets.

One post for a technician is vacant. Effort will be madeto fill in this post within this financial year. It shouldbe filled in by a laboratory technician.

Mr. M.Z. Mtsambiwa has already started data collection forhis project on the pre-recruitment stages of kapenta.

Mr. H. Nyaruwa has completed a Master of Science programmein food science and technology at Humberside (UK).

I am amazed however that Mr. Nyaruwa worked on beef andtropical fruits for his Diploma and MSc dissertationsrespectively. My understanding was that he should havespecialised on aspects of fish for his research projects.This is a high degree of irresponsibility, on his part.

Miss R.A. Sanyanga has completed a Master of Scienceprogramme in Applied Hydrobiology and Fisheries atUniversity of Wales, Institute of Science and Technology.

Miss P. Chifamba is on a Master of Science trainingprogramme in Fisheries Biology and Management at theUniversity of Wales in Bangor.

Mr Newman Songore is on a 9 months Diploma in FisheriesManagement course at Humberside (U.K). It is unlikely thatthis 9 months diploma programme will be recognised by thePublic Service Commission in terms of increasing Newman'sremuneration. He was reluctant to go on a 2 year programmewhich would have been equivalent to 'A' Levels. PSCrecognises a minimum of 2 years in diploma training.

The training programmes are soon coming to an end, and theirusefulness is fairly apparent. Those who have alreadycompleted training show much competence and interest intheir programmes.

There is a strong need at the Institute or at Head Officefor a fisheries economist. It is unlikely of course thatanother post will be given to the Institute. Hence theInstitute will continue relying on the Centre for AppliedSocial Sciences particularly in socio-economic programmeswith the inshore fishing communities.

(c) Staff accommodation

The house that was being constructed by NheweyembwaInvestments is not yet completed. NORAD has expressed

5

interest to provide funds for the completion of the houseand Ministry of Construction approval has now been grantedfor completion of the house, though this took a long time.Chances are that the biologist coming under theZambia/Zimbabwe SADCC Fisheries Project will occupy theh o u se

The house for the co-ordinator has finally been paid for.

There is need for a guest house that will accommodateproject staff from outside the Institute. For exampleZambian personnel have been staying in hotels and this willprove to be expensive in the long run.

Institute expansion

A site for a new Institute has been located. This site isabove CMED workshops in Kariba. Physical Planning(Mashonaland West) have already surveyed the site. Ministryof Construction was working on topographic surveys whenprogress was halted The outcome of the money promised byMinistry of Finance under PSIP is yet unknowns Hence thisproject is likely to be delayed.

Meanwhile the expansion of the Institute at the presentsite is under-way. A pre-fabricated laboratory and a pre-fabricated block of office will be constructed hopefullysoon. Tender arrangements for construction are beingorganised. Funds for thìs are provided within theNORAD/DANIDA funded programme.

In addition to this there is need to rehabilitate the mainInstitute office block. Because this block is not dumpproofed, ceilings and prefabricated divisions betweenoffices are almost collapsing. Ministry of PublicConstruction and National Housing have estimated the job at$44 846.50. This will be on a cost recoverable basis.Funds are available within thisfinancial year. Approvalby Committee of Management is awaited. As tendering is notnecessary in this case, the job could be commenced quitesoon.

Funds have also been budgeted for a sub station at Binga.This will be in the form of a guest house for officers andone for junior staff including an office, a laboratory anda store. This will facilitate working on the Binga side ofthe lake.

Vehicles and Vessels

The landrover 110 has been out of uso for almost 6 monthsnow. The problem started when the timing chain snapped and

6

the vehicle was taken to a local garage for the service.After that the vehicle lost compression and was losing oilin a bad way. The engine was removed and taken to Leylandfor repairs. It took Leyland almost four months to do thejob.

This created transport problems for the Institute which wasnow running on 2 vehicle - the Daihatsu pick up and thePeugeot 504 station wagon. It became necessary to hire avehicle from Hertz from time to time. CMED in Kariba orChinhoyi could not help.

The Peugeot 504 has run down tremendously for a vehiclehardly 2 years old. The shortage of vehicles at theInstitute has created a situation which necessitated theuse of the 504 for jobs heavier vehicles should have done.Also, there is the problem of too many drivers some who arenot necessarily careful. Arrangements are being made torestrict usage of the 504 by only a few drivers. This willof course await the arrival of the project vehicles.

Three vehicles a Toyota 5 tonne tuck, a Toyota land cruiserand a Hi-Lux twin cab are expected to arrjve any time now.These have been bought under the NORAD/DANIDA sponsoredprogramme. The problem of vehicle shorted will bealleviated.

The Pelican research vessel is running well. Mr. R. Brasted,the designer and consultant on the new research vessel hasbeen asked to do a survey on the Pelican. This is called forsince the Pelican has been running for over 14 years now andis heavily relied upon for most Institute researchactivities.

Construction of the new research vessel finally took offthis year and considerable progress has already been made.Construction is being carried out by Morrison Brothers inKwekwe. Mr Brasted has been on site on 2 occasions tosupervise the laying of the frame and the plating of sheetsteel. There was some delay in the acquisition of sheetsteel. Foreign currency was eventually provided by NORADfor this purpose. The engine, echo-sounder,radar etc. havebeen ordered and are expected to be fitted in January 1991.Because of the delay in acquiring sheet steel, the vesselis expected to be completed by June 1991 instead of April1991.

Because of increased research activities under theNORAD/DANIDA funded programme, two 6m run-about boats havebeen planned for. These are also being designed by MrBrasted. They will be constructed o-F aluminium which ischeaper and more durable than fibre glass. A number ofinflatables and outboard engines will also be bought.

(g) Fisheries manaqement

7

Fish poaching is still a very serious problem and is likelyto be a problem for a long time. It is necessary to take anumber of measures without delay. It is a fairly difficultproblem to handle because it is very much related to thedifficult economy in the country. Fishing and in our case,gill-netting is regarded as an occupation of last resort.That is when every thing else has failed, people will looktowards fishing for economic relief. There are low capitalinputs required and people will always look at the fisheryresourcein water bodies as inexhaustible.

The price differentials between the fishing village on LakeKariba and markets in Harare also create a problem. Fishsells for Z$ 2O0 /kg at the fishing camps and at $3.50 inKariba and weil over $6.00 in Harare.

It has been profitable for a lot of people in Kariba toorganise buying from fishing camps and selling at a numberof outlets in Kariba. These outlets include Willards, BlueWater Charters. Irvin and Johnson etc. I have come acrosssome women and men who have left their jobs to engage inthis trade on a full time basis. All they need is a hawker'slicence from Nyami Nyami or Binga Councils.

This trade between the fishing camps and outlets in Karibastarted when Willards set up their factory in Kariba andwere allowed to buy fish at the factory instead of havingtheir own boats buying from the fishing camps like Irvinand Johnson did. It also became profitable for unlicencedpeople to acquire gill-nets, use these and sell their fishto Willards as well. The police and the army have beeninvolved in this as well. Kapenta fishermem found itprofitable to fish in the prohibited shallow waters for thepurpose of catching tiger fish to sell to Willards. Oneneeds 10 kg of tiger fish per day to fetch about $1 000.00a month when selling at $3.50/kg This is a lot more moneythan these fishermen get on their kapenta returns evenwhen p.aid on commission.

It is also common practice for people to come to Karibafrom Harare on weekends, hire a boat. buy fish from thefishing camps or set gill-nets and sell the fish either inKariba or take it to Harare. Gill-net manufacturingcompanies were previously requested to sell gill-nets onlyto licenced fishermen. But this doesn't solve the problemanymore as many gill-nets are available from traders buyingthese from Mozambique.

Attempts to ask Willards to buy from only licenced traderswere fruitless as these would sell on behalf of unlicencedtraders. The Institute also requested the assistance andunderstanding of the Nyami Nyami Council particularly inallowing the Institute to veto the applications for hawker'slicences. Nyami Nyami Council failed to appreciate the needfor this.

Fish trading on hawker's licences was instituted when thefishery began, because it was anticipated it would be on asmall scale and was viewed as opening the areas close tolake shore to a cash economy. At the time it was only Irvinand Johnson who was the only and large commercial trader onfish. But today there are fairly large companies operatingon hawker's licences from Nyami Nyami.

In an attempt to get over these problems, companies havebeen stopped from buying fish on their premises. But theycan buy directly from fishinc camos.

fish on Kariba be instituted. This will be issued to a fewindividuals or companies and will have a number ofconditions attached to it, for example that all licenceholders should buy fish only after it has been landed andrecorded at the fishing camps. At the moment most buyersare intercepting fishermen before the fish is landed andrecorded. Our enumeration system will go to waste if thisis not rectified. Licencing a few individuals will make itfairly easy to monitor their operations.

As the economy will get worse and hit the poor people reallyhard, severe poaching problems need to be anticipated. Longterm measures that National Parks anticipate to take pertainto the establishment of fish wholesalers. Fishermen willthen be asked to sell their produce only to these and thenthese will then market the fish. Co-operatives orindividuals could establish these.

Mean while National Parks will make it a requirement forall fish buyers even though they may have the appropriatehawker's licence to obtain a fish selling permit from theDepartment.

Also in connection with these is the need to gazette anumber of fishing regulations. These were sent to HeadOffice almost 2 years ago. We put so much effort inapprehending and taking fish poachers to courts. We havelost these cases because we do not have these regulations.In the interest of our anti-poaching effort will this belooked into seriously. The Chief Ecologist (Aquatic)promised to handle this.

h) Cage culture pro.iect

Last year's report on this project was much more technicalhaving been based on growth rates and stocking density data.When Miss R.A. Sanyanga who was monitoring the cages leftfor training in September 1989, there was no one availablefrom the Institute to keep monitoring. Willards Foods (Pvt)Ltd who have the cage set up were not in a position tocollect biological data.

9

Willards have come forward with a request to add 8 morecages, over a period of eight months. They have 8 at themoment. There was some initial reservation based on thefact that any expansion should be based on a fairly soundscientific basis.

There is demand from many quarter-s to go cage culture.Location of cages demands areas that meet a number ofrequirements. these being sufficiént potection from strongwinds, good currents and areas sufficiently safe fromcrocodiles, amongst a number of other requirements.Expansion of cages will in no doubt create conflict withother activities particularly power boating and kapentafishing. Kapenta fishing is conducted at night. Powerboating is intense during public holidays.

A study proposal has therefore been prepared. It is beingpresented to SAREC for funding. The objectives are (i) tomake a feasi.bi 1 ity study of the potential of aquaculture onthe Lake and provide the technological as wel1as ecologic-economic background information necessary for thedevelopment and management of the cage culture system (ii)assess potential conflict with other lake users and minimisethis (iii) assess potential environmental impacts and (iv)study a number of biological parameters eg growth, feedconversion ratios and production potential.

It had been hoped that an ecologist within National Parkscould take up the project but this may not be feasible asthe fisheries Branch is short staffed. Besides allInstitute staff have projects with the NORAD fundedprogramme. It may therefore be necessary to obtain somebodyfrom outside National Parks at the level of either aresearch fellow or somebody reading up for a Ph.D.

In the light of this development, and also because Willardshave already made an investnent, they could be allowed toexpand, whilst the study is taking place. This will infact facilitate assessment of the impact of the cages onthe environment. At the same time because Willards havenot been paying a licence fee they should be requested toset aside 3 or 4 cages for experimental work, within theproposed study.

i) The 1990 Kariba International Tiger Fish Tournament

An economic evaluation of the 1989 Kariba InternationalTiger Fish Tournament (K.I.T.F.T) based on expenses relatingto travel costs, lodging fees, food and booze costs, costsof bait and boat hire, entrance or competition fees as wellas sponsorship showed that over the 3 day competitionmembers belonging to 309 teams spent an average of $288,00to catch 1 tiger fish or $100.00 to catch 1 kg of tigerfish.

10

Another evaluation using a similar approach was carried outin 1990 and showed astounding results. In 1990, over the3 day period, it cost $211.00 to catch 1kg of tiger fish or$491.00 to catch 1 tiger fish.

Such a vast difference in cost over a period of one yearreflects in a way the inflationary tendency of the dollarbut more on the extent to which people are prepared to spendfor recreation. In a way this also reflects the value ofthe tiger fish as a sport fish and it needs not beemphasized that angling pressure on this fish will increaseover time particularly as the South African politicalsituation improves and as more South Africans will visitthis country. There is no doubt also that the movementpattern of the fish needs to be Investigated as much as thefish also needs protection from heavy fishing.

j) Re-organisation of some inshore fishing camos

Following the expiring of the Irvin & Johnson (Pvt) Ltdlease on some inshore fishing concessions and the problemscreated by fishermen operating from Parks land particularlyin assisting rhino poachers (whether at gun point orwillingly) and the intense recreational usage of theMatusadona shoreline, there is high pressure from Tashingathat King's Camp fishermen be re-located This issue couldawait and be discussed in the context of the Lake shoreplanning studies towards the beginning of the second quarterof 1991.

11

ZAMBIA/ZIMBABWE SADCC FISHERIES PROJECT

by

Dr D.S.C. Lewis

GENERAL

Although the agreement between Norway, Denmark, Zimbabwe andZambia was not signed until 14 December 1990, the release ofadditional funds by the donors for the pre-phase enabled theProject to make consìderable progress in 1990. All seniorstaff at LKFRI underwent overseas training and in-countrytraining for both senior and junior staff took place duringthe year. A number of activities which required theappointment of consultant specialists took place in 1990.These included the convening of two planning workshops, alake-shore planning study, the establishment of a fisheriesdata-base and the conducting of a frame survey, thesupervision of the construc1ion of the new research vesseland the design of work boats and an evaluation of the socio-economic component of the Project. Agreement was reached inOctober on the appointment of Mr Helge Paulsen from theDanish Institute for Fisheries and Marine Research asfisheries biologist for the Project. Mr Paulsen and hisfamily are due to arrive in January 1991. During the year acontract was drawn up with the procurement agency, Techprofor the sourcing and procurement of Project equipment and anumber of consignments of essential equipment were received.

2 TRAINING

The Project provided funding for a number of trainingcourses during 1991. These are detailed in the Head ofStation's report.

3. WORKSHOPS AND MEETINGS

A planning workshop for the artisanal component of theProject, combined with a training course on the applicationof objective orientated project planning, was held in March1990. Dr Hasan Moinuddin was contracted as facilitator forthe workshop. The workshop succeeded in clarifying theobjectives of the Artisanal Sub-project and a comprehensivereport was produced Which listed objectives, indicators,external factors, outputs and activity plans for theartisanal component.

In view of the success of the March workshop, a secondworkshop, also overseen by Dr Moinuddin, was organised forSeptember 1990. This concentrated on the commercial fisherycomponent of the Project and also attempted to clarify

12

objectives and identify indicators, external factors etc.In addition the workshop served as a forum for debatebetween the fisheries managers (LKFRI) and the operators ofthe fishery and the perspectives of both parties wereexpressed and discussed. A report on the workshop whichcovers all of the issues discussed and recommends follow-upactions was prepared.

Meetings of the Project Steering Committee were held in May,August and October of 1990. Many issues were discussed anddecided upon at the meetings and all of these are fullydocumented in the minutes.

LAKE SHORE PLANNING STUDY

In April 1990 J.M. Hutton (Pvt) Ltd were contracted toundertake the lake-shore planning study for the Zimbabweanside of the Lake. The initial contract was for a year butit became apparent during the study that the many complicat-ed issues involved required both more time and some inputsnot anticipated when the initial contract was drawn up. TheSteering Committee thus agreed to an extension of thecontract period by four months and to the inclusion of twoadditional activities namely, an investigation of somespecific intrinsic socio-economic issues considered ofrelevance to the preparation of the lake-shore plan and theprovision of assistance with the detailed planning of theKariba Town Council area.

Good progress has been made and it is anticipated that adraft plan and report will be available by mid 1991.

FISHERIES DATA-BASE

Mr Villi Thorsteinsson from the Marine Research Institute inReykjavik, Iceland was contracted by the Project throughICEIDA to set up a fisheries data-base for Lake Kariba, totrain staff in data-base management and to supervise theimplementation of a frame survey. Mr Thorsteinsson spentthree months in Kariba (August to September) during whichtime he trained staff in the use of the relational data baseOpen Access III, set up the data base on the LKFRI comput-ers, prepared data-forms (in conjunction with staff fromCASS) and conducted a frame survey, on the Zimbabwe side ofthe Lake. After his departure, the frame survey wascontinued on the Zambian shore. Miss Rudo Sanyanga wasselected to manage the data-base on the Zimbabwe side andboth she and her Zambian counterpart are due to make a visitto the Marine Research Institute in Iceland for discussionson the frame survey results with Mr Thorsteinsson andfurther training in February 1991.

13

CONSTRUCTION OF NEW RESEARCH VESSEL

In June 1990, Marine Architect, Mr Robert Brasted wascontracted by the Project to supervise the construction ofthe now 11m research vessel by the Morrison Bros boat yardin Kwekwe. Mr Brasted made two visits to Zimbabwe in 1990(June/July; October/November) to supervise the laying of thekeel and the plating of the hull and full reports wereprepared at the end of 'each of his visits. The main enginefor the vessel has now arrived and is being fitted byMorrisons. All the required fittings have been selected byMr Brasted and orders have been placed with the suppliers.A consolidated consignment is expected to arrive in January1991 and Mr Brasted is due to make a further two visits tosupervise the final construction work and fitting out and tolake test the vessel. Mr Brasted has also been contractedto prepare detailed construction plans for aluminium workboats for the Project. Tenders have already been receivedfrom three companies for their construction and a selectionwill be made during Mr Brasted's next visit.

SOCIO-ECONOMIC COMPONENT

During 1990 negotiations took piace with both CASS and theInstìtute of African Studies at the Universìty of ZambiaBecause of the complexity of the proposed socio-economiccomponent of the Project and the high proposed costs of someof the activities, NORAD appointed an external consultant(Mr Doif Noppen of the Nordïc Consulting Group) to appraisethe various proposais put forward by the Project, CASS andlAS and to provide recommendations for the Socio-economicResearch and Monitoring Programme (SERMP). A report wassubmitted in December 1990 and meetings are scheduled forJanuary 1991 to discuss the recommendations.

Because of the delay in the implementation of the SERMP, anadditional element was added to the lake-shore planningstudy (see section 4), and Mr Jeremy Jackson was appointedto undertake a field study and prepare a report for theLake-shore planning consultant Mr Jackson also spent sometime with the data-base team and held discussions on thecoordination of field data collection exercises

PROCUREMENT AGENT

At the beginning of 1990, the procurement agency Techpro (adivision of ZAL Holdings) were selected from a shortlist ofthree by the Steering Committee to act as a sourcing andprocurement agent for Project equipment. A contract hasbeen drawn up and is due to be signed by both parties inJanuary 1991. Techpro have already placed a number oforders and arranged for shipment and have, to date proved tobe very efficient.

14

FISHERIES BIOLOGIST

At a meeting held between NORAD and senior members of DNPWLMand DoF in October 1990, the matter of the appointment of afisheries biologist for the Project was discussed. Thecandidate originally selected was no longer available totake up the post and it was therefore necessary to decidewhether to readvertise the post or to re-examine the shortlist in order to determine whether there was anothercandidate suitable for the post. The latter option waschosen and Mr Heige Paulsen from the Danish Institute forMarine and Fisheries Research was selected for the post. MrPaulsen was contacted and agreed to accept the offer. He isdue to take up his post at LKFRI in January 1991

EQUIPMENT

Despite the delay in the signing of the Project agreement,funds have been made available by the donors to enable theProject to purchase essential equipment. Equipment acquiredduring 1990 include workshop materials (flip charts andeasels, pin boards etc), sampling nets, a Wild stereomicroscope and assorted fittings, an engine for the researchvessel and an additional computer. Orders have been placedfor boat equipment and fittings, portable computers,accessories and consumables and vehicles. These items areexpected to arrive in early 1991.

STAFF ACTIVITIES

There has been extensive communication and cooperationbetween the staff of LKFRI and D0F during the year. BothProject Co-managers have made numerous visits to theneighbouring country for meetings to discuss management ofthe Project, budgets and work programmes. Staff from Zambiaspent considerable time at LKFRI during the setting up ofthe fisheries data-base and a joint team of LKFRI and D0Fstaff undertook a comprehensive frame survey of theartisanal fisheries along both shores of the Lake.

Although an integrated Project research programme has yet tobe finalised, individual staff have conducted their ownprogrammes and have outlined the activities they arescheduled to undertake in 1991 and the equipment that theywill require.

PROPOSED ACTIVITIES FOR 1991

With the Project agreement now signed, full scale Projectactivities can commence at the beginning of 1991.

15

On 4th and 5th March 1991, the annual meeting betweenDNPWLM, D0F and NORAD will be held. At this meeting thebudget for the period Jan 1991 - June 1992 will be present-ed, progress of the Project to date will be evaluated, theactivity schedule for the following 18 months will bediscussed and issues relating to the management of theProject will be tabled.

From 20 January to 16 February, an FAO - IFIP stockassessment training course will take place in Kariba.Although this is not being organised by the Project, itscontent is highly relevant and all available research staffare expected to attend. It is proposed that lengthfrequency data on kapenta from Lake Kariba will be used fordemonstrating the methodologies covered during the coursethus the exercises could provide valuable information on theKariba stocks.

During his visit to the Renewable Resources Assessment Groupat Imperial College, London at the end of 1990, Mr MorrisMtsambiwa, in conjunction with Dr Tony Pitcher, prepared acomprehensive work plan for the study on pre-recruitmentstages of Kapenta. Dr Pitcher is scheduled to pay a visitto Kariba in mid 1991 in order to finalise the details ofthe proposed sampling programme.

Having completed the ist frame survey of the artisanalfisheries of Lake Kariba at the end of 1990, Ms RudoSanyanga and her Zambian counterpart Mr Justin Lupikishawill spend two weeks with the data-base consultant, MrThorsteìnsson, to decide upon the sampling frame to be usedduring 1991. Regular sampling will then commence on theirreturn in March 1991. In view of the rapid changes whichappear to be taking place within the lake-shore fishingcommunities, it is proposed that the frame survey berepeated in 1991 when M Thorsteinsson returns for a furtherthree month visit to Zimbabwe and Zambia. During his visithe will also provide further training on the operation ofthe data-base and assist Project staff overcome any problemsthat have arisen since his last visit.

An economic evaluation of the kapenta fishery of Lake Karibais scheduled for the first half of 1991. This evaluationmay be jointly funded by the Project and FAO/IFIP who haveagreed to arrange for a workshop on fisheries bio-economicsto be held during the year. The terms of reference for theeconomic survey are being finalised and will be submitted toNORAD for approval at the forthcoming annual meeting.

A visit by hydro-acoustics expert, Torfin Lindern is plannedfor the latter half of 1991. It is anticipated that regularhydrographic surveys to assess fish distribution andseasonal abundance will be an important component of theProject and Mr Lindern will be required to assist in theestablishment of a suitable sampling and analysis programme.

16

A project monitoring workshop is planned for September 1991Dr Hasan Moinuddin, who served as facilitator at bothplanning workshops, has been invited to run the workshop,which will assess the progress of the Project and provideinstruction to Project staff in project evaluation andappraisal

Discussion is currently taking place as to whether or not atechnical workshop is required for the purpose of establish-ing a coordinated research programme for the project and torelate research activities to management requirements. Thedebate centres on whether it is possible to prepare such aprogramme with the input of Project staff, the newlyappointed fisheries biologist and various visiting experor whether it would be preferable to hold a workshop towhich various experts in the relevant fields could beinvited. The matter will be discussed further in 1991 and aproposal put to NORAD at the annual meeting.

In preparation for the commencement of the socio-economicresearch and monitoring programme, advertisements are beingprepared for recruitment of the research fllows who willundertake the in-depth S-E studies. These will be discussedwith CASS and it is hoped that appointments can be made intime for the in-depth studies to start in mid 1991.

17

P.C. CHIFAMBA

TITLE: Comparative study of growth of Limnothrissa rnjodon(Boulenger) in Lake Kariba.

OBJECTIVES

The objectives of the project were

to determine the extent of variation of growth rate ofkapenta throughout the lake.

to compare the two methods of age determination; lengthbased and counting of daily rings on otolith.

WORK ACCOMPLISHED

Preliminary work on age determination of kapenta using dailyrings was carried out. The growth parameters Loo, K and t1obtained differed from those obtained by other workers who usedlength based methods for age determination. A higher asymptoticlength and lower K value was found in this study. There are twopossible reasons for the differences. One is that theseparameters are very variable in time or as strongly suspected.that the difference is inherent in the methods used. Thisindicated a need to compare the methods of age determination.

A monthly sampling programme was initiated in May, 1989. Samplesof kapenta were collected at night using a small diameter net(3m) fitted on the Pelican. There are 14 sampling stationsapproximatély equidistant on the long axis of the lake. Fishcollected from each station were divided into two lots. Age inone lot is to be determined using length frequency analysis anddaily ring counts in the other.

For the length frequency analysis length and weight of individualfish were measured and the frequency histogram drawn. These willthen be used to determine growth parameters. Samples for dailyring counts are being kept frozen for later analysis, when thetechnique of embedding and grinding otolith have been mastered.

Ari attempt was made to validate the periodicity of the ringswhich so for are presumed to be daily rings. Six trips havebeen made in an attempt to capture live kapenta. It seemspossible to keep the fish alive as one kapenta was successfullycaptured and survived in the aquarium for a month.

18

FUTURE WORK

Samples being kept frozen will be processed by counting dailyrings and growth parameters calculated for the all stations.Growth rates between the areas will be compared. The growthparameters from the length frequency histograms will becalculated using various programmes. These growth parameter-swill be compared with these obtained from counting daily rings.This might explain differences between results in the preliminarywork and, values obtained by other workers.

Further attempts at capturing live kapenta for validating growthrings will be made. Initial rings will be marked using radio-active tetracyclines. Another short project follows.

TITLE: An analysis of the effects of fishing location andgear on kapenta catches on Lake Kariba.

OBJECTIVES

The objective of the analysis were:

to investigate the type of equipment on fishing vesselson Lake Karibaassess the extent to which different equipment on boatsaffected catch.

RES J L T S

Out of the 196 questionnaires sent to companies 138 werereturned. The forms showed that 30% of the boats were non-mobileand that most boats used nets of 7 and 8m diameter. The smallestnet size was 5m. Almost half the vessels had echo-sounders andthat these were fitted only on mobile boats.

Sengwa is by far, at present, the best fishing ground and hasthe lowest fishing pressure. -The amount of fish captured by thevessels is differert (5% significance level, ANOVA). Mobility,echo sounders and size of net all acted in an additive manner toincrease fish catches. Larger nets tended to catch more fishthan smaller nets although an anomaly was observed at 7mdiameter.

CONCLUSION

The implication of the results is that fishing effort isregulated throughout by the number of fishing vessels. All boatsare assumed to have the same amount of fishing power when fishingeffort is calculated. However, in future the fishing efficiencyof vessels is likely to rise and perhaps counteract the effectsof keeping the number of vessels constant in order to maintainconstant fishing effort. This has to be borne in whendetermining the number of vessels that can be allowed on thelake.

19

Machena

TITLE HYDRO-ACOUSTIC SURVEYS IN LAKE KARIBA.

Research Proposal

OBJECTIVES

To establish a programme for a regular monitor-ing of thestocks of kapenta as well as their size distribution.

To monitor fish migrations on the lake.

To attempt to establish environmental factors that triggerthe vertical migrations of kapenta.

JUSTIFICATION

The project is part o-F a larger project see.king to establishjoint research programmes with Zambia to obtain data on stocksand on the population dynamics of kapenta. It is hoped that theinformation generated will enable the formulation of jointmanagement strategies. The pelagic fishery is shared betweenZambia and Zimbabwe and has been managed separately.

The kapenta stocks of Kariba represent the highest socio-economically yielding fishery in Zimbabwe and allocatingresources for research in this fishery is much justified.

METHODS

a) Biomass and lenqth frequency analysis

The hydroacoustic technique as used in the pre-project phase byLindern (1989) will be used. This entails working along a numberof transects randomly selected to cover the lake. Along eachtransect the density of fish will be assessed using a portableecho-sounder Simrad EY-M. It operates at 70kHz with an outputof loo Watt. The echo signals are recorded on tape and lateranalysed by a computer. The hydroacoustic data acquisitionsystems HADAS will be used to digitize and analyze the recordedecho--signals. Through the integration of both single fish echoesand school echoes, the technique will give the total fish densityin an area (Lindern, 1989).

The data obtained will be used to estimate population size aswell as size distribution of the fish.

20

Fish miqrations

The same technique as described above will be used to get anindication of the extent of horizontal migrations, of thekapenta.

The technique will also be used to detect and monitor events inspawning areas. For this part another echo-sounder with a higherfrequency e.g. 200 kHz will be used.

Events triqqerinq daily vertical miqrations of the kaDenta

What triggers kapenta to carry out diurnal vertical migrations?Theories proposing zooplankton migrations and changing lightquality have been proposed.

For this part kapenta depth distribution will be correlated withspectral light quality as well as dissolved oxygen levels.Correlations with zooplankton biomass and species diversity isdemanding and assistance could be sought for this from UniversityLake Kariba Research Station.

REFERENCE

Lindern, T. 1989. Results from the acoustic survey of LakeKariba. A report prepared for the Zambìa/ZimbabweSADCC Fisheries Pro.iect. University of Oslo.

21

22

M.Z. MTSAMBIWA

TITLE: The pr-e-recruitment ecology of the freshwater sardineLimnothrissa miodon (BO u 1 e n g e r) i n Lake Ka r i ba.

OBJECTIVES:

To determine the stage in the development of larvaewhen ring deposition in kapenta otolith commence.

To validate periodicity of the increment deposition

To identify spawning grounds and time for kapenta

To evaluate the life history parameters ( e.g. growthand mortality rates, age at recruitment) for larvaland juvenile kapenta.

To undergo training towards a Ph. D. degree during thecourse of this study.

WORK ACCOMPLISHED

Due to late arrival of plankton nets and otherequipment none of the above objectives was accomplishedin detail. However two trial sampling on the R/VPelican was carried out in the Sibilobilo Lagoon duringAugust and September 1990. The August trials yieldeda lot of different types of zooplankton from bothsurface and 2 metre deep tows. Neither eggs norkapenta eggs were captured during tour. The Septembertrials yielded some larval kapenta (the smallestmeasured was 10 mm). It is possible that the Augusttrials yielded no larvae or juveniles because they werenot present in the system. Could it be possible thatits a hatching time for the larvae that appeared inSeptembe r?

Another possibility is that the August trialsexperienced a lot of errors from the research personnelas all were doing it for the first time. Escapementand avoidance of gear by the larvae and the juvenilescould result from the tow speed being too slow and fromdisturbance by the moving vessel. Tow distance andtime are also crucial as it was observed that daytimesampling yielded a very small variety if any ofzooplankton from the depths sampled i.e. surface and2 m deep. This could imply vertical migrations linkedto reduction of predation during the day, a pattern orstrategy observed for several planktonic organismswhich come up to the surface to feed at night.

This project was registered in October, Ph.D degree

with the Renewable Resources Assessment Group underImperial College of the University of London. Thethree years split PhD programme requires that thecandidate is in attendance at Imperial College for aminimum of 12 months during the 36 months. Dr TonyPitcher of the above mentioned group is the supervisorwho will visit the candidate for a specified periodannually. He has proposed that the candidate be inresidence at Imperial College as follows:

1990/91 Academic Year - 3 months (October toDecember 1990) this has been accomplished, hence

the temporary suspension of the samplingprogramme. The candidate made contacts with theLarval Ecology Team at DAFS in Aberdeen Scotlandand also attended a 2 weeks course quantitativeplankton sampling at the Freshwater EcologyWindermere Laboratory during the stay at Imperial.

1991/92 Academic Year - 3 months for initialanalysis of the data collected for the rest of

1990/91 Academic year. October to December 1991is proposed for activity as it will allow thecandidate to also register for his 2nd year atImperial College.

1992/93 Academic Year - 6 months for final writeup with constant consultation with thecandidate's supervisor. The last 6 months i.e.May-September 1993 is suggested for this following15 months of data collection and of the write up.

WORK TO BE DONE

The initial stage will involve attaining objectives (i) and (ii)

as these ensure that correct information about the age of thefish is used when determining other parameters about its lifehistory.

The move encompassing aspect of the programme will investigate

the following.

Kapenta spawning grounds, spawning times and thefrequency with which they spawn. It is generallyaccepted that kapenta has two spawning times within a year. What is the relative strength of thespawning population? Is there any overlap betweenthe two peaks thought to exist. These aspectshave a bearing on thé extend to which the fishcould be exploited.Kapenta larval dispersal and survival.

Determination of age-at-recruitment and whatenvironmental factors trigger this

(i)

(iii) Kapenta growth rates and factors influencing

23

these.

(iv) The candidate and his supervisor will need toexchange visits as suggest above.

The sampling programme will be designed in such a manner thatmost work will be accomplished concurrently.

24

REPORT ON SHORT COURSE IN ZOOPLANKTON QUANTITATIVE SAMPLINGMETHODS HELD AT THE FRESHWATER BIOLOGY LABORATORY WINDERMEREFROM 19 TO 30 NOVEMBER 1990.

BY

M.Z. MTSAMBIWA.

The aim of the two weeks course was to familiarise me , on mysupervisor Dr Pitcher's recommendations, on methods toquantitatively sample planktonic organisms. This was designed toprepare me for sampling freshwater sardine eggs and larvae andzooplankton as the prey for the larvae and juveniles on LakeKariba.

The course, which was tailored and tutored by Dr Glen George ofThe Freshwater Biology Laboratory Windermere as per ourrequirements, covered the following topics ( also see AppendixI):

Freshwater plankton composition, their biology and thestructure and function of selected Cladoceran andcopepods species.

Zooplankton population dynamics within a simple foodchain structure i.e. phytoplankton - zooDlankton -f ish larvae and methods of estimating biomass andproduct i on.

Basic physics and hydrodynamics o-F thermally stratifiedlakes and their influence on zooplankton spatialdistribution.

Zooplankton vertical distribution and its significance.

Methods of sampling zooplankton, eggs and fish larvae.

Practical work involved:

Sampling of planktonic organisms from Lake Windermerefrom a FBA research boat using nets, tubes and thecollection of physical parameters such as dissolvedoxygen, temperature, pH and conductivity at differentdepths.

Taxonomy : sorting out and classification o-F differentplanktonic organisms and their preservation.

The use of a PROFILER II designed at FBA tocollect information on DO, quantitative lightpenetration at different depths, temperature,conductivity and pH which are all recorded ontoa computer for later retrieval and analysis.

25

Most of the theoretical aspects of the course treated the tooicsdealt with during the Aberdeen visit of 24-25 October 1990 inmore detail with special reference to freshwater ecosystems. Theonly difference between the two is climatic in that Windermereis in the temperate region while Kariba is tropical.

Of the topics covered on the course this report will deal mainlywith the onerelated to Methods of sampling zooplankton and larval fish andthe preservation of such specimens.

The objective here was to outline and compare in more detail themost commonly employed methods of sampling planktonic organismsi.e. zooplankton, eggs and fish larvae. Emphasis before asampling method is selected is placed on the following factors:

the size range of the animals to be sampled, theirability to avoid different collection devices withage, body shape and behaviour (feeding migrations,predator avoidance strategies etc).

characteristics of the environment i.e. shallow ordeep lake ( whether stratified or mixed), weedy orsandy or muddy littoral zones.

the objectives of the investigation i.e. qualitative(systematic; distributional; life histories andcommunity structure) OR quantitative (production;biomass; population dynamics or physiological studies).

Different devices and sampling procedures must therefore beemployed when studying different organisms, lakes and problemsas no single device can make quantitative collections of thecomplete spectrum of organisms. Criteria for selection will alsodepend on the volume of water to be sampled, the depth of thestrata, the actual kinds of organisms and whether integrated orpoint samples are needed.

The devices for quantitative plankton collection can beconveniently divided into two basic categories:

1) Devices based on COLLECTION of water samples i.e.pumps, bottles and tubes. In this case the most directmethod for collecting plankton is to remove it from aknown volume of water sample collected in a bottle.The general feature is that only a small amount ofwater is sampled and the method is efficient for smalland less motile organisms. Therefore a large numberof replicates are needed in order to have a precisecensus of the organism and to determine representativefigures. As a result bottle sampling is not recommendedin large, deep and oligotrophic lakes but is suitablefor small, shallow and eutrophic lakes. Tubes can beconsidered as a sort of long bottles, usually flexibleand collect an integrated sample when lowered throughthe water column. Their length must be equal to the

26

27

maximum depth to be sampled and find their greatestapplication in shallow environments or littoral areasrich in vegetation. Two types of pumps are presentlyused i.e. hand pumps and motorised pumps.

2) Devices based on the FILTRATION of planktonic organismsfrom water directly in the field i.e. plankton nets.towed plankton samplers and plankton traps. Planktontraps present some advantages over bottles in that theyhave rapid closing systems large mouths which reducethe possibility of avoidance reactions and they allowthe simultaneous collection and filtration of a largevolume of water. They can also be easily handled by asingle person in a small boat and some have remoteclosing systems. Traps present one of the highestdegrees of efficiency especially with regard to thespecies that display the strongest avoidance reactions.Perhaps the simplest and mostly widely used planktonnet is a conical net with a plankton bucket at itslower end. Others have a reducing cone forward of themouth and a simple closure system. The shape andstructure greatly affects the quantity of water thatcan pass through the mouth of a net. Other factorsaffecting net efficiency include the fabric used toconstruct the net (gauze), mesh size, porosity, speedof sampling, avoidance by target organisms, escape ofsampled organisms and clogging. Thus choice of net willdepend on some of these factors, for example the fabrictype used in net construction has a marked effect onthe selectivity of the net, filtration efficiency(ratio of the volume of water filtered by a planktonnet to the volume swept by the mouth) and clogging.

The following were some of the measures recommended to get aroundthe problems associated with reducing the efficiency of a deviceor method selected for sampling:

Plankton gauze should have the following properties;- the meshes should be square and the mesh aperture

uni form- the material of the strands should be stiff enough

to resist bending or stretching but flexible enoughto allow self cleaning action, preferably nylon orsi 1k.

- the nature of the weave should prevent strands fromsliding out of place and should prevent the meshesfrom distorting diagonally.the porosity should not change when the net isimmersed in water.

- the gauze should resist clogging andallow completeremoval of material after use.

Towing speeds should be high enough to reduce evasiontime but not too fast to cause damage to the capturedorganisms. The speed should also be kept constantthroughout the tow.

Contrast between sampler and background should bereduced by us-i ng dark-coloured samplers, shiny surfacesshould be eliminated especially on moving parts thatwould cause cyclic reflections.

Samplers should be free of forward obstructions suchas tow lines and bridles and should be streamlined asmuch as possible.

Samplers should be fitted with enough filtering surfaceto remain 85% efficient throughout the tow. SeeAppendix II for a proposed mounting of bongo nets infront of two connected canoes driven by an outboardengine or in front of an ordinary kapenta rig.

Samplers should be evaluated in the context of theenvironmental conditions and the purposes for whichthey are used.

PRESERVATION OF PLANKTON SAMPLES.

The use of 4% formaldehyde as a preservative was stronglyrecommended. However it is absolutely necessary that the formalinbe neutralised before use by the addition of either borax (sodiumtetraborate), marble chips or another suitable buffering agent.

Most commercial formal-in is often contaminated with ironcompounds which on neutralisation would produce a brownprecipitate of iron hydroxide which in turn spoils the sampleby sticking to the surface of the organisms and obliteratingtheir finer structures. The alternative is analytical gradeformal-in which should be stored in glass or plastic containers.

To avoid dissolution of calcareous plankton in highlyconcentrated samples, the strength of buffered formal-in shouldbe doubled in such samples.

The following references were also recommended:

Bagenal, T.B. and Nellen, W. 1980. Sampling eggs, larvae andJuvenile fish. In: Backiel,T. and Welcomme, R.L. (edsJ:Guidelines for sampling fish in inland waters. EIFAC Tech.Pap. 33:13-36.

28

Bottreli, H.H., Duncan,A., Gliwitz,Z.M.. Grygierek, E., Herziq.A. ,Hillbrícht- Ilkowska,A., Kurasawa,H. , Larsson, P. andWeglenska, T. 1976. A review of some problems in zooplanktonproduction studies. Norw.J. Zool. 24:419-456.

Dabrowiski, KR., 1981.The spawning and early life history ofthe pollan (Coregonus pollan Thompson) in Lough Neagh,Northern Ireland. Int. Revue ges Hydrobioll. 66(3):299-236.

George, D.G. 1974. Dispersion patterns in the zooplanktonpopulations of a eutrophic reservoir. J. Anim. Ecol. 43:537-551.

-and Edwards. R.W. 1976. The effect of wind on thedistribution of chlorophyll A and crustacean plankton ina shallow eutrophic reservoir. J. Appl.Ecol.13:667-690.

1981. Wind induced water movements in the South Basin ofWindermere,. Freshwater Biology 11:37-60.

- 1981. Zoopiankton patchiness. Rep. freshwat. Biol. Ass.49:32-44

- and Harris. G.P. 1985. The effect of climate on long-termchanges in the crustacean zooplankton biomass of LakeWindermere 316(28):536-539.

Hewitt, D.F., Lund, J.W.G. and Smyly, W.J.P. 1990. Therelative effects of enrichment and climatic change on thelong-term dynamics of Daphnia in Esthwaite Water, Cumbria.Freshwater Biology 23:55-70.

Gulland, J.A. 1980. General concepts of sampling fish. In:Backiel, T. and Welcomme, R.L.(eds.): Guidelines forsampling fish in inland waters. EIFAC tech. Pap. 33:13-36.

He Xi and Lodge, D.M. 1990. Using minnow traps to estimate fishpopulation size: the importance of spatial distribution andrelative species abundance. Hydrobiologia 190:9-14.

Heath,M. and Dunn, J. 1990. Avoidance of a midwater frame trawlby herring larvae J. Cons. mt. Explor. Mer. 47:140-147.

Hewitt, DP. 1979. Test to confirm quantitative sampling of youngfish by the Beganel buoyant net. Freshwater Biology 9:339-

341.

Kerfoot, W.C. 1985. Adaptive value of vertical migration:comments on the predation hypothesis and somealternatives.In:Migration: Mechanisms and adaptivesignificance (ed. M.A.Rankin), pp 91-113 Contributions in Marine Sciences 27,University of Texas. Port Akansas.

29

Kramer, D., Kalin,M.J., Stevens, E.G Thrailkill. J.R., Zweifel.J.R. 1972. Collecting and processing data on fish eggs andlarvae in the California Current Region. NOAA Tech. Rep.

NMFS Circ. 370.

Lambert, W. 1989. The adaptive significance of diel verticalmigration on zooplankton Functional Ecology 3:21-27.

Longhurst, AR. 1976. Vertical migration. In: Cushing. D.H. andWalshJ.J(eds.) The Ecology of theSeas. Oxford Blackwoll467 pp

Matarese, A.C., Kendall,A.W.Jr., Blood, D.M. and Vinter, 'B.M.19? Laboratory Guide to Early Life History Stages ofNortheast Pacific Fishes. NMFS. NOAA.

McNaught, DC. 1966. Depth control by planktonic Cladocerans inLake Michigan. Pub. 15. Great Lakes Research Division. TheUniversity of Michigan.

Nellen, W. and Schnack, D. 1975. Sampling problems and methodsof fish eggs and larvae investigations with specialreference to inland waters. EIFAC Tech. Pap.(23) Suppl.1.Vol.2:529-531

Tranter, D.J. (ed). 1968. Zooplankton Sampling. IJNESCO Monoqr.Oceanogr. Methodol., 2:174p.

Vil.janen, M. 1987. Evaluation of modified seine and bongo netsfor quantitative sampling of vendace (Coregonus albulaL) larvae. Aqua Fennica 17(2):179 -191.

30

(Q

(j . t

I

- a r t S - r . r t

CrQCQC L-'i -\QJ. - ot».

S o

31

/ 4

o

H. Nyaruwa

REPORT ON TRAINING

POST GRADUATE TRAINING - HUMBERSIDE POLYTECHNIC, UK.

1.1 LEVEL: a) Postgraduate diploma in Food Studiesand

b) MSc in Food Studies :- Post harvestFood technology

1.2 The Postgraduate diploma in food studies has food science,food technology and resource utilization as main courses.Specialist courses include Dairy Science, Fish Science,,Vegetable and Fruit Science and Meat Science. Students arerequired to choose one specialist course. Then a studenthas to write a dissertation on a particular subject. whichcould be of his/her choice, sponsors choice or collegechoice. The dissertation is based on literature review.

The MSc in Food Studies has depth studies in foodchemistry/biochemistry, food technology, including foodchemical engineering, resource utilization and a specialistcourse as well.

The project involves practical work and literature surveyleading to the submission of a bound thesis. The resourceutilization course has a management, economic and marketinginput at this level.

13 Postgraduate DiDloma Dissertation

It was on the quality aspects of beef. It looked at theeffects of cold, heat, storage time and environmental gaseson the nutritional levels and shelf-life of beef.

(b) MSc Proiect

This was on the overall influence of harvesting practises,distribution practises and marketing strategies on the overallquality of exotic fruits at the point of sale. These practiseswere then compared with standard practises required by the BS5750 (ISO 9000).

32

H. NYARUWA

TITLE: POSTHARVEST FISH TECHNOLOGY IN LAKE KARIBA, ZIMBABWE

Research Proposal

OBJECTIVES

(i To set up a technically suitablefish processing methodin the fishing villages in terms of material,efficiency yield, time and preservation of nutrition

(ii) To assess and identify the extent and causes of fishpostharvest losses.

JUSTIFICATION

No comprehensive studies on post-harvest fish processing havebeen done on the Zimbabwean inshore fishery. The rudimentarynature and general inefficiency of fish processing methods usedinthefishing villages, coupled with the unreliability of thelake-based fish collection system suggest that there are high

post harvest losses. Bourdillon j. (1985) suggestimprovements in fish processing methods as a way of reducing

post harvest fish losses. Improvements in the fish processingmethods will result in better quality processed fish.

METHODS:

Setting up of a processing method suitable for the local

environment

The screen and solar method

White and black polvthene will be used to constructthe drying house. . Vent-s to let moisture out of thehouse will be guarded with screen-s to prevent insectentry e.g. blowflies.

This method will only be suitable in the sunny dryperiods, about 8 months of the year. The method willbe set uo at the institute for demonstration purposesand to investigate technical problems.

Smoking method

A combination of a Chorkor and kiln methods will beused to set up a fish smoking unit.

This will be set up at the institute for demonstrationand to identify technical problems. This will then beset up in the fishing villages using local material as

far as possìble.

33

(iii) Assessment an.d reduction of postharvest losses

Physical, economical and nutritional losses will beidentified. The principal causes will be identified.

Data collected will be used to determine and categorisecauses and to prevent the losses.

REFERENCES

Bourdillon, M.FC., Cheater, AP. and Murphree, M.W. (1985)Studies of fishing on Lake Kariba. Mambo Press. Gweru.

34

R. A. SANYANGA

TITLE: ASSESSMENT OF THE ABUNDANCE OF INSHORE FISH STOCKS ANDEVALUATING THE EFFECTS OF FISHING PRESSURE ON THEBIOLOGY OF COMMERCIALLY IMPORTANT SPECIES ANDECOLOGICAL STUDIES ON SYNODONTIS ZAMBEZENSIS

OBJECTIVES.

To assess the inshore stocks of the lake.

To evaluate the effectiveness of protected or"reserve" areas through the assessment of a number ofbiological parameters.

To test the hypothesis that fish migrate from reserveareas to stock areas depleted from heavy fishing.

To evaluate inshore fish species and biomassdistribution and factors causing these.

(y) To study the bioloqical aspects of Synodontiszambezensis in relation to fecundity. growth rates.feeding habits natural mortality etc. to assess itsability to increase its population in heavily fishedareas.

(vi) To study the tolerance of S. zambezensis to low oxygenlevels and low temperatures.

(vìi) To study strategies adopted if any by commercial fishspecies in response to heavy fishing pressure. Thiswill be done by comparing in some detail thepopulation dynamics, growth rates, fecundity, size atfirst maturity, sex ratios etc of 2 principlecommercial species - Serranochromis codringtoni andOreoch romis mortimeri.

WORK ACCOMPLISHED.

Fish data (catch per unit effort, species composition, length,mass and sex) were collected in November/December, 1989,February and July, 1990 in four areas of the lake. Two areas(Fothergill and Chete) are unfished and the other two areas(Gache Gache and Sengwa) are fished (Fig. 1). In each area datawere collected from vertical transects running from shore todeep water. Along each transect three 42 m long Lundgren surveynets were set at 3m depth; three at 6m depth; one at 12m depthand one at 20m depth. Each net has 14 different panels withmesh size varying from 6.25 to 75mm.

At each transect environmental variables slope, transparency anddepth were also recorded.

35

The catch in each 42m long net consisted a single sample and onegill-net set over night (6.00pm to 6.00am) was considered oneeffort.

To identify the relationship between fish species compositionand species abundance with 3 major environmental variables -depth, slope and transparency, as well as 4 nominal variables -presence of fishing, absence of fishing, presence of vegetationand absence of vegetation, the data sets were analysed withDetrended Canonical Correspondence Analysis (DCCA) (ter Braak,1986, 1987). DOCA is multivariate direct gradient analysistechn i que.

It was not possible to use gill-nets to assess absoluteabundance of fish although the Lundgren survey nets have beenused for this purpose in small Swedish lakes (Degerman et al1988). Catch per unit effort (OPIJE) by species was thereforetaken as an index of abundance. The variables depth, slope andabUndance, were quantitative. The DOCA analysis was performedonly on the data from the Chete and Sengwa areas.

The other analyses that follow were performed on the fish fromthe Gache Gache and Fothergill areas.

Fish length-weight relationship was found by regressing lengthon weight. Bar graphs were constructed showing length frequencycomposition and the frequency of OPUE by species at differentdepth zones. The Minitab "two sample" t-test was carried out tocompare mean lengths and OPUE of species at different -sites.Significance was at the 5% level

In the study 25 out of a possible 41 species were recorded.However, only 16 species were recorded in the Fothergill andGache Gache areas.

In terms of distribution which is shown in Fig.2, the bulk ofthe species showed a preference for shallow water with onlySynodontis Zarnbezens/s being found in large quantities in deepwater. This means that S. zarnbezensis must have ability totolerate low dissolved oxygen tensions, and low temperatures.However, this fish was limited to the bottom and was not pelagicprobably because of its feeding habits.

Although other species' showed reduced CPUE in fished areascompared to protected areas, S. zarnbezensis had highest OPIJE infished areas, The high levels of CPIJE of S. zambezensis in thefished areas contributed to reducing the differences in CPUEbetween fished and unfished areas so that in the end thedifferences were not significant. This showed a strong tendencyfor S.zambezen5is to rapidly increase in biomass in fished areas.

At the same time, commercially cropped species showed a severedepletìon of small and large length classes (See Fig.3). Thiscould be due to cropping of mature and large individualscontributing to poor recruitment.

36

37

The study therefore shows that fishing is having an impact orthe ecology and biology of some species.

This work was done as part of an MSc programme at the Universityof Wales, Cardiff (October 1989 to October, 1990).

The other aspects of the programme covered statistics andcomputer modelling fisheries bioloqy, toxicology, waterpollution and control, water management, aquatic invertebratebiology and aquatic microbiology.

A desk study which follows was also part of the MSc programme.The study which was entitled: Analysis of length frequency basedmethods in estimation of growth parameters had the followingobjectives.

to compare estimates of growth parameters L andZ/K obtained from the same data but using fourdifferent length -frequency analysis techniques.The data used were collected in the Samn Sea andCangan bay (Phillipines) by Corpuz et al

(1985)

to get an insight into why different techniquesgive different parameters on the same data shouldthese parameters be different

to investigate the reliability of the Bevertonand Holt and the Wetherall method under varying,knife-edge selection length L. and constantasymptotic length L.

A summary of the study follows:

Four length based me,thods were compared in their estimation forZ/K and L. These were the ELEFAN, the Beverton & Holt, theWetherall and Pauly's method. The ELEFAN estimates for Z/K andL were calculated by Corpuz etal., (1985). The author used thesame raw data to recalculate Z/K and L with the rest of themethods. Because no predicted values were known comparisons weremade using non-parametric Mann-Whitney Statistics, linearcorrelation, and the coefficient of variation (CV %).

The findings showed that none of the methods was perfect. eachhad its own limitations. The ELEFAN proved to be robust over arange of data qualities. The BH method was probable the best ofthe four for estimating Z/K. It was also found that theWetherall and Pauly's methods gave erroneous results (severalmagnitudes high) for fish which grow in excess of 40 cm inlength. Recognition of error in L and Z/K estimates could beeasily judged in the case of the Wetherall and Pauly's methods,but was difficult to detect with the ELEFAN. It was found thatdata structure which conformed to the ssumptipns of any methodin question generally produced reliable estimates of L and Z/K.

WORK TO BE DONE

Comparisons between fished and non-fished areas will continueinto the remaining protected areas and fished areas. TheZambian shore line is not subdivided into fished and protectedzones, the whole shore line is heavily fished. It would be ofinterest to extend the study to cover the Zambian shore line aswel 1.

In conjunction with Mr. Mabaye (University Lake Kariba ResearchStation) radio tracking and floy tagging of tiger fish and somecichlids will be done to assess their mobility and the extent towhich they could migrate in the lake.

After the broad assessment of the abundance,autecologicalstudies on S. zambezensis will be carried out. This follows therealisation that this species increases its population rapidlyin areas commercial species have been cropped and seems to bethe only original species in the Middle Zambezi Rive systemwhich has shown ability to colonise greater depths. It couldalso be exploited commercially.

These studies will be carried out in conjunction with otherInstitute activities e.g. data base management.

REFERENCES

Carpuz, A., Saeger, J. & Sambilay. V. 1985Population parameters of commercially importanty fishes inPhilippine waters. Tech. Rep. Dept. Mar. Fish. 6, 1-99.

Dergerman, E. , Nyberg, P. & Appelberg, M. 1988.Estimating the number of species and relative abundance offish in oligotrophic Swedish Lakes using multi-mesh gill-nets. Nardic J. Freshw. Res. 64: 91-100.

Ter Braak, C.J.F. 1986. Canonical correspondence analysis: anew eyen vector technique for multivariate direct gradientanalysis. Ecology 67: 1167-1179.

Ter Braak, C.J.F. 1987. Canoco - a Fartran program forcanonical community ordination by (partial) (detrended)(canonical) correspondence analysis (version 2.1) TNOInstitute of Applied Computer Science. Wageningen. TheNeatherlands. Research Station Report etc.

3$

studied.

a

q

a

z

.-

- s;

7

J

s'r')

Key

[Eli Fisnecj area

Non sne area

a GacnebFothergU

C Sengwa

d

Figure 1. LocatiOn of the fished and unf-ished areas that were

S»

Cusita

a

i /

TF?ANSPAREN CY

S. zamo

L. miodI

S. nebu

FIS HEI

sO. ni acro

I R dan

O. morts

s

w

5. macro

I S. codr

S T rend

Figure 2.Ordination diagram based on canonical correspondence analysis of fish species(G) with respect three quantitative envirönmental variables shown by arrows andfour nominal variables - presence of fishing, absence of fisning, presence ofvegetation, absence of vegetation () L maid z Limnothrissa miodon, P darzPharincochrornis, dar1ingi;»B.late= B v'cinus lateralis; H.discz Hion000tanivrusdischorhvnéus O. mdrti Oreochromis. mortinieri; D. schen Distichodus schenca;M. longz Hortnyrus lonqi rostri's; B. 'irnbe= Brvcinus miden; H. foskz Hvdroc'inusforskahiii; L. aiti LabecT'altiveTis; L. cyiiz Labéc cviindricus; C. gariezCl arias oariepinus; E: depr= Eutrooius deoressirostris; S. rnacroz Serranochromismacrocaohalus;S.. codrz Serranodirocnis codrinotoni; T. rend= Tilapia rendafli;M.delizMormyroos deliciosus; O. macro= Oreocnramis macroceohalus; S. nebuSynodontis nebulosus; M. elect= Halaoterurus electricus; S. zamb Svnodontiszarnbezensis.

40

_ tTOs< eL.atti

Lii

Lii>

D. schenMonÇ SLOPE

B.irn:

O L.cytiC.garie

I H. de'i

H. discS 3arbus

DE PT H

SO. OD

40.00

20. OD

10. OD

0.00b

/

7/

w //

Id

jr4

Figure 3. The length frequency distribution oi rrn0mjccodrin.qtoni between fished and unfished areas.

41

4 a 12 20 24 Z8 32 35

LENGTH CLASSES (04)urrr T sr3 j /j r î s

g 30.00

PUB LI CATI ON S

Chifamba, P, 1990. Preference of Tilapia rendalli (Boulenger)for some species of aquatic plants J. Fish Bio]. 36: 701-705

Machena. C., Kautsky, N. & Lindmark, G. 1990. Growth andProduction of Lagarosiphonilicifolius in Lake Kariba- a manmadetropical lake. Aquatic Botany; 37: 1-15

Mtsambiwa, M.Z. . & Pitcher, T.J.Estimation of hatching periods for kapenta Limnothrissamiodon(boulenger) based on otolith and readings from larvae andjuveniles. In prep.

Nyaruwa, H. , 1990. An integrated management system duringharvesting, distribution and marketing to improve thequality of exotic/tropical fruits M.Sc thesis. HumbersidePolytechnic U.K.

Nyaruwa, H., 1989. Some quality aspects o-F beef. Postgraduatediploma dissertation. Humberside Polytechnic, U.K.

Sanyanga, R.A. 1990. Analysis of length-frequency based methodsin estimation f growth parameters Loo and Z/K and Acomparison of gill-net catches from a fished and anunfished area of Lake Kariba (Zimbabwe). M.Sc thesis.University of Wale.s College of Cardiff.

42

Aquatic Botaiu'. 37 (1990)1-15Elsevier Science Publishers B.V., Amsterdam - Printed in The Netherlands

Growth and production of Lagarosiphonilicifotius in Laké Kariba - a man-made tropical

lake

Cecil Machena', Nils Kautsky2 and Gunilla Lindmark3'Lake Kariba Fisheries Research Institute, P.O. Box 75, Kariba, Zimbab%,e and Institute of

Ecological Botany, Uppsala University, P.O. Box 559, S-751 22 Lfopsala ('Sweden)2Askö Laboratory, Institute of Marine Ecology and Department of Zoology,

University Stockholm, S-106 91, Stockholm ('Sweden)3lnstitute of Limnolo8y, University of Lund, Box 65, S-22 1 00, Lund (Sweden,)

(Accepted for publication 31 October 1989)

ABSTRACT

Machena. C., Kautskv, N. and Ltndmark, G., 1990. Growth and production of Lago rosipiton ilicifol-lus in Lake Kariba - a man-made tropical lake. Aquat. Bot., 37: l-15.

Two different methods, the diurnal oxygen curve method in plexiglass enclosures, and the incre-mental growth techntque vere used to complement production estimations of Lago rosipho,t ilicifolit,sOberm. in Lake Kar,ba. The incremental groth technique gave a production rate of 16.4 mg g ' dryweight (DW) day' (7.5 mg C g-' DW day-') which was 6.5 times higher than the rate of 1.16 mgC g' DW day' obtained for the community from the diurnal oxygen curve method. The differenceis largely accounted for by the total respiration of the community, which is inherent to the enclosuretechnique, and alsoby the fact that the measurerbents were carried out for short periods oftime. Usingthe production rate from growth measurements, the biornass production of Lagarosiphon was calcu-lated as 1.6 g D\V m2 day' and to 598.6 g m2 year-'.

The average community FIR ratio of 1.16 indicates that the community is operating close to steady-state, and shows a high degree of self maintenance.

Lagarosipho,t has an annual turnover rate of 2.4 but the populations are perennial (i.e. grow allyear round). The main growing season is in summer, from October to February. Cohorts that developduring the summer, branch extensively at the end of this growth period, and while occluding light tothe bottom, die off in the process. The resulting drifting mats fragment and constitute the major re-productive modules. Although flowering was observed, sexual reproduction does not appear to beimportant.

It is concluded that the difference in the production values between the community taken as a wholearid individual plants reflects the role different components of the ecosystem play in the metabolismof the macrophyte comritunity. For this reason the value of 16.4 mg g-' day' calculated from markedindividuals is a more realistic growth rate of Lagarosiphon plants in this study.

0304-3770/90/S03.50 © 1990 Elsevier Science Publishers B.V.

C. MACHENA ET AL.

INTRODUCTION

This study was carried out to estimate the productivity of a Lagarosiphonilicifolius Oberm. dominated community and to follow the growth and phen-ology of Lagarosiphon in Lake Kariba.

A number of standard methods for primary production measurements havebeen recommended (Vollenweider, 1974; Kemp et al., 1986). In the presentstudy we use two of them: changes in growth over time in marked plants, andin situ studies of oxygen production and its seasonal variation in plexiglassenclosures.

Benthic macrophytes (i.e. macrophytes living on the bottom of the lake)form productive communities (Cattaneo and Kaiff, 1980; Morgan and Kit-ting, 1984) and are important for fisheries. Their contribution to the totalbiomass production is generally related to the size and morphometry of thewater body. For some shallow lakes e.g. Swartvlei (Howärd-Williams, 1978)and Lake Chad (Lévêque et al., 1983) hydrolittoral (i.e. the zone extendingfrom the water-line to the depth at which plants are rooted) production ex-ceeds pelagic planktonic algal production. However, for deeper lakes the op-posite usually holds.

Benthic macrophytes are also important in the lake as.a substrate for micro-and macrofauna and epiphytic algae (McLachlan and McLachlan, 1971;Wetzel & Hough, 1973; Pelican et al., 1978; Howard-Williams and Allanson,1981; Morgan & Kitting, 1984; Carpenter and Lodge, 1986). Machena andKautsky (1988) argued that the development of the submerged vegetation inLake Kariba facilitated the development of other benthic communities, par-ticularly gastropods and bivalves.

As Lagarosiphon is the dominant species regarding both biomass (Mach-ena and Kautsky, 1988) and area covered (Machena, 1987), productionstudies on this species may give an indication of the role of macrophytes inthe ecology of the lake.

MATERIAL AND METHODS

Study area

Lake Kariba (l6°28' to l8°06'S;26°04' to29°03'E) is an artificial lakeformed after damming of the Zambezi River. It is located in Southern Africabetween Zambia and Zimbabwe and is the third largest man-made lake in theworld. Its features and morphoogy are presented in Table 1. The main axisof the lake has a SWNE orientation following the old Zambezi River basin.(For details see Balon and Coche, 1974).

Two sites were chosen, one close to the Fisheries Research Institute at theeastern end of the lake where community metabolism studies were carried out

LAG.4ROSIPJIONfLIC/FOLIL'S!N LAKE KARIBA 3

TABLE 1

Main features and morphology of Lake Kariba at a water level of4S5 m a.s.l. Partly recalculatedfrom Balon and Coche (1974), and potential colonisable areas of submerged macrophytes

Length 277.0 kmWidth (mean) 19.4 kmDepth (mean) 29.2 mDepth (maximum) 120.0 mArea 5364.0 km2Average bottom area per

1 m depth intervalbetween 0-15m 105.2 km2 m

Total bottom area betweenO and 5 m deep 526.1 km2

Volume l56xl0°m3Percentage of bottom area

between O and 15m deep 23.5%Total length of shoreline 2668.0 km

and the other half-way along the Zimbabwean side of the lake where the growthof individual plants was followed. Site 1 is on an exposed shore, which isregularly subjected to turbidity from wave action. Studies at the site were car-ried out at a depth of I m. Site 2 was situated 4 m deep and about 1 kmoffshore. The standing crop of Lagarosiphon was similar at both sites and alsosimilar to the mean standing crop for the lake within the colonised zone be-tween depths of O and 5 m (Machena and Kautsky, 1988).

Measurement of gro wth rates

Growth rate was measured on individual1y marked Lagarosiphon ilicijoliusplants that had grown from fragmented shoots. In October when the plants.were small (unbranched with an average length of 20±2.65 cm), a total of40 plants in 4 (0.5x0.5 m) quadrats were individually marked. Pieces ofsteel rods were pushed into the ground alongside each identified ramet (Har-per, 1977) to distinguish these from plants that would establish later duringthe study. The positions of the plants in each quadrat were then marked onroughened plastic sheets (for use under water) on which all data wererecorded.

The marked plants were assumed to be of the same age and are treated asone cohort. The total length of each shoot and its branches were measured incm once every month from October 1986 until the plants senesced iii April1987.

Lagarosiphon is rooted and grows vertically with little branching until itreaches close to the water surface where it branches extensively thereby in-

4 C. MACHENA ET AL.

creasing its photoreceptive surface area. A length:weight regression was cal-culated using plants collected from the same population as the marked plants

Y=3.04X-5.5 (n=36; r20.96) (1)

where Y=dry weight (mg) and X=total length ofshoot and branches (cm).This equation was used to calculate the weight increments for each month.Relative growth rates were calculated using Eqn. (2) as outlined in Hunt(1978)

Rloge W2log W,

- T,T,where W, is dry weight in g at time T, in days and PV2 is dry weight at timeT2 and e is the base of the natural logarithm.

The mineral ash (inorganic) content of the plants, which amounted to17.5%, (as measured once in the early part of the study) was subtracted fromthe weight increments. This provided values for changes in the organic weight(Hunt, 1978) expressed per unit dry weight of the plant biomass and wasfound to be similar to the value of 1 5% reported for Cera tophyllurn dernersurnL. by Lipkin et al. (1986).

Areal photosynthetic rates were calculated using Eqn. (3):

Pr=Cû mg g' day'xBwhere Pr=2areal production; C=changes in organic weight per unit weight ofplant per unit time (mg g-' day-1); Bbiomass (g DW rn2) of Lagarosi-phon. Changes in weight in this paper refer to organic weight. The carboncontent of Lagarosiphon was assumed to be 46% of the total organic weight(ash-free DW) (Davies, 1970; Westlake, 1965 in Solander. 1982); the ashcontent was determined by ignition°n a muffle oven at 550°C. Denny (1985)gives the carbon contents of leaf tissues of Potarnogeton pectinatus L. between42 and 49%. Annual variation in the carbon content of the plants was as-sumed to be insignificant (Solander, 1982).

Measurement of community metabolism

Oxygen production and consumptionThe community metabolism of Lagarosiphon was calculated using diurnal

oxygen curves as described in Jansson and Wulff (1977) (see also Dybern etal., 1976). The communities were enclosed in plexiglass cylinders in situ witha lid on top, which enclosed a surface area of 0.066 m2 and a water volume of22.4 1. The lid had two outlets, one for withdrawing water samples and theother one for inserting the oxygen probe. The cylinder was pushed gently intothe substrate to ensure a-substrate seal -and was held in position by four steelbars to prevent tipping of the cylinder from wave action.

TA

BLE

2

Laga

rosi

phon

ilic

i/bliu

s co

mm

unity

met

abol

ism

in p

lexi

glas

s en

clos

ures

. Car

bon

equi

vale

nts

of p

rodu

ctio

n an

d re

spira

tion

valu

esar

c gi

ven

in p

aren

thcs

cs

Per

iod

Lag,

inG

ross

pro

duct

ion'

Are

al2

of s

tudy

3cy

linde

r(C

l': m

g 02

g- d

ay-

)G

ross

pro

duct

ion

(g D

\V)

(/1

1': g

02

m2

4,y

Are

al2

com

mun

ityN

et c

omm

unity

Com

niun

ityre

spira

tion

prod

uctio

nre

spira

tion

lI'/il

R(,

tR: g

02 m

2 da

y')

(1W

: mg

C g

' day

-')(C

R: m

g 02

g' d

ay-')

'Pro

duct

ion

or r

espi

ratio

n va

lues

er u

nit w

eigh

t of p

lant

obt

aine

d by

div

idin

g ch

ange

s in

oxy

gen

leve

ls in

the

encl

osur

e by

the

wei

ght o

f pla

nt h

arve

sted

in th

een

clos

ure.

2A,'e

al v

alue

s of

pro

duct

ion

and

resp

iratio

n ob

tain

ed b

y ex

trap

olat

ing

tota

! val

ues

with

in th

e en

clos

ure

to m

23R

cfer

s lo

Fig

. 3; L

ag. =

L. i

licifo

lius.

A 2

3-25

/II/8

55.

919

.93

(7.4

9)1.

78 (

0.67

)1.

55 (

0.58

)0.

9817

.31

(6.5

1) 1

.15

1326

-28/

01/8

67.

330

.9(1

1,6)

3.4

(1.3

)3.

1(1

.2)

0.9

28.4

(10.

7) 1

.09

C 1

5-17

/02/

869.

031

.7(1

1.9)

43! (

1.6)

6.1

(2.3

)4.

944

.9(1

6.8)

0.7

11)

10-

12/0

5/86

2.9

29.6

(11.

13)

1.30

(0,

49)

1.8

(0.4

1)4.

8616

.66

(6.2

7) 1

.20

E 0

7-09

/06/

863.

824

.75

(9.3

0)1.

19 (

0.44

)0.

96 (

0.37

)3.

116

.68

(6.2

)1.

24F

26-

28/0

7/86

8.7

22.9

(8.6

1)3.

02 (

1.11

)1.

7(0

.6)

3.6

13.0

9(4

.92)

4.77

G 1

5-17

/03/

888.

79.

48(3

.56)

1.24

(0.

47)

.33

(0.5

0)-0

.24

10.1

2(3

.80)

0.93

6 C. MACHENA ET AL.

Diurnal measurements were begun at dusk and continued for about 37 h,thus covering a full day and the nights before and after. A battery poweredbilge pump stirred water within the cylinder -ither continuously or at inter-vals. Dissolved oxy.gen and temperature readings were taken every hour usinga Yellow Springs 58 oxygen meter with a combined oxygen and temperatureprobe. At the end of each experiment (carried out during the periods indi-cated in Table 2), the enclosed plants and animals were collected, separatedby species and dried to constant weight.

Oxygen production and consumption were calculated from diurnal oxygenmass time curves. The respiration rate was calculated as the best fitting slopeon night consumption. This was extrapolated to 24 h to obtain total dailyrespiration. When the best fitting slope on night consumption was continued,the height between this slope and the mass time curve after 24 h of incuba-tion, gave daily gross oxygen production.

Commudity production and consumption rates were calculated on a unitarea basis. As Lagarosiphon was the dominant producer, production and con-sumption rates were also calculated on a Lagarosiphon unit weight basis.

pH, conductivity, turbidity and light measurementThe physical parameters pH, conductivity, turbidity and photosynthetic

active radiation (PAR), were measured in the enclosure to help elucidate bi-ological and physical processes in the enclosures. Water samples were takenfrom the lake medium and siphoned from the experimental cylinders every 3h and pH, conductivity and turbidity were measured immediately.

Conductivity was determinéd with a portable Crison meter (model 523)and pH was measured with a Crison portable pH meter 506 and a combinedpH electrode. Turbidity was measured as NTU (normal turbidity units) \vitha Hach model 16 800 portable tubidimeter. Light penetration (photosyn-thetic active radiation PAR, 400-700 nm) adjacent to and at about the centreof the cylinders was measured with a LI-COR-188B integrating quantum sen-sor.(submersible) and meter. Total daily PAR was derived by planimetricintegration of hourLy PAR values.

RESULTS

Growth measurement

The growth curve (DW) and phenologv of individual shoots of Lagarosi-phon are shown in Figs. i and 2. The plant weight increased rapidly from amean of 61.6 mg per shoot in October to a mean of 225 mg per shoot in Feb-ruary, when the shoots matured.

The production rate was consistently high between October and Januarywhen flowering and rapid branching occurred. By February the shoots were

LG.I ROSIPHON ILICIFOLI LS I N LAKE KAR IBA 7

300-

250-

g 200-- FLOWERING

BRANCHiOGH 150- /g too-D

50-

o

120

H-

oL 100

AS O N DJ F MAMJFig. 1. Growth of Lagarosiphon ilicifolius L. in Lake Kariba. The onset of branching and deathis indicated. Vertical bars are SE of means.

WATER SURFACE

o

GE AT H

00

EU

ó0 IH-ozN

20

o

BRANCHING ANDFLOWERING

DRIFTINGFRAGMENTS

V

-

Fig. 2. A diagranrnatic representation of the growth cycle of Lagarosiplton iIicitblitts in LakeKariba showing a marked density of mâture shoots at the water surface.

IH-o

80D

60

40

20

8 C. MACHENA ETAL,

extensively branched and attained a maximum length of 160 cm. Senescencebegan in February, as evidenced by shoots rotting from the substratum, andby April all plants were detached or dead. As th'e marked shoots were assumedto be of the same age, the growth period of the cohort can be estimated at 5months.

As the Lagarosiphon populations are perennial in Lake Kariba, this gives aturnover of 2.4 for overall annual net production. Rich et al. (1971) reportedan annual turnover of 1.5 and 3.0, using two different techniques for macro-phytes in a Michigan marl lake, and projected a possible range of 0.5 to 5 formacrophytes in general.

The mean net production rate for measurement of individual shoots was16.4 mg g' DW day-'. By dividing the daily production rate by 10 (hoursof day- light), the hourly value of 1.7 mg g' DW was obtained. In terms ofnet carbon production this is equivalent to 7.5 mg Cg-' DW day'. Calcu-'lations of areal production rate were made using a mean biomass of Lagaro-siphon of 100 g m2 (for the colonised zone of 0-5 m water depth given inMachena and Kautsky (1988). A mean areal production rate of l.6x l0 mgDW m2 day-', equivalent to 754 mg C m2 day' was obtained.

Community metabolism

The in situ enclosure experiments were performed during different seasons(Table 2) to cover variation in environmental conditions and developmentphases of Lagarosiohon (Fig. 3).

The diurnal changes in pH, PAR, turbidity, temperature, conductivity anddissolved oxygen are presented in Fig. 3 A-G. As expected, values for tem-perature, pH and oxygen were lowest in the early morning and peaked late inthe afternoon, reflecting normal ph'sical and biological activity. The pH var-ied by 0.5-1.3 units during 24 h and the total pH span was 6.7-9.5. The in-crease in pH during photosynthesis is due to the utilisation of carbon dioxidein the production process. The diurnal amplitude for water temperature was2°4 ° and the total temperature range covered by the experiments was 22°-33°C.

Conductivity values ranged between 109 and 145 ,u S cm' during the day.The changes in conductivity are more complex and less consistent, as severalinterlinked factors and community components are involved in the mineral-isation process. PAR varied irregularly and was inversely related to turbidity(Fig. 3). However, the low PAR values in the early part of the incubationobtained in March 1988 were dte to the light sensor being shaded by vegeta-tion. From variations in the oxygen production and consumption in the en-closure experiments, values for gross production and community respirationhave been estimated (Table 2). The variation in the weight of the plants indifferent enclosures was considered when calculating the productiôn and res-

A

B

C

G

o-

mg 1

o

4

2

8

o

4

2

8

o

4

2

9

8

o

4

2

10

8

o

4

2

12

lo

CONDUC-TIVITY

120

100

110

loo

150

140

130

120

110

120

110

100

90

80

130

120

110

100

90

130

20

lo

00

o

80

120

110

00

o

T C/\ H

-30

28 100-208

208O1O

7

E-

-20

-20:-

-10 o'

E ::

-29

: -

-22 -208:

-21 -lo -- 7-

-28

8

o

4

2

F

-a--k-

OXYGENTEMPERATUREpH

o

-- CONDUCTIVITY 4

PARLL1 TURBIDITY

LAG,IROSIPJION ILICIPOLIUS I LAKE KARIBA 9

18 24 06 12 18 24 00HOURS

Fig. 3. Diurnal variations of dissolved oxygen, pH, temperature and conductivity within theenclosure and of photosynthetic active radiation (PAR) and turbidity adjacent to the enclosureduring different experiments. A: 23-25 Novehìber 1985; B: 26-28 January 1986; C: 15-17 Feb-ruary 1986; D: 10-12 May 1986; E: 7-9 June 1986; F: 26-28 July 1986; G: 15-17 March 1988.

D

E

piration values. For comparison, values from the different experiments areexpressed as oxygen as well as carbon per unit dry weight of Lagarosiphon(Table 2).

The net community production rate varied between 4.9 and 4.86 mg Cg' day' over the period with a mean value of 1.16 mg C g' DW day.Given a mean standing biomass of Lagarosiphon of 100g m2 between depthsof O and 5 m (Machena and Kautsky, 1988) the net Lagarosiphon commu-nity production was 0.12 g C m2 day-' or calculated for a year 42g C m2year', which is 6.5 times less than the value obtained using direct measure-ment of changes in the weight of marked individual plants.

Production:respiration ratios varied between 0.7 and 1.7 (Table 2). Theratio is lowest in February and increases to a maximum in July after which itbegins to decline. Overall P:R ratios for the total community metabolismwere 1 . 16, indicating a community close to a steady-state.

DISCUSSION

The growth pattern of Lagarosiphon in Lake Kariba is similar to that ofPotarnogeton pectinatus in Lake Swartvlei (South Africa) (Howard-Wil-liams, 1978). Both species have perennial populations with cohorts in differ-ent stages of development more or less during the whole year. Maximum pro-duction of Potamogeton occurs in summer followed by a massive die-off ofshoots and, like Lagarosiphon in Lake Kariba, growth and decay continue allyear round. In Lake Kariba, the biomass ratio of senescent to young shoots ishigh (there are fewer young shoots than mature shoots) between Februaryand October, so that even though the P R ratio is high during other seasons,the production may be higher in summer.

The mean production rate of Lagarosiphon derived from the enclosure ex-periments (over 7 months; Table 2) of 1.16 mg C g-' DW da'-'is 6.5 timeslower than the rate of 7.5 mg C g-' DW day' calculated from growth studiesof marked plants. The large difference is due to the fact that the two tech-niques show different aspectsof production, and the result from the enclosureexperiments is a mean over different seasons. For example. the May produc-tion rate (Table 2) is fairly comparable with the rate derived from changes inplant length. However, the inclusion of results from both studies provides anenhanced understanding of community processes (cf. Kemp et al., 1986). Inthe enclosure experiments, net primary production is highly influenced by thetotal respiration of the community. In Lake Kariba, sediment respiration(measured by comparing respiîation in a Lagarosiphon enclosure and in an-other enclosure covering bare sediment patches) is high, often accounting forasmuch as 42% of the respiration in enclosures (Machena., et aL, 1989). Thisis probably one of the explanations for the lower net production rates foundin the enclosure experiments.

Io C. MACHENA ET AL.

L4GJROSIPHO,VILICIFOLILSIN LAKE KARII3A

However, from an ecosystem point of view, the measurement of commu-nity metabolism is attractive because it reflects the production of the entiresystem as well as its successional states. On the other hand, intermittent mea-surements of growth increments or harvesting give an integrated measure(over time) ofthe net production reflected as plant biomass. Growth incre-ments or harvesting techniques though, will not take into account losses dueto fragmentation and exudation.

The large differences in the results from the two techniques could also beexplained by the fact that the short-term measurements in the enclosures didnot reflect the continuously changing environmental conditions in the opensystem.

Light and turbidity changed continuously during incubation ( Fig. 3 A-G).The enclosures were set up in shallow water and on a number of occasions,the wind blew offshore at night and onshore during the day. This caused tur-bidity during the day and limited light penetration. Hence, in view of thesechanging environmental conditions any extrapolation of short-term results toobtain production (yield) values for longer periods will thus present errors(Lipkin et al., 1986). On the other hand, measurement of growth gives thecumulative net production over an entire range of changing environmentalconditions and therefore a more representative yield (Lipkin et al., 1986).

The production rate of Lagarosiphon in Lake Kariba calculated to be 16.4mg g-' DW day' or 7.5 mg C g-' DW day-' (0.8 mg C g-' DW h-') israther low compared with literature data for other macrophyte species. Wes-tlake (1975) concluded that optimum production rates- for a great variety ofspecies are usually between 2 and 10 mg C g' DW h-' and a mean value of4 would be typical. However, the rates given by Westiake are maximum val-ues; these are often taken under optimum light conditions and at the peak ofthe growing season and may not rlect a consistent rate of growth throughoutthe growing season, which should be lower than the maximum values. It shouldbe pointed out that the hourly rate of growth, which we compare with Wes-tlake's. has been calculated assuming production over a 10-h period of day-light thereby underestimating our rate of Lagarosiphon production in LakeKariba; if the maximum rate were calculated it may well fall within the rangegiven by Westlake (1975). Also the photosynthetic rates of Lagarosiphoncould be underestimated from the measurement in growth changes, as pho-tosynthetic tissues represent only a portion of the whole plant. though thewhole plant will be respiring.

Our growth rate value of 16.4 mg g-' DW day-' for Lagarosiphon is alsolower than the value of 25.7 ng g-' DW day-' for Potarnogeton ihunbergiiCharm and Schlecht. calculated from the value of 0.18 g g-' DW week'given in Denny (1985). Denny also gives a growth rate of 0.46 g g-' DWweek ' (65.7 mg g' DW day-') for P. schweinfurthii A. Bennett. The valuesgiven by Denny (1985) for P. thunbergii and P. schweinfurthii indicate high

rates of production consistent with the growth of tropical species. However,these plants were grown in ponds under idegi conditions. Furthermore thevalues incorporate changes in weight ofshoots rhizomes and roots. Contraryto this, growth changes of Lagarosiphon in Lake Kariba were followed indeeper water ( depth varied from 3 to 4 m ) and light could have been limitingas the maximum biomass of Lagarosiphon in Lake Kariba occurs betweendepths of 2 and 3 m (Machena and Kautsky, 1988). Measurements werestarted at a depth of 4 m to prevent exposure of the plants during the studyfollowing annual fluctuations in the lake level. At the end of the study depthwas 3 m. Lake Kariba has a mean annual drawdown of 3 m.

Total annual dry organic matter production ofLagarosiphon in Lake Kar-ibaranges from 12.8 to 861.3 gTDW m2year' (mean of492.8 gDW m2),with aLagarosiphon biomass in the lake ranging from 2.6 to 174.8 g DW m2(Machena and Kautsky, 1988). The within lake variability of macrophytestanding crop and production per unit area are about two orders of magni-tude, which is similar to findings in other lakes (Carpenter and Lodge, 1986).

The bio-energetics trend of the community is indicated by the month tomonth P:R ratios (Table 2). The P:R ratio for January is 1.09 and indicatesa system close to steady-state. Maintenance metabolism is significant andequals the gross primary production. This ratio coincides with flowering andextensive branching of plants. In February the community progresses to het-erotrophy with a P : R ratio of less than one. The excess of consumption overproduction is indicative of the declining state of the plants at this time of theyear when there is a large scale die-off ofshoots. The situation is reversedafter May, when the system becomes autotrophic. The P: R ratios are similarto those recorded in other studies e.g. Hannan and Dorns (1970). Murrayand Wetzel (1987) give an account of similar processes in the seagrass (Zos-tera marina L. and Ruppia inartima L.) communities in Chesapeake Bay,U.S.A.

An annual summation of gross production and respiration of the Lagaro-siphon community in this study gives an average ratio of 1.16. This indicatesa high degree of self-maintenánce throughout the whole study period. As theplants are the major producers of the community, this shows that a large pro-portion of the community production is stored in plant tissues that are notreadily utilised by herbivores. This is consistent with our observation thatthere is little grazing on Lagarosiphon.

All plants seemed to grow from fragmented shoots and we saw no evidenceof plants developing from seeds although much flowering was observed be-tween November and February '(summer). Sexual reproduction does not seemto play an important role in the maintenance of many submerged macrophytepopulations (Sculthorpe, 1967; Haag, 1983; Riemer, 1984; Kautsky, 1987),although seeds may be important to ensure survival in the event of disaster(van Wijk, 1983; Riemer, 1984). Mitchell (1969, 1970) also found that Sal-

¡2 C. MACHENA ET AL.

LIGA ROS/PI/QN ILICIPOLIIJS I N LAKE KA RIBA 13

vinia molesta D.S. Mitchell (a floating water fern) in Lake Kariba producessterile spores and reproduces vegetatively. Length ofLagarosïphon varied froma few centimetres on unproductive sites to about 4 m in deep and shelteredareas ofthe lake (Machena, 1987).

Not all plants flowered, branched and died in summer; cohorts which de-veloped in summer, invested considerable energy in vegetative propagation.During the other seasons, little branching and flowering were evident thoughdifferent cohorts are still going through a similar cycle. The community pre-sents a dynamic flux comprising phases ofbuilding up, high maintenance anddie-off.

The steady-state is attained during the extensive branching period and pro-gresses to heterotrophy at the massive die-off period. After the die-off thereis increased light penetration through the water column that enhances growthof other cohorts leading to a positive energy balance, which is maintained forthe remainder of the year.

The die-off results in the formation of entangled floating mats of La gaio-siphon. At this stage the plant fragments have already started to develop roots,which enable a rapid establishment after contact with the substratum is made.These plant fragments either continue drifting, or sink and establish. In ter-restrial communities vegetative propagation is efficacious for maintenanceand new establishment around established plants (Grime. 1979). In the caseof Lagarosiphon, vegetative propagation is important for colonisation of newareas and this might reduce the need for seeds.

The average annual water fluctuation in the lake is about 3 m, and as thelake level rises following the summer rains, terrestrial vegetation is inun-dated. It has been observed that those plant fragments which drift to the shorewith the current are trapped in the flooded vegetation and sink to the bottomwhere they establish new plants. By the time the flooded vegetation rots (3-4 months), a population of Lagarosiphon has already been-e\tablished. Therapid annual growth, canopy formation at the water surface and the massivevegetative reproduction by stem fragments are competitive strategies (sensuGrime, 1979 and Kautsky, 1988), and these could account for the dominanceof Lagarosiphon over other submerged macrophyte species in the lake(Machena, 1987).

The big difference in the production values between the community takenas a whole and individual plants reflects the role other components of theecosystem play in the metabolism of the macrophyte community. Sedimentrespiration in the Lagarosiphon community can be high (C. Machena et al.,1989) and this may overshadow the production of the plants in the commu-nity. For this reason the valus of 16.4 mg DW g day could be a morerealistic growth rate of Lagarosiphon.

ACKNOWLEDGEM ENTS

This study was carried out with the support of SAREC, the Swedish Agencyfor Research Cooperation with Developing Countries. We thank Professor E.van der Maarel, Professor C. den Hartog, Dr. C. Skarpe, Dr. A-M. Jansson,Dr. L. Kautsky and Dr. J.F. Talling for their comments on the manuscript.T. Moyo helped with the diving. D. Mwaita typed the manuscript, M. van derMaarel-Versluys made editorial checks. This paper is published with the permission of the Director of National Parks and Wild Life Management,Zimbabwe.

REFERENCES

Balon. E.K. and Coche. A.G., 1974, Late Kariba: A Man-made Tropical Ecosystem in CentralAfrica. Junk, The Hague.

Carpenter, S.R. and Lodge, DM., 1986. Effects ofsubmerged macrophytes on ecosystem pro-cesses Aquat. Bot., 26: 34l-37O.

Cattneo, A. and Kalif. J., 1980. The relative contribution ofaquatie macrophytes to the pro-duction olmacrophyte beds. Limnol. Oceanogr., 25: 280-289.

Coche, A.G., 1968. Description ofthe physico-chemical aspects ofLake Kariba. an impound-ment in ZambiaRhodesia. Fish. Res. Bull. Zambia, : 200-267.

Davies. OS., 1970. Productivity of macrophytes in Marion Lake, British Columbia. J. Fish.Res. Board Can., 27: 7 1-78.

Denny, P., 1985. The structure and functioning of African euhydrophvte communities. Thefloating-leaved and submerged vegetation. In: P. Denny (Editor). The Ecology and Manage-ment of African Wetland Vegetation. Junk, The Hague, pp. 125-151.

Dybern, B.I., Ackerfors, H. and Elmgrcn, R. (Editors), 976. Recommendations on methodsfor biological studies in the Baltic Sea. Bâltic Mar. Biol., i: l-98

Grime, J.P., 1979. Plant Strategies and Vegetation Processes. Wiley, Chichester, 222 pp.Haag, R.W, 1983. Emergence of seedlings of aquatic macrophytes from lake sediments. Can. J.

Bot.,61: 148-.156.Hannan, H.J-1. and Dorns. T.C., 1970. Succession of macrophyte community in a constant tem-

perature river. Limnol. Oceanogr., 15: 442-453,Harper, J.L., 1977. Population Biology of Plants. Academic Press, London.Howard-Williams, C., 1978. Growth and production of aquatic macrophytes in a south temper-

ature saline lake. Verh. mt. Verein. Limnol., 20: 1153-1158.Howard-Williams, C. and Allanson, B.R., 1981. An integrated study on littoral and pelagic pri-

mary production in a Southern African coastal lake. Arch. Hvdrobiol., 92: 507-. 534.Hunt, R., 1978. Plant Growth Analysis: Studies in Biology, no. 96. E. Arnold, London, 67 pp.Jansson, B.O. and Wulff, F., 1977. Ecosystem analysis of a shallow sound in the northern Baltic

-. a joint study by the Askö group. Contrib. Askö Lab. Univ. Stockholm, 18: 1.160.Kautsky, L,, 1987. Life cycles of three jiopulations of Poiwnogezon pectinatus L. at different

degrees of wave exposure in the Askö Area, Northern Baltic proper. Aquat. Bot., 27: 177-186.

Kautsky, L., 1988. Life strategies of aquatic soft bottom maerophytes. Oikos. 53: 126-.135.Kemp, W.M.. Lewis, M.R. and Jones, T.W., 1986. Comparison of methods for measuring pri-

14 c. MACHE?JA ET AL.

L4GAROSIPHOV IL!CIFOLI US N LAKE KARIBA 15

mary production by the submerged macrophyte Poainogeton perfoliatus L. Limnol. Ocean-ogr.,31: 1322-1334. :

Lévêque, C., Dejoux, C. and Lauzanne, L., 1983. Lake Chad, ecology and productivity of ashallow tropical ecosystem. J.P. Carmouze, J.R. Durand and C. Lévêque (Editors). Junk,The Hague.

Lipkin, Y., Beer, S. Best, E.P.H., Kairesalo, T. and Salonen, K., 1986. Primary production ofmacrophytes: terminology, approaches and a comparison ofmethods. Aquat. Bot., 26: 129-142.

Machena, C., 1987. Zonation ofsubmerged macrophyte vegetation and its ecological interpre-tation in Lake Kariba. Vegetatio, 73: 1 1 1-1 19.

Machena, C. and Kautsky, N., 1 988. A quantitative diving survey of benthic vegetation andfauna in Lake Kariba - a tropical man-made lake. Freshwater Biol., I 9: 1-14.

Machena, C., Kautsky, N. and Lindmark, G., 1989. Metabolism and nutrient dynamics in La-garosiphon!/ic(folius in tropical Lake Kariba. In: Ecology ofthe Hydrolittoral MacrophyteCommunities in Lake Kariba, Ph. D. Thesis, University of Uppsala.

McLachlan, A.J. and McLachlan, S.M., 1971. Benthic fauna and sediments in the newly createdLake Kariba (Central Africa). Ecology, 52: 800-809.

Mitchell, D.S., 1969. The ecology of vascular hydrophytes on Lake Kariba. Hydrobiologia, 34:448-460.

Mitchell, D.S., 1970. Autecological studies of Sal vinia aurticulara Aubl. Ph.D. thesis. Univer-sity of London.

Morgan, M.D. and Kiuing, CL., t 984. Productivity and utilization of the seagrass Hciloduleirightii and its attached epiphytes. Limnol. Oceanogr., 29: 1066-1076.

Murray, L. and Wetzel, R.L., 1987. Oxygen production and consumption associated with themajor autotrophic components in two temperate sca grass communities. Mar. Ecol. Progr.Ser., 38: 23 1-239.

Pelican, J. Hudec, K. and Stastny, K., 1978. Animal populations in fish pond littorals. In: D.D.Dvkyjová and J. Kvêt (Editors), Ecological Studies 28 Pond Littoral Ecosystems. Structureand Function. Springer, pp. 74-79.

Rich, P.H., \Vetzel, R.G. and Van Thuy, N., 1971. Distribution, production and role of aquaticmacrophvtes in a southern Michigan marl lake. Freshwater Biol., 1: 3-21.

Riemer, D.N., 1984. Introduction of Freshwater Vegetation. AVI Publishing Co., Westport, CT.Sculthorpe, C.D., 1967. The Biology of Aquatic Vascular plants. E. Arnold, London, 610 pp.Solander, C.D., 1982. Production of macr6phytes in two small, sub-arctic lakes in northern

Sweden. In: 3.3. Symoens, S.S. Hooper and P. Compere (Editors), Studies of Aquatic Vas-cular Plants. Royal Botanical Society of Belgium, pp. 181-186.

Van Wijk, R.T., 1983. Life cycles and reproductive strategies of Potamogeton pectina! u.s L. inthe Netherlands and the Camargue (France). Proceedings International Symposium ofAquatic Macrophytes, Nijmegen, 18-23 September, 1983.

Vollenweider, R.A., 1974. A Manual on Methods for Measuring Primary Production in AquaticEnvironment. IBM Handbook, no. 12. Blackwell, London, 213 pp.

Westlake, D.F., 1975. Primary production of freshwater macrophytes. In: J.P. Cooper (Editor),Photosynthesis and Productivity in Different Environments. Cambridge University Press,pp. 189-206.

Weizel, R.G. and Hough, R.A., 1973. Productivity and role of aquatic macrophytes in lakes. Anassessment. Pol. Arch. 1-lydrobiol., 20: 9-19.

J. Fish BioI. (1990) 36, 701-705

Preference of Tilapia rendaii (Boulenger) for some species ofaquatic plants

P. C. CHIFAMBALake Kariba Fisheries Research Institute, Box 75, Kariba, Zimbabwe

(Received 31 August 1989, Accepted 27 November 1989)

Preference of juvenile red-breasted bream, Tilapia renda/li (Boulenger), for the macrophytesCeratophyllum detnersurn L., Lagarosiphon ilicifolius (Oberm), Najas pedinata (Magnus) andValisneria aethiopica (Frenzi) was determined. Fish were offered a mixture of the four macro-phytes in equal quantities to determine selection. Preference was shown for V. aethiopica, whilstC. dernersum was least selected. This preference order was checked by examination of stomachcontents of fish from the same population in the wild and the differences are discussed.

Key words: Ti/apia renda/li; herbivory; food preference.

I. INTRODUCTION

Vegetation selectivity of herbivorous fish is usually studied with respect to biologi-cal control of aquatic weeds. The herbivorous red-breasted bream, Ti/apia renda/li(Boulenger), has demonstrated its potential in controlling aquatic macrophytes byremoving them completely from some impoundments (Junor, 1969). However, insome situations the fish has had no such impact on aquatic macrophytes whichthrive to nuisance level. Plant preference of the fish, which has not been examinedfor specific impoundments, could be an important factor.

The basic factors determining the quantity of individual food items in the diet arerelative availability and selection of some items by the fish (Horn et al., 1982;Mwebaza-Ndawula, 1984; Prejs & Prejs, 1987). Selection is based on palatabilityand nutritive qualities (Westoby, 1974, 1977). Items with high structural content,polyphenolic compounds and calcareous material are avoided (Horn et al., 1982).Energy and caloric value and protein content are regarded as the most importantcomponents in determining nutritive value (Westoby, 1974; Mattson, 1980).Plants with either high energy and protein content or both are favoured.

The objectives of the present study were to determine selectivity of T. renda/lifed on four different macrophytes, Ceratophyllum demersum L., Lagarosiphonilicifolius Oberm, Valisneria aethiopica Frenzl and Najas pedinata (Pari)Magnus. Nitrogen content, total organic matter and water content of plants wasmeasured to determine if there was any relationship between plant selection andthese parameters. Choice under aquaria conditions was compared with selectionin the wild deduced from stomach content analysis.

IL MATERIALS AND METHODSSelection trials were carried Out in a glass aquarium of volume OE35 m3. Water was

aerated using an aerator, type RENA 301. The temperature, dissolved oxygen and pH were701

00221 112/90/050701+05 503.00/0 © 1990 The Fisheries Society of the British Isles

702 P. C. CHIFAMBA

monitored throughout the experiment. A YS model 58 dissolved oxygen meter was usedto measure both dissolved oxygen and temperature. and for pH a Hach digital pH was used.Readings were taken once a day at 13.30 hours.

Tanks were stocked with fish. length range 56-1 17 cm. collected from the wild inLake Kariba. The fish were individually weighed and the standard length measured. Theywere acclimatized for 20 days. during which they were fed with the four macrophytes.Ceratophyllum dernersum, Na/as pectinata, Lagarosiphon ilicUblius and Valisneriaaethiopica, which were to be used in the feeding trials.

Macrophytes for feeding fish were collected from the lake by SCUBA divers. Only thestems with leaves were collected, except for V. aethiopica where roots were necessary toimprove keeping quality.

The plant material was washed thoroughly under running water to remove any attachedmaterial. Macrophytes with a visible layer of epiphytes were not used. After this cleaning.V. aerhiopica was clipped to remove excess roots (some roots and stem were necessary forholding plants onto the feeding trays). As much water as possible was removed, by blottingfor V. aethiopica, and by tossing in a feeding tray until water stopped dripping for the otherplants.

Material was then weighed to the nearest gramme using a Meuler PC 8 000 digitalbalance. An equal quantity of each macrophyte (between 25 and 30 g) secured onto feedingtrays was presented to the fish. The trays ensured that macrophytes remained submerged inan upright position and also simplified retrieval of remaining feed. Feed was changed eachday at 07.00 hours.

For comparison of wild and aquarium feeding, stomach contents of fish caught all overKariba basin (the north-eastern end of Lake Kariba) were examined. The stomachs werepreserved in 10% formahn for later analysis. All fish stomach contents were analysedwithin 4 weeks of capture; to facilitate this they vere diluted with water. A dissectingmicroscope was used when necessary, otherwise materials were identified by unaided visualinspection. All plant species and food items were recorded. From these the percentageoccurrence of each food was calculated as:

Number of stomachs with food item% Occurrence = x 100

i otal number of stomachs analysedThe water content and organic matter for the different macrophytes were worked out afterdrying plants in a 1000 C oven to constant weight and incineration at 450_5000 C to con-stant weight. Protein content was calculated from the Kjeidahl nitrogen, using the factor of51 as given by Grodzinski et al. (1975) for leaf protein. All assays were performed intriplicate.

III. RESULTS

Results of the aquarium selection trial are given in Fig. 1. There was a significantdifference (P < 0O01, ANOVA) between the mean weight of each macrophyteconsumed. Fish preferred V. aethiopica, N. pedinata, L. ilicfoIius and C.den2ersum in that order.

Percentage occurrence of food items consumed in the wild is given in Fig. 2.Considering stomachs containing aquatic plants only, the plants used in the pres-ent study occurred in the following proportions: V. aet/liopica 359%; L. ilicfolius385%; N. pedinata 256%; C. dernersum 0%. There was significant difference(P < OE001, 72) between what is consumed in the wild and what was selected in theselection trial, although C. dernersum was again the least consumed.

The results of the chemical analyses are given in Table I. The percentage organicmatter, ash and nitrogen content of plants differed significantly (P<00l,ANO VA) and that of water differed at P<005 (ANOVA).

PLANT FOOD PREFERENCES OF TILAPIA RENDALLI 703

V aeth/opica422%

L. Ilicifo/lus273%

Detritusl09%

N. pedinata3180/

FIG. 1. Proportions of four species of macrophytes consumed by Ti/apia renda/li in the selection trial.

Unidentifiedmacrophyles170/ Fila me n tous

algae147 0/

C. demersum92%

V. cern/op/ca245%

L. il/cifo//ugl68%

N. pectina/aI 8 2%

Insect larvae9I 0/

FIG. 2. Percentage occurrence of food items in the stomachs of Tilapia renda//i collected from the wild.

IV. DISCUSSION

Fish showed a preference for V. aethiopica (Fig. 1) which had a significantlyhigher protein and ash content (Table I). This agrees with the expected diet selec-tion strategy, whereby food item selection by a general herbivore is geared tomaximizing intake of nutrients, particularly nitrogen, and energy (Tan, 1970;Westoby, 1974; Mattson, 1980). Tan (1970) indicated that the superiority ofHydrilla verticillata over other plants tested as food was due to its soft nature (lowfibre content) and high ash content. In the present case V. aethiopica would beregarded as superior because of its relatively high ash and protein content.

The order of preference for the other macrophytes does not relate to proteincontent. Usually selection is due to an interaction of factors, although nutrient

704 P. C. CHIFAMBA

TABLE 1. Percentage water, organic matter, ash and protein content ofmacrophytes used inthe study

Type of plant % Water % Organicmatter %Ash % Protein

(dry wt.)

content may be dominant. Other factors such as palatability, particularly withrespect to fibre content, have been shown to affect selection (Westoby, 1977; Prejs,1984). Plants with a lot of structural material and polyphenolic compounds areselected against (Horn et al., 1982). Horn el al. (1982) showed that morphologicalfeatures such as high surface to weight ratio and life span were important in dietselection. A more comprehensive chemical analysis of the plants used in the studyas well as an assessment of leaf toughness, spatial availability of leaves on plantsand other parameters measured by Prejs (1984) should be carried out in order toexplain some aspects of selection in the present study.

C. dernersum, which was the least preferred of the four macrophytes, was notobserved in the stomachs of fish from the wild; it appears that the plant is beingavoided by the fish. The proportions in which the other plants, V. aethiopica, N.pedinata and L. ilic?folius, were consumed in the wild differ significantly(P<000l, x2) from those obtained in the selection trial (Fig. 2); this differencecould be explained by relative abundance of the macrophytes in the wild.

Selection in the wild is governed not only by nutrition level but also byabundance (Horn el al., 1982; Mwebaza-Ndawula, 1984) and is expected to be acompromise between preference and selection, so as to balance the cost of ingestion(including cost of search) of a particular food item and the benefit derived from it(Prejs, 1984). In the Kariba basin, L. ilicfolius, N. pedinata and C. den2ersum arethe dominant macrophytes (Machena, 1987). Since the diet of T. renda/li in thewild is strongly related to abundance (Caulton, 1977), the higher proportions of L.ilic?folius and N. pedinata in the diet were expected. It seems there was a positiveselection in favour of V. aet/liopica, since its abundance in the wild was low.

I am indebted to the management of the Prawn Farm who provided fish for theexperiment. I thank all members of the Kariba Underwater Research Society, especiallyRoni McDonald and Helen Head (my 'buddies') and members of the Army who bravedcrocodiles in Kariba collecting macrophytes; also Mr N. Moyo (my supervisor), Mr E.Dhlomo and Mr H. Masundire of the University of Zimbabwe, my colleagues at LakeKariba Fisheries Research Institute, and Miss Deliah Mwaita for the typing.

Lagarosiphon ilicfoliusLeaves 917 7.4 09 86Whole 918 7.4 07 l06Cera tophyllum dernersumLeaves 953 4l l30Whole 932 62 OE6 86Valisneria aethiopicaLeaves 93.7 5.4 08 l51Najas pectinataLeaves 948 4.5 125Whole 945 52 OE4 98

PLANT FOOD PREFERENCES OF TILAPIA RENDALLI 705

References

Caulton, M. S. (1977). A quantitative assessment of the daily ingestion of Panicurn repensby Tilapia rendalli Boulenger (Cichlidae) in Lake Kariba. Trans. Rhod. Scient. Ass.56, 38-42.

Grodzinski, W., Klekowski, R. Z. & Duncan, A. (ed.) (1975). Methods for EcologicalBioenergetics. IBP Handbook No. 24. Oxford: Blackwell Publications.

Horn, M. H., Murray, S. N. & Edwards, T. W. (1982). Dietary selectivity in the field andfood preferences in the laboratory for two herbivorous fishes (Cebidichihys violaceusand Xiphister mucosus) from a temperate intertidal zone. Mar. Biol. 67, 237-246.

Junor, F. J. R. (1969). Tilapia melanopleura Dum. in artificial lakes and dams in Rhodesiawith special references to its undesirable effects. Rhod. J. Agric. Res. 7, 6 1-69.

Machena, C. (1987). Zonation of submerged macrophyte vegetation in Lake Kariba,Zimbabwe and its ecological interpretation. Vegetatio 73, 111-119.

Mattson, W. J. Jr (1980). Herbivory in relation to plant nitrogen content. Ann. Rev. Ecol.Syst. 11, 119-161.

Mwebaza-Mdawula, L. (1984). Food and feeding habits of Clarias inossambicus from fourareas in the Lake Victoria basin, East Africa. Env. Biol. Fishes 10(1/2), 69-76.

Prejs, A. (1984). Herbivory of temperate freshwater fishes and its consequences. Env. Biol.Fishes 10 (4), 281-296.

Prejs, S. & Prejs, K. (1987). Feeding of tropical freshwater fishes: seasonality in resourceavailability and resource use. Oecologia (Berl.) 71, 397 404.

Tan, X. T. (1970). Composition and nutritive value, of some grasses, plants and aquaticweeds tested as diets. J. Fish Biol. 2, 253-257.

Westoby, M. (1974). An analysis of diet selection by generalist herbivores. Am. Nat. 108(96), 290-304.

Westoby, M. (1977). What are the biological bases of varied diets. Am. Nat. 112, 627-63 1.