EAST AFRICAN COMMON SERVICES ORCANIZATION

EAST AFRICAN FRESHWATER ,FISHERY

RESEARCH ORGANIZATION

ANNUAL REPORT

1961

) I

1962 PUBLISHED J3Y AUTllOltrry

'pRINTED BY DIA1IlOND PRESS LTD., NAJRom.

Price Shs. 4/.

EAST AFRICAN FRESHWATER FISHERY

RESEARCH ORGANIZATION

ANNUAL REPORT

1961

P.O. BOX 343,

J I NJA,

UCANDA

I

CONTENTS

Staff

Launches

Laboratory & Housing

Library

Meetings and Visits

Visiting 'Workers

Scientific Work of the Organization

. Fisheries of Lake Victoria

Commercial records

Population analysis

Rational exploitation

Other species of Tilapia

Small mesh gill net fisheries

Bagrus docmac Investigation

Lates Investigation

Breeding studies on T. zillii & T. nigra

The culture of Tilapia in ponds

Migration of River Fish

Hydrology and Algology

Gear Research

Buoy releases

Vertical gill nets

Publications

Page

1

2

3

3

3

4

4

4

5

5

7

8

8

10

11

12

13

13

14

14

IS

16

..a-~'

II

11

Appendix A

Preliminary Notes on the Relationship between

Feeding and Growth Rate in the Siluroid Fish

Bagrus docmac (Forsk)

H. Y. Elder ......

";.1

Appendix B

Nile Perch Investigation

E. L. Hamblyn ......

Appendix C

Breeding Studies on Tilapia zillii and Tilapia nigra

C. C. Cridland

Appendix D

Pond Culture Studies on Tilapia nigra

V. D. van Someren

Appendix E

River Fish Migration

V. D. van Someren

Appendix P

Report 011 Limllological \-\fork

J. F. Tailing ......

Appendix G

Buoy Release Trials

J. D. Roberts ......

Appendix H

A Note on Lake Rudolf

E. L. Hamblyn ......

19

23

29

33

36

40

43

...... 46

f

EAST AFRICAN FRESHWATER FISHERY RESEARCH

ORCANIZATION

ANNUAL REPOR~ 1~1

STAFF

Director:

V. D. VAN SOMER EN, M.B.E., B.Se., Ph.D.

Research Officers:

c. C. CRIDLAND, Dip Ferm. H. Y. ELDER, n.se.

E. L. HAMBLYN, B.Se. One vacant.

Experimental Fisher·ies Officer:

J. D. ROBERTS

Secretary (Part-time)

MRS. E. D. KEMP

Senior Field 0 [ficer:

R. 1. M. BAXTER

A ccounting Assistant:

L. F. D'COSTA

Foreman Artizan:

PRAGJI DOSSA MESV ANIA

GENF.RAL

The above shows the staff position at the close of the year. Mr. D. J. Garrod proceeded on U.K. leave in May, and resigned in November in order to take up an appointment with the Ministry of Agriculture and Fisheries Laboratory at Lowestoft 10 England, where he will carryon population dynamics studies in marine fisheries. His vacancy is not yet filled.

Mr. J. D. Roberts was promoted to a new C.1 Scale post of Experi­mental Fisheries Officer, in July 1961, and it was possible to upgrade Mr. Pragji Mesvania to the grade of Foreman Artizan also in July. .

The vacant post of Senior Field Officer (Laboratory, Stores and Housing) was filled by the appointment of Mr. R. 1. M. Baxter, formerly of the Uganda Tsetse Control Department, in September. He has already made a considerable impact on a backlog of store checking and maintenance work which is essential to the smooth running of the Organization.

2

This has been the last year during which the Organization has functioned under the aegis of the East 1\ frica High Commission as one 0 f it's nOll-sel f­contained services.. On Dcceillber 9th 1961, thc E.A. High Commission waS succeeded by the E. L\. Common Serv ices Or;~an~zation. A COll1mis::;ion Oil

the future of thelZeseilrch Services was set up in September, 196 I, under the chairrnanship of Prof. A. C. I:r;\zer, and its Tecomll1cndations, which at the time of writing are stiil lIi~del- consideration by the Covernmcnts. will effect a fundamental change in the administration and financing of this and other research organizatio!ls formerly under the E.A.H.C. They have recommended continuance of r,"search 011 an East African basis, uncler all E.A. Scientific Natura! l-":cs0t!rces Council with its own executive, ultimately controlled hy a l'vfinisterial Committee of the Common Services Organization.

Mr. F. L. Hamhlyn has been on detachment at Butiaba on Lake Albert for much of the year on the Nile perch investigation. hut had to withdraw from tllere in November and evacuate all the equipment because of the tremendous rise in level of Lake Alben which threatened to swamp the store and lahoratol-y hut he had lJeen u~ing. He is now stationed at J]eadqllartcrs in Jinja once more, cng;'tged on writing up his results.

No. 1 launch was out of COlllllJis;.;ion for over a month in 1unc, as an unfortunate accident strippeu hath propellers on a sunken reef' off Dagusi Island in late May. Opportunity was howeveI" f~d(en to order and refit two new propellers of slightly different pitch ordered by air from U.K. which are more dhcienl!y suited to the two Perkins S,6M diesel engines with which the launch is powered, and the illlpro\,(~lTlellt in her performance is marked. These new Perkins engines WlTe fitted iu November 1~59, and engine revolu­tions at full throttle were 1650 r.p.m., giving a speed of approximately 151 knots. The engines are however derated at this altitude, and correspondence with Messrs. Thornycrofts led to the repJacenl'~nt propellers having a pitch of 14" instead of the old 161:". The I1tting of these resulted in maximum engine revolutions of ZOOt) r.p.m .. giving a top speed of !7-17~ knots. Stop­ping the revoilltiuns dOWll to 1850 r.p.nl. now !~ives a crui~;ing speed identical to previolis top speed, a simil;lr fuel consumption and engines running more smoothly \'{elI within their limits. The illll11cdiate 'wild-its arc therefore extra po-wer ,mel speed availahle, and tiiC lung tcnn benc!its will be the engines cruising lightly lO;tded, giving longer Ii fc ;mu Jess wear and tear. This laullch ran sonIC 4.600 nliles on duty duri;.; the year in 57 trips and 2::;7 running hours and has othen-vise gil'Cn trouble free se.rvice.

No. 2 launch has gi I.'en CO',lLinu()us trouhle free service throughout the year. apart fr0111 the usual m;\inti:'il;111Ct.: ';jirpillg for repainting, running some 20 hours en dUly, nlainly (;11 Joel! tril's of very short duration. l"or a ....ariety of reasons it was not transferred to I,ake Alhert for the Nile perch investi­gatiun, remaining at Jinja.

No order has yet heen placed for the much larger neVi vessel for long range wurk. It is impossible to obtain adequate clear after-deck space for -working, I-"ith 'tween--decl; living accommodation, Ivithin the length and heam of hui! which can be transpoTted as a v:hole unit by rail. Vor this reason, the A frican Marine and (;ener;l! Engineering Co. Ltd. at Mumhasa have hee.n approached to estimate for :l welded steel hull craft of 55' LOA to specifications laid down by E.:\.I:.F,P'.O., to be assembled in sections on the Lake.

The preliminary estimates II-ere in excess of the Sllm of £14,000 available, and did not include ~,everal of the items cOIl~idered essential, such as v.p. propeller etc. The Colonial Office ha Ie now ulldertaken to sec if the whole

3

capital cost cannot be met, for a complete craft, by a 100% grant, so that the estimates are now being re-cast for consideration. This however will still leave the recurrent costs of additional crew, running, maintenance etc. to be met territorially to the extent of perhaps £4,000 p.a., and the position is still uncertain at the time of writing.

The Bedford Pick-up truck ran some 5,500 miles on duty during the year, its total mileage being 26,600. New front springs were fitted in November, and while the mechanical' condition is sound, the body work is showing signs of inevitable wear and tear. Though capacious, it is a far from satisfactory vehicle for safari work, for its road-holding capabilities in wet weather are very poor. It should he replaced with a 4-wheel drive vehicle with a trailer before long if finance can be made available.

During the year an additional grant was obtained from CD. & VI/. sources to assist in maintenance of the eight staff houses, laboratory building, work­shop and aquarium, together with the erection of a small museum to house specimen collections and display work in progress. A further survey of the structure of the buildings by the High Commission Property and Maintenance

. Officer has however indicated that the fabric of the Laboratory, Aquarium and houses is in a much worse state with subsidence and termite damage than was at first realised, and there is no doubt that the wear and tear of a tropical environment for 15 years, coupled with bad initial construction, is now going to be a much more expensive business to rectify than at first estimated.

I t has been possible however to construct several useful additions to the working facilities during the year, notably a waterproof cabinet for the Mettler Balance in the fish room and a properly paved courtyard on which the numerous outside experimental aquaria are now raised on Dexion stands, making for much cleaner and more erficient working. Repainting of Labora­tory and houses has been completed throughout, resulting in a fresher and more cheerful appearance.

The work in the Lihr.ary by Mr. Elder and Mrs. Kemp has resulted in the establishment of first-class working and reference facilities, but much still remains to be done. The rising cost of periodical subscriptions, volume binding, additional shelving etc. all exceed the Library expenditure which has heen pegged at £200 for several years, and application has been made for capital to other sources to improve the financial position; no sum is yet availahle. The Library facilities are immensely important, since this is now practically the only comprehensive hydrohio1ogical Library, south of the Sahara, still in working condition.

The Director attended the 15th and 16th meetings of the Agricultural and l'isheries Research Council in Nairobi in January and September and the Inland Fisheries Research Coordinating Committee meeting in Nairobi in October.

In May, the Director also spent a fortnight in Ethiopia at the invitation of the President of the University College of Addis Ahaba to advise on the siting of a new (primarily) hydrobiological station, and was able to see a great deal of the interesting Ethiopian Rift Valley lakes and Blue Nile Gorge, and his thanks are gratefully rendered to the Ethiopian authorities for making this possible.

He also attended the annual Joint Fisheries Research Organization Advisory Committee meeting at Samfya in Northern Rhodesia in October, and spent a further ten days in the Federation seeing someth.ing of the

.;..

4

flshery research in progress on Lakes Nyasa, Tanganyika and Bangweulu. The assistance given by ].F.RO. and territorial st~dT in making this journey possible is sincerely acknowledged.

Various members of the stafr were able to pay visits to other parts of Lake Victoria, notably the MViam:a area, also to Lake Albert, Lake Naivasha and the Inland Fishery Research Station in Kenya where, by courtesy of the Chief Fisheries Othcer, Kenya, a good deal of useful experimental \vork on Nile perch, Tilapia culture anc! river trapping is still in progress under our supervision. We are much indehted to all the officers of the Uganda Fisheries Department, the Kenya Fishery Division and the Tanganyika Department of Ag-riculture for facilities and help readily granted at all these places.

Dr. and Mrs. J. F. Tailing completed their year's visiting work and returned to the F.B ..A. Laboratories in England in early September. Their contribution to an understanding of the hydrology and al'gology of Lake Victoria has been outstanding, and we bade farewell to such pleasant colleagues with much regret. They were also able to accomplish an astonishing amount of work on Lake Albert and the Lakes of Western Uganda, and analyse samples from Ethiopia, Tanganyika, Kenya and the Federation, in addition.

There was the usual number of distinguished visitors to the Laboratory during the year, and we were especially pleased to be able to assist Prof. H. Heller of the Uni versity of Bristol for a few days with the collection of pituitaries from Protopterus and Polypterus on which he is carrying out chromatographic studies.

SCIENTIFIC \VORK OF TilE ORGANIZATION

In this report the same practice is being followed as for last year, each Research Officer giving a 1110re detailed account of his results, in the Appen­dices, than has previously been the case.

There is hO\.ve\'cr no separate appendix on the fisheries of Lake Victoria as such. the results obtained by Mr. Garrod bdore he left being summarised in this introductory section only.

FlSI1ERIES OF LAKE VIcroRIA

At the time of writing, it has not heen possible to complete the normal six-monthly analyoes of commercial statistics, even for the first part of 1961 under revic\v. The Ug;mda records are complete only up to March, though the full year's results ,\re to hand from Kenya and Tanganyika recording stations.

Up to the end of 1960, however, the recOl-ds showed no substantial cha1we which would necessitate revision of advice put forward earlier on the need''to reirnpose control, in order to stabilise the lisheries and obtain an optimum :;ust;tinable yield, instead of a yield which fluctuates and causes fishermen to move from inshore to offshore and vice versa.

In Uganda waters the general conclusions to be drawn from analyses of landings is that in some areas the fish stocks may be recovering as a result of the reduction in fishing effort when fishing became uneconomical three years ago. Nevertheless the ligures show that the yield is not yet as g-reat as when 5" nets were in use; and if 4~" nets continue in use, there may he an incn:ase in fishing effort in this area which wi.!l prevent a full recovery.

5

In spite of a low catch per net when 5" nets only were in use prior to 1957, it is now known that the fish stocks were actually under-exploited. However the subsequent use of smaller meshes has led to overfishing, and in the areas recorded the yield will never be substantially greater than when 5" nets were in use, but the basic stocks of fish have been seriously reduced. There is no doubt that a much better yield would have been obtained by increased fishing with 5" nets, instead of reducing mesh size in an effort to maintain yield. It is known that T. variabilis is caught in greater numbers in 4·r mesh nets, and at Bukakata for example where the numerical yield is equal to that obtained in 1955, prior to derestriction, this is because of the increased catches of this species which have offset the decline in T. esculenta catches; but whereas in 1955 the total catch was valued at Shs. 31,659/-, in 1960 approximately the same catch was valued at Shs. 18,398/- only, because the weight and value of T. variabilis from 4i-" nets is less than that of T. esculenta. This does not appear an economic advance in any way, especially in view of the damage done to the basic fish stocks which results from such use 0 f small mesh nets.

In Kenya waters, there was a particularly strong year class of T. esculenta entering the fishery, as had been forecast in the previous Annual Report, which enabled a temporary improvement to take place in some areas of the Gul f fisheries. Outside the Gulf, the 4i-" net records now show that, as had been predicted by E.A.F.F.R.O., there are no substantial reserves of other species which would offset the damage to 1'. esculenta fisheries which is inevitable when 4i-" net fishing is intensive.

In Tanganyika waters the catch per net is still declining in all areas, but here the retail price of fish is lower than elsewhere and the fishermen tend to move before the stocks become as seriously depleted as in other parts of the Lake.

During- the year, it has been noted that offshore fisheries have increased, but neither the territorial departments, nor E.A.F.F.RO., have any statistics to show if the development of these has been adequate to offset losses from declining catches inshore, and it is vital to obtain further data on this. It can be nothing short of disastrous however to fish new areas with 4r' nets only, and it must be realised that a mesh which returns the maximum catch of fish is by no means the optimum net to use, if the yield is to be sustained. This cannot be emphasised too strongly in view of recent developments to open up further areas to fishing.

As a corollary to this general analysis of commercial returns, Mr. Garrod also during the year completed three years intensive study of a "unit" fishery in the N. Buvuma area, which is known to be reasonably self-contained and unaffected by migration of fish.

From the examination of several thousand fish at weekly intervals at Waigalla landing, it has been possible to calculate the basic parameters of fish groyvth, fishing effort, natural and fishing mortality, fecundity and net selectivity, and combine them into a mathematical model which not only shows the existing state of the fishery, but enables forecasts to be made of the effect of varying the effort and mt:sh. This has proved extremely accurate when checked against known landings, and for the first time for any tropical freshwater fishery, has enabled a rational method of exploitation to be proposed in order to obtain an optimal sustainable yield from such a fishery.

6

It is thus known for example that in this N. Buvluna area of 29,235 acres, there are at present being fished 26,000 x 4f' nets per month, or 0.9 nets per acre per month, or 877 nets per night. This is resulting in a catch of 18.2 ,lbs. of fish per acre per year, of a total value of £27,000.

It is also evident however that this effort is probably decreasing the breeding stock at a rate which recruitment is unable to equalise, thus resulting in a state bordering on biological overfishing. In order' to preserve a sufficient recruitment rate, the model shows that if, for administrati ve con­venience, the exclusive use of 4i" nets is permitted to continue, the number of these in use must be reduced to 21,900 per month, or 0.75 nets per acre per month or 730 nets per night. This v"ill result in a reduction of the catch to 15.2 Ibs. of fish per acre per year of a,yalue of £25,000.

The model however also shows that the maximum yield from this area could be increased to 20 lbs. of fish per acre, value £30,000, from a standing stock of 1,100 tons of fish, and this could be sustained without damaging the breeding stock by fishing a mixture of 25% 4:}" nets and 75% 5" nets. This involves reducing the number of 4t" nets to 11 ,lOOper month, or 390 nets per night, and increasing the number Of 5" nets to 35,082 per month or 1,170 per night.·

It should however also be realised that if this rational technique is adopted, there would, firstly, be a temporary decline in catches while stocks adjusted themselves to the new equilibrium; and secondly, that while the overall total yield would be greater with the increased effort, there would also, as a corollary, be a lower catch per net. Thus though the total income of fishermen as a group would be greater, any individual fisherman's profit would be less. An increase in the price of fish to the fishermen would offset the latter and also cover increased capital cost of the additional nets required, while the formation of cooperatives could spread the overall greater profit more evenly.

These figures apply only to the Tilapia esculenta stocks in this area, but since there is no reason to suppose that the biology of this species varies markedly from place to, place on the Lake, these recommendations may also be applicable to other \,Init fishery areas of Lake Victoria as "veil. There is a physical limitation on the number of nets which can be fIshed from each type of canoe in use, hence it is also possible to control the effort to a great extent by a canoe li<:ensing system,: which in effect will impose a limitation On the number of nets which can he fIshed from anyone landing. area.

The records since de-restriction have sho\vn in fact that the major damage to the stocks occurred as a re:iu'lt of increase in effort with small mesh nets. These smaller mesh nets incv,itably caught· more fish than the 5" nets, since they were ,exploiting a hitherto untouched source of smaller size groups 0 f fish; and because more fish could be landed, more fishermen were induced to fish. Thefishery in Lake Victoria is unique in the fact that fish are sold by numbers ratber than by weight; arjd hence the usual economic brakes did not apply when these greater numbers of smaller fish were landed, and no price check resulted from the increase in numbers of fishermen. Thus, at least effort control must at first be reimposed, and this. must then be coupled with a return to larger meshes and a revision of marketing practice. Only thus could the fishery again be stabilised at sustainable maximum productivity.

I

7

The use however of any nets smaller in mesh than 41", in any proportion whatever, would be disastrous in its destruction of breeding stock. For a short period at the end of 1960,4" net catches were recorded in Kenya waters, and these showed catches of up to 10 fish per net. Coupled with seining which also takes place in these waters, the destruction of small breeding fish by such methods is far more than the stocks will ever stand.

It should be noted that in a fishery for such mouth-brooding Tilapia, one of the major curbs to the successful continuance of unrestricted fishing is the limited density dependent survival ability of restricted brood sizes. A mouth-brooder can only incubate a few hundred eggs and fry in the female's mouth, and since T. esculenla very rarely grow above 35 cm, there is obviously a low upper limit to the number of fry which can survive at each spawning­an upper limit which is far less than in the random pelagic or demersal spawning fish which form the basis of great marine fisheries. The efficiency of the mouth-brooding habit itsel f imposes a limitation on the number of fry which can survive in anyone stock of fish, in spite of a higher spawning frequency, and predation by man can reach levels far above the biological capabilities for recruitment of such fish.

In this connection, the data analysed by Garrod have shown that any question of maintenance of even this one small area by artificial stocking is economically impractical; the mortality calculations show that over 300 million fry a month would be required to mailltain the stock at its present level of exploitation, and the pond area required for this is prohibitively large.

OTHER SPECIES OF TILAPIA

It has not yet been possible fully to analyse the progress of the other species of Tilapia introduc,ed in the last few years. T. nilotica are still a rarity; T. zillii are forming a small but significant percentage of inshores water catches only. In the N. Buvuma area, catches of these valued £752 in the early months of 1961, but there is no distinct upward trend in the catches of this species in this area sufficient to offset the loss of T. esculenta. It is disappointing to report moreover that this species is not showing the promise of early introductions as far as size attained is concerned. The most suitable gear for cropping this species appears now to be 4" nets, and the use of this size mesh should be out of the question.

The on'ly introduced Tilapia for which sufficient recapture data are available for calculation of growth rate is T. leucosticta. Garrod has analysed the data available by the technique developed by Gulland and Holt, and has shown that there is a sexual dimorphism in growth, males reaching a maximum size of 30.5 cms and females 28.0 cms. These again are lengths more suitable for exploitation by gear less than 5" mesh.

The results from Dr. Fryer's study of T. variabilis are now available. These have shown that this species again is most suitably exploited by 4t" nets, is extremely slow growing, and probably would not stand heavy exploi­tation. It is moreover a fish of less value than T. esculenta because of its inferior keeping qualities. This study brought out most clearly the astonish­ing homing ability of this species to its breeding beaches. Some of Dr. Fryer's marked fish are still being returned, over two years after first marking, and almost always from the area in which they were tagged as breeding fish. This aspect of their behaviour will be the subject of special study in 1962 by an American worker from Prof. Hasler's laboratory in \tVisconsin, where fish orientation and homing studies are a particular subject of investigation.

8

SMALL-MESH GILL NET FISHEHU:S

In Tanganyika waters in partjcular, small-meshed gill net fisheries are more widespread than elsewhere, but the records for Synodontis fisheries show indecisive trends hecause records were organised properly only in 1958; but although catchcs are good at some landings, the catch per net has fallen and the usual disturbing tendency to use smaller mesh nets is evident as fish become scarcer.

However, where sufficient data are available, the most serious trend discernible is the decline of the important Labeo victorianus fishery, which in some areas has ceased altogether as being uneconomic even for African fishermen. Aggregated data for all Labeo fisheries show a decline in catch per net from a peak of 7.3 in 1954 to 0.9 in 1960, and the Kagera River area in particular has shown a very marked decline. _ It would be difficult to refute the suggestion that this is again due to uncontrolled overfishing, especially in the Kagera area where the fish are caught on their upstream spawning migration by the peculiarly deadly technique of floating gill nets down the river. Biological data on growth, fecundity, survival etc. of this species are however too scanty to enable any estimate to be made of true productivity and yield obtainable, but the present facts of species density- as indicated by catch per net-are sufficient to cause concern as to its future as a major component of Victoria fisheries.

BAGRUS INVESTIGATION -

Bagrus docmac coninues to be the object of a fishery of great importance on the Lake. Mr. Elder has continued his studies on this species throughout the year, emphasis being on its role as one of the major predators on the H aplochYomis stocks in Victoria, which indeed form its food almost exclu­sively. H aplochromis, which have been described as "useless" fish because of their small size, are thus nevertheless the indirect basis of a very large and valuable fishery and are very far from useless and unexploited.

The main work of collection of material in the field has now been com­pleted, although it is hoped to continue the monthly sampling of commercial catches until June 1962 when this officer is due to go on home leave. The analysis of collected material for age, growth, fecundity, and the observations on ecology, should he complete by June, 1962.

Commercially, Bagrus is the most important "non-cichlid" fish in Lake Victoria, forming approximately one sixth of the total yearly yield from the Lake. This is a remarkable fact, considering that Bagrus is a piscivorous predator. Since it depends largely on the H aplochromis, it was thought worthwhile to investigate some of the relationships between predator and prey, in order to establish the impact of predation of Bagrus on these H aplochromis stocks.

Experiments were therefore started in April to study the controlled growth of Bagrus under aquarium conditions. A technique was evolved for force-feeding the fish under water, the fish being held in a net at the surface of the water. Initially, macerated Hapiochromis flesh was introduced into the stomach through a wide-mouthed syringe, but this was abandoned hecause of thc ease with which the fragments were regurgitated, and the difficulty of collecting the regurgitated material. Subsequently, therefore, the fish were fed with whole, live Haplochromis, introduced into the stomach with scissor forceps. This latter method has the advantages that the prey fish is not easily regurgitated because of the many hackward pointing spines, and that the food is in the form naturally taken by Bagrus. Such forcc­

9

feeding has advantages over allowing the Bagrus to capture their own prey, in that the prey fish can be weighed accurately to 0.1 gm immediately before feeding, and that feeding I-ate can be varied at will. Bagrus used in these experiments were anaesthetised weekly in order to weight and measure them.

Present evidence indicates that Bagrus locates its prey probably by chemotactile perception, and that it feeds nocturnally while the prey fish are asleep. In daylight hours it remains inactive and hides under objects away from the light. Accordingly, the experimental fish were fed after dark, and were provided with cover during the day.

In the first part 0 f these experiments, fish over a wide size range were kept for periods ranging from three to seven months, and were fed at rates just sufiicient to maintain them at a constant weight. This showed that the smaller fish require relatively more food than the larger. Thus at the extreme sizes of Bagrus investigated, it was found that a fish of 65 gm total weight required 90 milligrams of prey per gram body weight per week, while one of 940 gm total weight required only 25 milligrams.

The plot of total body weight against the milligrams of prey required per week per gram body weight of the predator is a function of the metabolic rate, and the curve so obtained, over the range investigated, is in form similar to, and probably related to, the asymptotic growth curve. These findings are of fundamental importance in consideration of predatorjprey relation­ships, since clearly the age structure of a given predator population has an important bearing on the predation pressure which that population exerts.

In the second part of the experiments, Bagrus were kept in conditions ranging from complete starvation to feeding at rates probably in excess of anything achieved under natural conditions. These have shown that for any given fish the metabolic rate varies directly with the feeding rate, and seems to he independent of the condition factor of the fish. There is some indica­tion that the metabolic I<t{e of a starved fish decreases logarithmically, but consistent results have not yet been obtained.

A third part of the experiments on the growth of Bagrus under different temperature conditions has just been completed, while experiments on the growth of fish kept artificially active, and more direct estimates of metabolic rate by rate of oxygen uptake, are still in progress.

These can he expected to establish the parameters within which the normal growth and feeding rate of Bagrus in Lake Victoria lie, and hence 8. rational estimate of its role as an existing predator on the population of Haplochromis already in the Lake. It will be recalled that suggestions have a1ready been made by Uganda authorities that these H aplochromis populations might form the food basis for an exotic predator, the Nile perch, Lates, whose introduction has been proposed. This Organization has always advocated a policy of caution with regard to this, and a thorough investiga­tion of all aspects before any deliberate steps are taken. This study of Bagrus, an important predator already in the Lake, is one aspect of such an investigation, since it appears quite possible that the abundance of Bagr~ts itself. an important food fish for man, may in fact be limited by the stocks of Haf)lochromis, in spite of the apparent density of the latter. Nor must it he forgotten that there are several other predatory fish in the Lake already, such as Clarias mossambicus-again an important food fish for man-and several of the Haplochromis species themselves. In addition to these facts. it is probable that the present shoals of Haplochromis playa large part in the cycle 0 f nutrient material in the Lake, a feature 0 f ma jor importance where the prime crop is composed of herbivores such as the Tilapia species.

10

LA'n:s INVESTIGATIO),"

Hovvever, the other aspect of such an investigation is equally clearly a study of Latcs as it occurs in its natural environment elsewhere and accumu­lation of as much data as possible 011 its biology.

This investigation is being carried out by Mr. E. L. Hamblyn, who was able to establish a temporary labcratory early in the year at Butiaha on Lake Albert, in which Nile perch are the object of a large commercial fishery. The accomrnodation had to be evacuated in October owing to an exceptionally high rise in the level of Lake Albert which flooded the sand spit on which the stores and laboratory were erected. This officer had therefore to return to Headquarters at Jinja and continue with analysis of data already collected, and carry out further laboratory experiments on perch kept in controllcd conditions.

Nevertheless the field work has established a great deal of hitherto unknown information on the breeding habits of perch, and the biology of the early stages. Lates appear to be demersal spawners, fry and eggs (the latter almost certainly being I,ales eggs) being recorded near the bottom of the vYhite Nile between Lake Rubi and Rhino Camp in April in flowing water. In the main Lake Albert, very small fry have been located near the bottom in water not more thail fifteen feet deep in August, but trapping of fry and fingerlings and analysis of length frequencies indicate that breeding must occur in the Lake throughout many months of the year if not all.

Morphological examination of many more adult fish has clarified the various stages of ovary development, though "running" females have not been recorded; the appearance of comparable stages of the ovary varies £lccording to the age of the fish.

Lates of all sizes are present in the Butiaba Harbour area for most if not all of the year, and it seems possible that in this lacustrine environment, the pattern of breeding seasons is probably different from that in the riverine environment. On the other hand, fry and small fingerlings have also been recorded in deep water up to 130' and it is possible that these may have been carried to this deep area by currents from the shallow water in which they were spawned.

The presence of fish neveloping in hoth deep and shallow water has a bearing on the differentiation of Lates into deep and shallow water sub­species, but it is now helieved that in fact such subspecies are not valid on morphological and ecological grounds, though physiological differences may occur which in the course of time may lead to definite speciation.

Fish up to 40 cm show more dependence on Caridina prawns as food than Vias at f'irst supposed, and Lates is only primarily piscivorous when more than 10 Ihs. in weight; but whether tales can exist on prawns alone for a prolonged time is not known. The prey of Lates is now known to comprise manv other kinds of {-ish as a result of examination of more stomachs, and this 'fish will in fact take allY animal which conforms with the prey behaviour pattern described in last year's report.

Experiments are still in prog-ress in ponds at Sagana to determine the impact of predation of Lates of known size on known populations of Tilapia, and figures have heen obtained for such a relationship. It does not appear that in fact Laics has a marked effect on growth of surviving Tilapia, but in trying to -establi sh such a predator/prey balance there are a great many natural variables to be considered which are of the utmost importance, and

11

it cannot be said that the mere introduction of Lates into an overstocked pond will have the desired effect of increasing the yield of large size Tilapia. The problem is exceedingly complex and cannot be solved so simply. The use of an improved angling technique allied to "trotting" using Alestes as a live bait is briefly described.

BREEDING STUDIES ON Tilapia zillii ANlJ Tilapia nigra

These are the two species of Tilapia which are widely used in East Africa for pond culture work. T. zillii is a "guarder", or substratum spawner, which lays several thousand eggs at a brood, the parent fish standing guard over the eggs, which adhere in strings to a hard substrate, and over the shoals of fry which hatch from them. The adult fish are very largely feeders on macrophytic vegetation, and their ability to crop down many types of submerged and emergent pond vegetation is well known; it is for this reason they have been introduced into Lake Victoria in the hope that they will clear much of the fringing swamp vegetation where established.

T. nigra is a mouth-brooder, the few hundred eggs being taken into the female's mouth as soon as laid, and incubated there until hatched; the alevins or fry remain in the female's mouth until about the 17th day after laying, and the fry are then released periodically to feed themselves, being taken back into the female's mouth when disturbed, and at night time. This species is primarily a bottom grazer, and shows a marked sexual dimorphism in growth which renders it a very suitable subject for monosex culture, the males growing much larger than the females and growing markedly well in sexual isolation even a fter maturity.

As was described in last year's Annual Report, the breeding behaviour of T. zillii has been studied by Mr. Elder; and brood size, frequency of breeding. survival and growth of fry, by Mr. Cridland. The successful establishment of breeding pairs of T. nigra in aquaria at Jinja has now enabled Mr. Cridland to amass a great deal of data on the spawning behaviour, fecundity and frequency of breeding of this species as well, all of which has hitherto been completely unknown.

Successful spawning of T. nigra requires many hours of patient observa­tion and a careful technique which is described by this officer in his attached report. The male constructs a conically depressed nest in the sand, which. when a ripe female is introduced into the same compartment. is of considerable valence for her. After varying periods of time, and a certain amount of stylised courtship behaviour by both fish, the female lays her eggs in a steady stream at the bottom of the nest, this action occupying only a few seconds. The female immediately picks up the eggs in her mouth, at the same time as the male releases his milt, and afterwards the female, with eggs in her mouth, can be seen sucking in milt from the vicinity of the male's genital papilla; there seems little doubt that in this species, fertilisation is "internal" in the buccal cavity of the female. The eggs are then "chewed" bv the female, who immediately develops a ventral buccal pouch. •

The breeding colouration of both male and female is striking and is never seen unless a pair is observed actually spawning. All these various aspects will be described in full elsewhere.

The female is also capable of making a nest on her own, laying eggs in it and picking them up by herself and "incubating" them even if no male is present and they are unfertilised.

12

Survival and development a f eggs has been followed either in an artificial "mouth" in the laboratory, eggs being forcibly removed from the female, or else they have been allowed to develop naturally in her mouth.

Hatching of eggs at an average temperature of 23.2°C takes 4.3 days; the yolk sac of the alevins is absorbed after 9.8 days, and fry commence feeding before the yolk sac is fully absorbed. Fed on nymphal Artemia shrimps, they reach a size of 17.0 mm in 30 days, but even after this period the sexual dimorphism in growth is not yet apparent.

The number of eggs laid is clearly a function of the size of the female, though the relationship is not directly proportional. The smallest female which has laid fertile eggs is only 7.5 cm.

From 15 pairs, 48 broods have been recorded over a period of six months and 25 days. The shortest interval between broods of anyone pair has been 11 days, the longest 65 day~. The frequency of breeding shows a relationship to periods of maximum water temperature, and sunlight appears to exert a marked stimulating effect; dull, cold or rainy weather inhibiting breeding.

\iVhen the female is allowed to retain fertile eggs in her mouth, the interval between first picking up the eggs and first exit of fry from her mouth is astonishingly constant at 16-17 days. \Vhen the female is allowed to retain her eggs, the intervals between broods is only slightly lengthened. The female will retake free-swimming fry into her mouth at intervals, for a period of only seven days after first exit.

As soon as the female has picked up eggs, and they are fertilised, the male turns on her after 10 minutes, and drives her away, so that in the confined space of an aquarium, the two fish must be separated as soon as spawning is complete otherwise the female may be severely damaged by the male.

The survival of eggs artiflcially hatched in the laboratory has varied from 13.5% to 100%, with an average of 60.7%. The reason for this variability is not known. Survival of fry 30 days after hatching has varied from 34.5% to 96.2%, with an average of 67.3%.

All -thi~ breeding data now raises a great many questions of interest, mainly on the aspects of physiolcigiC<1.1 control of spawning, and the effect of the various stages of pre-and post-sp:tvvning b::haviour, on the stimu!<ltion or inhibition of final ovulation. To be ahle deliberately to control the spawning periodicity of any Tilapia, in order to reduce the frequency of breeding, would be a r,l;,jor break-through in the use of these fish for cultural purposes. It does not "ppear impo:;o,ible, ancI a Nuffield grant for two year:; to :t:.A.F.F.KO. has enabled the appointment of an officer to commence such studies early in 1962. Factors affecting fry survival and biology will also be examined.

Equivalent data have now also been collected for pairs of T zillii, and the comparative aspects are detailed in MI'. CricIland's report.

THE CULTURE OF TILAPIA IN PONDS

During the year, Dr. van Someren hilS continued ~tudies on the culture of. T nigra in ponds at Sagilna, being ably assisted in this by the Chief Fisheries Officer, Kenya, and the resident Fishery Officer at Sagana.

rl

13

While monosex culture, using males only, undoubtedly gives high yields of uniformly large sized fish, the problem of management of the overstocked dam in which breeding has occurred, resulting in a runted population, remains acute. It has already heen shown that the removal of a large quantity of fingerlings at regular intervals by netting will so decrease the biomass of the fish present as to allow of enhanced growth of survivors. But netting is a laborious technique, requiring capital equipment and labour, hence experi­ments have been directed to selective cropping of fry by traps which are selective for fry only. Such traps require little attention and can be operated by one man, but although experiments are still in progress, the various trapping methods so far tried have not succeeded in removing fry in quanti­ties sufficient to make any significant difference to the growth of survivors, since it appears more difficult to trap fry than adult fish, for reasons not yet fully known; the fry show a great deal of trap avoidance which is not evident in large fish, and the latter can in fact be decimated at an astonish­ingly high rate by the use of vertical opening traps.

Experiments on the growth of separated male T. nigra, and thus on population of a known density, held in a pond subdivided into various sized compartments by wire netting barriers, have shown that the growth of these is very sensitive to the area of bottom available per unit fish and not to the total water volume which surrounds the fish. Since T. nigra is primarily 'I a grazer, the results, although not unexpected, are 0 f considerable interest.

MIGRATION OF RIVER FISH

For four years an inclined grid trap has been operated on the small Ragati River at Sagana and Dr. van Someren has now analysed the fish migration records through this trap in relation to the daily hydrological observations on the river itself. These have shown the very high yields of katadromous fish such as eels, fish which are of high nutritive value, which can be obtained even by an inefficient trap on such a small river. The time of migration of these, and of Labeo cylindricus, are intimately related to the flood cycles and concurrent hydrological phenomena, though eels and Labeo show different migration patterns relative to each other. All Labeo and Barbus passing the trap_ in each direction have been tagged with silver wire and plastic tags, and though returns have been small they have shown a somewhat random pattern of movement; and analysis of the recapture data has shown a very slow growth rate in Labeo but a faster growth in Barbus tanensis. The results are being written up for publication.

HYDROLOGY AND ALGOLOGY

Dr. and Mrs. J. F. Tailing completed a 12 months survey of hydrological and algoJogical phenomena at the EAFFRO Open Lake Station south of Buvuma Island in Lake Victoria. These have shown that stratified condi­tions developed gradually in the Lake between August and November and continued with a sharp and deep thermal discontinuity till June, when the Lake waters again became mixed. Phytoplankton density reached its minimum at the time of maximum discontinuity, but increased again after the mixed state supervened in July, although phosphates and nitrates showed little

Ilincrease. Throughout the year nitrate was the nutrient most conspicuously in low concentration.

. Low oxy/?en values below the thermal discontinuity were rarely main­tained, and eVidence was obtained of a horizontal variation of thermal struc­ture in May, suggesting that an indirect communication between the bottom layers and the atmosphere may occur t.hrough horizontal water movements coupled with tilting.

The cyclical abundance of various phytoplanktonic algae has been fo,j]owed through the year, and while marked seasonal changes of temperature

14

and illumination cannot he direct causes of the cyclical variations found, there may be connections with seasoml! I-ainf'll] am! the extent of vertical and horizontal water movements. The seasonal cycle of }felosira in particular may be caused by a redistribution of this diatom from sba]low marginal areas rather than by vertical water movements.

Estimations of photosynthetic productivity have so far not shovvn any very unproductive season and tile average areal productivity ha~; been main­tained at values several times grcaLcr than the highest measured for spring diatom maxima in English lakes.

Comparative photosynthesis expel-iments have been conducted in <l

variety of other lakes in Uganda, showing conditions different from those in Victoria, such as Edward, (;-:orge, Mulche, Albert ;wd Bunyoni.

At the same time Dr. and Mrs. Tailing were able to analyse a variety of other Ethiopian and East African lakes, these providing an interesting back­ground of comparative limnology to their main project; and some features of outstanding interesl hoth as regards chemical comjJosition and ;dgal com­position have been discovered in these other lake~'.

The importance of seasonal algal studies to such phytoplankton feeders as the Tita-pia, and the effect of thermal stratification and oxygen level varia­bilities to bottom dwellers such as I3Ct!/i'bls and Latc.I, cannot he overemphasis­ed, for on them may depend much of the success or failure of local fisheries apart from predation by man; the fish themselves are sensitive to the environ­ment in which they live and cydic;t\ changes of the nature found by Dr. and Mrs. Tailing may markedly influence migration or survival. This is axioma­tic for the basic study of any fishery or fish popUlation, and far too little is yet known about this. Hydrological work on Victoria and other Jakes has been distressingly discontinuous over the past years and these studies have pin-pointed the necessity for full time study of the continuance of these fundamental phenomena, particularly on a synoptic basis.

GEAR RESEARCH

No final decision was taken during the year concerning the new vessel, and it was not possible to do any of the work required on- ring-netting and trawling etc. which is still required, IJoth for fish popubtion analysis and economic development.

Two aspccts of minor gear resell-eli hr\\'c 1>:':(:11 undcrti-lkcn, the rcsuits of which are probably of major inlrnrl:\lI,~e al1 ovcr tropical Afl-ica where gill net fishing is stamhrd practice.

Buoy Releases: Durin;:; the lalter pa.rt of 1960 ;md the earlier part of 1961, theft of our own experimental nl.,:ts set nightly ill the vicinity of Jinja in the N apo!con Gu! f, reached alarming proportions, over £150 worth of nets being stolen at night with liltle or no ch;)lic(~ of recovery owing- to local social conditions, even \-vith marked nets.

Gill nets are nearly always set on the bottom, with indicator buoys at the surface at intervals along the fleet, attaclied by ropes to the 'l.reig-hted nets. It is these buoys which give the net position Jvvay, making theft ;dJ too ea~;y.

In fact 011 one occ;)sion when one end of c'. set net \-vas being Ii fted late at night to l1lal-k fish, it was found that the other end was at the same time ;)Iso being lifted and loaded into a Canoe of thieves in the darkness!

15

Therefore if these indicator buoys could also be sunk at the time 0 f setting, to be released again to the surface just about dawn after a predeter­mined interval, the position of the nets would then be completely hidden from potential thieves. Some form of mechanical or soluble link holding a bight in the buoy rope would achieve this, the link fracturing after a set interval, releasing the bight and allowing the floats to surface. .

Mr. E. L. Hamblyn studied this problem and after trial of numerous devices, evolved a soluble link of a gelatin/glycerine mixture which dissolved slowly in water and factured under load after some hours. This was by far the most promising approach, but could not be developed further with II the chemical resources at our disposal. The problem was therefore pursued in cooperation with the E.A. Industrial Research Organization, who finally found that squares of ordinary carpenters' glue 20 x 20 x 7 mm in size with a 4 mm central hole (through which link strings were passed to secure a bight in the buoy rope) would fracture after periods of 6--8 hours at the normal Victoria water temperatures. These were given trials by Mr. Roberts, using made-up or standard native ambatch floats in Victoria and his report II is appended. Numerous settings without loss were then made, but it became

IIapparent that (a) the fracture period was not long enough, and (b) there was considerable variability in fracture time according to the source of the glue, and probably its moisture content before use.

It is essential to produce articles of absolutely standard performance and hence the matter is again in the hands of E.A.I.R.O. who, together with Messrs. Bulleys Tanneries, are now concentrating on producing a standarised product of reliable performance, for marketing. These links, though expend­able, must be extremely cheap to produce in quantity for local sale. Theft is common among the fishermen who are as dishonest among themselves over their own property as they are with ours, and they would readily use such theft prevention devices if cheap and reliable enough. Net thieving is prevalent all over A fric~ wherever gill nets are used, and the market would be wide and lucrative. ­

I t cannot be emphasised too strongly however that fishery officers and others concerned must not use or market any glue which is locally available, until a standardised reliable product can be put on the market. There could be nothing worse than such a method being unreliable or going off at "half­cock" as far as African fishermen are concerned, and advice is being sought to protect the Organization's interest in the development of this device.

As an instructive sidelight however on what may happen, in the restricted waters of Napoleon Uulf. the fishermen have realised that they are being baffled somehow and after a time have taken to observing the area where the launch has been setting nets, then going there after dark and dredging for the concealed nets with a stone attached to a long rope-with unfortunately an occasional modicum of success.

VERTICAL GILL NETS

The standard practice for netting Tilapia is to set horizontal nets sunk to the bottom during the night. Surfaced nets, or daylight settings, are usually very unproductive. Nevertheless it is known that there is a changing pattern of vertical distribution of fish, which may vary seasonally. These movements are usually known in an intuitive fashion by indigenous fishermen, but in order to find out more about this phenomenon Mr. Roberts has been experimenting with gill nets hung vertically.

Standard 4.}" or 4" gill nets were adapted for such vertical setting by lightly buoying both selvedges with cork floats, lacing one end of the net to a length of 2" conduit, and the other to a length of bamboo, and inserting

,i,

16

bamboo stretchers 8' in length across the nets at intervals of 10-12'. They were set in deep water in the Buvuma Channel at approximately 70', the nets being cut in equal halves gi ving 73' for the 4f' nets and 70' for the 4" nets. The nets were set separately but loosely linked by a head rope, using the buoy releases to prevent theft. On setting, the head of the nets could be: seen quite clearly just below the surface, indicating full depth coverage.

Nets were set in the late afternoon, lifted early in the morning, and the fish not removed until the nets were spread on shore and measurements taken.

The catches were poor, and there is no doubt that a standard gill net fished thus is not efficient, for the diamonds of the mesh are horizontal and not vertical; there was evidence from slime patches on the meshes that several fish had hit the net and failed to gill themselves with a mesh stretched thus.

Nevertheless from three settings of 2 x 41" nets and 2 x 4" nets in 70' of water over a mud-gravel bottom in October, T. esculenta were gilled at 5', 10' and 12' from the surface; Barbus at 39'; T. variabilis at 42' ; Bargus at 43', 53' and 54'; and a Monwyrus at 51'. It is of the greatest interest to record that no fish whatever were caught less than 18' off the bottom, and that on the same nights, local fishermen were found fishing nets for T. esculenta in the standard fashion at only 15' depth in this 70' of water; they clearly knew the depths at which these fish were at that time after somewhat stormy weather.

The use of this gear would be of great value to determine at which depths to set standard nets, and although catches will be small, continued use will give some indication over a period of time of the vertical distribution of commercial species.

PUBLICATIONS The following is a list of publications appearing during the year written

by members or past members of E.A.F.F.R.O., or by visiting research workers. The serial numbers are continued from the Annual Report for 1960.

133. FRYER, G. The Spermatophores of Dolo ps ranarum (Crustacea: Branchiura): their Structure, Formation and Transfer. Quart. J. Micr. S·ci. 101, (4), 56 (1960). p. 407---432.

134. CORBET, P. S. The Food of Non-Cichlid Fishes in the Lake Victoria Basin, with Remarks on their Evolution and Adaptation to Lacustrine Conditions. Proc. Zool Soc. Land. 136 (1), 1961, p. 1-101.

US, VAN SOMEREN, V. D. The Culture of Tilapia nigra (Gunther) in \VHITEHEAD, P. J. Ponde;: Pt. V. The Effect of Progressive

Alterations in Stocking Density on the Growth of Male T. nigra. Ii. Atr Agric. For. J. 26 (3) 1961,

136. EVANS, J. H. Growth of Lake Victoria Phytoplankton in Enriched Cultures. Nature 189, 4762 (1961) p. 417.

137. BEAUCHAMP, R. S. A.

138. VAN SOMEREN, V. D. WHITEHEAD, P. J.

139. VAN SOMEREN, V. D.

140. GARROD, D. J.

141. GARROD, D. J.

142. TJONNELAND, A.

143. VAN SOMEREN, V. D. WHITEHEAD, P. J.

144. FRYER, G.

145. FRYER, G.

146. ELDER, H. Y. GARROD, D. J.

147. EVANS, J. H.

148. FRYER, G.

17

East African Fisheries Research. Corona, 13 (2), 1961.

The Culture of Tilapia nigra (Gunther) in Ponds. VI-The Effect of Selective Crop­ping Methods on Breeding Populations. E. Afr. Agric. For. 1. 26 (4), 1961.

Seasonal Flood Fisheries in Kenya Colony. CCTA/CSA Hydrobiology and Inland Fisheries 4th Symposium, May, 1961.

The need for Research on the Population Dynamics a f Tropical Species, with special reference to "Ningu" (Labeo victorianus) from Lake Victoria. CCTA/CSA Hydro­biology &- Inland Fisheries 4th Symposium, May, 1961.

The Application of a Method for the Esti ­mation of Growth Parameters from Tagging Data at Unequal Time Intervals. N. Atlantic Fish Marking Symposium, May, 1961. 851, No. 41.

Light Trap Catches of Neoperla spio (New­man) (Insecta, Plecoptera) at Jinja, Uganda. Contr. Fac. Sci. Univ. Addis Ababa. Ser. C. (Zool), No. 1., 1961.

The Culture of Tilapia nigra (Gunther) in Ponds. Pt. VII-Survival and Growth of Tagged and Untagged Males of Different Sizes. E. Afr. Agric. For. 1. 27 (I) 1961.

The Parasitic Copepoda and Branchiura of the Fishes of Lake Victoria and the Victoria Nile. Proc. Zool. Soc. Lond. 137 (I) 1961, p. 41-60.

Larval development in the Genus Chono­peltis (Crustacea: Branchiura). Proc. Zool. 50c. Lond. 137 (1) 1961, p. 61-69.

A Natural Hybrid of Tilapia nigra and Tilapia leucosticta from Lake Naivasha, Kenya Colony. Nature 191, 4789 (l961) p. 722-724.

A Phytoplankton MUlti-Sampler and its Uses in Lake Victoria. Nature 191, 4783 (1961) p. 53-55.

Variation and Systematic Problems in a Group of Lernaeid Copepods. Crustaceana, 2, (4) 1961 p.275-285.

149. COTT, H. B.

150. FRYER, G.

,,;_. 151. GARROD, D. ].

152. GARROD, D. J.

153. GARROD, D. J.

154. FRYER, G.

18

Scientific Results of an Inquiry into the Ecology a.nd Economic Status of the Nile Crocodile (Crocodilus niloticus) in Uganda ,~nd Northern Rhodesia. Tra.1;ls: Zool. Soc, Lond, 29(4) 1961.

The Developmental History of Mutela bourguignati (Ancey) Bourguignat (Mol­lusca.: Bivalvia). Phil. Trans, R. SN, Land. Ser. B. 7.11, 244, (1961) p. 259-298.

Some Effects of Changes in Mesh Legisla­tion upon the Fishing Industry of Lake Vic­toria. . 3rd SymposiU11t Hydrobiology&­Inland Fisheries. Problems of Major Lakes, Lusaka 1960. CCTAjCSA No. 63.

The Rational Exploitation of the Tilapia esculenta stock 0 f the North Buvuma Jsland Area, Lake Victoria. E. A{r, (1gric. For. .J. 27 (2) 1961. p. 69-76. .

History of the Fishing Industry of Lake Victoria in Relation Jo Expansion of Mar­keting FaCilities. E. Afr, Agric. For. 1. 27 (2) 1961, p. 95-99.

Observations on the Biology of the Cichlid Fish Tilapia variabilis Boulenger in the Northern Waters 0 f Lake Victoria (East Africa). Rev. Zool. Bot. Afr. 64 (1-2), 1961.

19

APPENDIX A

PRELiMINARY NOTES ON THE RELATIONSHiP BETWEEN FEEDINC AND CROWTH RATE IN THE SILUROID FISH

BAGRUS DOCMAC (FORSK) by H. Y. Elder

As mentioned in the preceding part of this Report, experiments were begun in April to clarify the relationship between Bagrus and its main prey in the natural environment of L. Victoria, the cichlid fish of the genus H aplochromis.

Some of the results of the first two parts of these experiments, the relationship between size and basic food requirement and between feeding rate and growth rate are discussed in more detail in this appendix.

The fish used in the experiments were obtained from L. Victoria by a variety of methods, principally by gill net but also by seine net, traps and long lines. They were kept under controlled and measurable conditions in indoor aquaria and were first kept for an acclimatisation period of a few Jays to two weeks prior to incorporation into an experiment.

The fish were individually isolated in large aquaria holding approximately 1,150 litres, with two fish, separated by a partition, to each tank. The tanks were unheated and had a mean temperature of 22°C, with a diurnal variation normally within the range of plus or minus one degree. Temperatures were measured with maximum and minimum thermometers and all tanks in the aquarium house had the same temperature and showed the same pattern of fluctuation. In addition to diffused natural daylight of low intensity for a constant period of 12 hours per day, the tanks were illuminated for an 8 hour period within the hours of daylight by ordinary 60W electric bulbs. For the purposes of Parts I and II of the experiment, the physical conditions of water volume, light and temperature were considered to be uni form for all fish. The water was changed at periods from two to four weeks depending on the rate 0 f accumulatjon 0 f faeces and detritus.

Use was made of the strong thigmotactic tendencies of Bagrus, each fish being provided with a roofing tile under which it invariably remained almost inactive during the hours of illumination. The tiles were placed in a similar position in each tank.

Under aquarium conditions the fish show a diurnal rhythm of activity, becoming active at dusk and early evening at \vhich time, by preference, they normally hunt and feed. That Bagrus in the free state in L. Victoria show the same thigmotactic tendencies and diurnal rhythm of activity and feeding is supported by evidence from experimental fish catch figures and examina­tion of fish stomachs. Bagrus are taken in gill nets almost exclusively in dusk to dawn settings and the fish in the stomachs of captured fish normally show degrees of digestion corresponding to capture early on the night of the set or to a period 24 hours previously. During these experiments 'there­fore the fish were fed just after darkness. As described in the main section of this report, an under water force feeding technique was adopted for these experiments, this having the advantages that the quantity of food given (live Haplochromis spp.) can be determined accurately and that the amount fed can be varied as desired. Occasionally the fish did regurgitate food, in which cases the regurgitated remains were oven dried before weighing to convert the food to a dry weight basis, thus avoiding errors due to weight gain by water absorption. All fish were weighed and measured weekly, being anaesthetised in MS 222 Sandoz at a concentration of 250 p.p.m.

2Q

FIg. I.

7 ,;.<;

II~

Wt. (9).

10 12 14 16 .a 20 22 24 26 28 30 32 34 36

WEEKS

1000

Fig. 2.

Wt. (g).

o 10 20 ]() ",0 50 60 70 80 90 100

mg/9/ week

-----------

21

90

700 FiSh No.2.

600

Wt. Fig. 3. (0).

500

400 Fish No.3.

300I-,.--................,...,_..............-.-.................................-...............-.......................-.-.......~-.-...... o 2 4 6 8 10 12 14 16 19 20 22 24 26 28 30 32 34 36 3J

WEEKS.

Fig. 1. Opposite. A -detailed comparison between the weekly fluctuations

in weight (upper plot) and the weekly total of food given (lower

plot) for a fish being fed at "basic requirement level".

Fig. 2. Opposite. Curve showing the relationship between size of fish

(in grams) and the basic food requirement (in milligrams food per

gram predator per week).

Fig. 3. Above. Smoothed weight growth curves for two fish showing the

"net efficiency" of food conversion obtained for both fish during

each of two different rates of feeding.

-

11

----

22

The high degree of control and the constancy 0 f the physical conditions under which the fish were kept was made at the sacrifice of haling to work with relatively small numbers of fish. \VhcTcvcr possible thcrdorc experI­ments were conducted in duplicate and with controls. 'Cnder tht:~:.e condition" the techniques proved to be very sensitivc. This is illustrakJ by the high degree of correlation which was achieved for every fish belll'een the millor fluctuations of weekly avtTage feeding rate and growth, :Fig. 1. For the purposes of simplification in this account, only smoothed growth eun'e.:; are given for those other fish figured (Fig. 3.). l;

Due to the relatively short duration of these ex[)crimcnts the growth in weight has yielded more information than growth in length. In this preli ­minary report only growth in weight is considered.

In part I of the experiment, 10 determine the rehtionship between "maintenance food requirement" and the size of the fish, six fish varying between 60 g and 900 g were fed at rates which were found hy trial a!ii1 error to just maintain them al their weight at capture for period!; between 24 and 30 weeks. The maintenance requirement W;lS then calculated as the number of grams of food required per gram predator reI' week for the period a f the experiment. 'the results an: plottc~ ill Fig. 2 which shows an inverse relationship between maintenance requirement :md size of l,sh.

In Part II of the experiment, to determine the dlecls of variation in the feeding rate on the growth rate, fish were kept at feedinl'" rales varying­hetween complete starvation and rates probably exceeding any attaillt·C: normally under natural conditions. Amongst thOse tish being fed at rates in excess of the maintenance requirements and therefore showing f£Jsitive growth, an inverse relationship was found between growth and eA',ciency of utilisation of food, or "net efficiency". This Jatter is the percentage efficiency of utilisation of that amount of food ing~..;Led in excess of the lllaintenanct' requirement, or in other words the food availahle for growth. This can be demonstrated by a comparison either between two fish of the same size being fed at different rates or for anyone fish between two periods during which it was fed at significantly different rates. Table 1 and Fig. 3 gil'e thc lid efficiency for two different feeding rates for each of two different fish.

In the calculation of "net elTiciency" the ligures for "basic food re­quirement" were read from the curve obtained inPat-t I (Fig. 2.) of these experiments.

TABLE 1. Net efficiency of two different feeding rates for two diJlerent fish

Fish Period Av. weekly Basi(, fooL! "rt.. food A,'. ,",cekly Ne,L Ko. weeks ",t. of food rO(lnin~rnellt H,'"il:d>le jll(~renl(~)lt, "fficil'll('Y

(illc.) (g.) (g.) for I!:l'o\\'th ill wcig;ht % (g.) (g.)

1-8 108.3 25.0 S:U 20.9 25.1 2 23-38 30.4 25.2 5.2 2,f) 50.0 .- - ­

1-12 47.4 22.9 24.8 5.9 23.8 3 13-21 72.5 23.4 49.1 9.9 20.2

~._--

These results, some of which are heing repeated, tugelher with the results of Parts II, I V and V of the experiment, feeding rale,; ;\11(\ gTOI';lh rates of fish kept under different tel1lper:ltures. feeding' r:ttcs and i~rnwtJ:

rates of fish kept active, and mort: direct CSt.il;WtCS of lildaholi';l11 hy lhe oxygen uptake method, will be publ i~heJ in grealer dCl'ai I el~;e\'ihere.

23

APPE DIX B

NILE PERCH INVESTICATION

by E. L. }iamblyn

The year \I';\S marked by the establishment of a temporary laboratory at Bntiaha which allowed work to proceed more smoothly and systematically than hitherto.

The more important findings are noted below in general terms while a detailed accollnt of the knO\vil biology of Lalcs will be pnblishec1 elsevvhere.

Latcs fry up to 12 I11lll total length occur in the White Nile oetween Lake Ruhi and Hhino C,U11jJ in early April. They ranged in size from () I11m-12 mnl and were caught by a specially de:igncd net constructed of wire mosquito gauze which filtered out the fry frolll the water. Caught at the same time with the fry were a small number of dead eg-gs, and while it is imflos,'ible to state ,vith certainty that these eggs were Lates eggs the possi­bility that this .is so is high. ;1I1d it is unfortunate that no developing eggs \ven: found, The !Jet was fished, usually overnight, from a launch which lay moored along the river bank, so that only a very small fraction of the ri\cr water W;IS tiltere in relation to the river's volume. Judging by the frequency of fry takel) in the net the number of fry carried along the river hed mllst i.lc cnonnous. It was not possible to obtain records for any len;.,rth of ti'rll:, so it is not known for how long the 1.>reeding season lasts and with what particular environmental conditions, if <Lny, it is correlated. Evidence from. Iscwhere, not presented here, doe' suggest that the breeding season is Iimited in the ri "erine environment and may be associated with migratory movements of adult fish.

The situation in fhe lacustrin~ nvironment presents a contrast. Here frv of the same size as those recorded from the White Nile have been dis­co~'ered at different locations and at diHerent times of the year. At Ntoroko ;It the south end of the Jake fry were caught over a mixed sand and mud bottom between the outflow of the-'\Nassa and Musisi Rivers in mid-February. At Katel1'a on the ea,;t shol'e of the lake. a lilLie ciistance south of the month

f the ~vlurehison Nile, specimens were ~lhtaincd in late March. These were fund in a bay with a wtedy bottOm, At Dutiaba fry were prescnt through­out mosl of Augllst alld Septemher where they were found over a weedy bottom in the Fisheries ,Harbour, It should I,. noted here, as with the r 'corels from thc \,Vhite N iie. that the observations are of the isolated kind a" it \Vas not po,;sible to S<llllllle the stations continuously or c\'ell at frequent intcn;l!::; Wilh the .xceplioll of Dtltiaha where continuous daily observations '.'.'Cn'11la<1e, En'l1 here nO records exist ahout the date of the first appearance of fry as the Illodified l'eterdale trap::; caug"ht fry the first time they were .u:~ed, so that fry llltiSt have been presf.'nt for some time.

All the I'eeonls ha ve reference to shallow inshore water where the depth IS nOl greater tllan twellty feet, thus it i,; of great interest to record the (aptme of fry fmnl !:)() ft. of water some 11liles offshore at Mlima (ahout i'/\e miles north oJ Buhuka) ill mid-February, The fry of Mlima occurred tOi:.etl".cr with a 2 em indi\'idual an<1 a l() em specimen, and were included in ;l catch containing many hundreds of prawns alld some deep water Haplocltro­Jm~\, As pl'awns are known to exist between the bottom and the 90 ft. mark, :It the same tilHe of day Cit anotlIel' station, it is fairly certain that th se

~

24

captures were made on or near the bottom. The gear used was a large filter net of the same type as used in the White Nile, but operated as a trawl for a few minutes at a time, with the launch dri fting gently.

It is clear from the Butiaba evidence that the breeding season includes the whole of August. Evidence from size frequency of the captures from Danish eel traps and large Peterdales at this station which trap fish up to 30 cm indicate that breeding must occur throughout many months of the year if not all. Insufficient is known to suggest times of breeding maxima if they occur.

BEHAVIOUR OF FRY AND FINGERLINGS

The fry captured in Butiaba Harbour fall into two groups. One group consisting of individuals up to 12 mm occupied the bottom few feet and seldom less than 10 ft. below the surface. They were outside of the reed and papyrus beds, where attempts to capture them were unsuccessful. The second size group between 12 mm and up to nearly 20 mm Jived a solitary existence amongst the epiphyte clad piles of the harbour installations all the way from the top few inches to the limit of vision down through the water, usually about six feet. These fish showed the characteristic attitudes and behaviour in hunting shown by larger individuals and were many times observed darting around the algal filaments to capture the food associated with this community.

Fish larger than 20 mm were never observed in this habitat, but do occur leading a solitary existence among-st the fringing vegetation of shallower waters. They appear to live alone until they reach about 10 cm and exhibit the most beautiful colouration of a disruptive nature which is seen at its maximum development at this stage of the life history. The colour pattern is of brown marbling on a yellow ground extending over the spines of the pelvic and anal fins. The fin colours are later lost while the intensity of the colours generally fades with ag-e. Marbling is retained by many fish up to 30 cm but is rare above this length. Care is needed in describing and interpreting colour pattern in this species as changes with reference to light are evident in aquaria and changes are rapid after death. Fish in the 2-10 cm size group were never common in the large Peterdale traps although they are well within the selection range and were only occasionally caught in the infrequent and inefficient small seine hauls made through the fringing­vegetation.

Fish between 10 cm and 15 cm may be seen in numbers of two to five in the shallow water a little way from the fringing vegetation in depths between five and six feet. At this size the first signs of shoaling appear which become more marked with increasing size. Shoaling behaviour goes in step with the occupation of deeper water, shoals of larger fish occurring with increasing depth. Up to about 20 cm. shoals of up to SO fish have been observed under the hulls of vessels moored near the papyrus, while larger shoals of several hundred individuals may be seen swimming about the piles of the boathouse. These latter shoals were composed of fish up to 30 cm. Shoals of similar numbers and size were to be seen near the seaward mooring­chains of the railway dock.

Shoals of larger fish were present in Butiaba Harbour in May. Specimens from the shoal were caught in a Danish eel trap which took fish both by day and by night. As many as 20 fish were trapped at a time and shoals were often observed formed around the restrained but unalarmed fish in the trap which acted as a focus for the shoal. This group of fish left the

2S

harhour during June and have not been caught in comparable concentrations in 'later months. The traps however, caught fish up to 25 cm when set in shallow water near the fringing vegetation during the night. This has been the case from May to September and it is known that fish of this size may be caught in the harbour in most months of the year.

Fish between 25 cm and 30 cm fall within the selection range of a 3" gill net and large numbers were taken off Butiaba Island between April and September. Small gill nets of 11-" and 2" mesh captured Lates between 16 cm and 22 cm during August. The gill nets were set in water from 20'-70' deep off the east shore of Butiaba Island so that small fish of the same order of size as those found in the Fisheries Harbour coexist in relatively deep water over a mud and sand bottom. The tentative conclusion is that fish of all sizes are present in this area for most if not all of the year so that in the lacustrine environment as distinct from the riverine environ­ment hreeding occurs during most of the year.

The largest fish observed in Butiaba Harbour was about 201b. and was seen in company with smaller fish. At this time small fry were present and it may be that this large fish was a female ready to breed.

The fry and one fingerling from the Mlima deep water station may be contrasted with the inshore forms. The fingerling exhibited none of the brilliant colouring of the inshore fish of similar size, while the 20 mm fry was nearer in appearance to the pattern exhibited by 15 mm fry from inshore and riverine localities (which are exactly comparable in colour pattern, at least very soon after death). It is probable that these fish originated from eggs spawned in shallow water and were carried by the current to the position where they were caught. This hypothesis (for which there is some inde­pendent hydrological evidence) has an important bearing on the problem of the validity of the two subspecies Lates niloticus albertianus and Lates niloticus macrophthalmus. The original separation was made on morpholo­gical and ecological characters which became suspect when larger series of fish were examined. This doubt has been confirmed by the present work and it is now considered that the separation is no longer justified on either ground. There may, however, be physiological differences and it is my present opinion that this species is in a state of almost discernible speciation in the lacustrine environment.

GoNAD STATES

It was reported last year that some mystery surrounded the whereabouts of "running" females in all the environments where Lates were found. Examination of many more female fish has revealed a further stage in ovary maturation not previously recognised. In this condition the ovary is cream yellow in colour, fully obovate in section and the dorsal lobes only indistinctly defined. The blood vessels are not very obvious. The organ is firm to the touch and has a "full" appearance. This stage is termed "ready" as distinct from "running" as a female has never been recorded showing eggs about the genital papilla or passing eggs after capture.

The stage in maturation previous to "ready" is recorded as ripe. In this condition, which is much more commonly met with than the later stage described ahove, the ovary is yellow-orange in colour with well defined dorsal lobules and rather sharply obovate. The blood vessels are quite sharply defined and the organ is soft to the touch. It should be noted here that a "ready" ovary from which eggs have already been voided may be slightly oedematous.

,;.s

26

The foregoing remarks apply to young females which have probably not bred more than once. In older fish the opacity of the investing mem­branes disguises the true colour of the contents when the developing condi­tion can be confused with the spent condition.

In the spent ovary spirally arranged fibres can be easily seen in the semi-opaque investing membranes. The colour of the contents is dark red and the colour of the organ as a whole may be pink. But for the fibres seen in the investing membranes of the spent ovary the condition corresponds to that of the developing organ. The two conditions may be immediately differentiated by smears which have been used when doubt existed after macroscopic examination.

The quantity of fat in the fatty body and mesenteries varies consider­ably. There is usually little in ripe and ready fish, but considerable amounts may be present in fish with developing or with "spent recovering" ovaries. Large quantities of mesenteric fat may be present in fish in which breeding is in progress which is thought to be used as the breeding season progresses.

FOOD

The list of prey animals has been further extended and it is established that Lates will take any animal that moves and continues to move in accord­ance with the principle described in last year's report.

A most important element in the food of Lates in Lake Albert is the prawn Caridina sp. Lates throughout the size range feed on these animals which are eaten in great numbers. Caridina is an abundant element of the benthos in Lake Albert where they occur in great numbers between the bottom and the 90 ft. level off Kaiso and Mlima and, judging by the results of stomach content analysis, their numbers must be considerable for most of the year. These prawns also occur in the shallow water of the bay, where they are eaten by Lates from 15 cm upwards.

Lales, which feed primarily on prawns, produce characteristic faeces which are buoyant on voiding or rapidly become so. These faeces may be single boli or two or three together connected by a mucous sheath. The colour is a characteristic pink which is definitive for the presence of Lales when no other evidence is available. The rate of destruction of these faeces depends on conditions at the water surface. The faecal masses maintain their form and buoyancy for over 24 hours in still water under the conditions prevailing in Butiaba Harbour in May.

vVhen prawns are heavily utilised changes in the gut may occur. The pyloric caeeae become yellio"v pink and gas bubbles are common in the mid gut. The condition appears to be pathological only in appearance. Whether Lates can successfully maintain itsel f on prawns alone for a prolonged time is unknown.

The nature of the faecal masses also indicates conditions in the stomach as f?eces of different colour and shape derived from the digested remains of a prey are evacuated in the order in which digestio!l and maceration of the prey occur. A time lag occurs when the stomach is empty and movement of material down the gut may stop, leaving a residue in the hind gut. This residue, derived from the last sta~es of digestion of the previous meal, is voided before those faeces derived from the first stages of digestion of the next meal. These observations are of value ill feeding experiments and where fish may not be killed and eviscerated.

27

PI~EIJL\TION STUDIES

The nile perch has been suggested as the controlling predator .in pond culture of Tilapia where optimum yields of large Tilapia are the objective. The theoretical action of the predator is to prevent a large population of undersized Tilapia developing. This may be achieved by the continual removal of a fraction of the fry population so that the resulting Tilap·ia population is held at a level w here the optimum use is made 0 f the basic produ\;tive capacity of the body of water. In other words the desired ailYl is :the creation of the optimum "lebensraum" for the Tilapia, which can be achieved in a number of ways provided the fertility of the body of \vater IS know.n. The techniques used include monosex culture where the absolute number of individuals is known and increase of the population is impossible, and various mechanical methods of removing the estimated undesirable f.i-action of the population. In the predator-prey approach the predator acts flS a substitute for control by mechanical means, e.g. netting and trapping. If the, predator is itself a valuable /ish a further advantage accrues.

A preliminary experiment has been made using Lates in this capacity where the predation pressure exerted by eight Lales was tested against two thousand Tilapia nigra (Giinther) over a period of one month in a one acre pond. The Tilapia used were vulnerable to predation by all the Lates, except one, throughout the period .of the experiment. At the end of this experimentJ;here were 748 Tilapia survivors (37.5%) against a survival in the adjoining control pond, in which Lates were not present, of 1,441 (72%). The gain in weight of Lates was 1.25 lb. for the group having an average \veight at the start of 1.73 lb. Using the artificial standard of number of fry consumed per day !Jer predator we have a figure near three.

For comparative purposes the large mouth hlack bass were used as a predator in approximately the same pro!Jortion as Lates against a similar prey population in another experiment. The results were rather similar to those for Lates. Survi-v.al of prey was 869 (43%), seven black bass surviving of the original eight, one dying on the second day of the experiment. The average weight of the black bass at the start was 2.1 Ibs.

Experiments in progress indicate that predation pressure tends to flatten the expected population curve so that the discrete modes characterising parent and subsequent generations are no loager discernible. Also there appears to be an overall reduction in the total prey population. In one case where the parent prey population consisted of individuals outside the preda­tion range of Lates better growth of the parents was obtained than in the parent population in the control pond. It is not proposed to discuss these results here because the experiments are still in progress. It should be realised, however, that there are many considerations of great importance which it will be difhcult to assess. Chief of these is the reproductive capacity of Lates in this situation; the ultimate growth attained by Lates which widens the range of vulnerability of the Tilapia; the inter-relations between the growt.h rates of the t~o species; and the competition for the same plankton orgamsms as food In the early stages of development. One point of paramount importance, basic to pond culture, is the assessment of the water qualities of the pond before any riecisions are marie concerning the species of fish to be used.

ANGLING

. In recent year~ anglers have ~ol11l11ented that some of the trolling grounds In Lakt: Alben \vhlch formerly)', Ided good catches are no longer so produc­tive. One such g-round is that at Hutial>a Island. During August a long

';J

28

line was fished in this area with considerable success. At the same time angling results were poor even when lines were trolled in close proximity to the long line. No firm conclusions may be drawn from these observations because of their very limited duration.

Nevertheless it may prove profitable to reconsider angling techniques. The use of an artiilcial trolled lure superseded the use of live Tilapia because of the obvious advantages of such a lure. The method of predation in Lates requires only that the prey shall move and continue to make movements so there is no reason why the lure should be made to move bodily through the water as is the present practice. The most obvious method that presents itsel f is based on the traditional float fishing method for pike but omitting the sensitive float and using a single hook. A small hook bound to the top hal f of the shank of the holding hook would serve to attach the prey fish used as bait. The use of a small hook and the limited buoyancy of the float would alJow the use of Alestes as bait. These fish are very common in the inshore waters of the harbour and are readily captured in quantity with a small coJlapsable trap or obtainable from native fishermen. They are tough, easily handled, and easily kept in a keep net. It should be remembered that Lates seldom holds its prey in the mouth, which obviates the need for refinements of the gear. This method is suggested specifically for the capture of large fish and involves the principle of selecting the most likely area. Large Lates are solitary lurkers except during the ~reeding season when they may concentrate. As there is no clearly defined breeding season at Butiaba it is probable that Lates are present on these grounds throughout the year, so that this tackle fished a few feet above the weed beds in Butiaba Bay near the reeds around the island might prove most successful.

29

APPENDIX C

BREEDINC STUDIES ON TlLAPIA ZlLLlI AND TlLAPIA NIGRA

by c.c. Cridland

Broods were recorded from seven pairs of T. zillii which had been reared from previous experiments in the Laboratory, observations on these being made for a period of 19 months.

Broods from 15 pairs of T. nigra were recorded over a period of six and a hal f months, these fish being deri ved from the stock at the Inland Fishery Research Station at Sagana in Kenya.

Breeding tanks for both were situated in the open, partly shaded by a tree, each containing 132 litres of water. Each tank was divided into two equal parts by a nylon netting partition, and the male and female of each pair were kept separate in each half until ripe. Ripeness in T. zillii is indicated by certain behaviour traits which will be described elsewhere, and in the female T. zillii by a protrusion of the white genital papilla, apparently indicating ovulation has occurred, and by cet'tain behaviour postures. Male T. nigra remain in a state of ripeness for considerable periods, this being indicated by a striking colour pattern in which the silver-coloured iris is prominent on a dark blackish-brown colouration of the anterior part of the fish, certain behaviour postures, and usually a protrusion of the white genital papilla.

Ripe female T. nigra will usually spawn within a few hours of papilla protrusion, often by themselves if no male is present-the eggs being picked up and mouth-brooded even if infertile; but being expelled again in 2-3 days in this event. Spawning of T. zillii is however a more protracted process. The breeding behaviour of both will be described in full elsewhere.

For spawning, the- females were placed when ripe in the male com­partment; if introduced before fully ripe, males would attack and kill the female in both species, though in two cases the females killed the male. Although herbivorous by nature, both species thrived on a diet of oligochaete worms with additional lettuce for T. zillii, and Daphnia and Artemia nymphs for both. The water in the aquaria was changed frequently, and da,ily water temperatures recorded.

In the case of T. zit/ii, male and female were removed, weighed and measured after spawning. Eggs were removed from the breeding tanks, being scraped gently off the stone or glass on which they had been laid" and hatched in an "artificial mouth" in the laboratory-a large beaker into which water was run gently through a filter funnel to keep the eggs in gentle motion.

The same technique was employed with T. nigra but, after spawning the female was allowed to retain eggs in her mouth for a period up to twelve haUl'S to enSl1l'e full fertilisation, with a trickle of water running through the tank. After this period eggs were forcibly removed from her mouth and hatched in "artificial mouths" in the laboratory. In other cases the female was allowed to retain eggs in her mouth until hatching and fillal exit of the fry some sixteen days later.

Fry from both species were reared in the laboratory in aerated round. glass. bowls, containing 1(; Htres of water, and fed on nymphal Artemw shflmps.

....

30

Nine broods of artificially hatched T zillii were recorded for 30 days after hatching, and four broods left with the female in the aquarium. Out of seven broods, 100 fry were used in each batch for recording lengths, with 100 fry as replacements, and of the former a random sample of 20 fry were measured daily.

\Vith T. nigra 20 fry were measured daily, 60 being kept as a replace­ment from each brood, any dead fry in the batch measured daily being replaced with others of the same size in order to maintain the fry tanks at a constant population density. Where broods had been left with the female, fry were counted and measured the day they were first released from her mouth, six broods being recorded in this fashion.

Both species spawned only in sunshine, at any time of the day. In T. zillii the spawning process takes one to two hours; in 1'. nigra spawning is complete in a much shorter time, only about half an hour for the whole proceeding, of which egg laying itself occupies little over hal£-a-minute.

1'. zillii eggs are laid in adhesive strings. attached to a large stone, or on the bottom of the tank, or to the glass sides of the aquarium. 1'. nig'ra eggs are laid in a pile in a shallow sand depression made usually by the male, and picked up immediately by the female together with milt, either shed freely over the eggs by the male, or deliberately sucked in by the female from the vicinity of the male's genital papilla.

T. ZILLII

Hatching of eggs at an average temperature of 23.2°C in the laboratory took an average of 2.2 days, the maximum time was 2.4 days and the minimum 1.9 days. The average length of alevins when hatched was 4.03 mm, the maximum length 4.23 mm and the minimum length 3,89 mm. The yolk sac of the alevins was absorbed after 7.1 days. The average length of fry at the time the yolk sac was absorbed was 6.58 mm, maximum 7.31 mm and minimum 5.93 mm. Fry commenced feeding before the yolk sac was fully absorbed.

The average hatching time of eggs left with the female at an average temperature of 25.5°C (maximum 28.9°C, minimum 22.2°C) was two days. The alevins were not molested by the female, but as soon as the fry com­menced to swim freely, they were often eaten by the female. No counts or measurements were done on eggs, alevins or fry in cases were the brood was left with the female.

Fry from brood 2 of pair No.4 were fed on Daphnia and copepods. They made a poor growth, reaching an average length of 10.6 mm at the end of 30 days, and the survival rate after that time was the lowest of all the fry studied, due to overcrowding and unsuitable food. Ostracods were also tried as food but found to be unsuitable as the young fry were killed by the crustacea. All other broods studied up to 30 days were fed on nymphal Arternia shrimps; they reached an average size of 13.1 mm in 30 days.

From seven pairs, 35 broods were recorded over a period of 19 months, and eggs from 30 broods counted. The shortest interval between broods of anyone pair was 15 days and the longest 106 days.

In four broods the females were allowed to retain fertile eggs, and the average interval between broods was the same as for those where the eggs were removed.

31

The number of eggs laid is clearly a function of the size of the female, although the relationship is not directly proportional. The smallest female, which laid fertile eggs, was 11.0 cm. She produced 2218 eggs in her first brood. The largest female was 18.4 cm and the number of eggs from her first recorded brood was 6702.

The brood size increased with successive broods. The average number of eggs per gram body weight of the female was 65.5. The average size of the eggs was slightly smaller in the larger broods. Data are as follows :-

Av. length Av. width Total number of eggs from five large

broods 28,290 1.53 mm 1.10 mm. Total number of eggs from five small

broods 9,639 1.56 mm 1.12 mm

Alevins hatched from large eggs did not attain greater size than alevins hatched from smaller eggs, nor did the fry attain faster growth at the end of 30 days.

The percentage survi val at hatching did not differ from the early to the later broods, but on average there was a lesser survival of fry from the earlier broods than from the later ones after 30 days.

The survival of eggs hatched in the laboratory varied from 7.9% to 68.4% with an average of 43.1%.

Survival of fry 30 days after hatching varied from 53.7% to 95.4% with an average of 82.7%..

The average length and 'weight gained for males and females during the experiments were:­

Males 2.54 cm and 38.1 g. I'emales 2.40 cm and 21.9 g.

T. NIGRA

Hatching of eggs in the laboratory at an average temperature of 23.2°C toole 4.3 days; the yolk sac of the alevin was absorbed after 9.8 days, and fry commenced feeding before the yolk sac was fully absorbed. Fed on nymphal Artemia shrimps, they rcached a size of 17.0 mm in 30 days inside the laboratory. III the breeding tanks in the open where females were allowed to retain fertile eggs in' their mouth, the average temperature was 25.5°( during the time of hatching.

The number of eggs laid is also clearly a function of the size of the female, although the relationship is not directly proportional. The smallest female which laid fertile eggs was only 7.5 cm and produced 44 eggs. The largest female was 17.0 cm, and the number of eggs from her first recorded brood was 692. The brood size increased with' successive broods. The average number of eggs per gram body weig'ht of the female was 10.4. The average length of eggs from large broods was smaller than eggs from small broods, but the average width of eggs from large broods was'larger. Data taken from the five largest and !lve smallest broods are as follows:-

Av. length Av. width Total number of eggs from five

largest broods 3,793 2,39 mm 1,98 mm Total number of eggs from hve

smalkst broods 355 2.41 mm 1,84 mm

32

Large eggs did not necessarily hatch into large alevins, but alevins originating from large eggs attained a larger size after commencing feeding than a!evins originating from small eggs. The faster growth continued throughout the experiment up to 90 days. Data are as follows :-

Av. size 30 days 60 days 90 days of eggs

Fry under 9.0 mOl at L. 2.18 mm 16.25 mm 23.6 mOl 27.0 mOl

.; start of feeding: W. 1.64mm

Fry over 9.0 mm at L. 2.50 mm 19.20 mOl 25.0 mOl 29.3 mOl start of feeding: W. 1.88mm

The survival of eggs artificially hatched in the laboratory varied from 13.5% to 100%, with an average of 60.7%. The reason for this variability is not known.

Survival of fry reared in the laboratory, 30 days after hatching, varied from 34.5% to 96.2% with an average of 67.3%. When females were allowed to retain fertile eggs in their mouth, the survival rate of fry from the day of first exit of fry to approximately 30 days after hatching varied from 51.8% to 90.1%, with an average of 70.3%. ­

Broods reared in the laboratory grew to an average length of 17.0 mOl in 30 days, 24.4 mm in 60 days and 28.5 mm in 90 days. Broods left with females grew to an average length of 16.8 mOl in approximately 30 days, 30.5 mm in 60 days and 48.5 mm in 90 days.

The average length attained in 30 days for both was approximately the same, but a considerable increase in growth occurred in 60 and 90 days of fry reared with the females in the open. Fry should not be kept in "green" water after they are six weeks old, since the algae (Euglena sp.) appears to cause their death if eaten. The tank should be flushed with clean water every day.

Sexual dimorphism in growth occurred in fish from pair No.5, 1st brood, at 90 days. Out of 47 fish, 34 were males with an average length of 58.2 mOl. The average length of 13 females was 53.8 mm. It must be noted that a high mortality occurred in the brood when 60 days old; it is presumed that a greater number of females died, since this sex ratio is abnormal. The average length of males and females from this brood was 57.0 mm. The largest male was 62.0 mm, the smallest 54.0 mm. The largest female was 58.0 mOl, the smallest 50.0 mOl. T. nigra fry keep the "'Iilapia spot" on their dorsal fin up to 60 days old, after which it disappears.

The average length and weight gained for males and females during the experiment were:­

Males 2.0 em and 22.5 g.

Females 1.S em and 11.1 g.

t J .. · -- - 1

33

APPENDIX D

POND CULTURE STUDIES ON TILAPIA NI&RA

by V. D. van Someren

Prior to assuming the Directorship of E.A.F.F.R.O. in 1960, I had been engaged on an intensive programme on the pond culture and growth studies of Tilapia nigra at the Inland Fishery Research Station at Sagana in Kenya Colony, and through the courtesy of Major D. F. Smith, Chief Fisheries Officer, Kenya Colony, it has been possible to continue these on under the immediate supervision of NIr. C. E. P. Watson, Fishery Officer i/c the Sagana Station, and carry them to a conclusion.

CROPPING EXPERIMENTS

The culture of T. nigra in ponds is best effected by monosex methods, in order to obtain yields of fish all of a uniformly large size after a period of time. It is usual to grow males only, stocking them on a basis of sex separation carried out either by genital examination of immature fish, or on colour of the pelvic fin in mature fish taken from a mixed stock pond. The technique and subsequent growth features have been fully described by van Someren and Whitehead (E. Afr. Agric. For. J., 1959-61).

It has always been realised however that there are many ponds and dams in which this technique is not possible, because they are already heavily stocked with breeding fish of both sexes. In such cases some form of differential cropping is required in order to reduce the biomass of small fish and allow greater growth of survivors. The effect of differential removal of various age classes by hand sorting of seine net catches has already been described (van Someren & Whitehead, lac. cit.) For use in an ordinary farm pond however, -seine netting, while effective, is not ideal for it requires relatively expensive capital equipment, and labour to operate the net.

Some form of simple trapping is therefore considered the best method, the traps costing little to construct and being operated with the minimum attention by one person, the best catches usualJy being made at night time. One paper on this is already in press (van Someren. The Culture of Tilapia nigra (Gunther) in Ponds. Pt. VIII-The Effect of Cropping Breeding Populations by Trapping. E. Afr. Agric. For. J.). This describes the results of using two forms of wire mesh traps-a simple funnel entrance trap, and a "Peterdale" trap with a long vertical entrance slit, in a 4-acre dam heavily stocked with mixed breeding T. nigra. The entrance of these traps was designed to retain only the larger fish over 15 cm and the daily operation of these two traps led to a decline of the exploitable stockof large fish at the very high rate of one sixth by numbers per day.

Within only eight days practically all the larger fish of edible size had been removed, there being a marked decrease in all length frequency classes ­subsequently recorded. Over a period of 24 days, this dam yielded a crop of 90.1 Ibs. per acre, the proport~on of males in the catch increasing as the catch decltned. Such trapptng IS less effective during periods of high wind.

Eleven months later, no fish larger than IS cm were recorded in numerous seine net samQles, the trapping having presumably led to a greater survival of the unexplolted fingerlings and a consequent density dependent suppression of their growth rate.

Ii

34

I

i

,u.

I'

As has been repeatedly shown therefore, in small ponds of this size, removal of only the large edible fish will not maintain the supply of large fish, and it is therefore essential to remove periodically large quantities of the trash fingerlings which, if left in numbers, will never grow to a reasonable size. New traps have therefore been constructed with a vertical opening no wider than -Y', which have been proved to be selective of fish from 8 cm downwards only. Three ponds of one acre each have been allowed to accumulate equivalent populations of mixed breeding T. nigra and these have been fished respectively with one fry trap each at (a) daily intervals, all fry caught being killed and removed: (b) at weekly intervals, all fry caught being killed; and (c) at weekly intervals, all fry being returned alive, the latter being used as a control pond. In addition, each pond is also fished once weekly with a trap designed to retain larger fish only, in each case these being measured and weighed, to check on growth, and returned alive. In all cases also, all fry have been measured and weighed before killing or returning. This experiment has now been in progress for 4 months at the time of writing, and results to date have been inconclusive.

In the pond which is fished once daily, only some 1700 odd fry, weighing some 26 lbs. have been trapped and killed in this period; an average of 19 per trap per day; 400 fry weighing some 6 lbs. in the once weekly fishing and killing; an average of 21.5 per trap per day; and in the control pond, an average of 6.8 per day.

Adult fish catches have averaged 31.5 per on<;e weekly catch in the daily fry kill pond; 23.5 in the weekly fry kill pond, and 12.5 in the control pond. Average length of adults were respectively 22.0 cm, 22.0 cm and 22.6 cm after four months, being 19.3, 19.2 and 19.2 respectively at the com­mencement. However a check seine haul in each showed the population structure in each to be remarkably similar at the third month.

It is evident from these current results that extraction of fry at this rate is of no use whatever in reducing the total biomass of fry to allow of greater growth of larger fish.

Although preliminary trials of these fry traps showed catches of 150 and more per night, it is apparent that when fished consistently, even at weekly intervals in the same pond, fry and fingerlings appear to exhibit some form of trap avoidance even though care has been taken to fish the trap in a different part of the pond each time. It is also evident however that this trap avoidance is not shown by the larger adult fish which are caught once a week in consistent quantities- ill spite of the fact that they could learn avoidance as they are returned alive each week.

\iVork is still in progress to modi fy the trap and its technique of use in order to increase fry catches. Simplicity of trapping has obvious advantages over the complexity of netting, and some of the new techniques are lllore promising.

LEBENSRAUM EXPERIMENTS

It has always been assumed that the rate of growth of an individual T. nigra and the size achieved at any given moment, is probably a function of the grazing area accessible to it, irrespective of the volume of water by which it is surrounded, since T. nigra is not primarily a filter feeder but a grazer by habit. Within any area of bottom, the area available to an individual fish is therefore presumably a function of the stocking density of fish per unit area.

J

35

This supposition is in certain respects an over-simplification, since many other factors such as depth, turbidity (and the effect of these on temperature of the water) and the nature of food available clearly play a part; this renders growth comparisons from pond to pond of equivalent size difficult of interpretation.

In anyone pond, however, this supposItion should be susceptible of some form of proof, and preliminary experiments have been conducted on this. One half-acre pond was divided in half by a r mesh wire netting barrier, and one of these halves again divided in the ratio of 3: 1 by a cross barrier. It was filled with about 18" of water to about 6" below the top of the netting, and the netting was so bedded in the bottom that intercommu­nication was impossible. In the first trial, sexed immature fish were used. in order to eliminate population changes by breeding and these were stocked into the pond in the numbers of 50, 40 and 10 into each successively smaller division, thus giving each fish the same unit area of bottom grazing in each division, though all were in contact with the same volume of water.

Such experiments are not easy to control because of the differential survival in the different compartments. In this experiment survival rates were 82%, 83% and 900/0, and increments after seven months were 7.7 cm, 8.5 cm and 8.0 cm respectively.

A second experiment was then set up in the same compartments, stocking 50 fish (sexed males) into each, instead of proportionately to bottom area as before. Survival after four months (when the experiment was terminated) was 40%, 84% and 90% respectively, an unfortunate discrepancy probably clue to a high predation rate by birds in the largest compartment. Increments after four months were however 8.2 ern, 6.7 cm and 4.6 cm.

The relative uni(ormity of growth in the first experiment where fish density per unit bottom area was approximately the same in all compartments and the marked differences in the second experiment where fish density increased per unit bottom area is, in spite of survival differences, quite striking, and yet all fish in all compartments were in contact with the same total volume of water in the pond.

The experiments are now being repeated on a replicate basis in order to eliminate the variability, but these preliminary results at least show the importance of bottom area to a grazing fish, total water volume not being of the same significance. They also demonstrate the value of monosex growth as a research method in eliminating density variations due to breeding.

36

APPENDIX E

RIVER FISH MICRATION

by V. D. van Someren

For the last four years an inclined grid trap has been maintained on the Ragati River which flows through the Sagana Fishery Research Station in Kenya. For a number of reasons, mainly financial, it was not possible to make this trap operable at full efflciency throughout the flood period of every year since it was constructed in 1958, nor has the part for trapping upstream migrants heen entirely successful. However it has now been possible to collate all data collected so far on the migration of fish at this trap, and the full results are being published elsewhere. What follows is a condensed summary of the results.

The hydrological regime and fish fauna of this river have been described elsewhere (van Somercn, E. Afr. Agric. For. J. 24(4), April, 1959). Briefly, like most east of Hift rivers in Kenya, the Ragati is subject to twice yearly floods, resulting from higher catchment rainfall in the months of April, May, June (the "long" rains) and in November, December (tlie "short" rains). Intervening months are those of low rainfall on the catchment, and low water flows. The highest floods are those of April to June, their peak period varying within 2-3 weeks each way of mid-May in the four years under study.

The November/December floods are of shorter duration, with peaks usually less than half the height of the earlier floods, except for those of 1961 which reached hitherto unrecorded levels with the unprecedented1y heavy rainfall all over East Africa at this time. In fact these all-time high floods destroyed the gauge at the trap and records of water height have not been possible after the second week of October.

The fish migration records are usually incomplete for heights over 170 on the gauge, because the water overrides the grid at this height and presum­ably an unknown number of fish are lost to the records.

Attention was concentrated mainly on katadromous fish, because of the inefficiency of the upstream traps, and of these the eels, Anguilla nebulosa labiata, and the cyprinid Labeo cylindricus are by far the most important. Analyses of the results have shown the importance of the early, relative.!y small Hoods supervening after a period of low watcr, particularly in the early months of the year. These result in large downstream movements of cels on their spawning migration, thoug-h practically none of them show evidence of "silvering" or gonad maturation (the site is about 400 miles by river from the sea). There is unfortunately no evidence of how m~lI1Y eels travel at the peak of the floods because thcy, in particular, can slide over the inefficient trap, though it is believed that the bulk of them travel in the early rising water, the migration decreasing as the river rises further. .1\11 arc trapped at night only, and there is a subsidiary migration in the short rain floods, curiously enough not in the early part of the flood period, but later in its duration-giving the impression that eel migration probably occurs mainly in the period December to April, in two successive waves at this particular point in the Tana River system. Lower down the timing may well be different.

37

Figures for the years under study are as follows :­

Year Ibs. No. Average Wt.

1958 57.3 33 1.7

1959 196.2 96 2.1

1960 272.2 120 2.3

1961 347.3 117 2.9

(to August)

Trapping was incomplete in the early part of 1958; the trap was under repair from April 1959 to March 1960 because of high water, and the trap was again broken from October, 1961, so that too much importance cannot be attributed to the variation shown above. But the weight of eels which can be caught out of a very small river like the Ragati, even with an inferior trap of this nature, is the feature of importance, for they are highly valuable nutritious food.

The downstream migration of Labeo follows a slightly different pattern. Not so many travel in the early floods, the bulk of the fish travelling at the peak period or shortly afterwards. Those travelling at the peak periods tend to be spent adults, which have travelled up for spawning in the early floods and are later probably swept down as spent fish, while those trapped in the receding floods are usually small immature fish of about 15 cm, probably progeny resulting fr~m spawning in previous years higher up the river.

All La-beo and Barbus passing the trap in each direction have been tagged and released alive in the direction of travel, hence it has not been possible either to sex them fully or determine gonad state, but the tagging results have resulted in a number of recoveries which tend to show a some­what random movement compared with the unidirectional movement of eels. Thus some fish trapped down have been released below the trap and have been caught in the downstream trap again, showing they must have travelled up again in the short intervals between capture and recapture up to 70 days. Unfortunately the inefficiency of the upstream trap has precluded following these movements more closely.

The recapture percentage has been less than 1%, probably because many fish have not been trapped at very high flood levels, but the recovery data have shown an apparently very slow growth in this species 0 f Labeo, even in the smaller immature fish which might be expected to show noticeable increments, and the pattern is very variable, as the following table shows.

r

38

Labeo c')dindricus: Growth and Recapture Data, Ragati River I

TaR No. Date Length Date Days Increment Remarks Tagged at tagging Recapture free cms

cms I

71/G 7.7.58 20.7 15.9.58 70 0.0 Male

.;.4.

7/G 20/G 40/G 77/G

19.8.58 4.3.58

25.4.58 12.7.58

23.0 29.0 30.0 23.0

23.10.58 23.3.58 15.4.59

23.10.58

65 8

355 103

1.0 0.8( ?) 0.0 0.9

immature.

Female npe.

865/G 11.8.60 20.0 4.2.61 178 1.5 Female 374/G 263/G 369/G

14.11.60 16.7.60 23.9.60

17.0 24.4 16.9

8.2.61 19.5.61 2.6.61

87 297 252

0.0 0.1 2.7

immature.

Female 381/G 380/G 437/G

21.11.60 21.11.60

11.4.61

17.7 21.0 24.4

17.5.61 19.9.61 25.4.61

177 302

14

0.3 0.0 0.0)

immature.

Male 1.5.61 6 0.1) spent.

411/G 388/G

8.4.61 6.2.61

20.0 23.7

9.4.61 8.2.61

1 2

0.0 0.0

448/G 469/G 430/G

18.4.61 29.4.61 10.4.61

17.6 23.9 17.4

20.4.61 5.5.61

19.5.61

26

40 .

0.0 0.3 0.0

Female npentng.

505/G 8.6.61 26.4 9.6.61 1 0.0 540/G 6.10.61 11.3 10.10.61 4 0.0 597/G 529/G

14.9.61 26.6.61

19.6 18.0

10.10.61 7.7.61

26 12

0.0 0.0

I have shown elsewhere (van Someren, loco cit.), that in this species, males are mature at about 23 cm, and females at 29 cm, so that marked growth would not be expected after these sizes. The data however is still being analysed.

The very few recaptures of Barbus tanensis have however indicated a faster Rrowth and a more positive form of upstream/downstream movement at flood times (though the total of all caught is remarkably small) and in this species I have also shown (loc. cit.) that males mature at about 30 cm and females at 36 cm, so that most of the fish tagged would still be in the faster growing immature stages. The data is as follows, though again further analysis is still required.

Barbus tanensis: Growth and Recapture Data, Ragati River

Tag No. Date Length Date Days Increment Remarks Tagged at tagging Recapture free cms.

CI11S.

15/G 44/G 311/G

28.2.58 24.4.58 30.3.60

29.0 25.5 33.0

22.11.58 22.11.58 26.1.61 9.4.61

267 212 302 73

3.0 2.1 0.6) 0.4)

Male ripe.

19.6.61 71 1.0) 367/G 446/G

23.9.60 15.4.61

24.0 22.0

21.6.61 16.4.61 13.7.61

271 1

89

4.2 0.0) 0.5)

J

1

39

The day by day analysis of trap records is not yet complete, but the daily catches have shown the usual relationship to the complex of rising­water/falling water temperature/decrease in conductivity/increase in velocity/increase in turbidity, found in other migratory fish studies in Kenya rivers; the influence of any particular one of these is difficult to determine, since they are almost simultaneous changes and interactions in the environ­ment. The one feature common to all downstream migration is that it occurs at night in all three species, probably due to loss of visual fixes in the darkness in the external environment.

,u.

40

APPENDIX F

Report on Limnological Work during a Visit to EAFFRO

betwee.n August .1960, and September 1961,

by Dr. J. F. Tailing

. Work ,has been carried out on several aspects of algal productivity in African lake. waters, in relation to a background of physical and chemical limnology. Routine sampling of the offshore waters of Lake Victoria was maintained as the most detailed approach, but the collection of comparative data from other lake waters was also made. My wi fe has been responsible for the chemical analysis of plant nutrients, including nitrate, phosphate, silica and-in less detail-total phosphorus, iron and manganese.

By September, 1961, the observations on the annual cycle of Lake Victoria were almost completed. Stratified conditions developed gradually between August and November, and continued with a sharp and deep thermal discontinuity, until the end of June. The density of the phytoplankton, as judged by extracted pigments, then reached its lowest value, but increased considerably a fter the transition to a mixed state in July. This increase may have followed a greater availability of nutrients to the surface layers, but the measured concentrations of phosphate and nitrate showed little increase and the small rise observed in silica occurred about a level (4 mgl 1) unlikely to limit production. A subsequent fall in siI-ica was associated with an increased production of diatoms, particularly Melosira nyassensis var. victoriae. Nitrate has continued throughout the year to b~ the nutrient most conspicuously in low concentration, below 10 micro g N0 3 N/1 in the euphotic zone. Near the end of the stratified phase phosphate phosphorus had fallen to concentrations below 5 micro gil. Ammonia nitrogen has never been detected in the surface layers by the crude analytical method used (direct Nesslerisation), but appeared in small amounts in the bottom layer on at least two occasions when the oxygen concentration there fell below 0.5 mg/l. It is remarkable that such low oxygen values were rarely main­tained for many weeks in spite of an apparently sharp--i f small-thermal discontinuity. An indirect communication between the bottom layers and the atmosphere may occur through strong horizontal water movements coupled with tilting, or local breakdown, of the thermal discontinuity. Such horizontal variation of the thermal structure was found during a circular cruise around the lake in May. Water at temperature below 24.6°C was then exposed at the surface along the western shore, although else where such temperatures were only found in the deep layers below a thermal dis­continuity.

The quantitative estimation of individual algal species was continued throughout the year, based on samples from the surface, middle (30m) and bottom (60m) of the lake. Although seasonal (and unseasonal?) changes of cell numbers have occurred, a dominating role has generally been main­tained by colonial blue-green algae, probably several small-celled species of the genus Aphanocapsa. At least five times this dominance has been challenged by less persistent maxima of other species, notably Nitzschia acicularis (January-February and July-August), Melosira nyassensis (July­September), Anabaena fios-aquae (November-January) and !vficrocystis spp. (July). Pronounced seasonal changes of temperature and illumination cannot be the direct causes, but there are possible connections with the seasonal incidence of rainfall and the extent of vertical and horizontal water movements. Conspicuous increases of Melosira, Nitzschia and Microcystis (also several desmids and Entomostraca) followed by the breakdown of thermal stratification in the open lake in late June, and the earlier loss of stratification in some semi-enclosed coastal waters was previously shown by

I

42

wife found in such lakes, often exceeding 500 mg/l, do not seem to have been previously noticed. The density of the phytoplankton generally increase very conspicuously in the more shallow lake waters, as was emphasised by Ross in 1953. In the shallow waters of relatively low alkalinity ( < 11 meq/l) very narrow filamentous species of Myxophyceae often rise to abundance. Examples include the Kavirondo Gulf in Lake Victoria, Lakes Zwei and Awassa in Ethiopia, and in Lake Naivasha (Kenya), there is a most remarkable development (ca. 500,000 filaments/ml) of a very small spirilli­form organism, possibly a species of Spiru/ina. Small tlagellates generally appear to be few in numbers, but recently have been found in some quantity in samples from three lakes-Mutanda, Mulehc and Bunyoni-in Kigezi (W. Uganda).

Lakes Mutanda and Bunyoni have also provided interesting examples of pronounced chemical stratification, favoured by their position in deep and steep sided valleys. In Bunyoni there are indications of meromixis, with a small accumulation of calcium and bicarbonate in a completely deoxygenated layer (15m), rich in ammonia and hydrogen sulphide. Elsewhere we have only studied stratification in any detail in Lakes Victoria and Albert. In the latter it is unexpectedly variable considering the small temperature range involved. A complex hydrology is indicated by strong tilting of the isotherms observed in three longitudinal sections of the lake, with a suggestion of profile-bound density currents linked with periodic deoxygenation. Lake George can show a remarkable diurnal stratiflca!ion of temperature and oxygen, with the latter measured as 216% at the surface and 48% at 1.8 m late one calm afternoon in June. This stratification appears to be a more extreme development of the type previously studied in Gebel Aulia reservoir of the White Nile.

The excellent facilities of the E.A. Freshwater Fisheries Research Organization, Jinja, have been placed freely at our disposal during the year. We are greatly indebted to the Director, Dr. V. D. van Someren, and staff for much additional assistance. In particular Mr. E. HambJyn gave us much help in the work on Lake Albert. I am most grateful to the Colonial Office for making available a CD. & W. grant for supplementary expenses, and to the Council and Director of the Freshwater Biological Association for permission to undertake work in Africa.

-~'.. ­....,..

~-l~ -

43

APPENDIX G

BUOY RELEASE TRIALS

hy J. D. Roberts

Trials with the two types of links supplied by E.A. Industrial Research

Organization, and using sealeel one gallon cans as floats produced results a's

follows ; ­

1.

Site Temp. Float Link Size Submerged l<eleased Time Expired

linja 'Pier.

One gall.

20 20

mm mm

x x

Hours Hours Hours

Lake can 7 mIn. Victoria with 4 mm hole

1 0800 1425 6.25

2 "

0800 1430 6.30

3 0800 1500 7.00

4 "

0800 1520 7.20

5 " 0800 1540 7.40

6 "

0800 1630 8.30

Similar release gear was used on gill-net fleets set in Napoleon Gulf, and

gave adequate cover between setting at 1800-1830 hours and the early hours

of the following morning. It \vas felt however that the minimum time of

release might be inadequate to CO\'er early settings or leave the buoys exposed

during periods of hl'ight moonlight. The ideal would be to make the links

in varying sizes to cover the fishermens' requirements. This is only possible

if the glue used is always of the same consistency; buying sheets 0 f carpen­

ters' glue is not the answer, as the following experiments indicate.

Gear similar to that used by A frican fishermen was used for these trials.

I.e. Ambatch floats, coil' rope and stones for sinkers. To lengthen the life

of the links and simplify measurements, thc squares wcre madc 1" x 1"

with a 3/16" hole.

----

44

2. Links made from locally purchased glue of Dutch manufacture.

Site Temp. Float Link Size Submerged Released Time Expired

~---

Jinja Pier,

25°C -t-

Ambatch I" x I" x 3/16"

Hours Hours Hours

Lake with 3/16" Victoria hole

1 0800 1400 6.00" " "

2 0800 1425 6.25" " "

3 0800 1630 8.30" " "

4 0800 1700 9.00" " "

5 0800 1730 9.30" " " 6 0800 1730 9.30

" " " 7 0800 1800 10.00" " " 8 0800 1800 10.00" " " 9 .. .. 0800 18DO 10.00"

The floats were numbered in the order in which they surfaced as above and used for succeeding trials using the same numbers.

3. Glue purchased locally, manufactured in Holland.

Site Temp. Float Link Size Submerged Released Time Expired

Jinja 25°C Ambatch I" x I" Hours Hours Hours Pier, -t- x 5/16" Lake with ~"

Victoria hole

I 0815 1600 7.45" " " 2 0815 1555 7.40" " " 3 0815 1630 8.15" " " 4 0815 1700 9.00" " " 5 0815 1430 6.15

" " " 6 0815 1545 7.30

" " " 7 0815 1630 8.15

" " " 8 0815 1535 7.20

" 9

" ..

" .. .. 0815 1330 5.15

45

It will be noted that differences in the size and weight of the roughly cut

floats was more than offset by the variation in the links.

4. Glue purchased locally, origin unknown.

Site Temp. Float Link Size Submerged Released Time Expired

Jinja 25°C Ambatch 1" x 1" Hours Hours Hours Pier, -+- x 5/16" Lake with i" Victoria hole

1 0830 1115 2.35

2 0830 1020 1.50

3 0830 1130 3.00"

4 0830 1320 4.50

5 0830 1215 3.45

6 0830 1400 5.30

7 0830 1040 2.10"

8 0830 1205 3.35

9 0830 0920 0.50"

Failure in every case was due to string cutting through the link.

5. Glue purchased locally, manufactured in Holland.

Site Temp. Float Link Size Submerged Released Time Expired

Jinja Pier,

25°C -+-

Ambatch I" X I" x i" with

Hours Hours Hours

Lake 3/16" whole Victoria

At 2000 hours all buoys were still sub­

1-9 0915 merged. Surfacing took place between 2000 hours and 0630 hours on following morning, i.e. 12~ to 211 hours submerged.

Similar links were used on two occasions on nets set in the Buvuma Channel. These were set at 1630 hours. In each case only one link had parted at 0700 hours the following morning, the other link in each case remaining intact.

46

APPENDIX H

A NOTE ON LAKE RUDOLF

by E. L. Hamblyn

In January a further expedition was made to Lake Rudolf with the help and cooperation of the Kenya Department of Game Forests and Fisheries. Four days were spent at Loyongalani and experimental fishing was done in EI Molo Bay a few miles to the north of the camp.

EI Molo Bay is situated (Lat. 2° 46' N. Long. 36° 39' E) on the eastern shore of the lake nearly opposite South Island. The bay is an inlet guarded by a rocky island in line with the lake shore with gently sloping but limited strands on all sides but to seaward. To the north and south belts of emergent vegetation occupy the shore, while off the eastern strand is a raised bank c,.rrying reeds. The bay is the home of the El Molo tribe who eke out a precarious existence which depends on cattle and primiti ve fishing. The area inland is semi-desert and desert, as is the shoreline to the north and south. The cattle are dependent on the emergent vegetation of the bay which in turn depends on the weather conditions prevailing in the area. During the whole of our visit a strong wind estimated at Force 3-4 blew from the south west passing over the open lake. This wind is said to be typical of this area, blowing most of the season through and making the lake un­navigable for the primitive dual hull canoes of reeds used by the El Molo. The lake is dangerous to navigation for small launches and places of shelter are few. These considerations alone will militate against the establishment of a lake fishery and must be taken into account along with the isolation of this water.

The direct result of the wind which varies in strength diurnally is to catr~e tidal effects on the bay shoreline, which may move an estimated forty yards according to the wind strength. This has direct consequences for the fish fauna for which the vegetation belt is a nursery for some species and hahitat to others. No matter how abundant the vegetation the wind force i.s sufficient to ensure high oxygenation at all times. In addition the cattle contribute greatly to the fertility of the water hy direct manuring of the vegetation. This vegetation belt is the most important source of food for them. At "high water" the animals graze up to their hocks in water while at low water they are almost dry shod. -:.,­

Fingerling fish occur here in enormous numbers. One haul of a 12 ft. -,~

mosquito seine behind a small spit formed by wind and current action along the south shore produced hundreds of adult Hap!ochromis rudo!fianus, some of which carried eggs in their mouths, together with fingerling Ti!apia zillii and T. ni!otiw. A few Sari!ius n£loticus and Engrau!icypris stellae were also present. The HalJ!ochromis specimens which were measured varied from 4.1 cm to 8.6 em total length.

The south entrance to the bay is flanked by rocks to seaward and a pebble ;;trand giving way to sharp volcanic gravel along the shore line. A deep weed bed begins off the shore after the main force of the open lake water is spent and runs parallel to the shore in a widening belt about five vards off the shoreline. Presumably the weeds cannot survive the dessication of "low water" as the substratum seems quite suitable for them. The area about the weed hed was the scene of experimental gill net fishing. The results of these sets are shown in Table 1 together with the results recorded from other gear with the exception of the l1lo;;quito seine.

---=- tLc

47

Fishing Date Time Gear Catch 1. 11.1.CJl Overnight Gill net fleet 5 Tilapia nitotica

5" x 1, 3~" x 1, 2~" x 1, 2" x 1. Danish eel trap. Nil.

2. 12.1.61 Overnight Gill net fleet. 189 S;/tlOdontis schall

3. 14.1.61 1000 hours 50 yard se1l1c. 1 Hydrocyon forskali 3 Lates niloticus many H aplochrorni.\ many Tilapia (small)

4. 14.1.61 Overnight S" gill net. 2 Labeo lorii 11 Synodontis schalt

5. 15.1.61 Overnight 5" gill net. Danish eel trap.

1 Labeo lorii 20 S. schall 3 T. galileae

TABLE I: Record of Fishing and JI/umbers of Fish Caught at El M 010 Bay

The catches reflect the environment fished. In Fishing 1 the nets were set along- the open water side of the weeds parallel to the shore in about 6 ft. of water. This set caught the only Lates and T. (lilotica caught by gill nets. This contrasts with the large numbers of S, schall caught in all other gill net sets which were fished in the open water at about 20 ft. at \'arious angles to the shore. The Danish eel tmp was not very successful in the places tried, while the seine net used to collect rates from the weed bed was most inefl'icient due to the foul ground. These observations have implications for the development \!f fisheries in this lake. There are no suitable grounds for beach seines fished commerciall v in the area and there are serious dis­advantages to the use of light twi;le gill nets witbin the bay. It wiU he noticed that the gill net fleet Wi;lS not fished after the large haul of Synodontis caught in Fishing 2. These Machochid fish are dangerously armed with serrated pectoral spines which may be locked at right angles to the body. This behaviour, coupled with the strong dorsal spine, causes them to become almost inextricably entangled in a gill net which is ruined by a heavy catch. Thus any bottom gilL net fishery where these fish are abundant necessitates the use of heavy ply nets. The situation i. similar to the deepwater fisheries of Lake Albert where the most highly d veloped fishery of the Greeks of the former B Igian Congo employed heavy twine and 7i" mesh nets. The use of this gear presupposes availability of considerable capital.

The species caught were:­�Synodontis schall (Block .chneider)� H ydrocyon forslwli Cm'ier� Labeo lorii Heckel� Barilius niloticus (Joannis)� E ngraulic}' pris stellae \;\1orthington� Titapia nito/ira (Linne)� Tilapia =illii (Gervais)� Tilapia galileae (Artedi)� Haplochrorn.is rltdot fianus Trewavas� I.ates niloticus�