UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Analysis of growth and mortality from daily growth increments in the otoliths of dagaa (Rastrineobola.. Njiru, Murithi 1995

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata

Download

Media
ubc_1995-0389.pdf [ 4.88MB ]
[if-you-see-this-DO-NOT-CLICK]
Metadata
JSON: 1.0074806.json
JSON-LD: 1.0074806+ld.json
RDF/XML (Pretty): 1.0074806.xml
RDF/JSON: 1.0074806+rdf.json
Turtle: 1.0074806+rdf-turtle.txt
N-Triples: 1.0074806+rdf-ntriples.txt
Original Record: 1.0074806 +original-record.json
Full Text
1.0074806.txt
Citation
1.0074806.ris

Full Text

ANALYSIS OF GROWTH AND MORTALITY FROM DAILY GROWTH INCREMENTS IN THE OTOLITHS OF DAGAA (RASTRINEOBOLA ARGENTEA) IN NYANZA GULF, LAKE VICTORIA, KENYA by MURITHI NJTRU B.Sc, Moi University, Kenya, 1989 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES Department of Animal Science We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA July 1995 © Murithi Njiru, 1995 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis "for scholarly purposes may be granted by the head of my department or by his or her representatives. It is • understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of The University of British Columbia Vancouver, Canada Date DE-6 (2/88) Abstract Prior to 1960, Rastrineobola argentea was of little economic importance in terms of catches in Lake Victoria. Catches have increased in the last 15-20 years and it is now become the second most important commercially targeted fish species. Growth and mortality parameters of R. argentea, a tropical cyprinid, were estimated using growth increments in otoliths and length-frequency analysis to gather more biological data on the species. The Gompertz growth curve yielded the best fit for the juvenile population. Growth and population parameters for the commercial catch show a growth rate coefficient (K) of 1.8 yr"1 with LOT of 5.0 cm standard length (SL) in Nyanza Gulf and K of 1.5 yr"1 with Lx of 6.5 cm SL in the open waters site. Instantaneous growth rates decreased with age, with fish from open waters showing a more gradual decline. The weight (g)-length (mm) relationship is W = 5.562L33. Fish immersed in 600 mg/L of tetracycline hydrochloride showed its incorporation within 12 hours and increments were likely formed daily. Juvenile fish mortality ranged from 11.3 to 29.9 yr"1. Total mortality (Z) for adults estimated from length-converted catch curves was 4.0 and 4.8 yr"1 for Nyanza Gulf and the open waters respectively. Fishing mortality (F) estimated from catch and biomass was 0.98 yr_1 for Nyanza Gulf, while that of open waters from length-converted catch curve was 1.4 yr"1. The exploitation rate is 0.25 and 0.29 for Nyanza Gulf and open waters respectively. Two annual breeding peaks were observed in both Nyanza Gulf (May/ October) and open waters (May/November). Length and age at recruitment for L50% for Nyanza Gulf was 11.6 mm and 34.5 mm in open waters, corresponding to age of 44 and 175 days respectively. In comparison with published data on the growth and mortality of some small pelagics of African waters, R. argentea had low values of L^,, K, M and Z. n TABLE OF CONTENTS Contents Page Abstract ii Table of Contents iiList of Figures v List of Tables vii Acknowledgements vii1 INTRODUCTION The indigenous fishery of Lake Victoria 1 The Rastrineobola argentea fishery 2 Distribution, local names and nomenclature 3 Food and feeding habits 4 ReproductionWhy this project was chosen 5 Why use otoliths?The purpose of the study 7 Study area2. MATERIAL AND METHODS Samples collection 8 Choice of otolith 10 Otolith preparation , 10 Growth models 2 Growth performance 3 Powell-Wetherall's plotInstantaneous growth rate 14 Age validation. 5 Mortality 17 Age-catch curve.Length-converted catch curve 18 Natural mortality 9 Fishing mortality and biomassRecruitment pattern 20 Estimation of gear selection and probability of capture 2iii Estimation of age at recruitment 20 Relative yield-per-recruit analysis 1 3. RESULTS Growth 23 Size distributionAge distributionOtolith morphology 4 Growth models 3Growth performance 7 Instantaneous growth rateAge validation 43 Age-catch curveLength-converted catch curve 47 Natural mortalityFishing mortality and biomass 8 Recruitment pattern 53 Estimation of selection ogiveEstimation of age at recruitment 5Relative yield-per-recruit analysis 9 4. DISCUSSION 63 5. IMPLICATION OF STUDY FOR THE RASTRINEOBOLA FISHERY 71 REFERENCES 74 APPENDIX 8iv List of Figures Figure 1. Map showing the Nyanza Gulf of Lake Victoria, the inshore sampling site A and the open waters sampling site B 9 Figure 2. Length-frequency distribution of juvenile R. argentea from Nyanza Gulf of Lake Victoria 25 Figure 3. Length-frequency distribution of juvenile R. argentea from open waters of Lake Victoria 6 Figure 4. Length-frequency distribution of R. argentea commercial catch from Nyanza Gulf 27 Figure 5. Length-frequency distribution of R. argentea commercial catch from open waters 8 Figure 6. Age-frequency diagrams of R. argentea from Nyanza Gulf 29 Figure 7. Age-frequency diagrams of R. argentea from open waters 30 Figure 8. Light microscope photographs of R. argentea lapilli 31 Figure 9. Scanning electron micrographs of R. argentea lapilli showing daily increments 32 Figure 10. Relationship between standard length and number of increments on lapillus of R. argentea from Nyanza Gulf with fitted growth curves 34 Figure 11. Relationship between standard length and number of increments on lapillus of R. argentea from open waters with fitted growth curves 35 Figure 12. Powell-WetheraH's plot for R. argentea from Nyanza Gulf and open waters 38 Figure 13. Growth curve estimate for R. argentea using fixed Loo and fitted von Bertalanffy growth curve 39 Figure 14. Weight - length relationship of R. argentea 41 Figure 15. Relationship between logarithm of weight and against natural logarithm length of R. argentea 4v Figure 16. Instantaneous growth of R. argentea from Nyanza Gulf and open waters 42 Figure 17. Photographs of lapillus of R. argentea immersed in 600 mg/L TC 45 Figure 18. Age-catch curve of R. argentea from Nyanza Gulf 49 Figure 19. Age-catch curve of R. argentea from open waters 50 Figure 20. Length converted-catch curve of R. argentea 1 Figure 21. Recruitment pattern of R. argentea 54 Figure 22. Estimation of selection ogive from length-converted catch curve 56 Figure 23. Selection curve of R. argentea 57 Figure 24. Relative yield-per-recruit as a function of exploitation rate 60 Figure 25. Relative yield-per-recruit when assuming knife-edge and selection ogive 61 vi list of Tables Table 1. Growth parameters, von Bertalanffy, Gompertz, Linear models fitted to Nyanza Gulf and open water data of length-at-age 36 Table 2. Growth parameters ( K yr"1, t0 yr) and phi prime (())) estimates with fixed LM 40 Table 3. Concentration of tetracycline hydrochloride aclministered to R. argentea, exposure time and survival time in rearing tank 44 Table 4. Growth parameters (L^,, K) and phi prime of R. argentea compared with other published data 46 Table 5. Growth parameters (L^, K) and phi prime of R. argentea when increments are assumed to be deposited daily, twice a day and once every three days in Nyanza GulfTable 6. Mortality estimated (Z) of juvenile R. argentea, 95 % confidence interval (95% C.I.) and coefficient of determination (r2) from age-catch curve 52 Table 7. Total (Z) of adult R. argentea, 95% confidence interval (95% C.I.) and coefficient of determination (r2) from length-converted catch curve 52 Table 8. Month and percent of recruitment of R. argentea 55 Table 9. Probability of capture of R. argentea 55 Table 10. Estimated age and length of R. argentea at 25, 50 and 75% recruitment 60 Table 11. Maximum exploitation (Efflax) when assuming knife-edge and using selection ogive 62 vn Acknowledgements I wish, to express my deep appreciation and thanks to my supervisor Dr. T.J. Pitcher, for his guidance and encouragement in this study as well as his critical comments on the manuscript. My deepest thanks to Dr. D. Pauly who provided me with relevant literature for this study, and gave advice and criticism. Thanks to Dr. J.D. McPhail, J.O. Manyala and Mr Mwangi for generously providing their time, equipments and laboratory space. I cannot forget Drs. G. Iwama and T. Sinclair for providing initial direction and encouragement. I am indebted to Canadian International Development Agency (CIDA) and the Government of Kenya for providing the funding. The entire staff of Kenyan High commission (Ottawa) did a wonderful job coordinating the whole programm and especially Mr Ngoma and Mr Labatt. I would also like to thank Dr. E. Okemwa, Kenya Marine and Fisheries Research Institute, and the University of British Columbia for giving me this opportunity to do my post graduate study. My sincerest thanks goes to my wife Francisca, and my mother for their love, support and encouragement throughout this study. It will be a human oversight to forget my comrades in arms who helped me in several ways, A. Bundy, T. Hutton, M, Essen, R. Bonfil and R. Chuenpagdee. A.Tautz and G. Langton of the Fisheries Centre, University of British Columbia did marvellous job in coordinating the programm. Last but not least special thanks goes to M.Weiss and Elaine for helping with light and scanning microscope and processing of photographs. Its almost over. Vlll 1. INTRODUCTION The indigenous fishery of Lake Victoria The earliest surveys of Lake Victoria conducted by Worthington (1929) and Graham (1929) showed that the native multispecies fauna of the lake were dominated by haplochromine fishes with more than 500 species. The most abundant and highly cherished fish species were Oreochromis esculentus (Tilapia) and Bagrus docmac (catfish). Excellent catches were obtained using simple gears with little fishing effort (Kudhongania et al., 1992). With the introduction of synthetic fibre gill nets in the 1940s yields increased temporarily. In the late 1950s it was unprofitable to use the recommended 127 mm mesh nets and so fishermen used smaller gill mesh size. This lead to overexploitation of immature fish and endangered the recruitment process. With the decline of the more valued species, exploitation of haplochromines was intensified. Beach seine use was increased and haplochromine catch increased from 19% in 1958 to 37% in 1970 (Kudhongania et al., 1992). Unfortunately, the beach seines damaged some haplochromine and tilapiine stocks especially the eggs, fry, breeding and nursery areas (Welcomme, 1964). In the 1970s the fishery was still dominated by haplochromines and trawling in 1971 showed that traditional commercial fish species were continuously declining (Kudhongania and Cordone, 1974). Nile perch {hates niloticus) were introduced in the early 1960s to convert the bony, "trash" haplochromines into Nile perch flesh which started to dominate the catch by the early 1980s (Kudhongania et al., 1992). Since the 1980s Lake Victoria fisheries are dominated by two introduced species L. niloticus, O. niloticus and by the indigenous Rastrineobola argentea (CIFA, 1988). 1 The Rastrineobola argentea fishery Prior to the 1960s R. argentea was of little economic importance, forming insignificant proportions of fish landed from Lake Victoria (Chitamwemba, 1992; FAO, 1992). Catches of R. argentea have undergone explosive changes in the last 15-20 years in Lake Victoria (Manyala et al., 1992). In the Kenyan waters portion of the Lake, R. argentea landings increased to 30 % of the total fish landing by weight in 1985 as compared to 4.5% in 1969 (CIFA,1988), making it the second commercially important fish after Nile perch (FAO, 1992). Recent figures indicate an increase to 38.5 % in 1986 for R. argentea, a decline in Nile perch from 62.3% to 54.4 % and Nile tilapia from 2.4 to 1.7% (Asila et al, 1991). In 1991 R. argentea contributed 31%, Nile perch 31%, and Nile tilapia 15% of the total fishing landing by weight (FAO, 1992). Annual catches for dagaa were 0.24 t.km"2 yr"1 in 1971-1972 and 4.23 t.km"2 yr"1 in 198985-1986 (Morearf al, 1993), rising to 8.2 t.km"2 yr"1 in 1987 and were 24.3 t.km"2 yr"1 in 1989 (Manyala et al, 1992). The Rastrineobola fishery is currently in the inshore areas and around the numerous islands of Lake Victoria. Wanink (1989) reports an increasing biomass and mean size with increase in depth up to a maximum between 10-20 meters in the Mwanza Gulf in Tanzania. Methods used to capture R. argentea may or may not utilize "light fishing" (Manyala et al, 1992). Fishing by light attraction is done on moonless nights, where kerosene pressure lamps are used to concentrate the fish (Mous et al, 1991; Chitamwemba, 1992). Lamps are anchored with a sinker and, after some time, they are moved slowly towards the beach, bringing the fish within the reach of nets. The fish are either scooped by lift nets or towed to the beach by a beach/mosquito seine. In the scoop net fishery, the lamps are hauled close to the canoe after attracting R. argentea, then the fish are scooped with hand nets into the canoe. Beach seines up to 100 m made of nylon 2 and having a stretched mesh size of 4-12 mm are used in Kenyan waters (Manyala et al., 1992). In Ugandan waters two mesh size nets of 10 and 5 mm are used. Smaller mesh size although preferred by fishermen captures immature fishes of R. argentea and non-target species of O. niloticus and L. niloticus (Ogutu-Ohwayo et ai, 1988). R. argentea is processed by sun drying on the beaches (Chitamwemba, 1992), and a small percent (20%) is sold fresh (Katunzi, 1992). The piscivorous Nile perch introduced in the 1960s has produced radical changes in Lake Victoria's ecology. The perch is believed to have eliminated most of the haplochromine cichlids, and now R. argentea contributes 10-20 % of its diet (Ogari and Dadzie, 1988). Over-fishing and fishing with the wrong gears are believed to be among other factors affecting R. argentea population structure. The mean size and age at maturity of fish now caught in Ugandan and Kenyan waters have decreased and this has posed a question of the future of the Rastrineobola fishery in Lake Victoria (Wandera, 1992). Distribution, local names and nomenclature The small cyprinid R. argentea is endemic to Lakes Victoria, Kyoga and Nabugabo in Uganda. R. argentea is locally known as omena in Kenya, dagaa in Tanzania, and mukene in Uganda (Manyala et al., 1992; Wandera, 1992). Dagaa has a short life span of 1-2 years and its total length (TL) rarely exceeds 100 mm (Wanink, 1989). R. argentea was previously placed in the genus, Engraulicypris, until revised by Howes (1980, 1984) and reassigned to the genera Rastrineobola. The following is the species nomenclature. Class: Osteichthyes Family: Cyprinidae Order: Cypriniformes Species: Rastrineobola argentea 3 Food and feeding habits Zooplankton, mainly copepods, form the major diet of R. argentea, although aquatic insect larvae and pupae, mainly of chaoborids and chironomids, are also eaten (Wanink, 1989; Wandera,1992). R. argentea is a visual feeder (Wanink, 1988a, 1989) and Wandera (1992) found that the major feeding time for R. argentea occurred during daylight hours, while the least feeding was at night. An examination of a twenty four hours feeding cycle by Corbet (1961) and Wandera (1992) revealed that, during day time zooplankton formed the main food, while at night insect larvae and /or pupae were the main food. Juveniles (< 30 mm) fed only during the day almost exclusively on copepods and early instar of chironomid larvae (Wandera 1992). The feeding pattern of R. argentea is related to the species distribution over time and space. HEST (1988) indicated that adult R. argentea stay near the bottom of the lake during daytime and move to the surface at night while the juveniles and parasitized adults stay at the surface throughout. Zooplankton undergoes a similar diel migration (Katunzi, 1992). This migration has also been associated with depth-related abiotic factors such as dissolved oxygen in the water column and light penetration (Katunzi, 1992). Okedi (1982) suggests that R. argentea could be the most abundant species in waters less than 10 meters deep in Tanzanian waters. Reproduction Graham (1929) reported that R. argentea spawns in Lake Victoria producing planktonic eggs. Preliminary studies by Okedi (1973) revealed that the species breeds in the months of June, July and August and that fecundity increases with size. Wandera (1992) found that R. argentea breeds throughout the year with peaks after the two rainy seasons April-4 May and August-September, findings which differ to that of Okedi (1973). The periods after rainy seasons in Lake Victoria are associated with the lake's turnover when the lake completely or partially mixes. The subsequent algal bloom provides a lot of food for growing R. argentea larvae. Availability of enough food resources to the species encourages gamete production (Wandera, 1992). It is therefore not surprising that there may be two principal spawning peaks after the rainy season. Females mature at 43-44 mm and males 40-41 mm standard length (SL) and all individuals above 47 mm SL were mature in Ugandan waters. Manyala et al (1992) estimates the fecundity of dagaa at 1350 eggs for specimens of 60 mm total length (TL) and 170 eggs for specimens of 41 mm TL. Why this project was chosen Despite R. argentea's commercial and ecological importance very little biological information on the species is available. R. argentea plays an important role in the lake ecosystem other than being a cheap source of protein for both humans and animal feeds. In the absence of once abundant haplochromine cichlids, R. argentea is the major food of the Nile Perch (Ogari and Dadzie, 1988; Manyala et al., 1992) . It therefore serves as a bridging role in the transfer of energy from invertebrates to higher trophic levels (Wandera, 1992). Available information on R. argentea from the scientific literature is scarce and it is therefore difficult to make reliable assessments of its population dynamics and its fishery potential. It is not possible to understand biological changes that have occurred in the species during the last thirty years due to lack of data from the past (Mannini,1992; Wandera and Wanink, 1995). The available data which can be used for management are the annual catches, catches from different beaches and catch effort (Manyala et al., 1992). The findings from these 5 investigations apply only to small areas and are based on different sampling methods. Unfortunately, this information lacks biological parameters, although annual landings records date back to 1968 in Kenyan waters (Manyala et al., 1992) and in Tanzanian waters to 1979 (Katunzi, 1992). Recent work on R. argentea includes those of Wanink (1989; 1988a), Manyala (1991), Katunzi (1992), Manyala al (1992) Wandera (1992), Wandera and Wanink (1995). R. argentea is a renewable resource and requires sound management policies for mamtaining a sustainable stock. Such steps can only be taken adequately if the biology of the fish is sufficiently known. Why use otoliths The structures which encode age information in fishes are fin rays, vertebrae, opercular bones, scales, statoliths and otoliths (Jones, 1992). Otoliths have several advantages over other hard structures. Unlike other hard parts which get reabsorbed under conditions of food deprivation and severe stress, otoliths continue to grow throughout fish life (Neilson and Geen, 1984; Campana, 1985; Jones, 1992). A further advantage of using otolith, microstructure examination for age determination is that otoliths are often the first calcified structures that appear during the early development of teleosts. In most instances scales which are commonly used do not record daily events, they cannot be used to age fish under one year, they can be lost, regenerated, and deposition ceases to occur at older ages. Therefore, production of daily otolith increments facilitates age determination of bony fishes with potentially higher accuracy than scales. The daily increments consist of a continuous zone with calcium carbonate and a discontinuous zone with proteinaceous material (Amossaini and Pitcher, 1988). Disadvantages of using otoliths are that they entail killing the fish, they can be difficult 6 to read, thus giving false age, they are time consuming, they require training and specialized equipment which may be lacking in most developing countries (Jones, 1992). The purpose of this study The specific aims of this project on Rastrineobola argentea are to: (1) use otolith readings to estimate growth rates (2) validate daily growth increments in the otolith (3) use otolith readings to estimate mortality rates (4) compare growth rate and mortality of fish in the Nyanza Gulf with those in the open lake (5) investigate recruitment Study area The major portion of the Kenya waters of Lake Victoria (Figure 1) is a narrow gulf, known to various authors by several names. The Victoria Nyanza (Graham, 1929), Kavirondo Gulf (Copley, 1953), the Winam Gulf (Okach, 1982) and Nyanza Gulf (Ogari and Dadzie, 1988) all refer to the same place. The Kenya portion of the lake has an irregular shoreline of about 300 km and comprises only 6% (3 755 km2) of the entire lake area of 68 800 km2 (Rabour, 1991). The Nyanza Gulf has an area of approximately 1 920 km2 with a length of about 60 km and a width varying from 6 to 30 km (Manyala et al., 1992). The Gulf lies between latitudes 34° 13' and 43° 52' east and 0° 4' and 0° 32' south of the equator. The Nyanza Gulf is a swallow bay with an average depth of 6-8 m, and a maximum of 43 m at the open waters, and has an elevation of 1136 m above sea water (Rabour, 1991). 7 2. MATERIALS AND METHODS Sample collection Samples for juvenile R. argentea were collected three times per month from Kichmjio (site A Figure 1) using a 1 x 500 um mesh size planktonic net from January to May 1994. Comparison samples were taken once a month from open waters, near Mbita (site B Figure 1) from February to May 1994. The net was tied on a canoe and pulled for 15 minutes. Fish captured were transferred immediately into 95% ethanol for subsequent sorting and analysis. Handling specimens this way minimized shrinkage and physical damage due to net capture (Thorrold, 1988). On landing, the standard length of fish was taken with calipers to the nearest 0.5 millimeters. Commercial catch from landings of randomly selected boats was collected from Kichinjio from January to May 1994 and from open waters in February. Total sample for juvenile fish collected were 925 and 236 for Nyanza Gulf and open water respectively. For the commercial catch , 236 were sampled from Nyanza Gulf and 55 fish from open waters. 8 I Figure 1. Map showing the Nyanza Gulf of Lake Victoria, the sampling A site and the open waters sampling site B. 9 Choice of otolith Like most fish, R. argentea has three pairs of otoliths: sagittae, astericus and lapilli. Of the three types of otolith in R. argentea, the sagitta is the smallest, very weak and brittle which makes it unsuitable for age determination. The astericus is the largest, but has strong lines with interruptions and individual rings intersect with one another without any pattern (Campana and Neilson, 1985; Muth et al., 1988). Studies on cyprinids have found that lapillus and sagitta otolith are formed around the time of hatching (Victor and Brothers, 1981) or they are present in the first 4 days of development (Muth et al., 1988) while the astericus developed in the third week. The lapillus, the second largest otolith, shows clear bands in a concentric pattern, and was used here for aging R. argentea. Otolith preparation Otolith preparation is necessary to enhance the distinction between the incremental and discontinuous zones that comprise each bipartite growth increment (Campana and Neilson 1985; Alhossaini and Pitcher 1988). Grmdings act to reduce refractive effects and increase light transmission through the otolith (Campana and Neilson, 1985). The fish were placed in a drop of water on a microscope slide. The lapilli were then teased out with sharpened needles under a dissecting microscope. The otoliths were then picked and placed on a clean microscope using the tip of a wetted needle. If there was a lot of attached tissues the otolith was immersed in weak solution of sodium hypochlorite for several minutes to dissolve or loosen resistant material (Brothers, 1987). A small drop of crystalbond thermosetting plastic resin (Hall, 1992) was placed on a standard microscope slide and kept liquid on a hot plate. Each lapillus was carefully 10 placed on top of the resin drop. The plastic resin set in 10 seconds after the slide was removed from the hot plate, allowing enough time to position the otolith to the surface of the slide. If necessary, the crystalbond resin was reheated to reposition the lapillus. Each otolith was ground until the nucleus was reached using abrasive paper of 1200 grit and polished with carborundum aluminium 9-10 um and cleaned under running water. Grmding was done on both sides to form a thin section. Grmding both sides required more time but produced a superior quality preparation. Caution was taken not to overgrind the otolith as this process is irreversible. Daily increments were counted for both left and right lapillus under immersion oil using a compound microscope light microscope (LM) at 1000X. Counting was done from anterior to posterior end of the otolith along the most clear path with a hand counter. The hatch check near the nucleus was used as a reference point from which growth increments were counted. For each otolith, counting was done thrice and the mean of the counts was taken as the final count. An otolith was rejected if incremental counts between or within pairs of lapillus differed by more than three. Incremental width were measured in microns under LM at 300x and photographed using a Zeiss III RS photomicroscope. Preparation for scanning electron microscope (SEM) followed the same procedure as for LM up to the polishing stage with carborundum aluminium 9-10 um. Otoliths were then cleaned under deionized water before acid etching. The acid etching is a crucial step because the optimum etching media and time have to be determined by trial and error (Secor et ai, 1992). Solutions of 0.1 M (pH 7.3) and 0.01 M (pH 7.4) of diasodium emylmechaminetetraacetate (EDTA) were applied gently with a micropipet to sections with otoliths. After carefully measured time, which ranged from 5 seconds to 6 minutes, the solution was removed by placing the slide and affixed section in a beaker of deionized 11 water. The slide was gently dabbed dry with wipes, carefully avoiding touching the otolith. After etching the section was carefully removed with a tip of syringe needle from the slide by reheating thermoplastic glue and attaching it to a scanning electron microscope (SEM) stub with double sided sellotape. The sections were coated with gold (100 Angstroms) in a Nanotech Semprep II Sputter coater for 4 minutes and examined at 20 kV under SEM. Incremental width was measured from SEM photographs by reference to a recorded scale bar. Growth models Estimation of a parameterized growth model is considered to be a standard product of otolith micro structure examination (Campana and Jones, 1992). Although growth models vary, the rationale for their preparation is to allow prediction of an expected mean size or growth rate at some age and/or to facilitate comparison of estimated growth with other published estimates (Ricker, 1975; Campana and Jones, 1992). Three models were used to describe the growth of R. argentea : (i) Gompertz model: the generalized equation of the model (Zweifel and Lasker, 1976) is as follows; Lt = LMexp[- exp(-k(t-t0))] Where: Lt = predicted length (mm) at age t, LM= the mean length fish would reach if they were to grow indefinitely, t0 = age at length = 0, k = growth coefficient of dimension 1/t of the Gompertz equation 12 (ii) von Bertalanffy model (VBGF) (Gulland,1983): the generalized equation is Lt= LJl-exp(-K(t-t0 ))] Where: Lt = predicted length (mm) at age t, LOD= the mean length fish would reach if they were to grow indefinitely t„ = age at length = 0, K = the rate at which length tends towards the asymptote of dimension 1/t. (ii) Linear model (Campana and Jones, 1992) Lt = a+bt Where: Lt = predicted length (mm) at age t a = size of fish at age= 0 (Y intercept) b = growth rate (slope of the line) t = time in days The Excel 4.0 solver function was used to fit Gompertz and von Bertalanffy models using least squares which also provided estimates for r2. A type 1 linear regression was used for the linear model. Growth performance Pauly (1979a) found that different stocks of the same species have similar values of phi prime (())'), defined; <|r = Log10 K + 2 * Log10 L^,. Where K and L^, are from Bertalanffy model. Using the commercial sample, growth parameters for population for Nyanza Gulf and open waters site were estimated from which <jr could be calculated. 13 Powell - WemeraU's Plot was calculated using the Powell - Wetherall's plot (Sparre and Venema, 1992) using the equation, Li - L'j = a + bL' Where: L == mean length of fish calculated from a given cutoff length point Lj L'; = the lower class limit from which L{ is calculated a = y intercept b = slope of the line The Lw values thus obtained were used to estimate K and tD using von Bertalanffy growth curve for the fish population from Nyanza Gulf and open waters. Instantaneous growth rate Instantaneous growth rate is calculated by the following formula (Ricker, 1975, Pitcher and Hart, 1982), G = (Ln w2 - Ln w^/ tj- tt Where: Ln Wj= natural logarithm of weight at the start of time interval Ln w2= natural logarithm of weight at the end of time interval t2 & t, = times corresponding to w, and w2 Data from both Nyanza Gulf and open were grouped in sets of ten with a range of 10 days. Mean length (mm) and mean age (days) for each was calculated. Mean length was converted to mean weight (g) using the equation, W = alb 14 where: W = net weight of fish (mg) 1 = standard length (mm) The fitted regression equation is Ln W = b(Ln L) + Ln a. Constants a and b were estimated from the intercept and slope respectively. Age validation using tetracycline hydrochloride To verify that the increments observed in the otoliths of R. argentea were deposited daily, tetracycline hydrochloride (TC) was used. Tetracycline is incorporated into calcium structures of fish during growth (Thorrold, 1988). This can be restricted to one day's increment on the otolith thus enabling an accurate identification of the treatment date (Alhossaini and Pitcher, 1988). Fish were first kept in large holding tanks (500 Litres) with fresh lake water for 2 days to acclimatize. The holding tanks were positioned outside at Kisumu, Kenya, exposed to normal photoperiod but sheltered to avoid drastic temperature fluctuations which might stress the fish.The aim was to provide fish with conditions close to those found in the lake for them to recover from capture and transportation stress. Fish were fed twice daily on aquarium food and wild zooplankton captured with 20 um mesh size from Lake Victoria. Recovery from stress allows fish to resume normal growth again before being exposed to TC. The surviving fish were then divided for TC treatment. Choosing the right concentration treatment and duration is critical for tetracycline (GjC&aeter et al., 1984). Several trials were tried ranging from 100-600 mg/1 TC with exposure time of 2-24 hours. The test solution was prepared using distilled water because TC combines to calcium ions in hard water, hindering the uptake of TC by otolith (Muth et ai, 1988). Addition of 15 TC lowered distilled water pH from 7.3 to 4.3. TC test solution was adjusted to pH 7.8-8 with tris-buffer (Hetler, 1984) which corresponded to the lake water. During treatment fish were not fed. After treatment the fish were returned to 500 litres holding tanks to resume normal growth. 3 tanks each with 2 untreated fish distributed among tanks with treated fish acted as control. Water in the tanks was replaced with fresh lake water three times a week. Removed otoliths were mounted on epoxy-resin (thermoplastic glue) and kept in light proof slide holders to prevent degradation of tetracycline marks (Hall, 1992). TC treated otoliths were viewed under fluorescent ultraviolent light (UV 360 nm) and bright light illumination with a compound microscope. Under UV light an ocular marker was aligned with the fluorescent band on the otolith. Incremental rings were counted under bright light illumination from the marker to the edge of the otoliths to verify if they do correspond to the number of days since the fish was tagged and the time it was killed. Otoliths were photographed using a Zeiss III RS photomicroscope. Age validation using growth performance index Second approach to validation was to compare the values of the growth using the growth performance index ((j)') (Pauly 1979a, Sparre and Venema, 1992). For the same L^, values of K were fitted to the ages and size of the fish under three assumptions about ring deposition rate: 1) one ring per day 2) two rings per day 3) one ring per 3 days For each assumption, was calculated. The three (j)' estimates were then compared with the mean calculated from the literature on this species. This was necessary because 16 number of increments of otolith rings after immersion in TC did not correspond exactly to the number of days the fish survived. Mortality The general equation for mortality is Where: Z = instantaneous rate of mortality Ntl = number of fish at time tt Na= number of fish at time t 2 tx and tj are times 1 and 2 also we have Z = F + M where: Z = the total mortality per year F = fishing mortality per year M = natural mortality per year (Pitcher and Hart 1982; Gulland, 1983) Age-catch curve For the juvenile population fish abundance was converted to the natural logarithm of abundance and plotted against age which was read directly from otoliths. Conversion resulted in a straight line with a negative slope which was fitted through ordinary least square regression (Ricker,1975). Starting data points for the regression were established 17 using Robson and Chapman (1961) criteria (see Appendix 4). The equation used was, Ln N; = a + bt; Where Ln N;= natural logarithm of number of fish of age t;. a = y intercept b = slope of regression line (Z) (Pauly, 1984c) Length-converted catch curve Length frequency data was converted to their corresponding ages by means of set growth parameters (L^ and K of VBGF). Abundance at age decreased exponentially, making the slope an expression of mortality (Essig and Cole, 1986). Total mortality (Z) was estimated as the negative slope of: LnN/At; = a + bt;' LnNj = natural logarithm of number of fish in length class i. At = time fish needs to grow through length class i t'i = is the age corresponding to the midpoint of length class i a = y intercept b = is the regression slope (estimation of Z) Selection of points to be used in the regression followed Pauly's (1984c) rules (Appendix 5 ). 18 Natural mortality Natural mortality (M) was estimated following Pauly's empirical formula (Pauly, 1980a), linking the natural mortality with the von Bertalanffy parameters, K (yr"1), (cm) and mean annual temperature (T °C) of water in which fish stock live. Log 10(M) = -0.0066 - 0.279 log 10 Lx+ 0.6543 log 10 k + 0.463 log 10 T Pauly's empirical formula was modified for schooling fish such as R. argentea (Mannini, 1992) by multiplying M by 0.8 so the estimation becomes 20 % lower (Sparre and Venema, 1992). Fishing mortality and biomass Fish mortality is defined by F = Z - M and can thus be obtained by subtraction, however it was here also necessary to estimate F from F= C/B and P/B = Z (Allen, 1971), because Nyanza Gulf data gave an F of zero, where: F = fishing mortality per year C = catch per year B = the average biomass during the period considered P = production Z = total mortality Biomass was estimated using the same C (4.23 t.km2) and P (17.3 t.km2 yr"1) as in the ECOPATH II model (Moreau et al, 1993), but with a Z of 4.0 yr"1 from this study instead of 2.2 yr"1 . The ECOPATH II model is structured around a system of linear equations which estimates biomass and food consumption of various species of an aquatic system by analysis of flows between the elements of the ecosystems (Christensen and Pauly, 1992). 19 Exploitation rate was estimated by E = F/Z. . Recruitment pattern. The recruitment pattern is estimated from length frequency data by a method which involves; 1) backward projection onto the time axis of a set of length-frequency data; 2) summation of each month of the frequencies projected onto each month; 3) subtraction, from each monthly sum, of the lowest monthly sum to obtain a zero value where apparent recruitment is lowest; and 4) expressing monthly recruitment in percent of annual recruitment (Pauly et al, 1984). Estimation of gear selection and probability of capture When estimating mortality the left hand side of the catch curve is not considered because juveniles are not yet fully exploited or recruited. Backward extrapolation of the catch curve estimates the number of juveniles which ought to have been caught, had it not been for incomplete selection and recruitment. To obtain probabilities of capture, the number of fish in each length class caught are divided by the expected numbers (Pauly et al, 1984). Estimation of age at recruitment Age at recruitment was estimated using inverse von Bertalanffy equation: tf = tn-l/k* ln(l-L/LJ Where: t; = age in years, 20 tc = age at length = 0, K = the rate at which length tends towards the asymptote, L00= the mean length fish would reach if they were to grow indefinitely, Lj = length (mm) at age ti5 (Sparre and Venema, 1992). Relative yield-per-recruit analysis Beverton and Holt (1957) presented an equation requiring only three input parameters, M/K, U (= 1-Lc/LM ), and E (= F/Z) to access the effect of different exploitation rates a fishery. Beverton and Holt's relative yield per recruit model is defined by: (Y/R)' = E*UM/K*(l-3U/l+m+3U2 +U3/l+3m) Where: (Y/R)' = relative yield (g) per recruit m = (1-E)/(M/K) and E=F/Z E= the fraction of deaths caused by fishing (exploitation rate) F = fishing mortality per year Z = total mortality per year M = natural mortality per year K = the rate at which length tends towards the asymptote of dimension 1/t. U = 1-Lc/LM L„ = length (mm) of fish at first capture Lx= the mean length fish would reach if they were to grow indefinitely, (Sparre and Venema, 1992) 21 Plot of (Y/R)' was done by; 1) Assuming knife edge selection The knife-edge selection assumes all fish below length at first capture (Lc) escape through the mesh of the net, while fish above Lc are assumed to be suddenly exposed to full fishing mortality (F) which remains constant for the rest of the cohort life (Pauly, 1994). 2) Selection ogive. The fraction of fish retained in a fishing net depends on mesh size, fish size and their availability. Plot of fraction of fish retained by a net against length gives a sigmoid curve. Retention by a net of young fish is low because of their small sizes and they have not been fully recruited, while very old fish are in low numbers in fishing grounds. The intermediate size is the one mostly retained (Sparre and Venema, 1992). 22 3. RESULTS GROWTH The aims of the growth study were to use ageing of otoliths and length frequency analysis to determine and compare growth parameters of dagaa in the Nyanza Gulf and in the open waters of Lake Victoria. Size distribution Length-frequency diagrams for the Nyanza Gulf and open waters samples are presented in Figures 2 and 3, and from commercial catches Figure 4 and 5. Comparison of the two sets of length frequency data for the juvenile population indicates incomplete recruitment to my sampling gear by fish less than 3 mm and avoidance of sampling gear by fish more than 25 mm. In Nyanza Gulf, there is incomplete recruitment to the commercial gear by fish less than 10 mm and a diminishing of catch of fish more than 40 mm. In open waters, fish of less 29 mm were absent in the commercial catch. Appendix 1 presents catch in numbers for the entire sampling time. Age distribution. Figures 6 and 7 presents age-frequency diagrams for the Nyanza Gulf and open waters and data used is in appendix 2. Nyanza Gulf sample is dominated by fish of ages 20-70 days and lacks fish older than 140 days. In open waters the dominant age groups are 35-70 and 10-25 days in February and March respectively and lacks fish older than 110 days in February and 80 days in March. Nyanza Gulf sample had older juveniles compared to open waters. 23 Otolith morphology When observed under a light microscope (LM) the lapillus showed conspicuous marking arranged in concentric patterns radiating from the nucleus (Figure 8a) and right and left lapillus had equal number of increments. Increments were easily read in most otoliths, only 25 (5.6%) were rejected due to either the error of reading, imprecision being greater than 3 rings (10; 2.2 %) or because the otolith could not be clearly read due to nondaily rings (15; 3.4 %). Counts were done for 326 fish from the Nyanza Gulf and 120 fish from open waters, ranging from 3 mm to 25.4 mm SL, and ages ranging from 8-135 days. Only otoliths etched for 2 minutes or more in 0.1 M EDTA produced recognizable increments on SEM, while application 0.01 M EDTA produced no changes on the surface of otoliths. The best etched lapillus was the one where 0.1 M EDTA was applied for 2 minutes (Figure 9a). Application of 0.1 M EDTA for 5 and 6 minutes over etched the lapilli and only a few increments could be noticed in some areas probably where there was less etching. No marked changes in increment in morphology was evident both under LM and SEM, although in some otoliths a narrowing and subsequent widening of increments occurred in the first few increments near the nucleus (Figure 8a & 9b) and at the edge of older juveniles under (Figure 8b). Subdaily increments (Figure 8b & 9a) which are morphologically similar to daily increment were more easily distinguished in older juveniles. Daily increments appear in a regular sequence, they are more prominent and do not merge with others like subdaily increments. Careful adjustment of the LM focus could correctly interpret daily increments from subdaily. Use of SEM did not alleviate the subdaily increments problem (Figure 9a). 24 i 3 20 -cr £ 0 ] < rr [-1 r-l January ~|_r „ 1—1—1—1—1 1—1 1 •0 CO 05 ! ! T 1 1 t 1 I P'H CM ir> oo »- ^ *— «- CM CM Length(mm) & 60-§ 40 -I 20-Pi 0 - i, r-H February rn~TTTl r1 rr i n i—i i ~")—11 1—!—"— "0 CD 00 —I—I—1—I—!—!—1—T——T—1—PT—!™rn—t CM Ift 00 «- it r- T- T— CM CM Length (mm) Frequency o o o o '.. „ . r March Frequency o o o o 1—i—i—i—i—i—i—\—i—i—i—i—i—i—i—i—\—i—i—i—i—ii—i—i •"3<O05C\im00'-^t-i- <- CM CM Length (mm) Frequency O in O u •— April TL Frequency O in O u cococncMtooo'-'* «— *— CM CM Length (mm) 5> 4T CD 3 2 --cr 2 o i i i i fT May i i < ' 1 i i i i i i i tt, u co I ! 1 1 I .1 CO 03 i i t i i i i i i i i i i i i i i CM Ifl 00 *f r- i- CM CM Length (mm) Figure 2. Length-frequency distribution of juvenile R. argentea from Nyanza Gulf of Lake Victoria. Juvenile were sampled using a 500 um mesh size planktonic net from the shore of Lake Victoria. 25 February 8 20 T 10 l I I l 44- -"IvThTr-n rrfi n CO CD (35 CM un oo <-CM . CM Length (mm) March & 40 3 20 cr u 111 M 1111111111111 00 CD OT CM UO 00 CM CM Length (mm) April £ 6 8 4{ & 2-h 0 •44 -T4- i i i i i i i i i i i i CO CD CO CM in 00 <-T- CM CM Length (mm) May 8 4J §- 2 I i i I i i I i I I I i CO CD cn CM LO 00 CM CM Length (mm) Figure 3. Length-frequency of juvenile R. argentea from open waters of Lake Victoria. Juvenile were sampled using a planktonic net (500 microns) from the offshore of L.Victoria. 26 January cu cr UH 25 20 15 10 5 mi Mill HI n rm n-H O CO CD cn CN LD 00 CN CM CN CN CO CO CO i- (v. -=r -a-Length (mm) February CD cr cu PH 10 T 8 --6 -4 -2 --0 + in co t--H h-ococDajcNmoot-'^N-T-cMCNCMCNcococo'^r'^r'sr Length (mm) March o « <u cr (U 10 8 + 2 + 4-+-4 I ' I 4 44 -=t 1^- o i- T- CM CO CM CO CN CM CN CO IS1 CO 00 CO f+-Length (mm) Figure 4. Length-frequency distribution of R. argentea commercial catch from Nyanza Gulf. Adult fish samples were obtained from randomly selected commercial boats fishing at same area juvenile samples were collected. 27 February o c CU 3 cr cu Uc 10 8 + 0 I I I I I I I l i I I I I I I I I I I I I I l I 4-un CD co CM to CM cn CM CO CO CO in NT It Mil! CO in Length (mm) Figure 5. Length-frequency distribution of R. argentea commercial catch from open waters. Adult fish samples were obtained from randomly selected commercial boats fishing at same area juvenile were collected. 28 January s tu u PH 15 j 10 -5 -oJfl 111 n 1111111111111 CM CO CO CM lO CO CO O 00 o> CO CM CM J2 <D Age (days) February & 15 T [ -irrfM V rfffl 11 rfh ITh i M n n i OOOCDTt-CMOCOCD'^CMOCOCDTrCMO i-TNntminiDNcooioioi-Nn Age (days) March § cr cu )H PH MIHI rHHThn.mlTlflhr 11111111 COCD-tCMOOOCO'VCMOCOCO-^-CMO i-cMco-*inincot-~.coo)cnor-cMco Age (days) Figure 6. Age-frequency of juvenile R. argentea from Nyanza Gulf of Lake Victoria. Juvenile were sampled using a planktonic net (500 microns), age read direct from otoliths and frequency obtained from fish which otoliths had been read. 29 February S3 =s cr 6 T 4 2 0 IIIIIIIII CD CN mil (ill nil! 1111111111II111 •"3" CO CN o LO co m CD CD O oo CO CD o CN CN O CO Age (days) March CL> cr CD 15 10 4-5 ryr 1 o PT! 11111 m n i M n Trl 1111111111111 n1111111111111111111111111111 II i O CO CD T- T- CN ro o LO co LO CD CD 1^-CN co o cn co CD co -=r o T- o CO Age (days) Figure 7. Age-frequency of juvenile R. argentea from open waters of Lake Victoria. Juvenile were sampled using a planktonic net (500 microns), age read direct from otoliths and frequency obtained from fish which otoliths had been read. 30 Figure 8. Lapillus otolith of R. argentea. a) Arrow indicates hatch check for a 42 days old dagaa (10.1mm). Scale bar = 10 um. b) Subdaily increments (arrows) and nucleus (n) for a 52 old dagaa (15.3 mm) At the edge of otolith increments are not clearly visible, refocusing helped solve the problem. Scale bar 10 um. 31 Figure 9 Scanning electron micrographs of R. argentea lapillus showing daily increments. a) Arrow indicates subdaily growth. Scale bar 4 um. b) Nucleus (n) which lacks any increment. Scale bar 20 um. 32 Growth models The plot of the fitted models for length versus the age for juveniles is presented in Figures 10 and 11 for Nyanza Gulf and open waters respectively. The Gompertz growth model yields the best fit to the data in Nyanza Gulf and open waters respectively (r2 = 0.86 & 0.93), although a significance test of the r2 was not done. The r2 value derived from regression and total sums of squares was the highest. The worst fit model for Nyanza Gulf and open waters was von Bertalanffy growth model (r2 = 0.82 & 0.89). The Gompertz growth model produced higher k values and lower asymptotic length in both Nyanza Gulf and open waters compared to von Bertalanffy growth model (Table 1). The fitted equations for the Gompertz growth model were: Lt = 24 exp[- exp(-8.7(t-31.3))] for Nyanza Gulf and, Lt = 60.4 exp[- exp(-3.5(t-95.7))] for open waters. 33 Age(days) Figure 10. Relationship between standard length and number of increments on lapillus ofR. argentea from Nyanza Gulf (dots), together with fitted growth curves (solid line), a) Gompertz growth curve, b) von Bertalanffy growth curve, c) Linear regression. Details in the text. 34 a) 30 j E E 20 -£ Ol 10 -CD 0 - -+- -+-50 100 Age (days) 150 30 j E E 20 -x. at 10 -CD _l 0 - -+-50 100 Age(days) 150 30 j (mm] 20 -.c at 10 -cu -J 0 -50 100 Age(days) 150 Figure 10. Relationship between standard length and number of increments on lapillus of R. argentea from Nyanza Gulf (dots), together with fitted growth curves (solid line), a) Gompertz growth curve, b) von Bertalanffy growth curve, c) Linear regression. Details in the text. 35 Table 1. Growth parameters, von Bertalanffy, Gompertz, and Linear models fitted to Nyanza Gulf and open waters data of length-at-age. The parameters are defined in the text. Model Nyanza Gulf Open waters von Bertalanffy L^ (mm) 35.8 560.6 K (yr1) 2.9 0.13 tQ(days) -7.1 -17.6 r2 0.82 0.89 Gompertz L*, (mm) 24.9 60.4 k (yr-1) 8.7 3.5 t0 (days) 31.3 95.7 r2 0.86 0.93 Linear regression a 3.9 3.3 b 0.17 0.19 r2 0.84 0.92 36 Growth performance Using Powell and Wetherall's plot, Lw of 5.0 cm and 6.5 cm were obtained from the length-frequency data for Nyanza Gulf and open waters sampling sites (Figure 12) and points used for estimation of L^, are shown in appendix 3. The resulting von Bertalanffy growth curves for Nyanza Gulf and open waters are shown in Figure 13 and growth parameters on Table 2. Instantaneous growth rate Plotted weight (g) against length (mm) graph is shown in Figure 14 and the fitted regression to the natural log transform is shown in Figure 15. Equation for regression is; Y = 5.56 2+ 3.29X, and therefore the length-weight relationship is; W = antilog (-5.562)L329. Instantaneous growth (G) rates for Nyanza Gulf and open waters are shown in Figure 16. The graphs show a general decrease in instantaneous growth rate with age. In Nyanza Gulf there is a sharp decline of G to about 55 days, but it then increases slightly to 0.02 at round 99 days. In open waters the decline is gradual and starts at a lower G (0.07), but does not go as low as in Nyanza Gulf. 37 a) POWELL - WETHERALL PLOT 20 o 15 o o 1 M 10 o o o o o 5 0 5 25 Cutoff length (L' ;HM) 45 b) POWELL - WETHERALL PLOT 14 o o 1 7 o M o 0 30 44.25 58.5 Cutoff lensrth (L' ;HH) Figure 12. Powell-Wetherall's plot for R. argentea using commercial samples from, a) Nyanza Gulf ( = 5.0 cm), b) open waters ( = 6.5 cm). ° Point used in analysis o Point not used 38 Figure 13. The von Bertalanffy growth curve for adult R. argentea based on otolith readings and a fixed of, a) 5.0 cm in Nyanza Gulf, b) 6.5 cm in open waters. The first part of the curve (dots) presents age read directly from the otolith. 39 Table 2. Growth parameters (K yr1, t0 days) and phi prime ((j)) estimates with fixed L, (cm) together with their standard error (SE) for R. argentea. Site K SE t„ SE § SE Nyanza 5.0 1.8 0.129 -0.029 0.006 1.7 0.086 Gulf Open 6.5 1.4 0.087 - 0.027 0.005 1.7 0.062 waters 40 Figure 14. Weight (g) and length (mm) relationship of R. argentea from Nyanza Gulf. Ln.length (mm) Figure 15. Relationship between natural logarithm of weight plotted against natural logarithm of length (cicrles) of R. argentea showing a fitted regression line (solid line) 41 Figure 16. Instantaneous growth of R. argentea from Nyanza Gulf and open waters. 42 Age validation using tetracycline Table 3 shows the concentration of tetracycline hydrochloride (TC) used on R. argentea, the number of fish used and the hours of exposure. Only fish exposed to 600 mg/1 TC for 12 hours and 21 hours were marked (Figure 17). The appearance of a second ring at the edge of the lapillus (Figure 17a) is due to a second immersion in 600 mg/1 TC for 12 hours after the fish survived the 30 days. After the second immersion the fish lived for three days. Under high magnification (lOOOx) 30 increments on the lapillus of fish immersion in 600 mg/1 TC for 21 hours were found, suggesting the increments were likely deposited daily. Fish immersed in 600 mg/1 TC for 12 hours had no visible rings under a light microscope. Figure 17b shows a clear check on the lapillus of the same fish exposure to 600 mg/1 TC for 21 hours under bright light illumination. There was higher mortality in fish exposed to low and high concentration (100-200 and 600 mg/1 TC) compared to medium concentration (300-400 mg/1 TC), which could not be accounted for. There was no stress marks (checks) evident for fish immersed in TC, suggesting TC did not affecting fish growth. Age validation using growth performance index Phi prime (Pauly, 1979a) from this study and from other authors has a mean and standard deviation (SD) of 1.65+/-0.084 (Table 4). When increments are assumed to be formed two per day or one per every three days, the estimated phi prime does not lie within this mean +/-SD (Table 5), suggesting that the otolith rings seen were mostly likely deposited daily. 43 Table 3. Concentration of tetracycline hydrochloride administered to R. argentea, exposure time and survival time in holding tanks. TC Fish exposure survival comments (mg/1) no time (hrs) time (days) 100 1 4 15 died* 200 3 2 17 died* 200 1 6 1 died* 200 1 24 26 killed 300 2 18 25 killed 400 2 24 30 killed 500 1 24 6 died* 500 1 24 27 killed 500 1 6 26 killed 500 1 6 8 killed 600 3 21 30 killed+ 600 2 12 30 killed+ 600 1 12 7 died 600 1 6 7 died* * Dead fish with mould covering which I could not establish its origin. One suggestion could be that water was contaminated through fish feed. + Fish which had fluorescent marks under ultraviolet light. Fish were killed to remove otolith for further processing. 44 Figure 17. Lapillus of a 24.1 mm R. argentea immersed in 600 mg/L TC. b) Under ultraviolet light showing tetracyclme fluorescent bands, b) Same as photography (a) under bright illumination. Arrows indicates a check correspondmg to the tetracycline fluorescent band. Scale bar 10 um. 45 Table 4. Growth parameters (LM, K) and phi prime of R. argentea from this study compared with other published data. Source LM(SLcm) K yr"1 Phi prime Wanink ,1989 5.2 1.1 1.5 Wandera, 1992 6.5 0.92 1.6 Wandera and Wanink, 1995 6.5 0.92 1.6 Njiru 1995 (Nyanza Gulf) 5.0 1.8 1.7 Njiru 1995 (Open waters) 6.5 1.4 1.7 Mean+/-SD 1.62-/+0.084 Table 5. Growth parameters (L^,, K) and phi prime of R. argentea when increments assumed to be deposited twice a day and once per three days in Nyanza Gulf. Otolith ring L^cm) K yr"1 Phi prime assumed formation Daily 5.0 1.8 1.7 Twice daily 5.0 3.6 2.0 1 per 3 days 5.0 0.63 1.2 46 Age-catch curve Figure 18 and 19 present age-catch curves (age vs log frequency) for Nyanza Gulf and open waters respectively. All the regression slopes used to calculate total mortality rate (Z) were significantly different from zero ( P < 0.05). The Nyanza Gulf age-catch curves gave a Z of 0.038 day_1(13.9 yr1) in January, 0.031 day-1 (11.3 yr1) in February, and 0.041 day-1 (15.0 yr-1) in March (Table 6). The open waters age-catch curve gives a mortality (Z) of 0.026 day"1 (9.5 yr"1) in February and 0.082 day-1 (29.9 yr"1) in March (Table 6). Length-converted catch curve The length-converted catch curve give a total mortality (Z) of 4.0 yr-1 for Nyanza Gulf and 4.8 yr-1 for open waters (Figure 20). Data used for analysis is shown in appendix 5 and results for total mortality, their 95% confidence interval and coefficient of determination in table 7. The regression slopes were significantly different from zero at P < 0.05. Natural mortality The mean annual surface water temperature from the study area was 25 ° C. Pauly's equation (Pauly, 1980a) gave natural mortality of 4.1 yr"1, 3.4 yr1 and at 20% reduction natural mortality (Sparre and Venema, 1992) of 3.2, 2.7 yr-1 in Nyanza Gulf and open waters respectively. 47 Fishing mortality and biomass Nyanza Gulf data give Z of 4.0 yr"1, M of 4.1 yr"1, and F of - 0.1 yr"1 by subtraction and open water data give Z of 4.8 yr"1, M of 3.4 yr"1 and F of 1.4 yr"1. This F from open waters is compatible with F for Nyanza Gulf that can be calculated from the ECOPATH II model of Lake Victoria (Moreau et al, 1993) where, P = 17.3 t.km"2 yr"1, C = 4.23 t.km2 yr"1, F = 0.54 yr"1, B (P/Z) = 7.9 t.km2, and Z (Z = P/B) is 2.2 yr"1. Redoing estimation of biomass (B = P/Z. Allen, 1971) using Z of 4.0 yr"1 with the same catch (C ) and production (P), biomass becomes 4.3 t.km2 and estimation of F (F = C/B. Allen, 1971) has 0.98 yr"1. Exploitation rate (E = F/Z) is 0.29 for open waters and 0.25 for Nyanza Gulf. Appendix 6 presents a summary for estimates of Z, M, F and E from age-catch curve, length-converted catch curve and ECOPATH II model. 48 January c 3 cr CD C 5.000 j 4.000 I 3.000 2.000 1.000 0.000 o o 20 40 60 Age days 80 100 February Figure 18. Age catch curve of R. argentea from Nyanza Gulf. Ages were obtained by reading otoliths as outlined in chapter 2 and frequencies from the samples of fish which had been read. • Points included in analysis • Points not included Points used were established using Robson and Chapman (1961) criteria (Appendix 4). 49 February 3 - • §2.5-§ 2 - • S'I-S - ^^^^^ • i±3 1 - • CO.5 -o c I I I I I 20 40 60 80 100 Age (days) March 5 T 0 10 20 30 40 Age (days) Figure 19. Age catch curve ofR. argentea from open waters. Ages were obtained by reading otoliths as outlined in chapter 2 and frequencies from the samples of fish which had been read. • Points included in analysis • Points not included Points used were established using Robson and Chapman (1961) criteria (Appendix 4). 50 Figure 20. Length converted-catch curve of R. argentea from, a) Nyanza Gulf, b) open waters. Length was converted to age by inverse von Bertalanffy equation as outlined in text. '» Points included in analysis o Points not included Points to be used were established using Pauly 1984c method (Appendix 4). 51 Table 6. Mortality estimated (Z) of juvenile R. argentea, 95 % confidence interval (95% C.I.) and coefficient of determination (r2) from age-catch curve. Site Z (yr1) Z (day1) 95% C.I r2 Nyanza January 13.8 0.038 0.019-0.057 0.84 February 11.3 0.031 0.018-0.044 0:85 March 15.0 0.041 0.0083-0.073 0.75 Average 13.4 0.037 Open waters February 9.5 0.026 0.0095-0.042 0.83 March 29.9 0.082 0.0412-0.123 0.99 Average 19.7 0.054 Table 7. Total mortality (Z) of adult R. argentea, 95% confidence interval (95% C.I.) and coefficient of determination (r2) from length-converted catch curve. Site Z 95% CI r2 Nyanza 4.0 3.45-4.62 0.95 Open waters 4.8 1.4-8.1 0.72 52 Recruitment pattern As indicated by the analysis R. argentea recruitment goes on throughout the year with two peaks in both Nyanza Gulf and open waters (Figure 21). In Nyanza Gulf there is one prolonged major recruitment period from January to September with a peak around May, and the minor starts in September to November peaking in late October. In open waters the two peaks occur in similar months like those in Nyanza Gulf, with peaks in May and late November. Months with high recruitment in Nyanza Gulf are January, May, September and November and in the open waters January, June and December (Table 8). Estimation of selection and probability of capture Figures 22 shows estimation of selection ogive from length-converted catch curve by extrapolation of juvenile total mortality. Figure 23 and table 10 shows 25 % of the R. argentea enters the fishery at 8.9 mm, 50 % at 11.6 mm and 75 % at 14.0 mm in Nyanza Gulf. In open waters, 25% of R. argentea enters the fishery at 32.1 mm, 50 % at 34.5 mm and 75 % at 36.7 mm. Estimation of age at recruitment Estimation of age at recruitment using t0 of - 0.029 and - 0.027 for Nyanza Gulf and open waters respectively are presented in table 10. In Nyanza Gulf 25% of R. argentea enters the fishery at 29 days, 50% at 44 days and 75 % at 54 days. In open waters 25% of R. argentea enters the fishery at 160 days, 50% at 175 days and 75 % at 194 days. 53 a) 30 r Loo = 50, K = 1.8, C = 0, MP = 0, t0 =-.029 b) 20 r Sep Oct Nov Dec JTan Feb MAJP A.pr> May Jun Jul Loo =65, K = 1.4, C = 0, WP = 0, t0 =-.027 I 1 ! Figure 21. Recruitment pattern of R. argentea from, a) Nyanza Gulf, b) open waters. 54 Table 8. Month and percent of recruitment of R. argentea. Absolute time Percent Recruitment (months) Nyanza Gulf Open waters January 10.0 9.3 February 1.2 5.6 March 6.1 8.2 April 8.3 7.7 May 9.9 9.1 June 9.4 17.2 July 8.2 9.4 August 7.8 0.0 September 13.5 7.6 October 3.4 7.6 November 22.3 9.0 December 0.0 . 9.2 Table 9. Probability of capture of R. argentea. Probability (%) of Length capture (mm) Nyanza Gulf Open waters 25 8.9 32.1 50 11.6 34.5 75 14.0 36.7 55 a) b) Figure 22. Estimation of selection ogive from length-converted catch curve for R. argentea from, a) Nyanza Gulf, b) open waters. o; Points not used * Points used a Points showing extrapolation of juvenile total mortality (Z) from adult Z 56 a) Length classes (nn) b) Length classes (MM) Figure 23. Selection curve of R. argentea from, a) Nyanza Gulf, b) open waters. Details in the text. 57 Table 10. Estimated age and length of R. argentea at 25, 50 and 75% recruitment Recruitment (%) Length (mm) Age (years) Nyanza Gulf 25 8.9 0.029 (29 days) 50 ' 11.6 0.12 (44 days) 75 14.0 0.15 (54 days) Open waters 25 32.1 0.43 (160 days) 50 34.5 0.48 (175 days) 75 36.7 0.53 (194 days) 58 Relative yield-per-recruit model From the estimated parameters of L^,, K, tQ and M the yield per recruit was calculated for different exploitation rates (E). Figure 24 and table 11 shows maximum exploitation rates (Emax ) using different length at first capture (Lc) increases as Lc increases. When Lc is 12 mm in Nyanza Gulf, Emax which can sustain the fishery is at 0.48 and at Lc of 20 mm, Emax is at 0.67. In open waters at Lc of 32 mm, Emax is 0.83 and Lc of 40 mm has Eraax of 1. Plot of relative-yield per-recruit against exploitation rate when assuming knife edge selection gives a higher Emax than in case of selection ogive (Figure 25 and Table 1 lb). 59 Lc = 20 mm Lc = 17 mm Lc =12 mm Exploitation Figure 24. Relative yield-per-recruit as a ruction of exploitation rate for three values of mean length at first capture (Lc), a) Nyanza Gulf, b) open waters. 60 a) 2 r Hi •4 v 1.5 *» •*« (• s tl 5 .3 91 < / CO a 5 e B - .23 3 .73 Exploitation rate I Exploitation rate Finure 25 Relative y,eld-per-recruit showing the deference in exploration rate when a^Lg ta^e-edge (A) and usmg selection ogtve (B) for, a) Nyanza Gulf, b) open waters. 61 Table 11. Maximum exploitation (Emax) for R. argentea, a) assuming knife-edge Lc (mm) Lc/Loo F max Nyanza Gulf 12 17 20 0.24 0.34 0.40 0.48 0.59 0.67 Open waters 32 40 43 0.49 0.54 0.62 0.83 0.94 1.00 b) Using selection ogive Loo(mm) M/K E max Nyanza Gulf Open water 50 65 2.3 2.3 0.43 0.64 62 5. DISCUSSION Clear growth increments consisting of alternating light and dark bands were visible in the lapillus of R. argentea. Studies on cyprinids have found that lapillus and sagitta otoliths are present in the first 4 days of development while the astericus developed in the third week (Victor and Brothers, 1982; Muth et al., 1988), suggesting that growth increment on the lapillus provided the closest estimate of age in R. argentea. Descriptions of growth and mortality during the juvenile life of R. argentea using otoliths have not been previously reported. In the present study growth was tested against three growth models, the von Bertalanffy, Gompertz, and Linear regression. The r2 value derived from regression and total sums of squares from the least-squares regression was used as a criterion in choosing goodness of fit. For juvenile R. argentea, the Gompertz model was the best fit for Nyanza Gulf (r2 = 0.86) and open waters (r2 = 0.93), with estimated asymptotic length (Lx) of 24.9 mm and 60.4 mm respectively. Linear regression had a better fit in both Nyanza Gulf (r2 = 0.84) and open waters (r2 = 0.92) compared to von Bertalanffy model ( r2 = 0.82 & 0.89). For the von Bertalanffy, L^ for open waters of 560.6 mm and K of 0.13 seems unrealistic. Larvae fish do not grow according to the von Bertalanffy model (Sparre and Venema, 1992) and hence these might not be a realistic estimate of growth parameters. Generally, fitting a model is not only a statistical procedure, but requires a decision about whether or not these particular data are suitable (Alhossaini, 1989). Unfortunately there are no published growth estimates for juvenile R. argentea with which these data can be compared, but several studies have shown that Gompertz model and von Bertalanffy are equally suitable in describing growth of larval fish (Laroche et ai, 1982; Thorrold, 1988; Alhossaini, 1989), and selection of the appropriate growth model is 63 generally based on goodness of fit (Ricker, 1979). The von-Bertalanffy growth model however, is generally used more to describe growth of adult fish (Ricker 1979). This study obtained growth coefficient, K, of 1.5 and 1.8 yr-1 for commercial catch which is in agreement with published values of K for the species. Mannini (1992) reports a K of 0.57-1.1 yr1, while Wandera and Wanink (1995) from their studies got K of 1.0-1.2 yr1. One reason for their low K may be due to the authors converting length into age thereby underestimating growth of younger fish. Size is a poor indicator of cohort membership and age estimated from daily increments ageing technique is a better estimate of age for R. argentea. Counts of daily increments provide more accurate estimates of R. argentea age and growth than has previously been available. This information allows computation of age- -dependent mortality rates resulting in more accurate estimates of larval mortality in the lake. The most significant advantage of using otolith ageing technique is the ability to produce individual rather than population statistics which have not been available for R. argentea. However, there are drawbacks to the use of otoliths increment data for estimating age and mortality. Most apparent is the extensive effort required to extract them, prepare and count growth increments. Increments are inherently difficult to locate and this problem increases with size of the fish, where increments disappear or are difficult to see on the edge and near the nucleus of otolith because of refraction of the transmitted light (Campana, 1992). Overgrinding of lapillus lead to loss of microstructure and inaccurate age determination may have been caused by nondaily deposition of rings, or failure to detect all rings within an otolith due to the resolution problem of light microscopy (Campana and Jones, 1992; Jones, 1992). Increment width in this study ranged from 1.3 64 um near the nucleus to 4.5 um towards the edge of the lapillus indicating that LM resolution was adequate for counting the increments. Theoretical resolution of a LM is 0.2 um although for practical applications it is really close to 1 um (Neilson, 1992). SEM micrographs allowed for an accurate increment width measurement because they were not subjected to refractive effects that distort an imagine under LM (Campana, 1992). Chemical marking techniques commonly use compounds such as tetracyclme (TC) or oxytetracycline (OTC) which is incorporated into growing calcified tissues within a day of application (Thorrold, 1988; Geffen, 1992). In this study TC was incorporated within 12 hours after R. argentea was immersed in TC. Incorporation occurred only in 600 mg/1 TC and this was attributed to immersion instead of injection of TC. Geffen (1992) recommends juvenile and adult fish be marked by injection. In their studies Secor et al (1991) found that incorporation rates of OTC by immersion was low and time for exposure had to be increased up to 40 hours. Injection with TC was not possible because R. argentea were very fragile and a few minutes exposure lead to their death. Of the two marked fish, the fish immersed in 600 mg/L TC for 21 hours and lived for 33 days had 30 countable increments, while the one immersed for 12 hours had no increment beyond the fluorescence band. This suggested that R. argentea likely deposits increments daily. If the assumption of a daily periodicity of increments is correct, one would expect a reasonable agreement between the values of the growth parameters derived from the present work and those given in the literature. This is in fact the case. The values of phi prime (Pauly, 1979a) from this study were 1.7 and are in agreement with 1.5,1.6 and 1.7 from Wanink (1989), Wandera (1992), Wandera and Wanink (1995) respectively. When daily increment are assumed to be deposited more than once per day phi prime does not lie within the calculated mean and standard deviation (1.62 +/- 0.084). 65 It would therefore be safe to conclude the increments seen were actually daily rings. Tetracycline although most commonly used in marking otoliths (Thorrold, 1988; Geffen, 1992), is toxic, cause disorders in the digestive system, and inhibits protein synthesis affecting growth in animals (Winstein, 1975). It is possible that R. argentea immersed in 600 mg/L TC for 12 hours growth was affected and deposition of daily increments did occur, but they were beyond the resolution of light microscope. Yoklavich and Boehlert (1987) failed to observe daily, rings in most otoliths of Sebastes melanops after injection with TC. Hetler (1984) failed to determine increment beyond the fluorescence band after immersion of fish in oxytetracycline hydrochloride (OTC). The lapillus of fish immersed in TC showed low contrast between the continuous and discontinuous zones under a light microscope. The poor deposition of daily rings in R. argentea could be related to TC immersion or constant temperature. Such faint increments are a common phenomenon among several species in the laboratory, (Campana 1984b; Alhossaini and Pitcher, 1988) found fish reared in constant temperature in the laboratory were found to produce faint increments, whereas fish subjected to diel temperature cycle are characterized by more easily observed growth increments (Neilson and Geen, 1985). While a given procedure may be preferred in certain situation, light microscopy (LM) and scanning electron microscope (SEM) will generally produce increments counts of similar accuracy and precision if increment are of sufficient width (>1 um) (Campana, 1985). In this study increment ranged from 1.3-4.5 um and structures confused with daily growth increments under LM, such as subdaily increments were also confused when viewed on SEM. Subdaily increments have been reported in several other species (Laroche et ai, 1982; Campana, 1984b; Alhossaini and Pitcher, 1988). In these studies, use of SEM provided no further confirmation on subdaily increments, proofing that LM was adequate 66 and likely produced accurate number of increments as could have SEM. A strong dose of TC usually results in mortality and in a wide diffuse band, whose central area, corresponding to the exact time of marking is difficult to locate. Small doses, on the other hand, result in rings that are difficult to detect (Geffen, 1992). In this study strong dose (600 mg/1 TC) did not produce any diffuse bands and weaker doses (100-500 mg/1 TC) produced no rings. Instantaneous growth rate shows a general decreased with age. In the open waters there is more gradual decline whereas in Nyanza Gulf the decline is gradual. As a fish grows, the growth rate per unit length slows down.This a normal trend expected in fish as they grow (Bicker, 1975). Mortality was estimated by age-catch curve and length-converted catch curve. The age-catch was applied to the juvenile population while length-converted catch curve to commercial catch. Age-catch curve gave high juvenile total mortality (Z) ranging from 7.3-26.2 yr1. The juveniles have not yet migrated to the fishing grounds and they are exposed only to natural mortality (M), while the adult population is exposed to both M and total mortality (Z). Values of total mortality (Z) got for adult population with the length-converted catch curve of 4.0 yr-1 for Nyanza Gulf and 4.8 yr-1 for open waters were in agreement with what is published. Wandera and Wanink (1995) got a range of 3.9-4.4 yr-1, Manyala et al. (1992), 3.1 yr1, and Wandera (1992), 3.6 yr"1. Pauly's formula (Pauly, 1980a) gave a natural mortality (M) of 4.1 and 3.4 yr_1 in Nyanza Gulf and open waters respectively. The value of M for Nyanza Gulf is more than Z which is not suppose to be the case and that of open waters is higher than published figures. Manyala (1991) and Mannini, (1992) reports that, R. argentea has an M ranging 67 from 0.68-0.80 yr-1 and 1.8-2.9 yr-1 respectively. This study was carried out for only 3 months and the data may not be a true reflection of the status of the whole population of R. argentea. Direct measurements of M on the other hand is often impossible to obtain, and a method like Pauly's formula (1980a) is "qualified guess" (Sparre and Venema,1992), which is influenced by mean water temperature and asymptotic length (L^). Results on mortality and growth indicate that adult R. argentea has high total mortality (Z) and growth performance (K). However, when Lake Victoria R. argentea Z and K are compared with other small pelagic species inhabiting African lakes they are found to be low (Mannini, 1992). Limnothrissa miodon from lake Tanganyika has Z of 4.4-7.4 yr-1 and K of 0.96-2.5 yr-1 and in Lake Kariba the species has Z of 8.6-13.8 yr-1 and K of 3.1 yr-1. For Stolothrissa tanganicae in Lake Tanganyika Z is 5.2-5.5 yr-1 and K is 2.6-2.89 yr-1 (Mannini, 1992; Wandera and Wanink, 1995). Turner (1982) gave values of Z for Engraulicypris sardella in Lake Malawi ranging from 2.2-5.0 yr-1 and K of 2.6 yr-1. In most tropical fish stocks recruitment continues all year round with seasonal oscillations and small pelagics like E. sardella (Rufli and van Lissa,1982) and Stolothrissa tanganicae (Roest, 1978) have more than one spawning peak during a year. Using ELEFAN II (Electron Length Frequency Analysis, Pauly, 1987) this study shows R.argentea breeds throughout the year with two peaks which corresponds to the two rainy seasons from April-June and October-December. These findings are similar to those of Wandera (1992) who found R. argentea to breed from April-May and August-September, but differ from results of Wandera and Wanink (1995) who found only one major breeding period year in October-November. During rainy seasons Lake Victoria completely or partially mixes subsequently providing plenty of food resources and encourages R. argentea gamete production (Wandera, 1992). 68 Recruitment of fish to the fishing area is size dependent. Every size of fish is not fully represented at fishing grounds and the probability that it is retained by the fishing gear is a product probability of its presence in this fishing ground and its retention by the meshes (Sparre and Venema, 1992). In Nyanza Gulf and open waters, 50% of R. argentea of 11.6 mm (44 days) and 34.5 mm (175 days) can be retained by present commercial catch gears. Wanink (1989) reports that R. argentea can lives for 1-2 years, female mature at 43-44 mm and male at 40-41 mm (Wandera, 1992). Therefore, present commercial gear (5-8 mm mesh size) in Kenyan waters of Lake Victoria are catching immature R. argentea. The fishing mortality (F) for R. argentea from Nyanza Gulf is 0.98 yr-1 and an exploitation rate (E) of 0.25, while F in the open waters is 1.4 yr-1 and E of 0.29. The plot of relative yield-per-recruit (Y/R)' against exploitation (E) shows the fishery can withstand more exploitation than this, and increase of mesh size and thus increase in length of fish at first capture (Lc) produces a higher exploitation rate (Emax). When assuming knife-edge selection, Emax is higher than in selection ogive. In knife-edge selection assumption that fish at Lc are suddenly exposed to fishing mortality does not apply in real life situation. This method tends to overestimate numbers of older fish and underestimate young fish retained by a net. For management purposes, it is important to be able to determine changes in yield per recruit for different values of exploitation (E). The (Y/R)' can provide this information (Sparre and Venema, 1992). The model requires fewer parameters and is suitable for accessing the effects of mesh size. The model forms a direct link between fish stock assessment and fishery resource management. However, the (Y/R)' model should be used with caution because the values of E producing maximum relative yield per recruit could also reduce the parental stock to a 69 level at which no recruits are produced. In small tropical fish like R. argentea the value of E which maximize relative yield-per-recruit are generally high. Like in this study E of 0.7 (Lc = 20 mm) and 1.0 (Lc = 40 mm) in Nyanza Gulf and open waters respectively are quite high and therefore using only (Y/R)1 analysis for management can be very misleading (Pauly 1979b; Pauly and Martosubroto 1980). 70 4. IMPLICATION OF STUDY FOR RASTRINEOBOLA FISHERY The knowledge of more accurate age, growth and mortality of juvenile R. argentea is fundamental to sustainability of the present Rastrineobola fishery. Information on age structure can be used to clarify the effects of changes in the environment, growth and survival of the juveniles, resulting in improved understanding of factors affecting recruitment success. In adults the information can be used to determine the effects of fishing on the stocks, to understand life history events, and to maximize yield while ensuring future stocks of dagaa are maintained. If dagaa is to be cultured, knowledge of growth rates of cultured versus wild fish can be useful in determining the feasibility, potential, and profitability of rearing the fish in captivity. Growth rate and maximum length of R. argentea in Lake Victoria has decreased significantly over the last 15 years (Manyala, 1991; Wandera and Wanink 1995). In Mwanza Gulf the modal length of dagaa decreased by 18% between 1982 and 1987 (Wanink,1988a). Unfortunately, the early work on growth of R. argenteahad never been published for comparison. In Ugandan waters Rufli and van Lissa (1982) have cited a fork length of 105 mm (95 mm SL). The growth rate was said to resemble the values found for E. sardella. Wandera and Wanink (1995) assumed a K value of 2.74 yr1 (the two mean value given for E. sardella) for the dagaa population studied by Proude and Stoneman (1973). This early study demonstrates that R. argentea had a very high growth rate and attained a greater length than reported by scientists working on the species from the late 1980s to present. This reduction in growth rate and size could be due to a number of reasons. It could be due to overfishing, high predation, intra-specific competition for food or due to abrupt changes in the environment (Katunzi, 1992). In the current situation, overfishing, use of 71 illegal gears and predation pressure by the Nile perch are considered to be the main cause of Rastrineobola stock reduction (Katunzi, 1992; Manyala et al., 1992). This study has demonstrated using Beverton and Holt relative yield-per-recruit model that fishing is a major factor in reduction in size of dagaa at first capture. Fast growing species with high turn over rate are capable of reacting to higher fishing pressure (Katunzi, 1992). In consequence, there should be a stimulation of the reproductive output in order to sustain survival. This tends towards attaining reproductive maturity at a smaller adult body size shorter life span, smaller eggs and faster growth rates. These changes have not been reported in Lake Victoria R. argentea (Wanink, 1989; Katunzi,1992). If such a scenario occurs the recovery of R. argentea could be hampered by heavy fishing pressure, use of illegal gears, Nile perch predation and pollution of the lake. Studies have shown that the Rastrineobola fishery is shifting from 10 mm to 5 mm meshes (Manyala et al., 1992) and the latter captures immature individuals of the species. This study has shown there is 50% probability of R. argentea of 11.6 mm (44 days) aand 34.5 mm (175 days) in Nyanza Gulf and open waters respectively being retained by commercial gears. These are very immature fish for a species reported to mature between 40-43 mm and live for 1-2 years (Wandera, 1992). Therefore the present form of fishing is highly destructive to R. argentea recruitment success. Continual use of smaller mesh sizes would lead to further reduction of R. argentea size/length at first capture (Lc) and consequently the fishermen would in turn reduce their mesh sizes to target the smaller dagaa. This will decrease the immediate and long-term catches of the species. Beverton (1959) had predicted a similar trend due to use of smaller mesh gill-nets to capture the dwindling stocks of Oreochromis esculentus, Bagrus docmac, and Barbus altianus, once most cherished traditional species in Lake Victoria. 72 When this problem of small mesh sizes is compounded by the effects of Nile perch predation and pollution of the lake (Ochumba, 1988), it could spell doom to Rastrineobola fishery and the entire Lake Victoria fishery. Good estimation of growth and mortality is a vital management tool and implementation of findings (like increase in mesh sizes) which will save this species are long over due. The collapse of Rastrineobola would not only be disastrous to Lake Victoria ecosystem, but to the people and the economies of the three riparian states. Rhetoric speeches should cease and those empowered with authority should seek immediate solution. 73 REFERENCES Allen, R.R.1971. Relation between production and biomass. J. Fish. Res. Board Can. 28: 1573-1581. Alhossaini, M., and Pitcher, T.J. 1988. The relationship between daily rings, body growth and environmental factors in plaice, Pleuronectes platesa L., juvenile otoliths. J. Fish. Biol. 33: 409-418. Alhossaini, M. 1989. Growth and mortality of 0-group Plaice, Pleuronectes platesa L., using otolith micro structure. Phd.Thesis.University of Bangor. Wales. Asila, A.A., Dache, S.O., Rabour, CO. 1990. A case study of the influence of beach seines and mosquito seine on the fisheries of the Nyanza Gulf: a socio-economic review. In: A Symposium organized by the IFIP under the framework of the CIFA Sub-committee for Lake Victoria, 25-27 April, Kisumu, Kenya. UNDP/FAO Regional Project for Inland Fisheries Planning (IFIP). I: 1-18. Beverton, R.J.H. 1959. Report on the state of the Lake Victoria Fisheries. Mimeo. Fisheries Laboratory. Lowestoft. Beverton, R.J.H and Holt, S.J. 1957. On the dynamics of exploited fish populations. Fish. Invest. Minist. Agric. Fish. Food G.B. (2 Sea Fish.), 19: 533 pp. Brothers, E.B. 1987. Methodological approaches to the examination of otoliths in aging studies, p. 319-330. In: Summerfelt, R.C. and Gordon, E.H. (eds.) The age and growth of fish, The Iowa State University Press. Iowa. Campana, S.E. 1992. Measurements and interpration of the microstucture of fish otoliths, p 59-71. In: Stevenson, D.K. and Campana, S.E. (eds). Otoliths microstructure examination and analysis. Can. Spec. Publ. Fish. Aquat Sci. 117. Campana, S.E. 1984b. Interactive effects of age and environmental modifiers on the production of daily growth increments in otoliths of plainfin midshipman, Porichthys notatus. Fish. Bull. U.S. 82: 165-177. Campana, S.E. 1985. Microstructure of fish otoliths. Can. J. Fish. Aquat. Sci. 42: 1014-1032. Campana, S.E., and Neilson, J.D. 1985. Microstructure of fish otoliths. Can. J. Aquat. Sci. 42: 1015-1032. Campana, S.E., and Jones, CM. 1992. Analysis of otolith microstructure data, p 73-100. In: Stevenson, D.K. and Campana, S.E. (eds.). Otoliths microstructure examination and analysis. Can. Spec. Publ. Fish. Aquat. Sci. 117. 74 Chitamwemba, D.B.R. 1992. The fishery of Rastrineobola argentea in the southern sector of Lake Victoria p 51-61. In. Mannini, P. (ed.) The Lake Victoria dagaa {Rastrineobola argentea). Report of the first meeting of the working group on Lake Victoria Rastrineobola argentea, 9-11 December 1991, Kisumu, Kenya. UNDP/FAO Regional Project for Inland Fisheries Planning (IFIP). Christensen, V and Pauly, D. 1992. A guide to the ECOPATH II program (version 2.1). ICLARM soft ware 6, 72. International Centre for Living Aquatic Resources Management. Manila, Philippines. CIFA, 1988. Report of the fourth session of the Sub-committe for the Development and Management of the Fisheries of Lake Victoria. Kisumu, Kenya, 6-10 April 1987. FAO. Fish. Rep. 388: 112 pp. Copley, H. 1953. The tilapia fishery of Kavirondo Gulf. J. E. Afri. Nat. Hist. Soc. 94: 1-5 Corbet, P.S. 1961. The food of non-cichlids fishes in the Lake Victoria basin with remarks on their evolution and adaptation to lacustrine conditions. Proc. Zool. Soc. Lond. 316, 11: 1-101. Essig, R. J. and Cole, F.C 1986. Methods of estimating larval fish mortality from daily increments in otoliths. Trans. Amer. Fish. Soc. 115: 34-40. FAO. 1992. Report of the Sixth Session of the CIFA Sub-Committee for the Development and Management of the Fisheries of Lake Victoria, Jinja, Uganda 10-13 February 1992. FAO Fish. Rep. 475. 2pp. Gj<£saeter, J., Dayaratne, P., Bergstad, OA, Gjosaeter, H., Sousa, M.I., and Beck, I. M. 1984. Ageing tropical fish by daily growth in the otoliths. FAO Fish. Circ. 776: 54pp. Geffen, AJ. 1992. Validation of otolith increment deposition rate, p 101-113. In. Stevenson, D.K. and Campana, SE. (eds). Otolith microstructure examination and analysis. Can. Spec. Publ. Fish. Aquat. Sci. 117. Graham, M. 1929. The Victoria Nyanza and its fisheries. A report on the fishing survey of Lake Victoria, 1927-1928. Crown Agents, London, 225 pp. Gulland, J.A. 1983. Fish stock assessment, a manual of basic methods. John Wiley and Sons. New York. Hall, D.L. 1992. Age validation and aging methods for stunted brook trout. Trans. Am. Fish. Soc. 120: 644-649. HEST. 1988. Reports from the Haplochromine Ecology Survey Team (HEST) and the 75 Tanzania Fisheries Research Institute (TAFIRI) operating in Lake Victoria. Leiden, The Netherlands, 39: 10 pp. Howes, G.J. 1984. A review of the anatomy, taxonomy, phylogeny and biogeography of the African neoboline cyprinid fishes. Bull. Br. Mus. (Nat. Hist.) Zool. 47(3) 151-185. Howes, G.J. 1980. The anatomy and, phylogeny and classification of the bariliine cyprinid fishes. Bull. Br. Mus. nat. Hist. (Zool.) 37(3): 129-198. Hetler, W.F. 1984. Marking otoliths by immersion of marine fish larvae in tetracycline. Trans. Am. Fish. Soc. 113: 370-373. Jenkins, G.P., and Davis, T.L.O. 1990. Age, growth rate, and growth trajectory determined from otolith microstucture of southern bluefin tuna Thannus maccoyii larvae. Mar. Ecol. Prog. Ser. 63: 93-104. Jones, CM. 1992. Development and application of otolith increment technique, p 1-11. In. Stevenson, DK and Campana, S.E. (eds). Otolith microstructure examination and analysis. Can. Spec. Publ. Fish. Aquat. Sci. 117. Katunzi, E.F.B. 1992. Biological and fishery aspects of Rastrineobola argentea in the southern part of Lake Victoria, p 51-6. In. Mannini, P (ed.) The Lake Victoria dagaa {Rastrineobola argentea). Report of the first meeting of the working group on Lake Victoria Rastrineobola argentea, 9-11 December 1991, Kisumu, Kenya. UNDP/FAO Regional Project for Inland Fisheries Planning (IFIP). Kudhongania, A.W. and Cordone, A.J. 1974. Past trends, present stocks and possible future state of the fisheries of the Tanzania part of Lake Victoria. Afri. J. Trop. Hydrobiol. Fish. 3: 167-181. Kudhongania, A.W., Twongo T. and Ogutu-Ohwayo, R. 1992. Impact of the Nile perch fisheris of Lake Victoria and Kyoga. Hydrobiologia 232: 1-10. Laroche, J.L., Richardson, S.L., and Rosenberg, A.A. 1982. Age and growth of a Pleuronectid, Parophrys vetulus, during the pelagic larval period in Oregon coastal water. Fish. Bull. U.S. 80: 93-104. Mannini, P. 1992. Some characteristics of small pelagic species and possible affinities with the population of Lake Victoria Rastrineobola argentea, p 62-84. In: Mannini, P. (ed.) The Lake Victoria dagaa {Rastrineobola argentea). Report of the first meeting of the working group on Lake Victoria Rastrineobola argentea, 9-11 December 1991, Kisumu, Kenya. UNDP/FAO Regional Project for Inland Fisheries Planning (IFIP). 76 Manyala, J.O. 1991. Population dynamics of Rastrineobola argentea (Pellegrin) 1904 (Pisces: Cyprinidae) in the Winam Gulf of Lake Victoria Kenya. Master Thesis. University of Nairobi. Kenya. Manyala, J.O. Nyawade, CO., and Rabour, CO. 1992. The dagaa {Rastrineobola argentea PELLEGRIN) fishery in the Kenyan waters of Lake Victoria: a national review and proposal for future research, p 18-35. In: Mannini, P.(ed.) The Lake Victoria dagaa {Rastrineobola argentea). Report of the first meeting of the working group on Lake Victoria Rastrineobola argentea, 9-11 December 1991, Kisumu, Kenya. UNDP/FAO Regional Project for Inland Fisheries Planning (IFIP). Moreau, J., Ligtvoet, W., and Palomares, M.L.D. 1993 Trophic relationship in the fish community of Lake Victoria, Kenya, with emphasis on the impact of Nile perch {Lates niloticus), p 144-152 In: Christensen, V. and Pauly, D (eds) Trophical models of aquatic ecosystems. ICLARM Conf. Proc. 26, 390 pp. Mous, P.J. Budeba, Y.L., Temu, M.M. , and van D'2nsen,W.L.T. 1991. A catch effort data recording system for the fishery of the small pelagic Rastrineobola argentea in the southern part of Lake Victoria p 335-348. In: Cowx, I.G. (ed.) Catch sampling strategies: their application in freshwater management. Fishing News Books. London. Muth, R.T., Nesler, T.P., and Wasowicz, A.F. 1988. Marking cyprinids with tetracycline. Am. Fish. Soc. Symp. 5: 89-95. Neilson, J.D. 1992. Sources of error in otolith microsture examination, p.115-121. In: Stevenson, D.K and Campana, S.E. (eds). Otolith microstructure examination and analysis. Can. Spec. Publ. Fish. Aquat. Sci. 117. Neilson, J.D., and Geen, G.H. 1985. Effects of feeding regimes and diel temperature cycles on otolith increment formation in juvenile chinook salmon (Oncorhynchus tshawytscha). Fish. Bull. U.S. 83: 91-101. Ochumba, P.B.O 1988. Periodic massive fish kills in the Kenyan portion of Lake Victoria, p 47-60. In: CIFA, Report of the 4th Session of the Sub-committee for Development and Management of Fisheries of Lake Victoria, Kisumu, Kenya 6-10 April 1987. FAO. Fish. Rep. No 388. Ogari, J., and Dadzie, S. 1988. The food of the Nile perch, Lates niloticus (L.) after the disappearance of haplochromine cichlids in the Nyanza Gulf of Lake Victoria (Kenya). J. Fish. Biol. 32: 572-577. Ogutu-Ohwayo, R., Twongo, T., Wandera, S.B., and Balirwa, J.S., 1988. Fishing gear selectivity in relation to their manufacture and to the management of fisheries of the Nile perch, the Nile tilapia and Rastrineobola argentea (mukene) in Lake's Victoria 77 and Kyoga. A guide to fishnet manufacturers, fisheries managers and fishermen. UFFRO occasional pap. No 16, 16 pp. Okach, J.I.O. 1982. Reproductive biology and feeding ecology of catfish, Bagrus docmac Forskalii. (Pisces, Bagridae) in Winam Gulf of Lake Victoria. Master Thesis. University of Nairobi. Kenya. Okaronon, J.O. 1992. The changing fisheries of Lake Victoria, Uganda, p. 34. In: Report of the Sixth Session of the CIFA Sub-Committee for the Development and Management of the Fisheries of Lake Victoria, Jinja, Uganda, 10-13 February 1992, FAO Fish. Rep. 475. Okedi J. 1973. Preliminary observations on Engraulicypris argentus (Pellegrin 1904) from Lake Victoria. EAFRO Annual report 1973, 3-4 p. Okedi J. 1982. Standing crop and biomass estimates of Lake Victoria "Dagaa" (Engraulicypris argentus Pellegrin). Mimeo, 6 pp. Pauly, D. 1979a. Gill size and temperature as governing factors in fish growth: a generalization of von Bertalanffy's growth formula. Ber. Inst. Meereskd. Christian-Albrechts-Univ. Kiel. 63: 156 pp. Pauly, D. 1979b. Theory and Management of Tropical Multispecies Stocks: a Review with Emphasis on the Southern Asian Demersal Fisheries. ICLARM Stud. Rev. 1: 35 pp. Pauly, D. 1980a. On the interrelationship between natural mortality, growth parameters, mean environmental temperature in 175 fish stocks. J. Cons. int. Explor. Mer, 39(3): 175-192. Pauly, D. 1984c. Fish population dynamics in tropical waters: a manual for use with programmable calculator. ICLARM study and reviews 8,325 pp Pauly, D. 1987. A review of the ELEFAN system for analysis of length-frequency data in fish and aquatic inveterbrates, p 7-34. In: Pauly,D . and Morgan, GR. (eds) Length-base methods in fisheries research. ICLARM. Con. Proc. 13, 486 pp. Pauly, D. 1994. On the sex of fish and the gender of scientist. A collection of essays in fisheries science. Chapman and Hall. London. 250 pp. Pauly, D., Ingles, J. and Neal, R. 1984. Application to shrimp stock of objective methods for the estimation of growth, mortality and recruitment- related parameters from length-frequency data (ELEFAN I and II), p 220-234. In: Gulland, J.A. and Rothschilds, B.J. (eds.) Panaeid shrimps - their biology and management. Fishing News Books. Farnham, Surrey. England. 78 Pauly, D. and Martosubroto, P. 1980. The population dynamics of Nemipterus marginatus off Western Kalimantan, South China Sea. J. Fish. Biol. 17: 263-273. Pitcher, TJ. 1993. Impact of changes in African Lakes. Fish management science Research Programme. Final report on ODA research project R4683. 198 pp. Pitcher, T.J and Hart, P.J.B 1982. Fisheries ecology. London. Croom Helm. 414 pp. Proude, P.D. and Stoneman, J. 1973. Monthly length frequency records of Engraulicypris argetus in Lake Victoria with notes on food and sexual maturity (Mimeo). Fisheries Department, Entebbe, Uganda. Rabour, CO. 1991. Catch and sampling effort techniques and their application in fresh water fisheries management: with specific reference to Lake Victoria, Kenya waters, p 373-381. In: Cowx, I.G. (ed.) Catch effort sampling strategies: their application in freshwater fisheries management. Oxford University press. Ricker, W.E. 1975. Computation and interpretation of biological statistics of fish population. Bull. Fish. Res. Bd. Canada 191: 382 pp. Ricker, W.E. 1979. Growth rate and models, p 677-743. In: Hoar, W.S., Randall, D.J., and Brett, J.R. (eds.). Fish physiology. Vol.8: Bioenergetics and Growth. Academic Press. New York. Robson, D.S., and Chapman, D.G 1961. Catch curves and mortality rates. Trans. Amer. Fish. Soc. 90: 181-189. Roest, F.C. 1978. Stolothrissa tanganicae: population dynamics, biomass evolution and life history in the Burundi waters of lake Tanganyika. In: Welcomme, R.E. (ed.) Symposium on river and floodplain fisheries in Africa, Bujumbura, Burundi, November 21-23 1977. Review and experience papers. CIFA Tech. Pap. 5: 42-61 Rufli, H. and van Lissa, J. 1982. Age and growth of Engraulicypris sardella in Lake Malawi, p.85-97. In: Biological studies on the pelagic ecosystem of Lake Malawi. FI/MLW/75.019, Tech. Rep.l. FAO. Rome. Secor, D.H., Dean, J.M., and Laban, E.H. 1992. Otolith removal and preparation for micrcostructure examination, p 19-57. In: Stevenson, D.K. and Campana, S.E. (eds). Otoliths microstructure examination and analysis. Can. Spec. Publ. Fish. Aquat. Sci. 117. Sparre, P., and Venema, SC., 1992. Introduction to tropical fish assessment. Part 1. Manual. FAO Fisheries Technical paper, Rev. 1. 306.1: 376 pp. 79 Thorrold, S.R. 1988. Estimating some early life history parameters in tropical clupied, Herklotsichthys castelnaui, from daily growth increments in otoliths. Fish. Bull. 87: 73-83. Turner, J.L. 1982. Analysis of the catch and effort data of a purse seine fishery for Engraulicypris sardella at the southern end of Lake Malawi, p 109-114. In: Bilogical studies on the pelagic ecosystem of Lake Malawi. FLDP/MLW/75/019. Wandera, S.B. 1992. A study of Rastrineobola argentea in Ugandan Lakes, p 36-50. In: Mannini, P. (ed.) The Lake Victoria dagaa {Rastrineobola argentea). Report of the first meeting of the working group on Lake Victoria Rastrineobola argentea, 9-11 December 1991, Kisumu, Kenya. UNDP/FAO Regional Project for Inland Fisheries Planning (IFIP). Wandera, S.B. and Wanink, JH. 1995. Growth and mortality of Dagaa {Rastrineobola argentea, Fam. Cyprinidae) in Lake Victoria. Naga, the ICLARM quarterly. 42-45. Wanink, JH. 1988a. The recent changes in the zooplanktorous/insectivorous fish community of the Mwanza Gulf. HEST/TAFIRI Rep. 51, 25p Leiden, The Netherlands. Wanink, JH. 1989. The ecology and the fisheries of dagaa Rastrineobola argentea (Pellegrin) 1904. In. Fish Stock and Fisheries in Lake Victoria. A handbook to the HEST/TAFIRI & FAO/DANIDA regional seminar, Mwanza January/February 1989. Report of the Haplochromis Ecology Survey Team (HEST) and the Tanzania Fisheries Research Institute (TAFIRI) no.53 Leiden, The Netherlands, RUL. Winstein, K.H. 1975. Antimicrobial agents: Tetracycline and Chloramphenicol, p 1183-1200. In: Goodman, LS. and Gilman, A. (eds.) The phermocological basis of therapeutics. Macmillan Publishing Co. New York. Welcomme, R.L. 1964. The habitat and habitat preference of the young of Lake Victoria Tilapia (Pisces-Cichlidae) Revue. Zool. Bot. afr. 70: 1-28. Worthington, E.B. 1929. A report on the fishing survey of Lakes Albert and Kyoga. London: Crown Agents. 136 pp. Yoklovich, M.M. and Boehlert, G.W. 1987. Daily growth increments in otoliths of juvenile black rockfish, Sebastes melanops: an evaluation of autoradiography as anew method of validation. Fish. Bull. 85: 826-832. Zweifel, J.R. and Lasker, L. 1976. Prehatch and post hatch growth of fishes- a general model. Fish. Bull. U.S. 74: 609-621. 80 Appendix 1 Table ALL Length-frequency samples of juvenile R. argentea from Nyanza Gulf. Length (mm) January February March April May Total 3 0 1 0 0 0 1 4 1 0 0 0 0 2 5 16 1 12 0 0 29 6 20 4 19 2 2 47 7 30 10 24 13 6 83 8 11 29 31 16 7 90 9 33 70 8 10 6 133 10 28 55 25 11 8 126 11 15 53 17 1 2 98 12 15 25 10 0 0 51 13 17 26 14 0 0 58 14 4 17 5 0 0 26 15 7 7 12 0 0 26 16 21 15 6 0 0 42 17 3 14 9 0 0 26 18 2 14 17 0 0 33 19 4 4 4 0 0 12 20 0 7 6 0 0 13 21 0 5 17 0 0 22 22 0 7 4 0 0 11 23 1 5 2 0 0 8 24 0 8 1 0 0 9 25 0 1 0 0 0 1 26 1 4 1 0 0 6 Total 229 382 244 66 31 952 81 Table A 1.2. Length-frequency samples of juvenile R. argentea from open waters sampling site. Length (mm) February March April May Total 3 0 0 0 0 0 4 0 0 0 0 0 5 0 12 0 0 12 6 0 17 1 2 20 7 0 23 4 6 33 8 3 31 1 6 41 9 19 0 6 5 30 10 13 0 5 4 22 11 14 0 5 1 20 12 8 0 0 0 8 13 12 0 1 0 13 14 5 0 0 0 5 15 1 0 0 0 1 16 7 0 0 0 7 17 4 0 0 0 4 18 5 0 0 0 5 19 2 0 0 0 2 20 3 0 0 0 3 21 0 0 0 0 0 22 3 0 0 0 3 23 3 0 0 0 3 24 2 0 0 0 2 25 0 0 0 0 0 26 2 0 0 0 2 Total 106 83 23 24 236 82 Table A 1.3. Length frequency samples of R. argentea commercial catch from Ny Gulf. Length (mm) January February March Total 5 0 1 0 1 6 0 0 0 0 7 0 3 0 3 8 0 3 0 3 9 1 3 1 5 10 0 7 3 10 11 5 4 3 12 12 5 8 3 16 13 3 10 3 16 14 7 1 4 12 15 22 2 2 26 16 5 4 6 15 17 1 9 5 15 18 5 4 5 14 19 2 6 3 11 20 5 5 9 19 21 4 7 4 15 22 10 4 2 16 23 5 8 3 16 24 11 4 1 16 25 16 4 3 23 26 9 1 1 11 27 6 5 2 13 28 3 3 5 11 29 3 3 3 9 30 7 2 1 10 31 6 2 2 10 32 6 5 1 12 33 6 2 2 10 34 3 2 0 5 35 2 9 1 12 36 1 4 2 7 37 2 11 2 15 38 0 13 1 14 39 1 15 3 19 40 0 11 0 11 41 1 4 0 5 83 Table A 1.3. Continual Length January February March Total (mm) 42 1 3 1 5 43 1 1 1 2 44 0 6 4 10 45 1 4 1 6 46 0 4 0 4 47 0 0 0 0 48 0 0 0 0 49 0 0 0 0 50 0 1 0 1 51 0 2 0 1 52 0 2 0 2 53 0 2 0 2 54 0 0 0 0 55 0 0 0 0 56 0 0 0 0 57 0 1 0 1 58 0 3 0 3 Total 166 83 24 236 Table A 1.4. Length frequency samples of R. argentea commercial catch from open waters. Length (mm) February (mm) Length February 37 5 48 0 38 7 49 0 39 7 50 1 40 9 51 2 41 3 52 2 42 3 53 2 43 0 54 0 44 5 55 0 45 3 56 0 46 4 57 1 47 0 58 1 Total 55 84 Appendix 2 Table A2. Age-frequency samples of juvenile R. argentea from Nyanza Gulf and open waters sampling site. Nyanza Gulf Open waters Age Frequency (days) Jan Feb March Feb March 10 1 0 0 0 3 12 3 0 0 0 3 14 11 1 0 0 5 16 2 1 0 0 4 18 7 2 1 0 5 20 7 5 0 0 14 22 10 4 0 0 7 24 8 3 2 0 6 26 1 0 1 0 3 28 6 3 2 0 1 30 3 5 1 1 0 32 3 4 2 1 1 34 7 3 2 0 0 36 3 8 1 6 0 38 8 7 0 1 1 40 8 8 1 3 3 42 2 7 0 4 0 44 3 12 0 5 3 46 5 6 2 6 0 48 1 12 1 2 0 50 15 11 4 2 0 52 3 7 . 3 2 0 54 3 8 3 0 0 56 4 5 0 1 0 58 0 3 1 2 0 60 2 6 4 3 0 62 2 7 3 3 0 64 2 3 1 2 0 66 1 1 1 0 0 68 2 11 4 5 0 70 3 •7 7 0 0 72 1 1 5 0 0 85 Table A2. Continual Nyanza Gulf Open waters Age Frequency (days) Jan. Feb. March Feb. Mar 74 0 0 2 0 1 76 0 3 1 0 0 78 0 4 0 0 0 80 2 3 4 1 0 82 0 0 2 2 0 84 3 3 1 1 0 86 0 0 1 0 0 88 0 2 0 0 0 90 2 8 3 3 0 92 0 0 0 0 0 94 0 1 0 0 0 96 1 3 0 3 0 98 0 4 0 0 0 100 3 4 2 0 0 102 0 0 0 0 0 104 0 0 0 2 0 106 0 0 3 2 0 108 0 1 1 0 0 110 1 3 1 0 0 112 0 1 0 0 0 114 0 0 0 0 0 116 0 3 1 0 0 118 0 2 0 0 0 120 0 1 1 0 0 122 0 0 0 0 0 124 0 0 0 0 0 126 0 0 0 0 0 128 0 0 0 0 0 130 0 1 0 0 0 132 0 0 0 0 0 134 0 1 0 0 0 86 Appendix 3 Data for estimation of Lx using the method of Powell-Wetherall Table A3.1. Nyanza Gulf. L(mean)-L' L' 18.3 5.0 15.9 7.5 13.8 10.0 12.4 12.5 11.1 15.0 10.3 17.5 9.1 20.0 7.4 22.5 7.1 25.0*** 6.3 27.5 5.6 30.0 4.9 32.5 3.3 35.0 3.4 37.5 1.8 40.0 1.8 40.2 1.2 45.0 Table A3.2. Open waters. L(mean)-L' L' 12.0 30.0 9.8 32.5 8.0 35.0 6.8 37.5 6.6 40.0 6.7 42 5*** 5.9 45.0 6.4 47.5 4.4 50.0 3.8 52.5 3.2 55.5 1.2 57.5 Regression line fitted from this point 87 Appendix 4 Robson and Chapman method (1961) Its a least square method providing a statistical test whether the points lie on the same line. The method estimates Z from progressive subsets of the whole catch curve, dropping the youngest age from the analysis each time, until the test becomes non-significant. 1 On a scatter plot age is plotted against natural logarithm of number at age (Frequency). 2. Estimation of Mortality: Age which to start is coded zero, then the other subsequent ages coded / in order up to the maximum age (i max). The catches for each coded age are ni • Calculate: T = S in; from i = 1 to i max N = Z n; from i = 0 to i max Survival percentage (SV) is estimated as: SV = T/(N+T-1) So the instantaneous mortality rate (Z) is given by Z = -Ln( survival) Confidence limits may be attached by calculating the variance S2(SV) = SV (SV- [(T-l)/(N+T-2)] and standard deviation S(SV) = VS2(SV) which may be used with student's t for (N-l) d.f. to set confidence limit win the usual way. 3. The validity of the analysis on the current set of ages can be checked using chi-square which tests whether the first age group (n0) fits with the rest of the curve. Calculate: h = (N-n0)/N and b = [T(T-l) (N - 1)]/ [ (N(N+T-1)2 (N+T-2)] then X20bs = (SV-hJVb which is checked with chi-squarecrit tables or chart with 1 d.f. (NB.95% value of chi-square^, is 3.841). 4. If the test gives a significant result, the whole analysis is moved on by one age and repeated (go back to (2) above) until the test becomes non-significant. 5. Regression is started from this non-significant point including all the other points after it. 88 Appendix 5 Table A5.1. Length-frequency from Nyanza Gulf used for length-converted catch curve. Midlength Jan. Feb. March (mm) 6.25 2.5 8.75 1.0 7.5 2.0 11.25 7.5 15.0 7.5 13.75 12.5 15.0 8.5 16.25 27.5 10.5 10.5 18.75 7.5 14.5 10.5 21.25 14.0 14.0 14.5 23.75 21.0 14.0 5.0 26.25 28.0 7.5 5.0 28.75 9.0 8.5 9.0 31.25 16.0 4.5 3.5 33.75 12.0 5.5 2.5 36.25 4.0 1.5 4.0 38.75 2.0 14.5 5.0 41.25 1.5 3.0 0.5 43.25 1.5 2.0 1.0 46.25 1.0 1.0 Table A5.2. Length-frequency from open waters used for length-converted catch curve. Midlength February (mm) 31.25 1.5 33.75 6.0 36.25 10.5 38.75 16.0 41.25 12.8 43.75 6.3 46.25 7.8 48.75 1.3 51.25 4.0 53.75 3.0 56.25 0.75 58.75 3.3 89 Data points used in regression a)Nyanza Gulf Owena (coMMercial) (Wt.Mode : la > 49 o O T o o Z W o • 31 o 0 o C Ol 25 * 3 * * A * o * • • 1 «• 6. 25 26.25 46.25 Midlensrth (MM) b) Open waters dasraa <R. air»srentea) 18 -o z o 3t 0 o c 9 Of • T • o g i • 31 25 45.5 59.75 Midlensrth (MM) * Used data points o Points not used Selection of points to be used in tile regression (Pauly, 1984c). 1. The first point in the regression should be the point immediately to the right of the highest point on the catch curve plot. Points were chosen by soft ware ELEFAN II. 2. Any point within 5 % of L^should be discarded as these will generate unrealistically high age. 3. One single outlier may be rejected within the region where the straight line to be fitted. 90 Appendix 6 Table A6. Annual estimates of total mortality (Z), fishing mortality (F), natural mortality (M) and exploitation rate (E) from age-catch curve, length-converted catch curve and ECOPATH II from Nyanza Gulf and open waters. age len ecop Pauly age len age len Zl Z2 Z3 M FI F2 El E2 Gulf 13.4 4.0 2.2 3.2 10.2 0.80 0.76 0.20 Open 19.7 4.8 2.2 2.7 17.0 2.10 0.86 0.44 age = estimates from age-catch curve, len = estimates from length-converted catch curve, eco = estimate from ECOPATH II model. Pauly = estimate of M from Pauly equation. Note in the above table Z = F-M. M used to estimate F accounted for schooling behaviour (20% reduction). M = 3.2 yr-1 in Nyanza Gulf and 2.7 yr-1 in open waters. 91 

Cite

Citation Scheme:

    

Usage Statistics

Country Views Downloads
United States 16 2
India 8 0
France 4 0
Japan 3 0
Kenya 3 0
China 3 15
Iceland 2 0
Philippines 1 0
Ireland 1 0
Germany 1 0
Netherlands 1 0
City Views Downloads
Unknown 17 2
Clarks Summit 6 0
Ashburn 4 0
Tokyo 3 0
Wilkes Barre 3 0
Mountain View 3 2
Shenzhen 2 15
Nairobi 2 0
Akureyri 1 0
Cork 1 0
Beijing 1 0

{[{ mDataHeader[type] }]} {[{ month[type] }]} {[{ tData[type] }]}
Download Stats

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0074806/manifest

Comment

Related Items