TR 91 - 60 ASPECTS OF THE BIOLOGY, ECOLOGY AND POPULATION DYNAMICS OF GALEICHTHYS FELICEPS (VALENCIENNES) AND G. ATER (CASTELNAU) (PISCES: ARIIDAE) OFF THE SOUTH-EAST COAST OF SOUTH AFRICA THESIS Submitted in fulfilment of the Requirements for the Degree of DOCTOR OF PHILOSOPHY of Rhodes University by ROBIN LEWIS TILNEY December 1990 i ACKNOWLEDGEMENTS The Port Alfred municipality provided laboratory facilities and free accommodation at the Lagoon. Nature conservator Kevin Reynolds introduced me to the fishing community and provided invaluable logistic support and gill-netting expertise throughout the sampling period. The boat-owners in the commercial fishery gave me open access to the landing wharf and their catches. Roger and Pat Rathbone, Tony Welsh, Garth, Maureen, Claude and Mr . Dennis McLellan, Mannetjie van der Merwe, Stephanus Steyn, Mr. Gerald Schultz and Granville Guthrie were particularly helpful. The fishermen themselves were always tolerant of my presence on the wharf regardless of the often over-crowded conditions and their haste to process the catches. Sea Fisheries Inspector Trevor Harty was particularly knowledgeable about the fishery and provided welcome support throughout the project . Marc Griffiths helped me with my sampling programme and critically read several draft chapters of the thesis . Andy Scholtz provided excellent technical assistance and computing advice. His expertise was boundless. Lil Haigh prepared specimens for the ontogenetic study and critically read the provided expert chapter. photographic Jim Carnbray and Robin Stobbs assistance. Robin Cross and Shirley Pinchuck assisted with the electron microscopy . Andre Punt and Sarah Radloff provided mathematical and statistical expertise respectively. Malcolm Smale and Colin Buxton made Galeichthys material from their Algoa Bay sampling programme available to me. Colin also critically read the feeding chapter and contributed to many valuable discussions. Paul Skelton, Tony Ribbink, Mike Bruton, ii phil Heemstra, Leonard Compagno, Alan Whitfield and Ofer Gon were always generous with their time and expertise. Librarians Margaret Crampton, Eve Cambray and Sheila Villett were efficient, helpful and friendly. The JLB Smith Institute 'Fishlit' data service was particularly useful. Artist Dave Voorveldt did the original illustrations for the frontispiece and provided artistic advice on several occasions. Department secretary Linda Coetzee assisted with typing and computing. Funding was provided by SANCOR. My mother, Heather, pointed me in the direction of Ichthyology & Fisheries Science in January 1983 and saved me from being a school teacher. She has provided constant love and support for as long as I can remember. My partner, Lorna Burns, endured this seemingly Sisyphean endeavour courageously. She has been my harshest critic and strongest supporter. She also proof-read the final copy. My supervisor, Tom Hecht, was masterful in his guidance of this thesis. He recognised the potential of the data where I did not and sent me back to the drawing board time and again . He would squeeze blood from a stone I I thank you all sincerely. iii ABSTRACT This thesis represents a detailed investigation into aspects of the biology, ecology and population dynamics of two endemic ariid species, Galeichthys feliceps and G. ater, off the southeast coast of South Africa. The two species are exploited as a by-catch in the commercial ski-boat fishery off Port Alfred, a fishery dominated by highly fecund sparid and sciaenid species. They collectively constitute approximately 10% of the total annual catch in terms of landed mass . ~ feliceps outnumber G. ater in the catches by a ratio of 3:1. The investigation was designed to provide the biological data required for stock assessment and to determine optimum management strategies for the two populations. The implications of their K-selected life-history styles for exploitation received particular attention. While the two species were sympatric and had similar depth distributions they were found to be allopatric with respect to their foraging habitats. ~ feliceps foraged over sandy and muddy substrata in marine and estuarine environments. G. ater fed only on reef-associated species and did not utilise estuaries. Their feeding-associated morphologies were identical and both species preyed primarily on crustaceans (brachyuran crabs and isopods), echiurids, molluscs and polychaetes. The diet of ~ consumed. The two ater was broader in terms of the number of species species are mouth-brooders with low fecundity. ~ feliceps and G. ater produced a mean of 49 and 32 eggs each, per annum. The buccal incubation period was determined to be in the region of 140 days for ~ feliceps. Embryos hatched after approximately 75 - BO days and the young began exogenous feeding thereafter. The young fed intra-buccally on detritus provided by the parent. Adult buccal mucus may also have been iv used as a food source. Young were released at a total length of ± 55rnm. Adult males ceased feeding whilst mouth-brooding. Body musculature, abdominal fat and liver reserves provided energy for basal metabolism and males lost approximately 28% of their body mass during buccal incubation. Females expended less reproductive energy than males. Catches were dominated by mature fish (76% in ~ feliceps and 97% in G. ater). Females were significantly more abundant in catches during the spawning and mouth-brooding period. female to male feliceps and sex ratios were 1 . 65: 1 and 2.23: 1 for The ~ ater respectively. ~ Age and growth studies revealed that the two species mature at advanced ages (10 and 9 years for G. feliceps and 9 and 7 years for ~ ater males and females respectively). They are long-lived, reaching ages in excess of 18 years in G. feliceps and in excess of 15 years in ~ ater. Females live longer than males and grow larger. Yield-per-recruit and spawner biomassper-recruit analyses demonstrated that G. ater were exploited below FO.1 at a level where spawner biomass-per-recruit was reduced to between 45% and 65% of the unexploited level. The ~ ater stock was not adversely affected by current levels of fishing effort. For ~ feliceps, both sexes were exploited beyond FO.1 where spawner biomass-per-recruit was reduced to between 30% and 22% of the unexploited level. G. feliceps were shown to be sensitive to relatively low levels of exploitation, a phenomenon attributed to their highly Kselected life-history style. Should the species become targeted for in the future, effort restrictions in the form of a closed season during the spawning and mouth-brooding period would prove effective in population sex ratio. reducing effort and conserving the v TABLE OF CONTENTS CHAPTER 1 - INTRODUCTION 1 CHAPTER 2 - FEEDING 7 Introduction 7 Materials & Methods 8 13 Results Anatomical adaptations for feeding 13 Diet: Offshore samples 20 Sub-tidal gully samples 23 Estuarine samples 24 Eleuthero-embryo sample 27 28 Feeding seasonality Discussion 32 Feeding morphology 32 Diet, diet overlap, diet shift and the probability of lepidophagy 35 Habitat utilisation and resource partitioning CHAPTER 3 - REPRODUCTION 37 43 Introduction 43 Materials & Methods 46 Results 49 Gonad description 51 Sizes at sexual maturity 55 Spawning seasonality 56 Seasonali ty of with reproduction energy reserves ...... . associated . 56 Relationship between fecundity and body dimensions 60 Reproductive energy investment 64 Speculations on spawning behaviour 66 Discussion 67 vi CHAPTER 4 - EARLY ONTOG ENY 84 Introduction 84 Materials & Methods 86 Results 90 Descriptive early ontogeny of G . feliceps 93 Cleavage phase 93 Embryonic Phase 96 Free-embryo phase Discussion 117 128 Endogenous nutrition 133 Duration of the buccal incubation period 133 Saltatory ontogeny or otherwise? 135 CHAPTER 5 - AGE AND GROWTH Introduction .... Materials and Methods 137 137 138 Annulus validation 140 Reading of otolith sections 140 Results 143 Nature of the growth zones 143 Annulus validation 144 Discussion 157 CHAPTER 6 - POPULATION DYNAMICS 164 Introduction 164 Materials & Methods 170 Mortality determination 170 Selectivity 173 Yie1d-Per-Recruit Ana~ysi Results 174 176 Mortality estimation 176 Sex ratios 183 . . . Yield-per-recruit analyses Discussion 184 194 vii CHAPTER 7 - GENERAL DISCUSSION 202 Reproduction 203 Mouth-brooding 206 Evolution of mouth-brooding. 208 The barbel fishery - Implications of K-selection 212 BIBLIOGRAPHY . .. . .. . . . . . . . . . . . . . 218 APPENDIX I 256 APPENDIX II 266 APPENDIX III 272 APPENDIX IV 278 CHAPTER 1 - INTRODUCTION Fishes of the family Ariidae are predominantly marine and occur circum-globally in shallow coastal waters of tropical, sub-tropical and temperate oceans. Several species are also estuarine and some penetrate the mid- and upper reaches of rivers where they may live and reproduce in fresh water (Rimmer & Merrick 1983). It is a speciose family, comprising over 150 species (Rimmer & Merrick QQ cit.), many of which form an important component of both artisinal and commercial fisheries in many tropical and sub-tropical regions of the world (Singh & Rege 1968; Tobor 1969; Wongratana et al. 1974; Drnitrenko 1975; Taylor & Menezes 1977; Etchevers 1978; Jayaram & Dhanze 1978; Warburton 1978; Silas et al. 1980; Dan 1981; Taylor & Van Dyke 1981; Jayaram & Kailola 1983; Muncy & Wingo 1983; Cortes 1984; Menon 1984; Reis 1986; Bawazeer 1987; Brothers & Mathews 1987; Euzen 1987; Coates 1988; Pauly & Thia-Eng 1988). The siluriformes are thought to have evolved in fresh water in the South American and African tropics (Greenwood et al. 1966), although Roberts (1973) has suggested that they arose independently in Africa, Asia and South America. While the Ariidae are predominantly a marine family, there is evidence to suggest that they also evolved in fresh water (Lundberg 1975a in Grande & Lundberg 1988). While the area of greatest ariid radiation appears to be the Indo-Pacific Archipelago (Wongratana et al. 1974), much of their present day distribution is probably attributable to secondary radiation which would have been facilitated by their marine habit. Relatively little is known about the evolutionary status of the Ariidae and it has been suggested that they lie at the base of siluroid evolution along with the Diplomystidae (Regan 1911; Shelden 1937; Gosline 1944 in Hassur 1970; Alexander 1965). More recent studies have demonstrated that 1 they have attained an advanced form and that approximately half of all catfish fami l ies are more generalised than the ariids (Tilak 1965; Greenwood et al. 1966; Hassur 1970). Phylogenetic relationships of siluroid fishes have traditionally been based on skeletal structures such as cranial morphology, the Weber ian apparatus and the associated swim-bladder, the pectoral girdle and pelvic girdle and the caudal fin skeleton (e . g. Shelden 1937; Alexander 1965; Tilak 1965, 1971 ; Chardon 1968; Jayaram & Singh 1984; Howes 1985; Grande & Lundberg 1988) . The results of these studies have, however, been inconsistent and largely inconclusive. Recent karyological studies (LeGrande 1980, 1981; Fitzsimons et al. 1988) concur with Gosline's (1975) findings (which are based largely on osteology), that the Ariidae, along with the Bagridae, Doradidae, Ictaluridae and Pimelodidae, form a group which reflects the ancestral condition from which living catfishes evolved. stress, however, that their interpretations karyological data are largely speculative, They do of the and their conclusions tentative. Gosline (1975) has argued that the high incidence of parallel evolution wi thin the siluroid group will foil any attempts to accurately reconstruct catfish phylogeny. Striking features common to all ariids are their mouthbrooding habit and extremely low fecundity (Rimmer & Merrick op. cit . ), while the size of their eggs is unsurpassed amongst the teleosts. These reproductive characteristics would place them near the K, or precocial extreme, of the r/K (MacArthur & Wilson ,1 967) or a1tricial/precocial (Balon 1979) life-history style continua . This study of two ariid species occurring along the South African south-east coast therefore provided an opportunity to test, amongst other hypotheses, the predictability of the life-history model assumptions with respect to biological traits associated with K-strategist or precocial animals (slow growth, longevity, large average body size, low natural mortality 2 and vulnerability to sources of unnatural mortality such as fishing), (Adams 1980) . The two South African ariid species, Galeichthys feliceps Valenciennes and ~ ater Castelnau are endemic to the coasts of South Africa and Namibia where they frequent shallow coastal waters down to a depth of approximately 60 meters. The colloquial common name for both species is barbel, although G. ater, because of its muddy colouration, is also known as "vuiljassie" (; dirty jacket). The existing distributional records for the two species in the literature are somewhat inaccurate. Taylor (in Smith & Heemstra 1986, p.213) records the distribution of estuaries and rivers from Walvis distribution of ~ ~ feliceps as 'Sea, Bay to Natal'. The fe1iceps in rivers is restricted by their intolerance of salinities below 8ppt. (Whi tfie!d et al. 1981; Bennett 1985) . The barbel occurring off the coasts of Transkei and Natal grow to a large size and do not enter estuaries, and it is thought that they may represent a distinct Galeichthys Institute of limit of ~ species Ichthyology, (P. pers . C. Heemstra, comm.). JLB Smith The true eastern feliceps distribution therefore appears to be in the vicinity of East London. The distribution of Galeichthys ater is recorded as 'South coast to Port Alfred' (Taylor op. cit . p.212), although specimens have recently been collected off the Namibian coast as far north as Swakopmund (pers. obs.). Galeichthys ater and feliceps appear to share a ~ similar distribution between Swakopmund in the west and East London in the east. Estuaries play an important role in the life-cycle of G. feliceps, while G. ater are exclusively marine . The two species have previously been known under three generic names, Arius Valenciennes, Galeichthys Cuvier and Valenciennes, and Tachysurus Lacepede. In a recent revision of the South African ariids, Taylor (1986) reinstated Galeichthys as the valid generic name. Wheeler & Baddokwaya 3 (1981) demonstrated that Tachysurus was not an ariid catfish since it had a rayed adipose fin, nasal barbels and only one nostril on each side of the head . The genus Arius differs from Galeichthys in having seven anterior fused vertebrae (five in Galeichthys), and an epiotic lamina broadly fused with the transverse process of the 4th fused vertebra (remote from fused vertebral complex and epiotic not forming part of the skull roof in Galeichthys), (Taylor op. cit . ) . While three species, namely ater Castelnau, G. feliceps ~ Valenciennes and G. ocellatus Gilchrist and Thompson were originally described, G. ocellatus was subsequently found to be a junior synonym of feliceps. ~ Prior to this study little was known about the biology of ei ther species. Some information stomach content analyses had and aspects of estuarine occurring Coetzee Pool 1984, Marais 1984, Marais & ~ feliceps been published on reproduction in & Baird (Marais & 1980, Venter 1987). However, nothing was known about ~ feliceps in the marine environment, and Galeichthys ater had not been studied at all. This study was originally motivated by the presence of the two species in the catches of the commercial linefishery at Port Alfred. It was evident that, in terms of reproduction, they were vastly different from the other species in the fishery, the majority of which were highly fecund broadcast spawners. The two ariids are mouth-brooders with very low fecundity . A preliminary assessment of the fishery during 1984 demonstrated that while G. feliceps and G. ater represented what was essentially a by-catch, they nevertheless formed a substantial component of the annual catches at Port Alfred (Hecht & Tilney 1989). The question immediately arose as to how they were able to withstand the same rate. of exploitation as the more fecund target species in the fishery . presumed Significantly, highly K-selected and notwithstanding their life-history style, 4 they had continued to persist in catches long after some of the more fecund sparid species had virtually disappeared through over-fishing. As a result of this finding it seemed obvious that further biological and ecological investigation, in conjunction with a study of the dynamics of the two populations, would serve to enhance our understanding as to how highly precocial or K-selected fish species respond to exploitation. This study represents an investigation into the biology and ecology of ~ feliceps and ~ ater off the south-east coast of South Africa, in the vicinity of Port Alfred (Fig. 1). Aspects investigated in detail were reproduction, ontogenetic development, growth rate and population age structure, population dynamics and resource partitioning. The last was prompted by the similar overall gross morphology and simultaneous presence in the commercial catches of the two species, which suggested a similar habitat preference. As logistic obstacles precluded in situ investigation of their micro-habitat preferences or of the degree of inter-specific competition for essential resources, a feeding study was undertaken to investigate the nature of resource partitioning between them. The thesis is structured in the following way. Chapter 2 is an investigation into feeding and resource partitioning, Chapter 3 deals with reproduction and Chapter 4 is a study of the early ontogeny and mouth-brooding behaviour of ~ feliceps. The age and growth and population dynamics of the two species are explored in Chapters 5 & 6, and the thesis concludes with a general discussion (Chapter 7). 5 sr-. - -~ - _, N 30' 1 o INDIAN OCEAN 10 km 20 27" Figure 1 : Study area . Sampling was conducted primari ly on the commercial fishing quay in Port Alfred on a near-daily basis . The fishery operated between Great Fish Point in the east and Cape Padrone in the west . Estuarine samples were taken from the Kowie River, while some feeding data from the Great Fish and Mtati Rivers were also incorporated. Data from samples collected during a small-mesh trawl survey of Algoa Bay (Buxton et al. 1984) were incorporated, while feeding and growth information for ~ ater juveniles were obtained from rotenone samples taken between Shelly Bay and Chelsea Point west of Port Elizabeth (Smale & Buxton 1989). 6 E CHAPTER 2 - FEEDING Introduction Resource partitioning is widely accepted as being the prime mechanism allowing the co-existence of species assemblages and communities and it is evident that species commonly partition one of three resources, namely space, food or time of resource utilisation (see review by Schoener 1974). Since Galeichthys feliceps and G. ater could not be observed in situ, it was hoped that detailed feeding studies would provide an insight into the degree of interspecific interaction between them in their natural similar environment . morphologically it As was the two species were highly envisaged that they would compete for the available resources and that these would be partitioned between them. An attempt at integrating functional morphology and ecology was made by investigating the feeding associated morphologies of the two species. Aspects such as dentition, gape size, gill raker morphology, lateral line complexity, olfactory rosette structure and the function of the circum-oral barbels were explored. The aims of this feeding study were therefore to ascertain the micro-habi tats, diets, and feeding strategies of the two species, to investigate whether resource partitioning occurred, and if so, to establish the nature of resource partitioning between them. The feeding of G. feliceps has previously been studied in the Swartvlei estuarine system (Coetzee & Pool 1984) and the Gamtoos, Krom, Sundays and Swartkops estuaries (Marais 1984). The general feeding habits of G. feliceps have also been described by Day et al. (1981), Smith & Heemstra (1986) and Van der Elst (1981), but prior to this study (Tilney & Hecht 1990) 7 no published work existed on the feeding habits of G. ater, or of ~ feliceps in the marine environment. Elsewhere in the world, recent ariid feeding studies have been publ i shed on Galeichthys caerulescens (Yanez-Arancibia 1977) in Mexican coastal lagoons, Tachysurus tenuispinis (Mojumder & Dan 1979) in Indian coastal waters, Netuma barba, ~ planifrons and Genidens genidens (Araujo 1984) and Netuma barba (Reis 1986a), thalassinus revealed in (Euzen that all Brazilian 1987) coastal lagoons, in the Persian Gulf. ariids are benthic and The predators primarily over soft substrata on crustaceans, molluscs, fish and in some instances, detritus. Arius studies feeding polychaetes, Materials & Methods Galeichthys feliceps (FL 232-360mm) and ~ ater (FL 195-322mm) were collected from the catches of the Port Alfred commercial ski-boat fishery. The fishing grounds extend from Great Fish Point in the east to Cape Padrone in the west (Fig. 1). Samples were collected on a monthly basis between March 1984 and March 1987. Stomachs of juvenile ~ feliceps (FL 46-145mm) from the marine environment were obtained from the collection made on the R. V . Thomas !L.. Davie during an inshore small mesh trawl survey of the Cape south coast during 1982 1984). Stomach content data for G. ater 127mm) were obtained from rotenone (Buxton et al. juveniles collections (FL 48 - made in sub-tidal gullies between Schoenmakerskop and Marine Drive, Port Elizabeth (Smale &, Buxton, 1989) between September 1984 and September 1985. To determine the extent of feeding in estuaries by G. feliceps, gill nets were deployed in the Kowie River on a monthly basis between March 1984 and March 1987. Bottom-set gill nets (25, 40, 60, 75, 100 & 150 mm stretched mesh) were used at three localities at approximately 3.5, 4.5 and 5.5 km from the mouth 8 of the estuary. Gill netting was conducted at neap tides and preliminary trials revealed that the most effective sampling period was between dusk and dawn. Nets were checked and fish removed every two hours in order to prevent predation by the mangrove crab (Scylla serrata), cuttlefish (Sepia officinalis vermiculata), and isopods (Exosphaeroma was injected into the abdominal ~) . cavities of Formalin (10%) fish to halt digestion of stomach contents. Juvenile feliceps ~ (FL 45-5Smm) smaller than those caught with gill nets were captured at night in shallow marginal areas of the Kowie river between December and March using a cast net and a search light. The feeding behaviour of G.feliceps eleuthero-embryos during the buccal incubation phase was investigated by removal of young from mouth-brooding adults, caught in the gill nets. Feeding data for ~ feliceps (FL 47 - 38Smm) collected from the Great Fish and Mtati estuaries between April and July 1983 were also incorporated. The material was collected by the late B. E. Trowe of the Department of Ichythyology & Fisheries Science, Rhodes University. The entire digestive tract was removed and preserved in 10% formalin shortly after sampling . During analysis each stomach and intestine between 0 was designated a subjective (empty) and 5 (maximum distension), fullness index and intestine length was measured between the pyloric sphincter and the vent . By making a longitudinal incision of the stomach to expose the contents, large prey items were removed using forceps, while small food particles and remains were flushed into a petri-dish using water. taxa under Stomach contents were examined and sorted into a binocular microscope. Intact prey items were identified to species level and prey remains to the lowest taxon possible . 9 Stomach contents were analyzed using four methods: a.) Numerical occurrence: The number of individuals of each prey type in all stomachs was expressed as a percentage of the total number recorded. b.) Frequency of occurrence: The number of stomachs in which each prey item occurred was expressed as a percentage of the total number of stomachs in the sample. c.) Gravimetric method: Prey were blotted and weighed before drying to constant mass at 60°c to determine the wet : dry weight ratio (Hynes 1950; Hyslop 1980). d . ) Energy contribution: Published energy values for individual taxa (Table 1) were used to convert gravimetric data into energy content in kilo joules per gram (kj/g). The energy contribution of each food type for all stomachs was expressed as a percentage of the total energy (Whitfield 1980). Since the frequency with which a particular prey organism occurs in the diet is a measure of a degree of dependence by the predator on that organism, the product of the percentage frequency of occurrence (%FO) and percentage energy (%E) values was expressed as an index, termed the Energy Index (EI), as follows: EI = %FO x %E Vacuity indices based on the percentage ratio of empty stomachs to the total number of stomachs in the sample were calculated. A low vacuity index is an indication of a voracious predator (Euzen 1987) . Feeding seasonality was investigated in the marine samples. Data for each quarter were grouped according to season and sex. The seasonal trend of stomach fullness was similarly determined using the stomach fullness indices. Statistical methods were used in an attempt to quantify diet overlap between the two species in the marine environment . Measurements of niche or resource 10 overlaps have been demonstrated by several authors using different techniques, namely indices overlap chi-squared test in (Schoener combination 1970), the with overlap Cochran's measure of Morisita (1959 as modified by Horn 1966 in Zaret & Rand 1971) and Spearman rank correlation coefficients in conjunction with t-tests (Fritz 1974). Table I. Wet:dry weight ratios and energy values of food taxa used in the construction of Energy Indices . References Dry WI. Energy (%) (Kj/g) 26 14.47 Whitfield (1980) Calianassa kraussi 21 21 .24 Whitfield (1980) Uf!0gebia africana 30 14.49 Hanekom (1980) 25 13.30 Whitfield (1980) Isopoda 28 14.62 Whitfield (1980) larvae 33 17.94 Blaber (1979) Macrura 17 20.30 Whitfield (1980) Mysidacea 11 19.44 Cockroft & Mclachlan (1987) 18 21.70 This study Taxa Crustacea Amphipoda Anomura Brachyura Cockroft & Mclachlan (1987) Echiurida Ochaetostoma c8t!ense Mollusca Cephalopoda 22 22.58 Buchan & Smale (1981) Pelecypoda - flesh 7 20.28 Berry (1978) Mucus 5 15.05 Gorlick (1980) Polychaeta 28 18.57 Whitfield (1980) Teleostei Scales 18 24.40 Whitfield (1980) 31 8.52 Whitfield & Blaber (1978) Since stomach content analyses revealed that 61% of .2..... feliceps juveniles between 54-225 rom (FL) had scales in their stomachs to the exclusion of any other teleost remains, two experiments were conducted in order to ascertain whether .2..... feliceps juveniles are lepidophagous or whether the scales were 11 scavenged. Experimental procedure was adapted from Whitfield & Blaber (1978) . Experiment 1. Four ~ feliceps juveniles (FL 65-85mm) were starved for 48 hours and put into an aerated 50 litre glass aquarium . Scales (n=125) removed from living Liza richardsonii, Rhabdosargus holubi and pomadasys commersonni were introduced. After one hour the remaining scales were counted . Experiment 2. Three individuals from each of the above mentioned species (FL 80-150mm) were placed into a second tank together with four similarly starved ~ feliceps juveniles . After 120 hours the ~ feliceps digestive tracts were examined for the presence or absence of scales. In an attempt to assess the relative importance of sight, smell and taste in the location of food, aquarium observations were a conducted series of preliminary using G. feliceps juveniles. Individuals which had been deprived of either their sight, smell or taste were placed into separate, 100 litre aquaria . An untreated animal held in a fourth aquarium was used as a control. Food-finding efficiency was measured by recording the time taken to locate and ingest a 3mm 2 particle of food introduced into the water at a point distant from the animal . A 15 minute interval between successive tests was used to ensure that the fish were not engaged in appetitive searching behaviour prior to food introduction . Morphological adaptations for feeding were investigated as follows: The dentition and circum-oral barbels of G. feliceps and ~ ater were examined using scanning electron microscopy. Procedures according to Cross (1985) for the preparation and processing of material for electron microscopy were followed. Gill raker size and spacing were measured using vernier calipers while the structure of the olfactory rosettes and the extent of the lateral line were determined using a binocular 12 microscope. Stomach pH of f reshly sacrificed and dissected animals was measured using Merck pH indicator strips (pH 0- 14) . Results Anatomical adaptations for feeding Galeichthys species are equipped with three pairs of circumoral barbels, one maxillary and two dentary. They are capable of a certain degree of movement and can be held erect or laid flat against the body . In the erect position, mandibular barbels are held at 90° to the body. the four The longer outer pair are projected laterally at an angle of approximately 25°, so that when searching for food, the tips of all four mandibular barbels are trailed across the substratum . The long maxillary barbels are extended slightly anteriorly and at an angle of approximately 45° from the vertical (Fig. 2a & b) . The tips of the maxillary barbels do not touch the substratum during foraging activity and presumably serve to detect prey that has moved off the bottom. Taste buds constitute a large proportion of the surface area of the barbels of both species (Griffiths 1984; Andrew 1987 ) , (Plate Ia & b), and were found to be most abundant distally, on the leading edges of the barbels . The density of taste buds differed between the two species. Galeichthys feliceps barbels held approximately 90/rom 2 and Q..... ater 40/rom 2 (Andrew 1987). The structure of the taste buds was identical in the two species and was similar to those found on other benthic feeding fishes, e.g . flatfishes (Livingston 1987). The taste buds were raised, oval structures covered with short microvilli around the periphery and longer microvilli toward the centre (Plate Ic & d) . 13 (a) ds 20mm mx f..-md o-md (b) mx Fiqure 2. (a) Head-on and (b) lateral view of Galeichthys demonstrating the positions of the maxillary (rnx), inner (imd) and outer (o-md) mandibular barbels during foraging . ds=Dorsal spine; ps=pectoral spine . The mouth is slightly subterminal, of intermediate size, and has fleshy lips. During prey ingestion the body is orientated at approximately 35° from the horizontal, the angle required to bring the mouth into contact with the substratum. Aquarium observations of feeding animals demonstrated that prey was often detected using the mandibular barbels which were trailed over the substratum behind the mouth. When this occurred, the animal came to an immediate halt by abducting the pectoral fins, which also acted as a pivot about which the body tilted in order to assume the feeding posture described above. 14 Plate I. (a) Taste bud (tb) distribution on circum-oral barbels of G. feliceps and (b) G. ater. (c) Taste bud structure in ~ feliceps and (d) G. ater illustrating their similarity. sv=Short villi; lv=long villi. While the tail moved upwards and slightly anteriorly, the head moved downwards and slightly posteriorly, resulting in the mouth being positioned directly over the detected prey item. 15 Ingestion of benthic prey occurred while the animal was stationary. The suck and grasp method of prey capture common to many modern siluroids (Gosline 1973) was employed, in which a lowering of the hyoid complex sets up a negative pressure in the vicinity of the prey when the mouth is opened (Alexander 1970; Lauder & Clarke 1984; Osse et al. 1985 and Van Leeuwen & Miller 1985) . This results in a net inflow of water into the buccopharyngeal cavity, carrying the prey with it into the mouth . The oesophagus is short and distensible, leading to a muscular J-shaped stomach. The stomach environment is acidic, ranging between 2 . 0 and 6.0 in feeding G. feliceps. the pH Feeding ceases during mouth brooding (see Chapter 4), and stomach pH varies between 8.0 and 9.0 in these individuals. The anteriorly directed pyloric sphincter leads into a thin-walled intestine. The mean ratio of body length (FL) to intestine length for all size classes is 1:1 . 78 (±0.24) for k:.. fe1iceps and 1:1.79 (±0 . 21) for G.ater. Since very little processing of prey occurs within the buccal cavity, the teeth are small and serve to grasp prey before it is swallowed whole. The tooth structure was similar for the two species (Plate IIa & b) . Several rows of small, recurved, villiform teeth occur on the dentary, pre-maxilla, palate and on two pharyngeal plates on the roof of the mouth. The gill rakers, which are short and stout (mean length 2.97mm ±1.37mm), occur on both anterior and posterior margins of the gill arches and on the posterior margin of the pharyngeal fourth gill surface arch . opposite the Excurrent water expelled between the first gill arch and the operculum is strained through proportionately longer gill rakers found on the anterior margin of this gill arch. Gill rakers of adjacent gill arches are interdigitating, resulting in an effective, but inefficient, low sieve-potential filtering mechanism (Lagler et al. 1977). 16 Plate II. Scanning electron micrograph of the maxillary teeth of (a) G. feliceps and (b) G. ater juvenile showing the similarity in structure. It is likely that they serve primarily to protect the delicate respiratory ingested, lamellae live food. against The abrasion mean gap and size damage between by the opposing, interdigitating gill rakers is O.23mm (±O.22mm) in adult ~ feliceps, enabling the filtration of larger zooplankton (e . g. mysids and macruran shrimps). 17 The oval nasal c hambers are divided into two nares by a dividing bridge of skin, which when flattened completely covers the posterior naris and restri cts water flow into the olfactory chamber . The septum is supported by a cartilaginous rod and may be raised to form a ridge that directs water into the anterior naris, through the nasal chamber and out via the posterior naris. The numbers of sensory lamellae forming the olfactory rosettes ranges between 10 in juveniles and 60 in adults of both species and are accompanied by a change in shape of the olfactory rosettes from ovoid to elongate . The laterally positioned eyes are of moderate size, ranging between 4.9% (±0.12%) and 3.4% (±0.1%) of total body length in juvenile and adult ~ feliceps respectively . In ~ ater the eyes are 4.4% (±0.23%) of total body length in juveniles and 3 . 7% (±O. 24%) in adults. They are capable of lateral and vertical movement within their sockets, affording a wide field of periscopical and a certain degree of binocular vision. Galeichthys are relatively photophobic and their vision in daylight appears to be poor, as suggested by their inability to detect live prey from a distance of more than a few centimeters (pers. obs.). While their sight is probably more effective at lower light ,intensities, the extent to which it i s used in prey capture is unknown. The lateral line between the caudal peduncle and the pectoral girdle gives rise to regularly spaced, oblique dorsal and ventral branches, of which the latter are longer and may be branched. In the region anterior to the pectoral girdle the lateral line branches extensively, dorsally over the nape area, ventrally over the cleithral region and anteriorly onto the head. The dense network on the head extends over the opercular, infra-orbital and cranial areas extending anteriorly to the upper lip. The lateral line system increases in extent with growth, reaching maximum complexity at approximately the size of sexual maturity (Fig . 3) . It is not known whether Ampullae of Lorenzini are present in the lateral line system, although 18 they are thought to occur in clariids (Lissman & Machin 1963) and plotosids (Bullock 1973). 20mm Figure 3. Diagrammatic reconstruction of the lateral line system of Galeichthys. hll=Head lateral line; tll=trunk lateral line. The caudal fin structure differed in the two species and may be a reflection of different mobility (speed) requirements during foraging. In G. feliceps the caudal fin is deeply forked with the upper lobe being longer than the lower, while G. ater has a feebly forked caudal, verging on emarginate (Fig. 4). 19 ad c cp an 20 mm ad Q. alec cp c an Figure 4. A comparison of the caudal fin structure between similar sized G. feliceps and G. ater. ad=Adipose fin; an=anal fin; c=caudal fin; cp-caudal peduncle . Dietz Offshore samples The principle taxa in the diet of G. feliceps were brachyura (49% of total polychaeta energy (17%), in (Fig. the diet), echiurida (27%) and Sa). Two species of crab namely Thaumastoplax spiralis and Goneplax angulata made up the bulk of the crustacean component . A single species of echiurid, Ochaetostoma capense occurred in the diet. Day (1974) remarked that T. spiralis and and share a capense have a commensal relationship ~ common burrow, explaining the co-occurrence of these two animals in G. feliceps stomachs. The most commonly consumed polychaete was the sedentary Sternaspsis scutata (Appendix Ia) . Galeichthys dwellers. feliceps The crabs ~ prey exclusively angulata and L on soft spiralis burrowers (Barnard 1950; Day 1974), O. capense and 20 substratum ~ are sand scutata burrow in mud (Day 1967, 1974) and the bivalve Phaxas decipiens burrows in sandy and muddy substrata (Kilburn & Rippey 1982). The most important taxa in the diet of G. ater were brachyurans, isopods, polychaets and cephalopods (Fig. 5b). The Cape rock crab Plagusia chabrus and it's megalopae were the most common brachyuran in the diet, while Dehaanius dentatus and Macropodia falcifera were frequently encountered. The anomuran Galathea dispersa and the palinuran Scyllarides elizabethae (juveniles) were also important. A number of large chelae, probably the result of tackling crabs too large to ingest, also occurred. The polychaetes were generally difficult to identify as a consequence of their rapid digestion in the stomach. Most, however, belonged to the group errantia, and included Eunice Lysidice ~, natalense and Platynereis dumerilii. The sedentary Pherussa .!ill..:.. was also well represented. Juvenile Octopus vulgaris were the most abundant mollusc in especially ~ the diet, while the Cymodose ~ isopods, valida, and caridean shrimps were other important prey items (Appendix Ib) . In contrast to ~ both reef- and chabrus reef feliceps, the prey of ~ soft substratum-dwelling (5.5% of dietary energy) and dwellers, while Q.... capense dispersa (2.9%), the palinuran ~ ~ (5 %) ater consisted of species. Plagusia dentatus (2.8%) are occurs in mud. ~ elisabethae (1.5%) and the various isopods (10.3%), polychaetes (12.8%) and cephalopods (12.8%) occur over both hard and soft substrata (Day et al. 1970; C . Buxton, Dept. Ichthyology & Fisheries Science, Rhodes University, pers . comm.). Prey common to both species included Q.... capense, four crabs ~ angulata, !L.. orbiculare, Phi lyra punctata and Atelecyclus septemdentatus, an isopod Synidotea hirtipes, the cephalopod O. vulgaris and a polychaete Pherussa .!ill..:... While there were only eight common species in their diets, they shared 13 out of a total of 20 food taxa. 21 (a) , ISOP _ ~D o-A~ MY,SIDACEA ANOMURA TELEOSTEI ' CEPHALOPODA MACRURA (b) - - - - - -GASTROPODA Figure 5. Diets in the marine environment of (a) G. feliceps adults _ Wedges represent percentage energy contribution of prey_ FL=232-380mm, n=402_ (b) G. ater adults_ FL=185-322mm, n=285. The diet of ~ feliceps species, namely ~ was narrow, with 80% comprising spiralis (23.3%), ~ angulata (19.7%), capense (27%) and the sedentary polychaete ~ ater on the other hand fed more four ~ ~ scutata (9.3%)_ broadly, the largest contribution by a single species being that of P. chabrus (5_5%). The relative numbers of prey species taken by the two species are presented in Table 11_ 22 Table II. A comparison of the number of species in the diet of G. feliceps and G. ater sampled in the marine environment. G. atar G. feliceEs Common No. of species No. of species Taxon No. of species Brachyura 8 1 15 4 Echiurida 1 1 Isopoda Mollusca Polychaeta 2 3 6 22 6 Total 20 49 5 ----- 8 -------- The data obtained from the R. V. Thomas B . Davie inshore trawl survey showed that ~ feliceps juveniles in this environment fed primarily on ~ capense (Fig. 6a). Thaumastoplax spiralis, cephalopods, anomurans, mysids, teleost scales, polychaetes and amphipods made up the rest of the diet . Copious amounts of • mucus were also present in many stomachs (Appendix IC). The origin of this material is unclear, although the possibility exists that juveniles may remove mucus from the surface of other fish. ~ Sub-tidal gully samples ater juveniles inhabiting sub-tidal r e efs fed largely on amphipods, isopods and polychaetes (Fig. 6b). The large majority of amphipods were from the sub-order gammaridea, while two isopods, namely Munna sheltoni and Parisocladus perforatus were important. Most of the polychaetes in the diet were from the group errantia. As with of mucus were found in the feliceps, substantial amounts ~ stomachs (Appendix Id). 23 of ~ ater juveniles (0 ) TELEOST SCALES ECHIURIDA ~-tf@.; ~AMPHIOD POLVr" "o-TO (b) MACRURA Figure 6. Diets in the marine environment of (a) G. feliceps juveniles. FL;47-175mm, n;48. (b) G. ater juveniles. FL;81121mm, n;41. Estuarine samples While hook and line surveys revealed that adult ~ feliceps of both sexes frequent estuaries within 1-2 km of the mouth throughout the year, the gill net survey conducted in the Kowie estuary between 2.5-5.5 km upstream revealed that sexually mature G. feliceps do not forage upriver. A large component of the gill net sample did however consist of mouth-brooding males, which utilise this environment as a refuge between September and March . Stomachs from all mouth-brooding fishes (n;250) were empty , containing only traces of greenish bile-stained mucus . In addition, the diameter of their digestive tracts was considerably reduced, indicating a 24 complete cessation of feeding activity during the oral incubation phase (see also Chapter 4). In contrast to adults, juvenile G. feliceps were found to forage in estuaries throughout the year. component in the diet of juvenile ~ The most important feliceps in the Kowie estuary was the burrowing mudprawn U. africana (Fig. 7). Three species of crab, ~ edwardsii, orbiculare and ~ spiralis ~ made up the brachyuran component, Exosphaeroma hylocoetes and .!l<...... truncatitelson the isopod component, while the teleosts were represented by Atherina breviceps and unidentified remains (Appendix Ie). MACRURA U. africana MYSIDACEA ... ......... ~ .k rausi SPERMATOPHYTA MOLLUSCA Figure 7. Diet of G. feliceps estuary. FL=125-215mm, n=112. juveniles in the Kowie River In the Great Fish River estuary, data accumulated over a four month period between April and July 19B3 revealed that both juvenile and adult G. feliceps foraged in this estuary within 1-2 km from the river mouth . Sand prawn, crabs, amphipods and teleost scales were important in the diet of juveniles (Fig . Ba) . Other taxa in their diet were mysids , polychaetes, macruran shrimps and prawns, isopods and the mud prawn !L.. 25 africana. The crabs in the diet included C. edwardsii and orbiculare (Appendix If). The adults consumed mainly and teleosts (Fig. 8b). The mud prawn U. ~ africana, ~ kraussi crabs, penaeid prawns and mysid shrimps were also present (Appendix Ig) . (a) C. krau .. 1 MVSIDACEA U. atrlcan. 180PODA BRACHYURA ...... .!-'ACRURA TELEOST SCALES AMPHIPODA (b) c .• , •• ,,' --- MY81DACEA - -- Figure 8. Diets of G. feliceps in the Fish River estuary. (a) Juveniles FL=45-180mm, n=46. (b) Adults FL=255-385mm, n=23. Only juvenile G. feliceps were encountered in the Mtati River estuary which is closed off from the sea, opening for a brief period (2-3 weeks) during spring each year. They fed mainly on isopods, teleosts and teleost scales (Fig. 9). Sand prawn, 26 ~ kraussi, crabs, amphipods and Zostera ~ also occurred (Appendix Ig). Sample sizes for the Fish and Mtati Rivers were small, but yielded results consistent with those from the Kowie River and the literature (Coetzee & Pool 1984; Marais 1984). AMPHIPODA TELEOST 8CALEB Figure 9. Diet of G. FL;17S-214mm, n;38. feliceps in the Mtati River estuary. Eleuthero-embryo sample contained mucus and stomachs of eleuthero-embryos The detritivorous material. The latter comprised spermatophyte remains, crustacean exoskeleta, poriferan spicules, mucus and sand (Appendix Ih). Gravimetric data was used to express the diet graphically (Fig. 10). Ga1eichthys feliceps do not release their young to feed during the incubation phase (see Chapter 4), as is the case in some mouth brooding cichlids (Fryer & 11es 1972). Instead, the adult picks up detritus from the substratum and the young feed within the buccal cavity. This behaviour was presence of confirmed yolk and during teleost 27 aquarium remains in observations. The eleuthero-embryo stomachs indicate that a degree of sibling cannibalism may occur wi thin the adult buccal cavity. The eleuthero-embryos may als o feed on adult buccal mucous during the mouth-brooding phase . EXOSKELETA REMAINS WOODY REMAINS Figure 10. Diet of G. feliceos free-embryos during the mouthbrooding phase. FL=38-45mm, n=57. Feedinq seasonality The seasonal contribution of important prey in the diets of feliceps and G. ater are presented seasonal variation in mean stomach in Figure fullness 11, is and ~ the plotted in Figure 12. The seasonal variation in food composition was similar for both sexes in G. feliceps. Brachyura were important all year round, being most well represented ~ n summer and spring. Polychaetes were abundant in autumn and echiurids in winter. Molluscs were a relatively minor represented in component summer. In in their ater, ~ diet but brachyura, were well isopoda and polychaeta were important throughout the year, with polychaetes being particularly well represented in spring. Cephalopods, teleosts and macrura were less important in the diet, with no seasonal pattern . 28 ., '0 c: 7 6 oS <I) ::J 0 5 .t= !:: x 4 0 3 w ~ >a: w z <!l 2 w 0 if ~ AUTUMN SUMMER _ BRACHYURA ~ SPRING WINTER SEASONS ECHIURIDA CJ MOLLUSCA ~ - --, . ----- POLYCHAETA .- -- . - . .~ 7~- - (bl -, 6 5 x w o z ~ a: w z w if S! SUMMER ~ c3' c3' 9 AUTUMN c3' 9 WINTER SEASONS BRACHYURA ~ ISOPODA POLYCHAETA §..... TELEOSTEI SPRING CJ CJ S! CEPHA LOPODA MACRURA Figure 11. Feeding seasonality in (al G. feliceps and (b) ater males and females in the marine environment . ~ There was little seasonal variation in mean stomach fullness in G. feliceps. The fullness indices were slightly higher in summer and autumn, while males had fuller stomachs in summer and winter than did females. The maximum stomach fullness recorded, expressed gravimetrically as a percentage of total body mass, was 4.05% . The maximum stomach fullness value recorded for G. ater was 4 . 01%. The mean stomach fullness indices remained constant throughout the year . Male stomachs 29 5 >< w 0 z en en w z (a) 4 3 ..J ..J => LL :z: 2 U < ~ 0 1 f- en 0 5 >< w 0 z en en w z ..J SUMMER AUTUMN WINTER SEASONS SPRING SUMMER WINTER AUTUMN SEASONS SPRING b) 4 3 ..J => LL :z: 2 U < ~ 0 1 f- en 0 _ MALE ~ FEMALE Figure 12. Seasonal stomach fullness indices of (a) G. feliceps and (b) G. ater males and females in the marine environment. were fuller than female stomachs in spring. While G. feliceps stomachs were generally fuller than those of ~ ater throughout the year, the stomach vacuity indices of 14% and 11% for the two species respectively indicate that they fed at similar intensities. Comparisons of stomachs with food vs . 30 empty stomachs for the two species in the marine environment revealed significant seasonal differences for G. feliceps (X2 = 20.94, df = 3, P < .01 ), in which stomachs with food occurred most frequently in summer (95.5 %), followed by spring (92 . 6%), autumn (86.2%) and winter (73.1%). There were no significant seasonal differences at the 5% level for (X2 ater ~ = 2.66), in which stomachs with food occurred as follows: Summer (90.5%), spring (86.2%), autumn (82.1%) and winter (90.6%) . In the marine environment there was a marked shift in dietary preference with growth between juveniles and adults. The bulk of the juvenile ~ feliceps diet consisted primarily of echiurid worms (Fig. 6a), while in adults, crabs, polychaetes and echiurids were all major food items (Fig. Sa). Amphipods were an important dietary component in ~ ater juveniles (Fig . 6b). In adults they were absent, being replaced by crabs and cephalopods (Fig. 5b). In estuaries the diets of juveniles and adults were similar (Figs . 7, 8 & 9), although the sample of adults in this environment was small. The lepidophagy question Experiment 1. The ~ feliceps juveniles responded immediately to the introduced scales by engaging in appetitive searching . After one hour the experimental animals had reached satiation and 89% of the scales had been consumed. Experiment 2. All four G. feliceps digestive tracts were empty after 120 hours. Visual observations indicated that their behaviour appeared to be unaffected by the introduced fish. The importance of sight, smell and taste in food location a) The untreated control animals responded to introduced food after a few seconds by going into an arousal phase (Kleerekoper 1982) of rapid undirected locomotion until the food was located. b) In the absence of olfactory perception, animals failed to be aroused into a searching mode unless the food was introduced 31 into their immediate environment, suggesting that olfaction is necessary for the perception of distant stimuli in c) ~ feliceps. The removal of maxillary and mandibular barbels appeared to delay the final pin-pointing and ingestion of food particles by searching animals and suggested that barbels are used in the location of food in the immediate environment. d) Animals deprived of vision were more successful at finding food than control animals, suggesting that daylight may act as an inhibitor to feeding. Discussion Feeding morphology The specialised sensory structures of siluroids are believed to have evolved largely as a result of their nocturnal, benthic habits (Gosline 1973). The elaborate network of head lateral line canals, well developed olfactory rosettes and the three pairs of taste and touch sensitive circum oral barbels in Galeichthys constitute an effective nocturnal detection system for epi- and endobiontic prey (terminology after Peres 1982). Bardach & Atema (1971) mechano- as well as a suggest that fish taste buds have a chemosensory function. Lagler et al. (1977) state that touch sensitivity in silurid barbels arises from the nerves innervating the taste buds, which radiate to the surrounding regions and render them touch sensitive. The senses of taste and touch in Galeichthys may therefore work in unison in the selection of food before ingestion. Touch sensitivity would also compensate for the total or partial loss of vision during nocturnal foraging, or foraging in turbid environments. While the number and size of the olfactory lamellae in fish increase with age (Kleerekoper 1969), until a certain stage of growth is reached (Pfeiffer 1963, 1965 in Yamamoto 1982), the relationship between the number 32 of olfactory lamellae and olfactory acuity is unclear (Pipping 1926, 1927; Wunder 1957 in Yamomoto 1982) . Bond (1979) states that fish with elongated rosettes are thought to have acute powers of olfaction. The presence of an alarm substance in the epidermis of siluroids (Pfeiffer 1962), which is released on injury and serves as an olfactory intraspecific warning device, also suggests that olfaction may be acute in Galeichthys . It is not known whether the movement of water through the nasal chambers is entirely passive and dependent on forward movement of the animal, or whether active mechanisms such as cilia or accessory sacs are present . The former have been reported in ictalurids (Parker 1910, in Caprio 1982), in which olfaction was found to be a distant food-finding sense, more sensitive than the organs of taste. Caprio gustatory (1982) receptors, has since which are shown both that olfactory chemosensory and systems, detect a different but overlapping spectrum of amino acids with high sensitivity . He suggests that amino acids occurring in the prey of catfish might preferentially activate the taste system, while the olfactory sense socially relevant stimuli. is probably Bakhtin more (1976) sensitive to investigated the cellular morphology of the olfactory lining in several teleosts and elasmobranchs and described the mechanisms involved in olfactory reception and memory . The sense of sight in Galeichthys is probably geared towards movement detection at low light intensities, in which any moving object smaller than or equal in size is a visual feeding signal and any larger moving object is a visual defensive signal . Three types of visual cells occur in teleosts, namely rods, single cones and double cones (Protasov 1970). Most pelagic fish have an abundance of rods and double cones for colour vision, while bathypelagic fish tend to have rods only. Ali & Anctil (1976) found rods and single cones in the retinas of five siluroid families indicating an adaptation to low light intensity environments, but with a retention of the ability to distinguish colour . This ability may be important to Galeichthys during daylight hours, when visual cues may be used 33 for adopting cryptic colouration (Protasov 1970). If the premise that eye size parallels the development of the visual capabilities of fish is correct (Protasov 1970), it must be assumed that vision in Galeichthys is fairly well developed, sinc e their eye size approx i mates those of visual reef and soft substratum predators such as Petrus rupestris, Epinephelus guaza and Argyrosomus hololepidotus (Smale 1983) . While their sensory structures would equip them equally well for foraging during daylight hours or in clear water, the two main reasons for their nocturnal habit are probably to catch nocturnally active prey and to avoid predation while foraging (Sih 1987) . The robust, erectable pectoral and dorsal spines fulfill a defensive role by increasing the effective body size although they do not appear to deter predators large enough to accommodate them, and they are preyed upon by a wide variety of predatory species throughout their life-cycle. Galeichthys feliceps range over sandy and muddy substrata and probably rely on cryptic colouration and inactivity to avoid predation during daylight hours, since this environment does not provide physical shelter. They have dark, muddy coloured dorsal and lateral surfaces and lighter ventral surfaces, giving rise to obliterative shading when the inc i dent light is from above. Obliterative shading shadow-like (Thayer's renders objects principle), optically resulting in flat and effective diurnal camouflage (Lagler et al. 1977). Galeichthys ater are darker in colour, have uniformly distributed pigmentation and probably seek refuge in caves and crevices by day. Aquarium observations have shown that both species are inactive during daylight hours . By night they are active and constantly on the move. The obvious differences in caudal fin structure between the species gives advantage of cause this for speculation phenomenon . 34 as to the adaptive Working on the premise that deeply forked caudal fins are related to swinuning speed, although not necessarily to swinuning strength (Bond 1979), they would be advantageous for feliceps ~ because of their extensive foraging areas and the need for predator evasion in a refuge-poor environment. On the other hand, G. ater forage over reefs which are smaller in extent and refuge-rich. They may have capitalised on their more stenotopic existence in this environment by decreasing the size of their caudal fin. Diet, diet lepidophagy Both overlap, feliceps and ~ diet ~ shift and the of ater consumed a wide variety of benthic prey. Their maximum indices of fullness body weight probability respectively), were (4.05% and 4.01% of comparatively low. Marais (1984) reported maximum indices of fullness of 11.5% and 10.1% for feliceps in the Sundays and Swartkops estuaries. These ~ values were recorded during post-flood conditions when low salinities had driven U. africana from their burrows, making them highly accessible to predators. stomach fullness index of In addition, a maximum 18.5% was recorded in the Krom estuary. The low values obtained in this study in the marine environment are either a reflection of low prey densities, or are the result of considerable gut evacuation having occurred prior to capture. The mean indices of stomach vacuity for ~ feliceps of 14% and 3.3% in the marine and estuarine environments respectively, and 11% for a high ~ ater in the marine environment, are indicative of feeding intensity. While Galeichthys are primarily nocturnal feeders, their presence in the Port Alfred conunercial linefishery reveals the opportunistic nature of their feeding behaviour, since fishing occurs during daylight hours only. Diet width is thought to vary in response to factors such as prey quality, habitat productivity and the degree of intra- and interspecific competition (Hughes 1980). While the narrow diet of ~ feliceps could be ascribed to a prey-rich environment, 35 it might equally be a result of intense interspecific competition, particularly from benthic feeding elasmobranchs, of which there are several in the study area . The wider diet of ~ ater can probably be attributed to the greater invertebrate species diversity occurring over reefs (Russell 1982). However, without quantitative data relating to prey abundance, prey selection or the degree of diet overlap between the many co-occurring species in the two environments, the above assumptions remain speculative . Statistical quantification of diet overlap using Spearman rank correlation coefficients and the overlap measure of Schoener (1970) was unsuccessful. The data proved to be unsuitable as a result of the low number of common prey taxa at the species level, while comparisons at higher taxonomic levels would have been meaningless. The diet shift with growth between juveniles and adults in both species probably occurs in response to several factors, including progressive loss of the ability to filter zooplankton, change in gape size enabling ingestion of larger prey taxa and migration from nursery areas. Hoese (1966) reported scale and fin biting in juvenile Galeichthys felis Linnaeus (synonymous with Arius felis) in the Gulf of Mexico, and suggested that while this type of feeding was relatively minor, the ingested mucus might provide an important auxiliary energy source. Aquarium observations during the present study revealed that, on release from the adult buccal cavity after termination of the incubation phase, ~ feliceps juveniles frequently bit the body surface and fins of adults in the tank. In addition to the energy gained in this way, it has been suggested that juvenile fish might gain ~ immunological benefits from this habit (Hildemann 1962). feliceps juveniles were also observed feeding on fins and protruding faecal matter of Liza richardsonii under aquarium 36 conditions . This behaviour appeared to occur in response to olfactory or gustatory stimuli rather than a visual one, since no attempt at pursuit was made by ~ feliceps once the fish, presumably the source of the stimulus, immediate environment. This argument moved out of is supported their by the results of Experiment 1 above, in which it was apparent that G. f e liceps was strongly chemosensorally motivated by the presence of fresh scales. It is conceivable that in both species the sharply pointed, recurved teeth would enable the removal of scales, fins and mucus from other fish. However, it is apparent that lepidophagous fish, such as some cichlids from Lake Tanganyika (Liem & Stewart 1976) generally have a single row of large specialised teeth which facilitate the removal of scales from other fish . Similarly, Terapon jarbua, a marine scale-eating teleost (Whitfield & Blaber 1978), has an outer row of enlarged conical teeth on both jaws which are interdigitating and enable the efficient removal of scales. While it appears that ~ f eliceps is able to remove scales from other fish, it has not evolved specialised dentition to facilitate this habit. It is probable that the large majority of scales found in the stomachs of the estuarine caught samples were scavenged. The scales lost by other species caught in the gill net samples would have been readily accessible to G. feliceps prior to their own capture. It is also probable that G. feliceps removed scales directly from fish trapped in the nets, an opportunist feeding action which essentially amounts to scavenging. Finally, the results of Experiment 2 above, the absence of observed visually motivated approaches towards potential hosts , and the lack of specialised dentition indicate that lepidophagy is not a reality in this species . Habitat utilisation and resource partitioning The absence of foraging adult ~ feliceps in the Kowie River estuary gill net samples indicate that they probably prefer 37 deeper, marine environments . This is supported by the observation that the larger individuals caught at sea generally occur at greater depths (up to a maximum of approximately 60m) than the smaller sub-adults and juveniles which seldom occur deeper than 20 to 30m (Buxton, et al. 1984), and which forage extensively in estuaries. However, adult G. feliceps do venture into shallower water (including the mouth regions of estuaries) to forage at night, or diurnally when the water is turbid. Marais & Baird (1980), Marais (1981, 1983a & b) and Coetzee & Pool (1984) found that the bulk of G. feliceps catches in the Swartkops, Sundays, Krom and Gamtoos estuaries (south-east coast), and in the Swartvlei system (south coast) respectively, occurred in the mouth regions between the months of November and February, and that these were largely mouth brooding males. Coetzee & Pool (QQ cit) found that the majority of G. feliceps occurring upstream, beyond the mouth region in the Swartvlei system, were sub-adults and juveniles, a result consistent with that of this study. Estuaries therefore form an extension of the inshore habitat for G. feliceps juveniles and sub-adults, and the extent to which they utilise estuaries is largely dependent on the prevailing turbidity, as suggested by the data of Marais (1984) for the Swartkops, Sundays, Krom and Gamtoos estuaries. The Swartkops estuary produced the lowest cpue for Galeichthys and was the least turbid, while the Gamtoos estuary was the most turbid and yielded the highest cpue . However, ~ feliceps are unable to tolerate salinities much below 8 ppt. (Whitfield et al. 1981; Bennett 1985) and are thus prevented from foraging in estuaries during flood conditions, when food is often abundant. Estuarine predators of ~ feliceps such as Argyrosomus hololepidotus and Elops machnata (Marais 1984) and Platycephalus indicus (this study) probably also influence their estuary utilisation patterns. Galeichthys ater are restricted to reef habitats and do not occur in estuaries. As with in the size distribution of wi th increasing depth. ~ ~ feliceps, there is an increase ater, in hook and line catches, Juvenile 38 ~ ater utilise near-shore, sub-tidal reefs as nursery areas at the termination of the mouth brooding phase, where they feed prolifically on amphipods and isopods, b e fore moving to deeper habitats. A prel i minary feeding study on the Galeichthys species occurring off the Transkei and Natal coasts suggested that it forages over both hard and soft substrata. The warm waters of the east coast are home to a wide variety of sharks, many of which prey on Galeichthys, including five carcharhinid species, Carcharinus limbatus, L obscurus, L brevipinna, C . leucas and C. amboinensis, the spotted ragged tooth Eugomphodis taurus, three hammerheads, Sphyrna lewini, ~ zygaena and mokarran, ~ the tiger, Galeocerdo cuvier and the great white, Charcharodon carcharias (J. Cliff, Natal Sharks Board, pers. comm . ). This species is largely inactive during daylight hours when it remains within the confines of caves and other sources of refuge offered by reefs (Van der Elst 1981; M. Griffiths, Dept . Ichthyology & Fisheries Science, Rhodes University, pers. comm . ) . While this diurnal inactivity may be a predator evasion tactic, it is more probably due to their largely nocturnal habit. While this species appears to utilise both both reef and soft substratum environments, it does not frequent estuaries, as was evidenced by its absence from the gill net surveys of Wallace (1975), conducted in several east coast estuaries. It is probable that the southern distribution of this species is limited by water temperature . Grossman (1982) has questioned whether resource partitioning is necessary for co-existenc,e in stochastic communi ties, and has, along with other authors (Moyle et al. 1982) emphasised the importance of long-term monitoring in order to determine the nature of the community under investigation. Since the duYation of this study was not long enough to allow such an evaluation, it was impossible to demonstrate resource partitioning observed between ~ whether feliceps and ~ the ater was a real (permanent) phenomenon. However, according to the theory of island biogeography (MacArthur & Wilson 1967), K39 selected species assemblages are a characteristic of stable environments. If the presence of K-selected species can be relied upon as an environmental indicator, the near-shore habitat along the eastern Cape coast may be assumed to be a deterministic one. Therefore, resource partitioning may well have played an important role in shaping community structure in this region. Since the large majority of ariid species world-wide appear to forage over soft sediments (Taylor & Menezes 1977; Taylor & Van Dyke 1981; Jayaram & Kailola 1983), it seems logical to assume that this is their preferred habitat type. It was surprising therefore that G. ater chose to forage almost exclusively over reef environments . Either this species was competitively excluded from the preferred habitat during the course of its co-evolution with ~ feliceps, or it was pre-adapted to reef environments prior to its sympatry with ~ feliceps. The evolutionary significance of interspecific competition (and by inference, resource partitioning) has been the subject of much debate (e.g. MacArthur 1972; Schoener 1974, 1983; Connell 1980, 1985; Roughgarden 1983; Davic 1985; Maurer 1985). While character displacement (Brown & Wilson 1956 in Price 1975), or habitat shift (Schoener 1974) invariably results when two normally allopatric, morphologically and behaviourally similar species find themselves in sympatry, it has not been conclusively shown that such character displacement may subsequently become genetically fixed (Connell 1980). In the absence of this proof, Connell (QQ cit.) argues that interspecific competition should not be viewed as a mech'a nism which directs co-evolution. Mayr (1970) has argued that co-evolution of sympatric, congeneric species over vast time periods has often resulted in a high degree of resource partitioning. This led to the initial speculation that resource partitioning between the two morphologically and behaviourally similar Galeichthys species 40 was likely to be intricate. However, while morphological similarity resulted in similar feeding modes and prey taxa in Galeichthys, their separation by habitat preference isolated them from each other to the extent that in terms of range overlap, they effectively existed in allopatry. It is likely, however, that there may be a certain degree of aggressive competitive interaction in the areas abutting reefs, where both species probably forage. Hixon (1980) found this to be the case between overlapping populations of two predominantly allopatric California surfperches, Embiotoca lateralis and ~ jacksoni (Embiotocidae) . Largely as a result of their morphological and behavioural similarity, the diets of these two allopatric species were also very similar and resulted in considerable competitive interactions for food and territories in sympatry, at the interface of their respective habitat ranges. manipulative experiments Hixon (2£ Using demonstrated that cit~) ~ lateralis competitively excluded ~ jacksoni from the preferred foraging environment, namely the productive shallow reef areas. The extreme nature of resource partitioning demonstrated by the two Galeichthys species, namely habitat separation, is surprising in the light of the great many ariid species occurring in apparent sympatry elsewhere in the world. For example, many of the estimated 77 species which occur in the Indo-Pacific Archipelago (Wongratana et al. 1974), have overlapping distributional ranges. Although little is known about the feeding habits of the majority of species, it is generally stated that they feed predominantly on invertebrates and small fish, and that in many fisheries several species may be caught simultaneously, using the same gear (Wongratana et al. 1974; Taylor & Menezes 1977; Taylor & Van Dyke 1981; Jayaram & Kailola 1983). Araujo (1984) did not detect resource partitioning of any nature amongst three ariid species in a Brazilian coastal lagoon and suggested that food was not a limiting factor within their common habitat. However, he did find significant differences in feeding-associated morphological features such as gape size, eye diameter and 41 intestine length, suggesting that food overlap might decrease if food became less abundant, a phenomenon that has been demonstrated by many authors (see Ross 1986 for review). While detailed feeding studies over long periods of time invariably reveal the presence of resource partitioning, it is equally clear that the majority of species feed opportunistically and that in times of food abundance considerable diet overlaps will occur, particularly amongst morphologically similar congeneric species. The rigid and apparently permanent habitat separation demonstrated by G. feliceps and G. ater in this study is therefore unusual, although direct observations of their foraging behaviour may reveal considerably more interaction between them than their diets suggest. In conclusion, it was apparent that the feeding associated morphologies of the two species were highly similar. The taste and touch sensi ti ve circum-oral barbels, elaborate head lateral line network and well differentiated olfactory rosettes clearly indicated that they were adapted for benthic foraging in conditions of low light intensity. While the major food taxa of the two species were identical, species level identification of prey revealed a rigid partitioning of food and, consequently, foraged of entirely space between over soft them. sediments Galeichthys while ~ feliceps ater fed predominantly over reefs and marginally over soft sediments. The wider diet of ~ ater was probably attributable to a greater species diversity over reefs, although this was not quantitatively demonstrated. 42 CHAPTER 3 - REPRODUCTION Introduction outstanding feature of fish reproduction is the immense range of fecundities exhibited within the group as a whole. At An the one extreme there are certain elasmobranchs (e.g. Sgualus ~) 'which produce between two and three young every second year (Gilbert & Gilbert 1980), and at the other there are teleost pelagic spawners such as the ocean sunfish (Mola mola) , which produce up to 28 million eggs per annum (Lagler et al. 1977) . Mortality in fish populations is characteristically high during egg and larval stages and in an evolutionary sense the number of eggs produced by a fish is probably linked to the expected mortality between spawning and recruitment into the parent population (Nikolsky 1969; Cushing 1975; Sissenwine 1984). Two major factors influencing survival during early ontogenetic development in fish are availability of energy for growth and vulnerability to predation. Broadcast spawners characteristically produce large numbers of small eggs with an abbreviated embryonic phase and a prolonged, free-floating, energy-gathering dependent upon larval chance stage . Larval encounters survival with food is largely and passive avoidance from predators, and is characteristically low. As egg size and degree of parental care in fishes increases, fecundity tends to decrease (Lagler et al. 1977). This is particularly well demonstrated in the gradation of reproductive guilds (sensu Balon 1975a, 1981a,b, 1984) from egg hiding, egg guarding, external egg bearing (e.g. oral gestation), to internal bearing in the form of viviparity. The trend is for greater amounts of energy to be invested into fewer young, each with a higher chance of survival through to recruitment into the reproductively active adult population (Oppenheimer 1970). 43 Studies of reproduction in ariids (see Rimmer & Merrick 1983 also Mansueti & Hardy 1967; Jones et al . for review, 1978; Lara-Dominguez et al. 1981; Menon 1984; Mishima & Tanji 1985; Rimmer 1985a, 1985b; Reis 1986; Coates 1988; Yanez-Arancibia & Lara-Dominguez 1988) have demonstrated that they are a highly K-selected group. The largest egg diameter reported for the family is 20mm (Rimmer & Merrick Q£ cit.), a size unsurpassed amongst teleosts (Atz 1958), while their average fecundity at approximately 60 eggs per annum (Rimmer & Merrick Q£ cit.) is amongst the lowest exhibited amongst teleost fishes. The specialised reproductive modes referred to above thought to have evolved in response to stable, are predictable environments (Duellman 1989), where properties such as rapid population growth are not a prerequisite for survival. In the fisheries context the resilience of a population to exploitation is of over-riding importance, both in terms of the economics of the fishery and the survival of the stock. Species that are r-selected generally have high fecundities and rapid population regeneration rates. Their high growth potential renders them particularly resistant to the effects of fishing pressure and the sustainable yields exhibited by fisheries for these species are essentially supported by surplus production (Garrod & Horwood 1984) . The K-selected (MacArthur & Wilson 1967) or precocia1 (Balon 1979) species on the other hand tend to have low fecundities and generate little surplus population growth sensitive to rates production . and environmental vulnerable to over-fishing as a They exhibit result they perturbations (Adams 1980). are and slow highly extremely Examples of their vulnerability include the destruction of cichlid populations after introduction of Victoria the Nile perch into Lakes Kyoga and (Ribbink 1987), the rapid and considerable decline in populations of elasmobranchs (Holden 1977; Compagno In press) and marine mammals (Allen & Kirkwood 1988) following the initiation of intensive fisheries for them. 44 Considering the above, and the fact that the two ariid species formed a considerable component of the fishery at Port Alfred, there was a need to assess their reproductive potential and, hence, their vulnerability to fishing pressure. The models used in this study to assess the effects of exploitation on the population dynamics of the two species (see Chapter 6), were initially developed for the cod, herring and plaice stocks of the North Atlantic (Beverton & Holt 1957). Although the relationship between the numbers of spawners and recruits was poorly understood in these populations it was assumed that because of their extremely high fecundity, recruitment variability would not be limiting to population growth . In the formulation of the models, recruitment was therefore assumed to be constant and independent of stock size (Beverton & Holt QQ cit.). Yield in the fishery was calculated as a function of fishing mortality rate and the age of entry (Smi th 1988) . The only reproductive into the fishery information required by the models was the size at sexual maturity. When considering the exploitation of strongly K-selected species, however, recruitment is likely to be intimately linked with the size of the reproductively active parent stock . Thus the assumption of constant recruitment theoretically does not hold for such populations. As recruitment was not determined for the two populations under study, the conventional yield per-recruit model was still the best available method for determining their population responses to exploitation. In the light of the model's limitations it was anticipated that a thorough knowledge of the reproductive biology of the two species would be of considerable importance when interpreting the yield-per-recruit curves at the end of the study. The large egg size, low egg number and paternal mouth-brooding habit of ~ feliceps has been commented on by several authors 45 in the literature (Smith 1961; Marais & Baird 1980; Day et al. 1981; Van Der Elst 1981; Coetzee & Pool 1984; Taylor 1986; Marais & Venter 1987). These studies were, however, incomplete and inadequate to meet the above requirements. This chapter represents the first detailed investigation into the reproduction of ~ feliceps and G. ater using data collected over a three year period, and covers the following aspects: fecundi ty, Spawning seasonality, sizes at sexual maturity, spawning behaviour and differential reproductive energy investment by the sexes. Two other important aspects of reproduction, the early ontogeny and duration of mouthbrooding, are investigated in Chapter 4 for G. feliceps. Materials & Methods Biological samples obtained from the commercial fishing wharf were collected over three years and processed on a monthly basis. The following information was recorded: body, gonad, fat and liver mass, fork length, sex, gonad maturity, egg size and egg number. Sizes at 50% sexual maturity were determined using an accumulative technique in which the frequency of mature gonads encountered during the spawning season was plotted against size class. The size class in which at least 50% of fish exhibited mature gonads was adjudged to be the size at 50% maturity, and was determined independently for the sexes. Gonadosomatic (GSI) indices were used to determine the period and frequency of spawning, and were calculated as a function of body weight as follows: GSI = GONAD MASS X 1000 / BODY MASS 46 Hepatosomatic were (HSI) determined and abdominal fat-somatic in order to follow (FSI) trends indices in energy accumulation and expenditure . For the calculation of these two indices a correction factor for gonad mass was incorporated in order to avoid masking of the trends. The formulae were as follows: HSI or FSI = LIVER or FAT MASS X 1000 / (BODY MASS - GONAD MASS) As males do not feed while mouth-brooding (see Chapter 4), the condition factor (CF) was used to determine the extent of the weight loss that occurred during this period. Condition factors for females were calculated for comparative purposes. The condition factors were calculated as follows: CF = (BODY MASS - GONAD MASS) x 1000 / ~ (FORK LENGTH b) where b = the exponent derived from the natural log regression of length against weight (3.08 for 3.04 for ~ ~ feliceps and ater). The plots of mean monthly indices were used to exhibit seasonal trends. The abscissa of all plots began and ended with the months of July and June respectively in order to avoid graphic interruption of the spawning period. In order to detect energy utilisation trends in mouth-brooding individuals which are absent from commerci_al catches, the above indices were calculated independently for mouth-brooding ~ feliceps males captured in the Kowie estuary between October and March each year using gill nets. Mouth-brooding ater ~ males could not be captured and the physiological effects of mouth-brooding in this species could therefore not be determined. Adult male fish which were caught in the fishery during the usual mouth-brooding season were individuals which for some 47 reason had not spawned, possibly due to poor condition or through failure to procure a mate. If a certain proportion of the reproductively active male population adopted sneaking as an alternative reproductive strategy they would also be caught during this time. All indices were determined only for individuals equal to or larger than the size at 50% sexual maturity. In an attempt to explain why males and not females are the mouth-brooders in ariids, drawn up for feliceps . Using adiabatic bomb calorimetry the ~ energy content (Kj/g) a reproductive energy budget was of mature yolky eggs, hyaline eggs, testes, fat, flesh and liver tissue were determined. This data enabled the calculation of the approximate reproductive energy investment for the sexes. Similar sized males and females were used in the analyses . In ariids two types of eggs are released at spawning. The few large, yolky eggs which are subsequently fertilized by the male are accompanied by large numbers of small, hyaline eggs of unknown function (Rimmer & Merrick 1983). In the determination of fecundity the mean number of yolky eggs spawned per annum was used. function plotted To of determine whether fecundity increased as a body size the mean number of mature eggs was against fork length. Relative fecundity (RF) was determined in order to enable comparisons to be drawn between the two species and was calculated as follows: RF = GONAD MASS / (BODY MASS - GONAD MASS) The mean number of hyaline eggs and their total average weight relative to the yolky eggs in the ovaries was also determined. It was hoped that the determination of the amount of energy they contained might reveal something about their possible function. 48 Mouth-brooder buccal cavity volume was measured in order to reveal whether males were physically able to carry some or all of the eggs from one female, or whether they were able to carry the egg compliment necessary since it from more than was found one that female. This mouth-brooding was males invariably spat out some or all of their brood when they became entangled in gill nets, and an accurate assessment of the number of young they incubated could not be made . Buccal volume was determined using mouth-brooding males caught in gi l l nets. The buccal volume in the male is increased by lowering the hyoid complex during mouth-brooding . As the buccal cavi ty remained distended after death it was a simple matter to pour liquid plaster-of-Paris into their mouths. After the castes had dried they were dissected out and shaped using a file. This was to remove plaster-of-Paris where it had penetrated between the gill arches. The volumes of the castes were determined by immersion into a calibrated container filled with water. Secondary sex characters were observed in both species and were briefly described. Although spawning was not observed in either species, their proposed spawning behaviour was speculated upon on the strength of the secondary sex characters and the reproductive data. Results ~ Three secondary sex characters were noted in G . feliceps and ater, while a fourth was peculiar to ~ ater. a . ) The pelvic fins of adult females were significantly longer and wider than those of males (Table III) . It was envisaged that they played an active role during spawning. Captured riperunning females were sometimes seen to splay their pelvic fin rays and to hold the fins erect , forming a cup around the vent. It is possible that the eggs, which are adhesive once spawned, are held in the pelvic fin cup whilst the male fertilizes them. 49 Table I I I . Student t-tests d e monstrating significant sexrelated differences in the l engths and wi dths of pelv ic fins in G. f e lice 2s and G. ater. .._ ---------Females Males n mean sd n mean sd df t-test sig dif (95%) (yI N) G. FELICEPS PELVIC LENGTH 23 10.530 PELVIC WIDTH 8.609 23 1.496 0.951 40 14.261 1.245 61 -4.9699 YES 40 12.719 1.541 61 -6.0219 YES 13.554 12.123 1.985 1.844 71 71 -3 .5503 -4.5897 YES G. ATER PELVIC LENGTH 25 10.927 8.879 25 PELVIC WIDTH 0.902 0.913 48 48 YES b.) In males, the pharyngeal and palatine tooth-patches became completely embedded in a film of dense, congealed mucus during the mouth-brooding phase. Teeth on the pre-maxilla and dentary were free of this mucus and remained exposed. c.) The shape of the cleithrum in sexually mature individuals was different for males and females. The posterior margin of the post-humeral process was rounded in females and angular in males (Fig . 13). d.) Galeichthys ater females developed a fatty growth on their pectoral spines during the spawning season, which was not observed in the females of G . felice2s or on males of either species (Plate III) . Males and females of both species are able to produce sound by moving their pectoral spines rapidly backwards and forwards. The sound is produced in a stridulatory manner when the pectoral spine condyle moves across the ridged surface of the socket in the pectoral girdle (pers. obs.) . 50 (Ql (b) I 20mm Figure 13. Left lateral view of a) male, and b) female pectoral girdle (pectoral fin removed), demonstrating the sexual dimorphism in the shape of the post-humeral process of the cleithrum. Plate III. Dorsal aspect of male (left) and female (right) G. ater showing the fatty growth (f), on the pectoral spine of the latter during the spawning season . 51 Gonad deBcription The following macroscopic description is devoted to the seasonal changes occurring in the gonads of sexually mature fish. Gonadal development was identical in the two species and a single description suffices for both. The seven-stage macroscopic gonad maturity scale used is a modification of the six-stage scale used by Nikolsky (1963) . Stage 1 - Immature The testes were thin, thread-like and translucent, extending from the vent posterior edge anteriorly of the approximately swim bladder. half The way to the ovaries were distinguished from the testes only by a gradual thickening at their distal end. Eggs were not visible to the naked eye. Stage 2 - QuieBcent/Recovering Testes were thin, and flattened in cross section, creamy- beige in colour and were attached along their entire length, both to one another and to the dorsal lining of the coelomic cavity, by a medial mesorchium. The outer lateral edges had a frilled appearance. The stage 2 recovering testes were similar in appearance, although considerably longer, extending almost to the swim bladder. The ovaries, distally lobate in the region of the developing eggs, were translucent and attached to the dorsal body wall by a mesovarium. The two ovaries unite proximally to form a single oviduct. The distally situated eggs (± 2mm-3mm in diameter) were pale orange in colour and interspersed with two smaller egg-types, some pale yellow and some creamy-white, Proximally, O.2mm-O.3mm) both approximately 1.5mm in diameter. the ovarian wall bore masses of very small white eggs. The walls of Stage 2 (± recovering ovaries were thickened and opaque. Stage 3 - Developing The testes were longer, extending all the way to the swim bladder, pinkish in colour and semi-circular in cross section with the ventral surface being rounded. They were proximally 52 thin in the region of the long deferent duct (Loir et al. 1989), with only the distal half of the testes being thickened. The outer lateral edges were still frilled in appearance. The lobate distal region of the ovaries contained yolky eggs of approximately 4mm-6mm in diameter. Stage 4 - Active The testes had a reddish tinge . The proximal third of the testes remained underdeveloped and more deeply reddened than the distal two thirds, which were swollen and approximately 9mm in width (Plate IVa). The ovaries were greatly enlarged and occupied approximately 50% of the body cavity. The ovarian walls had stretched and become more translucent making the large (± 8mm-9mm) yolky eggs and the small, creamy-white eggs, termed hyaline eggs (Rimmer & between them clearly visible. Merrick 1983), interspersed The follicular blood vessels were prominent against the yellow yolk of the developing eggs (Plate IVb). Stage 5 - Ripe The testes had lost their redness and were pale and translucent (Plate IVa). Distally the girth of each testis had increased to a width of approximately 12mm. When incised they released colourless milt. The ovaries were greatly enlarged and contained eggs of approximately 11mm-14mm in diameter (Plate IVb). Females with ripe ovaries could be identified externally by their distended ventral abdominal wall. Stage 6 - Ripe-running Testes in the condition ripe-running translucent and on application of were completely slight pressure released colourless milt. Ripe-running ovaries were identified as those in which the eggs had (strdtum germinosum), been released from their follicles (Plate IVc). The yolky and hyaline eggs were easily extruded upon application of slight pressure to the abdominal wall . A plug of small hyaline eggs (each ± 0.7mm in 53 diameter) was extruded first, and were followed by the yolky eggs (± 13mm) , interspersed with larger hyaline eggs (± 1. 8mm) . B A :to ' - ,., J c Plate IV. G. feliceps gonads in (a) stage 4 and 5 testes, (b) stage 4 and 5 ovaries, (c) stage 6 ovary showing loose ripe and hyaline eggs after release from the follicles. 54 Stage 7 - Spent Spent testes were deeply reddened and flaccid. simi larly small, flaccid. Distally, amongst the Ovaries were empty follicles, pale yellow eggs of approximately 2. 5mm in diameter were v isible . These probably represented the following season's compliment of yolky eggs. No hyaline eggs were visible to the naked eye. Four different egg sizes were present in ripe ovaries. Situated distally the large (± 13mm in diameter), orange-yellow yolky eggs were the most obvious . Closely associated with them were a similar number of small (± 2.5mm) pale yellow, immature yolky eggs. An abundance of hyaline eggs (± 1.75mm) also occurred amongst the yolky eggs. Hyaline eggs of a distinctly smaller size class (± 0.7mm) were distributed throughout the mid- and proximal regions of the ovary. They were particularly densely distributed on the ovigerous folds of the tunica albuginea in the vicinity of the oviduct. The sizes and colours of the different egg-types are presented in Table IV . Table IV. The position, size and colour of the egg-types found in Galeichthys feliceps and G. ater ovaries . Egg type Position Colour in ovary .._- small hyaline large hyaline immature yolky ripe yolky --- Diameter (mm) sd 0.15 0.19 0.21 0.44 proximal distal distal white pale yellow 0.69 1.75 2.60 distal orange-yellow 13.19 wh ite 55 Sizes at sexual maturity feliceps males and females reached 50% sexual maturity at ~ approximately 315mm and 295mm (FL) respectively (Fig. 14). In G. ater 50% maturity was reached at approximately 235mm in both sexes (Fig. 15). Galeichthys ater males did not demonstrate 100% maturity at any size. This would appear to suggest that all individuals did not spawn each year. A similar phenomenon was observed amongst the larger female size classes. Spawning seasonality The spawning season was gonadosomatic (GSI's) indices determined and revealed occurred once a year in both species. reached a peak in development in from In the monthly that spawning feliceps gonads ~ September and occurred between September and December (Fig. 16) . In spawning ~ ater spawning occurred between August and October (Fig. 17). Seasonality of energy reserves associated with reproduction The mean distinct monthly fat-somatic sex-specific trends . indices (FSI' s) In males, fat demonstrated deposits were accumulated during autumn and winter prior to the spawning season (Fig. 18). In females, fat reserves were accumulated during summer after spawning and declined steadily during the winter months (Fig. 19) . 56 (a) 90 90 70 GO 50 <0 '0 20 o 290 270 250 330 350 3 30 350 SIZE CLASS (mm) F.L. o N_ 2' 1 ( b) '00 ,- - ~-a 90 90 70 GO 50 <0 )0 20 10 o ~ 250 - ~-r_ W_ 270 -_,~ -,_. 290 -4 "0 SIZE CLASS Crnm) F.L . o N",239 Figure 14. Size at 50·% sexual maturity for males (± 3l5rnrn) and (b) females (± 295rnrn). 57 (a) G. feliceps (a) '00 90 80 70 , 60 ':: cr ~ ~ so - - - -- - ---- --- ----- ~ ~ .0 30 20 10 0 200 220 2'0 2.0 230 250 260 270 290 280 FO.c:lK LENGTH (rrm) N • 0 '29 ( b) 100 -r - -~a - ~- - -_, 90 80 70 30 20 10 a 4r 200 ~-_$ 2'0 -,L_.r 220 >30 2 .0 250 FO~K o 260 270 280 290 300 '"a LENGTH Cnm) N = 135 Figure 15. Size at 50% sexual maturity for (a) G. ater males (± 235mm) and (b) females (± 235mm) . 58 (0) 7 , - - -- - - - - -- - - - - - - -- - - - - - - - - -- - - - - - - - - - - - - - - - - -- - - - - - - , 6 , 2 "UL AUG SEP OCT NOV DEC " AN FEB MAR AP R " AN FEB MAR APR MAY "UN MAY "UN IIiIONTHS (b) "0 '20 "0 '00 90 80 70 60 50 40 30 20 TO 0 "UL ACG SEP OCT NOV OEC MONTHS 16. Monthly gonadosomatic index values for (a) Sh feliceps males and (b) females, demonstrating peak gonadal development during September. Vertical bars represent 1 SD. Figure 59 (0) 9 , -- - - - - - - - - -- -- - - -- - - - - - - - - - - - - - - -- - - - - - - - - - - - - - - - - -- - - - - - - - , 6 7 6 5 4 3 2 JUL AUG SEP OCT DEC JAN FEB MAR APR MAY JUN JAN FEB MAR APR MAY JUN MONTHS ( b) "'0 70 0 90 eo 70 60 SO 40 30 20 70 0 JUL AUG SEP OCT NOV DEC MONTHS Figure 17. Monthly gonadosomatic index values for (a) G. ater males and (b) females, demonstrating peak gonadal development during August. Vertical bars represent 1 SD. 60 (a) ,. ,-~ 9 8 7 6 5 • , 2 o 4- - r- SEP JUL -r,. OCT DEC NOV JAN FEB MAR AP R ~ MAY JUN MONT HS 6. MOUTH- BROODERS ( b) ,. ,-----------------------------------------------------------, 9 8 7 6 S • 1-----j---J.. , 2 o -r.4, JUL -r~ AUG SEP OCT NOV -_r.,~ DEC FEB MAR APR MAY . ..JUN MONTHS Figure 18. Monthly fat-somatic index values for (a) G. feliceps males (mouth-brooders plotted separately) . Vertical bars represent 0.5 SD, and (b) females. Vertical bars represent 1 SD. 61 (0) ,. " 22 20 '" '6 ,. " '0 " 6 < 2 0 cUL MONTHS ( b) 2. 2< 22 20 '" '6 ,. " '0 " 6 4 2 0 cUL MONTHS Figure 19. Monthly fat-somatic index values for (a) G. ater males and (b) females. Vertical bars represent 1 SD. 62 The mean monthly hepatosomatic indices (HSI's) for G. feliceps males exhibited a peak during September and October, at the onset of spawning. The HSI' s for estuarine caught mouth- brooding individuals demonstrated that liver mass reached a low point at the culmination of the buccal incubation period (Fig . 20a). Mouth-brooding ~ ater males were not captured and the monthly HSI's therefore do not reflect a marked decline until January when individuals which had completed their mouth- brooding began to be caught in the fishery again (Fig. 2la). The HSI's for females did not fluctuate substantially (Figs. 20b&2lb). Condition factors for the sexes followed much the same trend as the fat and hepatosomatic indices. In ~ feliceps male body condition declined markedly during the mouth-brooding period (Fig. 22a). Females reached peak condition in May prior to the onset of gonad maturation, although their condition did not fluctuate markedly (Fig. 22b). Mouth-brooding G. ater males were not captured and the monthly condition factor curve is therefore not an accurate reflection (Fig. 23a). G. females did not vary markedly in condition (Fig. 23b). 63 ater (0) .. 2. 25 23 22 "20 19 18 '. " 15 "13 12 11 10 JUe I\AC)NTHS + MOUT H- BROODER S ( b) . 2. 25 21 23 22 20 19 '€ ,. "~ ---- 18 r----. ~ V :----- I I --- " 15 " - 13 12 11 10 JUL AUG SEP OCT DEC JAN FEB APR JUN MONTHS Figure 20. Monthly hepatosomatic index values for (a) marine and estuarine caught G. feliceps males and (b) females. Vertical bars represent 1 SD. 64 (0) 2. 27 26 25 24 23 22 ,.,. "'0 Vl :r: n '6 " '4 "'2 "'0 JUL AUG MONTHS ( b) 2. 27 26 2S ' 4 23 22 ,.,. "20 Vl I n ,6 " ,. " " "'0 J UL MONTHS Figure 21. Monthly hepatosomatic index values for (a) G. ater males (mouth-brooders not represented), and (b) females . Vertical bars represent 1 SD . 65 ( 0) 0 . 012 0 . 011 - 1 0 . 01 ~ \i ~ z Q 0 . 009 - 0.008 - I-~ ~ 0 9 I ~ '-------., O . OO? 0 . 006 JUL AUG OCT DEC JAN FEB MAR MAY JUN MQNT I-IS ( b) 0.012 , - - - - - - - - - - - - - - - - -- -- -- - - -- -- -- -- -- -- -- - - - - - - - - - - - - - - -______, o ~ t:; 011 0 . 01 ~ ~ 0 . 009 "8 0 z 0.008 0.007 0 . 006 4-._,r~ JUL AUG OCT DEC JAN FEB MAR MAY JUN MONTHS Figure 22 . Monthly condition factors for (a) G. feliceps males (mouth-brooders plotted separately, standard deviations plotted below the curve for clari ty), and (b) females. Vertical bars represent 1 SD. 66 (a) 0 . 016 0 . 015 0 . 0 14 0 0" g u 0.012 ~ 0.011 ~ !t "§ 0 0 0' 0 . 009 0 . 008 0 . 007 0 . 006 "UL A\Xi SEP OCT MONTHS ( b) 0 . 016 0 . 015 0 . 014 0 .0 13 15 t:; ~ I--- v 0 .01 2 !t ~ 0 . 011 ~ 0 is 0 . 01 U 0 009 0 . 008 0 . 007 0 . 006 "UL A\Xi SEP OCT NOV DEC "AN FEB MAP APR MAY "UN MONTHS Figure 23. Monthly condition factors for (a) G. ater males (mouth-brooders not represented), and (b) females. Vertical bars represent 1 SD. 67 Relationship between fecundity and body dimensions The influence of fish size and mass on the number and size of eggs i s reflecte d in the regression analyses in Table v. Table V . Regression equations for the re l ationships between fecundity and fork length, body mass and egg size for G. feliceps and G. ater . _ -.-.... .. _. __...-._.__.._ .__ _--_.._ ....._..._ ...._.._._---------- _ .._---_.__. Constant s,• . y J( coefflclent S.&. ,2 x .. Fl vs. Egg no. G. fe lce ~ G. ater Body wI. '19. Egg no. G. fencees G. ater Body '0'11:. vs. Gonad wi. Go feficeps G. ater 21.8104 12.5538 33.4452 41 .9648 42.5726 -13.8787 8.7817 0.0082 0.0488 0.02 12.3805 0.0256 8.7085 26.4583 0.0953 0.0222 -1.7656 14.3878 0.0607 -1.9821 0.0214 0.2613 0.11 0.07 0.28 0.10220,0520 0.03 0,0104 0.04 .().Q163 0.011 8 0.02 Egg diameter vs. Egg no. G. felice!:!:, G, ater 73.3454 52.7478 5.6282 -1.9369 0.1535 0.58 10.7913 Considering the r' values it could be concluded that there was no increase in fecundity with increasing fish size or mass. There was also no relationship between fish mass and gonad mass. The negative slopes (x-coefficients) of the relationships between egg diameter and egg number for the two species indicated that as eggs increased in size (matured), their numbers decreased (Figs 24 & 25). This meant that not all of the eggs originally present in, for example stage 3 ovaries, actually reached full maturation. This indicated that some developing eggs are resorbed during gonadal maturation . This wa s corroborated by the significant difference found between the mean number of eggs in stage 3 and stage 5 gonads in feliceps and in ater (Table VI). ~ Table VI. Test for significant differences between the number of developing eggs in stage 3 and stage 5 gonads for G. feliceps and G. ater. STAGE J OVARIES n M.~ E9. ~ G. leUc~s '" 22 ~. ., STAGE 5 OVARIES n -----._- 57.90 9.72 66 39 .n ' .64 " '" df Egg no. 49.27 31.54 12.79 4.00 107 58 M~ - -- - - - - - - - - - - - - - - - - - --------.---- .. _-------, 68 140$1 alg dif {95')1o leveQ 2.0007 YES 2.733 1 YES --------_. __._--- ~ 90 0 0 80 0 0 0 0 ! ~ 0 0 0 0 0 0 0 0 0 0 70 ~ 0 0 0 0 0 0 0 0 60 0 so 0 0 0 0 0 0 0 0 0 0 0 0 0 8 0 0 0 8 0 0 0 0 B 8 0 0 ~ § 0 0 8 0 0 0 0 0 0 0 0 0 8 0 0 0 ~ 0 § 0 0 0 0 0 0 <0 0 0 0 0 0 8 8 § 30 0 20 4 2 6 10 8 DjAMeT~ EGG 12 14 CnYn) Figure 24. The relationship between egg number and egg size (i. e. gonadal maturation) in G. feliceJ;1s. 60 55 0 50 B 4S 0 ~ w m 40 ~ ~ § 0 3> 0 B B 0 0 0 0 B 0 0 0 0 B 0 0 0 ~ B 0 30 0 0 8 B 0 0 0 0 0 25 0 0 0 0 § 80 0 B ~ 0 B 8 0 0 20 " 3 5 7 9 EGG DIAMETER 0 N .. 11 13 " 1'~ Figure 25. The decrease in egg numbers associated with gonadal maturation in G. ater . 69 A comparison of the average sizes and numbers of ripe eggs and of the relative fecundities in feliceps and G. ater gonads ~ is presented in Table VII . Table VII. Test for significant differences in egg size, egg number and relative fecundity (RF), between G. feliceps and G. ater. G. ATER Mean sd G. FELICEPS Mean sd n n dl sig dif t-test (P>0.05) Egg size Egg No. RF 12.36 49.27 0.119 4.85 12.79 0.037 72 66 82 12.55 31 .54 0.099 5.9 4.60 0.024 46 34 39 116 58 119 0.0936 2.7331 1.6159 NO YES NO The non-viable hyaline eggs were not included in the fecundity analyses for obvious reasons. However, they did constitute a substantial proportion of the mass of the gonads (± 22%), and the possibility that they played a role in reproduction could not be ignored. The mean number of hyaline eggs determined for three G. feliceps ovaries was found to be 23 013 (± 8 436). Assuming similar egg numbers for both ovaries, the total number of hyaline eggs per female was in the vicinity of 46 000. The buccal volume of mouth-brooding males was found to be slightly larger than the volume of an average egg compliment from a single female (Table VIII). Although the mean number of embryos found in the mouths of brooding males (x = 53), exceeded the mean number of eggs from individual females (x = 49), this difference was not significant at the 95% level (tstatistic = -0.678, df = 90). This indicated that males probably incubated all of the eggs from one female, and that 70 the buccal cavity was large enough to accommodate the increase in embryo size following hatching. Parameters used in the determination of brood volume in G. feliceps . Mean buccal volume for males is also presented. Table VIII. --.------ - - - -- - - - ----------------- -------- - - -- - no. eggs egg vol. per female (mil no. embryos calculated brood vol. brooded sd 49.27 12.79 n 77 1.02 0.18 256 53.33 8.43 15 Mean 50.26 calculated buccal vol. (mil 64.38 8.63 13 Reproductive energy investment In females, monthly hepatosomatic indices and condition factors revealed that liver and body weights remained relatively constant throughout the year, and they were not considered to be sources of reproductive energy. While the fat-somatic indices indicated that fat was accumulated during summer and autumn, it appeared to during winter. be utilised The energy contained for gonadal maturation in ripe ovaries would therefore be representative of the entire reproductive energy investment in females . There were no significant differences between mean body mass (t-test = 0.7019; df 0.3269; df = = 89; P > 0.05) or liver mass (t-test - - 89; P > 0 . 05) between similar sized males and females, suggesting that males did not accumulate body protein or liver reserves prior to mouth-brooding. There was, however, a significant difference in mean abdominal fat weight between the sexes at the time of spawning (t-test = 4 . 8015; df P < 0.05), indicating that males = 89; actively accumulated fat reserves prior to the mouth-brooding period. 71 In males, there was a signif i cant difference in mean fat weight (t-test = 5 . 291; df = 80; P < 0.05), liver weight (t-test = 6 . 6199; df = 80; P < 0.05) and body weight (t-test = 4 . 8887; df 80; = P < 0.05) before and after mouth-brooding . By implication,therefore, these were all sources of energy which were utilised during the mouth-brooding period. These, together with the energy contained in ripe testes, were considered to be representitive of the total reproductive energy investment in males. The energy and moisture content of the gonads, fat, liver and flesh of sexually mature G. feliceps between 315mm and 375mm (FL), are presented in Table IX . Table IX . Energy (Kj/g) and moisture (percentage by weight) contained in the gonads, fat, liver and flesh of G. feliceps . Sample ._- - Sample % Moisture content no. Mean sd - _ . _ - - - - - - - - - - - - - - - - ---- Dry weight Wet weight energy (Kj/g) energy (Kjlg) sd Mean _ Mean ---- -- --- . 7 52.953 2.717 27.866 1.650 HYALINE EGG 7 93.048 2.245 19.294 1.123 1.341 TESTES 76.016 5.225 27.453 2.125 6.584 RIPE EGG 12 13.11 FAT 13 16.468 4.546 36.077 0.926 30.136 LIVER 9 75.091 2.416 25.817 2.212 6.431 FLESH 3 79.930 1.020 23.655 0.978 4.748 -------- The amount of energy utilised determined by converting during mouth-brooding the post-mouth-brooding losses was in body, fat and liver weights into units of energy, using the data from Table IX . This energy, together with the energy contained in the testes, was considered to be representative of the reproductive energy investment by males . 72 The proportional body tissue inputs into mouth-brooding (flesh), liver and fat energy were 74.72%, 3.99%, 19.78% respectively. Males lost approximately 23.83% of their body mass, 50 .7 4 % of their liver mass and 76.4% of their accumulated fat reserves during mouth-brooding. hyaline eggs : spent ovary was The weight ratio of ripe eggs found to be 63 . 303% : 21.757% 14.941% (n = 11) . This enabled the separate extrapolation of the yolky and hyaline energy components from total gonad weights. The energy contained in the yolky eggs amounted to 98.91% I while the hyaline eggs contained the remaining 1.09%. Reproductive energy investment by the sexes is presented in Table X. These data revealed that females invested approximately 35% less energy into reproduction than the males . Table X. Reproductive energy investment in male and female G. feliceps. Pre-spawnlng mass M ~ .~ BODY MASS 585.63 FAT MASS 7.17 LNER MASS GONAD MASS 10.84 2.67 Post.mouth.brooding mass '" M.an '" 72 .49 5.35 2.27 446.05 53.63 1.35 1.91 5.34 0.63 1.40 1.74 0 .11 Weight loss Reproductive Energy (KI) (g) 139.58 5.82 5.50 2.04 682.676 175.392 35.371 13.431 TOTAL ~ Females YOLKY EGG MASS HYAUNE EGG MASS 43.72 15.03 6.14 2.57 0 0 TOTAL 0 0 43.72 15.03 573.169 20.155 ~ .._--_.... Speculations on spawning behaviour While courtship and spawning behaviour were not observed, a certain amount could be deduced from the reproductive data presented above and from the secondary sex characters. The following spawning scenario is envisaged: Males and females exchange recognition signals, possibly of an acoustic nature, 73 before pairing off. The female releases all of her eggs at once into the expanded pelvic fin cup. The eggs are adhesive and the egg mass is held between the pelvic fins while the male orientates himself appropriately and releases sperm. He then turns and takes the entire egg mass into his mouth. The sexes separate and the male seeks out a environment in which to mouth-brood. sheltered, protected The observation that 12% of the ~ feliceps males and 16% of the ~ ater males with ripe gonads, lacked abdominal fat reserves, led to the speculation that sneaking might be employed as an alternative reproductive strategy in the two species. This behaviour was however not confirmed . Discussion Although there was a considerable overlap in the spawning seasons of the two species, peak spawning commenced one month earlier in G. ater than it did in ~ feliceps. This phenomenon was observed during four successive spawning seasons and appeared to be consistent. While it is known that the onset of seasonal gonadal maturation in fish is controlled endogenously under the influence of exogenous environmental factors (Bye 1984), the reason for the temporal difference in spawning between these two closely related sympatric species is unclear . The timing of gamete release after completion of gonad ripening may be triggered by an environmental stimulus such as food availability, or potential food availability for the young in accordance with Cushing's (1975) 'Match/Mismatch' hypothesis, although this theory has recently been challenged (Sinclair 1988) . It is more probable that the temporal separation in spawning time evolved prior to, or during the speciation process. If latitude than ~ ~ ater originated in a more northerly feliceps, it may have been geared to spawn slightly earlier in the year. The specific mate recognition system (Paterson 1978) would also have evolved prior to, or 74 during the speciation process, which presumably occurred in allopatry, species and would have resulted in the maintenance integrity during their subsequent sympatry. of Out of curiosity, an attempt was made at hybridising gametes from feliceps males and ater females ~ in the laboratory. ~ The experiment was unsuccessful and the zygotes ceased development after the 8-cell stage. The cycle of accumulation and metabolism of visceral fat reserves occurred in response to different demands in the sexes. In females the fat reserves reached a peak in May and a low point in October-November, indicating that they were probably utilized for ovarian maturation, a common occurrence in fishes (e . g . Nikolsky 1963; MacKinnon 1972; Shul'man 1974 in De Vlaming et al. 1978; Delahunty & De Vlaming 1980; Pierce et al. males 1980; Fishelson et al. 1985; Flath & Diana 1985). In the fat reserves were accumulated during autumn and winter when gonad development was also occurring, and reached a peak in spring. Fat reserves declined rapidly after spawning, reaching a low at the culmination of the mouth-brooding phase, indicating that they contributed toward metabolic demands during this period. Guillemot et al . energy (1985) reported a similar phenomenon in several scorpaenid species in which accumulation of fat reserves occurred simultaneously with gonadal recrudescence, and served as energy for maintenancemetabolism during the post-spawning period when food was scarce. It appears that the female scorpaenids, which were from a live-bearing genus, and male ariids both accumulated fat reserves at the expense of increased growth in order to ensure their subsequent survival during the period of incubation. The hepatosomatic index for the females of both species reached a low point in August (Figs. 20b & 21b), indicating that liver reserves may also have been involved in gonadal maturation. Liver reserves were also utilised quite considerably by males during the oral incubation period. The hepatosomatic indices for ~ feliceps males (Fig. 20a) show that liver weight 75 decreased rapidly after the onset of mouth-brooding, indicating early mobilisation of the liver lipid store. Hepatosomatic indices also suggested that non-spawning males tended to begin resorbing liver energy approximately one month after the peak spawning period. In non-brooding ater males the liver weight ~ continued to increase during the summer months (Fig. 21b). The graph shows a sharp drop-off after December when the post- mouth-brooding individuals re-entered the samples after recommencement of feeding . The liver mass as a proportion of body mass in both species was low (approximately 1.9%), indicating that it is probably not an important energy storage organ. The finding that the bulk of the mouth-brooding energy requirements in ~ feliceps were derived from body musculature (74.7%), suggested that the muscle lipid content for this species was probably high. This was confirmed by Marais & Venter (1987), who found that lipid levels varied between 3.9% (in juvenile males) young embryos), in and 7.4% (in mouth-brooders feliceps ~ body musculature. incubating Shchepkin (1971a,b), exhibited two different methods of energy storage in horse mackerel and biochemical analyses scorpionfish of respectively. Using liver and muscle lipid content he demonstrated that the former had a small proportional liver weight (1.5% of body mass) and a low liver lipid content (12%), compared to the latter (7% and 25% respectively). However, the muscle lipid content in horse mackerel (3%) was higher than that of scorpionfish (1 . 2%-2.7%) . ~ feliceps and horse mackerel therefore exhibit similar energy storage mechanisms. Fat and protein are thus tha two main forms of energy utilised during mouth-brooding in involves a assimilation, number of catabolism feliceps. ~ processes and Protein accumulation including anabolism, and digestion, metabolically speaking, is an expensive energy source. Fat accumulation may be achieved through directly digestion and through pinocytosis absorption (lIes (Phillips 1984), 1969). It metabolically inactive and is deposited almost unchanged. 76 or is Fat would therefore be an energetically more economical energy source to accumulate, and males might be expected to utilise fat rather than protein as an energy source during mouthbrooding. It is clear from the results above that while fat is accumulated prior to mouth-brooding, the reserves are insufficient to meet the demands of the buccal incubation period and large amounts of protein are also metabolised. Judging from the reproductive energy dynamics of G. feliceps it would seem improbable that an individual could manage to mouth-brood successfully without the stored fat reserves. Males lost an average of 24% of their body weight during mouthbrooding and without their high energy fat reserves (30.14 Kj/g), a proportionately greater amount of low energy body protein (4.75 Kj/g), would be needed to yield the same amount of energy. In the absence of fat reserves it is estimated that the total body weight loss would increase to approximately 34%. The duration of mouth-brooding in ~ feliceps is approximately 140 days (see Chapter 4). Males do not feed while incubating and the significance of feeding cessation is highlighted by the work of Iles (1984), who found that the maintenance of an active epithelium in the alimentary canal was the largest nonreproductive source of protein loss in herring. The long period of starvation during mouth-brooding in ~ feliceps leaves the males considerably emaciated at the end of it, and the monthly condition factors for ~ feliceps mouthbrooders (Fig. 22a) clearly demonstrate the marked seasonal drop-off in condition . From the above information it is clear why it is the males and not the females that perform the parental care role. Energetically speaking it is unlikely that females would be able to produce the full compliment of eggs, mouth-brood them to term and also recover in time for the next spawning season. 77 If it is true that males are not able to mouth-brood unless they have accumulated abdominal fat reserves, the observation that 12% of feliceps and 16% of ~ ater males with fully ~ mature testes did not have any fat, would imply that they intended the to spawn without assuming burden of mouth- brooding. This suggests that these males may employ sneaking as an alternative reproductive strategy. common in fish (e.g. Gross 1984; Sneaking is fairly Chan 1987; Noakes et al. 1989), and also occurs in other animal groups (e.g. isopods, Shuster 1989 and grasshoppers, Alexander & van Staaden 1989). The implications of possible sneaking behaviour in ariid reproduction will be discussed at greater length in the General Discussion (Chapter 7). The plots of fork length against accumulative percentage sexual maturity indicated that mature (Fig. feliceps spawned each year ~ 14) . However, approximately 10% of the mature ater ~ male population failed to do so (Fig. 1Sa). Closer examination of the data revealed that where ~ ater individuals of mature size exhibited undeveloped testes during the spawning season, they had also not accumulated any fat reserves. Whether the lack of fat was a cause, or a symptom, of spawning failure or whether it was unrelated to the latter is not known. conceivable that if the duration of mouth-brooding in was even longer than that found for ~ It is ater ~ feliceps, some of the males might not have been able to recover condition in time for the successive spawning season. A certain proportion of ~ ater males may therefore only spawn every alternate year. In ~ feliceps a small percentage of males which had both fat reserves and ripe gonads also failed to spawn. These individuals may simply have been unable to find a mate. Females in the process of resorbing ripe eggs were also encountered in the samples toward the end of the spawning season, indicating that this phenomenon was common to both sexes . This finding would tend to corroborate the spawning behaviour scenario 78 described earlier in which it was env isaged that males and females spawned in pairs , as opposed to groups . ~ feliceps was thought to be monogamous since the number o f eggs produced by females was not significantly different from the number of embryos in the average male brood. In addition, all of the embryos being brooded by individual males were always at the same stage of development . Dmitrenko (1970), using a volumetric comparison of egg-batch and male buccal cavity, also deduced that males carried the eggs from a single female in Arius thalassinus from the Arabian Sea. However, he found that females of the species produced up to three batches of eggs during one spawning season. This is the only account in the literature of serial spawning in an ariid, and if this mode of spawning does in fact occur, it would probably be manifest in a sex ratio which heavily favoured males. This is because each male would most likely only be capable of brooding one batch of eggs each year, as indicated by the results from this study . Dmitrenko (2£ cit.) does not present sex ratio information, although Bawazeer (1987) demonstrated that the sex ratio for thalassinus ~ in Kuwaiti waters differed Significantly from unity in all size classes, but that neither sex was consistently favoured. The literature presents conflicting results with respect to ariid fecundity vs. body size relationships. Some species exhibit positive correlations (Etchevers 1978; Mishima & Tanji 1985; Rimmer 1985a; Reis 1986a; Coates 1988), while others reveal no correlation (Tobor 1969; Mishima & Tanji 1985) . The results of this investigation supports the latter (Table V). Since 50% sexual maturity occurred at 295mm, and the maximum recorded size for approximately 80% ~ feliceps females in the study was 375mm, of their growth attainment of sexual maturity. A was completed before closer examination of the l iterature revealed that for species in which there was no correlation between body length and fecundity, between 78% and 80% (n = 3) of their growth was completed prior to sexual 79 maturity. For species in which there were positive fecundity body size relationships, sexual maturity was r e ached earlier, when between 40% and 77% (n = 7) of their growth had been completed. Few authors have discussed the function of the hyaline eggs in ariids and only two suggestions have been put forward in the literature. Firstly, that they serve as nourishment for the male prior to commencement of mouth-brooding (Gunter 1947). Secondly, that they effect a lowering of the specific gravity of the egg mass and prevent it from sinking into the 'ooze ' of the substratum (Dmitrenko 1970). Menon (19B4) found hyaline eggs in the buccal cavities of four mouth-brooding ariid species from Indian waters. In all cases, sampling must have occurred soon after spawning as in each i nstance the eggs were still adhesive, a property which is lost after a few hours (see Chapter 4). Menon (op cit.) observed two distinct sizes of hyaline eggs (O . 2Smm-0 . Bmm and 1.9mm-3 . 3mm), along with the ripe yolky eggs in the mouths of adult males. Dmitrenko (1970) encountered partly digested 'yolkless' hyaline eggs in the stomachs of two mouth-brooding male A.:.. thalassinus, indicating that these eggs were indeed ingested . However, in the present study it was found that the hyaline eggs were of little nutritive value (Table IX), and also that they were negatively buoyant . Therefore neither of the above suggestions regarding their function seems convincing. While no concrete suggestions could be made on the strength of this study, the following ideas came to mind: a) Hyaline eggs may be the source of a chemical stimulus inhibiting feeding in males , although yolky eggs could serve this function equally well, b) as the hyaline eggs are released ahead of the yolky eggs at spawning they may serve as egg dummies to stimulate males to commence fertilization, c) the hyaline ancestral eggs condition, may be a and may BO remnant be from a more representative fecund of an evolutionarily incomplete process of change to a less fecund animal. Siluroids are thought to have evolved in fresh water (Greenwood et al. 1966), and the vast majority of present day fresh water siluroids marine representatives are highly fecund. (Ariidae, Since all of Aspredinidae & the Plotosidae) are either mouth-brooders or guarders with low fecundities, it seems that there may have been a shift toward lower fecundities in these families. Biochemical analysis of the constituents of hyaline eggs and a microscopic examination of their cellular structure might contribute toward our understanding of their reproduction. Gibaeva & Errnolina ( 1972 in Rimmer 1983), detailed differences in the cell role in & Merrick structure of the follicular epithelium in hyaline and yolky oocytes during the later stages of oogenesis in Arius felis, although no explanation as to the functional significance of the finding was presented. The reasons for the secondary sex characters observed in feliceps and ~ ater are also open to speculation . ~ Rimmer (1985a) found that the epithelium of the palatine tooth patches in Arius graeffei thickened and grew over the teeth during incubation thereby protecting the young from damage . The mucus coating on the pharyngeal and palatine tooth patches in ~ feliceps and ~ ater could serve a similar function. The need for such protection was highlighted by the frequent coughing motions exhibited by adults during mouth-brooding, which lead to considerable movement of the embryos within the buccal cavity (pers. obs.). The embryos were separated from the teeth on the pre-maxilla and dentary by an oral valve, and the latter were not mucus-coated. Sexual dimorphism in the pelvic fins has been described in several ariids (Lee 1937; Dmitrenko 1970; Rimmer 1985b). The pelvic fins of females were larger than those of the males in all instances. In some species, fleshy hook-like structures 81 appeared on the adaxial (dorsal) surfaces of female pelvic fins during the spawning season . These protuberances had the effect of orientating the pelvic fins in a vertical plane a n d formed a trough on either side of the vent. Some authors speculated that the female clasped the male in her enlarged pelvic fins during spawning, while others suggested that the fins were modified to hold the eggs while the male fertilized them. While the method of egg transfer between female and male has not yet been observed in ariids, it is most probable that the enlarged pelvic fins in females serve to hold the adhesive egg mass at spawning until the male has fertilized them and is able to turn around and take them up into his mouth. It seems unlikely that the pelvic fin modification would have evolved if the females laid their eggs directly onto the substratum as suggested by Gerard (1958 in Dmitrenko 1970). Sexual dimorphism in the cleithra of ariids has not been previously reported in the literature. In G. feliceps and ~ ater it is probably associated with sex-specific sound production or reception. Rapid backward and forward movement of the pectoral spines sets up a stridulation between the pectoral spine condyles and their sockets . Viewed laterally, the trajectory of the pectoral spine tip during repeated anterior-posterior motions, is elliptical. On the posterior stroke the tip of the spine is lowered slightly in a ventral direction, while on the anterior stroke it is elevated slightly . Sound is produced during the posterior stroke, when the many fine, transversely orientated ridges on the ventral surface of the superior condyle come into contact with the rugose, juxtaposing surface of the pectoral groove within the socket (terminology after Halstead 1978) . The left and right pectoral spines are moved backwards and forwards alternately, resul ting in the production of continuous sound . Sound was occasionally produced by the animals when they were handled, and the considerable vibration in the vicinity of the cleithra which accompanied it, could be clearly felt in the fingers. The post-humeral process of the cleithrum forms a close association 82 with the air bladder in that the medial surface of the former abuts directly onto the lateral surface of the latter. The air bladder in ariids is associated with both sound reception and sound amplification (Tavolga 1962, 1971), and it is envisaged that the sexual dimorphism of the cleithrum may playa role in one or both of the above . The difference in shape might modify the acuity of sound perception or the tone of the sound produced in the sexes, implying that sound communication may play an important role in courtship. The fatty deposit that appears on the female pectoral spines of ~ ater during the spawning season might also be associated with sound communication. The fat deposit would increase the density of the spine, which in turn would alter the pitch of the sound produced during stridulation. As G. feliceps females do not develop this fatty deposit it is conceivable that this phenomenon in G. ater may serve as an effective specific mate recognition system (sensu Paterson 1978) for reproduction in dark or highly turbid environments. In conclusion, it is evident that ~ fe1iceps and ater have ~ very similar reproductive strategies. While G. ater has a significantly lower fecundity, the relative fecundities of the two species are the same. function of body size. Fecundity does not increase as a Reproduction in both species is energetically more expensive for the male and it is possible that a small percentage of males may adopt sneaking as an alternative reproductive strategy. Energy requirements for maintenance metabolism during mouth-brooding are derived chiefly from body musculature and coelomic fat reserves, while a small proportion comes from the liver. Pair spawning occurs and the males incubate the entire brood from a single female. Sound communication may playa role in reproduction. 83 CHAPTER 4 - EARLY ONTOGENY Introduction The paternal mouth-brooding habit of leads to the reproductively active ~ feliceps and ~ male population ater being excluded from the fishery for an undetermined period each year. Since fishing effort at Port Alfred is relatively constant throughout the year (Hecht & Tilney 1989), there is a danger that differential exploitation of the sexes might lead to an imbalance in the sex ratios. As both species are monogamous, skewed sex ratios would have a detrimental effect on their population dynamics. If skewed sex ratios should arise then, from a fisheries management perspective, knowledge of the buccal incubation period would be imperative for it to be rectified. Buccal incubation periods for ariids in the literature vary between six and nine weeks (for review see Rimmer & Merrick 1983). Investigation of source material revealed, however, that these had not been reliably established and were no more than estimates (Gudger 1916; Lee 1937; Merriman 1940; Atz 1958; Tobor 1969; Rimmer 1985a). While the aforementioned authors described aspects of early development in ariids, none of them had monitored development for the duration of the mouth- brooding period. This study represents the first documented account of ariid ontogeny from activation through to release of young from the adult buccal cavity. Since attempts to induce spawning in captive animals failed, the duration of the buccal incubation period had to be determined using captured mouth-brooding animals which were transferring to aquaria in the laboratory. As the age of the embryos being incubated by wild-caught mouth-brooders was unknown it was necessary to artificially spawn adult fish and 84 record the time for embryonic development through to the stages identified from wild-caught broods . Ontogeny is a genetically determined hierarchical sequence of dynamic developmental steps, generally considered to start at activation and end at death. The rate and direction of differentiation and growth is also influenced by the extranuclear environment (Lovtrup 1974, 1984; Jantsch 1980; Hall 1983; Caro & Bateson 1986), so that developmental events are orchestrated through genetic, epigenetic and environmental interplay to produce an organism with maximum survival potential. It has been suggested that ontogeny proceeds in a saltatory fashion (Balon 1981c, 1984, 1986, 1989). Central to this theory is the idea that early development proceeds as a sequence of steps during which development is gradual, separated by periods of accelerated change, termed saltations. While the theory was not central to this study it was taken into consideration in the interpretation of the observed sequences of events. While the primary objective of the study was to determine the duration of the buccal incubation period, it was anticipated that a knowledge of the nature and timing of developmental events occurring during the early ontogeny of G. feliceps would allow parallels to be drawn with mouth-brooding species from other fish families. The study concentrated on the early development of only. ~ ater were excluded because of the ~ feliceps difficulty of capturing mouth-brooding individuals in their sub-tidal reef habitat. 85 Materials & Methods The holding system The holding facility comprised a recirculating system in which water was gravity fed from a 1000 liter header tank, through an ultra-violet filter to the incubators and then into a 2m X 0.8m X O.Sm gravel filter. An electric pump, activated by a float-switch, moved the water from the filter back to the header tank . A 15 hour light 9 hour dark photoperiod was maintained. The salinity within the system varied between 3Sppt and 38ppt. Water temperature fluctuated according to ambient thermal changes and ranged between 20.0°C and 26.0°C, with a mean of 2l.9°C. This was similar to the mean river temperature (2l.00C) range for the same time period, although the temperature in the river was greater, being between lS.SoC and 2S.0°C. The water flow rate through the incubators was held constant at approximately 25 litres per minute. Air stones were used to oxygenate the incoming water. Three different incubation chamber designs were used (Fig. 26). The funnel type was effective although at the desired flow rate the embryos began to tumble. Tumbling did not occur in the buccal cavity (pers. obs.), and it was considered inadvisable for the incubation of teloleci thaI eggs (E. Balon pers. comm.). This system was therefore abandoned. The second system comprised a series of mesh-bottomed trays stacked one on top of the other and held in a rectangular chamber in which water was introduced from above. Although this system functioned well, it was awkward in that access to the lower trays necessitated the removal of the upper trays. The third design comprised a 1.Sm section of lSOmm diameter PVC piping, blocked off at either end and with a series of large holes drilled along a horizontal plane allowing the insertion of plastic specimen bottles. A 2Smm tube at either end acted as the inlet and outlet respectively. Sections of the sides and bottoms of the specimen bottles were cut away and replaced with fine mesh gauze, 86 allowing water to pass (a) (b) 0 0 0 C> C> '" 0 '" '" '" 0 '" '" '" '" '"c , '" C> ':> C> '" '" '" C> ,~ ~ ~, .~> -J c, Q C> .." ,,:,> C> "" (e) • Figure 26 . Incubation chamber designs . (a) Funnel type, (b) tray type and (c) bottle type. Arrows indicate water flow. (Not drawn to scale . ) 87 through them. Different batches of embryos could be held independently in the incubator and all embryos were easily accessible. A problem encountered in all three incubator designs was that of invertebrate infestation, which proved to be a major stumbling block in the study. Sessile cilliates, copepods and turbellarians settled on, or burrowed into the eggs, causing high mortalities. Preventative measures including the use of fine mesh filters, ultra-violet water sterilisers and formalin and fresh water flushing proved largely ineffective . Embryos were removed from captured mouth-brooding G. feliceps on a monthly basis over a period of approximately five months. Several broods at different stages of development were observed, enabling the generation of an overall picture of the early ontogeny. A consequence of this method, however, was that only approximate ages could be assigned to the various developmental steps . The approximate duration of the entire oral incubation phase was determined by holding captured mouth-brooding fish in aquaria. Mouth-brooders were caught early on in the spawning season and in each instant two to three embryos were removed from the buccal cavity in order to establish the degree of embryonic development, and hence approximate age. Five mouthbrooders held in aquaria incubated their young to term. In order to elucidate the timing and duration of the activation and cleavage phases which were seldom encountered in the broods of wild-caught fish, ripe male and female £.:.. feliceps were caught off Port Alfred on hook and line and transported to the laboratory alive. Eggs were stripped and fertilised using macerated testes from sacrificed males . Sea water was added until the egg mass was submerged and then left undisturbed for a few minutes before individual eggs were pried apart by hand and transferred to the incubators. After three to four days the outer adhesive membrane began to flake off. At this time all 88 eggs were individually stripped of this layer by gentle rubbing between the finger and thumb. Considering the size of the eggs (mean = 12.4mm) this was an easy and effective method to use. They were kept underwater at all times and were not damaged during the cleaning process. Embryos were submerged in a petri-dish and viewed under a Nikon binocular microscope using either transmitted or reflected light, the latter from a cold light source. The embryos were viewed vertically (i.e . from above) and illustrated with the aid of a camera-Iucida attachment at variable magnification between 6.6X and 30X. In initial sketches coloured pencils were used as an aid for recording the direction of blood flow. Red was used to denote blood moving away from the heart, while blue was used to indicate vessels carrying blood toward the heart. Selected steps in development were photographed using AGFAXR 100 colour film and FUJI-CHROME 100 ASA colour slides. Some embryos were excised from the egg envelope and orientated horizontally under the microscope to facilitate observation of ventrally situated structures. Tricaine methane sulfonate (MS222) was used to anaesthetise excised embryos, inhibiting movement during photography and illustration. Skeletal development was monitored using a clearing and staining technique. The methods of Taylor & Van Dyke (1985) and Balon & Flegler-Balon (1985), specifically designed for use with small fish, were followed. Ventilation rates were monitored as follows: inhalations per minute was the number of recorded for non-brooding adult males, for mouth-brooders carrying unhatched embryos and for mouth-brooders carrying hatched embryos. The terminology of Balon (1975) was used to describe the intervals of development. The different phases, cleavage (C), embryo (E) and free embryo (F) were divided into recognisable 89 steps, each of which were assigned a code (Balon 1981a). For example, E'6 would represent the second step in the embryonic phase and the sixth step since activation. Developmental steps were subjectively identified on the strength of marked ontogenetic events and did not necessarily conform to saltations in ontogenetic development. The timing of the ontogenetic events was initially recorded in days, hours and minutes. For example 0: 12: 30 represents 12 hours and 30 minutes. Development was initially monitored every four hours. After approximately 36 hours when the 64-cell stage was reached cleavage could no longer be accurately described, development was very slow and the frequency of observations was therefore reduced. Subsequent development was recorded on a daily basis. The skeletal nomenclature adopted was after Gregory (1959). Results The yolky and hyaline eggs became strongly adhesive on being spawned. Manual stripping of gravid females resulted in a plug of small hyaline eggs (diameter 0.69mm ± 0.15mm, n = 50) being released first, which were followed by the yolky oocytes. The latter were interspersed with hyaline eggs larger than those which formed the initial egg plug (diameter 1.75mm ± 0.19mm, n = 50). The mature eggs were slightly oval in shape and measured approximately 15.65mm (± 0.80mm, n 0.44mm, n = = 50) X 13.19mm (± 50) (Plate V). Apically, the adhesive outer membrane was arranged into several bulges (usually between 5 ,and 6), which formed a stellate arrangement around the micropyle (Fig. 27). A depression in the vitelline membrane was visible immediately beneath the micropyle. At this stage the female pronucleus was randomly orientated relative to the micropyle. The yolky oocytes were bright yellow, which may be indicative of a high carotenoid respiratory pigment content (Balon 1977, 1979). The dense yolk was uniformly distributed and did not contain oil globules. 90 Plate V. Photograph demonstrating the nature of the hyaline eggs (he) and the pronucleus (pn). 6.6X Magnification. (~ - ~- ~ -am - - -m - ~-, -he - - pn Smm Figure 27. Detail of the mature egg before activation showing the stellate arrangement of folds in the adhesive membrane (am), forming the micropyle (m), pronucleus (pn), and the small, interstitial hyaline eggs (he), spawned along with the yolky eggs. Electron micrographs revealed that the outer surface of the egg envelope was rugose in appearance, while the inner surface was 91 dominated by a dense arrangement of pores, each of which was 0.6 approximately 24.8 apro~imtely ~m ~m in diameter. The thick (Plate VI). egg envelope was Plate VI. Electron micrographs of the egg envelope. a) Rugose outer surface, b) hatching enzyme damage to the pitted inner surface and c) cross section through the egg envelope. 92 After artificial fertilisation of eggs in the laboratory the longest period of embryonic development observed was 28 days, corresponding to approximately one fifth of the total mouthbrooding period (see later). The embryos died at this time due to cilliate, copepod and turbellarian infestation. These timeseries data were crucial to the study as they provided a foundation for the subsequent ageing of wild-caught broods, which were used to describe ontogenetic development beyond the age of 28 days. Descriptive early ontogeny of (Note: C = cleavage phase, E ~ = feliceps embryo phase, F = free-embryo phase.) Cleavage phase Step Cl 1 Activation: Period 0:00:00 - 0:19:00: After introduction of male gametes the egg hardened and a perivitelline space formed apically at one end of the egg at approximately 0:12:30. The yolk moved freely within the egg envelope and in all eggs the pronucleus was orientated directly beneath the micropyle. Cytoplasmic streaming was not observed. Step C2 2 Cleavage: Period 0:19:00 - 2:22:00, after which the blastodisc constituted a ball of very small blastomeres (morula), raised above the yolk surface: The first mitotic cell division occurred after 0:19:00. The second division occurred at right angles 0: 23: 20. to the first, The third division was parallel to the first occurred at approximately 1: 03: 30. at and Each successive division resulted in a decrease in the size of the blastomeres. After the 8-cell stage division was often asynchronous and cells were difficult to count. The 15, 32 and 54 - cell stages occurred at approximately 1:05:00, 1:08:45 and 1:12:15. After 2:22:00, 93 the blastomeres were minute and arranged in a near-spherical ball-shaped morula . This morula was situated in a depression in the yolk (Fig. 28), which was now punctate and covered with small, freckle-like spots . The outer, pre viously adhesive membrane, began breaking down and sloughing off. The diameter of the blastodisc had remained constant at approximately 2.1mm. 2mm f. td. e. ~ '. ( ~ ." ... '- ', . - .. . .... ~ fBJ · · -': . . ',: . . . .. ; . ";' g. I~ ---c"""---=_______ pvs .~ , -- - - - m 5mm Figure 28. Cell division after fertilisation . a) Fertilised zygote, b) 2-cell (00:19:00), c) 4 - cell (00:23:00), d) 8-cell (01:03:30), e) 32-cell (01:08:00), and f) morula (02:22:00) stages . g) Lateral view of morula (m), recessed into the yolk, and the peri-vitelline space (pvs). 94 Step C3 3 Epiboly and embryogenesis: Period 3:00:00 - 10:00:00, after which the syncytium had commenced migration around the yolksac and the embryonic shield had become evident: After approximately 04:00:00 to 04:20:00, the blastodisc had become flattened over the yolk. The germ ring was distinct as a zone of granular appearance around the blastodisc. Epiboly commenced after seven days when the syncytium began expanding over the yolk surface, within the boundary of the expanding germ ring (Fig. 29a). Embryogenesis commenced after nine days when the germ ring had advanced approximately one third of the way around the yolk mass . The embryonic shield became evident and lengthened as the syncytium advanced over the yolk, the presumptive caudal region remaining at the edge of the germ ring (Fig. 29b). There was no evidence of organic differentiation during this step. Smm (al {bl I ' ,,> - bd . .. " ,. ' Figure 29. a) Flattened blastodisc (bd) surrounded by the germ ring (gr) (04:20:00), b) early embryogenesis (9 days) demonstrating the embryonic shield (es) and the expanding germ ring. The yolk surface is speckled. 95 Embryonic phase Step E14 Commencement of organogenesis: Peri od 10 - 23 days, after which the full compliment o f somites was present, the tip o f the tail wa s free and cephal i c d i fferentiation had commenced. A simple heart, a few rudimentary vitelline blood vessels and a tube-gut had appeared: The presumptive notochord and the firs t somite s appeared after approximately 10 days. At 14 days the germ ring covered half of the yolk and approximately 1 3 pairs of somites were visible. The presumptive notochord was well defined . By the 18th day the optic vesicles had differentiated, the otocysts were distinct and 25 pairs of somites 'were visible (Fig. 30). The pericardial cavity began to form beneath t he embryo in the region between the optic vesicles and the otocysts, in a depression in the yolk. Differentiation of the brain into left and right halves occurred in the region anterior and adjacent to the otocysts . 3mm ;.--- -~- ov pnc 1/4Y------ -ot " , - - - -- - -- gr Figure 30. Beginning of cephalic differentiation at the 25somi te stage ( 1 8 days). Presumptive notochord (pnc), optic vesicles (ov ) and otocysts (ot). Epiboly approximately one third complete. 96 By day 20, ep i boly was two thirds complete and the tail was free from the yolk-sac and had lifted above it. The tail was short, thick and rounded . The full compliment of somi tes, approximately 45, were visible . The optic vesicles were well defined. The hind-brain area showed invagination and differentiation into left and right hemispheres. The otic capsules were distinct and oval in shape, although otoliths were not visible. The embryo was 5.4mm long . The head length (measured up to the first somite) was 1 . 9mm, and the postanal trunk length was 1.9mm. Removal of the egg envelope and manipulation of the embryo at 22 days revealed a simple tube-like heart pulsing at 24 beats per minute, although no blood cells were visible. Olfactory placodes were present. The mid- and hind-brain demonstrated considerable invagination. Three pairs of rudimentary gill pouches protruded ventro-laterally immediately posterior to the otic capsules . Two pairs of otoliths were visible in the otic vesicles. Epiboly was three quarters complete. The somites had taken on a chevron-like shape and resembled muscle myomeres. Step E2 S Period 24 - 26 days: Completion of epiboly, first appearance of red blood cells, development of vitelline circulation and additional differentiation: At 24 days, the first muscular contractions were witnessed in the form of periodic tail lashing and of the trunk w~igln and head region. A fairly large rugose area around the embryo interspersed with an incomplete network of shallow grooves and islands marked circulatory the plexus expansion of over yolk. the the primordial The first vitelline established vitelline blood vessels formed a simple loop on either side of the embryo, although no red blood cells were visible (Fig . 31). 97 Q, ~-c - - vp .... ; " -, - - r'--- - y p b. 2mm fb mb hb gp mm Figure 31. a) Completion of epiboly (24 days), marked by closure of the yolk plug (yp). First appearance of vitelline blood plexus (vp) . b) Magnified image of (a) above showing cephalic differentiation into fore- (fb), mid- (mb) and hindbrain (hb). Lenses (1) and 3 pairs of gill pouches (gp) are evident . The somites have taken on the shape of muscle myomeres (mm) • After approximately 25 days, red blood cells were visible for the first time. Large blood sinuses occurred in the vitelline plexus, which had branched considerably and covered approximately one third of the yolk surface . It comprised left and right lateral vitelline veins, arising from the left and right branches of the anterior and posterior cardinal veins, and transported blood from the embryo onto the yolk surface. Here they branched profusely before draining into anterior left and right vitelline veins which returned the blood to the sinus venosus. In addition, the posterior vitelline vein, arising 98 from the caudal vein, appeared for the first time. Blood vessels were also visible in the embryo, extending anteriorly to the branchial and cranial regions, where blood flow through the branchial posteriorly arches to the was tip evident of the for the tail. first The time, and forebrain had differentiated into two distinct lobes and further invagination and differentiation of the hindbrain had occurred. Dorsal and ventral fin folds were present. The intestine was present as a simple, straight tube (Fig. head length 1.9mm and 32). Embryo length was 7.2mm, trunk length 3.2mm. post~nal 3mm 1 +i;4!~- , , \' \' "\ - - - - rQV V •• IJ -"~ \ ,+ - W- PV " Figure 32. 25 day-old embryo . Note the left and right lateral vitelline veins (rlvv) which radiate extensively before draining into the left and right anterior vitelline veins (ravv) respectively . The posterior vitelline vein (pvv) has begun to differentiate . 99 StepE 3 6 Period 27 29 days: Vascular proliferation, retinal pigmentation, opening of the mouth and nostrils, appearance of barbel and pectoral fin buds: After 27 days the eyes had become prominent on either side of the head and the retina had become faintly pigmented. The forebrain lobes were very prominent and were supplied with superficial blood vessels. Maxillary barbel buds were present as were the nasal apertures. Ventrally the rudimentary mouth was open, although incapable of movement, highly vascularised tube-gut . and lead into a Prominent dorsal and ventral intestinal blood vessels were present. A white substance was visible in the intestine. Four gill arches were visible but gill lamellae had not yet formed. An opercular fold, situated anterior to the gill arches had become prominent. The heart was differentiated. The nature of the branchial by-pass system could not be determined due to the opacity of the head region, a phenomenon which obstructed subsequent identification of subcutaneous blood vessels. Hemi-spherical pectoral fin buds were visible for the first time. Considerable vascularisation had occurred in t.he abdominal region ventral to the notochord, presumed to be the renal plexus, where a network of very fine blood capillaries had resulted in a reddening of the area. A red organ, possibly the presumptive spleen was situated adjacent to the anterior region of the gut. The anterior left and right vitelline veins began to anastomose in the region anterior to the head, a process which lasted between 4 to 5 days before completion. The posterior vitelline plexus had branched and expanded considerably, and the entire vitelline plexus covered approximately half of the yolk surface . The caudal fin fold was pointed, and the notochord straight (Plate VII; Fig. 33). The embryo was 8.9mrn long. The head length (to anterior of pectoral fin bud) was 2.0mrn. The postanal trunk length was 3.4mrn. 100 From this point onwards, ontogenetic development was recorded using embryos from wild-caught mouth-brooders only. B Plate VII. a) Photograph of an embryo (egg envelope removed) at approximately 25 days showing the extent of the vitelline plexus (Mag. 6.6X). b) Higher magnification demonstrating the proliferation of blood vessels in the head and abdomen (25X). 101 3mm ~ - - an Figure 33. Commencement of anastomosis (an) of anterior left and right vitelline veins (27 days). Expansion of posterior vitelline plexus. Appearance of pectoral fin buds (pb) and retinal pigmentation. Caudal fin-fold (cf) is pointed. Step E4 7 Period 30 - 34 days: Notochordal flexion, chondrification of the branchial basket, complete anastomosis of anterior left and right vitelline veins, expansion of posterior vitelline plexus: After approximately 30 days anastomosis of the anterior left and right vitelline veins was well advanced and the vitelline plexus had expanded considerably. The posterior vitelline circulatory system had linked laterally on the yolk surface with the anterior system for the first time, marking the beginning of the capture of the anterior circulation and the disappearance of the left and right lateral vitelline veins. The highly vascularised presumptive liver was visible adjacent t o , and in close association with, the left lateral vitelline vein (hepatic vitelline vein). The notochord had undergone dorsal flexion in the caudal region, and the caudal fin had 102 taken on a heterocercal chondrification for the shape. first The branchial basket time, of four pairs showed of gill arches. The large lapillus otoliths were prominent. The embryo was 9. 9mm long, with a head length of 2. 4mm and a postanal trunk length of 3.7mm After approximately 32 days, the anastomosis of the anterior left and right vitelline veins was complete and the left anterior vitelline vein dominated the flow of blood returning to the heart. As a result of a considerable increase in the size of the presumptive kidney, deeply reddened posteriorly and and swollen. covered the the renal plexus had become The first operculum two gill had extended arches . Dorsal segmental arterioles and venules were visible in the trunk and tail regions. Some of the presumptive pleural ribs were visible as seven lateral vascularised extensions over the yolk. The gall bladder was prominent and was attached to the liver midlaterally (Fig. 34) . The sub-intestinal artery, a branch of the coeliaco-mesenteric artery, was large and prominent and joined the circulation of the inferior caudal vein in the vent region, at the origin of the posterior vitelline vein. The embryo was 11.1 mm long, with a head length of 2.9mm and a postanal trunk length of 4.2mm . 103 3mm ""-"i"""=- - - - cf pCV acV 4--.F"'r-- c V -,' aVv- ~ ,. ,=:-\ -da --" li- --" gb k Figure 34. Anastomosis of left and right anterior vitelline veins (32 days) has resulted in a single anterior vitelline vein (avv) . The anterior (acv) and posterior cardinal veins (pcv) unite to form the lateral vitelline veins . The dorsal aorta (da) and caudal vein (cv) are prominent. The posterior vitelline plexus has expanded considerably and has linked up laterally with the anterior vitelline plexus. Considerable vascularisation in the vicinity of the presumptive kidney (k). Intestine (i) a simple tube. Presumptive liver (li) and gall bladder (gb) developing in close association with left lateral vitelline vein (also known as hepatic vitelline vein). Maxillary barbel buds (mb) and an opercular fold are prominent. Four gill arches visible . Caudal fin heterocercal in shape. Step ESe Period 35 - 54 days: First ossification, fin differentiation, increased branchial circulation, loss of lateral left and right vitelline veins, first pigmentation, appearance of swim bladder: After approximately 35 days the maxillary barbels had elongated and protruded posteriorly beyond the eyes. The operculum had extended posteriorly and covered the first two gill arches. The subclavian veins were visible in the pectoral fin buds, which had become slightly pointed posteriorly . The intestine 104 had developed a slight bend anteriorly, indicating an increase in length and the onset of the stomach differentiation. Oss i fication was evident for the first time. The left and right had begun ossification, marking the pectoral girdle as the area of earliest skeletal maturation (Fig. 35) . Meckels cartilage was prominent and the notochord was strongly clei~hra cartilaginous . Embryo length was 11.lmm, head length 2.9mm and postanal trunk length 4.8mm . 2mm P9 - --M. sc - - -'" Figure 35. Ventral view of stained and cleared specimen (35 days) showing first ossification in the pectoral girdle (pg). Meckels cartilage (Mc) and spinal column sc) well chondrified. After approximately 38 days the anastomosis of the left and right anterior vitelline veins on the ventral aspect of the yolk was complete, and the single sub-intestinal vitelline vein now collected and transported blood anteriorally to the duct of Cuvier and the heart. The position of the gall bladder had changed. It was now situated ventral to the antero-laterally directed liver and contained a green substance, presumably bile. The dorsal fin fold had become enlarged in the region of the presumptive dorsal fin. The circulation in the caudal area had become more directed caudal loops, elaborate with heralding the condition. The the addition beginning of pectoral fin buds of ventrally the homocercal had elongated considerably in a posterior direction. The regions of the brain were well defined and a complex network of cutaneous blood vessels was visible on the head. The large lapillus otoliths 105 were distinct in the otic capsule (Plate VIlla operculum covered three of the branchial arches. & b). The The fully formed mouth opened and closed synchronously with opercular movements. The mandibular barbels were elongate. Embryo length 12.2rnrn, head length 2.9rnrn, and postanal trunk length S.9rnrn. Plate VIII. a) Embryo at approximately 38 days (egg envelope removed) showing differentiation of the dorsal and anal fins from the fin folds. Caudal fin shows early homocercal condition. The posterior vitelline plexus has joined with the anterior vitelline plexus laterally and now carries most of the vitelline blood (6.6X). b) Higher magnification (2SX) showing the proliferation of cutaneous blood vessels on the head, prominent lapillus otoliths and reduction in size of left lateral vitelline vein. 106 Two days later, at approximately 40 days, pectoral, dorsal and caudal fin lepidotrichia were visible for the first time. The operculum had completely covered the gill arches and the heart was nestled between the posterior-most left and right pair. Some differentiation of the stomach was visible as an L-shaped convolution in the anterior region of the gut. The white substance in the intestine extended approximately halfway to the vent. The swim bladder was visible for the first time, nestled anterior to the kidney, dorsal to the stomach. A yellow pigmented organ was visible in the coelomic cavity situated to the right of and adjacent to the stomach, probably the spleen. The liver was embryo, large and anteriorly directed lateral to the extending anterior to the heart. The spherical gall bladder was filled with bile. Dark pigmentation began to appear on the ventral rudimentary surface of gonad material the kidney. was What may have been visible as paired threads running medio-ventrally to the kidney, ending near the vent. After 43 days, melanophores appeared on the head for the first time, being concentrated over the optic lobe and cerebellar areas. Simple olfactory rosettes were visible through the olfactory nares. The pectoral and dorsal spines and rays were differentiating. The pelvic fin buds had appeared. More of the blood from the caudal vein began flowing into the posterior cardinal vein, largely by-passing the posterior vitelline vein and the vitelline vitelline vein respiratory (hepatic complex. vitelline vein) The left lateral had moved medio- laterally underneath the embryo together with the liver. The hepatic vein now lead directly into the duct of Cuvier. The right lateral vitelline vein had also decreased in size, indicating that some of the blood from the right anterior and posterior cardinal veins was flowing directly into the ductus Cuvieri for the first time (Fig. 36). 107 3mm 411;- - - cf Ii- -fl' / - - --+ \ tH~"¥1A,- ~f-r,>.1tI"\ ~j:-ti!f df' -~ - lo op Pf I"'-jllbl'!/'l-- - d f f 'l -,e?~ i- S bv Figure 36. The left anterior vitelline vein has moved anteromedially underneath the embryo along with the liver (43 days), and no longer serves a respiratory function. Right lateral vitelline vein reduced in size. Posterior vitelline plexus now dominant . Segmental blood vessels (sbv) prominent. Vascularisation of dorsal fin fold (dff). Advanced differentiation of the brain. Lapillus otoliths (10) large. Operculum complete, pectoral (pf), dorsal (df) and caudal fins (cf) rayed. Caudal fin fully homocercal. In one embryo, blood in the right lateral vitelline vein was seen to have reversed its flow, carrying blood from the vitelline plexus into the embryo. This may however have arisen as a result of damage to the vitelline circulatory system during handling. The by-passing of the vitelline circulation system was an indication that the gills were actively involved in respiration. However, examination of the gills revealed the presence of short, stumpy primary lamellae only. There were no secondary lamellae present. Ossification was evident proximally in the pectoral and dorsal spines and pterygiophores, in the operculae, dentary, the four 108 outermost branchiostegals, the maxillae and the centra of all the vertebrae excepting the caudal-most three (Fig . 37; Table XI). The embryo was 13.9mm long . The head length was 3.0mm and the postanal trunk length was 6.9mm. Figure 37 . Ventral view of stained and cleared specimen (43 days) showing degree of ossification. Pectoral girdle (pg), dentary (dt), maxillae (mx), operculae (op), 4 pairs of branchiostegal rays (br), vertebral centra, dorsal, pectoral and caudal fin pterygiophores (pty) . Basi-occipital (bo) partly ossified . Ceratohyal (chy) and articular (ar) chondrified . After approximately 50 days, lepidotrichia appeared in the pelvic fins . Dorsally, pigmentation extended laterally to the eyes and posteriorly to the nape region . Segmental arterioles and venules were prominent in the dorsal and ventral fin folds posterior to the differentiated dorsal and anal fins. Additional ossification appeared in the superior limb of the post-temporal, interopercular, frontal, para sphenoid and exoccipi tal . All the vertebral centra showed some ossification. The pre-anal vertebrae showed ossification of the neural and haemal spines. 109 XI: Summary of the degree of skeletal ossification during step ESe, after approximately 43 days. Table DEGREE Of OSSIFICATION NONE SUGHT MODERATE ADVANCED Pectoral spine Pectoral fin pil!ll'/glophorea Clelthrum- ventral limb post humoral process Ceratohyal HypohyaJ Epihyal UrohyaJ 8ranchiostegllls Den\aly Operculum Preopercle lnlflropercle Hyomaodlbutat Quadrale Metapt81}'gOid Vertebnll contra Vertebral spInes Fused vertebral compte)! Transverse process 4th vertebra Weberlan osslcles DORSAL NEUROCRANIAL SKELETON Dermosupraocclpltal Pterotic SphenoUc Post temporal Sup. post temporal Pre-frontal Frontal MaxilllllY Dermethmold VENTRAL NEURcx::RANlAl SKELETON In!. post temporal Basioccipital Pro-otic Parasphenold Orbllosphenold Prevomer Vomer CAUDAl SKELETON Pre ural centra Uranurs.1 Parhypural Hypurals Caudal pleryglophores 6 Step E 9 Period 55 75 days: Growth, skeletal ossification, lateral line development and dermal pigmentation after completion of fin differentiation: 110 Pigmentation was extensive after approximately 55 days, covering the entire cranial, nape and trunk areas, reaching the caudal visible. peduncle . The Few eyes s ubdermal had features become were surrounded by subsequently cranial and neurocrania I elements and were no longer prominent on either side of the head. The full compliment of pterygiophores was present in both the paired and unpaired fins . The dorsal and ventral fin folds were prominent and well vascularised . The embryo was large and the trunk reg i on was recurved so that the caudal fin was situated anteriorly, adjacent to the head (Fig. 38) • 3m m ~ ~V i-If I'<-\ - - - b m df Figure 38. Dorsal view of an embryo after approximately 55 days. Differentiation of pelvic fin (pvf) marks the completion of fin differentiation. Considerable pigmentation of dorsal regions of head, nape and trunk . Vitelline circulation completely dominated by the posterior vitelline plexus. Dorsal (dff) well vascularised. Maxillary (mb) and mandibular barbels fully formed . 11 1 Lateral line vesicles were present on the head, extending to the posterior margin of the cleithrum. The head lateral line was represented by a single anteriorly directed dorso-Iateral branch on either side of the head, extending above the eye and ending medially to the nostril, length 18. 8mm, at the premaxilla . Embryo head length 3. 8mm and postanal trunk length 10.0mm. Ossification proceeded rapidly during this period and the degree of ossification is presented in Figure 39 and Table XII. (a) 2mm (b) fr- - f,--/hyo pteop _ pop ____~ · ~"WI - So( ~f61_spt ({ Figure 39. a) Dorsal view of stained and cleared specimen (± 55 days) illustrating the degree of ossification (represented by blackened areas). Supra-occipital (soc), supra-post-temporal (spt), cleithrum (cl), pterotic (pte), operculum (op), preoperculum (pop), hyomandibular (hyo) and frontal (fr) well ossified. b) Ventral view: Maxillae (rnx), pre-maxillae (prnx) , dentary (dt), articular (ar), hypohyal (hhy), epihyal (ehy), 6 pairs of branchiostegal rays (br), clei thrum (cl), basioccipital (bo), transverse process of 4th vertebral centrum (trp), pleural ribs (pr) and neural and haemal vertebral arches. Ceratohyal (chy), coracoid (co), inferior post-temporal (ipt) and caudal skeleton incompletely ossified. 112 Table XII: Degree of skeletal ossification during step E6 9 , just prior to hatching (approximately 65 days). NONE DEGREE OF OSSIFICATION MODERATE ADVANCED SUGHT PECTORAL & NEURAL SKELETON Pectoral spine Pectoral tin pleryglophofea Clellhrum • ventral 11mb . post humeral process Ceratohyal HypohyaJ Epihyal UrOhyal Branchioslegllis Oentary Operculum Proopercle Inleropercle HyomandibulOl Ouo.drll1e Melaplerygold Vertebral cenlra Vertebral spines Fused vertebral complex Transverse process 4th vertebra Weberlan osslcles OORSAL NEUROCRANIAl SKELETON Oermosupraocclphlll Plerollc Sphenollc Post temporal Sup. post temporal Pre-frontal Frontal N.... Dermethmold VENTRAL NEUAOCRANIAL SKElETON Inl. post temporal Basloccipjtal Pro-otic Parasphenoid Orbitosphenoid Prevomer Vomer CAUDAl SKELETON Preural centra UronuraJ Parhypural HyPurais Caudal plerygIophores After approximately 65 days pigmentation was dense and covered the entire dorsal surface including the fins, less densely distributed . The dorsal were devoid of pigmentation. largely undifferentiated. where it was and ventral fin-folds The digestive tract was still The pyloric sphincter was present 113 and the stomach was still L-shaped. The intestine was a simple, straight tube through to the vent . The liver had increased in size but still had only a single lobe and the spleen was large and orange in colour. The primary lamellae on the gill arches had elongated and secondary lamellea were clearly visible. The swim bladder was nestled between the kidney and the head kidney and was dumb-bell shaped. The embryo had grown considerably in length and in girth. The total length was 24.2mm, head length 4.3mm and post anal trunk length 13.2mm. Several days prior to hatching the embryo appeared extremely constricted within the vitelline membrane, being doubled over wi th the caudal f in extended over and beyond the head. The heavily vascularised dorsal and ventral fin folds were s till prominent. The vitelline plexus, now originating entirely from the posterior vitelline vein, was profusely branched and a fine network of capillaries, covering the entire yolk-sac, was interspersed between the larger vessels (Fig. 40). Summary of the vitelline circulatory development The left and right lateral vitelline veins arise from the paired anterior and posterior cardinal veins which meet and flow adjacent to one another laterally over the yolk for a short while capillaries. before branching into a complex network of These capillaries eventually collect into the large left and right anterior vitelline veins, which meet just prior to entering the sinus venosus and the heart. The return flow from the caudal peduncle is via the caudal vein. While most of the blood from the caudal vein flows into the posterior vitelline vein (which later becomes the sub-intestinal vein), some also enters the posterior cardinal vein . Later in the development, the blood flow from the caudal vein is re-routed and flows primarily into the posterior cardinal vein, which in turn drains into the duct of Cuvier. The intestinal and gastric veins lead into the liver via the hepatic portal system, and the hepatic vein carries blood venosus and the heart. 114 from the liver into sinus 3m m . : ,. :,': : :; :': ? ' · :;_( : :~y;Yr II Figure 40. Embryo shortly before hatching (± 70 days). Extremely profuse posterior vi telline plexus, well vascularised dorsal (dff) and ventral fin folds (vff), dense pigmentation throughout, lateral line (11) prominent . Embryo large and constricted wi thin egg envelope, caudal peduncle and fin doubled back and situated dorsal to the head The left and right anterior vitelline veins anastomose in the region immediately anterior to the head after approximately 27 days, resulting approximately 30 - in blood flow 31 days), being redirected (after to the left anterior vitelline vein, which expands to accommodate the drainage from the entire vitelline plexus . As the vitelline network expands the anastomoses of the two branches of the anterior vitelline vein continues in a zip-like fashion subsequently mid-ventral plane in the mid-antero (Plate IX), until the entire yolk is completely enveloped by the vitelline plexus. 115 and rvv_ -Ivv Plate IX. Antero-ventral aspect of an embryo (± 30 days) demonstrating the progressive anastomosis of the anterior left (lvv) and right (rvv) vitelline veins in a ventral direction. The completion of this anastomosis results in the formation of a single sub-intestinal vitelline vein. At this stage the left lateral vitelline vein (hepatic vitelline vein) has disappeared from view along with the liver which has moved antero-medially inwards and underneath the embryo. The blood flow in the right lateral vitelline vein becomes captured by, and incorporated into, the expanding posterior vitelline vein network and also disappears. The anterior and posterior cardinal veins, which initially fed the right lateral vitelline vein, now flow into the hepatic portal system and the duct of Cuvier respectively. A few days later, some of the blood from the caudal vein bypasses the posterior vitelline plexus and is re-routed into the posterior cardinal vein. This may be indicative of the completion of the branchial circulatory system. The posterior vitelline plexus continues to proliferate, however, and forms a dense network of capillaries around the entire yolk-sac 116 lead into the sub-intestinal vitelline vein . The vitelline plexus is finally enveloped by the expanding ventral body wall. Free-embryo phase Step FllO Period 75 - 100 days: Hatching, straightening of the trunk region, more growth and ossification, continued differentiation of the digestive system and the onset of exogenous feeding: Hatching occurred after approximately 75 to 80 days, as evidenced by the ejection of egg envelopes from the buccal cavity in aquarium-held mouth-brooders. Embryos hatched headfirst through a tear in the egg envelope which commenced in the region immediately dorsal to the head. The dorsal and ventral fin folds in the caudal region were still present at hatching. The lateral line extended from the caudal peduncle to the head where it branched and completely encircled the eye and nostril. Additional branches extended anteriorly across the frontal to the upper lip, postero-dorsally across the pterotic and medially across the sphenotic . The left and right branches on the head did not meet, however . The stomach had differentiated and resembled the adult U-shape. The intestine exhibited considerable elongation and was thrown into two loops, one beneath the stomach and the other posteriorly, near the vent . The swim bladder differentiation was complete and the characteristic heart shape had been reached. The pneumatic duct was large and entered the oesophagus antero-ventrally to the swim bladder. The primary lamellae on the gill arches were considerably elongated, although they appeared on only four of the gill arches. Total length at hatching was 32.5mm (± 1.5mm, n = 35). Head length 5.8mm, postanal trunk length 15.9mm. Ossification of the skeleton was skeletal elements were at moderately advanced. least partly ossified, exception of the pleural ribs (Table XIII; Fig. 41). 117 All with the Table XIII: Skeletal ossification during step FllO, immediately following hatching (Approximately 80 days). NIL DEGREE OF OSSIFICATION SUGKT MODERATE ADVANCED PECTOAAL & NEURAL SKELETON Poetora] spino Pectoral fln pleryglophores CleUhrum • venllal limb Clelthru m . post humeral process Ceralooyal HypohyaJ Epihyal UrohyaJ Branchloslegals Dentary Operculum Preoporc!. Inleropercla Hyomandlbu!ar Quadrate MeiaptOIY90ld Vertebral centra Ver1ebral splnas Fused vertebral comptex Transverse process 4th vertebra Weberllll'lossicles DORSAL NEUROCRANIAl SKElETON Derm09upraocclpitai Pterolic Sphenotic Post temporal Sup. post tomporaJ Pra-frontal Frontal MAXillary Dermethmold VENTRAl NEUROCRANIAL SKELETON Int. post temporal Basioccipital Pro-otic Paruphenold Orbitosphenoid Prevomer Vomer CAUDAL SKELETON Preural centra Uronural PamypuraJ HypuraJa Caudal p\eryglcphore8 ------------------------------------------------------- - --- 118 (a) de - - - - - -fr _- ______ sph Smm (b) mx d pop iop ehy ~f co bo ps ip Figure 41. a) Dorso-lateral view of cleared and stained specimen ( ± 90 days): Premaxilla (prnx), dermethmoid (de), maxilla (rnx), frontal (fr), articular (ar), nasal (na), sphenotic (sph), pterotic (pte), supra-occipital (soc), hyomandibular (hyo), epihyal (epy), interopercle (iop), preopercle (pop), operculum (op), cleithrum (cl), superior post-temporal (spt), dorsal and pectoral sp~ne. Caudal skeleton incompletely ossified. b) Ventral view of head region: Premaxillae (prnx) and dentary with teeth, maxillae (rnx), hypohyal (hhy), urohyal (uhy), ceratohyal (chy), epihyal (epy), coracoid (co), basi-occipital (bo), inferior post-temporal (~pt), pectoral spine (ps). Lapillus, sagittal and astericus otoliths prominent. Weberian ossicles incompletely arranged . Tripus (tr) prominent. 119 The size of the free-embryo yolk-sac at hatch i ng was c onsiderable (Fig . 42), and was found to severely limit their swimming ability. 10m m , .'\ Figure 42. Lateral view of a free-embryo shortly after hatching (± 90 days), showing full differentiation of all external structures . The yolk-sac is large and well vascularised. Step F211 Period 100 - 140 days: Completion of skeletal ossification, more growth, absorption of yolk-sac, elaboration of lateral line system, intensification of pigmentation and release from the buccal cavity: After approximately 110 days, the ventral body wall had grown halfway around the yolk-sac. The dorsal and ventral fin folds had disappeared and the embryo had darkened in colour. Internally the liver had increased considerably in size and occupied the area anterior and dorsal to the stomach. Dorsally it abutted against the greatly enlarged swim bladder . Food particles were evident in the stomach while the intestine had elongated considerably to form several loops wi t hin the abdominal cavity. At 27mm it was approximately two-thirds of the free embryo length (44mm TL), (Fig . 43) . 120 10mm Figure 43. Lateral view of a free-embryo after approximately 120 days showing advanced absorption of yolk-sac and proliferation of the head lateral line. • Release of juveniles occurred once the yolk sac had been totally absorbed and the ventral body wall had knitted, after approximately 140 days. Total length at release was 54.3mm (± 2.0mm, n = 50). Traces of yolk were still visible internally, around the viscera. The juveniles were morphologically similar to the adults at this time although structures such as the olfactory rosettes, secondary gill lamellae and head lateral line were incompletely developed, and would not show complete development until sexual maturity (Fig. 44). 10mm Figure 44. Lateral view of a juvenile ~ feliceps at time of release from the adult buccal cavity (± 140 days). Pigmentation is dense, yolk-sac completely absorbed and head lateral line (hll) more extensive. 121 Skeletal ossification at the time of liberation was almost complete. The only bones still exhibiting some cartilaginous material were in the cranium and neurocranium (Table XIV; Fig. 45) • Table XIV: Degree of ossification at the time of release from the adult buccal cavity (Approximately 140 days). DEGREE OF OSSIFICATION NONE SUGHT PECTORAL & NEURAL SKELETON PedOf8i spine Pectoral fin pterygiophOfes Cleflhrum - ventmJ 11mb . post humeral process Ceratohyal HypohyeJ Epihyal Urohyal 8rl!l.m;:hiostegals Oentasy Operculum Preopercle (n(aroperele Hyomandibular Quadrate Metapterygold Vertebral centra Vertebral spines Fused vertebral complex Transverse process 41h vertebra Weborian osslcles DORSAl NEUROCRANtAL SKELETON DermoBupraoct:lpllal Plerotte Sphenotic Post temporal Superior post temporal Pre-fronlal Frontal Maxillary Dermethmold Palatine VENTRAL NEUAOCRANIAL SKELETON Inf. post temporal Basioccipital Pro-otic Parasphenold OrbitosphenoId Prewmer Vomer CAUDAL SKELETON PreuraJ cenlra Uronural Perhypural HypuraJs Caudal pt8f)'giophOfes 122 NKJDEAATE ADVANCED .------. (a) (b) 3mm Ie de -- - - pmx dt or ph~O P lOP eI eoe op 8~e trp (c) puhU1 ~ ns1 1mm ~ pu2 - -- 2 _ _ por ~ ---- -hs1 Figure 45. Prepared dry skeleton of a juvenile at time of release (± 140 days). a) Dorsal view of head region: Lateral ethmoid (Ie), dermethmoid (de), maxillae (mx), premaxillae (pmx) , dentary (d) , pala tine (pal) , articular (ar) , metapterygoid (mpt), quadrate (qua), hyomandibular (hyo), frontal (fr), pre-opercular (pop), inter-opercular (iop), frontal (fr), sphenotic (sph), pterotic (pte), supra-occipital (soc), clei thrum (cl), opercular (Op), epiotic (ep), transverse process of 4th fused vertebra (trp). b) Ventral view: Hypohyal (hhy), urohyal (uhy), ceratohyal (chy), epihyal (ehy), 6 prs . branchiostegal rays (br), coracoid (co), pro-otic (pot), pterotic (pte), inferior post-temporal (ipt), basioccipital (boc), exoccipital (eoc), tripus (tr) . c) Caudal skeleton: Preural centra (pu2, pu3), fused complex ural centrum (pul + ul), first neural spine (nsl), epural (epu), uronural (uro), hypurals (hyp 1-5), parhypural (par), first haemal spine (hsl) . 123 Artificial fertilisation and incubation of ~ feliceps eggs and embryos together with evidence gleaned from a series of wild~caught broods indicated that the incubation period up to the time of hatching was in the region of 75 to 80 days. The incubation time subsequent to hatching was less clear. Field evidence demonstrated that the size of embryos at release from the adult buccal cavity was approximately 54mm (± 5rnm, n = 53). Aquarium-held mouth-brooders demonstrated an incubation period between hatching and release from the buccal cavity of exactly 65 days (n 2). Approximately 21 days after hatching = adult fish showed an interest in the and ingested food for the first time, marking the onset of exogenous feeding by the freeembryos. Free-embryos held in the incubators were fed regularly from the time of hatching and reached the 'release length' after approximately 35-45 days, 20 days earlier than mouth- brooded free-embryos. This indicated that the super abundance of exogenous food available to the incubator-reared free embryos might have artificially accelerated their growth rate. The youngest embryos mouth-brooders that were, were using incubated the data by aquarium-held from artificially fertilised and incubated embryos, estimated to be approximately 14 days old at capture. The incubation times through to release of young in these mouth-brooders (n = 2), was 129 and 131 days respectively . This was representative of a total incubation period between fertilisation and release of young of approximately 143-145 days. The batches of embryos being incubated by the two mouth- brooders referred to above were judged to be the same age at the time of capture. Notably, in these two mouth-brooders,both hatching (evidenced by the ejected egg envelopes seen in the aquarium) and release of young occurred exactly two days earlier in the one individual than it did in the other. This was an indication that, under the particular set of environmental conditions present in the laboratory, ontogenetic development in ~ feliceps proceeded at a very definite rate. 124 The growth in total, head and postanal trunk length, between the 20th day (25 somite stage) and the time at release from the buccal cavity (± 140 days) is presented in Figure 46. The early ontogeny was marked by irregular growth until approximately the 50th day (coinciding with the completion of differentiation), after which it was consistent. The relatively few data points used in the construction of the curves may have resulted in artificially smooth growth rates. 60 '0 40 ? ~ ~ ~ w 30 ~ 20 '0 a 0 20 AGE 0 TOTAL LENGTH + C~yS) HEAO LENGTH o TRUNK LENGTH Figure 46. Growth in total, head and postanal trunk length between day 20 and day 140 . Growth was irregular until approximately the 50th day, coinciding with the completion of differentiation, after which it was consistent. The changes in adult ventilation rates associated with mouthbrooding are graphically presented in Figure 47. Ventilation rates did not change gradually over time in accompaniment with embryonic development. They were relatively constant for the duration of the embryonic phase and then exhibited a sudden, significant decline after hatching (adjusted Chi-sq.= 8.605, 125 df=l, P > 0.01), (the ventilation rate of the non-brooding, actively feeding control animals was used as the expected value in the contingency table for the Chi-Square test). This decline was probably enabled by an increased oxygen availability to the free-embryos after shedding of barrier to oxygen diffusion . respiratory apparatus maximal oxygen was extraction the Also, egg envelope, the embryonic functional from the at a former branchial hatching, enabling incurrent water in the parent buccal cavity. As the ventilation rate of mouth-brooders carrying control free-embryos animals, metabolic was significantly lower than that of was also indicative of this oxygen requirement (for basal a much metabolism) lower during buccal incubation than during periods of active feeding. 75 t '0 65 Z 60 " VI z Q 55 ~ 'i" ~ ~ so 45 40 t , 35 CONTROL ---e-- n = :50 HATCHED UNHATCHED n = "1 "1 ~ n • 102 Figure 47. Ventilation rates (inhalations min- I , ± I SD) for feeding adults (control), mouth-brooders carrying embryos, and mouth-brooders incubating free-embryos . The major developmental events describing the stepped continuum in G . feliceps early ontogeny are summarised in Table XV. 126 Table XV: Sununary of Galeichthys feliceps. Developmental step the early ontogenetic development of Important events Age (Days:Hrs:Mlns) Activation 0:00:00 - 0:19:00 C'1 Cleavage C'2 0: 19:00 - 2:22:00 Epiboly & Embryogenesis 3:00:00 - 10:00:00 Formation of perivitelline space after 0:12:30. First cell division at 0: 19:00. Morula at 2:22:00. Adhesive membrane sloughs off. C33 Blastodisc and germ ring present atter 4 days. Epiboly commenced after? days and embryogenesis after 9 days. Germ ring 1/3 of the way around yolk. Organogenesis First somites after 10 days. Optic vesicles & otaeysts after 18 days. 10 - 23 days Full compliment of somites E'4 (45) after 20 days. Cephalic differentiation . Formation of tube-heart & pericardia! cavity. 3 prs. rudimentary gill pouches. Epiboly 3/4 complete. 24 - 26 days Left & right lateral vitelline veins after 24 days. First embryo movement - tail lashing and trunk wriggling. Epiboly complete. Rrst red blood cells after 25 days. Dorsal & ventral fin folds. TUbe-intestine. 27 - 29 days Prominent eyes & retinal pigmentation . Nasal apertures, rudimentary mouth, maxillary barbel buds, pectoral fin buds. 4 gill arches. Differentiated heart. Vascularised renal plexus. Anterior left & right vitelline veins begin anastomosis. Posterior vitelline plexus expands. Vitelline plexus covers '/2 yolk surface. Caudal 30 - 34 days fin fold pointed. Anterior & posterior vitelline plexuses linked. Uver developing in association with left lateral vitelline vein. Dorsal flexion of notochord in heterocercal caudal fin. Chondrification of branchial basket. lapillus otoliths prominent. Gall bladder after 32 days. Subintestinal artery prominent and receives blood from caudal vein. 35·54 days First ossification in cleithrum. Meckels cartilage prominent. Vitelline plexus completely envelopes yolk. Synchronous mouth and opercular movements. Pectoral, dorsal & caudal fin lepidotrichia after 40 days. Swim bladder present. Stomach differentiation . Gall bladder filled with green bile. First melanophores on head after 43 days. Pelvic fin buds appear. Blood from caudal vein bypasses posterior vitelline vein plexus. 55 • 75 days Paired fin differentiation complete. Dorsal & ventral fin folds well vascu larised . lateral line vesicles on trunk & head . Extensive pigmentation dorsally. 2 gill lamellae. Embryo doubled over with caudal fin covering head region. Posterior vitelline plexus profusely branched. F'1O 75· 100 days Hatching at 75-80 days. Gastro-intestinal differentiation complete. Exogenousfeeding commences. Ossification moderately advanced, all skeletal elements partly ossified. 100 - 140 days Period 0"1 growth and ossification. Fin folds disappear. Yolk-sac 50% resorbed after 110 days. Yolk-sac 100% resorbed after 130 days. Release of juveniles after 140 days. 127 Discussion In a review of reproduction in ariids Rimmer & Merrick (1983) found that the longest recorded incubation period in the literature was nine weeks with the average for six different species feliceps, being eight weeks. The incubation at between 18 and 21 weeks, period is thus for ~ considerably longer than those previously recorded for ariids. However, it is apparent periods were from the based on literature that the estimates rather above than incubation experimental evidence. Ward (1957) working on Galeichthys felis (synonymous with Arius felis), found that after artificial fertilisation the two-cell and 16-cell stages of development were reached after four and 11 hours respectively. In G. stages were reached after 19 and 29 hours, feliceps these demonstrating a considerably slower development rate in the cleavage phase. Since most ariids occur in tropical and sub-tropical waters, and since temperature is one of the strongest influences on the rate of development (Blaxter 1969; Kuftina & Novikov 1986), the long incubation period demonstrated by G. feliceps may be a function of its temperate environment. The advanced stage of development at hatching appears to be peculiar to, and universal amongst, ariids (Gudger 1916, Mansueti & Hardy 1967, AI-Nasiri & Shamsul Hoda 1977, Jones et al. 1978, Rimmer 1985b). However, Balon (l975b, 1977) suggested that in the evolution from substratum guarders to mouth- brooders, earlier hatching was probably induced in the latter in response to two factors, namely the small peri-vitelline space associated with eggs having a large yolk to cytoplasm ratio, and the low oxygen tension of the surrounding water in the buccal cavity. Evidence generated during the present study would indicate that the above may not be true for ~ feliceps . Firstly, the small peri-vitelline space did not appear to inhibit growth since as 128 the embryo increased in size, so the yolk decreased in volume. While the embryo did appear to be somewhat constricted and cramped close to the time of hatching, when virtually no evidence of a peri-vitelline space remained, the embryo was then at an advanced stage of development. Secondly, it may be argued that the buccal cavity is an oxygen-rich environment. Balon (1981b, p.61) argued that 'some mouth-brooding cichlids, apogonids and probably ariids (Gunter 1947), have their buccal cavities packed to such a degree by the eggs that churning is impossible. The brooding parent respires entirely via the gill apertures'. Under such circumstances the buccal cavity would experience very little flushing at all and would undoubtedly be oxygen-deficient. However, this statement appears to be based on observations of one species of cichlid only (Balon 1977), and certainly does not occur in mouth-brooding li.... feliceps, in which ventilation is brought about by the branchial pump mechanism Observations of (Fry 1957) mouth-brooding in the normal fashion. individuals in aquaria demonstrated that embryos situated in the front of the mouth were rocked to-and-fro by a considerable current of inhalant water during breathing movements. Theoretically, water entering the adult buccal cavity would accelerate upon passing over and between the obstructive embryos, much like air does when passing over the upper, curved surface of an aeroplane wing . This phenomenon, together with the increased ventilation rate demonstrated by mouth-brooders in this study, would suggest that the buccal cavity is probably a well-oxygenated environment. What was apparent from this study, however, was that the ventilation rate of mouth-brooders declined dramatically after hatching of the embryos, indicating that oxygen requirements of embryos were considerably higher than those of free-embryos. The high ventilation rate exhibited before hatching is probably energetically expensive for the mouth-brooder and early hatching should therefore be desirable. As the branchial system is operational, although not yet fully differentiated after 129 approximately 43 days, it is surprising that hatching does not occur until approximately 75 or 80 days. More recently Balon (In Press) has argued that in the evolution of mouth-brooding 'Hatching is at first accelerated with the prolongation of brooding time but the larger yolk ultimately facilitates the incorporation of more carotenoid pigments with the ability to provide an endogenous oxygen supply, and so hatching can again be delayed'. This argument is not convincing, however, and in the absence of experimental evidence amounts to mere speculation. The timing and stage of development at hatching may be influenced by a host of factors. These include the following: the necessity to get rid of accumulating metabolic wastes; the increasing osmoregulatory capabilities associated with organic differentiation; changes in metabolic oxygen requirements at the onset of a new developmental phase; decreased susceptibility to parasite infestation, bacterial or fungal infection; and the need to commence exogenous feeding. Balon (1985) is of the opinion that hatching is an event of little consequence in ontogeny unless it coincides with the onset of exogenous feeding. In ~ feliceps the timing of the onset of exogenous feeding is unknown, although mouth-brooding adults do not appear to pick up detritus or food until several days after hatching has occurred . Free-embryos hatched in incubators will commence feeding immediately if food is available. This would suggest that the digestive system is operational at hatching. ' The advanced fin differentiation demonstrated at hatching in feliceps is surprising since the free-embryo has no need of ~ them during the extended buccal incubation phase subsequent to hatching. The only manoeuvering ability required is for orientation within the buccal cavity so that the embryo is able to face the inhalant water current . 130 This requirement is probably universal amongst mouth-brooders in which free-embryos are incubated. Certainly in some cichlids (Balon 1977), hatching occurs when fin differentiation is slight. The advanced fin differentiation at hatching is therefore probably a function of delayed development of some other structures or organs which require the protection afforded by the vitelline membrane. The development of the vitelline blood plexus differs from that of mouth-brooding cichlids in that the primary vitelline plexus is symmetrical and stems from left and right lateral branches of the paired anterior and posterior cardinal veins. Early in the development the blood is returned to the heart by anterior left and right vitelline veins , which later anastomose to form a single, large anterior vitelline vein leading into the duct of Cuvier, and the heart. In Labeotropheus ~ the vitelline plexus stems from the preanal network of the inferior caudal vein and the sub-intestinal vein and returns to the heart via a single anterior vitelline vein . Later on in the development the hepatic vitelline vein also proliferates on the left-hand side of the yolk, resulting in an asymmetrical vitelline plexus (Balon 1977). The more extensive system found in ariids is probably indicative of the high oxygen and nutrient demand of the very large embryo. Chondrification of the branchial basket commenced after approximately 30 days and blood was seen to flow through the branchial arches for the first time. Gill lamellae appeared after approximately 43 days and coincided with an increased blood flow from the caudal vein into the posterior cardinal vein. This development resulted in a reduction in blood flow to the vitelline plexus and probably marked the onset of combined branchial and vitelline respiration . Strictly speaking, it meant only that oxygenated blood from the vitelline plexus was being diverted through the branchial arches for the first time. It is unlikely that the rudimentary gills assimilation would function at have performed much this time . of Considering an the oxygen small perivitelline space and the consequent restricted volume of 131 perivitelline fluid, it is indeed unlikely that true branchial respiration occurred before hatching. The ossification of the skeletal system was considerably delayed and was the last major developmental event to occur. The white substance that appeared in the intestine fairly early in development (after approximately 27 days), increased in size and volume with time and was passed through the vent into the perivitelline space after about 45 days . This substance was deeply stained by Alizarin red, indicating a high calcium phosphate content (Taylor & Van Dyke 1985). It may therefore have constituted an accumulation of excess calcium during the period of development before the onset of ossification. Assuming this hypothesis to be correct, one would expect this calcium to be resorbed at the commencement of ossification, rather than being excreted into the perivitelline space (Plate X). While chemical gas chromatography could composition of the be pellet, used to reveal acquisition sufficiently large sample may prove to be problematic. Plate x. Galeichthvs feliceps embryo during step demonstrating the nature of the brittle, whi te pellet excreted into the perivitelline space. 132 of the a Endogenous nutrition The yolk of embryonic te l eosts is completely enclosed within the syncytium (also known as the periblast) which is situated between, and interfaces with, the yolk on the inside and the vitelline blood capillaries on the outside . All substances derived from the yolk must therefore pass through the syncytium before it can enter the embryonic blood system and be utilised by the embryo (Bachop & Schwartz 1974). Mobilisation of stored energy in the yolk begins with digestion of yolk platelets through the action of hydrolytic enzymes found within the yolksac syncytium (Hamlett et al. 1987). The main source for tissue format i on in the yolk is protein (Blaxter 1969). In order for it to be utilised by the embryo it must first be converted into amino acids, and then resynthesised into proteins within the embryo. While most of the amino acids would be resynthesised into protein, a certain amount would be needed for energy and would be catabolised. The result of this catabolism is the production of ammonia, which is highly toxic . Smith (1957) found that the egg envelope in salmonids was largely impermeable to the end products of nitrogen metabolism and that these accumulated in the vitelline fluid until hatching . The problem of ammonia toxicity is overcome in many teleosts by breaking it down into a non-toxic form such as urea through enzymatic action in association with the ornithineurea cycle. Urea may then be stored until hatching (Rice & Stokes 1974). Read (1968 in Blaxter 1969) suggested that the retention of urea served an osmoregulatory function in the early embryos of oviparous and ovoviviparous elasmobranchs. Ariids, which have a high proportion of protein in their yolk and an extensive embryonic period, might be expected to have either an egg envelope permeable to ammonia or the ability to convert and store it as urea. Duration of the buccal incubation period It is probable that rates of differentiation and growth within the incubators occurring may have in the natural differed considerably environment . 133 from those In the absence of a control developmental series (which by definition is an impossibility since the assimilation of such a series would involve at least some manipulation of natural events), they were impossible to detect. However, the project was conducted in order to determine the sequence of ontogenetic events and to elucidate the approximate overall duration of the incubation phase . The former has been accepted as being hierarchical in nature (Maynard Smith 1983 in Greenwood 1989; Balon 1986; Greenwood 1989), and are unlikely to have varied from those occurring in the natural environment, while the latter was more difficul t to establish. Although the duration of the oral incubation phase up to hatching (75-80 days) was determined with a reasonable amount of certainty, the period of free- embryo incubation was less certain. Embryos which were hatched and reared in incubators and fed to satiation each day reached the mean length at release (54mm) at 35 to 45 days after hatching (n = 4 broods). However, they might have grown considerably faster than naturally incubated young, which would probably not have access to as much high protein food. On the other hand, the two aquarium-held mouthbrooders (64-day post-hatching incubation period) might have delayed releasing their young due to the absence of some natural stimulus such as current or a suitable environment in which the young could shelter. The true post-hatching incubation period is therefore likely to fall somewhere between the two extremes of 35 and 64 days. Taking the mid-point would give a total incubation period, from activation to release of juveniles, of approximately 125 to 130 days. The duration of the mouth-brooding phase in some cichlids has been found to be determined by an innate, internal timing mechanism (Mrowka 1981 in Mrowka 1985), in which free embryos are released after a pre-determined time interval. Aquariumheld mouth-brooders demonstrated that the same mechanism may apply in ~ feliceps, although no experimental work has been done to test the hypothesis . Since the young are not released 134 periodically during incubation as they are in many cichlids, the parent has no visual cue as to the stage of development of its young . However, as their yolk supply becomes depleted, the free embryos are likely to become more demanding in terms of their exogenous food supply. Since it is likely that the source of much of the mucus found in the stomachs of free embryos was the adult buccal cavity itself, the intensity of mucus consumption may also serve as a trigger marking the end of the incubati on period. Saltatory ontogeny or otherwise? The theory of saltatory ontogeny in fish was originally propounded by Vasnetsov 1953 and Kryzhanovsky et al. 1953 (in Balon 1979), and has subsequently enjoyed considerable mention in the literature (see Balon 1986 for review). However, Greenwood (1989) has recently questioned the validity of the theory and has argued convincingly in favour of an alternative interpretation of ontogeny, namely that of a stepped continuum. Ontogenetic development in a stepped continuum would progress at varying speeds in which the so-called saltations could be likened to periods of accelerated development rather than leaps across discontinuities, as the saltatory ontogeny definition implies. While isolated examples of proposed saltations in the early ontogeny of fish have been presented (e.g. Balon 1979, 1981c; Paine & Balon 1984), they refer mainly to behavioural saltations involving sudden habitudinal shifts of organisms in their environment. While there seems to be little doubt that ontogenetic development is stepped, evidence for saltations (other than behavioural ones), said to occur during thresholds between developmental steps, appears to be lacking. Balon (1981c) has conceded that the identification and verification of saltations biochemical, would require physiological, detailed morphological elaboration and of ethological aspects of ontogeny, information which has yet to be presented in the literature . The writer's opinion is that there is no need to assume, as Balon (1985) does, that saltations are an inherent component of stepped development. 135 In altricial species particularly, ontogeny is often marked by sudden transitions from one niche to another (Paine & Balon 1984), or by a change from 'jerky to fluent swimming' 1985). While such behavioural changes are visible, (Balon and may indeed be likened to saltations, they are not evident in the ontogenetic development of all (1985) studied the species. early ontogeny in Cunningham the & Balon cyprinodontiform Adinia xenica, in which all of the embryological development occurred within the egg envelope . Although they had difficulty identifying thresholds in the ontogenetic development of this species, the observed ontogeny was nevertheless held to be sal tatory. Much of the early ontogenetic development in ~ feliceps also occurred within the egg envelope, and while steps were identified changes, on the strength of marked morphological ontogenetic development appeared to be gradual. It should be emphasised that the present study was not conducted in enough detail to test the theory of saltatory ontogeny, although it was evident from this investigation that Balonian ontogenetic saltations are considerably more cryptic than their definition implies. In conclusion, while artificial fertilisation and incubation was met with limited success, overall ontogenetic development was satisfactorily recorded using several developmentally overlapping broods from wild-caught mouth-brooders. As a result of the slow and apparent gradual ontogenetic differentiation, developmental steps were difficult to identify . Hatching and ossification were considerably delayed and young attained a large proportional body size before being released from the parent buccal cavity. This strategy confers a substantial survival advantage on the young. The prolonged mouth-brooding period has probably been facilitated by the paternal mouth~ brooding habit. The ontogenetic characteristics exhibited by feliceps are representative of an evolutionarily advanced mouth-brooding strategy. 136 CHAPTER 5 - AGE AND GROWTH Introduction The growth factors rate of fish controlling the is one of dynamics the of primary biological exploited populations (Baranov 1918, Peterson 1903 & Russell 1931 in Beverton & Holt 1957; Ricker 1975; Gulland 1977; Cushing 1981; Everhart & Youngs 1981; Summerfelt 1987). While stock assessment may be performed without age and growth data, using surplus production models (Graham 1935, 1938, Schaeffer 1954, 1957 1981), their inclusion into dynamic pool models in Cushing (Beverton & Holt 1957), introduces an additional biological element and improves Age and growth studies are therefore an accuracy. important prerequisite for accurate stock assessment. Several iterative methods are available which mathematically express the growth of fishes (for review see Butterworth et al. 1989), with parameters which may be incorporated directly into stock assessment models. extensively in fisheries Pitcher & Hart 1982), The dynamic management pool (Beverton models, used Holt 1957; & utilise the following age and growth data: age at recruitment; age at sexual maturity; and maximum age reached. Parameters of the growth equation are also used in the determination of asymptotic weight, and natural and total mortality' coefficients. Growth data also form the basis for virtual population analysis (Gulland 1965 in Pope 1983), in which population age structure is used in the determination of annual recruitment and fishing mortality. The choice of depends on the nature fitted curve been the appropriate method and fitting procedure modeled of the error distribution about the (Hughes 1986) . Once the error distribution has using the absolute, 137 relative or transformed logarithmic model (Hughes op cit.), the most appropriate growth function may be selected. For an error model to be acceptable, the criteria of randomness and homoscedasticity of the residuals must be satisfied. These are statistical tests used to determine the goodness and validity of fit 1989) . The growth functions available (Punt & Hughes include the four- parameter Schnute model (Schnute 1981; Schnute 1985 in Hughes & Punt 1988), and several sub-models thereof 1988; Punt & Hughes 1989). The age composition and growth rate of ~ were Firstly, determined for two reasons. (Hughes & Punt feliceps and to ~ test ater the predictions of the r- and K-selection theory, namely that Kselected species Secondly, should exhibit slow growth and longevity. to provide the parameters required for the stock assessment models. Materials and Methods Preliminary investigation into the use of various hard parts for ageing indicated that the lapillus otoliths were the most suitable structures to use. Other structures investigated were pectoral and dorsal spines, vertebrae, operculae, and sagittal and astericus otoliths. Spine and vertebral sections yielded indistinct growth zones in which it was difficult to identify growth checks. Opercular growth checks were not visible with the naked eye and these structures were rejected. The lapillae were large, thick and semicircular (Hecht & Hecht 1981), and exhibited relatively distinct growth zones upon sectioning . Fish lengths and lapillus otolith dimensions were positively correlated for both species (Table XVI), indicating that otolith size could be used as a reliable indicator of fish age . In addition, the number of otolith growth rings increased with otolith size, suggesting a relationship between the number of rings and fish age. 138 Several techniques were initially employed in the attempt to improve the clarity and 'readability' of otolith growth zones. These included: a) Sectioning Otoli ths were embedded in resin rods and secti oned laterally through the nucleus using a double-bladed diamond-edged saw (Williams & Bedford 1974; Rauck 1976; Beamish 1979). b) Burning - Whole and sectioned otoliths were exposed to a low heat intensity alcohol flame until browned (Christensen 1964; Hecht & Smale 1986) . c) Staining Whole otoliths and otolith subjected to ninhydrin staining (Schneppenheim sections & were Freytag 1980) . d) Heating - Whole and sectioned otoliths were heated to 1S0 a C in glycerine (Freytag 1980). e) Immersion - Whole and sectioned otoliths were immersed and viewed in a variety of liquids including glycerine, methyl salicylate, water and xylene (Williams & Bedford 1974; Hecht & Smale 1986) . f) Polishing - Otolith sections were polished using 1600 grade carborundum paper and g) Daily growth grinding paste (Hecht & Smale 1986). SO~m rings Scanning electron microscopy was employed in order to determine the feasibility of using daily growth ring counts in the identification of annuli (Victor & Brothers 1982; Radtke & Targett 1984; Campana & Neilson 1985; Radtke 1987; Pulfrich & Griffiths 1988) . The final procedure adopted for viewing the otoliths was as follows: resin unburnt otoliths were embedded in a clear polyester using metal or plastic moulds. The flexibility of plastic moulds made from longitudinally halved 2Smm diameter plastic piping expedited the removal of hardened resin rods . The mould surface was coated with a releasing agent (petroleum jelly, wax-based effi0edding. polish Otoliths diamond edged saw, were or conventional sectioned using grease) a before double-bladed with sections varying in thickness from 0.3Smm to O.SOmm. Otolith sections were fixed onto glass slides using 'D.P.X. Mountant' and viewed under a binocular microscope 139 between lOX and 30X magnification. The sections were v iewed under transmitted light . While growth zones were visible in whole otoliths it has been shown that in long-lived species, the use of whole otol i ths frequently results in age under-estimation (Beamish 1979; Bennett et al . 1982; Beamish & McFarlane 1983, 1987) . Annulus validation Several methods were used in the attempt to identify annual growth zones . a) Edge increment analysis was conducted using otolith sections. The month during which the outermost growth zone was at its broadest was identified and the width recorded. width of the outermost growth zone, in otoliths The collected during successive months, was then measured and expressed as a percentage of the maximum observed growth zone width. A plot of the monthly increments revealed whether or not the growth zones were annual. b) Otoliths from laboratory hatched and reared one year-old feliceps juveniles confirmed the presence of a juvenile ring ~ within the nucleus and enabled the identification of the first annulus in otoliths from 'wild' samples. c) Growth of ~ ater juveniles taken in monthly rotenone samples (Smale & Buxton 1989) revealed the size of one yearold fish. The nature of the first annulus in otolith sections could therefore be ascertained . Reading of otolith sections Otolith sections were counted three times without reference to size or sex of the animal at intervals of approximately two weeks. Where all three counts differed the secti on was rejected from the study, while if two out of the three counts were the same a fourth count was made . If after the fourth reading three counts matched, the age was accepted as valid. 140 The numbers of otolith sections used in the studies were 598 for feliceps, of which 53 (9%) were rejected, and 326 for ~ ater, of which 22 (7%) were rejected. The sexes were aged ~ separately as length frequency histograms suggested that females of both species grew to a larger size, and were older, than males. Back-calculation was employed in order to fill sampling gaps for certain lengths age at age classes. for four, For ~ five feliceps back-calculated and six year-old fish were included in the growth curve, while for G. ater back-calculated lengths at three, four, five and six years of age were used. Since the relationship between fish length and otolith radius was non-linear, the formula of Ricker & Lagler (1942 in Bagenal & Tesch 1978), was used: Sn (S Sn/S) = where Sn = adjusted distance to the nth annulus, S = average otolith radius for a fish of the observed length, Sn = radius of annulus ' n' , S = total otolith radius . The parameter estimates obtained from the adjusted relative error model indicated that either the Schnute (Schnute 1981) or the von Bertalanffy (von Bertalanffy 1934 in Pauly 1981) model could be used to model the growth of G. feliceps males and females, and of ~ ater females. An F-test was used in each instance to determine the preferred model (Punt & Hughes 1989) . The Schnute model produced the most reasonable fit for ~ ater males, and the F-test showed that it was also the preferred model for ~ ater females (F=12. 713; Bertalanffy Special model was df=l, 202). selected for male df=l, 308) and femal e (F=1.73, df = l, 329) 141 ~ The von (F=O. 0369; feliceps. The von Bertalanffy Special parameters) growth models (3-parameters) are described and Schnute by the (4- following equations: VON BERTALANFFY SPECIAL l(t) - L. I I 1 - exp (-K (t - t o) ) where l(t) - the length at age t, L. - the theoretical maximum length, K - growth rate parameter, t o - the age at which the fish would commence growing from zero length. SCHNUTE ~(t) where l(t) 1 1 lz a b - ["b" --- -a( t - U b_ ~ b) 2 1 - e 1 -a(t 1 - e 2 ',' J lib - t l ) the length at age t, the value of l(t) at time t-t 1 , the value of l(t) at time t-t z , parameter, parameter. The observed length-at-age data were processed using 'PC-YIELD' (Hughes & Punt 1988). However, requirements for randomes~ the data failed to meet the of residuals and homoscedasticity for either the absolute error or the random error models. After additional analysis of the data a successful transformation was obtained (A . Punt Dept . Applied Mathematics, University of comm.), Cape Town, pers. which yielded homoscedastic residuals, using an adjusted relative error model. 142 The adjusted relative error model used was as follows: where lt = the length of a fish, at = its age, A lt = the model estimate of the length, b = a pseudo - parameter, (t = a normally distributed error. Variance estimates of the growth model parameters were determined using the jack-knife re-sampling method (Hughes & Punt 1988). Since the dynamic pool models generally deal with biomass rather than length data (Beverton & Holt 1957; Gulland 1977, 1978; Cushing 1981; Ricker 1981; Pitcher & Hart 1982), growth in weight was calculated as follows. W. was determined from the L. value by For ~ feliceps substitution in the length-weight relationship, and was then used in the von Bertalanffy equation to calculate growth in weight. For ~ ater mean length at age data were transformed into mean weight at age data using the length-weight relationship. Results Nature of the growth zones The otolith growth zones did not conform to the characteristic hyaline-opaque pattern found in many teleost species native to the South African coast (Botha 1971; Hecht & Baird 1977; Nepgen 1977; Payne 1977; Geldenhuys 1978; Wallace & Schleyer 1979; Coetzee & Baird 1981; Buxton & Clarke 1985; Bennett & Griffiths 14 3 1986; Buxton & Clarke 1986; Buxton 1987; Clarke 1988; Griffiths 1988; Pulfrich and Griffiths 1988; Buxton & Clarke 1989). They were composite, comprising two or more indistinct sub-zones of varying opacity, similar to those found by Griffiths & Hecht (1986) and Crozier (1989) for Lophius upsicephalus and h piscatorius respectively. The nuclei of both species appeared opaque when viewed under transmi tted light . A hyaline c heck was present wi thin the nucleus. This check may have corresponded to the completion of yolk absorption and the onset of exogenous feeding (Drnitrenko 1975), (Plates XIa The hyaline & b). checks between successive growth zones were generally more distinct than those occurring within the growth zones (Plate XIIa) . After sexual maturation, the growth zones became narrow and less easily distinguishable (Plate XIIb) . In otoliths from older fish counts were best made along the medial edge of the section where growth zones were narrow and growth checks most easily defined. AnnuluB validation Edge increment analysis using otolith sections revealed that in ~ feliceps the annual growth zone commenced in September and reached maximum width in August of the following year . The analyses were determined for otolith sections with three and four growth zones. In ~ ater growth zones commenced in August and were complete by July of the following year. Fish with between five and nine growth zones were used in the analyses (Figs . 48 & 49) . The analyses revealed that the growth zones represented annuli. 144 Plate XIa. Otolith section from a 1+-year old G. juvenile: n=nucleus, j=juvenile ring, x=annulus. feliceps Plate XIb. Otolith section from a 1+-year old G. ater juvenile de~onstraig the nature of the annulus: n=nucleus, . j=juvenile ring, x=annulus. 145 Plate Xlla. Otolith section from an 8-year old G. female demonstrating the nature of the annuli (-). feliceps Table Xllb. Otolith section from a IS-year old G. ater male demonstrating the deterioration in annulus structure towards the otolith edge: - =annuli . 146 '00 90 80 '0 60 50 40 '0 20 '0 "AN MONTHS o N _ S? Figure 48. Monthly edge increment analysis in G. feliceps using otolith sections from fish aged at three and four years. '00 90 80 '0 60 50 4 0 '0 20 '0 "AN MONTHS O N .. 61 Figure 49. Monthly edge increment analysis in G . ater using o~lith sections from fish aged at between five and nine years . 147 Known age feliceps hatched and reared in aquaria o v er a ~ two-year period were used to confirm early growth rates (Fig. 50). Similarly, growth rates of young -of-the-year ater ~ obta i ned from monthly rotenone samples were used to conf i rm the size of one year-old f i sh, as established from otoliths (Fig. 51) . This information enabled the identification of the first annulus and confirmed the presence of a juvenile ring . '6 0 150 - .' '40 ... 130 - '20 ~ v "0 '"\1 '00 ~ w / ~ ~ f\ ~ ,+' 90 .+ .+---- 80 70 /+-+..- 60 50 -V ,+-~ // 40 DE C FEB APR JUN Ae.G OCT DEC FEB APR JUN AloG OCT DEC MONTHS Figure 50. Growth of laboratory hatched and reared G. feliceps juveniles . As spawni ng occurred in September , the curve illustrates that lengths (FL) of approximately 75 and 145mm were attained after 12 and 24 months respectively . ~ feliceps males and females were aged to 16 and 18 years respecti vely. Difficul ty was experienced wi th the i nterpretation of growth zones near the edge in some of the larger otoliths . Since these sections were rejected from the study, both species may have been under- aged by between three and five years. 148 " 0 , -_______________________ __ _ _ __ __ _ _ _ _ _ _ _ _ _ _ __ _ _ _ _ __ _- , 90 eo '0 60 '0 4J AN ~ _r-, FEB ~MAR APR MAY JUN -,_r JUL -r_. A UG 5EP OCT NOV ~ DEC MONTHS Figure 51. Growth of Q..:.. ater juveniles caught in monthly rotenone samples. Since spawning occurred in August, the curve illustrates that a length (FL) of approximately 95mm was attained after one year. Q..:.. ater males and females were aged up to 15 years although some otoliths which may have been up to 18 years of age were rejected on the strength that three successive counts did not concur. Clarity of growth zones also tended to deter iorate near the o tolith edge in many of the older specimens, precluding accurate interpretation. There were no significant differences between the sexes in the relatio nships between fork length and otolith radius, otolith length or body mass (t-test, p<O. OS), for either species (Appendix IIa). The regression equations for the relationships between fork length and otolith length, otolith radius and body mass are presented in Table XVI and the parameters of the growth models in Table XVII . 149 Table XVI: Relationships between fo rk length and otolith length, otolith radius and body mass for G . feliceps and G. ater. REGRESSION: G. ATER G. FELICEPS FL FL vs OR: (r2 FL FL vs OL: FL vs Wt: 45.27xORA1 .18 ~ FL (r2 ~ ~ 51 .58xOR A 1.01 ~ 0.953) ~ 0.97) FL = 16.91xOL A 1.14 17.19xOLA1.19 (r> ~ 0.983) (r2 Wt ~ 9.55E-Q6x FL A 3.08 Wt ~ 1.38E-Q5xFL A 3 .04 (r> (r> ~ 0.9) ~ 0.g8) ~ 0.981) --_._-_._------------ Table XVII: Parameters and standard errors of the von Bertalanffy and Schnute growth models for G. feliceps and G. ater respectively. ------------------------ - - - - - - - - - -- - - -- - MALES (S.E.) FEMALES (S.E.) G. FELICEPS = L ~ t 0 ~ = K 389.895 (15.182) -1.406 (0.227) L ~ t 0.147 (0.017) K 0 407.847 (5.490) -1.834 (0.0694) 0.119 (0.00502) ~ ~ ~ G. ATER a ~ b ~ 11 ~ 12 ~ L~ t ~ 0 K 0.274 (0.00549) a 1.072 (0.00575) b 69.289 (1.532) 358.176 (0.801) 11 ~ 12 ~ (1.221) L~ 279.344 N/A ~ ~ 0.274 (0.00549) ~ ~ 0.0739 (0.0448) 2.265 (0.569) 46.3 (104.221) 292.62 (2.450) 348.407 (39.693) -0.141 (0.364) 0.074 (0.0448) ~ t0 = K ~ - - - ------------ The age-length keys for the two species are presented in Tables XVIII, XIX, XX & XXI. The mean observed lengths at age between feliceps were not significantly different. In 150 the ~ sexes in ~ ater the older age classes were significantly different in size (P<0.05), with females being the larger sex (Appendix lIb). However, the growth curves for the sexes were found to be significantly different in both species ater F=19.218; df=4, (~ 412), feliceps F=30.492; df=4, 641: and were therefore ~ plotted separately. The growth curves for males and females of both species are shown in Figures 52, 53, 54 & 55 . Mean observed fork lengths (± 1 SD), are plotted together with the modeled curves. Back-calculation was conducted for ages 0 - 9 only since the annuli after sexual maturity were not validated. Student's ttest showed no significant differences in the back-calculated lengths at age between the sexes of either species (P< 0.05) (Appendix Irc) . Student's t-tests of mean observed vs . back-calculated lengths ~ at age yielded significant differences for 7, 8 & 9 year-old feliceps, suggesting the presence of Rosa Lee's phenomenon (Bagenal 1978). The observed and back-calculated values are presented in Table XXII. Growth in weight for ~ feliceps was described using the von Bertalanffy growth model as follows. Males: wit) = 928.7411 - exp(-0.16(t + 2.35) Females: wit) = 1064.16 11 13.083 exp(-0.16(t + 1.85) 13. 083 For G. ater mean weight at age data were used to construct the curve. Growth in weight is graphically presented in Figures 56 & 57. 151 Table XVIII: Age-length key for G. feliceps males. Fork lengths in mm. are used throughout. SIZE ALL AGES AGE CLASS --- --- ------ ------------- ---------- ------ ----- --------------------------------- ------0 16 40-59 14 60-79 80 - 99 4 100-119 2 5 6 7 8 9 10 11 lZ 13 14 15 16 17 18 16 14 5 13 7 120-139 3 140-159 160-179 13 3 19 14 180-199 200-219 10 10 220-239 3 5 4 3 5 4 240-259 260 - 279 2 8 10 280-299 10 19 Z4 14 8 11 13 3 9 300-319 320-339 3 7 9 6 13 11 4 8 17 26 37 61 17 4 9 306 3 11 Z 18 13 340-359 Z 10 Z 6 2 3 14 8 3 360-379 ALL SIZES 34 21 36 20 12 14 21 13 19 30 32 16 < 00 350 300 ! ~ 250 ~ 200 '50 '00 50 0 0 , Figure 52. Calculated growth curve (using the von Bertalanffy Special model) and mean observed lengths at age (±l SD), for G. feliceps males . 152 Table XIX: Age-length key for G. feliceps females. ALL SIZE AGES AGE CLAS S --- -- ---- ------------ -------------- --- ------------ ------------- -- ----- ---------------0 40- 59 16 60-79 14 80 -99 4 2 4 5 6 7 • 9 10 11 12 13 14 15 16 17 I. 16 14 12 7 100-119 120 -139 140- 159 160 -179 3 5 12 9 2 21 11 11 13 180-199 200-219 220 -239 2 10 6 22 24 23 9 6 4 • 260 - 279 4 6 4 4 280-299 1 2 240-259 1 300 -3 19 320 - 339 340 -359 360-379 19 15 9 12 2 6 6 6 13 • 2 6 4 13 6 4 17 7 9 3 37 14 5 3 6 9 8 1 53 36 6 7 7 1 330 ALL SIZES 34 21 34 25 19 16 23 10 14 27 19 16 19 20 9 9 8 40 0 3>0 300 ~ 250 ~ ~ ~ 200 '50 '00 50 0 0 2 Figure 53. Calculated growth curve (using the von Bertalanffy Special model) and mean observed lengths at age (± 1 SD), for G. feliceps females. 153 Table XX: Age-length key for G. ater males. ALL SIZE AGE CLASS AGES ------------- ------- ----- ----- --- --- ------ ---------- --- ---- -- --- ----- --- -a 60-79 80-99 2 3 5 7 6 8 9 10 11 12 13 14 15 11 6 1 100-119 120-139 2 5 7 140- 159 4 6 4 160-179 180-199 200-219 220-239 240-259 260-279 280-299 300-319 ALL SIZES 4 7 4 6 4 1 6 7 5 5 4 12 18 7 15 15 15 14 10 16 18 19 18 7 11 8 7 12 12 16 16 10 69 38 6 4 15 8 5 8 9 5 2 1 5 23 23 13 10 5 3 6 212 300 280 260 + 2 40 ! \l"w + 2.0 200 '00 ~ ~ '60 ,.0 ,.0 '00 80 60 0 , 2 3 4 • 6 7 AGE Cye e. ... ~) 8 9 '0 " ,. " ,. " Figure 54. Calculated growth curve (using the Schnute model), and mean observed lengths at age (± 1 SD), for G. ater males. 154 Table XXI: Age-length key for G. ater females . ALL SIZE CLASS AGE AGES ------------------------------------- ------------------------------------0 7 60-79 80-99 11 100-119 120-139 2 3 4 5 6 7 8 9 10 11 12 13 14 15 7 11 6 4 6 3 140- 159 5 160-179 180-199 200-219 3 6 5 8 14 2 3 3 3 220-239 2 8 4 3 3 2 8 2 3 240-259 260-279 280-299 300-319 9 3 2 9 7 15 10 15 9 2 1 24 26 26 11 12 21 51 54 4 ALL SIZES 18 7 11 6 6 14 6 4 12 11 12 7 25 12 203 320 300 200 260 2<0 ~ 220 I ~ ~ 200 ~ '00 '00 ",0 '20 '00 00 60 . 0 0 3 • 5 6 AGE 7 o 9 10 11 12 13 14 15 (yeor-s) Figure 55. Calculated growth curve (using the Schnute model), and mean observed lengths at age (± 1 SD), for G. ater females. 155 Table XXII: Tests for significant differences between mean observed and back-calculated fork lengths at age in G. feliceps and G. ater . __........__.__ ._-_ ................ _........._.-_. __...._.._.._.._.._----_._---------_._.._----_ .._--- .._--_ ...... MEAN FL (BACK-GALC) SO T·TEST 27 27 80.3 9.74 -2. 7123 117.5 15.73 '5' 179.1 17.74 18.38 203.3 224.2 16.63 14.56 -0.0759 0.5785 -0.1507 -0.3327 0.0581 MEAN FL AGE .. _._. N (OBSERVED) SO 34 20 36 45 32 63.1 14.86 116.9 11.08 '55 178.1 200.4 3' 45 2. 33 224.7 9.42 8.87 16.73 18.14 250.2 278.2 17.53 27 27 27 27 27 27 295.5 310.5 'Z6 26 11.7 25 N SIG ..- "'F I') __._--_._-_.._---_._ ...._ -_._-_........ _-_ ...._..--------_..._ -_. __ ...._...__.._---------._--_.......- ....._. G. FEUCEPS , 0 2 3 4 5 6 7 8 • 0 , 2 3 4 5 6 7 • " 19.45 241.9 257.8 273.4 266.5 13.84 1.0242 2.3267 3.1447 3.7499 14.38 14.2 14.98 G. ATER 80.3 18 7 ,.,.. " 23 14.56 28 204.3 '6 20 30 22' 15.0 1 15.35 13.39 9.61 239.6 249.6 9.32 7.23 168.8 189.1 22 25 25 25 25 25 24 10.98 6.87 111.9 .._._----_. __...- '5 9.07 116.9 9.61 144.9 169.2 188.9 'Z3 13.19 13.58 13.36 206.• 218.7 2312 242.1 ,. 2' 83.6 -0.5647 -0.7182 -0.1026 -0.0736 0.0236 .Q.3380 0.3488 '0.26 8.96 9.44 '.4344 1.4227 .00 900 700 600 ~ ~ 500 'i' <e 400 !l! 300 200 100 0 0 2 4 6 o 9 AGE (yee. ... s) MA LE + 10 12 14 16 19 FEMALE Figure 56. Calculated weight at age f o r male and female G . feliceps. 156 ·50 . 00 350 300 n ~ ~ 250 'i' '" w 200 ~ "0 '00 50 0 0 • 2 e 6 AGE o " '0 (yean!) + MALE " FEMALE Figure 57. Cal c ulated weight at age for male and female G . ater . Discussion While the largest G . reasonably well with feliceps the aged maximum in this sizes study compare recorded in the literature it is apparent that for G. ater this was not the case . The largest ~ ater individuals sampled in the study were considerably smaller than maximum sizes recorded for the s pecies . The largest spec imens of both species obtained in samples off Port Alfred were a 435mm (TL) female and a 350mm (TL) female ~ feliceps ater. Castelnau (1861) reported that ~ G. feliceps and G. ater reacned total lengths of approximately 45cm and 55cm respectively in Cape waters. Taylor (1986), on the other hand, mentions maximum sizes of 55cm and 45cm for ~ feliceps and G. ater respectively, although the length of an illustrated ~ ater presented is given at 48cm cit.). The modal size for ~ ater in the Port Alfred population was considerably smaller than that of ± 370mm, (Taylor QP ~ feliceps (± 310mm vs . [TL) ). However, a single preserved I 157 ~ ater specimen found in the JLB Smith Institute collection (sampled off Cape Point in 1975), was found to be 40cm (TL), which is cons iderably larger than that of the largest G. ater specimen sampled in the Port Alfred area. Follow-up sampling in the Cape Point area will have to be done in order to establish whether the species does indeed grow to a larger size there, and if so, whether this is the result of faster growth or longevity. Ariid age and growth studies in the literature yielded results which varied considerably with respect to both growth rate and longevity (Table XXIII). Table XXIII: Comparative ariid age and growth data from the literature. SPECIES & AGEING METHOD MAXIMUM AGE REFERENCE Arius couma 5 &. proops &. proops 3.5 Meunier!!.&. (1985) Meunier et & (1985) 3 Lecomte 3.5 Dan (1980) Pectoral fin ray section: !! ~ (1986) Whole otolith & Whole operculum: Tachysurus tenuispinis Whole otolith : Bagre bagre 6 Costa & Juras (1981 /82) G. caeru lescens 6 Warburton (1978) A. thalassinus A. thalassinus Galeichthys ater G. fenceps 9-11 Dmitrenko (1975) Netuma barba 6 Singh & Rege (1968) 52 Cort6s (1984) Otolith section: 19 Bawazeer (1987) 15 This study 18 This study 36 Reis (1986b) Whole vertebrae & Length frequency: l.~ Length frequency: Acicpsis bonillai 158 Since all known ariids are mouth-brooders (Rimmer & Merrick 1983) with a low fecundity, a characteristic of K-selected fish (Adams 1980), they might generally be expected to exhibit other K-selected characteristics such as slow growth and longevity. The wide range of maximum observed ages exhibited in ariid age and growth (3.5 studies 53 is years) therefore unexpected. Noticeably, studies using whole otoliths, operculae and vertebrae resulted in consistently lower ages being obtained than those in which otolith sections were used. This may indicate that age was under-estimated in the former studies, rather than it being over-estimated in the latter. Certainly for Arius proops and A. couma (Meunier et al. 1985), maximum observed sizes of approximately 72 and 80 cm (SL) at ages 3.5 and 5 years respectively, suggest growth rates equivalent to those of many extremely fast growing pisciverous (game-fish) species (Van der Elst 1981). Considering the ariid diet (omnivory) and K-selected life-history characteristics, such rapid growth rates are highly unlikely for this family. In order to compare results obtained using whole versus sectioned otoliths, the method of Warburton (1978), an otolith check technique using whole otoliths, was tested in the present study. It was found that many growth checks were obscured in older whole otoliths due to a stacking effect near the growing edge, a phenomenon that has been recorded in otoliths from several species (Blacker 1974; Clarke 1988). These checks were elucidated only upon viewing otolith sections. Bennet et al. (1982) were confronted with this phenomenon while ageing the scorpaenid Sebastes diplora, in which whole and sectioned otoliths yielded markedly different counts. Using radiometric analysis, in which the magnitude of the zloPb/ 226 Ra disequilibrium was determined as a function of otolith weight, they were able to show that the counting of growth zones in sections provided the best estimate of fish age, and that counts using whole otoliths lead to under-estimation in older fish. 159 While the cessation of feeding during mouth-brooding by adult males may have affected growth, this was not manifest in the pattern of growth ring deposition, as there was no apparent difference in appearance between mature male and female otolith sections. The continued growth of otoliths during periods of starvation has been documented by several authors (see review in Campana & Neilson 1985), and it has been found that although mean increment width decreases during starvation, the extent with which starvation influences increment width appears to be related to metabolic rate. Increments were found to be affected to a greater extent in animals with higher metabolic rates (Campana 1983a in Campana & Neilson 1985). This may explain the similarity in annulus appearance and size between males and females in Galeichthys, since sexual maturation occurs at an advanced age when metabolic rate is likely to have slowed considerably. Otolith increment deposition is under the control of an endogenous circadian rhythm, which operates regardless of starvation (Campana & Neilson QQ cit.). The use of fluctuations in hepatosomatic indices (HSI) is a good indicator of short-term changes in growth rate (Adams & McLean 1985). Other indices such as relative condition factors and visceral fat-somatic indices (FSI) are less sensitive to short term changes and respond better to long term changes such as those occurring annually. This had direct bearing on male ~ feliceps which showed dramatic fluctuations in both HSI and FSI during mouth-brooding. The growth rates of males is likely to decrease during the mouth-brooding period whereas females, which do not show a marked fluctuation in HSI or FSI, might be expected to maintain a normal growth rate. Strictly speaking, the presence of Rosa Lee's phenomenon in the data for ~ feliceps should have precluded the use of back- calculated lengths in the modeling process. However, since data was missing for four year-classes (3, 4, 5 & 6 year-old fish), the use of the back-calculated lengths at age was considered preferable to modeling their growth with missing year-classes. 160 The missing year-classes may have been due to their inaccessibility to fishermen . For example, the zone just behind the breaker line where they may well occur, is largely out of reach to both the shore angler and the boat fisherman . The observation that Rosa Lee's phenomenon was present in the data might, however, have been an artefact. Since the species is slow growing, resulting in very little difference in size between the successive age classes, it is unlikely that selective natural or fishing mortality would have occurred, i.e. that there would have been greater survival of the smaller sized fish of a given age group . calculated data were derived from a (Males, n~1; Females, In addition, the back- sub-sample of otoliths rather than from the entire sample n~16) (n~645). For back-calculation to be used with confidence, two criteria have to be satisfied. A good correlation should exist between otolith dimension and fish length, and measurement of annulus radius should occur between the nucleus centre and some consistent point along the annulus edge. During sectioning of otoliths can lead even slight variations from the centre of the nucleus to fairly substantial errors in subsequent measurements of annulus radii. Errors also arise as a result of the three-dimensional nature of otolith sections, the third dimension being the thickness of the section . When an otolith section is cut too thick, a different set of measurements would arise depending on which of the two lateral surfaces was used. Bartlett et al. (1984) emphasised the importance of obtaining the best possible relationship indicator (otolith) size. between fish length and They found that the traditionally used regression techniques which generally yield significant relationships can be improved using analysis of covariance. Using ANOVA, in which the relationship between fish length and otolith size is determined separately for each age class, the accuracy of back-calculated data can also be improved . 161 The growth curves indicated that both species were slow growing and late maturing, a predicted trend in K-selected fish (Adams 1980). Ages at 50% sexual maturity for females were 9 and 8 years, ' and for feliceps males and ~ ater, ~ 8 and 7 years respectively, a result which suggested that these species might be particularly vulnerable to recruitment over-fishing. In feliceps a difference in average body mass between the ~ sexes was apparent at an early age, with females being the heavier sex. The disparity in weight increased with age and males tended maturity. In toward ~ asymptotic growth soon after sexual ater it was also evident that the onset of sexual maturity in males lead to a leveling off of growth at an earlier age than it did in females. Age determination using otoliths invariably involves a level of subjective reliabili ty of decision making age growth and (Brothers studies 1987), depends and the largely on factors such as the experience of the investigator, familiarity with the life history style of the species under study and an understanding of the relationship between the biology of the animal and the physiological processes of organic and inorganic material (1974), deposition within the otolith. interpretation should precede ageing of fish to telecommunication, As stated by Sych counting, He equated in which the recorded information stored within the black box/structure being used (the age of the fish) needed to be decoded before it could be received/understood by the addressee/reader. In addition, he stated that the information encoded on the structure assumed different forms depending on the nature of the fish population and the conditions to which it had been exposed. He emphasised the importance of setting up, and adhering to, a decision rule which relates to the investigator's concept of annual rings as distinguished from account spawning the non-annual rings, season and and which any other takes into important environmental events. He provided a technique for testing the effectiveness of a formulated decision rule, but conceded that 162 the only way of verifying a decision rule was to compare the results with a sample of otoliths from known age fish. However, Beamish & McFarlane (1983) have shown that other verification techniques such as mark and recapture, edge increment analysis and cohort analysis may be equally effective in the absence of known-age reference samples. In the present study, annulus structure was confidently established up to the age at sexual maturity for both species. Beyond this age the structure changed and became increasingly difficult to interpret. It is therefore acknowledged that for large fish ages may have been either over or under estimated. While mark and recapture studies were attempted using oxytetracycline injection of sexually mature animals (Wild & Foreman 1980; Leaman & Nagtegaal 1987), none of these were recaptured. The data is therefore used with constraint until such time as all annuli have been validated. In conclusion, the rejection of a small proportion of the larger otolith sections from the study due to poor readability may have resulted in an incomplete representation of age classes for both species. The results demonstrated that ~ feliceps and ~ ater are slow growing and long-lived. Females tend to live longer and exhibit faster growth in terms of weight (both species) and length (G . ater) after the age at sexual maturity . For management purposes the vi tal information required was knowledge of the age at recruitment into the fishery and age at sexual maturity, both of which can be determined using the established growth curves. 163 successfully CHAPTER 6 - POPULATION DYNAMICS Introduction The Port Alfred fishery dates back to the early 1900's when several sparid species such as Roman Chrysoblephus laticeps, dageraad cristiceps, red stumpnose ~ gibbiceps and poenskop ~ Cymatoceps nasutus formed the bulk of the catch. By the mid1960's, however, they had become rare and effort was largely redirected to the kob Argyrosomus hololepidotus (Sciaenidae), and two deeper water sparids, the silver Argyrozona argyrozona and the panga pterogymnus laniarius (Hecht & Ti1ney 1989). While the contribution of barbel (G. feliceps & ~ ater) to the catches prior to this study is unknown, it would appear that they have always formed a by-catch in the fishery (Port Alfred Commercial Fishermen's Association, pers. comm.), and their exploitation is likely to have increased steadily as a function of the continued expansion of the fishery. During a although fishery, the preliminary assessment in 1984 feliceps and ~ ~ it was found that ater occurred as a by-catch in the they collectively constituted approximately 10% of total annual catch in terms of landed mass. This represented a harvest of between 35 and 40 tons per annum. The ratio of G. feliceps to ~ ater in catches was approximately 3:1 (Hecht & Tilney op. cit.). While the fishery operated at various depths between 30 and 100 meters, the two ariids were generally not caught deeper than 60 meters. As a consequence of their familial life-history characteristics, particularly their low fecundity and extended parental care behaviour, it was anticipated that they would be vulnerable to over- exploitation in the short term if they were to become actively targeted for in the fishery. The objective of this stock assessment was to determine the impact of exploitation on the two K-selected barbel species in a fishery dominated by highly fecund r-selected sparids and sciaenids. 164 With the ever increasing demand for protein, it is probable that both feliceps and ~ ater will become targeted for in ~ the foreseeable future . They have high catchability in that they are voracious feeders, have a large gape and will take a baited hook under virtually any circumstances. They are commonly caught in conjunction with kob which forage over similar substrata and occur in the same depth-range . The weight of barbel in an average daily catch is far exceeded by that of kob, a larger species. The average kob:barbel weight ratio is 4.8:1. However, the catch in terms of numbers of fish landed is far more equable (1.3:1), and it is clear that on this basis the ariids, which have a fecundity several orders of magnitude lower than that of kob (pers. obs.), require special protection. It was anticipated that an investigation into their population dynamics at this relatively early stage in their commercial exploitation would enable the formulation of suitable management strategies before extreme, exploitationrelated, population declines occurred. An understanding of the species' response to present levels of fishing effort will expedite future management policy decisions in which emphasis should be placed on acceptability to the fishery and ease of implementation. An array of stock assessment models are available to the fisheries biologist and the type of model used depends on the nature of estimates the or information required sustainable yield (e.g. estimates). total biomass The models are particularly useful for predicting the responses of exploited populations to changes in fishing effort (Gulland 1977; Cushing 1981). Using the information from the models the fisheries biologist is able to propose, in terms of fishing effort, how the stocks should be managed. Subsequently, it is the task of fisheries administrators to weigh up the constraints of the resource against the socia-economic demands of the user community and to marry the two in the formulation of a management strategy for the fishery as a whole. 165 There are essentially two groups of models used in fish stock assessment. These are the surplus production, surplus yield or Schaefer type models (Schaefer 1954; Gulland 1977; Cushing 1981), and the dynamic pool, analytic or Beverton-Holt type models (Beverton & Holt 1957, Ricker 1975; Gulland QQ cit.; Cushing QQ cit.). The two categories of models differ considerably with respect to their basic data requirements. The surplus production models treat the population as a single entity, are based on a mathematical function describing total biomass regeneration (population growth), and require catch and effort data only. The dynamic pool models, on the other hand, treat the population as the sum of its individuals and deal explicitly with biological parameters affecting population abundance including growth, mortality and reproduction. The surplus production models are generally used when biological data are lacking, and when catch and effort data from several successive years are available. The dynamic pool models may be applied to data assimilated from one, or many years of fishing, and are considered to be more realistic due to their incorporation of detailed biological information (Ricker 1975, Gulland 1977,1978,1985, Cushing 1981, Everhart & Youngs 1981, Pitcher & Hart 1982). A computer software programme using a Beverton & Holt-like dynamic pool model has been developed by the Department of Applied Mathematics at the University of Cape Town, specifically for use by the SANCOR 1 Marine Linefish community (Hughes 1986, Butterworth et al . 1989, Punt 1989, Punt & Hughes 1989). The assumptions of the model are as follows. a) Recruitment is constant from one year to the next, b) the stock is in an equilibrium state (biomass and agestructure are the same from one year to the next), c) the population is closed (there is no immigration into or emigration from the stock), (Butterworth et al. 1989). 1 South African Committee for Oceanographic Research. 166 The model utilises the total mortality rate of successive age classes in association with ages at sexual maturity and recruitment i nto the fishery to provide estimates of yieldper-recruit (Y/R), yield-per-recrui t in terms of biomass. model is the spawner A sub- model of the biomass-per-recrui t model (SB/R), which determines the proportion of spawners that remain in the fished population relative to the unexploited cond i tion, as a function of age at first capture tcr and mortality are fishing mortality F (Butterworth et al. 1989) . Independent estimates of total and natural required in order to determine the extent of fishing mortality . For the sake of convenience, (which apply over a instantaneous mortality rates short period of time during which the numbers in the population do not vary significantly), are used. This is done in order to avoid the algebraic complexity involved in the determination of mortality when the numbers of fish dying numbers from one particular cause dying from any other cause are affected by the (Gulland 1985). Another advantage of instantaneous rates of mortality is that they may be added to, or subtracted from, one another. Since by tradition all barbel caught in the Port Alfred commercial fishery are the property of the crew members who caught them (Plate XIII), and since boat owners generally only weigh the proportion of the catch that is sold to whole-salers, only a very small percentage of the total annual barbel catch is ever recorded. The obligatory linefish catch return statistics furnished to the Sea Fisheries Research Institute in Cape Town could therefore not be used to determine annual landings in the Port Alfred area. Although sampling of total boat catches was undertaken in an attempt to estimate the percentage contribution by barbel, fishing regularly occurs at depths beyond that at which barbel occur, preventing accurate extrapolation of the total annual barbel catch on this basis . Estimates of abundance could therefore not be determined, nor 167 could virtual population analyses be utilised for estimation of trends in recruitment or mortality. Plate XIII. Fisherman from the Port Alfred commercial fishery with the day's barbel catch. G. feliceps are held in his left hand and G. ater in his right hand. A single estimate of total instantaneous mortality was determined for each population using combined length-frequency samples for the period 1985-1987 . However, since fishing effort increased by approximately 17% during the study period from 2808 to 3375 boat-days per annum (Hecht & Tilney 1989), the estimates of instantaneous total mortality obtained for the fishery were probably slightly 168 underestimated . Strictly speaking, total mortality estimates should be determined using data from one year only. However, in the light of the slow growth exhibited by the two species, model and estimation, the considerable the use of variance the assumptions of the attached to mortality combined data from three successive years was considered acceptable. Yield per recruit can, apart from several constant factors for example Woo, to and K, be mathematically described by three parameters, one of which describes the fish, and two which are characteristic of the fishery. a) The parameter relating to the fish is the ratio of the growth coefficient (K), and natural mortality (M). The rate at which a fish approaches maximum size is dependent on both K and M, which therefore measure the rate at which the fish lives. In the calculation of yield curves, the M/K ratio tends to operate as a single parameter. A small M/K ratio indicates that the fish has a good chance of completing much of its potential growth before dying of natural causes, and the stock will contain many relatively large fish. A large M/K ratio suggests that fish are dying before much of their growth is completed and the stock should be fished at a relatively higher fishing effort and small recruitment size in order to gain the best yield from a given level of recruitment (Gulland 1985). b) The parameters relating to the fishery are the amount of fishing, expressed as the ratio of fishing to natural mortality F/M, which is also the exploitation ratio E, where E = F/(F+M) c) The relative size at first capture c, where c = lJloo = 1-exp{-K(tc-t o)} 169 The Beverton-Holt equation for yield-per-recruit is therefore essentially a function of fishing mortality and age at first capture. MaterialB & MethodB Since the growth curves for the sexes significantly different for both G. were found to be feliceps and G. ater the yield-per-recruit analyses were determined separately for the sexes (Beverton et al. 1984). This was considered particularly important for ater in which the size distribution of males ~ in the fishery was noticeably smaller than that of females . In addition, the male mouth-brooding habit of the barbel species is likely to cause differences in catchabili ty, and hence mortality, between the sexes at certain times of the year. As males do not feed whilst mouth-brooding they will theoretically not be subject to fishing mortality during this period, which lasts between four and five months. This provided additional motivation for modeling the sexes Department of Applied Mathematics, separately (A. Punt, University of Cape Town, pers. conun.). In the previous chapter it was shown that the Schnute growth model provided the curves of best fit for observed G. ater growth. However, since the von Bertalanffy growth parameters, K and L., are extensively used in dynamic pool-type modeling, growth of ~ ater was recalculated using the von Bertalanffy growth model. Mortality determination Totel mortality (Z): Several techniques are available for the determination of total mortality. Three were used in this study. a) Catch-curves: Age-length keys were constructed using two different methods. In the first, 170 a sub-sample of fish were aged, and the age-length key normalised with respect to the length frequency distribution of the total sample using matrix multiplication (Hughes 1986, Butterworth et al. 1989). The n o rmalised age-length key was then used in the construction of a catch-curve, in which the natural log of fish number was plotted against age as follows: logaN = a + bt where the value of 'b', with the sign changed is representative of total mortality, Z (Ricker 1975). An example of the method used in the normalisation of the age-length keys is provided in Appendix IlIa. In the second instance the von Bertalanffy growth equation was used to determine the mean relative ages of fish in successive size classes (Pauly 1983). These were then plotted against the adjusted number of fish in each size class as follows: log.(N/dt) = a + bt where dt = the time taken to grow from the lower (t l ) to the upper limit (t 2) of a given size class, t = the relative age corresponding to the mid-range of the length class in question. An example of the above methodology is provided for §..:.. feliceps males in Appendix IIIb. b) The Robson & Chapman (1961) method which computes survival S, in which the number of animals alive at successive ages are determined using the formula: s = T/n+T-1 = a statistic, n = sample size . where T 171 Mortality is determined from survival as follows: Z = -Log, S. c) A method proposed by Beverton & Holt (1956, in Butterworth et al . 1989) which determines total mortality from the relationship between the age at recruitment to the fishery and the mean age of fish in the catch as follows : Z = [1 + l/a - a f )] where a = mean age of all fully recruited fish sampled, af = age at full recruitment into the fishery . Natural mortality (M): Natural mortality incorporates death through predation, disease and other natural causes, and is readily determined using several different equations, four of which are used here. a) Pauly (1980). Pauly found that natural mortality was correlated with longevity, and hence K, and also with size and water temperature . He expressed these interrelationships in the form of a multiple regression as follows: Log 10 M = -0.0066 -(0.279 Log lO L.) + (0 . 6543 Log 10 K) + ( 0 . 4634 Log 10 T) where L. = theoretical maximum length, K = Brody growth coefficient, T = mean annual bottom temperature. 172 b) Gunderson & Dygert (1988). This is an empirical formula derived from regression of gonad index against values of M obtained from the literature to give the relationship: M where = (1.68 x WGSI) + 0 . 03 WGSI = a wet gonadosomatic index (gonad wt./body wt . ) c) Rikhter & Efanov (1977). This is also an empirically derived formula based on the relationship between natural mortality and age at sexual maturity as follows: M = 1.521/(0.7tm ) where ~ - 0.155 age at 50% maturity. = d) Roff (1984). An empirical formula relating natural mortality to growth rate and age at sexual maturity as follows: M = 3 x K[exp(-KT))/1 - exp(-KT) where K T = Brody growth coefficient, age at sexual maturity . Fishing mortality (F): Instantaneous rates of fishing mortality were derived by subtracting natural mortality estimates from the total mortality estimates, i.e. from the equation: F = Z - M. Selectivity Although the selectivity curves for barbel were logistic, knife-edge selectivity was assumed for the fishery to enable the use of the computer software programme PC-VONBERT (Punt 173 1989) for the calculation of percentage survival, yield-perrecruit and spawning biomass-per-recruit. The implications of the above are discussed later. Yield-Per-Recruit Analysis The Beverton-Holt (1957), like yield-per-recruit models used by the PC-VONBERT programme, analyse the trade-off between the increase in mass of individual fish and the decrease in the size of a cohort with time. The following assumptions apply: a) Recruitment is constant from one year to the next, b) the stock biomass and age structure is in equilibrium, c) there is no immigration of individuals into or emigration out of the stock (Hughes & Punt 1988). The following sub-models were used: a) Yield-per-recruit. -Mt Y/R = -k(t c 1 Fw~e (M + F) where -2k(t c - to) 3e 3e (M + F + k) + = w. theoretical maximum weight, tc to c - to) e (M + F + 2k) Y/R = -3k(t c - to) (M + F + 3k) yield-per-recruit (in terms of mass), age at first capture (50% recruitment), = theoretical age at which length = 0, K = Brody growth coefficient, F = instantaneous rate of fishing mortality, M = instantaneous rate of natural mortality. 174 b) The rate of change of numbers in a cohort with ( survival) , dN(t)/dt = (-M - S(t)F) x N(t) where N(t) = the number of t-year old fish in the population, M = the instantaneous rate of natural mortality, F = the instantaneous rate of fishing mortality, S(t) = the selectivity of the fishing gear on fish of age t . c) Spawner biomass-per-recruit. (i) for . SB/R :I: tm~c -Mt _ F(t _ m W e feliceps males), (~ t ) m c '" [1 3e- --- {M+Fl (ii) for t m<t c k m (t - t ol (M+F+k) -3k(t m - t 0 + - - - -(M + F' + 2k) feliceps females, (~ e ~ (M 11 + F + 3k) ater males & females) : SB/R :: W ... ~ -M t -M t m (e - e kto -(M + kIt -(M + kit e m ,; I _ 3e . ~ - e J (M + k) M - 2kt O 3e + where SB/R . [e -(M + 2k ltm - ( M + 2k)t - e (M+2k) J - 3kt O e -(M + 3klt . [e - = spawning biomass-per-recruit, tm = age at 50% sexual maturity, Ym = R = yield in terms of mass, recruitment 175 m - ( M + 3kltj - e (M + 3kl 1 Y m +. PR time The programme also calculates the following management variables: a) FMSY ' the level of fishing effort at which maximum sustainable yield i s attained. b) FO. I the , level of fishing effort at which the marginal yield-per-recruit drops to 10% of its value for the unexploited stock. Increasing F beyond FO. 1 provides a very small return in terms of yield-per-recruit in relation to the increased costs associated with a higher F . c) F so • ss , the level of fishing mortality at which 50% of the population spawner biomass remains. Results Mortality estimation The percentage length-frequency distributions and catch-curves for ~ feliceps and ater males and females are presented in ~ Figures 58, 59, 60 & 61 . The regression lines used to determine total mortality are plotted on the catch-curves. The lengthfrequency curves for ~ feliceps indicated that males and females were equally represented in all of the size classes. In contrast, the ~ ater length-frequency distribution was bi- modal and indicated that females survived to a much larger mean size than males. This bi-modal size distribution was probably a function of the slower growth rate exhibited by males. Males reached sexual maturity at approximately the same size, although a year later in life than females. The disparity in growth rate might have been a response to the high energetic requirements associated with mouth-brooding in males. 176 (a) ,. " " ,. "'2 "'0 9 B 7 •5 • 3 2 , a 2 '0 2'0 250 270 SIZE (b) CLASS (mm) IZZI F .L . N=837 ·,-----------------------------------------------------------, 5 • AGE (Years) Figure 58. (a) Percentage size frequency distribution and (b) catch-curve and regression line for G. feliceps males. 177 (0 ) '. ,. " " " "'0 9 B , 6 5 4 3 • ° 2'0 s ize CLASS (mm) F L. 12::Zl N_S,a ( b) AGE (T'ear"S) Figure 59. (a) Percentage size frequency distribution and (b) catch-curve and regression line for G. feliceps females. 178 (0) '6 " ,. " " " '0 9 e 7 6 5 • 3 2 0 200 2.5 2 30 275 260 275 290 30S 32 0 S IZE CL ASS Cmm) FORK LENGTH fZZl N::56"l (bl • 2 o 4- -r 6 e -.r~ '0 AGE (Ye", ... s) " '6 Figure 60. (al Percentage size frequency distribution and (b) catch-curve and regression line for G . ater males. 179 (a) 03 '2 " '0 9 B , .. 6 5 , , 2 0 200 2'0 2" 24 5 260 290 275 305 320 S. ZE CLASS (!T'ITl) FOI=lK LENGTH N=586 ezJ (b) 5 , -- -_ _ _ _ _ _ _ _ _ _ _ __ __ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __ __ _ _ __ _ _ _ _ _- , , , 2 o , 9 13 AG E ('(ears) " 19 Figure 61. (al Percentage size frequency distribution and (bl catch-curve and regression line for G. ater females . 180 The i nstantaneous t otal and natural mortality rates are presented in Tables XXIV & XXV, and demonstrate t he co nsiderable va ri a ti on i n the e stimates obtai ned. In Table XX IV , t he Ri c ker ' a nd Ric ker 2 me thods of t o tal mortality e s timation are those using normalised age-length keys and total length- f reque ncy distributions respect i vely . Table XXIVI Instantaneous total mortality rates (Z), using four different methods . Vari ance estimates appear i n parentheses . CHAPMAN·ROBSON (1961) ~ ~MALE S BEVEATON & HOlT AlCKER ' (1975) 0 .369 (0.024) 0.183 (0.106) §.;. FEUCEPS FEMALES 0.266 (O.O Hi) Q;. 0.445 (0.041) 0.229 (0.067) 0.234 (O.O86) 0.380 (0.17 4) ~MALES Q;, ATEA 0.613 (0.047) FEMALES R1C1<ER" (1956) (1975) 0.249 (0. 042) 0.223(0.039) 0.390 (0.075) 0.284 (0.070) 0.285 (0.064) 0.318 (0.056) 0.334 (O.O32) 0.257 (0,043) Table XXV I Estimates of instantaneous natural mortality rates (M), using four different methods. - - - - -------------------- - - - GUNDERSON & DYGERT (1988) PAULY AIKHTEA & EFANOV AOFF (1980) (1977) (1 984) 0_200 0.137 0.248 FEMALES 0.200 .§:..£§! MALES 0. 170 0.1 70 .§.:. ATER FEMALES 0. 101 0 .21 1 0.279 0.279 0.133 0.321 0. 136 0.188 0.130 0.239 .§:~MALES 2:. FEUCEPS There were no noticeable trends in total or natural mortality estimation using the different methods, i . e. one method did not result o ther . in consistently lower or higher estimates than any It was decided that the use of mean values representative of all the methods used in mortality estimation should be avoided, as thes e would simply have compounded the variances of the individual 181 methods wi t hout increasing precision. Instead, a result given by o ne method, for Z and M respecti vely, was chosen for use in the yield-per-recrui t models. The models we re therefore inte rpreted with due cognisance of the h i gh degree of variance associated with the mortality estimation. For Z, the method of Ricker 2 (Table XXIV) was chosen for the following reasons: a) The Ricker l method, which is based on normalised age-length keys, produced catch-curves in which the oldest age classes were inaccurately represented. This was a result of poor sampling procedure in which differences between the sizefrequency of the sub-sample of aged fish and the total sizefrequency sample from the fishery, gave rise to considerable bias in the older age classes during the normalisation procedure (Appendix IlIa). b) The Robson-Chapman method was based on the same data set as (a) above, and was therefore rejected. c) The Beverton-Holt method did not consider either the agefrequency or length-frequency distributions and was therefore considered to be inferior to the Ricker 2 method. d) The Ricker 2 method utilised the descending limb of the length-frequency distribution curve of the total sample, and was therefore the most representative method for the determination of z. For natural mortality, the estimates derived using the Pauly (1980) method were used. Although the variance in the value of M obtained using this method ranges between 0.3 and 3 times the estimate (Butterworth et al. 1989), it was chosen because of its general use in the literature. In order to compensate for the variance associated with M, stock simulation was performed 182 for natural mortality values 25% above and below the formula estimate. The chosen values of Z, M and F, together with the age and growth parameters used in the yield-per-recruit analyses, are presented in Table XXVI. Table XXVI: Parameters used in the yield-per-recruit analyses (to ; age at 50% recruitment, t m ; age at 50% sexual maturity). K l. 2:. FEUCEPS §.:. MALES FEUCEPS FEMALES .2;. ATEA MALES Q.;. ATEA FEMALES 38a.e 0.1 4 7 407.5 285.9 321.3 0.120 0.193 0.132 __._._.__._.._._._--_._.. ..... z M F 0.253 0 .217 0.123 0.124 9.46 9.82 0.390 0.137 11.01 8.9 1 0.101 -1.289 9.'18 8.78 0.3 18 0.334 -2.554 11.97 7.4 1 0.257 0 .1 33 ·1.406 -1.813 0.211 ..._---- - - - - - -- Sex ratios Because males were effectively excluded from the fishery during the mouth-brooding period, sex ratios were determined independently for the period encompassing spawning and mouthbrooding. They were also determined separately for the sexually mature sector of the populations (Table XXVII). The sex ratios for sexually mature animals revealed that there were significant feliceps, This differences throughout the year but only during the spawning season for i ndicated that the mature ~ feliceps for ~ ~ ater. male population might be considerably smaller than that of the mature female population, or that a proportion of the mature male population occupy a habitat in which they are not exploited by the fishery. Sex ratios for entire populations revealed that while there was, as expected, a significant difference between the numbers of male and female ~ ater caught during the spawning and 183 mouth-brooding period, t his was not the case for G. feliceps. A high juvenile component in the catches masked the absence of mouth-brooding males during this time. The percentage representation of mature animals in the catches is presented in Table XXVIII for the two species. Table XXVII: Sex ratios for G. feliceps and G. ater. Adjusted Chi-square tests were used to determine significant differences between the sexes at the 95% level. Mature pop . Chi·sq. 819. dif. Whole pop. Chi ·sq. (Male:Female) (df~ 1) (Yes/No) (Male:Female) (df _._._----_. ~ ~ 1) Si9. dif. (Yes/No) FELICEPS Spawning season 1:1.65 18.888 Yes 1: 1. 11 1.262 No Off season 12.360 Yes 1:1.03 0.074 No Spawning season 1:2.23 62.77 Yes 1:2.10 28.09 Yes Off season 0.705 No 1.11: 1 1.035 No 1:1.51 ~ATER 1.09:1 ----_. Table XXVIII: The percentage representation of mature G. feliceps and G. ater males and females in catches. Females Males -------G. FELICEPS Spawning season 56% Off season 60% 82% 88% G. ATER Spawning season Off season 94% 99% 96% 97% -------- Yield-per-recruit analyses The percentage recruit survival curves indicate that for ~ feliceps, between 68% and 71% of the recruits from any 184 particular year-class will hav e succumbed through n atural causes before they are recruited into the fishery (Fig. Similarly for ~ ater, 62). between 79% and 85% of the fish die before they are exposed to exploitation (Fig. 63). The curves indicate that both species might be fished more efficiently if they were recruited into the fishery at an earlier age . It should be noted, however, that in the extrapolation of survival curves back to age zero, the model uses total mortality values which were determined using only the larger size classes. Mortality amongst the smallest size classes is in reality far higher, and the percentage survival values determined by the model are probably over-estimates. The plots of yield-per-recruit (Y/R) vs. age at first capture (tc ) also indicated that t c should be reduced. They demonstrated that maximum yields for feliceps would be attained at ages ~ between 5 and 7 years for males, and 6 and 9 years for females (Fig . 64). ~ ater males and females should be recruited at ages between 2 and 4 years, and 3 and 5 years respectively for maximum sustainable yields to be realised (Fig. 65). The spawner biomass-per-recruit (SB/R) vs. (F) curves indicate that 0.5K, the optimum level ~ f i shing mortality feliceps are being fished beyond (Schaefer 1954; Butterworth et al. 1989) at which 50% of the initial spawner biomass remain (Fig. 66). Males are fished at between 0.25K and 0 . 33K, and females between O. 22K and O. 24K. Galeichthys ater males are fished between 0 . 43K and 0 . 67K, and females between 0.5K and 0 . 56K, indicative of a more healthy spawner biomass (Fig. 67) . The yield-per-recruit vs. resul ts of fishing effort curves reflect the the spawner biomass-per-recrui t present levels of fishing effort for ~ curves, in that feliceps are higher than the recommended FO.1 value in both the male and female populations (Fig. 68). Females are exploited more heavily than males. In ~ ater,however, present fishing effort occurs below the FO.1 value for both sexes (Fig. 69) . 185 (0) 1C C + . - - - - - - - - - - - - - - - -- -------t- 80 --' « :::> a: 60 ::J en !:: ::J a: .0 () w a: ;I' 20 12 B 16 AGE (Years) - Unfished 100+-- (b) x Fished -----------------+ 80 --' « > > a: 60 ::J en .... ::J a: .0 () w a: ;I' 20 O+-~._ -,_~ o 8 12 16 ~+ AGE (Years) -Unfished x Fished Figure 62. Percentage recruit survival vs. age fished and unfished (a) G. feliceps male and populations. 186 curves for (b) female leo (0) 80 ..J « :: > a:: ::;) 60 en >::;) a:: 40 () w a:: ... 20 12 8 16 AGE (Years) - Unfished (b) 10~ x Fished -~ 80 ..J < > > II: ::;) 60 en !:: ::;) II: () w 40 II: ... 20 o+-~.r o 12 8 16 AGE (Years) - Unfished x F ished Figure 63. Percentage recruit survival curves vs. age for the fished and unfished populations. (al G. 187 ater male and (bl female 12°+--- - - - - - - - - - - -- - - - - - - - - - t (Q) >- 90 OJ a:: u w a:: a:: 60 w <l. 0 ..J W >- 30 O+-~ o 3 6 9 ~ 12 -~+ 15 18 AGE AT FIRST CAPTURE Ie -M,O,14 (b) + M,O,10 xM'O,17 160+-------------------- - - - - - - -- - - -- - - - T >- 120 OJ a:: u w a:: a:: 80 w <l. 0 ..J W >- 40 0 1 ~0-5.12 AGE AT FIRST CAPTURE te -M'O,10 +M'O,08 XM'O,13 Figure 64. Yield per recruit vs. age at first capture (te) curves for the (a) G, feliceps male and (b) female populations, plotted using three different values for natural mortality (M). 188 (a) 50'+ - - - - - - - - - - -- - --------j- .0 !:: ::::J II: U w 30 II: II: W 0.. Cl ...J 20 !!! >- 10 01-r~36'9=:2;I58L AGE AT FIRST CAPTURE te -M,O,21 +M,O,16 90+-- - -- - - - -- - - - - - - - - t - (b) t- :; 60 II: U W II: II: W 0.. Cl ...J !!! 30 >- o+-~ a 3,8 7.6 11" 15,7 19,0 22,8 AGE AT FIRST CAPTURE te -M,O,13 +M'O,10 Figure 65. Yield per recruit vs. age at first capture (te) curves for (a) G. ater males and (b) females, plotted using three different values for natural mortality (M). 189 (a) 1560 (J) ~ 1170 ~ o iii <!) z z ;: 780 < ll. (J) o+-~_.L o 0,10 0,30 FISHING MORTALITY - F 0,5 0 +M~O,10 ( b) 3000 (J) (J) < ~ o 2000 iii <!) z z ;: < g; 1000 o+-~ o -M~O,10 0,10 - ~- 0,30 FISHING MORTALITY - F 0,20 ~- 0,40 +M~O,08 Figure 66. Spawner biomass per recruit vs. fishing mortality (F) curves for (a) G. feliceps males and (b) females, plotted for three values of natural mortality (M). F p = present fishing effort. 190 (0) 480+---------------------------------------+ (f) 360 (f) ~ < Q CD "z 240 z 3 < 0- (f) ___ ________ . _____ ______ ___ ____ O,5K 120 o+- o - .-~ 0,10 -~_. 0,20 0,30 FISHING MORTALITY - F -M=0,21 .M=0,16 xM=0,26 (b) 1440 (f) ~ 1080 ~ o CD "zz 720 3 < O,5K 0- (f) 360 O+- o -+L~ 0,20 0,10 0,30 FISHING MORTALITY - F -M=0,13 + M=O,IO xM=0,17 Figure 67. Spawner biomass per recruit vs. fishing mortality (F) curves for (a) G. ater males and (b) females, plotted for three values of natural mortality (M). Fp = present fishing effort. 191 120 ( a) 100 .... ::> 80 a: (J w a: a: 60 w Q. 0 ...J w 40 >20 0 0 0,10 0,20 0,30 0,50 O ,~ O 0,60 FISHING MORTALITY - F -t-,bO,14 • M,O,IO x M'0,17 1 6 0 - t - - - - - - - - -- -- - -- -- - - - j - (b) t:: 120 ::> a: w a: a: 80 FOI (J I ' w Q. o ...J ~ >- 40 0,10 0,20 0,30 0,40 0,50 FISHING MORTALITY - F + M'0,08 Figure 68. Yield per recruit vs, fishing mortality (F) curves for (a) G. feliceps males and (b) females, plotted for three different values of natural mortality (M). Fp = present fishing effort, 192 50 (a) FOI 40 I' >::> a: (J w 30 a: a: w Fp Il. 20 C ...J W ,. 10 0,10 0,20 0,30 FISHING MORTALITY - F + f'h 0,16 -M~0,21 x M~0,26 (b) 80 !:: ~ 60 (J w a: a: w Il. Fo" 40 c ...J W ,. 20 ~o ~o ~o FISHING MORTALITY - F +M~O,' Figure 69. Yield per recruit vs . fishing mortality (F) curves for (a) G. ater males and (b) females, Fp = present fishing effort . 193 Discussion The e ffect of gear on a fished population, i.e. the pattern of fishing mortality, is termed selectivity . Three types of selectivity are recognised, namely knife-edge, logistic and normal (Punt & Hughes 1989). The simplest method, and the one used in the available computer software programmes, is knifeedge selectivity which assumes that fish become available to the fishery at a particular age and that no fish younger than this age are caught. Assuming knife-edge selectivity and constant recruitment from one year to the next the catch-curve will show an exponential decline, the slope of which is determined by the total mortality rate (Butterworth et al. 1989, Punt & Hughes 1989). The recruitment, or selectivity curves, for the two barbel species in the linefishery were logistic. Starting with fish approximately 200mm in length, catches increased smoothly with increasing fish size and reached a maximum in the larger sizes classes. However, the calculation of total mortality using catch-curves was undertaken using only the size classes larger than that at which peak recruitment occurred. Several age classes were thus eliminated from the mortality estimation and in so doing, knife-edge selectivi ty was artificially imposed on the data. In order to improve precision by taking the full range of age data into account using a more realistic model, additional parameters need to be estimated. Butterworth et al. (op. cit.), however, state that the use of such models, in which increased numbers of parameters require simultaneous estimation from the same data set, may not necessarily lead to increased precision . Punt (Dept . Applied Mathematics, University of Cape Town, pers. comm . ) emphasised the importance of determining whether there is a substantial difference between assuming knife-edge or logistic selectivity. For the purposes of this preliminary stock assessment, knife-edge selectivity was accepted in the light of available its simplicity and yield-per-recrui t ease models. of application However, the to the final estimates of Z, and hence F, are likely to have poor precision. 194 Catch-curves may be considerably biased if incorrectly constructed (Pauly 1983), or if the numbers of fish caught at any age included in the subsequent regression are low (Butterworth et al. 1989), and they should be used with cauti on. Skewed catch curves lead to poor estimates of total and, hence, fishing mortality. Ricker (1975) cautions against the use of age-length keys unless the fish used fo r the age determination come from the same stock, during the same season and using the same fishing gear as those used to take the length-frequency samples. In he recommends that more sh~rt, effort be invested in ageing large samples of fish rather than in generating massive length-frequency data-banks, unless the latter are to be put to some other additional purpose. In this study catch-curves derived from the normalisation of age-length keys using length-frequency samples did result in biases in numbers at age for the oldest age-classes, and had to be rejected. The frequency of aged fish in the age-length key should exactly reflect the total size-frequency distribution of the sampled population if accurate catch-curves are to be constructed. Pauly (1983) warned that when utilising lengthfrequencies to construct catch-curves (Ricker 2 method in Table I), the incorporation of fish sizes close to that of the population asymptotic size should be avoided since the ages obtained for these fish may be grossly over-estimated. He emphasised the importance of visual examination of the plotted catch curve in order to identify the portion of the descending limb that may be used to calculate Z effectively. Several workers have developed models which correlate natural mortality to life history parameters such as age at sexual maturity, costs of reproduction, growth rate and maximum age (see Vetter 1988 for review). Since these models all produce a single estimate of M for the stock, they are essentially no more than broad estimates (Vetter op. cit.), and should be used with caution when applied to fishery models. In general, the results of this study have shown that the higher the estimates of M, the lower the estimates of 195 y/~x and the age at first capture. More accurate methods for determining M utilise catch analyses and predation methods (Vetter op. cit . ), both of which are based on cohort analysis and which require accurate catch data from several successive years . The duration of the present study precluded the use of cohort analysis. The accuracy of the yield-per-recruit curves will vary according to the degree of confidence with which the component parameters were established. The evaluation of stock condition is therefore largely reliant on representative estimates of growth rate and mortality. While the growth model parameters are fairly accurate in that they are able to realistically describe observed growth patterns, the widely ranging mortality estimates generated by the methods used in this study clearly are not. Two obvious deductions can be drawn from the above. Either the methods are poor and fail to adequately reflect real mortality, or the input data inadequate. \~er In all probability, both deductions apply. The results of the yieldper-recruit analyses should therefore be considered with the limitations of the mortality estimates in mind and be used with caution. Nevertheless, they are acceptable for use in preliminary stock assessments which seek to detect trends in population structure and to provide guide-lines for management alternatives. The results suggest that the feliceps spawner biomass has ~ been depleted beyond the recommended (O.5K) level (Fig. 67), an indication of growth-overfishing (Pitcher & Hart 1982). Both sexes are fished close to the critical level of O.2K, below which recruitment is believed to be detrimentally affected in r-selected fish stocks (Clark et al. 1985). This is an indication that G. feliceps is fairly heavily exploited. Figure 69 indicates that they are also being exploited above Fa.lt which is used as a management guideline when the yield-perrecruit curve is asymptotic. This is a fairly conservative strategy since it is always lower than the FMSY value. Hughes (1986) mentions, however, that 196 since the yield-per-recrui t model assumes recruitment to be independent of spawner biomass (which it is not), the FO.1 management strategy should be adopted since it has a smaller risk factor attached to it. This is particularly important when modeling strongly K-selected species in which recruitment is likely to be strongly dependent upon spawner biomass. The ater population appears to be in a healthy condition in ~ terms of spawner biomass-per-recruit and yield-per-recruit, although the females are slightly more heavily exploited than the males (Figs. 68 & 70). In the estimation of yield-per-recruit, the model determines the trade-off individual between fish available to the against the fishery rate the as of decline a increase in result of in numbers mass of of fish natural mortality (Butterworth et al. 1989). Since the growth rate of fish is generally fastest during the juvenile phase and since the younger cohorts contain greater numbers of individuals, the yield-per-recruit model will tend to predict higher sustainable yields at relatively young recruitment ages (te)' When natural mortality is high it is especially advantageous to harvest fish at an early age, and the model demonstrated this in Figure 66 for ater males and females. The model predicts that maximum ~ sustainable yields (MSY) will be realised at low te values, and that the te should decrease as M increases. Since M is higher for males than it is for females, a lower te is recommended for males (± 2 years for males and ± 4 years for females). As fishing intensity and thus fishing mortality (F) increases, the model will compensate by increasing (te) in order to prevent recrui tment overfishing. For and the model estimate of feliceps ~ te is F was relatively high therefore also set at a proportionately higher value (± 7 years for females and ± 6 years for males), (Fig. 65). The yield-per-recruit models were originally designed to assess highly r-selected stocks, in which egg and larval mortality has 197 been shown to be density-dependent (Cushing 1988). This means that even for a fairly small spawner biomass, recruitment is likely to be more-or-less constant from one year to the next. In strongly K-selected species, however, recruitment is more intimately linked with spawner biomass and fluctuations in the latter will probably significantly affect the former. A weakness of the yield-per-recruit model for assessing Kselected populations is, therefore, its failure to consider fecundity as a significant parameter . The literature depletion of considerable has the and provided abundant spawner rapid stock decline proof that invariably in catches excessive leads of to a K-selected species. Examples include trawl fisheries for mouth-brooding cichlids in the African Great Lakes (Fryer 1984, Witte & Goudswaard 1985, Ribbink 1987), several elasmobranch fisheries (Holden 1977; Compagno In Press), and the fisheries for marine mammals (Allen & Kirkwood 1988) . Al though ariids support active artisinal and shallow water commercial fisheries throughout the tropical and subtropical seas, information in the literature relating catch and effort trends are sparse. There is some evidence to suggest that fisheries for barbel stocks using gill nets and trawlers have resulted in unacceptably high mortality rates pauly & Thia-Eng 1988). Silas et al. (Cortes 1984, (1980) report that purse seine vessels off the coast of Goa are able to identify and target on aggregations of mouth-brooding Tachysurus maculatus (Ariidae). In one month, vessels landed an estimated 528 tons of adults and 38 tons of embryos, a potentially disastrous fishing strategy. Dmitrenko (1970) reported similar activities in the Arabian Sea where spawning aggregations of the ariid Arius thalassinus are actively sought out and trawled for. It is clearly evident that protection of the spawner biomass should be a priority in fisheries directed at species with low fecundity. In the management of the 198 ~ felicep s population the age at first capture (te ) should, therefore, be set at above 8 years, corresponding to the approximate age at 50% sexual maturity. Although the sex ratios for mature feliceps indicate that ~ significantly more females are caught in the fishery, this is not the case when juvenile fish are also considered (Table XXVII). However, as approximately 71% of the catch is comprised of mature fish there is a danger that females are being exploited too heavily. It may, therefore, become necessary to impose a closed season for ~ feliceps during the spawning season. This would effectively guard against excessive catches of females. While 97% of the ~ ater catch is made up of mature fish, the off season sex ratio does not differ significantly from unity and indicates that the sexes are more equally exploited . As the two species do not appear to suffer from barotrauma on being caught, probably a result of their phisostomous swim bladders, they lend themselves to the implementation of size limi ts as a management strategy. When fishes having phisoclistous swim bladders are hauled rapidly to the surface the expansion of air in the swimbladder commonly results in the stomach being everted and blown out through the mouth, and the intestine through the anus. Severe eye embolisms may also occur (Buxton 1987). Practical experience has shown that barbel may be returned to the sea after capture with a high degree of success provided they have not been hooked in or through the stomach or gills. Since barbel may be reproductively active for a period of 7 years or more after reaching sexual maturity, protection of the spawner stock would yield considerable returns in terms of future recruitment . The spawner stock could be protected using either minimum or maximum size l i mits. For ~ feliceps the size at 50% sexual maturity data (Figs. 14 & 15), indicate that a minimum size 199 limit of approximately 320-330mm (FL) would provide a degree of spawner protection to both sexes. However, the length- frequency histograms (Figs. 59a & 60a) indicate that the above minimum size limit would effectively exclude approximately 42% of the presently harvested biomass from the fishery. On the other hand, the imposition of this size limit would serve the dual purpose of reducing fishing effort and protecting the spawner biomass. The imposition of a maximum size limit at approximately 300310mm (FL) would also have a major effect on the fishery. Although it would offer complete protection to the spawner stock it would exclude approximately 86% of the presently exploited biomass from the fishery. A considerable reduction in the present to would then be required in order to realise previous yields . However, evidence suggests that the younger age-classes are more abundant in shallower areas, which explains their poor representation in present catches. Fishing boats tend to avoid areas shallower than approximately 30 meters due to the abundance of small "nuisance" these depths such as Boopsoidea inornata, species at Spondyliosoma emarginatum (Sparidae) and Pomadasys olivaceum (Haemulidae). The imposition of minimum size limits as a management strategy for .Ii.:.. feliceps would therefore be more practical for the fishery, and certainly more acceptable to fishermen. It should be noted that most barbel presently caught in the fishery are clubbed "senseless" by fishermen before being boated. This is to reduce the danger of injury from the poisonous pectoral and dorsal spines. The effectiveness of size limits as a management alternative for barbel would therefore depend on whether or not fishermen could be persuaded to refrain from the above practice . Barbel have large, bony heads, a stout pectoral girdle and heavy pectoral and dorsal spines, which together comprise a considerable proportion of their body mass. Since fish wholesalers will purchase only the 'head-off gutted weight', the 200 marketable mass of ~ feliceps is a mere 40% of the total body mass. This results in an average marketable mass of approximately 1909 and 204g for males and females respectively. The present target species in the commercial fishery are purchased whole (gutted and gilled only), and have the following mean body weights: kob (Argyrosomus hololepidotus, Sciaenidae) - l4llg, silver (Argyrozona argyrozona, Sparidae) 594g, panga (pterogymnus laniarius, Sparidae) 447g. It is evident that in comparison with the above species, barbel are poor candidates for commercial exploitation. To summarise, the models indicate that while the G. ater population shows no ill effects from exploitation, ~ feliceps is presently being over-exploited . While the yield-per-recruit models indicate that higher sustainable yields would be realised if the ages at first capture were reduced it was argued that the imposition of minimum size limits, set above the size at sexual maturity, would be a more suitable management strategy. As the spawner biomass for ~ feliceps was shown to be unacceptably low, protection of the spawner stock was considered to be a priority. It was shown that females were more heavily exploited than males, and it was suggested that a closed season during the spawning and mouth-brooding period would be an effective way of preventing excessive catches of the former. The study demonstrated a sensitivity for ~ feliceps to fairly low levels of exploitation. 201 CHAPTER 7 - GENERAL DISCUSSION During the course of the study several phenomena were encountered which could not be adequately explained using the data at hand . Examples include the reproductive function of the hyaline eggs, the purpose of the seasonal accumulation of tissue on the pectoral spines of h ater females, and the functional significance of the sexually dimorphic cleithra in both species. Other potentially rewarding avenues for future investigation are the nature of courtship and spawning behaviour, the extent of sound communication during courtship, the possible utilisation of sneaking as an alternative reproductive tactic in males, the cause of the strongly bimodal length-frequency distribution for the sexes in h ater, and the degree of interaction between the species at the interphase between reefs and sand, with the premise that one of the two environments is preferred by both species. It would also be interesting to determine whether or not the buccal cavity i s an oxygen-poor environment during mouth-brooding, and the value of yolk carotenoids as an endogenous oxygen supply for embryos . Finally, an investigation into the natural mortality rates for the younger age-classes would provide a means of assessing the degree to which the extended parental care behaviour confers a survival advantage on the young. A study of the biology of Galeichthys feliceps and G. ini tiated with chapters in penetrative the information hand would have investigations contained in the enabled more to be ater preceding challenging conducted. However, and the establishment of fisheries management policies for the South African linefish stocks, which relatively little is known, biological information that are diverse and of which requires the type of base-line this provide. 202 study has attempted to In the following pages aspects of the reproduction in ~ feliceps and ~ ater, particularly their mouth-brooding behaviour, will be discussed. The role of the two species in the Port Al f red c ommerc ial linefishery will also be debated in the light of their K-selected life-history traits. Reproduction The attainment of an optimal life-history strategy is dependent upon the allocation of energy resources amongst maintenance, growth and reproduction (Gadgil & Bossert 1970), in order to produce a phenotype which realises the highest rate of increase in population size under the given conditions of the environment they inhabit (Reznick 1985) . Environmental factors (food and space availability; predators; disease) will be the major determinant of natural mortality in K-selected species, whereas in r-selected density dependent. The species mortality criterion for is predominantly success in natural selection is the number of surviving offspring produced by a parent (Crow & Kimura 1970 in Adams 1980) and the best reproductive strategy is a compromise between two conflicting demands: production of the largest possible number of offspring (r-selection), and production of offspring with the highest possible fitness (K-selection) (Pianka 1970, 1972). A larger allocation of resources to reproduction at anyone age leads to better reproductive performance and is interpreted as a profit function. However, there will be a resultant decrease in subsequent growth, survival and reproductive contribution, which is considered a cost function. Natural selection tends to adjust the reproductive effort as an animal ages so that the overall 1966; fitness Gadgil & of the life-history is maximised Bossert 1970) . The cost of (Williams reproduction is defined as the trade off between the amount of energy invested in reproduction (fecundity and parental care) and the consequent reduction in growth rate, longevity or capacity for subsequent reproduction (Reznick 1985 ) . 203 Both Galeichthys species mature at an advanced age and at a proportionately large size when their growth rate and scope for growth (Brett 1979) are considerably reduced and the amount of energy required for maintenance of growth is low. There is a definite switch at the age of sexual maturity from a growth phase to a reproductive phase, particularly amongst males. The decline in growth Galeichthys males rate after sexual maturity exhibited by and their shorter average life-span is suggestive of a higher reproductive investment relative to that of females. In addition, the marked size-frequency dimorphism exhibited by ater, males being the smaller sex, indicates ~ that the cost of reproduction for males in this species may be higher than it is in ~ feliceps. A comparison of their relative reproductive investment could be used to demonstrate whether mouth-brooding is more expensive for ~ ater than it is for G. feliceps males. If it is not, the disparity in growth rate and longevity between the sexes in ~ ater is likely to be genetically based. An interesting phenomenon in Galeichthys reproduction is the way in which resources are allocated to growth and reproduction in the sexes. Both species have a reproductive life-span of approximately nine years. Optimal life-history theory dictates that under these circumstances resources should be allocated equitably between reproductive and non-reproductive activities (Gadgil & Bossert 1970). While this appears to be the strategy employed by females, males tend to invest larger amounts of energy into reproduction at the expense of future growth and survival. Since fecundity does not increase as a function of size in Galeichthys females, the only reproductive advantage to be gained from being large is in mate or territory acquisition. It may be that in Galeichthys it is the females which hold territories or compete for mates and the males which se lect mates . This role-reversal has been demonstrated in several fish species (Turner 1986; Svensson 1988). The larger average female size would support such an argument . Males lose approximately 28% of their total body weight during the 4 1 / 204 2- month mouth-brooding pe riod and contribute approximate l y more energy to reproduction than females (see Table 1/ 3 X). Foll owing mouth-brooding a recovery period of approximately three months is required before they regain their pre-spawning condition . As 7 1 / 2 months of eac h year are devoted e ither directly or indirectly to reproduction, males would appear to have little excess energy defending territories, available for establishing and or for other energetically demanding courtship-related activities . Species in which one of the sexes assumes the role of parental care invariably utilise courtship rituals to assess the f itness and preparedness of their prospective mates, a phenomenon often referred to as sexual selection (Emlen & Oring 1977; Otte 1979; Baylis 1981; Barnard 1983; Turner 1986; Chan 1987), and it is therefore likely that some form of courtship is employed by Galeichthys. Since it is the males that care for the young and since each male is able to care for one brood only, kinship theory would have it that in order to make the energy expenditure worthwhile, they should be absolutely sure of their parentage over the incubated brood (see Werren et al . 1980 for review, also Baylis 1981) . In contradiction to this argument there was evidence which suggested that sneaking behaviour occurred as an alternative reproductive strategy in a proportion of the male population of both species (the absence of fat reserves in males with ripe testes). Sneaking in fish has been documented for several nest-building spe c i es (Werren et al. 1980; Gross 1982, 1984; Turner 1986; Chan 1987), although the construction of nests, the holding of territories or the division of the male population into dominant and sub-dominant individuals may not be pre-requisites for this tactic to be employed. It could conceivably be effective in any situation in which at l e ast some courtsh i p between pa i rs was involved . 205 Gross (1984) argued that sneaking might be an evolutionarily stable strategy in one of two ways: either as a subordinate tactic wh ic h enabled individuals to make the best of a bad situation (e.g . because they were unable to secure a territory), or as an equivalent tactic in which case their success should decrease as sneaker frequency increased. The latter situation would theoretically enable an equilibrium to exist between sneakers and conventional parents and is probably the more likely of the two scenarios. It is also possible that sneaking frequency may be controlled by a genotype- environmental intera c tion (Gross op cit). The alleged sneakers in Galeichthys were not noticeably smaller than parental males . This would imply that they were not simply first-time spawners which had been unable to accumulate fat reserves in time for the spawning season. If fishing has created an imbalance in the population sex ratios and females are in short supply, sneaking may be the only way in which a proportion of the male population may hope to achieve fertilisation. Mouth-brooding The progressive demonstrated increase within the in reproductive mouth-brooding specialisation guild is largely manifest in the apportionment of increasingly dense yolk into fewer, larger ova (Balon 1981c, 1989; Blumer 1982). This phenomenon appears to have arisen in response to the survival advantage gained by increasing the size of offspring at the culmination of parental care (Oppenheimer 1970; Zale 1987; Balon In Press). Since egg size is p c.s itively correlated with development time (Ware 1975) mouth-brooding may have evolved in order to provide a safe environment for the embryo during the extended development period. The most advanced form of mouth-brooding appears to be that exhibited by the cichlid Cyphotilapia frontosa, in which free-embryos are fed within the buccal cavity, harboured until yolk absorption is complete, and released to fend for themselves for the first time as juveniles (Balon QQ cit.). This is also the category into which the ari i ds fall. The yolk density of ripe 206 ~ feliceps ova is lower than that occurring in frontosa which also has a far higher ~ concentration of lipids, although feliceps yolk has a higher ~ protein concentration (Table XXIX). These figures indicate that while C. frontosa yolk has a far higher energy content due to its high lipid fraction, it has a smaller average brood size (17 vs. 49 per clutch) and a far shorter incubation period than feliceps ~ (54 vs. 140 days). The average egg size and the size of the young at release, expressed as a percentage of parent size, 13.5% in is 2.8% and 10.7% in feliceps. ~ frontosa and 3 . 7% and ~ The increase in size between hatching (onset of exogenous feeding), and release is 48% in and 63% in frontosa ~ feliceps. The two species therefore have quite ~ different strategies of energy investment. In the cichlid, emphasis has been placed on the production of small clutches of eggs with high energy yolk resulting in rapid development and a short clutches, incubation period. lower-energy incubation period. Since G. feliceps high-protein ~ has yolks and larger egg a longer frontosa yolk has a relatively low protein content much of the protein needed for growth appears to be derived through intra-buccal feeding and coeval sibling cannibalism (sensu Hecht & Appelbaum 1988). Hatching occurs early in this species, after five to six days (Balon 2Q cit.). Table XXIX. Proximal yolk composition of the cichlid Cyphotilapia frontosa and the ariid G. feliceps. MOISTURE LIPIDS PROTEINS REFERENCE --------------- ----- - - - --- ----- --C. frontasa 44% 35% 20% Balon (In Press) G. feliceos 55% 7% 29% Marais & Venter (1987) 207 Mouth-brooding has been classified as a low-cost strategy in fish (Mrowka & Schierwater 1988), which capitalises on energy investments made during the preceding non-reproductive, feeding phase. The mouth-brooding adult generally ceases feeding activity, moves into a sheltered environment and becomes highly inactive. Energy expenditure is restricted and geared mainly towards churning of the eggs and the maintenance of an increased ventilation rate until hatching. Mrowka & Schierwater (op cit.) found that non-brooding, feeding individuals had a significantly higher metabolic rate than brooding or starved individuals. In the present study the significant difference in ventilation rates exhibited by non-brooding, feeding males as opposed to males carrying free-embryos is probably also indicative of a sharp decline in adult metabolic rate during mouth-brooding (see Figure 47). The high ventilation rate in adults carrying embryos, on the other hand, is probably necessary to compensate for the inefficient method of embryonic ventilation (oxygen diffusion across the egg 25~m-thick envelope) (Fry 1957). After hatching and the onset of branchial respiration in free-embryos, oxygen assimilation is more efficient and a lower rate of water turn-over is required . It is conceivable that the energy saved in the evolution from substrate guarding (in which considerable energy is expended by the adult in ventilating and protecting broods), to mouthbrooding is redirected into larger eggs and yolk of a higher density. Evolution of mouth-brooding. The evolution of parental care in fishes has been discussed by several authors (Werren et al. 1980; Baylis 1981; Gross & Sargent 1985; Sargent & Gross 1986) and in most species it is the male which performs this task (Oppenheimer 1970; Blumer 1982). The opinions expressed in the literature are unanimously in favour of substratum brooding having been the common precursor to mouth-brooding in all species (Iles & Holden 1969; Breder 1933, Myers 1939 in Oppenheimer 1970; Lowe 1959 in 208 Thys van den Audenaerde 1970; Thys van den Audenaerde 1970; Balon 1975a, 1977, 1984; Baylis 1981; Mrowka 1984). As male reproductive output is not limited by their rate of gamete production they should, in accordance with Game Theory (Maynard Smith 1977), try and mate with as many females as possible. This is most effectively achieved by defending territories that include favourable spawning sites, a factor which incidentally predisposes them to the evolution of parental care at no extra cost. Females on the other hand should select the f i ttest males and not engage in parental care. This is because female fecundity generally increases with accelerating returns as body size increases, and they stand to lose more in terms of future reproductive output through loss of body growth, than males (Sargent & Gross 1986). Mouth- brooding probably evolved from monogamous substratum spawners in which there was division of labour between the sexes. This certainly appears to have been the case in the African cichlids in which all species currently e xhibiting substratum guarding are strictly monogamous, at least for the duration of one complete breeding cycle (Fryer & lIes 1972). Apogonids are an exception and males may simultaneously incubate the eggs from several females (Thresher 1984) . In the evolution of mouthbrooding the sex that was the active guarder would probably also assume the role of buccal incubation, and mouth-brooders may be either male, female, or bi-parental (Baylis 1981). Buccal incubation is comparatively rare in fishes and has been reported to definitely occur in only seven teleost families (Apogonidae, Ariidae, Belontiidae, Cichlidae, Cyclopteridae, Opisthognathidae and Osteoglossidae) all of which demonstrate male care, although females also brood in three of the families. Gill chamber incubation occurs wi thin the family Amblyopsidae in which females are reported to carry the eggs, although this is not mouth-brooding in the true sense of the word. Lucio ce phalids (Oppenheimer 1970; Blumer 1982) and malapterurids (Blumer QQ cit . ) are also reported to mouth-brood 209 although (Breder it has yet to be substantiated in these families & Rosen 1966). Mouth-brooding is therefore predominantly a male trait (Baylis 1981). Mouth-brooding behaviour has been extensively studied only in the apogonids, speciose ariids families and and cichlids. while the All three apogonids and are highly ariids are primarily marine, cichlids occur almost exclusively in fresh water. In all ariids mouth-brooding is paternal while in apogonids it is predominantly paternal with bi-parental care occurring in four species (Breder & Rosen 1966; Thresher 1984). Amongst the African cichlids it is largely maternal with paternal mouth-brooding occurring in only three tilapia species (Lowe 1959 in Breder & Rosen 1966; Thys van den Audenaerde 1970). Bi-parental mouth-brooding occurs in four species, also tilapines (Iles & Holden 1969). While bi-parental and paternal mouth-brooders are generally monogamous (Lowe 1959 in Breder Rosen 1966; Fryer & Iles 1972), maternal mouth-brooding cichlid species are generally polygamous and exhibit both & polygyny and polyandry. Since male gametes are energetically inexpensive to produce, males stand to lose far less if their gametes are unsuccessful than do females. In a situation of bi-parental or female mouthbrooding the male might therefore be inclined to abandon the role of parental care in favour of seeking out additional females with which to mate. This would explain the relative paucity of bi-parental mouth-brooders and the prevalence of polygyny amongst cichlid maternal mouth-brooders. Maternal bearing is considered to be evolutionarily the most stable while bi-parental and paternal mouth-brooding are considered unstable (Gross & Sargent 1985). In some apogonids, eggs of a relatively small size (0.4mm) have been retained enabling many hundreds and even thousands to be incubated simultaneously by the male, often from several different females (Thresher 1984). In this polygynous situation 210 male mouth-brooding may be evolutionarily more stable . Amongst the ariids, in which few eggs are produced and in which paternal care is prolonged, male reproductive output is very limited and mouth-brooding is expected to be evolutionarily unstable. In producing such large eggs requiring extended development periods, females may have effectively' coerced' the males into enduring longer and longer periods of mouth-brooding (sensu Gross & Sargent 1985). It is unlikely that an evolutionary trend toward fewer, larger eggs will continue much further in ariids because a point in the duration of mouthbrooding will eventually be reached beyond which males will not be able to survive. It would appear that Galeichthys males are already energetically over-extended during the incubation period (see pp.204-205) and have probably reached a state of "reproductive recklessness" (sensu Calow 1979) with respect to their costs of reproduction. In the light of the concomitant restriction of male reproductive output associated with monogamous paternal mouthbrooding its prevalence amongst teleosts is surprising . Baylis (1981) argues that monogamous male care may be stable because increased survivorship of young offsets their reduced primary fecundity. In addition, he argues that solitary male or female care would have evolved only if it imparted maximum and equal reproductive advantage on both sexes. This implies that maximum reproductive success will have been achieved in Galeichthys when anyone brood is fertilised and successfully reared in one breeding season, and males should therefore have no inclination to fertilise additional broods. 211 The barbel fishery - Implications of K-selection The K-selected traits of ~ feliceps and ~ ater which are of relevance to their management are presented in Table XXX. Table XXX. The K-selected life-history traits of G. feliceps and G. ater. Ufe-history G. FELICEPS G. ATER traits Males Females Males Females Age at 50% maturity (Yrs) 9.8 8.9 8.8 7.4 Maximum age (Vrs) > 16 >1 8 >15 >15 Growth rate (K) 0.1 47 0.120 0.193 0.132 Natural mortality rate (M) 0.1 37 0.101 0.211 0.133 Mean Fecundity 49.3 Mouth-brooding duration (Days) While most 31.5 ? ±: 140 r-selected species have high fecundities and density-dependent mortality of egg and larval stages (Cushing density1988), K-selected species have low fecundity, independent pre-recruit mortality (Adams 1980) and a small population compensatory capacity, or 'scope for compensation' (Garrod & Horwood 1984). Rather than producing a surplus of young they have adopted physiological and behavioural devices to overcome the vulnerability of their early life-history stages . A primary contributing factor to their population stability appears to be their consistently high recruit survival from one year to the next (i.e. low natural mortality). Garrod & Horwood (1984), in their particle size theory, suggest that natural mortality is essentially a reflection of the relative abundance of particles of different sizes in the environment (either predator or prey) and that the trajectory of natural mortality is common to all species of equal size in an ecosystem. This implies that although precocial species care for their young 212 until they are of a large size they will be as vulnerable as any other organism of similar size once they are released to fend for themselves. By d e laying the point of entry into the general mortality trajectory precocial species provide protection during the most vulnerable ontogenetic stages. The larger the young when released, the fewer predators there will be that are large enough to eat them. Sissenwine (1984) stated that predation was probably the major cause of natural mortality in fish. If this assumption, and the particle size theory of Garrod & Horwood (1984) are accepted, large body size would confer a considerable survival advantage on an individual. Galeichthys are able to use their erectable pectoral and dorsal spines to simulate this, a trait which effectively compensates for their comparatively slow growth rate. In addition, although they are prey to a wide variety of animals including fish, sharks, rays, seals and birds, their defensive weaponry is likely to render them less attractive as a food source throughout their life-cycle than other species of equable body size. This phenomenon was experimentally demonstrated by Hoogland et al. (1957 in Barnard 1983) who found that pike (Esox lucius), when provided with minnows, tenspined sticklebacks selectively, first and on three-spined the minnows, sticklebacks, then the fed ten-spined sticklebacks and lastly on the three-spined sticklebacks, which have large, locking spines. As K-selected species are generally highly competitive their populations are stable (Pianka 1970) and if managed properly, should theoretically be able to support sustained harvests in the long term (Adams 19?0). However, Holden (1974, 1977) has shown that stocks of several elasmobranchs which catches in trawl fisheries form by- for more fecund teleosts in the North Atlantic and in the North Sea, have crashed . Holden (2£ cit.) attributed these population declines to their low reproductive potential and lack of reproductive flexibility with respect to both the age at sexual maturity and fecundity. A major problem associated with these trawl fisheries was that 213 elasmobranch-directed effort could not be reduced even after it had become apparent that the populations were in a decline. While it would be possible to impose effort restrictions on barbel catches in the Port Alfred fishery, this study has shown that Galeichthys are sensitive to even mild exploitation and it is questionable whether they could be profitably exploited at the low levels of effort required to ensure their long-term survival. As a consequence of their low fecundity, recruitment in K- selected populations is directly related to spawner biomass and inversely proportional to the number of (Holden 1974, spawners harvested 1977). As the yield-per-recruit models assume that recruitment is independent of spawner biomass they are inappropriate for use with K-selected populations (Holden QQ cit.). A more appropriate method used by Holden & Meadows (1964 in Holden 1977) for e1asmobranch populations utilises growth rate, age at sexual maturity for female fish, brood size, the sex ratio of the young and the duration of the breeding cycle to determine the number of female young that an initial cohort of females (e.g. prevailing total 1000 individuals) mortality rate could produce under the for the stock. If the hypothetical female cohort is unable to replace itself during its lifetime then fishing effort is too high. As an exercise, the ~ feliceps data were applied to this model and yielded results which varied widely depending on the natural mortality rate used. As the natural mortality rate in a cohort is highest at the smallest sizes and decreases with increasing size it is meaningless to use a single value of M for all ages in the calculations. Using the three different values of M (0 . 078, 0.101, and 0.126) which were used in the Beverton-Holt spawner biomass-per-recrui t and yield-per-recrui t models the model produced the following results: for an initial cohort of 1000 females the model predicted that at the age of 18 years approximately 51 016, 40 758 and 32 957 female offspring would have been produced for the three values of M respectively. The above methodology is presented in Appendix IV . These figures 214 indicate that at the present total mortality rate ~ feliceps cohorts would more than replace themselves during their lifetime. This result suggests that the stock is not overexploited, and it is contrary to the results of the BevertonHolt spawner biomass-per-recruit models which indicated that the spawner biomass has declined considerably. It is interesting to note that although the Beverton-Holt model is thought to be inappropriate for use with K-selected species it provided more conservative results for the G. feliceps female population than the model of Holden & Meadows (2Q cit . ) above. Multi-species management approaches are generally considered preferable to the independent modeling of single species (Gulland & Garcia 1984) although this argument applies more strongly to trawl fisheries than it does to linefisheries. Species-specific management regulations such as size limits, bag limits and closed seasons are effectively employed in linefisheries and allow a high degree of control over fishing mortality (Smale & Buxton 1985; Van der Elst 1985; Bennett & Griffiths 1986; Buxton 1987; Huntsman & Waters 1987; Griffiths 1988; Hecht & Tilney 1989). This is particularly true for ariids which are seemingly unaffected by barotrauma and can be successfully returned to the water after capture . Several management alternatives are therefore open for the regulation of the Port Alfred barbel fishery. The spawner biornass-per-recruit and yield-per-recruit vs. fishing effort curves demonstrated that the G. ater stock has not been adversely affected by present levels of fishing effort. The following discussion is thus directed at the ~ feliceps stock. While the study has exploited the shown that G. feliceps may be imposition of restrictive regulations overin the fishery at present are likely to be unpopular. This is because barbel currently have little commercial value and are regarded 215 largely as taking a "nuisance" management detecting the in the action critical collapse will occur. now pOint fishery. lies beyond in The danger in not the which difficulty imminent of stock As barbel are slow maturing and since predominantly mature fish are caught in the fishery recruitment failure will not be detected until several years after the event when it may be too late to rectify the damage. When Kselected stocks collapse, recovery is slow and seldom complete because once their numbers have been cropped below the level at which they are able to occupy their spacial and foraging niches, these are rapidly filled by opportunistic, altricial species which are highly successful colonisers (Hsu 1982 in Bruton 1989). An important point for consideration in the interpretation of yield-per-recrui t and spawner biomass-per-recrui t curves in this study was that a large percentage of the adult males were excluded from the catches for a period of approximately four months each year while mouth-brooding. The sex ratio and population structure reflected in the catches may not have been a true representation of the population structure in the ocean and the catch-curves used will have yielded mortality values for males which were over-estimates. However, the sex ratios did indicate that catches during the spawning and mouth-brooding females (female:male ratio and iL:.. ater respectively, = period significantly favoured 1.65:1 and 2.23:1 for iL:.. feliceps see Table XXVII), suggesting that some sort of effort restriction should be imposed to reduce barbel catches during this period. Restrictions in the form of a closed season, bag limits or size limits would all prove effective. Given the fact that barbel are caught in association with kob, their catches are directly influenced by the amount of effort that is directed at the latter. In recent years there has been a decline in kob-directed effort in the Port Alfred fishery 216 (Hecht & Tilney 1989) and barbel are experiencing a temporary reprieve from previous, higher levels of exploitation. Kob formed the major component of the total annual landings in the fishery between the 1960's and the mid 1980's. catches were surpassed for the first In 1987 kob time by those of two sparids Argyrozona argyrozona and pterogymnus laniarius, which are caught in deeper water (Hecht & Tilney .QQ cit.). The imposition of restrictions on barbel catches at this stage may therefore be premature. As barbel form an important source of revenue and protein for the largely impoverished fishermen who catch them, there is a socio-economic consideration which also weighs against the imposition of catch restrictions at present. Should increasing market prices result in barbel exploitation becoming profitable it will be necessary to re-assess the stocks before arriving at a suitable management strategy. The differential exploitation of the sexes is seen as an important area for regulation. Protection of the spawner biomass would also be a priority and could be achieved using minimum size limits set at a fork length of approximately 320-330mm . However, as this would exclude a large proportion of the stock presently available to the fishery, it may prove unacceptable to fishermen. The most effective strategy would therefore be to impose a between closed season during the mouth-brooding period September and December. This would limit effort, protect the population sex ratio, and have the added advantage of being easy to control. 217 BIBLIOGRAPHY Adams, P . B. (1980). Life history patterns in marine fishes and their consequences for fisheries management. Fish. Bull . 78 (1):1-12 . & McLean, R.B. (1985). Estimation of largemouth bass, Micropterus salmoides Lacepede, growth using the liver somatic index and physiological variables . ,L.. Fish Biol. 26: 111-126 . Adams, S.M. Alexander, A.J. & van Staaden, M. (1989). Alternative sexual tactics in male bladder grasshoppers (Orthoptera, Pneumoridae). In: Alternative life-history styles of anima l s, M.N. Bruton (ed.), Kluwer Academic Publishers, Dordrecht. pp.26l-277. Alexander, R.McN. (1965). Structure and function in the catfish. ,L.. Zool. Lond. 148:88-152. Alexander, R . McN. (1970). Mechanics of the feeding action of various teleost fishes. ,L.. Zool. Lond. 162: 145-156. Ali, M.A. & Anctil, M. (1976). Retinas of Fish. An Atlas, Springer-Verlag, Berlin. 284pp. Allen, K.R. & Kirkwood, G.P . (1988). Marine mammals. In: Fish population dynamics, (Second edition). J.A. Gulland (ed . ), John Wiley & Sons, New York. pp.251-269 . Al-Nasiri, S.K. & Hoda, S . M. S . (1977) . Notes on the developmental stages of sea catfish, Arius thalassinus (Rupp . ) from the Arab Gulf. Bull. Biol. Res. Cent. 9:41-49. 218 Andrew, T.G. (1987). A comparison of the sensory importance and use i n feeding of the barbels of Ga1eichthys feliceps and ~ ater. Unpublished honours degree project, Rhodes University, Grahamstown, South Africa. Araujo, F.G. (1984). Habitos alimentares de tres Bagres Marinhos (Ariidae) no Estuario da Lagoa dos Patos (RS), Brasil. Atlantica, Rio Grande. 7: 47-63. Atz, J . W. (1958) . A mouthful of babies. Anim. Kingd. 61:182- 186 . Bachop, W.E. Schwartz, & F.J. (1974). Quantitative nucleic acid histochemistry of the yolk sac syncytium of oviparous teleosts: Implications for hypotheses of yolk utilization. In: The early life-history of fish . J.H.S. Blaxter (ed.), SpringerVerlag, New York . pp.345-353. Bagenal, T.B. & Tesch, F.W. (1978). Age and growth. In: Methods for assessment of fish production in fresh waters. T. Bagenal (ed.), Third edition, Blackwell Scientific Publications, London . pp.101-136. Bakhtin, Ye.K. (1976). Morphology of the Olfactory Organ of ~ Some Fish Species and a Possible Functional Interpretation. Ichthyol. 16: 786-804. Balon, E.K. (1975a). Reproductive guilds of fishes: A proposal and definition. Balon, E.K. development. ~ Fish . Res. Bd. Can . 32 (6):821-864 . ~ (1975b). Terminology of intervals in fish Fish. Res. Bd. Can. 32 (9):1663-1670. Balon, E.K . (1977). Early ontogeny of Labeotropheus Ahl, 1927 (Mbuna, Cichlidae, Lake Malawi), with a discussion on advanced protective styles in fish reproduction and development. Env. Biol. Fish. 2 (2):147-176 . 219 Balon, E.K. (1979). The juvenilization process in phylogeny and the altricial to precocial forms in the ontogeny of fishes. Env. Bio l . Fish. 4:193-198. Balon, E.K. (1981a). Additions classification of reproductive and ammendments to the styles in fishes. Env . Biol.Fish. 6 (3/4):377-389. Balon, E.K. (198lb). About processes which cause the evolution of guilds and species. Env. Biol. Fish. 6:129-138. Balon, E.K. precocial (1981c). forms Saltatory processes and in the ontogeny of fishes. altricial AIDer. zool. to 21 (2) :573-596. Balon, E.K. (1984). Patterns in the evolution of reproductive styles in fishes. In: Fish reproduction: Strategies and tactics, G.W. Potts & R.J. Wootton (eds.), Academic Press, London. pp.35-53. Balon, E.K. (1985). The theory of saltatory ontogeny and life history models revisited. In: Early life histories of fishes. New developmental, ecological and evolutionary perspectives. E . K. Balon (ed.), Dr. W. Junk Publishers, Dordrecht. pp. 1 3- 30. Balon, E.K. (1986). Saltatory ontogeny and evolution. Rev. Biol. - B. Forum 79 (2):151-190. Balon, E.K. (1989). The epigenetic mechanisms of bifurcation and alternative life-history styles. In: Alternative lifehistory styles of animals. M.N. Bruton (ed.), Kluwer Academic Publishers. pp.467-501. 220 Balon, E.K. (In press). Probable evolution of the coelacanth's reproductive style: lecithotrophy and orally feeding youngs in cichlid fishes and in Latime ria chalumnae. In: evolution of coelacanths. J.A. Musick & M.N. Biology and Bruton (eds.). Env. BioI. Fish . Balon, E.K. & Flegler-Balon, C. (1985) . Microscopic techniques for studies of early ontogeny in fishes: problems and methods of composite descriptions. pp. 33-55. In: Early life histories of fishes: New developmental, perspectives, E.K. Balon ecological (ed.) . Dr . W. and evolutionary Junk Publishers, Dordrecht. 280pp. Bardach, J.E. & Atema, J. (1971). The sense of taste in fishes . In: Handbook of sensory physiology, Vol. ~ L. M. Beidler (ed.). Springer Verlag, Berlin. Barnard, C.J. (1983). Animal behaviour. Ecology and evolution, Croom Helm, London & Canberra. 339pp. Barnard, K.H. (1950). Descriptive catalogue of South Decapod Crustacea. Ann. ~ African Afr. Mus. 38: 1-837. Bartlett, J.R., Randerson, P.F . & Ellis, D.M. (1984). The use of analysis of covariance in the back-calculation of growth in fish. ~ Fish BioI. 24:201-213. Bawazeer, A.S. (1987). The fishery biology and management of the stock of chim, the ,giant sea catfish (Arius thalassinus) in Kuwait waters. Kuwait Bull. Mar. Sci. 1987(9):87-100. Baylis, J.R. (1981). The evolution of parental care in fishes, with reference to Darwin's rule of male sexual selection. Env. BioI. Fish. 6(2):223-251 . 221 Beamish, R.J. (1979). Differences in the age of Pacific hake (Merluccius productus) using whole otoliths and sections of otoliths. ~ Fish. Res. Bd. Can. 36:141-151 . Beamish, R.J. & McFarlane, G. A . (1983). The forgotten requirement for age validation in fisheries biology. Trans. Amer. Fish. Soc. 112:735-743. Beamish, R.J. & McFarlane, G.A. (1987). Current trends in age determination methodology. In: Age and growth of fish. Summerfelt & G.E. Hall (eds.), R.C. Iowa State University Press. pp.15-42. Bennett, B.A. (1985) . A mass mortality of fish associated with low salinity conditions in the Bot River estuary. Trans. Soc. ~ ~ Afr. 45 (3/4): 437-447 . Bennett, B.A . & Griffiths, C.L. (1986). Aspects of the biology of galjoen Coracinus capensis (Cuvier) off the South-Western Cape, South Africa. Bennett, J.T., ~ Afr. Boehlert, Mar. Sci. 4:153-162. ~ G.W. & Turekian, K.K. (1982). Confirmation of longevity in Sebastes diploproa (Pisces: Scorpaenidae) from 2l0 Pb / 226 Ra measurments in otoliths. Mar. Biol. 71:209-215. Berry, P . F. (1978). Reproduction, growth and production in the mussel Perna perna Africa. Invest. ~ (Linnaeus), on the east coast of South oceanogr. Res. Inst. 48:1-28. Beverton, Holt, R.J.H. & S.J. (1957). On the dynamics of exploited fish populations. Fish. Invest . London, Ser. 1119:1533. 222 Beverton, R.J.H. (Rapporteur), Cooke, J . G., Csirke, J . B., Doyle, R.W., Hempel, G. , Holt, S.J., MacCall, A.D., Policansky, & Wiebe, P.H. (1984) . Dynamics of single species. Group Report. In: Exploitation of marine communities, R.M. May (ed.). Dahlem D.J., Roughgarden, J., Shepherd, J.G., Sissenwine, M.P. Konferenzen. Berlin: Springer-Verlag. pp.13-58 . Blaber, S . J.M. (1979). The biology of filter feeding teleosts in Lake St Lucia, Zululand. J. Fish Biol. 15:37-59. Blacker, R.W. (1974). Recent advances in otolith studies. In: Sea fisheries research, F.R. Harden-Jones (ed.). Elek Science, London . pp.67-90. Blaxter, J.H.S. (1969). Development: Eggs and larvae. In: Fish Physiology, W.S. Hoar & D.J. Randall (eds.). Volume III. Academic Press, New York. pp.177-252. Blumer, L.S. (1982) . A bibliography and categorization of bony fishes exhibiting parental care. Zool. J. Linn. Soc. 76:1-22. Bond, C. E. (1979). Biology of Fishes. Saunders College Printing, Philadelphia. 514pp. Bothe, L. hakes (1971). Growth and otolith morphology of the Cape Merluccius Investl. ~ capensis Div. Sea Fish. Cast. ~ and ~ paradoxus Franca. Afr. 97:1-32. & Rosen, D.E. (1966). Modes of reproduction in fishes. T.F.H. Publications, New Jersey. 941pp. Breder, C.M . Jr. Brett, J.R. (1979). Environmental factors and growth. In: Fish physiology . Volume VIII, Bioenergetics and growth. W.S. Hoar, D.J. Randall & J. R. Brett (eds.), Academic Press, New York, pp.599-675 . 223 Brothers, E . B. (1987). Methodological approaches to the examination of otoliths in ageing studies . In: Age and growth of fish. R.C . Summerfe1t & G.E. Hall (eds.), Iowa State Univ . Press. pp.319-330 . Brothers, E.B. & Mathews, C.P. (1987). Application of otolith microstructural studies to age determination of some commercially valuable fish of the Arabian Gulf. Kuwait Bull . Mar. Sci. 1987 (9):127-157 . Bruton, M.N. (1989). The ecological significance of alternative life-history styles. animals, \ M.N. In: Alternative life-history styles of Bruton (ed.). Kluwer Academic Publishers, Dordrecht. pp.503-533. Buchan, P. R. Smale, & M. J . (1981). Estimates of biomass, consumption and production of Octopus vulgaris Cuvier off the east coast of South Africa. Invest . ocaenogr. Res . Inst. ~ 50: 1-9. Bullock, T.H. (1973). Seeing the World through a New Sense: Electroreception in Fish. Am. Sci . 61: 316-325 . Butterworth, D.S., Punt, A.E., Borchers, D.L. , Pugh, J.B ., & Hughes, G.S. (1989). A manual of mathematical techniques for linefish assessment. ~ Afr. Natl . Sci. Progr. ~ No. 160 . Scientia Printers, CSIR, Pretoria . 89pp. Buxton, C.D. (1987). Life history changes of two reef fish species in exploited and unexploited marine environments in South Africa. Unpublished Ph . D. thesis, Rhodes University, Grahamstown. 220pp. Buxton, C . D. & Clarke, J.R. (1985) . Notes on the diet of Pterogymnus 1aniarius (Cuvier) (Pisces: Sparidae). Zool. 20 (2):68-71 . 224 ~ Afr. ~ Buxton, C . D. & Clarke, J . R. (1986) . Age, growth and f~eding of the blue hottentot Pachymetopon aeneum (Pisces: Sparidae) with notes on reproductive biology . Afr. ~ Zool. 21(1):33~ 38. Buxton, C.D . & Clarke, J . R. (1989). A survey of the recreational rock-angling fishery at Port Elizabeth, on the south-east coast of South Africa. Afr . ~ ~ Mar. Sci . 8:183- 194 . C . D. , Smale, M.J., Wallace, J .H. & Cockroft, V.G. (1984). Inshore small-mesh trawling survey of the Cape south coast. Part 4. Contributions to the biology of some Teleostei Buxton, and Chondrichthyes. Bye, V.J . timing (1984). of ~ Afr . Zool. 19:180-188 . ~ The role of environmental factors in the reproductive strategies and tactics, cycles. G. W. In: Potts & Fish R.J . reproduction: Wootton (eds.). Academic Press, London. pp.187-205. Calow, P. (1979). The cost of reproduction - A physiological approach. Biol. Rev. 54:23-40 . Campana, S . E. & Neilson, J.D. otoliths. Can. Caprio, J. ~ (1985) . Microstructure of fish Fish . Aquat. Sci. 42:1014-1032. (1982). High sensitivity and specificity of olfactory and gustatory receptors of catfish to amino acids. In: Chemoreception in fishes. Developments in aquaculture and fisheries science, volume T.J. ~ Hara (ed.), Elsevier Scientific Publishing Company, Oxford. pp . 109-134. Caro, T.M. & Bateson, P. (1986). Organization and ontogeny of alternative tactics. Anim . Behav . 34:1483-1499 . Castelnau, De M. Le Comte, F. (1861 ) . Memoire sur les poissons de L'Afrigue Australe. Bailliere, Paris . 78pp. 225 Chan, T-Y. (1987). The role of male competition and female choice in the mating success of a lek-breeding southern African cichlid fish Pseudocrenilabrus philander (Pisces: Cichlidae). Unpublished M.Sc. Thesis, Rhodes University, Grahamstown. 109pp. Chardon, M. (1968). Anatomee comparee de L'Apparei1 de Weber et des structures connexes chez les Siluriformes. Musee Royal de L'Afrigue Centrale ~ Tevuren, Belgigue Annales - Serie In ~ - Sciences Zoologigues ~ No. 169 . Christensen, J.M. (1964). Burning of otoliths, a technique for age determination of soles and other fish. J. Cons. into Explor. Mer. 29:73-81. Clark, (1985). C . W., Charles, A. T., Beddington, J . R. & Mangel, M. Optimal capacity decisions in a developing fishery. Mar. Resource Econ. 2(1):25-53. Clarke, J.R. Aspects of (1988). the biology of the musselcracker, Sparodon durbanensis, and the bronze bream, Pachvmetopon grande, (Pisces: Sparidae), with notes on the eastern Cape recreational rock-angling and spear fisheries . Unpublished 141pp . Coates, D. M.Sc. Thesis, Rhodes University, Grahamstown. (1988). Length-dependent changes in egg size and fecundi ty in females, and brooded embryo size in males, of fork-tailed catfishes (Pisces: Ariidae) from the Sepik River, ~ Papua New Guinea, with some implications for stock assessments. Fish Biol. 33:455-464. Cockroft, A.C. & McLachlan, A. (1987). Nitrogen regeneration by the surf zone penaeid prawn Macropetasma africanum. Mar. Biol. 96:343-448 . 226 Coetzee, D.J. & Pool, R.C. (1984). Stomach content analysis of the sea barbel, Galeichthyes feliceps (Valenciennes in C & V), from the Swartvlei system, southern Cape. Afr. ~ Zool. ~ 20 (1):33-37. Coetzee, P.S. Baird, & D. (1981). Cheimerius nufar (Ehrenberg, St Croix Island, Algoa Bay. Compagno, L.J.V. (In Age, 1820) press) . food of (Sparidae), collected off Afr. ~ growth and Zool. 16(3):137-143. ~ Shark exploitation and conservation. NOAA, NMFS Tech. Bull. Connell, J.H. (1980). Diversity and the coevolution of competitors, or the ghost of competition past. Oikos 35:131- 138. Connell, J.H. (1985). On testing models of competitive coevolution. Oikos 45:298-299. Cortes, O.D.G. (1984). Estimacion del crecimiento y mortalidad del chivo cabezon AriopSis bonillai (Miles, 1945) (Pisces: Siluriformes: Ariidae), en la Cienaga Grande de Santa Marta, Colombia. An. Inst. Inv. Mar. Punta de Betin 14:67-84. Costa, M.de L. & Juras, I.de A.G.M. (1981/82). Determinacao da idade e crescimento do bandeirado, Bagre bagre (Linnaeus, 1766) . Sao Luis - Estado do Maranhao. Biol. Lab. Hidrob, Sao Luis, MA-BRASIL4(1):17-50. Cross, R. (1985). The preparation of biological material for electron microscopy. Rhodes University, Unpublished manual. Crozier, w.w. (1989). Age and growth of angler-fish (Lophius piscatorius L.) in the North Irish Sea. Fish. Res. 7:267-278. 227 Cunningham, J.E . R. & Balon, E.K. (1985). Early ontogeny of Adinia xenica (Pisces, Cyprinodontiformes): 1. Early development of e mbryos in hiding. Env. Biol. Fish. 14 (2/3): 115-166. Cushing, D.H . (1975). Marine ecology and fisheries, Cambridge University Press, Oxford. 278pp. Cushing, D.H. (1981). Fi sheries Biology . A Study in Population Dynamics. Second edition, The University of Wisconsin Press Ltd., Wisconsin. 295pp . Cushing, D.H. (1988). The study of stock and recruitment. In: Fish population dynamics, (Second edition). J.A. Gulland (ed.). John Wiley & Sons, New York. pp . 105-128 . Dan, S.S . (1981). Age and growth in the catfish Tachysurus tenuispinis (Day). Ind. ~ Fish. 27(1/2):220-235. Davic, R . D. (1985). In search of the ghost of competition past. Oikos 45:296-298. Day, J.H. (1967) . A monograph of the Polychaeta of southern Africa: Part 2 . Sedentaria. Brit. Mus . Nat . Hist. Lond. 878pp. Day, J.H . (1974) . A Guide To The Marine Life On South African Shores. A.A. Balkema, Cape Town. 300pp. Day, J.H., Blaber, S . J.M. fishes. In: & Wallace, J.H. (1981). Estuarine Estuarine ecology with particular reference to southern Africa, J.H. Day (ed.). A.A. Balkema, Cape Town. (1970). The benthic pp .1 97-221. Day , J .H ., Fie ld , J.G. & Penrith, M.J. ~ fauna and fi s hes of False Bay, South Af rica . Trans. Afr. 39, Part 1 : 1-108. 228 ~ Soc . De Vlaming, V.L., Kuris, A. & Parker, F.R.Jr. (1978). Seasonal variation of reproduction and lipid reserves in some subtropical Cyprinodontids. Trans. Am. Fish. Soc. 107(3):464472. Delahunty, G. & De Vlaming, V.L. (1980). Seasonal relationships of ovary weight, liver weight and fat stores with body weight in the goldfish, Carassius auratus (L.). ~ Fish Biol . 16:5- 13. Dmitrenko, Ye.M. (1970). Reproduction of the sea catfish [Arius thalassinus (Rupp.)] in the Arabian Sea. ~ Ichthyol. 10:634- 641. Dmitrenko, catfish Ye.M. Arius (1975). thalassinus, Peninsula (India). Duellman, Size-age composition of W.E. ~ in the vicinity of the giant Kathiawar Ichthyol. 15:695-702. (1989). Alternative life-history styles in anuran amphibians: evolutionary and ecological implications . In: Alternative life-history styles of animals, M.N. Bruton (ed.), Kluwer Academic Publishers, Dordrecht . pp.l01-126. Emlen, S . T. & Oring, L. W. (1977). Ecology, sexual selection, and the evolution of mating systems. Science 197(4300):215- 223. Etchevers, S . L. (1978). Contribution to the biology of the sea catfish, Arius spixii ,(Agassiz) (Pisces-Ariidae), south of Margarita Island, Venezuela. Bull. Mar. Sci. 28(2):381-385. Euzen, O. (1987). Food habits and diet composition of some fish of Kuwait. Kuwait Bulletin of Marine Science. 9:65-85. Everhart, W. H. & Youngs, W. D. (1981). Principles of fishery science. Cornell Univ. Press, New York. 349pp. 229 Fishelson, L., Montgomery, W.L. & Myrberg, A.A. (1985). A new f at surgeonfishes body associated with the gonad of (Acanthuridae: Te1eostei). Mar. BioI. 86:109-112 . & Korth, J.W. (1988). Karyology of the marine catfish Bagre marinus (Ariidae) with an analysis of chromosome numbers among siluriform fishes. Fitzsimons, ~ J.M., LeGrande, W.H. Ichthyol. 35 (2):189-193. ~ Flath, L . E. & Diana, J.S. (1985) . Seasonal energy dynamics of the alewife in southeastern Lake Michigan. Trans. Amer. Fish. Soc. 114:328-337. Freytag, G. (1980). marmorata. Cybium 3e Fritz, E . S. Spearman (1974). rank Problems in ageing Notothenia rossii (8):43-51. B~rie Total correlation diet comparison coefficients. in fishes Copeia by 1974 (1):210-214. Fry, F . E.J. (1957). Aquatic respiration of fish. In: The physiology of fishes, Volume I, Metabolism . M.E. Brown (ed.). Academic Press Inc., New York, pp.1-63. Fryer, G. (1984). The conservation and rational exploitation of the biota of Africa's great lakes . In: threatened natural (ed.), National Scientific habitats. A.V. Programmes Hall Report 92, Conservation of South Africa CSIR, Pretoria. pp . 135-154. Fryer, G. & lIes, T.D. (1972) . The cichlid fishes of the Great Lakes of Africa. Oliver & Boyd, Edinburgh. 641pp . Gadgil, M. & Bossert, W.H. (1970) . Life historical consequences of natural selection . Amer. Nat. 104 (935):1-24. 230 Garrod, D.J. & Ho rwood, J.W. (1984). Reproductive strategies and the response to exploitation. In: Fish reproduction. Strategies and tactics, G. W. Potts Academic Press, London. pp . 367-384. & R. J. Wootton (eds.), Geldenhuys, N.D . (1978). Age determination of the South African round herring Etrumeus micropus and length and age composition of the commercial catches, 1965-1973. Investl. Brch . ~ ~ Sea Fish. Afr . 115:1-16. Gilbert, C.K. & Gilbet, P . W. (1980). Shark. In: Academic American Encyclopedia, 17:242-244. Arete Publishing Co., Princeton, N.J. Gorlick, D.L. (1980). Ingestion of Host Fish Surface Mucus by the Hawaiian Cleaning Wrasse, Labroides phthirophagus (Labridae), and its Effect on Host Species Preference. Copeia 1980 (4)1 863-868. Gosline, W.A. (1973). Considerations Regarding the Phylogeny of Cypriniform Fishes, with Special Reference to Structures Associated with Feeding. Copeia 1973 (4):761-776. Gosline, W.A. (1975). The palatine-maxillary mechanism in catfishes, with comments on the evolution and zoogeography of modern siluroids. Occ. Calif. Acad. Sci . 120:1-31. ~ Graham, M. (1935). Modern theory of exploiting a fishery, and application to North Sea trawling. ~ Cons. Expl. Mer 10:264274 . Graham, M. (1938). Rates of fishing and natural mortality from the data of marking ments. ~ Cons. Int. Expl . Mer 13:76-90. 231 • Grande, L. & Lundberg, J.G. (1988). Revision and redescription of the genus Astephus (Si1uriformes: Icta1uridae) with a discussion of its phylogenetic relationships . ;L. Vert. Paleont. 8 (2):139-171. Greenwood, P.H. (1989). Ontogeny and evolution: saltatory or otherwise? In: Alternative life-history styles of animals. M.N. Bruton (ed.) . Kluwer Academic Publishers, London. pp.245-259. Greenwood, (1966). P.H., Rosen, Phyletic D.E., studies Weitzman, of S.H. teleostean Myers, & fishes, G.S. with a provisional classification of living forms . Bull. Amer. Mus. Nat. Hist. 131 (4):341-455. Gregory, W.K . (1959). Fish skulls. A study of the evolution of natural mechanisms. Eric Lundberg, Florida. 481pp. Griffiths, M.H. (1984). Electron microscopical observations of the tastebuds of ~ feliceps (Pisces: Ariidae). Unpublished Honours Degree Project, Rhodes University, Grahamstown, South Africa. Griffiths, M.H. (1988). Aspects of the biology and population dynamics of the geelbek Atractoscion aeguidens (Cuvier) (Pisces: Sciaenidae) off the South African coast. Unpublished M.Sc. Thesis, Rhodes University, Grahamstown. 149pp . Griffiths, M.H. & Hecht, T. (1986). A preliminary study of age and growth of the monkfish Lophius upsicephalus (Pisces:Lophiidae) on the Agulhas Bank, South Africa. ~ Afr. ;L. Mar. Sci. 4:51-60 . Gross, M.R . (1982). Sneakers, satellites, and parentals, polymorphic mating strategies in North American sunfishes. Tierpsychol. 60:1-26. 232 ~ Gross, M.R. (1984). Sunfish, salmon, and the evolution of alternative reproductive strategies and tactics in fishes. In: Fish reproduction. Strategies and tactics, G. W. Potts & R. J . Wootton (eds.). Academic Press, London. pp . 55-75 . Gross, M.R . & Sargent, R.C. (1985). The evolution of male and female parental care in fishes. Amer. Zool. 25:807-822. Grossman, G.D. (1982). Community regulation and patterns of resource partitioning. In: Gutshop Siemenstad ~. G.M. Cailliet & C.A. (eds.). Washington Sea Grant Program, WSG-WO 82- 2, University of washington, Washington . pp.166-177. Gudger, E.W. (1916). The gaff-topsail (Felichthys felis). A sea catfish that carries its eggs in its mouth. Zoologica II (5):125-158. Guillemot, P.J., Larson, R.J. & Lenarz, W.H. (1985). Seasonal cycles of fat and gonad volume in five species of northern California rockfish (Scorpaenidae: Sebastes). Fish. Bull. 83 (3):299-311. Gulland, J.A . (1977). The analysis of data and the development of models. In:Fish Population Dynamics. J.A. Gulland (ed.). John Wiley & Sons, London. pp.67-95. Gulland, J.A. (1978). Assessment of a fishery. In: Methods for assessment of fish production in fresh waters, IBP Handbook No.3., T. Bagenal (ed.), Blackwell Scientific Publications, London. pp.274-288. Gulland, J.A. (1985). Fish stock assessment. A manual of basic methods. John Wiley & Sons, Chichister. 223pp. 233 Gulland, J.A . & Garcia, S. (1984). Observed patterns in muli tspecies fisheries. In: Exploitation of marine communi ties. R.M. May (ed.). Dahlem Konferenzen . Berlin: Springer-Ve rlag. pp.155-190 . Gunderson, D.R. & Dygert, P.H. (1988). Reproductive effort as a predictor of natural mortality rate. Cons. into Explor . ~ Mer 44:200-209. Gunter, G. (1947). Observations catfish, Galeichthys felis on breeding of (Linnaeus). the marine Copeia 1947 (4):217- 223. Hall, B.K. evolution. Holder & (1983). In: Epigenetic control in Development and evolution, development B.C. Goodwin, and N. C. C . Wylie (eds . ). Cambridge Uni versi ty Press, London. pp. 353-377 . Halstead, B.W. (1978). Poisonous and venemous marine animals of (Revised the world, edition). Darwin Press, Princeton. 1043pp. Hamlett, W.C . , Schwartz, F.J. & Liberato J.A. DiDio. (1987). Subcellular organization of the yolk syncytial-endoderm complex in the preimplantation yolk sac of the shark, Rhizoprionodon terraenovea. Cell Tissue Res. 247:275-285. Hanekom, N.M . (1980). A study of two thalassinid prawns in the non-spartina regions of the Swartkops Estuary.. Unpublished Ph.D. thesis, University of Port Elizabeth, South Africa. Hassur, R.L . (1970). Studies on the osteology of catfishes, Order Siluriformes. Ph.D. Dissertation, Stanford University. University Microfilms International, Michigan . 133pp . 234 Hecht, T . & Appelbaum, S. (1988). Observations on intraspecific aggression juvenile and coeval Clarius gariepinus controlled conditions. Hecht, T. & Baird, D. the panga, sibling cannibalism by larval (Clariidae: Pisces) and under Zool. Lond. 214:21-44. ~ (1977). Contributions to the biology of pterogymnus laniarius (Pisces Sparidae): Age, growth and reproduction. Zoologica Africana 12 (2):363-372. Hecht, T. & Hecht, A. (1981). A descriptive systematic study of the otoliths of the Neopterygean marine fishes of South Africa, Part IV, Siluriforroes and Myctophiforroes. Trans . Soc. ~ Hecht, ~ Afr. 44 part 3:401-440. T. & Smale, M.J. (eds.), (1986). Proceedings of a workshop on age determination and growth modelling of South African marine linefish. Investgtl. No. 21, JLB Smith ~ Institute of Ichthyology, Grahamstown. 40pp. Hecht, T. & Tilney, R.L . (1989). The Port Alfred fishery. A preliminary investigation. Hildemann, W.H. (1962). ~ Afr. ~ Mar. Sci. 8:103-117. Immunogenic studies of poikilothermic animals. Amer. Nat. 96:195-204. Hixon, M.A. (1980) . Competative Interactions California Reef Fishes of the Genus Erobiotoca. between Ecology 61 (4) :918-931. Hoese, H . D. (1966). Ectoparasitism by juvenile sea catfish, Galeichthys felis. Copeia 1966 (4): BBO-BB1. Holden, M.J. (1974). Problems in the rational exploitation of elasmobranch populations and some suggested solutions. In: Sea Fisheries Research F.R. Harden-Jones London . pp.117-13B. 235 (ed.). Logo Press, Holden, M.J. (1977). Elasmobranchs. In: dynamics, J.A. Gulland (ed.), John Wiley Fish population & Sons, London. pp.187-215. Howes, G. J. (1985) . The phylogenetic relationships of the electric catfish family Malapteruridae ( Teleostei: Siluroidei). Nat. Hist. 19:37-67. ~ Hughes, G. S. (1986). Examining methods of fitting age / length data von to the Bertalanffy growth curve with a view to applying a simplified version of the Beverton and Holt yield per recruit model . Department of Applied Mathematics, University of Cape Town, Internal SANCOR Report. 70pp. Hughes, G.S. Punt, & A.E. (1988). PC-Yield User's Guide. Internal SANCOR Linefish Programme document. 50pp. Hughes, R. N . ( 1980). Optimal Foraging Theory in the Marine Context. Oceanogr. Mar. Biol. Ann. Rev. 18:423-481. Huntsman, G.R. & Waters, J.R . (1987). Development of management plans for reef fishes - Gulf o f Mexico and U.S. South Atlantic. In: Tropical snappers and groupers -Biology and fisheries management . J.J. Polovina & S . Ralston (eds.). Westview Press, Boulder, Colorado. pp . 533-560. Hynes, H.B.N. (1950). The food of fresh-water sticklebacks (Gasterosteus aculeatus and Pygosteus pungitius), with a review of methods used in studies of the food of fishes. ~ Anim . Ecol. 19 (1): 36-58. Hyslop, E.J. (1980). Stomach contents analysis - a review of methods and their application. Iles, T.D. ~ Fish BioI. 17:411-429. (1984). Allocation of resources to gonad and soma in Atlantic herring Clupea harengus L. In: Fish reproduction. 236 Strategies and tactics, Academic Press, London. pp.331-347 . l Ies, T.D. & Holden, M.J. (1969). Bi - parenta1 mouth brooding in Ti1apia ga1i1ea (Pisces, Cich1idae). Zool. [ ~ Lond. 158:327-333. Jantsch, E. (1980). The self-organizing universe. Scientific and human implications Qi the emerging paradigm of evolution. Pergamon Press, Oxford. 343pp . Jayaram, K.C. & Dhanze, J.R. (1978). Siluroid fishes of India, Burma and Ceylon. 22. A preliminary review of the genera of the family Ariidae (Pisces: Siluroidea). Matsya 4:42-51. Jayaram, K.C. & Kailola, P. (1984). Ariidae . In: Fischer, W. & G. Bianchi (eds.). FAO species identification sheets for fishery purposes . . Western Indian Ocean; Prepared and printed with the (Fishing support of 2ll. ~ the Danish International Development Agency (DANIDA). Rome, FAO of the UN, Vola 1-6: pag. var. Jayaram, K.C. & Singh, R. (1984). Contributions to the study of bagrid fishes 16. The skull of Chrysichthys auratus (Pisces, Bagridae). Rev. Zool. afro 93 (3):606-626. Jones, P.W. , Martin, F.D. & Hardy, J.D.Jr. (1978). Family Ariidae . In: Development of fishes of the mid-Atlantic bight. atlas of Acipenseridae An ~ larval and juvenile stages. Volume L through Ictaluridae. U. S. Fish & Wildlife Service. BioI. Servo Program. FWS!OBS-78!12. pp.301-307. Kilburn, Africa. R. & Rippey, E. (1982). Sea Shells of Southern Macmillan, Johannesburg. 249pp. K1eerekoper, University H. (1969) . Olfaction in Fishes. Press, Bloomington, London . 222pp. 237 Indiana Kleerekoper, H. (1982) . The role of olfaction in the orientation of fishes . In: Chemoreception in fishes. Developments in aguaculture and fisheries science, volume T.J. Hara (ed.), Elsevier Scientific Publishing ~, Company, Oxford. pp.201-225. Kuftina, N. D. Novikov, & G.G. (1986). Embryonic growth and utilization of yolk storage proteins during early ontogeny of cod, Gadus morhua, at different temperatures. ~ Ichthyol. 26 (5):76-88. Lagler, K.F . , Bardach, J.E., Miller, R.R. & May Passino, D.R . (1977). Ichthyology. John Wiley & Sons, New York. 506pp. Lara-Dominguez, A.L., Yanez-Arancibia, A. & Linares, F.A . (1981). Biologia y ecologia del bagre Arius melanopus GUnther en la Laguna de Terminos, sur del Golfo de Mexico (Pisces: Ariidae). An. Inst. Cienc. del Mar y Limnol. Univ. Nal. Aut6n. B (1):267-304. Mexico Lauder, G.V. & Clark, B. D. (1984). Water flow patterns during prey capture by teleost fishes. Leaman, B.M. & Nagtegaal, ~ BioI. 113:143-150. D.A. ~ (1987). Age validation and revised natural mortality rate for yellowtail rockfish. Trans. Amer. Fish. Soc . 116:171-175. Lecomte, F., Meunier, F.J. & Rojas-Beltran, R. (1986). Donnees preliminaires sur la croissance de deux teleosteens de Guyane, Arius proops (Ariidae, Siluriformes) et Leporinus friderici (Anostomidae, Characoidei). Cybium 10:121-134 . Lee, G. (1937). Oral gestation in the marine catfish, Galeichthys feliceps. Copeia 1937 (1):49-56. LeGrande, W. H. (1980). The chromosome complement of Arius felis (Siluriformes, Ariidae). Japan. 238 ~ Ichthyol. 27:82-84 . LeGrande, W.H . (1981) . Chromosomal evolution in North American catfishes (Siluriformes: Ictaluridae) with particular emphasis on the madtoms, Noturus. Copeia 1976:388-391. Liem, K.F. & Stewart, D.J. (1976). Evolution of the scale eating fishes of Lake Tanganyika: A generic revision with a description of a new species. In: Fish Physiology, Volume Bioenergetics and Growth, W.S. Hoar, ~ & D.J. Randall (eds . ), Academic Press, New York, pp.71-119. Lissman, H.W. & Machin, K.E. (1963). Electric receptors in a non-electric frsh (Clarias). Nature, Lond. 199:88-89. Livingston, M. E. specializations of Morphological (1987). five New Zealand relation to feeding behaviour . and sensory flatfish species, in Fish Biol . 31:775-795. ~ Loir, M., Cauty, C., Planquette, P. & La Bail, P.-Y. (1989). Comparative study of the male reproductive tract in 7 families of South American catfishes. Aguat. Living Resour. 2 (1):45- 56. L0vtrup, S. (1974). Epigenetics. A treatise .Q!! theoretical biology, John Wiley & Sons, London . 547pp. L0vtrup, S. (1984). Ontogeny and phylogeny from an epigenetic point of view. In: Human development, W. Edelstein, J.A. Meacham & H. Sinclair (eds.). S. Karger AG, Basel. pp.249-261. MacArthur, R.H. (1972) . Geographical Ecology. Harper & Row, New York. 269pp. MacArthur, R.H . & Wilson, E.O. (1967) . The theory of island biogeography, Princeton Univ. Mon. 239 ~ Biol. 1:1-203. MacKinnon, J.C. (1972). Summer storage of energy and its use for winter metabolism and gonad maturation in American plaice (Hippoglossoides platessoides). L Fish. Res. Bd. Can. 29:17491759. Mansueti, A.S. & Hardy, J.D.Jr. (1967). Ariidae sea-catfish. In: Development of fishes of the Chesapeake Bay region. An atlas of ~ larva and juvenile stages. Part 1....:.., E. E. Deubler Jr. (ed.), Natural Resources Institute, University of Maryland. pp.155-157. Marais, J.F.K. (1981). Seasonal abundance, distribution, and catch per unit effort using gill-nets, of fishes in the Sundays estuary. Afr. L ~ zool. 16 (3):144-150. Marais, J.F.K. (1983a.). Seasonal abundance, distribution and catch per unit effort of fishes in the Krom estuary, South Africa. Marais, Afr. L ~ J.F.K. zool. 18 (2):96-102. (1983b.). Fish abundance and distribution in the Gamtoos estuary with notes on the effect of floods. L ~Afr. Zool. 18 (2) :103-109. Marais, J.F.K. (1984). Feeding ecology of major carnivorous fish from four eastern Cape estuaries. ~ Afr. L zool. 19 (3):210-223 . Marais, J.F.K. & Baird, D. (1980). Seasonal abundance, distribution and catch per unit effort of fishes in the Swartkops estuary. ~ Afr. L Zool. 15:66-71. Marais, J.F.K. & Venter, D.J.L. (1987). Changes in body composition of sea-catfish associated with growth and reproduction. Symposium, Poster C-67, Stellenbosch. 6th Abstracts SANCOR document. 240 National Oceanographic of and papers posters, Maurer, B.A. (1985). On the evolutionary roles of interspecific competition. Oikos 45:300-302. Maynard Smith, J. (1977). Parental investment: A prospective analysis . An i m. Behav. 25:1-9. Mayr, E. (1970). Populations, Species and Evolution. An abridgement of Animal Species and Evolution. Belknap Press of Harvard University Press, Massachusetts. Mojumder, P. & Dan, S . S. (1979). Studies on food and feeding habits of catfish Tachysurus tenuispinis (Day) . Ind. Fish. ~ 26 (1/2):115-124. Menon, N.G. (1984). Observations on the intraovarian ova of a few tachysurids from Indian waters. Indian J . Fish. 31 (2):250256. Merriman, D. (1940). Morphological and embryological studies in two species of marine catfish, Bagre marinus and Galeichthys felis. Zoologica 25 (13):221-248 . & Rojas-Beltran, R. (1985). Mise en evidence de doubles cycles annuels de croissance sur le squelette de quelques teleosteens de Guyane . Bull Soc. Zool. Meunier, F.J., Lecomte, F. Fr. 110 (3):285-289 . Mishima, M. & Tanji, S. (1985). Fecundidade e incuba9ao dos Bagres marinhos (Osteichthyes, Ariidae) do complexo estuarino Lagunar de Cananeia (25 S, 48 W). Mojumder, P . & Dan, S.S. ~ Inst. Pesca 12 (2):77-85. (1979). Studies of food and feeding harits of catfish Tachysurus tenuispinus (Day). Ind. (Cochin) . 26 (1/2):115-124 . 241 ~ Fish. Moyle, P.B., Herbold, B. Daniels, & R.A. (1982). Resource parti tioning in a non-coevol ved assemblage of estuarine fishes. In: ..'JU.. Gutshop washington Sea G.M. Cailliet Grant Program, C.A. & WSG-WO Siemenstad 82-2, (eds.). University of Washington, Washington. pp.178-189. Mrowka, W. (1984). Is the parental-care behaviour of Aeguidens paraguayensis (Cichlidae) optimal? Behav. 89 (1/2): 128-146. Mrowka, W. (1985). Brood deprivation and the control mouthbrooding in Pseudocrenilabrus multicolor (Cichlidae). of ~ Tierpsychol 67:34-44. & Schierwater, B. Mrowka, W. mouthbrooding in a cichlid (1988). fish. Energy expendature for Behav. Ecol. Sociobiol. 22: 161-164. Muncy, R.J. & Wingo, W.M. (1983). Species profiles: Life histories and environmental requirements of coastal fishes and invertebrates (Gulf of Mexico) - sea catfish and gaff topsail catfish. Division of Biological Services, Fish and Wildlife Service, FWS/OBS-82/11.5, U. S. Dept. Interior, Washington . Nepgen, C.S.De Pachymetopon V. blochii argyrozona (Val.) ~ (1977). and biology the of the silverfish hottentot Argyrozona along the Cape south-west coast. Sea Fish. Brch. Nikolsky, G.V. (Val.) The ~ Investl. Afr. 105:1-35. (1963). The Ecology of Fishes. Academic Press, London and New York. 352pp. Nikolsky, G. V. (1969). Theory of Oliver & Boyd, Edinburgh. 323pp. 242 fish population dynamics. Noakes, D.L.G., Skulason, S. & Snorrason, S.S. (1989) . Alternative life-history styles in salmonine fishes with emphasis on arctic c harr, Salvelinus alpinus. In: Alternative life-history styles of animals, M. N. Bruton Academic Publishers, Dordrecht. pp.329-346. (ed.), Kluwer Oppenheimer, J . R . (1970). Mouthbreeding in fishes. Anim. Behav. 18:493-503. Osse, J .W.M . , Muller, M. & Van Leeuwen, J.L . (1985). The in fish. In: suction feeding Vertebrate analysis of Morphology, (Duncker .f!. Fleisher, eds. ) , Fortschritte der Zoologie, Band lQ, Gustav Fisher Verlag, Stuttgart. pp.217221. Otte, D.P. theory. (1979). Historical development of sexual selection In: Sexual selection and reproductive competition in insects, M.S. Blum & N. A. Blum (eds.). Academic Press, New York. pp.1-18. Paine, M.D. & Balon, E.K. (1984). Early development of the northern logperch, Perc ina caprodes semifasciata, according to the theory of saltatory ontogeny. Env . BioI. Fish. 11 (3):173-190. Paterson, H.E . H. (1978). More evidence against speciation by reinforcement. Afr. pauly, D. ~ (1980). ~ Sci. 74:369-371. On the interrelationships between natural mortali ty, growth parameter, and mean environmental temperature in 175 fish stocks. pauly, D. and ~ Cons. Int. Explor. Mer. 39 (2)1175-192. (1981) . The relationship between gill surface area growth performance in fish: a generalization of von Bertalanffy's theory of growth. Meeresforschung 28:251-282. 243 Pauly, D. (1983) . Some simple methods for the assessment of tropical fish stocks. FAO Fish. Tech. Pauly, D. & Thia-Eng, C. ~ (234):52pp. ( 1988) . The overfishing o f marine resources: Socioeconomic background in Southeast Asia. AMB I O 17 (3):200-206. Payne, A.I.L. (1977) . Stock differentiation and growth of the southern african kingklip Genypterus capensis. Sea Fish. Brch. Peres, J.M. ~ Investl. ~ Afr. 113:1-30. (1982) . General Features Assemblages in Pelagial and Benthal. of Organismic In: Marine Ecology. o. Kinne (ed . ) . Volume 5, Part 1. John Wiley & Sons, New York. 642pp . Pianka, E . R . (1970). On r- and K-selection. Amer. Nat . 104:592597. Pianka, E.R. (1972). rand K selection or band d selection? Amer. Nat. 106 (951) : 581-588. Pfeiffer, W. (1962). The fright reaction of fish. BioI . Rev. 37: 495-511. Phillips, A.M.Jr. (1969) . Nutrition, digestion and energy utilizat i on . . In: Fish physiology Volume 1. . W.S. Hoar & D.J. Randall (eds.). Academic Press, London . pp . 391-432. Pierce. R.J., Wissing, T.E : , Jaworski, J.G . , Givens, R.N. & Megrey, B . A . (1980). Energy storage and utilization patterns of gizzard shad in Acton Lake, Ohio. Trans . Amer. Fish . Soc. 109:611-616. Pitche r, T . J . & Hart, P.J . B. (1982) . Fisheries Ecology . Croom Helm Ltd., London . 414pp . 244 Pope, J.G. (1983). An investigation of the accuracy of virtual population analysis using cohort analysis. In: Key Papers on Fish Populations. D. H. Cushing (ed.). IRL Press Ltd., Oxford. pp.292-301. Price, P.W. (1975). Insect Ecology. John Wiley & Sons, New York. 514pp. Protasov, V. R . (1970). Vision and near orientation of fish. Academy of Sciences of the USSR, 17 5pp. Israel Program for Scientific Translations, Jerusalem. Pulfrich, A. & Griffiths, C.L. (1988). Growth, sexual maturity and reproduction in the hottentot Pachymetopon blochii (Val.). Afr . ..L.. mar. Sci. 7: 25-36 . ~ Punt, A. (1989). PC-VONBERT User's Guide. Department of Applied Mathematics, University of Cape Town, Cape Town. Internal SANCOR Report. 13pp. Punt, A. & Hughes, G. (1989). PC-YIELD II User's Guide. Benguela Ecol. Progrm. ~ ~ Afr. 18: 60pp. Radtke, R.L. (1987). Age and growth information available from the otoliths of the Hawaiian snapper, Pristipomoides filamentosus. Coral reefs 6:19-25. Radtke, R.L. & Targett, T.E. (1984). Rhythmic structural and chemical patterns in otoliths of the Antarctic fish Notothenia larseni: Their application to age determination. Polar Biol. 3:203-210. Rauck, G. ( 1976). A technique of sawing thin slices out of otoliths. Ber. dt. wiss. Kommn. Meeresforsch . 24:339-341. 245 Regan, C.T. (1911) . The classification of the f i shes of the order Ostariophysi. 2 . Siluroidea . Ann. MlliL.. nat. Hist . 8 (8) :553-577. Reis, E.G. (1986a) . Reproduct i on and Feeding Habits of the Marine Catfish Netuma barba (Siluriformes, Ariidae) in the Estuary of Lagoa Dos Patos, Brazil. Atlantica, Rio Grande 8: 35-55. Reis, E.G. (1986b). Age and growth of the marine catfish, Netuma barba (Siluriformes, Ariidae), in the estuary of the Patos Lagoon (Brasil). Fish. Bull. 84 (3) : 679-686. Reznick, D. (1985). Costs of reproduction: an evaluation of the impirical evidence. Oikos 44:257-267. Ribbink, A.J.R. (1987). African Lakes and conservation scenarios and suggestions. Env. their fishes: Biol. Fish 19 (1):3-26. Rice, S.D. & Stokes, R. M. wastes in the eggs and gairdneri Richardson. In: J . H.S. Blaxter, (ed.), New (1974). Metabolism of nitrogenous alevins of rainbow trout, Salmo The early life history of fish. York. Springer-Verlag. pp.325-337. Ricker, W.E . (1975) . Computation and interpretation of biological statistics of fish populations. Bull. Fish . Res. Bd. Can. 19111-382. Ricker, W.E. (1981). Changes in the average size and age of Pacific salmon . Can. ~ Fish. Aguat . Sci. 38 (12):1636-1656 . Rikhter, V.A. & Efanov, V.A. (1977). One of the approaches of estimating natural mortality of fish populations. Trudy Atlant. NIRO 73:77-85. 246 Rimmer, M.A. (1985a). Reproductive cycle of the fork-tailed catfish Arius graeffei Kner & Steindachner (Pisces : Ariidae) from the Clarence River, New South Wales. Aus. ~ Mar . Freshw. Res. 36:23-32. Rimmer, M.A. (1985b) . Early development and buccal incubation in the fork-tailed catfish Arius graeffei Kner & Steindachner (Pisces: Ariidae) from the Clarence River, New South Wales. Aust. ~ Mar. Freshw . Res. 36:405-411. Rimmer, M.A. & Merrick, J.R. (1983). A review of reproduction in the fork-tailed catfishes (Ariidae). Proc. Linn. Soc. ~ 107 (1):41-50 . ~ Roberts, T.R. In: (1973). Interrelationships of ostariophysians . Interrelationships of fishes. & C. Patterson (eds.) Zool. P.H. Greenwood, R.S. Miles Linn. Soc. 53:373-395 . ~ Robson, D.S. & Chapman, D.G . (1961). Catch curves and mortality rates. Trans . Am . Fish. Soc. 90:181-189. Roff, D.A. (1984). The evolution of life history parameters in teleosts. Can. Ross, S.T . assemblages: ~ (1986). A Aguat . Sci. 41:989-1000. Resource review of partitioning field studies . in fish Copeia 1986 (2):352-388. Roughgarden, J. (1983). Competition and theory in community ecology . Am. Nat. 122:583-601. Russel, B.C. (1982). The food and feeding habits of rocky reef fish of north-eastern New Zealand. Res. 17:121-145. 24 7 ~ Zeal. ~ Mar . Freshwater Sargent, R.C. & Gross, M.R. (1986) . Williams' principle: An explanation o f parental care in teleost fishes. In: The behaviour of teleost fishes, T.J . Pitcher (ed.). Croom Helm, London & Sydney. pp.275-293 . Schaefer, M. B. (1954) . Some aspects of the dynamics of populations important to the management of commercial marine fisheries. Bull. Inter-Amer . Trop. Tuna Comm. 1:27-56 . Schneppenheim, R. & Freytag, G. (1980). Age determination by staining otoliths of Notothenia rossii marmorata with ninhydrin. Cybium 3e s~rie (8):13-15. Schnute, J. (1981). A versatile growth model with statistically stable parameters. Can. ~ Fish. Aguat. Sci. 3811128-1140. Schoener, T. W. (1970). Nonsynchronous spatial lizards in patchy habitats. Ecology 51:408-418. Schoener, T.W. (1974). overlap of Resource partitioning in ecological communities. Science 185:27-39. Schoener, T.W. (1983). Field experiments competition. Am. Nat. 122:240-285. on interspecific Shchepkin, V.Ya. (1971a). Dynamics of lipid composition of the scorpionfish [Scorpaena porcus (L.)] in connection with maturation and spawning. ~ Ichthyol. 11:262-267. Shchepkin, V.Ya. (1971b). The dynamics of lipid composition in the Black Sea horsemackerel [Trachurus mediterraneus ponticus (Aleev)] in relation to maturation of the gonads and spawning. ~ Ichthyol. 11:587-591. Shelden, F . F. (1937). Osteology, myology, and evolution of the nematognath pelvic girdle. Ann. ~ ~ Sci. 37:1-96 . 248 probable Acad. Shuster, S.M. (1989). Male alternative reproductive strategies in a marine isopod crustacean (Paracerceis sculpta): the use of genetic markers to measure differences success among a-, Sih, A. and y-males . Evolution 43 (8):1683-1698. ~- (1987) . in fertilization Predators and Prey Evolutionary and Ecological Overview. Lifestyles: In: An Predation: Direct and Indirect Impacts QQ Aguatic Communities. W.C . Kerfoot & A. Sih (eds.). University Press of New England, Hanover. pp.203-224 . Silas, E.G., Parameswaran Pillai, P., Dhulked, C. Muthiah & Syda Rao, G. (1980) . Purse seine fishery - Imperative need for regulation. Mar. Fish. Infor, Serv o T ~ E Ser. (24):1-9. Sinclair, M. (1988). ' Marine populations. An essay on population regulation and speciation, Washington Sea Grant Program, University of Washington Press, Seattle. 252pp. Singh, V.D. & Rege, M.S. (1968). Observations on age and growth of Tachysurus sona (Ham.) . .J:..:.. Bombay Nat. Hist. Soc. 65 (1): 7587. Sissenwine, M.P. (1984). Why do fish populations vary? In: Exploitation of marine fish communities. R . M. May (ed.) . Dahl em Konferenzen, Springer-Verlag, Berlin. pp.59-94. Smale, M.J. (1983). Resource partitioning by six top predatory teleosts in eastern Cape coastal waters (South Africa). Unpublished Ph. D. thesis, Rhodes Uni versi ty, Grahamstown, South Africa. 284pp. Smalo, M. J . & Buxton, C.D . (1985). Aspects of the recreational ski-boat fishery off the eastern Cape, South Africa. .J:..:.. Mar. Sci. 3:131-144. 249 ~ Afr . Smale, M.J. Buxton, & C.D. (1989). The subtidal gully fish community of the Eastern Cape and the role of this habitat as a nursery area. Smith, J.L.B. Afr. ~ (1961). zool. 24 (1):58-67. ~ The .§.§ fishes of southern Africa. Central News Agency Ltd., Cape Town. 580pp. Smith, M.M. & Heemstra, P.H. (eds.). (1986). Smith's Sea Fishes (First Edition). MacMillan, South Africa: 1047pp. (1957). Early development and hatching. In: Smith, S. Physiology of fishes. M.E. Brown, (ed.), Volume I. New York, Academic Press. pp.323-359. Smith, T.D. (1988). Stock assessment methods: The first fifty years. In: Fish population dynamics. J.A. Gulland (ed.). John Wiley & Sons, New York. pp.1-33. Summerfelt, R.C. (1987). Preface. In: Age and growth of fish. R.C. Summerfelt & G.E. Hall (eds.), Iowa State Univ. Press, pp.xii-xiv. Svensson, reversed I. (1988). pipefish Reproductive species costs (Syngnathidae). in ~ two sex-role Anim. Ecol. 57:929-942. Sych, R. (1974). The sources of errors in ageing fish and consideration of the proofs of reliability. In: Ageing in fish. T . B. Bagenal (ed.), Unwin Brothers Ltd . , Surrey, England. pp . 78-86. Tavolga, W.N. (1962). Mechanisms of sound production in the ariid catfishes Galeichthys and Bagre. Bull. Amer. Mus . Nat. Hist . 124 (3):5-30 . Tavolga, W. N. (1971). Acoustic orientation in the sea catfish, Galeichthys felis. Ann.~ ~ Acad. Sci. 188:80-97. 250 Taylor, W. R. (1986). Family No. 59: Ariidae . In : Smith's sea f i s h es , M.M. Smi t h & P.C. Heemstra ( eds.). Macmi l lan South Africa (Pty) Ltd., Johannes burg. pp . 211-213. Taylor, W.R . & Menezes, N. A. (1977 ) . Ariidae . In: FAO species identification sheets f or fishery purposes. Western Central Atlantic (fishing area 111, 1: Fischer, W., Bianchi, G. & Scott, W. B. (Eds . ). Rome, FAO of the UN . Taylor, W.R . & Van Dyke, G. (1981). Ariidae . In: FAO species identification sheets for fishery purposes . Eastern Central Atlanti c (fishing area li.... .!1. i n part), 1: Fi scher, W., Bianchi, G. & Scott, W.B . (Eds.). Canada funds-in-trust, Ottowa, Department of Fisheries and Oceans, Canada, by arrangement with the FAO of the UN . Taylor, W.R. & Van Dyke, G. C. (1985) . Revised procedures for staining and clearing small fishes and other vertebrates for bone and cartilage study. Cybium 9 (2)1107-119. Thresher, R.E . (1984) . Reproduction in reef Publications, Neptune City. 399pp. fishes. T . F . H. Thys van den Audenaerde, D.F.E. (1970). The paternal mouthbrooding habit of Tilapia (Coptodon) discolor and its special significance. Rev. Zool . Bot. Afr . 82 (3/4):285-300. Tilak, R. Tilak, R. The comparative morphology of the (1965). osteocranium and the Weberian apparatus of Tachysuridae (Pisces: Siluroidei). ~ Zool . 146:150-174. (1971). A study of the osteocranium, Weberian Apparatus and the girdles of Chaca chaca (Hamilton): Family Chacidae, Siluroidei. Zool. Anz., Leipzig 186 (5/6):417-435 . 251 Tilney, R.L. & Hecht, T. (1990). The food and feeding habits of two co-occurring marine catfish Galeichthys feliceps and ~ ater (Osteichthyes: Ariidae) along the south east coast of South Africa. ~ Zool. Lond . 221:171-193. Tobor, J.G. (1969). Species of the Nigerianariid catfishes, their taxonomy, distribution and preliminary observations of the biology of one of them. Bull. Inst. fr o Afr. noire 31 Ser. A (2)1643-658. Turner, G. (1986). Teleost mating systems and strategies. In: The behaviour of teleost fishes, T. J. Pitcher (ed.) . Croom Helm, London & Sydney. pp.253-274. Van Leeuwen, J.L. & Muller, M. (1985). Prey capture in In: Vertebrate Morphology, Duncker & Fleisher Fortschritte der Zoologie, Band 30 Gustav Fisher fish. (eds.). Verlag, Stuttgart. pp.229-232. Van der Elst, R.P. (1981). A guide to the common of southern Africa. C. Struik, Cape Town. 367pp. Van der Elst, R . P . ~ fishes (1985). An appraisal of the new fishing laws. Ski Scene. 9:26-32. Vetter, E.F. (1988). Estimation of natural mortality in fish stocks. A. review. Fish Bull . 86:25-43. & Brothers, E.B. (1982). Age and growth of the fallfish Semotilus corporalis with daily otolith increments as a method of annulus verification. Can. ~ Zool. 60:2543-2548. Victor, B.C. Wallace, J.H. (1975). The estuarine fish of the east coast of South Africa. 1. Species composition and length distribution in the estuarine and marine environments . 2. Seasonal abundance and migrations. Invest. ~ oceanogr. Res. Inst. 40: 1-72. 252 Wallace, J.H. & Schleyer, M.H. (1979). Age determination in two important species of South African angling fishes. The kob (Argyrosomus hololepidotus Lacep.) and (Pomadasys commersonni Lacep.). Trans. the spotted grunter Soc. ~ ~ Afr. 44, Part 1:15-26. Warburton, K. (1978). Age and growth determination in a marine catfish using an otolith check techique. J. Fish Biol. 13:429434. Ward, J.W. (1957). The reproduction and early development of the sea catfish, Galeichthys felis, in the Biloxi (Mississippi) Bay. Copeia 1957 (4):295-298. Ware, D.M. natural (1975). Relation between egg mortality of larval fish. J. size, Fish. growth, Res. Bd. and Can. 32:2503-2512. Weatherly, A.H. growth. ~ (1976). Factors affecting maximation of fish Fish. Res. Bd. Can. 33:1046-1058. & Shine, R . the evolution of male parental care. ~ 631. Werren, J.H., Gross, M.R. (1980) . Paternity and theor. biol. 82:619- Wheeler, A. & Baddokwaya, A. (1981). The generic nomenclature of the marine catfishes usually referred to the genus Arius (Osteichthyes-Siluriformes) . Whitfield, A.K. Nat. Hist. 15:769-773. ~ (1980). A quantitative study of the trophic relationships within the fish Community of the Mhlanga Estuary, South Africa. Estuar. Coast. Mar. Sc. 10:417-435. Whitfield, A.K. & Blaber, S.J.M. (1978). Scale-eating habits of the marine teleost Terapon jarbua (Forskal). 12: 61-70. 253 ~ Fish Biol. Whitfield, A . K., Blaber, S.J.M . & Cyrus, D.P. (1981). Salinity ranges of some southern African fish species occurring in estuaries. Afr. ~ zool . 16 (3) : 152-155 . ~ Wild, A . & Foreman, T . J . (1980). The relationship between otolith increments and time for yellowfin and skipjack tuna marked with tetracycline . Inter-Am . Trop . Tuna Comm. 17:509541. Williams, G. C . (1966). Natural selection, the costs of reproduction, and a refinement of Lack's Principle. Amer. Nat. 100:687-690. Williams T. & Bedford, B . C . age determination. In: (1974). The use of otoliths for Ageing of fish. T.B . Bagenal (ed.) . Unwin Brothers, Surrey, England. pp.114-123. Witte, F. & Goudswaard, P . C. (1985). Prospects of the haplochromine fishery in southern Lake Victoria. In: CIFA report of the third session of the sub-committee for the development and management of the fisheries of Lake Victoria. Jinja, Uganda, 4-5 October 1984 . Fish . Rep. 335, FAO, Rome . pp.81-88. Wongratana, T . , Bathia, U. & Taylor, R . (1974). Ariidae. In: FAO species identification sheets for fishery purposes. Eastern Indian Ocean (fishing area .21.l and Western Central Pacific (fishing 211. ~ 1: Fischer, W & Whitehead, P.J.P. (Eds . ). Rome, FAO of the UN. Yamamoto, M. (1982). Comparative morphology of the peripheral olfactory organ in te1eosts . In: Chemoreception in fishes. Developments in aguaculture and fisheries science, volume T.J. Hara (ed.), Elsevier Scientific Oxford. pp.39-59 . 254 Publishing ~, Company, Yanez-Arancibia, A. (1977) . Biological and ecological studies and perspectives of culture of sea catfish Galeichthys caerulescens (Gunther) in the coastal lagoons on the Pacific coast of Mexico. Proceedings of the eighth annual meeting of the World Mariculture Society. 8: 117-133. Yanez-Arancibia, A. & Lara-Dominguez, A.L . (1988). Ecology of three sea catfishes (Ariidae) in a tropical coastal ecosystem Southern Gulf of Mexico. Mar. Ecol. Prog. Ser. 49:216-230. Za1e, A.V. (1987). Growth, survival, and foraging abilities of early life history stages of blue tilapia, Oreochromis aureus, and largemouth bass, Micropterus salmoides. Env. Biol. Fish. 20 (2):113-128. Zaret, T.M. & Rand, A . S . (1971). Competition in tropical stream fishes: Support for the competitive Ecology 52:336-342. 255 exclusion principle. APPENDIX I Stomach content analyses Appendix Ia: ","0 WT TOTAL 65.67 350.70 TOTAL 0,44 0.24 0.06 0.38 0.02 0 .22 0.25 0.06 0.19 5.43 1.00 1.16 1.37 3.80 0.91 0.25 0 .20 TOTAL 42.29 seEl:temdentalus 0.06 14.96 1.17 0.06 325.13 1.79 130.73 2.52 0.02 G. FEUCEPS: MARINE % ENERGY %F.O. E.I. ---------_.._---_........_-------------_._-------------------_..._------------CRUSTACEA AMPHIPODA Paramoera capensis Unidentified TOTAL ANOMURA Calianassa kraussi Unidentified BRACHYURA A\elc~u9 Gonael&)( ansulata Hymenosome orbiculate Mursla crisllmanus Nautilocorysles ocanata Ovallpes punctatus Phllyra punctata 0.45 3.82 84.01 0.04 0.75 46.87 18.85 2 .93 Unidentified chelae 1.17 6.94 Unidentified megaJopae 2.66 2.76 5.24 0.39 2.20 2.66 5.85 0.62 2.51 3.36 3.30 0.06 11.42 11.28 0.14 1.71 13.92 2.33 0.22 0.13 0.30 Qc;haetosloma capense 12.31 9.78 153.33 144.07 2:. capense 2.53 ' .26 3.63 10.59 4.73 4.50 TOTAL Synidotea hirlipes Unldenlified TOTAl MACRUAA Carldea Pallnura 0.25 18.16 3.98 0.25 3222.26 383.58 2,49 0.97 16.97 ~melnri 61.34 1.24 6.99 15.94 ThaumasloP!ax splrali$ ISOPODA 0.09 1,49 1.63 154.51 2.03 0.25 1.49 22.27 25.62 639.66 '.23 6.72 1.04 4.94 5.75 O.so 1.99 2.72 2.49 6.47 0.25 12.23 6.22 1.51 MYSIDACEA Unidentified STOMATOPODA Lyslo,quilia capensis ECHIURIDA TOTAl . proboscis TOTAL MOllUSCA CEPHALOPODA Oetopus~ Sepia !E.:. 0.71 0.06 0 .23 0.65 0.32 O.SO 25.97 34.01 25.87 969.66 3.98 12.21 1.02 2.74 0.25 2.98 GASTROPODA Unldenlified ~decip j ens Unidenllfled PQlYCHAETA TOTAL ERRANTlA Total Dlopalra cuprea Glycera corwoluta Marphysa macintoshi Trypanosyllis gemmullfera Unidentified SEDENTARLA Pherussa Total !E:- Sternaspsis $Cutaia Unidenlified 1.74 6.19 0.38 1.62 0.58 5." 0." 6.97 2.24 2.66 16.69 5.10 90.72 32.01 20.45 7.22 699.59 118.71 0.06 0.26 0.06 0.32 4.40 11.59 0.43 0.86 3.74 6.10 30.52 14.43 0.25 0.25 0.25 0.25 14.43 58.71 13.24 0.58 10.30 0.71 1.31 52.76 4.84 1' .89 0.06 0.31 PELECYPODA 0.65 20.88 0.75 12.19 2.'19 SIPUNCUU DA Unidentified 0.25 256 121.87 Appendix Ia contd. %NO WT 0.13 0.58 1.55 0.91 5.4<1 0.1 4 0.65 5.30 % ENERGY %F.O. E.I. SPERMATOPHYTA Unidentified detritus TElEOSTEI Scales TOTAl Unidentified 1.01 25.00 10.76 32.06 34.06 CRUSTACEAN REMAINS UNIDENTIFIED REMAINS STOMACH NUMBER 402 EMPTY STOMACH NUMBER 58 SIZE RANGE (F.L) 232-36Omm MAXIMUM STOMACH INDEX 4.38 1.49 2.24 4.0'% : 257 5.06 2.53 Appendix Ib : --_.._..__..__......._---_.......----..•.......__ ._---.__._----..-..--_.-... _....._---------_.-.._...._._-_....__._-----.... _--wr %NO '*' ENERGY %F .O. E.!. .._.....-.....__........... ... _.. _-.... _....... _.... .....-.------_.._.._--..._-_.._-_.- ..._._.........__...._--..........._....._...._--_._---------..._G.ATER: MARINE ALGAE Unidentified 0.22 1.18 0.59 3.14 BRYOZOA Unidentified CRUSTACEA ANOMURA TOTAL 1.19 '1.47 10.19 Unidentified 0.80 1.28 17.92 92.19 7.35 1.77 9.97 BRACHYURA TOTAL Atel!£lclus see:lemdenlalus Cryptodrorniopsls spon91osa ~ 0.69 0.30 1.59 ~ Dromidlopsls cornuta Goneplax atlQulata Hymenosoma orbiculare Matuta !e:. Macropodia falcitera formosa Parapllumnus pi sifer Phityra punctata Portumnus blgultatus Unidentified chelae 0.10 1.59 1.53 2.65 0.68 1.11 2.08 2.60 1.83 0.55 19.33 5.15 2. 40 3.98 Unidentlfled crabs 9.58 23.92 M:. P l a9u' .E: 1.99 Galathea disperse l a~ ~ - megalop8e ISOPODA 0.69 0.50 0.30 0.20 1.29 4.16 6.06 3.53 33.83 2.43 40.78 1.96 1379.72 4.76 3.29 1.09 5.10 16.80 0.72 1.45 1.81 1.09 2.54 0.36 6.39 1.09 0.36 4.3 1 6.52 15.58 17.84 29.75 0.10 0 .36 Arcturella corniger 0.50 0.50 0.79 0.30 0 .10 0.10 0.40 0.25 1,45 1.09 3.26 2.17 0.36 0.99 3.96 1.35 9.74 0.30 0.50 0.10 0.20 0 .72 1.23 .!22!!!. metallice. 0.99 0.10 0 .04 NeastacJlla t(an gu illa 0.20 0.03 !p:' Paracillacea 0.10 0 .02 0.20 0.20 0. 10 0.10 0.10 0 .57 Cymodoce amplmons £. telrath ele £.~ £:.~ .£. val ida .£~ CymodocaJla !p:. Exospilaeroma antikraussi Exosphaeloma !p:. Haliophasma foveolata ~ !e:. ~glanuos SVnidotea hirtipes ~ ~ variegata Unidentified MACRURA Carldea TOTAL Penaelda PAUNURA Scynarus ellsabethae ECHINODERMATA 12.10 0.65 0.32 0.68 0.57 0.01 0.16 0.36 0.40 1.15 2.12 5.29 0.40 2.97 1.68 15. t4 5 12.64 0.36 1.81 0.36 0.72 0.36 0.72 0.72 0.36 0.36 0 .12 0.01 9.03 12.44 11.29 42.35 0.36 0.36 2.17 11 .23 1.09 0.72 0.72 0. 10 7.33 17.25 16.85 27.56 0.36 3.62 0.10 £.~ 22.97 11.89 2.53 Anlarcturus kladoehorous ~!e:. TOTAL 0.30 0.50 0.69 0.30 3.79 3.37 4.27 3.87 1.75 0.36 17.03 22.45 22.75 t .45 95.79 2.75 4.60 88.08 HOlOTHUROIDEA Unidentified 1.45 ECHIURIDA Ochaelosloma caeense HYDROZOA TOTAL Gymnanglum arcuatum Sertularella !e:. 5.88 0.10 5.10 0.72 0.09 0.36 0.01 0.36 258 29.96 Appendi x Ib contd. %NO wr % ENERGY % F.O. E.t. ..- ........----------._.--- -----.- ...-------_._------------------------_._--_._---------------------------------_..... _-------_ ....._-._.... _....._-_._-_..._-MOLLUSCA CEPHALOPODA TOTAL 6.14 36.11 Total 308 0.10 30.44 2.78 14.57 8.70 !£:. Octopus !2:. Loligo Sepia !2:. GASTROPODA 0.20 Total ~!?:. Nassa scopularcus Unidentified PELECYPOOA Unidentified POLYCHAETA EARANTIA Eunice !2:. TOTAL Total Eupnrosine capen!>ls lys i d i ce~ Platynereis dumerill! Unidentified SEDENTARIA Tolal FlabeUJgera affinfs Pherussa !2:. Unidentified 2.18 1.19 0.10 0.89 12.32 15.04 U)O 13.33 0.36 7.61 0.72 5.97 6.87 6.15 5.50 6.33 2.75 0,37 0.36 3.2tl 0.89 0.20 3.26 13.38 0.10 29.98 22.91 2.97 11.00 200.56 0.10 15.49 11.84 33.33 33.33 0.36 0.36 0.36 0.36 0.10 0.05 0 .10 0.\0 10.60 0.16 2.36 0.10 7.07 0 .16 1.59 0.69 3. 15 3.76 0.'10 2.28 '.26 18.36 0.30 0. 22 l.08 0.10 l.29 2.58 9. 11 9.03 0.22 1.08 4.71 l.08 9.07 12.35 12.55 11.76 0.12 19.81 42.27 17.38 518.39 394.65 29.35 3.65 8.23 0.36 5.07 30.01 2.17 SIPUNCUUOA Themisle !EE.:. TELEOSTEI TOTAl G0811DAE Unidentified OPHICHTHYIDAE Unidentified TELEOST REMAINS Scales CRUSTACEAN REMAINS UNIDENTIFIED REMAINS 1.45 8.01 STOMACH NUMBER 285 EMPTY STOMACH NUMBER 30 SIZE RANGE (F.L) 195-322mm MAXIMUM STOMACH INDEX '1.05% 259 5.88 47.14 Appendix Ie: %NO wr TOTAL 83.92 0.51 TOTAL OJ<) 0.01 0.15 0.1 5 0.01 2.06 0,01 4. 17 G.FELICEPS: T.B. DAVIE - MARINE % ENERGY %F .O. E.!. ----_ ..--_.. -._........-----....-.-..... _-_.. -....... -...................--.._-_ ......._..............._.•......._----_..•..•. _---.------...........-•.•............._--_.._--------- CRUSTACEA AMPH IPODA Cunicus profundus Unidentified ANOMURA TOTAL d. Clibansrlus lons llarsus Unidentified 8AACHYURA TOTAL Thaumasloplax splralis Unidentified chelae TOTAL 0.44 0.12 0,15 0.0< 0.08 0.29 95.83 0,24 4.02 4 .1 7 6.25 LOO 25.15 2.08 6.25 5.59 12.50 3.69 8.33 0.29 0.27 0.18 0.09 1.49 52.08 2.08 52.08 77.52 1.61 1.32 69.84 30.74 79.24 0.11 Mesopodopsis afrieanus 0 . 15 0.01 Unidentified 79.09 0.11 4.68 4.62 76.48 35.42 2708.65 0.15 0.12 3.77 2.08 7.85 1.23 5.85 46.00 268.89 1.75 0.02 0.73 16.67 12.11 9.51 8.63 0.88 0.11 1.85 37.50 69.55 MYSIDACEA ECHIURIDA Ochaetostoma capense MQUUSCA CEPHALOPODA MUCOUS POLYCHAETA ERRANTI,A TELEOSTEI TOTAL Seal.. Spines CRUSTACEAN REMAINS 0.01 4.16 UNIDENTIFIED REMAINS 0.29 35,42 STOMACH NUMBER 48 EMPTY STOMACH NUMBER 0 SIZE RANGE (FL) 47· 145mm 260 Appendix rd: G. ATER JUVENI LES: SUB-nCAL GULLY: CRUSTACEA 1.67 TOTAl Gammarid ea 0." 0.65 Caprellidea 0.03 BRACHYUAA Unidentified TOTAL ISOPODA Clrolana !£:. Dynamenella huttonl .Q;~ E)(oSphaeroma %ENERGY .!f:. Gnathia africana Isoc.adus ollon MACRURA Macrobrac:hlum esuldons OSTRACODA Unidentified 97.50 92.68 24.39 3257.04 0.04 1.34 2.44 3.27 0.65 35.16 62 2197.42 !£:. .~O 2.44 0.01 2.44 0.'" 0.01 0.05 0.01 0.02 0.03 7.32 2 .44 2.44 2.44 .... 2.44 12.20 9.76 0.10 0.04 1.39 0.08 2.72 HYDROZOA Sartularia E.!. 33.41 0.36 Parisocladus perforatus %F .0. 100.00 0.01 Cymodocella pustulate ~ wr TOTAL AMPHIPOOA £:.~ %NO 0,01 2.44 .... 3.39 13.29 2.44 MOllUSCA GASTROPODA Orystele !e:. MUCOUS 0.01 1.\7 12.50 14.60 0.31 2.37 36.59 813.59 0.43 22.49 65.00 1481.60 POLYCHAETA ERRANTIA Unidentified STOMACH NUMBEA 41 EMPTY STOMACH NUMBER 1 SIZE RANGE (Fl) 61·12tmm 261 Appendix Ie: G.FELICEPS: KOWIE RIVER · JUVENILES _ ,. ENERGY %F .O. %NO. WT 40.84 tOO.96 89.13 0.21 0.01 1.09 6 .'13 74.58 0.32 4.40 6.11 70.18 15.86 7.82 E.!. ......................_......- ................. -.. -...---------_ .._--- ._... _---------_. __ ........_-------_._-_ .._---_._-----... .... _--_....._-_....._......_-----------_ . TOTAL CRUSTACEA AMPH IPODA Unidentified TOTAL ANOMURA Callianassa Kraussi ~ TOTAL BRACHYURA Clelstostoffill edwatdslJ Hymenosoma Ofbiculare Thllumaslof)lax spiraJis Unidentified chelae 3.26 40.22 7 .67 45.65 19.57 4 .61 22.37 14.07 3.27 1.78 30.43 150.12 54.27 3.00 0.43 4.16 2.27 22.83 51.78 0.87 1.62 aO.43 11.38 2.41 0 .1 4 0.21 0.21 0.11 0.11 4.07 0.06 0.10 0.05 0.05 1.04 1.09 1.09 1.09 11.96 2.47 3.54 0.39 0.58 8.70 10.87 1.61 0.64 0.75 0.21 0.72 0041 9.78 0.60 0.09 0.03 0.34 3.26 5.43 5.36 3.75 1.81 0.87 0.37 0.31 3.22 0.21 0.16 0.04 2.17 3.00 0.12 9.78 2.57 0.22 Phy1lodose madeirensis 2.14 0043 0.17 0.05 5043 Unidentified 2.44 8.70 TOTAL Gymodose .!ee;. S. afrieana £!!2!!!!! nuviolilis Exosphaeroma antikraussl E. hylocoetes E. truncalltelson Unidentified MACAURA TOTAL Macropetasma africanum Unidentified Carldea Unldentltted Penaelda MYSIDACEA TOTAL Mesopodopsls slabberi Unldenlified MOllUSCA GASTROPODA TOTAL PELECYPODA POlYCHAETA TOTAL 10.50 2930.27 2.17 0,84 ISOPODA ~!E£:. 43.48 3.22 72.86 49.25 3.26 2.17 3.99 1.11 2.17 0.50 6.52 5.43 \. 99 1.09 0.04 0.19 11.96 9.78 0.48 1." ERRANTlA SPERMATOPHYTA TOTAl 4.35 Delritus 0.01 Zoslera .!e:. 2.43 1." 52.63 13.22 6.95 0.54 0.11 0.11 0040 TElEOSTEI TOTAL Atherina breviceps CaHrogoblus agulhensfs . Uza.!e:. """ Spines Unidentified TUNICATA CRUSTACEAN REMAIN S 0.75 0.17 LtO ".00 1.09 30.43 9.78 3.26 15.22 9.78 112 EMPTY STOMACH NUMBER 20 MAXIMUM STOMACH INDEX 9.98% SIZE RANGE (F.l.) 35.87 7.61 5.96 16.95 UNIDENTIFIED REMS STOMACH NUMBER 0.18 A." 12.11 14.35 3.26 1.09 0.12 2.53 1.04 •.95 38.80 1.09 7.61 125·215mm 262 33.38 Appendix If: _---_........_...__ ....._-_ .....__ ._-_...... __..._... FISH RIVER: JUVEN ILES -_.._._-_._-_.._--_. __... ,,"0 TOTAL __...._-_._--_._...__......._ ...._------_.-_..... wr %ENERGY %F.O. E.1. 82.20 66.23 9.9 1 49.52 1.63 15.88 40.00 635.07 Calianass8. klaussi 1.03 3.51 985.50 0.08 0.04 40.32 0.65 24.44 Upogebia afrlcana 2.22 1.45 2.25 0.72 0.16 1.35 3.59 1.00 0.19 2.39 22.13 33.33 737.74 0.40 0.02 0.18 8.S9 1.81 7.08 0.72 6.36 0." 0.15 0.48 5.60 31.11 174.33 MYSIDACEA Unidentified OSTRACODA 5.84 0.34 1.65 Unidenlified 0.24 0.15 0.06 0.05 0.67 2.22 1.49 33.55 1.87 12.72 62.22 1045.53 %NO wr CRUSTACEA AMPHIPODA Unidentified ANOMURA TOTAL BRACHYURA Cleistostoma edY.tardsli Hymenosoma orbiculare Unidentified 11 . " 2.22 26.67 ISOPODA Unidentified TOTAL MACRURA Caridea Penaeida POLYCHAETA Unidentified TELEOSTEI Seal.. EMPTY STOMACH NO " SIZE RANGE (FLJ 4~ STOMACH NO -1 11.11 20.00 17.78 32.88 6.67 80m Appendix Ig: FISH RIVER: MATURE FISH - - ------_._% ENERGY %F.O. E.I. CRUSTACEA TOTAL 95.65 12.99 ANOMURA TOTAL 15.21 13.04 2.17 9.31 15.22 0.76 3.00 23.08 69.23 10.87 2.27 '.00 30.77 246.15 54.35 0.65 1.00 15.36 15.38 Callianassa kraussl ~ 89.13 ".00 '.33 0.98 46.15 38.46 2030.77 7." BRACHYURA Unidentified MACRURA Penaelda MYSIDACEA Unidentified SPERMATOPHYTA Zostera !I!:. 0.15 15.38 TELEOSTEI 4.35 Unidentified 2.31 UNIDENTIFIED REMAINS STOMACH NO EMPTY STOMACH NO SIZE RANGE (FLJ 10.00 23 10 _mm 263 44.00 15.36 676.92 Appendix Ih : G,FEUCEPS: MTATl RIVER · JUVENILES "NO wr -----------------------------------------_.._---------------------_...-------_..._-----_...... TOTAL CRUSTACEA _-_.... _..._... ..._.._-------------_.. ..-_..._---% ENERGY %F.O, E.I. _ , 86.84 9A5 3 .44 3.16 0.10 0.37 22.86 8.48 0.12 0.67 2.92 8.57 25.03 0.49 2.12 20.00 2.63 15.79 42.50 0.37 0.65 0.03 C.52 5.47 2.00 '.06 40.00 322.48 C.20 C.C3 0.10 2.83 0 .27 0.08 0.10 '.28 C.04 1.49 2.63 0 .04 0.01 2.83 0.04 0.02 0.10 2.83 0.27 0.14 0.66 22.86 15. 12 <174.49 6306.2 7 AMPH IPODA Unidentified ANOMURA Callanassa kraussll TOTAl BRACHYUAA Hymenosoma orbiculare Unidentified 0 . 12 [SOPODA Unidentlr!ed MACRUAA Carldea TOTAL MOLLUSCA GASTROPODA Asslminea bllasclala PELECYPODA Unidentified POLYCHAETA Unidentified SPERMATOPHYTA Zostera !E.:. TELEOSTEI TOTAl 90.43 C.CB 29.52 Catfr2Soblus multifaselalus 1.49 6.44 Unidentified 0.45 3.79 16.39 2.63 28.95 S"~ 89.90 24.25 63.06 100.00 ' C.69 UNIDENTIFIED REMAINS STOMACH NUMBER 38 EMPTY STOMACH NUMBER :; SIZE RANGE (fl) 175·214mm 264 100.00 36." 16.95 Appendix Ii: wr ..wr """a 0 .001 1.471 3.704 0.020 29.412 37.037 0.008 11.765 14.815 0.006 8.824 7.407 0.006 8.824 14.815 0.003 4.412 7.407 MOllUSCA GASTROPODA 0.002 2.941 3.704 MUCOUS 0.004 5.882 33.333 0.001 1.471 3.704 0.012 17.647 74.074 0,011 0.009 18.176 13.235 66.667 33.333 0.003 0.002 4.412 3.704 2.941 3.704 G.FEUCEPS: FREE EMBRYOS ALGAE Unidentified Remains TOTAl CRUSTACEA COPEPODA Unidentified ISOPOOA Unidentified remains EXOSKELETA Unidentified HYDROZOA Unidentified remains PORIFERA Spicules SANa SPERMATOPHYTA leafy remains Woody remains TELEOSTEI Unidentified rem~n s Yolk STOMACH NO 27 EMPTY STOMACH NO o SIZE RANGE (Fl) 3O·49mm 265 APPENDIX II Appendix IIa: G. FELICEPS: OTOLITH RADIUS AND OTOLITH LENGTB AT SIZE T-TEST, MALES VS FEMALES OTOLITH RADIUS MALES SIZE CLASS 250-259 260-269 270-279 280-289 290-299 300-309 310-319 MEAN OT . RAD SO 5 5 7 2 9 9 7 4.32 4.33 4.37 0.145 0.146 4.87 0. 5 37 4.81 0 . 373 320-329 11 14 4 3 350-359 360-369 MEAN N 330-339 340-349 FEMALES 0.194 4.95 O.31l1 5.25 5.22 5.38 0.268 0.424 0.576 6.12 0.527 5.67 0.612 N 3 2 3 1 5 8 5 16 15 15 7 3 SIG DIP OT. RAD SO t-TEST 4.47 0 . 157 -0 . 58517 4 .54 4.7 0.127 -0.92278 0.444 -0.27017 0.592986 0 0 . 319 0 . 284989 4.73 0.244 4.48 4.48 5.82 0.811391 0.31 0.325018 0.506 -0.02745 0.375 0.056592 0.443 0 . 549602 5.79 0.364 -0.1 7817 6.31 0.655 5.111 5.23 5 . 36 df (Y- */N- -) 6 5 8 I 12 15 10 25 27 17 8 OTOLITH LENGTH MALES SIZE CLASS 250-259 FEMALES MEAN N OT. LN. MEAN SO OT. LN. 0.282 0.428 0.38 1 3 2 3 10.19 0.244 9.95 10 . 63 0.071 t-TEST -0 . 23429 1. 029956 0.693 -0.37780 I 10.02 0 1.331244 I 5 8 5 16 15 15 7 3 10. 67 0.3 16 0.39 3 348 11.16 0.383 -0.31999 11.65 11. 74 12.28 12.44 12 .75 12.7 0.211 -0.23394 0.474 0.594 0.562 1. 029 0.46 0.153341 -0.68999 -0.06599 0. 184134 12 15 10 25 27 17 8 2 80-289 5 5 7 2 290-299 9 10.86 0.092 0.55 3 00-309 9 11 0.646 310-319 320-329 330-339 340-349 350-359 360-369 7 11 14 4 3 11.56 11.82 11.99 12.4 0.446 0.858 0 . 537 12.95 0.545 260 - 269 270-279 10.1 10.38 10 .35 10.17 SIG OlF SO N 0.515 266 df 6 5 8 (Y'"'*JN- -) Appendix IIa contd. : G. ATER: MEAN OTOLITH RADIUS AND OTOLITH LENGTH AT SIZE T-TEST, MALES VS FEMALES OTOLITH RADIUS MALES FEMALES MEAN SIZE CLASS 220-229 230-239 240-249 250-259 260-269 270-279 280-Z89 N 3 3 24 25 18 5 4 OT. RAD. SO N MEAN OT. RAD. SIC DIP SO t-TEST 4.39 0.354 4.36 0.417 0.37 6 5 4 . 31 0.228 0.109628 4.78 4.58 0.283 4.73 0.316 11 5.0 4 0.454 5.07 0.372 0.437 0.261 9 27 21 18 9 4 4.9 0.368 0.623029 -1.11272 0.562720 5.05 0.416 5 .4 0.458 5.3 5.2 290-299 300-310 310-319 5 .61 0.48 5.66 0.478 0.183 6.06 <Y- */N--) df -0.50988 7 27 34 25 30 23 t-TEST df 0.601981 OTOLITH LENGTH MALES SIZE CLASS 220-229 230-239 240-249 250-259 260-269 270-279 28D-Z89 290-299 300-310 310-319 N 3 3 24 25 18 5 4 MEAN OT. LN. FEMALES MEAN SO 9.67 0.494 10.13 10.47 0.748 10.85 0.405 0.369 11.4 11.5 11-42 0.579 0.476 0.637 N 6 5 11 9 27 21 18 9 4 SIC DIFF OT. LN. SO 9.91 0.488 10.51 10.88 11. 12 0.503 0.372 0.509 11.46 0.584 11.89 0.54 0.539 0.402 0.389 12.12 12.25 12.81 267 0.251720 - 0.07520 -0.10671 0.781158 0.077507 -0.73196 7 27 34 25 30 23 (Y -*/N--) Appendix IIa contd. : G. FELICEPS: MEAN WEICHT AT SIZE T-TE5T: MALES VS FEMALES MALES FEMALES SIG DIP SIZE CLASS N 16 15 14 13 26 21 30 40 56 29 14 5 250-259 260-269 270-2 79 280 - 289 290-299 300-309 310- 319 320-329 330-339 340-349 350-359 360-369 MEAN WT SD 251.4 18.94 278 30.71 314.3 356.2 33.83 401. 5 40.24 437.5 475.4 35.86 38.11 522 50.1 48.51 564.9 55.04 627.8 661.1 73.54 45.9 678.4 65.11 370-379 N 10 8 16 16 27 41 39 56 51 37 29 10 3 MEAN WT 238.9 268.9 SD 1:-TE8T 320.3 12. 83 19.09 30.55 321.4 26.15 1.450344 395 426.7 485.3 516 591.5 44.55 0.278996 0 . 521261 -0.42434 0.976036 0.417386 -0.254 6 6 639.4 41.35 45.97 52 68.13 73.8 698.7 63.85 -0.31744 -1.05271 725 . 9 36 . 49 - 0.85357 763.3 14.57 0.288356 -1.11574 df (Y- */N- -) 24 21 28 27 51 60 67 94 105 64 41 13 C. ATER: MEAN WEIGHT AT SIZE T-TEST: MALES VS FEMALES HALES FEMALES SIC DIP SIZE CLASS 200-209 210-219 220 - 229 230-239 240-249 250-259 260 - 269 270-279 280-289 290-299 300-309 310-319 320- 329 N 2 5 7 12 47 60 56 32 10 MEAN WT 175.6 182.1 SD N SD t-TEST df 4.65 6.91 2 7 196.3 37.4 239 22.4 263.8 292.2 322.6 24.46 11 30.56 22 27 60 69 44 25 341. 7 366.8 MEAN WT 37.54 44.4 43.3 13 2 165.5 21.92 0.647578 217 4 27.52 0.909791 30.69 53 . 65 34.58 30.85 46.58 41.23 43.58 1 . 218058 0.628912 -0.13611 241.3 279 324.9 361.5 395.9 434.2 476.4 492.6 268 66.79 21. 21 -1. 20203 -0.9 5684 7 17 56 80 81 90 77 (Y-*/N- - ) Appendix lIb: G. FELICEPS: MEAN OBSERVED LENGTH AT AGE T-TEST: MALES VS FEMALES FEMALE MALE srG DIP AGE N MEAN FL so N o 34 63 .1 Ilj.861 1 20 36 21 12 15 116.9 156 11.081 9.549 178.7 208.8 7.836 17.309 233.3 257.6 18 . 512 15.786 285.2 296.4 15.182 13 .823 311.9 11.657 316.9 329 . 7 12.438 8.986 329.8 336.9 340.4 8.773 4.051 10 14 27 19 17 19 20 6.539 9 324.9 324.4 333.3 342.3 349 350.8 6 . 726 6.2 9 346.9 9 351. 5 2 3 4 5 6 7 22 10 11 12 16 19 30 32 16 14 13 8 14 15 16 17 18 4 8 9 4 9 34 20 36 24 20 16 23 7 t-TEST SD MEAN FL o o 63.1 14 .861 116.9 154 11.081 8.978 0.457995 177 .6 9.781 0.208963 195.4 1.151833 216.6 14.551 13.923 243.1 20.277 267 15. 294.3 309 320.7 11. :.11 362.9 ~56 11.769 8.233 16.913 13.045 7.941 10.38 6.274 5.804 5.336 (Y-*/N--) df I. 432607 1.348264 1.468821 0.239117 0.466664 -0.63472 0.532090 0.702688 0.717615 -0.18687 0.268816 -0.12370 66 38 70 43 30 29 43 24 31 55 49 31 31 26 11 11 16 373 G. ATER: MEAN OBSERVED LENGTH AT AGE T-TEST: HALES VS FEMALES FEMALE MALE SIG DIF AGE N o 18 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 7 15 15 15 14 10 16 19 24 23 14 10 5 3 6 MEAN FL SO 80.3 10 .9 75 MEAN FL N 18 80.3 Ill. 9 6.669 o 149.3 13 .207 8 175.1 196.3 2D9.6 223 10.453 -0.82D66 -D.94942 10.563 12.9 -0.86013 -1.13816 10.746 10.893 9 . 126 7.146 7.818 5.88 7.253 -0.33084 0.890583 0.190398 -0.32587 -1.30096 -1.87873 -3 .18841 8.476 -2.D8355 7.014 3.668 -1.57516 -2.56912 198.9 219.6 241 . 5 11.973 9.153 14 8.189 6.123 4 254.9 257.6 8.155 8 . 723 264 268 4 .566 272.7 284.5 8.083 4 . 301 4.506 df D 7 6.669 14 . 727 13.846 16.36 8.477 10.975 t-TEST 11 111.9 140.2 165 185.7 250 251 SD 7 6 11 12 12 7 25 26 26 12 232 248.9 252.8 262.3 270.3 278.1 281 287.2 295 269 (Y- */N- -l 34 12 24 21 20 26 14 18 28 34 33 19 33 • 29 27 16 * * Appendix lIe: G. FELICEPS & G. ATER: BACK- CALCULATED MEAN FORK LENGTH AT AGE T-TEST : MALES VS. FEMALES G. FELICEPS AGE 0 2 3 4 5 6 7 8 9 MALES FEMALES MEAN MEAN N BeLe PL II II 77 .5 113 .4 16.28 SO 12.14 II 149.4 19.25 11 11 11 181.5 19.8 208.5 16.65 230.8 14.93 II 248.1 13.84 II 261. 3 275.3 15.57 16.22 288.8 16.58 10 10 N 16 16 16 16 16 16 16 16 16 15 BeLe FL SO t-TRST df 25 25 25 25 25 25 25 25 24 24 82.2 9.12 -0.564.42 120.3 -0.55943 152.1 15.21 17.18 177 .4 17. 8 1 0 .278326 200.8 221.5 16.64 14.13 0.590542 0.817074 238.9 13.49 0.859454 255.4 13.48 0.5 18537 272.2 13.2 0.261392 285 14.21 0.302307 -0.18922 SIG DIP (Y-* /N- -) G. ATER AGE 0 1 2 3 4 5 6 7 8 MALES FEMALES MEAN MEAN N BeLe FL 15 15 15 15 15 14 12 85.3 116 9 .02 10.45 141-8 166.6 12.71 14.58 186.9 14.58 16.21 12.46 10.86 II II 204.6 215.7 226.6 236 . 5 SO 10.97 N 10 10 10 10 10 10 9 SIG DlF SO t-TEST 8CLC FL 82.3 117.7 148 8.77 11.88 171.8 11.79 -0.48302 191 209 12.58 -0 . 36976 9.12 0.405097 -0.21665 -0. 61760 10.51 -0.39771 8.06 8 22,1.7 235.7 7.05 -0.66309 -1.09347 4 247.6 7.9 -1. 00746 271 df 23 23 23 23 23 22 19 17 13 (Y-*/N-- ) APPENDIX III A) An example of the method used in the normalisation of agelength keys and in the construction of catch-curves using the Ricker! method for ~ feliceps males: i) An age-length key is constructed using the sub-sample of aged fish as follows: SIZE CLASS ALL AGE 5 6 210-219 2 3 220-229 2 1 5 1 0 <I 7 8 a 10 11 12 13 14 15 16 4 200-209 3 230-2a92 240-249 250-209 260-269 270-279 280-289 290-299 JOO.309 AGES 3 • 3 9 2 2 • 3 10 5 1 5 , 2 7 2 3 4 3 5 4 310-319 10 9 14 20 20 30 3 320-329 330-339 340-349 4 3 7 6 5 10 1 " • 5 2 2 3 4 10 6 6 2 3 2 0 2 2 2 , 9 9 14 8 4 4 9 195 350-359 31 :l6<>O69 ALL SIZES 9 15 22 16 18 28 32 16 272 Appendix III contd. ii) Through row-wise division, the number of fish of each age in a size-class is divided by the total number of fish in the size-class. SIZE 5 CLASS 200-209 0.75 0.25 210·219 0.33 220-= 0.66 230-239 0.22 0.50 0.33 0.55 0.33 240-249 250-259 260-269 270-279 280-289 200-200 , 7 , AGE 9 ALL 10 11 12 13 14 15 , 0.00 0 .22 0.68 0.30 O.BO 0 . 10 0.14 0.71 0.14 10 7 10 0.50 0.20 0.30 • 0.11 0." 0.22 021 020 0.20 0.15 0.25 0.30 '20-<329 0.20 0.35 0.33 330-339 0.03 31D-319 340-349 " 0.21 0.30 0.25 0.100.10 20 20 0.43 0.100.13 30 0.16 0.32 0.19 0.19 0.10 0 .1 10.22 0.22 0.00 0.22 0.22 350-359 0. 11 0.22 0.66 1.00 36D-369 AlL • 15 AGES , , •, 0.1a ,= SIZES 16 22 16 18 28 32 16 14 8 4 273 " • "• • 1 195 Appendix III contd. iii) Using row-wise multiplication of the table by the total size-frequency sample, followed by column-wise single value is obtained for each age. addition, This represents a the normalised fish number at age. The natural logs of these values are plotted against age in the construction of a catch-curve . SlZE 5 CLASS 200-209 210-21 9 220-229 230-239 0.00 0.33 0.87 0.44 • 7 '0 " '2 •4 13 15 •• 2 • -,.. 24.50 9,80 • 31 35 49 57 75 14.70 8.33 38.00 12.88 16.07 21.42 21.42 16.07 10.95 18.25 21 .90 21.90 290-299 300·309 73 16.00 28.00 20.00 8.00 '.00 42.6655.4612.8017.06 310-319 320-329 330-339 340-349 TOTAl L·F 0.44 1.11 2.67 5.33 .0 ". 4.35 2 1.77 43.54 26.12 26.12 13.06 135 11.22 22.44 22. 44 0.00 22.44 22.44 101 5.33 10.66 32.00 48 350-359 360-369 loge: 9 0 9.30 18.60 3. 10 5.00 25.00 5.00 270-279 280-289 AGE 0.00 0.50 0.17 0.33 0.00 240·249 250-259 • 13.00 13 ._--_._---- ---------'.44 18.91 80.3782.92 83.04 118.3 135.2 75.57 73.64 48.57 18.39 33.11 67. 44 0.37 2.94 4.39 4.42 4.42 4.n 4 .91 4.33 4.30 3." 2.91 274 3.'" 4.21 Appendix III contd . IV) Catch-cur ve construc ted using the normalised age-length key. Note the b i ased values obtained for the older age class e s . • 5 l 5 2 5 2 1.5 0. 5 10 AGE ( Yean.: ) Figure 1: Catch-curve Ricker l method . " constructed 275 " using the Appendix III contd. B) An example of the metho d used i n the construction of c at c hcurves using the Ricker 2 method de scribed for G . feliceps males: -_._---_......_---_._...... __ __..--._-----_._--_... .-. ------- Data range (.) SIZE ClASS lower Upper Mid· N "'Nt" " 11 used in ca!ch-cUlve Umit Umit Rang • ... .... _--- .. _---------- - - - - - - - - _ .. _------ _._._--_ 210 270 260 219 229 238 2<9 259 269 279 28. 290 299 300 310 320 330 340 350 360 309 319 329 220 230 240 250 260 33' 349 359 369 214.5 224.5 234.5 244.5 254.5 264.5 274.5 284.5 294.5 304 .5 31 4.5 324.5 334.5 344.5 354.5 364.5 1 1 2 • 31 35 49 57 7S 73 80 128 135 10 1 48 13 5.26 5.55 6.06 650 8.91 7.48 8.02 5.61 6.02 6. 46 8.92 7.43 7.91 0.35 5.43 5.83 0.394 6 .26 0.421 6.71 0.452 7.20 7.72 4.23 DABS 4.53 4.59 4.76 4.62 8.55 9.2 0.531 8.28 0.581 9.26 9.98 10.78 9.91 0.642 10.7 0.7 18 8.' 9.59 10.33 11.6 0.813 11.18 11.69 12.63 0.938 12.15 13.85 1.108 1.353 1.739 2.437 8 .61 12.74 13.99 15.51 15.34 17.25 17.47 19.91 1.05 0 .37 13.27 14.63 16.32 18.58 0.99 1.62 2.94 4.27 4.59 4.92 4.80 4.31 3.32 1.67 _ . _-- - - - - - - - - - - - - - - - -- -- - - - - -- - - -- -- - -- - -_ . N = The number of fish in the size frequency distribution sample tl dt = = = the age of fish at the lower limit of the size class, the age of fish at the upper limit of the size class, the time needed to grow from the lower to the upper limit of a given size class t = the relative age corresponding to the mid-range of the size class in question. t2 276 Appendix III contd. Since growth contain more is not age linear, groups the than larger size smaller size classes classes. will To compensate for this a correction factor is used in which the numbers of fish in each age class are divided by 'dt'. The natural logs of these corrected ages are then plotted against relative age in the construction of a catch-curve. 6~-, 11 AGE ( Yea r s) IJ 15 " Figure 2: Catch-curve constructed using the total size-frequency distribution. 277 APPENDIX IV An example of the Ho lde n & Meadows (1964 in Holden 197 7) model as applied to G. felic e ps female data. The model estimates t he number of female young b o rn to an initial cohort of 1000 fish . The natural mortality rate was 0.101 and total mortality rate was recruitment o nwards attained fe cundity Fecundity the above cohort at from the age of (11 years) . Age at sexual maturity was at approximately 9 years (see Table XXVI). Mean was 49 and the sex ratio of young was taken as 1:1. did not increase as a function of body size. Given data the following stock sizes can be determined: 9 years = 1000e- a . oao = 402.93 cohort at 10 years cohort at 11 years The 0 . 318 cohort = 402.93e- a . 1a1 = 364.22 = 364. 22e- a . 31B = 265.01 survival and number of female embryos were calculated for values of M 25% above and below the Pauly (1980) estimate. These values of M were also used in the BevertonHolt yield-per-recruit and spawner biomass-per-recruit models. Age in yeBfS No. of fish No. of female embryos __.._._...._------_.._- _ _-- M - 0. 101 M - 0.078 M - 0.126 M - 0.101 M - 402.93 4 95.~ 321.74 9871.78 12141.96 10 364.22 11 265.Q1 283.65 206.33 8923.39 6492.75 12 13 192.82 140.30 150.17 4724.09 3437.35 14 15 102.08 74.27 458.4 1 333.54 242.68 178.58 128.48 93.48 16 17 54.04 39.32 68.02 49.50 16 28.61 36.01 .._._----_.. __ ._--_.._-_._._._. • Total number of temale embryos 0.708 M - 0.126 .. .. 112:31.05 7882.83 6949.43 79.5 57.64 2500.96 8171.79 5945.66 4326.21 3147.76 1819.62 2290.26 1947.75 1417.08 42.09 1323.98 963.34 700.95 1666.49 1212.75 882.25 1031.21 750. 19 558.60 40758.21 51016.18 31957.82 109.61 30.62 22.28 _._-----------------_._-_._-_.._-_._---- 278 5056.31 3678.17 2685.45