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