ASSESSMENT OF ECOLOGICAL RISKS FROM
EFFECTS OF FISHING TO PIKED SPURDOG
(SQUALUS MEGALOPS) IN SOUTH-EASTERN
AUSTRALIA
JUAN MATÍAS BRACCINI
Presented for the degree of Doctor of Philosophy
School of Earth and Environmental Sciences
The University of Adelaide
January 2006
DECLARATION OF AUTHORSHIP
The work contained within this thesis is my own, except where otherwise acknowledged. It
contains no material previously published or written by another person, except where due
reference is made.
There is no material within this thesis, which is accepted for the award of any other degree
or diploma in any university.
I consent to this thesis being made available for photocopying and loan under the
appropriate Australian copyright laws.
JUAN MATÍAS BRACCINI
13 January 2006
Cover image: The piked spurdog, Squalus megalops (photo by Kelvin Aitken).
1
CONTENTS
DECLARATION………..…………………………………………………………………....1
ABSTRACT….………………………………………………………………………………5
ACKNOWLEDGEMENTS
Thesis………………………………………………………………………………7
Chapter..…………………………………………………………………………....9
CHAPTER 1: General Introduction………………………………………………………..12
CHAPTER 2: Notes on population structure of the piked spurdog (Squalus megalops) in
south-eastern Australia
Preamble…………………………………………………………………………..20
2.1 Abstract..……………………...…………………………...………………..…21
2.2 Introduction…………………...…………………………...…………..………21
2.3 Materials and Methods...……...…………………………...……………..……22
2.4 Results………………………...…………………………...……………..……24
2.5 Discussion………..…………...…………………………...………………..…28
CHAPTER 3: Total and partial length–length, mass–mass and mass–length relationships for
the piked spurdog (Squalus megalops) in south-eastern Australia
Preamble………………………………………………………………………..…32
3.1 Abstract..……………………...…………………………...………………..…33
3.2 Introduction…………………...…………………………...………………..…33
3.3 Materials and Methods...……...…………………………...………………..…34
3.4 Results………………………...…………………………...………………..…35
3.5 Discussion………..…………...…………………………...………………..…40
2
CHAPTER 4: Sources of variation in the feeding ecology of the piked spurdog (Squalus
megalops): implications for inferring predator–prey interactions from overall
dietary composition
Preamble…………………………………………………………………….….…44
4.1 Abstract..……………………...…………………………...…………..………45
4.2 Introduction…………………...…………………………...………………..…45
4.3 Materials and Methods...……...…………………………...………………..…47
4.4 Results………………………...…………………………...………………..…54
4.5 Discussion………..…………...…………………………...………………..…68
CHAPTER 5: Determining reproductive parameters for population assessments of
chondrichthyan species with asynchronous ovulation and parturition: piked
spurdog (Squalus megalops) as a case study
Preamble………………………………………………………………………..…86
5.1 Abstract..……………………...…………………………...………………..…87
5.2 Introduction…………………...…………………………...………………..…87
5.3 Materials and Methods...……...…………………………...………………..…90
5.4 Results………………………...…………………………...………………..…97
5.5 Discussion………..…………...…………………………...…………………111
CHAPTER 6: Comparison of deterministic growth models fitted to length-at-age data of the
piked spurdog (Squalus megalops) in south-eastern Australia
Preamble…………………………………………………………………………118
6.1 Abstract..……………………...…………………………...…………………119
6.2 Introduction…………………...…………………………...…………………119
6.3 Materials and Methods...……...…………………………...…………………121
6.4 Results………………………...…………………………...…………………130
6.5 Discussion………..…………...…………………………...…………………140
CHAPTER 7: Hierarchical approach to the assessment of fishing effects on non-target
chondrichthyans: case study of Squalus megalops in south-eastern Australia
Preamble…………………………………………………………………………148
7.1 Abstract..……………………...…………………………...……………....…149
7.2 Introduction…………………...…………………………...…………………149
3
7.3 Materials and Methods...……...…………………………...…………………151
7.4 Results………………………...…………………………...…………………158
7.5 Discussion………..…………...…………………………...…………………162
CHAPTER 8: General Discussion…………………………..…………………………….174
APPENDIX A: Permission to reproduce published chapters..…………………………….184
REFERENCES……………………...………………………..…………………………….188
4
ABSTRACT
Target species in some Australian shark fisheries are adequately managed, but there has
been little attention given to non-target shark species and there is limited information on
the biology of their local populations. Among this group of non-target species, the piked
spurdog (Squalus megalops) is of special interest because it is a dominant and ecologically
important species with high natural abundance. Hence, the main purpose of the present
research was to improve knowledge of the basic biology of this species and to provide
essential data for its management, sustainable use and conservation.
Squalus megalops had a complex population structure, segregating by sex, size and
breeding condition. The sex ratio was biased towards females and there was sexual size
dimorphism with females attaining a larger maximum size than males. Conversion factors
from partial lengths to total length and from partial masses to total mass were determined
due to the common commercial fishing practice of eviscerating, beheading and finning
sharks. Comparisons of total and partial length–length and mass–length relationships
between males and females using different ranges of size showed that there was no effect
of size range on measurements reflecting only somatic growth (fork and carcass lengths;
carcass, pectoral fin and caudal fin masses). However, for variables reflecting somatic and
reproductive growth (total and liver masses), different outcomes can be expected when
different ranges of size are compared.
Examination of dietary composition revealed that S. megalops is an opportunistic predator
that consumes a wide range of prey items. High variability was found when overall
importance of prey items was estimated. Dietary composition varied in space and time,
exhibiting differences among regions, seasons and size classes. Therefore, the intrinsic
natural variability in the dietary composition of S. megalops and its spatial and temporal
variation in diet suggest that information on the ecological relationships among species is
likely to be missed when predator–prey interactions are only inferred from overall diet.
Reproductive parameters were determined for population assessment. For both sexes,
length-at-maturity differed depending on the criterion adopted for defining maturity.
Mature males are capable of mating throughout the year. Females have a continuous
5
asynchronous reproductive cycle. The sex ratio of embryos is 1:1 and litter size and nearterm embryo size increase with maternal length. Females have an ovarian cycle and
gestation period of two years. Although all females are mature at 600 mm, only 50% of
them are in maternal condition, contributing to annual recruitment each year. Hence, for
chondrichthyan species with reproductive cycles of two, three or more years, if maturity
ogives are used in population assessments instead of maternity ogives, the models will
over-estimate recruitment rates.
Age and growth information was also determined for population assessment. Precision
estimates, the relationship between spine total length and body length, edge analysis, and
agreement between counts on the inner dentine layer and the enameled surface support the
use of the first dorsal fin spine for the age estimation of S. megalops. Based on goodnessof-fit criterion, the best growth model for males and females was a two-phase von
Bertalanffy function. However, model selection cannot be based on quality of statistical fit
only and results should be interpreted with caution. Regardless of the model used, the
growth rate of S. megalops, particularly of females, is very low, even within the range of
growth rates reported for shark species.
A three-levelled hierarchical risk assessment approach was trialed to evaluate the
suitability of the approach for S. megalops. Integration of qualitative, semi-quantitative,
and quantitative biological and fishing impact data showed that S. megalops is potentially
highly susceptible to the effects of fishing. A qualitative assessment indicated that the only
fishing related activities to have moderate or high impact on S. megalops were those
associated with ‘capture fishing’ of the otter trawl, Danish seine, gillnet and automatic
longline methods. A semi-quantitative assessment ranked S. megalops at risk because of its
low biological productivity and, possibly, its catch susceptibility from cumulative effects
across the separate fishing methods. Finally, a quantitative assessment showed that
population growth is slow even under the assumption of density-dependent compensation
where the fishing mortality rate equals the natural mortality rate. Therefore, conservation
and management for sustainable use of S. megalops will require a close control of fishing
mortality due to the low capacity of this species to withstand fishing pressure.
6
ACKNOWLEDGMENTS
Thesis acknowledgments
It is not an easy task to start acknowledging people for their help and contribution to my
thesis without failing to name them all. If I have to start with someone, that shall be
Bronwyn Gillanders and Terry Walker, my supervisors. I will be forever grateful for, first,
giving me the opportunity to pursue this graduate degree and second, their advice and
encouragement since the very beginning. I could not have done it without their support. It
has been a great honour to work side by side.
I am very thankful to the members of two laboratories for helping in the lab, for collecting
samples and more importantly, for making my stay in Australia a wonderful life
experience, creating a fun research environment and making me feel home away from
home. In the Southern Seas Ecology Laboratories of The University of Adelaide are Sean
Connell, Melita de Vries, Travis Elsdon, Meegan Fowler-Walker, Andrew Irving, Bayden
Russell, Jarrod Stehbens, Ben Brunton, Paris Goodsell, Justin Rowntree, Elaine Vytopil,
Andrew Munro, Beth Hammond, Simon Hart, and Dan Gorman. In the Chondrichthyan
Laboratory of the Primary Industries Research Victoria are Justin Bell, Matt Reardon,
Fabian Trinnie, Megan Storrie, Dave Phillips, and Javier Tovar-Avila.
Special thanks to Cynthia Awruch and Javier Tovar-Avila, my fellow Latin American
shark biologists and friends, for the many discussions on the biology of sharks and some
other more philosophical topics. It has been a pleasure to undertake my Ph.D. along with
you. Thanks to Peter Risley, my “sample provider” and friend, for sharing his endless
knowledge on the ocean in general and sharks in particular aboard the ‘Ester-J’, the most
wonderful shark fishing vessel. My gratitude also to Ken Graham and Ross Daley for
helping with the logistics.
Special thanks also go to Emil and Meegan, the Southern Ocean and its amazing surf, and
the words of guidance of Bob Marley’s songs, for helping me to discover my spiritual side,
which has been crucial not just for the completion of this thesis but mostly for my journey
in life.
7
How can I express enough gratitude to my friends José Facelli and his wife, Evelina? I will
be forever grateful for their help in the tough process of settling in a new country. For
being there since the very first moment, many thanks.
To mum and dad, my brother Fede and sister Agus, what can I tell you? Simply that
without your support, kindness and love I would have never accomplished anything in life.
Finally, I thank Alicia above all, for giving up friends and family to get on board the ship
of adventure with me and move to the other side of the world. Without her unconditional
love, constant support, patience and laugh, this Ph.D. would be nothing but a dream. For
this, and for inspiring me every single day of my life, thank you.
8
Chapter acknowledgments
Peter Risley, Glenn Richardson and the crew of the fishing vessel ‘Nungurner’ provided
help in sample collection (Chapters 2, 3, 4, 5, and 6).
Martin Gomon, Gary Poore, Mark Norman, Robin Wilson, Winston Ponder, and Greg
Rouse helped in the identification of prey items (Chapter 4).
Mariano Koen Alonso assisted with data analyses (Chapter 4).
Sandy McFarlane and his colleagues from the Pacific Biological Station and Gill Watson
and Malcolm Smale from Port Elizabeth Museum at Bayworld re-read a subsample of
spines (Chapter 6).
Sarah Irvine commented on ageing techniques (Chapter 6), and Miguel Araya and
Vladimir Troynikov commented on growth models (Chapter 6).
Enric Cortés and Malcolm Haddon advised on Monte Carlo simulation modelling (Chapter
7).
Anne Gason and Masaaki Machida assited with data management (all chapters).
This research was supported by an International Postgraduate Research Scholarship and a
University of Adelaide Postgraduate Research Scholarship to Matías Braccini and an
Australian Fisheries Research and Development Corporation grant (FRDC 2002/033) to
Terry Walker. Bronwyn Gillanders was supported by an Australian Research Council QEII
Research Fellowship. Funding for the field and laboratory components was provided by
Sea World Research and Rescue Foundation, Royal Zoological Society of New South
Wales, Nature Foundation SA, and Royal Zoological Society of South Australia to Matías
Braccini.
9
The piked spurdog, Squalus megalops (photo by Kelvin Aitken).
10
CHAPTER 1
11
CHAPTER 1
GENERAL INTRODUCTION
Global fish production has been increasing since the 1950s, but this increase is mainly due
to aquaculture, as production from capture marine fisheries has remained relatively stable.
Currently, about 50% of the worldwide major marine fish stocks are fully exploited,
another 15–18% are overexploited and 9–10% of stocks have been depleted or are
recovering from depletion (FAO 2000a). Whereas depleted and recovering teleost and
invertebrate stocks may have ample potential for recuperation, this may not be the case for
sharks and other chondrichthyans.
Currently, shark species are facing an increasing risk of depletion due to the combination
of anthropogenic impacts such as fishing overexploitation and their unique life-history
strategies. Sharks are harvested by industrial, artisanal, traditional, and recreational
fisheries around the world and these fisheries have contributed to their decline (Walker
1998). Industrial fisheries directed at one or a small number of species of shark have
seldom been sustainable, although Walker (1998) concluded that some species can be
fished sustainably. However, other researchers (e.g. Holden 1974; Compagno 1990a;
Musick et al. 2000; Stevens et al. 2000) claim that most industrial shark fisheries have
been characterised by a rapid decline in catch rates or by a complete collapse of the fishery
after the initial exploitation.
The decline in shark captures is masked by unreliable and underestimated shark-fishing
statistics. A large proportion of world catches of sharks are not recorded in official fishery
statistics because they are taken incidentally (retained non-targeted catch or by-product)
while fishing for other target species (Bonfil 1994). Although official figures report about
700,000 T of elasmobranchs caught at the end of the 1980s, the actual level is at least in
the range of 1.0–1.35 million T including unreported by-product (Bonfil 1994). Moreover,
there is a high component of illegal fishing (Compagno 1990a; Bonfil 1994).
Even more worrisome, shark species are mostly taken in multispecies fisheries directed at
more highly valued teleosts and in most fisheries part or the entire shark catch is discarded
12
(discarded non-targeted catch or by-catch), mostly dead or severely damaged, at the point
of capture (FAO 2000b). Shark by-catch is not included in fishery statistics. For instance,
Bonfil (1994) reported that in the high seas the estimated annual incidental capture of
sharks and rays at the end of the 1980s was 250–270,000 T/year, but another 230–240,000
T/year could have been discarded (by-catch). Besides sharks being taken as by-product and
by-catch, there may also be some collateral mortality of sharks that are not actually
captured, associated with the fishing gear (Hall 1996). Such mortality does not form part of
the fishery catch statistics and is unaccounted fishing mortality.
In addition to the impact of fisheries, there are several reasons why the conservation of
sharks is of particular concern. First, most shark species have reproductive and growth
characteristics and feeding habits that make their vulnerability to the effects of fishing high
compared with teleost fish. Sharks mature late, exhibit low fecundity, long gestation
periods and high longevity, and many species have slow growth (Holden 1973). Hence,
sharks have low intrinsic rates of population increase and low resilience to fishing
mortality (Hoenig and Gruber 1990; Smith et al. 1998). Second, most shark fisheries have
been managed by population models designed for more productive animals such as teleost
fishes, increasing the vulnerability of shark populations to overexploitation (Musick et al.
2000). Third, sharks are predators at or near the top of the marine food web and hence they
are indirectly affected by the impacts on the species below them in the marine food web.
Finally, assessment of shark populations is severely impaired by a lack of biological
information (Cortés 1998a), especially for non-commercial species.
Therefore, the effects of fishing on most shark populations that are either targeted directly
or caught incidentally remain mostly unidentified and unquantified. As a result, there is an
increasing concern about the sustainability of shark stocks. International agreements reflect
the concern about the sustainable management of sharks; the Food and Agriculture
Organisation of the United Nations (FAO) developed the International Plan of Action for
the Conservation and Management of Sharks (IPOA-Sharks) to ensure the conservation
and management of sharks (targeted and non-targeted species) and their long-term
sustainable use (FAO 2000b).
As a signatory to the IPOA-Sharks, Australia has developed the Australian National Plan
of Action for the Conservation and Management of Sharks to ensure that all Australian
13
shark species are managed sustainably regardless of fishery or jurisdictional boundaries.
This need for management is also identified in the Fisheries Management Act 1991 and the
Commonwealth Environment Protection and Biodiversity Conservation Act 1999; this
latter Act highlights the need for strategic assessment of fisheries operating under
Commonwealth jurisdiction for ecological impacts on 1) target and by-product species, 2)
by-catch species, 3) threatened, endangered and protected species, 4) marine habitats and
5) marine food webs. Hence, different types of survey, assessments and biological studies
are required to meet the terms of the management commitments.
Australian sharks have been exploited by non-industrial and industrial fisheries. Local
sharks have been fished by Australian aborigines as a source of food. Since European
settlement in Australia, sharks have been captured to consume their meat and to extract oil
from their livers for lighting and medicinal purposes (Walker 1998). Local sharks have
also been utilized as fertilizer, as a source of Vitamin A and squalene and their leather has
been used for making bags, wallets, shoes, and other garments (Walker 1998). Target and
non-target shark species are captured in several different fisheries. Sharks are taken by
foreign vessels in Australian waters, are captured in two shark control programs
(Queensland and New South Wales), are targeted by recreational and game fishers, and are
caught as by-catch or by-product or both in more than 70 other commercial fisheries.
Furthermore, approximately 3,900 T of whole sharks were potentially finned and discarded
during 1998–99 (Rose and McLoughlin 2001). However, most of the Australian recorded
shark catch is taken by directed fisheries in the Southern and Eastern Scalefish and Shark
Fishery (SESSF). This fishery comprises the South East Trawl Fishery, Great Australian
Bight Trawl Fishery, and Gillnet Hook and Trap Fishery (GHATF). Of these fisheries, the
GHATF is the most important in terms of shark landings, contributing one-third of
Australia’s shark catch (Anonymous 2002). The GHATF deploys demersal gillnets and
longlines to harvest several species of temperate-water sharks (Walker 1999). The primary
target species are the gummy (Mustelus antarcticus) and the school (Galeorhinus galeus)
sharks. Walker (1999) reported that during the period 1970–97 these two species
comprised 88% of the catch; the remaining 12% were by-product species comprising saw
shark (Pristiophorus nudipinnis and P. cirratus) (7%), elephant fish (Callorhynchus milii)
(2%), and several other species of shark such as bronze whaler (Carcharhinus brachyurus),
whiskery (Furgaleus macki), dusky (C. obscurus), broadnose sevengill (Notorynchus
14
cepedianus), blue (Prionace glauca), mako (Isurus oxyrinchus), and angel (Squatina
australis) sharks.
Currently, few studies have addressed the stock assessment of Australian sharks and these
studies have only been focused on target species (e.g. Walker 1992; Punt and Walker 1998;
Simpfendorfer et al. 2000). The models used in these studies are complex and require
extensive data collection. However, shark populations can be depleted before collecting
sufficient data to undertake reliable stock assessment (Compagno 1990a). Although
complex models and comprehensive long-term monitoring data are required to reduce
uncertainty, in the short-term, ecological risk assessment methods based on simpler data
sets and judgement can provide advice on management of species at risk (Walker 2004).
Therefore, a different approach is required for the evaluation of non-target species, in
which long time-series of data have generally not been recorded. This new approach
should involve the development and implementation of ecological risk assessment methods
of each non-target shark species.
A risk assessment is a tool that allows quantifying risks to then allocate efforts and
regulations to protecting species identified as at risk. An ecological risk assessment
approach was originally applied to by-catch elasmobranch species (sharks and rays) in
northern Australia by Stobutzki et al. (2002) and the methodology was taken a step further
by Walker (2004) to make it more compatible with more comprehensive fishery
assessment methods. For the assessment of risks to a species, information on the biological
productivity and catch susceptibility is needed. Biological productivity is a measure of the
population turnover potential or how fast a population can recover after overfishing. Catch
susceptibility is a measure of the extent of the impact of a fishing method. In addition,
knowledge on the feeding ecology of a species is also needed, as this provides information
on the position and function of shark species in marine ecosystems. This knowledge is also
important as potential negative effects on the prey of a shark may have negative food web
cascading effects through trophic interactions. Furthermore, due to growing awareness of
the need of a multispecies approach to fisheries management (Gulland 1978; Botsford et
al. 1997), feeding ecology information is needed to determine interactions among
components of the ecosystem.
15
Within the group of non-target Australian sharks, which neither the ecological risk nor the
status of the stock have been assessed, the piked spurdog (Squalus megalops) is of special
concern as it is the most caught by-catch shark species by trawling operations on the
continental shelf and slopes of New South Wales, Victoria, Tasmania and South Australia
(Walker and Gason 2006). Also, a proportion of the captured sharks is landed and retailed
as ‘flake’. To date, the current marketed catch within dogfish species in south-eastern
Australia is mostly S. megalops (Daley et al. 2002) and 16 T were marketed in 2004 from
the GHATF (Walker and Gason 2005).
Squalus megalops is a member of the family Squalidae and it is one of the most abundant
shark species of southern Australia (Bulman et al. 2001; Graham et al. 2001). It is a
demersal species that is distributed off southern and eastern Australia, from Carnarvon
(Western Australia) to Townsville (Queensland), including Tasmania (Last and Stevens
1994). It has also been reported off the coasts of Brazil (Vooren 1992) and South Africa
(Bass et al. 1976) and there are unconfirmed reports off Indo China, New Caledonia and
New Hebrides (Last and Stevens 1994). This species inhabits the continental shelf and
upper continental slope (depths <510 m) in warm temperate and tropical areas (Last and
Stevens 1994). In eastern Australia, the abundance of S. megalops has not declined after 20
years of fishery exploitation. Hence, due to the depletion of many harvested shark species
and a decline in abundance of most other shark species in southern Australia (Graham et
al. 2001), S. megalops will inevitably become a more sought after species. However, due
to little knowledge on its biology and on the effect that fishing has on its local
population(s), the conservation status of S. megalops is uncertain.
The aim of the present study was to contribute to the ecological risk assessment of S.
megalops, an important by-catch shark species taken in south-eastern Australia. To
investigate the biology and ecological risk of this species, the specific objectives of this
research were to:
determine the population structure of S. megalops in south-eastern Australia
(Chapter 2),
determine total and partial length–length, mass–mass and mass–length relationships
(Chapter 3),
analyse stomach contents to determine the feeding ecology of this species and its
role in marine ecosystems (Chapter 4),
16
describe the reproductive biology of this species with emphasis on the information
needed for quantitative population studies (Chapter 5),
estimate the age and growth rate of S. megalops captured in south-eastern Australia
(Chapter 6), and
use qualitative, semi-quantitative and quantitative techniques for the assessment of
ecological risks of S. megalops (Chapter 7).
Notes on chapter style
Each chapter of this thesis that presents original data (Chapters 2–7) has been written in a
style suitable for publication in a scientific journal and can be read as a separate study.
Each chapter is preceded by a preamble that briefly describes the content of the chapter,
presents information on the publication status of the chapter at the time of thesis
submission (where applicable), and describes the contributions of all co-authors to the
research therein. Tables and figures appear within the text and all references cited in this
thesis are compiled at the end of the thesis and not at the end of each chapter.
17
The shark gillnet vessel ‘Ester-J’ (photo by the author).
18
CHAPTER 2
19
CHAPTER 2 PREAMBLE
Chapter 2 describes the population structure of S. megalops in south-eastern Australia and
shows the segregation pattern and sexual size dimorphism that the species exhibits in this
area. At the time this thesis was submitted (January 2006), this chapter was under peerreview with the journal Ciencias Marinas, with myself as senior author, and Bronwyn M.
Gillanders (The University of Adelaide) and Terence I. Walker (Primary Industries
Research Victoria) as co-authors.
I was responsible for sampling, analysing and interpreting the data, and for writing the
manuscript. Bronwyn M. Gillanders and Terence I. Walker supervised development of
research, data interpretation and manuscript evaluation.
20
CHAPTER 2
NOTES ON POPULATION STRUCTURE OF THE PIKED SPURDOG (SQUALUS
MEGALOPS) IN SOUTH-EASTERN AUSTRALIA
2.1 ABSTRACT
The population structure of Squalus megalops was studied in south-eastern Australia. A
total of 929 sharks was analysed. The sex ratio was biased towards females. There was
sexual size dimorphism; females attained a larger maximum size than males. Analyses of
sex ratio and length-frequency distributions of selected fishing shots suggested that S.
megalops had a complex population structure. Despite the opportunistic nature of the
sampling design, individual analysis of selected fishing shots suggested that S. megalops
segregated by sex, size and breeding condition. Small females and males segregated from
large females. In addition, large females in the first year of pregnancy seemed to be
separated from those in the second year of pregnancy. The segregation pattern exhibited by
this species needs to be considered in management plans.
2.2 INTRODUCTION
Most dogfish species (Squalidae) have complex population structures. Segregation in time
and space by sex, size and mature condition is a common feature of this group. Pregnant
and ovulating females of the roughskin dogfish (Centroscymnus owstoni) segregate from
immature specimens (Yano and Tanaka 1988). Size and sexual segregation have been
reported for the southern laternshark (Etmopterus granulosus) (Wetherbee 1996), the black
dogfish (Centroscyllium fabricii) (Yano 1995), the leafscale gulper shark (Centrophorus
squamosus), the Portuguese dogfish (Centroscymnus coelolepis) (Clarke 2000) and the
birdbeak dogfish (Deania calcea) (Clark and King 1989; Clarke et al. 2002a). Among
Squalus species, the spiny dogfish (S. acanthias) exhibits a complex population structure
related to its reproductive cycle. Off New Zealand, parturition, ovulation and mating occur
in deep water whereas pregnant females spend the first year of gestation in shallow waters
before migrating back to deep water during the second year of pregnancy (Hanchet 1988).
A complex size structure is also reported for the piked spurdog (S. megalops). In South
Africa, this species forms large schools often segregated by sex and size (Compagno
21
1990b). In New South Wales, southeast coast of Australia, large females segregate from
males and juveniles, aggregating in deeper waters and different regions (Graham 2005). In
Australia, many of the harvested species of sharks have been depleted and the abundance
of most other shark species has declined at least in southern Australia (Graham et al.
2001). However, S. megalops has a high natural abundance (Bulman et al. 2001; Graham
et al. 2001) so this shark will inevitably become a more sought after species. Given that the
current marketed catch within dogfish species in south-eastern Australia is mostly S.
megalops (Daley et al. 2002), a better understanding of the population structure of this
species is needed.
The aim of the present study was to determine the population structure of S. megalops in
south-eastern Australia. Due to complex segregation patterns found in Squalus species,
emphasis is put on separation between sexes, sizes and breeding condition.
2.3 MATERIALS AND METHODS
Male and female S. megalops were collected opportunistically from the by-catch of shark
gillnet and demersal trawl fishing vessels operating in the Australian Southern and Eastern
Scalefish and Shark Fishery during October 2002–April 2004. Samples were mainly
collected from Robe, Lakes Entrance and Ulladulla (Fig. 2.1). Shark gillnet fishing gear
consisted of monofilament netting of 6½-inch mesh-size, ∼4000 m long, and 2.4 m deep
deployed for 4–8 hours during night and day. Demersal trawl fishing gear consisted of
otter trawl or Danish seine nets. Otter trawl nets had a headline length of 24–50 m with a
stretched codend mesh-size of 90 mm and Danish seine nets had a headrope length of 30–
59 m with a stretched codend mesh-size of 38 mm. Trawling operations lasted for 3–4
hours during night and day. Fishing depth varied with location and fishing gear between 21
and 238 m. Each individual S. megalops was measured (total length, TL) to the nearest
millimetre. The reproductive condition of males and females was determined following
Braccini et al. (2006; see Chapter 5 for description of reproductive stages of males and
females).
The sex ratio and the length-frequency distribution of males and females were determined
for the entire sample and per fishing shot when depth information was available and
sample size per shot was ≥10. For the entire sample, a Chi-square test with Yates’
continuity correction and a Kolmogorov-Smirnov test (KS; Zar 1999) were used to test for
22
differences in the sex ratio and the length-frequency distribution of males and females,
respectively.
129º E
141º E
150 º E
200 m
Ulladulla
•
37º S
Robe
Lakes
Entrance
•
•
Australia
41º S
Tasmania
N
200
0
200 Miles
Figure 2.1. Map of sampling area (shaded).
23
2.4 RESULTS
Data from a total of 929 sharks were analysed. For the entire sample, the sex ratio (males :
females) was significantly biased towards females (1 : 3.49,
2
= 284.39, d.f = 1, P <0.001).
By fishing gear, the sex ratio was also biased towards females (1 : 25.34,
1, P <0.001 for shark gillnets, and 1 : 2.83,
2
2
= 133.07, d.f =
= 175.64, d.f = 1, P <0.001 for the demersal
trawl nets). Size of sharks ranged from 274–470 mm TL for males and 270–635 mm TL
for females. There was sexual dimorphism in maximum size; mean TL (± s.e.) of males
was 404 (2) mm whereas mean TL of females was 480 (3) mm. The length-frequency
distribution was significantly different between males and females (KS, dMAX = 0.685,
nmales = 207, nfemales = 722, P <0.001) (Fig. 2.2a). Shark gillnets mainly captured large
females (Fig. 2.2b) whereas otter trawl and Danish seine nets caught males and females of
a broader range of sizes (Figs. 2.2c, d).
Analyses of sex ratio and length-frequency distributions of selected fishing shots for the
trawl method (18 fishing shots in total) suggested that S. megalops had a complex
schooling pattern. In some cases, males and small females (<460 mm TL) were trawled
together (Fig. 2.3a; Table 2.1, shots 146, 155), but in other shots mainly small immature
females (Fig. 2.3b; Table 2.1, shot 205) or large females were captured (Fig. 2.3c; Table
2.1, shots 497, 509, 540). Among large females, those in the first year of pregnancy tended
to be separated from females in the second year of pregnancy (Table 2.1, shots 497, 509,
540). In one shot, a school of large males was caught (Fig. 2.3d; Table 2.1, shot 491).
Analyses of length-frequency distributions of selected fishing shots for the gillnet method
(8 fishing shots in total) also revealed a complex schooling pattern as only large females
were captured (Figs. 2.3e, f; Table 2.1, shots 467, 521); however, it may be possible that in
these shots small females and males were not sampled by the gillnet due to the sizeselectivity of this fishing gear.
24
25
LE
LE
LE
LE
Queenscliff
Robe
Robe
Ulladulla
Ulladulla
146
509
155
205
491
467
521
497
540
March
February
March
January
February
April
March
February
February
Month
Tn
Tn
Gn
Gn
Tn
Ds
Ds
Ds
Ds
Gear
164
210.5
90
87
75
40.5
38
68.5
73
depth (m)
Mean
0
0
0
0
1
3
3
0
22
I
1
2
0
2
33
5
6
0
5
M
Male
1
3
1
23
0
29
7
10
28
I
LE (Lakes Entrance); Ds (Danish seine net); Tn (trawl net); Gn (gillnet).
Location
Shot
23
0
20
17
0
2
2
29
2
Year 1
3
9
5
5
0
1
0
4
1
Year 2
Female
Number
0
0
1
3
0
3
2
1
0
O & P-p
1 : 27
433 (0)
429 (17)
⎯
⎯
1:6
416 (0)
416 (2)
⎯
1 : 24
406 (6)
1 : 4.4
434 (4)
⎯
⎯
1 : 1.2
386 (5)
Male
506 (5)
491 (16)
573 (5)
528 (7)
⎯
438 (7)
476 (12)
500 (5)
421 (8)
Female
Mean size (± s.e.)
1 : 1.1
ratio
Sex
Table 2.1. Sample details for selected fishing shots showing the number of immature (I) and mature (M) males and the number of immature (I),
first year of pregnancy (Year 1), second year of pregnancy (Year 2), and ovulating and post-partum (O & P-p) females.
Number of specimens
70
60
50
40
30
20
10
0
10
20
30
40
50
60
70
40
n = 207
(a)
(b)
20
10
0
10
20
males
females
30
n = 152
n = 722
40
40
40
(c)
n = 16
(d)
30
30
20
20
10
10
0
0
10
10
20
20
30
n=6
30
n = 115
40
n = 185
30
n = 455
40
260 300 340 380 420 460 500 540 580 620
260 300 340 380 420 460 500 540 580 620
Total length (mm)
Figure 2.2. Length-frequency distribution of Squalus megalops for (a) the entire sample,
and by fishing gear: (b) shark gillnet, (c) otter trawl net, and (d) Danish seine net.
26
16
16
(a)
12
n = 27
12
8
8
4
4
0
0
4
4
8
8
(b)
12
12
n = 35
n = 31
16
16
Number of specimens
16
(c)
12
n=0
16
12
8
8
4
4
0
0
4
4
8
8
(d)
females
12
n = 44
16
n=0
16
16
16
(e)
n=2
12
8
8
4
4
0
0
4
4
8
8
12
n = 34
males
12
12
n=8
(f)
n=0
12
n = 48
16
260 300340380420 460 500 540580 620
n = 27
16
260 300 340 380 420 460 500 540580620
Total length (mm)
Figure 2.3. Length-frequency distribution of some of the selected fishing shots (location,
gear and season): (a) 146 (Lakes Entrance, Danish seine net, summer), (b) 205 (Lakes
Entrance, Danish seine net, autumn), (c) 509 (Lakes Entrance, Danish seine net, summer),
(d) 491 (Queenscliff, Danish seine net, summer), (e) 467 (Robe, gillnet, summer), and (f)
521 (Robe, gillnet, autumn).
27
2.5 DISCUSSION
The present study provides evidence of a complex population structure for S. megalops in
south-eastern Australia. Although the opportunistic nature of the sampling design did not
allow for the effects of time, region, depth or sampling gear to be rigorously tested,
individual analysis of selected fishing shots suggested that S. megalops segregates by sex,
size and breeding condition. Sexual and size segregation is a common characteristic of
many shark species where juveniles, adult males and adult females separate into different
groups (Springer 1967). Male and small female S. megalops were sampled together and did
not occur with large females. Furthermore, in one of the shots, a school of only large
mature males was collected and on other occasions, schools of only large females were
captured. Graham (2005) reported a similar pattern off New South Wales where large
females and males occur in different regions and depths. In addition, large females in the
first year of pregnancy seemed to be separated from those in the second year of pregnancy.
Similarly, large female S. acanthias segregate by breeding condition. Pregnant females
spend the first year of pregnancy in shallow waters, perhaps due to warmer water
requirements for early embryo development, and migrate to deeper offshore waters during
the second year of pregnancy (Hanchet 1988).
Female S. megalops attained larger sizes than males. Sexual size dimorphism is frequently
observed in sharks and it is more common among viviparous species where for females,
due to their more energetically demanding reproductive mode, there is a strong selection
pressure for a larger body size (Sims 2003). Many species that have sexual size
dimorphism also exhibit sexual segregation (Sims 2003). In sexually dimorphic mammals,
males attain a larger size and seek habitats with higher food availability, whereas females
prefer habitats safe from predation (Main et al. 1996). Sex-specific habitat use has been
reported for several shark species (e.g. McLaughlin and O'Gower 1971; Sims et al. 2001).
Female scalloped hammerhead sharks (Sphyrna lewini) select habitats with more abundant,
energy-rich prey (Klimley 1987). Large female S. megalops had a different diet and
consumed more energy-rich prey than males and small females during summer and autumn
(Braccini et al. 2005; Chapter 4). Hence, if large females have different energetic
requirements, selection of different diet quality may lead to sexual size segregation (Main
et al. 1996). In this way, large females may inhabit areas with higher food availability
while males and small females trade off food preference for areas with fewer predators
(Bowyer 2004). However, other hypotheses, such as migration, differences in swimming
28
capabilities, male-avoidance, or absence of aggression between similar sized individuals
have also been proposed to explain segregation among sharks (Springer 1967; Sims 2003).
Given that S. megalops is the most commonly taken by-catch shark species by demersal
trawlers in south-eastern Australia (Walker and Gason 2006), further information is needed
on the location of parturition areas, and the spatial distribution of juveniles, males and
females in different breeding condition. A more rigorous sampling design would allow the
extent of the segregation pattern of S. megalops to be determined and testing the
hypotheses proposed to explain this phenomenon.
Small S. megalops were not collected by the sampling gears. The small length-classes are
often missing in dogfish studies (Clarke 2000). Gillnets select for larger-sized specimens,
but demersal trawl nets are likely to catch the smaller S. megalops as it was shown that,
when present, the small size-classes are retained by the 90-mm mesh-size codend (Graham
2005). Hence, small individuals probably occur outside the trawling grounds, being
unavailable to the trawl gear. A pelagic phase has been proposed for juvenile S. megalops
(Compagno et al. 1991). This life strategy would decrease predation risk as predation by
larger sharks and teleosts most likely occurs near the seabed (Graham 2005).
The sex ratio suggests that either females are more common in the population or they are
more vulnerable to fishing than males. In the latter case, this must be considered in the
management of this species, as the selective removal of females may have a
disproportionate effect on the reproductive output of the population. Furthermore, the
segregation pattern of S. megalops also needs to be considered in management plans.
Different management of males and females has already been proposed for mammals with
sexual segregation (Bowyer 2004).
29
One of several trawlers that assisted in sampling (photo by the author).
30
CHAPTER 3
31
CHAPTER 3
TOTAL AND PARTIAL LENGTH–LENGTH, MASS–MASS AND MASS–
LENGTH RELATIONSHIPS FOR THE PIKED SPURDOG (SQUALUS
MEGALOPS) IN SOUTH-EASTERN AUSTRALIA
3.1 ABSTRACT
Common commercial fishing practices of eviscerating, beheading and finning sharks create
the need for using conversion factors from partial lengths to total length and from partial
masses to total mass. In the present paper, these conversion factors were calculated for
Squalus megalops. In addition, total and partial length–length and mass–length
relationships of male and female S. megalops were compared using different ranges of
size. There was no effect of size range on measurements reflecting only somatic growth
(fork and carcass lengths; carcass, pectoral fin and caudal fin masses) but for variables
reflecting somatic and reproductive growth (total and liver masses), different outcomes can
be expected when different ranges of size are compared.
3.2 INTRODUCTION
Fisheries taking sharks are common throughout the world. Given that commercial shark
species are normally beheaded, eviscerated and landed in one of two forms: with fins
attached (‘carcass’) or without fins attached (‘trimmed carcass’), only partial lengths and
masses can be recorded after landing (FAO 2000b). Furthermore, due to increases in
worldwide demand for shark fins, in many fisheries only the fins are retained whereas the
rest of the animal is discarded. Due to these fishing practices, relationships between partial
lengths and total length and between partial masses and total mass of shark are needed to
determine the length and mass composition of captured sharks. Therefore, conversion to
live weight and length equivalent units using appropriate conversion factors is an essential
requirement for fisheries monitoring programmes and stock assessments.
Size relationships and size conversion factors have several biological applications and are
commonly used in fishery management. Size relationships, particularly total mass–total
length relationship, are commonly reported in biological studies of sharks (e.g. Stevens and
McLoughlin 1991; Kohler et al. 1995). Many studies test for differences between sexes in
33
these relationships; in some cases, significant differences are found (e.g. Chiaramonte and
Pettovello 2000; Walker 2005), whereas other studies show no differences (e.g. Bridge et
al. 1998; Francis and Stevens 2000). Many species of sharks exhibit sexual dimorphism in
maximum size, females being larger and heavier than males (e.g. Cortés 2000). For these
species, size relationship comparisons are thus made between groups of different ranges of
size so similarities or differences in these relationships may be an artefact of comparing
smaller individuals (males) with larger individuals (females).
In the present study, length–length and mass–length relationships of male and female piked
spurdogs (Squalus megalops), an abundant shark of southern Australia (Graham et al.
2001), were compared using different ranges of size. In addition, due to the common
fishing practice of eviscerating, beheading and finning sharks, conversion factors from
partial lengths and partial masses to total length and total mass were determined.
3.3 MATERIALS AND METHODS
Male and female S. megalops were collected from the by-catch of shark and demersal trawl
fishery vessels operating in the Australian Southern and Eastern Scalefish and Shark
Fishery during October 2002–April 2004. Total (TL), fork (FL) and carcass (CL) lengths
were measured to the nearest millimetre. Fork length was measured from the tip of the
snout to the caudal fork and CL was measured from the fifth gill-slit to the precaudal pit.
Total (TM), carcass (CM), liver (LM), pectoral fins (PFM) and caudal fin (CFM) masses
were recorded to the nearest gram. All length and mass measures were recorded in the
laboratory. Differences between sexes were tested by Student t-test on the slopes and
intercepts of the linear regression of FL and CL against TL and the linear regression of ln
(TM), ln (CM), ln (LM) ln (PFM), and ln (CFM) against ln (TL) or ln (CL) (Kleinbaum et
al. 1988). A factor is used to correct for biases caused by natural logarithmic
transformation (Beauchamp and Olson 1973).
Squalus megalops showed sexual dimorphism in maximum size, ranging from 274–470
mm TL (86–465 g TM) and 270–635 mm TL (84–1411 g TM) for males and females,
respectively. Hence, samples of different ranges of size were selected for statistical
comparisons. The following groups were compared: males (n = 207), all females (n = 721)
and small females (≤470 mm TL, n = 297). Geometric mean regressions (Ricker 1973)
were used to determine conversion factors from partial lengths and partial masses to total
34
length and total mass and from total length and total mass to partial lengths and partial
masses.
3.4 RESULTS
There were no significant differences in the FL–TL, CL–TL, CM–TL, PFM–TL, CFM–TL,
and CM–CL relationships between males and all females and between males and small
females (t-test, P >0.05 for comparisons of slopes and intercepts). Therefore, sexes and
sizes were pooled for calculation of conversion factors, shown in Table 3.1. The
conversion factors estimated are applicable to the size range analysed (270–635 mm TL),
which covers most of the population size range, with the exception of neonates (TL <270
mm).
There were significant differences in the TM–TL relationship between males and all
females (t-test, d.f. = 902, t = 5.06, P <0.05 for comparison of slopes and t = 5.01, P <0.05
for comparison of intercepts; Fig. 3.1). However, when animals of the same size range
were compared (males and small females), no differences were detected (t-test, d.f. = 500, t
= 1.78, P >0.05 for comparison of slopes and t = 1.74, P >0.05 for comparison of
intercepts; Fig. 3.1). There were no differences in the LM–TL relationship between males
and all females (t-test, d.f. = 873, t = 0.89, P >0.05 for comparison of slopes and t = 0.86, P
>0.05 for comparison of intercepts; Fig. 3.1), but significant differences were detected
between males and small females (t-test, d.f. = 481, t = 4.51, P <0.05 for comparison of
slopes and t = 4.47, P <0.05 for comparison of intercepts; Fig. 3.1). To standardize for the
effects of size, CM and LM were expressed as a proportion of TM. Carcass mass expressed
as a proportion of TM (CMP) decreased with TL for all females (Fig. 3.2), whereas the
CMP–TL relationship showed no trend for males and a slight decrease for small females
(not shown). Liver mass expressed as a proportion of TM (LMP) increased with TL for
small females (Fig. 3.2). For males and all females, the LMP–TL relationship showed no
trend (not shown).
35
Table 3.1. Conversion factors derived from geometric mean regressions. Estimated
parameters for converting (a) partial lengths and partial masses to total length and total
mass and (b) total length and total mass to partial lengths and partial masses. Values for
parameters (and standard error) derived from the equation Y = a + b X.
Variables
X
n
b ± s.e.
a ± s.e.
Y
(a)
Fork length
Total length
547
1.138 ± 0.005
5.736 ± 1.857
Carcass length
Total length
490
1.587 ± 0.017
26.764 ± 4.419
Carcass mass
Total mass
851
1.939 ± 0.011
–58.518 ± 3.537
Pectoral fin mass
Total mass
351
64.437 ± 0.962
–87.998 ± 9.119
Caudal fin mass
Total mass
352
82.529 ± 0.948
–136.357 ± 7.562
Total length
Fork length
547
0.878 ± 0.004
–4.972 ± 1.651
Total length
Carcass length
490
0.630 ± 0.007
–16.901 ± 2.959
Total mass
Carcass mass
851
0.516 ± 0.003
(b)
30.230 ± 1.676
Total mass
Pectoral fin mass
351
0.016 ± 2.31 × 10
1.367 ± 0.123
Total mass
Caudal fin mass
352
0.013 ± 1.43 × 10–4
1.467 ± 0.076
a and b are parameters and n is sample size.
36
–4
1500
Total mass (g)
males
all females
small females
1000
500
0
200
Liver mass (g)
150
100
50
0
200
300
400
500
600
700
Total length (mm)
Figure 3.1. Predicted relationship between total mass and total length and between liver
mass and total length for males, all females and small females. Values for parameters are
given in Table 3.2.
37
Table 3.2. Estimated parameters (and standard error) for the relationship between total
mass (TM) and total length (TL) and between liver mass (LM) and total length (TL) for
males, all females and small females, derived from the equation TM = a c TLb and
LM = a c TLb.
a (s.e. range)
b (± s.e.)
c
n
r2
Males
2.15 (1.44–3.20) × 10–6
3.124 (0.07)
1.003
205
0.91
All females
2.54 (2.18–2.96) × 10–7
3.482 (0.03)
1.005
699
0.97
Small females
8.09 (5.76–11.40) × 10–7
3.290 (0.06)
1.006
297
0.92
Males
7.15 (1.34–38.20) × 10–8
3.257 (0.28)
1.034
196
0.41
All females
1.05 (0.59–1.87) × 10–8
3.587 (0.09)
1.065
679
0.69
Small females
1.03 (0.41–2.63) × 10–11
4.743 (0.16)
1.031
287
0.76
Shark group
TM–TL
LM–TL
a and b are parameters, c is the (Beauchamp and Olson 1973) correction factor for
logarithmic transformation, n is sample size and r2 is square of correlation coefficient.
38
1.0
all females
CMP
0.8
0.6
0.4
0.2
0
0.20
small females
LMP
0.16
0.12
0.08
0.04
0
200
300
400
500
600
700
Total length (mm)
Figure 3.2. Relationship between carcass mass as a proportion of total mass (CMP) and
total length (TL) for all females and between liver mass as a proportion of total mass
(LMP) and TL for small females with 95% confidence intervals around the mean (– – –)
and 95% predicted intervals around the data (…..….). All females: CMP = 0.758 (0.01) –
3.723 (0.21) × 10–4 TL, n = 660, r2 = 0.32, and small females: LMP = –0.049 (0.01) +
3.211 (0.28) × 10–4 TL, n = 279, r2 = 0.32.
39
3.5 DISCUSSION
There were no sex or size effects in the FL–TL, CL–TL, CM–TL, PFM–TL, CFM–TL, and
CM–CL relationships. These length and mass measures reflect structural size and somatic
growth with little trade-off between somatic and reproductive growth. Otherwise, the
relatively larger increase in reproductive tissue experienced by adult female sharks (e.g.
Yano 1995) would be coupled with a decrease in their somatic tissue, particularly carcass
mass, expecting differences in the CM–TL and CM–CL relationships of all females
compared with males or small females. Hence, for measurements that only reflect somatic
growth (e.g. partial lengths, fin masses), comparing different ranges of size had no effect
on the relationships between these variables and TL.
Total mass and LM reflect somatic growth and reproductive investment. As the costs of
reproduction are different between males and females (Stearns 1992), different outcomes
can be expected when testing for differences between sexes if different ranges of size are
compared. This is of particular concern for species that exhibit sex and size segregation,
such as S. megalops (Graham 2005; Chapter 2), for which the full size range of the
population may not be adequately represented. Male and small female S. megalops had a
similar TM–TL relationship, but this relationship was different from the TM–TL
relationship of all females. Thus, if sampling is biased towards particular size-classes due
to size-selectivity of the sampling gear or size or sex segregation of sharks, comparisons
between sexes may not reflect real differences or similarities in the TM–TL relationship.
Hence, given the opportunistic sampling nature of most biological studies of sharks and the
small sample size of many studies, care must be taken when determining mass–length
relationships. If the size range is not fully represented, mass–length relationships may be
biased, affecting predictions of population assessments that use these relationships as
inputs to the models. Likewise, the LM–TL relationship of S. megalops differed between
the sexes depending on the ranges of size compared. For some squalid sharks ( Yano 1995;
Clarke et al. 2001) and other elasmobranchs (e.g. Craik 1978), size of liver varies with
reproductive stage, being relatively smaller for pregnant females. Liver lipid reserves are
used for vitellogenesis (Craik 1978); hence, an increase in liver lipids and liver mass is
expected for females entering first maturation. This was reflected in the larger slope of the
LM–TL relationship and the increase in LMP with TL for small females. This pattern was
obscured when small and large females were pooled as no trend was observed for this
relationship when using all females and also no differences were found in the LM–TL
40
relationship of males and all females. Therefore, the relationships between variables that
reflect somatic growth and reproductive dynamics and TL are affected by the ranges of
size used.
Most life-history parameters used in shark stock and demographic assessments are
determined as a function of TL or TM (e.g. maturity and maternity ogives, fecundity). Also
some shark fisheries use minimum and maximum size limits to regulate the catch.
However, commercial shark species are normally beheaded, eviscerated and finned so only
the mass and length of the carcass or the mass of the fins can be recorded after landing. It
is, therefore, essential to determine how these partial lengths and masses can be converted
to TL or TM (FAO 2000b). When measurements reflect only somatic growth, conversion
factors to TL or TM can be determined by pooling sexes and sizes, but for measurements
that reflect both somatic and reproductive growth, conversion factors should be determined
for sexes and sizes separately. Although many studies provide TM–TL relationships, few
present conversion factors to allow calculating TL or TM from partial length or partial
mass measures. Geometric mean regressions were used to determine conversion factors for
several length–length and mass–mass relationships for S. megalops. These conversion
factors are essential for assessment of this species. Given the depletion of many of the
harvested species of sharks and a decline in abundance of most other shark species in
southern Australia (Graham et al. 2001), S. megalops will inevitably become a more
sought after species.
41
Some of the prey items found in the stomach of Squalus megalops (photos by the author).
42
CHAPTER 4
43
CHAPTER 4
SOURCES OF VARIATION IN THE FEEDING ECOLOGY OF THE PIKED
SPURDOG (SQUALUS MEGALOPS): IMPLICATIONS FOR INFERRING
PREDATOR–PREY INTERACTIONS FROM OVERALL DIETARY
COMPOSITION
4.1 ABSTRACT
Sources of variation in dietary composition were examined in the piked spurdog (Squalus
megalops). This species is an opportunistic predator that consumed a wide range of prey
items. When importance of prey was measured by weight or occurrence, S. megalops
preyed largely on molluscs and teleosts. However, when number of prey was considered,
the main items were crustaceans. A bootstrap analysis showed that considerable variability
can be expected in the importance of prey items in the species’ overall diet. Regional,
seasonal and ontogenetic differences in dietary composition were found, but there were no
differences between mature and immature sharks or between males and females. The
spatial and temporal variation in diet exhibited by S. megalops and the intrinsic natural
variability of the dietary composition of this opportunistic predator suggest that studies that
infer predator–prey interactions from overall diet are likely to miss information on the
ecological relationships among species and thus account for only part of these interactions.
4.2 INTRODUCTION
The feeding ecology of marine animals has been studied to determine the ecological roles
and position of animals within foodwebs and to understand predator–prey interactions
(Caddy and Sharp 1986; Pauly et al. 1998; Cortés 1999). Interactions among species affect
population dynamics and also cause indirect ecological effects (Alonzo et al. 2003).
Hence, if interactions among species were determined, ecosystems could be managed with
higher certainty (Yodzis 1994). Traditional single-species fishery management ignores
fishery impacts on ecosystems (Agardy 2000). As an alternative, ecosystem-based fishery
management has been proposed to account for such impacts (Gulland 1978; Caddy and
Sharp 1986; Fulton et al. 2003). Many ecosystem models use dietary information as a
proxy for the interactions among species (e.g. Christensen 1995; Walters et al. 1997;
45
Yodzis 1998). However, most models use overall diet data, ignoring many sources of
variation that can affect the dietary composition of predators.
Natural systems are dynamic and vary in time and space (Paine 1988). It is, therefore,
expected that diet of predators, and hence predator–prey interactions, may also vary in time
and space. Trophic interactions are determined by the size of predators and their prey
(Floeter and Temming 2003), but little is known about predator–prey size relationships of
large marine predators such as sharks. Also for sharks, the effects of time and space and
their interactions with other potential sources of variation in their diet, such as sex or
maturity condition, have been little studied. Although some studies have reported regional,
seasonal, or ontogenetic differences in diet (see Wetherbee and Cortés 2004, for a review),
many studies on the diet of sharks have been limited to simple lists of prey items (Heithaus
2004). Moreover, variation in diet has often been reported qualitatively with little statistical
support (Ferry and Cailliet 1996; Cortés 1997; Wetherbee and Cortés 2004). Hence, a more
rigorous and quantitative approach is required to study the feeding ecology of sharks.
The piked spurdog (Squalus megalops) is a suitable species to test for the effects of
potential sources of variation in the dietary composition of predators, as it is a very
abundant shark in southern Australia (Jones 1985; Bulman et al. 2001; Graham et al.
2001). Squalus megalops inhabits waters of the continental shelf and upper continental
slope to 510 m (Last and Stevens 1994). Off South Africa, females grow larger (782 mm
total length, TL) than males (572 mm TL) and attain 50% maturity at 15 years, and 50% of
males are mature at 9 years old (Watson and Smale 1999). Given its high natural
abundance, which has remained stable since it was first surveyed (Graham et al. 2001), S.
megalops is a dominant and ecologically important species (Bulman et al. 2001) that is
likely to make an important contribution to the structure and function of an ecosystem.
Nevertheless, information on its feeding habits is scarce. Its overall diet has been described
for animals caught off South Africa and eastern Australia, where it preys mainly on
teleosts and cephalopods, but also consumes crustaceans and elasmobranchs (Bass et al.
1976; Ebert et al. 1992; Bulman et al. 2001). Although those studies offer a preliminary
description of the diet of this shark, more quantitative analyses are needed.
The purpose of this study was to investigate the effects of several sources of variation in
the feeding ecology of S. megalops. The specific objectives were to: (i) quantify its overall
46
dietary composition and account for how much variability would be expected when
calculating overall prey importance; (ii) examine relationships between prey and predator
size; and (iii) test for the effects of region, maturity condition, sex, season and ontogenetic
variation on its dietary composition.
4.3 MATERIALS AND METHODS
Sampling
Squalus megalops were obtained from the by-catch of shark and trawl vessels operating in
the Australian Southern and Eastern Scalefish and Shark Fishery (Fig. 4.1). Samples were
collected monthly between October 2002 and April 2004, with the exception of the period
July–September (Table 4.1), when S. megalops seems to move off the fishing grounds and
weather conditions restricted sampling. The specimens were sexed, measured (TL ±1 mm)
and weighed on an electronic balance (±0.1 g). Maturity of males was determined on the
basis of clasper calcification, condition of testes and vas efferens, and presence of semen in
seminal vesicles. Maturity of females was determined on the basis of the condition of
oviducal glands and ovarian follicles, and the presence of in utero eggs or embryos.
Diet and data analyses
Diet was studied by prey identification and analysis of stomach contents. The stomach of
each fish was removed, and the contents were identified to the lowest taxon practical.
When possible, to correlate size of prey and predator, body width (BW) of worms, TL of
fish, mantle length (ML) of cephalopods, and shield length (SL) of hermit crabs were
measured to the nearest millimetre. Where these lengths could not be measured, TL of fish,
ML of cephalopods, and SL of hermit crabs were estimated from hard tissue pieces found
in stomach contents by linear and allometric relationships determined by regression, using
a personal reference collection and the fish and crustacean reference collections of the
South Australian Museum, Australia, and Museum Victoria, Australia. Prey items that
digest more speedily than other prey items or soft-bodied prey may be under-represented if
the more persistent hard parts are included in the analyses (Bigg and Fawcett 1985; Bigg
and Perez 1985). Hence, hard parts (e.g. beaks, vertebrae, chelipeds) were only used for
estimating prey item size and describing the overall dietary spectrum, but they were
excluded from further analyses.
47
129º E
141º E
150 º E
WWP
EWP
NSW
200 m
Ulladulla
t
En
m
ro
nc
ra
e
ow
H
pe
Ca
e
s
ke
La
Robe
sP
on
ils
W
•
37º S
t
on
•
•
•
or
y
•
Australia
41º S
Tasmania
N
200
0
200 Miles
Figure 4.1. Map of sampling area showing the three biogeographic regions and ports –
west of Wilsons Promontory (WWP), east of Wilsons Promontory (EWP), New South
Wales (NSW).
48
Taxonomic classification of prey items does not account for differences in habitat
utilization of a predator. Therefore, data analyses were carried out by main zoological
group (Polychaeta, Sipuncula, Crustacea, Mollusca, Chondrichthyes, Teleostei) and
ecological group separately. The ecological groups considered were benthic infauna (prey
species living in the sediment), benthic epifauna (prey species living on the sediment
surface), benthic (prey species living on the bottom), demersal benthic (prey species living
near the bottom but not linked to it), demersal pelagic (prey species with extensive diel
vertical migration), and pelagic (prey species living in the upper layers of the water
column).
Overall diet
Stomach fullness (SF) and number of prey found in each stomach were recorded to
determine the feeding pattern of S. megalops. Stomach fullness was recorded using a
quarterly scale (0, empty; 1, 0–25% filled; 2, 26–50% filled; 3, 51–75% filled; 4, 76–100%
filled). Chi-square tests with Yates’ continuity correction (Zar 1999) were used to test for
differences in the distribution of SF.
To obtain a precise description of the overall diet of a predator, it is important to determine
the minimum number of stomachs required (Ferry and Cailliet 1996; Cortés 1997). The
number of S. megalops collected was tested to determine whether sufficient sharks were
sampled. Items such as sponges, hydroids, and algae were considered incidental, and were
excluded from the analysis. The cumulative number of randomly pooled stomachs was
plotted against the cumulative diversity of stomach contents. Diversity was calculated
using the pooled quadrat method based on the Brillouin Index of diversity (HZ; Pielou
1966). To ensure that curves reached an asymptotic value, 10 random orders of stomachs
(curves) were calculated (Koen Alonso et al. 2002). Diversity curves were considered
asymptotic if at least two previous values to the total sample diversity were in the range of
asymptotic diversity ±0.05 (Koen Alonso et al. 2002). Diversity curves were calculated for
each combination of factors considered in the analyses of variation in dietary composition.
49
Table 4.1. Sampling sites (see Fig. 4.1), collection time, and sample sizes collected for the
spatial, temporal, ontogenetic, maturity condition, and sexual components of the study
(sample sizes for the analyses may be smaller because of the occurrence of empty
stomachs).
Factor
Sample
Site
Collection time
West Wilsons Promontory (WWP)
Robe
Autumn 2004
36
East Wilsons Promontory (EWP)
Lakes
Autumn 2004
60
Ulladulla
Autumn 2004
41
Lakes
December 2002,
116
Entrance
February 2003, 2004
size
Spatial (large females 471 mm TL)
Entrance
New South Wales (NSW)
Temporal (seasonal)
Summer
Autumn
March 2003, 2004,
98
April, May 2003
Winter
June 2003
24
Spring
October 2002,
71
November 2003
Size (ontogenetic)*
Small male (≤400 mm TL)
Lakes
Spring 2002, 2003,
30
Small female (≤400 mm TL)
Entrance
Summer 2003, 2004,
51
Autumn 2003, 2004
92
Medium-sized females (401–470 mm TL)
Large male (401–470 mm TL)
37
Large female (471 mm TL)
100
Maturity condition
Immature
Lakes
Spring 2002, 2003,
174
Mature
Entrance
Summer 2003, 2004,
131
Autumn 2003, 2004
50
Table 4.1. Continued…
Factor
Sample
Site
Collection time
Male
Lakes
Spring 2002, 2003,
67
Female
Entrance
Summer 2003, 2004,
242
size
Sexual
Autumn 2003, 2004
*
Squalus megalops has a tendency to segregate by sex/size and this was reflected in the
size frequency distribution of some of the fishing shots analysed. Hence, the size classes
compared are based on this segregation pattern.
51
No single method of analysis of stomach contents completely describes the diet of a
predator (Hyslop 1980); hence, the importance of prey items was evaluated using
percentage weight (%W), percentage number (%N), percentage frequency of occurrence
(%FO), and percentage Index of Relative Importance (%IRI; Pinkas et al. 1971; Cortés
1997). Bootstrap methods (1000 replicates) were used to estimate confidence intervals
(2.5th and 97.5th percentiles) around the dietary parameters (mean %W, %N, %FO, and
%IRI; Haddon 2001). From the original data matrix, random samples of the observations
(i.e. each individual stomach) with replacement were generated to obtain the probability
distribution of the dietary parameter estimates for each prey item.
Predator–prey size relationship
The relationship between prey size and shark size was determined using the Spearman rank
correlation coefficient (rs). The length variables for the different taxonomic groups were
considered. Relative and cumulative frequency histograms of prey size:predator size ratios
were plotted to examine the patterns of prey size consumed by S. megalops (Bethea et al.
2004). For this latter analysis, only teleost and cephalopod prey were used.
Variation in dietary composition
Regional comparisons of diet were made for large females (471–650 mm TL) collected in
autumn (Table 4.1). A one-way non-parametric multivariate analysis of variance (NPMANOVA) using Bray–Curtis distances (Anderson 2001) on weight and number data for
sharks collected at the same time (autumn 2004) was used to test for regional effects on the
diet of S. megalops. Weight and number data were transformed to fourth root and
standardized to z-scores to minimize differences attributable to stomach size. Region was
treated as a fixed factor. Equal sample sizes were used (n = 30 for the analysis of
zoological groups, n = 28 for the analysis of ecological groups). If significant differences
were found, a posteriori pairwise comparisons were made (Anderson 2001).
Maturity condition was evaluated, and sexual, seasonal, and ontogenetic comparisons were
made on sharks collected from Lakes Entrance between October 2002 and March 2004
(Table 4.1). Non-parametric multidimensional scaling (nMDS) on Bray–Curtis similarity
measures using fourth root transformed data (Clarke 1993) were used to visualize patterns
of variation in dietary composition. Mean percentage weight and number of zoological and
ecological groups were used.
52
The relative and interactive effects of maturity condition, sex, season, and size were
evaluated in a similar way to the regional analysis using weight and number. Squalus
megalops is sexually dimorphic, females attaining larger sizes than males; hence, separate
analyses were undertaken for each sex to investigate the effects of maturity condition on
dietary composition. The effects of maturity condition (mature, immature) and season
(summer, autumn, spring; Table 4.1) were investigated using individuals within the 382–
406 and 433–509 mm TL range for males and females, respectively. These ranges covered
the sizes of the smallest mature and largest immature specimen of each sex. For the
analysis of males, season was not included as a factor because of the low number of
replicates for any season, except summer. Hence, the analysis was done using data
collected only during the latter season. For females, maturity condition was treated as fixed
and orthogonal to the random factor season (i.e. every level of the factor “maturity”,
mature or immature, is present in every level of the factor “season”, summer, autumn, or
spring; Table 4.1). Similar sample sizes (n = 7 for males, n = 8 for females) were used for
each combination of factors.
To test for sexual, ontogenetic (size), and seasonal differences, sharks of similar size (<471
mm TL) were used in a three-way NP-MANOVA (factors: sex, size, and season). Sex
(males, females) and size (small and large males, small and medium-sized females) were
treated as fixed and orthogonal to the random factor season (summer, autumn, spring;
Table 4.1). Equal sample sizes (n = 6) were used for each combination of factors. As small
and large males and small and medium-sized females had similar diets (see “Results”),
data were pooled to test for ontogenetic and seasonal differences between small (<471 mm
TL) and large (471 mm TL) animals. A two-way NP-MANOVA (factors: size and
season) with equal sample sizes (n = 26) was used for each combination of factors. Finally,
winter samples could only be collected for small specimens, so to include winter in the
seasonal study, a one-way NP-MANOVA was undertaken for small S. megalops using a
balanced design (n = 24).
53
4.4 RESULTS
The stomach contents and fullness of 937 S. megalops were examined. In all, there were 77
small males (274–400 mm TL), 105 small females (270–400 mm TL), 129 large males
(400–470 mm TL), 193 medium-sized females (401–470 mm TL), and 433 large females
(471–650 mm TL).
Overall diet
Of the 937 stomachs examined, 603 (65.3%) contained food, from which >60% contained
a single prey item. For stomachs with >1 item, the number of prey items ranged from two
to ten. For stomachs with prey, the distribution of stomach fullness was relatively even
(~25%) and there were no significant differences among the frequency of individuals in
each SF category ( 2 = 2.150, n = 603, P = 0.542).
Of the 603 stomachs with food, 111 were excluded because they contained only hard parts,
sponges, hydroids, algae, or unidentified material. The prey diversity curve for the overall
diet reached a stable level at about 350 stomachs (Fig. 4.2a), so the sample size of 492 was
large enough to describe the overall diet of S. megalops.
The stomachs contained 107 taxonomic levels of prey item: six polychaetes, two
sipunculids, 29 crustaceans, 17 molluscs, 47 fish, remains of sea lion, and other items such
as echiurids, algae, sponges, hydroids and brittle stars (Appendix 4.a). Arrow squid (family
Ommastrephidae) was the dominant prey item, contributing the highest values of %W
(20.03%), %N (7.54%), %FO (8.76%), and %IRI (32.05%). Octopus (Octopus spp) was
the second most important prey item by weight (12.55%), frequency of occurrence (7.66%)
and relative importance (19.37%). The third major prey was fish of the family Triglidae
(gurnards) in terms of weight (9.77%), number (5.33%), frequency of occurrence (5.97%)
and relative importance (12.00%). Shrimps (Caridea) and hermit crabs (Diogenidae) were
important by number (6.88% and 5.90%, respectively), but not in terms of weight or
occurrence.
A similar pattern was observed when data were analysed by main zoological group
(Appendix 4.a). Molluscs were the most important item by weight (56.43%), frequency of
occurrence (35.89%), and relative importance (50.31%). However, the most numerous
54
items were crustaceans (31.61%). Teleosts were the second most important item in terms
of weight (38.32%), frequency of occurrence (34.03%), and relative importance (37.27%).
When data were analysed by ecological group, the most important group by weight was
demersal pelagic prey (40.25%), followed by benthic (36.95%), and demersal benthic
(11.04%) prey (Appendix 4.a). In contrast, benthic epifauna dominated by number
(41.15%) and frequency of occurrence (29.41%), followed by benthic prey (21.10% by
number and 25.35% by frequency of occurrence). Finally, for %IRI, the main ecological
group was benthic prey (33.96%), followed by benthic epifauna (30.70%), demersal
pelagic (26.52%) and demersal benthic (6.27%) prey. Pelagic and benthic infauna were
less important.
Irrespective of analysing prey items by zoological or ecological group, considerable
variability was found around the estimation of overall mean prey importance (Appendix
4.a). For important prey such as molluscs or teleosts, there was ~20% of variability within
the upper and lower 95% confidence intervals. However, for less important prey such as
crustaceans, variability was ~50%. When the mean values obtained from bootstrapping
were compared with those obtained from point estimates of overall diet, variability ranged
from 1–14% (not shown). A similar pattern was observed for ecological groups.
Predator–prey size relationship
Squalus megalops consumed prey of a wide range of sizes (Fig. 4.3). More than 60% of
teleosts and cephalopods consumed were less than 30% and 24% of S. megalops total
length (TL), respectively, but S. megalops also consumed fish and cephalopods up to 60%
of its TL.
No correlation was found between predator TL and shield length of hermit crabs (rs =
0.119, n = 65, P >0.05), TL of teleosts (rs = 0.157, n = 39, P >0.05), or body width of
worms (rs = 0.273, n = 14, P >0.05). However, there was a positive correlation between
predator TL and mantle length of cephalopods (rs = 0.455, n = 43, P <0.05) (Fig. 4.3).
55
Diversity (HZ)
6
6
(a)
5
5
4
4
3
3
2
2
1
1
0
0
0
6
100
200
300
400
0
500
6
(b)
5
5
4
4
3
3
2
2
1
1
0
0
0
(c)
10
20
30
40
10
20
30
40
10
20
30
40
(d)
0
Number of stomachs
Figure 4.2. Cumulative diversity (HZ) of prey items for (a) the overall diet of S. megalops
and for the three regions analysed: (b) west of Wilsons Promontory, (c) east of Wilsons
Promontory, and (d) New South Wales. The straight lines indicate the range of asymptotic
diversity ±0.05.
56
16
100
12
75
8
50
4
25
0
0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Finfish TL/predator TL
100
16
(b)
12
75
8
50
4
25
0
0
Cumulative frequency (%)
Relative frequency (%)
(a)
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Cephalopod ML/predator TL
Cephalopod ML (mm)
400
(c)
300
200
100
0
300
400
500
600
700
Predator TL (mm)
Figure 4.3. Changes in prey size with predator size. Distribution of prey size:predator size
ratios for (a) teleosts and (b) cephalopods. (c) Relationship between cephalopod mantle
length (ML) and predator total length (TL) and 95% confidence limits. ML = 0.6894 TL –
218.68; r2 = 0.37. Open bars = relative frequencies at 0.02 intervals. Filled circles =
cumulative frequencies at 0.02 intervals.
57
Variation in dietary composition
Prey diversity for sharks collected from WWP (~3.10; Fig. 4.2b) and NSW (~2.64; Fig.
4.2d) was lower than for sharks from EWP (~3.85; Fig. 4.2c), suggesting a more diverse
diet at EWP. The prey diversity curves reached a stable level for each of the three regions
analysed, indicating that the sample was large enough to describe the diet of sharks from
each region.
There was a regional pattern in the diet of S. megalops (Table 4.2). Significant differences
in dietary composition were found between sharks collected from WWP and EWP,
irrespective of the use of weight or number of zoological or ecological groups (Fig. 4.4;
pairwise comparisons). Significant differences were also found between sharks collected
from WWP and NSW when weight of zoological group and weight or number of
ecological group were used (Fig. 4.4; pairwise comparisons). No differences were found
between the diets of sharks collected from EWP and NSW (Fig. 4.4; pairwise
comparisons). For EWP and NSW, S. megalops consumed mainly teleosts, molluscs and
crustaceans, and also small amounts of worms and chondrichthyans for EWP. However,
for WWP, S. megalops preyed largely on molluscs and, to a lesser extent, teleosts. For
ecological groups, S. megalops collected from WWP preyed mostly on demersal pelagic
prey, whereas those collected from EWP and NSW preyed mostly on benthic organisms.
Most prey diversity curves (not shown) showed asymptotes or trends towards an asymptote
for each combination of maturity condition and season. Irrespective of the use of weight or
number of a zoological or an ecological group, there were no significant differences in
dietary composition between immature and mature S. megalops (Table 4.3). Therefore,
immature and mature sharks were pooled for subsequent analyses.
Most prey diversity curves (not shown) showed asymptotes or trends towards an asymptote
for each combination of sex, size and season. A significant seasonal pattern in the dietary
composition of S. megalops was found for the three-way analysis, but there were no sexual
or ontogenetic differences (Table 4.4). Therefore, both sexes and sizes (small and large
males, and small and medium-sized females) were pooled for subsequent analyses.
58
Table 4.2. NP-MANOVA testing for the effects of region (east of Wilsons Promontory,
west of Wilsons Promontory, New South Wales) on the weight and number of zoological
(Polychaeta, Sipuncula, Crustacea, Mollusca, Chondrichthyes, and Teleosts) and
ecological groups (benthic infauna, benthic epifauna, benthic, demersal benthic, demersal
pelagic, and pelagic) in the diet of S. megalops.
Weight
Factor
Number
d.f.
F
P
F
P
Region
2
4.965
<0.001
3.580
0.011
Residual
87
Zoological group
Ecological group
Region
2
Residual
81
6.561
<0.001
6.990
<0.001
59
2.0
1.00
(a)
WWP
EWP
NSW
1.5
0.75
1.0
0.50
0.5
0.25
0.0
BE BI BN DB DP PE
2.0
(b)
Mean number
Mean weight (g)
(c)
0.0
BE BI BN DB DP PE
1.00
(d)
1.5
0.75
1.0
0.50
0.5
0.25
0.0
0.0
PO SI CR MO CH TE
PO SI CR MO CH TE
Prey category
Figure 4.4. Main prey groups found in the diet of S. megalops collected from west of
Wilsons Promontory (WWP), east of Wilsons Promontory (EWP), and New South Wales
(NSW). Mean weight of fourth root transformed data (±s.e.) of prey sorted by (a)
ecological and (b) zoological group, and mean number of fourth root transformed data
(±s.e.) of prey sorted by (c) ecological and (d) zoological group. BE, benthic epifauna; BI,
benthic infauna; BN, benthic; DB, demersal benthic; DP, demersal pelagic; PE, pelagic.
PO, polychaetes; SI, sipunculids; CR, crustaceans; MO, molluscs; CH, chondrichthyans;
TE, teleosts.
60
Table 4.3. NP-MANOVA testing for the effects of maturity condition (mature, immature)
and season (summer, autumn, spring), females only, on the weight and number of
zoological (Polychaeta, Sipuncula, Crustacea, Mollusca, Chondrichthyes, and Teleosts)
and ecological groups (benthic infauna, benthic epifauna, benthic, demersal benthic,
demersal pelagic, and pelagic) in the diet of male and female S. megalops.
Zoological group
Weight
Factor
d.f.
F
P
Ecological group
Number
F
P
Weight
F
P
Number
F
P
Male
Maturity
1
Residual
12
1.045 0.394
1.094
0.391 0.488
0.730 0.543 0.738
Female
Maturity
1
3.122 0.080
3.551
0.080 1.779
0.307 1.742 0.289
Season
2
1.843 0.077
1.800
0.107 1.374
0.190 1.370 0.206
Maturity × season
2
0.559 0.804
0.471
0.827 0.460
0.912 0.385 0.930
Residual
42
61
Table 4.4. NP-MANOVA testing for the effects of sex (male, female), size (small and
large for males, small, medium-sized, and large for females), and season (summer, autumn,
spring) on the weight and number of zoological (Polychaeta, Sipuncula, Crustacea,
Mollusca, Chondrichthyes, and Teleosts) and ecological groups (benthic infauna, benthic
epifauna, benthic, demersal benthic, demersal pelagic, and pelagic) in the diet of S.
megalops.
Zoological group
Weight
Factor
d.f.
F
Ecological group
Number
P
Weight
Number
F
P
F
P
F
P
3-way analysis for sharks <471 mm TL
Sex
1 0.505
0.619
0.508
0.602
2.622
0.190
2.764
0.189
Size
1 0.741
0.520
0.667
0.537
3.938
0.120
4.205
0.120
Season
2 5.130 <0.001 5.579
<0.001 3.963 <0.001
4.378
<0.001
Sex × size
1 0.315
0.782
0.211
0.841
1.161
0.390
1.085
0.425
Sex × season
2 0.750
0.582
0.826
0.516
0.583
0.817
0.569
0.804
Size × season
2 1.980
0.077
2.172
0.062
0.565
0.841
0.545
0.840
2 2.227
0.051
2.245
0.057
1.660
0.094
1.693
0.098
6.557
0.017
Sex × size ×
season
Residual
60
2-way analysis for all sizes including large females (471 mm TL)
Size
1 4.407
Season
2 6.935 <0.001
6.983 <0.001
8.978 <0.001
8.220 <0.001
Size × season
2 6.051 <0.001
7.826 <0.001
1.984
2.495
Residual
62
150
0.059
4.189
0.094
5.157
0.018
0.048
0.017
Prey diversity curves for each size–season combination reached a stable level and had
similar values of diversity, except for small sharks collected in winter that showed lower
values (Fig. 4.5). After including all sizes in the analysis, significant ontogenetic and
seasonal effects were detected. Also, a significant interaction between size and season was
found for weight and number of prey items for both zoological and ecological groups
(Figs. 4.6, 4.7; Table 4.4). The ordination showed two separate groups when zoological
data and ecological number data were used (Fig. 4.6). Large sharks tended to be separated
from small ones, and samples collected in summer and autumn were separated from those
collected in spring. However, no clear visual pattern was observed when the analysis was
done for ecological groups using weight data. Large and small S. megalops had different
diets in summer and autumn but similar diets in spring (Fig. 4.7; pairwise comparisons). In
summer and autumn, large sharks consumed mainly molluscs, whereas small sharks
consumed mainly crustaceans. For ecological groups in summer, large S. megalops preyed
mainly on demersal pelagic prey whereas small sharks preyed on benthic organisms. In
spring, both size classes had a similar feeding pattern, consuming mainly teleosts, followed
by molluscs and crustaceans. By ecological group, large and small sharks collected in
spring preyed mainly on benthic organisms. When winter was included in the seasonal
analyses, the seasonal pattern was similar (Table 4.5). Dietary composition in summer,
autumn and winter was similar, but significant differences were observed among these
three seasons and spring (pairwise comparisons).
63
5
5
4
4
3
3
2
2
1
1
Large-summer
Small-summer
0
Diversity (HZ)
0 10 20 30 40 50 60 70 80
0
5
5
4
4
3
3
2
2
1
0
10
20
30
40
1
Large-autumn
Small-autumn
0
0
0
10 20 30 40 50 60
0
10
20
30
40
5
4
3
2
1
Small-winter
0
0
10
20
30
5
5
4
4
3
3
2
2
1
1
Small-spring
Large-spring
0
0
10
20
30
40
50
0
0
10
20
30
Number of stomachs
Figure 4.5. Cumulative diversity (HZ) of prey items for small and large sharks from each
season. The straight lines indicate the range of asymptotic diversity ±0.05.
64
(a)
(c)
Stress 0.03
(b)
Stress 0.01
SSSu
LSSp
LSAu
LSSu
(d)
Stress 0.04
SSSp
SSAu
Stress 0.02
Figure 4.6. Non-parametric multidimensional scaling (nMDS) ordination of the stomach
contents of small sharks from spring (SSSp), autumn (SSAu), and summer (SSSu), and
large sharks from spring (LSSp), autumn (LSAu), and summer (LSSu). Mean %W of (a)
ecological and (b) zoological group, and mean %N of (c) ecological and (d) zoological
group.
65
2
2
(a)
(g)
summer
1
1
0
0
BE BI BN DB DP PE
2
BE BI BN DB DP PE
2
(b)
(h)
autumn
1
autumn
1
0
0
BE BI BN DB DP PE
BE BI BN DB DP PE
2
2
(c)
(i)
spring
spring
1
0
BE BI BN DB DP PE
2
(d)
summer
Mean number
1
Mean mass (g)
summer
0
BE BI BN DB DP PE
Large
Small
2
(j)
summer
1
1
0
0
PO SI CR MO CH TE
2
PO SI CR MO CH TE
2
(e)
autumn
(k)
autumn
1
1
0
0
PO SI CR MO CH TE
2
PO SI CR MO CH TE
2
(f)
spring
(l)
spring
1
1
0
0
PO SI CR MO CH TE
PO SI CR MO CH TE
Prey category
Figure 4.7. Size and seasonal effects in the diet of large and small S. megalops caught in
summer, autumn, and spring. Mean weight of fourth root transformed data (±s.e.) of prey
sorted by ecological (a, b and c) and zoological group (d, e and f), and mean number of
fourth root transformed data (±s.e.) of prey sorted by ecological (g, h and i) and zoological
group (j, k and l). BE, benthic epifauna; BI, benthic infauna; BN, benthic; DB, demersal
benthic; DP, demersal pelagic; PE, pelagic. PO, polychaetes; SI, sipunculids; CR,
crustaceans; MO, molluscs; CH, chondrichthyans; TE, teleosts.
66
Table 4.5. NP-MANOVA testing for the effects of season (summer, autumn, winter,
spring) on the weight and number of zoological (Polychaeta, Sipuncula, Crustacea,
Mollusca, Chondrichthyes, and Teleosts) and ecological groups (benthic infauna, benthic
epifauna, benthic, demersal benthic, demersal pelagic, and pelagic) in the diet of small
(<471 mm TL) S. megalops.
Zoological group
Weight
Factor
Season
d.f.
F
P
Ecological group
Number
F
P
3 6.274 <0.001 7.366 <0.001
Weight
F
Number
P
4.314 <0.001
F
P
5.483 <0.001
Residual 92
67
4.5 DISCUSSION
Dietary studies of sharks commonly report a high proportion of empty stomachs and few
prey items per stomach, most of them in advanced stages of digestion (Wetherbee et al.
1990; Ebert et al. 1992; Simpfendorfer et al. 2001a). Therefore, many shark species are
considered intermittent feeders. For these species, short periods of active feeding are
followed by longer periods of reduced predatory activity (Wetherbee et al. 1990;
Wetherbee and Cortés 2004). The present study supports this hypothesis. Almost 35% of
stomachs examined were empty, and for stomachs with prey, >60% contained a single prey
item, suggesting that feeding is intermittent. However, further research on the feeding
duration, total digestion time, and gastric evacuation rates using captive S. megalops would
allow estimates of feeding frequency and feeding periodicity.
There was a wide range of food items in the stomachs of S. megalops, which meant that
many stomachs were needed to describe overall diet. When diversity curves have been
used to determine the sample size required for a precise description of the diet of sharks,
most studies have found stable levels of diversity at <200 stomachs sampled (Carrassón et
al. 1992; Gelsleichter et al. 1999; Koen Alonso et al. 2002; Morato et al. 2003; Bethea et
al. 2004). However, prey diversity was high for S. megalops, and at least 350 stomachs had
to be sampled to describe its overall diet. Squalus megalops can be considered a generalist
and opportunistic feeder given that portions of large teleosts, cephalopods, and sharks were
found in many stomachs, and that they consumed abundant prey such as arrow squid
(Triantafillos et al. 2004) and gurnards (Triglidae; M. Gomon, pers. comm.). Other studies
also suggest sharks are generalist and opportunistic feeders that consume the most
abundant prey (Wetherbee et al. 1990; Hanchet 1991; Ellis et al. 1996; Koen Alonso et al.
2002).
Overall, results differed when average prey importance was analysed using weight,
number, or frequency of occurrence of prey groups. If importance of prey is to be deduced
on the basis of weight or frequency of occurrence, S. megalops preyed largely on molluscs
and teleosts. However, if number of prey is to be used, the main items were crustaceans.
Analyses done by ecological group showed that S. megalops was a versatile predator that
used a wide range of habitats. The most important items by weight were demersal pelagic
and benthic prey, whereas benthic epifauna and benthic prey were the most consumed
items by number and occurrence. Therefore, number, weight, and frequency of occurrence
68
measures provided different information on feeding habits (MacDonald and Green 1983;
Bigg and Perez 1985; Cortés 1998b). Ferry and Cailliet (1996) suggest using multiple
measures when prey items differ in size. For generalist and opportunistic feeders that
consume a wide range of prey, like S. megalops, the use of multiple measures allows a
better representation of overall diet.
Irrespective of which diet descriptor was used, the bootstrap analysis showed a wide range
of variability around the estimate of overall importance of prey. In general, studies on the
diet of sharks obtain samples opportunistically, and in many cases small sample sizes are
collected. However, as sharks are considered opportunistic predators (Wetherbee et al.
1990), large sample sizes would be needed for a comprehensive description of diet. Also,
many studies have reported a high proportion of empty stomachs (Wetherbee et al. 1990),
and some studies only described diet in terms of number or occurrence of prey, whereas
other studies only used weight. However, for S. megalops, number, occurrence, and weight
of prey showed different patterns of importance of prey. Therefore, a combination of small
sample size, high proportion of empty stomachs, the use of different descriptors of
importance of prey, and the opportunistic predatory nature of many shark species, is likely
to result in high variability in the dietary composition and hence in evaluation of predator–
prey interactions. Accurate characterization of predator–prey interactions inferred from
diet data is crucial for ecosystem-based models and in their increasing use as tools for
fisheries management. However, if overall diet data do not incorporate a measure of the
natural variability in dietary composition exhibited by many shark species, predatory
interactions and hence model predictions may be misleading. For example, if overall diet
data are used to describe the predatory relationships of S. megalops in southern Australia,
the main interactions will be with molluscs, in terms of %W, or with crustaceans, in terms
of %N. However, the main interactions will be with teleosts, if sampling is done only in
spring, or with molluscs, if only large sharks are collected in summer and autumn, or with
crustaceans, if only small sharks are collected in summer and autumn. The same pattern of
variability is reported for other shark species. Simpfendorfer et al. (2001b) compared the
diet of tiger sharks (Galeocerdo cuvier) from four sites off Western Australia. Overall, the
main predatory interactions by %FO were with turtles, teleosts, and sea snakes. However,
for one site, North West shelf, the interactions with teleosts and sea snakes were not as
important as with dugongs, and for another site, Ningaloo, tiger sharks interact almost
exclusively with turtles. The observed variability in the diet of sharks is particularly
69
relevant when using overall diet data as a descriptor of their predator–prey interactions,
because the use of overall data may obscure site-, size-, or sex-specific interactions. Also,
given that ecosystem-based models tend to use %W data from overall diet as inputs, the
occurrence of a few heavy prey items, for example, may overestimate the importance of
the interaction between the predator and those particular prey, and underestimate the
importance of the interactions with other prey.
Size-dependent predation can regulate population and community level dynamics (Brooks
and Dodson 1965), but size-selective feeding has been little studied in sharks. In the
present study, S. megalops preyed on a wide range of prey sizes (4–60% of its TL) and,
except for cephalopod items, the total length of S. megalops was not correlated with size of
prey. Other studies found that shark diets consisted of relatively small prey (in most cases,
<36% of the sharks TL), and that prey size was correlated to predator size (Cortés et al.
1996; Scharf et al. 2000; Bethea et al. 2004). However, the present study showed that S.
megalops had little size preference for prey, supporting the belief that this shark is a
generalist and opportunistic predator.
Predation can be highly variable in space and time (Bax 1998). There was regional,
seasonal, and ontogenetic variation in the diet of S. megalops, and this pattern was
consistent despite analyses being conducted on weight or number of zoological or
ecological prey groups. Variation was not explained by the effects of sex or maturity
condition, but this could be due to the low number of replicates for each combination of
factors (e.g. n = 6 for the sex × size × season analysis), and hence low statistical power
(Ferry and Cailliet 1996). Some authors have found differences in the diet of sharks
between sexes (Hanchet 1991; Stillwell and Kohler 1993; Simpfendorfer et al. 2001b;
Koen Alonso et al. 2002) and maturity condition (Koen Alonso et al. 2002). However,
some of these studies may have confounded the effects of sex or maturity condition with
other factors such as space and time because, although samples were obtained
opportunistically across a wide spatial and temporal scale, space and time were not
considered in the analyses.
Feeding plasticity of sharks results in regional, seasonal, and ontogenetic variation in diet
that complicates an accurate description of their feeding ecology (Wetherbee and Cortés
2004). However, most studies on the feeding ecology of sharks have described only overall
70
dietary composition. Some studies have reported regional, seasonal, or ontogenetic
variation (Jones and Geen 1977a; Lyle 1983; Laptikhovsky et al. 2001; Simpfendorfer et
al. 2001a; Ebert 2002), but most of them have done so qualitatively (Wetherbee and Cortés
2004). When a quantitative approach was taken (Cortés et al. 1996; Simpfendorfer et al.
2001b; Vögler et al. 2003; White et al. 2004), region, season, or ontogeny were evaluated
independently of each other even though samples were collected across wide spatial and
temporal scales. When sampling is opportunistic across wide spatial and temporal scales, if
the interactive effects of space and/or time are not considered, it is likely that differences in
diet attributed to a certain factor (e.g. size) are unknowingly confounded by the effects of
other factors (e.g. region) not included in the analysis. Furthermore, if a factor is analysed
independently but many factors are involved, the analysis should, at least, be undertaken
on standardized data to remove the effects of the other factors not considered.
Standardized data for the effects of season, sex, and size showed regional variation in the
diet of large females collected in autumn. Sharks from WWP fed largely on demersal
pelagic prey (mainly ommastrephid squid), but those from EWP and NSW had a more
varied diet, also consuming benthic prey (teleosts and crustaceans). A demersal pelagic
diet implies that a demersal shark such as S. megalops undergoes vertical feeding
migrations to exploit pelagic prey such as squid or preys on squid while aggregated near
the seabed (Roper and Young 1975). These findings suggest that S. megalops would have
different patterns of habitat utilization in different areas, interacting in different ecological
communities and acting as an energy linkage between them. Although squid occur across
the three regions (Norman and Reid 2000), information on their abundance at a lower scale
(regional level) is scarce. Several other shark species show regional variation in dietary
composition, switching between prey types with changes in prey availability (Medved et
al. 1985; Cortés and Gruber 1990; Stillwell and Kohler 1993; Simpfendorfer et al. 2001b).
Therefore, it is unclear whether the regional differences found in the diet of S. megalops
reflect different patterns in feeding and habitat utilization or rather the natural pattern of
prey availability. In any case, the present findings reinforce the importance of considering
spatial variation as a common phenomenon affecting the feeding ecology of sharks.
Large and small S. megalops exploited different resources during part of the year. In
summer and autumn, large sharks preyed mostly on demersal pelagic prey (mainly
ommastrephid squid), whereas small sharks consumed mainly benthic crustaceans. These
71
ontogenetic differences may be attributed to morphological limitations of small sharks (e.g.
gape-limited), better foraging ability of large fish, or differences in the habitat occupied by
the two size classes. In spring, however, both size classes had a more varied diet,
consuming mainly benthic organisms. Demersal pelagic prey such as squid occur
throughout the year, but they show large, unpredictable fluctuations in abundance
(Anderson and Rodhouse 2001). Therefore, the decline in squid consumption shown by
large S. megalops during spring may be due to a decline in the availability of squid.
Collection of data on the seasonal variation in the abundance of squid in the studied area is
needed for a better understanding of the seasonal pattern exhibited by large S. megalops.
Seasonal and ontogenetic variation in diet is common, and it has been reported for a related
species, the spiny dogfish (S. acanthias) (Jones and Geen 1977a; Hanchet 1991; Koen
Alonso et al. 2002), and for many other shark species (e.g. Cortés and Gruber 1990;
Simpfendorfer et al. 2001b; White et al. 2004). Cortés et al. (1996) found an interaction
between season and size of shark on the diet of the bonnethead shark (Sphyrna tiburo).
However, no other study on the diet of sharks has analysed the interaction of these factors
when samples from different seasons and size classes were compared. In the present study,
an interaction between size and season was found; large and small S. megalops had
different diets in summer and autumn, but consumed similar prey items in spring.
Therefore, the differences found in the dietary composition of large and small S. megalops
suggest that large and small individuals would exhibit, at least during part of the year,
different predator–prey interactions and ecological roles within the marine ecosystem.
Hence, if only the overall diet data are used in an ecosystem model as a proxy for the
predator–prey interactions of S. megalops, some of the interactions exhibited by this
species throughout its lifespan would be ignored.
In conclusion, high variability was found when the overall importance of prey items was
estimated. Furthermore, the dietary composition of S. megalops varied in space and time,
exhibiting differences among regions, seasons, and size classes. Therefore, the intrinsic
natural variability in the dietary composition of S. megalops, and the spatial and temporal
variation in diet exhibited by this opportunistic predator, suggest that studies that infer
predator–prey interactions from overall diet are likely to miss information on the
ecological relationships among species and therefore account for only part of these
interactions. Understanding predator–prey interactions is required for long-term strategic
ecosystem management (Bax 1998). Hence, given that natural variability is intrinsic to
72
ecological systems, the natural variability of predation should be considered when
predatory interactions are used to model ecosystem dynamics.
73
74
74
0.00
0.06
0.04
0.40
Lumbrineris spp
Nereididae
Eunicidae
Aphroditidae
0.02
2.33
Crustacea
0.11
Sipunculus robustus
Echiura
0.58
Unid. Sipuncula
0.67
0.05
Lumbrineridae
0.62
0.15
Sipuncula
%W
%N
%FO
%IRI
1.70
0.00
0.00
0.12
0.22
0.07
0.00
0.00
0.00
0.01
0.04
0.24
95%
3.22
0.07
0.32
1.22
1.10
0.89
0.11
0.16
0.01
0.09
0.30
1.04
95%
0.00
0.00
0.63
0.90
0.12
0.00
0.00
0.00
0.25
0.60
1.72
31.61 28.15
0.13
0.24
1.38
1.79
0.63
0.24
0.38
0.13
0.75
1.25
3.26
95%
34.86
0.39
0.63
2.17
2.63
1.21
0.64
0.86
0.39
1.36
2.06
4.39
95%
0.00
0.00
0.77
1.11
0.15
0.00
0.00
0.00
0.29
0.72
3.05
22.11 19.15
0.15
0.28
1.60
1.98
0.74
0.30
0.43
0.15
0.88
1.49
4.32
95%
24.79
0.57
0.73
2.55
2.91
1.47
0.75
0.91
0.46
1.60
2.47
5.65
95%
11.99
0.00
0.01
0.42
0.08
0.10
0.01
0.03
0.00
0.10
0.28
0.27
9.70
0.00
0.00
0.16
0.03
0.02
0.00
0.00
0.00
0.02
0.09
0.15
95%
14.22
0.02
0.05
0.82
0.15
0.25
0.05
0.08
0.01
0.22
0.56
0.40
95%
Mean Lower Upper Mean Lower Upper Mean Lower Upper Mean Lower Upper
Unid. Polychaeta
Polychaeta
Prey
Taxonomic groups
Appendix 4.a.
Overall dietary compositions. Prey item sorted by (upper panel) taxonomic groups and (lower panel) ecological groups. Mean percentage weight
(%W), mean percentage number (%N), mean percentage frequency of occurrence (%FO), and mean percentage Index of Relative Importance
(%IRI), and 95% confidence intervals. Unid.: unidentifiable; n = 492.
75
75
%W
%N
%FO
%IRI
0.06
0.04
0.36
0.00
0.01
0.02
0.01
0.01
0.01
0.02
0.10
0.01
0.00
0.14
0.02
Decapoda
Caridea
Palaemonidae
Alpheidae
Brachyura
Leucosiidae
Ebalia intermedia
Portunidae
Pilumnus spp
Dendrobranchiata
Solenoceridae
Haliporoides sibogae
Penaeidae
Penaeus spp
0.00
0.02
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.09
0.01
0.00
95%
0.06
0.32
0.00
0.01
0.27
0.07
0.04
0.02
0.02
0.04
0.04
0.01
0.74
0.08
0.14
95%
0.26
1.85
0.13
0.38
0.26
0.25
0.25
0.12
0.12
0.38
0.12
0.26
6.88
0.99
0.99
0.00
0.62
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
2.49
0.36
0.13
95%
0.63
3.84
0.38
0.86
0.65
0.78
0.76
0.39
0.38
0.87
0.39
0.66
12.22
1.77
2.32
95%
0.30
1.31
0.15
0.45
0.31
0.15
0.15
0.15
0.15
0.44
0.15
0.30
1.62
1.19
0.73
0.00
0.57
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.74
0.45
0.15
95%
0.75
2.23
0.45
1.01
0.75
0.46
0.45
0.45
0.46
1.02
0.45
0.75
2.56
2.08
1.46
95%
0.01
0.35
0.00
0.02
0.01
0.01
0.01
0.00
0.00
0.02
0.00
0.01
1.57
0.16
0.10
0.00
0.09
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.44
0.05
0.01
95%
0.04
0.83
0.02
0.08
0.05
0.03
0.03
0.01
0.01
0.08
0.02
0.04
3.18
0.37
0.30
95%
Mean Lower Upper Mean Lower Upper Mean Lower Upper Mean Lower Upper
Unid. Crustacea
Prey
Appendix 4.a. Continued…
76
76
%W
%N
%FO
%IRI
0.03
0.00
0.34
0.10
0.59
0.21
0.02
0.03
0.03
0.01
0.06
0.04
0.04
0.01
Anomura
Diogenidae
Dardanus arrosor
Strigopagurus strigimanus
Paguristes spp
Paguristes sulcatus
Pagurus spp
Distosquilla miles
Austrosquilla osculans
Isopoda
Cirolanidae
Cirolana spp
Cirolana capricornica
0.00
0.01
0.01
0.00
0.00
0.00
0.00
0.00
0.12
0.19
0.02
0.16
0.00
0.00
95%
0.02
0.08
0.10
0.14
0.02
0.08
0.06
0.05
0.34
1.13
0.21
0.56
0.01
0.11
95%
0.25
1.00
0.87
0.62
0.12
0.12
0.37
0.25
5.30
2.48
0.86
5.90
0.12
0.12
0.00
0.38
0.24
0.13
0.00
0.00
0.00
0.00
3.02
1.40
0.27
3.62
0.00
0.00
95%
0.63
1.77
1.65
1.26
0.38
0.39
0.86
0.63
7.85
3.72
1.50
8.25
0.39
0.39
95%
0.29
1.17
0.88
0.73
0.15
0.14
0.43
0.29
3.53
2.64
1.03
3.79
0.14
0.14
0.00
0.45
0.29
0.15
0.00
0.00
0.00
0.00
2.35
1.59
0.42
2.41
0.00
0.00
95%
0.73
2.05
1.64
1.46
0.57
0.45
1.03
0.74
4.95
3.85
1.86
5.18
0.45
0.45
95%
0.01
0.16
0.11
0.07
0.00
0.00
0.02
0.01
2.60
1.08
0.13
3.15
0.00
0.00
0.00
0.04
0.02
0.01
0.00
0.00
0.00
0.00
1.32
0.53
0.03
1.70
0.00
0.00
95%
0.04
0.35
0.26
0.17
0.02
0.02
0.08
0.04
4.42
1.84
0.30
4.90
0.01
0.02
95%
Mean Lower Upper Mean Lower Upper Mean Lower Upper Mean Lower Upper
Palinuridae
Prey
Appendix 4.a. Continued…
77
77
0.26
0.63
0.02
2.74
Volutidae
Fasciolariidae
Turbinidae
Unid. Cephalopoda
0.90
20.03
Ommastrephidae
Todarodes filippovae
1.44
Octopus berrima
9.56
0.78
Octopus warringa
Nototodarus gouldi
5.53
Octopus pallidus
12.55
0.09
Philine angasi
56.43
0.15
Octopus spp
%W
%N
%FO
%IRI
0.06
4.92
13.84
0.44
0.19
1.94
7.73
1.32
0.00
0.22
0.10
0.02
0.00
50.66
95%
2.10
15.38
26.55
2.73
1.55
10.02
18.03
4.74
0.04
1.13
0.49
0.19
0.41
62.47
95%
0.51
2.83
7.54
1.25
0.75
1.52
6.47
2.87
0.36
2.27
1.51
0.74
0.37
0.12
1.81
5.76
0.50
0.24
0.74
4.70
1.84
0.00
1.06
0.75
0.23
0.00
31.45 28.84
95%
1.00
4.01
9.40
2.12
1.37
2.42
8.23
4.03
0.79
3.81
2.43
1.37
0.89
34.44
95%
0.59
3.41
8.76
1.48
0.90
1.80
7.66
3.40
0.43
1.92
1.78
0.89
0.45
0.14
2.13
6.73
0.60
0.29
0.88
5.78
2.14
0.00
0.90
0.87
0.29
0.00
35.89 33.17
95%
1.20
4.75
11.09
2.44
1.66
2.80
9.54
4.77
0.92
2.97
2.85
1.76
1.01
38.35
95%
0.11
5.65
32.05
0.53
0.19
1.69
19.37
2.54
0.02
0.75
0.42
0.10
0.03
50.31
0.01
2.67
23.78
0.18
0.04
0.61
12.67
1.37
0.00
0.29
0.17
0.02
0.00
44.57
95%
0.30
9.39
41.99
1.11
0.43
3.32
26.35
4.19
0.06
1.42
0.80
0.24
0.10
54.64
95%
Mean Lower Upper Mean Lower Upper Mean Lower Upper Mean Lower Upper
Cephalaspidea
Mollusca
Prey
Appendix 4.a. Continued…
78
78
1.27
0.05
0.01
Ommastrephes bartramii
Histioteuthis spp
Bivalvia
0.03
0.31
0.28
38.32
Urolophus spp
Narcine tasmaniensis
Rajidae
Osteichthyes
0.04
0.48
Urolophidae
Anguilliformes
0.57
Mustelus antarcticus
2.58
0.05
Squalus spp
Unid. Osteichthyes
0.03
Unid. Chondrichthyes
1.64
0.70
Chondrichthyes
%W
%N
%FO
%IRI
0.00
1.43
32.71
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.50
0.00
0.00
0.16
0.00
95%
0.14
4.14
43.92
0.75
0.80
0.08
1.47
1.91
0.17
0.10
3.22
0.05
0.15
3.32
2.17
95%
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.52
0.00
0.00
0.36
0.00
0.13
4.61
0.00
3.24
30.34 26.95
0.24
0.49
0.12
0.25
0.12
0.13
0.13
1.54
0.13
0.13
0.88
0.37
95%
0.39
6.05
33.51
0.65
1.17
0.39
0.66
0.39
0.45
0.49
2.46
0.39
0.39
1.53
0.84
95%
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.83
0.00
0.00
0.31
0.00
0.15
5.45
0.00
3.83
34.03 30.74
0.29
0.45
0.14
0.29
0.15
0.14
0.15
1.67
0.14
0.15
1.05
0.44
95%
0.46
7.18
36.98
0.75
0.91
0.45
0.76
0.45
0.45
0.46
2.86
0.45
0.45
1.80
1.03
95%
0.00
5.23
37.27
0.02
0.05
0.00
0.03
0.01
0.00
0.00
0.08
0.00
0.00
0.30
0.06
0.00
3.21
32.73
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.03
0.00
0.00
0.06
0.00
95%
0.02
7.74
42.93
0.07
0.14
0.02
0.11
0.08
0.02
0.02
0.18
0.01
0.02
0.75
0.23
95%
Mean Lower Upper Mean Lower Upper Mean Lower Upper Mean Lower Upper
Todaropsis eblanae
Prey
Appendix 4.a. Continued…
79
79
%W
%N
%FO
%IRI
0.13
0.05
0.17
0.24
2.03
0.54
0.46
1.49
0.24
0.54
0.18
1.25
0.27
0.22
0.17
Scolecenchelys breviceps
Serrivomer spp
Congridae
Clupeidae
Paraulopus nigripinnis
Myctophidae
Macrouridae
Caelorinchus spp
Lepidorhynchus denticulatus
Macruronus novaezelandiae
Cyttidae
Cyttus australis
Macroramphosidae
Macroramphosus scolopax
0.01
0.01
0.00
0.00
0.00
0.13
0.00
0.21
0.00
0.00
0.74
0.00
0.01
0.00
0.00
95%
0.42
0.65
0.89
3.34
0.62
1.09
0.64
3.09
1.15
1.49
3.68
0.77
0.41
0.12
0.33
95%
0.51
0.50
0.12
0.24
0.13
0.85
0.24
0.77
0.36
0.25
1.51
0.12
0.50
0.36
0.37
0.12
0.11
0.00
0.00
0.00
0.24
0.00
0.25
0.00
0.00
0.61
0.00
0.12
0.00
0.00
95%
1.02
1.08
0.38
0.64
0.49
1.69
0.63
1.43
0.85
0.64
2.58
0.38
1.04
0.87
0.86
95%
0.60
0.59
0.14
0.29
0.15
0.87
0.28
0.92
0.44
0.29
1.47
0.15
0.59
0.44
0.44
0.14
0.14
0.00
0.00
0.00
0.15
0.00
0.30
0.00
0.00
0.59
0.00
0.14
0.00
0.00
95%
1.21
1.29
0.45
0.75
0.45
1.65
0.75
1.64
1.02
0.74
2.53
0.45
1.20
1.04
1.03
95%
0.06
0.06
0.01
0.06
0.01
0.16
0.02
0.28
0.05
0.03
0.70
0.01
0.05
0.02
0.03
0.01
0.00
0.00
0.00
0.00
0.03
0.00
0.07
0.00
0.00
0.26
0.00
0.00
0.00
0.00
95%
0.15
0.16
0.04
0.24
0.03
0.38
0.07
0.63
0.14
0.11
1.41
0.04
0.14
0.08
0.09
95%
Mean Lower Upper Mean Lower Upper Mean Lower Upper Mean Lower Upper
Scolecenchelys spp
Prey
Appendix 4.a. Continued…
80
80
%W
%N
%FO
%IRI
0.30
9.77
0.11
0.34
1.37
0.27
0.13
0.37
0.22
0.39
0.18
0.88
2.38
1.88
0.50
Triglidae
Lepidotrigla spp
Lepidotrigla mulhalli
Lepidotrigla modesta
Chelidonichthys kumu
Platycephalidae
Neoplatycephalus conatus
Scorpaenidae
Helicolenus percoides
Perciformes
Acropomatidae
Apogonops anomalus
Gempylidae
Thyrsites atun
0.00
0.53
1.32
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.17
0.00
0.00
6.02
0.00
95%
1.27
3.56
3.71
2.59
0.61
1.24
0.75
0.89
0.39
0.84
3.21
0.87
0.37
13.83
0.98
95%
0.24
1.35
4.48
0.25
0.12
0.24
0.13
0.38
0.25
0.13
0.75
0.36
0.13
5.33
0.12
0.00
0.58
2.81
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.13
0.00
0.00
3.83
0.00
95%
0.62
2.29
6.40
0.63
0.40
0.78
0.39
0.85
0.64
0.39
1.53
0.84
0.39
7.07
0.39
95%
0.29
1.45
4.12
0.29
0.15
0.14
0.15
0.45
0.30
0.15
0.74
0.43
0.15
5.97
0.15
0.00
0.60
2.80
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.15
0.00
0.00
4.25
0.00
95%
0.74
2.33
5.61
0.74
0.47
0.45
0.46
1.01
0.76
0.45
1.45
1.02
0.45
7.71
0.46
95%
0.03
0.63
3.77
0.04
0.01
0.01
0.01
0.05
0.01
0.01
0.21
0.04
0.00
12.00
0.01
0.00
0.22
2.15
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.03
0.00
0.00
7.51
0.00
95%
0.10
1.22
5.88
0.16
0.03
0.06
0.04
0.13
0.05
0.04
0.54
0.12
0.02
17.14
0.04
95%
Mean Lower Upper Mean Lower Upper Mean Lower Upper Mean Lower Upper
Centriscops spp
Prey
Appendix 4.a. Continued…
81
81
%W
%N
%FO
%IRI
0.01
0.00
Unid. Ophiuroidea
0.06
Argentina elongata
Unid. Algae
0.12
Apogonidae
0.00
3.50
Scomber australasicus
Unid. Hydrozoa
0.87
Trachurus declivis
0.00
0.40
Trachurus sp.
Unid. Porifera
0.28
Carangidae
0.79
1.43
Parequula melbournensis
Otariidae
0.29
0.00
0.00
0.00
0.00
0.00
0.00
0.00
1.56
0.00
0.00
0.00
0.32
0.00
95%
0.00
0.04
0.00
0.01
2.67
0.18
0.37
6.06
2.93
1.28
0.87
2.92
0.98
95%
0.12
0.12
0.12
0.12
0.12
0.25
0.13
1.62
0.13
0.13
0.23
1.12
0.12
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.80
0.00
0.00
0.00
0.25
0.00
95%
0.38
0.38
0.39
0.49
0.38
0.65
0.40
2.59
0.39
0.41
0.63
2.17
0.38
95%
0.15
0.15
0.14
0.15
0.14
0.30
0.15
1.93
0.15
0.15
0.29
0.90
0.15
0.00
0.00
0.00
0.00
0.00
0.00
0.00
1.02
0.00
0.00
0.00
0.29
0.00
95%
0.45
0.45
0.45
0.58
0.45
0.74
0.46
3.02
0.45
0.45
0.74
1.66
0.57
95%
0.00
0.00
0.00
0.00
0.02
0.01
0.01
1.32
0.02
0.01
0.02
0.31
0.01
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.54
0.00
0.00
0.00
0.07
0.00
95%
0.01
0.01
0.01
0.02
0.10
0.05
0.02
2.47
0.11
0.05
0.07
0.66
0.04
95%
Mean Lower Upper Mean Lower Upper Mean Lower Upper Mean Lower Upper
Sillago flindersi
Prey
Appendix 4.a. Continued…
82
82
1.24
36.95
11.04
40.25
Benthic infauna
Benthic
Demersal benthic
Demersal pelagic
6.49
4.02
Pelagic
%W
%N
%FO
%IRI
10.00
48.22
14.75
43.97
1.99
5.25
95%
3.64
32.30
7.50
30.18
0.65
2.87
95%
4.38
16.41
11.10
21.10
5.86
41.15
95%
4.22
8.59
6.08
2.85
19.51 13.62
14.02
24.53 17.70
7.71
46.80 35.61
95%
5.19
20.25
12.24
25.34
7.56
29.41
7.21
23.54
14.93
28.91
9.69
32.78
95%
3.54
17.26
9.84
21.90
5.62
26.12
95%
95%
1.78
0.80
8.42
4.49
1.31
2.08
0.79
26.52 32.34 21.17
6.27
33.96 40.28 28.08
1.24
30.70 35.62 25.72
95%
Mean Lower Upper Mean Lower Upper Mean Lower Upper Mean Lower Upper
Benthic epifauna
Prey
Ecological groups
Appendix 4.a. Continued…
83
Photo of pregnant female and mature male (photos by the author).
84
CHAPTER 5
85
CHAPTER 5
DETERMINING REPRODUCTIVE PARAMETERS FOR POPULATION
ASSESSMENTS OF CHONDRICHTHYAN SPECIES WITH ASYNCHRONOUS
OVULATION AND PARTURITION: PIKED SPURDOG (SQUALUS MEGALOPS)
AS A CASE STUDY
5.1 ABSTRACT
Population assessments of chondrichthyan species require several key parameters of their
reproductive biology, which were estimated for Squalus megalops. Length-at-maturity
differed depending on the criterion adopted for defining maturity. In the case of males,
length-at-maturity was smallest when condition of seminal vesicles was adopted as a
maturity criterion. For females, length-at-maturity was smallest when the largest follicle
diameter >3 mm was adopted as the criterion for maturity; this was appropriate only as an
indicator of the onset of maturity. Mature males are capable of mating throughout the year.
Females have a continuous asynchronous reproductive cycle. The sex ratio of embryos is
1:1 and litter size and near-term embryo length increase with maternal length. Females
have an ovarian cycle and gestation period of two years. This was reflected in the
differences found between the maturity and maternity ogives. Although all females are
mature at 600 mm, only 50% of them contribute to annual recruitment each year. Hence,
for chondrichthyan species with reproductive cycles of two, three or more years, if
maturity ogives are used in population assessments instead of maternity ogives, the models
will over-estimate recruitment rates.
5.2 INTRODUCTION
Depleted stocks of many teleost and invertebrate species have high potential for recovery
but this is generally not the case for many species of Chondrichthyes (sharks, rays and
chimaeras). Chondrichthyans have several biological characteristics that make them
susceptible to fishing overexploitation. Chondrichthyans are mostly long-lived predators
that have few offspring, producing close stock–recruitment relationships and slow stock
recovery when overfished (FAO 2000b). In other words, chondrichthyan populations tend
to have lower reproductive rates and lower natural mortality rates, and hence lower
biological productivity, than teleost and invertebrate populations. Consequently, only a
87
small proportion of chondrichthyan populations can be removed annually if catches and
populations are to remain sustainable (Walker 1998). Fisheries targeting chondrichthyan
species have been assessed by population models designed for teleosts, often resulting in
inappropriate techniques being applied to these animals (Musick et al. 2000; Walker 2004).
At present, assessment of chondrichthyan populations is limited by a lack of biological
information (Cortés 1998a), especially for non-targeted-species.
Information on the reproductive biology of chondrichthyans is crucial for quantitative
analysis of their populations. Measures of the biological productivity of chondrichthyan
species, derived from reproductive and natural mortality rates, are required for stock
assessments, demographic assessments and ecological risk assessments (Walker 2004).
Furthermore, these measures are required in species assessments by wildlife conservation
organizations such as the International Union for Conservation of Nature and Natural
Resources (IUCN; Hilton-Taylor 2000). All these assessments use the same information
for representing key parameters of reproduction: sex ratio at birth, the relationship between
the number of offspring and maternal age or size of animals (litter size) and, sometimes
useful for fisheries assessments, the relationship between the proportion of animals in
mature condition at any time and the age or size of animals (maturity ogive). Essential for
all types of assessment is the relationship between the proportion of the female population
contributing to annual recruitment (i.e. females in maternal condition) and the age or size
of animals (maternity ogive; Walker 2005).
Maternal condition is not usually considered or it is incorrectly equated to mature or
pregnant conditions in most demographic studies. Maternity ogives of chondrichthyan
species can be markedly different from maturity ogives and pregnancy ogives. For
example, off southern Australia, the parturition frequency of school shark (Galeorhinus
galeus) is triennial so at most one-third of the mature female population contributes to
recruitment at the beginning of the following year. Also, length-at-maternity is much larger
than length-at-maturity for this species (Walker 2005). Hence, although all females attain
maturity at ~1600 mm total length, about one-third of them are in maternal condition in
any year. Thus, for chondrichthyan species with complex breeding cycles exceeding one
year of duration, if maternity ogives are ignored and only maturity curves are considered in
the analysis, population models would over-estimate recruitment rates and bias
assessments.
88
Methods for appropriate determination of maternity parameters needed for population
assessments of chondrichthyans have only been established for viviparous species with
synchronous reproductive cycles (Walker 2005). Although most viviparous
chondrichthyans have synchronous mating, gestation and parturition (Hamlett and Koob
1999), in some species with complex breeding cycles, mating, gestation and parturition are
asynchronous (Yano and Tanaka 1988; Yano 1993, 1995). For these species, a different
approach is required for determining the parameters of maternity ogives for population
assessments.
Complex asynchronous breeding cycles of several years duration have been reported for
several squalid sharks (Squalidae) (Yano and Tanaka 1988; Yano 1995; Watson and Smale
1998). The piked spurdog (Squalus megalops) provides for a case study of the reproductive
information needed for quantitative population studies as this species may have a long and
complex ovarian cycle and gestation period (Watson and Smale 1998; Graham 2005).
Hence, this shark may have different patterns of maturity condition and maternity
condition. Furthermore, S. megalops has a high natural abundance in southern Australia
(Bulman et al. 2001; Graham et al. 2001) and, although this shark is among the most
caught by-catch species (Walker et al. 2005), its abundance has remained stable off New
South Wales since it was first surveyed (Graham et al. 2001).
The present paper is part of a broad-scale program for the assessment of ecological risk
from the effects of fishing on the population status of S. megalops and other
chondrichthyan species impacted by fishing in southern Australian fisheries. The specific
objectives of the present study were to: (i) describe the reproductive biology of this species
with emphasis on the information needed for quantitative population studies; (ii) analyse
temporal variation in reproductive condition of mature males; (iii) determine sex ratio of
embryos; (iv) evaluate litter and offspring size–maternal length relationships; (v) determine
the periodicity of the ovarian cycle and gestation period of females; (vi) determine the
maturity ogive as a function of length for each sex; and (vii) determine pregnancy and
maternity ogives as a function of female length.
89
5.3 MATERIALS AND METHODS
Sampling
Specimens of S. megalops were obtained from the by-catch of the Australian Southern and
Eastern Scalefish and Shark Fishery in waters off Robe, Lakes Entrance and Ulladulla,
Australia (Fig. 5.1). Samples from Lakes Entrance and Ulladulla were caught by
commercial bottom trawl fishing vessels, whereas samples from Robe were caught by
commercial shark fishing vessels using gillnets of 6½-inch (165 mm) mesh-size. Samples
were collected monthly between October 2002 and April 2004, with the exception of the
August–September period, when S. megalops seemed to move off the fishing grounds and
weather conditions hampered sampling operations. The specimens were sexed, measured
(total length, TL, ±1 mm), weighed (body mass, TM) on an electronic balance (±1 g), and
dissected to investigate their reproductive biology. Mass of liver (LM), gonads (GM) and
seminal fluid in seminal vesicles (SFM) were also weighed (±0.1 g).
Analyses of males
For males, different criteria were used to investigate maturity condition. Length of the left
clasper (CL) was recorded from the join in skin near the anus to the distal end (±1 mm).
Macroscopic inspection of condition of clasper calcification (CI), testes (GI), seminal
vesicles (VI), seminal fluid (VC), and vas deferens (VD) was undertaken to investigate
further maturity by adopting four indices of maturity condition (using Walker’s scale,
modified for this species; Walker 2005) (Table 5.1).
Temporal variation in reproductive condition of mature males was investigated by
recording the seminal vesicle fullness (VF) using a quarterly scale (0, empty to 4, full) and
seminal fluid coloration and consistency (VC: 1, clear to 3, cloudy and thick). For each
season, the gonadosomatic index (GSI = 100 GM TM-1), the hepatosomatic index (HSI =
100 LM TM-1) and the spermatosomatic index (SSI = 100 SFM TM-1) for males with VI =
2 were also calculated. Data were not analysed by month due to small sample sizes for
some months. Winter samples of mature males were not considered due to small sample
size (n = 2). Temporal variation in HSI, GSI and SSI was tested by ANOVA.
90
Reproductive cycle of females
For females, macroscopic inspection of condition of ovaries, oviducal glands and uteri was
undertaken to investigate sex ratio of embryos, litter size, growth of embryos, periodicity
of the ovarian cycle and gestation period, and mature, pregnant and maternal conditions.
Indices were adopted for recording the condition of ovaries (GI), oviducal glands (OI) and
uteri (UI) (using Walker’s scale, modified for this species; Walker 2005) (Table 5.1).
Maximum width of the left uterus was measured (±1 mm) to investigate the dynamics of
the reproductive cycle. For pregnant females, number of in utero eggs or embryos and the
sex, uterus (left or right), stage of development (in utero egg only, embryo with external
yolk or embryo only), and total length (TLE, ±1 mm) of each embryo were recorded.
Sex ratio of embryos
Chi-square tests with Yates’ continuity correction were applied to pregnant females (UI =
4 and 5) to test two hypotheses. (1) The number of in utero eggs and embryos in the left
uterus equalled the number in the right uterus, and (2) the sex ratio of in utero embryos
was 1:1.
Litter and offspring size–maternal length
The relationship between the number of in utero eggs or embryos (litter size) and maternal
TL and the relationship between total length of near-term embryos (TLE) (offspring size)
and maternal TL were represented by a linear regression model (Walker 2005). Embryos
were considered near-term when the external yolk sac was completely absorbed. Females
were collected from different regions (Fig. 5.1) and may exhibit geographic variation in
their reproductive parameters (Wourms 1977). ANCOVA was used to test for the effects of
region on the linear relationships between maternal TL and litter and offspring size.
91
129º E
141º E
150 º E
200 m
Ulladulla
• Robe
37º S
Lakes
Entrance
•
•
Australia
41º S
Tasmania
N
200
0
200 Miles
Figure 5.1. Map of sampling area showing the three regions compared (shaded) and ports.
92
Ovarian cycle
To determine the ovarian cycle, the diameter of the 20 largest follicles in each ovary was
measured (±1 mm) to obtain the largest follicle diameter (LFD) for females caught
throughout the year. Because the diameters of the largest follicles varied between
individual animals and uterus condition, temporal variation of follicle growth was
examined separately for each uterus condition defined in Table 5.1 (Walker 2005). Due to
the asynchronous nature of the reproductive cycle (see “Results”), only pregnant females
(UI = 4 and 5) were used to estimate the ovarian cycle. Pregnant females were arbitrarily
classed in 5 categories based on the size of the embryo they carried: 0 for females carrying
in utero eggs corresponding to 0 mm TLE, 1 for females carrying embryos <30 mm TLE, 2
for females carrying embryos 30–99 mm TLE, 3 for females carrying embryos 100–199
mm TLE, and 4 for females carrying embryos 200 mm TLE. Based on the assumption
that LFD from different year classes have the same growth pattern, pregnant females in
class 0 from late 2002, classes 1 and 2 from early 2003, class 3 from late 2003, and class 4
from early 2004 were used in a linear model to determine annual growth rate of follicles
(Walker 2005). Data from different regions were pooled and no comparisons among
regions were performed as no samples were collected from Robe and Ulladulla during
May–November.
Gestation period
Gestation period was determined by plotting the percentage of females in uterus condition
UI = 4–6 against month and TLE against Julian day. Based on the assumption that embryos
from different year classes have the same growth pattern, TLE from the same subset of
data selected for the ovarian cycle analysis was used in a growth model. The Gompertz
curve is one of the best models of embryonic fish growth (Ricker 1979), hence, this model
was fitted to the data subset, holding L at 230 mm, the average size at birth. Data from
different regions were pooled and no comparisons among regions were conducted as no
samples were collected from Robe and Ulladulla during May–November.
Maturity, pregnancy and maternity ogives
Logistic models were used to determine the proportion of mature males and females at any
TL and, for females, the proportion in pregnant and maternal conditions (Punt and Walker
1998; Roa et al. 1999). Model parameters and the ogives with 95% confidence intervals
(CI) were estimated by the method of maximum likelihood of the probit procedure using
93
Table 5.1. Indices used for staging reproductive condition. Maturity conditions
corresponding to each index are also listed (modified from Walker 2005). Maturity
condition: immature (I), mature (M), and uncertain (U).
Organ
Index
Description
Maturity
Male
Clasper
Testis
Seminal vesicles
CI = 0
Pliable with no calcification and without hooks
I
CI = 1
Partly calcified with small hooks
I
CI = 2
Rigid and fully calcified with evident hooks
M
GI = 1
Undifferentiated thin tissue strip
I
GI = 2
Thickened tissue strip becoming lobular
I
GI = 3
Enlarged with evident testicular lobules
M
VI = 1
Thin translucent walls and seminal fluids absent
I
VI = 2
Thickened opaque walls and seminal fluids
M
present
VI = 3
Thickened opaque walls and seminal fluids
M
absent
Vas deferens
VD = 1 Thin line along dorsal surface of abdominal
I
cavity
VD = 2 Thickened line that begins to coil
I
VD = 3 Enlarged and fully coiled
M
GI = 1
Largest follicles hyaline and of diameter <3 mm
I
GI = 2
Largest follicles white and of diameter 3–14 mm
I
GI = 3
Largest follicles yellow with yolk and of
M
Female
Ovary
diameter 15 mm
Oviducal gland
Uterus
OI = 1
Indistinct from anterior uterus
I
OI = 2
Distinct but only partly formed
I
OI = 3
Enlarged and kidney-shaped
M
UI = 1
Uniformly thin tubular structure
I
UI = 2
Thin tubular structure partly enlarged posteriorly
I
UI = 3
Enlarged tubular structure partly narrow
U
anteriorly
94
Table 5.1. Continued…
Organ
Index
Uterus
UI = 4
Description
In utero eggs present without macroscopically
Maturity
M
visible embryos present
UI = 5
In utero embryos macroscopically visible
M
UI = 6
Enlarged tubular structure distended (post-
M
partum)
Clasper index (CI); gonad index (GI); seminal vesicle index (VI); vas deferens index (VD);
oviducal gland index (OI); uterus index (UI).
95
the SAS statistical package (SAS Institute, Cary, North Carolina, USA) (Walker 2005).
Given the uncertainty in determining maturity in males (Conrath 2004), indices for each of
four separate methods were used and the results compared. A male was classed as in
mature condition if CI = 2, GI = 3, VI = 2–3, or VD = 3; otherwise it was classed as
immature. Logistic curves and associated parameters were determined for each maturity
criterion. Females had a complex reproductive cycle (see “Results”), hence, as for males,
different maturity criteria were adopted and the results compared. A female was classed as
in mature condition if GI = 3, OI = 3, or UI = 4–6; otherwise it was classed as immature. In
addition, the effect of adopting three alternative maturity criteria based on LFD was
assessed: LFD >3 mm (Walker 2005) (for onset of maturity), LFD 15 mm (present study)
and LFD 20 mm (Yano and Tanaka 1988; Watson and Smale 1998). Logistic curves and
associated parameters were determined for each maturity criterion. Regional comparisons
of samples from Lakes Entrance and Robe were undertaken using the logistic procedure of
the SAS statistical package (SAS Institute, Cary, North Carolina, USA) (Walker 2005).
Samples from Ulladulla were not included as all sharks were in mature condition and
hence the maturity ogive was not calculated.
For the analysis of pregnant females, a female was classed as in pregnant condition if UI =
4–5; otherwise it was classed as non-pregnant. For the maternity analysis, a female was
classed as in maternal condition if, had it survived, it would have given birth by the end of
2003 or early 2004; i.e. it contributed to annual recruitment at the beginning of 2004.
Hence, for each pregnant female, parturition time was calculated using the embryo growth
model. Pregnancy and maternity ogives were determined using logistic models. Model
parameters and the ogives with 95% CI were estimated by the method of maximum
likelihood of the probit procedure using the SAS statistical package (SAS Institute, Cary,
North Carolina, USA) (Walker 2005). Given that parturition frequency is biennial (see
“Results”), for SAS probit analysis of maternity condition, the parameter Pmax (maximum
proportion of animals in maternal condition) was altered from 1.0 to 0.5. The SAS output
was then multiplied by 0.5 to obtain the parameters of the maternity ogive, with 95% CI
(Walker 2005). Assuming that sampling was not biased for pregnant or non-pregnant
females, pregnancy and maternity ogives were determined using pooled data from different
regions; hence, no regional comparisons were made for these ogives.
96
5.4 RESULTS
Analyses of males
A total of 207 male S. megalops (274–470 mm TL) were collected for reproductive
analyses. Male sharks were mostly sampled from Lakes Entrance (Fig. 5.1) so no regional
comparisons were performed. Clasper length (CL) showed a sigmoid relationship with TL.
Claspers grew gradually in animals <350 mm TL, followed by rapid growth until 385 mm
TL and 27 mm CL (CI = 2), which several indicators suggest is the TL for the onset of
maturity.
Maturity ogives differed depending on the maturity criterion adopted (Fig. 5.2e). The ogive
based on maturity condition of seminal vesicles (VI = 2 or 3) showed a value of TL at
which 50% of the population was mature (L50) (with 95% CI) of 373 (368, 377) mm (Fig.
5.2a). This value was considerably smaller than those obtained using other maturity
criteria. The ogives based on vas deferens (VD = 3), clasper (CI = 2) and gonad (GI = 3)
conditions were in reasonable agreement and provided values of L50 of 392 (388, 396), 393
(390, 395) and 398 (395, 401) mm, respectively (Figs. 5.2b–d).
Irrespective of which maturity index was used, males classed as mature were found
throughout the year (Fig. 5.3). Statistical testing of the frequency of males in different GI,
CI, VI and VD conditions was not carried out due to the opportunistic nature of the
sampling design and possible size and sex aggregation of the sharks; however, mature
males (GI = 3, CI = 2, VI = 2–3 or VD = 3) were collected from each season (Figs. 5.3a–
d). The frequency of vesicle fullness (VF) of mature males with VI = 2 was similar
throughout the year; most mature males had full seminal vesicles (VF = 4) (Fig. 5.3e)
containing cloudy and thick seminal fluids (VC = 3) (Fig. 5.3f). There were no seasonal
differences in the maturity condition of mature males. Mature males had similar values of
mean HSI throughout the sampling period (ANOVA: F2, 117= 2.19, P = 0.1164). Also, no
seasonal variation was found in the mean GSI (ANOVA: F2, 114= 0.94, P = 0.3939) and
mean SSI (ANOVA: F2, 112= 1.14, P = 0.3241). It appears that mature males are capable of
mating throughout the year.
97
1.0
1.0
Proportion mature
(a)
(b)
0.8
0.8
0.6
0.6
0.4
0.4
0.2
0.2
0.0
0.0
1.0
1.0
(d)
(c)
0.8
0.8
0.6
0.6
0.4
0.4
0.2
0.2
0.0
0.0
300
1.0
350
400
450
500
Total length (mm)
(e)
0.8
0.6
0.4
VI
VD
CI
GI
0.2
0.0
300
350
400
450
500
Total length (mm)
Figure 5.2. Male length-at-maturity ogives. Proportion of male population in mature
condition versus total length with 95% confidence intervals (- - -) determined from (a)
seminal vesicle condition (VI), (b) vas deferens condition (VD), (c) clasper condition (CI),
(d) testis condition (GI), and (e) comparison of mean ogives for VI, VD, CI and GI. Values
of parameters and statistical values for the equation P = Pmax (1 + e -ln (19) ((l - l50 ) / (l95 - l50 )) ) −1 used
in the probit analysis are as follows:
Maturity criterion
Seminal vesicle condition
Vas deferens condition
Clasper condition
Testis condition
L50 (CI)
373 (368, 377)
392 (388, 396)
393 (390, 395)
398 (395, 401)
L95 (CI)
403 (399, 409)
428 (422, 437)
417 (413, 422)
436 (430, 444)
Pmax
1
1
1
1
n
201
172
207
207
ML
– 113.62
– 141.71
– 169.12
– 246.96
P
***
***
***
***
where l is total length (TL), P is the proportion of animals at TL l, L50 and L95 are
parameters, Pmax is an asymptotic constant, n is the total number of animals, ML is
maximum likelihood, and P is the probability of statistical significance (***P <0.001).
98
n=122
n=33
n=4
n=122
n=48
100
(a)
n=4
n=48
(b)
80
80
60
60
GI
1
2
3
40
20
CI
0
1
2
40
20
0
0
Summer Autumn Winter Spring
n=121
n=31
n=4
Summer Autumn Winter Spring
n=45
n=92
100
n=31
n=4
n=45
100
(c)
Percentage
n=33
100
(d)
80
80
60
60
VI
40
1
2
3
20
VD
1
2
3
40
20
0
0
Summer Autumn Winter Spring
n=97
n=30
Summer Autumn Winter Spring
n=48
n=97
100
n=30
n=48
100
(e)
(f)
80
80
60
VF
1
2
3
4
40
20
60
VC
1
2
3
40
20
0
0
Summer
Autumn
Summer
Spring
Autumn
Spring
Season
Figure 5.3. Percentage of male sharks at different maturity condition collected from
different seasons. (a) Gonad index (GI), (b) clasper index (CI), (c) vesicle index (VI), (d)
vas deferens index (VD), (e) vesicle fullness (VF) for males with VI = 2, and (f) seminal
fluid coloration (VC) for males with VI = 2 and VF = 4. Sample sizes are shown above
bars.
99
Reproductive cycle of females
Analyses of 722 female S. megalops (270–635 mm TL) suggested that females had a
continuous asynchronous reproductive cycle. There was an increase in LFD with uterus
width for females ovulating for the first time (UI = 1–3) (Fig. 5.4a). Ovulation occurred
once LFD reached ~40 mm and uterus reached a width of ~18 mm. For pregnant females
carrying in utero eggs (UI = 4) or embryos at an early-stage of development (UI = 51),
LFD was small. The follicles enlarged throughout gestation synchronously with embryonic
growth and were ready to be ovulated when embryos were near-term, indicating that
fertilization and the gestation of a subsequent litter can occur immediately after parturition.
This was supported by the significant correlation between LFD and TLE (r = 0.954, n =
544, P <0.0001) (Fig. 5.4b).
Sex ratio of embryos
Macroscopically visible in utero eggs and embryos were examined in 308 pregnant
females (UI = 4–5). Most of these females (72.7%) carried only one egg or embryo per
uterus and no female was observed to carry eggs and embryos at the same time. Significant
differences were found in the number of in utero eggs and embryos between the left uterus
and the right uterus ( 2 = 15.882, d.f. = 1, P <0.001). Of a total of 668 eggs and embryos
counted, 386 (57.8%) were present in the right uterus. However, when the analysis was
performed on pregnant females carrying up to two in utero eggs or embryos, no significant
differences were found between the left uterus and right uterus ( 2 = 3.561, d.f. = 1, P =
0.059). A total of 450 embryos was sexed of which 154 (34.2%) were male, 145 (32.2%)
were females, and 151 (33.6%) were classed as “unknown sex” due to their early stage of
development. The sex ratio of embryos was not significantly different from 1:1 ( 2 = 0.214,
d.f. = 1, P = 0.644).
Litter and offspring size–maternal length
Litter size was recorded for 308 pregnant females. All females carrying only one embryo
were excluded from the analysis because it was assumed that they had aborted embryo(s)
due to stress of capture. This assumption was supported by the occasional presence of
embryos on the deck of the vessels (J. M. Braccini, pers. obs.) and the fact that these
females had empty uteri with stretched and vascularized walls, suggesting the loss of one
or more embryos. Uteri containing one or two embryos had turgid walls, indicating that
abortion had not occurred. Regional differences detected in the relationship between
100
60
n = 488
Largest follicle diameter (mm)
(a)
Parturition
50
Ovulation and
fertilization
UI
40
1
2
30
Embryonic
and follicle
growth
3
4
20
6
10
51
52
Follicle
growth
53
54
Embryonic growth
0
0
10
20
30
40
50
60
Uterus width (mm)
50
Largest follicle diameter (mm)
(b)
40
30
20
10
n = 544
0
0
50
100
150
200
250
Total length of embryo (mm)
Figure 5.4. Relationships between largest follicle diameter (LFD) and (a) uterus width for
females in different uterus condition (UI) and (b) total length of embryo (TLE) with 95%
confidence intervals (- - -) and predicted intervals (….....) (see Table 5.1 for UI definition).
LFD = 0.153 TLE + 9.303; r2 = 0.931.
101
5
(a)
n = 273
(b)
n = 62
Litter size
4
3
2
1
Total length of near-term embryo (mm)
0
300
250
200
150
100
400
500
600
700
Total length (mm)
Figure 5.5. Relationship between maternal total length (TL) and (a) litter size and (b) total
length of near-term embryos (TLE) with 95% confidence intervals (- - -) and predicted
intervals (….....). Litter size = 0.00711 TL – 1.503; r2 = 0.330; TLE = 0.203 TL + 100.6; r2 =
0.587.
102
litter size and maternal TL (ANCOVA: F2, 274= 3.87, P = 0.022) were considered an
artefact of the sampling method. Samples from Lakes Entrance and Ulladulla were
collected by bottom trawl nets whereas those from Robe were collected by gillnets of 6½inch (165 mm) mesh-size. For S. megalops, 6½-inch gillnets selected for large-sized
females (J. M. Braccini, pers. obs.) and this is likely to have created apparent regional
differences in the litter size–maternal TL relationship caused by sampling bias or lengthselective fishing mortality. The results were therefore presented pooling the three regions.
Litter size for most females was two (69.3%) or three (30.0%) and only a few of them
carried four (0.7%) in utero eggs or embryos. Litter size showed a linear relationship with
TL (F1, 273 = 132.38, P <0.001) (Fig. 5.5a).
Embryo length (TLE) was recorded for 62 near-term embryos (191–244 mm TLE) and the
mean relative length-at-birth (with 95% CI) was 38.5 (35.6, 42.4) % of maternal TL. No
regional differences in the relationship between TLE and maternal TL (ANCOVA: F2, 60=
0.67, P = 0.515) were detected, so samples collected from different regions were pooled
for subsequent analyses. Near-term embryo length increased linearly with maternal TL (F1,
60 =
85.40, P <0.001) (Fig. 5.5b).
Ovarian cycle
Largest follicle diameter (LFD) was recorded for 658 females and ranged from 1–49 mm.
Females with uterus condition (UI) = 1 always had small follicles (Fig. 5.6a). Females with
UI = 2 showed a wide range of LFD (1–39 mm) at any time (Fig. 5.6b), indicating that
follicles can approach full size before the uteri were fully developed. Females with UI = 3
or 6 were observed carrying large follicles at all times of the year (Figs. 5.6c, d),
suggesting that follicle enlargement and ovulation are not temporally synchronous between
animals. For animals with UI = 4, LFD was relatively small indicating that ovulation was
complete; no animals were observed in the process of ovulation (Fig. 5.6e). Wide variation
of LFD (6–49 mm) was observed for pregnant females carrying embryos (UI = 5) (Fig.
5.6f). Females carrying small embryos had small follicles whereas females carrying nearterm embryos had large follicles, suggesting that ovulation immediately follows
parturition. Furthermore, only a small percentage of mature females (8.3%) were in UI = 6
condition and they all carried large follicles, indicating a short period between pregnancies.
There was a linear relationship between LFD and Julian Day (F1, 104 = 709.64, P <0.001)
103
50
50
Largest follicle diameter (mm)
(a)
n = 107
40
40
30
30
20
20
10
10
0
0
50
(b)
n = 235
(d)
n = 25
50
(c)
n = 16
40
40
30
30
20
20
10
10
0
0
50
50
(e)
(f)
n = 89
40
40
30
30
20
20
10
10
0
0
n = 186
0
100 200 300 400 500 600
0
UI
51
52
53
54
100 200 300 400 500 600
Julian day
2003
1/10/2002
2004
Figure 5.6. Ovarian cycle. Relationship between largest follicle diameter and Julian day
for (a) uterus index (UI) = 1, (b) UI = 2, (c) UI = 3, (d) UI = 6, (e) UI = 4, and (f) UI = 5.
104
Largest follicle diameter (mm)
50
40
30
UI
20
4
52
53
54
10
0
0
100
200
300
400
500
600
Julian day
2003
1/10/2002
2004
Figure 5.7. Hypothetical follicle development curve for females in uterus condition (UI) =
4 and 5. Linear model fitted to selected subset of data with 95% confidence intervals (- - -)
and predicted intervals (….....). Largest follicle diameter = 0.07 Julian day + 8.797; r2 =
0.874.
105
(Fig. 5.7), indicating linear follicular growth. Annual growth of LFD was 24 mm y–1,
suggesting an ovarian cycle of ~19 months.
Gestation period
A total of 423 embryos was measured and each of 152 in utero eggs was assigned a TLE
value of 0 mm for determining gestation period. S. megalops was an asynchronous breeder
in which ovulation, parturition and mating did not occur at any particular time of the year.
Females carrying in utero eggs (UI = 4) or near-term embryos (UI = 54) were observed
throughout the year, providing further evidence of no pattern of temporal periodicity in the
reproductive cycle (Figs. 5.8a, b). Furthermore, embryos at different stages of development
could be found at all times of the year. Based on the Gompertz growth model, annual
growth of embryos was 170 mm y–1, suggesting a gestation period of ~2 years (Fig. 5.8c).
Maturity, pregnancy and maternity ogives
Maturity ogives of females differed depending on the maturity criterion adopted (Fig.
5.9g). When the maturity criterion used was LFD >3 mm (onset of maturity), length at
which 50% (L50) of the animals were in mature condition (with 95% CI) was 459 (457,
461) mm (Fig. 5.9a). When LFD 15 mm and uteri condition (UI = 4, 5 or 6) were used,
L50 was 484 (481, 487) mm and 486 (485, 488) mm, respectively (Figs. 5.9b, e). The
ogives for LFD 20 mm and oviducal gland condition (OI = 3) were similar (Figs. 5.9c, f)
and showed a larger value of L50 of 495 (492, 498) and 495 (491, 499) mm, respectively.
Finally, L50 based on ovarian condition (GI = 3) was 477 (475, 479) (Fig. 5.9d).
The criterion used to test for the effects of region on the maturity ogives of females was
LFD 15 mm. This criterion was preferred to other criteria because follicles of 15 mm
diameter were yellow, indicating that vitellogenesis was well advanced and because the
ogive and the value of L50 obtained were in reasonable agreement with most of the other
criteria considered. Significant differences were found in the maturity ogives of females
from Lakes Entrance and Robe (P <0.0001) (Fig. 5.9h). The value of L50 for females from
Lakes Entrance was 478 (475, 482) mm, whereas the value of L50 for females from Robe
was 514 (506, 523) mm. However, as in the case of the litter size–maternal TL
relationship, there could be apparent differences due to the effects of length-selectivity of
the 6½-inch gillnet used in Robe. Such length-selectivity might distort the maturity ogive
by the effects of sampling bias and length-selective fishing mortality.
106
31 62100 52 7 4
Percentage
100
2
8 47 27
(a)
80
60
UI
40
4
20
5
N.D.
6
0
1 2 3 4 5 6 7 8 9 10 1112
Month
250
(b)
n = 575
200
150
Total length of embryo (mm)
100
50
TLE
0 mm
<30 mm
30-99 mm
100-199 mm
>199 mm
0
250
(c)
n = 97
200
150
100
50
0
0
100 200 300 400 500 600
Julian day
2003
1/10/2002
2004
Figure 5.8. Gestation period. (a) Distribution of different stages of maturity of females in
uterus condition (UI) = 4–6 during the year. (b) Length of embryos (TLE) collected during
the sampling period. (c) Hypothetical growth curve with Gompertz model fitted to the
(-0.009 t)
) 2
selected subset of data. TLE = 230 e (-8.068 e
; r = 0.9. N.D.: no data. Sample sizes
are shown above bars.
107
1.0
0.8
1.0
(a)
0.8
0.6
0.6
0.4
0.4
0.2
0.2
0.0
0.0
1.0
Proportion mature
0.8
1.0
(c)
0.8
0.6
0.6
0.4
0.4
0.2
0.2
0.0
0.0
1.0
0.8
(d)
1.0
(e)
0.8
0.6
0.6
0.4
0.4
0.2
0.2
0.0
0.0
1.0
0.8
(b)
(f)
1.0
(g)
0.8
LFD3
LFD15
LFD20
GI
UI
OI
0.6
0.4
(h)
Lakes
Entrance
0.6
0.4
0.2
Robe
0.2
0.0
0.0
300
400
500
600
300
400
500
600
Total length (mm)
Figure 5.9. Female length-at-maturity ogives. Proportion of female population in mature
condition against total length with 95% confidence intervals (- - -) determined from (a)
largest ovarian follicle diameter (LFD) >3 mm (LFD3), (b) LFD 15 mm (LFD15), (c)
LFD 20 mm (LFD20), (d) ovary condition (GI), (e) uteri condition (UI), and (f) oviducal
gland condition (OI). (g) Comparison of mean ogives for LFD3, LFD15, LFD20, GI, UI,
and OI, and (h) comparison between maturity ogives for females from Lakes Entrance and
Robe with 95% confidence intervals (- - -) based on the maturity criterion LFD15. Values
of parameters and statistical values for the equation
P = Pmax (1 + e -ln (19) ((l - l50 ) / (l95 - l50 ))) −1 used in the probit analysis are as follows:
Maturity criterion
LFD > 3 mm
LFD 15 mm
LFD 20 mm
Ovaries condition
Uteri condition
Oviducal gland condition
L50 (CI)
459 (457, 461)
484 (481, 487)
495 (492, 498)
477 (475, 479)
486 (485, 488)
495 (491, 499)
L95 (CI)
491 (488, 495)
554 (547, 563)
577 (570, 586)
527 (522, 531)
534 (530, 538)
573 (564, 584)
Pmax
1
1
1
1
1
1
n
706
616
615
621
719
584
ML
– 390.86
– 647.38
– 1131.75
– 879.03
– 1231.29
– 563.97
P
***
***
***
***
***
***
where l is total length (TL), P is the proportion of animals at TL l, L50 and L95 are
parameters, Pmax is an asymptotic constant, n is the total number of animals, ML is
maximum likelihood, and P is the probability of statistical significance (***P <0.001).
108
Length at which 50% of the female population was in pregnant condition was 495 (492,
497) mm (Fig. 5.10a); however, at any length, at most 50% of the female population was in
maternal condition (Fig. 5.10b). The TL-at-maternity and TL-at-pregnancy were larger
than TL-at-maturity (Fig. 5.10c). Although all females were mature at 600 mm, only half
of the population was in maternal condition and hence contributing to annual recruitment.
109
Proportion pregnant
1.0
(a)
0.8
0.6
0.4
0.2
0.0
Proportion maternal
0.5
(b)
0.4
0.3
0.2
0.1
0.0
1.0
(c)
Proportion
0.8
0.6
0.4
Maturity ogive
Pregnancy ogive
Maternity ogive
0.2
0.0
300
400
500
600
700
Total length (mm)
Figure 5.10. Female length-at-pregnancy and maternity ogives. Proportion of female
population in (a) pregnancy and (b) maternal conditions against total length with 95%
confidence intervals (- - -). (c) Comparisons between maturity, pregnancy and maternity
ogives. Values of parameters and statistical values for the equation
P = Pmax (1 + e -ln (19) ((l - l50 ) / (l95 - l50 )) ) −1 used in the probit analysis are as follows:
Condition
Pregnant
Maternal
L50 (CI)
495 (492, 497)
531 (528, 534)
L95 (CI)
554 (548, 560)
626 (618, 635)
Pmax
1
0.5
n
720
522
ML
– 777.19
– 1983.39
P
***
***
where l is total length (TL), P is the proportion of animals at TL l, L50 and L95 are
parameters, Pmax is an asymptotic constant, n is the total number of animals, ML is
maximum likelihood, and P is the probability of statistical significance (***P <0.001).
110
5.5 DISCUSSION
Given uncertainty as to the best descriptor of maturity of male sharks (Conrath 2004), the
results for four indices were compared in the present study. When condition of seminal
vesicles was used, L50 was considerably smaller than when conditions of gonads, vas
deferens or clasper calcification were used. Walker (2005) also found a smaller value of
L50 when comparing the condition of seminal vesicles with gonad condition or clasper
calcification in G. galeus. These findings suggest that seminal vesicle condition might
class some males as mature even though they may not be capable of mating as, for
example, they may not have fully functional claspers. Watson and Smale (1998) and
Graham (2005) found similar values of L50 to those obtained in the present study based on
conditions of the gonads, vas deferens and clasper calcification. For male S. megalops,
these criteria for maturity condition gave similar values of L50 and similar maturity ogives,
suggesting that any of these criteria could be used to determine maturity.
Irrespective of maturity criterion, mature males were observed in all seasons and none of
GSI, HSI and SSI exhibited seasonal variation, indicating that males are in mating
condition throughout the year. A similar pattern is reported for male S. megalops from
South Africa (Watson and Smale 1998) and several other shark species (Parsons and Grier
1992). This would be advantageous for species that inhabit environments with little
variation in environmental cues (e.g., deep sea and tropics) or where mate location may be
difficult (e.g., deep sea and open ocean) or both (Wourms 1977; Parsons and Grier 1992).
Squalus megalops inhabits waters of the continental shelf and upper continental slope to
510 m (Last and Stevens 1994) so it cannot be considered a deepwater shark. However,
most squalid species occur in deeper waters on the continental slope (Last and Stevens
1994); hence, the apparent lack of seasonality in the reproductive cycle of male S.
megalops may be an ancestral trait.
Females have a continuous reproductive cycle. Following ovulation, follicles begin to
undergo vitellogenesis again concurrently with embryonic growth and are ready for
ovulation and fertilization immediately after parturition. A similar pattern is reported for S.
megalops from New South Wales, Australia (Graham 2005) and South Africa (Watson and
Smale 1998) and for other species of Squalus (Kibesaki 1954; Jones and Geen 1977b;
Chen et al. 1981) although some female spiny dogfish (S. acanthias) have a resting period
between pregnancies (Jones and Geen 1977b; Hanchet 1988). The few observed mature
111
female S. megalops in the present study that were not pregnant all carried enlarged follicles
ready for ovulation, suggesting a very short period between pregnancies.
The sex ratio of embryos is 1:1. A 1:1 embryo sex ratio is also reported for S. megalops
from South Africa (Watson and Smale 1998) and for other squalid species (Hanchet 1988;
Yano 1995). A 1:1 embryo sex ratio is expected for a sexually balanced population,
assuming that males and females have similar mortalities. Less straightforward is,
however, the distribution of embryos between uteri. When all pregnant females were
considered in the analysis, a larger proportion carried eggs or embryos in the right uterus
than in the left uterus, but analysis of females carrying up to two eggs or embryos showed
that eggs or embryos were carried in similar numbers between the two uteri. Space in the
body cavity of viviparous sharks is important during embryonic development (Bass 1973),
particularly for species carrying relatively large-sized embryos, like S. megalops. In this
shark, the stomach is positioned on the left side of the body cavity; thus, when carrying
more than two embryos, space would be maximized if females hold more embryos in the
right uterus.
The litter size and embryo length of S. megalops increased with maternal TL. The pattern
of increasing number of embryos (Hanchet 1988; Yano and Tanaka 1988; Taniuchi et al.
1993) and length of near-term embryos (Hanchet 1988; Guallart and Vicent 2001) with
maternal TL is reported for other squalid species. This pattern is also observed in S.
megalops from South Africa and it may be related to an increase in space in the body
cavity (Watson and Smale 1998).
Ovulation and parturition in S. megalops exhibit no pattern of temporal periodicity,
suggesting that this shark is an asynchronous breeder. Most viviparous chondrichthyans
have synchronous mating (Hamlett and Koob 1999), although in a few species mating is
asynchronous (Yano and Tanaka 1988; Yano 1993, 1995). For chondrichthyans with
synchronous mating, the largest follicle diameter (LFD) and the size of the embryos are
recorded through time to determine the ovarian cycle and gestation period. However, this
method cannot be applied for species with asynchronous mating given that follicles or
embryos at very different stages of development are found at all times of the year. The
ovarian cycle of three deepwater squalid species could not be determined using this
method (Yano and Tanaka 1988). Watson and Smale (1998) used a similar approach to
112
estimate the gestation period of S. megalops without any success. In the present study,
ovarian cycle and gestation period were determined using the linear and Gompertz growth
models respectively, on a subset of data from different years. These models gave a good fit
to the data and allowed an approximate determination of the ovarian cycle and gestation
period. For other shark species, the linear (Walker 2005) and Gompertz (Hanchet 1988)
models have been used successfully for determining periodicity of ovarian cycle and
gestation period, respectively.
Squalus megalops has an ovarian cycle and gestation period of ~2 years. The periodicity of
the ovarian cycle and gestation are crucial for defining maternal condition of female
chondrichthyans; they need to be determined for population assessments of chondrichthyan
species. Most viviparous sharks have gestation periods of approximately a year (Stevens
and McLoughlin 1991; Hamlett and Koob 1999). However, for species producing largesized follicles, such as most squalid species (Chen et al. 1981; Hanchet 1988; Guallart and
Vicent 2001), ovarian cycle and gestation period are two, three or more years. Given that
the ovarian cycle and gestation period in S. megalops are biennial and that development of
follicles and embryos occurs concurrently, it is expected that parturition frequency for the
population is also biennial.
The different criteria used to calculate the maturity ogive of females are in reasonable
agreement in most cases. The condition of the reproductive tract and ovaries has been
commonly recorded to determine maturity of female chondrichthyans (Jones and Geen
1977b; Hanchet 1988; Watson and Smale 1998) though Walker (2005) proposed
measuring the diameter of the largest follicle (LFD) as an objective criterion of maturity
condition least prone to observer bias. To determine the onset of maturity of G. galeus, he
classed females as having reached the onset of maturity if LFD was >3 mm. In the present
study, the smallest value of L50 was obtained using this criterion. For S. megalops, follicles
<15 mm diameter were white, whereas follicles >15 mm were yellow, indicating that
vitellogenesis began at about this size. Furthermore, the ogive and the value of L50 obtained
using the criterion LFD 15 mm were in reasonable agreement with most of the other
criteria considered, suggesting that vitellogenesis starts when other reproductive structures
begin development. Thus, LFD 15 mm criterion was adopted for regional comparisons.
113
Differences in the maturity ogive of females from Lakes Entrance and Robe were found.
Spatial differences in size-at-maturity could occur when different age or size classes from
different locations are sampled or from length-selective fishing mortality (Walker 2005). In
the present study, regional differences in size-at-maturity could be a result of lengthselectivity of the 6½-inch gillnet used off Robe, selecting for the largest females and
possibly distorting the maturity ogive. Graham (2005) reported similar values of length-atmaturity for females collected from New South Wales. In South Africa, female S.
megalops also showed a similar length-at-maturity (Watson and Smale 1998) despite these
authors collecting a larger range of sizes (the largest female being 782 mm TL). This
suggests that females from New South Wales, South Africa and south-eastern Australia
would have similar maturity parameters. Taniuchi et al. (1993) reported spatial variation in
the length-at-maturity of female shortspine spurdogs (S. mitsukurii) from four different
locations off Japan and attributed it to differences in local environmental conditions.
However, they collected a different range of sizes from each location and their samples
from each location were obtained from different depths and years. Given that females of
squalid species can be segregated by stage of maturity and size (Yano and Tanaka 1988),
the geographical differences reported by these authors may be apparent and another
example of how using females of different size classes can distort maturity ogives.
The length at which 50% of the female population was pregnant was slightly larger than
the length at which 50% was mature. This suggests that once females attain maturity most
of them become pregnant soon after first ovulation and parturition there after. These
findings further support the hypothesis of a continuous breeding cycle. However, for
population assessment models it is important to distinguish the mature condition from
pregnant and maternal conditions. For species with reproductive cycles of several years
duration, a more critical relationship is the proportion of females in maternal condition.
Only half of the pregnant female population is in maternal condition in any year and
contributes to annual recruitment. The size of a population depends on the rates of birth,
death and migration. For viviparous chondrichthyans, birth rate can be calculated from the
number of females in the population, its fecundity rate and the proportion of females
contributing to annual recruitment (Walker 2005). Thus, for chondrichthyan species with
one year continuous reproductive cycles, calculation of population size can be performed
using maturity or maternity ogives as all mature females contribute to annual recruitment
114
each year. However, for species with a reproductive cycle of more than one year, such as S.
megalops, population size would differ depending on which ogive is used.
In conclusion, determining maternity ogives from information on the timing of ovulation,
period of gestation and parturition frequency is more complex for asynchronous species
than, as shown by Walker (2005), for synchronous species. Most squalid species are
deepwater asynchronous breeders with reproductive cycles of several years duration. Also,
many species are endemic and have restricted distributions. Given these biological and
ecological attributes, they are particularly vulnerable to fishing overexploitation.
Consequently, their populations require special management and a different approach to
determine reproductive parameters for population assessments. Reproductive parameters of
S. megalops were determined, despite this shark having an asynchronous reproductive
cycle. Mature males and females are capable of mating throughout the year and females
have a 2-year continuous cycle. Thus, although all females are mature at 600 mm TL, only
50% of them are in maternal condition, contributing to annual recruitment each year.
Hence, for chondrichthyan species with reproductive cycles of two, three or more years, if
maturity ogives are used in population assessments instead of maternity ogives, models
will over-estimate recruitment rates.
115
Injecting live specimens with OTC (photo by the author).
116
CHAPTER 6
117
CHAPTER 6 PREAMBLE
Chapter 6 compares different deterministic growth models fitted to length-at-age data
collected from first dorsal fin spines of S. megalops and discusses the implications of
sampling bias, length-selective fishing mortality, length-selective migration, and bias in
age estimation in the selection of the best growth model. At the time this thesis was
submitted (January 2006), this chapter was under peer-review with the journal Marine
Ecology Progress Series, with myself as senior author, and Bronwyn M. Gillanders (The
University of Adelaide), Terence I. Walker (Primary Industries Research Victoria), and
Javier Tovar-Avila (Primary Industries Research Victoria) as co-authors.
I was responsible for sampling, analysing and interpreting the data, and for writing the
manuscript. Bronwyn M. Gillanders and Terence I. Walker supervised development of
research, data interpretation and manuscript evaluation, and Javier Tovar-Avila read a subsample of 50 spines for evaluation of between-reader variability and between-reader bias
and also helped in manuscript evaluation.
118
CHAPTER 6
COMPARISON OF DETERMINISTIC GROWTH MODELS FITTED TO
LENGTH-AT-AGE DATA OF THE PIKED SPURDOG (SQUALUS MEGALOPS)
IN SOUTH-EASTERN AUSTRALIA
6.1 ABSTRACT
Age and growth estimates of Squalus megalops were derived from the first dorsal fin spine
of 452 sharks, ranging from 274–622 mm total length. Age bias plots and indices of
precision indicated the ageing method was precise and unbiased. Edge analysis of the
enameled surface of whole spines and similarities in the banding pattern laid in the
enameled surface of spines and in spine sections support the hypothesis of annual band
formation. Five growth models were fitted to length-at-age data from which a two-phase
von Bertalanffy model produced the best fit. However, model selection cannot be based on
quality of statistical fit only. Length-at-age data might not be representative of real growth
due to a combination of sampling bias, length-selective fishing mortality and/or bias in age
estimation. Regardless of the growth model used, growth rate of females (0.034–0.098
years–1) was very low, making S. megalops highly susceptible to fishing overexploitation.
6.2 INTRODUCTION
The most commonly used model to describe growth of elasmobranchs has been the von
Bertalanffy function (von Bertalanffy 1938) despite criticism (Knight 1968; Roff 1980). As
indicated by Carlson and Baremore (2005), few studies on elasmobranch growth have
examined alternative models and most studies simply fitted the von Bertalanffy function to
the data without much concern about the quality of the fit or the biological meaning of the
results. Hence, a range of growth models should be compared to determine the function
that provides the best description of the growth process (Haddon 2001).
Age and growth rates have been mainly studied for commercially important
elasmobranchs, such as spiny dogfish (Squalus acanthias), blue (Prionace glauca), gummy
(Mustelus antarcticus), and school (Galeorhinus galeus) sharks; however, little is known
about the age and growth of non-commercial squalid sharks (Squalidae). Although dogfish
are amongst the most abundant demersal sharks of temperate seas (Compagno 1984), most
119
of the ageing studies on this family have focused on S. acanthias (e.g. Holden and
Meadows 1962; Ketchen 1975; Beamish and McFarlane 1985). For this species, maximum
age varied widely, with a reported maximum age of up to 80 years (McFarlane and
Beamish 1987a). For other species of Squalus, age and growth rate have been estimated for
the shortspine spurdog (S. mitsukurii) in the North Pacific Ocean (Wilson and Seki 1994;
Taniuchi and Tachikawa 1999) and the longnose spurdog (S. blainvillei) in the
Mediterranean Sea (Cannizzaro et al. 1995). Age and growth rate information of the piked
spurdog (S. megalops) were estimated for sharks from South African waters (Watson and
Smale 1999). Males and females had different growth rates and maximum ages; the largest
male was 572 mm total length (TL) and 29 years old whereas the largest female was 782
mm TL and 32 years old.
Squalus megalops is a demersal species that is distributed off southern and eastern
Australia, from Carnarvon (Western Australia) to Townsville (Queensland), including
Tasmania (Last and Stevens 1994). However, the distribution of this species needs further
revision as it has also been reported off the coasts of Brazil (Vooren 1992) and South
Africa (Bass et al. 1976) and there are unconfirmed reports off Indo China, New Caledonia
and New Hebrides (Last and Stevens 1994). This species inhabits the continental shelf and
upper continental slope (depths <510 m) in warm temperate and tropical areas (Last and
Stevens 1994). Squalus megalops has a high natural abundance in southern Australia
(Bulman et al. 2001; Graham et al. 2001) and, although this shark is one of the major bycatch shark species in the area (Walker et al. 2005), its abundance has remained stable off
New South Wales (southeast coast of Australia) since it was first surveyed in 1976–77
(Graham et al. 2001). At present, the lack of biological data hampers a classification of the
conservation status of this species (Castro et al. 1999).
Until the present study, the age and growth rate of the Australian population(s) of S.
megalops remained unknown. Given that age and growth parameters can vary among
regions (e.g. Parsons 1993; Taniuchi and Tachikawa 1999) due to differences in
environmental conditions (Francis 1988), age and growth information for S. megalops
from Australian waters is required for population assessment of this species. Therefore, the
purpose of the present study was to estimate the age of S. megalops captured in southeastern Australia and compare different growth models to determine which model provides
the best fit to the growth data.
120
6.3 MATERIALS AND METHODS
Sampling
Specimens of S. megalops were obtained from the by-catch of demersal trawl and shark
gillnet vessels operating in the Southern and Eastern Scalefish and Shark Fishery in waters
off south-eastern Australia. Samples were collected monthly between October 2002 and
April 2004, with the exception of the August–September period, when S. megalops seemed
to move off the fishing grounds and weather conditions restricted sampling. The specimens
were sexed and measured (total length, TL, ±1 mm).
Spine and vertebrae preparation
Two portions of the spinal column (post cranial and thoracic vertebrae) and the first and
second dorsal fin spines (DFS) were removed and stored frozen for age estimation. Spines
were extracted by cutting horizontally just above the vertebral column to ensure that the
spine base and stem were intact (Beamish and McFarlane 1985). Soft tissue was removed
by immersing vertebrae and spines in hot water (55º C) for 0.5–1 minutes and trimming off
the skin, flesh, and cartilage with a scalpel. Cleaned spines were then rubbed with a cloth
to highlight the bands on the enameled surface. For vertebrae, the remaining tissue was
removed by soaking them in 4% sodium hypochlorite solution. Soaking time varied with
the size of the vertebrae between 10 and 20 minutes to avoid ‘over-bleaching’. Spines were
air-dried and stored in paper envelopes, whereas vertebrae were stored in a freezer.
Different measurements were recorded on the spines using electronic calipers to the nearest
0.01 mm following Ketchen (1975) (Fig. 6.1a).
Age estimation
A pilot study was carried out to determine which of the two structures—vertebrae or
spines—was more appropriate for age estimation. Whole and cross-sectioned vertebrae
(∼300 µm) were used. The sections were obtained by embedding vertebrae in epoxy resin
and sectioning with a GemmastaTM lapidary saw with a diamond-impregnated blade.
Sections were cleaned using ethanol and water, air-dried and mounted on glass slides using
epoxy resin. Sections were examined under transmitted light using a Leica IM 4.0 digital
image system on a Leica DMLB compound microscope. Vertebrae of S. megalops are
poorly calcified, with very fragile intermedialia, so, from sectioned vertebrae, only the
corpus calcareum was used. No clear banding pattern was observed on whole or sections
of vertebrae; hence, different stains (alizarin red S, silver nitrate, cobalt nitrate, ninhydrin)
121
were used to enhance readability. However, no improvement in readability was observed,
so only spines were used for further age estimation.
Bands laid on the external enameled surface of spines were counted using a dissecting
microscope (10×) and reflected light (Fig. 6.1b). A band was defined as an alternating
opaque and translucent zone or ridge or both, present on the enameled surface (McFarlane
and Beamish 1987a; Watson and Smale 1999). Solid bands on the leading edge of the
spine that were split on the trailing edge were counted as a single band (Watson and Smale
1999). Bands visible only as a dark mark on the leading edge but similar in thickness to
neighbouring bands that did continue to the trailing edge were also counted as single bands
(Watson and Smale 1999). A subjective readability score was assigned to each spine
following Officer et al. (1996) (Table 6.1). After three readings, spines having a readability
score of 4 (ambiguous band counts) or 5 (no band counts possible) were not used for
further analysis (<9% of readings).
The spines of 10 full-term embryos were examined to determine pre-birth bands. No bands
were observed and it was assumed that the first band was laid just prior to or after birth
(birthmark) (e.g. Holden and Meadows 1962; Moulton et al. 1992). Thus, the total number
of bands was calculated as the total number of bands counted minus one.
To determine whether the first or the second DFS was more appropriate for age estimation,
a random sub-sample of first and second DFS from 61 animals was read two times
separated by a minimum of one month without knowledge of length of animals. The
readability scores of the first and second DFS were compared. The coefficient of variation
(CV; Chang 1982; Campana et al. 1995) and the index of average percentage error (APE;
Beamish and Fournier 1981) were calculated to evaluate precision between readings (first
vs second reading) and between methods (first vs second DFS). Age bias plots (Campana
et al. 1995) were used to detect count differences between the two methods. The first DFS
showed better readability scores and higher precision between readings (see “Results”),
hence, this spine was used for age estimation.
122
Table 6.1. Readability scores assigned to readings of spines (modified from Officer et al.
1996).
Readability score
Description
1
Band count unambiguous with clear bands.
2
Band count unambiguous but bands of diminished clarity.
3
Two band counts possible but indicated count is most likely.
4
More than two interpretations possible; count is best estimate.
5
No band count possible; unreadable.
123
(a)
STL
NWP
SBD
(b)
(c)
ID
Figure 6.1. Dorsal fin spine of Squalus megalops. (a) Lateral view of worn second DFS
showing the different measurements recorded following Ketchen (1975): spine total length
(STL), spine base diameter (SBD) and no wear point (NWP). (b) Anterior view of first
DFS (6×) showing 20 bands on the enameled surface. (c) Section of first DFS (100×)
showing 17 bands on the inner dentine (ID) layer.
124
All first DFS were read three times by a single reader (first reader) separated by a
minimum of one month between readings. A second reader read a random sub-sample of
50 spines for evaluation of between-reader variability and between-reader bias. To evaluate
within- and between-reader precision, the CV and the APE index were calculated. Age bias
plots were used to detect systematic count differences between the first and second reader.
To accept a count for age estimation, the counts of at least two of three readings had to be
identical. If counts from two of the three readings differed, spines were recounted a fourth
time and the same procedure was applied. If the difference persisted, the spine was
discarded (<9% of readings).
For worn spines, Ketchen’s (1975) correction method was adopted to ensure that bands
were not missing. The relationship between band counts and spine base diameter (SBD)
was estimated for unworn spines from males (n = 45) and females (n = 46). For worn
spines, the diameter of the spine at the most distal point of no wear (NWP) was then
measured (Fig. 6.1a). From the band counts–SBD relationship, the number of bands
corresponding to the diameter at the NWP of worn spines was calculated and added to the
original count of bands (Ketchen 1975).
Verification and validation
A random sub-sample of spines from 10 female and 10 male sharks was sent to two
international experts who agreed that the spines were appropriate to estimate the age of S.
megalops. The annual periodicity of band deposition on whole spines was investigated by
analysing the edge of their enameled surface (Holden and Meadows 1962; Nammack et al.
1985; Taniuchi and Tachikawa 1999). The edge of spines collected throughout the year
was classified as dark, light or wide light following Holden and Meadows (1962) (Table
6.2).
For 89 sharks, counts on spine sections were compared with counts on the external
enameled surface. Serial sections (∼300 µm) were taken from the tip of each first DFS to
determine the optimal position of sectioning. The same method for sectioning vertebrae
was used to obtain spine sections. The inner dentine layer showed the clearest banding
pattern and was hence used for counting (Fig 6.1c) (Maisey 1979; Clarke et al. 2002a).
Maximum count of bands in the inner layer was found at the apex of the pulp cavity,
representing the optimal position of sectioning. The sections were examined under
125
Table 6.2. Definition of edge type of whole spines for the edge analysis following Holden
and Meadows (1962).
Edge type
Description
Dark
Dark band observed at the edge of the enamel.
Light
Dark band just formed and a light band of a width less than half the
width of the light band between the last two dark bands observed at
the edge of the enamel.
Wide light Light band observed at the edge of the enamel equal to or more than
half the width of the light band between the last two dark bands.
126
transmitted light using a Leica IM 4.0 digital image system on a Leica DMLB compound
microscope. Within the internal dentine layer, a band was defined as a pair of dark
(opaque) and light (translucent) concentric rings (Irvine 2004). Counting started at the pulp
cavity (centre) and continued outwards to the junction between inner and outer dentine
layers (Irvine 2004). Age bias plots (Campana et al. 1995) were used to detect count
differences between external (enameled surface) and internal (sections) counts.
Validation of the periodicity of band deposition was attempted by injecting captive S.
megalops with oxytetracycline (OTC). Twelve male and 12 female S. megalops of
different sizes were measured, implanted with roto-tags for individual identification,
injected with OTC at 25 mg per kg body mass (McFarlane and Beamish 1987b), and kept
in a 27,000-L outdoor aquarium. Captive sharks were subjected to natural variation in
water temperature and photoperiod and they were fed on a diet of squid and fish twice a
week.
Growth estimation
Based on the assumption that external bands were formed annually, for each sex, several
alternative growth models were fitted to the length-at-age data: the traditional von
Bertalanffy growth model (VBGM; von Bertalanffy 1938), a two-parameter modified form
of VBGM (2VBGM; Fabens 1965), a two-phase von Bertalanffy growth model
(TPVBGM; Soriano et al. 1992), the Gompertz growth model (Ricker 1975), and a twoparameter modified form of the Gompertz growth model (2Gompertz; Mollet et al. 2002)
(Table 6.3). Model parameters were estimated by least-squares non-linear regression.
Akaike’s Information Criterion (AIC) was used to determine the model that best fitted the
length-at-age data (Buckland et al. 1997; Burnham and Anderson 2002).
AIC = n ln (σˆ 2 ) + 2p
where n = sample size;
Residual sum of squares
; and
n
p = number of parameters.
σˆ =
127
Comparisons among the AIC values of the different growth models enabled the best model
for each sex to be selected, i.e. those models with the lowest AIC values. For model
comparisons, the delta AIC (∆AIC) and Akaike weights (wi) were calculated. The ∆AIC is
a measure of each model relative to the best model and is calculated as:
∆AIC = AICi – minAIC
where AICi = AIC value of model i; and
minAIC = AIC value of the best model.
Akaike weights represent the probability of choosing the correct model from the set of
candidate models and are calculated as:
wi =
exp(-∆AIC/2)
¦
R
r =1
exp(-∆AIC/2)
where R = number of candidate models.
Once the best model was determined, the growth curves of males and females were
compared by a Chi-square test on likelihood ratios (Kimura 1980; Cerrato 1990).
128
Table 6.3. Summary of growth models fitted to length-at-age data.
Model
Equation
VBGM
Lt = L∞(1 – e–k (t – t°))
2VBGM
Lt = L∞(1 – be–k t), b = (L∞ – L0)/ L∞
TPVBGM
Lt = L∞(1 – e–kAt (t – t°)), At = 1 – h/((t – th)2 + 1)
Gompertz
Lt = L∞ e–e(–k(t – t°))
2Gompertz Lt = L0(eG(1–e(–k t))), G = ln(L∞ / L0)
Meaning
Lt = mean length at time t
of terms
L∞ = theoretical asymptotic length
k = growth coefficient
t0 = theoretical age at zero length
h = magnitude of the maximum differences between VBGM and
TPVBGM
th = age at which transition between the two growth phases occurs
L0 = mean length at birth (214 mm for males and females)
129
6.4 RESULTS
Age estimation
The first dorsal fin spines (DFS) provided better readability scores and more precise
readings than the second DFS. Most first DFS had a readability score of 2 (51.5%) or 3
(42.4%), whereas most of the second DFS had a readability score of 3 (62.1%) or more
(Fig. 6.2a). Differences between the readings for first and second DFS varied by up to five
bands, but differences were mostly ±1 count (Fig. 6.2b). Mean coefficient of variation
(CV) and index of average percentage error (APE) among readings were 7.53% and
5.15%, respectively, for the first DFS and 9.03% and 6.39%, respectively, for the second
DFS, indicating more precise counts were obtained when the first DFS was used. Mean CV
and APE between the first and second DFS were 14.04% and 9.93%, respectively.
Agreement between first and second DFS decreased with the number of bands counted,
indicating the use of the second DFS systematically underestimates the number of bands
(Fig. 6.2c). The first DFS was used for age estimation because it showed a clearer
readability pattern, higher precision between readings, and an overall higher number of
bands.
The relationship between the first DFS length and total length was linear (DFS length =
2
0.086 TL – 2.739; r = 0.9) and there was no significant difference between males and
females (t-test, t = 1.77, d.f. = 98, P >0.05 for comparison of slopes, and t = 1.07, d.f. = 98,
P >0.05 for comparison of elevations). The increase in DFS length with total length shows
that spines grow throughout life, indicating this structure is useful for age estimation.
A total of 493 first DFS was examined of which 41 (8.3%) were rejected because they did
not conform to the selection criteria (i.e. readability score 3 and identical counts from at
least two of three or four readings). Band counts from 163 males (274–470 mm TL) and
289 females (287–622 mm TL) were used for age estimation. Within-reader precision
among readings was high; mean CV and APE were 6.99% and 5.25%, respectively.
Overall, for each band class CV was low, showing the lowest values for mid-band classes
(14–23 band class) (Fig. 6.3). A similar pattern was observed for APE. Mean CV and APE
between readers were 11.35% and 8.03%, respectively. There were no appreciable
systematic differences between readers (Fig. 6.4). Worn spines (4.9% males and 15.9%
females) were corrected for missing bands using the equations derived from the
130
50
Number of spines
(a)
DFS
40
1
2
30
20
10
0
1
2
3
4
5
6
Readability score
30
Number of spines
(b)
DFS
1
20
2
10
0
-5 -4 -3 -2 -1 0 1 2 3 4 5
Count difference
2
Mean count from second
DFS ±95% CI
22
20
(c)
2
18
16
14
12
1
7
6 5
2
4
10
8
6
4
3
3
1
5 8
11
1
4 6 8 10 12 14 16 18 20
Count from first DFS
Figure 6.2. Comparison between first and second dorsal fin spines (DFS) for 61 sharks. (a)
Distribution of readability scores assigned to readings of each spine. (b) Distribution of
differences between two readings on first and second dorsal fin spines. (c) Age bias plots.
The solid line is the 1:1 relationship. Sample sizes are given above each corresponding
count.
131
30
CV (%) ±s.e.
1
20
7
3
15
30
10
14
40
34
21
12
41
27 31
16
29
5
16
4
7
1
6
2
8
3
1 2
0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28
Number of bands
Figure 6.3. Mean values of coefficient of variation (CV) for each band class read by the
principal reader. Sample sizes are given for each corresponding band class.
132
relationship between the number of bands and SBD of unworn spines (Counts male = 0.394
SBD 3.064; r2 = 0.83; Counts female = 0.965 SBD 2.083; r2 = 0.94).
Verification and validation
The nature of the enamel edge varied with month (Fig. 6.5). At no time of the year were
the spines entirely of one edge type. Most sharks collected between May and July
(autumn–winter) had dark edges, whereas the highest percentage of light edges occurred in
sharks collected between October and December (spring–early summer). For spines with
wide light edges, the highest percentage was found in sharks collected between January
and April (summer–autumn). This annual edge pattern suggests that dark rings form during
the cold period of the year. In addition, there was good agreement between counts on spine
sections and counts on the external enameled surface (Fig. 6.4).
After a period of five months, captive sharks died due to a sudden increase in water
temperature (from 12º C in late winter to 22º C in early summer). Also, the sharks showed
signs of stress when approached, bumping the sides of the tank. Hence, validation of
annual band deposition on spines of captive S. megalops was not achieved. Unlike other
shallow-water sharks previously kept in similar conditions, S. megalops is a mid-water
species (depths <510 m) which seems to be more sensitive to changes in water temperature
than its shallow-water counterparts.
Growth estimation
Males ranged 1–15 years in age, whereas females reached a maximum age of 28 years.
Most males were 11–12 years whereas most females were 13–14 (Fig. 6.6). The value of
growth parameters of S. megalops were estimated separately from five models fitted to the
length-at-age data (Table 6.4). Growth models fitted the data well, with females showing
higher values of coefficient of determination (r2 ≥0.88) than males (r2 ≥0.72).
For males, the TPVBGM was the best of the five growth models fitted with an Akaike
weight (wi) of 0.54 (Table 6.4). However, VBGM and Gompertz growth model followed
rather closely (wi = 0.24 and 0.19, respectively), revealing a certain degree of uncertainty
regarding the best model for fitting length-at-age data of males. Estimates of asymptotic
length (L∞) varied among models. Traditional VBGM and TPVBGM predicted a slightly
133
Mean count from Reader 2 ±95% CI
32
30
28
26
24
22
20
18
16
14
12
10
8
6
4
2
0
2
22
1
1
1
3
3
1 1
3 1
3
1
2
2
3
3 22
3 1
2
3
1
1
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32
Mean count from sections ±95% CI
Count from Reader 1
30
28
26
24
22
20
18
16
14
12
10
8
6
4
2
0
3 2
33
2
44
5
333
3
3
3 4
3 3
2
3
2
35
5
35
5
2
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30
Count from enameled surface
Figure 6.4. Age bias plot for the comparison of band counts between readers and band
counts on spine sections and on the external enameled surface. The solid line is the 1:1
relationship. Sample sizes are given above each corresponding count.
134
100
Percentage edge type
80
28 54 25 35 9
27 7
36 31 45
60
40
Edge type
WL
20
N.D.
0
L
D
Jan Feb Mar Apr May Jun Jul Aug SeptOct Nov Dec
Month
Figure 6.5. Monthly variation in the type of enamel edge; wide light (WL), light (L) and
dark (D). Sample sizes are given for each corresponding month; N.D.: no data.
135
30
male
Number of sharks
female
20
10
0
0
2
4
6
8 10 12 14 16 18 20 22 24 26 28 30
Age class (years)
Figure 6.6. Age distribution for male and female S. megalops.
136
larger value of L∞ (L∞ = 455 mm TL) than Gompertz growth model, 2VBGM, and
2Gompertz growth model (L∞ = 449, 440, and 433 mm TL, respectively). Estimates of
growth coefficients (k) also varied among models, with 2Gompertz growth model and
2VBGM producing the highest values (k = 0.252 and 0.198 years–1, respectively).
For females, the best fitting model was the TPVBGM as the probability of choosing this
model as correct (wi) was 0.95 (Table 6.4); other models showed lower values of wi
indicating they do not fit the length-at-age data as well. There was high variability in the
estimates of L∞ among models. The 2Gompertz growth model, 2VBGM, and Gompertz
growth model predicted a lower value of L∞ (L∞ = 632, 699, and 717 mm TL, respectively)
than the predicted value of TPVBGM (L∞ = 756 mm TL) or VBGM (L∞ = 829 mm TL).
Estimates of k also varied widely among models, with 2Gompertz and traditional
Gompertz growth models producing the highest values (k = 0.098 and 0.063 years–1,
respectively). Traditional VBGM produced the lowest growth coefficient value (k = 0.034
years–1).
Given that the TPVBGM was the best fitting model, this function was used to draw the
growth curves of males and females (Fig. 6.7). Likelihood ratio tests indicated significant
differences between the growth curves (P <0.001) for the two sexes. Males (k = 0.158
years–1) grew faster than females (k = 0.042 years–1) (Table 6.4). Predicted length-at-age of
males was initially higher but after they attained 8 years growth of males and females was
similar up to ∼10 years when transition between growth phases occurred (Fig. 6.7; Table
6.4). After age 10 growth of males slowed down whereas growth of females continued
with length increasing steadily throughout their lifespan.
137
138
k (years–1)
0.252 (0.02)
433 (7)
–3.54 (1.68)
t0 (years)
10.5 (0.60)
th (years)
0.172 (0.05)
0.094 (0.08)
h
k (years–1)
–4.86 (2.10)
t0 (years)
449 (17)
0.158 (0.05)
k (years–1)
L∞ (mm)
455 (19)
L∞ (mm)
0.198 (0.02)
k (years–1)
2Gompertz L∞ (mm)
Gompertz
TPVBGM
440 (9)
–5.72 (2.50)
t0 (years)
L∞ (mm)
0.144 (0.05)
k (years–1)
2VBGM
455 (21)
L∞ (mm)
VBGM
Estimate
Parameter
Model
0.72
0.74
0.75
0.73
0.74
r
2
861.11
852.27
850.17
856.19
851.85
AIC
Males (n = 157)
10.94
2.10
0
6.02
1.68
∆AIC
<0.01
0.19
0.54
0.03
0.24
wi
0.098 (0.01)
632 (18)
0.026 (0.81)
0.063 (0.01)
717 (59)
10.2 (0.60)
0.087 (0.05)
–9.77 (2.60)
0.042 (0.01)
756 (88)
0.056 (0.01)
699 (33)
–10.83 (2.82)
0.034 (0.01)
829 (126)
Estimate
0.88
0.89
0.89
0.88
0.89
r2
9.50
∆AIC
6.49
0
1804.37 38.90
1771.92
1765.43
1791.11 25.68
1774.94
AIC
Females (n = 274)
<0.01
0.04
0.95
<0.01
0.01
wi
Table 6.4. Growth estimates (with 95% confidence intervals) and model selection criterion for male and female S. megalops. Akaike’s
information criterion (AIC); AIC differences between models (∆AIC); Akaike weights (wi); and sample size (n). Refer to Table 6.3 for meaning
of parameters.
700
(a)
650
600
550
500
450
400
350
Total length (mm)
300
250
n = 157
200
700
650
(b)
600
550
500
450
400
350
300
250
n = 274
200
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30
Age (years)
Figure 6.7. Two-phase von Bertalanffy growth model fitted to length-at-age data derived
from counts on the first dorsal fin spine of (a) male and (b) female S. megalops. Estimates
of model parameters are given in Table 6.4.
139
6.5 DISCUSSION
The first DFS of S. megalops showed clearer readability, higher precision between
readings and an overall higher number of bands than the second DFS. Most ageing studies
on squalid sharks use whole second DFS because it is larger and the tip of the first DFS
tends to be worn down (Cailliet and Goldman 2004). Although some studies have used the
first DFS as a check (e.g. Nammack et al. 1985), few have compared the suitability of the
first and second DFS for ageing (but see Holden and Meadows 1962; Irvine 2004). Given
the structure of the first and second DFS is similar (e.g. Holden and Meadows 1962;
Clarke et al. 2002a), preference for first or second DFS for age estimation should be a
question of readability. Sections of the first DFS of the birdbeak dogfish (Deania calcea)
(Clarke et al. 2002a) and the gulper shark (Centrophorus squamosus) (Clarke et al. 2002b)
provided better readability than sections from the second DFS. For S. megalops off South
Africa, Watson and Smale (1999) only used the second DFS, rejecting, after several
readings, 12% of the spines for age estimation whereas in the present study only 8% of the
spines used were rejected, probably reflecting the better readability of first DFS. Therefore,
for S. megalops, the clearer pattern of bands observed in the first DFS made this structure
easier to read and hence more suitable for age analyses than the second DFS.
The low values of the CV and the APE index for the within- and between-reader analyses
suggest high precision for the age assessment of S. megalops. These two indices assume
that variability among observations of individual fish can be averaged over all age classes,
obscuring differences in precision (Hoenig et al. 1995). However, when calculations were
made for each age class (within-reader only), between-age-class variability was low. A
trend for an increase in within-reader precision was observed for the mid-range age classes,
suggesting these classes are easier to read. The age bias plot indicated no bias in the age
estimation of Reader 1 (principal reader). Few studies on the age and growth of squalid
sharks provide estimates of precision and bias, despite their importance for any ageing
studies (Campana 2001). However, when such estimates are produced (Holden and
Meadows 1962; Ketchen 1975; Nammack et al. 1985; Clarke et al. 2002a, b; Irvine 2004),
most studies report high reproducibility of age estimates, further supporting the use of
spines as a precise approach for ageing squalid sharks.
The analysis of the enamel edge of whole spines supports the hypothesis of annual band
formation in S. megalops. The peak in dark bands observed during late autumn–winter
140
(May–July), followed by the peak in light bands during spring–early summer (October–
December), and by the peak in wide light bands during early autumn (March–April) is the
expected pattern for annual band deposition. A similar pattern is reported for S. acanthias
off north and west Scotland (Holden and Meadows 1962) and off north-eastern USA
(Nammack et al. 1985). For S. acanthias, the timing of light and dark band formation on
the enamel edge of spines was validated using mark-recapture oxytetracycline (OTC)tagged sharks (Tucker 1985).
Band counts on the inner dentine layer (spine sections) of first DFS of S. megalops were in
good agreement with counts on the enameled surface, verifying the age estimates obtained
from counts on the enameled surface of spines. Counts of bands on the inner dentine have
been used for age estimation of other squalid sharks (e.g. Clarke et al. 2002a, b; Irvine
2004). For S. acanthias, comparisons of internal (sections) and external (enameled surface)
counts agreed within ±2 bands (Holden and Meadows 1962). However, for the deepwater
velvet dogfish (Centroscymnus crepidater), Portuguese dogfish (C. coelolepis), and New
Zealand lantern shark (Etmopterus baxteri), the number of external bands exceeded the
number of internal bands in sharks older than 3–5 years (Irvine 2004). As for S. acanthias,
sections of first DFS of S. megalops confirmed the use of the enameled surface of whole
spines as a method of age estimation as both methods yielded similar counts.
Several authors (e.g. Beamish and McFarlane 1983; Cailliet 1990; Campana 2001) have
stressed the need for validation of the temporal periodicity of band deposition and of the
absolute age for accurate age estimation. Captive rearing of OTC-tagged S. megalops was
conducted as an attempt to validate the periodicity of band deposition. However, as with a
similar study in South Africa (Watson and Smale 1999), holding the sharks in captivity
was of limited success. More rigorous methods of age validation, such as the bomb
radiocarbon method (Druffel and Linick 1978), is not applicable for S. megalops as
samples of sharks born during the period of 14C increase (1955–1970) are not available.
Annual deposition of bands on spines of squalid sharks has been validated for S. acanthias
and this was on both sections and the enameled surface of whole spines (Beamish and
McFarlane 1985; Tucker 1985; McFarlane and Beamish 1987a). Although it does not
necessarily follow that these bands are annual in other squalid sharks (Clarke et al. 2002b),
for S. megalops, the most parsimonious interpretation of available evidence (edge analysis
and comparisons of internal and external bands) points to annual formation of bands.
141
However, further research is needed to confirm the annual pattern of band deposition and
absolute age of S. megalops. A pilot tagging study showed promising results (9 sharks
recaptured out of 617 tagged; Brown et al. 2000); hence, a large-scale release-recapture
program of known age and marked sharks or OTC-tagged sharks (Beamish and McFarlane
1985; Campana 2001) would provide information on periodicity of band deposition.
Growth model selection is not a straightforward process. Based on the goodness-of-fit
criterion used (AIC), the best model for both sexes was the TPVBGM. The use of a
TPVBM allows dividing growth into two phases, such as changes in habitat (e.g. from
coastal to off-shore waters), feeding habits (e.g. from a planktivorous to a piscivorous diet)
or energy allocation (e.g. from energy allocated to growth, as in juveniles, to energy
allocated to growth and reproduction, as in adults). For male S. megalops the change in
growth rate (∼10 years) corresponded with size at maturity whereas for females, the change
(∼10 years) was slightly before the size females start to mature (Braccini et al. 2006;
Chapter 5). However, model selection is a matter beyond the quality of statistical fit
(Haddon 2001). Proper description of the growth process also requires biological realism.
In addition to determining the best goodness-of-fit, the quality of the data used in the
fitting process and the shape of the growth curve are of similar importance in the
description of the growth process, particularly when samples are collected from
commercial fishing operations. When growth models are fitted to these type of data the
resulting parameters may be different from those obtained from a more representative
sample (Haddon 2001). Most studies on age and growth of elasmobranchs obtain samples
from commercial fisheries and assume that data are unbiased. However, the length–age
scattergram and the growth curve of S. megalops, particularly females (see Fig. 6.7 present
study; Watson and Smale 1999), and other shark species (e.g. Moulton et al. 1992;
Cannizzaro et al. 1995) does not asymptote suggesting that either samples are not
representative of the entire population or those species do not exhibit an asymptotic
maximum length and hence asymptotic models are not adequate to describe growth. Linear
growth for long-lived species, such as sharks, has only been reported for juveniles (e.g.
Simpfendorfer 2000) or when sampling is not representative of all size classes (e.g.
Wintner 2000). Hence, it is more likely that the observed scattergram and the subsequent
growth curve derived are a result of unrepresentative data due to a combination of several
factors. Length-selective sampling bias and length-selective fishing mortality of gillnets
142
can cause distortions to growth curves (Moulton et al. 1992; Walker et al. 1998). In the
present study, most of the samples were collected from demersal otter trawl and Danish
seine vessels. Although the selectivity of the mesh of the trawl-codend is not adequately
understood, length-selective sampling bias and fishing mortality might contribute to
explaining the shape of the growth curve and the scatter of data points for S. megalops.
Parturition is likely to occur outside the fishing grounds (Graham 2005). Larger neonates
and juveniles are thought to move to the fishing areas to feed on trawl discards or prey
exposed by trawl operations disturbing sediments, as S. megalops is an opportunistic
predator that consumes a wide range of prey items (Braccini et al. 2005; Chapter 4).
Hence, the larger neonates and juveniles would then become available to sampling by the
trawling gear and thereby create a bias in the size distribution of neonates and juveniles
sampled due to a higher probability of collecting large fast-growing individuals rather than
small slow-growing individuals. For intermediate-aged sharks, the large fast-growing
individuals of an age class may have been selectively removed from the population by
fishing (Walker et al. 1998; Haddon 2001). This length-selective removal of the largest
sharks of the available age classes by fishing has a biasing effect when sampling these age
classes. Alternatively, due to the strong size segregation exhibited by S. megalops (Graham
2005; Chapter 2), if certain size class or size classes, for example the largest females of
each age class, occurred outside the trawling areas only the smallest individuals would be
available for sampling and the size-at-age frequency distribution of the age classes would
be biased. Finally, age underestimation of older fish due to difficulties in band counting
and/or poor representation of older fish in the sample due to their low natural abundance
would also explain the lack of an asymptote in the growth curve (McFarlane and Beamish
1987a; Watson and Smale 1999). Hence, even when growth of S. megalops follows a
VBGM or other growth model, the collected length-at-age data may be variously affected
by sampling bias, length-selective fishing mortality, length-selective migration, and bias in
age estimation.
Precision estimates, the relationship between spine total length and TL, edge analysis, and
agreement between counts on the inner dentine layer and the enameled surface support the
use of the first DFS for the age estimation of S. megalops. Based on goodness-of-fit
criterion, the best growth model for males and females was the TPVBGM. However,
model selection cannot be based on quality of statistical fit only and given that length-atage data might not be representative of real growth, results should be interpreted with
143
caution. Regardless of the model used, the growth rate of S. megalops, particularly of
females, is very low, even within the range of growth rates reported for shark species
(0.03–1.337 years–1; Cailliet and Goldman 2004). The reproductive output of this species is
also low as litter size is at most four and the reproductive cycle is almost two years
(Braccini et al. 2006, Chapter 5). These characteristics indicate S. megalops has low
biological productivity and therefore higher risk to the effects of fishing than species with
higher biological productivity.
144
145
Lakes Entrance, some of the fishing vessels that operate in the mighty waters of the
Southern Ocean (photo by the author).
146
CHAPTER 7
147
CHAPTER 7 PREAMBLE
In Chapter 7 I applied a hierarchical framework for the assessment of the effects of fishing
on Squalus megalops by integrating qualitative and quantitative modelling tools. This is
the first study to use expert judgement, the precautionary approach and stochastic matrix
modelling to determine fishing effects on a chondrichthyan species. At the time this thesis
was submitted (January 2006), this chapter was under peer-review with the Canadian
Journal of Fisheries and Aquatic Sciences, with myself as senior author, and Bronwyn M.
Gillanders (The University of Adelaide) and Terence I. Walker (Primary Industries
Research Victoria) as co-authors.
I was responsible for sampling, analysing and interpreting the data, and for writing the
manuscript. Bronwyn M. Gillanders and Terence I. Walker supervised development of
research, data interpretation and manuscript evaluation.
148
CHAPTER 7
HIERARCHICAL APPROACH TO THE ASSESSMENT OF FISHING EFFECTS
ON NON-TARGET CHONDRICHTHYANS: CASE STUDY OF SQUALUS
MEGALOPS IN SOUTH-EASTERN AUSTRALIA
7.1 ABSTRACT
A three-levelled hierarchical risk assessment approach was trialed using piked spurdog
(Squalus megalops) to evaluate the suitability of the approach for chondrichthyan species.
At level 1, a qualitative assessment indicated that the only fishing-related activity to have
moderate or high impact on S. megalops was ‘capture fishing’ by otter trawl, Danish seine,
gillnet and automatic longline methods. At level 2, a semi-quantitative assessment ranked
S. megalops at risk because of its low biological productivity and, possibly, its catch
susceptibility from cumulative effects across the separate fishing methods. Finally, at level
3, a quantitative assessment showed that population growth is slow even under the
assumption of density-dependent compensation where the fishing mortality rate equals the
natural mortality rate. Although published information indicates relative abundance has
been stable in several regions of southern Australia, it is concluded that given its low
biological productivity, changed fishing practices leading to increased fishing mortality
could quickly put S. megalops at high risk. The hierarchical approach appears particularly
useful for assessment of chondrichthyan species in data-limited fisheries. This approach
allows for a management response at any level, optimising research and management
efforts by identifying and excluding low-risk species from data intensive assessments.
7.2 INTRODUCTION
Globally only limited attempts have been made to manage populations of chondrichthyan
species impacted by the effects of fishing (Anderson 1990; Bonfil 1994). The few
exceptions have focused on target species, such as gummy shark (Mustelus antarcticus)
(Walker 1998) and dusky shark (Carcharhinus obscurus) (Simpfendorfer 1999) in
southern Australia. Whereas biological parameter estimates and time series of catch,
fishing effort, and other monitoring data have been collected to enable stock assessment for
sustainable use of the high valued species, there is a paucity of such information for nontarget species (Bonfil 2004). Given the comparatively low biological productivity and
149
often high catch susceptibility of chondrichthyan species (Stobutzki et al. 2002; Walker
2004), management initiatives are needed long before sufficient data can be collected for
stock assessment (Walker 2004). Concerns about widespread depletion of chondrichthyan
populations led the United Nations Food and Agricultural Organization to develop an
International Plan of Action for the Conservation and Management of sharks (IPOAsharks; FAO 2000a) that was implemented during 1999.
More recently, to address the concerns of uncertainty associated with the wider impacts of
fishing on marine ecosystems, Australia is developing and implementing a broad process
for ecological assessment in a risk framework. This process is referred to as ecological risk
assessment and explicitly identifies five ecological components for analysis: target species,
non-target species comprising by-product (predominantly retained) and by-catch
(predominantly discarded), threatened species, fish habitats, and ecological communities.
Within each component, for each type of fishing method separately, the approach involves
a three-levelled hierarchical process of assessment, with increasing data requirements and
complexity when progressing from level 1, through level 2, to level 3 assessment. Level 1
assessment involves expert judgement and determines whether or not there is a need to
progress to level 2 or, alternatively, to implement a management response. Level 2
assessment is semi-quantitative and determines whether or not there is a need to progress
to level 3 or, alternatively, to implement a management response. When progressing from
one level to the next, depending on costs, there is the choice of either immediately
initiating a management response to ameliorate risk of adverse effects or, alternatively,
proceeding to invest in research and monitoring to enable a higher level of assessment
(Hobday et al. 2004). This approach to ecological risk assessment would be exhaustive if
all components were taken to level 3 and would require excessive costs in a multi-species
fishery.
The purpose of the present study is to simply take a single chondrichthyan species—the
piked spurdog (Squalus megalops)—for which sufficient data are available to undertake an
assessment at each of the three levels and to evaluate the suitability of the approach for
chondrichthyan species in general. Squalus megalops falls within the non-target
component and is one of the most abundant and widespread chondrichthyan species
impacted by Australia’s Southern and Eastern Scalefish and Shark Fishery (SESSF). The
species is mostly taken as by-catch by demersal trawl on the continental shelf and upper
150
slope (~600 T/year) (Walker and Gason 2006). Small quantities are also taken on the upper
slope by automatic longline targeted at teleost species and on the continental shelf by
gillnets deployed in the targeted shark fishery for M. antarcticus (Walker et al. 2005;
Walker and Gason 2006). Following long-term declines in abundance of shark and
chimaera populations off New South Wales, southeast coast of Australia (Graham et al.
2001), quota reductions on target and by-product shark and chimaera species, and growing
consumer demand for shark meat, some large individuals of S. megalops are beginning to
be retained for marketing (Walker and Gason 2005).
7.3 MATERIALS AND METHODS
Squalus megalops was assessed using the Australian process for ecological risk assessment
as part of the non-target component at each of the three hierarchical levels. Progressing
through the three levels, level 1 involved qualitative assessment based on expert
judgement, level 2 involved semi-quantitative assessment, and level 3 involved fully
quantitative assessment based on available data. The approaches to level 1 and level 2
assessments are explicitly described (Hobday et al. 2004) and Australia has extensive
experience with stock assessment of target and by-product species in a risk framework
(level 3 assessment). However, the process is not presently explicit for level 3 assessment
of non-target species where catches are not known precisely. The present study applies an
approach which assumes no knowledge of catches and catch rates.
Level 1 assessment
Level 1 assessment was a qualitative analysis for each fishing method that may impact S.
megalops based on expert knowledge. Adopting the precautionary approach, the species
was assigned the highest risk value if there was uncertainty about risk judgement (Hobday
et al. 2004). The fishing methods in the SESSF that may have an effect on S. megalops in
southern Australia are otter trawl and Danish seine nets, demersal shark gillnets, demersal
shark longlines, automatic longlines, droplines, and traps and pots. Within these fishing
methods, seven associated activities, explained in Table 7.1, may impact S. megalops. Each
fishing-related activity within the non-target component was assessed using spatial and
temporal scale, intensity and consequence analysis (Hobday et al. 2004). This approach
involved assigning a score to each of spatial and temporal scale, intensity and consequence
of the fishing activity (explained in Table 7.2). Fishing methods with fishing activities with
151
a consequence score ≤2 were eliminated from further assessments, whereas methods with
higher scores were assessed more in-depth at level 2.
Table 7.1. Description of the potential impacts of different fishing activities on Squalus
megalops. Direct impacts are impacts causing damage or mortality, whereas indirect
impacts are impacts altering the habitat of the species (adapted from Hobday et al. 2004).
Impact
Fishing activity
Direct
Capture (damage or mortality due to gear deployment, including discards)
Cryptic mortality (unaccounted damage or mortality due to interactions
with fishing gear)
Gear loss (damage or mortality without capture due to interactions with
gear lost from the fishing vessel)
Indirect
Species translocation (introduction of species to the habitat of the assessed
species)
On board processing and catch discarding (discard of unwanted parts of
target species or unwanted organisms from the catch)
Provisioning (use of bait or burley)
Pollution (introduction of chemical and physical pollutants from fishing
vessels, such as exhaust, oil spills, detergents, rubbish, lost gear, noise)
152
Table 7.2. Description of the score values for the spatial and temporal scale, intensity and
consequence of the fishing activities (adapted from Hobday et al. 2004).
Score
1
Spatial
Temporal
scale (nm)
scale
<1
Decadal
Intensity
Consequence
Negligible (remote
Negligible (impact
likelihood of detection)
unlikely to be
measured)
2
1–10
Every
Minor (seldom occurring
Minor (minimal
several
and rare to detect)
impact on stock size,
years
3
4
10–100
100–500
Annual
structure or dynamics)
Moderate (moderate at
Moderate (medium
broader scale, or severe but
impact on stock size,
local)
structure or dynamics)
Quarterly Severe (severe and
Major (wider and
occurring often at broad
longer term impact on
scale)
stock size, structure or
dynamics)
5
500–1000
Weekly
Major (occasional but very
Extreme (serious
severe and localized or
impact on stock size,
frequent and widespread
structure or dynamics
but less severe)
with long time period
to restore to
acceptable levels)
6
>1000
Daily
Catastrophic (local to
Intolerable
regional severity or
(widespread and
continual and widespread)
irreversible impact on
stock size, structure or
dynamics)
153
Level 2 assessment
For level 2, the species was assessed based on its biological productivity and catch
susceptibility. Biological productivity can be inferred from the reproductive rate or the
natural mortality rate of a species assuming that, immigration and emigration being equal,
there has to be a balance between reproductive rate and natural mortality rate for a
population to remain in equilibrium (Walker 2004). Species with low reproductive rate and
low natural mortality have low biological productivity and hence are at higher risk from
the effects of fishing than species with high biological productivity. For S. megalops,
natural mortality (M) was used as a proxy for biological productivity. Natural mortality
was estimated by five indirect life-history methods described elsewhere (e.g. Cortés 2002)
and the mean value was used. The indirect methods of Pauly (1980), Hoenig (1983), and
Chen and Watanabe (1989), and two methods by Jensen (1996) use parameters estimated
from the von Bertalanffy growth model (VBGM) and maximum age information which for
S. megalops were obtained from Chapter 6. Pauly’s (1980) method also uses information
on the mean value of water temperature (14.6º C) that was taken from
http://www.marine.csiro.au. Based on empirical data, Walker (2004) devised a scale for M
categorization where values of M ≤0.16, between 0.16 and 0.38, and ≥0.39 correspond to
low, moderate, and high biological productivity, respectively. This scale was used for
biological productivity categorization of S. megalops.
Catch susceptibility, a measure of the extent of the fishing impact of each fishing method,
is the product of availability (proportion of the spatial distribution of the population that is
fished by the fishing method), encounterability (proportion of the available population
encountered by one unit of fishing effort), selectivity (proportion of the encountered
population that is captured by the fishing gear) and post-capture mortality (proportion of
captured animals that die) (Walker 2004). Each of these fishing parameters ranges from 0
to 1; hence, catch susceptibility also ranges from 0 to 1. Fishing parameters with assigned
values of 0.33, 0.66, and 1.00 (upper value for each one-third range) are designated low,
moderate, and high, respectively (Table 7.3) (Walker 2004). Based on expert judgement
and the precautionary approach (i.e. high if unknown), the availability, encounterability,
selectivity, post-capture mortality, and catch susceptibility of S. megalops were determined
for each fishing method.
154
A species identified as having low biological productivity with moderate to high catch
susceptibility would be considered to be at high risk and need to be assessed at level 3. A
species with moderate to high biological productivity and low catch susceptibility would
be considered to be at low risk and require no further assessment.
Level 3 assessment
Level 3 assessment involved a quantitative data-intensive analysis. Application of biomass
dynamic models or more complex models requires time series of catch, fishing effort and
relative abundance data, which are not available for most shark species, particularly bycatch species. Given that demographic analyses require only life-history parameters, which
are commonly available for many species, this approach was used to make the assessment
more compatible and applicable to a broader suite of fisheries. The assessment involved
the quantitative estimation of population growth rate, elasticities, rebound potential and
population doubling time.
Population growth rate (λ = er) and elasticities (effect of a proportional change in a vital
rate on population growth rate) were estimated using a birth-flow Leslie matrix (Caswell
2001). Elasticities of fertility, juvenile survival, and adult survival, are normally obtained
by summation of matrix element elasticities across relevant age classes (e.g. Caswell 2001;
Carlson et al. 2003). However, in the present study, age-at-maturity of S. megalops was not
assumed to be knife-edge, rather, it was determined from an ogive produced by Braccini et
al. (2006) and Chapter 5. Hence, for each age class, the survival elasticity of juveniles and
adults was the product of the elasticity and the proportion of juvenile and adult females
within each class. The total juvenile and adult elasticity was then obtained by summing
across the different age classes. Population rebound potential (rz) and doubling time (TD =
ln (2) rz–1) were estimated by the method of Smith et al. (1998), which incorporates density
dependent compensation of adult M through preadult survival. Rebound potential is
calculated at the population level of maximum sustainable yield (Smith et al. 1998).
To account for uncertainty in life-history parameter values, a probability density function
(pdf) was developed for each life-history parameter following the approach of Cortés
(2002) and Carlson et al. (2003). The pdfs were then used in a Monte Carlo simulation
(with 10,000 iterations) to incorporate stochasticity in the estimation of population
parameters. Each iteration involved the random selection of a set of life-history parameter
155
Table 7.3. Description of the score values of fishing parameters for three arbitrary risk
categories (adapted from Walker 2004). Catch susceptibility is the product of availability,
encounterability, selectivity, and post-capture mortality.
Parameter
Risk category
Low
Availability
Moderate
High
0.33 (fishery ranges
0.66 (fishery ranges
1.00 (fishery ranges
<one-third of species
one-third–two-thirds
>two-thirds of
range)
of species range)
species range)
0.66 (0.33–0.66
1.00 (0.67–1.00
Encounterability 0.33 (0–0.33
probability of species probability of species
probability of
encountering the
encountering the gear,
species encountering
gear, e.g. pelagic
e.g. pelagic species
the gear, e.g.
species encountering
encountering mid
bottom-dwelling
bottom trawl net)
water trawl net)
species encountering
bottom trawl net)
Selectivity
0.33 (0–0.33
0.66 (0.33–0.66
1.00 (0.67–1.00
probability of species probability of species
probability of
being caught by the
being caught by the
species being caught
gear, e.g. filter-
gear, e.g. fast-
by the gear, e.g.
feeder species taking
swimming species
species with
a baited hook)
taken by bottom trawl
protruding structures
net)
taken by gillnet)
Post-capture
0.33 (0.67–1.00
0.66 (0.33–0.66
1.00 (0–0.33
mortality
probability of
probability of survival
probability of
survival after
after capture, e.g.
survival after
capture, e.g.
discarded species with capture, e.g. retained
discarded bottom-
a fragile structure and
dwelling species with ram-jet ventilation)
target and byproduct species)
spiracles and robust
structure)
Catch
susceptibility
156
0–0.33
0.33–0.66
0.67–1.00
values and the calculation of λ, elasticities, rz, and TD. In this way, prediction intervals
(2.5th and 97.5th percentiles) were obtained from the probability density distribution for
each of the estimated population parameters. Simulations were run using Microsoft Excel
spreadsheets equipped with the add-in PopTools (http://www.cse.csiro.au/poptools/) and a
risk assessment software (Crystal Ball; Decisioneering Inc.) lent by E. Cortés (NOAA,
Southeast Fisheries Science Centre, Panama City, Florida, USA).
Life-history parameters needed for the estimation of population parameters were obtained
from the literature. Average litter size, the relationship between fecundity and total length
(TL) of female, embryo sex ratio, length-at-maturity, and length-at-maternity (relationship
between the proportion of females in maternal condition, i.e. contributing to annual
recruitment, and TL) were obtained from Braccini et al. (2006) and Chapter 5. Average
litter size was assumed to follow a normal pdf with a mean (and s.d.) of 2.32 (0.48) and a
lower and upper bounds of 2 and 4 reflecting the range of litter sizes reported. The
fecundity–TL relationship was predicted from the linear equation: litter size = 0.0071
(0.001) TL – 1.503 (0.549) and was represented by a normal pdf. The embryo sex ratio
reported for S. megalops was 1:1 so a 0.5 factor was used to half the litter size–TL function
and obtain the number of female embryos per female. Length-at-maternity was predicted
from the logistic equation, proportion maternal = 0.5 (1 + e –ln (19) ((TL –531 (18)) / (626 (47) – 531
(18))) –1
) and was represented by a normal pdf. Growth parameters and maximum age were
obtained from Chapter 6. The growth equation for females was used to transform the
relationships between the reproductive variables and TL to relationships between
reproductive variables and age. Natality-at-age was calculated as the product of 0.5
(embryo sex ratio), the age-at-fecundity, and the age-at-maternity functions. Two extreme
case scenarios were considered. The first is deemed as the worst-case scenario, whereas the
second scenario is the more optimistic.
In the first scenario, growth parameters (and s.e.) L∞ = 756 (45) mm, k = 0.042 (0.005)
years–1 and t0 = –9.77 (1.30) years produced by a two-phase VBGM were used as the most
likely values in a normal pdf. Age at 50% maturity was represented by a triangular pdf
with 15 as the likeliest value and ±2 years as the lower and upper bounds. These values
were derived from a length-at-maturity curve (Braccini et al. 2006; Chapter 5) and the
growth curve produced by the two-phase VBGM. Maximum age was represented by a
157
linearly decreasing pdf scaled to 1, with the likeliest value of 28 (oldest shark aged;
Chapter 6) and the lower bound set by arbitrarily adding 30% to the likeliest value, i.e. 36
(Cortés 2002). Annual survivorship-at-age for the Leslie matrix was assumed to be
uniform, ranging from 0.862 to 0.936, estimated by the Hoenig (1983) and Jensen (1996)
methods, respectively. Adult M for the Smith et al. (1998) method was also assumed to
have a uniform pdf ranging from 0.066 to 0.149. Maximum sustainable yield for the Smith
et al. (1998) method was assumed to occur at total mortality (Z) = 1. 5 M.
In the second scenario, growth parameters L∞ = 699 (17) mm and k = 0.056 (0.005) years–1
produced by a two-parameter VBGM growth model were used as the most likely values in
a normal pdf. Age at 50% maturity was represented by a triangular pdf with 14 as the
likeliest value and ±2 years as the lower and upper bounds. These values were derived
from the length-at-maturity curve and the growth curve produced by the two-parameter
VBGM growth. Maximum age was represented as in the first scenario, with the difference
that the maximum bound was set to 50% of the likeliest value, i.e. 42. Annual
survivorship-at-age for the Leslie matrix was assumed to be uniform, ranging from 0.930
to 0.936, estimated by the Chen and Watanabe (1989) and Jensen (1996) methods,
respectively. Adult M for the Smith et al. (1998) method was also assumed to have a
uniform pdf ranging from 0.066 to 0.072. Maximum sustainable yield for the Smith et al.
(1998) method was assumed to occur at Z = 2 M.
A Spearman rank correlation (rs) was used to measure possible correlation between lifehistory parameters and λ. For each scenario, the correlation between the simulated M, k,
the slope of the fecundity curve, and TL at 50% maturity of the length-at-maternity
relationship with the forecasted λ was determined.
7.4 RESULTS
Level 1 assessment
Level 1 assessment of the fishing methods indicated that the only fishing-related activity to
have moderate or higher consequences on the sustainability of S. megalops was that
associated with ‘capture fishing’ (Appendix 7.a). Other activities had either negligible or
minor consequences. Among methods, ‘capture fishing’ of shark longlines, droplines, and
traps and pots had a negligible effect on S. megalops due to their recent decline to ~0 effort
158
in 2004 (Walker and Gason 2006). Hence these methods were not assessed at a second
level. Conversely, ‘capture fishing’ for the otter trawl, Danish seine, shark gillnet, and
automatic longline methods had a consequence score >2 so these fishing methods need to
be assessed at level 2.
Level 2 assessment
Estimates of natural mortality (M) had a mean value of 0.085, ranging from 0.066, for one
of Jensen’s (1996) methods, to 0.149 (Hoenig 1983), indicating that S. megalops had low
biological productivity (i.e. M ≤0.16). Catch susceptibility of S. megalops varied
depending on the fishing method (Appendix 7.b). For the shark gillnet method, given the
low selectivity, and moderate availability and post-capture mortality, catch susceptibility
was low (catch susceptibility = 0.66 × 1.00 × 0.33 × 0.66 = 0.14). For the automatic
longline method, catch susceptibility was also low, as availability was low (catch
susceptibility = 0.33 × 1.00 × 1.00 × 0.66 = 0.22). For the otter trawl and Danish seine
methods, there is uncertainty regarding the extent of their spatial overlap with S. megalops
distribution. If availability for these fishing methods was designated low, catch
susceptibility was also low (catch susceptibility = 0.33 × 1.00 × 1.00 × 1.00 = 0.33).
However, if availability was considered moderate, catch susceptibility was also moderate
due to high encounterability, selectivity and post-capture mortality (catch susceptibility =
0.66 × 1.00 × 1.00 × 1.00 = 0.66). In addition, cumulative effects across the four fishing
methods may cause an underestimation of total catch susceptibility; hence, based on S.
megalops low biological productivity, this species was classed as at high risk and needed
to be assessed at level 3.
Level 3 assessment
For the first scenario (worst-case), population growth rate (λ) averaged 0.975 (median
0.974, 95% CI 0.974–0.976) (Fig. 7.1). Fertility elasticities averaged 0.046 (0.045, 0.046–
0.046), elasticities of juvenile survival were 0.630 (0.629, 0.628–0.631) and those of adults
were 0.325 (0.325, 0.323–0.326). Rebound potential (rz) averaged 0.031 years–1 (0.029,
0.030–0.032) and population doubling time (TD) was 28.2 years (23.5, 27.9–28.5) (Fig.
7.1). There was a negative correlation between M and λ (rs = – 0.685, P <0.001) and a
positive correlation between the growth coefficient (k) and λ (rs = 0.474, P <0.001) but
159
TL50 (TL at 50% maturity) and the fecundity slope showed no correlation with λ (rs = –
0.139, P >0.001 and rs = 0.284, P >0.001, respectively).
For the second scenario (more optimistic), λ averaged 1.018 (median 1.019, 95% CI
1.017–1.019) (Fig. 7.1). Fertility elasticities averaged 0.045 (0.045, 0.045–0.045),
elasticities of juvenile survival were 0.397 (0.395, 0.396–0.398) and those of adults were
0.559 (0.560, 0.558–0.560). Rebound potential averaged 0.035 years–1 (0.032, 0.035–0.035)
and TD was 23.7 years (21.4, 23.5–23.9) (Fig. 7.1). There was a positive correlation
between the fecundity slope and λ (rs = 0.638, P <0.001) and between k and λ (rs = 0.342,
P <0.001) but TL50 and M showed no correlation with λ (rs = – 0.314, P >0.001 and rs = –
0.111, P >0.001, respectively).
160
0.12
0.12
Scenario 1
0.10
Scenario 2
0.10
0.08
0.08
0.06
0.06
0.04
0.04
0.02
0.02
0
0
6 8 0 2 4 6 8 0 2 4 6 8 0
0.8 0.8 0.9 0.9 0.9 0.9 0.9 1.0 1.0 1.0 1.0 1.0 1.1
6 8 0 2 4 6 8 0 2 4 6 8 0
0.8 0.8 0.9 0.9 0.9 0.9 0.9 1.0 1.0 1.0 1.0 1.0 1.1
Probability density
Population growth rate
Population growth rate
0.12
0.12
0.10
0.10
0.08
0.08
0.06
0.06
0.04
0.04
0.02
0.02
0
0
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10
Rebound potential
Rebound potential
0.12
0.12
0.10
0.10
0.08
0.08
0.06
0.06
0.04
0.04
0.02
0.02
0
0
0
10 20 30 40 50 60 70 80 90 100
Doubling time (years)
0
10 20 30 40 50 60 70 80 90 100
Doubling time (years)
Figure 7.1. Probability density distribution of population growth rate, rebound potential
and population doubling time for scenarios 1 and 2 (n = 10,000 simulations).
161
7.5 DISCUSSION
Protection and management of marine resources should be based on the integration of
qualitative and quantitative methods (Cortés 2004) simply because management based
solely on quantitative information is of limited application to data-poor fisheries (Johannes
1998). Qualitative, semi-quantitative, and quantitative data together in a hierarchical
assessment framework showed that S. megalops is potentially highly susceptible to the
effects of fishing. At a qualitative level (level 1), the hierarchical assessment allowed
screening out of fishing methods and fishing-related activities considered to pose no risk to
this species. This indicates that research effort should be allocated on those methods and
fishing activities (‘capture fishing’) leading to moderate or higher impacts on S. megalops.
Qualitative expert judgement, for example knowledge possessed by artisanal fishers, has
been valuable for successful management of tropical fisheries in developing countries
(Johannes 1998). Furthermore, expert knowledge is commonly used in Bayesian (e.g.
Martin et al. 2005) and fuzzy logic (Cheung et al. 2005) modelling for conservation
assessments. Hence, quantitative information need not be an exclusive condition for sound
management; a qualitative assessment of chondrichthyans would allow ruling out of
species ranked no risk allowing focus on those in risk. The value of this approach is
twofold. First, it is usually not practical to undertake long-term studies on non-target
chondrichthyan species due to an urgent need for their effective management. Second, the
Australian ecological risk assessment process is valuable at making research more costeffective and at prioritising research funding involving high costs associated with
collecting quantitative fishing data (Dulvy et al. 2003); the process also provides the option
for a management response at any level. In particular, this approach would be useful for
developing countries where two-thirds of reported landings of chondrichthyans occurred
(Bonfil 1994), resources for monitoring fishery impacts are limited (Johannes 1998) and
plans of management have not been implemented.
A semi-quantitative assessment (level 2) ranked S. megalops at high risk given its low
biological productivity and the cumulative catch susceptibility to the fishing methods.
Shark gillnets and Danish seines are used on the continental shelf whereas otter trawlers
operate on the upper slope throughout the SESSF and on the shelf off New South Wales,
far eastern Victoria, and eastern Tasmania. When considered separately, each of the three
fishing methods has low availability, but the three methods together have medium
availability and therefore may increase the catch susceptibility of S. megalops. In addition,
162
the availability to the otter trawl method would be high if the population is predominantly
distributed on the upper continental slope. Demersal trawling may create a higher food
supply by disturbing sediments and exposing prey, attracting S. megalops to the fishing
grounds as this species of shark is an opportunistic predator that consumes a wide range of
prey items (Braccini et al. 2005; Chapter 4). When applied to a large number of species,
the advantage of this approach is to allow low-risk species to be excluded from the dataintensive quantitative analysis of level 3, such that research and management efforts can be
directed where most needed (Stobutzki et al. 2002; Hobday et al. 2004). Even for datapoor fisheries, where species-specific information on biological productivity or catch
susceptibility is not available, this level of analysis can be applied. Information can be used
from studies from other areas or on closely related species (Walker 2004) with the caution
that there may be some degree of geographical variation in life-history parameters (e.g.
Parsons 1993) and catch susceptibility parameters.
A quantitative assessment (level 3) showed that population growth of S. megalops is slow
even under the assumption of density-dependent compensation after a fishing exploitation
rate equal to the natural mortality rate. For both the best- and worst-case scenarios, the
stochastic estimations of rebound potential (rz) and population doubling time (TD) are low
even within the range reported for shark species (0.017–0.136 years–1 and 5.1–41.5 years,
respectively) and are similar to the values reported for the spiny dogfish (S. acanthias)
from the north-western Atlantic (Smith et al. 1998). Stochastic population growth rate (λ)
was also slow, placing S. megalops towards the “slow” end of the spectrum along a
continuum of life-history traits of sharks (Cortés 2002). For the worst-case scenario, most
simulated λ values fell below 1 suggesting population decrease even under no fishing
mortality. Squalus acanthias and the shortspine spurdog (S. mitsukurii) also show λ values
below 1 and would not possess the biological attributes to restore λ to its original level
after moderate exploitation (Cortés 2002). For the worst-case scenario proposed for S.
megalops, λ values below 1 would result from the negative correlation between λ and
natural mortality (M) and the positive correlation between λ and growth coefficient (k).
Within this scenario, some of the permutations of the combined M and k that can be
obtained from their respective distributions (in this case high M and low k values) resulted
in λ values below 1 (McAllister et al. 2001). Elasticities of juvenile and adult survival
were higher than fertility elasticities indicating that λ is more sensitive to the survival of
163
juveniles and adults. As for S. acanthias and S. mitsukurii (Cortés 2002), for the worst-case
scenario, juvenile S. megalops had a higher elasticity than adults and management actions
should focus on the protection of juveniles. For the best-case scenario, due to the simulated
longer lifespan (42 years) and the larger number of adult age classes, elasticity of adults
was higher for what management actions should focus on this latter group. In all, the
estimated values of rz, TD and λ indicate that recovery time of S. megalops population after
fishing overexploitation is expected to be very long. Although the relative abundance of S.
megalops has remained stable off New South Wales (Graham et al. 2001), probably due to
refuge areas where ground is untrawlable (Graham 2005), this is currently the most caught
by-catch shark species taken by demersal trawlers in south-eastern Australia. Trends in
catch per unit effort from onboard scientific observer data suggest the population has been
stable during 1992–05 (Walker and Gason 2006). Nevertheless, the aggregating behaviour
of S. megalops (Graham 2005; Chapter 2), in combination with its slow growth rate
(Chapter 6) and low reproductive output (Braccini et al. 2006; Chapter 5), makes this
species potentially vulnerable to the effects of fishing. For species with similar life-history
traits, such as S. acanthias, stock depletions have been well documented (e.g. Holden
1977). Hence, any increase in catch susceptibility through targeting, increased retention of
catch, or change in gear design such as reducing mesh-size of shark gillnets could quickly
increase the risk of stock depletion for S. megalops. Conservation and management for
sustainable use of S. megalops will require a close control of fishing mortality due to the
low biological productivity of this species.
In summary, the hierarchical ecological risk assessment approach adopted in Australia
allows for the effective evaluation of the effects of fishing on non-target chondrichthyan
species and the identification of species at risk. Under this approach, research and
management effort can be prioritised and directed to where it is most needed. The
hierarchical approach allows for management response at any level as an alternative to
undertaking the research required to proceed to the next level of assessment. Hence, this
approach is particularly useful for fishery management organizations for the assessment of
data-limited fisheries. For example, the Australian Fisheries Management Authority
(AFMA) is presently applying such a framework for the evaluation of 14 fisheries as the
basis for determining priorities for research, fishery monitoring, and management (Hobday
et al. 2004). There is no doubt about the need for management of chondrichthyans,
particularly for those taken as by-catch in multispecies fisheries. The question is how to
164
make research and management more cost-effective and priority-driven when
chondrichthyan resources are being depleted and there are insufficient time or funding
available for comprehensive data collection on all species.
165
166
166
Gear loss
1
1
Cryptic
mortality
6
1
Capture
6
6
scale
activity
6
6
6
scale
Presence Spatial Temporal
Fishing
a.1 Otter trawl and Danish seine methods
1
2
4
Intensity
1
2
3
Consequence
1
2
1
Certainty
retrieve them.
which creates a high incentive not to loose and to
Trawl net loss is a rare event due to their high price,
this impact.
stock. Certainty low due to difficulties in measuring
but it is considered to have a minor impact on the
injured from encounters with trawl nets may happen
Cryptic mortality caused by escapement of animals
long-term relative abundance information.
2001); however, for the other regions there is no
abundance in one region of the fishery (Graham et al.
there has been no long-term decline in relative
range of the fishery. Moderate impact on stock as
S. megalops is captured by trawl nets throughout the
Rationale
Appendix 7.a.
Level 1 assessment of the fishing methods having a potential impact on S. megalops. Presence score: 1, present; 2, absent. Certainty score: 1
certain (data exist and considered sound); 2 uncertain (no data or considered poor or conflicting; see Table 7.2 for score value description). a.1
Otter trawl and Danish seine methods, a.2 shark gillnet method, and a.3 automatic longline method.
167
167
2
Moderate to large quantities of organisms are
0
1
Provisioning
Pollution
6
6
2
1
1
1
considered negligible.
ability >scale of the hazard. Impact on stock is
megalops is a mobile species with an avoidance
risks as this hazard only affects a small area and S.
Chemical or physical pollutants might have minor
Provisioning does not occur in this method.
extent of this impact is uncertain due to lack of data.
probability of being taken by the gear. However, the
(Braccini et al. 2005; Chapter 4), increasing the
3
on the stock through habitat modification.
shots but it is unlikely to have a measurable impact
nets can be translocated up to several miles between
Invertebrate and vertebrate species caught in trawl
discarding
3
1
Rationale
attract opportunistic species, such as S. megalops
6
1
Certainty
and catch
6
1
Consequence
discarded in trawling operations. This practice can
1
6
Intensity
processing
On board
translocation
6
Species
1
scale
activity
scale
Presence Spatial Temporal
Fishing
Appendix 7.a.1 Continued…
168
168
1
Cryptic
0
1
Provisioning
Pollution
discarding
and catch
processing
On board
1
1
Species
translocation
1
Gear loss
mortality
6
1
Capture
6
6
6
6
6
scale
activity
6
6
6
6
6
6
scale
Presence Spatial Temporal
Fishing
a.2 Shark gillnet method
2
3
1
1
2
3
Intensity
1
3
1
1
2
4
Consequence
1
1
2
1
1
2
1
Certainty
Refer to Table a.1.
Refer to Table a.1.
Refer to Table a.1.
Refer to Table a.1.
Refer to Table a.1.
Refer to Table a.1.
females could have a major impact on the stock.
Selective removal of significant numbers of large
the range of the fishery (Walker et al. 2005).
only large female S. megalops are taken throughout
target larger-sized species (e.g. M. antarcticus) so
Shark gillnets are 6–6½-inch mesh-size designed to
Rationale
169
169
1
Cryptic
discarding
and catch
processing
On board
1
6
6
1
Species
translocation
6
1
Gear loss
mortality
6
1
Capture
6
scale
activity
6
6
6
6
6
scale
Presence Spatial Temporal
Fishing
a.3 Automatic longline method
2
1
1
2
3
Intensity
2
1
1
2
3
Consequence
1
1
1
2
2
Certainty
effect on the stock.
Discards in this fishery are minimal, having a minor
Refer to Table a.1.
Refer to Table a.1.
Refer to Table a.1.
considered moderate.
abundant; hence, the impact on the stock is
(continental slope) where S. megalops is not very
and temporal scale but only in deepwater
This fishing method is used across a broad spatial
Rationale
170
170
1
1
Provisioning
Pollution
6
6
scale
activity
6
6
scale
Presence Spatial Temporal
Fishing
Appendix 7.a.3 Continued…
2
2
Intensity
1
2
Consequence
1
1
Certainty
Refer to Table a.1.
S. megalops stock.
the stock of squid and hence minor indirect effect on
harvested at low levels with minor consequence for
The most commonly used bait is squid, which is
Rationale
171
171
Av
L–Mo This method is confined to the continental shelf and upper slope overlapping in part with S.
Cs
longline
Automatic
gillnet
Shark
L
Mo
H
H
H
L
Mo
Mo
L
L
Selectivity is high given hook size does not affect catch (Walker et al. 2005).
S. megalops is likely to take the baited hook given its opportunistic feeding behaviour.
spatial distribution. Encounterability is high due to depth distribution of gear and species as
This method is confined to the continental slope with very low overlap with S. megalops
Table 6, Walker et al. 2005).
distribution of gear and species. Selectivity is low due to 6–6½-inch mesh-size of gillnet (see
moderate overlap with S. megalops spatial distribution. Encounterability is high due to depth
This method is confined to the continental shelf, mostly inside the 120–150 isobath, with
mortality is also high due to trauma caused inside the net.
species. Selectivity is high given uncertainty in selectivity of trawl nets. Post-capture
H
Pc
seine
H
Se
Rationale
megalops spatial distribution. Encounterability is high due to depth distribution of gear and
H
Ec
Parameter
& Danish
Otter trawl L–Mo
method
Fishing
Appendix 7.b.
Catch susceptibility for otter trawl and Danish seine method, shark gillnet method, and automatic longline method. Risk categories of
availability (Av), encounterability (Ec), selectivity (Se), post-capture mortality (Pc), and catch susceptibility (Cs) are L (low, 0.33), Mo
(moderate, 0.66), and H (high, 1.00).
Sunset aboard the fishing vessel ‘Ester-J’ after a day of sampling (photo by the author).
172
CHAPTER 8
173
CHAPTER 8
GENERAL DISCUSSION
Due to increasing risk of depletion of shark species worldwide, the lack of long-term timeseries of data, and urgent need for management actions, a different approach from that
usually adopted for target species is required for the assessment of non-target species. This
thesis applies the concept of hierarchical assessment for evaluating ecological risks to
Squalus megalops by the main fisheries operating in south-eastern Australia. The thesis
provides quantitative information that enables several key features of the life-history and
population status of this species to be determined. This was achieved by studying S.
megalops using different approaches: (i) analysing its population structure and
morphological relationships (Chapters 2 and 3), (ii) quantifying its feeding ecology and the
factors that contribute to variation in its diet (Chapter 4), (iii) determining the reproductive,
and age and growth parameters needed for its population assessment (Chapters 5 and 6),
and (iv) using the life-history information in a modelling framework for the assessment of
ecological risks (Chapter 7).
The present chapter provides a summary of the key findings presented in this thesis. For
expanded discussion of the points made here, and for further details that have been
omitted, the reader should consult the relevant text in Chapters 2 through to 7.
8.1 POPULATION STRUCTURE AND MORPHOLOGICAL RELATIONSHIPS
Most dogfish species have complex population structures, exhibiting segregation by sex,
size and mature condition (e.g. Hanchet 1988; Yano and Tanaka 1988; Yano 1995;
Wetherbee 1996; Clarke 2000). It is, therefore, prudent to consider this pattern in the
management of these species. Despite the opportunistic nature of the sampling design,
individual analysis of selected fishing shots suggested that S. megalops also has a complex
population structure, segregating by sex, size and breeding condition. In addition, given
that females attain a larger maximum size than males, implementation of maximum legal
sizes could result in a population dominated by females, affecting the biological
productivity of the population. Finally, the sex ratio suggested that females may be more
vulnerable to fishing than males probably due to an overlap between the distribution of
174
females and of the various fisheries operating in south-eastern Australia. Hence, as for
mammals with sexual segregation (Bowyer 2004), the strategies developed for the
conservation of S. megalops should incorporate the segregation pattern in management
plans, considering males and females as separate entities.
Commercial shark species are normally beheaded, eviscerated and landed with or without
fins attached; hence, only partial lengths and partial masses can be recorded after landing
(FAO 2000b). Furthermore, in many fisheries only the fins are retained, whereas the rest of
the animal is discarded. Therefore, relationships between partial lengths and total length
and between partial masses and total mass of sharks are needed to determine the length and
mass composition of captured sharks. Although many studies provide total mass–total
length relationships, few present conversion factors to allow calculating total mass or total
length from partial mass or partial length measurements. However, these relationships and
conversion factors are essential for fisheries monitoring programmes and population
assessments. Given the depletion of many of the harvested species of sharks and a decline
in abundance of most other shark species in southern Australia (Graham et al. 2001), S.
megalops will inevitably become a more sought after species. Hence, the conversion
factors from partial lengths to total length and from partial masses to total mass were
determined. When measurements reflect only somatic growth, conversion factors to total
length or total mass can be determined pooling sexes and sizes, but, for variables reflecting
somatic and reproductive growth (total and liver masses), different outcomes can be
expected when different ranges of size are compared. Due to the sexual size dimorphism of
S. megalops, conversion factors should be determined for sexes and sizes separately.
8.2 DIETARY COMPOSITION AND SOURCES OF VARIATION
Several shark species play an important role in the structure of marine communities
(Cortés and Gruber 1990; Ellis et al. 1996; Cortés 1999; Stevens et al. 2000) and the
exchange of energy between upper trophic levels (Wetherbee et al. 1990); however, there
are few quantitative studies on the feeding ecology of sharks and no studies account for the
natural variability in the diet of marine predators. This thesis provides evidence of different
sources of variation in the feeding ecology of S. megalops, an opportunistic predator with a
diverse diet that consumes a wide range of prey items.
175
The use of multiple descriptors of dietary composition allowed a better representation of
the overall diet of this opportunistic predator. When prey importance was inferred from the
weight or frequency of occurrence of prey items, S. megalops preyed mainly on molluscs
and teleosts. However, when number of prey was used, the most important prey items were
crustaceans. Therefore, given that number, weight, and frequency of occurrence measures
provide different information ( MacDonald and Green 1983; Bigg and Perez 1985; Cortés
1998b), multiple measures should be used for the description of feeding habits, particularly
when prey items differ in size (Ferry and Cailliet 1996).
There was a wide range of variability in the overall diet of S. megalops. The bootstrap
analysis showed that importance of prey varied by up to 50% within the upper and lower
95% confidence intervals. Variability in the overall diet of shark species has not been
previously reported. However, due to the combination of small sample sizes and
opportunistic sampling by many diet studies, high proportion of empty stomachs, and the
opportunistic predatory nature of many shark species, high variability in diet and hence in
predator–prey interactions is expected. Predator–prey interactions are used as linkages
between species in ecosystem modelling. However, model outputs may not be accurate
when overall diet data do not account for the variability in the composition of diet.
The dietary composition of S. megalops varied in space and time, and exhibited differences
among regions, seasons and size classes. When spatial and temporal effects in the diet of
sharks have been tested (Simpfendorfer et al. 2001b; Vögler et al. 2003; White et al.
2004), region, season or ontogeny have been evaluated independently of each other and the
interaction of these factors was not taken into account even though samples were collected
across wide spatial and temporal scales. Apart from Cortés et al. (1996), no other study on
the diet of shark has tested for the interaction of factors. This thesis showed an interaction
between size and season where large and small S. megalops had different diets in summer
and autumn, but consumed similar prey items in spring. Hence, large and small individuals
would have, at least during part of the year, different ecological roles within the marine
ecosystem. Therefore, the intrinsic natural variability in the dietary composition of S.
megalops and its spatial and temporal variation in diet suggests that information on the
ecological relationships among species is likely to be missed when predator–prey
interactions are only inferred from overall diet.
176
8.3 REPRODUCTIVE AND GROWTH PARAMETERS
Reproductive and age and growth parameters were estimated for the quantitative
population assessment of S. megalops. The reproductive and growth traits of S. megalops
are typical of species with low resilience, making this shark species highly susceptible to
the effects of fishing overexploitation.
Male and female S. megalops are capable of reproducing throughout the year, exhibiting
no pattern of temporal periodicity in their reproductive cycle. Continuous asynchronous
cycles have been reported for other dogfish species (Yano and Tanaka 1988; Yano 1995);
however, gestation period and ovarian cycle of those species have not been determined.
This thesis provides a method for the estimation of these parameters in asynchronous
breeders, such as S. megalops. For this species, ovarian cycle and gestation period are ~2
years. Periodicity of ovarian cycle and gestation period are crucial for defining maternal
condition of female chondrichthyans. Although most viviparous sharks have gestation
periods of approximately a year (Stevens and McLoughlin 1991; Hamlett and Koob 1999),
species producing large-sized follicles, such as most squalid species (Chen et al. 1981;
Hanchet 1988; Guallart and Vicent 2001), have ovarian cycles and gestation periods of
two, three or more years. Hence, for shark species with asynchronous breeding and
reproductive cycles of several years duration, if maturity ogives (i.e. proportion of females
in mature condition at any size) are used in population assessments instead of maternity
ogives (i.e. proportion of females contributing to annual recruitment), models will overestimate recruitment rates, leading to overly optimistic predictions.
Age and growth information needed for the population assessment of S. megalops was
derived from bands laid on the first dorsal fin spine. Precision estimates, the relationship
between spine total length and body length, edge analysis, and agreement between counts
on the inner dentine layer and the enameled surface support the use of this structure for the
age estimation of S. megalops. Most studies on age and growth of dogfish also use dorsal
fin spines for age estimation (e.g. Holden and Meadows 1962; Ketchen 1975; Clarke et al.
2002a). Different growth functions were fitted to length-at-age data to determine the best
growth model. From the examined models, a two-phase von Bertalanffy function provided
the best goodness-of-fit. However, the results should be interpreted with caution and might
not be representative of real growth because of the effects of sampling bias, lengthselective fishing mortality and bias in age estimation on length-at-age data. Data quality is
177
particularly relevant when samples are obtained from commercial fisheries, as in most
studies on age and growth of sharks, because growth parameters may be different from
those obtained from a more representative sample (Haddon 2001). Regardless of the model
used, the growth rate of S. megalops (particularly of females) is very low even within the
range of growth rates reported for shark species; hence, it is expected that population
recovery will be slow following fishing overexploitation.
8.4 ECOLOGICAL RISK ASSESSMENT
A hierarchical framework was applied to S. megalops for the assessment of ecological risks
from effects of fishing. By integrating qualitative and quantitative biological (reproduction,
age and growth) and fishing impact data, this approach showed that S. megalops is
potentially highly susceptible to the effects of fishing. A qualitative assessment indicated
that the only fishing-related activity to have moderate or high impact on this species was
that associated with ‘capture fishing’ of the otter trawl, Danish seine, gillnet and automatic
longline methods. Hence, other fishing methods or fishing-related activities were screened
out from further analyses. At the next assessment level, a semi-quantitative evaluation
ranked S. megalops as at risk given its low biological productivity and, possibly, its catch
susceptibility from cumulative effects across the separate fishing methods. The qualitative
and semi-quantitative assessments make use of qualitative expert judgement for efficient
risk identification (Hobday et al. 2004). Hence, quantitative data need not be an exclusive
condition for sound management of sharks as qualitative and semi-quantitative assessments
allow ruling out of species at low risk and focus on those at medium or high risk. This
approach ensures that research is more cost-effective and priority-driven. In particular,
such a method is useful for developing countries where two-thirds of reported landings of
sharks occurs (Bonfil 1994), resources for monitoring fishery impacts are limited
(Johannes 1998), and where plans of management have not been developed.
A quantitative assessment showed that population growth of S. megalops is slow even
under the assumption of density-dependent compensation where the fishing mortality rate
equals the natural mortality rate. The rebound potential and population doubling time are
low and similar to the values reported for S. acanthias from the north-western Atlantic
(Smith et al. 1998). For the worst-case scenario, most of the simulated population growth
rates suggested population decrease even under no fishing mortality. In this case, S.
megalops would not have the biological potential to restore population growth to its
178
original level after exploitation (Cortés 2002). Squalus megalops is currently the most
caught by-catch shark species taken by demersal trawlers in south-eastern Australia
(Walker and Gason 2006). Therefore, its conservation and management for sustainable use
will require a close monitoring and control of fishing mortality due to the low biological
productivity of this species.
Application of the hierarchical framework presented in this thesis allows research and
management efforts to be optimised by identifying and excluding low-risk species and
low-risk fisheries from data intensive assessments. Under this scheme, research and
management efforts can be prioritised and directed to where most needed. By making
research more cost-effective and priority-driven, this approach is particularly useful for
fishery management organizations for the assessment of data-limited fisheries.
8.5 MANAGEMENT
History has shown that marine resources are far from being unlimited and, in fact, are
usually overexploited. To quote Ludwig et al. (1993), “Although there is considerable
variation in detail, there is remarkable consistency in the history of resource exploitation:
resources are inevitably overexploited, often to the point of collapse or extinction” (p. 17).
Unfortunately, management of marine resources has commonly been implemented after
overexploitation of resources.
Management of sharks has been no exception. Concerns over the worldwide decline in
most harvested shark stocks led to the development of the International Plan of Action for
the Conservation and Management of Sharks. The Plan of Action requires each signatory
country to develop a national plan for the management of all chondrichthyan species taken
in the local fisheries. Currently, only Australia, USA, New Zealand, South Africa, and
Canada have developed management plans for some shark species (IUCN 2002).
Management in most other countries has been hindered by the lack of biological and
fishing information on most species, especially for fisheries in developing countries. It is
within this context that hierarchical assessments of ecological risks are of significant
relevance. As summarized in the words of Johannes (1998), “…the key management
question should not be what data do we need to make sound management decisions but
rather, what are the best management decisions to make when such data are unobtainable”
(p. 245).
179
The apparent stable population structure and relative abundance of S. megalops off New
South Wales after 20 years of commercial trawling suggests that this species inhabits
untrawlable grounds that provide large refuge areas (Graham 2005). Also, commercial
trawlers targeting more valuable species tend to avoid aggregation areas of undesirable
species, such as S. megalops. Automatic longliners only operate on the continental slope,
with very low spatial overlap with S. megalops distribution. Finally, only the largest
animals are retained by gillnets of 6- and 6½-inch mesh-size used in the commercial shark
gillnet method (pers. obs.) and most of the population is protected from this fishing method
by the legal minimum mesh-size of 6 inches (since 1975) (Walker et al. 2005). Hence,
current fishing mortality from these fishing methods is expected to be low and this would
explain the high relative abundance of this shark. However, at the same time, S. megalops
is the most commonly captured by-catch shark species by demersal trawlers in southeastern Australia (Walker and Gason 2006) and has started to be landed by the GHATF at
an increasing rate (from 5 T in 2002 to 16 T in 2004) (Walker and Gason 2005). In
addition, due to long-term declines in abundance of shark and chimaera populations in
southern Australia (Graham et al. 2001), quota reductions on target and by-product shark
and chimaera species (Walker and Gason 2005), and growing consumer demand for shark
meat, S. megalops will inevitably become a more sought after species. As previously
shown, the life-history traits, population growth rate, and rebound potential of S. megalops
only allow for low harvesting rate and any increase in fishing mortality should be closely
monitored for this species to avoid depletion.
8.6 FUTURE RESEARCH
Several areas need further research.
•
There is solid evidence for New South Wales (Graham 2005) and for south-eastern
Australia, from selected fishing shots, of sexual and size segregation. Females in
the first year of pregnancy seem to be segregated from females in the second year
of pregnancy. Small individuals may have a pelagic phase and would not be
collected by bottom trawl and gillnets. Given that S. megalops is the most
commonly taken by-catch shark species by demersal trawlers in south-eastern
Australia (Walker and Gason 2006), further information is needed on the location
of parturition areas, and the spatial distribution of juveniles, males, and females in
various breeding conditions. Therefore, a more rigorous sampling design would
180
allow determining the extent of the segregation pattern of S. megalops in southeastern Australia and testing the hypothesis proposed for the depth distribution of
small sharks. Finally, the areas of large aggregations of individuals, particularly of
large females, need to be determined and considered in management plans.
•
In this thesis, I tested for different sources of variation in the feeding ecology of S.
megalops (Chapter 4). However, regional effects could only be analysed for large
females collected in autumn, and samples from only three seasons could be
compared for the seasonal analysis. Future research will benefit from further
collection of samples from the sexes, sizes and seasons missing. The use of captive
sharks would allow estimates of feeding frequency and periodicity. Important for
fisheries management would be to incorporate variability in dietary composition
into ecosystem models to understand the implications of diet variability on model
predictions.
•
Age and growth parameters are vital for population assessment. I estimated the age
and growth rate of S. megalops, but, perhaps in part due to sampling bias, length-atage data might not be representative of natural growth. A more representative
sample, obtained from fishery-independent sources, would allow determining an
unbiased growth curve for the population. Also, band periodicity was only partly
validated; hence, further research, on both validation of periodicity of band
formation and absolute age, is essential. A large scale tagging program might allow
for the age validation of this species and also the estimation of its natural mortality.
In this thesis, I fitted different growth models to age data, from which a model
other than the Von Bertalanffy growth model produced the best fit. My results
encourage the use of a suite of growth functions to determine the model that best
represents the data.
•
The risk assessment provided a step towards the conservation and sustainable
management of S. megalops. However, the population dynamics of this species are
still uncertain due to several factors. Time series of catch and effort, and fisherydependent estimates of relative abundance are only available for the otter trawl
method for the past 13 years (1992–05), and there is little information for other
fishing methods used in southern Australia. There are no accurate estimates of
181
natural and fishing mortality and a poor understanding of the selectivity of the trawl
and hook gears. This information is not only lacking for S. megalops, but also for
most shark species, particularly non-target ones.
8.7 CONCLUSIONS
The data presented in this thesis and the applied risk assessment approach have shown that
the life-history characteristics of S. megalops of aggregating behaviour, low reproductive
output, slow growth rate, and high longevity make this species particularly vulnerable to
the effects of fishing. Although published information indicates relative abundance has
been stable in several regions of southern Australia, given the low biological productivity
of S. megalops, changed fishing practices leading to increased fishing mortality could
quickly put this species at high risk. Hence, effective management of S. megalops will
require a close control of fishing mortality.
182
APPENDIX A
183
APPENDIX A
SUMMARY
Appendix A provides documentation from publishers that gives permission to reproduce
chapters in this thesis that were published manuscripts (or accepted for publication) at the
time of thesis submission (January 2006). This information applies to Chapters 3, 4, and 5.
Chapters under peer-review with journals at the time of thesis submission do not require
such information.
CHAPTER 3
Chapter 3 was accepted for publication in the journal Fisheries Research in 2006.
Permission from Elsevier Ltd. to reproduce this chapter is provided below.
Copy of email received:
Dear Dr Braccini
We hereby grant you permission to reproduce the material "Total and partial length–length,
mass–mass and mass–length relationships for the piked spurdog (Squalus megalops) in
south-eastern Australia" at no charge in your thesis subject to the following conditions:
1. If any part of the material to be used (for example, figures) has appeared in our
publication with credit or acknowledgement to another source, permission must also be
sought from that source. If such permission is not obtained then that material may not be
included in your publication/copies.
2. Suitable acknowledgement to the source must be made, either as a footnote or in a
reference list at the end of your publication, as follows:
"Reprinted from Publication title, Vol. number, Author(s), Title of article, Pages No.,
Copyright (Year), with permission from Elsevier".
3. Reproduction of this material is confined to the purpose for which permission is hereby
given.
4. This permission is granted for non-exclusive English rights only. For other languages
please reapply separately for each one required. Permission excludes use in an electronic
form. Should you have a specific electronic project in mind please reapply for permission.
184
5. Should your thesis be published commercially, please reapply for permission.
Yours sincerely
Marion Moss
Senior Rights Assistant
Global Rights Department
Elsevier Ltd
The Boulevard
Langford Lane
Kidlington, OX5 1GB, UK
CHAPTER 4
Chapter 4 was published in the journal ICES Journal of Marine Science in 2005.
On behalf of Elsevier Ltd., permission from the Editor-in-Chief (Dr Andrew I.L. Payne) to
reproduce this chapter is provided below.
Copy of email received:
Hello Matias
Such permissions are normally given by the Technical Editor at ICES, Bill Anthony, but
because I know he is away for a few days, I will give you the permission you require,
copied to him.
The ICES Journal of Marine Science simply asks that due prominence be given on the
paper/chapter to the fact that the paper is published in the ICES Journal, giving the page
numbers and volume please. I am assuming that the chapter will match the paper exactly.
If not, words to the effect that the chapter is "based on and is similar to a paper published
in the ICES Journal, title to be given, with the page numbers and volume also given, will
suffice.
Good luck
Dr Andrew I.L. Payne
(Editor-in-Chief: ICES Journal of Marine Science)
International Fisheries Consultant
CEFAS
Pakefield Road
Lowestoft
Suffolk NR33 0HT, UK
www.cefas.co.uk
185
CHAPTER 5
Chapter 5 was accepted for publication in the journal Marine and Freshwater Research in
2006.
Permission from CSIRO Publishing to reproduce this chapter is provided below.
Copy of email received:
Dear Matias
Thank you for your request.
Permission is granted for you to include in your PhD thesis the below mentioned
manuscript provided that:
(a) full acknowledgement is given to the journal, the authors and CSIRO PUBLISHING as
the copyright holder and publisher;
(b) it is clearly indicated that the article is scheduled to be published in next year's issue 1
of MFR;
(c) a pointer to the journal's website (http://www.publish.csiro.au/journals/mfr) is included
in your thesis so that interested parties will be able to obtain further information about the
journal
(d) use is only for reproducing the manuscript as one of the chapters of your thesis.
With best wishes
Carla Flores
Rights & Permissions
CSIRO PUBLISHING
150 Oxford Street (PO Box 1139)
Collingwood, Victoria 3066
tel: +61 3 9662 7652
fax: +61 3 9662 7555
186
REFERENCES
187
REFERENCES
Agardy T (2000) Effects of fisheries on marine ecosystems: a conservationist's perspective.
ICES Journal of Marine Science 57, 761–765.
Alonzo SH, Switzer PV, Mangel M (2003) An ecosystem-based approach to management:
using individual behaviour to predict the indirect effects of Antarctic krill fisheries
on penguin foraging. Journal of Applied Ecology 40, 692–702.
Anderson CIH, Rodhouse PG (2001) Life cycles, oceanography and variability:
ommastrephid squid in variable oceanographic environments. Fisheries Research 54,
133–143.
Anderson ED (1990) Fishery models as applied to elasmobranch fisheries. In
Elasmobranchs as living resources: advances in biology, ecology, systematics, and
the status of the fisheries. (Eds HL Pratt Jr., SH Gruber, T Taniuchi) pp. 473–484.
NOAA Technical Report, NMFS 90: Seattle.
Anderson MJ (2001) A new method for non-parametric multivariate analysis of variance in
ecology. Austral Ecology 26, 32–46.
Anonymous (2002) The Australian national plan of action for the conservation and
management of sharks. Department of Agriculture, Fisheries and Forestry: Australia.
Bass AJ (1973) Analysis and description of variation in the proportional dimensions of
Scyliorhinid, Carcharhinid and Sphyrnid sharks. Oceanographic Research Institute
Investigational Report 32: Durban.
Bass AJ, D'Aubrey JD, Kistnasamy N (1976) Sharks of the east coast of southern Africa.
VI. The families Oxynotidae, Squalidae, Dalatiidae and Echinorhinidae.
Oceanographic Research Institute Investigational Report 45: Durban.
188
Bax NJ (1998) The significance and prediction of predation in marine fisheries. ICES
Journal of Marine Science 55, 997–1030.
Beamish RJ, Fournier DA (1981) A method for comparing the precision of a set of age
determinations. Canadian Journal of Fisheries and Aquatic Sciences 38, 982–983.
Beamish RJ, McFarlane GA (1983) The forgotten requirement for age validation in
fisheries biology. Transactions of the American Fisheries Society 112, 735–743.
Beamish RJ, McFarlane GA (1985) Annulus development on the second dorsal spine of
the spiny dogfish (Squalus acanthias) and its validity for age determination.
Canadian Journal of Fisheries and Aquatic Sciences 42, 1799–1805.
Beauchamp JJ, Olson JS (1973) Corrections for bias in regression estimates after
logarithmic transformation. Ecology 54, 1403–1407.
Bethea DM, Buckel JA, Carlson JK (2004) Foraging ecology of the early life stages of four
sympatric shark species. Marine Ecology Progress Series 268, 245–264.
Bigg MA, Fawcett I (1985) Two biases in diet determination of northern fur seals
(Callorhinus ursinus). In Marine mammals and fisheries. (Eds JR Beddington, RJH
Beverton, DM Lavigne) pp. 284–291. George Allen & Unwin: London.
Bigg MA, Perez MA (1985) Modify volume: a frequency-volume method to assess marine
mammal food habits. In Marine mammals and fisheries. (Eds JR Beddington, RJH
Beverton, DM Lavigne) pp. 277–283. George Allen & Unwin: London.
Bonfil R (1994) Overview of world elasmobranch fisheries. FAO Fisheries Technical
Paper 341: Rome.
Bonfil R (2004) Fishery stock assessment models and their application to sharks. In
Elasmobranch fisheries management techniques. (Eds JA Musick, R Bonfil) pp.
205–240. Asia Pacific Economic Cooperation: Singapore.
189
Botsford LW, Castilla JC, Peterson CH (1997) The management of fisheries and marine
ecosystems. Science 277, 209–515.
Bowyer TR (2004) Sexual segregation in ruminants: definitions, hypotheses, and
implications for conservation and management. Journal of Mammalogy 85, 1039–
1052.
Braccini JM, Gillanders BM, Walker TI (2005) Sources of variation in the feeding ecology
of the piked spurdog (Squalus megalops): implications for inferring predator–prey
interactions from overall dietary composition. ICES Journal of Marine Science 62,
1076–1094.
Braccini JM, Gillanders BM, Walker TI (2006) Determining reproductive parameters for
population assessments of chondrichthyan species with asynchronous ovulation and
parturition: piked spurdog (Squalus megalops) as a case study. Marine and
Freshwater Research 57, in press.
Bridge NF, Mackay D, Newton G (1998) Biology of the ornate angel shark (Squatina
tergocellata) from the Great Australian Bight. Marine and Freshwater Research 49,
679–686.
Brooks JL, Dodson SI (1965) Predation, body size, and composition of plankton. Science
150, 28–35.
Brown LP, Bridge NF, Walker TI (2000) Summary of tag releases and recaptures in the
southern shark fisheries. Marine and Freshwater Research Institute Report 18:
Queenscliff.
Buckland ST, Burnham KP, Augustin NH (1997) Model selection: an integral part of
inference. Biometrics 53, 603–618.
Bulman C, Althaus F, He X, Bax NJ, Williams A (2001) Diets and trophic guilds of
demersal fishes of the south-eastern Australian shelf. Marine and Freshwater
Research 52, 537–548.
190
Burnham KP, Anderson DR (2002) Model selection and multimodel inference: a
practical information-theoretic approach. Springer-Verlag: New York.
Caddy JF, Sharp GD (1986) An ecological framework for marine fishery investigations.
FAO Fisheries Technical Paper 283: Rome.
Cailliet GM (1990) Elasmobranch age determination and verification: an updated review.
In Elasmobranchs as living resources: advances in biology, ecology, systematics, and
the status of the fisheries. (Eds HL Pratt Jr., SH Gruber, T Taniuchi) pp. 157–165.
NOAA Technical Report, NMFS 90: Seattle.
Cailliet GM, Goldman KJ (2004) Age determination and validation in chondrichthyan
fishes. In The biology of sharks and their relatives. (Eds JF Carrier, JA Musick, MR
Heithaus) pp. 399–447. CRC press: Boca Raton.
Campana SE (2001) Accuracy, precision and quality control in age determination,
including a review of the use and abuse of age validation methods. Journal of Fish
Biology 59, 197–242.
Campana SE, Annand MC, McMillan JI (1995) Graphical and statistical methods for
determining the consistency of age determinations. Transactions of the American
Fisheries Society 124, 131–138.
Cannizzaro L, Rizzo P, Levi D, Gancitano S (1995) Age-determination and growth of
Squalus blainvillei (Risso, 1826). Fisheries Research 23, 113–125.
Carlson JK, Baremore IE (2005) Growth dynamics of the spinner shark (Carcharhinus
brevipinna) off the United States southeast and Gulf of Mexico coasts: a comparison
of methods. Fishery Bulletin 103, 280–291.
Carlson JK, Cortés E, Bethea DM (2003) Life history and population dynamics of the
finetooth shark (Carcharhinus isodon) in the northeastern Gulf of Mexico. Fishery
Bulletin 101, 282–292.
191
Carrassón M, Stefanescu C, Cartes JE (1992) Diets and bathymetric distributions of two
bathyal sharks of the Catalan deep sea (western Mediterranean). Marine Ecology
Progress Series 82, 21–30.
Castro JI, Woodley CM, Brudek RL (1999) A preliminary evaluation of the status of shark
species. FAO Fisheries Technical Paper 380: Rome.
Caswell H (2001) Matrix population models: construction, analysis and interpretation.
Sinauer: Sunderland.
Cerrato RM (1990) Interpretable statistical tests for growth comparisons using parameters
in the von Bertalanffy equation. Canadian Journal of Fisheries and Aquatic Sciences
47, 1416–1426.
Chang WYB (1982) A statistical method for evaluating the reproducibility of age
determination. Canadian Journal of Fisheries and Aquatic Sciences 39, 1208–1210.
Chen C, Taniuchi T, Nose Y (1981) Some aspects of reproduction in the pointed-snout
dogfish Squalus japonicus taken off Nagasaki and Choshi. Bulletin of the Japanese
Society of Scientific Fisheries 47, 1157–1164.
Chen SB, Watanabe S (1989) Age dependence of natural mortality coefficient in fish
population dynamics. Nippon Suisan Gakkaishi 55, 205–208.
Cheung WL, Pitcher TJ, Pauly D (2005) A fuzzy logic expert system to estimate intrinsic
extinction vulnerabilities of marine fishes to fishing. Conservation Biology 124, 97–
111.
Chiaramonte GE, Pettovello AD (2000) The biology of Mustelus schmitti in southern
Patagonia, Argentina. Journal of Fish Biology 57, 930–942.
Christensen V (1995) A model of trophic interactions in the North Sea in 1981, the year of
the stomach. Dana 11, 1–28.
192
Clark MR, King KJ (1989) Deepwater fish resources off the North Island, New Zealand:
results of a trawl survey, May 1985 to June 1986. New Zealand Fisheries Technical
Report 11: Wellington.
Clarke KR (1993) Non-parametric multivariate analyses of changes in community
structure. Australian Journal of Ecology 18, 117–143.
Clarke MW (2000) Aspects of the biology of three exploited deepwater sharks
Centrophorus squamosus, Centroscymnus coelolepis and Deania calceus
(Elasmobranchii: Squalidae) from the continental slopes of the Rockall Trough and
Porcupine Bank. Ph.D. thesis. University College: Dublin.
Clarke MW, Connolly PL, Bracken JJ (2001) Aspects of reproduction of the deep water
sharks Centroscymnus coelolepis and Centrophorus squamosus from west of Ireland
and Scotland. Journal of the Marine Biological Association of the United Kingdom
81, 1019–1029.
Clarke MW, Connolly PL, Bracken JJ (2002a) Catch, discarding, age estimation, growth
and maturity of the squalid shark Deania calceus west and north of Ireland. Fisheries
Research 56, 139–153.
Clarke MW, Connolly PL, Bracken JJ (2002b) Age estimation of the exploited deepwater
shark Centrophorus squamosus from the continental slopes of the Rockall Trough
and Porcupine Bank. Journal of Fish Biology 60, 501–514.
Compagno LJV (1984) FAO species catalogue: sharks of the world. An annotated and
illustrated catalogue of shark species known to date. (1) Hexanchiformes to
Lamniformes. FAO Fisheries Synopsis 125, Vol. 4: Rome.
Compagno LJV (1990a) Shark exploitation and conservation. In Elasmobranchs as living
resources: advances in the biology, ecology, systematics, and the status of the
fisheries. (Eds HL Pratt Jr., SH Gruber, T Taniuchi) pp. 391–414. NOAA Technical
Report, NMFS 90: Seattle.
193
Compagno LJV (1990b) Alternative life-history styles of cartilaginous fishes in time and
space. Environmental Biology of Fishes 28, 33–75.
Compagno LJV, Ebert DA, Cowley PD (1991) Distribution of offshore demersal
cartilaginous fish (Class Chondrichthyes) off the west coast of southern Africa, with
notes on their systematics. South African Journal of Marine Science 11, 43–139.
Conrath CL (2004) Reproductive biology. In Elasmobranch fisheries management
techniques. (Eds JA Musick, R Bonfil) pp. 133–164. Asia-Pacific Economic
Cooperation: Singapore.
Cortés E (1997) A critical review of methods of studying fish feeding based on analysis of
stomach contents: application to elasmobranch fishes. Canadian Journal of Fisheries
and Aquatic Sciences 54, 726–738.
Cortés E (1998a) Demographic analysis as an aid in shark stock assessment and
management. Fisheries Research 39, 199–208.
Cortés E (1998b) Methods of studying fish feeding: reply. Canadian Journal of Fisheries
and Aquatic Sciences 55, 2708.
Cortés E (1999) Standardized diet compositions and trophic levels of sharks. ICES Journal
of Marine Science 56, 707–717.
Cortés E (2000) Life history patterns and correlations in sharks. Reviews in Fisheries
Science 8, 299–344.
Cortés E (2002) Incorporating uncertainty into demographic modelling: application to
shark populations and their conservation. Conservation Biology 16, 1048–1062.
Cortés E (2004) Life history patterns, demography, and population dynamics. In Biology
of sharks and their relatives. (Eds JF Carrier, JA Musick, MR Heithaus) pp. 449–
469. CRC press: Boca Raton.
194
Cortés E, Gruber SH (1990) Diet, feeding habits and estimates of daily ration of young
lemon sharks, Negaprion brevirostris (Poey). Copeia 1990, 204–218.
Cortés E, Manire CA, Hueter RE (1996) Diet, feeding habits, and diel feeding chronology
of the bonnethead shark, Sphyrna tiburo, in southwest Florida. Bulletin of Marine
Science 58, 353–367.
Craik JCA (1978) An annual cycle of vitellogenesis in the elasmobranch Scyliorhinus
canicula. Journal of the Marine Biological Association of the United Kingdom 58,
719–726.
Daley RK, Stevens JD, Graham KJ (2002) Catch analysis and productivity of the
deepwater dogfish resource in southern Australia. Fisheries Research and
Development Corporation Project 1998/108: Hobart.
Druffel EM, Linick TW (1978) Radiocarbon in annual coral rings of Florida. Geophysical
Research Letters 5, 913–916.
Dulvy NK, Sadovy Y, Reynolds JD (2003) Extinction vulnerability in marine populations.
Fish and Fisheries 4, 25–64.
Ebert DA (2002) Ontogenetic changes in the diet of the sevengill shark (Notorynchus
cepedianus). Marine and Freshwater Research 53, 517–523.
Ebert DA, Compagno LJV, Cowley PD (1992) A preliminary investigation of the feeding
ecology of squaloid sharks off the west coast of southern Africa. South African
Journal of Marine Science 12, 601–609.
Ellis JR, Pawson MG, Shackley SE (1996) The comparative feeding ecology of six species
of shark and four species of ray (Elasmobranchii) in the north-east Atlantic. Journal
of the Marine Biological Association of the United Kingdom 76, 89–106.
Fabens AJ (1965) Properties and fitting of the von Bertalanffy growth curve. Growth 29,
265–289.
195
FAO (2000a) The state of world fisheries and aquaculture. FAO: Rome.
FAO (2000b) Fisheries management. 1 Conservation and management of sharks. FAO
Technical Guidelines for Responsible Fisheries 4 supplement 1: Rome.
Ferry LA, Cailliet GM (1996) Sample size and data analysis: are we characterizing and
comparing diet properly? In Feeding ecology and nutrition in fish: Proceedings of the
symposium on the feeding ecology and nutrition in fish, international congress on the
biology of fishes, San Francisco, 14–18 July 1996. (Eds D MacKinlay, K Shearer)
pp. 71–80. American Fisheries Society: San Francisco.
Floeter J, Temming A (2003) Explaining diet composition of North Sea cod (Gadus
morhua): prey size preference vs. prey availability. Canadian Journal of Fisheries
and Aquatic Sciences 60, 140–150.
Francis MP, Stevens JD (2000) Reproduction, embryonic development, and growth of the
porbeagle shark, Lamna nasus, in the southwest Pacific Ocean. Fishery Bulletin 98,
41–63.
Francis RICC (1988) Are growth parameters estimated from tagging and age-length data
comparable? Canadian Journal of Fisheries and Aquatic Sciences 45, 936–942.
Fulton EA, Smith ADM, Johnson CR (2003) Effect of complexity on marine ecosystem
models. Marine Ecology Progress Series 253, 1–16.
Gelsleichter J, Musick JA, Nichols S (1999) Food habits of the smooth dogfish, Mustelus
canis, dusky shark, Carcharhinus obscurus, Atlantic sharpnose shark,
Rhizoprionodon terraenovae, and the sand tiger, Carcharias taurus, from the
northwest Atlantic Ocean. Environmental Biology of Fishes 54, 205–217.
Graham KJ (2005) Distribution, population structure and biological aspects of Squalus spp.
(Chondrichthyes: Squaliformes) from New South Wales and adjacent Australian
waters. Marine and Freshwater Research 56, 405–416.
196
Graham KJ, Andrew NL, Hodgson KE (2001) Changes in relative abundance of sharks and
rays on Australian South East Fishery trawl grounds after twenty years of fishing.
Marine and Freshwater Research 52, 549–561.
Guallart J, Vicent JJ (2001) Changes in composition during embryo development of the
gulper shark, Centrophorus granulosus (Elasmobranchii, Centrophoridae): an
assessment of maternal-embryonic nutritional relationships. Environmental Biology
of Fishes 61, 135–150.
Gulland JA (1978) Fishery management: new strategies for new conditions. Transactions
of the American Fisheries Society 107, 1–11.
Haddon M (2001) Modelling and quantitative methods in fisheries. Chapman & Hall/CRC:
Boca Raton.
Hall MA (1996) On bycatches. Reviews in Fish Biology and Fisheries 6, 319–352.
Hamlett WC, Koob TJ (1999) Female reproductive system. In Sharks, skates, and rays.
The biology of elasmobranch fishes. (Ed. WC Hamlett) pp. 398–443. The Johns
Hopkins University Press: Baltimore.
Hanchet S (1988) Reproductive biology of Squalus acanthias from the east coast, South
Island, New Zealand. New Zealand Journal of Marine and Freshwater Research 22,
537–549.
Hanchet S (1991) Diet of spiny dogfish, Squalus acanthias Linnaeus, on the east coast,
South Island, New Zealand. Journal of Fish Biology 39, 313–323.
Heithaus MR (2004) Predator–prey interactions. In Biology of sharks and their relatives.
(Eds JF Carrier, JA Musick, MR Heithaus) pp. 487–521. CRC press: Boca Raton.
Hilton-Taylor C (2000) 2000 IUCN red list of threatened species. IUCN The World
Conservation Union: Gland.
197
Hobday A, Smith ADM, Stobutzki I (2004) Ecological risk assessment for Australian
Commonwealth fisheries. Final report stage 1. Hazard identification and preliminary
risk assessment. Australian Fisheries Management Authority Report R01/0934:
Canberra.
Hoenig JM (1983) Empirical use of longevity data to estimate mortality rates. Fishery
Bulletin 82, 898–903.
Hoenig JM, Gruber SH (1990) Life-history patterns in the elasmobranchs: Implications for
fisheries management. In Elasmobranchs as living resources: advances in the
biology, ecology, systematics, and the status of the fisheries. (Eds HL Pratt Jr., SH
Gruber, T Taniuchi) pp. 1–16. NOAA Technical Report, NMFS 90: Seattle.
Hoenig JM, Morgan MJ, Brown CA (1995) Analysing differences between two age
determination methods by tests of symmetry. Canadian Journal of Fisheries and
Aquatic Sciences 52, 364–368.
Holden MJ (1973) Are long-term sustainable fisheries for elasmobranchs possible?
Rapports et Proces-Verbaux des Reunions du Conseil International pour
l’Exploration de la Mer 164, 360–367.
Holden MJ (1974) Problems in the rational exploitation of elasmobranch populations and
some suggested solutions. In Sea Fisheries Research. (Ed. FR Harden Jones) pp.
117–137. Paul Elek: London.
Holden MJ (1977) Elasmobranchs. In Fish population dynamics. (Ed. JA Gulland) pp.
187–216. John Wiley: London.
Holden MJ, Meadows PS (1962) The structure of the spine of the spur dogfish (Squalus
acanthias L.) and its use for age determination. Journal of the Marine Biological
Association of the United Kingdom 42, 179–197.
Hyslop EJ (1980) Stomach contents analysis–a review of methods and their application.
Journal of Fish Biology 17, 411–429.
198
Irvine SB (2004) Age, growth and reproduction of deepwater dogfishes from southeastern
Australia. Ph.D. thesis. Deakin University: Warrnambool.
IUCN (2002) The role of CITES in the conservation and management of sharks. IUCN
Species Survival Commission Sharks Specials Group and TRAFFIC.
Jensen AL (1996) Beverton and Holt life history invariants result from optimal trade-off of
reproduction and survival. Canadian Journal of Fisheries and Aquatic Sciences 53,
820–822.
Johannes RE (1998) The case for data-less marine resource management: examples from
tropical nearshore finfisheries. Trends in Ecology and Evolution 13, 243–246.
Jones BC, Geen GH (1977a) Food and feeding of spiny dogfish (Squalus acanthias) in
British Columbia waters. Journal of the Fisheries Research Board of Canada 34,
2067–2078.
Jones BC, Geen GH (1977b) Reproduction and embryonic development of spiny dogfish
(Squalus acanthias) in the Strait of Georgia, British Columbia. Journal of the
Fisheries Research Board of Canada 34, 1286–1292.
Jones GK (1985) An exploratory dropline survey for deepsea trevalla (Hyperoglyphe
antarctica) in continental slope waters off South Australia. Department of Fisheries
Research Paper 15: Adelaide.
Ketchen KS (1975) Age and growth of dogfish Squalus acanthias in British Columbia
waters. Journal of the Fisheries Research Board of Canada 32, 43–59.
Kibesaki O (1954) Studies on the sharks from the East China and Yellow Seas-2. Maturity
of the S. brevirostris. Bulletin of the Seikai National Fisheries Research Institute.
Nagasaki 5, 36–46.
Kimura DK (1980) Likelihood methods for the von Bertalanffy growth curve. Fishery
Bulletin 77, 765–776.
199
Kleinbaum DG, Kupper LL, Muller KE (1988) Applied regression analysis and other
multivariable methods. PWS-Kent Publishing Company: Boston.
Klimley AP (1987) The determinants of sexual segregation in the scalloped hammerhead,
Sphyrna lewini. Environmental Biology of Fishes 18, 27–40.
Knight W (1968) Asymptotic growth: an example of nonsense disguised as mathematics.
Journal of the Fisheries Research Board of Canada 25, 1303–1307.
Koen Alonso M, Crespo EA, Garcia NA, Pedraza SN, Mariotti PA, Mora NJ (2002)
Fishery and ontogenetic driven changes in the diet of the spiny dogfish, Squalus
acanthias, in Patagonian waters, Argentina. Environmental Biology of Fishes 63,
193–202.
Kohler NE, Casey JG, Turner PA (1995) Length-weight relationships for 13 species of
sharks form the western North Atlantic. Fishery Bulletin 93, 412–418.
Laptikhovsky VV, Arkhipkin AI, Henderson AC (2001) Feeding habits and dietary overlap
in spiny dogfish Squalus acanthias (Squalidae) and narrowmouth catshark
Schroederichthys bivius (Scyliorhinidae). Journal of the Marine Biological
Association of the United Kingdom 81, 1015–1018.
Last PR, Stevens JD (1994) Sharks and rays of Australia. CSIRO Publishing: Melbourne.
Ludwig D, Hilborn R, Walters C (1993) Uncertainty, resource exploitation, and
conservation: lessons from history. Science 260, 17, 36.
Lyle JM (1983) Food and feeding habits of the lesser spotted dogfish, Scyliorhinus
canicula (L.), in Isle of Man waters. Journal of Fish Biology 23, 725–737.
MacDonald JS, Green RH (1983) Redundancy of variables used to describe importance of
prey species in fish diets. Canadian Journal of Fisheries and Aquatic Sciences 40,
635–637.
200
Main MB, Weckerly FW, Bleich VC (1996) Sexual segregation in ungulates: new
directions for research. Journal of Mammalogy 77, 449–461.
Maisey JG (1979) Finspine morphogenesis in squalid and heterodontid sharks. Zoological
Journal of the Linnean Society 66, 161–183.
Martin TG, Kuhnert PM, Mengersen K, Possingham HP (2005) The power of expert
opinion in ecological models using Bayesian methods: impact of grazing on birds.
Ecological Applications 15, 266–280.
McAllister MK, Pikitch EK, Babcock EA (2001) Using demographic methods to construct
Bayesian priors for the intrinsic rate of increase in the Schaefer model and
implications for stock rebuilding. Canadian Journal of Fisheries and Aquatic
Sciences 58, 1871–1890.
McFarlane GA, Beamish RJ (1987a) Validation of the dorsal spine method of age
determination for spiny dogfish. In Age and growth of fish. (Eds RC Summerfelt,
GE Hall) pp. 287–300. State University Press: Ames.
McFarlane GA, Beamish RJ (1987b) Selection of dosages of oxytetracycline for age
validation studies. Canadian Journal of Fisheries and Aquatic Sciences 44, 905–909.
McLaughlin RH, O'Gower AK (1971) Life history and underwater studies of a Heterodont
shark. Ecological Monographs 41, 271–289.
Medved RJ, Stillwell CE, Casey JG (1985) Stomach contents of young sandbar sharks,
Carcharhinus plumbeus, in Chincoteague Bay, Virginia. Fishery Bulletin 83, 395–
402.
Mollet HF, Ezcurra JM, O’Sullivan JB (2002) Captive biology of the pelagic stingray,
Dasyatis violacea (Bonaparte, 1832). Marine and Freshwater Research 53, 531–541.
201
Morato T, Solá E, Grós MP, Menezes G (2003) Diets of thornback ray (Raja clavata) and
tope shark (Galeorhinus galeus) in the bottom longline fishery of the Azores,
northeastern Atlantic. Fishery Bulletin 101, 590–602.
Moulton PL, Walker TI, Saddlier SR (1992) Age and growth studies of gummy shark,
Mustelus antarcticus Günther, and school shark, Galeorhinus galeus (Linnaeus),
from southern Australian waters. Australian Journal of Marine and Freshwater
Research 43, 1241–1267.
Musick JA, Burgess G, Cailliet G, Camhi M, Fordham S (2000) Management of sharks
and their relatives (Elasmobranchii). Fisheries 25, 9–13.
Nammack MF, Musick JA, Colvocoresses JA (1985) Life history of spiny dogfish off the
Northeastern United States. Transactions of the American Fisheries Society 114,
367–376.
Norman M, Reid A (2000) A guide to squid, cuttlefish and octopuses of Australasia.
CSIRO Publishing and the Gould League of Victoria: Melbourne.
Officer RA, Gason AS, Walker TI, Clement JG (1996) Sources of variation in counts of
growth increments in vertebrae from gummy shark, Mustelus antarcticus, and school
shark, Galeorhinus galeus: implications for age determination. Canadian Journal of
Fisheries and Aquatic Sciences 53, 1765–1777.
Paine RT (1988) Food webs: road maps of interactions or grist for theoretical
development? Ecology 69, 1648–1654.
Parsons GR (1993) Age determination and growth of the bonnethead shark Sphyrna tiburo:
a comparison of two populations. Marine Biology 117, 23–31.
Parsons GR, Grier HJ (1992) Seasonal changes in shark testicular structure and
spermatogenesis. Journal of Experimental Zoology 261, 173–184.
202
Pauly D (1980) On the interrelationships between natural mortality, growth parameters,
and mean environmental temperature in 175 fish. Journal du Conseil pour
L'Exploration de la Mer 39, 175–192.
Pauly D, Trites AW, Capuli E, Christensen V (1998) Diet composition and trophic levels
of marine mammals. ICES Journal of Marine Science 55, 467–481.
Pielou EC (1966) The measurement of diversity in different types of biological collections.
Journal of Theoretical Biology 63, 131–144.
Pinkas LM, Oliphant S, Iverson ILK (1971) Food habits of albacore, bluefin tuna and
bonito in Californian waters. California Fish and Game 152, 1–105.
Punt AE, Walker TI (1998) Stock assessment and risk analysis for the school shark
(Galeorhinus galeus) off southern Australia. Marine and Freshwater Research 49,
719–731.
Ricker WE (1973) Linear regression in fishery research. Journal of the Fisheries Research
Board of Canada 30, 409–434.
Ricker WE (1975) Computation and interpretation of biological statistics of fish
populations. Bulletin of the Fisheries Research Board of Canada 191, 1–382.
Ricker WE (1979) Growth rates and models. In Fish physiology 8. (Eds WS Hoar, DJ
Randall) pp. 677–743. Academic Press: New York.
Roa R, Ernst B, Tapia F (1999) Estimation of size at sexual maturity: an evaluation of
analytical and resampling procedures. Fishery Bulletin 97, 570–580.
Roff DA (1980) A motion for the retirement of the von Bertalanffy function. Canadian
Journal of Fisheries and Aquatic Sciences 37, 127–129.
Roper CFE, Young RE (1975) Vertical distributions of pelagic cephalopods. Smithsonian
Contributions to Zoology 209, 1–51.
203
Rose C, McLoughlin K (2001) Reviews of shark finning in Australian fisheries. Bureau of
Rural Sciences: Kingston.
Scharf FS, Juanes F, Roundtree RA (2000) Predator–prey relationships of marine fish
predators: interspecific variation and effects of ontogeny and body size on trophic–
niche breadth. Marine Ecology Progress Series 208, 229–248.
Simpfendorfer CA (1999) Demographic analysis of the dusky shark fishery in southwestern Australia. In Life in the slow lane: ecology and conservation of long-lived
marine animals. (Ed. JA Musick) pp. 149–160. American Fisheries Society
Symposium 23: Bethesda.
Simpfendorfer CA (2000) Growth rates of juvenile dusky sharks, Carcharhinus obscurus
(Lesueur, 1818), from southwestern Australia estimated from tag-recapture data.
Fishery Bulletin 98, 811–822.
Simpfendorfer CA, Donohue K, Hall. NG (2000) Stock assessment and risk analysis for
the whiskery shark (Furgaleus macki (Whitely)) in south-western Australia.
Fisheries Research 47, 1–17.
Simpfendorfer CA, Goodreid A, McAuley RB (2001a) Diet of three commercially
important shark species from Western Australian waters. Marine and Freshwater
Research 52, 975–985.
Simpfendorfer CA, Goodreid AB, McAuley RB (2001b) Size, sex and geographic
variation in the diet of the tiger shark, Galeocerdo cuvier, from Western Australian
waters. Environmental Biology of Fishes 61, 37–46.
Sims DW (2003) Tractable models for testing theories about natural strategies: foraging
behaviour and habitat selection of free-ranging sharks. Journal of Fish Biology 63
(Supplement A), 53–73.
204
Sims DW, Nash JP, Morritt D (2001) Movements and activity of male and female dogfish
in a tidal sea lough: alternative behavioural strategies and apparent sexual
segregation. Marine Biology 139, 1165–1175.
Smith SE, Au DW, Show C (1998) Intrinsic rebound potentials of 26 species of Pacific
sharks. Marine and Freshwater Research 49, 663–678.
Soriano M, Moreau J, Hoenig JM, Pauly D (1992) New functions for the analysis of twophase growth of juvenile and adult fishes, with application to Nile perch.
Transactions of the American Fisheries Society 121, 486–493.
Springer S (1967) Social organization of shark populations. In Sharks, skates, and rays.
(Eds PW Gilbert, RF Mathewson, DP Rall) pp. 149–174. The Johns Hopkins Press:
Baltimore.
Stearns SC (1992) The evolution of life histories. Oxford University Press: New York.
Stevens JD, Bonfil R, Dulvy NK, Walker PA (2000) The effects of fishing on sharks, rays,
and chimaeras (chondrichthyans), and the implications for marine ecosystems. ICES
Journal of Marine Science 57, 476–494.
Stevens JD, McLoughlin KJ (1991) Distribution, size and sex composition, reproductive
biology and diet of sharks from northern Australia. Australian Journal of Marine and
Freshwater Research 42, 151–199.
Stillwell CE, Kohler NE (1993) Food habits of the sandbar shark Carcharhinus plumbeus
off the U.S. northeast coast, with estimates of daily ration. Fishery Bulletin 91, 138–
150.
Stobutzki IC, Miller MJ, Heales DS, Brewer DT (2002) Sustainability of elasmobranchs
caught as bycatch in a tropical prawn (shrimp) trawl fishery. Fishery Bulletin 100,
800–821.
205
Taniuchi T, Tachikawa H (1999) Geographical variation in age and growth of Squalus
mitsukurii (Elasmobranchii: Squalidae) in the North Pacific. In Proceedings of the
5th Indo-Pacific fish conference, Noumea, New Caledonia, 3–8 November 1997.
(Eds B Séret, J-Y Sire) pp. 321–328. Societe Francaise d’Ichtyologie: Paris.
Taniuchi T, Tachikawa H, Shimizu M, Nose Y (1993) Geographical variations in
reproductive parameters of shortspine spurdog in the North Pacific. Nippon Suisan
Gakkaishi 59, 45–51.
Triantafillos L, Jackson GD, Adams M, McGrath Steer BL (2004) An allozyme
investigation of the stock structure of arrow squid Nototodarus gouldi (Cephalopoda:
Ommastrephidae) from Australia. ICES Journal of Marine Science 61, 829–835.
Tucker R (1985) Age validation studies on the spines of the spurdog (Squalus acanthias)
using tetracycline. Journal of the Marine Biological Association of the United
Kingdom 65, 641–651.
Vögler R, Milessi AC, Quiñones RA (2003) Trophic ecology of Squatina guggenheim on
the continental shelf off Uruguay and northern Argentina. Journal of Fish Biology
62, 1254–1267.
von Bertalanffy L (1938) A quantitative theory of organic growth (inquiries on growth
laws. II). Human Biology 10, 181–213.
Vooren CM (1992) Reproductive strategies of eight species of viviparous elasmobranchs
from southern Brazil. Bulletin de la Societe Zoologique de France 117, 303–312.
Walker TI (1992) Fishery simulation model for sharks applied to the gummy shark,
Mustelus antarcticus Günther, from Southern Australian waters. Australian Journal
of Marine and Freshwater Research 43, 195–212.
Walker TI (1998) Can shark resources be harvested sustainably? A question revisited with
a review of shark fisheries. Marine and Freshwater Research 49, 553–572.
206
Walker TI (1999) Southern Australian shark fishery management. In Case studies of the
management of elasmobranch fisheries. (Ed. R Shotton) pp. 480–514. FAO Fisheries
Technical Papers 378/2: Rome.
Walker TI (2004) Management measures. In Elasmobranch fisheries management
techniques. (Eds JA Musick, R Bonfil) pp. 285–321. Asia Pacific Economic
Cooperation: Singapore.
Walker TI (2005) Reproduction in fisheries science. In Reproductive biology and
phylogeny of Chondrichthyes: sharks, rays and chimaeras. (Ed. WC Hamlett) pp.
81–127. Science Publishers Inc.: Enfield.
Walker TI, Gason AS (2005) GHATF monitoring database management 2004/0. Final
report for Project No R03/1383. Australian Fisheries Management Authority:
Canberra.
Walker TI, Gason AS (2006) Shark and other chondrichthyan byproduct and bycatch
estimation in the South East Fishery trawl and non-trawl sectors. Final report to the
Fisheries Research and Development Corporation for Project No 2001/007. Primary
Industries Research Victoria: Queenscliff.
Walker TI, Hudson RJ, Gason AS (2005) Catch evaluation of target, by-product and bycatch species taken by gillnets and longlines in the shark fishery of south-eastern
Australia. Journal of Northwest Atlantic Fishery Science 35, 505–530.
Walker TI, Taylor BL, Hudson RJ, Cottier JP (1998) The phenomenon of apparent change
of growth rate in gummy shark (Mustelus antarcticus) harvested off southern
Australia. Fisheries Research 39, 139–163.
Walters C, Christensen V, Pauly D (1997) Structuring dynamic model of exploited
ecosystems from trophic mass-balance assessments. Reviews in Fish Biology and
Fisheries 7, 139–172.
207
Watson G, Smale MJ (1998) Reproductive biology of shortnose spiny dogfish, Squalus
megalops, from the Agulhas Bank, South Africa. Marine and Freshwater Research
49, 695–703.
Watson G, Smale MJ (1999) Age and growth of the shortnose spiny dogfish Squalus
megalops from the Agulhas Bank, South Africa. South African Journal of Marine
Science 21, 9–18.
Wetherbee BM (1996) Distribution and reproduction of the southern lantern shark from
New Zealand. Journal of Fish Biology 49, 1186–1196.
Wetherbee BM, Cortés E (2004) Food consumption and feeding habits. In Biology of
sharks and their relatives. (Eds JF Carrier, JA Musick, MR Heithaus) pp. 225–246.
CRC press: Boca Raton.
Wetherbee BM, Gruber SH, Cortés E (1990) Diet, feeding habits, digestion, and
consumption in sharks, with special reference to the lemon shark, Negaprion
brevirostris. In Elasmobranchs as living resources: advances in biology, ecology,
systematics, and the status of the fisheries. (Eds HL Pratt Jr., SH Gruber, T Taniuchi)
pp. 29–47. NOAA Technical Report, NMFS 90: Seattle.
White WT, Platell ME, Potter IC (2004) Comparisons between the diets of four abundant
species of elasmobranchs in a subtropical embayment: implications for resource
partitioning. Marine Biology 144, 439–448.
Wilson CD, Seki MP (1994) Biology and population characteristics of Squalus mitsukurii
from a seamount in the Central North Pacific Ocean. Fishery Bulletin 92, 851–864.
Wintner SP (2000) Preliminary study of vertebral growth rings in the whale shark,
Rhincodon typus, from the east coast of South Africa. Environmental Biology of
Fishes 59, 441–451.
Wourms JP (1977) Reproduction and development in chondrichthyan fishes. American
Zoologist 17, 379–410.
208
Yano K (1993) Reproductive biology of the slender smoothhound Gollum attenuatus,
collected from New Zealand waters. Environmental Biology of Fishes 38, 59–71.
Yano K (1995) Reproductive biology of the black dogfish, Centroscyllium fabricii,
collected from waters off western Greenland. Journal of the Marine Biological
Association of the United Kingdom 75, 285–310.
Yano K, Tanaka S (1988) Size at maturity, reproductive cycle, fecundity, and depth
segregation of the deep sea squaloid sharks Centroscymnus owstoni and C. coelolepis
in Suruga Bay, Japan. Nippon Suisan Gakkaishi 54, 167–174.
Yodzis P (1994) Predator–prey theory and management of multispecies fisheries.
Ecological Applications 41, 51–58.
Yodzis P (1998) Local trophodynamics and the interaction of marine mammals and
fisheries in the Benguela ecosystem. Journal of Animal Ecology 67, 635–658.
Zar JH (1999) Biostatistical analysis. Prentice Hall: London.
209
View publication stats