Content uploaded by Patrick Campbell
Author content
All content in this area was uploaded by Patrick Campbell on Jun 03, 2014
Content may be subject to copyright.
MARINE ECOLOGY PROGRESS SERIES
Mar Ecol Prog Ser
Vol. 361: 267–278, 2008
doi: 10.3354/meps07399
Published June 9
© Inter-Research 2008 · www.int-res.com*Email: lisa.natanson@noaa.gov
Ontogenetic vertebral growth patterns in the
basking shark Cetorhinus maximus
Lisa J. Natanson
1,
*
, Sabine P. Wintner
2
, Friederike Johansson
3
, Andrew Piercy
4
,
Patrick Campbell
5
, Alessandro De Maddalena
6
, Simon J. B. Gulak
7
, Brett Human
8
,
Franco Cigala Fulgosi
9
, David A. Ebert
10
, Farid Hemida
11
, Frederik H. Mollen
12
,
Stefano Vanni
13
, George H. Burgess
4
, Leonard J. V. Compagno
14
,
Andrew Wedderburn-Maxwell
15
1
National Marine Fisheries Service (NMFS), Northeast Fisheries Science Center, NOAA, 28 Tarzwell Drive, Narragansett,
Rhode Island 02882-1199, USA
2
Natal Sharks Board, Private Bag 2, 4320 Umhlanga Rocks, and Biomedical Resource Unit, University of KwaZulu-Natal,
PO Box X54001, Durban 4000, South Africa
3
Vertebrate Section, Göteborg Natural History Museum, Box 7283, 402 35 Göteborg, Sweden
4
Florida Program for Shark Research, Florida Museum of Natural History, University of Florida, PO Box 117800, Gainesville,
Florida 32611, USA
5
Department of Zoology, Natural History Museum, Cromwell Road, London SW7 5BD, UK
6
Banca Dati Italiana Squalo Bianco, Via L. Ariosto 4, 20145 Milan, Italy
7
Pelagic Observer Program, South East Fisheries Science Centre, National Marine Fisheries Service, NOAA,
75 Virginia Beach Drive, Miami 33149, Florida, USA
8
Marine Science and Fisheries Centre (Biodiversity Project), PO Box 467, PC 100 Muscat, Sultanate of Oman
9
Dipartimento Scienze della Terra, Università di Parma, Parco Area delle Scienze 152/A, 43100 Parma, Italy
10
Pacific Shark Research Center, Moss Landing Marine Laboratories, 8272 Moss Landing Road, Moss Landing,
California 95039, USA
11
Laboratoire Ecologie et Environnement (Projet Halieutique), Faculté des Sciences Biologiques (FSB),
Université des Sciences et Techniques Houari Boumedienne (USTHB), BP 32, El Alia, 16111 Bab Ezzouar, Alger, Algeria
12
Elasmobranch Research, Meistraat 16, 2590 Berlaar, Belgium
13
Sezione di Zoologia ‘La Specola’, Museo di Storia Naturale dell’Università, Via Romana 17, 50125 Florence, Italy
14
Shark Research Center, South African Museum, PO Box 61, 8000 Cape Town, South Africa
15
Umhlanga Radiology, Radiology Department, Umhlanga Hospital, Private Bag X09, 4320 Umhlanga Rocks, South Africa
ABSTRACT: Age and growth of the basking shark Cetorhinus maximus (Gunnerus) was examined
using vertebral samples from 13 females (261 to 856 cm total length [TL]), 16 males (311 to 840 cm TL)
and 11 specimens of unknown sex (376 to 853 cm TL). Vertebral samples were obtained worldwide
from museums and institutional and private collections. Examination of multiple vertebrae from
along the vertebral column of 10 specimens indicated that vertebral morphology and band pair (alter-
nating opaque and translucent bands) counts changed dramatically along an individual column.
Smaller sharks had similar band pair counts along the length of the vertebral column while large
sharks had a difference of up to 24 band pairs between the highest and lowest count along the col-
umn. Our evidence indicates that band pair deposition may be related to growth and not time in this
species and thus the basking shark cannot be directly aged using vertebral band pair counts.
KEY WORDS: Basking shark · Cetorhinus maximus · Age · Growth · Vertebrae
Resale or republication not permitted without written consent of the publisher
Mar Ecol Prog Ser 361: 267–278, 2008
INTRODUCTION
The basking shark Cetorhinus maximus (Gunnerus)
has been the target of various fisheries worldwide as far
back as the 18th century (Castro et al. 1999). Perfor-
mance of historical basking shark fisheries suggests that
the species is particularly susceptible to overfishing
(Castro et al. 1999). The basking shark is assessed as
‘Vulnerable’ on the International Union for the Conser-
vation of Nature and Natural Resources (IUCN) Red List
of Threatened Species and is listed in Appendix II of the
Convention on International Trade in Endangered
Species of Wild Fauna and Flora (CITES) (http://sea.
unep-wcmc.org. Little is known about the biology of this
species, including age and growth (Castro et al. 1999).
The age of a shark is commonly determined by coun-
ting alternating opaque and translucent band pairs
deposited in vertebrae. Access to basking shark verte-
brae, however, is very limited, especially those for
which accurate morphological information is available.
This has hampered attempts to elucidate age and
growth in this species and previous results have been
questioned (Parker & Stott 1965, Pauly 2002).
Previous studies of basking shark age have relied
on small sample sizes and suspect methodology. Mat-
thews (1950) presented a growth curve based on length
frequency information from 13 individuals, but stressed
that it was a tentative growth curve. Parker & Boeseman
(1954) reanalyzed Matthews’ (1950) data with an addi-
tional 41 data points obtained from the literature and
museums. Parker & Stott (1965) re-examined the Parker
& Boeseman (1954) data using a graphical approach and
an empirical growth formula, added data, and attempted
verification with vertebral counts from 5 specimens. Us-
ing all the graphical analyses and vertebral band pair
counts, Parker & Stott (1965) concluded that there were
7 pre-birth band pairs and that subsequent band pairs
were deposited biannually. These 3 papers contained
data adjustments (adding lengths to the original data
points), questionable assumptions (arbitrary monthly
grouping of animals), or plotting and/or calculation er-
rors. Pauly (1978) questioned the biannual deposition
theory, reanalyzed all the previous data using several
growth models and concluded that the previous analyses
were erroneous and the 7 pre-birth bands ‘bizarre’.
While these studies used the best data available at the
time, the sample sizes were quite limiting. We attempted
to determine age for Cetorhinus maximus based on
vertebral band pair counts from sources worldwide.
MATERIALS AND METHODS
Sampling. Vertebral samples from 40 basking sharks
were obtained from museums and institutional and pri-
vate collections. Samples came from several oceans
and seas, including the North Atlantic, South Atlantic,
Mediterranean, Adriatic, Southwest Indian, North Sea
and Northeast Pacific. Vertebrae were preserved using
methods of the collectors or institutions, and included:
ethanol and/or formaldehyde fixation, drying, and
freezing. Samples ranged from 1 vertebra to the entire
vertebral columns from each specimen. Maturity con-
dition of individual samples was based on the data pro-
vided by the collector (see Table 1).
Morphological measurements were inconsistent. In
many cases it is unknown if the measurement was
taken over the body or was a straight line (caliper)
measurement; additionally, length type was not always
specified. Often the location of the vertebrae along the
column was not specified.
Total length (TL, cm) with the tail in natural position
was used throughout this study. In specimens where
the type of length was not given, the measurement or
estimate was assumed to be TL. A variety of morpho-
logical measurements was available for these speci-
mens, including fork and pre-caudal lengths (FL and
PCL, respectively) and girth from behind the pectoral,
dorsal and pelvic fins. These data and data from the
literature (Bigelow & Schroeder 1948, Matthews 1950,
Wood 1957, Siccardi 1961, Chen 1963, Springer &
Gilbert 1976, Cadenat & Blache 1981, Izawa & Shibata
1993, Soldo 1999) were used to calculate the following
conversions:
TL = 1.09 × FL + 12.01
(r
2
= 0.996, n = 16; 261 to 846 cm TL)
TL = 1.20 × PCL + 15.53
(r
2
= 0.995, n = 17; 261 to 855 cm TL)
FL = 1.14 × PCL – 8.32
(r
2
= 0.999, n = 20; 261 to 846 cm TL)
Calculated TL values are indicated throughout with
a
.
The TL–weight (WT, kg) relationship was generated
using data from the collected specimens and published
data (Bigelow & Schroeder 1948, Wood 1957, Siccardi
1961, Springer & Gilbert 1976, Cadenat & Blache 1981,
Howes 1998, Soldo 1999, Serena et al. 2000, Zuffa et
al. 2001, Schwartz 2002) and references contained
therein;
WT = 9.0073 × 10
–6
× TL
2.94
(n = 59, range 249 to 1000 cm, 95% CI on exponent =
2.74 to 3.12, r
2
= 0.944)
Vertebral measurements were made by the collec-
tion institutions using guidelines for whale sharks
following Wintner (2000). Measurements included
length, dorsal and lateral diameter of each centrum, as
well as diameter at the angle change (Wintner 2000).
The angle change on the centrum face was regarded
as the birthmark (Wintner 2000). Analyses were per-
268
Natanson et al.: Vertebral growth in the basking shark
formed separately on vertebrae obtained from 3 differ-
ent regions: trunk, abdomen and tail. All vertebral
dimensions (VD, cm) were related to TL to determine
the proportional relationship between somatic and
vertebral growth so that the correct back-calculation
method could be chosen (Goldman 2004).
The relationship between lateral diameter (LD, cm)
and TL was calculated to determine the best method
for back-calculation of length-at-age data and to con-
firm the interpretation of the birth band. Regressions
were fitted to the data by sex and an analysis of covari-
ance (ANCOVA) was used to test for difference be-
tween the sexes. Based on recommendations in Cailliet
& Goldman (2004), multiple back-calculation methods
were examined to find the most appropriate biological
and statistical fit.
The analyses included the quadratic-modified Dahl-
Lea back-calculation method:
TL
i
= TL
c
× [(a + bLD
i
+ cLD
i
2
)/(a + bLD
c
+ cLD
c
2
)]
where a, b and c are the quadratic fit parameter esti-
mates, LD
i
are the lateral diameters at band i and LD
c
is the lateral diameter at capture; as well as the linear-
modified Dahl-Lea method:
TL
i
= TL
c
× [(a + bLD
i
)/(a + bLD
c
)]
where a and b are the linear fit parameter estimates.
A minimum of 1 vertebra and, if available, multiple
vertebrae, from different parts of the vertebral column
was removed for processing. Every fifth vertebra from
whole columns of 3 specimens was analyzed for
changes in morphology, size and band pair count.
X-radiography and sectioning were used to enhance
the visibility of band pairs.
X-radiography. X-radiographic settings varied be-
tween the laboratories due to different machines, film
and vertebral sizes. The diameter at the angle change
was marked on each image. Band pairs were counted
and measured twice by 1 reader (results hereinafter
referred to as X-radiography). Images were scanned
into the computer using a CreoScitex EverSmart Jazz
1
scanner with EverSmart Jazz Scanning Software ver-
sion 3. Band pairs (consisting of 1 opaque and 1 trans-
lucent band) were then counted and measured by
1 reader using digital image analysis software Image
Pro
®
Discovery (hereinafter referred to as digital
image analysis). X-radiographs were produced for
each vertebra along the column of Specimen 36.
Sectioning. Sectioning was only performed on verte-
brae from the western North Atlantic (Specimens 7 to
17, 37 to 40). One vertebra from each animal and every
fifth vertebra from 2 whole columns were processed.
The vertebrae were sectioned using either a Diamond
Pacific Model TC-6 trim saw with a diamond blade or a
standard hacksaw. Each centrum was sectioned
through the center at the point where the lateral diam-
eter measurement was taken. The resulting bow-tie
sections were directly photographed with a millimeter
scale using a Nikon Coolpix 5700 camera system or
with an MTI CCD 72 video camera attached to an
SZX9 Olympus stereomicroscope using reflected light.
Sections were preserved in 70% ethanol. Band pairs
were counted and measured on the images using
Image Pro 4 software. Measurements on the sections
were made along the same plane as would have been
measured on the vertebral face so that comparisons
could be made for both the angle change (birth band)
and lateral diameter.
Bias. Aging bias and precision were analyzed by
looking at percent agreement (± 1 band pair) and
contingency tables. Chi-squared tests of symmetry
(McNemar 1947, Bowker 1948, Hoenig et al. 1995,
Evans & Hoenig 1998) were performed to determine
whether differences between readers were biased or
due to random error (Cailliet & Goldman 2004). Addi-
tionally, bias graphs and coefficient of variation (CV)
were examined (Chang 1982, Campana et al. 1995).
RESULTS
Vertebrae were obtained from 40 specimens: 13 fe-
males (261 to 856
a
cm), 16 males (311 to 840 cm) and
11 specimens of unknown sex (376 to 853 cm)
(Table 1). Vertebrae from 37 basking sharks (261 to
856
a
cm) were X-rayed, and X-radiographs were
scanned for digital image analysis. Vertebrae from
12 of the 37 were also sectioned. Vertebrae from
3 other sharks were processed exclusively with sec-
tioning. Nineteen specimens either lacked morpholog-
ical measurements or were from either the extreme
cranial or the tail area, and so could not be used for
age-related analyses. The 21 samples used for age-
related analyses consisted of 9 females (261 to 716 cm),
11 males (311 to 840 cm) and 1 specimen of unknown
sex (480 cm; Table 1). Band counts from all specimens
were examined for comparison between processing
techniques and the differences in band counts along
the vertebral column.
Morphology of the vertebrae
Basking shark vertebrae are generally round but
vary slightly in shape and morphology depending on
their location along the vertebral column. The position
269
1
Reference to trade names does not imply endorsement by the
National Marine Fisheries Service
Mar Ecol Prog Ser 361: 267–278, 2008270
Table 1. Cetorhinus maximus. Data from samples used in this study including processing methods, number of vertebrae per sample, morphometrics and band pair counts.
Specimens 9 and 36 had weights of 262 and 48 kg, respectively. M: male, F: female, U: unknown; TL: total, FL: fork, PCL: pre-caudal lengths; nm: no measurement;
I: immature, Ma: mature; Med: Mediterranean Sea, North: North Sea
Speci- Sex Length Matu- No. Vertebra(e) Capture Storage Vertebral prepa- Band pair count
men (cm) rity of taken from location ration method of largest vertebra
TL FL PCL vertebrae X-radio- Section X-radio- Section
obtained graph graph
Used in final analysis
1 M 665 – – – 1 Before first Adriatic/Med Formol X – 31 –
dorsal
2 M 840 – – – 1 – Adriatic/Med Dried X – 31
d
–
7 F 458 420 – I 1 Above the gills NW Atlantic 70% ethanol, X X 14 13
then frozen
8 F 350
a
310 – – 1 Above the gills NW Atlantic 70% ethanol, X X 9 9
then frozen
9 M 362 320 – I 1 Above the gills NW Atlantic 70% ethanol, X X 10 10
then frozen
10 M 392 356 – I 1 Above the gills NW Atlantic 70% ethanol, X X 12 12
then frozen
12 M 452
a
411.5 – I 1 Above the gills NW Atlantic Frozen X X 14 14
13 M 770
a
695.5 – Ma 1 Above the gills NW Atlantic Frozen X X 33 30
14 F 575 533.4 – I 1 Above the gills NW Atlantic Frozen X X 19 15
15 F 686 615 – I 2 Between gills NW Atlantic Frozen X X 30 24
and dorsal
16 M 522 444.5 – I 2 Above the gills NW Atlantic Frozen – X 14 13
19 M 685 601 540 I
a
1 No. 20 SE Atlantic Formalin fixed, X – 21 –
monospondylous
50% isopropyl alcohol (?)
20 F 319 273 250 I 1 No. 17 SE Atlantic Formalin fixed, X – 9 –
monospondylous
50% isopropyl alcohol (?)
26 M 311 – – I 1 From caudal Skagerrak/North 80% ethanol X – 8 –
part of trunk
27 U 480 – – – 1 From trunk? Kattegatt/North 80% ethanol X – 12 –
28 M 345 – – – 1 Transverse cross Skagerrak/North Formalin fixed, X – 11 –
section of trunk vert. 80% ethanol
31 M 550 – – – 3 Big vert. Kattegatt/North Dried X – 18 –
36 F 261 228.8 205 I Column (56) Behind the head SW Indian Frozen X – 6 –
37 F 706 595 – – 2 Abdomen NW Atlantic Frozen X X 25 24
38 F 714
a
644 – Ma Column (85) Abdomen NW Atlantic Frozen – X – 27
40 F 457 403 – I Column (50) Abdomen NW Atlantic Frozen – X – 13
Not used in final analysis
3 U 376 – 302.3 – 2 Cervical (head NW Atlantic Dried X – 11 –
region)
4 U nm – – – 2 Too high NE Atlantic 75% methy- X – Unreadable –
lated spirit
5 U 853 – – – 2 Tail North 75% methy- X – 23 –
lated spirit
6 U nm – – – 2 – NE Atlantic Dried X – 26 –
11 F 856
a
– 700 – 1 Tail NW Atlantic 70% ethanol, X X 23 –
then frozen
Natanson et al.: Vertebral growth in the basking shark
of the basopophyses (following the ter-
minology of Walker 1975) provides only
a coarse indication of the origin of each
vertebra (Fig. 1). Cranial vertebrae
(Nos. 1 to ~10) can be distinguished by
basopophyses that are directed ventro-
laterally. Abdominal vertebrae (Nos. ~10
to ~30) are generally the largest and the
basopophyses are directed laterally. In
trunk vertebrae (Nos. ~30 to ~55), the
basopophyses start to come together
ventrally, while in tail vertebrae (Nos.
~55 to tail tip), they combine to form the
hemal arch (Fig. 1). Vertebrae from all
regions of the body contain distinct,
evenly spaced band pairs. Unless other-
wise noted, the following analyses use
only samples obtained from above the
gill area (n = 21).
Each of the 3 vertebral measurements
showed significant positive relationships
to TL (Fig. 2). The shape of the vertebrae
change as the shark grows, with the ver-
tebrae often becoming wider at the ven-
tral surface. It is due to this shape
change that we used multiple VD mea-
surements. Though we compared all the
VD measurements that we made to TL,
we used the LD measurement for the
back-calculation as the lateral growth is
symmetrical. The relationship between
LD and TL was slightly curvilinear. A
linear regression gave a significant fit to
the data (TL = [37.05 × LD] +162.70; r
2
=
0.933); however, a quadratic equation
produced a slightly better fit (TL = 49.06
+ [63.42 × LD] + [–1.26 × LD
2
]; r
2
= 0.943)
based on a lower mean square error
(1948.35 and 1750.46, respectively) and
a t-test that showed the third parameter
to be significantly different from 0 (t =
1.77; df = 21, p = 0.09) An ANCOVA
using ln-transformed data was used to
compare the LD–TL relationships be-
tween the sexes. As there was no signif-
icant difference between either the
slopes or intercepts (p = 0.30 and 0.42,
respectively), the data were combined.
It was still necessary to compare the
mean back-calculated length-at-age
from both the linear- and quadratic-
modified Dahl-Lea back-calculation
equations to determine if the better
statistical fit equated to a better bio-
logical fit.
271
Table 1 (continued)
Speci- Sex Length Matu- No. Vertebra(e) Capture Storage Vertebral prepa- Band pair count
men (cm) rity of taken from location ration method of largest vertebra
TL FL PCL vertebrae X-radio- Section X-radio- Section
obtained graph graph
17 U nm – – – 10 Dorsal NW Atlantic Frozen X X 22 15
18 M nm – – I 1 – Med Frozen X – 19 –
21 M 777 – – Ma 1 Tail NE Pacific Dried X – 20 –
22 F nm – – – 1 Midback region NE Pacific Dried X – Unreadable (44) –
23 F nm – – Ma 3 – NE Pacific 70% ethanol X – 20 –
24 M nm – – I 1 – Thyrrhenian/ Formalin/alcohol (?) X – 28 –
Med
25
c
M 545 492 439 I 1 Transverse cross Skagerrak/North Formalin fixed X – 12 –
section of tail vert.
29 M nm – – – 1 Just behind Kattegatt/North 80% ethanol X – 12 –
second dorsal fin
30 F 415 – – – 1 Beneath Kattegatt/North 80% ethanol X – 11 –
second dorsal tail
32 U nm – – – 1 – – Dried X – 16 –
33 U nm – – – 2 Trunk Skagerrak/North Formalin fixed, X – 13 –
80% ethanol
34 U nm – – – 1 Tail Kattegatt/North Dried X – 28 –
35 U nm – – – 1 – – Formalin fixed, X – 33 –
60% ethanol
39 U nm – – – 2 Abdomen NW Atlantic Frozen X X 25 26
a
Calculated TL
b
Almost mature
c
Jagerskold sample
d
This specimen was difficult to read using X-radiography by eye, thus, the digital image analysis count was used
Mar Ecol Prog Ser 361: 267–278, 2008
Whole columns
Examination of vertebrae from different positions
along the column of 3 specimens of various lengths
(261, 457 and 714 cm) showed that the vertebral
dimensions decline as vertebral number increases
(Fig. 3a). This trend is minor in small animals and dra-
matic in larger animals. Analyses of the VD–TL rela-
tionship were performed separately on vertebrae
obtained from 3 different regions: abdomen, trunk and
tail (as described in ‘Results; Morphology of the verte-
brae’). Comparison of the regressions of these 3 re-
gions indicated that there was a significant difference
in these relationships for all 3 VD measurements and
all 3 regions (ANOVA; p < 0.05); therefore, conversions
of VD to TL can only be used if the vertebrae are taken
from the same region of the body.
Band pair counts differed along the vertebral col-
umn, particularly in the largest individual. Counts
were low at the cranium, increased to a plateau and
272
Fig. 1. Cetorhinus maximus. Photographs of 5 vertebrae from
the vertebral column of Specimen 40 (total length, TL = 457
cm). Photographs are not to scale. Lateral diameters (LD, mm)
are shown for each vertebra. B = basapophysis. Terminology
follows Walker (1975). Number refers to position of the
vertebra along the vertebral column
0
100
200
300
400
500
600
700
800
900
0 2 4 6 8 10 12 14 16 18
Vertebral measurement (cm)
Total length (cm)
Dorsal diameter (DD)
Lateral diameter (LD)
Vertebral length (VL)
TL = –1.26 × LD
2
+ 63.42 × LD + 49.06
r
2
= 0.943, n = 21
Fig. 2. Cetorhinus maximus. Relationship between vertebral
dimensions and total length (TL). Quadratic fit is shown for
the lateral diameter relationship
0
20
40
60
80
100
120
140
160
180
a
b
0 10 20 30 40 50 60 70 80 90
0 10 20 30 40 50 60 70 80 90
Lateral diameter (mm)
Sample 40 (457 cm TL)
Sample 36 (261 cm TL)
Sample 38 (714 cm TL)
0
5
10
15
20
25
30
Vertebral number
Band pair count
Sample 40 (457 cm TL)
Sample 38 (714 cm TL)
Sample 36 (261 cm TL)
Fig. 3. Cetorhinus maximus. Relationship between (a) lateral
diameter or (b) band pair number and vertebral number for
3 specimens. Specimens 36 (261 cm) and 40 (457 cm) are head
to pre-caudal pit, while Specimen 38 (714 cm) starts just
behind the head (estimated at vertebra no. 5) and extends to
the tip of the tail. TL: total length
Natanson et al.: Vertebral growth in the basking shark
then decreased toward the caudal region (Fig. 3b).
Specimen 38 (714 cm) was missing the first few verte-
brae of the column, thus the extreme cranial vertebral
counts are missing. However, it is clear that the counts
along this column follow the described pattern, as the
first vertebra counted (~ No. 5) was lower than the sub-
sequent 2 (~ Nos. 10 and 15), followed by a steep
decline in band pair count to the caudal region. Band
pair count along the column changes ontogenetically.
The band pair counts of vertebrae of a small basking
shark (261 cm) (from plateau to pre-caudal pit) never
differed by more than 1. However, counts from the
457 cm (from plateau to pre-caudal pit) and 714 cm
(from plateau to tip of tail) specimens differed by up to
3 and 20 band pairs, respectively. Additionally, com-
parisons of counts between cranial and caudal verte-
brae from the same specimen of 8 additional sharks
showed clear differences in band pair counts in the
larger specimens (Table 2).
Angle change
The diameter of the angle change differed depend-
ing on the location of the vertebra along the vertebral
column. Due to the subjectivity of the location of the
angle change on the face of the vertebrae, the angle
change measurement between different researchers
was not consistent. The angle change is generally pre-
sumed to represent the birth band (Casey et al. 1985)
and those bands deposited prior to the angle change
pre-birth bands; although we do not have validation,
we will use this terminology. Average birth VD and
numbers of pre-birth bands were derived from the
vertebral sections only.
Using the sections only (n = 11), the number of pre-
birth bands ranged from 6 to 10 (mean ± 95% CI = 8.4 ±
0.81). The average angle change diameter occurred at
5.26 cm (95% CI, ±0.47 cm). Mean length at birth cal-
culated using the quadratic-modified Dahl-Lea back-
calculation method ranged from 87 to 321 cm (mean ±
95% CI = 196.8 ± 57.2 cm), while that calculated using
the linear-modified Dahl-Lea back-calculation method
ranged from 243 to 376 cm (mean ± 95% CI = 301.2 ±
30.9 cm).
Bias and precision of band pair counts
Band pairs were visible in all vertebrae using all
techniques (Fig. 4). In some cases, dried vertebrae
were distorted and interpretation was difficult. In addi-
tion, band pairs at the centrum edges of large verte-
brae were difficult to discern using X-radiography.
Two vertebrae were considered unreadable using
273
Table 2. Cetorhinus maximus. Data on samples with >1 verte-
bra from an individual vertebral column. Number in paren-
theses is approximate number along the column, if known.
Large, medium and small are relative to available vertebrae
from the sample and not to area along the column. TL: total
length, na: not available
Specimen TL (cm) Large Medium Small
3 375.9 10 (2) na 11 (35)
5 853.4 26 (40) na 20 (45)
15 686 30 (10) na 22 (45)
31 550 19 (40) 18 (50) 8 (60)
4na27na15
6na27na15
23 na 13 13 13
33 na 12 na 11
Fig. 4. Cetorhinus maximus. Comparison of X-radiographed
and sectioned vertebrae from Specimen 13. A split band is
noted on the edge of the sectioned vertebrae; this would show
on an X-radiograph as 2 band pairs (see ‘Results; Bias and
precision of band pair counts’)
Mar Ecol Prog Ser 361: 267–278, 2008
X-radiography (Table 1). Sections provided very dis-
tinct band pairs on fish of all lengths.
Comparison of repeated counts using X-radiography
indicated no systematic bias between counts 1 and 2 of
1 reader (Fig. 5). The individual CVs between counts
fluctuated around the mean at 4.0. Percent agreement
to within 1 band pair occurred in 80% of the samples
(56 of 70). Additionally, the McNemar (1947), Bowker
(1948) and Evans & Hoenig (1998) chi-squared tests of
symmetry gave no indication that differences between
counts 1 and 2 were systematic rather than due to ran-
dom error (χ
2
test, p > 0.05). This level of precision was
considered acceptable, and the second count was used
in the growth curve analyses. Though the precision
(repeatability) of the counts was high, the accuracy
(reflection of the true age) was not.
Comparison of counts using vertebral sections
yielded a slightly lower CV (2.5%). While the low sam-
ple size limits the value of the analyses, the bias graphs
indicated no bias between counts. Percent agreement
to within 1 band pair occurred in 83% of the samples
(10 of 12). Additionally, the McNemar (1947), Bowker
(1948) and Evans & Hoenig (1998) chi-squared tests of
symmetry gave no indication that differences between
counts 1 and 2 were systematic rather than due to ran-
dom error (χ
2
test, p > 0.05). The sectioned counts were
then compared to those obtained using digital image
analysis and X-radiography.
While good agreement was found between counts
obtained by a single method, counts differed depend-
ing on the method used. Percent agreement to within
1 band pair between digital image analysis and X-radi-
ography occurred in 50% of the samples (21 of 42).
However, 21% (9) of readings disagreed by 4 or more
band pairs. The Bowker (1948) and Evans & Hoenig
(1998) chi-squared tests of symmetry gave no indica-
tion that differences between these methods were sys-
tematic rather than due to random error (χ
2
test, p >
0.05). In contrast, the McNemar (1947) test did indicate
that there was a systematic error (χ
2
test, p < 0.05).
Graphical comparison of these counting methods sup-
ported the McNemar (1947) results, as it was apparent
that counts from digital image analysis were higher
than those obtained from X-radiography alone. The
tests of symmetry using sectioning related to digital
image analysis (n = 8) gave no indication that differ-
ences between these methods were systematic rather
than due to random error (χ
2
test, p > 0.05). Percent
agreement to within 1 band pair between digital image
analysis and sectioning occurred in 50% of the sam-
ples (4 of 8). However, when sectioning was compared
to X-radiography (n = 12), the Evans & Hoenig (1998)
result indicated a significant difference (χ
2
test, p <
0.05), though the McNemar (1947) and Bowker (1948)
tests did not. Percent agreement to within 1 band pair
between sectioning and X-radiography occurred in
58% of the samples (7 of 12). Both methods that rely on
X-radiography produced higher counts than section-
ing, based on graphical comparison, though the small
sample sizes limit the usefulness of these analyses
(Fig. 6).
Sectioning produced high quality images on all sizes
of samples. Many splits in the opaque band were evi-
dent in sections of large vertebrae. These splits oc-
curred at the edge of the corpus calcareum at what
would be the face of the vertebra (Fig. 4). In an X-radi-
ograph of the vertebral face, these splits would appear
274
–20
–10
0
10
20
30
40
50
60
70
80
51015
20
25
30
35
Number of band pairs, second reading
Mean number of band pairs,
first reading
4
1
3
4
2
5
6
8
4
3
1
12
4
2
41
3
2
2
2
2
1
1
1
1
x = y
Fig. 5. Cetorhinus maximus. Pairwise comparison of vertebral
counts from 2 replicate band counts from X-radiographed
specimens. Each error bar represents the 95% CI for the
mean count from Reading 1 related to all fish with a given
count in Reading 2. Sample number is presented above each
error bar. The 1 to 1 equivalence line is also presented
70
60
50
40
30
20
10
0
–10
–20
5
10
15
20
25 30 35
Band pair count based
on sectioned samples
Mean number of band pair
count, by method
X-radiography n = 12
Digital image analysis n = 8
2
2
2
x = y
Fig. 6. Cetorhinus maximus. Pairwise comparison of vertebral
counts from X-radiography and digital image analysis. Each
error bar represents the 95% CI for the mean count from
reads based on X-radiography and digital image analysis to
all fish of a given count in the reading based on sections.
n = 1 except where noted. The 1 to 1 equivalence line is also
presented
Natanson et al.: Vertebral growth in the basking shark
as distinct bands and thus lead to over-counting. Addi-
tionally, X-radiographs of a whole vertebra penetrate
through both faces of the vertebral cone. Thus, shad-
ows of the band pairs on the backside will appear to be
on the face that is being counted, which can lead to
over-counting. Due to the inability of the software to
distinguish between a shadow and a band, the digital
image analysis method appeared to consistently over-
estimate the band pair count. The human reader would
also tend to overestimate counts in the larger fish when
using X-radiographs due to the difficulty of ‘reading’
the edges of the large specimens and the splits that
would not be distinguished in the X-radiograph. The
lower CV from the sectioning method also indicates
that this method is more repeatable. Unfortunately,
due to the difficulty in obtaining Cetorhinus maximus
vertebrae to section, we had to primarily use the
X-radiography method in our analyses. Band pair
counts for 21 sharks that were used in the final analy-
sis included 11 that were based on sections and 10 ba-
sed on X-radiographs. Only 4 of the 10 with X-radi-
ographs had band pair counts >17. Since this was the
point where counts obtained from the 2 processing
methods started to diverge (with the X-radiograph
counts slightly higher), only the 4 samples with
>17 band pairs were likely to be affected by processing
differences, thus these are highlighted in the figures.
Band pair counts and total length
For both sexes, the number of band pairs increased
with increasing length (Fig. 7a). The maximum num-
ber of band pairs for males and females were 30 and 27
(Specimens 13 and 38), respectively, for those where
vertebral location in the spinal column was known.
The highest number of band pairs in any sample was
33 using X-radiography (Specimens 13 and 35) and 47
from digital image analysis (Specimen 35). The small-
est specimen (No. 36) had 6 band pairs.
DISCUSSION
It is evident upon examination of published literature
on basking shark vertebral morphology and age and
growth that authors have known that the basking
shark vertebrae are irregular compared to other spe-
cies of elasmobranch. Hasse (1882) and Ridewood
(1921) highlighted 2 points about the vertebral centra
of the basking shark that are critical: (1) that ‘...in pass-
ing back from the root of the caudal fin the concentric
lamellae dwindle...’ (Ridewood 1921) and (2) that the
basking shark vertebral centra are different from other
lamnoids such that the adult hind caudal vertebra are
more recognizable as lamnoid in character than the
trunk and anterior caudal vertebrae of the young.
Ridewood (1921) also noted that the band pair counts
change along the vertebral column. The difference in
appearance of basking shark vertebrae from the white
shark Carcharodon carcharias, shortfin mako Isurus
oxyrinchus, salmon shark Lamna ditropis, porbeagle
Lamna nasus or thresher Alopias vulpinus is clear and
verifies the differences Hasse (1882) and Ridewood
(1921) mentioned. Additionally, in the porbeagle,
shortfin mako and thresher, all vertebrae along the col-
umn have the same number of band pairs (Natanson et
al. 2002, 2006, B. Gervelis [NMFS] unpubl. data). Even
while proposing the biannual band pair deposition
hypothesis, Parker & Stott (1965) cast doubt on it for
the same reasoning we provide: ‘the irregularities that
occur in a single vertebra, the reduced number of rings
275
900
800
700
600
500
400
300
200
100
0
0 5 10 15 20 25 30 35 40 45
0 5 10 15 20 25 30 35 40 45
Maturity 701 cm TL
(Matthews 1950)
Maturity 701 cm TL
(Matthews 1950)
1100
1000
900
800
700
600
500
400
300
200
100
0
Band pair count
Total length (cm)
21 good samples
Lein & Aldrich (1982)
Maturity
Other studies
Springer &
Gilbert (1976) tail
a
b
Fig. 7. Cetorhinus maximus. (a) Growth data based on verte-
bral band counts from 21 specimens. Open diamonds are
samples processed using X-radiography and having counts
>17. (b) Growth data including samples from the literature
and this analysis
Mar Ecol Prog Ser 361: 267–278, 2008
in the caudal region and the apparent existence of
seven rings at birth are features that do not obviously
harmonise with the idea of annual increases of two
rings.’ At the time, Parker & Stott (1965) were unable to
provide an alternate hypothesis.
While several of the previous studies that looked at
band pairs in the basking shark noted the difference
in band pair count along the vertebral column (Ride-
wood 1921, Parker & Stott 1965, Desse & Du Buit
1971), others have not found a difference along the
column (Jägerskiöld 1915, Izawa & Shibata 1993). The
disparity between these studies is in the length and
condition of the specimens examined. Izawa & Shi-
bata (1993) examined a 260 cm specimen which had 8
band pairs. In the present study we also examined a
specimen (no. 36) of 260 cm TL, which had 6 to 7
band pairs. The difference in counts between these
same-sized specimens could be due to individual vari-
ation or the fact that the specimen from the present
study was emaciated and in very poor condition. If
band pair deposition is related to girth and body sup-
port, the decreased number of band pairs in our spec-
imen may be due to its being underweight and thus
not having the girth to require support. This may also
explain the low number of band pairs and lack of dif-
ference along the length of the vertebral column in
the Jägerskiöld (1915) specimen. Photographs pre-
sented in that publication show a thin, underweight
specimen. This specimen would be expected to show
some band pair count variation along the vertebral
column based on our observations along the column
of a 550 cm specimen which had band pair counts
ranging from 8 to 18 for small and large vertebrae,
respectively. Jägerskiöld’s (1915) specimen had fewer
bands than would be expected on the largest verte-
brae based on length (our 457 cm specimen had 12
band pairs and our 550 cm specimen had 18). The
count is not in question as our Sample 25 is Jäger-
skiöld’s (1915) sample and we obtained the same
count (12). While the lower band pair count could be
due to individual variation, combined with the lack of
variation along the column, this suggests that this
specimen is anomalous. Ridewood (1921) and Desse &
Du Buit (1971), examined specimens of various stages
(an adult, a young individual and an individual of
15 m TL), and did find a difference in band pair count
in vertebrae taken from along the vertebral column.
Parker & Stott (1965) noted Ridewood’s (1921) find-
ings of changes in counts along the column and
though their specimens widely ranged in length
(475 to 877 cm), they sampled from 1 area along the
column to maintain consistency. Since none of these
previous authors examined whole columns from a
range of specimen lengths, they did not take note of
the ontogenetic changes in band pair number.
Vertebral growth zone deposition in the basking
shark is similar to that of the angel shark Squatina cal-
ifornica (Natanson & Cailliet 1990). The angel shark,
the gummy shark Mustelus antarcticus and the school
shark Galeorhinus galeus have all been found to have
a change in band pair count along the vertebral col-
umn (Ridewood 1921, Natanson & Cailliet 1990, Offi-
cer et al. 1996). In both the angel shark and the bask-
ing shark, the band pair counts along the vertebral
column are similar in the young, and progressively
become variable along the column so that in adults,
counts are low in the cranial vertebrae, increase to a
plateau at the anterior abdominal area and decrease
toward the tail (Natanson & Cailliet 1990; their Fig. 2).
Additionally, Natanson & Cailliet (1990) documented
that angel sharks are born with ∼7 band pairs. Parker &
Stott (1965) first suggested that basking sharks had
7 or possibly 8 band pairs at birth, which corresponds
to results from the present study indicating an average
of 8 pre-birth band pairs. Most elasmobranch age
studies do not indicate if band pair counts along the
vertebral column were examined; however, this is an
important step in determining the usefulness of the
vertebrae as an aging structure. Consistency in loca-
tion of vertebral sampling is important not only for cal-
culating the relationship between vertebral dimension
and body length (Natanson et al. 2006, Piercy et al.
2006) but also for comparison of band pair count.
Though many studies have shown that band pair
counts are consistent along the vertebral column
(Natanson et al. 2002, 2006, Joung et al. 2004, Knee-
bone 2005, Bishop et al. 2006, Piercy et al. 2006,
B. Gervelis [NMFS] unpubl. data), the results of those
studies that do show a difference highlight the need to
examine band pair counts along the column of every
species aged.
Angel sharks deposit band pairs relative to growth
rather than time (Natanson & Cailliet 1990) and the
similarities in vertebral growth between the angel
shark and the basking shark suggest that vertebral
growth in basking shark might also be related to
somatic growth; however, age validation for the bask-
ing shark has not been completed. The consistency of
the growth bands suggests that they might be related
to a structural component of the vertebral column and
the increase in structural bands in the thickest part of
the shark also suggests a strengthening component.
Band pair deposition in the angel shark appears to be
related to growth in girth and therefore to provide
physical support for the growing vertebrae in the
widest part of the animal (Natanson & Cailliet 1990).
Unfortunately, girth data were not available for most of
the samples in the present study for comparison.
Counting band pairs in the basking shark is not diffi-
cult regardless of preparation method. There is a high
276
Natanson et al.: Vertebral growth in the basking shark
level of repeatability of counts within each method.
Results between methods, however, indicate that use
of X-radiography overestimates band pair number in
specimens with >17 band pairs. Most of the previous
studies on the basking shark appear to use either sec-
tions or whole vertebrae, and thus band pair counts
between studies are comparable if the vertebrae are
taken from the same location of the vertebral column.
Often this is not the case, which explains the discrep-
ancy between Springer & Gilbert’s (1976) Sarasota
basking shark ‘age’ (16) and Eastern Atlantic speci-
mens of similar length (22). Springer & Gilbert (1976)
were comparing the counts from a pelvic vertebra to
counts from a pre-caudal vertebra. A count difference
of >7 band pairs could be obtained from these 2 areas
of the vertebral column in a shark this length (based
on counts from a similar length shark; Specimen 38;
Table 2).
Overall, band pair counts from the present study
seem to fit those reported in the literature, although
there is count variation in specimens of similar lengths.
This difference could be due to the unknown location
of some of these vertebrae along the respective verte-
bral columns or to individual variation in somatic
growth (Fig. 7b). Lein & Aldrich (1982) presented a
maximum and a minimum count for their samples.
They noted the change in counts along the column and
the maximum count came from the largest vertebra
from a specimen (J. Lein pers. comm.). Additionally,
they were counting on the face of fresh (unpreserved)
unprepared vertebrae (J. Lein pers. comm.). Due to the
similarity in location of these vertebrae to our 21 sam-
ples of known location, we can compare the counts
between the studies and thus enhance the sample size.
Changes in band pair width led to Parker & Stott’s
(1965) conclusion that 7 band pairs were deposited
prior to birth. Because they assumed no difference in
pre-gestational and post-gestational deposition, they
estimated a 3.5 yr gestation. Current information on
the pre-birth bands for the angel shark gives us more
insight as to possibilities for the basking shark. The
angel shark has 7 pre-birth band pairs yet has only a
10 mo gestation period. The 7 pre-birth bands are pre-
sumably what is needed mechanically to support the
neonate at the birth length. The change in band spac-
ing between pre- and postnatal basking shark vertebra
noted by Parker & Stott (1965) is possibly related to the
change in growth rate between the rapidly growing
embryo and the slower growing neonate. Our finding
of 8 pre-birth bands corresponds to an average length
at birth of 197 cm TL using the quadratic-modified and
301 cm TL using the linear-modified Dahl-Lea back-
calculation equations. Currently, the birth length of
this species is not well documented. Matthews (1950)
suggested a birth length of 6 feet (183 cm TL) based on
a lack of observations of smaller individuals. Springer
& Gilbert (1976) stated that no free-swimming young
<200 cm have been recorded; however, the smallest
free-swimming specimen is 180 cm (Lein & Fawcett
1986). Historically, Templeman (1963) reported on a
200 cm postnatal specimen from 1833. These appear to
be the smallest reliably reported specimens on record.
The few measured small free-swimming specimens
are similar in length to those mentioned in the anecdo-
tal reports and suggest a reasonable estimated birth
length for Cetorhinus maximus of approximately 180
to 200 cm, which is similar to our quadratic-modified
Dahl-Lea estimate (197 cm TL). This suggests that the
quadratic-estimated back-calculation is, both statisti-
cally and biologically, the preferred method for esti-
mating age at previous length in this species.
The current evidence suggests that the basking
shark may deposit band pairs relative to structural
morphology rather than time. Regardless, it is evident
that while the vertebrae from the basking shark grow
proportionally to the body length and have distinct
band pairs, the band pairs may not be formed relative
to an increment of time, and thus the vertebrae from
this species may not be useful for determining age.
Ongoing studies into the validation of the band pairs
should provide a more definitive picture of the verte-
bral deposition in this species.
Acknowledgements. The assistance of the following people is
gratefully acknowledged: D. C. Bernvi, B. C. Delius, P. Dey-
nat, G. Nilson, C. Stenberg and 2 anonymous persons who
assisted in organizing X-radiography in a private hospital in
Algiers. We thank B. Runsten and J. Fong for providing
X-radiographs for this study. We thank S. F. J. Dudley for
comments on an earlier draft of the manuscript. W. Ledwell
and J. Lein provided invaluable information. K. Goldman con-
tributed much needed back-calculation expertise. G. Cailliet
and G. Skomal assisted by reviewing early versions of the
manuscript and acting as sounding boards for us in this study.
LITERATURE CITED
Bigelow HB, Schroeder WC (1948) Fishes of the western North
Atlantic. 1, Sharks. Mem Sears Found Mar Res 1:59–576
Bishop SDH, Francis MP, Duffy C, Montgomery JC (2006)
Age, growth, maturity, longevity and natural mortality of
the shortfin mako (Isurus oxyrinchus) in New Zealand
waters. Mar Freshw Res 57:143–154
Bowker AH (1948) A test for symmetry in contingency tables.
J Am Stat Assoc 43:572–574
Cadenat J, Blache J (1981) Requins de Mediterranée et d’At-
lantique, plus particulièrement de la Côte Occidentale
d’Afrique. Faune Trop ORSTOM 21:149–152
Cailliet GM, Goldman KJ (2004) Age determination and vali-
dation in chondrichthyan fishes. In: Carrier J, Musick JA,
Heithaus MR (eds) Biology of sharks and their relatives.
CRC Press, Boca Raton, FL, p 399–447
Campana SE, Annand MC, McMillan JI (1995) Graphical and
statistical methods for determining the consistency of age
determinations. Trans Am Fish Soc 124:131–138
277
➤➤
➤
Mar Ecol Prog Ser 361: 267–278, 2008
Casey JG, Pratt HL Jr, Stillwell CE (1985) Age and growth of
the sandbar shark (Carcharhinus plumbeus) from the
western North Atlantic. Can J Fish Aquat Sci 42:963–975
Castro JI, Woodley CM, Brudek RL (1999) A preliminary eval-
uation of the status of shark species. FAO Fish Tech Pap
380:1–72
Chang WYB (1982) A statistical method for evaluating the
reproducibility of age determination. Can J Fish Aquat Sci
39:1208–1210
Chen JTF (1963) A review of the sharks of Taiwan. Biol Bull
Dep Biol Tunghai Univ (Ichthyol Ser 1) 19:1–102
Desse G, Du Buit MH (1971) Diagnostic des pièces rachidi-
ennes des Téléostéens et des Chondrichthyens. II Chon-
drichthyens. Expansion Scientifique Française, Paris
Evans GT, Hoenig JM (1998) Testing and viewing symmetry
in contingency tables, with application to readers of fish
ages. Biometrics 54:620–629
Goldman KJ (2004) Age and growth of elasmobranch fishes.
In: Musick JA, Bonfil R (eds) Elasmobranch fisheries man-
agement techniques. FAO Fish Tech Pap 6:97–132
Hasse JCF (1882) 1879–85 Das natuerliche System der
Elasmobranchier auf Grundlage des Baues und der
Entwicklung ihrer Wirbelsaeule. Eine morphologische
und palaeontologische Studie. Gustav Fischer, Jena, Ger-
many
Hoenig JM, Morgan MJ, Brown CA (1995) Analyzing differ-
ences between two age determination methods by tests of
symmetry. Can J Fish Aquat Sci 52:364–368
Howes CA (1998) Basking sharks Cetorhinus maximus (Gun-
nerus) stranded on the Yorkshire coast. Naturalist 123:
101–102
Izawa K, Shibata T (1993) A young basking shark, Cetorhinus
maximus, from Japan. Jpn J Ichthyol 40:237–245
Jägerskiöld LA (1915) On a basking shark — Cetorhinus maxi-
mus (Gunnerus) — found in Sweden. Göteborgs Kongl
Vetensk Vitterh Samh Handl 16:3–13
Joung SJ, Liao YY, Chen CT (2004) Age and growth of the
sandbar shark, Carcharhinus plumbeus, in northeastern
Taiwan waters. Fish Res 70:83–96
Kneebone J (2005) Validated age and growth estimates for the
tiger shark Galeocerdo cuvier, in the western North
Atlantic. Ms thesis, University of New Hampshire, Durham
Lein J, Aldrich D (1982) The basking shark (Cetorhinus maxi-
mus) in Newfoundland. Department of Fisheries, Govern-
ment of Newfoundland and Labrador, St. John’s
Lein J, Fawcett L (1986) Distribution of basking sharks,
Cetorhinus maximus, incidentally caught in inshore fish-
ing gear in Newfoundland. Can Field Nat 100:246–252
Matthews LH (1950) Reproduction in the basking shark,
Cetorhinus maximus (Gunner). Philos Trans R Soc Lond A
234:247–316
McNemar Q (1947) Note on the sampling error of the differ-
ence between correlated proportions or percentages.
Psychometrika 12:153–157
Natanson LJ, Cailliet GM (1990) Vertebral growth zone depo-
sition in Pacific angel sharks. Copeia 1990:1133–1145
Natanson LJ, Mello JJ, Campana SE (2002) Validated age and
growth of the porbeagle shark (Lamna nasus) in the west-
ern North Atlantic Ocean. Fish Bull (Wash DC) 100:
266–278
Natanson LJ, Ardizzone D, Cailliet GM, Wintner S, Mollet H
(2006) Validated age and growth estimates for the shortfin
mako, Isurus oxyrinchus, in the North Atlantic Ocean. In:
Goldman KJ, Carlson JK (eds) Proc Am Elasmobranch Soc
21st Mtg, July 2005, Tampa, FL. Environ Biol Fish 77:
367–383
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 determi-
nation. Can J Fish Aquat Sci 53:1765–1777
Parker HW, Boeseman M (1954) The basking shark, Cetorhi-
nus maximus, in winter. Proc Zool Soc Lond 124:185–194
Parker HW, Stott FC (1965) Age, size and vertebral calcifica-
tion of the basking shark, Cetorhinus maximus (Gun-
nerus). Zool Meded 40:305–319
Pauly D (1978) A critique of some literature data on the
growth, reproduction and mortality of the lamnid shark
Cetorhinus maximus (Gunnerus). ICES CM 17:1–10
Pauly D (2002) Growth and mortality of the basking shark
Cetorhinus maximus and their implications for manage-
ment of whale sharks Rhincodon typus. In: Fowler SL,
Reed TM, Dipper FA (eds) Elasmobranch biodiversity,
conservation and management. IUCN, Gland, p 199–208
Piercy AN, Ford TS, Levy LM, Snelson FF Jr (2006) Analysis
of variability in vertebral morphology and growth ring
counts in two Carcharhinid sharks. Environ Biol Fish
77:401–406
Ridewood WG (1921) On the calcification of the vertebral
centra in sharks and rays. Philos Trans R Soc Lond B
210:311–407
Schwartz FJ (2002) Basking sharks, Cetorhinus maximus,
family Cetorhinidae, recorded in North Carolina waters
1901–2002. J NC Acad Sci 118:201–205
Serena F, Vacchi M, Notarbartolo di Sciara G (2000) Geo-
graphical distribution and biological information on
the basking shark, Cetorhinus maximus in the Tyrrhen-
ian and Ligurian seas. In: Séret B, Sire JY (eds) Proc
Eur Elasmobranch Assoc 3rd Mtg, Boulogne-sur-Mer,
France, May 1999. Société Française d’Ichtyologie and
Institut de Recherche pour le Développement (IRD),
Paris, p 47–56
Siccardi EM (1961) ‘Cetorhinus’ en el Atlantico Sur (Elas-
mobranchii: Cetorhinidae). Rev Mus Argent Cienc Nat
Bernadino Rivadavia Inst Nac Investig Cienc Nat Zool 6:
61–101
Soldo A, Peharda M, Onofri V, Glavic N, Tutman P (1999)
New record and some morphological data of the basking
shark, Cetorhinus maximus (Gunnerus, 1765), in the east-
ern Adriatic. Ann Ser Hist Nat 9:229–232
Springer S, Gilbert PW (1976) The basking shark, Cetorhinus
maximus, from Florida and California, with comments on
its biology and systematics. Copeia 1976:47–54
Templeman W (1963) Distribution of sharks in the Canadian
Atlantic (with special reference to Newfoundland waters).
Bull Fish Res Board Can 140:1–77
Walker WF (1975) Vertebrate dissection, 5th edn. WB Saun-
ders, Philadelphia
Wintner SP (2000) Preliminary study of vertebral growth rings
in the whale shark, Rhincodon typus, from the east coast of
South Africa. Environ Biol Fishes 59:441–451
Wood FG Jr (1957) Southern extension of the known range of
the basking shark, Cetorhinus maximus (Gunnerus).
Copeia 1957:153–154
Zuffa M, Soldo A, Storai T (2001) Preliminary observations on
abnormal abundance of Cetorhinus maximus (Gunnerus,
1765) in the central and northern Adriatic Sea. Ann Ser
Hist Nat 11:185–192
278
Editorial responsibility: Kenneth Sherman,
Narragansett, Rhode Island, USA
Submitted: July 26, 2007; Accepted: January 9, 2008
Proofs received from author(s): May 21, 2008
➤
➤
➤
➤
➤
➤
➤
➤
➤
➤
