foxandhound

C1adistir.s ( 1992)
8:295-318
THE PHYLOGENETIC RELATIONSHIPS AMONG
REQUIEMAND HAMMERHEAD SHARK% INFERRING
PHYLOGENY WHEN THOUSANDS OF EQUALLY
MOST PARSIMONIOUS TREES RESULT
Gavin J. P. Naylor’
’ Department of Biology University of Michigan,
Ann Arbor, Michigan 48109-1048,
U.S.A.
Rectivedfwpublication 20 August 1992; accepted 6 October1 992
Abstmct-Protein
variation among 37 species of carcharhiniform
sharks was examined
at 17 presumed loci. Evohttionary
trees were inferred from these data using both cladistic character and a
distance Wagner analysis. Initial cladistic character
analysis resulted in more than 30 000 equally
parsimonious
tree arrangements.
Randomization
tests designed
to evaluate the phylogenetic
information
content of the data suggest the data are highly significantly
different from random
in spite of the large number of parsimonious
trees produced.
Different starting seed trees were
found to influence
the kind of tree topologies
discovered
by the heuristic
branch swapping
algorithm
used. The trees generated
during the early phases of branch swapping on a single seed
tree were found to be topologically
similar to those generated
throughout
the course of branch
swapping. Successive weighting increased
the frequency
and the consistency with which certain
clades were found during the course of branch swapping, causing the semi-strict consensus to be
more resolved. Successive weighting also appeared
resilient to the bias associated with the choice
of initial seed tree causing analyses seeded with different
trees to converge on identical final
character weights and the same semi-strict consensus tree.
The summary cladistic character
analysis and the distance Wagner analysis both support the
monophyly
of two major clades, the genus Rhizopionodon and the genus Sphyma.. The distance
Wagner analysis also supports the monophyly
of the genus Gzrcharhinus. However, the cladistic
analysis suggests that Grrcharhinus is a paraphyletic
group that includes the blue shark Rionnc~
+Lcn.
Introduction
Carcharhiniform
sharks comprise about 55% of the approximately 350 living
shark species. The Carcharhiniformes
are currently divided into eight families:
The cat sharks (Sciliorhinidae),
the finback catsharks (Proscillidae), the false catsharks (Pseudotriakidae),
the barbelled houndsharks (Leptocharidae),
houndsharks (Triakidae) , the weasel and snaggletoothed
sharks (Hemigalidae) , the
requiem sharks (Carcharhinidae)
and the hammerhead
sharks (Sphyrnidae)
(Compagno, 1988). Of these eight families, the requiem sharks and the hammerhead sharks are generally better known due to the large size of several of their constituent species, their occasional implication in shark attacks and their increasing
commercial importance to fisheries.
Cawharhinus contains more species than any other genus within the family
Carcharhinidae.
It comprises 30 species whose collective ranges spread over
tropical, sub-tropical
and temperate
regions
(Garrick,
1982; Bigelow and
Schroeder, 1948; Compagno, 1984). Many species within the genus are phenotypically similar and are often taxonomically confused. The confusion is frequently
compounded by extensive intraspecific variation among geographically isolated
populations. As a result, a number of synonyms have been used to describe species
Oi$N-3007~y”/040?9.5+24$08.00/0
0 1992 The Willi Hennig
Sorietv
within
the genus.
Recent
revisions
by Garrick
(out of more than 100) of these nominal
The marked
most species
heterogeneity
1985),
within species
indicate
and morphological
of Cu~&~rhi~~us has allowed intra-generic
from morphological
1982),
(1982,
that only 30
species are valid.
data for only the most obvious
uniftirmity
relationships
sibling
among
to be proposed
species
pairs (Garrick,
i.e. C. v&-C.
dussumni, C. ~PUCU--C.nmboinunsis, C. limhntus-C:. nmh&v-hynchoidps and C. ntnhlyrh?rn~hos-C1 whdk.
The establishment
of phylogenetic
relationships beyond
Baranes
these pairs had not been rigorously
and Shahrabany-Baranes,
metric
data)
(Lavery,
mltil
1992).
genetic
Lavery surveyed
The
present
relationships
within
taxa using allozyme
included
attempted
(with the exception
1986, who used phenetic
allozyme
variation
study represents
the genus
data derived
clustering
in 21 carcharhiniform
an effort
to investigate
Cmrharhiwus, and among
from 37 (:archarhiniform
of
of morphotaxa
the phylo-
its closely related
taxa, 1 1 of which were
in the survey by Lavery (1992).
Materials and Methods
Sharks were collected
and
December
America,
1987
Belize,
Philippines
in the field and from fish markets
from
south
the Bahamas,
and
Trinidad,
the northwestern
taken by me with the exception
Suiso of the Sea Grant
of the Virginia
possible
different
carcharhiniform
Localities
and samples
Samples
of heart,
shark in the field
stored
at -75°C.
equal volume
for
Islands.
of Marine
of Curcharhinus
nitrogen
and skeletal
2 grams
until
Approximately
Supernatants
which were collected
were
frozen
the
were
by Mark
C. altimus collected
by
A total of 21 (out of a
16 species
from
eight
were
other
surveyed.
1.
muscle
were removed
placed
from
in separate
to the laboratory
where
each
cryotubes
they were
half of each tissue sample was homogenized
water and centrifuged
1984
of North
of specimens
to Carcharhinus,
of each),
transported
February
coasts
in the Red Sea,
Samples
Science.
and
related
sizes are given in Appendix
liver, kidney
Gulf
Hawaii, and one
that are closely
of distilled
10 minutes.
Institute
between
and
the gulf of Aqaba
in Oahu
(approximately
and kept in liquid
east
of two C. gdupagensis
species
genera
the
Hawaiian
Program
Dr J. Musick
30)
Australia,
in an
at 0°C and 12000 g (10 000 rpm)
at -75°C
until
required
for electro-
phoresis.
Supernatant
Connaught
fractions
Starch
seven different
sumed
were run in horizontal
and 7% sucrose,
buffers
as outlined
loci were initially screened,
scored.
A list of optimum
presented
in Table
1. The
buffers
relative
by weight.
by Selander
starch
gels made
up with 12.5%
Gels were prepared
et al. (1971).
More
using one of
than
of which 17 showed activity clearly enough
and
tissues
mobility
directly from primary gels or from secondary
Richardson
et al. (1986) when electromorphs
used for the
of electromorphs
“line up”
were seen
17 presumed
was scored
35 preto be
loci is
either
comparison
gels sense
to have closely similar
mobilities.
PHYL~(;ENUK ANAIXIS
Data collected in the present study were subjected to both cladistic character and
distance Wagner analysis (Farris, 1972) using Cavalli-Sforza
and Edwards (1967)
PHYLOGENYOF
Presumed
2%
REQUIEM AND HAMMERHEAD SHARKS
Table 1
tissues used and electrophoretic
loci scored,
Protein
Locus
acronym”
Aspartate amino transferase
Aspartate amino transferase
Carbonate
dehydratase
Dipeptidasel
Esterase
Esterase
Isocitrate dehydrogenase
Glutamate
dehydrogenase
Glycerol-Sphosphate
dehydrogenase
Leucine amino peptidase
Lactate dehydrogenase
Lactoyl glutathione
lyase
Malic enzyme
Malic enxyme
Malate dehvdrogenase
Superoxide
dismutase
Superoxide
dismutase
GOT-l
GOT-2
CA
PEP
EST-l
EST-02
IDH
GDH
a<;PD
IAP
LDH
CL0
ME-l
ME-2
MDH
SOD-l
SOD-2
conditions
Enzyme
Tissue
commission No.* type’
2.6.1.1
2.6.1.1
4.2.1.1
3.4.13.1
3.1.1.1
3.1.1.1
1.1.1.42
1.3.1.3
1.1.1.8
3.4.1 I
1.1.1.27
4.4.1.5
1.1.1.40
1.1.1.4f~
I. 1.37
1.15.1.1
1.15.1.1
H
H
R
H
H
H
H
K
M
H
H
I<
M
M
H
H
H
1
Stain
rrfrrrnc-C’
Electrophoretic
conditions”
I
I
I
I
.5
6
2
2
5
5
2
3
,F
i
i
i
I
:1
I
I
‘,
1
I
7
?
?
3
1
-I
L,
‘>
;
‘1
i,
’ I.ocus acronyms follow Richardson
et al. (19%).
‘Nomenclature
Committee of the International
Union of Biochemistry
(1978).
’ H, heart muscle; K, kidney; M, skeletal muscle.
’ 1. Tris-ritrate
EDTA, 250 V, 5 h; 2, Tris-versene-borate.
200 V. 6 h; 3, Tris-citrate pH 6.7. 150 c’, 5.j 11:
4. Poulik, 2.50 V, 5-6 h; 5. Tris-citrate
pH 8.0, 130 V, 5 h: 6. I.ithium hydroxide,
350 V, 6-7 h:
7, TrisHCl, 250 V. 5 h.
’ 1, Selander et al. (1971); 2, Sicilian0 and Shaw (1976): 3. Shaw and Prasad (1970); 4, Hal-r-is and
Hopkinson
(1976); 5, Recipes routinely used in the laborator-\ of R. Highton and outlined in H~tlgt~
(1986).
’ CJycyl lrurine used as substrate for Dipeptidasr.
chord
distance.
the combined
group
in both
to the carcharhinids
characters
(Compagno,
A cladistic
was treated
analyses.
and sphyrnids
macrostwma were
Hemigaleus
Hemigaleids
based
are regarded
on a number
used
analysis
3.0k (Swofford,
as a character.
was carried
1990)
out
using
on a Macintosh
Electromorphs
character
states. In the few cases where
common
electromorph
of morphological
was entered
within
the
computer
program
IIci. Each presumptive
loci were treated
heterozygotes
were encountered.
as the character
state.
locus
as unordered
the most
As apparent
ht*tcro-
zygosity over all presumed
loci was low (less than 20/u), and most samples
represented
by five specimens,
this method of coding proved satisfactorv
but two cases. These
specimen
which
cases specimens
with other
were both cases where
was heterozygous
were heterozygous
(seed “A”)
ment branch
the taxon
was represented
for the presumed
for one unique
taxa. They were thus scored
A heuristic
wise addition
as
as the sister
1988: 255-256).
parsimony
PAUP, version
microstoma and
Hemigaleus
outgroup
as possessing
locus
under
electromorph
were
~OI-
all
by’a single
study.
both
In
and one shared
the shared electromorph.
procedure
was used to search for most parsimonius
trees. The step(simple) option of PAUP was used to generate
the initial setd tree
while the tree bisection
swapping.
subsequent
branch
survey and
because
reconnection
Only tree topologies
swapping.
Because
electrophoretic
(TBR)
of minimum
a large number
mobilitv
option
was used to implr-
length were retained
of taxa were included
characters
were coded
f’ol
in the
withotlt
XI\
298
G. J. I’. NAYLOR
presupposed transformation sequence (unordered),
most parsimonious trees (MPTs) were generated.
a large number
(37267)
of
EVALUATION
OF MULTIPLE
MOST PARSIMONKWS
TREES
The phylogenetic infiation
content of the data
The phylogenetic information content of the data set was evaluated using the
randomization tests of Archie (1989). Character state assignments were randomly
permuted within each character of the original data matrix to remove phylogenetic
information (Archie, 1989). Fifty different randomized matrices were generated,
each one reflecting the character state distribution of the original data. Each
matrix was subsequently subjected to the heuristic (TBR) tree searching procedure
of PAUP 3.0k. Searches were left to run to exhaustion or were stopped once 10 000
MPTs had been found. The tree length derived from the original data
matrix was contrasted with the histogram of tree lengths derived from the 50
randomized matrices (Fig. 2) and subjected to a one-sample t-test.
Seed tree bias
The effect of different starting seed trees on the outcome was investigated by
contrasting results using stepwise addition (seed “A”, set “A”) with those obtained
from a second seed tree, “B” constructed using random addition (set “B”). This
second tree “B” was subjected to TBR branch swapping until 20 000 MPTs had
been generated. A majority rule consensus of the 20000 seed “B” trees was
constructed and contrasted with the majority rule consensus of the seed “A” trees.
Signal consistency during branch swapping
S&samples
of trees produced by branch swapping were taken to test whether
the same “kinds” of trees were produced throughout the course of a heuristic
search. Each sub-sample comprised a block of 5000 trees produced sequentially
by the branch swapping algorithm (i.e. tree Nos l-5000, 5001-10000,
lOOOl15 000, 15 001-20 000). Majority rule consensus trees were constructed for each
block of 5000 trees and compared with one another and with the majority rule
consensus tree for all 20 000 trees using partition distances (Penny and Hendy,
1985). This procedure was carried out separately for trees in set A and for trees in
set B.
The consistency of the hierarchical signal among s&samples
was also investigated graphically sense Naylor (1992). Each sub-sample of 5000 trees was subjected
to the plot group frequency option of PAUP. This option orders “groups found”
(clades) by their frequency of occurrence in a sample of MPTs. The “groups
found” in subsamples were re-ordered to coincide with the order found for the
entire 20 000 tree set. The frequency of occurrence of “groups found” was plotted
for each of the four subsamples and compared. This procedure was carried out
separately for both tree sets A and B.
Successive weighting
Successive weighting (Farris, 1969, 1989; Carpenter, 1988) was adopted in an
attempt to reduce the number of most parsimonious trees produced. Autapomorphic character states (electromorphs)
were removed from the data set and
PHYLOGENYOF REQUIEM AND HAMMERHEAD SHARKS
299
recoded as missing values to ensure that the resealed consistency indices used in
successive weighting were not artificially inflated (see Farris, 1989). Characters
(presumed loci) were weighted according to the best fits of their resealed consistency indices to the initial 20 000 trees. TBR branch swapping was carried out on
the initial seed tree with the new character weights enforced. After 20 000 new trees
had been generated, the characters were re-weighted according to their new best
fits. This protocol was re-iterated until character weights stabilized, at which point
the semi-strict consensus of the resultant set of trees was computed. The procedure
was applied to tree sets A and B, separately.
The two final sets of successively weighted trees (one from set A, the other from
set B) were sub-sampled as outlined above. Hierarchical signal constancy among
subsamples within tree sets, and between tree sets, was contrasted with its counterpart in the unweighted analysis.
The topology of the distance Wagner tree was entered into PAUP to determine
its length.
Results and Discussion
A total of 201 electromorphs was detected over the 17 presumed loci surveyed.
Of these 116 (58%) were found to be unique (autapomorphic)
to individual taxa.
The electromorphs
detected were alphabetically coded in the character state
matrix to reflect their relative electrophoretic
mobilities and are presented in
Appendix 2a. Those most anodal are denoted by “A” while the most cathodal are
denoted by “Z”. (Because 27 alleles were found at the PEP locus the most cathodal
allele was scored using the integer symbol “l”.)
The rooted distance Wagner network (Fig. 1) had a cophenetic correlation
coefficient of 0.929, an Fvalue of 7.1 (Prager and Wilson, 1976) and a length of 92
steps when fitted to the character matrix.
32 767 most parsimonious trees (MPTs) of length 86 steps (computed excluding
autapomorphies) and retention index (RI) 0.892 (Farris, 1989) were generated by
branch swapping (TBR) on seed tree “A”. This number represents a ceiling
imposed by computer memory rather than a complete set of MPTs. 20 000 MPTs
of length 86 steps and Rl 0.892 were generated by branch swapping on the seed
tree “B” at which point TBR branch swapping was deliberately interrupted.
Results of the randomization tests indicate that the MPT length of 86 steps is 31.44
standard deviations shorter than the mean MPT length of 159.18 found for the 50
randomized data sets (Fig. 2), suggesting that the data matrix has a high overall
phylogenetic content even though it yields a large number of MPTs.
The majority rule consensus tree (Margush and McMorris, 1981) for set A was
resolved into 27 component forks [Fig. 3(a)]. The majority rule consensus tree for
set B was resolved into (29) component forks [Fig. 4(a)] which differed in topo
logical structure from that of set A.
In the unweighted implementation,
majority rule trees for sub-samples within a
tree set yielded closely similar topologies to the majority rule tree for all trees in
a set. This similarity in “hierarchical
signal” throughout the course of branch
300
(;..I.
P. NAM.OR
C.sorrah
Caltimus
- C.plumbeus
I
Cleucas
Cmelanopterus
lmacrorhinus
T.obesus
N.brevirostris
P.glauca
S.lewini
Stiburo
S.tudes
S.mockarran
-
Szygaena
G.cuvieri
Racutus
R.lalandii
Rporosus
R.terranovae
H.macrostoma
H.microstoma
,
0.0
I
0.1
1
I
I
I
0.2
0.3
0.4
0.5
I
0.6
I
0.7
DISTANCE FROM ROOT
Fig. 1. Distance Wagner tree (Farris, 1972) of 37 species of carcharhiniform
sharks. The tree was
constructed
using Cavalli-Sforra
and Edwards (1967) chord distance and was rooted with the two species
of Hemigakus as an outgroup.
The tree topology had a length of 92 steps when fitted to the original
character
state matrix.
PHYLOGENYOF
301
REQUIEM AND HAMMERHEAD SHARK!3
lo-
0
(
I
90
80
I
110
I
100
”
I
120
---r----
-4
140
130
16”
150
MPT length
Fig. 2. Histogram
plot of most parsimonious
tree lengths derived from 50 rdrldom permutation\
ot
the data matrix (Archie, 1989). The original data matrix yields an MPT length of86 steps which is :%I.+I
standard deviations shorter than the mean of 159.18 steps found for the permuted
data sets, wggcw~ng
that the original data matrix has latent hierarchical
structure.
swapping
on
frequency
plots
a single
starting
tree
(Fig. 5). The
minor
reflected
in the
among
sub-samples
contrasts
with the variation
partition
distance
distances
to be about five times smaller than among
corresponds
The
[Fig. 6(a)].
to different
topological
search
The
“islands
similarity
tree set, suggests
heuristic
between
is also
variation
partition
of trees”
that only a portion
indices
clearly
trees
among
of the total number
the majority
searches
on different
rule consensus
set also shows that majority
a single
branch
in majorit):
rule consensus
signal produced
less resolved
swapping;
rare topologies,
as more
by contrast,
reflects
a
f~-orrl ;I
over all tht
to use heuristic
srarch-
of executing
incomplett
to conclusion
on a singlt
rule trees among
more
throughout
seed tree than does the semi-strict
becomes
within
advised to use the several seeds option.
similarity
of the hierarchical
one search
set
(1991).
suhsamples
of trees resulting
the alternatives
seed trees or executing
seed tree, one is better
The topological
between
tree
This probahh
as put forward by Maddison
rule
to estimate
and has to choose
group
tree sets
show within
tree set distances.
trees in a tree set. It follows from this that, if one is obliged
ing methods
within
tree sets A and B in the pair-wise matrix of tree
in majority
are required
sub-sample
“rare”
accurately
the course
consensus.
groups
the majority
the predominant
sub-samples
reflects
of branch
(The
art- found
rule
the constant\
swapping
semi-strict
during
consensus,
hierarchical
within ;I tic’<’
being
signal
on
consensus
the courst’
of’
unaffectrd
hi
throughout
branc.11
swapping).
Successive
weighting
resulted
in the stabilization
‘The final weights for tree set A were identical
are
presented
topology
in Appendix
for both
less sensitive
to bias arising
sets will be required
2b).
The
tree sets, suggesting
semi-strict
consensus
that the successive
from different
to test the generality
of weights after three iterations.
to those of tree set B (final weights
seed trees.
tree
weighting
(Further
had
the
same
procedure
work on other
i\
data
of this supposition).
Sub-sampling
of the final set of successively weighted trees from both tree stats
revealed a different hierarchical
signal profile during branch swapping than was
observed
for the unweighted
implementation
as described
by Naylor (1992).
Frequency distributions
of groups found (Fig. .5) had broad plateau regions at the
top of curves corresponding
to higher
numbers
of groups
found
in 100%
of tree4
_
Lmacrorhinus
C.acronotus
C.albimarginaius
C.attimus
-[
C.plumbeus
C.fafciformis
Cgalapagensis
Cbngimanus
- Cobscurus
-- P.gfauca
- Cperezi
Camblymynchos
Cwheeleri
C.brachyurus
Cbrevipinna
c.leucas
C.limbatus
Cmacbti
Cmelancpterus
C.porOSuS
Csealei
Csorrah
____
-
-
H.macrostoma
H.microstoma
C.acronotus
C.albimarginatus
C.attimus
C.plumbeus
C.amblyrhynchos
C.wheeleri
C.brachyurus
C.brevipinna
C.falciformts
Cgalapagensis
Cobscurus
C&don
C.leucas
C.limbatus
C.bngimanus
Cmacbtt
Cmelanopterus
Cparezi
C.porosus
Csealei
Csorrah
Gcuvieri
L.macrorhinus
N.brevirostris
P.gtauca
R.acutus
Rtalandei
Rporosus
Rterranovae
S.lewinr
Smockarran
S.ttburo
Studes
Szygaena
T obesus
-4
r
---I
-
-
-
S.lewini
Smockarran
Szygaena
Stiburo
S.tudes
N.brevirostris
Tobesus
L.macrorhinus
G.cuvierr
Racutus
R.lalandei
R porosus
R.terranovae
Cbrachyurus
C.brevipinna
Csorrah
C.leucas
C.limbatus
C.melanopterus
-----I C.a”imus
Columbeus
PHYLOGENYOF
REQUIEM AND HAMMERHEAD SHARKS
303
as seen in the unweighted analysis. However, while the within-tree set distances are
comparable to those seen for the unweighted case, the among-tree set distances
are roughly half those seen for the unweighted case. Invoking Maddison’s (1991)
“islands of trees” analogy once again, it would appear (for this data set at any rate)
that successive weighting has drawn islands in the “archipelago”
closer together
without affecting their size.
(a) Majority rule consensus - No weighting
I
I
I
1
(W Semi-strict consensus - No weighting
(cl Semi-strict consensus - Successive welghting
Fig. 4. (a) Majority rule consensus tree of the first 20 000 MPTs generated by branch swapping on
seed B. (b) Semi-strict consensus tree of the first 20 000 MPTs generated by branch swapping on seed B.
(c) Semi-stl-ict consensus tree of successively weighted trees generated by branch swapping on seed B.
Final (asymptotic) weights used are given in Appendix 2b. Note that the topology for Fig. 5(c) is the
same as that of Fig. 4(c), reflecting the convergence of tree topologies toward the same solution.
G..J. P. NAYI.OK
set A
0
100
group categories
set 0
200
0
100
200
group
categorias
Fig. 5. Hierarchical
signal consistency
among subsamples
of trees. Each plot depicts the frequencies
of corresponding
groups (clades) found in each of four sub-samples.
Each subsample
comprised
5000
trees (l-5000,5001-10
000, 10 001-15 000,15 001-20 000). Note that the cuxves produced
by successive
weighting have characteristically
extended
plateau regions and truncated
tails, reflecting,
respectively,
an increased number of clades found in 100% of the trees and a decreased number of rare clades found
among all the trees. Both factors cause the strict consensus tree derived from successively weighted trees
to be more resolved.
The semi-strict consensus of the successively weighted analysis was chosen as the
most reliable summary of the phylogenetic signal in the data on the grounds that
it was: (a) uninfluenced by seed tree bias; and (b) the most resolved tree that was
entirely consistent across all s&samples
(the majority rule trees varied across
subsamples for some of the groupings found).
VARLATION BETWEEN ANALV~ES
While the retention indices for the MPTs generated in the parsimony analysis
are high (0.892) and the cophenetic correlation coefficient is high for the distance
Wagner tree, the conclusions concerning relationships vary between analyses. The
semi-strict consensus tree derived from the successively weighted character analysis
is much less resolved than is the distance Wagner tree. This is largely because the
semi-strict consensus tree is a particularly stringent summary of the MPTs found. As
a result, there are clades found in the distance Wagner tree that appear unresolved
in the semi-strict consensus tree, but that are supported by a large percentage of
PHYLOGENYOF
the MPTs used to generate
conservative,
only
successively
weighted
The
genera
genera
Rhizopionodon,
(the
than
sharks
(Sphyma)
topological
resolution
reef white-tip
weighted
is not
character
seen
congruence
However,
Relationships
within
Carcharchinus
species pairs are defined
The
among
of the blue
then the genus
However,
when
shark
P. gkzuca is included,
the genera
analysis. Interestingly,
Wagner
clade using the constraints
clade
(0.79%
increase
loci scored
the genus
option
in 9.5% of
Somts
plumbru.s-
[C.
abollt
rrla-
it has experienced
outside
weighted
implementation,
based on parsimony
that independent
accelerated
teeth
referable
1905),
(l.andini,
to
whereas
referable
19’7’7). If indeed
but rather
the sister group
must be explained.
(One
explanation
membc~- of
cxvolrltion rclativc,
at so many loci and gi\rell
extra
to make
steps in the utl-
in tree length
systematic
How&r,
support
it is notcworth\~
Compagno
(Jarrharhinus
appear
in
the
to Prionace do not
is of course,
that did not develop
( l!)Xti)
of‘ sharks.
Middle
appear
Ptionacv is riot a derived
a clade containing
in the,
statemcnls
the idea that thr bIu~-
Eoctsnc,
until
member-
I hc~
\vithitl
(2~rcharhinrr.c. I.O.WJC~O~C,
Tn’aenodon, Negaption, as suggested by the distance Wagner analysis,
spicuous absence of its fossil teeth during the Eocene.
Oligocene,
descendant
of an old lineage
until the Pliocene.)
tht.
of the presunlecl
a derived
molecular
position.
first
of’
fi-onI
to 62 I.53
link I’. g/ouca to the obscures group
Carcharhinus
teeth
derivrd
61 664
at eight
1% increase
and fossil evidence
teeth
from
cloes not require
less than
species within the genus
serrated
(‘.
in the number
of P. glauca from
length
that, if it is indeed
as to its true phylogenetic
that its broad
0/~sr11171.~,
from the (,‘awhwhilir~u~
the exclusion
it may be premature
shark might be a derived
C.
If the weights
in tree
Carcharhinus
and causes
morphological
containing
is no increase
P. glauco is autapomorphic
Carcharhinus,
If I? &IILC~~is
group in both analyses.
analysis but as the sister taxon to tht,
there
then
2a). This suggests
(2rrrharhinu.r
tree.
analyses.
7dm!4~i].
is problematic.
implementation).
in an increase
in length).
implementation,
Pliocene
in all
Loxodon, Negaption and Trkzenodon in thta
of PAUP,
(Appendix
that its placement
(Stromer,
was found
if P. glawa is excluded
are implemented,
results
weighted
Similarly
shark I
Wagner
are disagreements
to its sister taxa. Given that P. glaum is autapomorphic
suggests
shark),
lemon
of the successiveh
between
it falls in a clade
Carcharhinus,
(in the unweighted
weighting
(Jarcharhinus
Relation-
tiger
tree.
1’tionace &zutn
distance
successive
(Rhirtr
Carcharhinus
tree
forms a monophyletic
Carrharhinus
clade containing
steps required
outside
consensus
But there
and C. longimanus in the character
galapagensk
analyses.
Negaprion (the
considerably
brarhyww].
sharks
these sub-groups.
placement
omitted,
between
in both trees, i.e. [C. amb@-hynrhos-4.
[C. brev ip’znna-C.
C. altimus],
tionships
sharpnose
this basal placement
vq
to be consistent
basally on the distance
the consensus
too
of thr
reliable.
tend
Galpocerdo (the
are placed
semi-strict
the MPTs that were used to generate
tree
clades in all trees and havtt
concordance.
shark),
in the
analyses.
and
(the slit-eye shark),
Loxodon
consensus
Carcharhinus
monophyletic
trees. Rhizojwionodon and Gakocedro, are placed
This
At the risk of seeming
semi-strict
here as phylogenetically
other
also show some
Sphyma,
Triaenodon
consensus.
in the
fall into distinct
of within-clade
among
found
Hammerhead
for example,
a high degree
ships
within
analyses.
pionodon),
the semi-strict
groups
tree sets are regarded
relationships
between
and
those
30.5
REQUIEM AND HAMMERHEAD SHARKS
that I? ,&ucn
its unique
then thy CX)I~and hlioc.t~~~c~
is the moclt~r~1
tooth
morphologic
C;..]. P. NAMX)K
(a) PARTITION
MAJORITY
(DERIVED
a
b
METRIC
DISTANCES
RULE CONSENSUS
FROM UNWEIGHTED
c
d
BETWEEN
TREES
DATA SET)
e
f
g
h
t
I
I
:22
I
23
25
:23
I
I
24
26
\\
Consensus
Tree
a
b
:
:!
g
h
\\
__ -4 I_*_.
:_23 _ _ _ _2_4_ _ _2P_
KEY:
\.
MPTs
Used
l-5000
5001-10000
10001-15000
15001-20000
l-5000
50001-10000
10001-15000
15001-20000
\
TREE
SET A
TREE
SET B
>
Fig. 6. (a) Partition metric distances (Penny and Hendy, 198.5) among majority rule trees constructed
from s&samples of trees generated by branch swapping on unweighted data sets. Trees a to d are derived from branch swapping on seed tree “A”, while trees e to h are derived from branch
swapping on seed tree “B”. The partition distances show that groups of trees derived by branch
swapping on a given seed tree are more similar to one another (triangular outlines) than they
are to trees generated from a different seed (rectangular outline). This observation is concordant with
Maddison’s (1991) finding of islands of trees.
CORRESPONDENCETO THE STRATIGRAPHK:
SEQIJENCE
The remainder
of the discussion addresses the tree topology
semi-strict consensus
tree of successively weighted trees [Figs
Because carcharhiniform
1984) the first recorded
depicted in the
3(c) and 4(c)].
sharks have such an excellent fossil tooth record (Maisey,
stratigraphic
appearance
of each taxon is presented to
PHYLOGENYOF
M
REQUIEM AND HAMMERHEAD SHARKS
307
PARTITION
METRIC DISTANCES
BETWEEN
MAJORITY RULE CONSENSUS TREES
(DERIVED FROM SUCCESSIVELY WEIGHTED
b
a
c
d
e
DATA SET)
f
g
h
j&,
0
‘.
1
\
\
0
2“.
'5
.
,
\
\
0
\
I
15
4
z“,
I________----->
:Is_
_
_;3_ _ - _,;_ - - - -
11:
I
I
y.
13;
I
I
I4“,
,
I
,
I
;16
11
'14
I
9
11
11
13:
:_w
___l_J__ _13_- ._JJ1
KEY:
Consensus
Tree
a
b
C
d
e
f
9
h
:6
I
I
I
I
I
I
:_4___
.
\
.
\
4“.
.
\
.
\
_4____4_1‘>
MPTs
Used
l-5000
5001-10000
10001-15000
15001-20000
\
l-5000
50001-10000
10001-15000
\
I 5001-20000
TREE SET A
TREE
SET B
J
Fig. 6. (b) Partition metric distance (Penny and Hendy, 1985) among majority rule trees constructed
from sub-samples of trees generated by branch swapping after asymptotic successive weighting. Trees a
to d are derived from branch swapping on seed tree “A “, while trees e to h are derived from branch
swapping on seed tree “B”. The partition distances show groups of trees derived by branch swapping on
a given seed tree to be more similar to one another than to trees generated from a different seed. Note,
however, than the partition distances between different tree sets are considerably smaller for the successively weighted case than they are for the unweighted case depicted in Fig. 6(a).
determine
whether
the cladogram
branching
appearance times.
The tiger shark Caleocerdo cuvieri emerges
most basal to the selected
outgroup
sequence
is concordant
as one of seven polytomous
Hemigakus.
As mentioned
previously,
with first
branches
Galeomdo
and Rhiwprionodon form the two most basal branches in 95% of the successively
weighted MPTs, perhaps suggesting a more basal position for these two taxa than
SO8
actually
depicted
first recorded
Capetta,
in the semi-strict
from
The
four
species
cluster
1988).
basal placement
Eocene,
of sharpnose
traits
(Compagno,
with
clade
species
similarities
hammerhead
from
that
incorrectly
the
same
similar
than
makes
shapes.
seven
the
on
the
teeth
argument
between
snout
occurring
Within
1988;
pers.
The
fossil
from
fossil
the origin
tooth
ancestor.
represent
and
Rhiqtrrionodon
would
that
the
the lateral
vestige
and
origin
the two diminutive
Loxodon are so similar
appear
(Capetta,
in the Middle
The genus
Fossil
hierarchical
of Rhizoprionodon
1987).
1905).
this period
Teeth
teeth
The
are
species
of Rhiroptionodon
of Nigeria
first
most commonly
within
the genus
or Scoliodon.
Triaenodon
at the
and Sphyrna. Teeth
in size
have not yet been
Teeth
small
and
to my
from Negaprion first
1955).
from
the
recorded
C. egertoni (Capetta,
RELATIONSHIPSWITHINTHECLADECONTAINING
1987;
of sphyrnids
and Scoliodon, both
a large clade that appears
recorded
a
shape
independently
Triaenodon are quite
(White,
tooth
shark,
in the fossil record
from
onward is the C. leucaslike
Relationships
white-tip
in the fossil record.
Carcharhinus comprises
Ca&arhinus
(Stromer,
to those
from
trowel-like
Sphyma tibwo and S. tudes
level as Galeocerdo, RhiqtAonodon
reported
Eocene
reef
be
(1988)
of an incipient
with the origin
species
the
macrorhinus,
first appearances
have not been
been
(Capetta,
Scoliodon
shark, Negapn’on breuirosttis, all branch
knowledge
of
might
Compagno
with the very similar
be consistent
to the
of sphyrnids
obsesus and the Lemon
established
have
of the hammerheads
He suggests
in conjunction
subsequent
that reliable
group
teeth
It is possible,
horizons
evidence.
Loxodon
and shape
pair
distinct
fossil
1987).
form a sub-clade.
The slit-eye shark,
from
This
monophyletic
earliest
(Capetta,
at earlier
This
the hammerheads,
same polytomous
R. @rosus and R.
counts.
R. acutus.
for the development
obs.)
at any time
(Lower
sequence
are markedly
a clearly
Miocene
of Scoliodon might
hammerheads
Compagno,
form
found
ridges
taxa. The
branching
by morphology.
polytomy.
Lower
supposed
cephalofoil.
The
only by vertebra1
branch
pre-anal
as Physogaleus, Scoliodon or Rhizqiwionodon due to their
Scoliodon or Rhizo@ionodon-like
flanges
and
with the early appearance
in the fossil record.
If this is the case, then
is currently
hammerhead
mono-
two primitive
in the more derived
and relatively primitive
either
a convincing
a distinctly
furrows
is corroborated
sphyrnid
identified
strikingly
1952:
labial
is consistent
(Sphyrnidae)
from
small
of (GnLocPrcto are
(Arambourg,
occurs
with R. Zalandii while all three
sharks
Sphyrna are recorded
however,
teeth
Rhi.zojwionodon shares
distinguishable
from the largely circumtropical
branching
Fossil
of Morocco
Rhizofwionodon form
Galporerdo: long
of its teeth
the Rhizoptionodon
terranovae are sibling
The
sharks
None of these characters
1987)
have many cranial
tree.
Eocene)
from the polytomy.
of Rhiqbrionodon
Capetta,
consensus
(Lower
1987).
branching
morphological
within
the Ypresian
198 1; Capetta,
phyletic
older
P. N.Z\TL.OR
(;.,I.
to include
Middle
Eocene
P. gluuca.
of Egypt
in the fossil record
from
1987).
CARCHARHINUSAND PRIONACE
Carcharhinus are only partially
resolved.
C. UCTP
notus is placed as the sister taxon to all other members
of the genus. C. isodon
branches next, after which there is then a polytomy containing
11 branches. Eight
of these
11 branches
give rise directly
to terminal
taxa. Two of the branches
give
PHYLOGENYOF
rise to the
species
pairs
wheeler-i). One branch
REQUIEM A.ND HAMMERHFm
taxa.
All eight sharks in this resolved
length.
All but one of the eight
tudinally
branch
between
containing
C. galapagensis,
PREVI~LXYPR~POSED
Springer
(1950,
suggested
pointed
ridge-backed
and
that
members
pectoral
treatment
characters
that
of a group
group,
include
large ridge-backed
tips
assessment
(1982)
rounded
except
The
of the genus
group
Garrick
(1982)
sub-groupings
(1982)
forms.
rounded
tips
with
(Garrick,
that both
his
of specirs
too
1982).
be regarded
Garrick
as the c-entral
C. perezi and (:. long-
obscun~s group,
th us excluded
was erected
C. son-ah, C. falr@rnis,
Eulamia group but included
Compagno
Hr
with
(1988)
to
teeth
(.A rtlb-
C. longimmnrts,
concurred
with
(;. perezi from the group and included
here
suggest
that the large ridge-C. long-imc~nuv, (.‘. /al&
(.‘. ppfpzi) ;irc’ all members
of a monophvlcti(
P. glauca.
and Compagno
(1988)
seal& group.
character
(1988)
have postulated
based on morphological
suggest
several
in the semi-strict
analysis. Compagno’s
tentati\.c
Both (&rick
he closely
is supported
co~~sc~xas
inclusion
other
similarity.
that C. po)us~~.sshould
This close relationship
tree but does not appear
sively weighted
be divided
into those
the inclusion
might
to as the
data presented
within Carcharhinus
the C. dussumm’-C.
Wagner
those
(C. ol~~~rus, (A galapagtxsis,
C. altimus and
Compagno
and
and
C. altimus, (I plumhew,
fin morpholoE.
allozyme
tha: also includes
be divided
with each other
that he removed
C. albimarginatus.
,formis, C. plumbpus,
clad<
sharks that had, for the most part, broad triangular
backed
members
should
necessitated
he referred
Garrick
in spite of its distinct
sensu lato might
as it was considered
C. galapagensis
marginatus and C. sealei from Springer’s
oarrick’s
off next,
polytomous
CAK~IIMHIN~V
(Eulamia)
groups
included
which
(C. pmzi excepted).
forms
to be aligned
which
This
to an apical
Carrharhinus
was abandoned
C. obscures and
manus.
longi-
C. altimus and C. plumbewc
(Carcharhin2c.s sense stricto) and ridge-backed
and smooth-backed
in other
proposed
than 2 m total
ridge running
C. faltiformis, branches
the sister taxon
suggested
forms
dorsal
to more
(1.
eight
C. obscurus, C. longimanus and P. glauca.
1951)
Pterolumniops. This
diverse
forms
in turn that the ridge-backed
first
the clade.
INTERRELATIONSHIPSWTHIN
into smooth-backed
clade containing
have an interdorsal
pair. The silky shark,
by C. perezi, which
(C. amblyrhynchos,
well-resolved
clade are large, growing
(I? gh,cu)
the two dorsal fins. Within
off basally as a species
followed
and
C. bruchyurus)
(C. breuipinna,
gives rise to an internally
309
SHARKS
linked
to
by the distance
tree or the succc~-
of‘ (I mdoti
and (:. wwrclh
in this group is not supported.
Garrick
and
suggested
(A ZI&JP~P~~
on
characteristics.
This
that
the
C. albimar@natus
basis
suggested
tree but does not appear
of overall
was closely
morphology,
relationship
in the semi-strict
related
tooth
is supported
co~~sc~~s~~s
to CL ambly~-h~rld/o.~
_
shape
and
vcr-tehal
by the distance
tree. Compagno’s
L$‘agnt.rtreatment
of (:. perezi as the sister group to C. clmb1vrhvnchn.sand (.L 7rhwlti
is not supported.
The suggestion by both authors that- the sibling species C. 1irnbatu.c and C.‘. atnhi~rhvnrhoidus are most closely related
beuipinnn f-rills closest
to CAbryuipirzna is not supported.
to C. brachyurus. The
striking
similarity
In this study. (1.
between
cl. litd/atw
and C. brvc~ipinna in body shape, color pattern, tooth shape and behavior appears to
be the result of convergent
evolution rather than of rlose rrlatedness.
Support for
G.,J. P. NAM.C)K
310
this supposition
is seen in the disparate
to color pattern.
tips. These
diffusely
C. &&ztusjuveniles
markings
pigmented
unpigmented
The
present
recent
(pers.
but develop
with age (Garrick,
been dipped
diminish
182),
ontogenies
in intensity
obs.).
of the two species with respect
are born with distinctive
black pigmented
fin
with age so that large adults appear onl)
C,‘. brpvipinnn juveniles
are born
In contrast,
an increasingly
intensive
so that large adults often
black
appear
pigment
on the fin tips
as though
their fins had
in black ink (pers. obs.).
allozyme
survey
by Lavery
study. In his study, Lavery
tunately,
the
two data
evidence
for
the
sets cannot
phylogeny.
scored
(1992)
is closely
38 allozyme
comparable
loci for 21 taxa.
to the
Unfor-
be readily combined
to evaluate the total
are thus restricted
to topological
Comparisons
C.acronotus
Csealei
Cdussumeri
C.porosus
C.titzoyensis
Calbimarginatus
Camblyrhynchos
C.wheeleri
C.altimus
Cperezi
C.plumbeus
Cgalapagensis
C.obscurus
C.longimanus
C.brachyurus
C.brevipinna
Climbatus
Camblyrhynchoides
Cfalciformis
C.sorrah
C.leucas
Camboinensis
C.melanopterus
Ccautus
Cbornensis
Rhizoprionodon
Fig. 7. Phylogenetic hypothesis for sharks of the genus Carcharhinus after Garrick (1982). Topology
deduced from Garrick’s text.
PHYLOGENYOF
REQUIEM AND HAMMERHEAD SHARK!S
311
similarities
and differences
among inferred
trees. In Lavery’s study, one most
parsimonious
tree resulted with a length of 215 steps and a retention
index of
0.545. The monophyly
of the genus Curcharhinus was not supported
in his study.
Instead, the genus was seen to be paraphyletic
including
members
of two other
genera (Negapn’on amtudiem and Galeocerdo cwieri) and one member of an entirely
different family (Hemipistis ehgutus) . A re-analysis of Lavery’s data shows that if
the genus Curcharhinus is constrained
to be monophyletic,
five most parsimonious
trees of length 219 steps result. This represents
a four-step
increase
over his
presented
most parsimonious
tree. (Lavery reported
that “placing
these three
species in their recognized
positions outside the genus Curcharhinus results in a tree
C.acronotus
C.brachyurus
C.falciformis
C.melanopterus
C.cautus
Csignatus
Nasolamia
C.albimarginatus
Caftimus
C.galapagensis
C.longimanus
C.obscurus
C.plumbeus
Cleucas
Camboinensis
Glyphis
Lamiopsis
Negaprion
Prionace
Camblyrhynchos
C.perezi
C.wheeleri
C.brevipinna
Climbatus
Camblyrhynchoides
C.leidon
Cisodon
C.porosus
C.sealei
Csorrah
C.fitzroyensis
E
Cdussumeri
C.bomensis
Cmacloti
Chemiodon
Triaenodon
Fig. 8. Phylogenetic hypothesis for sharks of the genus Carcharhinus
Topology deduced from Compagno’s text and presented cladogram.
atter Compagno
(1988).
(i. J. I’. NAYLOK
3 12
almost
10% longer
ment]
I was unable
[either
228 or 229 steps,
to reproduce
depending
this finding).
on the precise
By reciprocal
sion of Neguption breuirosttis, Galeocerdo cuuieri and Hw@zkus
the present
study results in an increase
As pointed
out by Lavery
as they examine
polling
different
the taxa from
phylogenetic
structure
(1992),
portions
both
in tree length
the two studies
within Carrharhinus
do not overlap
phylogeny.
is recommended
within the genus
the incluin
of two steps.
of the Currhnrhinus
studies,
arrange-
contrast,
to more
extensively
A larger
survey,
confidently
assert
Curchnrhinus.
CIASSIFICATION
Results
from
both
the
the Carcharhinidae,
hammerhead
cladistic
the
herein
(genus
of the Carcharhinidae,
are in agreement,
Proposals
suggested
to the
(1934,
1939,
that “the
prime
reason
without
sufficient
forms]
and Springer
for these
and Eulumia
dependent
accelerated
more,
(1982).
to,
molecular
the
shows
that reflects
evolution
hypothesized
a
or
largely
supported
of
cladogram
a formal
re-classification
relationships
being collected
(1982)
established
in
intra-generic
of
s.1.
the data.
within
available.
be
If a
of
Further-
the
genus
monophyletic
several monotypic
Curchurhinus
data.
ad hoc incidences
strictly
put
are strongly
by the available
any
known
affinities
proposed
to explain
becomes
other
Curcharhinus
is excepted.)
of relationships
tree,
that
concerning
(1853))
Garrick
from,
cladogram,
must be invoked
ommended
evidence
forms]
the presented
unresolved
1951).
distinction
the tentative
of the genus would involve naming
are currently
re-classification
s.1. into two genera,
classification
further
the
data presented
to become
some of the groupings
and are only tenuously
is adopted
because
Curchurhinus
characters,
within
on the basis of only one species for
[ridge-backed
However,
(1950,
taxa failing
of Curchurhinus
Results of this study partially corroborate
forward by Garrick
classification
the
a tentative
Sphyrnini
cladistic
that
includes
have been made by Owen
Curcharhinus
1943)
reference
(Sp rm g er’s division
[smooth-backed
analysis
tribe
that Compagno’s
the genus
Whitley
and
species.”
which
proposed
data and the molecular
usage is that for the most part they were erected
each
group
(1988)
indicate
based on morphological
assigned
I suggest
analysis
be adopted.
to subdivide
Gill (1962),
were
character
Compagno
As both the morphological
hammerheads
and
is a paraphyletic
Sphyna).
hammerheads
Carcharinidae.
Wagner
as it now stands,
sharks
reclassification
in which
distance
re-
taxa. It is rec-
postponed
until
DNA sequence
data
for this purpose.
Acknowledgments
Several people assisted
particularly
like to thank
allowed
me to accompany
in the collection
the commercial
them
fishing
of material for this study. I would
fishermen
and scientists who either
or actually
They are: from the east coast of North America:
N. Sanders,
J. Musick and J. Colvocoresis;
collected
specimens
for me.
D. De Maria, R. Bakey, E. Sanders,
from Hawaii: E. Schellenburger,
S. Raiser
and M. Suiso; from the Red Sea: A. Sulliman and “Muhammed”.
Others who
generously
provided either direct assistance or advice were R. Apilado, “Mungpert” and D. Cadra (Philippines),
J. Parrish and Sea Life Park (Hawaii), P. Coutin
PHYLOGENYOF
REQUIEM
AND HAMMERHEAD
SHARKS
3 I3
and P. Showers (Sierra Leone), R. Fox, D. Doubilet and H. Fraise (Australia),
M. Doty, L. Cohen, T. Bowman (Belize), S. Blum, J. Casey, R. Shipp, J. Dendo,
G. Nelson, L. Abele, V. Springer, G. Burgess, J. Castro, J. Ott, E. Silverfine, D. Carr,
J. Jibrin, S. Wilkes, S. Fitzgerald, D. Chipman and E. Knurek (U.S.A.).
I thank Carla Hass for patiently teaching me electrophoretic
techniques and
E. Clark, M. Doty, R. Highton, C. Miter, G. Vermeij, R. Voss, G. Wilkinson and
four anonymous reviewers for their helpful suggestions for the improvement of the
manuscript. I would like to give special thanks to R. Highton for the use of his
electrophoresis laboratory, to G. Vermeij for his persistent encouragement and to
E. Clark for first suggesting I pursue this project. Financial support was provided by
The WIN Foundation, The University of Maryland Foundation, The University of
Maryland Computer Science Center, D. Shen and the National Science Foundation
(grant BSR 87-08121 to G. J. Vermeij and grant BSR 83-07115 to R. Highton).
REFERENCES
ARAMBOUR~;,C. 1952. Les vertebres fossiles des gisements
de phosphates
(Manx-AlgrrirTunisie). Serv. Geol. Maroc, Notes et Mem. 92: l-372.
ARCHIE,J. W. 1989. A randomization
test for phylogenetic
information
in systematic data,.
Syst. Zool. 38: 239-252.
BARANES,
A. AND 0. SHAHRABANYBARANES.1986. A numerical taxonomic study of the northern
Red Sea carcharhinids.
Proc. 2nd Int. Conf. Indo-Pacific Fishes, pp. 246255.
BIGFLOW,H. B. AND W. C. SCHROEDER. 1948. Fishes of the western
North Atlantic. Part I
Mem. Sears Found. Mar. Res. Yale Univ. Press, New Haven, Connecticut.
CAPETTA,H. 1981. Additions a la faune de selaciens fossiles du Maroc. 1: sur la presence dtss
genres Heptrunchias, Alo@
et Cldonlor!+s dans L’Ypresien des Ouled .Abdoun.
CAobios 14: 563-575.
CAPE?TA, H. 1987. Chondrichthes
II. Mesozoic and (:enzoic Elasmobranchii.
Vol. Jb. tiandbook of Paleoichthyology.
Gustav Fischer Verlag, Stuttgart, West Germany.
C~ENTER, J. M. 1988. Choosing among multiple equally parsimonious
cladograms. <:ladistics
4: 291-296.
CAVAI.I.I-SFOR%A,
I.. 1,. ANDW. F. EDWARDS.1967. Phylogenrtic
analysis: Models and estimation
procedures.
Evolution 21: 550-570.
COMPA(;NO,I,. J. V. 1984. FAO species catalogue. Vol. 4. Sharks of the world. An annotated and
illustrated catalogue of shark species known to <late. Part 2. Oarcharhiniformcs.
F.40
Fish. Synop. 125: 251-655.
COMPA(:NO, L. J. V. 1988. Sharks of the Order Carcharhiniformrs.
Princrton
Ilni\. Press.
Princeton,
New Jersey.
FARRIS,J. S. 1969. A successive approximation
approach
to character weighdng. Svst. ZooI.
18: 256268.
Farris, J. S. 1972. Estimating
phylogenedc
trees from distance ma1rict.s. ,4m. Nat. IO(i:
645-668.
FARRIS,J. S. 1989. The retention
index and the rcscaled consistrncv
index. (:ladistics I,:
417-419.
GARRICK,,J. A. F. 1982. Sharks of the genus Carch~mhinrr.s.NOAA
Trrh. Rep., NMFS. (1it.c..
445.
GARRKK,J. ,4. F. 1985. Additions to a revision of the shark genus Carcharhirms: Synotlyrny OI
,-\ptionodon and Hypoprion, and description
of a new species of C’wchnrhirru~.
N( )A..1
Tech. Rep., NMFS. Circ. 34.
Gr1.1..T. 1862. Analytical synopsis of the order of squali: and revision of the nomenclature
OI
the genera. Ann. Lyceum Nat. Hist. 7: 367-408.
HARRIS, H. AND D. A. HOPKINSON. 1976. Handbook
of Enzyme Electrophoresis
in fluman
(Genetics. North Holland Publishing, Amsterdam,
The Netherlands.
314
G. J. P. NAYLOR
HEDGES,S. B. 1986. An electrophoretic analysis of holarctic hylid frog evolution. Syst. Zool.
35: l-21.
LANDINI,W. 1977. Revizione degli ittiodontoliti pliocenici della collezione Lawley. Palaeontographia Italica. 70: 92-134.
LIVERY, S. 1992. Electrophoretic analysis of phylogenetic relationships among australian
carcharhinid sharks. Aust. J. Mar. Freshwater Res. 43: 97-108.
MADDISON, D. R. P1991. The discovery and importance of multiple islands of most-parsimonious
trees. Syst. Zool. 40: 315-328.
MASSEY,J. G. 1984. Higher elasmobranch phylogeny and biostratigraphy. Zool. J. Sot. 82:
33-54.
MARGUSH,T. ANDF. R. MCMORRIS.1981. Consensus n-trees. Bull. Math. Biol. 43: 239-244.
NAYLOR,G. J. P. 1992. Plotting frequency distibutions of Phylogenetic groupings found
among sets of most parsimonious trees. Cladistics 8: 161-164.
NOMENCLATURE
COMMITTEEOF THE INTERNATIONAL
UNION OF BIOCHEMISTRY.ed. Enzyme
Nomenclature 1978. Academic Press, New York.
OWEN,SIR RICHARD.1853. Descriptive Catalogue of the Osteological Series Contained in the
Museum of the Royal College of Surgeons of England, Vol. 1. Taylor and Francis,
London
PENNY,D. ANDM. D. HENDY.1985. The use of tree comparison metrics. Syst. Zool. 34: 75-82.
PRAGER,E. M. ANDA. C. WILSON. 1976. Congruency of phylogenies derived from different
proteins. A molecular analysis of the phylogenetic position of cracid birds. J. Mol.
Evol. 9: 45-57.
RICHARDSON,
B. J., P. R. BAVERSTOCKAND M. ADAMS. 1986. Allozyme Electrophoresis. A
Handbook for Animal Systematics and Population Studies. Academic Press,
Orlando, Florida.
SELANDER,R. K., M. H. SMITH,S. Y. YANG,W. E. JOHNSON,ANDJ. B. GENTRY.1971. Biochemical
polymorphism and systematics in the genus Peromyscus. I. Variation in the old-field
mouse (Peromyscus polionotus). Stud. Genet. VI. Univ. Texas Publ., 7103: 49-90.
SHAW,C. R. ANDR. PRASAD.1970. Starch gel electrophoresis of enzymes-a compilation of
recipes. Biochem. Genet. 4: 297-320.
SICILIANO,M. J. ANDC. R. SHAW.1976. Separation and visualization of enzymes on gels. In
I. Smith (ed.). Chromatographic and Electrophoretic Techniques, Vol. 2, Fourth
edn. Heinemann, London, pp. 185-209.
SPRINGER,S. 1950. A revision of North American sharks allied to the genus Curcharhinus.
Amer. Mus. Nov., 1451; 1-13.
SPRINGER,S. 1951. Correction for “A revision of North American sharks allied to the genus
Curcharhinus”. Copeia 3: 244.
STROMER,E. 1905. Die fischreste des mittleren und oberen eocans von Agypten. Beitr. Pal.
Oesterr. 18: 37-58, 163-92.
SWOFFORD,
D. L. 1990. Phylogenetic Analysis Using Parsimony (PAUP), version 3.0k. Computer
software. Illinois Natural History Survey, Champaign.
WHITE, E. I. 1955. On Lyamna eulybuthrodon. Blake. Ann. Mag. Nat. Hist. 8: 191-193.
WHITLEY,G. P. 1934. Notes on some Australian sharks. Mem. Queensl. Mus. 10: 180-200.
WHITLEY,G. P. 1939. Taxonomic notes on sharks and rays. Aust. J. Zool. 9: 227-262.
WHITLEY,G. P. 1943. Ichthyological descriptions and notes. Proc. Linn. Sot. N.S.W., Sydney.
PHYLOGENYOF
REQUIEM AND HAMMERHEAD SHARKS
Appendix
315
1. Material Examined
Curcharhinus acronotus: 454791, 60516.3 (LORAN), Charleston, S. Carolina, N.
America (372, 391); Bulls Bay, Nr Charleston, S. Carolina, N. America (373);
Dauphin Is., Alabama, N. America (163); Mayport, Jacksonville, Florida, N.
America (442).
Curcharhinus dbimurginatus: Mercedes, Camarines Norte, Philippines (287)
Curc~urhinus ultimus: Norfolk Canyon, 60 miles east of Virginia Beach, N. America
(110, 111, 116, 122,569).
Curcharhinus umblyrhynchos; Mercedes, Camarines Norte, Philippines (288, 289);
Nihoa Island, (N. W. Hawaiian Islands), Hawaii (545); Tumalutab Is., Zamboanga City, Mindanao, Philippines (327,328).
Curcharhinus bruchyurus; Saint Kilda, Vincent Gulf, S. Australia (003).
Curcharhinus bre+innu; Mayport, Jacksonville, Florida, N. America (433) ; Ponce
Inlet, Daytona, Florida, N. America (445, 450, 452, 461).
Carchurhinus fulc+rmis:
Moro Gulf: S. of Mindanao (via Arena Blanco, Zamboanga), Philippines (329, 330, 332, 333, 339); Port of Spain fish mkt. Trinidad
(474,506,506,508,509).
Curcharhinus gulupugensis: Raena Point, Oahu, Hawaii (128, 130, 132) ; Nihoa Island,
(N. W. Hawaiian Islands), Hawaii (544); seaward side of Rabbit Island, Oahu,
Hawaii (572).
Curchurhinus is&n:
Bulls Bay, Nr Charleston, S. Carolina, N. America (409, 410,
411,41la,
411b).
Carchurhinus Zeucus; Dauphin Is., Alabama, N. America (100); Mayport, Jacksonville,
Florida, N. America (436); Ponce Inlet, Daytona, Florida, N. America (457,458);
sand shoal inlet, eastern shore, Virginia, N. America (370).
Curcharhinus limbatus: Dauphin Is., Alabama, N. America (154, 156); Ponce Inlet,
Daytona, Florida, N. America (444, 446); sand shoal inlet, eastern shore, Virginia, N. America (368).
Curcharhinus longimanw: 19 10 N, 160 51 W (off Hawaii) (574); 19 15 N, 160 49 W
(off Hawaii) (541, 542, 543); off Nihoa Island, (N. W. Hawaiian Islands), Hawaii.
(573).
Carchurhinus mdoti: Turtle Islands, Sulu Sea. (Navotas mkt, Manila), Philippines
(284).
Curcharhinus melunoperus: Anda, Pangasinan.
(Alaminos mkt), Philippines (261,
262); Straights of Tiran, Gulf of Aqaba, Red Sea (005,006).
Curchurhinus obscurus: Dauphin Is., Alabama, N. America (95,99); Norfolk Canyon,
off Virginia Beach, Virginia, N. America (567, 568); Ocean City, Maryland,
N. America (246).
Curcharhinusperezi:
Deep Water Caye, Grand Bahama Is., Bahamas (233, 234, 235);
Glover’s Reef, Dangriga, Belize (179,180).
Curchurhinus plumbeus: Makapu Point, Oahu, Hawaii (555, 556); Maunalua Bay, S.
Oahu, Hawaii (570,571);
Mayport, Jacksonville, Florida, N. America (437).
Curcharhinus pmosus: Manzanilla Bay, eastern Trinidad (495,496,497,
498,499).
Curchurhinus seuki: Turtle Islands, Sulu Sea. (Navotas mkt., Manila), Philippines
(273,274,275,
276,277).
Curchurhinus sorruh: Navotas mkt, Manila, Philippines (origin probably Palawan)
(259, 358); S. China Sea coast of Palawan (Navotas mkt Manila), Philippines
(296,297) ; Turtle Islands, Sulu Sea. (Navotas mkt, Manila), Philippines (300).
316
G.J. P. NAYLOR
Curcharhinus whpelen‘: El Tur, South Sinai (in Straits of Tiran).
side of Tiran Is. at “Melba” Egypt (008).
Galpocerdo ruvieri: Dauphin
Is., Alabama,
Hawaii (552,553,554);
Hemigalew
Ocean
mucrostoma: Turtle
N. America
City, Maryland,
Islands,
(248);
E<gypt (004):
Makapu
N. America
South
Point,
Oahu,
(239).
Sulu Sea. (Navotas mkt. Manila),
Philippines
(314, 315, 316, 317, 318).
microstoma: Visayan Sea. (Cebu City mkt), Philippines
Hemigahs
Loxodon mucrmhinus: Navotas Fish landing,
Sulu Sea. (Navotas mkt, Manila),
Philippines
Neguption breuirosttis: off Cosgrove,
Panama
City, Florida,
Manila,
Florida
N. America
NewJersey,
Rhiqbrionodon
217);
Turtle
N. America
acutus: Freetown,
Islands,
Sulu
(324, 325).
(258) ; Turtle
Islands,
(285, 285).
Keys, Florida,
N. America
(558,
560);
(75).
Prionace gluucu: 19 15 N, 160 49 W. (200
Pleasant,
Philippines
m SSW of Oahu,
Hawaii
(540) ; Point
(416).
(Sierra
Sea,
Fisheries
(Navotas
mkt,
by catch),
Manila),
Sierra
Leone
Philippines
(215.
(272.
312,
313).
Rhiwpionodon
lalundii: Port of Spain fish mkt. Trinidad
(485, 486, 487, 488, 489)
Rhizoptionodon porosus: Port of Spain fish mkt, Trinidad
Rhizoptionodon
N. America
terrunouae: 454’791.
(37.5, 376, 390);
Sphyma huini: American
Is., Alabama,
Panama
Shoals,
N. America
City, Florida,
$Lh~rnu mockawan:
Jacksonville,
Is., Alabama,
Florida
(257);
Dauphin
Is.,
N. America
City.
Charleston,
N. America
Keys, Florida,
Ocean
N. America
Florida,
(LORAN),
60516.3
Dauphin
(511).
N. America
Maryland,
S. Carolilla,
(164).
(539);
Dauphin
N. America
(24.5):
(08 1) .
Alabama,
(443);
N.
America
Panama
(101,
City, Florida,
256):
Mayport.
N. America
(089,
090).
~phymu tihro: Manzanilla
L$dqwzu hdes: Manzanilla
mkt, Trinidad
Bay, eastern
Trinidad
(500, 50 1) .
Bay, eastern
Trinidad
(502,
Sphyrna zygaenu: Ocean
City, N. America
Triaenodon obesus: Between
Island,
503, 504);
Port of’ Spain fish
(49 1) .
Tiran
(N. W. Hawaiian
(244).
and Sinafir
Islands),
Hawaii
Is., Gulf of Aqaba,
(546,
551);
South
Egypt (010);
Nihoa
side of Tiran
Egypt (009).
(Three-digit
numbers
in parentheses
correspond
to assigned
field numbers)
Is.,
PHYLOGENY OF REQUIEM AND HAMMERHEAD SHARK!3
Appendix
2a. Character Matrix Based on Electromorph
Autapomorphic
Presumed locus”
123456789
C. acronotus
C. albimarginalus
6 altimus
EGKFKI
EFMKENBDJ
M
E
I
E
C
J
E
F
Q
E
F
H
E
E
M,J
ELMJ
E
G M
C H M
E
H D
C. amblyrhynrhos
C. brachyuruh
C. bmtipinna
C faln@rnu
C. galapagpnsis
Cz’.isodon
C. leuras
C. limbatus
C. longimanus
(.‘. marloll
(: melanoptenl,
(.A obscuruc
(.‘. p flezi
C plumba,Puc
C. porosus
C. st7ab-i
C: .sorrah
(.Y VJhrPlen
C. cuvien
FL macrostoma
H mirrostomn
I.. macrorhinus
N. brpoiro.~tns
I’. glauca
R. acufu.\
R. lalandii
R. porosus
R. twranovar
.s. kwini
S. mockarran
.S. tibum
$. tudrr
S. zygaeno
7: obessu.r
E.J
D
E
E
A
H
12
MJ
F
K
F
F
E
H F
M
E
I
E
B
N
B
C 0
E
H M
E
C J
F
?
A
HMBFAF
H N B
F
E
L.
F
ERKMI
E
I.
M
F
D C
A D C
A D C
A D C
G H G
G H G
C
K G
G K G
G H G
F
E E
12
13
14
I5
(;
0
A
D
B
If
C
K
I
N
N
K
S
I‘
Ii
K
(:
(:
I’
0
1:
I’
V
\’
X
LV
H
\I’
1
%
%
I.
1:
bl
M
R
1
(:
F
F
F
J
F
H
F
C
(;
F
F
F
F
F
F
E’
F
E
G’
B
B
D
D
D
D
c;
B
G
(;
K
B
E
E
E
C
C
(:
C:
c
(:
(:
<:
F
I
I
I
I
I
I
I
I
H
”
(;
G
(:
D
G
(:
(:
(:
B
0
B
B
L
hl
E
J
B
I
B
H A
<; I.
c:
I.
F
0
B
P
B
P
B
P
B
A N
(1 P
A S
A S
(:
H
X4 E
C
(:
(:
B
C
C
(:
(1
(:
(:
A
.A
<:
c:
I
F
F
F
F
G
c:
E
H
G
C:
I
I
I
(;
I
;
(;
(;
J
‘ii
I
<;
I
I
I
E
E
E
(:
I
1
D
D
D
D
B
F
.4
B
F
I
(;
B
(k
A
D
D
(i
<.
B
C:
B
B
B
B
B
B
B
B
B
B
B
B
B
B
B
B
B
D
<;
I
A
B
B
E
E
F
E
H
H
H
H
H
B
F
D
D
D
F
F
D
D
D
F
D
D
F
F
F
D
D
D
D
G
A
A
H
Z
F
E
E
E
E
<:
B
(:
(:
B
F
J
D
D
D
D
D
D
D
D
D
D
1
D
I>
J
‘D
1)
I)
D
K
H
H
(:
E
A
B
B
B
B
(;
(;
(;
(;
G
F
F
F
F
F
K
I
N
N
N
N
N
D
K
f:
N
B
F
F
(:
B
K
H
H
H
H
B
D
C
C
A
F
J
G
G
G
C:
H
H
H
H
H
E
.J
II
I
I>
K
I
K
K
F
F
K
D
H
I
E
P
K
MJ
10
C
N
N
?
E
E
N
K
N
J
1
Including
D
K
D
K
K
I
I
K
K
F
G
F
F
Differences,
States
N H <: I
(:
I
D
D
(:
(:
G
F
(;
1;
D
(;
“ (I) EST-I; (2) EST-2; (3) GOT-I: (4) GOT-2; (5) LDH-H: (6) ME-l: (7) ME-2 (8) SOD-I:
(10) (YGPD; (11) MDH; (12) ID”: (13) CA; (14) PEP; (15) GDH; (16) GI.0: (17) IAP.
!
b
1)
(:
(:
(:
;
b:
16
I7
:i
c:
(1
H
(:
D
(:
(;
”A
.\
A
11
A
.\
A
.1
:i
(:
E
”
.I
.i
.\
1
.\
‘(:
(:
(’
\
\
-\
(
\
(’
”\
c.
\
\
1
)
\
J
F
\
\
(
,\
(:
.\
(.
\
,\
(.
(:
(:
A
K
.A
:I
A
I
:\
.\
:\
(:
(
1\
1%
\
.\
\
\
,i
4
\
(
\
I;
\
(91 %)1)-L’:
G. J. P. NAYLOR
318
Appendix 2b. Character Matrix Numerically Coded with Autamorphic States Removed
[Final, asymptotic character weights used for successive weighting (Fan-is, 1969) are
@en for each locus]
Presumed locusa
1
c. aLronotus
C. albimqinatus
C. altimus
C. amblyrhynchos
C. brachyurus
C. breoipkzna
C. /al@rmis
C. galupagensis
C. isodon
c. leuca.5
C. limbatus
C. kmgimanus
C. ma&i
C. melunoptews
C. obscwus
C. perezi
C. plumbeus
c. porosus
C. se&i
C. sorrah
C. wheeleri
G. cwieri
H. macrostoma
H. microstoma
L. mucwrhinus
N. brevirostris
P. giuuca
R acutus
R lalandii
R porosus
R terranovae
s. lewini
S. mockarran
s. t&u70
S. tudes
S. zygakma
T. obesus
2
5
?
2
7
4
?
2
5
?
142323
2455261??
3
1
4
275276325?
165?33
2145161253?
4
2
4
?
2
7
?
3
2
5
3
142???
2
4
?
2733253142?
235463?
2
5
2
1 6
2954661253162433
2
5
5
?
752253142?
?
6527622531426?3
2
6
?
2
3
?
2
3
?
3142633
5
4
6
6
1
2
?
253162433
?
?
?
?
?
6125,l
4
2
6
?
5
7
6
?
25314??
295466
1 2
5
3
1
6
2
6
?
4761???
162433
2752763253165133
2
?
?
2
3
?
?
2
5
3
1
4
?
1
?
2352253142?35
265273?
2
5
?
1
4
214516
1 2
5
3
1
4
3???5??
23,?????1
5?12??11
3
2
?
1
5
?
12???1
1
3
2
?
33?2???2?3????
3
3
?
5
?
4
?
2531’???3
2955????5?1
6
32234?1
322251?
1
223471322251?
1
223471321?51821
1223471
1251821
3
2
463?5144
13333?13
4
6
3
?
323??3
5
2
?
4
4
?
4
8
3
1 5
1
4
?
?
3333512
4
8
3
1 5
1
4
?
13333512
4
6
3
?
442323?13
5
2
?
3
3
?
2
2
?
225316???4
2
3
4
5
6
7
8
9
10
11
12
13
14
15 16
2
?
3
3
2
2
3
3
2
1
3
?
3
?
2
?
?
5
3
2
?
2
4
3
3
3
2
3
3
3
2
2
?
2
3
3
3
3
4
4
7
7
2
2
2
6
6
4
?
?
?
2
2
3
1
1
a (1) EST-l, wt = 1000; (2) EST-2, wt = 489; (3) GOT-l, wt = 1000: (4) GOT-2, wt = 564; (5) LDH-H,
wt = 533; (6) ME-l, wt = 762; (7) ME-2, wt = 467; (8) SOD-l, wt = 643; (9) SOD-2, wt = 1000; (10) aGPD,
wt = 333; (11) MDH, wt = 1000; (12) IDH, wt = 778; (13) CA, wt = 1000; (14) PEP, wt = 1000: (15) GDH,
wt = 1000; (16) GLO, wt = 729.