A phylogeny of extant and fossil buckler dory ishes,
family Zeidae (Zeiformes, Acanthomorpha)
by
Francesco Santini (1)(5), James C. tyler (2),
alexandre F. Bannikov (3) & Dorin-Sorin BaCiu (4)
AbsTrAcT. - a data set of 45 putatively informative morphological characters (mostly osteological, and a few external
features) is analyzed for 12 extant and fossil species belonging to the buckler dory ish genera Zeus and Zenopsis, which
together constitute the family Zeidae (Zeiformes, acanthomorpha), and for two outgroup taxa. Zeus consists of two extant
and three fossil species, whereas Zenopsis consists of four extant and three fossil species. Both genera date back to at least
the oligocene (36 Mya). the phylogenetic analysis of the full data set (extant + fossil) provides strong support for the
monophyly of the Zeidae but only weak support for the monophyly of Zenopsis, and it calls into question the existence of a
clade formed by the extant and fossil taxa assigned to Zeus. additional phylogenetic analyses were performed: (1) the most
incomplete fossil taxon, Zeus robustus, was excluded; (2) all fossil taxa were excluded; and (3) the new extant species of
Zenopsis was excluded. All of these analyses conirm a strongly supported clade formed by Zeus + Zenopsis and of a much
less strongly supported clade formed by the extant and fossil species of Zenopsis. the analyses do not provide support for a
clade formed by extant and fossil species previously assigned to the genus Zeus, even though when the fossil species are
excluded from the analysis, the two extant species of Zeus appear as each other’s sister groups in one of three equally parsimonious cladograms.
résumé. - une hypothèse phylogénétique pour les Zeidae actuels et fossiles (Zeiformes, acanthomorpha).
une matrice des 45 caractères morphologiques (principalement ostéologiques) a été constituée pour 12 espèces actuelles et fossiles des deux genres de Zeidae, Zeus et Zenopsis et pour deux extra-groupes. le genre Zeus comprend deux espèces actuelles et trois fossiles, et le genre Zenopsis en comprend quatre et trois respectivement. les deux genres datent de
l’oligocène moyen (36 millions d’années). l’analyse phylogénétique du jeu des données de toutes les espèces supporte
bien la monophylie des Zeidae, mais très peu celle de Zenopsis ; le clade formé par les espèces actuelles et fossiles du genre
Zeus n’est toutefois pas soutenu. Des analyses complémentaires ont été effectuées, après avoir éliminé soit Zeus robustus,
le fossile le moins connu, soit tous les fossiles, soit la nouvelle espèce de Zenopsis. toutes ces analyses soutiennent l’existence d’un clade des Zeidae et d’un clade Zenopsis. il n’y a pas de support pour un clade Zeus.
key words. - Zeiformes - Zeidae - Fossils - osteology - Phylogeny.
Zeidae (buckler dories, including the John dory or St.
Peter’s fish) are moderate- to deep-bodied acanthomorph
ishes. The family is composed of 12 species (six extant and
six extinct) that are presently classiied in two genera. The
genus Zeus linnaeus, 1758, includes the extant species Zeus
faber linnaeus, 1758 and Zeus capensis valenciennes, 1835,
and the extinct species Zeus robustus Gorjanovi¢kramberger, 1891 (oligocene of Slovenia), Zeus jerzmanskae Baciu et al., 2005 (oligocene of Poland) and Zeus primaevus Scarabelli, 1859 (Miocene of italy and algeria). the
genus Zenopsis Gill, 1863, includes the extant species Zenopsis nebulosus (Schlegel, 1847), Zenopsis conchifer (lowe,
1852), Zenopsis oblongus Parin, 1989, and Zenopsis sp. (to
be described by u. yamada, t. nakabo and D. Bray – D.
Bray, pers. comm.), and the extinct species Zenopsis hoernesi Gorjanovi¢-kramberger, 1891 (oligocene of Slovenia),
Zenopsis clarus Daniltshenko, 1960 [oligocene of the Carpathians (Poland, romania) and northwestern Caucasus
(russia)] and Zenopsis tyleri Baciu & Bannikov, 2001 (oligocene of romania).
The irst comprehensive phylogenetic hypothesis for the
order Zeiformes (dories) (tyler et al., 2003) showed that the
extant Zeiformes form a strongly supported monophyletic
group composed of six clades of familial status: the Cyttidae, oreosomatidae, Parazenidae, Zeniontidae, Grammicolepididae, and Zeidae, with this last family appearing in
the terminal, most-derived clade of dories, as the sister group
to the Grammicolepididae. However, none of the extinct taxa
(1) Département Systématique et Évolution, uMr 7138, Muséum national d’Histoire naturelle, 43 rue Cuvier, 752311 Paris CeDex 05,
FranCe.
(2) national Museum of natural History, Smithsonian institution, Washington, DC 20560-0106, uSa. [tyler.jim@nMnH.Si.eDu]
(3) Paleontological institute, russia academy of Sciences, Profsoyuznaya 123, Moscow 117997, ruSSia. [aban@paleo.ru]
(4) Muzeul de Stiinte ale naturii, Str. Petru rares nr. 26, 5600 Piatra neámt, roMania. [dsbaciu@ambra.ro]
(5) Present address : Department of Zoology, university of toronto, 25 Harbord Street, toronto, ontario M5S3G5, CanaDa.
[francesco.santini@utoronto.ca]
Cybium 2006, 30(2): 99-107.
Santini et al.
Phylogeny of Zeidae
table i. - list of specimens of extant species examined for this work. For fossil materials, see Baciu et al. (2005). [liste des spécimens d’espèces actuelles examinés dans
cet article. Pour les matériels fossiles, voir Baciu et al. (2005)].
taxons
Museum number
CaS 38409
CaS 76856
uSnM 187864
uSnM 307305
Xenolepidichthys
aMnH 29455
CaS 38403
CaS 38406
uSnM 320013
uSnM 320015
uSnM 320016
Zenopsis conchifer CaS 47401
FMnH 67090
uSnM 117280
aMnH 4451
aMnH 56447
aMnH 56833
Zenopsis nebulosus aMnH 92291
aMnH 95024
aMnH 95028
Zenopsis oblongus* uSnM 285048 (paratype)
uSnM 353898
uncatalogued specimen from
Shirshov institute of
oceanology
Zenopsis sp.
aMS.i 22825-019
aMS.i 22826-004 (paratypes)
Faku 64803 (holotype)
Faku 64804 (paratype)
Zenion hololepis
number of
specimens
Sl (mm)
in lot
1
90
2
70-73
3
50-73
2
45-48
4
70-75
1
70
1
75
1
64
3
59-77
2
59-64
1
54
3
44-58
2
51-81
1
126, D
1
368, D
1
405, D
1
325, D
1
335, D
1
340, D
1
272 r
347 (CS, in poor
1
condition)
1
1
2
1
1
255 r
r
r
r
r
173-193,
r (1 CS,
Zeus capensis
uSnM 330849
8
in poor condition)
Zeus faber
uSnM 307842
2
48-67
uSnM 320014
2
59-89
uSnM 320063
1
80
aMnH 22707
1
300, D
aMnH 91448
1
230, D
aMnH 95055
1
270, D
uSnM 176975
1
215, D
223, r
uSnM 328597
1
*this species is unique among Zeidae in having the third interneural space vacant
(rather than the fourth) and only three dorsal-fin pterygiophores anterior to the neural
spine of the fourth abdominal vertebra (rather than four pterygiophores).
and only three of the six known extant species of zeid ishes
(Zeus faber, Zenopsis conchifer, and Zenopsis nebulosus)
were included in that (mostly osteological) analysis because
of the lack of suficient material. Baciu et al. (2005) have
since published a revision of the fossil record of the Zeidae,
including the description of new species and detailed redescription of the osteology of several others. the availability
of these new data for the fossil zeids, in addition to the fact
that osteological material for the least known extant zeid
100
taxa has become available for the irst time,
makes it possible to produce the irst phylogenetic hypothesis for the Zeidae that
includes all of the species currently assigned
to this family.
mATErIALs AND mETHODs
Fourteen taxa are included in this analysis, including 12 zeids and two outgroups.
in addition to the zeid taxa and the two outgroups already analyzed in tyler et al.
(2003), three additional extant zeid species
(Zeus capensis, Zenopsis oblongus, Zenopsis sp.) are included. the six extinct taxa of
Zeidae included in this analysis were
described with reconstructions in Baciu et
al. (2005); the reader is referred to that
paper for the descriptions and measurements of those species. Minor differences
from the data in that paper are based upon
our re-interpretation of the number of vertebral segments or anal pterygiophores in a
few specimens. table i lists all examined
species, and for extant species it includes
museum number, number of specimens in
each lot, and standard length (Sl) when
available. Museum abbreviations follow
leviton et al. (1985).
Phylogenetic analysis
external morphological and osteological characters were obtained from direct
observation of the specimens listed in table
i; the character list is reported in appendix
i. the characters were analyzed following
the principles of phylogenetic systematics
(kitching et al., 2000). a matrix for all the
specimens examined was irst constructed
using WinClaDa (nixon, 2002). this
matrix was subsequently analyzed using
nona (Goloboff, 1999) and is shown in
table ii. Following tyler et al. (2003) one
species of Zeniontidae and one of Grammicolepididae were used as outgroups. the zeniontid Zenion
was selected as the first outgroup, following the protocol
suggested by nixon and Carpenter (1993). all characters
were assigned equal weight (1), and all multistate characters
were analyzed as unordered. Heuristic searches, with random addition of taxa, the tBr + tBr branch swapping
option of nona, and 10,000 replications were performed.
tree length (l), consistency index (Ci), and retention index
Cybium 2006, 30(2)
Santini et al.
Phylogeny of Zeidae
table ii. - Data set of 45 morphological characters for the 14 species in this analysis (12 Zeidae plus two outgroups). numbers in parentheses at the right side of the table indicate the number of characters that could be determined, when some characters are unknown for a taxon.
[Jeu de données de 45 caractères morphologiques pour les 14 espèces utilisées dans cette analyse. les nombres entre parenthèses à droite
du tableau indiquent le nombre de caractères qui ont pu être déterminés dans les cas où quelques-uns d’entre eux restent inconnus pour un
taxon.]
Character number and state
Taxon
5
10
15
20
25
30
35
40
45
Zenion
Xenolepidichthys
00000 00000 00000 00000 00000 00000 000-- 00000 00000
00011 00022 10100 00000 00000 00000 020-- 03021 01010
Zeus
Zeus
Zeus
Zeus
Zeus
11110
11110
?????
?0???
?1?1?
faber
capensis
robustus
jerzmanskae
primaevus
11111
11111
?????
?1??1
??1?1
01101
00101
???0?
1??01
???11
Zenopsis conchifer 1p122 11112 01111
Zenopsis nebulosus 1p122 11112 01111
Zenopsis oblongus 10?22 11111 11111
Zenopsis sp.
11?2? ?1??1 ?1?1?
Zenopsis clarus
?1?2? ?11?1 10?11
Zenopsis tyleri
?1??? ?11?1 1??11
Zenopsis hoernesi ????? ????? ????1
Zeus faber : 39(0,1)
Zeus robustus : 37(0,1)
Zenopsis conchifer : 2(0,1); 39(0,1);
Zenopsis nebulosus : 2(0,1)
Zenopsis oblongus : 42(1,2)
Zenopsis clarus : 41(1,2)
11111
11110
?1?10
???10
?1?11
11120
11120
1112?
11112
11111
11101
11101
?1???
11?0?
11?0?
11101
11101
?????
11101
11101
112p1
11201
?p11?
11301
11311
22211
22210
112??
122??
122??
(15)
(29)
(30)
11011
11111
11111
?1110
?1?11
?1?11
?1???
11110
11111
10110
11110
11112
11112
011??
11111
11111
11?11
?1?1?
11?1?
11?1?
?1???
13310
13310
13310
1?310
13210
13210
132?0
122p1
12311
12201
?1211
12301
12301
122??
p2122
12122
1p122
0212?
p21?2
221?2
11???
(30)
(34)
(32)
(15)
41(1,2)
(ri) are provided for each analysis (kluge and Farris, 1969;
Farris, 1989). When more than one most-parsimonious tree
was obtained, a strict consensus tree was calculated. Character evolution (appendix ii) was studied using the delayed
transformation (Deltran) option of WinClaDa,
because most characters were scored as unknown (“?”) for at
least some taxa, and the use of the accelerated transformation (aCCtran) option would have necessitated hypothesizing the presence of certain character states within lineages, for which there is no evidence that these states have ever
been present. the decay index (Bremer, 1988, 1994) was
calculated using nona (Goloboff, 1999). Cladograms for
publication were produced using treevieW (Page, 1996).
unknown character states in the fossil taxa are indicated
with a question mark “?”. inapplicable characters are indicated with a horizontal dash “-”. it should be remembered
that although tree-building programs treat dashes and question marks in the same way, they are conceptually different.
rEsuLTs
analysis of the full data set produces two equally parsimonious trees (ePts hereafter) (Fig. 1). the topology of the
resultant strict consensus tree strongly supports the monoCybium 2006, 30(2)
phyly of the Zeidae, with a high decay index of nine, and
illustrates the existence of a clade formed by the extant and
extinct species of Zenopsis, whereas Zeus appears to be paraphyletic. Furthermore, the two extant species of Zeus, Z.
faber and Z. capensis, never appear to form a monophyletic
group, and Z. capensis appears to be a more basal taxon than
Z. faber. the relationships of the oligocene Zeus robustus,
whose fossil record is based upon highly incomplete materials, appear to be problematic because its placement is significantly variable among the cladograms in igure 1. The relationships for the species of Zenopsis are stable despite the
weak decay index for all groups within the Zenopsis clade.
This low support might be a relection of the presence of fossil taxa, which have many unknown character states (see also
Santini and tyler, 2004). in the Zenopsis clade, the yet-tobe-described species Zenopsis sp. appears as the most basal
lineage. Subsequently, two subclades can be identiied: one
formed by the extant Zenopsis conchifer and Zenopsis nebulosus; and one formed by Zenopsis oblongus plus the three
extinct species. Within this last clade, Zenopsis oblongus is
the sister lineage to Zenopsis hoernesi + (Zenopsis clarus,
Zenopsis tyleri).
exclusion of the highly incomplete Zeus robustus from
the data set does not help to resolve the relationships among
the various species assigned to Zeus. on the contrary, it
101
Phylogeny of Zeidae
Santini et al.
Figure 1. - equally parsimonious trees
(ePts) 1 and 2 (l = 85, Ci = 0.76, ri =
0.80), and strict consensus (SC) tree
produced by the analysis of the full data
set of 12 extant and extinct species of
Zeidae plus two outgroups. Decay
index is indicated above the tree
branches of the SC tree. Cladogram #2
was selected for character optimization
in igure 5. [ePts 1 et 2 (l = 85, CI =
0,7,6 RI = 0,80), et arbre de consensus
strict (SC) produits par l’analyse de
toutes les informations pour 12 espèces
de Zeidae vivants et fossiles, et deux
extra-groupes. l’indice de Bremer est
indiqué sur les branches de l’arbre SC.
le deuxième cladogramme a été utilisé
pour l’optimisation des caractères dans
la igure 5.]
Figure 2. - Strict consensus tree of the ive EPTs (L = 0.83, CI =
0.78, ri = 0.82) produced by the analysis of the data set of extant
and extinct species of Zeidae plus two outgroups when Zeus robustus is not included. Decay index is indicated above the tree branches. [arbre de consensus strict des 5 ePts (l = 0,83, CI = 0,78, RI =
0,82) produit par l’analyse du jeu de données contenant les espèces
actuelles et fossiles de Zeidae et deux extra-groupes, à l’exclusion
de Zeus robustus. l’indice de Bremer est indiqué sur les branches
de l’arbre.]
102
increases the number of EPTs to ive. The strict consensus
tree (Fig. 2) shows that although both the monophyly of the
Zeidae and of Zenopsis are still supported, with a decay
index of 11 and 2 respectively, within Zenopsis only the
clade of Z. conchifer and Z. nebulosus and the clade of Z.
clarus and Z. tyleri are recovered, albeit with a low decay
index of 1. these last two clades appear in a polytomy with
the other three species of Zenopsis.
When only the extant taxa are included in the analyzed
data set, three ePts are recovered (Fig. 3). the consensus
tree supports the monophyly of Zenopsis, even though the
relationships among the various taxa are less well resolved
than in the analysis with the fossils, whereas Zeus still appears
as paraphyletic. Within Zenopsis, Z. sp. and Z. oblongus
appear in a polytomy with the Z. nebulosus + Z. conchifer
clade. the support for the monophyly of the Zeidae is very
high (decay index of 19), whereas that of the monophyly of
Cybium 2006, 30(2)
Santini et al.
Phylogeny of Zeidae
Figure 3. - ePts 1 through 3 (l = 0.70,
Ci = 0.92, ri = 0.95) and the strict consensus tree (SC) produced by the analysis of the data set of six extant species
of Zeidae plus two outgroups only.
Decay index is indicated above the tree
branches of the SC tree. [ePts 1-3 (l =
0,70, CI = 0,92, RI = 0,95) et arbre de
consensus strict (SC) produits par
l’analyse du jeu de données de six
espèces actuelles de Zeidae et seulement deux extra-groupes. l’indice de
Bremer est indiqué sur les branches de
l’arbre SC.]
exclusion from the analysis of Zenopsis sp., with many
unknown internal characters, recovers one ePts (Fig. 4). in
this case, the monophyly of the extant Zeus is again not
recovered because Zeus capensis appears as the most basal
zeid, whereas Zeus faber is the sister group to Zenopsis.
DIscussION
Figure 4. - Most parsimonious tree (l = 0.67, Ci = 0.97, ri = 0.98),
produced after analysis of the data set of extant species of Zeidae
plus two outgroups, when Zenopsis sp. is not included. Decay index
is indicated above the tree branches. [arbre le plus parcimonieux
(l = 0,67, CI = 0,97, RI = 0,98) produit de l’analyse du jeu de données des espèces actuelles de Zeidae avec deux extragroupes, à
l’exclusion de Zenopsis sp. l’indice de Bremer est indiqué sur les
branches de l’arbre.]
Zenopsis is relatively good (decay index of 3, which corresponds to slightly more than 4% of the tree length).
Cybium 2006, 30(2)
Phylogenetic analysis
one result of this work is that the inclusion of even very
incomplete fossil materials does not prevent the inference of
a phylogeny when a phylogenetic signal is present in the data
set (a conclusion already supported by Santini and tyler,
2003, 2004; tyler and Santini, 2005). the monophyly of the
Zeidae and of Zenopsis are recovered in all analyses with
and without the inclusion of the fossils. also, the relationships among the extant species of Zenopsis do not change
103
Phylogeny of Zeidae
Santini et al.
Figure 5. - EPT 2 from igure 1 selected
for the study of character evolution.
letters below branches correspond to
letters in the character evolution section
in appendix ii. [ePt 2 de la figure 1
utilisé pour l’étude de l’évolution des
caractères. les lettres au-dessous des
branches correspondent aux lettres
dans la partie de l’appendice II qui
traite de l’évolution des caractères.]
when the fossils are included in the analysis. the situation is
different for Zeus, in which the monophyly of the group is
not recovered in the analyses when the fossils are included.
When only the extant species of Zeus are analyzed, the
monophyly of the group is supported, but with a weak decay
index (see following section on classification for further
comments).
a second result is that according to the topology of the
trees recovered, the radiation of the Zeidae is likely much
older than previously thought. the fact that the three oligocene species of Zenopsis appear to be very derived taxa
within the Zenopsis clade indicates that the radiation of the
various lineages in this group had already occurred by
35 Mya.
Classiication
the two extant species of Zeus appear to form a monophyletic group in only one of the trees produced by the analyses of extant species alone (Fig. 3). a similar result is produced when the fossils are included in the analyses. the
present results do not allow us to recover a monophyletic
Zeus, a genus that was irst created in 1758 by Linnaeus and
has been recognized as valid ever since. Because in one of
the analyses the two extant species appear to be each other’s
sister groups, and because the decay index for the relationships among most species of Zeus is always very weak, we
prefer to retain the current generic classification. We thus
continue to recognize the generic name Zeus as a valid taxon
pending further work, which either may reveal support for
the existence of a clade formed by all fossil and extant Zeus
or may ind increased support for the paraphyly of this group
and, hence, the recognition of additional generic categories.
104
Paleobiogeography of the Zeidae
the only putative eocene zeid was originally described
from the tertiary of Georgia as Platax (?) colchicus by Simonovich et al. (1875); however, Bogatshov (1933) concluded
that this ish instead belongs to Zeus and he noted that the
marls in Georgia in which this species was found are of
eocene age. the two type specimens cannot be located.
Danilchenko (1960) mentioned the opinion of Bogatshov
and hypothesized that the Georgian ish is in fact a Zenopsis.
Baciu et al. (2005) considered this taxon to be a Zenopsis
nomen dubium; however, on the basis of the phylogeny presented herein and on the other known fossil zeids, we think it
is likely that the zeids irst appeared in the Eocene. Five of
the six additional fossil zeid species are in fact distributed in
the oligocene of the Central Paratethys (Slovenia, Poland,
romania) and eastern Paratethys (Caucasus), with several
of these species from the rupelian, the oldest part of the oligocene (approximately 30-36 Mya). Zeus primaevus, the
single species from the late Miocene (5-11 Mya), is distributed only in the Mediterranean basin (algeria, italy, Spain).
the extant zeid species occur in the atlantic, indian, and
Pacific oceans and are mostly distributed throughout the
coastal waters of the continents, with only Zeus faber occurring in the Mediterranean and Black Seas.
Most of the oligocene zeids are geographically distributed in a single location, probably corresponding to isolated
basins. For example, Zeus robustus is known only from Slovenia; Zeus jerzmanskae from Poland; Zenopsis hoernesi
from Slovenia; and Zenopsis tyleri from romania. the only
exception is Zenopsis clarus, which is widely distributed
throughout the Paratethys, with fossil specimens known
from russia, romania, and Poland. the upper Miocene
Cybium 2006, 30(2)
Santini et al.
Phylogeny of Zeidae
Zeus primaevus is known from deposits in italy, oran (algeria), and Spain (Baciu et al., 2005), and its more expansive
distribution might indicate that the ichthyofauna had by then
been homogenized across the Mediterranean region. the
recent species Zeus faber was described by Bassani (1905)
from the Pleistocene deposits of taranto (southern italy). We
have not examined materials from taranto, but it is possible
that these Pleistocene specimens are referable to the extant
Zeus faber, which likely belongs to a phylogenetically very
ancient lineage. no fossils are known for any of the other
extant species of Zeidae.
Acknowledgements. - Many individuals, as listed in Baciu et al.
(2005), made the materials of their institutions available to us for
examination, and we thank them all. We additionally thank n.v.
Parin, Shirshov institute of oceanology, Moscow, who donated rare
specimens of Zenopsis oblongus to the uSnM collection; D. Bray,
Museum of victoria, and t. nakabo, kyoto university Museum,
for radiographs of the new species of Zenopsis (herein referred to
as Zenopsis sp.). S. Zehren, university of alabama, provided literature on zeiforms. D. Goujet and M. verán, MnHn, Paris, helped
arrange for a visit to the paleoichthyological collections of the
MnHn. our work in Paris was also facilitated by loaned materials
and radiographs sent by S. raredon, l. Palmer, and J. Williams,
Smithsonian institution, Washington, D.C.; n.n. Parin, international academic agency “nauka,” Moscow; and n. Micklich, Hessisches landesmuseum, Darmstadt. M. Hautecœur, MnHn, radiographed some specimens of Zeidae on loan to FS.
excavations in the east Carpathians in 2002 by D.-S. Baciu, during which important specimens of Zenopsis were discovered, were
funded by grant no. 7312-02 from the national Geographic Society,
Washington, D.C.
our joint work on Zeiformes has been made possible by two
Short term visitor Fellowships awarded to D.-S. Baciu and a.
Bannikov by the Smithsonian institution for collaborative research
with J.C. tyler, and by a nato life Science and technology collaborative linkage grant (lSt. ClG. 978836) to J. tyler, a. Bannikov, D.-S. Baciu, and F. Santini. F. Santini was supported by a
Marie Curie Fellowship for a project on “Paleontological and
molecular approaches to the phylogeny of acanthomorpha
(Pisces).”
this manuscript was greatly improved by the comments of D.
tyler, n. Micklich, G. lecointre, who also helped with the French
translation of the abstract, and two anonymous reviewers.
rEFErENcEs
BaCiu D.-S., Bannikov a.F. & J.C. tyler, 2005. - revision
of the fossil ishes of the family Zeidae (Zeiformes). Boll. Mus.
Civ. Storia Nat. Verona (Geologia, Paleontologia, Preistoria),
29: 95-118.
BaSSani F., 1905. - la ittiofauna delle argille marnose plistoceniche di taranto e di nardò (terra d’otranto). atti R. accad. Sci.
Fis. Mat. Napoli, ser. 2, 12: 1-60.
BoGatSHov v.v., 1933. - Data on the tertiary Fish Fauna of the
Caucasus. tr. azerb. Neft. Issled. Inst., Geol. Otd., 15: 1-62. [in
russian]
Cybium 2006, 30(2)
BreMer k., 1988. - the limits of amino acid sequence data in
angiosperm phylogenetic reconstruction. evolution, 42: 795803.
BreMer k., 1994. - Branch support and tree stability. Cladistics,
10: 295-304.
DANILCHENKO P.G., 1960. - Bony ishes of Maikop deposits of
the Caucasus. tr. Paleontol. Inst. akad. Nauk SSSR, 78: 1-208.
[in russian]
FarriS J.S., 1989. - the retention index and rescaled consistency
index. Cladistics, 5: 417-419.
GoloBoFF P., 1999. - nona (no naMe) version 2. tucumán,
argentina: published by the author. available via http://www.
cladistics.com.
kitCHinG i.J., Forey P.l., HuMPHrieS C.J. & D.M. WilliaMS, 2000. - Cladistics, 2 nd edit. 228 p. oxford: oxford
university Press.
kluGe a.G. & J.S. FarriS, 1969. - Quantitative phyletics and
the evolution of anurans. Syst. Zool., 18: 1-32.
leviton a.e., GiBBS, r.H., Heal e. & C.e. DaWSon, 1985.
- Standards in herpetology and ichthyology: Part i. Standard
symbolic codes for institutional resource collections in herpetology and ichthyology. Copeia, 3: 802-832.
MaDDiSon W.P., 1989. - reconstructing character evolution on
polytomous cladograms. Cladistics, 5: 365-377.
nixon k.C., 2002. - WinClaDa ver. 1.00.08. ithaca, ny: published by the author. available via http://www.cladistics.com.
nixon k.C. & J.M. CarPenter, 1993. - on outgroups. Cladistics, 9: 413-426.
PaGe r.D., 1996. - treevieW: an application to display phylogenetic trees on personal computers. Comp. appl. Biosc., 12:
357-358.
Santini F. & J.C. tyler, 2003. - a phylogeny of the families of
fossil and extant Tetraodontiform ishes (Acanthomorpha, Tetraodontiformes), upper Cretaceous to recent. Zool. J. linn.
Soc., 139: 565-617.
Santini F. & J.C. tyler, 2004. - the importance of even highly
incomplete fossil taxa in reconstructing the phylogenetic relationships of the tetraodontiformes (acanthomorpha: Pisces).
Int. Comp. Biol., 44: 349-357.
SiMonoviCH S., BatSeviCH l. & a. Sorokin, 1875. - Geologic Description of Parts of the kutaisi, lechkhumi, Zugdidi
and Senaki Districts of kutaisi Province. 191 p. Tilis: Materialy dlya geologii kavkaza. [in russian]
tyler J.C. & F. Santini, 2005. - a phylogeny of the fossil and
extant zeiform-like ishes, Upper Cretaceous to Recent, with
comments on the putative zeomorph clade (acanthomorpha).
Zool. Scripta, 34:157-175.
tyler J.C., o’toole B. & r. WinterBotttoM, 2003. Phylogeny of the genera and families of Zeiform ishes, with
comments on their relationships with tetraodontiforms and
caproids. Smithson. Contrib. Zool., 618: 1-110.
Reçu le 10 mai 2005.
accepté pour publication le 10 septembre 2005.
105
Santini et al.
Phylogeny of Zeidae
Appendix I
character list
the character list is modified and reduced from tyler et al.
(2003) to compensate for the different taxonomic sampling. two
new characters, numbers 34 and 35 in our list, and some additional
character states have been added. Characters are arranged according to body region.
cranial
1. Basisphenoid: present (0); absent (1).
2. Frontal, supraocular serrations: present (0); absent (1).
3. otolith, shape: moderate to large size, rounded or slightly to
deeply indented on one or both sides, or oblong with humps
(0); tiny, trilobed (bow-tie shaped) (1).
4. lachrymal, size/depth: large, deep, height about one to four
times in the length (0); moderate, height about ive to seven
times in the length (1); slender (2).
5. infraorbitals, number of (well-developed elements exclusive
of the lachrymal, dermosphenotic, and of variable rudiments):
ive to eight (0); four or less (1); nine or more (2).
6. Dermosphenotic: a distinctly separate ossiication from the
sphenotic, sometimes relatively free from the skull (0); fused
or highly consolidated with the sphenotic (1).
7. Premaxilla, alveolar process: ventrally rounded or moderately
indented to form a pair of blunt lobes (0); deeply bifurcated
ventrally (1).
8. Symplectic, ventral lange: present (0); absent (1).
9. Dentary, cartilages (on lateral surface of dentary): two cartilages of moderate size, lying sequentially one behind the
other, of about the same size or the irst only slightly shorter
than the second (0); two cartilages of moderate size, each
attached anteriorly to the dentary and lying sequentially one
behind the other, the irst shorter than the second (1); absent
or unconsolidated (2).
10. Dentary, serrations on the lower border of: multiple serrations
behind the symphysis (0); a single barb near the symphysis
(1); none (2).
11. Ceratohyal, notches on the lower border of: prominent notches for the heads of some of the branchiostegal rays in the anterior group (0); no prominent notches (1).
12. Ceratohyal-epihyal articulation: exclusively through cartilage
(0); with bony interdigitated articulations, at least in specimens of large size (1).
13. epihyal, depth of the anterior end of: equal, or about equal, to
the depth of the adjacent part of the ceratohyal (0); distinctly
less deep than the adjacent part of the ceratohyal (1).
Vertebral column and median ins
14. First vertebra in the caudal peduncle with a modiied neural or
haemal spine: third preural centrum, Pu3 (0); second preural
centrum, Pu2 (1).
15. First vertebra, dorsal extension of the neural spine when the
neural arch and spine are plastered to the skull: the neural
spine with a long dorsal portion free from the skull beyond
the curvature of the supraoccipital and the exoccipitals (0);
the neural spine extending only slightly, or not at all, dorsally
above its attachment to the skull (1).
16. Baudelot’s ligament, placement of the proximal attachment
of: to the irst vertebra (0); to the exoccipital (1).
17. Ossiied ribs: present only on the last few abdominal vertebrae (0); present on most of the abdominal vertebrae behind
the fourth (1).
106
18. Ossiied epineurals: present on most of the abdominal vertebrae or their ribs (0); present on only a few of the anterior
abdominal vertebrae (1).
19. epurals, number: two (0); one (1).
20. Pu2, extra-caudal ossicle in the haemal spine of: absent (0);
present, in at least some specimens (1).
21. vacant interneural spaces, number of groups of (when two or
more spaces are vacant): two (0); three or four (1).
22. Dorsal-in pterygiophores, number of anterior to the neural
spine of the fourth abdominal vertebra: two or three (0); four
(1).
23. Supraneurals, number of: one (0); none (1).
24. Second anal-in spine, length of: very short, much less than
one-half the length of the first spine (0); moderate to long,
more than one-half the length of the irst spine to almost as
long (1); longer than irst spine (2).
25. anal-fin pterygiophores, number of anterior to the haemal
spine of the third caudal vertebra: ive or six (0); seven (1);
eight or nine (2).
Paired-in girdles
26. Supracleithrum, ventral end of: simple (0); deeply bifurcate
(1).
27. Cleithrum, posterior edge: without a posterodorsal prong
above the articulation with the postcleithrum (0); cleithral
process present as a prong above the articulation with the
postcleithrum (1).
28. extrascapulars: one long bone, sometimes forming an open
tube, more or less closely held to the skull (0); two tubular
bones, not closely held to the skull, except at large specimen
sizes (1).
29. Pelvic-in spines: present (0); absent (1).
30. Pelvic-in rays, anterolateral processes of the medial (lower)
surfaces of: absent (0); present as prongs from the medial surfaces of the ray bases (1).
31. Pelvis, posterior process of behind pelvic-in base: short to
moderate in length, and in shape a moderate to broad plate or
lattened shaft (0); long and rod-like, moderately separated
from its opposite member along the midline (1).
scales
32. Scales, on most of the body: moderate to small, spiny
“ctenoid” (spinoid) (0); moderate to small, cycloid (1); greatly elongate vertically (2); absent (excluding enlarged bucklerlike scales), or only in the lateral line (3).
33. Scales, buckler-like (greatly enlarged midline scales): absent
(0); present midabdominally and from the far rear end of the
spinous dorsal in (from no further forward than the last dorsal spine) to the end of the soft dorsal-fin base (1); present
midabdominally and from the posterior region of the spinous
dorsal in (from under the last two or three dorsal spines) to
the end of the soft dorsal-in base (2); present midabdominally
and from the front to middle regions of the spinous dorsal in
to the end of the soft dorsal-in base (3).
34. Buckler-like scales, in addition to major spiny process, accessory spiny process: present (0); absent (1); not applicable,
when buckler-like scales absent (-).
35. Buckler-like scales, radiating striations: present (0); absent
(1); not applicable, when buckler-like scales absent (-).
36. Scales, along the bases of the dorsal- and anal-in rays: absent
along the bases of the rays, but spiny processes present on the
scales alongside the lateral expansions of the distal ends of
the dorsal- and anal-in pterygiophores (0); absent from along
Cybium 2006, 30(2)
Santini et al.
Phylogeny of Zeidae
the bases of the rays, and the scales nearby without spiny projections and not extending beyond the lateral expansions of
the distal ends of the dorsal- and anal-in pterygiophores (1).
meristic data
37. vertebrae, total number of: 27 or 28 (0); 29 to 32 (1); 33 to 36
(2); 37 or 38 (3).
38. abdominal vertebrae, number of: ten or 11 (0); 13 (1); 14 (2);
15 (3).
39. vertebrae, number of in the caudal peduncle (posterior to the
last vertebra whose neural or haemal spine supports a pterygiophore): ive (0); three or four (1); eight (2).
40. Procurrent caudal-in rays, number of (including the number
in both the dorsal and ventral sides, if different): two or three
(0); one (1).
41. Dorsal-in spines, number of: six or seven (0); eight or nine
(1); ten or more (2).
42. vacant interneural spaces, total number of below the spiny
and anterior part of the soft dorsal-in base, posterior to the
irst dorsal-in pterygiophore: two (0); four (1); ive (2).
43. Anal-in spines, number of: two (0); three (1); four (2).
44. Pectoral-in rays, number of: 15 or 16 (0); 13 or 14 (1); 11 or
12 (2).
45. Pelvic-in elements, total number of: seven (0); eight (1); six
(2).
Appendix II
character evolution
The dificulties produced by the use of consensus trees in the
study of character evolution are well known (e.g., see Maddison,
1989). in order to investigate the evolution of the various characters, we selected one of the two ePts obtained from the analysis of
the full data set. We selected ePt number 2 (Fig. 1) because in this
cladogram Zeus robustus, one of the oldest fossil zeids, is represented as being a stem zeid, and we judge this interpretation of the
data to be more reliable, on the basis of the stratigraphic criterion,
than any of the others.
The selected tree is shown in igure 5. All internodes have been
labeled with letters in order to more easily list the character states
that support the various clades. the character optimization was performed using Deltran.
A: 4(0→1) lachrymal moderate, height about five to seven
times in the length, convergent in Xenolepidichthys; 13(0→1) anterior end of epihyal distinctly less deep than adjacent part of ceratohyal, convergent in Xenolepidichthys; 17(0→1) ossiied ribs present on most abdominal vertebrae behind fourth; 19(0→1) one
epural; 21(0→1) three or four groups of vacant interneural spaces;
22(0→1) four dorsal-in pterygiophores anterior to neural spine of
fourth abdominal vertebra; 23(0→1) no supraneural; 24(0→2) second anal-in spine longer than irst spine; 27(0→1) cleithral process
present as prong above articulation with postcleithrum; 40(0→1)
one procurrent caudal-fin ray, convergent in Xenolepidichthys;
41(0→1) eight or nine dorsal-fin spines; 42(0→1) four vacant
interneural spaces, convergent in Xenolepidichthys and in Zenopsis
hoernesi; 43(0→2) four anal-in spines; 44(0→1) 13 or 14 pectoralin rays, convergent in Xenolepidichthys.
b: 7(0→1) alveolar process of premaxilla deeply bifurcated
ventrally; 10(0→1), serration on lower border of dentary consists
of a single barb near the symphysis; 15(0→1) first neural spine
extends only slightly, or not at all, dorsally above its attachment to
skull; 26(0→1) ventral end of supracleithrum deeply bifurcate;
Cybium 2006, 30(2)
31(0→1) posterior process of pelvis behind pelvic-fin base long
and rod-like, moderately separated from its opposite member along
midline; 32(0→1) moderate to small cycloid scales; 33(0→1)
buckler-like scales present midabdominally and from far rear end
of spinous dorsal in to end of soft dorsal-in base; 36(0→1) scales
absent from along bases of rays, and scales nearby without spiny
projections and not extending beyond lateral expansions of distal
ends of dorsal- and anal-in pterygiophores; 37(0→1) 29 to 32 vertebrae; 42(1→2) ive vacant interneural spaces.
c: 1(0→1) basisphenoid absent; 2(0→1) supraocular serrations
of frontal absent; 3(0→1) tiny, trilobed otolith; 6(0→1) dermosphenotic fused or highly consolidated with the sphenotic; 8(0→1) ventral lange of symplectic absent; 9(0→1) two cartilages of moderate
size, each attached anteriorly to dentary and lying sequentially one
behind the other, the irst shorter than second; 16(0→1) Baudelot’s
ligament attached to exoccipital; 18(0→1) ossiied epineurals present on only a few of anterior abdominal vertebrae; 28(0→1) extrascapulars two tubular bones, not closely held to skull, except at large
specimen sizes; 30(0→1) anterolateral processes of medial surfaces
of pelvic-in rays present as prongs; 38(0→2) 14 abdominal vertebrae.
D: 12(0→1) ceratohyal-epihyal articulation with bony interdigitated articulations, at least in specimens of large size; 20(0→1)
extra-caudal ossicle in haemal spine of Pu2 present.
E: 14(0→1) Pu2 is first vertebra in caudal peduncle with a
modiied neural or haemal spine; 24(2→1) second anal-in spine
moderate to long, more than one-half length of irst spine to almost
as long, convergent in Zeus jerzmanskae; 39(0→1) three or four
vertebrae in caudal peduncle, convergent in Zeus robustus.
F: 4(1→2) lachrymal slender; 29(0→1) pelvic-fin spines
absent; 33(1→3) buckler-like scales present midabdominally and
from front to middle regions of spinous dorsal fin to end of soft
dorsal-in base; 34(0→1) accessory spiny process of buckler-like
scales absent; 35(1→0) radiating striations of buckler-like scales
present; 43(2→1) three anal-in spines; 44(1→2) 11 or 12 pectoralin rays.
G: 5(0→2) nine or more infraorbitals; 32(1→3) scales (excluding enlarged buckler-like scales) absent from most of body, or only
lateral line scales present; 37(1→2) 33 to 36 vertebrae; 45(0→2)
six pelvic-in elements.
H: 11(0→1) no prominent notches on lower border of ceratohyal, convergent in Xenolepidichthys and Zeus jerzmanskae;
39(1→0) ive vertebrae in caudal peduncle.
I: 33(3→2) buckler-like scales present midabdominally and
from posterior region of spinous dorsal in to end of soft dorsal-in
base.
J: 25(0→2) eight or nine anal-fin pterygiophores anterior to
haemal spine of third caudal vertebra, convergent in Zeus jerzmanskae; 38(2→3) 15 abdominal vertebrae, convergent in Zeus jerzmanskae, Zeus primaevus, and Zenopsis nebulosus.
K: 10(1→2) no serrations on lower border of dentary, convergent in Xenolepidichthys.
autoapomorphic features are not informative with regard to the
phylogenetic relationships of the species analyzed, but we include
them here for diagnostic purposes: for Zeus robustus 38(0→1),
39(0→1); for Zeus jerzmanskae 11(0→1), 24(2→1), 25(0→2),
38(0→3); for Zeus capensis 41(1→2); for Zeus faber 41(1→2),
45(0→1); for Zeus primaevus 25(0→1), 38(2→3); for Zenopsis sp.
20(1→0), 41(1→0); for Zenopsis oblongus 2(1→0), 22(1→0); for
Zenopsis hoernesi 21(1→0), 42(2→1); for Zenopsis tyleri,
41(1→2); for Zenopsis conchifer 18(1→0); for Zenopsis nebulosus
25(0→1), 38(2→3).
107
A phylogeny of extant and fossil buckler dory fishes,
family Zeidae (Zeiformes, Acanthomorpha)
by
Francesco SANTINI (1)(5), James C. TYLER (2),
Alexandre F. BANNIKOV (3) & Dorin-Sorin BACIU (4)
ABSTRACT. - A data set of 45 putatively informative morphological characters (mostly osteological, and a few external
features) is analyzed for 12 extant and fossil species belonging to the buckler dory fish genera Zeus and Zenopsis, which
together constitute the family Zeidae (Zeiformes, Acanthomorpha), and for two outgroup taxa. Zeus consists of two extant
and three fossil species, whereas Zenopsis consists of four extant and three fossil species. Both genera date back to at least
the Oligocene (36 Mya). The phylogenetic analysis of the full data set (extant + fossil) provides strong support for the
monophyly of the Zeidae but only weak support for the monophyly of Zenopsis, and it calls into question the existence of a
clade formed by the extant and fossil taxa assigned to Zeus. Additional phylogenetic analyses were performed: (1) the most
incomplete fossil taxon, Zeus robustus, was excluded; (2) all fossil taxa were excluded; and (3) the new extant species of
Zenopsis was excluded. All of these analyses confirm a strongly supported clade formed by Zeus + Zenopsis and of a much
less strongly supported clade formed by the extant and fossil species of Zenopsis. The analyses do not provide support for a
clade formed by extant and fossil species previously assigned to the genus Zeus, even though when the fossil species are
excluded from the analysis, the two extant species of Zeus appear as each other’s sister groups in one of three equally parsimonious cladograms.
RÉSUMÉ. - Une hypothèse phylogénétique pour les Zeidae actuels et fossiles (Zeiformes, Acanthomorpha).
Une matrice des 45 caractères morphologiques (principalement ostéologiques) a été constituée pour 12 espèces
actuelles et fossiles des deux genres de Zeidae, Zeus et Zenopsis et pour deux extra-groupes. Le genre Zeus comprend deux
espèces actuelles et trois fossiles, et le genre Zenopsis en comprend quatre et trois respectivement. Les deux genres datent
de l’Oligocène moyen (36 millions d’années). L’analyse phylogénétique du jeu des données de toutes les espèces supporte
bien la monophylie des Zeidae, mais très peu celle de Zenopsis ; le clade formé par les espèces actuelles et fossiles du genre
Zeus n’est toutefois pas soutenu. Des analyses complémentaires ont été effectuées, après avoir éliminé soit Zeus robustus,
le fossile le moins connu, soit tous les fossiles, soit la nouvelle espèce de Zenopsis. Toutes ces analyses soutiennent l’existence d’un clade des Zeidae et d’un clade Zenopsis. Il n’y a pas de support pour un clade Zeus.
Key words. - Zeiformes - Zeidae - Fossils - Osteology - Phylogeny.
Zeidae (buckler dories, including the John dory or St.
Peter’s fish) are moderate- to deep-bodied acanthomorph
fishes. The family is composed of 12 species (six extant and
six extinct) that are presently classified in two genera. The
genus Zeus Linnaeus, 1758, includes the extant species Zeus
faber Linnaeus, 1758 and Zeus capensis Valenciennes, 1835,
and the extinct species Zeus robustus Gorjanovi -Kramberger, 1891 (Oligocene of Slovenia), Zeus jerzmanskae Baciu et
al., 2005 (Oligocene of Poland) and Zeus primaevus Scarabelli, 1859 (Miocene of Italy and Algeria). The genus
Zenopsis Gill, 1863, includes the extant species Zenopsis
nebulosus (Schlegel, 1847), Zenopsis conchifer (Lowe,
1852), Zenopsis oblongus Parin, 1989, and Zenopsis sp. (to
be described by U. Yamada, T. Nakabo and D. Bray – D.
Bray, pers. comm.), and the extinct species Zenopsis hoerne si Gorjanovi -Kramberger, 1891 (Oligocene of Slovenia),
Zenopsis clarus Daniltshenko, 1960 [Oligocene of the
Carpathians (Poland, Romania) and northwestern Caucasus
(Russia)] and Zenopsis tyleri Baciu & Bannikov, 2001
(Oligocene of Romania).
The first comprehensive phylogenetic hypothesis for the
order Zeiformes (dories) (Tyler et al., 2003) showed that the
extant Zeiformes form a strongly supported monophyletic
group composed of six clades of familial status: the Cyttidae, Oreosomatidae, Parazenidae, Zeniontidae, Grammicolepididae, and Zeidae, with this last family appearing in
the terminal, most-derived clade of dories, as the sister
group to the Grammicolepididae. However, none of the
(1) Département Systématique et Évolution, UMR 7138, Muséum national d’Histoire naturelle, 43 rue Cuvier, 752311 Paris CEDEX 05,
FRANCE.
(2) National Museum of Natural History, Smithsonian Institution, Washington, DC 20560-0106, USA. [tyler.jim@NMNH.SI.EDU]
(3) Paleontological Institute, Russia Academy of Sciences, Profsoyuznaya 123, Moscow 117997, RUSSIA. [aban@paleo.ru]
(4) Muzeul de Stiinte ale Naturii, Str. Petru Rares nr. 26, 5600 Piatra Neámt, ROMANIA. [dsbaciu@ambra.ro]
(5) Present address : Department of Zoology, University of Toronto, 25 Harbord Street, Toronto, Ontario M5S3G5, CANADA.
[francesco.santini@utoronto.ca]
Cybium 2006, 30(2): 99-107.
SANTINI ET AL.
Phylogeny of Zeidae
Table I. - List of specimens of extant species examined for this work. For fossil materials, see Baciu et al. (2005). [Liste des spécimens d’espèces actuelles examinés dans
cet article. Pour les matériels fossiles, voir Baciu et al. (2005)].
extant zeid taxa has become available for
the first time, makes it possible to produce
the first phylogenetic hypothesis for the
Zeidae that includes all of the species currently assigned to this family.
MATERIALS AND METHODS
Fourteen taxa are included in this analysis, including 12 zeids and two outgroups.
In addition to the zeid taxa and the two outgroups already analyzed in Tyler et al.
(2003), three additional extant zeid species
(Zeus capensis, Zenopsis oblongus, Zenop sis sp.) are included. The six extinct taxa of
Zeidae included in this analysis were
described with reconstructions in Baciu et
al. (2005); the reader is referred to that
paper for the descriptions and measurements of those species. Minor differences
from the data in that paper are based upon
our re-interpretation of the number of vertebral segments or anal pterygiophores in a
few specimens. Table I lists all examined
species, and for extant species it includes
museum number, number of specimens in
each lot, and standard length (SL) when
available. Museum abbreviations follow
Leviton et al. (1985).
extinct taxa and only three of the six known extant species of
zeid fishes (Zeus faber, Zenopsis conchifer, and Zenopsis
nebulosus) were included in that (mostly osteological) analysis because of the lack of sufficient material. Baciu et al.
(2005) have since published a revision of the fossil record of
the Zeidae, including the description of new species and
detailed redescription of the osteology of several others. The
availability of these new data for the fossil zeids, in addition
to the fact that osteological material for the least known
100
Phylogenetic analysis
External morphological and osteological characters were obtained from direct
observation of the specimens listed in table
I; the character list is reported in appendix
I. The characters were analyzed following
the principles of phylogenetic systematics
(Kitching et al., 2000). A matrix for all the
specimens examined was first constructed
using WINCLADA (Nixon, 2002). This
matrix was subsequently analyzed using
NONA (Goloboff, 1999) and is shown in
table II. Following Tyler et al. (2003) one
species of Zeniontidae and one of Grammicolepididae were
used as outgroups. The zeniontid Zenion was selected as the
first outgroup, following the protocol suggested by Nixon
and Carpenter (1993). All characters were assigned equal
weight (1), and all multistate characters were analyzed as
unordered. Heuristic searches, with random addition of taxa,
the TBR + TBR branch swapping option of NONA, and
10,000 replications were performed. Tree length (L), consistency index (CI), and retention index (RI) are provided for
Cybium 2006, 30(2)
SANTINI ET AL.
Phylogeny of Zeidae
Table II. - Data set of 45 morphological characters for the 14 species in this analysis (12 Zeidae plus two outgroups). Numbers in parentheses at the right side of the table indicate the number of characters that could be determined, when some characters are unknown for a taxon.
[Jeu de données de 45 caractères morphologiques pour les 14 espèces utilisées dans cette analyse. Les nombres entre parenthèses à droite
du tableau indiquent le nombre de caractères qui ont pu être déterminés dans les cas où quelques-uns d’entre eux restent inconnus pour un
taxon.]
each analysis (Kluge and Farris, 1969; Farris, 1989). When
more than one most-parsimonious tree was obtained, a strict
consensus tree was calculated. Character evolution
(appendix II) was studied using the delayed transformation
(DELTRAN) option of WINCLADA, because most characters were scored as unknown (“?”) for at least some taxa, and
the use of the accelerated transformation (ACCTRAN)
option would have necessitated hypothesizing the presence
of certain character states within lineages, for which there is
no evidence that these states have ever been present. The
decay index (Bremer, 1988, 1994) was calculated using
NONA (Goloboff, 1999). Cladograms for publication were
produced using TREEVIEW (Page, 1996). Unknown character states in the fossil taxa are indicated with a question
mark “?”. Inapplicable characters are indicated with a horizontal dash “-”. It should be remembered that although treebuilding programs treat dashes and question marks in the
same way, they are conceptually different.
RESULTS
Analysis of the full data set produces two equally parsimonious trees (EPTs hereafter) (Fig. 1). The topology of the
resultant strict consensus tree strongly supports the monoCybium 2006, 30(2)
phyly of the Zeidae, with a high decay index of nine, and
illustrates the existence of a clade formed by the extant and
extinct species of Zenopsis, whereas Zeus appears to be
paraphyletic. Furthermore, the two extant species of Zeus, Z.
faber and Z. capensis, never appear to form a monophyletic
group, and Z. capensis appears to be a more basal taxon than
Z. faber. The relationships of the Oligocene Zeus robustus,
whose fossil record is based upon highly incomplete materials, appear to be problematic because its placement is significantly variable among the cladograms in figure 1. The relationships for the species of Zenopsis are stable despite the
weak decay index for all groups within the Zenopsis clade.
This low support might be a reflection of the presence of fossil taxa, which have many unknown character states (see
also Santini and Tyler, 2004). In the Zenopsis clade, the yetto-be-described species Zenopsis sp. appears as the most
basal lineage. Subsequently, two subclades can be identified:
one formed by the extant Zenopsis conchifer and Zenopsis
nebulosus; and one formed by Zenopsis oblongus plus the
three extinct species. Within this last clade, Zenopsis
oblongus is the sister lineage to Zenopsis hoernesi + (Zenop sis clarus, Zenopsis tyleri).
Exclusion of the highly incomplete Zeus robustus from
the data set does not help to resolve the relationships among
the various species assigned to Zeus. On the contrary, it
101
Phylogeny of Zeidae
SANTINI ET AL.
Figure 1. - Equally parsimonious trees
(EPTs) 1 and 2 (L = 85, CI = 0.76, RI =
0.80), and strict consensus (SC) tree
produced by the analysis of the full
data set of 12 extant and extinct species
of Zeidae plus two outgroups. Decay
index is indicated above the tree
branches of the SC tree. Cladogram #2
was selected for character optimization
in figure 5. [EPTs 1 et 2 (L = 85, CI =
0,7,6 RI = 0,80), et arbre de consensus
strict (SC) produits par l’analyse de
toutes les informations pour 12 espèces
de Zeidae vivants et fossiles, et deux
extra-groupes. L’indice de Bremer est
indiqué sur les branches de l’arbre SC.
Le deuxième cladogramme a été utilisé
pour l’optimisation des caractères
dans la figure 5.]
Figure 2. - Strict consensus tree of the five EPTs (L = 0.83, CI =
0.78, RI = 0.82) produced by the analysis of the data set of extant
and extinct species of Zeidae plus two outgroups when Zeus robus tus is not included. Decay index is indicated above the tree branches. [Arbre de consensus strict des 5 EPTs (L = 0,83, CI = 0,78, RI =
0,82) produit par l’analyse du jeu de données contenant les espèces
actuelles et fossiles de Zeidae et deux extra-groupes, à l’exclusion
de Zeus robustus. L’indice de Bremer est indiqué sur les branches
de l’arbre.]
102
increases the number of EPTs to five. The strict consensus
tree (Fig. 2) shows that although both the monophyly of the
Zeidae and of Zenopsis are still supported, with a decay
index of 11 and 2 respectively, within Zenopsis only the
clade of Z. conchifer and Z. nebulosus and the clade of Z.
clarus and Z. tyleri are recovered, albeit with a low decay
index of 1. These last two clades appear in a polytomy with
the other three species of Zenopsis.
When only the extant taxa are included in the analyzed
data set, three EPTs are recovered (Fig. 3). The consensus
tree supports the monophyly of Zenopsis, even though the
relationships among the various taxa are less well resolved
than in the analysis with the fossils, whereas Zeus still
appears as paraphyletic. Within Zenopsis, Z. sp. and Z.
oblongus appear in a polytomy with the Z. nebulosus + Z.
conchifer clade. The support for the monophyly of the Zeidae
is very high (decay index of 19), whereas that of the mono-
Cybium 2006, 30(2)
SANTINI ET AL.
Phylogeny of Zeidae
Figure 3. - EPTs 1 through 3 (L = 0.70,
CI = 0.92, RI = 0.95) and the strict consensus tree (SC) produced by the analysis of the data set of six extant species
of Zeidae plus two outgroups only.
Decay index is indicated above the tree
branches of the SC tree. [EPTs 1-3 (L =
0,70, CI = 0,92, RI = 0,95) et arbre de
consensus strict (SC) produits par
l’analyse du jeu de données de six
espèces actuelles de Zeidae et seule ment deux extra-groupes. L’indice de
Bremer est indiqué sur les branches de
l’arbre SC.]
Exclusion from the analysis of Zenopsis sp., with many
unknown internal characters, recovers one EPTs (Fig. 4). In
this case, the monophyly of the extant Zeus is again not
recovered because Zeus capensis appears as the most basal
zeid, whereas Zeus faber is the sister group to Zenopsis.
DISCUSSION
Figure 4. - Most parsimonious tree (L = 0.67, CI = 0.97, RI = 0.98),
produced after analysis of the data set of extant species of Zeidae
plus two outgroups, when Zenopsis sp. is not included. Decay
index is indicated above the tree branches. [Arbre le plus parci monieux (L = 0,67, CI = 0,97, RI = 0,98) produit de l’analyse du
jeu de données des espèces actuelles de Zeidae avec deux extra groupes, à l’exclusion de Zenopsis sp. L’indice de Bremer est
indiqué sur les branches de l’arbre.]
phyly of Zenopsis is relatively good (decay index of 3, which
corresponds to slightly more than 4% of the tree length).
Cybium 2006, 30(2)
Phylogenetic analysis
One result of this work is that the inclusion of even very
incomplete fossil materials does not prevent the inference of
a phylogeny when a phylogenetic signal is present in the
data set (a conclusion already supported by Santini and
Tyler, 2003, 2004; Tyler and Santini, 2005). The monophyly
of the Zeidae and of Zenopsis are recovered in all analyses
with and without the inclusion of the fossils. Also, the relationships among the extant species of Zenopsis do not
103
Phylogeny of Zeidae
SANTINI ET AL.
Figure 5. - EPT 2 from figure 1 selected
for the study of character evolution.
Letters below branches correspond to
letters in the character evolution section in appendix II. [EPT 2 de la figure
1 utilisé pour l’étude de l’évolution des
caractères. Les lettres au-dessous des
branches correspondent aux lettres
dans la partie de l’appendice II qui
traite de l’évolution des caractères.]
change when the fossils are included in the analysis. The situation is different for Zeus, in which the monophyly of the
group is not recovered in the analyses when the fossils are
included. When only the extant species of Zeus are analyzed,
the monophyly of the group is supported, but with a weak
decay index (see following section on classification for further comments).
A second result is that according to the topology of the
trees recovered, the radiation of the Zeidae is likely much
older than previously thought. The fact that the three
Oligocene species of Zenopsis appear to be very derived
taxa within the Zenopsis clade indicates that the radiation of
the various lineages in this group had already occurred by
35 Mya.
Classification
The two extant species of Zeus appear to form a monophyletic group in only one of the trees produced by the analyses of extant species alone (Fig. 3). A similar result is produced when the fossils are included in the analyses. The present results do not allow us to recover a monophyletic Zeus,
a genus that was first created in 1758 by Linnaeus and has
been recognized as valid ever since. Because in one of the
analyses the two extant species appear to be each other’s sister groups, and because the decay index for the relationships
among most species of Zeus is always very weak, we prefer
to retain the current generic classification. We thus continue
to recognize the generic name Zeus as a valid taxon pending
further work, which either may reveal support for the existence of a clade formed by all fossil and extant Zeus or may
find increased support for the paraphyly of this group and,
hence, the recognition of additional generic categories.
104
Paleobiogeography of the Zeidae
The only putative Eocene zeid was originally described
from the Tertiary of Georgia as Platax (?) colchicus by
Simonovich et al. (1875); however, Bogatshov (1933) concluded that this fish instead belongs to Zeus and he noted
that the marls in Georgia in which this species was found are
of Eocene age. The two type specimens cannot be located.
Danilchenko (1960) mentioned the opinion of Bogatshov
and hypothesized that the Georgian fish is in fact a Zenopsis.
Baciu et al. (2005) considered this taxon to be a Zenopsis
nomen dubium; however, on the basis of the phylogeny presented herein and on the other known fossil zeids, we think it
is likely that the zeids first appeared in the Eocene. Five of
the six additional fossil zeid species are in fact distributed in
the Oligocene of the Central Paratethys (Slovenia, Poland,
Romania) and Eastern Paratethys (Caucasus), with several
of these species from the Rupelian, the oldest part of the
Oligocene (approximately 30-36 Mya). Zeus primaevus, the
single species from the late Miocene (5-11 Mya), is distributed only in the Mediterranean basin (Algeria, Italy,
Spain). The extant zeid species occur in the Atlantic, Indian,
and Pacific Oceans and are mostly distributed throughout the
coastal waters of the continents, with only Zeus faber occurring in the Mediterranean and Black Seas.
Most of the Oligocene zeids are geographically distributed in a single location, probably corresponding to isolated
basins. For example, Zeus robustus is known only from
Slovenia; Zeus jerzmanskae from Poland; Zenopsis hoernesi
from Slovenia; and Zenopsis tyleri from Romania. The only
exception is Zenopsis clarus, which is widely distributed
throughout the Paratethys, with fossil specimens known
from Russia, Romania, and Poland. The Upper Miocene
Cybium 2006, 30(2)
SANTINI ET AL.
Phylogeny of Zeidae
Zeus primaevus is known from deposits in Italy, Oran (Algeria), and Spain (Baciu et al., 2005), and its more expansive
distribution might indicate that the ichthyofauna had by then
been homogenized across the Mediterranean region. The
recent species Zeus faber was described by Bassani (1905)
from the Pleistocene deposits of Taranto (southern Italy). We
have not examined materials from Taranto, but it is possible
that these Pleistocene specimens are referable to the extant
Zeus faber, which likely belongs to a phylogenetically very
ancient lineage. No fossils are known for any of the other
extant species of Zeidae.
Acknowledgements. - Many individuals, as listed in Baciu et al.
(2005), made the materials of their institutions available to us for
examination, and we thank them all. We additionally thank N.V.
Parin, Shirshov Institute of Oceanology, Moscow, who donated
rare specimens of Zenopsis oblongus to the USNM collection; D.
Bray, Museum of Victoria, and T. Nakabo, Kyoto University Museum, for radiographs of the new species of Zenopsis (herein referred
to as Zenopsis sp.). S. Zehren, University of Alabama, provided literature on zeiforms. D. Goujet and M. Verán, MNHN, Paris, helped
arrange for a visit to the paleoichthyological collections of the
MNHN. Our work in Paris was also facilitated by loaned materials
and radiographs sent by S. Raredon, L. Palmer, and J. Williams,
Smithsonian Institution, Washington, D.C.; N.N. Parin, International Academic Agency “Nauka,” Moscow; and N. Micklich, Hessisches Landesmuseum, Darmstadt. M. Hautecœur, MNHN, radiographed some specimens of Zeidae on loan to FS.
Excavations in the East Carpathians in 2002 by D.-S. Baciu, during which important specimens of Zenopsis were discovered, were
funded by grant no. 7312-02 from the National Geographic Society, Washington, D.C.
Our joint work on Zeiformes has been made possible by two
Short Term Visitor Fellowships awarded to D.-S. Baciu and A.
Bannikov by the Smithsonian Institution for collaborative research
with J.C. Tyler, and by a NATO Life Science and Technology collaborative linkage grant (LST. CLG. 978836) to J. Tyler, A. Bannikov, D.-S. Baciu, and F. Santini. F. Santini was supported by a
Marie Curie Fellowship for a project on “Paleontological and
molecular approaches to the phylogeny of Acanthomorpha
(Pisces).”
This manuscript was greatly improved by the comments of D.
Tyler, N. Micklich, G. Lecointre, who also helped with the French
translation of the abstract, and two anonymous reviewers.
REFERENCES
BACIU D.-S., BANNIKOV A.F. & J.C. TYLER, 2005. - Revision
of the fossil fishes of the family Zeidae (Zeiformes). Boll. Mus.
Civ. Storia Nat. Verona (Geologia, Paleontologia, Preistoria),
29: 95-118.
BASSANI F., 1905. - La ittiofauna delle argille marnose plistoceniche di Taranto e di Nardò (Terra d’Otranto). Atti R. Accad. Sci.
Fis. Mat. Napoli, ser. 2, 12: 1-60.
BOGATSHOV V.V., 1933. - Data on the Tertiary Fish Fauna of the
Caucasus. Tr. Azerb. Neft. Issled. Inst., Geol. Otd., 15: 1-62. [In
Russian]
Cybium 2006, 30(2)
BREMER K., 1988. - The limits of amino acid sequence data in
angiosperm phylogenetic reconstruction. Evolution, 42: 795803.
BREMER K., 1994. - Branch support and tree stability. Cladistics,
10: 295-304.
DANILCHENKO P.G., 1960. - Bony fishes of Maikop deposits of
the Caucasus. Tr. Paleontol. Inst. Akad. Nauk SSSR, 78: 1-208.
[In Russian]
FARRIS J.S., 1989. - The retention index and rescaled consistency
index. Cladistics, 5: 417-419.
GOLOBOFF P., 1999. - NONA (NO NAME) version 2. Tucumán,
Argentina: published by the author. Available via
http://www.cladistics.com.
KITCHING I.J., FOREY P.L., HUMPHRIES C.J. & D.M.
WILLIAMS, 2000. - Cladistics, 2nd Edit. 228 p. Oxford:
Oxford University Press.
KLUGE A.G. & J.S. FARRIS, 1969. - Quantitative phyletics and
the evolution of anurans. Syst. Zool., 18: 1-32.
LEVITON A.E., GIBBS, R.H., HEAL E. & C.E. DAWSON, 1985.
- Standards in herpetology and ichthyology: Part I. Standard
symbolic codes for institutional resource collections in herpetology and ichthyology. Copeia, 3: 802-832.
MADDISON W.P., 1989. - Reconstructing character evolution on
polytomous cladograms. Cladistics, 5: 365-377.
NIXON K.C., 2002. - WINCLADA ver. 1.00.08. Ithaca, NY: published by the author. Available via http://www.cladistics.com.
NIXON K.C. & J.M. CARPENTER, 1993. - On outgroups. Cladis tics, 9: 413-426.
PAGE R.D., 1996. - TREEVIEW: An application to display phylogenetic trees on personal computers. Comp. Appl. Biosc., 12:
357-358.
SANTINI F. & J.C. TYLER, 2003. - A phylogeny of the families of
fossil and extant Tetraodontiform fishes (Acanthomorpha,
Tetraodontiformes), Upper Cretaceous to Recent. Zool. J. Linn.
Soc., 139: 565-617.
SANTINI F. & J.C. TYLER, 2004. - The importance of even highly
incomplete fossil taxa in reconstructing the phylogenetic relationships of the Tetraodontiformes (Acanthomorpha: Pisces).
Int. Comp. Biol., 44: 349-357.
SIMONOVICH S., BATSEVICH L. & A. SOROKIN, 1875. - Geologic Description of Parts of the Kutaisi, Lechkhumi, Zugdidi
and Senaki Districts of Kutaisi Province. 191 p. Tiflis: Materialy dlya geologii Kavkaza. [In Russian]
TYLER J.C. & F. SANTINI, 2005. - A phylogeny of the fossil and
extant zeiform-like fishes, Upper Cretaceous to Recent, with
comments on the putative zeomorph clade (Acanthomorpha).
Zool. Scripta, 34:157-175.
TYLER J.C., O’TOOLE B. & R. WINTERBOTTTOM, 2003. Phylogeny of the genera and families of Zeiform fishes, with
comments on their relationships with tetraodontiforms and
caproids. Smithson. Contrib. Zool., 618: 1-110.
Reçu le 10 mai 2005.
Accepté pour publication le 10 septembre 2005.
105
SANTINI ET AL.
Phylogeny of Zeidae
Appendix I
Character list
The character list is modified and reduced from Tyler et al.
(2003) to compensate for the different taxonomic sampling. Two
new characters, numbers 34 and 35 in our list, and some additional
character states have been added. Characters are arranged according to body region.
Cranial
1. Basisphenoid: present (0); absent (1).
2. Frontal, supraocular serrations: present (0); absent (1).
3. Otolith, shape: moderate to large size, rounded or slightly to
deeply indented on one or both sides, or oblong with humps
(0); tiny, trilobed (bow-tie shaped) (1).
4. Lachrymal, size/depth: large, deep, height about one to four
times in the length (0); moderate, height about five to seven
times in the length (1); slender (2).
5. Infraorbitals, number of (well-developed elements exclusive
of the lachrymal, dermosphenotic, and of variable rudiments):
five to eight (0); four or less (1); nine or more (2).
6. Dermosphenotic: a distinctly separate ossification from the
sphenotic, sometimes relatively free from the skull (0); fused
or highly consolidated with the sphenotic (1).
7. Premaxilla, alveolar process: ventrally rounded or moderately
indented to form a pair of blunt lobes (0); deeply bifurcated
ventrally (1).
8. Symplectic, ventral flange: present (0); absent (1).
9. Dentary, cartilages (on lateral surface of dentary): two cartilages of moderate size, lying sequentially one behind the
other, of about the same size or the first only slightly shorter
than the second (0); two cartilages of moderate size, each
attached anteriorly to the dentary and lying sequentially one
behind the other, the first shorter than the second (1); absent
or unconsolidated (2).
10. Dentary, serrations on the lower border of: multiple serrations
behind the symphysis (0); a single barb near the symphysis
(1); none (2).
11. Ceratohyal, notches on the lower border of: prominent notches for the heads of some of the branchiostegal rays in the anterior group (0); no prominent notches (1).
12. Ceratohyal-epihyal articulation: exclusively through cartilage
(0); with bony interdigitated articulations, at least in specimens of large size (1).
13. Epihyal, depth of the anterior end of: equal, or about equal, to
the depth of the adjacent part of the ceratohyal (0); distinctly
less deep than the adjacent part of the ceratohyal (1).
Vertebral column and median fins
14. First vertebra in the caudal peduncle with a modified neural or
haemal spine: third preural centrum, PU3 (0); second preural
centrum, PU2 (1).
15. First vertebra, dorsal extension of the neural spine when the
neural arch and spine are plastered to the skull: the neural
spine with a long dorsal portion free from the skull beyond
the curvature of the supraoccipital and the exoccipitals (0);
the neural spine extending only slightly, or not at all, dorsally
above its attachment to the skull (1).
16. Baudelot’s ligament, placement of the proximal attachment
of: to the first vertebra (0); to the exoccipital (1).
17. Ossified ribs: present only on the last few abdominal vertebrae (0); present on most of the abdominal vertebrae behind
the fourth (1).
106
18. Ossified epineurals: present on most of the abdominal vertebrae or their ribs (0); present on only a few of the anterior
abdominal vertebrae (1).
19. Epurals, number: two (0); one (1).
20. PU2, extra-caudal ossicle in the haemal spine of: absent (0);
present, in at least some specimens (1).
21. Vacant interneural spaces, number of groups of (when two or
more spaces are vacant): two (0); three or four (1).
22. Dorsal-fin pterygiophores, number of anterior to the neural
spine of the fourth abdominal vertebra: two or three (0); four
(1).
23. Supraneurals, number of: one (0); none (1).
24. Second anal-fin spine, length of: very short, much less than
one-half the length of the first spine (0); moderate to long,
more than one-half the length of the first spine to almost as
long (1); longer than first spine (2).
25. Anal-fin pterygiophores, number of anterior to the haemal
spine of the third caudal vertebra: five or six (0); seven (1);
eight or nine (2).
Paired-fin girdles
26. Supracleithrum, ventral end of: simple (0); deeply bifurcate
(1).
27. Cleithrum, posterior edge: without a posterodorsal prong
above the articulation with the postcleithrum (0); cleithral
process present as a prong above the articulation with the
postcleithrum (1).
28. Extrascapulars: one long bone, sometimes forming an open
tube, more or less closely held to the skull (0); two tubular
bones, not closely held to the skull, except at large specimen
sizes (1).
29. Pelvic-fin spines: present (0); absent (1).
30. Pelvic-fin rays, anterolateral processes of the medial (lower)
surfaces of: absent (0); present as prongs from the medial surfaces of the ray bases (1).
31. Pelvis, posterior process of behind pelvic-fin base: short to
moderate in length, and in shape a moderate to broad plate or
flattened shaft (0); long and rod-like, moderately separated
from its opposite member along the midline (1).
Scales
32. Scales, on most of the body: moderate to small, spiny
“ctenoid” (spinoid) (0); moderate to small, cycloid (1); greatly elongate vertically (2); absent (excluding enlarged bucklerlike scales), or only in the lateral line (3).
33. Scales, buckler-like (greatly enlarged midline scales): absent
(0); present midabdominally and from the far rear end of the
spinous dorsal fin (from no further forward than the last dorsal spine) to the end of the soft dorsal-fin base (1); present
midabdominally and from the posterior region of the spinous
dorsal fin (from under the last two or three dorsal spines) to
the end of the soft dorsal-fin base (2); present midabdominally and from the front to middle regions of the spinous dorsal
fin to the end of the soft dorsal-fin base (3).
34. Buckler-like scales, in addition to major spiny process, accessory spiny process: present (0); absent (1); not applicable,
when buckler-like scales absent (-).
35. Buckler-like scales, radiating striations: present (0); absent
(1); not applicable, when buckler-like scales absent (-).
36. Scales, along the bases of the dorsal- and anal-fin rays: absent
along the bases of the rays, but spiny processes present on the
scales alongside the lateral expansions of the distal ends of
the dorsal- and anal-fin pterygiophores (0); absent from along
Cybium 2006, 30(2)
SANTINI ET AL.
Phylogeny of Zeidae
the bases of the rays, and the scales nearby without spiny projections and not extending beyond the lateral expansions of
the distal ends of the dorsal- and anal-fin pterygiophores (1).
Meristic data
37. Vertebrae, total number of: 27 or 28 (0); 29 to 32 (1); 33 to 36
(2); 37 or 38 (3).
38. Abdominal vertebrae, number of: ten or 11 (0); 13 (1); 14 (2);
15 (3).
39. Vertebrae, number of in the caudal peduncle (posterior to the
last vertebra whose neural or haemal spine supports a pterygiophore): five (0); three or four (1); eight (2).
40. Procurrent caudal-fin rays, number of (including the number
in both the dorsal and ventral sides, if different): two or three
(0); one (1).
41. Dorsal-fin spines, number of: six or seven (0); eight or nine
(1); ten or more (2).
42. Vacant interneural spaces, total number of below the spiny
and anterior part of the soft dorsal-fin base, posterior to the
first dorsal-fin pterygiophore: two (0); four (1); five (2).
43. Anal-fin spines, number of: two (0); three (1); four (2).
44. Pectoral-fin rays, number of: 15 or 16 (0); 13 or 14 (1); 11 or
12 (2).
45. Pelvic-fin elements, total number of: seven (0); eight (1); six
(2).
Appendix II
Character evolution
The difficulties produced by the use of consensus trees in the
study of character evolution are well known (e.g., see Maddison,
1989). In order to investigate the evolution of the various characters, we selected one of the two EPTs obtained from the analysis of
the full data set. We selected EPT number 2 (Fig. 1) because in this
cladogram Zeus robustus, one of the oldest fossil zeids, is represented as being a stem zeid, and we judge this interpretation of the
data to be more reliable, on the basis of the stratigraphic criterion,
than any of the others.
The selected tree is shown in figure 5. All internodes have been
labeled with letters in order to more easily list the character states
that support the various clades. The character optimization was performed using DELTRAN.
A: 4(0 1) lachrymal moderate, height about five to seven
times in the length, convergent in Xenolepidichthys; 13(0 1) anterior end of epihyal distinctly less deep than adjacent part of ceratohyal, convergent in Xenolepidichthys; 17(0 1) ossified ribs present on most abdominal vertebrae behind fourth; 19(0 1) one epural; 21(0 1) three or four groups of vacant interneural spaces;
22(0 1) four dorsal-fin pterygiophores anterior to neural spine of
fourth abdominal vertebra; 23(0 1) no supraneural; 24(0 2) second anal-fin spine longer than first spine; 27(0 1) cleithral process
present as prong above articulation with postcleithrum; 40(0 1)
one procurrent caudal-fin ray, convergent in Xenolepidichthys;
41(0 1) eight or nine dorsal-fin spines; 42(0 1) four vacant
interneural spaces, convergent in Xenolepidichthys and in Zenopsis
hoernesi; 43(0 2) four anal-fin spines; 44(0 1) 13 or 14 pectoral-fin rays, convergent in Xenolepidichthys.
B: 7(0 1) alveolar process of premaxilla deeply bifurcated
ventrally; 10(0 1), serration on lower border of dentary consists
of a single barb near the symphysis; 15(0 1) first neural spine
extends only slightly, or not at all, dorsally above its attachment to
skull; 26(0 1) ventral end of supracleithrum deeply bifurcate;
Cybium 2006, 30(2)
31(0 1) posterior process of pelvis behind pelvic-fin base long
and rod-like, moderately separated from its opposite member along
midline; 32(0 1) moderate to small cycloid scales; 33(0 1)
buckler-like scales present midabdominally and from far rear end
of spinous dorsal fin to end of soft dorsal-fin base; 36(0 1) scales
absent from along bases of rays, and scales nearby without spiny
projections and not extending beyond lateral expansions of distal
ends of dorsal- and anal-fin pterygiophores; 37(0 1) 29 to 32 vertebrae; 42(1 2) five vacant interneural spaces.
C: 1(0 1) basisphenoid absent; 2(0 1) supraocular serrations
of frontal absent; 3(0 1) tiny, trilobed otolith; 6(0 1) dermosphenotic fused or highly consolidated with the sphenotic; 8(0 1)
ventral flange of symplectic absent; 9(0 1) two cartilages of moderate size, each attached anteriorly to dentary and lying sequentially one behind the other, the first shorter than second; 16(0 1)
Baudelot’s ligament attached to exoccipital; 18(0 1) ossified
epineurals present on only a few of anterior abdominal vertebrae;
28(0 1) extrascapulars two tubular bones, not closely held to
skull, except at large specimen sizes; 30(0 1) anterolateral processes of medial surfaces of pelvic-fin rays present as prongs;
38(0 2) 14 abdominal vertebrae.
D: 12(0 1) ceratohyal-epihyal articulation with bony interdigitated articulations, at least in specimens of large size; 20(0 1)
extra-caudal ossicle in haemal spine of PU2 present.
E: 14(0 1) PU2 is first vertebra in caudal peduncle with a
modified neural or haemal spine; 24(2 1) second anal-fin spine
moderate to long, more than one-half length of first spine to almost
as long, convergent in Zeus jerzmanskae; 39(0 1) three or four
vertebrae in caudal peduncle, convergent in Zeus robustus.
F: 4(1 2) lachrymal slender; 29(0 1) pelvic-fin spines
absent; 33(1 3) buckler-like scales present midabdominally and
from front to middle regions of spinous dorsal fin to end of soft dorsal-fin base; 34(0 1) accessory spiny process of buckler-like
scales absent; 35(1 0) radiating striations of buckler-like scales
present; 43(2 1) three anal-fin spines; 44(1 2) 11 or 12 pectoralfin rays.
G: 5(0 2) nine or more infraorbitals; 32(1 3) scales (excluding enlarged buckler-like scales) absent from most of body, or only
lateral line scales present; 37(1 2) 33 to 36 vertebrae; 45(0 2)
six pelvic-fin elements.
H: 11(0 1) no prominent notches on lower border of ceratohyal, convergent in Xenolepidichthys and Zeus jerzmanskae;
39(1 0) five vertebrae in caudal peduncle.
I: 33(3 2) buckler-like scales present midabdominally and
from posterior region of spinous dorsal fin to end of soft dorsal-fin
base.
J: 25(0 2) eight or nine anal-fin pterygiophores anterior to
haemal spine of third caudal vertebra, convergent in Zeus jerzman skae; 38(2 3) 15 abdominal vertebrae, convergent in Zeus jerz manskae, Zeus primaevus, and Zenopsis nebulosus.
K: 10(1 2) no serrations on lower border of dentary, convergent in Xenolepidichthys.
Autoapomorphic features are not informative with regard to the
phylogenetic relationships of the species analyzed, but we include
them here for diagnostic purposes: for Zeus robustus 38(0 1),
39(0 1); for Zeus jerzmanskae 11(0 1), 24(2 1), 25(0 2),
38(0 3); for Zeus capensis 41(1 2); for Zeus faber 41(1 2),
45(0 1); for Zeus primaevus 25(0 1), 38(2 3); for Zenopsis sp.
20(1 0), 41(1 0); for Zenopsis oblongus 2(1 0), 22(1 0); for
Zenopsis hoernesi 21(1 0), 42(2 1); for Zenopsis tyleri,
41(1 2); for Zenopsis conchifer 18(1 0); for Zenopsis nebulosus
25(0 1), 38(2 3).
107