DSpace at VNU: The skull of Hagiangella goujeti Janvier, 2005, a high-crested acanthothoracid (Vertebrata, Placodermi) from the Lower Devonian of northern Vietnam

This article was downloaded by: [Aston University]
On: 18 January 2014, At: 11:58
Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,
37-41 Mortimer Street, London W1T 3JH, UK

Journal of Vertebrate Paleontology
Publication details, including instructions for authors and subscription information:
/>
The skull of Hagiangella goujeti Janvier, 2005, a highcrested acanthothoracid (Vertebrata, Placodermi) from
the Lower Devonian of northern Vietnam
Vincent Dupret
Janvier

d e

a f

b

b

c

, Ta Hoa Phuong , Tong-Dzuy Thanh , Nguyen Duc Phong , Philippe

& Gaël Clément

d

a

Université Lille 1—Sciences et Technologies , FRE 3298 Géosystèmes du CNRS, 59655,
Villeneuve d’Ascq cedex, France
b

Department of Geology , Vietnam National University , 334 Nguyen Trai street, Thanh Xuan
District, Ha Noi City, Viet Nam
c

Vietnam Institute of Geosciences and Mineral Resources (VIGMR) , Km9+300, Nguyen Trai
Street, Thanh Xuan District, Ha Noi City, Viet Nam
d

Muséum National d’Histoire Naturelle , UMR 7207 du CNRS, CP38, 47 rue Cuvier 75231 Paris
cedex 05, France
e

Palaeontology Department , The Natural History Museum , Cromwell Road, London, SW7
5BD, United Kingdom
f

Department of Evolutionary Organismal Biology , University of Uppsala , Norbyvägen 18A,
742 36, Uppsala, Sweden
Published online: 09 May 2011.

To cite this article: Vincent Dupret , Ta Hoa Phuong , Tong-Dzuy Thanh , Nguyen Duc Phong , Philippe Janvier & Gaël Clément
(2011) The skull of Hagiangella goujeti Janvier, 2005, a high-crested acanthothoracid (Vertebrata, Placodermi) from the
Lower Devonian of northern Vietnam, Journal of Vertebrate Paleontology, 31:3, 531-538, DOI: 10.1080/02724634.2011.558148
To link to this article: />
PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained
in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the
Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and
are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and
should be independently verified with primary sources of information. Taylor and Francis shall not be liable for
any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever
or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of
the Content.
This article may be used for research, teaching, and private study purposes. Any substantial or systematic
reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any
form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://
www.tandfonline.com/page/terms-and-conditions


Journal of Vertebrate Paleontology 31(3):531–538, May 2011
© 2011 by the Society of Vertebrate Paleontology

ARTICLE

THE SKULL OF HAGIANGELLA GOUJETI JANVIER, 2005, A HIGH-CRESTED
ACANTHOTHORACID (VERTEBRATA, PLACODERMI) FROM THE LOWER DEVONIAN OF
NORTHERN VIETNAM
∗ ,1,

VINCENT DUPRET,

† TA HOA PHUONG,2 TONG-DZUY THANH,2 NGUYEN DUC PHONG,3 PHILIPPE JANVIER,4,5

4

¨ CLEMENT
´
and GAEL
´
`
Universite´ Lille 1—Sciences et Technologies, FRE 3298 Geosyst
emes
du CNRS, 59655 Villeneuve d’Ascq cedex, France;
2
Department of Geology, Vietnam National University, 334 Nguyen Trai street, Thanh Xuan District, Ha Noi City, Viet Nam,
; ;
3
Vietnam Institute of Geosciences and Mineral Resources (VIGMR), Km9+300, Nguyen Trai Street, Thanh Xuan District, Ha Noi
City, Viet Nam, ;
4
´
Museum
National d’Histoire Naturelle, UMR 7207 du CNRS, CP38, 47 rue Cuvier 75231 Paris cedex 05, France, ;
;
5
Palaeontology Department, The Natural History Museum, Cromwell Road, London SW7 5BD, United Kingdom

Downloaded by [Aston University] at 11:58 18 January 2014

1

ABSTRACT—The acanthothoracid Hagiangella goujeti Janvier, 2005, has been described exclusively on the basis of isolated
thoracic plates from the Lochkovian (Lower Devonian) Khao Loc Formation of Tung Vai, Ha Giang Province, northern
Vietnam. It is characterized by a very high, triangular median crest on the median dorsal plate, and has been referred to the
Acanthothoraci on the basis of the morphology of its fused anterolateral, spinal and anterior ventrolateral plates, and the

characteristic stellate ornamentation of the group. Isolated plates of H. goujeti are relatively abundant at Tung Vai and no
other placoderm taxon from this locality seems to share the same type of ornamentation. However, the skull of this species
remained elusive. Here we report two well-preserved skull roofs from Tung Vai, which we refer to H. goujeti. They display the
same stellate ornamentation and small size as the previously described plates of the thoracic armor of this species. This new
material shows that the head of H. goujeti is surprisingly short (i.e., possibly lacking dermal rostral and pineal elements), in
contrast to the elongate and narrow skull of all other acanthothoracids. The combination of unique characters (e.g., presence
of two pairs of posterior pit lines, two pairs of central and paranuchal plates, etc.) suggests a possible sister group relationship
to the placoderm assemblage Petalichthyida + Ptyctodontida + Arthrodira.

INTRODUCTION
The Placodermi McCoy, 1848, (‘armored fishes’) are the most
diversified vertebrates of the entire Devonian period (Janvier,
¨ 1944, is con1996). Among them, the Acanthothoraci Stensio,
sidered a paraphyletic ensemble of generalized placoderms, to
which the different clades of placoderms are related (Goujet,
1984a; Janvier, 1996). So, in the paraphyletic family Palaeacan¨ 1944, ‘Radotina’ prima is the sister group
thaspididae Stensio,
to the Arthrodira Woodward, 1891 (including the Phyllolepida
¨ 1934), on one hand, and to the Petalichthyida Jaekel,
Stensio,
1911, plus the Ptyctodontida Gross, 1932, on the other (see Janvier, 1996:fig. 4.57). A group including Brindabellaspis Young,
1980, Romundina Ørvig, 1975, and Palaeacanthaspis Brotzen,
1934, is the sister group to the Antiarchi Cope, 1885; a last ensemble comprising Radotina kosorensis Gross, 1950, Kosoraspis
Gross, 1959, and Kimaspis Mark-Kurik, 1973a, is more closely
related to the Rhenanida Broili, 1930. The supposedly monophyletic family Weejasperaspididae White, 1978, would be the
sister group of a large clade that includes the Palaeacanthaspididae, Arthrodira, Petalichthyidae, and Ptyctodontida. More
recently, Goujet and Young (2004) have considered the Acanthothoraci monophyletic. Given the present work, we follow a
non-monophyletic hypothesis.
*Corresponding author. †Current address: Department of Evolution¨
ary Organismal Biology, University of Uppsala, Norbyvagen

18A, 742 36
Uppsala, Sweden,

The recently described Hagiangella goujeti Janvier, 2005, from
the Lochkovian of northern Vietnam was not assigned to any
particular acanthothoracid family, because the original material consisted only of some isolated thoracic armor fragments.
The new skull roof material described herein shows that this
taxon is distinct from the Palaeacanthaspididae and represents
the sister group to either the Ptyctodontida, or at least, the clade
((Ptyctodontida, Petalichthyida) Arthrodira).

GEOLOGICAL SETTING, MATERIALS, AND METHODS
The previously described postcranial material of Hagiangella
gougeti, including the holotype, was collected at Tung Vai, Ha Giang Province, Vietnam, and comes from two outcrops that have
been named “Tung Vai 1” (N23◦ 03.43, E104◦ 54.91, altitude 954
m) and “Tung Vai 2” (N23◦ 03.18, E104◦ 54.51, altitude 954 m),
respectively (Racheboeuf et al., 2005) (Fig. 1). The skull roofs described herein all come from Tung Vai 2, from which also comes
the holotype of the species. They are referred to H. goujeti because no other placoderm taxon from the same locality bears the
same, characteristic stellate ornamentation.
The specimens have been prepared by cleaning the natural
mould of the skull roof with dilute hydrochloric acid. An elastomer cast of the natural mould was made and whitened with
magnesium oxide for photography.
Institutional Abbreviation—BT, Geological Museum, Hanoi,
Vietnam.

531


532

JOURNAL OF VERTEBRATE PALEONTOLOGY, VOL. 31, NO. 3, 2011
median pit line; N, nuchal plate; occ, occipital cross-commissure;
p.PaN, posterior paranuchal plate; p.pop, posterior postorbital
process; PaN, paranuchal plate; Pi, pineal plate; pmc, postmarginal sensory line canal or groove; PPi, postpineal plate; ppl,
posterior pit line; ppl1, first (anterior) posterior pit line; ppl2,
second (posterior) posterior pit line; Prm, premedian plate; PrO,
preorbital plate; PtO, postorbital plate; R, rostral plate; SM,
submarginal plate; soc, supraorbital sensory line canal or groove;
s.p, pit for presumably cutaneous sensory organ; sov, supraorbital
vault; tpl, transverse pit line (in Denison, 1978:fig.16A).
SYSTEMATIC PALEONTOLOGY
Class PLACODERMI McCoy, 1848
¨ 1944
Order ACANTHOTHORACI Stensio,
Family HAGIANGELLIDAE, fam. nov.
Diagnosis—The same as for the type genus and type species,
by monotypy.
Type Genus—Hagiangella Janvier, 2005 (see Racheboeuf
et al., 2005:534 for taxonomic disclaimer).

Downloaded by [Aston University] at 11:58 18 January 2014

Genus HAGIANGELLA Janvier, 2005 (see Racheboeuf et al.,
2005:534 for taxonomic disclaimer)
Diagnosis—The same as for the type species, by monotypy
(Racheboeuf et al., 2005).
HAGIANGELLA GOUJETI Janvier, 2005
Holotype—An isolated median dorsal plate (BT-186;
Racheboeuf et al., 2005:fig. 4A–D).
New Materials—Two isolated skull roofs, BT231 (Fig. 2A) and

BT232 (Fig. 2B–C), respectively.
Locality and Age—Tung Vai 2, Ha Giang Province, Northern
Vietnam; base of the Khao Loc Formation, Lochkovian (Early
Devonian).
Diagnosis (emended from Racheboeuf et al., 2005)—A small
Acanthothoraci with a long spinal plate, and a very high, vertical,
and triangular process of the median dorsal plate prolonged by
a posterior blade. The ventrolateral lamina of the shoulder girdle is expanded medially. As in other acanthothoracids, the ornamentation of the armor consists of stellate tubercles. The rostral and pineal plates are not fused to the postethmoid part of the
skull roof, as is the case of the pineal plate in Radotina prima (see
Denison, 1978:fig. 22C). There is no evidence of a naso-capsular
dermal entity involving the rostral, pineal and premedian plates.
The anterior surface of the preorbital plates is vertical and ornamented.
DESCRIPTION

FIGURE 1. Geographic location of Tung Vai 1 and 2 in northern Vietnam (modified after Tong-Dzuy et al., 1995:text-fig. 1).

Anatomical Abbreviations—a.PaN, anterior paranuchal plate;
a.pop, anterior postorbital process; C, central plate; C1, first
(anterior) central plate; C2, second (posterior) central plate; cc,
central sensory line canal or groove; d.e, external foramen for
the endolymphatic duct; ioc, infraorbital sensory line canal or
groove; lc, cephalic main lateral sensory line canal or groove;
lim.sov, limit of the supraorbital vault; M, marginal plate; mpl,

Only the postethmoid ossification of the skull roof is preserved;
no evidence of rostral, pineal, or premedian plates has been
found in the material. These dermal ethmoid and pre-ethmoid
elements have maybe been detached after the death of the animal (as occurs in most acanthothoracids and other ‘loose-nose’
placoderms; e.g., Kujdanowiaspis, in which the postethmoid ossification is most often preserved), rather than having been simply
absent. Nevertheless, the ornamented anterior surface of the preorbital plates contradicts this latter hypothesis.

The plate boundaries are very difficult to identify, most sutures
being unclear. The radiation center of each plate (except that of
the nuchal plate) seems to be slightly elevated. This feature, along
with the concentric pattern of the tubercles and the lack of tubercles along plate sutures (likely due to the differential growth of
the plates), was used to identify the plate boundaries. The sensory line system consists of deeply open grooves, except around
the nuchal plate radiation center, where the grooves appear much
shallower. We have not identified the occipital cross-commissure,


533

Downloaded by [Aston University] at 11:58 18 January 2014

DUPRET ET AL.—SKULL OF HAGIANGELLA FROM DEVONIAN OF VIETNAM

FIGURE 2. Hagiangella goujeti, Janvier, 2005. Skull roof in dorsal view (left side) and schematic interpretation (right side). A, BT231 in external
view. B, BT232 part. C, BT232 counterpart. Scale bar equals 1 mm.


Downloaded by [Aston University] at 11:58 18 January 2014

534

JOURNAL OF VERTEBRATE PALEONTOLOGY, VOL. 31, NO. 3, 2011

FIGURE 3. Comparison of different placoderm skull roofs. A, Hagiangella goujeti Janvier, 2005 (schematic reconstruction of the skull roof in external
view after BT231-232; the right part of the reconstruction takes into account the possibility that the marginal plate extends mesially to the infraorbital
and main lateral sensory line grooves; dotted line indicates the limit—in internal view—of the supraorbital vault.). B, Ctenurella gladbachensis Ørvig,
1960 (redrawn after Ørvig, 1962:fig. 1A). C, “Radotina” prima (Barrande, 1872) (redrawn after Gross, 1958; Westoll, 1967; Denison, 1978:fig. 22C).
D, Romundina stellina (redrawn after Ørvig, 1975; Denison, 1978:fig. 22G). E, Kimaspis tienshanica Mark-Kurik, 1973a (redrawn after Mark-Kurik,

1973a:text-fig. 2, and Denison, 1978:fig. 22D, who reconstructed the ethmoid and re-identified C2 and a.PaN—Elga Mark-Kurik, pers. comm.). D,
Sensory line system indicated in light grey. Not to scale.

but two pairs of posterior pit lines are most probably present (see
Discussion).
One pair of sensory line grooves is puzzling (ppl1, Fig. 1A–C;
see below for discussion of homologies), because it does not cross
any radiation center; on the contrary, it runs between the anterior
and posterior central plate radiation centers towards the point of
convergence between the supraorbital sensory line and the (second) posterior pit line on the nuchal plate. These grooves may be
suggestive of a plate boundary, but we consider that they could
also be part of the sensory line system, because some grooves
do not necessarily cross the radiation centers in Radotina (see
Mark-Kurik, 1973a). Nevertheless, the skull roof of some species
of Radotina consists of plates but also of many tesserae, as in Kimaspis but contrary to Hagiangella.
The skull roof of H. goujeti is shorter than in any other Acanthothoraci, especially in the postorbital area. Its median anterior
edge is straight and vertical, and covered with stellate tubercles.
Therefore, we assume that the rostral, pineal, and premedian
plates were lacking, or that a gap separated them from the preor-

bital plates, as in the Antiarchi-related Romundina stellina (see
Ørvig, 1975). The orbits are very large (almost half of the skull
roof length); the supraorbital vault is visible in ventral view in
specimen BT232 (sov, Fig. 2B). The posterior part of the skull
roof is deeply embayed medially, and somewhat elevated, most
probably to fit the very high median dorsal plate (see Racheboeuf
et al., 2005:figs. 4A–D, 5). Nevertheless, the posterolateral corner
of the skull roof (i.e., the posterior edge of the paranuchal plates)
does not extend as posteriorly as in others species of Acanthothoraci (Fig. 3).
The boundary between the two adjacent preorbital plates

(PrO, Fig. 2) is very difficult to discern. They form the anterior
half of the orbital edge and are crossed by the straight supraorbital sensory line grooves (soc, Fig. 2) that extend until the level
of the radiation center of the nuchal plate (N, Fig. 2). They contact the postpineal plate posteriorly (Ppi, Fig. 2) in the midline.
The postorbital plates (PtO, Fig. 2) constitute the posterior
part of the orbital edge. In Radotina they are crossed by the
infraorbital sensory line grooves, but this is harder to assess in


Downloaded by [Aston University] at 11:58 18 January 2014

DUPRET ET AL.—SKULL OF HAGIANGELLA FROM DEVONIAN OF VIETNAM
Hagiangella because the boundary between the postorbital and
marginal plates is not discernible. In other words, there is a possibility that the infraorbital groove could be borne by the marginal
plate, as in the Ptyctodontida; nevertheless, this is doubtful
because the angle of the infraorbital groove is well marked
(about 90◦ ) in Hagiangella, compared to the slight bow observable in the Ptyctodontida (see Fig. 3B). No central sensory groove
or canal is visible.
The postpineal plate is of uncertain shape, but its approximate
location can be identified owing to a different tuberculation (i.e.,
more dense) in this area, especially in specimen BT231 (Ppi, Fig.
2A).
The lateral margin of the marginal plate (M, Figs. 2A, C, 3A)
is unknown, and its boundary is indistinct medially. It seems that
on specimen BT232 (Fig. 2C), the plate extends into a lateral
blade, circling the orbit posteriorly, and following the infraorbital sensory line groove. These features are comparable to those
observed in the Ptyctodontida. The marginal plate is crossed by
the postmarginal sensory line groove (pmc, Figs. 2A–B, 3A). The
internal view of specimen BT232 exposes two neurocranial processes, respectively identified as the anterior and the posterior
postorbital process (a.pop, p.pop, Fig. 2B).
The paired central plates (C1, C2, Fig. 2) surround the nuchal

plate laterally. The anterior central plates contact the preorbital,
postorbital, posterior central, anterior paranuchal, nuchal, and
most probably the postpineal plates. The anterior central plate
is not crossed by any sensory line groove. The posterior central
plate constitutes the posterior edge of the skull roof, together
with the nuchal and the posterior paranuchal plates. The boundary between the anterior and posterior central plates and the
nuchal plate is not discernible.
The shape of the nuchal plate (N, Fig. 2) cannot be determined,
because the boundary with the central plates is not clear. It contacts anteriorly the postpineal plate anteriorly (which, as noted,
shows a denser tubercle distribution; Ppi, Fig. 2A). The radiation
center of the nuchal plate, situated in the middle of the plate, is
slightly depressed; the supraorbital sensory groove and the two
pairs of posterior pit lines become shallower and most probably converged at this point. Because the course of these sensory grooves where they converge is not as obvious as in other
grooves, we believe that this confluence was more superficial.
The anterior and posterior paranuchal plates (a.PaN, p.PaN,
Fig. 2) are crossed by the two pairs of posterior pit lines. Only
their mutual boundaries, as well as those with the postorbital and
central plates, are discernible. The boundary with the marginal
plate is unknown. The posterior paranuchal plate does not extend
posteriorly contrary to its homologue in other Acanthothoraci.

DISCUSSION
Sensory Line Groove Homology in Hagiangella goujeti
Although the homology between the supraorbital, infraorbital, cephalic main lateral, and postmarginal sensory grooves
of Hagiangella and those of other placoderms is clear, this is
not the case for the posterior pit line and the occipital crosscommissure, which cross the posterior central, anterior, and posterior paranuchal and nuchal plates. Moreover, determining homologies is made more difficult by the fact that a real central
groove and a median pit line are not observed on the material,
so that the posterior pit line and the occipital cross-commissure
cannot be identified by a process of elimination. However, first,
because the median pit line runs parallel to the central sensory

groove in most placoderms that display these two sensory lines,
and second, because no similar structure is observed in the material of Hagiangella goujeti, we consider that the latter lacks the
central sensory groove and the median pit line.

535

In most placoderms, the occipital cross-commissure runs either
through a gap behind the nuchal plate or in the posterior part of
the nuchal plate, but rarely at the level of the nuchal plate radiation center. On the contrary, when an element of the sensory line
system crosses the radiation center of the nuchal plate, it may be
the supraorbital grooves, the central grooves, or the median and
posterior pit lines (as in the Phyllolepididae and Wuttagoonaspididae).
The occipital cross-commissure is documented in acanthothoracids (e.g., Mark-Kurik, 1973a, 1994), as in all other placoderm
groups. Usually, this groove runs from the posterior paranuchal
(or single paranuchal) plate radiation center and crosses either
the posterior-most part of the nuchal plate or through the nuchal
gap or extrascapular plates. The origin of this occipital crosscommissure in Acanthothoraci is situated posteriorly in the posterior extension of the posterior paranuchal plate, and is directed
toward the nuchal gap.
The posterior pit line, when single, runs between the
paranuchal and the central or nuchal plate radiation centers.
When two pairs of posterior pit lines occur, the first (anterior)
one runs between the central and the marginal radiation centers (as is the case in Yiminaspis shenme, see Dupret, 2008:fig.
2A–B), or its course is visible (if not restricted) on the central
plate only (see ppl.a of Romundina stellina in Ørvig, 1975:fig.1A;
or Lunaspis broilii, Gross, 1961; or pp1 in Eurycaraspis incilis
Liu, 1991:fig. 1) but still oriented onto the marginal plate (see
also Dupret, 2008, for discussion). Additionally, when two pairs
of paranuchal plates are present, the second (posterior) posterior pit line always runs between the nuchal and the anterior
paranuchal plates. Nevertheless, in the basal arthrodire Yiminaspis shenme, in which only one pair of paranuchal plates occurs,
the second posterior pit line is anchored in the posterior part of

the paranuchal plate, in the immediate vicinity of the occipital
cross-commissure (the latter running further posteriorly from the
nuchal plate radiation center).
Consequently, as far as Hagiangella is concerned, and because the supraorbital groove has been identified and the central groove is absent, the posterior-most sensory line component
(ppl2, Figs. 2, 3A), running between the posterior paranuchal
and the nuchal plate radiation centers, can be homologized with
the second posterior pit line rather than with the occipital crosscommissure.
As for the more anterior sensory line component (ppl1, Figs. 2,
3A), it should not be homologized with a median pit line, because
in most cases it parallels the course of the central groove. Hence,
this component is considered homologous with the first posterior
pit line of the Petalichthyida (though incomplete in this taxon)
and of Yiminaspis shenme (see Dupret, 2008:fig. 2), thereby entailing the presence of two pairs of posterior pit lines on the skull
roof of Hagiangella goujeti.
The presence of two pairs of posterior pit lines is not unique
among placoderms. The Petalichthyida also possess two pairs of
posterior pit lines, although the anterior one is never complete.
Recently, two pairs of posterior pit lines have been identified
in the basal arthrodire Yiminaspis shenme Dupret, 2008 (Fig.
2A–B); it is nevertheless noteworthy that these are associated
with one pair of central and paranuchal plates. In Yiminaspis
shenme, the first (anterior-most) posterior pit line runs from the
radiation center of the marginal plate to the sensory groove confluence at the level of the radiation center of the nuchal plate.
This pattern is very similar to that of the first posterior pit line
of Hagiangella goujeti. It is also noteworthy that the lateral end
of the first posterior pit line lies close to the postmarginal groove
in both Hagiangella and Yiminaspis. The second (posterior-most)
posterior pit line of Hagiangella runs between the radiation center of the paranuchal plate and the confluence of the sensory
grooves at the level of the radiation center of the nuchal plate,
as in its homologue in Yiminaspis shenme. Therefore, the two

536

JOURNAL OF VERTEBRATE PALEONTOLOGY, VOL. 31, NO. 3, 2011

medial transverse sensory grooves in Hagiangella goujeti can be
homologized with the first and second posterior pit lines of Yiminaspis shenme.
Affinities of Hagiangella with the Ptyctodontida

Downloaded by [Aston University] at 11:58 18 January 2014

Although the skull roof of Hagiangella goujeti is far from displaying a typical ptyctodontid-like pattern, some of its features
compare to either this group or closely related ones:
1. The orbits are very large, as in the Ptyctodontida, and represent about one half of the skull roof length. Large orbits, though possibly related to lifestyle, occur mostly in the
Ptyctodontida and some Acanthothoraci (e.g., Romundina
stellina), and is probably an apomorphic character state (yet
possibly homoplastic, i.e., acquired by convergence). Nevertheless, large orbits may represent a juvenile feature. The latter interpretation may be supported by the presence of only
one tubercle generation in Hagiangella goujeti (D. Goujet,
pers. homologized). Although the number of specimens is not
large enough to assess any of these hypotheses, because the
skull roof specimens are about half the size they should be to
fit the thoracic armor material (see Racheboeuf et al., 2005),
we favor the juvenile specimen hypothesis. For this reason, we
do not include the large orbits in the diagnosis.
2. The skull roof of the Ptyctodontida is laterally composed by
slender blades (i.e., the marginal plates) that surround the orbits posteroventrally. Such a feature is unknown in the Acanthothoraci. A possible exception may be Romundina stellina,
in which a slight and tiny process surrounds the orbit posteroventrally, but it is unclear whether the plate in question
is the postorbital or the marginal plate (see Denison, 1978:fig.
22G; Fig. 2D; however, the marginal plate in Romundina stellina does not appear to extend onto the orbital edge). Moreover, this area of the skull roof is not well preserved in Hagiangella goujeti. Nevertheless, one specimen (BT232, Fig. 2B–C)

suggests the presence of large infraorbital processes for the
marginal plate at least, thus displaying a condition similar to
that observed in the Ptyctodontida.
3. The central sensory line is lacking in Hagiangella goujeti, as
well as in the Ptyctodontida and in the Petalichthyida, the
latter two forming a clade according to Goujet and Young
(1995). For these authors, the absence of this central sensory
line represents a derived condition.
4. The sensory line system of the skull roof of Hagiangella goujeti seems to consist of grooves, rather than enclosed canals
that open to the exterior via pores, though it is possible that
a thin bone layer roofed those canals; the superficial dermal bone layer has been partly destroyed during fossilization (Fig. 2). The possession of canals plus pores is present
in the Petalichthyida and the Ptyctodontida (it is one of the
synapomorphies considered by Goujet and Young, 1995, for
this clade), as well as in the palaeacanthaspid Kimaspis tienshanica Mark-Kurik, 1973a (which Mark-Kurik considered as
close to “the hypothethical ancestor of placoderms”; MarkKurik, 1973a:329). It is uncertain whether a closed sensory line
canal condition is derived relative to the presence of grooves
(as in the Arthrodira), but it is possible that this combination
of canals and pores is inherited from a single common ancestor, shared the Ptyctodontida, the Petalichthyida, and possibly
Kimaspis and Hagiangella (if the presence of sunken canals vs.
grooves can be determined in new, better-preserved material).
5. The supraorbital sensory groove and the two pairs of posterior pit lines meet at the radiation center of the nuchal
plate in all Ptyctodontida, except in Rhamphodopsis threipandli Watson, 1934 (attributed to a nuchal plate instead of
a postpineal plate by Long, 1997). This feature also occurs in basal Arthrodira (i.e., the Wuttagoonaspididae Wutta-

goonaspis fletcheri Ritchie, 1973, and Yiminaspis shenme), and
in some more derived Arthrodira (i.e., the Phyllolepididae).
In phyllolepidids, however, this confluence occurs on the ‘centronuchal’ plate, recently considered as consisting of the fused
central plates only, the nuchal plate being absent (see Dupret
and Zhu, 2008, for details). It is nevertheless noteworthy that
Young (2005) considers this sensory line confluence on a medial dermal element as a primitive condition for placoderms.

6. No endolymphatic foramen or duct has been identified in the
skull material of Hagiangella. This structure is absent (presumably lost) in the Ptyctodontida.
7. The trunk armor is known from median dorsal plates, possible posterior dorsolateral plates, and a single dermal element formed from the fusion of the anterior ventrolateral,
anterolateral, and spinal plates (Racheboeuf et al., 2005).
Racheboeuf et al. pointed out the resemblance between the
median dorsal plate of Hagiangella and that of the ptyctodontid Rhamphodopsis and some high crested groenlandaspidid
arthrodires. Recent phylogenetic analyses show the Groenlandaspididae and Phlyctaenididae as basal taxa among the
Phlyctaenii (e.g., see Dupret, 2004; Dupret et al., 2007, 2009;
Dupret and Zhu, 2008). Some groenlandaspidid taxa (i.e.,
Mulgaspis Ritchie, 2004) also retain the presence of anteroventral plates that were once considered typical for the
actinolepidoid arthrodires (Miles, 1973) until their discovery in some Chinese quasipetalichthyids (Liu, 1991). Also,
the internal side of the median dorsal plate shows a pair
of strong ridges, that Racheboeuf et al. (2005) compare to
the Petalichthyida and the posterior median dorsal plate of
the Antiarchi; this structure is also comparable to that in
some Actinolepidoidei (e.g., Actinolepis magna, Mark-Kurik,
1973b:text-fig. 3B; Erikaspis zychi, Dupret et al., 2007) and
Phlyctaenii (e.g., see Dicksonosteus arcticus, Goujet, 1984b:fig.
56, bcp.d), and may be related to the attachment of the axial
muscles of the trunk (Goujet, 1984b).
The trunk armor of the earliest ptyctodonts (i.e., Tollodus brevispinus Mark-Kurik, 1977, from the Lochkovian of Kotelny
Island, New Siberian Archipelago; N.B., first mentioned and
illustrated as ‘ptyctodont’ in Mark-Kurik, 1974; see below)
shows interesting comparisons with that of Hagiangella. A
comparison between the trunk armor of Hagiangella and other
Acanthothoraci is given in Racheboeuf et al., 2005.
The median dorsal plate of both Hagiangella and Tollodus
shows a broad and flat basis from which extends a flattened
triangular blade in Hagiangella (Racheboeuf et al., 2005:figs.
4A–F, 5), and a ‘spiny’ process in Tollodus (Mark-Kurik,

1974:text-fig. 1(11), pl.1 fig. 3; Mark-Kurik, 1977:figs. 1–3). In
later ptyctodonts, the median dorsal blade is short based and
presents a high sub-vertical spiny process.
The anterior dorsolateral plate of Tollodus already shows
the features of the later ptyctodonts (short, deep unornamented overlap blades for the median dorsal and anterolateral
plates; Mark-Kurik, 1974:pl.1 fig.2, text-fig. 1(3); Mark-Kurik,
1977:fig. 4). Nevertheless, the shape of the articular condyle
(for the craniothoracic dermal articulation) seems to be much
simpler than the hollow one of the later ptyctodonts.
The anterolateral plate of Hagiangella is as long as high,
whereas that of Tollodus is much higher than long (MarkKurik, 1977:fig. 5), as in later ptyctodonts. Lastly, the ventral
armor in Tollodus is composed of the possibly fused interolateral and anterior ventrolateral plates (Mark-Kurik, 1977:fig.
1), whereas it is considered that the later ptyctodonts only possess a pair of interolateral plates (e.g., Long, 1997). The ventral armor of Hagiangella shows a pair of well-developed anterior ventrolateral plates (of which a medial expansion of is
reminiscent of the Phlyctaenididae), but no interolateral (the
latter being also absent in all Acanthothoraci; Racheboeuf et
al., 2005:figs. 6, 7). It is hence possible, though speculative,


DUPRET ET AL.—SKULL OF HAGIANGELLA FROM DEVONIAN OF VIETNAM
to suggest that the anterior ventrolateral plate of Hagiangella
and the possibly fused interolateral and anterior ventrolateral
plates of Tollodus are homologous.
The ventral armor of Tollodus and other ptyctodonts is very
short, compared to that of Hagiangella. Also, in lateral view
the trunk armor of Tollodus (Mark-Kurik, 1977:fig. 1B) is at
least twice as short as that of Hagiangella, and hence is more
‘ptyctodont’ than that of Hagiangella.
The spinal plate of Hagiangella is strong and long, extends beyond the posterior edge of the anterior ventrolateral plate, and
belongs to a unit composed of this plate with the anterolateral and the anterior ventrolateral plates (similar to that in the
acanthothoracids Romundina stellina, Palaeacanthaspis vasta,

and Kosoraspis peckai; Racheboeuf et al., 2005). The spinal
plates of Tollodus and later ptyctodonts are very reduced.

Downloaded by [Aston University] at 11:58 18 January 2014

Phylogenetic Position and Systematics of Hagiangella goujeti
Despite many common features shared by Hagiangella and
the Ptyctodontida, we cannot consider Hagiangella goujeti as a
Pyctodontida s.s. because it lacks the diagnostic characters of this
group. Alternative hypotheses would consider Hagiangella as the
sister group of the Ptyctodontida (Fig. 4, hypothesis 1). Alternatively, because Hagiangella possesses two pairs of central and
paranuchal plates, it could also be considered as closely related
to the group Ptyctodontida + Petalichthyida (Fig. 4, hypothesis
2). Moreover, because two pairs of paranuchal plates occur in
the Petalichthyida, and occasionally in the most primitive members of the Arthrodira (e.g., Yujiangolepis liujingensis Wang et
al., 1998; see revision in Dupret et al., 2009), Hagiangella could
also be considered as the sister taxon to the group Ptyctodontida
+ Petalichthida + Arthrodira (Fig. 4, hypothesis 3).
The presence of two pairs of central plates was previously regarded as typical for the Acanthothoraci, although some possess
one pair, as is the case for ‘Radotina’ prima, which is considered
as most closely related to the group Arthrodira + Petalichthyida
+ Petalichthyida (Goujet, 1984a; Janvier, 1996). Hence, Hagiangella may equally be considered as the sister group of this latter
group (including ‘Radotina’ prima) (Fig. 4, hypothesis 4). However, a close relationship between ‘Radotina’ prima and Hagiangella seems unlikely (different skull roof, trunk armor, and sensory line patterns; compare Fig. 3A and C).

537

Lastly, with regards to the stellate tubercle ornamentation,
although considered for some time as exclusive to the acanthothoracids, it appears that many groups of placoderms show
this ornamentation on the scale cover—but not the dermal
armor—though not necessarily the most primitive taxa (e.g., the

scales of arthrodires of the ‘buchanosteid-type’ of Turner and
Murphy, 1988, or of the rhenanid Ohioaspis tumulosa, in Burrow
and Turner, 1999:figs. 3D–F, 4C).
Considering the numerous problems of homology and character coding for the Placodermi in general (notably the large number of non-applicable data), an extensive phylogenetic analysis
will be the subject of a subsequent article. The phylogeny proposed in Figure 4 has to be considered as showing provisional
phylogenetic hypotheses described above.
Regarding the familial attribution of Hagiangella, and considering that we have pointed out some features it shares with the
Petalichthyida, the Ptyctodontida, and the basal Arthrodira, we
have for some time considered the possibility that it could be referred provisionally to the family “Palaeacanthaspididae.” However, as already stressed in Introduction, this family is considered as paraphyletic, relative to the Arthrodira and the group
Petalichthyida + Ptyctodontida. Therefore, we propose here the
erection of a new family, for the only genus Hagiangella.
CONCLUSIONS
The skull roof of Hagiangella goujeti Janvier, 2005, displays a
typical acanthothoracid set of characters (e.g., stellate ornamentation, two pairs of central and paranuchal plates), but also features that recall conditions observed in the Petalichthyida and the
most primitive Arthrodira (e.g., two pairs of posterior pit lines,
two pairs of paranuchal plates) and the Ptyctodontida (e.g., wide
postorbital blade most probably contributed by the marginal
plate). This leads us to consider four possible phylogenetic relationships for Hagiangella, including one in which Hagiangella is
the sister taxon to the ensemble Arthrodira + Petalichthyida +
Ptyctodontida.
ACKNOWLEDGMENTS
V.D. contributed to this article during a temporary position at
the USTL (Villeneuve d’Ascq, France; credits to the ECLIPSE
Project—Terrestrialization of the INSU/CNRS [Marco Vecoli],
of which this publication is a contribution); the article has been
finalized at the University of Uppsala. Field work was funded
´
by the Departement
Histoire de la Terre of the Museum National d’Histoire Naturelle, Paris. Ta Hoa Phuong and Tong Dzuy
Thanh are grateful to the National Foundation for Sciences and

Technology Development of Viet Nam for the effective support to the Projects 105.01.79.09 and 105.06.60.09. The authors
thank Elga Mark-Kurik (Institute of Geology, Tallinn University
´
of Technology, Estonia) and Daniel Goujet (Departement
His´
toire de la Terre, Museum
National d’Histoire Naturelle, Paris,
France) for their reviews, comments, and corrections.
LITERATURE CITED

FIGURE 4. Relationships among Placodermi (modified after Janvier,
1996:fig. 4.57). The numbers 1 to 4 are related to the possible phylogenetic
position of Hagiangella goujeti (branch in dashed lines).

`
`
` partie.
Barrande, J. 1872. Systeme
silurien du centre de la Boheme.
1ere
´
´
Recherches paleontologiques,
supplement
au vol. i., 647 pp.
¨
Broili, F. 1930. Uber
Gemundina Sturtzi
Traquair. Abhandlungen
¨

der Bayerischen Akademie der Wissenschaften Mathmatischnaturwissenschaftliche Abteilung 6:1–24.
Brotzen, F. 1934. Die silurischen und devonischen Fischvorkommen in
Westpodolien II. Paleobiologica 6:111–131.
Burrow, C. J., and S. Turner. 1999. A review of placoderm scales, and
their significance in placoderm phylogeny. Journal of Vertebrate Paleontology 19:204–219.
Cope, E. D. 1885. The position of Pterichthys in the system. American
Naturalist 19:289–291.


Downloaded by [Aston University] at 11:58 18 January 2014

538

JOURNAL OF VERTEBRATE PALEONTOLOGY, VOL. 31, NO. 3, 2011

Denison, R. H. 1978. Placodermi; in H.-P. Schultze (ed.), Handbook of
Paleoichthyology 2. Gustav Fischer Verlag, Stuttgart, New York,
128 pp.
Dupret, V. 2004. The phylogenetic relationships between actinolepids
(Placodermi: Arthrodira) and other arthrodires (phlyctaeniids and
brachythoracids). Fossils and Strata 50:40–55.
Dupret, V. 2008. First wuttagoonaspid (Placodermi, Arthrodira) from
the Lower Devonian of Yunnan, South China. Origin, dispersal and
palaeobiogeographic significance. Journal of Vertebrate Paleontology 28:12–20.
Dupret, V., and M. Zhu. 2008. The earliest phyllolepid (Placodermi,
Arthrodira), Gavinaspis convergens, from the late Lochkovian
(Lower Devonian) of Yunnan (South China). Geological Magazine
145:257–278.
Dupret, V., D. Goujet, and E. Mark-Kurik. 2007. A new genus of placoderm (Arthrodira: ‘Actinolepida’) from the Lower Devonian of
Podolia (Ukraine). Journal of Vertebrate Paleontology 27:266–284.

Dupret, V., M. Zhu, and J.-Q. Wang. 2009. The morphology of Yujiangolepis liujingensis (Placodermi, Arthrodira) from the Pragian of
Guangxi (South China) and its phylogenetic significance. Zoological Journal of the Linnean Society 157:70–82.
Goujet, D. 1984a. Placoderm Interrelationships: a new Interpretation,
with a short review of Placoderm Classifications. Proceedings of the
Linnean Society of New South Wales 107:211–243.
Goujet, D. 1984b. Les poissons placodermes du Spitsberg—Arthrodires
Dolichothoraci de la formation de Wood Bay (Devonien inferieur),
´
´ es).
´
Cahiers de Paleontologie
(section vertebr
Editions du CNRS,
Paris, 284 pp.
Goujet, D., and G. C. Young. 1995. Interrelationships of placoderms revisited. Geobios Memoire Special 19:89–95.
Goujet, D., and G. C. Young. 2004. Placoderm anatomy and phylogeny:
new insights; pp. 109–126 in G. Arratia, M. V. H. Wilson, and R.
Cloutier (eds.), Recent Advances in the Origin and Early Radiation
of Vertebrates. Verlag Dr. Friedlich Pfeil, Munich.
Gross, W. 1932. Die Arthrodira Wildungens. Geologische und Palaeontologische Abhandlungen 19:1–61.
¨
Gross, W. 1950. Die paleontologische
und stratigraphische Bedeutung
¨
der Wirbeltierfaunen des Old Reds und der marinen altpalaozoischen Schichten. Abhandlungen der Deutschen Akademie der Wissenschaften zu Berlin, Mathematisch-naturwissenschaftliche Klasse
1:1–130.
¨
¨
Gross, W. 1958. Uber
die alteste

Arthrodiren-Gattung. Notizblatt
¨ Bodenforschung zu Wiesbaden
des Hessischen Landesamtes fur
86:7–30.
Gross, W. 1959. Arthrodiren aus dem Obersilur der Prager Mulde.
Palaeontographica (A) 113:1–135.
Gross, W. 1961. Lunaspis broilii und Lunaspis heroldi aus dem
¨
Hunsruckschiefer
(Unterdevons, Rheinland). Notizblatt Hessisches
¨ Bodenforschung zu Wiesbaden 89:17–43.
Landesamtes fur
¨
¨
Jaekel, O. ed 1911. Die Wirbeltiere. Eine Ubersicht
uber
die fossilen und
lebenden Formen. Berlin, 252 pp.
Janvier, P. 1996. Early Vertebrates. Clarendon Press Edition; in H.
Charnock, J. F. Dewey, R. A. Price, S. Conway Morris, A. Navrotsky, B. J. Skinner, and E. R. Oxburgh (eds.), Oxford Monographs on
Geology and Geophysics, Volume 1. Oxford Science Publications,
Oxford, U.K., 393 pp.
Liu, Y.-H. 1991. On a new petalichthyid, Eurycaraspis incilis gen. et sp.
nov., from the middle Devonian of Zhanyi, Yunnan; pp. 139–177 in
M.-M. Chang, Y.-H. Liu, and G.-R. Zhang (eds.), Early Vertebrates
and Related Problems of Evolutionary Biology. Science Press, Beijing.
Long, J. A. 1997. Ptyctodontid fishes (Vertebrata, Placodermi) from the
Late Devonian Gogo Formation, Western Australia, with a revision of the European genus Ctenurella Ørvig, 1960. Geodiversitas
19:515–555.
Mark-Kurik, E. 1973a. Kimaspis, a new palaeacanthaspid from the Early

Devonian of Central Asia. Eesti NSV Teaduste Akadeemia Toimetised, Keemia. Geoloogia 22:322–330.
Mark-Kurik, E. 1973b. Actinolepis (Arthrodira) from the Middle Devonian of Estonia. Palaeontographica 143:89–108.
Mark-Kurik, E. 1974. Discovery of new Early Devonian fish localities
in the Soviet Arctic. Eesti NSV Teaduste Akademia Toimetised
23:330–335.
Mark-Kurik, E. 1977. The structure of the shoulder girdle in early
ptyctodontids; pp. 61–70 in V. V. Menner (ed.), Ocherki po filogenii
i sistemaike iskopaemykh myb i beschelyustryck.

Mark-Kurik, E. 1994. Placoderms and their stratigraphic significance; pp. 123–132 in V. N. Karatajute-Talimaa and R. G.
Matukhin (eds.), Stratigraphy and Fauna of the Lower Devonian deposits of the Tareya key section (Taimyr). Nedra, St.
Petersburg.
McCoy, F. 1848. On some new fossil fish of the Carboniferous period.
Annals and Magazine of Natural History 2:1–10.
Miles, R. S. 1973. An actinolepid arthrodire from the Lower Devonian
Peel Sound Formation, Prince of Wales Island. Palaeontographica
143:109–118.
Ørvig, T. 1960. New finds of Acanthodians, Arthrodires, Crossopterygians, Ganoids and Dipnoans in the Upper Middle Devonian Calcareous flags (Oberer Plattenkalk) of the Bergisch
¨
Gladbach—Paffrath Trough (Part 1). Palaontologisches
Zeitschrift
34:295–335.
Ørvig, T. 1962. Y a-t-il une relation directe entre les arthrodires
´
ptyctodontides et les holocephales?
Extrait des Colloques internationaux du Centre National de la Recherche
`
´
Scientifique—Problemes
actuels de Paleontologie—Evolution

´ es
´ 104:49–61.
des Vertebr
Ørvig, T. 1975. Description, with special reference to the dermal
skeleton, of a new Radotinid arthrodire from the Gedinnian of
Arctic Canada. Extrait des Colloques internationaux du Cen`
tre National de la Recherche Scientifique—Problemes
actuels de
´
´ es
´ 218:41–71.
Paleontologie—Evolution
des Vertebr
Racheboeuf, P. R., P. Janvier, T. H. Phuong, J. Vannier, and W. ShangQi. 2005. Lower Devonian vertebrates, arthropods and brachiopods
from northern Vietnam. Geobios 38:533–551.
Ritchie, A. 1973. Wuttagoonaspis gen. nov., an unusual arthrodire from
the Devonian of Western New South Wales, Australia. Palaeontographica 143:58–72.
Ritchie, A. 2004. A new genus and two new species of groenlandaspidid arthrodire (Pisces: Placodermi) from the Early-Middle Devonian Mulga Downs Group of western New South Wales, Australia.
Fossils and Strata 50:56–81.
¨ E. 1934. On the Placodermi of the Upper Devonian of
Stensio,
East Greenland. I. Phyllolepida and Arthrodira. Meddelelser om
Grønland 97:1–58.
¨ E. 1944. Contributions to the knowledge of the vertebrate
Stensio,
fauna of the Silurian and Devonian of Podolia II—note on two
¨ Zoologi
Arthrodires from the Downtonian of Podolia. Arkiv for
35:1–83.
Tong-Dzuy, T., P. Janvier, P. Ta Hoa, and N. Doan. 1995. Lower Devonian biostratigraphy and vertebrates of the Tong Vai valley, Vietnam. Palaeontology 38:169–186.

Turner, S., and A. Murphy. 1988. Early Devonian vertebrate microfossils
from the Simpson Park Range, Eureka County, Nevada. Journal of
Paleontology 62:959–964.
Wang, S.-T., J. Pan, and J.-Q. Wang. 1998. Early Devonian fishes
from central and southern Guangxi and correlation of the vertebrate biostratigraphy in south China. Vertebrata Palasiatica 36:58–
69.
Watson, D. M. S. 1934. The interpretation of arthrodires. Proceedings of
the Zoological Society, London:437–464.
Westoll, T. S. 1967. Radotina and other tesserate fishes. Journal of Linnean Society (Zoology) 47:83–98.
White, E. I. 1978. The larger arthrodiran fishes from the area of the Burrinjuck Dam, NSW. Transactions of the Royal Society of London
34:149–162.
Woodward, A. S. 1891. Catalogue of the Fossil Fishes in the British Museum of Natural History. Part II. Containing the Elasmobranchii
(Acanthodii), Holocephali, Ichthyodorulites, Ostracodermi, Dipnoi,
and Teleostomi (Crossopterygii), and Chondrostean Actinopterygii, Volume II. British Museum of Natural History, London,
567 pp.
Young, G. C. 1980. A new Early Devonian placoderm from New
South Wales, Australia, with a discussion of placoderm phylogeny.
Palaeontographica (A):10–76.
Young, G. C. 2005. A new phyllolepid placoderm occurence (Devonian fish) from the Dulcie Sandstone, Georgina Basin, central Australia. Proceedings of the Linnean Society of New South Wales
126:203–214.
Submitted January 12, 2010; accepted January 22, 2011.
Handling editor: Zerina Johanson.