Page 1
CHARACTERIZATION OF THE PLACODERM
(GNATHOSTOMATA) ASSEMBLAGE FROM THE
TETRAPOD-BEARING LOCALITY OF STRUD
(BELGIUM, UPPER FAMENNIAN)
by S�EBASTIEN OLIVE1,2*, GA€EL CL�EMENT3, EDWARD B. DAESCHLER4 and
VINCENT DUPRET5,6*1OD Earth and Life History, Royal Belgian Institute of Natural Sciences, Rue Vautier 29, 1000, Brussels, Belgium; e-mail: [email protected] & Diversity Dynamics Lab, Department of Geology, Li�ege University, B18, Boulevard du Rectorat, Sart-Tilman, 4000, Li�ege, Belgium3CR2P, UMR 7207 CNRS/MNHN/Paris6 CR2P, CP 38, 57 rue Cuvier, F75231, Paris Cedex 05, France; e-mail: [email protected] of Natural Sciences of Drexel University, 1900 Benjamin Franklin Parkway, Philadelphia, PA 19103, USA; e-mail: [email protected] of Evolution and Development, Department of Organismal Biology, Uppsala University, Norbyv€agen 18A, SE 752 36, Uppsala, Sweden6Current address: Research School of Physics and Engineering, College of Physical and Mathematical Sciences, Building 84, The Australian National University,
Acton, ACT 2601, Australia; e-mail: [email protected]
*Corresponding authors
Typescript received 14 April 2015; accepted in revised form 27 June 2015
Abstract: The placoderm fauna of the late Famennian tet-
rapod-bearing locality of Strud, Belgium, is studied on the
basis of historical and newly collected material. It includes
the previously described antiarch Grossilepis rikiki, the groen-
landaspidid Turrisaspis strudensis sp. nov. and the actinolepi-
doideid Phyllolepis undulata. P. undulata is thoroughly
described and joins the list of the valid Phyllolepis species
confidently diagnosed. A morphometrical analysis performed
on the centronuchal and anterior ventrolateral plates of the
Phyllolepis material demonstrates that there is only one
species of Phyllolepis in Belgium (thus, Phyllolepis konincki
becomes a junior synonym of P. undulata), that P. rossimon-
tina (Pennsylvania) is a synonym of P. undulata and that the
unity of the genus Phyllolepis is strongly supported, although
the characterization of several species within this genus is
blurred. The strong resemblance between the faunal compo-
sitions in Strud and Red Hill (Pennsylvania, USA) suggests
important faunal exchanges between these regions of the
Euramerica landmass.
Key words: Famennian, palaeobiogeography, Phyllolepis,
Red Hill, Strud, Turrisaspis.
S INCE its rediscovery in 2004, the tiny quarry of Strud
(Namur Province, Belgium) has been the subject of
intensive excavation and the driving force for a reconsid-
eration of the Famennian palaeontological record in
Belgium (see Olive et al. 2015a for a summary). The site
has yielded a very abundant flora (Table 1; Prestianni
et al. 2007) but also an exquisitely preserved invertebrate
fauna (Table 1) including one of the oldest putative
insects (Garrouste et al. 2012, 2013; H€ornschemeyer et al.
2013) and continental crustaceans (Gueriau et al. 2014a,
b; Lagebro et al. 2015). The vertebrate fauna from Strud
(Table 1) consists of sarcopterygian taxa (Lohest 1888a,
b; Leriche 1931; Cl�ement et al. 2004; Cl�ement and
Boisvert 2006), as well as yet undescribed acanthodian
and actinopterygian taxa. The placoderm fauna has been
partly described recently, with the antiarch Grossilepis
rikiki Olive, in press. Phyllolepid placoderms have been
considered in the past (Lohest 1888b; Leriche 1931), but
need a complete re-evaluation, which was one of the aims
of this article.
Phyllolepid placoderms are easily recognizable by the
presence of a ‘centronuchal’ plate of still-debated
homologies (Dupret and Zhu 2008) and a ridged dermal
ornamentation. Their most basal member, Gavinaspis, is
dated from the Lochkovian of China (Dupret and Zhu
2008). Other phyllolepids are known from the Middle
Devonian of Gondwana with the genera Austrophyllolepis,
Placolepis, Cobandrahlepis, Cowralepis and Yurammia. Late
Devonian (Famennian) phyllolepids from Euramerica all
belong to the genus Phyllolepis (Young 2005a, b, c;
Dupret and Zhu 2008; Long and Daeschler 2013).
The arthrodire groenlandaspidid genus Turrisaspis was
first described by Daeschler et al. (2003) from the late
Famennian Catskill Formation at the Red Hill site in
Pennsylvania, USA. They justified the erection of a new
genus based on the high shape of the median dorsal plate
© The Palaeontological Association doi: 10.1111/pala.12190 981
[Palaeontology, Vol. 58, Part 6, 2015, pp. 981–1002]
Page 2
and the tall and narrow plates of the lateral trunk shield.
To date, it has been considered to be endemic to Pennsyl-
vania. A short history of the family Groenlandaspididae is
available in Olive et al. (2015a).
Here, we present the placoderm assemblage found in the
Famennian tetrapod-bearing locality of Strud. The nomen-
clatural issue of Belgian Phyllolepis species is resolved and
the species definition within this genus questioned. A new
species of the genus Turrisaspis is recognized in Strud.
Institutional abbreviations. ANSP, Academy of Natural Sciences
of Philadelphia, Pennsylvania, USA; IRSNB, Institut royal des
Sciences naturelles de Belgique, Brussels, Belgium; NHMD, Nat-
ural History Museum of Denmark, Copenhagen, Denmark;
PALULG, Palaeontological collections of the Universit�e de Li�ege,
Li�ege, Belgium; UCL, Universit�e Catholique de Louvain-
la-Neuve, Louvain-la-Neuve, Belgium.
Anatomical abbreviations. ADL, anterior dorsolateral plate; AL,
anterior lateral plate; AMV, anterior median ventral plate; APi,
anterior pineal plate; AVL, anterior ventrolateral plate; bhy, basi-
hyal; CeNu, centronuchal plate; chy, ceratohyal; IL, interolateral
plate; MD, median dorsal plate; PDL, posterior dorsolateral plate;
PMV, posterior medioventral plate; PN, postnasal plate; PNu,
paranuchal plate; PrO, preorbital plate; Psph, parasphenoid;
PtO, postorbital plate; PVL, posterior ventrolateral plate; Sp,
spinal plate.
GEOLOGICAL SETTING
The stratigraphical section at Strud belongs to the Haine–Sambre–Meuse Overturned Thrust Sheets (HSM OTS),
which is a Variscan structural element of the Namur–Dinant Basin (Denayer et al. in press; Fig. 1A). During
the late Famennian times, this region was located along
the south-eastern margin of Euramerica, and in particular
in the London–Brabant Peninsula (Ziegler 1990). The
observed facies in the HSM OTS are typically continental
(Thorez et al. 2006). The Strud section is composed of a
TABLE 1 . Strud flora and fauna.
Vertebrata
Placodermi
Arthrodira
Phyllolepididae
Phyllolepis undulata
Groenlandaspididae
Turrisaspis strudensis
Antiarcha
Bothriolepididae
Grossilepis rikiki
Actinopterygii
Actinopterygii indet.
Sarcopterygii
Dipnomorpha
Dipnoi
Soederberghia cf. S. groenlandica
Jarvikia sp.
Dipteridae indet.
Porolepiformes
Holoptychius sp.
Tetrapodomorpha
‘Osteolepiformes’
Litoptychius-like
Tristichopteridae indet.
Cosmine-covered ‘osteolepididae’
Tetrapoda
Ichthyostega-like tetrapod
Acanthodii
Acanthodii indet.
Arthropoda
Hexapoda
Insecta
Strudiella devonica
Crustacea
Malacostraca
Schramidontus labasensis
Tealliocaris walloniensis
Branchiopoda
Strudops goldenbergi
Anostraca indet.
Conchostraca indet.
Chelicerata
Merostomata
Eurypterida indet.
Plantae
Tracheophyta
Zosterophyllopsida
Barinophytales
Barinophyton citrulliforme
Filicopsida
Zygopteridales
Rhacophyton condrusorum
Progymnospermopsida
Archaeopteridales
Archaeopteris halliana
(continued)
TABLE 1 . (Continued)
Gymnospermopsida
Condrusia rumex
Pseudosporogonites hallei
Moresnetia zalesskyi
Incertae sedis
indet. microsporangia
Sphenopteris flaccida
Sphenopteris modavensis
Vertebrata: Cl�ement et al. (2004), Cl�ement and Boisvert (2006),
Cl�ement and Prestianni (2009), Olive (in press). Arthropoda:
Garrouste et al. (2012), Gueriau et al. (2014a, b), Lagebro et al.
(2015). Plantae: Prestianni et al. (2007), Cl�ement and Prestianni
(2009), Denayer et al. (in press).
982 PALAEONTOLOGY , VOLUME 58
Page 3
road-cut exposure and of two disused quarries. The
smallest quarry (the northern one) was exploited during
the second half of the nineteenth century. It had been
abandoned and subsequently rediscovered in 2004, thanks
to geologists from Li�ege University (Laurent Barchy and
Jean-Marc Marion). In this quarry, the beds are in reverse
succession and the dipping is 80° southward. The fossilif-
erous strata of this quarry (lithological unit UL7) belong
to the Upper Devonian Evieux Formation (Denayer et al.
in press; Fig. 1B, C). They were first considered to be
upper middle Famennian (GF Biozone, Prestianni et al.
2007) because of the absence of some typical palynologi-
cal markers. However, the plant association and the
lithostratigraphy, as well as a revision of the palynological
markers, argue preferentially for a late Famennian age
(VCo ‘rad’ zone, Denayer et al. in press). The sedimentol-
ogy of the lithological unit 7 corresponds to the filling
sequence of a channel in an alluvial plain (Garrouste
et al. 2012; Denayer et al. in press). Placoderm remains
are found throughout the lithological unit UL7, except in
its uppermost part (Fig. 1C). The top of the sequence
corresponds to the final phase of the channel filling. This
corresponds to fresh or brackish water ponds seasonally
dried and flooded, which permitted the preservation of
delicate invertebrates and small actinopterygian remains
unpreserved in the rest of the sequence. Heavier fossils,
such as placoderm remains, are found in the immediately
underlying beds.
The placoderm material discussed herein comes from
Strud, except for a few specimens from Modave, Ch�evre-
mont and Evieux. All these localities belong to the Mont-
fort/Evieux Formation, Famennian. Modave and Evieux
belong to the Dinant synclinorium, whereas Ch�evremont
belongs to the Vesdre area, both being Variscan (Hercy-
nian) structural elements (Fig. 1A). The fossil material
consists of disarticulated head and thoracic elements,
except for one phyllolepid specimen partly preserved in
articulation (PALULG.2014.01.29.16 from Ch�evremont).
MATERIAL AND METHODS
The material used in the morphometric analysis includes
specimens from various international collections. The
material from the ANSP palaeontological collections and
the material from Belgian institutions have been directly
measured, and the rest of the material has been measured
based on figures in the literature (including a few lost
Belgian specimens used by Lohest (1888a) and Leriche
(1931)). Measurement methods are those of Daeschler
et al. (2003, fig. 10) for median dorsal plates of Turrisas-
pis, and those of Figure 2 for anterior ventrolateral and
centronuchal plates of Phyllolepis. As noticed by Young
(2005b, p. 209), Stensi€o (1939, p. 7) did not clearly
defined the anterior ventrolateral plate measurements he
used. Young (2005b, c) clearly defined some of those
measurements but did not take into account all features
used by Stensi€o such as the lengths of the medial or of
the posterior margins for instance. Therefore, we decided
to redefine the measurements used by Stensi€o (1939).
Measurements are given in Olive et al. (2015b, appendices
1–2). Material recently collected has been used in this
analysis as well as Lohest’s and Leriche’s figured and mea-
surable material. Not all measurements were possible
depending on plate preservation (see Olive et al. 2015b,
appendices 1–2) ‘NM’ for nonmeasurable). Thus, sample
size may vary from one graphic to another.
SYSTEMATIC PALAEONTOLOGY
This published work and the nomenclatural act it contains have
been registered in ZooBank: http://zoobank.org/References/
835B725E-7824-442D-B283-2CBC614EE92C
Order ARTHRODIRA Woodward, 1891
Suborder PHLYCTAENIOIDEI Miles, 1973
Infraorder PHLYCTAENII Miles, 1973
Family GROENLANDASPIDIDAE Obruchev, 1964
Genus TURRISASPIS Daeschler et al., 2003
Type species. Turrisaspis elektor Daeschler et al., 2003. Red Hill,
Clinton County, Pennsylvania, USA, Duncannon Member, Cats-
kill Formation, late Famennian (Fa2c substage), Late Devonian.
Turrisaspis strudensis sp. nov.
Figure 3
1965 Tiaraspis sp.; Gross, abb. 4A, taf. 2, figs 2–3
(median dorsal plate PALULG A/2005)
2005 Groenlandaspis thorezi (partim); Janvier and
Cl�ement, pl. 2C (median dorsal plate PALULG
A/2005, the same median dorsal plate that was
attributed to Tiaraspis sp. by Gross (1965)).
LSID. urn:lsid:zoobank.org:act:D4668CEA-9F3C-4BA8-8769-
C45B9579C6E1
Derivation of name. In reference to the locality of Strud, where
the material was found.
Holotype. IRSNB P.9447, a posterior dorsolateral plate (Strud).
Other material. Strud: APi: IRSNB P.9448; MD: IRSNB P.9449,
IRSNB P.9450, IRSNB vert 32.438-001, PALULG A/2005,
UCL.P.V.L.10.532; AL: IRSNB P.9451–9452, IRSNB vert 32.220-
OL IVE ET AL . : P LACODERMS FROM STRUD 983
Page 4
A
B
C
984 PALAEONTOLOGY , VOLUME 58
Page 5
001–002, IRSNB vert 32.438-002; AVL: IRSNB P.9453–9454,IRSNB vert 32.164-003, IRSNB vert 32.220-003–004, IRSNB vert
32.745-001; PMV: IRSNB P.9455; PVL: IRSNB P.9456–9457,IRSNB vert 32.745-002; Sp: IRSNB P.9458, IRSNB vert 32.438-
003.
Modave: MD: PALULG A/2005.
Diagnosis. Turrisaspis species with 90° angle of the bend
in the main lateral sensory line groove on the PDL;
overlapping area of the PDL for the ADL posteriorly long;
PVL with a large posterior part of the ventral lamina; Sp
with a large anteriorly directed spinelet at the middle of
the lateral margin.
Localities and horizons. Strud, Haltinne District, Gesves Com-
mune, Namur Province, Belgium; Evieux Formation (Denayer
et al. in press), VCo ‘rad’ zone, late Famennian, Upper Devo-
nian.
Modave, Li�ege Province, Belgium; upper part of the Evieux
Formation, Famennian, Upper Devonian.
Morphological description of the skull roof
Anterior pineal plate. (APi, Fig. 3A) The anterior pineal plate is
a trapezoidal plate with the posterior margin concave and
shorter than the anterior one. The pineal foramen is located at
the geometric centre of the plate. The ornamentation consists of
numerous closely spaced, small tubercles.
Morphological description of the thoracic armour
Median dorsal plate. (MD, Fig. 3B–E) The median dorsal plate
is very high and narrow. The tip of the plate is pointed and
forms an angle varying between 25° (specimen IRSNB P.9449,
Fig. 3C) and 39° (largest specimen UCL.P.V.L.10.532, Fig. 3B)
depending on the specimen. Three of four median dorsal plates
(IRSNB P.9449, P.9450 and PALULG A/2005, Fig 3C–E) are
roughly the same size, and presumably of the same ontogenetic
stage but differ in the tip angle. It is thus unclear whether a rela-
tionship between the size of the plate and tip angle exists. How-
ever, it seems by comparison with the largest median dorsal
plate (UCL.P.V.L.10.532, Fig. 3B) that the angle opened with
growth, as observed in Groenlandaspis riniensis (Long et al. 1997,
p. 258). The posterior edge of the plate displays a gentle slope
and is covered with large, pointed spinelets whose apices are of
a different colour to the plate itself (i.e. darker; on
UCL.P.V.L.10.532 only, Fig. 3B), suggesting that it is likely to be
a different type of tissue. The same observation is made for the
tips of several tubercles covering the same MD. The anteroven-
tral projection of the median dorsal plate is much narrower than
the posterior one. The tip of the plate can also bear spinelets, as
the posterior edge does (see specimen PALULG A/2005, Fig. 3E).
The anterior margin of the median dorsal plate is straight or
slightly concave and devoid of spinelets. The ventral margin of
the plate is deeply indented by overlap areas of the anterior and
posterior dorsolateral plates. The ornamentation of the sides of
the median dorsal plate consists of small tubercles organized in
sinuous rows parallel to the ventral margin.
A
B
F IG . 2 . Measurement methods for the Phyllolepis plates. A,
CeNu. B, AVL. Abbreviations: L, length; L.ant, anterior length;
L.mg.ant, anterior margin length; L.mg.lat, lateral margin length;
L.mg.med, medial margin length; L.mg.post, posterior margin
length; W, width.
F IG . 1 . Location, geological and stratigraphical framework of the Strud locality and location and geological framework of the Mod-
ave, Evieux and Ch�evremont localities. A, geological map of southern Belgium with the location of the Strud, Modave, Evieux and
Ch�evremont localities. B, lithological and lithostratigraphical column of the Strud locality. C, schematic block diagram of the Strud
channel with the location of placoderm remains. Abbreviations: HSM OTS, Haine–Sambre–Meuse Overturned Thrust Sheets; Lux, Lux-
emburg; UL, lithological unit. Modified after Denayer et al. (in press). Colour online.
OL IVE ET AL . : P LACODERMS FROM STRUD 985
Page 6
A
F
K L O
Q
R
NM
G H I
J
P
B C D E
F IG . 3 . Turrisaspis strudensis sp. nov. A–D, F–Q, from Strud, and E, from Modave. A, APi in external and internal views,
IRSNB P.9448a. B, MD in left external view, UCL.P.V.L.10.532. C, MD in left external view, IRSNB P.9449a. D, MD in left
external view, IRSNB P.9450a. E, MD in right external view, PALULG A/2005. F, right PDL in external view, IRSNB P.9447a.
G, right PDL in external view, IRSNB P.9447b. H, right PDL in external view, IRSNB P.9447, interpretative drawing. I, right AL
in external view, IRSNB P.9451. J, left AL in external view, IRSNB P.9452a. K, left AVL in external and internal views, IRSNB
P.9453a. L, left AVL in external and internal views, IRSNB P.9454. M, PMV in external view, IRSNB P.9455a. N, left PVL in
external view, IRSNB P.9457a. O, left PVL in external view, IRSNB P.9456a. P, proximal part of a right Sp in ventral view,
IRSNB P.9458a. Q, distal part of a right Sp in dorsal view, IRSNB P.9458b. R, cross section of a right Sp, IRSNB P.9458a,
indicated by a double arrow on P. Dashed line for assumptions; dashed line plus dots for information provided by counterparts.
Abbreviations: a.b, annular bourrelet/thickening; ad.cr, anterodorsal crest of anterior lateral plate; av.cr, anteroventral crest of
anterior lateral plate; av.s, anteroventral sulcus; emb.ADL, embayment of median dorsal plate for anterior dorsolateral plate;
emb.PDL, embayment of median dorsal plate for posterior dorsolateral plate; f.pi, pineal foramen; gr.Sp, ventral groove of spinal
plate; ins.Sp, insertion area for spinal plate; lat, laterally; oa.ADL, overlap area for anterior dorsolateral plate; oa.AL, overlap area
for anterior lateral plate; oa.AMV, overlap area for anterior medioventral plate; oa.AVL, overlap area for anterior ventrolateral
plate; oa.IL, overlap area for interolateral plate; oa.MD, overlap area for median dorsal plate; oa.PMV, overlap area for posterior
medioventral plate; oa.PVL, overlap area for posterior ventrolateral plate; obs.l, obstantic lamina of anterior lateral plate; pbr.l,
postbranchial lamina of anterior lateral plate; pd.cr, posterodorsal crest of anterior lateral plate; pec.n, pectoral notch. Scale bar
represents 1 cm. Colour online.
986 PALAEONTOLOGY , VOLUME 58
Page 7
Posterior dorsolateral plate. (PDL, Fig. 3F–H) The only available
posterior dorsolateral plate (part and counterpart) is slightly
deformed but appears narrower than the material from Pennsyl-
vania (although this could be due to an earlier ontogenetic
stage). The posterodorsal process of the ornamented area is pro-
nounced and angles upward, forming an apron and thus resem-
bling the condition observed in the material of Groenlandaspis
potyi (see Olive et al. 2015a, fig. 2) as well as that of Turrisaspis
elektor from Pennsylvania (although to a lesser extent; see
Daeschler et al. 2003, fig. 9E). The posterior edge of the over-
lap area for the MD plate is situated anteriorly to the posterior
edge of the plate, while it is at the same level in T. elektor.
Below this apron, the posterior edge of the plate is convex
(while concave in T. elektor). The area overlapped by the med-
ian dorsal plate is high and well preserved on the counterpart
and does not reach the posterior margin (i.e. the apron is free).
The main lateral sensory line groove forms an angle of exactly
90° on the holotype. The area overlapped by the anterior
dorsolateral plate extends more sharply and posteriorly than
in T. elektor. The ornamentation consists of small, evenly
distributed tubercles.
Anterior lateral plate. (AL, Fig. 3I–J) The shape of the AL seems
to vary during ontogeny. The anterodorsal edge is rounder on the
smallest specimen than on the largest plate; this ensures a slightly
convex aspect of the dorsal edge of the smallest plate, while the
largest one displays a straighter dorsal edge. The anteroventral
and posterodorsal crests are sharper on the largest plate. The
ventral edge is straight and longer than the dorsal edge. The
posterior margin is concave, and the posterodorsal corner of the
plate is the most posterior part of the plate. The ornamentation
consists of numerous small tubercles distributed randomly.
Anterior ventrolateral plate. (AVL, Fig. 3K–L) The AVL is the
most frequently found plate and the best preserved. The
V-shaped depression is well defined and opens on the pectoral
notch, which forms a large embayment. The ornamentation of
the AVL consists mostly of small scattered tubercles.
Posterior median ventral plate. (PMV, Fig. 3M) The PMV is
enlarged in its posterior part with wide overlapped areas for all
neighbouring plates. The exposed part of the plate is elongate
and covered with small tubercles.
Posterior ventrolateral plate. (PVL, Fig. 3N–O) The ventral lam-
ina of the posterior ventrolateral is elongated. On the cranial
margin, the area overlapped by the anterior ventrolateral plate
is crescent-shaped and quite long. The plate presents the typi-
cal sigmoid flexure and mutual overlap seen in the Phlyctaenii
(except for the family Phlyctaenididae): the right plate over-
laps the left one, while the latter protrudes a short process
overlapping the right plate. This sigmoid flexure is sharper
and simpler (one process) on T. strudensis than it is on
T. elektor where it is less pronounced (see Daeschler et al.
2003, fig. 9R). The angle between the ventral and lateral lami-
nae cannot be measured because of the strong compression of
the specimen. The ventral surface displays a well-marked
annular bourrelet of Goujet (1984) circling the posterior ring
of the thoracic armour. The dermal ornament consists of scat-
tered tubercles.
Spinal plate. (Sp, Fig. 3P–R) IRSNB P.9458 is broken into two
parts, the posterior one corresponding roughly to the ‘free’ part
of the plate and the anterior one contacting the AVL. The latter
displays two types of ornamentation on the lateral side, divided
by a large anteriorly pointed spinelet at mid-length. Anterior to
the large spinelet, the visible ornamentation consists of three
rows of large tubercles; past the spinelet, two rows remain with
smaller and spiniform tubercles. The free posterior medial mar-
gin displays medially oriented spinelets larger and more widely
spaced than the lateral ones, as in Turrisaspis elektor. The rest of
the plate is smooth. In cross section, the angle made by the
asymptotic lines of the ventral and dorsal laminae is 30°(Fig. 3R). The continuation of the anteroventral sulcus (of the
interolateral plate) is visible on the ventral side of the spinal
plate (see Goujet 1984; Dupret 2003, 2010).
Remarks
The material is referred to the genus Turrisaspis based on
the high and narrow median dorsal and posterior dorso-
lateral plate, as well as a short posterior projection
(‘apron’) of the posterior dorsolateral plate which does
not allow the overlap area for the median dorsal plate to
reach the posterior corner of the plate, a feature shared
by Turrisaspis and Africanaspis (Long et al. 1997). The
attribution to Africanaspis is not considered because the
typical lateral ridge on the posterior dorsolateral plate is
not present in the plate from Strud. Moreover, in Africa-
naspis the dorsal process of its posterior dorsolateral plate
is much lower than that of T. strudensis.
Several characteristics of T. strudensis differ from the
description of T. elektor: (1) the angle of the sensory line
groove on the posterior dorsolateral plate being 90° in
T. strudensis rather than 70° in T. elektor; (2) the over-
lapping area of the posterior dorsolateral plate for the
anterior dorsolateral plate is longer posteriorly in T. stru-
densis; (3) the spines on the posterior margin of the
median dorsal plate are more widely spaced and not as
numerous in the Belgian specimens; (4) the posterior part
of the ventral lamina of the posterior ventrolateral plate is
larger in T. strudensis; and (5) the mainly smooth spinal
plate of T. strudensis displays larger spinelets on the free
posterior medial margin and with a very large spinelet at
the middle of the lateral margin. Although point 3 may be
due to individual variation within T. elektor according to
Daeschler et al. (2003, p. 56: ‘individual variation is also
evident in the median dorsal plate sample with respect to
details of the serrations along the caudal margin’), we
assert here that the other features diverging from the type
species of the genus Turrisaspis are relevant and numerous
enough to justify the erection of a new species.
OL IVE ET AL . : P LACODERMS FROM STRUD 987
Page 8
Olive et al. (2015a) questioned the attribution of the
median dorsal plate published by Gross (1965, abb. 4A)
(Fig. 3E) to Tiaraspis, and later redescribed by Janvier and
Cl�ement (2005) as Groenlandaspis thorezi. Olive et al.
(2015a) did not agree with that assumption because: (1) ‘the
specimens do not match with each other’; and (2) they ‘con-
cur with Gross (1965) who recognized the ventral lamina as
an overlap area’. The newly described material of Turrisaspis
from Strud allows the attribution of this problematic median
dorsal plate to this taxon, based on the numerous shared fea-
tures (i.e. high and narrow shape, posterior margin with a
gentle posteroventral slope and with pointed spines, ventral
margin deeply indented by the posterior dorsolateral plate,
and ornamentation made of small tubercles arranged in
sinuous rows parallel to the ventral margin).
Suborder ACTINOLEPIDOIDEI Miles, 1973
Infraorder PHYLLOLEPIDIDA Stensi€o, 1934
Family PHYLLOLEPIDIDAE Woodward, 1891
Genus PHYLLOLEPIS Agassiz, 1844
Type species. Phyllolepis concentrica Agassiz, 1844. Perthshire and
Rosebrae beds, Morayshire, Scotland, Famennian.
Phyllolepis undulata Lohest, 1888b
Figures 4–7
1888b Phyllolepis undulatus Lohest, p. 48.
1888b Phyllolepis corneti Lohest, p. 48.
1888b Pentagonolepis konincki Lohest, p. 52.
1931 Phyllolepis undulata; Leriche, p. 9.
1939 Phyllolepis konincki; Stensi€o, p. 19.
2005 Phyllolepis rossimontina Lane and Cuffey.
2013 Phyllolepis konicki [sic]; Long and Daeschler, fig. 6.
2013 Phyllolepis undiulata [sic]; Long and Daeschler, fig. 6.
Type specimen. PALULG.2014.04.08.10, median dorsal plate, lec-
totype defined by Leriche (1931, pl. 1, figs 3, 3a) and figured
before by Lohest (1888b, pl. 10, fig. 5).
Other material. Strud: PrO: IRSNB P.9459; PN: IRSNB P.9460,
IRSNB vert 32.745-003–004; PtO: IRSNB P.9461, IRSNB vert
32.220-005; PNu: IRSNB P.9462–9463, IRSNB vert 31.264-001–005, IRSNB vert 31.913-003–005, IRSNB vert 32.164-004, IRSNB
vert 32.164-018, IRSNB vert 32.220-006–011, IRSNB vert 32.438-
004–009, IRSNB vert 32.745-005; CeNu: IRSNB P.9464–9468,IRSNB vert 31.264-006–015, IRSNB vert 31.595-001, IRSNB vert
31.913-006, IRSNB vert 32.164-005–007, IRSNB vert 32.220-012–021, IRSNB vert 32.438-010–014, IRSNB vert 32.745-006,
PALULG.2014.04.08.19, PALULG.2014.04.08.20; bhy, chy & Psph:
IRSNB P.9469; MD: IRSNB P.9470–9473, IRSNB vert 31.264-016–030, IRSNB vert 31.595-002–009, IRSNB vert 31.913-007–008,IRSNB vert 32.048-001, IRSNB vert 32.164-008–009, IRSNB vert
32.220-022–031, IRSNB vert 32.438-015–023, IRSNB vert 32.745-
007–008, PALULG.2014.04.08.11, PALULG.2014.04.08.17; ADL:
IRSNB P.9474, IRSNB vert 31.913-009, IRSNB vert 32.164-010,
IRSNB vert 32.220-032–034, IRSNB vert 32.438-024–026, IRSNBvert 32.745-009; AL: IRSNB P.9475; AVL: IRSNB P.9476–9482,IRSNB vert 31.264-031–049, IRSNB vert 31.595-010–011, IRSNBvert 31.913-010–015, IRSNB vert 32.048-002, IRSNB vert 32.164-
011–013, IRSNB vert 32.220-035–036, IRSNB vert 32.438-027–032, PALULG.2014.04.08.16, PALULG.2014.04.08.15, PALULG.
2014.04.08.18; PVL: IRSNB P.9483–9484, IRSNB vert 31.264-050–056, IRSNB vert 31.595-012–013, IRSNB vert 32.048-003, IRSNB
vert 32.164-014–017, IRSNB vert 32.220-037–041, IRSNB vert
32.438-033–040.
Ch�evremont: Articulated specimen: PALULG.2014.01.29.16;
CeNu: PALULG.2014.04.08.20; MD: PALULG.2014.04.08.12; AVL:
PALULG.2013.05.3.11, PALULG.2014.01.29.19; PVL: PALULG.20
14.01.29.17, PALULG.2014.01.29.18.
Evieux: MD: PALULG.2014.04.08.13; AVL: PALULG.2014.04.
08.14.
Red Hill: See Lane and Cuffey (2005), Long and Daeschler
(2013) and Olive et al. (2015b, appendices 1–2).
Tioga site: CeNu: ANSP 23360 (Olive et al. 2015b,
appendix 1).
F IG . 4 . Phyllolepis undulata, from Strud. A, right PrO in external view, IRSNB P.9459. B, left PtO in external view, IRSNB P.9461a.
C, right PN in internal view, IRSNB P.9460. D, right PNu in external view, IRSNB P.9462. E, PNu in external and internal views,
IRSNB P.9463a. F, CeNu in external and internal views, IRSNB P.9467. G, CeNu in external view, IRSNB P.9466a. H, CeNu in exter-
nal view, IRSNB P.9464. I, CeNu in external and internal views, IRSNB P.9465a. J, CeNu in internal view, IRSNB P.9468. K, MD in
external view, IRSNB P.9470. L, MD in external and internal views, IRSNB P.9471. M, MD in internal view, IRSNB P.9472. N, MD in
internal view, IRSNB P.9473. O, left ADL in external view, IRSNB P.9474a. P, right AL in ventral view, IRSNB P.9475. Q, right PVL
in external and internal views, IRSNB P.9483. R, left PVL in internal view, IRSNB P.9484. Abbreviations: ab.at, abdominal muscle
attachment of median dorsal plate; bu, central bulge of centronuchal plate; csc, central sensory canal; de, central depression on cen-
tronuchal plate visceral surface; dsc, dorsal sensory canal; ifc, infra-orbital canal; ins.Sp, insertion area for spinal plate; lc, main lateral
line; oa.ADL, overlap area for anterior dorsolateral plate; oa.AL, overlap area for anterior lateral plate; oa.CeNu, overlap area for cen-
tronuchal plate; oa.MD, overlap area for median dorsal plate; oa.PDL, overlap area for posterior dorsolateral plate; oa.PNu, overlap
area for paranuchal plate; oc.os, occipital ossification attachment of the centronuchal plate; pbr.l, postbranchial lamina of anterior lat-
eral plate; ppl, posterior pit line; soc, supraorbital canal; syn.at, synarcual attachment of the median dorsal plate; vsc, V-shaped canal
of postnasal plate. Scale bars represent 1 cm. Colour online.
988 PALAEONTOLOGY , VOLUME 58
Page 9
Diagnosis. This is the first formal diagnosis for this spe-
cies, as diagnoses were not compulsory when it was first
named and described in 1888.
V-shaped sensory line canal of the postnasal plate
reaching the centronuchal plate; centronuchal plate shape
displaying a variability stretching from longer than broad
to broader than long (shape variation continuous);
anterolateral margin of the centronuchal plate either well
defined or included in the convex anterior margin which
may be sinusoidal; external surface of the centronuchal
plate bulged in the middle; visceral surface of the cen-
tronuchal plate with a tiny depression in the centre and
with a longitudinal median groove running from the cen-
tre to the posterior margin and corresponding to the
occipital ossification; basihyal elongate and oak-leaf-
shaped; visceral surface of the median dorsal plate with
A
F
K
L
O Q R
NM
G H I
J
P
B C D E
OLIVE ET AL . : P LACODERMS FROM STRUD 989
Page 10
an anteroposterior ridge corresponding anteriorly to the
synarcual attachment area; anterior ventrolateral plate
slightly longer than broad in the majority of cases; ante-
rior ventrolateral plate following an isometric growth;
ornamentation made of concentric parallel lines on all
plates and of a largely meshwork reticulated pattern in
the anterior part of the centronuchal plate.
Localities and horizons. Strud, Haltinne, Gesves, Namur Pro-
vince, Belgium, Evieux Formation (Denayer et al. in press), VCo
‘rad’ zone, late Famennian, Upper Devonian.
Evieux, Esneux, Li�ege Province, Belgium, Montfort/Evieux
Formation, Famennian, Upper Devonian.
Ch�evremont, Li�ege Province, Belgium, Montfort/Evieux For-
mation, Famennian, Upper Devonian.
Red Hill, road cut on route 120, 1 mile west of Hyner, Clin-
ton County, Pennsylvania, USA, Duncannon Member, Catskill
Formation, late Famennian (Fa2c substage), Late Devonian.
Tioga site, construction zone on route 15, just south of Tioga,
Tioga County, Pennsylvania, USA, Catskill Formation, late
Famennian, Late Devonian.
Morphological description of the skull roof
Preorbital plate. (PrO, Fig. 4A) The only known preorbital plate
is poorly preserved. The supraorbital canal is well marked, and
the overlap area for the centronuchal plate is narrow.
Postnasal plate. (PN, Fig. 4B) The postnasal plate is pentagonal
in shape and displays the typical V-shaped sensory line canal,
which reaches the area overlapped by the centronuchal plate.
Postorbital plate. (PtO, Fig. 4C) The sole specimen of a postor-
bital plate displays the anterior part only. The intersection
between the main lateral line, central sensory and infra-orbital
canals is observable. The areas overlapped by the centronuchal
and paranuchal plates are narrow.
Paranuchal plate. (PNu, Figs 4D–E, 5) The paranuchal plate is
characteristic of the genus Phyllolepis with the anterior extremity
of the main lateral line located at about half-length of the lateral
side and separating a larger area in front of the main lateral
line and a smaller one posteriorly (Long 1984). The postnuchal
process is thin, and the area overlapped by the centronuchal
plate is narrow. No median process (interpreted as a vestigial
endolymphatic duct; see Dupret and Zhu 2008) has been identi-
fied. The ornamentation consists of well-defined concentric
parallel lines.
Centronuchal plate. (CeNu, Figs 4F–J, 5) The centronuchal plate
of P. undulata is slightly broader than long in the majority of
specimens (73%) and sometimes as long as broad (21%). How-
ever, two of the smallest plates of the sample (IRSNB P.9467
and IRSNB vert 32.164-007, Fig. 4F) display a centronuchal plate
longer than broad (6%). The anterolateral corner is either well
defined (Phyllolepis-like) or part of the curved and convex ante-
rior edge (Placolepis-like), which can occasionally be sinusoidal
(see Discussion for further details). A low bulge is consistently
found at the radiation centre in all specimens. The supraorbital
and the central sensory canals are well defined contrary to the
posterior pit lines, which are faint impressions at best or present
in more superficial layers of the dermis. On the smallest individ-
uals, the supraorbital canals and posterior pit lines are consis-
tently invisible, and the central sensory canals are faint. The
visceral surface is smooth except for the shallow central depres-
sion (which corresponds externally to the bulge). Moreover, a
longitudinal median groove runs from this central depression to
the posterior margin as in Phyllolepis orvini (Stensi€o 1934, p. 46),
and may correspond to the occipital ossification attachment
(Ritchie 2005, fig. 13) The ornamentation consists of a large
reticulated pattern in the anterior half and straight ridges, parallel
to posterior margin, in the posterior half.
A B C
F IG . 5 . Phyllolepis undulata, from Ch�evremont. A, partly articulated specimen in external and internal views, PALULG.2014.01.29.16.
B, specimen partly in connection in external and internal views, PALULG.2014.01.29.16, interpretative drawing. C, MD, in external
view, PALULG.2014.04.08.12. Abbreviations: AVL, anterior ventrolateral plate; CeNu, centronuchal plate; dsc, dorsal sensory canal;
MD, median dorsal plate; PNu, paranuchal plate; PVL, posterior ventrolateral plate. Scale bars represent 1 cm. Colour online.
990 PALAEONTOLOGY , VOLUME 58
Page 11
Shape variations of the centronuchal plate are numerous; that
is straight or convex posterolateral margin, anterolateral corner
either well defined (Phyllolepis-like) or part of the curved and con-
vex anterior edge (Placolepis-like). However, the morphometrical
analysis (see below) does not permit discrimination between
different forms and the variations seem continuous. Therefore, all
centronuchal plates are gathered under P. undulata.
Parasphenoid, basihyal and ceratohyals
Three elements of the visceral skeleton are recovered in anatomi-
cal position on the part and counterpart of specimen IRSNB
P.9469 (Fig. 6) and are preserved as extremely flattened, dark
brown bone. The basihyal is on the part, and the parasphenoid
and ceratohyals are on the counterpart. The outlines of the skull
roof are very difficult to discern.
A B
C D
F IG . 6 . Phyllolepis undulata, from Strud. Parasphenoid, basihyal and ceratohyals. A, bhy in ventral view, part, IRSNB P.9469a. B, chy
and Psph in dorsal view, counterpart, IRSNB P.9469b. C, close-up of the bhy in ventral view, part, IRSNB P.9469a. D, three elements in
connection, IRSNB P.9469, interpretative drawing. Abbreviations: bhy, basihyal; chy, ceratohyal; m.ri, median ridge of basihyal; o.g, obli-
que groove of basihyal; Psph, parasphenoid. Scale bars represent: 1 cm (A–B, D); 2 mm (C). Colour online.
OL IVE ET AL . : P LACODERMS FROM STRUD 991
Page 12
The parasphenoid and ceratohyals are preserved as three
small discs; the parasphenoid is approximately four times
bigger than each ceratohyal. The basihyal shows a complex
outline, oak-leaf-like; the anterior tip is slender, while the
posterior part is broad. The only visible reliefs on the basihyal
consist in a median ridge in the anterior half, and an oblique
groove just posterior to it on the right side. The four
elements connect to each other when print and counterprint
are assembled.
Morphological description of the thoracic armour
Median dorsal plate. (MD, Figs 4K–N, 5) The median dorsal
plate is slightly broader than long in P. undulata. Only IRSNB
P.9471 (Fig. 4L) and PALULG.2014.04.08.12 (Fig. 5C) display
respectively the left or the right dorsal sensory line groove run-
ning from the anterolateral corner to the centre of the plate. The
visceral surface exhibits a longitudinal ridge surrounded by a
pair of shallow triangular depressions, forming a hourglass shape
with the thinner part in the middle or in the anterior half of the
plate (as figured by Stensi€o 1934, pl. 10). The anterior portion
of the hourglass shape most likely corresponds to the synarcual
attachment area, and the posterior portion may correspond to
abdominal muscle attachment. The ornamentation consists of
well-defined concentric parallel lines.
Anterior dorsolateral plate. (ADL, Fig. 4O) The anterior dorso-
lateral plate is only known by its smooth visceral side. The over-
lap area for the median dorsal plate is broader than that for the
anterior lateral plate; the overlap area for the paranuchal plate is
narrow.
Anterior lateral plate. (AL, Fig. 4P) The anterior lateral plate is
known from a single specimen. It is quadrangular in shape,
longer than high. The lateral margin is slightly convex, and the
posterior, lateral and anterior margins are straight.
Anterior ventrolateral plate. (AVL, Figs 5A–B, 7) The anterior
ventrolateral plate is slightly longer than broad in the majority
of specimens (87%) and sometimes as long as broad (9%) or
slightly broader than long (4%). The antero-mesial angle
(sensu Young 2005c, fig. 2B) is protruding anteriorly; this
corner can be straight or slightly rounded, implying a limited
contact with the minute anterior median ventral plate (if
present). The area overlapping the posterior ventrolateral plate
is very narrow (Fig. 7E). Growth series are available thanks to
the great number of specimens. The anterior ventrolateral
plate follows an isometric growth (Fig. 8A). The ornamenta-
tion consists of well-defined parallel lines, more widely spaced
anteriorly.
Posterior ventrolateral plate. (PVL, Figs 4Q–R, 5A–B) The poste-rior ventrolateral plate is triangular in shape. The antero-mesial
angle varies between 60 and 70°. The area overlapped by the
anterior ventrolateral plate is narrow mesially and slightly wider
laterally. The ornamentation consists of well-defined parallel lines.
Remarks
Concerning the paranuchal plates, IRSNB P.9463
(Fig. 4E) is very different from all other paranuchal plates
found in Strud (IRSNB P.9462 illustrating the most com-
mon shape; Fig. 4D). IRSNB P.9463 displays a straight
lateral margin and the portion of the plate anterior to the
lateral line canal is shorter than the posterior portion, a
condition absent in other paranuchal plates of Strud and
in other Phyllolepididae species (Long 1984, fig. 2). The
species with a paranuchal plate that is the most similar to
IRSNB P.9463 is Phyllolepis woodwardi, which shares a
similar posterolateral angle and posterior and posterolat-
eral straight edges, but the portion of the plate anterior
to the lateral line canal is much longer in P. woodwardi
(Stensi€o 1939; but for skull roof and paranuchal plate
comparisons, see Long 1984, figs 2A, 25; Ritchie 2005, fig.
20) than in P. undulata. As it is the largest of all phyl-
lolepid plates of the locality, IRSNB P.9463 may corre-
spond to an adult specimen. Whatever the case, as the
sample is small and as the variations could be due to
intraspecific variation, this plate is placed in P. undulata
for convenience.
Long and Daeschler (2013, p. 37) advocated ‘that the
identification of phyllolepid placoderms from disarticu-
lated material is only possible if a large associated
assemblage is known, as for Placolepis from Braidwood,
Australia (Ritchie 1984) or Phyllolepis species from East
Greenland (Stensi€o 1934, 1936, 1939)’. This is now the
case for P. undulata of which most plates of the trunk
shield are known. Long and Daeschler (2013) added
‘the identification of genera from isolated plates
remains uncertain based only on plate shape and
ornamentation unless one of the diagnostic plates is
present (marginal or paranuchal)’. Even if no marginal
plates of P. undulata are currently recognized, the para-
nuchal plates are numerous and correspond (except one
doubtful plate, see above) to Phyllolepis (paranuchal
with the anterior extremity of the main lateral line
located at about half-length of the lateral side (Long
1984)).
Long and Daeschler (2013, p. 37) listed the valid Phyl-
lolepis species that can be confidently diagnosed (i.e.
P. woodwardi, P. orvini, P. nielseni, P. rossimontina (here
put in synonymy with P. undulata) and P. thomsoni).
P. undulata can be added to this list.
DISCUSSION
How many Phyllolepis species in Belgium?
For the material found in Belgium, Lohest (1888b)
erected the genus Pentagonolepis and two new species of
992 PALAEONTOLOGY , VOLUME 58
Page 13
Phyllolepis: P. undulatus and P. corneti. At that time, no
articulated specimens of Phyllolepididae were known,
which led Lohest to define Pentagonolepis konincki with
anterior ventrolateral plates, Phyllolepis undulatus with
median dorsal plates and Phyllolepis corneti on the basis
of a sole centronuchal plate. Thanks to the discovery of a
complete specimen of Phyllolepis (Woodward 1915), Ler-
iche (1931) realized that the various plates were in fact
different parts of the same animal; Leriche thus put these
Lohest taxa in synonymy (under the name Phyllolepis
undulata and no longer ‘undulatus’, which was grammati-
cally incorrect). Later, on the basis of the material figured
by Lohest (1888b), Leriche (1931) and Stensi€o (1939)
assessed that there were two species of Phyllolepis in Bel-
gium. He assigned the material studied by Leriche and
Lohest to Phyllolepis undulata and to Phyllolepis konincki.
According to Stensi€o, P. konincki displayed a shorter and
broader centronuchal plate than that of P. undulata, a
median dorsal plate with the anterolateral margin slightly
forward directed, and an anterior ventrolateral plate
longer, in proportion to its width, than in P. undulata.
He also mentioned the ornamentation as a discriminating
character, with narrow ridges mostly situated further
apart in P. undulata than in P. konincki. This specific dis-
crimination character has recently been questioned by
Long and Daeschler (2013, p. 37), because ‘the nature of
A
F G
B
C
D E
F IG . 7 . Phyllolepis undulata, from Strud. Anterior ventrolateral plates. A, left AVL in external and internal views, IRSNB P.9476a.
B, left AVL in external and internal views, IRSNB P.9480. C, left AVL in internal view, IRSNB P.9479. D, left AVL in internal view,
IRSNB P.9477. E, left AVL in internal view, IRSNB P.9482. F, right AVL in external and internal views, IRSNB P.9481. G, left AVL in
external and internal views, IRSNB P.9478a. Abbreviations: oa.PVL, overlap area for posterior ventrolateral plate; pec.n, pectoral notch.
Scale bar represents 1 cm. Colour online.
OL IVE ET AL . : P LACODERMS FROM STRUD 993
Page 14
A
F
G H
B
C D
E
994 PALAEONTOLOGY , VOLUME 58
Page 15
the dermal ornamentation is quite variable within popula-
tions of phyllolepid species’.
A morphometrical analysis was carried out on cen-
tronuchal and anterior ventrolateral plates (Fig. 8) to
solve the problem of the number of Phyllolepis species
from Belgium. The majority of the material comes from
the locality of Strud, but a few plates from Ch�evremont
(Li�ege Province, Famennian) found in collection were
added to the morphometrical analysis. The centronuchal
and anterior ventrolateral plates were chosen for analysis
because they display a high degree of variability within
the Belgian material and are the most commonly encoun-
tered elements in the Famennian deposits of Belgium.
Moreover, those plates were used by Stensi€o (1939) to
justify the presence of two species of Phyllolepis in Bel-
gium. Thus, they appear to be the most appropriate
material to study the Phyllolepis diversity in Belgium.
Measurements of the centronuchal and anterior ventro-
lateral plates clearly show a distribution of points suggest-
ing that the size and shape variations are continuous
(R2 stretching from 0.8469 to 0.9434, Fig. 8). This result
argues against the interpretation of more than one species
within the sample. For instance, the overall shape of the
centronuchal plate (shorter and broader for P. konincki)
and of the anterior ventrolateral plate (longer and nar-
rower in P. konincki), used by Stensi€o as discriminating
characters (respectively Fig. 8A, D), are actually characters
within a range of continuous variations. Thus, we assume
that variations around the best-fit lines illustrate
intraspecific and ontogenetic variation. Consequently, we
assert that there is a sole species of Phyllolepis in Belgium.
The specific epithet ‘undulata’ is retained to follow
Leriche (1931), who acted as first reviewer, and thus
follow the article 24.2.1 of the International Code of
Zoological Nomenclature and the recommendation 24A,
which allow to ‘best serve stability and universality of
nomenclature’. Pentagonolepis konincki, Phyllolepis corneti,
Phyllolepis undulatus and Phyllolepis konincki are all
considered to be synonyms of P. undulata.
Phyllolepis rossimontina as a junior synonym of
Phyllolepis undulata
Lane and Cuffey (2005) described Phyllolepis rossimontina
from the Famennian deposits of the Catskill Formation,
at the Red Hill site, in north-central Pennsylvania. From
the same site, they recognized other specimens attributed
to various species of Phyllolepis: P. aff. concentrica,
P. aff. nielseni, P. aff. undulata and P. aff. woodwardi.
Later, several of these specimens were identified by Long
and Daeschler (2013) as specimens of P. rossimontina (see
Olive et al. 2015b, appendices 1–2 for details). To clarify
the situation, all the measurable anterior ventrolateral and
centronuchal plates of phyllolepids from the Catskill
Formation (Red Hill and Tioga sites) were integrated in
the morphometrical analysis (Fig. 9).
Figure 9 plots the different measurement ratios used.
The distribution of points for all graphs suggests that the
size variations are continuous, arguing against an inter-
pretation of more than one taxon within the sample,
based on the size. All points representing P. rossimontina
and the other species noticed by Lane and Cuffey (2005),
that is P. aff. nielseni, P. rossimontina?, P. sp. and
P. aff. concentrica, as well as unpublished material found
in the collection of the ANSP, fit within the distribution
of points from those representing P. undulata sensu this
article. This analysis suggests the presence of only one
species in the article of Lane and Cuffey (2005), which
corresponds to P. undulata found in Belgium.
Lane and Cuffey (2005) compared P. rossimontina with
P. undulata sensu Stensi€o (1939) and P. konincki sensu
Stensi€o (1939). According to them, P. rossimontina differs
from P. undulata by: (1) a centronuchal plate slightly
broader than long (roughly square in P. undulata); (2)
the curvature of the posterolateral margins of the cen-
tronuchal plate more pronounced; and (3) the posterolat-
eral corners of the same plate flared outward slightly.
Thanks to the morphometrical analysis (Fig. 9A), it is
clear that the first criterion is not correct, considering the
intraspecific and ontogenetic variations, and the fact that
the sample follows the best-fit line. The second and third
criteria are not directly explained by the morphometrical
analysis but as there are, within the sample, centronuchal
plates broader than long and centronuchal plates longer
than broad (for instance IRSNB P.9464, Fig. 4G), cen-
tronuchal plates displaying curved posterolateral margins
and flared posterolateral corners and some others not, it
could be related to intraspecific variation.
Concerning the anterior ventrolateral plate, Lane and
Cuffey (2005, p. 120) argued that P. undulata also differs
from P. rossimontina by: (1) a broader and more angular
shape; (2) a straight posterolaterally slanting lateral mar-
gin; (3) a sharp angle where the lateral and posterolateral
plate margins join; (4) a proportionally longer posterome-
dial margin; (5) a less well-pronounced curvature to the
F IG . 8 . Graphs of measurement ratios for CeNu (A–C) and AVL (D–H) of Phyllolepis specimens from Belgium. A, length vs width.
B, anterior length vs length. C, posterior margin length vs anterior margin length. D, length vs width. E, anterior length vs length. F,
lateral margin length vs medial margin length. G, medial margin length vs width. H, posterior margin length vs lateral margin length.
Abbreviations: L.ant, anterior length; L.mg.ant/lat/med/post, anterior/lateral/medial/posterior margin length. Colour online.
OL IVE ET AL . : P LACODERMS FROM STRUD 995
Page 16
A
F
G H
B
C D
E
996 PALAEONTOLOGY , VOLUME 58
Page 17
posterolateral margin; and (6) a more posterolaterally
slanting anterior margin. Our morphometrical analysis
shows that criteria 1 and 4 may be invalidated by contin-
uous variation (Fig. 9D, H). The other criteria, even if
not quantified by our morphometrical analysis, are not
sufficient to discriminate between species, as Figure 7
shows anterior ventrolateral plates displaying a strong
variability concerning the criteria 2, 3, 5 and 6.
Lane and Cuffey (2005, p. 120) also compared the cen-
tronuchal plate of P. rossimontina with that of P. konincki
sensu Stensi€o (1939). According to them, that of P. rossi-
montina differs from that of P. konincki in lacking the
ornamentation of concentric ridges. In agreement with
Long and Daeschler (2013), it should not be considered
as a diagnostic character. Moreover, in P. undulata sensu
this article, both ornamentation types are found in the
same specimens, that is concentric ridges and meshlike
ornamentation. Regarding the anterior ventrolateral plate,
Lane and Cuffey (2005, p. 123) argued that P. konincki
sensu Stensi€o (1939) differs from P. rossimontina by: (1)
a much more pronounced curve in the posterolateral
plate margin; (2) a better defined angle where the ante-
rior and lateral plates margins are joined; and (3) a pos-
terolaterally slanting lateral plate margin. As for the
comparison with the anterior ventrolateral plate of P. un-
dulata sensu Stensi€o (1939), Figure 7 illustrates anterior
ventrolateral plates of the Belgian species with a strong
variability for all these criteria, which are therefore insuf-
ficient to discriminate P. rossimontina from P. konincki
sensu Stensi€o (1939).
Therefore, and considering the results of our morpho-
metrical analysis, we consider that P. rossimontina is actu-
ally the same as P. undulata sensu this article. According
to the nomenclatural rule of priority, only the taxon
P. undulata is retained, whereas P. rossimontina becomes
a junior synonym of P. undulata.
Other Phyllolepis species, a preliminary analysis
The figured and measurable anterior ventrolateral and
centronuchal plates of other species (all Famennian in
age) of the genus Phyllolepis have also been integrated in
this study (Fig. 10): P. tolli Vasiliauskas, 1963 (Latvia);
P. concentrica Agassiz, 1844 (Scotland); P. woodwardi
Stensi€o, 1939, and P. orvini Heintz, 1930 (East Green-
land); P. nielseni Stensi€o, 1939 (East Greenland); and
P. thomsoni Long and Daeschler, 2013 (Pennsylvania,
USA). Several remarks can be made:
1. Concerning the centronuchal measurements, Phyllolepis
woodwardi approaches (Fig. 10B–C) or even overlaps
(Fig. 10A) the cloud of points represented by the speci-
mens of P. undulata (which includes all specimens of
P. rossimontina), but diverges more from that group
when considering the AVL measurements (Fig. 10D–E,G–H). However, measurements for the only available
AVL have been taken on an enlarged photography
without scale information (i.e. Stensi€o 1939, text-fig. 2)
and the point representing the AVL of P. woodwardi in
all graphs should therefore appear closer to the cloud
of P. undulata. In any case, the sample of P. woodwardi
is too limited to draw a conclusion.
2. Concerning P. nielseni, measurements made on the
AVL overlap those of P. nielseni and P. undulata
(Fig. 10D–H). Measurements made on the cen-
tronuchal plates show P. nielseni representative
slightly above the point cloud of P. undulata
(Fig. 10A–C). However, as for P. woodwardi (see
above), the sample is too limited.
3. On all graphs (Fig. 10A–H), all specimens of P. orvini
group together, except a centronuchal plate (Stensi€o
1936, pl. 2, fig. 1; Fig. 10A–B), which could either be
misidentified, or have a wrong scale on the publica-
tion, or even be a juvenile form. The points repre-
senting P. orvini clearly fall outside the range of
variation of P. undulata.
4. The only specimen (an AVL) known of P. tolli falls
within the cloud point of P. orvini. This would argue
for a synonymy between both taxa (P. tolli being
junior synonym of P. orvini).
5. Recently, Long and Daeschler (2013) described Phyl-
lolepis thomsoni from the late Famennian (expansa
zone) of Pennsylvania, USA. Only one centronuchal
plate is usable for the morphological analysis. Like for
P. nielseni, it approaches the P. undulata cloud, but
the size of the sample does not permit any further
conclusions.
6. P. concentrica is represented by a single AVL, well iso-
lated of all other species on the graphs.
An interesting feature appears when the best-fit line is
drawn for all species. The R2 ranges between 0.8964 and
0.996 (Fig. 10) and suggests that both size and shape vari-
ations are continuous for all species of Phyllolepis. It con-
firms that the genus Phyllolepis is strongly supported.
F IG . 9 . Graphs of measurement ratios for CeNu (A–C) and AVL (D–H) of Phyllolepis specimens from Belgium and Pennsylvania. A,
length vs width. B, anterior length vs length. C, posterior margin length vs anterior margin length. D, length vs width. E, anterior
length vs length. F, lateral margin length vs medial margin length. G, medial margin length vs width. H, posterior margin length vs
lateral margin length. Abbreviations: L.ant, anterior length; L.mg.ant/lat/med/post, anterior/lateral/medial/posterior margin length;
Pennsyl., Pennsylvania. Colour online.
OL IVE ET AL . : P LACODERMS FROM STRUD 997
Page 18
A
F
G H
B
C D
E
998 PALAEONTOLOGY , VOLUME 58
Page 19
However, recognizing different species within this genus
is more difficult. Three morphogroups may be identified:
one includes P. orvini and P. tolli, another one includes
P. concentric, and a third one includes P. undulata,
P. nielseni, P. woodwardi and P. thomsoni.
Faunal and floral affinities between Belgium and
Pennsylvania
The Red Hill site is a road-cut exposure in the upper part
of the Duncannon Member of the Catskill Formation in
Clinton County, Pennsylvania, USA. The fossil remains at
Red Hill come from a broad but narrow sequence of
strata (Cressler et al. 2010a), which are attributed to the
VCo palynozone (sensu Streel et al. 1987) within the
Famennian stage.
Daeschler and Cressler (2011, table 1) established a
list of the Red Hill fauna and flora contents. The verte-
brate fauna comprises phyllolepid and groenlandaspid
placoderms, gyracanthid acanthodians, primitive chon-
drichthyans, palaeoniscid actinopterygians, dipnoan,
rhizodontid, megalichthyid and tristichopterid sarcoptery-
gians, as well as early tetrapods. The antiarch placoderm
Bothriolepis and the porolepiform Holoptychius have not
been reported from the main fossiliferous zone at Red Hill
although they are very well represented in many other sites
of the Catskill Formation (Daeschler et al. 2003). Their
absence in Red Hill may be due to palaeoenvironmental
conditions (Cressler et al. 2010a). The palaeoenvironment
of the Red Hill site consists of a meandering stream sys-
tem, with frequent avulsion events, in which floodplain
ponds were present (Sevon 1985; Berg et al. 1993; Woo-
drow et al. 1995; Daeschler and Cressler, 1999, 2011;
Cressler et al. 2010a). The presence of terrestrial arthro-
pods (Shear 2000) and of well-preserved terrestrial plants
(Cressler 1999; Cressler et al. 2010b) reinforces the non-
marine interpretation of the Red Hill site (Daeschler et al.
2003).
Kenrick and Fairon-Demaret (1991) and then Cressler
et al. (2010b) and Prestianni and Gerrienne (2010)
noticed that the localities of the Belgian Evieux Forma-
tion, including Strud, display plant assemblages that are
very similar to those from the North American sites,
especially Red Hill. Both sites were located along the
southern margin of the Euramerica and date from the
same VCo palynozone (sensu Streel et al. 1987).
An equivalent observation may be made for the
vertebrate faunas. Phyllolepididae, Groenlandaspididae,
Acanthodii, Actinopterygii, Dipnoi, Megalichthyididae,
Tristichopterididae and Ichthyostegalia are found in both
localities. Similarities even reach the generic level, when
considering Turrisaspis (known in these two localities
only) and even the species level for Phyllolepis undulata
(see above). The main differences between the two locali-
ties are the chondrichthyans that are present in Red Hill,
but lacking in Strud, and the presence of antiarchs and
porolepiforms in Strud, which are absent in Red Hill
(although present in other localities of the Catskill Forma-
tion). An accurate quantification of the fauna similarities
between the Strud and Red Hill localities is difficult to
evaluate so far because taxonomic identification is still in
progress for some taxa at Strud and Red Hill. However,
even if not quantified, the similarities of fauna and flora
between these two late Famennian localities seem strong.
Exchanges between these sites over a single landmass (i.e.
Euramerica) would have been facilitated for plants with
aeolian spore dispersion. The close similarity of the verte-
brate faunas from these sites may be explained by faunal
exchanges. Several of the vertebrate taxa found in both
localities are considered to be non-marine; for example,
groenlandaspidids are at least considered cautiously as
indicators of non-marine rocks in the Famennian
(Daeschler et al. 2003). Thus, either there was a non-
marine link between Pennsylvania and Belgium during the
late Famennian, or the shared taxa were more tolerant of
marine/brackish conditions than previously thought
(Schultze and Cloutier 1996; Anderson et al. 1999).
Long and Daeschler (2013, p. 41) discussed the co-
occurrences of Phyllolepis species with early tetrapods,
bothriolepid antiarchs and groenlandaspid arthrodires in
several localities (notably P. tolli from Latvia, P. nielseni
from East Greenland, ‘P. rossimontina’ and P. thomsoni
from Pennsylvania). The placoderm fauna from Strud
represents a similar association. They also argued that the
presence of Phyllolepis in the Northern Hemisphere dur-
ing the late Famennian ‘may serve as an indicator of
favorable paleoecological conditions in the search for
early tetrapods’ (Long and Daeschler 2013, p. 41), a
remark already made by Young (2005c) and confirmed
herein by the association of Phyllolepis and an ichthyoste-
gid-like tetrapod in the late Famennian of Strud (Cl�ement
et al. 2004). However and to date, the absence of tetrapod
remains among several thousand specimens of vertebrate
F IG . 10 . Graphs of measurement ratios for CeNu (A–C) and AVL (D–H) of all Phyllolepis species. A, length vs width. B, anterior
length vs length. C, posterior margin length vs anterior margin length. D, length vs width. E, anterior length vs length. F, lateral mar-
gin length vs medial margin length. G, medial margin length vs width. H, posterior margin length vs lateral margin length. Abbrevia-
tions: Green., Greenland; L.ant, anterior length; L.mg.ant/lat/med/post, anterior/lateral/medial/posterior margin length; Pennsyl.,
Pennsylvania; Scot., Scotland. Colour online.
OL IVE ET AL . : P LACODERMS FROM STRUD 999
Page 20
fossils in the Kl�unas fossil site, Latvia (T�ervete Fm, mid-
dle Famennian), where Phyllolepis tolli, Bothriolepis ornata
and B. jani have been found together, is a counterexam-
ple of the presence of Phyllolepis as indicator of tetrapods
(Vasilkova et al. 2012).
Acknowledgements. We gratefully acknowledge the Belgian Fed-
eral Science Policy Office for the research financial support
(Doctoral Fellow to SO) and the Jessup Fund (ANSP, Philadel-
phia, USA) for the funding provided to SO for the ANSP collec-
tion visit. This research received support, under the number
DK-TAF-2253, from the SYNTHESYS Project (http://www.syn-
thesys.info/), which is financed by European Community
Research Infrastructure Action under the FP7 Integrating Activi-
ties Programme. The authors thank J. Denayer (University of
Li�ege, Belgium; the University of Queensland, Australia) for his
remarks concerning the geological framework as well as E. Bour-
don (University of Copenhagen) for her help. The ANSP team
and G. Cuny (University of Copenhagen, now University of
Lyon) are thanked for their warm welcome in their respective
institutions. A. Folie, D. Nolf and A. Dreze (IRSNB, Brussels, Bel-
gium) granted access to the collection of the IRSNB and provided
technical support. E. Poty (ULg) granted access to the collection
of the ULg and M.-C. Van Dyck (UCL, Louvain-la-Neuve,
Belgium) to the collection of the University of Louvain-la-
Neuve. We thank the Gesves local council staff for providing
technical support and excavation permission in Strud. We are
indebted to the successive Strud field teams, who helped us since
2004. The authors thank the editor and the reviewers for their
fruitful remarks. The photographs were provided by W. Miseur
(IRSNB) and Philippe Loubry (CNRS/MNHN/UPMC-Paris 6,
Paris, France). This article is a contribution to the ANR TERRES
2010-BLAN-607-03 project.
DATA ARCHIVING STATEMENT
Data for this study are available in the Dryad Digital Repository:
http://dx.doi.org/10.5061/dryad.qb550
Editor. Zerina Johanson
REFERENCES
AGASSIZ, L. 1844. Monographie des poissons fossiles du Vieux
Gr�es Rouge ou syst�eme d�evonien (Old Red Sandstone) des Isles
Britanniques et de Russie. Jent and Gassman, Neuchatel, 171
pp.
ANDERSON, M. E., LONG, J. A., EVANS, F. J., AL-
MOND, J. E., THERON, J. N. and BENDER, P. A. 1999.
Biogeographic affinities of Middle and Late Devonian fishes of
South Africa. Records of the Western Australian Museum,
suppl. 57, 157–168.BERG, T. M., McINERNEY, M. K., WAY, J. H. and
MACLACHLAN, D. B. 1993. Stratigraphic correlation chart
of Pennsylvania. Commonwealth of Pennsylvania, Bureau of
Topographic and Geologic Survey, Pennsylvania, General
Geology Report 75.
CL�EMENT, G. and BOISVERT, C. A. 2006. Lohest’s true
and false ‘Devonian amphibians’: evidence for the rhyn-
chodipterid lungfish Soederberghia in the Famennian of
Belgium. Journal of Vertebrate Paleontology, 26, 276–283.-and PRESTIANNI , C. 2009. Fauna, flora, and paleoenvi-
ronment of the Upper Devonian tetrapod-bearing locality of
Strud, Belgium. 105–115. In GODEFROIT, P. and LAM-
BERT, O. (eds). Tribute to Charles Darwin and Bernissart
iguanodons: New perspectives on vertebrate evolution and Early
Cretaceous ecosystems. Institut royal des Sciences naturelles de
Belgique, Brussels, 145 pp.
-AHLBERG, P. E., BLIECK, A., BLOM, H., CLACK,
J. A., POTY, E., THOREZ, J. and JANVIER, P. 2004.
Palaeogeography: Devonian tetrapod from Western Europe.
Nature, 427, 412–413.CRESSLER, W. L. III 1999. Site-analysis and floristics of the
Late Devonian Red Hill locality, Pennsylvania, an Ar-
chaeopteris-dominated plant community and early tetrapod
site. Published PhD thesis, University of Pennsylvania,
Philadelphia, 156 pp.
- DAESCHLER, E. B., SLINGERLAND, R. and
PETERSON, D. A. 2010a. Terrestrialization in the Late
Devonian: a palaeoecological overview of the Red Hill site,
Pennsylvania, USA. 111–128. In VECOLI , M., CL�EMENT,
G. and MEYER-BERTHAUD, B. (eds). The terrestrializa-
tion process: modelling complex interactions at the biosphere–geosphere interface. Geological Society, London, Special Publi-
cation, 339, 192 pp.
-PRESTIANNI , C. and LEPAGE, B. A. 2010b. Late
Devonian spermatophyte diversity and paleoecology at Red
Hill, north-central Pennsylvania, U.S.A. International Journal
of Coal Geology, 83, 91–102.DAESCHLER, E. B. and CRESSLER, W. L. 1999. Sampling
an early continental ecosystem: Late Devonian bone beds at
Red Hill, Clinton County, PA. Journal of Vertebrate Paleontol-
ogy, 19, 41A.
--2011. Late Devonian paleontology and paleoenviron-
ments at Red Hill and other fossil sites in the Catskill Forma-
tion of north-central Pennsylvania. 1–16. In RUFFOLO,
R. M. and CIAMPAGLIO, C. N. (eds). From the shield to
the sea: geological trips from the 2011 Joint Meeting of the GSA
Northeastern and North-Central Sections. Geological Society of
America, Field Guide, 20, 116 pp.
- FRUMES, A. C. and MULLISON, C. F. 2003.
Groenlandaspidid placoderm fishes from the Late Devonian
of North America. Records of the Australian Museum, 55,
45–60.DENAYER, J., PRESTIANNI , C., GUERIAU, P., OLIVE,
S. and CL�EMENT, G. in press. Stratigraphy and deposi-
tional environments of the Late Famennian (Late Devonian)
of Southern Belgium and characterization of the Strud locality.
Geological Magazine.
DUPRET, V. 2003. Etude anatomique des genres Kuj-
danowiaspis Stensi€o 1942 et Erikaspis nov. gen. (Placodermi
Arthrodira “Actinolepida”) du D�evonien inf�erieur de Podolie
(Ukraine). Nouvelle proposition de la phylog�enie des Arthro-
1000 PALAEONTOLOGY , VOLUME 58
Page 21
dira. Unpublished PhD thesis, Mus�eum national d’Histoire
naturelle, Paris, 403 pp.
-2010. Revision of the genus Kujdanowiaspis Stensi€o, 1942
(Placodermi, Arthrodira, “Actinolepida”) from the Lower
Devonian of Podolia (Ukraine). Geodiversitas, 32, 5–63.-and ZHU, M. 2008. The earliest phyllolepid (Placodermi,
Arthrodira), Gavinaspis convergens, from the late Lochkovian
(Lower Devonian) of Yunnan (South China). Geological Maga-
zine, 145, 257–278.GARROUSTE, R., CL�EMENT, G., NEL, P., ENGEL, M.
S., GRANDCOLAS, P., D’ HAESE , C., LAGEBRO, L.,
DENAYER, J., GUERIAU, P., LAFAITE , P., OLIVE,
S., PRESTIANNI , C. and NEL, A. 2012. A complete insect
from the Late Devonian. Nature, 488, 82–85.------------
-2013. Garrouste et al. reply. Nature 494, E4–E5.GOUJET, D. 1984. Les poissons placodermes du Spitsberg –Arthrodires Dolichothoraci de la Formation de Wood Bay
(D�evonien inf�erieur). Editions du CNRS, Paris, 284 pp.
GROSS , W. 1965. €Uber die placodermen-gattungen As-
terolepis und Tiaraspis aus dem Devon Belgiens und einen
fraglichen Tiaraspis-rest aus dem Devon Spitzbergens. Bul-
letin de l’Institut royal des Sciences naturelles de Belgique, 61,
1–19.GUERIAU, P., CHARBONNIER, S. and CL�EMENT, G.
2014a. First decapod crustaceans in a Late Devonian continen-
tal ecosystem. Palaeontology, 57, 1203–1213.---2014b. Angustidontid crustaceans from the Late
Devonian of Strud (Namur Province Belgium): insights into
the origin of Decapoda. Neues Jahrbuch f€ur Geologie and
Pal€aontologie, Abhandlungen, 273/3, 327–337.HEINTZ, A. 1930. Oberdevonische fischreste aus €Ost-
Gr€onlands. Skrifeter om Svalbard og Ishavet, 30, 31–46.H €ORNSCHEMEYER, T., HAUG, J., B�ETHOUX, O.,
BEUTEL, R. G., CHARBONNIER, S., HEGNA, T. A.,
KOCH, M., RUST, J., WEDMANN, S., BRADLER, S.
and WILLMANN, R. 2013. Is Strudiella a Devonian insect?
Nature, 494, E3–E4.JANVIER, P. and CL�EMENT, G. 2005. A new groenlandas-
pidid arthrodire (Vertebrata: Placodermi) from the Famennian
of Belgium. Geologica Belgica, 8, 51–67.KENRICK, P. and FAIRON-DEMARET, M. 1991.
Archaeopteris roemeriana (G€oppert) sensu Stockmans, 1948
from the Upper Famennian of Belgium: anatomy and leaf
polymorphism. Bulletin de l’Institut royal des Sciences naturelles
de Belgique, Sciences de la Terre, 61, 179–195.LAGEBRO, L., GUERIAU, P., HEGNA, T. A., RABET,
N., BUTLER, A. and BUDD, G. E. 2015. The oldest noto-
stracan (Upper Devonian Strud locality, Belgium). Palaeontol-
ogy, 58 (3), 497–509.LANE, J. A. and CUFFEY, R. 2005. Phyllolepis rossimontina
sp. nov. (Placodermi) from the Uppermost Devonian at Red
Hill. North-Central Pennsylvania. Revista Brasileria de Paleon-
tologia, 8, 117–126.LERICHE, M. 1931. Les poissons famenniens de la Belgique –Les faci�es du Famennien dans la r�egion gallo-belge – Les
relations entre les formations marines et les formations
continentales du D�evonien sup�erieur sur la bordure
m�eridionale du continent Nord-Atlantique. M�emoires de la
Classe des Sciences de l’Acad�emie Royale de Belgique, 4, 1–72.LOHEST, M. 1888a. D�ecouverte du plus ancien amphibien
connu et de quelques fossiles remarquables dans le Famennien
sup�erieur de Modave. Annales de la Soci�et�e G�eologique de
Belgique, 15, 129–137.-1888b. Recherche sur les poissons des terrains pal�eozo€ıques
de Belgique. Poissons des Psammites du Condroz, Famennian
sup�erieur. Annales de la Soci�et�e G�eologique de Belgique, 15,
112–203.LONG, J. A. 1984. New phyllolepids from the Victoria and the
relationships of the group. Proceedings of the Linnean Society
of New South Wales, 107, 263–308.-and DAESCHLER, E. B. 2013. First articulated phyl-
lolepid placoderm from North America, with comments on
phyllolepid systematics. Proceedings of the Academy of Natural
Sciences of Philadelphia, 162, 33–46.-ANDERSON, M. E., GESS , R. and HILLER, N. 1997.
New placoderm fishes from the Late Devonian of South
Africa. Journal of Vertebrate Paleontology, 17, 253–268.MILES , R. S. 1973. An actinolepid arthrodire from the Lower
Devonian Peel Sound Formation, Prince of Wales Island.
Palaeontographica, 143, 109–118.OBRUCHEV, D. V. 1964. Agnatha and fish. In ORLOV, Yu.
(ed.). Fundamentals of paleontology, Vol. 11. Moscow, 522 pp.
OLIVE, S. in press. Devonian antiarch placoderms from Bel-
gium revisited. Acta Paleontologica Polonica. doi:10.4202/app.
00015.2013
-PRESTIANNI , C. and DUPRET, V. 2015a. A new spe-
cies of Groenlandaspis Heintz, 1932 (Placodermi, Arthrodira),
from the Famennian (Late Devonian) of Belgium. Journal of
Vertebrate Paleontology, 35 (4). doi:10.1080/02724634.2014.
935389
-CL�EMENT, G., DAESCHLER, E. B. and DUPRET,
V. 2015b. Data from: Characterization of the placoderm
(Gnathostomata) assemblage from the tetrapod-bearing local-
ity of Strud (Belgium, upper Famennian). Dryad Digital
Repository. doi:10.5061/dryad.qb550
PRESTIANNI , C. and GERRIENNE, P. 2010. Early seed
plant radiation: an ecological hypothesis. Geological Society,
London, Special Publication, 339, 71–80.-STREEL, M., THOREZ, J. and GERRIENNE, P.
2007. Strud: old quarry, new discoveries. Preliminary report.
43–47. In STEEMANS, P. and JAVAUX, E. (eds). Recent
advances in palynology. Carnet de G�eologie, Brest, Memoir
2007/01, 73 pp.
RITCHIE , A. 1984. A new placoderm, Placolepis gen. nov.
(Phyllolepidae), from the Late Devonian of New South Wales,
Australia. Proceedings of the Linnean Society of New South
Wales, 107, 321–353.-2005. Cowralepis, a new genus of phyllolepid fish (Pisces,
Placodermi) from the Late Middle Devonian of New South
Wales, Australia. Proceedings of the Linnean Society of New
South Wales, 126, 215–259.SCHULTZE, H-P. and CLOUTIER, R. 1996. Comparison
of the Escuminac Formation ichthyofauna with other late
Givetian/early Frasnian ichthyofaunas. 348–368. In
SCHULTZE, H-P. and CLOUTIER, R. (eds). Devonian
OLIVE ET AL . : P LACODERMS FROM STRUD 1001
Page 22
Fishes and Plants of Miguasha, Quebec, Canada. Verlag Dr.
Friedrich Pfeil, M€unchen, 374 pp.
SEVON, W. D. 1985. Nonmarine facies of the Middle and Late
Devonian Catskill coastal alluvial plain. 79–90. In WOO-
DROW, D. L. and SEVON, W. D. (eds). The Catskill Delta.
Geological Society of America, Special Paper, 201, 246 pp.
SHEAR, W. A. 2000. Gigantocharinus szatmaryi, a new trigono-
tarbid arachnid from the Late Devonian of North America
(Chelicerata, Arachnida, Trigonotarbida). Journal of Paleontol-
ogy, 74, 25–31.STENSI €O, E. 1934. On the Placodermi of the Upper Devonian
of East Greenland. I. Phyllolepida and Arthrodira. Meddelelser
om Grønland, 97, 1–58.-1936. On the Placodermi of the Upper Devonian of East
Greenland. Supplement to Part I. Meddelelser om Grønland,
97, 1–53.-1939. On the Placodermi of the Upper Devonian of East
Greenland. Second supplement to Part I. Meddelelser om
Grønland, 97, 1–33.STREEL , M., HIGGS, K. T., LOBOZIAK, S., RIEGEL,
W. and STEEMANS, P. 1987. Spore stratigraphy and corre-
lation with faunas and floras in the type marine Devonian of
the Ardenne-Rhenish Regions. Review of Palaeobotany and
Palynology, 50, 211–229.THOREZ, J., DREESEN, R. and STREEL , M. 2006. Fa-
mennian. Geologica Belgica, 9, 27–45.VASILIAUSKAS, V. 1963. Phyllolepis tolli sp. nov. and some
problems concerning the stratigraphy of the Famennian
deposits in the Baltic area. 407–429. In GRIGELIS , A. A. and
KARATAJUTE-TALIMAA, V. N. (eds). Geology of Lithua-
nia. Lietuvos TSR Mokslu akademija, Vilnius. [in Russian]
VASILKOVA, J., LUKSEVICS , E., STINKULIS , G. and
ZUPINS, I. 2012. Taphonomy of the vertebrate bone beds
from the Kl�unas fossil site, Upper Devonian T�ervete Forma-
tion of Latvia. Estonian Journal of Earth Sciences, 61, 105–119.
WOODROW, D. L., ROBINSON, R. A. J., PRAVE, A. R.,
TRAVERSE, A., DAESCHLER, E. B., ROWE, N. D. and
DELANEY, N. A. 1995. Stratigraphic, sedimentological and
temporal framework of Red Hill 41 (Upper Devonian Catskill
Formation) near Hyner, Clinton County, Pennsylvania: site of
the oldest amphibian known from North America. p. 16. In
WAY, J. (ed.). Field Trip Guide, 60th Annual Field Conference
of Pennsylvania Geologists. Lock Haven University, Lock Haven.
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. British Museum of Natural
History, London, 567 pp.
-1915. Preliminary report on the fossil fishes from Dura
Den. Reports of the British Association for Advancement of
Science, Australia, 84, 122–123.YOUNG, G. C. 2005a. An articulated phyllolepid fish (Placo-
dermi) from the Devonian of central Australia: implications
for non-marine connections with the Old Red Sandstone con-
tinent. Geological Magazine, 142, 173–186.-2005b. A new phyllolepid placoderm occurrence (Devonian
fish) from the Dulcie Sandstone, Georgina Basin, central Aus-
tralia. Proceedings of the Linnean Society of New South Wales,
126, 203–213.-2005c. New phyllolepids (placoderm fishes) from the
Middle-Late Devonian of Southeastern Australia. Journal of
Vertebrate Paleontology, 25, 261–273.ZIEGLER, P. A. 1990. Geological atlas of Western and Central
Europe, Second edition. Shell Internationale Petroleum
Maatschappij B.V. & Geological Society of London, Avon,
239 pp.
1002 PALAEONTOLOGY , VOLUME 58