Phylogenetic relationships of Upper Jurassic (Middle Volgian) plesiosaurians (Reptilia: Sauropterygia) from the Agardhfjellet Formation of central Spitsbergen, Norway

Abstract

Plesiosaurian relationships have been subject to considerable debate at both higher and lower levels within the date. Recent fieldwork in the Uppermost Jurassic of the Agardhfjellet Formation on central Spitsbergen has uncovered numerous ichthyosaurians and plesiosaurian remains, including three new plesiosauroids, one new pliosauroid referrable to Pliosaurus, and a previously described, re-classified taxon. The phylogenetic relationships of these five taxa were investigated based on data sets previously constructed for global plesiosaurian relationships. Two specimens from the British Kimmeridge Clay Formation and a Russian Volgian (uppermost Jurassic) taxon referrable to Pliosaurus were added to the data matrix to improve the phylogenetic resolution of this genus. The results yielded a tree topology closely conforming to the traditional plesiosauroid and pliosauroid dichotomy, nesting Leptocleidia within the latter. Pliosaurus forms a monophyletic clade containing all currently recognised species. The three new long-necked taxa form a monophyletic sister group to the Cretaceous Elasmosauridae. These results should, however, be considered preliminary pending the discovery of more complete cranial material and adult specimens.

Full text

277
Introduction
In the past decade, our knowledge of plesiosaurian
(Sauro pterygia: Plesiosauria) relationships has improved
with the application of cladistic methodologies. How-
ever, a broad consensus regarding plesiosaur relation-
ships remains elusive. In particular, the first global phylo-
genetic analyses have called into question long-standing
assumptions of higher-level relationships within the
clade, including the validity and definition of the two
traditional clades, Plesiosauroidea and Pliosauroidea
(O’Keefe, 2001; Druckenmiller & Russell, 2008; Ketchum
& Benson, 2010). Furthermore, there has been consider-
able debate regarding species-level relationships within
certain clades, such as Elasmosauridae (Sato, 2002;
O’Keefe, 2004; Groβman, 2007; Druckenmiller & Russell,
2008; Ketchum & Benson, 2010), Pliosauridae (O’Keefe,
2004; Druckenmiller & Russell, 2008), and Rhomaleo-
sauridae (O’Keefe, 2004; Smith & Dyke, 2008; Ketchum
& Benson, 2010). An updated matrix devised to clarify
aspects of pliosaurid relationships was published recently
by Ketchum & Benson (2011).
In 2004, a new locality for Upper Jurassic marine reptiles
was discovered on the high-arctic Norwegian archipelago
of Svalbard (Hurum et al., this volume). The finds were
recovered from the dark-grey to black mudstone of the
Slottsmøya Member of the Agardhfjellet Formation ,
which is Volgian (Tithonian) to Valanginian in age
(Nagy & Basov, 1998; Collignon & Hammer, this volume ;
Hammer et al., 2011). Numerous skeletons of plesiosauri-
ans and ichthyosaurs have been found and excavated over
the course of seven field seasons. The process of collect-
ing and describing these new taxa is ongoing, but work
to date allows the description or redescription of several
plesiosaurian taxa that are unique to this area, including
Colymbosaurus svalbardensis (Knutsen et al., this volume
(d)), Spitrasaurus wensaasi (Knutsen et al., this volume
(c)), S. larseni (Knutsen et al., this volume (c)), Djupe-
dalia engeri (Knutsen et al. this volume (a)) and Pliosau-
rus funkei (Knutsen et al. this volume (b)).
The objectives of the present study are to conduct a pre-
liminary cladistic analysis in order to place the new Sval-
bard OTUs into a preliminary phylogenetic context and
to test its effect on the topology and stability of broad-
scale relationships within Plesiosauria. The new material
from Svalbard is significant in that it contributes valu-
able new information on both long- and short-necked
morphotypes to an ever-growing data set on plesiosaur
Druckenmiller, P.S. & Knutsen, E.M. Phylogenetic relationships of Upper Jurassic (Middle Volgian) plesiosaurians (Reptilia: Sauropterygia) from
the Agardhfjellet Formation of central Spitsbergen, Norway. Norwegian Journal of Geology, Vol 92, pp. 277-284. Trondheim 2012, ISSN 029-196X.
Plesiosaurian relationships have been subject to considerable debate at both higher and lower levels within the clade. Recent fieldwork in the Upper-
most Jurassic of the Agardhfjellet Formation on central Spitsbergen has uncovered numerous ichthyosaurians and plesiosaurian remains, including
three new plesiosauroids, one new pliosauroid referrable to Pliosaurus, and a previously described, re-classified taxon. The phylogenetic relati-
onships of these five taxa were investigated based on data sets previously constructed for global plesiosaurian relationships. Two specimens from
the British Kimmeridge Clay Formation and a Russian Volgian (uppermost Jurassic) taxon referrable to Pliosaurus were added to the data matrix
to improve the phylogenetic resolution of this genus. The results yielded a tree topology closely conforming to the traditional plesiosauroid and
pliosauroid dichotomy, nesting Leptocleidia within the latter. Pliosaurus forms a monophyletic clade containing all currently recognised species.
The three new long-necked taxa form a monophyletic sister group to the Cretaceous Elasmosauridae. These results should, however, be considered
preliminary pending the discovery of more complete cranial material and adult specimens.
Patrick S. Druckenmiller, University of Alaska Museum, 907 Yukon Drive, Fairbanks, AK 99775, United States. E-mail: psdruckenmiller@alaska.edu.
Espen M. Knutsen, Natural History Museum (Geology), University of Oslo, Pb. 1172 Blindern, 0318 Oslo, Norway. Present address, Haoma Mining
NL, Bamboo Creek Mine, P.O. Box 2791, South Headland, 6722 WA, Australia. , E-mail: espenkn80@gmail.com.
Patrick S. Druckenmiller & Espen M. Knutsen
NORWEGIAN JOURNAL OF GEOLOGY Phylogenetic relationships of Upper Jurassic (Middle Volgian) plesiosaurians
Phylogenetic relationships of Upper Jurassic
(Middle Volgian) plesiosaurians (Reptilia: Sauropterygia)
from the Agardhfjellet Formation of central
Spitsbergen, Norway

278
where xi is the selected datum, and n is the number of
desired states (not including state “0”). xs is rounded up
to the nearest integer to give the state. Following Thiele
(1993), these meristic characters were each given a
weight of one, and treated as ordered. Discrete charac-
ters were treated as unordered and given a weight equal
to number of states available for the meristic characters
(Thiele, 1993).
Analysis of the data matrix
The data matrix was analysed using PAUP* version
4.0b10 (Swofford, 2002). Simosaurus gailardoti, Augusta-
saurus hagdorni and Cymatosaurus were selected as out-
group taxa, following Druckenmiller & Russell (2008),
and Ketchum & Benson (2010). Tree searches were per-
formed using PAUPRat (Sikes & Lewis, 2001) which
implements the Parsimony Ratchet strategy described
by Nixon (1999). Similar to Ketchum & Benson (2010),
twenty independent batch files were constructed using
PAUPRat. Each batch file performed 200 heuristic search
iterations, with the random seed number being gener-
ated by the system clock. ‘WTMODE’ was set to multi-
plicative, with the percentage of characters being per-
turbed to 15%. The best trees from each batch were com-
bined into one tree-file using PAUP. The twenty inde-
pendent PAUPRat batch files yielded 313 most parsimo-
nious trees (MPTs) with a tree length of 18804. These
trees were used as starting points for a heuristic search
in PAUP implementing the ‘TBR’ branch swapping algo-
rithm, number of repetitions to 200, and ‘maxtrees’ to
increase automatically. This resulted in 149679 MPTs
with a tree length of 16484. These MPTs were used to
calculate the strict consensus tree (Figure 1A).
Following the method implemented by Ketchum & Ben-
son (2010), and described by Wilkinson (2003), ‘wild-
card’ taxa were identified based on comparisons of the
strict Adams consensus trees (Figure 1) constructed
from the 14976 MPTs and found in the analysis of the
un-pruned data matrix. This resulted in the removal of
Colymbosaurus svalbardensis and Pliosaurus sp. (BRSMG
Cc332) from the data matrix and then reanalysing the
matrix as described above. This strict reduced consensus
tree is presented in Figure 2.
Bootstrap support was calculated using 2000 replicates
and the heuristic search options described above. Max-
Trees were set to 100 to reduce computation time. Fol-
lowing Ketchum & Benson (2010), calculation of Bremer
support was not performed due to the large size of the
data sets (Müller, 2004, 2005), and because it is computa-
tionally expensive. Character reconstructions were per-
formed using both ACCTRAN and DELTRAN character
optimisation, and unambiguous character definitions of
nodes were found by comparison of the two.
morphology. New data on British Pliosaurus are also
incorporated into the analysis, and the first species-level
analysis of the genus is presented. Finally, these data help
to fill a temporal gap in our knowledge of plesiosaurs
from the latest Jurassic (Volgian-Tithonian).
Institutional Abbreviations:
BRSMG Bristol City Museum and Art Gallery, Bristol,
UK.
CAMSM Sedgwick Museum, Cambridge, UK.
NHMUK Natural History Museum, London, UK.
PIN Palaeontological Institute, Russian Academy of
Sciences, Moscow, Russia.
PMO Natural History Museum, University of Oslo,
Norway (Palaeontological collections).
Material and methods
Construction of data matrix The data matrix (supplemen-
tary material) was constructed in Mesquite 2.71 and is a
modified version of Benson et al. (2011), which in turn is
largely based on the matrix of Ketchum & Benson (2010).
Five Svalbard taxa were added to the matrix (Supple ment
1). Pliosaurus funkei Knutsen et al. (this volume (b)) is
based on a composite scoring of complementary mate-
rial from both the holotype (PMO 214.135) and the
referred specimen (PMO 214.136). Three new species of
long-necked plesiosaurians (described in this volume)
were all scored from the holotype specimens: Djupeda-
lia engeri Knutsen et al. (this volume (a); PMO 216.839);
Spitrasaurus wensaasi Knutsen et al. (this volume (c);
PMO 219.718); and S. larseni Knutsen et al. (this vol-
ume (c); SVB 1450). The scores for Colymbo saurus sval-
bardensis (Persson, 1962) Knutsen et al. (this volume (d))
are based on the holotype specimen (PMO A 27745) col-
lected in 1931.
As a result of a broader review of the Upper Jurassic plio-
saurids (Knutsen, this volume), additional data on Plio-
saurus species was also added into the analysis (Supple-
ment 1). Pliosaurus macromerus (NHMUK 39362, sensu
Knutsen, this volume), and P. rossicus (PIN 304/1, holo-
type) were newly incorporated into the matrix, and the
scores for P. brachydeirus were modified from Benson
et al. (2011; Supple ment 2). In the original analysis by
Ketchum & Benson (2010), BRSMG Cc332 was scored
to represent P. brachyspondylus; however, recent work
indicates that this specimen cannot with certainty be
referred to any existing species and is listed in this matrix
as Pliosaurus sp. (Knutsen, this volume). In this analysis
P. brachyspondylus (sensu Knutsen, this volume) is repre-
sented by CAMSM J.35991 (Supple ment 2).
Meristic characters were divided into twenty-six states by
using the formula provided by Thiele (1993):
xs=[(xi–xmin)⁄(xmax–xmin )]n
P. S. Druckenmiller & E. M. Knutsen NORWEGIAN JOURNAL OF GEOLOGY

279
NORWEGIAN JOURNAL OF GEOLOGY Phylogenetic relationships of Upper Jurassic (Middle Volgian) plesiosaurians
Spitrasaurus wensaasi (80%)
Hauffiosaurus longirostris (77%)
Plesiosaurus macrocephalus (73%)
Kimmerosaurus langhami (78%)
Hauffiosaurus tomistomimus (39%)
Kaiwhekea katiki (73%)
Edgarosaurus muddi (37%)
Marmornectes candrewi (64%)
Polycotylus latipinnis (84%)
OUMNH J.02247 (64%)
Thalassomedon haningtoni (61%)
Leptocleidus superstes (72%)
Brancasaurus_brancai (48%)
Thalassiodracon hawkinsi (39%)
Rhomaleosaurus victor (56%)
Occitanosaurus tournemirensis (41%)
Terminonatator pontiexensis (56%)
Maresaurus coccai (67%)
Macroplata tenuiceps (54%)
Tricleidus seeleyi (41%)
Cymatosaurus (56%)
NHMUK R2439 (84%)
Pliosaurus andrewsi (60%)
Leptocleidus capensis (48%)
Cryptoclidus eurymerus (31%)
Muraenosaurus leedsii (17%)
Eurycleidus arcuatus (70%)
Kronosaurus queenslandicus (68%)
Manemergus anguirostris (76%)
Callawayasaurus colombiensis (39%)
Plesiopleurodon wellesi (73%)
OUMNH J.28585 (85%)
Eopolycotylus rankini (71%)
Plesiosaurus dolichodeirus (23%)
Elasmosaurus platyurus (87%)
Djupedalia engeri (77%)
Hydrotherosaurus alexandrae (63%)
Trinacromerum bentonianum (54%)
Palmulasaurus quadratus (83%)
Nichollssaura borealis (36%)
Pliosaurus brachyspondylus (85%)
QMF 18041 (43%)
Dolichorhynchops herschelensis (41%)
Augustasaurus hagdorni (46%)
NHMUK 49202 (45%)
Colymbosaurus svalbardensis (96%)
Colymbosaurus svalbardensis (96%)
Umoonasaurus demoscyllus (64%)
Seeleyosaurus guilelmiimperatoris (25%)
Rhomaleosaurus megacephalus (46%)
Styxosaurus snowii (50%)
Pliosaurus macromerus (84%)
Eromangasaurus australis (83%)
Liopleurodon ferox (18%)
Hydrorion brachypterygius (49%)
Archaeonectrus rostratus (69%)
Peloneustes philarchus (7%)
Strateosaurus taylori (55%)
FHSMVP321 (69%)
Aristonectes parvidens (73%)
Simosaurus gaillardoti (23%)
Attenborosaurus conybeari (75%)
Rhomaleosaurus zetlandicus (56%)
Microcleidus homalospondylus (38%)
Libonectes morgani (41%)
Pliosaurus brachydeirus (84%)
Pliosaurus funkei (composite; 79%)
Hauffiosaurus zanoni (58%)
Brachauchenius lucasi (62%)
Dolichorhynchops osborni (15%)
Pliosaurus sp. (BRSMG Cc332; 53%)
Pliosaurus sp. (BRSMG Cc332; 53%)
Spitrasaurus larseni (72%)
Pliosaurus rossicus (86%)
Thililua longicollis (75%)
Simolestes vorax (27%)
Attenborosaurus conybeari
Pliosaurus andrewsi
Plesiosaurus macrocephalus
Polycotylus latipinnis
Palmulasaurus quadratus
Eopolycotylus rankini
Elasmosaurus platyurus
Kaiwhekea katiki
Plesiopleurodon wellesi
Seeleyosaurus guilelmiimperatoris
OUMNH J.02247
Maresaurus coccai
Trinacromerum bentonianum
Dolichorhynchops osborni
Leptocleidus superstes
Plesiosaurus dolichodeirus
Liopleurodon ferox
Marmornectes candrewi
Microcleidus homalospondylus
Callawayasaurus colombiensis
Muraenosaurus leedsii
Brachauchenius lucasi
Hydrorion brachypterygius
Hauffiosaurus tomistomimus
Umoonasaurus demoscyllus
Djupedalia engeri
Thalassomedon haningtoni
Spitrasaurus wensaasi
Dolichorhynchops herschelensis
Occitanosaurus tournemirensis
Simosaurus hgaillardoti
Cymatosaurus
Colymbosaurus svalbardensis
Colymbosaurus svalbardensis
Thalassiodracon hawkinsi
Pliosaurus brachyspondylus
NHMUK R2439
Eromangasaurus australis
Hauffiosaurus longirostris
Pliosaurus sp. (BRSMG Cc332)
Pliosaurus sp. (BRSMG Cc332)
Terminonatator pontiexensis
Rhomaleosaurus zetlandicus
FHSMVP321
Augustasaurus hagdorni
Tricleidus seeleyi
Kimmerosaurus langhami
NHMUK 49202
Kronosaurus queenslandicus
Macroplata tenuiceps
Edgarosaurus muddi
Pliosaurus macromerus
Simolestes vorax
QMF 18041
Hauffiosaurus zanoni
Styxosaurus snowii
Eurycleidus arcuatus
Pliosaurus funkei (composite)
Peloneustes philarchus
OUMNH J.28585
Aristonectes parvidens
Libonectes morgani
Manemergus anguirostris
Pliosaurus brachydeirus
Pliosaurus rossicus
Rhomaleosaurus megacephalus
Nichollssaura borealis
Hydrotherosaurus alexandrae
Brancasaurus brancai
Strateosaurus taylori
Spitrasaurus larseni
Thililua longicollis
Cryptoclidus eurymerus
Rhomaleosaurus victor
Leptocleidus capensis
Archaeonectrus rostratus
A B
Figure 1. Strict (A; tree length=16886, CI=0.32, HI=0.68, RI=0.61) and Adams (B) consensus trees of 14976 most parsimonious trees resulting
from the analysis of the un-pruned data matrix. ‘Wildcard’ taxa are marked in red. Percentages give the proportion of missing data for each
operational taxonomical unit (OTU).

280 P. S. Druckenmiller & E. M. Knutsen NORWEGIAN JOURNAL OF GEOLOGY
46%
56%
23%
75%
45%
73%
54%
69%
77%
39%
58%
46%
56%
56%
67%
84%
84%
79%
85%
84%
86%
27%
64%
64%
07%
18%
69%
62%
68%
55%
39%
23%
25%
85%
49%
41%
38%
80%
72%
77%
87%
41%
56%
50%
63%
39%
61%
17%
41%
31%
78%
36%
48%
64%
48%
72%
37%
54%
73%
75%
73%
73%
83%
41%
71%
15%
76%
83%
70%
60%
43%
84%
Brancasaurus brancai
Plesiosaurus macrocephalus
Rhomaleosaurus zetlandicus
Polycotylus latipinnis
Nichollssaura borealis
Dolichorhynchops osborni
Maresaurus coccai
Pliosaurus brachydeirus
Thililua longicollis
NHMUK 49202
Styxosaurus snowii
Cymatosaurus
Manemergus anguirostris
Spitrasaurus larseni
NHMUK R2439
Attenborosaurus conybeari
Pliosaurus rossicus
Brachauchenius lucasi
Hydrorion brachypterygius
Occitanosaurus tournemirensis
Simosaurus gaillardoti
Cryptoclidus eurymerus
Hauffiosaurus tomistomimus
Kronosaurus queenslandicus
Leptocleidus superstes
Liopleurodon ferox
Kimmerosaurus langhami
Seeleyosaurus guilelmiimperatoris
Eurycleidus arcuatus
Pliosaurus andrewsi
OUMNH J.02247
Leptocleidus capensis
Strateosaurus taylori
OUMNH J.28585
Macroplata tenuiceps
Djupedalia engeri
Dolichorhynchops herschelensis
Spitrasaurus wensaasi
Eopolycotylus rankini
Pliosaurus brachyspondylus
Muraenosaurus leedsii
Rhomaleosaurus megacephalus
Trinacromerum bentonianum
Hydrotherosaurus alexandrae
Pliosaurus macromerus
Kaiwhekea katiki
Hauffiosaurus zanoni
Libonectes morgani
Thalassiodracon hawkinsi
Plesiosaurus dolichodeirus
Rhomaleosaurus victor
FHSMVP321
Aristonectes parvidens
Terminonatator pontiexensis
Hauffiosaurus longirostris
Tricleidus seeleyi
QMF 18041
Microcleidus homalospondylus
Augustasaurus hagdorni
Plesiopleurodon wellesi
Palmulasaurus quadratus
Peloneustes philarchus
Pliosaurus funkei (composite)
Archaeonectrus rostratus
Elasmosaurus platyurus
Simolestes vorax
Marmornectes candrewi
Umoonasaurus demoscyllus
Thalassomedon haningtoni
Edgarosaurus muddi
Eromangasaurus australis
Callawayasaurus colombiensis
1
2 - Neoplesiosauria
1 - Plesiosauria
3
2
3- Plesiosauroidea
4
4- Pliosauroidea
9
9- Pliosauridae
11
11- Pliosaurus
10
10- Leptocleidida
12
12- Lepticleididae
13
13- Polycotylidae
8
6
5
7
8- Elasmosauridae
99
51
51
66
66
66
90
Figure 2. Strict consensus tree of 3718 most parsimonious trees (tree length=16582, CI=0.32, HI=0.68, RI=0.61). Circled numbers label nodes,
numbers below lines are bootstrap values, percentages give the proportion of missing data for each operational taxonomical unit (OTU). The
Svalbard taxa are in bold.

281
NORWEGIAN JOURNAL OF GEOLOGY Phylogenetic relationships of Upper Jurassic (Middle Volgian) plesiosaurians
analysis. Benson et al. (2012) did not recover a sister
group relationship between Pliosauridae and Rhomaleo-
sauridae. The addition of new OTUs and modified scores
for some taxa resulted in several topological differences
in Pliosauridae compared to that of Ketchum & Benson
(2010). In the current analysis Marmornectes candrewi is
recovered (Figure 2) as the most basal pliosaurid, and not
NHMUK R2439, which nests with Pliosaurus. The other
significant difference is in the sister group relationships
of Brachaucheninae (Brachauchenius+Kronosaurus)
FHSM VP321); al though it is consistently the most
derived clade within Pliosauridae, it was recovered as the
sister taxon to Pliosaurus (P. brachydeirus+P. brachyspon-
dylus) in Ketchum & Benson (2010) but is the sister
taxon of Liopleurodon in this study. Brachaucheninae
also differs here by its inclusion of ‘Pliosaurus’ andrewsi.
As discussed elsewhere (Knutsen, 2012), ‘P.’ andrewsi is
not referrable to Pliosaurus (it lacks teeth with a subtri-
hedral cross- section, among other features) and awaits
formal taxonomic revision.
The present study is the first species-level analysis of
all valid Pliosaurus taxa including P. brachydeirus, P.
brachyspondylus, P. rossicus, P. macromerus and the new
Svalbard taxon, P. funkei (Knutsen et al., 2012a). Plio-
saurus is recovered as a monophyletic, but un resolved
clade, along with NHMUK R2439. Pliosaurus (exluding
NHMUK R2439) is unambiguously supported by four
synapomorphies including the absence of a notochordal
pit in the occipital condyle (81:0), teeth with a subtrihe-
dral cross-section (109:2), dorsal neural spines that are
less or equal in height to the centrum (137:0), and the
radius is proximodistally shorter than its anteroposte-
rior width (162:4). The inclusion of NHMUK R2439 in
this clade is likely erroneous. This Oxford Clay specimen
(Callovian) is missing data for characters 81, 109 and
137, and is the only OTU in the clade scored for charac-
ter 162 (scored 4) except for P. funkei, which is scored 2.
The breaking apart of a monophyletic Pliosaurus by
inclusion of BRSMG Cc332 may indicate paraphyletic
relationships within the genus, or be a result of lacking
completeness of character scorings for other members of
the genus.
The new Svalbard taxon Pliosaurus funkei clearly nests
with other Pliosaurus taxa in this analysis, but based on
the strict pruned tree alone, the exact nature of relation-
ships among the different species of the genus requires
further study. A putative relationship between P. funkei
and P. macromerus and P. rossicus is indicated in the
strict reduced and Adams trees. The similarity between
these two taxa and P. funkei was also noted in the com-
parative analysis of cranial material in the description
of P. funkei (Knutsen et al., 2012). Resolving species-
level relationships within Pliosaurus remains problem-
atic; the scorings for all of the specimens in this ana-
lysis are highly incomplete (ranging from 53-86%), and
one of the more cranially complete specimens, BRSMG
Results and Discussion
The pruning of two ‘wildcard’ taxa (Colymbosaurus
svalbard ensis and Pliosaurus sp. (BRSMG Cc332))
reduced the number of MPTs from 14976 to 3718, and
reduced tree length by 304 fewer steps (Figure 2). The
strict reduced tree fully resolves the three new Svalbard
plesiosauroid taxa into a single clade and recovers Plio-
saurus funkei within a monophyletic, but unresolved
Pliosaurus (except ‘P’. andrewsi).
Similar to the analysis of Ketchum & Benson (2010),
the monophyly of Plesiosauria (node 1) is well sup-
ported, with a bootstrap value of 90. Two plesiosaurians,
Attenboro saurus and NHMUK 49202, are successive sis-
ter taxa to Neoplesiosauria (node 2). Two major clades
are recovered in Neoplesiosauria; Plesiosauroidea (node
3) and Pliosauroidea (node 4). However, the composi-
tion of these two clades most closely corresponds to the
traditional definition of Plesiosauroidea and Pliosauroi-
dea (Welles, 1943; Brown, 1981; Druckenmiller & Rus-
sell, 2008; Smith & Dyke, 2008), in marked contrast to
the analysis of O’Keefe (2001) and Ketchum & Benson
(2010). Specifically, this difference is due to the position
of Leptocleidida (Polycotylidae+Leptocleididae), which
is recovered here within Pliosauroidea, but is nested
within Plesiosauroidea in Ketchum & Benson (2010).
Leptocleidia was similarly recovered in Pliosauro idea
in Benson et al. (2011), despite minor alterations to the
matrix of Ketchum & Benson (2010). Given the low
bootstrap support (less than 50 percent) for Pliosauroi-
dea and Plesiosauroidea in the present analysis, and the
high degree of plasticity in the systematic position of
Leptocleidia (O’Keefe, 2001; Druckenmiller & Russell,
2008; Ketchum & Benson, 2010), the status and composi-
tion of these major clades must remain uncertain.
Pliosauroidea (node 4) is supported unambiguously by
ten synapomorphies (3:G-H, 60:0, 63:1, 91:1, 95:C, 99:1,
112:A, 118:C, 140:1, 146:1). A paraphyletic assemblage of
Early to Middle Jurassic pliosauroids is recovered basal to
Leptocleidia (Node 10) and its sister group Pliosauridae
(Node 9). Leptocleidia (node 10) is unambiguously sup-
ported by nine synapomorphies (54:1, 67:1, 68:1, 78:1,
83:0, 125:2, 153:B-C, 167:1, 174:1) and is composed of
Polycotylidae and Leptocleididae (Druckenmiller & Rus-
sell, 2008; Ketchum & Benson, 2010). A certain degree
of incongruity in the topology and resolution in Lepto-
cleidia exists between this analysis and that of Ketchum
& Benson, (2010), but are not discussed further here.
Pliosauridae is supported by nine unambiguous synapo-
morphies: (95:H, 97:1, 98:0, 102:0, 107:0, 123:0, 124:1,
153:6, 168:2). State 98:0 (surangular is mediolaterally
thick and broadly rounded in dorsal view) is unique to
Pliosauridae. Four synapomorphies supported Pliosau-
ridae in Ketchum & Benson (2010), of which only one
state, 123:0 (ventral surface of cervical centra is flat or
only slightly convex) was shared in common with this

282 P. S. Druckenmiller & E. M. Knutsen NORWEGIAN JOURNAL OF GEOLOGY
wide (162:4-5). Because of the small amount of cranial
material, state 75:0 was only scored for Djupedalia and
109:1 for Spitrasaurus larseni. Thus, the shared presence
of a greatly elongate neck and short but wide epipodials
appears to have had a large influence on this grouping.
However, both of these postcranial characters exhibit
high degrees of homoplasy (ci = 0.34) and bootstrap sup-
port for this grouping was less than 50 percent.
Node 5, including four British plesiosauroid taxa, is sup-
ported by nine unambiguous synapomorphies: presence
of bone surface ornamentation around the orbit margin
(34:1); the presence of an anterior interpterygoid vacu-
ity (59:1); the parasphenoid is wide and blunt posteri-
orly in ventral view between the posterior interpterygoid
(68:2); basisphenoid contributes to the articulation sur-
face of the basioccipital tuberosities for contact with the
pterygoids (71:0); 4-5 teeth in the mandibular symphysis
(112:9); atlas centrum participates in the rim of the rim
of the atlantal cup (114:1); axis rib articulates partly with
the atlantal centrum (117:1); ventral plates of the scapu-
lae meet along the midline (145:1); and the posterolat-
eral margin of the coracoid forms an angled cornu that
extends lateral to the glenoid fossa (151:1). Most of the
synapomorphies of node 5 are unknown in the Svalbard
taxa. The first four of these are cranial features that are
unknown in any of the Svalbard taxa, in only one taxon
(Spitrasaurus larseni) is the number of symphyseal teeth
known, and in only Djupedalia engeri is the contribution
of the atlantal centrum to the atlantal cup known. Like-
wise, the early ontogenetic state of the Svalbard mate-
rial makes it impossible to score for features of the pec-
toral girdle. In the light of these observations, it is appar-
ent that missing data plays an influential role in the cur-
rent topology. For these reasons, we view these results as
preliminary pending discovery of new specimens from
later ontogenetic stages and with key cranial material. The
future inclusion of the British Late Jurassic taxon Colym-
bosaurus trochanterius (which has been shown to exhibit
similarities to C. svalbardensis; Knutsen et al., 2012 (d)) in
the data matrix might also aid in understanding the affini-
ties of the Svalbard taxa as well as other Jurassic plesio-
sauroids.
Node 7 forms an unnamed clade consisting of the three
Svalbard plesiosauroid taxa Spitrasaurus wensaasi, S.
larseni and Djupedalia engeri. The other currently recog-
nised Svalbard plesiosauroid, Colymbosaurus svalbard-
ensis, is known only from a small amount of material
from the pelvic girdle and hind limb and was removed
as a wildcard taxon in this reduced analysis (Figure 2).
Spitrasaurus and Djupedalia are supported unambigu-
ously by two characters: dorsal neural spines are less or
equal in height to the dorsal centrum (137:0), and the
radius is relatively proximodistally even shorter than for
node 6 (162:2). In the present analysis S. larseni and D.
engeri are recovered as sister taxa, to the exclusion of S.
wensaasi, suggesting that Spitrasaurus is paraphyletic.
However, this result must be viewed with caution. S.
Cc332, was identified as a wildcard taxon in the pres-
ent analysis. Furthermore, none of the characters recog-
nised as diagnostic for the different Pliosaurus species
by Knutsen (2012) were considered here, as it would
require extensive revisions of the data matrix beyond
the scope of this study. Ongoing work to describe new
material , taxonomic clarifications of existing specimens,
and a reanalysis of phylogenetically important characters
within the genus are needed.
Plesiosauroidea is unambiguously supported by three
synapomorphies (82:1, 83:0, 115:1). At the base of the
clade is a paraphyletic assemblage of several Early Juras-
sic plesiosauroids, including Plesiosaurus dolichodei-
rus. Three derived plesiosauroid clades are recovered.
The newly described taxa from Svalbard, Spitrasau-
rus wensaasi , S. larseni and Djupedalia engeri (node 7;
Knutsen et al., 2012a and 2012c), form a monophyletic
group that is the sister taxon to Elasmosauridae (node
8). This unnamed clade, in turn, is the sister taxon to
four taxa (Muraeonosaurus (Tricleidus (Cryptoclidus
+ Kimmero saurus))) (node 5). Support for all of these
nodes is low (bootstrap values less than 50 percent).
In general terms, the topology among derived plesio-
sauroids in congruent with stratigraphic distribution;
Node 5 includes Middle to Late Jurassic forms, and the
relatively derived Svalbard and elasmosaurid taxa are
Late Jurassic and Cretaceous, respectively. However, the
clade Cryptoclidia, sensu Ketchum & Benson (2010) was
not recovered by exclusion of Leptocleidia (see above).
Although a re-evaluation of global relationships among
plesiosaurians is beyond the scope of this study, it is
likely that ongoing morphological work to better under-
stand Middle to Late Jurassic long-necked taxa, includ-
ing putative basal cryptoclidians and the new Svalbard
taxa, will play a pivotal role in resolving the systematic
relationships of Leptocleidia and Cryptoclidia (Ketchum
& Benson, 2010) with either plesiosauroids or pliosau-
roids (Benson et al., 2011).
The definition of Elasmosauridae has long been a con-
tentious issue (O’Keefe, 2001; Druckenmiller & Rus-
sell, 2008; Ketchum & Benson, 2010). In this analysis,
it is supported by a single synapomorphy (presence of a
ventral notch in the articular face of the cervical centra
in anterior view; 122:1), but is otherwise composed of
the same taxa (although in a somewhat different topol-
ogy) as that recovered in Ketchum & Benson (2010).
Putative members of the clade, Muraenosaurus (Druck-
enmiller & Russell, 2008) and Brancasaurus (O’Keefe,
2004), were not identified as elasmosaurids here. The sis-
ter group relationship between Elasmosauridae and the
new Svalbard taxa (node 6) is supported by four unam-
biguous synapomorphies: absence of a distinct notch in
the suspensorium for articulation with the paraoccipi-
tal process (75:0), teeth with oval cross-section (109:1),
a high number of cervical vertebrae (118:P) and a radius
that is proximodistally shorter than anteroposteriorly

283
NORWEGIAN JOURNAL OF GEOLOGY Phylogenetic relationships of Upper Jurassic (Middle Volgian) plesiosaurians
Acknowledgements – We thank R.B.J. Benson for sharing his data
matrix and helpful discussions on plesiosaurian phylogeny; Tommy
Wensås, Stig Larsen, Magne Høyberget, Øyvind Enger and Lena Kris-
tiansen for spending their summer holidays excavating marine repti-
les on Svalbard every summer over the last seven years, and May-Liss
Knudsen Funke for preparing the specimens; H.T. and R. Schassberger
and H.S. and J.K. Druckenmiller; our financial sponsors ExxonMobil,
Fugro, OMW, Spitsbergen Travel, Powershop, Hydro, StatoilHydro,
National Geographic, and Brian Snyder for funding our annual Arctic
fieldwork. Constructive reviews were provided by R.B.J. Benson and R.
Forrest. This work is part of a Ph.D. study funded by the University of
Oslo, Norway.
Supplementary material:
The following supplementary material accompanies the
online version of this paper: http://www.geologi.no/njg/
Supplement 1 - Character scores for added taxa.
Supplement 2 - Modified character scores for Pliosaurus
(in bold).
References
Benson, R.B.J., Ketchum, H.F., Noè, L.F. & Gomez-Pérez, M. 2011:
New information on Hauffiosaurus (Reptilia, Plesiosauria) based
on a new species from the Alum Shale Member (Lower Toarcian:
Lower Jurassic) of Yorkshire, UK. Palaeontology 54, 547-571.
Benson, R.B.J., Evans, M., & Druckenmiller, P.S. 2012: High Diversity,
Low Disparity and Small Body Size in Plesiosaurs (Reptilia, Sau-
ropterygia) from the Triassic–Jurassic Boundary. PLoS ONE 7(3),
e31838.
Brown, D.S. 1981: The English Upper Jurassic Plesiosauroidea (Repti-
lia) and a review of the phylogeny and classification of the Plesio-
sauria. Bulletin of the British Museum (Natural History), Geology
35, 253-347.
Collignon, M. & Hammer, Ø. 2012: Lithostratigraphy and sedimen-
tology of the Slottsmoya Member at Janusfjellet, Spitsbergen: evi-
dence for a condensed section. Norwegian Journal of Geology 92,
89-101.
Druckenmiller, P.S. & Russell, A. P. 2008: A phylogeny of Plesiosauria
(Sauropterygia) and its bearing on the systematic status of Lepto-
cleidus Andrews, 1922. Zootaxa 1863, 1-120.
Großmann, F. 2007: The taxonomy and phylogenetic position of the
Plesiosauroidea from the Lower Jurassic Posidona Shale of south-
west Germany. Palaeontology 50, 545-564.
Hammer, Ø., Nakrem, H.A., Little, C.T.S., Hryniewicz, K., Sandy, M.R.,
Hurum, J.H., Druckenmiller, P.S., Knutsen, E.M. & Høyberget, M.
2011: Hydrocarbon seeps from close to the Jurassic- Cretaceous
boundary, Svalbard. Palaeogeography, Palaeoclimatology, Palaeo-
ecology 306, 15-26.
Hurum, J.H., Nakrem, H.A., Hammer, Ø, Knutsen, E.M., Drucken-
miller, P.S., Hryniewicz, K. & Novis, L.K. This volume. An
Arctic Lagerstätte - the Slottsmøya Member of the Agardhfjellet
Formation (Upper Jurassic - Lower Cretaceous) of Spitsbergen.
Norwegian Journal of Geology 92, 55-64.
Ketchum, H.F. & Benson, R.B.J. 2010: Global interrelationships of
Plesiosauria (Reptilia, Sauropterygia) and the pivotal role of taxon
sampling in determining the outcome of phylogenetic analyses.
Biological Reviews 2009, 1-32.
Ketchum, H.F. & Benson, R.B.J. 2011: A new pliosaurid (Sauroptery-
gia, Plesiosauria) from the Oxford Clay Formation (Middle Juras-
sic, Callovian) of England: evidence for a gracile, longirostrine
grade of Early-Middle Jurassic pliosaurids. Special Papers in Pala-
eontology 86, 109-129.
Knutsen, E.M. 2012: A taxonomic review of the genus Pliosaurus
(Owen, 1841a) Owen, 1841b. Norwegian Journal of Geology 92,
larseni and D. engeri unambiguously cluster on the basis
of the relative length of the cervical centra (120:0), which
has been argued to vary ontogenetically in elasmosaurids
(O’Keefe & Hiller, 2006; Brown, 1981), and one quanti-
tative character: the relative length to width dimensions
of the radius (162:1). This topology is somewhat a prod-
uct of the gap weighting method employed here (Thiele,
1993) in which different scores may be derived from very
slight differences in the actual dimensions of features. For
example, the actual values for character 162 in S. larseni
and S. wensaasi are 0.714 and 0.711, respectively, but they
are assigned two different scores (0 and 1). Additionally,
all three specimens are juvenile specimens and the val-
ues recorded for the radius are only approximate and are
intended to show that this element is wider than long.
Finally, most of the features used to diagnose both genera,
such as the posterior displacement of the cervical neural
spines, the medial contact of the postzygapophyses and
the presence of preaxial accessory ossicles, are not charac-
ters used in this phylogenetic analysis. We feel that subse-
quent analyses that incorporate new characters will more
faithfully resolve the relationships between these taxa.
Conclusions
1) A preliminary phylogenetic analysis was conducted
after incorporating new data on long- and short
necked taxa from the latest Jurassic of Svalbard into
the most recent and largest data matrix available
(Ketchum & Benson, 2010; Benson et al., 2011).
2) After removal of wildcard taxa, a strict consensus tree
recovered a topology conforming closely to the tradi-
tional concept of Plesiosauroidea and Pliosauroidea.
A monophyletic Leptocleidia sensu Ketchum & Ben-
son (2010) was recovered but was not nested within
Plesiosauroidea. However, tree support for this result
is low and the stability of a traditionally-conceived
Pliosauroidea remains uncertain.
3) The first species-level analysis of Pliosaurus is con-
ducted based on a recent re-examination of mate-
rial. All five currently recognised species of Pliosau-
rus form a monophyletic, but unresolved clade along
with NHMUK R2439. However, the strict reduced
consensus tree indicates possible paraphyletic rela-
tionships within the genus, which may result from
high percentages of missing data for most of the taxa.
4) Three new species of Svalbard long-necked taxa form
a monophyletic group that is the sister group to Cre-
taceous Elasmosauridae. However, relationships of
the Svalbard taxa to Middle and Late Jurassic British
taxa is difficult to fully assess due to missing data.
5) Results from the present analysis are best considered
preliminary pending new discoveries of additional
cranial material and adult-aged specimens.

284 P. S. Druckenmiller & E. M. Knutsen NORWEGIAN JOURNAL OF GEOLOGY
259-276.
Knutsen, E.M., Druckenmiller, P.S. & Hurum, J.H. 2012 (a): A new
plesiosaurian taxon (Reptilia – Sauropterygia) from the Agardh-
fjellet Formation (Middle Volgian) of central Spitsbergen, Norway.
Norwegian Journal of Geology 92, 213-234.
Knutsen, E.M., Druckenmiller, P.S. & Hurum, J.H. 2012 (b): A new
species of Pliosaurus (Sauropterygia: Plesiosauria) from the Middle
Volgian of central Spitsbergen, Norway. Norwegian Journal of Geo-
logy 92, 235-258.
Knutsen, E.M., Druckenmiller, P.S. & Hurum, J.H. 2012 (c): Two new
species of long-necked plesiosaurians (Reptilia: Sauropterygia)
from the Late Jurassic (Middle Volgian) Agardhfjellet Formation of
central Spitsbergen. Norwegian Journal of Geology 92, 187-212.
Knutsen, E.M., Druckenmiller, P.S. & Hurum, J.H. 2012 (d):
Re description and taxonomic clarification of “Tricleidus” svalbard-
ensis based on new material from the Agardhfjellet Formation
(Middle Volgian). Norwegian Journal of Geology 92, 175-186.
Müller, K. 2004: PRAP- computation of Bremer support for large data
sets. Molecular Phylogenetics and Evolution 31, 780-782.
Müller, K. 2005: The efficiency of different search strategies in esti-
mating parsimony jackknife, bootstrap, and Bremer support. BMC
Evolutionary Biology 5, 58.
Nagy, J. & Basov, V.A. 1998: Revised foraminiferal taxa and biostrati-
graphy of Bathonian to Ryazanian deposits in Spitsbergen. Micro-
paleontology 44, 217-255.
Nixon, K.C. 1999: The parsimony ratchet, a new method for rapid par-
simony analysis15. Cladistics 15, 407-414.
O’Keefe, F.R. 2001: A cladistic analysis and taxonomic revision of the
Plesiosauria (Reptilia: Sauropterygia). Acta Zoologica Fennica 213,
1-63.
O’Keefe, F.R. 2004: Preliminary description and phylogenetic position
of a new plesiosaur (Reptilia: Sauropterygia) from the Toarcian of
Holzmaden, Germany. Journal of Paleontology 78, 973-988.
O’Keefe, F.R. & Hiller, N. 2006: Morphologic and ontogenetic patterns
in elasmosaur neck length, with comments on the taxonomic uti-
lity of neck length variables. Paludicola 5, 206-229.
Persson, P.O. 1962: Plesiosaurians from Spitsbergen. Norsk Polarinsti-
tutt 1962, 62-68.
Sato, T. 2002: Description of plesiosaurs (Reptilia: Sauropterygia) from
the Bearpaw Formation (Campanian–Maastrichtian) and a phylo-
genetic analysis of the Elasmosauridae. Ph.D., University of Cal-
ga r y.
Sikes, D.S. & Lewis, P.O. 2001: PAUPRat: PAUP implementation of the
parsimony ratchet. Beta software, Version 1. Distributed by the aut-
hors, Department of Ecology and Evolutionary Biology, University
of Connecticut, Storrs.
Smith, A.S. & Dyke, G.J. 2008: The skull of the gigant predatory
plio saur Rhomaleosaurus cramptoni: implications for plesiosaur
phylogeny . Naturwissenschaften 95, 975-980.
Swofford, D. 2002: PAUP*: Phylogenetic analysis using parsimony
(* and other methods). Version 4.0b10. Sinauer Associates, Sunder-
land, Massachusetts.
Thiele, K. 1993: The Holy Grail of the perfect character: the cladistic
treatment of morphometric data. Cladistics 9, 275-304.
Welles, S.P. 1943: Elasmosaurid plesiosaurs, with description of new
material from California and Colorado. University of California,
Memoirs 13, 125-254.
Wilkinson, M. 2003: Missing data and multiple trees: stability and
support . Journal of Vertebrate Paleontology 23, 311-323.