Content uploaded by Stefania Nosotti
Author content
All content in this area was uploaded by Stefania Nosotti on May 18, 2017
Content may be subject to copyright.
Stefania Nosotti
Sezione di Paleontologia, Museo Civico di Storia Naturale di Milano
Tanystropheus longobardicus (Reptilia, Protorosauria):
re-interpretations of the anatomy based on new specimens
from the Middle Triassic of Besano (Lombardy, northern Italy)
Volume XXXV - Fascicolo III
Novembre 2007
Memorie della Società Italiana di Scienze Naturali
e del Museo Civico di Storia Naturale di Milano
© 2007 Società Italiana di Scienze Naturali
Museo Civico di Storia Naturale di Milano
Corso Venezia , 55 - 20121 Milano
In copertina: Tanystropheus longobardicus, MSNM BES SC 1018, skull. Watercolor by Massimo Demma.
Registrato al Tribunale di Milano al n. 6694
Direttore responsabile : Anna Alessandrello
Responsabile di redazione: Stefania Nosotti
Graca editoriale: Michela Mura
Stampa: Litograa Solari, Peschiera Borromeo - Novembre 2007 ISSN 0376-2726
INDEX
INTRODUCTION ................................................. Pag. 4
INSTITUTIONAL ABBREVIATIONS .......... Pag. 6
ANATOMICAL ABBREVIATIONS .............. Pag. 6
MATERIALS .................................................... Pag. 7
METHODS ....................................................... Pag. 8
DESCRIPTION OF THE SPECIMENS ........ Pag. 9
Specimen MSNM BES SC 265 ........................ Pag. 9
Skull ................................................................... Pag. 10
Axial skeleton ..................................................... Pag. 13
Appendicular skeleton ........................................ Pag. 15
Specimen MSNM BES SC 1018 ...................... Pag. 20
Skull ................................................................... Pag. 20
Axial skeleton ..................................................... Pag. 27
Appendicular skeleton ........................................ Pag. 29
Specimen MSNM V 3663 a b ........................... Pag. 38
Specimen MSNM BES 215 .............................. Pag. 38
Specimen MSNM BES 351 .............................. Pag. 41
Specimen MSNM V 3730 ................................. Pag. 42
DISCUSSION ................................................... Pag. 44
Discussion and interpretation of the skull anat-
omy of Tanystropheus: a new reconstruction ... Pag. 44
The axial skeleton: an overall description ...... Pag. 62
Cervical vertebrae and ribs ................................. Pag. 62
Dorsal vertebrae and ribs .................................... Pag. 67
Sacral vertebrae .................................................. Pag. 69
Caudal vertebrae ................................................. Pag. 69
The mobility of the vertebral column in Tanys-
tropheus .............................................................. Pag. 70
Gastralia ............................................................. Pag. 71
Could Tanystropheus walk? ............................. Pag. 71
The pes in Tanystropheus ................................... Pag. 71
The hindlimb: terrestrial versus aquatic locomo-
tion ...................................................................... Pag. 73
Tanystropheus mode of life: on what side of the
shoreline? .......................................................... Pag. 76
IMPLICATIONS FOR PHYLOGENETIC
ANALYSIS ........................................................ Pag. 80
CONCLUSIONS ............................................... Pag. 82
AKNOWLEDGEMENTS ................................ Pag. 84
REFERENCES ................................................. Pag. 85
Stefania Nosotti
Tanystropheus longobardicus (Reptilia, Protorosauria):
re-interpretations of the anatomy based on new specimens
from the Middle Triassic of Besano (Lombardy, northern Italy)
Abstract - After more than one century since the rst report of Tanystropheus longobardicus from the Middle Triassic of Besano,
new specimens from the same outcrops are described. These specimens include two articulated skeletons and an isolated pes, all from
small-sized individuals, and fragmentary remains of larger individuals, i.e. a skull with some associated cervical vertebrae, an isolated
dorsal vertebra and isolated cervical ribs.
This new material conrms the presence of T. longobardicus in the Besano Formation, which previously yielded few evidence.
Moreover, it includes remarkably complete and well preserved specimens which provided the opportunity of a new interpretation of
the anatomy of Tanystropheus, formerly described on the basis of a rich sample from the Swiss Grenzbitumenzone.
The description presented here applies to small-sized individuals of Tanystropheus, traditionally interpreted as the juveniles of
T. longobardicus. However, the point is raised that they might represent the adults of a different species, demonstrating the presence
of two taxa among the Swiss and Italian material referred to T. longobardicus. The holotype, and the single known specimen, of the
small-sized Tanystropheus meridensis from the Meride Limestone is also considered and re-interpreted, leading to the conclusion that
this species is probably a junior synonym of T. longobardicus.
Comparisons of the specimens of Tanystropheus from the Besano Formation with those from the equivalent Grenzbitumenzone
helped to nd the problematic elements of the classical reconstruction.
A new reconstruction of the skull of Tanystropheus is presented based on a three-dimensional clay model, with a re-interpretation
of the pre-orbital region, the skull roof, and the lower jaw. The reconstruction of the temporal region of the skull is shown to be highly
problematical. Finally, the new specimens conrm the presence of a sclerotic ring in Tanystropheus.
In the postcranial skeleton, the more important new information concern the morphology of the appendicular skeleton, which is
remarkably well preserved in the new specimens. In particular, preservation of complete and perfectly articulated manus and pedes for
the rst time yields unequivocal evidence on their morphology.
The anatomy of the appendicular skeleton, in particular that of the hindlimb, is discussed in the context of locomotion mode. An
overall view of previous studies on the mode of life of Tanystropheus is presented and discussed. According to these results, Tany-
stropheus should be regarded as a marine protorosaur, with close terrestrial ancestors, living in shallow waters. The feeding strategy of
Tanystropheus is discussed, on the assumption that it likely was a slow, axial or paraxial swimmer with a stiff neck.
In conclusion, the new information obtained from the specimens described here is evaluated in the context of the recent cladistic
analyses of protorosaurian relationships, highlighting the bearing of systematic anatomical work of original materials on the descrip-
tion and coding of phylogenetically informative characters.
Key words: Tanystropheus, Tanystropheus longobardicus, Middle Triassic, Besano, northern Italy, new specimens, Tanystropheus
meridensis, anatomy, skull, pes, mode of life, systematic.
Riassunto – Tanystropheus longobardicus (Reptilia, Protorosauria): re-interpretazioni dell’anatomia basate su nuovi esemplari
provenienti dal Triassico medio di Besano (Lombardia, Italia settentrionale).
A distanza di oltre un secolo dalla prima segnalazione di Tanystropheus longobardicus negli strati fossiliferi del Triassico medio
di Besano (Varese), vengono qui descritti nuovi esemplari provenienti dai medesimi aforamenti. Si tratta di due scheletri in larga
parte articolati e di un piede isolato ascrivibili ad individui di piccola taglia e di resti frammentari di individui di grandi dimensioni, in
particolare un cranio associato ad alcune vertebre cervicali, una vertebra dorsale isolata e frammenti di coste cervicali.
Questo consistente nucleo di esemplari costituisce una ricca documentazione della presenza di T. longobardicus negli strati fos-
siliferi di Besano. Tale presenza era precedentemente testimoniata solamente dall’olotipo, distrutto nel corso della Seconda Guerra
Mondiale, e da due esemplari frammentari conservati nelle Collezioni del Paläontologischen Institut und Museum der Universität,
Zürich (PIMUZ). La completezza e l’ottimo stato di conservazione dei nuovi materiali consente una re-interpretazione dell’anatomia
scheletrica di Tanystropheus, descritta nel secolo scorso sulla base di svariati esemplari aforati sul versante svizzero del giacimento.
This monograph is dedicated
to Giorgio Teruzzi and Anna Alessandrello,
for their unconditional trust and support
to Massimo Demma,
whose generosity, patience, dedication and extreme professionalism
made an important contribution to this publication
to Dr. Rupert Wild,
who established a milestone in Tanystropheus research
STEFANIA NOSOTTI
4
La nuova descrizione dell’anatomia di T. longobardicus qui presentata è applicabile ad individui di piccola taglia, che non supe-
rano i due metri di lunghezza complessiva. Nell’interpretazione classica, tali esemplari rappresenterebbero gli stadi giovanili della
specie. Viene tuttavia avanzata l’ipotesi che in alternativa essi rappresentino gli adulti di una specie differente e che nel materiale
italiano e svizzero proveniente dalla Formazione di Besano siano in realtà presenti due specie, l’una di piccola taglia, l’altra di taglia
medio-grande. Viene anche preliminarmente re-interpretato l’olotipo, ed unico esemplare conosciuto, di Tanystropheus meridensis,
proveniente dal Calcare di Meride, concludendo che esso è probabilmente sinonimo di T. longobardicus.
Il dettagliato confronto dei nuovi esemplari provenienti dalla Formazione di Besano con quelli dell’equivalente Grenzbitumenzone
conservati nelle Collezioni del PIMUZ ha permesso di ottenere un’informazione più completa e di individuare gli aspetti problematici
della ricostruzione classica.
Viene presentata una nuova ricostruzione del cranio di Tanystropheus ottenuta tramite la realizzazione di un modello tridimen-
sionale. La regione pre-orbitale del cranio, il tetto cranico e la mandibola sono stati re-interpretati con risultati innovativi e ragione-
volmente attendibili, mentre la ricostruzione della regione temporale, rivelatasi particolarmente problematica, resta ipotetica e viene
pertanto proposta come un punto di partenza per ulteriori discussioni. I nuovi esemplari hanno tra l’altro denitivamente confermato
la presenza di un anello sclerotico nell’occhio di Tanystropheus.
Circa lo scheletro postcraniale, gli esemplari di Besano hanno fornito nuove rilevanti informazioni. In particolare, l’eccezionale
conservazione degli arti, unica nell’ambito di tutto il materiale conosciuto, ha permesso per la prima volta di accertare inequivocabil-
mente alcune delle loro caratteristiche anatomiche.
L’anatomia dello scheletro appendicolare, soprattutto del piede, viene interpretata in relazione alla locomozione terrestre ed
acquatica. Questo, ed altri aspetti cruciali per l’interpretazione del modo di vita di Tanystropheus, vengono presentati e discussi, con
un excursus che include anche gli studi più recenti e il confronto con altri vertebrati estinti ed attuali. Nell’interpretazione dell’autore
Tanystropheus era un rettile marino nell’intero arco di vita, tuttavia non altamente specializzato per la vita acquatica in quanto stret-
tamente imparentato con antenati francamente terrestri. Poiché Tanystropheus viene considerato un nuotatore lento dotato di un collo
piuttosto rigido, uno degli aspetti più problematici nell’interpretazione del suo modo di vita rimane quello della strategia utilizzata
per la cattura di prede presumibilmente molto sfuggenti. Contenuti stomacali a pesci e cefalopodi sono stati infatti rinvenuti in alcuni
esemplari nelle Collezioni del PIMUZ.
Inne l’informazione inedita ottenuta con lo studio dei nuovi esemplari viene valutata nel contesto delle più recenti analisi cla-
distiche concernenti le relazioni logenetiche dei protorosauri e si dimostra come in alcuni casi la descrizione e la codicazione dei
caratteri siano basate su illazioni speculative presentate come dati certi, risultando di conseguenza errate e fuorvianti. Viene pertanto
sottolineata l’importanza dello studio anatomico di materiale originale per una descrizione e codicazione accurata dei caratteri lo-
geneticamente informativi.
Parole chiave: Tanystropheus, Tanystropheus longobardicus, Triassico medio, Besano, Italia settentrionale, nuovi esemplari,
Tanystropheus meridensis, anatomia, cranio, pes, modo di vita, sistematica.
The genus Tanystropheus was erected by Hermann
von Meyer (1847-1855) in 1852 (Quenstedt, 1963; Wild,
1976; I.C.Z.N., 1981), to comprise the species T. conspic-
uus, based on isolated bones from the Upper Muschelkalk
of Bayreuth. These elements were interpreted by von
Meyer as strikingly elongate caudal vertebrae of a reptile.
Von Meyer did not explain the ethymology of the name
Tanystropheus, which means “long ribbon”, derived from
the Greek “tany-”= prex, meaning “long”, “stretched”
and “stróphos”= “strap”, “rope”, “ribbon”, referring to
the slender, elongate shape of the bones. According to
von Meyer, the same bones had earlier been interpreted
by Georg zu Münster as belonging to the limb of a new
reptile, for which he had proposed the name Macroscelo-
saurus (with no indication of the relevant specic epithet,
if any). However, von Meyer gave no reference to any
publication by Münster. He introduced the name Tanys-
tropheus maintaining that the name given by Münster was
no longer accepted, with no further explanation. Since
then, the name Tanystropheus has almost exclusively
been used (Wild, 1976: 124), while only few authors
(Broili, 1915: 51; Kuhn, 1934: 118) considered the use of
the name Tanystropheus against the principle of priority,
the senior synonym being Macroscelosaurus.
Oskar Kuhn (1934) referred to “Macroscelosaurus
Münster, 1834” but no mention of Macroscelosaurus has
ever been found either in the paper of 1834, or in any
other known published paper by Münster (Kuhn, 1963:
5; Wild, 1974: 147; 1976: 124). Possibly, as suspected by
Wild R. (pers. comm., 1998), Münster proposed the name
in an unpublished private letter to von Meyer. Seemingly,
von Meyer rst published Macroscelosaurus, as a junior
synonym of Tanystropheus. Following an application by
Wild (1975; 1976), the International Commission on Zoo-
logical Nomenclature (I.C.Z.N., 1981) under its plenary
powers conserved Tanystropheus von Meyer, [1852] as
the valid name for the genus and suppressed Macroscelo-
saurus von Meyer, [1852]. Thus von Meyer was accepted
as the author, and 1952 as the publication date of both
names.
In 1886 Francesco Bassani reported a new reptile from
the Middle Triassic of Besano (Varese Province, Lom-
bardy, Italy), and named it Tribelesodon longobardicus,
in reference to its tricuspid dentition. The specimen,
housed in the Collections of the Museo di Storia Naturale
di Milano, consisted of incomplete remains of a skull
and some associated postcranial elements, preserved on
a part and counterpart plate (Nopcsa v., 1923: 161, pl.
II). Bassani gave a very short preliminary description
of the specimen, which at the time was very poorly pre-
pared (Peyer, 1931: 83), and tentatively concluded that it
represented a ying reptile. As he did not include in his
description any illustration, it is difcult to understand
the reasons of his interpretation. Bassani also referred to a
“second specimen” but Peyer (1931: 94) pointed out that
a second Tribelesodon specimen could not be located in
Milano, and conjectured that the “second specimen” was
in fact a poorly preserved specimen of Macrocnemus.
Bassani never published the monographic description of
Tribelesodon, and for more than fteen years the reptile
from Besano lay almost forgotten in the Collections of the
Museo di Storia Naturale di Milano.
The rst detailed study of the holotype of Tribelesodon
longobardicus was undertaken by Franz von Nopcsa in
INTRODUCTION
TANYSTROPHEUS LONGOBARDICUS: RE-INTERPRETATIONS OF THE ANATOMY BASED ON NEW SPECIMENS FROM BESANO 5
1902 and 1904, and supported by grants from the Wiener
Akademie der Wissenschaften (Nopcsa v., 1923: 161).
The results of the study, and the photographs of the speci-
men were not published until 1923.
In the meantime, Gustav von Arthaber (1921) pub-
lished a short description and a preliminary reconstruc-
tion of the skull of Tribelesodon. Von Nopcsa provided
von Arthaber with a photograph of the skull, that was also
published. At that time Tribelesodon was still consid-
ered a ying reptile. Later, von Arthaber (1922: g. 3 a)
published a composite photograph of the whole speci-
men, obtained by overlapping the negatives of the part
and counterpart taken by von Nopcsa in 1902 (Nopcsa v.,
1923: 161), and the reconstruction of the skull alreadly
published in 1921.
Von Nopcsa (1923) continued to consider Tribelesodon
as a ying reptile (Nopcsa v., 1923: g. 6). His reasoning
based on the interpretation of some problematic, very
elongate bones as segments of a forelimb with a long digit
that he presumed supported a wing membrane. Once this
assumption had been made, it inuenced von Nopcsa’s
interpretation of many of the other preserved elements.
Even so, von Nopcsa did consider the possibility that the
problematic bones could be elongate vertebrae, similar to
the cervical vertebrae of the pterosaur Doratorhyncus, or
to the caudal vertebrae of Tanystrophaeus (sic!) from the
Upper Muschelkalk of Bayreuth. However, he ultimately
rejected these latter possibilities.
In 1927 additional remains of Tanystropheus were
found in the Grenzbitumenzone of Monte San Giorgio
(Cava Tre Fontane, Switzerland). These remains, thought
to belong a single individual, comprised a few bones
preserved on two plates. Later, in 1929, Bernard Pey-
er’s excavations in the Grenzbitumenzone of Monte San
Giorgio (Valporina, Switzerland) yielded the rst almost
complete small skeleton of Tanystropheus (neotype, i.e.
specimen PIMUZ T 2791), together with other fragmen-
tary specimens.
In 1931 Peyer published the description of all this new
material. On the basis of the complete specimen he was
able to identify the problematic elongate bones of Tanys-
tropheus conspicuus from the Upper Muschelkalk of Bay-
reuth as cervical vertebrae. At the same time, after a close
re-examination of the holotype of Tribelesodon, Peyer
showed that Tribelesodon and Tanystropheus were conge-
neric, and he referred all the specimens from Besano and
Monte San Giorgio to the same species: Tanystropheus
longobardicus. Peyer proposed amending the diagnosis
of the genus Tanystropheus - up to then based solely on
the isolated vertebrae of T. conspicuus - and designat-
ing T. longobardicus as the type species. Moreover, he
suggested designating the complete skeleton (specimen
PIMUZ T 2791) and the whole material found in 1929
(sic!) as the neotype of T. longobardicus. Wild (1974:
148, 151) correctly argued that both proposals had to
be rejected, the type species of the genus Tanystropheus
being T. conspicuus by monotypy and the original type
being at the time neither lost or destroyed (irrespective
of the “neotype” designated by Peyer consisted of more
than one specimen, therefore was per se invalid). In the
light of the new discoveries, Peyer gave a re-description
of the holotype of Tribelesodon. This remains the best
account of this specimen, which was later destroyed by a
re caused by the bombing of the Museo di Storia Natu-
rale di Milano during World War II. T. longobardicus was
reconstructed by Peyer as a terrestrial reptile, probably
living close to the water. He interpreted the long neck as
an adaptation towards catching prey from the shore. Peyer
proposed Tanystropheus as the unique representative of
the “Tanysitrachelia” (Peyer, 1931: 89), a new suborder
nested within the Sauropterygia (this group then includ-
ing Trachelosauria, Nothosauria and Plesiosauria).
In 1974 Rupert Wild reviewed all the material of T.
longobardicus, including both specimens published by
Peyer in 1931, as well as undescribed ones resulting from
later excavations conducted by Peyer and Emil Kuhn-
Schnyder. On the basis of a total of 27 specimens - all but
one housed in the Collections of the Paläontologisches
Institut und Museum der Universität in Zürich (PIMUZ) -
Wild (1974) described in detail the skeletal anatomy of T.
longobardicus, with special emphasis on its skull, which
had remained very poorly known after Peyer’s descrip-
tion. In addition, he described the isolated vertebrae of T.
conspicuus from the German Muschelkalk of Bayreuth.
As the holotype of Tribelesodon longobardicus (=Tanys-
tropheus longobardicus) had been destroyed, Wild could
validly designate specimen PIMUZ T 2791 as the neo-
type of T. longobardicus. Wild (1974) gave a detailed
description of the anatomy, ontogenetic development and
the adaptation of T. longobardicus. He considered it to
be terrestrial during juvenile stages but predominantly
aquatic as an adult. Wild (1974) regarded Tanystroph-
eus as a “highly specialized lacertilian”, and included
it in the Tanysitrachelida (1974: 147), as an infraorder
of the “Lacertilia”. Later, Wild (1980a) rejected the
infraorder Tanysitrachelida in favour of the infraorder
Prolacertiformes, the latter representing one of the two
main branches of the Lacertilia (Wild, 1980a: g. 13).
The paper by Wild (1974) remains a benchmark study of
Tanystropheus.
Since then, new species of Tanystropheus have been
described (Jurcsák, 1975; Wild, 1980a; Rieppel, 2001),
and fragmentary specimens have been referred to the genus
(Jurcsák, 1976, 1978, 1982; Vickers-Rich et al., 1999;
Rieppel 2000, 2001; Dalla Vecchia & Avanzini, 2002;
Dalla Vecchia, 2000, 2006; Sulej, 2004; Renesto, 2005),
revealing a western Tethyan distribution and a stratigraphi-
cal occurrence from the lower Anisian to the late Norian.
The validity of the different Tanystropheus species
is summarized below but a full revision of the genus is
beyond the scope of this paper.
Wild (1980b: 204; 1987: 39) considered the possible
synonymy of T. longobardicus and T. conspicuus. In addi-
tion, he questioned the validity of T. biharicus (Jurcsák,
1975), considering this species a probable junior synonym
of T. longobardicus (Wild, 1980a: 12). Fraser & Riep-
pel (2006) recently erected Amotosaurus rotfeldensis for
specimens from the Upper Buntsandstein of the Black
Forest (Germany) previously assigned to “Tanystropheus”
antiquus (Ortlam, 1967; Wild, 1980b; Wild & Oosternik,
1984). These authors also provisionally retained the status
of Tanystropheus antiquus for the specimens from the
Lower Muschelkalk of Poland, originally described by
von Huene (1907-1908) but they questioned its validity.
Renesto (2005: 386) questioned the assignment of T. fossai
(Wild, 1980a) to Tanystropheus, maintaining that the holo-
type of T. fossai lacks unequivocal characters justifying its
referral to that genus. Re-examination of the holotype of
6STEFANIA NOSOTTI
T. meridensis (Wild, 1980a) led Fraser and colleagues
(Fraser et al., 2004; Fraser & Rieppel, 2006: 866) to the
conclusion that this specimen cannot be distinguished
from the smallest specimens of T. longobardicus, and they
include it in this taxon. Detailed comparison of the skull
of T. meridensis with the new specimens described here
conrms the statement that T. meridensis and T. longo-
bardicus should be considered conspecic (Fraser et al.,
in preparation). Renesto (2005) recently described a new
specimen (MCSN 4451) of Tanystropheus from Switzer-
land which, like T. meridensis, was collected in the Lower
Meride Limestone of Ladinian age. However, due to the
incomplete preservation of MCSN 4451, together with the
uncertain status of the single specimen of T. meridensis
(holotype: Wild, 1980a), Renesto preferred to consider the
new specimen as T. cf. longobardicus. A preliminary per-
sonal examination of MCSN 4451, however, did not reveal
striking differences from the new specimens, suggesting
that it might well be referred to T. longobardicus. Interest-
ingly, Fraser et al. (2004; contra Wild, 1974 and Tschanz,
1988) suggested that there might be two separate taxa
represented in the material of T. longobardicus from the
Grenzbitumenzone in the PIMUZ Collections but further
studies are necessary. At present, T. haasi (Rieppel, 2001)
is the only unquestioned species of the genus.
After the publication of the monograph by Wild
(1974), several authors discussed, and sometimes drasti-
cally re-interpreted, both the adaptations of Tanystropheus
(Tschanz, 1985, 1986, 1988; Rieppel, 1989; Taylor, 1989;
Ford, 2002; Renesto, 2005) and its phylogenetic relation-
ships (Benton, 1985; Evans, 1988; Gauthier et al., 1988;
Benton & Allen, 1997; Jalil, 1997; Dilkes, 1998; Peters,
2000a; Rieppel et al., 2003). By contrast, little additional
anatomical investigations have been undertaken, such that
the descriptions given by Wild (1974; 1980a) remain the
most complete ever published.
New, remarkably complete and well preserved speci-
mens of T. longobardicus were collected in the 1990ies
in sediments outcropping near Besano (Varese Province,
Lombardy, Italy). These specimens, described in this
paper, offer the rst opportunity for new interpretations of
the anatomy of Tanystropheus since Wild’s (1974) mono-
graph. The new material also documents the presence of
Tanystropheus in the Middle Triassic of Besano. Very
few and incomplete specimens were previously known
from this locality. The holotype of Tribelesodon longo-
bardicus (see above) is now lost, leaving just two more,
very fragmentary specimens housed in the PIMUZ Col-
lections (specimens PIMUZ T 2782 and PIMUZ T 2788,
see Wild, 1974: tab. 1). Three of the new specimens can
be referred to large-sized individuals, the rst time that
specimens equivalent in size to the larger specimens in
the Grenzbitumenzone have been recorded in the Besano
Formation.
INSTITUTIONAL ABBREVIATIONS
MCSN = Museo Cantonale di Scienze Naturali, Lugano,
Switzerland.
MCSNB = Museo Civico di Scienze Naturali “E. Caf”,
Bergamo, Italy.
MFSN = Museo Friulano di Storia Naturale, Udine, Italy.
MGB = Museu Geologia de Barcelona, Spain.
MSNM = Museo di Storia Naturale, Milano, Italy.
PIMUZ = Paläontologisches Institut und Museum der
Universität, Zürich, Switzerland.
YPM = Yale Peabody Museum, New Haven, Connecticut,
USA.
ANATOMICAL ABBREVIATIONS
a = angular
ac = atlas centrum
af = articular facet
afo = adductor fossa
ai = atlas intercentrum
ana = atlas neural arch
arf = articular fossa
art = articular
as = astragalus
at = anterior tubercle
axi = axis intercentrum
bo = basioccipital
bs + ps = basisphenoid-parasphenoid complex
c = cervical vertebra
ca = calcaneum
caf = capitular articular facet
cbI = ceratobranchial I
cc = cristae cranii
cd = caudal vertebra
ce = centrale
ch = chevron
cl = clavicle
co = coronoid
cor = coracoid
cr = cervical rib
cti = cristae temporales inferiores
d = dorsal vertebra
dc = distal carpal
de = dentary
dr = dorsal rib
dt = distal tarsal
e = epipterygoid
eo = exoccipital
f = frontal
fc = bular condyle
fe = femur
= bula
f-lf = lateral ange of the frontal
fm = foramen magnum
fp = fossa parietalis
g = gastralia
h = humerus
hs = horizontal shelf
i = ilium
7
icl = interclavicle
iob = impression of the olfactory bulb
is = ischium
j = jugal
l = lacrimal
lj-pp = lower jaw post-dentary part
ll = lateral lamina of the surangular
llo = lateral lobe of the main body of the frontal
lvm = latero-ventral margin
m = maxilla
Mca = Meckelian canal
mc = metacarpal
ml = medial lamina of the surangular
mlo = medial lobe of the main body of the frontal
mt = metatarsal
mth = maxillary tooth/teeth
n = nasal
nc = neural canal
ns = neural spine
of = obturator foramen
op = opisthotic
p = parietal
p-stp = supratemporal process of the parietal
pa = preatlas
pap = pre-acetabular process
pf = postfrontal
ph = phalanx
pl = palatine
ple = pleurapophysis
pl-mp = maxillary process of the palatine
plt = palatine tooth/teeth
pm = premaxilla
pmt= premaxillary tooth/teeth
po = postorbital
pra = prearticular
prf = prefrontal
prf-l = lobe of the prefrontal
prf-oa = orbital aspect of the prefrontal
prf-pp = palatine process of the prefrontal
prf-va = postero-ventral aspect of the prefrontal
prz = prezygapophysis
pt = pterygoid
pz = postzygapophysis
pzc = postzygapophyseal canal
pzp = postzygapophyseal process
pu = pubis
q = quadrate
r = radius
ra = radiale
raf = articular facet for the rib
s = sacral vertebra
sa = surangular
sb = sesamoid bone
sc = scapula
scp = sclerotic plate
so = supraoccipital
sp = splenial
sq = squamosal
st = supratemporal
stp = stapes
t = tibia
taf = tubercular articular facet
tc = tibial condyle
tp = transverse process
u = ulna
ul = ulnare
v = vomer
vk = ventral keel
vt = vomerine tooth/teeth
MATERIALS
All the new specimens described in this paper are
housed in the Paleontological Collections of the Museo di
Storia Naturale di Milano.
Five specimens come from the outcrops of the
Besano Formation, or Grenzbitumenzone in the Swiss
geological literature (Röhl et al., 2001). On the basis of
its ammonite fauna, Rieber (1973) referred the Grenzbi-
tumenzone to the uppermost Anisian-lowermost Ladin-
ian. One more specimen was collected in fossiliferous
levels of the Meride Limestone of Ladinian age, which
is separated from the Besano Formation by a dolomitic
band (Furrer, 1995).
1) Specimen MSNM BES SC 265, Tanystropheus
longobardicus.
Complete skeleton. Sasso Caldo (SC) quarry, Besano
(BES), Varese Province, Lombardy, northern Italy.
Besano Formation, Lower Ladinian (curionii Zone),
Middle Triassic.
2) Specimen MSNM BES SC 1018, Tanystropheus
longobardicus.
Incomplete skeleton. Sasso Caldo (SC) quarry, Besano
(BES), Varese Province, Lombardy, northern Italy.
Besano Formation, Lower Ladinian (curionii Zone),
Middle Triassic.
3) Specimen MSNM V 3663 a b, Tanystropheus
longobardicus, gift by Cesare Ferrario.
Incomplete skull and associated cervical vertebrae
on a part and counterpart. Spoil from the Vallone mine,
Besano, Varese Province, Lombardy, northern Italy.
Besano Formation, Anisian-Ladinian boundary, Middle
Triassic.
4) Specimen MSNM BES 215, Tanystropheus cf.
longobardicus.
Isolated dorsal vertebra. Rio Ponticelli quarry, Besano
(BES), Varese Province, Lombardy, northern Italy. Besano
Formation, Anisian-Ladinian boundary, Middle Triassic.
5) Specimen MSNM BES 351, Tanystropheus longo-
bardicus.
Cervical ribs. Spoil from the Vallone mine, Besano
(BES), Varese Province, Lombardy, northern Italy. Besano
Formation, Anisian-Ladinian boundary, Middle Triassic.
6) Specimen MSNM V 3730, Tanystropheus cf. longo-
bardicus, gift by Sergio Rampinelli.
Pes, with partial natural mould and bone remains of the
epipodials. Besano, Varese Province, Lombardy, northern
Italy. Meride Limestone, Ladinian, Middle Triassic.
In addition, the material of Tanystropheus longo-
bardicus in the PIMUZ Collections is re-examined.
TANYSTROPHEUS LONGOBARDICUS: RE-INTERPRETATIONS OF THE ANATOMY BASED ON NEW SPECIMENS FROM BESANO
8STEFANIA NOSOTTI
METHODS
In the description of the materials the terminology of
Wild (1974) is adopted (unless otherwise stated) in the
description of processes, foramina and other anatomical
details of each bone. New terms are introduced in quota-
tion marks.
In describing elements of the pectoral and the pelvic
girdles, the terms medial and lateral are adopted to dif-
ferentiate between the two main surfaces, except for the
interclavicle and the coracoids. The main surfaces of the
latter elements are referred to as dorsal and ventral.
In describing limb elements, to avoid confusion in the
terminology relative to their orientation, the adopted ter-
minology is specied as follows. The normal position of
the humerus and the femur is taken to be directed straight
outwards from the body, with the dorsal surface correlated
with the extensor musculature, and the ventral surface
correlated with the exor musculature. The side facing
anteriorly is referred to as the anterior side, the side facing
posteriorly is referred to as the posterior. As the precise
orientation of the epipodial segment remains elusive, the
normal position of this segment is arbitrarily taken to be
one in which the epipodials lie side by side on a vertical
plane perpendicular to the median sagittal plane. The sur-
faces that, in this position, face anterior or posterior are
referred to as the anterior (extensor musculature) or the
posterior (exor musculature) surface of the bones. The
side closer to the median sagittal plane is referred to as the
medial, the other side as the lateral. Finally, the anatomical
position of the manus and pes is taken to be one in which
they point forward and are in contact with the ground.
The surface in contact with the ground is referred to as the
palmar and the plantar respectively (exor musculature),
the other one is referred to as the dorsal (extensor mus-
culature). As is the case of the epipodials, the manus and
pes have a medial and a lateral side. In describing move-
ments of the ankle and the metatarsals, the terminology of
Brinkman (1980) is adopted. Flexion of the ankle is meant
to be a decrease in the angle at the ankle, and dorsiexion
of the metatarsus is a metatarsal movement associated
with exion of the ankle joint; extension of the ankle is
an increase in the angle at the ankle, and plantarexion of
the metatarsus is a metatarsal movement associated with
extension of the ankle.
Skeletal elements in the different specimens are briey
described under the heading of each specimen. An overall
anatomical description, comparisons and discussion con-
cerning the skull, the axial and appendicular skeleton are
reported in the section “Discussion”.
Specimens of T. longobardicus in the PIMUZ Col-
lections have estimated overall lengths between 53 and
535 cm (Wild, 1974: tab. 1). Wild (1974) considered the
small specimens to represent juveniles and the large ones
adults of the same species, and conjectured that reproduc-
tive maturity corresponded approximately to a size of 2
m total length (on the basis of presence/absence of post-
cloacal bones and of the interpretation of the allometric
growth curves). He considered certain differences in the
dentition, the shape of the bones of the skull, and in the
postcranial skeleton to be due to ontogenetic variation.
Recently, Fraser et al. (2004) raised the possibility that
the differences observed between the smallest and the
larger individuals might instead be related to the pres-
ence of two separate taxa among the Grenzbitumenzone
(=Besano Formation) material. Consequently, I will refer
to the PIMUZ specimens as “small-sized” and “large-
sized” specimens, rather than to juveniles and adults or
even as two distinct species, pending the formal revision
of taxonomy. Following Wild (1974), I maintain a demar-
cation size of approximately 2 m total length between
small and large individuals, considering that individuals
over 2 m lack distinct tricuspid teeth and dentition on the
palatine and pterygoid.
See Tab. 1 for the correspondence of the letters used
by Wild (1974) to designate the specimens in the PIMUZ
Collections with their catalogue number. In the table, the
specimens’ overall length is also reported.
Letter used
by Wild
Number in PIMUZ
Collection
Overall length
(cm)
a T 2791 87
b T 2779 78*
c T 2793 535*
d T 2482 140*
e T 2795 105*
f T 2792 190*
h T 2485 135*
i T 2787 250
k T 2817 295
l T 2480 225*
m T 2818 420
n T 2782 160*
o T 2780 85*
p T 2790 335*
q T 2819 365*
r T 2484 100*
x T 1270 210*
y T 2789 330*
Table 1 – Correspondence of the letters used by Wild
(1974) with the respective catalogue numbers in the
Tanystropheus longobardicus specimens in the PIMUZ
Collections, and overall length of each specimen.
(* = estimated). After Wild, 1974: tab. 1.
9
TANYSTROPHEUS LONGOBARDICUS: RE-INTERPRETATIONS OF THE ANATOMY BASED ON NEW SPECIMENS FROM BESANO
DESCRIPTION OF THE SPECIMENS
Specimen MSNM BES SC 265
(Figs. 1-9, 49 C-D, 54, 56 A, 58, 65, Pls. I-II, Tabs. 2-5, 8)
The specimen consists of an almost complete, and
mostly articulated skeleton (Fig. 1, Pls. I-II), lying on its
left lateral side. The skull is separated from the vertebral
column, and is exposed in ventral view. The vertebral
column is exposed in right lateral view, with some disar-
ticulation and rotation in the sacral and proximal-caudal
regions. Where disarticulated, the ribs are still in close
proximity to the corresponding vertebrae. Gastral ribs
cluster in the trunk region. The elements of the pectoral
and pelvic girdles are disarticulated but still lie in an
arrangement indicating their original position. The fore-
Fig. 1 – Tanystropheus longobardicus, MSNM BES SC 265. Scale bar 50 mm. Preparation: Sergio Rampinelli. Photo: Luciano
Spezia.
10 STEFANIA NOSOTTI
limbs lie side by side ventral to the trunk, the right fore-
limb overlapping the left one. The right hindlimb has been
raised as a consequence of the twisting of the vertebral
column, while the left hindlimb lies ventral to the trunk.
The estimated overall length of the specimen is 110 cm.
Skull
(Figs. 2-3, 49 C-D)
The complete skull is preserved in ventral view. Its
length, measured from the tip of the left lower jaw (=esti-
mated anterior extent of the premaxilla) to the posterior
margin of the parietals, is 5.4 cm. As the parietals are dis-
placed posteriorly, a more reliable estimate of the length
of the skull might be the distance between the tip of the
left lower jaw and the maximum posterior extent of the
left squamosal, which is 5.1 cm.
The cranial elements are heavily crushed and elements
of the neurocranium are randomly displaced posteriorly,
thereby exposing the ventral surface of the parietals. The
lower jaws rotated as a unit, with their anterior ends dis-
placed to the right, and the posterior ends to the left, rela-
tive to the sagittal plane. Almost all the teeth of the upper
jaws are still in situ, while the majority of the mandibular
teeth are missing.
Premaxilla
The right premaxilla and part of the left one are concealed
by the lower jaws. The area of contact between the right
premaxilla and maxilla is not exposed. The left premaxilla
still contacts the left maxilla but the suture between the two
elements cannot be distinguished. Consequently, the number
of premaxillary teeth cannot be ascertained. As in specimen
MSNM BES SC 1018 the premaxilla bears six teeth (p. 21),
I infer that the complete premaxillary dentition is seen on the
left premaxilla of MSNM BES SC 265. The teeth are coni-
cal and have nely striated enamel. The tooth implantation
is subthecodont (sensu Romer, 1956; Edmund, 1979; Wild,
1974; Motani, 1997; Zaher & Rieppel, 1999).
Maxilla
Both maxillae are preserved. The right maxilla is
almost completely exposed in ventro-medial view, cov-
ered by the right lower jaw only very anteriorly. Only part
of the jugal process is missing. The left maxilla is com-
pletely preserved and exposed in medial view but it no
longer contacts the jugal. Almost all the ankylosed maxil-
lary teeth, and/or the replacement ones, are preserved in
situ on both maxillae. There are ten teeth in situ in the
left maxilla and two empty alveoli anterior to them: the
complete maxillary dentition then includes 12 teeth. The
rst unequivocally tricuspid tooth on the left maxilla is
the seventh. The sixth has a distinct posterior cusp but
only a very rudimentary anterior one. The fth only has
at best very rudimentary posterior cusp. Twelve teeth are
also preserved on the right maxilla. In this case the rst
unequivocally tricuspid tooth is the fourth.
The raised lingual margins of the maxillae, dorsal to
the teeth, should frame the choanae. The latter, however,
cannot be clearly identied in the dermal palate.
Vomer
The vomers are entirely concealed by the lower jaws
but in between the latter, some isolated vomerine teeth
are exposed. They are very tiny and conical, with striated
enamel.
Palatine
Both palatines are partially exposed. However,
because of the poor preservation, no details can be dis-
cerned on left palatine. The maxillary process of the right
palatine is clearly visible but it no longer contacts the
maxilla. It shows a notched antero-lateral margin fram-
Fig. 2 – Tanystropheus longobardicus, MSNM BES SC 265, skull. Scale bar 10 mm. Photo: Luciano Spezia.
11
TANYSTROPHEUS LONGOBARDICUS: RE-INTERPRETATIONS OF THE ANATOMY BASED ON NEW SPECIMENS FROM BESANO
Fig. 3 – Tanystropheus longobardicus, MSNM BES SC 265, skull. Drawing and pencil: Fabio Fogliazza.
ing the choana postero-medially, and a weakly concave
postero-lateral margin that framed the suborbital fenestra
antero-medially. A conical tooth preserved on the left side
is tentatively interpreted as a single palatine tooth.
Pterygoid
The right pterygoid exhibits the transverse process
emerging close to the right lower jaw as a stout element,
no longer contacting the ectopterygoid. Remnants of the
quadrate process are possibly preserved further posteri-
orly, lateral to the parietal, and facing the right quadrate.
No tooth-attachment sites can be identied. The left
pterygoid could not be identied.
Ectopterygoid
No ectopterygoids could be identied in MSNM BES
SC 265.
Frontal
Both frontals are partly exposed in ventral view, roof-
ing the orbits with their horizontal lateral anges (pp.
48-49).
12 STEFANIA NOSOTTI
Parietal
Both parietals are broadly exposed in ventral view,
and only partly concealed by the overlapping mandi-
bles. Posteriorly, they are separated along their sutural
contact. The morphological features of the ventral sur-
face of the parietal plate that were described by Wild
(1974: 11, g. 2) can be clearly identied. A medial,
narrow fossa parietalis is framed by the cristae cranii.
They extend laterally and anteriorly into the cristae
temporales inferiores. These in turn form the thickened
lateral margin of the parietals. Posterior supratemporal
processes project laterally at an angle of approximately
145°-150°. They bear an elongate facet for the reception
of the squamosals and perhaps the supratemporals (see
below).
Supratemporal
Paired small bony bars which overlap the parietals
posteriorly might represent the supratemporals. However,
because they are displaced, their relationships with the
neighbouring bones are unclear and their identication
remains equivocal.
Prefrontal
On the right and left side of the skull there are two
protuberances which are here interpreted as the palatine
processes of the prefrontals.
Squamosal
Both squamosals are preserved. The right squa-
mosal is partly concealed by overlapping bones. The
left squamosal is fully exposed. Anteriorly, it forms a
sharply pointed postorbital process which is in close
proximity to the jugal. Medially and posteriorly, the
left squamosal is very broad, probably because of
compression, and a parietal or supratemporal process
cannot be clearly identified. The left squamosal does
not meet the supratemporal process of the parietal but
this is here considered to be an artefact of preserva-
tion.
The shape of the squamosal in MSNM BES SC
265 can be described as follows. The bone forms a
“posterior ramus” projecting laterally from the skull
roof. Then it sharply bends antero-laterally, forming
an “anterior ramus”. Where the bone bends, a ventrally
projecting triangular lamina forms the quadrate process.
At approximately two thirds of its length, the anterior
ramus of the squamosal bends medially and tapers into
the postorbital process. The ventral side of the squa-
mosal is grooved.
Postorbital
No postorbitals could be identied in MSNM BES
SC 265.
Jugal
Both jugals are preserved. The left one, exposed in
lateral view, offers more anatomical detail.
The jugal has a triradiate shape, with a long
and slender suborbital process (=maxillary process
sensu Wild, 1974: 11) and two posterior processes:
a slender, free-ending quadratojugal process extend-
ing posteriorly, and a broad, long postorbital process
extending dorsally. The ventral margin of the jugal is
strongly thickened, forming a ridge along the lateral
surface of the bone. The lateral surface of the pos-
torbital process exhibits an elongate, dorso-ventrally
oriented groove.
Quadrate
The right quadrate is badly crushed and few details
can be discerned. It is considered to be exposed in lat-
eral view, with the articular condyle facing the lower
jaw and the cephalic condyle facing the quadrate
process of the squamosal. Only remnants of the left
quadrate were tentatively identied on the left side of
the skull.
Epipterygoid
An isolated, slender bony rod lying posterior to the
elements of the neurocranium is interpreted as an epip-
terygoid (Wild, 1974: 14, gs. 1, 8 b). Its presumed ven-
tral end is broken but clearly expanded.
Ceratobranchials
Lateral to the left exoccipital is a ceratobranchial I
(Wild, 1974: gs. 21, 87). The expanded proximal end of
the bone is directed laterally.
Neurocranium
Elements of the neurocranium are randomly displaced
posterior to the parietals. As most of them are badly
crushed and distorted, their interpretation must remain
equivocal.
Both exoccipitals, fused to the partially preserved
opisthotics, are exposed just posterior to the parietals, and
still frame a very deformed foramen magnum.
The very poorly preserved supraoccipital lies further
posteriorly, probably in occipital view.
The basisphenoid-parasphenoid complex is rather
well preserved in ventro-lateral view. The cultriform
process, projecting from the sphenoidal plate anteriorly,
is distinct. The right alar process and the anterior part
of the basisphenoidal plate were pushed outwards. The
basipterygoid process is visible on both sides of the
basicranium.
Posterior to the basisphenoid-parasphenoid complex
lies a severely compressed and distorted element, which
was tentatively identied as the basioccipital.
Lower jaw
The mandibular rami separated at the symphysis
and shifted towards the median sagittal plane, so that
they are both exposed in lateral view. The right man-
dibular ramus was probably stretched during fossiliza-
tion, because it is a little longer than the left one. The
mental foramina open at the bottom of small pits which
are arranged in a series on the lateral side of the den-
taries. The dentaries still bear some conical and sharply
pointed teeth at a variety of developmental stages. The
presence of poorly distinct cusps on the posteriormost
preserved tooth on the left lower jaw remains equivo-
cal. The dorsally convex suture between the dentary
and surangular is present on both mandibular rami. The
same is true for the sutures dening the articular; on the
left side the prearticular is also identiable, ventral to
the articular. The presence of a coronoid remains uncer-
tain (p. 61, Fig. 49).
13
TANYSTROPHEUS LONGOBARDICUS: RE-INTERPRETATIONS OF THE ANATOMY BASED ON NEW SPECIMENS FROM BESANO
Axial skeleton
(Figs. 1, 4-8, 54, 56 A, 58, Pls. I-II, Tabs. 2-3)
Cervical vertebrae and ribs (Figs. 4, 54, 56 A, 58, Pls. I-II,
Tab. 2)
Twelve cervical vertebrae are preserved in right lat-
eral view. Vertebrae two through eleven are complete.
Only remnants of the atlas are preserved. Part of its right
neural arch, is still articulated with the axis (Fig. 56 A).
Two elements posterior to the parietals are tentatively
interpreted as the remnants of the left atlas neural arch
and the articulated proatlas (Fig. 3). These elements,
however, are very poorly preserved. Moreover, the pre-
sumed proatlas is larger relative to the axis than the same
element in MSNM BES SC 1018 (Fig. 12). The neural
spine on the 12th vertebra (Fig. 4) is damaged and it is
overlapped by the left scapula. Cervical vertebrae three
through ten are deeply grooved because of the collapse
of their bony walls inside the hollow vertebral centra
(Wild, 1974: 81).
Cervical vertebrae two through twelve are still articulated,
except for the eight and ninth element. At the joints between
the 11th and the 12th cervical vertebrae and that between the
12th cervical vertebra and the rst dorsal, the cervical verte-
bral column is bent backwards by almost 180° (Figs. 1, 4,
Pls. I-II). It is no longer articulated with the skull.
Compression during fossilization makes it impossible
to take reliable measurements of the vertebrae, apart from
their length (Tab. 2).
Number
of cervical
vertebra
MSNM BES
SC 265
Centrum length
(mm)
MSNM BES
SC 1018
Centrum length
(mm)
2 13.8 17.0*
3 37.2 48.0*
4 40.5 59.8
5 40.0* 60.0*
6 41.8 59.7
7 46.1 69.8
8 57.3 82.8
9 62.0 89.3
10 57.9 n. m.
11 35.0 35.5
12 19.2 -
Table 2 – Tanystropheus longobardicus, length of the cer-
vical vertebrae in MSNM BES SC 265 and MSNM BES
SC 1018 (* = estimated; n. m. = not measurable).
Fig. 4 – Tanystropheus longobardicus, MSNM BES SC 265, twelfth cervical vertebra and dorsal vertebrae one through three. Water-
color: Massimo Demma.
The cervical ribs lie close together forming a thick,
paired bundle that runs parallel to the cervical column
along the cervical series seven through twelve (Fig. 1, Pls.
I-II). Anterior to the seventh cervical, with the stronger
dorsal bending of the cervical column, the rib bundle has
14 STEFANIA NOSOTTI
become detached from the vertebrae and projects forward
in a straight line, while the cervical column is bent back-
wards. However, one of the ribs of the fourth pair and all
the ribs of the rst through third pairs follow the bending
of the vertebral column, and lie randomly displaced to its
left side. The ribs of the atlas and the axis cannot be dis-
tinguished from one another.
Dorsal vertebrae and ribs (Figs. 4-5, 7-8, Pls. I-II)
Thirteen dorsal vertebrae are preserved, some of
them very poorly. Dorsal vertebrae one through eleven
are exposed in right lateral view. The rst dorsal is more
or less in its original position relative to the 12th cervical
(Fig. 4). The latter is tilted backwards, so that the zyga-
pophyses of the two vertebrae are slightly shifted relative
to each other. Dorsal vertebrae one through three (Fig. 4)
and ve through nine (Fig. 7) still form articulated series.
The fourth dorsal is somewhat displaced relative to the
preceding and the subsequent vertebrae. The tenth ver-
tebra is exceptionally poorly preserved. The 12th and 13th
dorsal vertebrae (“lumbars”, Wild, 1974) are no longer
articulated. The rst is exposed in antero-lateral view, the
second in right lateral view (Fig. 5). The pleurapophysis
(sensu Wild, 1974; see note on p. 69) of the 12th dorsal is
broken away and lies ventral to the 13th dorsal. It is not
clear if it is complete. As preserved, it is 12.7 mm long
and 1.3 mm deep.
Fig. 5 – Tanystropheus longobardicus, MSNM BES SC 265, dorsal
vertebrae eleven through thirteen. Watercolor: Massimo Demma.
Measurements of the centrum length are generally
precluded because of poor preservation. The length of the
second dorsal vertebral centrum, which is the best pre-
served of the dorsal series, is approximately 11 mm and
that of the 11th is approximately 10 mm.
Few of the dorsal ribs are preserved. They are disar-
ticulated and scattered but some of them still lie close to
the corresponding vertebra. The ribs of the rst and third
pairs are dichocephalous (Fig. 4), while the remaining
preserved ribs are holocephalous.
Sacral vertebrae (Figs. 6, 8, Pls. I-II)
Two sacral vertebrae are poorly preserved next to the
right ilium and partly overlapped by the right femur. They
exhibit stoutly built pleurapophyses (sensu Wild, 1974; Fig. 6 – Tanystropheus longobardicus, MSNM BES SC 265, sacral and
proximal caudal vertebrae. Scale bar 20 mm. Photo: Massimo Demma.
15
TANYSTROPHEUS LONGOBARDICUS: RE-INTERPRETATIONS OF THE ANATOMY BASED ON NEW SPECIMENS FROM BESANO
see note on p. 69). They remain articulated together but
completely separate from the dorsal and the caudal series.
The centra lie at right angles to the rest of the vertebral
column, and are exposed in right lateral view. The pleu-
rapophysis of the presumed rst sacral is complete, while
that of the presumed second sacral is probably truncated.
The distal end of the complete pleurapophysis is greatly
expanded and bears a wide triangular surface that con-
tacted the ilium.
Caudal vertebrae (Figs. 1, 6, 8, Pls. I-II, Tab. 3)
The caudal series begins posterior to the right femur.
It is estimated that there are at least 40 caudals (p. 70).
With the possible exception of the vertebral pairs three-
four, four-ve, and ve-six, the caudal vertebrae form a
complete articulated series.
Caudals one through three (Figs. 6, 8) are damaged
and deformed. The rst caudal is poorly preserved in
right lateral view. The distally truncated right pleu-
rapophysis (sensu Wild, 1974; see note on p. 69) is
crushed upwards and there is a remnant of the neural
arch. The centra of the second and third caudals are
partially rotated around their longitudinal axis and are
exposed in latero-ventral view. The right pleurapophy-
sis of the second caudal and both pleurapophyses of
the third are preserved. The right pleurapophysis of the
third caudal is 11 mm long and 2 mm deep. The right
postzygapophysis of the second caudal and the corre-
sponding prezygapophysis of the third are exposed in
ventral view.
Distal to the third caudal, the series is exposed in
right lateral view. The fourth and fth caudals (Figs.
6, 8) are largely concealed by the overlapping bones of
Table 3 – Tanystropheus longobardicus, length of the
proximal caudal vertebrae in MSNM BES SC 265
(* = estimated; n. m. = not measurable).
Number of
caudal vertebra
Centrum length
(mm)
1 11.0*
2 10.0*
3 10.0*
4 n. m.
5 n. m.
6 n. m.
7 11.1
8 12.6
9 13.2
10 13.5
11 14.2
the right hindlimb and the cervical series. Remnants of
their pleurapophyses project vertically out of the slab,
indicating that these vertebrae lie on their left side. The
distal end of the pleurapophyses of these vertebrae was
probably broken away. Indeed there is a fragment of
one such pleurapophyses preserved ventral to the fth
caudal.
Caudals six through eleven (Fig. 6) are very well pre-
served (the sixth one is, however, partially overlapped
by the cervical ribs bundle). Distal to the 11th element,
the caudals still form an articulated series (27.5 cm in
length). The centra are the only preserved portion of
these vertebrae, and the articulation between them is
difcult to discern.
The measurements of the proximal caudal vertebrae of
MSNM BES SC 265 are given in Tab. 3.
Gastralia
The gastralia randomly cluster in the trunk region.
Few of them are preserved.
Appendicular skeleton
(Figs. 1, 7-9, 65, Pls. I-II, Tabs. 4-5, 8)
The most interesting features of the appendicular skel-
eton of MSNM BES SC 265 are the presence of sesamoid
bones in the elbow and the knee joints, and the left tarsus,
whose morphology is discussed on page 72.
Pectoral girdle (Fig. 7)
The elements of the pectoral girdle are displaced and
only partially preserved and exposed. Only the right
scapula and the interclavicle are missing.
The left scapula lies dorsal to the 11th and 12th cervical
vertebrae and is exposed in lateral view. Its short pedun-
cle bears the articular facet that contributes to the glenoid
fossa. The antero-dorsal margin of the bone is heavily
corroded but the posteriorly projecting process of the
scapular blade is well preserved.
The coracoids are preserved ventral to the ante-
riormost dorsal vertebrae, partially overlapping each
other. They are plate-shaped bones, with a short and
stocky glenoid process. A supracoracoideal foramen is
not clearly identiable. The two coracoids are exposed
in ventral view, because the articular facet that con-
tributes to the glenoid fossa is visible. Thus, the more
dorsal of the two is the left coracoid, shifted across the
right one.
The clavicles are partially exposed ventral to the rst
dorsal rib. They are rod-shaped, curved elements offering
little anatomical detail. The more ventral element might
be the medial part of the left clavicle, and the other one
the lateral part of the right clavicle (Wild, 1974: 103, gs.
65, 91).
Forelimb (Fig. 7, Tabs. 4, 8)
The forelimbs are fairly well preserved. However,
some of their elements are missing and few anatomical
detail of the preserved bones can be described. They lie
side by side, the right limb uppermost and overlapping the
left one. Because of the displacement of the bones of the
pectoral girdle, the limbs are no longer articulated with
the latter.
16 STEFANIA NOSOTTI
Fig. 7 – Tanystropheus longobardicus, MSNM BES SC 265, forelimbs. I-IV = Metacarpals. Scale bar 25 mm. Photo: Luciano Spezia.
Pencil: Fabio Fogliazza.
17
TANYSTROPHEUS LONGOBARDICUS: RE-INTERPRETATIONS OF THE ANATOMY BASED ON NEW SPECIMENS FROM BESANO
The right forelimb (Fig. 7) is almost completely
preserved. However, its elements are disarticulated and
displaced relative to one another. The right humerus is
preserved in dorsal view. The epipodials are no longer
articulated with the humerus and are displaced more or
less as a unit. Given the relative position of the radius
and the ulna, the radius is preserved in anterior view and
the ulna, probably, in antero-lateral view. The distal end
of the ulna is displaced over that of the radius. The very
poor preservation of the manus hinders its interpretation
but it seems reasonable to assume that it is preserved in
dorsal view. The manus is approximately in its original
position relative to the forearm but no longer articulated
with it. Only two elements of the carpus, possibly the
ulnare and the medial distal carpal (homology unknown),
are poorly preserved. Differently from metacarpals I-IV,
metacarpal V cannot be reliably identied. Metacarpal I
is displaced proximally and no longer aligned with the
others. By comparison with MSNM BES SC 1018, all
the phalanges are preserved but they are scattered, and
their proximo-distal original sequence cannot be reliably
reconstructed.
The left forelimb (Fig. 7) is not complete, because
the distal ends of the epipodials and the manus are not
preserved. The humerus - exposed in dorsal or dorso-
posterior view - lies along the same longitudinal axis as
the forearm, and is still articulated with it. A tiny sesam-
oid bone is identiable in the elbow joint. The epipodials
lie side by side exposed in anterior view. The proximal
end of the ulna overlaps the radius, and its distal end is
overlapped by the radius.
The proximal ends of the humeri have a at articular
surface. Distally, the ulnar condyle is more prominent
than the radial one. The epipodials have slightly and
equally expanded proximal ends, with at or slightly
convex articular surfaces. The ulnar margin of the radius
and both margins of the ulna are concave.
Measurements of the elements of the forelimbs in
MSNM BES SC 265 are given in Tab. 4. Length ratios
between different elements of the forelimbs are given in
Tab. 8.
The right ilium is well preserved, albeit partly con-
cealed by the large pleurapophysis of the rst sacral
vertebra. The dorsal iliac blade projects far posteriorly,
ending in a blunt tip. Anteriorly, it is short and rounded
and it is delimited by a supra-acetabular buttress from
the wide ventral acetabular fossa. As the acetabular
fossa extends farther anteriorly than the dorsal blade,
there is a waisted area between the two parts. A badly
crushed and deformed element ventral to the 13th dorsal
vertebra is tentatively interpreted as the left ilium. Two
shallow fossae might represent facets for the sacral pleu-
rapophyses.
The right pubis is largely concealed by the overlap-
ping left ilium and only its ventralmost portion, antero-
posteriorly expanded, is exposed.
The ischia lie posterior to the right femur. The right
ischium is partly concealed by the overlapping caudal
vertebrae. The anterior portion of the ischiadic plate and
acetabular process is exposed. The left ischium is partly
concealed by the overlapping right femur. The ventral
portion of the ischiadic plate and part of the acetabular
process are exposed.
Hindlimb (Figs. 8-9, Tabs. 5-8)
The hindlimbs of MSNM BES SC 265 are nearly
complete and articulated. The left hindlimb lies ventral
to the trunk, as expected for a dead animal lying on its
left side. By contrast, as a result of the twisted sacral
and proximal-caudal regions of the vertebral column,
the right hindlimb has been raised and turned upside
down. Because of the displacement of the bones of the
pelvic girdle, the limbs are no longer articulated with
the latter.
The right hindlimb (Fig. 8) is almost complete but
its elements are partially disarticulated and displaced.
The femur exposes its antero-ventral side. The articu-
lar surface on its proximal end is grooved. The distal
end displays poorly differentiated condyles, the tibial
condyle being slightly more prominent than the bu-
lar one. The epipodials are no longer articulated with
either the femur or the tarsus. They are preserved in
medial, slightly anterior view. The proximal end of
the tibia overlaps that of the bula. Distally, the bula
was displaced posteriorly. The proximal and the distal
ends of the tibia are slightly expanded and convex. The
distal end of the bula is not expanded. The pes is pre-
served in dorsal view. The only identiable elements
of the tarsus are the calcaneum and the astragalus (Fig.
9), the latter poorly preserved and displaced relative to
the calcaneum. All the metatarsals are preserved; their
proximal ends are no longer aligned, and metatarsal I is
displaced medially. Metatarsals II through IV and some
phalanges are partially concealed by the overlapping
cervical vertebral column. By comparison with MSNM
BES SC 1018, all the phalanges, except the ungual pha-
lanx of digit ve, were identied, and the phalangeal
formula is 2,3,4,5,4. The phalanges are mostly disar-
ticulated but their proximo-distal sequences can be still
identied (Fig. 9).
The left hindlimb (Fig. 8) is better preserved than the
right one. Its elements are all identiable and only the
distal phalanges of the digits are partially displaced. The
femur is exposed in anterior view. The shaft is gently,
sigmoidally curved. The proximal end is expanded and
Table 4 – Tanystropheus longobardicus, measurements of
the forelimbs in MSNM BES SC 265 (n. m. = not measur-
able).
Length (mm)
Right Left
Humerus 50.3 49.5
Radius 35.3 n. m.
Ulna 32.8 n. m.
Pelvic girdle (Fig. 8)
The elements of the pelvic girdle are displaced and
only partially preserved and exposed.
The elements of the right side of the girdle are all pre-
served and are thicker than the corresponding elements of
the left side. They maintain their original position, having
simply moved away from one another, and are exposed in
lateral view. The ischium, and possibly the ilium - both
exposed in medial view - are the only identiable ele-
ments of the left side of the girdle.
18 STEFANIA NOSOTTI
Fig. 8 – Tanystropheus longobardicus, MSNM BES SC 265, hindlimbs. Scale bar 30 mm. Photo: Luciano Spezia. Pencil: Fabio Fogliazza.
19
TANYSTROPHEUS LONGOBARDICUS: RE-INTERPRETATIONS OF THE ANATOMY BASED ON NEW SPECIMENS FROM BESANO
weakly convex. The distal end is deected ventrally.
Only the tibial condyle is exposed, and its rounded con-
tour indicates that the articular surface for the tibia was
sub-cylindrical in shape. The crus is still articulated with
the femur, and rotated on its longitudinal axis together
with the articulated pes. The crus is thus exposed in
postero-medial view and the pes in plantar view. A
small sesamoid bone is preserved in the knee joint. The
tibia has expanded ends, with the proximal end at, and
the distal end slightly convex. The proximal end of the
bula is concealed by the tibia, but its weakly convex
and not expanded distal end slightly overlaps the tibia.
The pes is displaced laterally relative to the crus. The
tarsus (Figs. 9, 65) comprises four ossied elements in
close juxtaposition: astragalus, calcaneum, distal tarsal
three and distal tarsal four. The astragalus and the cal-
caneum meet along a straight line and the astragalus
slightly overlaps the calcaneum proximally with a lat-
eral process. A foramen for the perforating artery (Wild,
1974: 116) cannot be identied. The astragalus is a
rather stocky element, elongate transversally relative to
the longitudinal axis of the crus. Distally it forms a shal-
low fossa. The calcaneum has a weakly polygonal shape.
Distally, it articulates with metatarsal V. Distally and
medially a line seemingly indicating a contact between
the calcaneum and distal tarsal four (Fig 65) represents
a crack at the bottom of a shallow fossa. The proximal
margin of distal tarsal four is not distinct. As preserved,
distal tarsal four partly overlaps the calcaneum (Fig. 9)
and it is probably a larger element than it appears. Distal
tarsal four contacts metatarsal V laterally and the proxi-
mal head of metatarsal IV distally. Distal tarsal three is
displaced and overlaps distal tarsal four. The metatarsals
and the phalanges are not greatly displaced relative to
one another: metatarsal I through III are displaced proxi-
mally, and some phalanges are no longer articulated but
their sequence in each digit remains clearly identiable
(Fig. 9). Metatarsals II through IV, and some phalanges
of the digits are partially concealed by the overlapping
cervical vertebral column. The proximal ends of the
metatarsals overlap, each lateral element overlapping
the medial one. By comparison with MSNM BES SC
1018, all the phalanges are preserved, the phalangeal
formula being 2,3,4,5,4.
Measurements of the elements of the hindlimbs in
MSNM BES SC 265 are given in Tab. 5. Length ratios
between different elements of the hindlimbs are given in
Tab. 8.
Fig. 9 – Tanystropheus longobardicus, MSNM BES SC 265, pedes. I-V = metatarsals. Drawing: Fabio Fogliazza.
20 STEFANIA NOSOTTI
Specimen MSNM BES SC 1018
(Figs. 10-29, 49 E, 53 B, 54, 56 B, 59, 62, Pls. III-IV,
Tabs. 2, 5-6, 8)
The specimen consists of a partially articulated and
beautifully preserved skeleton lying on its right lateral
side. The skull is not properly articulated with the neck,
because the occipital elements are displaced but lies next
to it. The neck is tilted backwards and is separated from
the trunk. However, cervical vertebrate two through ten
form an articulated series, except for a separation of the
fourth from the fth. The trunk region comprises a series
of dorsal vertebrae, isolated dorsal ribs, and clustering
gastral ribs. The preserved proximal caudal vertebrae
are scattered on the slab, while some of the distal caudals
are articulated and form two isolated series. The pectoral
girdle with both forelimbs, and the left pelvic girdle with
the left hindlimb, clearly lie in their original position rela-
tive to the trunk. The right pelvic girdle and hindlimb are
disarticulated, and their elements randomly displaced.
The estimated overall length of the specimen is 140 cm.
Skull
(Figs. 10-13, 49 E, 53 B)
In spite of incomplete preservation, the skull of
MSNM BES SC 1018 provides some superb anatomi-
cal details, and probably represents the best example of
the overall morphology of the skull in all of small-sized
specimens of T. longobardicus to date. It is exposed in
left ventro-lateral view, so that both mandibular rami
are exposed, the left one in lateral view, the right one in
medial view. The length of the skull from the anterior end
of the left premaxilla to the posterior end of the left lower
jaw is 5.8 cm.
The skull is crushed, and the majority of bones are
displaced to a variable degree. In particular, many of the
unidentied elements are randomly displaced posteriorly.
By contrast, elements preserved in the antorbital region
are still mostly articulated. Dermal bones completing the
circumorbital series, and forming the skull roof are only
slightly displaced. Dermal bones of the palate are partly
exposed in between the mandibular rami. The lower jaws
Table 5 – Tanystropheus longobardicus, measurements of the hindlimbs in MSNM BES SC 265 and MSNM
BES SC 1018 (* = estimated; n. m. = not measurable).
MSNM BES SC 265 MSNM BES SC 1018
Right hindlimb (mm) Left hindlimb (mm) Left hindlimb (mm)
Femur 72.5 70.8 n. m.
Tibia 61.4 62.1 74.2
Fibula 62.0* 58.0* 74.0
Metatarsal I 20.6 20.7 27.8
Metatarsal II n. m. n. m. 34.6
Metatarsal III 31.4 30.8 37.7
Metatarsal IV 32.0* 29.0* 36.4
Metatarsal V 08.0* 08.5 11.1
Digit I
Phalanx I n. m. 14.43 13.6
Ungual n. m. 05.7
Digit II
Phalanx I n. m. n. m. 14.3
Phalanx II 07.3 07.4 09.3
Ungual n. m. 04.9 05.5
Digit III
Phalanx I 14.0 13.0* 17.2
Phalanx II 07.0 07.0* 08.9
Phalanx III n. m. 06.0* 07.2
Ungual 04.5 04.0* 05.1
Digit IV
Phalanx I n. m. 15.9 19.5
Phalanx II 07.7 07.6 09.3
Phalanx III 04.7 04.7 05.6
Phalanx IV 03.8 03.6 04.4
Ungual 03.0* 03.1 03.7
Digit V
Phalanx I 28.7 27.8 34.6
Phalanx II 09.5 09.1 12.4
Phalanx III n. m. 04.4 05.6
Ungual n. m. 03.0 03.4
21
TANYSTROPHEUS LONGOBARDICUS: RE-INTERPRETATIONS OF THE ANATOMY BASED ON NEW SPECIMENS FROM BESANO
are complete and in situ. Almost all the teeth of the upper
and lower left jaws are in situ, while those of the right side
are mostly displaced.
Premaxilla
Only the left premaxilla is exposed but it is com-
plete. It possesses a long and slender maxillary process
contacting the dorsal margin of the maxilla, and a short
but clearly delimited nasal process. Posterior to the
nasal process, the entire dorsal margin of the premaxilla
is free, and presumably framed the external naris latero-
ventrally (Figs. 36-37). Tiny foramina open anteriorly
on the dorsolateral surface of the premaxilla, possibly
transmitting branches of the medial ethmoidal nerve
(Oelrich, 1956).
The premaxilla possesses six teeth - ve of which are
preserved in situ - most probably representing the com-
plete premaxillary dentition. The teeth are long, conical
and slightly recurved, with nely striated enamel. The
third one is the longest. The premaxillary teeth interlock
with the anteriormost dentary teeth of the lower jaw.
Maxilla
Only the left maxilla is exposed. It is complete, and
still articulated with the premaxilla. It is a triangular
bone, bifurcating posteriorly into a long, tapering ven-
tral jugal process, and into a shorter, yet well developed,
dorsal process (processus frontalis, sensu Wild, 1974:
g. 2). The emarginated region between the two proc-
esses receives the lacrimal. Dorsal to the lacrimal, the
posterior margin of the dorsal process contacts the pre-
frontal. The dorsal margin of the jugal process no longer
contacts the jugal. Posterior to the contact with the maxil-
lary process of the premaxilla, the dorsal margin of the
maxilla is free. Originally, it presumably articulated with
the nasal (p. 46).
Fourteen maxillary teeth are preserved in situ. A gap
between the two posteriormost teeth suggests that the
full maxillary tooth count might be of 15. All the maxil-
lary teeth are tricuspid. The lateral surface of the maxilla
exhibits a series of shallow fossae laterally delimited by
bulges. Each of these fossae corresponds to a tooth posi-
tion.
Vomer
Remnants of the vomer are tentatively identied ante-
rior to the right palatine but because of very poor pres-
ervation they provide no further detail. However, some
isolated teeth are interpreted as vomerine teeth. They are
tiny and pointed, with striated enamel.
Palatine
Anterior to the pterygoid, two more dentigerous bones
are interpreted as the palatines. The element that lies
immediately anterior to the pterygoid, with four large
alveoli (diameter approximately 1.4 mm), is interpreted
as the left palatine exposed in ventral view. It still con-
tacts the pterygoid, albeit probably slightly displaced
(the anteriormost tooth position of the pterygoid is partly
covered with bone). The maxillary process is apparently
crushed against the dentigerous region of the bone. The
element anterior to the left palatine is interpreted as the
right palatine exposed in ventral view. It also exhibits four
tooth alveoli approximately as large as those of the left
palatine. None of the preserved isolated teeth were identi-
ed as palatine teeth.
Pterygoid
Only the left pterygoid is exposed between the man-
dibular rami, and it is preserved in ventral view. It shows a
well developed transverse process, and a slender, tapering
quadrate process. It bears a row of at least 12 tooth posi-
tions whose diameter becomes smaller posteriorly (from
approximately 0.9 to 0.7 mm). No pterygoid tooth is pre-
served in situ, and I was not able to single out pterygoid
teeth among the isolated preserved teeth.
Ectopterygoid and Nasal
No ectopterygoids or nasals could be identied in
MSNM BES SC 1018.
Frontal
Both frontals are well preserved, albeit separated.
The left frontal rotated upwards on its longitudinal
axis, exposing the ventral side. It exhibits a large axe-
shaped lateral ange, which is medially delimited by
a prominent, sinuously curved ridge. Therefore, the
ange is concave near the median sagittal plane, and
deepened into the cranial cavity. The anterior, lateral,
and posterior margins of the ange are free. Medial
to the lateral ange, the main body of the left frontal
extends between the sinuous margin bordering the lat-
eral ange and a straight, yet irregular, medial margin
originally contacting that of the right frontal. The ante-
rior end of the main body of the frontal is bifurcate,
with a lateral and medial lobes interlocking with the
prefrontal. A notch in the anterior margin of the lateral
ange demarcates laterally the anterior end of the main
body of the frontal. Posteriorly, the left frontal still
contacts the parietal but the latter is badly crushed and
provides no information about the nature of the contact
between the two bones.
The right frontal is displaced into the orbit as an iso-
lated element, and it is exposed in dorsal view. Its lateral
ange is partly concealed by the left frontal and prefron-
tal. The main body of the right frontal is thick, and raised
above the lateral ange. As in the left frontal, a notch
between the anterior end of the main body of the frontal
and its lateral ange is distinct. A notch in the postero-
medial end of the right frontal is interpreted as the antero-
lateral margin of the parietal foramen (p. 51).
Parietal
Both parietals are preserved. The left one is crushed
and broken into small pieces although is approximately in
its original position, posterior to the left frontal. The right
parietal, exposed in ventral view, is displaced, yet still in
close proximity to the right frontal. Posteriorly it is over-
lapped by the left postorbital. Its straight lateral margin is
only slightly thickened. Its straight medial margin (crista
cranii, sensu Wild, 1974: g. 2) and oblique antero-medial
margin are thickened. Posteriorly, the antero-medial
margin is recessed into a subtriangular fossa which bor-
dered the parietal foramen postero-laterally.
Supratemporal
No supratemporals could be identied in MSNM BES
SC 1018.
22 STEFANIA NOSOTTI
Fig. 10 – Tanystropheus longobardicus, MSNM BES SC 1018, skull. Scale bar 10 mm. Photo: Luciano Spezia.
23
TANYSTROPHEUS LONGOBARDICUS: RE-INTERPRETATIONS OF THE ANATOMY BASED ON NEW SPECIMENS FROM BESANO
Fig. 11 – Tanystropheus longobardicus, MSNM BES SC 1018, skull. Watercolor: Massimo Demma.
24 STEFANIA NOSOTTI
Prefrontal
Only the left, completely preserved, prefrontal is
exposed. The strong compression renders the interpre-
tation of its complex three-dimensional shape difcult
(p. 52, Fig. 37). I assume that the posteriormost part of
the prefrontal rotated upwards on a longitudinal axis
together with the frontal, thus exposing the postero-ven-
tral aspect.
The prefrontal is a large bone extending dorso-ven-
trally between the skull roof and the lacrimal, and forming
almost entirely the anterior margin of the orbit, as well
as its anterior wall. Dorsally, the prefrontal contacts the
frontal with a frontal process tting into the bifurcate
anterior end of the main body of the frontal. Ventrally, the
prefrontal is slightly shifted relative to the overlapping
lacrimal, and the facet receiving the lacrimal is clearly
visible. Anteriorly, the prefrontal contacts the dorsal proc-
ess of the maxilla.
Laterally, the prefrontal develops a sinuous crest close
to the orbital margin, which is dorsally expanded into a
rounded “lobe” (p. 51). In the fossil, this lobe was raised
and is exposed in ventral view. The dorsalmost part of the
prefrontal, medial to the crest and facing anteriorly, is not
visible.
Nothing can be said about the presumed contact
of the prefrontal with the nasal, because no nasals are
preserved in MSNM BES SC 1018 (but see p. 52 for
discussion).
Lacrimal
Only the left lacrimal is exposed. It is a small, elongate
subtriangular element. Antero-ventrally it is received into
the posterior concavity of the maxilla, and contacts the
dorsal margin of its jugal process. Dorsally, it is sutured
to the prefrontal, entering the anterior margin of the orbit
to a very limited extent. Posteriorly, the lacrimal contacts
the anterior end of the suborbital process of the jugal. A
lacrimal foramen is positioned close to the contact of the
lacrimal with the maxilla.
Postfrontal
Remains of the left postfrontal are probably preserved
between the left parietal and supraorbital process of the
postorbital (see below).
Squamosal and Postorbital
The identication of the squamosal and postorbital in
MSNM BES SC 1018 is problematic.
An arched element forming the posterior margin of the
orbit was identied as the left postorbital. Remnants of
bone projecting posteriorly from the arch are interpreted
as part of the postorbital (squamosal process). Conse-
quently, this element appears to be triradiate. The precise
extent of the antero-dorsal supraorbital process (=orbital
process sensu Wild, 1974: g. 2) of the postorbital cannot
be determined, because this area of the skull is badly
crushed. It is assumed that this area comprises remnants of
the left parietal, postfrontal and postorbital (supraorbital
process) but the inter-relationships of the three elements
cannot be discerned and are only tentatively represented
in Fig. 12. Ventrally, the postorbital forms a bipartite
jugal process. Each ramus of the jugal process tapers
into a blunt tip. The anterior ramus of the jugal process is
in close proximity to the groove on the postorbital proc-
ess of the jugal. Postero-dorsally the postorbital forms a
squamosal process, that is no longer articulated with the
squamosal. Based on this interpretation of the postorbital,
the squamosal would not be preserved in MSNM BES SC
1018.
Alternatively, the remnants of bone assumed to be
part of the postorbital (squamosal process) might rep-
resent the left squamosal, and the postorbital would be
an arched rather than a triradiate element. I also consid-
ered that an unidentied element preserved dorsal to the
postorbital might represent a squamosal. However, by
comparison with other specimens of Tanystropheus and
related taxa, these latter interpretations of the squamosal
and postorbital appear to be less supported (see discus-
sion on pp. 53-55).
Jugal
Both jugals were identied but the right one is con-
cealed to a large extent, and only part of its thickened
ventral margin is exposed.
The left jugal is perfectly preserved in lateral view,
closely resembling in shape that described in MSNM
BES SC 265. However, the jugal of MSNM BES SC
1018 provides more anatomical detail. In particular, the
longitudinally elongate groove on its postorbital process
is far more distinct and a crest delimits the groove pos-
teriorly. Posterior to the crest a shallow fossa is clearly
visible.
Quadrate
The isolated left quadrate, badly crushed and not com-
plete, is situated posterior to the jugal. It is considered to
be in lateral view, with a concave posterior margin and
an expanded dorsal cephalic condyle. The medial lamina
is partly exposed. The isolated right quadrate was tenta-
tively identied close to the right mandibular ramus but
this element is strongly distorted and does not warrant
description.
Ceratobranchials
Ceratobranchial I lies in close proximity to the right
lower jaw. It is very similar in shape to that described for
MSNM BES SC 265.
Neurocranium
Among the bones of the neurocranium only the
supraoccipital was identied. It is exposed as an isolated
element in dorsal view and exhibits a distinct sagittal crest
(=crista occipitalis, Wild, 1974: 14).
Scleral plates
Isolated scleral plates are preserved in MSNM
BES SC 1018. Four, with a subrectangular shape, are
positioned dorsal to the skull. Other elements, mostly
with a very irregular shape and sometimes with a pitted
surface, lie inside the orbit. They all apparently repre-
sent crushed and poorly preserved scleral plates. It is
impossible to determine the full number of the scleral
plates.
Lower jaw
Both mandibular rami are rather well preserved, the
left in lateral view and the right in medial view. They are
still close to each other at the mandibular symphysis.
25
TANYSTROPHEUS LONGOBARDICUS: RE-INTERPRETATIONS OF THE ANATOMY BASED ON NEW SPECIMENS FROM BESANO
Fig. 12 – Tanystropheus longobardicus, MSNM BES SC 1018, skull. Red teeth: isolated maxillary or dentary teeth. The extent of
the left parietal and postfrontal, and the dorsal extent of the left postorbital (supraorbital process) are arbitrarily chosen. Drawing:
Massimo Demma.
26 STEFANIA NOSOTTI
Fig. 13 – Tanystropheus longobardicus, MSNM BES SC 1018, lower jaws. Drawing: Massimo Demma.
27
TANYSTROPHEUS LONGOBARDICUS: RE-INTERPRETATIONS OF THE ANATOMY BASED ON NEW SPECIMENS FROM BESANO
In the left mandibular ramus (Figs. 13, 49 E) the
dentary is complete. As previously mentioned, the
anteriormost conical dentary teeth interlock with the
premaxillary teeth. The rst dentary tooth, preserved in
the right lower jaw, cannot be seen on the left side but
four subsequent teeth are distinct. One more conical
tooth is almost completely concealed by the two pos-
teriormost premaxillary teeth. Thus there was a total
of six conical teeth anteriorly. The tooth immediately
posterior to the sixth dentary tooth is unequivocally
tricuspid, as are all the subsequent teeth. These teeth
are overlapped to a large extent by the corresponding
teeth of the maxilla but they can be readily counted.
There are eleven tricuspid teeth preserved, and two
empty tooth positions (eighth and 11th), giving a total
of 13 tricuspid teeth. The dentary tricuspid teeth are not
interlocking with but lingual to the maxillary tricuspid
teeth. Shallow fossae delimited by bulges on the side
of the dentary mark the position of each tooth alveolus.
Posterior to the dentary, the left mandibular ramus is
crushed and fragmented. The oblique line anteriorly
delimiting the crushed posterior part of the lower jaw
probably represents the suture between the dentary and
the surangular (p. 59, Fig. 49). The conguration of
the postdentary elements is not entirely clear in the left
lower jaw. Only the prearticular can be clearly identi-
ed. It has shifted ventrally relative to the articular.
A facet on the articular was for the reception of the
prearticular. I also tentatively identied part of the
splenial below the surangular. It is uncertain whether a
coronoid was present or not (p. 61).
The right mandibular ramus is completely preserved
(Figs. 13, 53 B). The Meckelian canal runs along the
entire length of the dentary. The dentary teeth are mostly
displaced. The splenial is a large bone forming to a large
extent the ventral margin of the lower jaw. It extends
far anteriorly, roong part of the Meckelian canal, and
posteriorly it underlies the angular and the articular.
Between the anterior and posterior slender extensions,
the splenial forms a large part of the medial side of the
lower jaw. Posteriorly, the prearticular is clearly identi-
able as a ribbon-shaped sheet of bone overlapping the
articular ventrally. The sutures between the articular,
surangular, angular and the putative coronoid cannot
be determined, and I can only assume the approximate
position of these elements. However, the adductor fossa
bordered by these bones is perfectly preserved, as is the
articular fossa for the quadrate.
Axial skeleton
(Figs. 10-12, 14-20, 54, 56 B, 59, Pls. III-IV, Tab. 2)
Cervical vertebrae and ribs (Figs. 10-12, 14, 54, 56 B, 59,
Pls. III-IV, Tab. 2)
Eleven cervical vertebrae are preserved in left lateral
view. Vertebrae two through nine are complete. Some parts
of the atlas lie as isolated elements close to the axis. In
particular, both atlas neural arches are preserved in lateral
view (Figs. 10-12, 56 B). Their slightly dissimilar shape
is probably due to compression. The left atlas neural arch
overlaps the axis, and it is only slightly displaced relative
to the axis prezygapophysis. The right neural arch is an
isolated element preserved close to the axis. It is clearly
sutured to another element that is tentatively interpreted
as the proatlas. Two polygonal, articulated elements pre-
served dorsal to the left neural arch of the atlas (Figs. 10-
12) might be the atlas centrum and intercentrum but this
interpretation remains very doubtful. The tenth vertebra
is incomplete, lacking its posterior half. However, the
postzygapophysis of the tenth cervical is still preserved
in articular contact with the prezygapophysis of the 11th
(Fig. 14). The 11th cervical is completely preserved as an
isolated element. Partial collapsing of the bony walls into
the hollow vertebral centra (Wild, 1974: 81) is apparent in
the cervical series.
Fig. 14 – Tanystropheus longobardicus, MSNM BES SC 1018, elev-
enth cervical vertebra in left lateral view. Pz c10: postzygapophysis of
the tenth cervical vertebra. Scale bar 10 mm. Photo: Massimo Demma.
Cervicals two through ten form an articulated series,
with the exception of the fourth and fth (Fig. 59).
Although the cervical vertebral column of MSNM BES
SC 1018 is no longer articulated with the dorsal verte-
bral column - the elements of which are mostly isolated
scattered elements - it is clearly tilted backwards, as
commonly observed in other Tanystropheus specimens.
Although the cervical series is not properly articulated
with the skull, it maintains its approximate original posi-
tion relative to it.
Because of compression it is impossible to take reli-
able measurements of the vertebrae, apart from their
length (Tab. 2).
The third through tenth pairs of cervical ribs are pre-
served in close proximity to the corresponding centra,
sometimes still in articulation. Their shafts are exten-
sively broken. Three anteriormost shorter elements dis-
placed ventral to the axis were identied as the ribs of the
rst two cervical vertebrae. Other rib heads and fragments
of their shafts are scattered on the slab.
Dorsal vertebrae and ribs (Figs. 15-18, 20, Pls. III-IV)
Thirteen dorsal vertebrae are preserved. Five of them,
identied as ve through nine or six through ten, form a
partially articulated series still approximately positioned
in the trunk region (Pls. III-IV, Figs. 15, 20). They are
exposed in left lateral view. Three additional isolated
28 STEFANIA NOSOTTI
dorsal vertebrae are tentatively interpreted as one of the
rst three anteriormost dorsal, the fourth, and the 11th.
The anteriormost dorsal (Fig. 16) is exposed in left lateral
view: its centrum is 14.8 mm long. Both pre- and postzyg-
apophyses and the posterior end of the neural spine are
broken. The putative fourth dorsal (Fig. 17) is exposed in
right lateral view: its centrum is 15.4 mm long. A fracture
crosses the posterior area of the neural arch and continues
along the border of the posterior end of the centrum. The
putative 11th dorsal (Fig. 18) is exposed in right lateral
view: its centrum is 14.6 mm long.
Fig. 15 – Tanystropheus longobardicus, MSNM BES SC 1018, mid-
dorsal vertebra (see discussion on p. 68) in left lateral view. Scale bar 5
mm. Photo: Massimo Demma. Fig. 16 – Tanystropheus longobardicus, MSNM BES SC 1018, anteri-
ormost dorsal vertebra (see discussion on pp. 67-68) in left lateral view.
Scale bar 10 mm. Photo: Massimo Demma.
Fig. 17 – Tanystropheus longobardicus, MSNM BES SC 1018, pre-
sumed fourth dorsal vertebra (see discussion on p. 69) in right lateral
view. Scale bar 5 mm. Photo: Massimo Demma.
Fig. 18 – Tanystropheus longobardicus, MSNM BES SC 1018, pre-
sumed eleventh dorsal vertebra (see discussion on p. 69) in right lateral
view. Scale bar 5 mm. Photo: Massimo Demma.
The preserved, disarticulated dorsal ribs mostly cluster
in the trunk region. They are stout, longitudinally grooved
elements. All of them are holocephalous. A single, short
dichocephalous rib is preserved close to the isolated right
femur (Pls. III-IV). It was clearly associated with one of
the anteriormost dorsals.
29
TANYSTROPHEUS LONGOBARDICUS: RE-INTERPRETATIONS OF THE ANATOMY BASED ON NEW SPECIMENS FROM BESANO
Sacral vertebrae (Fig. 19)
One sacral vertebra is preserved as an isolated ele-
ment exposed in right lateral view. It bears a stout, albeit
crushed, pleurapophysis (sensu Wild, 1974; see note on
p. 69). The centrum is slender and the low neural spine
raises posteriorly into a rounded process.The amphicoe-
lous centrum of this vertebra is 14 mm long.
Caudal vertebrae (Pls. III-IV)
Twelve caudal vertebrae are scattered on the slab.
A group of six caudal vertebrae exposed in lateral or
ventro-lateral view is preserved in the same area as the
scattered elements of the right hindlimb. When measur-
able, the centrum length of these vertebrae is approxi-
mately 12 mm, although the caudal to the left of the
anteriormost preserved dorsal (Pl. IV, d1-3?) has a centrum
length of 16.4 mm. The centrum is amphicoelous. The
shape of these vertebrae is difcult to interpret but the
presence of pleurapophyses (sensu Wild, 1974; see note
on p. 69), sometimes very wide at the base, indicates that
they are most likely proximal caudals.
An additional proximal caudal vertebra is preserved
in antero-lateral view close to the distal part of the right
tibia. The right pleurapophysis is distally truncated and a
line at the very base of it might be interpreted as a suture
(see note on p. 69). The orientation of the prezygapo-
physeal articular facets is sub-horizontal. Ventral to the
amphicoelous centrum, the chevron is preserved. It is a
Y-shaped element and the two rami are proximally con-
nected by a transverse bony bar.
Three more caudal vertebrae lie in the area between the
cervical vertebral column and the left femur. Two of them
lie to the right of the rst articulated series of distal caudal
vertebrae. They are very deformed and their shape is dif-
cult to interpret. One of the two appears to preserve pleura-
pophyses and on this basis it might be assumed to be part of
the proximal caudal vertebral column. A third caudal lies to
the right of cervical ten, and is preserved in left lateral view.
The centrum is 16.9 mm long. By direct comparison with
MSNM BES SC 265, this vertebra is tentatively identied
as the ninth caudal. A pleurapophysis is lacking but a shal-
Fig. 19 – Tanystropheus longobardicus, MSNM BES SC 1018,
sacral vertebra in right lateral view. Scale bar 5 mm. Photo: Massimo
Demma.
low, yet distinct tiny fossa is present on its lateral surface.
This small fossa is seen in other isolated caudals of MSNM
BES SC 1018 ventral to the pleurapophyses.
Finally, two vertebrae preserved in anterior view are
identied as distal caudal vertebrae. They are very small
and lack pleurapophyses. One is close to the mid-dorsal
series in the trunk region. The second is partially over-
lapped by the proximal part of the right tibia and bula.
Its centrum is weakly amphicoelous; only the pedicel of
the neural arch is preserved.
Like MSNM BES SC 265, the distal caudals form two
articulated series in which the centra are the only pre-
served vertebral elements. The articulation between the
centra, however, is hardly identiable. The rst series is
5 cm long and is very poorly preserved. The second is 20
cm long and better preserved. Remnants of bone proximal
to the rst series probably represent other distal caudal
vertebrae but they are exceptionally poorly preserved.
Gastralia
The gastralia are preserved in their original position,
ventral to the trunk. They are closely packed, and, in spite
of disarticulation of the different elements, they clearly
form a “skeletal unit”. The anterior elements are preserved
in ventral view, while the posterior ones appear to be in
anterior view. It is estimated that the gastral skeleton
includes about 30 units, each composed by four elements.
Appendicular skeleton
(Figs. 20-29, 62, Pls. III-IV, Tabs. 5-6, 8-9)
Specimen MSNM BES SC 1018 is unique amongst
all known specimens of Tanystropheus with respect to the
beautiful preservation of the limbs. By contrast, the pre-
served elements of the pectoral and the pelvic girdles are
displaced, and some of them are fragmentary or broken.
Pectoral girdle (Fig. 20)
All the elements of the pectoral girdle, with the pos-
sible exception of the interclavicle, can be identied.
The left scapula is complete, and exposed in lateral view.
The thick ventral peduncle bears the articular facet contribut-
ing to the glenoid cavity. Dorsal to the peduncle, the scapular
blade is strongly expanded and fan-shaped, projecting more
posteriorly than anteriorly. It exhibits concentric striation.
The right scapula is badly broken but its overall shape is still
recognizable. It is preserved in medial view.
The two coracoids lie side by side, concealed partly by
the left humerus. They are plate-shaped elements. Their
glenoid area is not exposed. The element to the left is
probably the left coracoid in ventral view, and the one to
the right is the right coracoid in dorsal view.
The two clavicles are preserved in between the scapu-
lae. The element overlapping the left scapula is most
likely the left clavicle in ventral view, while the other
is the right clavicle in anterior or posterior view (Wild,
1974: 103, gs. 65, 91).
An element overlapped by the right clavicle is tenta-
tively interpreted as the interclavicle. Two distinct proc-
esses which embrace the clavicles project laterally from
the anterior plate. This presumed interclavicle has a shape
different from the only other interclavicle reported in the
material of Tanystropheus (Wild, 1974: 103, g. 64).
30 STEFANIA NOSOTTI
Forelimb (Figs. 21-23, Tabs. 6, 8-9)
Both forelimbs are superbly preserved in complete
articulation. They are without doubt the best preserved
forelimbs of all known Tanystropheus longobardicus
specimens.
The right humerus (Fig. 21) is exposed in anterior
view. Because of the twisting of the two articular ends
relative to each other (Wild, 1974: 106, g. 67), the
expansion of the proximal end and the prole of the distal
end are evident. Although the distal end is crushed, the
rounded shape of the two condyles is clear. The forearm is
preserved in anterior view, and the manus in dorsal view.
The radiale and the lateral distal carpal are each broken
into two pieces.
The left forelimb (Fig. 22) is stretched, the humerus,
exposed in ventral view, lying along the same longitudi-
nal axis as the forearm, exposed in anterior view. The
manus is exposed in dorsal view. Because of compres-
sion, the proximal end of the humerus lies on the same
plane as the distal one. A crack crosses the proximal end
of the humerus but all other elements are intact. A sesam-
oid bone is preserved in the elbow joint.
The humeri have a more or less straight shaft and
expanded ends. The proximal end is slightly convex, the
distal one is differentiated into two condyles.
Fig. 20 – Tanystropheus longobardicus, MSNM BES SC 1018, pectoral girdle and mid-dorsal vertebra in left lateral view. Scale bar
15 mm. Photo: Massimo Demma.
The epipodials are of sub-equal length, and have
approximately equally expanded ends. In anterior view,
the proximal end of the radius slightly overlaps that of the
ulna, and vice versa distally. The radius is stouter than the
ulna. Its ulnar margin is concave. Both ends are slightly
expanded and have at articular surfaces. The ulna is a
slender element with concave margins and expandend
ends, the proximal slightly more so than the distal one.
The articular surfaces are distinctly convex, particularly
the distal one. A spatium interosseum is present.
The carpus includes four ossied elements (Fig. 23).
The proximal two, the radiale and the ulnare, meet each
other along a straight line and enclose a well-dened per-
forating foramen. Following Wild (1974), the two distal
carpal elements are referred to as the medial and the lat-
eral one, because their homology is at present unknown.
Contra Wild (1974: 109, g. 70), the largest distal carpal
is the lateral one. It is a roundish element positioned
between the ulnare, proximally, and the proximal heads
of metacarpals III and IV, distally. Medially, the lateral
distal carpal contacts a rounded medial distal carpal. The
latter is in juxtaposition with the proximal heads of meta-
carpals II and III. There is a wide gap between the medial
distal carpal and the radiale, and the medial region of the
carpus is unossied.
31
TANYSTROPHEUS LONGOBARDICUS: RE-INTERPRETATIONS OF THE ANATOMY BASED ON NEW SPECIMENS FROM BESANO
Fig. 21 – Tanystropheus longobardicus, MSNM BES SC 1018, right forelimb. Scale bar 20 mm. Photo: Luciano Spezia.
32 STEFANIA NOSOTTI
Fig. 22 – Tanystropheus longobardicus, MSNM BES SC 1018, left forelimb. Scale bar 20 mm. Photo: Luciano Spezia.
33
TANYSTROPHEUS LONGOBARDICUS: RE-INTERPRETATIONS OF THE ANATOMY BASED ON NEW SPECIMENS FROM BESANO
Fig. 23 – Tanystropheus longobardicus, MSNM BES SC 1018, left
manus in dorsal view. Broken lines indicate the contours of the caudal
vertebra which was erased to highligth the carpus. I-V = metacarpals.
Scale bar 5 mm. Photo: Massimo Demma.
Metacarpals I and V are far shorter than the others,
with V being the shortest. Metacarpals II through IV are
distinctly longer. Metacarpal III is the longest, followed
by IV and then II. The proximal ends of metacarpals
II through IV overlap. In dorsal view each medial ele-
ment overlaps the lateral one. The phalanges do not
look as thick and shortened as in Wild’s reconstruction
(1974: g. 70). The phalangeal formula of the manus is
2,3,4,4,3 (see Tab. 9 for comparison with other protoro-
saurian taxa).
Measurements of the elements of the forelimbs in
MSNM BES SC 1018 are given in Tab. 6. Length ratios
between different elements of the forelimbs are given in
Tab. 8.
Pelvic girdle (Figs. 24-25, Pls. III-IV)
The left elements of the pelvic girdle (Fig. 24) are
better preserved than the right ones, albeit fragmentary.
The ilium, ischium and pubis approximately maintain
their position relative to one another but have separated.
They are exposed in lateral view. Fragments of the right
elements of the pelvis are scattered in the same area.
The left ilium has a posteriorly projecting dorsal blade
and a wide, ventral acetabular portion. The ilium contrib-
utes most to the acetabulum. A distinct waist separates
the two parts. The dorsal blade tapers posteriorly into a
blunt process. Anteriorly, it is concealed to a large extent
by a fragment of the supposed right pubis but it is clearly
far shorter than posteriorly. A short pre-acetabular proc-
Length (mm)
right left
Humerus n. m. 67.5
Radius 44.6 44.9
Ulna 43.8 44.1
Metacarpal I 7.3 7.6
Metacarpal II 13.0 13.0
Metacarpal III 14.2 14.0
Metacarpal IV 13.4 13.3
Metacarpal V 6.6 6.6
Digit I
Phalanx 1 5.1 5.4
Ungual 4.0 4.3
Digit II
Phalanx 1 5.6 5.8
Phalanx 2 4.2 4.0*
Ungual 4.5 4.0*
Digit III
Phalanx 1 6.6 6.9
Phalanx 2 4.3 4.3
Phalanx 3 3.9 3.8
Ungual 3.9 4.0
Digit IV
Phalanx 1 6.2 6.6
Phalanx 2 4.2 4.5
Phalanx 3 3.9 3.8
Ungual 3.7 3.9
Digit V
Phalanx 1 5.7 5.7
Phalanx 2 4.6 4.6
Ungual 3.7 3.7
Table 6 – Tanystropheus longobardicus, measurements
of the forelimbs in MSNM BES SC 1018 (* = estimated;
n. m. = not measurable).
34 STEFANIA NOSOTTI
Fig. 24 – Tanystropheus longobardicus, MSNM BES SC 1018, pelvic girdle. Scale bar 15 mm. Photo: Massimo Demma.
ess or tubercle is present anteriorly. The posterior part of
the dorsal blade deepens into a fossa, dorsally sulcated by
a groove. The acetabular fossa is dorsally framed by the
acetabular buttress. Its ventral margin is not identiable
but posterior to it the articular facet for the ischium is dis-
tinct. The right ilium (Pls. III-IV) is displaced close to the
11th cervical vertebra, and is exposed in lateral view. Only
its post-acetabular portion is preserved. The iliac blade
has the same shape as the left side.
The left pubis is complete. It is differentiated into an
acetabular part and a ventral blade. The latter is antero-
posteriorly expanded and its surface exhibits concentric
striation. In the acetabular region, a distinct articular
peduncle for the ischium projects posteriorly. Conse-
quently, the posterior margin of the pubis, bordering
the thyroid fenestra, is strongly concave. A wide, oval
obturator foramen pierces the pubis in this area. Ante-
rior to the peduncle, the acetabular facet of the pubis
is distinct. Anterior to the acetabular facet, the margin
of the pubis is damaged. In this area it contacted the
pre-acetabular part of the ilium and apparently formed
a pubic tuberosity. Contra Wild (1974: 112, g. 71), a
processus lateralis similar to that observed in the extant
squamates was not identied in the pubis of MSNM
BES SC 1018. Dorsally, the anterior margin of the pubis
is straight, ventrally, is very slightly convex. A fragment
of bone overlapping the anterior part of the iliac blade is
tentatively interpreted as the proximal part of the right
pubis.
Only the fan-shaped, ventralmost part of the left
ischium is preserved, broken in two pieces. Fragments of
the same region of the right ischium are preserved close to
the left pubis. The surface of the ischiadic blades displays
concentric striation.
The left side of the pelvic girdle of MSNM BES SC
1018 is reconstructed in Fig. 25.
35
TANYSTROPHEUS LONGOBARDICUS: RE-INTERPRETATIONS OF THE ANATOMY BASED ON NEW SPECIMENS FROM BESANO
Fig. 25 – Tanystropheus longobardicus, reconstruction of the pelvic
girdle (left side in lateral view) based on MSNM BES SC 1018. Water-
color: Massimo Demma.
Hindlimb (Figs. 26-29, Pls. III-IV, Tabs. 5, 8-9)
The preserved elements of the right hindlimb (Fig.
26, Pls. III-IV) are disarticulated and scattered. The right
femur is preserved in antero-dorsal view. The shaft is prox-
imally broken, at approximately four fths of the length
of the bone. Another piece lies to the right of this larger
piece. The isolated proximal end of the femur (Fig. 26)
lies ventral to the trunk. The articular surface is convex.
The distal end is deected ventrally, giving the shaft a
gently curved shape. The tibial condyle widely conceals
that for the bula. The rounded shape of the condyles sug-
gests that they formed sub-cylindrical articular surfaces
for the tibia. The right tibia is broken into two pieces. One
represents approximately two thirds of the overall length
of the bone and includes the proximal end. It is displaced
to a position dorsal to the trunk, and is exposed probably
in dorsal view. The crushed distal end with the distalmost
part of the shaft is preserved on the other side of the trunk.
Both ends of the tibia are expanded. The proximal end of
the right bula with the proximalmost part of its shaft is
still articulated with the tibia. The larger part of the shaft
with its distal end lies close to the femoral head. Disartic-
ulated elements of the right pes are preserved in the same
area as the femoral shaft. Distal tarsal three and distal
tarsal four together with metatarsals I through IV lie close
to one another, and there are some scattered phalanges.
The astragalus and calcaneum are the only articulated ele-
ments of the right pes. The astragalus and calcaneum unit
is exposed in plantar-distal view. The medial swollen part
of the astragalus was crushed on its lateral part, exposing
its distal surface. Consequently, the astragalus is strongly
deformed.
All the elements of the left hindlimb (Fig. 27) are pre-
served to the right side of the neck. Although not properly
articulated, the femur, crus, and pes are preserved close
to one another in their original proximo-distal sequence.
The left femur is exposed in antero-dorsal view. Its proxi-
mal head is broken into small pieces but it is still clearly
facing the acetabular part of the left ilium. The shaft is
gently, sigmoidally curved. The two rounded distal con-
dyles are exposed, the tibial widely concealing that for the
bula (Fig. 29). The tibia and the bula rotated slightly
clockwise on their longitudinal axis and fell as a unit. The
two bones lie thus parallel to each other and are probably
exposed in antero-medial view. In this view, the proximal
end of the tibia is more expanded than the distal one. The
medial margin of its shaft is slightly concave. The bula
is a slender bone with a sigmoidal shape and no expanded
ends. The tibia and the bula are equal in length. The pes
(Fig. 28) is exposed in plantar view. All the elements of
the tarsus are preserved (Fig. 62) but the articulated astra-
galus and calcaneum unit was displaced and turned upside
down, so exposing its dorsal side. The astragalus and
calcaneum meet along a straight line and enclose a well-
dened foramen for the perforating artery (Wild, 1974:
116). The articular facet for the bula formed by the two
bones is exposed on the proximal side of the astragalus
and calcaneum unit. On the distal side, both astragalus and
calcaneum contribute to a fossa, proximally delimited by
a sharp margin. The medial part of the astragalus is partly
overlapped by metatarsal V. Distal tarsal three and distal
tarsal four contact metatarsals III and IV respectively but
distal tarsal four does not meet the proximal end of meta-
tarsal V. Therefore some dislocation probably occurred.
The metatarsals and the phalanges are completely pre-
served in articulation. The proximal ends of the metatar-
sals overlap, each lateral element overlapping the medial
one. A small but clearly delimited ventral tubercle is seen
close to the lateral margin of metatarsal V. The phalangeal
formula of the pes is 2,3,4,5,4 (see Tab. 9 for a comparison
with other protorosaurian taxa).
Measurements of the elements of the left hindlimb in
MSNM BES SC 1018 are given in Tab. 5. Length ratios
between different elements of the hindlimbs are given in
Tab. 8.
Fig. 26 – Tanystropheus longobardicus, MSNM BES SC 1018, proxi-
mal end of the right femur. Scale bar 5 mm. Photo: Massimo Demma.
36 STEFANIA NOSOTTI
Fig. 27 – Tanystropheus longobardicus, MSNM BES SC 1018, left hindlimb. Scale bar 30 mm. Photo: Luciano Spezia.
37
TANYSTROPHEUS LONGOBARDICUS: RE-INTERPRETATIONS OF THE ANATOMY BASED ON NEW SPECIMENS FROM BESANO
Fig. 28 – Tanystropheus longobardicus, MSNM BES SC 1018, left pes in plantar view. Astragalus and calcaneum in dorsal view.
I-V = metatarsals. Scale bar 10 mm. Photo: Roberto Appiani.
38 STEFANIA NOSOTTI
Fig. 29 – Tanystropheus longobardicus, MSNM BES SC 1018, left
hindlimb, knee joint. Scale bar 10 mm. Photo: Massimo Demma.
Specimen MSNM V 3663 a b
(Figs. 30-32)
Specimen MSNM V 3663 is a skull with ve associ-
ated cervical vertebrae preserved on a part (V 3663 a), and
a natural mould with remains of bone and teeth on a coun-
terpart plate (V 3663 b). It is notable for its large size and
the preservation of some elements of the temporal region
and of the vomerine dentition.
By comparison with the skulls of MSNM BES SC 265
and MSNM BES SC 1018, the overall length of the indi-
vidual represented by this specimen is estimated to have
been at least 3 m. In addition, the length of the most com-
plete cervical vertebra in MSNM V 3663, representing
an element of the series six through nine, is 17 cm. This
length corresponds to that of the cervical vertebrae six-
seven in specimens estimated to be approximately 3.5 m
long (PIMUZ T 2790 and PIMUZ T 2819) in the PIMUZ
Collections (Wild, 1974: tab. 3).
The length of the skull, measured from the posterior-
most extent of the parietals (as preserved) to the anterior-
most extent of the right lower jaw, is 12 cm. The skull is
exposed in ventral view and very poorly preserved. It is
heavily crushed and the surface of the bone is worn away.
Few elements of the skull could be identied. In the pre-
orbital region there is a small part of the premaxilla with
conical premaxillary teeth, part of the right maxilla with
conical maxillary teeth, and the vomers. The latter exhibit
a series of tooth alveoli: seven on the right vomer and ve
on the left. Teeth are present in some of these. These teeth
are very small and apparently conical but their tips are
broken. The anteriormost of the preserved teeth on the left
vomer has a basal diameter of 1.3 mm.
Some elements of the skull roof, circumorbital series
and of the temporal region are also exposed. Remains
of the parietals and frontals can be discerned but they
are indistinct. By contrast, the right postfrontal, still
articulated to the fronto-parietal plate, can be clearly
identied. It is articulated with another element which is
interpreted as a fragmentary postorbital. Finally, the right
jugal is exposed in medial view. Posterior to the elements
described, a partial natural mould and remains of bone
probably represent the squamosal, and indicate the poste-
riormost extent of the skull.
The anteriormost part of the right dentary, exposed in
lateral view, overlaps the premaxilla. Another fragment of
the right lower jaw is preserved just anterior to the right
jugal. The fragmented, anteriormost part of the left den-
tary and a partial natural mould of the left lower jaw are
seen on the left side of the skull.
On the counterpart plate, a partial natural mould of
the right lower jaw and remains of it can be identied.
Moulds and remains of the premaxillary and maxillary
teeth are distinct as well.
The preserved cervical vertebrae are crushed and
fragmented and their morphology cannot be described
in detail. It can be only stated that they have the typical
elongate shape of cervicals three through ten. One pair of
fragmentary, originally articulated vertebrae is preserved
close to the skull. Another segment of the cervical verte-
bral column is represented by three, originally articulated
elements: the posterior end of one vertebra, a second more
complete vertebra and the anteriormost part of a third.
Fragments of cervical ribs run ventral to the vertebrae.
Specimen MSNM BES 215
(Fig. 33)
Specimen MSNM BES 215 is an isolated mid-dorsal
vertebra in posterior view (p. 68). It is signicant for its
large size and the peculiar morphology. Although embed-
ded in a slab, the vertebra is fairly three-dimensionally
preserved.
The vertebra lacks the dorsal part of the neural spine.
Its height, measured from the ventral surface of the cen-
trum to the maximum dorsal extent of the broken neural
spine, is 7.4 cm. The height of the centrum is approxi-
mately 3 cm. By comparison with the height of the cen-
trum of the mid-dorsal vertebrae of MSNM BES SC 1018
and the large T. conspicuus (Wild, 1974: g. 54), the
overall size of the individual to which MSNM BES 215
belonged is estimated to have been at least 5 m.
The centrum has an elliptic dorso-ventrally elongate
shape. Its articular surface is worn away but a thickened
39
TANYSTROPHEUS LONGOBARDICUS: RE-INTERPRETATIONS OF THE ANATOMY BASED ON NEW SPECIMENS FROM BESANO
Fig. 30 – Tanystropheus longobardicus, MSNM V 3663 a, plate. Scale bar 100 mm. Preparation: Fabio Fogliazza. Photo: Luciano Spezia.
margin and a shallow central fossa are identiable. Short
and stocky transverse processes project laterally from
the area between the centrum and the very base of the
neural arch. The neural canal is pear-shaped and lled
with matrix. Above the canal there are short postzyga-
pophyses with at articular surfaces that slightly slant in
a dorso-lateral/medio-ventral plane. These wide articular
surfaces reach the area just above the neural canal. A short
bony shelf develops above the neural canal, and forms the
oor of a very shallow postzygapophyseal trough (sensu
Rieppel, 2001), within which paired postzygapophyseal
canals (sensu Rieppel, 2001: gs. 2-3) can be seen,
albeit lled with matrix. This latter feature is interesting
because postzygapophyseal canals were never previously
described in the post-cervical vertebrae of Tanystropheus
(p. 68). For this reason MSNM BES 215 is here consid-
ered as Tanystropheus cf. longobardicus, albeit on the
basis of its stratigraphical position it could be ascribed to
T. longobardicus.
The neural spine of MSNM BES 215 is broken dor-
sally. A median longitudinal ridge runs along it, and
diverges ventrally above the neural canal and between the
at articular surfaces of the postzygapophyses, framing
the postzygapophyseal canals medially. The dorsal liga-
ment would have been inserted on this ridge (Wild, 1974:
g. 38 e).
40 STEFANIA NOSOTTI
Fig. 31 – Tanystropheus longobardicus, MSNM V 3663 a, plate, skull. Scale bar 20 mm. Photo: Roberto Appiani. Drawing: Massimo
Demma.
41
TANYSTROPHEUS LONGOBARDICUS: RE-INTERPRETATIONS OF THE ANATOMY BASED ON NEW SPECIMENS FROM BESANO
Fig. 32 – Tanystropheus longobardicus, MSNM V 3663 b, counterplate. Scale bar 10 mm. Photo: Luciano Spezia.
Fig. 33 – Tanystropheus cf. longobardicus, MSNM BES 215, mid-dorsal
vertebra in posterior view. Scale bar 10 mm. Photo: Luciano Spezia.
Specimen MSNM BES 351
(Fig. 34)
Specimen MSNM BES 351 comprises isolated frag-
ments of the shafts of cervical ribs belonging to an
individual larger in size than MSNM BES SC 265 and
MSNM BES SC 1018. This can be inferred from the
diameter of the shafts which is approximately 2 mm in
MSNM BES 351, and 0.5-1.0 mm in MSNM BES SC
265 and MSNM BES SC 1018. The diameters of the cer-
vical ribs in specimen PIMUZ T 2819 (estimated overall
Table 7 – Tanystropheus cf. longobardicus, measurements
of MSNM V 3730.
Length (mm)
Metatarsal I 28.0
Metatarsal II 36.0
Metatarsal III 38.0
Metatarsal IV 37.8
Metatarsal V 12.2
Digit I
Phalanx I 12.8
Ungual 6.4
Digit II
Phalanx I 14.3
Phalanx II 8.8
Ungual 6.4
Digit III
Phalanx I 22.6
Phalanx II 16.9
Phalanx III 9.1
Ungual 6.2
Digit IV
Phalanx I 20.6
Phalanx II 10.0
Phalanx III 6.2
Phalanx IV 4.0
Ungual 4.0
Digit V
Phalanx I 38.4
Phalanx II 12.7
Phalanx III 4.9
Ungual 4.4
42 STEFANIA NOSOTTI
and distal tarsal three is far smaller than distal tarsal four.
Distal tarsal three lies medial to the distal head of the
astragalus and is in close juxtaposition with the proximal
end of metatarsal III. Distal tarsal four lies in the lateral
embayment of the astragalus and contacts the proximal
head of metatarsal IV distally. It contacts distal tarsal
three medially and articulates with the proximal rounded
head of metatarsal V laterally.
The metatarsals are tightly packed, and their proximal
ends overlap. On the proximal end of metatarsal IV there is
a distinct facet for the reception of metatarsal V. The bony
wall of metatarsals I through IV is partly collapsed. Lateral
to the contact with the calcaneum the surface of metatar-
sal V is recessed into a shallow fossa. Distally, metatarsal
V forms a rounded articular head for the rst phalanx of
digit ve. Lateral to that there is a distinct rounded proc-
ess, possibly representing a lateral plantar tubercle (p. 73).
An “outer process” with a distinct tuberosity is developed
proximo-laterally (p. 73). The lateral margin of metatarsal
V is concave and thickened.
The metatarsals and the phalanges of digits one and
two form articulated series. The ungual phalanx of digit
three is displaced, as are the three articulated distal
phalanges of digit four and the two articulated distal
phalanges of digit ve. Finally, the rst phalanx of digit
ve is displaced across the distal ends of metatarsals III
and IV, and lies at right angles to the second phalanx of
digit ve. Like T. meridensis, MSNM V 3730 was col-
lected in the Meride Limestone of Landinian age. T meri-
densis is probably conspecic with T. longobardicus (see
“Introduction”). However, pending the format revision of
taxonomy, MSNM V 3730 is provisionally considered as
Tanystropheus cf. longobardicus.
Fig. 34 – Tanystropheus longobardicus, MSNM BES 351, fragments of cervical ribs. Scale bar 30 mm. Photo: Luciano Spezia.
length 3.65 m, Wild, 1974: tab. 1) are 1.0-2.0 mm proxi-
mally and thickening to a maximum between 2.0 and 3.0
mm (Wild, 1974: 60, tab. 4).
Specimen MSNM V 3730
(Figs. 35, 61, Tab. 7)
Specimen MSNM V 3730 is an isolated pes with all
the elements preserved, and mostly in articulation. A
partial natural mould and remains of the disarticulated
epipodials are also present but are poorly preserved and
provide no information on their detailed morphology.
On the basis of the phalanges, which bear longitudinal
grooves along their shafts and prominent distal condyles
separated by a groove, this specimen most probably rep-
resents a right pes in plantar view. By comparison with
MSNM BES SC 1018, the overall size of the individual
to which the pes belonged is estimated to have been
approximately 1.50 m.
The tarsus of MSNM V 3730 is unique for all known
Tanystropheus material in preserving all the elements in
their original position and articulation. The calcaneum
is a polygonal element with roundish and thickened
lateral margin. Its concave distal margin matches the
rounded proximal head of metatarsal V. The astragalus is
a proximo-distally elongate element, obliquely oriented
within the tarsus. Its proximal end meets the calcaneum
in a straight line. Half way along it there is an elon-
gate foramen for the passage of the perforating artery
(Wild, 1974: 116). Distally, the astragalus expands into
a rounded head, that does not contact metatarsal II. The
distal ossications of the tarsus are roundish elements,
43
TANYSTROPHEUS LONGOBARDICUS: RE-INTERPRETATIONS OF THE ANATOMY BASED ON NEW SPECIMENS FROM BESANO
Fig. 35 – Tanystropheus cf. longobardicus, MSNM V 3730, right pes in plantar view. I-V = metatarsals. Scale bar 10 mm. Preparation:
Giuseppina Damiano. Photo: Luciano Spezia.
44 STEFANIA NOSOTTI
Fig. 36 – Tanystropheus longobardicus, reconstruction of the skull in the small-sized specimens. See discussion on pp. 44-62.
Drawing: Massimo Demma.
DISCUSSION
Discussion and interpretation of the skull anatomy of
Tanystropheus: a new reconstruction
In spite of the large number of specimens of T. longo-
bardicus recovered during excavations in the Besano
Formation (=Grenzbitumenzone in the Swiss geologi-
cal literature) the strong compression of these fossils
together with the frequent disarticulation of various
elements have been limiting factors in our interpreta-
tion. This is especially true for the skull. Skulls of T.
longobardicus in the PIMUZ Collections that are more
or less complete are those of the large-sized specimens
PIMUZ T 2790 (Wild, 1974: pl. 14) and PIMUZ T
2819 (Wild, 1974: gs. 5-6, pls. 15-16), and that of the
small-sized specimen PIMUZ T 2791 (Wild, 1974: gs.
3-4; Fig. 38 in this paper). The latter was rst briey
described by Peyer (1931). Wild (1974) gave a more
complete description, mainly based on X-radiographic
plates (Wild, 1974: g. 3). However, this specimen is
severely crushed, and its interpretation very difcult.
45
TANYSTROPHEUS LONGOBARDICUS: RE-INTERPRETATIONS OF THE ANATOMY BASED ON NEW SPECIMENS FROM BESANO
Fig. 37 – Tanystropheus longobardicus, three-dimensional reconstruction of the skull in the small-sized specimens. See discussion on
pp. 44-62. Watercolor: Massimo Demma.
Another quite complete small-sized skull, PIMUZ T 3901
(Fig. 39), is somewhat better preserved. This specimen,
however, was referred by Wild (1980a) to a separate spe-
cies, T. meridensis (see “Introduction” for discussion).
The new specimens from Besano described here, par-
ticularly the remarkably well preserved MSNM BES SC
1018, offer valuable new evidence on the morphology of
the skull in Tanystropheus.
In providing this new reconstruction of the skull (Figs.
36-37), three points should be considered.
Firstly, the reconstruction of the skull of Tanystro-
pheus presented here is only applicable to small-sized
representatives of T. longobardicus (or, alternatively, to
the representatives of a small-sized species of Tanystroph-
eus, see p. 8) and also to T. meridensis. Differences in the
shape and/or size of skull elements between small- and
large-sized PIMUZ specimens were observed but are not
discussed in any detail.
Secondly, the ventral and occipital views of the skull
were not reconstructed in detail. The elements of the
dermal palate and of the neurocranium are too poorly pre-
served and/or exposed in the new specimens, so they add
nothing further to the previous descriptions. The extent of
the dermal palate and the occiput in the model is consist-
ent with the reconstruction of Wild (1974: g. 8).
Thirdly, in the words of Underwood (1970: 2): when
looking at the reconstruction, “I hope that readers will
accept the more extended ights of speculation as intended
to provoke discussion rather than to state positions rmly
held”.
The drawings in Figs. 36 and 37 were developed from
a clay model, that was used to test the proportions and
the nature of the contacts between adjacent elements of
the skull in the three dimensions. The nal reconstruction
resulted from a complex procedure, in which different
actions were performed simultaneously, each affecting
each other, so that slight adjustments were continu-
ously necessary. As a rst step, an unnished model was
made approaching the general proportions of the skull in
MSNM BES SC 1018 and in PIMUZ T 2484, as recon-
structed by Wild (1974: gs. 7 a, 8). Indeed, PIMUZ T
2484 shows very similar proportions of the skull bones
to those of MSNM BES SC 1018. Precise paper tem-
plates of the bones were then made from MSNM BES
SC 1018 (Figs. 10-13) and PIMUZ T 2484 (Figs. 40, 43,
48, 50-52), and applied to the rough model, which was
consequently modied. The three-dimensional propor-
tions of the individual skull elements and their respec-
tive contacts were nally adjusted working directly on
the model, paying attention to how each change affected
the overall conguration of the skull. The squamosal was
reconstructed based on MSNM BES SC 265 (Figs. 2-3)
and PIMUZ T 2791 (Fig. 38).
Premaxilla and Maxilla
The premaxilla and maxilla of MSNM BES SC 1018
are very similar in proportions and shape to the same
elements in PIMUZ T 2484 (Wild, 1974: gs. 2, 82, 83,
pl. 18; Fig. 43 in this paper), in PIMUZ T 2791 (Wild,
1974: g. 4; Fig. 38 in this paper) and in T. meridensis
46 STEFANIA NOSOTTI
(Wild, 1980a: g. 1; Fig. 39 in this paper). There are
some distinctive features common to all these speci-
mens but not present in the large-sized individuals. The
premaxilla forms a long and slender maxillary process,
and a short, yet distinct nasal process. The maxilla is
triangular in shape anteriorly - without a premaxillary
process (sensu Wild, 1974: 9, g. 82) - and forms a very
well developed dorsal process.
The nature of the sutural contact between the premax-
illa and the maxilla is quite clear, and invariably the same
in all the skulls of the small-sized specimens. As can be
seen in PIMUZ T 2484 (Fig. 43), the antero-lateral sur-
face of the maxilla has a deep furrow, matching the maxil-
lary process of the premaxilla. T. meridensis (Fig. 39), in
which the contact between the premaxilla and maxilla can
be seen both laterally (left side of the skull) and medially
(right side of the skull), conrms that the maxillary proc-
ess of the premaxilla overlaps the maxilla laterally, tting
into the maxillary furrow. According to Wild (1974: 9; see
also reconstructions in gs. 7 a, 8, 10 a) the dorsal proc-
ess of the maxilla were overlapped by the nasal and the
prefrontal. However, the nature of the contact between
the three bones remains not completely clear, and cannot
be seen in either MSNM BES SC 1018 or in any PIMUZ
specimen. My hypothesis is that the dorsal process of the
maxilla was not concealed by the nasal and prefrontal (see
“Nasal” and “Prefrontal”).
Nasal
In the majority of known T. longobardicus speci-
mens the nasals are not preserved. A fragmentary, iso-
lated nasal is preserved in the large specimen PIMUZ
Fig. 38 – Tanystropheus longobardicus, PIMUZ T 2791, skull. Scale bar 10 mm. Photo: Heinz Lanz, PIMUZ.
T 2818 (Wild, 1974: pl. 12), and both nasals are com-
pletely preserved, albeit isolated, in the small PIMUZ
T 2484 (Fig. 40). Finally, in T. meridensis the nasals
are preserved in situ but severely crushed (Fig. 39).
According to Wild (1974: 9), a loose junction of the
nasals with the frontals would account for the nasals
being invariably present as isolated elements (or com-
pletely missing). It seems likely that the nasals had
very loose contacts with all the neighbouring bones.
Consequently, the nature of these contacts is very dif-
cult to reconstruct.
According to Wild’s hypothesis (1974: g. 8) the
nasals overlapped the dorsal process of the maxillae, and
were overlapped by the prefrontals and the frontals. In
the known Tanystropheus specimens, however, there is
no evidence for the nasal overlapping the dorsal process
of the maxilla. The latter is fully exposed both in PIMUZ
T 2791 (Fig. 38) and in MSNM BES SC 1018. In these
specimens no nasal is preserved. However, I assume that
the nasals in T. meridensis are preserved in situ (Fig.
39), crushed medially against the dorsal process of the
maxilla which is also fully exposed (contra Wild, 1980a:
5, I clearly identied a well developed dorsal process of
the maxilla in T. meridensis). This condition of preserva-
tion suggests that the nasal was positioned medial to the
dorsal process of the maxilla, rather than overlapping it.
It is possible that the lateral margin of the nasal and the
dorsal margin of the maxilla established a syndesmotic,
unossied contact. It is difcult to ascertain whether the
nasals were overlapped by the prefrontals and the fron-
tals, but I assume that such was the case (see “Frontal”
and “Prefrontal”).
47
TANYSTROPHEUS LONGOBARDICUS: RE-INTERPRETATIONS OF THE ANATOMY BASED ON NEW SPECIMENS FROM BESANO
Fig. 39 – Tanystropheus meridensis, PIMUZ T 3901, skull. Scale bar 10 mm. Photo: Heinz Lanz, PIMUZ.
48 STEFANIA NOSOTTI
Fig. 40 – Tanystropheus longobardicus, PIMUZ T 2484, nasals. Scale
bar 15 mm. Photo: Heinz Lanz, PIMUZ.
The general shape and proportions of the nasals
were drawn based on those of PIMUZ T 2484 (Fig. 40)
but slightly modied to t them into the model and to
obtain elongate external nares. Elongation of the exter-
nal nares was inferred from T. meridensis (Fig. 39), in
which the entire length of the maxillary process of the
premaxilla forms the latero-ventral margin of the exter-
nal naris.
Frontal
In MSNM BES SC 1018 the frontals are paired bones,
as they are in the small-sized PIMUZ specimens (Wild,
1974: 10), in which the suture between the two elements
is partly identiable.
In all specimens of T. longobardicus the frontals
display large lateral axe-shaped anges (die Orbital-
lamellae or orbital laminae sensu Wild, 1974: 10).
Wild (1974: 10) stated that these anges were attened
because of compression but were originally slanting
latero-ventrally, (Wild, 1974: gs. 7 a, 8). By contrast,
I think that the condition of preservation reects the
original shape of the frontals. In all the PIMUZ speci-
mens, isolated frontals are invariably preserved with
horizontal lateral anges (Figs. 41-43). In the complete
skulls, preserved in lateral or dorso-lateral view, there
is no evidence for a vertical orientation of the anges.
Fig. 41 – Tanystropheus longobardicus, PIMUZ T 2482, fronto-parietal plate in ventral view. The arrow points to the fronto-parietal
suture. Scale bar 6 mm. Photo: Nicholas C. Fraser.
49
TANYSTROPHEUS LONGOBARDICUS: RE-INTERPRETATIONS OF THE ANATOMY BASED ON NEW SPECIMENS FROM BESANO
Fig. 42 – Tanystropheus longobardicus, PIMUZ T 2787, fronto-parietal plate in ventral view. Scale bar 10 mm. Photo: Nicholas
C. Fraser.
In T. meridensis (Fig. 39) the two frontals rotated as a
single unit and they clearly lie on one plane, with the
dorsal surface exposed. The same is probably true for
PIMUZ T 2791 (Wild, 1974: gs. 3-4; Fig. 38 in this
paper), although this specimen is difcult to interpret
because of extensive crushing. In MSNM BES SC 1018
(Figs. 10-12) there is no evidence that the lateral anges
were oriented latero-ventrally. The left frontal was
raised upwards, exposing the entire ventral surface of
the lateral ange, rather than extending into the orbit as
it would have been expected if the lateral ange formed
a medial wall to it. In MSNM BES SC 265 (Figs. 2-3),
which is preserved in ventral view, the lateral anges of
the frontals are seen as bony laminae roong the orbits.
The horizontal orientation of the lateral anges of the
frontals in the model (Figs. 36-37) is consistent with the
general proportions of the skull.
The frontal plate of MSNM BES SC 1018 was
reconstructed duplicating the left frontal (exposed in
ventral view). As a result, the frontal plate in ventral
view (Fig. 44) closely resembles that of PIMUZ T 2787
(Fig. 42). Paired sinuous ridges characterize the ventral
surface of the frontals (Figs. 41-42, 44). These ridges
prosecute posteriorly on the ventral surface of the pari-
etals (see “Parietal”). Contra Wild (1974: 10). I do not
consider these ridges to mirror structures on the dorsal
surface of the frontal, in particular the supraorbital
ridges or die Supraorbitalkanten sensu Wild (1974:
10, g. 2). According to Wild, the supraorbital ridges
on the dorsal surface of the frontals formed the dorsal
margin of the orbits and of the vertical orbital lami-
nae and they were impressed upon the ventral surface
through compression. It is not clear what constitutes
the cristae cranii frontalis considered by Wild (1974:
10) a feature of the ventral surface of the frontals.
Unequivocal evidence that the sinuous ridges repre-
sent real features of the ventral surface of the frontal
plate is provided by the correspondence of the skeletal
anatomy to that of the soft parts. By comparison with
casts of the endocranial cavity in extinct and extant
reptiles (Hopson, 1979), it is clear that the shape of
the sinuous ridges on the frontal plate corresponds to
the outline of the forebrain (see also Wild, 1974: 10).
This is further conrmed by the presence on the ventral
surface of the anterior process of the frontal plate (see
below) of the oval impressions of the olfactory bulbs
50 STEFANIA NOSOTTI
Fig. 43 – Tanystropheus longobardicus, PIMUZ T 2484, fronto-parietal plate in dorsal view. The arrows point to articular facets, see
discussion on p. 51. Scale bar 10 mm. Photo: Nicholas C. Fraser.
(Figs. 41-42). Hopson (1979: gs. 1-2) compared
endocasts made from the skull of small and medium-
sized specimens of the extant Caiman crocodilus and
observed that changes in endocast proportions reect
changes in forebrain shape during ontogeny. In particu-
lar, in smaller individuals the cerebral cast is propor-
tionally wider and in larger individuals it is narrower.
Interestingly, the shape of the sinuous ridges on the
ventral surface of the frontal plate in the small PIMUZ
T 2482 is slightly different from the larger PIMUZ
T 2787. This might be due to individual variation or
alternatively be related to changes in forebrain shape.
In the frontal plate of T. meridensis (Fig. 39), exposed
in dorsal view, Wild (1980a) described a thick supraor-
bital ridge continuous with the orbital margin of the pre-
frontal. I interpret this ridge as the medial contact between
the two frontals. The medial relief on the frontal plate of
T. meridensis is at least partly due to compression of the
two frontals against the other. However, the right frontal
of MSNM BES SC 1018, with a strongly thickened and
sculptured medial margin, suggests that there was a raised
area where the two frontals met.
A tripartite anterior process (nasal process sensu Wild,
1974: 10) formed by the two frontals such as seen in
PIMUZ T 2484 (Fig. 43) is apparently not conspicuous in
MSNM BES SC 1018. The shape of this process results
from the contact of the anterior ends of the frontals. As
described in MSNM BES SC 1018 (p. 21, Fig. 12), the
anterior end of the main body of the frontal is bifurcate,
forming two lobes. In this specimen the lateral lobe is
seen lateral (dorsal, as preserved) and the medial lobe
is seen medial (ventral, as preserved) to the interlocking
prefrontal. The medial lobe in MSNM BES SC 1018 is
apparently very short, so that if the anterior end of the
left frontal is duplicated, a very short anterior process
of the frontal plate results. I emphasize, however, that
an elongate triangular fragment of bone is preserved in
MSNM BES SC 1018 anterior to the medial lobe (Fig.
11), and this might be interpreted as an anterior continu-
ation of the lobe itself, which as a consequence would be
longer. Considering also that a tripartite, well developed
anterior process of the frontal plate is invariably present
in all PIMUZ specimens (Figs. 41-43), I reconstructed it
in the model.
51
TANYSTROPHEUS LONGOBARDICUS: RE-INTERPRETATIONS OF THE ANATOMY BASED ON NEW SPECIMENS FROM BESANO
Fig. 44 – Tanystropheus longobardicus, reconstruction of the fronto-
parietal plate in ventral view based on MSNM BES SC 1018. Water-
color: Massimo Demma.
The dorsal surface of the anterior process can be seen
in PIMUZ T 2484 (Fig. 43). It shows paired articular
facets on each side, one between the medial and lateral
lobes, the other in the notch between the anterior process
and the lateral ange of the frontal. Wild (1974: 10, g.
2) thought that the fronto-parietal plate of PIMUZ T 2484
was preserved in ventral view, and related the presence
of the facets on the anterior process of the frontal plate
to a contact with the nasals and prefrontals. According to
Wild (1974: 9-10), both these elements were overlapped
by the frontals. He assumed that the anterior process of
the frontal plate “t[ted] in between the nasals on the
dorsal side”, and that it was almost completely covered
by the prefrontals on the ventral side. However, Wild did
not state precisely which elements were received in the
articular facets on the anterior process, and concluded that
the nature of the contacts between the frontals, nasals and
prefrontals was very complex.
The medial articular facets on the dorsal side of the
anterior process of the frontal plate in PIMUZ T 2484
might be for the receptions of the nasals. However, the
medial margin of the nasals of PIMUZ T 2484 (Fig. 40)
is not notched posteriorly. This indicates that if the articu-
lated nasals had overlapped the anterior process of the
frontal plate they would have completely concealed it. By
contrast, evidence from PIMUZ T 2484 is that the median
part of the anterior process between the two articular
facets was exposed on the dorsal side of the skull roof.
Consequently, I assume that the nasals were overlapped
by the anterior process of the frontal plate. However, the
nature of the contact between the nasals and the frontals
remains unclear.
In MSNM BES SC 1018, the prefrontal and the fron-
tal are still fully articulated (p. 24) in an interlocking
suture, exposed in ventral view. The frontal process of
the prefrontal is received into a notch between the medial
and lateral lobes of the anterior process of the frontal. In
dorsal view, this notch corresponds to the margin of the
articular facet between the two lobes. I assume that the
frontal process of the prefrontal was directed medially
and contacted the anterior process of the frontal in an
oblique suture. Dorsally, the prefrontal overlapped the
medial facet of the anterior process, ventrally, the margin
of the prefrontal met the margin of the medial facet.
The lobe of the prefrontal (p. 24, Figs. 12, 37) appar-
ently does not reach the facet on the notch between
the anterior process and the lateral ange of the fron-
tal. Moreover, the margin of the lobe is thickened and
rounded, suggesting that it was probably free. At the same
time, if the lobe of the prefrontal did not reach the lateral
facet on the anterior process of the frontal plate, the latter
facet remains unexplained.
Based on the well preserved right parietal of MSNM
BES SC 1018, the frontals tted posteriorly in a V-shaped
embayment formed by the parietals (Fig. 44). As a result,
the parietal foramen is partly delimited by the frontals. A
similar morphology is seen in PIMUZ T 2482 (Fig. 41). By
contrast, in PIMUZ T 2484 (Fig. 43), preserved in dorsal
view, the frontals contact the parietals in an unequivocal
interdigitating suture which is positioned anterior to the
parietal foramen. These differences in the fronto-parietal
contact are probably due to individual variation.
The posterior contact of the frontals with the postfron-
tals is discussed under the heading “Postfrontal”.
Parietal
The parietals of MSNM BES SC 1018 and MSNM
BES SC 265 conform to Wild’s statement (1974: 11)
that the parietal plate is formed by two separate ele-
ments. They are completely fused in the large-sized
specimens in the PIMUZ Collections, while a suture
between the two elements is still partly identifiable in
the small-sized specimens.
The parietals were reconstructed in ventral view
(Fig. 44), largely based on MSNM BES SC 1018
(right parietal, Figs. 10-12). The reconstructed pari-
etals are very similar to those of PIMUZ T 2482 (Fig.
41; see also Wild, 1974: fig. 2, where, contra Wild,
the parietals drawn in ventral view are clearly those of
PIMUZ T 2482, not of PIMUZ T 2484). The assump-
tion that the parietals of PIMUZ T 2482 are preserved
in ventral view is consistent with the shape of the
frontals in the same view, as interpreted above. As for
the frontals, I consider that the condition of preserva-
tion of the parietals reflects their original morphol-
ogy, and that the relief on the ventral surface of the
parietals is not a direct result of compression causing
the relief on the dorsal surface to be impressed upon
it. The morphology of the ventral surface of the pari-
etals conceivably mirrors that of the soft parts in the
cerebellar region of the brain (see also “Frontal” and
Wild, 1974: 11).
52 STEFANIA NOSOTTI
In the model, the contact between the parietals and
the frontals was reconstructed based on MSNM BES SC
1018 (see “Frontal”), while that between the parietals
and the postfrontals (see “Postfrontal”) was inferred
from PIMUZ T 2484 (Fig. 43). The lateral supratempo-
ral processes were reconstructed on the basis of MSNM
BES SC 265 and the PIMUZ specimens. The supratem-
poral processes overlap the squamosals and possibly
the supratemporals (see “Supratemporal”), as can be
inferred from the presence of a facet on their ventral
surface in MSNM BES SC 265 (Figs. 2-3) and PIMUZ
T 2482 (Fig. 41), and from the articulated squamosal in
the large PIMUZ T 2819 (Wild, 1974, gs. 5-6). Finally,
in the parietals in dorsal view I assumed the presence
of lateral obliquely slanting anges (die Temporalügel
or parietal laminae sensu Wild, 1974: 11) but there is
no clear evidence for them either in the new specimens
or in the small-sized PIMUZ specimens. They are only
unequivocally present in the large PIMUZ T 2819 (Wild,
1974: gs. 5-6).
The ratio between the length of the parietals relative
to the overall length of the fronto-parietal plate result-
ing from the reconstruction in MSNM BES SC 1018 is
congruent with that observed in PIMUZ T 2484 (Fig.
43). In the larger PIMUZ T 2482 (Fig. 41) and PIMUZ T
2787 (Fig. 42) the ratio is progressively higher. The rela-
tive length of the frontal and the parietal changes, with
the parietal increasing and the frontal decreasing with
increasing overall size.
Supratemporal
Wild (1974: 12) claimed the presence of supratempo-
rals in T. longobardicus on the basis of an element inter-
preted as a supratemporal preserved in situ in the large-
sized PIMUZ T 2819. Paired rod-shaped small bones
might be interpreted as supratemporals in MSNM BES
SC 265 (p. 12, Figs. 2-3). For the present the occurrence
of supratemporals in T. longobardicus is equivocal and
it is therefore impossible to describe the nature of their
contacts with the neighbouring bones.
Prefrontal
The prefrontal of MSNM BES SC 1018 is very similar
to the prefrontal of T. meridensis (Fig. 39) but different
from the prefrontals as described and gured by Wild
(1974: 10, gs. 8-9) in the specimens of T. longobardicus
in the PIMUZ Collections.
In Wild’s reconstructions (1974: gs. 8-9), the pre-
frontals extend in a antero-posterior direction and only
form the antero-dorsal margin of the orbits. Further ven-
trally, the anterior margins of the orbits are formed by the
lacrimals. The prefrontals overlap both the dorsal process
of the maxillae and the nasals with a process (anterior
process) extending anteriorly far beyond the fronto-
nasal contact. Medially, they contact the frontal (frontal
process), extending posteriorly beyond the fronto-nasal
contact. The “orbital lamina” of the prefrontal is con-
tinuous with the “orbital lamina” of the frontal. Wild also
maintained that the prefrontals reached the palatine with
a palatine process.
Among the small-sized PIMUZ specimens, Wild
(1974: 10) identied a prefrontal preserved in situ in
PIMUZ T 2485 and an isolated prefrontal in PIMUZ T
2795. It is on the latter that he based the drawing in g. 2
and the reconstruction in g. 8. However, in Wild’s draw-
ing of PIMUZ T 2485 (Wild, 1974: pl. 8) a prefrontal is
not identiable and in the radiograph of PIMUZ T 2795
(Wild, 1974: pl. 6) it is difcult to understand which ele-
ment Wild was referring to as the prefrontal.
In MSNM BES SC 1018 (Figs. 10-12) and in T.
meridensis (Fig. 39) the prefrontal mostly extends in
a dorso-ventral direction and forms almost the entire
anterior margin of the orbit. Based on MSNM BES
SC 265 (p. 12, Figs. 2-3), I conrm that the prefrontal
reached the palatine ventrally. In MSNM BES SC 1018
and in T. meridensis the prefrontal is located posterior
to the dorsal process of the maxilla, and not overlapping
it. Based on the interpretation of the contact between
the frontal and the prefrontal (see “Frontal”), the latter
element had to overlap the nasal in order to reach the
anterior process of the frontal plate. The reconstruction
presented here differs from Wild’s (1974: gs. 8-9) in
that the prefrontal does not extend far anteriorly beyond
the fronto-nasal contact. Finally, the prefrontal is recon-
structed in contact with the anterior process of the fron-
tal plate only, positioned mainly anterior to the latter,
not lateral to it. In the new reconstruction the “orbital
laminae” of the frontal are horizontally, not vertically
positioned as was assumed by Wild. Consequently, the
orbital surface of the prefrontal contacts the ventral sur-
face of the frontal more medially, and there is no conti-
nuity between an “orbital lamina” of the prefrontal and
an “orbital lamina” of the frontal.
Among the large-sized PIMUZ specimens, Wild
(1974: 10) identied the prefrontals in PIMUZ T 2787.
According to Wild, the right prefrontal is isolated but
complete (Wild, 1974: g. 1), and it is very similar in
shape to the prefrontal of the small-sized specimens
(Wild, 1974: g. 2). Wild also maintained that the left
prefrontal of PIMUZ T 2787 (Fig. 42) is preserved in
situ, and that it still contacts the maxilla and the frontal.
However, he stated that it was not possible to unequivo-
cally identify the margins of the three bones. Indeed, I
could not identify any contact between the maxilla and
the presumed prefrontal. Wild assumed that the prefron-
tal widely overlapped the ventral surface of the anterior
process of the frontal but the presumed prefrontal of
PIMUZ T 2787 does not. An alternative interpretation
might be that the “prefrontal” in PIMUZ T 2787 is in fact
a nasal but also in this case the nature of the contact with
the frontal is not clear.
Lacrimal
A lacrimal was only reported by Wild (1974: 10, gs.
1, 2) for PIMUZ T 2795 and PIMUZ T 2787, and in both
specimens it is an isolated bone. The lacrimal of PIMUZ
T 2795 was gured by Wild (1974: g. 2) as a triangular
bone with no surface detail. The shape of the lacrimal
of PIMUZ T 2787 as gured by Wild (1974: g. 1) is
comparable in shape to that of MSNM BES SC 1018,
although the lacrimal foramen is more posteriorly located
in PIMUZ T 2787. The lacrimal of MSNM BES SC
1018 documents for the rst time a lacrimal of T. longo-
bardicus preserved in situ, conrming Wild’s reconstruc-
tion (1974: g. 8 b). A lacrimal of approximately the same
shape, size and position as that in MSNM BES SC 1018
was identied by Wild (1980a: g. 1) in the skull of T.
meridensis (Fig. 39).
53
TANYSTROPHEUS LONGOBARDICUS: RE-INTERPRETATIONS OF THE ANATOMY BASED ON NEW SPECIMENS FROM BESANO
Postfrontal
The new specimens provide very limited information
on the morphology of the postfrontal, and no postfron-
tals are preserved in the majority of PIMUZ specimens.
However, in PIMUZ T 2484 postfrontals are preserved
still sutured to the frontals (Fig. 43; see also Wild, 1974:
g. 2). This specimen best claries the position, shape
and contacts of the postfrontals in dorsal view. The post-
frontals are subtriangular elements anteriorly contacting
the horizontal anges of the frontals, and medially the
parietals (Fig. 43). The exact same conguration can be
seen on the radiograph of the slightly larger PIMUZ T
2482 (Fig. 45). An articulated postfrontal is probably also
preserved in the small PIMUZ T 2791 (Wild, 1974: 18,
gs. 3-4; Fig. 38 in this paper) but this skull is severely
crushed and the information it provides is limited. Finally,
a postfrontal is probably preserved in T. meridensis (Wild,
1980a; Fig. 39 in this paper).
Fig. 45 – Tanystropheus longobardicus, PIMUZ T 2482, fronto-pari-
etal plate, radiograph. The arrow points to the postfrontal. Courtesy
PIMUZ.
In the large PIMUZ T 2819 (Wild, 1974: gs. 5-6), the
left postfrontal appears to be at least partially preserved
but it provides little detail. Wild (1974) claimed the pres-
ence of a postfrontal in PIMUZ T 2787 (Fig. 42), but
there is no evidence of a suture between the frontal and
the postfrontal as shown in Wild’s g. 1. Evidence from
other specimens indicates that the presumed postfrontals
in PIMUZ T 2787 represent the posterior part of the hori-
zontal anges of the frontals. In the large MSNM V 3663
(p. 38, Fig. 31) the right postfrontal is preserved, sutured
to the skull roof in a manner similar to PIMUZ T 2819.
The postfrontal is therefore a very short element pro-
jecting laterally and ventrally, and sutured to the frontal
and parietal. Its shape is similar in small- and large-sized
specimens. However, the nature of its contact with the
postorbital is not completely clear, due to the difculties
encountered in the interpretation of the temporal region of
the skull (see below).
Squamosal and Postorbital
Wild (1974: 11) described the squamosal as a sickle-
shaped element with a wide anterior postorbital process
overlapping the postorbital, a pointed posterior parietal
or supratemporal process overlapped by the supratem-
poral process of the parietal, and a quadrate process
overlapping the cephalic condyle of the quadrate. Wild’s
description was evidently based on the small PIMUZ T
2791 (Wild, 1974: gs. 2-4; Fig. 38 in this paper), the
larger PIMUZ T 2787 (Wild, 1974: g. 1, pl. 9; Fig. 42
in this paper) and the large PIMUZ T 2819 (Wild, 1974:
gs. 5-6).
The general shape of the left squamosal in MSNM
BES SC 265 (p. 12) is very similar to that of the squa-
mosal in Wild’s (1974) g. 2 (this element is not from
PIMUZ T 2484, as stated by Wild but clearly corresponds
to the isolated right squamosal of PIMUZ T 2791, see Fig.
38 in this paper). I suggest that this shape represents real
features of the bone, and that the squamosal originally
had a complex three-dimensional shape (Fig. 46) which
cannot be seen in the fossils because the bone is crushed
on the slab. The orientation of the posterior and anterior
rami of the squamosal - and of its postorbital process - in
the three dimensions shown in Fig. 46 might well pro-
duce the shape of the compressed squamosal described
for MSNM BES SC 265 (p. 12). As the left squamosal of
MSNM BES SC 265 is preserved in its original antero-
posterior orientation, there is unequivocal evidence for
the pointed process of the squamosal being the postorbital
process (contra Wild, 1974: 11, g. 2).
Wild (1974: 11, g. 2) interpreted the postorbital as an
elongate, subtriangular and attened bone. Its long and
pointed ventral jugal process overlapped the jugal, and
the expanded dorsal end formed an anterior supraorbital
process (orbital process sensu Wild, 1974), overlap-
ping the postfrontal, and a posterior squamosal process,
interlocking with the squamosal. Among the small-sized
specimens, Wild identied isolated postorbitals in
PIMUZ T 2484 (Wild, 1974) and T. meridensis (Wild,
1980a). While the element interpreted by Wild (1980a:
g. 1; Fig. 39 in this paper) as a postorbital in T. meriden-
sis is probably a quadrate, it is not clear which element of
PIMUZ T 2484 was identied by Wild as the postorbital.
In the caption of g. 1 Wild (1974) stated that the isolated
postorbital illustrated was from PIMUZ T 2484 but he
did not indicate which bone he identied as the postor-
bital in the “exploded” skeleton (Wild, 1974: pl. 18). It is
possible that it is the element preserved to the right side
of the fronto-parietal plate (labelled “po?” in Fig. 43),
which is here considered as the right postorbital. In the
same specimen there are remains of a bone overlapped
by the left prefrontal which conceivably represent the left
postorbital (see Fig. 43 in this paper and Wild, 1974: pl.
18, where the remains are labelled “postfrontal”). In the
large-sized specimens in the PIMUZ Collections, Wild
reported a postorbital for PIMUZ T 2819 (Wild, 1974:
g. 6). This fragmentary element, however, is in all prob-
ability a squamosal (Kuhn-Schnyder, 1967: g. 3). Wild
(1974) probably based the description of the postorbital
54 STEFANIA NOSOTTI
Fig. 46 – Tanystropheus longobardicus, reconstruction of the left
squamosal in the small-sized specimens (see discussion on p. 53).
A) Dorsal view. B) Ventral view. Drawing: Massimo Demma.
on the larger-sized PIMUZ T 2787 (Wild, 1974: g. 1,
pl. 9), in particular on the element labelled “po?” in
Fig. 47. Two additional elements preserved in PIMUZ
T 2787 (Fig. 42) were interpreted by Wild as the squa-
mosals. The element labelled “po” in Fig. 42, however,
is similar in shape to the putative postorbital described
for MSNM BES SC 1018 (p. 24). Contra Wild (1974), I
suggest that it represents a postorbital. Its shape is differ-
ent from that of the element labelled “sq” in Fig. 42. In
this latter element a laterally projecting posterior ramus
and an antero-laterally oriented anterior ramus - with a
tapering postorbital process bending medially - can be
clearly seen. Therefore, I concur with Wild (1974) that
this second element is indeed a squamosal. The preserved
part of the right postorbital in MSNM V 3663 (Fig. 31)
is conceivably the dorsal one, extending between the
postfrontal and the squamosal, while the jugal process is
assumed to be truncated.
The contacts of the squamosal and postorbital with the
neighbouring bones are not completely clear.
Wild (1974: 11, 18) stated that both in the small-
sized PIMUZ T 2791 (Wild, 1974: gs. 3-4) and in the
large-sized PIMUZ T 2819 (Wild, 1974: gs. 5-6) there
were squamosals preserved in situ, and he maintained
that the articular relationships between the squamosal
and the neighbouring bones, as indicated in the recon-
structions presented in gs. 8 and 9 (Wild, 1974), were
derived from these specimens. Contra Wild, I suggest
that these elements only contact the parietal. Assuming
Wild correctly identied the left squamosal in PIMUZ T
2791 on the stereo-radiographs (Wild, 1974: gs. 3-4),
this element is oriented in the direction of the ascend-
ing process of the jugal, in a manner similar to the left
squamosal in MSNM BES SC 265. It does not contact
the postorbital antero-ventrally. In fact, a postorbital
as described by Wild cannot be identied in PIMUZ
T 2791. The element identied by Wild as the right
squamosal in PIMUZ T 2819 is broken anteriorly, and
again a right postorbital was not described by Wild in
this specimen.
As in all the available material of T. longobardicus
there are no articulated squamosals and postorbitals, the
nature of the contact between these two elements remains
unclear. However, the element tentatively identied as
the right postorbital in PIMUZ T 2484 (Fig. 43) exhibits
a posterior triangular and longitudinally elongate facet.
This might be a facet for the reception of the postorbital
process of the squamosal.
Wild emphasized (1974: 11) that the articular rela-
tionship between the postorbital and the postfrontal was
not clear. Assuming that the fronto-parietal plate and the
postfrontals in PIMUZ T 2484 (Fig. 43) were preserved in
ventral view, Wild conjectured that the postorbital over-
lapped the dorsal surface of the postfrontal, because on
the ventral surface of the latter a facet for the reception of
the postorbital cannot be seen. Contra Wild, I assume that
the fronto-parietal plate and the postfrontals in PIMUZ
T 2484 are preserved in dorsal view. This would suggest
that the postorbital was overlapped by the postfrontal.
However, the presumed left postorbital is very fragmen-
tary, and the nature of the contact between the postfrontal
and the postorbital cannot be stated precisely. Articulated
postfrontals and postorbitals are not preserved in other
PIMUZ specimens. Specimens MSNM BES SC 265 and
55
TANYSTROPHEUS LONGOBARDICUS: RE-INTERPRETATIONS OF THE ANATOMY BASED ON NEW SPECIMENS FROM BESANO
Jugal
A jugal is preserved in lateral view both in MSNM
BES SC 265 and in MSNM BES SC 1018; in the latter
specimen the bone is particularly well preserved in situ.
The comparison with the small- and large-sized
specimens in the PIMUZ Collections revealed that a
much taller dorsal or postorbital process characterizes
the small-sized specimens by comparison with large-
sized ones.
Fig. 47 – Tanystropheus longobardicus, PIMUZ T 2787, presumed postorbital (see discussion on p. 54). Scale bar 20 mm. Photo:
Nicholas C. Fraser.
MSNM BES SC 1018 add no information. Apparently,
only MSNM V 3663 preserves a postfrontal in articu-
lation with the postorbital (p. 38). According to Wild
(1974: 11), a loose junction of the postorbital with the
neighbouring elements would account for the postorbital
being mostly preserved as an isolated element. In MSNM
V 3663, however, the postfrontal and the postorbital seem
to be steadily sutured (Fig. 31).
No articulated postorbitals and jugals are preserved in
the PIMUZ material and the conguration of the contact
between the two elements was here inferred from MSNM
BES SC 1018 (p. 24).
The result of my interpretation of the temporal region
of the skull is shown in Figs. 36-37. The posterior ramus
of the squamosal and the posterior part of its anterior
ramus lie on a plane slanting downwards. The anterior
part of the anterior ramus of the squamosal (postorbital
process) is deected ventrally and medially, where it con-
tacts the posterior side of the postorbital. The postorbital
is represented as a triradiate element. Dorsally and ante-
riorly it contacts the postfrontal. Ventrally, the bid jugal
process of the postorbital contacts the jugal. The anterior
ramus of the jugal process ts into the elongate fossa on
the dorsal process of the jugal.
Comparison with the skull of the closely related Mac-
rocnemus (Rieppel & Gronowsky, 1981: g. 1; Premru,
1991: g. 3; Renesto & Avanzini, 2002: g. 2 A) and the
more distantly related Prolacerta (Modesto & Sues, 2004:
g. 3) seems to lend further support for the presence of
a triradiate postorbital and a short, triangular postfrontal
in Tanystropheus. Both in Macrocnemus and Prolacerta
the two elements are in loose contact. Interestingly, in
the uncatalogued Macrocnemus PIMUZ specimen “Alla
Cascina” (Rieppel & Gronowsky, 1981: g. 1), the pos-
torbital process of the squamosal contacts a long, hori-
zontal posterior process of the postorbital along its entire
ventral margin.
56 STEFANIA NOSOTTI
A feature that was not mentioned by Wild (1974)
because it cannot be seen in the PIMUZ material is the
presence of an elongate groove close to the orbital margin
of the dorsal process. This groove is posteriorly delimited
by a crest, which itself is positioned anterior to a shallow
elongate fossa.
Quadrate
The left quadrate of MSNM BES SC 1018 is badly
crushed but apparently similar in shape to the quadrate
of T. meridensis (Wild, 1980a: g. 1; Fig. 39 in this
paper).
Wild (1980a) claimed that the shape of the quadrate
was one of the diagnostic characters of the new species
T. meridensis. According to him, the quadrate of T. meri-
densis would be taller, and more slender and concave
posteriorly than in T. longobardicus.
Among the small-sized specimens of T. longobardicus
in the PIMUZ Collections both quadrates are preserved
as isolated elements in PIMUZ T 2484 (Wild, 1974: g.
2). One of the two (left, Fig. 43), exposing its medial side
according to Wild, is badly crushed. The other (right, Fig.
40), exposing its lateral side according to Wild, provides
more detail. As compared to the quadrate of T. meriden-
sis, the quadrate of T. longobardicus is apparently shorter
and wider only because of the presence of a wide medial
lamina. If the ratio between the height of the quadrate
and the length of the maxilla is calculated, the height
of the quadrate is approximately 50% of the length of
the maxilla in T. meridensis, and approximately 40% in
PIMUZ T 2484. In Wild’s reconstruction (1974: g. 8)
based on PIMUZ T 2484, the height of the quadrate is
approximately 45% of the length of the maxilla. Fitting
the lower jaw into the model it became apparent that a
height of the quadrate of approximately 47% of the length
of the maxilla was required for the quadrate to bridge the
distance between the squamosal and the articular fossa
of the lower jaw (assuming that the lower jaw was not
distorted). Signicantly, the height of the quadrate so
obtained falls within the range of 40-50% of the length
of the maxilla observed in the PIMUZ specimens. As the
left quadrate of MSNM BES SC 1018 is badly crushed
and fragmentary, it is difcult to evaluate the presence of
a medial lamina and its extent. The medial lamina of the
quadrate in T. meridensis is apparently very short. How-
ever, part of the medial lamina might still be enclosed in
the matrix in both these specimens. Finally, the quadrate
of PIMUZ T 2791 (Wild, 1974: gs. 3-4; Fig. 38 in this
paper), which is preserved in situ and exposed in lateral
view, exhibits the same concave posterior margin as the
quadrate of T. meridensis.
In conclusion, the quadrate of T. longobardicus prob-
ably had a shape similar to that described by Wild (1974)
for the PIMUZ specimens, with an even wider medial
lamina in the large-sized specimens compared to the
small-sized ones. There is no appreciable differences in
the shape and height of the quadrate in T. longobardicus
and T. meridensis.
According to Wild (1974: 12, 25) the quadrate of
Tanystropheus was streptostylyc. It was loosely articu-
lated with the squamosal and, probably, the supratempo-
ral, and had ligamentous connections with the opisthotic
and the pterygoid. This interpretation cannot be veried
in the new specimens.
Epipterygoid
The epipterygoid is not preserved in MSNM BES SC
1018. According to Wild (1974: 14, g. 1), it is preserved
only in the larger specimens among the material housed
at the PIMUZ. An isolated epipterygoid is preserved in
MSNM BES SC 265 (p. 12, Figs. 2-3), and, as stated by
Wild, it is a rod-shaped element with expanded ends. As
the contacts of this bone with other elements of the skull
cannot be ascertained, it was not included in the recon-
struction.
Neurocranium
Only some elements of the neurocranium are pre-
served in the new specimens. In particular, in MSNM
BES SC 265 the basisphenoid-parasphenoid complex is
clearly identiable, while the other elements are badly
crushed, and their interpretation is very difcult (p. 12). In
MSNM BES SC 1018 I could only identify the supraoc-
cipital (p. 24).
To assess whether the extent of the occiput recon-
structed in the model was consistent with the proportions
of other regions of the skull I attempted a reconstruction
of the supraoccipital-exoccipitals-opisthotics complex
in MSNM BES SC 1018 by direct comparison of its
supraoccipital with that preserved in PIMUZ T 2484
(Fig. 43). I then extrapolated the proportions of the exoc-
cipitals and opisthotic from their dimension in PIMUZ
T 2484.
The lateral end of the reconstructed paroccipital proc-
esses of MSNM BES SC 1018 resulted very close to the
quadrate but did not quite reach it. Wild (1974: g. 8)
came to the same conclusion. However, my reasoning
was based on a paper template of the supraoccipital-exoc-
cipitals-opisthotics complex, which does not take into
account the thickness of the bone. Therefore a contact
of the paroccipital process with the quadrate cannot be
excluded.
Sclerotic ring
Wild (1974: 17) drew attention to the presence in
the PIMUZ material of small, thin bony plates with a
polygonal or rounded shape, and tentatively interpreted
them as sclerotic plates. In T. meridensis (1980a: gs. 1-2)
he again identied ten thin roundish-oval bony elements
positioned within the orbit as sclerotic plates.
MSNM BES SC 1018 conrms the presence of a
sclerotic ring in Tanystropheus. This specimen is the only
one in all the available material that shows isolated plates
with a denitive shape. In particular, four plates preserved
dorsal to the skull indicate a subrectangular shape. Other
elements, mostly with quite irregular shape, are preserved
within the orbit.
The proposed reconstruction of the sclerotic ring
(Figs. 36-37) should be considered provisional. Firstly,
the possible size of an eye was determined on the size
of the reconstructed orbit, followed by an estimate of
the diameter of the cornea. While arranging the subrec-
tangular sclerotic plates around the cornea, I established
that the plates should follow the eye curvature, and that
18 plates were required to maintain the contact of the
plates orbitally (sensu Underwood, 1970). As a result,
the plates partly overlapped corneally (sensu Under-
wood, 1970). The reconstructed pattern of overlap is
arbitrary.
57
TANYSTROPHEUS LONGOBARDICUS: RE-INTERPRETATIONS OF THE ANATOMY BASED ON NEW SPECIMENS FROM BESANO
Fig. 48 – Tanystropheus longobardicus, PIMUZ T 2484, lower jaws. Scale bar 10 mm. Photo: Rosi Roth, PIMUZ.
Lower jaw
Within the small-sized T. longobardicus speci-
mens in the PIMUZ Collections, complete lower jaws
in situ are preserved in PIMUZ T 2791 (Wild, 1974:
figs. 3-4; Fig. 38 in this paper). In this specimen the
right lower jaw overlaps the left one to a large extent.
Since the skull is also heavily crushed, PIMUZ T
2791 provides very little information for the recon-
struction of the lower jaw. The left mandibular ramus
is preserved in situ, exposed in lateral view, in the
small T. meridensis (Wild, 1980a: fig. 3; Fig. 39 in
this paper). Wild (1980a: 7; Fig. 49 F in this paper)
maintained that the configuration of the different ele-
ments in the lower jaw of T. meridensis was similar to
T. longobardicus.
In other “exploded” skulls of small-sized T. longo-
bardicus specimens in the PIMUZ Collections, more or
less complete lower jaws are preserved as isolated ele-
ments. Both dentaries, post-dentary parts, and isolated
bones of the lower jaws are preserved in PIMUZ T 2484
(Figs. 48, 50-52). On the basis of this specimen Wild
(1974: 40, gs. 15-16 a) reconstructed the lower jaw in
lateral view for the small-sized T. longobardicus speci-
mens. Specimen PIMUZ T 2482 (Wild, 1974: pl. 5)
preserves the complete left lower jaw in medial view
(Wild, 1974: g. 27 b), as well as part of the dentary of
the right lower jaw in lateral view. Indeed, PIMUZ T
2482 provides the most detail of the medial surface of
the lower jaw within the PIMUZ material but it is not
exhaustively informative. The fragmentary mandibular
ramus of PIMUZ T 2779 (Peyer, 1931: g. 3 in pl. 13,
text-g. 14) is too poorly preserved and adds no sub-
stantial information. Consequently, Wild reconstructed
the lower jaw in medial view for the small-sized T.
longobardicus specimens only tentatively (Wild, 1974:
g. 16 b), based on the conguration of the bones on
the lateral side of the lower jaw and on isolated man-
dibular elements (Wild, 1974: 40, g. 15) from PIMUZ
T 2484 (coronoid and prearticular) and PIMUZ T 2795
(splenial).
Complete mandibular rami in lateral view are pre-
served in the large-sized specimens PIMUZ T 2787
(right lower jaw; Wild, 1974: pl. 9) and PIMUZ T
2819 (left lower jaw; Wild, 1974: figs. 5-6), while
in PIMUZ T 2793 (Wild, 1974: figs. 18-19) the right
lower jaw is heavily damaged, and the left one is
broken into two pieces. The lateral surface of the
lower jaw of the large-sized specimens of Tanystro-
58 STEFANIA NOSOTTI
Fig. 49 – Lower jaws in small-sized specimens of Tanystropheus in lateral view, not to scale. A) Tanystropheus longobardicus, small-
sized specimens, reconstruction of Wild (after Wild, 1974: g. 16 a). B) Tanystropheus longobardicus, reconstruction of the author,
drawing: Massimo Demma. C) Tanystropheus longobardicus, MSNM BES SC 265, left lower jaw, photo: Luciano Spezia. D) Tanys-
tropheus longobardicus, MSNM BES SC 265, right lower jaw, reected horizontally for easier comparison, photo: Luciano Spezia.
E) Tanystropheus longobardicus, MSNM BES SC 1018, left lower jaw, photo: Luciano Spezia. F) Tanystropheus meridensis, PIMUZ
T 3901, reconstruction of Wild (after Wild, 1980a: g. 3). G) Tanystropheus meridensis, PIMUZ T 3901, left lower jaw, as interpreted
by the author, photo: Heinz Lanz, PIMUZ. Pink lines: suture dentary-surangular. Green lines: presumed coronoid. Orange lines: suture
articular-prearticular. See discussion on pp. 57-61.
59
TANYSTROPHEUS LONGOBARDICUS: RE-INTERPRETATIONS OF THE ANATOMY BASED ON NEW SPECIMENS FROM BESANO
Fig. 50 – Tanystropheus longobardicus, PIMUZ T 2484, post-dentary part of the left lower jaw in lateral view. Scale bar 5 mm. Photo:
Rosi Roth, PIMUZ.
pheus was reconstructed by Wild (1974: 40, fig. 17)
on the basis of PIMUZ T 2819, the medial surface
on the basis of radiographs of the right lower jaw of
PIMUZ T 2787, and of a small fragment of the left
lower jaw of the same specimen. Wild (1974: fig. 19)
also provided a reconstruction of the lateral surface of
the lower jaw of PIMUZ T 2793 but this is similar to
that of PIMUZ T 2819.
As reconstructed by Wild, the nature of the contacts
between the elements of the lower jaw, is similar in both
small- and large-sized specimens of Tanystropheus.
A new reconstruction of the lower jaw in lateral view
is presented here, based on MSNM BES SC 1018 (Figs.
13, 49 E), MSNM BES SC 265 (Figs. 2-3, 49 C-D), and
PIMUZ T 2484 (Figs. 48-52), which is partly re-inter-
preted. The conguration of the mandibular elements on
the medial side of the lower jaw remains not completely
clear but the right lower jaw of MSNM BES SC 1018
(Fig. 53 B) provides new information on the shape of the
splenial and prearticular.
By comparison with Wild’s reconstruction (1974:
g. 16; Fig. 49 A in this paper), my reconstruction
of the lower jaw in lateral view (Fig. 49 B) differs on
three points. First, the surangular overlaps the dentary.
Second, the angular is wider. Third, the splenial con-
tributes to a larger extent to the ventral margin of the
lower jaw.
Wild (1974: 38, figs. 15-16) maintained that the
dentary overlapped the surangular, observing that
the surangular was remarkably grooved anteriorly to
establish a strong contact with the dentary. However,
on both mandibular rami of MSNM BES SC 265,
an oblique, dorsally convex suture is unequivocally
present (Fig. 49 C-D), which can be identified as the
contact between the dentary and the surangular. The
same suture occurs in MSNM BES SC 1018 (left lower
jaw, Fig. 49 E), albeit less distinct, and in T. meriden-
sis (Fig. 49 G). In both isolated dentaries of PIMUZ
T 2484 (Fig. 48) an oblique, dorsally convex line sets
off a depressed area, suggesting that the dentary was
overlapped by the surangular. In the post-dentary iso-
lated part of the left lower jaw of the same specimen
(Figs. 49, 51 B), the anterior part of the surangular
displays a deepened surface dorsal to a sinuous line.
I interpret this line to represent the suture between
the dentary and surangular on the lateral surface of
the lower jaw. I assume that, if observed anteriorly,
the surangular would be excavated by a deep groove
separating a medial and a lateral bony laminae, which
are compressed one over the other in the fossil. The
medial lamina would overlap the medial surface of the
dentary, the lateral would overlap the lateral surface of
the dentary. In other words, the dentary would fit into
the groove between the two laminae. On the medial
surface of the right lower jaw of MSNM BES SC 1018
(Fig. 53 B), however, the nature of the contact between
the dentary and the surangular is not clear. Perhaps, a
coronoid overlapped the area where the two bones met
(see below).
Joining paper templates of the dentary and of post-
dentary part of the jaw in PIMUZ T 2484, I observed
that where the two elements met, on the ventral margin
60 STEFANIA NOSOTTI
Fig. 51 – Tanystropheus longobardicus, PIMUZ T 2484, lower jaws. The best preserved dentary is shown in association with the best
preserved post-dentary part of the lower jaw. A) Right dentary in lateral view, reected horizontally. B) Post-dentary part of the left
dentary in lateral view. Photo: Rosi Roth, PIMUZ.
Fig. 52 – Tanystropheus longobardicus, PIMUZ T 2484, post-dentary
part of the right lower jaw in medio-lateral view. Scale bar 5 mm.
Photo: Rosi Roth, PIMUZ.
Fig. 53 – Lower jaw in medial view, in small-sized specimens of Tanystropheus longobardicus. A) Reconstruction of Wild
(modied after Wild, 1974: g. 16 b). B) Right mandibular ramus of MSNM BES SC 1018 as interpreted by the author, drawing:
Massimo Demma.
of the lower jaw there is a notch (Fig. 51). I assume
that this notch accommodated the splenial, as stated by
Wild (1974: g. 16). Indeed, Wild found poor evidence
for the shape of the splenial in the PIMUZ material.
Based on the right lower jaw of MSNM BES SC 1018
(Fig. 53 B), I conclude that this element extended along
the ventral margin of the lower jaw far more posteri-
orly than it was assumed by Wild (Figs. 49 A-B, 53).
Evidence from MSNM BES SC 1018 also indicates
that the isolated element close to the posterior part
of the right lower jaw in PIMUZ T 2484 (Wild, 1974:
g. 15; Figs. 48, 52 in this paper), interpreted by Wild
as a prearticular, is more probably a fragment of the
splenial.
The nature of the contact between the surangular
and the angular in lateral view cannot be inferred from
the new specimens but it is clear in PIMUZ T 2484. In
61
TANYSTROPHEUS LONGOBARDICUS: RE-INTERPRETATIONS OF THE ANATOMY BASED ON NEW SPECIMENS FROM BESANO
Dentition
As ascertained by Wild (1974), the dentigerous bones
in the small-sized specimens of Tanystropheus are the
premaxilla, the maxilla, the dentary, the pterygoid, the
palatine and the vomer. This is fully conrmed by the
new specimens, which also conrm the presence of ante-
rior conical teeth and posterior tricuspid teeth in the upper
and lower jaws.
Wild considered small-sized specimens to have ve
conical premaxillary teeth and the large-sized ones to
have six (Wild, 1974: 47, g. 80). However, on the basis
of the new specimens there is no difference in the number
of premaxillary teeth between small- and large-sized
specimens of Tanystropheus. As the interpretation of the
premaxillary dentition in MSNM BES SC 1018 is une-
quivocal, the premaxillary dentition was reconstructed on
the basis of this specimen.
The left maxilla of MSNM BES SC 1018 bears 14
teeth, all ankylosed. Specimen MSNM BES SC 265
only has 12 tooth positions in the maxilla. The number
of maxillary teeth documented by Wild in the small-
sized specimens in the PIMUZ Collections was 11-13
(Wild, 1974: 47, g. 80). As Wild counted 12 or pos-
sibly 14 maxillary teeth in the large-sized specimens,
there is again no appreciable difference in the number
of maxillary teeth in specimens of different size. While
the 14 maxillary teeth in MSNM BES SC 1018 are all
tricuspid, the anteriormost tricuspid maxillary tooth
identied in MSNM BES SC 265 is the fourth (right
maxilla). Wild considered the anteriormost maxillary
teeth to be conical in the small-sized PIMUZ specimens
(Wild, 1974: 47, g. 80). It therefore seems likely that
there is some variability but I based the reconstruction
of the maxillary dentition on MSNM BES SC 1018, i.e.
with 14 tricuspid teeth.
The left dentary of MSNM BES SC 1018 bears
19 teeth, with the anteriormost six conical, and the
remaining ones tricuspid. In MSNM BES SC 265 the
anteriormost preserved teeth are conical but the actual
number of conical and tricuspid teeth, as well as the
total number of dentary teeth, cannot be ascertained.
Wild (1974: 47, g. 80) stated that the small-sized
specimens in the PIMUZ Collections have at least
eight, or even 11, anterior conical teeth, followed by
tricuspid teeth, for a total of 17-19 dentary teeth. Again,
both the left (Figs. 50, 51 B) and the right (Figs. 48,
52) isolated post-dentary parts of the lower jaws of
this specimen I identied the suture between the two
elements as an oblique, dorsally convex line, which is
positioned more dorsally than assumed by Wild. In all
probability, he identied as a suture the line delimiting
a wide ventral area characterized by a striking sculp-
tured surface for muscular attachment. This was prob-
ably for the insertion of the posteroventral 1b (super-
cialis layer of the external adductor) described by
Rieppel & Gronowsky (1981), and assumed by these
authors to be present also in Macrocnemus (Rieppel &
Gronowsky, 1981: g. 6 B), a taxon closely related to
Tanystropheus.
Posteriorly, as can be seen more clearly in the iso-
lated post-dentary part of the left lower jaw of PIMUZ
T 2484 (Fig. 51 B), the surangular meets the articular
in an interdigitating suture, and the angular meets both
the articular and prearticular. The nature of the con-
tact between the surangular, angular, and articular on
the medial surface of the lower jaw in Tanystropheus
remains elusive (Fig. 53 B).
In the left lower jaws of MSNM BES SC 265 and
MSNM BES SC 1018 (Fig. 49 C-E) it is conrmed
that the prearticular formed the posteriormost part of
the ventral margin of the lower jaw without, however,
completely underlying the retroarticular process. The
prearticular is dorsally delimited from the articular by
a straight or dorsally slightly concave suture (Figs. 49,
51-52). In MSNM BES SC 1018 there is evidence for
the prearticular being exposed to a lesser extent than
supposed by Wild on the medial surface of the lower jaw
(Fig. 53).
The coronoid is an element difficult to identify in
the material of Tanystropheus. Wild (1974: fig. 15,
pl. 18; Fig. 48 in this paper) interpreted as the right
coronoid in lateral view an isolated element of sub-
triangular shape in PIMUZ T 2484. This element has
a distinct peduncle and an expanded portion entirely
occupied by a concave articular facet. Wild (1974:
37, figs. 15-16; Fig. 53 A in this paper) interpreted
the peduncle as directed posteriorly and ventrally and
the coronoid probably overlapping the surangular,
the dentary and the splenial. He also stated that “The
slightly convex dorsal lamina of the coronoid rises
only insignificantly above the dentary” (Wild, 1974:
37). This interpretation is difficult to understand. In
my opinion there is no evidence of the peduncle being
directed ventrally and posteriorly. I think that the only
way to interpret the presumed coronoid of PIMUZ T
2484 is to suppose that the peduncle was directed dor-
sally and posteriorly, forming a rudimentary coronoid
process, and that the concave facet on the expanded
part of the bone was in contact with the medial surface
of the lower jaw. Based on this interpretation, the pre-
sumed coronoid in PIMUZ T 2484 would be the left
one in lateral view.
In the new specimens there is no clear evidence
of the coronoid morphology. However, a rod-shaped
element can be seen close to the dorsal margin of the
mandible in the left lower jaw of MSNM BES SC 265
(Fig. 49 C). In the left lower jaw of MSNM BES SC
1018, an approximately rod-shaped but curved ele-
ment that corresponds in position to the rod-shaped
element of MSNM BES SC 265 is preserved (Fig.
49 E). A curved isolated element similar in shape and
position to that observed in MSNM BES SC 1018 is
also seen in T. meridensis (Fig. 49 G). This element
was interpreted by Wild (1980a) as an ectopterygoid.
Finally, on the right lower jaw of MSNM BES SC
1018 (Fig. 53 B), preserved in medial view, there is a
distinct rod-shaped element similar to that observed in
the left lower jaw of MSNM BES SC 265 and in the
same position as all the others. In the right lower jaw
of MSNM BES SC 1018 I tentatively identified as part
of the coronoid also an area ventral to the rod-shaped
element described above, which might correspond to
the expanded part of the isolated bone interpreted by
Wild as a coronoid in PIMUZ T 2484. The rod-shaped
element observed in different specimens in lateral
view might represent the peduncle of the same bone,
forming a rudimentary coronoid process.
62 STEFANIA NOSOTTI
there seems to be some variability but I reconstructed
the dentary dentition based on MSNM BES SC 1018.
If, as stated by Wild (1974: 47, g. 80), the large-sized
specimens in the PIMUZ Collections had 19-20? teeth
on the dentary, there is no appreciable difference in the
number of dentary teeth in specimens of different size.
The conical teeth of the upper and the lower jaw are
interlocking (Ford, 2002), while the tricuspid teeth of
the upper jaw are positioned labial to the corresponding
teeth of the lower jaw.
At least 12 tooth-attachment sites are present on the
left pterygoid of MSNM BES SC 1018. Wild (1974:
13-14) counted 14 alveoli in the pterygoid of PIMUZ
T 2484 (Fig. 43), but emphasized that this number is
variable (he counted less than 13 alveoli in PIMUZ
T 2482). No information can be given regarding the
shape and size of pterygoid teeth based on the new
specimens. According to Wild (1974: 13, 48) the ptery-
goid teeth were all conical and forming a shagreen-like
dentition.
The elements identied as the left and the right
palatine in MSNM BES SC 1018 each bear four large
alveoli. Wild (1974: 48) counted ve large alveoli
plus a smaller one in the small-sized specimens in the
PIMUZ Collections. He reported no palatine teeth pre-
served in situ in the PIMUZ specimens but mentioned
an isolated conical tooth in PIMUZ T 2484 with a
basal diameter matching the size of the palatine alveoli
(Wild, 1974: g. 24 e). This tooth is similar in shape to
the palatine tooth tentatively identied in MSNM BES
265 (p. 11).
In both MSNM BES SC 1018 and MSNM BES
SC 265 only isolated vomerine teeth were identied.
According to Wild (1974: 47-48) the total number of
vomerine teeth in the small-sized PIMUZ specimens
is 9-12. The tiny vomerine teeth identied in the new
small-sized specimens are similar in shape to those
described by Wild (1974: g. 24 d). The same is true
for the vomerine teeth of the large-sized MSNM V 3663
(p. 38, Fig. 31), some of which are preserved still in
situ. Based on this specimen the vomer bore at least ten
alveoli.
The axial skeleton: an overall description
(Figs. 1, 4-8, 14-20, 30, 33-34, 54-60, Pls. I-IV, Tabs. 2-3)
As emphasized by Wild (1974: 52), the remarkable
compression and/or incomplete preservation of the ver-
tebral column in the T. longobardicus material from the
Grenzbitumenzone (=Besano Formation) hampers the
detailed reconstruction of vertebral morphology. Speci-
mens MSNM BES SC 265 and MSNM BES SC 1018
support Wild’s statement and some crucial issues, such
as the precise orientation of the zygapophyseal articular
facets, remain elusive. Thus, as pointed out by Wild
(1974: 52), examination of T. longobardicus specimens is
of little value for understanding the functional morphol-
ogy of the vertebral column, particularly of the cervical
series. On the other hand, Wild (1974: 67-94, gs. 39-62)
did describe in detail some isolated, three-dimensionally
preserved vertebrae of T. conspicuus from the Upper
Muschelkalk of Bayreuth. It was on this material that
he based his arguments for the mobility of the vertebral
column and the posture of the neck. Maintaining that the
vertebrae of T. conspicuus were very similar to those of
T. longobardicus, Wild (1974: 52) argued that the same
conclusions could be applied to both species and recon-
structed the vertebrae of Tanystropheus based on T. con-
spicuus (Wild, 1974: g. 38).
As discussed below, the vertebral morphology in
the new specimens of T. longobardicus is similar to
but not exactly the same as T. conspicuus. Moreover,
differences in vertebral morphology can be observed
between smaller- and larger-sized individuals of T.
longobardicus in the PIMUZ Collections. Interestingly,
Dalla Vecchia (2006) recently described isolated, three-
dimensionally preserved vertebrae of Tanystropheus
cf. longobardicus from the Middle Triassic of northern
Friuli (Italy). These vertebrae belong to large-sized
individuals and some of their features are different
from those described for the new specimens. These dif-
ferences in vertebral morphology might alternatively
represent ontogenetic variation within the same taxon
or taxonomic variation in separate species, and can be
evaluated only with a complete revision of the genus
Tanystropheus. On one hand, T. longobardicus and T.
conspicuus might be synonymous (Wild, 1980b: 204;
1987: 39), while the small- and large-sized T. longo-
bardicus specimens in the PIMUZ Collections might
represent two separate taxa (Fraser et al., 2004).
Cervical vertebrae and ribs
(Figs. 1, 7, 14, 30, 34, 54-59, Pls. I-IV, Tab. 2)
All T. longobardicus specimens, those described here
included, conrm that the cervical vertebral column com-
prises 12 vertebrae (Wild, 1974: 50, although 13 cervicals
were represented by Wild in pl. 1).
The cervical vertebrae three through ten are typi-
cally extremely elongate and slender, and the increase
in centrum length is not concomitant with an increase in
height.
The length/height ratio can be estimated taking the
measurement for height at the middle of the centrum,
where the neural spine is absent or weakly developed
as a low keel. In the new specimens, the centra of cervi-
cal vertebrae three through ten are at least nine times as
long as the height of the vertebra at the middle of the
centrum. The highest ratios occur in cervicals eight and
nine, with centra 13-15 times as long as the vertebral
height.
In Fig. 54 the pattern of vertebral lengthening along
the cervical column in MSNM BES SC 265, MSNM BES
SC 1018 and in the PIMUZ specimens is compared by
plotting the ratios “length of cervical vertebrae two (axis)
through twelve / length of the axis”. The pattern obtained
is similar for the new specimens and those in the PIMUZ
Collections, irrespective of the overall size of the indi-
vidual specimens. As stated by Wild (1974: g. 29), there
is a peak of vertebral lengthening between the second and
third cervical vertebrae. After a further, yet moderate,
increase in length between the third and fourth vertebrae,
vertebral lengthening is negligible between the fourth and
sixth vertebrae. A gradual, yet continuous, increase in
length is observed again between the sixth and ninth ver-
tebrae, the latter being the longest (with the single excep-
63
TANYSTROPHEUS LONGOBARDICUS: RE-INTERPRETATIONS OF THE ANATOMY BASED ON NEW SPECIMENS FROM BESANO
Fig. 54 – Ratios “length of cervical vertebrae two (axis) through twelve/length of the axis” in Tanystropheus longobardicus (measure-
ments for the PIMUZ specimens after Wild, 1974: tab. 3). The specimens are listed from the smallest-sized one to the largest-sized one.
tion of PIMUZ T 2818). Posterior to the ninth vertebra,
the trend is reversed, with the vertebral length slightly
decreasing in the tenth vertebra relative to the ninth, and
very steeply so between the tenth and the 12th.
Wild (1974: 52) stated that with increasing age
(=overall length) of the individuals the difference in
length between the eighth, ninth and tenth cervical ver-
tebrae becomes smaller and smaller but this is apparently
only true for PIMUZ T 2818, in which cervical vertebrae
eight through ten reach very similar lengths, the eighth
being the longest. Wild (1974: 52) also maintained that
in specimens over 2 meters overall length (i.e. his “adult”
individuals) the mutual rate of lengthening of the cervical
vertebrae changes. In particular, with increasing overall
length, the difference in length of vertebrae four through
ten would decrease, because of the positive allometric
growth in the third and fourth vertebrae, and slower or
no additional growth in the ninth vertebra. However,
Wild’s statement is not conrmed by the pattern shown
in Fig. 54, which shows instead that the mutual rate of
lengthening of the cervical vertebrae is not affected by the
overall length of the specimens. I think that the pattern
of lengthening of the cervical vertebrae in specimens of
different overall length does not contradict the hypothesis
that small- and large-sized specimens of Tanystropheus
represent different ontogenetic stages of the same species.
However, contra Wild (1974: 52), I do not think that it
supports the hypothesis either.
Both in the PIMUZ material and in the new specimens
the atlas is never preserved as a complete element. The
atlas-axis complex of T. longobardicus was reconstructed
by Wild (1974: g. 31) on the basis of the large-sized
specimen PIMUZ T 2819 (Fig. 55). This specimen pre-
serves the paired neural arches of the atlas, and other
isolated elements interpreted by Wild as the proatlas, the
atlas centrum, and the atlas intercentrum (Wild, 1974:
g. 6, pl. 16). The right atlas neural arch of PIMUZ T
2819 is very similar in shape to the left atlas neural arch
of MSNM BES SC 1018 (Figs. 10-12, 56 B). Both are
exposed in lateral view. The neural arch is an L-shaped
element with a posterior and a ventral ramus. In PIMUZ
T 2819 the posterior ramus is more developed than the
ventral one, while in MSNM BES SC 1018 the two rami
approximately reach the same length. The postzygapoph-
ysis takes the form of a rib-shaped, horizontal relief pro-
jecting backwards from the posterior ramus of the neural
arch. According to Wild (1974: 55, g. 31), the neural
arch projected anteriorly into a pointed process, the cornu
atlanticum but the presence of this process on the left atlas
neural arch of PIMUZ T 2819 is not clear. By contrast,
I assume that the right atlas neural arch of MSNM BES
SC 1018 is articulated anteriorly with the proatlas (Figs.
10-12). Based on this assumption it is unlikely that the
two isolated elements interpreted by Wild as the proatlas
in PIMUZ T 2819 (Fig. 55) were correctly identied. As
an alternative I suggest that the element labelled “pa?”
in Fig. 55 and partially overlapped by the left quadrate is
the proatlas in articulation with the right atlas neural arch.
Wild considered this fragment of bone to be part of the
left quadrate. I do concur with Wild on the identication
of the atlas centrum and intercentrum in PIMUZ T 2819
(see also Figs. 10-12).
64 STEFANIA NOSOTTI
Fig. 55 – Tanystropheus longobardicus, PIMUZ T 2819, atlas and axis on the interpretation of Wild, 1974. pa?: proatlas on the inter-
pretation of the author. Scale bar 20 mm. Photo: Heinz Lanz, PIMUZ.
The morphology of the axis is well documented in
the known material of Tanystropheus, and is apparently
the same in both small- and large-sized specimens (Wild,
1974: gs. 30-31). The reconstruction given by Wild
(1974: g. 31), based on the large-sized PIMUZ T 2819
(Fig. 55), is consistent with the new specimens (Fig. 56).
Apart from the atlas, the axis is invariably the shortest
vertebra of the cervical column. The articular surfaces
of the centrum are not inclined. The anterior surface is
ventrally bevelled, receiving an elongate intercentrum. A
latero-ventral margin (die Lateroventralkante or margo
inferior, sensu Tschanz, 1986: 60) running along the
entire length of the centrum delimits a well developed
ventral keel (=hypapophysis sensu Wild, 1974: 55). As
in the subsequent cervicals (see below), the latero-ventral
margin bifurcates anteriorly into two sharp cristae for the
rib articulation. As compared to that of the subsequent
cervicals, the neural arch of the axis is well developed,
raising anteriorly into a rounded neural spine. It has an
anterior tubercle for the insertion of the atlanto-occipital
ligaments (Wild, 1974: 56; Fig. 56 A in this paper). The
prezygapophyses are apparently short and stocky (Wild,
1974: g. 31) and do not project beyond the anterior end
of the centrum. According to Wild (1974: 56), they origi-
nally had horizontal articular surfaces but this cannot be
veried in the new specimens. The articular relationship
between the atlas neural arch and the axis, as interpreted
by Wild (1974: g. 31), is conrmed by MSNM BES SC
265 (p. 13, Fig. 56 A). An articular facet at the basis of the
axis neural arch, as seen in PIMUZ T 2819 (Fig. 55) prob-
ably received the atlas neural arch. The postzygapophyses
of the axis are elongate and project beyond the posterior
end of the centrum.
The morphology of cervicals three through ten is simi-
lar in the new specimens, and it is particularly clear in the
more three-dimensionally preserved vertebrae of MSNM
BES SC 1018. The anterior and posterior articular surfaces
of the centra are not inclined. By contrast, Wild (1974: 95)
ascertained that in T. conspicuus these surfaces are inclined
anteriorly to a variable degree. As is typical for Tanystroph-
eus, strongly keeled margins run along the centrum (Wild,
1974; Tschanz, 1985, 1986). However, a latero-dorsal
margin (die Laterodorsalkante or margo lateralis, sensu
Tschanz, 1986: 61) running along the entire length of the
centrum and continuing onto the dorsal margins of the pre-
and the post-zygapophysis was not observed in the new
specimens. The neural spine is poorly developed.
The third and fourth cervicals of MSNM BES SC 1018
(Fig. 57 A) have a strongly keeled latero-ventral margin
running along the entire length of the centrum, approxi-
mately at its half-height. On this basis the cross section in
the middle of the centrum would have been sub-circular.
The latero-ventral margin is posteriorly bid. Anteriorly it
bifurcates into two sharp cristae separated by a groove, for
the rib articulation. This latter feature is very distinct also
in the cervical vertebrae of MSNM BES SC 265 (Fig. 58).
TANYSTROPHEUS LONGOBARDICUS: RE-INTERPRETATIONS OF THE ANATOMY BASED ON NEW SPECIMENS FROM BESANO 65
Fig. 56 – Tanystropheus longobardicus, axis. A) MSNM BES SC 265,
right lateral view. Scale bar 5 mm. B) MSNM BES SC 1018, left lateral
view. Scale bar 5 mm. Photos: Massimo Demma.
The dorsal crista continues onto the ventral margin of the
prezygapophysis. A keeled ventral margin (die Ventral-
kante or margo ventralis, sensu Tschanz, 1986: 59) was
probably not developed in these vertebrae. The neural
spine is only anteriorly developed in the third cervical,
where it consists of a low ridge projecting into a pointed
tip rising up from the dorsal surface of the centrum. In the
fourth cervical it is rudimentarily developed also posteri-
orly. Anteriorly, the neural spine bears a sharply pointed
process, and ventral to it is a sub-triangular fossa bordered
by raised margins.
In MSNM BES SC 1018, posterior to the fourth
cervical, the raised latero-ventral margin runs along the
centrum from the prezygapophysis, becoming lateral to
the at ventral surface of the centrum, and levels out in
its posterior part (Tschanz, 1986: 60) (Fig. 57 B). On this
basis the cross section in the middle of the centrum would
have been sub-triangular (Tschanz, 1986: 60). A ventral
keel seems to be developed anteriorly in cervicals eight
through ten. In cervicals ve through ten the neural spine
extends along the entire length of the centrum as a con-
tinuous, yet very low, dorsal keel, which becomes higher
anteriorly and posteriorly (Fig. 57 B).
In the specimens of T. longobardicus of the PIMUZ
Collections Wild (1974: 58) observed that from the sixth-
seventh cervical on backwards the neural spine grows sig-
nicantly higher anteriorly and is gradually reduced pos-
teriorly. In cervicals three through ten of the larger speci-
mens a dorsal keel is absent, and the neural spine is almost
completely reduced posteriorly (Tschanz, 1986: 62). A
neural spine only anteriorly developed was described
by Dalla Vecchia for the large, three-dimensionally pre-
served cervical MFSN 31579 (T. cf. longobardicus, Dalla
Vecchia, 2006: g. 8), tentatively identied as the ninth
(Dalla Vecchia, 2006: g. 17). Dalla Vecchia suggested
that this feature might be an apomorphy indicating that
the new material from Friuli represents a new species of
Tanystropheus.
Fig. 57 – Reconstruction of an anterior cervical vertebra (A) and of a mid-cervical vertebra (B) in small-sized specimens of Tanystro-
pheus longobardicus. Left lateral view. Not to scale. Watercolor: Massimo Demma.
66 STEFANIA NOSOTTI
Both pre- and postzygapophyses of the cervical verte-
brae three through ten project well beyond the anterior and
posterior ends of the centrum. A strongly keeled margin
delimits the postzygapophysis ventro-laterally, extending
anteriorly up to approximately one half the length of the
centrum. In MSNM BES SC 1018 the postzygapophyses
have a nely striated surface for muscular insertion (die
Intertransversal-Muskulatur, Wild, 1974: 58).
In MSNM BES SC 265 and MSNM BES SC 1018,
the 11th vertebra (Fig. 14) is similar in its general shape to
that gured by Wild (1974) in g. 97 a (PIMUZ T 1270).
The centrum is remarkably shorter than that of the preced-
ing cervicals, and the neural arch higher, yet not as high
as in the corresponding cervical of T. conspicuus (Wild,
1974: g. 97 b-c). The anterior and posterior articular
surfaces of the centrum are not inclined. Two distinct
facets for the rib articulation are borne on two tubercles
positioned antero-ventrally on the centrum and separated
by a groove. A latero-ventral margin runs from the dorsal
tubercle along approximately three-fourths of the centrum,
delimiting a ventral keel. Both pre- and postzygapophyses
project beyond the articular surfaces of the centrum. The
neural spine runs along the entire length of the centrum as
a low keel, raising anteriorly and posteriorly, so that the
outline of the neural spine in lateral view is concave. As
is also the case in the preceding cervicals, the neural spine
deepens anteriorly into a sub-triangular fossa bordered by
raised margins.
In MSNM BES SC 265 (Figs. 4, 7) the 12th cervi-
cal is very short, the length of its centrum only slightly
exceeding that of the axis (Tab. 2). Albeit damaged, its
neural arch is clearly much higher than in the preceding
cervicals. As emphasized by Renesto (2005), the poste-
rior articular surface of the centrum is antero-posteriorly
inclined. Similar to the 11th vertebra, two antero-ventral
tubercles separated by a groove bear distinct facets for the
rib. A latero-ventral margin runs from the dorsal tubercle
along almost the entire length of the centrum, delimiting
a ventral keel. By contrast with the postzygapophyses, the
prezygapophyses project only very slightly beyond the
anterior end of the centrum. This features are consistent
with the description given by Wild (1974: 80, g. 50) for
the 12th cervical of T. conspicuus.
Little can be added to our understanding of the articu-
lation of the cervical vertebrae on the basis of the new
specimens. As expected, the prezygapophyses are dor-
sally overlapped by the postzygapophyses. Observation of
the disarticulated vertebrae four and ve in MSNM BES
SC 1018 (Fig. 59) reveals that the postzygapophyseal
articular facet was probably sub-vertical in this region.
The postzygapophyseal articular facet of the fourth cervi-
cal has an oval, antero-posteriorly elongate shape and is
positioned ventral and medial to the postzygapophyseal
process (sensu Rieppel, 2001). The latter projects pos-
teriorly beyond the articular facet, roong it, and in the
articulated vertebrae would lie on a sub-horizontal shelf
borne on the prezygapophysis. This shelf is seen on the
fth cervical, medial to the dorso-lateral margin of the
prezygapophysis.
In the large, three-dimensionally preserved cervical
MFSN 31579 (Dalla Vecchia, 2006: 37, g. 8) the postzi-
gapophyseal facets slant in a dorso-lateral/ventro-medial
plane, forming an angle of 45° with the medial plane of
symmetry. The prezygapophyseal facets appear to be
vertical in MFSN 31579 but they face dorso-medially in
other cervicals comprised in the material studied by Dalla
Vecchia (2006: 37-38).
In both MSNM BES SC 265 and MSNM BES SC
1018, the cervical ribs clump in tight, paired bundles
ventral to the cervical column. Both specimens do not
add new information on the well-known morphology of
the cervical ribs in Tanystropheus. The head of each rib
in dorsal view is axe-shaped, with an anterior and poste-
rior process. The latter is present also in the anteriormost
cervical ribs.
Wild (1974: 98) emphasized that the articular facets
for the ribs borne on the centrum are different between
T. conspicuus and T. longobardicus. He described the
Fig. 58 – Tanystropheus longobardicus, MSNM BES SC 265, articula-
tion between the fth and sixth cervical vertebrae. Scale bar 5 mm.
Photo: Massimo Demma.
Fig. 59 – Tanystropheus longobardicus, MSNM BES SC 1018, articu-
lation between the fourth and fth cervical vertebrae. Scale bar 5 mm.
Photo: Massimo Demma.
67
TANYSTROPHEUS LONGOBARDICUS: RE-INTERPRETATIONS OF THE ANATOMY BASED ON NEW SPECIMENS FROM BESANO
articular facets of the latter as elliptical and positioned
in grooves. According to Wild (1974: 59) the tuberculum
and capitulum in the rst eight ribs posteriorly formed
a single synapophysial articular surface, these ribs thus
being functionally olocephalous. This cannot be con-
rmed in the new specimens. However, an isolated rib
head preserved close to the tenth cervical vertebra in
MSNM BES SC 1018 exhibits a completely separated
tuberculum and capitulum. The 11th cervical of the same
specimen unequivocally has two articular facets for the
rib (Fig. 14), as do the 11th and 12th cervicals in MSNM
BES SC 265 (Figs. 4, 7). In the latter specimen, the tuber-
culum and capitulum of the 12th cervical rib are distinctly
shorter and higher, as is typical of a dichocephalous rib
(Fig. 4).
In the specimens described here, as mentioned
above, the tuberculum and capitulum of the function-
ally holocephalous ribs articulate on two divergent crests
- separated by a deepened triangular surface - positioned
ventro-laterally on the centrum. In MSNM BES SC 265,
this feature is clearly seen in vertebrae of the series two
through seven (Fig. 58 A). In MSNM BES SC 1018, it
can be clearly seen in the third and fourth vertebrae. In the
sixth vertebra the articulation of the rib with the centrum is
apparently more ventral. In fact, moving posteriorly there
seems to be a progressive ventral shift in the position of
the rib articulation. The articular area on the centrum in
the mid- and posterior cervicals cannot be described on
the basis of MSNM BES SC 1018.
The extremely thin and elongate shafts of the cervical
ribs are closely packed in MSNM BES SC 265, and it is
difcult to ascertain either their length or number in the
different regions of the cervical vertebral column. The
same is true for MSNM BES SC 1018, in which the shafts
of the cervical ribs are also extensively broken. In MSNM
BES SC 265 one of the fourth ribs is isolated, and its shaft
apparently intact. Its length exceeds that of three cervical
vertebrae.
Dorsal vertebrae and ribs
(Figs. 1, 4-5, 7, 15-18, 20, 33, 60, Pls. I-IV)
All the known specimens conrm that the dorsal verte-
bral column of T. longobardicus consists of 13 vertebrae.
However, Wild (1974) distinguished 11 dorsal vertebrae
and two “lumbar” vertebrae. Indeed, the morphology of
the last two dorsals is different from that of the preceding
ones (see bolow).
Wild (1974) described the dorsal vertebrae in T. longo-
bardicus on the basis of the large-sized specimen PIMUZ
T 2817 and the smaller PIMUZ T 1270. He maintained
that in the other specimens in the PIMUZ Collections
these vertebrae were too badly crushed to permit a recon-
struction.
Common features of the dorsal vertebrae are the ven-
trally concave centra and tall neural spines, as compared
to those of the cervical vertebrae. The neural spines are
dorsally attened and reinforced by a stout, rough ridge
for muscular insertion.
The dorsal vertebrae can be grouped into an ante-
riormost region, comprising vertebrae one through
three; an intermediate region, comprising vertebrae
ve through ten; and a posterior or “lumbar” region
comprising the last two dorsals. The fourth and 11th
dorsal vertebrae cannot be clearly assigned to anyone
of these regions (see below).
In Fig. 60 A the anteriormost dorsal vertebrae are
reconstructed on the basis of MSNM BES SC 265 (Fig. 4)
and MSNM BES SC 1018 (Fig. 16). The height meas-
ured from the ventral surface of the centrum to the top of
the neural spine is approximately equal to the length of
the centrum. The centrum is deeply concave and keeled
ventrally. The posterior articular surface slants antero-
posteriorly. Typically, the anteriormost dorsal vertebrae
bear two articular facets for dichocephalous ribs. The
wide, sub-circular articular facet for the capitulum of
the rib is borne on a distinct tubercle positioned ventrally
and anteriorly on the centrum. A rounded articular facet
for the tuberculum is borne on a stout but short trans-
verse process. The transverse process is positioned on
the anterior half of the centrum, where the neural arch
is fused to the centrum but is somewhat more ventral
Fig. 60 – Reconstruction of an anteriormost dorsal vertebra (A)
and of a mid-dorsal (B) in small-sized specimens of Tanystropheus
longobardicus. Right lateral view. Not to scale. Watercolor: Massimo
Demma.
68 STEFANIA NOSOTTI
in the rst dorsal. A stout bony bridge extends between
the transverse process and the prezygapophysis. A short
buttress supports the transverse process ventrally, both
anteriorly and posteriorly. This arrangement is similar to
that described by Wild (1974: 61) for the rst ve dorsals
in T. longobardicus. The prezygapophyses are short and
project dorsally and only slightly beyond the anterior end
of the centrum. The postzygapophyses are located high
on the neural arch, separated from the centrum by a deep
notch. They do not project beyond the posterior end of the
centrum. The orientation of the zygapophyseal articular
facets is sub-vertical. The neural spine takes the form of
a truncated pyramid, with anteriorly and posteriorly pro-
jecting processes at the top and a at dorsal surface. It is
approximately as long as tall.
The morphology of the anteriormost dorsal verte-
brae in MSNM BES SC 265, MSNM BES SC 1018,
and the small-sized T. longobardicus specimens in the
PIMUZ Collections (Peyer, 1931: pls. 1-3) differs from
that reconstructed by Wild (1974: gs. 38 c, 52) for T.
conspicuus. In the latter species, the height of the anteri-
ormost dorsals, measured from the ventral surface of the
centrum to the top of the neural spine, is approximately
twice the length of the centrum. The neural spines are
sub-rectangular in shape and taller than long (Renesto,
2005: 378).
The centrum of the mid-dorsal vertebrae (Fig. 60 B) is
poorly preserved both in MSNM BES SC 265 (Fig. 7) and
MSNM BES SC 1018 (Figs. 15, 20) but on the basis of the
material in the PIMUZ Collections it is ventrally concave
(see also Renesto, 2005: g. 4; Dalla Vecchia, 2006: g.
10). The length of the centrum is equal to or exceeds the
height measured from the ventral surface of the centrum
to the top of the neural spine. Neither the anterior nor
posterior articular surfaces of the centrum are inclined.
The single, sub-oval articular facet for an holocephalous
rib is borne on a short transverse process. However, it
might look shorter than it really was since the specimens
have been greatly crushed (see for comparison specimen
MCSN 4451, Renesto, 2005: g. 4 A). The transverse
process is positioned close to the base of the neural arch,
approximately mid way along the centrum. A short, ridged
buttress supports the transverse process antero-ventrally
(centrodiapophyseal lamina sensu Dalla Vecchia, 2006).
Short prezygapophyses, positioned at the same level as the
transverse processes, project very slightly beyond the end
of the centrum. The lateral margin of each prezygapophysis
takes the form of a laterally projecting crest reaching the
transverse process posteriorly (Figs. 15, 20, 60 B). The
postzygapophyses are located low on the neural arch and
project only slightly beyond the end of the centrum. The
orientation of the zygapophyseal articular facets is inter-
preted as sub-horizontal (see also Tschanz, 1985: 174). The
neural spines are sub-rectangular and approximately twice
as long as tall (Renesto, 2005: 378).
The morphology of the mid-dorsals of the new speci-
mens from Besano is thus different in some respects from
that of the mid-dorsals of T. conspicuus (Wild, 1974: gs.
38 d-e, 54) and of other large-sized specimens of Tanys-
tropheus such as PIMUZ T 2818 (T. longobardicus, Wild,
1974: pl. 13) and MFSN 31596 (T. cf. longobardicus,
Dalla Vecchia, 2006: g. 10). In these large dorsals the
height measured from the ventral surface of the centrum
to the top of the neural spine is twice or more the length of
the centrum. The neural spine is antero-posteriorly short
and approximately as long as tall. This was linked by
Wild (1974: 61) to the need for larger and higher insertion
areas for the musculature in the large-sized individuals.
A diapoprezygapophyseal lamina (sensu Dalla Vecchia,
2006) runs obliquely from the transverse process on the
lateral side of the prezygapophysis. The transverse proc-
ess is rather high on the neural arch. Specimen MFSN
31596 has quite long and dorso-ventrally attened trans-
verse processes with antero-posteriorly expanded ends,
which according to Dalla Vecchia (2006: 43) might be an
apomorphy of the taxon represented by the material from
Friuli.
Specimen MSNM BES 215 described here (pp. 38-
39) also represents a large-sized individual. The short
postzygapophyses, positioned low on the neural arch
and provided with sub-horizontal articular facets, sug-
gest that this vertebra is in all probability a mid-dorsal.
As in MSNM BES SC 265 and MSNM BES SC 1018, the
transverse process projects laterally from the area where
the neural arch meets the centrum. By contrast, although
the neural spine of MSNM BES 215 is broken, the height
of its preserved part suggests that it was very tall. Finally,
this dorsal vertebra is unique for all known Tanystropheus
material for the presence of postzygapophyseals canals in
the postzygapophyseal trough. Wild (1974: 81) described
a deep postzygapophyseal trough (die postzygapophyseale
Grube) and postzygapophyseal canals (der Doppelkanal;
das Röhrenpaar sensu von Meyer, 1847-1855) in the cer-
vical vertebrae of T. conspicuus. Rieppel (2001) described
a similar feature in the cervical vertebrae of T. haasi. In
this species, postzygapophyseal grooves in the oor of the
postzygapophyseal trough extend far anteriorly into the
neural arch, separated from one another by a thin vertical
bony septum, and separated from the neural canal by an
equally thin horizontal septum. As formerly pointed out
by Wild (1974: 81) the postzygapophyseal canals end
blindly. Wild (1974: 81) stated that a postzygapophyseal
trough could be seen also in the dorsal, lumbar and caudal
vertebrae of Tanystropheus. However, postzygapophyseal
canals have been never described in the post-cervical
vertebrae.
Based on the small-sized Tanystropheus specimen
MCSN 4451 Renesto (2005: g. 4) assumed that the
nature of the rib articulation changes gradually along
the dorsal vertebral series and described an intermedi-
ate condition in which two articular facets for the rib are
very close to each other but still not conuent. Specimen
MSNM BES SC 265 unequivocally shows that from the
fth dorsal on backwards there is a single articular facet
for the rib, borne on the transverse process. A close look
at MCSN 4451 reveals that the ventral articular facet on
the vertebra gured by Renesto (2005) in g. 4 A is in fact
an artefact of fossilization, and that there is but a single
articular facet for the rib. I was not able to conrm the
presence of two articular facets close to each other in the
vertebra gured by Renesto (2005) in g. 4 B. Indeed
the vertebra is badly crushed and difcult to interpret.
In MSNM BES SC 265 the fourth dorsal is very poorly
preserved. However, as far as can be judged from the
morphology of the transverse process and prezygapo-
physis, it has a single articular facet for the rib just like
all subsequent dorsals. By contrast, on the basis of speci-
mens PIMUZ T 2817 and PIMUZ T 2787, Wild (1974)
69
TANYSTROPHEUS LONGOBARDICUS: RE-INTERPRETATIONS OF THE ANATOMY BASED ON NEW SPECIMENS FROM BESANO
considered the fourth dorsal had two articular facets for
the rib, while from the fth on there was a single articular
facet. An intermediate condition similar to that described
by Renesto (2005) for MCSN 4451 might be present in an
isolated dorsal of MSNM BES SC 1018, that I tentatively
interpret as the fourth (Fig. 17). The strongly concave and
keeled centrum, the antero-dorsally projecting prezygap-
ophysis, and the shape of the neural spine are typical char-
acters of the anteriormost dorsal vertebrae. By contrast,
the postzygapophysis is very similar in shape to that of
the subsequent dorsals, probably bearing a sub-horizontal
articular facet. This vertebra has an unequivocal single
articular facet for the rib on the transverse process, with
a stout, oblique, and distinctly projecting crest extending
between the transverse process and the prezygapophysis.
Another crest extends from the transverse process to the
antero-ventral area of the centrum. Along this crest there
may be a second articular facet. If this is the case, this
vertebra has two facets for the rib articulation showing a
condition similar to that described by Renesto (2005). I
conclude that there is individual variation in the position
of the facets for the rib articulation on the anteriormost
dorsal vertebrae. However, as a rule the mid-dorsal verte-
brae typically have a single facet borne on the transverse
process.
The morphology of the 11th dorsal vertebra is dif-
ferent from both preceding and subsequent vertebrae.
Although this vertebra is not well preserved in MSNM
BES SC 265 (Fig. 5), it is very similar to an isolated,
well preserved vertebra in MSNM BES SC 1018 (Fig.
18). The centrum of the 11th dorsal bears a single, wide,
roundish facet for the rib articulation borne on a stout
and short transverse process. The transverse process is
positioned very close to the prezygapophysis at the base
of the neural arch. Both anteriorly and posteriorly, the
transverse process is supported by short ventral but-
tresses. The anterior buttress is strongly ridged. The
prezygapophysis is shaped as in the mid-dorsal verte-
brae, with a presumed sub-horizontal articular facet. The
postzygapophysis is positioned high on the neural arch,
and in shape is more like the anteriormost dorsals than
the mid-dorsals.
Dorsal vertebrae 12 and 13 (“lumbar vertebrae”), are
poorly preserved in MSNM BES SC 265 (Fig. 5) but are
strikingly different from the preceding ones. The neural
spine is square and distinctly shorter than in the mid-
dorsals. The postzygapophysis is positioned high on the
neural arch. Finally, these vertebrae bear long pleurapo-
physes1.
In both MSNM BES SC 265 and MSNM BES SC
1018, the dorsal ribs are mostly disarticulated and
scattered, and many of them are missing alltogether.
Observation of the preserved elements suggests that the
morphology of the dorsal ribs in the new specimens is
similar to that described by Wild (1974: 63, g. 35). In
MSNM BES SC 265, the rst dorsal rib is preserved in
association with the rst dorsal vertebra (Fig. 4). The
tuberculum is short and thick, the capitulum is longer
and more slender. The shaft is distally overlapped by
the second dorsal vertebra so that its full length cannot
be ascertained. An anteriormost, isolated dichocephal-
ous rib is also preserved in MSNM BES SC 1018 (p.
28), which is shaped like the rst dorsal rib of MSNM
BES SC 265. Its shaft does not seem broken, although
it appears rather short. The second dorsal rib is not pre-
served in MSNM BES SC 265 but the third is. It is not
completely articulated with the third vertebra but the
tuberculum and the capitulum lie very close to the cor-
responding vertebral articular facets. This rib is slender
with a short tuberculum and a longer capitulum. The
shaft is gently curved. In the new specimens it is not
possible to ascertain whether the fourth dorsal rib was
still dichocephalous but on the basis of the morphology
of the fourth dorsal vertebra in MSNM BES SC 265 (see
above) it probably was not.
Sacral vertebrae
(Figs. 1, 6, 8, 19, Pls. I-IV)
The sacral vertebrae are readily identied based on
their expanded pleurapophyses1 that articulated with the
ilium. Specimen MSNM BES SC 265 (Figs. 6, 8) con-
rms that there are two sacral vertebrae in Tanystropheus.
The isolated sacral vertebra preserved in MSNM BES
SC 1018 (Fig. 19) is interesting, because it displays the
vertebra in lateral view. In MSNM BES SC 1018, the
overall shape of the vertebra is quite different from that
of the sacral vertebrae described by Wild for T. longo-
bardicus (Wild, 1974: 62) and T. conspicuus (Wild, 1974:
gs. 55, 56, tab. 6). These display a shorter, less slender
centrum and a tall, sub-rectangular and antero-posteriorly
short neural spine. A large, three-dimensionally preserved
sacral vertebra referred to Tanystropheus cf. longo-
bardicus (MFSN 31552) was recently described by Dalla
Vecchia (2006: g. 11). It does not preserve the neural
spine but the centrum is very similar to that of the sacrals
in T. conspicuus.
Caudal vertebrae
(Figs. 1, 6, Pls. I-IV, Tab. 3)
The description of the caudal vertebrae is based on
MSNM BES SC 265, which preserves the caudal series in
its original proximo-distal sequence (Figs. 1, 6). The iso-
lated caudals in MSNM BES SC 1018 do not add relevant
information and neither specimen provides any real detail
of the caudals distal to the 11th (Fig. 1, Pls. I-IV).
The centra of the rst three caudal vertebrae are
approximately as long as those of the second and the
11th dorsal vertebrae. The length of the subsequent centra
steadily increases proceeding distally along the caudal
series (Tab. 3). Up to the eight vertebra, the centra are
1 Wild (1974: 62) used the term “pleurapophysis” to indicate a rib fused to the transverse process. As emphasized by Romer (1956:
276), once a rib is completely fused to the transverse process without a discernible suture between the two, it is difcult to establish
whether it represents a rib fused to the transverse process or an elongate transverse process. Both Renesto S. (pers. comm., 2006)
and I observed that there is sometimes a line of separation visible at the very base of the wall of the neural arch, without even a short
transverse process (see for example the third caudal in MSNM BES SC 265). However, it is difcult to tell whether it is a suture or a
fracture. In the absence of additional arguments clarifying this issue, I will adopt the term “pleurapophysis” as used by Wild.
70 STEFANIA NOSOTTI
strongly concave ventrally. Caudals one through seven
have well developed latero-ventral margins, delimiting
a ventral keel in the sixth and seventh. The articular sur-
faces of the centra are not inclined.
The rst neural arch that can be observed is that of the
sixth caudal. The neural spine is lower than in the mid-
dorsals, and dorsally attened. Distal to the sixth vertebra,
the neural spine progressively shortens and shifts posteri-
orly. The pre- and postzygapophyses lie at approximately
the same level. The prezygapophyses project well beyond
the anterior end of the centrum, while the postzygapo-
physes are shorter, extending only to approximately the
posterior end of the centrum. A crest connects the pre-
and postzygapophyses on the lateral side of the neural
arch. This crest is very well developed in vertebrae seven
through nine but reduced in the tenth vertebra and almost
absent in the 11th. The zygapophyseal articular surfaces
appear sub-horizontal (see also Tschanz, 1985: 174; Dalla
Vecchia, 2000: g. 2; 2006: g. 12).
In MSNM BES SC 265, the rst through eighth
caudals display pleurapophyses (see note on p. 69), that
increasing in length from the rst to the third element in
the caudal series. The pleurapophyses of the fourth and
fth caudals are broken and cannot be measured. A frag-
ment of one such pleurapophyses (Pl. II) suggests that
they were still rather long. Distal to the fth vertebra, the
length of the pleurapophyses is signicantly reduced. In
the eighth caudal the pleurapophysis is rudimentary, and
there are no pleurapophyses on the subsequent caudals.
The number of caudal vertebrae with more or less well
developed pleurapophyses is therefore consistent with the
number given by Wild (1974: 62) of seven-eight for the
small-sized specimens of T. longobardicus. As pointed
out by Wild, the position of the pleurapophysis on the
centrum shifts diagonally ventrally and posteriorly pro-
ceeding distally along the caudal series. The pleurapophy-
ses of the proximal caudal vertebrae are stoutly built (see
also the proximal caudals in MSNM BES SC 1018, Pls.
III-IV), as shown by their wide antero-posterior base, and
they are strongly dorso-ventrally attened.
None of the preserved caudal vertebrae in the new
specimens is similar in shape to the proximal caudals of
T. conspicuus (Wild, 1974: gs. 38 f, 57-61). The latter
possess very tall and antero-posteriorly short, sub-rectan-
gular neural spine. The postzygapophysis is very short
and apparently bears an oblique articular facet. However,
given the rather poor preservation of caudals one through
ve in MSNM BES SC 265 and of the proximal isolated
caudals in MSNM BES SC 1018, the overall shape of
these vertebrae cannot be described and compared to the
specimens in the PIMUZ Collections. The shape of the
sixth caudal vertebra in MSNM BES SC 265, and of those
distal to it, is more like that of the mid-caudal vertebra of
T. conspicuus in Wild’s g. 62. Dalla Vecchia (2000: g.
2, MFSN 25761; 2006: g. 12, MFSN 31549) described
proximal isolated caudal vertebrae of Tanystropheus from
the Middle Triassic of northern Italy, belonging to rather
large individuals. The shape of the neural spine of these
vertebrae is also similar to that described by Wild for the
proximal caudals of T. conspicuus.
The rst preserved chevron in MSNM BES SC 265 is
that between the seventh and eighth vertebrae, and the last
between the 11th and the 12th. As stated by Wild (1974: 67,
g. 37), the shape of the chevrons changes moving dis-
tally down the vertebral series. Their ventral rami become
shorter, and antero-posteriorly expanded in lateral view. An
isolated chevron is preserved in MSNM BES SC 1018 (p.
29, Pls. III-IV), showing that the two branches of the chev-
rons meet ventrally, forming a median spine, and are dor-
sally connected by a transverse bar in a condition slightly
different from that shown in Wild’s (1974) g. 67.
In MSNM BES SC 265, the vertebral series distal to
the 11th caudal vertebra is complete but it cannot be ascer-
tained how many vertebrae are present. On the basis of its
length (p. 15), the series apparently comprises a minimum
of 30 vertebrae of a length comparable with that of the
proximal caudals but this number was surely far higher
if the progressive shortening of the centra towards the
tip of the tail is taken into account. In fact, the estimated
length of the distalmost caudal vertebrae of MSNM BES
SC 265 is 2.5 mm. Unfortunately, the caudal series is not
complete and poorly preserved in MSNM BES SC 1018.
According to Wild (1974: 62), the overall number of
caudal vertebrae was over 50.
The presence of fracture planes indicative of caudal
autotomy (Wild, 1974: 94) cannot be conrmed on the
caudal vertebrae of MSNM BES SC 265.
The mobility of the vertebral column in Tanystropheus
The description of the pre- and postzygapophyseal
articular facets in the three-dimensionally preserved ver-
tebrae of T. conspicuus (Wild, 1974: 96-98) remains the
best starting point for any understanding of the mobility
of the vertebral column in Tanystropheus.
At present, the only detailed descriptions and interpre-
tations of the axial locomotory system in Tanystropheus
are those of Wild (1974) and Tschanz (1985; 1986; 1988).
These authors, however, came to different conclusions.
Analysing the shape and the orientation of the zygapo-
physeal facets, and evaluating the role played by the very
thin and elongate cervical ribs in the mechanics of the cer-
vical vertebral column, Wild (1974) asserted that the neck
of Tanystropheus was very mobile and, when on land,
could assume an S-shaped conguration.Wild recognized
three distinct regions in the neck. In the anterior region,
he considered sagittal movements predominated but with
some lateral exion. In the mid-region, sagittal exion
predominated, while in the caudal region both sagittal and
lateral exion were possible.
Wild (1974) also stated that, compared to the cervical
series, the mobility of the dorsal and caudal series was
less. He assumed that the anteriormost dorsal column
had the capability to ex dorsally and laterally but pro-
ceeding posteriorly towards the sacral region the overall
mobility of the dorsal vertebral column decreased. In the
caudal region as a whole he considered dorsiexion to be
reduced and to be absent alltogether in the proximal cau-
dals. From the mid-caudal region on, Wild thought that
sagittal and lateral movements of the tail were possible.
However, according to Wild, the overall mobility of the
tail did not suggest that it was used for propulsion.
By contrast, Tschanz (1985; 1986; 1988) gave a differ-
ent interpretation of the morphology described by Wild and
postulated that the neck of Tanystropheus had little mobil-
ity. On the other hand, Tschanz thought that there was rea-
sonable lateral movement in both the trunk and tail.
71
TANYSTROPHEUS LONGOBARDICUS: RE-INTERPRETATIONS OF THE ANATOMY BASED ON NEW SPECIMENS FROM BESANO
Gastralia
(Pls. I-IV)
Only a few elements are preserved in MSNM BES SC
265 and while the majority are present in MSNM BES SC
1018, it is difcult to determine their precise number and
arrangement. In the anteriormost gastral rows two medial
elements meet along the midline. They form an angle,
with the vertex facing anteriorly. They are thicker medi-
ally and taper laterally into pointed ends. These elements
appear longer than those in the reconstruction of Wild
(1974: g. 36). Additional elements are thinner, curved,
and pointed at both ends. They possibly represent the lat-
eral elements described by Wild. Wild counted 25 gastral
rows but judging from MSNM BES SC 1018, this number
might have been higher.
Could Tanystropheus walk?
The pes in Tanystropheus
The new specimens offer remarkable new information
on the morphology of the pes in Tanystropheus.
The reconstruction of the tarsus is largely based on
the perfectly articulated tarsus of MSNM V 3730 (Figs.
35, 61) and the disarticulated, yet perfectly preserved ele-
ments in MSNM BES SC 1018 (Fig. 62).
The tarsus of Tanystropheus comprises four ossi-
ed elements. The calcaneum has a polygonal, laterally
rounded shape with thickened lateral margin. Brinkman
(1981: 7) claimed the presence of a “laterally directed
dorsoventrally compressed tuber on the calcaneum”
in Tanystropheus. However, there is neither a proces-
sus lateralis (sensu Schaeffer, 1941: 442) or a lateral,
Fig. 61 – Tanystropheus cf. longobardicus, MSNM V 3730, right tarsus
in plantar view. I-V = metatarsals. Drawing: Massimo Demma.
archosauromorph-like calcaneal tuber. The astragalus
is an elongate element with an oblique orientation. The
astragalus and calcaneum meet along a straight line. Their
mutual mobility was precluded, and the two elements can
be considered as a single unit (Wild, 1974: 118). In some
Tanystropheus specimens, such as MSNM BES SC 265
(Fig. 65) and PIMUZ T 2480 (Fig. 63), the astragalus has
a proximal lateral process overlapping or indenting with
the calcaneum. However, Tanystropheus does not show
the diagnostic features of a complex concave-convex
articulation between the astragalus and the calcaneum
(contra Brinkman, 1981: 7).
Dorsally and distally, as previously noted by Wild
(1974) in the partially articulated tarsus of PIMUZ T
2480 (Fig. 63) and conrmed by MSNM BES SC 1018
Fig. 62 – Tanystropheus longobardicus, MSNM BES SC 1018, left pes
in plantar view, particular of the tarsus. The astragalus and calcaneum
are exposed in dorsal view. I-V = metatarsals. Scale bar 5 mm. Photo:
Massimo Demma.
Fig. 63 – Tanystropheus longobardicus, PIMUZ T 2480, right tarsus in
dorsal view. Scale bar 5 mm. Photo: Heinz Lanz, PIMUZ.
72 STEFANIA NOSOTTI
(Fig. 62), the astragalus and calcaneum both contrib-
ute to an embayment, in which the astragalus receives
distal tarsal four, and the calcaneum receives metatar-
sal V. In MSNM V 3730 (Figs. 35, 61) the tarsus is
exposed in plantar view (p. 42). Medially, the astra-
galus is recessed into a shallow fossa, like in the left
tarsus of MSNM BES SC 265 (p. 19). However, an
embayment formed by the astragalus and calcaneum
proximally delimited by a sharp margin cannot be
seen. At the same time, distal tarsal four and metatarsal
V are not overlapped by the astragalus and the calca-
neum. Consequently, distal tarsal four and metatarsal V
must have possessed rounded proximal articular heads
mirroring the concavity of the embayment (Fig. 64).
The presence of a hinge-like mesotarsal joint in Tanys-
tropheus is thus conrmed (Wild, 1974, after Kuhn-
Schnyder, 1960). Wild thought that the presence of a
mesotarsal joint was one of the characters indicative of
lepidosaurian afnities. However, the mesotarsal joint
in Tanystropheus is quite different from that observed
in the extant squamates, in which the astragalocalca-
neum and the distal tarsals have a complex, highly spe-
cialized morphology, permitting rotatory movements
of the crus on the pes, as well as exion-extension
(Rewcastle, 1980). The fourth ossied element of the
tarsus, distal tarsal three, is positioned medial to distal
tarsal four, between the astragalus, proximally, and
metatarsal III, distally.
I ascertained that the specimen of Tanystropheus
recently described by Renesto (2005: pl. 2 H) corre-
sponds to the description given above with respect to
the shape of the tarsal elements and the nature of their
articulation.
The morphology of the tarsus in MSNM BES SC 265
(p. 19, Figs. 9, 65), preserved in plantar view, looks dif-
Fig. 64 – Reconstruction of the tarsus in small-sized specimens of
Tanystropheus longobardicus. A) in dorsal view; B) in plantar view.
Drawing: Massimo Demma.
Fig. 65 – Tanystropheus longobardicus, MSNM BES SC 265, left
tarsus in plantar view. I-V = metatarsals. Scale bar 5 mm. Photo: Mas-
simo Demma.
ferent from MSNM BES SC 1018 (Fig. 62) and MSNM V
3730 (Figs. 35, 61), and is more difcult to interpret. The
astragalus appears to have a slightly different shape and
is not obliquely oriented within the tarsus. The articular
relationships of distal tarsal three and distal tarsal four
with the astragalus and calcaneum, and between each
other, remain unclear.
Wild (1974: 118) conjectured that the tarsus of
Tanystropheus originally comprised elements that were
not ossied. This cannot be conrmed on the basis of
the new material. However, in MSNM BES SC 265 the
region medial to the astragalus is difcult to interpret: it
neither looks like bone nor matrix (Fig. 65, Pl. I). It is
possible that it represents a chondrogenic focus or a lm
of organic matter, indicative of the potential for chon-
drication of the centrale. Indeed, in other protorosaurs,
including Langobardisaurus and Macrocnemus, the area
distal and/or medial to the astragalus is occupied by an
ossied centrale (Fig. 66). In MSNM BES SC 265 there
is an undoubted gap medial and distal to the astragalus.
The same gap is present in the perfectly articulated tarsus
of MSNM V 3730. I suggest that this gap is indicative of
a cartilaginous element keeping the metatarsals separate
from the astragalus and providing some exibility and
elasticity for the medial area of the tarsus. In some ple-
siosaur limbs there is also indication of the presence of
elements which apparently do not ossify (Forrest R., pers.
comm., 2007).
In the well preserved specimens described here,
metatarsals I-IV are tightly packed, their proximal ends
overlapping partly. In MSNM V 3730 (Figs. 35, 61) an
obliquely oriented articular facet is visible on metatarsal
IV, suggesting that metatarsal V overlapped it. If similar
articular facets were present on all the metatarsals, rela-
tive movements of the metatarsals could not be lateral,
as stated by Wild (1974: 118) but must mostly have been
dorso-plantar. In other words, mutual movements of the
metatarsals would have resulted in an arched, ventrally
concave metatarsus, rather than in a spreading of the foot
due to abduction of the digits.
73
TANYSTROPHEUS LONGOBARDICUS: RE-INTERPRETATIONS OF THE ANATOMY BASED ON NEW SPECIMENS FROM BESANO
The morphology of metatarsal V (in plantar view) is
best seen in MSNM V 3730 (see description on p. 42;
Figs. 35, 61). Metatarsal V is generally described, in
Tanystropheus and its sister-taxa, as “hooked” (Olsen,
1979; Jalil, 1998; Renesto & Avanzini, 2002; Renesto
et al., 2002). The use of this term has been quite sub-
jective over the years, as recently discussed by Riep-
pel et al. (2003: 370). “Hooking” of metatarsal V in
Tanystropheus refers only to the fact that the roundish
proximal head of metatarsal V articulates both with the
calcaneum, proximally, and distal tarsal four, medially.
A rounded process lateral to the distal end of metatar-
sal V might represent a lateral plantar tubercle for the
insertion of the gastrocnemius muscle (Robinson, 1975;
Brinkman, 1980; Rieppel, 1989; Renesto et al., 2002).
In addition, an “outer process” (sensu Robinson, 1975)
with a distinct tuberosity is developed proximo-later-
ally. This might represent the same process described
by Rieppel (1989) for Macrocnemus, for the possible
insertion of muscles involved in the rotation of the
foot.
The phalangeal formula of the pes in Tanystropheus
(Tab. 9) is primitive. The rst phalanx of digit ve takes
the form of a metatarsal.
Fig. 66 – A) Langobardisaurus tonelloi (MFSN 1921), left pes in plantar
view, redrawn after Renesto et al., 2002: g. 5 B. B) Langobardisaurus
pandoli (MCSNB 2883), left tarsus in plantar view, redrawn after
Renesto, 1994: g. 7. C) Langobardisaurus pandoli (MCSNB 2883),
right tarsus in plantar view, redrawn after Renesto, 1994: g. 8 B.
D) Macrocnemus bassanii (PIMUZ T AIII/208), left tarsus in dorsal
view, redrawn after Rieppel, 1989: g. 8 f. Not to scale.
The hindlimb: terrestrial versus aquatic locomotion
Once it was ascertained that Tanystropheus was
not a ying reptile, as originally hypothesized by
Bassani (1886) and von Nopcsa (1923), Peyer (1931)
interpreted Tanystropheus as a lizard-like, terrestrial
reptile.
Later, Wild (1974) suggested a terrestrial habit for
the juveniles but a predominantly aquatic habit for
the adults. Wild considered the hindlimb to be the pri-
mary organ for locomotion in water, mainly because
of a paddle-shaped, webbed foot. According to Wild,
skeletal correlates for aquatic adaptation were the
wide sites for the insertion of adductor and retractor
limb musculature on the ischium and the femur, the
elongation of the metatarsus, and the presence on the
metatarsals of oblique articular facets permitting the
digits to be spread. Wild suggested that digit five, in
particular, had great lateral mobility, and because of
its long, metatarsal-like first phalanx, it contributed
to a large extent to the distal widening of the surface
of the supposed webbed foot. At the same time, Wild
considered Tanystropheus to be capable of moving
on land as an adult, although he did not explain how
in any detail. In essence, he envisaged a lizard-like
gait, in which the mesotarsal joint in the tarsus of
Tanystropheus played a role that is substantially simi-
lar to the function of the same joint in extant squa-
mates (Rewcastle, 1980). Following Kuhn-Schnyder
(1960), Wild (1974: 109-110, 142) thought that the
forelimb was used not only for locomotion on land
but also for digging. Putative skeletal correlates for
this adaptation were an alleged shortened and wide
manus with thickened phalanges and strong claws
(but see p. 33).
Interpretations of Tanystropheus as a fully aquatic
reptile emphasized the role of the hindlimb in support-
ing aquatic locomotion (Tschanz, 1985; 1986), or as
the main organ of propulsion in water (Taylor, 1989).
However, Renesto (2005) recently questioned an
aquatic mode of life for Tanystropheus and discussed
the morphology of the limbs. According to Renesto
(2005: 388), the architecture of the hindlimb is com-
pletely different from that of any appendage-propelled
aquatic vertebrate and lacks any evident adaptation
for swimming in open waters. He maintained that the
hindlimb was the main propulsive force in locomotion
in Tanystropheus, because the forelimb is relatively
very short and the carpus poorly ossied. In fact,
forelimbs considerably shorter than the hindlimbs is
an archosauromorph feature shared by Tanystropheus
with the relatively well known protorosaurian taxa
Langobardisaurus (Renesto et al., 2002) and Macroc-
nemus (Rieppel, 1989), and with the poorly described
Tanytrachelos (Olsen, 1979) and Cosesaurus (Ellen-
berger, 1977; Sanz & López-Martínez, 1984; Peters,
2000a) (Tab. 8).
The new specimens described here confirm that
the architecture of the pelvis and hindlimb in Tanys-
tropheus is very similar to that of the presumed
terrestrial taxa Macrocnemus (Rieppel, 1989) and
Langobardisaurus (Renesto, 1994; Renesto et al.,
2002). However, major differences, are observed in
the pes.
74 STEFANIA NOSOTTI
The pelvic girdle of Tanystropheus is small compared
to the size of the hindlimb and is very similar in shape
and proportions (relative to the hindlimb) to that of Mac-
rocnemus (Rieppel, 1989: g. 4 B). Langobardisaurus
(Renesto et al., 2002: g. 3 A) has a distinct, posteri-
orly projecting process on the ischium, which is absent
in Macrocnemus and Tanystropheus. Contra Renesto
(2005) and Rieppel (1989), Tanystropheus does have a
preacetabular process (Fig. 25) similar to that of Lango-
bardisaurus (Renesto et al., 2002), yet it remains smaller
than that of Macrocnemus. In the latter, the preacetabu-
lar process is robust and turned outward in an unusual
manner (Rieppel, 1989: 384, g. 6).
The femur of Tanystropheus (Figs. 1, 8, Pls. I-IV) is
long and slender, and only slightly expanded proximally.
There is no medial deection of its proximal articular
head (Fig. 26). This feature indicates a sprawling gait.
The distal end of the femur displays rounded condyles
(Fig. 29). The tibia and bula are of sub-equal length
(Figs. 1, 8, Pls. I-IV, Tab. 5); a spatium interosseum is
present. The tibia has slightly and equally expanded ends,
the slender bula has no expanded ends. The proximal
articular surfaces of the tibia and the bula are at or
slightly convex. The knee joint (Fig. 29) is apparently
hinge-like and fully functional.
The tarsus of Tanystropheus differs from that of Mac-
rocnemus and Langobardisaurus (Fig. 66) in having only
four ossied elements (Fig. 61) and a distinct mesotarsal
joint between the proximal and distal elements (Fig.
64). With the exception of PIMUZ T 2480 (Fig. 63), on
Table 8 – Length ratios between different elements of the
forelimb and the hindlimb in the protorosaurs Tanystro-
pheus longobardicus, Tanytrachelos ahynis, Langobardi-
saurus tonelloi, Cosesaurus aviceps and Macrocnemus
bassanii. The length of the manus and pes is consid-
ered to be the length of the longest unit “metacarpal/
metatarsal+phalanges”.
1) Tanystropheus longobardicus, MSNM BES SC 265.
2) Tanystropheus longobardicus, MSNM BES SC 1018.
3) Tanytrachelos ahynis, YPM 7622, Olsen, 1979.
4) Langobardisaurus tonelloi, MFSN 1921, Muscio G.,
1997 and pers. comm., 2006.
5) Cosesaurus aviceps, MGB-V1, Sanz & López-Mar-
tínez, 1984.
6) Macrocnemus bassanii, MSNM BES SC 111.
Length ratios 1 2 3 4 5 6
humerus/femur 0.7 - 0.9 0.7 0.6 0.8
humerus/radius 1.4 1.5 - 1.3 - 1.1
humerus/manus - 2.0 - - - 1.8
radius/manus - 1.4 - - - 1.6
femur/tibia 1.2 - 1.8 1.1 1.2 0.9
femur/pes 1.1 - 1.2 1.2 - 0.9
tibia/pes 1.0 0.9 0.7 1.1 - 1.0
which Wild based the reconstruction of the pes (Wild,
1974: g. 77), the elements of the tarsus in the Tanystro-
pheus material in the PIMUZ Collections are invariably
displaced (Wild, 1974: gs. 74-76). This might suggest a
looser articulation than that observed in the very compact
tarsus of Macrocnemus and Langobardisaurus. A dislo-
cation of the crus and tarsal ossications seems to be the
general rule. It is indicative of a looser crurotarsal joint
in Tanystropheus, than in Macrocnemus (Rieppel, 1989:
g. 8; Fig. 66 D in this paper) and Langobardisaurus
(Fig. 66 A-C), although overall the ankle joint is quite
similar in all three taxa. In Langobardisaurus the tibia
and the bula t loosely into concavities formed by the
proximal borders of the astragalus, calcaneum and cen-
trale (Renesto et al., 2002). In Macrocnemus (Rieppel,
1989), the bula is received in an embayment formed
by the astragalus and calcaneum. The tibia articulates
with the astragalus, bearing a distinct articular facet on
its medial side. This facet forms the proximal part of an
embayment completed by the centrale and distal tarsal
one, within which the tibia is accomodated during the
propulsive phase of the stride (Rieppel, 1989). In Tanys-
tropheus, the articular facet formed by the astragalus
and calcaneum for the bula is similar to that described
in Macrocnemus. There is no identiable facet for the
tibia on the astragalus. However, in MSNM BES SC 265
(Fig. 65) the tibia, even though it is displaced medially,
still contacts the astragalus. The presence of a carti-
laginous centrale in the tarsus of Tanystropheus, as dis-
cussed above, is consistent with the conguration of the
ankle joint in other protorosaurs, in which a contact of
the tibia with the centrale is observed (Benton & Allen,
1997; Dilkes, 1998).
Dislocation of the tarsus and metatarsus is commonly
observed in the fossil material of Tanystropheus. By con-
trast, the metatarsals are in most cases preserved in asso-
ciation, with their proximal ends overlapping. As in other
protorosaurs, this suggests that the metatarsals formed a
functional unit. Metatarsal V has an outer process and a
lateral tubercle similar to the presumed terrestrial Macroc-
nemus (Rieppel, 1989) and Langobardisaurus (Renesto et
al., 2002). The presence in the foot of Tanystropheus of
a metatarsal-like rst phalanx on digit ve is a character
shared with other protorosaurs, in particular Tanytrach-
elos, Langobardisaurus and Cosesaurus. In Tanystroph-
eus, however, the articulated phalanx exceeds the length
of metatarsals I-IV, while in Tanytrachelos, Lango-
bardisaurus and Cosesaurus it is shorter than the longest
metatarsal (metatarsal III in Tanytrachelos, Olsen, 1979:
g. 4 A; metatarsal IV in Langobardisaurus, Renesto et
al., 2002: g. 5, and Cosesaurus, Ellenberger, 1977: g.
12). The phalangeal formula of the pes in Tanystropheus
(2,3,4,5,4) is the same as in Tanytrachelos, Langobardis-
aurus and Macrocnemus (Tab. 9).
Comparing the limbs of Tanystropheus with those of
other protorosaurs, and, in particular, those of the recently
discovered Chinese protorosaur Dinocephalosaurus (Li,
2003; Li et al., 2004), Renesto (2005) concluded that
the morphology of the limbs of Tanystropheus rules
out any kind of aquatic locomotion. He also empha-
sized that none of the authors who had proposed a fully
aquatic mode of life for Tanystropheus outlined which
mode of propulsion was adopted by this reptile. In fact,
Tschanz (1985; 1986) did suggest that Tanystropheus
75
TANYSTROPHEUS LONGOBARDICUS: RE-INTERPRETATIONS OF THE ANATOMY BASED ON NEW SPECIMENS FROM BESANO
Table 9 – Comparison of the phalangeal formulae of the manus and pes in the protorosaurs Tanystropheus
longobardicus, Tanytrachelos ahynis, Langobardisaurus, Cosesaurus aviceps, and Macrocnemus bassanii.
was an axial-subundulational swimmer (sensu Braun &
Reif, 1982; 1985), in which locomotion resulted from
undulatory movements of the trunk and tail, addition-
ally supported by paddling of the hindlimbs. However,
Tschanz did not discuss in detail the morphology of the
hindlimb in Tanystropheus and did not explain how this
paddling was possible.
Assuming an aquatic mode of life for Tanystropheus,
it is still most likely that it returned to land for reproduc-
tion, in which case it must have been capable of moving
on land in some way. However, nobody has ever tried
to explain in detail how this reptile walked. Although a
detailed analysis of locomotion is beyond the scope of
this paper, some points are raised below.
The absence of articular cartilages in the fossil mate-
rial limits a complete understanding of the full range of
movements possible at any joint. The shape of the femur
apparently indicates a sprawling gait for Tanystropheus.
Rieppel (1989) pointed out that with a sprawling gait
there is a requirement for some rotation of the lower
limb relative to the femur, and that in extant squamates
this rotation does not take place within the knee joint, as
demonstrated by Rewcastle (1980) but in the mesotarsal
joint. In Macrocnemus, in the absence of such a joint, as
well as of an archosauromorph-like articulation between
the astragalus and calcaneum, Rieppel (1989) stated
that the rotation took place between the crus and the
proximal tarsal ossications (p. 74, Fig. 67). As a result
of this movement, the distal metatarsal heads would be
aligned at right angles to the body axis, thereby distribut-
ing the body weight during the propulsive phase of the
stride. Renesto et al. (2002) assumed a similar rotation
movement in Langobardisaurus. In the latter, however,
the presence of a larger and more transversally elongate
centrale, the lack of distal tarsal one, and the presence of
a second axis of rotation between the centrale-astragalus-
calcaneum and the distal tarsals-metatarsals, would have
resulted in a more anterior orientation of the pes than in
Macrocnemus (Renesto et al., 2002).
Ongoing research indicates a digitigrade stance in
the pes of many protorosaurs, including Tanytrachelos
(Peters, 2000b: gs. 9 B, 15 A), Cosesaurus (Peters,
2000b: g. 16), Macrocnemus (Avanzini & Renesto,
2002) and Langobardisaurus (Renesto et al., 2002), and
to a bipedal posture, during rapid locomotion - as previ-
ously stated by Rieppel (1989) for Macrocnemus - or even
while standing and walking.
For Tanystropheus, a digitigrade stance in the pes is
the only plausible conguration while walking as well.
Any analysis of terrestrial locomotion in Tanystropheus
must take into account the metatarsal-like proportions
of the rst phalanx of digit ve during pedal plantar-
exion (sensu Brinkman, 1980: 278). As mentioned
above, the length of this phalanx, when articulated with
metatarsal V, exceeds that of the longest (III) metatarsal.
Intuitively, it is clear that only the elements of the pes
distal to the distal end of the metatarsal-like phalanx
could contact the ground. Furthermore, the metatarsus
of Tanystropheus is asymmetrical, although not in the
same manner as Macrocnemus, Langobardisaurus and
Cosesaurus. In the latter taxa there is a regular increase
in length from metatarsal I through metatarsal IV. In
Tanystropheus, the longest metatarsal is metatarsal III,
followed by metatarsal IV, metatarsal II and metatarsal
I. Tanytrachelos (Olsen, 1979: g. 4) apparently has a
similar metatarsal conguration. Again, however, the
length of the articulated metatarsal-like phalanx of digit
ve in Tanytrachelos does not exceed the length of the
metatarsals. Analysing the hinge lines (Peters, 2000b)
and comparing the pes of Tanytrachelos with the Gwyn-
nedichnium trackmaker, Peters (2000b: gs. 9 B, 15 A)
suggested that the pes of Tanytrachelos could adopt both
a plantigrade and digitigrade conguration. Peters also
drew hinge lines in the pes of Tanystropheus, as recon-
structed by Wild, briey discussing the pattern obtained
(Peters, 2000b: 30, g. 15 B). Nevertheless, he was
unable to come to a denitive conclusion about the pos-
sible conguration for the pes in this protorosaur. A pre-
liminary examination of the hinge lines on the superbly
preserved left pes of MSNM BES SC 1018 shows a
slightly different pattern than that gured by Peters in
the plantigrade conguration. The medial and transver-
sal set of lines are more distal (Ma, Ta and Tb in Peters’
g. 15 A were not identied in the pes of MSNM BES
SC 1018), and clearly distal to the articulation of the
metatarsal-like phalanx of digit ve and the subsequent
phalanx. The whole lateral set of lines is too tangential
to have a functional meaning.
Any conguration of the pes in Tanystropheus requires
its re-orientation to bring a set of hinge lines oriented at
right angles to the direction of locomotion. While the
ankle joint in Tanystropheus is similar to that of Macroc-
nemus and Langobardisaurus, in the absence of an ossi-
ed centrale, if any, in Tanystropheus, there is no solid
Phalangeal formula Manus Pes
Tanystropheus longobardicus 2,3,4,4,3 2,3,4,5,4
Tanytrachelos aynis Fraser N. C., pers. comm., 2007 2,3,4,4,3 2,3,4,5,4
Langobardisaurus Renesto et al., 2002 2,3,4,5,3 2,3,4,5,4
Cosesaurus aviceps Ellenberger, 1997 1,2,3,3,1? 2,3,4,5,3
Macrocnemus bassanii Rieppel, 1989 2,3,4,5,3 2,3,4,5,4
76 STEFANIA NOSOTTI
embayment constraining the dislocation of the tibia along
the medio-distal margin of the tarsus. This dislocation is
responsible for rotation and consequently the outward ori-
entation of the pes. The main movement permitted at the
crurotarsal joint in Tanystropheus was probably exion
and extension of the ankle.
In contrast to Macrocnemus and Langobardisaurus,
a (poorly specialized) mesotarsal joint is present in the
tarsus of Tanystropheus. In the reconstruction presented
here (Fig. 64, p. 72), such a joint would have permitted
the dorsiexion of the unit comprising the distal tarsals
and metatarsals on the unit comprising the astragalus
and calcaneum. A little mutual rotation of the two units
might have been possible but I doubt that the extent of
this movement permitted the re-orientation of the pes
required for the propulsive phase of the stride.
On the other hand, the prevalent movement of ex-
ion-extension at the ankle and at the mesotarsal joint in
Tanystropheus, together with the unossied medial part of
the tarsus and the loose joints between the tarsal elements
- and also between the tarsus and the crus - are consistent
with a paddling motion. The overall exibility of the pes
was apparently more important than the mobility at each
joint.
Skeletal correlates for an efcient pedal plantar-
exion (Rieppel, 1989) include a well differentiated
“hooked” (p. 73) metatarsal V, with an outer process
and distinct lateral and medial tubercles, and elongate
metatarsals I-IV forming a single functional unit. The
occurrence of some of these features in Tanystropheus
might be interpreted as plesiomorphic and inherited
from a common ancestry within the terrestrial proto-
rosaurs.
Tanystropheus mode of life: on what side of the
shoreline?
The skeletal anatomy of Tanystropheus is unique and
there is no equivalent in either extant or extinct animals.
The recently described Chinese protorosaur Dinocepha-
losaurus (Li, 2003; Li et al., 2004) has an overall form
closely recalling Tanystropheus. However, in Dinocepha-
losaurus the proportions of the limbs, with shortened
epipodials, the poorly ossied carpus and tarsus, the
rounded astragalus and calcaneum, and the poor differen-
tiation of metatarsal V, unequivocally indicate the adapta-
tion to an aquatic mode of life.
Critical issues in the interpretation of the mode of life
for Tanystropheus include the general proportions and
shape of the body, skeletal anatomy, mobility of the long
neck and static problems related to it, locomotion (see
also the preceding section), and feeding.
Tanystropheus does not have a compact and stream-
lined body (Taylor, 1989). The trunk is very short com-
pared with the extremely long neck and the tail. The latter
is neither deep and/or laterally compressed. The gracile
limbs are not paddle-shaped, and neither shortened nor
broadened. The pes is very long relative to the crus and the
femur, but similar proportions are observed in the related,
presumed terrestrial taxa Langobardisaurus and Macroc-
nemus (Tab. 8). Except for the disproportionately elongate
neck, the overall form of the skeleton of Tanystropheus is
not strikingly different from that of the latter taxa.
Nevertheless, Rieppel (1989) underlined many
skeletal correlates indicative of terrestrial habits for
Macrocnemus which are absent in Tanystropheus, and
emphasized that the overall degree of ossication is less
in adult Tanystropheus than in Macrocnemus. This might
well indicate that Tanystropheus was an aquatic animal
throughout its life. In fact, the mild degree of skeletal
paedomorphosis observed in Tanystropheus might have
important functional effects. As discussed in the preced-
ing section, the morphology of the pes in Tanystropheus
does not suggest a marked adaptation to an aquatic mode
of life, yet it renders it difcult to imagine efcient ter-
restrial locomotion. Moreover, the minimal degree of
ossication of the carpus and the reduced size of the
forelimb suggest that the forelimb was not a major con-
tribution to any kind of locomotion (Renesto, 2005).
Renesto (2005) emphasized that the extremely long,
yet scarcely exible, neck of Tanystropheus seems unsuit-
able in any environment, and summarized an overall view
of previous studies (Renesto, 2005: 387, g. 10). Accord-
ing to Peyer (1931), Wild (1974) and Kummer (1975) the
neck of Tanystropheus was rather mobile and held hori-
zontally or considerably raised.
The axial locomotory system in Tanystropheus was
more recently interpreted by Tschanz (1985; 1986) draw-
ing a comparison with extant reptiles (Iguana and Vara-
nus), and the neck was concluded to have been almost
inexible. Tschanz compared the neck of Tanystropheus
to an architectural construction in which the interspinal
ligaments and the intervertebral muscles represent a ten-
sion boom between two neighbouring vertebrae, while the
bundles of cervical ribs represent a strut. The nal result
would be a neck incapable of dorsal exion. Tschanz
maintained that the neck could only be held horizontally
and not raised above the level of the shoulder. According
to Tschanz, this conguration is strongly indicative of a
fully aquatic mode of life for Tanystropheus. Buoyancy
would have supported the long neck in the water and
would have reduced the hydrostatic stress on the cardio-
vascular system.
By contrast, according to the most recent interpreta-
tion by Renesto (2005), there would be evidence indicat-
ing that the neck was more exible and mobile than sug-
gested by Tschanz. The shafts of the cervical ribs, running
in tight bundles ventral to the vertebral column, would
have at least partially constrained the dorsoventral ex-
ion of the neck but there would have been some limited
exibility, so that the neck was not completely rigid. As
a result of the hollow vertebral centra, it was very light
and, even in the absence of a strong epaxial musculature,
it could have been supported on land, held in an inclined
posture. Drawing a comparison with the long-necked
azdarchid pterosaurs, Renesto (2005) conjectured that
Tanystropheus likewise possessed a muscle/ligament
system allowing the neck to be raised above the horizon-
tal plane.
According to Renesto (2005), the form of the 12th
cervical vertebra in Tanystropheus would account for the
inclined posture of the neck. Its forward-slanting poste-
rior surface, when in juxtaposition with the at anterior
surface of the rst dorsal vertebra, would produce an
angle at the base of the neck. However, the presence of
an asymmetrical intervertebral disc (Tschanz, 1986: 65)
or a slightly dorsally curved vertebral column in the trunk
77
TANYSTROPHEUS LONGOBARDICUS: RE-INTERPRETATIONS OF THE ANATOMY BASED ON NEW SPECIMENS FROM BESANO
region might well result in a horizontally stretched neck,
even in the presence of the forward-slanting posterior sur-
face of the 12th cervical vertebra.
Concerning the static problems related to the presence
of a long neck in the terrestrial environment, Renesto
(2005) emphasized that the whole anterior portion of the
body of Tanystropheus would have been lighter than the
posterior one, and speculated that a large muscular mass
at the base of the tail shifted the centre of mass posteri-
orly, thus stabilizing the body. He cited as evidence for
the presence of this muscular mass the massive patches
of black matter and a large number of small spherules
comprising tiny calcium carbonate crystals at the base
of the tail. It is not clear, however, whether this type of
preservation is related to peculiar taphonomical condi-
tions characterizing the levels of the Lower Meride
Limestone from which the specimen - collected from an
isolated block - is alleged to have come (Renesto, 2005:
378), and whether it is observed in other fossil vertebrates
from the same levels. This would support Renesto’s inter-
pretation. I concur with Renesto that the presence of the
long pleurapophyses of the proximal caudal vertebrae is
related to muscular insertions. At the same time, many
extant reptiles, including lizards, have caudal lipid stores.
In the viviparous skink Niveoscincus metallicus caudal fat
bodies comprise the majority (55-78%) of the fat reserves,
and the 90-95% of caudal fat occurs within the proximal
third of the tail (Chapple & Swain, 2002). Finally, stabi-
lization of the body due to the weight of a posterior, large
muscular mass might also be equally advantageous in an
aquatic animal.
Analysing the patterns of aquatic locomotion within
the Sauropterygia and discussing the probable steps in
the origin of aquatic locomotion among reptiles, Car-
roll & Gaskill (1985) pointed out that in the early stages
of aquatic adaptation propulsion in water is mainly
achieved through lateral undulations of the axial skel-
eton. In the majority of secondarily aquatic reptiles,
use of limbs is disadvantageous in water because of the
constraints imposed by the morphology and stereotyped
pattern of movement inherited from close terrestrial
ancestors. Therefore, the emphasis put on the forelimbs
for propulsion that is observed in nothosaurs and plesio-
saurs can be interpreted as secondary, and subsequent to
a progressive reduction of the forelimbs to limit drag in
the early stages of aquatic adaptation. Extant crocodiles
and the marine iguana also swim with lateral undula-
tions of the trunk and tail, keeping the forelimbs close
to the body and mostly trailing the hindlimbs, which
are occasionally thrust backwards in a paddling motion
(Carroll & Gaskill, 1985). On the other hand, Braun
& Reif (1985) stated that discontinous propulsion,
because of its low efciency at medium to high Rey-
nolds-numbers, plays practically no role in sh. At the
same time, discontinuous propulsion such as paddling
occurs in semi-aquatic tetrapods which can also walk,
y, jump, and even climb. Usually, adaptation to rowing
and paddling requires minor anatomical changes of the
terrestrial locomotory apparatus (Braun & Reif, 1985).
In the light of the discussion above, Tschanz’s (1985;
1986) interpretation of Tanystropheus as an axial-subun-
dulational swimmer (sensu Braun & Reif, 1982; 1985),
which might have also used the hindlimbs as paddles,
seems consistent.
Other long-necked reptiles, such as the plesiosauroid
plesiosaurs, have been interpreted as relatively slow
swimmers living in coastal environments. While they had
evolved efcient propulsive mechanisms, plesiosaurs, in
particular long-necked plesiosauroids, have a poor hydro-
dynamic prole (Massare, 1988): the long neck, increases
drag. In the diagram of “neness ratios” published by Mas-
sare (1988: g. 2), some plesiosauroids do not fall into the
category representing body shapes which minimize the total
drag and which are optimal for fast, continuous swimmers.
According to Massare, many of the plesiosauroids had a
slower cruising swimming speed for a given size than fusi-
form-shaped marine reptiles such as ichthyosaurs and plio-
sauroids. Forrest R. (pers. comm., 2007) emphasizes that,
from a hydrodynamic point of view, a stiff neck is essential
and there is a growing consensus that the neck exibility in
some plesiosauroids was akin to a stiff shing rod (see also
Ford, 2002). Interpreting Tanystropheus as an axial-subun-
dulational swimmer, Tschanz (1985; 1986) underlined that
in this kind of swimmer the anterior part of the body must
be rigid, and that the stiffened part of the body in Tanystro-
pheus is represented by the elongate neck.
The new Tanystropheus specimens conrm that the
articular surfaces of the pre- and postzygapophyses of the
dorsal vertebrae posterior to the second-third and of the
caudal vertebrae are oriented more or less horizontally,
and therefore permit lateral movements. While Tschanz
(1985) stated that the caudal haemapophyses are mod-
erately elongate, I concur with Renesto (2005) in saying
that, on the basis of skeletal evidence, the tail in Tanystro-
pheus is not powerful, enlarged and laterally attened as it
would be expected in a tail-propelled tetrapod. However,
albeit highly speculative, the hypothesis that the shape
of the tail in the living animal was different from that
indicated by the skeletal morphology cannot be excluded.
Thus, for istance, it is possible that ns composed of
soft tissue were present in the living animal. According
to Renesto (2005), the elongate pleurapophyses of the
proximal caudal vertebrae would have hindered lateral
undulation of the tail. However, the pleurapophyses in
specimen MSNM BES SC 265 (Fig. 6) are not too long to
completely preclude lateral movement, and, if the length
of the tail is taken into account, little lateral movements at
any vertebral joint might well have produced an appreci-
able lateral movement of the tail as a whole.
In the light of the discussion above, I regard Tanys-
tropheus as an aquatic protorosaur with close terrestrial
ancestors, living in shallow waters (maximum a few ten
meters deep, Rhöl et al., 2001) and in all probability
returning to land for reproduction. I concur with Tschanz
(1985; 1986) that Tanystropheus was a slow, not highly
specialized swimmer, relying on lateral undulations of the
trunk and tail for aquatic propulsion, perhaps enhanced by
paddling with the hindlimbs.
A critical, still unclear issue in the interpretation of the
mode of life for Tanystropheus is feeding, in particular
method of capture and processing prey. Stomach contents
were reported by Wild (1974: 51, 142) in some large-
sized specimens of Tanystropheus. They consist of sh
and hooklets from cephalopods’ arms. This suggests that
Tanystropheus had the capability to seize elusive prey.
However, it is difcult to imagine how it could catch such
preys both in water and from the shoreline, especially if
the neck was stiff.
78 STEFANIA NOSOTTI
As discussed by Massare (1988), the swimming capa-
bilities of Mesozoic marine reptiles have implications for
the mode of predation. Plesiosauroid plesiosaurs provide
a model of feeding strategies that might be adopted by
an aquatic, long-necked animal. Because of their slower
continuous swimming speeds, plesiosauroids had the
option for one of three strategies: pursue slow prey, eat
sessile prey or use an ambush technique to capture prey
(Massare, 1988). According to Massare, the possibility of
an ambush mode of attack is more speculative for the ple-
siosauroids than for the crocodiles and mosasaurs. How-
ever, very long-necked plesiosauroids might have been
ambush predators, the prey being caught before the large
body was even detected in dark, murky water (Massare,
1988). There is also evidence that the four ipper “gait”
of plesiosaurs, while not especially efcient for normal
locomotion, was very well suited for rapid acceleration
required in the ambush method of predation (Long Jr. et
al., 2006). McHenry et al. (2005) recently found stom-
ach contents dominated by benthic invertebrates in two
Australian elasmosaurid specimens, suggesting that even
structures as specialized as the elasmosaurid neck are not
necessarily indicative of very specialized niches. This is
conrmed by the wide range of tooth forms found in the
long-necked plesiosaurs (Forrest R., pers. comm., 2007).
Contra Wild (1974: 142), it seems very unlikely that
an aquatic Tanystropheus was a fast swimmer pursuing
slower prey. An ambush method of capturing prey, requir-
ing fast starts and rapid acceleration of the body, i.e. short
bursts of swimming, also seems improbable for Tanystro-
pheus (but see Ford, 2002).
An alternative ambush method of capturing prey was
hypothesized by Peters (2005) for the recently described,
aquatic protorosaur Dinocephalosaurus. According to
Peters, this animal was a sit-and-wait predator, hiding in
bottom silt and snatching passing sh from below. Peters
considered the same model applicable to Tanystropheus as
well. Li et al. (2004) argued that extending the head verti-
cally would have been impossible for Dinocephalosaurus
because the hydrostatic pressure would have prevented
lung ination. Peters (2005) suggested that this problem
could have been overcome by gulping a small bubble of
air and carrying it to the bottom in the throat sac before
passing it to the lungs under equalized pressure. Contra
Peters, LaBarbera & Rieppel (2005) maintained that
Dinocephalosaurus was unlikely to have been a benthic
ambush predator, because its orbits are not facing dorsally
and the morphology of the cervical vertebrae indicate that
dorsiexion of the neck was impossible. Moreover, such
motion would generate high drag forces on the neck that
would tend to drag the body of the animal in the direction
opposite to the motion of the head.
Li et al. (2004) maintained that a form of suction
feeding was a possible strategy for Dinocephalosaurus.
Moderate lateral exion of the neck, followed by rapid
straightening and splaying outward of the ribs, would
produce an increase in the esophageal volume, that in
turn would create suction. This model was questioned by
Peters (2005) and Demes & Krause (2005) who denied
the possibility of an expansion of the esophagous as
suggested by Li et al. (2004). They also argued that the
feeding mechanism invoked by the latter authors would
require physiologic adaptations for removal of salt from
the swallowed water and/or some mechanism of second-
ary water expulsion. However, LaBarbera & Rieppel
(2005) objected to the arguments of Peters (2005) and
Demes & Krause (2005), noting that reptilian physiol-
ogy automatically implies the presence of some form
of salt gland. Interestingly, fossil evidence of salt glands
was recently reported by Fernández & Gasparini (2000)
for the marine metrorhynchid crocodiliform Geosaurus.
LaBarbera & Rieppel also suggested that excess water
might have been expelled through the orices between
the fang-like grasping teeth of Dinocephalosaurus.
Taylor (1989) emphasized that the stiffened neck of
Tanystropheus seems inappropriate for an ambush strat-
egy of prey capture, and suggested that the residual lateral
exibility of the neck might have been used for lunging
the head over the prey. Tschanz also (1985; 1986) sug-
gested that lunging the head over the prey might have
been the feeding strategy of Tanystropheus, following a
model of “kinetic inertial feeding”. Given the extremely
elongate neck, a minor dorsal bending in its posterior part
would have permitted a sufcient retraction of the head.
The energy stored by the bent cervical ribs could then
be used to accelerate the head and manipulate the prey
deeper into the pharynx.
Because of the stiffened neck, it is difcult to imagine
Tanystropheus feeding on fast moving prey. Both a model
of kinetic inertial feeding (sensu Tschanz, 1985; 1986) or
suction feeding (sensu Li et al., 2004) implies the capa-
bility of some exion of the neck. The model suggested
by Tschanz seems more plausible, given the sub-vertical
orientation of the pre- postzygapophyseal articulation of
the last cervical and rst dorsal vertebrae. This indicates
that the neck could be moved in the sagittal plane, while
its lateral movement could be achieved only with a lateral
movement of the trunk. Because of the poor knowledge of
the cranio-cervical joint, movements of the head relative
to the neck cannot be estimated.
A functional interpretation of the dentition might be
relevant for our understanding of the feeding strategy
adopted by Tanystropheus. Wild (1974: 50-51) compared
the dentition in the large-sized specimens of Tanystroph-
eus (his “adult” individuals) with that of marine predators
such as the nothosaurs. Ford (2002) rst observed that the
premaxillary and anterior dentary teeth of Tanystropheus
are interlocking and suggested that in this aspect - and in
dorsally placed nares - the skull of Tanystropheus resem-
bles those of sh-eating pterosaurs and sauropterygians
rather than terrestrial diapsid. Indeed, the anterior conical
interlocking teeth of Tanystropheus seem well-suited for
catching prey.
The posterior tricuspid dentition of small-sized Tanys-
tropheus specimens was interpreted by Wild (1974: 50) as
evidence for predominant insectivorous habits. According
to Wild, the tricuspid teeth were an adaptation for grip-
ping small and fast moving prey, catched with the long,
conical anterior teeth. The shagreen-like pterygoid teeth
and the pointed vomerine and palatine teeth conrmed the
functional interpretation of the dentition of Tanystropheus
as a device for capturing and crushing prey.
Based on the comparison with the tricuspid dentition
of the extant marine iguana, Cox (1985) alternatively con-
jectured an herbivorous diet for Tanystropheus. However,
as pointed out by Taylor (1989), the absence of a very
mobile neck would make it difcult for Tanystropheus to
hunt for insects or tear algae from the rocks.
79
TANYSTROPHEUS LONGOBARDICUS: RE-INTERPRETATIONS OF THE ANATOMY BASED ON NEW SPECIMENS FROM BESANO
A new model for the interpretation of the tricuspid
dentition in Tanystropheus is provided by the extant pin-
nipeds. These mammals tipycally display cheek teeth (or
postcanines) little differentiated from one another which,
particularly in the Phocidae, display a considerable vari-
ability in shape and degree of cusp development (Miya-
zaki, 2002a). In general, cusps are related to gripping of
slippery prey. Tricuspid teeth are observed in many spe-
cies. The species of the genus Pusa feed on sh (P. sibir-
ica), sh and crustaceans (P. caspica), or on small sh and
a wide variety of small pelagic amphipods, euphausiids,
and other crustaceans (P. hispida) (Miyazaky, 2002b).
The Harbor Seal (Phoca vitulina) feeds opportunistically
on many kinds of shes, shellshes and squids (Schef-
fer, 1969). In some species, such as the Crabeater Seal
(Lobodon carcinophaga), the cheek teeth have a highly
modied shape with complex elongate cusps to trap and
strain krill (Miyazaki, 2002a).
In a fully aquatic Tanystropheus the posterior tricus-
pid teeth might have well been an adaptation to gripping
and piercing slippery, tough-skinned prey, and possibly
cutting esh. Wild (1974: 50) observed the similarity
of the tricuspid teeth of Tanystropheus with those of the
primitive esh-eating cynodonts, but argued that they
had not a shearing function. He emphasized that the tri-
cuspid teeth in the PIMUZ specimens show horizontal
rather than oblique wear facets, suggesting a tooth to
tooth occlusion. However, oblique wear facets can be
seen in some maxillary teeth in MSNM BES SC 265.
Moreover, based on the new specimens from Besano,
the tricuspid teeth of the upper jaw are positioned labial
to those of the lower jaw. The upper and lower set might
perhaps create a shear pattern similar to the carnassial
shear of Carnivora.
Finally, some authors claimed that an interpreta-
tion of the mode of life for Tanystropheus based on the
understanding of the function of its long neck, might be
misleading (Taylor, 1989). Given the positive allometric
growth of the neck in Tanystropheus (Tschanz, 1988;
Wild, 1974), Taylor (1989) considered the great length
of the neck might be an indirect consequence of the
evolution of large size. In this case the large size might
be an adaptation for a reason other than feeding, such
as sexual competition or defence. Taylor concluded that
perhaps Tanystropheus had a long neck simply because it
was a big animal, and it survived in spite of it rather than
because of it.
Tschanz (1988: 1002) also stated that “structures with
no recognizable adaptive value may be more reasonably
explained as results of allometric growth”. Analysing the
ontogenetic development of the neck in T. longobardicus
and Macrocnemus bassanii, Tschanz observed that the
positive allometric growth of the neck has different
growth parameters for the two taxa. In particular, there
would be evidence for a relatively longer neck and for
decelerated growth of it in Tanystropheus, in comparison
to Macrocnemus. According to Tschanz, this would make
sense assuming the occurrence in the ancestry of Tanys-
tropheus, morphologically exemplied by M. bassanii, of
heterochronic processes, such as hypermorphosis and pre-
displacement, accounting for an extreme elongation of the
neck with increasing body size. In the absence of decel-
erated growth of the neck, the accelerated body growth
would have produced an elongation of the neck rendering
it a functionally inappropriate or unadapted structure.
Discussing the elongation of the neck in different Tanys-
tropheus species, Wild (1987) apparently maintained as
well that an extremely long neck was unadapted, when
recognizing the specialization of the cervical vertebral
region as one of several “evolutionary trends” character-
izing the history of Tanystropheus and potentially leading
to extinction.
The hypothesis that some structures characterized by
unexplained but demonstrated positive allometric growth
developed simply as a consequence of a positive selective
pressure for a large body size was often put forward to
explain “odd” morphology. As in the emblematic example
of the huge antlers of the Irish Elk, peculiar anatomical
traits were traditionally considered as inadaptive. Dis-
cussing the origin and function of “bizarre” structures,
and, in particular, the case of the Irish Elk, Gould (1974)
sharply commented on the interpretation maintaining that
the enormous antlers are a passive consequence of selec-
tion for larger bodies as follows: “Curiously, this standard
contention has not escaped the dead hand of the orthoge-
netic explanation that it bravely claimed to replace. For
it assumes that the antlers were disadvantageous per se,
and that selection preserved them only because it favored
the total phenotype of larger bodies and antlers. Instead
of an immutable trend, we now have an immutable cor-
relation. Thus, the assumption of deleterious antlers was
transported bodily from the orthogenetic to the allometric
argument…”. Gould argued that in the case of the Irish
Elk, the “allometric antidote to orthogenesis” rested more
upon dogmatic assumptions than “upon a rm foundation
of careful and copious data”, and cited evidence for a
primary selective advantage for the antlers, as structures
for display conditioning the female choice or establishing
dominance through ritualized encounters between com-
peting males.
Zarnik (1925) strongly claimed an adaptive meaning
of the long necks of plesiosaurs: “One has to exclude the
notion that these animals were the product of accidental
orthogenesis which we will call hypermorphology; hyper-
morphological disharmony between form and function
would led to their extinction (as for example, the ammo-
nites with untwined spiral in the Cretaceous Period)”.
Recently, Noè (2006) emphasized that the possession of
a long neck in the marine environment poses a number of
functional, biomechanical, ecological and physiological
problems. Nevertheless, its continued presence through-
out the long evolutionary story of the plesiosaurs dem-
onstrates that a long neck can be a successful adaptation
to life in water. It’s also worth noting that the tendency to
hyperelongation of necks occurred in at least two, possi-
bly three distinct evolutionary lineages in plesiosaur evo-
lution (Forrest R., pers. comm., 2007; O’Keefe & Robin,
2001; Smith, 2003).
Finally, elongation of the neck is generally interpreted
as an adaptation related to feeding strategies. In fact,
“unhortodox” hypotheses on the functional meaning of
neck elongation were put forward for some long-necked
animals. This is the case for neck elongation in extinct
sauropods (Senter, 2007) and in extant giraffe (Sim-
mons & Scheepers, 1996), which traditionally have been
thought to have evolved under the pressure of competition
for food, but have alternatively been indicated as driven
by sexual selection.
80 STEFANIA NOSOTTI
IMPLICATIONS FOR PHYLOGENETIC ANALYSIS
Rieppel et al. (2003) recently presented a synthesis of
the most recent/computer supported cladistic analyses of
protorosaurian interrelationships, obtained by combining
the data published by Benton & Allen (1997), Jalil (1997)
and Dilkes (1998). Major difculties affecting the results
of this analysis centre on a very poor knowledge of many
protorosaurian taxa, as well as the description and coding
of the characters included in the analysis. These problems
are often related to objectivity in assessing large list of
characters, within a large list of taxa, and with direct
observation of all the original material.
It should probably be accepted that some characters
will remain only provisionally phylogenetically informa-
tive. However, the importance of rst hand systematic
anatomical work of the original material cannot be over
emphasized. The new specimens described here conrm
this, providing new information on some poorly dened
or previously undescribed characters of Tanystropheus.
The discussion given in this paper also emphasizes how
the description of some characters is (inevitably) based
more on subjective interpretation of the material, rather
than on strong evidence offered by the material itself.
The characters coded by Benton & Allen (1997;
BA=characters by Benton & Allen), Jalil (1997;
J=characters by Jalil) and Dilkes (1998; D=characters by
Dilkes) are re-considered here on the basis of the descrip-
tions and discussions presented in this paper. Contrary
to Benton & Allen, Jalil did not differentiate between T.
longobardicus and T. meridensis but used the genus as
terminal taxon, while Dilkes considered T. longobardicus
only. I assigne characters states for T. meridensis on the
basis of preliminary personal observation, provisionally
maintaining it as a separate species. However, the holo-
type, and the single known specimen, of T. meridensis
preserves only the skull and the rst six cervical verte-
brae. Overlapping characters used by Benton & Allen,
Jalil and Dilkes are reported in parentheses. If not men-
tioned, overlapping characters have been coded the same
by all authors. All the characters of the different authors
are provisionally considered but an important next step is
to exclude characters that might be more easily affected
by inaccurate scoring (see also discussion in Rieppel et
al., 2003). These comprise ill-dened characters, and par-
ticularly those including qualitative terms, that are there-
fore highly subjective. Characters that may be inuenced
by size, ontogenetic stage, individual variation or vagar-
ies of preservation also should be re-evaluated. They are
particularly relevant to ongoing discussions concerning
the different species of Tanystropheus (pp. 5-6).
The coding for the following characters of the skull is
conrmed or emended. Benton & Allen’s coding for BA4,
BA7, BA10, BA12 (J43, D35), and BA15 is conrmed
both for T. longobardicus and T. meridensis. As a quadra-
tojugal is conrmed to be absent in Tanystropheus (BA12,
J43, D35), BA11 (J27), and J39 should be coded not
applicable. Characters BA1, BA5 (J50), BA6, and BA14
(J62) are correctly coded for T. longobardicus and should
be coded the same for T. meridensis. Dilkes’ coding for
D15, overlapping with BA5-6, is accepted. Charac-
ter state 1 (nasals longer than frontals) for BA2 (J61,
D18) cannot be evaluated either on the basis of the new
material or for T. meridensis. However, PIMUZ T 2484
conrms Benton & Allen’s coding for T. longobardicus,
while character state 0 (nasals shorter than frontals) for
T. meridensis is rejected. Nasals tapering anteromedially
(J49) or with anterior process at the midline (D13) can
only be conrmed on the basis of PIMUZ T 2484. Wild
(1974) argued for the presence of a supratemporal, and
this element is possibly present in MSNM BES SC 265
and in T. meridensis. Both the presence and interpreta-
tion of the supratemporal remain uncertain, however,
and Benton & Allen’s coding for BA13 (D31) is only
provisionally accepted. Jalil’s coding for J14 (D29) for
T. longobardicus (postparietal absent) is accepted and
is conceivably applicable to T. meridensis as well. Any
coding for J3 (postparietal large or small) is consequently
not applicable. Character J63, coded as unknown by Jalil,
is probably better coded 0 (no teeth recurved and later-
ally compressed) as per Rieppel et al. (2003) and Dilkes
(D58-59), both for T. longobardicus and T. meridensis.
Jalil’s coding for skull characters J2, J5, J8, J13, J15-16
(J16, D37), J18, J25 (D8 is described more in detail but
this renders it not applicable to Tanystropheus), J28, J42,
J68, and J69 is conrmed, both for T. longobardicus and
T. meridensis. The coding for J17 (large exposure of the
angular) is correct for T. longobardicus, while it is based
only on Wild’s (1980a) reconstruction for T. meridensis.
However, “large” or “restricted” is subjective. Dilkes’
coding for the following characters of the skull is con-
rmed for T. longobardicus: D1, D4-7, D9-10, D16-17,
D19-21, D27, D30, D32-34, D36, D46, D49-50, D54-57,
D60-66, D69-72, D75-76, and D127. The same character
states as for T. longobardicus are conceivably applicable
to T. meridensis for these characters. Many of Dilkes’
characters describe the dentition and, due to the pres-
ence of specialized dentitions in protorosaurs (presence
of tricuspid teeth in Tanystropheus; see also the dentition
of Langobardisaurus in Renesto & Dalla Vecchia, 2000),
more of these characters should be included in future
analyses of protorosaurs interrelationships. I think that
evidence from MSNM BES SC 1018 and T. meridensis
is consistent with Dilkes’ coding for D125 (prefrontals
separate along midline). Based on the new specimens of
T. longobardicus and on T. meridensis the correct charac-
ter state for D73 is 0. None of the states described for D74
is applicable to Tanystropheus, in which the retroarticular
process is formed by the angular and the prearticular
without fusion of the two elements. Dilkes’ coding for
D45 is rejected, because the state of this character cannot
be assessed in Tanystropheus.
Some characters of the skull should be coded differ-
ently in small- and large-sized specimens of Tanystro-
pheus. Character BA3 is coded 1 (fronto-parietal suture
straight) by Benton & Allen, probably based on the
large-sized specimen PIMUZ T 2819 but in small-sized
specimens (not clear in T. meridensis) the suture is inter-
digitating or V-shaped. Character BA16, coded 0 (ptery-
goid ange teeth present) for both T. longobardicus and
T. meridensis, should be coded as unknown in the latter.
The derived state for this character in T. longobardicus
is only applicable to the large-sized individuals. Dilkes
coded palatine and pterygoid (palatine ramus of) teeth as
absent (D67-68) but this is only true for the large-sized
specimens (unknown in T. meridensis). By contrast,
81
TANYSTROPHEUS LONGOBARDICUS: RE-INTERPRETATIONS OF THE ANATOMY BASED ON NEW SPECIMENS FROM BESANO
character state for D136 (crown of marginal teeth tri-
cuspid) is applicable only to the small-sized specimens.
Dilkes’ coding for D25 (parietals fused with loss of
suture) is only applicable to large-sized Tanystropheus
specimens. The suture is still partly present in small-
sized specimens or the two parietals are completely
separated. Dilkes’ coding for D26 (sagittal crest on the
parietal table present) and D28 (median border of pari-
etal drawn downwards to form ventrolateral ange) is
probably based on the large PIMUZ T 2819. In the new
reconstruction of the skull for the small-sized specimens
the parietal table can be described as “constricted with-
out a sagittal crest” and a ventrolateral ange is present
(but see p. 52). However, preservation of the dorsal sur-
face of the parietal is very poor in all known small-sized
specimen of Tanystropheus.
The new specimens demonstrate that the reconstruc-
tion of the skull for small-sized specimens of Tanystro-
pheus is highly speculative in some areas, and that both
Wild’s reconstruction and that presented in this paper
cannot be considered denitive statements on skull mor-
phology. Jalil’s coding for J1 might be correct based on
both reconstructions but the nature of the contact nasal-
prefrontal remains hypothetic both for T. longobardicus
and T. meridensis. The coding for J4 and BA9 (J54) is
provisionally accepted for T. longobardicus but the shape
of the squamosal and the nature of its contacts with neigh-
bouring bones have been only tentatively interpreted.
The same characters should be coded as unknown for T.
meridensis. Jalil’s coding for J26 (D11-12) (external naris
elongate anteroposteriorly and close to the midline) is
probably correct, based only on T. meridensis. Benton &
Allen’s coding for BA8 (posterior process of postorbital
does not extend beyond back of lower temporal fenestra)
and Dilkes’ coding for D23 (ratio of lengths of anteroven-
tral and posterodorsal processes of postorbital >1.0) are in
all probability correct for T. longobardicus, while a pos-
torbital was not identied in T. meridensis. A postorbital
and parietal contact was probably absent (D22, coded
unclearly) and in all probability the postfrontal entered
into the margin of the upper temporal fenestra as stated by
Dilkes (D24). The coding for some characters concerning
the general shape of the skull (J38 and J64, similar to D2)
and the shape of the upper temporal fenestra (D3) remains
uncertain.
The coding for some characters of the skull cannot be
evaluated either on the basis of the new material or for
T. meridensis and remains based on Wild’s (1974) recon-
structions of T. longobardicus. Dilkes’ coding for char-
acters D38-44, D53, D126, D130, and D142 cannot be
conrmed in the new specimens because the bones of the
dermal palate, the basicranium, and the occipital region
are poorly preserved. However, as Wild (1974) thought
that cartilaginous tissue was interposed between the
paroccipital process and suspensorium, the nature of their
contact should not be regarded as “strong” (J7), and the
correct character state for D52 should be 0 (paroccipital
process ends freely). A stape is only tentatively identied
in MSNM BES SC 1018 (Fig. 12) and adds no informa-
tion (see character J6). The prootic is not preserved in any
of the new specimens and very poorly preserved in the
PIMUZ material (see characters J70, and D 47-48). Jalil
and Dilkes coded differently the position of the occipital
condyle relative to the craniomandibular joint (J66, ante-
rior to the craniomandibular joint; D51, even with it) but
as far as can be judged from Wild’s (1974) reconstruction
the correct coding is that of Dilkes. Although Wild (1974)
stated that a septomaxilla is present in Tanystropheus, I
agree with Dilkes’ coding for D14 as unknown. Evidence
for the presence of this element seems to be very poor.
The coding for the following characters of the axial
skeleton is conrmed or emended. Benton & Allen’s
coding for BA17-19 (J56), BA20 (J40, D82), BA22
(J41), and BA25 (J57, D137) is conrmed for T. longo-
bardicus. Only the coding for BA20 and BA22 can be
conrmed for T. meridensis. Character BA19 should
be coded as unknown for T. meridensis, as all the other
characters of the axial skeleton considered by Benton &
Allen, Jalil, and Dilkes. Trunk intercentra are absent in
T. longobardicus. Consequently, the coding for BA24 is
rejected, and that for J67 and D80 is conrmed. Char-
acter BA21 (ovoid spine-table on top of neural spine) is
ambiguous. The shape of the neural spines of the dorsal
vertebrae is discussed on pages 62 and 68. Benton &
Allen’s description (BA23=tall and rectangular) is appli-
cable to the large-sized specimens of Tanystropheus. For
the small-sized specimens this character is better captured
by Dilkes (D85), where he codes the “dorsal neural spine
height” as “low with height < length”. Jalil’s coding for
characters J20, J29 (D83), J36 (D77), J37 is conrmed.
Character J30 should be better described, because only
the posteriormost cervical ribs have completely separated
tuberculum and capitulum. In the anterior cervical ribs the
tuberculum and capitulum are posteriorly conuent and
these ribs are considered to be functionally olocephal-
ous. The character “well-developed transverse processes
of trunk vertebrae” (J31) needs to be re-described. The
transverse processes in the new specimens from Besano
are short but they might appear shorter than they really
were because of compression (see p. 68). Moreover, while
the “lumbar” vertebrae supposedly have long pleurapo-
physes, these might alternatively be considered as fused
ribs or very long transverse processes (see note on p. 69).
The same is true for the pleurapophyses of the proximal
caudals (see Dilkes’ character D89). Jalil’s coding of J19
(no holocephalous dorsal ribs) is rejected. In fact, most of
the dorsal ribs in Tanystropheus are holocephalous, and
only the anteriormost three or four are dicocephalous.
This character is better described and correctly coded by
Dilkes (D86). Dilkes’ coding for characters D78-79, D81,
D87, D90, D92, D128, D131-134 is conrmed. “Proximal
caudal neural spine height” (D88) cannot be assessed in
the new specimens but Dilkes’ coding is applicable to the
vertebrae of T. conspicuus as described by Wild (1974).
The coding for the following characters of the appen-
dicular skeleton is conrmed or emended (all unknown
in T. meridensis). Benton & Allen’s coding for the
following characters is conrmed: BA26 (J44), BA29
(J47), BA30 (J10, J33 and J51), BA31 (J46), BA33-
35, BA36 (J22, D100), BA37-39, BA40 (J59, D115),
BA41 (J35, D116), BA42 (J53), BA43 (J58), BA44-45
and BA47 (for BA40 see comments in Rieppel et al.,
2003: 370; for BA41 and BA42 see discussion on pp.
71-72 of this paper). Benton & Allen’s coding of BA32
(metacarpal three equal in length to, or longer than,
fourth) is conrmed, while Jalil’s coding of the same
character (J55) is rejected. The presence of an entepi-
condylar groove or foramen on the humerus (BA27)
82 STEFANIA NOSOTTI
was reported by Wild (1974) for T. conspicuus but it
is not conrmed for T. longobardicus (Wild, 1974; see
also the specimen of T. cf. longobardicus described by
Renesto, 2005), including the new specimens. Contra
Benton & Allen, both Jalil (J32) and Dilkes (D107)
coded an entepicondylar foramen absent. Character
state 0 for BA28 (radius length relative to the humerus)
is questionable. In the well preserved specimens
described here the radius is 66-70% the length of the
humerus. The desumption of the shape of metatarsal V
(BA46, J12, D122) is a character still causing confu-
sion (Rieppel et al., 2003: 370), and clearer descrip-
tions are needed. Jalil’s coding is conrmed for J9
(D93), J11 (D121), J23, J24 (D114), J45, J52, and J71
(D138). An ectepicondylar foramen was never reported
for Tanystropheus. Both Jalil (J21) and Dilkes (D108)
coded it as absent. I reject Jalil’s coding for J48 (ilium
with reduced contribution in the acetabulum), and
Dilkes’ coding for D105 (relative contribution of pubic
elements to acetabulum approximately equal) because
the new specimens demonstrate that the element pri-
marily contributing to the acetabulum is the ilium (Fig.
25). A concave-convex astragalo-calcaneum articula-
tion is denitely absent in Tanystropheus (see discus-
sion on p. 71): Jalil’s coding for J34 is rejected, and
Dilkes’ coding for D113 conrmed. Dilkes’ coding is
conrmed for D94, D97, D99, D101-102, D104, D106,
D109, D112, D119-120, D123-124, D135 and D144.
Dilkes’ coding for characters describing the shape of
the interclavicle (D96 and D98) is based on Wild’s
interpretation of a single, fragmentary element (Wild,
1974: g. 64). An interclavicle is only tentatively iden-
tied in MSNM BES SC 1018 (Fig. 20). The presence
of a processus lateralis on the pubis (D103) established
by Wild (1974) is not conrmed in the new specimens
(p. 34). Presence and position of a centrale (D117-118)
is correctly coded by Dilkes as unknown. The presence
of a cartilaginous centrale in Tanystropheus remains
highly conjectural (see discussion on p. 72). However,
an ossied centrale contacting the tibia is present in the
related taxa Langobardisaurus and Macrocnemus.
Finally, coding for BA48 (J60) (postcloacal bones)
cannot be conrmed in the new specimens but the pres-
ence of postcloacal bones in Tanystropheus was estab-
lished by Wild (1974) in the PIMUZ material.
At present, there is a general consensus on the archo-
sauromorph, particularly Archosauria, afnities of Tanys-
tropheus. The inclusion of Protorosauria (the priority of
Protorosauria Huxley, 1871 over Prolacertiformes was
established by Chatterjee, 1986) within the Archosauro-
morpha was formerly put forward by Gow (1975), and
was consolidated in the 1980ies through the work of
several palaeontologists, exhaustively summarized by
Benton (1985) and Evans (1988). The strongest argu-
ments against the inclusion of the Protorosauria inside the
Archosauromorpha remain those of Wild (1974; 1980 a).
In a preliminary paper (Nosotti, 1999), I maintained that
the tiny elements observed in the elbow and the knee
joints in MSNM BES SC 265 should be interpreted
as secondary epiphyseal centers of ossication. This
character was recognized by Benton (1985) and Evans
(1988) as a lepidosaurian synapomorphy but was also
described in pterodactyloid pterosaurs (Bennett, 1993),
and discussed in the protorosaur Macrocnemus (Premru,
1991; Rieppel, 1989). Through a closer examination of
these elements and taken together with the nature of the
articular ends of the long bones in the new Besano speci-
mens, I concluded that my earlier interpretation should
probably be rejected.
Dilkes’ (1998) cladistic analysis of protorosaurs inter-
relationships recently raised the possibility that Proto-
rosauria as conventionally conceived are paraphyletic.
This result was conrmed by the re-analysis of Rieppel
et al. (2003). This latter paper should be consulted for an
update on the more recent views on the relationships of
Tanystropheus with other protorosaurian taxa. The sister
group relationships of Tanystropheus and Tanytrachelos
(Olsen, 1979), the two being the representatives of the
Family Tanystropheidae, is, however, unanimously con-
rmed in all cladistic analyses that include both these taxa
(Benton, 1985; Evans, 1988; Jalil, 1994; Benton & Allen,
1997; Peters, 2000a; Rieppel et al., 2003). Tanytrachelos
ahynis, from the Upper Triassic (Carnian) of the USA
was interpreted by Olsen (1979) as an aquatic animal. At
present, Tanytrachelos remains poorly described. More
information on the anatomy of Tanytrachelos would
provide the opportunity for interesting comparisons with
Tanystropheus and perhaps offer new insights into the
supposed aquatic radiation of the (terrestrial) protorosaurs
represented by the Tanystropheidae.
CONCLUSIONS
1) New material from the Middle Triassic of Besano
adds considerable detail to the knowledge of the anatomy
of Tanystropheus longobardicus. Re-interpretations of the
anatomy presented in this paper are only applicable to
small-sized representatives of T. longobardicus and also
to T. meridensis, that cannot be distinguished from the
smallest specimens of T. longobardicus and is regarded
here as conspecic. Three larger-sized Tanystropheus
specimens are comprised in the new material and their
morphology is described and discussed.
2) A new reconstruction of the skull is presented, devel-
oped from a clay three-dimensional model. Major novel-
ties in the interpretation of the pre-orbital region concern
the nature of the contacts between the different elements
and the shape of the prefrontals. The shape, location and
contacts of the nasals are reconstructed by comparison
with Tanystropheus material in the PIMUZ Collections
and remain partly speculative. The contacts between
the nasals, prefrontals and frontals are only tentatively
reconstructed. The fronto-parietal plate is re-interpreted:
the lateral anges of the frontal are horizontally rather
than vertically oriented and the features distinguishing
the dorsal and ventral surfaces of the fronto-parietal plate
are described. Major difculties were encountered in the
reconstruction of the temporal region: the postorbital and
the squamosal are re-described and a new tentative inter-
pretation of their contacts with the neighbouring bones is
given. The presence of a sclerotic ring in the orbit is de-
nitely conrmed. The shape and contacts of the elements
of the lower jaw are partly re-interpreted.
83
TANYSTROPHEUS LONGOBARDICUS: RE-INTERPRETATIONS OF THE ANATOMY BASED ON NEW SPECIMENS FROM BESANO
Fig. 67 – Tanystropheus longobardicus. Watercolor: Fabio Fogliazza.
84 STEFANIA NOSOTTI
Acknowledgements
I am deeply grateful to the volunteers of the Paleon-
tological Group of Besano who excavated in the Besano
quarries for the last three decades. They provided me with
the superbly preserved material which rst gave me the
opportunity to undertake this exciting enterprise.
I warmly thank: the Soprintendenza per i Beni
Archeologici della Lombardia for assistance in obtaining
permission for excavations at the Sasso Caldo site near
Besano; the Regione Lombardia, Assessorato alle Cul-
ture, Identità, Autonomie della Lombardia for nancial
support for the preparation of the specimens and research;
Giovanni Tafuni (Ospedale Maggiore di Milano) for
the radiographs; Giuseppina Damiano (MSNM), Fabio
Fogliazza (MSNM) and Sergio Rampinelli for the skil-
ful preparation of the specimens; Roberto Appiani, Mas-
simo Demma, Nicholas C. Fraser, Heinz Lanz (PIMUZ),
Rosi Roth (PIMUZ) and Luciano Spezia (MSNM) for
the photographs; Massimo Demma and Fabio Fogliazza
(MSNM) for the line drawings, watercolors and pencils;
Massimo Demma, Simone Maganuco and Michela Mura
for the editing of illustrations; Kathleen Histon for the
translation of the dedication; Giacomo Bracchi (MSNM
and Museo Civico di Storia Naturale di Piacenza) for
careful reading of the proofs.
I extend my warmest thanks to: Hugo Bucher and
Heinz Furrer (PIMUZ) for allowing me access to the
specimens in their Collections and Heinz Lanz (PIMUZ)
for his generous assistance and hospitality during my visit
in Zürich; Rudolf Stockar (Museo Cantonale di Storia
Naturale, Lugano) for allowing me access to specimen
MCSN 4451 in his care.
I am deeply indebted to the reviewers Nicholas
C. Fraser (Virginia Museum of Natural History, Martin-
sville), Simone Maganuco (MSNM and Dipartimento di
Scienze della Terra, Università degli Studi di Firenze)
and Olivier Rieppel (Field Museum of Natural History,
Chicago) for comments, criticism and discussion, which
greatly improved the original manuscript. N. C. Fraser
and O. Rieppel also edited the English text, greatly
improving its style and the clarity of the contents. I thank
S. Maganuco also for the generous and constant support,
which helped me to overcome the numerous difculties
encountered in completing my job.
Silvio Renesto (Università degli Studi dell’Insubria,
Varese) provided helpful comments and criticism on a
very early draft of this paper.
My interpretation of the temporal region of the skull
was greatly improved through discussion with N. C. Fraser
and S. Maganuco. I proted also from discussion with
Cristiano Dal Sasso (MSNM), who offered valuable sug-
gestions and hints for further investigations. Likewise I am
indebted to Richard Forrest for discussion on the mode of
life of Tanystropheus, for sharing information on his origi-
nal work, and giving access to unpublished data and litera-
ture on plesiosaurs. Giovanni Pasini (Museo dei Fossili di
Besano) offered helpful suggestions and was always gener-
ously available for discussion. Over the years I have greatly
beneted from discussions with Rupert Wild (Staatliches
Museum für Naturkunde, Stuttgart). His outstanding mono-
graph on Tanystropheus was an important reference for my
job. Giuseppe Muscio (MFSN) shared with me information
on his original work. The friend Marco Auditore provided
much needed access to literature.
I thank for help and discussion David Martill (Uni-
versity of Portsmouth) and Lars Schmidt (Department
of Geology, University of California), as well as the fol-
lowing colleagues from MSNM: Giorgio Bardelli, Valter
Fogato, Paola Livi, Maurizio Pavesi, Michela Podestà,
Fabrizio Rigato, Stefano Scali.
My hearthly thank to Michela Mura (Graphic design,
MSNM) for her friendly assistence and skilful editing of
this monograph.
I acknowledge nancial support for the publication of
this monograph from Cinehollywood s.r.l.
Linguistic mistakes, errors of fact or interpretation,
along with any inadvertent omissions are my own.
3) A re-description of the axial skeleton is given,
mainly based on the fully articulated MSNM BES SC
265. Differences in vertebral morphology between small-
and large-sized specimens of Tanystropheus are assessed
through comparison with three-dimensionally preserved
vertebrae referred to Tanystropheus conspicuus and
Tanystropheus cf. longobardicus.
4) The superbly preserved limbs of MSNM BES SC
1018 provide for the rst time unequivocal information
on the elements of the manus and pes and their articular
relationships. A re-description of the tarsus based on
the new specimens is given. The presence of sesamoid
bones in the elbow and knee joints is reported for the
rst time.
5) Discussion on the morphology of the hindlimb
highlights the difculties in the interpretation of ter-
restrial locomotion in Tanystropheus. It is concluded
that, although the morphology of the hindlimb is not
strikingly specialized for locomotion in water, it is
more consistent with an aquatic mode of life for Tanys-
tropheus.
6) Based on the overall skeletal anatomy, Tanystroph-
eus is regarded as an aquatic protorosaur with close ter-
restrial ancestors, living in shallow waters. At present, the
more consistent model for aquatic locomotion in Tanys-
tropheus seems to be that suggested by Tschanz (1985;
1986), who envisaged Tanystropheus as an axial-subun-
dulational swimmer relying on lateral undulations of the
trunk and tail for aquatic propulsion, perhaps enhanced
by paddling with the hindlimbs. Major difculties are
encountered in the assessment of the method of capture
and processing the prey adopted by Tanystropheus. The
cheek dentition of extant pinnipeds is proposed as a new
model for the understanding of the functional signicance
of the tricuspid dentition in Tanystropheus.
7) The notion that the extremely elongated neck of
Tanystropheus was inadaptive is rejected, based also on
comparison with plesiosauroid plesiosaurs.
8) Description and coding of the phylogenetically
informative characters used in the most recent cladistic
analyses are discussed in the light of the new information
provided by the new material.
85
TANYSTROPHEUS LONGOBARDICUS: RE-INTERPRETATIONS OF THE ANATOMY BASED ON NEW SPECIMENS FROM BESANO
REFERENCES
Arthaber G. v., 1921 – Studien über Flugsaurier auf
Grund der Bearbeitung des Wiener Exemplares von
Dorygnathus Banthensis Theod. sp. Denkschriften
der Akademie der Wissenschaften in Wien, Mathema-
tisch-Naturwissenschaftliche Klasse, Wien, 97 (1919):
391-464.
Arthaber G. v., 1922 – Über Entwicklung, Ausbildung
und Absterben der Flugsaurier. Paläontologische Zeit-
schrift, Berlin, 4: 1-47.
Avanzini M. & Renesto S., 2002 – A review of Rhyncho-
sauroides tirolicus Abel, 1926 ichnospecies (Middle
Triassic: Anisian-Ladinian) and some inferences on
Rhynchosauroides trackmaker. Rivista Italiana di
Paleontologia e Stratigraa, Milano, 108 (1): 51-
66.
Bassani F., 1886 – Sui fossili e sull’età degli schisti bitu-
minosi triasici di Besano in Lombardia. Atti Società
Italiana di Scienze Naturali, Milano, 29: 15-72.
Bennett S. C., 1993 – The ontogeny of Pteranodon and
other pterosaurs. Paleobiology, Chicago, 19 (1): 92-
106.
Benton M. J., 1985 – Classication and phylogeny of the
diapsid reptiles. Zoological Journal of the Linnean
Society, London, 84: 97-164.
Benton M. J. & Allen J. L., 1997 – Boreopricea from the
Lower Triassic of Russia, and the relationships of the
prolacertiform reptiles. Palaeontology, London, 40:
931-953.
Braun J. & Reif W.-E., 1982 – A new terminology of
aquatic propulsion in vertebrates. Neues Jahrbuch für
Geologie und Paläontologie-Abhandlungen, Stuttgart,
164 (1-2): 162-167.
Braun J. & Reif W.-E., 1985 – A survey of aquatic loco-
motion in shes and tetrapods. Neues Jahrbuch für
Geologie und Paläontologie-Abhandlungen, Stuttgart,
169 (3): 307-332.
Brinkman D., 1980 – Structural correlates of tarsal and
metatarsal functioning in Iguana (Lacertilia; Iguani-
dae) and other lizards. Canadian Journal of Zoology,
Ottawa, 58: 277-289.
Brinkman D., 1981 – The origin of the crocodiloid tarsi
and the interrelationships of thecodontian archosaurs.
Breviora Museum of Comparative Zoology, Cam-
bridge, Massachusetts, 464: 1-23.
Broili F., 1915 – Beobachtungen an Tanystropheus cons-
picuus H. v. Meyer. Neues Jahrbuch für Mineralogie,
Geologie und Paläontologie, Stuttgart, II: 51-62.
Carroll R. L. & Gaskill P., 1985 – The nothosaur Pachyp-
leurosaurus and the origin of plesiosaurs. Philosophi-
cal Transactions of the Royal Society, London, B 309
(1139): 343-393.
Chapple D. G. & Swain R., 2002 – Distribution of energy
reserves in a viviparous shink: does tail autotomy
involve the loss of lipid stores? Austral Ecology, 27
(5): 565-572.
Chatterjee S., 1986 – Malerisaurus langstoni, a new diap-
sid reptile from the Triassic of Texas. Journal of Verte-
brate Paleontology, Northbrook, 6 (4): 297-312.
Cox B., 1975 – The longest-necked lizard? Nature,
London, 254: 654-655.
Dalla Vecchia F. M., 2000 – Tanystropheus (Archosauro-
morpha, Prolacertiformes) remains from the Triassic of
the Northern Friuli (NE Italy). Rivista Italiana di Pale-
ontologia e Stratigraa, Milano, 106 (2): 135-140.
Dalla Vecchia F. M., 2006 – Resti di Tanystropheus, Sau-
rotterigi e “Rauisuchi” (Reptilia) nel Triassico medio
della Val Aupa (Moggio Udinese, Udine). Gortania
– Atti del Museo Friulano di Storia Naturale, Udine,
27 (2005): 25-48.
Dalla Vecchia F. M. & Avanzini M., 2002 – New scat-
tered remains of Triassic reptiles from northeastern
Italy. Bollettino della Società Paleontologica Italiana,
Modena, 41 (2/3): 215-235.
Demes B. & Krause D. W., 2005 – Suction feeding in a
Triassic protorosaur? Science, Washington, 308: 1112-
1113.
Dilkes D. W., 1998 – The Early Triassic rhynchosaur
Mesosuchus browni and the interrelationships of
basal archosauromorph reptiles. Philosophical
Transactions of the Royal Society, London, B 353
(1368): 501-541.
Edmund A. G., 1969 – Dentition. In: The Biology of the
Reptilia. Gans C. (ed.). Academic Press, London &
New York, 1: 117–200.
Ellenberger P., 1977 – Quelques précisions sur l’anato-
mie et la place systématique trés spéciale de Cose-
saurus aviceps. (Ladinien supérieur de Montrall,
Catalogne). Cuadernos de Geología Ibérica, Madrid,
4: 169-188.
Evans S., 1988 – The early history and relationships of
the Diapsida. In: The phylogeny and classication of
the tetrapods. Benton M. J. (ed.). Clarendon Press,
Oxford, 1: 221-260.
Fernández M. & Gasparini Z., 2000 – Salt glands in a
Tithonian metriorhynchid crocodyliform and their
physiological signicance. Lethaia, Oslo, 33: 269-
276.
Ford T., 2002 – A new interpretation of the skull of Tanys-
tropheus. Journal of Vertebrate Paleontology, North-
brook, 22 (3, supplement): 53A.
Fraser N. C., Nosotti S. & Rieppel O., 2004 – A re-evalu-
ation of two species of Tanystropheus (Diapsida, Pro-
torosauria), from Monte San Giorgio, southern Alps.
Journal of Vertebrate Paleontology, Hanover, 24 (3,
supplement): 60A.
Fraser N. C. & Rieppel O., 2006 – A new protorosaur
(Diapsida) from the Upper Buntsandstein of the Black
Forest, Germany. Journal of Vertebrate Paleontology,
Northbrook, 26 (4): 866-871.
Furrer H., 1995 – The Kalkschieferzone (Upper Meride
Limestone; Ladinian) near Meride (Canton Ticino,
Southern Switzerland) and the evolution of a Middle
Triassic intraplatform basin. Eclogae geologicae Hel-
vetiae, Basel, 88: 827-852.
Gauthier J., Kluge A. G. & Rowe T., 1988 – Amniote
phylogeny and the importance of fossils. Cladistics,
London, 4: 105-209.
Gould S. J., 1974 – The origin and function of “bizarre”
structures: antler size and skull size in the “Irish Elk”,
Megaloceros giganteus. Evolution, 28: 191-220.
Gow C. E., 1975 – The morphology and relationships
of Youngina capensis Broom and Prolacerta broomi
Parrington. Paleontologia Africana, Johannesburg,
18: 89-131.
86 STEFANIA NOSOTTI
Hopson J. A., 1979 – Paleoneurology. In: Biology of the
Reptilia. Gans C. (ed.). Academic Press, London, 9A:
39-146.
Huene F. v., 1907-1908 – Die Dinosauriers der Euro-
päischen Triasformation, mit Berücksichtigung der
aussereuropäischen Vorkommnisse. Geologische und
Palaeontologische Abhandlungen, Supplement-Band
1. Gustav Fischer, Jena.
Huxley T. H., 1871 – A Manual of the Anatomy of Verte-
brate Animals. J. & A. Churchill, London.
I.C.Z.N., 1981 – Opinion 1186. Tanystropheus H. v.
Meyer, [1852] (Reptilia) conserved. Bulletin of
Zoological Nomenclature, London, 38 (3): 188-
190.
Jalil N.-E., 1997 – A new prolacertiform diapsid from the
Triassic of North Africa and the interrelationships of
the Prolacertiformes. Journal of Vertebrate Paleontol-
ogy, Northbrook, 17 (3): 506-525.
Jurcsák T., 1975 – Tanystropheus biharicus n. sp. (Rep-
tilia, Squamata) o nouă specie pentru fauna Triasică a
României. Nymphaea, Oradea, 3: 45-52.
Jurcsák T., 1976 – Noi descoperiri de reptile fosile în Tria-
sicul de la Aleşd. Nymphaea, Oradea, 4: 67-105.
Jurcsák T., 1978 – Rezultate noi în studiul saurienilor
fosili de la Aleşd (Bihor, România). Nymphaea,
Oradea, 6: 15-60.
Jurcsák T., 1982 – Occurrences nouvelles des Sauriens
mèsozoïques de Roumania. Vertebrata Hungarica,
Budapest, 21: 175-184.
Kuhn O., 1934 – Sauropterygia. In: Fossilium catalogus.
I, Animalia. Quenstedt W. (ed.). W. Junk., Berlin, Pars
69.
Kuhn O., 1963 – Sauria. In: Fossilium catalogus. I, Ani-
malia. Westphal F. (ed.). W. Junk., Berlin, Pars 104
(Supplementum I).
Kuhn-Schnyder E., 1960 – Hand und Fuss von Tanystro-
pheus longobardicus (Bassani). Eclogae geologicae
Helvetiae, Basel, 52 (1959) (2): 921-941.
Kuhn-Schnyder E., 1967 – Das Problem der Euryapsida.
Mitteilungen aus dem Paläontologischen Institut der
Universität Zürich, Zürich, 49: 335-348.
Kummer B., 1975 – Biomechanik fossiler und rezenter
Wirbeltiere. Natur und Museum, Frankfurt am Main,
105: 156-167.
LaBarbera M. & Rieppel O., 2005 – Suction feeding in a
Triassic protorosaur? Science, Washington, 308: 1112-
1113.
Li Chun, 2003 – First record of Tanystropheid Reptile
(Order Protorosauria) from the Middle Triassic of
China. Acta Geologica Sinica (English Edition), Bei-
jing, 77 (4): 419-423.
Li Chun, Rieppel O. & LaBarbera M. C., 2004 – A triassic
aquatic protorosaur with an extremely long neck. Sci-
ence, Washington, 305 (5692): 1931.
Long J. H. Jr., Schumacher J., Livingston N. & Kemp M.,
2006 – Four ippers or two? Tetrapodal swimming
with an aquatic robot. Bioinspiration & Biomimetics,
Bristol, 1: 20-29.
Massare J. A., 1988 – Swimming capabilities of Mesozoic
marine reptiles: implications for method of predation.
Paleobiology, Chicago, 14 (2): 187-205.
McHenry C. R., Cook A. G. & Wroe S., 2005 – Bottom-
feeding plesiosaurs. Science, Washington, 310 (5745):
75.
Meyer H. v., 1847-1855 – Die Saurier des Muschelkalkes:
mit Rücksicht auf die Saurier aus Buntem Sandstein
und Keuper. In: Zur Fauna der Vorwelt, zweite Abthei-
lung. Heinrich Keller, Frankfurt a.m.
Miyazaki N., 2002a – Teeth. In: Encyclopedia of
Marine Mammals. Perrin W. F., Würsig B. & The-
wissen J. G. M. (eds.). Academic Press, San Diego:
1227-1232.
Miyazaki N., 2002b – Ringed, Caspian, and Baikal Seals.
In: Encyclopedia of Marine Mammals. Perrin W. F.,
Würsig B. & Thewissen J. G. M. (eds.). Academic
Press, San Diego: 1033-1037.
Modesto S. P. & Sues H-D., 2004 – The skull of the Early
Triassic archosauromorph reptile Prolacerta broomi
and its phylogenetic signicance. Zoological Journal
of the Linnean Society, London, 140: 335-351.
Motani R., 1997 – Temporal and spatial distribution
of tooth implantations in ichthyosaurs. In: Ancient
Marine Reptiles. Callaway J. M. & Nicholls E. L.
(eds.). Academic Press, 81-103.
Münster G. zu, 1834 – Vorläuge Nachricht über einige
neue Reptilien im Muschelkalke von Baiern. Neues
Jahrbuch für Mineralogie, Geognosie, Geologie und
Petrefaktenkunde, Stuttgart: 521-526.
Muscio G., 1997 – Preliminary note on a specimen of Pro-
lacertiformes (Reptilia) from the Norian (Late Trias-
sic) of Preone (Udine, North-Eastern Italy). Gortania
- Atti del Museo Friulano di Storia Naturale, Udine,
18 (1996): 33-40.
Noè L., 2006 – The role of the plesiosaurian long neck.
A new model. Journal of Vertebrate Paleontology,
Northbrook, 26 (3, supplement): 105A.
Nopcsa F. v., 1923 – Neubeschreibung des Trias-Pte-
rosauriers Tribelesodon. Paläontologische Zeitschrift,
Berlin, 5: 161-181.
Nosotti S., 1999 – New ndings of Tanystropheus longo-
bardicus (Reptilia, Prolacertiformes) in the Middle
Triassic of Besano (Lombardy, Northern Italy). In:
Third International Symposium on lithographic lime-
stones. Renesto S. (ed.). Rivista del Museo civico di
Scienze Naturali “E. Caf”, Bergamo, 20 (supple-
mento): 117-120.
Oelrich T. M., 1956 – The Anatomy of the Head of Cteno-
saura pectinata (Iguanidae). Miscellaneous Publica-
tions, Museum of Zoology, University of Michigan,
94: 9-122.
O’Keefe F. R. & Robin F., 2001 – A cladistic analysis and
taxonomic revision of the Plesiosauria (Reptilia: Sau-
ropterygia). Acta Zoologica Fennica, 213: 1-63.
Olsen P. E., 1979 – A new aquatic Eosuchian from the
Newark Supergroup (Late Triassic-Early Jurassic) of
North Carolina and Virginia. Postilla, New Haven,
176: 1-14.
Ortlam D., 1967 – Fossile Böden als Leithorizonte für die
Gliederung des Höheren Buntsandsteins im nördlichen
Schwarzwald und südlichen Odenwald. Geologisches
Jahrbuch, Hannover, 84: 485-590.
Peters D., 2000a – A reexamination of four prolacer-
tiforms with implications for pterosaur phylogen-
esis. Rivista Italiana di Paleontologia e Stratigraa,
Milano, 106 (3): 293-336.
Peters D., 2000b – Description and interpretation of inter-
phalangeal lines in tetrapods. Ichnos, Chur & New
York, 7 (1): 11-41.
87
TANYSTROPHEUS LONGOBARDICUS: RE-INTERPRETATIONS OF THE ANATOMY BASED ON NEW SPECIMENS FROM BESANO
Peters D., 2005 – Suction feeding in a Triassic protoro-
saur? Science, Washington, 308: 1112.
Peyer B., 1931 – Die Triasfauna der Tessiner Kalkalpen II.
Tanystropheus longobardicus Bass. sp. Abhandlungen
der Schweizerischen Palaontologischen Gesellschaft,
Basel, 50: 9-110.
Premru E., 1991 – Beschreibung eines neuen Fundes von
Macrocnemus bassanii Nopcsa (Reptilia, Squamata,
Prolacertiformes) aus der Grenzbitumenzone (Anis/
Ladin) von Besano, Italien. Diplomarbeit an der
Philosophischen Fakultät II der Universität Zürich,
Zürich: 1-57.
Quenstedt W., 1963 – Clavis bibliographica. In: Fossilium
catalogus. I, Animalia. Westphal F. (ed.). W. Junk.,
Berlin, Pars 102.
Renesto S., 1994 – A new prolacertiform reptile from
the Late Triassic of Northern Italy. Rivista Italiana
di Paleontologia e Stratigraa, Milano, 100 (2): 285-
306.
Renesto S., 2005 – A new specimen of Tanystropheus
(Reptilia Protorosauria) from the middle Triassic of
Switzerland and the ecology of the genus. Rivista
Italiana di Paleontologia e Stratigraa, Milano, 111
(3): 377-394.
Renesto S. & Avanzini M., 2002 – Skin remains in a juve-
nile Macrocnemus bassanii Nopcsa (Reptilia, Prolac-
ertiformes) from the Middle Triassic of Northern Italy.
Neues Jahrbuch für Geologie und Paläontologie.
Abhandlungen, Stuttgart, 224 (1): 31-48.
Renesto S. & Dalla Vecchia F. M., 2000 – The unusual
dentition and feeding habits of the prolacertiform rep-
tile Langobardisaurus (Late Triassic, Northern Italy).
Journal of Vertebrate Paleontology, Northbrook, 20
(3): 622-627.
Renesto S., Dalla Vecchia F. M. & Peters D., 2002
– Morphological evidence for Bipedalism in the Late
Triassic Prolacertiform Reptile Langobardisaurus.
Senckenbergiana lethaea, Frankfurt am Main, 82 (1):
95-106.
Rewcastle S. C., 1980 – Form and function in lacertil-
ian knee and mesotarsal joints; a contribution to the
analysis of sprawling locomotion. Journal of Zoology,
London, 191: 147-170.
Rieber H., 1973 – Cephalopoden aus der Grenzbitumen-
zone (Mittlere Trias) des Monte San Giorgio (Kanton
Tessin, Schweiz). Schweizerischen Pa1äntologischen
Abhandlungen , Basel, 93: 1-96.
Rieppel O., 1989 – The hind limb of Macrocnemus bas-
sanii (Nopcsa) (Reptilia, Diapsida): development and
functional anatomy. Journal of Vertebrate Paleontol-
ogy, Northbrook, 9 (4): 373-387.
Rieppel O., 2000 – Middle Triassic marine vertebrates
from the northern Gondwanan shelf. Zentralblatt für
Geologie und Paläontologie, Stuttgart, I (9-10): 1269-
1284.
Rieppel O., 2001 – A new species of Tanystropheus
(Reptilia: Protorosauria) from the Middle Triassic of
Makhtesh Ramon. Israel. Neues Jahrbuch für Geolo-
gie und Paläontologie. Abhandlungen, Stuttgart, 221
(2): 271-287.
Rieppel O. & Gronowski, 1981 – The loss of the lower
temporal arcade in the diapsid reptiles. Zoological
Journal of the Linnean Society, London, 72: 203-
217.
Rieppel O., Fraser N. C. & Nosotti S., 2003 – The mono-
phyly of Protorosauria (Reptilia, Archosauromorpha):
a preliminary analysis. Atti della Società italiana di
Scienze naturali e Museo civico di Storia naturale in
Milano, Milano, 144 (II): 359-382.
Robinson P. L., 1975 – The functions of the hooked fth
metatarsal in lepidosaurian reptiles. Colloque interna-
tional C.N.R.S., Paris, 218: 461-483.
Röhl H. J., Schmid-Röhl A., Furrer H., Frimmel A., Osch-
mann W. & Schwark L., 2001 – Microfacies, geoche-
mistry and palaeoecology of the Middle Triassic
Grenzbitumenzone from Monte San Giorgio (Canton
Ticino, Switzerland). Geologia Insubrica, Lomazzo,
6 (1): 1-13.
Romer A. S., 1956 – Osteology of the Reptiles. The Uni-
versity of Chicago Press, Chicago.
Sanz J. L. & López-Martínez N., 1984 – The prolacertid
lepidosaurian Cosesaurus aviceps Ellenberger & Vil-
lalta, a claimed “protoavian” from the Middle Triassic
of Spain. Geobios, Lyon, 17: 747-753.
Schaeffer B., 1941 – The morphological and functional
evolution of the tarsus in amphibians and reptiles.
Bulletin of the American Museum of Natural History,
New York, 78 (6): 395-472.
Scheffer V. B., 1969 – Seals, Sea Lions, and Walruses.
Stanford University Press, Stanford.
Senter P., 2007 – Necks for sex: sexual selection as an
explanation for sauropod dinosaur neck elongation.
Journal of Zoology, London, 271 (1): 45-53.
Simmons R. E. & Scheepers L., 1996 – Winning by a
neck: sexual selection in the evolution of giraffe.
The American Naturalist, Chicago, 148 (5): 771-
786.
Smith A. S., 2003 – Cladistic analysis of the Plesiosauria
(Reptilia: Sauropterygia). M. Sc. unpublished Thesis,
University of Bristol, Department of Earth Sciences,
Faculty of Sciences.
Sulej T., 2004 – Krasiejow. The remarkable discovery of
Triassc pre-dinosaurs. Przygod Studio, Opole.
Taylor M. A., 1989 – Neck and neck. Nature, London,
341 (6244): 688-689.
Tschanz K., 1985 - Tanystropheus - An unusual reptilian
construction. In: Konstruktionsprinzipien lebender
und ausgestorbener Reptilien. Konzepte SFB 230.
Stuttgart, (4): 169-178.
Tschanz K., 1986 – Funktionelle Anatomie der Halswir-
belsäule von Tanistropheus longobardicus (Bassani)
aus der Trias (Anis/Ladin) des Monte San Giorgio
(Tessin) auf der Basis vergleichend morphologischer
Untersuchungen an der Halsmuskulatur rezenter
Echsen. Ph. D. Thesis Universität Zürich, ADAG
Administration & Druck AG., Zürich.
Tschanz K., 1988 – Allometry and heterochronic proc-
esses in the neck-growth of the Triassic prolacertiform
reptiles Tanystropheus and Macrocnemus. Palaeontol-
ogy, London, 31: 153-179.
Underwood G., 1970 – The Eye. In: The Biology of the
Reptilia. Gans C. & Parson T. S. (eds.). Academic
Press, London & New York, 2: 1-97.
Vickers-Rich P., Rich T., Rieppel O., Thulborn R. A. &
McClure H. A., 1999 – A Middle Triassic vertebrate
fauna from the Jilh Formation, Saudi Arabia. Neues
Jahrbuch für Geologie und Paläontologie Abhandlun-
gen, Stuttgart, 213 (2): 201-232.
88 STEFANIA NOSOTTI
Stefania Nosotti - Museo Civico di Storia Naturale di Milano, Sezione di Paleontologia, Corso Venezia 55, 20121 Milano, Italy.
e-mail: stefanianosotti@yahoo.it
Tanystropheus longobardicus (Reptilia, Protorosauria): re-interpretations of the anatomy based on new specimens from the Middle
Triassic of Besano (Lombardy, northern Italy)
Memorie della Società Italiana di Scienze Naturali e del Museo Civico di Storia Naturale di Milano
Volume XXXV – Fascicolo III
Wild R., 1974 – Die Triasfauna der Tessiner Kalkalpen
XXIII. Tanystropheus longobardicus (Bassani) (neue
Ergebnisse). Schweizerische Paläontologische Abhan-
dlungen, Basel, 95 (1973): 4-162.
Wild R., 1975 – Tanystropheus H. v. Meyer, 1855 (Rep-
tilia): request for conservation under the plenary
powers. Z.N.(S.) 2084. Bulletin of Zoological Nomen-
clature, London, 32 (2): 124-126.
Wild R., 1976 – Tanystropheus H. v. Meyer, [1852]
(Reptilia): revised request for conservation under
the plenary powers. Z.N.(S.) 2084. Bulletin of
Zoological Nomenclature, London, 33 (2): 124-
126.
Wild R., 1980a – Die Triasfauna der Tessiner Kalkal-
pen XXIV. Neue Funde von Tanystropheus (Rep-
tilia, Squamata). Schweizerische Paläontologische
Abhandlungen, Basel, 102: 4-43.
Wild R., 1980b – Tanystropheus (Reptilia: Squamata) and
its importance for stratigraphy. Mémoires de la Société
Géologique de France, n.s., Paris, 139: 201-206.
Wild R., 1987 – An example of biological reasons for
extinction: Tanystropheus (Reptilia, Squamata).
Mémoires de la Société Géologique de France, n. s.,
Paris, 150: 37-44.
Wild R. & Oosterink H., 1984 – Tanystropheus (Reptilia,
Squamata) aus dem Unteren Muschelkalk von Win-
terswijk, Holland. Grondboor en Hamer, Oldenzaal,
5: 142-148.
Zaher H. & Rieppel O., 1999 – Tooth implantation and
replacement in squamates, with special reference to
mosasaur lizard and snakes. American Museum Novi-
tates, New York, 3271: 1-19.
Zarnik B., 1925 – K ethologiji plesiosaurija, sa prinosima
mehanici kraljeznice u recentnih sauropsida. Zagreb.
