Biol. Rev. (2002), 77, pp. 333–401 " Cambridge Philosophical Society
DOI : 10.1017\S1464793102005924 Printed in the United Kingdom
333
Snake phylogeny based on osteology, soft
anatomy and ecology
MICHAEL S. Y. LEE and JOHN D. SCANLON
Department of Palaeontology, The South Australian Museum & Dept. of Environmental Biology, University of Adelaide,
North Terrace, Adelaide 5000, Australia (e-mail : lee.mike!saugov.sa.gov.au, scanlon.john!saugov.sa.gov.au)
(Received 23 April 2001 ; revised 7 February 2002)
ABSTRACT
Relationships between the major lineages of snakes are assessed based on a phylogenetic analysis of the most
extensive phenotypic data set to date (212 osteological, 48 soft anatomical, and three ecological characters).
The marine, limbed Cretaceous snakes Pachyrhachis and Haasiophis emerge as the most primitive snakes :
characters proposed to unite them with advanced snakes (macrostomatans) are based on unlikely
interpretations of contentious elements or are highly variable within snakes. Other basal snakes include
madtsoiids and Dinilysia – both large, presumably non-burrowing forms. The inferred relationships within
extant snakes are broadly similar to currently accepted views, with scolecophidians (blindsnakes) being the
most basal living forms, followed by anilioids (pipesnakes), booids and booid-like groups, acrochordids
(filesnakes), and finally colubroids. Important new conclusions include strong support for the monophyly of
large constricting snakes (erycines, boines, pythonines), and moderate support for the non-monophyly of the
‘ trophidophiids ’ (dwarf boas). These phylogenetic results are obtained whether varanoid lizards, or
amphisbaenians and dibamids, are assumed to be the nearest relatives (outgroups) of snakes, and whether
multistate characters are treated as ordered or unordered. Identification of large marine forms, and large
surface-active terrestrial forms, as the most primitive snakes contradicts with the widespread view that snakes
arose via minute, burrowing ancestors. Furthermore, these basal fossil snakes all have long flexible jaw
elements adapted for ingesting large prey (‘ macrostomy ’), suggesting that large gape was primitive for
snakes and secondarily reduced in the most basal living foms (scolecophidians and anilioids) in connection
with burrowing. This challenges the widespread view that snake evolution has involved progressive,
directional elaboration of the jaw apparatus to feed on larger prey.
Key words : Snakes, serpentes, squamates, phylogeny, cladistics, feeding, morphology.
CONTENTS
I.
II.
III.
IV.
V.
Introduction...........................................................................................................................
Snake monophyly..................................................................................................................
Snake outgroups....................................................................................................................
Terminal taxa, character coding, and data availability ........................................................
Character list .........................................................................................................................
(1) Osteology.........................................................................................................................
(a) Skull..........................................................................................................................
(b) Mandible ..................................................................................................................
(c) Dentition...................................................................................................................
(d) Hyoid apparatus .......................................................................................................
(e) Vertebrae..................................................................................................................
( f ) Pelvis and hindlimb ..................................................................................................
334
337
338
340
349
349
349
369
371
373
373
375
Michael S. Y. Lee and John D. Scanlon
334
VI.
VII.
VIII.
IX.
X.
XI.
XII.
XIII.
(2) Soft anatomy ...................................................................................................................
(a) Musculature..............................................................................................................
(b) Visceral organs.........................................................................................................
(c) Palate and nasal region............................................................................................
(d) External morphology ................................................................................................
(3) Ecological traits ...............................................................................................................
Characters not employed. ......................................................................................................
Parsimony analyses ................................................................................................................
Phylogenenetic results and evolutionary implications ............................................................
(1) Basal fossil snakes and modern snakes.............................................................................
(2) Scolecophidians and alethinophidians .............................................................................
(3) Anilioids and macrostomatans........................................................................................
(4) Xenopeltids and ‘ core macrostomatans ’.........................................................................
(5) Booids and ‘ advanced snakes ’ .........................................................................................
Sensitivity to outgroup and evolutionary assumptions ...........................................................
Acknowledgements.................................................................................................................
References ..............................................................................................................................
Appendix 1. Abbreviations ....................................................................................................
Appendix 2. Specimens examined .........................................................................................
I. INTRODUCTION
Snakes are one of the most successful and speciose
groups of vertebrates (Stanley, 1979). Despite the
conservativeness of their body form – all are longbodied and limb-reduced – they exhibit great species
richness and have invaded a broad range of
environments (e.g. Greene, 1997). It might be
expected that the combination of high taxonomic
and ecological diversity coupled with morphological
constraints might create problems for interpreting
their interrelationships (e.g. Heise et al., 1995).
However, empirical analyses of relationships between and within major snake lineages based on
morphology have shown that this view is overly
pessimistic – character congruence and phylogenetic
resolution in recent studies do not appear to be
clearly worse than in studies of other vertebrates
(e.g. Cundall, Wallach & Rossman, 1993 ; Kluge,
1993 a, b ; Underwood & Kochva, 1993 ; Slowinski,
1994, 1995 ; Wallach & Gu$ nther, 1998 ; Scanlon &
Lee, 2000 ; Tchernov et al., 2000).
Very early taxonomic works on snakes, as on all
other groups, divided snakes into grades rather than
clades, and often focused on only a few anatomical
characters or regions. Dume! ril (1806) for instance,
divided snakes into a primitive group, the Homodermes, characterized by a lack of enlarged ventral
scales, and a derived group, the He! te! rodermes,
characterized by possession of such scales. As many
of the most striking adaptations of snakes involve
their feeding, other early classifications often focused
on the jaw apparatus. Merrem (1820) placed snakes
with rigid jaws into the primitive grade Typhlini,
375
375
375
379
379
380
380
381
382
384
386
388
389
390
391
392
393
398
399
and snakes with dilatable jaws into the derived
group Glutones. Several workers (Mu$ ller, 1831 ;
Dume! ril & Bibron, 1844 ; Stannius, 1856 ; Cope,
1864, 1900) erected another dichotomy between
those snakes where the jaw is short and the
supratemporal small or absent (Microstomata,
Angiostomata or Scolecophidia) and those where
the jaw is long and suspended from the end of a
long, posteriorly projecting bone now recognised as
the supratemporal (Macrostomata or Eurystomata).
Dume! ril (1853) used the dentition to subdivide
snakes further into Opote! rodontes [worm-like forms
lacking teeth on either the maxilla or mandible – the
same group Dume! ril and Bibron (1844) had named
Scolecophidia], Aglyphodontes (non-venomous
forms with ungrooved teeth on both upper and lower
jaws), Opisthoglyphes (with grooved teeth – i.e.
fangs – at the rear of the maxilla), Prote! roglyphes
(with grooved teeth on the front of the maxilla,
usually followed by small smooth teeth), and
Solenoglyphes (with only a pair of large hollow fangs
on the reduced maxilla).
The most prolific early worker was Cope (e.g.
1893, 1894, 1895), whose numerous contributions
culminated in a posthumously published monograph
(1900) in which snakes were classified mainly on
lung and hemipenial morphology, with brief consideration of maxillary mobility and reduction of the
pelvis and hindlimb. Cope suggested that the large
second lung, pelves and hindlimbs in booids (‘ peropods ’) indicated their basal position within snakes.
Anilioids (‘ tortricines ’) were also quite primitive,
with only slightly smaller lungs, pelves and hindlimbs. He noted that hemipenial characters were
Snake phylogeny
generally too variable to diagnose large lineages, and
emphasised their utility to diagnose genera or groups
of similar genera.
Boulenger (1893) and Janensch (1906) made
further important contributions to snake systematics
based on osteology and palaeontology. Nopcsa
(1923) integrated Boulenger’s classification of living
snakes with Janensch’s classification of fossil forms,
and proposed the grade concepts of Cholophidia,
Angiostomata, and Alethinophidia. Cholophidia
included several groups of extinct, aquatic snakes
which he regarded as ancestral to living groups (i.e.
stem snakes) ; Angiostomata included blindsnakes,
regarded as very specialized but also the most
primitive living snakes ; and Alethinophidia included
all the remaining families, divided informally into
‘ primitive ’ groups (Xenopeltidae, Ilysiidae, Uropeltidae, and Boidae sensu Boulenger) and ‘ specialized ’
groups (Colubridae, Viperidae and Amblycephalidae). Nopcsa (1923) accompanied this classification with a branching diagram of proposed
relationships (effectively, a cladogram) and explicit
hypotheses of evolutionary processes, placing snakes
within the radiation of varanoid ‘ lizards ’. Hoffstetter
(1939, 1955) continued to recognize Cholophidia
and Angiostomata (the latter with Dume! ril and
Bibron’s name Scolecophidia), and added a new
layer of formal classification by naming the two
grades within Alethinophidia, Henophidia and Caenophidia. Romer (1956) accepted the same basic
groups but used names denoting lower ‘ rank ’, at the
family or superfamily level ; he recognized superfamilies Typhlopoidea (l Angiostomata or Scolecophidia), Booidea (l Henophidia), and Colubroidea
(l Caenophidia). However, following McDowell
and Bogert (1954), he placed the extinct cholophidians ‘ provisionally ’ within Varanoidea as a single
family Palaeophidae, stating that ‘ there is no certain
evidence that they were snakes rather than advanced
snake-like lizards ’ (p. 568).
The ‘ modern ’ age of snake systematics can be
considered to have been initiated by Underwood
(1967), who produced a comprehensive classification
of snakes based on what remains the most comprehensive set of phenotypic features considered to
date, including external and internal morphology,
histology, cytology and biochemistry. This study was
published before cladistic analysis was widely used,
and as a result there was no accompanying character matrix or branching diagram. However, the
favoured arrangement can be reconstructed from the
text and the proposed taxonomy. Scolecophidians
were interpreted as monophyletic, and the most
335
basal snake lineage. Within scolecophidians, typhlopids and anomalepidids were interpreted as more
closely related than either were to leptotyphlopids.
The remaining extant snakes, henophidians (booids)
and caenophidians (colubroids), formed a clade, the
Alethinophidia. Henophidians were interpreted as a
paraphyletic and largely unresolved assemblage
basal to a monophyletic Caenophidia. Some resolution within the Henophidia was proposed,
however. Aniliids and cylindrophids (including
Anomochilus) were proposed to form a clade. Boines,
tropidophiids, and bolyeriines were recognized as
another clade, with pythons and Loxocemus being
close relatives. Although formally placed within the
henophidian grade (in contrast to earlier systems),
Acrochordus was recognized as being more closely
related to caenophidians than to other henophidians.
McDowell (e.g. 1972, 1974, 1975, 1979, 1987)
revised snake classification at a similarly comprehensive level based on morphology. Among scolecophidians, he proposed that Typhlopidae and
Leptotyphlopidae shared a common ancestry after
divergence from within Anomalepididae. In place
of Henophidia (or Boidae of older classifications
such as Boulenger’s) he recognized a greater number
of distinct alethinophidian lineages (superfamilies
Anilioidea, Booidea, Tropidophioidea, Bolyerioidea,
Acrochordoidea) of equal rank with Colubroidea,
regarding the new groupings as ‘ surviving minor
lines from the early radiation that also gave rise to
the Boidae and Colubridae (sense of Boulenger, 1893
and Romer, 1956) ’ ; recognition of such groups at
superfamilial level had already been suggested by
Hoffstetter (1968). McDowell’s (1975, 1987) concept
of Anilioidea includes Xenopeltis and Loxocemus, considered closer to boids in all other classifications.
The most recent analyses of snake relationships
include morphological studies by Rieppel (1988 ; no
explicit numerical analysis), Kluge (1991 ; details
not yet published), Cundall et al. (1993) and a
molecular study by Heise et al. (1995). An allozyme
study (Dowling et al., 1996) has been shown to
contain virtually no phylogenetic information
(Buckley, Kearney & De Queiroz, 2000) and will
not be discussed further. The morphological studies
agreed on the basal position and monophyly of
the scolecophidians, but disagreed on relationships
between typhlopids, leptotyphlopids and anomalepidids. The molecular study also recognized the
basal position of scolecophidians among living
snakes, but weakly suggested their paraphyly. All
morphological studies agreed that henophidians
were a paraphyletic assemblage of snakes that fell
336
between scolecophidians and colubroids. However,
the proposed interrelationships among the henophidians differed widely between studies ; with
relationships among anilioid and booid taxa, and the
monophyly of each group, being actively debated.
Rieppel (1988) revived the name Macrostomata for
a clade comprising alethinophidians other than
anilioids ; this concept has been widely accepted and
there is general agreement on its monophyly and
content with regard to extant taxa, though as
discussed below the inclusion of certain fossils is
contentious.
The fossil record of snakes has until recently not
shed much light on these controversies, since it is
dominated by vertebrae which contain only a limited
number of phylogenetically informative characters
(Hoffstetter & Gasc, 1969 ; Gasc, 1976 ; Underwood,
1976 ; Rage, 1984, 1987 ; McDowell, 1987 ; Szyndlar
& Bo$ hme, 1996 ; Averianov, 1997). Most completely
known forms (i.e. those with good cranial material)
are very similar to living members of modern
‘ families ’ and thus shed little light on the interrelationships between these families. However, there are
a small number of well-preserved forms which are
clearly different from all modern snakes, and the
relationships of these remain contentious. Dinilysia
has been interpreted as an aniliid (l ‘ ilysiid ’ ;
Woodward, 1901), a basal ‘ booid ’ that retains some
primitive features found in aniliids (Romer, 1956 ;
Estes, Frazzetta & Williams, 1970 ; Rage, 1977,
1987), a stem snake or ‘ cholophidian ’ (McDowell,
1972, 1974, 1987 ; Scanlon, 1996), a basal scolecophidian (raised as a possibility although not formally
supported by McDowell, 1987), or a basal alethinophidian (or equivalently, sister taxon to Alethinophidia after divergence from Scolecophidia ; Rieppel,
1988 ; Kluge, 1991 ; Zaher, 1998 ; Tchernov et al.,
2000). Madtsoiids have been interpreted as booids
(Simpson, 1933 ; Underwood, 1976 ; Rage, 1984,
1987 ; Barrie, 1990), basal alethinophidians (Scanlon, 1993 ; Rage, 1998), or basal snakes (clearly
implied by Hoffstetter, 1961, despite his classification
of Madtsoiinae within Boidae ; McDowell, 1987 ;
Scanlon, 1996 ; Scanlon & Lee, 2000). The limbed
marine form Pachyrhachis has been described as a
probable aigialosaurian lizard (Haas, 1979 ; Rage,
1987 ; Rieppel, 1988), a stem snake (McDowell,
1987 ; Scanlon, 1996 ; Caldwell & Lee, 1997 ; Lee &
Caldwell, 1998 ; Scanlon & Lee, 2000), and a
macrostomatan snake (Zaher, 1998 ; Rieppel &
Zaher, 2000 b ; Tchernov et al., 2000). The very
similar Haasiophis was recently described as a
macrostomatan (Tchernov et al., 2000), but this
Michael S. Y. Lee and John D. Scanlon
interpretation is questionable (see below). The
recently described Podophis has been considered to
provide additional support for the interpretation of
these forms as stem snakes rather than macrostomatans (Rage & Escuillie! , 2000). Janensch (1906)
assigned Archaeophis proavus to its own family without
clear affinities, but most authors have regarded it,
more recently along with A. turkmenicus, as close
relatives or members of the Palaeophiidae (Nopcsa,
1923 ; Hoffstetter, 1955 ; Romer, 1956 ; Rage, 1984,
1987 ; McDowell, 1987 ; Tatarinov, 1988 ; Averianov, 1997). However, the affinities of palaeophiids
themselves are uncertain, as they are known (apart
from Archaeophis) almost entirely from vertebrae.
Palaeophiids (and by implication Archaeophis) have
been interpreted as stem snakes (Nopcsa, 1923 ;
Hoffstetter, 1939 ; McDowell, 1972), as either snakes
or ‘ advanced snake-like lizards ’ (McDowell &
Bogert, 1954 ; Romer, 1956), booids (Hoffstetter,
1959 ; Rage, 1984, 1987), or acrochordoids
(McDowell, 1979, 1987).
Thus, for the last century there has been agreement on the broad outline of snake phylogeny, with
scolecophidians being the most basal extant forms,
henophidians representing an intermediate grade
and caenophidians being the most derived clade.
However, the relationships within each of these
assemblages remain contentious, and no scheme is
widely accepted. Some systems have avoided the
problem of henophidian paraphyly by excluding
caenophidians from otherwise broad analyses of
henophidians (e.g. Underwood, 1976 ; Wallach &
Gu$ nther, 1998). Furthermore, the affinities of important fossil taxa remain debated : in particular,
whether forms such as Pachyrhachis, Haasiophis,
Dinilysia, madtsoiids and Archaeophis are basal to all
living snakes (i.e. stem snakes) and thus primitive
forms that might shed light on snake origins (Coates
& Ruta, 2000). McDowell (1987) is the only recent
author explicitly to have recognized such an assemblage of stem snakes that lie outside the radiation
of extant forms, to which he applied Nopcsa’s (1923)
term Cholophidia. Most other authors have instead
tried to ‘ shoehorn ’ all snakes into either the
Scolecophidia or Alethinophidia, and possible stem
snakes such as Dinilysia and madtsoiids were thus
included in Alethinophidia without rigorous phylogenetic analysis, e.g. based on overall similarity to
certain booids (e.g. Simpson, 1933 ; Rage, 1984,
1987) or single characters such as ‘ quadrate rotation ’ (Rieppel, 1988). However, some early fossils
have been described by Rage and colleagues as
‘ Serpentes incertae sedis ’ (rather than being placed
Snake phylogeny
into either Alethinophidia or Scolecophidia), showing that some workers are open to the possibility that
these fossils could lie outside the scolecophidianalethinophidian dichotomy, and thus be stem snakes
(Rage & Prasad, 1992 ; Rage & Richter, 1994 ; Rage
& Werner, 1999).
Here, a phylogenetic analysis of snakes is carried
out based on all informative traits that could be
identified during examination of skeletal morphology of all snake lineages. Osteological traits used by
previous workers are also reevaluated and, if clearly
defined and cladistically informative, included.
Recent soft anatomical data sets (Cundall et al.,
1993 ; Wallach & Gu$ nther, 1998) are also compared
and combined with these skeletal data. The study
was initiated at this time for several reasons. Firstly,
there has been much recent work on snake phylogeny, with no clear consensus emerging. Also, the
recent work on the anatomy of early fossil snakes
Pachyrhachis (Scanlon, 1996 ; Lee & Caldwell, 1998),
Haasiophis (Tchernov et al., 2000) and madtsoiids
(Barrie, 1990 ; Scanlon, 1996, 1997 ; Rage, 1998 ;
Scanlon & Lee, 2000), means that they are now
much better known, and allows their inclusion.
Furthermore, recent analyses of relationships within
diverse snake lineages (e.g. colubroids, booids) now
allow the primitive condition in these clades to be
inferred with more confidence, thus potentially
resolving their affinities to other groups more
robustly. Finally, new analyses of squamate phylogeny based on several different sets of phenotypic
and molecular traits have supported the view that
the nearest relatives to snakes are varanoids and in
particular, mosasauroids (Schwenk, 1988, 1993 ;
Scanlon, 1996 ; Lee, 1998, 2000 a ; Caldwell, 1999).
Identification of the nearest outgroups to snakes
therefore allows a more accurate interpretation of
the direction of character evolution within snakes.
II. SNAKE MONOPHYLY
Although there are many superficially snake-like,
limb-reduced tetrapods, snakes are diagnosed by a
large number of highly distinctive derived characters, and their monophyly has been long recognized.
Other limbless tetrapods were removed from snakes
by systematists of the early 19th century. Caecilians
were recognized as distinct by Oppel (1811), legless
lizards by Mu$ ller (1831), and amphisbaenians by
Gray (1825) (see review in Rieppel, 1988). The
modern concept of snakes, at least with respect to
living taxa, has been almost universally accepted for
over a century. The only significant challenge to this
337
notion was the proposal that typhlopids are not
closely related to other snakes (McDowell & Bogert,
1954 ; Bogert, 1961 ; Goin & Goin, 1962) ; however,
the proposed evidence supporting this separation has
been strongly refuted (Underwood, 1957, 1967 ;
McDowell, 1967, 1974) and this arrangement is no
longer accepted by any workers. There is now also
consensus that certain limbed fossil taxa (notably
Pachyrhachis and Haasiophis) are closely related to, if
not nested within, living snakes (see Coates & Ruta,
2000). We thus use the vernacular term ‘ snakes ’ to
include these fossil taxa, so that the taxonomic
content of ‘ snakes ’ remains stable regardless of
which hypothesis is accepted. A more restricted
definition of ‘ snakes ’ (i.e. a crown-clade definition)
is not adopted because it generates instability with
respect to the inclusion of these problematic fossil
taxa (see also Lee, 2001).
All snakes, so defined, form a well-corroborated
clade whose monophyly has never been seriously
questioned in recent decades. Listed below are the
derived features that are present in snakes but absent
in other squamates and in rhynchocephalians. As
with most diverse clades, it is difficult to find
autapomorphies that have not been lost or modified
in some aberrant lineages (Hoffstetter, 1968). Hence,
some of these characters might not be present in all
snakes, but the isolated absences are most parsimoniously interpreted as subsequent elaborations or
secondary losses. The most detailed discussions of
these autapomorphies remain those of Underwood
and Bellairs (Bellairs & Underwood, 1951 ; Underwood, 1967, 1970 ; Bellairs, 1972) ; recent work
(cited below) has identified some additional osteological traits.
Skeletal traits
Descending process of frontal meeting sphenoid,
i.e. ‘ parabasisphenoid ’ (Bellairs & Underwood,
1951 ; Lee & Caldwell, 1998).
Descending flanges of parietal enclosing trigeminal ganglion, and ophthalmic branch of trigeminal
nerve (Bellairs & Underwood, 1951 ; Underwood,
1967 ; Bellairs, 1972 ; Bellairs & Kamal, 1981). In
other lepidosaurs, the parietal flanges are either
absent or do not enclose the trigeminal ganglion and
nerve. The position of the neural structures cannot
be determined in fossil taxa, but based on the
similarity of the skeletal elements, can be inferred to
be similar to those in extant snakes.
Crista circumfenestralis present around base of
stapes, housing extension of perilymphatic system
338
(Estes et al., 1970 ; Rieppel, 1988). Apparent reversal
in anomalepidids. The small crista in Acrochordus,
sometimes interpreted as primitive and lizard-like,
has recently been interpreted as uniquely specialized
(Rieppel & Zaher, 2001).
Postorbital with long ventral process forming
entire posterior margin of orbit (Lee & Caldwell,
1998). This character is tentative as it is only
applicable in snakes where the postorbital is present
and well developed (e.g. Pachyrhachis, Dinilysia,
booids). In other squamates, and rhynchocephalians, the postorbital, when present, never has a
long ventral process.
Mobile septomaxilla-maxilla contact (Lee &
Caldwell, 1998).
Vomer medial to palatine (Lee & Caldwell, 1998).
Palatine with long, transverse process that closely
approaches its counterpart in the midline (Lee &
Caldwell, 1998).
Dentary with two or fewer mental foramina (Lee
& Caldwell, 1998). Apparent reversal in some
madtsoiids, Acrochordus and a few colubroids, which
possess up to three foramina (Hoffstetter, 1960 ;
Hoffstetter & Gayrard, 1965 ; Cundall, 1981 ; Scanlon, 1996).
Teeth ankylosed to the rims of discrete sockets
(Lee & Caldwell, 1998). In other squamates and in
rhynchocephalians discrete sockets are either absent,
or when present (mosasauroids) the teeth are deeply
implanted (see Scanlon & Lee, 2000).
Epiphyses absent on axial and appendicular
skeleton (Haines, 1969 ; Rieppel, 1988).
Shoulder girdle and forelimb completely absent
(Bellairs & Underwood, 1951 ; Underwood, 1967).
All other squamates and rhynchocephalians, even
highly limb-reduced forms, retain at least vestiges of
the shoulder girdle (Camp, 1923).
Maxilla barely, if at all, overlapping the lateral
surface of the prefrontal (McDowell, 1974).
Soft anatomical traits
Trabeculae cranii partly (scolecophidians) or
completely (alethinophidians) separate in the orbital
region. In scolecophidians the trabeculae fuse for a
short distance anterior to the frontals ; in alethinophidians the trabeculae are fully separated by a
sagittal keel on the parasphenoid rostrum. Other
lepidosaurs have trabeculae extensively fused in the
orbital region, forming a cartilaginous interorbital
septum (Bellairs & Underwood, 1951 ; Underwood,
1967 ; McDowell, 1967 ; Bellairs, 1972 ; Rieppel,
1983, 1988).
Michael S. Y. Lee and John D. Scanlon
Rods functional and densely present in retina
(Walls, 1940 ; Underwood, 1970). Lizards lack rods,
while the putative ‘ rods ’ in Sphenodon appear to be
modified cones (Underwood, 1970) and, furthermore, are sparse and functionally insignificant
(Estes, De Queiroz & Gauthier, 1988).
Ciliary body of eye lacking muscles (Walls, 1940 ;
Underwood, 1970).
Gall bladder lies far behind liver, with the
connecting (cystic) duct being long (Underwood,
1967).
Right systemic arch much larger than left (Bellairs
& Underwood, 1951 ; direction of asymmetry incorrectly stated in Estes et al., 1988).
Right kidney distinctly anterior to left kidney
(Bellairs, 1972).
Many further likely autapomorphies of snakes
have been noted by these and later workers (Romer,
1956 ; Dowling & Duellman, 1978 ; Estes et al., 1988 ;
Rieppel, 1988 ; Scanlon, 1996 ; Lee, 1998), but as
acknowledged in these papers, these traits occur in
some other squamates (and thus might be synapomorphies of a more inclusive clade consisting of
snakes and some ‘ lizard ’ lineages), and\or occur in
rhynchocephalians (and thus might be primitive for
squamates as a whole). These traits include some
‘ classic ’ snake features such as the retractile, deeply
forked tongue, loss of the external ear, loss of the
temporal arches, mobile symphysial tips, body
elongation, and reduction of the pelvic girdle and
hindlimb.
III. SNAKE OUTGROUPS
Until recently, cladistic investigations of snake
phylogeny were hampered by their uncertain position within Squamata, and resulting uncertainty
over the most appropriate outgroup to use. Snakes
have been proposed to be most closely related to
scincids (e.g. Senn & Northcutt, 1973), amphisbaenians (Rage, 1982), a dibamid-amphisbaenian
clade (Hallermann, 1998 ; see also Greer, 1985 ;
Evans & Barbadillo, 1998 ; Rieppel & Zaher, 2000 b),
pygopodids (Iordansky, 1990), varanoids (Nopcsa,
1923 ; Camp, 1923 ; McDowell & Bogert, 1954 ;
McDowell, 1967, 1972 ; Bellairs, 1972 ; Schwenk,
1988, 1993 ; Lee, 1998), or placed outside lizards
altogether (Boulenger, 1891 ; Hoffstetter, 1962,
1968 ; Underwood, 1970 ; Rieppel, 1983, 1988).
These possible outgroups are very different from
each other, and adoption of any particular outgroup
might have a major impact on inferred relationships
Snake phylogeny
339
M
os
as
au
ro
id
ea
A
us
a
ot
m
r
n
e
us
a
th
od
an
l
r
n
Va
La
He
Ingroup (Snakes)
A
B
C
D
E
A
m
ph
is
ba
en
D
ib
ia
am
id
ae
B
A
Ingroup (Snakes)
B
C
D
E
Fig. 1. The two alternative hypotheses of the position of
snakes within squamates. (A) Varanoid relationships
(e.g. McDowell, 1972 ; Schwenk, 1993 ; Lee, 2000 a).
(B) Amphisbaenian-dibamid relationships (Hallermann,
1998 ; Zaher and Rieppel, 1999).
within snakes. For instance, if the small, worm-like
dibamids were used to infer the primitive condition
in snakes, many characters correlated with small size
and burrowing would be primitive for snakes – this
would predispose scolecophidians to emerge as the
most basal snakes. By contrast, if the large, macrophagous varanoids were used, this would conversely
predispose ‘ booid-like ’ taxa to be the most primitive
snakes.
The only two ideas currently widely accepted are
that snakes are related to varanoid lizards, or are
related to amphisbaenians and dibamids. Several
recent cladistic studies imply that snakes are most
closely related to varanoids, and in particular, to the
marine mosasauroids (Schwenk, 1988, 1993 ; Scanlon, 1996 ; Cooper, 1997 ; Lee, 1998). Tchernov et al.
(2000) agree that the nearest relatives of snakes are
‘ varanoids ’, which they define as Lanthanotus, Varanus, and mosasauroids (contra the prior definition
which also includes Heloderma ; Estes et al., 1988).
However, they did not specify whether mosasauroids
were thought to be more closely related to snakes or
to terrestrial varanoids. Caldwell (1999) concurs on
the snake-mosasaur relationship, but excludes both
from varanoids. By contrast, some recent studies
have suggested that amphisbaenians and dibamids
might be the nearest relatives of snakes (Hallermann,
1998 ; Rieppel & Zaher, 2000 b). However, the
characters supporting this arrangement appear to be
largely correlated with limb reduction, body elongation and miniaturization ; addition of fossils disrupts this grouping (Lee & Caldwell, 2000 ; Coates &
Ruta, 2000).
Thus, while varanoids appear to be the most
appropriate outgroup for snakes, an amphisbaeniandibamid outgroup cannot yet be totally discounted.
For this reason, separate analyses were undertaken
here using either possible outgroup, to see whether
outgroup assumptions affected inferred relationships
within the ingroup (snakes). The ‘ varanoid ’ outgroup arrangement assumed in this analysis is shown
in Fig. 1 A ; supporting evidence is presented in the
studies cited above. The first (most proximal)
outgroup is Mosasauroidea, the next outgroup is
Varanidae (Varanus and Lanthanotus), and the third
outgroup is Heloderma. Based on this phylogenetic
hypothesis and the distribution of character states in
these taxa, a composite ‘ varanoid ’ root was constructed. For some characters the primitive character
state was uncertain due to variability in the
outgroups : in such cases the varanoid root was coded
with all states which might conceivably have been
primitive. In other cases, polarity could not be
determined because characters were inapplicable
in all of the outgroups : in such cases the varanoid
root was coded as unknown (?). ‘ Dolichosaurs ’, a
morphologically diverse assemblage of small marine
squamates, are also closely related to mosasauroids,
and perhaps even more closely to snakes (as first
suggested by Nopcsa, 1908), but are not employed
here as outgroups since they are still only poorly
known (e.g. Lee & Caldwell, 2000).
The ‘ amphisbaenian-dibamid ’ outgroup arrangement adopted is shown in Fig. 1 B ; with the snake
tree being rooted with an amphisbaenian-dibamid
clade ( fide Hallermann, 1998 ; Rieppel & Zaher,
2000 b). More distant outgroups were not easily
identifiable since the position of the amphisbaeniandibamid-snake clade relative to other squamates
differs in these studies, being either basal to
scleroglossans (Hallermann, 1998) or nested within
anguimorphs (Rieppel & Zaher, 2000 b).
Although the character codings for all the outgroup taxa are presented, these outgroups were
never all employed at once, because this led to
problems with calculation of branch support via
constraint trees. Rather, the ingroup trees were
340
rooted either with the single composite varanoid
outgroup (with the specific varanoid taxa, and the
amphisbaenian-dibamid outgroup, being deleted),
or with the amphisbaenian-dibamid clade (with all
varanoid taxa being deleted). These two very
different outgroup arrangements provide a severe
test of whether the retrieved snake phylogeny is
affected by different outgroup assumptions. However, the relative validity of these alternative
outgroup hypotheses lies outside the scope of this
analysis, and can only be evaluated in the context of
a phylogenetic analysis of all squamates.
IV. TERMINAL TAXA, CHARACTER CODING,
AND DATA AVAILABILITY
Recent morphological cladistic studies of snakes
(Kluge, 1991 ; Cundall et al., 1993 ; Scanlon & Lee,
2000 ; Tchernov et al., 2000) have used as terminal
taxa extant snake groups for which there is good
evidence for monophyly. These taxa (often corresponding to ‘ families ’ in recent systems) are mostly
again used here, and the relevant studies that provide
diagnoses cited below. In addition to the extant
taxa, fossil snake lineages that are represented by
good cranial material (Pachyrhachis, Haasiophis,
madtsoiids, Dinilysia) are also included in the
analysis. Forms such as Simoliophis, Lapparentophis,
Podophis, Pachyophis, Mesophis, palaeophiids and Coniophis are known largely from isolated vertebrae
and\or poorly preserved body fossils, and thus not
included. The only well-preserved specimen of
Archaeophis, in the Palaeontological Institute (Moscow) remains to be described in detail and could not
be examined for this study ; hence this taxon also was
not included. A preliminary evaluation of all these
taxa, based on published descriptions, revealed that
they consisted of over 70 % missing data and their
inclusion did not change the relationships between
the other better known forms.
Nearly all the terminal taxa considered here are
very morphologically uniform with respect to the
characters used in this study (and often of very low
taxonomic diversity as well), and were thus invariant
for most characters. The exception is Colubroidea ;
fortunately, the position of colubroids high in the
snake tree and close to acrochordids is wellcorroborated (see Section VII(5)) and not of major
concern here. For characters variable within terminal taxa, coding follows the ‘ inferred ancestral
condition ’ (l ‘ groundplan ’) approach (Yeates,
1995 ; Wiens, 1998 ; Bininda-Edmonds, Bryant &
Russell, 1998). The ancestral condition in the
Michael S. Y. Lee and John D. Scanlon
terminal taxon is inferred based on the distribution
of character states within the best available phylogeny for that terminal taxon (see below). This
approach has been shown to provide more reliable
results than coding variable taxa with all observed
states, regardless of the distribution of the character
states, or by simply assuming that the most common
state is primitive, even if it does not occur in any
basal lineages (Bininda-Edmonds et al., 1998 ; Wiens,
1998). Where a character is highly variable in basal
members of the terminal taxon, all the states which
occur in these basal members (and thus might be
primitive) are entered. It should thus be emphasized
that each taxon is coded with the inferred ancestral
condition(s), and derived members of any particular
taxon might have character states other than those
coded.
Pachyrhachis. Diagnosis : Lee & Caldwell (1998).
Monotypic.
Haasiophis. Diagnosis : Tchernov et al. (2000).
Monotypic.
Dinilysia. Diagnosis : Estes et al. (1970), McDowell
(1987). Monotypic.
Madtsoiidae. Diagnosis : Hoffstetter (1961), Scanlon (1992, 1996), Rage (1998). Eight (or nine)
named genera and 14 (or 15) species, mostly
represented entirely by vertebrae and thus poorly
known. Cranial character codings for madtsoiids are
derived mainly from Wonambi, the most completely
known taxon (Scanlon & Lee, 2000).
Leptotyphlopidae. Diagnosis : List (1966), Underwood (1967), Wallach & Ineich (1996). Two genera
with around 80 species, all adequately known species
morphologically uniform.
Typhlopidae (Figs 2 A, 3 A, 4 A). Diagnosis :
Underwood (1967), McDowell (1974, 1987), Rage
(1987), Rieppel (1988), Wallach & Ineich (1996).
Seven genera with around 220 species, all adequately
known species morphologically uniform.
Anomalepididae. Diagnosis : Haas (1964, 1968),
List (1966), Robb & Smith (1966), Wallach &
Ineich (1996). Four genera with 15 species, all
adequately known species morphologically uniform.
Cylindrophis (Figs 2 B, 3 B, 4 B). Diagnosis : Cundall
et al. (1993). A single genus with seven species, all
morphologically uniform.
Anilius. Diagnosis : McDowell (1975, 1987). Monotypic.
Uropeltidae. Diagnosis : Underwood (1967),
McDowell (1975, 1987 – as Uropeltinae), Dowling
& Duellman (1978). This taxon is speciose (approximately 44 species in eight genera) but morphologically rather uniform ; for variable characters, the
Snake phylogeny
341
pm
B
A
pm
C
pm
sm
sm
na
prf
na
prf
prf
mx
fr
na
fr
mx
pof
mx
fr
pal
ec
ec
pa
pt
pa
pa
pro
pro
qa
so
pt
so
eo
qa
pro
qa
so st
eo
eo
pm
D
pm
E
F
st
pm
sm
sm
sm
na
na
prf
prf
na
spo
mx
mx
fr
pof
pa
prf
fr
fr
mx
fr
pof
pof
pa
ec
ec
ec
pa
pt
st
pro
pro
st
pro
so
so
eo
st
qa
so
eo
qa
eo
qa
pt
Fig. 2. Skulls of six divergent lineages of snakes in dorsal view. (A) Ramphotyphlops sp. (Typhlopidae), (B) Cylindrophis ruffus, (C) Xenopeltis unicolor, (D) Liasis stimsoni orientalis (Pythoninae), (E) Bolyeria multocarinata (Bolyeriinae),
(F) Xenodermus javanicus (Colubroidea). Based on (A) AMS R19116, (B) AMNH 85647, FMNH 13100, (C) NMNH
122782, (D) QM J28416 (E) BMNH 70n11n30n4A, (F) Bogert (1964). See Appendix 1 for anatomical and institutional
abbreviations.
phylogeny of Cadle et al. (1990) has been used to
infer the primitive state.
Anomochilus. Diagnosis : Cundall & Rossman
(1993), Cundall et al. (1993). A single genus, with
two very similar species.
Loxocemus. Diagnosis : Underwood (1967),
McDowell (1975, 1987). Monotypic.
Xenopeltis (Figs 2 C, 3 C, 4 C). Diagnosis : Underwood (1967), McDowell (1975, 1987). A single
genus with two very similar species.
Michael S. Y. Lee and John D. Scanlon
342
A
pm
C
B
pm
pm
sm
na
sm
prf
vo
vo
sm
vo
mx
mx
fr
mx
pal
pal
pal
ec
ec
pa
pbs
pt
pt
pt
pbs
pbs
qa
pro
bo
qa
qa
bo
op
bo
op
op
pm
D
F
E
pm
pm
sm
vo
pm
na
sm
sm
vo
vo
mx
pal
mx
mx
pal
pal
mx
bs
ec
ec
bs
ec
bs
pt
pt
pt
bo
bo
qa
qa
bo
bo
qa
st
op
Fig. 3. Skulls of six divergent lineages of snakes in ventral view. (A) Ramphotyphlops sp. (Typhlopidae), (B) Cylindrophis ruffus, (C) Xenopeltis unicolor, (D) Liasis stimsoni orientalis (Pythoninae), (E) Bolyeria multocarinata (Bolyeriinae),
(F) Xenodermus javanicus (Colubroidea). Sources as in Fig. 2. See Appendix 1 for abbreviations.
Pythonines (Figs 2 D, 3 D, 4 D). Diagnosis :
McDowell (1975), Kluge (1993 a). Eight genera
with 31 species (some recently recognized : Harvey et
al., 2000), moderately diverse morphologically. For
variable characters, the phylogeny of Kluge (1993 a)
has been used to infer the primitive state ; the
alternative arrangement of Underwood & Stimson
(1990) gives different codings for a few characters
but makes very little difference to the analyses.
Erycines. Diagnosis : Hoffstetter & Rage (1972),
Snake phylogeny
343
A
pa
D
fr
prf
st
na
pro
pm
qa
pbs
mx
pt
qa
sta
pt
cor
E
pro
so
pof?
pa
pbs
pro pa
eo so
pof
fr
prf
st
pt
mx
den
prf
na
sm
pm
qa
ec
com
B
fr
ec
cor
sm na
st
sta
qa
pm
mx
mx
pbs ec mx
pt
den
lat cr
den
com
com
F
C
pro
so
pa
prf
na
sm
pm
VII
den
spl
pof fr spo
pro
sm
cor
com
pa
so
st so pro pa
pof
fr
prf
na
qa
fr
prf
st
sta
pm
na
sm
pm
qa
pt pbs
ec
mx
pt
sm
ec
mx
com
cor
den
den
com
Fig. 4. Skulls of six divergent lineages of snakes in right lateral view. (A) Ramphotyphlops sp. (Typhlopidae), (B) Cylindrophis ruffus, (C) Xenopeltis unicolor, (D) Liasis stimsoni orientalis (Pythoninae), (E) Bolyeria multocarinata (Bolyeriinae),
(F) Xenodermus javanicus (Colubroidae). Sources as in Fig. 3 except (C) based on BMNH 1947n1n1n10, 1947n1n1n12. See
Appendix 1 for abbreviations.
Kluge (1993 b). Thirteen extant species assigned to
two to four genera, moderately diverse. Also numerous fossil forms, not considered here as they are
very incompletely known. We provisionally accept
Calabaria as an erycine (Kluge, 1993 b) rather than a
pythonine as in most other classifications, but not its
synonymy with Charina. For variable characters, the
phylogeny of Kluge (1993 b) has been used to infer
the primitive state.
Boines. Diagnosis : McDowell (1979), Kluge
(1991). Five genera with 35 species, moderately
diverse morphologically. For variable characters,
the phylogeny of Kluge (1991) has been used to infer
the primitive state.
Tropidophiinae. Diagnosis : McDowell (1975),
Zaher (1994). Two genera, one monotypic and one
with approximately 19 species, morphologically
uniform.
Ungaliophiinae. Diagnosis : McDowell (1975),
Zaher (1994). Two genera with three species,
morphologically uniform.
Bolyeriinae (Figs 2 E, 3 E, 4 E). Diagnosis :
McDowell (1975). Two monotypic genera, morphologically quite uniform.
Acrochordus. Diagnosis : Hoffstetter & Gayrard
(1965), McDowell (1979). Single genus with three
extant and one extinct species, all very similar. We
do not include any of the poorly known fossil taxa
344
Table 1. Character-by-taxon matrix compiled for this study
(The taxon ‘ varanoids ’ represents the composite rooting taxon used in analyses assuming varanoid affinities of snakes. -l inapplicable. ? l unknown.
Polymorphism and uncertainty are shown as follows : A l 0 and 1 ; a l 0 or 1 ; B l 1 and 2 ; b l 1 or 2 ; C l 0, 1 and 2 ; D l 0 and 2 ; d l 0 or 2 ; E
l 0 and 3 ; F l 2 and 3 ; f l 2 or 3 ; G l 1, 2 and 3 ; H l 1 and 3 ; J l 0, 1 and 3.)
20
30
40
50
60
70
0AA00A0-A0
011001--00
010001-000
0100000000
0A10000000
0010?00000
0AA0000000
0000000000000000000000000100000000100000000000000000000000000A0-
-011A00-30
-000110030
-00000003-00000000-00000003-000-----0
-0000000E0
0030001010
1030001110
00000-1000
00000-0000
00000-0100
00000--000
00000-A000
A000212--1
000021D001
0100000A20
0100000020
0000000000
0A000000A0
0A000000C0
B1-2A0--B1
21-D00--10
00-010--100-000--120-010--100-D00--000-0A0--A-
--001C00B--011000B--00000000
--00000000
--00000000
--00000000
--00000000
1110????00
0b??????00
?b??????00
?b??????00
-1011A0211
-101100211
-2-1?B-211
0110101100
0110201100
0110201000
0110200000
0110B00000
0B00?00000
A2-02B-000
A1A0BC0000
02-0B2-000
1110100000
0110B00000
02-0200000
A2-00B-000
01101A0A00
?100000???
?10???00??
?b00A00a10
?b00?00110
0100000A11210-1110D210-11-00100000110
2100000010
0100A00110
1100000010
1101000002101000010
010110011A
A101100B11
1101100211
1101100110
A101110A10
0101100010
0101110B10
CBAA10A010
?02???????
?02?000b2?
002??0023?
?02??00?1?
-011100110
-011100100
-011100120
0000100010
00210000J0
00211000B0
0021000010
-12110A0-0
01211000?0
002A01A011
0A2001A021
00210000H1
01200000B0
0110001-00
0220001--0
1120001-10
12DAAA1-10
00000???0?
0000????0?
0??00??00?
001000001?
00100000000121--0000121--000010000000010000000
0010000000
00100--000
2110110111
2110110111
001110A011
001110101A
001110101A
00111010A1
0011101011
0010101011
102-1--011
A0211--0A1
0000?00020
10????0?00
10?0????00
10001?1000
0000-12--1
0A00-12--A
0-11-10110
0100112--1
00A0112--1
01101101B1
0110012--1
0210112--1
0110010121
02100100-0
0210A10100
0110110100
0110110100
0110110000
0110100100
0-11010110
0B11A101A0
01?2?0--0?
?1?2????1?
?1021???10
a10210--10
21?A10--12
23?000--02
20?000--12
2101110010
2101110110
2201110021
2201110121
2201110011
2201110111
1H01110010
1A0D110A10
B10B111-00
1101111-10
1111110A00
2111111-A0
2210111-1A
2G1C111-A?
a1011a??00
???000??01
a10101100a
a101101000
2001A20-12001AA0-22001020-11011010-1B01AAB0-B1011110002
1010010002
1111010101
0110010100
01AAAB0-00
01011B0100
0101A10100
010AAB0-01
0101A20-0A
01001B0-01
01A1020-01
01AAA20-0A
Michael S. Y. Lee and John D. Scanlon
Outgroups
Amphisbaenia
Dibamidae
Heloderma
Lanthanotus
Varanus
Mosasauroidea
Varanoid outgroup
Ingroup
Pachyrhachis
Haasiophis
Madtsoiidae
Dinilysia
Leptotyphlopidae
Typhlopidae
Anomalepididae
Anomochilus
Uropeltidae
Cylindrophis
Anilius
Xenopeltis
Loxocemus
Pythoninae
Boinae
Erycinae
Ungaliophiinae
Tropidophiinae
Bolyeriinae
Acrochordus
Colubroidea
10
Snake phylogeny
Outgroups
Amphisbaenia
Dibamidae
Heloderma
Lanthanotus
Varanus
Mosasauroidea
Varanoid outgroup
Ingroup
Pachyrhachis
Haasiophis
Madtsoiidae
Dinilysia
Leptotyphlopidae
Typhlopidae
Anomalepididae
Anomochilus
Uropeltidae
Cylindrophis
Anilius
Xenopeltis
Loxocemus
Pythoninae
Boinae
Erycinae
Ungaliophiinae
Tropidophiinae
Bolyeriinae
Acrochordus
Colubroidea
80
90
100
110
120
130
140
--0A100A00
--01-00000
0000100001
0000000000
0000100000
0000000200
0000A00D00
0000001000
0000?01010
0000000000
00000?0100
0000000000
0000?00000
0000000000
0000-00100-00-0011-1-0-00000
-000-00000
-1-0-0000-0-0-0000A
-A00-00000
000B000000
-00100A010
0000000000
0000000000
0000000000
0000000000
0000000000
0-0--1C000
0-0--00000
0-1--00000
0-0--00000
0-0--00000
0-0--00000
0-0--00000
10100001-0
101000-1-0
1010000000
00100????0
0000000000
0000000000
00A0000000
0-000-A0-0-0-001010
0-00000000-00000000-00000000-000000A0-0000000-
??01-10f??
0001-000??
0?????0???
00?0000???
--0-2-1100
--00201102
--00201102
--?0101d03
--10100201
0010100201
0010100201
0010100211
0011-00211
1111-10211
1111-10211
1111-A0211
0111-00211
A111-00211
0111-00211
1211-1031A
A211?103AA
000???b???
??????10??
??????1??0
?0???010?0
0010001000
1010000000
1010000001
0000101010
A000102010
0000102010
0000101110
0000111112
0000111112
0000111012
000011101B
000011101A
000011101A
0000111110
0000111012
0101212002
010121B00B
?001??????
???1?????1
?011?10100
0010-10000
0000-10100100-1-100100-2-101002010011002010010
1001010011
1001010111
-101010010
-101010010
-A01A1A010
01011111A0
A10111111A
11010110A0
0001A10000
-001A10001
-001010001
0A01AB000A
020?0?00?0
01011?0010
0101?0?010
1101000010
2-120-12-2-120-12-2-120-1020
00020001-0
0001000020
0001000020
0001000020
110B01002A
0201010020
11011-0030
12011-003A
1101A10030
1001A100F0
1101A10020
1202010020
1202010021
1C02A10021
?00??0A0?0
0?0??00??1
1010?10101
00000100?2
---0012-00
---0012-A0
0100012-00
0001112-10
0000111110
1A00111110
0001-0111A
1010001-11
1010001-11
1111000011
1111000011
10AA00C01A
1A11002-1A
1011-02-11
111100B011
111A001-11
11100A2-11
??????????
?0????????
00100A0010
00000????0
20100??1-0
001A0001-0
001000?1-0
10010????1
111100???1
11A1001011
11A1001011
0001000001
2011100001
20A1110001
20A1A10001
20A1100001
20A10????1
2011000001
0001000011
20A1A100A1
20AA0A0001
????0??00?
0????0110?
0?010?0?10
0000010110
1000212010
10002-2010
10002120-0
10?1112010
100211B010
10A1111010
1001111000
1000112100
1001122100
1AA1B21B10
1011B21200
1011BD11A0
10111D11A0
1112121101
1011122B00
1110022101
111CB2BB01
345
346
Table 1 (cont.)
160
170
180
190
200
210
-0001-0010
0-001-0000
00000-0010
00000-0000
00000-0010
00000-0000
00000-00A0
0D1------0
021------0
0000-01000
0000001100
0001-00000
0101002200
0A0A00CC00
0100A0A01A
01-000001A
0000000000
0000001001
0000000000
0000A01000
000000A000
01000AA0001000000000001A0-A?
0000100-00000100-0000100000000AA0000-
1-???0A111
1-???A011A
0200000000
0000000000
1-00000000
0A00?00100
0A00000A00
0-01001010
0-0101??12
0-01100000
0-00100000
0-00200000
00AA001000
000AC0A000
A0A000100B
000000100C
0000000000
0000000000
0000000000
0000000A00
0000000A00
00???00?21
0?????0121
110???0A21
20?1??0?20
2001100101
20A1101001
B001101001
21111?0120
21211001C1
B111100120
2001100110
2001110121
2000010121
2A00020120
2000020120
2000020120
200001012A
2000010121
200001012A
2000010021
20000B012A
?1010??000
??01001000
0101????00
?101??A100
010A012200
0201112010
001-----10
0000001200
0000000200
0100001100
011-00--00
1000111200
1001?1Ab00
AA0A0A0000
1A0A0AAC00
100A00CD0D
100000A002
1001001201
1000001100
100000A-02
1D0A0AC-02
000001?0?1
011000?1?1
?11?01B001
0???22??11
0100000011
0110000010
0110010010
0110101010
010A10C011
0110112011
02-0110011
1201112011
1211102011
121011B111
12A01B1111
12001A1111
---01DB111
121A122111
12111A1111
---0a22111
---12B-A1A
00??000?10
00??000110
00??000110
01??000?11
011---110011--1-10011--1110111-01100111-A1A0A1
011-A0001A
0101100011
0100000010
0101000011
01A1200011
011-20001A
011-D000A1
011-000011
011-000011
011-200010
011-000010
0A1-0A001A
0a??00?0?0
00???10???
01??????00
02???????0
1-0003A1?1
1-000G11A1
1-100F11?1
1-01031???
1-01021111
0B010F1101
0101031001
0001020011
0201030011
0B01030000
0AA10F0000
A2A103AAA0
0C01130000
0A01220000
0A11A30000
0C012F01A0
00012G0000
0001010??0
??????????
1B111A0200
01111100?1
0200011012
0200011012
0200011012
??????????
02A10110AB
0211011001
0211011001
1211111002
1211111002
1211221101
121122AA01
12A12BA1A1
12111B1001
12111BA001
121112000B
12110B0002
A2111B0D?B
100???0100
1110???100
11100010??
11????10??
0011--1010
0011--1012
0011--101F
??????10?0
1010121013
1011--1010
1010111012
1110111013
1110111011
2110111010
2110111010
1A10111010
G110111010
3110111010
3110111013
3110110013
3110111013
Michael S. Y. Lee and John D. Scanlon
Outgroups
Amphisbaenia
Dibamidae
Heloderma
Lanthanotus
Varanus
Mosasauroidea
Varanoid outgroup
Ingroup
Pachyrhachis
Haasiophis
Madtsoiidae
Dinilysia
Leptotyphlopidae
Typhlopidae
Anomalepididae
Anomochilus
Uropeltidae
Cylindrophis
Anilius
Xenopeltis
Loxocemus
Pythoninae
Boinae
Erycinae
Ungaliophiinae
Tropidophiinae
Bolyeriinae
Acrochordus
Colubroidea
150
Snake phylogeny
Outgroups
Amphisbaenia
Dibamidae
Heloderma
Lanthanotus
Varanus
Mosasauroidea
Varanoid outgroup
Ingroup
Pachyrhachis
Haasiophis
Madtsoiidae
Dinilysia
Leptotyphlopidae
Typhlopidae
Anomalepididae
Anomochilus
Uropeltidae
Cylindrophis
Anilius
Xenopeltis
Loxocemus
Pythoninae
Boinae
Erycinae
Ungaliophiinae
Tropidophiinae
Bolyeriinae
Acrochordus
Colubroidea
220
230
240
250
260
0C????????
1A????????
00????????
00????????
00????????
00????????
00-0000000
????????00
?0?????100
??????????
??????????
??????????
??????????
0000000-00
00???0?0??
120??00002
??????????
??????????
????0?????
??????????
00C0000D0C
???0???0??
0??022?-??
???????000
???????001
???????000
???????0??
D0A0CD0000
??30000???
??30000???
000000000000??0000000??00AA0
??????????
0000000000
000
000
010
??0
010
11?
A10
00????????
?0????????
??????????
??????????
1100000100
1200010100
1200001100
1200100000
-220100000
11B0100000
1110100001
-201100000
11011000?0
1101111001
1101111101
1101111101
110111110?
1B0111110?
-201111101
-201111110
-2011A1111
??????????
??????????
??????????
??????????
??00001000
?0000010A1
?000000011
1000010000
1000010000
1000010000
1000011000
1001011100
1001011100
0001011100
0011011100
0011011100
?011001011
?011001D11
1011001200
0111101101
01111A1BAA
??????????
??????????
??????????
??????????
12200002AB
1BA0000AAC
121000010B
122?111212
ABA010A10A
0110100101
0110100201
0000100101
0000100101
0001100100
000110A100
00A110A100
1200111100
1201111100
11A011?B00
1200100000
ABC1101CAC
???????0??
???????0??
???????0??
???????0??
2B10B10010
21CACBA011
21A0B2A010
2221221020
1111C1A020
1211220021
22B1120011
2211010021
2201000021
2201000121
221100A110
2B010001BA
2?01??1110
2?00??A110
B110A1A11A
1100001-20
C2CACCAA2A
??????????
??????????
??????????
??????????
1-F0011001-F0010001-30010001-10111???
1-2012100A010111000220111111
1-10220111
0010220111
0A102B1A11
0A1021AA1C
0A10211012
01102112??
021121121A
0A112B1211
0111000212
1-1A2BACAC
?1?
???
???
???
000
000
000
0??
101
111
11?
011
011
011
111
111
111
111
011
111
A1A
347
348
referred to ‘ Acrochordoidea ’ by either McDowell
(1979, 1987) or Rage (1984, 1987).
Colubroidea (Figs 2 F, 3 F, 4 F). Diagnosis : Underwood (1967 : as ‘ Caenophidia ’), McDowell (1987).
More than 2100 species in nearly 400 genera, very
diverse. For variable characters, the primitive state is
based on the condition found in basal taxa identified
by Kraus & Brown (1998), as well as viperids
(traditionally considered basal colubroids : e.g.
Rage, 1987 ; Cadle, 1988 ; J. B. Slowinksi and R.
Lawson, in prep.). The recent study by Gravlund
(2001) resulted in a very large basal polytomy in
colubroids (if clades supported by only one step are
collapsed), and is thus relatively uninformative with
respect to identification of basal lineages.
During this study, skeletal material representing
almost all ingroup terminal taxa, and diverse
representatives of the outgroups was examined. Only
Anomochilus could not be examined for skeletal
characters (although an alcohol-preserved whole
specimen was consulted), and codings for skeletal
characters for this taxon are based on Cundall &
Rossman (1993). Similarly, only a brief and informal
examination of Haasiophis was possible since the
material was then still to be described, and codings
for that taxon are thus largely based on the published
descriptions and figures in Tchernov et al. (2000),
and on a photograph of the whole specimen
(Wildham, 2000), with reinterpretation of certain
elements as noted below. The material examined is
listed in Appendix 2 ; multiple members of each
diverse terminal taxon, focusing on basal forms, were
examined in an attempt to capture adequately the
primitive state.
For some characters of fossil taxa, codings are
adopted that reflect partial knowledge and thus
maximise the use of available information. For
instance, character 2 concerns the premaxilla ascending process (0, long and contacting frontals ; 1,
intermediate in length and not contacting frontals ;
2, extremely reduced or absent). The premaxilla is
not preserved in madtsoiids or Dinilysia ; however,
the morphology of the preserved skull roof elements
indicates that the premaxilla could not have contacted the frontals. These taxa must therefore have
exhibited either state 1 or 2, and are scored
accordingly.
All phylogenetically informative skeletal characters identified were employed. In addition, phenotypic characters used by previous authors were
evaluated and, if informative for the current taxonomic sample, also included. Particularly valuable
sources for such characters were Haas (1930 a, b,
Michael S. Y. Lee and John D. Scanlon
1931 a, b, 1973), Underwood (1967, 1976), Hoffstetter & Gasc (1969), McDowell (1972, 1975, 1987),
Rage (1984), Rieppel (1977, 1979 b, 1988), Groombridge (1979 a–c), Bellairs & Kamal (1981), Kluge
(1991, 1993 a, b), Cundall et al. (1993), Wallach &
Gu$ nther (1998) and Cundall & Greene (2000).
However, all osteological and external morphological characters were directly checked against
specimens, i.e. the original codings were not accepted
uncritically. The internal soft anatomical characters,
however, were largely based on the published
descriptions and character sets – such traits only
accounted for 40 of the 263 characters employed in
this analysis. For brevity, in the character descriptions, CXX refers to character XX in Cundall et al.
(1993) ; WXX refers to character XX in Wallach &
Gu$ nther (1998) ; and TXX refers to character XX of
Tchernov et al. (2000). Many characters are illustrated, as indicated in the descriptions ; however, it
should be mentioned that for a few characters, the
illustrations appear to show a different character
state to that actually scored, due to either the angle
of view, or overlap of elements.
Many multistate characters used in previous
analyses (notably Cundall et al., 1993) have been
recoded into two or more distinct characters more
adequately to capture potential homologies. For
instance, the unordered multistate character 7 in
Cundall et al. (1993) – interolfactory processes absent (0), present and unfused to subolfactory
processes (1) and present and fused to subolfactory
processes (2) – does not provide evidence to unite
taxa with states 1 and 2, since either state can be
derived parsimoniously from state 0. However, taxa
with states 1 and 2 share a homology, the presence of
the interolfactory process. Recoding this character
into two binaries – process present\absent, process
unfused\fused – will more adequately capture potential homologies. Taxa with the process will be
united by the first recoded character, while variation
in the form of the process is reflected in the second
recoded character (treated as inapplicable in taxa
lacking the process). Further discussion of the
advantages of such coding is given in Wilkinson
(1995).
Many of the characters in this analysis are
multistate. Two sets of analyses were performed,
where such characters were treated as either ordered
(if possible), or unordered. In the ordered analysis,
transformation series are derived that minimise the
amount of evolutionary change between character
states. The extremes in a morphocline were coded as
being derivable from each other only via inter-
Snake phylogeny
mediate stages (i.e. the transformation from 0 to 2
entails two steps, 0- 1 and then 1- 2). Only
characters that formed no clear morphoclines were
treated as unordered. Parsimony-based arguments
in favour of this approach are presented in Wilkinson
(1992) and Slowinski (1993) : because such codings
discriminate against large changes within a character (e.g. those between extremes in a morphocline),
they result in cladograms that entail less overall
evolutionary change than cladograms constructed
by coding all multistate characters as unordered. As
in almost all recent morphological cladistic analyses,
all state changes were weighted equally, i.e. not
scaled to the number of character states.
The following multistate characters were considered to form clear morphoclines, with an extreme
state being primitive, and in the ‘ ordered ’ analysis
were ordered 0–1–2 or where applicable, 0–1–2–3
etc. : 2, 5, 6, 11, 12, 18, 22, 23, 34, 42, 49, 59, 61, 66,
69, 70, 72, 75, 87, 90, 96, 101, 102, 104, 108, 117,
134, 135, 137, 138, 141, 146, 149, 152, 157, 158, 160,
162, 166, 167, 182, 185, 186, 192, 195, 196, 200, 201,
210, 212, 213, 232, 233, 238, 240, 241, 242, 243, 245,
246, 249, 252, 253, 255, 258.
The following multistate characters were considered to form clear morphoclines with an intermediate state being primitive, and in the
‘ ordered ’ analysis were thus ordered 1–0–2 or where
applicable, 1–0–2–3 : 8, 31, 47, 78, 121.
The following multistate characters did not form
clear morphoclines and even in the ‘ ordered ’
analysis had to remain unordered : 28, 29, 33, 45, 51,
52, 54, 60, 80, 85, 94, 109, 120, 136, 165, 175, 198,
206, 228.
In the ‘ unordered analysis ’, all multistate characters were treated as unordered. Ordered and
unordered analyses were performed using both
possible outgroup hypotheses, making a total of four
analyses. These four analyses thus tested the sensitivity of phylogenetic inferences to treatment
(ordering) of multistate characters, and\or outgroup
assumptions.
The full data matrix (Table 1) is available
on TreeBASE (http:\\www.herbaria.harvard.edu\
treebase\) in Nexus format.
V. CHARACTER LIST
(1) Osteology
(a) Skull
1. Premaxilla. 0, anterior surface convex or
straight ; 1, anterior surface concave. Used by Kluge
349
as pr
A
B
as pr
nk
nk
pal pr p
pal pr p
mx
E
D
C
mx
mx
ec
ec
ec pr
pt
ec
pt
ec pr
pt
groove
groove
Fig. 5. Jaw and palatal elements. Right lateral view of
premaxilla of (A) Boa constrictor (Boinae), (B) Liasis
olivaceus (Pythoninae). Dorsal (internal) view of right
palatal elements of (C) Anilius scytale, (D) Loxocemus bicolor,
(E) Candoia aspera (Boinae). Based on (A) BMNH
1964n1243, (B) AR 8422, (C) AMNH 85981, (D) AMNH
19393, (E) AMNH 74992.
(1993 a, character 2) for Pythoninae. While some
uropeltids and colubrids have state 1, basal forms
have state 0. Scolecophidians have a highly modified
and reoriented premaxilla (see character 9) and are
coded as inapplicable.
2. Ascending process of premaxilla. 0, long and
contacting frontals, i.e. extends entire snout-frontal
distance. 1, intermediate in length and not contacting frontals (Fig. 5 A). 2, extremely reduced or
absent (Fig. 5 B). Ordered 0–1–2. We adopt Cundall
et al.’s (1993) proposed homologies for the premaxillary processes, and regard the ascending
process as any median process extending dorsally
and exposed between the horizontal laminae of the
nasals. This is invariably the more dorsal if two
median processes are present, and is the homologue
of the single process in ‘ lizards ’. This character is a
subdivided version of C1, which combined three
distinct premaxillary characters into a single character : shape of ascending process, size of ascending
process, and size of nasal process. Kluge (1991 :
character 7) also employed this character but
considered the ascending process a neomorph of
350
boines : he stated that the process is absent in most
outgroups to boines including all pythons, and does
not mention (Kluge, 1993 a) the small but distinct
ascending process in pythons such as Morelia and
Liasis spp. (Scanlon, 2001 ; Fig. 5 b ; see also figures in
Frazzetta, 1975). Typhlopids and leptotyphlopids
have state 1, but the process appears shorter because
the ventral portion merges with the lateral flanges
above the naris (see character 4). Although the
premaxilla is poorly or not preserved in Haasiophis,
Dinilysia and madtsoiids, the morphology of the skull
roof indicates it did not contact the frontals (i.e. not
state 0), hence these taxa are coded with state 1 or 2.
Also, although there is variability within colubroids,
basal forms have state 1.
3. Ascending process of premaxilla (excluding
lateral flange, see next character). 0, process transversely expanded, partly roofing external nares (Fig.
3 A). 1, process narrow or spine-like, separating but
not roofing external nares (Fig. 2 B–F). This is a
subdivided version of C1 (see character 2). This
character is inapplicable in taxa with an extremely
reduced ascending process.
4. Ascending process of premaxilla. 0, without
lateral flange (Fig. 2 B–F). 1, with lateral flange
forming dorsal margin of external naris (Fig. 3 A).
This character refers to a discrete flange above the
external naris, and is thus distinct from the previous
character, which refers to the dimensions of the main
(vertical) body of the premaxilla. Because the
derived state is recognisable in anomalepidids despite
their reduced ascending process, this character is
considered applicable for all states of character 2.
5. Nasal keel (process) of premaxilla. 0, absent.
1, moderately developed, short flange. 2, well developed, long process (Fig. 5 A, B). Ordered 0–1–2.
Corresponds to C2 and part of C1. This character
refers to a median sagittal flange that extends
posteriorly and is not exposed dorsally, and is the
more ventral if two processes are present. It cannot
be coded in mosasauroids, where the modified
premaxilla and vestigial nasals make it difficult to
assign any of these character states.
6. Palatal (vomerine) process of premaxilla (Fig.
5 A, B). 0, extensive overlapping contact with vomer.
1, non-overlapping, point contact with vomer. 2, not
in contact with vomer. Ordered 0–1–2. C3, adapted
from Underwood (1976, character 38).
7. Premaxilla-vomer contact. 0, flat overlap.
1, well-defined facet. T3, but unlike Tchernov et al.
(2000) we consider this character inapplicable in
taxa lacking a sizeable contact (see previous character).
Michael S. Y. Lee and John D. Scanlon
8. Premaxillary palatal foramina. 0, paired (Fig.
3 B–D). 1, single (Fig. 3 F). 2, multiple (Fig. 3 A).
Ordered 1–0–2. When paired, these are the ventral
openings of the premaxilla channels (e.g. Kluge,
1993 a).
9. Main body of premaxilla. 0, on anterior end of
the snout (Fig. 3 B–F). 1, on ventral surface of snout
(Fig. 3 A).
10. Snout shape. 0, tapering anteriorly in front of
orbits (Fig. 2 B–F). 1, spherical, expanded in front
of orbits (Fig. 2 A).
11. Posterior margin of lateral process of premaxilla, in palatal view. 0, oriented anterolaterally
(Fig. 3 B, E, F). 1, oriented transversely, perpendicular to midline (Fig. 3 A, C, D). 2, oriented
posterolaterally. Ordered 0–1–2. States 0 and 1 were
used by Kluge (1993 a, character 4) for pythonines.
T2 is very similar, but refers to the orientation of the
entire lateral process rather than the posterior
margin alone ; as these traits are highly correlated
T2 is not included here as an additional character.
12. Maxilla-premaxilla contact. 0, close, suture or
strong abutting contact. 1, close but not abutting,
connected by short ligament (Fig. 3 B–E). 2, loose,
widely separated (Fig. 3 A, F). Ordered 0–1–2. As in
C28, after Mahendra (1938). In Pachyrhachis, the
slight separation between the maxilla and premaxilla
and the smoothly rounded (rather than sutural)
anterior end of the maxilla indicate state 1. In
Dinilysia and Wonambi (Madtsoiidae) the rounded,
pitted surface on the anterior tip of the maxilla
indicates attachment to the premaxilla via a ligament
(Scanlon, 1996) and thus either state 1 or 2.
13. Anterior (premaxillary) process of maxilla.
0, well developed, forming ventral margin of external
naris (Fig. 3 B–D). 1, poorly developed or absent,
maxilla excluded from ventral margin of external
naris (Fig. 3 A, F). Reduction of the premaxillary
process, allowing maxillary kinesis to be even more
independent of the snout unit, can be seen as a
continuation of the (ordered) transformation series
in the preceding character.
14. Dorsal (ascending or prefrontal) process of
maxilla. 0, well developed (Fig. 4 A, B). 1, poorly
developed or absent (Fig. 4 C–F). T29.
15. Anteromedial maxillary flange. 0, present,
small horizontal shelf on medial surface of anterior
end of maxilla (Fig. 6 A). 1, absent, anterior end of
maxilla without such shelf (Fig. 6 B). Variation in
this feature in anguimorphs is discussed by Rieppel
(1980, character 7) ; a similar process occurs in some
snakes, figured in seven species of Leptotyphlops by
List (1966, plates 3 and 12), and in a uropeltid
Snake phylogeny
351
A
am fl
B
C
pal pr m
C
D
pal pr m
D
ec pr
Fig. 6. Jaw elements. Ventral view of right maxilla of (A) Cylindrophis ruffus, (B)Trachyboa gularis (Tropidophiinae).
Medial view of left splenial of (C) Loxocemus bicolor, (D) Candoia aspera (Boinae). Based on (A) AMNH 85647,
(B) AMNH 28982, (C) AMNH 44902, (D) AMNH 107142.
(Melanophidium) and Xenopeltis by Rieppel (1977 ;
‘ pma ’ in his Fig. 11). Cannot be scored in taxa with
an anteriorly reduced maxilla.
16. Lateral maxillary foramina. 0, present.
1, absent. This character refers to large discrete
foramina on the middle part of the maxilla, and does
not include any small indistinct foramina at anterior
or posterior tips.
17. Maxilla. 0, alveolar (tooth) row oriented
longitudinally (Fig. 3 B–F). 1, alveolar (tooth) row
oriented transversely (Fig. 3 A).
18. Maxilla-palatine articulation. 0, located anteriorly, at or in front of anterior orbital margin.
1, located beneath anterior half of orbit. 2, located
posteriorly, at same level as centre of orbit or further
posterior. Ordered 0–1–2. Modified from Kluge
(1991, character 22) who suggested that his similar
character (‘ most of the palatine process of the
maxilla … occurs posteriorly, within the orbit ’) may
be considered a synapomorphy of boines, erycines
and pythonines. In Dinilysia most of the palatine
lateral process (and most of the maxilla-palatine
articulation) lies beneath the anterior half of the
orbit (Estes et al., 1970).
19. Palatine process of maxilla. 0, absent, medial
margin of maxilla smooth or with (at most) indistinct
swelling (Fig. 2 A, 3 C). 1, present, medial margin of
maxilla with distinct process (Fig. 3 B, D–F).
20. Palatine process of maxilla. 0, does not
approach pterygoid, palatine broadly enters suborbital fenestra (Fig. 3 B, D–F). 1, contacts pterygoid, excluding palatine from suborbital fenestra.
Inapplicable in taxa lacking a well-developed palatine process. Similar to a character used by Kluge
for boines and erycines (Kluge, 1991, character 50 ;
Kluge, 1993 b, character 41) defined slightly differently in terms of the position of the maxillary process
of the palatine, but which can not always be scored
without disarticulating the palatal bones.
21. Palatine process of maxilla. 0, dorsomedial
surface pierced by a large foramen. 1, not pierced.
T30. Unlike Tchernov et al. (2000), we consider this
character inapplicable in Xenopeltis, which lacks
the palatine process (as do scolecophidians and the
outgroups). Contrary to Tchernov et al. (2000), the
foramen is present in Loxocemus and bolyeriines
(M. S. Y. Lee, personal observations).
22. Ectopterygoid flange of maxilla. 0, maxilla
without distinct posteromedial (ectopterygoid) expansion or flange (Fig. 6 A). 1, maxilla with weak
but distinct posteromedial (ectopterygoid) expansion or flange (Fig. 6 B). 2, maxilla with large
posteromedial (ectopterygoid) expansion or flange.
Ordered 0–1–2. Based on W8. A distinct flange is
present in some pythons (Morelia and Python spp.,
e.g. Scanlon, 2001) but not in basal forms according
to Kluge’s (1993 a) phylogeny. This character cannot
be scored in taxa where the maxilla is greatly
reduced posteriorly (see next character).
23. Posterior extent of maxilla. 0, does not reach
middle of orbit. 1, reaches middle of orbit, or slightly
further. 2, extends past posterior margin of orbit.
Ordered 0–1–2. Discussed by McDowell and Bogert
(1954). Most mosasaurs have state 0, but aigialosaurs
appear to have state 1. The posterior end of the
maxilla is hidden by the jugal in all known
aigialosaur skulls, but the dentary tooth row in both
Aigialosaurus and Opetiosaurus extends to near the
middle of the orbit (Gorjanovic-Kramberger, 1892 ;
Kornhuber, 1901). In lizards generally, the maxil-
352
lary tooth row is rarely much shorter than that of the
dentary, and there seems no reason to suppose it was
in aigialosaurs ; we therefore adopt the reconstruction of Russell (1967, Fig. 98 b) rather than Carroll
& deBraga (1992, their Fig. 7 b).
24. Nasal. 0, does not closely approach lateral
process of premaxilla. 1, extends anteriorly to almost
reach lateral process of premaxilla. Variation within
pythonines and colubroids suggests that state 1 is
correlated with reinforcement of the snout in
burrowing forms.
25. Horizontal lamina of nasal. 0, narrow anteriorly, tapering to a point beside premaxilla (Fig.
2 D). 1, horizontal lamina of nasal wide anteriorly, at
most tapering only slightly to a blunt anterior end
(Fig. 2 A–C, E). Not applicable in taxa where the
nasals are vestigial or absent. The distribution of this
character in varanoids and primitive snakes was
discussed by McDowell & Bogert (1954) and Rieppel
(1983) and in pythonines by Kluge (1993 a, character 8). This character is equivalent to T4 (nasal
broad anteriorly and approaching septomaxilla, or
narrow anteriorly and not approaching septomaxilla). However, because the posterior limit of the
septomaxilla appears to be independently variable,
we do not involve this structure in our character
definition.
26. Horizontal lamina of nasal. 0, posterior margin
wide (Fig. 2 A–C, E). 1, posterior margin narrow,
tapering to a posteromedial point (Fig. 2 D). This
character does not include the median vertical
flange, which can extend past the wide posterior
margin of the horizontal lamina. Modified from
character 11 of Kluge (1993 a) for Pythoninae, after
Frazzetta (1959). Not applicable in taxa where the
nasals are vestigial or absent.
27. Nasal-frontal contact. 0, horizontal laminae
of nasals and frontals in contact (Fig. 2 A–D).
1, horizontal laminae of nasals and frontals not in
contact (Fig. 2 F). This and the following character
represent an alternative analysis of the variation
described by C5, modified from character 14 of
Kluge (1991).
28. Nasal-frontal contact. 0, nasals overlap frontals
dorsally. 1, frontals overlap nasals dorsally. 2, clasping junction, nasal fits into anterior groove in
frontal and is thus overlapped and underlapped by
frontal. Unordered ; inapplicable if the horizontal
laminae of nasals and frontals are not in contact (see
preceding character).
29. Nasal-frontal boundary : 0, concave posteriorly
in dorsal view. 1, approximately straight and
transverse. 2, convex posteriorly. 3, W-shaped, nasals
Michael S. Y. Lee and John D. Scanlon
project posteriorly into embayments in frontals.
Unordered. Partly equivalent to character 12 of
Kluge (1993 b) for erycines. ‘ Boundary ’ rather than
‘ contact ’ is used to make this character applicable in
taxa where the nasals and frontals are separated by
a small gap (see above). This character cannot be
scored in mosasauroids, where the nasals are vestigial
or absent and furthermore do not meet one another
to form a continuous nasofrontal boundary.
30. Descending laminae of nasals. 0, not enlarged
(shallow) anteriorly (Fig. 7 A). 1, distinctly enlarged
(very deep) anteriorly (Fig. 7 B).
31. Anterior process of prefrontal. 0, moderately developed, triangular flange (Fig. 4 B, D–F).
1, greatly reduced. 2, greatly elongated process (Fig.
4 C). Ordered 1–0–2. Loxocemus and Xenopeltis possess
two anterior processes of the prefrontal, the longer
(anteroventral) process more closely resembles the
single process of other snakes topologically and
structurally, and homology has accordingly been
assumed. The shorter dorsal lappet has thus been
interpreted as a neomorph (see next character).
32. Anterodorsal lappet of prefrontal, extending
along lateral margin of nasal. 0, absent or indistinct
(Fig. 2 A, B, D–F). 1, distinct process (Fig. 2 c). See
previous character for discussion.
33. Prefrontal-nasal contact. 0, prefrontal separated from nasal by external naris. 1, prefrontal contacts nasal (Fig. 2 A–E). 2, prefrontal separated from
nasal by ragged fissure that is not part of external
naris (Fig. 2 F). 3, prefrontal separated from nasal by
frontal-maxillary contact. Unordered. Similar to
T23, which combined the very different states 0, 2
and 3 into a single character state (prefrontal
separated from nasal) ; cf. discussion of ‘ retraction of
the external nares ’ in McDowell & Bogert (1954),
Pregill, Gauthier & Greene (1986, character 2) and
Estes et al. (1988, characters 2 and 4).
34. Prefrontal-maxilla contact. 0, anterior process
and ventrolateral margin of prefrontal contact
maxilla (Fig. 4 B, C, E). 1, anterior process of
prefrontal does not contact maxilla and projects
freely, only ventrolateral margin of prefrontal contacting maxilla (Fig. 4 D, F). 2, anterior process and
ventrolateral margin of prefrontal do not contact
maxilla (Fig. 4 A). Ordered 0–1–2. T22, with an
extra state (2) added.
35. Prefrontal-maxilla contact on facial region.
0, tight or interdigitating, relatively rigid. 1, flat or
slightly convex surfaces, allowing rocking or sliding
motion. In Wonambi (Madtsoiidae) the prefrontal is
unknown but the maxilla has a deeply corrugated
surface indicating an interdigitating suture (Scanlon,
Snake phylogeny
353
D
A
na
fr
sm
pbs
sm
vo
pm
B
E
na
fr
sm
sm
vo
pbs
pm
C
F
na
fr
pbs
pm
sm
sm
vo
G
I
fr
pa
prf
fd
eo
eo
bo
pbs
of
J
H
eo
fr
pa
eo
bo
prf
pbs
of
Fig. 7. Orbit, snout and occiput. Right lateral view of snout region, with maxilla and prefrontal omitted, of
(A) Cylindrophis ruffus, (B) Corallus caninus (Boinae), (C) Macropisthodon rudis (Colubroidea). Right lateral view of dorsal
process of septomaxilla of (D) Typhlops punctatus (Typhlopidae), (E) Casarea dussumieri (Bolyeriinae), (F) Cerastes cerastes
(Colubroidea). Right lateral view of orbital region of (G) Anilius scytale, (H) Trachyboa gularis (Tropidophiinae). Posterior view of occipital condyle of (I) Xenopeltis unicolor, ( J) Cylindrophis ruffus. Based on (A) AMNH 12872, 85647,
(B) AMNH 57816, (C) AMNH 34513, (D) BMNH 1975n567, (E) BMNH 1992n995, (F) BMNH 2n3n1a, (G) AMNH
85981, (H) AMNH 28982, (I) BMNH 1947n1n1n12, ( J) BMNH 1930n5n8n47.
Michael S. Y. Lee and John D. Scanlon
354
C
B
A
prf
fr
fr
fr
prf
prf
la for
la for
pal
mx
pal
mx
pt
med
ft
lat ft
mx
pal
pt
la duct
E
D
la duct
la for
la for
medial
medial
anterior
anterior
Fig. 8. Orbital elements. Anterior region of right orbit viewed from posterolateral aspect in (A) Anilius scytale,
(B) Loxocemus bicolor, (C) Candoia aspera (Boinae). Ventral view of right prefrontal of (D) Corallus enydris (Boinae),
(E) Bolyeria multocarinata (Bolyeriinae). Based on (A) AMNH 85980, (B) AMNH 44902, (C) AMNH 107142,
(D) BMNH 1904n1228, (E) BMNH 70n11n30n4A.
1996). Degree of interdigitation and mobility in
Dinilysia and anilioids has been poorly documented
but is mentioned by Estes et al. (1970), Frazzetta
(1970), Cundall & Rossman (1993), and Cundall
(1995).
36. Antorbital buttress of prefrontal. 0, lateral foot
process does not contact palatine (Fig. 8 A, C).
1, lateral foot process contacts palatine (Fig. 8 B).
T16, modified. The lateral foot process is the ventral
process that extends lateral to the lacrimal foramen
(e.g. Frazzetta, 1966, Cundall & Rossman, 1993).
Tchernov et al. (2000) subdivided state 1 into two
states, lateral foot process contacts palatine and
maxilla, or palatine only. However, the latter state
was attributed to Xenopeltis only, and the character is
thus cladistically uninformative in the context of the
current taxon set. When the lateral and medial
processes are fused ventral to the lacrimal foramen
(Fig. 8 A, B), their boundaries can be difficult to
determine. This character is also inapplicable in
varanoids, where the area of the lateral foot process
is occupied by the lacrimal (a condition that is not
comparable to any condition in the ingroup). All
scolecophidians were coded as unknown by Tcher-
nov et al. (2000). The condition in typhlopids and
anomalepidids is indeed difficult to score because
this region is highly modified – however, leptotyphlopids are generalized enough to score and exhibit
state 0.
37. Antorbital buttress of prefrontal. 0, medial
foot process does not contact maxilla. 1, medial foot
process contacts maxilla. Not applicable in anomalepidids and typhlopids, which have a highly modified
maxilla.
38. Lateral process of palatine. 0, does not reach
lateral edge of maxilla (Fig. 8 A, C). 1, reaches
lateral edge of maxilla (Fig. 8 B).
39. Outer orbital (lateral) margin of prefrontal, in
lateral view. 0, slants anteroventrally (Fig. 4 B).
1, vertical (Fig. 4 C–F). T14. The prefrontal is not yet
known in madtsoiids ; however, because the ventral
(maxillary) facet for the prefrontal is well anterior to
the dorsal (frontal) facet, state 0 can be provisionally
inferred (Scanlon & Lee, 2000).
40. Prefrontal lacrimal duct roof. 0, absent
(Fig. 8 D). 1, present, a horizontal flange extending
anteriorly from lacrimal foramen (Fig. 8 E). Kluge
(1993 a : character 20) considered this trait within
Snake phylogeny
pythonines, and the structures involved are well
illustrated by Frazzetta (1966, fig. 18). This area is
rarely exposed in articulated skulls, but may sometimes be visible in anterior view through the external
naris.
41. Prefrontal-frontal contact. 0, approximately
straight. 1, curved, prefrontal fitting into deep
embayment in frontal. This character relates to the
boundaries of the elements as exposed in dorsal view ;
in most if not all cases, state 1 results from a process
of the prefrontal (dorsal lappet) overlapping the
frontal.
42. Prefrontal-frontal contact. 0, oriented approximately parasagittally, prefrontals contact only lateral margins of frontals and are widely separated
(Fig. 2 A). 1, oriented anteromedially, prefrontals
contact anterolateral margins of frontals and are
moderately separated (Fig. 2 B, E, F). 2, oriented
anteromedially or transversely, prefrontals closely
approaching or contacting one another (Fig. 2 C, D).
Ordered 0–1–2. A more finely subdivided version of
C13 (frontal-prefrontal sutures widely separated\not
widely separated medially). The divisions between
character states are often difficult as the relevant
suture is often curved : state 1 includes straight
oblique contacts as well as those where the prefrontal
contact curves anteromedially around the anterolateral corner of the frontal. This character cannot
be coded in anomalepidids or Acrochordus, where a
lateral process of the frontal makes a point or very
narrow contact with the prefrontal.
43. Prefrontal-frontal contact. 0, prefrontal
sutured to or tightly buttressed against frontal.
1, prefrontal moveably articulated to frontal. This
represents a modification of C8 (prefrontal extends
between frontal and maxilla\prefrontal sutured between frontal, nasal, and maxilla\prefrontal articulated between frontal and maxilla). The first two
states of Cundall et al.’s (1993) character are difficult
to interpret and are here combined into a single
character state (0). Coded as ‘ ? ’ in Madtsoiidae
since the prefrontal is unknown, although a mobile
joint could be inferred in Wonambi because the
maxilla-prefrontal suture (see character 35) has
strong interdigitation and seems likely to have been
rigid, while the dorsal and lateral facets on the
anterior frontal are smooth.
44. Antorbital (vertical) buttress of prefrontal.
0, broad, extends medially underneath lateral descending flanges of frontal. 1, narrow, does not
extend medially to reach lateral descending flanges
of frontal. Character 17 of Kluge (1993 b) for Erycinae, slightly reworded.
355
45. Lacrimal foramen. 0, completely enclosed by
prefrontal (Fig. 8 A, B). 1, between prefrontal and
palatine. 2, between prefrontal and maxilla (Fig.
8 C). Unordered. A more finely subdivided version of
character 19 of Kluge (1993 a : lacrimal foramen
incompletely\completely enclosed by bone) and T15
(which considered only states 0 and 2). Not
applicable in taxa lacking the foramen, i.e. scolecophidians.
46. Jugal. 0, present. 1, absent. Character 32 of
Estes et al. (1988) is partly equivalent. In order to
avoid using phylogenetic hypotheses that are currently under test to make a priori assumptions about
homology and non-homology (see character 47), the
suborbital ossification of bolyeriines (Cundall &
Irish, 1989) is here provisionally coded as a jugal
because it is similar in position and morphology to
the jugals in other taxa ; however, incongruence with
other characters indicates that it is probably a
neomorph. The curved rod-like element below and
behind the eye in anomalepidids has been considered
either a postorbital (Haas, 1964) or a jugal (List,
1966). It is embedded in the postorbital ligament
without contacting the maxilla ; Haas (1964) points
out that the jugal, when reduced in lizards, remains
in contact with the maxilla (while the postorbital
does not), so the bone is here considered a postorbital. A very similar reduced postorbital (rod-like,
with only a loose association with the parietal)
occurs in some fossorial colubroids (e.g. some
Simoselaps ; Scanlon & Shine, 1988). The putative
jugal in Pachyrhachis was reinterpreted by Tchernov
et al. (2000) as the ‘ ectopterygoid ’, because it
‘ broadly overlaps ’ the ‘ posterior end of the maxilla
… as in macrostomatan snakes ’. This interpretation
is questionable because the element concerned lies
entirely in front of the postorbitofrontal on both
sides, and almost entirely in front of the pterygoids,
whereas the ectopterygoid in snakes lies mainly or
entirely behind the postorbitofrontal (when present),
and contacts the pterygoid usually well behind the
palatine-pterygoid contact. By contrast, the element
is in the appropriate position to be a jugal : in front
of the postorbitofrontal and on the ventral margin of
the orbit (broadly overlapping the suborbital part of
the maxilla, as the jugal does in most squamates).
Although Tchernov et al. (2000) did not identify
jugals in Haasiophis, a reinterpretion of the circumorbital elements suggests they might be present. The
elements interpreted as ‘ postorbitals ’ by Tchernov et
al. (2000) are the correct shape (sliver-shaped) and
in the correct position (anteroventral region of orbit,
nearly reaching the ventral ‘ foot ’ of the prefrontals)
356
to be the jugals. Furthermore, there are other
preserved elements (not identified by Tchernov et al.,
2000) which might correspond to the real posterior
orbital elements, i.e. the postorbital and\or postfrontal (see postfrontal ; next character). Until the
identity of these elements is convincingly demonstrated, the condition of the jugal in Haasiophis
should be considered unknown.
47. Postorbitofrontal ossification(s) in adults.
0, one discrete ossification, conventionally termed
the postorbitofrontal (Fig. 4 B, D–F). 1, two discrete
ossifications, conventionally termed the postfrontal
and postorbital. 2, no discrete ossifications (Fig.
4 A, C). Overlaps with C9 and T26 (postorbital
present\absent). Ordered 1–0–2 (two, one, no
ossifications). The ‘ supraorbital ’ (see Section VI) is
here treated as a different element to the postorbitalpostfrontal complex, and thus not considered in
scoring this character. We regard as a ‘ postfrontal ’
any circumorbital ossification that straddles the
frontoparietal suture and lies dorsal to the postorbital
(largely or entirely separating it from the frontoparietal unit), and as a ‘ supraorbital ’ any circumorbital ossification that lies anterior to the frontoparietal suture and the postorbital (allowing the
latter robustly to contact the frontoparietal unit).
Thus, the posterodorsal orbital element in Dinilysia is
considered to be a postfrontal (Estes et al., 1970 ;
Rieppel, 1977), and this taxon has two postorbitofrontal ossifications. By contrast, the dorsal orbital
elements in pythonines and the erycine Calabaria
are supraorbitals (Rieppel, 1977 ; Kluge, 1993 b ;
Cundall et al., 1993), and these taxa have a single
postorbitofrontal ossification. In contrast, Tchernov
et al. (2000 ; contra Rieppel, 1977) interpreted the
element in Dinilysia as the supraorbital because they
considered this taxon an alethinophidian and ‘ believe the postfrontal to be absent in alethinophidians ’. This interpretation ignores the similarity
of the element in Dinilysia to the postfrontal of lizards
and dissimilarity to the supraorbital of pythonines
and Calabaria. It is circular reasoning to interpret
homologies based on a priori phylogenetic assumptions that one is attempting to test : it is more
objective to hypothesise homologies based on topological and structural comparisons and test them by
character congruence. We interpret the posterior
orbital element in Dinilysia as a putative postfrontal,
based on topological (at frontoparietal suture) and
structural (forked medially) similarities to the postfrontal of lizards. Although postorbitals were identified in Haasiophis (Tchernov et al., 2000), these
appear to be jugals (see previous character). How-
Michael S. Y. Lee and John D. Scanlon
ever, real postorbitals might be present, and this
taxon is thus coded as unknown. On the left side of
Haasiophis, there is a rectangular element positioned
directly over the maxilla-ectopterygoid junction,
and extending anteromedially to underlap the
frontal-parietal contact. This is shown in both the
photograph and interpretive drawing, but not
identified. On the right side, a similar distinct
element is visible on the photograph, but interpreted
in the drawing as the anterior part of an implausibly
large right ectopterygoid (implying the right
element extends much further anteriorly, and is
twice as large, as the well-preserved left element).
The elements are in the correct position, and are
approximately the right shape, to be the true
postorbitofrontals.
48. Postorbitofrontal ossification(s). 0, forked
medial margin, anterior and posterior rami tightly
clasping frontoparietal suture (Fig. 2 D). 1, not
forked medially, without distinct anterior and
posterior rami, abutting skull roof (Fig. 2 B, E, F).
T28, with polarity reversed. Not applicable in taxa
lacking postorbitofrontal ossification(s) ; see previous
character. From Underwood (1976), Estes et al.
(1988) and Kluge (1993 a). Tchernov et al. (2000)
assigned the unforked state to varanids, which have
the forked state. Dibamids were coded as unknown,
but some retain a postfrontal, which is forked.
Tchernov et al. (2000) also assign the forked state to
boines and some erycines, but these have the
unforked state. Pachyrhachis was also reconstructed
without a forked postorbitofrontal, contrary to the
original description (Haas, 1979 ; confirmed by Lee
& Caldwell, 1998). The putative anterior ramus
identified by Haas was instead identified by Tchernov et al. (2000) as part of the descending flange of
the parietal. However, only the identification of the
extreme tip of the anterior ramus is questionable. As
stated and illustrated in previous descriptions, the
more proximal part of the anterior ramus and the
entire posterior ramus are continuous with the rest of
the postorbitofrontal, and these two rami are clearly
exposed on both sides (Haas, 1979, Figs 4 and 6 ; Lee
& Caldwell, 1998, Fig. 3). The dorsal surface of the
skull of Pachyrhachis has been embedded in resin and
not further prepared since the earlier descriptions
(Haas, 1979 ; Lee & Caldwell, 1998) ; thus, Tchernov
et al.’s (2000) reinterpretation presumably cannot
be based on additional preparation. Conceivably,
X-rays might have revealed that the long anterior
and\or posterior rami of the postorbitals consisted
of resistant matrix rather than bone, and that
the identification of Haas (1979) was incorrect.
Snake phylogeny
However, no evidence supporting this reinterpretation was presented.
49. Prefrontal and postorbitofrontal ossification(s). 0, widely separated, frontal broadly enters
orbit (Fig. 2 E, F). 1, narrowly separated, frontal
narrowly enters orbit. 2, in contact, frontal excluded
from orbit (Fig. 2 B). This character cannot be coded
in taxa lacking posterior orbital ossifications, or in
taxa where the prefrontal and postorbitofrontal are
separated by a supraorbital (i.e. pythonines and
some erycines). Ordered 0–1–2.
50. Lateral process of parietal. 0, lateral process
distinct (Fig. 2 A, D). 1, lateral process absent (Fig.
2 B, C). The lateral process of the parietal is situated
immediately behind the posterior orbital ossification
(s), when these elements are present. Equivalent to
T35 (parietal ‘ distinctly expanded ’, and ‘ not expanded … laterally at anterior end ’), with polarity
reversed. Tchernov et al. (2000) incorrectly code
varanoids, Haasiophis, Dinilysia, typhlopids and
anomalepidids as lacking the expansion.
51. Posterior orbital margin. 0, complete, closed
by postorbital contacting jugal. 1, complete, closed
by postorbital contacting ectopterygoid-maxilla unit
(Fig. 4 F). 2, incomplete (Fig. 4 A–D). Unordered.
This character is very similar to T27 (postorbital
separated\closely approaches or contacts ectopterygoid to form posterior orbital margin), but has been
rephrased slightly because in some snakes (e.g. some
pythons : Frazzetta, 1966 ; Kluge, 1993 a) and some
amphisbaenians (Berman, 1973) the postorbital
contacts the maxilla directly, anterior to the ectopterygoid, to complete the orbital margin. The
illustrated pythonine is atypical ; basal pythons have
a closed orbital margin (Kluge, 1993 a).
52. Frontal shape. 0, frontals gradually tapering
anteriorly. 1, frontals rectangular, at most slightly
constricted in middle (Fig. 2 D–F). 2, Frontals
gradually tapering posteriorly (Fig. 2 B, C).
3, Frontals greatly constricted in middle (Fig. 2 A).
Unordered.
53. Anterior tab of frontal. 0, distinct and welldefined. 1, poorly defined or absent. T20, after
Frazzetta (1966). This character refers to a transverse ridge on the anterior margin of the frontal
which underlaps the prefrontal and nasal (Frazzetta,
1966, Figs 11 and 12) ; the term we use reflects
analogous similarity to the frontal and parietal
‘ tabs ’ in other squamates (Estes et al., 1988).
Tchernov et al. (2000), referring to it as a ‘ preorbital
ridge ’, coded Dinilysia as unknown and bolyeriines
with state 0 ; we find they have states 0 and 1,
respectively.
357
54. Frontal-parietal contact (dorsal aspect).
0, mostly straight and transverse, slight median
notch in frontals at most (Fig. 2 A, F). 1, anteriorly
concave, i.e. frontals extending posteriorly into
broad median embayment in parietals (Fig.
2 B, C, E). 2, complex W or M shape (Fig. 2 D).
Unordered. Underwood (1976, character 47) used
the ratio of depth to width of the embayment in the
anterior margin of the parietal (state 1 here) as a
quantitative character in Booidea. The suture in
Haasiophis is W-shaped, with a small median anterior
projection of the parietal into the frontal : this is
shown in the specimen photographs but not the
drawing in Tchernov et al. (2000), which depicts a
smooth U-shaped suture. Pachyrhachis also has a
complex suture (contra Scanlon & Lee, 2000).
55. Subolfactory (lateral descending) processes of
frontal. 0, not contacting one another ventromedially. 1, meeting ventromedially, below medial
descending processes of frontal if present. Variation
in this feature has often been discussed in anguimorphs, e.g. McDowell & Bogert (1954).
56. Medial descending processes of frontal.
0, absent. 1, present. This and the following character
are based on C7 and T24 (interolfactory processes
absent\present but unfused to subolfactory processes\present and fused to subolfactory processes)
split into two binaries. The ‘ interolfactory pillar ’
of the frontal (Cundall et al., 1993) is formed by
the medial descending process and the portion of
the subolfactory (lateral) process which encircles the
olfactory peduncle and extends dorsomedially to
meet the medial descending process. This character
was initially identified by Underwood (1967) and
McDowell (1967, 1975).
57. Medial descending processes of frontal.
0, meeting with subolfactory (lateral descending)
frontal processes at the mesial frontal suture.
1, fused to subolfactory frontal processes, mesial
frontal suture obliterated.
58. Mesial frontal suture. 0, on ventral portion of
interolfactory pillar. 1, on middle of interolfactory
pillar. Kluge (1991) commented on further variation
in position of the suture in relation to his character
15 (l character 40 of Underwood, 1976 : nasalfrontal contact dorsal\dorsal & ventral\ventral to
suture), which is not used here since the ‘ dorsal ’ and
‘ ventral ’ states only occur in some booid lineages
and are uninformative in the context of the current
taxon set.
59. Length of main body of parietal (i.e. excluding
supraorbital or posterior processes). 0, short, at
most 40 % skull (snout-occiput) length (Fig. 2 A, F).
Michael S. Y. Lee and John D. Scanlon
358
A
C
B
eo
st
so
pa
pa
pa
pro
jr
eo
op
pro
V2
lat
pro
V3
st
st
bo
cc
vcp pbs
vca
qa
eo so
eo so
D
E
ba pr
keel
F
ba pr
pbs
pbs
vcp
vcp
bo
keel
vcp
pbs
bo
bo
Fig. 9. Braincase elements. (A) Anilius scytale, right lateral view of braincase. Dorsal view of braincase of (B) Anilius
scytale, (C) Exiliboa placata (Ungaliophiinae). Ventral view of parabasisphenoid of (D) Anilius scytale, (E) Loxocemus
bicolor, (F) Acrantophis dumerili (Boinae). Based on (A) AMNH 85980, (B) AMNH 85982, (C) AMNH 102892,
(D) AMNH 85981, 85982, (E) AMNH 44902, (F) BMNH 92n2n29n20.
1, intermediate, between 40 and 55 % of skull length
(Fig. 2 C–E). 2, long, at least 55 % of skull length
(Fig. 2 B). Ordered 0–1–2.
60. Suture between frontal and parietal descending flanges. 0, in lateral view, suture between frontal
and parietal extends approximately vertically, or
slightly anterodorsally (Fig. 4 F). 1, suture greatly
inclined anterodorsally, i.e. closer to the horizontal
than the vertical (Fig. 4 C). 2, suture curved,
extending vertically in its ventral portion and
becoming horizontal more dorsally (Fig. 4 A).
Unordered. Modified from character 28 of Kluge
(1993 a) for Pythoninae. Not applicable in lizards
which lack the suture, thus polarity is indeterminate.
61. Optic foramen. 0, posteriorly located, posterior
border forming a deep notch in parietal (Fig. 4 D).
1, intermediate position, posterior border formed by
straight margin of parietal (Fig. 4 C). 2, anteriorly
located, posterior border within frontal (Fig. 4 A).
Ordered 0–1–2. This character includes T33 (parietal not notched\notched for optic foramen), in
turn based on Underwood (1967) ; Kluge recognizes
further distinctions within state 0 in boines (Kluge,
1991, characters 30-31) and erycines (Kluge, 1993 b,
character 23). Tchernov et al. (2000) coded Dinilysia
with an unnotched parietal bordering the foramen ;
however, the reconstructions in Estes et al. (1970)
showed a notched parietal. Dinilysia has therefore
been coded with both possible states (0 or 1).
Similarly, in Pachyrhachis and the madtsoiid Wonambi,
the morphology of the frontals indicates that the
foramen was partly bordered by the parietals, and
they must have therefore exhibited either state 0 or
1. Haas (1930 a, 1964, 1968) described the condition
in scolecophidians ; McDowell (1967) contrasted this
with the alethinophidian condition, and concluded
that the ventral completion of the enclosure of the
braincase probably took place independently in the
two groups.
62. Optic foramen. 0, opening faces anterolaterally (Fig. 4 C). 1, opening faces laterally (Fig.
4 D–F).
63. Anterior (supraorbital) process of parietal.
0, absent or poorly developed (Fig. 2 A, D–F).
1, enlarged, extending along at least 50 % of lateral
margin of frontal (Fig. 2 B, C). C11 (see also their
comments on C10), T32. The condition in Haasiophis
is uncertain, as the shape of the frontoparietal suture
Snake phylogeny
depicted in the photographs appears be different
from that reconstructed by Tchernov et al. (2000)
(see character 54 above). Furthermore, even if this
reconstruction is correct, the supraorbital process
is short on the right (state 0), but long on the left
(state 1).
64. Posterior border of parietal. 0, with distinct
median notch (Fig. 2 E). 1, without distinct median
notch (Fig. 2 A–D, F). This and the following
character are very similar to T37 (posterior parietal
margin embayed posteriorly\straight\pointed posteriorly), which is accordingly not employed here.
There might be a case for treating T37 as a distinct
character if Tchernov et al.’s (2000) codings only
considered the lateral margins of the posterior
border, ignoring any median structures, but this is
not clear from the character descriptions.
65. Posterior border of parietal. 0, without median
projection over supraoccipital (Fig. 2 A, C, E, F).
1, with median projection over supraoccipital (Fig.
2 B, D).
66. Posterolateral (supratemporal or suspensorial)
process of parietal. 0, well developed, posterolateral
margin of parietal with a distinct flange. 1, reduced,
posterolateral margin of parietal with a triangular
corner (Fig. 9 B). 2, absent, posterolateral margin of
parietal rounded (Figs 2 F, 9 C). Ordered 0–1–2.
The condition found in Dinilysia and anilioids is
discussed in detail by Estes et al. (1970). This is a
more finely subdivided version of T34 (supratemporal process of parietal distinct\not distinct).
67. Descending flange of parietal. 0, without
horizontal crest. 1, with very large horizontal crest,
extending from orbital region towards prootic.
Barrie (1990) noted the similarity of Wonambi
(Madtsoiidae) to Dinilysia in this feature.
68. Descending flange of parietal. 0, does not
contact anterior margin of base of basipterygoid
process. 1, broadly contacts anterior margin of base
of basipterygoid process. T36. This character is
inapplicable in taxa lacking basipterygoid processes,
which is presumably why scolecophidians, Anomochilus and some advanced snakes were coded as
unknown by Tchernov et al. (2000). However, many
outgroup and anilioid taxa retain basipterygoid
processes yet were also coded as unknown ; they all
exhibit state 0. Pythons cannot be coded objectively
because the region is highly modified due to the
presence of the basisphenoid fenestra (see Kluge,
1993 a).
69. Supratemporal. 0, large, quadrate contacts
supratemporal but not otic capsule (Fig. 2 B–F).
1, vestigial, quadrate contacts supratemporal and
359
otic capsule. 2, absent, quadrate contacts otic capsule
only (Fig. 2 A). Ordered 0–1–2. C16, also related to
C2 (quadrate associated with supratemporal or otic
capsule) and T38 (supratemporal present\absent),
which have therefore not been employed here. Some
amphisbaenians retain a vestigial temporal element
(Gans, 1978), here considered the supratemporal
rather than the squamosal based on its dorsomedial
position.
70. Supratemporal. 0, posterior end projecting
greatly beyond otic capsule (Fig. 2 D). 1, posterior
end projecting slightly beyond otic capsule (Fig.
2 C, E, F). 2, posterior end not projecting beyond otic
capsule (Fig. 2 B). Ordered 0–1–2. Similar to T39
(supratemporal without free posterior process\with
simple free process\with dilated free process\with
hooked free process). However, the last two states of
T39 are not uniformly present in more than one
terminal taxon and are cladistically uninformative here, and thus not employed. Similar transformation series have been recognized by Estes et al.
(1970), McDowell (1975), Underwood (1967, 1976),
Groombridge (1979 a) and Kluge (1991). Contrary
to Tchernov et al. (2000), we treat this character as
inapplicable in taxa where the supratemporals are
vestigial or absent.
71. Supratemporal. 0, does not substantially cover
dorsolateral surface of prootic (Fig. 2 B, C). 1, covers
almost the entire dorsolateral surface of prootic (Fig.
2 D–F). Not applicable in taxa with vestigial (or
absent) supratemporals ; see also discussion under
character 136.
72. Supratemporal. 0, anterior tip well behind
anterior margin of prootic (Fig. 2 B, C). 1, anterior
tip slightly behind anterior margin of prootic (Fig.
2 D, E). 2, anterior tip in line with or in front of
anterior margin of prootic (Fig. 2 F). Ordered 0–1–2.
Not applicable in taxa with vestigial (or absent)
supratemporals. Tchernov et al. (2000) state in their
character list that Pachyrhachis is coded as unknown
because ‘ the supratemporals are displaced anteriorly ’, but in the data matrix it is coded with ‘ 1 ’ ;
we favour the former assessment. Loxocemus is coded
with state 1 by Tchernov et al. (2000) but has state
0 (e.g. Underwood, 1976 ; Rieppel, 1977).
73. Quadrate. 0, without small ossification (‘ stylohyal ’) on medial surface, contacting stapes. 1, with
such ossification. This character is related to, but
distinct from, character 146 (l T59).
74. Dorsoposterior (l suprastapedial) process of
quadrate. 0, distinct, large (Fig. 4 B, C). 1, indistinct,
small or absent (Fig. 4 D–F). T56. In leptotyphlopids, the stapedial shaft appears to have become
360
parallel with, and thus indistinct from, the suprastapedial process ; this potential ‘ loss ’ of the suprastapedial process fails the similarity test for homology
with the true losses in other taxa, and leptotyphlopids have been scored as inapplicable.
75. Dorsoposterior (suprastapedial) process of
quadrate. 0, projects posteroventrally, forming acute
angle with quadrate shaft. 1, projects posteriorly,
forming approximately a right or slightly obtuse
angle with quadrate shaft (Fig. 4 B, C). 2, projects
posterodorsally, forms very obtuse angle with quadrate shaft in lateral view (Fig. 4 A). Cannot be scored
in taxa lacking a prominent process. Ordered 0–1–2.
The angle measured is that formed in lateral view by
posterior margin of the quadrate shaft, not by the
external crests (if present).
76. Length of quadrate shaft (i.e. excluding
suprastapedial process). 0, short, maximum length
along shaft no more than 25 % of snout-occiput
length. 1, long, more than 25 % of snout-occiput
length. T57 is a similar character (quadrate shaft
longer\slightly shorter\much shorter than suprastapedial process) ; however, this definition is problematic because the length of the shaft is compared
to the suprastapedial process, which is proportionately much more variable and often totally absent
(see character 74).
77. Cephalic condyle of quadrate. 0, situated
dorsally, approximately level with dorsal margin of
prootic (Fig. 4 B–F). 1, situated ventrally, well below
level of dorsal margin of prootic (Fig. 4 A). C27,
T58, reworded.
78. Quadrate shaft. 0, inclined slightly anteroventrally. 1, inclined greatly anteroventrally (Fig. 4 A).
2, vertical (Fig. 4 B–D). 3, inclined posteroventrally
(Fig. 4 F). Ordered 1–0–2–3. This represents T55
(quadrate slanting anteroventrally\vertical\slanting posteroventrally), with their first state subdivided. The orientation of the quadrate influences
the position of the articular condyle for the mandible,
and hence mandible length, which has thus not been
included (see C34, Section VI). Discussed previously
by Kamal (1966). Although the preserved quadrates
of Pachyrhachis are angled slightly posteroventrally,
they might have been vertical in life since their lower
ends are displaced posteriorly (the lower jaws have
been pushed backwards relative to the upper jaws).
Similarly, figures of Anomochilus in Cundall and
Rossman (1993) depict the quadrate as either
slightly anteroventral, or vertical. These taxa have
each been coded with their two possible states.
79. Septomaxilla. 0, projects anterolaterally, overlapping lateral process of premaxilla and\or anterior
Michael S. Y. Lee and John D. Scanlon
tip of the maxilla (Fig. 2 B). 1, does not project
anterolaterally, not overlapping lateral process of
premaxilla or anterior tip of maxilla (Fig. 2 C–E).
This feature has been mentioned (Haas, 1930 a ;
Rieppel, 1977 ; Cundall and Shardo, 1995) but not
yet employed in a numerical phylogenetic analyses ;
at least in some colubroids, it may have an important
functional role in kinesis among snout bones (‘ rhinokinesis ’ ; Cundall and Shardo, 1995). In Cylindrophis,
Anomochilus and some Leptotyphlops the septomaxillary
process overlaps both the premaxilla and the maxilla
at their contact ; in some other taxa with state 1 the
maxilla and premaxilla are separated and the
septomaxilla overlaps only one or the other.
80. Dorsolateral flange of septomaxilla. 0, blunt,
without spine, expansion or calcified ligament (Fig.
7 F). 1, with spine projecting posterolaterally (Fig.
7 E). 2, with posterior expansion projecting posteromedially towards frontal (Fig. 7 D). 3, with calcified
ligament. Unordered, since the structures in states 1,
2 and 3 are so different that homology should not be
assumed. This character is a modification of Kluge’s
(1993 a) presence-absence character (49) for
pythons, and is similar to T6 (spine absent\short\
long). However, as the end of the spine is easily
broken in skull preparations, we do not employ the
distinction spine short\long. Also, although Tchernov et al. (2000) coded Anomochilus with a short spine,
it lacks a spine but instead possesses a calcified
ligament (Cundall and Rossman, 1993, Figs 5, 6).
T8 (posterior ventral edge of spine not in front\in
front of opening of Jacobson’s organ) is also
correlated with the length of the spine, and is not
considered as a separate character here.
81. Septomaxilla. 0, maxilla, but not septomaxilla,
contributes to posterior border of the external naris
(Fig. 4 B–F). 1, septomaxilla with lateral flange
contributing to the posterior border of the external
naris (Fig. 3 A, 4 A). In burrowing groups with
state 1, the septomaxilla partly or completely replaces the maxilla in reinforcing the snout laterally.
82. Septomaxilla-frontal contact. 0, posteromedial
flange of septomaxilla short, not contacting frontal
(Fig. 7 A, B). 1, posteromedial flange of septomaxilla
long, contacting frontal adjacent to midline on lower
part of interolfactory pillar (Fig. 7 C). T7. Initially
identified by Underwood (1967) and discussed by
McDowell (1975). The contact between the large
posterolateral flange of the septomaxilla, and the
frontal, is not considered homologous.
83. Fenestra for duct of Jacobson’s organ. 0, faces
ventrally. 1, faces posteroventrally. McDowell
(1972) noted that scolecophidians have the opening
Snake phylogeny
of the duct into the mouth (fenestra vomeronasalis
externa) facing more posteriorly than other snakes,
and this also applies to the bony fenestra concerned
here.
84. Vomer. 0, does not enter lateral margin of
fenestra for Jacobson’s organ (Fig. 3 A–E). 1, forms
posterior part of lateral margin of fenestra for
Jacobson’s organ (Fig. 3 F). T10 (after Groombridge, 1979 b), slightly rephrased. Tchernov et al.’s
(2000) definition (lateral margin formed by septomaxilla only, or septomaxilla and vomer) is inapplicable in the outgroups, which have a third state
(the lateral margin formed by maxilla).
85. Vomeronasal nerve. 0, does not pierce the
ridge on the vomer forming the posterior wall of the
vomeronasal organ. 1, pierces ridge via a single large
foramen (sometimes with one or two additional
small foramina). 2, pierces ridge through a cluster of
numerous small foramina. T11 rephrased, after
Groombridge (1979 b). This character has been more
precisely defined here to avoid confusion with the
ventral foramen on the vomer. Unordered.
86. Medial fenestra in vomeronasal cupola. 0,
posterior ends of sagittal flanges of vomer and
septomaxilla with small or no contact, and large
intervening fenestra or embayment. 1, posterior ends
of sagittal flanges of vomer and septomaxilla in
extensive contact, with small or no intervening
fenestra. C6, T9 ; see Cundall & Rossman (1993,
Fig. 6 B).
87. Palatine length (excluding posteromedial
process). 0, short anteroposteriorly (much shorter
than vomer). 1, intermediate in length anteroposteriorly (as long as vomer). 2, long anteroposteriorly
(much longer than vomer). Ordered 0–1–2. Modified from Pregill et al. (1986, character 31 ; vomer
shorter than or subequal to palatine\nearly twice
length of palatine) with addition of a state.
88. Horizontal (palatal) lamina of vomer. 0,
posterior end narrow, tapering to a point ; choana
wide (Fig. 3 A, B, D–F). 1, posterior end expanded ;
choana narrow (Fig. 3 C). Similar to T12 (horizontal
lamina posteriorly long and parallel-sided\short and
tapering), which has thus not been employed here.
After Groombridge (1979 b).
89. Vertical (posterior dorsal) lamina of vomer. 0,
small or absent. 1, well developed. T13, after
Groombridge (1979 b). The state labels assigned by
Tchernov et al. (2000) are here reversed because the
primitive state is their state 1, i.e. process small or
absent, the condition found in the outgroups (which
they coded as unknown). Scolecophidians, which
they also coded as unknown, have state 0.
361
90. Palatine-vomer contact. 0, medial (choanal)
process of palatine with extensive contact with vomer
(Fig. 3 A, B). 1, tiny point contact. 2, no contact
(Fig. 3 D–F). Ordered 0–1–2. A more finely subdivided version of T47 (contact well-defined\tiny or
absent). It, and the following, also represent a
subdivision and redefinition of C20 (choanal process
of palatine with simple contact\complex contact\no
contact with vomer). Previously discussed for Cylindrophis in Haas (1931 b), Xenopeltis in Bellairs (1949),
and boines in Kluge (1991 : character 51). The
palatines of madtsoiids (Scanlon, 1996, 1997 ; J. D.
Scanlon, unpublished observations) show facets
indicating extensive articulation, although the
vomer is unknown.
91. Palatine-vomer articulation. 0, medial (choanal or vomerine) processes of palatines do not project
ventromedially to separate vomers. 1, medial processes of palatines project ventromedially to separate
the posterior portions of the vomers. This is
presumably equivalent to the presence\absence of
state 1 of C20 ; a ‘ complex ’ articulation between the
choanal process and vomer. Not applicable in taxa
without a palatine-vomer contact (see previous
character). Polarity of this character is difficult to
determine because the palatine-vomer contact in the
outgroups is sutural and immobile rather than
mobile.
92. Medial (choanal or vomerine) process of
palatine. 0, anteroposteriorly broad plate of bone
(Fig. 3 A, B, D, E). 1, narrow finger-like process
(Fig. 3 C, F). T48. Many of the codings in Tchernov
et al. (2000) are not accepted here. Haasiophis (coded
with a narrow process) cannot be scored because the
structure is broken and largely obscured ; however,
the small exposed fragments are extensive anteroposteriorly, suggesting a broad process. Conversely,
Pachyrhachis and Dinilysia (both scored as unknown),
and madtsoiids (not considered by Tchernov et al.,
2000) all have a well-preserved, broad process (Estes
et al., 1970 ; Scanlon, 1997 ; Lee & Caldwell, 1998 ;
Scanlon & Lee, 2000). Leptotyphlopids, bolyeriines
and Acrochordus were scored with a narrow process,
but have a broad process. Similarly, pythonines and
colubroids were scored with a narrow process, but
both conditions are widespread in each group.
93. Medial (choanal or vomerine) process of
palatine. 0, without distinct, large anterior flange.
1, with distinct, large plate-like anterior flange,
abutting vomer posterolaterally. Because these large
anterior flanges do not project between the vomers
(Estes et al., 1970 ; Scanlon & Lee, 2000), this
character is independent of character 91 (medial
362
processes of palatine separating vomers). In Heloderma and Varanus, the entire medial process is
oriented anteriorly and the presence of a separate
anterior process cannot be objectively determined.
However, among other outgroups, Lanthanotus lacks
the process (as do most ‘ lizards ’), while in mosasaurs
the palatine and vomer are usually fused and the
character cannot be scored.
94. Anterior process of palatine. 0, no anterior
process, only medial (choanal or vomerine), lateral
(maxillary) and posteromedial (pterygoid) processes
present (Fig. 3 A). 1, narrow (‘ dentigerous ’) process
present, in addition to medial, lateral and posteromedial processes (Fig. 3 B–F). 2, wide horizontal
plate present. Unordered, since the three states do
not form a morphocline. This is equivalent to C19,
where the process was called the ‘ subvomerine
process ’. However, Cundall et al.’s (1993) term is
ambiguous because the medial (choanal) process can
also underlap the vomer, and ‘ anterior process ’ is
used here instead. Lee (1997, character 58) coded
mosasauroids with an anterior process. However, the
anteromedial process in mosasauroids, which extends
between the choanae, is homologous with the
choanal process in snakes and other squamates, and
not with the anterior process. This character is also
similar to T41 (anterior dentigerous process absent\
present). Tchernov et al. (2000) coded Pachyrhachis as
unknown, but it clearly has a large anterior process
which bears teeth (Haas, 1980 a ; Lee & Caldwell,
1998).
95. Anterior process of palatine. 0, contacting vomer-septomaxilla complex (Fig. 3 B, C).
1, not contacting vomer-septomaxilla complex (Fig.
3 D–F). T43, first used by McDowell (1975). Inapplicable if there is no anterior process of any kind
(see previous character). We recognize a palatinevomer contact in some taxa additional to those with
distinct articulatory facets (McDowell, 1975). As
noted by McDowell, even slight contact results in the
choanal passage being completely encircled by bone.
Contrary to Tchernov et al. (2000), the contact is
present in ungaliophiines, Acrochordus, some pythons,
and some colubroids (Hoffstetter & Gayrard, 1965 ;
McDowell, 1975, 1979 ; this study). Also, although
Cundall & Irish (1989) illustrated bolyeriines without a contact, and Tchernov et al. (2000) code them
accordingly, material examined of both genera seems
to show a tiny contact and bolyeriines are here coded
with both states.
96. Palatine-maxilla contact. 0, palatine sutured
to maxilla. 1, palatine meets maxilla in a loose joint.
2, palatine does not contact maxilla. Ordered 0–1–2.
Michael S. Y. Lee and John D. Scanlon
Similar to C21, T45, with an extra state (0) added to
account for the condition in the outgroups.
97. Lateral (maxillary) process of palatine.
0, situated in middle or anterior end of main body of
palatine (Fig. 3 B, C, E). 1, at posterior end of main
body of palatine (Fig. 3 D). T44, after Kluge (1991).
Tchernov et al.’s (2000) definition refers to the
‘ posterior end of the palatine ’, but their character
codings imply that the character refers to the end of
the main body of the palatine, excluding the medial
pterygoid process (Kluge, 1991, 1993 a). Tchernov et
al. (2000) coded Haasiophis with state 1 ; however, the
lateral process is largely obscured and cannot be
assumed to have extended to the posterior end of the
palatine. Haasiophis is here coded as unknown.
Leptotyphlopids, coded as unknown, have state 0
(e.g. List, 1966).
98. Lateral process of palatine. 0, pierced by
foramen. 1, lacking foramen. This foramen carries
the palatine (l sphenopalatine, l maxillary branch
of trigeminal) nerve in extant taxa. Similar to T46
(nerve pierces palatine\between palatine and prefrontal) ; previously employed by Underwood (1976,
character 56) and Kluge (1991, character 52), with
variation noted by Frazzetta (1959) and McDowell
(1975). State 0 applies to taxa with a deep or
constricted notch (e.g. Anomochilus ; Cundall &
Rossman, 1993) as well as those with an enclosed
foramen.
99. Articulation of palatine with pterygoid.
0, short. 1, long. T49, the long articulation results
mainly from the ‘ medial pterygoid process ’ of the
palatine of Kluge (1991, 1993 a). Although Haasiophis
was coded by Tchernov et al. (2000) with state 1, the
putative medial pterygoid process of the palatine
(preserved on only one side) is not continuous with
the palatine, and is as close to the parasphenoid as to
the pterygoid. It could therefore represent the edge
of the descending flange of the frontal. Haasiophis is
thus coded as unknown.
100. Pterygoid tooth row curvature. 0, concave
medially (Fig. 3 C, D). 1, straight to slightly convex
medially (Fig. 3 B, E, F). Estes et al. (1970) contrasted the straight pterygoid tooth row in ‘ recent
snakes ’ (Cylindrophis and Anilius) with the strongly
curved condition in Dinilysia and Lanthanotus.
101. Ectopterygoid process of pterygoid (Fig.
5 C–E). 0, well developed, a large rectangular or
triangular lateral process (Fig. 3 B). 1, poorly
developed, a small rounded lateral flange (Fig.
3 C–F). 2, absent (Fig. 3 A). Ordered 0–1–2. Part of
a subdivision of T52 (ectopterygoid process distinct
and articulates dorsally with ectopterygoid\process
Snake phylogeny
indistinct and articulates laterally with ectopterygoid\process indistinct and articulates dorsally with
ectopterygoid) into two separate characters, development of process (this character), and position of
ectopterygoid articulation (character 102). It cannot
be scored in dibamids, where the ectopterygoid is
anterior rather than lateral to the pterygoid. The
large recurved ectopterygoid processes of Haasiophis
(similar to those of madtsoiids and some anilioids ;
Scanlon, 1997) were thought by Tchernov et al.
(2000) to be parts of the prootics ; however, they are
continuous with the pterygoid and articulate with
the ectopterygoid. Published photographs (Haas,
1979, 1980 a) show that Pachyrhachis also possesses
large ectopterygoid processes (Scanlon, 1996), but
these have not been recognized by most authors
(Haas, 1979, 1980 a ; Lee & Caldwell, 1998) who
have considered them part of the large coronoid
processes which they closely overlap.
102. Ectopterygoid attachment to pterygoid.
0, anterior to basipterygoid process. 1, lateral to
basipterygoid process. 2, posterior to basipterygoid
process. Ordered 0–1–2. Not applicable in taxa
lacking a structural equivalent of the basipterygoid
process (including ‘ sphenoid wing ’).
103. Pterygoid quadrate ramus. 0, robust, platelike (Fig. 3 C–F). 1, gracile, rod-like (Fig. 3 A).
104. Pterygoid medial margin. 0, with distinct
medial spur in region of basicranial articulation.
1, with smooth medial bulge. 2, with straight margin.
Ordered 0–1–2. Estes et al. (1970, p. 43) mentioned
the ‘ internal process ’ of the pterygoid in Dinilysia,
but did not comment on conditions in other taxa.
McDowell & Bogert (1954, p. 62) presumed the
notch between the process and the quadrate ramus
of the pterygoid in mosasaurs is for the Eustachian
passage. However, the wide basicranial articulation
in mosasaurs (and by implication snakes) has also
been interpreted as correlated with loss of the
Eustachian passage (Russell, 1967).
105. Pterygoid quadrate ramus. 0, with a shallow
groove along ventromedial surface, or no groove
(Fig. 5 C). 1, with a very deep groove along
ventromedial surface, becoming dorsomedial posteriorly (Fig. 5 D, E). Kluge (1993 b, character 43)
distinguished three states of this character in erycines, after previous observations by others (e.g.
Frazzetta, 1966 ; Hoffstetter & Rage, 1972). Almost
identical to T51 (quadrate ramus ‘ robust, rounded
or triangular in cross-section, but without
groove ’\‘ blade-like and with distinct longitudinal
groove ’), which has not been employed as a separate
character here. Tchernov et al.’s (2000) character
363
states are also more ambiguous because neither
applies to the numerous taxa with blade-like rami,
but lacking a distinct groove. We adopt their coding
of Haasiophis with a groove, but note that the rami
are badly damaged on both sides. The ‘ groove ’ in
the left ramus resembles a crack in both the
photograph and the interpretative drawing, while
the photograph and interpretative drawing of the
fragment of the right element are rather dissimilar.
Both rami in Pachyrhachis are much better preserved
and show no grooves (Lee & Caldwell, 1998) but
this taxon was coded as unknown by Tchernov et al.
(2000).
106. Pterygoid quadrate ramus. 0, vertical or
oblique sheet. 1, approximately horizontal sheet.
Estes et al. (1970, p. 58) equated the condition in
Dinilysia (‘ vertical plate with a sharp ventral edge ’)
with that in lizards and Cylindrophis, contrasting it
with Anilius and boids ‘ which have the ventral edge
rotated medially ’. The condition in pythonines and
boines is indeterminate due to the great curvature of
the quadrate ramus.
107. Pterygoid quadrate ramus. 0, terminates
near jaw joint (Fig. 3 B–F). 1, projects posteriorly
well past jaw joint (Fig. 3 A). C24, reworded slightly.
See also Section VI (T64).
108. Ectopterygoid. 0, large (Fig. 2 B–F). 1, small.
2, absent (Fig. 3 A). Ordered 0–1–2. This, and the
following four characters, overlap with C22 which
combined both presence and contacts of the ectopterygoid into a single character (ectopterygoid in
contact with maxilla and pterygoid\in contact with
maxilla\in contact with pterygoid\not in contact
with either element\absent). That character was
also uninformative because it was unordered and
each derived state present in only one terminal
taxon.
109. Ectopterygoid-pterygoid contact. 0, clasps
pterygoid on both dorsal and ventral surfaces.
1, simple overlap on only ventral surface of pterygoid (Scanlon & Lee, 2000, Fig. 1 a). 2, simple overlap
on only dorsal surface of pterygoid (Fig. 5 C, D).
3, simple contact on only lateral edge only of pterygoid (Fig. 5 E). Unordered. Overlaps with C22 (see
previous character), and also part of a subdivision of
T52. The character state conditions in T52 (listed
under character 101) include only contact of
ectopterygoid with anterodorsal, mostly lateral, or
dorsal surfaces of pterygoid, and thus do not account
for the clasping contact typical of lizards, and the
ventral-only contact seen in madtsoiids and Haasiophis. State labels here corrected from Scanlon & Lee
(2000, legend to their Fig. 3).
364
110. Ectopterygoid-maxilla contact. 0, posterior
tip of the maxilla abuts ectopterygoid (Fig.
4 B, D, E). 1, posterior tip of the maxilla is lifted off
ectopterygoid and projects freely (Fig. 4 C). Discussed by McDowell (1975). This cannot be coded
in taxa lacking an ectopterygoid, and also in some
colubroids with highly aberrant maxillae (e.g.
viperids). However, in all basal colubroids with
primitive maxillae, the element usually projects at
least a slight (e.g. Xenodermus) to great (e.g. Fimbrios)
distance beyond the ectopterygoid. Note that the
drawings of Xenodermus reproduced in Figs 2 F, 3 F
and 4 F do not clearly show the free projection
observed in the material examined.
111. Ectopterygoid-maxilla contact. 0, anterior
end of ectopterygoid restricted to posteromedial
edge of maxilla (Fig. 5 C). 1, ectopterygoid invades
significantly the dorsal surface of the maxilla (Fig.
5 D, E). T54. Tchernov et al. (2000) coded Haasiophis
with state 1, but it appears to have state 0. The left
ectopterygoid is well preserved, medial to the
coronoid process, and is correctly identified in
ventral view (but unlabelled in dorsal view) : it
barely overlaps the posterior end of the maxilla. The
right element is poorly preserved ; what they interpret as the anterior portion overlapping the
maxilla makes this element twice as large as the left,
and is a misidentified postorbitofrontal or maxillary
fragment. Pachyrhachis was also coded as possessing a
large dorsal overlap, but the amount of overlap is
not visible in either specimen (Lee & Caldwell,
1998).
112. Ectopterygoid shape. 0, distal end of ectopterygoid with single anterior process projecting
dorsally along maxilla (Fig. 5 C). 1, distal end of
ectopterygoid with two anterior processes projecting
dorsally along maxilla (Fig. 5 D, E). Not applicable
in taxa where this region of the maxilla is invaded by
the jugal, or in amphisbaenians and dibamids (where
the ectopterygoid either has a complex suture with
the maxilla, or projects ventrally beneath the
maxilla). The form of the maxillary end of the
ectopterygoid was discussed by Frazzetta (1966) and
McDowell (1975) in Pythoninae ; Kluge used variation in this area in analyses of Boinae (1991,
character 32), Pythoninae (1993 a, character 38) and
Erycinae (1993 b, characters 24–25). In Haasiophis,
the anterior margin of both ectopterygoids is
obscured ; note that the ‘ anterior margin ’ of the
right ectopterygoid was misidentified in Tchernov et
al. (2000) (see previous character).
113. Lateral edge of the ectopterygoid. 0, straight
or slightly curved, lacking distinct angulation.
Michael S. Y. Lee and John D. Scanlon
1, distinctly angulated, a distinct ‘ corner ’ present
between the anterior (parasagittally oriented) and
posterior (posteromedially oriented) portions of the
lateral margin. State definitions expanded from T53
(lateral edge straight\angulated). Tchernov et al.
(2000) coded taxa with a slightly curved (convex)
ectopterygoid margin as possessing state 0, and we
have therefore expanded state 0 to include this
morphology. However, contrary to Tchernov et al.
(2000), Pachyrhachis has no angulation (Lee &
Caldwell, 1998), nor does Haasiophis if the wellpreserved left element is considered (Tchernov et al.,
2000, Fig. 1). The postulated angulation in the right
element is the angle between the smoothly curved
edge of the true ectopterygoid and the straight edge
of another element (see character 111). Varanus,
Lanthanotus and mosasauroids lack the angulation
(contra Tchernov et al., 2000).
114. Cultriform process. 0, anterior one-third
broad posteriorly and tapering anteriorly (Fig. 9 E).
1, anterior one-third narrow throughout (Fig.
9 D, F). Not applicable in the outgroups, because
they all possess a much shorter cultriform process
(i.e. the part corresponding to the anterior one-third
in snakes is not ossified). Variation among primitive
snakes is discussed by Estes et al. (1970) and Rieppel
(1977).
115. Interchoanal keel of cultriform process.
0, absent. 1, present, a sagittal flange extending
ventrally between the medial processes of the
palatines. Not applicable in taxa with a short
cultriform process. This is also difficult to score in
taxa with an extremely narrow anterior cultriform
process (e.g. Anilius, tropidophiines, some boines,
some erycines), because the distinction between keel
and cultriform process is uncertain. This corresponds
to C17, except that the structure was termed an
interchoanal ‘ process ’ in that study. Because of its
morphology, the term ‘ keel ’ appears to be more
descriptive. T80 (interchoanal process absent\
broad-based\narrow-based) appears to be equivalent to this character, but the meaning of broad
and narrow here is not clear, because the keel is
always a blade-shaped structure. It presumably does
not refer to the width of the cultriform process itself
(parabasisphenoid rostrum), since that trait is
covered by T81. In some Liasis (Pythoninae), the
anterior tip of the parasphenoid bears a distinct
ventral keel similar to that illustrated for Anomochilus
by Cundall & Rossman (1993), but this is not
present in more basal pythons (Scanlon, 1996).
116. Parabasisphenoid transverse width behind
frontal descending flanges. 0, narrow, without
Snake phylogeny
concave ventral surface. 1, broad and ventrally
concave. A modification of T81 (parabasisphenoid
behind optic foramen narrow\broad and ventrally
concave). We use the posterior limit of the frontal
flanges as a landmark, rather than the optic foramen
which is absent in the outgroups and variable in
position within the ingroup (see character 61).
Under Tchernov et al.’s (2000) definition, taxa
lacking an optic foramen (i.e. the outgroups) should
be scored as inapplicable, although these were scored
with state 0.
117. Basipterygoid process. 0, prominent, i.e. a
pedicel or flange projecting far laterally with distinct
distal facet (Fig. 9 D, F). 1, weak, consisting of a crest
or mound without a distinct distal facet (Fig. 9 E).
2, absent. Ordered 0–1–2. This character is a muchused though contentious one in snake phylogeny.
These states seem to correspond to the following
states in T82 : basipterygoid processes distinct\
shallow and elongate facet present\facet absent. The
character also overlaps with character 56 of Kluge
(1991 : parasphenoid wing absent or weak\large
without facet\large with facet), character 51 of
Kluge (1993 a : ‘ pterygoid process ’ of sphenoid
absent\short\tall), and character 33 of Lee (1997 :
basipterygoid process long\short). Kluge used different terms for the facet in boines (Kluge, 1991) and
erycines (Kluge, 1993 b). Prominence of the process
was assigned a score from 0–100 by Underwood
(1976, character 63).
The basipterygoid process of lizards is considered
by some authors (e.g. McDowell, 1967, 1975) to be
homologous in whole or part with both the ‘ basipterygoid process ’ and ‘ parasphenoid wing ’ in some
snakes ; others (e.g. Rieppel, 1988 ; Kluge, 1991)
deny any such homology, but their arguments based
on details of embryology and soft tissues are
problematic if extended to fossils, or recent taxa
represented by dry skulls. The term ‘ basipterygoid
process ’ is used here for any ventrolateral projection
in the basicranial region of the sphenoid. We code
Haasiophis as possessing a basipterygoid process on
the basis of statements by Tchernov et al. (2000).
However, their assertion that this projection is
homologous with the process in booids, but not that
in lizards, cannot be supported because it is only
poorly visible in X-rays. It is thus difficult to see how
it could be argued to be more similar (and thus more
likely homologous ?) with a booid process than with
a lizard process. A more reasonable interpretation
would be to treat the very similar processes in
lizards, booids and Haasiophis as potentially homologous, and let character congruence arbitrate.
365
Kluge (1991, 1993 a), applying such a test, considered it ‘ likely that such a projection was lost early
in snake history, and the prominence in more derived
snakes requires a different name in order to
underscore its independent history ’ (Kluge, 1993 a,
p. 25). While our phylogenetic results support
Kluge’s (1993 a) interpretation of character evolution for extant snakes, the terminological proposal
would necessitate wholesale renaming of homoplasious morphological traits and instability in these
names with each new proposed phylogenetic arrangement. For present purposes we prefer to name
traits on the basis of topological relationships,
without necessarily implying homology.
118. Distal surfaces (facets or crests) of basipterygoid processes. 0, long axes oriented obliquely, or
transversely in ventral view (Fig. 9 F). 1, long axes
oriented parasagitally in ventral view (Fig. 9 D).
Difficult to score in many taxa with reduced
basipterygoid processes. The extensive variation in
orientation of the facet or crest has apparently never
been used as a phylogenetic character.
119. Parabasisphenoid (l sphenoid). 0, sphenoid
wing absent, no triangular dorsolateral prominence
lateral to alar process of dorsum sellae. 1, sphenoid
wing present as triangular prominence distinct from
alar process, extending up anterior margin of prootic
below the trigeminal notch. Similar to T79 (‘ parabasisphenoid wings absent\present ’), but we attempt to base the character on explicit structural
criteria applicable to an external view of the
articulated skull and independent of the course of the
vidian canal (e.g. character 124). The alar process
(Oelrich, 1956 ; Russell, 1967 ; Bellairs & Kamal,
1981, Fig. 41 ; ‘ clinoid process ’ of Rieppel & Zaher,
2000 a) is the upturned lateral end of the dorsum
sellae, and in many lizards (e.g. Varanus, and some
but not all mosasaurs ; Lingham-Soliar, 1995, Rieppel & Zaher, 2000 a) it is extensively exposed in
lateral view, anterior to the prootic, and thus with
topographic relationships superficially equivalent to
the sphenoid wing of alethinophidians (McDowell,
1967, 1974 ; ‘ parabasisphenoid wing ’ of Tchernov et
al., 2000). In snakes the dorsum sellae forms a low
transverse ridge rather than a prominent crest (see
characters 128, 129 below) but, still being geometrically saddle-shaped, retains upturned lateral
prominences. These structural equivalents of the
alar processes have apparently never been identified
previously in snakes, and are usually hidden laterally
by the prootic, but are readily identifiable in an
oblique internal view of the sphenoid (e.g. Scanlon
& Lee, 2000, Fig. 2 b). The scolecophidian and
366
alethinophidian conditions are contrasted by
McDowell (1967), who suggested that the triangular
prominence of alethinophidians was formed by a
‘ basipterygoid process that has expanded upward
and also fused with the epipterygoid ’, whereas the
smooth border in scolecophidians reflects the loss of
both basipterygoid process and epipterygoid. However, contrary to McDowell (1967), List’s (1966)
figures show a prominence between parietal and
prootic in a number of scolecophidians : such flanges
exist in some typhlopids (e.g. Rhinotyphlops schlegelii,
Typhlops punctatus, M. S. Y. Lee, personal observations ; T. muelleri, Rieppel, 1979 a, Fig. 1). McDowell
(1987) also showed a triangular prominence in a
drawing of the braincase of Dinilysia, which is
consistent with our observations although no such
structure was shown by Estes et al. (1970) ; Dinilysia
is here coded as unknown.
120. Ventral surface of parabasisphenoid.
0, smooth posteriorly, lacking keel (Fig. 9 D). 1, with
median keel in posterior region, at level of posterior
openings of vidian canals (Fig. 9 E, F). 2, with pair of
parasagittal keels. Unordered. This character is
distinct from the ‘ interchoanal keel ’ of the cultriform
process, which is situated much more anteriorly
(character 115). Variation in depth of the keel was
used by McDowell (1975) and Kluge (1991, character 57 ; 1993 a, character 61) in pythons and boines.
T 77 (basioccipital and parabasisphenoid without\
with sagittal ridges), with the latter state subdivided.
121. Basioccipital-parabasisphenoid suture. 0, positioned midway between fenestra ovalis and trigeminal foramen. 1, posteriorly positioned, closer to
fenestra ovalis than to trigeminal foramen. 2, anteriorly positioned, closer to trigeminal foramen
than to fenestra ovalis. Ordered 1–0–2. T78 (suture
under fenestra ovalis\under trigeminal foramen),
with an extra state added to accommodate the many
taxa where the suture lies in an intermediate
position. Haasiophis is coded as unknown because
neither the fenestra ovalis nor the trigeminal foramen
is exposed. Tchernov et al. (2000) misidentified part
of the pterygoid as the prootic margin of the
trigeminal foramen, and thus coded Haasiophis with
a posterior suture (see character 136).
122. Basioccipital. 0, with short posterolateral
flanges (Fig. 9 E, F). 1, with long posterolateral
processes (Fig. 9 D).
123. Posterior opening of vidian canal. 0, within
basisphenoid, not bordered by prootic. 1, partly
bordered by prootic (i.e. on basisphenoid-prootic
suture : Fig. 9 A) or entirely within prootic. T69
subdivides state 1 into two conditions. However, this
Michael S. Y. Lee and John D. Scanlon
subdivision is not phylogenetically informative in the
context of these terminal taxa because one of the
conditions (vidian canal opening entirely within
prootic) is an autapomorphy of uropeltids. Also,
many of Tchernov et al.’s (2000) codings are
incorrect. They coded Dinilysia as having a vidian
canal on the suture, but it actually lies entirely within
the basisphenoid (Estes et al., 1970). Similarly,
Cylindrophis, Anilius, pythons, erycines, boines, Acrochordus and bolyeriines were coded as uniformly
possessing a canal on the suture, but in all these
groups the ‘ within basisphenoid ’ condition is common (and, in bolyeriines, ubiquitous). Conversely,
basal colubroids were coded as all having the ‘ within
basisphenoid ’ condition, but many have the canal
on the suture.
124. Vidian canal. 0, does not open intracranially.
1, opens intracranially, emerging on internal surface
of sphenoid (primary opening) then emerging
externally on sphenoid-parietal suture (secondary
opening). Rieppel (1979 b) considered state 1 plesiomorphic for snakes, but the presence of state 0 in
lizards, madtsoiids, Dinilysia and most scolecophidians suggests otherwise.
125. Vidian canals. 0, symmetrical. 1, asymmetrical, left larger than right or vice versa. Used by
Underwood (1976, character 66).
126. Hypophysial pit (sella turcica). 0, without
distinct anterior bony boundary. 1, bounded anteriorly by distinct ridge. This and the following
three characters are discussed and\or illustrated for
some taxa by Rieppel (1979 a, b).
127. Cerebral carotid artery. 0, opens into
posterior region of hypophysial pit, near posterior
transverse wall. 1, opens into middle region of
hypophysial pit, well away from posterior transverse
wall.
128. Dorsum sellae (crista sellaris). 0, well developed. 1, greatly reduced.
129. Dorsum sellae (crista sellaris). 0, oriented
anterodorsally, overhanging the posterior portion of
hypophysial pit. 1, oriented dorsally, not overhanging hypophysial pit. Not applicable if dorsum sellae
is weakly developed.
130. Laterosphenoid. 0, absent, V2 and V3 exits of
trigeminal foramen confluent. 1, present, fuses to
prootic forming vertical bar between exits of V2 and
V3 (Fig. 9 A). C14, T65. The development and
homology of the ‘ laterosphenoid ’ is discussed by
various authors including Kamal & Hammouda
(1965), Rieppel (1977), and Bellairs & Kamal
(1981). Cundall et al. (1993) noted that presence of
the laterosphenoid is variable in some alethino-
Snake phylogeny
phidians (e.g. between the two sides of the skull in
some pythonines, bolyeriines, and colubroids), but it
appears to be primitively present in all these groups.
Tchernov et al. (2000) coded the laterosphenoid as
present in Haasiophis, but we regard this as unknown
because the structure is not shown in their photographs nor in their interpretive drawings (the lateral
projections they identify as part of the ‘ prootics ’ are
the ectopterygoid processes of the pterygoids, see
character 101).
131. Alar process of prootic. 0, long distinct
process projecting anteriorly well past trigeminal
(V) foramen. 1, short process not projecting past
trigeminal (V) foramen. Rieppel & Zaher (2000 a)
deny the presence of an alar process in snakes but
apparently did not consider any fossil snakes, and
their developmental criteria for recognising such a
structure are difficult to apply. Photographs of
Haasiophis show a long alar process of the prootic
between the supratemporal and the parietal ; these
are visible on both sides and are symmetrical, but are
interepreted as parts of the parietal by Tchernov et
al. (2000, Fig. 1). The elements interpreted as
‘ prootics ’ by these authors are the lateral flanges of
the pterygoid : they are clearly continous with the
pterygoid (as shown by both photographs and
specimen drawings) and articulate with the ectopterygoid.
132. Trigeminal foramen, anterior margin.
0, closed by parietal at least medially ; upper and
lower anterior processes of prootic may touch superficially lateral to the parietal. 1, closed by prootic,
deep contact or fusion of prootic processes excludes
parietal from opening. Reworded slightly from T70.
Contrary to Tchernov et al. (2000), some basal
pythonines have state 1.
133. Exit foramen for the facial (VII) nerve
(hyomandibular branch, if distinct). 0, located
outside the opening for the mandibular branch of the
trigeminal nerve (V3, or V2jV3). 1, located within
the opening. T67. Tchernov et al.’s (2000) codings
are largely adopted here, but with the caveat that
because the posterior limit of the trigeminal foramen
usually lacks a distinct bounding ridge, ‘ inside ’ vs.
‘ outside ’ may be too subjective for repeatability.
Varanoids and dibamids were scored as unknown by
Tchernov et al. (2000), but they exhibit state 0 [e.g.
Rieppel (1980) for varanoids ; and Rieppel (1984)
and Greer (1985) for dibamids]. Also, Loxocemus was
scored with state 1, but exhibits state 0 (e.g. Rieppel,
1977).
134. Sulcus connecting exit foramen of palatine
branch of facial (VII) nerve with posterior opening
367
of vidian canal. 0, weakly recessed, with shallow and
smooth margins. 1, deeply recessed, with sharply
defined anterior and often also posterior margins
(Fig. 9 A). 2, embedded, closed laterally forming a
tunnel in prootic. Ordered 0–1–2. Similar to T68
(sulcus with anterior and posterior margins distinct\
posterior margin indistinct\embedded), but rephrased because the anterior and posterior margins
are usually similarly developed. Our codings are
very different to those of Tchernov et al. (2000),
possibly because of the slightly different criteria
used. Difficult to code in dibamids, where the exit of
the VII nerve and the vidian canal are closely
juxtaposed. See also character 123 above.
135. Crista circumfenestralis. 0, juxtastapedial
recess bordered by crests which may extend directly
laterally but do not converge. 1, juxtastapedial
recess surrounded by crests which converge to partly
enclose stapedial footplate, much of footplate remains exposed laterally (Fig. 9 A). 2, juxtastapedial
recess surrounded by crests which converge to largely
enclose stapedial footplate. Ordered 0–1–2. This
character represents a more finely divided version of
T74 (posterior enclosure of the juxtastapedial recess
absent\present). The crista circumfenestralis has
been discussed by Estes et al. (1970), McDowell
(1974, 1975, 1979, 1987), Rieppel (1979 a, c, 1988),
and variation among erycines used in the analysis of
Kluge (1993 b, character 31). As in other osteological
characters, we prefer to use purely osteological
landmarks rather than soft tissue arrangements
which are not determinable in fossils, such as the
extent of the juxtastapedial sinus (cf. Rieppel, 1988).
136.
Supratemporal-supraoccipital
contact.
0, supratemporal and supraoccipital separated by
dorsal exposures of parietal and exoccipital.
1, supratemporal and supraoccipital separated by
dorsal exposures of prootic, parietal and exoccipital
(Fig. 9 B). 2, supratemporal and supraoccipital in
contact (Fig. 9 C). Unordered. Not applicable in
taxa lacking a supratemporal or a discrete supraoccipital. This character is related to T66 (prootic
dorsally exposed\fully concealed by supratemporal
or parietal). However, the prootic is not fully
concealed in nearly all of the taxa Tchernov et al.
(2000) code with their second state (the main portion
remains exposed lateral to the parietal-supratemporal unit). Thus, their character presumably refers
only to the exposure or concealment of the prootic
medial to the parietal-supratemporal unit (state 1
here). Rieppel (1977) termed this structure in
anilioids the ‘ intercalary element ’, but examination
of partially disarticulated skulls reveals it is part of
368
the prootic. He claimed the presence of two
intercalary elements in Anilius : the more anterior is
actually a sliver of the prootic, the more posterior is
part of the exoccipital which is separated by a groove
from the rest of the element ; this groove was
presumably mistaken for a suture. These dorsal
prootic exposures were also illustrated in Loxocemus
(Rieppel, 1977, Fig. 17), but the taxon was coded by
Tchernov et al. (2000) as lacking the exposure. In
Haasiophis, slivers of prootic appear to be present
medial to the supratemporal, but were instead
identified as parts of the parietal by Tchernov et al.
(2000). Conversely, what Tchernov et al. (2000)
identified as the prootics are the lateral (ectopterygoid) processes of the pterygoid (see character
101). Also, although they coded Pachyrhachis with
state 1, the supratemporals appear displaced anteromedially and this character is not determinable.
Under character T40, Tchernov et al. (2000)
acknowledged that the relative position of the
supratemporal cannot be determined due to such
displacement.
137. Paroccipital process. 0, long process. 1, distinct flange. 2, indistinct bump or absent. Ordered
0–1–2. The paroccipital process is reduced in all
snakes relative to the condition in typical ‘ lizards ’,
but the above states can still be distinguished within
snakes. Estes et al. (1970, p. 53) considered the
condition in Dinilysia as distinct from both lizards
and anilioid snakes, but also as a possible intermediate. T 73 seems to refer to the paroccipital
process : posterior opisthotic process, located above
the stapes and below the supratemporal, is absent\
present. However, the codings are puzzling, because
Loxocemus was coded as having the process (it is
absent), while Dinilysia was coded as lacking the
process and lizards as unknown (they both have a
long paroccipital process). Also, Pachyrhachis was
coded with the process, but this region is unknown.
138. Supraoccipital. 0, external (dorsoposterior)
surface with no, or very weak transverse ridge (Fig.
9 B). 1, external surface with moderate transverse
ridge (Fig. 9 C). 2, external surface with very high
transverse ridge. Ordered 0–1–2.
139. Supraoccipital dorsal exposure. 0, long,
sagittal dimension more than 50 % transverse dimension (Fig. 9 C). 1, short, sagittal dimension less
than 50 % transverse dimension (Fig. 9 B). The
supraoccipital is fused to the opisthotic-exoccipital in
some typhlopids and some uropeltines ; however,
those with distinct supraoccipitals exhibit state 1. In
taxa with paired supraoccipitals, the transverse
dimension is the distance across both supraoccipitals.
Michael S. Y. Lee and John D. Scanlon
Underwood’s (1976) character 50 is somewhat
similar (lateral extension of supraoccipital : taper\
parallel sided\expanded).
140. Supraoccipital-prootic contact. 0, narrow,
less than half supraoccipital-parietal contact.
1, broad, subequal to or as long as supraoccipitalparietal contact. T62. Varanoids, amphisbaenians
and dibamids were coded as unknown by Tchernov
et al. (2000). This character is inapplicable in
varanoids because the supraoccipital meets the
parietal ventrally rather than (as in snakes) posteriorly, and in amphisbaenians because the prootic
is not a discrete element (Gans, 1978). However, the
character can be scored in dibamids, which have a
supraoccipital-parietal contact similar to snakes and
exhibit state 0 (Rieppel, 1984 ; Greer, 1985).
141. Exoccipital separation dorsal to foramen
magnum. 0, exoccipitals widely separated above
foramen magnum. 1, exoccipitals with point contact
above foramen magnum (Fig. 2 B). 2, exoccipitals in
extensive median contact above foramen magnum
(Fig. 2 A, C–F). Ordered 0–1–2. State 2 has long
been interpreted as diagnostic of living snakes (e.g.
Gilmore, 1938), but state 1 occurs in a few derived
colubroids (Rage, 1984, Rieppel, 1988), and a
(probably juvenile) Cylindrophis maculatus (Estes et al.,
1970, Fig. 9). The exoccipitals are shown clearly
to be separated above the foramen magnum in
Haasiophis (Tchernov et al., 2000, Fig. 1), supporting
a previous interpretation of Pachyrhachis (Lee &
Caldwell, 1998). However, this primitive character
in Haasiophis was dismissed as ‘ an artifact of
preservation ’ (Tchernov et al., 2000, p. 2011), even
though all elements in that region appear undisturbed (their Fig. 1). Furthermore, these authors
code Haasiophis for all other characters in this region
apparently without reservations about postmortem
disturbance : the contacts of the supratemporal,
supraoccipital, prootic, and opisthotic (T40, T62,
T66, T72).
142. Exoccipital separation ventral to foramen
magnum. 0, exoccipitals separated below foramen
magnum, not in contact along dorsal midline of
occipital condyle (Fig. 7 I). 1, exoccipitals in contact
below foramen magnum, along dorsal midline of
occipital condyle (Fig. 7 J). Used for erycines by
Kluge (1993 b, character 30). This and the following
character represent the subdivision of T 76 (exoccipitals separated & fovea dentis deep\exoccipitals
meeting and fovea dentis shallow\exoccipitals fused
and fovea dentis absent) into two characters,
exoccipital contact and fovea dentis. If only the
exoccipital contact is considered, T76 can be seen to
Snake phylogeny
have subdivided our state 1 into two states (exoccipitals meeting but unfused\fused) : however, this
division is uninformative here because the fused state
is an autapomorphy of a single terminal taxon. State
1 is present among typhlopids only in some Ramphotyphlops (Greer, 1997), so state 0 (found in other
Ramphotyphlops, and other genera) is presumably
primitive. Not applicable in dibamids, where the
exoccipitals fuse with the basioccipital (Greer, 1985).
143. Occipital condyle. 0, dorsal surface deeply
concave, i.e. with deep ‘ fovea dentis ’ (Fig. 7 I).
1, dorsal surface slightly concave at most, i.e. with
shallow or no ‘ fovea dentis ’ (Fig. 7 J). T76 (see
above) subdivided state 1 into two states, fovea
dentis shallow\absent. However, this subdivision is
uninformative here because the latter substate
characterises only a single terminal taxon.
144. Stapedial shaft. 0, straight. 1, angulated.
T60, after McDowell (1975). Tchernov et al. (2000)
coded Pachyrhachis with state 0 and Dinilysia with
state 1. Because of the uncertainty over the identity
of the stapes in Pachyrhachis (Zaher & Rieppel, 1999)
this character is here treated as unknown. Contrary
to Tchernov et al. (2000), Loxocemus has a straight
shaft.
145. Stapedial shaft. 0, slender and longer than
diameter of stapedial footplate. 1, thick and not
longer than diameter of footplate. T61.
146. Distal end of stapes. 0, associated with dorsal
tip of suprastapedial process of quadrate. 1, associated with ventral end of suprastapedial process and
dorsal end of quadrate shaft, i.e. cephalic condyle.
2, associated with middle or ventral half of quadrate
shaft. Ordered 0–1–2. Not applicable in the outgroups, where the stapes is not directly associated
with the quadrate (see character 73 above). This
character corresponds to C33. C25 (stapes associated\not associated with quadrate cephalic condyle)
is redundant with this one and thus not used here.
T59 is similar : ‘ stylohyal (distal tip of dorsal
stapedial process) fuses to the posterior tip of the
suprastapedial process of quadrate (0), or it fuses to
the ventral aspect of a reduced (embryonic) suprastapedial process (1), or the stylohyal fuses directly
to the shaft of the quadrate (2) ’. Tchernov et al.’s
(2000) definition is problematic because it involves
(unnecessarily) features of embryology not available
for any fossils or most extant taxa. Furthermore,
scolecophidians were coded with state 0, yet lack a
stylohyal which fuses to the quadrate (and thus
should be treated as inapplicable if fused stylohyal
position, rather than stapes orientation, is used).
Booids were coded with state 1, yet the stapes (and
369
stylohyal) is positioned in the middle of the quadrate
shaft in these taxa, well below the suprastapedial
process if one is considered to exist (i.e. state 2).
(b) Mandible
147. Dentary length. 0, dentary long, more than
40 % of main mandible length, i.e., length excluding
retroarticular process (Fig. 4 B–F). 1, dentary short,
less than 40 % of main mandible length (Fig. 4 A).
This represents a modification of C29 (dentary
35 % or 35 % of skull length). As defined, C29
(dentary length relative to skull) was largely redundant with C34 (mandible length relative to
skull). C29 is accordingly modified here to make it
more independent, although clearly length of the
dentary, and total mandible length, will always be
partly correlated.
148. Mental foramina on lateral surface of
dentary. 0, two or more (Fig. 4 A, C). 1, one
(Fig. 4 B, D–F). Discussed by Hoffstetter (1960),
McDowell (1987). No snakes have more than three
(Hoffstetter & Gayrard, 1965 ; McDowell, 1979 ;
Rage & Prasad, 1992 ; Scanlon, 1997). Pachyrhachis is
coded as unknown, as the conflict among reports
(beginning with Haas, 1979, 1980 a) shows that
the dentaries are too poorly preserved for reliable
observation. Madtsoiidae are now coded as polymorphic, since recent preparation of dentaries of
Wonambi naracoortensis confirms that the foramina are
single ( J. D. Scanlon, personal observations, 2001),
though multiple in all other known madtsoiid
specimens. The specimen of Xenopeltis illustrated in
Fig. 4 C is the only one seen with two small foramina
in addition to the usual single large one, so this taxon
is scored with state 1.
149. Posterolateral margin of dentary. 0, notch
absent, posterolateral margin of dentary straight or
slightly concave, dorsoposterior and ventroposterior
processes indistinct (Fig. 4 A). 1, with shallow notch,
processes short. 2, with deep notch, processes long
(Fig. 4 B–F). Ordered 0–1–2. T87 (‘ dentigerous
process of dentary absent\short\long ’) overlaps with
this and the next character and is thus not used here.
150. Dentary posterior margin. 0, dorsal posterior
process does not extend much further than ventral
posterior processes (Fig. 4 B, D, E). 1, dorsal process
extends much further posteriorly than ventral process (Fig. 4 C, F). Discussed by McDowell (1987).
151. Posteromedial shelf of dentary. 0, not exposed
medially. 1, exposed medially. Presence of a shelflike process level with the posterodorsal limit of the
the intramandibular septum is more widespread
370
(e.g. in Pachyophis, Lee, Caldwell & Scanlon, 1999 b) ;
this character relates to whether it is exposed, or
covered by the splenial and\or coronoid in the
articulated mandible.
152. Meckel’s canal (groove). 0, lacks floor
anteriorly, open ventrally anterior to level of anterior
inferior alveolar foramen. 1, floored by a horizontal
ventral lamina for its full length. 2, enclosed
anteriorly, with ventral and medial lamina. Ordered
0–1–2. Modified from Estes et al. (1988, character
57). State labels have been corrected from the matrix
of Scanlon & Lee (2000).
153. Splenial. 0, splenial present as discrete
element. 1, splenial not present as discrete element.
Absence in Anomalepididae follows Haas (1964),
not List (1966).
154. Splenial size. 0, small, extends no more than
50 % of distance from intramandibular joint to
symphysis. 1, large, extends more than 50 % of
distance from intramandibular joint to symphysis.
Taxa lacking a discrete splenial (previous character)
are scored as inapplicable, since this could occur via
fusion directly from either state. Scored in madtsoiids
based on the splenial facet on the ventromedial
margin of the dentary.
155. Splenial-angular joint. 0, vertical in medial
view. 1, highly oblique in medial view. Only scorable
in taxa with a straight, simple joint.
156. Foramen within splenial (l inferior alveolar
foramen). 0, present (Fig. 6 D). 1, absent (Fig. 6 C).
157. Dorsal margin of splenial. 0, deeply notched,
posterior region of notch bordered dorsally by
anterodorsal spine (Fig. 6 D). 1, moderately notched,
posterior region of notch not bordered dorsally by
anterodorsal spine (Fig. 6 C). 2, smooth, not notched.
Ordered 0–1–2.
158. Splenial-coronoid contact. 0, posterior end of
splenial in broad contact with coronoid. 1, posterior
end of splenial only just reaches coronoid. 2, posterior
end of splenial does not contact coronoid. Not
applicable in taxa lacking either element. Ordered
0–1–2. Contact versus separation of splenial and
coronoid in booids was used as a binary character by
Kluge (1991, character 60 ; 1993 b, character 49).
159. Splenial lateral exposure. 0, Anterior portion
of splenial not exposed laterally (Fig. 4 B–F).
1, Anterior portion of splenial greatly exposed laterally (Fig. 4 A).
160. Coronoid. 0, coronoid large and distinct (Fig.
4 A–D). 1, coronoid greatly reduced and sometimes
fused to compound. 2, coronoid never present as
distinct element (Fig. 4 E, F). Ordered 0–1–2. Similar to W16 (coronoid present\reduced\absent), a
Michael S. Y. Lee and John D. Scanlon
more finely subdivided version of T84 (coronoid
present\absent) ; see also Kluge (1991, character 60 ;
1993 b, character 49). The large coronoid in Xenopeltis is very closely applied to the surangular, and
has been previously erroneously described as absent
(e.g. Rieppel, 1977 ; Frazzetta, 1999). The reduced
element in tropidophiines is similarly closely applied
and often partly fused to the surangular (e.g.
McDowell, 1975, Fig. 4). A distinct anteromedial
shelf of the compound in Acrochordus and many
colubroids resembles the fused coronoid in tropidophiines, so that a likely homologue of the coronoid
bone can be identified in many taxa with state 2.
161. Coronoid. 0, with posteroventral process or
expansion. 1, without posteroventral process or
expansion. T85. Tchernov et al. (2000) coded
Pachyrhachis as unknown and Haasiophis with state 1.
However, the coronoid in Pachyrhachis has a distinct
expansion : even though the more medial parts of the
putative coronoid might be the lateral flange of
the pterygoid, the unequivocally identifiable part
of the coronoid shows state 0 (see the long left sliver
in Lee & Caldwell, 1998, Fig. 4). Similarly, Haasiophis has a posteroventral expansion ; although it was
coded as lacking the expansion, other taxa with a
smaller expansion (e.g. Anomochilus ; Cundall &
Rossman, 1993) were coded as possessing it. Also,
tropidophiines and erycines were coded as unknown (l inapplicable ?), yet tropidophiines and
some erycines possess coronoids (see previous
character) and exhibit state 1.
162. Coronoid lateral exposure. 0, coronoid
overlaps lateral surface of surangular and is exposed
in lateral view. 1, coronoid does not overlap lateral
surface of surangular, but projects dorsally beyond it
and is thus well exposed in lateral view (Fig. 4 A, B).
2, coronoid entirely medial to surangular and is
largely covered in lateral view (Fig. 4 C, D). Ordered
0–1–2. This represents a more finely subdivided
version of C30 (coronoid exposed\not exposed
laterally). A character similar to states 1 and 2 has
been used by Kluge (1991, character 62 ; 1993 a,
character 68), but states 0 and 1 have only been
distinguished by Scanlon (1996).
163. Coronoid-angular contact. 0, coronoid and
angular separated by prearticular, or prearticular
portion of compound bone. 1, coronoid contacts
angular. Character 61 of Kluge (1991).
164. Coronoid process. 0, well developed, distinct
projection lateral to adductor fossa (Fig. 4 A, B, D).
1, poorly developed or absent, smooth rounded crest
at most (Fig. 4 C, E, F). T86 is ‘ coronoid process
formed by coronoid only\coronoidjcompound\
Snake phylogeny
compound only (i.e. coronoid absent) ’. We have
subdivided this character into two characters
(coronoid process, and coronoid bone), since a
process is present in some taxa which lack the
coronoid bone, and a coronoid bone is present in
some taxa lacking a process. Note that the prominent
flange medial to the adductor fossa in some colubroids (Fig. 4 F), formed by the prearticular rather
than the surangular lamina of the compound, fails
the topological test of homology and is not considered equivalent to the coronoid process in more
basal snakes. Note also that state 2 in T86 (coronoid
process formed by compound only, coronoid absent)
is redundant with T84 (coronoid absent) ; our
recoding avoids such problems.
165. Surangular eminence. 0, compound postdentary element without dorsal eminence (Fig. 4 A).
1, surangular portion of compound with dorsal crest
or process lateral to adductor fossa (Fig. 4 B–E).
2, prearticular portion of compound with ascending
process medial to adductor fossa (Fig. 4 F). Unordered.
166. Adductor fossa. 0, posterior region exposed
medially, prearticular dorsal margin lower than
surangular dorsal margin (Fig. 4 A). 1, posterior
region exposed dorsally only ; prearticular about
equal in height to surangular (Fig. 4 C, D). 2, posterior region exposed laterally only, prearticular
higher than surangular (Fig. 4 F). Ordered 0–1–2.
The orientation of the anterior half may be correlated with the presence or absence of the coronoid
process (which extends back to form a lateral wall to
the adductor fossa), hence restriction of this character to the posterior part of the fossa. T88 [‘ medial
margin of Meckel’s (adductor) fossa forms a distinct,
dorsally projecting crest\low and smooth ’] is very
similar and thus not employed here.
167. Anterior surangular foramen. 0, situated
posteriorly, below apex of coronoid process or more
posterior (Fig. 4 A). 1, situated anteriorly, between
apex and anterior limit of coronoid process (Fig.
4 D). 2, situated far anteriorly, in front of anterior
limit of coronoid process (Fig. 4 B, C). Ordered
0–1–2. Not applicable in taxa lacking a well-formed
coronoid process.
168. Lateral crest of compound element, extending
anteriorly from articular cotyle along ventrolateral
surface of mandible. 0, absent (Fig. 4 A, B).
1, present (Fig. 4 E). The crest represents the attachment of superficial adductor and\or pterygoideus muscles (e.g. Frazzetta, 1966).
169. Articular-surangular fusion. 0, articular and
surangular not fully fused in region of articular facet.
371
1, articular and surangular fully fused in region of
articular facet. Since various portions of the prearticular-surangular suture appear to vary independently, it can be problematic to describe the
elements simply as ‘ separate ’ or ‘ fused ’ (cf. characters 66, 67 of Rieppel, 1980).
170. Retroarticular process length. 0, long, longer
than articular facet (Fig. 4 A, F). 1, short, not longer
than articular facet (Fig. 4 B, D). This and the next
character represent a subdivision of C32 and T89
(retroarticular process long, short and simple, short
with dorsal flange, absent). Cundall et al. (1993,
character 32) recognized elongation as a synapomorphy of anomalepidids and typhlopids only, but
do not give a specific quantitative criterion ; the
antero-posterior length of the facet provides a
convenient standard of comparison.
171. Dorsal flange of retroarticular process.
0, absent. 1, present. Refers to a distinct dorsal
expansion at the lateral edge of the distal end of the
retroarticular process. A subdivision of C32, see
previous character.
(c) Dentition
172. External grooves and ridges on tooth bases.
0, present, surface of bases of mature tooth crowns
with vertical ridges and grooves. 1, absent, surface
of bases of mature tooth crowns smooth. The occurrence of fluted teeth in snakes generally was
proposed by McDowell & Bogert (1954, p. 58),
questioned by Underwood (1957), denied by Rieppel (1983) and not mentioned by Estes et al. (1988,
character 86). Among snakes, weak grooves and
ridges occur in Pachyrhachis (Lee & Caldwell, 1998)
and strong grooves and ridges in at least some
madtsoiids (Scanlon & Lee, 2002). Teeth are
primitively simple in colubroids, although some
derived forms have reacquired fluted teeth (Vaeth,
Rossman & Shoop, 1985). Haasiophis and Archaeophis have strongly fluted tooth crowns [fluting is
shown as extending to the tooth base in the
interpretive drawings in Tchernov et al. (2000),
although this level of detail is not clear in the
accompanying photographs].
173. Premaxillary teeth. 0, present. 1, absent. C4,
T1. Tchernov et al. (2000) stated that Haasiophis
lacks premaxillary teeth based on X-rays : however,
they coded this taxon as possessing teeth in their data
matrix, and the part of the premaxilla exposed
ventrally (their Fig. 1) appears to show alveoli. For
these reasons, this taxon is treated as unknown.
174. Premaxillary tooth number. 0, three or more
Michael S. Y. Lee and John D. Scanlon
372
C
B
A
sub for
par for
haemal keel
prz pr
F
D
E
sub for
G
prz pr
I
H
sub for
par for
haemal keel
K
J
pz for
L
pz for
Fig. 10. Vertebrae. (A-C) Trunk vertebra of Nanowana godthelpi (Madtsoiidae) in (A) right lateral, (B) anterior, and
(C) ventral view. (D-F) Trunk vertebra of Typhlops lumbricalis (Typhlopidae) in (D) right lateral, (E) anterior, and
(F) ventral view. (G-I) Trunk vertebra of Tropidophis melanurus (Tropidophiinae) in (G) right lateral, (H) anterior,
and (I) ventral view. Posterior view of trunk vertebra of ( J) Nanowana godthelpi (Madtsoiidae), (K) Eunectes murinus
(Boinae), (L) Xenopeltis unicolor. Based on (A-C) QM F19741, (D-F) AMNH73230, (G-I) AMNH 93002, ( J) QM
F19741, (K) BMNH 66n8n14n349, (L) BMNH 1930n5n8n41.
alveoli on each side of the midline. 1, one or two
alveoli on each side. Used in some previous analyses
of booids (Underwood, 1976, character 35 ; Kluge,
1993 a, character 1).
175. Maxillary teeth. 0, nearly uniform in size, at
most only slightly larger in middle of tooth row, with
uniform gradation. 1, distinctly larger near middle
of tooth row, smaller anteriorly and posteriorly.
2, distinctly larger near anterior end of tooth row,
smaller in middle and posteriorly. The first and last
one or two teeth are not considered, since they are
nearly always slightly smaller than the more central
teeth. Unordered, since each state can evolve directly
into any other either by enlargement or diminution
of some teeth or ‘ migration ’ of enlarged teeth (e.g.
Jackson & Fritts, 1997). Another derived state
(teeth distinctly larger near posterior of tooth row,
smaller in middle and anteriorly) is restricted to
some colubroids, hence uninformative for the present
analysis.
Snake phylogeny
176. Maxillary teeth. 0, nine or more alveoli.
1, eight or fewer alveoli. Underwood (1976, character
52) ; Kluge (1993 a, character 12 ; 1993 b, character
13, with five and four states respectively). Separate
character states in meristic traits such tooth counts
are always rather arbitrarily delineated. Here, the
boundaries between character states for such characters have been chosen so that the variability within
most terminal taxa falls within a single character
state, i.e. to minimize polymorphic taxa.
177. Dentary teeth. 0, eight or more alveoli.
1, seven or fewer alveoli. Underwood (1976, character 60) ; Kluge (1993 a, character 64 ; 1993 b,
character 50, each with four states).
178. Alveoli (in middle of maxilla and dentary).
0, not expanded transversely. 1, wider transversely
than anteroposteriorly. This feature (state 1) is
clearly visible in Haasiophis (Tchernov et al., 2000)
and numerous madtsoiid specimens (Scanlon, 1996,
1997) ; apparently absent in Podophis (Rage &
Escuillie! , 2000) and not determinable in Pachyrhachis
(the only alveoli clearly visible occlusally appear to
be expanded, but are near the anterior end of the
right maxilla ; Haas, 1980 a). At least in madtsoiids,
state 1 corresponds to a medial expansion of the
tooth base covered by bone of attachment when fully
ankylosed ; the more conical crown occupies the
lateral part of the alveolus (Scanlon, 1996).
179. Palatine teeth. 0, absent. 1, present. C18.
180. Palatine teeth. 0, nine or more alveoli.
1, eight or fewer alveoli. Palatine tooth number was
also used for booids by Underwood (1976, character
58) and Kluge (1993 a, character 50, three states),
but most of the variation they describe falls within
our state 1.
181. Pterygoid teeth. 0, present. 1, absent. C23,
T50.
182. Pterygoid teeth. 0, twelve or more alveoli.
1, eleven to nine alveoli. 2, eight or fewer alveoli.
Ordered 0–1–2. Pterygoid tooth number has been
used in analyses of booids by Underwood (1976,
character 59) and Kluge (1993 a, character 57 ;
1993 b, character 42).
(d) Hyoid apparatus
183. Median (basihyal) element. 0, present, uniting hyoid cornua. 1, absent. C36, after Underwood
(1976, character 64) and Kluge (1991, character
68). Wallach & Gu$ nther (1998), after Groombridge
(1979 c), used an unordered four-state character
(W20 : diverging and separated\semiparallel and
373
separated\semiparallel and united\parallel and
united), here split into 183 and 185.
184. First branchial arch elements. 0, present.
1, absent, replaced by caudal extensions of the lateral
edge of the basihyal (if present). C37.
185. Hyoid cornua. 0, diverging sharply posteriorly. 1, diverging only slightly posteriorly. 2, parallel. Ordered 0–1–2. W20. Based on Langebartel
(1968 ; see also Underwood, 1976, character 65) and
Groombridge (1979 c). The homology of the divergent processes in lizards and snakes is assumed
here but has been questioned (see Cundall et al.,
1993, character 35).
(e) Vertebrae
186. Number of presacral vertebrae. 0, less than
120. 1, 120–160. 2, 160–200. 3, over 200. Ordered
0–1–2–3. The number and boundaries of states are
arbitrary, but chosen such that all states are occupied
by one or more taxa, and most terminal taxa can be
assigned a single state (cf. character 176).
187. Number of caudal vertebrae. 0, more than
20. 1, fewer than 20.
188. Dorsoposterior process on atlas neural arch,
overlying axis neural arch. 0, present, well developed. 1, absent or very weak.
189. Second (axis) intercentrum. 0, not fused to
anterior region of axis centrum, suturally connected
at most. 1, fused to anterior region of axis centrum.
190. Neural spine height. 0, well-developed process (Fig. 10 A, G). 1, low ridge, or absent (Fig.
10 D).
191. Posterior margin of neural arch. 0, shallowly
concave in dorsal view. 1, with deep, V-shaped
embayment in dorsal view exposing much of centrum in front of condyle. Comparisons are restricted
to mid-trunk elements, in order to avoid intracolumnar variability.
192. Zygosphene buttress. 0, with deeply concave
anterior edge, i.e. deeply notched between zygosphenes. 1, with shallowly concave anterior edge, i.e.
slightly notched between zygosphenes. 2, with
straight or slightly sinuous anterior edge, i.e. not
uniformly concave between zygosphenes. Ordered
0–1–2. Madtsoiidae is coded as polymorphic, but
state 2 is restricted to Alamitophis (Scanlon, 1993,
1997).
193. Condyles of mid-trunk vertebrae. 0, oval,
sagittal dimension much less than transverse diameter (Fig. 10 E). 1, round, sagittal dimension
similar to transverse dimension (Fig. 10 B, H). The
expression ‘ sagittal ’ rather than ‘ vertical ’ dimension
is used to make this character applicable in taxa
374
where the condyles are obliquely oriented (next
character) : otherwise, in such forms, a reduction in
‘ vertical ’ dimension of the condyles would be scored
even when the condyles remain round. Comparisons
of this and the next character are restricted to midtrunk elements, in order to avoid intracolumnar
variability which remains inadequately described for
all taxa.
194. Condyles of mid-trunk vertebrae. 0, facing
very dorsally, ventral edge (at most) of condyle
surface exposed in ventral view (Fig. 10 F). 1, facing
posteriorly, or posterodorsally, much of condyle
surface exposed in ventral view (Fig. 10 C, I).
195. Precondylar constriction of centrum.
0, absent or very weak (Fig. 10 F). 1, moderate (Fig.
10 C, I). 2, strong. Ordered 0–1–2.
196. Orientation of zygapophyses of mid-trunk
vertebrae. 0, steeply inclined medially, 30m or more
from the horizontal. 1, moderately inclined medially,
between 15m and 30m from the horizontal. 2, not
inclined medially, less than 15m from horizontal.
Ordered 0–1–2. This character attempts quantitatively to encode variation previously recognized
(e.g. Hoffstetter & Gasc, 1969 ; Rage, 1984). Comparisons are restricted to mid-trunk elements, in order
to avoid intracolumnar variability.
197. Paracotylar foramina (foramen on anterior
surface between cotyle and transverse process).
0, present on most or all vertebrae (Fig. 10 B, H).
1, present on no, or few, vertebrae (Fig. 10 E). Modified from Underwood (1976, character 77) and
Kluge (1991, character 69).
198. Parazygantral foramina (foramen on posterior surface of neural arch, between zygantrum
and postzygapophyseal facets). 0, absent on all
vertebrae (Fig. 10 L). 1, numerous small pits in
parazygantral area (Fig. 10 K). 2, large, single
foramen present on each side (Fig. 10 J). Unordered.
Discussed by Simpson (1933), Hoffstetter (1961),
Hoffstetter & Gasc (1969), Smith (1976), Holman &
Case (1992).
199. Subcentral foramina. 0, uniform throughout
column, small and paired in most vertebrae (Fig.
10 C, I). 1, irregular, being either small and paired,
absent, or single and large, in different vertebrae
(Fig. 10 F). See Mahendra (1935) for discussion.
200. Prezygapophyseal process. 0, absent (Fig.
10 C). 1, present as a small process extending
laterally from prezygapophyseal facet (Fig. 10 I).
2, present as a prominent process extending laterally
from prezygapophyseal facet (Fig. 10 F). Ordered
0–1–2. Based on W18.
Michael S. Y. Lee and John D. Scanlon
201. Hypapophyses. 0, present on anterior eight
cervicals or fewer. 1, present up to at least cervical
ten, but absent in mid- and posterior trunk.
2, present throughout trunk, but poorly developed in
posterior trunk. 3, present throughout trunk, well
developed throughout. Ordered 0–1–2–3. Based on
W17, and discussions by Underwood (1967), Bogert
(1968) and Hoffstetter (1968) for snakes, and
Hoffstetter & Gasc (1968) for varanoids.
202. Ventral surface of centra. 0, mid-trunk
vertebrae with smooth, transversely convex ventral
surface (Fig. 10 F). 1, mid-trunk vertebrae bearing
single median haemal keels (Fig. 10 C, I). This keel
may merge posteriorly with the hypapophysis, if
present.
203. Lymphapophyses. 0, fewer than three forked
free ribs or lymphapophyses. 1, three or more freeending cloacal vertebrae with lymphapophyses.
Discussed by Gasc (1966) and Hoffstetter & Gasc
(1969). In Pachyrhachis there appears to be at least
one forked rib, and no more than two, but it is not
clear whether they are articulated or fixed.
204. Caudal vertebrae. 0, with posteroventral
projections. 1, without posteroventral projections.
Discussed by Hoffstetter & Gasc (1969) ; the nature
of the projections is covered in the following two
characters.
205. Posteroventral elements of caudals. 0, articulate with centrum. 1, fuse with centrum. Discussed for marine varanoids by Nopcsa (1903). Until
recently state 0 had not been reported in snakes, but
articulated chevrons are now known in Podophis
(Rage & Escuillie! , 2000) ; in madtsoiids they
generally separate from the short pedicels and have
not been identified as isolated elements (Scanlon,
1996), but one caudal vertebra has been found with
a partially fused chevron in Wonambi (Scanlon &
Lee, 2000).
206. Posteroventral elements of caudals. 0, distally
fused (chevrons). 1, distally separated (haemapophyses). 2, single median element (caudal hypapophyses). Unordered. These states are discussed by
Hoffstetter & Gasc (1969).
207. Ribs. 0, tuber costae (l tuberculum, tuberculiform process, dorsal process) of rib absent or
weakly developed. 1, tuber costae well developed.
The plesiomorphic state of this character was given
considerable weight by Nopcsa (1923) as implying
pachyophiids and palaeophiids were basal to other
snakes, whereas Hoffstetter (1955 ; Hoffstetter &
Gasc, 1969) considered reduction of the process to be
secondary in these aquatic forms.
Snake phylogeny
208. Ribs. 0, slender throughout body. 1, thickened and heavily ossified (pachyostotic) in middle
region of body. The primitive state of this character
is coded as either 0 or 1 because, although most
outgroups have state 0, some mosasauroid-like forms
which might be the nearest relatives of snakes have
state 1 (Lee & Caldwell, 2000).
209. Cervical region. 0, present, at least 20 anterior
vertebrae bearing ribs distinctly shorter and thinner
than ribs of other trunk vertebrae. 1, absent, only the
first few anterior vertebrae bear short, thin ribs.
( f ) Pelvis and hindlimb
210. Pelvic girdle. 0, three elements present. 1, two
elements present. 2, single element present. 3, no
elements present. More finely subdivided than C38
(pelvic vestiges present\absent) ; ordered 0–1–2–3.
211. Pelvis. 0, external to sacral or cloacal ribs.
1, internal to sacral or cloacal ribs.
212. Hindlimb. 0, hindlimb present, with distinct
femur, tibia and fibula. 1, hindlimb vestigial, with
one bone (femur) only, sometimes with a single distal
spur. 2, hindlimb absent. Ordered 0–1–2. Variation
has been summarized by Essex (1927), Groombridge
(1979 c, p. 465) and List (1966).
(2) Soft anatomy
(a) Musculature
213. Origin of adductor mandibulae externus, pars
superficialis. 0, from braincase only. 1, from both
braincase and temporal tendon. 2, from temporal
tendon only. Ordered 0–1–2. C39.
214. Bony insertion of adductor mandibulae externus,
pars superficialis. 0, on dorsal edge of lower jaw
(coronoid process or dorsal edge of compound bone).
1, on ventrolateral surface of lower jaw (external
surface of surangular portion of compound bone).
This and the following represent a subdivision of
C40.
215. Muscular insertion of adductor mandibulae
externus, pars superficialis. 0, on adductor externus medialis.
1, on adductor externus profundus. C40 (part).
216. Adductor mandibulae externus, pars levator anguli
oris. 0, present. 1, absent. C41. Cundall et al. (1993)
subdivided ‘ presence ’ into two conditions, but this
subdivision was cladistically uninformative in both
their analysis and this one because one of these
conditions was an autapomorphy of Xenopeltis.
217. Insertion tendon of adductor mandibulae externus,
pars medialis. 0, prominent. 1, reduced or absent.
C42.
375
218. Quadrate tendon of adductor mandibulae externus, pars profundus. 0, prominent. 1, reduced or
absent. C43.
219. Origin of m. pterygoideus (adductor mandibulae
internus). 0, single. 1, double. C44 (see also Cundall
& Greene, 2000).
220. Retractor vomeris. 0, absent. 1, present. C45.
221. Origin of retractor pterygoidei. 0, posterolateral
to vidian nerve. 1, anteromedial to vidian nerve.
C46.
222. Pterygoid attachment to basicranium. 0, via
strong ligaments. 1, via cid-muscles. T83. The fossil
taxon Dinilysia is coded with state 0 by Tchernov et
al. (2000), but it is here coded as unknown because
it is unclear how the condition can be inferred based
on purely osteological data.
223. Intermandibularis anterior. 0, undivided, only
pars anterior present. 1, divided, pars anterior and pars
posterior present. This and the following character
represent a recoding of C47 into two characters.
224. Intermandibularis anterior, pars anterior.
0, covered entirely by genioglossus. 1, exposed ventrally medial to genioglossus.
225. Intermandibularis anterior, pars anterior. 0, inserts
on oral mucosa and tongue sheath. 1, inserts on
interramal pad. C48. Cundall et al. (1993) state that
this character is inapplicable in taxa with an
undivided intermandibularis anterior. However, Groombridge (1979 a) showed that the undivided intermandibularis anterior corresponds to the intermandibularis
anterior pars anterior in those taxa with a divided
muscle, and thus, such taxa can be scored.
226. Intermandibularis posterior, pars posterior.
0, present. 1, absent. C49.
227. Omohyoideus. 0, present. 1, absent. C50.
(b) Visceral organs
228. Trachea. 0, right of heart midline. 1, dorsal to
heart. 2, left of heart midline. Cannot be coded in
the outgroups, which have short, paired tracheae
that do not reach the level of the heart. Unordered,
because states 0 and 2 could conceivably evolve
directly from each other via a change in symmetry so
that asymmetry might be homologous. W34.
229. Free tips on tracheal rings. 0, absent. 1,
present. W32. Amphisbaenians and dibamids exhibit
state 0 (V. Wallach, personal communication).
230. Tracheal lung. 0, absent. 1, present. W33.
Amphisbaenians and dibamids exhibit state 0 (V.
Wallach, personal communication).
231. Tracheal entry into right lung. 0, subterminal, near cranial end. 1, terminal at cranial end.
C51.
376
Michael S. Y. Lee and John D. Scanlon
Table 2. Characters diagnosing each branch in Fig. 11A
(Optimisation was via delayed transformation [as in most recent studies : e.g. Gauthier, Kluge and Rowe (1988),
Bryant and Russell (1992), Lee (1998)] ; however, unambiguous changes (i.e. those which occur under both delayed
and accelerated transformation) are indicated in bold type. The consistency index is listed in parentheses after each
synapomorphy. All changes are from 0 to 1 unless otherwise indicated.)
Clades
Snakes : 12 (1n0), 23 (0n333 : 0 2), 52 (0n429), 54 (0n4 : 0 2), 87 (0n4), 94 (0n5), 102 (0n286), 149
(0n5 : 0\1 2), 150 (0n2), 170 (0n333), 179 (0n333), 196 (1n0), 201 (0n6).
Haasiophis, Madtsoiidae, Dinilysia, and modern snakes : 41 (0n5), 59 (0n333), 104 (0n5), 109
(1n0), 148 (0n333), 162 (0n667), 163 (0n2), 178 (0n5), 186 (0n5), 202 (0n333), 203 (1n0).
Madtsoiidae, Dinilysia, and modern snakes : 19 (0n333), 64 (0n2), 96 (1n0), 116 (0n333), 139 (0n25),
141 (1n0), 166 (0n222), 182 (0n333), 192 (1n0), 193 (0n5), 207 (0n5), 208 (1n0 : 1 0).
Dinilysia and modern snakes : 18 (0n286), 29 (0n5 : 3 1), 33 (1n0), 45 (0n2), 101 (0n5), 136 (0n667),
141 (1n0 : 1 2), 144 (0n333), 158 (0n25), 169 (1n0), 172 (1n0), 180 (0n333), 200 (0n5).
Modern snakes : 5 (0n333), 24 (0n25), 41 (0n5 : 1 0), 46 (0n5), 48 (0n333), 49 (0n333), 51 (0n667 :
0 2), 66 (0n4), 75 (1n0), 109 (1n0 : 1 2), 117 (0n333), 131 (1n0), 135 (0n667), 137 (0n333), 186 (0n5 :
1 3), 190 (0n5), 192 (1n0 : 1 2), 197 (0n5), 209 (1n0), 211 (1n0), 212 (0n4), 227 (0n333), 242 (0n4),
249 (0n4), 253 (0n6), 256 (0n5).
Scolecophidia : 3 (0n5 : 1 0), 4 (1n0), 8 (1n0 : 0 2), 9 (1n0), 10 (1n0), 23 (0n333 : 2 1), 25 (0n25), 28
(0n667), 54 (0n4 : 2 0), 60 (0n5 : 0 2), 61 (0n667 : 1 2), 62 (0n5 : 1 0), 69 (0n667), 75 (1n0 : 1 2),
77 (0n5), 78 (0n75), 83 (1n0), 94 (0n5 : 1 0), 98 (0n2), 101 (0n5 : 1 2), 103 (1n0), 104 (0n5 : 1 2), 107
(1n0), 117 (0n333 : 1 2), 128 (1n0), 135 (0n667 : 1 2), 137 (0n333 : 1 2), 145 (0n333), 149 (0n5 :
2 0), 156 (0n5), 157 (0n667 : 1 2), 167 (0n333 : 1 0), 173 (0n333), 177 (0n5), 179 (0n333 : 1 0),
181 (0n5), 188 (0n333), 193 (0n5 : 1 0), 194 (1n0 : 1 0), 199 (1n0), 200 (0n5 : 1 2), 201 (0n6 : 1 0),
202 (0n333 : 1 0), 204 (0n5), 218 (0n333), 231 (0n333), 232 (0n286 : 0 2), 251 (0n25), 253 (0n6 :
1 2), 262 (0n5 : 1 0).
Typhlopidae and Anomalepididae : 12 (1n0 : 1 2), 13 (1n0), 16 (0n333), 17 (1n0), 19 (0n333 : 1 0),
34 (0n5 : 0 2), 35 (0n5), 55 (1n0 : 1 0), 80 (1n0 : 0 2), 81 (1n0), 87 (0n4 : 1 0), 92 (0n333), 147
(1n0), 148 (0n333 : 1 0), 159 (1n0), 170 (0n333 : 1 0), 176 (0n5), 187 (0n5), 210 (0n231 : 0 2), 212
(0n4 : 1 2), 230 (0n333).
Alethinophidia : 42 (0n5), 43 (0n333), 54 (0n4 : 2 1), 56 (1n0), 73 (1n0), 78 (0n75 : 0 2), 80 (1n0), 85
(1n0), 89 (0n5), 99 (0n5), 119 (1n0), 124 (1n0), 130 (1n0), 134 (0n333), 154 (0n333 : 1 0), 165 (0n667),
178 (0n5 : 1 0), 184 (1n0), 205 (1n0), 206 (1n0), 215 (1n0), 226 (0n5), 235 (1n0), 242 (0n4 : 1 2), 244
(0n5), 255 (0n5), 260 (0n667), 263 (1n0).
Anilioidea : 5 (0n333 : 1 2), 50 (0n333), 62 (0n5 : 1 0), 63 (0n5), 70 (0n4 : 0 2), 91 (0n5), 101 (0n5 :
1 0), 102 (0n286 : 1 0), 118 (0n5), 121 (0n4), 122 (0n5), 127 (1n0), 129 (0n333), 145 (0n333), 150
(0n2 : 1 0), 187 (0n5), 202 (0n333 : 1 0), 213 (0n667), 221 (0n333), 232 (0n286), 246 (0n333 : 1 2),
257 (0n25), 261 (0n2).
Cylindrophis, Anomochilus, and Uropeltidae : 7 (1n0), 115 (1n0), 142 (0n5), 143 (1n0), 173 (0n333),
188 (0n333), 227 (0n333 : 1 0), 245 (1n0 : 1 2), 249 (0n4 : 1 2).
Anomochilus and Uropeltidae : 8 (1n0), 47 (0n333 : 0 2), 69 (0n667), 94 (0n5 : 1 2), 152 (0n5 :
1 0), 158 (0n25 : 1 2), 166 (0n222 : 1 0), 171 (1n0), 176 (0n5), 181 (0n5), 212 (0n4 : 1 2), 251
(0n25).
Macrostomata : 14 (1n0), 35 (0n5), 39 (0n5), 40 (1n0), 68 (1n0), 79 (1n0), 86 (1n0), 90 (0n333), 92
(0n333), 106 (1n0), 111 (0n333), 113 (0n5), 116 (0n333 : 1 0), 120 (0n5), 138 (0n5), 139 (0n25 : 1 0),
146 (1n0), 151 (1n0), 152 (0n5 : 1 0), 161 (1n0), 162 (0n667 : 1 2), 191 (0n5), 195 (0n4), 214 (1n0), 224
(1n0), 233 (0n667 : 1 0), 245 (1n0 : 1 0), 255 (0n5 : 1 2), 259 (0n5).
Xenopeltidae : 11 (0n25), 18 (0n286 : 1 0), 22 (0n5), 25 (0n25), 31 (1n0 : 0 2), 32 (1n0), 36 (1n0), 38
(1n0), 50 (0n333), 52 (0n429 : 1 2), 60 (0n5), 63 (0n5), 88 (0n333), 90 (0n333 : 1 2), 137 (0n333 :
1 2), 156 (0n5), 164 (0n333), 167 (0n333 : 1 2), 189 (0n5), 200 (0n5 : 1 2), 210 (0n231), 221 (0n333),
228 (0n5), 249 (0n4 : 1 2), 250 (0n2), 256 (0n5 : 1 2), 258 (0n5).
‘ Core macrostomatans ’ : 15 (1n0), 34 (0n5), 37 (1n0), 49 (0n333 : 1 0), 51 (0n667 : 2 1), 61 (0n667 :
1 0), 72 (1n0), 74 (0n25), 114 (0n25), 121 (0n4 : 0 2), 133 (1n0), 136 (0n667 : 1 2), 144 (0n333 :
1 0), 168 (0n5), 173 (0n333), 190 (0n5 : 1 0), 196 (1n0 : 1 2), 216 (0n5), 217 (0n5), 218 (0n333), 220
(0n333), 223 (0n5), 240 (1n0 : 1 0), 243 (0n4 : 1 0), 248 (1n0), 257 (0n25), 261 (0n2).
Snake phylogeny
Table 2. (cont.)
Clades
Booidea : 6 (0n5), 20 (1n0), 30 (1n0), 71 (0n5), 95 (1n0), 97 (0n5), 109 (1n0 : 2 3), 125 (0n5), 146 (1n0 :
1 2), 150 (0n2 : 1 0), 195 (0n4 : 1 2), 198 (1n0), 228 (0n5), 234 (0n333), 246 (0n333 : 1 0).
Pythoninae and Boinae : 26 (1n0), 42 (0n5 : 1 2), 76 (0n333), 105 (0n5), 112 (0n333), 117 (0n333 :
1 0), 126 (0n5), 175 (0n667 : 0 2), 201 (0n6 : 1 2).
‘ Advanced snakes ’ : 22 (0n5), 24 (0n25 : 1 0), 70 (0n4), 117 (0n333 : 1 2), 160 (0n5), 185 (0n667),
226 (0n5 : 1 0), 230 (0n333), 231 (0n333), 232 (0n286 : 0 2), 236 (0n333), 237 (0n333), 247 (0n5), 252
(0n5), 258 (0n5 : 0 2).
Tropidophiinae, Bolyeriidae and Caenophidia : 27 (1n0), 53 (1n0), 66 (0n4 : 1 2), 92 (0n333 :
1 0), 99 (0n5 : 1 0), 201 (0n6 : 1 3), 244 (0n5 : 1 0), 254 (1n0).
Bolyeriidae and Caenophidia : 51 (0n667 : 1 2), 57 (0n333), 90 (0n333 : 1 2), 100 (0n25), 102
(0n286 : 1 2), 104 (0n5 : 1 2), 112 (0n333), 137 (0n333 : 1 2), 180 (0n333 : 1 0), 197 (0n5 : 1 0),
210 (0n231 : 0 3), 212 (0n4 : 1 2).
Caenophidia : 21 (1n0), 33 (1n0 : 1 2), 44 (0n5), 72 (1n0 : 1 2), 76 (0n333), 78 (0n75 : 2 3), 82 (1n0),
84 (1n0), 85 (1n0 : 1 2), 89 (0n5 : 1 0), 110 (1n0), 132 (0n5), 140 (0n5), 160 (0n5 : 1 2), 185 (0n667 :
1 2), 219 (1n0), 222 (1n0), 225 (1n0), 228 (0n5), 236 (0n333 : 1 0), 249 (0n4 : 1 2).
Terminal branches
Pachyrhachis : 1 (0n667), 49 (0n333 : 0 2), 64 (0n2), 65 (0n2), 74 (0n25), 76 (0n333), 78 (0n75 :
0 2\3), 102 (0n286 : 1 2), 166 (0n222).
Haasiophis : 28 (0n667 : 0 1\2), 29 (0.5 : 3 2), 70 (0n4), 74 (0n25), 100 (0n25), 105 (0n5), 120 (0n5),
137 (0n333), 138 (0n5), 168 (0n5).
Madtsoiidae : 28 (0n667 : 0 2), 66 (0n4), 67 (0n5), 93 (0n5), 98 (0n2), 111 (0n333), 113 (0n5), 118 (0n5),
120 (0n5), 129 (0n333), 134 (0n333), 142 (0n5), 191 (0n5), 195 (0n4), 198 (1n0 : 0 2).
Dinilysia : 39 (0n5), 47 (0n333), 65 (0n2), 67 (0n5), 93 (0n5), 94 (0n5 : 1 0), 120 (0n5 : 0 2), 123 (0n25 :
1 0), 138 (0n5), 150 (0n2 : 1 0), 165 (0n667 : 0 2), 166 (0n222 : 1 2), 182 (0n333 : 1 2), 195
(0n4).
Leptotyphlopidae : 47 (0n333 : 0 2), 50 (0n333), 66 (0n4 : 1 2), 108 (0n4 : 0 2), 121 (0n4 : 0 2),
158 (0n25 : 1 2), 163 (0n2 : 1 0), 166 (0n222 : 1 0), 233 (0n667 : 1 2), 238 (0n5 : 1 2), 257
(0n25).
Typhlopidae : 11 (0n25), 29 (0n5 : 1 0), 47 (0n333 : 0 2), 52 (0n429 : 1 3), 59 (0n333 : 1 0), 69
(0n667 : 1 2), 108 (0n4 : 0 2), 152 (0n5 : 1 2), 155 (0n5), 158 (0n25 : 1 0), 166 (0n222 : 1 0), 216
(0n5), 250 (0n2).
Anomalepididae : 2 (0n2 : 1 2), 6 (0n5 : 0 1\2), 29 (0n5 : 1 2), 43 (0n333), 44 (0n5), 52 (0n429 :
1 0), 66 (0n4 : 1 2), 90 (0n333), 96 (1n0 : 1 2), 112 (0n333), 152 (0n5 : 1 0), 153 (0n5), 183 (0n5),
217 (0n5), 227 (0n333 : 1 0), 229 (0n333), 246 (0n333 : 1 2), 253 (0n6 : 2 3).
Anilius : 11 (0n25), 18 (0n286 : 1 0), 45 (0n2 : 1 0), 47 (0n333 : 0 2), 52 (0n429 : 1 2), 58 (0n333),
59 (0n333 : 1 2), 60 (0n5), 64 (0n2 : 1 0), 88 (0n333), 98 (0n2), 100 (0n25), 114 (0n25), 116 (0n333 :
1 0), 139 (0n25 : 1 0), 149 (0n5 : 2 1), 153 (0n5), 162 (0n667 : 1 2), 167 (0n333 : 1 0), 175
(0n667), 210 (0n231 : 0 2), 220 (0n333), 238 (0n5 : 1 2), 250 (0n2), 252 (0n5 : 0 2), 253 (0n6 : 1 2),
258 (0n5), 259 (0n5).
Cylindrophis : 25 (0n25), 52 (0n429 : 1 2), 59 (0n333 : 1 2), 60 (0n5), 65 (0n2), 87 (0n4 : 1 2), 100
(0n25), 111 (0n333), 167 (0n333 : 1 2), 204 (0n5), 241 (0n333 : 2 1), 250 (0n2).
Anomochilus : 5 (0n333 : 2 1), 23 (0n333 : 2 0), 24 (0n25 : 1 0), 25 (0n25), 43 (0n333 : 1 0), 77
(0n5), 80(1n0 : 1 3), 104 (0n5 : 1 2), 108 (0n4), 114 (0n25), 117 (0n333 : 1 2), 122 (0n5 : 1 0), 123
(0n25 : 1 0), 137 (0n333 : 1 2), 170 (0n333 : 1 0), 177 (0n5), 179 (0n333 : 1 0), 213 (0n667 : 1 0),
231 (0n333), 232 (0n286 : 1 2), 233 (0n667 : 1 2), 236 (0n333), 237 (0n333), 238 (0n5 : 1 2), 239
(1n0), 240 (1n0 : 1 2), 243 (0n4 : 1 2), 247 (0n5), 261 (0n2 : 1 0).
Uropeltidae : 11 (0n25 : 0 2), 18 (0n286 : 1 0), 42 (0n5 : 1 0), 58 (0n333), 87 (0n4 : 1 2), 134
(0n333 : 1 2), 143 (1n0 : 1 2), 150 (0n2), 157 (0n667 : 1 0), 163 (0n2 : 1 0), 186 (0n5 : 3 2), 189
(0n5), 206 (1n0 : 1 2), 210 (0n231 : 0 3), 213 (0n667 : 1 2), 241 (0n333 : 2 1), 242 (0n4 : 2 1),
246 (0n333 : 2 1), 253 (0n6 : 1 2), 256 (0n5 : 1 2), 262 (0n5 : 1 0).
377
378
Michael S. Y. Lee and John D. Scanlon
Table 2. (cont.)
Terminal branches
Xenopeltis : 19 (0n333 : 1 0), 42 (0n5 : 1 2), 47 (0n333 : 0 2), 70 (0n4), 123 (0n25 : 1 0), 134
(0n333 : 1 0), 145 (0n333), 155 (0n5), 158 (0n25 : 1 2), 163 (0n2 : 1 0), 174 (1n0 : 1 0), 180 (0n333 :
1 0), 182 (0n333 : 1 0), 186 (0n5 : 3 2), 210 (0n231 : 1 3), 212 (0n4 : 1 2), 251 (0n25).
Loxocemus : 3 (0n5 : 1 0), 11 (0n25 : 1 2), 45 (0n2 : 1 0), 49 (0n333 : 1 2), 58 (0n333), 61 (0n667 :
1 0), 64 (0n2 : 1 0), 74 (0n25), 101 (0n5 : 1 0), 102 (0n286 : 1 2), 121 (0n4 : 0 2), 125 (0n5), 136
(0n667 : 1 2), 144 (0n333 : 1 0), 154 (0n333), 166 (0n222 : 1 0), 182 (0n333 : 1 2), 243 (0n4 :
1 0), 246 (0n333 : 1 0).
Erycinae : 2 (0n2 : 1 2), 6 (0n5 : 1 2), 11 (0n25), 18 (0n286 : 1 2), 57 (0n333), 59 (0n333 : 1 0), 98
(0n2), 163 (0n2 : 1 0), 182 (0n333 : 1 2), 260 (0n667 : 1 2).
Pythoninae : 2 (0n2 : 1 2), 5 (0n333 : 1 2), 45 (0n2 : 1 0), 48 (0n333 : 1 0), 90 (0n333 : 1 2), 139
(0n25), 157 (0n667 : 1 0), 158 (0n25 : 1 0), 218 (0n333 : 1 0), 223 (0n5 : 1 0), 249 (0n4 : 1 2),
250 (0n2), 261 (0n2 : 1 0).
Boinae : 24 (0n25 : 1 0), 29 (0n5 : 1 2), 54 (0n4 : 1 0\2), 65 (0n2), 98 (0n2), 102 (0n286 : 1 2), 138
(0n5 : 1 2), 243 (0n4).
Ungaliophiinae : 1 (0n667), 11 (0n25), 57 (0n333), 91 (0n5), 97 (0n5), 102 (0n286 : 1 0), 158 (0n25 :
1 0), 160 (0n5 : 1 2), 229 (0n333).
Tropidophiinae : 16 (0n333), 23 (0n333 : 2 1), 29 (0n5 : 1 0), 48 (0n333 : 1 0), 59 (0n333 : 1 0),
88 (0n333), 90 (0n333 : 1 0), 132 (0n5), 134 (0n333 : 1 2), 140 (0n5), 154 (0n333), 158 (0n25 : 1 2),
166 (0n222 : 1 2), 167 (0n333 : 1 2), 185 (0n667 : 1 2), 186 (0n5 : 3 2), 229 (0n333), 234 (0n333),
252 (0n5 : 1 2).
Bolyeriidae : 2 (0n2 : 1 2), 5 (0n333 : 1 2), 18 (0n286 : 1 0), 22 (0n5 : 1 2), 34 (0n5 : 1 0), 46
(0n5 : 1 0), 64 (0n2 : 1 0), 65 (0n2), 121 (0n4 : 2 0), 123 (0n25 : 1 0), 129 (0n333), 160 (0n5 :
1 0), 164 (0n333), 175 (0n667 : 0 2), 183 (0n5), 221 (0n333), 228 (0n5 : 0 2), 230 (0n333 : 1 0),
232 (0n286 : 2 1), 242 (0n4 : 2 1), 243 (0n4), 261 (0n2 : 1 0).
Acrochordidae : 2 (0n2 : 1 2), 5 (0n333 : 1 0), 6 (0n5 : 0 1\2), 16 (0n333), 31 (1n0), 45 (0n2 : 1 0),
49 (0n333), 52 (0n429 : 1 2), 54 (0n4 : 1 0), 71 (0n5), 87 (0n4 : 1 2), 117 (0n333 : 2 1), 126 (0n5),
134 (0n333 : 1 0), 135 (0n667 : 1 0), 148 (0n333 : 1 0), 166 (0n222 : 1 2), 167 (0n333 : 1 2), 188
(0n333), 195 (0n4 : 1 0), 200 (0n5 : 1 2), 207 (0n5 : 1 0), 220 (0n333 : 1 0), 237 (0n333 : 1 0),
238 (0n5 : 1 0), 241 (0n333 : 2 1), 242 (0n4 : 2 1), 246 (0n333 : 1 0), 255 (0n5 : 2 0), 256 (0n5 :
1 0), 257 (0n25 : 1 0), 260 (0n667 : 1 2).
Colubroidea : 18 (0n286 : 1 0), 22 (0n5 : 1 2), 114 (0n25 : 1 0), 164 (0n333), 165 (0n667 : 1 2),
182 (0n333 : 1 0), 234 (0n333), 251 (0n25).
232. Left lung, bronchus and orifice. 0, present,
large 1, present, small. 2, absent. Ordered 0–1–2.
Modified from C52, with intermediate state based
on Underwood (1967 ; 1976, character 20) and
McDowell (1975).
233. Right lung-snout distance. 0, 44 % of
snout-vent length. 1, 44–38 % of snout-vent length.
2,
38 % of snout-vent length. Ordered 0–1–2.
C59 ; in this and the following visceral characters,
boundaries of quantitative states are those used by
Cundall et al. (1993).
234. Intercostal arteries. 0, spanning single segment. 1, spanning multiple segments. Wallach &
Gu$ nther (1998, Table 2, character ‘ ICA ’).
235. Thymus glands. 0, one pair. 1, two pairs.
Wallach & Gu$ nther (1998, Table 2, character ‘TG’).
236. Vena cava position. 0, midventral. 1, lateral
(right). Wallach & Gu$ nther (1998, Table 2, character ‘ VC ’). Amphisbaenians and dibamids exhibit
state 0 (V. Wallach, personal communication).
237. Limb of pancreas extending to spleen.
0, present. 1, absent. C54.
238. Heart-snout distance. 0, 35 % snout-vent
length. 1, 35–26 % snout-vent length. 2,
26 %
snout-vent length. Ordered 0–1–2. C58.
239. Posterior extension of liver. 0, absent.
1, present. C53.
240. Liver length. 0,
24 % snout-vent length.
1, 24–32 % snout-vent length. 2, 32 % snout-vent
length. Ordered 0–1–2. C60.
241. Liver-snout distance. 0, 60 % snout-vent
length. 1, 60–50 % snout-vent length. 2,
50 %
snout-vent length. Ordered 0–1–2. C61.
242. Liver-gall bladder interval. 0, k2 % snout-
Snake phylogeny
vent length. 1, k2 % to j7 % snout-vent length.
2, 7 % snout-vent length. Ordered 0–1–2. C62.
243. Gall bladder-snout length. 0, 64 % snoutvent length. 1, 64-75 % snout-vent length. 2, 75 %
snout-vent length. Ordered 0–1–2. C63.
244. Kidneys. 0, nonlobed. 1, lobed. W43.
Amphisbaenians and dibamids exhibit state 0 (V.
Wallach, personal communication).
245. Right kidney-snout length. 0, 86 % snoutvent length. 1, 86–90 % snout-vent length. 2, 90 %
snout-vent length. Ordered 0–1–2. C64.
246. Left kidney-vent distance. 0, 10 % snoutvent length. 1, 10–6 % snout-vent length. 2, 6 %
snout-vent length. Ordered 0–1–2. C65.
247. Ileocolic (rectal) caecum. 0, present. 1, absent. C56. A caecum is present in some colubroids
(anonymous referee, personal communication).
(c) Palate and nasal region
248. Nasal gland and aditus conchae. 0, not roofed
by prefrontal. 1, roofed by prefrontal. T19, after
McDowell (1975). This character can vary independently of presence or absence of an anterior
flange of the prefrontal (character 31) and is thus
treated as a distinct character. McDowell (1975)
regarded state 1 as diagnostic of Booidea (Pythoninae, Boinae, Erycinae only) ; Tchernov et al. (2000)
cited McDowell but attributed state 1 to numerous
other groups of advanced snakes, so either a criterion
distinct from that of McDowell was actually used, or
some codings in these studies were incorrect. We
nevertheless provisionally adopt the codings in the
later study pending a more detailed discussion of
character states from these authors.
249. Ectochoanal cartilages. 0, left and right
ectochoanal cartilages well separated, no choanal
arc of superficial palate. 1, cartilages approach
midline anteriorly, choanal arc present but deeply
emarginate posteriorly. 2, cartilages with long
median approximation, choanal arc present and
uninterrupted. Ordered 0–1–2. This and the next
two characters (Scanlon, 1996, characters 255–257)
are based on McDowell (1972) and Groombridge
(1979 c). ‘ States ’ A to H of Groombridge (some of
which were used by Wallach & Gu$ nther, 1998,
W13) combine several independently varying characters (see also characters 4, 9, 13, 14 and 18 in
cladogram, Groombridge, 1979 c, Fig. 10), which we
separate here.
250. Lateral folds of superficial palate. 0, no
distinct lateral folds. 1, distinct lateral folds present.
In alethinophidians the lateral fold lies immediately
379
medial to the dentigerous bar of the palatine. In
scolecophidians (and Lanthanotus), the latter landmark is absent but the fold appears to lie in an
otherwise similar position.
251. Vomerine flaps. 0, present, ventral projection
of the vomerine superficial palate. 1, absent, no
ventral projections of the vomerine superficial palate.
State 0 apparently characterises all extant varanoids
(McDowell, 1972, Figs 6, 8, 12). The distinction
drawn by Groombridge (1979 c) between vomerine
flaps and palatal lobes does not seem to be clear-cut,
and they are here considered the same character.
252. Vomerine flaps. 0, without posterolateral
processes, no posterior bifurcation. 1, with short
posterolateral processes forming shallow posterior
bifurcation. 2, with long posterolateral processes
forming extensive posterior bifurcation. Ordered
0–1–2, and inapplicable in taxa without vomerine
flaps.
(d ) External morpholog y
253. Eyes. 0, exposed, moveable eyelids present.
1, covered by separate scale, a transparent spectacle
(brille). 2, covered by a normal cranial scale with
transparent ‘ window ’. 3, covered by a normal
cranial scale without transparent ‘ window ’. Ordered 0–1–2–3. C57 further subdivided.
254. Dorsal scales. 0, smooth. 1, keeled.
255. Midventral scales. 0, undifferentiated.
1, slightly expanded transversely, remaining much
narrower than body width. 2, greatly expanded
transversely, approaching body width. Ordered
0–1–2. Contrary to Cundall et al. (1993) and
Wallach & Gu$ nther (1998), Anomochilus has very
slightly expanded ventral scales.
256. Subcaudals. 0, undifferentiated. 1, single
row. 2, paired row. Unordered.
257. Anal shields. 0, two or more. 1, single.
258. Hemipenis. 0, single, undivided. 1, partly
divided (bilobed distally only). 2, fully divided.
Ordered 0–1–2. Wallach & Gu$ nther (1998, Table 2,
character ‘ HP ’).
259. Sulcus spermaticus. 0, simple, not bifurcated.
1, bifurcated. This and the following character represent a recoding of the character ‘ SS ’ in Wallach &
Gu$ nther (1998, Table 2). Both states occur in
Varanus (Branch, 1982).
260. Sulcus spermaticus. 0, bifurcations centripetal.
1, bifurcations centrolineal. 2, bifurcations centrifugal. Ordered 0–1–2. Not applicable in taxa lacking
a bifurcated sulcus. Those Varanus with a bifurcation
exhibit the centripetal state (Branch, 1982).
380
(3) Ecological traits
261. Reproductive mode. 0, oviparous. 1, viviparous. C55. Mosasauroids have been confirmed to
exhibit viviparity because some basal forms (aigialosaurs) contain several large advanced embryos
(Caldwell & Lee, 2001).
262. Prey. 0, entirely invertebrates. 1, partly or
entirely vertebrates. Based mainly on Pregill et al.
(1986), Greene (1983, 1997), Cundall & Greene
(2000), Schwenk (2000). Pachyrhachis is coded as 1
based on the presence of fish tooth plates in its gut
area (Haas, 1979).
263. Constriction. 0, absent. 1, present. This is
probably partly correlated with the previous character, since snakes which take invertebrate prey do not
employ constriction, because such prey is either too
small or too weak (Cundall & Greene, 2000).
However, because the converse is not true, this
character has been treated as distinct. Data based
on Greene & Burghardt (1978), Greene (1983),
Cundall (1987), and Cundall & Greene (2000).
VI. CHARACTERS NOT EMPLOYED
Cundall et al. (1993)
C10. Supraorbital orbital ossification. 0, absent.
1, present. This was used subsequently by Tchernov
et al. (2000, character 25) and Scanlon & Lee (2000,
character 46). Based on topological and structural
criteria, the element in Dinilysia is not a supraorbital
(contra Tchernov et al., 2000) but is a postfrontal
(Rieppel, 1977 ; Estes et al., 1970 ; see character 47).
Accordingly, a supraorbital ossification is only
present in pythonines (and a single erycine) and this
character is not phylogenetically informative for the
terminal taxa employed here.
C34. Length of mandible. 0, 80–100 % skull
length. 1, less than 80 %. 2, more than 100 %. This
character is here omitted because it is redundant
with other characters included. The total length of
the mandible is the combination of the distance from
symphysis to cotyle (related to characters 70, 76 and
78, i.e. length of the free posterior process of the
supratemporal, and length and orientation of quadrate), and the length of the retroarticular process
(character 170).
Scanlon & Lee (2000)
S36. Antorbital buttress of prefrontal. 0, medial
foot process contacts palatine. 1, medial foot process
Michael S. Y. Lee and John D. Scanlon
does not contact palatine. This character is not
cladistically informative for the terminal taxa employed here.
Tchernov et al. (2000)
T5. Descending laminae of nasals. 0, articulate
with medial frontal ‘ pillars ’. 1, separated from
medial frontal pillars. T5, after Underwood (1976).
Cannot be scored in taxa lacking medial frontal
pillars (see characters 54–56). Bolyeriines, coded as
unknown by Tchernov et al. (2000), exhibit state 0
(Cundall & Irish, 1989). Not cladistically informative, because the derived state is only uniformly
present in one terminal taxon in this study.
T17. Antorbital buttress of prefrontal. 0, medial
foot process absent. 1, medial foot process low.
2, medial foot process high. The medial foot process
is the portion of the buttress medial to the lacrimal
foramen (e.g. Frazzetta, 1966 ; Cundall & Rossman,
1993). Contrary to Tchernov et al. (2000), the
medial foot process is present in all ingroup taxa,
hence the distinction between states 0 and (1j2) is
cladistically uninformative here. All the outgroups,
and leptotyphlopids, have a medial foot process.
This is similar in shape and position to that in snakes,
so it is surprising that Tchernov et al. (2000) coded
these taxa as lacking the process. The distinction
between medial foot process ‘ low ’ or ‘ high ’ is not
clarified beyond the single word descriptions, and
the medial process appears to be in a similar,
ventromedial position in all taxa where it is present.
We thus do not employ this character until states 1
and 2 are described more precisely.
T18. Antorbital buttress of prefrontal. 0, medial
foot process not expanded ventrally around lacrimal
duct. 1, medial foot process expanded ventrally
around lacrimal duct. This character is redundant
with the arrangement of bones around the foramen,
which was covered by T15 (character 45 here).
T21. Supraorbital process of parietal. 0, does not
participate in suspension of prefrontal. 1, participates
in suspension of prefrontal. The derived state is not
present in any uropeltids (e.g. Rieppel, 1977), contra
Tchernov et al. (2000), and is thus uniformly present
in only one terminal taxon, rendering this character
cladistically uninformative in both their and the
current analysis.
T31. Anteromedial foramen on maxilla. 0, absent.
1, present. T31. Tchernov et al. (2000) coded the
foramen as absent in the outgroups and a range of
snakes. However, it is present in the outgroups and
all snake taxa except colubroids, and is hence
Snake phylogeny
cladistically uninformative here. In some taxa (e.g.
Acrochordus), the foramen is very small, while in
others (Varanus, some pythonines) it has shifted
slightly and lies on the dorsomedial edge (rather
than medial surface) of the anterior end of the
maxilla. The foramen on the medial maxillary
surface in typhlopids is also provisionally considered
homologous, although this foramen is in the middle
rather than anterior portion of the element due to its
great anteroposterior shortening.
T42. Palatine-ectopterygoid contact. 0, present.
1, absent. State 0 was coded as present only in
anomalepidids by Tchernov et al. (2000) ; however,
these have the derived state (e.g. Haas, 1964, 1968).
Thus, this character is invariant across all snakes.
T63. Atlanteal crest of exoccipitals. 0, lateral
margin straight or with lateral notch. 1, with a
foramen. State 1 was scored only in Uropeltinae and
some Cylindrophis, and this character is thus not
cladistically informative as the derived state is
uniformly present in only one terminal taxon.
T64. Ventral margin of crista circumfenestralis. 0,
without large contribution from opisthotic-exoccipital, and does not articulate with pterygoid. 1, with
large contribution from opisthotic-exoccipital, and
articulates with pterygoid. This character conflates
two distinct traits, the composition of the crista
circumfenestralis and its ‘ articulation ’ with the
pterygoid. However, the first trait is uninformative
here since the opisthotic-exoccipital contributes
greatly to the crista circumfenestralis in all snakes.
The second trait, the articulation of the braincase
with the pterygoid, has not been properly described
(cf. Frazzetta, 1970). Anilius, Cylindrophis and Anomochilus were coded with state 1 ; the only detailed
description of cranial kinesis in any of these (Cundall,
1995 : Cylindrophis) provides no explicit information
on the extent or nature of their contact and how it
differs from other snakes.
T71. Alignment of trigeminal foramina. 0, V2
and V3 at approximately the same horizontal level.
1, V2 distinctly dorsal to V3. This character is
cladistically uninformative for the terminal taxa
employed here.
T72. Opisthotic-supratemporal contact. 0, distal
end of the opisthotic (paroccipital process) abuts
along its entire lateral margin against the medial
aspect of the supratemporal. 1, opisthotic-supratemporal contact is much reduced or absent. Despite
the character description, the taxa coded with state
1 in Tchernov et al.’s (2000) matrix do not have a
reduced opisthotic-supratemporal contact : the entire distal end of the opisthotic remains in contact
381
with the supratemporal. However, in these taxa the
supratemporal is much enlarged and thus not all of
it can be in contact with the paroccipital process.
Thus, this character effectively refers to the freely
projecting supratemporal process, and is redundant
with T39 (represented here by character 70).
T75. Fenestra pseudorotunda. 0, ‘ absent ’.
1, ‘ present ’. State 1 is actually present in all snakes,
and the character is thus invariant.
VII. PARSIMONY ANALYSES
As discussed above, four analyses were performed :
using all possible combinations of both outgroup
arrangements (varanoids, or an amphisbaeniandibamid clade) and both treatments of multistate
characters (ordered and unordered characters).
Analyses were performed on Macintosh G4s using
PAUP* 4 (Swofford, 1999) and associated programs,
MacClade (Maddison & Maddison, 2000) and
TreeRot 2 (Sorenson, 1999). Non-default orderings
of characters (i.e. 1–0–2 etc.) were accomplished
using stepmatrices in PAUP. Support for each
grouping was evaluated by branch support (Bremer,
1988) and bootstrapping (Felsenstein, 1985) using at
least 1000 replicates. Also, the significance of the
branch support for each clade was evaluated by
using Templeton’s (1983) nonparametric test to
compare the optimal tree with the best tree lacking
the clade (Lee, 2000 b). AutoCladeS (Ericksson,
2001) was used to generate PAUP command files for
these calculations. However, the exact nonparametric test cannot be performed in the current beta
versions of PAUP with ordered characters (presumably due to a bug), and the Kishino-Hasegawa
approximation was used ; the command files generated by AutoCladeS were modified accordingly. The
P values generated were halved since the one-tailed
test is more appropriate than the two-tailed test,
although some other assumptions are problematic
(Goldman, Anderson & Rodrigo, 2000). Where
there were multiple MPTs or constrained trees, all
comparisons were performed and the P values
averaged.
In most (but not all) cases, there was good
correlation between both values (Figs 11, 12) ; the
exceptions are discussed below. The most parsimonious tree(s) were found using branch-and-bound
searches. Branch support and bootstrap values were
calculated using heuristic searches employing 100
random addition sequences, with the PAUP command files generated by TreeRot modified accordingly. Some characters were rendered uninformative
Michael S. Y. Lee and John D. Scanlon
Er
yc
in
Bo ae
in
a
Py e
th
o
U nin
ng
a
al e
Tr iop
op
h
id iin
Bo op ae
h
ly
er iina
iin
e
ae
Co
lu
br
oi
de
a
s
is
lu is
us
ph
hi
lt
c
m
o
e
r
o
e
s
p
c
u ind
ili
om no xo
l
An Cy
An Xe Lo
a
si
ly
i
n
Di
U
ro
pe
lti
da
e
is
ch his
a
h op
i
yr
ch aas
a
P
H
+
+
A
+M
ad
+ tso
iid
ae
Le
pt
o
A typ
no
h
m lop
a
Ty
le id
ph pid ae
lo
i
pi dae
da
e
382
3,83
0.09
6,71,0.13
17,100
<0.01
2,50,0.28
Xenopeltinae
6,88,0.09
Scolecophidia Anilioidea
32,100,<0.01 4,62,0.15
us
rd
o
ch
ro
c
A
Booidea
10,99,<0.01
Caenophidia
14,100,<0.01
6,82,0.10
3,77,0.13
Advanced snakes: 8,91,0.11
Core Macrostomata: 15,100,<0.01
Macrostomata: 16,100,<0.01
Alethinophidia: 8,84,0.16
Modern snakes: 10,100,<0.01
3,74,0.27
2,67,0.21
Er
yc
in
Bo ae
in
a
Py e
th
o
U nin
ng
a
al e
Tr iop
op
h
id iin
Bo op ae
h
ly
er iina
iin
e
ae
Co
lu
br
oi
de
a
is
ch his
a
h op
i
yr
ch aas
a
P
H
+
+
B
+M
ad
+ tso
iid
ae
Le
pt
o
A typ
no
h
m lop
a
Ty
le id
ph pid ae
lo
i
pi dae
da
e
U
ro
pe
lti
da
e
4,91,0.08
s
is
s
lu
s
ph
hi
lti mu
c
o
e
r
o
e
s
p
d
c
u
om
ili
no xo
lin
An
Cy An Xe Lo
a
si
ly
i
n
Di
us
rd
o
ch
ro
c
A
13,100,<0.01
5,79,0.19
3,73,0.13
11,99,<0.01
5,87,0.07
16,100,0.02
22,100
13,99,0.07
<0.01
1,39,0.42
1,46,0.42
1,30,0.42
3,81,0.35
6,97,0.17
1,58,0.41
1,50,0.37
3,83,0.28
16,100
<0.01
6,93,0.09
4,82
0.08
Fig. 11. Snake phylogeny based on the present data set, when rooted with the (preferred) varanoid outgroup. Fossil
taxa indicated byj. Branch support, bootstrap percentage, and clade significance values shown after each clade.
Weakly supported clades (either branch support 3, or bootstrap 70 %) are indicated with thin lines, moderately
supported clades (branch support at least 3 AND bootstrap at least 70 %) are shown with normal lines, strongly
supported clades are shown with thick lines (branch support at least 6 AND bootstrap at least 90 %). (A) With
multistate characters ordered according to morphoclines, (B) with multistate characters unordered.
in some analyses (due to choice of outgroup or lack
of ordering). Because these characters can artificially
inflate the consistency index or reduce bootstrap
results, they were excluded when calculating these
measures. Polymorphism in terminal taxa was
treated as uncertainty (over the primitive state)
rather than polymorphism (in the ancestral lineage)
when calculating tree lengths.
VIII. PHYLOGENETIC RESULTS AND
EVOLUTIONARY IMPLICATIONS
The phylogenetic results are here discussed and
compared to previous studies. A formal taxonomy is
not erected since a molecular analysis of higher-level
snake phylogeny based on nuclear c-mos and
mitochondrial cytochrome b DNA sequence data is
Snake phylogeny
2,54
0.27
1,42,0.37
Er
yc
in
Bo ae
in
ae
Py
th
on
in
U
ae
ng
al
i
o
Tr
op phi
in
id
ae
Bo op
hi
ly
er ina
iin
e
ae
s
is
lu
s
us
ph
hi
lti
c
m
o
e
r
o
e
s
p
d
c
u
om
ili
no oxo
lin
L
An
An Xe
Cy
16,100
0.02
us
rd
o
ch
ro
c
A
3,84
0.09
4,82,0.18
14,100,<0.01
6,80,0.10
3,74,0.13
10,100
<0.01
18,100,0.02
7,81,0.15
Co
lu
br
oi
de
a
U
ro
pe
lti
da
e
a
si
ly
i
n
Di
Le
pt
ot
yp
A
hl
no
op
m
id
Ty ale
a
ph pid e
lo
i
pi dae
da
e
is
ch his
a
h
p
io
yr
ch aas
a
P
H
+M
ad
+ tso
iid
ae
+
+
A
383
8,86,0.09
16,100,0.03
12,99,0.07
3,46,0.27
4,48,0.30
1,26,0.38
4,73,0.03
Er
yc
in
Bo ae
in
ae
Py
th
o
U nin
ng
a
al e
i
o
Tr
op phi
in
id
a
Bo oph e
ly
iin
er
iin ae
ae
s
is
lu is
us
ph
hi
lt
c
m
o
e
r
o
e
s
p
c
u ind
ili
om eno xo
l
Lo
An Cy
An
X
a
si
ly
i
n
Di
13,100
0.03
5,89,0.03
us
rd
o
ch
ro
c
A
4,85
0.22
Co
lu
br
oi
de
a
is
ch his
a
p
h
io
yr
ch aas
a
H
P
U
ro
pe
lti
da
e
B
M
ad
ts
oi
id
ae
Le
pt
o
A typ
no
h
m lop
Ty ale ida
ph pid e
i
lo
pi dae
da
e
4,63,0.30
13,100,<0.03
6,79,0.16
11,100
0.03
3,71,0.27
5,84,0.18
12,99,0.08
14,100,0.03
10,99,0.11
4,78,0.24
2,77,0.39
Fig. 12. Snake phylogeny based on the present data set, with an amphisbaenian-dibamid clade as the outgroup. Fossil
taxa indicated by j. Branch support, bootstrap percentage, and clade significance values shown after each clade,
line thicknesses as in Fig. 11. (A) With multistate characters ordered according to morphoclines, (B) with multistate
characters unordered.
currently being performed ( J. B. Slowinski & R.
Lawson, in prep.). A combined analysis of these data
sets, along with 12S and 16S rRNA (Heise et al.,
1995), is in preparation and a formal taxonomy
should be based on all these data.
Because ordering multistate characters when they
have clear morphoclines is preferable to leaving
them unordered (e.g. Slowinski, 1993 ; Wilkinson,
1992), and varanoids appear to be the more
appropriate outgroup than an amphisbaeniandibamid clade (e.g. McDowell, 1972 ; Schwenk,
1988 ; Lee, 2000 a), we first discuss the results of the
ordered analysis employing a varanoid outgroup.
Characters diagnosing each branch are listed in full
in Table 1, and selected clades and characters are
discussed below. The other three analyses produced
broadly similar results and are discussed afterwards.
The focal analysis produced a single tree of length
L l 741, consistency index CI l 0n49, retention
index RI l 0n67 (Fig. 11 A ; see Table 2 for complete
list of supporting synapomorphies). These indices
are well within the typical range for phenotypic data
384
matrices with this number of terminal taxa (Sanderson & Donoghue, 1996 ; Wimberger & de Queiroz,
1996), indicating that convergent evolution in snakes
(at least for these characters) is not obviously greater
than in other taxa. The four included fossil taxa
emerged as basal snakes, falling outside the extant
radiation of snakes, in agreement with some previous
studies but not others (see Introduction). The
relationships of extant forms were broadly similar to
those of most recent studies (see Introduction), with
scolecophidians being the most basal living snakes
and the remaining forms comprising a monophyletic
Alethinophidia. Among alethinophidians, anilioids
are most basal, followed by the booid-like taxa, with
bolyeriines, Acrochordus, and colubroids being the
most derived (deeply nested) forms. Within these
broad areas of agreement, however, there are
substantial disagreements between the various studies, as discussed below.
(1) Basal fossil snakes and modern snakes
In addition to the characters discussed in the
introduction supporting snake monophyly, a number of characters used in the analysis (putative
synapomorphies for subgroups of snakes) optimised
instead as synapomorphies of snakes as a whole,
subsequently reversed or modified in some taxa.
These include (Table 2) : mobile maxilla-premaxilla
contact (character 12) ; palatine with anterior
process (94) and bearing teeth (179), deep posterolateral notch on dentary (149) ; vertebral zygapophyses inclined less than 30m (196), hypapophyses
present at least to tenth presacral vertebra (201).
Pachyrhachis emerges as the most basal snake, lying
outside all other snakes with strong support (branch
support 4, bootstrap 91 %, clade significance 0n08 ;
hereafter abbreviated respectively as BR, BO and
CS). The relatively low branch support, but high
bootstrap, for this clade probably reflects the large
amount of missing data in the fossil forms : because of
the relatively low number of characters determining
relationships among these taxa, it may only require
a few characters for a high bootstrap value to appear
(Lee, 2000 b). The most compelling characters
uniting Haasiophis, madtsoiids, and Dinilysia with
modern snakes include : coronoid excluded from
lateral surface of postdentary bone (162) ; more than
120 presacral vertebrae (186), trunk vertebrae with
median haemal keels (202), and at least three
transitional (‘ sacral ’ or cloacal) vertebrae with
forked, free lymphapophyses (203).
The remaining three fossil lineages (Haasiophis,
Michael S. Y. Lee and John D. Scanlon
madtsoiids and Dinilysia) are robustly positioned
between Pachyrhachis and modern (extant) snakes.
There is strong evidence placing them above
Pachyrhachis (see above) and also strong evidence
placing them below modern snakes. The monophyly
of modern snakes, to the exclusion of all the fossil
forms, is highly corroborated (BR 10, BO 100 %,
CS 0n01) by such traits as : ectopterygoid overlaps
pterygoid dorsally, not ventrally (109) ; basipterygoid processes reduced or absent (reversals in some
booids) (117), alar process of prootic reduced (131),
crista circumfenestralis converging around stapedial
footplate (135), paroccipital processes reduced
(137) ; paracotylar foramina mostly absent (197) ;
pelvic vestiges internal to ribs (211), hindlimb
vestigial (212).
These results support recent studies that placed
pachyostotic, limbed marine forms as the most basal
or primitive snakes (McDowell, 1987 ; Scanlon,
1996 ; Caldwell & Lee, 1997 ; Scanlon & Lee, 2000 ;
Rage & Escuillie! , 2000). This is consistent with the
original description of Pachyrhachis (Haas, 1979,
1980 a, b) which considered them not true snakes but
‘ snake-like varanids ’, i.e. varanids that resembled
snakes either because of convergence or because they
were on the stem lineage of snakes. By contrast, it has
recently been argued that pachyophiids are not
outside of other snakes but nested deeply within
snakes (Zaher, 1998 ; Zaher & Rieppel, 1999, 2000 ;
Tchernov et al., 2000 ; see also Greene & Cundall,
2000). The characters from these conflicting groups
of studies were all reassessed, and either included in
the current character set, or excluded for explicit
reasons (see Section VI). Most of the proposed
characters supporting a basal position for these
marine snakes remain compelling and are listed
above. However, the proposed macrostomatan
characters (Zaher & Rieppel, 1999 ; Tchernov et al.,
2000) are problematic. Zaher & Rieppel (1999)
presented alternative interpretations of four elements, and suggested that these reinterpretations
support macrostomatan affinities of Pachyrachis.
(1) Pachyrachis is reconstructed with an unforked postorbitofrontal, an interpretation that is questionable
(see character 48). Furthermore, even if accepted,
the statement that this represents a macrostomatan
character is untenable, since anilioids also have a
simple element (e.g. Cylindrophis), while conversely
there is a forked element in most pythons (Kluge,
1993 a). (2) The putative jugal is interpreted instead
as the ‘ ectopterygoid ’, which is also questionable
(see character 46). Also, even if the identification as
the ectopterygoid were accepted, the suggestion that
Snake phylogeny
the ‘ ectopterygoid ’-maxilla overlap represents a
synapomorphy uniting Pachyrhachis and macrostomatans is not tenable, as madtsoiids and some
anilioids also exhibit this trait while most scolecophidians lack a discrete ectopterygoid and so cannot
be coded for either state (see character 111). It was
further suggested that (3) the putative squamosal
identified by Lee & Caldwell (1998) instead represents the shaft of the stapes, and that (4) the putative
stapes (Haas, 1980 a ; Lee & Caldwell, 1998) instead
represents the distal end of a long paroccipital
process. However, no evidence was presented supporting these two reinterpretations. As emphasised
previously (Lee & Caldwell, 1998), identification of
the putative squamosal was very uncertain, and this
character is thus not employed here. If the hypothesized ‘ squamosal ’ is indeed actually the stapes, then
the related reinterpretation of the putative stapes as
the paroccipital process is plausible. However, even
if these reinterpretations are correct, and the original
identifications (Haas, 1980 a ; Lee & Caldwell, 1998)
erroneous, the statement that a long stapedial shaft
and a pronounced paroccipital process represent
‘ macrostomatan features ’ of Pachyrhachis cannot be
accepted. A long narrow stapes and large paroccipital process are found throughout squamates
(including the nearest outgroups to snakes) and are
presumably primitive for snakes. These symplesiomorphies therefore cannot be used to group Pachyrhachis and macrostomatans. Furthermore, macrostomatans, though possessing the long narrow stapes as
suggested by Zaher & Rieppel (1999), actually lack
a sizeable paroccipital process. The quadrate in
macrostomatans is suspended entirely by the supratemporal, and the paroccipital process is greatly
reduced or completely absent (see character 137).
Thus, none of the four Pachyrhachis-macrostomatan
synapomorphies proposed by Zaher & Rieppel
(1999) is compelling. For two, the proposed reinterpretation of the preserved element is less likely than
the previous interpretation (Haas, 1979 ; Lee &
Caldwell, 1998). For the remaining two, the proposed reinterpretation is possible (although evidence
has not yet been presented) but the distribution of
the character across snakes is not clear-cut. In this
respect, it is interesting that none of these four
characters was included in Zaher & Rieppel’s (1999)
synapomorphy scheme for snakes (their Fig. 2), even
though a Pachyrhachis-macrostomatan node was
presented and (other) supporting characters for this
node listed.
Most recently, Tchernov et al. (2000) have again
suggested that Pachyrhachis, and a new limbed snake
385
Haasiophis, are advanced forms that have re-evolved
legs. However, in that study, postcranial characters
were explicitly ignored, which is a critical omission
since many postcranial characters support an alternative, basal position for both Pachyrhachis and
Haasiophis, e.g. notched zygosphenes (character 192),
absence of dorsal processes on rib heads (207), welldeveloped pelvis, differentiated hindlimb with tibia,
fibula, and tarsals (212). Also, many apparently
primitive cranial characters of Pachyrhachis and
Haasiophis were ignored or dismissed. For instance,
the exoccipitals are preserved clearly separated
above the foramen magnum in Haasiophis but this
was rejected as a preservational artefact (see character 141). Conversely, there is acceptance of some
very dubiously preserved derived traits that might
unite Haasiophis with macrostomatans. The premaxilla is coded as toothless, yet the element is
barely exposed (see character 173). A ‘ laterosphenoid ’ (character 131) is coded as present in
Haasiophis ; however, this region is not exposed and
no laterosphenoid is shown in the interpretive
drawings. There are also a priori assumptions of
homology where the apparently primitive condition
of Haasiophis is instead interpreted as secondarily
derived, without explanation. For instance, the
‘ well-developed … basipterygoid
processes ’
of
Haasiophis (character 77), although poorly known
only from X-rays, are coded as neomorphs homologous to the parasphenoid wings of some macrostomatans but not homologous to the similar
processes of lizards (character 117). If apparently
primitive characters in Haasiophis are arbitrarily
interpreted as secondarily derived in this manner,
then no amount of evidence can lead to the
conclusion that Haasiophis is a basal snake. Finally,
some primitive features of Haasiophis were missed
because the relevant elements were misidentified.
For instance, there appears to be a jugal but this is
identified by Tchernov et al. (2000) as a ‘ postorbital ’
(character 46). These problems mean that the
phylogenetic conclusions of Tchernov et al. (2000)
are suspect, and that the original placement of these
forms outside of modern snakes (e.g. Haas, 1979 ;
Caldwell & Lee, 1997) is more likely.
The placement of madtsoiids and Dinilysia as basal
snakes, outside the modern radiation or crownclade, has also been supported recently (McDowell,
1987 ; Scanlon, 1996 ; Scanlon & Lee, 2000). The
main alternative hypothesis of madtsoiid relationships – a close affinity with booids (Simpson, 1933 ;
Rage, 1984 ; Barrie, 1990) – was based only on
subjective phenetic comparisons. The recent place-
386
ment of Dinilysia with living alethinophidians
(Rieppel, 1988 ; see also Tchernov et al., 2000) was
based on a single character (quadrate orientation)
which is here demonstrated (by character incongruence) to be convergent. Interestingly, the two
limbed marine snakes do not form a clade, but rather
comprise the two most basal snake lineages ; this is
consistent with a marine origin for snakes (e.g.
Nopcsa, 1923 ; Lee & Caldwell, 2000). The large size
and absence of burrowing adaptations in the next
two snake lineages (madtsoiids and Dinilysia) also
argue against a subterranean origin (Scanlon & Lee,
2000).
Bellairs & Underwood (1951, p. 223) noted that
Nopcsa’s (1923) theory of an aquatic origin of snakes
was ‘ based on three suppositions : that reduction of
the limbs and elongation of the neck and body may
occur in association with aquatic adaptation, that
the snakes were derived from reptiles closely related
to the early aquatic Platynota, and that the first
known snakes were themselves aquatic. We do not
feel that the last two of these assumptions have been
adequately substantiated. The most serious objection
to Nopcsa’s theory is that raised by Walls’s [1940]
interpretation of the ophidian eye … ’. In fact,
Nopcsa (1923) had already obviated objections to
the first assumption by noting the numerous parallel
origins of snake-like (i.e. eel-like) body form among
fully aquatic ‘ fish ’. The second question, of the
phylogenetic relationship of snakes to other squamates, remains contentious even 50 years later, but
recent analyses have greatly strengthened support
for close relationship to varanoids, and particularly
the aquatic mosasauroids and ‘ dolichosaurs ’ (see
Section III). The phylogenetic pattern found here
within snakes, with early, aquatic lineages basal to
all other (terrestrial and fossorial) forms, supports
the third ‘ assumption ’. This pattern, with the most
basal snake lineages represented only by fossils, also
weakens Walls’s (1940) inferences regarding eye
structures. Arguments from ophthalmic structure
cannot be extrapolated to a clade more inclusive
than the one represented by extant forms whose eyes
are available for study : the most such evidence could
ever imply is that the latest common ancestor of
Scolecophidia and Alethinophidia was fossorial.
Extrapolating it to the common ancestor of the more
inclusive clade consisting of all snakes is not justified.
Furthermore, comparisons of eyes of diverse vertebrate taxa reveals that the eyes of snakes are not
particularly similar to the eyes of fossorial forms, but
are equally similar to those of aquatic forms (C.
Caprette, unpubl. data).
Michael S. Y. Lee and John D. Scanlon
The three basal snake lineages (Pachyrhachis,
Haasiophis, Madtsoiidae) had long and flexible jaw
elements, suspended by long supratemporals, and
were thus already well-adapted for swallowing largediameter prey (Scanlon & Lee, 2000 ; Tchernov et
al., 2000). Indeed, one specimen of Pachyrhachis
contains remains of a very large ingested fish (Haas,
1979). Furthermore, mosasauroids (the likely sister
group of snakes) also possessed long flexible jaws and
large gape, and there is also direct evidence from gut
contents that at least some of them consumed
relatively large prey items (Russell, 1967). Thus,
interpretation of snake evolution as largely one of
progressive refinement of the feeding apparatus for
ingesting large prey (e.g. Greene, 1983 ; Rieppel,
1988 ; Cundall & Greene, 2000) needs to be critically
reexamined. According to the present study, the
most parsimonious interpretation is that wide gape
and large prey are primitive for snakes (Lee, Bell &
Caldwell, 1999 a) : this conclusion holds even if
amphisbaenians and dibamids, rather than mosasauroids, are assumed to be the nearest relatives of
snakes. Large gape has been secondarily lost in
scolecophidians and anilioids, most likely in connection with the need to consolidate the skull for
head-first burrowing, and for feeding in confined
spaces (see Andrews et al., 1987). Similar (and
unequivocal) reductions have occurred in burrowing
colubroid snakes (Savitsky, 1983), and indeed in
other burrowing squamate lineages such as amphisbaenians. The highly mobile intramandibular and
symphysial joints are often interpreted as adaptations for increasing gape, and their presence in gapelimited and relatively basal leptotyphlopids is anomalous if large gape is assumed to be an innovation
unique to higher snakes (Cundall & Greene, 2000).
However, if these joints and large gape were
primitive for snakes as a whole, and leptotyphlopids
are secondarily gape-limited, as suggested here, this
would explain the retention of such joints and their
co-option for the novel function of rapidly shovelling
small invertebrate prey into the mouth.
(2) Scolecophidians and alethinophidians
Modern snakes form two monophyletic sister groups :
the fossorial, small, worm-like scolecophidians, and
the generally larger, less fossorial, and more typically
snake-like alethinophidians. Scolecophidia is easily
the most strongly supported clade in this analysis
(BR 32, BO 100 %, CS 0n01), corroborated by
numerous traits, such as : ascending process of
premaxilla expanded laterally to form part of dorsal
Snake phylogeny
rim of external naris (4) ; multiple premaxillary
foramina (8) ; premaxilla on ventral surface of snout
(9) ; snout spherically expanded anterior to the
orbits (10) ; anteriorly positioned optic foramen
(61) ; supratemporal vestigial or absent (69) ; quadrate articulated low on braincase (77), strongly
inclined anteroventrally (78) and with suprastapedial process forming very obtuse angle with shaft
(75) ; fenestra for duct of Jacobson’s organ opening
posteroventrally (83) ; pterygoid lacks ectopterygoid
process (101) and with slender, rod-like quadrate
process (103) that extends posteriorly well past
quadrate-articular joint (107) ; crista sellaris greatly
reduced (128) ; footplate of stapes completely enclosed by crista circumfenestralis (135) ; posterolateral margin of dentary lacks notch for compound
bone (149) ; splenial neither pierced by a foramen
(156) nor notched dorsally (157) ; surangular foramen posterior to coronoid process (167) ; teeth few
on dentary (177) and absent on palatine (179) and
pterygoid (181) ; atlas neural arch with weak or no
dorsoposterior process (188), vertebral condyles
depressed (193) and strongly oblique (194), subcentral foramina irregular (199), prezygapophyseal
processes very prominent (200), caudal haemapophyses absent (204) ; tracheal entry to right lung
terminal (231), left lung entirely absent (232) ; eyes
reduced and covered by scales (253). The rapid jaw
oscillations of typhlopids and leptotyphlopids (Kley
& Brainerd, 1996 ; Cundall & Greene, 2000) might
represent another scolecophidian synapomorphy –
however, homology is uncertain because these
movements are exhibited by the upper jaw in
typhlopids, the lower jaw in leptotyphlopids, and
have yet to be confirmed in anomalepidids.
Many of the above traits are correlates of
burrowing (Lee, 1998 ; Rieppel & Zaher, 2000 b) :
downweighting these non-independent traits might
reduce support for this clade. Nevertheless, the
monophyly of scolecophidians has been almost
universally accepted in recent times (Underwood,
1967 ; Rieppel, 1988 ; Cundall et al., 1993 ; Scanlon &
Lee, 2000 ; Tchernov et al., 2000). The only recent
challenge to scolecophidian monophyly was a molecular study (Heise et al., 1995). However, that study
lacked adequate taxon sampling of basal snakes since
it was aimed mainly at resolving relationships among
derived snakes (colubroids) : for instance, only a
single anilioid was sequenced. Furthermore, the
published tree was based on distance rather than
cladistic analyses. On this tree the putative paraphyly of scolecophidians was only weakly supported
(bootstrap 67 %) and the status of the group should
387
perhaps have been more conservatively interpreted
as unresolved.
Within scolecophidians, typhlopids and anomalepidids are very strongly united to the exclusion of
leptotyphlopids (BR 17, BO 100 %, CS 0n01) by
traits such as : maxilla almost free of prefrontal
(34, 35), highly mobile and widely separated from
premaxilla (12) and lacking anterior process (13),
palatine process (19) and lateral foramina (16) ;
alveolar row oriented transversely (17) ; subolfactory
processes of frontal separated ventrally (55) ; dorsolateral flange of septomaxilla contributing to margin
of external naris (81) and with posteromedial
expansion projecting towards frontal (80) ; palatine
very short (87) and with slender choanal process
(92) ; dentary very short (147) ; splenial largely
exposed laterally (159) ; retroarticular process elongate (170) ; pelvis reduced to a single pair of elements
(210) ; hindlimb absent (212) ; tracheal lung present
(230). This hypothesis, and some of these supporting
characters, have been proposed before (e.g. Underwood, 1967 ; Rieppel, 1988 ; Cundall et al., 1993 ;
Scanlon & Lee, 2000). Some other studies, however,
have united leptotyphlopids and typhlopids to the
exclusion of anomalepidids (McDowell, 1987 ; Heise
et al., 1995). However, these studies either used
subjective, non-cladistic evaluation of morphological
traits (McDowell, 1987) or distance analysis of
molecules (Heise et al., 1995), methods now widely
seen as unreliable.
All other living snakes are united as the Alethinophidia (BR 8, BO 84 %, CS 0n16). This clade is
corroborated by numerous highly distinct novelties,
such as : frontals with medial descending processes
forming interolfactory pillars (56) ; quadrate with
‘ stylohyal ’ ossification on medial surface articulating
with tip of stapes (73) ; septomaxilla dorsolateral
edge with posterior spine (80) ; vomeronasal nerve
pierces vomer via a single large foramen (85) ;
posterior dorsal lamina of vomer well developed
(89) ; sphenoid with dorsolateral wing at parietalprootic boundary (119) ; vidian canal opens anteriorly on internal surface of sphenoid wing (124) ;
laterosphenoid present and fused to prootic, separating foramina for maxillary and mandibular branches
of trigeminal nerve (130) ; coronoid process formed
mainly by surangular lamina of compound (165) ;
first branchial arch elements absent (184) ; haemapophyses fused to centra of caudal vertebrae (205) and
free distally, not forming chevrons (206) ; m. adductor
mandibulae externus pars superficialis inserts on adductor
externus profundus (215) ; two pairs of thymus glands
(235). This grouping has also been widely accepted,
388
based in part on these characters (Underwood,
1967 ; Rieppel, 1988 ; Cundall et al., 1993 ; but see
Dowling & Duellman, 1978) and molecular sequences (Heise et al., 1995). However, despite the
compelling synapomorphies and its almost universal
acceptance, statistical support for Alethinophidia is
not as high as expected, due to numerous conflicting
characters that unite some burrowing forms (notably
uropeltids and Anomochilus) with scolecophidians.
These characters, however, are mostly losses and
reductions correlated with subterranean adaptation
(Lee, 1998 ; Rieppel & Zaher, 2000 b). By contrast,
many of the characters supporting alethinophidian
monophyly are derived evolutionary novelties. If
simple losses and reductions are weighted less than
complex presences (e.g. Bryant, 1989 ; Neff, 1986),
then the support for Alethinophidia would increase
to (arguably) more ‘ realistic ’ levels.
(3) Anilioids and macrostomatans
Anilioids emerge as basal to all other alethinophidians (macrostomatans) ; this arrangement (i.e.
macrostomatan monophyly) is strongly supported.
However, whether anilioids are the monophyletic
sister group, or paraphyletic stem group, of macrostomatans is largely unresolved. They emerge in the
analysis as monophyletic, as proposed long ago by
Underwood (1967) ; but this is only weakly supported (BR 4, BO 62 %, CS 0n15). Many characters
supporting anilioid monophyly are present (but
inferred to be convergent) in other taxa – e.g.
supratemporal not projecting posteriorly beyond
otic capsule (70), choanal processes of palatines
projecting ventromedially between posterior parts
of vomers (91), distal facets of basipterygoid processes elongated parasagitally (118), basioccipitalsphenoid suture located posteriorly, below fenestra
ovalis (121). Others are found in the outgroups
but optimise as reversals diagnosing anilioids – e.g.
ectopterygoid process of pterygoid well developed
(101) and located anterior to basipterygoid process
(102). Two characters may be of more value but
require further investigation : basioccipital with
elongate posterolateral processes (122), and cerebral
carotid foramina lateral to hypophysial pit rather
than within it (127). An albumin immunological
distance study (Cadle et al., 1990) found support for
uropeltidsjCylindrophis (while Anilius, Xenopeltis and
a scolecophidian were much more distant and
unresolved), which agrees with the present analysis
and Tchernov et al. (2000), and contrasts with the
paraphyletic arrangement of Cundall et al. (1993).
Michael S. Y. Lee and John D. Scanlon
Similarly, relationships among anilioids are only
weakly resolved, apart from an Anomochilus-uropeltid
clade (BR 6, BO 71 %, CS 0n13). The relatively poor
support for monophyly of, and relationships among,
anilioids is again due to a conflicting signal of
burrowing characters uniting uropeltids and Anomochilus with scolecophidians. These confounding
characters would reduce support for branches between the uropeltid-Anomochilus clade and scolecophidians, i.e. Anilioidea and Alethinophidia. These
branches have relatively low bootstrap support
compared to branch support, as would be predicted
of clades supported and contradicted by large
opposing suites of characters (Lee, 2000 b). Other
recent studies have also been ambivalent regarding
anilioids : monophyly was tentatively proposed by
Rieppel (1988 ; see also Tchernov et al., 2000) based
on a single character now known to be present in
most other snakes (the perilymphatic foramen),
while paraphyly was proposed by Cundall et al.
(1993) but again, weakly supported.
Monophyly of the Macrostomata is extremely
strongly supported (BR 16, BO 100 %, CS 0n01)
by such traits as : maxilla with reduced dorsal
process (14) and in mobile contact with prefrontal
(35) ; orbital margin of prefrontal vertical (39) ;
lacrimal duct roofed by horizontal lamina of prefrontal (40) ; septomaxilla not projecting anteriorly
to overlap premaxilla and\or anterior tip of maxilla
(79) ; medial fenestra in vomeronasal cupola reduced
or absent (86) ; palatine choanal process not or
barely contacting vomer (90) ; pterygoid quadrate
ramus nearly horizontal (106) ; lateral margin of
ectopterygoid with distinct angle (113) ; stapes
associated with shaft of quadrate, not suprastapedial
process (146) ; dentary with exposed posteromedial
process (151) ; coronoid without posteroventral expansion (161) and largely hidden laterally by
surangular (162) ; neural arches with deep, V-shaped
posterior embayment (191) ; adductor externus superficialis muscle inserts on ventrolateral surface of
mandible (214) ; anterior portion of intermandibularis
anterior muscle not covered by genioglossus medially
(224) ; transversely expanded ventral scales (255).
Macrostomata is here used in the broader sense to
include Xenopeltis and Loxocemus. These taxa are, to
some extent, intermediate between basal snakes and
other macrostomatans, and thus their inclusion or
exclusion from Macrostomata is somewhat arbitrary
(Cundall & Greene, 2000). Following earlier studies
(e.g. Rieppel, 1988 ; Scanlon & Lee, 2000), we
provisionally include them, and refer to the more
restricted clade which excludes them as ‘ core
Snake phylogeny
macrostomatans ’. A division between macrostomatans (sensu lato) and other modern snakes
(anilioids and scolecophidians) was made by early
workers including Mu$ ller (1831, coining the names
Macrostomata and Microstomata), Stannius (1856),
and Cope (1864, 1900). This highly distinct clade,
and some of the supporting characters, have been
identified in other recent studies (Rieppel, 1988 ;
Cundall et al., 1993 ; Scanlon & Lee, 2000 ; Cundall
& Greene, 2000). Tchernov et al. (2000) retrieved a
similar clade, but also embedded Pachyrhachis and
Haasiophis inside it based on (problematic) evidence
discussed above. By contrast, the only recent studies
which contradict macrostomatan monophyly both
have methodological shortcomings (see above) :
McDowell (1975, 1987), linked xenopeltids and
anilioids, and Heise et al. (1995) united the only
erycine and only anilioid sampled, a highly heterodox arrangement that has no support from any
other study.
While the monophyly of Macrostomata is well
supported by many unequivocal synapomorphies,
only two of them (pertaining to the maxillaprefrontal contact) are clearly related to enlarged
gape. The diagnosis given above does not include
many other obvious gape-related features that
distinguish macrostomatans from scolecophidians
and anilioids : macrostomatans have long, flexible
jaws suspended by long supratemporals, and an
ability to ingest large prey. These are usually
interpreted as unique innovations of macrostomatans. However, as discussed above, these traits are
also found in basal fossil snakes and in mosasauroids,
and thus might be retained primitive features of
snakes which have been lost in scolecophidians and
reduced in anilioids, possibly in connection with
burrowing habits.
(4) Xenopeltids and ‘ core macrostomatans ’
Xenopeltis and Loxocemus form a well-supported clade
(BR 6, BO 88 %, CS 0n09) that is the sister taxon to
all other macrostomatans, which are here informally
termed ‘ core macrostomatans ’. Diagnostic characters include : prefrontal with greatly elongated
anteroventral process (31) and distinct anterodorsal
process (32) ; lateral process of palatine contacts
lateral foot process of prefrontal (36) and extends to
lateral margin of maxilla (38) ; vomer broad
posteriorly (88) and not contacting palatine choanal
process (90) ; paroccipital process greatly reduced
(137) ; splenial not pierced by a foramen (156) ;
389
coronoid process greatly reduced (164) ; second
intercentrum fused to centrum of axis (189) ; prezygapophyseal processes prominent (200) ; pelvic girdle
reduced (210) ; cartilaginous choanal arc uninterrupted (249) ; subcaudal scales paired (256). Underwood (1976, p. 167) named this clade the Xenopeltinae and noted additional synapomorphies – the
possession of an (enlarged) occipital scale and
tubercles on the belly. This association, however, has
been widely contradicted. Cundall et al. (1993) and
Tchernov et al. (2000) placed Xenopeltis and Loxocemus
as successive outgroups to core macrostomatans, i.e.
not monophyletic but in the same region of the tree.
Underwood (1967) suggested a similar basal macrostomatan position for Xenopeltis but placed Loxocemus
within core macrostomatans, as booids. Similarly,
Heise et al. (1995) placed Loxocemus within core
macrostomatans as booids, and in particular with
pythonines (Xenopeltis was not sampled).
Core macrostomatans form a very strongly supported clade (BR 15, BO 100 %, CS 0n01),
diagnosed by characters such as : anterior tip of
maxilla lacks medial flange (15) ; prefrontal with
anterior process (apex) free of dorsal surface of
maxilla (34) but medial foot process in contact with
maxilla (37) ; supratemporal extends nearly or as far
anteriorly as prootic (72) ; foramen in prootic for
hyomandibular branch of facial nerve located within
opening for trigeminal (133) ; compound mandibular element with ventrolateral ridge for muscle
insertion (168) ; zygapophyses of trunk vertebrae
close to horizontal (196) ; levator anguli oris absent
(216), insertion tendon of adductor externus medialis
muscle reduced (217), quadrate tendon of adductor
externus profundus muscle reduced (218) ; retractor
vomeris muscle present (220) ; intermandibularis anterior
muscle divided into anterior and posterior portions
(223) ; liver relatively short (240) ; nasal gland and
aditus conchae roofed by prefrontal (248). This
grouping was also very strongly supported in the
only three detailed cladistic analyses of higher snake
phylogeny-Cundall et al. (1993), Scanlon & Lee
(2000) and Tchernov et al. (2000) – who identified
some of these synapomorphies. Cundall & Greene
(2000, p. 296) reviewed the evidence for this clade
and suggested that its diagnostic synapomorphies
‘ all correlate with increased gape size ’. Most of the
above traits are indeed gape-related, being related to
greater flexibility of the jaw skeleton and sophistication of the jaw adductor musculature. However,
a few characters are from other organ systems (e.g.
characters 133, 196, 240, 248) and thus potentially
independent. As noted above, an intuitive mor-
390
phological analysis (Underwood, 1967) and a phenetic molecular study (Heise et al., 1995) both placed
Loxocemus within core macrostomatans.
(5) Booids and ‘ advanced snakes ’
All recent cladistic analyses (Cundall et al., 1993 ;
Scanlon & Lee, 2000 ; Tchernov et al., 2000) have
placed the large, constricting Booidea (erycines,
pythonines, and boines) as basal to ‘ advanced
snakes ’, an informal term here used in the same sense
as Kluge (1991), to refer to ungaliophiines, tropidophiines, bolyeriines and caenophidians. However,
whether booids are the monophyletic sister group, or
paraphyletic stem group, of advanced snakes has
been contentious. In this analysis (and Scanlon &
Lee, 2000), they emerge as very strongly monophyletic (BR 10, BO 99 %, CS 0n01), based on
traits such as : palatal process of premaxilla largely
or entirely separated from vomer (6) ; palatine
process of maxilla contacts pterygoid (20) ; descending lamina of nasals deep anteriorly (30) ; dorsolateral surface of prootic mostly covered by supratemporal (71) ; anterior process of palatine free of
septomaxilla and vomer (95) ; ectopterygoid attaches
to lateral edge rather than dorsal surface of pterygoid
(109) ; vidian canals asymmetrical (125) ; distal end
of stapes associated with middle or ventral region of
quadrate shaft (146) ; vertebrae with strong precondylar constriction (195), and small scattered pits or
foramina in parazygantral area (198) ; intercostal
arteries span multiple segments (234). The only
strong advocate for booid monophyly in recent times
has been McDowell (1987, p. 28), who stated that
the Booidea were a ‘ sharply defined group without
clear relationships to any other ’ ; some of the
characters listed in his diagnosis could be seen as
synapomorphies of this group : palatine separated
from vomer beneath choanal passage (cf. 95) ; large
‘ facial wing ’ of the prefrontal conceals lateral nasal
gland (cf. 248) ; and ‘ tabular ’ (l supratemporal)
without lateral lobe and well separated from rim of
juxtastapedial fossa (not used in subsequent analyses,
and also omitted here due to difficulty of recognizing
presence and absence of lateral lobe as discrete
states). By contrast, the cladistic morphological
analysis of Cundall et al. (1993) left booids in an
unresolved polytomy, while that of Tchernov et al.
(2000) found weak evidence for paraphyly. In both
these studies, boines and erycines were united but
also very weakly. The phenetic molecular analysis of
Heise et al. (1995) found booids to be polyphyletic,
linking erycines with uropeltids, pythonines with
Michael S. Y. Lee and John D. Scanlon
Loxocemus, and boines with advanced snakes. Within
booids, pythonines and boines are here united to the
exclusion of erycines (BR 3, BO 83 %, CS 0n09)
based on characters such as : horizontal lamina of
nasal narrow posteriorly (26) ; prefrontals in neartransverse contact with frontals and approaching
midline (42) ; ectopterygoid with two anterior
processes (112) ; hypophysial pit bounded anteriorly
by a distinct ridge (126) ; prominent haemal keels or
hypapophyses present in middle and posterior trunk
(201). The extraordinary elasticity of intermandibular tissues in boines and pythonines (Cundall &
Greene, 2000) might represent a further shared
innovation (synapomorphy).
The small booid-like groups (ungaliophiines,
tropidophiines, bolyeriines) emerge as more closely
related to caenophidians than to booids. The clade
containing all these taxa is here informally termed
‘ advanced snakes ’ (as in Kluge, 1991), and is
strongly supported (BR 8, BO 91 %, CS 0n11).
Diagnostic characters include : basipterygoid processes absent (117) ; coronoid bone greatly reduced
(160) ; hyoid cornua parallel (185) ; intermandibularis
posterior muscle with distinct pars posterior (226) ;
tracheal lung present (230) ; tracheal entry to lung
terminal (231) ; left lung absent (232) ; ileocolic
caecum absent (247) ; hemipenis strongly bilobed
(258). Previous cladistic studies (Kluge, 1991 ;
Cundall et al., 1993 ; Scanlon & Lee, 2000 ; Tchernov
et al., 2000) have identified this clade and some of the
supporting synapomorphies.
Among advanced snakes, ungaliophiines, tropidophiines and bolyeriines are successive sister groups to
caenophidians, an arrangement having moderate
support. In the present analysis, tropidophiines,
bolyeriines and caenophidians are moderately
strongly united (BR 3, BO 77 %, CS 0n13) to the
exclusion of ungaliophiines by characters such as :
horizontal laminae of nasals and frontals separated
(27) ; preorbital ridge of frontal weak or absent (53) ;
kidneys not lobed (244) ; dorsal scales keeled (254).
Within this clade, bolyeriines and caenophidians are
strongly united (BR 6, BO 82 %, CS 0n10) by
characters including : posterior margin of orbit
incomplete (51) ; pterygoid with straight medial
margin (104) ; ectopterygoid with bifurcated anterior end (112) ; paroccipital process vestigial (137) ;
nine or more palatine teeth (180) ; pelvis and
hindlimb totally absent (210, 212). Various arrangements of the small booid-like groups have been
found in other studies, differing in the relative
position of ungaliophiines, tropidophiines and
bolyeriines among themselves or along the caeno-
Snake phylogeny
phidian stem. Monophyly of Tropidophiidae (sensu
McDowell, 1975, comprising ungaliophiines and
tropidophiines) was supported by Kluge (1991) and
Wallach & Gu$ nther (1998) and assumed by Cundall
et al. (1993). By contrast, Zaher (1994), Tchernov
et al. (2000) and the present study suggest that
‘ tropidophiids ’ are paraphyletic, although the exact
affinities of the component groups differ in all these
studies.
Caenophidians include Acrochordus and the
highly diverse Colubroidea, which contains many
non-venomous forms as well as all venomous forms
(elapids, atractaspidids, opisthoglyphous colubrids,
and viperids). Caenophidia is very strongly supported in the present analysis (BR 14, BO 100 %,
CS 0n01) as well as all other recent analyses
(Rieppel, 1988 ; Cundall et al., 1993 ; Scanlon & Lee,
2000 ; Tchernov et al., 2000), and has been proposed
repeatedly by earlier studies that used intuitive
phylogenetic methods (e.g. Underwood, 1967 ;
Groombridge, 1979 c). Synapomorphies include :
dorsal surface of palatine process of maxilla lacking a
large foramen (21) ; prefrontal separated from nasal
(33) and with narrow antorbital buttress (44) ; supratemporal extends anteriorly as far or further than
prootic (72) ; quadrate shaft long (76) and inclined
posteriorly (78) ; septomaxilla contacts interolfactory
pillar of frontal (82) ; lateral margin of fenestra for
Jacobson’s organ formed partly by the vomer (84)
which is pierced posteriorly by a cluster of foramina
for the vomeronasal nerve (85), and has a welldeveloped, vertical posterior process (89) ; posterior
end of maxilla free of ectopterygoid (110) ; coronoid
absent as a distinct element (160) ; pterygoideus muscle
with divided origin (219) ; pterygoid bone lacks
direct ligamentous connection to braincase (222) ;
pars anterior of intermandibularis anterior muscle inserts
on interramal pad (225). The alternative view that
Acrochordus is a very basal snake distantly related to
colubroids (e.g. McDowell, 1979, 1987 ; Rage, 1987)
is based on a small number of ‘ key ’ characters such
as the development of the crista circumfenestralis or
the number of mental foramina ; these traits have
been included here and demonstrated to be homoplastic based on incongruence with other characters.
Recent molecular studies have further affirmed the
close relationship between Acrochordus and colubroids
(Kraus & Brown, 1998 ; Gravlund, 2001). Our usage
of Caenophidia includes Acrochordus as well as
colubroids (as in Rieppel, 1988). Excluding Acrochordus from Caenophidia (Underwood, 1967) would
make it synonymous with Colubroidea and thus
redundant. Given the extremely strong evidence for
391
close relationships between Acrochordus and colubroids, it seems reasonable to use the more general
interpretation of Caenophidia and apply it to this
clade, rather than synonymizing Caenophidia with
Colubroidea and coining a new name.
IX. SENSITIVITY TO OUTGROUP AND
EVOLUTIONARY ASSUMPTIONS
Differences in outgroup assumptions, or character
coding, might have biased the current, and previous
analyses of snake phylogeny (e.g. see Coates & Ruta,
2000). As a result, the two most plausible outgroup
hypotheses (varanoids or amphisbaeniansjdibamids), and the two alternative methods of character
coding (ordered or unordered), were employed,
producing a total of four analyses. The ‘ focal ’
analysis with the varanoid outgroup and ordered
characters is described in detail above ; the other
three analyses are now briefly compared to this
result to ascertain its sensitivity to alternative
outgroup assumptions and character coding
methods.
The analysis which retained the varanoid outgroup, but treated all characters as unordered,
produced a single most parsimonious tree (Fig. 11 B ;
L l 709, CI l 0n51, RI l 0n67) very similar to the
single tree found in the focal analysis. The only
difference is that the four ‘ anilioid ’ taxa form a
paraphyletic ‘ comb ’ at the base of Alethinophidia
rather than being united as a clade. They diverge
from the alethinophidian stem in the same sequence
as in the analysis of Cundall et al. (1993, Fig. 2), but
this region of the tree is weakly supported in both
studies. The branch support and bootstrap frequencies for most of the shared clades are somewhat
lower than in the previous analysis. The subterranean-origin hypothesis is again contradicted by
the basal position of the aquatic and large-terrestrial
fossil taxa (as in the focal analysis). However, the
paraphyly of ‘ anilioids ’ would strongly support the
idea that fossorial habits developed in an ancestor of
all modern snakes and were lost only in macrostomatans (as suggested by Cundall et al., 1993).
The analysis with the amphisbaenian-dibamid
outgroup, and ordered characters, produced a
broadly similar result. The two most parsimonious
trees (L l 766, CI l 0n47, RI l 0n67) and the strict
consensus (Fig. 12 A) were almost identical to that of
the focal analysis, except in two areas. The four fossil
taxa form a basal clade, rather than a paraphyletic
392
assemblage (although they still lie outside all other
snakes as before), and relationships among the most
basal alethinophidians are unresolved (Anomochilus,
Uropeltidae, and the remaining alethinophidians
form a trichotomy). Relationships within scolecophidians, and macrostomatans, are identical to those
in the focal analysis. When the amphisbaeniandibamid outgroup was used and characters treated
as unordered, a very similar but less resolved tree
results. 7 equally parsimonious trees are found (L l
729, CI l 0n49, RI l 0n67 ; consensus, Fig. 12 B),
and the strict consensus collapses some regions that
are poorly supported in the ordered analysis. The
four fossil taxa, and the four anilioids plus macrostomatans, form four- and five-way polytomies, and
there is also a trichotomy between the fossil clade,
Scolecophidia, and Alethinophidia.
The results of this study, then, hardly change
when characters are treated as unordered rather
than ordered – only some poorly supported clades
collapse. Use of an alternative (amphisbaeniandibamid) outgroup results in a more significant
change. Scolecophidians shared many burrowing
adaptations with the amphisbaenian-dibamid outgroup : all these forms are small, worm-like, subterranean forms with small jaws, reduced eyes and
consolidated skulls lacking many elements. There is
thus a signal that pulls scolecophidians towards the
base of the snake tree, and the status of the fossil taxa
as a comb of successive outgroups to extant snakes is
therefore weakened (ordered analysis) or disappears
(unordered analysis). The base of the snake tree is
effectively a three-way polytomy between the fossil
forms, scolecophidians, and alethinophidians. However, the amphisbaenian-dibamid outgroup is less
well supported than the varanoid outgroup, based
on extrinsic evidence (higher-level analyses of
squamate phylogeny, see Section I). Furthermore,
even in the two analyses using the less supported
amphisbaenian-dibamid outgroup, there is support
for the position of the fossil taxa outside alethinophidians (BR 4\2, BO 73\77, CS 0n3\0n39), and
strong support for their position outside macrostomatans (BR 12\10, BO 99\99, CS 0n07\0n11). It
is only their basal position with respect to scolecophidians that becomes questionable. The monophyly
of, and relationships within, macrostomatans remain
constant in all four analyses (Figs 11, 12), which is
consistent with the intuitive expectation that perturbation of the outgroup should most strongly
influence basal relationships within the ingroup.
Thus, the four analyses refute the proposed ‘ high ’
placement (Tchernov et al., 2000) of Pachyrhachis and
Michael S. Y. Lee and John D. Scanlon
Haasiophis within alethinophidians in general, and
within macrostomatans in particular.
Finally, we acknowledge that, as with any large
phylogenetic study, there will undoubtedly be many
character codings that contain errors or at least
questionable interpretations. We hope that future
workers will evaluate and refine these results, not
just by highlighting apparent mistakes, but also by
presenting their own hypotheses backed up by their
own data. It is very easy to ‘ nitpick ’ other people’s
work, but in order to refute a hypothesis conclusively,
one should present an alternative, better corroborated hypothesis along with the supporting evidence. This takes boldness, since it entails subjecting
oneself to similar scrutiny. Science proceeds not just
by refutations, but more importantly, by conjectures ; good scientists should do both.
X. ACKNOWLEDGEMENTS
We wish to acknowledge in particular the discussion and
encouragement provided by Garth Underwood, whose
work formed much of the background and inspiration for
this study. We also thank the following for extensive
discussion and comments, though they do not necessarily
agree with all (or even many) of our conclusions : Michael
Caldwell, Harry Greene, David Cundall, Chris Caprette,
Van Wallach, Jean-Claude Rage, Herndon Dowling and
the late Joe Slowinski. We are grateful to the following
people for access to and assistance with specimens in
collections under their care : Darrel Frost, Tom Trombone, Linda Ford, Charlotte Holton (AMNH) ; Allen
Greer, Ross Sadlier (AMS) ; John Wombey (ANWC) ;
Colin McCarthy, Garth Underwood, Nick Arnold, Mark
Wilkinson, Barry Clarke (BMNH) ; Robert Carroll
(RM) ; Joe Slowinksi (CAS) ; Alan Resetar, Olivier
Rieppel (FMNH) ; Eitan Tchernov (HUJ) ; Ruggero
Calligaris, Flavio Bacchia (MCSNT) ; Van Wallach, Jose
Rosado (MCZ) ; Hermann Kollman (NHW) ; Kevin de
Queiroz, George Zug (NMNH) ; John Coventry (NMV),
Patrick Couper, Jeanette Covacevich, Andrew Amey
(QM) ; Richard Estes, Tod Reeder (REE) ; Mark
Hutchinson, Steve Donnellan (SAM) ; Jenny Clack, Ray
Symonds (UMZC) ; Michael Archer, Henk Godthelp
(UNSW, AR) ; Ken Aplin (WAM). Financial support to
M. L. came from the Australian Research Council, the
Fulbright Foundation, the Australian Vice-Chancellors
Committee ; and to J. S. from the Australian Commonwealth Postgraduate Awards scheme, University of
New South Wales Faculty of Biological and Behavioural
Sciences (Sydney), Rheinische Friedrich-Wilhelms Universita$ t (Bonn), Hessiches Landesmuseum (Darmstadt)
and the Joyce W. Vickery Scientific Research Fund.
Snake phylogeny
XI. REFERENCES
A, R. M., P, F. H., C, A. &  Q, A.
(1987). The ecological cost of morphological specialization :
Feeding by a fossorial lizard. Oecologica 73, 139–145.
A, A. O. (1997). Paleogene sea snakes from the eastern
part of Tethys. Russian Journal of Herpetology 4, 128–142.
B, D. J. (1990). Skull elements and associated remains of
the Pleistocene boid snake Wonambi naracoortensis. Memoirs of
the Queensland Museum 28, 139–151.
B, A. d ’A. (1949). Observations on the snout of Varanus,
and a comparison with that of other lizards and snakes.
Journal of Anatomy 83, 116–146.
B, A. d ’A. (1972). Comments on the evolution and
affinities of snakes. In Studies in Vertebrate Evolution (eds. K. A.
Joysey and T. S. Kemp), pp. 157–172. Oliver and Boyd,
London.
B, A. d ’A. & K, A. M. (1981). The chondrocranium and the development of the skull of recent reptiles. In
Biology of the Reptilia, Vol. 11 (eds. C. Gans and T. S. Parsons),
pp. 1–263. Academic Press, New York.
B, A. d ’A. & U, G. (1951). The origin of
snakes. Biological Reviews 26, 193–237.
B, D. S. (1973). Spathorhynchus fossorium, a middle Eocene
amphisbaenian (Reptilia) from Whamming. Copeia 1973,
704–721.
B-E, O. R. P., B, H. N. & R, A. P.
(1998). Supraspecific taxa as terminals in cladistic analysis :
implicit assumptions of monophyly and a comparison of
methods. Biological Journal of the Linnean Society 64, 101–133.
B, C. M. (1961). Die Amphibien und Reptilien Europas
[review]. Copeia 1961, 252–256.
B, C. M. (1964). Snakes of the genera Diaphorolepis and
Synophis and the colubrid subfamily Xenoderminae (Reptilia,
Colubridae). Senckenbergiana Biologica 45, 509–531.
B, C. M. (1968). A new genus and species of dwarf boa
from southern Mexico. American Museum Novitates 2354, 1–38.
B, G. A. (1891). Notes on the osteology of Heloderma
horridum and H. suspectum, with remarks on the systematic
position of the Helodermatidae and on the vertebrae of the
Lacertilia. Proceedings of the Zoolocgical Society of London 59,
109–118.
B, G. A. (1893). Catalogue of the snakes in the British
Museum (Natural History), Vol. I. Taylor & Francis Ltd.,
London.
B, W. R. (1982). Hemipenial morphology of platynotan
lizards. Journal of Herpetology 16, 16–38.
B, K. (1988). The limits of amino-acid sequence data in
angiosperm phylogenetic reconstruction. Evolution 42, 795–
803.
B, H. N. (1989). An evaluation of cladistic and character
analyses as hypothetico-deductive procedures, and the consequences for character weighting. Systematic Zoology 38, 214–
227.
B, H. N. & R, A. P. (1992). The role of phylogenetic analysis in the inference of unpreserved attributes of
extinct taxa. Philosophical Transactions of the Royal Society of
London, Series B (Biological Sciences) 337, 405–418.
B, L., K, M. &  Q, K. (2000). Slowlyevolving protein loci and higher-level snake phylogeny : a
reanalysis. Herpetologica 56, 324–332.
C, J. E. (1988). Phylogenetic relationships among ad-
393
vanced snakes : a molecular perspective. University of California
Publications in Zoology 119, i–x, 1–77.
C, J. E., D, H. C., G, C. & G, D. F.
(1990). Phylogenetic relationships and molecular evolution in
uropeltid snakes (Serpentes : Uropeltidae) : allozymes and
albumin immunology. Biological Journal of the Linnean Society
40, 293–320.
C, M. W. (1999). Squamate phylogeny and the
relationships of snakes and mosasauroids. Zoological Journal of
the Linnean Society 125, 115–147.
C, M. W. & L, M. S. Y. (1997). A snake with legs
from the marine Cretaceous of the Middle East. Nature 386,
705–709.
C, M. W. & L, M. S. Y. (2001). Live birth in
Cretaceous marine lizards (mosasauroids). Proceedings of the
Royal Society of London, Series B (Biological Sciences), 268,
2397–2401.
C, C. L. (1923). Classification of the lizards. Bulletin of the
American Museum of Natural History 48, 289–481.
C, R. L. & DB, M. (1992). Aigialosaurs : midCretaceous Varanoid lizards. Journal of Vertebrate Paleontology
12, 66–86.
C, M. & R, M. (2000). Nice snake, shame about the
legs. Trends in Ecology and Evolution 15, 503–507.
C, W. E. (1997). Independent evolution of squamate
olfaction and vomerolfaction and correlated evolution of
vomerolfaction and lingual structure. Amphibia-Reptilia 18,
85–105.
C, E. D. (1864). On the characters of the higher groups of
Reptilia Squamata – and especially of the Diploglossa. Proceedings of the Academy of Natural Sciences of Philadephia
1864, 224–231.
C, E. D. (1893). Prodromus of a new system of the nonvenomous snakes. American Naturalist 27, 477–483.
C, E. D. (1894). On the lungs of the Ophidia. Proceedings of
the American Philosophical Society 33 (145), 217–225, pls 9–16.
C, E. D. (1895). The classification of the Ophidia. Transactions of the American Philosophical Society 18, 186–219, pls 14–33.
C, E. D. (1900). The crocodilians, lizards and snakes of
North America. Reports of the U.S. National Museum 1898,
153–1249, pls 1–36.
C, D. (1981). Cranial osteology of the colubrid snake
genus Opheodrys. Copeia 1981, 353–371.
C, D. (1987). Functional morphology. In Snakes : Ecology
and Evolutionary Biology (eds. R. A. Seigel, J. T. Collins and
S. S. Novak), pp. 106–140. Macmillan, New York.
C, D. (1995). Feeding behaviour in Cylindrophis and its
bearing on the evolution of alethinophidian snakes. Journal of
Zoology, London 237, 353–376.
C, D. & G, H. W. (2000). Feeding in snakes. In
Feeding : Form, Function, and Evolution in Tetrapod Vertebrates (ed.
K. Schwenk), pp. 293–333. Academic Press, San Diego.
C, D. & I, F. J. (1989). The function of the
intramaxillary joint in the Round Island boa, Casarea
dussumieri. Journal of Zoology, London 217, 569–598.
C, D. & R, D. A. (1993). Cephalic anatomy of
the rare Indonesian snake Anomochilus weberi. Zoological Journal
of the Linnean Society 109, 235–273.
C, D. & S, J. D. (1995). Rhinokinetic snout of
thamnophiine snakes. Journal of Morphology 225, 31–50.
C, D., W, V. & R, D. A. (1993). The
systematic relationships of the snake genus Anomochilus.
Zoological Journal of the Linnean Society 109, 275–299.
394
D, H. G. & D, W. E. (1978). Systematic Herpetology : A Synopsis of Families and Higher Categories. HISS
Publications, New York.
D, H. G., H, C. A., H, S. B. & H, R.
(1996). Snake relationships revealed by slowly-evolving
proteins : a preliminary survey. Journal of Zoology, London 240,
1–28.
D! , A. (1806). Analytische Zoologie. Aus der FranzoW sischen, mit
ZusaW tzen, von L. F. Froriep. Verlag des Landes-IndustrieComptoirs, Weimar [not seen, cited in Rieppel 1988].
D! , A. (1853). Prodrome de la classification des reptiles
ophidiens. Memoires de l’AcadeT mie scientifique de Paris 23,
399–535.
D! , A. M. C. & B, G. (1844). ErpeT tologie generale ou
Histoire naturelle complete des Reptiles, Vol 6. Librairie
Encyclope! dique de Roret, Paris.
E, T. (2001). AutoCladeS version 1n0. Computer
Program ; available from www.bergianska.se\personal\
TorstenE\.
E, R. (1927). Studies in reptilian degeneration. Proceedings of
the Zoological Society of London 4, 879–945.
E, R., D Q, K. & G, J. A. (1988).
Phylogenetic relationships within Squamata. In Phylogenetic
Relationships of the Lizard Families (eds. R. Estes and G. Pregill),
pp. 119–281. Stanford University Press, Stanford, California.
E, R., F, T. H. & W, E. E. (1970). Studies
on the fossil snake Dinilysia patagonica Woodward : Part 1.
Cranial morphology. Bulletin of the Museum of Comparative
Zoology, Harvard 140, 25–74.
E, S. E. & B, J. (1998). An unusual lizard
(Reptilia : Squamata) from the Early Cretaceous of Las
Hoyas, Spain. Zoological Journal of the Linnean Society 124,
235–265.
F, J. (1985). Confidence limits on phylogenies : an
approach using the bootstrap. Evolution 39, 783–791.
F, T. H. (1959). Studies on the morphology and
function of the skull in the Boidae (Serpentes), Part 1. Cranial
differences between Python sebae and Epicrates cenchris. Bulletin
of the Museum of Comparative Zoology, Harvard 119, 453–472.
F, T. H. (1966). Studies on the morphology and
function of the skull in the Boidae (Serpentes), Part II.
Morphology and function of the jaw apparatus in Python sebae
and Python molurus. Journal of Morphology 118, 217–296.
F, T. H. (1970). Studies on the fossil snake Dinilysia
patagonica Woodward. II. Jaw machinery in the earliest
snakes. Forma et Functio 3, 205–221.
F, T. H. (1975). Pattern and instability in the evolving
premaxilla of boine snakes. American Zoologist 15, 469–481.
F, T. H. (1999). Adaptations and significance of the
cranial feeding apparatus of the sunbeam snake (Xenopeltis
unicolor) : Part 1. Anatomy of the skull. Journal of Morphology
239, 27–43.
G, C. (1978). The characteristics and affinities of the
Amphisbaenia. Transactions of the Zoological Society of London 34,
347–416.
G, J. P. (1966). Les rapports anatomiques du membre
pelvien vestigial chez les squamates serpentiformes. I. Anguis
fragilis (Anguidae, Lacertilia) et Python sebae (Boidae,
Ophidia). Bulletin du MuseT um national d’Histoire naturelle, Paris
(2) 37, 916–925.
G, J. P. (1976). Snake vertebrae – a mechanism or merely a
taxonomist’s toy ? Morphology and Biology of Reptiles (eds. A.
d’A. Bellairs and C. B. Cox), pp. 177–190.
Michael S. Y. Lee and John D. Scanlon
G, J., K, A. G. & R, T. (1988). Amniote
phylogeny and the importance of fossils. Cladistics 4, 105–209.
G, C. W. (1938). Fossil snakes of North America.
Geological Society of America Special Paper 9, 1–96.
G, C. & G, O. (1962). Introduction to Herpetology. W. H.
Freeman, San Francisco.
G, N., A, J. P. & R, A. G. (2000).
Likelihood-based tests of topologies in phylogenies. Systematic
Biology 49, 652–670.
G-K, C. (1892). Aigialosaurus, eine neue
Eidechse aus die Kreideschiefern der Insel Lesina mit
Ru$ cksicht auf die bereits beschriebenen Lacertiden von
Comen und Lesina. Glasnik huvatskoga naravoslovnoga drustva
(Societas historico-naturalis croatica) u Zagrebu 7, 74–106.
G, P. (2001). Radiation within the advanced snakes
(Caenophidia) with special emphasis on African opistoglyph
colubrids, based on mitochondrial sequence data. Biological
Journal of the Linnean Society 72, 99–114.
G, J. E. (1825). A synopsis of the genera of reptiles and
amphibians, with a description of some new species. Annals of
Philosophy, New Series 10, 193–217.
G, H. W. (1983). Dietary correlates of the origin and
radiation of snakes. American Zoologist 23, 431–441.
G, H. W. (1997). Snakes : The Evolution of Mystery in Nature.
University of California Press, Berkeley.
G, H. W. & B, G. M. (1978). Behaviour and
phylogeny : constriction in ancient and modern snakes. Science
200, 74–76.
G, H. W. & C, D. (2000). Limbless tetrapods and
snakes with legs. Science 287, 1939–1941.
G, A. E. (1985). The relationships of the lizard genera
Anelytropsis and Dibamus. Journal of Herpetology 19, 116–156.
G, A. E. (1997). The Biology and Evolution of Australian
Snakes. Surrey Beatty & Sons, Sydney.
G, B. C. (1979 a). Comments on the intermandibular muscles of snakes. Journal of Natural History 13, 477–498.
G, B. C. (1979 b). On the vomer in Acrochordidae
(Reptilia : Serpentes), and its cladistic significance. Journal of
Zoology, London 189, 559–567.
G, B. C. (1979 c). Variations in morphology of the
superficial palate of henophidian snakes and some possible
systematic implications. Journal of Natural History 13, 447–475.
H, G. (1930 a). U= ber das Kopfskelett und die Kaumuskulatur
der Typhlopiden und Glauconiiden. Zoologische JahrbuW cher,
Abteilung fuW r Anatomie 52, 1–94.
H, G. (1930 b). U= ber die Kaumuskulatur und die Scha$ delmechanik einiger Wu$ hlschlangen. Zoologische JahrbuW cher, Abteilung fuW r Anatomie 52, 95–218.
H, G. (1931 a). Die Kiefermuskulatur und die Scha$ delmechanik der Schlangen in vergleichender Darstellung.
Zoologische JahrbuW cher, Abteilung fuW r Anatomie 53, 127–198.
H, G. (1931 b). U= ber die Morphologie der Kiefermuskulatur
und die Scha$ delmechanik einiger Schlangen. Zoologische
JahrbuW cher, Abteilung fuW r Anatomie 54, 333–416.
H, G. (1964). Anatomical observations on the head of
Liotyphlops albirostris (Typhlopidae, Ophidia). Acta Zoologica
45, 1–62.
H, G. (1968). Anatomical observations on the head of
Anomalepis aspinosus (Typhlopidae, Ophidia). Acta Zoologica 49,
63–139.
H, G. (1973). Muscles of the jaws and associated structures
in the Rhynchocephalia and Squamata. In Biology of the
Snake phylogeny
Reptilia, Vol. 4 (eds. C. Gans and T. S. Parsons), pp. 285–490.
Academic Press, London.
H, G. (1979). On a new snakelike reptile from the lower
Cenomanian of Ein Jabrud, near Jerusalem. Bulletin du
MuseT um national d’Histoire naturelle, Paris (4) 1 C, 51–64.
H, G. (1980 a). Pachyrhachis problematicus Haas, snakelike
reptile from the lower Cenomanian : ventral view of the skull.
Bulletin du MuseT um national d’Histoire naturelle, Paris (4) 2,
87–104.
H, G. (1980 b). Remarks on a new ophiomorph reptile from
the lower Cenomanian of Ein Jabrud, Israel. In Aspects of
Vertebrate History (ed. L. L. Jacobs), pp. 177–192. Museum of
Northern Arizona Press, Flagstaff, Arizona.
H, R. W. (1969). Epiphyses and sesamoids. In Biology of the
Reptilia, Vol. 1 (eds. C. Gans, A. d’A. Bellairs and T. S.
Parsons), pp. 81–115. Academic Press, London and New
York.
H, J. (1998). The ethmoidal region of Dibamus
taylori (Squamata : Dibamidae), with a phylogenetic hypothesis on dibamid relationships within Squamata. Zoological
Journal of the Linnean Society 122, 385–426.
H, M. B., B, D. G., A, L. K. & C, P. T. (2000). Systematics of pythons of the Morelia
amethistina complex (Serpentes : Boidae) with the description
of three new species. Herpetological Monographs 14, 139–185.
H, P. J., M, L. R., D, H. G. & H, S. B.
(1995). Higher-level snake phylogeny inferred from mitochondrial DNA sequences of 12S rRNA and 16S rRNA genes.
Molecular Biology and Evolution 12, 259–265.
H, R. (1939). Contribution a l’e! tude des Elapidae
actuels et fossiles et de l’oste! ologie des ophidiens. Archives du
MuseT um d’Histoire naturelle de Lyon 15, 1–78.
H, R. (1955). Squamates de type moderne. In TraiteT
de PaleT ontologie, Vol. 5 (ed. J. Piveteau), pp. 606–662. Masson,
Paris.
H, R. (1959). Un serpent terrestre dans le Cretace!
infe! rieur du Sahara. Bulletin du SocieteT geT ologique de France 7,
897–902.
H, R. (1960). Un dentaire de Madtsoia (serpent
ge! ant du Pale! ocene de Patagonia). Bulletin du MuseT um national
d’Histoire naturelle, Paris (2) 31, 379–386.
H, R. (1961). Nouveaux restes d’un serpent boide!
(Madtsoia madagascariensis nov. sp.) dans le Cretace! supe! rieure
de Madagascar. Bulletin du MuseT um national d’Histoire naturelle,
Paris (2) 33, 152–160.
H, R. (1962). Revue des re! centes acquisitions
concernant l’histoire et la syste! matique des squamates.
Colloques internationaux du Centre national de la Recherche scientifique
104, 243–279.
H, R. (1968). [Review of ] A Contribution to the
Classification of Snakes, by Garth Underwood. Copeia 1968,
201–213.
H, R. & G, J. P. (1968). Observations sur le
squelette cervical et spe! cialement sur les hypapophyses des
sauriens varanoides actuels et fossiles. Bulletin du MuseT um
national d’Histoire naturelle, Paris (2) 39, 1028–1043.
H, R. & G, J. P. (1969). Vertebrae and ribs of
modern reptiles. In Biology of the Reptilia, Vol. 1 (eds. C. Gans,
A. d’A. Bellairs and T. S. Parsons), pp. 201–310. Academic
Press, London and New York.
H, R. & G, Y. (1965). Observations sur
l’osteologie et la classification des Acrochordidae (Serpentes).
395
Bulletin du MuseT um national d’Histoire naturelle, Paris (2) 36,
677–696.
H, R. & R, J. C. (1972). Les Erycinae fossiles de
France (Serpentes, Boidae). Compre! hension et histoire de la
sous-famille. Annales de PaleT ontologie (VerteT breT s) 58, 81–124, pls
1–2.
H, J. A. & C, G. R. (1992). A puzzling new snake
(Reptilia : Serpentes) from the Late Paleocene of Mississippi.
Annals of Carnegie Museum 61, 197–205.
I, N. (1990). Evolution of Complex Adaptations : Jaw
Apparatus of the Amphibians and Reptiles. Nauka, Moscow [In
Russian].
J, K. & F, T. H. (1997). Observations of a grooved
anterior fang in Psammodynastes pulverulentus : does the mock
viper resemble a protoelapid? Journal of Herpetology 30,
128–131.
J, W. (1906). U= ber Archaeophis proavus Mass., eine
Schlange aus dem Eoca$ n des Monte Bolca. BeitraW ge zur
PalaW ontologie und Geologie Oq sterreich-Ungarns und des Orients 19,
1–33, 2 pls.
K, A. M. (1966). On the process of rotation of the quadrate
cartilage in Ophidia. Anatomische Anzeiger 118, 87–90.
K, A. M. & H, H. G. (1965). The development
of the skull of Psammophis sibilans. I. The development of the
chondrocranium, II. The fully formed chondrocranium, III.
The osteocranium of a late embryo. Journal of Morphology 116,
197–310.
K, N. & B, E. L. (1996). Feeding by mandibular
raking in a snake. Nature 402, 369–370.
K, A. G. (1991). Boine snake phylogeny and research
cycles. Miscellaneous Publications of the Museum of Zoology,
University of Michigan 178, 1–58.
K, A. G. (1993 a). Aspidites and the phylogeny of pythonine
snakes. Records of the Australian Museum, Supplement 19, 1–77.
K, A. G. (1993 b). Calabaria and the phylogeny of erycine
snakes. Zoological Journal of the Linnean Society 107, 293–351.
K, A. (1901). Opetiosaurus bucchichi. Eine neue fossile
Eidechse aus der unteren Kreide von Lesina in Dalmatien.
Abhandlungen der k. k. geologischen Reichsanstalt 17 (5), 1–24,
3 pls.
K, F. & B, W. M. (1998). Phylogenetic relationships
of colubroid snakes based on mitochondrial DNA sequences.
Zoological Journal of the Linnean Society 122, 455–487.
L, D. A. (1968). The hyoid and its associated
muscles in snakes. Illinois Biological Monographs 38, 1–156.
L, M. S. Y. (1997). The phylogeny of varanoid lizards and
the affinities of snakes. Philosophical Transactions of the Royal
Society of London, Series B (Biological Sciences) 352, 53–91.
L, M. S. Y. (1998). Convergent evolution and character
correlation in burrowing reptiles : towards a resolution of
squamate phylogeny. Biological Journal of the Linnean Society 65,
369–453.
L, M. S. Y. (2000 a). Soft anatomy, diffuse homoplasy, and
the relationships of lizards and snakes. Zoologica Scripta 29,
101–130.
L, M. S. Y. (2000 b). Tree robustness and clade significance.
Systematic Biology 49, 829–836.
L, M. S. Y. (2001). Snake origins and the need for scientific
agreement on vernacular names. Paleobiology 27, 1–6.
L, M. S. Y. & C, M. W. (1998). Anatomy and
relationships of Pachyrhachis problematicus, a primitive snake
with hindlimbs. Philosophical Transactions of the Royal Society of
London, Series B (Biological Sciences) 353, 1521–1552.
396
L, M. S. Y. & C, M. W. (2000). Adriosaurus and the
affinities of mosasaurs, dolichosaurs, and snakes. Journal of
Paleontology 74, 915–937.
L, M. S. Y., B, G. L. J. & C, M. W. (1999 a).
The origin of snake feeding. Nature 74, 915–937.
L, M. S. Y., C, M. W. & S, J. D. (1999 b). A
second primitive marine snake : Pachyophis woodwardi from the
Cretaceous of Bosnia-Herzegovina. Journal of Zoology, London
248, 509–520.
L-S, T. (1995). Anatomy and functional morphology of the largest marine reptile known, Mosasaurus
hoffmani (Mosasauridae, Reptilia) from the Upper Cretaceous,
Upper Maastrichtian of The Netherlands. Philosophical Transactions of the Royal Society of London, Series B (Biological
Sciences) 347, 155–180.
L, J. C. (1966). Comparative osteology of the snake families
Typhlopidae and Leptotyphlopidae. Illinois Biological Monographs 36, 1–112.
M, D. R. & M, W. P. (2000). MacClade 4 :
Analysis of Phylogeny and Character Evolution. Sinauer Associates,
Sunderland, Massachusetts.
M, B. C. (1935). The subcentral foramina of the
Squamata. Current Science (Bangalore) 4, 320–322.
M, B. C. (1938). Some remarks on the phylogeny of the
Ophidia. Anatomischer Anzeiger 86, 321–368.
MD, S. B. (1967). Osteology of the Typhlopidae and
Leptotyphlopidae : a critical review. Copeia 1967, 686–692.
MD, S. B. (1972). The evolution of the tongue of snakes,
and its bearing on snake origins. Evolutionary Biology 6,
191–273.
MD, S. B. (1974). A catalogue of the snakes of New
Guinea and the Solomons, with special reference to those in
the Bernice P. Bishop Museum. Part I. Scolecophidia. Journal
of Herpetology 8, 1–57.
MD, S. B. (1975). A catalogue of the snakes of New
Guinea and the Solomons, with special reference to those in
the Bernice P. Bishop Museum. Part II. Anilioidea and
Pythoninae. Journal of Herpetology 9, 1–80.
MD, S. B. (1979). A catalogue of the snakes of New
Guinea and the Solomons, with special reference to those in
the Bernice P. Bishop Museum. Part III. Boinae and
Acrochordoidea (Reptilia, Serpentes). Journal of Herpetology
13, 1–92.
MD, S. B. (1987). Systematics. In Snakes : Ecology and
Evolutionary Biology (eds. R. A. Seigel, J. T. C. Collins and
S. S. Novak), pp. 1–50. MacMillan, New York.
MD, S. B. & B, C. M. (1954). The systematic
position of Lanthanotus and the affinities of the anguimorphan
lizards. Bulletin of the American Museum of Natural History 105,
1–154.
M, B. (1820). Versuch eines Systems der Amphibia. Johann
Christian Krieger, Marburg.
M$ , J. (1831). Beitra$ ge zur Anatomie und Naturgeschichte
der Amphibien. Zeitschrift fuW r Physiologie 4, 190–275.
N, N. A. (1986). A rational basis for a priori character
weighting. Systematic Zoology 35, 110–123.
N, F. (1903). U= ber die Varanusartigen Lacerten Istriens.
BeitraW ge zur PalaW ontologie und Geologie Oq sterreich-Ungarns und des
Orients 15, 31–42, 2 plates.
N, F. (1908). Zur Kenntnis der fossilen Eidechsen. BeitraW ge
zur PalaW ontologie und Geologie Oq sterreich-Ungarns und des Orients
21, 33–62, 1 pl.
Michael S. Y. Lee and John D. Scanlon
N, F. (1923). Eidolosaurus und Pachyophis. Zwei neue
Neocom-Reptilien. Palaeontographica 65, 99–154.
O, T. M. (1956). The anatomy of the head of Ctenosaura
pectinata (Iguanidae). Miscellaneous Publications of the Museum of
Zoology, University of Michigan 94, 1–122jfigs 1–59.
O, M. (1811). Die Ordnungen, Familien und Gattungen der
Reptilien, als Prodrom einer Naturgeschichte derselben. Joseph
Lindauer, Munich.
P, G. K., G, J. A. & G, H. W. (1986). The
evolution of helodermatid squamates, with description of a
new taxon and an overview of Varanoidea. Transactions of the
San Diego Society of Natural History 21, 167–202.
R, J. C. (1977). La position phyle! tique de Dinilysia patagonica,
serpent du Cretace! supe! rieur. Comptes rendus de l’AcadeT mie des
Sciences, Paris 284, 1765–1768.
R, J. C. (1982). La phyloge! nie des Le! pidosauriens (Reptilia) : une approche cladistique. Comptes rendus de l’AcadeT mie des
Sciences, Paris 294, 399–402.
R, J. C. (1984). Handbuch der PalaW oherpetologie, Vol. 11.
Serpentes. G. Fischer, Stuttgart.
R, J. C. (1987). Fossil History. In Snakes : Ecology and
Evolutionary Biology (eds. R. A. Seigel, J. T. C. Collins and
S. S. Novak), pp. 57–76. MacMillan, New York.
R, J. C. (1998). Fossil snakes from the Paleocene of Sa4 o Jose!
de Itaboraı! , Brazil. Part I. Madtsoiidae, Aniliidae. Palaeovertebrata 27 (3–4), 109–144.
R, J. C. & E! , F. (2000). Un nouveau serpent bipe' de
du Ce! nomanien (Cretace! ). Implications phyletiques. Comptes
rendus de l’Academie des Sciences, Paris, Sciences de la Terre et des
PlaneZ tes 330, 513–520.
R, J. C. & P, G. V. R. (1992). New snakes from the
late Cretaceous (Maastrichtian) of Naskal, India. Neues
Jahrbuch fuW r Geologie und PalaW ontologie Abhandlungen 187, 83–97.
R, J. C. & R, A. (1994). A snake from the Lower
Cretaceous (Barremian) of Spain : the oldest known snake.
Neues Jahrbuch fuW r Geologie und PalaW ontologie Monatshefte 1994,
561–565.
R, J. C. & W, C. (1999). Mid-Cretaceous (Cenomanian) snakes from Wadi Abu Hashim, Sudan : The earliest
snake assemblage. Palaeontologia Africana 35, 85–110.
R, O. (1977). Studies on the skull of the Henophidia
(Reptilia : Serpentes). Journal of Zoology, London 181, 145–173.
R, O. (1979 a). The braincase of Typhlops and Leptotyphlops
(Reptilia : Serpentes). Zoological Journal of the Linnean Society
65, 161–176.
R, O. (1979 b). The evolution of the basicranium in the
Henophidia (Reptilia : Serpentes). Zoological Journal of the
Linnean Society 66, 411–431.
R, O. (1979 c). The perilymphatic system of the skull of
Typhlops and Acrochordus, with comments on the origin of
snakes. Journal of Herpetology 14, 105–108.
R, O. (1980). The phylogeny of anguinomorph lizards.
Denkschriften der Schweizerischen Naturforschenden Gesellschaft 94,
1–86.
R, O. (1983). A comparison of the skull of Lanthanotus
borneensis (Reptilia : Varanoidea) with the skull of primitive
snakes. Zeitschrift fuW r zoologische Systematik und Evolutionsforschung
21, 142–153.
R, O. (1984). The cranial morphology of the fossorial
lizard genus Dibamus with a consideration of its phylogenetic
relationships. Journal of Zoology, London 204, 289–327.
R, O. (1988). A review of the origin of snakes. Evolutionary
Biology 22, 37–130.
Snake phylogeny
R, O. & Z, H. (2000 a). The braincases of mosasaurs
and Varanus, and the relationships of snakes. Zoological Journal
of the Linnean Society 129, 489–514.
R, O. & Z, H. (2000 b). The intramandibular joint
in squamates, and the phylogenetic relationships of the fossil
snake Pachyrhachis problematicus Haas. Fieldiana, Geology New
Series 1507, 1–69.
R, O. & Z, H. (2001). The development of the skull
in Acrochordus granulatus (Schneider) (Reptilia : Serpentes),
with special consideration of the otico-occipital complex.
Journal of Morphology 249, 252–266.
R, J. & S, H. M. (1966). The systematic position of the
group of snake genera allied to Anomalepis. Natural History
Miscellanea 184, 1–8.
R, A. S. (1956). Osteology of the Reptiles. University of
Chicago Press, Chicago.
R, D. A. (1967). Systematics and morphology of American mosasaurs (Reptilia, Sauria). Bulletin of the Peabody Museum
of Natural History 23, 1–237.
S, M. J. & D, M. J. (1996). The relationship
between homoplasy and confidence in a phylogenetic tree. In
Homoplasy : The Recurrence of Similarity in Evolution (eds. M. J.
Sanderson and L. Hufford), pp. 67–89. Academic Press, San
Diego.
S, A. H. (1983). Coadapted character complexes among
snakes : fossoriality, piscivory, and durophagy. American
Zoologist 23, 397–409.
S, J. D. (1992). A new large madtsoiid snake from the
Miocene of the Northern Territory. The Beagle, Records of the
Northern Territory Museum of Arts and Sciences 9, 49–60.
S, J. D. (1993). The phylogenetic position of the
Madtsoiidae (Serpentes). Second World Congress of Herpetology,
Adelaide, South Australia. Collected abstracts, 230–231.
S, J. D. (1996). Studies in the palaeontology and
systematics of Australian snakes. Ph.D. thesis, University of
New South Wales, Sydney. 648 pp.jAppendix.
S, J. D. (1997). Nanowana gen. nov., small madtsoiid
snakes from the Miocene of Riversleigh : sympatric species
with divergently specialised dentition. Memoirs of the Queensland
Museum 41, 393–412.
S, J. D. (2001). Montypythonoides : the Miocene snake
Morelia riversleighensis (Smith and Plane, 1985) and the
geographical origin of pythons Memoirs of the Association of
Australasian Palaeontologists 25, 1–35.
S, J. D. & L, M. S. Y. (2000). The Pleistocene serpent
Wonambi and the early evolution of snakes. Nature 403,
416–420.
S, J. D. & L, M. S. Y. (2002). On varanoid-like
dentition in primitive snakes. Journal of Herpetology 36,
100–106.
S, J. D. & S, R. (1988). Dentition and diet in
snakes : adaptations to oophagy in the Australian elapid genus
Simoselaps. Journal of Zoology, London 216, 519–528.
S, K. (1988). Comparative morphology of the lepidosaur tongue and its relevance to squamate phylogeny. In
Phylogenetic Relationships of the Lizard Families : Essays Commemorating Charles L. Camp (eds. R. Estes and G. Pregill),
pp. 569–598. Stanford University Press, Stanford, California.
S, K. (1993). The evolution of chemoreception in
squamate reptiles : a phylogenetic approach. Brain Behaviour
and Evolution 41, 124–137.
S, K. (2000). Feeding in lepidosaurs. In Feeding. Form,
397
Function and Evolution in Tetrapod Vertebrates (ed. K. Schwenk),
pp. 175–291. Academic Press, San Diego, California.
S, D. G. & N, R. G. (1973). The forebrain and
midbrain of some squamates and their bearing on the origin
of snakes. Journal of Morphology 140, 135–152.
S, G. G. (1933). A new fossil snake from the Notostylops
beds of Patagonia. Bulletin of the American Museum of Natural
History 67, 1–22.
S, J. B. (1993). ‘ Unordered ’ versus ‘ ordered ’ characters. Systematic Biology 42, 155–165.
S, J. B. (1994). A phylogenetic analysis of Bungarus
(Elapidae) based on morphological characters. Journal of
Herpetology 28, 440–446.
S, J. B. (1995). A phylogenetic analysis of the New
World Coral Snakes (Elapidae : Leptomicrurus, Micruroides, and
Micrurus) based on allozymic and morphological characters.
Journal of Herpetology 29, 325–338.
S, M. J. (1976). Small fossil vertebrates from Victoria Cave,
Naracoorte, South Australia. IV. Reptiles. Transactions of the
Royal Society of South Australia 100, 39–51.
S, M. (1999). TreeRot Version 2. Computer program and
documentation. Distributed by the author, Boston University.
S, S. M. (1979). Macroevolution : Pattern and Process. W. H.
Freeman, San Francisco and New York.
S, H. (1856). Zoonomie der Amphibien, 2nd ed. Von Veit,
Berlin.
S, D. L. (1999). PAUP* Version 4 – Phylogenetic Analysis
Using Parsimony (*and Other Methods). Computer program and
documentation. Sinauer, Sunderland, Massachusetts.
S, Z. & B$ , W. (1996). Redescription of Tropidonotus atavus von Meyer, 1855 from the upper Oligocene of Rott
(Germany) and its allocation to Rottophis gen. nov. (Serpentes,
Boidae). Palaeontographica Abteilung A 240, 145–161.
T, L. P. (1988). Structure of the skull in the sea snake
‘ Archaeophis ’ turkmenicus from the Lower Eocene of Turkmeniya. Paleontological Journal 1988, 73–79.
T, E., R, O., Z, H., P, M. J. &
J, L. L. (2000). A fossil snake with limbs. Science 287,
2010–2012.
T, A. R. (1983). Phylogenetic inference from restriction site endonuclease cleavage site maps with particular
reference to the humans and apes. Evolution 37, 221–244.
U, G. (1957). Lanthanotus and the anguinomorph
lizards : a critical review. Copeia 1957, 20–30.
U, G. (1967). A Contribution to the Classification of
Snakes. British Museum (Natural History), London.
U, G. (1970). The eye. In Biology of the Reptilia,
Vol. 2 (eds. C. Gans and T. S. Parsons), pp. 1–97. Academic
Press, London.
U, G. (1976). A systematic study of boid snakes. In
Morphology and Biology of Reptiles (eds. A. d’A. Bellairs and
C. B. Cox), pp. 151–175. Academic Press, London.
U, G. & K, E. (1993). On the affinities of
the burrowing asps Atractaspis (Serpentes : Atractaspididae).
Zoological Journal of the Linnean Society 107, 3–64.
U, G. & S, A. F. (1990). A classification of
pythons (Serpentes, Pythoninae). Journal of Zoology, London
221, 565–603.
V, R. H., R, D. A. & S, W. (1985). Observations of tooth surface morphology in snakes. Journal of
Herpetology 19, 20–26.
W, V. & G$ , R. (1998). Visceral anatomy of the
Michael S. Y. Lee and John D. Scanlon
398
Malaysian snake genus Xenophidion, including a cladistic
analysis and allocation to a new family (Serpentes : Xenophidiidae). Amphibia-Reptilia 19, 385–404.
W, V. & I, I. (1996). Redescription of a rare
Malagasy blind snake, Typhlops grandidieri Mocquard, with
placement in a new genus (Serpentes : Typhlopidae). Journal
of Herpetology 30, 367–376.
W, G. L. (1940). Ophthalmological implications for the
early history of snakes. Copeia 1940, 1–8.
W, J. J. (1998). The accuracy of methods for coding and
sampling higher-level taxa for phylogenetic analysis : a
simulation study. Systematic Biology 47, 397–413.
W, M. (2000). Untitled photograph of Haasiophis
terrasanctus. Trends in Ecology and Evolution 15, cover of issue 12.
W, M. (1992). Ordered versus unordered characters.
Cladistics 8, 375–385.
W, M. (1995). A comparison of two methods of
character construction. Cladistics 11, 297–308.
W, P. &  Q, A. (1996). Comparing behavioural and morphological characters as indicators of phylogeny. In Phylogenies and the Comparative Method in Animal
XII. APPENDIX 1. ABBREVIATIONS
Anatomical abbreviations
am fl
as pr
ba pr
bo
cc
com
cor
den
ec
ec pr
eo
fd
fjo
fr
jr
la duct
la for
lat
lat cr
lat ft
med ft
mx
na
nk
of
op
pa
pal pr m
pal pr p
par for
pbs
pm
pof
prf
pro
prz pr
anteromedial flange of maxilla
ascending process of premaxilla
basipterygoid process
basioccipital
crista circumfenestralis
compound bone of mandible
coronoid
dentary
ectopterygoid
ectopterygoid process
exoccipital
fovea dentis
fenestra for duct of Jacobson’s organ
frontal
juxtastapedial recess
lacrimal duct
lacrimal foramen
laterosphenoid
lateral crest of compound element
lateral foot process of prefrontal
medial foot process of prefrontal
maxilla
nasal
nasal keel of premaxilla
optic (II) foramen
opisthotic
parietal
palatine process of maxilla
palatal process of premaxilla
paracotylar foramen
parabasisphenoid (sphenoid)
premaxilla
postorbitofrontal
prefrontal
prootic
prezygapophysial process
Behavior (ed. E. P. Martins), pp. 206–233. Oxford University
Press, Oxford.
W, A. S. (1901). On some extinct reptiles from
Patagonia, of the genera Miolania, Dinilysia, and Genyodectes.
Proceedings of the Zoological Society of London 1901, 169–184.
Y, D. (1995). Groundplans and exemplars : paths to the
tree of life. Cladistics 11, 343–357.
Z, H. (1994). Les Tropidopheoidea (Serpentes : Alethinophidia) sont-ils re! ellement monophyletique ? Arguments en
faveur de leur polyphyletisme. Comptes Rendus de l’Academie des
Sciences, SeT rie III (Sciences de la Vie) 317, 471–478.
Z, H. (1998). The phylogenetic position of Pachyrhachis
within snakes (Squamata, Lepidosauria). Journal of Vertebrate
Paleontology 18, 1–3.
Z, H. & R, O. (1999). The phylogenetic relationships of Pachyrhachis problematicus, and the evolution of
limblessness in snakes (Lepidosauria, Squamata). Comptes
Rendus de l’Academie des Sciences, Sciences de la Terre et des PlaneZ tes
329, 831–837.
Z, H. & R, O. (2000). A brief history of snakes.
Herpetological Review 31, 73–76.
Snake phylogeny
pt
pz for
qa
sm
so
spl
spo
st
sta
sub for
vca
vcp
vo
V2
V3
VII
399
pterygoid
parazygantral foramen
quadrate
septomaxilla
supraoccipital
splenial
supraorbital
supratemporal
stapes
subcentral foramen
vidian canal (anterior opening)
vidian canal (posterior opening)
vomer
foramen for maxillary branch (V2) of trigeminal nerve
foramen for mandibular branch (V3) of trigeminal nerve
foramen for palatine (VII) nerve.
Institutional abbreviations
AMNH
AMS
ANWC
AR
BMNH
BSP
CAS
DGM
FMNH
FU
HUJ
MCSNT
MCZ
NHW
NMNH
NMV
NTM
QM
REE
RM
SAM
UMZC
WAM
American Museum of Natural History, New York.
Australian Museum, Sydney.
Australian National Wildlife Collection, Canberra.
Archer Reference collection, University of New South Wales, Sydney (and temporary registration
for QM palaeontological material).
The Natural History Museum, London.
Bayerische Staatssammlung fu$ r Pala$ ontologie und Historische Geologie, Munich (cast examined
in RM courtesy of R. Carroll).
California Academy of Sciences.
Departamento Nacional de Produc: a4 o Mineral, Rio de Janeiro, Brazil (specimens examined courtesy
of J. C. Rage).
Field Museum of Natural History, Chicago.
Flinders University of South Australia, Adelaide, Department of Zoology.
Hebrew University, Jerusalem.
Museo Civico di Storia Naturale, Trieste.
Museum of Comparative Zoology, Harvard University, Cambridge MA.
Naturhistoriches Museum, Wien.
National Museum of Natural History, Washington.
Museum of Victoria, Melbourne.
Northern Territory Museum of Arts and Sciences, Darwin.
Queensland Museum, Brisbane.
Richard Estes Collection, San Diego State University, San Diego.
Redpath Museum, McGill University, Montreal.
South Australian Museum, Adelaide.
University Museum of Zoology, Cambridge.
Western Australian Museum, Perth.
APPENDIX 2. SPECIMENS EXAMINED
j indicates fossil taxa.
Heloderma. Heloderma horridum (AMNH 56439,
57863, 57868, 64128, 71664, 118700, 11870 ; MCZ
5008, 5009, 5010, 13154). Heloderma suspectum
(AMNH 56432, 66998, 71082, 71864, 72646, 72908,
72999, 73771, 74777, 74778, 109521, 110174,
118698, 139670, 142627 ; UMZC R9319, R9320,
R9321).
Varanidae. Lanthanotus. Lanthanotus borneensis
(FMNH 134711).
Varanus. Varanus albigularis (BMNH 1974n2480).
Varanus bengalensis (BMNH 1930n1n10n2, 1974n2479).
Varanus exanthematicus (BMNH 1920n1n20n3660, 1974n
2480). Varanus giganteus (UMZC R9586, R9587).
Varanus gilleni (BMNH 1910n5n28n13). Varanus gouldii
(BMNH R1983n1132 ; WAM R28281, R70345).
Varanus griseus (BMNH 71n6n6n2, 1974n2481-83).
400
Varanus indicus (BMNH 1932n7n19n2).Varanus komodoensis (BMNH 1934n8n2n2, 1985n1226 ; NMNH
228163, 101444). Varanus mertensi (BMNH
1983n1132 – as V. gouldii). Varanus niloticus (BMNH
1970n183 ; UMZC R9551). Varanus panoptes
(AMS R100500 ; QM J48291). Varanus salvator
(BMNH 1972n2160-62). Varanus tristis (WAM
R106054). V. cf. tristis (QM J69781). Varanus varius
(BMNH 1987n2154 ; ANWC REPS13 ; QM J51467).
jMosasauroidea. jAigialosauridae. Aigialosaurus dalmaticus (BSP 1902II501). Opetiosaurus buccichi (NHW unnumberred specimen). Carsosaurus
marchesetti (MCSNT 11430-11432). jMosasauridae. Clidastes sp. (AMNH 192, 1548 ; BMNH
R2946 ; NMNH 3765, 3778, 11627, 11719). Globidens
sp. (NMNH 4993). Halisaurus sp. (NMNH 418442).
Liodon sp. (AMNH 1401). Mosasaurus sp. (AMNH
1391). Platecarpus sp. (AMNH 126, 127, 1488, 1491,
1820, 14788, 14800 ; BMNH R4002 ; NMNH 3774,
3791, 18274). Tylosaurus sp. (AMNH 1543, 4909 ;
BMNH R3616, R35622, R35625 ; NMNH 3764).
Amphisbaenia. Amphisbaena alba (MCZ 165208).
Bipes biporus (MCZ 141205, 145823, 83227 ; CAS
145152). Trogonophis weigmani (MCZ 60011). Anops
kingi (MCZ 43363). Rhineura floridana (MCZ 55615).
Dibamidae. Anelytropsis papillosus (AMNH
64023). Dibamus leucurus (MCZ 109854). Dibamus
novaeguineae (AMNH 86717).
jMadtsoiidae. Gigantophis garstini (BMNH
R8343, R3188, R3010, R8344). Madtsoia cf. bai
(BMNH R2976). Madtsoia camposi (DGM 1313a-c).
Nanowana godthelpi (QM F19741, F23066-7, F3137986 ; AR 10842). Nanowana schrenki (QM F23043,
F23051, F23082, F31391-92, F31394-95). Wonambi
barriei (QM F23038, F23047-48, F23077-78, F4018994). Wonambi naracoortensis (SAM P16166-68,
P26528, P27777, P30178, P31785, P31801 ; FU
1762). Yurlunggur camfieldensis (holotype partial skeleton, NTM P various numbers). Yurlunggur unnamed
spp. (QM F19740, F23036, F23041, F23046, F2304950, F23060, F30560, F36441 ; AR 6021, 10401,
10684, 10775, 11054, 12192, 13901, 14176 ; NMV
P186652).
jDinilysia. Dinilysia patagonica (BMNH
3154 :1 – cast ; MCZ unnumbered – 6 cast).
jHaasiophis. Haasiophis terrasanctus (HUJ EJ695).
jPachyrhachis. Pachyrhachis problematicus (HUJ
3659, 3775).
Leptotyphlopidae.
Leptotyphlops
blandfordii
(BMNH 69n8n28n58-61 ; as Glauconia). Leptotyphlops
conjunctus (MCZ 30065 ; BMNH 1954n1n11n84). Leptotyphlops dulcis (MCZ 39681). Leptotyphlops emini
Michael S. Y. Lee and John D. Scanlon
(FMNH 56374). Leptotyphlops humilis (AMNH 73716,
66170 ; REE 1089, 2034). Leptotyphlops macrolepis
(BMNH 1904n6n30n5). Leptotyphlops nigricans (AMNH
5156 ; MCZ 21473). Leptotyphlops scutifrons (BMNH
1962n550).
Typhlopidae. Acutotyphlops subocularis (MCZ
65997 – catalogued as Ramphotyphlops). Rhinotyphlops
schlegelli (REE1902 ; MCZ 38551). Ramphotyphlops
australis (AMS unnumbered). Ramphotyphlops lineatus
(FMNH 31371). Ramphotyphlops sp. (AMS R19116).
Rhinotyphlops acutus (BMNH 52n9n13n267, 1930n5n8n7
– catalogued as Typhlops). Rhinotyphlops lalandei
(BMNH 1930n5n8n8 – catalogued as Typhlops). Typhlops angolensis (AMNH 11633 ; FMNH 98952).
Typhlops bothriorhynchus (BMNH 60n3n19n1079). Typhlops diardii (BMNH 1930n5n8n1-5, 1933n9n12n1-n2).
Typhlops lineolatus (MCZ 131480). Typhlops lumbricalis (AMNH 73230). Typhlops punctatus (BMNH
1911n6n9n2, 1975n567 ; MCZ 7293, 48063). Typhlops
reticulatus (AMNH 3001).
Anomalepididae. Anomalepis aspinosus (MCZ
14785-cs). Helminthophis praeocularis (AMNH 38123cs). Liotyphlops wilderi (guentheri) (BMNH 87n12n29n8
(A)). Typhlophis squamosus (AMNH 25051-cs).
Anilius. Anilius scytale (AMNH 5613, 85980-82 ;
BMNH 58n8n23n48 ; REE 1326, ND43 ; FMNH
11175, 35688, 35683 ; MCZ 2984, 5478).
Cylindrophis. Cylindrophis ruffus (AMNH 85647 ;
BMNH 1930n5n8n47 ; FMNH 13100, 131780,
179033 ; MCZ 3699). Cyliindrophis lineatus (AMNH
12872 ; BMNH 1947n1n1n8). Cylindrophis maculatus
(FMNH 142395).
Uropeltidae. Brachyophidium rhodogaster (MCZ
18073).
Melanophidium
wynaudense
(BMNH
1930n5n8n126-127 ; MCZ 24739). Melanophidium punctatum (BMNH 1930n5n8n118). Platyplecturus madurensis
(BMNH 1930n5n8n112-116 ; MCZ 18046). Plecturus
(Platyplectrurus) perrotetii (MCZ 3875 – catalogued as
Platyplecturus). Pseudotyphlops philippinus (MCZ
34889). Plecturus (Pseudoplecturus) canaricus (BMNH
74n4n29n940-943). Rhinophis blythii (MCZ 4233).
Rhinophis philippinus (MCZ 18036). Rhinophis sanguineus (MCZ 3854). Uropeltis woodmasoni (BMNH
1930n5n8n75 – catalogued as Silybura nigra). Uropeltis
pulneyensis (BMNH 1930n5n8n72, vertebra, catalogued
as Uropeltis). Uropeltis ceylanica (AMNH 43343 ;
BMNH 1930n5n8n81-88, 1930n5n8n91a-b, 1930n5n8n9496, 1964n967 – some catalogued as Silybura brevis ;
MCZ 3852). Uropeltis smithi (BMNH 1930n5n8n5657 – as grandis). Uropeltis ocellatus (BMNH
1930n5n8n78-79 – as Silybura ; MCZ 3873).
Anomochilus. Anomochilus leonardi (BMNH
1952n1n2n63-a).
Snake phylogeny
Xenopeltis.
Xenopeltis
unicolor
(BMNH
1930n5n8n41, 1947n1n1n10, 1947n1n1n12 ; unnumbered
specimen labelled ‘ Xenopeltis concolor, Java S-d ’ ;
FMNH 11524, 53289, 138682, 122000, 131713,
178975 ; MCZ 3114 ; NMNH Field Herp 122782).
Loxocemus. Loxocemus bicolor (AMNH 19393,
44902, 71392, 110151, 141141 ; BMNH 82n8n7n16 ;
MCZ 5037, 27831).
Pythoninae. Aspidites ramsayi (AMS R132964-1 ;
QM J23629, J63378). Aspidites melanocephalus (QM
J5616, J30786). Liasis (Antaresia) childreni (AMS
R132962 ; SAM R26973). Liasis (Antaresia) maculosus
(QM J132961). Liasis (Antaresia) stimsoni orientalis
(QM J28416). Liasis (Apodora) papuana (AMS
R16488). Liasis (Bothrochilus) boa (AMS R132966).
Liasis (Leiopython) albertisii (AMS R16796). Liasis
fuscus (l mackloti) (AMS R41872 ; QM J268988).
Liasis olivaceus (AMS R132963 ; AR 8422, 9373,
9374). Morelia amethistina (QM J15815, J51148 ;
AMS R4908). Morelia oenpelliensis (AMS R93417).
Morelia spilota (QM J22191 ; AMS R132965). Morelia
viridis (QM J22455, J2108). Python sebae (BMNH
1964n1100, unnumbered mounted specimen). Python
molurus (AMS R366 ; BMNH 1940n4n28n1, 1972n
2168), Python reticulatus (BMNH 1964n1251, 1972n
2169).
Boinae. Acrantophis dumerili (BMNH 92n2n29n20 ;
NMNH 497683). Boa constrictor (BMNH 59n7n36n3536, 1952n1n2n60, 1964n1243 ; MCZ 19878). Candoia
aspera (AMNH 74992, 107142 ; FMNH 13915,
21731). Candoia bibroni (FMNH 22997). Candoia
carinata (FMNH 217065, 250097). Candoia sp.
(FMNH 250845). Corallus caninus (AMNH 57816).
Corallus enydris (BMNH 1904n1228, MCZ 42794).
Corallus hortulanus (AMNH 74832). Epicrates inornatus
(MCZ 37298). Epicrates cenchria (BMNH 62n6n13n1,
1964n1215, 1969n833). Eunectes murinus (BMNH
66n8n14n349 ; NMNH 220302). Sanzinia madagascariensis (NMNH 220313).
Erycinae. Calabaria (Charina) reinhardtii (FMNH
19478, 31372, 191123 ; MCZ 49104, 22501). Charina
bottae (AMNH 63487, 64945 ; NMNH 533578).
Charina (Lichanura) trivirgata roseofusca (MCZ 8966).
Charina (Lichanura) t. trivirgata (AMNH 141145,
75285). Eryx elegans (AMNH 143763). Eryx jaculus
(BMNH unnumbered). Eryx johni (AMNH 102222 ;
BMNH 1930n5n8n31 ; NMNH 348599). Eryx miliaris
(AMNH 143771, 143772). Eryx conicus (MCZ
150436).
Ungaliophiinae. Exiliboa placata (AMNH
102892) ; Ungaliophis panamensis (AMNH 62639,
401
76305 ; MCZ 56051). Ungaliophis continentalis
(AMNH 10293, 93813).
Tropidophiinae. Trachyboa gularis (AMNH
28982). Tropidophis canus (AMNH 73066, 45839).
Tropidophis haetianus (l maculatus) (FMNH 31341).
Tropidophis melanurus (AMNH 46690, 93002 ; MCZ
8965, 33499, 38173). Tropidophis pardalis (FMNH
233).
Bolyeriinae. Bolyeria multocarinata (BMNH
70n11n30n4A). Casarea dussumieri (BMNH 1992n99495, 70n11n30n4C ; MCZ 49135).
Acrochordus. Acrochordus arafurae (FMNH
250808 ; QM J51619). Acrochordus javanicus (BMNH
1964n970 ; FMNH 51712, 51711, 98780 ; REE 1943 ;
QM J23718). A. granulatus (FMNH 13182, 232795,
242177, 242178 ; REE 1985).
Colubroidea. Viperidae. Agkistrodon contortrix
(MCZ 50226). Bitis arietans (BMNH 1964n1249).
Bitis nasicornis (BMNH unnumbered). Causus maculatus (BMNH 1975n675). Causus rhombeatus (BMNH
unnumbered ; MCZ 49558, 49559). Cerastes cerastes
(BMNH 2n3n1a). Crotalus sp. (QM J28228, J28227).
Daboia russelli (BMNH 1974n5201 – catalogued as
Vipera). Lachesis muta (MCZ 50222). Macrovipera
lebetina (BMNH 1930n5n8n943 – catalogued as
Vipera). Trimeresurus trigonocephalus (BMNH 1972n22).
Vipera berus (BMNH 59n9n6n442, unnumbered). Colubridae. Achalinus meiguensis (FMNH 18777). Achalinus spinalis (FMNH 24896). Achalinus niger (NMNH
133989). Achalinus werneri (AMNH 67196 – catalogued as loochooensis). Boiga irregularis (QM J22173).
Cerberus australis (QM J23630). Clelia clelia (MCZ
31651). Coluber caspius (MCZ 37003). Dendrelaphis
punctulatus (QM J47449). Fimbrios klossi (FMNH
71698). Macropisthodon rudis (AMNH 34513 ;
FMNH 24952). Natrix natrix (BMNH 58n8n23n32,
1920n1n20n2729-30, 1930n5n8n167). Nerodia cyclopion
(BMNH 1969n2974 – catalogued as Natrix). Nerodia
erythrogaster (BMNH 1969n2956 – catalogued as
Natrix). Nerodia fasciata (MCZ 176472). Nerodia
sipedon (BMNH 1969n2945 – catalogued as Natrix).
Rhabdophis subminiatus (BMNH 1930n5n8n248). Sinonatrix aequifasciata (FMNH 24762). Sinonatrix annularis (FMNH 24766). Sinonatrix percarinata (FMNH
18751). Tropidonophis doriae (BMNH 1904n3n17n10 –
catalogued as Amphiesma). Xenochrophis piscator
(BMNH 1930n5n8n220, 1930n5n8n186 – catalogued as
Nerodia). Xenochrophis triangulifera (FMNH 128409,
98942, 210104). Xenodermus javanicus (FMNH 67427).
Xenodon rhabdocephalus (MCZ 38717). Xenodon severus
(MCZ 32750).