Annotated checklist of the recent and extinct pythons (Serpentes, Pythonidae), with notes on nomenclature, taxonomy, and distribution

Abstract McDiarmid et al. (1999) published the first part of their planned taxonomic catalog of the snakes of the world. Since then, several new python taxa have been described in both the scientific literature and non-peer-reviewed publications. This checklist evaluates the nomenclatural status of the names and discusses the taxonomic status of the new taxa, and aims to continue the work of McDiarmid et al. (1999) for the family Pythonidae, covering the period 1999 to 2010. Numerous new taxa are listed, and where appropriate recent synonymies are included and annotations are made. A checklist and a taxonomic identification key of valid taxa are provided.


Introduction
Pythons (family Pythonidae) represent a family of non-venomous basal snakes within the superfamily Pythonoidea Fitzinger, 1826(sensu Vidal et al. 2007, Vidal and Hedges 2009. Although present in Europe during the Miocene, and probably since the late Eocene (Szyndlar and Rage 2003), pythons are now restricted to the warmer regions of the Old World, ranging from Africa through South and Southeast Asia, Indo-Malaysia and New Guinea, to Australia (Kluge 1993, Scanlon 2001, Rawlings and Donnellan 2003, Rawlings et al. 2008). More than two thirds of the currently recognized extant species are found in the Australo-Papuan region (Kluge 1993, Scanlon 2001, Rawlings et al. 2008, where they have the greatest level of morphological and genetic diversity (Heads 2002, Rawlings andDonnellan 2003), and a high degree of endemism (Harvey et al. 2000, Rawlings et al. 2004). Whereas two Asian species (P. molurus and P. bivittatus) range north of the Tropic of Cancer, an African and at least seven Australian species extend their ranges south of the Tropic of Capricorn. Pythons occur in a variety of habitats, from desert and savanna, to subtropical and tropical rainforest (Kluge 1993) and into seasonally fl ooded grasslands and paddifi elds. Most species are terrestrial, some are arboreal (Kluge 1993) and a few are semi-aquatic. Th e 40 recognized extant species range in maximum adult length from 0.61 m to 10.0 m, and include the longest extant snake species.

Taxonomic changes since 1999
McDiarmid et al. (1999) has become a standard reference for snake taxonomists. Since then python systematics has received considerable attention as new phylogenetic and geographical evidence has become available. Aside from the descriptions of new genera, species, and subspecies (Table 1), the most noteworthy action was the split of the genus Python by Rawlings et al. (2008), placing two Asian taxa, reticulatus and timoriensis, into Broghammerus.
One author, the amateur herpetologist Raymond T. Hoser of Victoria, Australia, has caused considerable confusion in python taxonomy over the last decade by describing numerous taxa (6 new genera and subgenera, 4 new species, and 19 new subspecies) in the non-peer-reviewed literature without providing adequate descriptions for his proposed new taxa (for discussions see Aplin 1999, Wüster et al. 2001, Williams et al. 2006, 2008, Schleip 2008). Hoser rarely included important taxonomic information or data on scale counts, numbers of specimen examined, statistics, or the results of DNA analysis. Moreover, Hoser designated several types without ever having apparently examined them. Although not mandatory, the International Code of Zoological Nomenclature (ICZN 1999), hereafter termed the "Code", recommends that only specimens personally examined by the author should be designated as types (Recommendation 73B). Furthermore, Hoser himself (1996Hoser himself ( , 1997 considered failing to examine type specimens "sloppy taxonomy". Another recommendation (Recommendation 73C), which states which data should be provided with the holotype, is often not followed by this author. In general, inadequate descriptions inevitably lead to problems in clearly assigning specimens to established taxa, and are expensive and time-consuming for subsequent workers who have to re-examine the type material in order to make taxonomic decisions, instead of being able to rely on adequate original descriptions.
In general, the professional herpetological community has rarely accepted Hoser's taxa (Wüster et al. 2001, Aplin 2002, Williams et al. 2006, 2008, Schleip 2008, Zaher et al. 2009) unless one of his numerous names turns out to be valid and a senior synonym based on more exacting scientifi c work carried out by professional researchers, as was the case with Broghammerus.

Approach and Scope
Th e primary objective of this taxonomic checklist is to provide an overview of the taxa in the family Pythonidae, and to establish their nomenclatural status under the provisions of the Code and their current taxonomical status based on published works and knowledge. It is, however, beyond the scope of this list to propose re-classifi cations or re-arrangements of genera that lack fully resolved phylogenetic relationships. Although this checklist can only be a snapshot in time, it is intended to continue the work of McDiarmid et al. (1999) for the family Pythonidae over the past decade and provide updates to the list compiled by Henderson and Powell (2007). For taxa described during the past decade type species (for generic names) or type specimens (for specifi c names) are provided along with their type localities. Recently designated neotypes are also provided. Where new distributional information is available, this is included with the relevant citation. However, in contrast to the work of McDiarmid et al. (1999) and Henderson and Powell (2007), this checklist also contains extinct taxa.
Taxa are hierarchically arranged by indentation, and are presented in alphabetical order at the level of genera, species, and subspecies, although, in the case of sub-  species, the nominate subspecies precedes other subspecies, which then are listed in alphabetical order. Annotations are made directly below the relevant taxon, unless otherwise stated. Synonyms before the year 1999, and remarks on valid taxa, unless new data are available, can be found in McDiarmid et al. (1999). A key to the extent genera, species and subspecies recognized within the family of Pythonidae is provided in Appendix 2.

Interpretation and application of the Code
Th e Code rules on issues regarding nomenclatural acts and works, and aims to "provide the maximum universality and continuity in the scientifi c names of animals compatible with the freedom of scientists to classify animals according to taxonomic judgments" (ICZN 1999). Due to its universality, the wording of the Code leaves considerable room for interpretation. For the assessment of the nomenclatural status of published names, and for the purpose of nomenclatural stability, the Code was here strictly applied to all names. In any case of ambiguous wording, the authors have consulted the glossary of the Code as suggested in the Code's "Explanatory Note", and as stated in article 89. In the checklist we use the abbreviation "APP" (application): APP1. "Characters": To be available a name must "be accompanied by a description or defi nition that states in words characters that are purported to diff erentiate the taxon" (ICZN 1999: Art. 13.1.1). A description in the meaning of the Code is "a statement in words of taxonomic characters of a specimen or a taxon" (ICZN 1999: glossary entry for "description"), and a defi nition is "a statement in words that purports to give those characters which, in combination uniquely distinguish a taxon" (ICZN 1999: glossary entry for "defi nition"). Th e glossary defi nes the word taxon as a "taxonomic unit, whether named or not: i.e., a population, or group of populations of organisms which are usually inferred to be phylogenetically related and which have characters in common which diff erentiate (…) the unit (e.g., a geographic population, a genus, a family, an order) from other such units" (ICZN 1999: glossary entry for "taxon"). Th is latter statement clearly excludes distribution itself as a character to diff erentiate taxa and that complies with article 13.1.1, since it requires characters to diff erentiate a "geographic population" from other such units. Many taxonomists are likely to accept a geographic population, especially an insular population, only separated from other such populations by distribution, at subspecifi c rank. However, the Code does not distinguish between specifi c and subspecifi c rank in its requirements (Arts. 45.1, 45.2), and therefore subspecies must also be distinguishable by characters other than by their isolated locality or distribution.
APP2. "Generalized statements": Generalized statements such as "separated by distribution" or "separated by analysis of DNA" or relative statements such as "usually (but not always) has" do not constitute a character in the sense of article 13.1.1 (APP1). Analysis of DNA clearly describes a method although genomic diff erences are of diagnostic value, and distribution itself is not a character, as it is not intrinsic to any specimen within the taxon. Th erefore, these are not attributes of an organism (see glossary for character). Moreover, strictly following the glossary defi nition of the word description, the Code would require that a taxon must be uniquely distinguished from other taxa and generalized statements do not imply uniqueness.
APP3. "Priority": Article 23.3.5 requires the replacement of an unavailable name with the oldest available synonym (senior synonym).
APP4. "Incorrect subsequent spelling": Article 33.3 states that "any subsequent spelling of a name diff erent from the correct original spelling, other than a mandatory change or an emendation, is an "incorrect subsequent spelling"; it is not an available name and, like an incorrect original spelling (…), it does not enter into homonymy and cannot be used as a substitute name" (ICZN 1999). For speciesgroup names article 11.9.3.2 states that they are "deemed to have been published in combination with the correct original spelling of the generic name, even if it was actually published in combination with an emendation or incorrect spelling of the generic name" (ICZN 1999). Th erefore, incorrect subsequent spellings are corrected to the original spelling.
APP5. "Nomen dubium" (pl. nomina dubia): According to the glossary of the Code a nomen dubium is "a name of unknown or doubtful application" (ICZN 1999). Th is glossary defi nition leaves a wide scope for applying the term. A nomen dubium may be a lost type specimen or a type that lacks important diagnostic features so that a name cannot be applied to a specimen with clarity. Melville (1980Melville ( , 1984 noted that considering a name as nomen dubium is a matter of taxonomic decision and not a nomenclatural one. Moreover, Mones (1989) revealed that this term was fi rst used for a taxon which was accompanied by an insuffi cient description. He states that the term "(…) denotes ignorance, incapability to interpret the facts, insuffi cient diagnosis, or actual poorness of the type specimen" (Mones, 1989: 232). We agree with the above mentioned views and, hence, insuffi cient information on the holotype (Recommendations 72E, 73A, see Introduction) that obviously was randomly chosen from an online database of a natural history museum and was not examined by the author (Recommendation 73B) along with an insuffi cient diagnosis or defi nition of taxonomic characters (see Art. 13.1.1, Recommendation 13A) may make a name be considered a nomen dubium. However, the name remains available, and a subsequent revision or re-description of the taxon may establish its validity.
APP6. "Nomen nudum" (pl. nomina nuda): For generic names to be available, the Code requires "the fi xation of a type species in the original publication" (ICZN 1999: Art. 13.3). All names must be "explicitly indicated as intentionally new" (ICZN 1999: Art. 16.1). Generic names, as well as specifi c and subspecifi c names, to which no characters were provided that comply with article 13.1.1 (see APP1, APP2), are deemed a nomen nudum, and therefore considered unavailable.
APP8. "Unavailable name": A name is regarded as unavailable under the provisions of the Code, if either the requirements for publication or the requirements for availability are not met. Th is seems to be the case for names published by Hoser in his selfpublished Australasian Journal of Herpetology. Although the journal's website states that several hard copies were placed in libraries to comply with the Code, these authors were unable to locate hard copies from any major European or North American library, or obtain such from the publisher when fi rst issued (also see Recommendations 8B-D that to be regarded as published works they "must be obtainable, when fi rst issued (…)", and "must have been produced in an edition containing simultaneously obtainable copies by a method that assures numerous identical and durable copies". Neither requirements was fulfi lled. Later (20 May 2009), a colleague requesting original printed hard copies directly from the publisher only receiving single-sided, black and white versions of the online papers, printed on a domestic laser printer and bound by a large staple on the upper left hand corner (V. Wallach, pers. comm.). On his website, the publisher states "both print (fi rst print run) and online are identical including use of color". Th erefore, the hard copy received by our colleague was apparently "printed on demand". Article 9.7 states that "copies obtained on demand of an unpublished work [Art. 8], even if previously deposited in a library or other archive" do not constitute published work. Th e publisher disseminates the articles via the internet as PDFs downloadable from the journal's website, and appears to rely on the trust of subsequent workers, that paper copies do exist (e.g., Zaher et al. 2009). However, the dissemination of PDFs over the internet does not currently constitute "published works" (Art. 9.6). Since no hard copies of the relevant second issue (Hoser 2009) were obtainable when fi rst issued, and requested hard copies were "printed on demand", this work must be regarded as "not published" under the provisions of the Code (Arts. 8.1. 2, 8.1.3, 8.6, 9.6, 9.7) and the names therein are deemed unavailable (also see Wallach et al. 2009). Th e names, however, are listed for the completeness of the list but are not part of the formal synonymy. tio in this taxon "is generally at least 60% light colour to 40% or less darker blotches" (Hoser 2004), and by larger average size.
perthensis (Stull, 1932) Smith, 1985. Hoser (2000 separated this taxon from "other subspecies" by distribution (APP1, APP2), and from "other A. stimsoni" by color. Hoser (2000) cited without acknowledgment a statement made by Ehmann (1992) and quoted by Kend (1997: 148) and added to the statement, "the snout has a less box-like anterior when compared with other A. stimsoni" (Hoser 2000). However, since he considered A. stimsoni a synonym of "A. saxacola" (see comments on A. s. orientalis) and A. stimsoni stimsoni a separate subspecies, it is not clear what he means by A. stimsoni. Th is taxon is placed in the synonymy of A. stimsoni orientalis until further research has assessed its validity. Subsequent workers (e.g., Sonnemann 2007) have not recognized this taxon.
Type locality: Wyndham, Western Australia. Remarks: Hoser (2000) provided the same erroneous accession number for the holotype as was already provided by Smith (1985) in his original description; WAM R51208 is the number for a skink, Eremiascincus isolepis (fi de Mecke et al. 2009) (Doughty, pers. comm.). Hoser (2000) separated this taxon from the nominate form by lower loreal, subocular, and parietal scale counts (see Barker and Barker 1994: 1-2). Th e same is stated to be diagnostic for A. m. davieii, which makes them indistinguishable from each other, as already noted by Aplin (2002: 55-56) Barker and Barker [1994: 5] in support of this claim), color darkening above the eye in adults, and "from all other Womas by distribution" (Hoser 2000: 10) (APP1, APP2). Because of the vague description of this taxon, specimens cannot be unambiguously assigned to this taxon. Th e name is placed into the synonymy of A. ramsayi. For further comments see A. r. richardjonesii.
Type locality: near Port Hedland, Western Australia. Remarks: Aplin (2002) considered this taxon a nomen nudum. We disagree because Hoser (2000) provided characters that purport to diff erentiate this taxon from the "main race". Nevertheless, this taxon is indistinguishable from A. r. panoptes, as both taxa share the diagnostic characters and are only separated by "vast distance" (Hoser 2000) (APP1, APP2) (also see Wüster et al. 2001). Without further data, these taxa must be treated as synonyms, with A. r. panoptes having priority.

Genus Aspidoboa Sauvage, 1884 [synonym of Python]
Remarks: Hoser (2004) resurrected this genus to include the species of the Python curtus complex (sensu Keogh et al. 2001). As demonstrated by Rawlings et al. (2008) after exclusion of the taxa reticulatus and timoriensis (see Broghammerus), the genus Python forms a monophyletic grouping, including the taxon brongersmai. Since Keogh et al. (2001) demonstrated that brongersmai is the sister taxon to curtus and breitensteini, separating these three taxa from the genus Python would result in the non-monophyly of the genus. It is our opinion that the recognition of Aspidoboa at subgeneric rank only causes confusion and is unnecessary in a low-diversity genus as Python.

Synonyms:
Austroliasis Hoser, 2000 (incorrect subsequent spelling, APP4) Remarks: Hoser (2004) used the correct spelling rather than his earlier incorrect spelling of this taxon as "Austroliasis" (see below) but also included the species of the amethistinacomplex (sensu Harvey et al. 2000) and furthermore added timorensis (APP4) Peters, 1877. Nevertheless, Hoser only listed this genus without comment or evidence for its resurrection. Remarks: Hoser (2000) intending to resurrect Australiasis Wells and Wellington, 1984 created an incorrect subsequent spelling. Under the rules of the ICZN, this name is not an available name (Art. 33.3, ICZN 1999). See Australiasis.

Genus Bothrochilus Fitzinger, 1843
Remarks: Rawlings et al. (2008) identifi ed a sister-group relationship of this monotypic genus with Leiopython, which they also considered monotypic. Th ey proposed synonymy of Leiopython with Bothrochilus, with the latter being the senior synonym. Also see comments on Leiopython.  Pauwels et al. (2003). O' Shea and Lazell (2008) reported a specimen from Itbayat Island, Batanes Province, Philippines, the northeastern-most record for the taxon.
Remarks: Hoser (2004) describes the nominate form as "largish regional race with brownish head, much the same colour as the lighter dorsal body markings, although light-headed specimens are known and several colour variants and distinct colour mutations are also known".
Remarks: Hoser (2004) asserts that size and color separate this subspecies from the nominate subspecies. Th e statement "generally smaller race" is as unspecifi c as the statement made for the size of the nominate form (see remarks there). Th erefore, a diff erentiation of both based on size is nearly impossible. Hoser (2004) describes the color of this subspecies by stating "it rarely has a head lighter than the body as in some other variants of Broghammerus, such as those from Bali or parts of Th ailand" (Hoser 2004 Type locality: Nakhom Ratchasima, Central Th ailand. Remarks: Hoser (2004) separated this subspecies from the nominate subspecies by size, stating that this would be "a large race". However, the nominate form was also claimed to be "largish", hence, the former statement cannot separate this taxon from the nominate form. No further characters are provided to separate this taxon from other subspecies. Th e name is herein treated as nomen dubium (APP5) and is assigned to the synonymy of the nominate form. In accordance with articles 11.9.5 and 32.5.2.2, the name was corrected to comply with this principle. Th e author separates this taxon from the nominate form (referred to as "normal reticulatus") only by "larger average adult size" (Hoser 2004 Lang and Vogel (2006), and O' Shea (2007), but was overlooked by Henderson and Powell (2007) (Henderson 2009, pers. comm.). However, the relevant paper was cited in the list of references by the latter authors. Type locality: "Djamplong", South Timor, Indonesia. Th e MCZ online collection database provides the following information on the locality: "Djamplong, S Timor Indoaustralia, Indonesia, Timor Timur?, Nusa Tenggara".
Remarks: Hoser (2004) separated this subspecies from the nominate subspecies, referred to as "typical reticulatus", by color, stating that this taxon is "usually a brightly coloured subspecies" (Hoser 2004). However, the author clearly stated that several color variants are know within the nominate form as well. Hoser did not provide other characters that would indicate whether the specimen is assignable to this taxon or the nominate form. Th e name is assigned to the synonymy of B. r. reticulatus. Type locality: Selayar Island, Indonesia. Remarks: Although this taxon was recognized by subsequent workers (e.g., De Lang andVogel 2006, O'Shea 2007), it was overlooked by Henderson and Powell (2007) (Henderson 2009, pers. comm.). However, the latter authors cited the relevant work in the list of references. Type locality: Buitenzore (believed a misspelling of Buitenzorg, the Dutch colonial name for Bogor), Java, Indonesia.
Remarks: Hoser (2004) provided characters to separate this taxon, but he attempts to distinguish this subspecies from the species "Broghammerus reticulatus", which includes the subspecies itself. Th is taxon is indistinguishable from other subspecies based on the original description and is therefore treated as nomen dubium (APP5), and placed in the synonymy of the nominate form.
However, specimens from Bali, West Malaysia, Jaya, West Kalimantan, and Vietnam form a clade as demonstrated by Auliya et al. (2002).

timoriensis (Peters, 1876) Synonyms:
Austroliasis timorensis ( Remarks: Doubts were casted in literature that this species occurs on Timor (e.g., Barker and Barker 1996, McDiarmid et al. 1999, O'Shea 2007 because no records other than the type specimen are known from Timor, and this reported occurrence is likely incorrect.

Genus Chondropython Meyer, 1874 [synonym of Morelia]
Remarks: Hoser (2000) resurrected this genus for viridis, and later added two new subspecies viridis shireenae Hoser, 2004 and viridis adelynhoserae Hoser 2009 (APP8). Rawlings et al. (2008) demonstrated that two lineages ("northern" and "southern" lineage) along with the taxon M. carinata, form a subclade within the clade that represented Morelia. Although Chondropython is the oldest available name for this subclade, we see no advantage in resurrecting taxa as subgenera in a low-diversity genus.

azureus (Meyer, 1874) Synonyms:
Chondropython viridis (Schlegel, 1872) -Hoser, 2000 Chondropython viridis viridis ( Rawlings and Donnellan (2003) revealed the existence of a sibling species pair within the green tree python. Th e authors found a genetic divergence of about 7% in mitochondrial DNA (cytochrome b gene) between the northern and southern lineages, separated by the Central Mountain Range that extends in an east-west direction through New Guinea (also see comments on M. viridis). Rawlings and Donnellan (2003) revealed the existence of two species, one from north of the central cordillera, the other from the south, including the Aru Island and Australian populations. Nevertheless, within the southern lineage the Australian material formed a well supported clade whereas material from Aru Island clustered with that from Merauke and Timika. Th e authors state that "a determination of the species status of the northern and southern lineages awaits a more thorough assessment of divergence at nuclear genes based on wider geographic sampling than we could achieve herein with allozymes" (Rawlings and Donnellan 2003: 42). In 2008, Rawlings et al. (2008: 604) referred to the northern populations as the "unnamed sibling taxon of M. viridis". However, it is not yet evident that only a single taxon occurs on Aru Island, and that the published type locality for M. viridis is correct. Th e name azureus Meyer 1874 would be available for the northern linage, having its type locality on Biak Island, one of the localities from which specimens of "M. viridis N[orth]" were analyzed by Rawlings et al. (2008) and hence a strong candidate for the taxon name, based on priority. Since the types are presumed lost, we call for the designation of a neotype. Remarks: Rawlings and Donnellan (2003) revealed a genetic distance of about 3% in mitochondrial DNA (cytochrome b gene) between the Normanby Island specimen and all other specimens examined from the southern parts of New Guinea. However, this analysis was based on a single museum specimen from Normanby Island. Further research is needed to ascertain the taxonomic status of this population. However, this name is considered unavailable (APP8, also see introduction).
Type locality: Cape York, Queensland, Australia. Remarks: Hoser (2004) stated that the "white or other markings along the vertebra" are not diagnostic for this subspecies, but that "a very thin line or line of dots along the spine" is "generally a diagnostic trait for adults of this subspecies", although the author further states, that he had seen specimens with and without such markings. Furthermore, he noticed that "vertebral markings decline with age". With the latter comments, the author himself invalidated the utility of vertebral markings as a diagnostic character. In the absence of other characters, this taxon is apparently indistinguishable from the nominate form. Furthermore, as demonstrated by Rawlings and Donnellan (2003: 36), "all of the Australian haplotypes, which form a single lineage, are nested among the southern New Guinean haplotypes". We therefore placed this taxon in the synonymy of M. viridis (see comments there).

Synonyms:
Helionomous Gray, 1841 (nomen nudum) -Hoser, 2004 (nomen nudum APP6) Remarks: Th e type species for Heleionomus Gray, 1842 is H.variegatus [= Python natalensis]. Th e resurrection of the genus Heleionomus for Python sebae and P. natalensis is unwarranted because the actual status of natalensis and sebae has not been fully resolved and, furthermore, separation from Python would compromise monophyly of the genus Python. Rawlings et al. (2008) showed a sister-group relationship between sebae and molurus and that the genus Python (after exclusion of reticulatus and timoriensis) forms a monophyletic group. Th is genus is therefore assigned to the synonymy of Python.

Genus Helionomus Gray, 1842 [nomen nudum (APP6), incorrect subsequent spelling (APP4)]
Remarks: Hoser (2004) obviously intended to resurrect the genus Heleionomus Gray, 1842 but changed the name to "Helionomus". Th is constitutes an incorrect subsequent spelling (Art. 33.3). However, the name Helionomus was already used by Gray (1841) listed in the index for Boidae, but no species was assigned to this name, and it is therefore considered a nomen nudum. Also see Heleionomus.

Type species: Liasis fuscus Peters, 1873
Remarks: Hoser (2000) established this genus for the separation of the water pythons (L. fuscus and L. mackloti) from the olive pythons (L. olivaceus), both currently referred to Liasis. He distinguished the two genera by the number of mid-body rows, stating that "Liasis usually has over 60" (Hoser 2000) (APP2). Barker and Barker (1994: 35) provided a range of 58-63 mid-body scale rows for L. olivaceus barroni. According to Rawlings et al. (2004Rawlings et al. ( , 2008, the taxa fuscus and mackloti are closely related to each other and since L. mackloti Duméril and Bibron is the name-bearing type of Liasis (by subsequent designation [see Stimson and McDowell (1986) and Opinion 1514, ICZN, 1988]) and because Gray (1849) (2000) for the Torres Strait islands and New Guinean populations. However, Rawlings et al. (2004) demonstrated that specimens from Queensland, the Torres Strait islands (Saibai), and New Guinea form a well-supported clade, which was considered the sister group to the clade comprising the Northern Territory and Indonesian populations. Since Queensland is the type locality of L. fuscus Peters, the resurrection of this name is unwarranted as it is a junior synonym to L. fuscus. Th e name cornwallisius is therefore placed into the synonymy of L. fuscus. fuscus jackyae (Hoser, 2004) [nomen dubium, synonym of L. fuscus] Holotype: WAM R13882.
Type locality: Kalumburu, Western Australia. Remarks: Hoser (2004) claimed that several diagnostic features separated this taxon from others, but discusses only one (APP2); he stated that "in Katrinus fuscus fuscus (from coastal Queensland) the upper lips are pale with a little brown peppering. However, in K. fuscus jackyae (from the NT and WA) the lips are usually darker with more dark brown peppering or even blotches" (Hoser 2004). He continued that this subspecies would intergrade with K. f. cornwallisius around the Gulf of Carpentaria. Th e name is herein considered a nomen dubium (APP5).

Genus Leiopython Hubrecht, 1879
Remarks: Recent studies revealed that Bothrochilus and Leiopython form a clade. Th us, since Rawlings et al. (2008) considered both genera monotypic, they proposed "the use of a single generic name (Bothrochilus) for this species pair" (Rawlings et al. 2008: 613). Later, Schleip (2008) demonstrated that this genus is not monotypic. Rawlings et al. (2008) had used material from L. hoserae for their genetic analysis (GeneBank accession number U69835, Western Province, PNG at Mawatta). Until further molecular genetic data clarify the relationships of the taxa involved, and in deference to nomenclatural stability, we are reluctant to synonymize Leiopython with Bothrochilus.
Remarks: Henderson and Powell (2007) listed only Leiopython albertisii Peters & Doria, 1878. Hoser (2000 incorrectly ascribed albertisii to Gray 1842. Th e taxon was named in honor of Italian naturalist Luigi Maria D'Albertis, who made a name for himself in New Guinea. D'Albertis was only a few months old in 1842 and would, therefore, have been an unlikely recipient of Gray's dedication. Furthermore, Hoser repeatedly used an incorrect spelling for the species albertisii by omitting the terminal -i (APP4, article 33.4).

Synonyms:
Leiopython albertisi barkerorum Hoser, 2009 (APP4, APP8, see introduction) Remarks: Hoser (2000) diff erentiated this subspecies only by remote distribution (APP2). Other characters mentioned by Hoser (2000) were said to overlap with the nominate form. Wüster et al. (2001) and Schleip (2008) therefore considered the name a nomen nudum. Furthermore, since the name honours two persons, it should have been suffi xed with -orum. In 2009 Hoser re-described this taxon with the name emended to albertisi barkerorum. However, the name is considered not published under the Code (APP8).  Brongersma, 1953. Remarks: Hoser (2000 established this monotypic genus for Morelia boeleni stating "while the Boelen's Python (boeleni) has close affi nities with the carpet pythons, there is no evidence before this author to suggest that the relationship is any closer than that between the Green (viridis) and carpets. Th us if viridis is entitled to be placed in a separate genus to the carpets, so too should be boeleni" (Hoser 2000: 21-22). Rawlings et al. (2008) supported monophyly of the genus Morelia including M. boeleni. Lenhoserus would therefore only be a subgenus within Morelia. Other authors have not adopted this name, and, in adherence to nomenclatural stability, we regard Lenhoserus as a subjective junior synonym of Morelia (see comments there).

Genus Liasis Gray, 1842
Synonyms: Katrinus Hoser, 2000 Remarks: Scanlon and Mackness (2002) considered the gender of Liasis Gray feminine because Gray's (1842) original use of the combination Liasis olivacea implied it to be feminine. However, Gray (1842) also used the masculine gender for Liasis amethystinus [=Morelia amethistina (Schneider)] within Liasis. Hence, Gray did not clearly indicate his intentions concerning the gender of Liasis. Despite this discordance in gender, the ICZN had used it plenary power (Art. 81.1, ICZN, 1999) to fi x a type species for Liasis (Opinion 1514, ICZN 1988, and additionally (but perhaps not deliberately) fi xed the gender as masculine. Th e name and gender also entered the List of Available Names in Zoology (also see Art. 80.6, 80.7, ICZN

Genus Morelia Gray, 1842
Synonyms: Lenhoserus Hoser, 2000Chondropython Meyer, 1874-Hoser, 2000Nyctophilopython Wells & Wellington, 1985-Hoser, 2000Montypythonoides Smith & Plane, 1985-Scanlon, 2001 Australiasis Wells & Wellington, 1984-Hoser, 2004 Remarks: Hoser (2000) proposed the splitting of this genus into several genera. He created a new genus, Lenhoserus Hoser (see comments there) (for M. boeleni), and resurrected Australiasis Wells & Wellington (for M. amethistina and B. timoriensis), but created an unavailable name (APP6) ("Austroliasis") by incorrect subsequent spelling (APP4). Later, Hoser (2004) used the correct spelling Australiasis Wells & Wellington, added the species recognized by Harvey et al. (2000) and additionally resurrected the taxon duceboracensis Günther 1879 (see comments there). Furthermore, he resurrected Chondropython Meyer 1874 (for M. viridis) and Nyctophilopython Wells & Wellington (for M. oenpelliensis). However, phylogenetic studies (Rawlings et al. 2008) revealed that this taxonomic action is unwarranted. Although Rawlings et al.'s (2008) maximum parsimony analysis showed Morelia to be diphyletic (but monophyletic in Bayesian analysis), the separation of the amethistina-complex (sensu Harvey et al. 2000) and of oenpelliensis from the bredli/ spilota-clade would in any case be unwarranted and would nullify the monophyly of this grouping. Th e resurrection of Chondropython would only be warranted at subgeneric rank with the inclusion of the two recognized lineages of the green tree python (sensu Rawlings and Donnellan, 2003) and of M. carinata. However, Rawlings and Donnellan (2003) and Rawlings et al. (2008) avoided such placement because the phylogeny was not fully resolved (see comments for Chondropython azureus) Th e placement of M. boeleni as a separate monotypic genus is also unwarranted. We do not see any value in dividing such a small genus, and in the interests of nomenclatural stability, we place Australiasis, Lenhoserus, Chondropython, and Nyctophilopython in the synonymy of Morelia.

amethistina (Schneider, 1801) Synonyms:
Austroliasis amethistina ( Remarks: Harvey et al. (2000) identifi ed three races within the species, two from the mainland of New Guinea, separated by the Central Mountain Range (also see remarks on M. clarki), and another race from New Ireland (see remarks on M. duceboracensis). Th is is consistent with other species found in this region (e.g., the two lineages of the green tree python (sensu Rawlings and Donnelan, 2003), and L. albertisii/L. hoserae). According to Harvey et al. (2000), the holotype of A. amethistina is lost. We call for the designation of a neotype.

riversleighensis (Smith & Plane, 1985) -Scanlon, 2001
Remarks: Smith and Plane (1985) documented signifi cantly lesser curvature in the teeth of this taxon, to that found in species of the genera Python and Morelia, and because "…of the slight curvature of the dentary teeth…" (Smith and Plane 1985: 194) the authors considered this taxon more closely related to Morelia than to Python.
bredli (Gow, 1981) Remarks: Fyfe (2007)  kinghorni Stull, 1933 Distribution: For range extensions in Queensland see Augusteyn (2004) and Fearn and Trembath (2006).  Schwaner et al. 1988) on the ground of "higher incidence of scale anomalies" to the ventral scales. It can be argued that anomalies do not make good diagnostic characters, and these anomalies were already described in detail by Schwaner et al. (1988). Hoser (2004) further claims that this taxon may be distinguished from M. mippughae "by having lanceolate-shaped dorsal scales as opposed to more rhomboidal-shaped dorsal scales" (also see comments on M. mippughae). According to Schwaner et al. (1988: 15), and in support of Smith (1981), "specimens of imbricata have distinctly elongated, lanceolate-shaped, posterior dorsal scales. M. s. variegata usually have the rhomboidal condition". Furthermore, Schwaner et al. (1988) also reported that specimens from the St. Francis Island exhibited reduced ventral and subcaudal scale counts and a shorter tail than specimens from other populations. Hoser (2004) stated that this taxon is distinguishable from "all other Morelia by colouration and patterns" (Hoser 2004), but contradicted this statement when stating that this taxon is "highly variable in individual colouration and pattern", and that "this species cannot be defi nitively separated from other Morelia on the basis of scalation alone as these properties (ventral counts and the like) may overlap with other Morelia" (Hoser 2004). Based on this description, it is unlikely that specimens can be correctly assigned to this species unless they were known to originate from the type locality. Schwaner et al. (1988: 14) and Pearson et al. (2002) assigned the St. Francis Island population to the subspecies M. s. imbricata. We concur with this and relegate this taxon to the synonymy of M. s. imbricata. Mense (2006), Henderson and Powell (2007), and Swan (2007) did not list this taxon. Type locality: Iron Duchess, Middleback Ranges, South Australia.
Remarks: Hoser (2004) separated this taxon from its relative M. macburniei "by a lower incidence of scale anomalies" (Hoser 2004) of the ventral scales. Th is is meaningless because most populations will show few anonomalies, hence, using the "normal state" as a character does not diff erentiate this taxon from others. Hoser (2004) continued that this taxon has "more rhomboidal-shaped dorsal scales as opposed to having lanceolate-shaped dorsal scales" (Hoser 2004), which is, according to Schwaner et al. (1988) also true for M. s. variegata (also see remarks on M. macburniei). It is further separated from its closest relative M. s. metcalfei by its color pattern and from all other Morelia by coloration and patterning. Hoser claimed that "a suite of characteristics" separate this taxon from its closest relatives M. macburniei and M. metcalfei, but failed to enumerate characters other than those cited above. Hence, the name cannot clearly be assigned to a specimen and this name is therefore considered a nomen dubium (APP5). Mense (2006), Henderson and Powell (2007), and Swan (2007) Tedford's (1967) Site B, Riversleigh, northwestern Queensland, Australia. Late Oligocene -early middle Miocene (Scanlon 2001).
Remarks: Smith and Plane (1985) described the two extinct species riversleighensis and antiquus from Australia. Kluge (1993) synonymized antiqua (name amended for gender by Scanlon 1992) with olivaceus Gray due to the lack of auta-pomorphies and great overall similarity and riversleighensis with spilota Lacépède. Scanlon (2001) synonymized antiqua with riversleighensis.

spilota (Lacépède, 1804) Synonyms:
Morelia riversleighensis (Smith & Plane, 1985) -Kluge, 1993 Remarks: Hoser (2000Hoser ( , 2004 listed several subspecies of M. spilota at specifi c rank, without comment. Since no new evidence is available, these authors continue to treat them all as subspecies. Th ese authors also treat the taxon M. harrisoni described by Hoser (2000) Hoser (2000: 24) described this taxon at specifi c rank but considered it "similar in most respects to the others in the genus Morelia" separating it from M. s. cheynei, M. s. variegata, and M. s. mcdowelli "by distribution" (APP1, APP2), and further stating that specimens of this taxon "tend to have a lower average ventral and subcaudal scale count than Morelia cheynei, Morelia variegata and Morelia macdowelli, however the sample seen is too small to conclude if this trend is general" (Hoser 2000: 25). Hoser's concept of this taxon comprises several populations throughout New Guinea. Th e author referred to Barker and Barker (1999) for further diagnostic characters. Barker and Barker identifi ed several diff erent and distinct populations from New Guinea, which Hoser (2000) placed within this catch-all taxon. For the "Port Moresby" population Barker and Barker (1999) stated that they "exhibit some characteristics of both variegata and mcdowelli (…). Most Port Moresby carpets have longitudinally expanded lateral pale blotches and bold facial stripes from the eye to the nasal scale, as do mcdowelli in northern Queensland. Th e patterns on the tops of the heads are similar to variegata". For the "Irian Jaya" (now West Papuan or Papuan) population they stated that "[a]t 2 and 3 years of age, some are even as black and gold as M. s. cheynei", but did not provide further information on the "Trans-fl y" (PNG) or the "Northern New Guinea" populations. Hoser (2000: 25) fi nally stated that "Morelia harrisoni can best be defi nitively separated from the other species of Morelia by DNA analysis" (APP2). To the authors's best knowledge, no such analysis has been carried out. Since the diagnostic characters provided by Hoser (2000) and by Barker and Barker (1999) overlap with those for other taxa of the Morelia spilota complex, this taxon is likely to be confused with them. We consider this taxon a subspecies inquirenda (APP7). Henderson and Powell (2007) did not recognize this taxon. Mense (2006) discussed this taxon as a subspecies of M. spilota, and O' Shea (2007: 134) wrote: "Papuan Carpet Python Morelia spilota ssp. Th e status of all New Guinea Carpet Pythons is controversial (…). Th e New Guinea populations are fragmented and isolated, and their taxonomic status and relationships have yet to be determined with certainty". Until further research has established otherwise, these authors treat this taxon as a subspecies of M. spilota, as proposed by Mense (2006) and Flagle and Stoops (2009 Wells andWellington (1984, 1985), this name comprised all the taxa now recognized at subspecifi c rank, excluding the nominate subspecies and M. s. imbricata, but including the New Guinean populations. Now M. s. variegata is taxonomically restricted to Northern Australia (Kend 1997, Mense 2006 Wilson and Heinsohn (2007) Remarks: Rawlings and Donnellan (2003) found molecular evidence for cryptic diversity within M. viridis, resulting in two genetically distinct races. Th e type locality for M. viridis is Aru Island, which applies to the "southern lineage", including Australian specimens (Rawlings and Donnellan 2003) (also see Chondropython azureus). However, Rawlings and Donnellan (2003: 42) noted that "(…) the east/west limits of the distribution of the two lineages may not necessarily be at the extreme ends of the central cordillera or the island", and hence, there may be even more lineages present. Finally, due to the absence of molecular genetic data from the holotype, the type locality Aru Island cannot defi nitely be confi rmed.  Head (2005) reported remains of an indeterminate python from Miocene-age strata of the Siwalik Group of Pakistan. From the known distribution of extant species, this is likely to be a species of Python.

Genus
Remarks: Hoser (2004) split this genus into several genera, e.g., Aspidoboa Sauvage (for breitensteini, brongersmai, and curtus), Broghammerus Hoser (for reticulatus), and Shireenhoserus Hoser (for anchietae and regius). However, Hoser (2004) overlooked Enygrus Wagler, 1830 (also see McDowell 1979: 9-10, 28), which makes Shireenhoserus a subjective junior synonym of Enygrus Wagler. He further intended to resurrect Heleionomus Gray, 1842 (for sebae and natalensis) but spelt the name as "Helionomus". Th is constitutes an incorrect subsequent spelling (APP4), although the name Helionomus was already listed in Gray 1841 but is considered a nomen nudum (see remarks for Heleionomus). Only molurus and bivittatus would have remained within Python. Evidence from genetic studies reveal that with the exception of reticulatus and timoriensis, which were placed into Broghammerus (see comments there) by Rawlings et al. (2008), no further splitting of the clade Python is indicated. Furthermore, the phylogenetic relationships of several species (e.g., regius and anchietae, molurus and bivittatus, and sebae and natalensis) have not been fully resolved (e.g., Douglas et al. 2010: fi g. 4-6). Other groups (e.g. the curtus-group sensu lato) are currently under study.  Shea (1998), for distribution in Asia see Pauwels et al. (2003), Barker (2008, 2010). Barker and Barker (2010) considered records of the occurrence of bivittatus in the Sichuan Province deviant due to complete isolation from the natural range of bivittatus and therefore excluded the province from the range of occurrence. Records from Sumatra and Borneo are believed to be incorrectly identifi ed (Haile 1958, Groombridge andLuxmoore 1991). Remarks: Jacobs et al. (2009) considered this taxon a valid species. Evidence for this placement was already provided by O' Shea (1998Shea ( , 2007 and Barker and Barker (2008) who pointed out that isolated populations of bivittatus do exist within the distributional range of molurus along the southern Nepalese border and in north-east India as reported from Assam by O' Shea (2007). Jacobs et al. (2009) primarily referred to Barker and Barker (2008) when stating that the isolated populations appear to exist not only sympatrically but syntopically with molurus but maintain their own integrity by avoiding interbreeding. However, O'Shea (pers. obs.) has observed the species inhabiting diff erent habitats. P. molurus appears to occur in dry sandy woodland whereas bivittatus prefers riverine forests and fl ooded grasslands. O'Shea had not observed the two species occurring sympatrically or syntopically. Jacobs et al. (2009: 12) stated that de Rooij (1917) had assumed the type locality of Kuhl's (1820) concept of bivittatus, which was based on unverifi ed pictures by Seba, to be in Indochina rather than in the Sundaland and that the populations occurring between China and Java may be considered P. molurus sondaica (sic) Werner, 1899. Nevertheless, according to Jacobson et al (2009), Mertens (1930 fi xed the type locality to Java without the designation of a neotype, which has led to nomenclatural problems. Mertens (1930) as well as Werner (1909Werner ( , 1930 and Pope (1935) assumed that Schlegel (1837) rather than Kuhl (1820) had introduced the name bivittatus. According to Jacobs et al. (2009), Mertens (1930 was aware that Schlegel's (1837) composite concept of P. bivittatus included several python taxa, namely those from India (P. molurus) and from Africa (P. sebae), respectively.  -Constable, 1949 Python molurus pimbura -Deraniyagala, 1955 Python molurus molurus -Stimson, 1969 Distribution: First reported from Nunavil (Th enmarachi), Jaff na Peninsula, Sri Lanka by Abyerami and Sivashanthini (2008).
Remarks: Hoser (2004) resurrected this taxon from the synonymy of P. m. molurus without providing reasons for this action. Deraniyagala (1945) separated the subspecies from P. m. molurus based on lower subcaudal scale counts and the irregular shape of the lateral markings. Dorsal midbody scale rows and ventral scale counts overlap those of the nominate subspecies. Constable (1949: 124) did not follow this placement and synonymized this taxon with the nominate subspecies, which was followed by Stimson (1969). A second paper by Deraniyagala (1955: 6) provided a more detailed description of the subspecies. Th erein, he stated that this taxon is also separated from the nominate form "in generally possessing three preoculars instead of two" or four as stated by Wall (1921: 47) for some Indian populations of the nominate form. Th ere appears to be a range in preocular scale counts across India, from three in the northeast, to four in the northcenter, and two in northwest (O'Shea pers. obs.) but this data, from only a few specimens, requires further verifi cation. Contrary to his fi ndings in 1945, Deraniyagala (1955) reports this taxon to have "more subcaudals" than the nominate form, obviously a typographic error according to the scale count data provided therein. It seems likely that subsequent workers overlooked this latter work, since neither Stimson (1969) nor McDiarmid et al. (1999 or Henderson and Powell (2007) cited it. Several subsequent workers accepted the placement to the synonymy of the nominate form, but no further studies have been conducted on the molurus-complex. However, besides the lower subcaudal scale counts and the higher number of preoculars, the pink surface of the head may also constitite a morphological diff erence. Boulenger (1890Boulenger ( , 1893 and MA Smith (1943) recorded two preoculars for Python molurus, while Wall (1921) records three preoculars for specimens from Ceylon. Since Sri Lanka is a known biodiversity hot spot with a high level of endemism, this allopatric population may represent a cryptic species. Because of the evidence provided by Deraniyagala (1955), these authors tentatively list this taxon as a valid subspecies and call for further research regarding its true status (APP7).

natalensis A Smith, 1840 Synonyms:
Python natalensis A Smith -Broadley, 1999 Helionomus natalensis (A Smith) -Hoser, 2004 (nomen nudum, also see remarks on Python) Python natalensis A Smith -Henderson and Powell, 2007 Distribution: Notes on the distribution of this species can be found in Alexander (2007).
Remarks: McDiarmid et al. (1999) refer to A. Smith 1833. According to Branch and Bauer (2005), the name "Python Natalensis" already appeared in A. Smith (1833) as well as in A. Smith (1838) but without a description. Th e name appeared again in A. Smith (1840), but this time was accompanied by a plate. Gray (1842) also cites A. Smith (1840) as do Branch and Bauer (2005). Elevated to specifi c rank by Broadley (1999).
Remarks: Szyndlar and Rage (2003: 72-73) considered this name a nomen dubium as it is indistinguishable from other (extinct) Python.

Genus Rawlingspython Hoser, 2009 [unavailable name (APP8)]
Type species: Liasis perthensis Stull, 1932 Remarks: Hoser (2009) Bocage, 1887 Remarks: Hoser (2004) established this genus for the smaller African taxa P. anchietae and P. regius. Hoser (2004) overlooked the older name Enygrus Wagler, 1830 (see remarks under Python), relegating Shireenhoserus as a junior synonym. Moreover, after relocation of the two Asian taxa reticulatus and timoriensis the genus Python now forms a clade, including P. regius. Th e phylogenetic relationship between P. regius and P. anchietae has not yet been examined and separation would result in non-monophyly. Hence, the recognition of this genus is unwarranted and it is assigned to the synonymy of Python.

Discussion and Conclusion
In taxonomy, there have always been "lumpers" and "splitters", but neither trend is helpful when taken to the extreme. "Splitters" could easily achieve monophyly by placing every single species in its own monotypic genus. Equally, lumping all taxa together into large unmanageable genera may obscure phylogenetic relationships and evolutionary diversity. Th us, a well-balanced "middle-ground" between "lumping" and "splitting" based on scientifi c evidence is the most desirable approach. In truth, Pythonidae is a relatively small family currently containing 40 extant species in nine genera, as defi ned here, yet it has been the subject of unprecedented attention by both professional and amateur taxonomists resulting in both papers that clarify and papers that confuse the phylogenetic relationships within the family. Whereas some subspecies may be recognized, erecting additional monotypic genera and creating subgenera within small genera is unwarranted and destabilizes taxonomy. Stable nomenclature, however, is most important for "unambiguous communication about biodiversity" and names must be clearly assignable to specimens to allow "unambiguous identifi cations" (Pyle and Michel 2008: 40). Since pythons are also highly desired by both the skin and pet trades an established and widely accepted taxonomy with associated nomenclature is essential if they are to be protected and conserved. Any unnecessary and unscientifi c deviations from a well-founded taxonomy can only serve to further threaten already vulnerable wild populations.