OF NEOGENE AND QUATERNARY SNAKES OF CENTRAL AND EASTERN EUROPE . PART 1 : SCOLECOPHIDIA , BOIDAE , COLUBRINAE

Remains of Neogene and Quaternary seoleeophidians, boids and «eolubrine» eolubrids, including snakes previously deseribed and those undeseribed yet, eoming fram Poland, Ukraine, Moldavia, Czeehoslovakia, Austria, Hungary, Romania, Bulgaria, and Greece are diseussed. The following taxa, including 7 extinet speeies, were reeognized: Seoleeophidia indet.; Boidae: Bransateryx septentrionalis, Bransateryx sp., Albaneryx volynicus, cf., Gongylophis sp., Eryx jaculus, Eryx sp., eL Eryx sp., Eryeinae indet.; Colubridae: Texasophis bohemiacus, Coluber dolnicensis, Coluber planicarinatus, Coluber viridij1avus, cf. Coluber viridiflavus, Coluber caspius, Coluber gemonensis, cf. Coluber gemonensis, Coluber sp., Coronella austriaca, Coronella sp., cf. Coronella sp., Elaphe kohfidischi, cf. Elaphe kohfidischi, Elaphe paralongissima, Elaphe longissima, cf. Elaphe longissima, Elaphe quatuorlineata, cf. Elaphe quatuorlineata, eL Elaphe situla, Elaphe sp., cf. Malpolon sp., Telescopus sp., «Colubrinae» indet.


Introduction
The present study, consisting of two separate parts including this paper and the subsequent one (Szyndlar, in press), is devoted to fossil snakes found in the strip of Europe lying between the Baltic and Aegean Seas, hereafter referred to as Central and East Europe.The discussed fossils are of Neogene and Quaternary age.Contrary to the situation in West Europe, pre-Miocene snakes have never been reported from the area *.Although snake remains have been quite often found in fossil materials from Central and East European sites, most of them were rarely identified below subordinal level or described in detail.For example, a recent catalogue of Pleistocene vertebrate faunas from Hungary compiled by Jánossy (1986) lists 38 localities yielding snake remains.Materials from almost one third of the numbered sites were defined as either «abundant» or «common» or they counted thousands of bones.Of them, however, only the fossils from Villány 3 were identified to specific level (after Kretzoi, 1956; see below), while those from the remaining 37 localities were defined as «Ophidia inde1.».

History oC research
The history of previous studies of fossil snakes from Central and East Europe is brief.The first fossil snake described from the area was Coluber podo- licus, found in Ukraine (von Meyer, 1844).Two other forms reported in the nineteenth century were Laophis crotaloides and Python euboicus, both described from the Miocene of Greece by Owen (1857) and Roemer (1870), respectively.According to Rage (1984), all these three snakes are nomina dubia.
Few next researches, resumed after forty years, were restricted to the area of Austro-Hungarian Empire (Kormos, 1911;Bolkay, 1913).Although somewhat surprising, the results included in Bolkay's paper have not lost their value, because until almost the present time they have often been cited, especially by herpetologists not familiar with paleontology, as the only source of information on Central European Neogene snakes.In part it perhaps resulted from the fact that Bolkay's paper was written in English, while most publications from the area were in German.A similar opinion can be addressed to the paper of von Szunyoghy (1932); it has become famous not on account of its paleontological contents, but for a long time it has served as a useful key for determining fossil remains of modern snakes.
No other papers devoted to Central and East European fossil snakes were published prior to the early 1950s.Since that date, a number of papers were published in several Central European countries.Of them, Poland was the only country where the research on fossil snakes was undertaken on a larger scale, thanks to studies initiated by Mlynarski (1960, and further papers) in the beginning of the 1960s.Few works with reference to other countries were published (see chapter «Localities» for full account).Among these papers, especially noteworthy is the work of Rabeder (1977), providing much data on the Pleistocene snake fauna of Austria.According to my best knowledge, fossil snakes have never been reported from Albania and Yugoslavia.

Z. SZYNDLAR
Prior to the beginning of the 1980s our knowledge on extinct snakes from the area was rather limited.It is to be noted that of more than forty papers devoted to Central and East European ophidian paleontology (i.e., those in which fossils were identified to generic level at least), about half were published after 1980.Most recent publications are those of Szyndlar, Zerova, and their co-authors (Szyndlar,  1984; Zerova et al., 1987, and other papers).Studies of Zerova are especially noteworthy because they cover the southwestern Soviet Union, an area unexplored since the time of von Meyer (1844).

Contents
In the present work I attempt to summarize all upto-date knowledge about Central and East European fossil snakes.Fossils previously described in the literature as well as those hitherto not reported are considered.An overwhelming part of the discussed fossils have been personally examined; those unstudied are usually either of little importance or are 10s1.The only important collection not examined by me is the classical material of Bolkay (1913).Unfortunately, I have not received access to this collection; opinions referred to it in the following text are then exclusively based on Bolkay's descriptions and illustrations.
Morphological descriptions included in this paper are rather parsimonious and they are focused mainly on ophidian vertebrae.Throughout the text they are, however, accompanied by references to the literature containing more detailed osteological descriptions of both living and fossil snakes.Despite these limitations, the chapters entitled «Systematic account», included in both parts of this study and containing descriptions of particular taxa and comments on their taxonomic status, are extensive enough.These chapters may be thought as reference-texts for ophidian paleontologists.The chapter «History of snakes in Central and East Europe», included in the subsequent part of this study (Szyndlar, 1991), summarizes the entire information about the composition and past distribution Qf the extinct snake fauna in the area, against the background of the recent fauna; this chapter is addressed to both paleoherpetologists and neoherpetologists.

Localities
Of more than one hundred fossil snake localities covered by the present study, collections from 92 sites were examined personally by the author.Of all the sites considered, seventy altogether are discussed in detail below.A number of upper Pliocene and Pleistocene localities of Poland and Ukraine, represented exclusively by living snake species occurring at present in these areas, are considered jointly throughout the text.Detailed descriptions of the Polish Pleistocene ophidian fauna can be found in Szyndlar (1984), while most upper Pliocene and Pleistocene snakes of Ukraine are currently being studied by Zerova (in prep.).
Below, the column «Localities» lists the sites from which fossil materials were discussed in the text.AII these sites are mapped on figure 1.For Neogene localities, the European Land Mammal Ages (Fahlbusch, 1976) and Mammal Zones (Mein, 1975) are given.The zone MN 17 is regarded, according to currently accepted views, as uppermost Pliocene.References to dating of the 10calities are placed at the beginning of the column; in cases for which no papers are cited, information on their age comes from the articles listed in the column «References».
The column «References» lists papers devoted to local snake faunas, both detailed descriptions and short mentions, if such papers exists.
The column «Material examined» lists fossils studied by the author and indicates their repository.

Poland
Localities: The fossil sites listed below are exclusively of Neogene age.Younger localities, when they are mentioned in the present paper, are generally referred to as «Polish Pleistocene».AII the sites discussed here were listed and briefIy described by Szyndlar (1984).Dating of the localities follows Nadachowski et al. (1989): (1) Przeworno 2 (upper Orleanian/middle Astaracian; MN 5-7).
Material examined: whole (IZAN) except for the classical material of von Meyer (1844) from Bolurubince that was lost (fide Rage, 1984).
References: The only snake materials described in detail come from the aboye two localities.Natricine remains from Dolcine were studied by Rage and Rocek (1983); a description of the remaining snake material was given afterwards by Szyndlar (1987a).Of the latter locality, sorne snakes were briefly described by Wettstein-Westersheimb (1955).

Hungary
Localities: Datings follow Fejfar and Heinrich (1983), Jánossy (1986), and Kordos (1987):  Bolkay (1913)  References: Although Hungary possesses perhaps the highest number of snake-bearing sites in Central and East Europe, ophidian fossils were identified to generic or specific level in four papers only.In the first of them, Kormos (1911) recognized (erroneously) two snake genera in the material from Polgárdi.Bolkay (1913) described cranial remains of seven snake taxa from Polgárdi, Beremend 1, Nagyharsány-hegy and Villány 3; most of the Bolkay's material (with addition of a few new fossils) was then critically commented by von Szunyoghy (1932).Kretzoi (1956) listed three snake species from Beremend 4, and Villány 3 and 6.
References: The only hitherto described snake material comes from the archeological site of Bacho Kiro (Mfynarski, 1982); moreover, remains of a Nalrix from Dorkovo were mentioned by Thomas el al. (1986).
Material examined: whole specimens (IZBAN and ZZSiD) except for that from Dorkovo.
References: The only four literature items concerning fossil snakes of the area are descriptions of a viperid and a python from the Greek Miocene, by Owen (1857) and Roemer (1870), respectively, as well as Schneider's (1975) study on the herpetofauna of Chios; the most recent publication is a short description of elapid remains from Tourkobounia 1 by Szyndlar and Zerova (1990).

Systematic account
This chapter summarizes basic data on snake fossils available from the discussed area; throughout the text, information about particular ophidian taxa is always arranged in the same formal.
Information about fossil collections, their systematic identification, localities, and geological age are presented separately for each different taxon.When possible, the entire fossil material is listed in detail and followed by catalogue numbers.Unless the discussed collection is catalogued, only an abbreviated name of its repository is given.In cases when the material was previously described in any form, appropriate references and synonymy are cited; synonyms concern exclusively fossils coming from the discussed area.It is also always indicated unless the material was seen only by the author.If fossils referred to the same taxa were recorded from more than one loca-litY, they are listed in chronological order and each site is provided with a serial number concordant with those given in the chapter «Localities» and mapped on figure 1; both synonyms and comments concerning fossils from particular localities (in the column «Remarks») are then preceded by an appropriate serial number.
It should be stressed that aH taxonomic re-allocations given in the present work refer almost exclusively to the fossils personally examined by me or, in a few cases, to those non-examined fossils for which descriptions given in the literature were sufficiently adequate.No changes were proposed in reference to those fossils quoted in the literature which were not described, not figured or described in an unsatisfactory manner.In most doubtful cases, however, critical comments are expressed in the column «Remarks» accompanying each taxon.
The usage of the qualifiers «cf.» foHows Estes (1987).The qualifiers generally refer to structural similarities but sometimes they are also used on account of stratigraphic or geographical reasons.Diagnoses of particular taxa are brief and they intend to indicate the most important differentiating features of vertebrae; there are also added references to the literature containing more detailed descriptions.Because one of the major purposes of the present paper is to serve as a guide to the East European fossils, it includes drawings of most species found in the area.Considering that the overwhelming majority of ophidian remains found wherever in fossil sites are vertebrae, the illustrations are almost exclusively of these elements.
Suborder SCOLECOPHIDIA Duméril et Bibron, 1844.Hitherto known fossil remains of scolecophidians are restricted to precaudal vertebrae.Because of the simple morphology of the vertebrae and of a great similarity of these elements even in members of different families, identification below the subordinallevel is usually regarded as an impossible task.Scolecophidian vertebrae are of minute size, about 2 mm long or smaller, devoid of neural spines and hypapophyses, with undivided paradiapophyses, and have strongly flattened cotyles and condyles.Scolecophidia indet.(fig.2).
Material: (37) Lower Miocene (MN 4) of Dolcine: one vertebra (DPFNSP 1316). (10)Late Miocene (MN 9) of Gritsev: one vertebra (IZAN).( 14 Remarks: Except for the scolecophidian from Dolcine (37), characterized by a different morphology of the zygosphenal roof (Szyndlar, 1987) (1984) considered this (probably lost) fossil a no~en dublUm.~ased on Roemer's description and figure (supra CI!., pI.XIII), It can be stated that the discussed remains belonged indeed to a boid snake but not necessarily to a member of the ~ubfamily.B?inae.I thus agree with Rage's opinion that the ongmal descnptlOn of Python euboicus is inadequate and little can be said about the diagnostic characters of the species.
Subfamily Erycinae Bonaparte, 1831.Erycine sn~kes are characterized by highly complicated morphology of theIr caudal vertebrae and most diagnoses have been based on these elements.Precaudal vertebrae, as characteristic for most boid snakes, are always relatively very short (i.e., their centra are wide~than long) and have reduced prezygapophyseal processes.Cramal elements have not been found in Central and East European fossil sites.
Genus Bransateryx Hoffstetter et Rage, 1972.Th~type species of this extinct genus, Bransateryx vireti, was descnbed from the Upper Oligocene and Lower Miocene of West Europe by Hoffstetter and Rage (1972).The diagnosis of the genus w~s then sh~rtly summarized by R~ge (1984: 24): the palatine retammg a medIal p.rocess;.the postenor caudal vertebrae high, very short, and provlded wlth several complex additional processes.!tshould .benoted, however, that homologous vertebrae of the IIvmg specles Eryx johni from southern Asia may display a si-mIlar morphology.
Dlagnostlc vertebral characters: The posterior caudal vertebrae of Bra'"!Sateryx septentrionalis differ from those of the type species by havmg, among other characters, a spherical neural spine longer p~erapophyses, distinct subcotylar tubercles, and hypap~phy ses wlth strong lateral processes.Trunk vertebrae differ from those of the type species by having a zygosphene with three lobes in dorsal view (and not concave).The centrum length of thunk vertebrae ranges between 3.45 and 3.90 mm.For a more detailed description see Szyndlar (1987).
Remarks: This species displays morphology intermediate between that of B. vireti from the French Miocene and the living genus Charina from North America; perhaps both genera represented the same evolutionary lineage (Szyndlar, 1987) or even they should be synonymized.
Genus Gongylophis Wagler, 1830.This living genus, at present occurring on the Indian Peninsula, closely resembles Eryx and has often been synonymized with the latter.Generic distinction of Gongylophis was demonstrated by Rage (1972) and recently confirmed by Tokar' (1989).Vertebrae of Gongylophis were briefly described by Rage (1972) and carefully figured by Hoffstetter and Rage (1972, fig. 4).There are only minor differences in morphology of isolated vertebrae of both genera.However, the posterior caudal vertebrae of Gongylophis have relatively low neural spines in comparison with Eryx, while Z. SZYNDLAR its trunk vertebrae differ from all but one members of Eryx by presence of a distinct haemal keel.The only keel-bearing member of the latter genus, Eryx colubrinus from Africa, also has very low neural spines on its posterior caudal vertebrae.Thus, identification of fossil remains as members of Eryx is possible, when both trunk and posterior caudal vertebrae are available.
Remarks: The allocation of trunk vertebrae in the genus Gongylophis was based on a set of various features, most important of them the presence of a distinct haemal keel.The centrum length of the largest trunk vertebra is 6.22 mm.No caudal vertebrae of this snake have been found in Dolnice.For further details see Szyndlar (1987).
Vertebrae of members of this genus were described and discussed in a number of works (e.g., Sood, 1941;Bogert, 1968, and others).Caudal vertebrae of Eryx, provided with additional processes, display a similar complex pattern as in Charina and Bransateryx; the neural spine of Eryx is, however, distinctly lower (except in Eryx johni; see aboye).
Remarks: (70): There is no significant difference between the caudal vertebrae from Chios figured by Schneider (1975, fig. 4) and those of the living Eryx jaculus (i.e., E. turcicus of Schneider).( 73): AIso, no differences can be noted in regard to the trunk vertebrae from Pili B. Although the fossils were not compared with all living members of the genus Eryx, the probability is low that the discussed remains may have belonged to another species than E. jaculus.The islands Chios and Kos (where Pili B is situated), are presently inhabited by this snake (Wettstein, 1953).
( 14)  Remarks: Lack of haemal keels on the trunk vertebrae indicates that the remains belonged to the genus Eryx.The vertebrae from particular localities may have belonged to different species; interrelationships among them and living species are currentiY being studied by Zerova (in prep.), while fossils from Cherevichnoie (Iower layer) ( 14) by Szyndlar and Zerova (in prep.)Remarks: Because of scarcity and/or fragmentary nature of the available material the allocation of the vertebrae to the genus Eryx cannot be fully demonstrated; it is not unlikely that the vertebrae may have belonged to a Gongylophis-like snake.The remains from particular localities presumably represented different species.
Traditionally, ophidian paleontologists have subdivided the Colubridae into two subfamilies, Colubrinae and Natricinae.The only criterion of this subdivision is, respectively, absence or presence of hypapophyses on postcervical thoracic vertebrae.Since this arrangement is, at least in part, inconsistent with snake systematics accepted by neoherpetologists, below I use these names in informal forms, Le., as «colubrines» and <<llatricines».

«Colubrines».
Colubrine snakes are most abundant in post-Paleogene fossil materials.It is very easy to distinguish «colubrine» vertebrae, devoid of hypapophyses throughout the postcervical precaudal region of the column, from other advanced snakes, yet proper identification to the generic level is in most cases hazardous.For instance, it is impossible to separate overall two of the most common European genera, Coluber and Elaphe, based on vertebral characters, unless they are e10sely similar to particular living members of these genera.Proper taxonomic allocation, based on similarity to living species, is possible with reference to relatively young fossils, but this is not the case of geologically older snakes.Fortunately, there are distinct differences between most skull bones of Coluber and Elaphe (cí.Szyndlar, 1985Szyndlar, , 1988)), in this case the taxonomic position of fossil colubrine species described on the basis of both cranial and axial elements is usually well grounded.
Regarding vertebrae, subdivision of colubrine snakes into two informal groups (disregarding their generic allocation), namely «small-sized colubrines» and «Iarge-sized colubrines», may facilitate taxonomic allocation of fossils (Szyndlar, 1984, fig. 6).The former group ineludes snakes with total length rarely exceeding 100 cm (usually much smaller); trunk vertebral centra seldom reach a length of 5 mm and the vertebrae are relatively elongate.The latter group ineludes snakes with total length often reaching 200 cm (sometimes even more); centra oflarge trunk vertebrae exceed a length of 6 mm (or more) and they are almost as long as wide.The large-sized colubrines comprise the folIowing living species, presently inhabiting Central and East Europe: Z. SZYNDLAR should be noted that vertebrae belonging to juveniles of large colubrines are easily distinguished from those of adults of smal! colubrines, among others on the basis of a relatively much higher diameter of their neural canals.Identification of vertebrae of large-sized colubrines is usual!y easier than in the case of smal!er species.Apart from morphological features, proportions oftrunk vertebrae (or, strictly speaking, proportions of the length and width of vertebral centra) may be helpfui in the identification process; 1 caution other workers, however, that setting too much weight on numerical (instead of morphological) features may lead to serious errors.
Identification of vertebrae belonging to smal!-sized colubrines is an especial!ydifficult task, on account of their very similar morphology.Figure 8 shows a set of trunk vertebrae of sorne smal! colubrine species in dorsal, lateral and ventral views; it can be seen from the drawing that al! vertebrae except that of Telescopus display very similar morphological patterns.The only possible differentiating features are the shape of the zygosphenal roofs along with the length and shape of the prezygapophyseal processes.On account of intraspecific variation, however, these slight differences may be obscure and, in consequence, differentiation at even generic leve! may be impossible.This is the main reason that in my own identifications of colubrine species, based on vertebrae only, in most cases 1 use the qualifiers «cf.» before (and not after) generic names.
Genus Texasophis Holman, 1977.Remains of this extinct genus, known exclusively from precaudal vertebrae, have been reported from the Oligocene and Miocene of North America and Europe; up to the present, five species of Texasophis have been described.The basic differentiating features of the genus, given by Holman (1977), are an elongated vertebral form, moderately vaulted neural arch, low neural spine, a very robust distinct haemal keel, and very distinct subcentral ridges.The taxonomic status of Texasophis was recently questioned by Zerova (1987: 13) who placed, without any additional comments, Texasophis meini (original!ydescribed from the French Miocene by Rage and Holman, 1984) into the modern genus Boiga.According to Zerova (pers. comm., 1989), her decision resulted from comparison of Texasophis with the living Boiga trigona-tumo Indeed, the later species is closely similar (though not identical with) to Texasophis meini; on the other hand, another species of Boiga examined by me, namely B. dendrophila strongly differs in its vertebral morphology from both B. trigonatum and Texasophis.
Diagnostic vertebral characters: This extinct species, known exclusively from the type locality, is most similar to T. meini from the French Miocene, but differs from the latter by having a much narrower haemal keel; it differs from most members of the genus by having the zygosphene (in dorsal view) straight rather than provided with three lobes.The centrum length of the holotype vertebra is 3.60 mm.For more detailed morphological description see Szyndlar (1987).
Remarks: Description of this extinct, species was based on a single fragmentary vertebra only.The bone displays a set of very peculiar features not ocurring among colubrine snakes, among others a relatively very short centrum and lack of paracotylar foramina; these features can be, however, pathologic in nature (Szyndlar, 1984).Considering both the absence of any additional material and the fact that the age of the fossil may be much younger than original!yascertained (cL remarks to Natrix parva, Szyndlar, 1991) the taxonomic status of this snake is uncertain.Szyndlar, 1984.)gene and Neogene; five of them were recognized nomina dubia by Rage (1984).The referral of the overwhelming majority of these fossils to the genus Coluber cannot be demonstrated (!).It should be noted that sorne authors used the generic name Coluber not in the strict systematic meaning but rather as a symbol indicating a «typical» colubrine snake.
Diagnostic vertebral characters: Vertebrae of this extinct species, known exclusively from the type locality, resemble those of larger members of the living genera Coluber and Elaphe.The diapophysis occurs posterior to the parapophysis; the haemal keel does not reach the subcotylar rim (it forms a distinct «step» immediately behind the diapophyses); the neural spine is longer than high; the prezygapophyseal processes are obtuse and somewhat shorter than the prezygapophyseal facets; the zygosphene is slightly convex (in the larger vertebra slightly concave) in dorsal view.The centrum length of the holotype vertebra is 7.56 mm and is 1.35 times longer than wide; the centrum of the larger vertebra is as long as wide.For more detailed morphological description see Szyndlar (1987).
Remarks: The decisive evidence for referring the fossil to the former genus is presence of a prominent supraangular crest on its compound bone; this feature being characteristic of Coluber and absent in Elaphe.C. dolnicensis is therefore the oldest fossil that may be referred with certainty to the genus Coluber.
Coluber planicarinatus (Bachmayer et Szyndlar, 1985) (fig.11).Diagnosis: Trunk vertebrae of C. planicarinatus differ from those of small-sized European colubrines by having a very broad and flat haemal keel and minute paradiapophyses.The neural spine is very low (three times longer than high); the prezygapophyseal processes are very short (more than twice as short as the prezygapophyseal facets); the zygosphene is slightly convex in dorsal view.The centrum lenght of the holotype vertebra is 4.03 mm and is 1.44 times longer than wide.
Remarks: This extinct species, known exclusively from the type locality, was originally described by Bachmayer and Szyndlar (1985) as a distinct fossil genus, Nanus, on the basis of a peculiar morphology of the trunk vertebrae.Discovery of a basiparasphenoid, referred to this snake and c10sely resembling those of the recent C. najadum-C.rubriceps group, caused removal of the fossil to the genus Coluber (Bachmayer and Szyndlar, 1987).The fossil snake is perhaps c10sely related to the aboye mentioned recent species but, considering scantiness of the available materials, it is not c1ear.For detailed morphological description see Bachmayer andSzyndlar (1985, 1987).
Remarks: Bolkay (1913: 223, 224) erected a new extinct species for the discussed quadrate because he was unable to «... identify it with the spp. of Zamenis occurring in our monarchy».Specific distinction of Ihis fossil was questioned by von Szunyoghy (1913) who recognized the quadrate as comparable with the living Zamenis dahli (i.e., Coluber najadum).Mlynarski (1961), although retaining von Szunyoghy's allocation, observed that the quadrate is similar rather to that of Natrix.Rage (1984) then stated that the taxonomic status of the species is doubtful.Considering that a single quadrate is an insufficient basis for erecting a new colubrine species, 1 agree with Rage's opinion.There are no significant differences in morphology of quadrates belonging to at least three European small members of Coluber, namely C. najadum, C. rubriceps, and C. gemonensis.Perhaps distinction of C. hungaricus may be confirmed by detailed research of other remains coming from the type locality, including vertebrae, which have never been investigated by the Hungarian authors.tal fragments, 4 ectopterygoid fragments, 33 quadrates and quadrate fragments, 17 compounds (UWPI 2350/3/1-61,64-73,68-89; not seen, flde Rabeder, 1977).( 55) Middle Pleistocene of Varbeshnitsa: 9 trunk vertebrae (IZAN).
Diagnostic vertebral characters: Trunk vertebrae can be differentiated from ~h?se of other large-sized European colubrines on the basls of a dlstmctly flattened and widening posteriorly haemal keel and a straight zygosphene in dorsal view (Szyndlar, 1984, fIg. 6); m very large snakes, the zygosphenal roof is concave in dorsal vlew.The prezygapophyseal processes are relatively long (almo.sta~lo~g as the prezygapophyseal facets) and acute; the neural spme IS shghtly longer than high.The centrum length of 60 largest trunk verte?rae from R~bielice Królewskie lA (formerly C. robertmertensl) ranges between 6.67 and 8.64 mm and it is 1.17 (± 0.05) times longer than wide on average (Szyndlar, 1984); in 65 vertebrae from Bad Deutsch Altenburg 20 it ranges between 4.88 and 6.48 mm and is 1.15 (± 0.05) times longer than wide.In absolute size and proportions, they closely resemble trunk vertebrae of most East European large-sized colubrines.
(6) A.few sku!Iremains from R~bie.liceKrólewskie lA, closely resembhng C: vmdifla.vus,were descnbed by M)'ynarski (1964) as a dlstmct extmct speCles, C. robertmertensi.Numerous both cranial and axial elements from the same locality were then reported by Szyndlar (1984), who referred them to C. robertmertensi, Szyndlar (1984),.following the opinion of Mlynarski (1964), noticed close slmllanties between the fossil species and the living C. vi- ridiflavus; the most important difference was, however, the absolute I~rger size of the former snake; moreover, sorne minor morphologlcal dlfferences were observed in the premaxilla maxillae and quadrates.Trunk vertebrae of C. robertmertensi differed frot ~ose of C. viridiflavus then available, apart from greater dimensl.ons,m havmg concave and not straight zygosphene.Recent studles reveal, however, that all these differences are exclusively of allometr~c nature because morphology of bones belonging to sma-lIer speclmens of C. robertmertensi is consistent with that of C. viridiflavus.Therefore, the fossil species is here synonymized with the living C. viridiflavus.It should be noted that there are sorne minor osteological differences observed between the nominative subs~ecies from West Europe and C. viridiflavus carbonarius inhabitmg, among others, northwesternmost Yugoslavia and southern Italy; the fossils from R~bielice Królewskie lA are consis-t~nt with the latter subspecies.( 46) The same similarities were pre-vlOusly pOlnted out by von Szunyoghy (1932), with reference to cramal remains from Beremend 1.
Remarks: (4) (7) Remains from the Polish Pliocene differ in sorne details. from C. virid!flavus; although vertebrae coming from these 10cahtIes are of medlUm size, their zygosphenal roofs are often concave, moreover, a part of vertebrae from W~ie 1 does not possess flattened haemal keels.
Diagnostic vert~bral characters: Trunk vertebrae of C. caspius are usually well dlfferentiated from those of other large-sized European colubrines by the following features: the centra are distinctly el.ongated (c.L below); the zygosphene, even in smaller spe-Clmens, IS concave m dorsal vlew; the haemal keel is distinctly high and sharp throughout most of ItS length.Its height diminishes immediately behind the cotyle rim and the keel becomes wider and flattened immediately before the condyle; the prezygapophyseal processes are relatively long (as long as the prezygapophyseal fa-cets) and acute.Mid-trunk vertebrae of the following two living specimens (ZZSiD 262, 10 vertebrae measured; and ZZSiD 326, 30 vertebrae measured) have a centrum length of 5.86-6.01mm (mean 5.90 ± 9.23) and 6.70-7.12mm (mean 6.97 ± 0.11), respectively; the ratio centrum length/width is 1.36-1.42(mean 1.39 ± 0.02) and 1.29-1.42(mean 1.34 ± 3.68), respectively.It should be stressed that vertebrae of similar (and even smaller) absolute dimensions of other large-sized species of Coluber and of Elaphe usually have the centra only a little longer that wide.Of few available fossil vertebrae, the largest one (from Varbeshnitsa) has the centrum length 6.55 mm; considering that C. caspius is the largest European snake, the absolute size of its vertebrae can be certainly much higher.
Most records of C. caspius from the area were based on cranial e1ements.(44) (45) (48) (52) (53) Nevertheless, sorne records from Hungarian and Romanian sites, based exclusively on compounds (von Szunyoghy, 1932), are not credible and cannot be accepted.Of them, of special importance is the presumed presence of C. caspius in the Miocene of Polgárdi.This element is, however, very similar to those of sorne other species of the genus Coluber.For example, the only known compound of C. dolnicensis from the Czech Miocene (see aboye) c10sely resembles that of the living C. caspius, but there are significant differences between the vertebral morphology of both snakes.Unfortunately, von Szunyoghy (1932) disregarded vertebrae throughout his study; the discussed compounds needs a prompt re-examination and it is also necessary to examine vertebrae from Polgárdi in order to confirm or refute von Szunyoghy's determination.
(41) Abundant fossil materials from Bad Deutsch Altenburg 20, consisting of numerous cranial elements and vertebrae, are currently being examined by Szyndlar and Rabeder (in prep.); the bones c10sely resemble those of the living C. caspius.
Except for the Austrian and Romanian localities, all remaining fossil sites of C. caspius are located either within the present range of the species or in the c10se vicinity of it.
Diagnostic vertebral characters: Trunk vertebrae of this living species are characterized by a slightly convex (almost straight) zygosphene in dorsal view and by relatively long and acute prezygapophyseal processes; these are the only features differentiating C. gemonensis from most other small-sized colubrines.Vertebrae of C. gemonensis are c10sely similar to those of C. rubriceps, C. najadum, and Eirenis; it is then very difficult to discriminate properly from one another fossil remains of all these snakes.In a pictorial key placed in one of my previous papers (Szyndlar, 1984;fig. 6), I suggested that the main feature differentiating trunk vertebrae of C. gemonensis from other small-sized colubrines is the presence os a well-developed haemal keel, the structure being absent or weakly developed in the remaining small colubrines.Unfortunately, this statement, based on a limited comparative collection, was largely untrue; absence of the keels is characteristic for younger specimens, while in older examples of most small-sized colubrines the structure is usually c1early visible.Therefore, the haemal keel is not useful for identifying particular species.The centrum length of trunk vertebrae of cf. C. gemonensis from two Bulgarian sites (see below) ranges between 3.42 and 4.58 mm, while the ratio centrum length/width is between 1.29 and 1.50.
Remarks: (41) (55) (56) A few vertebrae, belonging to smallsize colubrines are most similar to those of the living C. gemonensis, but this allocation is not quite certain (cf.remarks for C. gemonensis, aboye).
Remarks: (10) The remains from Gritsev, owing to the presence of abundant cranial remains, belonged undoubtedly to a new extinct species of the genus Coluber; (24) trunk vertebrae coming from sorne younger Ukrainian sites are comparable with a number of living species of the genus Coluber.AII these fossils are currently being studied by Zerova (in prep.).
Diagnostic vertebral characters: In its vertebral morphology, this living snake differs from other small-sized colubrines (cf.fig.8) by a set of the following features: strongly depressed neural arch, very short prezygapophyseal processes (twice to three times shorter than the prezygapophyseal facets); zygosphene of variable shape in dorsal view, usually with two distinct outer lobes and an indistinct (or absent) median lobe; the haemal keel is usually weakly developed; in dorsal view, trunk vertebrae are strongly narrowed in the middle of the centrum length (Szyndlar, 1984).Of the material available, the centrum length of trunk vertebrae does not exceed 3.00 mm; in 60 vertebrae from the Polish Pleistocene, it ranges between 2.78 and 2.96 mm and the centra are 1.34-1.53times longer than wide (mean 1.45 ± 0.05) (Szyndlar, 1984).
Remarks: Vertebrae of C. austriaca, although in most cases well differentiable from small-sized colubrines belonging to other genera, are very similar to those of another member of the genus Coronella, namely C. girondica from West Europe.Sorne differences between both species were observed by Szyndlar (1984, fig. 6), Le., basal portion of prezygapophyses more strongly built and parapophyses longer than diapophyses in C. austriaca.These observations were based on limited comparative material.More clear differences can be observed in sorne cranial elements; (6) apart from vertebrae, skull bones were the basis of identification of C. austriaca from several Polish Pleistocene localities (see Szyndlar, 1984, for details).( 53) Bolkay (1913: 225), in his report of this snake from Bra §ov, mentioned only that «the bones agree entirely with the recent and corresponding parts of Coronella aus- triaca Laur.»; the figure showing a basiparasphenoid (ibid.fig. 3) does not fully display features characteristic for C. austriaca.( 35) Redkozubov (1987), who based his report exclusively on vertebrae, also only mentioned about the presence of C. austriaca in the Moldavian locality of Chishmikioy, providing no comments on his find.
AII the aboye mentioned sites are located within the present range of C. austriaca.
Remarks: No comments on these finds were provided by Redkozubov (1987).Remarks: Generic allocations of these vertebrae, usually strongly damaged, is not fully demonstrated.
Remarks: This extinct species, described on the basis of both cranial and axial elements, displays a number of features characteristic for the genus Elaphe, among others; the basipterygoid processes reaching the lateral margins of the basiparasphenoid (condition unknown among members of Coluber except for C. viridiflavus); the prefrontal process of the maxilla slanting.posteriorly (conditions restricted to the genus Elaphe).The baslocclpltal of E. kohfidischi, with a peculiar basioccipital crest, does not resemble other European colubrines.The vertebrae, however, in the opinion of Bachmayer and Szyndlar (1985) being closest to those of the living E. longissima and extinct E. paralongissima, are actually more similar to Coluber caspius, especially in their elongation, morphology of the haemal heel and the concave zyg?sphene (but this structure displays a different pattern of allometnc vanation than in the living C. caspius).The allocation of the remams to the genus Elaphe, although most probable, is not fully demonstrated.For more detailed morphological description of this snake see Bachmayer andSzyndlar (1985. 1987).
Remarks: The vertebrae from Cherevichnoie most resemble those of E. kohfidischi from the type locality; on account of poor preservation, their systematic allocation cannot be fully demonstrated (Szyndlar and Zerova, in prep.).
. The ectopterygoid and basioccipital figured by Bolkay display indeed features characteristic for Elaphe: the external ramus of the ectopterygoid is of subquadrate shape (d: Bolka~, 1913; pI.XII: 6), while the basioccipital possesses a dlstmct median crest forked before the occipitocondylar tubercle.~oreo,:er, the basioccipital crest consists of two tubercles and IS devOld of a median process (cf.ibid., pI.XII: 8).These fe~tures can be observed indeed in E. longissima, however, also m sorne other.members of the genus.Regarding the palatine and quadra~e, thelr allocation in the genus Elaphe is not demonstrable.Palatmes of Elaphe and Coluber differs from each other by morphology of the vomerine process, the structure missing in the.Polgá~di fossi~~cf.ibid., pI.XII: 5).The quadrate, shown m the mner vlew (d.Ibld., pI.XII: 7), does not display the qua?rate crest (lying ~m. the .ou.ter side of the bone), the feature especIally helpful for dlstmgUlshmg Elaphe and Coluber from each other.The bones may have belonged to various snake taxa, thus the status of E. kormosl remams an open qu¡:stion.
Diagnostic vertebral characters: In most cases, this living snake may be easily distinguished from other large-sized European colubrines on the basis of vertebral morphology: the haemal keel is strongly flattened and it is not widened before the condyle; the zygosphene is distinctly concave; the prezygapophyseal processes are very short (twice shorter than the prezygapophyseal facets) and acute.Its trunk vertebrae display the lowest centrum length/width ratio among European colubrines and even smaller vertebrae are relatively short, e.g., in a living specimen (ZZSiD 231; 10 vertebrae measured), with the centrum length ranging between 4.35 and 4.55 mm, the centrum length/width ratio is 1.09-1.16(mean 1.12 ± 2.33).In largest fossil vertebrae (four examples), coming from Tourkobounia 2, the centrum length ranges between 6.00 and 7.68 mm and the centrum length/width ratio ranges between 1.06 and 1.18.E. quatuorlineata is the only European snake with hypapophyses (in cervical region of the column) directed forward and not backward (cf.Szyndlar, 1984, fig. 6).
Remarks: Systematic allocation of this single vertebra, somewhat damaged, cannot be fully demonstrated.Diagnostic vertebral characters: The main difference between this living European snake and other small-sized colubrines are very short and acute prezygapophyseal processes; the centrallobe of the zygosphene is triangle-shaped (cf.fig.8).The centrum length of 14 largest vertebrae from Rethymnon ranges between 3.73 and 4.63 mm (mean 4.10 ± 0.24).In comparison with other small colubrines, trunk vertebrae of E. situla are less elongate: the centrum length/width ratio of the aboye vertebrae is 1.15-1.37(mean 1.27 ± 0.83).
Remarks: AH the localities are located within the present range of E. situla.
Remarks: (10) Abundant remains from Gritsev, clearly referable to the genus Elaphe owing to the presence of a number of cranial elements, are currently being studied by Zerova (in prep.); this is the oldest certain record of this genus in Europe.(24) (25) Younger fossils from other Ukrainian sites tentatively referred to Elaphe by Zerova (1987), are at least in part referable to sorne living Europan members of this genus.
Diagnostic vertebral characters: Vertebrae of M. monspessula- nus, the only living European member of the genus, can be differentiated from other large-sized West Paleartic colubrines by the following features: usually a thin and sharp haemal keel, weakly widening before the condyle; prezygapophyseal processes relatively long (as long as the prezygapophyseal facets or somewhat longer) and acute; a straight zygosphene in dorsal view (in small individuals with three lobes), often with a minute median notch; parapophyses distinctly longer than diapophyses.Trunk vertebrae of this species can reach very large size; the largest specimen available (MNCN 820943) has the centrum length 8.70 mm.Moreover, they are more elongated than those of other large-sized European colubrines (except for Coluber caspius).Of 30 trunk vertebrae of a living example (ZZSiD 244), with the centrum length 6.70-7.40mm (mean 7.17 ± 0.21), the centrum length/width ratio is 1.24-1.45(mean 1.31 ± 0.06).The vertebrae from Tourkobounia 1 (4 specimens) are smaller; the centrum length is 4.73-5.47mm and the centrum lenght/width ratio is 1.37-1.48.Vertebrae of M. monspessulanus coming from all regions of the column were figured by Szyndlar (1984;fig. 4).
Remarks: The vertebrae from Tourkobounia 1 do not display fully the differentiating features of M. monspessulanus.The loca-litYis situated within the recent range of this species.
Remarks: The presumed presence of Malpolon in Polgárdi reported by Kormos (1911) was not confirmed by later students of the Polgárdi herpetofauna (i.e., Bolkay, 1913, andvon Szunyoghy, 1932).On the other hand, neither Bolkay nor von Szunyoghy examined vertebrae on which Kormos' report was based.In any case, Kormos provided no description of his find and his identification is doubtful.Snake vertebrae from Devínska Nová Ves, identified by Wettstein-Westersheimb (1955)  Diagnostic vertebral characters: Unlike other small-sized colubrines, trunk vertebrae of Telescopus fallax, the only living European member of the genus, can be easily differentiated on the basis of their peculiar morphology.The most important feature is that its parapophyses are twice longer than diapophyses; moreover, the neural spine is extremely low, the zygosphene is provided with three distinct lobes and the prezygapophyseal processes are very short and obtuse.The centrum of a trunk vertebra from Varbeschnitsa has the length 3.82 mm and is 1.44 times longer than wide.
(1): A unique ophidian remain from the locality of Przeworno 2, is a vertebral fragment of a medium-sized snake (centrum length 5.27 mm), similar to Elaphe or Coluber and not to Malpo-Ion (Szyndlar, 1984).
(44): Few fragmentary vertebrae from Polgárdi belonged to snakes of very small size; they probably represent a single species.
(60): Fragmentary vertebrae from Maramena 1 belonged to at least 2 very small colubrine species; one trunk vertebra belongs to a very large snake (centrum length about 9 mm).( 63): The material from Maritsa contains vertebral fragments belonging to 2 small colubrines, both closely resembling those from Maramena 1; caudal vertebrae present in the material may have also represented natricine snakes.
(3): Several small vertebral fragments (centrum length less than 4 mm) from Podlesice display a peculiar morphology in having a strongly downswept posterior portion of the haemal keel, deep furrows accompanying the lateral foraina and the condylar neck extending far behind the neural arch (Szyndlar, 1984).
(64): Vertebrae from Spilia 4 are very small and strongly damaged; caudal vertebrae may have also belonged to Nalrix.
(70): A single colubrine trunk vertebra from Chios was figured by Schneider (1975, fig. 3 A) but not discussed in the tex!.
(8): Of three unidentified vertebrae from Mala Cave, two belong to a very small (but adult) snake, while another one is a juvenile; the generic allocation is uncertain (Szyndlar, 1984).
(73): The few strongly damaged vertebral fragments from Pili B belong to a small-sized snake.

, the remaining aboye listed fossils do not differ both from one another or from the living European species, Typhlops vermicularis. It is highly probable, especially with refe-
. AII the fossil sites are located outside the present range of the genus.