The fossiliferous bone breccia yielding the fossils here studied is located on the East coast of Mallorca, close to Caló den Rafelino, near Cala Morlanda (Manacor; Fig. 1). The breccia consists of hardened red silts of a collapsed cave excavated in the Upper Miocene reef complex (Fornós, 1998, Pomar et al., 1990). This complex has a Late Tortonian-Messinian chronology (Bizon et al., 1973). Although the red silts are filling the whole collapse, the fossils are concentrated in a very small area, about 1 sq m (Fig. 1).

The stratigraphic evidence suggests a very Early Pliocene chronology for the Caló den Rafelino site (Bover et al., 2007, Quintana et al., in press). The fossiliferous breccia is the cemented floor sediment on a cave excavated in the limestones of the Upper Miocene. Up to now, no snakes have been found in the Mallorcan sites containing the mammalian Myotragus pepgonellae (attributed to Early or Middle Pliocene [MN14 or MN16; Moyà-Solà and Pons-Moyà, 1982, Pons-Moyà et al., 1979]) and their descendants. Thus, the Caló den Rafelino infilling should have been restricted to a very short period, posterior to the Upper Tortonian-Lower Messinian and previous to Middle Pliocene (MN16). The mammalian fauna of this deposit (Bover et al., 2007, Quintana et al., in press) suggests that the colonization event occurred during the MSC (5.6–5.32 My ago, according to [Clauzon et al., 1996, Krijgsman et al., 1999]). Consequently, a very Early Pliocene age is suggested for the site.

The whole fauna obtained at this site is currently under study. It includes birds, reptiles (two viperids, one of them described here, a colubrid, a chelonid, two lacertids and an anguid), and mammals (a bovid, a leporid, a glirid, a cricetid and an insectivore) (Bover et al., 2007 and Quintana et al., in press).

The material is curated at the Institut Mediterrani d’Estudis Avançats (IMEDEA), Esporles, Mallorca, Spain.

Order Serpentes Linné, 1758.

Family Viperidae Laurenti, 1768.

Genus Vipera Laurenti, 1768.

Vipera sp. “Oriental vipers complex”.

Material: Middle trunk vertebra (IMEDEA 90113) and a cervical vertebra (IMEDEA 90116).

Age: Tentatively very Early Pliocene.

The earliest representative of the “Oriental vipers complex” in Europe comes from the early Miocene (MN4)from Lisboa (Antunes and Rage, 1974). From the MN4 to the MN16 this complex had a wide distribution in central and southern Europe (Czech Republic, France, Germany, Russia, Hungary, Italy, Moldova, Austria, Ukraine, Iberian Peninsula, Greece). Szyndlar and Rage, 2002 published a complete record and chronology of deposits with Vipera “Oriental vipers complex”, while Bailon and Blain, 2007 mentioned the last representative of the complex in western Europe (MN16 from Cova Bonica, Iberian Peninsula). However, no representative of the group had been found on islands of the western Mediterranean. In island context, the only fossil records attributed to the “Oriental vipers complex” come from the Middle Pleistocene of Chios, where a Viperidae with a similar morphology to the extant forms V. palestina–V. lebetina has been obtained (Schneider, 1975), and from the Holocene (Pre-Neolithic) of Cyprus (V. lebetina, [Bailon, 1999]). Currently, V. schweitzeri, previously considered as a subspecies of V. lebetina, is found on Chios, while V. lebetina is present in Cyprus.

The initial separation of the Balearic block from continent occurred likely during the end of the Oligocene, c. 25 My ago (e.g. de Bruijn et al., 1979). Although Adrover et al., 1977 recorded varanids, anguids and snakes in the Balearic Islands Oligocene, no description of this material has been published yet. Otherwise, the earliest known chronology for the “Oriental Vipers complex” is the MN4, between 18 and 20 My ago. Thus, it seems reasonable to suggest that the snake from Caló den Rafelino cannot be considered as a relict taxon of the fauna that accompanied the Balearic block during its first separation from the European continent, and the colonization of Mallorca by this viperid should have been a more recent event. It probably occurred during some posterior connection between mainland and the island or by overseas colonization. In this sense, several episodes of faunal colonization of the Balearic Islands (and more specifically for Mallorca and Menorca) have been proposed according to the mammalian fossil succession (e.g. Bover et al., 2008, Quintana, 1998). A first colonization event occurred during the Langhian-Serravallian regression (Middle Miocene, c. 14.2 My ago), as documented by the faunal assemblage from Sant Llorenço and Santa Margalida in Mallorca (Adrover et al., 1985, Mein and Adrover, 1982) and Punta Nati 2 in Menorca (Quintana, 1998, Quintana and Agustí, 2007). In this Menorcan site, a viperid fang displays a morphology and size similar to those of the small European vipers has been attributed to Vipera sp. by Bailon et al., 2002.

During the Messinian Salinity Crisis (MSC) (Upper Miocene, between 5.6 and 5.32 My ago), the Mediterranean sea level dropped about 1500 m (Clauzon et al., 1996, Gautier et al., 1994, Krijgsman et al., 1999), establishing new connection between the mainland and the Balearic Islands and a second colonization event. A Messinian chronology has been proposed for the colonization of the Balearic Islands by Podarcis lilfordi and P. pityusensis (Alcover and Mayol, 1981, Alcover and Mayol, 1982, Kotsakis, 1981). Bover et al., 2007 suggested that the arrival to Mallorca of the fauna found in the Caló den Rafelino took place during the main regressive episode of the Messinian. No relationship between the faunal assemblage of the Middle Miocene from Mallorca (Adrover et al., 1985, Mein and Adrover, 1982) and the fauna from Caló den Rafelino apparently exists.

Island faunas furnish relevant documentation for ecologic and evolutionary studies. Body size changes (gigantism or dwarfism) are among the most remarkable features displayed by insular species.

The large size observed in the viperid from Caló den Rafelino suggests a putative case of island gigantism. Nevertheless, in the case of snakes, some further considerations should be taken into account. First, most of the reptilian species have a continuous (though increasingly slow) growth once sexual maturity is reached (Andrews, 1982). Second, snakes display a wide range of adult body size, even within the same population and among different populations, as a consequence of the ectothermic metabolism (avoiding the constraints promoted by the maintenance of a constant body temperature, as is the case in endothermic animals) (Pough, 1980). Third, in the fossil record for this group of vipers it can be found examples of continental populations of large size, although slightly smaller than the snake from Caló den Rafelino (Szyndlar and Rage, 1999). Currently, the maximum body size known for a continental V. lebetina (“Oriental vipers complex”) is 214 cm, with a high variability among different populations, while the insular subspecies V. lebetina lebetina from Cyprus and V. schweizeri from the Cyclades Islands are generally much smaller (Mallow et al., 2003).

Patterns of body size change in island snakes are unclear. Case, 1978 described a trend towards decreasing body size (dwarfism) in island snake populations, attributed to changes in food availability. Other papers described an opposite scenario, i.e., snake populations with larger body size on islands than on mainland (Forsman, 1991, Mori, 1994, Rodriguez and Drummond, 2000, Schwaner and Sarre, 1988). All these authors related the size of the snakes to the size of the consumed preys, the latter being different on islands from that consumed on mainland (“diet alteration hypothesis” [Boback, 2003]).

Boback, 2003 studied the body size of 30 snake species and he suggested that body size changes in snakes are bimodal and mainly influenced by the size of the available preys, whereas none of the physiographic or phylogenetic history variables seem to be key factors to determine the recorded size trends. The distribution of populations of giant or dwarf snakes differs significantly between the families Colubridae and Viperidae (Boback, 2003): while in viperids a reduction of size in island populations is frequently observed, the foraging style of colubrids (nest-robbing behaviour) may promote the increase of body size in island populations. However, almost all viperids studied by Boback, 2003 belong to the genus Crotalus, and it cannot definitively be established if the reduction in body size on islands affects all Viperidae genera or just species of Crotalus. According to Boback, 2003, island populations of Vipera schweizeri (V. lebetina in the Boback paper) from Cyclades Islands, the only representative of the “Oriental vipers complex” considered by this author, displays a remarkable decrease in body size. However, Boback and Guyer, 2003 proposed an empirical model that suggests a trend for snakes to acquire an optimal body size. In the case of snakes with continental and island populations, these authors noted that the species of small body size become larger in the islands, while the large species on the continents are smaller in the islands.

Several papers suggested that populations of snakes that reach a large body size live on islands with abundant endothermic preys, while the body size is ostensibly reduced in populations living on islands with only ectothermic animals (Aubret and Shine, 2007, Bonnet et al., 2002, Case, 1978, Hasegawa, 2003, Schwaner and Sarre, 1988).

With an unknown mainland ancestor, the interpretation of the large size of the Caló den Rafelino snake is controversial. Its size is close to the maximum end of the body size range known for the “Oriental vipers complex”, suggesting that an insular body size decrease should be excluded. In this sense, the emerging trend of body size change does not agree with the snake general model of Case, 1978 and Boback and Guyer, 2003, or of Boback, 2003 model for the Viperidae. According to these models, the viperid from Mallorca should have displayed a smaller size than its mainland ancestors.

In our view, the isolation of a viperid on Mallorca after the MSC, should have produced adaptations to a new diet, consuming preys of larger size, mainly endothermic vertebrates (birds and mammals), whose size and probably abundance have been increased through their insular evolution (at least for some of them, as evidenced by the existence in the deposit of rodents of large size). The great phenotypic plasticity of this group of snakes should have selected phenotypes successively adapted to the change of body size of the preys. A co-evolution process between the island populations of this group of vipers and endothermic preys existing in the island could have been the key factor to explain the large body size displayed by this fossil viper.

This process was probably favoured, although to a lesser degree, by the existence of favourable ecological conditions, such as the existence of a warm-temperate climate and perhaps a relatively dry and open landscape. Due to its large body size, the snake from Caló den Rafelino can be considered as one of the largest predators in Mallorca during the Early Pliocene.