NEW DATA ON THE VERTEBRATE ASSEMBLAGE OF FIUME SANTO (NORTH-WEST SARDINIA, ITALY), AND OVERVIEW ON THE LATE MIOCENE TUSCO-SARDINIAN PALAEOBIOPROVINCE
Abstract
Abstract: This paper reports the results of a study of more than 2400 specimens in the collection of fossil vertebrates from Fiume Santo in north-west Sardinia. This locality represents the westernmost documentation of the Tusco-Sardinian palaeobioprovince, which was in existence during the late Tortonian in the North Tyrrhenian area. During the Tortonian, the region that presently corresponds to southern Tuscany and the Sardinian Massif (Central Italy) was occupied by a complex of large islands characterised by endemic vertebrate populations. Morphological features, adaptations and trophic structure (e.g. low diversity and scarce carnivore taxa), mainly in the mammal faunas, attest to a long period of isolation from the continental palaeobioprovinces of the Mediterranean and Central Europe. The greatest numbers of fossil remains of the well-known endemic Oreopithecus fauna of the so called V0–V2 assemblages, are found at several sites in southern Tuscany (e.g. Montebamboli, Casteani, Baccinello). In Sardinia this endemic fauna occurs only at the Fiume Santo locality, from which 11 taxa have been recognized: Crocodylia indet., Chelonii indet., Oreopithecus bambolii, Mustelidae indet., Eumaiochoerus cf. E. etruscus, Umbrotherium azzarolii gen. et sp. nov., Tyrrhenotragus gracillimus, Bovidae gen. et sp. indet. (?Neotragini), Maremmia cf. M. lorenzi, Etruria viallii gen. et sp. nov., Turritragus casteanensis gen. et sp. nov. Analysis of these fossils and comparison with material from localities in Tuscany has led to a re-evaluation of the latter and to the description of three new endemic taxa among the ruminants.
The Fiume Santo site in north-west Sardinia (Sassari Province; Text-fig. 1) yields a rich assemblage of endemic fossil vertebrates. The fauna includes Oreopithecus, a highly derived extinct ape, and the bovids Maremmia and Tyrrhenotragus. The occurrence of these taxa characterises the faunal assemblages V0–V2 of the well-known Tortonian faunas (Turolian Mammal Age, Mammalian Neogene (MN) zones 11–12, ‘mammal zone’; Text-fig. 2) of the Maremma region in southern Tuscany (Hürzeler and Engesser 1976), where the most complete faunal succession crops out in the Baccinello-Cinigiano Basin (Lorenz 1968; Engesser 1989; Rook et al. 1996, 1999a; Bernor et al. 2001). The Sardinian site represents the westernmost documentation of the extinct insular fauna of the so-called Tusco-Sardinian palaeobioprovince (Oreopithecus faunal assemblages, or ‘OZF’Oreopithecus Zone faunas, as defined in Bernor et al. 2001). This province was in existence until the latest Tortonian in the North Tyrrhenian region, when this insular region began to be connected to the European mainland.
Map showing the location of Fiume Santo site.
Geochronological framework.
The Fiume Santo site was discovered in the early 1990s (Cordy and Ginesu 1994; Cordy et al. 1995) during deep excavations carried out for the construction of a parking area within the thermo-electric power station of the Italian State Electricity Company (now ENDESA Italia). The site has been declared protected by the local office of the Italian Ministry for Cultural Heritage and Archaeology, the ‘Soprintendenza per i Beni Archeologici per le province di Sassari e Nuoro’. The fossiliferous area is relatively large and within the industrial complex of the electricity company. In September 2001, the ‘Soprintendenza per i Beni Archeologici per le province di Sassari e Nuoro’ put the University of Florence in charge of the excavation and study of the site and its fauna. The present contribution presents the first results of an analytical study of the 1994–95 fossil sampling, a small portion of the entire Fiume Santo assemblage, which currently consists of more than 13,000 specimens.
At present, the geological context of the site is not satisfactorily known (Cordy et al. 1995). An extensive geological survey is being carried out (Benvenuti and Papini in progress) within the framework of the project.
Institutional abbreviations. Bac, Baccinello collection housed in the Naturhistorisches Museum, Basel; FS, field inventory register of Fiume Santo material; IGF, Natural History Museum, Geology and Palaeontology section, University of Florence; IR, formal inventory register of Fiume Santo material according to the ‘Inventario Regionale della Soprintendenza’.
Systematic Palaeontology
The faunal assemblage
The revision of the Fiume Santo fossil assemblage, the object of the present paper, has been based on the remains collected during the first excavations (1994–95), as noted above. It was carried out under the direction of one of us (LT) and stems from a study of more than 2400 specimens (of which at least one-quarter have not yet been identified), representing 11 vertebrate taxa: Crocodylia indet., Chelonii indet., Oreopithecus bambolii, Mustelidae indet., Eumaiochoerus cf. E. etruscus, Umbrotherium azzarolii gen. et sp. nov. (nomen nudum in Hürzeler and Engesser 1976), Tyrrhenotragus gracillimus, ?Neotragini gen. et sp. indet., Maremmia cf. M. lorenzi, Etruria viallii gen. et sp. nov. (nomen nudum in Hürzeler and Engesser 1976), and Turritragus casteanensis gen. et sp. nov. (for a complete list of the material analysed herein, see Supplementary Data file on the website of the Palaeontological Association: http://www.palass.org). The available material did not allow us to identify some of the taxa listed by Cordy et al. (1995), which includes two rodents (the murid Valerymys aff. V. turoliensis and a large Gliridae indet.). No micro-mammal remains were recovered from the material at our disposal. The same holds true for carnivores. Cordy et al. (1995) mentioned the occurrence of the endemic Ursidae ‘Hyaenictis’anthracitis, but our study revealed the occurrence of only two fragmentary remains attributable to the Mustelidae.
Although the fossil assemblage analysed represents only a small part of the whole Fiume Santo collection, our taxonomic study and general conclusions are significant enough to be presented here and will provide a basis for future palaeontological investigations at the site.
The fossils are not often well preserved. The enamel of teeth is perfectly preserved but dentine and roots are often chemically eroded or sometimes absent altogether. Bones generally appear ‘decorticated’ with the innermost part of cortical bone or even the trabecular bone usually being exposed (Text-fig. 3). These states of preservation suggest that the teeth and bones were in highly alkaline environments both before and after burial, which strongly altered the most organic components of skeletal elements (Fernández-Jalvo et al. 2002).
A, example of a fossiliferous assemblage from Fiume Santo; in the foreground is a right mandible of Umbrotherium azzarolii; × 0.5. B, Maremmia cf. M. lorenzi M3 with absent dentine; × 2. C, Maremmia cf. M. lorenzi humerus; arrow indicates an area where trabecular bone is exposed; × 0.5.
The identifications of mammals are based mainly on tooth morphology and, in the case of bovid remains, on horns. Postcranials are mostly badly fragmented, preventing their certain identification. A first estimate of the percentage occurrences of each taxon is provided in Table 1 and Text-figure 4. In Text-figure 4A the percentages have taken into account the entire set of identified fossils (horns, teeth and bones of mammals, teeth of crocodilians and shell fragments of chelonians). In Text-figure 4B only mammalian dentition and isolated teeth have been considered in order to avoid over-representation of the most common and easily identifiable taxon Maremmia. From these diagrams it is apparent that bovids are the best documented group, representing more than 75 per cent of the whole sample of mammal dentitions, whereas carnivores and suids are represented by only 2–4 fragments per taxon.
| Taxa | No. | % |
|---|---|---|
| A | ||
| Indet. | 661 | 26.72 |
| Chelonii indet. | 22 | 0.89 |
| Crocodylia indet. | 49 | 1.98 |
| Mustelide indet. | 2 | 0.081 |
| O. bambolii | 25 | 1.01 |
| E. cf. etruscus | 4 | 0.16 |
| Bovidae indet. | 429 | 17.33 |
| T. gracillimus | 45 | 1.77 |
| ?Neotragini indet. | 36 | 1.45 |
| M. cf. lorenzi | 853 | 34.46 |
| Etruria viallii | 35 | 1.41 |
| T. casteanensis | 14 | 0.56 |
| U. azzarolii | 301 | 12.17 |
| Total | 2475 | |
| B | ||
| Indet. | 86 | 5.03 |
| Mustelide indet. | 2 | 0.12 |
| O. bambolii | 25 | 1.46 |
| E. cf. etruscus | 4 | 0.23 |
| Bovidae indet. | 352 | 20.62 |
| T. gracillimus | 44 | 2.57 |
| ?Neotragini indet. | 33 | 1.93 |
| M. cf. lorenzi | 824 | 48.21 |
| Etruria viallii | 35 | 2.05 |
| T. casteanensis | 14 | 0.82 |
| U. azzarolii | 290 | 16.97 |
| Total | 1709 | |
- Bovidae indet. includes fragments (e.g. of enamel walls and bones) that cannot be referred with certainty to any of the small to middle-sized species.
Proportional occurrences of taxa from Fiume Santo based on A, total remains and B, mammalian dental remains.
The taxonomic composition and diversity of the Fiume Santo fauna is fully comparable with that of southern Tuscany (Baccinello V0–V2, Montebamboli, Casteani, Ribolla, Montemassi, Serrazzano), and all occurrences of mammal taxa at Fiume Santo are the only records of their presence outside Tuscany (see data in Rook et al. 2000). Although both faunas are characterised by the dominance of the bovid Maremmia, some differences in the relative abundance of taxa are evident: for example, the giraffid Umbrotherium azzarolii is the second most abundant form at Fiume Santo, whereas in southern Tuscany it is represented by only one specimen (the type specimen from Casteani). Another difference concerns the characteristics of the dentition of Maremmia, which suggest different ecological conditions between the regions (see discussion of bovid taxonomy below).
Class REPTILIA Laurenti, 1768
Remarks. The herpetofauna is represented by remains of crocodylians and chelonians. No amphibians or small reptiles have been identified in the 1994–95 samples, although Cordy et al. (1995) listed four unidentified forms: one chelonian, one anuran and two crocodylians.
Order CROCODYLIA Gmelin 1789
Crocodylia indet.Text-figure 5A–B
Herpetofauna from Fiume Santo. A–B, Crocodylia indet., isolated tooth IR#63925 in A, lingual and B, mesial views; × 1. C–D, Chelonii indet., right xiphiplastron FS1995#0579 in C, ventral (the dark area is concretion filling a shallow depression) and D, dorsal views; × 0.8.
Description. The crocodylians are represented exclusively by isolated teeth (49 recovered) of medium size (crown height c. 7–33 mm) and rather massive form (Text-fig. 5A–B). Differences in elongation and size indicate that they come from different parts of the teeth-bearing bones. The teeth are variably fractured and incomplete. The enamel is usually translucent and when the surface is well preserved, it is usually characterized by fine longitudinal ridges and irregularly scattered small pustules that may anastomose in the distal sector where they diverge in laterally, reaching the mesiodistal keels (without producing denticles or serrations). In specimens having a transverse constriction on the crown, the ridges are weakly expressed and the small pustules more in evidence.
Comparisons and discussion. Despite the poor phylogenetic value of isolated teeth (Brochu 2000), some taxonomic relevance cannot be denied in particular contexts of known age and geographic provenance, as in the case of the Fiume Santo assemblage. The crocodylians of the late Miocene Mediterranean islands are represented exclusively by members of the family Crocodylidae since the range of the alligatoroid Diplocynodon was restricted to the European mainland, possibly because some sort of intolerance to salt water limited its dispersal (see discussion in Delfino in press). The isolated teeth from Fiume Santo do not belong to this genus, whose Miocene representatives have teeth that are more acute with smaller crowns and are usually not ornamented by any ridges. It is, therefore, possible to doubt the presence of this genus at Fiume Santo. Cordy et al. (1995) listed two crocodylian taxa from this locality but the available isolated teeth suggest derivation from different parts of the teeth-bearing bones of a single taxon. According to a recent review of the Italian Cenozoic crocodylians (Kotsakis et al. 2004), Sardinia was inhabited by at least one long-snouted crocodylid of the Tomistominae during the late Miocene. Short-snouted crocodylids surely inhabited southern Tuscany (‘Crocodylus bambolii’ Ristori, 1890 is a crocodylid of uncertain affinity; Delfino and Rook 2008), and possibly the Apulo-Abruzzi Province, where non-Diplocynodon-like isolated teeth have been found at the locality of Scontrone (Rustioni et al. 1992), whereas Crocodylus cranial remains have been identified recently from the Gargano palaeo-archipelago (Delfino et al. 2007). Well-grounded evidence for Crocodylus in Sardinia is not yet available, even if the morphology of the isolated teeth reported here does not differ from that of a member of this genus. Further remains (cranial elements) are needed to support such a determination; hence the material from Fiume Santo is at present conservatively identified only at order level.
Order CHELONII Brongniart, 1800
Chelonii indet.Text-figure 5C–D
Description. Chelonians are represented by a few small carapace or plastron fragments. The most interesting specimen is fragment FS1995#0579, which is tentatively considered to be the external portion of a right xiphiplastron (Text-fig. 5C–D). Sutures are not preserved but a transverse ‘step’ on the ventral (external) surface may represent the boundary between the femoral and anal shields. Such a boundary is not visible on the dorsal surface of the element (in the lateral area covered by the shield), probably because of inadequate preservation. The surface immediately caudal to the ‘step’ is evidently slightly depressed in a way similar to that seen in modern Testudo graeca from Turkey (in the female the depression is also developed anterior to the furrow between the femoral and anal shields). Unlike the shape of shells of modern Mediterranean testudinids, the posterolateral edge of the xiphiplastron is not approximately rounded but straight. The maximum thickness of 12.5 mm is reached at the level of the proximal (broken) tip.
Comparisons and discussion. The chelonian fragments available are of little diagnostic value. Although their size rules out the presence of marine turtles and the proximal thickness of specimen FS1995#0579 could suggest a tortoise, all the remains are simply referred to Chelonii indet.
Class MAMMALIA Linnaeus, 1758Order PRIMATES Linnaeus, 1758Family HOMINIDAE Gray, 1825
Genus OREOPITHECUS Gervais, 1872
Type species. Oreopithecus bambolii Gervais, 1872, from Montebamboli (Grosseto, Central Italy), late Tortonian, late Miocene.
Oreopithecus bamboliiGervais, 1872Plate 1; Table 2
Figs 1–25. Oreopithecus bambolii from Fiume Santo. 1–4, IR#63902, right P3 in 1, distal, 2, occlusal, 3 proximal and 4, palatal views. 5–7, FS1995#0011, left M1 in 5, occlusal, 6, palatal and 7, buccal views. 8–10, IR#63900, left M2 in 8, occlusal, 9, palatal and 10, buccal views. 11–13, FS1995#0010, right M3 in 11, occlusal, 12, palatal and 13, buccal views. 14–16, FS1995#0009, right D4 in 14, occlusal, 15, lingual and 16, buccal views. 17–19, IR#64428, left P4 in 17, occlusal, 18, lingual and 19, buccal views. 20–22, IR#63905, left M1 in 20, occlusal, 21, lingual and 22, buccal views. 23–25, FS1995#0004, right M3 in 23, occlusal, 24, lingual and 25, buccal views. All × 3.
| Taxa | L | B | Ba | Bp |
|---|---|---|---|---|
| IR#63902 P3 | 7.1 | 9.2 | ||
| IR#63904 P4 | 6.8 | 9.2 | ||
| FS1995#0005 M1 | 8.5 | 8.0 | 7.3 | |
| FS1995#0011 M1 | 8.7 | 8.6 | 8.5 | |
| FS1995#0012 M1 | 10.5 | |||
| IR#63900 M2 | 9.4 | 9.3 | 9.1 | |
| IR#63903 M2 | 9.9 | 9.5 | 7.7 | |
| FS1995#0001 M2 | 8.9 | 8.6 | 7.9 | |
| FS1995#0002 M2 | 10.1 | 8.8 | 8.1 | |
| IR#63901 M3 | 9.7 | 9.2 | 7.5 | |
| FS1995#0006 M3 | *11.4 | 10.4 | ||
| FS1995#0010 M3 | 9.6 | 9.1 | 8.0 | |
| FS1995#0012 M3 | 10.5 | |||
| FS1995#0009 D4 | 7.4 | 5.4 | 4.6 | |
| FS1995#0003 P3 | 5.8 | 7.3 | ||
| IR#63906 P3 | 6.5 | 8.5 | ||
| IR#63906 P4 | 7.0 | 7.7 | ||
| IR#64428 P4 | 7.8 | |||
| IR#63905 M1 | 9.4 | 7.7 | 7.2 | |
| IR#64427 M1 | 7.3 | |||
| FS1995#0007 M2 | 10.5 | |||
| FS1995#0004 M3 | 11.4 | 9.2 | 8.8 |
Description
Twenty isolated teeth and one right fragmentary mandible bearing P3–P4. No postcranial elements have been identified. Apart from a short note on the fragmentary mandible IR#63906 by Cordy and Ginesu (1994, TEXT-FIG. 3., TEXT-FIG. 4.), none of these specimens has been described previously. A selection of the material is illustrated in Plate 1 and measurements of the specimens are given in Table 2.
Upper dentition. Upper premolars are represented by one right P3 (IR#63902) and one right P4 (IR#63904); they are typically bicuspid, with subequally developed cusps (the buccal one being slightly more robust). The P3 (IR#63902; Pl. 1, figs 1–4) is a virtually unworn tooth; a strong cingulum encircles its base, and is more developed on the lingual aspect. The P4 (IR#63904) is a deeply worn tooth, with the two cusps worn down to be connected distally around a dentine lake. A marked cingulum runs on the lingual side. The first upper molar crowns are longer than broad, with strongly rounded corners. Of the two M1s in the sample, one is nearly unworn (FS1995#0005), while the other (FS1995#0011; Pl. 1, figs 5–7) is in an advanced stage of wear, having dentine exposure on three main cusps (protocone, paracone and hypocone). The protocone (the largest of the four cusps) is relatively low and conical in shape. No metaconule is present along the crista obliqua. Mesial to the protocone there is a cingular shelf, which ends on the mesial face of the paracone. The mesial fovea is variable, being easily identifiable in FS1995#0011 and very small in FS1995#0005. In both M1s, distal to the metacone there is a relatively broad distal shelf, which in FS1995#0005 bears a distinct distal conule. The second molars are wider than the first. Of the four specimens, three are unworn or slightly worn (IR#63903, FS1995#0001, FS1995#0002) while one is at a relatively advanced stage of wear (IR#63900; Pl. 1, figs 8–10), with all four main cusps exposing a dentine tip. The mesial margin is characterized by the occurrence of a wide mesial cingulum, delimiting a variably sized mesial fovea. The metaconule is absent along the crista obliqua in all four M2s. Distal to the hypocone and metacone is a robust, broad distal shelf, which in IR#63903, FS1995#0001 and FS1995#0002 is characterised by the presence of a well-defined distal conule. The third upper molars are characterised by the narrowing of the distal portion of the tooth. All four M3s present in the sample are at a very low stage of wear. Two are complete (IR#63901, FS1995#0010; Pl. 1, figs 11–13), while the other two (FS1995#0006, FS1995#0012) are damaged in the distolabial part. All are characterised by a wide, robust mesial cingulum running mesial to the protocone and paracone, and by a distal shelf that, starting at the disto-internal margin of the metacone, goes around the hypocone base as far as the labial side of the tooth. The distal shelf is characterised by the occurrence of small cusplets on its ridge. A peculiarity of IR#63901 is the reduction of the hypocone that gives a ‘triangular’ shape to the tooth.
Lower dentition. The lower deciduous dentition of Orepithecus is known thanks to the left D4 of the type specimen (Gervais 1872; Hürzeler 1951). The right D4 from Fiume Santo (FS1995#0009; Pl. 1, figs 14–16) is virtually identical in its dimensions and morphology to the Monte Bamboli D4 illustrated and described by Hürzeler (1951, 1958). In respect to the type D4 (which is partially damaged with the hypoconid area missing), the Fiume Santo specimen is complete and shows a slightly more advanced stage of wear.
A crown fragment of a left I1 (FS1995#0013) is the only incisor present in our collection. The specimen lacks its base (cervix) and the buccal margin is more intact. The occlusal margin is unworn and characterised by the occurrence of a larger (distal) and a smaller (mesial) mammelon. The margin ridges on the lingual side are almost undetectable, the mesial one being more marked.
Lower premolars are represented by a right fragmentary mandible with P3–P4 (IR#63906), a left isolated P3 (FS1995#0003) and a left isolated P4 (IR#64428; Pl. 1, figs 17–19). Both P3s represent a bicuspid tooth, with a large protoconid and a smaller, but prominent metaconid located transversely opposite the protoconid. There is no evidence of wear produced by the honing action of the upper canine. The lingual aspect of the metaconid differs between IR#63906 and FS1995#0003, the latter one showing a concave profile. The mesial fovea is made by an extension of the lingual cingulum, which in Oreopithecus is expanded. It is worth mentioning that P3 in Oreopithecus is extremely variable, ranging from a relatively narrow, single-cusped sectorial tooth to a broad, bicuspid molariform tooth (Rook et al. 1996). In this respect the notation of Cordy and Ginesu (1994), which stresses the differences between IR#63906 and lower premolars from the Tuscan samples, especially for the P3, seems inappropriate.
Regarding the P4s (IR#63906 and IR#64428), both are at a comparable stage of wear (exposure of dentine tip on metaconid) and well preserved, although in IR#64428 a flake of enamel is missing on the distal wall of the paraconid. As for the P3s, the tooth is broad and elliptical in shape. The talonid basin forms a broad, but shallow triangular platform distal to the main cusps. The distal marginal ridge bears cusplets along the crest. In both, P4s have a continuous cingulum (stronger in IR#63906) that encircles the buccal aspect of the protoconid.
In common with the upper molars, the lower molars are characterized by four main cusps alternately arranged. Two M1s are present in the sample, one complete (IR#63905; Pl. 1, figs 20–22) and one represented by the distal part of the trigonid (IR#64427), both at a similar stage of wear, with dentine exposed on tips of the main cusps. The protoconid and metaconid are large and transversely aligned. A centroconid is situated in the midline of the tooth. The distal shelf is wide and runs around the metaconid forming a lingual cingulum. The only M3 (FS1995#0004; Pl. 1, figs 23–25) consists of the crown of a well-preserved, unworn tooth. It is relatively long but not as in other Oreopithecus specimens. The protoconid and metaconid are the larger and most elevated cusps. These are transversely aligned, a condition typical for Oreopithecus. From the apex of the protoconid, a short preprotocristid connects it to a relatively small protoconid. A centroconid is clearly positioned in the midline of the tooth. The talonid area is unusual because a distal fovea is not clearly delimitated. The hypoconid and entoconid are relatively large. Such a talonid morphotype falls within the range of the observed morphology of Oreopithecus.
The mandibular ramus fragment IR#63906 is poorly preserved with no inner or outer surface of the bone so that the thickness of the ramus cannot be estimated; the only feature that can be seen is the height of the ramus below the premolars (23.5 mm below P4), which is within the range observed in the Tuscan sample (contra Cordy and Ginesu 1994).
Comparisons and discussion. Oreopithecus was a large-bodied hominoid with some features that are typical of living apes, reflecting significant adaptations to vertical climbing (Harrison 1991), but also other characters in several parts of its skeleton that are more likely to be linked to bipedality (Köhler and Moyá-Solá 1997, 2003; Moyá-Soláet al. 1999, 2005; Rook et al. 1999b, 2004). The site of Fiume Santo, with its enormous record of undistorted bone remains, has the potential to yield undistorted post-cranial remains of Oreopithecus, which will eventually provide a critical contribution to the continuing debate about its locomotor behaviour (Moyá-Soláet al. 2005; Rook et al. 2004). At present, it is represented at Fiume Santo only by dental remains and the sample available for this study shows that no appreciable differences exist between it and that from southern Tuscany in either morphology or dimensions (Text-fig. 6). A more comprehensive study, including the material collected during site exploitation from 1996 until the present, will allow for more extensive comparisons and a better definition of the Fiume Santo Oreopithecus in due course.
Scatter diagrams comparing total length and posterior breadth of A, M2 and B, M3 of Oreopithecus bambolii samples from Fiume Santo (solid diamonds) and sites from southern Tuscany (open diamonds).
Order CARNIVORA Bowdich, 1821
Remarks. Predators are usually absent from insular mammalian assemblages or limited to forms with peculiar specializations in behavioural/feeding strategies (MacArthur and Wilson 1963; Sondaar 1977, 1987). The mammal assemblage of the so-called Oreopithecus faunal assemblages (OZF sensuBernor et al. 2001) is no exception in this regard. Those in southern Tuscany have yielded some carnivores, all represented by very limited remains. An isolated mandible from Montebamboli represents a bear (Indarctos anthracitis) showing unusual dental adaptation to an omnivorous diet (Weithofer 1888), and a maxillary fragment represents the mustelid Mustela majori (Ficcarelli and Torre 1967). In addition, and better documented, are the lutrine carnivores: these are known to be relatively diverse, being represented by three species of two endemic genera: Tyrrhenolutra helbingi (Baccinello V1), Paludolutra maremmana (Montebamboli), and Paludolutra campanii (Montebamboli, Baccinello V2) (Hürzeler 1987).
Among carnivores, Cordy et al. (1995) recorded only the bear ‘Hyaenarctos’anthracitis from Fiume Santo. Within the material studied, however, identified carnivore remains are limited to two very fragmentary specimens that are not identifiable beyond family level.
Suborder CANIFORMIA Kretzoi, 1943Family MUSTELIDAE Fischer de Waldheim, 1817
Mustelidae indet.
Description. Material attributable to Mustelidae indet. consists of two small dental fragments, one representing part of the anterior portion of possibly a left P4 paracone, the other a portion of the talon (with protocone, hypocone and lingual cingulum) of a left M1.
Comparisons and discussion. The fragmentary status and the limited sample do not allow much scope for comparison and discussion. We are confident that analyses of post-1995 recoveries and new site excavations at Fiume Santo will yield more evidence of the carnivores that characterise the faunal assemblage.
Order ARTIODACTYLA Owen, 1848Family SUIDAE Gray, 1821
Genus EUMAIOCHOERUS Hürzeler, 1982
Type species. Eumaiochoerus etruscus (Michelotti), 1861; from Montebamboli (Grosseto, Central Italy), late Tortonian, late Miocene.
Eumaiochoerus cf. E. etruscusText-figure 7
Eumaiochoerus cf. E. etruscus from Fiume Santo. A–C, FS1995#0196, right M3 in A, labial, B, lingual and C, occlusal views; × 2.
Description. The Suidae are represented by only two complete teeth: a right I3 and a right M3 (Text-fig. 7), the latter without a root but the occlusal surface is well preserved, without wear and with no eroded enamel. This specimen is subtriangular in shape with four main cusps: paracone (anterobuccal), metacone (posterobuccal), protocone (anterolingual), and hypocone (posterolingual). These are not opposed but alternating. The anterior margin has a strong cingulum with a central cusp (protopreconule). There is an accessory cusp (tetrapreconule) positioned centrally between the anterior and posterior pairs of the main cusps. Two small additional cusplets are present labially at the tetrapreconule. A second accessory cusp (pentapreconule) occurs between the posterior pair of main cusps and the talon. This last one is simple with a large central cusp divided by a small furrow, and three small pillars on the labial side.
Based on its dimensions (length, 31.53 mm; width, 20.77 mm) and morphology, this specimen can be referred to Eumaiochoerus etruscus following the redescription by Mazza and Rustioni (1997) of the Montebamboli specimens, the only significant difference being that the three grooves (‘furchen’) in the Fiume Santo specimen, normally present on the main cusps of Suidae (Hunermann 1968), are shallow in the paracone and metacone and completely lost in the protocone and hypocone. No conclusions are possible from the other fragmentary remains from Fiume Santo.
Comparisons and discussion. Van der Made (1999) considered the simplification of molars with the loss of ‘furchen’ to be a result of increasing enamel thickness, a derived feature occurring in suids under insular conditions. The endemic Sus sondaari from the Plio-Pleistocene of Sardinia lacks them. The fact that grooves are lacking in the Fiume Santo suid, unlike that from Montebamboli-Baccinello V2, suggests a more pronounced degree of endemic evolution in the former. The limited sample does not allow much scope for further comparison, so despite the close resemblance, we prefer to take a cautious approach and refer our material to Eumaiochoerus cf. E. etruscus.
Family GIRAFFIDAE Gray, 1821
Genus UMBROTHERIUM gen. nov.
1976 Umbrotherium Hürzeler and Engesser, p. 334 (nomen nudum)
Derivation of name. After one of the oldest peoples of Central Italy (Umbri; First Millennium AD) and Greek, therion, wild beast, a common reference to mammals.
Type and only species. Umbrotherium azzarolii sp. nov.
Diagnosis. Middle sized ruminant with fairly brachyodont dentition. Markedly rugose enamel wall. P3 and P4 not molarised, rectangular in outline, wider than long. Upper molars with a weak entostyle and cingulum. Mesostyle joins metastyle via a poorly developed cingulum; anterior lobes longer than posterior lobes. Cranial features, in particular whether or not ossicones occur, currently unknown.
Remarks. The remains originally referred to Umbrotherium by Hürzeler and Engesser (1976), and analogously to Etruria viallii (see below), were not described, rendering it a nomen nudum according to the ICZN (1999). In order to make the name available for use, we provide an extensive description and formal definition here.
Umbrotherium azzarolii Hürzeler and Engesser, 1976 was referred to the Giraffidae family by these authors. This attribution is followed here although it is not firmly grounded because the main diagnostic characters, such as bilobate lower canines and ossicones, are not documented from either southern Tuscany or Fiume Santo. Nevertheless, some features present in the dentition, such as the markedly rugose enamel, highly molarised P4 and fusion of enamel folds between cones/conids, which occurs in very worn teeth, are typical of giraffids.
Umbrotherium azzarolii sp. nov.Text-figure 8; Table 3
Umbrotherium azzarolii gen. et sp. nov. from southern Tuscany and Fiume Santo. A–C, IGF 14615, holotype, left maxillary fragment with M2–P3 from Casteani in A, occlusal, B, lingual and C, labial views; × 1. D–F, FS1995#0340, right P3 from Fiume Santo in D, occlusal, E, lingual and F, labial views; × 1. G–I, FS1995#0342, left partial mandible with M3–P3 from Fiume Santo in G, occlusal, H, lingual and I, labial views. J–L, FS1995#0343, right partial mandible with M3–P4 from Fiume Santo in J, occlusal, K, lingual and L, labial views; × 0.5.
| Specimen | Measurements | ol | ln | mw | h |
|---|---|---|---|---|---|
| IGF 14615 | M2–P3 (type) | ||||
| M2 | 23.0 | 21.3 | 21.3 | 14.2 | |
| M1 | 19.8 | 16.2 | 19.4 | 10.9 | |
| P4 | 13.6 | 11.3 | 17.4 | 14.4 | |
| P3 | 13.4 | 12.4 | 17.2 | 12.3 | |
| FS1994#0461 | M3 | 24.25 | 22.0 | 34.0 | 21.64 |
| IR#63976 | M3 | 22.7 | 23.0 | 20.7 | 24.3 |
| IR#63971 | M3 | 23.5 | 24.9 | 24.8 | 28.3 |
| IR#63975 | M3 | 21.8 | 20.0 | 25.0 | 26.0 |
| SN | M3 | *22 | *22 | 24.4 | |
| SN | M2 | *23 | *23 | 27.0 | |
| FS1995#0312 | M2 | 22.6 | 23.0 | 23.5 | 25.6 |
| FS1994#0437 | M2 | 25.8 | 23.0 | 24.0 | 24.2 |
| IR#63972 | M2 | 24.0 | 23.2 | 25.6 | 26.5 |
| FS1995#0311 | M1 | 21.2 | 20.4 | 21.3 | 17.6 |
| FS1995#0313 | M1 | 22.9 | 20.5 | 22.0 | 8.6 |
| FS1995#0310 | P2 | 12.0 | 12.7 | 14.4 | 22.3 |
| FS1995#0056 | M3 | 33.0 | 32.0 | 15.0 | 22.0 |
| FS1995#0140 | M3 | 31.3 | 31.6 | 15.0 | 16.3 |
| FS1995#0126 | M3 | 28.5 | 28.0 | 13.4 | 26.4 |
| FS1995#0210 | M3 | 29.2 | 29.5 | 15.6 | 11.1 |
| FS1994#0342 | M3–P2 | *106 | |||
| M3–M1 | 62.6 | 66.5 | |||
| P4–P3 | 37.9 | ||||
| M3 | 28.0 | 28.7 | 14.7 | 11.9 | |
| M2 | 19.8 | 19.4 | 14.2 | 8.4 | |
| M1 | 16.2 | 16.2 | 13.2 | 3.2 | |
| P4 | 15.0 | 13.6 | 12.0 | 10.4 | |
| P3 | 11.2 | ||||
| FS1994#0343 | M3–P4 | ||||
| M3–M1 | 66.2 | 71.8 | |||
| M3 | 28.5 | 28.0 | 13.4 | 26.4 | |
| M2 | 22.2 | 21.1 | 14.6 | 16.0 | |
| M1 | 17.7 | 17.7 | 14.8 | 12.7 | |
| P4 | 15.2 | 14.4 | 11.4 | 16.9 | |
| IR#63967 | M3 | *26 | *28 | 27.3 | |
| M2 | 22.3 | *23 | |||
| IR#63968 | M3 | 27.0 | 28.0 | 14.5 | 21.7 |
| M2 | 22.6 | 19.6 | 21.5 | ||
| IR#63969 | M3 | 28:4 | 29.4 | 15.3 | 30.1 |
| IR#63974 | M2 | 26.1 | 21.8 | 16.0 | 28.0 |
| IR#64468 | M2 | 22.0 | 22.0 | 14.1 | 15.0 |
| FS1995#0256 | M2 | 25.0 | 23.0 | 15.6 | 20.6 |
| FS1995#0449 | M2 | 23.9 | 21.3 | 12.9 | 22.6 |
| FS1995#0448 | M2 or M1 | 21.3 | 20.6 | 15.4 | 11.5 |
| FS1995#0208 | M1 | *22 | *20 | *14 | *11 |
| FS1995#0618 | M1 | 19.6 | 18.7 | 15.5 | 12.0 |
| FS1994#0126 | P2 | 8.4 | 9.8 | 6.0 | 9.3 |
| IR64472 | P2 | 10.1 | 10.9 | 8.4 | 12.1 |
- ol, occlusal length; ln, length at the neck; mw, maximum width; h, height at metastylid/mesostyle; *, inferred measurement on slightly broken or damaged elements.
1888 Antilope (Palaeoryx?) sp. Weithofer, p. 365.
1889 Antilope (Palaeoryx?) sp. Weithofer, pp. 57, 62.
1912 Antilope (Palaeoryx?) sp. n. Del Campana, p. 212, pl. 18.
1976 Umbrotherium azzarolii Hürzeler and Engesser, p. 334 (nomen nudum).
Derivation of name. After Prof. Augusto Azzaroli, palaeontologist and Emeritus Professor at the University of Florence.
Holotype. IGF14615, an upper left series with P3–M2, housed in the Natural History Museum, Geology and Palaeontology section, University of Florence (Text-fig. 8A–C).
Locality, horizon and age. Casteani (southern Tuscany), faunal assemblage V1; late Miocene, late Tortonian, late Turolian.
Other referred material. Fiume Santo, north-western Sardinia; see below.
Diagnosis. As for the genus.
Description
Upper dentition. There are no more features to report in addition to those already listed in the diagnosis apart from a strong concavity in the lingual walls, which is more evident in less worn teeth.
Lower dentition. Incisors show an asymmetric outline (e.g. 1995#0195). D4 has a protruding metasylid and two well-developed interlobal columns. P2 simple, short and large, with a triangular outline. P3 is not molarised and lacks a metaconid, fusion between hypoconid and entoconid occurs late, entoconid joins protoconid, paraconid and parastylid well developed (Text-fig. 8). P4 is molarised, with metaconid in the form of an anterior–posterior wall, which joins the paraconid and lacks the transverse connection to the protoconid. Entoconid and hypoconid are two obliquely orientated crests. A deep groove in labial side divides P4 into two unequal parts, anterior and posterior lobes, with the posterior one being strongly reduced. On lingual side, a less deep groove occurs (Text-fig. 8). P4 and P3 are both characterised by having a labially sharp hypoconid. Lingual walls of molars obliquely orientated, in particular the anterior one. In M3 the second and third lobes are separated by a clear step; the hypoconulid is high but narrow (Text-fig. 8). Ectostylids are absent; a feeble cingulum occurs only on anterolabial side of lower molars. The mandible is characterised by a fairly thick horizontal ramus (Text-fig. 8).
Comparisons and discussion. A substantial part of mammalian dental remains at Fiume Santo belongs to this puzzling form. In the preliminary notes of Cordy and Ginesu (1994) and Cordy et al. (1995) the occurrence of two giraffid species (differing in size) was reported. However, morphometric analysis of the material indicates the presence of only one form (Table 3; Text-fig. 9); some minor differences in size between dental elements may reflect sexual dimorphism.
Plot of length at the neck and maximum width of M3s of Umbrotherium azzaroli from Fiume Santo and some Miocene giraffid samples. Original data for Decennatherium pachecoi and Bigerbohlinia schaubi from Los Valles de Fuentiduena and Piera respectively, housed in the Palaeontological Institute ‘M. Crusafont’ in Sabadell; data for Decennatherium cf. D. pachecoi from Ravin de la Pluie and Nombrevilla are from Geraads (1979); data for Injanatherium arabicum from Al Jadidah are from Morales et al. (1987); data for Samotherium boissieri and Paleotragus rouenii are from Geraads (1978).
The advanced morphology of P4 is comparable to that which characterized giraffid species of late Miocene age referred to the subfamilies Giraffinae and Sivatheriinae, while that of P3 is very peculiar. The molarised P3, with a highly developed metaconid joined to the paraconid, seems to be an apomorphy in the Giraffinae (as in Paleotragus and Samotherium; Geraads 1978, 1979, 1986), while it is not recorded in Sivatherinae. Umbrotherium has a non-molarised lower third premolar (Text-fig. 8) lacking a metaconid; this could preclude a link to Paleotraginae, because it is untenable in the context of an evolutionary simplification process in the Tusco-Sardinian palaeobioprovince. On the other hand, a primitive P3 is present in Decennatherium pachecoi (Crusafont Pairó 1952; Morales and Soria 1981; pers. obs.) from the Vallesian site Los Valles de Fuentidueña (MN9) in Spain, and in Helladotherium from Greece.
Biometric comparison of Umbrotherium azzarolii shows that it is comparable in size to the medium-sized giraffids, such as the late Miocene Paleotragus (Text-fig. 9). However, this does not help to clarify phyletic relationships because the reduction in size may be a result of the evolution of Umbrotherium within the insular domain of the Tusco-Sardinian palaeobioprovince.
Family BOVIDAE Gray, 1821
Remarks. As reported above, the taxonomic diversity of bovids from Fiume Santo is fully comparable with that of assemblages V1 and V2 from continental Italy. Despite the fact that bovids are well represented and diverse at sites in southern Tuscany, only two genera have been described in detail in the literature: Maremmia and Tyrrhenotragus (Hürzeler 1983; Thomas 1984). Another endemic species was named Etruria viallii by Hürzeler and Engesser (1976) but was not accompanied by a formal description (and hence has been a nomen nudum hitherto), and other forms have been reported in faunal lists as undetermined Bovidae. We describe these ruminants here for the first time, taking into account both Tuscan and Sardinian material.
The reconstruction of morphologies of bovid taxa has been strongly limited by the fact that no dentitions have been found in direct association with skulls and horns, and the isolated horn remains of middle size are difficult to refer to Etruria, which is formally diagnosed below. Indeed, only for Maremmia and Neotragini, the largest and smallest bovids, is the association between dentitions and horn cores possible. The dentitions differ in size, morphology and height of dental crown (hypsodonty). On the whole, the species have hypsodont dentition, with basal pillar and goat-folds usually absent. Maremmia shows the greatest relative values of dental height (see below). Description of the dental features follows the terminology of Gentry (1992).
Tribe NEOTRAGINI (Sclater and Thomas, 1894)
Genus TYRRHENOTRAGUS Thomas, 1984
Type species. Tyrrhenotragus gracillimus (Weithofer, 1888) from Montebamboli, late Miocene, late Tortonian.
Diagnosis (after Thomas 1984). Small neotragini characterized by short and posteriorly orientated horn cores. Very hypsodont dentitions. Upper premolar series reduced, but with all dental elements. In upper molars styles are well developed, in lingual walls lobi are carenated; interlobal columns are lacking. M3 presents a strongly developed posterior lobe. Lower premolar series reduced, but complete. In P4 the metaconid is joined to the entoconid and entostylid. Lingual wall of lower molars slightly undulating, labial lobi carenated, in particular in upper part of wall. The third lobe of M3 is fairly developed.
Tyrrhenotragus gracillimus (Weithofer, 1888)Text-figure 10A–I; Table 4
A–I, Tyrrhenotragus gracillimus from Fiume Santo. A–C, FS1995#0039, right M2 in A, labial, B, lingual and C, occlusal views. D–F, FS1995#0022, right M3 in D, lingual, E, labial and F, occlusal views. G–I, FS1995#0035, right M3 in G, lingual, H, labial and I, occlusal views. J–O, Neotragini gen. et sp. indet. from Fiume Santo. J–L, FS1995#0071, right M1 in J, occlusal, K, lingual and L, labial views. M–O, FS1995#0018, left M3 in M, lingual, N, labial and O, occlusal views. All × 2.
| Taxa | ol | ln | mw | h |
|---|---|---|---|---|
| T. gracillimus | ||||
| FS1995#0039 M2 | 6.3 | 5.6 | 6.3 | 7.5 |
| FS1995#0330 M2 | 8.0 | 7.2 | 7.3 | 9.1 |
| FS1995#0255 M3 | 6.9 | 7.4 | 6.3 | 9.5 |
| FS1995#0333 M3 | 6.7 | 8.1 | 6.5 | 10.1 |
| FS1995#0067 M2 | 6.7 | 6.3 | 4.3 | 5.5 |
| FS1994#0537 M2 | 8.7 | 6.9 | 4.7 | 11.3 |
| FS1995#0275 M2 | 8.0 | 6.5 | 4.6 | 9.3 |
| FS1994#0621 M3 | 13.0 | 13.0 | 5.0 | 12.0 |
| FS1995#0035 M3 | 12.1 | 11.9 | 4.7 | 9.2 |
| FS1995#0022 M3 | 10.1 | 9.6 | 4.2 | 10.4 |
| IR#63964 M3 | 12.5 | 13.9 | 4.9 | 8.4 |
| Neotragini sp. indet. | ||||
| FS1995#0216 P4 | 7.2 | 6.2 | 8.8 | 12.6 |
| FS1995#0332 M1 | 9.3 | 8.6 | 8.7 | 11.0 |
| FS1995#0071 M1 | 9.7 | 8.5 | 8.0 | 12.2 |
| IR#63959 M1 or M2 | 10.6 | 8.3 | 9.0 | 14.2 |
| FS1995#0168 P3 | 7.0 | 6.8 | 3.8 | 3.8 |
| FS1995#0028 M1 or M2 | 9.2 | 8.0 | 5.3 | 6.0 |
| FS1995#0053 M1 or M2 | 8.2 | 8.2 | 5.3 | 7.6 |
| IR#63961 M2 | 9.3 | 7.8 | 4.9 | 12.0 |
| FS1995#0029 M2 | 10.5 | 8.2 | 5.6 | 4.3 |
| FS1995#0018 M3 | 15.0 | 14.7 | 6.0 | 7.7 |
| FS1995#0020 M3 | 13.4 | 13.5 | 5.4 | 9.9 |
| FS1994#0373 M3 | 14.6 | 14.0 | 5.7 | 11.3 |
| FS1995#0577 M3 | 13.6 | 14.7 | 5.7 | 13.0 |
| FS1995#0692 M3 | 12.9 | *13 | 5.5 | 10.4 |
Description. Upper molars are high-crowned. Lower molars have an undulating lingual wall, with well-developed parastylid. M3 has a variably developed third lobe, which frequently has an accessory posterior stylid (Text-fig. 10F).
Comparisons and discussion. Tyrrhenotragus gracillimus is present in both V1 and V2 assemblages in southern Tuscany. It is the smallest bovid species. According to Thomas (1984), dental and horn morphologies suggest that it is referable to the Neotragini.
?Neotragini gen. and sp. indet.Text-figure 10J–O; Table 4
Description. The Fiume Santo assemblage includes some specimens that are similar to Tyrrhenotragus, but larger (Text-fig. 11). They are provisionally referred to an undetermined taxon of Neotragini. Some fragmentary horn cores are small enough to belong to a neotragin representative, but larger than those of Tyrrhenotragus; they are short, laterally inserted on the skull and slightly concave.
Plot of occlusal length and maximum width of M3 of Tyrrhenotragus from Fiume Santo and Baccinello-Cinigiano, and Neotragini gen. et sp. indet. from Fiume Santo.
Comparisons and discussion. The possibility that the remains here referred to Neotragini indet. belong to Tyrrhenotragus males has been considered. Thomas (1981) reported an important size dimorphism for the presumed neotragin species Homoidorcas tugenium from the Ngorora Formation. A size difference of even 30 per cent among the dental remains from Fiume Santo appears too large, however, for this to be attributed to sexual dimorphism in Tyrrhenotragus (Text-fig. 11); these remains are, therefore, tentatively attributed to a new taxon. Analyses of sexual dimorphism in different groups of Recent African antelopes (Jarman 1974; Pérez-Barbería et al. 2002) indicate that in small species (such as those of the Neotragini) sexual dimorphism is reflected by reduced body size.
Material from the Baccinello-Cinigiano Basin stored in the Basel Natural History Museum includes some fragments of isolated teeth and horns that are also attributable to this form.
Tribe ?ALCELAPHINI (Rochebrune, 1883)
Genus MAREMMIA Hürzeler, 1983
Type species. Maremmia haupti (Weithofer, 1888); from Casteani (Grosseto, Central Italy), late Miocene, late Tortonian.
Diagnosis (emended after Hürzeler 1983 and Thomas 1984). Alcelaphini ‘aberrant’, with posteriorly orientated horn cores. Horn cores subcircular to oval in cross section with transverse ridges on anterior surface. Dentitions very hypsodont with poorly developed roots. Upper premolar series reduced, lacking P2. Upper molars with enamel islet on occlusal surface. Very reduced M1and very large M3. Overgrown lower incisors and canine, in the form of enamel lamellas. Lower molars with undulating lingual walls and an occlusal surface with an 8-shaped profile; internal cavities with mediolateral constriction. Lower premolar series strongly reduced, lacking P2 and P3. Among lower molars, M1 relatively small, M3 very large. Posterior part of horizontal ramus of mandible very high; the ascending ramus is vertically orientated.
Maremmia cf. M. lorenziHürzeler, 1983TEXT-FIG. 12., TEXT-FIG. 13.; Table 5
Maremmia cf. M. lorenzi from Fiume Santo. A, FS#63942, right and FS#63943, left horn cores in frontal view. B, FS#63942 in medial view. C–D, IGF 8186, right horn core of Maremmia sp. from Podere Passonaio (Baccinello-Cinigiano Basin) in C, frontal and D, lateral views. Arrows indicate the occurrence of sinuses inside frontals and pedicles. All × 0.5.
Photographs and drawings of Maremmia cf. M. lorenzi from Fiume Santo. A–C, FS1995#0488, left M3 in A, labial, B, lingual and C, occlusal views; × 0.5. D, FS1995#0692, left P4 in occlusal view; × 1. E, FS1995#0692, left unworn P4 in occlusal view; × 1.5. F–H, FS1995#0300, left M3 in F, occlusal, G, labial and H, lingual views; × 1. I–K, FS1994#0344, left mandible with M3–M1 in I, occlusal, J, labial and K, lingual views; × 0.5. L–N, FS1995#0247, right M2 in L, lingual, M, labial and N, occlusal views; × 1.
| ol | ln | mw | |||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| n | min | mean | max | s.d. | n | min | mean | max | s.d | N | min | mean | max | s.d. | |
| M3 | 37 | 15.7 | 21.4 | 26.7 | 3.14 | 38 | 16.1 | 23.1 | 29.8 | 2.87 | 35 | 11.6 | 14.1 | 18.4 | 1.61 |
| M2 | 20 | 15.9 | 18.8 | 27.6 | 2.40 | 20 | 13.5 | 16.9 | 27.5 | 2.97 | 19 | 10.5 | 12.9 | 17.3 | 1.60 |
| M1 | 9 | 14.0 | 16.9 | 19.8 | 1.75 | 9 | 11.0 | 13.1 | 15.5 | 1.63 | 8 | 9.8 | 11.3 | 13.0 | 1.05 |
| P4 | 2 | 10.2 | 10.2 | 10.2 | – | 2 | 7.5 | 8.5 | 9.5 | 1.41 | 2 | 7.5 | 7.6 | 7.7 | 0.14 |
| P3 | 1 | – | 8.0 | – | – | -- | – | – | – | – | 1 | – | 5.0 | – | – |
| M3 | 29 | 21.6 | 25.5 | 29.3 | 1.62 | 27 | 25.5 | 25.5 | 30.6 | 1.54 | 25 | 8.0 | 9 | 10.1 | 0.55 |
| M2 | 19 | 14.1 | 18.8 | 21.5 | 2.2 | 18 | 13.0 | 15.7 | 18.0 | 1.57 | 15 | 7.1 | 7.8 | 9.3 | 2.25 |
| M1 | 13 | 11.3 | 16.5 | 19.3 | 2.1 | 13 | 12.8 | 12.8 | 15.5 | 1.43 | 13 | 6.5 | 7.5 | 8.8 | 0.66 |
| P4 | 5 | 10.9 | 11.8 | 12.9 | 0.75 | 5 | 8.1 | 9.4 | 10.2 | 0.99 | 4 | 5.3 | 5.6 | 6.0 | 0.31 |
Description
The most abundant taxon at Fiume Santo, with 852 identified remains (horns, teeth and postcranial elements).
Horn cores. Horns present an anti-clockwise torsion on the right side, and a kind of hollowing inside the pedicles. This feature, which was judged to be absent from the highly deformed material from the lignitiferous level of southern Tuscany localities (Thomas 1984; Vrba 1997) is clearly present in the Fiume Santo assemblage (Text-fig. 12A–B) and in recently discovered remains from the Baccinello-Cinigiano area (Benvenuti et al. 2001; Text-fig. 12C–D). In some well-preserved specimens transverse ridges on the anterior surface of horn cores are evident. The cores are lateromedially compressed.
Upper dentition. Upper molars with strong styles. M3 is significantly more developed than M2 and M1, and shows a characteristically enlarged metasyle (Text-fig. 13A–C). The premolar series is reduced, the second premolar being absent.
Lower dentition. Molars have an occlusal surface with the highly distinctive 8-shaped profile and a stretched connection between lobes (Text-fig. 13). Lobes have cavities that are constricted centrally. M3 is significantly larger than M2 and M1. The lower premolar row is highly reduced with even the P3 absent. P4 is molarised with a fused paraconid and metaconid, and short hypoconid (Text-fig. 13D–E). The overgrown lower incisors are characterised by the presence of enamel only on the ventral side of the tooth. The horizontal ramus of the mandible progressively increases in height from anterior to posterior, under the M3 (Text-fig. 13J–K); the angular tuberosity for the insertion of the masseter muscle is prominent.
Postcranial bones. The largest bovid remains at Fiume Santo are attributable to this species. The humerus is characterized by a medial condyle with an expanded top and, on the ventral side, a deep bulge.
Comparisons and discussion. The Maremmia material from Fiume Santo closely resembles M. lorenzi from the V2 level at Baccinello-Cinigiano. According to Hürzeler and Engesser (1976) and Hürzeler (1983), Maremmia is represented in southern Tuscany by two species belonging to the same phyletic lineage, Maremmia haupti and M. lorenzi, occurring in V1 and V2 assemblages respectively. They differ in size and in the degree of dental series reduction, with a noticeable increase in upper and lower M3 in M. lorenzi. Thomas (1984) did not agree with the distinction between M. haupti and M. lorenzi, only recognizing M. haupti as a valid taxon. However, biometric comparison of material from Baccinello-V1, Fiume Santo and Baccinello-V2 underlines, in addition to the larger size of the latter two samples, the relatively more developed M3 in the V2 assemblages, which, in our view, justifies their separation at the species level. Moreover, Maremmia from Fiume Santo appears to be slightly larger when compared to the material from the V2 level of Baccinello-Cinigiano (Text-fig. 14). It is therefore referred to Maremmia cf. M. lorenzi.
Plot of length at the neck and maximum width of M3 in Maremmia haupti, M. lorenzi and Maremmia cf. M. lorenzi from Baccinello-Cinigiano Basin levels V1 and V2, and Fiume Santo. The holotype of M. lorenzi n. Bac93 from Trasubbie Creek, stored in Basel Naturhistorisches Museum, is indicated by the arrow.
The origin and evolutionary relationships of Maremmia are controversial. To Hürzeler (1983), the morphology of the horn cores and dentition (e.g. the 8-shaped profile of the occlusal surface and P4 morphology) suggested a link with the Alcelaphini. Thomas (1984) considered the origin of Maremmia to be closely associated with the middle Miocene Afro-Eurasian Caprotragoides, and Vrba (1997) believed that Maremmia is the sister taxon of Caprotragoides. Gentry (1997) considered the ‘puzzling’ European bovid Maremmia haupti to lie close to the ancestry of African Alcelaphini. Although the similarity with Alcelaphini is striking, the oldest known records of these bovids are latest Miocene–Early Pliocene in Africa (Gentry 1980) and, therefore, clearly post-date the existence of the Tusco-Sardinian palaeobioprovince. Hence, the evolutionary relationships of Maremmia remain unresolved.
Subfamily undetermined (?ANTILOPINAE)Tribe undetermined
Genus ETRURIA gen. nov.
1976 Etruria Hürzeler and Engesser, p. 333 (nomen nudum).
Derivation of name. From Etruria, Latin name of the region in Central Italy where the Etruscan civilization developed and the Oreopithecus fauna existed.
Type and only species. Etruria viallii sp. nov.
Diagnosis. Bovid similar to a gazelle in size. Lower molars lack basal pillars, with protruding metastylid, lobi triangular in shape with flat labial wall. Third lobe of M3 fairly well developed and with a central cavity; metastylid disappears with advanced wear stages. Lower premolars short, in particular P2, whose length represents about 18 per cent of length of premolar row; P3 triangular in shape with well-developed parastylid and paraconid; P4 somewhat primitive with an anterior valley between paraconid and metaconid, while metaconid, entoconid and endostylid tend to be fused lingually (Text-fig. 15).
Photographs and drawings of Etruria viallii gen. et sp. nov. from southern Tuscany and Fiume Santo. A–C, Bac1002, holotype, right fragmentary mandible with M3–P2 in A, occlusal, B, lingual and C, labial views; × 1. D–F, FS1995#0027, right M2 from Fiume Santo in D, lingual, E, occlusal and F, labial views, × 1.5.
Etruria viallii sp. nov.Text-figure 15; Table 6
| Taxa | ol | ln | mw | h |
|---|---|---|---|---|
| Casteani | ||||
| Bac 1002 M3–P2 (type) | 67.2 | 66.4 | ||
| M3–M1 | 43.0 | 40.1 | ||
| P4–P2 | 24.1 | 23.1 | ||
| M3 | 15.8 | 16.5 | 7.3 | 15.6 |
| M2 | 13.0 | 10.0 | 7.0 | 12.1 |
| M1 | 11.0 | 9.0 | 7.0 | 9.0 |
| P4 | 9.2 | 7.6 | 6.1 | 9.3 |
| P3 | 8.0 | 6.1 | 6.0 | 8.0 |
| P2 | 4.4 | 4.8 | 4.2 | 5.0 |
| Bac 1003 M3–M1 | 39.4 | 39.3 | ||
| M3 | 18.0 | 18.7 | 9.1 | 5.4 |
| M2 | 11.5 | 10.2 | 8.7 | 4.4 |
| M1 | 9.7 | 8.2 | 7.3 | 1.6 |
| P4 | 7.2 | 6.3 | 6.1 | 2.5 |
| P2 | 6.8 | 6.1 | 4.8 | 3.5 |
| Fiume Santo | ||||
| FS1995#0145 M2 | 12.9 | *11 | *11 | 14.9 |
| FS1995#0136 M1 | 12.7 | |||
| FS1995#0027 M2 | 13.1 | 11.6 | 6.5 | 14.7 |
| IR#64440 M2 | 15.2 | 13.7 | 16.1 | 7.5 |
| IR#64452 M3 | 17.3 | 18.3 | 5.1 | |
| IR#64452 M2 | 10.3 | 10.3 | 3.4 | |
| FS1995#0620 M3 | 18.5 | 17.4 | 7.9 | 12.7 |
| FS1995#0088 P3 | 7.5 | 6.6 | 4.5 | 9.0 |
| FS1995#0014 M3 | 19.1 | 17.5 | 7.6 | 6.6 |
| FS1995#0240 M3 | 18.0 | 17.6 | 11.5 | |
| FS1994#0520 P4 | 11.7 | 9.2 | 5.8 | 14.9 |
1976 Etruria viallii Hürzeler and Engesser, p. 333 (nomen nudum).
Derivation of name. In honour of Prof. Vittorio Vialli (1914–1983), eminent palaeontologist at Bologna University.
Holotype. Bac1002, a right fragmented mandible with complete tooth series P2–M3, housed in the Basel Natural History Museum (Text-fig. 15A–C); for measurements of specimen, see Table 6.
Locality, horizon and age. Baccinello (Grosseto district, southern Tuscany), lignite level, faunal assemblage V1; late Miocene, late Tortonian, late Turolian.
Other referred material. Fiume Santo, north-west Sardinia: see Supplementary Data file (http://www.palass.org).
Diagnosis. As for the genus.
Description. Upper dentition: P4 rectangular in outline (long and narrow); the anterior crista of the hypocone does not join the parastyle; the fragmentary upper molars have poorly developed styles. Lower dentition: the Fiume Santo material adds nothing to the features previously recorded from the type material.
Comparisons and discussion. Etruria viallii has the lowest degree of dental reduction and hypsodonty among the bovids of the Tusco-Sardinian palaeobioprovince. Unlike Maremmia spp. and Bovidae gen. et sp. nov. (see below), the dental roots are well developed. The dental characters described for E. viallii were reported by Gentry (1992) to be distinctive of the Antilopini tribe (e.g. Gazella).
Subfamily undeterminedTribe undetermined
Genus TURRITRAGUS gen. nov.
1918 Gazella hauptiDel Campana, p. 170, pl. 16, figs 4A–C, 6A–C.
1984 Maremmia haupti Thomas, p. 86.
Derivation of name. After the town of Porto Torres in north-west Sardinia, whose surroundings include the industrial complex in which the Fiume Santo site is located.
Type and only species. Turritragus casteanensis sp. nov.
Diagnosis. Bovid of small size, similar to Etruria vialli; with hypsodont dentition, roots very short. Highly reduced dental series, lacking P2 and P2–P3. Upper molars have a well-developed mesostyle, basal pillars are absent. Upper premolars rectangular in outline. Inner islets of upper and lower molars show an evident anterior–posterior compression. Lower molars have an undulating lingual wall and bear a metastylid only at a low level of wear. M3 bears a narrow hypoconulid. P4 not molarised and lacks transverse metaconid crest. Hypoconid separated from metaconid by a groove (Text-fig. 16).
Photographs and drawings of Turritragus casteanensis gen. et sp. nov. from Southern Tuscany and Fiume Santo. A–C, IGF 14631, holotype, left mandible with M3–P4 from Casteani in A, labial, B, lingual and C, occlusal views. D–F, IGF 11746, left mandible with M3–P4 from Casteani in D, labial, E, occlusal and F, lingual views. G–I, IGF 14633, left maxillary with M2–P3 from Casteani in G, occlusal, H, lingual and I, labial views. J–L, FS1995#0091, left M2 from Fiume Santo in J, labial, K, lingual and L, occlusal views. All × 1.
Remarks. The revision of the material from Casteani and other sites in southern Tuscany housed in the Natural History Museum of the University of Florence, and analysis of the Fiume Santo fossils allows us to recognise another new middle-sized to small bovid species. The remains from Casteani were ascribed to M. haupti by previous authors.
Turritragus casteanensis sp. nov.Text-figure 16; Table 7
| Taxa | ol | ln | mw | h |
|---|---|---|---|---|
| Casteani | ||||
| IGF 14633 M2–P3 | ||||
| M2 | 12.0 | |||
| M1 | 10.8 | *8 | 10.4 | 15.0 |
| P4 | *8 | 6.3 | 8.6 | 15.7 |
| P3 | 8.0 | 7.0 | 6.4 | 10.5 |
| IGF14631 M3–M1 | 46.1 | 47.4 | ||
| M3 | 19.2 | 20.0 | 8.0 | 25.0 |
| M2 | 13.7 | 12.8 | 7.3 | 10.6 |
| M1 | 10.8 | 7.9 | 6.7 | 11.3 |
| P4 | 9.6 | 8.3 | 4.3 | 12.3 |
| IGF11746 M3–M1 (type) | 44.0 | 48.0 | ||
| M3 | 17.8 | 19.6 | 6.7 | 30.3 |
| M2 | 13.7 | 12.5 | 7.3 | 8.4 |
| M1 | 10.3 | 9.0 | 7.0 | 7.0 |
| P4 | 8.8 | 8.3 | 4.4 | 7.0 |
| Fiume Santo | ||||
| IR#63950 M3 | 15.0 | 16.3 | 12.7 | 28.4 |
| FS1995#0091 M2 | 14.4 | 13.0 | 7.2 | 22.0 |
| IR#63937 M2 | 14.5 | 14.1 | 7.5 | 22.6 |
Derivation of name. After Casteani, type locality of the species.
Holotype. IGF 14631, a left mandible with complete tooth series P4–M3, housed in the Natural History Museum, Geology and Palaeontology section, University of Florence (Text-fig. 16A–C); for measurements, see Table 7.
Referred material from type locality. IGF 11746, left mandible with P4–M3 (Text-fig. 16D–F); IGF 14633, upper series P3–M2 (Text-fig. 16G–I) (Table 7); for material from the Fiume Santo site, see Supplementary Data file (http://www.palass.org).
Locality, horizon and age. Casteani (Grosseto district, southern Tuscany), Faunal assemblage V1; late Miocene, late Tortonian, late Turolian
Diagnosis. As for the genus.
Comparisons and discussion. The material from Casteani indicates another bovid species which, like Maremmia, has a markedly reduced dental series. The shortening of the dental series and the highly hypsodont teeth make Turritragus casteanensis, by comparison with Maremmia, very derived for its chronological position. A reduction of the premolar series, although with all elements present, is recorded in Aragoral mudejar from the Upper Vallesian of Spain, which is considered to be a primitive representative of the Caprinae. Very hypsodont bovid species were described by Robinson (1986) from late Middle Miocene deposits in Tunisia, and regarded as primitive representatives of the Rupicaprini. The absence of cranial parts and horn cores in T. casteanensis prevents precise determination of its systematic affiliation.
Very few isolated highly hypsodont teeth referable to T. casteanensis are present in the Fiume Santo collection (Text-fig. 16J–L). They do not add anything new to the diagnostic features recorded from the type material from Casteani.
Remarks on bovid species from the Tusco-Sardinian palaeobioprovince
As a result of our analysis it is possible to compile a list of dental differences, other than size, for distinguishing between the various species (Table 8; Text-fig. 17). The most significant features and profound changes in the masticator apparatus are found in Maremmia: a strong reduction of the premolar series and enlargement of lower and upper third molars, which indicates a tendency to chew posteriorly closer to the fulcrum of the mandible; and a shortening and deepening of the horizontal mandibular ramus. The diagnostic dental features of Maremmia also include a high degree of hypsodonty. Hypsodonty has been widely used as an indicator of feeding preferences in many mammal categories and is considered to be a functional adaptation to an abrasive diet (Janis 1995; Fortelius et al. 2002) and climatic parameters, in particular aridity (Damuth and Fortelius, 2001; Fortelius 2003). The calculated values of this index for the bovids of the Tusco-Sardinian palaeobioprovince (Table 8; h = height of unworn M2 divided by the length according to Fortelius et al. 2002; Fortelius 2003) indicate that it can be classified as a hypsodont species (h > 1.2) and suggests the occurrence of open, dry environments (for the inverse connection between hypsodonty and humidity, see Damuth and Fortelius 2001; Fortelius et al. 2002, 2003; Fortelius 2003; Eronen and Rook 2004; Eronen 2006).
| Taxa | T. gracillimus | Etruria viallii | T. casteanensis | M. cf. lorenzi |
|---|---|---|---|---|
| Characters | ||||
| Deepening of horizontal ramus, beneath molars | moderate | moderate | intermediate | strong |
| Premolar reduction | moderate1 | moderate1 | yes2 | yes2 |
| Hypsodonty | high (1.6) | moderate (1.3) | high (>1.7*) | very high (2.1) |
| Molarization of P4, with paraconid-metaconid fusion | no | no | no | yes |
| Stretched connection between lobes of lower molars | no | no | no | yes |
| Metasylid development | early | intermediate | very early | very early |
| Development of posterior lobe of M3 | intermediate | high | low | low |
| Compression of central cavities of molars | constricted anteroposteriorly | constricted anteroposteriorly | constricted anteroposteriorly | constricted centrally |
- 1, P2 is short but always present; 2, P2 is lacking; *, only one slightly worn M2 was available.
Plot of occlusal length and maximum width of bovid M3s from Fiume Santo.
With regard to the postcranium, 83 bovid remains have been identified in the Fiume Santo collection. The most common and best preserved skeletal element is the astragalus.
Specific characters, e.g. articular facets or areas for tendon/ligament insertion, are difficult to discern on the bones because the material is highly weathered; however, after a preliminary analysis, some differences in the shape of the tuber calcanei of the calcaneus and in the proximal articulation of the metatarsals suggest differences in locomotor behaviour among the various bovid taxa, and confirms the occurrence of more than one middle-sized species (Etruria and Turritragus). Two metacarpals (both recorded as Bac1013 in the collection of the Basel Natural History Museum; Text-fig. 18) from the Baccinello-Cinigiano Basin are characterised by a strong shortening (total length, c. 60 mm; proximal width, c. 14 mm; distal width, c. 16 mm). These dimensions and proportions allow their exclusion from Neotragini and Maremmia; it is possible that they are referable to Etruria or Turritragus casteanensis. The shortening of metapodials, which reflects an adaptation to rough, rocky ground, is well documented among bovids and frequently found in Caprinae and more primitive forms (Alcalá and Morales 1997). The strong correlation between size and morphology of distal limb bones and habitat is well established in Recent bovid communities. Limb morphology reflects speed, jumping capacity and ground conditions. However, in the case of the Fiume Santo bovids, more comprehensive ecological characterisation will only be possible following analysis of more abundant postcranial remains.
Bovidae indet., Bac1013, right metacarpal from Baccinello, anterior view; × 1.
The faunal assemblages of the late miocene tusco-sardinian palaeobioprovince
The faunal succession V0–V1–V2 within the sedimentary succession across the various southern Tuscan basins (OZF Oreopithecus Zone faunas in Bernor et al. 2001) is a classic example of a chronofauna according to the concept of Olson (1952), which aims to indicate the persistence and continuity through time of evolving animal assemblages under changing environmental conditions. The Fiume Santo vertebrate assemblage offers the opportunity to compare local populations belonging to the same time span as the V2 assemblage, within the Tusco-Sardinian chronofauna.
Overall, the Fiume Santo fauna is very similar to the V2 assemblage (Montebamboli and Baccinello V2 faunas). Nevertheless, there are some differences, such as the slightly derived features in the suid Eumaiochoerus cf. E. etruscus and in the antelope Maremmia cf. M. lorenzi, which could be a result of environmental differences, and perhaps also related to slight chronological differences. The dominance of bovids with a high degree of hypsodonty in the mammal assemblage from Fiume Santo suggests wide open spaces; forest-adapted forms are under-represented. The occurrence and/or abundance of taxa generally in the so-called Oreopithecus assemblages seem to have been strongly influenced by environmental conditions. The Fiume Santo time-equivalent (or evolutionary-equivalent) V2 assemblages in Tuscany vary in composition. While the suid Eumaiochoerus and the primate Oreopithecus are abundant at Montebamboli, both are extremely rare (Oreopithecus) or absent (Eumaiochoerus) in assemblages from the Baccinello-Cinigiano Basin (Rook et al. 1999a). Such differences are easily explained if the sedimentary and palaeoenvironmental context of the sites is taken into account: swampy and marshy-margined lake at Montebamboli versus flood-dominated fluvio-lacustrine sediments near an uplifted delta plain in the Baccinello-Cinigiano Basin (Benvenuti et al. 1999, 2001).
Zoogeography of tusco-sardinian late miocene bioprovince
A zoogeographic study of the Fiume Santo faunal assemblage was carried out by means of the Genus Faunal Resemblance Index (GFRI) between the Tuscan OZF (Oreopithecus faunal assemblages: Baccinello V0–V2, Montebamboli, Casteani, Montemassi, Ribolla, Serrazzano, Baccinello V0–V2) and Fiume Santo together with some European localities across the MN10–13 time interval. For this purpose, we based our analyses on the Neogene Old World (NOW) Database. This database was released to the public in 1996 and since then the number and geographic range of localities have been increased substantially. The latest public NOW dataset can be freely downloaded from the website http://www.helsinki.fi/science/now/.
According to the procedure followed by Bernor et al. (2001, 2004) we calculated both Dice and Simpson GFRIs. Dice’s index is the most commonly used faunal resemblance index (especially by zoologists), while Simpson’s index has long been used by palaeontologists and additionally adjusts for differences in sample size, which is a problem in the present analysis (see Bernor et al. 2001 and 2004 for a discussion of the method). The GFRI graph (Text-fig. 19) shows, as expected, very high values of resemblance (100 per cent) for the Fiume Santo and southern Tuscany assemblages, while the GFRI is very low for the rest of the faunas. Apart from the obvious conclusions regarding the endemic character of the Tusco-Sardinian palaeobioprovince, it is interesting to note that the GFRI index is 0 for the late Miocene faunal assemblages of Greece (Ouranopithecus Faunal Zone), meaning that no taxa (at generic level) are in common between the two faunas. Very low values characterise comparisons between the Oreopithecus Faunal Zone and other, older (Can Llobateres) and younger (Pikermi; Baccinello V3; Maramena) assemblages, owing to the common occurrence of a few genera of micro-mammals and carnivores (Mustela, Huerzelerimys, Kowalskia, Parapodemus).
Genus-level Faunal Resemblance Index (GFRI): pair-wise comparison between the set of localities under consideration and the Tuscany ‘Oreopithecus FZ’: A, Can Llobateres; B, Ouranopithecus FZ; C, Fiume Santo; D, Pikermi; E, Baccinello V3; F, Maramena; black bars, Dice index; white bars, Simpson index. Comparative data updated from Fortelius (2005).
It has been suggested that the survival in the latest Miocene (MN11–12) Tyrrhenian area of the extremely derived ape Oreopithecus bambolii (in contrast to the extinction pattern of European mainland hominoids since MN10), was linked to a particular combination of palaeogeographic and palaeoclimatic conditions. The terminal Miocene palaeogeographic changes in the Tyrrhenian area, a consequence of the intensive Messinian tectonism which produced the Apennine mountain chain, led to the end of the Tusco-Sardinian palaeobioprovince and the extinction of the Oreopithecus bambolii insular fauna (Rook et al. 2000).
The origin of the endemic vertebrates in the Tusco-Sardinian province has been discussed in detail in the literature. On the basis of the presence of the non-endemic murid Huerzelerimys vireti (Engesser 1989) in the older levels of the Baccinello-Cinigiano Basin (V0 assemblage), it appears that the entry of the ancestors to the endemic taxa may be dated close to the beginning of the Turolian (MN 11 unit; Mein 1999). However, because of the limited similarities, the correlation of endemic faunal successions with the late middle/early late Miocene continental European and African faunal successions is problematic. Undeniably, the systematics and, especially, estimates of phylogenetic relationships of insular taxa are affected by endemic evolutionary modification. In the case of long-lasting geographic isolation, the endemic changes could be so great as to render the identification of ancestral taxa very difficult, as it becomes increasingly hard to establish which characters are inherited from ancestral taxa and which are derived as a result of endemic evolution. The phylogenetic affinities of some of the mammals, such as Oreopithecus and the suid Eumaiochoerus, lie mainly with species from the European continent (Moyá-Solá and Köhler 1997; Bernor et al. 2001); however, those of other taxa, in particular the bovids, are more contentious and, as yet, poorly understood, although it has been suggested previously that Maremmia and Tyrrhenotragus were African in origin (Hürzeler and Engesser 1976; Hürzeler 1983; Thomas 1984; Abbazzi in press).
The Tusco-Sardinian palaeobioprovince chronofauna became extinct at the end of latest Miocene when the youngest endemic faunal assemblages were replaced by the dispersal of a completely renovated mammal fauna (the so called Bacinello V3) made up of continental taxa with fully continental European affinities (Hürzeler and Engesser 1976; Bernor et al. 2001) and different environmental tracking (Eronen and Rook, 2004), pointing to a renewed and definite biogeographical connection with Europe. The ultimate cause of this faunal turnover is attributed to the intense tectonic activity that affected the area during latest Messinian times and led to the uplift of the Apennine mountain chain (Boccaletti et al. 1990); the slopes of the newly emerged Apennines constituted a wide pathway for the dispersal of mammal communities. This faunal change marks the time when the Corso-Sardinian Massif was definitely isolated from southern Tuscany by the opening of the Tyrrhenian Sea and southern Tuscany became fully connected to the Apennine chain (Benvenuti et al. 2001; Rook et al. 2006).
Conclusions
The rock succession at Fiume Santo contains the richest and best fossil record of the youngest stage in the evolution of the chronofauna that in latest Miocene times (late Tortonian) characterised the North Tyrrhenian area, the so-called Tusco-Sardinian palaeobioprovince. During the latest Miocene, the North Tyrrhenian area, a region presently corresponding to southern Tuscany and the Sardinian Massif, was occupied by a complex of large islands characterised by endemic vertebrate populations. Morphological features, adaptations and trophic structure (e.g. low diversity and scarce carnivore taxa), mainly in the mammal faunas, attest to a long period of isolation from the continental palaeobioprovinces of the Mediterranean and Central Europe. The analysis of the Sardinian fossils and their comparison with material previously recovered from various localities in Tuscany, allowed us to re-evaluate the latter and to describe three new endemic taxa among the ruminants. Tusco-Sardinian vertebrate assemblages are characterized through time and across geographically separated areas (Tuscany vs. Sardinia at the level of V2 faunas), and consideration is given to zoogeographic patterns and evidence of evolution of an endemic chronofauna in insular conditions, and its extinction during the latest Miocene times.
Acknowledgments
Acknowledgements. This work was carried out under an agreement between the ‘Soprintendente ai Beni Archeologici per le Province di Sassari e Nuoro’ and the Vertebrate Palaeontology Group of the University of Florence. We are indebted to the Sassari Soprintendente, Dr Francesco Nicosia for the support. We thank all the personnel of the ‘Fiumesanto’ thermoelectric power station for their help during all phases of field work, in particular the former and present directors of the power station, Drs Francesco Capriotti and Marco Bertolino respectively, and the person responsible for external relations, Mr Mario Abeltino. For facilitating access to fossil collections in their care, we thank Burkart Engesser and Christian A. Meyer (Naturhistorisches Museum, Basel), Elisabetta Cioppi and Menotti Mazzini (Natural History Museum, Geology and Palaeontology Section, University of Florence), Walter Landini and Chiara Sorbini (Museo di Storia Naturale e del Territorio di Calci, Pisa), Salvador Moyá-Solá and Meike Köhler (Barcelona), Bernard Battail and France de Lapparent de Broin (Musée national d'Histoire naturelle, Paris). Among numerous people who have helped greatly with suggestions and discussion we mention Marisa Arca and Caterinella Tuveri (Nuoro), Raymond L. Bernor (Washington, DC), F. Clark Howell (Berkeley), Jorge Morales and Jan van der Made (Madrid), Salvador Moyá-Solá and Meike Köhler (Sabadell) and Maria Rita Palombo (Rome). The editorial work of David J. Batten (Manchester) is appreciated. Our research at Fiume Santo is part of a wider research program on the evolution of late Neogene mammal faunas co-ordinated by LR at the University of Florence. Field work at Fiume Santo was made possible thanks to the support of the National Geographic Society (grant #7484-03 to LR), the RHOI program at University of Berkeley (project NSF-BCS-0321893), and the logistic and economic support of ENDESA Italia. For the purposes of this paper, LA dealt with Bovidae and Giraffidae, MD with Herpetofauna, GG with Suidae, and LR with Primates, Carnivora and zoogeography. All contributed equally to the general discussion and concluding sections.
