Abstract: The infaunal irregular echinoid, Tithonia oxfordiana, is described and compared to congeneric species previously described from Upper Jurassic and lowermost Cretaceous strata. This new species characterizes a monospecific echinoid assemblage, which occurs only in some places where deep-marine middle Oxfordian deposits are exposed in south-east France. Specimens are closely packed and clearly concentrated at the top of small carbonate chemoherms; a close connection of the echinoids with the emission of reduced chemicals, which were oxidized by chemoautotrophic bacteria, is highly probable. Based on general test shape and plate architecture, T. oxfordiana probably was a deposit feeder on chemosynthetic organic matter produced by such bacteria. In view of the fact that T. oxfordiana is the sole species of the Jurassic genus Tithonia known from Oxfordian strata, it is postulated that chemoherms possibly acted as refugia for these peculiar echinoids, which have an episodic record between the Callovian and Valanginian.
A lmost three decades ago, unexpected and spectacular, modern ‘oases of life’ were discovered in deep-sea spreading centres (Lonsdale 1977, 1979; Corliss et al. 1979). Such occurrences are now well documented in very different geological contexts and may be schematically related to hydrothermal vents in oceanic rises, as well as to cold seeps in passive and active margins. In both cases, the emission of reduced chemicals (mainly H2S and CH4) and their oxidation by chemoautotrophic bacteria allow a chemosynthesis of organic matter. Bacteria commonly are hosted by symbiotic macro-organisms, such as tube worms and bivalves (mainly giant clams and mussels). Such localized primary production acts as the starting point of a more or less complex trophic web, which explains the oasis aspect of these deep-sea assemblages (Jannasch 1984; Cavanaugh 1985; Hessler 1985; Grassle 1986; Laubier 1988; Tunnicliffe 1988, 1991, 1992a; Carney 1994; Hessler and Kaharl 1995). Such peculiar ecosystems, based only on chemosynthesis, are now commonly recognized in the deep sea (Tyler et al. 2003).
Most fossil chemoherms, like their modern equivalents, contain abundant bivalves: lucinids in the fossil record and vesicomyids in present-day settings. Because of this similarity, fossil lucinids from chemoherms very probably were symbiotic with chemoautotrophic bacteria, at the base of the trophic web. But many questions still remain regarding the rarer organisms that co-occur with these bivalves. An interesting case is illustrated here using echinoderms. They are often rare and have never been recorded previously from fossil or modern seep communities. Here, we describe a unique fossil locality at which occurs a monospecific echinoid assemblage. The aim of the paper is threefold; first of all, this uncommon species is named and described, and its systematic position is clarified. Second, its ecology and place in the palaeoecosystem are outlined, and third, the probable role of sea floor seeps in the dispersal of this taxon and, more generally, of marine benthos is discussed.
Palaeogeography and tectonics
The chemoherms that yielded the echinoids are exposed near Beauvoisin (Drôme, south-east France; see Gaillard et al. 1985; Rolin et al. 1990) (Text-figs 1, 2). During the Jurassic, south-east France corresponded to a deep, subsiding basin (Baudrimont and Dubois 1977; Debrand-Passard 1984), related to the opening of the Ligurian Tethys Ocean (Lemoine 1985; Lemoine et al. 1986). Chemoherms occur in the deepest, central area of the basin, linked to faulting (Gaillard et al. 1985). Indeed, chemoherms are more or less situated along major faults that were probably active during the deposition of the Bathonian–middle Oxfordian Terres Noires Formation (Lemoine et al. 1982) (Text-fig. 1). A direct relationship between some pseudobioherms and small, active synsedimentary faults has been demonstrated at Beauvoisin (Gaillard and Rolin 1988). Clearly, faults favoured the circulation of fluids that fed the seeps. However, the nature and origin of these fluids are unknown. They may have a deep origin, related to crustal faults (Lemoine 1985; Lemoine et al. 1986), an intermediate provenance related to halokinesis (Mascle et al. 1988) or a superficial source related to simple undercompaction of the thick Terres Noires Formation that is rich in organic matter. The last-named scenario is the most probable; it would have led to methane generation and expulsion (Hedberg 1974). In this respect, a convincing comparison has been proposed between Beauvoisin and the modern-day passive margin of Congo-Angola, where pockmarks occur (Gay 2002).
At the Beauvoisin site, chemoherms occur in a thick (i.e. up to 2500 m) terrigenous unit referred to as the Terres Noires Formation and at the base of the overlying alternating limestone/marl unit named ‘Alternance argovienne’ (Text-fig. 3). Abundant ammonites have enabled a detailed biozonal scheme to be established, which is in accord with a synthetic overview established by the ‘Groupe Français d’Etude du Jurassique’ for the Jurassic of western Europe and the Mediterranean (Cariou et al. 1997). The Terres Noires Formation comprises dark, well-stratified marls that range in age between the Bathonian and middle Oxfordian. Only the uppermost portion of this unit is exposed at Beauvoisin; this corresponds to the upper lower Oxfordian and the lower middle Oxfordian (Bourseau 1977). Chemoherms are abundant here (about 40 sedimentary bodies), large (up to 15 m in height) and well exposed. The overlying limestone/marl alternation corresponds to the upper middle Oxfordian (transversarium Zone) and to a very homogeneous sedimentation across the entire basin (Gaillard et al. 1996, 2004). The latter comprises only rare, small chemoherms, and those yielded echinoids are all situated at the same level, between beds R3 and R4, in the schilli Subzone (Text-fig. 3).
These chemoherms are columnar or lens-like carbonate bodies (>90 per cent CaCO3), interbedded in limestone/marl alternations. Their morphology and petrography correspond to the general characteristics of previously described ‘pseudobioherms’ (Gaillard et al. 1985, 1992; Rolin et al. 1990). At Beauvoisin, only a composite chemoherm is well exposed; this has been studied in detail (Text-figs 4, 5). At the base, small columnar carbonate bodies occur, about two metres in height and with only rare infaunal bivalves. They are lateral equivalents of beds R3c, R3d and R3e of Gaillard et al. (1996) (see Text-figs 3–5). These small chemoherms are overlain by a larger, lens-like upper chemoherm that is richer in fossils. The precise stratigraphical position of this carbonate lens is less clear, because of poor outcrop conditions and closely spaced faults, but it is probably lateral to or immediately overlain by bed R4. A relationship between the genesis of this complex chemoherm and faulting has been demonstrated (Gaillard and Rolin 1988). Echinoids are very abundant, yet restricted to the top of the upper carbonate lens (>200 specimens collected on a small surface area, c. 5 m2). At Beauvoisin, only two other chemoherms occur in the same stratigraphical range, and they are also topped by similar echinoid assemblages. In summary, echinoids are confined to a certain stratigraphical level and occur on the top of chemoherms only.
Material and Methods
As far as macrofauna is concerned, the assemblage is composed of benthic (85 per cent) and nektonic biota (15 per cent; 400 specimens counted). Remains of nektonic biota consist mainly of ammonite shells (54 per cent), with fewer belemnite rostra (35 per cent) and aptychi (11 per cent). Benthic species include predominantly echinoderms and molluscs. Echinoid tests are very abundant (63 per cent), followed by bivalve shells (36 per cent) and rare gastropods (1 per cent). Some holothuroid sclerites and sponge spicules are also present. Although echinoids are rare in the Oxfordian basin of south-east France as well as in the peculiar environment of chemoherms, here they clearly predominate and are monospecific. In contrast, bivalves correspond to species commonly found in chemoherms (Rolin et al. 1990), all being infaunal lucinids that have recently been described as Beauvoisina carinata (Kiel et al. 2010), exhibiting a large size range of 10–90 mm. Whereas echinoids are concentrated at the top of the chemoherms, bivalves also occur within the chemoherms.
Microfaunal elements include abundant small benthic foraminifera, rarer planktonics (Globuligerina) and ostracods. Benthic foraminifera are represented mainly by spirillinids (Spirillina), nodosariids (Lenticulina, Dentalina, Nodosaria and Citharina) and agglutinated taxa (Textularia, Trochammina and Haplophragmoides) (Bouhamdi 2000; Bouhamdi et al. 2000, 2001).
Echinoids are generally well preserved, and test surfaces are not abraded and not normally encrusted by epibionts (e.g. rare serpulids). In most cases, tests are complete, but often slightly deformed by compaction; fragments can also be observed in the matrix (Text-fig. 6A). Tests are filled with a mud that is similar to the matrix in which they are embedded; pellets may be abundant (Text-fig. 6B). Spines, which are very small, are relatively abundant in the matrix around the tests. Most specimens collected had weathered out in outcrop, but some tests have also been observed in the rock column, without any preferred orientation. This may have resulted from bioturbation, similar to bivalves in chemoherms (Rolin et al. 1990). Clearly, echinoids are rarely, if ever, transported and thus are probably found in their original biotope.
No specimens smaller than 17 mm test length have been collected, which means that the available material may correspond to either preadult (17–25 mm) or adult individuals (25–35 mm). A univariate analysis of test length shows two well-separated peaks (Text-fig. 7). The absence of juveniles is very common in fossil echinoid populations and reflects palaeodemographic structures. Indeed, juveniles and adults of the same species usually live separated (Néraudeau 1991), the former being restricted to calmer and more stable environments than those inhabited by adults. Consequently, when a very localized population perishes in a sudden trophic, thermic or sedimentary event, only a very precise range of test size typifies the fossil assemblage (Néraudeau 1989, 1991, 1995). All materials are contained in the collections of Université de Lyon 1 (FSL).
In addition to direct observations of test shape and architecture, biometric and biostatistical analyses have been performed on the 105 best-preserved tests (of 210 specimens) collected from the top of the same chemoherm. The remaining specimens are weathered, broken or compacted. The biometric approach used classic dimensions measured on the test: antero-posterior length (LL), maximum width (LA), maximum height (HT), interval between trivium and bivium (BI), position of maximum width (LM), apex (AP) and peristome (PS) with regard to the front of the test, and position of the periproct with regard to the maximum height of the test (PP). Several percentage ratios were calculated to minimize the size effect and to illustrate the proportional variations of test shape: ambitus (LA/LL), elevation (HT/LL), trivium–bivium gap (BI/LL), relative position of apex (AP/LL), maximum width (LM/LL), peristome (PS/LL) and periproct (PP/HT) and test inflation ratio (2XHT/LL + LA). These dimensions and ratios enable calculation of parameters of univariate analysis, mean (M) and standard deviation (S), and bivariate analysis, correlation factor (R).
Discussion. Morphologically, the Middle to Late Jurassic – earliest Cretaceous genera Tithonia and Metaporinus are very close, which led some authors to consider the former to be a subgenus of Metaporinus (Lambert and Thiéry 1909–1925; Beurlen 1934; Mortensen 1950). However, on the basis of apical disc structure and general test morphology, Jesionek-Szymańska (1963) clearly distinguished Metaporinus from Tithonia and awarded the latter full generic rank, a decision followed by the majority of subsequent authors (Wagner and Durham 1966; Mintz 1966; Solovjev 1971; Smith 2005; Barras 2007; Saucède et al. 2007). Metaporinus typically has subpetaloid to petaloid paired ambulacra, with closely packed plates and elongated and circumflexed pore pairs (Smith 2005). Moreover, these tithoniid echinoids have a slightly oval, elongated, high test with a slight frontal depression while Tithonia typically is cordiform with a moderate anterior notch. The oral face of Metaporinus shows a strong bulge on the sternum, while in Tithonia, the oral surface is nearly flat with a slight depression on ambulacrum III and a slight bulge on the plastron. The aboral face in Tithonia generally is rounded, whereas it is shaped like a roof in Metaporinus. The anterior part of the apical system of Metaporinus is close to the anterior margin while that of Tithonia is more central. Posterior ambulacra I and V of Metaporinus are concave frontally, while those of Tithonia are strongly convex. The peristome is transversely elongated in Metaporinus, but round in Tithonia. The main distinguishing features of Tithonia are the marginal position of the periproct and the fact that the posterior ocular plates are in contact with it or only slightly separated (Smith 2005).
With reference to these features, specimens collected from the middle Oxfordian at Beauvoisin clearly belong to the genus Tithonia, and in view of the fact that none of the species described previously for this genus correspond, either morphologically or chronostratigraphically, to it, a new name is coined for it.
Derivation of name. Alluding to the Oxfordian age of the material.
Type material. Holotype (FSL 286 400) and three paratypes (FSL 286 401, 286 402 and 286 403), contained in the Centre Commun des Collections de Géologie C3G, Université de Lyon 1.
Type locality and type level. Natural outcrop on the southern flank of the Montagne de la Taillade at Beauvoisin (Drôme); ‘Alternance argovienne’ of the middle Oxfordian (transversarium Zone, schilli Subzone).
Diagnosis. Test very inflated, with a height/length ratio between 0.75 and 0.95; test distinctly acuminate but faintly declived; preadult ambitus subquadrangular, becoming wider and triangular at adult stage with width/length ratio about 1; bulgy adoral surface without anal groove adorally; peristome in feeble depression, subcircular in preadult, more pentagonal in adult; ethmophract and endocyclic apical system with oculars II and IV containing posterior gonopores and absence of genitals 1 and 4; periproct oval and pointed at top, surrounded by oculars I and V and genital 5; sparse tuberculation on adoral face with small tubercles, denser from ambitus to oral face; sternum densely tuberculate.
Description. Test length (LL) between 17 and 35 mm; medium-sized individuals (23–27 mm) with rounded ambitus, occasionally slightly subquadrangular or feebly rectangular for smallest ones (17–20 mm); largest specimens (32–35 mm) slightly wider than long, with a subtriangular outline and a slightly acuminated posterior part. Maximum width (LM) variable but generally situated halfway. LA/LL ratio from 0.9 (smallest specimens) to 1.14 (generally larger specimens).
Test height (HT) and profile: large specimens proportionally more flattened than small ones (HT/LL between 70 and 96 per cent); lateral profile strongly curved forwardly below the trivium and dropping suddenly beyond the bivium (Pl. 1, figs 3, 8, 13, 17; Text-fig. 8B). Top of the test more or less acuminated. Aboral face strongly convex, very gibbous and slightly laterally declived (Pl. 1, figs 4, 5, 9, 10, 14, 15, 18, 19).
Ambulacra: ambulacrum III slightly depressed, with feeble anterior notch, frontal groove less marked upwards and effacing near mid-height in more strongly inflated individuals. Ambulacra I and V relatively large but slightly narrower than ambulacra of the trivium (Text-fig. 8A, G). Paired ambulacra strongly deviated upwards when crossing test margin (mainly in the smallest specimens), with sinuous outline on the oral face. Ambulacral pores arranged in pairs at the base of the plates, in a lateral-external position, except in the phyllodes (single pores) (Text-fig. 8D, H). Rounded pores, the outer larger than the inner, very elongated laterally and irregularly arranged adorally.
Peristome: anterior, located near the first one-third of the oral face from the front margin. Subcircular to subpentagonal in outline and located in a slight depression (Pl. 1, figs 2, 7, 12; Text-fig. 8C). Surrounded by very narrow, laterally stretched basicoronals with two pores irregularly situated; bordered by small triangular or quadrangular plates of the phyllodes with single pore (Text-fig. 8C, H).
Aboral face: sparsely covered by small tubercles. Tubercles denser and larger on the margin and the base of the flanks.
Periproct: supramarginal, at two-thirds of test height from the adoral face, for small specimens (Pl. 1, fig. 10), but descending to mid-height in larger specimens (Pl. 1, fig. 19; Text-fig. 8B). Outline from drop shaped to oval, upper tip generally sharp (Pl. 1, fig. 5; Text-fig. 8F, I). Oculars I and IV (in lateral positions) and genital 5 (on the posterior part) partially and entirely collapsed inside the periproct, respectively (Text-fig. 8A–B, E–F, I).
Posterior face of the test: vertical, without subanal groove or flat area near the periproct. Anal area inclined posteriorly in small individuals, vertical in preadults and becoming subparallel to the front and perpendicular to the oral face in larger specimens (Pl. 1, figs 3, 8, 13, 17; Text-fig. 8B).
Apical system: observed only in specimens larger than 22 mm. Typical tithoniid structure with posterior ocular plates containing gonopores, rather than the posterior genital plates (Text-fig. 9G, I). Madreporite (genital 2) generally slightly larger than other apical plates and with numerous perforations (Text-figs 8D, 9). Apex either in a forward or near-central position independent of size, with AP/LL ratio varying between 0.26 and 5.4. Interval between trivium and bivium making up one-third or two-thirds of the total adoral face length, with BI/LL ratio varying from 0.37 to 0.70. Larger individuals having proportionally the largest trivium–bivium interval. Several individuals with a near-conical apex. Apex located immediately behind the apical system.
Remarks. Jesionek-Szymańska (1963) noted that during the ontogeny of this kind of disasteroid, genitals 1 and 4 were strongly reduced prior to joining lateral oculars II and IV, respectively, which grew progressively longer. However, she could not illustrate this through a lack of juveniles. However, in smaller and medium-sized T. oxfordiana, between the pores of ocular IV and genital 4, or ocular II and genital 1, a faint, short indistinct line that could be an incompletely resorbed plate suture can be seen. However, it is impossible to determine whether the resulting composite plates are built mainly out of oculars or of genitals.
Discussion. Since the original description of Nucleolites convexus (Catullo 1827) and the introduction of Tithonia, about 14 species have been assigned to this genus. After revision, only six of these are accepted as valid by Smith (2005), namely T. convexa (Catullo, 1827), T. transversa (d’Orbigny, 1853), T. praeconvexa (Jesionek-Szymańska, 1963), T. berriasensis (de Loriol, 1867), T. munsteri (Desor, 1848) and T. exile (d’Eichwald, 1865). None of these species correspond to the new species, either morphologically or chronostratigraphically (Table 1).
Table 1. Comparisons of Tithonia oxfordiana with other species of Tithonia.
Deep and prominent anterior notch
HT/LL = 71–82%
Nearly oval profile
LA/LL = 105%
HT/LL = 80%
Marked anal groove
Deep anterior notch
Nearly square to trapezoidal profile
LM/LL = 51%
Triangular and posteriorly strongly acuminated
HT/LL = 90%
Marked anal groove
Nearly square profile
Conspicuous anterior notch
HT/LL = 75%
Nearly oval profile
LA/LL = 91–105%
HT/LL = 61–85%
Round to cordiform
Slight anal groove
Nearly oval profile
LM/LL = 28–47%
LA/LL # 100%
HT/LL = 70–96%
No marked anal groove
Subquadrangular to subtriangular
LA/LL = 85–90%
HT/LL = 63–75%
Marked anal groove
Slightly elongated to subcircular
Nearly square profile
Tithonia praeconvexa, from the upper Callovian of the Kraków-Częstochowa Upland (southern Poland), is the stratigraphically oldest and best-studied member of the genus. The species has also been recorded by Solovjev (1971) from the upper Callovian of the Crimean Peninsula (Ukraine). On the basis of measurements supplied by Jesionek-Szymańska (1963) and illustrations in Solovjev (1971), the Polish and Ukrainian material is morphologically very close. Adults of T. praeconvexa appear to be slightly smaller (LL = 28 mm) than T. oxfordiana (LL = 35 mm), while the ambitus in the former is gently elongated to subcircular and subquadrangular at the preadult stage in the new species, becoming slightly subtriangular when adult (Table 1); during growth, the ambitus becomes progressively wider, and the adult width/length ratio is about 1.0. Tithonia praeconvexa is not as inflated as T. oxfordiana (Table 1). Test profile of T. oxfordiana is slightly acuminate but faintly declived, whereas it is round and subhorizontal in T. praeconvexa. The peristome in T. praeconvexa is closer to the anterior margin than in T. oxfordiana; in both species, the peristome is in a shallow depression, subcircular in small specimens but becoming more pentagonal in larger ones. Tithonia oxfordiana is characterized by a slightly bulgy adoral face while that of T. praeconvexa is flatter. Compared to T. praeconvexa, T. oxfordiana differs mainly by the absence of a marked anal groove on its posterior face. The anterior and posterior parts of the apical system are more central and more close set in T. praeconvexa than in T. oxfordiana. Similar to T. oxfordiana, T. praeconvexa has an anterior apical system with ocular II joined to genital 1 and ocular IV joined to genital 4. The periproct is surrounded by oculars I and V and genital 5; its shape is oval, sometimes slightly pointed at the top.
Tithonia transversa was first described from the Callovian of Escragnolles (Var, France) and later illustrated by Cotteau (1867–1874) as Metaporhinus transversus (d’Orbigny) and by Deecke (1928) as Tithonia transversa (d’Orbigny). Following Cotteau’s arguments, the age of T. transversa should be Berriasian rather than Callovian, and its morphology differs markedly from that of Jurassic species, in particular T. oxfordiana. It is large sized and has a general cordiform shape with a contracted posterior part, its width exceeding length, and with a truncated posterior face, a strongly marked anal groove and a deep anterior notch.
Tithonia convexa from the Tithonian of Monte-Veronese (Provincia Austro-Venete, Italy) is based on Catullo’s (1820–1822) original description and figure of Nucleolito convesso, later referred by Catullo (1827) as Nucleolites convexus. This is the type species of Tithonia, and as such, it is fairly well known (Cotteau 1870; Favre 1877; Cotteau et al. 1884; Deecke 1928; Beurlen 1934; Mintz 1966). About 15 specimens, inclusive of those illustrated by Cotteau from the Tithonian of Algeria (J00924, J01319), are housed in the collections of the Muséum national d’Histoire naturelle Paris. Morphologically, all of these closely match the original figures and the individual illustrated by Favre (1877) from the Tithonian of Talloires and Lémenc (Haute Savoie and Savoie, France). This species has a rather rounded ambitus in small specimens (13–19 mm test length; LA/LL = 91–94 per cent), but it becomes slightly cordiform and wider than long in large specimens (21–35 mm test length; LA/LL = 98–105 per cent). The test is generally slightly narrower posteriorly than anteriorly, but remains rounded and not acuminated, with the maximum width situated more anteriorly in small individuals (LM/LL = 28–36 per cent) than in large ones (LM/LL = 38–47 per cent). Tithonia oxfordiana is more acuminated with the maximum width situated much more anteriorly. Similarly, in small and large individuals, the test of T. convexa is moderately elevated, while it is much more elevated in T. oxfordiana (Table 1). The peristome is in an anterior position and prolonged by a weak anterior notch, slightly more marked in T. oxfordiana than in T. convexa; the periproct is fairly high on the posterior face at the top of an anal groove, better marked in T. convexa than in T. oxfordiana.
Tithonia exile (= T. heinzi Coquand, 1880) is characterized by a deep and prominent anterior notch; the test is not very elongated and modestly elevated. The posterior part of the test is broadly truncated. Such morphological features do not match any of the Late Jurassic species. Solovjev (1971) described several small specimens from the Barremian of Crimea (Ukraine) that, in general appearance, are morphologically very different from T. praeconvexa and T. oxfordiana in showing a generally rounded aspect, but are very close to T. convexa, except for a more anterior peristome and a flat rather than a rounded oral face.
Tithonia munsteri was originally described as Dysaster munsteri from the Berriasian of Mecklenburg (eastern Germany). Subsequently, it was transferred to Metaporinus (Agassiz and Desor 1847; Desor 1858; Cotteau 1860; Deecke 1928; Beurlen 1934) and Tithonia (Mintz 1966). The species is clearly different from T. oxfordiana, having an extremely highly elevated test and near-squarish profile, a triangular and posteriorly strongly acuminated ambitus, a marked anterior notch and a deep subanal groove.
Tithonia berriasensis, from the Berriasian of Ardèche (France), was originally described as a species of Metaporinus by de Loriol (1873), Deecke (1928) and Beurlen (1934), but transferred to Tithonia by Mintz (1966). It has a conspicuous anterior notch and a subanal rostrum, both features distinguish it from Jurassic species of Tithonia.
Mode of life
Tithonia has an inflated, acuminated test; such morphology is characteristic of infaunal echinoids that inhabit deep-water, muddy environments, such as the Middle Jurassic Metaporinus or the Cretaceous Archiacia, Claviaster and Infulaster (Smith 1984; Néraudeau and Moreau 1989; Smith and Zaghbib-Turki 1985). Similar to these conical echinoid taxa, Tithonia lacks a dentition allowing it to graze on organic films on gravels or a rocky bottom. Like the great majority of irregular echinoids, Tithonia was almost certainly a bulk sediment swallower (Smith 1984). Primitive disasteroids, such as Pygorhytis, have phyllodes and relatively large, unspecialized adoral pores similar to those of cassiduloids (Smith 1984). They lived in sands or gravels, ingesting large volumes of sediment to feed on minute organic material particles contained in it. On the other hand, the smaller size of pores and associated tube feet of Tithonia allowed it to pick up particles smaller than sands or gravels. On the basis of its small number of tube feet around the mouth, Tithonia is more closely similar to spatangoids than cassiduloids, and like other Late Jurassic disasteroids, Tithonia could have had specialized penicillate tube feet around the mouth that used mucous adhesion, rather than suction, to pick up particles (Smith 1984). Thus, Tithonia probably was able to collect both fine and large particles with equal ease. However, it was not able to ingest large quantities of sediment to judge from its small mouth and weakly developed phyllodes and pores. This potential alimentary handicap probably had no detrimental effect in environments where nutrient levels were high, such as at chemoherm surfaces. The exceptional nutrient levels of the chemoherm environment may explain the high population density of Tithonia at Beauvoisin, especially as the genus in general is rare and present only in low abundance in ‘normal’ Jurassic echinoid assemblages.
Trophic relations and position in the ecosystem
Two observations suggest that the echinoid fauna formed part of a chemoautotrophic ecosystem. First, it is the occurrence of the assemblage just above carbonate lenses that are clearly chemoherms, and second, it is the high biomass and high number of individuals with regard to the low number of species, which characterize such special ecosystems in the Recent (Grassle 1986, 1989).
Moreover, while echinoderms predominate in the errant megafauna of the deep-sea floor (Gage and Tyler 1991), they are either rare in or absent from the vicinity of hydrothermal vents or seeps (Grassle 1985; Laubier 1989; Desbruyères et al. 2006). At the present day, only a single site dominated by echinoderms is known; it is situated on the Mazagan escarpment off Morocco where very dense holothurian ‘schools’ have been observed between depths of 1500 and 3000 m and where seeps are suspected (Zibrowius et al. 1988). Some other sites with abundant holothurians are known (Hecker 1985; Desbruyères et al. 1994; Van Dover et al. 2003). The occurrence of ophiuroids in relatively high densities has also been noted (Hecker 1985; Van Dover et al. 2003; Stöhr and Segonzac 2004). Localities at which crinoids are abundant are unknown, but unusually, well-preserved crinoids have been described from possibly methanogenic carbonates of early Oligocene age (Burns et al. 2005). Abundant echinoids have never been observed near modern hydrothermal vents or cold seeps, and those taxa that have been seen are not endemic. Only some regular echinoids (Echinus alexandriDanielssen and Koren, 1883) have been collected in the Lucky Strike site some ten metres of the vent sites (Mid-Atlantic Ridge) and in the south-east Pacific Rise (38°S, 2200 m depth; see Desbruyères et al. 2001, 2006; E. Ramirez-Llodra and P. A. Tyler, unpubl. data). The presence of mud in the peristomial opening of E. alexandri indicates a deposit-feeding behaviour. Like E. alexandri, holothurians from the Mazagan escarpment off Morocco probably ingest mud with bacteria. Numerous spatangoid echinoids (Sarsiaster griegiiMortensen, 1950) have been collected at the Blake Ridge cold seep site off South Carolina (Van Dover et al. 2003); this hemiasterid normally feeds on sediment and presumably also has symbiotic gut bacterial flora and may be specialized for this sort of habitat. A similar symbiotic association may be suspected for Tithonia from Beauvoisin. Microbial mats thrive in the close vicinity of Recent hydrothermal vents and cold seeps, and the same phenomenon could be expected in fossil chemoherms. Therefore, microbial mats could also correspond to the main source of food for Tithonia. Perhaps, Tithonia had a deposit-feeding mode of life and picked up superficial sediment enriched with bacteria. This is probably the best hypothesis, so long as modern examples are unknown or poorly understood.
The restriction of Tithonia to the top surface of the chemoherms demonstrates that echinoid settlement was brief and coincided with the final phase of chemoherms in the basin. In this respect, the presence of these echinoids in such a peculiar environment can be explained in two ways. First of all, chemoherms, in their final phase of existence, emitted a toxic chemical substance that rapidly killed Tithonia leading to the mass occurrence of these uncommon echinoids. Prior to this, Tithonia probably did occur but never formed dense aggregations (mass death accumulations). This hypothesis is supported by the fact that the Tithonia population contains two cohorts (preadults and adults) and lacks juveniles and thus probably is an in situ thanatocoenosis, juveniles being absent for ecological and demographic reasons (Néraudeau 1991). However, this style of preservation requires a sudden and massive sedimentation event, burying the thanatocoenosis and preventing the numerous Tithonia tests from being destroyed and scattered post-mortem by currents and other organisms subsequently. Second, chemoherms, near the end of their existence, emitted a new kind of bacteria that constituted a novel and opportunistic food source for echinoids previously living elsewhere. It is difficult to imagine that the ‘arrival’ of Tithonia on the chemoherm surface occurred during their preadult and adult stages because other settlements of Tithonia are unknown around the chemoherms. It is more likely that an exceptional settlement of Tithonia larvae took place at the top of the chemoherms when these temporarily became a favourable environment as far as nutrient production is concerned. However, as with the previous hypothesis, a sudden and massive sedimentation event must have buried this palaeobiocoenosis in situ.
Whichever hypothesis is correct, it can be assumed that these echinoids and the associated bivalves constituted prey for nektonic and nektobenthic predators, such as belemnites (abundant), crustaceans and fishes (less abundant). Traces of predation are not observed on tests which are generally entire. Nevertheless, the echinoids studied probably constituted a step in the following trophic chain: bacteria – bivalves (symbiotic)/echinoids (symbiotic or swallowers) – cephalopods/crustaceans/fish.
Relation between chemoherms and the history of Tithonia
Representatives of the genus Tithonia were previously known from two stratigraphical levels only, namely the Callovian (i.e. T. praeconvexa) and Tithonian–Berriasian (i.e. T. convexa, T. transversa, T. berriasensis, T. munsteri and T. exile). There were no previous records from Oxfordian deposits. Few data are available on the precise environment of all of these species, but all occurred at the border of the Tethys Ocean, probably in relatively deep-water settings. The present material suggests that T. oxfordiana was an endemic species restricted to the unusual environments provided by chemoherms. We assume that during the Oxfordian, these chemoherms constituted refuge areas for these echinoids. Thus, the submarine sources and ‘oases of life’, such as the present Oxfordian example, possibly had a fundamental palaeobiogeographical role in the survival and dispersion of some taxa.
Finally, the new species, Tithonia oxfordiana, is the first well-documented example of an extinct echinoid, which was linked to the development of chemoherms.
Acknowledgements. We thank Daniel Desbruyères and Michel Segonzac (IFREMER) for documentation on Recent deep-sea hydrothermal vent faunas, Jérôme Thomas (Dijon) for photographs of specimens and Gérard Breton for the loan of echinoid material. The authors are also very grateful to Ken McNamara, Andrew Smith and an anonymous reviewer for constructive comments that improved the text, as well as to John W. M. Jagt for editorial handling.