Cyprus hydrothermal vent community: taphonomy, palaeoecology and endemism

The Cyprus vent communities contain examples of shelly taxa (gastropods, ammonites and possibly serpulids as ‘Troodos attached tubes’) and taxa forming organic tubes (as ‘Troodos collared tubes’ and ‘Troodos wrinkled tubes’) (Table 1). With the exception of the ammonite (see below), it is most probable that these taxa represent animals that were living in areas of active mineralization; around or on hydrothermal vent chimneys. Extremely rapid (a year or less) replacement of alvinellid and vestimentiferan worm tubes, and molluscan shells by sulphide minerals (mostly pyrite) and silica is known from vent chimneys at modern hydrothermal sites (Cook & Stakes 1995; Georgieva et al. 2015). This rapid replacement process is probably a prerequisite for preservation of vent macrofauna in the fossil record, because a number of sea-floor experiments have shown that the organic tubes of vestimentiferans, crab carapaces and bivalve shells are dissolved and/or microbially degraded in only a few years at modern vent sites (Kennish & Lutz 1999; Ravaux et al. 2003). For bivalve shells the rate of shell dissolution is controlled by mineralogy, with aragonitic microstructures dissolving before those made of calcite, and also by proximity to active venting, because vent fluid is usually acidic. Even bivalve shells away from active venting are estimated to dissolve entirely in c. 300 years, due to undersaturation of aragonite in the water at the experimental sites of Kennish & Lutz (1999), which are at 2600 m on the East Pacific Rise. In the context of the Cyprus vent communities the rate of carbonate shell dissolution would have been similar close to active vents, and likely to be more rapid away from the vents (compared with the modern experimental sites), because of the estimated 2500–5000 m water depth during Troodos ocean crust formation and the fact that carbonate compensation depths (ccd) in Cretaceous oceans were probably shallower than today, because of higher CO₂ atmospheric concentrations (global estimate of 3500 m during the Late Cretaceous; Zeebe & Tyrrell 2019). Furthermore, the first biogenic sediments deposited on the Troodos ocean crust (the middle–late Turonian to latest Santonian radiolarian cherts of the Perapedhi Formation) lack carbonate, showing that even c. 11 myr after the formation of the Troodos ophiolite the sea floor was still below the palaeo-ccd.

In modern vent communities the majority of animals with chemosymbionts live close to the active part of vent systems and are consequently those more likely to be preserved (see above). We use these observations and comparisons with modern taxa to reconstruct the Cypriot vent communities. The commonest element in the Cypriot communities, both in terms of abundance (Little et al. 1999b) and geographic distribution (Table 1), are the possible vestimentiferan tube worms (the Troodos collared tubes and Troodos wrinkled tubes of Georgieva et al. 2019). If these are indeed vestimentiferans then they would have been dependent for their nutrition on sulphide-oxidizing chemosymbionts, as are all modern vent siboglinids (Hilário et al. 2011; Karaseva et al. 2016). The Troodos attached tubes may have been the attached juveniles of siboglinids, or, if serpulids, would have been filter-feeders, although serpulids with methanotrophic symbionts have recently been described from hydrocarbon seeps in the Pacific Ocean (Goffredi et al. 2020). The Cypriot vent gastropods lived alongside the worm tube taxa, because they often co-occur in the same blocks of sulphide. We infer that the Cypriot vent abyssochrysoids were grazers on microbial mats, like the majority of extant vent abyssochrysoids. However, some modern vent abyssochrysoids have methantrophic and sulphide-oxidizing symbionts (e.g. species belonging to the genera Ifremeria and Alvinochoncha), housed in hypertrophied gills. These taxa are large and have a globular morphology to contain their expanded gill tissue (Dubilier et al. 2008). The Cypriot species of Paskentana, in particular P. dixoni sp. nov., approach the shape of some of the modern species of Ifremeria, but suggesting that these fossil species had chemosymbionts is speculative, given that there are no living representatives of Paskentana for comparison.

The conspicuous indentation on two of the upper whorls of the holotype of Paskentana xenophontosi from Memi (see below) could be attributed either to predation damage or to a growth defect. Another alternative is that it represents mechanical damage of a very thin shell with high organic content, leading to ductile rather than brittle damage behaviour. Similar folds have been seen in the shells of abyssochrysoid shells from Jurassic and Cretaceous hydrocarbon seeps in California, USA (Kiel et al. 2008a). In the context of hydrothermal vents, one further explanation for the indentations in Paskentana xenophontosi could be that this was an area where the shell was completely dissolved, leaving the more flexible periostracum to deform into the space left behind. This interpretation is supported by the common observation at modern vent sites of extreme shell dissolution of bivalve and gastropod shells (especially protoconchs) leaving only periostracum, even while animals are still alive (e.g. Warén & Ponder 1991; Warén & Bouchet 1993). Three of the Cyprus gastropod specimens from two vent sites (Hokkaidoconcha morisseaui (Fig. 4B) and Cyprioconcha robertsoni (Fig. 4I) from Kambia, and Desbruyeresia memiensis (Fig. 3U) from Memi) have smooth, domed structures on their shell surfaces, which are roughly circular and c. 1–2 mm in diameter. There are 1–4 of these per shell. These structures do not belong to the shells themselves, so are not palaeo-pathologies, but rather represent epizoans that attached to the shell surfaces. There are several possibilities as to their origin, including gastropod egg capsules, or the very early settlement structures of vestimentiferan worm tubes (Southward et al. 2005).

The presence of a single gaudryceratid ammonite in the Kambia VMS is surprising for two reasons. First, this is the first cephalopod discovered in any ancient vent community; and second, it is related to the water depth at which the Cypriot vent communities lived (see ‘Geological setting’ above). It seems highly improbable that the ammonite inhabited the Kambia vent site at depths of 2500–5000 m, because the thin shells of a gaudryceratid would have imploded even at much shallower depths. Indeed, Hewitt (1996) calculated an implosion depth of 328 m for gaudryceratids. Even modern Nautilus shells, which are substantially thicker than those of gaudryceratids, and have thicker septae, implode at a maximum depth of 785 m (Kanie et al. 1980). For this reason, we suggest that the gaudryceratid represents a post-mortem water-logged shell that sank from surface waters and settled fortuitously close to the active Kambia vent site, where its shell was rapidly replaced by pyrite. Thus, the ammonite was not part of that vent community.

Worm tubes are found in all six of the fossiliferous Cypriot VMS (Table 1). This was to be expected considering that they are the commonest macrofossils in Phanerozoic VMS (Haymon et al. 1984; Oudin et al. 1985; Boirat & Fouquet 1986; Little et al. 1997, 1998, 1999a, b, c, 2007; Revan et al. 2014; Georgieva et al. 2019); it also reflects the fact that in many modern vent communities, particularly those in the East Pacific, tube-forming polychaetes (e.g. siboglinids and alvinellids) are abundant, and also tend to live in proximity to areas of active mineralization (Shank et al. 1998; Hilário et al. 2011; Karaseva et al. 2016). Gastropods are found in three of the fossiliferous Cypriot VMS (Table 1). Their absence from the Peristerka and Sha VMS is likely to be due to collection failure given that fewer than 10 macrofossils were found at each site (Oudin & Constantinou 1984; Little et al. 1999b, table 1). More worm tube fossils were found in the Kapedhes VMS (n = 77; Little et al. 1999b, table 1), but the absence of other taxa may also be an artefact of collection failure, or of the selective preservation of only worm tubes at this site. There are no shared gastropod species between the Kinousa, Memi and Kambia VMS (Table 1). This is intriguing, because there would be an expectation of at least some shared species, given the small distances between the sites (Fig. 1; Little et al. 1999b, table 1) and the lack of obvious differences in fluid chemistry, substrate and palaeobathymetry. However, a major caveat is that there is no proof that any of the Cypriot VMS were active at the same time, and indeed they could have been separated temporally over as much as 3 myr (see ‘Geological setting’ above). In this case the faunal differences may reflect evolutionary changes in vent gastropod lineages over time.

Macroevolutionary considerations

The gastropod fauna of the Cypriot hydrothermal vent communities consists of five genera and eight species in three abyssochrysoid families: Provannidae, Hokkaidoconchidae and Paskentanidae. Provannids are ubiquitous at modern hydrothermal vent sites (e.g. Warén & Bouchet 1993; Desbruyères et al. 2006; Linse et al. 2019); hokkaidoconchids and paskentanids are extinct taxa. Species of Desbruyeresia, Hokkaidoconcha, Ascheria and Paskentana from Cyprus are the first representatives of these genera to be recorded from any hydrothermal vent sites, and those of the last-named genus are also the youngest representatives from any environment. Provannids, hokkaidoconchids and paskentanids are common constituents of Mesozoic hydrocarbon seep communities (Fig. 2), including representative species of the genera Desbruyeresia, Hokkaidoconcha, Ascheria and Paskentana (Kaim et al. 2008, 2014, 2017; Kiel et al. 2008a). Desbruyeresia is relatively rare in the fossil record. In addition to the Cypriot record, the genus is known from Cenomanian seep deposits in Japan (Kaim et al. 2008) and Eocene seeps in Washington State, USA (Hybertsen & Kiel 2018). Hokkaidoconcha is a cosmopolitan genus with the oldest occurrence in the Oxfordian seeps in France (Kiel et al. 2010) and the youngest from the Eocene seeps of Barbados (‘zygopleurid sp. A’ of Gill et al. 2005, reinterpreted as ?Hokkaidoconchidae by Kaim et al. 2008). Ascheria is known from Lower Cretaceous seeps of California and the Czech Republic (Kaim et al. 2013, 2014) and ranges temporally to Oligocene seeps in Peru (Kiel et al. 2020). The genus Paskentana might be as old as Late Triassic (i.e. the Paskentana-like gastropod of Kiel et al. 2017 from Turkey), and its youngest occurrences at seeps are from the Hauterivian of the Crimea and California (Kiel et al. 2008a, 2010; Kaim et al. 2014). We suggest that the latest Cenomanian to earliest Turonian age of the Cypriot vent species of Paskentana might indicate a habitat switch for this gastropod taxon from hydrocarbon seeps to hydrothermal vents between the Lower and Upper Cretaceous. However, the lack of other Cretaceous vent sites with gastropod fossils makes this inference quite speculative. From a wider palaeobiogeographic perspective, in the Mesozoic the representative species of Hokkaidoconcha, Ascheria and Paskentana occurred both at vents (Cyprus) and at seeps (France, Crimea and Czech Republic; Kiel et al. 2010; Kaim et al. 2013, 2014) in the Tethys, while Desbruyeresia has so far been found only at the Cypriot vents for this time period and area. All four genera occurred in seeps in the North Pacific, now California and/or Japan (Kaim et al. 2008, 2014), during the Mesozoic. Today, a high diversity of species of Desbruyeresia is known from the West Pacific vent sites and serpentinization-related seeps, and there is a single species on record from Indian Ocean vents (Sasaki et al. 2010; Chen et al. 2016).

Compared with modern vent sites and contemporary seeps, many ecologically important and speciose major taxa are missing from the Cyprus vent communities, in particular vetigastropod and neomphaline snails, bivalves and arthropods. It may be that these taxa were originally present, but have not been found in the Cypriot VMS deposits, either because of simple collection failure or through contemporary taphonomic loss. It seems very unlikely that the absence of bivalves has a taphonomic explanation, because bivalves are present in other (older) vent communities (Devonian; see Little et al. 1999c), and they can be preserved in modern vent deposits (CTSL pers. obs.) Contemporary seep deposits contained numerous bivalve taxa, particularly those with supposed chemosymbionts: solemyids, lucinids and thyasirids (Kiel et al. 2008b, 2010; Kiel 2013; Kaim et al. 2014; Hryniewicz et al. 2017). The absence of these families and of the extinct modiomorphid seep genus Caspiconcha may have been related to substrate availability, given that the Cypriot vent sites consisted exclusively of hard substrates (as deposits of metal sulphides and basaltic pillow lavas), whereas solemyids, lucinids, thyasirids and Caspiconcha were infaunal or semi-infaunal (Kelly et al. 2000; Kiel & Peckmann 2008; Jenkins et al. 2013, 2018), and would have required at least some soft substrate to burrow into. Indeed, at most modern vent sites solemyids, lucinids and thyasirids are either absent or highly localized to small patches of sediment (e.g. Glover et al. 2004; Desbruyerès et al. 2006; Oliver & Holmes 2006, 2007), the exception being sediment-covered vent sites, where soft substrate is abundant (e.g. Lartaud et al. 2010; Bell et al. 2016). The two chemosymbiotic bivalve groups that are nearly ubiquitous and often numerically dominant at modern vent sites are the large vesicomyid clams and bathymodiolin mussels (e.g. Taylor & Glover 2010), which are either epifaunal where no soft substrate is available, or semi-infaunal where there is. These taxa have Paleocene divergence estimates of 63 and 58 Ma, respectively (Vrijenhoek 2013), dates that are largely supported by the Eocene appearance of these taxa in seep deposits (Amano & Kiel 2007; Kiel & Amano 2013). In consequence, neither the vesicomyids nor the bathymodiolins would be expected to be present at Upper Cretaceous vent sites.

The lack of vetigastropods or neomphalines in the Cyprus vent communities also cannot be explained by taphonomic loss either, and these groups have numerous representative taxa in contemporary and other Mesozoic seep deposits (Kiel et al. 2008a; Kaim et al. 2009, 2014). Therefore, their absence from the Cyprus vents could have been a real phenomenon, perhaps a function of Late Cretaceous biogeography and/or environmental preference.

In contrast, taphonomic loss may explain the absence of arthropods from the Cyprus vent communities, because many modern vent arthropods, such as bresiliid shrimp, have only lightly calcified carapaces, with low preservation potential, and in experiments even the more robustly shelled taxa, such as bythograeid crabs, have been shown to disappear very quickly (36 days or less) in the modern vent environment (Ravaux et al. 2003). However, age is also a likely explanation for the absence of arthropods from the Cyprus vents because the arthropod groups that occur abundantly in modern vent communities appear not to have evolved by the Late Cretaceous. Phylogenetic analyses of bresiliid shrimp, bythograeid crabs and vent cirripedes resulted in Cenozoic divergence estimates for all of these (Vrijenhoek 2013; Herrera et al. 2015).

The Cypriot vent communities share worm tube macrofossils with other Neotethyan hydrothermal vent sites of Late Cretaceous age. The Troodos ophiolite lies at one end of a belt of Upper Cretaceous ophiolites that stretches from Cyprus, through Syria, Iraq and Iran, to the Semail ophiolite of Oman and the United Arab Emirates. These ophiolites probably represent a series of small back-arc or forearc basins lying above a subduction zone within the gradually closing Neotethys Ocean (Robertson et al. 1991). The Semail ophiolite formed between 96.1 and 95.5 Ma (Rioux et al. 2016), corresponding to the Cenomanian in version 2020/01 of the International Chronostratigraphic Chart (Cohen et al. 2020), and thus is around 1–2 myr older than the Troodos ophiolite. One of the many VMS in the Semail ophiolite has been shown to contain worm tube fossils (Bayda; Haymon et al. 1984; Haymon & Koski 1985). These are morphologically quite similar to the Troodos collared tubes and the Troodos wrinkled tubes (Georgieva et al. 2019), being sinuous with an external ornament of longitudinal ridges and concentric annulations, between 1 and 5 mm in diameter (Little et al. 1998), and may also have been formed by vestimentiferan worms. A little later in the Late Cretaceous, to the north of this belt of ophiolites in the Neotethys, were a series of Turonian–Santonian-aged volcanic arcs with associated hydrothermal activity, represented today by VMS in the southern Transcaucasus – eastern Pontides orogenic belt in Georgia and Turkey. Three of these VMS contain worm tube fossils. From the Madneuli VMS of Georgia are three smooth-walled tubes that are considerably larger than any of the Cypriot tubular fossils, being between 6.8 and 11.9 mm in diameter (Little et al. 2007). The Lahanos and Killik VMS of Turkey contain tubular fossil-like structures that are even larger than the Madneuli tubes, being 20–25 mm in diameter (based on Revan et al. 2014, figs 4c, 5). There are not sufficient morphological features in the tubular fossils from the Georgian and Turkish VMS to identify them with any certainty, but their presence, along with the Cyprus and Oman examples, provides a record of hydrothermal vent communities in the western part of the Neotethyan Ocean for c. 10 myr, from the Cenomanian to the Santonian, from a variety of tectonic settings (Little et al. 2007). The presence of worm tube fossils in the Upper Cretaceous Neotethyan vent communities is intriguing given the absence of large tube-dwelling worms (e.g. vestimentiferans) from modern vent sites in the modern-day Indian Ocean, and their occurrence, often in great abundance, in Caribbean, West and East Pacific vent communities (Karaseva et al. 2016). This, together with the distribution of modern and fossil species of Desbruyeresia (see above), is suggestive of a biogeographic link in the Late Cretaceous between the western Neotethys and the Pacific via spreading ridges(s) in the central and eastern Neotethys, since destroyed by the Cenozoic collision of India into Asia and the continued movement northwards of Africa and Arabia (e.g. Moix et al. 2008; Gibbons et al. 2015). This would have caused a regional extinction of hydrothermal communities, and the Neotethyan vent faunal combination of provannid, hokkaidoconchid and paskentanid gastropods, and tube worms is likely to have disappeared. However, only the discovery of other Mesozoic and Cenozoic vent communities from the Neotethys, Indian and West Pacific will corroborate this hypothesis.

Worm tubes

The worm tubes from the Cypriot VMS were described in detail by Georgieva et al. (2019). We here reiterate only the most important information. Georgieva et al. (2019) distinguished three groups of tubes: Troodos collared tubes from the Kambia, Kapedhes, Memi and Sha VMS; Troodos wrinkled tubes from the Kambia and Kapedhes VMS; and Troodos attached tubes from the Kinousa and Memi VMS. The collared and wrinkled tubes were considered by Georgieva et al. (2019) to be possible vestimentiferan siboglinids, while the attached tubes are possibly serpulids, although a siboglinid affinity could not be ruled out. The worm tubes figured by Oudin & Constantinou (1984) from the Peristerka VMS do not show sufficient characters to be compared to the worm tubes from the other Troodos VMS (Little et al. 1999).