Introduction
Eocene rocks from Seymour (Marambio) Island, at the northern tip of the Antarctic Peninsula (Fig. 1), contain a record of remarkable mollusc shell concentrations representing ecosystems with no Recent parallels, as they developed in a relatively warm (c. 14°C) shallow sea in a sunlight/darkness-stressed setting. Detailed study of fossil organisms as pointed out by Buick & Ivany (Reference Buick and Ivany2004), from environments no longer extant is necessary if we are to understand the nature and degree of environmental control on their life history.
Fig. 1. Map showing study localities.
In this paper we describe and analyse the association of encrusting and boring organisms developed on the hermitted shells of the gastropod Antarctodarwinella ellioti Zinsmeister, 1976, from the lower section of the La Meseta Formation exposed along the northern tip of Seymour Island (Fig. 1). Hermitted will be used instead of hermit crab-occupied shell (after Vermeij Reference Vermeij1978). The age of the rocks bearing the fossil fauna considered is Eocene, c. 52 Ma (Dutton et al. Reference Dutton, Lohmann and Zinsmeister2002).
Antarctodarwinella ellioti belongs in the Struthiolariidae (Fischer, 1884), a family of gastropods that evolved exclusively in the Southern Hemisphere since the Late Cretaceous. Modern representatives of this family still inhabit shelf environments in the circum-Antarctic realm. It is represented there by the genera Struthiolaria Lamarck, 1816, Pelicaria Gray, 1847, Tylospira Harris, 1897 and Perissodonta Martens, 1878. The Recent Struthiolaria papulosa (Martyn, 1784), Pelicaria vermis (Martyn, 1784), and Tylospira scutulata (Martyn, 1784) live in clean sandy environments in shallow to outer shelf marine settings. Being soft-bottom dwellers, most of these gastropods live generally buried in the sediment. They may be either deposit feeders, drawing their food directly from the substrate or else suspension-feeders, in which case they maintain the tips of their siphons above the sediment/water interface.
Encrusting and boring organisms are significant components of marine communities, interacting in different ways with either live or dead host substrates. They have proven to be useful tools in the understanding of the life habits and/or post-mortem histories of their hosts (Bottjer Reference Bottjer1982, Bordeaux & Brett Reference Bordeaux and Brett1990, Bien et al. Reference Bien, Wendt and Alexander1999, Taylor & Wilson Reference Taylor and Wilson2003, Parras & Casadío Reference Parras and Casadío2006).
The community of encrusting and boring organisms dominated by polychaetes and bryozoans recorded on the studied specimens of Antarctodarwinella ellioti - and their distribution on the shells - allow us to infer that they were occupied by hermit crabs.
Hermit crabs are a very specialized group with a wide geographic distribution. They put back into circulation empty gastropod shells that would otherwise likely be buried in the sediments (Walker Reference Walker1989, Gutiérrez et al. Reference Gutiérrez, Jones, Strayer and Iribarne2003). Hermit crab assemblages can be considered examples of facilitation (Bruno et al. Reference Bruno, Stachowicz and Bertness2003) in which crabs extend the range of associated species through positive interactions (Reiss et al. Reference Reiss, Knäuper and Kröncke2003, Bell Reference Bell2005). Over 180 Recent hermit crab species act as host for at least 550 invertebrate species, representing 16 phyla (Williams & McDermott Reference Williams and McDermott2004). Hermit crabs benefit from some symbionts, particularly bryozoans and hydractinians, through extension of shell apertures and by providing shelter from predators (Taylor et al. Reference Taylor, Schembri and Cook1989, Olivero & Aguirre-Urreta Reference Olivero and Aguirre-Urreta1994, Taylor Reference Taylor1994, Taylor & Schindler Reference Taylor and Schindler2004). However, hermit crabs are also negatively impacted (e.g. reduction of reproductive success) by several symbionts (Buckley & Ebersole Reference Buckley and Ebersole1994, Williams Reference Williams2000, Reference Williams2002, McDermott Reference McDermott2001). Hermit crabs are common members of high-energy, shallow marine faunas (Hazlett Reference Hazlett1981, Williams & McDermott Reference Williams and McDermott2004). They have been recorded in rocks as old as Lower Jurassic (Glaessner Reference Glaessner and Moore1969) but their fossil record is poor and generally restricted to isolated claws. In contrast, indirect evidences of the presence of hermit crabs are abundant from the Middle Jurassic onwards (Walker Reference Walker1992). Among the most important evidence used to recognize hermitted shells is the type and distribution of encrusting and boring organisms on their surface. In such shells, these organisms tend to colonize preferably the aperture periphery, apertural notch, callus, columella and the interior whorls of the shell. Boring or encrusting organisms that settle on these areas are mainly suspension-feeders that benefit from the water currents produced by the hermit crabs (Williams & McDermott Reference Williams and McDermott2004). However, the outer surface is colonized also (Walker Reference Walker1988, Reference Walker1998). Presumably colonization of the outer shell surface is equally probable in hermitted epifaunal and infaunal gastropod shells. However, only those infaunal gastropods shells that are occupied by hermit crabs will show colonization of the outer surface as such surfaces are not available to bionts when the gastropod is alive.
The hermit crab record in Antarctica is poor and restricted to the Upper Cretaceous on Snow Hill and James Ross islands (Aguirre-Urreta & Olivero Reference Aguirre-Urreta and Olivero1992, Olivero & Aguirre-Urreta Reference Olivero and Aguirre-Urreta1994).
Geological setting
The La Meseta Formation (Elliot & Trautman Reference Elliot, Trautman and Craddock1982) is a clastic sedimentary unit exposed in the northern third of Seymour (Marambio) Island and in a small area of Cockburn Island. It takes its name from the plateau lying at the northern end of Seymour Island, where the 720 m type section of this unit is exposed. The La Meseta Formation unconformably overlies the López de Bertodano (Upper Cretaceous), Sobral (Palaeocene), and Cross Valley (Palaeocene) formations. It is unconformably overlain by late Cenozoic glacial beds.
The La Meseta Formation represents the upper part of the filling of a retroarc basin named the James Ross Basin by del Valle et al. (Reference Del Valle, Elliot and Macdonald1992). It contains an abundant and well-preserved fauna of marine invertebrates and vertebrates, as well as terrestrial vertebrates and plants. The fossil content is a unique temporal window that allows the study of evolutionary and palaeobiogeographic events taking place during the early Palaeogene in the high latitudes of the Southern Hemisphere. The excellently preserved fossils occur mainly in remarkable accumulations dominated by mollusc shells. Such a unique fossil fauna was subject of numerous systematic studies (e.g. Wilckens Reference Wilckens1910, Zinsmeister & Camacho Reference Zinsmeister, Camacho and Craddock1982, Woodburne & Zinsmeister Reference Woodburne and Zinsmeister1982, Feldmann & Woodburne Reference Feldmann and Woodburne1988, Stilwell & Zinsmeister Reference Stilwell and Zinsmeister1992, Radwańska Reference Radwańska and Gaździcki1996, Bitner Reference Bitner and Gaździcki1996a, Goin et al. Reference Goin, Case, Woodburne, Vizcaino and Reguero1999, Casadío et al. Reference Casadío, Marenssi and Santillana2001, Hara Reference Hara and Gaździcki2001). However, very few studies deal with the nature and origin of the shell concentrations from taphonomic and palaeoecologic perspectives (Zinsmeister Reference Zinsmeister1987, Vizcaíno et al. Reference Vizcaíno, Reguero, Goin, Tambussi, Noriega and Casadío1998). Consequently, issues such as the taphonomic evolution biases of these concentrations remain virtually unknown.
Based on its fossil content, the La Meseta Formation was referred to the Lower Eocene–Upper Eocene or even Lower Oligocene by Stilwell & Zinsmeister (Reference Stilwell and Zinsmeister1992). Isotopic data suggest an Eocene age of c. 52 Ma (Dutton et al. Reference Dutton, Lohmann and Zinsmeister2002) for the lower part of the section. Elliot & Trautman (Reference Elliot, Trautman and Craddock1982) interpreted this formation as a tide-influenced delta system in which they recognized a lower section attributed to a prodelta, a middle section representing the delta front and an upper part probably characterizing a subtidal delta platform. Pezzetti (Reference Pezzetti1987) proposed a tidal-dominated delta as the origin of this formation, yet noted that only delta-front deposits could be recognized. Doktor et al. (Reference Doktor, Gaździcki, Marenssi, Porębski, Santillana and Vrba1988) suggested that the La Meseta Formation represented the progradation of a shallow marine environment, restricted by bars or barriers, under wave and tidal influence. Sadler (Reference Sadler, Feldmann and Woodburne1988) recognized the lens geometry of this unit, and he proposed that such internal lens structure together with the sedimentologic setting could correspond to lagoon and barrier environments. Marenssi (Reference Marenssi1995) and Marenssi et al. (Reference Marenssi, Santillana and Rinaldi1998a, 1998b) concluded that the La Meseta Formation represented an incised valley system, where sedimentation took place in deltaic, estuarine and wave-influenced tidal-shelf environments. Porębski (Reference Porębski2000) suggested that the La Meseta Formation represented a shelf valley-fill sequence, its development mainly governed by local subsidence along fault-controlled valley margins.
The overall depositional setting ranged from a prograding delta front to a storm-influenced sub aqueous delta plain dominated by tides after marine-flooding within an incised valley (Marenssi Reference Marenssi1995, Marenssi et al. Reference Marenssi, Santillana and Rinaldi1998a).
The studied samples come from the lower section of the La Meseta Formation (Fig. 2). These beds are included in the middle part of the Telm 3 (Sadler Reference Sadler, Feldmann and Woodburne1988) or in the upper part of the Acantilados Allomember (Marenssi et al. Reference Marenssi, Santillana, Rinaldi and Casadío1998b). This part of the succession is characterized by intercalated biogenic and sedimentologic concentrations where the brackish bivalve, Eurhomalea Cossmann, 1920, is dominant. Also present, but in greatly reduced numbers are other bivalves, gastropods and brachiopods. Biogenic concentrations consist of several dispersed to densely-packed and poorly-sorted lenses with fine sandstone matrix, ranging from 1 to 1.4 m thick, with sharp planar or undulating bottoms and sharp undulating tops. Most of the bivalves are articulated, many as the so-called “butterflied” bivalves. Fragmentation, abrasion, encrustation, and bioerosion are low. Bioturbated beds are common and consist of small-diameter (<2 mm) tubes perpendicular or inclined to stratification. These concentrations are interpreted as having formed in a delta plain environment during brief episodes of benthic colonization. Fossils are intercalated with sedimentologic concentrations formed by densely-packed to dispersed lens or beds, between 0.05 and 1.10 m thick, with fine to medium sandstone matrix and trough cross-stratification. Lithic clasts (up to 0.4 m), intraclasts (up to 0.5 m), concretions (between 0.05 and 0.2 m), and preserved burrows of the underlying beds are frequent. They are well- to poorly-sorted, and even bimodal (one mode represented by Eurhomalea spp. measuring 30 to 50 mm long, and the other one by juvenile veneroids and gastropods, all smaller than 5 mm long). The bottom and top surfaces are sharply undulose, and the bottom one is also sometimes bioturbated. Most of the bivalves are disarticulated, concordant or in chaotic position, sometimes arranged in stacking pattern. Fragmentation is low in some specimens to high in others; abrasion, encrustation and bioerosion are low. Eurhomalea species are all infaunal; thus some centimetres of sediment should have undergone erosion before the shells were finally deposited. The high proportion of disarticulated valves suggests that most individuals were dead at the time of exhumation. Furthermore, the scarcity of shell fragments suggests a single brief event of exhumation and exposure on the seafloor, as these large but thin and relatively fragile shells break easily. The distinctive stacking pattern and imbrications of disarticulated valves indicates highly turbulent conditions. Except for Antarctodarwinella ellioti, all invertebrate remains show few post-mortem borers and encrusters. This suggests minimal exposure time on the sea floor after exhumation. These concentrations are interpreted as composite or multiple-event concentrations deposited by tidal channels in a tidal-dominated sub aqueous delta plain.
Fig. 2. Stratigraphic section of the La Meseta Formation showing the beds containing Antarctodarwinella ellioti.
Materials and methods
Thirty-three specimens of Antarctodarwinella ellioti were collected from two localities, 64°13′40″S, 56°39′05″W and 64°13′34″S, 56°38′56″W (Fig. 1), most from biogenic concentrations. Specimens were analysed under a binocular microscope and the kind and location of encrusters and borers was recorded for each one.
The number and percentage of specimens bored or encrusted by different organisms was calculated. To test the hypothesis that a correlation exists between the sector of the shell considered and the presence/absence of the different encrusting/boring taxa - i.e. that there was preferential colonization - we defined three sectors reflecting distinct morphological features of the shells (Fig. 3). Sectors and percentage were: A = spire and abapertural exterior of last whorl (50%), B = spire and adapertural exterior of last whorl (25%), C = aperture interior (25%).
Fig. 3. Shell sectors on Antarctodarwinella ellioti. A = spire and abapertural exterior of last whorl, B = spire and adapertural exterior of last whorl, C = aperture interior. Aperture interior sectors: columella (COL), siphonal canal (SCAN), outer lip (OL) and apertural notch (APN).
The presence of boring and encrusting organisms in each sector was recorded, and the frequency of shells with different taxa of encrusting or boring organisms was determined for each sector. Contingency tables were performed and a Chi-square Independence Test was used to ascertain whether or not these frequencies were non-randomly distributed among sectors. Methodological restrictions of the Chi-square Independence Test requires that all expected frequencies be 1 or greater and at most 20% of them be less than 5. Therefore, we performed this test using only the taxa that met these assumptions.
Because some of the boring or encrusting organisms showed preferential placement or at least high frequencies on area C, we divided this area into four sectors (Fig. 3): columella (COL), siphonal canal (SCAN), outer lip (OL) and apertural notch (APN). To test the hypothesis that the frequency of each of the boring/encrusting taxa is the same for each sector, we recorded the presence of boring and encrusting organisms in each sector, and subsequently performed a Cochran Q Test. Software packages used for statistical analyses and graphics were Excel XP and STATISTICA, version 99.
All specimens are housed at the Departamento de Ciencias Naturales, Universidad Nacional de La Pampa (GHUNLPam), Santa Rosa, La Pampa, Argentina, under numbers GHUNLPam 25205–25237.
Results
The fauna of encrusting and boring organisms recorded in the shells of Antarctodarwinella ellioti includes polychaetes, bryozoans, balanids, algae, and phoronids? (Table I and Fig. 4). Each case is described and discussed below and the distribution of each particular encruster/borer is also discussed.
Table I. Presence/absence of boring and encrusting organisms in each sector of the shell.
Fig. 4. Percentage of shells bored or encrusted by different organisms, n = 33.
Encrusting organisms - polychaetes
The first reference recording Spirorbis Daudin, 1800 in the La Meseta Formation was by Wiedman & Feldmann (Reference Wiedman, Feldmann, Feldmann and Woodburne1988). These authors described material attached primarily to other larger calcareous tube hosts collected from the top beds of this lithostratigraphic unit (Telm 7 sensu Sadler Reference Sadler, Feldmann and Woodburne1988). Bitner (Reference Bitner1996b) later recorded Spirorbis encrusting brachiopods from different beds of the La Meseta Formation. Specimens referred to Spirorbis sp. and encrusting shells of Antarctodarwinella ellioti have a coiled calcareous tube, either evolute planispiral or trochoid, small (<0.2 cm diameter), with smooth outer and inner surfaces, with up to three whorls in the largest specimens, in which each successive whorl is larger than the previous whorl (Fig. 5d–f). Some specimens of Antarctodarwinella ellioti (GHUNLPam 25214, 25217, 25218, 25219, 25222, 25224, 25227, 25228, 25236) record mutual overgrowth of Spirorbis sp. and cheilostome bryozoans. In such cases, instead of acquiring their typically spiral shape, polychaetes elevate their apertures away from the encroaching zooids. A similar behaviour was described by Lescinsky (Reference Lescinsky1997) for Carboniferous “spirorbids” and by Stebbing (Reference Stebbing and Larwood1973) for Recent spirorbids growing adjacent to bryozoan colonies.
Fig. 5a. & b. Leptichnus isp. (GHUNLPam 25228), c. Coralline algae covering Spirorbis sp. (GHUNLPam 25205), d. & f. Spirorbis sp. (GHUNLPam 25205).
Studies of settlement patterns of encrusting communities on extant marine species revealed that Spirorbis is the first in the succession (Chalmer Reference Chalmer1982, Keough Reference Keough1983, Walker & Carlton Reference Walker and Carlton1995). Ryland & Sykes (Reference Ryland and Sykes1972) found a positive correlation between the abundances of spirorbids and bryozoans in Recent encrusting communities. Taylor (Reference Taylor1979) observed a similar correlation in middle Jurassic communities, as he described interspecific overgrowth of Spirorbis and several bryozoans taxa.
Numerous studies were published on the Spirorbis larval settlement behaviour and metamorphosis (Dirnberger Reference Dirnberger1990, Hamer & Walker Reference Hamer and Walker2001). Larvae of most species prefer settlement areas with poor illumination (Al-Ogily Reference Al-Ogily1985). In the White Sea, S. spirorbis (Linnaeus, 1758) is present between 1 and 20 m deep and is affected by significant fluctuations of salinity and temperature (Ushakova Reference Ushakova2003).
Spirorbid polychaetes are frequent inhabitants of hermitted shells (Williams & McDermott Reference Williams and McDermott2004). The most commonly colonized areas of the shells are the apertural notch, inner and outer lip, siphonal canal, and umbilicus (Walker Reference Walker1992, Walker & Carlton Reference Walker and Carlton1995).
Spirorbis sp. occurs in 48% of the sampled specimens of Antarctodarwinella ellioti (Fig. 4) and shows a closer association (76%) with sector C of the shell (Fig. 6). Comparing observed and expected frequencies (Table II) of Spirorbis on this sector of the shell reveals that Spirorbis shows higher observed frequencies than expected, suggesting a preference for the aperture interior of the shell. However, Spirorbis showed the same frequency on all the sectors of the aperture (Table III). Walker (Reference Walker1992) reported a similar pattern in Recent hermitted shells from Mexico and northern California.
Fig. 6. Segmented bar graph showing the distributions of taxa according to shell sector.
Table II. Contingency table of the shell sector and respective borers and encrusters for 33 randomly selected Antarctodarwinella ellioti specimens. Expected frequencies printed below observed frequencies, in brackets. γ2 and p values with 6 df.
γ2 = 27.15 P value = 0.000136
Table III. Presence/absence of boring and encrusting organisms in each sector of the aperture interior. Cochran Q test γ2 and p values with 3 df.
Encrusting organisms - bryozoans
Among the organisms living associated with hermit crabs, bryozoans are the best known. Many of them are multilayered, forming monticules, overgrowth of the gastropod shell aperture, and outward growth from the shell surface (Taylor Reference Taylor1994). However, not all bryozoans associated with hermit crabs form such a thick and conspicuous crust, as small multiserial encrusting colonies may also occur (Williams & McDermott Reference Williams and McDermott2004).
Encrusting bryozoans usually occur in the apertural notch, siphonal canal, and outer lip area of the shell (Walker Reference Walker1992). On living hermitted shells their preferred settlement location may be limited to the aperture side for a variety of reasons, such as less susceptibility to dehydration, and competition between encrusting species (Walker Reference Walker1992).
Encrusting cheilostome bryozoans on the studied shells of Antarctodarwinella ellioti are very poorly preserved and, in general, they are only revealed by small pits on the gastropod shell. This trace fossil is assigned to Leptichnus isp. and consists of pits that are sub-circular in cross section and are found in multiserial arrangements (Fig. 5a & b).
The oldest described Leptichnus is from the Upper Cretaceous, but the ichnogenus only became common during the Cenozoic. At least nine modern cheilostome taxa produce this trace (Taylor et al. Reference Taylor, Wilson and Bromley1999).
Cheilostome bryozoans (Leptichnus isp. or the organism itself) are the most abundant encrusting organisms on Antarctodarwinella ellioti. They are present on 52% of the gastropod shells (Fig. 4), and show a close association with sector C (94% of specimens) of the shell (Fig. 6). Observed frequencies are higher than expected in this sector (Table II), suggesting a preference for the aperture interior of the shell. However, its greater abundance in this sector could also be explained if the interior acted as “taphonomic refuges”. When found in area C, they show highest frequencies on the columella and on the outer lip (P < 0.001) (Table III).
Encrusting organisms - balanomorph Cirripedia
Encrusting barnacles are present on hermitted shells in modern habitats but only two appear to be truly commensal (Williams & McDermott Reference Williams and McDermott2004). They are not frequent on similar shells in the fossil record. They occur on the external shell surface and sometimes near the apertural notch (Walker Reference Walker1992, Fernandez-Leborans & Gabilondo Reference Fernandez-Leborans and Gabilondo2006).
Two species of balanomorph barnacles were identified from the La Meseta Formation (Zullo et al. Reference Zullo, Feldmann, Wiedman, Feldmann and Woodburne1988). Amongst the studied specimens of Antarctodarwinella ellioti, only one carries barnacles (GHUNLPam 25219). These are completely re-crystallized and thus impossible to identify further. They are small, with a maximum diameter of 0.4 mm, and occur in areas A and C. One specimen was recorded from the aperture notch of area C.
Encrusting organisms - coralline algae
A few specimens of Antarctodarwinella ellioti (6%, Fig. 4) show non-geniculate coralline algae with crusts only a few microns thick. Crusts show cylindrical to compressed, non-branching protuberances less than 300 µm high (Fig. 5c). Their growth form falls within the “warty” type of Woelkerling et al. (Reference Woelkerling, Irvine and Harvey1993).
Encrusting coralline algae are mainly restricted to normal marine water and are known to range from intertidal environments down to a depth of 250 m. However, the most common range in temperate-water environments is c. 20–40 m.
Boring organisms - polychaetes
Polydorids commonly inhabit gastropods shells, mainly in shells occupied by hermit crabs (Boekschoten Reference Boekschoten1966, Reference Boekschoten1967, Blake & Evans Reference Blake and Evans1973, Walker Reference Walker1988). They are commensals, feeding on materials captured by hosts or brought in through respiratory currents (Williams & McDermott Reference Williams and McDermott2004). According to Walker (Reference Walker1992), the most important feature suggesting that a shell was hermitted is the presence of the boring ichnogenus Helicotaphrichnus Kern, Grimmer & Lister, 1974. This boring is produced by spionids belonging to the genera Polydora Bosc, 1802 and Dipolydora Verrill, 1881. The boring produced by the extant D. commensalis (Andrews, 1891) is frequently associated with many species of hermit crabs and uniquely follows the columella of shells. According to Blake & Evans (Reference Blake and Evans1973) the spionid can only grow in hermitted shells and dies when the crab abandons the shell.
Helicotaphrichnus isp. and Caulostrepsis isp. (Fig. 7) are the most common boring ichnotaxa recorded and are present on 64% of the specimens of Antarctodarwinella ellioti (Fig. 4). Borings similar to those of Caulostrepsis isp. and Helicotaphrichnus isp. are produced by polychaetes of several families. Polydora is one of the better known extant boring Spionidae. The borings have a wall composed of mucus and fine sand particles. The infilling between the limbs consists of sand and detritus, cemented by mucus.
Fig. 7a. Caulostrepsis isp. (GHUNLPam 25228), b. Transverse section (GHUNLPam 25215) showing Helicotaphrichnus isp. (arrows).
Polychaetes that bore in hermitted shells occur more frequently around the aperture, generally on the columella, apertural notch, and outer lip (Walker Reference Walker1988, Reference Walker1992, Reference Walker1998). Blake & Evans (Reference Blake and Evans1973) and Kern et al. (Reference Kern, Grimmer and Lister1974) showed that boreholes produced on the columella by the spionid Dipolydora commensalis can be considered as the most important indicators of hermitted shells.
The boring polychaetes recorded in Antarctodarwinella ellioti show higher observed frequencies than expected in areas A and B (Table II). However, they are more abundant on area C, albeit not statistically significant. The frequency of polychaetes in area C is significantly higher on the columella than on other sectors of the interior aperture (P < 0.0001, Table III).
Boring organisms - bryozoans
Boring bryozoans included by Pohowsky (Reference Pohowsky1978) in the Order Ctenostomata Busk, 1852, comprise a group of ethologically defined marine species whose members live immersed in mainly calcareous substrates, with the lophophores being the only exposed parts of the organism (Warme Reference Warme and Frey1975). The substrates most frequently used by ctenostomes are gastropod and bivalve shells, although during the Palaeozoic they were colonizers of other organisms with calcareous parts, such as brachiopods and crinoids (Pohowsky Reference Pohowsky1978). Some taxa may show preference for a particular kind of shell. Several species inhabit live or dead mollusc shells. Silén (Reference Silén1947) pointed out that Penetrantia concharum Silén, 1946 inhabits only dead shells. It was suggested that larvae of this species avoid shells that still retain their periostracum (Pohowsky Reference Pohowsky1978). Information on the geographic range of living species of Ctenostomata is scanty and, according to Soule & Soule (Reference Soule and Soule1969), the known records are from latitudes lower than 64° for both hemispheres, and mainly from cold to warm temperate water.
Borings recorded on the shells of Antarctodarwinella ellioti and identified as ctenostomes consist of a regular system of tunnels and cavities. The tunnels are narrow, gently curved and branch off from a point in opposing directions at angles ranging from 20° to 40°. Primary openings are at the right or left side of the tunnels and are fusiform, tear-shaped or sub circular depending on the degree of wear. Their major axis attains a length of 0.25 mm and is aligned parallel to the tunnels. Cavities are placed in front of the bifurcation point and can be connected to each other by means of secondary subordinate tunnels (Fig. 8a–f). This trace is similar to borings produced by Terebripora d'Orbigny, 1847 and is referred to the ichnogenus Pinaceocladichnus. Most of the colonies are juxtaposed, rendering identification of ancestrulae difficult. In addition, wear contributes also to hamper accurate identification.
Fig. 8a. & b. Pinaceocladichnus isp. cutting through a polychaete boring (GHUNLPam 25207), c. Crab claw resting trace (arrow) delimited by Pinaceocladichnus isp. (GHUNLPam 25235), d–f. Pinaceocladichnus isp. (GHUNLPam 25235).
Pinaceocladichnus isp. appears in 39% of the studied specimens of Antarctodarwinella ellioti (Fig. 4) and shows higher observed frequencies than expected in areas A and B (Table II). When on area C, it appears more frequently on the columella (P < 0.01). Walker (Reference Walker1992) stated that boring bryozoans have been occasionally used to infer hermitted shells in the fossil record, and that in this case they penetrate areas thought to be inaccessible on living snails, such as the callus and parts of the columella.
Boring organisms–phoronids?
Only 3% of the studied specimens of Antarctodarwinella ellioti (Fig. 4) carried traces characterized by networks of multiapertural cylindrical borings of constant diameter and circular cross-section, showing a regular pattern involving a side branch and openings at regular intervals. This trace is assigned to Talpina isp. According to Bromley (Reference Bromley and Donovan1994), Talpina is produced by phoronids. Casadío et al. (Reference Casadío, Marenssi and Santillana2001) documented the presence of this trace in shells of Ostrea antarctica Zinsmeister, 1984, from the lower part of the La Meseta Formation.
Diversity and community succession
Ninety four percent of the studied gastropod shells show at least one boring or encrusting organism. Most of the specimens of Antarctodarwinella ellioti (79%) show between 1 and 3 epizoans and/or ichnotaxa. Only 6% specimens show no boring or encrustation at all on their surface and 15% show the maximum diversity of 4 taxa (Fig. 9).
Fig. 9. Percentage of shells of Antarctodarwinella containing 0 to 4 encrusting and boring taxa.
The segmented bar graph (Fig. 6) provides the conditional and marginal distributions of shell sectors for each one of the encrusters and borings. There is an association between some shell sectors and particular encrusters and borings. Thus, encrusting bryozoans and Spirorbis sp. have a closer association with sector C of the shell of Antarctodarwinella ellioti (94% and 76% respectively).
The contingency table (Table II) shows that there is an association between shell sector and presence of borings and encrusters, i.e. certain parts of the shell were preferentially colonized (P < 0.001). Contrasting the observed and the expected frequencies, it is possible to establish that encrusting bryozoans and encrusting polychaetes (Spirorbis sp.) show higher observed frequencies than expected in sector C, suggesting a preference for the aperture interior of the shell. Boring polychaetes and boring bryozoans show higher observed frequencies than expected in areas A and B, although this difference in frequency is smaller than the previous one.
As some of the boring or encrusting organisms recorded on shells of Antarctodarwinella ellioti showed a clear preference (encrusting polychaetes and encrusting bryozoans) for or at least a higher frequency (boring polychaetes and boring bryozoans) in area C (aperture interior) - suggesting that the shells were hermitted - there also is an association between some of the aperture interior areas and particular encrusters and borers. Thus, boring polychaetes, boring bryozoans, and encrusting bryozoans, do not show the same frequency in each area of the aperture interior (Table III); they are more frequent on the columella (P < 0.0001, P < 0.01 and P < 0.001 respectively). The encrusting bryozoans also appear to show a preference - albeit not as high as on the columella - for the outer lip.
Recruitment of encrusting bryozoans and Spirorbis sp. was simultaneous, as mutual overgrowth was observed. Also observed were algae covering specimens of Spirorbis sp. Boring polychaetes were among the first to settle on the shells, as colonies of bryozoans were observed crossing through the eroded polychaete borings (Fig. 8a & b).
Conclusions
The identities and placement of encrusting and boring organisms, together with the inferred infaunal habit of Antarctodarwinella ellioti, suggest that shell colonization began after the death of the gastropods.
A Chi-square Independence Test revealed that the community of encrusting and boring organisms recorded on Antarctodarwinella ellioti showed a clear preference for the aperture interior area of the shell. A subsequent Cochran Q Test indicated that the differences in frequency of encrusting and boring organisms as counted on the different aperture interior areas were statistically significant. Thus, boring polychaetes, boring bryozoans, and encrusting bryozoans are more frequent on the columella. Encrusting bryozoans also appear to show higher preference on the outer lip. This association of boring and encrusting organisms and their distribution on the shell support that the specimens of Antarctodarwinella ellioti were inhabited by hermit crabs.
The community of encrusting and boring organisms associated with shells of Antarctodarwinella ellioti is similar to those described in recent hermitted shells from temperate environments in mid-latitudes of the northern hemisphere (Stachowitsch Reference Stachowitsch1980, Walker Reference Walker1988, Walker & Carlton Reference Walker and Carlton1995). These similarities (predominance of spirorbids, spionids, and encrusting bryozoans) suggest that this type of community was already established in the Eocene. However, the Antarctic hermitted shells inhabited a temperate, shallow marine environment subject to long periods of darkness and low food availability, suggesting that one of the most important environmental factors conditioning the development of these communities was water temperature.
Acknowledgements
We thank the Instituto Antártico Argentino (IAA) for support during the field season, S. Santillana who provided invaluable field support and useful discussions, M.C. Martin for help with the statistical analyses, and P.D. Taylor and M.B. Aguirre-Urreta for reviews which improved the manuscript.
