Fossils, sediments, tectonics

Facies (2006) 52: 361–380 DOI 10.1007/s10347-006-0048-2

ORIGINAL ARTICLE

Margret Steinthorsdottir · Scott Lidgard · Eckart H˚akansson

Fossils, sediments, tectonics Reconstructing palaeoenvironments in a Pliocene–Pleistocene Mediterranean microbasin Received: 27 February 2005 / Accepted: 6 June 2005 / Published online: 29 April 2006 C Springer-Verlag 2006


Abstract Tectonic displacement and small-scale tsunamis apparently affected deposition of the Kolymbia limestone, Cape Vagia, Rhodes, Eastern Mediterranean. Coarse beds interrupt the sequential build-up of this Pliocene–Pleistocene bryomol limestone. Celleporid bryozoans, bivalves, and brachiopods dominate these beds. The palaeoecology of the thicket-forming Celleporaria palmata is re-evaluated and subsequently revised. The limestone comprises two parasequences in a transgressive systems tract, and deposition occurred at palaeodepths between 30 and 120 m. At intervals, tectonic movements lowered relative sea level and sent slumps of shallow-water fauna downslope. The depositional history was validated using independent sets of data: sediment structure and grain size, palaeobathymetry using bryozoan growth forms and occurrences of modern representatives of bryozoans and other taxa, basin configuration, and regional tectonics. Concordance of these lines of evidence provides a means of evaluating confidence in palaeoenvironmental inferences. Keywords Celleporaria . Bryozoans . Palaeoecology . Cool-water carbonates . Eastern Mediterranean . Pliocene–Pleistocene M. Steinthorsdottir · S. Lidgard (
) Field Museum of Natural History, 1400 South Lake Shore Drive, Chicago, IL 60605, USA e-mail: [email protected] Tel.: +1-312-665-7625 Fax: +1-312-665-7641 E. H˚akansson Geological Institute, University of Copenhagen, Øster Voldgade 10, 1350 Copenhagen K, Denmark Present address: M. Steinthorsdottir Department of Geology, Trinity College, Dublin 2, Ireland

Introduction Formation of carbonate sediments outside the tropical realm has become a widely observed phenomenon (Lees and Buller 1972; Nelson et al. 1988; James 1997; Wright and Burchette 1998; Mutti and Hallock 2003). Bryozoanrich facies are common in these sediments and beds formed mostly or entirely by bryozoan skeletons are not infrequent. Studies of cool-water carbonates in both modern and past environments are now revealing more subtle differences in tectonic settings, admixtures of skeleton-producing organisms, palaeoecological and taphonomic patterns, and resulting facies architectures. Factors influencing carbonate production are more numerous than previously thought and typical cool-water associations (foramol, rhodalgal, and bryomol) do occur in tropical settings (Pomar et al. 2004). Still, the principal models for cool-water carbonate deposition and the overwhelming majority of examples come from just two well-studied regional contexts: large open ocean platforms and shelves such as those in South Australia and New Zealand, and ramps such as those in the Western Mediterranean (Hayton et al. 1995; Henrich et al. 1995; James 1997; Pomar 2001; Pedley and Grasso 2002; Lukasik and James 2003), though the latter are less open, with lower wave energy, less tidal flux, shallower stormwave bases, and clearer water. Detailed geological studies of cool-water carbonate depositional histories in smaller, more closed basins and in palaeooceanographic settings with relatively sluggish circulation and normally lower energy regimes such as those of the Eastern Mediterranean are few in number. Similarly, while recent analyses of tectonic controls (Butler et al. 1997), taphonomic factors (Brachert et al. 1998; Yesares-Garcia and Aguirre 2004), and actualistic palaeobathymetry (Moissette 2000a, b; Scarponi and Kowalewski 2004) have progressed toward greater objectivity in interpreting cool-water carbonate deposition and facies architectures, few case studies have attempted to synthesize these different lines of evidence. The aim of the present study is to infer and validate the depositional history of a highly fossiliferous bryomol

362

limestone from the Pliocene–Pleistocene in a characteristic tectonic microbasin in the Eastern Mediterranean. A secondary aim is to infer the life mode and habitat of bryozoan thickets built by Celleporaria palmata in this limestone. Our approach is synthetic. We combine literature-based local and regional data to establish a tectonic context, then stratigraphical, palaeontological, and sedimentological data to reconstruct the palaeoenvironment and the depositional history. We first place the island of Rhodes in a regional geologic framework, and then summarize the stratigraphy and environment of the study area, Cape Vagia. We describe the stratigraphic column of the Kolymbia limestone, dividing it into facies and beds. The results of field and laboratory work are then listed with interpretations of these data. We infer the mode of deposition of each bed, based on sedimentological data and palaeoecological evidence derived from the fossils. The different fossil groups and the various sets of palaeoecologic/bathymetric indications are outlined separately in an attempt to convey their input objectively. The fossil assemblages are analyzed within this perspective and then evaluated using bathymetric data from extant members of the assemblages. The data are then brought together to reconstruct the depositional history of the Kolymbia limestone. The creation of the basin topography is inferred and the depositional frame, within which evidence from all organisms must fit, is established. Our interpretations are supported by a suite of independent lines of evidence that are congruent with each other, providing a validating test to our hypothesis.

Geological setting

Fig. 1 A The position of Rhodes in the Aegean Sea as well as the location of the study area, the bay south of Cape Vagia, on the east coast, between Rhodes’ two main cities (after Spjeldnaes and Moissette 1997). B The sketch map shows the general configuration of

the Cape Vagia microbasin, position of major faults, the distribution of the deposits and the most important topological features, as well as the position of localities 1–4 and the distribution of the studied samples

The Greek island of Rhodes in the Aegean Sea is at the eastern end of the sedimentary part of the Hellenic arc (Fig. 1A). The arc is one of the southernmost nappes of the Alpine system on the southern rim of the Anatolian Plate in an active tectonic regime (McKenzie 1978; Pirazzoli et al. 1996; Kontogianni et al. 2002). The tectonic framework is dominated by collision of the Arabian and African plates with Eurasia (Oral et al. 1995). The African plate is presently subducting below the Aegean plate along the Hellenic arc, involving complex vertical movements and crustal extension (Dewey et al. 1986). The Hellenic Trench in front of the subduction zone is followed northwards by a sedimentary arc and then a volcanic arc. The Aegean Sea and particularly the Hellenic arc are seismically active and among the most tsunamigenic areas in Europe (Perissoratis and Papadopoulos 1999; Dominey-Howes 2002). Compared to the Western Mediterranean, where there is scant evidence for seismically triggered slope instabilities, the Eastern Mediterranean region is characterized by frequent failures and sediment movements on bedding planes (Mienert et al. 2002). The Eastern Mediterranean is warm-temperate, transitional to the true reef environments of the tropics, as indicated by the presence of banks of the zooxanthellate coral Cladocora caespitosa in some protected situations in the Aegean (Henrich et al. 1995). In the Pliocene, the area may have been even warmer, as the sea-surface temperatures may have been slightly higher than today (Haywood et al. 2000; Nebout et al. 2004).

363

Rhodes consists mostly of Mesozoic and Caenozoic rocks, the local basement, which were folded and faulted late in the Alpine orogeny. A Late Pliocene and Pleistocene postorogenic sediment cover of varying thickness overlies the basement (Mutti et al. 1970; Meulenkamp et al. 1972; Hanken et al. 1996). Tilting occurred in the Late Pliocene, raising northeastern and lowering southwestern Rhodes. The area apparently became divided into blocks with differential vertical movements bounded by normal faults (Mutti et al. 1970; Pirazzoli et al. 1983, 1989; Hanken et al. 1996; but see Kontogianni et al. 2002). Further rise and differential tectonic movements continue to the present day (Meulenkamp et al. 1972; Benda et al. 1977). This regime exposes Rhodes to rapid, localized, tectonically driven sea-level changes in addition to regional PlioPleistocene glacial eustacy (Kovacs and Spjeldnaes 1999; Hansen 2001). In 1996, Hanken et al. proposed a new formal lithostratigraphy using facies mapping. They recognized three major transgression–regression cycles, subdividing the stratigraphic succession into three formations: the Lindos Acropolis, Rhodes, and Kritika formations (Fig. 2). There has been little Holocene sedimentation. A number of littoral notches, commonly between five and eight, have been incised within the 3.75 m zone above the present sea level, recording post-glacial sea-level fall as well as active coastal uplift. The maximum Holocene transgression is about 4 m (Pirazzoli et al. 1989; Hanken et al. 1996; Kontogianni et al. 2002). The Eastern Mediterranean in general is an enclosed, microtidal and calm sea. The subsurface topography of Rhodes is generally very steep, which in other settings has been reported as setting the base wave and swell-wave level at quite shallow depth, approximately 30–80 m (Hageman et al. 2003). The study area, the bay south of Cape Vagia, is a microbasin opening to the south (Figs. 1B and 3A). The exposed succession of Kolymbia limestone and Lindos Bay clay is draping or onlapping Jurassic meta-limestone basement (Hanken et al. 1996) that is hard, dark grey, and severely eroded. Karstification, marine mechanical erosion expressed by surf caves and notches, and marine bioerosion are apparent. The basement is locally broken into blocks that rest at the foot of solid basement ridges. Cape Vagia sensu stricto is approximately 45 m high, bound by a major normal fault to the ESE (see Fig. 1B). The basement headland to the west is about 15 m high with large basement blocks at the foot of the headland, forming cryptic environments (caves) between them. To the north the basement rises gradually to about 30–40 m, obscured by later deposits, vegetation, and building constructions. These basement highs were sites of a “carbonate factory” for the limestone deposited in the basin. The basin was divided into four localities, based on differences in facies architecture and composition (Fig. 1B). In the western part of the basin, the Kolymbia limestone is exposed in a small section close to sea level, succeeded by the Lindos Bay clay, which forms a steep cliff (locality 1). The largest portion of the succession is exposed laterally in the northern part of the bay (locality 2). Towards the tip

Fig. 2 The stratigraphy of the Plio-Pleistocene succession of Rhodes. Three formations are recognized. The deposit studied here is the Kolymbia limestone, part of the Rhodes formation (from Hanken et al. 1996)

of Cape Vagia, the Kolymbia limestone drapes the steep basement (locality 4). In the southwestern part of the basin, it onlaps large basement blocks (locality 3). For this study, the majority of samples are derived from locality 2, due to the superior exposures and stratigraphic range of that locality (Fig. 3B). The Kolymbia limestone is represented by four facies (Kl 1–4) and the Lindos Bay clay by one (Lbc 1). Out of numerous samples collected in the field, 11 samples were analyzed in detail. The position of the samples in the basin is marked in Fig. 1B and their distribution in the stratigraphical column is illustrated in Fig. 4. Material and methods A simple geological map was constructed since detailed topographical maps are unavailable (due to strenuous relations with nearby Turkey) (Fig. 1B). Three partial

364

Fig. 3 Two photographs showing: A the configuration of the Cape Vagia basin, with basement, deposits, and the main locality indicated, and B a close up of locality 2, showing the appearance of facies in

the field and the relationship between the Kolymbia limestone and the Lindos Bay clay

stratigraphic section logs were combined into a composite log, correlating the distinct beds across the basin (Fig. 4). Representative bulk samples of 0.5–1.5 kg were collected throughout the study area. The sediment has undergone diagenesis and some beds are reworked, making species identification difficult and time-consuming for some fossils, primarily bryozoans. Effort was therefore focused on clarifying the composition and origin of the fauna of the celleporid beds (bed 4 and 10), as well as the coarse shell beds, the oyster-pectenid bed (bed 6) and the Gryphus bed (bed 8). Two samples were included from the bryozoan limestone, which separates the coarser beds as well as constitutes the matrix in those. The most effective method for separating the fossils was treatment with a solution of Glauber’s salt (Na2 SO4 ·10H2 O), poured over a sample, which was then frozen and thawed 10 times. Samples were then sieved through four different mesh sizes (4, 2, 1, and 0.5 mm), dried, weighed, and studied using a microscope. The 1–2 mm fraction was in most cases split into four subsamples using a designated sample-splitting machine, minimizing the risk of losing information. The three largest fractions were sorted into different faunal and non-faunal groups, which were weighed in order to establish the relative abundance of each group. Bryozoans were classified to genus or species level when possible and 23 nominal species of cheilostomes and cyclostomes were examined using a scanning electron microscope (SEM). The larger fossils, the bryozoan C. palmata, the brachiopods Gryphus vitreus and Novocrania anomala, the bivalves Pecten jacobeus and Ostrea lamellosa were identified and information about their modern ecological preferences

gathered. Rhodoliths, echinoderms, and Lindos Bay clay bryozoans were examined for ecological purposes only. Stratigraphy The Kolymbia limestone is exposed along the east coast of Rhodes, including its type section at Cape Vagia. The lithology is dominated by highly fossiliferous limestone, mostly rudstone, sometimes grading into floatstone (Cuffey 1985) and reworked limestone facies associations. The limestone is generally fining upwards, with increasing content of terrigenous silt and clay. Ichnofossils are abundant, and together with body fossils and lithology, they indicate an increasing water depth towards the top (Hanken et al. 1996). Deposition of the limestone during a major transgression is also a conclusion of subsequent studies (Moissette and Spjeldnaes 1995; Spjeldnaes and Moissette 1997; Hansen 2001). The transgression moved across a steep, irregular coastline. Facies architecture, composition, and palaeontology were strongly influenced by local palaeo-relief. The sessile benthic communities provided local sources of carbonate grains for the deposition of carbonate sediments below wave erosion depths. As the transgression continued, carbonate production rates did not keep up with the rising sea level and drowning occurred, resulting in gradation of the Kolymbia facies group into the Lindos Bay clay. However, this drowning is likely to have occurred at slightly different times in individual basins, reflecting the size and distribution of local carbonate factories.

One bed; onlaps karstified basement at locality 2. Overlying a sequence boundary. Water-escape structures.

Two beds (4 and 10), in Kl 1 type matrix. Inclination ca. 18◦ . Large lateral propagation.

Three beds (2, 6, and 8), in matrix of Kl 1 type.

Gradual contact over 10–15 cm with underlying Kolymbia limestone.

Kl 2 Muddy calcarenite

Kl 3 Celleporid beds

Kl 4 Shell beds

Lbc 1 Terrigenous silts and clays

Texture and grain types

Rudstone to floatstone. Bed 2 dominated by O. lamellosa. Bed 6 dominated by P. jacobeus with O. lamellosa. Bed 8 dominated by G. vitreus. Stiff blue-greyish to greenish calcareous clay, with fine, mm-thick beds. A thin red ash bed 90 cm above the LBc–Kl boundary.

Dark brown-grey mudstone Mud content 85–95 wt.%, mainly reworked carbonate and terrigenous material. Probable end-member of facies Kl 1. Rudstone. Large bioclasts of C. palmata branches dominate. Bed 10 is thicker and denser, traceable across basin. Outsize bored basement clasts.

Brown-greyish rudstone to floatstone. 4 mm fraction as well. Brachiopods and bivalves are also abundant. Rhodoliths, echinoids, foraminifers, gastropods, and arthropod fragments occur as subdominant components. This abundance in fossils, many of them having modern representatives, provides an opportunity to extract palaeoecological information and to check environmental inferences from each group against the others. Bryozoans are by far the most diverse and abundant of the fossils in the Kolymbia limestone and are the main focus of the palaeoecological analysis. The celleporid thickets in the Kolymbia limestone represent a very different scale of occurrence from the rest of the bryozoan fauna. We treat them separately as well as together with other bryozoan occurrences. Other fossils from the limestone are mainly employed as additional validation tests of the inferred setting in which the limestone was deposited. Bryozoans The bryozoan fauna is consistent through facies in terms of species present and the approximate abundance of each of those species. About 23 species are the dominant components and perhaps five more species occur in smaller but persistent amounts. Spjeldnaes and Moissette (1997) reported 64 bryozoan species from the Kolymbia limestone at Cape Vagia, including those from a single bulk sample. They reported 12 species found exclusively encrusting the celleporids; these species were not included in this study. Disregarding these 12 species, about 30 species from their list are still unaccounted for. There are also some substantial discrepancies between the list determined in this study and that of Spjeldnaes and Moissette (1997). One of the most common bryozoan species in all the samples studied here (Umbonula sp. 1) is not present in their list. Furthermore, these authors only report one distinct bed with C. palmata in the Kolymbia limestone, whereas in fact there are two.

Vinculariiform

Adeoniform

•

• •

•

•

•

•

Cellariiform

•

•

•

Reteporiform

•

•

Facies 1

Crisia denticulata Onychocella angulosa Patinella radiata Puellina (C.) innominata Margaretta sp. aff. cereoides? Adeonella polystomella Diporula verrucosa Frondipora verrucosa Cupuladria doma Buskea dichotoma Hippellozoon mediterraneum Hornera frondiculata Chaperia annulus Cellaria salicornioides Tessaradoma boreale Setosellina capriensis Tervia irregularis Celleporaria palmata Umbonula monoceros Umbonula sp. 1 Metrarabdotos (porometra) helveticum canariense Smittina cervicornis •• • ••• ••• ••

•

• ••• • ••

••

• • • •

Proximal Relative abundance (species per sample) 4

Note. • rare; •• present; ••• frequent; •••• abundant

•

•

•

•

• •

•

• • •

Membraniporiform

Growth form

Celleporiform

•••• •

•••• ••

• •• •

•• ••• • • ••• • •

Distal 8

•• •••• • •••• •

•

•• • • ••

••

••• •• •• • •••• •

Proximal 16

Facies 3

•• ••• • •••• •••

•• •• •• •

••• •• •• •• •

3

•• •••• • •••• •••

•• •• • ••• •• •

• • • •• •• ••• •

20

•• •••• •• •••• •••

• •• • •• •

•

••• ••• •• • ••• ••

23

•• •

•• ••••

• •

• ••• •

• ••• • •

••• ••

26

Growth forms and relative abundances of bryozoan species among samples from the Kolymbia limestone, Cape Vagia, Rhodes, Greece

Lunulitiform

Table 2

• ••• •••• ••• •• •••

••• • • •• •••

Distal 25

••

•••• • ••• •

• ••• •• •• •••• ••••

••• • • •

• •

••• • • • •••

24

•••• ••• • ••• •••

••

•••• •• •• ••

••

•• ••

•• ••

18

• ••• • • ••

•

•••• •• • •

••• • •

• • •

19

Facies 4

368

369

Recognition to species level was impeded somewhat by the preservational state of fauna. Most bryozoan colonies are smothered in fine-grained sediment (mud-fraction) and are slightly re-crystallized. Bryozoans in a sample from a fossil cave are however in better preservational state and confirm the unexpectedly low number of species found in the samples from the basin. Twenty-two abundant and easily recognized bryozoan species were studied in detail and are listed in Table 2. An additional species, Metrarabdotos moniliferum, is very abundant in a sample from the cryptic cave, but present only sporadically in the basin. We apply several different techniques to infer environmental provenance from the bryozoans, including different approaches to comparative growth-form analyses and inference of palaeobathymetry using depth range overlaps from extant species. We consider data showing relative dominance of different colony growth forms, using species counts and relative abundance as separate metrics. We consider separately species occurrences in different limestone strata and facies, then apply depth ranges of living conspecifics using an assumption of uniformitarianism. Modern bathymetric range data are derived from published literature and studies of modern celleporid thickets occurring elsewhere. The results for the Kolymbia limestone are compared to results for bryozoans from the above lying Lindos Bay clay. We then summarize the results derived from this comprehensive study of the bryozoan fauna. Bathymetric and other ecological attributes of non-bryozoan taxa are used principally in post-hoc comparisons that either validate or dispute palaeoecological inferences drawn from the bryozoans. Bryozoan growth forms as ecological indicators The growth forms of bryozoan colonies correlate to varying extents with the environments in which they live (Stach 1936; Smith 1995). A number of different types of growthform classification schemes have been put forward over the years (e.g., Stach 1936; Lagaaij and Gautier 1965; Schopf 1969; McKinney and Jackson 1991; Hageman et al. 1997, 1998; Kaselowsky et al. 2005). One of the most commonly used is the archetypal classification, based on comparison to archetype growth forms, where each group is given the name of an archetypal bryozoan taxon. This rationale is simple and often yields interpretable results when applied to fossil material. The disadvantages include the somewhat arcane terms and lack of consensus on which growth forms to include in wider groups and which to assign to their own groups. Therefore, some archetypal classifications are minimalist, with fewer than 10 morphogroups while others are more complicated with more than 20 morphogroups. The most severe problem of this classification is that it presumes that bryozoan colony form is conservative or invariant within species, which usually it is not (McKinney and Jackson 1991). Nor have the palaeodepth predictions of archetypal classifications been confirmed in a wide range of modern environments. An archetypal classification is used in the

Fig. 6 An archetypal classification of bryozoan growth forms. Seven morphological forms are distinguished, each named after archetype forms (revised after Moissette 2000b)

present paper, due to its simplicity and prevalence, but also on grounds of precedence and comparison, as the classification has been applied to fossil bryozoan assemblages from the studied area before (Moissette and Spjeldnaes 1995; Spjeldnaes and Moissette 1997). An archetypal scheme modified after Moissette (2000b and references therein) is given later showing the most commonly used names for the seven colony forms occurring in this study (Fig. 6). If a predictive relationship between growth-form dominance in assemblages and (palaeo)environment can be established, it can be an effective tool in predicting palaeoenvironments in unknown settings, considering the great abundance of bryozoans today and in the fossil record. Relative species richness and abundance of bryozoan colony forms do change with changes in the environment, notably depth (McKinney and Jackson 1991). Bathymetric distributions of changing growth-form dominance patterns are thus used to interpret palaeobathymetry. This method was developed by Stach (1936) and has since been modified and applied by numerous workers (e.g., Schopf 1969; Hageman et al. 1998; Moissette 2000a, 2000b). However, two types of metrics, relative taxonomic richness and relative abundance, have sometimes been confused in the literature. Relative taxonomic richness usually refers to the proportion of species with different growth forms in a given environmental setting. Relative abundance usually refers to the proportion of skeletal mass, volume, or living biomass in different growth forms in a given environmental setting. Environmental settings for both these metrics have then been abstracted as ecological parameters such as depth, current regime, grain size, or bottom type. In this study, we apply both metrics in independent trials to infer palaeoenvironmental conditions. Moissette (2000b) used relative abundance of each colony form as a measure of the depth interval at which a given assemblage lived (rather than using taxonomic richness and presence/absence). These inferred preferences are used here.

370

Taxonomic patterns of colony form dominance The distribution of colony growth forms within the studied species is given in Table 2. Within the model of Moissette (2000b), the encrusting membraniporiform and celleporiform colonies prefer solid and loose substrate, respectively, whereas both the growth forms are typically more abundant in environments with moderate to strong currents and slow sedimentation. The membraniporiforms predominate in shallow water, the celleporiforms live in slightly deeper water. The adeoni-, vincularii-, and reteporiform colonies are more abundant on solid substrates, with low to moderate currents and slow sedimentation, at moderate depths. The vinculariiforms are typically more abundant in deep environments. The cellariiforms colonies prefer solid or loose substrate, low to moderate currents, slow to moderate sedimentation at shallow to moderate depths. The lunulitiforms are more abundant on loose substrates, endure moderate currents and slow to rapid sedimentation mostly at moderate depths. When our data are evaluated using this model, the inferred ecological conditions that result include a hard substrate, low to strong currents, slow sedimentation rate, and moderate depth. Abundance patterns of colony form dominance Abundances were determined for 11 samples in four abundance categories (Table 2). These categories, rather than continuous metrics, were used based on results from a previous study (Steinthorsdottir 2001). One of the conclusions of this earlier study is that whereas continuous metrics tend to show high variance and potential sample bias, categorical metrics reduce variance and show stronger consistency among workers. Abundances were determined from the three most studied facies, facies Kl 1 (bryozoan limestone), Kl 3 (celleporid beds), and Kl 4 (shell beds). Adeoni- and vinculariiform bryozoans dominate in all three. There is no clear trend in the distribution of species between the facies or samples, but there is a definite difference in the species abundance from sample to sample. When applying the growth-form model of Moissette (2000b) to the bryozoan growth forms using abundance as a parameter, instead of mere presence or absence of species, the environmental factors for adeoniform and vinculariiform bryozoans are accentuated. These are a preference for solid substrates, low to moderate currents, and slow sedimentation at moderate depth.

depth of 60 m and the largest overlap of bathymetric ranges occurs between 30 and 80 m. Some species have also been assigned a “preferred” bathymetric range, in which they are most commonly or abundantly found. Registration of number of species within each bathymetric interval was repeated, weighting the preferred bathymetric ranges double (i.e., counting the species twice within the preferred bathymetric range and once within the normal bathymetric range). The resultant largest overlap was also 30–80 m. Bryozoan reefs and thickets A growing literature exists on reefs and thickets constructed by bryozoans (e.g., Cuffey et al. 1977; Bradstock and Gordon 1983; Gordon et al. 1994; Henrich et al. 1995; Hageman et al. 1998; Batson 2000; Betzler et al. 2000; Pedley and Grasso 2002; Cranfield et al. 2003; Lukasik and James 2003; Cocito 2004; James et al. 2004). These structures range from micro-scale patch reefs (Scholz and Hillmer 1995) to large continuous build-ups, sometimes even dominating carbonate production completely (Hayton et al. 1995; Cocito and Ferdeghini 2001). The range of bryozoan constituents in terms of size, form, and phylogeny is also wide. The settings in which these structures occur are likewise variable and the controlling factors are many. Celleporaria sp. is a common builder of thicket-forming lump- or tree-shaped colonies (Bradstock and Gordon 1983; Gordon et al. 1994; Hageman et al. 1998; Batson 2000; Pedley and Grasso 2002; Grange et al. 2003). With more new studies emerging, the controlling ecological factors seem to become more complicated and therefore similar bryozoan structures appear in different geological settings. In a study of the environmental controls on Celleporaria, Hageman et al. (2003) demonstrate that celleporid thickets can be found at different depths and settings, but all are associated with mesotrophic conditions, moderate sedimentation rates, and relatively low energy (below swell-wave base). These authors also considered a mud-silt substrate to be an important controlling factor, but since the colonies grow as epibionts, the encrusted organisms as much as the depositional environment influence Celleporaria distribution patterns. Celleporid recruitment and growth strategy is seemingly an opportunistic utilization of resources when present (Hageman et al. 2003). In this paper, we use the celleporids in ecological interpretations, in line with the rest of the fossils. C. palmata thickets

Bathymetry based on extant bryozoan species’ ranges Known bathymetric ranges of extant species also present in the fossil assemblage were compared as a means of reconstructing palaeobathymetry (Fig. 7). The bathymetric ranges were divided into 10 m intervals to 260 m, below which only five eurybathymetric species prevail. The number of species in each interval was calculated and the largest overlap of species established. All species are present at the

Large-scale bryozoan thickets in the Kolymbia limestone are essentially monospecific, constructed by C. palmata. The colonies are celleporiform, constructed by chaotic, frontally budded zooids, forming massive, multi-layered arborescent structures (Fig. 8). The branches retain more or less the same thickness throughout their length, and do not fuse to three-dimensional structures or have heavy, more nodular bases. The celleporid branches are 1–4 cm

371

Fig. 7 Bathymetric ranges of extant species of bryozoans in the Kolymbia limestone. Darker intervals represent preferred bathymetric range. Largest overlap (dotted vertical bar) is between 30 and 80 m; all species are present at 60 m depth (data from Cheetham

thick and each segment is up to 15 cm long. The branch diameters are bimodal, 1–2 cm and 3–4 cm. Neither juvenile colonies, nor bases from which the bryozoans could have grown, were observed. Relatively few small, nodular colonies were found. A somewhat flattened stem hole occurs in the middle of many of the branches, with a slightly greater abundance in thicker branches than thinner ones. This cylindrical stem hole is uniformly about 2 mm thick and represents an organism, which the colonies encrusted and used as support for initial growth. The ecology of C. palmata interpreted from the literature and our observations can be summarized as follows: the celleporids initiated their growth upon cylindrical organisms, probably gorgonians, which supported and anchored them on the hard substrate of the basement. The habit of growing as epibionts on other organisms is a common strategy for bryozoans (e.g., Cocito et al. 2000; Hageman et al. 2000). Indeed, some species of Celleporaria seem to recruit only as epibionts (Hageman et al. 2003). Though the individual branches are now “broken”, this fragmentation may be due to their original configuration as pseudo-cellariiform (falsely jointed) as well as to compaction and pressure from overlying deposits. Some of the branches have been broken in situ, as indicated by the broken counterpart continuing on the other side of the neighboring branch. Despite careful observation and measurements, no preferred orientation for the celleporid bryozoan branches could be detected. Colonies anchored on firm substrates may have required

1967; Hayward and Ryland 1979, 1985; Moissette 1988; Moissette and Spjeldnaes 1995; Moissette et al. 2002; Pouyet and Moissette 1992; Ryland and Hayward 1977; Spjeldnaes and Moissette 1997; Zabala and Maluquer 1988)

Fig. 8 One of the most spectacular colonies of C. palmata in the Kolymbia limestone (photo: Margret Steinthorsdottir)

strong disturbance to be detached and transported more or less intact. Spjeldnaes and Moissette (1997) inferred a depth preference for the celleporids of 30–50 m. This inference was

372

based on modern bathymetric ranges of membraniporiform bryozoans species encrusting them. The epibiont attachment of colonies on gorgonians is similar to the extant celleporid species Turbicellepora avicularis, which builds pseudo-cellariiform colonies (Pouyet 1973). The gorgonian may have been similar to Lophogorgonia ceratophyta, which has the appropriate height, construction, and substrate preference in the Mediterranean. Several tracemaking species were present, among those Culicia woodii and Gastrochaena (Rocellaria) dubia indicate fairly shallow water (Bromley and Asgaard 1993; Tillbrook 1997). Bryozoans from the Lindos Bay clay Samples from the Lindos Bay clay were not analyzed in the present study, but data on the bryozoans from that deposit are given in Moissette and Spjeldnaes (1995). The authors infer depths of about 400 m for the lower parts of the clay, decreasing upwards to about 250 m, using data obtained from bryozoans, ichnofossils, and other fossils. However, when using their data on known bathymetric ranges of extant bryozoan species found in the Lindos Bay clay for constructing an overlap, in the same way as for the bryozoans from the Kolymbia limestone, the largest overlap was found to be between 100 and 600 m. When constructing an overlap using only bathymetric ranges of bryozoans found in the Cape Vagia section, the same overlap of 100–600 m was found. The difference in largest overlap in bathymetric ranges between the Kolymbia limestone and the Lindos Bay clay supports the deposition model of transgression at Cape Vagia. Palaeoenvironment inferred from bryozoans Although there is no overriding trend among different facies or samples, there are small variations in the abundances of species, some of which can help in the interpretation of facies and reconstruction of the depositional history. The consecutive bathymetric ranges of the studied bryozoan species range over 4300 m. Yet all ranges coincide at 60 m and the largest overlap of species is between 30 and 80 m. This result is also consistent with inferences from growth forms based on both taxonomic richness and abundance. The bulk of the bryozoan fauna present in the Kolymbia limestone at Cape Vagia thus probably lived in a range around 30–80 m. The full range of species is conservatively estimated to 10–120 m. The fauna is quite mixed, with shallow- and deep-water forms deposited together. Brachiopods There are a number of brachiopod species in the Kolymbia limestone. The most distinctive of these are G. vitreus, deposited in a shell-rich bed (facies Kl 4, bed 8) and distributed unevenly in small amounts in facies Kl 1

sediment, and N. anomala, encrusting on celleporids or basement. G. vitreus at Cape Vagia are rather thick-shelled and differ in this respect from specimens normally described in the literature. Typically this brachiopod lives on the shelf and in the bathyal zone on hard substrates. It is associated with N. anomala at depths from 120 to more than 1000 m in the Mediterranean today, perhaps with a bathymetric preference for approximately 150–300 m (Logan 1979; Logan et al. 2004). Emig (1989) and Emig and Garc´ıa-Carrascosa (1991) studied the distribution of G. vitreus along the Mediterranean continental margin and defined five zones of density versus depth. The G. vitreus biocoenosis (bathyal detritic sand, dominated by the brachiopod) is reported to form a belt at depths of approximately 100–250 m. The distribution of the biocoenosis is directly related to moderate bottom currents, moving perpendicular to the slope. The density per square meter of G. vitreus in bed 6 is approximately 100 specimens/m2 , thus intermediate to zones 2 and 3 of Emig (1989), at approximately 120–140 m depth. The bathymetric distribution of the thick-shelled specimens in the Kolymbia limestone may be shifted upwards relative to the normal thin-shelled form (Spjeldnaes and Moissette 1997). Bivalves A detailed palaeoecological study from the Pliocene of Southern Spain showed that in general, bivalve assemblages from shallow, coarse-grained deposits construct strongly ribbed, inflated shells, like the shell of P. jacobeus (Aguirre et al. 1996). P. jacobeus can sometimes predominate in coarse-grained sediments, such as coarseto-medium sand and calcirudites. When found in calcirudites, the other main components are oysters, coralline red algae, mytilids, arciids, serpulid worms, balanids, and bryozoans (Aguirre et al. 1996). P. jacobeus currently lives at depths of 25–183 m (Hallquist and Hansen 1997). The specimens in the Kolymbia limestone were probably living distally in the basin at depths of ∼ 50–120 m on facies Kl 1 sediment bottom. O. lamellosa typically constructs banks in shallow water. Tropical littoral oysters grow from the surface down to 10–20 m (Laborel and Laborel-Deguen 1996). The recent species Ostrea edulis, which is quite similar to O. lamellosa, lives in environments with normal salinity, in very shallow water (Laurain 1984). The oysters at Cape Vagia were living in shallow water at the slopes or on top of the basement carbonate factory. They encrusted basement and the shells of one another. Lithophaga lithophaga bores into hard, vertical to overhanging surfaces, creating the ichnofossil Gastrochaenolites torpedo. The Mediterranean species L. lithophaga is restricted to extremely shallow marine environments and is considered an excellent indicator of upwards co-seismic movements (Laborel and Laborel-Deguen 1996). Its optimal zone of occurrence is immediately below sea level, giving it a very narrow range in the microtidal Mediterranean Sea. It is abundant at depths from 0 to 1 m;

373

population density decreases as depth increases to about 10 m, below which it is rarely found. L. lithophaga is intolerant of sedimentation and the optimum orientation of the boring is therefore horizontal, whereas the preferred substrate is vertical (Bromley and Asgaard 1993). The ichnofossil G. torpedo is abundant in the basement limestone at Cape Vagia, indicating wide fluctuations of sea level within the basin. Gastrochaena (Rocellaria) dubia is represented by the characteristic “figure-of-8” shaped ichnofossil Gastrochaenolites isp. on C. palmata colonies in this study and in the Pliocene Coralline Crag (Tillbrook 1997). G. (Rocellaria) dubia is also indicative of shallow conditions, ranging down to several tens of meters (Bromley and Asgaard 1993). Rhodoliths Rhodoliths are dependent for growth on photosynthetic algae and bacteria and thus form within the euphotic zone. Morphologically the rhodoliths in the Kolymbia limestone are similar to Lithophyllum racemus, which lives on hard substrate and sandy bottom in shallow water in the Mediterranean today (Riedl 1983). Lithophyllum sp. in a study from the Holocene of the Mediterranean occurred at depths ranging from very shallow to just below 20 m. Coralline algal build-ups were associated with several large gorgonians (Sartoretto et al. 1996). In the Kolymbia limestone, the rhodoliths are always found abundantly in association with the celleporid thickets in facies Kl 3 and in somewhat smaller amounts in the other facies. They do not encrust the bryozoans, which may indicate a distinct and probably shallower provenance (cf. Ferdeghini et al. 2000; Corda and Brandano 2003; Pomoni-Papaioannou et al. 2003). They probably lived on the top and upper slope of the basement carbonate factory and were transported downslope together with the celleporids.

Depositional history Basin creation Steep normal faults bounding the Cape Vagia basin dip to the east; the cape is the footwall block and the basin bottom is the hanging wall block of the eastern fault (Fig. 1B). The overall geometry is that of a half-graben, creating a microbasin between the raised footwall blocks. Pre-transgressive karstification also sculpted the topography in the basin, creating a rugged surface. Large basement boulders at locality 3 are interpreted to have been displaced when the main normal faults became active. They were subsequently deposited at the foot of the next raised block. Downslope sliding or small-scale tsunamis triggered by earthquakes are known transport mechanisms in the Mediterranean (e.g., Papazachos et al. 1985). Mastronuzzi and Sans´o (2000) recorded boulders along the Ionian coast (southern Italy, in a setting somewhat similar to Rhodes), of up to 80 tonnes in weight, which had been transported at least 40 m. Mastronuzzi and Sans´o (2000) note that catastrophic events are a neglected but important agent in development of coastal morphology. Depositional framework We reconstructed the depositional framework and inferred palaeobathymetry of the Cape Vagia bay and the Kolymbia limestone by synthesizing topographical, sedimentological, and palaeoecological data. The highest basement ridge extends approximately 40 m above the most proximal part of the Kolymbia limestone (Fig. 9). At the top of the basement ridge, the greatest water depth is restricted to about 80 m. This depth is an approximate limit for oysters and

Echinoderms Echinoderms are abundant in the Kolymbia limestone, with regular echinoids especially abundant in the transported facies Kl 3 and Kl 4. Regular echinoids are particularly common today on hard substrate within the euphotic zone, where they graze on algae (Smith 1984). The presence of regular echinoids is further indicated by the presence of Gnatichnus isp., the stellate grazing trace fossil made by regular echinoids, on transported basement blocks in facies Kl 3 and Kl 4. Echinocyamus pusillus is present in large part of the Kolymbia limestone stratigraphic column. E. pusillus lives in the Mediterranean today at depths of approximately 20–50 m (Riedl 1983; Philippe 1984). Echinolampas sp. is also found in the Kolymbia limestone. The genus Echinolampas lives today only in subtropical seas. Generally they live between 10 and 500 m, with preference for 30–50 m (Philippe 1984).

Fig. 9 The depositional framework of the Kolymbia limestone. Water depth at the most proximal locations was between 10 and 80 m and at distal locations between 50 and 130 m

374

of peak abundances for the light-dependent algae creating rhodoliths. The Kolymbia limestone was deposited below wave base, inferred by us to be approximately 50 m. This is therefore the minimum depth at the proximal part of the deposit. The maximum water depth at the proximal part of the deposit is 120 m (80 m + 40 m), constrained by topographical evidence. Minimum water depth at the top of basement ridge is 10 m (50 m – 40 m). The maximum water depth in the most distal part of the basin is approximately 130 m. Currents across the sea floor are indicated both by the open geometry of the basin and banks of G. vitreus. All fossils found fit conformably into this depositional frame. Deposition of the Kolymbia limestone The deposition of the Kolymbia limestone in the Cape Vagia basin took place during a large-scale transgression, punctuated by smaller-scale tectonic movements and/or eustatic sea-level fluctuations. The overall trend of the stratigraphic section is retrogradational stacking with deepening and fining upward of sediments. Coarse event beds interrupt this pattern before disappearing towards the top of the section. The coarse beds contain shallow-water fauna. The Kolymbia limestone then grades into the deep-water clays and silts of the Lindos Bay clay. The first bed deposited at the onset of transgression was the muddy limestone of facies Kl 2. At the same time, oysters and bryozoans grew on the basement slope. Smallscale tectonic disturbance(s) are the inferred mechanism for downslope mass movement, transporting the oysters and bryozoans into deeper water and depositing bed 2. Water depth continued to increase and the bryozoan limestone of facies Kl 1 accumulated as bed 3. At the same time proximally in the basin, rhodoliths and oysters were living at or near the top of the cape ridge, while C. palmata inhabited the bottom of the slope. The water depth was approximately 50–60 m distally in the basin and 10–20 m proximally. Slumps and debris flows, again inferred to forcing by tectonic movement, carried fossils downslope into the basin, mixing shallow with deeper water fauna and depositing bed 4 of facies Kl 3. As transgression continued, bryozoan limestone (facies Kl 1) was again accumulated in the basin (bed 5). Water depth was now approximately 30–40 m proximally and 70–80 m distally. P. jacobeus formed large communities proximo-distally, with some G. vitreus distally. Echinoids were a common, but minor component. Oysters were living most proximally. Storms or small-scale tectonic event(s) again resulted in mixing faunal components in bed 6, representing facies Kl 4 shell beds. Pectenids and brachiopods were living on the sandsilt bottom initially whereas the oysters were transported. Depth was now approximately 80 m proximally and 120 m distally. Bryozoan limestone with some echinoids and other minor components accumulated again in bed 7. With rising water depth, the ecological conditions for G. vitreus became increasingly favorable and the brachiopod became abundant. A few oysters still inhabited the topographically highest sites in the basin.

Sea level was lowered up to 50 m between beds 8 and 10. It may be that the main mechanism involved was eustatic sea-level fluctuation. Tectonic movements probably also played a role in downslope movements. Shallower water sediments and fauna were mixed with the G. vitreus community at the basin bottom, forming bed 8. Water depth was now similar to that before the deposition of bed 4. Bryozoan limestone accumulated as the sea level rose, forming bed 9. Oysters and rhodoliths inhabited the most proximal basin, while C. palmata thickets again formed on the slope of the cape ridge. A small-scale tectonic movement may have again created a debris flow, transporting the faunal components to deeper water and mixing them with the bryozoan limestone, depositing bed 10 of facies Kl 3. After deposition of bed 10, transgression continued. Hereafter only bryozoan limestone (beds 11–15) was deposited until facies Kl 1 graded into the Lindos Bay clay. No beds of shallow faunal composition were deposited, but continuation of small-scale tectonic movements or storms are indicated by beds with a comparatively larger proportion of bryozoan branches (beds 11, 13, and 15). At the time of deposition of these five last beds, the water depth must have exceeded depth boundaries indicated by rhodoliths and oysters proximally in the basin, as well as celleporids, pectenids, and G. vitreus more distally in the basin. Beds 11–15 have not been studied in detail. However, it is likely that the bathymetric range of the bryozoan fauna in these beds will reflect their deeper water affinity and bridge the gap between the largest bathymetric range overlap of 30– 80 m for the underlying part of the Kolymbia limestone and the largest bathymetric range overlap of 100–600 m for the overlying Lindos Bay clay. Discussion Deposition of the celleporid beds The depositional model for the Kolymbia limestone inferred here is similar to interpretations given by Spjeldnaes and Moissette (1997) and Moissette and Spjeldnaes (1995). However, there are important differences between our models. Spjeldnaes and Moissette (1997) interpret the celleporid beds as autochthonous, living directly on the soft bottom. Their interpretation rests on two assertions; (1) the configuration of the thicket deposits rules out transport, and (2) the celleporid thickets are analogous ecologically to the “reefs” formed by the living azooxanthellate coral Lophelia. First, we do not find compelling evidence for the assertion that the configuration of branches precludes downslope transport. Slump deposits in poorly consolidated silty and clayey low-angle inclined seabeds are not uncommon phenomena in modern seas. Such deposits are perhaps most often related to seismic activity and may leave epibenthic skeletal remains largely intact in a different location. Synsedimentary block faulting, slumping, and mass sediment movements in debris flows are also well documented in tectonically active Mediterranean basins, particularly in

375

the Aegean (Perissoratis and Papadopoulos 1999; Brachert et al. 2001; Mienert et al. 2002; Lykousis et al. 2003). Outsize basement blocks up to 20 cm often accompany the relatively intact celleporid branches (colonies in excess of 25 cm, see Fig. 8), along with rhodoliths and oysters from shallow water. These blocks, rhodoliths, and large skeletal fragments, as well as intermediate skeletal grains, float together in a silt–clay matrix. The lack of sorting, huge grain size variance, and diverse provenance of basement and skeletal elements appear to be strongly indicative of transport en masse without significant winnowing, rather than in situ growth and deposition of celleporids on a more or less homogeneous soft seabed. Fossils can be transported some distance without showing major signs of damage, especially if transported in a mud slump. Large, branching colonies of some fragile bryozoans are seasonally detached and transported relatively intact by wave and current action from nearshore to subtidal offshore environments (Pederson and Peterson 2002). Second, while we also recognize some gross structural similarities between Celleporaria thickets and the rigid framework of the living deep- and cold-water coral Lophelia (Freiwald 2002), we note that these groups of organisms are phylogenetically, ecologically, and mechanically quite different. Living Lophelia reefs have recently been found in the Ionian Sea (Taviani et al. 2005), but they are growing on firmground basement substrates and at greater depths (500–700 m) than envisioned in the earlier model’s analogy. A much closer analogy on all bases of comparison exists with thickets formed by living and fossil congeneric bryozoan species (Cuffey et al. 1977; Bradstock and Gordon 1983; Gordon et al. 1994; Tillbrook 1997; Pedley and Grasso 2002; Hageman et al. 2003; Lukasik and James 2003; Taylor et al. 2004; and others). While species of Celleporaria in these studies have nodular or mound-like colony forms in addition to branching ones, there are some ecological commonalities associated with their occurrences in various shallow to outer shelf settings. Most occur below storm-wave base and at the low end of, or below, the euphotic zone. Some actualistic studies have indicated an increased occurrence of Celleporaria thickets in an inferred “mesotrophic” gradient of nutrient availability between oligotrophic and eutrophic conditions, or a tolerance for relatively low nutrient conditions. Hageman et al. (2003) report that some thicket-constructing species of Celleporaria in South Australia are tolerant of silty or muddy conditions, although occurrence as epibionts on fistulose sponges and on hard substrates are also reported. Celleporaria species from Tasman Bay, New Zealand, also occur in turbid conditions, though at a shallower depth range (10–35 m) in lower energy environments (Bradstock and Gordon 1983). Lower wave energy conditions, microtidal fluctuations, and sluggish bottom circulation are conditions that broadly characterize the Eastern Mediterranean today, and presumably began to develop following the post-Messinian reflooding of the entire basin. Spjeldnaes and Moissette (1997) suggest a depth of about 40 m for the celleporid thickets, indicated by known mod-

ern bathymetric ranges of epizoic bryozoan species. Presence of boulders of Jurassic limestone, as well as oysters and rhodoliths from shallow water, were explained by these authors as washed down into the thickets from steep basement slope by heavy storms or tsunamis after earthquakes. However, the same authors (Moissette and Spjeldnaes 1995) had, in a study on bathymetric indications provided by bryozoans, disregarded all samples from the Kolymbia limestone at Cape Vagia because of the mixing of deeper and shallow-water fauna. Their interpretation of the autochthonous deposition of the celleporid thickets also conflicts with interpreting the stratigraphically adjacent Gryphus bed, as G. vitreus is a deep-water species. If the celleporid thickets were embedded in situ, a depth decrease of approximately 100 m would have to have transpired in the time period represented by 30 cm of facies Kl 1 bryozoan limestone. The limestone shows no signs of such rapid shallowing. It would be expected that such a significant transition would provide a more apparent sedimentological signature. Rapid shallowing of sea level caused by eustatic sea-level fall and perhaps tectonism did occur between deposition of the Gryphus bed and of the uppermost celleporid bed. However, by replacing the habitat of the celleporid thickets to the basement slope in our model, the difference in depth (about 40 m) becomes more easily explained. We also believe that the intermixed rhodolith, oyster and celleporid gravel-sized grades, sand-sized bryozoan colonies and silt/mud matrix are explained more parsimoniously by a common shallow-water origin, with slump or debris flow transport en masse. The perhaps slightly lesser depths at Cape Vagia may be explained by different ecological conditions for the thicker shelled specimen of G. vitreus found at that locality, as also noted by Spjeldnaes and Moissette (1997). Large, robust branching colonies would have needed rather expansive bases to support and stabilize themselves on soft poorly consolidated silt–mud substrate, but such bases are absent. Where actualistic studies have shown large celleporids growing on muddy substrates, the colony forms have been lattice-like or globular rather than treelike. For example, Australian Celleporaria form thickets by originating their growth upon sponges, which stabilize them on the soft bottom (Hageman et al. 2003). These thickets form three-dimensional boxworks, by fusing branches and thereby constructing a self-supported structure. This is not the case here. We propose that the distribution of organisms upon which the Celleporaria colonies initiate their growth may be as important a factor as the nature of the sedimentary substrate per se. Indeed the colonies require an associated erect organism to be present before thickets can be formed, the so-called “ghost frame-builders” of Cocito et al. (2000). When no such organism is present, Celleporaria form lumps on the sea floor, which are more stable on a soft bottom, and do not form erect thickets. Spjeldnaes and Moissette (1997) interpret the cylindrical hole in the center of many celleporid branches to be the trace of an organism upon which the colony formed at the beginning of its growth, possibly a gorgonian. The celleporids are inferred by us to have initiated colony growth epizoically

376

on gorgonians or similar organisms. Epizoic growth is apparently a common, underappreciated facet of carbonate production and establishment of large individual colonies by bryozoans (Hageman et al. 2000). In addition, certain Celleporaria species are apparently selective of erect, upright substrates, and hollow cylindrical openings are conspicuous in both fossil and modern samples (Gordon et al. 1994; Hageman et al. 2003). Gorgonians would have provided the celleporids with support by anchoring on hard substrate, which, in the case of growth on soft bottom, could be provided by secondary hard bottoms such as basement blocks or oyster shells. However, such small substrates are fairly patchy and unstable on a soft, current-swept bottom, which does not seem to be in accordance with the large, widely distributed thickets. This problem can be avoided by accepting the transported nature of the celleporid thickets and inferring their provenance from the basement slope, where hard substrates would have been more common. Tectonic movements as depositional agent The coarse beds in the Kolymbia limestone, interpreted as tectonic event beds, occur at regular intervals separated by about 20–30 cm of limestone of normal facies Kl 1 composition. These intervening beds are all of more or less identical texture and composition. Such regularly spaced events, causing shallow and deeper water fauna to mix, can be attributed to periodically reoccurring tectonic movements. Even in recorded history, earthquakes are sometimes reported to occur at regular intervals, the length of which depends on the seismo-tectonic area considered (Pirazzoli et al. 1994; Perrissoratis and Papadopoulos 1999). Kontogianni et al. (2002) suggest a modified earthquake cycle model for Rhodes, reflecting small uplifts alternating with major uplifts and subsidence as a consequence of reactivation of different faults of the same fault zone. Earthquake-generated tsunamis are a well-known phenomenon in the Mediterranean region (Perrissoratis and Papadopoulos 1999; Pirazzoli et al. 1999; Massari and D’Alessandro 2000; Dominey-Howes 2002; Yalciner et al. 2002; Lykousis et al. 2003) and are reported to transport and deposit geological material, sometimes even massive boulders (Papazachos et al. 1985; Mastronuzzi and Sans´o 2000). However, deposits derived from tsunamis in the Aegean are often both thin and limited in lateral extension, probably due to coastal morphology (Dominey-Howes 2002; Papadopoulos et al. 2004). In this seismically active area Rhodes is still experiencing uplift. The vertical displacement was superimposed on a transgression and is recorded by Holocene notches, excavated by wave action at sea level (Pirazzoli et al. 1989, 1996, 1999). Some of the coarse beds may also be interpreted as storm deposits, e.g., bed 6, but the overall regularity of the thick, coarse beds seems to indicate a tectonic drive behind the events, rather than a climatological one.

Bryozoans as ecological indicators The evidence presented here on bryozoans, based both on their growth forms and on bathymetric information provided by extant species, validates the depositional frame set independently by topographic restrictions and ecological information from other organisms. The concordance of these independent suites of data contributes to a validating test of our depositional model. The consistency of two different types of data (colony growth forms, actualistic depth constraints combining a number of individual species) from the bryozoans is also significant. These data substantiate the value of bryozoans as palaeobathymetric indicators. The deeper water affinities of the Lindos Bay clay are also consistent with the largest bathymetric overlap of bryozoans. The bryozoans alone, however, would not provide as detailed information about each bed, and thereby the depositional history of the Kolymbia limestone, as does the combined information from bryozoans and other fossils at Cape Vagia. It is also relevant to note that diagenetic overprints, such as the loss of aragonitic forms noted above, can potentially reduce the precision of our palaeobathymetric estimates. For example, if patterns of bathymetric range overlaps from aragonitic bryozoans would have provided more tightly constrained palaeodepth estimates, and these forms were absent due to diagenetic dissolution, the remaining calcitic forms would represent only a subsample of the original fauna. This type of taphonomic overprint highlights the importance of utilizing multiple lines of palaeoecological inference. Considerable progress has been made by recent studies emphasizing quantitative multivariate methods, including statistical resampling, to enhance the objectivity, precision, and associated error margins of palaeoecological and palaeobathymetric interpretations of cool-water carbonates (e.g., Jimenez and Braga 1993; Brachert et al. 1998; Ferdeghini et al. 2000; Lukasik et al. 2000; Moissette 2000b; Hageman et al. 2003; Scarponi and Kowalewski 2004; Yesares-Garcia and Aguirre 2004). We concur with arguments supporting the value of these approaches, but also acknowledge that there is considerable cost in time and effort that cannot be afforded in many field studies. We suggest that all of these methodological advances still rely on one form or another of inductive reasoning. The resulting interpretations are essentially based on patterns of association or correlation, with some degree of uncertainty being unavoidable. That reasoning may be from the general to the specific, as when depth ranges of living taxa accumulated over regional or global scales are applied to specific case studies, in which case the variance associated with modern observations is factored out. The reasoning may also be from the specific to the general, as when results from a limited number of actualistic studies are assumed to be uniformly true in all cases, or when the ecological parameters of one species are extrapolated to all members of a genus or higher taxon. An implicit, but necessary, assumption of uniformitarianism may include both types of induction. This assumption may be especially relevant due to anthropogenic effects such as trawling and other

377

forms of bottom fishing, that have drastically altered patterns of occurrence of benthic species and are especially destructive to biogenic constructions of long-lived colonial animals (Watling and Norse 1998; Collie et al. 2000; Grange et al. 2003). The present study is no exception in all these regards. We hope to have shown that bryozoans do have value as palaeobathymetric and palaeoenvironmental indicators. At the same time, we believe that the confidence assigned to such palaeoecological interpretations or explanation is proportional to the degree of concordance of many independent lines of inference (Oreskes et al. 1994; Rykiel 1998; Salmon 1998).

• Ecological, bathymetric, sedimentologic, and tectonic lines of evidence provide independent validation of a depositional framework and history. Acknowledgements Richard Bromley, Rikke Bruhn, and Ulla Asgaard (University of Copenhagen, Denmark) provided useful discussions, and Patrick Wyse Jackson (Trinity College, Dublin, Ireland), the editor A. Freiwald and anonymous reviewers commented critically on the manuscript. We gratefully acknowledge their help. We also thank Hotelejer Anders M˚ansson og hustru’s mindelegat Foundation and the Danish-Icelandic Society for generous support of fieldwork.

References Conclusions • The Cape Vagia basin is an extensional microbasin, developed as a half-graben located between two major normal faults. The upper slope area was the site of a carbonate factory, producing coarse skeletal material for deposition toward the basin center, where finer material also accumulated. • The Kolymbia limestone constitutes four lithofacies. The stratigraphical column is fining upward, dominated by bryozoan limestone, interrupted at regular intervals by coarse beds, containing mixed shallow and deeper water fauna. The facies architecture of the limestone is retrogradational on intermediate slopes and draping on steep slopes. The Kolymbia limestone is bounded below by a type 1 sequence boundary and above by a marine flooding surface, placing it in the transgressive systems tract, where it comprises two parasequences. • The fauna of the coarse beds includes large brachiopods, bivalves, and the bryozoan C. palmata, forming thickets. The bryozoan colonies apparently initiated their growth epizoically, probably on gorgonians, which anchored them to the basement slope, where they grew at palaeodepths of ∼ 30–50 m. Detachment and downslope transport occurred during deposition. • The growth-form distribution of the 22 classified bryozoan species indicates a moderately deep-water habitat, in low to moderate currents and slow sedimentation on solid substrates. The subset of these species that are extant have their greatest bathymetric overlap between 30 and 80 m. • A depositional frame was constructed using the topography of the basin as well as ecological data derived from the fossils at Cape Vagia. The water depth at the time of deposition of the Kolymbia limestone was 10–80 m at the most proximal sites and 50–130 m close to the basin center. All the studied fossils and living representatives fit conformably within the depositional frame. • The deposition of the Kolymbia limestone took place during a major transgression, punctuated by smaller tectonic events, triggering downslope mass movements. • Bryozoans proved to be significant ecological indicators when interpreting the overall depositional history, but did not yield information as detailed as that provided by some other fossils.

Aguirre J, Braga JC, Jim´enez AP, Rivas P (1996) Substraterelated changes in pectinid fossil assemblages. Palaeogeogr Palaeoclimatol Palaeoecol 126:291–308 Batson P (2000) The Otago Shelf bryozoan thickets: aspects of their distribution, ecology and sedimentology. Marine Science Abstracts, University of Otago, New Zealand Benda L, Meulenkamp JE, van de Weerd A (1977) Biostratigraphic correlations in the Eastern Mediterranean Neogene, 3: correlation between mammal, sporomorph and marine microfossils assamblages from the Upper Cenozoic of Rhodes, Greece. Newslett Stratigr 6:117–130 Betzler C, Brachert TC, Braga J-C, Martin JM (1997) Nearshore, temperate, carbonate depositional systems (lower Tortonian, Agua Amarga Basin, southern Spain): implications for carbonate sequence stratigraphy. Sediment Geol 113:27–53 Betzler C, Martin JM, Braga JC (2000) Non-tropical carbonates related to rocky submarine cliffs (Miocene, Almeria, southern Spain). Sediment Geol 131:51–65 Brachert TC, Betzler C, Braga JC, Martin JM (1998) Microtaphofacies of a warm-temperate carbonate ramp (uppermost Tortonian lowermost Messinian, southern Spain). Palaios 13:459– 475 Brachert TC, Hultzsch N, Knoerich AC, Krautworst UMR, Stuckrad OM (2001) Climatic signatures in shallow-water carbonates: high-resolution stratigraphic markers in structurally controlled carbonate buildups (Late Miocene, southern Spain). Palaeogeogr Palaeoclimatol Palaeoecol 175:211–237 Bradstock M, Gordon DP (1983) Coral-like bryozoan growths in Tasman Bay, and their protection to conserve commercial fish stocks. N Z J Mar Freshw Res 17:159–163 Bromley RG, Asgaard U (1993) Endolithic community replacement on a Pliocene rocky coast. Ichnos 2:93–116 Bromley RG, Hanken N-M (2003) Structure and function of large, lobed Zoophycos, Pliocene of Rhodes, Greece. Palaeogeogr Palaeoclimatol Palaeoecol 192:97–100 Butler RWH, Grasso M, Gardiner W, Sedgeley D (1997) Depositional patterns and their tectonic controls within the Plio-Quaternary carbonate sands and muds of onshore and offshore SE Sicily (Italy). Mar Petrol Geol 14:879–892 Cheetham A (1967) Paleoclimatic significance of the bryozoan Metrorobdotos. Transact Gulf Coast Assoc Geol Soc 17:400– 407 Cocito S (2004) Bioconstruction and biodiversity: their mutual influence. Sci Mar 68(Suppl 1):137–144 Cocito S, Ferdeghini F (2001) Carbonate standing stock and carbonate production of the bryozoan Pentapora fascialis in the north-western Mediterranean. Facies 45:25–30 Cocito S, Ferdeghini F, Morri C, Bianchi CN (2000) Patterns of bioconstruction in the cheilostome bryozoan Schizoporella errata: the influence of hydrodynamics and associated biota. Mar Ecol Progr Ser 192:153–161 Collie JS, Escanero GA, Valentine PC (2000) Photographic evaluation of the impacts of bottom fishing on benthic epifauna. ICES J Mar Sci 57:987–1001

378 Corda L, Brandano M (2003) Aphotic zone carbonate production on a Miocene ramp, Central Apennines, Italy. Sediment Geol 161:55–70 Cranfield HJ, Manighetti B, Michael KP, Hill A (2003) Effects of oyster dredging on the distribution of bryozoan biogenic reefs and associated sediments in Foveaux Strait, southern New Zealand. Cont Shelf Res 23:1337–1357 Cuffey RJ (1985) Expanded reef-rock textural classification and the geologic history of bryozoan reefs. Geology 13:307– 310 Cuffey RJ, Fonda SS, Kosich DF, Gebelein CD, Bliefnick DM, Soroka LG (1977) Modern tidal-channel bryozoan reefs of Joulters Cays (Bahamas). In: Taylor DL (ed) Proceedings of the 3rd International Coral Reefs Symposium, vol. 2. Rosenstiel School Marine and Atmospheric Science, Miami, pp 340– 345 Dewey JF, Hempton MR, Kidd WSF, Saroglu F, Sengor AMC (1986) Shortening of continental lithosphere: the neotectonics of eastern Anatolia, a young continental collision zone. Geol Soc Spec Publ 19:3–36 Dominey-Howes D (2002) Documentary and geological records of tsunamis in the Aegean Sea region of Greece and their potential value to risk assessment and disaster management. Nat Hazards 25:195–224 Emig CC (1989) Distributional patterns along the Mediterranean continental margin (upper bathyal) using Gryphus vitreus (Brachiopoda) densities. Palaegeogr Palaeoclimatol Palaeoecol 71:253–256 Emig CC, Garc´a-Carrascosa MA (1991) Distribution of Gryphus vitreus (Born 1778) (Brachiopoda) on transect P2 (Continental margin, French Mediterranean coast) investigated by submersible. Sci Mar 55:385–388 Ferdeghini F, Acunto S, Cocito S, Cinelli F (2000) Variability at different spatial scales of a coralligenous assemblage at Giannutri Island (Tuscan Archipelago, northwest Mediterranean). Hydrobiologia 440:27–36 Freiwald A (2002) Reef-forming cold-water corals. In: Wefer G, Billett D, Hebbeln D, Jørgensen BB, Schl¨uter M, van Weering T (eds) Ocean margin systems. Springer, Berlin Heidelberg New York, pp 365–385 Gordon DP, Stuart IG, Collen JD (1994) Bryozoan fauna of the Kaipuke Siltstone, northwestern Nelson: Miocene homologue of the modern Tasman Bay coralline bryozoan grounds. N Z J Geol Geophys 37:239–247 Grange KR, Tovey A, Hill AF (2003) The spatial extent and nature of the bryozoan communities at Separation Point, Tasman Bay. Marine Biodiversity Biosecurity Report No. 4. Ministry of Fisheries, Wellington, 22 p Hageman SJ, Bone Y, McGowran B (1997) Bryozoan colonial growth-forms as paleoenvironmental indicators: evaluation of morphology. Palaios 12:405–419 Hageman SJ, Bock PE, Bone Y, McGowran B (1998) Bryozoan growth habits: classification and analysis. J Paleontol 72:418–436 Hageman SJ, James N, Bone Y (2000) Cool-water carbonate production from epizoic bryozoans on ephemeral substrates. Palaios 15:33–48 Hageman SJ, Lukasik J, McGowran B, Bone Y (2003) Paleoenvironmental significance of Celleporaria (Bryozoan) from modern and Tertiary cool-water carbonates of Southern Australia. Palaios 18:510–527 Hallquist M, Hansen T (1997) Kystnære Plio-Pleistocene aflejringer p˚a nordvestkysten af Rhodos: En sedimentologisk, taphonimisk og palæoøkologisk analyse. Unpublished Masters thesis, Geological Institute, University of Copenhagen Hanken N-M, Bromley RG, Miller J (1996) Plio-Pleistocene sedimentation in coastal grabens, north-east Rhodes, Greece. Geol J 31:271–298 Hansen KS (2001) Sedimentology of Pliocene–Pleistocene temperate water carbonates from Northeast Rhodes, Greece. Unpublished Ph.D. thesis, Geological Institute, University of Copenhagen, 311 pp

Hayton S, Nelson CS, Hood SD (1995) A skeletal assemblage classification system for non-tropical carbonate deposits based on New Zealand Cenozoic limestones. Sediment Geol 100:123–141 Hayward PJ, Ryland JS (1979) British ascophoran bryozoans. Academic, London, 312 pp Hayward PJ, Ryland JS (1985) Cyclostome bryozoans. The Bath Press, Avon, 147 pp Haywood AM, Sellwood BW, Valdes PJ (2000) Regional warming: Pliocene (3 Ma) paleoclimate of Europe and the Mediterranean. Geology 28:1063–1066 Henrich R, Freiwald A, Betzler C, Bader B, Sch¨afer P, Samtleben C, Brachert TC, Wehrmann A, Zankl H, K¨uhlmann DHH (1995) Controls on modern carbonate sedimentation on warm-temperate to Arctic coasts, shelves and seamounts in the Northern Hemisphere: implications for fossil counterparts. Facies 32:71–108 James NP (1997) The cool-water carbonate depositional realm. SEPM Spec Publ 56:1–20 James NP, Feary DA, Betzler C, Bone Y, Holbourn AE, Li QY, Machiyama H, Simo JAT, Surlyk F (2004) Origin of late Pleistocene bryozoan reef mounds; Great Australian Bight. J Sediment Res 74:20–48 Jimenez AP, Braga JC (1993) Occurrence and taphonomy of bivalves from the Nijar reef (Messinian, Late Pliocene, SE Spain). Palaeogeogr Palaeoclimatol Palaeoecol 102:239–251 Kaselowsky J, Scholz J, Mawatari SF, Probert PK, Gerdes G, Kadagies N, Hillmer G (2005) Bryozoans and microbial communities of cool-temparate to subtropical latitudes—paleoecological implications. I. Growth morphologies of shallow-water bryozoans settling on bivalve shells (Japan and New Zealand). Facies 50:347–361 Kontogianni VA, Tsoulos N, Stiros SC (2002) Coastal uplift, earthquakes and active faulting of Rhodes Island (Aegean Arc): modeling based on geodetic inversion. Mar Geol 186:299– 317 Kovacs E, Spjeldnaes N (1999) Pliocene–Pleistocene stratigraphy of Rhodes, Greece. Newslett Stratigr 37:191–208 Laborel J, Laborel-Deguen F (1996) Biological indicators of Holocene sea-level and climatic variations on rocky coasts of tropical and subtropical regions. Quat Int 31:53–60 Lagaaij R, Gautier YV (1965) Bryozoan assemblages from marine sediments of the Rhone delta, France. Micropaleontology 11:39–58 Laurain M (1984) Bivalves Ostr´eid´es. In: Pouyet S (ed) La faune du faci`es “Marnes Bleues”, Burdigalien du bassin de FauconMollans-Malauc`ene (sud-est de la France). Nouv Arch Mus d’Hist Nat Lyon 22:74–78 Lees A, Buller AT (1972) Modern temperate-water and warm-water shelf carbonate sediments contrasted. Mar Geol 13:M67–M73 Logan A (1979) The recent Brachiopoda of the Mediterranean Sea. Bull Inst Oc´eanogr Monaco 72/1434:1–112 Logan A, Bianchi CN, Morri C, Zibrowius H (2004) The present-day Mediterranean brachiopod fauna: diversity, life habits, biogeography and paleobiogeography. Scientia Marina 68 (Suppl. 1):163–170 Lukasik JJ, James NP (2003) Deepening-upward subtidal cycles, Murray Basin, South Australia. J Sediment Res 73:653–671 Lukasik JJ, James NP, McGowran B, Bone Y (2000) An epeiric ramp: low-energy, cool-water carbonate facies in a Tertiary inland sea, Murray Basin, South Australia. Sedimentology 47:851–881 Lykousis V, Sakellariou D, Roussakis G (2003) Prodelta slope stability and associated coastal hazards in tectonically active margins: Gulf of Corinth (NE Mediterranean). In: Locat J, Mienert J (eds) Submarine mass movements and their consequences. Kluwer, Dordrecht, pp 433–440 Løvlie R, Støle G, Spjeldnæs N (1989) Magnetic polarity stratigraphy of Pliocene–Pleistocene marine sediments from Rhodos, Eastern Mediterranean. Phys Earth Planet Inter 54:340–352 Massari F, D’Alessandro A (2000) Tsunami-related scour-and-drape undulations in Middle Pliocene restricted-bay carbonate deposits (Salento, south Italy). Sediment Geol 135:265–281

379 Mastronuzzi G, Sans´o P (2000) Boulders transport by catastrophic waves along the Ionian coast of Apulia (southern Italy). Mar Geol 170:93–103 McKenzie R (1978) Active tectonics of the Alpine–Himalayan belt: the Aegean Sea and surrounding regions. Geophys J R Astr Soc 55:217–254 McKinney FK, Jackson JBC (1991) Bryozoan evolution, 2nd edn. University of Chicago Press, Chicago, 252 pp Meulenkamp JE, de Mulder EFJ, van de Weerd A (1972) Sedimentary history and paleogeography of the Late Cenozoic of the island of Rhodes. Zeitschr Deutsch Geol Ges 123:541–553 Mienert J, Berndt C, Laberg JS, Vorren TO (2002) Slope instability of continental margins. In: Wefer G, Billett D, Hebbeln D, Jørgensen BB, Schl¨uter M, van Werring T (eds) Ocean margin systems. Springer, Berlin Heidelberg New York, pp 179– 193 Moissette P (1988) Faunes de Bryozoaires du Messinien d’Algerie occidentale. Doc Laborat G´eol Fac Sci Lyon 102:1–351 Moissette P (2000a) The use of Neogene bryozoans for a better understanding of the ecology of some Recent species. In: Herrere Cubila A, Jackson JBC (eds) Proceedings of the 11th International Bryozoology Association Conference, pp 191– 197 Moissette P (2000b) Changes in bryozoan assemblages and bathymetric variations. Examples from the Messinian of northwest Algeria. Palaeogeogr Palaeoclimatol Palaeoecol 155:305– 326 Moissette P, Spjeldnaes N (1995) Plio-Pleistocene deep-water bryozoans from Rhodes, Greece. Palaeontology 38:771–799 Moissette P, Spjeldnaes N, Georgiades-Dikeoulia E (2002) Highly diverse bryozoan faunas from the Plio-Pleistocene of the Greek island of Rhodes. In: Wyse Jackson PN, Buttler CJ, Spencer Jones ME (eds) Bryozoan studies 2001. A A Balkema, Lisse, pp 215–220 Mutti E, Orombelli G, Pozzi R (1970) Geological studies on Dodocanese islands (Aegean Sea). IX Geological map of the island of Rhodes (Greece). Explanatory notes. Ann Geol Pays Hellen 22:77–226 Mutti M, Hallock P (2003) Carbonate systems along nutrient and temperature gradients: some sedimentological and geochemical constraints. Int J Earth Sci 92:465–475 Nelson CS, Hyden FM, Keane SL, Leask WL, Gordon DP (1988) Application of bryozoan zoarial growth-form studies in facies analysis of non-tropical carbonate deposits in New Zealand. Sediment Geol 60:301–322 Nebout NC, Foucault A, Melieres F (2004) Vegetation markers of palaeoclimate cyclical changes in the Pliocene of Punta Piccola (Sicily, Italy). Palaeogeogr Palaeoclimatol Palaeoecol 214:55–66 Oral MB, Reilinger RE, Toks¨oz MN, King RW, Barka AA, Kinik I, Lenk O (1995) Global positioning system offers evidence of plate motions in Eastern Mediterranean. Eos 76:9–11 Oreskes N, Schrader-Frechetten K, Belitz K (1994) Verification, validation, and confirmation of numerical models in the Earth sciences. Science 263:641–646 Papadopoulos GA, Imamura F, Minoura K, Takahasi T, Kuran U, Yalciner AC (2004) Strong earthquakes and tsunamis in the East Hellenic Arc. Geophys Res Abstr 6:03212 Papazachos BG, Koutitas C, Hatzidimitrou PM, Karacostas BG, Papaioannou CA (1985) Source and short-distance propagation of the July 9, 1956 southern Aegean tsunami. Mar Geol 65:343–351 Pederson EJ, Peterson MS (2002) Bryozoans as ephemeral estuarine habitat and a larval transport mechanism for mobile benthos and young fishes in the northcentral Gulf of Mexico. Mar Biol 140:935–947 Pedley M, Grasso M (2002) Lithofacies modelling and sequence stratigraphy in microtidal cool-water carbonates: a case study from the Pleistocene of Sicily, Italy. Sedimentology 49:533–553 Perissoratis C, Papadopoulos G (1999) Sediment instability and slumping in the southern Aegean Sea and the case history of the 1956 tsunami. Mar Geol 161:287–305

Philippe M (1984) Echinides. In: Pouyet S (ed) La faune du faci`es “Marnes Bleues”, Burdigalien du bassin de Faucon-MollansMalauc`ene (sud-est de la France). Nouv Arch Mus Hist Nat Lyon 22:85–91 Pirazzoli PA, Laborel J, Stiros SC (1996) Earthquake clustering in the Eastern Mediterranean during historical times. J Geophys Res 101(B3):6083–6097 Pirazzoli PA, Montaggioni LF, Thommeret J, Thommeret Y, Laborel J (1983) Sur les lignes de rivage et la neotectonique a Rhodes (Grece) a l’Holocene. Ann Inst Oceanogr 58:89–102 Pirazzoli PA, Montaggioni LF, Sali`ege JF, Segonzac G, Thommeret Y, Vergnaud-Grazzini C (1989) Crustal block movements from Holocene shorelines: Rhodes Island (Greece). Tectonophysics 170:89–114 Pirazzoli PA, Stiros SC, Laborel J, Laborel-Deguen F, Arnold M, Papageorgiou S, Morhange C (1994) Late-Holocene shoreline changes related to palaeoseismic events in the Ionian Islands, Greece. The Holocene 4:397–405 Pirazzoli PA, Stiros SC, Arnold M, Laborel J, Laborel-Deguen F (1999) Late Holocene coseismic vertical displacements and tsunami deposits near Kynos, Gulf of Euboea, central Greece. Phys Chem Earth (A) 24:361–367 Pomar L (2001) Types of carbonate platforms: a genetic approach. Basin Res 13:313–334 Pomar L, Brandano M, Westphal H (2004) Environmental factors influencing skeletal grain sediment associations: a critical review of Miocene examples from the western Mediterranean. Sedimentology 51:627–651 Pomoni-Papaioannou F, Drinia H, Dermitzakis MD (2003) Neogene non-tropical carbonate sedimentation in a warm-temperate biogeographic province (Rethymnon Formation, Eastern Crete, Greece. Sediment Geol 154:147–157 Pouyet S (1973) R´evision syst´ematique de cell´epores (Bryozoa, Cheilostomata) et des esp`eces fossils Europ´eennes. Analyse du quelques populations a` cell´epores dans le N´eog`ene du Bassin Rhodanian. Doc Laborat G´eol Fac Sci Lyon 55:1–266 Pouyet S, Moissette P (1992) Bryozoaires du Pliocene d’Altavilla (Sicile-Italie): Revision de la collection Cipolla, novelles donnes, paleoecologie. Palaeontograph Abt A 223:19–201 Riedl R (1983) Fauna und Flora des Mittelmeeres, 3rd edn. Verlag Paul Parey, Hamburg, Berlin, 835 pp Rykiel EJ Jr (1998) Testing ecological models: the meaning of validation. Ecolog Model 90:229–244 Ryland JS, Hayward PJ (1977) British anascan bryozoans. Academic, London, 188 pp Salmon WC (1998) Causality and explanation. Oxford University Press, Oxford, 434 pp Sartoretto S, Verlaque M, Laborel J (1996) Age of settlement and accumulation rate of submarine “corallig`ene” ( − 10 to − 60 m) of the Northwestern Mediterranean Sea; relation to Holocene rise in sea level. Mar Geol 130:317–331 Scarponi D, Kowalewski M (2004) Stratigraphic paleoecology: bathymetric signatures and sequence overprint of mollusk associations from upper Quaternary sequences of the Po Plain, Italy. Geology 32:989–992 Scholz J, Hillmer G (1995) Reef-bryozoans and bryozoanmicroreefs: control factor evidence from the Philippines and other regions. Facies 32:109–144 Schopf TJM (1969) Paleoecology of ectoprocts (Bryozoans). J Paleontol 43:234–244 Smith AB (1984) Echinoid palaeobiology. George Allen and Unwin, London, 191 pp Smith AM (1995) Palaeoenvironmental interpretations using bryozoans: a review. Geol Soc Spec Publ 83:231–243 Spjeldnaes N, Moissette P (1997) Celleporid (bryozoan) thickets from the Upper Pliocene of the island of Rhodes, Greece. SEPM Spec Publ 56:263–270 Stach LW (1936) Correlation of zoarial forms with habitat. J Geol 44:60–65 Steinth´orsd´ottir M (2001) Bryozoan growth forms and their application as ecological indicators. Unpublished examination assignment for a one term palaeontology course (B16)

380 Taviani M, Remia A, Corselli C, Freiwald A, Malinverno E, Mastrototaro F, Savini A, Tursi A (2005) First geo-marine survey of living cold-water Lophelia reefs in the Ionian Sea (Mediterranean basin). Facies 50:409–417 Taylor PD, Gordon DP, Batson PB (2004) Bathymetric distributions of modern populations of some common Cenozoic Bryozoa from New Zealand, and paleodepth estimation. N Z J Geol Geophys 47:57–69 Tillbrook KJ (1997) Barnacle and bivalve associates of a bryozoancoral symbiosis from the Coralline Crag (Pliocene) of England. Tertiary Res 18:7–22 Van Wagoner JC, Posamentier HW, Mitchum RM, Vail PR, Sarg JF, Loutit TS, Hardenbol J (1988) An overview of the fundamentals of sequence stratigraphy and key definitions. SEPM Spec Publ 42:39–45

Watling L, Norse EA (1998) Disturbance of the seabed by mobile fishing gear: a comparison to forest clearcutting. Conserv Biol 12:1180–1197 Wright VP, Burchette TP (eds) (1998) Carbonate ramps. Geol Soc Spec Publ, London, 149 pp Yalciner AC, Alpar B, Altinok Y, Ozbay I, Imamura F (2002) Tsunamis in the Sea of Marmara—Historical documents for the past, models for the future. Mar Geol 190:445–463 Yesares-Garcia J, Aguirre J (2004) Quantitative taphonomic analysis and taphofacies in lower Pliocene temperate carbonatesiliciclastic mixed platform deposits (Almeria-Nijar basin, SE Spain). Palaeogeogr Palaeoclimatol Palaeoecol 207:83–103 Zabala M, Maluquer P (1988) Illustrated keys for the classification of Mediterranean Bryozoa. Treballs del Museu de Zoologia Barcelona 4, 294 pp