The Lower Visean Skelevatka Suite (Zones C₁ ^vb-d) exposed along Mokra Volnovakha River (Fig. 1B) and in the central fluxing limestone quarry, is divided into two parts. The lower part is composed of bioclastic, foraminiferal and algal packstone and grainstone. The upper part displays similar facies but is also contains black chert nodules and lenses (Fig. 2). Brachiopods and calcareous green algae are the most common elements, together with the tabulate and rugose corals. The solitary rugose corals Haplolasma subibicina and Amygdalophyllum brazhnikovae are common (Vassiljuk, 1960), as well as colonies of Siphonodendron affine tanaicum (possibly S. ondulosum), “Lonsdaleia” longisepta, Eolithostrotionella zhizhinae, Eolithostrotionella mariupoliensis (possibly Ceriodotia), Siphonodendron sp. (= Diphyphyllum lateseptatum in Vassiljuk, 1960), Dorlodotia fomitschevi, D. briarti and numerous colonies of the tabulate Syringopora geniculata. The systematic position of several corals (Eolithostrotionella) is still unresolved but some have been discussed by Ogar (2010), Denayer (2011) and Hecker (2012).

The Upper Devonian to Upper Visean Yılanlı Formation crops out in the Bartın and Zonguldak areas in northwestern Turkey (Fig. 1). The Lower Visean part is dominated by bioclastic grainstone to rudstone, commonly silicified in the Kokaksu section (suburbs of Zonguldak, Fig. 1C), from where the Turkish material originates (Fig. 2). The macrofossils are abundant in this stratigraphic unit: brachiopods (mainly productids), bellerophontids, syringoporid tabulate corals and numerous colonial rugose corals (Dorlodotia euxinensis, Ceriodotia bartinensis, C. petalaxoides, Siphonodendron ondulosum, S. martini) and the solitary rugosans Palaeosmilia murchisoni and Clisiophyllum multiseptatum. To date, no Amygdalophyllum is known form the Lower Visean of NW Turkey. This assemblage is typical of the upper part of the Lower Visean of northwestern Turkey and defined the rugose coral biozone RC5B of Denayer (2014), an equivalent of RC5β-γ of Poty et al. (2006). The foraminiferal content, notably numerous primitive Pojarkovella, indicates the biozone MFZ11B of Hance et al. (2011).

The type specimen, labelled 1405/11 is deposited in the National Museum of Natural History at the National Academy of Science of Ukraine in Kyiv; other Ukrainian material is housed in the University of Kiev; the Turkish specimen is housed in the collections of animal and human palaeontology of the University of Liege (Belgium). Nomenclature follows Hill (1981) unless specifically noted.

Family Amygdalophyllidae Grabau in Chi, 1935.

Vassiljukia gen. nov.

Etymology . This genus is dedicated to Prof. Dr. N. P. Vassiljuk who described the type species.

Type species . Lithostrotion columnariformis Vassiljuk, 1960 from the Lower Visean of the Donets Basin.

Diagnosis . Cerioid or sub-cerioid coral showing a variably thickened amygdalophylloid axial structure. Long major septa reaching the axis of the coral. Minor septa long, entering into the tabularium. Minor septa usually less thickened than the major septa. Dissepimentarium composed of concentric interseptal dissepiments and occasional second order lonsdaleoid dissepiments. Tabulae incomplete, conical, commonly thickened near the axial structure. Increase lateral and non-parricidal.

Discussion . The corallites show the characters of the solitary genus Amygdalophyllum Dun and Benson, 1920. The axial structure is a typically amygdalophylloid thickened spindle-shape structure, the major septa are thickened in the tabularium in the same way as in various species of Amygdalophyllum, and the dissepimentarium is composed of thickened concentric interseptal dissepiments. However, the colonial habitus is unusual in Amygdalophyllum. Amygdalophyllum Dun and Benson, 1920 is defined as a solitary coral (See Poty, 2007 for a recent emendation of the diagnosis) but several species can develop a limited increase with few offsets. The type specimen of A. dianthoides (McCoy, 1851) from the Lower Visean of England shows several offsets developed in the outer part of the calyx in a peripheral parricidal increase. Material recently collected in the type area also shows occasionally a limited increase (Boland, 2002). In these English corals, the offsets form a sub-cerioid proto-colony but no true colony is known in A. dianthoides. Similarly, A. minimum Webb, 1990 from the Upper Tournaisian–Lower Visean of eastern Australia develops non-parricidal lateral increase. The latter possibly gave rise to the genus Cionodendron Benson and Smith (1923) and the other Australian lithostrotionids (Denayer and Webb, 2015). However, A. minimum shows fasciculate proto-colonies rather than cerioid ones. Massive colonies of Australian lithostrotionids (Australastraea) are typically characterized by thamnasterioid, asteroid and aphroid habitus. Their internal characters are however very similar to those of Amygdalophyllum. Vassiljukia differs from these Australastraea by a true cerioid habitus (complete and continuous wall between neighbouring corallites), a less robust but more complex columella with distinct septal lamellae and higher diameter and number of septa. Rodríguez and Hernando (2005) described from the Late Visean of South Spain the genus Espielia which is either a colonial Amygdalophyllum or, as suggested by these authors, Rylstonia. Guadiatia Gómez-Herguedas and Rodríguez (2005) is another amygdalophyllid coral developing a colonial habitus by peripheral increase. Espielia, Cionodendron and Guadiatia differ from Vassiljukia by their fasciculate habitus. The attribution of the proto-colonies and limited increase of A. dianthoides to a distinct genus is debatable but we preferred to keep the species in Amygdalophyllum since individuals without offsets are clearly Amygdalophyllum. On the other hand, the Ukrainian and Turkish specimens are worth classifying in a new genus since their cerioid habitus is stable and means a further step in the corallite integration into the colony. From slightly older strata (C₁ ^v b), Vassiljuk (1960) figured another cerioid colony under the name Lonsdaleia longiseptata Lissitzin, 1925. This species does not belong to Lonsdaleia because the axial structure is not axophyllid but rather amygdalophylloid, with a thickened axial plate and gently upturned axial tabellae. It is very similar to Vassiljukia columnariformis but differs from the latter by its conspicuous lonsdaleoid dissepimentarium and possibly by a parricidal increase (see two juvenile corallites in the figured specimen). Unfortunately, there is only one colony available and without additional material it is not possible to decide whether “L.” longiseptata belongs to Vassiljukia (V. columnariformis or a distinct species) or to another distinct genus.

Vassiljukia columnariformis (Vassiljuk, 1960)

Fig. 3

* 1960 Lithostrotion columnariformis Vassiljuk, p. 90, pl. 22, figs. 4-4a.

Holotype . Specimen 1405/11, from the Lower Visean (C₁ ^vc) of the Mokra Volnovakha River valley, Donets Basin.

Additional material . One colony from the type locality (Ukraine) and one colony is known from Kokaksu (Zonguldak, NW Turkey).

Description.

Habitus and external character . The colonies are sub-cerioid to cerioid and count less than 10 polygonal corallites. The colonies are up to 5 cm in diameter and less than 3 cm high.

Internal characters in transverse section . The corallites are polygonal to sub-circular in transverse section (Fig. 3: 1a, 3a). Their width is 13.3 mm on average (maximum 22 mm) and their tabularium is 8.5 mm-wide (maximum: 10 mm). There are, in average 31 septa of each order (maximum 35). The major septa are long and reach the axial structure. They are spindle-shaped and slightly wavy in the dissepimentarium, straighter in the tabularium, except their axial ends that curve near the axial structure. Their width is greatest (0.8 mm) in the outer part of the tabularium. Their axial ends are sharp or rounded. The minor septa enter into the tabularium on 1 mm, their length varies between one-half and two-thirds of the major (Fig. 3: 3c). The cardinal and counter septa are connected to the axial plate but the cardinal one can also be shorter than the other septa. The axial structure is symmetrical and composed of 2.5-3 mm-long axial plate spindle-shaped or elongated, bearing about 20 septal lamellae, usually short and thick. The latter are confluent with the axial ends of the major septa (Fig. 3: 1b, 3b) or are intercepted by upturned axial tabellae (Fig. 3: 1c). The dissepimentarium is composed of 5 to 11 rows of concentric or rarely herringbone interseptal dissepiments. The inner row is thickened in continuity with the septa (Fig. 3: 1a) and closely spaced to the next row. Lonsdaleoid dissepiments are occasional in the outer zone (Fig. 3: 1a). The external wall is 0.4–0.5 mm-thick, wavy or festooned.

Internal characters in longitudinal section . The axial structure is continuous vertically and its axial plate is slightly sinuous (Fig. 3: 1b). The tabularium is slightly divided in Vassiljuk's holotype. The axial tabellae are cone- to tent-shape, upturned towards the axial structure. The periaxial tabellae are vesicular and sub-parallel to the axial ones. The dissepiments are vesicular and relatively flat but show an abrupt change in the orientation from gently inclined in the outer part of the dissepimentarium to steeply inclined near the tabularium (Fig. 3: 1b). There are c. 22 tabellae and 18 dissepiments per vertical centimetre.

Discussion . As stated earlier, the species shows corallites very similar to individual specimens of the solitary genus Amygdalophyllum. A. dianthoides (McCoy, 1851) which sometimes develops a limited increase shows larger corallites and has more septa (up to 40 septa for c. 20 mm in diameter) than Vassiljukia columnariformis. The proto-colonial A. minimum Webb (1990) differs from the present species by smaller corallites (less than 8 mm) and a fasciculate habitus. The largest species of massive Australian lithostrotionid, Australastraea parvicolumnare (Pickett, 1966) has similar dimension (15 mm wide) but has less septa (24-27 versus 31-35 for V. columnariformis). Moreover, the Australian species shows a discontinuous wall that produces a thamnastreoid habitus rather than a cerioid one.

Distribution . Vassiljuk's (1960) specimens come from the C₁ ^v c-d levels of the Donets Basin which corresponds to the Early Visean (Fig. 2). The Turkish specimen was collected at Kokaksu, in a horizon attributed to the Uppermost Moliniacian (Lower Visean) RC5B coral sub-zone (Denayer, 2014), together with Dorlodotia briarti, D. euxinensis, Ceriodotia bartinensis and Lithostrotion potii). Associated foraminifers indicate the MFZ11B zone of Hance et al. (2011).

As discussed above, Vassiljukia columnariformis differs from Amygdalophyllum dianthoides by a stable colonial habitus. But what is a stable colonial habitus? How can we distinguish a small colony from a solitary coral with a limited increase? Where does the coloniality begin? It is commonly admitted that all the colonial genera originated from solitary corals that developed, for any reason, an increase at different moment of their evolution. Examples are numerous (see Rodríguez and Somerville, 2010): Corwenia evolved from Dibunophyllum, Lonsdaleia evolved from Axophyllum, Palastraea evolved from Palaeosmilia, the latter producing a massive habitus while the others are fasciculate.

In theory, the first specimens of a colonial genus are those who produce one or several buds which have the ability to produce themselves other clonal buds and thus initiating a third generation of individuals. This ability is supposed to be genetically controlled (Fedorowski, 1978 and Fedorowski and Ogar, 2013). Nevertheless, before reaching this true colonial stage, the solitary coral can develop proto-colonies in which the second generation of individuals is unable to produce any buds (limited increase, see example in Table 1). These offspring-less individuals are the “lost structures” of Fedorowski (1978) and form, together with the initial corallite, a protocolony in the sense of Fedorowski and Ogar (2013). After these authors, the lost structures are characterized by a small diameter and morphological features different from those observed in the mature parent. Several proto-colonies have been documented in the fossil record and described either under a new generic name, or as a special case of the solitary taxa. Palaeosmilia murchisoni showing few offsets occurs in the Lower Visean of Belgium (Poty, 2010) whereas the true Palastraea (i.e. with a stable colonial habitus) appeared only in the Upper Visean. Obviously, the Lower Visean specimens are protocolonies and consequently, being still Palaeosmilia, they should not be classified within the genus Palastraea. Similarly, Amygdalophyllum minimum from the Tournaisian of eastern Australia (Webb, 1990) shows occasional offsets but the true coloniality is not acquired at this stage. Following this rule, the Lower Visean Amygdalophyllum dianthoides, producing only limited offsets, is a proto-colony and differs from Vassiljukia. Moreover, in the Donets Basin, the proto-colonial forms commonly co-occur with true solitary specimens of the same species in the same beds. They thus share a stratigraphic and geographic distribution and exemplify a sympatric evolution. On the other hand, the Givetian genus Heliophyllum of North America includes solitary, proto-colonial and true colonial forms, the latter displaying both fasciculate and massive habitus (e.g., various sub-species of H. halli, Oliver and Sorauf, 2002). This variability in the growth form is possibly the phenotypic expression of a genetic diversity as both habitus are found in the same beds.

The Lower Visean of the Donets Basin yielded many species of amygdalophyllid corals variously named and figured by Vassiljuk (1960). The re-investigation of her collection led to the identification of a new colonial genus. Vassiljukia columnariformis, originally described as a Lithostrotion but displays the diagnostic characters of the Amygdalophyllidae. Its cerioid habitus is interpreted as a genuine colonial facies and not just a “lost structure”. Interestingly, in NW Turkey, Vassiljukia columnariformis first appears in the same bed as Ceriodotia bartinensis (a cerioid Dorlodotia) and Lithostrotion potii (the oldest of the genus, cerioid form derived from Siphonodendron). The Lower Visean of the Donets Basin also yielded “Lonsdaleia” longiseptata, another cerioid massive coral similar and questionably synonym to V. columnariformis. Similarly, in South China, the first Visean cerioid corals (“Lithostrotionella”) derived from the fasciculate genus Heterostrotion (Poty et al., 2011) also appear at approximately the same time. The Lower Visean was hence a period of “ceriodization” of corals but no explanation is proposed so far for this observation.