Cornulites, Schlotheim, 1820

Herringshaw, Liam G., Thomas, Alan T. & Smith, M. Paul, 2007, Systematics, shell structure and affinities of the Palaeozoic Problematicum Cornulites, Zoological Journal of the Linnean Society 150 (4), pp. 681-699 : 697-698

CORNULITES – SHELLED PHORONID, SOLITARY

BRYOZOAN OR ASEPTATE CNIDARIAN?

Vinn (2005) hypothesized that phoronids were perhaps the closest living relatives of Cornulites and the tentaculitids, but the supporting evidence is slight. Morphologically, phoronids have a bulbous apex from which a thin, straight tube arises, quite unlike the gradually expanding morphology of Cornulites . More significantly, the phoronid tube is unmineralized, being formed of polymeric mucopolysaccharides ( Emig, 1982; Cohen, 2000) arranged as a central basophilic layer and two peripheral acidophilic layers. The central layer is formed of ‘numerous very thin parallel coats’ ( Emig, 1982: 39) but the hypothesis that this ‘microlamellar structure... secreted by the entire body surface’ (Vinn, 2005: 210) is homologous with the lamellar calcite skeleton of Cornulites is unconvincing. Phoronid tubes have neither the growth pattern nor the complexity of the cornulitid skeleton and, while brachiopods are accepted by most researchers as their sister-group and can have lamellar, pseudopunctate shells, using these features to create a composite suite of lophophorate shell characters and reconstruct cornulitids as ‘skeletal phoronids’ is entirely speculative. The only two groups with which Cornulites can be compared directly are bryozoans and cnidarians.

Dzik (1991, 1993) first raised the possibility that cornulitids were ‘close to the extinct ancestors of the Bryozoa’ ( Dzik, 1991: 128), having noted similarities between the putative Ordovician cornulitid Cornulitozoon and corynotrypid bryozoans. The characters upon which the hypothesis was based were that Cornulitozoon was of similar size and shape to corynotrypids, had the same non-porous wall structure and ‘funnellike apertural collars’ ( Dzik, 1991: 122), and an apical morphology virtually identical to that of the bryozoan ancestrula. He argued also that Siluro-Ordovician members of the two groups showed a closely similar original shell microstructure ( Dzik, 1991) and speculated that cornulitids diverged from bryozoans after the acquisition of a tubular, mineralized skeleton but prior to the development of small zooid size and a colonial life habit. This was supported broadly by Vinn & Mutvei (2005: 733–735), who noted that some bryozoans have a ‘vesicular wall structure... and an eggshaped embryonic shell’ similar to Cornulites .

Although the initial chamber of Cornulites certainly resembles that of bryozoans, its morphology is similar also to that seen in some corals (e.g. Stolarski, 2000: fig. 2D) and molluscs ( Vinn & Mutvei, 2005). Hence, the shape is more likely to reflect functional constraints than being indicative of a close phylogenetic relationship. Similarly, there are groups other than bryozoans that have lamellar low-magnesian calcite skeletons, including rugose corals ( Sandberg, 1975). Of the skeletal characters listed by Vinn (2005) as present in both cornulitids and bryozoans, psuedopuncta are the only feature for which a close analogue is lacking in zoantharians.

Tavener-Smith & Williams (1972) argued that the distribution of bryozoan pseudopuncta, which are developed most strongly in trepostomes, cystoporates and cryptostomes, indicated they are ‘restricted to [bryozoan] skeletons of coelocystic origin’ (Tavener- Smith & Williams, 1972: 156). From this it was deduced that pseudopuncta functioned as muscle bases that ‘improved the attachment of a highly folded, hypostegal epithelium to the skeletal surface’ ( Tavener-Smith & Williams, 1972: 156). Nothing is known of cornulitid soft tissues, but it is likely that their pseudopuncta served a similar function. Although no analogous microstructures have been described in corals, Wise (1970) and Muscatine, Tambutte & Allemand (1997) noted small pits and spines on the inner surface of some scleractinian skeletons, interpreted as the skeletal surface impressions of desmocytes that attached the soft tissues to the coral skeleton. The means by which the soft parts were attached to the lamellar calcite skeleton of Palaeozoic corals requires further investigation.

In terms of overall skeletal structure, cornulitid cellulae and tabulae are comparable with both the dissepiments and tabulae of corals and the cystiphragms and diaphragms of bryozoans. Their method of formation is comparable also, based on the studies of bryozoans by Tavener-Smith & Williams (1972) and of zoantharians by Wells (1969). There are no structures in cornulitids analogous to zoantharian septa, but this does not negate the possibility of a cnidarian affinity. Stolarski (2000) showed that the scleractinian species Guynia annulata was initially aseptate, whilst Fedorowski (1991: 417) argued it was ‘very probable’ rugose corals also lacked septa early in ontogeny. Additionally, septa are extremely weakly developed in early taxa such as Tabulaconus (see Debrenne, Gangloff & Lafuste, 1987).

Cornulitids were solitary organisms. Solitary taxa are known in many cnidarian groups, fossil and extant ( Scrutton, 1997), but not within the Bryozoa. While accepting that the ancestral bryozoan was probably solitary ( Dzik, 1991), it is more parsimonious to compare Cornulites with cnidarians. Furthermore, cornulitid shell symmetry is clearly radial, suggesting also a closer affinity to cnidarians. The pseudopuncta of Cornulites do resemble those of bryozoans, but we interpret this character as having been acquired convergently.