Abstract: Brachiopods are marine Lophotrochozoa whose soft parts are enclosed in a bivalved shell. Although brachiopods are represented by a rich record from the Early Cambrian to the present, the origin of their bivalved body plan remains controversial. The Early Cambrian organophosphatic tommotiids Micrina and Paterimitra from Australia have been proposed as stem brachiopods. Here, we describe their earliest ontogeny, indicating that tommotiids possessed bivalved planktotrophic larvae. The curious combinations of characters in Micrina and Paterimitra indicate that they may belong to the stems of the Linguliformea and Rhynchonelliformea, respectively. The bivalved shell of adult living brachiopods may represent a plesiomorphic character retained from planktic tommotiid larvae; the crown group body plan of the Brachiopoda may have evolved through the paedomorphic retention of a bivalved larval state.
The material comes from richly fossiliferous Australian lower Cambrian successions in the Northern Territory and South Australia (Laurie 1986; Bengtson et al. 1990). All examined material of Micrina comes from the Todd River Dolostone (Northern Territory locality (NT) 600; Laurie 1986), kindly provided by J. Laurie (Canberra), and the Wilkawillina Limestone (sample 92-20, close to National Museum of Victoria locality (NMVPL) 1594; Bengtson et al. 1990). All examined material of Paterimitra comes from the lower Cambrian of the Arrowie Basin, Flinders Ranges (sample MMF/0.0; Skovsted et al. 2009a).
Fossils were dissolved out of carbonate rocks by 10 per cent acetic acid and coated with gold for examination under scanning electron microscopes (SEM). All illustrated material is deposited in the South Australian Museum, Adelaide (SAMP), or in the Commonwealth Palaeontological Collections, Canberra (CPC).
Micrina from the Northern Territory and South Australia (Laurie 1986) is composed of two types of sclerites, high conical mitrals (Text-fig. 1A) and smaller low arcuate sellates (Text-fig. 1E), that were interpreted as a brachiopod-like bivalved animal by Holmer et al. (2008). Paterimitra from South Australia is more complicated, and the base has two bilaterally symmetrical sclerites: a pyramidal S1-sclerite (Text-fig. 2A) and a smaller saddle-shaped S2-sclerite (Text-fig. 2B). Partial scleritomes demonstrate that the asymmetrical L-sclerites surrounded the apertural margin of the S1-sclerite to form an open cone in which the symmetrical sclerites are joined together around a small basal tubular attachment opening (Skovsted et al. 2009a).
The apices of Micrina and Paterimitra sclerites preserve a succession of major growth discontinuities that are of consistent size and morphology. The mitral sclerite of Micrina has a circular to hemispherical growth disturbance, 200–300 μm in transverse diameter, the larval shell (Text-fig. 1A–C). The centre of the larval shell has a second growth disturbance, defining one pair of lateral inflated lobes with an outer rim, 100 μm wide, here termed the protegulum (embryonic shell). The protegulum has a median mound and transversely folded extension posteriorly. The surface of the mitral protegulum is wrinkled, but imperforate (Text-fig. 1A, D, J). The larval shell is perforated by 8–12 openings of setal tubes, four of which are consistently positioned close to the protegulum lobe margins in the configuration shown in Text-figure 1C. The larval shell and protegulum are also provided with numerous nick points (sense of Williams and Holmer 2002; Text-fig. 1C). In the sellate sclerite of Micrina, the protegulum is elongately oval, 100 μm wide and only 10–15 μm long; the larval shell represents a larger version of the protegulum; both the protegula and larval shells are invariably perforated by two pairs of setal tubes (Text-fig. 1E–G, I).
In the pyramidal S1-sclerite of Paterimitra, the larval shell is defined by a major growth disturbance, 300–500 μm wide. The shell has the same outline as the adult sclerite, and the anterior side of the sclerite possesses two low radial ridges defining a central semicircular indentation, which will be covered by a plate, the colleplax, through later ontogeny (Text-fig. 2A, C–D, J). The protegulum is saddle-shaped and smooth, usually 100 μm wide and about as long, forming the posterior protruding flange around what eventually develops into the attachment opening of the S1-sclerite (Text-fig. 2A, C–D, J). The larval shell of the Paterimitra S2-sclerite is acutely triangular, 150–200 μm wide, with the narrow, posterior, tapering end raised into an up-turned flange. The protegulum of the Paterimitra S2-sclerite is poorly defined, but generally smooth, and forming a half-moon-shaped plate, 50–180 μm wide, and about as long; the posterior margin has a shallow indentation (Text-fig. 2B, F–G, K).
Fossils of stem group and crown group brachiopods can record the ontogeny in ultrastructural detail back to the Cambrian (e.g. Chuang 1977; Holmer 1989; Williams et al. 1996; Freeman and Lundelius 1999, 2005; Popov et al. 2007, 2010). The earliest preserved stages of both Micrina and Paterimitra are here interpreted as embryonic shells or protegula, secreted by the embryo just before hatching; they are similar in size and morphology to protegula reported from extant rhynchonelliform and linguliform brachiopods (Freeman and Lundelius 1999, 2005). The mitral protegulum of Micrina had one pair of setal sacs enclosed by lateral lobes, whilst the ventral protegulum has two lateral setal tubes (Text-fig. 1H). This type of embryonic shell is most similar to that of the linguliforms, where similar evidence for setae has been recorded from the protegulum (e.g. Balinski 1997; Tapanila and Holmer 2006). The transversely folded posterior extension of the mitral protegulum in Micrina is comparable with the transversely folded linguliform attachment rudiments described by Balthasar (2004, 2009, fig. 2) and suggest that Micrina had an embryonic pedicle (Text-fig. 1H). The nick points indicate normal follicular mantle setae, and planktotrophic larval shells of this type are known from a numerous linguliforms (Williams and Holmer 1992). The stratiform shell of Micrina is also closely comparable to basal linguliform brachiopods in structure and indistinguishable from Mickwitzia as well as more derived linguliform brachiopods (Skovsted and Holmer 2003; Balthasar 2004).
The protegulum and larval shell of Paterimitra lack evidence of embryonic or larval setae. The larval shell of the pyramidal S1-sclerite of Paterimitra is identical to the ventral larval shell of the organophosphatic stem group rhynchonelliform Salanygolina, which has the same outline and a colleplax (Text-fig. 2C, I–J). Moreover, the colleplax is also present in the calcareous chileates, thus providing a link to the Rhynchonelliformea (Holmer et al. 2009). Both Paterimitra and Salanygolina also have identical posterior delthyrium-like flanges, which in Paterimitra becomes the attachment opening of the S1-sclerite. In contrast, the adult attachment of Salanygolina was achieved by the perforation and colleplax, a structure formed by the ventral mantle. Holmer et al. (2009) speculated that the earliest larval attachment was performed by an emerging posterior pedicle rudiment, which then atrophied. The function of the colleplax in Paterimitra is conjectural, but it may have housed a specialized type of ventral epithelium enclosing parts of the ventral mantle cavity (Text-fig. 2I). The protegulum of the Paterimitra S2-sclerite is of similar size to that of the dorsal protegulum of Salanygolina (Holmer et al. 2009), and both shells have flange-like structures (Text-fig. 2C–D, I–J). However, the larval shell of the Paterimitra S2-sclerite has no larval setal sacs (cf. Holmer et al. 2009).
Paterimitra and Salanygolina also share several potential homologous characters related to their skeletal secretion and surface imprints, which include distinctive polygonal microstructures and the development of peculiar shell-internal cavities; characters also shared with some paterinide brachiopods and the tubular tommotiid Eccentrotheca (Skovsted et al. 2008; Balthasar et al. 2009).
The evidence from Paterimitra and Micrina seems to suggest that they belong within the stems of the rhynchonelliform and linguliform clades, respectively (see also Skovsted et al. 2009a, 2011). We suggest that the bivalved adult brachiopod shell represents a plesiomorphic character retained from planktic tommotiid larvae; pending further phylogenetic analysis, the crown group body plan of the Brachiopoda may have evolved independently in living linguliform and rhynchonelliform brachiopods (see also Gorjansky and Popov 1986) through paedomorphic retention of a bivalved larvae.
Acknowledgements. This research was funded by grants from a Macquarie University Development Research Grant to CBS and GAB, the Swedish Research Council (VR) to CBS, GB and LH, as well as from the National Science Foundation of China (NSFC 40702005) to ZZ. Dr. John Laurie is thanked for providing specimens of Micrina. Reviews from Leonid Popov (Cardiff) and Mark Sutton (London) improved the manuscript.