Introduction

The Cambrian is one of the fulcrum points in the history of life. It is during the Cambrian that metazoans first start to appear in great numbers in the fossil record, which in turn precipitated a major change in the nature of the sea floor (Bottjer et al. 2000). Whereas the typical late Neoproterozoic sea floors are characterised by the presence of well-developed microbial mats and an absence of sediment mixing by vertically oriented burrowing, the Cambrian saw a marked increase in sediment disturbance (Gehling 1999; Droser et al. 1999) and the demise of microbial-stabilised firm grounds. The new Phanerozoicstyle seafloors resulted in relatively greater water content of seafloor sediment and a blurry water-sediment interface, which led to the first appearance of a mixed layer (Bottjer et al. 2000). The effect of this change in dominant substrate types on the ecology and evolution of benthic metazoans has been termed the Cambrian Substrate Revolution (CSR), and was a global event of primary importance for the early diversification of echinoderms (Bottjer et al. 2000). During the early Cambrian the majority of echinoderms rested directly on non bioturbated soft substrates as sediment stickers and sediment attachers, but with increased disturbance and the formation of a well developed mixed layer echinoderms were supposedly propelled to develop new substrate-adaptative morphologies and strategies (Dornbos 2006).

The group of echinoderms that made the most dramatic shift from firm bottom sediment attachers to hardground cementers during this interval are the edrioasteroids. Edrioasteroids are an extinct clade of small, sessile forms with a globular to discoidal theca lacking exothecal appendages. The majority of edrioasteroids reported from the middle Cambrian are fully plated forms that have little or no modification of their lower surface for attachment. Forms such as Stromatocystites (Smith 1985; Smith and Jell 1990; Parsley and Prokop 2004), Cambraster (Jell et al. 1985; Smith 1985; Zamora et al. 2007), Edriodiscus (Smith and Jell 1990) and Totiglobus (Bell and Sprinkle 1978) all have upper and lower surfaces fully plated, and are never found attached to shells or recognisable hard substrates. As speculated by Parsley and Prokop (2004) these were probably only loosely attached to the sea floor through mucous adhesion. By the Ordovician, however, these forms had disappeared and been replaced by either deep infaunal forms with a retractable peduncle (rhenopyrgids; Smith 1985; Holloway and Jell 1983) or fixed epifaunal isorophids with an uncalcified aboral surface and well developed peripheral ring that are always found attached to hardgrounds and/or shells (Guensburg and Sprinkle 1994; Sumrall et al. 2006). These have uniserial ambulacral floor plates in contrast to the biserial or quadriserial flooring plates of stromatocystitids and their kin.

Fig. 1.

A. Map showing the locality studied in the north part of Iberian Chains (Northeast Spain). After Gozalo and Liñán 1988. B. Map of Spain with the location of Iberian Chains. C. Synthetic stratigraphic column from Purujosa showing the levels from which the edrioasteroids were collected. CZ, Cantabrian Zone; ELAZ, East Lusitanian—Alcudian Zone; GCZ, Galician—Castilian Zone; IC, Iberian Chains; OMZ, Ossa-Morena Zone; SPZ, South Portuguese Zone; WAL, West Asturian—Leonese Zone.

There is, however, evidence that the upper Cambrian was a key period in edrioasteroid evolution. There are four records of edrioasteroids from this time interval. One indeterminate edrioasteroid from North America (Sumrall et al. 1997), a possible stromatocystitid from France (Ubaghs 1998) and two isorophid edrioasteroids from Australia, Chatsworthia and Hadrodiscus (Smith and Jell 1990), both of which have unplated aboral surfaces. Unfortunately, because the latter were recovered by acid preparation, it is not known whether these lived permanently attached or not. Furthermore, a new edrioasteroid from the middle Cambrian of China (Zhao et al. in press) shows intermediate characteristics. As a juvenile it appears very isorophid-like with a well-developed marginal ring forming a broad, aboral attachment area lacking aboral plates. However, it grows into a more stromatocystitid-like form that apparently adhered loosely to stabilised sediment.

We here report two new isorophid edrioasteroids from the early middle Cambrian (Cambrian Series 3) of Spain. These are not only the oldest members of this clade, but also demonstrate that the substrate revolution was already in progress, as both are clearly permanently attached to hard substrates.

Institutional abbreviations.—GM, Museum of Palaeontology, Guizhou University, Guiyang, China; MPZ, Zaragoza University Palaeontological Museum.

Geological setting and stratigraphy

The edrioasteroids were collected in the Purujosa 3 (Pur3) and Purujosa 4 (Pur4) sections in the northern part of the Iberian Chains (NE, Spain), 2 km south of Purujosa village (Fig. 1). Both specimens come from the uppermost part of the Murero Formation but from different stratigraphic levels. This formation ranges in age in the studied area from early Caesaraugustan to early Languedocian (middle Cambrian). The Murero Formation consists of a siliciclastic succession with some interbedded carbonate nodules (Liñán et al. 2002) and has been interpreted as deposited in transgressive conditions representing offshore environments (Álvaro and Vennin 1997; Gámez et al. 2009). Edrioasteroids are restricted to two layers. The older consists of green silstones very rich in complete echinoderms, including the edrioasteroid reported here Isorophida gen. et. sp. indet., the gogiid eocrinoid Gogia parsleyi, the cinctan Gyrocystis platessa, as well as isolated Ceratocystis-like stylophoran plates and trilobite moults mainly belonging to Eccaparadoxides brachyrachis (see Zamora et al. 2009 for more details). The trilobite fauna does not provide a precise constraint on the age of this horizon, but based on levels above and below the layer yielding the edrioasteroid is uppermost Caesaraugustan—lowermost Languedocian in age.

The younger level is a thin layer (< 1 m in thickness) of red lutites that are richly fossiliferous. Associated fauna includes echinoderms (the edrioasteroid Protorophus hispanicus gen. et sp. nov., the cinctan Gyrocystis platessa, the stylophorans Ceratocystis sp., plus two different undescribed cothurnocystids, and the blastozoans Gogia gondi, Lichenoididae indet., and Eocystites sp.), polimerid trilobites (Eccaparadoxides^? pradoanus, Eccaparadoxides brachyrachis, Conocoryphe heberti, Solenopleuropsis thorali, Solenopleuropsis marginata, S. verdiagana), agnostoid trilobites (Condylopyge sp., Peronopsis acadica, P. ferox, Plaurocterium sp., and Megagnostus sp.), and both orthid and lingulid brachiopods. Esteve et al. (in press) and Zamora (2010) provide further details on the fauna. The trilobite Solenopleuropsis thorali, indicates an age of earliest Languedocian (sensu álvaro and Vizcaïno 1998) for this level.

Middle Cambrian isorophid precursors

The two edrioasteroids recovered (Figs. 2, 3) are both rather small (4.3 mm and 11 mm in diameter) and have the body clearly divided into a peripheral rim and a central disc. Only the upper surface appears to be plated. The peripheral rim is circular and consists of a series of plates which become progressively smaller towards the edge of the rim. The largest innermost plates overlap on their oral face, but more or less abut to form a flat pavement aborally. The aboral surface of these plates lacks the radially aligned ridges and grooves that are commonly found in post-Cambrian isorophid edrioasteroids. Because of the arrangement of plates the peripheral rim is highest at its interior edge and slopes outwards.

The central disc is disrupted and largely missing in one specimen, but the other has a more or less undisrupted surface (Fig. 2A, B). The centre of the zone is covered by seven large, tall oral plates that appear to have abutted in life, although they are now slightly disrupted due to post-mortem collapse. Around the outside the most prominent elements are large U-shaped ossicles, three of which are clearly visible, one is partially visible and one further plate is presumably hidden by overlying platelets. These fill the space between the oral plates and the peripheral rim and are separated from each other by three or four small, somewhat imbricate plates. These U-shaped plates we interpret as ambulacral flooring plates. In close proximity to the U-shaped plates and overlying them are two or three small lath-like plates that, based on their shape and position, we interpret as disarticulated cover plates. In one interray there are a series of small lathe-like plates, again rather disrupted, that we interpret as the periproctal cone. The position of the periproct agrees with the inferred 2-1-2 pattern of oral plates in the central region.

Viewed from beneath (Fig. 2B, D) plate boundaries in the disc are difficult to make out, although the plates of the peripheral rim are clearly distinguished. While the central region beneath the oral plates and the periproct is clearly open and unplated, the outer part of the disc forms a flat calcified surface with few unambiguous plate boundaries. Large plates with a hemispherical outline coincide precisely in position with the U-shaped ambulacral plates seen in oral view, implying that these are the aboral view of the same plates.

A well developed peripheral rim is a characteristic of isorophid edrioasteroids, but is also seen in juvenile of the new edrioasteroid from China (Fig. 4B, C) (Zhao et al. in press). However, no edrioasteroid has ever been described with U-shaped terminal ambulacral flooring elements. The specimen that preserves disc plating is small, less than 5 mm in diameter, and thus potentially a juvenile. However, the ontogeny of the new edrioasteroid from China has been described in detail from individuals as small as 0.5 mm in diameter. At equivalent sizes (3–5 mm) Chinese specimens (Fig. 4B, C) clearly has paired adradial ambulacral, flooring plates that are visible both internally and externally, and a mouth frame constructed of six large integrated interradial plates (Zhao et al. in press). By contrast Protorophus has no clear mouthframe and single U-shaped terminal plates. It is this uniseriality of ambulacral flooring plates that suggests that Protorophus is closer phylogenetically to isorophids than to the new edrioasteroid from China.

Nor is our specimen likely to be a juvenile Cambraster. Cambraster is an edrioasteroid-like form that has a well developed marginal ring of stout plates which are bounded externally by a narrow peripheral rim of plates (see Jell et al. 1985; Smith 1985; Zamora et al. 2007). The ambulacra in adults are underlain by biserial flooring plates along almost all of their length, but the very tip of the ambulacra extends onto the marginal ring which has a corresponding U-shaped notch to house the distalmost coverplates.

Nevertheless, our specimen is not a juvenile Cambraster because firstly, the aboral surface is not fully plated, and secondly the complete ontogeny of Cambraster is now known (unpublished data) with specimens ranging from 6 to 35 mm in diameter. These clearly show that a continuous marginal ring of large plates is present from the earliest stages of growth, and is composed of its full adult complement of about 20 marginal plates by 6 mm diameter.

Fig. 2.

Edrioasteroid Protorophus hispanicus gen. et sp. nov., Purujosa (Zaragoza Province, Spain). Holotype MPZ2009/1233 in adorai (A) and aboral (B) views; camera lucida drawing of the adorai view (C) with interpretation of major anatomical structures; camera lucida drawing of the aboral face (D). Note the base of the flooring plates. Scale bar is 1 mm for all specimens. Photographs are taken from latex casts whitened with NH₄Cl sublimate. Abbreviations: LAO, left anterior oral plate; LBO, left branching oral plate; RAO, right anterior oral plate; RBO, right branching oral plate.

Attachment in echinoderms and the Cambrian substrate revolution

It seems highly likely that attachment at metamorphosis is primitive for echinoderms in general. Both crinoids and asteroids pass through an attached stage in their development, before becoming free living. Indeed it is the adoption of obligate attachment at the anterior and the subsequent body axis rotation that allowed echinoderms to break away from the straight-jacket of bilateral symmetry and explore new body-plans (Smith 2008). Larval attachment at metamorphosis was therefore probably prevalent and possibly obligate amongst Cambrian echinoderms. Early in the Cambrian attachment was retained into adulthood by some stalked forms. Cambrian gogiids for example have plated holdfasts that allow them to attach directly to the bottom or to shell debris on the sea floor and were sessile throughout life (Sprinkle 1973; Parsley and Zhao 2006; Zamora et al. 2009). The same is also true for early solutes (Daley 1996).

Cambrian edrioasteroids, such Stromatocystites, Totiglobus, Cambraster, and the new Chinese taxa, have been interpreted as facultative sediment attachers (Bell and Sprinkle 1978; Smith and Jell 1990; Dornbos 2006; Parsley and Prokop 2004; Zhao et al. in press). Yet the discovery of immediately post-larval individuals in the new edrioasteroid from China (Zhao et al. in press) suggest that attachment of juveniles was crucial even in these Cambrian groups. By the beginning of the Ordovician almost all edrioasteroids had switched to become sessile cementers throughout life, utilising a variety of different substrates. These include exposed concretions, hardgrounds, shell pavements, isolated skeletal debris, microbial buildups and encrusting bryozoan colonies, living organisms and pebbles (see Sumrall et al. 2006, for a synthesis).

Our small individuals are clearly firmly attached to skeletal debris. Zhao et al. (in press) also report that very occasionally the new Chinese edrioasteroid is found attached to skeletal debris (presumably as small individuals). What this implies is that the basic isorophid mode of life arose by retention of the early larval attachment phase into adulthood. While skeletal debris was sparse and hardgrounds lacking this early attachment phase was abandoned in favour of free-lying. However, as the Cambrian substrate revolution progressed and bioturbation increased in both depth and intensity during the Cambrian (Bottjer et al. 2000), a transition occurred in the dominance of subtidal soft substrates from firm Proterozoic-style soft substrates without a well-developed mixed layer to soupier Phanerozoic style soft substrate with a well-developed mixed layer. This substrate change greatly affected the evolution of echinoderms with the apparition of new body plans and different ecological strategies. Dornbos (2008) examined the fossil record of three echinoderm groups (edrioasteroids, eocrinoids, helicoplacoids) indicating that genera adapted to a typical Proterozoic style soft substrates as sediment attachers, sediment resters and shallow sediment stickers were taxonomically dominant during the Early and Middle Cambrian.

The Cambrian substrate revolution clearly affected the evolution of echinoderms but in the case of edrioasteroids, their history was more closely tied to the development of hard-grounds in the fossil record. In the lower Cambrian, when skeletal debris were less common and hardgrounds absent, edrioasteroids attached only as juveniles and became free-living or sediment attachers as they developed. The origin of isorophids as obligate encrusters presumably involved retention of this larval stage into adulthood and became possible only with the expansion of hardground environments from the late Cambrian onwards (for a history of hardgrounds see review by Taylor and Wilson 2003). Nevertheless, this evolutionary shift was already underway by the Middle Cambrian, as our discovery of an 11 mm diameter adult attached to a trilobite moult (Fig. 3A) demonstrates. The fact that such occurrences are extremely rare (we have just one example from a horizon that has yielded more than 150 cinctans, 17 gogiids and only one edrioasteroid) implies that opportunities for hardground colonisation were few and far between at this time.

Systematic palaeontology

Fig. 3.

Isorophida gen. et sp. indet., MPZ2010/857, Purujosa (Zaragoza Province, Spain). A. General view of specimen attached to a paradoxidid cranidium. B. Detail of the oral surface of the theca, note the well developed marginal ring and the disarticulated spines in the central theca. C. Aboral surface of the same specimen. Photographs are taken from latex casts whitened with NH₄Cl sublimate.

Fig. 4.

New edrioasteroid from the middle Cambrian of China (Zhao et al. in press). A. Adorai view of an adult specimen GM2103. B. Adorai view of a juvenile specimen GM1331 C. Aboral view of a juvenile specimen GM1670. B, C show a marginal ring. Photographs are taken from latex casts whitened with NH₄Cl sublimate. Courtesy of Colin Sumrall.