Background to skeletal evolution of the Asteroidea

Spencer and Wright (Reference Spencer, Wright and Moore1966, p. U9) subdivided the asterozoan skeleton into three categories: the axial or ambulacral; the adaxial or virgal/adambulacral/lateral; and the remainder, the extraxial. The biological and evolutionary significance of the echinoderm skeletal classification proposed by Spencer and Wright (Reference Spencer, Wright and Moore1966) was documented by Mooi et al. (Reference Mooi, David, Marchand, David, Guille, Féral and Roux1994), David and Mooi (Reference David, Mooi, Mooi and Telford1998), and Mooi and David (Reference Mooi and David1998, Reference Mooi and David2000, Reference Mooi and David2008), these authors preferring a two-fold axial and extraxial subdivision, the adaxial of Spencer and Wright (Reference Spencer, Wright and Moore1966) of uncertain assignment within axial-extraxial terminology (Mooi and David, Reference Mooi and David2000, p. 331). Early fossils have not been found to be helpful to evaluation of adaxial genesis; nevertheless, the adaxial concept of Spencer and Wright (Reference Spencer, Wright and Moore1966) has been found to be useful to understanding of the fossil record (Blake, Reference Blake2013, Reference Blake2018), and it is retained here. Extraxial ossicular expressions are ontogenetically relatively fluid as compared to those of axial and adaxial series, thereby likely contributing to complexities of certain occurrences noted here.

Diverse skeletal asterozoan remains from Early Ordovician strata are assigned to four class-level taxa (Blake, Reference Blake2013). Cambrian trace fossils suggestive of asterozoan origins have been recognized (Mikulás, Reference Mikulás1992), suggesting (but not demonstrating) that subphylum diversification might have begun prior to emergence of readily preserved skeletons (Schuchert, Reference Schuchert1915, p. 31; Sprinkle, Reference Sprinkle, Jangoux and Lawrence1983, p. 8; Blake and Guensburg, Reference Blake and Guensburg2015, p. 484).

In his comprehensive monograph of the Paleozoic Asterozoa, Spencer (Reference Spencer1914–1940) favored existence of a then-undetermined asterozoan group basal to the Asteroidea and Ophiuroidea; subsequently, Spencer (Reference Spencer1951) recognized the Somasteroidea as the first asterozoans, and since then assignment of the Somasteroidea to a stemward position has been widely but not universally accepted.

Following the thinking of both Spencer (Reference Spencer1951) and Fell (e.g., 1963), Spencer and Wright (Reference Spencer, Wright and Moore1966) favored a crinoid ancestry for the Asterozoa, whereas both Smith and Jell (Reference Smith and Jell1990) and Shackleton (Reference Shackleton2005) preferred an edrioasteroid hypothesis. Blake et al. (Reference Blake, Zamora and García-Alcalde2015b) and Blake and Guensburg (Reference Blake and Guensburg2015) accepted a stemward positioning for the Somasteroidea within the Asterozoa, but argued that no satisfactory subphylum outgroup is available.

The asterozoan skeleton consists of thousands of discrete elements partitioned among a comparatively small number of ossicular categories. Ossicular morphology varies not only among individuals and populations but also with ontogeny, such that an ossicle of given external dimension from a proximal arm interval of a small individual differs significantly from an ossicle of similar external dimensions from a distal arm interval of a conspecific large individual. Given the generally scanty fossil record of asterozoans, these complexities compromise understanding.

Ossicular arrangements of many fossil asterozoans are readily visualized as architectural floor plans; however, abutment surfaces between ossicles typically are inclined rather than perpendicular to any plane of view, with flexibility of the living and dying individual, sediment compaction, and weathering all locally altering ‘floor plan’ appearance and therefore potential reconstruction of life configuration. Specimens of Aerliceaster n. gen. (Figs. 1, 2) document concerns in that they retain near-life overall ossicular positioning, albeit with severe and misleading outcrop weathering.

The Somasteroidea and sourcing of asteroid morphology

Because no outgroup for the subphylum Asterozoa is identified (Blake and Guensburg, Reference Blake and Guensburg2015, p. 483), guidance to early asteroid history must be sought from within the Somasteroidea. For discussion purposes, the asterozoan skeleton is first separated into two basic categories, the axial-adaxial and the extraxial, the latter then with a three-fold subdivision consisting of the ambital framework, the dorsal or abactinal skeleton above the ambital framework, and within asteroids, but not somasteroids, the actinal skeleton between the marginal and adaxial series. The axial and adaxial category provided character complexes for evaluation of the Somasteroidea and the three derived asterozoan classes, the Asteroidea, Ophiuroidea, and Stenuroidea (Blake, Reference Blake2013; Blake and Guensburg, Reference Blake and Guensburg2015), as well as for three new ordinal concepts for Paleozoic Asteroidea, the Euaxosida, Hadrosida, and Kermasida (Blake, Reference Blake2018). At the current level of understanding, axial and adaxial ossicles provide little help for evolutionary sequencing within the Somasteroidea, whereas, and in contrast, the extraxial ambital framework is important, and the dorsal skeleton provides data useful for evaluation of the transition from Somasteroidea to Asteroidea.

Differentiated ossicles, termed ‘marginals,’ have been widely recognized among echinoderms. Because the term ‘marginal’ can carry implications of homology, the genetically neutral term ‘ambital framework’ was proposed (Blake, Reference Blake2013; Blake and Guensburg, Reference Blake and Guensburg2015). Marginal ossicular recognition in asteroids begins with uniformity of interpretation among authors; however, Blake (Reference Blake1978) argued that marginal series origin immediately proximal to the terminal ossicle provides an essential and objective criterion for identification of both marginals and, indirectly, of extraxial categories relative to the marginals. Beginning with the positional criterion, so-called ‘marginals’ of ophiuroids, edrioasteroids, and other echinoderms are not homologous with marginals of somasteroids and asteroids (Blake and Guensburg, Reference Blake and Guensburg2015). Shackleton (Reference Shackleton2005) rejected homology between asteroid ‘marginals’ and the commonly similarly termed series of ophiuroids while accepting ‘marginal’ homology between asteroids and edrioasteroids.

An ‘ambital framework’ was recognized for all five known somasteroid genera, but ‘marginal’ status definitively accepted for only three (Blake and Guensburg, Reference Blake and Guensburg2015). The ambital framework of Chinianaster and Thoralaster consists of many small, delicate, irregular ossicles forming an ‘ambital necklace’ (Blake and Guensburg, Reference Blake and Guensburg2015) that was judged likely, but not definitively, homologous with the ambital or marginal framework of the remaining three genera, and therefore accepted as stemward to true marginals. The ambital framework of Thoralaster is differentiated beyond that of Chinianaster, and therefore the latter is considered stemward.

Marginals of Villebrunaster are an irregular series of small ossicles that reaches the arm tip, it a logical derivative of an ambital framework such as that of Chinianaster. The more robust and clearly defined marginal framework of Ophioxenikos argues progression toward asteroid configuration, such as the marginals of Eukrinaster (Fig. 4.4, 4.5). A stemward positioning relative to Asteroidea for Archegonaster Jaekel is rejected because of morphology and stratigraphic position (Blake and Guensburg, Reference Blake and Guensburg2015, p. 483).

Ophiuroids lack marginals and therefore in this hypothesis, absence from Chinianaster and Thoralaster favors separation of the ophiuran clade prior to the Villebrunaster grade. The madreporite of Chinianaster lies within the virgal field, an expression consistent with closer linkage to the Ophiuroidea.

Because the above sequencing is based entirely on morphologic interpretations without reference to an outgroup, it is readily reframed. Villebrunaster morphology can be assigned to the stemward position, it giving rise to one branch leading toward Ophixenikos and the Asteroidea and a second branch exhibiting ambital framework reduction leading toward ophiuroids. Ambital framework ossicles occur in stenuroids sensu Blake (Reference Blake2013), suggesting that occurrence in at least this class might be paraphyletic or polyphyletic. MacBride (Reference MacBride1896, p. 398) placed asteroids stemward to ophiuroids in a diagram published early in the history of phylogenetic interpretation. At the beginning of his monograph, Spencer (Reference Spencer1914, p. 7) also favored a stemward positioning for asteroids. A stemward positioning for Villebrunaster might also be favored by hydropore-madreporite expression. Although a number of specimens of Villebrunaster are available, no madreporite has been recognized, and a madreporite is difficult to recognize among many early asteroids (perhaps in part for taphonomic reasons, see Kolataster n. gen., Fig. 3.1, 3.4); the absence of a recognized madreporite from Villebrunaster might reflect presence of only an uncalcified hydropore, the true madreporite emergent within the Asteroidea and therefore perhaps convergent, as suggested by Aerliceaster n. gen. (Fig. 2.5). Presence of a dorsal madreporite in Archegonaster favors at least possible presence of a clade close to the Asteroidea. The alternative possibilities above are consistent with complex branching within the Somasteroidea suggested by Mooi and David (Reference Mooi and David2000, fig. 6) as well as with occurrence of small, extinct early asterozoan clades (Blake, Reference Blake1995, Reference Blake2014; Blake et al., Reference Blake, Gahn and Guensburg2019). Either early diversification hypothesis is suitable for the evaluation of the Asteroidea because interpretation of this class begins with the Villebrunaster and Ophioxenikos branch.

Three dorsal skeletal expressions are recognized among somasteroids. That of Chinianaster, Thoralaster, and Villebrunaster consists of paxilliform ossicles linked by radiating connecting rods, the two ossicular types together forming a delicate open meshwork. A second dorsal expression is the closely fitted field of uniformly small and irregular ossicles of Ophioxenikos. The third, that of Archegonaster, is a field of tiny granules in which larger ossicles are lacking. Ophioxenikos is similar to Eukrinaster and many other stemward asteroids. The dorsal configuration of Archegonaster argues against a basal assignment for this genus. The fields of small, irregular ossicles typical of somasteroids also are similar to those of early ophiuroids and stenuroids.

As a part of the axial-adaxial system, the mouth frame is considered first below followed by the ambital framework, the dorsal skeleton, and accessories.

The axillary and mouth frame construction and function in Aerliceaster nexosus n. gen. n. sp

A problematic and debated ossicular occurrence among Paleozoic asteroids is that of the so-called axillary. The axillary characteristically is developed as an external, interbrachial, unpaired ossicle aligned with the inferomarginal series and abutting the mouth-angle pair (Fig. 1). The axillary is most readily recognized in taxa in which actinal ossicles are few or lacking, whereas, and in contrast, in taxa with an enlarged disk and many actinals, an axillary typically is difficult to recognize; in Eukrinaster (Fig. 4.4, 4.5), for example, marginal frame appearance only suggests presence of an axillary (Blake et al., Reference Blake, Guensburg, Sprinkle and Sumrall2007, figs. 2.6, 3.1), as does an appearance adjacent to the mouth-angle ossicle (Blake, Reference Blake2018, pl. 2.11). Difficulties of recognition cannot prove absence, at least in part because of size and shape issues but also because of expressions in other taxa, the axillary displaced from the arm margin onto the enlarged disk of Xenaster Simonovitsch (Blake and Hotchkiss, Reference Blake and Hotchkiss2004, fig. 1.4; Blake Reference Blake2018, pl. 13.2) suggesting potential ossicular loss, or at least loss of clear expression. Further, if axillaries are inferomarginal derivatives, generation in at least some lineages might have been step-wise rather than abruptly transformational, as suggested by the complex ventral disk expressions of Aerliceaster n. gen. Potential loss of clarity of expression is also suggested by occurrences within the crown group (Blake et al., Reference Blake, Bielert and Bielert2006, fig. 3; Blake et al., Reference Blake, Tintori and Kolar-Jurkovsek2017, fig. 2). Axillary ossicles have not been identified in somasteroids, although Fell (Reference Fell1963, fig. 8) did recognize a so-called ‘T-plate’ in Villebrunaster in the position of the axillary, an occurrence not subsequently confirmed. McKnight (Reference McKnight1975, p. 14), for example, argued presence of an odontophore (= axillary) as a key to asteroid recognition.

Mouth-frame construction and operation in both asteroids and ophiuroids was described and illustrated by Spencer (Reference Spencer1925, p. 263, fig. 186). In Spencer's view, in primitive forms, the mouth frame formed a trapping mechanism that, when opened, pulled mouth frame ossicles apart, and with closing, the ossicles grasped the prey and contracted toward the disk axis. Closure then created tension in the surrounding disk that gradually reopened the mouth frame as muscles relaxed.

All five interbrachia are exposed in one specimen of A. nexosus n. gen. n. sp. (Fig. 1) with axillary appearance from interbrachium to interbrachium ranging from well defined to somewhat obscure, the lack of distortion among the still closely fitted ossicles demonstrating original arrangement, without significant ossicular displacement that would hinder interpretation. Expressions of the better-preserved mouth-angle pair differ—one pair with a groove on each extended ossicle, a second pair comparatively retracted with robust edges of the mouth-angle ossicles abutting the adaxial, and in a third (situated between the other two), mouth-angle ossicles are rotated toward and into the mouth area. Although it is difficult to apply the operational interpretations of Spencer (Reference Spencer1925) to the mouth frame of Aerliceaster n. gen., disposition suggests mobility, an ability of the mouth-angle pair to extend and contract during feeding. The elliptical depression between the mouth-angle ossicular pair in this model was used for musculature to grasp and push particles into the mouth; although Spencer (Reference Spencer1925) used the term ‘prey,’ sediment feeding would be appropriate as well. The differing appearances of mouth-angle pair of the A. nexosus n. gen. n. sp. holotype are argued as having ‘frozen’ different feeding postures during the death event, the expressions likely aided by some tissue decay.

An elliptical depression between the proximal edges of ossicles of the mouth-angle pair, such as A. nexosus n. gen. n. sp. (Fig. 1.3), is common among Paleozoic asteroids, although variously obscured both by positioning at death and preservation (Spencer, Reference Spencer1922, p. 205 and associated discussions; Blake, Reference Blake2018, p. 24 for Schuchertia and Palasterina). The gap, commonly found in taxa with comparatively small accessories on the mouth-angle ossicles, seemingly would have served as effective gripping tools, in accord with the interpretations of Spencer (Reference Spencer1925).

Axillary/odontophore positioning on a small disk immediately adjacent to the mouth-angle pair arguably provided bracing; in contrast, life modes of genera with enlarged disks and small extraxial ossicular sizes might have favored flexibility and lesser need for bracing. The small actinal field of A. nexosus n. gen. n. sp. provided some flexibility rather than the rigidity of the single abutted axillary configuration. Two possible end-member hypotheses are enabled by A. nexosus n. gen. n. sp., either an enlarged actinal field reminiscent of the virgal field of somasteroids was being lost and a differentiated axillary was emergent, or the reverse, that axillary differentiation was declining with increasing actinal field size. Axillary transfer to the interior (i.e., traditional ‘odontophore’ status) would retain some bracing while restoring flexibility lost with axillary differentiation and small disk sizes.

Discussion here focuses on difficulties surrounding interpretation of presence of a single axillary ossicle, but potential presence of two or even multiple proximal inferomarginals variously incorporated into evolution of the axillary/odontophore configuration (Blake, Reference Blake2018) is a complexity consistent with above interpretations.

Ambital framework and dorsal ossicular expressions

Dorsal and body-margin ossicles of available Aerliceaster n. gen. are most clearly exposed in two arms of a single specimen (Fig. 2.1). Inferomarginals are recognized based on their occurrence as a distinctive series originating at the arm tip and positioning along the ambital margin. A lateral series of small, sub-rectangular ossicles abuts the inferomarginals and extends from the interbrachium to about the midarm position. The central portion of the arm is dominated by several rows of enlarged ossicles. The medial series of one arm is aligned, initially suggesting straightforward category determination, whereas rows of the second better-preserved arm are offset. Offset cannot be ascribed to growth abnormality because the concern here is interpretation of potential derivation of ossicular categories from the posited irregular ancestral configuration of somasteroids: ossicular ‘offset’ of Aerliceaster n. gen. could represent an emergent status of ossicular series differentiation. It also is not possible to argue taphonomic disruption because, although surfaces are heavily leached, ossicles appear in life positions both in this and the other specimens of the suite; any disruption was minor.

The medial series of the more complete arm extends from the arm tip, the apparent site of series genesis, to the disk, whereas tracking of the ossicular series reaching the arm tip of the second arm is ambiguous because of offset (Fig. 2.1, 2.2, 2.4) and irregularity of ossicular arrangement at the disk. Alignment at the disk at least suggests continuity with the lateral series of the better-preserved arm. Lateral series of neither arm appears to reach the arm tip, thereby failing the positional criterion for superomarginal series. For these reasons, in the taxonomic section above, the three series are described as midarm series, but not recognized as representing either superomarginal nor carinal series. Expression, however, is consistent with the suggestion of Shackleton (Reference Shackleton2005) of secondary superomarginal derivation, perhaps from abactinal series, here with alignment emergent.

Kolataster n. gen. (Fig. 3) is of a common overall asteroid shape. An inferomarginal series is present and dorsal ossicles are closely fitted; like Chinianaster and unlike Ophioxenikos, many ossicles are paxilliform and a second ossicular type is present. Dorsal ossicles of Kolataster n. gen. appear aligned into series as in Aerliceaster n. gen., but no differentiation of arm series (i.e., carinals and superomarginals) is recognized, and in that way Kolataster n. gen., like Aerliceaster n. gen., suggests an emergent status, but one in which series differentiation suggests separate evolutionary pathways.

Disk abactinals

Spencer (Reference Spencer1914, p. 34) endorsed the notion that “the most primitive Asterozoa have a leathery skin in which are embedded small irregular plates,” a view consistent with his later view that somasteroids, with their dorsal skeletons of comparatively uniform, small, and irregularly arranged skeletal elements, represent “ … the first stages in the differentiation of a starfish” (Spencer, Reference Spencer1951, p. 87). Dorsal disk ossicles of both the somasteroid Ophioxenikos and the asteroid Eukrinaster are uniform, small, closely fitted, irregular in arrangement, and not known to be differentiated into carinal or concentric disk series; the asteroid condition is considered plesiomorphic and derived from an ancestral somasteroid condition.

For Ordovician Hudsonaster, Schuchert (Reference Schuchert1915, p. 32) described a simple arrangement of comparatively few closely fitted dorsal disk ossicles, and Spencer (Reference Spencer1916, p. 64, Reference Spencer1918, p. 128) discussed and provided diagrammatic reconstructions of ossicular arrangements for both Urasterella and Hudsonaster. Comparatively few larger apical ossicles are present in the two, these arranged in one or two irregular but approximately pentameral cycles. Citing Stürtz (Reference Stürtz1893), Spencer (Reference Spencer1914, p. 34) suggested that although many “primitive” asteroids show the apical system of urasterellids, this arrangement is not ancestral “but is acquired secondarily and independently in many forms.” Although the term “primary circlet” was not used by Spencer (Reference Spencer1918), Spencer and Wright (Reference Spencer, Wright and Moore1966, p. U30) defined a dorsal disk primary circlet as a: “Ring of prominent ossicles on aboral surface, typically consisting of five radials and five interradials surrounding centrale.”

The configuration of Aerliceaster n. gen. is incompletely exposed, but a centrale and an enclosing ring are suggested, and surrounding ossicles are numerous, the complex seemingly similar to that described for Hudsonaster and Urasterella. Preservation of Kolataster n. gen. is incomplete, but the numerous similar ossicles, perhaps concentric in arrangement, is suggestive of Aerliceaster, Hudsonaster, and Urasterella. Collectively, dorsal disks of exemplars here are consistent with Spencer's (Reference Spencer1914) views of the ancestral condition of Asterozoa, the Somasteroidea representing the first of the Asterozoa (Spencer, Reference Spencer1951, p. 87) followed by ongoing potentially convergent ordering of the dorsal disk skeleton (Schuchert, Reference Schuchert1915; Spencer, Reference Spencer1916, Reference Spencer1918).

Many asteroids exhibit a regular five-fold apical ossicular configuration early in their ontogenetic skeletal development (see Fewkes, Reference Fewkes1888; Kano et al., Reference Kano, Komatsu and Oguro1974; Komatsu, Reference Komatsu1975), and comparable five-fold ossicular configurations about central axes are widely distributed among pentameral echinoderms (e.g., Spencer and Wright, Reference Spencer, Wright and Moore1966, fig. 26; Ubaghs, Reference Ubaghs and Moore1967, fig. 18). Such arrangement, however, is not considered to be of fundamental significance to echinoderm phylogenetic interpretation (e.g., Spencer, Reference Spencer1914, p. 34; Philip, Reference Philip1965).

Madreporite

Among the five known genera of Somasteroidea, a hydropore or madreporite has been recorded for only two. In Archegonaster, a madreporite is positioned dorsal to the single row of marginal ossicles, and in Chinianaster, the madreporite is located within the virgal field ventral to an ambital framework of tiny ossicles, these likely, but not definitively, homologous with the marginals of asteroids (Blake and Guensburg, Reference Blake and Guensburg2015). The stemward asteroid condition of water pore development is ambiguous because of differences among somasteroid occurrences.

Madreporic expressions are varied among early Paleozoic asteroids. A hydropore/madreporite has not been recognized in all early species, likely at least in part due to taphonomic disruption. Presumably, the ancestral condition is that of a simple opening surrounded by small ossicles. Small sample size limits interpretation of Aerliceaster n. gen., nevertheless preservational style likely could enable recognition of a clearly differentiated madreporite, were one present. The best-preserved dorsal disk of Aerliceaster n. gen. has an apparent circlet of small ossicles (Fig. 2.5) most like the anal pore of extant asteroids, but alternatively, it might represent a hydropore prior to madreporic differentiation. The tiny but clearly differentiated madreporite of Kolataster n. gen. (Fig. 3.1, 3.2, 3.4) demonstrates that the ossicle in fossil occurrences can be too small to be readily seen.

Actinal skeleton

The actinal skeleton of Asteroidea was posited as homologous with the virgal skeleton of the Somasteroidea by Shackleton (Reference Shackleton2005), whereas Blake and Hotchkiss (Reference Blake and Hotchkiss2004) and Blake and Guensburg (Reference Blake and Guensburg2015) argued virgal loss was followed by homoplastic derivation of actinals among asteroid clades. Actinal shapes and arrangements among interbrachia in Aerliceaster n. gen. are all but chaotic (Fig. 1), whereas actinal configurations among extant asteroids are regular, regularity having been argued as emerging independently among clades (Blake and Hotchkiss, Reference Blake and Hotchkiss2004, p. 361; Blake et al., Reference Blake, Donovan, Mah and Dixon2015a, p. 1049). Data for Kolataster n. gen. are not available.

Accessory ossicles

Accessory ossicles are the granules, spines, spinelets, and pedicellariae seated on the ossicles of the primary skeleton. Identification of granular ossicles either as ‘primaries’ or ‘accessories’ can be ambiguous. For example, in many extant Poraniidae, small granules occur in a dermal layer, but enlarged ossicles are lacking and therefore either ‘primaries’ were lost leaving only accessory granules, or, less likely, the ‘granules’ might be reduced primaries (‘less likely’ because intermediate reduced states have not been recognized in the Poraniidae, and size reduction leading to a uniform granular size suggests a more complex process than ossicular loss). The concern takes on pertinence for interpretation of early asterozoan history because only small granules have been reported from the dorsal disk of Archegonaster (Smith and Jell, Reference Smith and Jell1990, p. 764).

In Aerliceaster n. gen., some small ossicles are, or at least appear, fitted between adjacent marginal and the adaxial ossicles, and therefore argue a part of the primary skeleton, whereas others appear to rest on the adjacent ossicles and therefore suggest accessories. The complexities potentially challenge category determination and therefore also series identification among asteroids, for example, within the Mesopalaeasteridae sensu Blake (Reference Blake2018). Alternatively, differences might only document comparatively trivial local growth displacement rather than a broader challenge to recognition of ‘primary’ and ‘accessory’ ossicular recognition.