Axial / Extraxial theory
Abridged english v.
Contributions of the extraxial-axial theory to understanding the echinoderms
by Bruno DAVID(1) & Rich MOOI(2)
FIGURES ARE AVAILABLE ON THE EXTENDED FRENCH VERSION HERE
Although of bizarre morphology, echinoderms exhibit a number of derived features (apomorphies) that makes the recognition of the 5 extant and 21 extinct classes (Ubaghs, 1967a) relatively unproblematic : (1) a calcite skeleton made of a three-dimension network of trabeculae, the stereom ; (2) a water vascular system ; (3) pentaradial symmetry. The latter feature, embodied by the pentaradial model (fig.a) and exemplified by echinoids and asteroids, is described in texts as an organizing principle of the phylum (Brusca & Brusca, 1990 ; Enay, 1993). In this model, echinoderms show no trace of longitudinal segmentation evident in other phyla and the bilateral structure of the larvae is only awkwardly connected to adult pentamerism. The major problem with the pentaradial model is that it is characterized by the absence of a coherent system of homologies that can be applied to all echinoderm groups. This stems partly from the 200-year-old tradition of accentuated differences among the groups, rather than commonalities, and the frequent use of similarity in position alone as a basis for assigning homologies. Recently, we developed a new system of skeletal homologies called the extraxial-axial theory (EAT) which questions the usefulness of pentamerism in enhancing our understanding of the phylum.
The E.A.T. - The EAT is a model of architectural homologies based not just on similarity of position, but on major aspects of ontogeny (Mooi et al., 1994 ; David & Mooi, 1996 ; Mooi & David, 1997). There are two major components in the body wall of an echinoderm. Axial skeleton is associated with the water vascular system. New elements are added adjacent to the primary podium in a pattern called the ocular plate rule (OPR), with the oldest elements next to the mouth, and the youngest at the tip of the growing axial skeleton (fig.1c). Extraxial skeleton forms anywhere in the body wall outside the axial system, and comprises two sub-regions : perforate extraxial, which contains several types of body openings including the anus, the gonopores, and the hydropore ; and imperforate extraxial which closes off the coelom in the most aboral part of the body wall. Extraxial skeleton is most strongly expressed in plesiomorphic taxa like Camptostroma, in which the axial skeleton is restricted to thin strips (fig.2a). In the cystoids, imperforate extraxial skeleton forms a stem, and the perforate extraxial forms a calyx upon which the axial is mounted (fig.2b). Asteroids have lost imperforate extraxial skeleton, but the axial has become a conspicuous component of the oral surface (fig.2c). An echinoid is made up almost entirely of axial skeleton arranged in 5 growth zones (fig.2d). Asteroids and echinoids, in spite of their use in texts as exemplary echinoderms, have strongly departed from the skeletal topology of basal echinoderms.
Tests of the EAT
The first test explores the ontogenetic and genetic validity of the divisions among axial, perforate extraxial, and imperforate extraxial skeletons. The second tests the model in explanations of skeletal configurations left unclarified by the pentaradial model.
A) Embryological and genetic bases of the EAT - We rely on data only indirectly related to the empirical observations that led us to suggest the major subdivisions of the body wall. Embryology supports the EAT because the axial skeleton of adults arises solely from hydrocoel derivatives in the rudiment, whereas the extraxial skeleton is contributed solely from the non-rudiment part of the larva (fig.3). Therefore, the axial and extraxial components of the body of an echinoderm have radically different embryological origins (David and Mooi, 1996). Metamorphosis is uniquely pronounced in echinoids because the developing larva must discard or resorb so much of the non-rudiment larval body. Other echinoderms, such as the crinoids, retain extraxial skeleton in the adult, and do not reduce the relative contribution of the larval body (fig.3). Another independent data set based on analysis of homeobox genes (Lowe and Wray, 1997) shows that certain genes are expressed only in one type of skeleton, and not in the other, further illustrating the validity of the subdivisions inherent in the EAT.
B) The bizarre morphology of forms such as stylophorans can be better understood through application of the EAT. The pentaradial model, in emphasizing the lack of synapomorphies with other echinoderms, forces the stylophorans down the phylogenetic tree. However, the EAT shows that the aulacophore arises from the confluence of extraxial and axial components to form an arm like those found in crinoids (fig.4). Examination of teratological specimens, such as pentamerous echinoids or specimens in which aberrant plate morphologies have obscured "normal" configurations constitute an additional test of the EAT. These forms are found to make perfect sense when reinterpreted according the EAT and the OPR
Implications of the EAT
With the establishment of a new way of looking at echinoderm homologies, it has become possible to reconsider certain major traits in the evolution of the echinoderms.
A) Interambulacra - Previous works, relying on the pentaradial model, almost universally describe echinoids as being made up of 5 ambulacra alternating with 5 interambulacra. The interradial regions of other forms such as asteroids and edrioasteroids are also referred to as "interambulacra". However, the OPR orerates in the production of the interambulacra of echinoids, which are therefore axial in origin. In all other forms, interradial regions are made of extraxial skeleton and we can no longer consider interambulacra of echinoids to be homologous with the interradial regions of other echinoderms (fig.5).
B) Marginal rings - Some authors have suggested homologies between the ring of plates along the perimeter of some asteroids, and those found in certain edrioasteroids and edrioasteroid-like forms such as Stromatocystites (fig.6). The similarity of these structures has at least in part been the basis for derivation of asteroids from edrioasteroids (Smith 1990 ; Smith and Jell, 1990). The EAT reveals that the marginal ring of forms such as Stromatocystites marks the boundary between imperforate and perforate extraxial skeletons, and that the ring itself is made of imperforate extraxial elements. In asteroids, which lack imperforate extraxial skeleton, the marginal ring is embedded deep within the perforate extraxial region, and might even represent highly modified axial elements that have separated from the main growth zone of the axial region.
C) Arms and brachioles - Echinoderms of several clades have arm-like extensions, usually used as feeding structures and variously known as arms, brachioldes, free ambulacra, etc. This diversity of structure has suggested a wide variety of homology shemes and phylogenetic arrengements (Fell, 1963 ; 1963 ; Paul and Smith, 1984 ; Smith, 1984a ; 1990, Sumrall, 1997). Fell's (1963, 1965) hypothesis suggested that there were two subphyla, the Crinozoa and the Asterozoa in spite of the fact that he saw all appendages, whether brachioles or arms, as the product of a single evolutionary event (fig.7a). Smith (1984a) derived arms separately in the asteroids and the crinoids, and brachioles were represented by yet another event in the history of the cystoids, blastoids, and lepidocystoids (fig.7b). The EAT suggest that there are actually two major types of body extensions : "true" arms, and axial arms. The former is made of the confluence of axial and extraxial elements, which are arranged to support both the hydrocoel and the somatocoels (fig.8a). True arms are characteristic of asteroids, ophiuroids, and crinoids, and are probably homologous in all these groups. Even the pinnules of crinoids exhibit the construction typical of true arms. As suggested above, the aulacophore of stylophorans are also homologous with this type of arm (fig.4). In contrast, axial arms are made only of axial skeleton supporting the hydrocoele (radial canal), and lack extraxial elements or somatocoel extensions. They are instead constructed of coverplate-bearing flooring plates that extend outward from the body. The biserial arrrangement of these two plate series is characteristic of brachioles seen in lepidocystoids, and eocrinoids (fig.8b). Axial arms can take on a variety of configurations in addition to that of the earliest eocrinoids. The main trunk of the axial system can branch and become free to form brachioles in addition to the terminal brachiole (fig.8c). This branching can become very crowded and form complicated sets of brachioles mounted directly on top of the thecal plates as in some rhombiferan cystoids (fig.8d). Blastoids have not only greatly 'formalized" the arrangement of brachioles upon the theca, but have added complicated hydrospire cavities below the flooring plates that give rise to the brachioles (fig.8e).
D) Phylogeny - Detailed character analysis of the type presented above for interambulacra, marginal rings, and arms is yielding a picture of echinoderm relationships (fig.9). The phylogeny presented here is based on a large number of features, but we have focused only on those characters that play a prominent role in the present paper. The most plesiomorphic morphologies are to be found in taxa such as Camptostroma and Stromatocystites. Edrioasteroids form a clade of their own, but well down the tree relative to previous analyses that consider them to be more derived forms (Smith, 1984a ; 1988 ; Sumrall, 1997). On the other hand, we consider the homalozoans, and stylophorans in particular to be highly derived forms. Although many questions remain concerning the poorly preservec Arkarua of the Precambrian, its general morphology is not inconsistent with our view of the plesiomorphic echinoderms (fig.9). Our interpretation is also slightly more consistent with the stratigraphic data concerning edrioasteroids of the lower Cambrian (Sprinkler and Guensburg, 1997). The EAT's proposed homology of the arms of crinoids and asteroids and the separation of the former from the cystoid-like taxa implies that the older grouping of stemmed Pelmatozoa is paraphyletic and should be abandonned. This is again more consistent with the stratigraphic record, since crinoids and asteroids appear later than the eocrinoids as the base of the cystoid clade. Eothuria, once placed within the echinoids, is here considered to be below the node that includes other echinoids, holothurians and ophiocysioids. The construction of the test solely of axial plates is an apomorphy for the echinoids, since in Eothuria-like taxa, ophiocistioids, and holothurians, the "interambulacra" are constructed of perforate extraxial skeleton. One very important consequence of the EAT is that pentamery in intimately associated with the development of axial skeleton. Because the story of echinoderm evolution is generally on of reduction of extraxial skeleton, and increasing dominance of axial skeleton, pentamery is less prominently expressed in the very oldest taxa than in more derived forms. Pentamerism is actually superimposed on a basically linear arrangement of axial - perforate extraxial - imperforate extraxial skeletons. The classical references to pentaradial morphology as exemplified by asteroids and echinoids obscure an essential linearity expressed in both skeletal topology and coelomic arrangements that is more comparable with that seen in other phyla.