JOURNAL OF MORPHOLOGY 270:1364–1380 (2009) Skeletal Histology of Bothriolepis canadensis (Placodermi, Antiarchi) and Evolution of the Skeleton at the Origin of Jawed Vertebrates Jason P. Downs1,2* and Philip C.J. Donoghue3 1Department of Geology and Geophysics, Yale University, New Haven, Connecticut 06511 2Academy of Natural Sciences of Philadelphia, Philadelphia, PA 19103 3Department of Earth Sciences, University of Bristol, Bristol BS8 1RJ, UK ABSTRACT We used light microscopy and scanning been especially true in debate over the primacy of electron microscopy to compile a complete histological bone versus cartilage, and the developmental evo- description of the dermal skeleton of the antiarch placo- lution of the dermal skeleton. Though they may be derm, Bothriolepis canadensis. Placodermi is most often among the oldest extant lineages of skeletonizing cited as the sister group of crown group Gnathostomata, vertebrates, it has long been recognized that chon- but some recent authors propose that placoderms instead represent a paraphyly of forms leading to the drichthyans are far removed, both temporally and crown. In either phylogenetic scenario, comparative phylogenetically, from those extinct vertebrates in analysis of placoderm and gnathostome histological data which a mineralized skeleton first evolved and in allows us to address the primitive condition of both the which skeletal developmental systems were estab- gnathostome skeleton and the jawed vertebrate skeleton. lished (Heintz, 1929; Romer, 1942). This recogni- The results of this work support the interpretation that tion stems from the discovery that the mineralized the external skeleton of Bothriolepis canadensis is com- skeleton first arose in extinct jawless relatives of prised exclusively of cellular dermal bone tissue. The the jawed vertebrates, to whom they are related unique stratification of the antiarch thoracic skeleton by degree (Janvier, 1981). The study of these that has led to controversial interpretations in the past extinct stem gnathostomes (see Fig. 1 for meaning is explained by the nature of the articulations between adjacent elements. Skeletal features long thought to be of taxonomic concepts) has revealed that peculiar- gnathostome innovations are instead discovered to arise ities, such as the chimeric embryological composi- along the gnathostome stem. These innovations include tion of the vertebrate skeleton, betray its piece- secondary osteons, the systematic reconstruction of the meal evolutionary assembly over a protracted epi- skeleton in response to growth, and unfused, overlap- sode of early vertebrate phylogeny (Donoghue and ping joints that enable marginal growth while maximiz- Sansom, 2002; Donoghue et al., 2006). Thus, the ing the area of the articulation surface. The extensive vertebrate skeleton may be considered more appro- evidence for spheritic mineralization agrees with a priately and accurately as a series of distinct skel- model of the skeleton as one capable of a high growth etal systems characterized by their distinct devel- rate and active remodeling. Dermal skeletal develop- opmental and evolutionary origins: the dermal ment in both placoderms and osteichthyans is primarily skeletogenetic with only a minor odontogenetic contribu- skeleton, neurocranium, splanchnocranium (viscer- tion in some taxa. This demonstrates the problem inher- ocranium), appendicular and axial skeletons ent with assuming a broad application for those hypoth- (Donoghue and Sansom, 2002). eses of dermal skeletal evolution that are based on a Considerable effort has been expended in under- chondrichthyan model. Our results highlight the impor- standing the early evolutionary assembly of the tance of anatomical and ontogenetic context in the inter- vertebrate skeleton, but we remain no closer to pretation of fossil tissues. J. Morphol. 270:1364–1380, 2009. Ó 2009 Wiley-Liss, Inc. Contract grant sponsors: National Science Foundation, Geological KEY WORDS: vertebrate; placoderm; histology; dermal Society of America, Paleontological Society, Yale Institute for Bio- skeleton; skeletogenesis; bone spheric Sciences (Field Ecology and ECOSAVE divisions). *Correspondence to: Jason P. Downs, Academy of Natural Scien- ces of Philadelphia, 1900 Benjamin Franklin Parkway, Philadel- phia, PA 19103. E-mail: [email protected] INTRODUCTION Chondrichthyans are generally perceived to be Received 18 December 2008; Revised 21 April 2009; the most basal clade of vertebrates with a mineral- Accepted 24 April 2009 ized skeleton and, thus, they have been influential Published online 16 June 2009 in in attempts to uncover the nature of the primitive Wiley InterScience (www.interscience.wiley.com) vertebrate skeleton (Donoghue, 2002). This has DOI: 10.1002/jmor.10765 Ó 2009 WILEY-LISS, INC. THE SKELETON OF Bothriolepis canadensis 1365 The aim of this study has been to remedy this situation, to completely characterize the skeletal histology of a placoderm, providing a benchmark for future studies that will seek to establish the di- versity of skeletal histology within the group. Our study focuses on the antiarch, Bothriolepis cana- densis (Whiteaves, 1880). This is not because it is representative of placoderms (though they may be representative of the very earliest jawed verte- brates; Johanson, 2002; Brazeau, 2009) but because Bothriolepis has proven previously to be the best source of histological data for placoderms. There is an abundance of specimens representing all stages of ontogeny, many of which are fully articulated and preserve histological microstruc- ture with great fidelity. Thus, B. canadensis repre- Fig. 1. Evolutionary relationships among the principal sents the best possible candidate for establishing groups of stem- and crown-gnathostomes, applied following Hennig (1981) and Jefferies (1979), effecting an explanation of skeletal composition in placoderms. the concepts of gnathostome stem and crown, and how they relate to the concept of a clade of jawed vertebrates. Trees differ only in terms of the relations of placoderms and acanthodians Previous Research Into the Skeletal and the topology of the tree otherwise follows Donoghue et al. Histology of Antiarch Placoderms (2000) and Donoghue and Smith (2001). (A) The traditional view of placoderm monophyly as promulgated by Goujet and The difficulty in obtaining ontogenetic data for Young (1995, 2004), Young (1986), and Janvier (1996). (B) The fossil taxa continues to generate controversial tis- hypothesis of placoderm paraphyly as promulgated by Brazeau sue identifications and developmental interpreta- (2009) and Johanson (2002). Icons of representative fishes after tions for Antiarchi. Past interpretations of the Janvier (1996). clade’s skeletal microstructure are marked by a lack of anatomical and developmental context. understanding the condition from which the skele- Goodrich (1909) provided a diagrammatic figure tons of chondrichthyans and osteichthyans of the gross structure of the dermal skeleton of departed. In large part, this occurs because so lit- Bothriolepis canadensis which he described as tle is known concerning the nature of the skeleton composed of true bone. He noted that, although in the most basal jawed vertebrates, the placo- the surface is ornamented by tubercles, there is no derms. Placodermi is an extinct clade or grade of evidence that they were formed from fused den- jawed vertebrates that first appears in the mid Si- ticles. Goodrich recognized a vascular middle layer lurian (Wenlock, circa 428 Ma) and goes on to and a typically lamellated basal layer and his fig- dominate the Devonian (417-354 Ma) vertebrate ure (p. 206) shows clear evidence of planar discon- fossil record (Carr, 1995). Though the systematic tinuity in the middle vascular layer. position and interrelationships of the group Heintz (1929) provided some of the earliest pub- remain points of contention, recent analyses place lished microscopic images of the antiarch external Placodermi as either the sister group of crown skeleton. The description, limited to a single sec- Gnathostomata (Fig. 1A; Young, 1986; Goujet and tion through an unidentified ‘‘trunk armor plate’’ Young, 1995; Janvier, 1996) or a paraphyletic (Pl. XXIV: Heintz, 1929), demonstrates the particu- grade leading to the crown (Fig. 1B; Johanson, lar stratified nature of the external skeleton in 2002; Brazeau, 2009). In either phylogenetic con- Asterolepis sp. Heintz (1929) designated four dif- text, placoderms are integral to understanding the ferent zones of tissue without sharp distinctions condition of the skeleton at the origin of jawed ver- among them, a superficial and a basal compact, la- tebrates and immediately before the emergence of mellar bone tissue and two unique middle cancel- chondrichthyans and osteichthyans. lous bone tissues. The superficial and basal cancel- There is no good explanation for the dearth of lous tissues were respectively designated the information on the composition of the skeleton in Maschen-schicht (mesh layer) and the Kanal- placoderms. Placoderms are well represented in schicht (channel layer). the fossil record, with over 700 recognized species Though he did not publish microscopic images, (Carr, 1995). However, it is perhaps because placo- Erik Stensio¨ did observe thin sections to interpret derms are so richly represented in the fossil record the skeletal tissues in Bothriolepis canadensis and and possess such complex skeletons that palaeon- Remigolepis sp. and to describe the internal fea- tologists have not resorted to skeletal