The discovery of embryos in certain fossil fishes not only shows that internal fertilization and live birth evolved early in vertebrate history, but also raises questions about the origin of jawed vertebrates.
Every once in a while, a discovery comes along that puts our biological understanding of some extinct group of organisms on a much firmer footing. On page 1124 of this issue, Long and colleagues1 present such a discovery, one that may prove to have far-reaching implications for our understanding of early vertebrate evolution. For the second time in less than a year2, they have found preserved embryos in the body cavity of a placoderm fish.
Placoderms are extinct jawed fishes that lived from approximately 430 million years ago to the end of the Devonian 360 million years ago. Their heads and shoulder girdles are covered in bony armour; plates of bone also form the biting surfaces of the jaws, sometimes flat and covered with small bumps, sometimes developed into nasty, self-sharpening scissor blades. The largest of these fishes equalled the size of a great white shark and must have been formidable predators, but most were less than a metre in length.
Working with three-dimensional placoderm fossils from Gogo in Western Australia, Long and colleagues have discovered specimens of three different placoderms, Materpiscis2, Austroptyctodus2 and Incisoscutum1, that contain minute but perfectly preserved armour plates of the same species in their body cavities. When first discovered, they were thought to be stomach contents3. But the plates show no bite marks or etching by stomach acids, and are not mixed with bones from other species; they are the remains of unborn embryos. In Materpiscis, a curving tubular structure associated with one of them has been interpreted as an umbilical cord2.
Embryos in the body cavity imply internal fertilization. It was noted long ago4 that ptyctodonts, the placoderm subgroup to which Materpiscis and Austroptyctodus belong, have sexually dimorphic pelvic fins, somewhat like the 'claspers' used for internal fertilization in sharks. Arthrodires, the placoderm group that includes Incisoscutum, lack the sexually dimorphic external bones present on the pelvic fins of ptyctodonts. However, Long et al. argue that the partially preserved internal fin skeletons of their specimens indicate a shark-like structure, probably implying sexual dimorphism and internal fertilization. Ptyctodonts and arthrodires seem to be closely related, and so internal fertilization, and possibly live birth of young, are probably shared features retained from their common ancestor.
The living jawed vertebrates, or gnathostomes, fall into two groups, the Chondrichthyes and the Osteichthyes (Fig. 1). The Chondrichthyes (sharks, rays and ratfishes) all have internal fertilization, and many give birth to live young, whereas the ancestral condition for the Osteichthyes (ray-finned fishes, lobe-finned fishes and land vertebrates) is to spawn small eggs that are fertilized externally. Live-bearers tend to produce much fewer young than external spawners and have lower potential rates of population growth. This contrast in reproduction puts a new perspective on the ecology of the Gogo environment, a tropical reef5, where a wide diversity of placoderms coexisted with lungfishes and primitive ray-finned fishes that were probably externally fertilizing spawners. It is also interesting to note that the extinction of the placoderms at the end of the Devonian was followed by a major diversification of chondrichthyans. But it is to the study of gnathostome interrelationships that the discoveries of Long et al. may prove to be most pertinent.
Ideas about the origin of gnathostomes are currently in a state of flux. For much of the twentieth century, placoderms were regarded as relatives or possibly ancestors of chondrichthyans4, partly because they seemed to use internal fertilization. But recently the majority view has placed them in the gnathostome stem group6 — that is, the common ancestral lineage of the living jawed vertebrates. A new analysis by Brazeau7 suggests that placoderms may not be a natural group at all, but a 'paraphyletic array' spread out along the gnathostome stem (Fig. 1a; contrast with Fig. 1b). If that is correct, placoderms become extremely informative about the origin of jawed vertebrate morphology. This is where the evidence for internal fertilization and live-bearing in placoderms becomes important.
The ancestral mode of reproduction for osteichthyans seems to be external fertilization. The distribution of live-bearing among living vertebrates strongly suggests that internally fertilizing live-bearers are unlikely to give rise to externally fertilizing spawners, so we would not expect the osteichthyan stem lineage, or the gnathostome stem lineage below it, to contain a segment characterized by live-bearing. Brazeau's analysis7 places the ptyctodonts and arthrodires as successive branches off the gnathostome stem, implying the existence of such a segment unless the two groups have evolved live-bearing independently. However, only a minor change in the tree would be needed to join ptyctodonts and arthrodires together in a clade (that is, forming a single side branch), and thus make the offending stem segment disappear. A more important question is whether the most primitive placoderms, such as the antiarchs (bottom-feeding fishes with armoured pectoral fins), were also live-bearers, because this would undermine the case for the placoderms forming a paraphyletic segment of the gnathostome stem.
Long and colleagues1 argue that the antiarchs had external fertilization. They lack pelvic fins altogether, and fossils have been found of free-living juveniles that are small and undeveloped enough to correspond to the embryos of Materpiscis, Austroptyctodus and Incisoscutum. It may thus be that both internal fertilization and live-bearing evolved within the placoderms. Perhaps this was a unique innovation in one placoderm clade. Alternatively, could some placoderms be stem gnathostomes and others, those with internal fertilization, stem chondrichthyans? Possibly, but this conflicts with new evidence that the acanthodians (vaguely shark-like fishes, with fin spines and tiny scales, which became extinct about 250 million years ago) form a paraphyletic array encompassing the bases of the chondrichthyan and osteichthyan lineages (Fig. 1a)6.
The tangled skein of jawed-vertebrate origins continues to challenge researchers. But discoveries such as the placoderm embryos of Gogo are giving us the tools to gradually untangle it — as well as showing us intimate glimpses of life in a lost world.
Long, J. A., Trinajstic, K. & Johanson, Z. Nature 457, 1124–1127 (2009).
Long, J. A., Trinajstic, K., Young, G. C. & Senden, T. Nature 453, 650–653 (2008).
Dennis, K. & Miles, R. S. Zool. J. Linn. Soc. 73, 213–258 (1981).
Moy-Thomas, J. A. & Miles, R. S. Palaeozoic Fishes (Chapman & Hall, 1971).
Long, J. A. Swimming in Stone: The Amazing Gogo Fossils of the Kimberley (Fremantle, 2008).
Janvier, P. Early Vertebrates (Oxford Sci. Publ., 1996).
Brazeau, M. D. Nature 457, 305–308 (2009).
About this article
Zoological Journal of the Linnean Society (2014)
Nature Communications (2013)
99th Dahlem Conference on Infection, Inflammation and Chronic Inflammatory Disorders: Evolution of adaptive immunity in vertebrates
Clinical & Experimental Immunology (2010)
Molecular Biology (2009)