• DGF OnLine Series

Eric Buffetaut, CNRS, Laboratoire de Paléontologie des Vertébrés, Case 106, Université Paris 6, 4 place Jussieu, 75252 Paris Cedex 05, France.

Remains of the giant bird Gastornis were first discovered near Paris in 1855. Later finds near Reims resulted in a skeletal reconstruction by V. Lemoine which was largely erroneous, but was widely accepted, and hindered the recognition of close resemblances with the North American form Diatryma. Recent revisions have shown that the two forms are closely related, and it is suggested that they may even be congeneric. The biogeographical history and the possible origin of the Gastornithidae are discussed.

Historical introduction

Remains of the giant bird Gastornis, found by the physicist Gaston Planté (who later invented the lead battery still used in cars) were first reported from the Early Eocene “Conglomérat de Meudon”, near Paris, in 1855 (Prévost 1855; Hébert 1855). This find attracted much attention in France and abroad (Owen 1856). Gastornis parisiensis was not only a giant bird, it was also one of the oldest known birds at that time (Archaeopteryx had not yet been described). However, the scantiness of the available material made it difficult to discover its zoological affinities. Both Hébert (1855) and Milne-Edwards (1867) placed it close to the ducks, but Lartet (1855) thought it was closer to the waders, and Valenciennes (1855) thought he could detect resemblances with the albatross. A very different opinion was expressed by Bonaparte (1856), who thought it was a flightless bird related to Aepyornis. Despite these uncertainties, Gastornis was to become a well-known fossil, which was mentioned in many scientific and popular books and articles (see, for instance, Meunier, 1882).
It became even more famous when Gastornis remains were discovered by Victor Lemoine, a physician from Reims, among the Late Palaeocene vertebrate assemblage from Cernay, near Reims, which was at that time the oldest vertebrate fauna from the European Tertiary. Lemoine (1878, 1881) described various new elements of Gastornis, including what he thought were bones of the skull and jaws. Unlike the condition in other birds, the skull bones were unfused. The jaws bore tooth-like excrescences. Lemoine was clearly influenced by descriptions of Archaeopteryx, and, more importantly, by the recently published work of Marsh (1880) on the toothed birds Hesperornis and Ichthyornis from the Upper Cretaceous of Kansas, as well as by Huley’s ideas on the relationships between birds and reptiles. He interpreted Gastornis
Lemoine (1881) published a skeletal reconstruction of GastornisGastornis bones were subsequently reported from Belgium (Dollo, 1883) and England (Newton, 1885), they did not bring significant changes to Lemoine’s reconstruction. in evolutionary terms as a primitive bird still exhibiting many reptilian features. which showed a 2.70 metres tall, rather slenderly built bird, with a long, low toothed beak. This reconstruction was largely accepted as correct, and was widely reproduced in text-books and popular books alike. Although more

Fig.1. Lemoine’s skeletal reconstruction of Gastornis (a), published in 1881, was largely based on non-avian material (especially in the skull), but it was generally accepted and widely reproduced (here, in a popular paper by Meunier, 1882). This erroneous reconstruction, which differed conspicuously from the reconstruction of DiatrymaGastornis and Diatryma. published by Matthew and Granger in 1917 (b), was for a long time a serious obstacle to the recognition of the close similarities between

When the giant bird Diatryma was first described from the Lower Eocene of New Mexico by Cope (1876), resemblances with Gastornis were suggested by various authors, but the available material of Diatryma remained for many years so scanty that significant comparisons with the European form were difficult. When Matthew and Granger (1917) described a fairly complete Diatryma skeleton from the Lower Eocene of Wyoming, they produced a skeletal reconstruction which differed significantly from Lemoine’s reconstruction of Gastornis. They had some doubts about the validity of Lemoine’s reconstruction, however, and urged European palaeontologists to revise the French material. As this was not done, uncertainty continued to prevail about the relationships between Gastornis and Diatryma. Interestingly, after the publication of the description of DiatrymaDiatryma rather than to Gastornis (Switzerland: Schaub 1928; France: Gaillard 1939; Germany: Fischer 1962, 1978; Berg 1965), with the exception of Weigelt’s mention (1939) of Gastornis from the Palaeocene of Walbeck. by Matthew and Granger, most new finds of giant birds from the Palaeogene of Europe were referred to

Recent developments

In 1992, Martin published a revision of Gastornis, based on both old and newly discovered material, in which he showed that Lemoine’s skeletal reconstruction was largely based on non-avian material. In particular, there was no factual basis at all for the “reptilian” features of the skull as erroneously reconstructed by Lemoine. One of Martin’s conclusions was that there were many resemblances between Gastornisand Diatryma, which could both be placed in the family Gastornithidae, and that “the identification of Diatryma in Europe should be approached with caution” (Martin 1992, p.107).
Andors’s detailed revision of Diatryma (Andors 1988, 1992) led him to similar conclusions, although, pending a revision of the European material, he preferred to keep Diatryma and Gastornis in separate families of the order Gastornithiformes.
It is now clear that lack of recognition of the close similarities between Gastornis and Diatryma was largely a result of the uncritical acceptance of Lemoine’s erroneous reconstruction, which differed conspicuously from the much more reliable reconstruction of Diatryma by Matthew and Granger, and hindered a proper appreciation of the real affinities of Gastornis.

Are Gastornis and Diatryma congeneric ?

In view of the close similarities between Gastornis and Diatryma revealed by the recent revisions by Martin (1992) and Andors (1992), the question of a possible generic identity can reasonably be asked, all the more so that it would make sense from a palaeobiogeographical point of view. The study of unpublished Gastornis material in the palaeontology collection of the University of Reims (collected in the 1950s by an amateur palaeontologist, Mrs Lasseron) confirms the resemblances with Diatryma, including in parts of the skeleton which were hitherto unknown in Gastornis, such as the proximal end ot the tibiotarsus (Buffetaut, in prep.). Although generic identity seems likely – in which caseGastornis would clearly have priority – it can be proved (or disproved) only after a detailed comparative study of the available material from Europe and North America.

The biogeographical history of the Gastornithidae

From a biogeographical point of view, the generic identity of Gastornis and Diatryma would not be unexpected, since there are close faunal similarities between the Early Tertiary vertebrate faunas of Europe and North America, especially in the Early Eocene (Russell, 1975). WhetherGastornis and Diatryma turn out to be congeneric or not, the stratigraphic record of gastornithids, if it accurately reflects the temporal distribution of these birds in Europe and North America, may have some interesting implications for their biogeographical history. In North America, Diatryma is restricted to the Early Eocene (Wasatchian; Andors 1992). In Europe, gastornithids (including both Gastornis and the specimens referred to Diatryma) have a longer record, from the beginning of the Late Palaeocene (with the Walbeck specimens) to the Middle Eocene. This suggests that the family Gastornithidae may have originated in Europe, and that it reached North America during the important episode of faunal interchange between Europe and North America at the beginning of the Eocene. As suggested by the Diatryma remains from the Lower Eocene of Ellesmere Island (West & Dawson 1978), an Arctic route is the likeliest one for the dispersal of those giant birds.
The only evidence of gastornithids outside Europe and North America is the tibiotarsus fragment from the Lower Eocene of Honan, in China, described by Hou (1980) as Zhongyuanus xichuanensis. This specimen apparently indicates a wide Eurasian distribution of the family during the Early Eocene.

The problem of the origin of the Gastornithidae

There is no real consensus concerning the exact place of the Gastornithidae among birds. Widely different opinions have been offered since the first discovery of Gastornis in 1855. According to Martin (1992), the Gastornithiformes, considered as predators, were “very likely an early experiment in the evolution of large terrestrial birds”, without a special relationship with any modern avian group. Andors (1992) holds a different opinion and thinks that the Gastornithiformes, which he interprets as folivorous birds, have affinities with the Anseriformes. Be that as it may, the origin of the Gastornithidae is still unclear. As mentioned above, the earliest known representatives of the family are from the Upper Palaeocene of Europe, and they do not seem to differ markedly from the later, Eocene forms. Gastornithids have often been considered as an early “attempt” by birds to occupy a niche left vacant by the extinction of the dinosaurs, before the appearance of large mammals. However, evidence from the Upper Cretaceous of Europe now indicates that large birds were already present several million years before the end of the Cretaceous (Buffetaut et al. 1995). Newly discovered material (including a partial pelvis) from southern France suggests that these birds were in all likelihood terrestrial, and some resemblances with gastornithids are apparent. A possible origin of gastornithids among those newly discovered Late Cretaceous forms cannot be excluded, but only more detailed comparisons (which are in progress) will confirm or falsify this hypothesis.