Ibis (2003), 145, 382–391
B l a c k w e l l P u b l i s h i n g L t d . A n e w E o c e n e s w i f t - l i k e b i r d w i t h a p e c u l i a r f e a t h e r i n g
GERALD MAYR* Forschungsinstitut Senckenberg, Division of Ornithology, Senckenberganlage 25, D-60325 Frankfurt/M., Germany
A new taxon of swift-like birds is described from the Middle Eocene of Messel (Germany). It is tentatively assigned to the extinct family Jungornithidae and exhibits a completely unexpected feathering, which contrasts sharply with that of recent swifts. The short and rounded wings clearly show that it was not adapted to gliding, but might have caught its prey by sallying flights from a perch.The tail of the new taxon is very long and the tail feath- ers are broad and nearly symmetrical.The phylogenetic relationships between the Jungorn- ithidae and other apodiform birds are still not convincingly resolved. The early Oligocene genus Jungornis itself shares unique derived characters with hummingbirds which are, how- ever, absent in the Eocene genus Argornis and in the new taxon from Messel.
Swifts are classified into two recent families, the Australasian Hemiprocnidae (tree swifts) and the Apodidae (true swifts), which have a worldwide distribution. Both families differ in a number of osteological features, and especially concerning the structure of the wing skeleton, the Hemiprocnidae exhibit a less derived morphology than the Apodidae.
Most recent authors consider the Trochilidae (hummingbirds) to be the closest extant relatives of swifts and both taxa are usually united in a single order Apodiformes (e.g. del Hoyo et al. 1999). Monophyly of swifts and hummingbirds is not only supported by derived morphological characters, but also by biochemical and molecular analyses (Kitto & Wilson 1966; Cracraft 1981; Sibley & Ahlquist 1990; Johansson et al. 2001; Mayr 2002).
A recent phylogenetic analysis by Mayr (2002) provided strong evidence that theAegothelidae (owlet- nightjars) are the sister taxon of the Apodiformes.
THE EARLY TERTIARY FOSSIL RECORD OF APODIFORM BIRDS
Harrison (1984) classified E. vincenti in its own family, but Karhu (1988) even doubted its correct assignment to the Apodiformes.
The exact systematic position of the extinct Aegialornithidae Lydekker 1891 has also not yet been convincingly resolved. The family is generally considered to be closely related to the Hemiprocnidae (e.g. Harrison 1984; Karhu 1988, 1992), but this assign- ment is mainly based on – probably plesiomorphic – overall similarity. Members of the Aegialornithidae are especially abundant in the Upper Eocene fissure fillings of the Quercy, France (Mourer-Chauviré 1988), but Peters (1998) recently described a record of Aegialornis from Middle Eocene deposits of the Geiseltal, Germany. Olson (1999) tentatively assigned a tarsometatarsus from the Lower Eocene of North America to the family.
Most authors also classify Primapus Harrison & Walker 1975 from the Lower Eocene of England into the Aegialornithidae (e.g. Harrison & Walker 1975; Peters 1985; Karhu 1988; Mourer-Chauviré 1988). Harrison (1984), however, included Primapus in the Apodidae.
So far, no pre-Quaternary fossils of hummingbirds are known. However, swifts have a fairly extensive early Tertiary record, and apart from representatives of both recent families, three fossil families of swift- like birds have been identified.
The genus Cypselavus Gaillard 1908 is known from the Upper Eocene and Lower Oligocene of the Quercy, and is generally assigned to the Hemi- procnidae (Harrison 1984; Peters 1985; Mourer- Chauviré 1988).
Eocypselus vincenti Harrison 1984 from the Lower Eocene of England represents a tiny swift-like bird that is known from a few skeletal elements only.
The earliest certain representatives of the Apodidae belong to the genus Scaniacypselus Harrison 1984 which includes two species. Scaniacypselus wardi was described by Harrison (1984) from the early Eocene of Denmark and is known from isolated
© 2003 British Ornithologists’ Union