Brychaetus

Abstract: The geological, faunal and palaeoecological conditions of the marine deposits from lowermost Eocene in North Jutland are briefly reviewed as background for the descriptions of six species of osteoglossiform fishes from the Stolle Klint Clay and the overlying Mo-clay (Olst and Fur Formations respectively). Four of these primitive teleosteans are referred to new genera and species (one based on an almost complete skeleton, three others on skull material, one very incomplete), and the two most fragmentary specimens are referred to Brychaetus sp. and an indeterminate osteoglossiform. The phylogenetic relationships of these fossils are evaluated in the framework of two different models of osteoglossomorph phylogeny provided earlier by Taverne and Hilton. Despite differences in both data bases, methodologies and results given by the two models (the former based on c. 300 characters in an intuitive, qualitative phylogenetic parsimony analysis, the latter on 72 characters in a critical, rigorous, quantitative cladistic analysis) the phylogenetic positions of four of the fossil species are very similar in the two models concerning the relations to the recent forms. The other two species, rather fragmentary, but similar in many ways to the Eocene phareodonts (paraphyletic group), end up very differently in relation to extant forms in the two models. The phylogenetic systematics of all the marine, fossil osteoglossiforms (including Brychaetus, Opsithrissops and Monte Bolca forms) is evaluated as background for interpretation of their (palaeo-)biogeographic significance as marine members of a group, Osteoglossomorpha (of which the recent forms are prime examples of ‘primary division freshwater fishes’, and of which the extant osteoglossiforms have a classical ‘Gondwana distribution’. There are 9 marine, Eocene taxa (plus an otolith from the Maastrichtian of USA) and none of the 9 appear more closely-related to any other marine form in either model: they might constitute 9 separate migrations from freshwater into the sea. The phylogenetic results strongly suggest instead, that the extant osteoglossiforms have independently entered freshwater from the sea on two, perhaps even three occasions. This may have happened as late as the Eocene, and phareodonts could be yet another independent invasion of freshwater in the Late Cretaceous. The mormyriforms most likely had an independent invasion into freshwater (in one model even with notopterids as a separate migration from the sea by Mid or Early Cretaceous). Because all the closest outgroups of the Osteoglossomorpha are marine, the group obviously originated in the sea, probably by the Late Jurassic, and it is not impossible that Hiodontiforms in NE Asia and North America underwent another independent freshwater invasion very early in the Cretaceous. What then is wanting? The expected Cretaceous, marine osteoglossomorphs are not found (but note the above otolith). Geology, ‘ash-series’ and faunas This ‘ash-series’ comprises the main series of ash falls of the Palaeogene of NW Europe. The Palaeocene–Eocene ash falls were recognized as three distinctive ‘phases’ (Knox 1996, 1997; Knox & Morton 1988, 1990) of which the second one constitutes the main phase (phase 2.1–2.2b in Knox 1996, 1997 correction sheet p. 10x), which was deposited in the Danish Stolle Klint Clay (of the Olst Clay facies, Heilmann-Clausen 1996) and the overlying Mo-clay (Fur Formation proper, Pedersen & Surlyk 1983 as modified by Heilmann-Clausen et al. 1985; Bonde 1997; Heilmann-Clausen 2006). These two deposits are well exposed only in NW Jutland round the western Limfjorden region, but tiny exposures of ‘Mo-clay like’ layers with ash can be found in a few places further south in East Jutland, and Stolle Klint Clay seems to correspond to the basal part of the Olst Clay Formation exposed south of Limfjorden and in East Jutland (Andersen 1937; Nielsen 1994), NW Sealand (Petersen 1973) and also to the lower part of the Sele Formation in the North Sea (Schioler et al. 2007). Prinz & von Ermengen (1883) described and figured the so-called ‘black sand’ from limestone boulders of the Mo-clay as volcanic ash (tuffs), From: CAVIN, L., LONGBOTTOM, A. & RICHTER, M. (eds) Fishes and the Break-up of Pangaea. Geological Society, London, Special Publications, 295, 253–310. DOI: 10.1144/SP295.14 0305-8719/07/$15.00 # The Geological Society of London 2008. which, however was only recognized in Denmark by 1902. Professor N.V. Ussing immediately sent the young O.B. Boggild to all known localities with such layers in Jutland (Boggild 1903). Ussing also numbered the upper part of the ash series from þ1 to þ140 in a footnote to his major work on end moraines in Jutland (1907). After Ussing’s death in 1910, Boggild took over both the professorship and the investigation of the ash layers, which, after some delay, was published as his famous monograph in 1918. In this he continued the numbering of the ash layers downwards from 21 to 239 in the lower half of the Mo-clay (in which the ash layers are much less frequent). In those days many of the now-conspicuous sea cliffs were covered by scree and grass, due to less coastal erosion in the Limfjorden, and the industrial exploitation of the Mo-clay in the many clay pits had hardly begun early in the century, the initiation being the founding of ‘Skarrehage Molervaerk’ (now ‘Skamol’) by 1912 on the Isle of Mors (see Bonde 1987 at the 75th anniversary of Skamol; Pedersen et al. 1994). It is estimated from measurements made on Isle of Fur in the 1920s of the horizontal mine shaft into the Stolle Klint (1⁄4 cliff; originally dug about 1800 in search for coal, the black volcanic ashes) that, since then, at least 50 m of the seaward face of the cliff has eroded away. The Knudeklint, now type locality of Fur Formation (Pedersen & Surlyk 1983), was early in the 20th century mostly covered by grass and had a wide low foreland protecting part of the cliff from the sea. Stolle Klint Clay (bottom of Olst Fm) Stolle Klint Clay (1⁄4 SKC; which was informally named by Heilmann-Clausen 1996) is a blackish-greyish, laminated mudstone with a few thin, hardened and silicified beds, and it contains a few of the earliest and numbered ash layers 234 to 239 (Boggild 1918; Pedersen & Surlyk 1983; Heilmann-Clausen et al. 1985; Nielsen 1994). The 10–15 cm thick and whitish ash layer 233 is at the upper boundary of SKC. The fossils, plant debris, fishes, shrimps and insects are found mainly in the hardened layers, locally called ‘shale’ (Danish: ‘skifer’; and therefore likely to be confused with the silicified layers in the Mo-clay originally called ‘skifer’ by Gry (1940) from around the level of ash layer 220, low in the Fur Formation; see Bonde 1997; Andersen & Sjorring 1997). The SKC contains few, if any, diatoms (Homann 1991). The fish fauna in North Jutland is clearly from open marine water (Bonde 1997), but Kothe (1990) indicates that this basal part of the ash series, at least in the North German region, seems to show a brackish influence as based upon the dinoflagellates. No doubt the North Sea basin at this time (corresponding to the lower part of Sele Formation in the central North Sea, see Heilmann-Clausen 1985; Schioler et al. 2007) was very restricted and nearly or completely landlocked by the ‘Thulean landbridge’ (in the SW from Greenland via the Faroes to Scotland) combined with the closed ‘Channel’ and by the Spitzbergen-Barents shelf (in the north) in the narrow region between Greenland and the ‘Scandinavian Continent’ (Fig. 2; Bonde 1979, 1997; Schmitz et al. 1996; Ziegler 1990). This closure of the basin may be due to a combination of low global sea level and upheaval of the Thulean land-bridge area above the ‘Icelandic hot spot’ (Heilmann-Clausen 2006). Some indicate that the entire rift zone along the East Greenlandic coast and the Mohn Ridge may have been about one km above sea level (Larsen 1988) before the Fig. 1. Localities of the Mo-clay region (modified from Bonde 1997). N. BONDE 254

Abstract: The marine osteoglossid "Brychaetus" caheni from the Lower Paleocene of Cabinda (Africa) is re-studied. It does not belong to the genus Brychaetus because of differences in teeth and premaxilla shape. It is assigned to the new genus Ridewoodichthys on the basis of its caudal skeleton, which differs from that of all other known fossil and Recent Osteoglossidae.

Abstract: The author describes the snout osteology of the fossil Osteoglossidae Phareodus and Brychaetus. He shows that the two genera considerably differ in their naso-ethmoid region. Indeed, Phareodus encaustus and P queenslandicus possess a large dermethmoid contacting the frontals and separating the two nasals from each other, which is the primitive condition within teleosts. P. testis is a little more specialized. Its dermethmoid still separates partially the two nasals but it begins to lose its contact with the frontals. On the contrary, Brychaetus muelleri exhibits a specialized snout pattern. The two nasals, articulated with the frontals, are joined together on almost their entire length, except at their anterior edge where a small dermethmoid, largely separated from the frontals, is inserted between them, as in Osteoglossum and Scleropages. Thus, Brychaetus is a valid genus and not a synonym of Phareodus. The author also shows that the osteoglossid Musperia radiata, from the Eocene of Indonesia, displays the same advanced snout pattern as Brychaetus and the two Recent Osteoglossidae. The caudal skeleton of Musperia is described for the first time. The first preural and the first ural centra bear complete neural spines. The five hypurals are autogenous. There is a pair of uroneurals and no epural.