Early Devonian fishes from coastal De Long Strait, central Chukotka, Arctic Russia

ABSTRACT Calcareous and sandy deposits from the basal members of the Enmakaj and Pil'hikaj formations in coastal exposures along the De Long Strait in central Chukotka, Arctic far-eastern Russia, have yielded two assemblages of fossil fish comprising heterostracan plate fragments, turiniid and other thelodont scales, acanthodian scales and a partial tooth, typical of the Old Red Sandstone facies. Exceptional are acanthothoracid placoderm platelets characteristic of marine facies. In addition a sarcopterygian fragment have been found in Member 1 of the Enmakaj Formation. Some scale surfaces show an unusual, scoured preservation. A Lochkovian age, and most probably basal Lochkovian, is supported for the Enmakaj assemblage, and a somewhat later Lochkovian age is supposed for the Pil'hikaj assemblage. The palaeobiogeographic affinities of these assemblages based on the heterostracans and thelodonts are with other Lochkovian occurrences in Arctic regions such as Severnaya Zemlya, Spitsbergen and the northern and north-eastern Old Red Sandstone Continent in general.


INTRODUCTION
Limited and rarely described Devonian deposits occur in the far north-east of the Russian Arctic, in the northern part of the Chukchi Autonomous Okrug (district) on the southern coast of the De Long Strait (Proliv Longa in Russian), to the SSW of Wrangel Island (note that the name of the strait must be written "De Long Strait" because it has been named after the American explorer George Washington De Long; see Anon. 2010). e outcrop area is located between the mouth of the River Pegdymel' in the west and Cape Shmidt (Mys Shmidta in Russian, named after the Academician Otto Schmidt) to the east, i.e. between longitudes 174°E and c. 180°E (Fig. 1, upper map). Structurally it belongs to the northernmost part of the Kuyul tectonic division at the northern edge of the Chukotka Fold Belt (Bychkov & Gorodinsky 1992: fig. 1;Natal'in et al. 1999: figs 1, 7), which is a component of the Arctic Alaska -Chukotka Terrane (Till et al. 2010).
A review of earlier investigations of the studied area and a detailed description of the Devonian section were given by Rogozov & Vasil'eva (1968). ese authors first identified the occurrence of the Lower Devonian rocks in the area and subdivided the strata into formations based on lithology and fauna (Rogozov & Vasil'eva 1968: 152, fig.). e Devonian section Poberezh'e ('the Coast') is 4.5 km long, starting at the longitude of Cape Enmakaj (Enmykaj) and almost reaching the Pil'hikaj (Pilzykej) lagoon to the west. e Devonian rocks crop out along Tonnel'nyj Brook, and separate exposures also occur in the upstream part of the River Kuul' (Fig. 1, lower map). e Lower Devonian of the Coast section is subdivided into a lower, Enmakaj Formation and an upper, Pil'hikaj Formation, which are succeeded by the Middle Devonian Long Formation and the Upper Devonian Pegdymel' Formation. e latter formation is exposed in the basin of the River Pegdymel'.
A preliminary identification, including that of five different Early Devonian representatives of agnathans (heterostracans) and fishes which remained unpublished at that time, was given by the senior author (EMK) in 1976. e late Dr Svetlana Cherkesova (Institute of Arctic Geology, Leningrad now St Petersburg) gave information (pers. comm. to EMK 2002) on the stratigraphic range of the samples and the dating of the formations (Fig. 2), however, considering the age of the Pil'hikaj Formation as Pragian-Emsian, based on invertebrate data. She was in the field together with the team of Y. Rogozov and identified all the brachiopods listed in Rogozov & Vasil'eva (1968).
A full assemblage of fish specimens from two members (Fig. 2) will be described in this paper. Of the fish groups A. Blieck described Pteraspidomophi, S. Turner elodonti, C. J. Burrow Acanthodii, E. Mark-Kurik Placodermi and Sarcopterygii. Paleobiogeographic interpretation was given by AB, CJB and ST. GEOLOGICAL SETTING AND STRATIGRAPHY e general description of the Lower Devonian Enmakaj and Pil'hikaj formations comes from Rogozov & Vasil'eva (1968: 153-154; translated and slightly modified by EMK).
e Enmakaj Formation is 270 m thick with dominant sandstones. Characteristic for the formation are thin bedded sandstones, with an occurrence of cavernous leached sandstone with fossils and a thin macro-intercalation of sandstone and clayey shale. Such rock units reach a thickness of 10-15 m, with sandstone beds of 3-5 cm thick, and clayey shale layers of 1-1.5 cm. e formation is subdivided into three members: -the Lower Member (1) consists mainly of sandstone with thin layers of argillite and clayey shale.
In the upper part of the member, thin beds of fossil-bearing sandstone, siltstone, clayey shale and rarely limestone (as lenses or nodules) are intercalated. Fossils comprise brachiopods Atrypinella sp., Cyrtina sp., ostracodes, crinoids, plant and fish remains (described below); -the Middle Member (2) is a sandstone and shale unit with well-marked intercalation of thick and thin platy sandstone and clayey shale. is sandstone reveals ripple marks and trace fossils in a more clayey facies; -the Upper Member (3) is mainly a shale and limestone unit with siltstone interbeds. At the base of the unit there is a limestone with corals: Spongophyllum sp., Pseudoamplexus sp., Hexagonaria sp., Grypophyllum sp., Acanthophyllum cf. mansfieldense (Duncan, 1898), and brachiopods: Atrypinella ex gr. barba Khodalevich, 1939, "Camarotoechia" cf. haraganensis Amsden, 1958, Dubularia ex gr. thetis (Barrande, 1879, Carinatina ex gr. comata (Barrande, 1879), Cyrtina sp. (ex gr. dalmani Amsden, 1958) and Howellella sp. Leached sandstones, occurring in the upper part of the member, contain brachiopods: Parachonetes? sp., "Stegorhynchus" ex gr. nympha (Barrande, 1879), Dentatrypa sp. and Carinatina sp. Tentaculites, ostracodes, trilobites, bryozoans and other fossils have also been noted in all the above rocks. GEODIVERSITAS • 2013 • 35 (3) e boundary between the Enmakaj Formation and the overlying Pil'hikaj Formation may be defined in two different ways, either at the base of the siltstone-sandstone member overlying Member 3, or within the leached sandstone beds in the upper part of the Enmakaj Formation. e Pil'hikaj Formation conformably overlies the Enmakaj Formation. e Pil'hikaj Formation, with siltstones predominant, is 432 m thick. Characteristic for the formation is the occurrence of siltstone with calcite concretions containing fossils (fishes, etc.) and micro-intercalation of thin layers (a few millimeters in thickness) of sandstone, siltstone, clayey shale and limestone. In the formation three members are established: -the Lower Member (4) is a siltstone-sandstone unit with 20-30 cm thick, clayey shale interbeds and rare syngenetic calcite concretions. is unit often contains brachiopods of the family Pygnacidae, Carinatina ex gr. comata, Stegorhynchus ex gr.  (6) is a sandstone-siltstone unit with clayey shale interbeds. e stratified sandstone of this unit is well sorted.

MATERIAL AND METHODS
e fish specimens at our disposal are mostly tiny fragments or small isolated skeletal elements. ey came from two levels: -a) a grey calcareous siltstone sample from outcrop no. 73 of the Lower Member (1) of the Enmakaj Formation at the Tonnel'nyj (Tunnel'nyj) Brook, which yielded, after treatment with dilute acetic acid, several plate fragments (the largest one is 13 mm long) and acanthodian and thelodont scales. e fish remains are pyritized and etched, in several cases quite strongly; for instance, the thelodont and acanthodian scale crowns and even the bony bases were affected (Figs 7,9). e specimens with collection numbers of the Institute of Geology, Tallinn University of Technology (TUT, specimens GIT 580-1 to 27), were obtained from rock pieces of a siltstone sample taken in the field for the study of brachiopods and given to us by S. Cherkesova; -b) material of the Lower Member (4) of the Pil'hikaj Formation, from the Coast (Poberezh'e) locality, consists of a 7 cm long sample (no. 74-3) of grey, strongly cemented calcareous siltstone with fine cracks filled with calcite. On its surface the sample has an impression of heterostracan ornament. A bone fragment (about 5 cm long) with a double number 73-4 (74-3) from Tonnel'nyj Brook could come either from the Enmakaj Formation or, more probably, from the base of the Pil'hikaj Formation (based on information supplied by S. Cherkesova). e fragment was from the contact between two different rocks: grey siltstone and clayey shale. Because the siltstone had a carbonate cement, the ornament could be cleaned using dilute acetic acid. Both of these macroremains plus a small fragment are housed in the Palaeontological Museum, Russian Academy of Sciences, Moscow (collection no. with prefix PIN).
We have not seen the fish remains from a third level, in the Middle Member (5) of the Pil'hikaj Formation (Fig. 2), mentioned by Rogozov & Vasil'eva (1968).
Some macrospecimens were whitened with ammonium chlorite. All were photographed, using a digital camera Nikon D 200 or one attached

Specimen PIN 3845/4
is specimen is mostly preserved as a natural, external mould of a large fragment of a plate from the head carapace of a heterostracan (Fig. 3). Owing to the alignment of superficial ornamentation of oak-leaflike tubercles, the antero-posterior axis of the plate is orientated up-down on Figure 3, by comparison to some specimens figured and described by Dineley & Loeffler (1976: pl. 1: 1-3, pl. 4: 1, 3) where the narrower extremity of tubercles is orientated frontward. So, the minimum width of the plate is evaluated at c. 6 cm, with a preserved length of 22 mm. It does not show any plate or field boundary. e ornamentation is made of series of alternating tubercles, with bigger tubercles surrounded by fields of smaller ones (Figs 4; 5A-C). Bigger tubercles are oak-leaf-shaped, triangular with rounded lateral expansions. ey are 1.5 to 2 mm long (Figs 4; 5A-C). In most cases no boundary is seen on either side of these tubercles, except in a small area (Fig. 4C) where polygonal boundaries (either pentagonal or hexagonal) seem to surround some tubercles in a tessellated-like pattern. Each tessera would thus be made of a bigger central tubercle with smaller outer tubercles, with a total length of c. 1.5 mm. PIN 3845/4 has an ornamentation (or "exoskeletal ultrasculpture" sensu Märss 2006) which looks very like that of some traquairaspid heterostracans figured and described by Dineley & Loeffler (1976): ?Traquairaspis retusa Dineley & Loeffler, 1976(Dineley & Loeffler 1976: pl. 1: 1; fig. 5), Traquairaspididae indet. Type 2 (Dineley & Loeffler 1976: pl. 4, 1 and fig. 16). So, PIN 3845/4 does probably correspond to a fragment of a dorsal shield of a traquairaspid, and more precisely to its anterior part because the alignments of tubercles diverge rearward. e central part of PIN 3845/4 shows a series of small (< 1 mm long) triangular tubercles with only one expansion on both sides, that are surrounded by a single row of much smaller tubercles, each set of bigger and smaller tubercles being limited by a polygonal, either pentagonal or hexagonal, boundary which mimics c. 1.5 mm long tesserae (Fig. 4C). is pattern is reminiscent of Lepidaspis serrata Dineley & Loeffler, 1976 (figs 73-76 and pls 29, 32). However, no tessera of Lepidaspis Dineley & Loeffler, 1976 shows additional smaller tubercles, peripheral to the main central one (Dineley & Loeffler 1976: figs 73-75, pl. 32, 5-8) -except the tessera in their plate 32, 6 which bears a small lateral tubercle beside the main central denticulated one. So, PIN 3845/4 is probably not from Lepidaspis.

Specimen PIN 3845/3
is specimen is a 52 mm long fragment of a probable dorsal spine of a heterostracan. It is 23 mm wide at its (proximal) wider extremity, and 16 mm wide at its (distal) narrower extremity (Fig. 5D). It is laterally flattened and symmetrical. e narrower extremity is broken and thus shows a natural transverse section of the bone: its histology is typical of a heterostracan with a central cancellous, honeycomb-like layer where the cancellae have a mean width of 0.5 mm, while the outer layer, below the external tuberculated surface, is reticulated (with smaller cavities). e wider extremity of PIN 3845/3 is broken as well, and shows the same histological structure. e external ornamentation of its distal part, where the matrix has been removed, shows par-allel rows of alternating, narrow, long tubercles ( Fig. 5D) interspersed with much smaller tubercles (Fig. 5E, F). e bigger tubercles are 1 to 3 mm long × 0.5 mm wide. ey sometimes show a median longitudinal faint ridge (carina or crest) (Fig. 5F), and all have small lateral denticulated expansions that are always simple, rather sharp, and never subdivided (bifid). e smaller intermediate tubercles are very small, elongate, sharp and arranged as 1, 2 or 3 rows between the bigger tubercles ( Fig. 5E, F). No boundary of tessellated-like units is visible on this part of the specimen.
is ornamentation is very like the one of several traquairaspid taxa described by Dineley & Loeffler (1976), but different from each of them in its detailed features. e Traquairaspididae indet. Type 3 of Dineley & Loeffler (1976: pl. 4: 3, fig. 17) has bigger elongate and smaller intermediate tubercles, but the latter are relatively longer than on PIN 3845/3, and they are organised as "long, narrow […] twig-like ridges" (Dineley & Loeffler 1976: 40) fig. 18) also has intercalated smaller tubercles, but the bigger ones are much bigger (1.5 to 8 mm long × 0.8 mm wide), "round crested" and "commonly kinked, unbranched" (Dineley & Loeffler 1976: 41) unlike those of PIN 3845/3. PIN 3845/3 has the same shape as a specimen from the Lower Devonian of Spitsbergen which is the dorsal spine of a heterostracan, and was identified as Weigeltaspis heintzi Tarlo, 1964by Blieck (1983 pl. VI: 2-5; based upon the holotype preserved in the Palaeontological Museum of Oslo, Norway -specimen PMO D 2440/2441, and its probable counterpart preserved in the Muséum national d'Histoire naturelle, Paris -specimen MNHN.F.SVD900). However, specimen PIN 3845/3 is broken in its proximal part and is not preserved attached to its dorsal plate. e holotype of W. heintzi (Tarlo 1964: pl. IV: 6-7;1965: pl  Specimen GIT 580-5 is is a 15 mm long fragment of a dermal bony plate with elongate, sharp (higher than wide) tubercles ( Fig. 6A, B). ese tubercles show simple lateral denticulations. No intercalated smaller tubercles and no tessellated-like pattern are visible on this specimen. Tarrant (1991: 402) gives a diagnosis of the order Traquairaspidiformes (family Traquairaspididae of Dineley & Loeffler 1976) where the ornamentation of dorsal shields is "often of elevated, laterally serrated tubercles, commonly with narrow interstitial tubercles or ridges". Commonly does not mean always, so our specimen GIT 580-5 could be a traquairaspid. Being stratigraphically older than the specimens PIN 3845/2-4, the specimen 580-5 might represent a distinct taxon, but this cannot be demonstrated based upon such poorly preserved material.

Specimen PIN 3845/2
is specimen is probably a fragment from a plate, rather than a tessera, because its edges are broken (Fig. 6C). It shows two rather big, round-topped, c. 1.5 mm long tubercles with lateral simple or bifid denticulations. It shows no interstitial smaller tubercles, but a few small tubercles are visible on what is probably the outer edge of the specimen (upper part on Fig. 6C).

DISCUSSION
is material has been provisionally determined as "traquairaspid", "Traquairaspis sp. indet." or "Lepidaspis" on the handwritten labels and notes preserved with it. ese variations reflect the uncertainty regarding its taxonomic status. As described above, this heterostracan material is consigned to what is usually understood as traquairaspids (order Traquairaspidiformes or family Traquairaspididae, depending of authors' opinion). However, few articulated traquairaspids are known, and it is difficult to determine such fragmentary remains as the ones figured here (Figs 3-5; 6A-C). Rare exceptions are the specimens from the Lower Devonian of the Canadian Arctic, where most described species of Traquairaspididae (sensu Dineley & Loeffler 1976) have a fused dorsal shield enclosing orbital and branchial openings, and comprising rostral, pineal, orbital and median dorsal fields (see e.g., ?Traquairaspis mackensiensis Dineley & Loeffler, 1976[Dineley & Loeffler 1976 Brotzen,1933, following Tarlo (1965 who considered Weigeltaspis to be characterized by a "dorsal median plate long and narrow, with prominent median ridge in posterior half of plate". is material was attributed to the Traquairaspidiformes by Blieck (1983: 89) by comparison with Traquairaspis symondsi (Lankester, 1868) (in Dineley 1964: fig. 5), the type species of Phialaspis Wills, 1936in Tarrant (1991, and thus of the family Phialaspididae sensu Tarrant (1991). e second specimen of Weigeltaspis heintzi figured by Tarlo (1965: fig. 2C) shows a ?branchial plate, a dorsal disc and a field of tesserae, features that persuaded Tarlo to classify Weigeltaspis among his Psammosteiformes. e small area with a tessellated-like pattern on PIN 3845/4 ( Fig. 4C) recalls the ornament of Lepidaspis, but this latter genus was considered an agnathan vertebrate of uncertain affinities by its authors (Dineley & Loeffler 1976: 175 et seq.). All this seems to be very confusing. A revision of all material that has been called "traquairaspids" is needed. A conservative opinion is kept here, and the material from Chukotka is provisionally attributed to Traquairaspididae? indet. based upon its dermal ornamentation (exoskeletal ultrasculpture) of oak-leaf-like tubercles with mostly no tessellated pattern of the plates. Well developed dorsal structures such as a dorsal spine on Devonian agnathans is classically interpreted as an adaptation to Old Red Sandstone or Old Red Sandstone-like water environments, which occur by convergence in different higher taxa such as the heterostracans, osteostracans, galeaspids and pituriaspids (Janvier 1996: fig. 7.10.G).
Order CYATHASPIDIFORMES Berg, 1937 Family PORASPIDIDAE Kiaer, 1932 Poraspididae gen. et sp. indet.  DESCRIPTION is material is represented by three small fragmentary remains of dermal bony elements. e specimen GIT 580-4 ( Fig. 6G)  DESCRIPTION ese are two small bony elements, partly fractured, but corresponding to isolated tesserae. Specimen GIT 580-1 (Fig. 6E) is c. 1.8 mm long, with a single central, denticulated tubercle which shows a thin longitudinal crest. Specimen GIT 580-3 ( Fig. 6F . 4A). Both specimens can also be compared with broken portions of the superficial layer of traquairaspid elements (e.g., Fig. 4C), but in traquairaspids the base is usually much thicker than in GIT 580-1 and 3. So, affinities with Oniscolepis (senior synonym of Strosipherus) seem more likely.  (Dineley & Loeffler 1976: 190 DESCRIPTION e specimen is a diamond-shaped, 1.7 mm long tessera with a single central, narrow, elongate tubercle (c. 1.4 mm long). is tubercle has denticulated edges. Each denticulation is simple (undivided). Additionally, on each side of this central tubercle occurs a much smaller narrow tubercle (Fig. 6D). e base of the tessera is perforated by small foramina of the underlying (probably reticulated) layer. is tessera compares well with those of Lepidaspis serrata Dineley & Loeffler (1976: figs 74, 76, pl. 32: 6, 7), and especially with the tessera in their plate 32: 6, which bears a small lateral tubercle beside the main central denticulated one. However, because we have here a single tessera, it is difficult to compare with the great variability of shapes observed on Lepidaspis serrata tesserae (Dineley & Loeffler 1976), and so only tentatively assign GIT 580-2 to Lepidaspis.
( Fig. 6I) MATERIAL. -Specimen GIT 580-8: isolated tessera or platelet. DESCRIPTION is specimen is an isolated square-shaped element (either a tessera or platelet), partially broken, and with an altered surface. It bears a single, probably central, star-shaped tubercle (Fig. 6I). Because it is larger (c. 3.2 mm long for its preserved part) than all other tesserae that have been prepared from sample 73 of the Coast section, we suggest that it might correspond to a different taxon. Furthermore, with such a small sample (a single specimen), no thin section has been made, and thus its histology is not confirmed as being of a heterostracan. †Subclass THELODONTI Jaekel, 1911 Following Märss et al. (2007), we shall place the higher taxon as a Subclass equivalent to Heterostraci in the Class Pteraspidomorphi, although the full analysis of phylogenetic characters of this latter "clade" is as yet wanting and the interrelationships between the subclasses should be explored more fully.

REMARKS
As there were only a few specimens retrieved from the sample 73 of the Lower Member (1) of the Enmakaj Formation, it is difficult to determine the species of thelodont. All, however, have typical thelodontiform histology with a single or few pulp openings in the base.
Family TURINIIDAE Obruchev, 1964 REMARKS Several turiniid taxa have been described (e.g., Märss et al. 2007) but variation of scale form is not precisely known. Some of the scales from Chukotka are tentatively referred to three of the current turiniid species. Mark-Kurik E. et al. Genus Turinia Traquair, 1896

REMARKS
As there is only one articulated specimen of the type species Turinia pagei, from the Lower Devonian (lower Lochkovian) Lower Garvock Group of Scotland, and rare other patches of scales, we still cannot determine the full extent of variation in this taxon. e range of scales apparent on the macrofossils does not "match" the wealth of variation presented by isolated scales in beds of the same age (e.g., Gross 1967;Ørvig 1969a;Turner 1973: figs 8a, b, e, g, pl. 2;1982: pl. 97;Karatajūtė-Talimaa 1978;Märss & Ritchie 1998: fig. 49). is is one of the taxonomic problems to be accounted for when examining an assemblage of few scales.
Six of the scales, described below, would seem to be referable to one of the principal genera known in the Devonian, Turinia, by their platform-like crown with rounded undulating ridges or posterior extending lappets. e turiniid scales are of different age classes exhibiting from relatively shallow bases to deeper bases with small pulp openings of more mature ones (Märss et al. 2007).

DESCRIPTION
Specimen GIT 580-19 (Fig. 7A) is a mature scale seen here in antero-lateral view, with a high simple elliptical crown, pointed posterior, and a flat topped rounded anterior. ere are possible minor ridges on the posterior otherwise smooth neck. e condition of the scale surface, however, is poorly preserved (see Taphonomy discussion below). e deep base has an anterior extension of the base, which might have projected into a longer root. Scale GIT 580-16, seen in lateral view (Fig. 7B), is more typical of a head to cephalopectoral scale with a series of rounded undulations around the gently rounded crown. e upper surface is smooth. e base is not larger than the crown but extends slightly anteriorward; it too exhibits slight scalloping in its basal growth around a medium-sized central pulp cavity. ese scales are typical of many turiniid head or cephalopectoral scales and are tentatively placed in T. pagei. Alternatively they might belong to Turinia composita Karatajūtė-Talimaa, 2002 or Turinia polita Karatajūtė-Talimaa, 1978 (see below). DESCRIPTION e scale GIT 580-10 ( Fig. 7C) is wide and rhombic with the crown not quite as wide as the base. e anterior rim of the crown is gently scalloped into three sections. e large flat median part of the crown has two lateral lappet areas which expand posteriorly so that the crown ends in at least five points. A smaller extension to the lower left might be an artifact or a further lower lappet on that side. e posterior parts are not well preserved but probably extended well beyond the mid-posterior point of the crown. ere is a wide shallow grooved neck separating the crown from the relatively shallow base. As the basal view is not available the pulp opening type is unknown. e base is slightly thickened anteriorly into a short downwards-projecting narrow spur, which is broken off at the tip. Specimen GIT 580-11 (Fig. 7D)  sides. e extended parts are better preserved and number 9 or 10 with the posterior mid-point. e neck is wider and the base slightly shallower; it extends anteriorly into a short projection. As with the other scale, the basal view is not available and so the pulp opening type is unknown. Within the neck as seen in the lower right of Fig. 7D there are exposed parts of three fine concentric ridges, which might be evidence of growth lines within the basal tissue (cf. Märss et al. 2007). ese two scales with their wide laterally expanded crown with several separate posterior points are comparable with the body scales of Turinia composita (Karatajūtė-Talimaa 2002: fig. 1G, H) and to some extent with some of Turinia barentsia Blom & Goujet, 2002(Blom & Goujet 2002. As with the other scales, the surface is severely scoured or etched and opening of dentine and bony aspidine structure can be seen.
Turinia sp. cf. T. polita Karatajūtė-Talimaa, 1978 ( DESCRIPTION is relatively large scale, seen in dorsal crown view, is broken on either side of the crown. All that remains is the wide and flat elliptical mid-section of the crown and the anterior parts of two lateral segments, which may or may not have been separated from the mid-section by a deep channel on the one side. e remnants of an expanding lateral lappet are left mid-scale on the other side. e neck is trough-like and the base apparently not deep.
is specimen shows a typical turiniid body scale configuration, resembling the type species. It also resembles the mid-section of body scales of other turiniid taxa such as

DESCRIPTION
GIT 580-13 has a flat simple slightly subtriangular crown with a slight scalloping of anterior rim (Fig. 7G) and a shallow neck that merges with the rounded base that almost matches the crown in size. e ventral view shows the wide open, thelodontid type narrow base, which is a relatively thickened torus around a very large pulp cavity with some dental tubule openings within (Fig. 7H). Based on published occurrences, there are no exactly similar scales to this one from the Chukotka material or elsewhere. However, it is a placoid thelodont scale rather than a simple shark scale based on its rounded base and simple crown. It is not identical to but is generally comparable with those of Nikolivia gutta, known from several Lower Devonian (Lochkovian) assemblages, but shares the large open pulp cavity and smooth subrounded crown shape (cf.

DESCRIPTION
is is an unusual ovoid scale with a wide base, wider than the high, extremely flat and almost featureless crown (Fig. 7J, K). ere is a slight asymmetry to the crown with some curvature to one side. e base is also apparently not deep but it is overgrown with a small central pulp hole, denoting a mature state (Fig. 7I). ere is a lateral groove to one side leading to the scale rim.
e wide shape of this scale is unlike any Turinia or Nikolivia taxon seen elsewhere. It might be a pathological scale or broken with the crown top sheared off, and/or from the crown flatness it might be a ventral scale on the body. Some cephalopectoral scales of Turinia antarctica Turner & Young, 1992(Turner & Young 1992 show flat tops to the crowns that suggest this explanation. e presence of a groove on the base for a large canal is uncommon, and combined with the slight depression to the side of the scale, might indicate that this is a specialized scale associated with a pore-canal or lateral line (see Märss et al. 2007). ere is a possibility that this is a new form but until further material showing the variation is found, the identification of the taxon for this scale is tentative at best.

COMMENTS ON DISTRIBUTION OF THELODONTS
In general the majority of scales are typical of Lower Devonian assemblages around the Old Red Sandstone Continent, with Turinia and Nikolivia type scales. e turiniid scales fall within the variation range of the type species and co-occuring taxa such as Turinia composita, T. barentsia and T. polita, all found in the British, Baltic to Arctic localities. As so few scales were recovered from locality 73, Tonnel'nyj Brook, Chukotka and possibly the smallest scales were lost, the absence of such expected key taxa as Boreania minima Karatajūtė-Talimaa, 1985, which typically occur in the earliest Lochkovian (e.g., Talimaa 2000) is not surprising.
Based on published occurrences, there are no identical scales to the two nikoliviid-like scales in the new material from Chukotka. However, they are generally comparable with Nikolivia gutta known from the Lochkovian of Britain, the Baltic and Spitsbergen and perhaps others from Arctic Russia and Podolia (Ukraine).

DESCRIPTION
Four microremains from the Enmakaj Formation belong to placoderms. e best-preserved specimen ( Fig. 8A; GIT 580-22) is an isolated spindle-like skeletal element, twice longer than wide. It has a slightly oval central tubercle, placed asymmetrically, and around it smaller asymmetrical tubercles, forming three rows, except at one of the sides. At the opposite side the tubercles are partly laterally compressed and lamellar. Tubercles have narrow smooth ridges, which are most numerous on the central tubercle (up to 13). e ridges end at the sharp tips of tubercles. e second remain ( Fig. 8B; GIT 580-25) is rather fragmentary and shows spongy bone with four broken tubercles on it. Ornament of the tubercles is rougher than that of the previous specimen. e ridges are in cross section less sharp in comparison with those of the specimen GIT 580-22. One of the ridges bifurcates at its proximal end. Apices of the tubercles were differently directed. e specimen could be a fragment of a larger spindle-like element. Two remaining specimens of poor preservation show ornament of different type. e specimen GIT 580-23 (Fig. 8C) is probably a fragment of an exoskeletal plate with a slightly concave margin (?). It is covered with small stellate flat round or elongated tubercles. Number of short ridges varies from 7 to 10. Most of the tubercles are clearly separated from one another. e specimen GIT 580-24 (Fig. 8D) is a fragment, in which the ornament resembles somewhat that of the specimen 580-23 but is rougher. Stellate tubercles are closely backed. Short ridges end with minute rounded swellings. e specimens may belong to different forms.

COMPARISON AND STRATIGRAPHICAL DISTRIBUTION
Identification of the Lower Devonian placoderm microremains is complicated as in many cases the ornament of the carapace plates is not figured in details. However, there are exceptions, concerning acanthothoracids (also called as palaeacanthaspids or radotinids according to their earliest known representatives Palaeacanthaspis Brotzen, 1934 andRadotina Gross, 1950 Long, 1984 (New South Wales, Australia), carrying even finer nodules along ridges. In one of the scales (Long & Young 1988: fig. 9B) the ends of ridges are truncated. e Romundina type of ornament is recognized in several acanthothoracid tubercles from the Lochkovian-Pragian of southeastern Australia (Basden et al. 2000). Some acanthothoracids possess the radotinid type of ornament, consisting of conical stellate tubercles with 4-12 ridges. e latter ones bear rows of small round nodules. is kind of ornament occurs in a Pragian radotinid from the Armorican Massif, France (Goujet 1976).
Tubercles similar to those of the above radotinid, i.e. acanthothoracids, can be seen in some According to our interpretation these few placoderm remains from the Chukotka Enmakaj assemblage belong probably all to acanthothoracids. Acanthothoracids are particularly characteristic of the Lochkovian. ey are reported from Australia, North America and numerous regions of Eurasia, including the present day Arctic. Goujet (1998) mentioned that on the Prince of Wales Island, the Canadian Arctic, at least three different forms of these placoderms, one of them being Romundina, occur in the Lochkovian. In this region acanthothoracids can also be met together with actinolepid arthrodires.
In result of the reassessment of the paper by Mark-Kurik (1974) it can be said that two different acanthothoracids (one of them probably Romundina) come from the Pshenitsyn Formation of Kotelnyj Island, New Siberian Archipelago, Russia. Four trunk armour plates of latter material, i.e. MD, left ADL, right complex plate (AL + Sp + AVL) (Mark-Kurik 1974: fig. 1), and another complex plate (Mark-Kurik 1974: pl. II, fig.1) belong to a smaller acanthothoracid. e figures 1 to 7 of the same paper (Mark-Kurik 1974) show ornament of a larger acanthothoracid. e Pshenitsyn Formation is dated by Cherkesova (1988) as Lochkovian. e left ADL plate (Mark-Kurik 1974: figs 1-9, pl. II: 7) was erroneously identified as the equivalent plate of an arctolepid(?) arthrodire. is misinterpreation was repeated in the figure 5.3 of the paper of Blieck & Janvier (1993).
In the Lower Devonian of the Tajmyr Peninsula, westwards of the New Siberian Islands, acantho-thoracids occur on four levels of the Lochkovian Ust'-Tareya Regional Stage (Mark-Kurik 1994). ree levels are in the Uryum Beds, the fourth one belonging to the upper part of the Tolbat Beds (Mark-Kurik 1994: fig. 48). Acanthothoracid skull roof and trunk armour plates from Tajmyr were compared with those found in the northern part of the Siberian Platform (Norilsk area, Kureika and Koldy River outcrops) and the Timan-Pechora province (Vozej and Lekejyaga drill cores), NE of European Russia (Mark-Kurik 1994: figs 49, 50). According to Goujet (pers. comm. to EMK 1999)  Timan-Pechora province Goujet (1997) reported the presence of two forms, resembling the Saudi Arabia acanthothoracid, occurring together with a third one, practically undistinguishable from Romundina. In addition to acanthothoracids the actinolepid arthrodires have been found in the Lochkovian and Pragian of above province. Tsyganko et al. (2000) mentioned also the occurrences of radotinids in the Lochkovian Ovinparma Regional Stage of the same province. It can be concluded that acanthothoracids are more common in the Lochkovian than in the Pragian and Emsian. Garralepis sp. (Fig. 9A, B) MATERIAL DESCRIPTION e one scale GIT 580-21 has a kite-shaped crown 0.9 mm long with smooth straight lateral edges that converge at c. 60°, and three short ridges running back from the rounded anterior crown edge (Fig. 9A, B). e central area of the crown is eroded, but the sides are well preserved and slightly higher than the central area. e scale neck is very short anteriorly and slightly deeper laterally; the base is strongly convex, and deepest below the level of the anterior crown. Closely spaced grooves encircling the base mark the insertion of Sharpey's fibre layers.

COMPARISON
Although larger than the type scales from the Garra Formation (Lochkovian) of central New South Wales, Australia, the scale matches their simple morphology (Burrow 2002: figs 16E-G, 30A-H). e well-preserved lateral crown edges indicate this region is composed of orthodentine rather than mesodentine or enameloid; both latter tissue types are strongly eroded in the Chukotka microremains. Scales of Nostovicina lacrima (Valiukevičius, 1994) (Valiukevičius 1994) from the Lochkovian of Taimyr and Timan-Pechora, northern Russia, and more recently identified from coeval deposits in central New South Wales (Burrow 2002) have a similar shape, but have a mesodentine crown, whereas scales of Garralepis simplex Burrow, 2002 are characterized by having orthodentine crowns with enameloid in upper central layers of the growth zones. e simple but characteristic morphology, associated with non-mesodentinous histology, have not been identified in scales of any other known taxa; however, as the identification is based on a single scale, it is only tentatively assigned to Garralepis. DESCRIPTION e tooth is c. 2.5 mm high, subtriangular in cross-section with at least one sharp vertical carina, and multiple canals visible in the broken base (Fig. 9C). Surficial tissue along the visible carina is well preserved compared with the rest of the tooth, indicating it is probably orthodentine while the rest of the tooth is probably mesodentine. e tooth appears to be flattened, presumably labio-lingually.

COMPARISON
e cross-sectional shape of the tooth resembles that of the main cusps on dentigerous jaw bones and tooth whorls of ischnacanthiform acanthodians, e.g., the poracanthodid Zemlyacanthus menneri (Valiukevicius, 1992) (Valiukevičius 1992: pl. 4.2;8.1, 3) from the Lochkovian of Severnaya Zemlya. e relatively large size of the tooth is consistent with the size range of these structures also, rather than the much smaller palatine teeth that are also found in ischnacanthiforms (e.g., Valiukevičius 1992: fig. 4C). Both tooth whorls and dentigerous jaw bone cusps of Z. menneri show a dense reticulated network of canals in the tooth bases (e.g., Valiukevičius 1992: fig. 8), and tissue differentiation between the tooth carinae (orthodentine) and the rest of the tooth (osteodentine), characters that are also seen in GIT 580-26. None of the acanthodian scale taxa in the Chukotka assemblage have been associated with dentigerous jaw bones or tooth whorls at other localities. e tooth is tentatively assigned to the Ischnacanthiformes, as it could possibly be from other acanthodians with tooth whorls having pointed main cusps (see . DESCRIPTION ree scales conform to the three commonest morphotypes of Nostovicina guangxiensis, with all having equal width and length, a flat rhombic crown that is smaller than the base, and variably developed lateral crown edges (Fig. 9D-F). Scale GIT 580-18 (Fig. 9D) exemplifies the most common form of N. guangxiensis, with strong ridges extending back from the anterior edge, and short oblique ridges running down from the posterior corner of the crown; the scale is c. 1.3 mm long and wide. e crown on GIT 580-14 (Fig. 9E) is 0.7 mm long, heavily eroded and cracked, with scalloped anterior edges indicating the scale originally had four or five crown ridges. e lateral ledges join to form a posterior point extending slightly beyond the posterior corner of the base. Scale GIT 580-15 (Fig. 9F) is c. 1.4 mm long and wide, and is also very poorly preserved with worn remnants of one lateral ledge, a rounded anterior crown margin, and three ridges extending about a third the length of the crown.

COMPARISON
e lack of histological information hampers identification of the scales, with their shapes fitting within the broad range exhibited by those of Nostolepis striata Pander, 1856. Another taxon Nostovicina laticristata Valiukevičius, 1994(Valiukevičius 1994) from various Lochkovian circum-Arctic localities in northern Canada and Russia has similar morphotypes, but its scales are very small with a deep rounded base. Type scales of Nostovicina guangxiensis are from the Early Devonian of Guangxi, China (Wang 1992); the taxon is one of the commonest acanthodians in microvertebrate assemblages from Lochkovianearly Pragian deposits throughout southeastern Australia (Burrow 2002). e older teleostome Yealepis douglasi Burrow & Young, 1999 from the Ludlow of Victoria, Australia (Burrow & Young 1999), has scales with identical morphotypes and comparable size to the Early Devonian ones, but their histology is unknown. Scales with the same shape and histology as N. guangxiensis are also found in Silurian-Devonian boundary beds of the Birch Creek Section BCII, Roberts Mountains, Nevada, USA and the Klonk section, Czech Republic .

Incerti ordinis Incertae familiae Genus Cheiracanthoides Wells, 1944
Cheiracanthoides rarus Valiukevičius, 1994 ( DESCRIPTION e one scale GIT 580-20 is 1.0 mm wide, 1.0 mm long; the crown and base have a square outline (Fig. 9G, H). e crown surface is almost flat, curving down slightly along the anterior edges. Multiple short closely-spaced ridges probably ornamented these edges, although only those towards the centre of the scale are preserved with the rest of the crown being heavily eroded (Fig. 9G). e neck is concave and a constant depth of c. 0.15 mm all round. e base is convex, with a maximum depth of 0.4 mm. No pores are visible on the scale neck (Fig. 9H).

COMPARISON
e scale features are very poorly preserved, but the general shape and proportions match those of scales of Cheiracanthoides rarus from the Lochkovian of Taimyr, northern Russia. DESCRIPTION e Enmakaj assemblage contains a single small fragment of a porolepiform sarcopterygian (GIT 580-27; Fig. 10). It has a smooth cosmine covered surface penetrated by pore-canals. In cross section the upper cosmine layer consists of the fused goblet-shaped odontodes, some of them show-ing segments of narrow pulp canals. Flask-shaped cavities separate the odontodes and end higher up with pores (Fig. 10A). Below cosmine is a rather compact layer of spongiosa (Fig. 10B). e fragment can be provisionally identified as belonging to a species of Porolepis.

COMMENTS ON STRATIGRAPHICAL DISTRIBUTION
AND TAXONOMY e porolepidids are known from the Lower and Middle Devonian of many regions: Rhineland, Baltic area, Spitsbergen, Urals, ?Western USA, Canadian and Russian Arctic (including Novaya Zemlya), ?Vietnam (Ørvig 1957, 1969bMark-Kurik & Novitskaya 1977;Blieck & Janvier 1993;Vorobyeva 2004). In our case occurrences from the Lower Devonian are of particular interest. Earlier it was considered that porolepidids appeared in the Pragian (Siegenian) (Vorobyeva & Obruchev 1964). But re-assessment of age of several local stratigraphical units evidence that they were rather common already in the Lochkovian. Lochkovian stratigraphical units with Porolepis are: the Kureika Formation of the Siberian Platform Matukhin 1995), the Bely Kamen and Uryum Beds of the Ust-Tareya Regional Stage of Taimyr Peninsula (Cherkesova 1994), and the Pshenitsyn River Formation of Kotelnyj Island, New Siberian Archipelago (Cherkesova 1975(Cherkesova , 1988. Mark-Kurik (1974) mentioned the similarity of the Pshenitsyn River Formation fish assemblage to that of the Kureika Formation. Porolepis in the Enmakaj Formation of Chukotka came according to a later age dating also from the Lochkovian (Cherkesova, pers. comm. to EMK 1976) [see here the section "age and correlation"]. Porolepis is abundant in the Enmakaj Formation (Cherkesova 1973: 279). Two Siberian species have been formally defined: Porolepis taimyrica Vorobyeva, 1963and P. kureikensis Vorobyeva, 1963. Ørvig (1969b
e placoderm remains are also significant for Lower Devonian age. Acanthothoracids are generally known from the Lower Devonian but they are particularly characteristic of the Lochkovian. ey are known from the Lochkovian of Arctic Canada (Ørvig 1975;Goujet 1998), Kotelnyj Island, Arctic Russia (Mark-Kurik 1974, NW of the Siberian Platform (Norilsk area, Kureika and Koldy River outcrops), on four levels of the Lochkovian Ust'-Tareya Regional Stage of Taymyr Peninsula. Kolymaspis Bystrow, 1956, a specific acanthothoracid, was found in the Nelyudim Formation, Lochkovian of Eastern Yakutia (Denison 1978: 35;Matukhin 1995). Acanthothoracids occur in the Lochkovian of the Timan-Pechora province (Vozej and Lekejyaga drill cores), NE of European Russia (Mark-Kurik 1994). Tsyganko et al. (2000) mentioned also the occurrences of radotinids in the Lochkovian Ovin-  (Burrow 2002). Type specimens of Nostovicina guangxiensis are from the Early Devonian of Guangxi, China; the taxon is one of the commonest acanthodians in microvertebrate assemblages from Lochkovianearly Pragian deposits throughout southeastern Australia (Burrow 2002). Cheiracanthoides rarus is from the Lochkovian of Taimyr, northern Russia (Valiukevičius 1994).
us, most of the vertebrate microremains from Lower Member (1) of the Enmakaj Formation give a Lochkovian, and perhaps early Lochkovian age. Obruchev (1973: 194) briefly mentioned an undeterminable pteraspid find from the Chukotka coast (behind Tonnel'nyj), collected by Y. G. Rogozov. Novitskaya (1986: 119 and pl. XXIV: 1, 2) described a pteraspid fragmentary dorsal disc from "Chukotka, seashore behind Tonnel'nyj" as Pteraspidiformes indet. 1 (housed in the collection of the Institute of Palaeontology of the Russian Academy of Sciences, Moscow, specimen PIN 3845/1; "Heterostraci Pteraspidae" in Cherkesova 1973). Novitskaya (1986: 125) mentioned that in its shape and depth of the pineal notch the plate resembles that in Podolaspis. Later, she (Novitskaya 2004: 170: Pteraspidiformes indet. 2; misprint for Pteraspidiformes indet. 1) compared this specimen with equivalent skeletal elements of the pteraspids Podolaspis, Larnovaspis Blieck, 1984and "Belgicaspis" Zych, 1931(now Rhinopteraspis Jaekel, 1919see Blieck 1980see Blieck , 1984 fig. 4; see also Blieck 1984: fig. 74) who place the L/P boundary between the Old Red faunal zones I and II, above the Khmeleva 1 Member. So, after Novitskaya's (1986Novitskaya's ( , 2004 comparison of the Chukotka pteraspid to Podolian species, the Lower Member (4) of the Pil'hikaj Formation might be either middle-late Lochkovian or early Pragian in age. e latter age would thus not be in disagreement with Cherkesova (1973) who proposed a Pragian-Emsian age for the Pil'hikaj Formation. However, the specimen on which Novitskaya's (1986Novitskaya's ( , 2004 determination is based is a fragmentary dorsal disc with only part of its dermal bone preserved (Novitskaya 1986: pl. XXIV: 1, 2), thus without any generic diagnostic feature. Hence, this specimen does not give a precise dating for the lower Pil'hikaj Formation. So, let us consider the biostratigraphic informations given by the newly described material.
So, the traquairaspid heterostracan material from Lower Member (4) of the Pil'hikaj Formation supports a Lochkovian, and probably Middle to Upper? Lochkovian age, but not a Pragian age.
TAPHONOMICAL REMARKS e material from both the Lower Member (1) of Enmakaj Formation, and the Lower Member (4) of Pil'hikaj Formation has been collected in silici clastic rocks, either siltsone or fine-grained sandstone. In locality 73 (Lower Member (1) of Enmakaj Formation), the thelodont and acanthodian scales are very poorly preserved. It seems that they have been etched severely, and they show an unusual preservation. eir surface is scoured, producing a microcrystalline texture (Figs 6I; 7A, C-F; 9D-H). In some the dentine is reduced to a shell with copious holes (Fig. 6E). On the thelodont scale GIT 580-19 the outer durodentine or enameloid layer has been stripped away and out layer of base, leaving exposed orthodentine tubules (Fig. 7A). e acanthodian material also does not have good histological preservation (many hyphae and recrystallization: Fig. 9).
We are uncertain as to the cause of these features. It might be diagenetic with mild metamorphism that led to pyritization. Alternatively, have the microremains been through something else's gut with consequent acidic attack? Surely, acid preparation did not cause the poor preservation of remains. Quite possibly leaching is responsible for the poor state of some of the microremains, although placoderm remains and the porolepid fragment are not so badly preserved.
In the younger localities (74 and 73-4 [74-3]) from the Lower Member (4) of Pil'hikaj Formation, only macroremains of incompletely preserved heterostracans have been collected. e superficial ornamentation of their dermal plates is rather well preserved even if, in some places, the outer surface of the tubercles is worn, with probably disappearance of the outermost layer, which is classically attributed to enameloid, but might just be a durodentine, i.e. the outer untubuled layer of normal dentine.
PALAEOBIOGEOGRAPHIC INTERPRETATION e vertebrate remains that have been collected in the Early Devonian section along the De Long Strait, Chukotka, give various palaeobiogeographic signals, but most of the assemblage is typical for a detrital facies from the Old Red Sandstone Continent (ORSC). e traquairaspids from the upper level (Lower Member of the Pil'hikaj Formation) have strong affinities with the Lochkovian of other Arctic regions, that is both the Canadian Arctic (and in particular North-West Territories -NWT) and Spitsbergen. ese regions are classically reconstructed at the "northern" edge of the Old Red Sandstone Continent (Scotese 2002 "northern" margin of the ORSC (this continental drift would result in an oblique collision of the AACM with Laurussia -the ORSC -continuing until the end Devonian; see Cocks & Torsvik [2011: 30]). In such a configuration, the heterostracan faunas from Chukotka, NWT of Canada and Spitsbergen, would have been palaeobiogeographically connected through the shelf and narrow oceanic corridor that still separated the AACM and ORSC (Cocks & Torsvik 2011: fig. 17). is was also the case for the Early Devonian fish faunas of Severnaya Zemlya (an element of the Kara-Taimyr block) which had strong relations with Spitsbergen and the Canadian Arctic as well (see Blieck et al. 2002, section "palaeogeographical setting" for references and discussion). All these regions are representatives of what has been called the Arctic Province by Blieck & Janvier (1999: fig. 9.14) (Fig. 11). e microremains from the lower level (Lower Member of Enmakaj Formation) of Chukotka also show strong affinities with the ORSC. Similarly, the thelodont assemblage has strong affinities with the Lochkovian of other Arctic regions, especially in Russia and Spitsbergen. e general composition with turiniid and nikoliviid-like scales, such as Turinia composita and possible Nikolivia aligera suggests closest affinity with that from the Lochkovian Pod'emnaya Formation from the Matusevich River at October Revolution Island, Severnaya Zemlya Archipelago (Karatajūtė-Talimaa 2002). elodont assemblages within the Lochkovian further afield such as Arctic and western North America in general show a similar composition (e.g., Märss et al. 2007).
Investigation of vertebrate distribution in older Silurian sections in Siberia and elsewhere has shown the prevalence of their remains in deposits of the more shallow facies of marine palaeobasins (Märss & Einasto 1978;Karatajūtė-Talimaa & Predtechenskij 1995;Žigaitė et al. 2011). elodonts and certain other vertebrates are found in nearshore environments in deltaic-lagoonal zone to shallow-water marine environments of epicontinental basins (e.g., Turner 1999) and this would also seem to have extended into the Devonian.
Of the identifiable acanthodian scales, the one tentatively assigned to Garralepis, is perhaps the only unusual element as this taxon is otherwise known only from eastern Australia in Lochkovian times. Cheiracanthoides rarus has only been recorded previously from the mid-Lochkovian type locality in Taimyr, northern Russia (Valiukevičius 2000: fig. 1). Nostovicina guangxiensis, however, has a wider distribution both stratigraphically and geographically, being common in the Lochkovian-early Pragian of southeastern Australia (Burrow 2002) and in Silurian-Devonian boundary beds of the Birch Creek Section BCII, Roberts Mountains, Nevada, USA and the Klonk section, Czech Republic ). e type scales from younger strata in China tend to have a higher base than the older material, and a trend to a higher base is detectable in successively younger southeastern Australian occurrences; the Chukotka scale bases are relatively shallow, conforming to the older material. Ischnacanthiforms with dentigerous jaw bones and tooth whorls bearing tooth cusps like the Chukotka specimen have a nearly worldwide distribution in the Lochkovian. ese fish clearly had the ability to extend their range far wider than the other taxa, and can be treated as "pelagic".

CONCLUSIONS
Two Lochkovian assemblages of fossil fish comprising macro-and microfossils of heterostracans, turiniid and other thelodonts, acanthodians and acanthothoracid placoderms, typical of the Old Red Sandstone continental margins, is described for the first time from the basal members of the Enmakaj and Pil'hikaj formations in coastal exposures along the De Long Strait, central Chukotka, Arctic far-eastern Russia. In addition evidences of a sarcopterygian occurs in Member (1) of the Enmakaj Formation. e palaeobiogeographic affinities of these assemblages are most closely with other Arctic regions such as Severnaya Zemlya (October Revolution Island), Spitsbergen and the northern and north-eastern ORSC in general. More wideranging gnathostome taxa link the northeastern ORSC with East Gondwana.
Our knowledge of the macro-and microremains of certain taxa points to the necessity for a revision of both the traquairaspids and tessellated heterostracans as well as of Lochkovian thelodonts in the Arctic north and western ORSC. Further research into early placoderm microremains in the Silurian to Lower Devonian is also needed; little is known. and Valdek Mikli for the SEM micrographs: all these persons are from the Tallinn University of Technology (TUT, Estonia). Colleagues from the Palaeontological Museum, RAS, Moscow, and in particular G. Zakharenko, are thanked for the loan of specimens. Research work of EMK was financed by the Project SF0140020s08 of the Institute of Geology, TUT. Research work of AB was financed by both the French National IGCP Committee (IGCP Project 406 workshop in the Lithuanian Institute of Geology, Vilnius, 1999)  the Queensland Museum. e authors are grateful to Daniel Goujet and Annemarie Ohler (Muséum national d'Histoire naturelle, Paris), and to an anonymous reviewer for their valuable remarks. is is a contribution to both IGCP Project 591 " e Early to Middle Paleozoic Revolution: Bridging the Gap between the Great Ordovician Biodiversification Event and the Devonian Terrestrial Revolution" and IGCP Project 596 "Climate change and biodiversity patterns in the Mid-Paleozoic (Early Devonian to Late Carboniferous)".