Permineralised (hematite) wood remains come from yellow ferruginous sandstones of two localities, which are situated on the South of the Arkhangelsk region. The first one (outcrop K276), is near the Voya River (Fig. 1). Sandstone sample (50 × 20 × 30 mm) with small fragments resembling wood was split chemically, i.e. macerated with high-test hydrogen peroxide. Then “dispersed” wood fragments were selected (above 30 pieces) with a stereomicroscope Olympus SZ-6045. The other locality (borehole RPK3) is 40 km to the north from the first one (Fig. 1). The wood fragments have been extracted from the pieces of two very brittle specimens of stem. The biggest fragment of stem is up to 50 mm long and 65 mm wide, and was easily broken during the preparation. More than 15–20 stem pieces of each specimen were selected and studied.

All pieces of selected or macerated woods have been studied and photographed with a scanning electron microscope (CAMSCAN). Six to eight series of scanning electron microscope (SEM) for each specimen were made. Unfortunately, no other fossil remains (invertebrates, spores) were found in these localities.

This taxon is represented by two specimens of stem and a lot of small “dispersed” wood fragments from the sandstone sample (Plate I, Plate II and Plate III). The specimens (RPK3/4-1, RPK3/4-7) are from a very brittle decorticated stem. A lot of wood fragments from different places of the stem were studied with SEM. Main significant anatomical features of this stem are indicated in Table 1. The secondary xylem (Plate II5) and some tracheids of primary xylem (Plate I2, 4) are well-preserved in the stem, nothing is known of the pith. In oblique-radial section metaxylem elements have been observed in three-dimensional view (Plate I2, 4). Metaxylem tracheids are narrow (17–29 μm in diameter), hexagonal in shape. They have scalariform thickening both on radial and tangential walls. Secondary xylem is of pycnoxylic type. In radial section the polygonal (frequently hexagonal) narrow tracheids are above 800 μm long and 22–45 μm (average 32 μm) wide (Plate II5). The tracheids have mainly uniseriate and biseriate bordered pitting (Plate I3), rarely triseriate (Plate I1). The pits are round to hexagonal in shape (Plate I3), 7–13 μm (average 8–9 μm) in diameter. Pit apertures are inclined (Plate I1), slit-like (up to 1 μm wide, 7 μm long). There is no pitting on tangential walls. Xylem rays (Plate I6) are uni- and biseriate (14–20 μm wide), not high (commonly 10–15 cells or 170–355 μm high) and relatively long (530 μm in visible length). Ray cells are rectangular in shape (Plate II2). The width of the ray cells is 9–20 μm, their height is 17–33 μm (average 23 μm). In radial section, ray cells are 21–48 μm long. The distance between neighboring rays in radial section is 103–250 μm. There are 1–5 (frequently 2–3) contiguous pits in each cross-field (Plate I5). Mainly, they are arranged in two horizontal rows. The pits are 6.5–10 μm in diameter; rounded in shape with inclined narrow apertures. True growth rings are invisible.

The biggest “dispersed” wood piece from specimen K276/1 is up to 4000 μm long and 2100 μm wide (Plate III1). Secondary xylem is pycnoxylic type. Tracheids of secondary xylem (Plate II3) are narrow 13–50 μm (average 22–32 μm) wide, not so long (up to 640 μm long), hexagonal in shape (Plate III5). Tracheid pitting, only on radial walls, is predominantly biseriate (50%) and triseriate (12.5%), rarely uniseriate (37.5%). Multiseriate pits (Plate II1) are contiguous and alternate (araucarian), while uniseriate ones are free. Pits are rounded and hexagonal in outline, 5.2–12 μm (average 6–8 μm) in diameter, with inclined narrow-elliptical oblique apertures (up to 2 μm long and 8 μm wide). Rays are rare, only uniseriate (Plate III6). Most xylem rays (Plate III6) are short (3 cells or up to 120 μm high), several rays are larger (up to 16 cells or up to 380 μm high). The rays are 127–190 μm long in radial section, some rays exceed 225 μm long. Ray cells are rectangular or square (Plate III6), 23–72 μm long in radial section. Their height is 15–36 μm (commonly 30 μm). Cross-fields (Plate II4; Plate III3, 4) show 1–8 round-hexagonal pits (commonly 3–5 pits) 6–10 μm (7.5–9 μm) in diameter. Cross-fields pits are arranged in 1, frequently 2–3 rows. They are contiguous or free. Pit apertures are inclined elliptic or slit-like in shape.

Described above specimens show some identical characters: absence of pitting on tangential walls, uni-, rarely biseriate rays up to 15–16 cells high, 1–3 rows of tracheid pitting on radial walls, similar diameter of pits, from 1 to 8 round, hexagonal pits with inclined apertures of the pits in cross-fields. They correspond to a single taxon. Among Carboniferous plants of gymnospermous affinities, our wood most closely resembles in ray organization and pitting the species Eristophyton sp. described by Galtier and Scott (Galtier and Scott, 1994) from the Late Visean of France. As shown in Table 1 Eristophyton sp. is similar to described above wood remains in ray height (average 7.4 cells high), ray width (generally 1–2 cells wide), multiseriate pitting only on radial walls, cross-field pitting, etc. Therefore, the wood fragments under study are determined as Eristophyton sp., and their age is assumed to be Late Visean.

Recently a comprehensive review of the genus Eristophyton was given by A.L. Decombeix et al. (Decombeix et al., 2007). The Mississippian species of the genus were examined in that paper. The type species of the genus Eristophyton is E. beinertianum (Goepp.) Zal. from the Upper Visean of Poland, and the Upper Tournaisian of Great Britain. Another Mississippian species – E. waltonii (Lacey) Galtier et Scott comes from the Upper Tournaisian – Upper Visean of Scotland. E. beinertianum and E. waltonii have well-preserved primary xylem and a fragmentary data about characteristics of secondary xylem (tracheid pitting, cross-field, etc). Therefore it is extremely difficult to compare these two species (E. beinertianum and E. waltonii) with the studied material of Eristophyton sp. The third Mississippian species – E. fasciculare (Scott) Zal. described by Scott (Scott, 1902) and later Galtier et al. (Galtier et al., 1993) from the Visean – Namurian A of Scotland has well-preserved primary and secondary xylems. E. fasciculare shows mostly uniseriate, rarely bi-triseriate rays up to 35 cells high, and multiseriate tracheid. E. fasciculare differs (Table 1) from our specimens by higher rays, and absence of the tracheids with uniseriate pitting.

There are also several Early Carboniferous taxa resembling the wood fragments under study in some characteristics of secondary xylem, for instance in absence of pitting on tangential walls. As shown in Table 1 all these species show mainly uni-biseriate rays. Specimens under study resemble Endoxylon zonatum (Kidston) Scott (Lacey, 1953 and Scott, 1924) from the Visean of Scotland in uniseriate (rarely biseriate) rays and in uni-multiseriate pitting, although the rays of the latter are shorter (generally 1, rarely up to 9 cells high), shape of radial pits is oval to rectangular, and pit diameter varies (from 9 to 19 μm). Besides, pores in cross-fields are simple.

Bilignea solida Scott (Table 1) described by Scott (Scott, 1924) from Tournaisian – Namurian A of Scotland shows (Table 1) the similar uni-, biseriate rays with in average 1–10 cells high and multiseriate pitting. But the character of pits is different: its pits are larger in diameter and oval to rectangular in shape. B. resinosa Scott differs from Eristophyton sp. by presence of pits on tangential walls, and uni-, biseriate pitting.

Some similarity is observed between species of the genus Protopitys with specimens under study. P. buchiana Goepp. shows resembling height and width of rays, but distinguishes by uniseriate pitting, larger rectangular pits on radial walls, and a great number of simple pits in cross-fields. Another species – P. scotica Walton known from the Mid-Late Mississipian of Scotland (Galtier and Scott, 1994 and Walton, 1957) differs from our specimens (Table 1) by lower rays (generally 1–3 cells high). Besides, there is no information about its pitting in cross-fields.

In comparison with our specimens Paleoxylon bourbachensis Coulon et Lem. (Table 1) shows the same width in rays (1–2 cells) and multiseriate pitting. But main differences are in higher rays (up to 78 cells), and simple pores of pits in cross-fields. Stanwoodia kirktonensis Galtier and Scott (Galtier and Scott, 1991) from the Upper Visean of Scotland resembles (Table 1) our specimens in uni-, biseriate rays, and pitting (generally bi-triseriate). But S. kirktonensis shows shorter (1–10 cells high) rays, different character and number of cross-field pits (6–9 very small pits).

Dadoxylon ambiguum Witham sensu Frentzen known from the Late Visean of France (Galtier et al., 1998) resembles (Table 1) the wood of Eristophyton sp. in width of rays, pitting (uniseriate-multiseriate) of the tracheids and similar number of pits in cross-fields. In contrast, it is characterized by higher rays (up to 35 cells) and oval pits in cross-fields. The wood fragments under study differ from Tovoxylon alekseevii O. Orl. (Orlova, 2009) from the Upper Visean of northern Russia by absence of tangential pitting, taller rays (up to 16 cells instead 1 cells in high), character of the pitting on the radial walls (1–3 rows in contrast of uniseriate pitting), and inclined slit-like pores (instead simple round pores).

The specimens described in this paper extend the recognition of systematical composition and taxonomic variability of the plants of gymnospermous affinities found in the Carboniferous deposits of the northern Russia. As noted above, earlier the woods of P. bourbachensis, and T. alekseevii were described from the Lower Carboniferous deposits of northern Russia. Present data suggest that three stems of gymnospermous affinities occurred in the Early Carboniferous plant association of the North Russia. There are P. bourbachensis, T. alekseevii, and Eristophyton sp. Besides, the similarity between the Late Visean anatomically preserved plant assemblages of France (Galtier et al., 1998) and northern Russia is supported by the presence of two similar taxa – P. bourbachensis and Eristophyton sp. in both areas.