20 REMARKS. The list above includes all names generally recognized as synonyms of the widely reported Lepidodendron aculeatum, as well as all figured North American remains. The holotypes of Lepidodendron aculeatum (Sternberg 1820, Taf. VI, fig. 2) and Lepidodendron obovatum (Sternberg 1820, Taf. VI, fig. 1) both originated from the Radnice Member (Bolsovian), Kladno Formation, Bohemia, Czech Republic. Although most authors accept that these two species are synonymous, there has been no agreement on a preferred specific epithet. Andrews (1955, p. 178) mentioned that Lepidodendron dichotomum was the first species of Lepidodendron figured by Sternberg, implying that this would be the type species. However, he suggested that Lepidodendron obovatum might be a better type. On the other hand, Chaloner and Boureau (in Boureau 1967) pointed out that the problems surrounding the use of Lepidodendron obovatum made this species unsuitable. They regarded Lepidodendron aculeatum as more appropriate. We concur considering that Lepidodendron aculeatum has been used in a consistent manner, whereas the use of Lepidodendron obovatum has been more controversial. [We acknowledge that by admitting the synonymy of Lepidodendron aculeatum with Lepidodendron obovatum but not following Andrews’ choice of the latter as the correct name, we are contravening rules of priority (ICBN Article 11.5; http:// www.iapt-taxon.org/nomen/main.php?page=art11).]

From the Cumberland Basin, Bell (1944) figured as Lepidodendron aculeatum? two specimens representing quite substantial tree trunks. One of these specimens (Bell 1944, pl. XLIX, fig. 5) was available for us to study. Unfortunately, this is the least well preserved. It is the cast of a decorticated specimen that, although it shows the outline of the leaf cushions reasonably well, provides little detail of the leaf scar. It lacks a leaf trace and shows only vague parichnos markings. Only one leaf scar is clearly visible, at about one third the height of the elongate rhombic leaf cushion. Although preservation is poor, the size and shape of the keeled leaf cushions and the position of the relatively small leaf scar suggest that the Lepidodendron aculeatum determination by Bell is correct. The second specimen figured by Bell (1944, pl. L, fig. 3), although also poorly preserved, even more clearly belongs to this species.

21 In order to facilitate comparison with Lepidodendron bellii (= Lepidodendron obovatum sensu Presl in Sternberg 1838, non Sternberg 1820), we figure a well-preserved specimen of Lepidodendron aculeatum from most Langsettian strata in the Peñarroya Basin, southwestern Spain (Fig. 2).

22 COMPARISONS. Leaf cushions of Lepidodendron bellii are rhomboidal, with a marked horizontal asymmetry. The upper and lower ends of leaf cushions in this species are only slightly inflected in opposite directions, whereas in Lepidodendron aculeatum the cushions are fusiform, symmetrical and with acuminate apex and base that are distinctly inflected in opposite directions. The leaf scar is rhomboidal in both species, but is located in the upper third in Lepidodendron bellii, and a little above the middle in Lepidodendron aculeatum. Also, the length/breadth ratio of the leaf cushion is 3–4 in Lepidodendron aculeatum and up to 2.5 in Lepidodendron bellii. However, this ratio might vary depending on the position on the stem and is thus an unreliable character for species distinction.

23 STRATIGRAPHIC AND GEOGRAPHIC DISTRIBUTION. This is one of the most commonly (albeit not always correctly) cited species of Pennsylvanian Lepidodendraceae, with a range that includes most of the Namurian (from Chokierian upwards) as well as the entire Westphalian, where it is most frequent.

24 OCCURRENCE IN THE MARITIME PROVINCES, CANADA. Cumberland Basin (Nova Scotia): Bell (1944): locality 1983 (GSC 8562 — decorticated); Springhill (GSC 8558). Calder et al. (2006). Sydney Basin (Nova Scotia): Zodrow and McCandlish (1980). Calder et al. (1996).

25 OCCURRENCE IN THE UNITED STATES. Alabama: Gillespie and Rheams (1985, unfigured), Dilcher and Lott (2005), Dilcher et al. (2005). Arizona: Tidwell et al. (1992). Georgia: Gillespie and Crawford (1985), Gillespie et al. (1989). Illinois: Lesquereux (1879–1880), Janssen (1957), Langford (1958), Darrah (1969), Jennings (1984), Leary and Thomas (1989), Wittry (2006). Indiana: Canright (1959), Wood (1963), DiMichele and Dolph (1981), Willard et al. (1995). Michigan: Arnold (1934, 1949). New Mexico: Lucas et al. (2009), Lucas et al. (2013). Ohio: Cross et al. (1996). Pennsylvania: Wood (1860, 1869), Lesquereux (1858), Lesquereux (1879–1880), Oleksyshyn (1982). Rhode Island: Lesquereux (1879–1880). Utah: Tidwell (1967), Tidwell et al. (1974). West Virginia: Gillespie and Clendening (1962), Gillespie et al. (1966), Gillespie et al. (1978); Kosanke (1979).

26 HOLOTYPE. Pl. XII, fig. 2 of Bell 1966 (partially copied here as Fig. 3); Springhill Mines, Cumberland Basin, Nova Scotia.

27 DERIVATION OF NAME. After Walter A. Bell, in recognition of his fundamental contributions to Carboniferous paleobotany in Canada.

28 DIAGNOSIS. Leaf cushions rhomboidal, higher than wide, with rounded lateral angles and acute base and apex; upper part of area above the leaf scar relatively small, and that below large and with a keel. Leaf scars rhomboidal and situated high on the cushion. Infrafoliar parichnos small, vertically elongate.

29 DESCRIPTION (based on the specimen figured by Bell 1966). Leaf cushions rhomboidal, higher than wide, horizontally asymmetrical, with rounded lateral angles and an acute base and apex that are very slightly inflected in opposite directions. Dimensions: 28–30 mm long and 11–12 mm broad; ratio = 2.5. Leaf scars rhomboidal, placed in the upper third of the cushion and occupying less than one-third of cushion width; rhomboidal, with rounded margins and three small, rounded cicatricules arranged in a line. Dimensions: 3–4 mm long and 4–5 mm broad; ratio = 0.7–0.8. Infrafoliar parichnos vertically elongate, elliptical, small, but distinct. Keel well-marked below the leaf scar, with short (less than 1 mm) transverse markings. Upper part of the field small, with a short, poorly marked keel.

30 REMARKS. Although the holotype of Lepidodendron obovatum (as photographed by Němejc 1963, Thomas 1970 and Kvaček and Straková 1997) is conspecific with Lepidodendron aculeatum, other specimens figured as Lepidodendron obovatum are not. Jongmans (1929, p. 225–244) provided the first exhaustive synonymy of Lepidodendron obovatum. Of subsequent work, most important is Němejc’s (1947, p. 53) opinion that the specimen figured as Sagenaria obovata by Presl (in Sternberg 1838, Taf. LXVIII, fig. 6), does not belong to Lepidodendron aculeatum. Presl’s specimen was refigured by Thomas (1970, pl. 30, fig. 3 — erroneously cited in his plate caption as Presl in Sternberg 1838, pl. LXVIII, fig. 2), who assigned it, incorrectly, to Lepidodendron mannebachense. The holotype of Lepidodendron mannebachense (as photographed by Thomas 1970, pl. 30, fig. 4 and Kvaček and Straková 1997, pl. 33, fig. 6) possesses almost isodiametric leaf cushions, shown by Thomas (1970, fig. 7A) to lack infrafoliar parichnos. This specimen, originating from the Lower Rotliegend (Autunian) of Manebach in Thuringia (Germany), is different from the other remains attributed to Lepidodendron mannebachense by Thomas (1970). The latter, of Westphalian age, are characterized by more elongate leaf cushions showing infrafoliar parichnos.

The magnificent specimen from the roof of an unspecified coal seam at Springhill, Nova Scotia, figured by Bell (1966, pl. XII, fig. 2) as Lepidodendron aculeatum is regarded as conspecific with the specimen figured as Lepidodendron obovatum by Presl in Sternberg (1838, pl. LXVIII, fig. 6), and which we regard as different. The specimen from Springhill is here selected as the holotype of Lepidodendron bellii.

The possible synonyms in the list above refer to specimens (holotypes) that are poorly figured and which require revision of material not available to us.

31 STRATIGRAPHIC AND GEOGRAPHIC DISTRIBUTION. Presl’s specimen is from the Bolsovian of the Radnice Member, Kladno Formation, Bohemia. Crookall (1964) recorded this species (as Lepidodendron obovatum) throughout the Westphalian.

32 OCCURRENCE IN THE MARITIME PROVINCES, CANADA. Cumberland Basin (Nova Scotia): Bell (1944): Springhill (GSC 5813 — decorticated). Bell (1966): Springhill (GSC 14937 — holotype). Sydney Basin (Nova Scotia): Calder et al. (1996).

33 OCCURRENCE IN THE UNITED STATES. Alabama: Lesquereux (1879–1880), Gillespie and Rheams (1985), Dilcher and Lott (2005), Dilcher et al. (2005). Georgia: Gillespie et al. (1989). Illinois: Lesquereux (1879–1880), Janssen (1957), Jennings (1974), Wittry (2006). Indiana: Wood (1963), DiMichele and Dolph (1981). New Mexico: Lucas et al. (2013). Pennsylvania: Lesquereux (1879–1880). Utah: Tidwell (1967), Tidwell et al. (1974). West Virgina: Jongmans (1937); Gillespie et al. (1978).

34 DESCRIPTION. Leaf cushions smooth, only slightly raised, spirally arranged, and separated from one another by narrow grooves (c. 1 mm width); elongate rhomboidal, bilaterally symmetrical, with a very prominent keel (in the impression), present both above and below the leaf scar; upper and lower angles of cushions acute, with almost straight margins that are only very slightly inflected in opposite directions; lateral angles more rounded. Dimensions: 30–35 mm long and 11–13 mm broad, with the maximum breadth about the middle; length/breadth ≈ 2.7. Leaf scar situated a little above the middle of cushion, prominent, 3.5–4.5 mm broad, occupying one third of cushion width; its lateral angles prolonged into two well-marked straight, horizontal lines reaching the cushion margin. Vascular bundle prints and infrafoliar parichnos lacking. Ligule scar present at 2–3 mm above the leaf scar.

35 REMARKS. Bell (1944) figured and described as Lepidodendron jaraczewskii two very different specimens. One of these (Bell 1944, pl. LI, fig. 1) may be attributed to Omphalophloios anglicus (see below). The other specimen (Bell 1944, pl. LI, fig. 2 — Figs. 4a–b of the present paper) shows leaf cushions with a prominent keel both above and below the leaf scar and strongly marked straight lines from the edges of the leaf scar to the cushion margin. Lepidodendron jaraczewskii also possesses rhomboidal, elongate leaf cushions, but keels are less prominent and the lateral lines from the leaf scars curve downwards to reach the cushion margin. This is quite different to the pattern observed in the specimen from Nova Scotia.

Thomas (1970) assigned both of Bell’s specimens to Lepidodendron feistmantelii, an identification that we support only for one (Bell 1944, pl. LI, fig. 2 — Figs. 4a–b herein). Zalessky (1904, pl. IV, figs. 6, 10) based Lepidodendron feistmantelii on two specimens from the Donets Basin that show well-marked, rhomboidal, isodiametric, smooth leaf cushions with a marked keel and a centrally placed leaf scar. Zalessky (1904, pl. IV, fig. 9; pl. VIII, fig. 8) also figured two specimens under the name Lepidodendron veltheimii that show the same characters albeit with more elongate leaf cushions. This is regarded here as being within the intraspecific variation Zalessky (1904) considered Lepidodendron jaraczewskii, the specific name later used by Bell (1944), as conspecific with Lepidodendron veltheimii. However, the holotype of Lepidodendron veltheimii as photographed (upside down) by Kvaček and Straková (1997, pl. 54, fig. 4) has smaller leaf cushions with proportionately larger, transversally elongate leaf scars occupying most of the cushion width. Also, the arched lines that meet the cushion margin from the lateral sides of the leaf scar in Lepidodendron jaraczewskii are not present in Lepidodendron veltheimii.

36 STRATIGRAPHIC AND GEOGRAPHIC DISTRIBUTION. “Lepidodendron” feistmantelii is very rare. Zalessky’s material is from two different horizons in the Upper Bashkirian of the Donets Basin (C₂³) and Middle Moscovian (C₂⁶). In Great Britain, the species ranges from Langsettian to lower Bolsovian (see Thomas 1970). Bureau’s (1913) specimens come from the Namurian (Serpukhovian?) of Basse Loire, southern France.

37 OCCURRENCE IN THE MARITIME PROVINCES, CANADA. Pictou Coalfield (Stellarton Basin, Nova Scotia): Bell (1944): locality 2995 (GSC 8553) (Bell 1944, p.89 recorded the specimen from this locality as possibly originating from the Pictou Group).

38 REMARKS. A single specimen showing the imprint of the bark of what seems to have been a large tree was figured by Bell (1944, 1966) as Lepidodendron rimosum. We have not re-examined this specimen, but the rhomboidal, elongate leaf cushions, with rounded lateral angles and sharp bases and apices, slightly inflected in opposite directions, closely resemble those of Sternberg’s type (photographed by Stockmans 1935 and Kvaček and Straková 1997; see also the copy of Sternberg’s figure in Crookall 1964). The prominent, relatively small, rhomboidal leaf scars, placed a little above the middle of a cushion are also similar. Although the holotype of “Lepidodendron” rimosum has much larger interareas than the Canadian specimen, this is not necessarily significant for a specific distinction. The amount of separation between leaf cushions (i.e., the width of interareas) depends largely on their position on the stem, with the older parts commonly showing larger interareas. Indeed, the wide separation between leaf cushions in the holotype made Němejc (1947, p. 62) consider that it merely represented a developmental stage in the lower part of an old tree and that this character was useless for specific distinction. This species is too poorly understood for a precise generic attribution. Therefore, we retain it provisionally in Lepidodendron.

39 COMPARISONS. Bell’s (1944, 1966) specimen of “Lepidodendron” rimosum may be compared with the rare species “Lepidodendron” fusiforme (as figured by Crookall 1964). That species has elongate leaf cushions that are similar to those of “Lepidodendron” rimosum, but lack the very wide interareas in specimens of equivalent size (representing the older part of the tree). Also, the lateral angles of the cushions are more acute or only very slightly rounded in “Lepidodendron” fusiforme. The discussions presented by Němejc (1947) and Crookall (1964) show that “Lepidodendron” fusiforme and “Lepidodendron” rimosum have often been confused or regarded as synonyms.

40 STRATIGRAPHIC AND GEOGRAPHIC DISTRIBUTION. According to Crookall (1964), the species, although uncommon, is recorded throughout the entire Westphalian of Great Britain. Sternberg’s holotype originated from the Bolsovian of the Radnice Member, Kladno Formation, Bohemia. Lepidodendron taxandricum, which is regarded as a synonym, originated from the upper Westphalian A (upper Langsettian) of the Campine (Kempen) coalfield in Belgium.

41 OCCURRENCE IN THE MARITIME PROVINCES, CANADA. Cumberland Basin (Nova Scotia): Dawson (1868). Bell (1944): Springhill (GSC 9551). Bell (1966): Springhill (GSC 9551 — same as Bell, 1944). Pictou Coalfield (Nova Scotia): Bell (1962): locality 948 (GSC 810 — cf.); locality 990 (GSC 811 — cf.). Sydney Basin (Nova Scotia): Dawson (1868).

42 OCCURRENCE IN THE UNITED STATES. Illinois: Lesquereux (1866), Lesquereux (1879–1880), Langford (1958), Wittry (2006). Missouri: White (1899). Pennsylvannia: Wood (1860, 1869). Utah: Tidwell et al. (1974).

43 DESCRIPTION. Leaf cushions spirally arranged and separated by relatively large, unornamented interareas. They are fusiform, elongate and longitudinally symmetrical, with pointed upper and lower angles that are only very slightly inflected in opposite directions; lateral angles rounded. Dimensions: 15–20 mm long and 3–3.5 mm broad, with maximum width in the upper third; length/breadth ratio ≈ 5. Leaf scar prominent, rhomboidal, broader than long, with rounded upper and lower angles and acute lateral angles, occupying nearly the entire cushion width, 1.5–2 mm long and 2.8–3 mm broad; the three punctiform prints are placed in line, near the middle of the scar.

44 REMARKS. A single specimen from the Joggins section (locality 1343), cited as Lepidodendron rimosum by Bell (1944), but not figured, is included here as “Lepidodendron” cf. tijoui (Figs. 6a, b). Unfortunately, this specimen is not very well preserved due to the coarse grain size of the containing sediment. The elongate shape of the fusiform leaf cushions with acuminate, slightly inflected apex and base, and the position of the small leaf scars in the upper one-third of the cushion, as well as the presence of wide interareas, allow a general comparison with “Lepidodendron” rimosum. However, Bell’s specimen has smooth, apparently unornamented interareas, whereas the type material of “Lepidodendron” rimosum has more or less continuous, clearly marked lines parallel to the cushion margins. Bell’s specimen also shows a leaf scar that is a little broader than long and which occupies virtually the entire cushion width. In “Lepidodendron” rimosum it is longer than broad and occupies a little over one-third of the width. This suggests that “Lepidodendron” tijoui may be the better identification. Due to the relatively large size of both cushions and leaf scars in the holotype of “Lepidodendron” tijoui and the presence of interareas in the present specimen, we make the species attribution only tentatively.

Although the apparent absence of infrafoliar parichnos and relatively flat leaf cushions suggest a possible attribution to Diaphorodendron, we consider that this specimen shows insufficient detail for a precise generic attribution.

45 STRATIGRAPHIC AND GEOGRAPHIC DISTRIBUTION. The holotype of “Lepidodendron” tijoui is from St. Johns, Illinois.

46 OCCURRENCE IN THE MARITIME PROVINCES, CANADA. Cumberland Basin (Nova Scotia): Bell (1944): locality 1343 (one specimen without catalogue number).

47 OCCURRENCE IN THE UNITED STATES. Illinois: Lesquereux (1870), Janssen (1940). Missouri: White (1899). Pennsylvania: Wnuk (1985).

51 DESCRIPTION. Leaf cushions overlapping, rhomboidal in outline, much broader than long, with acute lateral angles, obtuse upper angle, and a rounded lower angle. Keel absent. Dimensions: 4–5 mm long and 6–10 mm broad; ratio = 0.5–0.6. Leaf scars occurring near the cushion apex, strongly protruding and squashed downwards on compression, transversely rhomboidal, elongate, with lateral angles very sharp and the upper and lower angles rounded, occupying approximately one third of cushion area. Dimensions: 2–3.5 mm long and 3–7 mm broad. In its lower part, the leaf scar bears three small rounded markings in line; only the central, larger cicatricule (vascular trace) is clearly visible in the leaf scar.

52 REMARKS. Bell (1944) figured several specimens from the Cumberland Basin as Lepidophloios laricinus, but two different species seem to be represented. One is characterized by the small leaf cushions and scars of Lepidophloios laricinus, whereas the other displays larger leaf cushions similar to those of Lepidophloios acadianus (a synonym of Lepidophloios macrolepidotus — see later). Bell (1944, 1966) also figured several specimens with ulodendroid and halonial (branch and strobilar) scars. Specimen GSC 4503 (Bell 1944, pl. LXI, fig. 1 — refigured here in Fig. 7) is a stem fragment with subcircular scars in a helicoidal pattern. It has transversely elongate leaf cushions, thus allowing its identification as Lepidophloios. This specimen is only slightly flattened, preserving most of its three dimensional aspect and thus allowing the helicoidal pattern of scars to be followed around the branch. These subcircular scars are set on protuberances. This is the Halonia condition that Jonker (1976), following Renier (1910), interpreted as slight elevations of cortical tissue supporting pedunculate strobili. This interpretation agrees with the reconstruction by Hirmer (1927, fig. 263). A central depression on the subcircular protuberance would mark the place of insertion of the strobilar stalk. Bell (1966, pl. XII, fig. 1) figured a similar specimen with halonial scars, but more poorly preserved. Both the latter specimen and the one figured here are from the same locality (1388) at Joggins.

A different, larger kind of scar of more elliptical shape and forming a depression on large branches or stems is exemplified by the negative print of a branch surface as figured by Bell (1944, pl. LVIII, fig. 4 — refigured here in Fig. 8 g). This kind of scar was interpreted by Renier (1910) and Jonker (1976) as corresponding to an adventitious branch. Adventitious branch scars seem uncommon, but they have been found on several different kinds of lycopsid stems, e.g., on Bothrodendron (Crookall 1964, pl. LXXIII, fig. 5) and Bergeria (Kidston 1893, pl. III, fig. 9, fig. 10, as Lepidodendron landsburgii — refigured as Lepidodendron ophiurus by Crookall 1964, pl. LXIII, fig. 1), as well as in Lepidophloios.

Some of the stem remains with ulodendroid scars are not clearly identifiable as Lepidophloios laricinus. For instance, specimen GSC 8556 (Bell 1944, pl. LX, fig. 5), which is unidentifiable either generically or specifically (see later), is a decorticated, poorly preserved stem fragment with rhomboidal leaf cushions, longer than wide, and an ulodendroid branch scar.

53 COMPARISONS. The leaf cushions of Lepidophloios macrolepidotus are much larger, up to four times the size of those of Lepidophloios laricinus. Also, the leaf cushions of Lepidophloios macrolepidotus are more equidimensional and do not protrude as much as those of Lepidophloios laricinus. The leaf scars of Lepidophloios macrolepidotus are more rhomboidal and occur at the extreme base of cushions. Moreover, the ligule pit in Lepidophloios macrolepidotus is more clearly separate from the leaf scar. According to Thomas (1977), the stomatal frequencies also differ, with 250 per mm² in Lepidophloios laricinus and 130 per mm² in cuticle preparations of material from the Joggins section. Thomas attributed the latter material to Lepidophloios acadianus (a synonym of Lepidophloios macrolepidotus).

Lepidophloios acerosus also possesses small, rhomboidal leaf cushions, but these are longer than broad, with a distinct keel (a feature absent in Lepidophloios laricinus); also its ligule pit occurs immediately above the leaf scar (in contrast to 1–1.5 mm above the leaf scar in Lepidophloios laricinus). According to Thomas (1977, p. 284), the cuticles of these two species are also different.

54 STRATIGRAPHIC AND GEOGRAPHIC DISTRIBUTION. Lepidophloios laricinus is quite common in Westphalian strata. It has been reported most often from the Langsettian and Duckmantian substages, and much more rarely from the (upper) Namurian (fide Crookall 1964). The type material is from the Radnice Member, Kladno Formation in Bohemia, of Bolsovian age. According to Wagner and Álvarez-Vázquez (2010), this species ranges from Langsettian to Cantabrian in the Iberian Peninsula, a longer range than is commonly accepted.

55 OCCURRENCE IN THE MARITIME PROVINCES, CANADA. Cumberland Basin (Nova Scotia): Dawson (1897). Bell (1944): locality 636 (GSC 8601); locality 1045 (GSC 8567 + one piece without number with cf.); locality 1386 (one piece without catalogue number); locality 1388 = 2990 (GSC 4503); locality 1982 (three pieces without catalogue number); locality 2989 (GSC 8561 + three pieces without catalogue number). Bell (1966): locality 1388 = 2990 (GSC 14930). Zodrow and McCandlish (1980). Minas Basin (Nova Scotia): Bell (1944): locality 77 (GSC 8216). Sydney Basin (Nova Scotia): Bell (1938): locality 498 (GSC 3373). Bell (1966): locality 1331 (GSC 14935). Minto Coalfield (Nova Scotia): Bell (1940): locality 2656 (GSC 10360); locality 2746 (GSC 10364).

56 OCCURRENCE IN THE UNITED STATES. Alabama: Gillespie and Rheams (1985), Lyons et al. (1985), Dilcher and Lott (2005), Dilcher et al. (2005). Colorado: Jennings (1980). Georgia: Gillespie and Crawford (1985), Gillespie et al. (1989). Illinois: Lesquereux (1866), Janssen (1940, 1957), Langford (1958), Jennings (1974), Boneham (1975), Gastaldo (1977), Wittry (2006). Indiana: Canright (1959). Missouri: White (1899), Basson (1968). Ohio: Cross et al. (1996). Utah: Tidwell et al. (1974). West Virginia: Jongmans (1937), Gillespie et al. (1978).

57 DESCRIPTION. Cushions imbricate, slightly over- lapping, rhomboidal, slightly broader than long, flat, no keel; lateral angles acute, upper angle obtuse, and lower angle well rounded. Dimensions: 13–25 mm long and 23–32 mm broad; ratio = 0.5–1.0. Leaf scars (situated on a protrusion of the cushion) transversely rhomboidal, with acute lateral angles and rounded upper and lower margins; showing three small, punctiform markings, not always clearly visible, fully aligned and close to the lower margin. Dimensions: up to 5 mm long and 8–10 mm broad. Ligule pit (often indistinct) at 4–5 mm above the leaf scar.

58 REMARKS. Bell (1944, p. 93) followed Kidston (1886, 1901) by including Lepidophloios acadianus in the synonymy of Lepidophloios laricinus, and suggesting that both Lepidophloios prominulus and Lepidophloios parvus might also belong to this species. However, Dawson (1897) defended the individual identity of his species on the basis of larger leaf cushion size, a different length/breadth ratio, and by the absence of a keel. Thomas (1977), who figured two specimens from Joggins, Nova Scotia, maintained Lepidophloios acadianus as a separate species on the basis of both cushion morphology and cuticle characteristics (larger epidermal cells and more stomata in Lepidophloios laricinus). He distinguished Lepidophloios acadianus from Lepidophloios macrolepidotus on its flatter leaf cushion and a more distinct ligule pit, which also appears more clearly separate from the leaf scar. Thomas (1977) described the epidermal cells as being of similar size in Lepidophloios acadianus, Lepidophloios macrolepidotus and Lepidophloios laricinus, but these cells have thicker anticlinal walls in Lepidophloios laricinus. He recorded different stomatal densities: 250 stomata per mm² in Lepidophloios laricinus, 50 per mm² in a specimen determined as Lepidophloios macrolepidotus, and about 130 per mm² in the cuticles from Joggins material identified as Lepidophloios acadianus by Dawson (seven specimens in the Kidston collection, Geological Survey, London — numbers 2318 to 2324). However, stomatal density is known to vary with the degree of humidity, and is thus of lesser importance for species identity. We regard Lepidophloios acadianus as synonymous with Lepidophloios macrolepidotus on morphological criteria.

Goldenberg (1862) believed Lepidophloios macrolepidotus and Ulodendron majus to be closely related. We agree that Ulodendron majus could be a Lepidophloios. Unfortunately, its holotype seems to be lost (fide Crookall 1966, p. 488), and the original illustration (Lindley and Hutton 1831, pl. 5) is inadequate for a definitive opinion. We regard several species of Lepidophloios introduced by Lesquereux (1870, 1884) as synonyms of Lepidophloios macrolepidotus.

59 STRATIGRAPHIC AND GEOGRAPHIC DISTRIBUTION. Goldenberg (1862) and Remy and Remy (1959) recorded Lepidophloios macrolepidotus from the Bolsovian of Saarland, Germany, and Thomas (1977) recorded it from the Westphalian A (Langsettian) of Yorkshire. The specimen of Lesquereux’s Lepidophloios dilatatus included in the synonymy above was from Cannelton, Pennsylvania, and is of Asturian age. The material described as Lepidophloios vaningenii by White (1899) is from Henry County, Missouri, and is of Bolsovian or early Asturian age.

60 OCCURRENCE IN THE MARITIME PROVINCES, CANADA. Cumberland Basin (Nova Scotia): Dawson (1868, 1897). Bell (1944): locality 1344 (one piece without catalogue number); locality 1388 = 2990 (GSC 4503); locality 1687 (GSC 4499); locality 1982 (GSC 8563 + three pieces without catalogue number); locality 1983 (one piece without catalogue number). Bell (1966): locality 1982 (GSC 8563 — same as Bell 1944). Thomas (1977): Joggins (seven specimens deposited in London).

61 OCCURRENCE IN THE UNITED STATES. Illinois: Lesquereux (1870), Lesquereux (1879–1884). Missouri: Lesquereux (1879–1880), White (1899). Ohio: Abbott (1968). Pennsylvania: Lesquereux (1879–1880).

62 REMARKS. Bell (1944, p. 94, pl. LX, fig. 2) recorded as Ulodendron a specimen (GSC 4492) representing an isolated ulodendroid branch scar. This specimen cannot be attributed to any particular lycopsid genus. This is also the case for another specimen (GSC 8556 — Bell 1944, pl. LX, fig. 5), which is a poorly preserved, decorticated stem remain showing rhomboidal leaf cushions and a large ulodendroid branch scar. It was included by Bell in Lepidophloios laricinus, as was the stem fragment recorded by Bell (1966, pl. XII, fig. 1) as “Lepidophloios laricinus forma Halonia tortuosa” (GSC 14930). All these decorticated stem remains are here regarded as indeterminable.

63 OCCURRENCE IN THE MARITIME PROVINCES, CANADA. Cumberland Basin (Nova Scotia): Bell (1944): locality 1388 = 2990 (GSC 4492); locality 1982 (GSC 8556). Bell (1966): locality 1388 = 2990 (GSC 14930).

67 REMARKS. As noted in the list of synonymy, Lepidostrobus variabilis is regarded generally as identical to Lepidostrobus ornatus (e.g., Němejc 1954). Bell (1944, p. 95) recorded several lycopsid strobili, not always well preserved, under the name of Lepidostrobus variabilis? and mentioned that more than one species might be represented. Bell (1944, pl. XLIX, fig. 6) figured a single fragmentary specimen from the Joggins section representing the upper part of a cylindrical, elongate strobilus with a broad (about 9 mm) central axis and long (12–14 mm) sporophylls with apparently rather short distal laminae. The specimen figured here as Fig. 9 c, from the Donald Reid collection (also from Joggins), shows similar characters. Both specimens generally fit the description of Lepidostrobus ornatus (as in Brack-Hanes and Thomas 1983). The specimen illustrated as Figs. 9a–b, from Inverness County, Nova Scotia, has a broad central axis and reflexed sporophylls and can be attributed only generically.

From these observations, it is apparent that the name Lepidostrobus ornatus covers strobili identified on similar morphological characters independent of spore contents.

68 STRATIGRAPHIC AND GEOGRAPHIC DISTRIBUTION. Lepidostrobus ornatus is widely recorded with a range from Chokierian to Asturian. The specimen illustrated as “Strobilus” by Parkinson (1804) and included in the species by Brongniart (1828a) is from the Coal Measures of Derbyshire, England, of Westphalian age.

69 OCCURRENCE IN THE MARITIME PROVINCES, CANADA. Cumberland Basin (Nova Scotia): Bell (1944): locality 1374 (GSC 8573); locality 1388 = 2990 (one piece without catalogue number); locality 1413C (two pieces). Donald Reid collection (1999): DRC-99 151 + DRC-99 152 + DRC-99 154. Sydney Basin (Nova Scotia): Calder et al. (1996). Minto Coalfield (New Brunswick): Bell (1940): locality 1145 (GSC 10556); locality 1153 (GSC 10663); locality 2656 (GSC 10361); locality 2789 (GSC 10526); locality 2791 (GSC 10373). Bell (1966): locality 1153 (GSC 10663).

70 OCCURRENCE IN THE UNITED STATES. Arizona: Tidwell et al. (1992). Illinois: Gastaldo (1977). Indiana: Canright (1959); Wood (1963). Pennsylvania: Lesquereux (1879–1880). Utah: Tidwell (1967), Tidwell (1975). West Virginia: Gillespie et al. (1978).

71 REMARKS. According to Chaloner (1953) Lepidostrobus olryi (now Flemingites olryi) is characterized by slender strobili. Chaloner obtained megaspores and microspores from several incomplete specimens of this species, which is also distinguished by short, closely adpressed sporophylls that are subtriangular with an acute apex.

All of Bell’s (1944) records of this species are unconvincing. The fragmentary and poorly preserved specimen figured in his pl. LVI, fig. 3 (GSC 9608) cannot be determined specifically. The specimen illustrated in his pl. LX, fig. 1 (GSC 9948) shows a Sigillariostrobus axis with incompletely preserved sporophylls that are longer than is apparent from the photograph. A better-preserved specimen (Bell 1944, pl. LVII, fig. 3 — GSC 8564), shows longer, less closely adpressed sporophylls than occur in Lepidostrobus olryi, and might belong to Lepidostrobus brongniartii.

The narrow strobili of Flemingites olryi are similar to those of Lepidostrobus brongniartii as figured by Němejc (1954), whose identification is subject to reservation. In the absence of spore data, the specimen illustrated by Bell (1944, pl. LVII, fig. 3) can only be assigned tentatively to Lepidostrobus brongniartii.

Chaloner’s (1953) brief synonymy list for Lepidostrobus olryi lacks Bell’s (1944) records, as does the list compiled by Crookall (1966).

72 STRATIGRAPHIC AND GEOGRAPHIC DISTRIBUTION. Lepidostrobus brongniartii has been recorded from several localities in the Westphalian and Stephanian of Europe.

73 OCCURRENCE IN THE MARITIME PROVINCES, CANADA. Cumberland Basin (Nova Scotia): Bell (1944): locality 1338 = 1686 (GSC 8564).

76 REMARKS. Bell (1944, pl. XLVIII, figs. 1–2) figures medium-sized leaves (about 50–55 mm long and 5 mm wide) each with a lanceolate lamina, being broadest at mid-length, a pointed apex, and a clearly marked central vein. These characters accord well with the description of this relatively common, easily recognizable species. We regard the three specimens from the Sydney coalfield recorded by Bell (1938, pl. XCVIII, figs. 7–9) as Lepidostrobophyllum lanceolatum var. constrictum as falling within the morphological range to be considered typical for this species. The varietal epithet refers to a constriction in the basal quarter of the lamina; we consider this feature insufficient to warrant varietal distinction.

77 COMPARISONS. Lepidostrobophyllum majus has a lamina that is larger, with its broadest part in the lower two thirds.

Lepidostrobophyllum lancifolium has a lamina with a sharply pointed apex and subparallel margins in its lower part; also, the lamina of this species widens at its base so as to constitute two small auricles.

78 STRATIGRAPHIC AND GEOGRAPHIC DISTRIBUTION. Lepidostrobophyllum lanceolatum is a relatively common species, recorded from Chokierian to Asturian strata. The type material comes from the Bensham seam at Jarrow Colliery, Durham, and is of Duckmantian age.

79 OCCURRENCE IN THE MARITIME PROVINCES, CANADA. Cumberland Basin (Nova Scotia): Bell (1944): locality 1685 (GSC 8206). Sydney Basin (Nova Scotia): Bell (1938): locality nº 2 colliery (GSC 3836 + GSC 4067); locality 513 (GSC 3304); locality 573 (GSC 3428); locality 715 (GSC 3420). Bell (1966): locality 1312 (GSC 14923); locality 1331 (GSC 14924). Calder et al. (1996). Minto Coalfield (New Brunswick): Bell (1940): locality 1108 (GSC 10788); locality 1140 (GSC 10772); locality 2819 (GSC 10659).

80 OCCURRENCE IN THE UNITED STATES. Alabama: Lyons et al. (1985). Illinois: Lesquereux (1870), Lesquereux (1879–1880), Langford (1958), Darrah (1969), Jennings (1970), Gastaldo (1977). Indiana: Canright (1959). Oklahoma: Read and Mamay (1964). Pennsylvania: Lesquereux (1879–1884), Oleksyshyn (1982). Rhode Island: Lesquereux (1879–1884), Lyons and Darrah (1979). Utah: Tidwell et al. (1974). West Virginia: Gillespie et al. (1978).

81 REMARKS. Although the specimen figured by Bell (1944, pl. LIII, fig. 3) from locality 562 in New Brunswick is incomplete (it lacks the pedicel) and small (48 mm long and 15 mm wide), the identification seems reasonable. It is based on the broad lanceolate shape of the lamina, with its widest part in the lower two-thirds and the margins more or less parallel in this part, converging in the upper part to a slightly acuminate apex.

Bell (1962, pl. XLVIII, fig. 5) shows a typical example from the Minto coalfield of New Brunswick, probably of Bolsovian age. This specimen (GSC 956) shows a lanceolate lamina 100 mm long (without pedicel) and 14 mm at maximum width in the middle third, tapering gradually to a pointed apex. It also shows a prominent (about 2 mm wide) single vein.

82 STRATIGRAPHIC AND GEOGRAPHIC DISTRIBUTION. According to Crookall (1966), this species is rare in the Westphalian B and C (Duckmantian and Bolsovian) of Great Britain, and relatively common in the Westphalian D (Asturian). In the Iberian Peninsula, Wagner and Álvarez-Vázquez (2010) recorded it as occurring from Langsettian to Stephanian B (sensu stricto).

83 OCCURRENCE IN THE MARITIME PROVINCES, CANADA. Cumberland Basin (Nova Scotia): Bell (1944): locality 1983 (one specimen without catalogue number). Tynemouth Creek (New Brunswick): Bell (1944): locality 562 (GSC 9589). Sydney Basin (Nova Scotia): Bell (1966): locality 1306 (GSC 14925). Minto Coalfield (New Brunswick): Bell (1940): locality 2616 (GSC 10466); locality 2832 (GSC 10363). Bell (1962): locality 1100 (GSC 956).

84 OCCURRENCE IN THE UNITED STATES. Alabama: Dilcher and Lott (2005), Dilcher et al. (2005). Illinois: Lesquereux (1879–1880), Janssen (1940), Janssen (1957), Langford (1958), Abbott (1963), Darrah (1969), Gastaldo (1977), Wittry (2006). Indiana: Abbott (1963); Willard et al. (1995). Kansas: Cridland et al. (1963). Massachusetts: Lyons and Chase (1976), Lyons (1984). Missouri: Abbott (1963); Basson (1968). Ohio: Cross et al. (1996). Utah: Tidwell (1967), Tidwell (1975), Tidwell et al. (1974). West Virginia: Gillespie et al. (1966), Gillespie et al. (1978).

85 DESCRIPTION. Lamina lanceolate, apparently thin, lacking in consistency, with cordate base and straight or slightly convex margins tapering gradually to a pointed apex. Simple vein strongly marked, wide (about 0.5 mm at base), reaching into the tip of the leaf, and bordered by a groove on both sides. Dimensions: 17–20 mm long and 6–7.5 mm at maximum width (in lower fourth of leaf length); ratio ≈ 2.7. Pedicel wedge-shaped, 6–7 mm long and 3–4 mm broad at the widest point (always narrower than the lamina).

86 REMARKS. Lepidostrobophyllum fletcheri Bell (1944, p. 96, pl. XLIX, figs. 1–4) was placed in synonymy with the distinctive species Lepidostrobophyllum morrisianum by Crookall (1966, p. 532), who also transferred the latter to the genus Lepidophylloides, which includes detached leaves of arborescent lycopsids. However, we agree with Tenchov (1987, p. 67) that the illustrations of both Lesquereux (1879) and Bell (1944) show this species to represent a sporophyll composed of a stalk, pedicel and lamina. The species should thus be included in Lepidostrobophyllum (as was done by Bell).

87 COMPARISONS. This species is clearly distinguished by the lamina having a lower part lacking consistency, a cordate base, and a pointed apex.

88 STRATIGRAPHIC AND GEOGRAPHIC DISTRIBUTION. Lepidostrobophyllum morrisianum is a rare element of European Westphalian floras. It is long-ranging, having been recorded from Langsettian to Asturian substages.

89 OCCURRENCE IN THE MARITIME PROVINCES, CANADA. Cumberland Basin (Nova Scotia): Bell (1944): locality 999 (GSC 8215 + GSC 8589 + GSC 9508 + GSC 9802 + one piece without catalogue number — together with Annularia ramosa and Eusphenopteris cf. neuropteroides).

90 OCCURRENCE IN THE UNITED STATES. Illinois: Lesquereux (1879–1880), Langford (1958), Wittry (2006).

94 DESCRIPTION (based on all the Canadian material available). Leaf cushions small, contiguous in medium-sized specimens, and with narrow interareas, up to 0.7 mm wide in the older ones. They are subrhomboidal, asymmetrical, higher than broad, with maximum width at about the middle. Lateral angles rounded, apex and base acute, very slightly inflected. Dimensions: 5–8 mm long and 3–4 mm broad; ratio = 1.7–2. Keel below the leaf scar faintly marked, unornamented or with short, transverse wrinkles. Leaf scars large in relation to cushion size, placed in the upper third of the cushion and occupying two thirds or more of cushion width. Leaf scars rhomboidal, broader than long, with upper and lower angles rounded and lateral angles acute, with three foliar prints (leaf trace and parichnos), more or less of equal size, in line, and situated in the lower part of the leaf scar. Dimensions: 2–3 mm long and 2.5–3.5 mm broad; ratio ≈ 0.8. Ligule pit placed just above the leaf scar.

95 REMARKS. Dawson (1868) figured and described a very fragmentary bark imprint from Bolsovian(?) strata of the Pictou coalfield in Nova Scotia as a new species, Lepidodendron decurtatum. Dawson’s drawing is reproduced here as Fig. 11b. Part of a photograph of the type, which is in the Dawson collection in the Redpath Museum, McGill University, Montreal, is reproduced here in Fig. 11a. It confirms the accuracy of Dawson’s drawing.

The characters described by Dawson (1868) include leaf cushions separated by shallow (narrow?) furrows, of rhombic ovate shape, obliquely acuminate below, nearly as broad as long, with transverse wrinkles, particularly on median line, with a rhombic leaf scar that contains three cicatricules (“vascular points” in Dawson’s description). Although the holotype of Lepidodendron decurtatum is fragmentary, its characters wholly agree with those recorded by Bell (1962) for Lepidodendron bretonense, a taxon which Bell (1938) had introduced previously as Lepidodendron dichotomum var. bretonensis from upper Asturian and lower Cantabrian strata of the Sydney coalfield, Breton Island, Nova Scotia. Bell (1938, 1962) did not refer to Lepidodendron decurtatum, a species that has not been mentioned since its introduction by Dawson (1868). Additional material of Lepidodendron bretonense was illustrated by Bell (1966) from the Minto coalfield in New Brunswick.

Specimens recorded by Bell (1944) as Lepidodendron dichotomum var. bretonensis from the Cumberland Basin, Nova Scotia, are all fragmentary and sometimes poorly preserved, e.g., the partly decorticated specimen refigured here in Figs. 12b–c. The holotype of Lepidodendron bretonense (Bell 1938, pl. XCV, fig. 9; pl. XCVI, fig. 1) from the Sydney Basin is refigured here (Figs. 12f, g, h) alongside a single specimen from the Guardo coalfield, northwestern Spain (Figs. 12d–e), testifying to the widespread occurrence of Diaphorodendron decurtatum.

Bell (1938) noticed the close resemblance of Lepidodendron dichotomum var. bretonensis with Lepidodendron andrewsii and Lepidodendron grigorievii. He noted that the only distinguishing character between Lepidodendron dichotomum var. bretonensis and Lepidodendron andrewsii is the presence of a keel below the leaf scar in some Canadian specimens of the former taxon. Both Lepidodendron andrewsii and Lepidodendron grigorievii are here regarded as likely synonyms of Diaphorodendron bretonense, and thus of Diaphorodendron decurtatum. Bell (1962) also compared Lepidodendron bretonense with the similar species Lepidodendron loricatum (originally described as Lepidodendron dichotomum by Arber 1903, 1914) and with Lepidodendron subdichotomum. The latter is a poorly characterized species (see Jongmans 1929, p. 317). According to Němejc (1947, p. 58), both species are possible synonyms of Lepidodendron dissitum, a species described from the Westphalian of Belgium (see Stockmans 1935), and recorded later from the Stephanian of northwestern Spain (Stockmans and Willière 1965). The Spanish records of Lepidodendron dissitum are discussed at length by Wagner and Talens (1985), who compared, but did not synonymize, this species with Lepidodendron bretonense. Bashforth (2005) transferred Lepidodendron bretonense to Diaphorodendron, which is correct in view of the absence of infrafoliar parichnos. However, he failed to refer to Lepidodendron dissitum and seems to have overlooked the comparison made by Wagner and Talens (1985). Lepidodendron dissitum also shows the absence of infrafoliar parichnos, which suggests that it should be assigned to Diaphorodendron. The material from northwestern Spain attributed to Lepidodendron dissitum comes mainly from Stephanian strata (up to and including Stephanian B), but also includes some remains of Asturian age. Wagner and Álvarez-Vázquez (2010, p. 316) identified the Spanish material as Diaphorodendron bretonense, but chose not to synonymize it with Lepidodendron dissitum, a species originally described from the Bolsovian of Belgium. The latter was redescribed by Stockmans (1935), who figured specimens which are quite similar to Diaphorodendron bretonense. However, slight differences are apparent. “Lepidodendron” dissitum from Belgium shows slightly more elongate leaf cushions, which in some specimens also display relatively wide, wrinkled interareas. Sauveur’s (1848, pl. LIX, fig. 3) illustration (a drawing) of the holotype of Lepidodendron dissitum suggests poor preservation. Unfortunately, the specimen cannot be located and may have been lost (C. Prestianni, personal communication, 2011). However, it is likely that the specimens illustrated by Stockmans (1935) from Belgium were assigned correctly to Sauveur’s species. In view of the (admittedly rather slight) differences noted, the present writers have opted for now to keep “Lepidodendron” dissitum separate from Diaphorodendron decurtatum. A more detailed study of the former species would be helpful.

Wagner (1965) tentatively identified material from the Stephanian in northwest Spain as Lepidodendron scutatum, but mentioned that Gómez de Llarena (1950) had used the name Lepidodendron dissitum. The latter name was adopted also by Stockmans and Willière (1965) for Spanish material. Wagner and Talens (1985) accepted the latter action, but compared with the material illustrated by White (1899) as Lepidodendron scutatum. In view of the poor illustration of the holotype of Lepidodendron scutatum (Lesquereux 1879, pl. LXIII, figs. 6, 6c), this taxon should be regarded as species dubia pending restudy of the type. The description by Lesquereux (1880, p. 369–370) seems to suggest that it might belong to Bergeria rather than Diaphorodendron. White (1899, p. 199) already expressed doubts as to the conspecifity of his material with the various specimens illustrated by Lesquereux.

Wnuk (1985) recorded several specimens as Lepidodendron bretonense from the Anthracite Field of Pennsylvania, U.S.A. Wnuk’s photographs show the right size and shape of leaf cushions for Diaphorodendron decurtatum, as well as a proper leaf scar in the top half of the cushion. In a drawing, Wnuk (1985, text-fig. 13) indicated the presence of foliar parichnos markings but no infrafoliar ones. The same drawing also shows the transverse wrinkles in the lower part of the leaf cushions, occupying the area of a keel that is not evident.

Two species described by Stockmans and Willière (1960) from the lower Westphalian of Belgium, Lepidodendron croftii and Lepidodendron paucipunctatum, are here regarded as synonyms of Diaphorodendron decurtatum. Stockmans and Willière (1960) only provided a brief diagnosis of these two species, which were not even compared with one another.

96 COMPARISONS. Lepidodendron dichotomum is similar to Diaphorodendron decurtatum in the size and shape of its leaf cushions, as well as in the position of its leaf scar. However, Lepidodendron dichotomum possesses infrafoliar parichnos (see the photograph of the holotype by Němejc 1946, pl. 1, figs. 1–3a — attributed, surprisingly, to Lepidodendron mannebachense by Opluštil 2010, fig. 5, who refigured the specimen). Infrafoliar parichnos are absent in Diaphorodendron decurtatum.

“Lepidodendron” dawsonii has slightly larger, more elongate, fusiform leaf cushions four to five times longer than wide. Also, the leaf scars are rhomboidal, with a nearly equal length/breadth ratio or with a slightly greater width, and a location just above the middle of the cushion. Infrafoliar parichnos are clearly absent in “Lepidodendron” dawsonii.

97 STRATIGRAPHIC AND GEOGRAPHIC DISTRIBUTION. The holotype of Lepidodendron decurtatum is from Bolsovian(?) strata of the Pictou coalfield, Nova Scotia. The type material of Lepidodendron andrewsii is from the uppermost Asturian or lowermost Cantabrian strata of Mazon Creek, Illinois. Lepidodendron wilmingtonii was based on Mazon Creek material from the Wilmington area. The type material of Lepidodendron grigorievi comes from the C₃¹ level of the Donetz Basin, of late Moscovian (late Asturian to early Cantabrian) age. The type specimens of Lepidodendron croftii and Lepidodendron paucipunctatum are from upper Westphalian A (upper Langsettian) strata of the Campine (Kempen) coalfield, Belgium. Diaphorodendron decurtatum has been recorded in Great Britain as Lepidodendron loricatum, Lepidodendron dichotomum and Lepidodendron obovatum, collectively from Langsettian to Bolsovian strata (see Crookall 1964; Thomas 1970). Bashforth (2005) recorded Diaphorodendron bretonense from Bolsovian strata of Newfoundland. In the Iberian Peninsula, Wagner and Álvarez-Vázquez (2010) recorded the species (as Diaphorodendron bretonense) from Asturian to Stephanian B strata.

98 OCCURRENCE IN THE MARITIME PROVINCES, CANADA. Cumberland Basin (Nova Scotia): Bell (1944): locality 999 (GSC 7506 + GSC 8666 + GSC 8547 — together with Lepidostrobophyllum hastatum + GSC 8548 — also with Lepidostrobophyllum hastatum + GSC 8549). Sydney Basin (Nova Scotia): Bell (1938): locality 513 (GSC 3307 + GSC 3345 + GSC 3348 — holotype of Lepidodendron bretonense); locality 720 (GSC 3427); locality 906 (GSC 3998); GSC 4491. Bell (1966): locality 504 (GSC 15059). Zodrow and McCandlish (1980). Calder et al. (1996). Pictou coalfield (Nova Scotia): Dawson (1868 — holotype). Bell (1940): locality 2657 (GSC 10927). Bell (1962): locality 977 (GSC 809); locality 1079 (GSC 812). Alison and Carroll (1972 — holotype). Newfoundland: Bashforth (2005).

99 OCCURRENCE IN THE UNITED STATES. Alabama: Gastaldo (1985). Illinois: Lesquereux (1879–1880), Noé (1925), Langford (1958, 1963), Wittry (2006), Taylor et al. (2009). Indiana: Abbott (1968), Boneham (1975), DiMichele in Eggert and Phillips (1982), Willard et al. (1995). Missouri: White (1899). Ohio: Abbott (1968), Cross et al. (1996). Pennsylvania: Oleksyshyn (1982), Wnuk (1985). West Virginia: Wagner and Lyons (1997).

103 DESCRIPTION. Leaf cushions flat, smooth, contiguous or separated by narrow grooves, obovate, with the broadest part in the upper third or upper half of cushion, straight or slightly inflected acuminate base, acute apex, and rounded lateral angles. Dimensions: 7–15 mm long and 2–4 mm broad; ratio ≈ 3.5. Keel absent or only faintly marked, smooth or (occasionally) with a few faint, short transverse lines. A small, irregular scar marks the position of the leaf base at the top of the cushion. Parichnos absent. Leaves linear-lanceolate, slightly curved near the base, single-veined, at least three or four times longer than the cushions, and inserted at 45–50° angle. Branches dichotomously forked at angles of 30–45° (measurements based on more examples than the few Canadian specimens).

104 REMARKS. Although Bell (1944, p. 80) mentioned the presence of Lepidodendron lanceolatum at several localities in the Cumberland Basin, he only figured one specimen (his pl. XLVIII, fig. 3 — later included by Bell 1962 in Lepidodendron pictoense), which is from the Minas Basin, Nova Scotia (Bell’s locality 1406, Moose River). Another specimen from the same locality (cited as 3100 but apparently the same as locality 1406) was figured by Bell (1966, pl. IV, fig. 13) and recorded as Lepidodendron pictoense. Crookall (1964, p. 290) referred Bell’s (1944) specimen to Lepidodendron ophiurus, as were those determined as Lepidodendron lycopodioides by Bell (1938) and Lepidodendron ophiurus by Bell (1940). Němejc (1947) included Lepidodendron lanceolatum in Lepidodendron acutum, a species synonymized herein with Bergeria dilatata.

The specimen in Bell (1944, pl. XLVIII, fig. 3) shows thin ultimate branches with attached leaves and terminal strobili. This material is closely similar to that figured here (Figs. 14c, d, f and 16a–b, g–h) from locality 1411, also at Moose River (Minas Basin). Bell (1944) compared his material with Lepidodendron simile, which we consider to be a synonym of Bergeria dilatata, and with Lepidodendron lycopodioides, a species that Němejc (1947) regarded as synonymous with Lepidodendron selaginoides. However, “Lepidodendron” selaginoides possesses fusiform, elongate leaf cushions with linear, very short leaves that are different from those of the Canadian specimens.

Zeiller (1888) and Kidston (1890, 1891) both referred Lepidodendron dilatatum (now Bergeria dilatata) to Lepidodendron ophiurus. The holotype of Lepidodendron ophiurus (Brongniart 1822, pl. IV, figs. 1a, b — drawings reproduced in Crookall 1964, text-fig. 93) shows three small leafy branches with rhomboidal leaf cushions that display a well-marked, unornamented keel. The enlargement figured by Brongniart (1822, fig. 1b) shows definite (rhomboidal) leaf scars with a single central marking, suggesting the absence of parichnos. Quite apart from the presence or absence of parichnos, the clearly differentiated leaf scars exclude the assignment of Lepidodendron ophiurus to Bergeria.

The synonymy list cites the many names applied to this species, both as synonyms and misidentifications. Bergeria dilatata is well represented in the Westphalian fossil record, allowing for a full description of adpression characters. Although our description is based primarily on the Canadian specimens, we also compared these with a large collection (several hundred specimens) from the Peñarroya Basin of early Westphalian age in southwestern Spain.

105 COMPARISONS. The presence of almost flat, smooth leaf cushions with small, irregular scars at or near the top of the cushion, and the absence of parichnos make Bergeria dilatata a distinctive species.

Bergeria worthenii has the field above and below the leaf scar ornamented with coarse, transverse, discontinuous wrinkles.

106 STRATIGRAPHIC AND GEOGRAPHIC DISTRIBUTION. The holotypes of both Lepidodendron dilatatum and Lepidodendron gracile originated in the Low Main coal seam, Felling Colliery, Durham coalfield, corresponding to lower Westphalian B (lower Duckmantian) according to Ramsbottom et al. (1978). The type material of Bergeria acuta, Bergeria angulata, Bergeria marginata, Bergeria quadrata and Bergeria rhombica all came from the Nýřany Member, Kladno Formation, central Bohemia, of Asturian/ early Cantabrian age. The holotypes of Lepidodendron lanceolatum and Lepidodendron scutatum came from the Clinton Coal, Missouri, probably of early Asturian age. In Great Britain, the species has been recorded (under various names) throughout the Westphalian. According to Wagner and Álvarez-Vázquez (2010), this species (recorded as Ulodendron acutum) ranges from Langsettian to Asturian in the Iberian Peninsula.

107 OCCURRENCE IN THE MARITIME PROVINCES, CANADA. Cumberland Basin (Nova Scotia): Bell (1944): locality 1053 (one piece with leafy branches without catalogue number); locality 1080 (GSC 9854 + GSC 9865 + GSC 9866 — together with Laveineopteris polymorpha and Calamites suckowii); locality 1340 (three pieces, without catalogue number, with small leafy branches); locality 1341 (one piece — with Zeilleria avoldensis); locality 1342 (one piece — leafy branches); locality 1344 (two pieces — leafy branches); locality 1375 (one piece — leafy branch); locality 1401 (six pieces — with Dorycordaites palmaeformis); locality 1406 (GSC 9913 + two pieces without catalogue number — also Flemingites russelianus); locality 1430 (three pieces — small leafy branches); locality 1439 (small leafy branch); locality 1491 (two pieces — poorly preserved); locality 1496 (one poorly preserved specimen associated with rootlets + two pieces associated with Calamites suckowii); locality 1497 (three pieces — fragmentary) locality 1498 (two pieces — leafy branches); locality 1983 (one piece); locality 2989 (two pieces); GSC 8995. minas basin (nova scotia): Bell (1944): locality 77 (GSC 332); locality 1411 (five pieces — also Flemingites russelianus); locality 2261 (without catalogue number). Bell (1966): locality 3100 (GSC 14929). Sydney Basin (Nova Scotia): Bell (1938): locality 514 (GSC 3305); locality 537 (GSC 3315); locality 922 (GSC 3999); locality 923 (GSC 4003); locality 929 (GSC 3978). Minto Coalfield (New Brunswick): Bell (1940): locality 1107 (GSC 10449); locality 2645 (GSC 10401); locality 2839 (GSC 10448). Bell (1966): locality 1107 (GSC 10449 — same as Bell, 1940).

108 OCCURRENCE IN THE UNITED STATES. Alabama: Gillespie and Rheams (1985); georgia: Gillespie and Crawford (1985). Illinois: Lesquereux (1879–1880), Darrah (1969). Kentucky: Lesquereux (1879–1880). Michigan: Arnold (1949). Missouri: Lesquereux (1879–1880, 1884), White (1899). Pennsylvania: Lesquereux (1879–1880, 1884), Wood (1860). West Virginia: Jongmans (1937), Gillespie et al. (1978).

109 DESCRIPTION. Leaf cushions contiguous, varying in outline from narrowly fusiform to obovate, with maximum width in the upper third; base elongate, acuminate, apex acute, lateral angles rounded. Dimensions: 10–13 mm long and 1.5–3 mm broad; ratio 5 to 7. No proper leaf scars, but a narrow irregular, transversely oval to punctiform scar in the upper third of cushion and ocupying almost the entire cushion width. Keel absent; field above and below the leaf scar occupied by relatively coarse, transverse, discontinuous wrinkles. Ligule pit situated at 0.5–1 mm above the leaf scar. Leaves linear-lanceolate, inserted at a narrow angle, rigid in aspect, in excess of 30 mm length, with an acuminate apex and a prominent vein.

110 REMARKS. Several specimens were figured as Lepidodendron wortheni by Bell (1944). Some of these (e.g., Bell 1944, pl. XLVII, fig. 2) are similar to the type material from Murphysboro, Illinois (see photographs in Janssen 1940, 1957, and Laveine et al. 2003). Bell’s specimens show almost total cover of leaf cushions by transverse wrinkles, leaving only a narrow, transversely oval strip of leaf scar. A ligule pit is visible immediately above the leaf scar. (Bell 1944, p. 90, mentions a punctiform ligule scar at 0.25 mm above the “leaf trace”, referring to his pl. XLVII, fig. 4 — a detail reproduced here as Fig. 15 a). Two of the specimens figured by Bell show branches with attached leaves: one of these (Bell 1944, pl. L, fig. 2) represents a thin distal branch with spreading leaves; the other (Bell 1944, pl. XLVII, fig. 2; partially reproduced here as Fig. 15 e) represents a larger branch.

111 COMPARISONS. Although various species belonging to different genera also show abundant transverse wrinkles covering most of the leaf cushions, Bergeria worthenii may be distinguished by large cushions with a generally convex surface, as well as by the absence of a proper leaf scar.

The similar Namurian (Serpukhovian) species Lepidodendron lossenii also possesses fusiform leaf cushions with short, transverse and irregularly placed wrinkles in the field above and below the leaf scar. However, its leaf cushions are always small and have a length/breadth ratio of ≈ 10.

White (1937) mentioned the general similarity of the Mississippian species Lepidodendron wedingtonense with Lepidodendron worthenii and its synonym Lepidodendron brittsii. However, Lepidodendron wedingtonense shows small, transversally elongate leaf scars with three relatively big cicatricules. This distinguishes it clearly from Bergeria.

112 STRATIGRAPHIC AND GEOGRAPHIC DISTRIBUTION. The type material of Bergeria worthenii comes from Murphysboro, Illinois, U.S.A., and is of early Asturian age (Peppers 1996, p. 61–62). Crookall (1964) recorded the species throughout the Westphalian of Great Britain, being rare in Westphalian A (Langsettian) and B (Duckmantian), and fairly common in Westphalian D (Asturian). According to Josten (1991) the species ranges from Langsettian to Bolsovian in the Ruhr district of western Germany.

113 OCCURRENCE IN THE MARITIME PROVINCES, CANADA. Cumberland Basin (Nova Scotia): Bell (1944): locality 666 = 1141 (GSC 10232 — part and counterpart); locality 1039 (GSC 9027 + GSC 9301); locality 1498 (cf. — three pieces without catalogue number). Bell (1966): locality 666 = 1141 (GSC 10232 — same as Bell, 1944). Sydney Basin (Nova Scotia): Bell (1938): locality 914 (GSC 4098); locality 922 (GSC 4063); locality 923 (GSC 3514 + GSC 4084).

114 OCCURRENCE IN THE UNITED STATES. Illinois: Lesquereux (1866), Lesquereux (1879–1880), Janssen (1940, 1957), Langford (1958), Crookall (1964), Laveine et al. (2003), Wittry (2006). Indiana: Canright (1959), Wood (1963), DiMichele and Beall (1990). Missouri: Lesquereux (1879-80), White (1899).

117 DESCRIPTION. Cylindrical strobilus up to 60 mm long and 15 mm broad, tapering only in the top part to end in a rounded apex. Sporophylls arranged in a low-angle spiral, rigid in aspect and closely adpressed. Sporophyll lamina lanceolate, with acute apex and a relatively broad, prominent central vein. Dimensions: up to 10 mm long and 1.3 mm wide.

118 REMARKS. The association of Flemingites russelianus with Lepidodendron acutum (now Bergeria dilatata) has been generally recognized (e.g., Chaloner 1953; Álvarez- Vázquez 1995). All the Canadian specimens included in this species are associated with branches of Bergeria dilatata. None of these display sporangial contents.

The incomplete strobilus figured by Jongmans (1937, pl. 15, fig. 25) as Lepidostrobus cf. goodei, most likely belongs to Flemingites russelianus, a conclusion reinforced by its association with Bergeria dilatata (recorded by Jongmans as Lepidodendron ophiurus).

119 STRATIGRAPHIC AND GEOGRAPHIC DISTRIBUTION. Binney’s material comes from Airdrie, Lanarkshire, Scotland, from rocks of Duckmantian age. Chaloner (1953) and Crookall (1966) recorded the species from Langsettian to Bolsovian in Great Britain. In the Peñarroya Basin, southwestern Spain, it occurs in upper Langsettian and upper Duckmantian/lower Bolsovian strata (Álvarez-Vázquez 1995, 2000), often in association with specimens of the relatively common Bergeria dilatata.

120 OCCURRENCE IN THE MARITIME PROVINCES, CANADA. Cumberland Basin (Nova Scotia): locality 1053 (GSC 9951 — fragmentary + GSC 9952 + GSC 9953 — together with Senftenbergia plumosa); locality 1406 (GSC 9913 — strobili associated with leafy branches of Bergeria dilatata + one specimen without catalogue number); locality 1411 (two pieces — together with leafy branches of Bergeria dilatata); locality 1496 (one specimen, part and counterpart, associated with Bergeria dilatata, Calamites suckowii and Dorycordaites palmaeformis).

121 OCCURRENCE IN THE UNITED STATES. Michigan: Arnold (1949).

124 DESCRIPTION. Ribs 5–8 mm wide, slightly convex, separated by zig-zagging longitudinal furrows. Straight or slightly arched, horizontal (transverse) furrows above the leaf scars which are closely spaced, hexagonal, straight-sided, with a slight rounding of the lower margin and a small notch in the upper. Dimensions: about 4 mm long and 5–6 mm broad; ratio = 0.6–0.8. Three small cicatricules arranged in line in the upper third of leaf scar; the central (leaf trace), punctiform to transversely elongate, and the lateral (parichnos) oval and slightly downflexed.

125 REMARKS. Bell (1944, pp. 91–92, pl. LIII, fig. 1) figured and described one specimen (GSC 5899) from a single locality at Springhill as Sigillaria mamillaris. This specimen is moderately well preserved, and shows characteristic hexagonal leaf scars as well as zigzagging furrows. Crookall (1966, p. 376) included Bell’s specimen in the synonymy of Sigillaria mamillaris.

Bell (1944, pl. XLV, fig. 2; pl. XLVI, fig. 1) also figured two specimens from locality 1338 in the Cumberland Basin under the name of Sigillaria elegans. Although Crookall (1966) includes these specimens in the synonymy of Sigillaria elegans, the larger leaf scars exclude Bell’s specimens from that species as it is commonly understood. They are here assigned to Sigillaria hexagona, a species generally included in the synonymy of Sigillaria elegans. This poses a taxonomic problem. Brongniart (1837, pl. 146, figs. 1, 1A; pl. 158, fig. 1) illustrated two very similar specimens as Sigillaria elegans and Sigillaria hexagona, respectively. Sigillaria hexagona is a species based on Schlotheim’s (1820) Palmacites hexagonatus. We regard both specimens illustrated by Brongniart as Sigillaria elegans sensu Brongniart (non Sternberg), this being the form with small hexagonal leaf scars that occurs from Chokierian to middle Langsettian. Brongniart (1837, pl. 155) further illustrated a specimen with larger leaf scars as Sigillaria hexagona, a determination we accept. However, in the text volume, Brongniart (1837, p. 439) included the specimens that he figured as Sigillaria hexagona with Sigillaria elegans. Three different localities in western Germany were mentioned by Brongniart (1837), without specifying the origin of each specimen. All these occur in lower Westphalian strata.

Sigillaria elegans was introduced by Sternberg (1825, Taf. LII, fig. 4) on the basis of upper Stephanian material from Wettin, Saxony. Jongmans (1932) discussed the taxonomic problem surrounding this species, and suggested that the middle to upper Namurian and lower Westphalian material that had been assigned to Sigillaria elegans by Brongniart might not be conspecific with the (Stephanian) holotype of that species. Jongmans (1932) also suggested that Sternberg’s Sigillaria elegans might be conspecific with Sigillaria brardii. However, the name Sigillaria elegans has been generally accepted for the small favularian Sigillaria with often poorly marked, relatively prominent leaf scars of hexagonal shape, which occurs in Namurian and lowermost Westphalian strata. It seems likely that this species (Sigillaria elegans sensu Brongniart, non Sternberg) may be conspecific with Sigillaria minima as figured and described by Brongniart (1837, pp. 435–436, pl. 158, figs. 2, 2A) from Namurian strata in the Vendée, southwestern France. This specimen (holotype) was refigured (photographically) by Bureau (1913, pl. XLI, figs. 3, 3A).

Although Bell (1940, 1944, 1966) recorded Sigillaria elegans from Canada, we regard his specimens as belonging to Sigillaria hexagona.

126 STRATIGRAPHIC AND GEOGRAPHIC DISTRIBUTION. Present (under different names) in upper Namurian and Westphalian strata.

127 OCCURRENCE IN THE MARITIME PROVINCES, CANADA. Cumberland Basin (Nova Scotia): locality 1374 (cf. — poorly preserved specimen without catalogue number). Bell (1944): locality 667 (one piece without catalogue number); locality 1338 = 1686 (GSC 6537 + GSC 6542 + GSC 8557 — together with Sigillaria scutellata); locality 1498 (part and counterpart — with Senftenbergia plumosa); Springhill (GSC 5899). Bell (1966): locality 1493 (GSC 14934). Calder et al.(2006).

128 OCCURRENCE IN THE UNITED STATES. West Virginia: Gillespie and Latimer (1960); Gillespie et al. (1966).

129 REMARKS. Bell (1940, pl. VIII, figs. 1, 2) illustrated as Sigillaria reticulata two specimens from a borehole in the Pictou coalfield (Stellarton Basin), Nova Scotia, that are typical for this species. The stem is not ribbed and displays well-marked longitudinal wrinkles, as well as relatively small, spaced-out, oval leaf scars. Two specimens from locality 2488, from the same coalfield (see our Fig. 18b-d) show the small, about 1.5–2 mm long and 3 mm broad, transversely oval leaf scars, 5 mm apart in vertical rows on a non-ribbed stem. Although not very well preserved due to the coarseness of the sediment, the short, more or less flexuous wrinkles are visible.

In contrast, the specimen figured by Bell (1944, pl. LVII, fig. 1) as Sigillaria reticulata? from Springhill (locality 1081), Nova Scotia, shows widely spaced oval leaf scars, about 3 mm long and 4.5 mm broad, which are separated by an irregularly striped area that we interpret as subepidermal and thus a preservational character. However, it also shows some leaf scars. This specimen resembles that mistakenly figured as Sigillaria brardii var. denudata by Arber (1912, pl. 12, fig. 10) from strata of Asturian age in the Forest of Dean, England. From the same locality (1081), Bell (1944, pl. LVI, fig. 4) figured as Sigillaria laevigata? a decorticated fragment displaying large, paired parichnos in vertical rows on a longitudinally striate stem, clearly in subepidermal preservation. This latter specimen is assigned here to Syringodendron sp.

Sigillaria vermiculata (Deltenre, 1924–1926, p. 84, pl. XVIII, figs. 9, 10) also encompasses non-ribbed material in which the stem is covered by longitudinal, more or less flexuous wrinkles, with oval leaf scars placed in alternating vertical rows. There is no apparent reason to separate this material from Sigillaria reticulata, which Deltenre figured from the same horizon as his type of Sigillaria vermiculata — Veine du Parc, Charleroi coalfield, Belgium.

130 COMPARISONS. Well-preserved specimens of Sigillaria reticulata with transversely oval leaf scars on a non-ribbed stem are sufficiently distinctive to avoid confusion with any other Sigillaria.

There is a faint resemblance with Asolanus camptotaenia, but the latter has smaller, subrhomboidal leaf scars, broader than long, with a characteristic cross-over ornament in between.

131 STRATIGRAPHIC AND GEOGRAPHIC DISTRIBUTION. Sigillaria reticulata is rare. Its holotype is from White River, Arkansas, U.S.A. In Europe, it has been recorded from the Langsettian of Yorkshire, England (Crookall 1966), the Duckmantian of Belgium (Deltenre 1924-1926), and the upper Duckmantian/lower Bolsovian of Peñarroya in southwestern Spain (Wagner and Álvarez- Vázquez 2010).

132 OCCURRENCE IN MARITIME PROVINCES, CANADA. Cumberland Basin (Nova Scotia): Bell (1944): locality 205 = 1081 (GSC 10943 — together with Calamites suckowii). Pictou Coalfield (Stellarton Basin, Nova Scotia): locality 2488 (two pieces without catalogue number). Bell (1940): locality 2684 (GSC 10741 + GSC 10803).

133 OCCURRENCE IN THE UNITED STATES. Alabama: Lesquereux (1879–1880). Arkansas: Lesquereux (1860), Lesquereux (1879–1880), Crookall (1966).

134 DESCRIPTION. Ribs 8–15 mm broad, slightly convex, separated by slightly undulate furrows. Leaf scars widely separated (13–20 mm apart), occupying half to two thirds of rib width; subhexagonal, with a notch in the upper margin, and well rounded lower margin. Dimensions: 5–12 mm long and 5–9 mm broad; ratio = 1–1.2. Vascular trace and parichnos markings in line and situated slightly above the middle of the leaf scar; leaf trace punctiform to oval, and parichnos scars oval, vertically elongate or slightly arched. Ligule pit about 1 mm above the leaf scar. Intervening rib surfaces between leaf scars smooth or ornamented with short, transverse lines occupying the entire width of rib.

135 REMARKS. The single specimen from Springhill, Nova Scotia, figured by Bell (1944, pl. LIII, fig. 2) as Sigillaria scutellata, shows the slightly convex ribs as well as the characteristic ornamentation consisting of several transverse or oblique markings that occupy the entire width of ribs below the subhexagonal leaf scars. As usual for this species, the ornamentation is only faintly preserved. Bell (1966, pl. IX, fig. 1) also figured another, more poorly preserved specimen from locality 205 at Springhill. The latter is re-illustrated partially as Figs. 17a, b.

Bell (1944, pp. 92–93, pl. LIV, fig. 1; pl. LV, fig. 1) figured and described two specimens from locality 1338 (GSC 8557 and GSC 8555) as a new species, Sigillaria fundiensis. He noted the resemblance with Sigillaria scutellata, but recorded the absence of transverse ornamentation as distinguishing the former species from the latter. However, sigillarian stem remains were often deposited after floating in water for some time, thus often producing a partial decay of superficial tissues. Indeed, such imperfect preservation is rather common. Accordingly, Sigillaria fundiensis is regarded here as based on imperfectly preserved remains of Sigillaria scutellata.

Bell (1944) also compared Sigillaria fundiensis with Sigillaria nudicaulis, a species that may be synonymous with Sigillaria scutellata. In fact, the holotype of Sigillaria fundiensis is strikingly similar to specimens figured as Sigillaria nudicaulis by Crookall (1966, pl. XCI, figs. 4–5a) from Shropshire, England.

136 COMPARISONS. Sigillaria polyploca Boulay has more convex ribs separated by strongly undulate furrows. Its leaf scars are subpentagonal and each has a clearly rounded lower margin.

Sigillaria mamillaris has flexuous or zigzagging furrows between ribs, which are only 4–10 mm wide. Its leaf scars are in vertical rows and 2–6 mm apart; they are hexagonal, with straight lateral margins and a more or less arched lower margin.

137 STRATIGRAPHIC AND GEOGRAPHIC DISTRIBUTION. According to Josten (1991), in western Germany Sigillaria scutellata ranges from Langsettian to mid-Bolsovian, and only rarely beyond. From Great Britain, Crookall (1966) recorded the species (as Sigillaria scutellata and Sigillaria nudicaulis) as being fairly common in Westphalian A (Langsettian) and B (Duckmantian), but rare in Westphalian C (Bolsovian) and D (Asturian).

138 OCCURRENCE IN THE MARITIME PROVINCES, CANADA. Cumberland Basin (Nova Scotia): Bell (1944): locality 205 = 1081 (GSC 14936 + two pieces without catalogue number); GSC 9897; locality 1081 (GSC 8568); locality 1338 = 1686 (GSC 8555 — holotype of Sigillaria fundiensis + GSC 8557 — together with Sigillaria hexagona). Bell (1966): locality 205 (GSC 14936 — same as Bell, 1944). Falcon-Lang et al. (2004). Sydney Basin (Nova Scotia): Zodrow and McCandlish (1980).

139 OCCURRENCE IN THE UNITED STATES. Illinois: Janssen (1957). Indiana: Wood (1963). Michigan: Arnold (1934, 1947, 1949). West Virginia: Gillespie et al. (1978).

140 DESCRIPTION. Slightly convex ribs, smooth or showing a longitudinal striation in decorticated condition, 6–7 mm wide, separated by straight, deep furrows. Leaf scars about 40 mm apart in vertical rows, occupying two thirds of rib width, subhexagonal, with a notch in the upper margin, a rounded lower margin, and lateral margins prolonged downwards by two short, slightly raised lines. Dimensions: 4–5 mm long and 4–4.5 mm broad; ratio = 0.8–1. Cicatricules placed in the upper half of the leaf scar, the central (vascular trace) punctiform, and the two lateral ones (parichnos) oval. A plume is visible above the leaf scars.

141 REMARKS. Sigillaria subrotunda is more commonly recorded as Sigillaria deutschii (= Sigillaria deutschiana), a synonym. Two specimens from localities 1339 and 1982 (both in the Joggins section) are included in this rare species. These specimens (at GSC Ottawa) were unrecorded by Bell.

142 COMPARISONS. Sigillaria scutellata shows longer and broader leaf scars that are less widely separated vertically. Also, the surface between leaf scars in Sigillaria scutellata is smooth or ornamented with short, transverse lines.

Sigillaria ovata possesses leaf scars of similar size and that are also well separated vertically, but of ovate shape with rounded margins; furthermore, the rib surfaces are smooth and unornamented.

Sigillaria rugosa possesses more elongate, oval to pyriform leaf scars, which are separated by rugose areas in between the scars.

143 STRATIGRAPHIC AND GEOGRAPHIC DISTRIBUTION. Sigillaria subrotunda is rare. The type material of both Sigillaria subrotunda and Sigillaria deutschii is from Saarland, western Germany. Crookall (1966) recorded the species (as Sigillaria cf. deutschiana) from the Westphalian B (Duckmantian) of Great Britain. The specimens figured and described by Deltenre (1924–1926) originated from the Duckmantian of Belgium.

144 OCCURRENCE IN THE MARITIME PROVINCES, CANADA. Cumberland Basin (Nova Scotia): Previously undescribed material: locality 1339 (without catalogue number — together with Eusphenopteris sauveurii + Dorycordaites palmaeformis and Cyperites sp.); locality 1982 (GSC 8566 — with Dorycordaites palmaeformis and Cyperites sp.).

147 REMARKS. Bell (1944) figured two specimens as Gymnostrobus salisburyi. One of these (Bell 1944, pl. LXI, fig. 2) is an incomplete but apparently cylindrical strobilus, up to 190 mm long and 30 mm wide, with sporangia that are perpendicularly inserted, about 10 mm long; sporophylls are not clearly visible. We regard it as a sigillarian strobilus, but cannot assign it more precisely. The other specimen figured as Gymnostrobus salisburyi (Bell 1944, pl. LXII, fig. 4), is clearly a rhizomorph attributable to Stigmaria ficoides (see later).

148 OCCURRENCE IN THE MARITIME PROVINCES, CANADA. Cumberland Basin (Nova Scotia): Bell (1944): locality 1388 = 2990 (GSC 10112).

153 REMARKS. Bell (1944) figured as Lepidophyllum sp. some fragmentary lycopsid leaves, 40–50 mm wide and up to 100 mm long. The long, narrow, single-veined leaves (as described by Bell, but not clear from his illustrations) are commonly referred to as Cyperites (or its synonym Sigillariophyllum). The name Cyperites is applied to long leaves that occur in several genera of arborescent Carboniferous lycopsids (e.g., Sigillaria, Lepidodendron sensu stricto, Lepidophloios, Omphalophloios, Polysporia).

Snigirevskaya (1958) replaced Lepidophyllum, the name most commonly applied to leaves only a few centimetres long (in contrast to the much longer, narrow, and parallel-sided leaves of Cyperites) by Lepidophylloides. She explained that the name Lepidophyllum was preoccupied by a living angiosperm genus.

154 STRATIGRAPHIC AND GEOGRAPHIC DISTRIBUTION. Cyperites is common throughout the Carboniferous.

155 OCCURRENCE IN THE MARITIME PROVINCES, CANADA. Cumberland Basin (Nova Scotia): Bell (1944): locality 1685 (GSC 8207).

158 REMARKS. This common species is characterized by a laterally extensive rooting system of repeated equal dichotomies of main roots at four corners of the stem base. Apart from the specimen that clearly has helicoidally arranged circular scars and attached rootlets figured by Bell (1966, pl. IX, fig. 2), it is likely that one of the two remains figured as Gymnostrobus salisburyi by Bell (1944, pl. LXII, fig. 4) also represents Stigmaria ficoides. As noted above, the other specimen (Bell 1944, pl. LXI, fig. 2) is a Sigillariostrobus. Jongmans (1930) compared Gymnostrobus salisburyi with Stigmaria and it was listed as a synonym of Stigmaria ficoides by Chaloner and Boureau in Boureau (1967, p. 675). Specimens with a similar state of preservation were figured as Stigmaria ficoides by Kidston (1902, pl. LII, fig. 3) and Arber (1920, pl. XL, fig. 1; pl. XLI, fig. 4).

159 COMPARISONS. Stigmaria stellata is distinguished by the stellate pattern of ridges radiating from each of the rootlet scars.

Stigmaria evenii has smaller circular to oval rootlet scars, about 3 mm diameter, set among relatively short, less regularly disposed, undulating ridges.

160 STRATIGRAPHIC AND GEOGRAPHIC DISTRIBUTION. Stigmaria ficoides is widely distributed in both Mississippian and Pennsylvanian strata. Its stratigraphic value is restricted to the recognition that it is a Carboniferous plant fossil.

161 OCCURRENCE IN THE MARITIME PROVINCES, CANADA. Cumberland Basin (Nova Scotia): Dawson (1868). Bell (1944): locality 666 = 1141 (GSC 10277). Bell (1966): locality 1343 (GSC 14932). Falcon-Lang et al. (2004). Sydney Basin (Nova Scotia): Bell (1938): locality Cranberry Head (GSC 4400). Zodrow and McCandlish (1980). St. John (New Brunswick): Dawson (1871). Falcon-Lang (2006). Newfoundland: Bashforth (2005).

162 OCCURRENCE IN THE UNITED STATES. Arizona: Tidwell et al. (1992). Colorado: Jennings (1980). Illinois: Lesquereux (1870), Noé (1925), Janssen (1940, 1957), Jennings (1970). Indiana: Canright (1959), Boneham (1975). Iowa: Rothwell (1984). Kentucky: Rothwell and Pryor (1991). Michigan: Arnold (1934, 1947, 1949). Ohio: Abbott (1968), Cross et al. (1996). Pennsylvania: Lesquereux (1879–1880), Oleksyshyn (1982). Utah: Tidwell (1967). West Virginia: Lesquereux (1879–1880), Gillespie and Latimer (1960), Gillespie and Clendening (1962), Gillespie et al. (1966), Gillespie et al. (1978).

165 DESCRIPTION. Leaf cushions contiguous, spirally arranged, rhomboidal, longer than broad, with maximum width in the upper third, rounded lateral angles, and acute base and apex. Keel absent. Dimensions: 19–22 mm long and 10 mm broad; ratio = 1.9–2.2. Leaf scars rhomboidal, placed in the centre of the cushions or slightly above, with sharp lateral angles and more rounded upper and lower ones, displaying a relatively big, rounded cicatricule (leaf trace) in its upper part. Parichnos absent. Dimensions: 3–4 mm long and 5–6 mm broad; ratio ≈ 0.6.

166 REMARKS. One of the lycopsid stem fragments figured by Bell (1944, pl. LI, fig. 1 — upside down) as Lepidodendron jaraczewskii, shows protruding leaf scars situated a little above the central part of the rhombic leaf cushion. Bell recorded the apparent absence of a leaf trace and of transpiratory (parichnos) scars. However, since his illustration inadvertently combines a specimen of “Lepidodendron” feistmantelii with one of Omphalophloios anglicus (a species that he did not consider), his comment regarding the absence of a leaf trace seems to refer to the specimen attributable to ”Lepidodendron” feistmantelii rather than that referable to Omphalophloios. The latter displays a vascular trace in at least some of the leaf scars. Bell considered that the absence of leaf trace and parichnos was due to the scars being covered over by a coaly substance, but this explanation is unconvincing. Parichnos are clearly absent, not covered over. The total absence of parichnos is a notable character of Omphalophloios (Brousmiche- Delcambre et al. 1995).

Kidston (1901) synonymized Omphalophloios cyclostigma with Lepidodendron anglicus and transferred the latter to Omphalophloios. However, Kidston regarded Omphalophloios as a rhizomorph akin to Stigmaria, a suggestion that may have been based on the absence of parichnos. Cambier and Renier (1912) figured stem remains with attached leaves and identified (correctly) the published remains of Omphalophoios as subaerial parts of a lycopsid tree.

167 STRATIGRAPHIC AND GEOGRAPHIC DISTRIBUTION. The holotype of Omphalophloios anglicus is from upper Asturian strata of Somerset, Great Britain, where this species is very rare (according to Crookall 1966). Lesquereux and White’s specimens originated from Clinton, Missouri, probably lower Asturian. Belgian material (Cambier and Renier 1912) comes from the Langsettian of the Charleroi Basin.

168 OCCURRENCE IN THE MARITIME PROVINCES, CANADA. Cumberland Basin (Nova Scotia): Bell (1944): locality 1337 (GSC 8712).

169 OCCURRENCE IN THE UNITED STATES. Missouri: Lesquereux (1879–1880), White (1898), White (1899).