Cretaceous (Albian–Aptian) Conifer Wood from Northern Hemisphere High Latitudes: Forest Composition and Palaeoclimate ⁎ M

Review of Palaeobotany and Palynology 143 (2007) 167–196 www.elsevier.com/locate/revpalbo

Cretaceous (Albian–Aptian) conifer wood from Northern Hemisphere high latitudes: Forest composition and palaeoclimate ⁎ M. Harland a, , J.E. Francis a, S.J. Brentnall b, D.J. Beerling b

a School of Earth and Environment, University of Leeds, LS2 9JT, UK b Department of Animal and Plant Sciences, University of Sheffield, S10 2TN, UK Received 19 January 2006; received in revised form 8 July 2006; accepted 16 July 2006 Available online 19 October 2006

Abstract

Permineralised conifer wood is abundant in Cretaceous (Albian–Aptian) sediments in high northern latitudes (N60°). The wood provides evidence of conifer-dominated forests that extended across the northern polar regions during greenhouse periods when the climate was warmer than today. This study investigates the composition of the Cretaceous (Albian–Aptian) high latitude Northern Hemisphere conifer forests using wood from Spitsbergen, and Ellesmere and Axel Heiberg islands in the Canadian Arctic Archipelago. Identification of the fossil woods indicates that the conifers included Pityoxylon, Piceoxylon, Laricioxylon, Protopi- ceoxylon, Palaoepiceoxylon, Taxodioxylon, Juniperoxylon, Protocedroxylon, Araucariopitys, Xenoxylon, Cupressinoxylon and Taxaceoxylon. This study shows that Spitsbergen was dominated by Taxodioxylon (25%) and in the Canadian Arctic Pityoxylon (33%) was dominant. Climate analysis of the conifers indicates that the northern Cretaceous (Albian–Aptian) forests of Svalbard grew in moist cool upland areas with warmer temperate areas in the lowlands, probably with rivers and/or swampy areas present. The forests of the Canadian Arctic probably grew under slightly cooler conditions than on Svalbard, similar to northern Canada today. © 2006 Published by Elsevier B.V.

Keywords: Cretaceous; fossil wood; Spitsbergen; Canadian Arctic; conifer; palaeoclimate

1. Introduction nuous sunlight (Creber and Chaloner, 1984; Chaloner and Creber, 1990; Vakhrameev, 1991; Spicer et al., 2002; Fossil evidence shows that during greenhouse periods Skelton, 2003). There are no modern analogues for these the climate was warm enough in the high northern forests in the Northern Hemisphere today, as the present latitudes (N60°) to allow forests to flourish under con- tree line does not extend far beyond 70°N, primarily ditions of elevated CO2, even under the unusual light determined by wind, temperature, moisture and land regime of long dark winters and summers with conti- availability (Krebs, 1972; Wilmking and Juday, 2005). Little work has been undertaken on the Cretaceous ⁎ Northern Hemisphere forests but a few reports suggest Corresponding author. Now at CASP, University of Cambridge, that conifers were a dominant element of the Northern West Building, 181a Huntingdon Road, Cambridge, CB3 0DH, UK. Fax: +44 1223 276606. Hemisphere forest communities (Arnold, 1953; Bannan E-mail address: [email protected] (M. Harland). and Fry, 1957; Spicer and Parrish, 1986; Spicer, 2003).

For example, the floras of Albian to Cenomanian age ceous floras of Spitsbergen contained the conifers Ar- from the Grebenka River region of northeastern Russia aucarites and Podozamites (Harland, 1997). and from the North Slope of Alaska appear to have been Polar forests potentially exerted a significant effect very similar. These floras were highly diverse, contain- on climate dynamics at both regional and global scales, ing early successional communities dominated by due to feedback processes such as albedo, the land Equisetites and Birisia ferns (Spicer and Herman, surface heat budget, and hydrological and carbon cycles 2001; Spicer et al., 2002) with mature stands rich in (Foley et al., 1994; DeConto et al., 2000; Beringer et al., conifers of Araucarites, Pagiophyllum, Pityophylum, 2005). Recent studies have shown that polar vegetation Podozamites, Cephalotaxopsis and Sequoia, particular- may have been critical for maintaining high latitude ly by the early Cenomanian in drier areas (Parrish and warmth, through feedbacks such as evapotranspiration Spicer, 1988; Spicer and Parrish, 1990; Spicer and and carbon cycling, without concomitant warming of Herman, 2001; Spicer et al., 2002; Spicer, 2003). Al- the lower latitudes (Foley et al., 1994; Upchurch et al., though less diverse, the forests of Axel Heiberg and 1998; DeConto et al., 2000). Therefore it is important to Amund Ringness islands in the Canadian Arctic appear understand the distribution and diversity of these forests to have been dominated by the conifers Cedroxylon and to allow them to be accurately included within vegeta- Piceoxylon (Bannan and Fry, 1957). The Early Creta- tion models for coupling to climate models.

Fig. 1. Geological setting of the Canadian Arctic Archipelago specimens. a) Present day map of Canada showing the location of Ellesmere and Axel Heiberg islands and the Sverdrup Basin in light grey (after Basinger, 1991). b) Enlarged map showing the location of the specimen collection areas (⁎). RR = Roll-rock valley, E = Eureka and BL = Buchanan Lake. c) Sedimentary log of Aptian–Albian Christopher Formation (compiled from Patchett et al., 2004; Hall et al., 2005). M. Harland et al. / Review of Palaeobotany and Palynology 143 (2007) 167–196 169

This paper focuses on Northern Hemisphere Creta- The fossil wood from the Canadian Arctic is pre- ceous (Albian–Aptian) fossil wood of Ellesmere and Axel served in the Sverdrup Basin. During the mid-Aptian rift Heiberg islands on the Canadian Arctic Archipelago and activity began (Ellesmere and Axel Heiberg islands; Spitsbergen, the main island of Svalbard. The biodiversity Patchett et al., 2004; Fig. 1a and b) and basaltic vol- and climatic interpretations are discussed. The results canism associated with the rifting led to the deposition have provided information about the polar forests to of thick, coarse, non-marine deposits throughout the provide a proxy for the verification of a new vegetation Sverdrup Basin (Fig. 1a). This was followed in the mid- model called the University of Sheffield Conifer Model Aptian and Albian by a marine transgression that (USCM, Brentnall et al., 2005). produced dark laminated mudstones of the Christopher Formation (Patchett et al., 2004; Fig. 1c) in which the 2. Geological background fossil wood was found. Some plate tectonic reconstructions for the early FossilwoodinboththeCanadianArcticand Mesozoic suggest that Svalbard and Ellesmere Island Svalbard was preserved within black mudstones that were close to each other, with Svalbard forming the represent deep marine conditions. The wood represents eastern extension to the Sverdrup Basin (Worsley et al., trees that lived on nearby land but were then washed as 1986). During the Early Cretaceous in the south of driftwood into adjacent marine basins. Wood in both Spitsbergen there was a gradual transition from pro- regions is preserved by calcite mineralization. delta shales through delta front to fluvial dominated

Fig. 2. Geological setting of the Svalbard specimens. a) Present day map of Svalbard, the grey area is Jurassic/Cretaceous sediments (modified from Worsley et al., 1986). b) Enlarged map showing the location of the specimen collection areas (⁎) around Lundstromdalen and Storknausen Peak. c) Generalised section log for Lundstromdalen (J. Francis, unpublished data). 170 M. Harland et al. / Review of Palaeobotany and Palynology 143 (2007) 167–196 sequences (Worsley et al., 1986). Overlying these specimens within one morphogenus displayed differences sediments, and partially laterally equivalent, are marine in structure these were separated by using a morphogenera sandstones, siltstones and shales that suggest a relatively number (e.g. Type F1,TypeF2). open marine shelf setting that would have provided the In order for comparisons to be made between the carbonate for the formation of the concretions in which specimens used here and previous descriptions and to the fossil wood was found (Worsley et al., 1986). The aid identification, it was necessary to calculate the per- permineralised wood was collected from the southern centage of certain features present within the wood parts of the main island of Spitsbergen, from the structure. For example, bordered pits within a single Lundstromdalen and Storknausen areas (Fig. 2a and b). specimen may be uniseriate, biseriate or multiseriate and The fossil wood represents the remains of trees that the percentage of each type can aid in identifying the drifted from the land surface into a marine basin and specimen to morphogenus level. Therefore all quantita- then preserved within dark laminated mudstones of the tive data are given in Table 1. Aptian–Albian Carolinefjellet Formation (Fig. 2c). Although the fossil woods are Cretaceous in age, and thus direct modern affinities are unknown, palaeocli- 3. Material and methods mate is inferred from the fossil wood based on the climate tolerances of the living types, bearing in mind In total 23 fossil wood specimens were examined that the fossils may have had different ecological toler- from the Northern Hemisphere sites. Ten of these spe- ance in the Cretaceous. cimens are from the Canadian Arctic Archipelago (Ellesmere and Axel Heiberg islands). Thirteen speci- 4. Fossil wood descriptions and identification mens were examined from Spitsbergen. All of the samples are stored at the School of Earth and Environment, Coniferales University of Leeds, UK. Family: Pinaceae Lindley Standard thin sections (transverse, radial longitudinal Morphogenus: Pityoxylon Kraus 1870 and tangential longitudinal orientations) were prepared Type species: Pityoxylon succiniferum Kraus 1870 for all of the wood specimens. The thin sections were Material: This morphogenus is described from four examined under an Olympus BH-1 transmitted light specimens from Ellesmere Island (RR111, RR122, E139 binocular microscope and images captured using a Leica and E140). IM1000 imaging system. Description: Three samples are probably derived from Structural features identified as being present or absent trunk material due to their straight ring boundaries with within conifer wood allows identification to morphogenus large radii of curvature in transverse section (RR111, level (Kräusel, 1949; Barefoot and Hankins, 1982). RR122 and E140, Fig. 3a). Sample E139 appears to be However, fossil conifer wood taxonomy and nomencla- derived from branch material as indicated by uneven ture poses considerable problems, due to the conservative widths of individual growth rings (compression wood) nature of wood structure, despite the efforts of several and highly curved ring boundaries (Fig. 3b). All samples authors to clarify the situation (Schweingruber, 1990; display distinct growth rings (Table 1). The transition Philippe, 1993; Bamford and Philippe, 2001; Wheeler and from earlywood to latewood is conspicuous (E139 and Lehman, 2005). Comprehensive keys are rare or rely on RR122) or gradual (E140 and RR111). Vertical resin the observation of fine detail. Natural variation within any canals may be present in all samples, however these particular genus or within the wood from different parts of samples have been subjected to intense fungal attack a single tree needs to be accounted for (Phillips, 1941; and these may be an artefact of that attack. Radial walls Jefferson, 1982; Schweingruber, 1990). The fossil wood of tracheids bear predominantly uniseriate bordered pits in this study was identified using published wood with some biseriate forms also present (Table 1 and identification keys and fossil wood descriptions (e.g. Fig. 3c). Where pits are biseriate they are oppositely or Phillips, 1941; Kräusel, 1949; Greguss, 1955; Shilkina, alternately arranged and the pits are generally touching 1967; Roy and Hills, 1972; Barefoot and Hankins, 1982; but some are spaced (Table 1). Rays in tangential section Stewart, 1983; Meijer, 2000; Fairon-Demaret et al., 2003; are mostly uniseriate but a few multiseriate forms are Falcon-Lang, 2005). No attempt was made to identify the also present. Several ray cells contain resinous material wood to species level as it was felt that this was not and a few appear to be pitted (Fig. 3d). All samples possible in several genera where there are only subtle appear to contain horizontal resin canals, ranging in size differences in wood structure between species, even in from ∼18 to ∼42 μm wide, most display multiseriate modern wood (Phillips, 1941). However, where several fusiform rays but again this may be an artefact of the Table 1 Summary of quantitative data collected from Northern Hemisphere fossil wood Pityoxylon Piceoxylon Laricioxylon Protopiceoxylon Palaepiceoxylon E139 E140 RR111 RR122 RR102 RR113 LD105 LD123 LD126 LD102 RR121 RR123 BL125

N° of rings present 49 29 63 49 6 43 13 49 46 5 37 22 59 167 (2007) 143 Palynology and Palaeobotany of Review / al. et Harland M.

Bordered pit type (%) Uniseriate 75 66 90 90 49 45 78 63 64 93 55 49 58 Biseriate 25 34 10 10 47 50 22 37 36 7 44 50 40 Triseriate 0 0 0 0 4 5 0 000112

Bordered pit arrangement (%) Opposite 93 97 74 68 86 69 100 85 69 100 91 91 76 Alternate 7 3 26 32 14 31 0 15 31 0 9 9 24 Touching 82 61 80 70 80 73 36 96 97 85 64 60 88 Spaced 18 39 20 30 20 27 64 4 3 15 36 40 12 Ray height 1–12 1–15 1–15 1–14 1–28 1–19 1–31 1–27 1–30 2–26 1–49 1–37 1–21 Mean ray height 5 5 5 5 9 6 11 7 9 9 14 13 7

Type F1 (Taxodioxylon) Type F2 (Taxodioxylon) Juniperoxylon Protocedroxylon Araucariopitys Xenoxylon Taxaceoxylon Cupressinoxylon LD131 LD129 LD133 LD101 LD120 LD108 LD130 SN25 4 LD132 E137 N° of rings present 21 16 12 6 35 36 11 20 27 77

Bordered pit type (%) Uniseriate 66 80 92 94 49 85 32 64 79 90 Biseriate 31 13 8 6 46 15 55 34 19 8 Triseriate 3 7 0 0 5 0 13 2 2 2 –

Bordered pit arrangement (%) 196 Opposite 62 49 25 100 48 95 13 81 57 71 Alternate 38 51 75 0 52 5 87 19 43 29 Touching 88 100 31 80 98 92 100 82 72 81 Spaced 12 0 69 20 2 8 0 18 28 19 Ray height 1–11 1–13 1–18 1–14 1–25 1–10 1–41 1–11 1–12 1–11 Mean ray height 4 4 7 4 7 4 19 5 5 4 171 172 M. Harland et al. / Review of Palaeobotany and Palynology 143 (2007) 167–196

Fig. 3. Pityoxylon. a) Transverse section showing straight, evenly spaced ring boundaries (RR111). b) Transverse section showing rings pinching out toward the left hand side (compression wood, highlighted by the black arrow) (E139). c) Radial section showing uniseriate bordered pits (RR111). d) Tangential section showing uniseriate rays, the arrow indicates a pitted ray cell (E140). e) Tangential section showing a horizontal resin duct (E140). f) Radial section showing ray tracheids (E139). preservation style (Fig. 3e). The rays are 1–15 cells high, lining resin ducts thin or thick, in cross-field 1–6pinoidor mean of 5 (Table 1). Cross-field pits are taxodioid, 1–4 piceoid or taxodioid pits, wood parenchyma present piceoid and pinoid types, arranged singly, one in each or absent, rays 1–45 cells high and uniseriate (Phillips, corner of the cross-field, or randomly arranged. Ray 1941; Greguss, 1955). The names Pinoxylon, Pinuxylon tracheids appear to be present in samples E140, E139 and and Piceoxylon have been used on the basis that it was RR122 (Fig. 3f). The horizontal walls of ray cells are thin possible to separate these morphogenera using their (compared to the vertical walls of tracheids) to slightly anatomical characteristics (Gothan, 1905). As the samples thickened (RR111) with smooth walls, although some described here are poorly preserved and the presence of pitting also present in places (RR111, RR122 and E139). resin ducts questionable, allocation to Pinuxylon, al- Spiral thickenings and septa are absent. though suspected, cannot be definite. The samples have Identification: These samples were initially identified as therefore been placed in the morphogenus Pityoxylon having affinities with the morphogenus Pinuxylon on the which refers to samples that exhibit anatomical character- basis of displaying the following characteristics: growth istics similar to modern Pinus, Picea, Larix, Pseudot- rings distinct, resin ducts present, walls of epithelial cells suga, and some other Abietineae (Seward, 1919). These M. Harland et al. / Review of Palaeobotany and Palynology 143 (2007) 167–196 173 samples are similar to Pityoxylon ruffordi Seward in uniseriate, biseriate or rarely triseriate bordered pits having bordered pits in either a single or double row, (Table 1 and Fig. 4d). Where multiseriate, they are op- where multiseriate generally oppositely arranged with 2– positely arranged with crassulae present (Table 1 and 4 cross-field pits although the rays in the species are Fig. 4e). The bordered pits are generally circular but higher than observed here reaching up to 30 cells (Seward, occasionally flattened along the contact when biseriate 1919). Pityoxylon ranges from the Jurassic to Recent and they are spaced or touching (Table 1). Resin ca- (Seward, 1919). Samples of Pityoxylon wood have been nals were observed in radial section in sample LD105 described from the Early Cretaceous of Sweden, the mid- (Fig. 4f). Rays seen in tangential section are generally Cretaceous of New Jersey and Staten Island, USA and the uniseriate, although some are biseriate (RR102 and United Kingdom and Late Cretaceous samples from RR113) or triseriate (RR102). Rays are 1–19 cells high Japan and France (Seward, 1919). As these samples were with mean values of between 6 and 11 cells (Table 1). initially assigned to Pinuxylon they were also compared Where rays contain resin ducts they have fusiform mul- to previously described Pinuxylon. Pinuxylon has previ- tiseriate cells around the ducts (Fig. 4g). Horizontal ously been reported from Early Cretaceous sediments of resin ducts have thick-walled epithelial cells. Cross- Franz-Josef Land, Russian High Arctic, confirming the fields contain 1–6 (RR113 and LD105) or possibly up to presence of this genus within high latitude forests at this 8 (RR102) pits. Pits in all samples are piceoid and time (Shilkina, 1967). A possible cone (Pityostrobus taxodioid (Fig. 4h), mostly arranged 2 or 4 side-by-side milleri spp. nov.) from Hokodz River Basin, northwestern or 4 in the corners. The horizontal walls of rays cells are Caucasus of Russia of Aptian age may confirm the thin and well pitted (Fig. 4d). Ray tracheids are present presence of the Pinaceae in this region (Falder et al., with dentate walls occasionally displaying minute pro- 1998). jections (Fig. 4d). Xylem parenchyma is present in late- Comparison with extant wood: The fossil wood of this wood at ring boundaries (LD105) or scattered (RR102 morphogenus has similarities to wood of the family and RR113). Spiral thickenings were observed in sample Pinaceae, which consists of Abies, Keteleeria, Pseudot- RR113. Septa were absent in all samples. suga, Tsuga, Picea, Pseudolarix, Larix, Cedrus and Pi- Identification: These samples were identified as nus (Phillips, 1941; Greguss, 1955; Roy and Hills, 1972). having affinities with the Abietoideae form of Pinaceae Of the Pinaceae Cedrus, Abies, Keteleeria, and Pseudo- (having highly pitted horizontal walls) due to the lack of larix all lack resin canals or only produce occasional typically “pinoid” cross-field pits and thin-walled epi- traumatic ducts. Picea, Pseudotsuga and Larix do pro- thelial cells distinctive of Pinus. In erecting the genus duce normal vertical and horizontal resin canals sur- Piceoxylon Gothan (1905) included characteristics of rounded by thick-walled epithelial cells. In Pinus the cells extant genera Picea, Larix and Pseudotsuga. Kräusel surrounding resin canals are thin-walled. As the samples (1949) continued to use the genus in this sense, as have examined here are poorly preserved with the presence of authors such as Bannan and Fry (1957).AsPseudot- resin canals questionable, it is not possible to place them suga may be readily distinguished from Picea and Larix with certainty within any particular genus based on resin by the presence of well developed spiral thickenings it is canal presence and epithelial cell types. Therefore it is felt that Roy and Hill's (1972) emended diagnosis only possible to refer to these samples as similar to mod- restricting Piceoxylon to include structural variability of ern Pinaceae. the wood of Picea and Larix only is more appropriate to Morphogenus: Piceoxylon Gothan, 1905 em. Roy and use here. The presence of distinct growth rings with Hills (1972) abrupt transition of earlywood to latewood, vertical and Type species: Picea beaufortense Roy and Hills (1972) horizontal resin ducts with thick-walls, tangential and Material: Three samples of this morphogenus were radial walls of vertical tracheids without spiral thicken- identified. One sample was from Spitsbergen (LD105) ings or checkings, radial walls with uniseriate or bi- and two from the Canadian Arctic (RR102 and RR113). seriate and opposite pits, crassulae present, cross-field Description: Two specimens are probably derived from pits generally piceoid, occasionally taxodioid and xylem mature stem material (LD105 and RR102, Fig. 4a) and parenchyma restricted to ring boundaries place samples one specimen branch material due to the presence of LD105, RR102 and RR113 within this amended diag- compression wood (RR113, Fig. 4b). Growth rings are nosis. These samples were compared to other previously distinct. The transition from earlywood to latewood is described Piceoxylon samples and similarities were noted abrupt (LD105 and RR113) or gradual (RR102). Vertical with P. christopheri and P. thomsonii (Bannan and Fry, resin canals with thick-walled epithelial cells are present 1957). Piceoxylon of Albian age has been reported by in all samples (Fig. 4c). The radial walls of tracheids bear Bannan and Fry (1957) from the same formation 174 M. Harland et al. / Review of Palaeobotany and Palynology 143 (2007) 167–196

Fig. 4. Piceoxylon. a) Transverse section of trunk wood showing straight, evenly spaced ring boundaries and resin canals (highlighted by black arrows, LD105). b) Transverse section of branch wood showing compression wood (white arrow). c) Thick-walled vertical resin canals in transverse section (LD105). d) Radial section showing uni- and biseriate bordered pits, ray tracheids (black arrow) and well pitted thickened horizontal walls of ray cells (LD105). e) Radial section showing opposite and alternately arranged biseriate bordered pits with crassulae (white arrow, LD105). f) Resin canals in radial section of sample LD105. g) Thick-walled horizontal resin canal in tangential section (LD105). h) Radial section showing piceoid pit (RR102). M. Harland et al. / Review of Palaeobotany and Palynology 143 (2007) 167–196 175

(Christopher Formation) on Amund Ringnes Island in the latewood is conspicuous (Table 1 and Fig. 5a). Vertical Canadian Arctic. Piceoxylon continued to be present in resin canals with thick-walled epithelial cells are present the Campanian–Maastrictian interval on Ellesmere Island (Fig. 5b). The canals sometimes contain resinous deposits when it appears to become more dominant (Falcon-Lang also seen in the transverse tracheids (particularly in et al., 2004). the latewood cells). The radial walls of tracheids bear Comparison with extant wood: Of the 8 genera of uniseriate or biseriate bordered pits (Table 1 and Fig. 5c). Abietoideae, Cedrus, Abies and Tsuga can be immedi- Where pits are biseriate they are mostly oppositely ar- ately disregarded because normal vertical and horizontal ranged but some are alternate with most touching but with resin ducts are absent from these genera (Phillips, 1941; occasional spaced forms (Table 1). Rays seen in tangential Greguss, 1955). Likewise, Keteleeria and Pseudolarix section are generally uniseriate with a few multiseriate. In may be discounted as they do not produce normal resin sample LD123 multiseriate, fusiform rays occur around ducts, only occasional traumatic cysts (Phillips, 1941). resin ducts (Fig. 5d). Rays are 1–30 cells high with mean The remaining genera, Pseudotsuga, Picea and Larix, of7forLD123and9inLD126(Table 1). In radial section share the common characteristics of horizontal and ver- cross-field pits are cupressoid or piceoid with 1–4pitsper tical resin ducts which have thick-walled epithelial cells cross-field arranged either in one row side-by-side, singly (Greguss, 1955). However, Pseudotsuga may be disre- or one pit in each corner (Fig. 5e). Ray tracheids may be garded in this instance because well developed, closely present, though sparse (Fig. 5f). The horizontal walls of spaced spiral thickenings were not observed in the ray cells appear to be thick and whilst in some places they samples described here. There are many common char- are smooth in others they are well pitted. Spiral thicken- acteristics shared by Larix and Picea, making differen- ings are present in sample LD126 (Fig. 5c) but septa are tiation difficult unless the preservation is exceptional. absent from both samples. However, Greguss (1955) suggested these genera may Identification: As LD123 and LD126 show character- be separated using several diagnostic features. For Pi- istics of both Picea and Larix they could be placed in the cea there are 8–10 epithelial cells surrounding the resin restricted genus Piceoxylon Gothan (1905) em. Bannan ducts, only occasionally paired bordered pits, walls of and Fry (1957). However, since they contain a relatively transverse tracheids minutely dentate and the presence high proportion of latewood cells and display character- of 2–6 generally piceoid pits per cross-field. In Larix istics most like extant Larix wood they are placed in the there are 10–14 epithelial cells surrounding the resin genus Laricioxylon Greguss 1967. The samples de- ducts, fairly frequently paired bordered pits and scribed here bear some resemblance to L. jarmolenkoi crassulae, walls of transverse tracheids mostly smooth (Blokhina, 1985) in the form of the bordered pits, the and 2–6(8–10) piceoid or taxodioid cross-field pits. presence of thick-walled epithelial cells and having Taking all of these characteristics into account, plus the rays up to 30 cells high, however the only sample LD126 arrangement of the cross-field pits, these samples appear has spiral thickenings similar to those described by to have characteristics of both genera. However, the Blokhina (1985). The type specimen (holotype) of Lar- proportion of latewood in these samples is low and the icioxylon is missing. No Laricioxylon appear to have resin ducts appear to be placed centrally within the rays, been reported previously from the Cretaceous. However, indicating a stronger affinity with Picea (Barefoot and it may be present but included in previous papers within Hankins, 1982). Extant Picea occur in temperate and the generalised name Piceoxylon. A cone with affinities colder regions up to the Northern Hemisphere limit of to extant Larix has been reported from the Early the Arctic Circle and to the southern limits of Taiwan Cretaceous, Shahai Formation, Yixian, Liaoning Prov- and Mexico, although most species occur in China ince, China (Shang et al., 2001), although this also has (Vidakovic, 1991). In warmer regions (subtropical) Pi- similarities to extant Picea and Pseudotsuga. cea occurs in high mountains e.g. Tibet and Assam in Comparison with extant wood: Sample LD123 contains the Himalaya (Vidakovic, 1991). both vertical and horizontal resin ducts that appear Morphogenus: Laricioxylon Greguss 1967 normal and not traumatic. Therefore Cedrus, Abies, Type species: Laricioxylon nogradense Greguss 1967 Keterleeria and Pseudolarix may be discounted, as Material: Two samples from Spitsbergen (LD123 and these all lack resin canals or only produce occasional LD126) were identified as belonging to this morphogenus. traumatic ducts. Pseudotsuga may also be discounted as Description: Both specimens are probably derived from this genus always displays well developed, closely branch material as indicated by the presence of compres- spaced spiral thickenings in its longitudinal and trans- sion wood and the curvature of ring boundaries. Growth verse tracheids which were not observed in these fossil rings are distinct and the transition from earlywood to samples. The difficulty of differentiating between Picea 176 M. Harland et al. / Review of Palaeobotany and Palynology 143 (2007) 167–196

Fig. 5. Laricioxylon. a) Transverse section of branch material showing compression wood marked by white arrow (LD133). b) Transverse section showing vertical resin canals (LD126). c) Radial section showing uniseriate, biseriate alternate (bottom white arrow) and opposite (top white arrow) bordered pits and spiral thickenings (right arrow) (LD126). d) Tangential section showing horizontal resin canal (LD123). e) Radial section showing cupressoid cross-field pit (LD123). f) Radial section showing ray tracheids (LD123).

and Larix then arises. LD123 and LD126 show similari- Type species: Palaeopiceoxylon transiens (Shimakura ties to both, however, the large proportion of latewood 1937) Kräusel (1949) and the frequency of paired bordered pits on radial Material: Only one specimen of this type was found, walls of tracheids indicates a stronger affinity with from Axel Heiberg Island (BL125). Larix (Greguss, 1955). Extant Larix is found in colder Description: This sample is probably derived from a regions of the Northern Hemisphere, occurring in North mature stem due to its straight ring boundaries and America, Asia and Europe (Vidakovic, 1991). Larix large radii of curvature in transverse section (Fig. 6a). grows mainly in mountainous regions between 200 and The rings do not contain compression wood but are 2400 m, although most prefer moist to boggy sites. evenly spaced along their length. This sample has 59 Family: ?Protopinaceae distinct growth rings and the transition from earlywood Morphogenus: Palaeopiceoxylon Kräusel (1949) to latewood is gradual. Vertical resin canals are present M. Harland et al. / Review of Palaeobotany and Palynology 143 (2007) 167–196 177

Fig. 6. Palaeopiceoxylon BL125. a) Transverse section showing straight ring boundaries and numerous resin ducts. b) Transverse section showing vertical resin duct. c) Radial section showing uni- and biseriate alternate and oppositely (black arrow) arranged bordered pits. d) Radial section showing resin ducts. e) Tangential section showing poorly preserved resin ducts. f) Radial section showing taxodioid cross-field pits. g) Radial section showing well pitted horizontal ray cells. with thick-walled epithelial cells (Fig. 6b). The radial may be crassulae present but preservation is poor and walls of tracheids bear uniseriate, biseriate and rare this cannot be definitely confirmed. Where pits are triseriate bordered pits (Table 1 and Fig. 6c). There multiseriate they are mainly arranged oppositely, 178 M. Harland et al. / Review of Palaeobotany and Palynology 143 (2007) 167–196 although some alternately arranged pits are also present bordered pits, with rare triseriate forms (Table 1 and (Table 1). The pits are generally touching with some Fig. 7c). Where pits are multiseriate they are arranged spaced (Table 1). Resin ducts are also present in radial alternately or oppositely sometimes with crassulae pres- section (Fig. 6d). Rays seen in tangential section are ent (Table 1). The majority of pits are touching, although generally uniseriate although some are biseriate. Some a few are spaced more than one pit diameter apart cells contain resinous material and walls are pitted. (Table 1). Rays seen in tangential section are mainly Horizontal resin canals with thick-walled epithelial uniseriate, although a few are biseriate. The rays are 1– cells are present (Fig. 6e). Rays are 1–21 cells high 25 cells high, mean of 7 cells (Fig. 7d). In radial section with a mean of 7. Preservation is poor, however, it cross-fields appear to contain 1–6, possibly 7, pits per appears there are 1–9 pits per cross-field either ar- field. The pits are arranged in various orientations in- ranged side-by-side, vertically or randomly. The pits cluding a single large pit per field, 6 organised in 2 appear to be piceoid, taxodioid and possibly with some horizontal rows and 4 arranged side-by-side. Most are pinoid (Fig. 6f). It cannot be determined whether ray taxodioid but there are also some cupressoid forms tracheids are present or not. The horizontal walls of ray (Fig. 7e). Horizontal ray walls are thick and very smooth cells are well pitted (Fig. 6g). Spiral thickenings and to well pitted (abietinaceous). Ray tracheids may be septa are absent. present although the preservation is poor. Xylem paren- Identification: Comparison with Kräusel's (1949) chyma is present. Spiral thickenings, septa, normal and scheme shows this specimen is similar to Palaeopi- traumatic resin canals are absent. ceoxylon due to the presence of the following Identification: Comparison to Kräusel's (1949) scheme characteristics: pits on radial walls of tracheids at least indicates that LD120 should be included within the in part circular; if arranged in several rows opposite, morphogenus Planoxylon due to its alternately arranged mainly separated by bars of Sanio, pits on radial walls of bordered pits, numerous small cross-field pits and pitted tracheids mainly of mixed type with all possible grada- horizontal walls (abietinaceous pitting). However Pla- tions of structure, horizontal and tangential walls of noxylon has since been separated from Protocedroxylon medullary ray cells very densely pitted (abietinaceous), based on the difference in tracheal pitting (Medlyn and resin canals in normal wood and horizontal and vertical Tidwell, 1986). Planoxylon has been retained for wood resin canals present. Palaeopiceoxylon appears to having typical araucarian pitting and vertical pairs of be representative of the Pinaceae, however, it retains pits in cross-fields, which does not match LD120. characters of pitting that indicate its affinities with Sample LD120 combines features of both araucarian Araucariaceae and Cordaitales, making it an intermedi- and abietinaceous conifers, placing it within the ate form. It is suggested that Palaeopiceoxylon could be morphogenus Protocedroxylon (Medlyn and Tidwell, included within the Protopinaceae, the generic name 1986). Protocedroxylon transiens has been identified covering all wood types with affinities to Pinaceae but from the Early Cretaceous of King Charles Land and the also retaining characteristics of more ancestral forms Late Jurassic or Early Cretaceous of Spitsbergen (Shilkina, 1967). The only other record of Palaeopi- (Medlyn and Tidwell, 1986) but no other occurrences ceoxylon is from the Early Cretaceous of Franz-Josef could be found. Protocedroxylon and Araucariopitys Land by Shilkina (1967). This indicates that Palaeopi- have been considered as taxonomical synonyms (Bam- ceoxylon was present at that time but was probably only ford and Philippe, 2001). Eckhold (1923) suggested that a minor component of the forests. these genera were taxonomical synonyms however he Morphogenus: Protocedroxylon Gothan (1910) kept the younger name Protocedroxylon. This situation Type species: Protocedroxylon araucarioides Gothan remains unclear and both names are still in use (Jeffrey, (1910) 1907; Medlyn and Tidwell, 1986; Falcon-Lang and Material: Only one sample was identified as belonging Cantrill, 2000; Poole and Cantrill, 2001). This situation to this morphogenus, LD120 from Spitsbergen. is worsened because the type fossil of Araucariopitys Description: Sample LD120 is probably derived from has not been available for study, therefore it is argued branch material as indicated by the presence of growth that, until both sets of type material can be compared, no rings (compression wood) and pronounced curvature of final decision can be made (personal communication, the ring boundaries. Thirty five distinct growth rings are Philippe). It was decided that, as both names are still present (Fig. 7a). The transition from earlywood to widely used, both could be used in this study by placing latewood is conspicuous. Tracheid cells, particularly in specimens within the form-genus that was most the latewood, contain resinous material (Fig. 7b). The comparable. radial walls of tracheids bear uniseriate and biseriate Morphogenus: Protopiceoxylon Gothan (1907) M. Harland et al. / Review of Palaeobotany and Palynology 143 (2007) 167–196 179

Fig. 7. Protocedroxylon LD120. a) Transverse section showing copious compression wood. b) Transverse section showing resin filled tracheids. c) Radial section showing uni- and biseriate opposite (bottom arrow) and alternate (top arrow) bordered pits. d) Tangential section showing resin filled ray cells. e) Radial section showing cross-field with cupressoid pit.

Type species: Protopiceoxylon extinctum Gothan in LD102 this seems to be restricted to ring boundaries. (1907) Samples RR121 and RR123 contain traumatic resin Material: The morphogenus is described from one ducts (Fig. 8b). The radial walls of tracheids bear uni- sample from Spitsbergen (LD102) and two from the seriate, biseriate or rare triseriate bordered pits (Table 1 Canadian Arctic (RR121 and RR123). and Fig. 8c). Where pits are multiseriate they are pre- Description: All three samples are probably derived dominantly arranged oppositely, although a small num- from mature stems as indicated by the ring boundaries ber are arranged alternately (Table 1). The pits are with large radii of curvature in transverse section mainly touching with some being spaced more than one (Fig. 8a). There is no compression wood present. The pit diameter apart (Table 1). Rays seen in tangential number of rings varies from 5–37 (Table 1). The transi- section are generally uniseriate with a few biseriate or tion from earlywood to latewood is conspicuous. Tra- rare multiseriate forms. The rays are generally high, cheid cells appear to contain dark resinous material and with means of between 9–14 per sample (Table 1 and 180 M. Harland et al. / Review of Palaeobotany and Palynology 143 (2007) 167–196

Fig. 8. Protopiceoxylon. a) Transverse section showing the relatively straight ring boundaries and traumatic resin ducts (RR123). b) Transverse section showing detail of vertical traumatic resin ducts (RR121). c) Radial section showing uni- and biseriate opposite (indicated by white arrow) bordered pits (RR121). d) Tangential section showing ray cells (LD102). e) Tangential section showing poorly preserved horizontal traumatic duct (RR123). f) Radial section showing piceoid cross-field pit (black arrow) and well pitted horizontal walls (white arrow) (RR121). g) Radial sections showing taxodioid (black arrow) and cupressioid (white arrow) cross-field pits (RR121). h) Radial section showing ray tracheids (RR121).

Fig. 8d). Some ray cell walls appear to be pitted. preservation (Fig. 8e). Cross-fields appear to contain 1– Traumatic resin ducts may be present in samples RR121 8 pits arranged side-by-side or in 2 rows. The pits and RR123 but this could also be an artefact of the present in cross-fields appear to be of mixed type, M. Harland et al. / Review of Palaeobotany and Palynology 143 (2007) 167–196 181 including taxodioid, cupressoid, piceoid and pinoid points to note in Protopiceoxylon wood are the abietinac- types (Fig. 8f and g). Horizontal walls of ray cells are eous pitting on the tracheid walls, abundant xylem thin and well pitted (abietoid) and ray tracheids are parenchyma particularly in autumn wood, as seen in present (Fig. 8f and h). Xylem parenchyma is present all samples here, and the occurrence of vertical resin but spiral thickenings and septa are absent. canals, observed in samples RR112 and RR123. These Identification: Shilkina (1967) indicated that the Early specimens show strong similarity to P. extinctum Gothan Cretaceous Pinaceae still retain some shared structural in possessing abietinaceous characters, intermediate characters of their affinity with Araucariaceae and Cor- between Cedroxylon and Pityoxylon, but is also similar daitales. This suggests that Pinuxylon, Palaeopiceoxylon, to P. edwardsi Stopes in having mostly uniseriate, Cedroxylon and probably Keteleerioxylon should be contiguous bordered pits. Protopiceoxylon has been grouped together in the Protopinaceae due to their com- described from the Jurassic of King Charles' Land and mon ancestry, however it is felt that there are sufficient Spitsbergen by Gothan (1907, 1910) and on Franz-Josef distinguishing features that can be used to separate these Land by Seward (1919). genera. In the case of the samples described here it was Family: Taxodiaceae Endl ex K. Koch initially felt that they should be placed in the morpho- Morphogenus: Taxodioxylon Hartig 1848 em. Gothan genus Cedroxylon. The samples described here provide (1905) a general match for the morphogenus Cedroxylon when Type species: Taxodioxylon goeppertii Hartig 1848 compared to Kräusel's (1949) scheme on the basis of Two sub-types of Taxodioxylon appear to be represented the following characteristics: pits on radial walls of tra- by LD129, LD131 and LD133 therefore this section gives cheids circular and, where multiseriate, oppositely ar- the general identification of Taxodioxylon, before the sub- ranged, never typically araucarioid, normal wood without types are described individually, and discusses previous resin canals, spiral thickenings absent and abietinaceous occurrences in the fossil record before comparing these ray pitting present. Samples RR121 and RR123 bear samples with extant wood of Taxodium. similarities to C. disjunctum (Bannan and Fry, 1957). Identification: Taxodioxylon is identified by the presence LD102 also bears some similarity to C. disjunctum, of the following features: distinct growth rings, smooth however it is lacking the alternate radial pitting seen in this to sparsely pitted horizontal walls of ray cells, 1–30 cells species. It has been suggested that Cedroxylon is an high rays with generally uniseriate but occasionally 1–2 illegitimate late synonym of Tiloxylon,eventhoughitis biseriate in the body, 1–6 taxodioid to glyptostroboid widely used (Bannan and Fry, 1957; Philippe et al., 1999; pits in the cross-fields and the absence of resin ducts Bamford and Philippe, 2001). Bamford and Philippe and spiral thickenings (Greguss, 1955; Fairon-Demaret (2001) further suggest that Cedroxylon is not repre- et al., 2003). In the absence of any other evidence sented in the Jurassic–Early Cretaceous of Gondwana these woods are notoriously difficult to assess because and cannot be used for this time period. However, Ce- of overlapping characteristics (Fairon-Demaret et al., droxylon was reported from Jurassic sediments on 2003). Included within the Taxodiaceae are Athrotaxis, Spitsbergen by Gothan (1910),suggestingthatitwas Crytomeria, Cunninghamia, Glyptostrobus, Sequoia already present prior to the Cretaceous. The discussion (inc. Sequoiadendron), Taiwania, Taxodium and Scia- below shows that this name has also been used for Early dopitys (Phillips, 1941; Greguss, 1955). Recently, this Cretaceous samples. Cedroxylon of Albian age has been classification has been changed using a combined reported by Bannan and Fry (1957) from the Christopher morphological and molecular approach (Gadek et al., Formation of Amund Ringnes Island, Canadian Arctic 2000; Kusumi et al., 2000), which has placed Sciadopitys and is also represented in the Early Cretaceous of Franz- in the monotypic family of the Sciadopityaceae Luerss Josef Land (Shilkina, 1967). Seed cones of Cedroxylon (Fairon-Demaret et al., 2003). Sciadopitys is distin- have been found in the Albian–Cenomanian of the River guished from all other Taxodioideae by its lack of wood Kiya, Chulym–Yenisei Basin, Siberia and various other parenchyma and the presence of large, solitary, simple remains of late Senonian age from the Sym Suite, Sym pits (Phillips, 1941; Meijer, 2000).Thesefeaturesare River, Yilyui Basin, Siberia (Vakhrameev, 1991). How- different from those found in the fossil samples described ever, in light of the suggested illegitimacy of Cedroxylon here, therefore this genus is excluded from further in the Early Cretaceous it was decided that these samples consideration. Likewise Taiwania can be discounted as should be classified as Protopiceoxylon. Protopiceoxylon this genus displays only cupressoid cross-field pits and is described as being similar to Pityoxylon,exceptin not the taxodioid type seen in all other members of having only vertical resin canals in the normal wood this family (Phillips, 1941) and the fossils under (Seward, 1919). Edwards suggests that the most important consideration. 182 M. Harland et al. / Review of Palaeobotany and Palynology 143 (2007) 167–196

Northern Hemisphere Taxodioxylon wood has been Morphogenus: Type F1, Taxodioxylon c.f. gypsaceum reported from the Late Cretaceous (Campanian) Oldham (Göppert) Kräusel (1949) and (Maastrichtian) Edmonton formations (Ramanujam Material: Only one sample was used to describe this and Stewart, 1969; Ramanujam, 1971), both of Alberta, sub-type of the morphogenus Taxodioxylon, LD131 Canada. The Edmonton Formation has also yielded from Spitsbergen. cones and leaf impressions and compressions (Horrell, Description: Sample LD131 is probably derived from 1991). Taxodiaceae pollen of Aptian age has been a mature stem due to the straight ring boundaries. recovered from the Pechora Depression in Siberia There is no compression wood present, indicating that (Vakhrameev, 1991). Twigs of Sequoia and rare Glyp- this is trunk material. The preservation is poor. The tostrobus have also been recorded from the late Albian sample has 21 growth rings that are distinct with wavy and Cenomanian strata in the Southern Urals (Vakhra- boundaries and the transition from earlywood to late- meev, 1991). Cones of Taxodiaceae are recorded wood is conspicuous (Fig. 9a). The radial walls of tra- from Cenomanian Pâtût in Western Greenland (Seward, cheids bear mainly uniseriate bordered pits (Table 1 and 1926) and complete shoots from the Albian–Cenoma- Fig. 9b). Biseriate pits also occur and there are oc- nian Nanushuk Group from the North Slope of Alaska casional triseriate forms. Where pits are multiseriate (Spicer and Parrish, 1986). In Eastern Siberia leaf the majority are oppositely arranged; alternate arrange- impressions, shoots and cones of Sequoia and Glyptos- ment is less common (Table 1). Pitting is both spaced trobus have been found of late Albian–early Cenoma- and contiguous but contiguous pitting is more common nian age and various Taxodiaceae remains have been with only a few spaced more than one pit diameter reported from the Turonian–Santonian, Vilyui Basin apart (Table 1). Rays in tangential section are commonly (Nizhny Chyrimin Floral Complex and Sym suite) uniseriate but rare biseriate forms also occur, 1–11 and Khatanga Basin (Vakhrameev, 1991). Seeds belong- cells high, mean 4 cells. Many of the ray cells are ing to Glyptostrobus, Sequoia and Taxodium have filled with dark resinous deposits (Fig. 9c). In radial been recorded from European strata of Cenomanian section cross-fields contain 1–6 pits showing various age (Fairon-Demaret et al., 2003). Sequoia cones have arrangements; some contain 1 large pit whilst others also been reported from late Albian–Campanian strata have 2 in opposite arrangement, 3 diagonally arranged in the Chandler–Colville Region of Alaska (Smiley, or up to 6 in random orientation. The cross-field 1969; Spicer and Parrish, 1986; Vakhrameev, 1991). pits are dominantly taxodioid with rare glyptostroboid Twigs and shoots of Taxodium, Athrotaxis and Glyp- forms having very narrow borders, almost window- tostrobus are recorded from the Fort Union Forma- like (Fig. 9d). Horizontal walls of ray cells are thin tion of the USA (latest Late Cretaceous) and the and smooth. Xylem parenchyma is present but spiral Potomac Group of Puddledock, Virginia, USA (∼early thickenings/checkings are absent. There are septa pre- Albian) (Vakhrameev, 1991; Srinivasan, 1995). This sent (although not abundant, Fig. 9e) and rare ray evidence indicates that taxodiaceous trees were wide- tracheids. Resin canals and traumatic resin ducts are spread across the Northern Hemisphere throughout absent. the Cretaceous. Identification: This specimen was compared with Comparison with extant wood: LD131 resembles extant various species of Taxodioxylon and other taxodiaceous Sequoia in its thin and smooth horizontal walls of wood fossil woods (Phillips, 1941; Kräusel, 1949; Greguss, and ray parenchyma, uni-, bi- or triseriate radial tracheid 1955; Ramanujam and Stewart, 1969; Ramanujam, 1971; pitting and rare occurrences of thin-walled marginal ray Meijer, 2000; Fairon-Demaret et al., 2003)(Table 2). This tracheids. Notched bordered pits are common in Sequoia comparison indicates T. gypsaceum (Göppert) Kräusel but absent in this sample therefore it is discounted (Meijer, (1949) [syn. T. sequoianum (Mercklin) Gothan, 1905]is 2000). In the remaining specimens notched bordered pits most similar to LD131, due to there being in excess of 3 were not observed and rays are low therefore Sequoia pits per cross-field (taxodioid and glyptostroboid types), sempervirens can be discounted (Barefoot and Hankins, transverse walls of resin parenchyma mainly smooth 1982). Sequoiadendron is discounted for all specimens and traumatic resin ducts absent. Meijer (2000) also because in this genus bordered pits are always in a single mentions that ray heights can be low (1–8 cells), similar to vertical row even in wide tracheids (Meijer, 2000), which this sample. Because of poor preservation of the cross- does not compare with these fossil samples. Taxodium is field pits this material is referred to as Taxodioxylon c.f. mainly confined to the Northern Hemisphere in humid gypsaceum. temperate to subtropical regions of East Asia and North Morphogenus: Type F2 Taxodioxylon Hartig 1948 em America (Meijer, 2000). Gothan (1905) M. Harland et al. / Review of Palaeobotany and Palynology 143 (2007) 167–196 183

Fig. 9. Taxodioxylon LD131. a) Transverse section showing poorly preserved ring boundaries. b) Radial section showing uni (black arrow) and biseriate opposite (white arrow) bordered pits. c) Tangential section showing ray cells filled with dark resinous material. d) Radial section showing taxodioid cross-field pit. e) Radial section showing septa.

Material: This morphogenus sub-type is described from (Fig. 10d), although opposite pitting can be quite com- two samples from Spitsbergen LD129 and LD133. mon. Crassulae were observed in both samples (Table 1). Description: Sample LD129 (Fig. 10a) is probably Pitting is touching or spaced (Table 1). Rays observed in derived from branch material as indicated by compres- tangential section are generally uniseriate with only rare sion wood. Sample LD133 (Fig. 10b) is probably derived biseriate forms, 1–18 cells high with means of 4–7 cells from a small stem. Although the central rings show (Table 1 and Fig. 10e). Some cells contain dark resinous pronounced curvature, the outer rings are reasonably material. In radial section horizontal walls are thin and straight with no compression wood. Growth rings are unpitted, with ray tracheids, if present, being rare. The distinct in both samples (Fig. 10a and b). The transition number of pits per cross-field seems to vary from 1 to 2 in from earlywood to latewood is either abrupt (LD129) or LD129 (Fig. 10f) to 1–6 in LD133 (Fig. 10g). In both gradual (LD133). The radial walls of tracheids bear samples the majority of cross-field pits contain 1 large mainly uniseriate, occasionally biseriate or rarely tri- narrow bordered taxodiod pit. Occasional slit-like aper- seriate bordered pits (Table 1 and Fig. 10c and d). Where tures were observed in LD133. Xylem parenchyma is pits are multiseriate most are alternately arranged present, being abundant in the latewood in sample 184 M. Harland et al. / Review of Palaeobotany and Palynology 143 (2007) 167–196

Table 2 Salient features of different Cretaceous species of Taxodioxylon (after Meijer, 2000) Species Number of Nature of cross-field Traumatic resin Miscellaneous features cross-field pits pits ducts T. albertense (Penhallow) Shimakura 1937 1–3 Taxodioid/Cupressoid Absent Ray height 1–70 cells T. burgessii (Penhallow) Kraüsel (1949) 1–3 Taxodioid (always) Possible (vertical/radial) T. taxodii Gothan (1905) N3 Taxodioid Absent Parenchyma cross walls pitted T. gypsaceum (Göppert) Kraüsel (1949) N3 Taxodioid/Glyptostroboid Radial ducts Parenchyma cross walls smooth absent T. montanense (Torrey) Kraüsel (1949) N3 Taxodioid Radial ducts Parenchyma cross walls smooth present T. drummhellerense Ramanujam 2–4 Taxodioid/Glyptostroboid Parenchyma cross walls well and Stewart (1969) pitted T. antiquum Ramanujam 2–5 Taxodioid (single row) Ray height N80 cells, ray width and Stewart (1969) 1–4 cells T. cryptomerioides (Hartig) Gothan 1–4 Taxodioid/Cupressioid/ Absent Parenchyma walls smooth or (1905) Podocarpoid widely spaced pits T. multiseriatum Ramanujam 1–4 Taxodioid Absent Rays 2–70 cells high and Stewart (1969)b

LD129, but spiral thickenings, septa, resin canals and tiseriate forms were observed, 1–14 cells high, mean of 4 traumatic resin ducts are absent. cells. Many of the cells are filled with dark resinous Identification: Identification of taxodioid fossil wood deposits and tangential walls often show sieve-like puts emphasis on the types of cross-field pits present thickenings (Fig. 11d). In radial section horizontal walls (Kräusel, 1949; Fairon-Demaret et al., 2003) and on the of ray cells are thin or thickened and smooth to well form of the horizontal walls of the parenchyma cells pitted ray tracheids are present and dentate (Fig. 11e). (Fairon-Demaret et al., 2003). However, the poor preser- Tangential walls contain bead-like thickenings. Cross- vation of the cross-field areas in these samples makes it fields appear to contain 1–4 pits arranged singly or with difficult to determine these factors with certainty. There- one pit in each corner of the cross-field. The cross-field fore these samples have been placed in the general pits appear to be mainly cupressoid with some taxodioid morphogenus Taxodioxylon Hartig 1848 em Gothan, types also present (Fig. 11f). Xylem parenchyma is fre- 1905. quent, commonly with resin content but spiral thicken- Family: Cupressaceae nom.cons. ings/checkings are absent. Septa are present although Morphogenus: Juniperoxylon Houlbert 1910 in Lecoin- scattered, with poor preservation making them difficult tre em. Kräusel (1949) to recognize. Resin canals and traumatic resin ducts are Type species: Juniperoxylon turonense Houlbert 1910 absent. in Lecointre Identification: Comparison with Kräusel's (1949) Material: This morphogenus is present as one specimen scheme indicates that LD101 should be included within from Spitsbergen (LD101). Juniperoxylon due to the bordered pits being generally Description: LD101 is probably derived from branch circular and opposite, the lack of resin ducts, spiral material as indicated by the uneven widths present within thickenings and abundant parenchyma. The holotype individual growth rings (compression wood), its small of Juniperoxylon cannot be found and it has been sug- size and the pronounced curvature of the ring bound- gested that the holotype may display differences to aries. Six distinct growth rings are present (Fig. 11a). The the published description (personal communication, transition from earlywood to latewood is conspicuous. Philippe; Bamford and Philippe, 2001). Bamford and The tracheids are generally rounded so that intercellular Philippe (2001) consider that the name has been validly spaces frequently occur (Fig. 11b). The radial walls of published. Stewart (1983) indicates that it is extremely tracheids bear mainly uniseriate bordered pits, although difficult to distinguish between Cupressaceae and Taxo- occasional biseriate pits also occur (Table 1). Where pits diaceae from macrofossils, with determination resting are biseriate they are always oppositely arranged. Pits are on the decussate (Cupressaceae) or helical (Taxodia- mainly touching however pits spaced more than one pit ceae) arrangement of leaves. Stewart (1983) therefore diameter apart do also occur (Table 1). Rays seen in suggests a possible common origin for these families, tangential section are generally uniseriate but a few mul- with differentiation occurring in the Jurassic and the M. Harland et al. / Review of Palaeobotany and Palynology 143 (2007) 167–196 185

Fig. 10. Taxodioxylon. a) Transverse section of branch material showing compression wood (LD129). b) Transverse section showing relatively straight, evenly spaced ring boundaries (LD133). c) Radial section showing uniseriate bordered pits (LD133). d) Radial section showing biseriate alternate bordered pits (LD129). e) Tangential section showing ray cells (LD129). f) Radial section showing singly arranged glyptostroboid (top arrow) and taxodioid (bottom arrow) cross-field pits (LD129). g) Radial section showing multiple cross-field pits with arrow point to two side-by-side (LD133). emergence of modern genera in the Late Cretaceous Comparison with extant wood: Juniperoxylon is most and early Tertiary. However, Philippe (1994) des- similar to Juniperus, grouped within the Taxodiaceae, cribed a new species of Taxodioxylon, Taxodioxylon traditionally included within the Cupressaceae with lemoignei from the Oxfordian of France indicating Thujoideae and Cupressoideae (Greguss, 1955). Greguss that this family was well separated from the Cupressa- (1955) identifies an anatomically more or less indepen- ceae by the Middle Jurassic. Shoots are reported for dent group of Taxodioideae including Arceuthos, Cu- two species of Juniperus from the Late Cretaceous of pressus, Diselma, Fitzroya, Juniperus, Libocedrus and Canada (Stewart, 1983) and remains of middle Albian Pilgerodendron. Sample LD101 appears to be part of to Coniacian age of the Chandler–Colville and Kuk this group due to the presence of the following features: River regions of Alaska (Smiley, 1966, 1969). No other wood parenchyma is present, horizontal walls never descriptions of Juniperoxylon were found from the perfectly smooth, but varicose, slightly nodular or Cretaceous. thickened bead-like, ray cells likewise uneven, nodular, 186 M. Harland et al. / Review of Palaeobotany and Palynology 143 (2007) 167–196

Fig. 11. Juniperoxylon LD101. a) Transverse section showing compression wood. b) Transverse section showing interstitial spaces. c) Radial section showing uni- and biseriate alternate bordered pits. d) Tangential section showing resin filled cells. e) Radial section showing ray tracheid (top arrow) and smooth horizontal walls (bottom arrow). f) Radial section showing taxodioid cross-field pit.

or pitted, while tangential walls are commonly with Description: Sample E137 contains a lot of deformation bead-like thickenings. Transverse tracheids fit closely in transverse section making it difficult to determine together without intercellular spaces with only a few whether it is derived from branch or trunk material, exceptions (Wilson and White, 1986). Of the Northern however it has tentatively been classified as branch Hemisphere species Juniperus spp. can be distinguished (Fig. 12a). Sample LD132 is probably derived from a from others by this feature, having somewhat rounded small stem (Fig. 12b). Both samples have distinct growth tracheids, so that intercellular spaces are produced be- rings and the transition from earlywood to latewood is tween them at the cell corners (Phillips, 1941; Wilson conspicuous. There appears to be a high proportion of and White, 1986). Sample LD101 displays this feature, dark resinous cell contents present (Fig. 12b). The radial indicating a strong affinity with the Juniperoideae. Ju- walls of tracheids bear predominantly uniseriate bordered niperus is widely distributed in the Northern Hemi- pits, however a few biseriate and rare triseriate forms also sphere, mostly confined to temperate regions of Europe occur (Table 1 and Fig. 12c). Where the pits are multi- and northern regions of Asia, China, America and Africa seriate they are generally oppositely arranged although (Vidakovic, 1991). alternate forms also occur (Table 1 and Fig. 12d). The Family: Cupressaceae nom.cons. bordered pits are predominantly touching but a few are Morphogenus: Cupressinoxylon nom.cons. Göppert 1850 spaced (Table 1). There are resin spools present in the Type species: Cupressinoxylon gothanii Kräusel 1920 radial tracheids of E137 (Fig. 12e). Rays seen in tan- Material: This morphogenus was identified in one gential section are predominantly uniserate with rare sample from Ellesmere Island in the Canadian Arctic examples with biseriate cells in the body, again there is a (E137) and one from Spitsbergen (LD132). lot of dark resinous material filling cells (Fig. 12f). M. Harland et al. / Review of Palaeobotany and Palynology 143 (2007) 167–196 187

Fig. 12. Cupressinoxylon. a) Transverse section showing poor preservation (E137). b) Transverse section showing reasonably straight ring boundaries (LD132). c) Radial section showing uniseriate bordered pits (LD132). d) Radial section showing biseriate oppositely arranged bordered pits (LD132). e) Radial section showing resin spools (E137). f) Tangential section showing beaded tracheid walls (black arrow) and resin filled ray cells (LD132). g) Radial section showing cupressoid cross-field pit (E137). Radial section showing taxodioid (white arrow) and narrow bordered glyptostroboid cross-field pits (LD132).

Sample LD132 has bead-like thickenings on the Identification: Under the classification scheme of tracheid walls in tangential section (Fig. 12f). Rays Kräusel (1949) it is not clear whether these samples are between 1–12 cells high with mean values of 4–5 should be identified as Cupressinoxylon or Taxodioxy- (Table 1). The cross-fields are poorly preserved in both lon. Poole et al. (2001) suggest that the presence of samples but E137 appears to contain 1–2pitsin predominantly smooth tangential walls and cross-field themainbodyand4–5 in marginal cells, LD132 regions characterized by cupressoid pits define the mor- seems to contain 1–6 pits. In LD132 most pits are phogenus Cupressinoxylon but the samples here contain arranged singly although some have 2–3 pits arranged mixed types of pits and the tangential walls in LD132 are side-by-side and several were observed with 6 in tiers bead-like. However Falcon-Lang (2003) had difficulty irregularly arranged. Horizontal walls are thin and distinguishing the two genera as his samples also smooth. Ray tracheids are present in sample E137. contained both cupressoid and taxodioid cross-field Cross-field pits appear to be piceoid, cupressoid or pits therefore he classified his samples as Cupressinox- taxodioid (Fig. 12g and h). A few scattered septa are ylon/Taxodioxylon type. Sample E137 described here present in sample LD132. Spiral thickenings and resin has many similarities to the Cupressinoxylon described ducts are absent. by Poole et al. (2003) from King George Island, 188 M. Harland et al. / Review of Palaeobotany and Palynology 143 (2007) 167–196

Antarctica, although there are differences including the appear to be the only family which can have bead-like lack of taxodioid cross-field pits in their sample. Sample tangential walls. Cupressus occurs in warm moderate LD132 appeared to have many similarities to Glyptos- and subtropical regions in the Northern Hemisphere troboxylon Kräusel (1949). It was however felt that the including the eastern Mediterranean, the Himalaya, samples described here had more similarities with Cu- China and in America from Oregon to Mexico. pressinoxylon and they were therefore included in this Family: ?Extinct Genera of Uncertain Affinity morphogenus. Cupressinoxylon has been described as Morphogenus: Araucariopitys Jeffrey in Hollick et an illegitimate renaming of Retinodendron but Bamford Jeffrey (1907) et al. (2002) proposed that Cupressinoxylon be con- Type species: Araucariopitys americana Jeffrey (1907) served. Cupressinoxylon has been widely used in a broad Material: This morphogenus was identified from sample sense for a genus comprising all woods of Cupressaceae LD108 from Spitsbergen. type (Francis, 1983; Poole et al., 2001; Bamford et al., Description: Sample LD108 is probably derived from a 2002; Falcon-Lang, 2003) whilst Retinodendron has not mature stem. Thirty six distinct growth rings are present in been used for over a century. Therefore conservation this sample (Fig. 13a). The transition from earlywood to of Cupressinoxylon would avoid further confusion. latewood is conspicuous. The tracheids frequently contain Cupressaceae remains are present across the Northern resinous material (Fig. 13a). The radial walls of tracheids Hemisphere within Cretaceous sediments. Wood has bear mainly uniseriate bordered pits with a few biseriate been reported from Early Cretaceous sediments of Franz- pits also present (Table 1 and Fig. 13b). Where the pits are Josef Land by Shilkina (1967). Seward (1926) reported multiseriate they are mostly oppositely arranged with only twigs and leaves that may be attributable to Cupressus rare alternate forms (Table 1). The majority of the pits are (although they may also be of Libocedrus or Thuja) from touching with only rare spaced forms (Fig. 13c). Rays Skansen, Western Greenland. It is suggested that the seen in tangential section are generally uniseriate although Cretaceous age may be questionable as the remains were some are partially biseriate with 1–2 paired cells in the recovered from loose blocks on a scree slope where both body (Fig. 13d). Rays are 1–10 cells high, mean 4 cells, Cretaceous and Tertiary sediments are now known to which are frequently filled with resinous material. Cross- occur. Unspecified remains of Cupressaceae have been field areas appear to contain 1–3 pits, usually arranged found in mid Albian age sediments of Lena Province, side-by-side although when 3 they are often arranged in 2 Siberia and Coniacian to Maastrichtian age of the rows. The pits appear to be cupressoid with some possibly Chandler–Coleville region of Alaska (Smiley, 1969; pinoid (Fig. 13e). Horizontal walls are thin and unpitted. Vakhrameev, 1991). Xylem parenchyma is present, commonly with resin con- Comparison with extant wood: According to Barefoot tents. Spiral thickenings are also present. Septa, normal and Hankins (1982), resin plugs only occur in Arau- and traumatic resin canals are absent. caria, Callitris, Cupressus spp., Dacrydium, Fitzroya, Identification: Several modern species of Araucaria and Juniperus, Libocedrus, Podocarpus, Saxegotheae, Agathis have uniseriate radial tracheid pitting with some Thuja and Widdringtonia spp. Sample E137 seems to biseriate and alternately arranged pitting also present. As fit best with Greguss' (1955) description of Cupressus: previously discussed it is suggested that there are growth rings distinct, tracheids in cross section angular problems with using the name Araucariopitys.However and rounded, resin ducts absent, wood parenchyma fre- as this name is still widely used it was decided that it could quent, commonly with resin content, in some species all be used for this specimen as it was most comparable walls of ray parenchyma smooth and thin, in others to previous descriptions of this morphogenus. Using horizontal walls pitted and tangential walls smooth, but Kraüsel's (1949) scheme for this specimen proved diffi- in most of them the horizontal walls are pitted while cult, depending on which characteristics were followed. the tangential walls are thickened, unevenly punctate LD108 seems to have affinities with Protocupressinoxy- or bead-like, or quite exceptionally dentate, horizontal lon, Protophyllocladoxylon or Araucariopitys. Protocu- walls of longitudinal parenchyma smooth or from slight- pressinoxylon shares characteristics with LD108 such as ly to markedly thickened, nodular, in cross-fields 1–3 the absence of resin canals and the presence of only a few (4–5) cupressoid, taxodioid or podocarpoid pits, spiral cross-field pits. It also contains cupressoid cross-field pits thickenings absent, transverse tracheids in some species as seen in LD108 (Kräusel, 1949). However it also rarely present, rays 1–30 cells high, tangential walls of contains taxodioid and glyptostroboid type cross-field pits ray cells smooth or with scalariform or sieve-like and very dense abietinaceous pitting in the horizontal thickenings. LD132 fits this description particularly as walls, not observed in LD108 (Kräusel, 1949). LD108 in Greguss' (1955) descriptions of Cupressaceae they also differs from Protocupressinoxylon in having some M. Harland et al. / Review of Palaeobotany and Palynology 143 (2007) 167–196 189

Fig. 13. Araucariopitys LD108. a) Transverse section showing straight ring boundaries and copious resin filled cells. b) Radial section showing uni- (black arrow) and biseriate oppositely arranged (white arrow) bordered pits. c) Radial section showing spaced (top arrow) and touching (bottom arrow) bordered pits. d) Tangential section showing a uni- and biseriate ray and spiral thickenings (white arrow). e) Radial section showing a cupressoid cross-field pit. alternately arranged bordered pits and biseriate rays LD108 has closest affinity with Araucariopitys in having (Francis, 1983). Therefore Protocupressinoxylon is dis- mixed type bordered pits (transitional forms) and smooth counted. Protophyllocladoxylon is discounted because, horizontal walls of ray cells. LD108 is very similar to the although LD108 contains the araucarian aspects of this Araucariopitys of late Albian age from Alexander Island, genus, other features do not match. Protophyllocladoxy- Antarctica, described by Falcon-Lang and Cantrill (2000), lon is characterized by the presence of araucarian-type although the tangential rays are shorter and strictly uni- cross-field pitting, the presence of podocarpoid or seriate in their sample, unlike LD108. Previous descrip- dacrydioid type pits and bordered pits rarely touching, tions of Araucariopitys do not mention spiral thickenings which excludes LD108 even though it shares the features (Falcon-Lang and Cantrill, 2000; Poole and Cantrill, of 1–3 oppositely arranged bordered pits and smooth 2001) however they do appear to be present in figures of horizontal walls of ray cells (Kräusel, 1949; Medlyn and Araucariopitys presented in Shilkina's (1967) paper on Tidwell, 1975). According to Kräusel's (1949) scheme, the fossil wood of Franz-Josef Land. Cretaceous age 190 M. Harland et al. / Review of Palaeobotany and Palynology 143 (2007) 167–196

Araucariopitys has mainly been described from Southern biseriate bordered pits with some uniseriate and a few Hemisphere sites such as the late Albian specimen of triseriate forms also being observed (Table 1 and Fig. Falcon-Lang and Cantrill (2000) from Alexander Island, 14b). Where the bordered pits are multiseriate they are Antarctica discussed above. Poole and Cantrill (2001) mainly arranged alternately with a minority being also described Araucariopitys from the Williams Point oppositely arranged (Table 1). The bordered pits are beds of Livingston Island, Antarctica of Cenomanian to always touching (Table 1). Rays seen in tangential section early Campanian age. Northern Hemisphere samples have are uniseriate, 1–41 cells high, mean of 19 cells (Fig. 14c). been reported by Shilkina (1967) from the Early Creta- In radial section cross-fields appear to contain mainly 2 ceous of Franz-Josef Land. pits arranged side-by-side or superimposed although up to Morphogenus: Xenoxylon Gothan (1905) 4 pits may be present in some areas. Cross-field pits Type species: Xenoxylon latiporosum (Cramer in Heer) seem to be mainly taxodioid with some cupressoid or Gothan (1910), basionym Pinites latiporosum Cramer podocarpoid, and occasional slit-like apertures were also in Heer 1868 observed (Fig. 14d). The horizontal walls are thin and Material: Sample LD130 from Spitsbergen was the only unpitted. Ray tracheids, spiral thickenings and horizontal specimen identified as this morphogenus. and vertical resin canals are absent. Xylem parenchyma Description: Sample LD130 is probably derived from a and septa are present (Fig. 14e). mature stem. There are 11 growth rings present in this Identification:ComparisontoKräusel's (1949) scheme sample (Fig. 14a). Growth rings are sometimes indistinct indicates that LD130 has an affinity with Xenoxylon due and the transition from earlywood to latewood is gradual to the presence of the following features: pits on radial (Fig. 14a). The radial walls of tracheids bear mainly walls of tracheids large, arranged alternately and cross-

Fig. 14. Xenoxylon LD130. a) Transverse section showing straight ring boundaries. b) Radial section showing biseriate alternate pitting. c) Tangential section showing high rays. d) Radial section showing cupressoid (black arrow) and taxodioid (white arrow) cross-field pits. e) Radial section showing septa. M. Harland et al. / Review of Palaeobotany and Palynology 143 (2007) 167–196 191 fields with large pits. The presence of some araucarian range of Middle Triassic to Maastrichtian (Medlyn and features combined with modern ones indicates that this is Tidwell, 1975; Spicer and Parrish, 1990), with the vast a transitional conifer form. Medlyn and Tidwell (1975) majority of examples occurring in pre-Cretaceous strata. indicate there are three fossil genera with possible affi- Philippe and Thévenard (1996) indicate, from a review of nities to the Podocarpaceae: Mesembrioxylon, Proto- Xenoxylon occurrences, that during the Cretaceous Xe- phyllocladoxylon and Xenoxylon. Mesembrioxylon is an noxylon was limited to cool, wet areas of the high northern illegitimate, artificial genus established to include extinct latitudes. Early Cretaceous species X. latiporosus and the genera with affinities to Podocarpaceae, replacing Early Cretaceous X. barberi are both reported from Gothan's genera Podocarpoxylon and Phyllocladoxylon Franz-Josef Land by Shilkina (1967).FoliageofXenox- (Medlyn and Tidwell, 1975). Sample LD130 contains ylon is unknown, although it has been suggested that there cupressoid and taxodioid cross-field pits not seen in may be a relationship with Elatides, Podozamites or Phyllocladoxylon or Podocarpoxylon (Kräusel, 1949). Baiera (Medlyn and Tidwell, 1975), none of which have Protophyllocladoxylon is very similar to Xenoxylon with been confirmed. Kräusel (1949) indicating that the only feature separating Family: ?Taxaceae Gray or Cephalotaxaceae Neger them is the nature of the pits in radial walls of tracheids: in Morphogenus: Taxaceoxylon Kräusel et Jain 1964 Protophyllocladoxylon they are typically araucarioid Type species: Taxaceoxylon rajmahalense Kräusel et Jain whilst in Xenoxylon they are much enlarged if in a single 1964 row. Sample LD130 shows similarities to both, having Material: Sample SN25 4 from Spitsbergen was identi- some araucarioid type pitting but also large non- fied as this morphogenus. araucarioid pits. However Medlyn and Tidwell (1975) Description: This sample is probably derived from branch indicate that araucarioid type cross-field pitting occurs in material as there is a large proportion of compression all species of Protophyllocladoxylon which excludes wood present (Fig. 15a). There are 20 distinct growth LD130 from this genus and indicates a closer affinity with rings present and the transition from earlywood to late- Xenoxylon. Xenoxylon spans a rather narrow geological wood is gradual. The radial walls of longitudinal tracheids

Fig. 15. Taxaceoxylon SN25 4. a) Transverse section showing compression wood. b) Radial section showing uni- and biseriate bordered pits. c) Radial section showing biseriate oppositely arranged bordered pits. d) Radial section showing 1, 2, 3 and 4 pits per cross-field mainly cupressoid (cupressoid pit marked by white arrow). e) Radial section showing spiral thickenings. 192 M. Harland et al. / Review of Palaeobotany and Palynology 143 (2007) 167–196 bear mainly uniseriate bordered pits although biseriate perfect fit to the specimen described here Greguss and rare triseriate forms also occur (Table 1 and Fig. 15b). (1955) notes that the description itself is not perfect and Where the bordered pits are multiseriate most are is unclear so it cannot be regarded as final and absolute oppositely arranged with a minor alternate component in all cases. Modern Taxus is largely distributed in the (Table 1 and Fig. 15c). The bordered pits are mainly Northern Hemisphere, particularly Europe, North touching with some being spaced greater than one pit Africa, Asia and North America (Vidakovic, 1991). diameter apart (Table 1). Rays seen in tangential section are uniseriate. Some dark resinous material was seen 5. Discussion within cells and some were also pitted. Rays are 1–11 cells high with mean of 5. The cross-field areas appear to Our analysis reveals that the forests of the Cretaceous contain 1–4 pits of piceoid, cupressoid or possibly (Albian–Aptian) Northern Hemisphere high latitude taxodioid type (Fig. 15d). The pits are either arranged regions had a conifer component of much higher diversity singly, vertically or horizontally side-by-side in 2 rows. than previously recognised (Seward, 1931; Bannan and The horizontal walls of ray cells are unpitted. Ray Fry, 1957; Harland, 1997). In total 12 morphogenera were tracheids are present (Fig. 15e). Abundant spiral thicken- identified: Piceoxylon, Laricioxylon, Protopiceoxylon, ings and some septa are present (Fig. 15f). Taxodioxylon, Pityoxylon, Palaeopiceoxylon, Taxaceox- Identification:UsingKräusel's (1949) scheme the ylon, Juniperoxylon, Protocedroxylon, Araucariopitys, closest match to the specimen described here is Taxox- Xenoxylon and Cupressinoxylon. ylon: pits on radial walls being at least in part circular, if Ten of these morphogenera were present on Spitsber- arranged in several rows mainly separated by bars of gen: Piceoxylon, Laricioxylon, Protopiceoxylon, Taxo- Sanio, pits on radial walls of tracheids never typically dioxylon, Taxaceoxylon, Juniperoxylon, Protocedroxylon, araucarioid, generally or predominantly circular and op- Araucariopitys, Xenoxylon and Cupressinoxylon. Taxo- posite, normal wood without resin canals, spiral thicke- dioxylon was the dominant morphogenus and represented nings present. Fossil woods of Taxaceae have variously 25% of the sample set. Only five morphogenera were been described as Taxoxylon, Taxaceoxylon and Tor- present on Ellesmere and Axel Heiberg islands on the reyoxylon (Roy, 1972). Taxoxylon is used in a more ge- Canadian Arctic Archipelago: Pityoxylon, Piceoxylon, neral sense whilst when the wood shows definite affinity Protopiceoxylon, Palaeopiceoxylon and Cupressinoxy- to Taxaceae Taxaceoxylon is used. The sample here is lon. The dominant morphogenus was Pityoxylon, forming quite similar to Taxaceoxylon mcmurrayensis described 33% of the sample set. by Roy (1972), even though the preservation is quite Very little work has previously been undertaken on the poor, therefore this sample has been included in Taxa- Cretaceous fossil flora from Spitsbergen and has been ceoxylon. Taxaceoxylon wood has been described from entirely based on leaf and shoot impressions (Vakhrameev, the Early Cretaceous sediments of the Lower Athabasca 1991; Harland, 1997) so the current study is the first River region of Alberta Canada. Although it is unclear comprehensive study of Spitsbergen fossil wood. Previous whether shoots and other unspecified remains described studies have shown that the flora of early part of the by Vakhrameev (1991) from Neocomian, Turonian and Cretaceous (Barremian–Aptian) was dominated by Cenomanian sediments in the Privelkhoyanje area of Ginkgo with rare cycadophytes (Harland, 1997). Seward Lena Province and Vilyui Basin area, Siberia, are Tax- (1931) indicated that, amongst the conifers, Elatides, oxylon or Taxaceoxylon it is clear that the Taxaceae were which may have araucarian affinity, was abundant within present in those regions during those time periods. the Early Cretaceous. Within the Early Cretaceous strata, Comparison with extant wood: When compared to Araucarites, Sequoiites, Podozamites and Pinites (Seward, extant wood this sample is most similar to Taxaceae and 1931; Harland, 1997) were also present in Spitsbergen, but Cephalotaxaceae (Taxales) described by Greguss (1955) these samples were collected from a scree slope which may as having the following features: growth rings distinct or have contained both Cretaceous and Tertiary sediments indistinct, tracheids rounded or angular in cross section, (Harland, 1997) so their age is uncertain. resin ducts absent in some genera only, occasionally Although there have been several studies on the Ter- wood parenchyma, in all tracheids relatively delicate tiary flora of the Canadian Arctic Archipelago (Basinger, spiral thickenings, in cross-fields 1–3(4–6) round or 1991; Francis, 1991; Greenwood, 1993 Basinger et al., obliquely positioned elliptical podocarpoid or cupres- 1994; Kumagai et al., 1995; Williams et al., 2003), the sioid pits, tangential walls of ray parenchyma always Cretaceous flora, as with that from Spitsbergen, has smooth, horizontal walls sporadically thickened, rays 1– received little attention (Bannan and Fry, 1957; Falcon- 28 cells high, biseriate in places. Although this is not a Lang et al., 2004). However, Bannan and Fry (1957) M. Harland et al. / Review of Palaeobotany and Palynology 143 (2007) 167–196 193 identified Cedroxylon wood from the Albian sediments of conifers with taxodiaceous affinities were identified the Christopher Formation from Strand Fiord Peninsula from the earlier Aptian to Albian samples used in the on Axel Heiberg Island. They also found Piceoxylon in current study. The presence of Piceoxylon also supports the Christopher Formation on Amund Ringnes Island. its previous identification within the Christopher Forma- Falcon-Lang et al. (2004) indicated that a flora of low to tion (Bannan and Fry, 1957). The conifers of the North medium diversity that included Ginkgo,fungi,cycads, Slope of Alaska and northeastern Russia appear to have angiosperms, ferns, lycopsids and bryophytes grew on been distinct from those found in this study on Spits- Ellesmere Island during the Campanian to Maastrichtian. bergen and Ellesmere and Axel Heiberg islands, where The conifers of Ellesmere Island were dominated by there was a much higher diversity during the Cretaceous Taxodiaceous types with subordinate Pinaceae and paly- (Albian–Aptian). Nonetheless, there were some appar- nomorphs of five taxa represented— Sequoiapollenites, ently common coniferous elements such as Xenoxylon Taxodiaceaepollenites, Podocarpites, Abietineaepolle- found in the Colville region of Alaska and on Spits- nites and Alisporites. bergen (Parrish and Spicer, 1988). Other Cretaceous high latitude Northern Hemisphere Nearest Living Relative analysis (NLR) of the conifer floras have previously been described, including those of flora suggests that Spitsbergen would have had moist, Albian to Cenomanian age from northeastern Russia and cool upland areas and warm temperate lowland envi- theNorthSlopeofAlaska(Spicer and Herman, 2001; ronments with rivers and/or swamps. The dominance Spicer, 2003). The Kukpowruk, Corwin and Killik of Taxodioxylon, which may resemble modern Taxo- Tongue floras of Alaska are similar in composition to dium, suggests that the forests were similar to those each other and to the fossil floras from the Bour–Kemuss, found near the coast of the southeast USA from Grebenka and Tyl' Formations of northeast Russia. Initial Louisiana to North Carolina where Taxodium distichum colonisation by Equisetites was apparently followed by (the Swamp Cypress) dominates warm temperate forests inclusion of Birisia ferns (Spicer and Herman, 2001; in wet, swampy areas (Moore, 1982). The Cretaceous Spicer, 2003). The next successional stage was a flora Taxodium type conifers may also represent a mesother- dominated by Arctopteris with Nilssonia. This gave way mal to megathermal microenvironment with mean to a period of environment-specific vegetation: Ginkgo annual temperatures around 24 °C (Moss et al., 2005). dominated riparian environments and interfluve areas Modern Taxodium has, however, been shown to be were dominated by Podozamites and Pityophyllum.The growing in relictual groups and its true environmental Corwin and Grebenka floras contain almost identical tolerances are not entirely known. The presence of Tax- diverse angiosperms assemblages with some taxa being aceoxylon (similar to modern Taxus) suggests that exclusive to this flora (e.g. Araucarites spp. and Schef- Spitsbergen was cool-temperate with temperatures fleraephyllum), although Podozamites conifers are pres- varying between 1–10 °C, only deviating outside of ent in all locations (Spicer and Herman, 2001). In this range for a few days per year (Vakhrameev, 1991; addition, the Grebenka flora contains the conifers Ce- Vidakovic, 1991). Laricioxylon, similar to modern phalotaxopsis, Araucarites, Pagiophyllum, Pseudolarix Larix, can tolerate very harsh environments, down to (?), Sequoia and Elatocladus (Spicer et al., 2002). Parrish −50 °C, but requires temperatures between −5 and 5 °C and Spicer (1988) also identified the conifer Xenoxylon in for the diplotene stage of reproduction. If temperature the Albian to Cenomanian strata of the Colville River area varies far beyond this range sterility can occur. These (North Slope of Alaska). cooler estimates are closer to those of Philippe and The sample identified and described as Araucariop- Thévenard (1996), who propose that the presence of itys in this study confirms Seward's (1931) finding that Xenoxylon may indicate mean annual temperatures of conifers of araucarian affinity were present in Spitsber- between 5 and 15 °C. gen during the Cretaceous. Previous studies that have The tree types present in the Canadian Arctic are identified Pinites and Sequoiites on Spitsbergen (Sew- similar to those found in Svalbard, but are dominated by ard, 1931; Harland, 1997) are more puzzling, because Pityoxylon with relatively narrow growth rings which the current study found no Pinuxylon or Sequoioxylon suggests that the climate was cool-temperate, between 3 wood at this locality, although representatives of the and 10 °C (mean annual temperature; Whittaker, 1975). Pinaceae (Piceoxylon, Laricioxylon, Pityoxylon) and Temperatures for the Cenomanian of the Colville region Taxodiaceae (Taxodioxylon and Juniperoxylon) families of the North Slope of Alaska have been estimated to were found to be present. Although the taxodiaceous have been 10±3 °C, from leaf-margin analysis (Parrish conifers dominated the Campanian–Maastrichtian for- and Spicer, 1988). This implies that the Aptian–Albian ests on Ellesmere Island (Falcon-Lang et al., 2004), no forests of the Canadian Arctic were growing in a similar 194 M. Harland et al. / Review of Palaeobotany and Palynology 143 (2007) 167–196 or slightly cooler climate than on the North Slope of Bamford, M.K., Philippe, M., 2001. Jurassic–Early Cretaceous Gond- Alaska during the Cenomanian. wanan homoxylous woods: a nomenclature revision of the genera with taxonomic notes. Rev. Palaeobot. Palynol. 113, 287–297. Bamford, M., Zijlstra, G., Philippe, M., 2002. (1529) Proposal to 6. Conclusions conserve the name Cupressinoxylon against Retinodendron (fossil Gymnopermae, Coniferales), with a conserved type. Taxon 51, Fossil wood is abundant in the Cretaceous sediments 205–206. of the Northern Hemisphere polar regions, representing Bannan, M.W., Fry, W.L., 1957. Three Cretaceous woods from the Canadian Arctic. Can. J. Bot. 35, 327–337. the remains of vegetation that was able to grow at high Basinger, 1991. The fossil forests of the Buchanan Lake Formation latitudes when the climate was much warmer than the (Early Tertiary), Axel Heiberg Island, Canadian Arctic Archipel- present. ago: preliminary floristics and palaeoclimate. Geol. Surv. Can. The Northern Hemisphere high latitude conifer Bull. 403, 39–65. forests were much more diverse than previously thought, Basinger, J.F., Greenwood, D.R., Sweda, T., 1994. Early Tertiary vegetation of Arctic Canada and its relevance to paleoclimatic being composed of Piceoxylon, Laricioxylon, Protopi- interpretation. In: Boulter, M.C., Fisher, H.C. (Eds.), Cenozoic ceoxylon, Taxodioxylon, Pityoxylon, Palaeopiceoxylon, Plants and Climates of the Arctic. NATO ASI Series. Springer- Taxaceoxylon, Juniperoxylon, Protocedroxylon, Arau- verlag Berlin, Heidelberg. cariopitys, Xenoxylon and Cupressinoxylon. In Svalbard Barefoot, A.C., Hankins, F.W., 1982. Identification of Modern and the dominant morphogenus was Taxodioxylon and in the Teritary Woods. Oxford University Press, Oxford, UK, p. 189. Beringer, J., Chapin III, F.S, Thompson, C.C., McGuire, A.D., 2005. Canadian Arctic Pityoxylon. Surface energy exchanges along a tundra-forest transition and Comparison of the fossil conifer diversity to modern feedbacks to climate. Agric. For. Meteorol. 131, 143–161. counterparts shows that the forests of Svalbard probably Blokhina, N.I., 1985. Larch wood from Tertiary deposits of the Siziman grew in moist cool upland areas and warm temperate Bight (Khabarovsk Region). Paleontological Journal Translated from: areas, probably with rivers and/or swampy areas present Drevisina listvennitsy iz tretichnykh otlozheniy bukhty Siziman (Khabarovsk kray). Paleont. Zhur. No.3. pp. 105–109, 1985, 3: 88–92. in the lowlands. The forests of the Canadian Arctic were Brentnall, S.J., Beerling, D.J., Osborne, C.P., Harland, M., Francis, J.E., probably growing in similar conditions to those of Valdes, P.J., Wittig, V.E., 2005. Climatic and ecological determinants

Svalbard but the dominance of Pityoxylon indicates that of leaf lifespan in polar forests of the high CO2 Cretaceous it would have been slightly cooler, probably similar to “greenhouse” world. Glob. Chang. Biol. 11, 1–19. northern Canada today. Chaloner, W.G., Creber, G.T., 1990. Do fossil plants give a climatic signal? J. Geol. Soc. (Lond.) 147, 343–350. The widespread occurrence of high latitude forests Creber, G.T., Chaloner, W.G., 1984. Influence of environmental makes them a valuable tool in assessing the climatic factors on the wood structure of living and fossil trees. Bot. Rev. 50 conditions prevailing at the time the trees grew and for the (4), 357–447. validation of the climate simulations from climate models. DeConto, R.M., Brady, E.C., Bergengren, J., Hay, W.W., 2000. Late Cretaceous climate, vegetation, and ocean interactions. In: Huber, B.T., Macleod, K.G., Wing, S.L. (Eds.), Warm Climates in Earth Acknowledgements History. Cambridge University Press, Cambridge, pp. 275–296. Eckhold, W., 1923. Die hoftüpfel bei rezenten und fossilen coniferen. Research funding for this project was provided by Jhrb. Preuss. Geol. Landesanst 42, 472–505. the Natural Environmental Research Council (NERC), Fairon-Demaret, M., Steurbaut, E., Damblon, F., Dupuis, C., Smith, T., UK, grant number NERC/S/J/2002/10896. MH would Gerrienne, P., 2003. The in situ Glyptostroboxylon forest of Hoegaarden (Belgium) at the initial Eocene thermal maximum like to thank Marc Philippe at the Laboratoire de Paléo- (55 Ma). Rev. Palaeobot. Palynol. 126, 103–129. botanique at the Université Lyon, France for many dis- Falcon-Lang, H.J., 2003. Growth interruptions in silicified conifer woods cussions on the draft paper regarding the taxonomy and from the Upper Cretaceous two medicine formation, Montana, USA: classification of fossil wood. JF thanks the Geologic- implications for palaeoclimate and dinosaur palaeoecology. Palaeo- – al Survey of Canada and the Australian Research Coun- geogr. Palaeoclimatol. Palaeoecol. 199, 299 314. Falcon-Lang, H.J., 2005. Intra-tree variability in wood anatomy and its cil for funding to collect the fossil wood, and to L. A. implications for fossil wood systematics and palaeoclimatic Frakes for research support during those projects. David studies. Palaeontology 48 (1), 1–13. J. Cantrill and an anonymous reviewer are thanked for Falcon-Lang, H.J., Cantrill, D.J., 2000. Cretaceous (Late Albian) their reviews. coniferales of Alexander Island, Antarctica. 1: wood taxonomy: a quantitative approach. Rev. Palaeobot. Palynol. 111, 1–17. Falcon-Lang, H.J., MacRae, R.A., Csank, A.Z., 2004. Palaeoecology References of Late Cretaceous polar vegetation preserved in the Hansen Point Volcanics, NW Ellesmere Island, Canada. Palaeogeogr. Palaeocli- Arnold, C.A., 1953. Silicified plant remains from the Mesozoic and matol. Palaeoecol. 212, 45–64. Tertiary of Western North America: II some fossil woods from Falder, A.B., Rothwell, G.W., Mapes, G., Mapes, R.H., Doguzhaeva, Northern Alaska. Mich. Acad. Sci. Pap. 38, 9–20. L.A., 1998. Pityostrobus milleri sp. nov., a pinaceous cone from M. Harland et al. / Review of Palaeobotany and Palynology 143 (2007) 167–196 195

the Lower Cretaceous (Aptian) of southwestern Russia. Rev. Moore, D.M., 1982. Green Planet: The Story of Plant Life on Earth. Palaeobot. Palynol. 103, 253–261. Cambridge University Press, p. 288. Foley, J.A., Kutzbach, J.E., Coe, M.T., Levis, S., 1994. Feedbacks Moss, P.T., Greenwood, D.R., Archibald, S.B., 2005. Regional and between climate and boreal forests during the Holocene epoch. local vegetation community dynamics of the Eocene Okanagan Nature 371, 52–53. Highlands (British Columbia-Washington State) from palynology. Francis, J.E., 1983. The dominant conifer of the Jurassic Purbeck Can. J. Earth Sci. 42, 187–204. Formation, England. Paleontology 26 (Part 2), 277–294 pls. 38–41. Parrish, J.T., Spicer, R.A., 1988. Middle Cretaceous wood from the Francis, J.E., 1991. The dynamics of polar fossil forests: Tertiary fossil Nanushuk Group, Central North Slope, Alaska. Paleontology 31 forests of Axel Heiberg Island, Canadian Arctic Archipelago. (1), 19–34. Geol. Surv. Can. Bull. 403, 29–38. Patchett, P.J., Embry, A.F., Ross, G.M., Beauchamp, B., Harrison, J.C., Gadek, P.A., Alpers, D.L., Heslewood, M.M., Quinn, C.J., 2000. Mayr, U., Isachsen, C.E., Rosenberg, E.J., Spence, G.O., 2004. Relationships within Cupressaceae sensu lato: a combined morpho- Sedimentary cover of the Canadian Shield through Mesozoic time logical and molecular approach. Am. J. Bot. 87 (7), 1044–1057. reflected by Nd isotopic and geochemical results for the Sverdrup Gothan, W., 1905. Zur anatomie lebender und fossiler Gymnosper- Basin, Arctic Canada. J. Geol. 112, 39–57. men-Hölzer. Abhandl. Preuss. Geol. Landasanst. 44, 1–108. Philippe, M., 1993. Nomenclature générique des trachéidoxyyles Gothan, W., 1907. Die fossilen Hölzer von König-Karls Land. Kungl. mésozoïques à champs araucarioïdes. Taxon 42 (1), 74–80. Svensk. Vet.-Akad. Handl. 42, 1–27. Philippe, M., 1994. Radiation précoce des conifers Taxodiaceae et bois Gothan, W., 1910. Die fossilen Holzreste von Spitzbergen. Kungl. affines du Jurassique de France. Lethaia 27, 67–75. Svensk. Vet.-Akad. Handl. 45, 1–56. Philippe, M., Thévenard, F., 1996. Distribution and palaeoecology of Greenwood, D.R., 1993. Stratigraphy and floristics of Eocene swamp the Mesozoic wood genus Xenoxylon: palaeoclimatological im- forests from Axel Heiberg Island, Canadian Arctic Archipelago. plications for the Jurassic of Western Europe. Rev. Palaeobot. Can. J. Earth Sci. 30, 1914–1923. Palynol. 91, 353–370. Greguss, P., 1955. Identification of Living Gymnosperms on the Basis Philippe, M., Zijlstra, G., Barbacka, M., 1999. Greguss's morphogen- of Xylotomy. Akademiai Kiado, Budapest. era of homoxylous fossil woods: a taxonomic and nomenclatural Hall, R.L., Macrae, R.A., Hills, L.V., 2005. Middle Albian (lower review. Taxon 48 (4), 667–676. Cretaceous) gastroplitinid ammonites and dinoflagellates from the Phillips, E.W.J., 1941. The identification of coniferous woods by their Christopher Formation (Dragon Mountain, Axel Heiberg Island, microscopic structure. Bot. J. Linn. Soc. LII, 259–320. Canadian Arctic Islands) and a revision of the genus Pseudogas- Poole, I., Cantrill, D., 2001. Fossil wood from Williams Point Beds, troplites Jeletzky, 1980. J. Paleontol. 79 (2), 219–241. Livingston Island, Antarctica: a late Cretaceous southern high Harland, W.B., 1997. The geology of Svalbard. Geological Society latitude flora. Paleontology 44 (6), 1081–1112. Memoir. J. Geol. Soc. London, vol. 17, p. 521. Poole, I., Hunt, R.J., Cantrill, D.J., 2001. A fossil wood flora from Horrell, M.A., 1991. Phytogeography and paleoclimatic interpretation King George Island: ecological implications for an Antarctic of the Maestrichtian. Palaeogeogr. Palaeoclimatol. Palaeoecol. 86, Eocene vegetation. Ann. Bot. 88, 33–54. 87–138. Poole, I., Mennega, A.M.W., Cantrill, D.J., 2003. Valdivian ecosys- Jefferson, T.H., 1982. Fossil forests from the Lower Cretaceous of tems in the Late Cretaceous and Early Tertiary of Antartcia: further Alexander Island, Antarctica. Paleontology 25 (4), 681–708. evidence from myrtaceous and eucryphiaceous fossil wood. Rev. Jeffrey, E.C., 1907. Araucariopitys, a new genus of araucarians. Bot. Palaeobot. Palynol. 124, 9–27. Gaz. 44, 435–444. Ramanujam, C.G.K., 1971. Fossil coniferous woods from the Oldman Kräusel, R., 1949. Die fossilen Koniferen Hölzer (Unter Auschluss Formation (Upper Cretaceous) of Alberta. Can. J. Bot. 50, 349–352. von Araucarioxylon Kraus). II: Kritische Untersuchungen zur Ramanujam, C.G.K., Stewart, W.N., 1969. Fossil woods of Taxodia- Diagnostik lebender and fossiler Koniferen-Hölzer. Palaeontol. ceae from the Edmonton Formation (Upper Cretaceous) of Alberta. Abh., Abt. B 89, 83–203. Can. J. Bot. 47 (1), 115–124. Krebs, C.J., 1972. Ecology: The Experimental Analysis of Distri- Roy, S.K., 1972. Fossil wood of Taxaceae from the McMurray Formation bution and Abundance. Harper and Row Publishers, New York, (Lower Cretaceous) of Alberta, Canada. Can. J. Bot. 50, 349–352. p. 694. Roy, S.K., Hills, L.V., 1972. Fossil woods from the Beaufort Kumagai, H., Sweda, T., Hayashi, K., Kojima, S., Basinger, J.F., Formation (Tertiary), northwestern Banks Island, Canada. Can. J. Shibuya, M., Fukaoa, Y., 1995. Growth-ring analysis of Early Ter- Bot. 50, 2637–2654. tiary conifer woods from the Canadian High Arctic and its paleo- Schweingruber, F.H., 1990. Microscopic wood anatomy: structural climatic interpretation. Palaeogeogr. Palaeoclimatol. Palaeoecol. variability of stems and twigs in recent and subfossil wood from 116, 247–262. Central Europe. Swiss Federal Institute for Forest, Snow and Kusumi, J., Tsumura, Y., Yoshimaru, H., Tachida, H., 2000. Phy- Landscape Research, Zurich, p. 226. Translated by Karen Baudais- logenetic relationships in Taxodiaceae and Cupressaceae sensu Lundström. stricto based on matK gene, chlL gene, trnL–trnF IGS Region, Seward, A.C., 1919. Fossil Plants: AText-book for Students of Botany and trnL intron sequences. Am. J. Bot. 87 (10), 1480–1488. and Geology. Cambridge University Press, Cambridge. Medlyn, D.A., Tidwell, W.D., 1975. Conifer wood from the Upper Seward, A.C., 1926. The Cretaceous plant-bearing rocks of Western Jurassic of Utah Part I: Xenoxylon morrisonense sp. nov. Am. J. Greenland. Philos. Trans. R. Soc. Lond., B 422, 57–175. Bot. 62 (2), 203–208. Seward, A.C., 1931. Plant Life through the Ages: A Geological Medlyn, D.A., Tidwell, W.D., 1986. New species of Protocedroxylon and Botanical Retrospect. Hafner Publishing Company, London, from the Upper Jurassic of British Columbia, Canada. Great Basin p. 603. Nat. 46 (3), 452–458. Shang, H., Cui, J.-Z., Li, C.-S., 2001. Pityostrobus yixianensis sp. Meijer, J.J.F., 2000. Fossil woods from the Late Cretaceous Aachen nov., a pinaceous cone from the Lower Cretaceous of north–east Formation. Rev. Palaeobot. Palynol. 112, 297–336. China. Bot. J. Linn. Soc. 136, 427–437. 196 M. Harland et al. / Review of Palaeobotany and Palynology 143 (2007) 167–196

Shilkina, A., 1967. The fossil woods of Franz-Josef Land. Translation Stewart, W.N., 1983. Paleobotany and the Evolution of Plants. from Palaeobotanica— Botanical Institute, V. L. Kamarova Acade- Cambridge University Press. my Nauk U.S.S.R 6 Series, vol. 8. Upchurch, G.R.J., Otto-Bliesner, B.L., Scotese, C.R., 1998. Vegetation- Skelton, P.W., 2003. Survey of the Cretaceous World. In: Skelton, P.W. atmosphere interactions and their role in global warming during the (Ed.), The Cretaceous World. Cambridge University Press, Cam- latest Cretaceous. Philos. Trans. R. Soc. Lond. 353, 97–112. bridge, pp. 9–41. Vakhrameev, V.A., 1991. Jurassic and Cretaceous Floras and Climates Smiley, C.J., 1966. Cretaceous floras from Kuk River Area, Alaska: of the Earth. Cambridge University Press, Cambridge, p. 318. stratigraphic and climatic interpretations. Bull. Geol. Soc. Am. 77, Vidakovic, M., 1991. Conifers morphology and variation. Graficki 1–14. zavod Hrvatske, Zagreb, p. 755. Smiley, C.J., 1969. Cretaceous floras of Chandler–Colville Region, Wheeler, E.A., Lehman, T.M., 2005. Upper Cretaceous–Paleocene Alaska: stratigraphy and preliminary floristics. AAPG Bull. 53, conifer woods from Big Bend National Park, Texas. Palaeogeogr. 482–502. Palaeoclimatol. Palaeoecol. 226, 233–258. Spicer, R.A., 2003. Changing climate and biota. In: Skelton, P.W. Whittaker, R.H., 1975. Communities and Ecosystems. Macmillan (Ed.), The Cretaceous World. Cambridge University Press, Cam- Publishing Co., Inc., New York, 385 pp. bridge, pp. 85–162. Williams, C.J., Johnson, A.H., LePage, B.A., Vann, D.R., Sweda, T., Spicer, R.A., Parrish, J.T., 1986. Paleobotanical evidence for cool 2003. Reconstruction of Tertiary Metasequoia forests. II. Struc- north polar climates in middle Cretaceous (Albian–Cenomanian) ture, biomass, and productivity of Eocene floodplain forests in the time. Geology 14, 703–706. Canadian Arctic. Paleobiology 29 (2), 271–292. Spicer, R.A., Parrish, J.T., 1990. Late Cretaceous–early Tertiary palaeo- Wilmking, M., Juday, G.P., 2005. Longitudinal variation of radial climates of northern high latitudes: a quantitative view. J. Geol. Soc. growth at Alaska's northern treeline— recent changes and possible (Lond.) 147, 329–341. scenarios for the 21st century. Glob. Planet. Change 27, 282–300. Spicer, R.A., Herman, A.B., 2001. The Albian–Cenomanian flora of Wilson, K., White, D.J.B., 1986. The Anatomy of Wood: Its Diversity the Kukpowruk River, western North Slope, Alaska: stratigraphy, and Variability. Stobart and Sons Ltd, London, p. 309. palaeofloristics, and plant communities. Cretac. Res. 22, 1–40. Worsley,D.,Aga,O.J.,Dalland,A.,ElverhØi,A.,Thon,A.,1986.Evo- Spicer, R.A., Ahlberg, A., Herman, A.B., Kelley, S.P., Raikevich, M.I., lution of an arctic archipelago— the geological history of Svalbard. Rees, P.M., 2002. Palaeoenvironment and ecology of the middle Den norske stats oljeselskap a.s, Stavanger, Norway, p. 121. Cretaceous Grebenka flora of northeastern Asia. Palaeogeogr. Palaeoclimatol. Palaeoecol. 184, 65–105. Srinivasan, V., 1995. Conifers from the Puddledock locality (Potomac Group, Early Cretaceous) in eastern North America. Rev. Palaeobot. Palynol. 89, 257–286.