The Miocene floras of Iceland and their significance for late Cainozoic North Atlantic biogeography more

boj_441.fm Page 369 Wednesday, November 23, 2005 3:37 PM Blackwell Science, LtdOxford, UKBOJBotanical Journal of the Linnean Society0024-4074The Linnean Society of London, 2005? 2005 149? 369417 Original Article MIOCENE FLORAS FROM ICELAND T. DENK ET AL. Botanical Journal of the Linnean Society, 2005, 149, 369–417. With 214 figures The Miocene floras of Iceland and their significance for late Cainozoic North Atlantic biogeography THOMAS DENK1*, FRIDGEIR GRÍMSSON2 and ZLATKO KVABEK3 1 Department of Palaeobotany, Swedish Museum of Natural History, Box 50007, 104 05 Stockholm, Sweden 2 Earth Science Institute, University of Iceland, Askja, Sturlugata 7, IS 101 Reykjavík, Iceland 3 Charles University, Faculty of Science, Albertov 6, CZ-128 43 Praha 2, Czech Republic Received October 2004; accepted for publication April 2005 A large number of plant macrofossils from several Middle to Upper Miocene localities from Iceland have been studied. The fossil material includes four ferns and fern allies, seven conifers, and about 40 species of flowering plants. Betula islandica and Salix gruberi are described as new species. Coniferous twigs previously ascribed to the genus Sequoia are shown to belong to Cryptomeria based on macro-morphological and epidermal features. Fossil plants from Iceland are compared with coeval fossil taxa from Europe and North America and with living plants. The main finding is that the Miocene flora of Iceland belongs to a widespread Neogene northern hemispheric floral type including plants whose representatives are restricted to East Asia, North America and to western Eurasia at the present time. Previously inferred conspicuous similarities to North American modern equivalents appear to be misleading. The type of vegetation in four plant-bearing sedimentary formations from the late Mid Miocene to Late Miocene, the 12 Ma Brjánslækur-Seljá Formation, the 10 Ma Tröllatunga-Gautshamar Formation, the 9–8 Ma Skar~sströnd-Mókollsdalur Formation, and the 7–6 Ma Hre~avatn-Stafholt Formation, corresponds to a humid temperate broadleaved (deciduous)–coniferous mixed forest dominated by Betulaceae, Fagaceae and Acer. Changes in species composition in the sedimentary formations reflect a shift from warm temperate (Cfa climate) to cool temperate (Cfb climate) conditions from the late Mid Miocene to the latest Miocene. This shift was connected to repeated phases of extinction and colonization. Specifically, one set of thermophilic taxa including Magnolia, Liriodendron, Sassafras and Comptonia went extinct between 12 and 10 Ma, and appears to have been replaced by another set of thermophilic taxa in the 10 Ma formation (Juglandaceae aff. Pterocarya/Cyclocarya, Rhododendron ponticum type). The 9–8 and 7–6 Ma formations are characterized by taxa that migrated to Iceland from Europe, such as Fagus gussonii, Betula cristata and Pterocarya fraxinifolia type. Although there is convincing evidence that plants colonized Iceland both from North America and Europe until 12 Ma, migration in the younger formations (9–8, 7–6 Ma) is suggested to have occurred mainly from Europe. © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417. ADDITIONAL KEYWORDS: climate – plant fossils – plant migration patterns. INTRODUCTION The plateau basalts of Iceland belong to the so-called Brito-Arctic Igneous Province (BIP) that formed in connection with the basaltic break-up of the North Atlantic near the Palaeocene–Eocene transition. The BIP comprises fossils found in intrabasaltic sediments at Svalbard (Spitsbergen), eastern Greenland, *Corresponding author. E-mail: thomas.denk@nrm.se Iceland, the Faeroes, Antrim (Ireland) and the Isle of Mull, and offshore at the Rockall Plateau and the Norwegian Sea (Boulter & Kvadek, 1989). Although most of them are of Palaeogene age, the Icelandic basalts are much younger, the oldest rocks being about 16–15 Ma (Moorbath, Sigurdsson & Goodwin, 1968; McDougall, Kristjansson & Saemundsson, 1984; Hardarson et al., 1997). As Iceland is situated in the North Atlantic approximately half way between North America and Europe it plays a key role as part of a potential land bridge between © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 369 boj_441.fm Page 370 Wednesday, November 23, 2005 3:37 PM 370 T. DENK ET AL. Localities included in the present study are the following (Fig. 1): A. North-western and western Iceland 1. Brjánslækur-Seljá Formation, 12 Ma: Brjánslækur and Seljá in Va~alsdalur localities. 2. Tröllatunga-Gautshamar Formation, 10 Ma: Tröllatunga, Húsavík, Gautshamar, Margrétarfell localities. 3. Skar~sströnd-Mókollsdalur Formation, 9–8 Ma: Mókollsdalur locality. 4. Hre~avatn-Stafholt Formation, 7–6 Ma: Fífudalur, Hestabrekka, Primilsdalur/Porvaldsdalur. B. North-eastern Iceland 5. Vindfell (‘Hre~avatn horizon’ according to Akhmetiev et al., 1978; 7–6 Ma). The ages of the localities are based on K–Ar, Ar–Ar dates and palaeomagnetic correlation (Moorbath et al., 1968; McDougall et al., 1984; Hardarson et al., 1997). Specimens from another locality, Selárdalur (15 Ma; Hardarson et al., 1997; Kristjansson, Hardarson & Audunsson, 2003), collected during a field trip in summer 2003 have not been included in the present study. They are mainly leaf impressions of Fagus sp., Tilia sp. and Ulmus sp. (cf. Grímsson & Denk, 2005; F. Grímsson, T. Denk & L.A. Símonarson, unpubl. data). Most of the fossils studied are cleavage impressions–compressions, and cleaved compressions with a very thin organic layer, and some are preserved as true impressions (cf. Chaloner, 1999). Maceration of cuticles was tried in all cases when leaf fossils were preserved as carbonized compressions. Most of the compressions were cracked into tiny pieces with an adhering layer of (diatomite) matrix. Therefore, fragments of lamina first were cleaned in HF (in plastic microcells), and then transferred into a drop of distilled water on a microslide. A routine maceration with Schulze solution and 5% KOH was employed (for more details see Kvadek, 2004), and the cuticles obtained were then embedded in a drop of glycerine on the same microslide. The time of treatment with Schulze solution varied from several seconds to 10 min. In general, the epidermal structure in the angiosperm samples examined is rarely preserved on the cuticle membrane, while conifer needles sometimes had comparatively thick cuticles. For the preparation of in situ pollen the same procedure as for cuticles was applied. The terminology used for leaf morphology follows Dilcher (1974), Spicer (1986) and Liu (1996). Europe and North America during the Late Cainozoic (Akhmetiev et al., 1978; Eldholm, Myhre & Thiede, 1994; Tiffney & Manchester, 2001). Today, welldeveloped stratigraphic frameworks are available for the North Atlantic area (for Iceland see McDougall et al., 1984) but almost no terrestrial vertebrate fossils are known from Iceland (Símonarson, 1990), and our understanding on how Mid to Late Miocene biota of Iceland were related to coeval biota in Europe and North America and to modern biota exclusively relies on plant fossils. More than a hundred years ago Oswald Heer (1868) studied plant macrofossils from the Miocene of Iceland as part of his monumental work on Cainozoic floras from high latitudes of the northern hemisphere (Heer, 1868–1883). Windisch (1886) gave a summary of works dealing with fossils from Iceland prior to Heer and from this time on Cainozoic plants from Iceland continuously have attracted the interest of palaeobotanists and botanists (e.g. Áskelsson, 1946, 1957; Lindquist, 1947; Manum, 1962; Friedrich, 1966; Akhmetiev et al., 1978; Friedrich & Símonarson, 1982, 2002). Previous workers agree on the close affinity of the Miocene flora of Iceland to modern (eastern) North American taxa (Friedrich, 1966; Akhmetiev et al., 1978; Friedrich & Símonarson, 1982; Mai, 1995). Comprehensive studies on the Miocene macroflora from Iceland, however, are missing. The Natural History Museum in Stockholm houses one of the largest collections of Miocene plant fossils from Iceland. Although A. G. Nathorst had a plan to work on this collection at the end of the 19th century he never managed to do so, and so the material has never been studied as a whole and was not included in the major review by Akhmetiev et al. (1978). The latter study was mainly concerned with stratigraphy and did not provide taxonomic revisions of plant taxa. To fill this gap, and to provide data for a better understanding of the biogeographical position of Iceland during the Late Cainozoic, we studied more than 3500 plant fossils from several Mid to Late Miocene localities of Iceland and revised previous taxonomic treatments. MATERIAL AND METHODS The plant material examined is housed at the Swedish Museum of Natural History (c. 3500 specimens). Additional specimens are stored at the Department of Earth Sciences, University of Aarhus, at the Geological Museum Copenhagen, and in Reykjavík (Icelandic Museum of Natural History and F. Grímsson). The fossils were collected by Th. Thoroddsen (1886, 1888), Dr Winkler München (1883), and G. G. Bár~arson (1910–1924), among others. Most of the localities were revisited and additional specimens collected in summer 2003 (T. Denk, F. Grímsson). SYSTEMATIC PALAEOBOTANY Information on localities in the synonymy lists includes the name used by the respective author(s) followed by the currently used locality names in square © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 boj_441.fm Page 371 Wednesday, November 23, 2005 3:37 PM MIOCENE FLORAS FROM ICELAND 371 Figure 1. Age and location of plant-bearing sediments on Iceland. Only those localities mentioned in the text are shown. (1) Selárdalur, latest Early Miocene, 15 Ma; (2) Seljá/Va~alsdalur, late Mid Miocene, 12 Ma; (3) Brjánslækur, late Mid Miocene, 12 Ma; (4) Tröllatunga, early Late Miocene, 10 Ma; (5) Ví~idalsá, early Late Miocene, 10 Ma; (6) Húsavík, early Late Miocene, 10 Ma; (7) Gautshamar, early Late Miocene, 10 Ma; (8) Margrétarfell, early Late Miocene, 10 Ma; (9) Tindafjall, Late Miocene, 9–8 Ma; (10) Mókollsdalur, Late Miocene, 9–8 Ma; (11) Langavatnsdalur, Late Miocene, 7–6 Ma; (12) Primilsdalur/Porvaldsdalur, Late Miocene, 7–6 Ma; (13) Fífudalur, Late Miocene, 7–6 Ma; (14) Hestabrekka, Late Miocene, 7–6 Ma; (15) Brekkuá, Late Miocene, 7–6 Ma; (16) Hre~avatn, Late Miocene, 7–6 Ma; (17) Tjörnes, Late Pliocene, c. 3 Ma; (18) Sandfell, Late Miocene, 7–6 Ma; (19) Vindfell, Late Miocene, 7–6 Ma; (20) Hólmatindur, early Late Miocene, 10 Ma. (modified after Sæmundsson, 1974, 1979; Aronson & Sæmundsson, 1975; Jóhannesson, 1980; Jóhannesson & Sæmundsson, 1989; Steinthórsson & Thorarinsson, 1997; Hardarson et al., 1997; Símonarson et al., 2000, 2002). brackets. Numbers of the specimens studied that are stored at the Swedish Museum of Natural History are available from http://www.nrm.se/pb/data/ welcomeeng.html and therefore are not listed in the systematic part. EQUISETACEAE Equisetum L. Equisetum sp. (Figs 2–4) ?Equisetum winkleri Heer, 1859: 317 (Gaulthame [Gautshamar], Sandafell [Sandfell]). Equisetum winkleri Heer; Heer, 1868: 140, pl. 24, figs 2–6 (Gaulthvamr [Gautshamar], Sandafell [Sandfell]). Equisetum sp. (Equisetum parlatorii Schimper?); Windisch, 1886, p. 26 (Tröllatunga). Equisetum sp. (cf. Equisetum parlatorii Heer; Schimper); Friedrich, 1966: 57, pl. 1, fig. 8 (Seljá in Va~alsdalur). Equisetum sp., Akhmetiev et al., 1978: 178, 181, pl. 7, figs 4, 7 (Vindfell), pl. 12, figs 10, 18 (Tjörnes), pl. 15, fig. 23 (Vididaljsau [Ví~idalsá]). Description: A few fragments of aerial stems with nodes and leaves in whorls, fused into a sheath, leaves around 5 mm long, 10–11 leaves per axis width; one underground rhizome, nodules without specific diagnostic characters. © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 boj_441.fm Page 372 Wednesday, November 23, 2005 3:37 PM 372 T. DENK ET AL. 2 3 4 5 6 7 8 9 10 11 12 13 © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 boj_441.fm Page 373 Wednesday, November 23, 2005 3:37 PM MIOCENE FLORAS FROM ICELAND 373 Figures 2–13. Figs 2–4. Equisetum sp. 2, 3, Gautshamar, S094412-02, S094413, fragments of aerial stems. 4, Húsavík, S094588, rhizome. Figs 5–9. Osmunda parschlugiana 5–8, Tröllatunga. 5, 6, S106766, apical part of frond, close-up of pinna 7. 8, S106764, isolated pinna, close-up of pinna showing venation pattern. 9, Húsavík, S106895, large pinna. Figs 10, 11. Pteridophyta indet. 1, Húsavík, S094732, fragment of pinna, close-up of two segments. Figs 12, 13. ‘Dryopteris’ sp., Brjánslækur, S134428, pinna, close-up showing venation. Scale bars = 5 cm in Fig. 9, 1 cm in Figs 4, 5, 10, 12, 13, and 0.5 cm in Figs 2, 3, 6–8, 11. Discussion: The material is very fragmentary. The number of leaves forming the sheath corresponds to that reported by Heer (1868) for E. winkleri. Material: Seljá in Va~alsdalur (Friedrich, 1966); Gautshamar, Húsavík; Ví~idalsá, Vindfell, Tjörnes (Akhmetiev et al., 1978). OSMUNDACEAE Osmunda L. Osmunda parschlugiana (Unger) Andreánszky (Figs 5–9) Osmunda heeri Gaud.; Akhmetiev et al., 1978: 178, 180, pl. 6, figs 1, 2, 8, 9 (Husavik [Húsavík]), pl. 11, fig. 5 (Holmatindur [Hólmatindur]). Description: Several fragmentary leaf apices and isolated pinnae; pinnae alternately arranged, up to 70 mm long and 20 mm wide, base asymmetric, slightly cordate, apex blunt, margin finely crenulate, lateral veins rather dense, usually branching twice; all veinlets ending in sinuses. Discussion: Previous records of Osmunda on Iceland were assigned to O. heeri (Akhmetiev et al., 1978). However, in Europe foliage of the same morphology has usually been assigned to O. parschlugiana (KovarEder, Kvadek & Ströbitzer-Hermann, 2004), which is typified by a pinna from Parschlug (Early–Middle Miocene of Styria). This genus was common in the Arctic Cainozoic and its records were published under different species names (see Boulter & Kvadek, 1989). All belong to the Osmunda regalis L.-type. Material: Húsavík, Tröllatunga, Margrétarfell; Hólmatindur (Akhmetiev et al., 1978). Pteridophyta gen. et spec. indet. 1 (Figs 10, 11) Description: Leaf fragment; 14 mm long, probably representing a small medial part of a pinna, only two pairs of alternate segments preserved, segments about 7 mm long and c. 3 mm wide, narrow oblong, margin entire, bluntly rounded at the apex, fused for 2 mm from the rachis, sinus narrow and sharp, primary vein almost perpendicular to the rhachis, secondaries very thin, hardly visible. Discussion: Akhmetiev et al. (1978: pl. 6, fig. 2) illustrated a fern fragment from Brjánslækur as Osmunda heeri resembling the present fossil, but differing in wider and rounded sinuses. This type of fern foliage is not rare among leptosporangiate ferns and cannot be attributed to a particular genus in sterile condition. Akhmetiev et al. (1978: 20) mentioned Woodwardia from the same locality (coll. Áskelsson), which may develop a similar venation pattern without areoles in very young leaves, and can be mistaken for other genera (Kvadek & Hurník, 2000). Even such anomalous Woodwardia fossils, however, differ from the described fragment by rounded sinuses. Material: Húsavík. Pteridophyta gen. et spec. indet. 2 (‘Dryopteris’ sp.) (Figs 12, 13) Description: Pinna of a frond; 37 mm long and 15 mm wide, becoming narrower towards the apex, segments of pinna oblong, maximal 10 mm long and 4 mm wide, forming an angle of about 30–40∞ with rachis, apex asymmetrical, bluntly acute, segments not fused at their base, sinuses sharp, margin entire, primary vein dividing segments into two unequal parts, secondary veins thin and hardly visible. Discussion: Fern fronds similar to the present one were figured in Akhmetiev et al. (1978: pl. 11, fig. 11) from Hólmatindur and referred to as Dryopteris linneaneiformis Iljinskaya. The nature of the asymmetrical apex of the segments is typical of some modern species of Dryopteris. However, without fertile structures it is difficult to assign this type of pinnae to particular modern fern genera. The primary vein dividing the segments in two unequal parts is not typical of Dryopteris. Material: Brjánslækur. PINACEAE Abies Mill. Abies steenstrupiana (Heer) Friedrich (Figs 14–18) Pinus steenstrupiana Heer, 1859: 318, partim (basionym). © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 boj_441.fm Page 374 Wednesday, November 23, 2005 3:37 PM 374 T. DENK ET AL. 17 14 15 16 18 21 19 20 22 25 27 23 24 26 28 29 © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 boj_441.fm Page 375 Wednesday, November 23, 2005 3:37 PM MIOCENE FLORAS FROM ICELAND 375 Figures 14–29. Figs 14–18. Abies steenstrupiana. 14, 15, Primilsdalur, S094922, leaf, S106897, scale. 16-18, Brjánslækur 16 S094057, scale. 17, S094032-01, seed. 18, S094013, seed. Figs 19–22. Larix sp. Tröllatunga, twig fragments with spur shoots. 19, S106536. 20-22, S106776-01. Figs 23–29. Picea sect. Picea, Brjánslækur. 23, S094008A, seed cone. 24, 25, S094051, seed cone, close-up of scales with entire margin. 26, S094002, seed. 27, S094063-03. 28, S093949. 29, S094059, seed cone. Scale bars = 5 cm in Figs 23, 24, 29, 1 cm in Figs 14–21, 25–28, and 0.5 cm in Fig. 22. Pinus steenstrupiana Heer; Heer, 1868: 144, pl. 24, figs 23–25, (?) 26 (Brjamsloek [Brjánslækur], Langavasdalr [Langavatnsdalur], Hredavatn [Hre~avatn]). Pinus steenstrupiana Heer; Windisch, 1886: 29 (Brianslaekr [Brjánslækur]). Abies steenstrupiana (Heer) Friedrich, 1966, comb. inval., p. 58, pl. 1, figs 6, 9, 10, text-fig. 13 (Brjánslaekur [Brjánslækur]). Description: Cone scales, winged seeds and leaves; scales broad flabelliform, proximal margin decurrent, 18 mm long and 25 mm wide, no bracts visible, with a distinct radial striation; seeds broadly winged, wing attached to almost half of the circumference of the seed, 10 mm wide at its apical margin, 5 mm at its proximal part, seed roundish to elliptical, c. 5 mm in diameter; leaves flat needles, linear, c. 17 mm long, oblong, apex rounded, becoming narrower towards base. Discussion: Friedrich (1966: 58) based his new combination on a publication by Heer (1868) although Pinus steenstrupiana was already validly published in Heer (1859: 318) based on isolated leaves and cone scales from Brjánslækur, Langavatnsdalur and Hre~avatn. Although Heer (1859) included Abies and Picea within Pinus, which was usage at that time, he made a clear distinction between firs (‘Weisstannen’), spruces (‘Rothtannen’) and true pines (‘Föhren’). We agree with the transfer suggested by Friedrich and validate here his combination. Material: Brjánslækur, Primilsdalur. Larix Mill. Larix sp. (Figs 19–22) Description: Long shoots with short, lateral spur shoots; small protrusions on long shoots indicating the position of leaves on long shoots. A narrow needle. Discussion: The record in the Miocene of Iceland may be connected either with many others in the Canadian Arctic Miocene (LePage & Basinger, 1991) or with those in Siberia because Larix did not reach Central Europe prior to the Pliocene (Mai, 1995). Material: Tröllatunga. Picea A. Dietr. Picea sect. Picea sp. (Figs 23–33) Pinus microsperma auct. (? non Heer, 1859); Heer, 1868: 142, pl. 24, figs 11–17 (Brjamslaek [Brjánslækur], Hredavatn [Hre~avatn]). Pinus aemula Heer, 1859: 318 (Brjamsloek [Brjánslækur]). Pinus aemula Heer; Heer, 1868: 143, pl. 24, fig. 20 (Brjamslaek, [Brjánslækur]). Pinus brachyptera Heer, 1859: 318 (Brjamsloek [Brjánslækur]). Pinus brachyptera Heer; Heer, 1868: 143, pl. 24, fig. 18 (Brjamslaek, [Brjánslækur]); Windisch 1886, p. 30 (Brianslaekr [Brjánslækur]). Picea microsperma (Heer) Friedrich, 1966, comb. inval., p. 60, pl. 1, fig. 11 (Brjánslaekur [Brjánslækur]). Picea sp. 1; Akhmetiev et al., 1978: 177, pl. 2, fig. 5 (Brjanslaekur [Brjánslækur]). Picea sp. 2; Akhmetiev et al., 1978: 177, pl. 2, figs 9, 11 (Brjanslaekur [Brjánslækur]). Picea breweriana Wats. fossilis Akhmetiev et al., 1978: 177, pl. 2, figs 6, 7 (Brjanslaekur [Brjánslækur]). Picea sp., Akhmetiev et al., 1978: 178–180, pl. 7, fig. 6 (Vindfell), pl. 8, fig. 4 (Tindar [Tindafjall]), pl. 9, fig. 1 (Mokollsdalur [Mókollsdalur]), pl. 12, figs 1–6, 9, 11, 15, 18, pl. 13, fig. 12 (Tjornes [Tjörnes]). Description: Pollen cone with Picea type pollen in situ, seed cones, winged seeds and uncertain leaf remains; pollen cone about 25 mm long, in late stage of maturity, with microsporophylls widely spaced, bisaccate pollen 61 ¥ 40 mm; seed cones up to 90 mm long and 25 mm wide, seed scales closely imbricate, distal margin entire, c. 11 mm wide, winged seeds 16–22 mm long, 4–7.6 mm wide, wing attached to upper part of seed. Discussion: Seed cones, pollen cones and winged seeds are referred to this entity, which may represent more than one single natural species. It has previously been assigned to Picea microsperma and compared to the modern P. glauca Voss. by Friedrich (1966), who overlooked the fact that the basionym, Pinus microsperma Heer, was published in 1859 (not only in 1868 as Friedrich quoted) based on a seed from the Swiss locality Le Locle (Heer, 1859: 161), and therefore not appropriate for the Icelandic record. Later its seed © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 boj_441.fm Page 376 Wednesday, November 23, 2005 3:37 PM 376 T. DENK ET AL. 32 33 30 31 34 35 36 37 38 41 42 43 39 40 44 © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 boj_441.fm Page 377 Wednesday, November 23, 2005 3:37 PM MIOCENE FLORAS FROM ICELAND 377 Figures 30–44. Figs 30–33. Picea sp., Tröllatunga. 30, 31, S106517A, male inflorescence. 32, 33, Bi-saccate pollen from the inflorescence shown in 30. Fig. 34. Pinus sp., Primilsdalur, GMK6731, seed. Figs 35, 36. Pseudotsuga sp., Primilsdalur, HPK 32a, 32b, seed cone showing conspicuous bracts with 3-lobed apex. Figs 37–44. Tsuga sp. 37–40, Brjánslækur. 37– 38, S093406-02, S094004, cuticle preparations from the leaves shown in Figs 39, 40. 39, 40, leaf fragments. 41, 42, Primilsdalur, GMK 6740, 6738, stalked leaves. 43, 44, Fífurdalur, S094080-02, leaf, close-up showing leaf petiole. Scale bars = 1 cm in Figs 30, 31, 34, 39–43, 2 cm in Figs 35, 36, 0.5 cm in Fig. 44, 100 mm in Figs 32, 33, and 50 mm in Figs 37, 38. cones were named P. breweriana S. Wats. fossilis by Akhmetiev et al. (1978). Examination of living species of Picea shows that the fossils are very similar to a number of North American species having cone scales with entire margined rounded apices, but also to northern populations of the European P. abies (L.) Karsten (P. abies ssp. obovata Ledeb.), and to at least ten East Asian species belonging to the section Picea. The circumscription of fossil spruce species requires much more complete material and certainly cannot be based on isolated seeds (e.g. LePage, 2001). Therefore, we leave the specific designation for the records from Iceland open. Material: Brjánslækur, Tröllatunga, Fífudalur, Vindfell, Mókollsdalur, Tjörnes, Tindafjall. Pinus L. cf. Pinus sp. (Fig. 34) Pinus thulensis Steenstrup, Heer, 1868: 141, pl. 24, fig. 21 (Hredavatn [Hre~avatn]). Description: One winged seed; 22 mm long, the elliptic seed 5.2 mm long and 3.4 mm wide, attached to the wing along half of its circumference, wing asymmetric. Remarks: Pollen data indicate the presence of Pinus in all Miocene sedimentary formations of Iceland (Akhmetiev et al., 1978). Seeds and leaves, however, are rare. The reason for this is probably that Pinus grew at some distance from the sedimentary basins. The seed from Primilsdalur resembles Pinus by its shape and the way that the wing is attached to the seed. Nevertheless, based on the present material possible affinities to Pseudolarix and Keteleeria cannot be ruled out. Material: Primilsdalur, GMK6731. Material stored at Geological Museum Copenhagen. Pseudotsuga Carrière Pseudotsuga sp. (Figs 35, 36) Description: Seed cone; 56 mm long and 37 mm wide when measured from the distal ends of the bracts, 24 mm wide when measured from the distal ends of the scales; peduncle 8.2 mm long and 3.5 mm wide; cone widest in its lower to upper–middle part and gradually tapering thereafter towards the distal region, shape oblong cylindrical; composed of numerous wide to narrow obovate cone scales, scales relatively long with smoothly rounded distal margin and wedge shaped below, scales much shorter than their bracts, outer surface of scales marked by fine cellular rows radiating from the proximal part towards the distal margin; three-pronged bracts projecting like tongues and pointing straight towards tip of the scales, bracts extending beyond the distal scale margin, bracts long acute-tipped with lateral wing-like extensions in their lower parts, the wings displaying extremely fine and dense venation. Discussion: The form of the bracts and their distinctive appearance is characteristic for the genus Pseudotsuga. The cone is the first representative of this genus from the Hre~avatn-Stafholt Formation and from Iceland as a whole. The fossil record of Pseudotsuga in Europe is restricted to a few uncertain leaf remains and no records of cones belonging to Pseudotsuga are known from the Cainozoic of Europe (Mai, 1995). The oldest macrofossil record in North America is from the Early Oligocene of Oregon (Schorn & Erwin, 2000). By the Miocene the genus was also present in Japan. Pseudotsuga consists of eight (to nine) extant species displaying an East Asian–western North American disjunct distribution. Material: Primilsdalur. Specimen nos HPK-32a, 32b, stored in Reykjavík (coll. F. Grímsson). Tsuga Carrière Tsuga sp. (Figs 37–44) Description: Single flat needle leaves; shortly petiolate or incomplete at base, apex acute or slightly notched. Hypostomatic, cells longitudinally orientated, narrow elongate, straight-walled, stomata in narrow bands, longitudinally arranged, 48–52 mm long, incompletely amphicyclic, with two (short) lateral subsidiary cells and two polar, slightly elongate cells. Discussion: These leaf remains may belong to different species. Whereas for the leaves from Brjánslækur epidermal features support their generic identity, the leaves from Fífudalur may belong to various taxa of Pinaceae, such as Tsuga, Pseudotsuga or Abies. © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 boj_441.fm Page 378 Wednesday, November 23, 2005 3:37 PM 378 T. DENK ET AL. Boulter (1970 – Neogene of Derbyshire) is based on anatomically studied foliage shoots. Sterile remains from both localities match well those from Iceland in morphology and cuticle structure. We follow the interpretation of Kilpper (1968) who did not directly assign the sterile foliage shoots to C. rhenana. Similar shoots of Doliostrobus (D. taxiformis var. sternbergii, cf. Kvadek, 2002) differ in completely amphicyclic stomata with narrow subsidiary cells and distinct crystal cavities in the cuticle. The occurrence of Cryptomeria in the Miocene of Iceland is the westernmost record of this genus, indicating the palaeofloristic connection of Iceland with the rest of Europe during the Neogene. At present a single species, Cryptomeria japonica (L. f.) D. Don., is native to Japan. Material: Brjánslækur. Material: Brjánslækur, Fífudalur; Primilsdalur (specimens GMK 6740, 6738 stored at the Geological Museum Copenhagen). CUPRESSACEAE (INCL. TAXODIACEAE) Cryptomeria D. Don Cryptomeria anglica Boulter (Figs 45–49) Araucarites sternbergii auct. (non Goepp.); Heer, 1859: 316, 317 (Brjamsloek [Brjánslækur]). Sequoia sternbergii auct. (non (Göpp.) Heer); Heer, 1868: 140, pl. 24, figs 7–10 (Brjamslaek [Brjánslækur]); Windisch, 1886: 28 (Brianslaekr [Brjánslækur]); Friedrich, 1966: 63, pl. 1, figs 5, 7 (non text-fig. 14 = Alnus) (Brjanslaekur [Brjánslækur]). Akhmetiev et al., 1978: 177, pl. 1, figs 1, 4, 8, 14 (Brjanslaekur [Brjánslækur]). Brjanslaekuria kryshtofovichii Sveshnikova, 1984: 264–266, pl. 1, figs 1–4, pl. 2, figs 1, 2 (Brjanslaekur [Brjánslækur]). Description: Sterile foliage shoots with helically disposed falcate needle leaves, usually patent, rarely appressed, with blunt apex and long decurrent base, around 10 mm long, quadrangular in cross-section, with lateral margin indicated as a line running parallel with and close to the adaxial edge of needles. Cuticle medium thick, showing quadrangular more or less elongate cells and two stomatal bands, stomata incompletely amphicyclic, densely arranged and irregularly orientated, roundish, 50–58 mm long, with a simple circle of subsidiary cells bordered by a faintly thicker peripheral line, guard cells forming distinct cuticular thickenings at their polar ends (= T-pieces). Discussion: This conifer was attributed to Sequoia sternbergii (= Doliostrobus sternbergii) based on associated, but detached cone-like structures by Heer (1865: text-fig. 161) and Friedrich (1966: text-fig. 14). However, we interpret these remains as the strobilelike infructescences of Alnus. The stomatal topography and structure of the needles studied provides strong support for the reassignment to Cryptomeria. Stomata in Sequoia and Sequoiadendron are fully amphicyclic with an inner circle of thicker subsidiary cells. The European records of Cryptomeria have been assigned to two more or less morphological species. Cryptomeria rhenana Kilpper (1968 – Miocene of Rhineland) is based on a seed cone, while C. anglica ANGIOSPERMS (Families are listed in alphabetical order; aquatic, herbaceous and graminoid plants appear at the end.) BETULACEAE Alnus Mill. Alnus sp. (Figs 50, 51) Description: Leaves petiolate, petiole 8+ mm long, lamina ovate to narrow ovate, sometimes slightly triangularly lobed, 5.5–11 cm long and 3.7–8.6 cm wide, apex acute or obtuse, base round, margin double dentate/serrate with small obtuse to acute teeth, close to the base margin entire, teeth of two sizes, i.e. primary and secondary teeth present, secondary and abmedial veins running into primary teeth, secondary teeth served by veinlets branching off from tertiary veins, up to 7 secondary teeth along the margin between two adjacent secondary veins, secondary venation craspedodromous, secondary veins up to 10 pairs, lowest ones (pectinal veins) often gently curved and subparallel to basal margin, sending off several abmedial branches, higher up secondary veins relatively straight, tertiary veins oblique to perpendicular to secondary veins, simple or forked, about 4–11 tertiary veins per 1 cm secondary vein, less distinct in larger specimens, course of the quaternary veins orthogonal, areoles imperfect, veinlets branched twice or three times. Figures 45–51. Figs 45–51. Cryptomeria anglica, Brjánslækur. 45–48, S093406, S093948A vegetative branches, closeups showing arrangement of leaves. 49, Cuticle preparation from the branch shown in Figs 47, 48. Figs 50, 51. Alnus sp., Hestabrekkusund, HBK 63. Scale bars = 5 cm in Figs 45, 47, 51, 3 cm in Fig. 50, 1 cm in Figs 46, 48, and 50 mm in Fig. 49. © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 boj_441.fm Page 379 Wednesday, November 23, 2005 3:37 PM MIOCENE FLORAS FROM ICELAND 379 49 45 46 47 48 50 51 © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 boj_441.fm Page 380 Wednesday, November 23, 2005 3:37 PM 380 T. DENK ET AL. (?) Alnus sp.; Friedrich, Símonarson & Heie, 1972: 8, pl. 1, fig. 4 (Mókollsdalur). Description: Leaves petiolate, petiole 7–16 mm long, lamina broad ovate to elliptic, 5–13 (to 15) cm long and 3.5–8 (to 9) cm wide, base obtuse to slightly cordate, apex acute to acuminate, opadial vein present, basalmost secondary veins almost perpendicular to primary vein, following secondaries forming angles between 30∞ and 45∞ to primary vein, 8–12 pairs, tertiary veins rather densely spaced, abmedial veins inserting teeth or forming loops from which short veinlets insert teeth, teeth of more or less equal size, small, with basal side as long as or slightly longer than apical side, apex acute. Discussion: Such broad leaves have often been reported from Sarmatian and Pannonian/Pontian deposits of Poland, Hungary, Austria, Moravia and Greece as A. cecropiaefolia (Ettingsh.) Berger. Knobloch (1969) mentioned the modern Mexican species A. pringlei Fernald as comparable with A. cecropiifolia, while later authors suggested Eurasian alders, e.g. A. glutinosa ssp. barbata (C. A. Meyer) Yaltırık (Kvadek, Velitzelos & Velitzelos, 2002). Also the North American A. rhombifolia Nutt. sometimes has leaves that resemble the fossil species in shape and show a dentate margin at the leaf base similar to the fossil. Material: Brjánslækur, Tröllatunga (?) Mókollsdalur. Alnus cf. kefersteinii (Goepp.) Goepp. (Figs 59–62) Alnus kefersteinii (Goepp.) Goepp.; Heer, 1868: 146, pl. 25, figs 4–9 (Hredavatn [Hre~avatn], Husawik [Húsavík]); Windisch, 1886: 35, partim (infructescences) (Brianslaekr [Brjánslækur]). Sequoia sternbergii auct. (non (Goepp.) Heer); Heer, 1865, partim, text-fig. 161 (Bramslaek [Bránslækur]); Friedrich, 1966: 63, partim, textfig. 14 (Brjánslaekur [Brjánslækur]). Description: Strobile-like infructescences, 1.5–2 cm long, medial and lateral bracteoles fused. Discussion: Infructescences of Alnus occur in all four horizons but cannot be linked to particular species because they were never found attached to twigs bearing leaves. In our material specimens from the younger Hre~avatn-Stafholt Formation are larger than specimens from the older formations. These infructescences lend credibility to the identification of leaves to this genus, because Alnus and Betula overlap in leaf architecture. Brjánslækur, Material: Hestabrekka. Húsavík, Hre~avatn, Discussion: Such lobed leaves are typical of a number of modern high latitude Alnus species in Europe and North America [e.g. Alnus incana (L.) Moench., A. incana ssp. tenuifolia (Nutt.) Breitung, A. rubra Bong.]. The fossil leaves co-occur with infructescences referred to as A. cf. kefersteinii (Fig. 62) that are larger than those from the older Brjánslækur-Seljá and Tröllatunga-Gautshamar formations. Material: Hestabrekkusund, Brekkuá (HBK-9, 16a, 16b, 34, 37, 41, 59, 61, 62a, 63, HHK-1, 2, 17a, 17b, 30, 44a, 44b, 49, HHSl-22a, 22b, 36a, 36b, 78a, and 78b). Material stored at Reykjavík (coll. F. Grímsson). Alnus aff. gaudinii (Heer) Knobloch & Z. Kvadek (Figs 52–55) Juglans Friedrich & Símonarson, 1983: fig. 10 (Brjánslækur). Description: Leaves petiolate, petiole rarely preserved, 4+ mm long, lamina narrow ovate, 5–11 cm long, 3–4.5 cm wide, serrate, base cordate or rounded, apex acute, secondaries pinnate, more densely spaced in the lower part, steeper and less dense towards the apex, in 8–11 pairs, curved towards the margin, semicraspedodromous, abmedial veins forming loops from which small veinlets insert teeth, teeth with long basal and short apical side. Carbonized tissue resistant to maceration so that only very thin adaxial cuticle fragments were obtained; cells quadrangular, straight-walled and faintly granular on outer periclinal walls. Discussion: Similar leaves occur in a number of Miocene localities of Europe, but the secondaries are generally denser and the base more acute. Knobloch & Kvadek (1976) first recognized their true generic affinity based on epidermal features. Alnus nitida (Spach) Endl. from the Himalayas has often been indicated as the living analogue. Foliage very similar to the Icelandic specimens is also found in A. subcordata C. A. Meyer from the southern Caspian Sea and A. japonica Sieb. et Zucc. Material: Brjánslækur. Alnus cecropiifolia (Ettingshausen) Berger (Figs 56– 58) Alnus kefersteinii auct. (non (Goepp.) Goepp.); Heer, 1868: 146, pl. 25, fig. 9b (Husawik [Húsavík]); Windisch, 1886: 35, partim (foliage) (Brianslaekr [Brjánslækur], Tröllatunga). Alnus sp.; Friedrich, 1966: 70, pl. 1, fig. 13, pl. 2, figs 10, 11, text-fig. 18 (Seljá in Vadalsdalur [Seljá in Va~alsdalur], Brjánslaekur [Brjánslækur]). © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 boj_441.fm Page 381 Wednesday, November 23, 2005 3:37 PM MIOCENE FLORAS FROM ICELAND 381 52 53 54 56 55 57 58 59 60 61 62 Figures 52–62. Figs 52–55. Alnus cf. gaudinii, Brjánslækur, S087465-01, leaf and venation details. Figs 56–58. Alnus cecropiifolia, Tröllatunga, S087416, leaf and venation details. Figs 59–62. Alnus cf. kefersteinii, infructescences. 59, 60, Gautshamar, S094150A, S094232. 61, Tröllatunga, S106556. 62, Hestabrekka, S106900. Scale bars = 5 cm in Figs 52, 56, and 1 cm in Figs 53-55, 57-62. © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 boj_441.fm Page 382 Wednesday, November 23, 2005 3:37 PM 382 T. DENK ET AL. B. pseudolumnifera and B. cristata. Dentition in B. cristata is quite similar to B. pendula Roth. Material: Mókollsdalur (Akhmetiev et al., 1978); Hre~avatn, Fífudalur, Hestabrekka, Primilsdalur/ Porvaldsdalur; Vindfell. Betula subnivalis Lindquist (Figs 71, 72) B. macrophylla auct. (non (Goepp.) Heer); Heer, 1868: 146–147, pl. 25, figs 13–15 (Hredavatn [Hre~avatn]). ? B. prisca auct. (non Ettingsh.); Heer, 1868: 148, pl. 25, figs 22–25 (Hredavatn [Hre~avatn]). B. forchhammeri Heer, 1868: 148, 149, pl. 25, figs 28, 29 (Hredavatn [Hre~avatn]). Betula sp. Lindquist 1947: 346, fig. 4: 3, 4 (Phrimils~alur [Primilsdalur]). Betula subnivalis Lindquist; Akhmetiev et al., 1978: pl. 10, fig. 6, pl. 11, 10, 14–16 (Hredavatn [Hre~avatn]. Betula sp., Friedrich, Símonarson & Heie 1972: 8, pl. 1, fig. 3 (Mókollsdalur). Betula sp. 1; Akhmetiev et al., 1978: pl. 11, figs 2–4 (Hredavatn [Hre~avatn]. Description: Scales of fruiting catkins, 6–7 mm long, 4–5 mm wide, trilobed, central lobe longer than lateral lobes, lobes elliptic, rounded at apex. Discussion: Such scales are comparable with a number of East Asian and North American species of Betula. Closest similarities are to B. lenta L., B. papyrifera var. cordifolia Fern. (North America) and B. potaninii Batal. (East Asia). Material: Hre~avatn (Primilsdalur/Porvaldsdalur). Betula islandica sp. nov. Denk, Grímsson & Kvadek (Figs 75–80) Corylus sp., Friedrich, 1966: 74, pl. 2, figs 1, 2, 7, 8, text-fig. 20 (Brjánslaekur [Brjánslækur]); Friedrich, 1968: pl. 2, fig. 2a, b (Brjánslækur). Betula sp., Friedrich & Símonarson, 1983, figs 6, 10 (Brjánslækur). Holotype designated herewith: Specimen no. S08742201 (figs 75–77); Paratypes specimen nos. S094058; S093963. Betula L. Betula cristata Lindquist emend. (Figs 63–70, 73, 74; Epitype Figs 63–65) Betula macrophylla auct. (non (Goepp.) Heer); Heer, 1868: 146–147, pl. 25, figs 11, 12, 16–19 (Hredavatn [Hre~avatn]). Betula cristata Lindquist, 1947: 346, fig. 1: 1–5, fig. 2: 1–5 (Phrimils~alur [Primilsdalur]. Betula sp., Friedrich, Símonarson & Heie, 1972: 8, pl. 3, fig. 1 (Mókollsdalur). Betula macrophylla auct. (non (Goepp.) Heer); Akhmetiev et al., 1978: pl. 9, figs 3, 10, 12 (Mokollsdalur [Mókollsdalur]), pl. 10, figs 1, 9, 11–13, pl. 11, fig. 1 (Hredavatn [Hre~avatn]. Description: Leaves petiolate, petiole 7+ mm long, lamina ovate to elliptic, 6–10+ cm long, 3.5–7.5 cm wide, serrate, base cordate, apex acute, secondary veins craspedodromous, 8–10 pairs, pectinal veins and their abmedial branches and external veins inserting teeth, teeth triangular with attenuate glandular apex. Discussion: Lindquist (1947) selected a few specimens from Primilsdalur (Hre~avatn–Stafholt Formation) to establish the new species, Betula cristata Lindq., which he compared with the modern B. pubescens Ehrh. (syn. B. callosa Notö, B. tortuosa Ledeb., B. coriacea Gunnarss.) from Eurasia. All the specimens he chose, however, are apices of larger leaves. Examining numerous birch leaves from different localities representing the Hre~avatn–Stafholt Formation clearly shows that these fragments belong to larger leaves, which have nothing in common with B. pubescens. Therefore, we emended the diagnosis of Lindquist (1947). Betula macrophylla (Goepp.) Heer (i.e. Alnus macrophylla Goepp., 1855) based on material from Sornica, Poland, was transferred into the new taxon (combination) Alnus adscendens by Zastawniak & Walther (1998). Although this transfer seems to be superfluous, this species certainly characterizes a type of betulaceous foliage that is different from that described above from Iceland. The leaves described here are comparable with B. pseudolumnifera Givulescu, a birch foliage from the Upper Miocene of southern and western Europe. The latter has been compared with the modern Japanese B. maximowicziana Regel (Kvadek et al., 2002). However, teeth are slightly more attenuate and more densely spaced in B. maximowicziana than in Figures 63–74. Figs 63–70, 73, 74. Betula cristata. Figs 63–67. Primilsdalur. 63–65, S094978, leaf, close-ups of leaf margin. 66, 67, S094898, leaf, close-up of leaf margin. Figs 68, 73. Hestabrekka, S094463, S094454, leaves. Figs 69, 70. Fífurdalur, S094094, leaf, close-up of leaf margin. Fig. 74. Primilsdalur/Porvaldsdalur, S094882, leaf. Figs 71, 72. Betula subnivalis. Primilsdalur/Porvaldsdalur. 71, S106896-01, fruit scales. 72, S094927, fruit scale. Scale bars = 5 cm in Figs 63, 66, 68, 69, 1 cm in 64, 65, 67, 70, 71, 73, 74, and 0.5 cm in 72. © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 boj_441.fm Page 383 Wednesday, November 23, 2005 3:37 PM MIOCENE FLORAS FROM ICELAND 383 64 66 63 65 67 70 68 69 71 72 73 74 © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 boj_441.fm Page 384 Wednesday, November 23, 2005 3:37 PM 384 T. DENK ET AL. 75 78 79 76 77 80 Figures 75–80. Figs 75–79. Betula islandica sp. nov., Brjánslækur. 75–77, S087422-01. Holotype. 78-79, S094058, leaf, close-up of leaf margin. Fig. 80. Betula sect. Costatae sp., Brjánslækur, S093963, fruit scales with long and narrow lobes. Scale bars = 5 cm in Figs 75, 78, 1 cm in Figs 76, 77, 79, 80. Type locality: Surtarbrandsgil gully, close to Brjánslækur, Brjánslækur-Seljá Formation, 12 Ma, north-western Iceland. Etymology: The species is named after Iceland. Diagnosis: Leaves petiolate, lamina ovate, large, coarsely serrate, base cordate, apex acute to acuminate, secondary venation craspedodromous, glandular teeth supplied by secondary veins, abmedial branches of pectinal veins, or external veins, or by veinlets branching off from tertiary veins, main teeth slightly © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 boj_441.fm Page 385 Wednesday, November 23, 2005 3:37 PM MIOCENE FLORAS FROM ICELAND larger than secondary teeth, with convex basal and apical sides and acuminate apex. Description: Leaves petiolate, petiole 6–16 mm long, lamina broad ovate, serrate, 8–12.5 cm long, 5–7 cm wide, base cordate, apex acute to acuminate, secondary veins craspedodromous (9 to) 10–11 pairs, inserting (primary) teeth, abmedial branches of pectinal veins, external veins and veinlets that branch off at an angle of 90∞ from tertiary veins inserting (secondary) teeth, teeth of more or less equal size, broadly triangular, basal and apical sides convex, apex acuminate, glandular. Cuticles reflect only straight-walled polygonal cells of the adaxial epidermis and 4–6-cellular bases of glandular trichomes scattered on veins, more common in smaller specimens. Discussion: This fossil species falls within the variability of the modern section Costatae (Regel) Koehne based on its large leaf size. Within the section Costatae B. islandica shows similarities to the modern North American Betula alleghaniensis Britt. (syn. B. lutea Michx.), and to the Eurasian species B. utilis D. Don and B. ermanii Cham. Material: Brjánslækur. Betula sect. Costatae (Regel) Koehne sp. (Fig. 80) Betula sp., Friedrich, 1966: 74, pl. 2, figs 5, 6 (Brjánslaekur [Brjánslækur]. Betula sp., Akhmetiev et al., 1978: pl. 1, fig. 8 (?Selárdalur). Description: Scales of fruiting catkins 11–12 mm long, trilobed, with narrow, c. 1 mm wide, linear lateral and central lobes, the central lobe slightly longer than the lateral ones. Discussion: These scales belong to section Costatae based on their long and narrow lobes. Mädler (1939) described very similar scales from the Pliocene of Germany as B. longisquamosa. Among living species, closest similarities are to the East Asian B. delavayi Franchet var. delavayi, and B. chinensis Maxim. var. fargesii Hu ex P. C. Li. Because of the presence of a single species based on leaves and on scales, both belonging to section Costatae, we suggest that these belong to a single biological species. Material: Brjánslækur. Corylus L. Corylus sp. (Figs 81–86) ? C. macquarrii auct. (non (Forbes) Heer); Heer 1868: 149, pl. 26, figs 1a, 2–4 (Brjamslaek [Brjánslækur], Hredavatn [Hre~avatn], Laugavatsdalur [Langavatnsdalur]). 385 Description: Leaves, petiole not preserved, lamina broad ovate to elliptic, serrate, 10–14 cm long, 6–8+ cm wide, base round to slightly cordate, apex acute, secondary veins craspedodromous, 8–11 pairs, with numerous abmedial branches, close to the margin abmedial veins are connected by a tertiary vein running parallel to the margin, from this tertiary vein short veinlets run into teeth, teeth triangular when inserted by secondary veins and almost round when inserted by higher-order veinlets, tooth apex acute to bluntly acute. Discussion: This foliage has nothing in common with that found in Palaeogene deposits from the Isle of Mull, Spitsbergen and Greenland (see Corylites hebridicus in Boulter & Kvadek, 1989; Kvadek et al., 1994). The latter has been called Corylus macquarri by Heer (1868) based on specimens from Lower Cainozoic deposits from Spitsbergen, and Greenland, and here included were also a few leaf fragments from Iceland. Of the latter, one specimen (pl. 26, fig. 2) may belong to Corylus, while the other fragments are Betulaceae with uncertain affinities. Later, C. macquarri from the Early Cainozoic of Greenland and Spitsbergen was included within the extinct betulaceous morpho-genus Corylites Gardner ex Seward et Holttum by Boulter & Kvadek (1989). Leaves of Corylites differ from Corylus by the more densely spaced and higher number of secondary veins, fewer secondary teeth between two primary teeth, as well as the absence of trichomes (observed in material from Mull, Boulter & Kvadek, 1989). Foliage unambiguously ascribable to Corylus appears to be rare in the fossil record of the Northern Hemisphere (cf. Mai, 1995). Leaves named Corylus avellana L. fossilis from the Pliocene of Germany (Knobloch, 1998) belong with certainty to the genus and are very similar to the fossils from Iceland. Among modern species the western Eurasian C. avellana L., and C. colchica Albov., and the East Asian C. chinensis Franch. have similar foliage. Material: Brjánslækur Hre~avatn. (?) Langavatnsdalur (?) BETULACEAE GEN. DIV. INDET. Betula prisca auct. (non Ettingsh.); Heer, 1868: 148, pl. 25, fig. 9a (Husawik [Húsavík]), fig. 20 (Sandafell [Sandfell]), specimens lost, originally in the Copenhagen collections, pl. 26, fig. 1b, c (Brjamslaek [Brjánslækur]). These specimens figured in Heer (1868) belong to different taxa, of which some might not belong to Betulaceae. © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 boj_441.fm Page 386 Wednesday, November 23, 2005 3:37 PM 386 T. DENK ET AL. 81 82 83 84 85 86 Figures 81–86. Corylus sp., Brjánslækur, leaves, close-ups of leaf margin. 81–83, S094065-01. 84, 85, S093938-01. 86, S094037. Scale bars = 5 cm in Figs 81, 84, 86, 1 cm in Figs 82, 83, 85. ERICACEAE Rhododendron L. Rhododendron aff. ponticum L. (Figs 87–90) ? (?) Rhododendron sp., Akhmetiev et al., 1978: 146, pl. 10, fig. 14 (Hredavatn [Hre~avatn]). Description: Leaves petiolate, petiole 8–22 mm long, lamina elliptic to obovate, entire, 5–15 cm long, 2.5–4.6 cm wide, base acute to rounded, sometimes slightly decurrent, apex bluntly acute with a pointed tip, secondary veins eucamptodromous to brochidodromous, 8–12 pairs, sometimes intersecondary veins hard to distinguish from secondary veins. Discussion: Without epidermal structures preserved it is difficult to decide whether the leaves from © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 boj_441.fm Page 387 Wednesday, November 23, 2005 3:37 PM MIOCENE FLORAS FROM ICELAND 387 91 92 88 87 93 94 89 90 95 96 Figures 87–96. Figs 87–90. Rhododendron aff. ponticum, Tröllatunga, S087459, S106760-01, S106760-02, S106740-01. Figs 91–94. Vaccinium sp. (evergreen), Tröllatunga. 91, 92, S106621, leaf, close-up of leaf margin showing glandular teeth. 93, 94, S106624, leaf, close-up of leaf margin. Figs 95, 96. ‘Arctostaphylos’ (deciduous), Tröllatunga, S106768, leaf, closeup of leaf margin. Scale bars = 5 cm in Figs 87–90, 1 cm in Figs 91, 93, 95, 3 mm in Figs 92, 94, 96. © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 boj_441.fm Page 388 Wednesday, November 23, 2005 3:37 PM 388 T. DENK ET AL. Tröllatunga belong to this genus or not. Comparison with modern species, however, appears to support the generic identification. Beside R. ponticum from the southern and eastern Black Sea region, a number of eastern North American and East Asian large-leaved species closely resemble the fossils. We doubt that the leaf figured in Akhmetiev et al. (1978) from the Hre~avatn-Stafholt Formation and referred to (?) Rhododendron belongs to the same plant type as that described here. Material: Tröllatunga. Vaccinium L. Vaccinium sp. (Figs 91–94) ? Vaccinium islandicum Windisch, 1886: 41, text figs 1–3 (no locality indicated). Description: Leaves petiolate, petiole 2+ mm long, lamina elliptic, dentate, 2–3 cm long, 0.7–1.5 cm wide, base rounded to acute, apex acute, secondary veins not clearly visible, teeth glandular, appressed, basal side much longer than apical side, at high magnification leaf surface densely beset with dark-shining dots, which appear to be glands. Discussion: Windisch (1886) based his new species on two specimens (part and counterpart). His description fits well with the specimens described here. However, his line drawings show spinose teeth that are quite different from our specimens, and do not resemble Ericaceae. Glands as found in the fossil leaves are typical of some evergreen species of Vaccinium. Material: Tröllatunga. cf. ‘Arctostaphylos’ sp. (Figs 95, 96) Phyllites vaccinioides Heer?, Heer, 1868: 154, pl. 27, fig. 13 (Hredavatn? [Hre~avatn]). Description: Lamina elliptic, dentate, no petiole preserved, c. 14 mm long, 8 mm wide, base acute?, apex rounded, slightly emarginate, secondary veins brochidodromous, about 7 pairs, first-order loops followed by second-order loops from which small veins insert the teeth, teeth small, appressed, glandular. Discussion: These leaves closely resemble deciduous species of Arctostaphylos and Vaccinium. Material: Tröllatunga. FAGACEAE Fagus L. Fagus gussonii Massalongo (Figs 97–101) ? Fagus deucalionis Unger; Heer, 1868: 149, pl. 25, fig. 32 (Brjamslaek [Brjánslækur]). Fagus sp., Friedrich, Símonarson & Heie, 1972: 8, pl. 1, fig. 3 (leaf), 5 (cupule), pl. 3, fig. 3 (cupule) (Mókollsdalur). Fagus orientalis Lipsky; Akhmetiev et al. 1978: pl. 8, fig. 10, pl. 9, fig. 9 (cupule) (Mokollsdalur [Mókollsdalur]). Fagus gussonii Massalongo, Grímsson & Denk, 2005: 43–50, pl. 10–13 (Mókollsdalur). Description: Leaves and cupules; leaves petiolate, petiole up to 13.5 mm long, lamina elliptic to broad ovate to obovate, crenulate, dentate, or teeth absent, 8–15+ cm long, 4–8 cm wide, base acute, asymmetrically rounded or slightly cordate, apex acute or acuminate, pairs of secondaries 10–16, secondary venation pseudocraspedodromous to semicraspedodromous, rarely craspedodromous, teeth with long basal side and short apical side, rarely small triangular teeth following a convex sinus. Cupules 1.5–2 cm long, peduncle thickened when inserting cupule valves, i.e. a connecting-piece sensu Denk & Meller (2001) present, valves beset with spine-like appendages. Discussion: Such leaves have been described from the Late Miocene of Spain, Italy, and Greece as F. gussonii (Massalongo & Scarabelli, 1859; Barrón & Diéguez, 1994; Kvadek et al., 2002). They are associated with large cupules as described above. Cainozoic cupules of Fagus lack species-diagnostic features and have been referred to a single species F. deucalionis Unger (Denk & Meller, 2001). As there is only a single type of leaves in the Mókollsdalur horizon, we refrain from a specific determination of the cupules. Both leaves and cupules of Fagus gussonii are comparable with the modern western Eurasian F. sylvatica L. (including F. orientalis Lipsky) and with the Chinese F. longipetiolata Seemen. Another type of Fagus foliage is the most common element in the Selárdalur locality (Grímsson & Denk, 2005), which belongs to the oldest plant-bearing sedimentary formation on Iceland (15 Ma; Hardarson et al., 1997; Kristjansson et al., 2003). The leaf from Brjánslækur figured in Akhmetiev et al. (1978: pl. 5, fig. 1) is too fragmentary to allow a closer determination. Based on the course of the secondary veins and their thickness it belongs to Betulaceae. Material: Mókollsdalur. Specimen M-243 stored at Department of Earth Sciences, University of Aarhus. © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 boj_441.fm Page 389 Wednesday, November 23, 2005 3:37 PM MIOCENE FLORAS FROM ICELAND 389 97 98 99 100 101 102 Figures 97–102. Figs 97–101. Fagus gussonii, Mókollsdalur, leaves. 97, S094861B. 98, S105948. 99, S105955. 100, S105944, close-up of leaf margin showing typical pseudocraspedodromous venation. 101, M-243, dentate leaf. 102, Mókollsdalur, S105957, imprint of a large cupule. Scale bars = 5 cm in 97–99, 101, 1 cm in 100, 102. JUGLANDACEAE cf. Pterocarya/Cyclocarya sp. (Figs 103–109, 112–114) ? Pterocarya paradisiaca (Unger) Iljinskaya; Akhmetiev et al., 1978: pl. 6, fig. 12 (Brjanslaekur [Brjánslækur]). Carya sp., Símonarson, 1991: 144, fig. 1 (Tröllatunga). Description: Lamina elliptic, serrate, 6–16 cm long, 3–6 cm wide, base acute to asymmetrically rounded, apex acute, a large leaflet with petiolule preserved, © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 boj_441.fm Page 390 Wednesday, November 23, 2005 3:37 PM 390 T. DENK ET AL. 104 106 107 103 105 111 108 109 110 112 113 114 Figures 103–114. Figs 103–109, 112, 114. Juglandaceae aff. Pterocarya/Cyclocarya (non Pterocarya fraxinifolia), Tröllatunga. 103, 104, S106525-01, leaf, close-up showing leaf margin and higher order venation. 105, S106722b, leaf. 106, 107, S106630-01, leaf, close-up of leaf margin. 108, 109, S106776-04, leaf, close-up of leaf margin. 112, 113, S106631, leaf, closeup of leaf margin. 114, S106722a-01, leaf. Figs 110, 111. ?Juglandaceae (aff. Juglans), Tröllatunga, S094044-02, leaf fragment, close-up of leaf margin. Scale bars = 5 cm in Figs 103, 105, 106, 110, 112, 114, 2.5 cm in Fig. 108, 1 cm in Figs 104, 107, 109, 111, 113. © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 boj_441.fm Page 391 Wednesday, November 23, 2005 3:37 PM MIOCENE FLORAS FROM ICELAND petiolule c. 7 mm long, secondary venation semicraspedo-brochidodromous, primary loops followed by secondary loops from which veins run into teeth, teeth with long basal side and short apical side, tooth apex acute to acuminate; tertiary veins perpendicular to secondary veins close to the margin, oblique towards the midvein. Discussion: In a short note Símonarson (1991) figured and described a leaf from Tröllatunga as Carya sp. (based on examination by Walter Friedrich). Most of the juglandaceous leaves from Tröllatunga appear to belong to a single species. It is, however, quite difficult to assign them with certainty to a particular genus within the Juglandaceae. In general, species of Pterocarya (and Cyclocarya) often have semicraspedobrochidodromous venation as found in the fossils, whereas secondaries in Carya and Juglans either split before reaching the margin, each branch inserting a tooth (craspedodromous), or form loops (brochidodromous to eucamptodromous) that reach almost to the margin, from which small tertiary veins run into the teeth. There are, however, some exceptions within Carya and Juglans with some species displaying some similarity to the fossils from Tröllatunga. Among Carya, C. cathayensis Sarg. sometimes shows semicraspedo-brochidodromous venation with the loops not reaching close to the margin so that they are followed by second-order loops. The same is true for some specimens of Juglans cinerea L. Among Pterocarya P. fraxinifolia Spach. has more widely spaced secondary veins and blunter teeth than the fossils, and primary loops reach closer to the margin. A number of East Asian species are comparable with the fossil type (e.g. P. rhoifolia Sieb. et Zucc., P. stenoptera DC. and P. tonkinensis (Franch.) Dode). In addition, the Icelandic fossils resemble the Eurasian fossil species Pterocarya paradisiaca (Ung.) Iljinsk. and, to some extent, also P. denticulata (Weber) Heer (i.e. Carya denticulata (Weber) W. Schimper; cf. Dorofeev & Iljinskaja, 1994). Ferguson (1971) pointed out the difficulties in trying to assign Juglandaceae-like leaves to a particular genus of the family and even to the family itself. According to Akhmetiev et al. (1978) both juglandaceous pollen and pollen ascribable to Juglans and Pterocarya are present in the Brjánslækur-Seljá, the TröllatungaGautshamar and the Skar~sströnd-Mókollsdalur formations. Whereas Pterocarya is usually represented by few pollen grains except for in some samples from Tröllatunga, Juglans pollen appears to be a codominant element in most of the pollen assemblages. Material: Tröllatunga, Húsavík, Mókollsdalur. Material from Mókollsdalur stored at Department of Earth Sciences, University of Aarhus. 391 cf. Juglans (Figs 110, 111) Juglans bilinica auct. (non Unger); Heer, 1868: 153, pl. 28, figs 14–17 (Brjamslaek [Brjánslækur]); Windisch, 1886: 49 (Brianslaekr [Brjánslækur). Description: Leaflet fragment, upper portion of lamina 7+ cm long, dentate, secondary veins irregularly spaced, curved, semicraspedodromous-brochidodromous, primary loops reaching close to margin, teeth with long basal and short apical side, tooth apex acute, c. 3 teeth along the sinus between two adjacent secondary veins, tertiary veins perpendicular to secondary veins, simple or branched. Discussion: Incompletely preserved leaves (leaflets) from Brjánslækur were assigned to J. bilinica by Heer (1868). Friedrich (1966) re-examined the original specimens of Heer’s publication and considered them either Juglans or Carya. Also Windisch (1886) was uncertain whether they were more similar to Juglans or to Carya and whether such leaf remains could be ascribed to Juglandaceae at all. Another leaf from Brjánslækur figured in Friedrich & Símonarson (1983) as Juglans is different from the specimens figured in Heer (1868) and appears to belong to Alnus gaudinii (see above). Juglans bilinica Unger has recently been transferred to Fraxinus on account of the cuticle structure of topotypic specimens (Kvadek & Hurník, 2000). Also the leaf described in the present study cannot be assigned to Juglandaceae with certainty. Although some modern species of Juglans, e.g. J. nigra L., display a similar type of secondary venation and dentition this leaf may also belong to another genus of Juglandaceae or even to Alnus gaudinii. Material: Brjánslækur. Pterocarya Kunth. Pterocarya sp. Pterocarya sp., Friedrich, Símonarson & Heie, 1972: 9, pl. 2, figs 1 (leaf with leaflets), 2 (winged fruit) (Mókollsdalur). Discussion: Leaves with leaflets and large winged fruits (wings 20 mm long and > 20 mm wide) of Pterocarya were reported from Mókollsdalur. Leaves are very similar to the modern P. fraxinifolia (Lam.) Spach. from the eastern Black Sea and southern Caspian Sea areas. They differ by their blunt teeth from the more abundant specimens from the TröllatungaGautshamar Formation referred to as to Pterocarya/ Cyclocarya (see above). The large samaras of the specimen from Mókollsdalur are comparable with some East Asian species of Pterocarya, such as P. macroptera Batalin s.l. © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 boj_441.fm Page 392 Wednesday, November 23, 2005 3:37 PM 392 T. DENK ET AL. Magnolia sp., Akhmetiev et al., 1978: pl. 4, fig. 1a (Brjanslaekur [Brjánslækur]). Description: Leaves, no petiole preserved, lamina elliptic to obovate, entire, 8–20 cm long, 2.5–7 cm wide, base rounded (to bluntly acute) or very-base oblong, apex acute, secondary venation brochidodromous, 9–14 pairs of secondary veins. The obtained cuticle structure is the same in very large and much smaller leaves, suggesting a large range of size variation. Anticlinal walls of cells are finely undulating both on the adaxial and abaxial leaf side. On a fragment of the costal area, dense rounded simple serial trichome bases, rarely with an attached long-terminal part have been observed. This is a feature of some Magnoliaceae, particularly Michelia and some species of Magnolia (Baranova, 1972). However, the oil cells widely spread in the Magnoliaceae have not been observed in the material at hand. Discussion: Such leaves closely resemble in foliage species of Magnolia L. (East Asia–Himalayas, North and Central America), Michelia L. and Manglietia Blume (both East Asia). Material: Brjánslækur. Liriodendron L. Liriodendron procaccinii Unger (Fig. 124) Liriodendron procaccinii Unger; Heer, 1865: 331, text-fig. 186a; Heer, 1868: 151, pl. 27, figs 5–8 (Brjamslaek [Brjánslækur]). Description: One samaroid fruit with a wing-like lanceolate style, c. 3 cm long; steep reticulate venation running from the pericarp towards the apex of the wing. Discussion: The genus consists of only two modern species, L. tulipifera L. in North America and L. chinensis Sargent in Central China. Liriodendron was a common element in the Cainozoic of the northern hemisphere. In Europe it persisted at least until the Late Pliocene (cf. Knobloch, 1998). Material: Brjánslækur. Material: Tröllatunga, Húsavík, Mókollsdalur (material stored at Department of Earth Sciences, University of Aarhus). LAURACEAE Laurophyllum sp. 1 (Figs 207, 208) Description: A few leaves, lamina elliptic, entire, petiole not preserved, base acute, apex not preserved, c. 7 cm long, 2 cm wide, secondary venation eucamptodromous, secondary veins widely and irregularly spaced, the lowest pair more acute than pairs above, 6–8 pairs; mesophyll tissue of one specimen (S094035) contains lens-shaped oil cells. Discussion: Leaf shape and type of venation plus the presence of oil cells indicate closer affinity to Lauraceae. Foliage of the modern Laurus nobilis L. resembles the fossil leaf. Material: Brjánslækur. Sassafras Trew. Sassafras ferrettianum Mass. (Figs 115–117) Sassafras, Friedrich, 1966: 81, pl. 3, fig. 5, pl. 4, figs 1, 3, text figs 25–27 (Brjánslaekur [Brjánslækur]). Akhmetiev et al., 1978: pl. 3, figs 4, 5 (Brjanslaekur [Brjánslækur]). Description: Leaves, petiole not preserved, lamina simple, ovate-elliptic, entire, 8–15 cm long, 6.5+ cm wide, base cuneate, apex bluntly acute, venation pedate-camptodromous, the basal secondary veins originating from a common point well above the base of the lamina, secondary veins brochidodromous. Oil cells 37–40 mm in diameter were observed in mesophyll tissue of one specimen. No other details of epidermal tissue are available. Discussion: Sassafras was a widespread element in the Cainozoic of Eurasia and North America. At present it comprises only three species displaying an East Asian–North American disjunction. In addition to unlobed leaves, three-lobed forms have been described from Brjánslækur by Friedrich (1966). Material: Brjánslækur. MYRICACEAE Comptonia L’Héritier Comptonia hesperia Berry (Figs 125, 126) Comptonia hesperia Berry; Friedrich, 1966: 68, pl. 1, figs 1, 2, 4, text fig. 17 (Brjanslaekur [Brjánslækur]). ? Dryopteris linneaneiformis Iljinsk.; Akhmetiev et al., 1978: pl. 11, fig. 11 (Holmatindur [Hólmatindur]). MAGNOLIACEAE ? Magnolia sp. (Figs 119–123) Magnolia cf. reticulata Chaney & Sanborn; Friedrich, 1966: 78, pl. 3, figs 1–4, text-figs 23, 24 (Brjánslaekur [Brjánslækur]). © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 boj_441.fm Page 393 Wednesday, November 23, 2005 3:37 PM MIOCENE FLORAS FROM ICELAND 393 115 116 117 118 Figures 115–118. Figs 115–117. Sassafras ferrettianum, Brjánslækur, leaves. 115, S093368-01. 116, S094021. 117, S093993 Fig. 118. cf. Magnolia, Brjánslækur, S094068-01-01, adaxial cuticle displaying undulate epidermis cells, preparation from leaf shown in Fig. 121. Scale bars = 5 cm in Figs 115–117, 50 mm in Fig. 118. © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 boj_441.fm Page 394 Wednesday, November 23, 2005 3:37 PM 394 T. DENK ET AL. 119 120 121 123 122 124 125 126 Figures 119–126. Figs 119–123. Magnolia sp., Brjánslækur, leaves. 119, S093996. 120, S094010-02. 121, S094068-01. 122, S094054. 123, S094018. Fig. 124 Liriodendron procaccinii, S094043-02, pericarp. Figs 125, 126. Comptonia hesperia, Brjánslækur, S094066-01, leaf, close-up showing secondary venation. Scale bars = 5 cm in Figs 119–123, 1 cm in Figs 124– 126. © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 boj_441.fm Page 395 Wednesday, November 23, 2005 3:37 PM MIOCENE FLORAS FROM ICELAND Description: Leaves elliptical, dissected, sessile, fernlike, base decurrent, apex missing, each lobe of the leaf supplied by two secondary veins, sinus between the lobes rounded to acute, apex of lobe acute. Discussion: These leaf remains are also similar to narrow forms of C. oeningensis A. Braun (~ Myrica oeningensis (A. Braun) Heer), which differs in producing incompletely dissected leaves (‘vindobonesis’-type). According to Zhilin (1980) and Friedrich (1966), it is extremely difficult to assess specific relationships of fossil Comptonia on the basis of few fragments without knowing the range of variation in the populations under study. Material: Brjánslækur. 395 cordate, apex acute (acuminate), secondary venation brochidodromous, venation pattern close to the margin not preserved, teeth triangular with basal side slightly longer than apical side. Discussion: The few leaves resemble Prunus in their shape, size and dentition. Material: Brjánslækur. Rosaceae gen. et spec. indet., type C (Figs 140, 141) Description: Leaf, petiole not preserved, lamina elliptical, serrate, 7.3+ cm long, 2.9 cm wide, base acute, apex acute, secondary venation brochidodromous, secondary veins branching close to the margin, one branch running to the subsequent secondary vein (forming a loop) the other branch curving towards the apex and sending two further small branches into the tooth and the sinus between two adjacent teeth. Discussion: We tentatively assigned this leaf to Rosaceae. We cannot exclude that it belongs to the same type of plant as do the leaves assigned to Rosaceae, gen. et spec. indet. Type B (? Prunoideae) described above. Partly similar leaves have been ascribed to Sorbus cf. uzenensis Huzioka by Knobloch (1998). Material: Brjánslækur. Crataegus L. cf. Crataegus sp. (Figs 142, 143) Description: One fragmentary leaf probably showing the apical part of the leaf, margin coarsely dentate, each tooth with a small subsidiary tooth at the basal side. Material: Fífudalur. ROSACEAE Rosaceae gen. et spec. indet., type A (? Prunoideae) (Figs 127–136) Ulmus? sp., Friedrich, 1966: 77, pl. 2, fig. 3 (Brjanslaekur [Brjánslækur]). Description: Leaves petiolate, one specimen with petiole preserved, petiole 1.6 cm long, lamina ovate or elliptic, serrate, 6–12 cm long, 3.5–6 cm wide, base rounded, symmetrical, apex (elongate) acute, secondary venation brochidodromous to craspedodromous (apically), from the loops small veins inserting teeth, teeth triangular with basal side slightly longer or as long as apical side, both sides ± convex, gland-like dark spots in sinuses between two teeth, sometimes teeth with glandular tip, occasionally very small subsidiary teeth visible. Discussion: Leaves from the Middle Miocene of North America displaying very similar tooth architecture have been assigned to Amelanchier Medik. (Schorn & Gooch, 1994). They differ from the Icelandic leaves by their entire margin at the basal part of the leaf. There are, however, also species of Amelanchier with teeth along the entire margin, such as Amelanchier asiatica Endl. ex Walp. Moreover, similar leaves from the Neogene of Central Europe have been ascribed to Cerasus (Adans.) Focke (= Prunus L.) by Knobloch (1998). We tentatively assign our leaves to Prunoideae within the Rosaceae, because of their overall similarity with Prunus. Material: Brjánslækur, Seljá, Tröllatunga, Hre~avatn. Rosaceae gen. et spec. indet., type B (? Prunoideae) (Figs 137–143) Description: Leaves, no petiole preserved, lamina elliptic, serrate, c. 9 cm long, 4 cm wide, base slightly SALICACEAE Populus L. Populus sp. 1 (ex group Populus tremula L.) (Figs 144– 148) Populus latior Al. Braun cf. forma transversa Heer; Friedrich, 1966: 65, text-fig. 15 (Seljá/Vadalsdalur [Seljá/Va~alsdalur]). ? Populus sp., Akhmetiev et al., 1978: pl. 6, fig. 3 (Husavik [Húsavík]). Description: Leaves petiolate, petiole 4.7 cm long, lamina roundish, dentate along entire margin, lamina of two ± complete leaves 8.5 ¥ 8 and 6 ¥ 6.4 cm, base slightly cordate, apex round, venation pinnate, c. 4 © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 boj_441.fm Page 396 Wednesday, November 23, 2005 3:37 PM 396 T. DENK ET AL. 127 128 129 130 131 132 133 134 135 136 Figures 127–136. Rosaceae gen. et spec. indet, type A (?Prunoideae). Figs 127–131. Brjánslækur. 127, 128, S094017, leaf, close-up of leaf margin showing higher order venation and dentition. 129-131, S093943, leaf, close-ups of margin. Figs 132, 133. Seljá, S134378-01, small leaf, close-up of leaf margin. Figs 134–136. Tröllatunga, S106524, leaf with long petiole, close-ups of leaf margin. Scale bars = 3 cm in Figs 127, 129, 132, 134, 1 cm in Figs 128, 130, 131, 136, 0.5 cm in Figs 133, 135. © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 boj_441.fm Page 397 Wednesday, November 23, 2005 3:37 PM MIOCENE FLORAS FROM ICELAND 397 138 139 137 141 140 142 143 Figures 137–143. Figs 137–139. Rosaceae gen. et spec. indet, type B (?Prunoideae), Brjánslækur, S094044-01, elliptic leaf, close-ups of leaf margin. Figs 140, 141. Rosaceae gen. et spec. indet., type C, Brjánslækur, S093734, leaf, close-up of leaf margin. Figs 142, 143. cf. Crataegus sp., Fífurdalur, S094081, leaf fragment, close-up of margin. Scale bars = 5 cm in Figs 137, 140, 2.5 cm in. Fig. 142, 1 cm in 138, 139, 141. © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 boj_441.fm Page 398 Wednesday, November 23, 2005 3:37 PM 398 T. DENK ET AL. 144 145 146 147 148 Figures 144–148. Populus sp. 1 (ex group Populus tremula L.) Seljá. 144, 145, S134370, leaf, close-up showing glandular teeth. 146, S134363, leaf, arrowhead indicates proximal end of petiole. 147, 148, S134369-01, leaf fragment, close-up showing higher order venation, glandular teeth. Scale bar = 5 cm 144, 146, 1 cm in 145, 147, 148. © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 boj_441.fm Page 399 Wednesday, November 23, 2005 3:37 PM MIOCENE FLORAS FROM ICELAND pairs of secondaries curved towards the apex, the basalmost pair (pectinal veins) with several abmedial branches, these forming loops from which smaller veins insert the teeth, teeth broad triangular to appressed, tooth apex glandular, teeth variable in size and density. A smaller lamina of this type (4 ¥ 4 cm) has been reported (Friedrich, 1966). Discussion: These leaves are distinct from another common European Neogene Populus, P. populina (Brongn.) Knobloch, by their more pronounced pinnate venation. They differ from leaves of the Miocene of Sornica, which Goeppert (1855) figured as P. balsamoides Goepp. and P. emarginata Goepp. by their smaller number of lateral veins. Furthermore, the leaves from Iceland differ from leaves from Öhningen described as P. latior A. Braun by Heer (1856) in their smaller length to width ratio and the dentate margin at the leaf base. They belong, however, to the section Populus, where they can be compared with the modern North American P. tremuloides Michx. and the European P. tremula L. The latter display similar variation in the size and density of dentition. Material: Húsavík. Seljá (Va~alsdalur), Brjánslækur (?) 399 Description: Leaves, lamina 6.7+ cm long and 5+ cm wide, the basal part missing, symmetrical, ?ovate, leaf apex acute to attenuate, margin incompletely preserved, finely crenulate to entire, venation pinnate, secondary venation camptodromous, secondary veins gradually diminishing towards apex, connecting to following secondaries by forming inconspicuous marginal loops, secondary veins alternate or opposite, diverging from the midvein at angles of 56–76∞ and originating at intervals of 4–8 mm in the middle part of the leaf, intersecondary veins originating from the primary midvein, 0–2 intersecondary veins between secondary veins, tertiary veins perpendicular to oblique to secondary veins, and usually forked, 5–6 tertiary veins per 1 cm secondary vein, marginal ultimate venation forming loops. Discussion: The type of venation and margin, along with the shape of the leaves are typical of the genus Populus. This is the first report of Populus from the Hre~avatn-Stafholt Formation. Similar leaves have not been described from the Miocene of Iceland before. Material: Hestabrekkusund (HHK-16 and 7b), material stored in Reykjavík (coll. F. Grímsson). Populus sp. 3 (Fig. 151) Populus sp. 2, Akhmetiev et al., 1978: pl. 7, fig. 3 (Hredavatn [Hre~avatn]). Populus sp. 2 (Figs 149, 150) 149 150 151 Figures 149–151. Figs 149, 150. Populus sp. 2, Hestabrekkusund, apical part of leaf, part and counterpart. Fig. 151. Populus sp. 3, female catkin. Scale bar = 5 cm in Figs 149, 150, 1 cm in Fig. 151. © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 boj_441.fm Page 400 Wednesday, November 23, 2005 3:37 PM 400 T. DENK ET AL. Description: Leaves and fruitlets; leaves petiolate, petiole rarely preserved, 1.8 cm long, dilated at proximal end, lamina elliptic, 4–15 cm long, 3–5 cm wide, entire or crenulate close to the leaf base and often dentate towards the apex, dentition sometimes inconspicuous, base rounded to acute, apex bluntly acute, secondary venation eucamptodromous to brochidodromous, curved towards the apex and running almost parallel to the margin, secondary veins irregularly spaced, intersecondaries and/or tertiary veins typically running perpendicular to primary vein, small veins originating from secondaries inserting teeth, teeth appressed, reduced to glandular tips, basal side much longer than apical side; cuticles smooth, adaxial with solitary simple trichome bases, abaxial densely hairy on veins, stomata visible as spindle-shaped traces of the guard cells. Capsules 10+ mm long, bottle-shaped with a narrow curved apical part, dehiscing by two recurved valves. Discussion: The Iceland willow differs from the Central European Salix varians Goepp. (incl. S. macrophylla Heer) by its indistinctly toothed margin, widely spaced secondaries and fewer intersecondaries. Hantke (1954) in his revision of the Schrotzburg flora interpreted less distinctly dentate and entire leaf forms (including S. tenera A. Braun) as extreme forms of S. lavateri A. Braun, which is a narrow-leaved species that has nothing in common with our new willow from Iceland. The preserved epidermal structure does not differ from S. lavateri in the type of the pubescence but a more detailed comparison is unreliable due to the poor state of preservation of the only compression from Brjánslækur studied. Specimens similar to S. gruberi have also been described from the Late Miocene of Alaska as S. kachemakensis Wolfe (Wolfe, 1966). Among modern species the North American S. scouleriana Barr. and particularly the European Salix caprea L. are similar to the Miocene willow from Iceland. Salix caprea shows a comparable variability in leaf size and shape, and has a wavy margin with irregular, shallow, bluntpointed teeth. Material: Brjánslækur, Fífudalur, Gautshamar, Húsavík, Seljá in Va~alsdalur, Hre~avatn, Vindfell. Description: Female catkin with several fruits; catkin approximately 6.4 cm long, capsules spirally arranged around a slender axis, axis around 0.3–0.76 mm wide, capsules 4.5–5.1 mm long, 2.6–3.4 mm wide, length to width ratio 1.3–1.9, capsules wide to narrow obovate, slightly asymmetrical, subsessile with short but relatively stout petiole, distal region of capsules marked by a notch, capsules dehiscing by two or three valves. Discussion: The form of the capsules and the way in which they are attached to the axis are characteristic of female catkins in the genus Populus. Akhmetiev et al. (1978) reported a similar specimen that most likely belongs to the same species. Material: Primilsdalur, Brekkuá (HBK-82b, HPK-36b, HPK-36c), material stored in Reykjavík (coll. F. Grímsson). Salix L. Salix gruberi sp. nov. Denk, Grímsson & Kvadek (Figs 152–162) Salix macrophylla auct. (non Heer, 1856); Heer, 1868: 146, pl. 25, fig. 3a, b (Gaulthvamr [Gautshamar], Hredavatn [Hre~avatn]); Windisch 1886: 34 (Husavik [Húsavík]). Salix varians auct. (non Goepp.); Windisch, 1886: 33 (Brianslaekr [Brjánslækur]). Salix tenera auct. (non A. Braun); Friedrich, 1966: 66, pl. 1, figs 3, 12, text-fig. 16 (Brjánslaekur [Brjánslækur], Seljá/Vadalsdalur [Seljá/Va~alsdalur]). Salix sp., Akhmetiev et al., 1978: pl. 5, figs 3, 6, 7 (Selja [Seljá/Va~alsdalur]), pl. 11, figs 6, 13, 18, 19 (Hredavatn [Hre~avatn]). ? Salix cf. glauca L.; Akhmetiev et al., 1978: pl. 8, figs 2, 9 (Mokollsadalur [Mókollsdalur]). Holotype designated herewith: Specimen no. S134360, figs 152, 153 (Seljá/Va~alsdalur), Paratypes specimen nos. S093954-01, S094075-02, S094107-01, S094628, S094633, S134358. Type locality: Seljá in Va~alsdalur, Brjánslækur-Seljá Formation, 12 Ma, north-western Iceland. Etymology: The species is named for Gerwin Francisco Gruber, who collected most of the specimens. Diagnosis: Leaves petiolate, lamina elliptic, margin entire, inconspicuously dentate or crenulate, teeth reduced, glandular, confined to apical part of leaf, secondaries curved towards apex, irregularly spaced, eucamptodromous to brochidodromous, some of the intersecondary and tertiary veins typically perpendicular to primary vein. SAPINDACEAE (INCL. ACERACEAE) Acer L. Acer crenatifolium Ettingshausen ssp. islandicum (Heer) comb. et stat. nov. (Figs 163–169) Vitis islandica Heer, 1859: Fl. Tert. Helv. III, p. 319 (basionym); Heer, 1866: 150, pl. 26, figs 1f, 7a ([Bramslaek] Brjánslækur). © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 boj_441.fm Page 401 Wednesday, November 23, 2005 3:37 PM MIOCENE FLORAS FROM ICELAND 401 154 152 153 155 156 157 158 159 160 161 162 Figures 152–162. Salix gruberi sp. nov. Figs. 152–155. Seljá. 152, 153, holotype, S134360, leaf, close-up of leaf margin, arrows indicate glandular teeth. 154, 155, S134358, leaf, close-up of leaf margin. Figs 156, 157, 159. Fífudalur. 156, 157, S094107, leaf, close-up of leaf margin. 159, S094075, two capsules or two halves of a single capsule. Fig. 158. Brjánslækur, S093954-01, leaf. Figs 160–162. Húsavík, S094628, S094633, leaves, close-up of leaf with arrows indicating glandular teeth. Scale bars = 5 cm in 152, 154, 156, 1 cm in 153, 155, 157–162. © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 boj_441.fm Page 402 Wednesday, November 23, 2005 3:37 PM 402 T. DENK ET AL. 163 164 165 166 167 168 169 174 172 170 171 173 Figures 163–174. Figs 163–170. Acer crenatifolium ssp. islandicum, Tröllatunga. 163, 164, S087458, leaf, close-up of leaf margin. 165, 166, S106774, leaf, close-up of leaf margin. 167, S106580-02, samara. 168, S106715, samara. 169, S10671702, samara. Figs 170–174. Acer askelsonii emend. 170, Primilsdalur, S106579-01, leaf. 171, 172, Gautshamar, S094385, leaf, S106898, pericarp, arrowheads indicating zone of attachment of two pericarps and proximal part of wing. 173, 174, Hestabrekka, S116469, pericarp, arrowhead indicating zone of attachment of two pericarps. Scale bars = 5 cm in Figs 163, 165, 170, 2 cm in 167–169, 173, 174, 1 cm in Figs 164, 166, 171, 172. © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 boj_441.fm Page 403 Wednesday, November 23, 2005 3:37 PM MIOCENE FLORAS FROM ICELAND Acer otopterix auct. (non Goepp.); Heer, 1866: 152, pl. 25, fig. 1a, pl. 28, figs 1, 3, 4 (Gaulthvamr [Gautshamar]), 2, 5–8 (Bramslaek [Brjánslækur]). Acer crenatifolium Ettingshausen; Windisch, 1886: 258 (Brianslaekr [Brjánslækur]); Friedrich, 1966: 84, pl. 5, figs 1–3 (Brjánslaekur [Brjánslækur]); Akhmetiev et al., 1978: 177, 178, pl. 3, fig. 1, pl. 4, fig. 1b (Brjanslaekur [Brjánslækur]), pl. 5, figs 4, 9, 10 (Selja [Seljá]). Acer islandicum Friedrich & Símonarson, 1982; nom. illegit. superfl., p. 159, pl. 1, figs 1–4, 6–8, pl. 2, figs 1–6, pl. 3, figs 5, 7, 8, pl. 4, figs 1, 4, pl. 5, figs 1– 4 (Brjánslækur, Tröllatunga). Acer sp. 1, Friedrich & Símonarson, 1982: 162, pl. 3, fig. 1, figs 3–6, 8, pl. 4, fig. 2 (samaras, Brjánslækur, Tröllatunga). Lectotype designated herewith: no. 6778 (Copenhagen coll.) illustrated in Heer, 1866: pl. 26, fig. 1f (Brjánslækur). Description: Leaves and samaras; leaves petiolate, petiole not preserved in most cases, 22+ mm long in one specimen, lamina variable in size, 4–15+ cm long, 3.5–14+ cm wide, palmate, 5-lobed, rarely 3-lobed, lobes serrate, leaf base cordate, apex acute, actinodromous, secondary veins craspedodromous, teeth rather coarse, regularly spaced, basal side slightly convex to straight, apical side straight to concave; obtained cuticle structure shows straight-walled polygonal cells of the adaxial epidermis and less distinct cells with anomocytic stomata of the abaxial epidermis, guard cells elliptic, with a large aperture, solitary simple trichomes on veinlets 180 mm long. Samaras 2.0–2.8 cm long, pericarp 8–10 mm long, 5–8 mm wide, elliptic, with a wide attachment scar, 5–7 mm wide, samaras forming angles of 30–80∞, peduncle 20+ mm long. Discussion: This subspecies is closely related to Acer crenatifolium ssp. crenatifolium, and indistinguishable from this Central European maple by leaf epidermal features (see Friedrich & Símonarson, 1982, and the present paper vs. Walther, 1972). It is part of a group of maples that was widespread during the Cainozoic in the Northern Hemisphere and is comparable with the modern section Rubra Pax that shows a disjunction between East Asia and North America. Based on epidermal features, Walther (1972) found also similarities of A. crenatifolium to the modern A. hyrcanum Fisch. & C. A. Mey. from the Balkans and northern Iran. The fossil species Acer tricuspidatum Braun & Agassiz (Bronn, 1838) differs in mostly densely hairy lower leaf surface and more quadrangular/elliptic shape of the guard cell pairs. As Walther (1972) stated, A. tricuspidatum matches in epidermal 403 features A. saccharinum L. rather than A. rubrum L. Hence, we do not support the reduction of A. crenatifolium to a form of A. tricuspidatum, as done by Procházka & Bù|ek (1975). The name Acer islandicum Friedrich & Símonarson (1982) is illegitimate, being superfluous at the time of its publication, because the authors cited under its synonyms also Vitis islandica Heer, which had priority. Material: Brjánslækur, Gautshamar, Húsavík, Tröllatunga, Mókollsdalur. Acer askelssonii Friedrich & Símonarson (Figs 170– 174) ? Platanus aceroides auct. (non Goepp.), Heer, 1868: 150, pl. 26, fig. 5 (Hredavatn [Hre~avatn]). Acer otopterix auct. (non Goepp.), Heer, 1868: 152, pl. 28, figs 9–11 (Hredavatn [Hre~avatn]), 12 (Tindarfell [Tindafjall]), 13 (Gaulthvamr [Gautshamar]). Acer sp. 2, Friedrich & Símonarson, 1976: 163, pl. 6, figs 1, 2 (Primilsdalur), 3 (Brekkuá). Acer sp. ex sect. Platanoidea Pax, Akhmetiev et al., 1978: 179, pl. 10, figs 3, 4 (Hredavatn [Hre~avatn]). Description: Leaves and samaras; leaves palmate, petiole not preserved, lamina 5-lobed, entire with one or two coarse teeth per lobe, 3–10 cm long, 3.5–8 cm wide, secondary venation craspedodromous or brochidodromous, leaf base cordate, apex attenuate, lobe apex attenuate, teeth triangular acute to acuminate; samaras 5–8 cm long, up to 2 cm wide, with large pericarp, 1.5–3.6 cm long, 0.8–2 cm wide, attachment scar 0.7–1.8 cm wide. Discussion: These leaf remains resemble the modern species A. platanoides L. and A. saccharum Marsh. While A. saccharum is the only North American species of a group of closely related species from western Eurasia and North America including A. hyrcanum and A. opalus Mill. among others (Tian, Guo & Li, 2002; G. W. Grimm & T. Denk, unpubl. data), A. platanoides belongs to a group of Eurasian species including A. campestre L. and A. laetum C. A. Mey. (syn. A. cappadocicum Gleditsch.). The fossil leaves co-occur with large samaras in the Hre~avatn horizon, which were described as Acer askelssonii by Friedrich & Símonarson (1976) and suggested to be most closely related to A. saccharinum among modern maples. The samaras of A. askelssonii, however, display a large zone of attachment of the two pericarps, which is not the case in A. saccharinum showing a reduced pointlike attachment scar. Material: Húsavík, Hre~avatn. Gautshamar, Mókollsdalur, © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 boj_441.fm Page 404 Wednesday, November 23, 2005 3:37 PM 404 T. DENK ET AL. SMILACACEAE Smilax L. Smilax sp. (Figs 175–181) Description: Fragments of two 3–5-veined acute leaf apices with distinct reticulate venation forming irregular narrow meshes, and a rounded base with five basal veins arising from the petiole with steep reticulate higher-order veins between them. None has yielded cuticle structures. Discussion: Similar leaf impressions occur in the Miocene of Europe and have been assigned to Smilax (e.g. Velenovsky, 1881: pl. 2, fig. 23), although the closely related Heterosmilax may also produce such type of foliage. Material: Brjánslækur, Tröllatunga. NYMPHAEACEAE ? Nuphar (Figs 198, 199) Description: Fragments of large round leaves showing some details of venation may belong to an aquatic plant. Similar leaves occur in Nuphar Sm. where secondaries run radially from the leaf base and pinnately from the primary vein with higher order veins running parallel and perpendicular between the secondary veins. Material: Tröllatunga. INCERTAE SEDIS Dicotylophyllum sp. 1 (herbaceous, ‘Brassicaceae’) (Figs 200–202) Description: Lamina elliptic or ovate, 5–9 cm long, 3–7 cm wide, petiole not preserved, base acute to round, apex acute to acuminate, secondary venation brochidodromous, followed by higher-order loops, course of venation very well preserved, leaf margin crenulate. Discussion: These conspicuous leaves may belong to a herbaceous plant, judging from the prominent higher order venation. Similar leaves are found in many families, such as Brassicaceae, Lamiaceae or Primulaceae, which makes a closer determination impossible. Material: Brjánslækur. ULMACEAE Ulmus L. Ulmus cf. pyramidalis Goepp. (Figs 182–188) Description: Two leaves, the lamina 5–7 cm long, 2–2.3 cm wide, base asymmetric, petiole 6–8 mm long, with simple coarsely dentate margin and partly forked secondaries belong undoubtedly to the Ulmaceae. We compare them to a common Miocene elm of Europe U. pyramidalis Goepp., from which they differ in slightly more widely spaced secondaries and a longer petiole. Material: Brjánslækur. GRAMINOID AND AQUATIC PLANTS POACEAE Phragmites Adans. Phragmites sp. (Figs 189–191) Phragmites oeningensis A. Braun; Akhmetiev et al., 1978: pl. 6, fig. 4 (Brjanslaekur [Brjánslækur]). Description: Leaves, rhizomes; sheaths and blades of leaves, and stems or rhizomes with distinct nodes and internodes belong to Phragmites. Material: Seljá in Va~alsdalur, Hre~avatn. Graminoid leaves (Cyperaceae) (Figs 192–197) Arundo sp.; Akhmetiev et al., 1978: pl. 2, fig. 12, pl. 7, fig. 5 (Brjanslaekur [Brjánslækur], Vindfell). Description: Fragments of leaves showing a distinct midrib and parallelodromous venation as found in many members of Cyperaceae. Material: Brjánslækur, Húsavík, Hre~avatn, Vindfell. Dicotylophyllum sp. 2 (‘Lonicera’) (Figs 203–206) ? Lonicera sp., Akhmetiev et al., 1978: pl. 8, fig. 1 (Mokollsdalur [Mókollsdalur]). Description: Leaves petiolate, petiole 3+ mm long, lamina elliptical, entire, 5.5–8+ cm long, 1.7–3 cm wide, base acute or round, apex attenuate, secondary veins irregularly spaced, incomplete brochidodromous (Figs 204, 206), intersecondaries present, main loops followed by additional loops. Discussion: These leaves are distinct by their incomplete brochidodromous venation. They are, however, impossible to assign with certainty to a particular modern family or genus. Similar leaf shape and type of venation can be observed in some species of Lonicera (Caprifoliaceae), e.g. L. alpigena L., which does not necessarily mean that there is a relation between the fossil leaves and the modern genus. © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 boj_441.fm Page 405 Wednesday, November 23, 2005 3:37 PM MIOCENE FLORAS FROM ICELAND 405 175 176 177 178 179 180 181 Figures 175–181. Smilax sp. Figs 175, 176. Brjánslækur, S093953-01, basal portion of lamina. Figs 177–181. Tröllatunga, apical portions of leaves, close-ups of leaf margin showing higher order venation. 177–179, S106746. 180, 181, S106749. Scale bars = 1 cm in Figs 175, 177, 180, 0.5 cm in 176, 178, 179, 181. © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 boj_441.fm Page 406 Wednesday, November 23, 2005 3:37 PM 406 T. DENK ET AL. 183 184 182 185 188 187 186 Figures 182–188. Ulmus cf. pyramidalis Brjánslækur. 182–185, S093964-01, leaf, close-ups of leaf margin and leaf base. 186–188, S094005, small leaf. Scale bar = 5 cm in 182, 186, 1 cm in 183-185, 187, 188. © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 boj_441.fm Page 407 Wednesday, November 23, 2005 3:37 PM MIOCENE FLORAS FROM ICELAND 407 192 189 190 191 193 194 195 196 197 Figures 189–197. Figs 189–191. Phragmites sp. Primilsdalur, S094976, leaf, S094909-01, leaf, S106795, rhizome. Figs 192–197. Cyperaceae gen. et spec. indet. Húsavík, leaves showing parallelodromous venation. 192, 193, S094580. 194, 195, S094542-01. 196, 197, S094542-02. Scale bars = 5 cm in Figs 189–191, 1 cm in Figs 192–196, 0.5 cm in Fig. 197. © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 boj_441.fm Page 408 Wednesday, November 23, 2005 3:37 PM 408 T. DENK ET AL. 198 199 200 201 202 Figures 198–202. Figs 198, 199. ‘Nuphar’, Tröllatunga, S106711, leaf, close-up of lamina showing higher-order venation. 200–202, Dicotylophyllum sp. 1, Brjánslækur, leaves, S093977-01, S094020, S094067. Scale bars = 5 cm. Material: Brjánslækur,?Mókollsdalur. Dicotylophyllum sp. 3 (‘Neolitsea’) (Fig. 209) Description: A single leaf, lamina elliptic, entire, 8+ cm long, 2.5 cm wide, no petiole preserved, base probably acute, apex bluntly acute, modified acrodromous, i.e. three primary veins merged along c. 1 cm from the leaf base, lateral primary veins connected with secondary veins in the upper third of the leaf and forming a brochidodromous pattern, short secondary veins originating from lateral primary veins and running into the leaf margin. Discussion: Similar leaves occur in the genus Neolitsea belonging to the Lauraceae. Without information about epidermal features, however, a closer comparison to modern genera is extremely hampered. Material: Brjánslækur. Dicotylophyllum sp. 4 (Figs 210–212) Description: Leaf fragment, elliptical, dentate, c. 8.5 cm long, 3 cm wide, base not preserved (? acute), apex elongate acute to acuminate, secondary venation brochidodromous, primary loops followed by a series of higher-order loops, teeth spinose with © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 boj_441.fm Page 409 Wednesday, November 23, 2005 3:37 PM MIOCENE FLORAS FROM ICELAND 409 203 204 205 206 207 208 209 210 211 213 214 212 Figures 203–214. Figs 203–206. Dicotylophyllum sp. 2, Brjánslækur, leaves, close-ups showing venation details. 203, 204, S093936-03. 205, S093952. 206, S093937. Figs 207, 208. Laurophyllum, Brjánslækur, S094035-01. 207, entire margined leaf. 208, cuticle displaying oil-cells. Fig. 209. Dicotylophyllum sp. 3, Brjánslækur, S094061, modified acrodromous venation. Figs 210–212. Dicotylophyllum sp. 4, Tröllatunga, S106673, leaf, close-ups showing leaf margin. Figs 213, 214. Dicotylophyllum sp. 5, Tröllatunga, S106780, leaf, close-up showing leaf margin. Scale bars = 5 cm in Figs 203, 205, 207, 209, 210, 3 cm in Fig. 213, 1 cm in Figs 204, 206, 212, 0.5 cm in Figs 211, 214, 50 mm in Fig. 208. © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 boj_441.fm Page 410 Wednesday, November 23, 2005 3:37 PM 410 T. DENK ET AL. 1981, 2002; Mudie & Helgason, 1983) that emphasized strong affinities to ‘eastern deciduous forests’ from North America, we found that taxonomic affinities of the late Mid Miocene to Late Miocene (12–6 Ma) Icelandic floras are not markedly North American. Rather, the Icelandic floras represent a typically northern hemispheric late Cainozoic type of vegetation with modern analogues in Europe and Asia as well as North America (Table 1). In principle, most of the taxa reported from Iceland do not indicate a particular migration route to Iceland. For example, the genera Sassafras and Magnolia that are found in the 12 Ma sedimentary formation in Iceland occurred in the whole northern hemisphere during the Cainozoic (Mai, 1995; Manchester, 1999) and display an East Asian–eastern North American disjunction at present. One taxon, endemic to North America at present, Comptonia, is not indicative of floral affinities of the Icelandic late Cainozoic vegetation either, because it was a widespread element in the European Cainozoic (Mai, 1995; Manchester, 1999). In contrast, a small number of plant taxa are strong indicators for one of the two possible migration routes (Table 1). Cryptomeria, which is restricted to East Asia at present, was widespread in the European Cainozoic, but has a highly ambiguous fossil record for North America (Aulenback & Lepage, 1998). Thus it points to a migration from Europe rather than from North America. A similar picture emerges when looking at modern taxa related to Alnus gaudinii (see Table 1). For the 10 Ma formation no clear pattern about migration routes emerges, because most of the fossils could not be identified below the genus or even subfamily level (Table 1), and modern higher taxa, to which the fossils can be assigned, show a Eurasian– North American distribution. For instance, Alnus cecropiifolia, which was a widespread element in the European late Cainozoic, appears to have taxonomic affinities to the North American modern A. rhombifolia and A. pringlei (Mexico, cf. Knobloch, 1969). In addition, a possible relation to the western Eurasian A. glutinosa ssp. barbata has been suggested by Kvadek et al. (2002). Relations to coeval North American species of Alnus are not quite clear. Also, Acer crenatifolium was widespread in the European Cainozoic, and clearly shows affinities to the modern North American A. rubrum. In addition, Walther (1972) found anatomical similarities to the modern A. hyrcanum. A clearer pattern emerges for the younger formations: Fagus gussonii is only known from the 9–8 Ma Mókollsdalur locality. This Late Miocene type of beech was confined to Europe and must have reached Iceland from Europe. The same may be true for glandular tip, basal side convex, much longer than apical side. Discussion: A distinct leaf type, which we are not able to assign to any living family. Material: Tröllatunga. Dicotylophyllum sp. 5 (‘Pyrus’) (Figs 213, 214) Description: Leaves, 5.5–6 cm long, 2.5 cm wide, ovate-elliptic, dentate, secondary venation brochidodromous, primary loops followed by secondary loops from which small veinlets insert the teeth, teeth spinose. Discussion: A distinct leaf type, which we are not able to assign to any living family. Dentition resembles some species of Pyrus L. within the Rosaceae. Material: Tröllatunga. ‘Ulmus’ diptera Steensstrup ‘Ulmus’ diptera Steensstrup ex Heer, 1859: 319 (Brjamsloek [Brjánslækur], Langavasdalr [Langavatnsdalur]). ‘Ulmus’ diptera Steensstrup ex Heer; Heer, 1868: 149, pl. 27, figs 1–3 (Brjamslaek [Brjánslækur], Laugavatsdalr [Langavatnsdalur], Hre~avatn [Hre~avatn]). Non Ulmus, affinities uncertain. Discussion: The three leaves figured in Heer (1868) do not belong to Ulmus. The leaf figured in pl. 27, fig. 1 superficially resembles Ulmus, but the numerous small teeth of almost equal size with a sharp apex are not found in elms. The leaf shown in pl. 27, fig. 2 may belong to Betulaceae. DISCUSSION TAXONOMIC AFFINITIES AND MIGRATION ROUTES Most previous ideas about the taxonomic identity of Miocene plants from Iceland go back to Oswald Heer’s classic work on Cainozoic floras (Heer, 1859, 1868). This involves misinterpretations of the generic identity of key taxa such as Cryptomeria anglica Boulter (this genus being confined to East Asia at present) that consistently has been mistaken for Sequoia sternbergii (genus endemic to North America at present) in the macrofossil record, and, as a consequence, also in the palynological record, where Sequoia and Cryptomeria pollen is not distinguishable in LM (cf. Áskelsson, 1946; Manum, 1962; Friedrich, 1966; Akhmetiev et al., 1978). In contrast to previous work (Áskelsson, 1957; Friedrich, 1966; Akhmetiev et al., 1978; Friedrich & Símonarson, © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 Table 1. Taxa reported for the Miocene of Iceland and their assumed modern analogues Hre~avatnStafholt Formation Extant area Modern analogue 7–6 Ma + A, Eb A, E BrjánslækurSeljá Formation ±12 Ma + 9–8 Ma ±10 Ma + + + + + + + TröllatungaGautshamar Formation Skar~sströndMókollsdalur Formation + + + A, E A, E A, E A, E boj_441.fm Page 411 Wednesday, November 23, 2005 3:37 PM Agea Equisetum sp. Osmunda parschlugiana Pteridophyta gen. 1 Dryopteris sp. Abies steenstrupiana Larix sp. Picea sect. Picea Equisetum O. regalis complex ? Dryopteridaceae? Abies Larix Picea sect. Picea Pinus sp. Pseudotsuga sp. Tsuga sp. Cryptomeria anglica Alnus sp. + + + + Leaves Organ Axis Fronds, leaflets Pinnae Pinnae Scales, seeds, leaves Spur shoots Cones, seeds, leaves, male inflorescences, in situ pollen Leaves Cones Leaves, cuticles Leafy axis, cuticles Leaves + + + A, E A, E A, E E A, E E Alnus cf. gaudinii Alnus cecropiifolia Female cones Leaves + + + + + + + + + + + + ?+ Bracts Leaves Bracts Leaves Leaves Leaves Leaves Leaves, cupules Leaflets Leaflets + + Leaves + + + + + + A, E A, E (A?), E A, E A, E E ?+ A, E A, E A, E A, E E ? A, E Alnus cf. kefersteinii (strobili) Betula cristata Betula subnivalis © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 Betula islandica Betula sectio Costatae sp. Corylus sp. Rhododendron aff. ponticum MIOCENE FLORAS FROM ICELAND ‘Vaccinium’ sp. (evergreen) ‘Arctostaphylos’ sp. (deciduous) Fagus gussonii Juglandaceae gen. 1 (Pterocarya/Cyclocarya) Juglandaceae gen. 2 (Carya??) Pinus Pseudotsuga Tsuga C. japonica A. incana, A. incana ssp. tenuifolia A. japonica, A. subcordata, A. nitida A. pringlei, A. rhombifolia, A. glutinosa Alnus (B. lenta?), B. maximowicziana B. lenta, B. papyrifera, B. potaninii B. sectio Costatae B. chinensis var. fargesii, B. delavayi Corylus R. ponticum; similar types in eastern N Am. Vaccinium (evergreen) Arctostaphylos (deciduous) F. sylvatica, F. longipetiolata Pterocarya, Cyclocarya Juglandaceae 411 a b The ages of the localities are based on K–Ar dates and palaeomagentic correlation (McDougall et al., 1984). A = North America, E = Eurasia. c † = Taxon is absent in the modern flora of the respective area. Bold type face indicates taxa that are indicative of a certain migration direction because they are restricted to either A or E since the Cainozoic. ‡No indication here for a certain migration direction because the genus had an A & E distribution during the Cainozoic. Table 1. Continued Hre~avatnStafholt Formation Extant area Modern analogue A†c, E BrjánslækurSeljá Formation Leaves, fruits Leaves Leaves Leaves Samaras Leaves Leaves + A, E A, E + + ?+ + + + + + + A, E A, E + Leaves Leaves Leaves Leaves Leaves Female catkin Leaves Leaves, samaras + + + + ?+ + + + + + + + + + + + + + + Leaves, samaras Leaves Leaves Leaves, rhizomes Leaves Leaves Leaves Leaves Leaves Leaves Leaves + + + + + + + + ? A, E A, E A, E A, E†‡ A, E + + P. fraxinifolia, P. macrocarpa Laurus nobilis Sassafras Magnolia Liriodendron C. peregrina Prunoideae Prunoideae Rosaceae TröllatungaGautshamar Formation Skar~sströndMókollsdalur Formation boj_441.fm Page 412 Wednesday, November 23, 2005 3:37 PM 412 T. DENK ET AL. Pterocarya sp. (European– East Asian type) Laurophyllum sp. 1 (‘Laurus’) Sassafras ferrettianum Magnolia sp. Liriodendron sp. Comptonia hesperia Rosaceae gen. et spec. indet. type A (?Prunoideae) Rosaceae gen. et spec. indet. type B (?Prunoideae) Rosaceae gen. et spec. indet. type C cf. Crataegus sp. Populus ex group P. tremula L. Populus sp. 2 Populus sp. 3 Salix gruberi Acer crenatifolium ssp. islandicum Acer askelsonii A, E A, E A, E A, E A, E A, E A, E A, E ? ? ? ? ?A, E Crataegus P. tremula, P. tremuloides ? ? S. scouleriana, S. caprea A. sect. Rubrum A. saccharum, A. platanoides Smilax Ulmus Phragmites Cyperaceae Nuphar Brassicaceae, Lamiaceae? Lonicera Neolitsea Pyrus Smilax sp. Ulmus pyramidalis Phragmites sp. Cyperaceae cf. Nuphar sp. Dicotylophyllum sp. 1 (herbaceous) Dicotylophyllum sp. 2 (‘Lonicera’) Dicotylophyllum sp. 3 (‘Neolitsea’) Dicotylophyllum sp. 4 Dicotylophyllum sp. 5 (‘Pyrus’) a © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 b The ages of the localities are based on K–Ar dates and palaeomagentic correlation (McDougall et al., 1984). A = North America, E = Eurasia. c † = Taxon is absent in the modern flora of the respective area. Bold type face indicates taxa that are indicative of a certain migration direction because they are restricted to either A or E since the Cainozoic. ‡No indication here for a certain migration direction because the genus had an A & E distribution during the Cainozoic. boj_441.fm Page 413 Wednesday, November 23, 2005 3:37 PM MIOCENE FLORAS FROM ICELAND Pterocarya (European–East Asian type), which has no fossil record in North America in the Late Miocene (Manchester, 1999), although most recently winged Pterocarya fruits have been reported from the Pliocene–Pleistocene of North America (Stults & Axsmith, 2004). Besides, a number of taxa that become dominant in the youngest formation studied (7–6 Ma) appear already in the 9–8 Ma Skar~sströndMókollsdalur Formation (Betula cristata) and the 10 Ma Tröllatunga-Gautshamar Formation (Acer askelssonii) but play a minor role in these older formations. Of these, B. cristata clearly shows affinities to the modern East Asian B. maximowizciana and the European late Cainozoic species B. pseudolumnifera, and therefore might have reached Iceland from Europe. This points to a migration from Europe for the 9–8 Ma plant assemblages, and to a shift in dominance from the 9–8 Ma Skar~sströnd-Mókollsdalur Formation (Fagus-Pterocarya) to the 7–6 Ma Hre~avatn-Stafholt Formation (Acer askelssonii, Alnus sp. aff. A. incana, Betula cristata, Salix gruberi, several conifers). Some taxa, which appear for the first time in the Hre~avatn-Stafholt Formation, did not necessarily migrate from Europe. Alnus sp. that cooccurs with relatively large strobili of A. cf. kefersteinii in the Hre~avatn-Stafholt Formation shows affinities to both North American and European modern species. In the case of Pseudotsuga it is completely unclear how and when it migrated to Iceland. The oldest fossil record of the genus is from the Early Oligocene of western North America; by the Early/Mid Miocene it appears to have reached Japan via Beringia (Schorn & Erwin, 2000). So far, all unambiguous fossils of Pseudotsuga occurred within the modern range of the genus (western North America, Japan, Taiwan; Schorn & Thompson, 1998). The Icelandic occurrence therefore points to a much wider range of this genus in the late Cainozoic. 413 Asia. These two taxa are absent in all localities belonging to the Hre~avatn-Stafholt Formation (7– 6 Ma), which are characterized by Betula cristata, Alnus sp. aff. A. incana, Acer askelsonii and several conifers. This points to an asymmetric pattern of plant migration to Iceland. While Europe and North America are suggested to have been source areas for the older (≥ 12 Ma) and probably also the 10 Ma formations (cf. Table 1), the Mókollsdalur plant assemblage (9–8 Ma) shows colonization mainly from Europe. An exception appears to be the North American hickory aphid, Longistigma caryae, from Mókollsdalur that has no fossil record except for the Icelandic one (Heie & Friedrich, 1971) and is confined to North America at present. Taxa typical of the youngest formation studied (7–6 Ma) either persisted from the older floras (e.g. B. cristata) or are not indicative of a certain migration direction (Alnus sp. aff. A. incana). A remarkable exception is Pseudotsuga, which has a fossil record outside Iceland that is centred around the Pacific area (see above). PALAEOCLIMATIC CONSIDERATIONS PATTERNS OF MIGRATION Plant fossil data indicate that thermophilic taxa such as Comptonia, Sassafras and Magnolia, which were typical of the older Brjánslækur-Seljá Formation (12 Ma) and partly of the 15 Ma Selárdalur-Botn Formation (F. Grímsson, T. Dent & L.A. Símonarson, unpubl. data), went extinct between 12 and 10 Ma. They were replaced by another set of less thermophilic taxa such as Juglandaceae aff. Pterocarya/Cyclocarya and Rhododendron ponticum type in the TröllatungaGautshamar Formation (10 Ma), or persisted (Smilax, Alnus cecropiifolia, Acer crenatifolium). The 9–8 Ma sediments of Mókollsdalur are devoid of distinctly thermophilic taxa but bear Fagus gussonii and Pterocarya (European–East Asian type), two taxa with affinities to the eastern Black Sea region and/or East Heer (1868) appears to have been the first to use a quantitative approach to estimate the palaeoclimate for the late Mid Miocene of Iceland. Using a key taxon, Liriodendron procaccinii, which he compared with the modern L. tulipifera from eastern North America, he concluded that the mean annual temperature (MAT) for Iceland was 11.5 ∞C during Miocene times. The modern climate of Iceland is a Cfc climate (mild with no dry season, cool summer) in the southern and western coastal areas, and an ET climate (polar tundra, no true summer) in the remaining parts according to Köppen’s classification (Köppen & Geiger, 1928; Stern, de Hoedt & Ernst, 2000). Our data suggest that a Cfa climate (humid warm temperate with no dry season, hot summers) persisted in Iceland at least until 10 Ma, and most probably until 9–8 Ma. The absence of key taxa such as Fagus gussonii and Pterocarya (European–East Asian type) in the Hre~avatn-Stafholt Formation may indicate the change from a warmer Cfa to a cooler Cfb (cooler summer than Cfa) climate. No dramatic change in climate is suggested by plant fossil data over the period 12 Ma to 7–6 Ma, although a hiatus appears to occur between 9–8 and 7–6 Ma (Cfa to Cfb). This scenario is in stark conflict with previously predicted palaeotemperatures for Eastern Iceland (Mudie & Helgason, 1983; Duncan & Helgason, 1998), suggesting a cooling of 10 ∞C from 10 to 9.5 Ma. This assumption was based on a regional palynoflora with thermophilic elements in the older sediments and cold elements such as Betula nana and Alnus viridis type pollen in the younger sediments of © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 boj_441.fm Page 414 Wednesday, November 23, 2005 3:37 PM 414 T. DENK ET AL. Further support for this scenario comes from studies from deep-sea sediments. Based on silicoflagellates and dinocyst assemblages recovered from ODP holes 642 and 643 in the Vøring Plateau east of Iceland, early high-latitude climate cooling has been suggested between 8.5 and 7.4 Ma, and 9 and 6.6 Ma, respectively (Ciesilski & Case, 1989; Mudie, 1989). Moreover, an earlier onset of glaciations in the western North Atlantic has been shown based on a comparison of ODP sites in Baffin Bay and the Norwegian Sea (Mudie, de Vernal & Head, 1990). Marine geological evidence shows that first pulses of ice-rafted debris in the northern hemisphere date back to 14–10 Ma, with earlier glaciations in Greenland than in the eastern North Atlantic (Thiede & Myhre, 1995; Thiede et al., 1998). These early glaciations may not have had immediate influence on the vegetation, but their diachrony may explain changing migration patterns to Iceland during the Late Cainozoic. While larger glaciers in southern Greenland may have prevented migrations to Iceland as early as 10 Ma, migration from the east may still have been feasible. A partly emerged Iceland–Faeroe ridge until 10 Ma (Thiede & Eldholm, 1983; Eldholm et al., 1994) may have provided the physical link for plants to migrate to Iceland. Migration of plants from Europe to Iceland at 9–8 Ma may have been enabled by island hopping, because relevant taxa (Fagus, Pterocarya) have no potential for wind or bird dispersal. Plant taxa of the later (Pliocene) sediments are wind and bird dispersed (Betulaceae, Ericaceae, Salicaceae; Akhmetiev et al., 1978). the profile studied (Mudie & Helgason, 1983). Cold elements occurred also in the older parts of the profile. These successive vegetation types were compared with different forest types in North America ranging from ‘North Carolina deciduous forests’ (MAT 13–15 ∞C) to ‘Canadian Atlantic region’ (MAT 3–5 ∞C, Mudie & Helgason, 1983). In fact, no clear distinction could be made between a post-fire vegetation and a true cold boreal vegeation in the younger sediments. In addition, these studies used very small sample sizes along the vertical profile and found much less pollen in the younger sediments (one of the samples contained only 16 palynomorphs in total; Mudie & Helgason, 1983). A similar phenomenon can be observed in the macrofossil record in different localities belonging to sediments of the same age in northwestern Iceland. Here the Brjánslækur-Seljá Formation is very rich at the type locality Brjánslækur (Cryptomeria, Comptonia, Magnolia, Liriodendron, Smilax, among many others), but much poorer at the close-by Seljá locality (Populus, Salix-dominated), both of which obviously represent different types of vegetation and/or different phases of succession towards a climax community. The same is true for the Tröllatunga and the Húsavík localities (10 Ma), where the first is rich in thermophilic hardwood species, while the latter is dominated by Salix and ferns. A drop in temperature as suggested by Mudie & Helgason (1983) would mean that the climate abruptly deteriorated at around 10 Ma, which would have led to extinction of thermophilic taxa. Subsequently, the climate would have become warmer again and warm temperate taxa would have colonized Iceland before 9–8 Ma. Although this would be in accordance with the pattern of repeated extinction followed by colonization detected by us, the drop of 10 ∞C seems too far fetched and would need support from other sequences, mainly the numerous ones from north-western Iceland (cf. Akhmetiev et al., 1978). ACKNOWLEDGEMENTS We should like to thank Guri Bugge, Gerwin Gruber and Thomas Mörs for help with collecting material. The Swedish Research Council is gratefully acknowledged for funding a field trip to Iceland in the summer of 2003 and the Swedish Polar Research Secretariat for kindly providing tents for the fieldwork. The Icelandic Research Fund for Graduate Students is thanked for providing a scholarship to Fri~geir Grímsson. Zlatko Kvadek received funding from the grant project GABR no. 205/04/0099. Finally, we should like to express our gratitude to Dieter Mai and Steve Manchester for their helpful reviews of the manuscript. EVIDENCE FROM OTHER DISCIPLINES In the present study we have outlined patterns of migration and extinction of plants during different periods of the Neogene and established source areas for plant taxa reaching Iceland at different times. We observed a series of colonizations and extinctions, first of thermophilic taxa and later of warm to cool temperate taxa. In contrast to previous work we showed that main source areas for immigrants to Iceland during the late Mid and earliest Late Miocene were both Europe and North America, while later immigration appears to have happened from Europe. Our results suggest that climate cooling was gradual rather than abrupt for the time interval 12 Ma to 7–6 Ma. REFERENCES Akhmetiev MA, Bratzeva GM, Giterman RE, Golubeva LV, Moiseyeva AI. 1978. Late Cainozoic stratigraphy and flora of Iceland. Moscow: Academy of Sciences of the USSR. Published in English by the National Research Council, Reykjavik, Iceland, September 1981. © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 boj_441.fm Page 415 Wednesday, November 23, 2005 3:37 PM MIOCENE FLORAS FROM ICELAND Aronson JL, Sæmundsson K. 1975. Relatively old basalts from structurally high areas in central Iceland. Earth and Planetary Science Letters 28: 83–97. Áskelsson J. 1946. Um gró~urmenjar í Pórishlí~arfjalli vi~ Selárdal. Andvari 71: 80–86. Áskelsson J. 1957. Myndir úr jar~fræ~i Íslands VI. Náttúrufrœ~ ingurinn 27: 22–29. Aulenback KR, Lepage BA. 1998. Taxodium wallisii sp. nov. – first occurrence of Taxodium from the Upper Cretaceous. International Journal of Plant Sciences 159: 367–390. Baranova M. 1972. Systematic anatomy of the leaf epidermis in the Magnoliaceae and some related families. Taxon 21: 447–469. Barrón E, Diéguez C. 1994. Neogene species of the genus Fagus L. from La Cerdaña (Lerida, Spain). Taxonomic conclusions and phylogenetic considerations. Anales del Jardin Botanico de Madrid 52: 21–32. Boulter MC. 1970. Cryptomeria – a significant component of the European Tertiary. Paläontologische Abhandlungen B III: 279–287. d Boulter MC, Kvadek Z. 1989. The Palaeocene flora of the Isle of Mull. Special Papers in Palaeontology 42: 1–149. Bronn HG. 1838. Lethaea Geognostica, oder Abbildungen und Beschreibungen der für die Gebirgs-Formationen bezeichnendsten Versteinerungen. Zweiter Band, das Kreide- und Molassen-Gebirge enthaltend. Stuttgart: E. Schweizerbart’sVerlagshandlung. Chaloner BW. 1999. Plant and spore compression in sediments. In: Jones TP, Rowe NP, eds. Fossil plants and spores: modern techniques. London: Geological Society, 36–40. Ciesilski PF, Case SM. 1989. Neogene paleoceanography of the Norwegian Sea based upon silicoflagellate assemblage changes in ODP Leg 104 sedimentary sequences. Proceedings of the Ocean Drilling Program, Scientific Results 104: 527–541. Denk T, Meller B. 2001. The systematic significance of the cupule/nut complex in living and fossil Fagus. International Journal of Plant Sciences 162: 869–897. Dilcher DL. 1974. Approaches to the identification of angiosperm leaf remains. Botanical Review 40: 1–157. Dorofeev P, Iljinskaja I. 1994. Fossil flowering plants of Russia and adjacent states, Vol. 3. Leitneriaceae-Juglandaceae. Komarov Botanical Institute, Saint Petersburg: Russian Academy of Sciences. Duncan RA, Helgason J. 1998. Precise dating of the Holmatindur cooling event in eastern Iceland: evidence for midMiocene bipolar glaciation. Journal of Geophysical Research 103, B6: 12397–12404. Eldholm O, Myhre AM, Thiede J. 1994. Cainozoic tectonomagmatic events in the North Atlantic: potential palaeoenvironmental implications. In: Boulter MC, Fisher HC, eds. Cainozoic plants and climates of the Arctic. NATO ASI Series, Vol. 127. Berlin: Springer, 35–55. Ferguson DK. 1971. The Miocene flora of Kreuzau, western Germany I. The leaf remains. Verhandelingen der Koninklijke Nederlandse Akademie van Wetenschappen, Afdelning Natuurkunde 60: 1–297. Amsterdam: London North-Holland Publishing Company. 415 Friedrich WL. 1966. Zur Geologie von Brjánslaekur (Nordwest-Island) unter besonderen Berücksichtigung der fossilen Flora. Sonderveröffentlichungen des Geologischen Institutes der Universität Köln 10, 1–110, Bonn: Wilhelm Stollfuss Verlag. Friedrich WL, Símonarson LA. 1976. Acer askelssonii n. sp., grosse Neogene Teilfrüchte aus Island. Palaeontographica B 155: 151–166. Friedrich WL, Símonarson LA. 1981. Die fossile Flora Islands: Zeugin der Thule- Landbrücke. Spektrum der Wissenschaft 10/1981: 22–31. Friedrich WL, Símonarson LA. 1982. Acer-Funde aus dem Neogen von Island und ihre stratigraphische Stellung. Palaeontographica B 182: 151–166. Friedrich WL, Símonarson LA. 1983. Fossile planter fra Island. Naturens Verden 9: 302–313. Friedrich WL, Símonarson LA. 2002. The Tertiary flora of Iceland – a witness of a migration route between Europe and North America. 6th European Palaeobotany–Palynology Conference. Athens August 29–September 2: 183–184. Athens: University of Athens. Friedrich WL, Símonarson LA, Heie OE. 1972. Steingervingar í millilögum í Mókollsdal. Náttúrufræ~ ingurinn 42: 4– 17. Goeppert HR. 1855. Die tertiäre Flora von Schossnitz in Schlesien. Görlitz: Heyn’sche Buchhandlung (E. Remer). Grímsson F, Denk T. 2005. Fagus from the Miocene of Iceland: systematics and biogeographical considerations. Review of Palaeobotany and Palynology 134: 27–54. Hantke R. 1954. Die fossile Flora der obermiozänen Oehninger-Fundstelle Schrotzburg. Denkschriften der Schweitzerischen Naturforschenden Gesellschaft 80: 27–118. Hardarson BS, Fitton JG, Ellam RM, Pringle MS. 1997. Rift relocation – a geochemical and geochronological investigation of a palaeo-rift in northwest Iceland. Earth and Planetary Science Letters 153: 181–196. Heer O. 1856. Flora Tertiaria Helvetica – Die tertiäre Flora der Schweiz, 2nd volume. Winterthur. Heer O. 1859. Flora Tertiaria Helvetica – Die tertiäre Flora der Schweiz, 3rd volume. Winterthur. Heer O. 1865. Die Urwelt der Schweiz. Zürich. Heer O. 1868. Flora fossilis arctica – Die fossile Flora der Polarländer enthaltend die in Nordgrönland, auf der Melville-Insel, im Banksland, am Mackenzie, in Island und in Spitzbergen entdeckten fossilen pflanzen. Zürich: Friedrich Schulthess. Heer O. 1868–1883. Flora fossilis arctica. 8 volumes. Zürich: Friedrich Schulthess. Heie OE, Friedrich WL. 1971. A fossil specimen of the North American Hickory Aphid (Longistigma caryae Harris) found in Tertiary deposits in Iceland. Entomologica Scandinavica 2: 74–80. Jóhannesson H. 1980. Jar~lagaskipan og próun rekbelta á Vesturlandi. Náttúrufrœ~ ingurinn 50: 13–31. Jóhannesson H, Sæmundsson K. 1989. Geological map of Iceland 1:500 000. Bedrock Geology, 1st edn. Reykjavík: Icelandic Museum of Natural History and Icelandic Geodetic Survey. © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 boj_441.fm Page 416 Wednesday, November 23, 2005 3:37 PM 416 T. DENK ET AL. Massalongo A, Scarabelli G. 1859. Studii sulla flora fossile e geologia stratigrafica del Senigalliese. Imola, Italy. McDougall I, Kristjansson L, Saemundsson K. 1984. Magnetostratigraphy and geochronology of Northwest Iceland. Journal of Geophysical Research 89: 7029–7060. Moorbath S, Sigurdsson H, Goodwin R. 1968. K–Ar ages of the oldest exposed rocks in Iceland. Earth and Planetary Science Letters 4: 197–205. Mudie PJ. 1989. Palynology and dynocyst biostratigraphy of the Late Miocene to Pleistocene, Norwegian Sea: ODP Leg 104, sites 642–644. Proceedings of the Ocean Drilling Program, Scientific Results 104: 587–610. Mudie PJ, de Vernal A, Head MJ. 1990. Neogene to recent palynostratigraphy of circum-arctic basins: results of ODP Leg 104, Norwegian Sea, Leg 105, Baffin Bay, and DSDP Site 611, Irminger Sea. In: Bleil U, Thiede J, eds. Geological history of the polar oceans: Arctic versus Antarctic. The Netherlands: Kluwer Academic Publishers, 609–646. Mudie PJ, Helgason J. 1983. Palynological evidence for Miocene climate cooling in eastern Iceland about 9.8 Myr ago. Nature 303: 689–692. Procházka J, B u |ek B. 1975. Maple leaves from the Tertiary ˚ of North Bohemia. Rozpravy Ústøedního Ústavu Geologického 41: 7–86. Sæmundsson K. 1974. Evolution of the Axial Rifting Zone in Northern Iceland and the Tjörnes Fracture Zone. Geological Society of American Bulletin 85: 495–504. Sæmundsson K. 1979. Outline of the geology of Iceland. Jökull 29: 7–28. Schorn HE, Erwin DM. 2000. The impression record history and ecological diversification of Pseudotsuga Carriere (Pinaceae) in western North America during the later half of the Cenozoic. Abstract F-4. Botany 2000, Oregon Convention Center 6–10 August, 2000 Portland, OR. Schorn HE, Gooch NL. 1994. Amelanchier hawkinsae sp. nov. (Rosaceae, Maloideae) from the Middle Miocene of Stewart Valley, Nevada, and a review of the genus in the Nevada Neogene. Paleobios 16: 1–17. Schorn HE, Thompson A. 1998. The genus Pseudotsuga: a revision of the fossil record and inferred paleo-biogeographical and migrational patterns. Integrative Paleontology and the Future, February 28, 1998. University of California, Berkeley: Museum of Paleontology, Abstract 27. Símonarson LA. 1990. Fyrstu landspendyraleifarnar úr íslenskum tertíerlögum. Náttúrufrœ~ ingurinn 59: 189–195. Símonarson LA. 1991. Hikkoría frá Tröllatunga. Náttúrufrœ~ ingurinn 60: 144. Símonarson LA, Akhmetiev M, Leifsdóttir ÓE, Friedrich WL, Eiríksson J. 2000. Rannsóknir á tertíerflóru í Dyrafir~i. Ágrip erinda og veggspjalda. Vorrá~stefna. Jar~frae~ifélag Íslands, Reykjavík, April 13: 31–33. Geoscience Society of Iceland: Reykjavík. Símonarson LA, Akhmetiev M, Leifsdóttir ÓE, Friedrich WL, Eiríksson J. 2002. The Upper Miocene DufansdalurKetilseyri plant bearing horizon in Northwest Iceland. Abstract Volume. The 25th Nordic Geological Winter Meeting. January 6th–9th, Reykjavík, Iceland: 189. Geoscience Society of Iceland: Reykjavík. Kilpper K. 1968. Koniferen aus den tertiären Deckschichten des niederrheinischen Hauptflözes, 3. Taxodiaceae und Cupressaceae. Palaeontographica B 124: 102–111. Knobloch E. 1969. Tertiäre Floren von Mähren. Brno: Moravské Museum Brno. Knobloch E. 1998. Der pliozäne Laubwald von Willershausen am Harz. Documenta Naturae 120: 1–302. d Knobloch E, Kvadek Z. 1976. Miozäne Blätterfloren vom Westrand der Böhmischen Masse. Rozpravy Ústøedního Ústavu Geologického 42: 1–131. Köppen W, Geiger R. 1928. Klimakarte der Erde, Wall-map 150 cm x 200 cm. Verlag Justus Perthes, Gotha. d Kovar-Eder J, Kvadek Z, Ströbitzer-Hermann M. 2004. The Miocene Flora of Parschlug (Styria, Austria) – Revision and Synthesis. Annalen des Naturhistorischen Museums Wien A 104: 45–161. Kristjansson L, Hardarson BS, Audunsson H. 2003. A detailed palaeomagnetic study of the oldest (approximate to 15 Myr) lava sequences in Northwest Iceland. Geophysical Journal of International 155: 991–1005. d Kvadek Z. 2002. Novelties on Doliostrobus (Doliostrobaceae), an extinct conifer genus of the European Palaeogene. Basopis Národního Muzea, rada P rírodovedná 171: 131–175. ˇ ˇ d Kvadek Z. 2004. Cuticular studies in the Early Oligocene flora of Flörsheim (Mainz Basin, Germany). Senckenbergiana Lethaea 84: 1–73. d Kvadek Z, Hurník S. 2000. Revision of Early Miocene plants preserved in baked rocks in the North Bohemian Tertiary. Acta Musei Nationalia Pragae B 56: 1–48. d Kvadek Z, Manum SB, Boulter MC. 1994. Angiosperms from the Palaeogene of Spitsbergen, including an unfinished work by A. G. Nathorst. Palaeontographica B 232: 103–128. d Kvadek Z, Velitzelos D, Velitzelos E. 2002. Late Miocene flora of Vegora, Macedonia, N Greece. Korali Publications, Athens. LePage BA. 2001. New species of Picea A. Dietrich (Pinaceae) from the middle Eocene of Axel Heiberg Island, Arctic Canada. Botanical Journal of the Linnean Society 135: 137–167. LePage BA, Basinger JF. 1991. Early Tertiary Larix from the Buchanan Lake Formation, Canadian Arctic Archipelago, and a consideration of the phytogeography of the genus. Geological Survey of Canada, Bulletin 403: 67–82. Lindquist B. 1947. Two species of Betula from the Iceland Miocene. Svensk Botanisk Tidskrift 41: 339–353. Liu Y-S. 1996. Foliar architecture of Betulaceae and a revision of the Chinese betulaceous megafossils. Palaeontographica B 239: 23–57. Mädler K. 1939. Die pliozäne Flora von Frankfurt am Main. Abhandlungen der Senckenbergischen Naturforschenden Gesellschaft 446: 1–202. Mai DH. 1995. Tertiäre Vegetationsgeschichte Europas. Jena: Gustav Fischer. Manchester SR. 1999. Biogeographical relationships of North American Tertiary floras. Annals of the Missouri Botanical Garden 86: 472–522. Manum S. 1962. Studies in the Tertiary flora of Spitsbergen, with notes on Tertiary floras of Ellesmere Island, Greenland, and Iceland. A palynological investigation. Norsk Polarinstitutt Skrifter 125: 1–127. © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 boj_441.fm Page 417 Wednesday, November 23, 2005 3:37 PM MIOCENE FLORAS FROM ICELAND Spicer RA. 1986. Pectinal veins: a new concept in terminology for the description of dicotyledonous leaf venation patterns. Botanical Journal of the Linnean Society 93: 379–388. Steinthórsson S, Thorarinsson S. 1997. Iceland. In: Moores EM, Fairbridge RW, eds. Encyclopedia of European and Asian Regional Geology. London, New York: Chapman, 341– 352. Stern H, de Hoedt G, Ernst J. 2000. Objective classification of Australian climates. Australian Meteorological Magazine 49: 87–96. Stults D, Axsmith BJ. 2004. A preliminary reinvestigation of the Citronella Formation flora of southern Alabama: age and paleoclimatic implications. 7th International Organization of Palaeobotany Conference. Bariloche, Argentina March 21– 26, 2004: 115–116. Bahia Blanca, Argentina: Arsa Gráfica. Sveshnikova IN. 1984. A new genus of the family Taxodiaceae from the Middle Miocene of Iceland. Yearbook of the All-Union Palaeontological Society 1: 263–269 (in Russian). Thiede J, Eldholm O. 1983. Speculations about the paleodepth of the Greenland-Scotland Ridge during late Mesozoic and Cenozoic times. In: Bott MHP, Saxow S, Talwani M, Thiede J, eds. Structure and development of the Greenland– Scotland Ridge: new methods and concepts. Plenum, New York, 445–456. Thiede J, Myhre AM. 1995. Non-steady behaviour in the Cenozoic polar north Atlantic system – the onset and variability of northern hemisphere glaciations. Philosophical Transactions of the Royal Society of London Series A – Physical Sciences and Engineering 352 (1699): 373–385. Thiede J, Winkler A, Wolfwelling T, Eldholm O, Myhre AM, Baumann KH, Henrich R, Stein R. 1998. Late Cen- 417 ozoic history of the polar North Atlantic – results from ocean drilling. Quaternary Science Reviews 17: 185–208. Tian X, Guo Z-H, Li D-Z. 2002. Phylogeny of Aceraceae based on ITS and trnL-F data sets. Acta Botanica Sinica 44: 714– 724. Tiffney BH, Manchester SR. 2001. The use of geological and paleontological evidence in evaluating plant phylogeographic hypotheses in the northern hemisphere Tertiary. International Journal of Plant Sciences 162: S3–S17. y Velenovsky J. 1881. Die Flora aus den ausgebrannten tertiären Letten von Vrtovic bei Laun. Abhandlungen der Königlichen Böhmischen Gesellschaft der Wissenschaften, Mathematisch-Naturwissenschaftliche Classe VI (11): 1–56. Walther H. 1972. Studien über tertiäre Acer Mitteleuropas. Abhandlungen des Staatlichen Museums für Mineralogie und Geologie zu Dresden 19: 1–309. Windisch P. 1886. Beiträge Zur Kenntnis der Tertiärflora Von Island. Inaugural-Dissertation behufs Erlangung der philosophischen Doctorwürde der Hohen philosophischen Facultät der Universität Leipzig. Gebauer-Schwetschke’sche Buchdruckerei, Halle a. S. Wolfe JA. 1966. Tertiary plants from the Cook Inlet Region, Alaska. Geological Survey Professional Paper 398-B: 1–32. Zastawniak E, Walther H. 1998. Betulaceae from Sornica near Wroclaw (Poland) – a revision of Goeppert’s original materials and study of more recent collections. Acta Palaeobotanica 38: 87–145. Zhilin SG. 1980. Notes on systematics of fossil plants. Myricaceae. In: Systematics and evolution of higher plants. Leningrad: Nauka, 9–20 (in Russian). © 2005 The Linnean Society of London, Botanical Journal of the Linnean Society, 2005, 149, 369–417 boj_441.fm Page 418 Wednesday, November 23, 2005 3:37 PM