See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/222297269 Phytolith types and type-frequencies in subalpine-alpine plant species of the European Alps Article in Review of Palaeobotany and Palynology · January 2004 DOI: 10.1016/j.revpalbo.2003.11.002 · Source: OAI CITATIONS READS 66 93 3 authors, including: Jean-Paul Theurillat Marco Madella 89 PUBLICATIONS 3,195 CITATIONS 171 PUBLICATIONS 1,673 CITATIONS University of Geneva SEE PROFILE University Pompeu Fabra & IMF-CSIC SEE PROFILE Some of the authors of this publication are also working on these related projects: North Gujarat Archaeological Project (NoGAP) View project All content following this page was uploaded by Marco Madella on 29 November 2016. The user has requested enhancement of the downloaded file. All in-text references underlined in blue are linked to publications on ResearchGate, letting you access and read them immediately. Review of Palaeobotany and Palynology 129 (2004) 39 – 65 www.elsevier.com/locate/revpalbo Phytolith types and type-frequencies in subalpine–alpine plant species of the European Alps Adriana L. Carnelli a,*, Jean-Paul Theurillat b, M. Madella c a Institut of Plant Science, University of Bern, Altenbergrain 21, CH-3013 Bern, Switzerland Centre Alpien de Phytogéographie, Fondation J.-M. Aubert, CH-1938 Champex, Switzerland and Laboratoire de Biogéographie, Département de Botanique et de Biologie végétale, University of Geneva, 1 ch. de l’Impératrice, CH-1292 Chambésy, Switzerland c The Cambridge Phytolith Project, The McDonald Institute for Archeological Research, University of Cambridge, Downing Street, Cambridge CB2 3ER, UK b Received 28 May 2002; accepted 11 November 2003 Abstract Biogenic silica extracted from 21 species commonly occurring in subalpine and alpine plant communities in the central Swiss Alps were examined using light and SEM microscopes; 19 species being screened here for the first time. An inventory of phytolith types is provided and type-frequencies are assessed. Light microscope photographs and SEM micrographs provided illustrations of the types described. The monocotyledons analysed belonged to the genera Calamagrostis, Festuca, Nardus, Poa and Carex. Monocotyledons yielded mainly types of epidermal origin (short cells, rods, cork cells, silicified stomata and trichomes). Dicotyledons analysed were from five genera of the Ericaceae family (Arctostaphylos, Calluna, Loiseleuria, Rhododendron, Vaccinium) and from one genus of the Betulaceae (Alnus). In dicotyledons, silicified epidermal jigsaw cells, stomata complexes and vessels were recovered. Conifer species of the genera Abies, Juniperus, Picea, Pinus and Larix were studied. Distinctive conifer cells were mainly silicified endodermids and transfusion tissues. Phytolith taxonomic diagnostic potential was tested by cluster analysis and principal component analysis (PCA). It was shown that grass and sedge species could be easily differentiated on the basis of phytolith types. In general, Ericaceae and conifers could also be distinguished on the basis of phytoliths; however, some species yielding mainly redundant types were not unequivocally identifiable. The aim of the present work is to provide a framework for phytolith-based plaeoecological studies at the treeline in the European Alps. D 2004 Elsevier B.V. All rights reserved. Keywords: biogenic silica; conifers; Cyperaceae; Ericaceae; Gramineae; phytolith 1. Introduction * Corresponding author. Tel.: +41-31-6314923. E-mail address: adriana.carnelli@ips.unibe.ch (A.L. Carnelli). 0034-6667/$ - see front matter D 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.revpalbo.2003.11.002 Opal-A is deposited in the tissues of many vascular plants. These deposits can be cast from the cells, from the infilling of the cell lumen or by 40 A.L. Carnelli et al. / Review of Palaeobotany and Palynology 129 (2004) 39–65 silicification of the cell walls (their shape resembling the outline of the original cell). The silicification can also be found as intercellular deposits in the interstitial spaces (Carnelli et al., 2001). The term phytoliths is commonly employed to describe silicified cells, either isolated or in tissues. The initial interest in phytolith type was focused on their diagnostic potential in plant taxonomy and phytoliths were observed in situ (Metcalfe, 1960). More recently, attention has been given to their applications in archeological and palaeoecological studies. Phytoliths released in soils, in sediments or in archeological sites, can be preserved and, eventually, analysed for palaeoecological reconstructions. The morphology of phytoliths occurring in several crop species has been more extensively described because of its interest for archeological studies (Blackman, 1969; Whang et al., 1998; Rapp and Mulholland, 1992; Ball et al., 1999). In general, grasses were the plants usually chosen for study (Twiss et al., 1969; Blackman, 1971; Palmer, 1976), and only a few non-grass species naturally occurring in a temperate climate have been investigated. Extensive works for American and Japanese species are available (Geis, 1973; Wilding and Drees, 1973; Klein and Geis, 1978; Kondo and Peason, 1981; Bozarth, 1993). For a more comprehensive list of phytolith studies in non-grass plants, see also Mulholland et al. (1992). Any attempt to interpret soil-borne phytolith assemblages should be based on a documented inventory relevant for the study area (Bozarth, 1993; Piperno, 1988). Data on species of the European Alps are not available, and we are aware of only one study of a mountain ecosystem, in the Caucasus (Blinnikov, 1994). This paper is the third of a series describing the scanning of a set of 21 species occurring in the subalpine – alpine belts of the European Alps. In this paper, the morphology of phytoliths is described. The content of biogenic silica has been already assessed from the same plant material (Carnelli et al., 2001) and the diagnostic potential of aluminium presence in biogenic opal has been tested (Carnelli et al., 2002). The aim of the present work is to supply an inventory of phytoliths of dominant alpine species on siliceous bedrock to provide a framework for phytolith-based palaeoecological studies at the treeline in the Euro- pean Alps. Therefore, light microscopy and scanning electron micrographs are supplied for phytolith identification. In addition, we used cluster analysis and principal components analysis (PCA) to test the hypothesis that taxonomic groups can be distinguished on the basis of phytolith types. The physiological role of silicon in vascular plants is still an open subject and ecological conditions may play a strong role in determining the degree of silicification of plant individuals and populations, therefore the present data should be validated with a larger data set of plant populations from different ecological settings. 2. Methods 2.1. Laboratory procedures and samples Phytoliths present in the above-ground tissues of 21 species commonly occurring on siliceous bedrocks in the subalpine and alpine vegetation belts in the Central Alps were analysed (Table 1). Plant material was collected in Val d’Arpette and at Furka Pass (Central Alps, Valais region, CH). A composite sample was collected from the aboveground tissues of several randomly selected plants from each species, from sites with rather homogeneous environmental conditions. Site locations, sampling and extraction procedures have been described elsewhere (Carnelli et al., 2001). The leaves and branches of woody species were analysed separately. Among the plants previously analysed for biogenic silica (Carnelli et al., 2001), species with very low biogenic silica content and rare identifiable phytoliths were not taken into account [i.e., leaf material from Alchemilla pentaphylla L. (Rosaceae), Geum montanum L. (Rosaceae), Leontodon helveticus Merat (Asteraceae), Salix herbacea Vill. (Salicaceae), Veronica bellidioides L. (Scrophulariaceae), and leaf and wood from Empetrum nigrum subsp. hermaphroditum (Hagerup) Böcher (Ericaceae)]. Silica bodies were observed and identified using a petrographic microscope equipped with phase contrast optic and polarised illumination at a magnification of  504 and  800 (oil immersion). Permanent microscope slides were mounted in A.L. Carnelli et al. / Review of Palaeobotany and Palynology 129 (2004) 39–65 Table 1 List of species and plant tissues from which biogenic silica was extracted (L = leaves, C = culms, N = needles, W = woody branches). Plant samples were collected in Central Swiss Alps, if not specified in Val d’Arpette, otherwise: Species Plant tissue Sampling sites altitude (m a.s.l.) MONOCOTYLEDONS Gramineae Calamagrostis villosa (Chaix.) Gmelin. Festuca halleri All. Festuca melanopsis Foggi, Rossi and Signori [ = F. puccinellii auct.] Festuca scabriculmis (Hackel.) Richter Nardus stricta L. Poa alpina L.a L, C L, C L, C 1900 2730 2090 L, C L L 2300 2300 2795 Cyperaceae Carex curvula All. Carex sempervirens Vill. L, C L, C 2470 2300 Ericaceae Arctostaphylos uva-ursi Spreng. Calluna vulgaris L. Loiseleuria procumbens Desf. Rhododendron ferrugineum Linn. Vaccinium myrtillus L. Vaccinium vitis-idaea L. L, L, L, L, L, L, W W W W W W 2420 2420 2370 2050 2180 1980 Betulaceae Alnus viridis (Chaix) DC L, W 2000 DICOTYLEDONS CONIFERS Pinaceae Abies alba Miller Juniperus nana Willd. Larix decidua Miller Picea abies (L.) Karsten Pinus cembra L. Pinus mugo Turra a N, N N, N, N, N, W W W W W 2050 1930 2050 2050 2050 2050 Furka Pass. EUKITT (mounting medium with refractive index of 1.5 at 20 jC). A count of 500 identifiable phytoliths per slide was carried out whenever possible. Few large fragments of silicified tissues (skeletons) were extracted from heavy silicified monocotyledons (e.g., Calamagrostis villosa): they were not considered in the count, this should not affect the representativeness of the samples since they were present in negligible quantities. For very poor samples, a standardised scanning of 3 h was 41 performed. The number of phytoliths counted was on average 300. Non-idioblastic silica particles, such as intercellular deposits or fragments, were not included in the count. Whenever possible the types were described quoting the botanical terminology for the plant cells from which they originated (Fahn, 1974), or employing the terminology commonly used in phytoliths studies, and referring to their shape. In this study, the phytoliths are identified by an acronym and an outline sketch (Table 2, Fig. 1). Since several trichome types are present, they are identified, for simplicity, as ‘‘TRI’’ followed by a number (Fig. 1). Light microscope photographs were taken using Kodak T MAX 400 asa professional film. A fraction of the same extract used for light microscopy was spread with the help of a paint-brush on double sided tape, mounted on aluminium stub, coated with gold and examined in a JEOL JSM 6400 scanning electron microscope operating at 15 kV, at a working distance of 15 mm and live time of 50 s. SEM photographs were taken with Agfa Pan APX 100 asa professional film. 2.2. Cluster analysis and principal components analysis Cluster analysis and PCA analysis were employed to test if taxonomically related species could be identified and grouped on the basis of the presence and frequency of types (STATISTICA 4.5 F for Windows). In woody species, the percentages of types were counted separately in leaf and wood tissues: for the analysis, wood and leaf type percentages were added together and the new sum normalised to 100%. For cluster analysis (Fig. 2), Ward’s method was applied (where the criterion of choice of linkage is the least increase in the sum of squared deviations from cluster means), and the algorithm of the Euclidean distance was used on normalised data (i.e. chord distance). However, in cluster analysis, the ordination can be biased by the arbitrary choice of data transformation and linkage algorithms. To test the strength of the ordination obtained, additional tests were also run with square root transformed data and presence/absence data. For comparison, PCA was also performed (Fig. 3) on square root transformed data. Finally, to 42 A.L. Carnelli et al. / Review of Palaeobotany and Palynology 129 (2004) 39–65 Table 2 Morphotype acronyms and description, anatomical origin (if known) and available literature Acronym Description Occurrence in plant taxa BLOCK RIDGED BULL CALL L CELL Block with a ridge in the middle, endoderm cells Bulliform cells (Parry and Smithson, 1958b) Long cells with sinuous edge CONE 1, CONE 2, CONE 3, CONE M P. abies, P. mugo M. C. vulgaris, V. myrtillus, V. vitis-idaea, R. ferrugineum C. curvula, C. sempervirens RECT PITTED Cones smooth or with sculpturing, distinctive of sedges (Ollendorf et al., 1987; Metcalfe, 1971). Isolated or in skeletons of 2, 3 or more Cork cells M. Jigsaw epidermal cells D, C Epidermal long cells M, D, C Epidermal long cells with grainy surface D, C Irregular waved blocky, mesophyll cells L. decidua Mesophyll cells L. decidua Mesophyll cells M, D, C Undulated hypodermal cells (Klein and Geis, 1978; Bozarth, 1993) P. abies Mesophyll palysade cells A. viridis Microhairs D, C Irregular rectangular with few pits M, D, C Polyhedral cells M, D, C A. alba, P. abies, P. cembra, Blocky polyhedrons transfusion cells often pitted (Klein and Geis, 1978; Bozarth, 1993) P. mugo Rectangular with many pits in rows D, C Rods: ROD BR ROD CW ROD FW ROD LS ROD LTS ROD MS Branched Rod cells (Parry and Smithson, 1958a) Rods long coarse wavy Rods long smooth fine wavy Rods long smooth (‘‘long cells’’ in Metcalfe, 1960) Rods long thin smooth Rods medium smooth N. stricta M, P. abies M, D M M, D, C M, D, C SPHE R SPHE S STO CYP; STO ERIC 1; STO ERIC 2; STO GRAM; STO CON; STO DIC TRACH Spherical rugose Spherical smooth Stomata complexes, respectively: Cyperaceae, Ericaceae, Gramineae, conifer and, dicotyledon type Tracheids D, C M, D, C Trichomas TRI 1/18 THREAD Trichomas, see Fig. 1 for variants Monocellular thread-like, non-segmented trichomas See Tables 3 – 5 M, C, D Trapezoids (short cell) TRA TRA CR TRA L TRA LOB TRA LTH TRA O TRA SB TRA V VES VES SP WAVY Trapezoids (‘‘short cells’’ in Metcalfe, 1960) Crenated based trapezoids Long trapezoids Lobated based trapezoids Long thin trapezoids Ornamented trapezoids Sinuous based trapezoids Narrow, chimney-like trapezoids Vessels Vessels with spiral thickening Ridge-like invaginations of the mesophyll Gramineae Gramineae Gramineae P. alpina, F. melanopsis Gramineae Gramineae Gramineae Gramineae A. viridis M, D, C L. decidua, P. abies CORK JGS L CELL L CELL GRAINY LARIX BLOCKY W LARIX SPIKY CELL MESO UND PALISADE PAPILLAE PLAT PERF POLY POLY CONIF D, C Occurrence in plants from this data set is summarised as: M (monocotyledons), D (dicotyledons) and C (conifers). For diagnostic morphotypes, the species or genus are listed. Outline drawings are in Fig. 1. A.L. Carnelli et al. / Review of Palaeobotany and Palynology 129 (2004) 39–65 Fig. 1. Outline sketches of phytolith morphotypes. 43 44 A.L. Carnelli et al. / Review of Palaeobotany and Palynology 129 (2004) 39–65 Fig. 2. Dendrogram resulted from the clustering analysis (Ward’s method, Euclidean distance, normalized data). The species are clustered on the basis of phytolith frequency. For the analysis, leaf and woody phytoliths were added. Two clusters are individuated: monocotyledons and dicotyledons with conifers. The latter consists of two clusters: Ericaceae and conifers with Alnus viridis. verify the robustness of the ordination, both cluster analysis and principal component analysis were run after adding similar types from all categories [respectively, all trapezoids (TRA), all rod-like cells (ROD), all Cyperaceae conical types (CONE), all trichomes (TRI)]. 3. Type descriptions and frequencies Line drawings of the types are displayed in Fig. 1 and an explanation of the acronyms in the text is given in Table 2. Percentage presence of the types and SEM micrographs and light microscope photos are given in Table 3 and Plates I– III for monocotyledons, in Table 4 and Plates IV – V for dicotyledons, and in Table 5 and Plates VI – IX for conifers. Descriptions of species types are divided into three sections: monocotyledons, dicotyledons (i.e., mainly Ericaceae) and conifers. 3.1. Monocotyledons (Table 3, Plates I –III) The species examined belong to Gramineae and Cyperaceae. Bulliform cells (BULL) occurred regularly in monocotyledons but were observed in very low percentages in the population here examined. Rods were common in grasses but not unique to this family, excepting ROD BR that was unique to Nardus stricta (Plate I, 5 –6). Trapezoidal types were unique to Gramineae (TRA, TRA L, TRA LOB, TRA O, TRA SB, TRA V, TRA CR; e.g., Plate I, 2 – 4); other taxonomically significant types were some trichome types (TRI 2, TRI 5, TRI 7; e.g., Plate II, 1 and 11 and stomata complexes (STO GRA; Plate I, 10). Unique to Cyperaceae types were cones (CONE 1, CONE 2, etc.; Plate III, 13 and 16), stomata complexes (STO CYP; Plate III, 13) and the trichome type TRI 18. A.L. Carnelli et al. / Review of Palaeobotany and Palynology 129 (2004) 39–65 45 Fig. 3. Scatter plot resulted from the PCA. Species are grouped on the basis of phytolith frequency. For the analysis, leaf and woody phytoliths percentage were added. Two groups are individuated: monocotyledons and conifers with dicotyledons; Carex species are isolated. Factor 1 explains 45% of variability of the data; factor 2 explains 17%. PCA loadings are given in Table 6. Variables V11, V12 and V15 are grouped with V9; variables V14, V16, V17, V20 and V21 are grouped with V10 and V19. Symbols: V1: Calamagrostis villosa, V2: Festuca melanopsis, V3: Festuca halleri, V4: Festuca scabriculmis, V5: Nardus stricta, V6: Poa alpina, V7: Carex sempervirens, V8: Care curvula, V9: Calluna vulgaris, V10: Vaccinium myrtillus, V11: Rhododendron ferrugineum, V12: Larix decidua, V13: Pinus mugo, V14: Pinus cembra, V15: Picea excelsa, V16: Abies alba, V17: Arctostaphylos uva-ursi, V18: Vaccinium vitis-idaea, V19: Juniperus nana, V20: Alnus viridis, V21: Loiseleuria procumbens. 3.2. Gramineae (Plates I– II) In Gramineae, epidermal cells, mesophyll cells, stomata complexes and trichomes were silicified. 3.2.1. Epidermal cells Elements of silicified epidermis were common in above-ground tissues of grasses (Plates I, 8, 13 – 14 and III, 3 –7). The most common types were ‘‘short cells’’ (Metcalfe, 1960), in the species examined here their general outline was trapezoidal (TRA), but several variants occurred (Fig. 1). The length of trapezoids was about 15 A for short trapezoids and about 30 A for long trapezoids, and width was about 10 A. Trapezoids (TRA) diagnostic of grasses, were idiomorphic and well silicified (Plates I, 2 –4, 7, 11 and II, 2, 4 –6, 13, 18). Short cells were among the first silica bodies to be described (Metcalfe, 1960), they were also defined as ‘‘hats’’ and ‘‘wavy – edged coastal rods’’ (Parry and Smithson, 1964), and as ‘‘nodular coastal rods’’ (Geis, 1978). Among the species examined here, trapezoids with a lobed base (TRA LOB; Plate I, 2 – 3) were dominant in Poa alpina (51%). Sinuous-based trapezoids (TRA SB) were common in Festuca melanopsis (Plate II, 18), Calamagrostis villosa and P. alpina (Plate I, 11), respectively, 50.4%, 27.4% and 23.9%. Rod-like long cells (ROD) were sub-epidermal columnar or rectangular cells. They were recovered isolated or in skeletons in all grass species, with percentages from 0.2% to 11%. Rods with narrow terminations were present (e.g., Plates I, 11 and II, 10). Rods were erratically filled with silica; the silicification may be complete or restricted to the cell walls. Rods varied in length (20 – 50 A), in width (5– 10 A) and in the extent to which the margins were sinuous (e.g., smooth, fine wavy, coarse wavy; see Fig. 1). Silicified projections may occur (Plate II, 8 – 9) (Hayward and Parry, 1975). Occasionally, rods presented concave ends 46 A.L. Carnelli et al. / Review of Palaeobotany and Palynology 129 (2004) 39–65 Table 3 Monocotyledon species: morphotypes counting as percentages and total number counted Calamagrostis villosa Morphotypes unique of monocotyledons BULL 1.2 CORK 11.4 STO GRAM 4.9 TRA 14 TRI 2 1.4 TRA LTH 0.6 ROD LS 4.7 TRA SB 27.4 TRA O – TRA L – TRI 7 5.5 TRI 5 0.4 TRA LOB – TRA V – CONE 1 – CONE 2 – CONE 3 – CONE M – STO CYP – TRA CR 2.4 ROD BR – TRI 18 – Festuca melanopsis 0.4 0.2 4 7.8 0.2 2 1.4 50.4 0.2 4.2 2.4 0.2 0.5 – – – – – – – – – Morphotypes occurring in monocotyledons and dicotyledons ROD FW – 1.4 TRI 1 0.4 2.2 TRI 3 6.7 0.7 TRI 4 – 0.2 TRI 16 – 4.7 TRI 15 0.2 0.9 TRI 6 – 0.9 TRI 17 – – TRI 12 – – TRI 8 – 0.2 TRI 13 – – TRI 9 – – TRI 10 – – Morphotypes occurring in monocotyledons and conifers ROD CW 5.3 7.4 Festuca halleri Festuca scabriculmis Nardus stricta 0.4 16.9 29.4 18.1 0.4 0.2 – – 2.8 – – – – – – – – 2.2 – – – – – – – – 2.1 1.9 – 3.6 0.2 0.2 4.4 – – – 0.2 2.5 0.2 – – – – 0.4 – – 0.2 – – – – 0.6 0.6 0.4 – – – – – – – – 0.7 0.2 0.2 0.2 0.7 0.2 – 0.2 0.6 9.8 – – – – – – – – – – – 4 – 0.9 6.6 0.9 3.8 – Carex sempervirens 0.2 17.2 0.2 46.9 0.2 0.9 0.2 2.1 1.9 3.5 1.2 – – 9.1 0.2 – – – 0.4 33.1 0.4 31.6 0.6 – Poa alpina 3.4 0.2 23.9 – 0.2 – – 51 – – – – – – – – – 0.9 1.7 0.6 – 0.4 – 0.2 – 0.2 0.4 – – – – – – – – – – – – 1.5 – – – – – – – – – – – – – 15.5 36.9 13.5 14.9 1.6 – – – 8.9 25.9 2.7 9.7 13.9 – – 3.6 – 1.2 – – 0.5 – – 4.6 – – 0.4 0.2 – – – – 9.0 9.5 11.0 1.3 7.0 0.2 – – – – – – – – 3.2 0.4 0.6 – – – 1.5 1.3 0.8 0.2 – – – – – 0.2 2.8 1.4 – Morphotypes occurring in monocotyledons, dicotyledons and conifers ROD MS 0.8 0.2 1.9 L CELL 2.2 6.9 0.2 POLY – – 3.7 ROD LTS 9.1 0.2 0.2 MESO 1.4 0.4 0.7 THREAD – – 0.2 PLAT PERF – – – SPHE R – – – VES SP – – – Tot. % 100 100 100 N. counted 507 552 431 100 497 0.2 – – 100 471 100 469 Carex curvula 0.2 – – 0.2 – 100 502 0.4 0.4 0.8 0.8 – – – 0.8 – 0.8 0.4 1.9 – 5 24.7 – – – – – 0.4 100 259 A.L. Carnelli et al. / Review of Palaeobotany and Palynology 129 (2004) 39–65 47 Plate I. Gramineae. SEM micrographs of silicified tissues from plant leaves; *light microscope photographs. Scale bar = 10 Am. (1) Poa alpina, trichoma.* (2) P. alpina, TRA LOB.* (3) P. alpina, TRA, TRA LOB.* (4) P. alpina, TRA V, VESS.* (5) Nardus stricta, ROD BR. (6) N. stricta, ROD BR.* (7) N. stricta, TRA.* (8) N. stricta, skeleton ROD CW, STO GRAM.* (9) N. stricta, TRI 2.* (10) Calamagrostis villosa, STO GRAM.* (11) C. villosa, TRA CW, ROD MS.* (12) C. villosa, EL CELL.* (13) C. villosa, ROD FW. (14) C. villosa, ROD CW. (15) C. villosa, ROD FW. 48 A.L. Carnelli et al. / Review of Palaeobotany and Palynology 129 (2004) 39–65 Table 4 Dicotyledon species: morphotypes counting as percentages and total number counted; L from leaves, W from wood Arctostaphylos uva-ursi Calluna vulgaris L L W Morphotypes unique of dicotyledons STO ERIC 1 8 15.7 STO ERIC 2 – – CALL L CELL – – TRI 11 – – TRI 14 – – STO DIC – – PALISADE – – VES – – W 27.4 – – – 1.1 – – – Loiseleuria procumbens Vaccinium myrtillus Vaccinium vitis-idaea Rhododendron ferrugineum Alnus viridis L L L L L – – 12.3 – – – – – W W W W 0.8 – – 1.3 – – – – 0.4 – – – – – – – – 32.7 2.8 – 0.2 – – – – 1.6 – – – – – – – 28.8 4.7 0.3 – – – – – 17.6 – – – – – – 2.9 – 12.8 – – – – – 23.5 – – – – – – – – – – – – 3.9 2.4 0.3 0.3 0.8 1.1 0.3 0.3 0.3 – – 0.8 0.3 – 0.5 – – – – – – – – – – – – – – 0.2 0.4 0.2 1.2 0.4 – 0.4 – 0.6 – – – – – – – – – – – – – – – – – 4.1 0.9 0.3 0.3 0.3 0.3 – – – – – – – – – – – – – – – – – – – – 0.7 2.2 0.5 3.9 2.9 1.2 2 1.2 3.2 – 1 – 0.5 – – – – – – – – – – – – – – – – – – – – – – – – – – 56.2 – – – – – 1.3 – – – – 0.4 28.8 – – 0.2 – – – – – – – – 45.6 – – 3.8 – – 28.6 – – – – – 8.6 – – 1.2 – – 20.6 – – – – – 3.6 8.0 3.3 1.2 – – Morphotypes occurring in monocotyledons, dicotyledons and conifers POLY 85.3 69.4 9.5 82.6 35 97.9 VES SP 0.8 – – 0.3 1.6 – MESO – – 0.7 – 0.5 – PLAT PERF – – – 0.3 – – TREAD – – 26.7 – – – L CELL – – – – – – ROD MS – – – – – – SPHE R – – – – – – ROD LTS 0.3 – – – – – Tot. % 100 100 100 100 100 100 N. counted 400 121 453 310 377 235 29 1.8 – – – 0.8 0.2 0.2 – 100 507 96.7 – – 1.7 – – – – – 100 120 6.6 1.6 0.3 0.3 – – – – – 100 320 52.7 – – 1.1 – – – – – 100 91 33.2 20.6 1.2 – – – – – – 100 407 55.9 – – – – – – – – 100 34 72.3 – 2.4 – – 1.5 0.3 0.3 – 100 336 Morphotypes occurring in monocotyledons and dicotyledons TRI 9 0.5 – – – TRI 8 1.3 – – – TRI 16 – – 0.2 – TRI 4 – – – – TRI 3 – – – – TRI 1 – – 0.9 – TRI 12 0.3 – 2.6 – TRI 13 – – – – TRI 15 – – 1.5 – ROD FW 0.3 – – – TRI 6 0.3 – – – TRI 10 0.3 – – – TRI 17 – – 8.2 – Morphotypes occurring JGS L CELL GRAINY PAPILLAE SPHE S RECT PITTED TRACH in dicotyledons and conifers 2.8 14.0 21.2 – – – – – – 0.3 – – – 0.8 – – – – 4.5 – – – – – in proximity of stomata (Plate I, 11). Branched rods (ROD BR; Plate I, 5– 6; about 50 A length) were unique to Nardus stricta (Parry and Smithson, 1958a), a small percentage (4%) was extracted either intact or fragmented. Cork cells are epidermal cells, crescent- or U-shaped (CORK; Plate II, 17; about 10 A). They were completely silicified and ubiquitous in the different grass species (0.2 – 33.1%); in 49 A.L. Carnelli et al. / Review of Palaeobotany and Palynology 129 (2004) 39–65 Table 5 Conifer species: morphotypes counting as percentages and total number counted; N from needles, W from wood Abies alba Larix decidua Picea abies Pinus cembra Pinus mugo Juniperus nana N N N N N N Morphotypes unique of conifers POLY CONIF 34.4 STO CON 5.4 WAVY – BLOCK RIDGED – UND – LARIX SPIKY CELL – LARIX BLOCKY W – W 0.4 – – – – – – W W W – – – – – – – 12.3 9.2 12.6 18.0 33 – – – – – – – – – 0.2 9.1 – – – – – – – – – – – – 71.1 – – 1.0 – – – 0.8 – – – – – – – 11.9 – – – – – Morphotypes occurring in monocotyledons and conifers ROD CW – – – – 0.8 – – – – – – Morphotypes occurring in dicotyledons and conifers RECT PITTED 1.1 0.4 – L CELL GRAINY 3.2 – – TRACH 11.8 – 4.8 PAPILLAE 6.5 – – SPHE S 1.1 – – JGS – – 1.6 – 0.8 – 2.7 – – 1.1 – – – – – 1.4 33.5 4.4 3.0 – – 0.3 – – – – – – 18.6 3.2 – 0.2 – 0.4 – – – – – – – – – – – dicotyledons and conifers 10.6 91.7 8.4 – 4.6 – 6.8 – 2.3 7.7 – – 1.3 – – 2.6 – – – – – – – – – – – 100 100 100 311 108 261 94.9 0.6 – – – – – – – 100 178 47.6 0.8 – – – – – – – 100 496 99.7 – – – – – – – – 100 320 4.4 – 0.2 – – 0.4 0.4 0.2 0.2 100 495 98.8 – – – – – – – – 100 250 Morphotypes occurring in monocotyledons, POLY 36.6 99.2 PLAT PERF – – THREAD – – L CELL – – MESO – – ROD MS – – SPHE R – – ROD LTS – – VES SP – – Tot. % 100 100 N. counted 93 256 – 0.3 20.9 – – 31.8 11.6 W – – – – – – Festuca halleri cork cells had crenate outline due to tiny projections. Fan-shaped bulliform cells (BULL) occurred in the adaxial surface of the leaf blade, they were silicified mainly in old leaves (Parry and Smithson, 1958b); BULL were rare in the species examined (0.2 – 1.2%). 3.2.2. Mesophyll cells Sponge-like mesophyll cells (MESO) were only occasionally silicified (0.2 – 1.4%). 3.2.3. Stomata Silicified stomata complexes (STO GRAM) were present in all grasses as epidermal fragments (silica 87.2 – – – – – – – – 100 109 skeletons) as well as isolated elements (25 A). They were characterised by the presence of guard-cells having their lumen more constricted in the middle. Stomata in Calamagrostis villosa (4.9%; Plate I, 10) and Festuca melanopsis (4%; Plate II, 15) had subsidiary cells with spiky edges. In Festuca scabriculmis stomata constituted 29.4% of the total phytoliths extracted (Plate II, 14). 3.2.4. Trichomes Trichomes having their lumen partially or totally infilled with silica were commonly recovered in grasses (Plates I, 1, 9 and II, 1, 11– 12). Tree trichome types were unique to grass species in this data set (TRI 50 A.L. Carnelli et al. / Review of Palaeobotany and Palynology 129 (2004) 39–65 A.L. Carnelli et al. / Review of Palaeobotany and Palynology 129 (2004) 39–65 2: Calamagrostis villosa, 1.4%; Festuca melanopsis, 0.2%; Festuca halleri, 0.4%; Nardus stricta, 0.6%; TRI 5: C. villosa, 0.4%; F. melanopsis, 0.2%; and TRI 7: C. villosa, 5.5%; F. melanopsis, 2.4%; F. halleri, 1.5%) (Table 3). 3.3. Cyperaceae (Plate III) In Cyperaceae, silicified cells were mainly epidermal cells and stomata complexes. 3.3.1. Epidermal cells The majority of phytoliths identified in sedges were conical (CONE; Plate III, 2– 5, 7– 8, 10– 13). CONE types mainly occurred in the epidermal cells that overlay the sclerenchyma occurring with the vascular bundles of the veins (Metcalfe, 1971; Ollendorf, 1992). From the top, a CONE appeared as a circle within a larger one, positioned on a pentagonal or hexagonal cell. From the side it was conical. They were recovered as single elements (CONE 1) or arranged in groups of two (CONE 2) or three (CONE 3) or more (CONE m) on one epidermal cell. These cells were counted as single elements, no matter the number of cones that they yielded (Table 3). Cells with several cones were very common in Carex sempervirens (65.3%). Carex sempervirens and Carex curvula presented both CONEs with psilate (smooth) surfaces (Plate III, 11), and with sculpturing (Plate III, 4, 13). The presence of surface ornamentation consisted of small subsidiary cones (satellites), scattered or clustered on the apex or periphery of the CONE. The ornamentation did not seem to be diagnostic at the species level in the population observed; however, more investigations are needed. In C. curvula, CONE types were about 5 Am in diameter. C. sempervirens tissues were more extensively silicified, and CONEs were generally bigger (about 10 Am) with more marked ornamentation. From the side, the psilate CONE could be confused with a conical trichome (TRI 9). CONEs were largely dominant in sedges; 51 the different types totalling 47.2% for C. curvula and 80.8% for C. sempervirens. Other epidermal cells were present in small percentages (ROD FW, ROD CW; Plate III, 12). The sum of ROD types was 10.8% and 3.1%, for Carex curvula and Carex sempervirens, respectively. Rodshaped phytoliths were reported for various species of Cyperaceae (Ollendorf et al., 1987). 3.3.2. Stomata In Cyperaceae, paracytic silicified stomata complexes (STO CYP; Plate III, 6, 14, 17) had clearly defined elongated dome-shaped subsidiary cells lying parallel to the stomata pore. Silicified stomata (measuring about 45 A in length) were present in both species, although more frequently in Carex curvula (13.9%), than in Carex sempervirens (1.6%). 3.3.3. Other cells Rare silicified mesophyll cells (MESO) (about 10 A) were present only in Carex sempervirens (0.2%); trichomes were also rare, in all up to 4%). Polyhedrons (POLY; Plate III, 2, 5) represented up to 24.7% of the phytolith production observed in Carex curvula, and were similar to the polyhedrons found in conifers, although they were generally smaller (about 30 A) and never yield bordered pits. 3.4. Dicotyledons (Table 4, Plates IV– V) The dicotyledons described here belong to the Ericaceae family, with the exception of Alnus viridis (Betulaceae). Studies of wild dicotyledons are scanty (Bozarth, 1992; Runge, 1996; Runge and Runge, 1997), and in particular the phytolith type of Ericaceae has never been investigated. Phytoliths extracted from leaves and wood were examined separately. Stomata complexes were characteristic in the Ericaceae (STO ERIC 1, STO ERIC 2; (Plates IV, 5 –6, 14, 17 and V, 5, 10). TRI 11 was present only in Loiseleuria procumbens and Vaccinium vitis-idaea (1.3% and 0.3% respectively) and TRI 14 occurred Plate II. Gramineae. SEM micrographs of silicified tissues from plant leaves; *light microscope photographs. Scale bar = 10 Am. (1) Festuca halleri, TRI 7. (2) F. halleri, ROD LS, TRA SB. (3) F. halleri, ROD CW, STO GRAM. (4) F. halleri, ROD FW, TRA. (5) F. halleri, ROD CW. TRA. (6) F. halleri, ROD CW, TRA.* (7) F. halleri, skeleton. (8) F. halleri, ROD FW with spiny projections.* (9) F. halleri, ROD FW. (10) F. halleri, ROD FW. (11) F. halleri, TRI 4. (12) Festuca scabriculmis, TRI 2.* (13) F. scabriculmis, STO GRAM.* (14) F. scabriculmis, STO GRAM. (15) Festuca melanopsis, STO GRAM. (16) F. melanopsis, ROD CW.* (17) F. melanopsis, CORK. (18) F. melanopsis, TRA SB.* 52 A.L. Carnelli et al. / Review of Palaeobotany and Palynology 129 (2004) 39–65 A.L. Carnelli et al. / Review of Palaeobotany and Palynology 129 (2004) 39–65 only in Calluna vulgaris and Vaccinium myrtillus (1.1% and 0.2%, respectively). The types most commonly occurring in leaves and wood were jigsaw epidermal cells (JGS; Plates IV, 13 and V, 7) these types are very common in dicotyledons in general. Silicified vessels and vessels with spiral thickening (VES, VES SP; Plates IV, 12, 15 –16 and V, 2) were also often recovered from leaf tissues, mainly in Rhododendron ferrugineum leaves (VESS SP, 20.6%), these types are not distinctive. 3.4.1. Epidermal cells Epidermal cells presented a wide morphometric variability. Two main groups were recognisable: cells with a sinuous outline (JGS and CALL L CELL) and cells with a polygonal outline (POLY). Silicified jigsaw epidermal cells were frequent (JGS; Plates IV, 13 and V, 7) and typical of dicotyledon cells. This type is described as anticlinal by Piperno (1988). Cells with a similar shape, however, occurred in a very low percentage in other groups (e.g., conifers; Plate VII, 10). They were usually recovered as silica skeletons. The highest percentages were recorded in Loiseleuria procumbens and Vaccinium vitis-idaea leaves (56.2% and 45.6%). Elongated and sinuous cells, probably of epidermal origin (CALL L CELL; Plate V, 8) were present in Calluna vulgaris wood (12.3%; Plate IV, 8), Vaccinium myrtillus leaves (2.8%), Vaccinium vitisidaea leaves (4.7%) and Rhododendron ferrugineum leaves (12.8%). Polyhedral cells (POLY) do not present distinctive morphological features; they are formed by fragments of silicified tissues of various histological origin. POLY are common in leaves (6.6 – 85.3%) and form most of the phytoliths in wood (from 52.7% to 97.9%). These types were dominant in Alnus viridis leaves and wood (55.9% and 72.3%). 3.4.2. Mesophyll cells Mesophyll cells (MESO) were only occasionally silicified in Ericaceae leaves (0.5 –0.7%; Plates IV, 2, 53 9 and V, 6, 12). In Alnus viridis, palisade mesophyll cells (PALISADE; 2.4%) were present. 3.4.3. Stomata Each stomata type (STO ERIC 1, STO ERIC 2; (Plates IV, 5 –6, 14, 17 and V, 4 –5, 10) may constitute up to 32.7% of phytoliths in leaves of the Ericaceae, and added together up to 65.4% in the leaves of Vaccinium myrtillus. They were also present in woody branches of most of the Ericaceae examined (0.4 – 23.5%). Calluna vulgaris and Rhododendron ferrugineum had similar stomata complexes consisting of many neighbouring cells which had no distinctive shape (STO ERIC 1) (this type also occurred in other Ericaceae species). The stomata complex may be covered by silicified cuticle. In R. ferrugineum, stomata were generally bigger (30 Am) than in C. vulgaris (20 Am). In the genus Vaccinium, the stomata apparatus consisted of a pair of guard cells with single subsidiary cells on each side placed parallel to the pore. In Vaccinium myrtillus, the stomatal edge had a double polar T-piece (STO ERIC 2; Plate IV, 14). 3.4.4. Trichomes In the species examined, several trichome types (TRI, Table 4) were regularly present in leaves although each usually showed low frequency (less than 2%). Only in Calluna vulgaris (total sum 14.5%) and Rhododendron ferrugineum (total sum 19.3%) trichomes were more abundant, while in the remaining heaths trichomes were less common (total from 2.7% to 6.5%). Trichomes typical of Ericaceae were TRI 11 and TRI 14. In C. vulgaris, surface hairs were simple uniseriate silicified hairs (THREAD, 26.7%; Plate V(11)). Arrow tip-like hairs (TRI 14) were present in this species (1.1%) and in Vaccinium myrtillus (0.2%). TRI 11 was present with low frequency in Loiseleuria procumbens (1.3%) and Vaccinium vitis-idaea (0.3%). 3.4.5. Vascular elements Vascular elements (VES, VES SP) were present in Ericaceae leaves (Plate IV, 3, 4, 15, 16) and wood Plate III. Cyperaceae. SEM micrographs of silicified tissues from plant leaves; *light microscope photographs. Scale bar = 10 Am. (1) Carex curvula, ROD CW. (2) C. curvula, CONE M. (3) C. curvula, CONE M. (4) C. curvula, CONE 2. (5) C. curvula, ROD LS. (6) C. curvula, STO CYP. (7) C. curvula, CONE M, side view.* (8) C. curvula, CONE M, top view.* (9) C. curvula, STO CYP.* (10) Carex sempervirens, ROD CW. (11) C. sempervirens, CONE 2, CONE 3. (12) C. sempervirens, ROD FW. (13) C. sempervirens, CONE 2. (14) C. sempervirens, STO CYP. (15) C. sempervirens, CONE 3.* (16) C. sempervirens, CONE 2.* (17) C. sempervirens, STO CYP.* 54 A.L. Carnelli et al. / Review of Palaeobotany and Palynology 129 (2004) 39–65 A.L. Carnelli et al. / Review of Palaeobotany and Palynology 129 (2004) 39–65 (Plate IV, 7– 8, 11– 12). VES SP were frequent in Rhododendron ferrugineum leaves (20.6%), but seldom in the leaves of the other Ericaceae studied (up to 1.8%). 3.4.6. Wood The most common types occurring in wood of Ericaceae were not idioblastic. Indeed, POLY were always more than 50%; other common types were JGS (in Arctostaphylos uva-ursi, 14%; Vaccinium vitis-idaea, 28.6%; Rhododendron ferrugineum, 20.6%) and stomata (STO) (A. uva-ursi, 15.7%; V. vitis-idaea, 17.6%; R. ferrugineum, 23.5%). 3.5. Conifers (Table 5, Plates VI –IX) Needle and wood phytoliths were examined separately. The types most frequently recovered were not idioblastic; they had a polygonal outline and were of variable size (POLY; e.g., Plate VI, 10 – 12). It must be stressed that this type although not distinctive of conifers was, however, the dominant type produced. Indeed, POLY were common in needles (4.4 – 87.2%) and dominant in wood (91.7 – 99.7%). In woody tissues, distinctive types were rare; nevertheless, the abundance of POLY types is itself characteristic of wood. Distinctive types in conifers were recovered mainly from needles. Silicified transfusion cells (POLY CONIF) were common in Pinus mugo needles (71.1%; Plate VI, 1– 2, 5, 9) and present in needles of Abies alba, Picea abies and Pinus cembra (34.4%, 12.3% and 0.2%, respectively). Silicified endodermal cells (BLOCK RIDGED) occurring in needles were diagnostic of P. abies (18%) and P. mugo (1%). Larix decidua could be distinguished on the basis of mesophyll tissues (LARIX BLOCKY W, 11.6%; Plate IX, 4 –6 and LARIX SPIKY CELL, 31.8%; Plate IX, 11). WAVY type (Plate IX, 1– 2) was present only in needles of P. abies and L. decidua (12.6% and 20.9%). Conifer stomata (STO CON; Plate VII, 6) were present with low frequency 55 (0.3 – 11.9%) in all species excepting P. mugo. UND type (hypodermal cells) was typical of P. abies (33% in needles). 3.5.1. Epidermal cells In conifers, cells in the inter-stomata files were elongated with smooth edges, similar to grass epidermal cells, albeit generally bigger in size. These cells, once fragmented, were likely to produce fragments with a polygonal outline (POLY) (Plate IX, 7 –9). Juniperus nana needles contained abundant POLY (Plate VIII, 7 – 11), probably in part also of epidermal origin. In this species, POLY had a grainy aspect when observed under a light microscope and, when observed at SEM, the surface appeared covered by many small spheres that could be either attached to the external surface or loose (Plate VIII, 6, 11). 3.5.2. Hypodermal cells Picea abies needles yielded characteristic hypodermal cells, with rectangular outline and with mineralised cell walls undulated on the long side and straight on the short side (UND; Plate VII, 3, 7– 9), finger-like projections may also occur. 3.5.3. Mesophyll cells Distinctive types were present in Larix decidua and Picea abies (WAVY; Plate IX, 1 –2). These types were probably ridge-like invaginations of the mesophyll, which have been described elsewhere as ‘‘suspended bridge-like’’ (Blinnikov, pers. comm.) or ‘‘square netting-like fragments’’ (Rovner, 1971). WAVY type was also recovered in Pinus sylvestris (Klein and Geis, 1978), Pinus banksiana and Picea glauca (Bozarth, 1993). LARIX BLOCKY W types probably also originated from silicified mesophyll cells (Plate IX, 4– 6). These cells were completely silicified; they had an undulating outline, grainy aspect and evident three-dimensional relief. LARIX SPIKY cells may be silicified mesophyll cells (31.8% in L. decidua needles). Plate IV. Ericaceae. SEM micrographs of silicified tissues from plant leaves (L = leaves, W = wood); *light microscope photographs. Scale bar = 10 Am. (1) Rhododendron ferrugineum (L), unknown. (2) R. ferrugineum (L), MESO, VESS SP. (3) R. ferrugineum (L), TRACH. (4) R. ferrugineum (L), VESS SP. (5) R. ferrugineum (L), STO ERIC 1.* (6) R. ferrugineum (L), STO ERIC 1. (7) R. ferrugineum (W), POLY. (8) R. ferrugineum (W), POLY. (9) R. ferrugineum (W), MESO. (10) Vaccinium vitis-idaea (W), POLY. (11) V. vitis-idaea (W), POLY. (12) V. vitisidaea (W), VESS.* (13) Vaccinium myrtillus (L), JGS.* (14) V. myrtillus (L), STO ERIC 2.* (15) V. myrtillus (L), VESS,* (16) Arctostaphylos uva-ursi (L), VESS. (17) A. uva-ursi (L), unknown. (18) A. uva-ursi (L), unknown. 56 A.L. Carnelli et al. / Review of Palaeobotany and Palynology 129 (2004) 39–65 Plate V. Ericaceae. SEM micrographs of silicified tissues from plant leaves (L = leaves, W = wood); *light microscope photographs. Scale bar = 10 Am. (1) Loiseleuria procumbens (L.), RECT PITTED. (2) L. procumbens (L), VESS. (3) L. procumbens (L), TRACH. (4) Calluna vulgaris (L), STO ERIC 1, MESO. (5) C. vulgaris (L), STO ERIC 1. (6) C. vulgaris (L), MESO. (7) C. vulgaris (L), JGS. (8) C. vulgaris (W), CALL L CELL. (9) C. vulgaris (W), papillated ROD. (10) C. vulgaris (L), STO ERIC 1.* (11) C. vulgaris (L), TRHEAD.* (12) C. vulgaris (L), MESO.* 3.5.4. Endodermal cells and vascular elements Ridged silicified endodermal cells were present in Picea abies (BLOCK RIDGED, 18%). Conifers produced cubic to polyhedral blocky elements, consisting of well-silicified transfusion elements, scattered or in pairs, often with bordered pits (POLY CONIF). Isolated or in skeletons, they were extremely abundant in Pinus mugo (71%; Plate VI(1– 2, 5, 9), less frequent in P. abies (Plate VII(11)), Abies alba, Pinus cembra and absent in Larix decidua. Silicified rectangular platelets with pits (RECT PITTED; Plate IX(3, 10) were rare (up to 1.4% in Pinus cembra needles). They originate from transfusion cells and tracheary systems. RECT PITTED were A.L. Carnelli et al. / Review of Palaeobotany and Palynology 129 (2004) 39–65 57 Plate VI. Conifers. SEM micrographs of silicified tissues from plant tissues (N = needles, W = wood); *light microscope photographs. Scale bar = 10 Am. (1) Pinus mugo (N), POLY CONIF. (2) P. mugo (N), POLY CONIF. (3) P. mugo (N), unknown. (4) P. mugo (N), unknown. (5) P. mugo (N), POLY CONIF. (6) P. mugo (N), unknown. (7) P. mugo (N), TRACH.* (8) P. mugo (N), TRACH.* (9) P. mugo (N), POLY CONIF.* (10) P. mugo (W), POLY. (11) P. mugo (W), POLY. (12) P. mugo (W), POLY. identified in all conifer wood and needle samples except for Larix decidua and Juniperus nana. These types are also present in dicotyledons. differentiate between species. Detailed descriptions of conifer stomata morphology are available elsewhere (Trautmann, 1953; Hansen, 1995). 3.5.5. Stomata Silicified stomata (STO CON) were detected occasionally in needles. In Picea abies (Plate VII, 6), Pinus cembra and Juniperus nana, stomata made up about 10% of total types. No attempt was made to 3.5.6. Wood Silicified cells in conifer wood were mainly nonidioblastic polyhedrons (POLY; Plates VI, 10 – 12, VIII, 2– 4 and IX, 12. Other types occurring in wood were POLY CONIF, RECT PITT and PLAT PERF. 58 A.L. Carnelli et al. / Review of Palaeobotany and Palynology 129 (2004) 39–65 Plate VII. Conifers. SEM micrographs of silicified tissues from plant tissues (N = needles, W = wood); *light microscope photographs. Scale bar = 10 Am. (1) Picea abies (N), UND. (2) P. abies (N), silica skeletons. (3) P. abies (N), UND. (4) P. abies (N), silica skeletons. (5) P. abies (N), TRACH. (6) P. abies (N), STO CONIF.* (7) P. abies (N), UND.* (8) P. abies (N), UND.* (9) P. abies (N), UND.* (10) P. abies (N), JGS.* (11) P. abies (N), POLY CONIF.* 3.6. Clustering analysis and PCA In the dendrogram shown in Fig. 2 (Wards’ method, Euclidean distance, normalized data), species are classified into two main clusters on the basis of phytolith frequency: the first cluster groups monocotyledons (grasses and sedges), and the second cluster groups dicotyledons and conifers. In the monocotyledon cluster, grasses were separated on the basis of short cells (TRA), bulliform cells (BULL) and stomata (STO GRAM). Among grasses, Poa alpina and Festuca scabriculmis were distinguished from other grasses because of the high frequency of TRA LOB, and STO GRAM, respectively. Within the monocotyledon cluster, Carex species were easily differentiated on the basis of CONE and STO CYP presence. Dicotyledons and conifers were assigned to the same cluster, mainly because of redundant polyhe- A.L. Carnelli et al. / Review of Palaeobotany and Palynology 129 (2004) 39–65 59 Plate VIII. Conifers. SEM micrographs of silicified tissues from plant tissues (N = needles, W = wood); *light microscope photographs. Scale bar = 10 Am. (1) Pinus cembra (N), silica skeleton. (2) P. cembra (W), POLY. (3) P. cembra (W), POLY with perforations scalariform. (4) P. cembra (W), POLY. (5) P. cembra (W), POLY. (6) Juniperus nana (N), SPHE S. (7) J. nana (N), POLY. (8) J. nana (N), POLY. (9) J. nana (N), POLY. (10) J. nana (N), POLY. (11) J. nana (N), POLY. (12) J. nana (N), TRACH.* dral types (POLY). However, two groups were differentiated: the first consisted of Ericaceae and the second of conifers and Alnus viridis. Ericaceae species were distinguished on the basis of the presence of the types STO ERIC 1, STO ERIC 2 and JGS. Conifers were grouped by distinctive types such as POLY CONIF. Alnus viridis was grouped with conifers rather than with Ericaceae because of the presence of epidermal long cells (L CELL GRAINY) and microhairs (PAPILLAE); types absent in the Ericaceae. It should be noted that Larix decidua yielded diagnostic types (LARIX BLOCKY W, LARIX SPIKY CELL) and that Pinus mugo was negatively differentiated from all conifers by the absence of silicified stomata (STO CON), and was characterised by the abundance of POLY CONIF. Thus, they were assigned to distinct branches of the cluster. The robustness of this classification was tested by using different data transformations (not transformed percentage data, root transformed data 60 A.L. Carnelli et al. / Review of Palaeobotany and Palynology 129 (2004) 39–65 Plate IX. Conifers. SEM micrographs of silicified tissues from plant tissues (N = needles, W = wood); *light microscope photographs. Scale bar = 10 Am. (1) Larix decidua (N), WAVY.* (2) L. decidua (N), WAVY. (3) L. decidua (N), TRACH.* (4) L. decidua (N), LARIX BLOCKY W.* (5) L. decidua (N), LARIX BLOCKY W. (6) L. decidua (N), LARIX BLOCKY W. (7) L. decidua (N), L CELL. (8) L. decidua (N), L CELL. (9) L. decidua (N), TRACH.* (10) L. decidua (N), LARIX SPIKY CELL. (11) L. decidua (W), POLY.* and, finally, with groups of similar types summed in one category, e.g., all trapezoids, not shown). No matter which transformation was applied, the two main clusters were maintained (monocotyledons, dicotyledons– conifers) and the classification of the monocotyledon cluster was unchanged. However, according to the transformation chosen, some species (Juniperus nana, A. viridis and Pinus cembra) were attributed differently within the second cluster. Discrepancies were as follows. Normalised (Fig. 2) and root-transformed data (not shown) gave the same results. If the analysis was run with non-transformed percentages, results were unchanged but the cluster grouping P. cembra, A. viridis and J. nana showed a weak affinity with the Ericaceae. With presence/ absence data, A. viridis and J. nana were not well classified within a cluster, and showed only a weak affinity with conifer and Ericaceae groups, respectively. Finally, when data of similar types were combined in categories (non-transformed percentage A.L. Carnelli et al. / Review of Palaeobotany and Palynology 129 (2004) 39–65 data; when several variants of a type were present in a species, they were added together: i.e., all trapezoids, rods, cones, trichomes and polygonal) P. cembra, A. viridis and J. nana were attributed to the Ericaceae cluster. In addition, when a test was run, keeping leaf and wood data separate (percentage data non-transformed, Euclidean distance; not shown), the classification was similar but a further cluster, grouping all wood samples, was present. This cluster showed an affinity between wood samples and samples from conifers needles of Juniperus nana and Ericaceae leaves of Arctostaphylos uva-ursi and Alnus viridis. PCA classification is a multivariate ordination technique, employed here to examine the degree of correspondence between types and taxa (Fig. 3, Table 6). The first axis of the principal components analysis explains the majority of the variability (45%). Such a high percentage can be explained by the strong diagnostic power of short cells that are unique and largely dominant in grass species. PCA run on square root transformed data confirmed the classification proposed by the cluster analysis: the species were divided into two main groups (monocotyledons and dicotyledons with coni- Table 6 PCA loadings for each species, labels are explained in Fig. 3 Species Factor 1 Factor 2 V1 V2 V3 V4 V5 V6 V7 V8 V9 V10 V11 V12 V13 V14 V15 V16 V17 V18 V19 V20 V21 0.2017522 0.1863757 0.0175233 0.1660261 0.0787601 0.0646531 0.0423945 0.4103878 0.8037481 0.9190497 0.8196555 0.7163232 0.7541657 0.8580057 0.7323243 0.8447424 0.9508866 0.7586679 0.8886325 0.8681059 0.9083434 0.7932636 0.6935168 0.8875576 0.7194143 0.8286935 0.6046596 0.0108916 0.0587367 0.0046828 0.0571643 0.0135351 0.0108368 0.0350921 0.0717067 0.0067158 0.0479644 0.0863568 0.0133594 0.1289849 0.0714046 0.0823253 61 fers). Carex species were separate. If the PCA was run with similar types added together in categories (as was done for cluster analysis) the classification did not vary, but points were less scattered (not shown). 4. Discussion In this study, biogenic silica extracted from 21 species either dominant or abundant at the subalpine – alpine ecocline were examined to obtain an inventory of phytolith types. Of the 21 species, 19 were studied for the first time for phytoliths. 4.1. Monocotyledons It is known that grasses and sedges yield typical phytoliths (e.g., Twiss et al., 1969; Mulholland and Rapp, 1992; Ollendorf, 1992). A morphological classification for short cells related to grass taxonomy has been proposed (Twiss et al., 1969). This classification is based on the fact that short cell morphology is in general distinctive in three of the five subfamilies occurring in the world: Festucoideaea, Panicoideae and Clorideae, the other two are Arundinoideae and Bambusoideae). All grasses examined here belonged to the Festucoideae and produced only short cells of trapezoidal shape (TRA sensu lato). Present data confirmed that the genera of this subfamily could not be distinguished on the basis of short cell morphology. In this set of grass species, the only types that were characteristic at species level were branched rods (ROD BR) that were unique to Nardus stricta (Parry and Smithson, 1958a). However, it should be noted that lobate trapezoids (TRA LOB), although not diagnostic, were largely dominant in Poa alpina. CONES were characteristic of the Cyperaceae family. However, a study on modern phytolith assemblages of subalpine – alpine grasslands showed that CONES are poorly preserved in soil (Carnelli, 2002; Runge, personal communication). Gramineae and Cyperaceae are known to produce a significant percentage of silicified above-ground tissues; in a previous study, it was shown that subalpine and alpine grasses and sedges from the Central Alps produced between 0.6% and 5.6% of silicification of dry plant mass. 62 A.L. Carnelli et al. / Review of Palaeobotany and Palynology 129 (2004) 39–65 4.2. Dicotyledons Phytolith studies often neglect wild dicotyledon species. The dicotyledons examined here belonged mainly to the Ericaceae. Species of this family are often dominant at the treeline, where they can form extensive heaths. We are aware of only one study mentioning the Ericaceae; it covers East African plants and discusses the species Vaccinium stanley (Runge, 1999). Subalpine –alpine heath species produced silicified cells mainly from the epidermis (JGS). These types are, however, commonly recovered in all dicotyledons. Distinctive stomata complexes (STO ERIC 1, STO ERIC 2) were present. The species studied accumulated low percentages of biogenic silica in leaves and wood, with a maximum value for Calluna vulgaris, in which silicified tissues were 0.68% of leaves’ dry mass (Carnelli et al., 2001). 4.3. Conifers Conifers are dominant in the subalpine forest and their contribution to phytolith assemblages cannot be neglected when investigating biogenic silica in plants at the treeline. A few species of conifers, mainly American, have been investigated (Klein and Geis, 1978), but data on European species are lacking: Picea abies, Abies alba, Juniperus nana, Pinus cembra and Pinus mugo were examined here for the first time. In conifers, silicification was detected both in needles and in wood. Larch epidermal phytoliths, varying greatly in size and shape but very thin and fragile, were earlier described for Larix laricina (Du Roi) K. Koch and Larix decidua Mill. (Klein and Geis, 1978). Klein and Geis (1978) after examining 15 taxa of Pinaceae stated that the identification of conifers was possible on the basis of tracheary elements, blocky polyhedrons and epidermal cells. However, on the basis of the present result, epidermal cells should not be taken as diagnostic, because they are similar to the long cells occurring in grass species, and fragmentation after deposition in the soil will make distinction impossible (Plate IX(7 –8)). Conifer-specific phytoliths (POLY CONIF), produced in tranfusion tracheids and endodermal cells, were very frequent in Pinus mugo. This type was also mentioned by Klein and Geis (1978) although fre- quency was not assessed. The same types were extracted from conifer needles (Abies balsamea, Picea glauca and Pinus banksiana) and topsoils in boreal forest (Bozarth, 1993). Non-idioblastic polyhedral phytoliths (POLY) were dominant in woody tissues. However, being highly redundant, they did not add any relevant taxonomic information and they should not be used for palaeoecological reconstructions. Among conifers, Larix decidua and Picea abies contained the highest proportion of biogenic silica (1.09% and 0.85% of dry mass, respectively) (Carnelli et al., 2001) and needles were extensively silicified, including epidermal tissues. In the remaining conifers, silicification was less extensive (about 0.01% of dry mass) and mainly endodermal cells were biomineralised. 4.4. Cluster analysis and PCA Cluster analysis and PCA results only partially supported the hypothesis that taxonomically related species can be identified on the basis of type presence and frequency. Indeed, while monocotyledon plants could be clearly identified, dicotyledons and conifers were less clearly separated. Both classification and ordination methods evidenced that the taxonomical power of phytoliths described here was limited for species yielding mainly redundant types such as Juniperus nana, Alnus viridis and Pinus cembra. These could not be unequivocally identified on the basis of phytoliths. In contrast, idioblastic types from key species may be diagnostic at family, genus or even at species level (e.g., Cyperaceae, Pinus mugo, Larix decidua and Calluna vulgaris). However, it is important to stress that, in general, herbaceous and woody species could be identified, the former yielding diagnostic types, the latter being defined by some specific types. This finding is relevant when applied to the study of assemblages of fossil phytoliths, e.g., in soils. 4.5. Potential relevance of phytoliths when studying the subalpine treeline The subalpine –alpine ecocline is a transition zone of heathland, stunted trees and meadows that occurs above the timberline (i.e., the limit of rather closed forest with trees 6– 8 m or taller). The present tim- A.L. Carnelli et al. / Review of Palaeobotany and Palynology 129 (2004) 39–65 berline in the Central Swiss Alps is located between 2000 and 2300 m a.s.l. However, its present elevation is largely anthropogenically determined (e.g., Theurillat et al., 1998; Tinner and Theurillat, 2003; Carnelli et al., in press). Land use since the Neolithic period has resulted in an artificial extension downward of alpine meadows, both for grazing and timber exploitation. Therefore, the uppermost limits attained by the forest during the Holocene, and even the natural occurrence of alpine meadows and heathlands are debated issues. The history and dynamic of vegetation at the treeline was investigated mainly by means of pollen, plant macrorest and soil charcoal analysis (e.g., Burga, 1991, 1995; Tinner et al., 1996; Wick and Tinner, 1997; Talon et al., 1998; Carcaillet and Brun, 2000; Carnelli et al., in press; Tinner and Theurillat, 2003). The analysis of soil-borne phytoliths may supply a new perspective in the reconstruction of the history of vegetation at the subalpine – alpine ecocline by enhancing the possibility of detecting taxa, such as monocotyledons, that are usually under-represented in palaeobotanical records. However, soil-borne phytolith assemblages were of limited use for palaeoecological reconstruction. (1) The diagnostic power of phytoliths was limited by the fact that similar types could, in some cases, be produced by plants of unrelated taxa. (2) The occurrence of a given type in plant tissues did not guarantee that it would be found in soil assemblages, since a differential preservation of phytoliths may occur in soil (Carnelli, 2002; Bartoli and Wilding, 1980). (3) Yet the combination of qualitative (i.e., the inventory of type frequencies) and quantitative (biogenic silica production) analysis may be a promising tool (Carnelli, 2002). It was shown in a previous study (Carnelli et al., 2001) that the heavily silicified aboveground tissues of grasses and sedges produce an annual biogenic silica input of 1.9 – 10.3 g m 2 in the subalpine and alpine meadows in the Central Alps. In comparison, the estimated annual production of biogenic silica from subalpine heaths would be between 0.19 and 1.13 g m 2, i.e., about one order of magnitude smaller than the predicted input of meadows, as well as the annual input of biogenic silica of a Larix decidua subalpine forest (2.4 g m 2 year 1) or of a Picea excelsa forest (0.7 g m 2 year 1). Thus, the combination of quantitative analysis of biogenic silica production and the inventory of type frequencies 63 may provide an estimate of the potential frequency of types in soil assemblages (Carnelli et al., 2001) and the comparison with findings in soil should help towards a better understanding of the taphonomical processes phytoliths undergo. In addition, the chemical composition of opal could be a valuable tool to source phytoliths lacking distinctive morphology: indeed, it was shown that the presence of aluminium in the reticule of biogenic silica was very frequent in woody species phytoliths but negligible in grass and sedge species (Bartoli and Wilding, 1980; Carnelli et al., 2002). It should be noted that, although conifers and dwarf shrubs produced low amounts of silica, phytoliths contained in persistent needles or leaves and in woody tissue persisted in the plant for several years. Therefore, they were less hydrated and often accumulated aluminium (Bartoli, 1985; Carnelli et al., 2002). For these reasons, once released in soil, they were likely to be better preserved than the more soluble phytoliths from yearly abscised Larix needles or grass leaves. 5. Conclusions In this paper, descriptions of species and related types were grouped under three taxa: conifers, dicotyledons (mainly Ericaceae) and monocotyledons. Interestingly, this framework reflected the three main physiognomical units of vegetation at the subalpine – alpine ecocline: coniferous forests, alpine heathlands (dominated by woody dicotyledons) and alpine meadows (dominated by monocotyledons). Our results showed that it was possible identify these three groups by using phytolith morphology to differentiate them and that there was a rationale for the utilisation of phytolith soil fossils in order to study past interactions between these plant communities. Indeed, quantitative and qualitative analysis of biogenic silica in subalpine –alpine soils may give a record of past altitudinal fluctuations of woody versus herbaceous vegetation during the Holocene, triggered by variations of climate and/or by the impact of human activities. This is now a subject of general concern in view of the predicted climate modification which is very likely to have an enormous influence on high alpine ecosystems (Theurillat et al., 1998; Theurillat and Guisan, 64 A.L. Carnelli et al. / Review of Palaeobotany and Palynology 129 (2004) 39–65 2001). Information about past ecosystem fluctuations is essential in order to understand patterns of vegetation resistance and resilience and thus be able to forecast future ecosystem responses. Acknowledgements This work was supported by the Swiss National Science Foundation (project FNRS 31-52911.97 to J.P. Theurillat). We thank Prof. Walter Wildi and Prof. Brigitta Ammann for logistic support and discussion; Rossana Martini, Philippe Clerc, Dario Zürcher and Luca Malgeri for help at SEM, light microscopy and with photographs, respectively. The English language was kindly revised by Julie Warrillow. The paper was improved, thanks to the revision of Freya Runge and Irwin Rovner. References Ball, T.B., Gardner, J.S., Anderson, N., 1999. Identifying inflorescence phytoliths from selected species of wheat (Triticum monococcum, T. dicoccon, T. dicoccoides, and T. aestivum) and barley (Hordeum vulgare and H. spontaneum) (Gramineae). American Journal of Botany 86, 1615 – 1623. Bartoli, F., 1985. Crystallochemistry and surface properties of biogenic opal. Journal of Soil Science 36, 335 – 350. Bartoli, F., Wilding, L.P., 1980. Dissolution of biogenic opal as a function of its physical and chemical properties. Soil Science Society America Journal 44, 873 – 878. Blackman, E., 1969. Observations on the development of the silica cells of the leaf sheath of wheat (Triticum aestivum). Canadian Journal of Botany 47, 827 – 838. Blackman, E., 1971. Opaline silica bodies in the range grasses of southern Alberta. Canadian Journal of Botany 49, 769 – 781. Blinnikov, M.S., 1994. Phytolith analysis and Holocene dynamics of alpine vegetation. Experimental investigations of alpine plants communities in the northwest Caucasus. In: Onipchenko, V.G., Blinnikov, M.S. (Eds.), Vëroffentlichungen des Geobotanischen Instituts der Eidgenossisch Technischen Hochschule, Stiftung Rübel, Heft, vol. 114. Zürich, pp. 23 – 40. Bozarth, S.R., 1992. Classification of opal phytoliths formed in selected dicotyledons native to the Great Planes. In: Rapp, G.J., Mulholland, S.C. (Eds.), Phytolith Systematics. Emerging Issues. Plenum, New York 350 pp. Bozarth, S., 1993. Biosilicate assemblages of boreal forests and aspen parklands. In: Pearsall, D.M., Piperno, D.R. (Eds.), Current Research in Phytoliths Analysis: Applications in Archeology and Paleoecology, vol. 10. University of Pennsylvania, Philadelphia, pp. 95 – 105. Burga, C.A., 1991. Vegetation history and palaeoclimatology of the Middle Holocene: pollen analysis of alpine peat bog sediments, covered formerly by the Rutor Glacier, 2510 m (Aosta Valley, Italy). Global Ecology and Biogeography 1, 143 – 150. Burga, C.A., 1995. Végétation et paléoclimatologie de l’Holocène moyen d’une ancienne tourbière située au front du Glacier du Rutor, 2510 m (Vallée d’Aoste, Italie). Revue de Géographie Alpine, 1. Carcaillet, C., Brun, J.-J., 2000. Changes in landscape structure in the northwestern Alps over the last 7000 years: lessons from soil charcoal. Journal of Vegetation Science 11, 705 – 714. Carnelli, A.L., 2002. Long term dynamics of the vegetation at the subalpine – alpine ecocline during the Holocene: comparative study in the Aletsch region, Val D’Arpette, and Furka Pass, (Valais, Switzerland). University of Geneva, PhD thesis no. 3378. Terre and Environnement, vol. 40, Imprimerie de la Section de physique, Genève. Carnelli, A.L., Madella, M., Theurillat, J.-P., 2001. Biogenic silica production in selected alpine plant species and plant communities. Annals of Botany 87, 425 – 434. Carnelli, A.L., Madella, M., Theurillat, J.-P., Ammann, B., 2002. Aluminum in the opal silica reticule of phytoliths: a new tool in palaeoecological studies. American Journal of Botany 89, 346 – 351. Carnelli, A. L., Theurillat, J.-P., Thinon, M., Talon, B., in press. Determination of the past uppermost treeline limit in the Central European Alps (Switzerland) based on soil and charcoal analysis. The Holocene. Fahn, A., 1974. Plant Anatomy Pergamon Press, Oxford. Geis, J.W., 1973. Biogenic silica in selected species of deciduous angiosperms. Soil Science 116, 113 – 119. Geis, J.W., 1978. Biogenic opal in three species of Gramineae. Annals of Botany 42, 1119 – 1129. Hansen, B., 1995. Conifer stomate analysis as a paleoecological tool: an example from the Hudson Bay Lowlands. Canadian Journal of Botany 73, 244 – 252. Hayward, D.M., Parry, D.W., 1975. Scanning electron microscopy of silica deposition in the leaves of barley (Hordeum sativum L.). Annals of Botany 39, 1003 – 1009. Klein, R.L., Geis, J., 1978. Biogenic silica in the Pinaceae. Soil Science 126, 145 – 156. Kondo, R., Peason, T., 1981. Opal Phytoliths in tree leaves—opal phytoliths in dicotyledonous angiosperm tree leaves. Research Bulletin of Obihiro University 12 (1), 217 – 230 (Japanese with English summary). Metcalfe, C.R., 1960. Anatomy of Monocotyledons: I. Gramineae Claredon Press, Oxford. Metcalfe, C.R., 1971. Anatomy of the Monocotyledons: II. Cyperaceae Claredon Press, Oxford. Mulholland, S.C., Rapp, G.J., 1992. A morphological classification of grass silica-bodies. In: Rapp, G.J., Mulholland, S.C. (Eds.), Phytolith Systematics. Emerging Issues. Plenum Press, New York, pp. 65 – 89. Mulholland, S.C., Lawlor, E.J., Rovner, I., 1992. Annotated bibliography of phytolith systematics. In: Rapp, G., Mulholland, S.C. (Eds.), Phytolith Systematics. Plenum Press, New York, pp. 277 – 322. A.L. Carnelli et al. / Review of Palaeobotany and Palynology 129 (2004) 39–65 Ollendorf, A.L., 1992. Toward a classification scheme of sedge (Cyperaceae) phytoliths. In: Rapp, G.J., Mulholland, S.C. (Eds.), Phytolith Systematics. Emerging Issues. Plenum Press, New York, pp. 91 – 111. Ollendorf, A.L., Mulholland, S.C., Rapp, G.J., 1987. Phytoliths from some Israeli sedges. Israel Journal of Botany 36, 125 – 132. Palmer, P.G., 1976. Grass cuticles: a new palaeoecological tool for East African lake sediments. Canadian Journal of Botany 54, 1725 – 1734. Parry, W.D., Smithson, F., 1958a. Silicification of branched cells in the leaves of Nardus stricta. Nature 182, 1460. Parry, W.D., Smithson, F., 1958b. Silicification of bulliform cells in grasses. Nature 181, 1549 – 1550. Parry, W.D., Smithson, F., 1964. Types of opaline silica depositions in the leaves of British grasses. Annals of Botany 28, 169 – 185. Piperno, D.R., 1988. Phytolith Analysis—An Archeological and Geological Perspective Academic Press, London. Rapp, G., Mulholland, S.C. (Eds.), 1992. Phytolith Systematics. Emerging Issues. Plenum Press, New York. Rovner, I., 1971. Potential of opal phytoliths for use in palaeoecological reconstruction. Quaternary Research 1, 343 – 359. Runge, F., 1996. Opal-Phytolithe in Pflanzen aus dem humiden und semi-ariden Osten AfriKas und ihre Bedeutung für die Klimaund Vegetationsgeschichte. Botanische Jahrbücher für Systematik 118, 303 – 363. Runge, F., 1999. The opal phytolith inventory of soils in central Africa—quantities, shapes, classification, and spectra. Review of Palaeobotany and Palynology 107, 23 – 53. Runge, F., Runge, J., 1997. Opal phytoliths in East African plants and soils. In: Pinilla, A., Juan-Tresserras, J., Machado, M.J. (Eds.), Estado Actual de los Estudios de Fitolitos en Suelos y Plantas, vol. 4. CSIC, Madrid, pp. 71 – 82. Talon, B., Carcaillet, C., Thinon, M., 1998. Etudes pédoanthracologiques des variations de la limite supérieure des arbres au 65 cours de l’Holocène dans les Alpes Francß aises. Géographie physique et Quaternarie 52, 195 – 208. Theurillat, J.-P., Guisan, A., 2001. Potential impact of climate change on vegetation in the European Alps: a review. Climatic Change 50, 77 – 109. Theurillat, J.-P., Felber, F., Geissler, P., Gobat, J.-M., Fierz, M., Fischlin, A., Küpfer, P., Schussel, A., Velutti, C., Zhao, G.-F., 1998. Sensitivity of plant and soil ecosystems of the Alps to climate change. In: Cebon, P., Dahinden, U., Davies, H.C., Imboden, D., Jaeger, C.C. (Eds.), Views from the Alps: Regional Perspectives on Climate Change. MIT Press, Cambridge, MA, pp. 225 – 308. Tinner, W., Theurillat, J.-P., 2003. Uppermost limit, extent and fluctuations of the timberline ecotone in the Swiss Central Alps during the past 11,500 years. Arctic, Antarctic and Alpine Research 35, 158 – 169. Tinner, W., Ammann, B., Germann, P., 1996. Treeline fluctuations recorded for 12,500 years by soil profiles, pollen, and plant macrofossils in the Central Swiss Alps. Arctic and Alpine Research 28, 131 – 147. Trautmann, W., 1953. Zur Unterscheidung fossiler Splatöffnungen der mitteleuropäischen Coniferen. Flora 140, 523 – 533. Twiss, P.C., Suess, E., Smith, R., 1969. Morphological classification of grass phytoliths. Soil Science Society of America Proceedings 33, 109 – 115. Whang, S.S., Kim, K., Hess, W.M., 1998. Variation of silica bodies in leaf epidermal long cells within and among seventeen species of Oryza (Poaceae). American Journal of Botany 85, 461 – 466. Wick, L., Tinner, W., 1997. Vegetation changes and timberline fluctuations in the Central Alps as indicators of Holocene climatic oscillations. Arctic and Alpine Research 4, 445 – 458. Wilding, L.P., Drees, L.R., 1973. Scanning electron microscopy of opaque opaline forms isolated from forest soils in Ohio. Soil Science Society of America, Proceedings 37, 647 – 650.