Phytoliths from the coastal savannas of French Guiana

Jennifer Watling; José Iriarte

The interpretation of Neotropical fossil phytolith assemblages for palaeoenvironmental and archaeological reconstructions relies on the development of appropriate modern analogues. We analyzed modern phytolith assemblages from the soils of ten distinctive tropical vegetation communities in eastern lowland Bolivia, ranging from terra firme humid evergreen forest to seasonally-inundated savannah. Results show that broad ecosystems – evergreen tropical forest, semi-deciduous dry tropical forest, and savannah – can be clearly differentiated by examination of their phytolith spectra and the application of Principal Component Analysis (PCA). Differences in phytolith assemblages between particular vegetation communities within each of these ecosystems are more subtle, but can still be identified. Comparison of phytolith assemblages with pollen rain data and stable carbon isotope analyses from the same vegetation plots show that these proxies are not only complementary, but significantly improve taxonomic and ecosystem resolution, and therefore our ability to interpret palaeoenvironmental and archaeological records. Our data underline the utility of phytolith analyses for reconstructing Amazon Holocene vegetation histories and pre-Columbian land use, particularly the high spatial resolution possible with terrestrial soil-based phytolith studies.

"Palms (Arecaceae) are a dominant feature of some landscapes in Africa and may provide different subsistence elements for humans such as construction materials, fabrics, fuel, food, medicine, and or- namentals. Palms have been identiﬁed through phytolith analyses in different localities and palae- oanthropological levels of Olduvai Gorge (northern Tanzania) dated to approximately 1.8 Ma. Thus, the presence of palms poses some interesting questions mostly related to the information they can provide in terms of ecological and vegetation reconstructions as well as the interaction between these plants and the possible use of them by hominins. We present here the study of modern soils from Serengeti National Park, Lake Eyasi and Lake Manyara, where palms are still present. The main goal is the reconstruction of palm landscapes and their interaction with other geographical factors (palaeolake, fresh water courses, etc.) through phytolith analyses and taking into account preservation conditions. Contrary to what was expected, characteristic palm phytoliths do not accumulate in the same amounts as observed in the palaeoanthropological samples from Olduvai Gorge. The short cell phytoliths of grasses, also common in the area, do not correspond to the grasses growing in the spots where the samples were collected. There are several reasons for these inconsistencies. Preservation of phytoliths is poor in disturbed areas, open grasslands and scarcely vegetated areas. Fresh watercourses may also inﬂuence in the number of phy- toliths and preservation caused presumably by erosion and water transport. This study reafﬁrms the idea that the absence in the archaeological and palaeontological/palaeoecological record of some phytolith morphotypes it is not always related to their absence on the past. Most importantly, the presence of plant groups is very signiﬁcant but the relative abundance is not easily interpreted because of preservation and disturbance."

Phytoliths recovered from pedological profiles in the Teotihuacan Valley, Mexico, as part of a paleoenvironmenal study of the region, are indicators of paleosols as well as fluctuations in the spatial distribution of soils and vegetation through time. In this analysis, grass (Poaceae) phytoliths represented by the subfamilies Pooideae, Panicoideae and Chloridoideae, are considered. A two-stage study included a cross-valley sampling

Quaternary International 287 (2013) 162e180 Contents lists available at SciVerse ScienceDirect Quaternary International journal homepage: www.elsevier.com/locate/quaint Phytoliths from the coastal savannas of French Guiana Jennifer Watling*, José Iriarte Department of Archaeology, College of Humanities, University of Exeter, Laver Building, North Park Road, Exeter EX4 4QE, United Kingdom a r t i c l e i n f o a b s t r a c t Article history: Available online 12 October 2012 Phytolith reference collections are a fundamental prerequisite for accurate interpretation of fossil phytolith assemblages used in reconstructing vegetation histories. As part of a multi-disciplinary research project studying Late Holocene landscape transformations associated with pre-Columbian raised ﬁeld complexes in the coastal savannas of French Guiana, phytolith production in selected plant species with high environmental and economic signiﬁcance native to the region was examined. A total of 49 families, 92 genera, and 108 species were analysed. Phytolith abundance in each specimen was rated qualitatively and morphotypes described following modern standards of nomenclature. Of the 92 non-Poaceae species tested, 37 contributed phytoliths that are diagnostic to at least the family level. Two of these are newly-discovered phytolith morphotypes isolated from Protium guianense (Burseraceae) and Thelypteris conﬂuens (Thelypteridaceae [Pteridophyta]) which have not been described previously. This work represents the ﬁrst systematic undertaking to establish a phytolith reference collection of French Guiana ﬂora. Results reinforce the usefulness of phytolith analysis for distinguishing ecologically signiﬁcant taxa, and therefore major vegetation formations. The creation of a comprehensive reference collection for French Guiana improves taxonomic resolution and has provided the necessary ground work for the interpretation of palaeoevironmental and archaeological records in the region. Ó 2012 Elsevier Ltd and INQUA. All rights reserved. 1. Introduction 1.1. Background information Phytoliths are opaline silica bodies produced in the leaves, stems, reproductive structures, and roots of plants. Monosilicic acid is drawn up through groundwater, transported to the aerial organs and deposited as solid silicon dioxide in cell wall inﬁllings, cell interiors and inter-cellular spaces (Pearsall, 2000; Piperno, 2006). The ever-increasing use of phytoliths in palaeoenvironmental studies, and the reﬁnement and standardisation of phytolith identiﬁcation, is now allowing researchers to tackle questions of climate change and past human land-use and diet (e.g. Fredlund and Tieszen, 1994, 1997; Alexandre et al., 1997; Blinnikov et al., 2002; Strömberg, 2004; Madella et al., 2005; Chandler-Ezell et al., 2006; Iriarte, 2006; Neff et al., 2006; Piperno et al., 2007; Zurro et al., 2009; Iriarte et al., 2010). The complementarity of phytolith and pollen data to reconstruct palaeoenvironments is particularly powerful and often allows ﬁner-grained interpretations of the nature and scale of past human impact on the environment (e.g. Kealhofer and Piperno, 1998; Denham et al., 2003; Iriarte et al., * Corresponding author. E-mail address: jgw203@ex.ac.uk (J. Watling). 1040-6182/$ e see front matter Ó 2012 Elsevier Ltd and INQUA. All rights reserved. http://dx.doi.org/10.1016/j.quaint.2012.10.030 2012; McMichael et al., 2012). Due to its resolution within grass (Poaceae) taxa, phytolith analysis is also a key tool to the study of grassland and savanna-forest mosaic dynamics over time (e.g. Fredlund and Tieszen, 1994; Alexandre et al., 1997, 1999; Strömberg, 2004). Reference collections of phytoliths from modern plants are crucial for advancing the discipline of phytolith analysis and for ensuring accurate, regionally sensitive interpretations of past vegetation. As such, reference collections now exist from all over the world (Kealhofer and Piperno, 1998; Runge, 1999; Lu and Liu, 2003a,b; Wallis, 2003; Kondo et al., 1994; Piperno, 2006; Mercader et al., 2009). Given the propensity of phytolith studies being carried out, reference collections available from tropical and subtropical South America are relatively few (Piperno, 1988, 2006; Piperno and Pearsall, 1998; Honaine et al., 2006; Iriarte and Alonso-Paz, 2009). This is particularly surprising considering the size and biological diversity of the area in question. In addition, many previous studies have had a strong focus on phytoliths from Poaceae, leaving non-Poaceae families somewhat understudied (Sendulksy and Labouriau, 1966; Campos de and Labouriau, 1969; Sondahl and Labouriau, 1970; Teixeira da Silva and Labouriau, 1970; Honaine et al., 2006). The same can also be said for other monocotyledonous families such as the Cyperaceae and Marantaceae which are known to produce highly speciﬁc phytoliths (Piperno, J. Watling, J. Iriarte / Quaternary International 287 (2013) 162e180 1989; Ollendorf, 1992; Honaine et al., 2009). While there is still a strong case for studying these families in order to reﬁne taxonomic resolution, much less attention has been given to phytolith production among woody dicotyledons, which represent the most taxonomically diverse group of plants in the region. This is despite previous analyses which have shown that diagnostic phytoliths can be retrieved from the reproductive structures of several families (Piperno, 1989, 2006). These factors considered alongside the heterogeneous nature of Neotropical vegetation formations makes creating regional collections in which all the major plant groups are represented all the more important. This study of phytoliths from species native to the French Guiana coast was carried out as part of international, cross-disciplinary research that is unravelling the nature and scale of Late Holocene landscape modiﬁcation associated with extensive pre-Columbian raised ﬁeld complexes in the region (Rostain, 2008, 2010; Iriarte et al., 2010, 2012; McKey et al., 2010). It represents one of few reference collections that exist for the Neotropics, and the very ﬁrst from French Guiana. Moreover, the seasonally-ﬂooded forestsavanna mosaic that characterises much of the study area is not a local phenomenon. Over 20% of the area termed as the Neotropics consists of grasslands (Da Silva and Bates, 2002). Therefore, the data from this study have the potential to be used by archaeobotanists and palaeoecologists working elsewhere. The taxa chosen for analysis were: Poaceae (4 subfamilies, 14 genera, 16 species), nongrass monocotyledons (6 families, 23 genera, 26 species), herbaceous dicotyledons (6 families, 8 genera, 8 species), woody dicotyledons (25 families, 43 genera, 46 species) and ferns (Pteridophytes) (11 families, 12 genera, 13 species). 1.2. Regional setting and vegetation The coastal savannas of French Guiana (Fig. 1) are bounded by a series of elongate sandy ridges (chéniers), representing Late Quaternary marine terraces that run parallel to the seashore (Prost, 1989; Plaziat and Augustinus, 2004). Annual rainfall in the region is between 2.5 and 4 m, and is highly seasonal, with most falling between December and July (Barrett, 2001). The vegetation is characterised by mangrove forest along the mudﬂats of the Atlantic sea shore, which is mud deposited sediments from the Amazon river, followed by a series of brackish and freshwater marshes, lowlying seasonally ﬂooded savannas and relatively higher, drier Fig. 1. Map showing the location of the French Guiana coastal plain. 163 savannas interspersed with ﬂooded and evergreen forest. Further inland, terra ﬁrme vegetation is tropical evergreen forest with areas of savanna. The mangrove forests of French Guiana stretch across 800 km2 of the coastal plain (Grenand et al., 2004). Dominant species include the white mangrove Avicennia germinans (Avicenniaceae), Laguncularia racemosa (Combretaceae) and Spartina brasiliensis (Poaceae). Where major rivers ﬂow out into the Atlantic, estuarine mangrove ecosystems dominated by Rhizophora racemosa (Rhizophoraceae) occur alongside Conocarpus erecta (Combretaceae) and large stands of “moucou moucous” Montrichardia arborescens (Araceae) (Granville, 1986). Behind the mangrove forests, areas of brackish water swamps and lagoons are dominated by the highly salt-tolerant sedge Eleocharis mutata (Cyperaceae), Cyperus articulatus (Cyperaceae), Amaranthaceae and Papilionoideae (Fabaceae) species, M. arborescens (Araceae) and Chrysobalanus icaco (Chrysobalanaceae) (Lindeman, 1953). Further inland, brackish water species are replaced by freshwater marsh vegetation. These formations, termed “wet to intermediate savannas” by Beard (1953) and “marsh savannas” by Granville (1986), are inundated for most of the year. Dominant species include Typha angustifolia (Typhaceae), Cyperaceae sedges (C. articulatus, C. giganteus, Scleria eggersiana, Rhynchospora spp. and Lagenocarpus guianensis), ferns (Blechnum serrulatum, (Blechnaceae), Thelypteris interrupta (Thelypteridaceae) and Acrostichum aureum (Polypodiaceae)), grasses (Echinochloa polystachya (Poaceae) and Leersia hexandra (Poaceae)) alongside a few arboreal species such as C. icaco and Mauritia ﬂexuosa (Arecaceae) (Lindeman, 1953; Granville, 1986). The relatively drier marsh savannas of the coastal plain are deﬁned predominantly by the presence of Panicoideae grasses which can tolerate seasonal ﬂooding (Axonopus spp., Echinolaena inﬂexa, Aristida tincta, Isachne polygonoides, Sacciolepis myuros, Panicum cyanescens) in association with Cyperaceae sedges (Fuirena umbellata and Cyperus spp.), Heliconia psittacorum (Heliconiaceae), Lentibulariaceae and Fabaceae herb species (Granville, 1986). It is in this environment that most of the pre-Columbian raised ﬁeld complexes are found (Rostain, 2008; Iriarte et al., 2010; McKey et al., 2010). In the upland savannas, a different assemblage of Panicoideae grasses occur including Axonopus aureus, Panicum cyanescens, Paspalum spp., Schizachyrum riedelii and Trachypogon plumosus, alongside Bulbostylis spp. and Rhynchospora spp. (Cyperaceae) (Hooke, 1971). Shrubs and small trees become common, including Curatella americana (Dilleniaceae), Burmannia sp. (Burmanniaceae), Byrsonima verbascifolia (Malpighiaceae) and Clusia nemorosa (Clusiaceae). Other herbaceous dicots include Melampodium camphoratum (Asteraceae) and members of the family Gentianaceae. The forest formations of the Guianan coast are as equally diverse as the savannas. Swamp (permanently inundated) forests dominated by Annona glabra (Annonaceae), Ficus spp. (Moraceae), Erythrina glauca (Fabaceae) and ferns T. interrupta and B. serrulatum occur behind the mangrove belt (Lindeman, 1953; Granville, 1986), while marsh (seasonally inundated) forests are common adjacent to river courses. The latter are characterised by stands of Euterpe oleracea (Arecaceae) known as “pinotieres” and other formations marked by the abundance of Virola surinamensis (Myristicaceae), Caryocar microcarpum (Caryocaraceae), Sloanea grandiﬂora (Elaeocarpaceae), Clusiaceae (Symphonia globulifera, Tovomita spp.), and Lecythidaceae (Eschweilera spp.) (Granville, 1986). Terra ﬁrme forest formations on the coastal plain are concentrated on the well-drained upper marine terraces. Dominant species include those from the families Fabaceae (Hymenaea courbaril, subfamily Caesalpinioideae, and Inga spp., subfamily Mimosoideae), Burseraceae (Protium heptaphyllum), Arecaceae 164 J. Watling, J. Iriarte / Quaternary International 287 (2013) 162e180 (Astrocaryum vulgare, A. murumuru, Oenocarpus bacaba), Chrysobalanaceae (Parinari campestris, Hirtella spp., Licania spp.), Simaroubaceae, Humiriaceae (Humiria spp.), Anacardiaceae (Tapiria guianensis) and Myrtaceae (Eugenia wullschlaegeliana). Herbaceous plants include Marantaceae (Ischnosiphon sp.), Phenakospermum guianensis (Strelitziaceae), various Piperaceae and the bamboo Parodiolyra luetzelbergii (Poaceae, Bambusoideae) (Lindeman, 1953). More detailed vegetation studies of the region can be found in Granville (1986), Hooke (1971), Lindeman (1953), and Beard (1953). regions of the world (e.g., Bozarth, 1992; Kondo et al., 1994; Runge, 1999; Pearsall, 2000; Lentfer, 2003; Piperno, 2006; Mercader et al., 2009). Owing to the sheer diversity of the ﬂora of the region, these primary selection criteria resulted in some families being underrepresented or unstudied (see Table 1). To remedy this, taxa which are ranked higher in their ability to produce phytoliths were prioritised over poor producers (e.g., Alismataceae, Pontederiaceae, Typhaceae), as were understudied families (e.g., Humiriaceae, Simaroubaceae). Some untested genera of ecologically signiﬁcant plants that are known to be poor producers were also analysed (e.g. Araceae, Malpighiaceae). Where families are known to produce diagnostic forms in speciﬁc plant parts, those parts were targeted for sampling, such as the leaves of Poaceae (Metcalfe, 1960), and the reproductive structures of the Cyperaceae and dicotyledons (Piperno, 1989, 2006). Particular attention was given to ferns because they can be diagnostic to speciﬁc wetland types (see Lindeman, 1953), they are good phytolith producers (Piperno, 2006, Table 1.1), and they have been largely understudied. 2. Materials and methods The plant species analyzed were chosen on the basis of two main criteria: (1) plant species with potential as environmental indicators due to their habitat speciﬁcity (Beard, 1953; Lindeman, 1953; Hooke, 1971; Granville, 1986), and (2) plant species abundance in the vegetation of the region with known phytolith production in related genera or families reported from other Table 1 Flora present in the study area. Sub-epidermal Monocotyledons 1 2 3 4 5 6 7 8 9 10 11 12 13 Araceae Montrichardia arborescens (L.) Schott leaf, stem, fruit Arecaceae Astrocaryum murumuru Mart. leaf, stem, fruit Attalea camopiensis (Glassman) Zona leaf, fruit Bactris maraja Mart. leaf, stem, fruit Desmoncus orthacanthos Mart. leaf, fruit Euterpe oleracea Mart. leaf fruit Geonoma maxima (Poit.) Kunth leaf, stem Oenocarpus bacaba Mart. leaf, stem, fruit Mauritia ﬂexuosa L. f. leaf fruit Socratea exorrhiza (Mart.) H. Wendl. leaf, fruit Cyperaceae Bulbostylis junciformis (Kunth) Lindm. leaf inﬂo B. lanata DC. seeds Cyperus articulatus L. leaf, inﬂo 14 C. giganteus Vahl inﬂo 15 Eleocharis mutata (L.) Roem. & Schult. inﬂo 16 Fuirena umbellata Rottb. inﬂo 17 Kyllinga brevifolia Nees inﬂo 18 Lagenocarpus guianensis Nees seeds 19 Remirea maritima Aubl. Environment Abundance Epidermis Globular Cyperaceae Troughed Conical/ Inter- Vesicular Hairs Hair Mesophyll MFBs Cystoliths Tracheary Stomata Other cones hat-shaped cellular inﬁllings bases elements MGV, FWM NP MF A p x x t A p A p A p A A p ech ech x t t A p ech t A C p ech ech A p ech C R p p EVF ech EVF x EVF x MF x EVF EVF A x MF, FWM x EVF x DS x x x x DS C BWM, FWM, MS A p ps x C p ps x C p ps x A p ps x A p gr x A p ps x t x FWM, MS x BWM, MS MS x t x t FWM FWM B x x J. Watling, J. Iriarte / Quaternary International 287 (2013) 162e180 165 Table 1 (continued ) Sub-epidermal Monocotyledons leaf, inﬂo 20 Rhynchospora gigantea Link inﬂo 21 R. triﬂora Bojer leaf, inﬂo 22 Scirpus asper J. Presl & C. Presl inﬂo 23 Scleria eggersiana Boeckeler inﬂo Heliconiaceae 24 Heliconia psittacorum Sesse & Moc. leaf, inﬂo Marantaceae 25 Ischnosiphon arouma (Aubl.) Körn. leaf, stem, inﬂo Strelitziaceae 26 Phenakospermum guianensis Aubl. leaf stem 27 28 29 30 31 32 33 34 Environment Abundance Epidermis Globular A a ps x x t x t x FWM A p gr R p ps A p gr x A p, a ps x A p gr A p gr, n, irr A A p p ech, gr, irr ech, gr, irr R p C p R p R p x FWM x FWM x FWM x t MS x EVF x t EVF Herbaceous dicotyledons Amaranthaceae Blutaparon vermiculare B (L.) Mears leaf, inﬂo Asteraceae DS Malampodium camphoratum (L. f.) Baker leaf, inﬂo Gentianaceae Schultesia brachyptera DS Cham. leaf, seeds Lentibulariaceae Utricularia reniformis A. MS St.-Hil. leaf, stem Phytolaccaceae Petiveria alliacea L. EVF leaf, stem, fruit Phytolacca rivinoides EVF Kunth & C. D. Bouche leaf, fruit Piperaceae Piper avellanum (Miq.) EVF C. DC. leaf, stem, inﬂo EVF Peperomia macrostachya (Vahl) A. Dietr. leaf, stem, inﬂo Woody dicotyledons Anacardiaceae 35 Tapirira guianensis Aubl. leaf fruit Annonaceae 36 Annona glabra L. leaf fruit 37 Guatteria guianensis (Aubl.) R. E. Fr. leaf fruit 38 Unonopsis stipitata Diels leaf, fruit, stem 39 Xylopia frutescens Aubl. leaf, fruit, stem Avicenniaceae 40 Avicennia germinans (L.) L. leaf Cyperaceae Troughed Conical/ Inter- Vesicular Hairs Hair Mesophyll MFBs Cystoliths Tracheary Stomata Other cones hat-shaped cellular inﬁllings bases elements R t x x ps x x t NP A p gr, ps C p gr, ps C NP pp gr C NP p A NP p p gr A p gr A p gr NP p x t x EVF x x x x x SF SF x x x t, scl x t x SF SF MGV (continued on next page) 166 J. Watling, J. Iriarte / Quaternary International 287 (2013) 162e180 Table 1 (continued ) Sub-epidermal Monocotyledons 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 fruit Burmanniaceae Burmannia bicolor Mart. leaf, stem, fruit Burseraceae Protium guianense (Aubl.) Marchand leaf fruit Cannabaceae Ampelocera edentula Kuhlm. leaf, stem, fruit Celtis iguanaea (Jacq.) Sarg. leaf seed Caryocaraceae Caryocar microcarpum Ducke leaf, stem, fruit Chrysobalanaceae Chrysobalanus icaco L. leaf fruit Parinari campestris Aubl. leaf, inﬂo Hirtella triandra Sw. leaf, fruit Licania macrophylla Benth. leaf, stem, fruit Clusiaceae Clusia nemorosa G. Mey. leaf, fruit Tovomita choisyana Planch. & Triana leaf, fruit Symphonia globulifera L. f. leaf, fruit Combretaceae Conocarpus erecta L. leaf, fruit Laguncularia racemosa (L.) C.F. Gaertn. leaf, fruit Terminalia guianensis L. leaf, stem, fruit Dilleniaceae Curatella america L. leaf fruit Davilla alata (Vent.) Benoist leaf fruit Doliocarpus guianensis (Aubl.) Gilg leaf, fruit Elaeocarpaceae Sloanea grandiﬂora Sm. leaf fruit Sloanea guianensis (Aubl.) Benth. leaf, fruit Erythroxylaceae Erythroxylum mucronatum Benth. leaf ,fruit Fabaceae (Caesalpinoideae) Dimorphandra ignea Ducke leaf, stem Hymenaea courbaril L. leaf, fruit Fabaceae (Mimosoideae) Environment Abundance Epidermis Globular Cyperaceae Troughed Conical/ Inter- Vesicular Hairs Hair Mesophyll MFBs Cystoliths Tracheary Stomata Other cones hat-shaped cellular inﬁllings bases elements NP DS C p gr, ps A C p* p gr A p A A p p* R p A A p A a, p ps R p ps R p ps R pp gr, ps EVF x EVF x x x x EVF x x x x x t MF gr BWM, FWM x gr, ps EVF x x t x t x EVF x EVF DS MF R gr, ps MF R p gr, ps C p C p gr C p gr x A A p p ps x x A R a, p p gr A pp gr A R p p R p A pp C p ps C p gr MGV MGV t MGV t, scl x DS x x EVF x x t x x t t EVF x x MF x MF gr t EVF x t, scl EVF EVF t x J. Watling, J. Iriarte / Quaternary International 287 (2013) 162e180 167 Table 1 (continued ) Sub-epidermal Monocotyledons 64 Inga stipularis DC. leaf, fruit 65 Inga cayennensis Sagot ex Benth. leaf, fruit Fabaceae (Papilionoideae) 66 Erythrina fusca Willd. leaf, stem, fruit 67 Sesbania exasperata Kunth leaf, fruit Humiriaceae 68 Humiria balsamifera J. St.-Hil leaf, fruit 69 Humiria ﬂoribunda (Mart.) Mart. ex Urb. leaf, fruit 70 Humiriastrum subcrenatum (Benth.) Cuatrec leaf, stem, fruit Icacinaceae 71 Emmotum fagifolium Desv. Ex Ham. leaf, fruit Lecythidaceae 72 Eschweilera apiculata Mart. ex DC. leaf, fruit Malpighiaceae 73 Byrsonima verbascifolia (L.) Rich. ex Juss. leaf, fruit Moraceae 74 Ficus guianensis Desv. ex Ham. leaf fruit 75 Sorocea muriculata Miq. leaf, stem, fruit Myristicaceae 76 Virola surinamensis (Rol. ex Rottb.) Warb. leaf, fruit Myrtaceae 77 Eugenia wullschlaegeliana Amshoff leaf, fruit Rhizophoraceae 78 Rhizophora racemosa G. Mey. leaf, stem, fruit Simaroubaceae 79 Picramnia guianensis (Aubl.) Jans.-Jac. leaf, fruit 80 Simarouba amara Aubl. leaf, fruit 81 82 83 84 85 Ferns Adiantaceae Pityogramma calomelanos L. leaf stem Azollaceae Azolla caroliniana Willd. leaf, stem Blechnaceae Blechnum serrulatum Rich. leaf, stem Dennstaedtiaceae Lindsaya stricta Dryand. leaf, stem Dryopteridaceae Dryopteris patula (Sw.) Underw. leaf, stem Environment Abundance Epidermis Globular Cyperaceae Troughed Conical/ Inter- Vesicular Hairs Hair Mesophyll MFBs Cystoliths Tracheary Stomata Other cones hat-shaped cellular inﬁllings bases elements EVF R p gr x x R p A p ps A p ps t C p gr t C p gr C p ps EVF t BWM FWM EVF EVF x t EVF t EVF NP MF R gr DS R p t A C p p gr gr, ps x x x A p gr x x A p gr, ps t C p gr t, scl R p C p gr C p gr A R a R p A p, a C sm SF x t EVF x t MF EVF MGV scl EVF x t EVF x x scl MF, FWM x t x x x BWM SF, FWM x x MS t MF, FWM NP (continued on next page) 168 J. Watling, J. Iriarte / Quaternary International 287 (2013) 162e180 Table 1 (continued ) Sub-epidermal Monocotyledons Environment Abundance Epidermis Globular Hymenophyllaceae 86 Trichomanes hostmannium (Kl.) Kze. leaf, stem Lycopodiaceae 87 Lycopodium cernuum L. pratilobie leaf, stem Oleandraceae 88 Nephrolepis biserrata Sw. leaf, stem Polypodiaceae 89 Acrostichum aureum L. leaf, stem 90 Ceratopteris thalictroides Hook. leaf, stem Salviniaceae 91 Salvinia auriculata Aubl. leaf Thelypteridaceae 92 Thelypteris conﬂuens (Thunb.) C. V. Morton leaf 93 Thelypteris interrupta (Willd.) K. Iwats. leaf, stem Cyperaceae Troughed Conical/ Inter- Vesicular Hairs Hair Mesophyll MFBs Cystoliths Tracheary Stomata Other cones hat-shaped cellular inﬁllings bases elements MF A p x A p R p C p C p C p R p, a x R p, a x DS x scl x x FWM FWM FWM BWM SF t x FWM, SF Key: p ¼ polyhedral epidermis, a ¼ anticlinal epidermis, pp ¼ pitted polyhedral epidermis, p* ¼ diagnostic polyhedral epidermal form, gr ¼ granulate, ps ¼ psilate, ech ¼ echinate, n ¼ nodular, irr ¼ irregularly angled/folded, t ¼ tracheid, scl ¼ sclereid, EVF ¼ evergreen forest, MF ¼ marsh forest, SF ¼ swamp forest, MGV ¼ mangrove forest, BWM ¼ brackish water marsh, FWM ¼ freshwater marsh, MS ¼ marsh savanna, DS ¼ dry savanna, B ¼ beach vegetation. The “Other” category shows the occurrence of other diagnostic phytoliths not mentioned in the table. Samples were collected from plant specimens at the Herbier de Guyane (CAY, Cayenne, French Guiana) and the Kew Royal Botanical Gardens (K, London) and curated at the Archaeobotany Laboratory at the University of Exeter, UK. Two specimens from the same target species were sampled for phytolith extraction. When possible, all individual components of the plant (e.g. leaf, culm, twig, inﬂorescence, and fruit) were sampled. For species in which phytolith production is not known, all plant parts were digested and mounted together. These parts were later mounted separately if any potentially diagnostic phytolith morphotypes were observed. Taxonomic names followed Tropicos Botanical Information System at the Missouri Botanical Garden (www.tropicos.org) and Poaceae subfamily assignments were made according to the Catalogue of New World Grasses (CNWG) (Judziewicz et al., 2000; Peterson et al., 2001; Zuloaga et al., 2003). Phytoliths were extracted from plant material by the wet oxidation method described in Piperno (2006) and were mounted on slides in Permount at University of Exeter. They were identiﬁed, counted and measured under a Zeiss Axioscope 40 light microscope at 500 magniﬁcation and photographed and measured using Carl Zeiss Axiovision software. For each specimen, at least 5 transects of the prepared slide were scanned for phytoliths. Phytolith production was measured qualitatively using categories devised by Wallis (2003): Abundant (A): a number of phytoliths observed in the majority of ﬁelds of view; Common (C): a small number of phytoliths observed in multiple ﬁelds of view; Rare (R): a small number of phytoliths observed with most ﬁelds of view containing no phytoliths; Nonproducer (NP): no phytoliths observed. When possible phytoliths were named and described following the ICPN descriptors deﬁned by Madella et al. (2005). 3. Results Results of the analysis are described below and summarized in Tables 1 and 2. Phytolith descriptions are split into two categories: phytoliths with limited to no taxonomic value and phytoliths with diagnostic value at and below family level. Within the ﬁrst category, the description of phytoliths follows plant taxonomy, described according to their anatomical origin. Diagnostic phytoliths are ordered according to families. Table 2 Poaceae sampled for phytolith analysis. Environment Poaceae Arundinoideae Aristida tincta Trin. And Rupr leaf, inﬂo Gynerium sagittatum (Aubl.) P. Beauv. leaf, inﬂo Abundance Bilobates A x Polybates Crosses Rondels Spooled/ horned towers Saddles Rondeloid/ saddleoid x x Saddeloid/ biloboid Elliptoid/ biloboid Irregular, complex Bulliforms Hairs MS x x FWM A x x x J. Watling, J. Iriarte / Quaternary International 287 (2013) 162e180 169 Table 2 (continued ) Environment Bambusoideae Olyreae Parodiolyra luetzelburgii (Pilg.) Soderstr. & Zuloaga leaf, stem, inﬂo Chloridoideae Distichlis spicata (L.) Greene leaf, inﬂo Spartina alterniﬂora Loisel. leaf, inﬂo Micrairoideae Isachne polygonoides (Lam.) Döll leaf, inﬂo Panicoideae Axonopus aureus P. Beauv. leaf, inﬂo Axonopus surinamensis (Hochst. Ex Steud.) Henrard leaf, inﬂo Echinochloa polystachya (Kunth) Hitchc. leaf, inﬂo Echinolaena inﬂexa (Poir.) Chase leaf, inﬂo Panicum cyanescens Nees ex Trin. leaf, inﬂo Paspalum parviﬂorum Rhode ex Flüggé leaf, inﬂo Paspalum pulchellum Kunth leaf, inﬂo Sacciolepis myuros (Lam.) Chase leaf, inﬂo Schizachyrium riedelii (Trin.) A. Camus leaf, inﬂo Trachypogon spicatus (Humb. & Bonpl. Ex Willd.) Nees leaf, inﬂo Abundance Bilobates A x Polybates Crosses Rondels Spooled/ horned towers Saddles Rondeloid/ saddleoid Saddeloid/ biloboid Elliptoid/ biloboid Irregular, complex Bulliforms Hairs x x x x EVF x x BWM A x BWM A x A x x MS x DS A x x x A x x x A x x x A x A x x A x x x x A x x x x A x x x x A x x x x A x MS FWM x MS x x x MS, DS DS DS MS MGV, DS DS x x EVF ¼ evergreen forest, MF ¼ marsh forest, SF ¼ swamp forest, MGV ¼ mangrove forest, BWM ¼ brackish water marsh, FWM ¼ freshwater marsh, MS ¼ marsh savanna, DS ¼ dry savanna, B ¼ beach vegetation. 3.1. Common phytoliths with limited taxonomic value 3.1.1. Epidermal phytoliths As shown in Table 1, siliciﬁed polyhedral (Fig. 2a) and anticlinal epidermal cells (Fig. 2b and c) are characteristic of monocotyledons, dicotyledons, and Pteridophytes, and they are well represented in both arboreal and herbaceous taxa. They are thus of no taxonomic value. As noted in previous studies (Geis, 1973; Bozarth, 1992; Piperno, 2006), siliciﬁcation seems to occur at random in most of the species studied. Many epidermal fragments demonstrate a mosaic of heavily siliciﬁed and slightly siliciﬁed cells, and silica tends to concentrate more consistently in the cell walls and lumina of epidermal hairs and hair bases. In some plant taxa, such as Tapirira guianensis (Anacardiaceae), C. nemorosa (Clusiaceae), Doliocarpus guianensis (Dilleniaceae), and Erythroxylum mucronatum (Erythroxylaceae), the epidermal cells appear as plates with pits or as aggregations of what seems to be ﬁbrous material (Fig. 2d). P. campestris (Chrysobalanaceae) produced thick, irregular phytoliths from the polyhedral epidermis (Fig. 2e). 3.1.2. Hairs and hair bases Trichomes or hair cells are among the most common and diverse of phytolith morphotypes and can be of considerable diagnostic signiﬁcance (Kondo et al., 1994; Piperno, 2006). However, the 170 J. Watling, J. Iriarte / Quaternary International 287 (2013) 162e180 Fig. 2. Phytoliths with little or no taxonomic value. a) Polyhedral epidermis from Humiriastrum subcrenatum, b) Anticlinal epidermis from Davilla alata leaf, c) Anticlinal epidermis from Scleria eggersiana inﬂo d) Decorated epidermis from Clusia nemorosa, e) Thick, irregular polyhedrals from Parinari campestris, f) Trichome from Rhynchospora gigantea inﬂo, g) Hair base from Annona glabra leaf, h) Hair base from Chrysobalanus icaco leaf, i) Hair base from Conocarpus erecta, j) Tracheid from Unonopsis stipitata, k) Sclereid from Erythroxylum mucronatum, l) MFB from Sloanea grandiﬂora leaf, m) Globular granulate phytolith from Burmannia bicolor, n) Globular psilate phytolith from Licania macrophylla, n) Trichome with V1 material from S. amara, o) Vesicular inﬁlling from Simarouba amara. Scales ¼ 20 mm. trichomes observed from the French Guiana ﬂora were very similar amongst unrelated taxa. The most common trichomes were V-shaped hair cell-wall siliciﬁcations that varied in size considerably within and among species (Fig. 2f). They were most common in woody dicots, but were also observed in herbaceous dicots, ferns, and amongst the Cyperaceae (see Table 1). Prickle hairs diagnostic of Poaceae were present in most of the grass species tested. Trichomes that were of more diagnostic potential are discussed below. Hair bases were observed exclusively in the dictoyledons, with the majority being of limited taxonomic value. They were found in T. guianensis (Anacardiaceae), Guatteria guianensis and A. glabra (Annonaceae), P. campestris (Chrysobalanaceae), D. guianensis (Dilleniaceae) and Simarouba amara (Simaroubaceae). These forms were largely spherical with a circular mark in the centre (Piperno, 1988) (Fig. 2g). C. icaco (Chrysobalanaceae) contributed a different type of hair base (Fig. 2h) that has stellate projections surrounding the centre. According to Metcalfe and Chalk (1983), stellate hairs occur in the bases of certain species of the Chrysobalanaceae and these are especially conspicuous in a few genera, including Chrysobalanus. This hair base is grouped into phytoliths with little taxonomic value as the full diagnostic potential of this morphotype is also unknown. Circular to acircular hair bases isolated from C. erecta (Combretaceae) had echinate projections emanating from the cell wall (Fig. 2i). The hair base was reported from Combretum imberbe (Mercader et al., 2009, Fig. 3as) but the diagnostic potential of this phytolith is as yet unclear until more studies are conducted within the family. More diagnostic hairbases were found in members of families Asteraceae, Dilleniaceae, Moraceae and Cannabaceae (see section 3.2). J. Watling, J. Iriarte / Quaternary International 287 (2013) 162e180 3.1.3. Mesophyll Mesophyll phytoliths of little taxonomic value were observed exclusively in woody dicots, including G. guianensis (Annonaceae), C. icaco (Chrysobalanaceae) Davilla alata (Dilleniaceae), S. grandiﬂora (Elaeocarpaceae) and Amplocera edentula (Cannabaceae). 3.1.4. Tracheary elements Tracheids of no diagnostic signiﬁcance were commonly found in monocots, dicots and a few ferns (see Table 1) (Fig. 2j). Sclereids were rarer and observed in woody dicots, including G. guianensis (Annonaceae), T. guianensis (Anacardiaceae), E. mucronatum (Erythroxylaceae) and Eugenia wullschlaegelina (Myrtaceae), but also in the fern Lycopodium cernuum (Lycopodiaceae). Sclereids isolated from E. mucronatum were far larger (up to 200 mm) and more abundant than those from the other species analysed in this study (Fig. 2k). Siliciﬁed sclerenchyma is sometimes known to take the form of multi-faceted bodies, categorised separately due to the multiple origins of this morphotype (see section 3.1.6). 3.1.5. Stomata Siliciﬁed stomata were frequently encountered in all types of plant tested in this study. Only the stomatal complexes found in the Arecaceae and Cyperaceae are of potential diagnostic value and these are brieﬂy described in section 3.2. 3.1.6. Multi-faceted bodies Multi-faceted bodies (MFBs) were recorded from one herbaceous dicot (Piper allevanum, Piperaceae) but were mostly found in woody dicots, including D. alata (Dilleniaceae), C. icaco (Chrysobalanaceae), S. grandiﬂora (Eleaeocarpaceae) (Fig. 2l) and Inga stipularis (Fabaceae). MFBs are known to originate from various tissues (Runge, 1999) and can be siliciﬁed terminal tracheids (Postek, 1981), siliciﬁed inter-cellular spaces between mesophyll cells (Metcalfe and Chalk, 1979) or sclerenchyma (Piperno, 2006). They are generally large and elongated silica bodies whose facets are most likely formed by contact with cell walls. C. icaco was the only species to produce MFBs over 100 mm in length, but with this exception this phytolith type offers little potential to differentiate among taxa. 3.1.7. Globular phytoliths Two types of globular phytolith, granulate (Fig. 2m) and psilate (Fig. 2n), are considered here as being of limited taxonomic value. The globular granulate category encompasses a range of decorated spherical and aspherical bodies whose surface decoration may be rugulose, verrucose or nodular (e.g., Piperno, 2006; Iriarte and Alonso-Paz, 2009). They have been considered by many researchers as being largely diagnostic of arboreal taxa (Amos, 1952; Geis, 1973; Scurﬁeld et al., 1974; Kondo et al., 1994; Piperno, 2006). Although this phytolith type was found in 27 of the 46 species of woody dicots sampled, it was also present in a signiﬁcant portion of non-woody plants, with over one-fourth of the monocots and onefourth of the herbaceous dicots documented as producers (Table 1.1). The dimensions of globular granulate phytoliths are thought to be of importance as a diagnostic trait, as previous studies found that dicots typically produce spheres 3e9 mm in diameter and monocots spheres 9e25 mm in diameter (Piperno, 1988, 59). In this study, the majority of globular granulate phytoliths from dicots and monocots all fell within a range of 5e11 mm in diameter. In addition, while globular granulate phytoliths from monocots included forms much larger than this (e.g., 15 mm in species of Heliconiaeae), two dicot species produced spheres just as large (D. guianensis, Dilleniaceae and C. nemorosa, Clusiaceae). This observed degree of overlap between unrelated taxa should be considered in future regional phytolith studies. However, when 171 found in abundance, globular granulate phytoliths are still a reliable indicator of tree cover in the past (Alexandre et al., 1997; Piperno, 2006). Like globular granulate bodies, globular psilate phytoliths are commonly produced in the leaves and stems of woody dicots but are also produced in herbaceous monocots (Carnelli et al., 2001; Piperno, 2006; Iriarte and Alonso-Paz, 2009). These phytoliths are the same as Piperno’s (1988) spherical smooth phytoliths, though a differentiation is made here between these and vesicular inﬁllings (Strömberg, 2003). Among the species tested, 20 were found to produce globular psilate phytoliths and these all ranged between 5 and 11 mm in diameter. Cyperaceae were also consistent producers of this form. 3.1.8. Vesicular inﬁllings Vesicular inﬁllings are spherical to aspherical bodies consisting of smooth, opaque silica with concentric lamination (Geis, 1973; Strömberg, 2003, 2004). In this study, they were found attached to the polyhedral epidermis of E. mucronatum (Erythroxylaceae) and V1-type siliciﬁcation (Strömberg, 2003, 2004) was seen inﬁlling trichomes in P. guianensis (Simaroubaceae) (Fig. 2o). Vesicular inﬁllings occur as isolated bodies in Picramnia guianensis (Fig. 2p). The diagnostic potential of these phytoliths is not yet fully understood, but they are most likely exclusive to woody dicots (C. A. Strömberg, personal communication, 2010). 3.2. Phytoliths with diagnostic value at the family level and below 3.2.1. Non-grass monocotyledons 3.2.1.1. Arecaceae. Proliﬁc phytolith production was observed in all palm species tested. Globular echinate phytoliths were abundant in all plants parts tested from Attalea camopiensis, E. oleracea, Geonoma maxima, M. ﬂexuosa and Socratea exorrhiza. These phytoliths are spherical to aspherical, have echinate surface decoration and originate in sub-epidermal tissue (Tomlinson, 1961; Bertoldi de Pomar, 1971; Prychid et al., 2004; Passos and Menconça, 2006; Piperno, 2006). Although globular echinate phytoliths are considered of little diagnostic signiﬁcance below family level, interesting size and morphological differences between the species warrant further studies. For example, diameters of globular echinate phytoliths of E. oleracea (Fig. 3a) (6.7 3.1 mm, mean sd; range 4.3e23.3 mm; N ¼ 100) displayed a signiﬁcantly higher maximum value (23.3 mm) compared to, for instance, O. bacaba (9.7 3.2 mm, mean sd; range 4.7e16.1 mm; N ¼ 100) which had the second highest maximum value (16.1 mm), although having the highest mean value overall (9.7 mm). The spines were much more clearly deﬁned in E. oleracea, O. bacaba and A. camopiensis than in some other species, e.g., M. ﬂexuosa, where surface decoration sometimes appeared almost granulate (Fig. 3b). More aspherical phytoliths were observed in G. maxima. These preliminary results show that a study that takes into account a greater number of variables and incorporates a much larger sample of regional palm species could have the potential to discriminate among different palm genera (see Fenwick et al., 2011). Conical to hat-shaped bodies were identiﬁed in Astrocaryum murumuru, Bactris maraja, Desmoncus orthacanthos and S. exorrhiza. Unlike the conical bodies of the Cyperaceae, Arecaceae hat-shaped phytoliths have a ﬂatter apex and occur as singular forms instead of in platelets (Tomlinson, 1961; Piperno, 1988, 2006). As with the globular echinate phytoliths, there were noticeable differences in the size of the hat-shaped phytoliths that were isolated. Those of A. murumuru were by far the largest (15.9 4.0 mm, mean sd; range 8.6e27.5; N ¼ 50) (Fig. 3c). S. exorrhiza yielded the second largest phytoliths (11.2 2.0 mm, mean sd; range 7.92e14.76; 172 J. Watling, J. Iriarte / Quaternary International 287 (2013) 162e180 Fig. 3. Phytoliths from non-grass monocots. a) Globular echinate phytoliths from Euterpe oleracea, b) Globular echinate phytoliths from Mauritia ﬂexuosa, c) Conical body from Astrocaryum murumuru, d) Stomata from E. oleracea leaf, e) Adaxial epidermis from M. ﬂexuosa leaf, f) Cones from Rhynchospora gigantea inﬂo, g) Polygonal cone from Cyperus articulatus achene, h) Polygonal cone from Kyllinga brevifolia achene, i) Polygonal cone from Remirea maritime achene, j) Polygonal cone from Scirpus asper achene, k) Polygonal cone from R. gigantea achene, l) Troughed bodies from Heliconia psittacorum, m) Irregularly angled/folded bodies from Ichnosiphon arouma, n) Nodular sphere from I. arouma, o) Conical body from I. arouma, p) Druse-like bodies from Phenakospermum guianensis. Scales ¼ 20 mm. N ¼ 30), yet these averaged only 11.2 mm in diameter by comparison. Differences in surface decoration were also observed, the area around the apices appearing more decorated in A. murumuru. More detailed study is again needed to determine whether signiﬁcance can be attached to these observations. The stomatal complexes isolated from the leaves of E. oleracea are also diagnostic of the Arecaceae. As Fig. 3d shows, the inner edges of the guard cells are distinctly angled and they are of uniform thickness (Carnelli et al., 2001, 43). M. ﬂexuosa displayed large numbers of elongated phytoliths with sinuous edges which are casts of cells making up the adaxial epidermis of the leaves (Fig. 3e) (Tomlinson, 1961; Passos and Madonca, 2006). This type of epidermis is distributed among members of the palm family (Tomlinson, 1961). 3.2.1.2. Cyperaceae. Nearly all species of Cyperaceae tested produced cone-shaped phytoliths that are diagnostic of the family, and originate from epidermal cells overlying sclerenchyma tissue (Mehra and Sharma, 1965; Metcalfe, 1971; Piperno, 1988, 2006; Ollendorf, 1992; Kondo et al., 1994; Honaine et al., 2009). Cone characteristics and production patterns were seen to differ substantially among genera. However, this was also the case within some genera. Conical phytoliths from Bulbostylis spp. and Rhynchospora spp. displayed the least amount of visible variation between species. Bulbostylis produced cones with satellites, rounded bases, pointed apices and psilate decoration whereas Rhynchospora phytoliths were similar, but had more abundant, smaller satellites covering the entire platelet (Fig. 3f). S. eggersiana was the only species to produce cones with a sinuous base. Although this cone feature is not exclusive to Scleria (Ollendorf, 1992), further investigation of sediments from the French Guiana coast may ﬁnd it to be of diagnostic signiﬁcance. One siliciﬁed stomatal complex was found in S. eggersiana that conformed to the type described by Carnelli et al. (2004) as J. Watling, J. Iriarte / Quaternary International 287 (2013) 162e180 distinctive of the family. These have well-deﬁned elongated, domeshaped subsidiary cells that lie parallel to the stomatal pore (Carnelli et al., 2004, 51; Guo et al., 2011, Fig. 4g.1). The achene pericarp of certain genera of Cyperaceae produce phytoliths diagnostic to genus level (Piperno, 1989, 2006; Honaine et al., 2009) including Cyperus/Kyllinga, Scirpus and Rhynchospora and these diagnostic bodies are commonly used in palaeoenviromental studies (e.g., Lu et al., 2006; Iriarte et al., 2010). Four- to eight-sided bodies with densely stippled surfaces and a central protuberance were found in C. articulatus (Denton, 1983; Piperno, 2006; Honaine et al., 2009), while Kyllinga brevifolia contributed densely stippled polygonal cones without straight walls or, in most cases, a visible central protuberance (Fig. 3geh). As there can be very little that differentiates this morphotype between genera (Piperno, 1989), it is often sensible to group them into a Cyperus/Kyllinga-type. Remirea maritima (syn. Cyperus pendunculatus, Institute of Systematic Botany, 2011) is the dominant sedge species found on French Guianan beaches and the achenes were found to produce Cyperus-type polygonal cones. While phytoliths from R. maritima and those from C. articulatus and C. gigantea shared the same key characteristics, it was noted that those from the former species were consistently larger than those of the two Cyperus spp. and displayed fewer, more elongated stipples, while also having a more obvious central protrusion (Fig. 3i). Without a larger regional study of inter-generic differences, these observations remain only tentative, so the inclusion of Remirea polygonal cones into a Cyperus/ Remirea/Kyllinga-type is proposed. The necessary addition of Remirea polygonal cones to the Cyperus/Kyllinga group reﬂects the ongoing debates among botanists concerning taxonomic classiﬁcations within the Cyperaceae (e.g. Goetghebeur, 1989). Scirpus asper contributed four- to seven- sided polygonal cones with deﬁned “buttresses” along the outer wall and a large, welldeﬁned protuberance (Fig. 3j). This is similar to other Scirpus spp. and a distinguishing characteristic for the genus (Schuyler, 1971; Piperno, 1989). Polygonal cones were also isolated from Rhynchospora gigantea, commonly in the form of four- or ﬁve-sided plates with sinuous edges, a central protuberance and very few stipples (Fig. 3k) (Piperno, 1989). None of the other genera tested, including Bulbostylis, Eleocharis, Fuirena and Scleria, were found to produce diagnostic achene bodies, conﬁrming previous studies (Piperno, 1989; Iriarte and Alonso-Paz, 2009). 3.2.1.3. Heliconiaceae. Highly decorated bodies with troughs were recorded from all parts of the H. psittacorum specimen (Fig. 3l). These phytoliths are produced in the sub-epidermis over bundlesheath ﬁbers and are diagnostic to genus level (Tomlinson, 1961; Prychid et al., 2004; Piperno, 2006). Where these phytoliths are present in soil samples amongst savanna vegetation, it would be reasonable to assume that they originate from H. psittacorum which is the only species of Heliconia that grows in the open savannas in French Guiana (Granville, 1986). 3.2.1.4. Marantaceae. Analysis of the leaf, stem and inﬂorescence of Ischnosiphon arouma yielded an abundant array of phytoliths, primarily different types of globular morphotypes diagnostic to the family level only. These came under the categories of irregularly angled/folded (Fig. 3m) (after Piperno, 1985), nodular (Fig. 3n) and granulate. The nodular morphotypes displayed the greatest size range, from less than 7 mm to up to 20 mm in diameter. The spheres with irregularly angled/folded surfaces all had diameters between 12 and 18 mm; whilst the granulate spheres were consistently between 5 and 8 mm in diameter. Conical-shaped bodies with nodular surface decoration originate from the inﬂorescence bracts of the majority of Marantaceae species (Fig. 3o) (Piperno, 2006). 173 3.2.1.5. Strelitziaceae. P. guianensis contains an abundance of druselike bodies which are globular, spiky and irregularly folded (Fig. 3p). These phytoliths are produced in all organs except the root and are formed in thin-walled silica cells adjacent to vascular bundles (Tomlinson, 1969; Prychid et al., 2004). Although similar to the irregularly angled/folded spheres of Marantaceae, the edges of the P. guianensis bodies are straighter and more acute than the rounded projections of the Marantaceae. These bodies will allow distinction to at least the family level. Small globular phytoliths with echinate decoration are also produced in P. guianensis. They are commonly no larger than 5 mm in diameter so a SEM is required to study their particular morphological characteristics, however they can be easily mistaken for the globular echinate bodies of Arecaceae from a light microscope. Globular granulate bodies of similarly small proportions were isolated from the species. 3.2.2. Poaceae The morphological phytolith classiﬁcations for the grasses selected in this study follow a system ﬁrst proposed by Twiss et al. (1969) and modiﬁed in later years to include aspects of threedimensional morphology (e.g. Brown, 1984; Fredlund and Tieszen, 1994; Piperno and Pearsall, 1998; Iriarte, 2003). In simplest terms, it has been observed that Panicoideae grasses produce lobate phytolith forms (bilobates and crosses), Chloridoideae grasses saddles or “battle-axe” forms, and Pooideae a large number of rondels and wavy trapezoids. However, many exceptions have been found, and those relevant to the French Guiana ﬂora are described below. Of the 16 species of Poaceae analysed, two produced bulliform phytoliths (Fig. 4a) and ten produced prickle-type hairs (Fig. 4b) diagnostic to the Poaceae family (Table 1). 3.2.2.1. Arundinoideae. As expected, the phytoliths isolated from the two species of Arundinoideae grass were found to display overlap with types typical of the subfamilies Panicoideae and Chloridoideae. The bilobates produced by Aristida tincta were diagnostic of the genus and have long, thin shafts and ﬂared, convex edges (Mulholland, 1989; Fredlund and Tieszen, 1994) (Fig. 4c). “Rondeloid/saddleoid” phytoliths exhibit characteristics of rondels in longitudinal view and of saddles in planar view (Fig. 4d). These morphotypes are also thought to be characteristic of the genus Aristida (Piperno and Pearsall, 1998). The phytolith assemblage from Gynerium sagittatum was characterised by elliptoid/biloboid forms that resemble rondels from side view but narrow bilobates in planar view (Piperno and Pearsall, 1998; Lu and Liu, 2003a) (Fig. 4e). Gynerium spp. grow in seasonally ﬂooded areas, though the same phytoliths are also produced by the bambusoid tribes Guaduinae and Chusqueinae, which grow in terra ﬁrme forest (Piperno and Pearsall, 1998). 3.2.2.2. Bambusoideae. One species of the bamboo subfamily was tested, P. luetzelbergii, which belongs to the subtribe Olyreae. This species produced a large number of “irregular, complex” shortcells that are characteristic of the subfamily (Piperno and Pearsall, 1998). These take the form of polylobates in planar view but are far less regular in appearance and trapezoidal in side view (Fig. 4f[1]). Crosses and bilobates with variant 3 and 8 attributes (i.e. having a nodular or blocky appearance) (Piperno, 2006) were also isolated from this species (Fig. 4f[2]). Many of the bilobates have fully developed conical projections along their length (variant 3) and resemble collapsed saddles from certain angles, illustrating the importance of rotating phytoliths during analysis to ascertain their 174 J. Watling, J. Iriarte / Quaternary International 287 (2013) 162e180 Fig. 4. Phytoliths from Poaceae. a) Bulliform cell from Axonopus aureus, b) Prickle-type hair from Aristida tincta, c) Aristida-type bilobate from Aristida tincta d) Rondeloid/saddleoid phytolith from A. tincta, e) Elliptoid/biloboid phytolith from Gynerium sagittatum, f) Irregular/complex phytoliths (1) from Parodiolyra luetzelbergii. Bilobate (2) to the right is variant 3/8, g) Rondeloid/saddleoid phytolith from Spartina alterniﬂora, h) Spooled/horned towers from Distichlis spicata, i) Tall saddles from Isachne polygonoides, j) Saddleoid/biloboid phytoliths from Isachne polygonoides, k) Bilobate from Axonopus aureus, l) Squat bilobate from Panicum cyanescens, m) Polybate from Paspalum parviﬂorum, n) Cross from P. parviﬂorum, o) Spooled/horned towers from Trachypogon spicatus, p) Fan-shaped rondel from Echinolaena inﬂexa. Scales ¼ 20 mm, except Fig. 4a where scale ¼ 50 mm). three-dimensional structure. These phytoliths are also diagnostic to sub-family level. 3.2.2.3. Chloridoideae. The classic saddle phytoliths of the Chloridoideae were notably lacking in the species tested in this study. While Spartina alterniﬂora produced saddles, rondels with spools were observed in greater abundance, as were phytoliths which display characteristics of the two. The morphotypes with rondeloid/saddleoid characteristics have been placed into this category. However, contrary to Aristida spp. rondeloid/saddeloid types, the planar face of S. alterniﬂora phytoliths was circular/oval, not saddle-shaped (Fig. 4g). Lu and Liu (2003a) have proposed a “rondel/saddle ellipsoid” category to describe the phytoliths of S. alterniﬂora. Distichlis spicata is a species adapted to brackish-water marshes. No saddles were observed in this species, though they have apparently been observed rarely by other researchers (Lu and Liu, 2003a,b). Again, abundant quantities of rondels were recorded which were rather small but elongated (as in S. alterniﬂora) but most had spools and an exaggerated, a-circular “rondel” face (Fig. 4h). These phytoliths resemble the chionochloid phytolith class (“towers”) of Kondo et al. (1994), and have been grouped with these morphotypes into a “spool/horned tower” type by Lu and Liu (2003a,b) and in the present study. 3.2.2.4. Micrairoideae. Isachne polygonoides yielded abundant quantities of tall saddles, as well as thick, blocky, irregular phytoliths which had both bilobate and saddle tendencies (Fig. 4i and j). Phytoliths from Isachne were ﬁrst described by Piperno and Pearsall (1998) who reported similar biloboid/saddeloid morphotypes in two other primitive Poaceae subfamilies, Arundinaceae and Bambusoideae. Isachne has since been re-classiﬁed into a reinstated J. Watling, J. Iriarte / Quaternary International 287 (2013) 162e180 Micrairoideae subfamily (Sánchez-Ken et al., 2007), and recent phylogenetic studies have demonstrated a close relationship between the Micrairoideae and the Arundinoideae (Edwards, 2011). Further study into phytolith production in more native Micrairoideae and Arundinoideae species in French Guiana are needed to discern the diagnostic potential of this morphotype in the region, however Lu and Liu (2003a,b) found biloboid/saddeloid phytoliths to be exclusive to the Isachne in their study of 250 grasses from China and the southeastern United States. 3.2.2.5. Panicoideae. The overwhelming majority of Poaceae that are native to coastal French Guiana belong to the subfamily Panicoideae. The Panicoideae are adapted to hot and humid conditions and are distributed widely in tropical climates (Twiss, 1992); in French Guiana they occur as components of freshwater marshes, marsh savanna and dry savanna (Paspalum spp., Schizachyrium spp., A. aureus) (Hooke, 1971; Granville, 1986). All species from the subfamily, apart from I. polygonoides, produced bilobates in various quantities (Fig. 4k) and it was noted that P. cyanescens and Schizachyrum riedelii both produce an unusual number of squat bilobates (Piperno and Pearsall, 1998) whose central shaft is almost absent (Fig. 4l). A detailed morphological assessment of lobate phytoliths like that done by Lu and Liu (2003b) was outside the scope of this paper. However, they also encountered squat bilobates in four Chinese Panicum species, yet the opposite (i.e. long central shafts) in Schizachyrum brevifolium, conﬁrming the variability of bilobate production among genera. Whilst bilobates are not exclusive to the Panicoideae (e.g. Brown, 1984; Mulholland, 1989; Piperno and Pearsall, 1998; Honaine et al., 2006), they can be distinguished from those produced by, for example, bambusoid and arundindoid grasses (Fredlund and Tieszen, 1994; Piperno and Pearsall, 1998). Polylobates, which are bilobates with one or more kinks in the shaft, were found to be common in several members of the Panicoideae (Fig. 4m). They were the most abundant morphotype in both Paspalum species tested (Paspalum parviﬂorum and Paspalum pulchellum) and in Echinchloa polystachya, far outnumbering bilobate forms (also observed by Lu and Liu, 2003b). Bodies with three to four lobes and indentations (crosses, Fig. 4n) were observed in all panicoid species apart from P. cyanescens and Trachypogon spicatus, a morphological pattern previously observed for other members of these genera (Piperno and Pearsall, 1998). Rondels, which are circular to oval phytoliths mostly associated with the grass subfamily Pooideae (Mulholland, 1989), are also found in other subfamilies including the Arundinoideae, Bambusoideae, Chloridoideae and Panicoideae. Spool/horned towers were encountered in T. spicatus, which were trapezoidal in side view, had a ﬂared circular base and a narrow top (Fig. 4o). While proportions differ between these and the spooled rondels of D. spicata (Fig. 4h), here they are grouped together following Lu and Liu (2003a,b). The dominant phytolith morphotypes from Echinolaena inﬂexa were rondels (trapeziforms c.f. Palmer et al., 1985) (Fig. 4p) (Piperno and Pearsall, 1998). These were noticeably large (frequently longer than 20 mm), and had curved, sinuate edges so as to appear fan-shaped in side view. When rotated, the narrowest face of the rondel is ovalshaped in planar view. This morphotype is believed to be diagnostic at least to genus level in French Guianan assemblages. 3.2.3. Dicotyledons 3.2.3.1. Annonaceae. Two of the four species of Annonaceae tested yielded large, irregular, multi-faceted bodies from the leaves that are diagnostic of the family (Piperno, 1988). Unonopsis stipitata produced spherical to aspherical forms ranging from 20 to 40 mm in diameter (Fig. 5a) whereas G. guianensis bodies were irregular, elongated and larger (40e90 mm) (Fig. 5b). These phytoliths are 175 thought to develop in contact with cell walls inside mesophyll cells or as intercellular spaces (Piperno, 1988; Runge, 1999). They were absent in A. glabra and Xylopia frutescens specimens tested. 3.2.3.2. Asteraceae. A number of hair bases with aspherical centres were isolated from M. camphoratum which were substantially larger than any others encountered in this study (Fig. 5c). Similar large hair bases were isolated from Asteraceae (Aspilla spp.) by Mercader et al. (2009, Fig. 4a). This family also produces hair bases consisting of two half-spheres joined together (Piperno, 1988), but this type was not encountered in this species. M. camphoratum also produced blocky polygonal phytoliths (Fig. 5d). These appear as irregular, ridged blocks with psilate surfaces. Such phytoliths occur in the epidermis of only a few dicotyledon families (Mercader et al., 2009). No trichomes or opaque perforated platelets diagnostic of the family were present in this species (e.g. Bozarth, 1992). 3.2.3.3. Burseraceae. The fruit of Protium guianense yielded small quantities of pentagonal bodies which originate in the fruit epidermis and are diagnostic of the genus Protium (Piperno, 1989) (Fig. 5e). These phytoliths have sinuous edges, a stippled surface and, commonly, a central hemispherical protuberance. Other distinctive phytoliths were isolated from the leaves of P. guianense (Fig. 5feh). They appear as what are informally termed “boney” bodies: elongated cylindrical bodies with psilate surface and verrucate/nodular decoration. The phytoliths have elaborated terminals that almost always consist of a bunch of irregular protrusions on one end and one ﬂatter end that is greater in diameter than the cylindrical stem to which it is attached to. These bodies are up to 90 mm long. The anatomical origin of this phytolith is unknown, but based on its observed position within the tissue, it appears to come from the leaf epidermis. This phytolith form was not encountered in a survey of the available literature, indicating that this is an undescribed morphotype. In the samples, it occurred in abundance, and this fact combined with it not having been described in previous studies of the genus Protium (e.g. Piperno, 1989), suggests that it is diagnostic at the species level. 3.2.3.4. Cannabaceae. The Cannabaceae, alongside the Moraceae, are producers of cystoliths. They are present in abundance in both Celtis iguanea and Ampelocera edentula, usually attached to large hair bases surrounded by substantial polyhedral epidermal complexes (Fig. 5i). Like those produced by the Moraceae, they took on spherical to aspherical forms and had rough, often tuberculate surface sculpting. Cystoliths from A. edentula are only diagnostic to the Cannabaceae/Moraceae families. However, C. iguanea cystoliths are much larger (up to 40 mm) and commonly adjoined to the stalk, a feature diagnostic of the genus in contrast to other members of these families (Honaine et al., 2005) (Fig. 5j). The seeds of C. iguanea were found to produce the densely stippled epidermal plates also characteristic of the genus (Bozarth, 1992; Piperno, 2006; Iriarte and Alonso-Paz, 2009). Whilst all measuring about the same size, these phytoliths take irregular shapes and sometimes appear more circular than elongated. All have well-deﬁned cell walls (Fig. 5k). 3.2.3.5. Dilleniaceae. Phytolith production varied signiﬁcantly among the three species of Dilleniaceae that were sampled. Only C. americana yielded the trichomes diagnostic of the family, which appear as uni-cellular, short, “deltoid-shaped” phytoliths with pointy tips and wide bases (Fig. 5l) (Piperno, 2006). Small numbers of the diagnostic multi-cellular hairbases were found in C. americana (Piperno, 2006) (Fig. 5m). 176 J. Watling, J. Iriarte / Quaternary International 287 (2013) 162e180 Fig. 5. Phytoliths from dicotyledons. a) Spherical multi-faceted phytolith from Unonpsis stipitata, b) Irregular multi-faceted phytolith from Guatteria guianensis leaf, c) Hairbase from Melampodium camphoratum, d) Blocky polyhedrals from M. camphoratum, e) Stippled body from Protium guianense fruit, feh) “Boney” bodies from P. guianense leaf, i) Hair base and polyhedral epidermal complex from C. iguanea leaf, j) Cystolith from Celtis iguanea, k) Stippled epidermis from C. iguanea seeds, l) Trichomes from Curatella americana, m) Hair base from C. americana, n) Cystolith from Sorocea muriculata, o) Trichomes from Ficus guianensis, p) Squat trichome from F. guianensis, q) Hair base from F. guianensis leaf. Scales ¼ 20 mm. The absence of hair or hair base phytoliths in D. alata shows some degree of intra-family variability (Piperno, 2006). 3.2.3.6. Moraceae. Cystoliths were isolated from both Moraceae species tested; Ficus guianensis and Sorocea muriculata. These originate in the leaf epidermis as wall outgrowths of lithocyst cells and are commonly highly siliciﬁed (Honaine et al., 2005). The typical spherical to aspherical, irregularly-shaped, roughly decorated cystoliths diagnostic of the Moraceae/Cannabaceae were found in both species (Fig. 5n). Most measured around 12 mm in diameter, but those from F. guianensis were sometimes twice as big. J. Watling, J. Iriarte / Quaternary International 287 (2013) 162e180 177 Fig. 6. Phytoliths from Pteridophytes. a) Epidermal phytolith from Thelypteris conﬂuens, b) Elongated undulating body from Pityogramma calomelanos, ced) Pitted bodies from P. calomelanos, e) Bowl-shaped phytoliths from Trichomanes hostmannium, feg) Aspherical granulate bodies from T. conﬂuens. Scales ¼ 20 mm. The Moraceae are also common producers of siliciﬁed trichomes (Metcalfe and Chalk, 1983; Kealhofer and Piperno, 1998; Wallis, 2003). Both species have elongated, conical, unicellular bodies with psilate surfaces that varied considerably in length (20e90 mm) (Fig. 5o). F. guianensis also yielded small quantities of squat conical trichomes that have broad circular bases up to 15 mm in diameter and 20 mm in height (Fig. 5p). This phytolith type is to date unique to the families Moraceae and Cannabaceae (Wallis, 2003). Unlike Wallis (2003), no tuberculate decoration around the base of these phytoliths was discerned. Isolated hair bases from F. guianensis are diagnostic of the genus Ficus. These are typically circular to a-circular and display regular striations that emanate from inside of the base to the periphery of the cell (Piperno, 2006) (Fig. 5q). 3.2.4.3. Thelypteridaceae. The phytoliths isolated from T. conﬂuens were aspherical, had highly granular surfaces akin to those of cystoliths and measured up to 50 mm in length (Fig. 6feg). This type of phytolith is unique, and to the authors’ knowledge has not been described before. Other genera within the Thelypteridaceae were tested previously for phytoliths by Mazumdar (2010, 2011) and were seen to produce various forms of epidermal plates, but this is the ﬁrst time the genus Thelypteris has been analysed. It is possible that the phytoliths from T. conﬂuens are even more highly folded versions of those isolated from Cyclosorus interruptus (Mazumdar, 2010, Fig. 1K), but phytolith production is still poorly understood in this family. Interestingly, this morphotype was not encountered in T. interrupta, suggesting possible species speciﬁcity. 3.2.4. Pteridophytes Typical anticlinal fern epidermal forms were observed in Thelypteris conﬂuens (Thelypteridaceae) and B. serrulatum (Blechnaceae) which are distinguished by their well-deﬁned undulating ridges (Piperno, 2006; Iriarte and Alonso-Paz, 2009) (Fig. 6a). 4. Discussion and conclusions 3.2.4.1. Adiantaceae. Two phytolith morphotypes were encountered from the leaves of Pityrogramma calomelanos (Adiantaceae). The ﬁrst were elongated bodies with two undulating ridges running parallel to each other (Fig. 6b). These phytoliths were described by Piperno (2006) as diagnostic of the family Polypodiaceae, to which P. calomelanos belonged until it was reclassiﬁed into the Adiantaceae. They are up to 100 mm long in P. calomelanos, but they can be up to 1000 mm in length (Piperno, 2006; Sundue, 2009; Mazumdar, 2011). Also encountered were large elongated bodies with opposing rounded projections and heavily pitted decoration. The phytolith type in Fig. 6c was found in the second specimen tested, and appears to be a more or less-digested version of the pitted phytolith in Fig. 6d. These have been observed in modern fernland soils in New Zealand (Kondo et al., 1994, Plate 27; Carter, 2002, Fig. 3G). 3.2.4.2. Hymenophyllaceae. The leaves of the tree-fern Trichomanes hostmannium were found to produce shallow, roughly bowl-shaped phytoliths diagnostic of this family (Piperno, 2006) (Fig. 6e). The aim of this study was to assess phytolith production in ecologically and economically signiﬁcant plants native to coastal French Guiana in order to provide a reference base for palaeoecological and archaeological work being conducted in the region. Of the 108 species analysed, over 80% were from previously untested species, genera or families. Plant species with phytoliths diagnostic to the family level or below accounted for 49% of the total sample, and 40% of non-Poaceae species. Phytolith production patterns observed in this study conform to those reported from other regions of the world (Piperno, 2006). The Poaceae species tested produced large numbers of speciﬁc phytoliths diagnostic to below family level, and all of the non-Poaceae monocots tested, apart from M. arborescens (Araceae), produced forms diagnostic to at least family level. There was little diagnostic phytolith production amongst the woody dicots analysed, with only 9 out of 46 species displaying diagnostic morphotypes. The majority of phytoliths produced by the woody dicot species are only attributable to “dicot” level (e.g. MFBs, sclereids). The herbaceous dicots were the group producing the lowest numbers of phytoliths and did not yield any forms diagnostic to family level or below; most produced very low numbers of phytoliths with limited taxonomic value (e.g. globular phytoliths). Phytolith production was found to be highly variable among the Pteridophyta included in the 178 J. Watling, J. Iriarte / Quaternary International 287 (2013) 162e180 study, but promising nonetheless. Three of the 13 fern species yielded phytoliths diagnostic to family level or below, and one (B. serrulatum) to the Pteridophyta division. Two previously undescribed phytolith morphotypes were isolated: one from the fern T. conﬂuens and the other from leaves of Protium guinanense. The latter is of particular signiﬁcance due to its unusual morphology and its absence in previous studies of the genus (Piperno, 1985, 1988, 2006) and both discoveries help to further enlarge the growing body of data for phytolith morphologies from neotropical plants. Analysis conﬁrmed the presence of important diagnostic morphotypes such as the druse bodies produced by P. guianensis (Strelitziaceae) and hair base and epidermal phytoliths from M. camphoratum (Asteraceae). In addition, this study has highlighted several issues of further inquiry, including: i) genus-level size and morphological differences in palm phytoliths, ii) hair base siliciﬁcation in the Chrysobalanaceae and iii) phytolith production in ferns. It is hoped that these issues will be explored in further investigations of phytoliths from the Neotropics. Results from this study have also brought to light important exceptions to the tripartite short-cell classiﬁcation system (Twiss et al., 1969) for ecologically-signiﬁcant Poaceae species in the Neotropics. Saddle phytoliths usually thought typical of chloridoid grasses, which in French Guiana grow in brackish-water marshes, were more commonly produced by two marsh savanna species A. tincta (Arundinoideae) and I. polygonoides (Panicoideae). Rather than producing abundant saddle phytoliths, the chloridoid species that were tested (D. spicata, S. alterniﬂora) instead yielded large quantities of rondeloid/saddeloid phytoliths and spooled rondels. While only 16 Poaceae species were included in the reference collection, the ones chosen for this study include those that are among the most environmentally speciﬁc in the region (Hooke, 1971; Granville, 1986). Had they remained untested, these kinds of important deviations would remain unknown and phytolith assemblages could be incorrectly interpreted. Assessing the usefulness of phytoliths for distinguishing vegetation types is made easier in this case due to having only selected plant species that occupy speciﬁc ecological niches. Of the nine major vegetation formations, all but two (beach and mangrove forest) were found to contain plants that produce phytoliths diagnostic to family, genus or species level. Terra ﬁrme forest is possible to distinguish by phytoliths diagnostic to Cannabaceae, Burseraceae, Marantaceae, Strelitziaceae and Bambusoideae species alongside a preponderance of typically dicotyledon phytoliths (e.g. globular granulate/psilate, MFBs, tracheary elements, vesicular inﬁllings). In the same way, swamp (permanently inundated) forest includes species of Annonaceae, as well as Ficus spp. and Thelypteris spp. that produce diagnostic morphotypes. Marsh forest indicators include Arecaceae and the ferns Trichomanes spp. and Pityrogramma spp. Freshwater marsh environments were identiﬁed in lower levels of raised ﬁeld sediments in French Guiana (Iriarte et al., 2010, 2012) and were characterised by high numbers of Cyperaceae and Marantaceae phytoliths, while greater quantities of Panicoideae grass morphotypes characterise marsh savannas, of which the genera Aristida and Echinolaena are speciﬁc components. Dry savannas would be more difﬁcult to distinguish in individual samples in the absence of diagnostic phytoliths from the Dilleniaceae, however the abundance of polylobates from Paspalum spp. may be worth investigating as an additional indicator. The next step will be to test these associations of particular taxa and ecosystems through the analysis of modern soil phytoliths, which will be the subject of a future study. Due to the variability of phytolith production patterns among species, it will invariably follow that some taxa will be over- or under-represented in a given fossil or soil phytolith assemblage. In palaeoenvironmental studies, combining phytolith and pollen analysis can often ﬁll the gaps left by major groups of plants, such as herbaceaous dicotyledons, and is key to more detailed reconstructions of certain vegetation types. In a recent study by the authors, phytoliths and pollen were used in tandem to ascertain the history and environmental legacy of the French Guianan raised ﬁelds. The decline over time of Cyperaceae and Marantaceae phytoliths and pollen and the increase of Poaceae is interpreted as the expansion of terra ﬁrme (non-ﬂooded) environments as a result of pre-Columbian mound construction in the seasonally ﬂooded savannas (Iriarte et al., 2012). Similarly, differences in ubiquity of phytolith production among diagnostic species must be considered; for instance, palm phytoliths are produced abundantly in all parts of the plant, whereas Celtis sp. stippled bodies are produced only in the seed. This means that, while measuring relative phytolith frequencies is an effective method to identify and compare vegetation dynamics over time, it is clear that an approach which takes into account the presence, absence and combinations of plant taxa should also be employed. In summary, this reference collection represents the ﬁrst systematic attempt to categorise phytoliths produced by plants native to the seasonally-ﬂooded savanna-forest ecosystems of coastal French Guiana. Its compilation has demonstrated the importance of comprehensive, regionally-speciﬁc comparative collections in order to maximise the accuracy of phytolith studies, not only for the study region, but for comparable Neotropical ecosystems. This analysis adds to the usefulness of phytoliths and proves that they can make signiﬁcant contributions to palaeoecological and archaeological debates in lowland South America. Acknowledgements This research was funded by the “Amazonie” interdisciplinary programme of INEE (National Institute of Ecology and Environment), CNRS, France. The authors would like to thank the following collaborators in Project Amazonie: Doyle McKey, Stéphen Rostain, Bruno Glaser, Jago Birk. Special thanks also goes to Doyle McKey and Ruth Dickau for their suggestions and comments on the manuscript, and to staff at the IRD Herbarium in Cayenne, in particular Marie-François Prévost, and Sara Edwards and others from Royal Botanical Gardens, Kew for helping with the collection of plant specimens. Thanks also goes to Dolores Piperno and Caroline Strömberg for their opinions on some of the phytoliths encountered during analysis and to Adam Wainwright of University of Exeter who prepared many of the slides used in analysis, and Mike Rouillard, also of University of Exeter, who prepared phytolith slides and the ﬁgures for the paper. References Alexandre, A., Meunier, F.C., Koud, J.-M., 1997. Phytoliths: indicators of grassland dynamics during the Late Holocene in intertropical Africa. Palaeography, Palaeoclimatology, Palaeoecology 136, 213e239. Alexandre, A., Meuneir, J.-D., Mariottie, A., Soubies, F., 1999. Late Holocene phytoliths and carbon-isotope record from a latosol at Salitre, South-Central Brazil. Quaternary Research 51, 187e194. Amos, G.L., 1952. Silica in Timbers. Bulletin No. 267. Commonwealth Scientiﬁc and Industrial Research Organization, Melbourne. Barrett, J., 2001. Atlas Illustré de la Guyane. IRD Editions, Amsterdam. Beard, J.S., 1953. The savanna vegetation of northern tropical America. Ecological Monographs 23, 149e215. Bertoldi de Pomar, H., 1971. Ensayo de clasiﬁcación morfológica de los silicoﬁtilitos. Ameghiana 8, 317e328. Blinnikov, M.A., Busacca, A., Whitlock, C., 2002. Reconstruction of the Late Pleistocene grassland of the Columbia Basin, Washington, USA, based on phytolith records in loess. Palaeogeography, Palaeoclimatology, Palaeoecology 177, 77e101. Bozarth, S., 1992. Classiﬁcation of opal phytoliths formed in selected dicotyledons native to the Great Plains. In: Rapp Jr., G., Mulholland, S.C. (Eds.), Phytolith Systematics: Emerging Issues. Plenum Press, New York, pp. 193e214. J. Watling, J. Iriarte / Quaternary International 287 (2013) 162e180 Brown, D., 1984. Prospects and limits of a phytolith key for grasses in the Central United States. Journal of Archaeological Science 11, 211e243. Campos de, A.C., Labouriau, L.G., 1969. Corpos diliceos de gramineas dos Cerrados-II. Pesquisas Agropecuarias Brasileiras 4, 143e151. Carnelli, A.L., Madella, M., Theurillat, J.-P., 2001. Biogenic silica production in selected Alpine plant species and plant communities. Annals of Botany 87, 425e434. Carnelli, A.L., Theurillat, J.-P., Madella, M., 2004. Phytolith-types and typefrequencies in subalpine-alpine plant species of the European Alps. Review of Palaeobotany and Palynology 129, 39e65. Carter, J.A., 2002. Phytolith analysis and paleoenvironmental reconstruction from Lake Poukawa Core, Hawkes Bay, New Zealand. Global and Planetary Change 33, 257e267. Chandler-Ezell, K., Pearsall, D.M., Zeidler, J.A., 2006. Root and tuber phytoliths and starch grains document manioc (Manihot escuelenta), arrowroot (Maranta arundinacea) and Llerén (Calathea sp.) at the Real Alto site, Ecuador. Economic Botany 60, 103e120. Da Silva, J.M., Bates, J.M., 2002. Biogeographic patterns and conservation in the South American cerrado: a tropical savanna hotspot. Bioscience 52, 225e234. Denham, T.P., Haberle, S.G., Lentfer, C.J., Fullagar, R., Field, J., Therin, M., Porch, N., Winsborough, B., 2003. Origins of agriculture at Kuk Swamp in the Highlands of New Guinea. Science 301, 189e193. Denton, M.F., 1983. Anatomical studies of the Luzulae group of Cyperus (Cyperaceae). Systematic Botany 8, 250e262. Edwards, E., 2011. New grass phylogeny resolves deep evolutionary relationships and discovers C4 origins. New Phytologist 193, 304e312. Fenwick, R.S.H., Lentfer, C.J., Wisler, M.I., 2011. Palm reading: a pilot study to discriminate phytoliths of four Arecaceae (Palmae) taxa. Journal of Archaeological Science 38, 2190e2199. Fredlund, G.G., Tieszen, L.T., 1994. Modern phytolith assemblages from the North American Great Plains. Journal of Biogeography 21, 321e335. Fredlund, G.G., Tieszen, L.T., 1997. Calibrating grass phytolith assemblages in climatic terms: application to Late Pleistocene assemblages from Lansas and Nebraska. Palaeogeography, Palaeoclimatology, Palaeoecology 136, 199e211. Geis, J.W., 1973. Biogenic silica in selected species of deciduous angiosperms. Soil Science 116, 113e130. Goetghebeur, P., 1989. Studies in Cyperaceae 9. Problems in the lectotypiﬁcation and infrageneric taxonomy of Cyperus L. Bulletin de la Societe Royale de Botanique de Belgique 122, 103e114. de Granville, J.J., 1986. Le littoral Guyanais, fragilité de l’environnement. Nature Guyanaise, Cayenne. Grenand, P., Moretti, C., Jacquemin, H., Prévost, M.F., 2004. Créoles, Wayãpi, Palikur. Pharmacopées traditionnelles en Guyane. IRD Editions, Paris, 624 pp. Guo, M., Jie, D., Liu, H., Luo, S., Li, N., 2011. Phytolith analysis of selected wetland plants from Chanbai mountain region and implications for palaeoenvironment. Quaternary International 250, 119e128. Honaine, M.F., Zucol, A.F., Osterrieth, M.L., 2005. Biomineralizaciones de sílice en Celtis tala (Celtidaceae). Boletin de la Sociedad argentina de Botánica 40, 229e239. Honaine, M.F., Zucol, A.F., Osterrieth, M.L., 2006. Phytolith assemblages and systematic associations in grassland species of the south-eastern pampean plains, Argentina. Annals of Botany 98, 1155e1165. Honaine, M.F., Zucol, A.F., Osterrieth, M.L., 2009. Phytolith analysis of Cyperaceae from the Pampean region, Argentina. Australian Journal of Botany 57, 512e523. Hooke, J., 1971. Les Savanes Guyanaises: Kourou. In: Memoire, vol. 44. ORSTOM, Paris. Institute of Systematic Botany, 2011. http://www.systbot.uzh.ch. Iriarte, J., 2003. Emergent cultural complexity in the wetlands of Uruguay during the Middle Holocene. Ph.D. thesis, University of Kentucky, USA. Iriarte, J., 2006. Vegetation and climate change since 14810 14C yr BP in southeastern Uruguay and implications for the rise of early Formative societies. Quaternary Research 65, 20e32. Iriarte, J., Alonso-Paz, E., 2009. Phytolith analysis of selected native plants and modern soils from southeastern Uruguay and its implications for palaeoenvironmental and archaeological reconstruction. Quaternary International 193, 99e123. Iriarte, J., Glaser, B., Watling, J., Wainwright, A., Birks, J., Renard, D., Rostain, S., McKey, D., 2010. Agricultural landscapes of coastal Amazonia: phytolith and carbon isotope analysis of raised ﬁelds from French Guiana savannas. Journal of Archaeological Science 37, 2984e2994. Iriarte, J., Power, M.J., Rostain, S., Mayle, F.E., Jones, H., Watling, J., Whitney, B.S., McKey, D.B., 2012. Fire-free land use in pre-1492 Amazonian savannas. Proceedings of the National Academy of Sciences of the United States of America 109, 6473e6478. Judziewicz, E.J., Soreng, R.J., Davidse, G., Peterson, P.M., Filgueiras, T.S., Zuloaga, F.O., 2000. Catalogue of new world grasses (Poaceae): I. Subfamilies Anomochlooideae, Bambusoideae, Ehrhartoideae, and Pharoideae. Contributions from the United States National Herbarium 39, 1e128. Kealhofer, L., Piperno, D.R., 1998. Opal phytoliths in Southeast Asian ﬂora. Smithsonian Contributions to Botany 88. Kondo, R., Childs, C., Atkinson, I., 1994. Opal Phytoliths of New Zealand. Mannaki Press, Lincoln, New Zealand. Lentfer, C.J., 2003. Plants, people and landscapes om prehistoric Papua New Guinea: a compendium of phytoliths (and starch). Ph.D. thesis, Southern Cross University, Lismore, Australia. Lindeman, J.C., 1953. The Vegetation of Coastal Surinam. Domplein 2-Utrecht. Lu, H.-Y., Liu, K.-B., 2003a. Phytoliths of common grasses in the coastal environments of southeastern USA. Estuarine, Coastal and Shelf Science 58, 587e600. 179 Lu, H.-Y., Liu, K.-B., 2003b. Morphological variations of lobate phytoliths from grasses in China and southeastern United States. Diversity and Distribution 9, 73e87. Lu, H.-Y., Wu, N.-Q., Yang, X.-D., Jiang, H., Liu, K.-B., Liu, T.-S., 2006. Phytoliths as quantitative indicators for the reconstruction of past environmental conditions in China I: phytolith-based transfer functions. Quaternary Science Reviews 25, 945e959. Madella, M., Alexandre, A., Ball, T., 2005. International code for phytolith Nomenclature 1.0. Annals of Botany 96, 253e260. Mazumdar, J., 2010. Phytoliths in ferns II: in some more Thelypterpoid ferns. Bionature 30, 13e17. Mazumdar, J., 2011. Phytoliths of pteridophytes. South African Journal of Botany 77, 10e19. McKey, D., Rostain, S., Iriarte, J., Glaser, B., Birk, J., Holst, I., Renard, D., 2010. PreColumbian agricultural landscapes, ecosystem engineers and self-organized patchiness in Amazonia. Proceedings of the National Academy of Sciences of the United States of America 107, 7823e7828. McMichael, C.H., Bush, M.B., Piperno, D.R., Silman, M.R., Zimmerman, A.R., Anderson, C., 2012. Spatial and temporal scales of pre-Columbian disturbance associated with western Amazonian lakes. The Holocene 22, 131e141. Mehra, P.N., Sharma, O.P., 1965. Epidermal silica cells in the Cyperaceae. Botanical Gazette 126, 53e58. Mercader, J., Bennett, T., Esselmont, C., Simpson, S., Walde, D., 2009. Phytoliths in woody plants from the Miombo woodlands of Mozambique. Annals of Botany 104, 91e113. Metcalfe, C.R., 1960. Anatomy of the Monocotyledons: I. Graminae. Oxford University Press, London. Metcalfe, C.R., 1971. Anatomy of the Monocotyledons: V. Cyperaceae. Oxford University Press, London. Metcalfe, C.R., Chalk, L., 1979. Anatomy of the Dicotyledons, second ed., vol. 1. Oxford University Press, London. Metcalfe, C.R., Chalk, L., 1983. Anatomy of the Dicotyledons, second ed., vol. 2. Oxford University Press, London. Mulholland, S.C., 1989. Phytolith shape frequencies in North Dakota grasses: a comparison to general patterns. Journal of Archaeological Science 16, 489e511. Neff, H., Pearsall, D.M., Jones, J.G., Arroyo de Pieters, B., Freidel, D.E., 2006. Climate change and population history in the Paciﬁc lowlands of Mesoamerica. Quaternary Research 65, 390e400. Ollendorf, A.L., 1992. Towards a classiﬁcation scheme of sedge (Cyperaeae) phytoliths. In: Rapp Jr., G., Mulholland, S.C. (Eds.), Phytolith Systematics: Emerging Issues. Plenum Press, New York, pp. 91e111. Palmer, P.G., Gerbeth-Jones, S., Hutchison, S., 1985. A scanning electron microscopy survey of the epidermis of East African grasses, III. Smithsmonian Contributions to Botany 55. Passos, M.A.B., de Menconça, M.S., 2006. Epiderme dos segmentos foliares de Mauritia ﬂexuosa L. f. (Arecaceae) em três fases de desenvolvimento. Acta Amazonia 36, 431e436. Pearsall, D.M., 2000. Palaeoethnobotany: A Handbook of Procedures. Academic Press, San Diego. Peterson, P.M., Soreng, R.J., Davidse, G., Filgueiras, T.S., Zuloaga, F.O., Judziewicz, E.J., 2001. Catalogue of new world grasses (Poaceae): II. Subfamily Chloridoideae. Contributions from the United States National Herbarium 39, 1e255. Piperno, D.R., 1985. Phytolith analysis and tropical paleo-ecology: production and taxonomic signiﬁcance of siliceous forms in New World plant domesticates and wild species. Review of Palaeobotany and Palynology 45, 185e228. Piperno, D.R., 1988. Phytolith Analysis: An Archaeological and Geological Perspective. Academic Press, San Diego. Piperno, D.R., 1989. The occurrence of phytoliths in the reproductive structures of tropical angiosperms and their signiﬁcance in tropical palaeoecology, palaeoethnobotany and systematics. Review of Palaeobotany and Palynology 61, 147e173. Piperno, D.R., 2006. Phytoliths: A Comprehensive Guide for Archaeologists and Palaeoecologists. Altimira Press, Oxford. Piperno, D.R., Pearsall, D.M., 1998. The silica bodies of tropical American grasses: morphology, taxonomy and implications for grass systematics and fossil phytolith identiﬁcation. Smithsonian Contributions to Botany 85. Piperno, D.R., Moreno, J.E., Iriarte, J., Holst, I., Lachniet, M., Jones, J.G., Ranere, A.J., Castanzo, R., 2007. Late Pleistocene and Holocene environmental history of the Iguala Valley, Central Balsas Watershed of Mexico. Proceedings of the National Academy of Sciences of the United States of America 104, 11874e11881. Plaziat, J.-C., Augustinus, G.E.F., 2004. Evolution of progradation/erosion along the French Guiana mangrove coast: a comparison of mapped shorelines since the 18th century with Holocene data. Marine Geology 108, 127e143. Postek, M.T., 1981. The occurrence of silica in the leaves of Magnolia grandiﬂora L. Botanical Gazette 142, 11874e11881. Prost, M.T., 1989. Coastal dynamics and chenier sands in French Guiana. Marine Geology 90, 259e267. Prychid, C., Rudall, P.J., Gregory, M., 2004. Systematics and biology of silica bodies in monocotyledons. The Botanical Review 69, 377e440. Rostain, S., 2008. Agricultural earthworks on the French Guiana coast. In: Silverman, H., Isbell, W. (Eds.), Handbook of South American Archaeology. Springer/Kluwere/Plenum, New York, pp. 217e233. Rostain, S., 2010. Pre-Columbian earthworks in coastal Amazonia. Diversity 2, 331e352. 180 J. Watling, J. Iriarte / Quaternary International 287 (2013) 162e180 Runge, F., 1999. The opal phytolith inventory of soils in Central Africa: quantities, shapes, classiﬁcation and spectra. Review of Palaeobotany and Palynology 107, 23e53. Sánchez-Ken, J.G., Clark, L.G., Kellogg, E.A., Kay, E.E., 2007. Reinstatement and emendation of subfamily Micrairoideae (Poaceae). Systematic Botany 32, 71e80. Schuyler, A.E., 1971. Scanning Electron Microscopy of achene epidermis in species of Scirpus (Cyperaceae) and related genera. Proceedings of the Academy of Natural Sciences of Philadelphia 123, 29e52. Scurﬁeld, G., Anderson, C.A., Segnit, E.R., 1974. Silica in woody stems. Australian Journal of Botany 22, 211e229. Sendulksy, T., Labouriau, L.G., 1966. Corpos siliceos de Gramineas dos Cerrados-I. Annais da Academia Brasileira de Ciencias 38, 159e185. Sondahl, M.R., Labouriau, L.G., 1970. Corpos siliceos de Gramineas dos Cerrados IV. Pesquisas Agropecuarias Brasileiras 5, 183e207. Strömberg, C.A., 2003. The origin and spread of grass-dominated ecosystems in the Late Tertiary of North America: preliminary results concerning the evolution of hypsodonty. Ph.D. thesis, University of California, Berkeley, USA. Strömberg, C.A., 2004. Using phytolith assemblages to reconstruct the origin and spread of grass-dominated habitats in the Great Plains during the Late Eocene to Early Miocene. Palaeogeography, Palaeoclimatology, Palaeoecology 207, 239e275. Sundue, M., 2009. Silica bodies and their systematic implications in Pteridaceae (Pteridophyta). Botanical Journal of the Linnean Society 161, 422e435. Teixeira da Silva, S., Labouriau, L.G., 1970. Corpos siliceos de Gramineas dos Cerrados-III. Pesquisas Agropecuarias Brasileiras 5, 167e182. Tomlinson, P.B., 1961. Anatomy of the Monocotyledons: II. Palmae. Clarendon Press, Oxford. Tomlinson, P.B., 1969. Anatomy of the Monocotyledons: III. Commelinales-Zingiberales. Clarendon Press, Oxford. Twiss, P.C., 1992. Predicted world distribution of C3 and C4 grass phytoliths. In: Rapp Jr., G., Mulholland, S.C. (Eds.), Phytolith Systematics: Emerging Issues. Plenum Press, New York, pp. 113e128. Twiss, P.C., Suess, E., Smith, R.M., 1969. Morphological classiﬁcation of grass phytoliths. Soil Science Society of America Proceedings 33, 109e115. Wallis, L., 2003. An overview of leaf phytolith production patterns in selected northwest Australian ﬂora. Review of Palaeobotany and Palynology 125, 201e248. Zuloaga, F.O., Morrone, O., Davidse, G., Filgueiras, T.S., Peterson, P.M., Soreng, R.J., Judziewicz, E., 2003. Catalogue of new world grasses (Poaceae): III. Subfamilies Panicoideae, Aristidoideae, Arundinoideae, and Danthonioideae. Contributions from the United States National Herbarium 48, 1e730. Zurro, D., Madella, M., Briz, I., Vila, A., 2009. Variability of the phytolith record in ﬁsherehunteregatherer sites: an example from the Tamana society (Beagle Channel, Tierra del Fuego, Argentina). Quaternary International 193, 184e191.

RELATED PAPERS

RELATED TOPICS

Log In

Phytoliths from the coastal savannas of French Guiana

Phytoliths from the coastal savannas of French Guiana

Related Papers

RELATED PAPERS

RELATED TOPICS