Wood and bark phytoliths of West African woody plants

Quaternary International xxx (2016) 1e18

Contents lists available at ScienceDirect

Quaternary International journal homepage: www.elsevier.com/locate/quaint

Wood and bark phytoliths of West African woody plants Lucia Veronica Collura*, Katharina Neumann Goethe University, Institute of Archaeological Sciences, Dept. III: Pre- and Protohistory, Norbert-Wollheim-Platz 1, D-60629, Frankfurt am Main, Germany

a r t i c l e i n f o

a b s t r a c t

Article history: Available online xxx

Long-term wood anatomical research has shown that 10% of the world's trees and shrubs produce silica in their wood, but silica production in bark had never been systematically investigated. We present here the results of the first comprehensive study on phytoliths in bark and compare them with data on wood phytoliths. We studied 103 bark samples from 92 species and 35 wood samples from 31 species mainly distributed in the West African savannas, altogether representing 34 plant families. The presence of silica in >90% of the studied bark samples indicates that silica production in bark is much more common than in wood. We developed a classification with three anatomical and five morphological classes and recorded their abundance in the processed phytolith samples. With a few exceptions, the phytoliths of bark and wood belong to different classes and can be clearly distinguished from each other. Wood produces Globulars s.l. and Aggregates with their own shape independent from the cells in which they had been formed. In bark, unspecific Blockies and Silica particles/Accumulations are omnipresent while 31% of the species have specific morphotypes with consistent morphologies. These phytoliths reflect the anatomy of the cells and tissues and develop either through silification of the cell lumen or the cell walls. They belong to the anatomical classes Sclerenchyma (with two subclasses Fibres and Sclereids), Cork/Parenchyma, and Cork aerenchyma. Phytoliths in bark and wood have taxonomic relevance, but the distribution is uneven on different taxonomic levels. Some Urticalean Rosids, Bignoniaceae and Capparaceae develop diagnostic phytoliths in the bark. Wood and bark phytoliths can be identified in special archaeological and palaeoecological contexts, but because they are from vegetative tissues, redundancy with similar morphotypes from other plant organs and taxonomic groups has to be considered in mixed assemblages. © 2016 Elsevier Ltd and INQUA. All rights reserved.

Keywords: Wood Bark Phytoliths West Africa Silica Plant anatomy

1. Introduction When they decay, trees and shrubs produce phytoliths that are deposited in soils and sediments. Which morphotypes are representative for woody plants in forests, open woodlands and savannas and how the woody plant cover should be quantified is a matter of debate. In Africa, the density of the woody plant cover is usually assessed through calculation of the index D:P (dicotyledons ¼ globular decorated vs. Poaceae short cell phytoliths, Alexandre et al., 1997; Barboni et al., 1999, 2007; Bremond et al., 2005, 2008; Neumann et al., 2009). The overall validity of the D:P for indicating tree cover densities has been €mberg (2003, 2004). Besides globulars, questioned by Stro

* Corresponding author. E-mail addresses: [email protected] (L.V. Collura), k.neumann@em. uni-frankfurt.de (K. Neumann).

€mberg (2004) proposed silicified sclerenchyma from the leaves Stro of tropical trees, Marantaceae and palm phytoliths, and short cells from Bambusoid grasses as indicators for forest. Neumann et al. (2009) and Garnier et al. (2013) used sclereids as additional representatives of the woody plant cover. In a study on phytolith assemblages in modern soils of Central African forests, Runge (1999) interpreted globulars and irregularly shaped phytoliths as forest indicators. Globulars, however, are widespread in different plant taxa and tissues, especially in monocots, e.g. the Zingiberales (Chen and Smith, 2013). The question arises: Which morphotypes are typical and diagnostic for the trees and shrubs themselves? After their decay, dicotyledonous trees and shrubs deposit large amounts of phytoliths from leaves, wood and bark in the soil. Comparative studies on the phytoliths of modern woody dicots are essential for unambiguously assigning single morphotypes in an unknown assemblage to dicot leaves, wood and bark, and assessing redundancy between them and with other taxonomic groups. However, these studies are still rare, especially in Africa. Runge

http://dx.doi.org/10.1016/j.quaint.2015.12.070 1040-6182/© 2016 Elsevier Ltd and INQUA. All rights reserved.

Please cite this article in press as: Collura, L.V., Neumann, K., Wood and bark phytoliths of West African woody plants, Quaternary International (2016), http://dx.doi.org/10.1016/j.quaint.2015.12.070

2

L.V. Collura, K. Neumann / Quaternary International xxx (2016) 1e18

(1996, 2000) investigated the leaf phytoliths of East African woody and herbaceous dicots, and Mercader et al. (2009) presented data on 90 woody species from the Miombo woodlands in Mozambique. While leaf phytoliths can be characterized by the presence of trichomes, epidermis fragments and special sclereids (Kondo and Peason, 1981; Postek, 1981; Baas et al., 1982), there is still much confusion on wood and bark phytoliths. Because silica in wood can present problems in the handling of tropical timbers, phytoliths in wood have long since been recorded in the frame of wood anatomical research (e.g. Amos, 1952; Scurfield et al., 1974; Ter Welle, 1976a,b; Richter, 1980; Metcalfe and Chalk, 1983). The presence of silica is a useful diagnostic character in wood identification (InsideWood, 2004 onwards, http://insidewood.lib.ncsu.edu; Wheeler, 2011) and Ter Welle (1976a,b) distinguished between globular and aggregate silica bodies. In contrast to wood, it is almost completely unknown which phytoliths are produced in the bark. A major problem of some comparative studies is that they do not treat wood and bark separately but as a unit, e.g. ‘wood/bark’ (Albert and Weiner, 2001), ‘twig’ (Iriarte and Paz, 2009), or ‘stem’ (Mercader et al., 2009). The lack of comparative data on the diagnostic morphotypes for the separation of wood and bark phytoliths hampers assessment of their redundancy in palaeoecological, archaeological and palaeoanthropological contexts when the distinction between wood and bark is a central point of the interpretation. A few examples may illustrate the dilemma. In a study of surface samples from soils in Olduvai Gorge/Tanzania, blocky parallelepipedal phytoliths were assigned to bark (Albert et al., 2006, Fig. 2h), but high percentages of this morphotype in soils could hardly be explained because the dominant vegetation in the area is an arid grassland with a very sparse tree cover. Were these phytoliths really from dicot bark? The use of wood and bark in prehistory and its role for human evolution is also an important issue. Wood phytoliths were reported on two 1.5 million year old Acheulean hand axes from Peninj/Tanzania, interpreted as wood working tools. The image of the ‘phytoliths’, however, does not show phytoliths, but polyhedral calcium oxalate crystals (Dominguez-Rodrigo et al., 2001, Fig. 2). In a recent study on phytoliths in the dental calculus of 2 million years old Australopithecus sediba (Malapa, South Africa), ‘dicotyledon wood/bark’ phytoliths were reported (Henry et al., 2012, Fig. 3b, and SOM Table 5), resulting in the interpretation that wood and bark were components of the diet of Australopithecus sediba. Because wood is not edible for humans, and bark is only consumed in boreal areas € (Swetnam, 1984; Ostlund et al., 2009), it is important to clarify if the phytoliths of the calculus can be unequivocally attributed to wood and/or bark, or if this is another case of redundancy with morphotypes from other origins. Wood and bark constitute two functionally different components of the plant stem, being responsible for the water and nutrient transport respectively (Esau, 1965). Therefore differences in the cellular composition of the tissues should become manifest in the production of different phytolith morphotypes. In this paper we present the results of a study on wood and bark phytoliths from modern reference material mainly from West Africa. By treating wood and bark separately, we show that they produce clearly different phytolith assemblages. We also assess redundancy with phytoliths from other plant parts and taxonomic groups and eventually evaluate their taxonomic value and the reliability of identification of wood and bark phytoliths in different ecological and archaeological contexts. Wood and bark are universal materials for numerous purposes in everyday life, e.g. as construction material, firewood, tools, ropes, textiles and containers. Identification of wood and bark phytoliths can open new perspectives for archaeology and

palaeoanthropology because phytoliths are the final remnants of wood and bark long after their decay in the soil.

2. Material and methods The wood reference collection in the Institute of Archaeological Sciences at Goethe University Frankfurt includes ca. 1200 wood samples, mainly from West and Central Africa. Most of them are backed up with corresponding herbarium specimens in the Herbarium Senckenbergianum (Frankfurt). 1123 anatomical slides of 795 species had been previously checked by K.N. and her collaborators in the course of wood anatomical studies, and 31 silicacontaining species from 15 families had been identified for the West African savanna region. 35 corresponding wood samples were selected for phytolith extraction. For detecting patterns of silica production in the wood, the following species were represented by more than one specimen: Lannea acida (4) and Trichilia emetica (2). Species with no observable silica in the anatomical slides were not included in the study. Numerous wood samples mostly from twigs collected by members of the Frankfurt team still had their bark adhering. Bark samples of 92 representative species from 34 plant families, including those with known silica in wood, were selected for phytolith extraction. The following species were represented by 2e3 samples each: Acacia tortilis, A. nilotica, Alchornea cordifolia, Annona senegalensis, Boscia senegalensis, Capparis tomentosa, Ficus ingens, Kigelia africana, Parinari curatellifolia and Trema orientalis. Every phytolith sample has a laboratory number (PHV), corresponding with the wood collection number. Plant family and species names follow The Plant List (http://www.theplantlist.org/). For species of the large Leguminosae family, the subfamily names Caesalpinioideae, Mimosoideae and Papilionoideae are indicated. For phytolith extraction, we used a modified dry- and wetashing method after Piperno (2006). The bark samples were washed in Alconox® to remove contaminations. 1e4 g of dried plant material, first weighed to ±0.001g, was ashed in a muffle furnace at 500  C for 8e12 h. For carbonate removal, the ash was treated with hydrochloric acid (HCl) for 10e15 min. at 95  C, washed with destilled water and centrifuged at 3000e3800 rpm. For removal of organic substances, the sample was treated with nitric acid (HNO3) and potassium chlorate (KlO3) for 1e3 h, centrifuged, washed three times with ethanol 95%, centrifuged and dried. The resulting phytolith samples were placed in a small glass tube and weighed. Silica content is indicated as % of dry weight, as conventionally used in wood and phytolith studies (Amos, 1952; Pettersen, 1984; Piperno, 2006). For a few wood and bark samples no quantitative data were available because silica content was too low to be measured, or the samples had been processed before the beginning of this study. After complete removal of water and ethanol with xylol, a small amount of the dried extracted sample was mounted on a microscope slide with Caedax. In a few cases immersion oil was used to enable 3D observation. We studied the slides with an optical light microscope Leica DMLS at magnifications of 200, 400, 630; photos were taken with a camera Leica DFC 320. To better recognize the surface structure and decoration, we studied some samples with a scanning electron microscope Hitachi S4500. We worked with 5 KV and the lower detector. Images of the phytoliths were taken with the photo system and program of Point Electronic, DISS5. Presence/absence of the different morphotype classes was noted in a semi-quantitative analysis. For assessing the abundance of the major morphotype classes in the bark samples, we used three categories after Iriarte and Paz (2009): (a) abundant, one or more in each image field; (c) common, one or more in each slide transect; (r) rare, on the order of one to three in each slide. For the Globulars

Please cite this article in press as: Collura, L.V., Neumann, K., Wood and bark phytoliths of West African woody plants, Quaternary International (2016), http://dx.doi.org/10.1016/j.quaint.2015.12.070

L.V. Collura, K. Neumann / Quaternary International xxx (2016) 1e18

s.l. (Class G) and Aggregates (Class H) in wood, we used the two categories (x) abundant, and (r) rare. 3. Classification, naming and description of morphotypes (Table 1) Whenever possible, we applied the ICPN protocol (ICPN Working Group, 2005) for naming and describing the morphotypes. The standard protocol developed by the ICPN suggests that the name of a morphotype should be formed by a maximum of three descriptors, listed in the following order: 1. shape, 2. texture and/or ornamentation and 3. anatomical origin when clear and beyond doubt. The protocol dictating the shape of the morphotype as first descriptor proved to be problematic. When phytoliths are studied in their anatomical context it becomes evident that silicified cells from one and the same tissue, e.g. parenchyma, cork or sclerenchyma, display a high variability of shapes. The exclusive use of the shape as first descriptor can lead to numerous apparently dissimilar morphotypes that are actually variants of a same cell or tissue type.

3

Many morphotypes in the studied bark samples reflect the anatomy of various cells and tissues which can be clearly identified. In this case it is appropriate to use the anatomical name as first descriptor. If the anatomical origin of the phytolith is not recognizable, the first descriptor is a term indicating the most conspicuous features of the morphotype. In most cases this is the shape as proposed by the ICPN. In the case of Elongates, we used a two-dimensional descriptor as first name because the three-dimensional features are often not clearly visible in the light microscope. An exception is class G. Aggregates which uses the texture as first descriptor. These silica bodies consist of numerous tiny nodules or grains and show very variable and irregular shapes. For the morphotypes of this class, we use the names ‘Nodular’ and ‘Granular’ as first descriptor to indicate that they are aggregates and not solid bodies with various surface decorations. The SEM study has shown that several phytoliths, especially fibres, different elongates and globulars also consist of tiny silica particles. However, these morphotypes have anatomical or morphological names because their anatomical origin or geometric shape is consistent and clearly identifiable.

Table 1 Morphotype classes, morphotypes and descriptions of wood and bark phytoliths. Class/Subclass

Morphotype

Description

Figure

Size (mm)

A. Sclerenchyma A.1. Fibre

Fibre pitted

Long and thin, sometimes with pointed end(s), psilate surface, with elongate, crest-like knobs in oblique position to the long axe of the phytolith (¼ silica filling of fibre pits) Long and thin, sometimes with pointed end(s), psilate, with large concave facets Long and thin, with pointed end(s), sometimes bent, psilate

5a, b

(i) elongate with granulate surface, consisting of an aggregation of small, smooth granules (ii) elongate columellate with parallel long sides (iii) elongate columellate facetate (iv) irregularly elongate columellate (v) irregularly elongate sporadically columellate (vi) blocky columellate (vii) blocky sporadically columellate (viii) irregular columellate (ix) irregular dendriform (i) elongate or irregular with thick silicified walls, non-silicified lumen and pit channels (ii) thin elongate psilate with long, sometimes dendritic projections (i) isodiametric with penta- to hexagonal faces, psilate, single and multicell (ii) tabular, penta- to hexagonal, psilate, single and multicell Rectangular (2D)/paralellepipedal (3D), L/ W  2, psilate, single and multicell

5e, g

l: 95e910, Runge (1999), p. 34 Plate III 4: A6 Rod with w: 6-10 short diagonal lines. €mberg (2003), p. 315 Fig. 4.12g: Elo-3 Stro Smooth, cylindric rod; Fig. 4.12h: Elo-8 Sinuous irregular rod; Fig. 4.12i: Elo-9 Faceted, straight rod; Fig. 4.12k: Elo-11 Rod with knobs; Fig. 4.12o: Elo-17 Thick trapezoidal elongate with oblique ridges/ knobs. € mberg l: 50e555, Runge (1999), p. 30 Plate I 2. Stro w: 12-40 (2003), p. 318 Fig. 4.14d: Scl-7. Large oblong S-body; Subtype 2; Fig. 4.14e: Scl-8. Compact irregular S-body; Subtype 3.

Fibre facetate Fibre psilate

A. Sclerenchyma A.2. Sclereid

Sclereid pitted

Sclereid psilate

B. Parenchyma/Cork (¼ P/C) P/C Blocky B.1. Blocky polyhedral

P/C Blocky paralellepipedal

P/C Blocky with irregular projections P/C Blocky cavate fimbriate P/C Blocky multilayered B. Parenchyma/Cork B.2. Parenchmyma strand

Parenchyma strand

C. Cork aerenchyma

Cork aerenchyma

D. Elongate

Polyhedral to paralellepipedal, psilate, with irregular projections at the edges, only in bark of Stereospermum kunthianum, single and multicell Polyhedral to paralellepipedal with a central cavity and fimbriate edges, psilate, only in bark of Celtis mildbraedii, single and multicell ±Paralellepipedal blockies consisting of numerous thin tabular layers, not seen as multicell, only in bark of Maerua crassifolia Multicell strand of elongate cells, rarely paralellepipedal blockies, with ± straight margins, psilate; also single elongate cells Tabular silicified tissue, various shapes, with irregular perforations of various size

5c 5d

5f, h 5i 5k 5j 5l 5m 5n 5o 5per

References

20e110 30e60 80e200

5s, t

100e230

6a; 5c

10e50

€mberg (2003), p. 316 Fig. 4.13c: One Stro sinuate edge elongate. €mberg (2003), p. 314 Fig. 4.11f: Cl-2. Stro Attached verrucate silica.

6ced; 7d 6b; 7f

40e60

6hek

10e15

7gel

25e50

7a, b

5e15

6eeg; 7e

l: 35e100, w: 20

7m, n

40e700

€mberg (2003), p. 319 Fig. 4.15g: Blo-6. Stro 3D Blocky polyhedron. Types Blo-1. to Blo€mberg (2003), p. 319 Fig. 4.15bef can 5. Stro also be included in this morphotype.

Runge (1996), p. 323 Fig. 10: Verkieseltes Pl€ attchen mit unregelm€ aßigen Perforationen.

4a, b (continued on next page)

Please cite this article in press as: Collura, L.V., Neumann, K., Wood and bark phytoliths of West African woody plants, Quaternary International (2016), http://dx.doi.org/10.1016/j.quaint.2015.12.070

4

L.V. Collura, K. Neumann / Quaternary International xxx (2016) 1e18

Table 1 (continued ) Class/Subclass

Morphotype

Description

Figure

E. Blocky unspecific F. Globular s.l.

Elongate, with unsilicified elliptic cavities (pits?) on the surface arranged perpendicular to the length of the body, reminding of a staircase Elongate facetate Elongate, with concave facets either in side or surface view, representing impressions of adjacent cells Elongate vesicular Elongate, thin, with large, spherical to elliptic concavities, only in wood of Capparis tomentosa Elongate with Cylindroid, with spherical to elliptic central or globular part(s) terminal swellings Elongate irregular Elongate, high variability in shape and size Blocky irregular Variable in shape and size Tabular irregular Tabular, thin, variable in shape and size Psilate sphere, single or composed, sometimes Globular/ Subglobular psilate; with dark core types 1e2 in Table 4

4e 4f, g 4hej 4kem 4n, o 3aed

12e430 5e19

5e55

3hej

10e50

Ellipsoid with a echinate surface, single or 3k, l composed, only in bark of Boscia species Smooth sphere with numerous cylindric 3s, t projections, only in bark of Celtis adolfi-fridericii

10e45

G. Aggregates

Ellipsoid/Lobate; types 3e5 in Table 4 Ellipsoid echinate; type 9 in Table 4 Globular cylindric projections; type 10 in Table 4 Nodular

H. Silica particles/ Accumulations

Granular Irregular aggregation of small, smooth granules 3q, r Small irregular silica particles sometimes occurring aggregated, 3u amorphous silica aggregations

Aggregation of small, smooth nodules with various shapes: globular, elongate or irregular

The second descriptor indicates the second most important feature of the morphotype, usually related to shape or ornamentation. If no appropriate terms were found in the ICPN, we used standard geometrical terms or Stearn's Botanical Latin (Stearn, 2008). To avoid unnecessary long names for the different variants of the major morphotypes, we did not use a third descriptor as part of the name, except in a few special cases. Instead, sub-types were

References

4c, d

Globular/ Sphere with various surface decorations, single 3eeg or composed Subglobular decorated; type 6 in Table 4

Elliptic to irregular (with two or more lobes), psilate

Size (mm) l: 50e250, w: 25e50

Elongate scalariform

3mep

Piperno (1988): Spherical smooth. Kondo et al. (1994), Plate 10: Spherical smooth. €mberg (2003), Runge (1999): Type B1. Stro p. 314 Fig. 4.11e: Cl-1 Smooth sphere and subsphere. Iriarte and Paz (2009), p. 111 Fig. 2c: Globular smooth. Mercader et al. (2009): Globular psilate. Garnier et al. (2013), p. 420 Plate 3, B: Globular psilate (e), Globular with dark core (f). Piperno (1988), p. 224 Plate 7: “irregularly angled or folded surface” and “spherical nodular”; p. 224 Plate 5: Spherical rugulose. Kondo et al. (1994), Plate 12e13: Spherical verrucose. Runge (1999): B2 “Spherical with a rough (not smooth) surface”; p. 37 € mberg (2003), p. Plate IV, 1: Type B6. Stro 314 Fig. 4.11c: Cl-7 Small rugulose sphere and subsphere. Piperno (2006), p. 196 Fig. 2.13. Neumann et al. (2009), p. 6 Fig. 2c: Globular verrucate; p. 6 Fig. 2d: Globular columellate. Iriarte and Paz (2009), p. 111 Fig. 2a. Mercader et al. (2009), p. 104 Fig. 6n: Globular verrucate oblong, ‘Large elongated sub-spheroid with irregular texture and processes’. Mercader et al. (2009), p. 100 Fig 3.l-n: Globular folded; p. 100 Fig. 3.o, q: Globular granulate. Garnier et al. (2013), p. 420 Plate 3, Ba-d: Globular decorate indiff.

5e10

10e168

Piperno (1988), p. 223 Plate 3e4: Spherical nodular. Kondo et al. (1994), Plate 15e16: €mberg (2003), p. 314 Spherical nodular. Stro Fig. 4.11i: Cl-8. Large nodular body. Neumann et al. (2009), p. 6 Fig. 2e: Globular nodular. Iriarte and Paz (2009), p. 111 Fig. 2eef: Globular granulate. Mercader et al. (2009), p. 100 Fig. 3r, s: Globular tuberculate, ‘Large spheroid beset with nodular processes’. Mercader at al. (2009), p. 94 Tab. 2.

10e33

described and classified with Latin letters, but not named sepa€mberg (2004). rately, similar as in Stro In total we established eight different classes. Classes A. Sclerenchyma, B. Parenchyma/Cork, and C. Cork aerenchyma are named after their anatomical origin. Classes D. Elongate, E. Blocky unspecific and F. Globular s.l. are geometric classes defined after their shape, and Class G. Aggregates after its texture. In the course of the

Please cite this article in press as: Collura, L.V., Neumann, K., Wood and bark phytoliths of West African woody plants, Quaternary International (2016), http://dx.doi.org/10.1016/j.quaint.2015.12.070

L.V. Collura, K. Neumann / Quaternary International xxx (2016) 1e18

study, we found that irregular agglomerations of silica particles are very common in bark, so that they deserve their own Class H. Silica particles/Accumulations. In the final evaluation of the data, it turned out to be useful to differentiate the morphotype classes into two groups: 1. specific morphotypes with consistent morphologies, 2. unspecific morphotypes. In the first case the phytoliths have either a clearly definable anatomical origin or repetitive geometrical characteristics (Classes A., B., C., D., F. and G.). On the contrary, morphotypes of class E. (Blockies unspecific) cannot be attributed to any tissue or cell type and display very irregular shapes. Class H. Silica particles/Accumulations also belongs to the group of unspecific phytoliths. Class A. Sclerenchyma comprises the two subclasses A.1. Fibres and A.2. Sclereids which are anatomically related due to their lignified walls (Esau, 1965, p. 203 ff.). Silicified fibres can occur single or in bundles and are then multicellular skeletons while silicified sclereids usually occur only as single cells. Parenchyma and cork (Class B. Parenchyma/Cork) are structurally and functionally different tissues in the bark. Silicified tissue fragments consisting of ±isodiametric polyhedral cells cannot always be safely attributed to one or the other (Roth, 1981, pp. 88 ff.). Therefore we merged them into one class. Silicified parenchyma and cork are often present as multicellular tissue fragments (skeletons); single cells are also common. Subclass B.1. Blocky comprises a wide range and variation of geometric cell shapes, from ±isodiametric polyhedrons to par€ mberg, 2003) describes allelepipedal. The term ‘blocky’ (after Stro parenchyma and cork cells with L/W  2. Subclass B.2. Parenchyma strand comprises multicellular elongate or blocky parallelepipedal phytoliths in parallel or consecutive arrangement. Single elongate phytoliths can be assigned to class B.2. if they occur in a specific context, accompanied by multicellular parenchyma strands.

5

All single Elongates which cannot be identified as fibres, sclereids or parenchyma/cork, are allocated to class D. Elongates. Class F. Globular s.l. comprises morphotypes with a more or less rounded shape and soft outlines, including globular, subglobular, ellipsoid and lobate forms. The definition of the class is broad because many transitional forms can be present together in the same species. These morphotypes are abundant in wood. Often the Globulars s.l. occur composed, i.e in multiples of two or more (¼ compound in Bowdery et al., 2001). 4. Results 4.1. Absolute silica content Among the 35 studied wood samples, 11.4% have silica contents per g dry weight of 2e<5%, 28.6% have 1e<2%, 11.4% have 0.5e1%, 37.1% have 0.05e0.5% (Table 2, Fig. 1). The four wood samples of Lannea acida show great variation in the silica content between 2.64% and 0.78%, the variation in the two wood samples of Trichilia emetica is 1.71% and 0.43% respectively. The highest amounts of silica were observed in the bark. Almost all bark samples have silica, although in very variable quantities. In 10.7% of the bark samples silica content is >5%, with Stereospermum kunthianum as the best producer (17.6%). 7.8% of the bark samples had 2e< 5% silica, 13.6% had 1e< 2%, 13.6% had 0.5e< 1%, 43.6% had 0.05e< 0.5%, and 2.9% had <0.05% (Table 3, Fig. 1). The species of which several different bark samples from different trees had been processed, showed considerable variation in their silica content. The three samples of Kigelia africana had 11.02%, 9.48% and 0.24% respectively, the two samples of Boscia senegalensis had 2.62% and 1.9% respectively. Specific phytoliths with consistent morphologies are present in 8 of 11 species with silica contents of >5%, and in 11 out of 22 species with silica contents of 1e5%. Among the species with silica contents <1%, only a few produce specific phytoliths.

Table 2 Absolute silica content in wood samples. * ¼ species with more than one processed sample. PHV no.

Wood coll. no.

Family

Species

Sample weight

Silica weight in g

% Silica/g dry weight

940 882 915 832 924 919 936 921 939 934 867 918 870 884 938 923 933 914 926 971 935 916 941 925 947 872 922 917 927

503 306 1491 851 1700 368.2 1749.2 1745 870 178.0 482.1 1590 433.2 221 1551 249.2 14.0 950 333 889.2 242.2 1649 506 1593 1554 1485 1391 932.2 480.2

Anacardiaceae Malvaceae Euphorbiaceae Anacardiaceae Chrysobalanaceae Anacardiaceae Sapotaceae Burseraceae Anacardiaceae Sapotaceae Leguminosae Caesalpinioideae Anacardiaceae Meliaceae Sapotaceae Anacardiaceae Capparaceae Sapotaceae Phyllanthaceae Meliaceae Putranjivaceae Sapotaceae Phyllanthaceae Anacardiaceae Dipterocarpaceae Malvaceae Olacaceae Burseraceae Phyllanthaceae Rubiaceae

Lannea fruticosa Cola cordifolia Sapium shirakiopsis Lannea acida* Maranthes polyandra Lannea acida* Vitellaria paradoxa Commiphora africana Lannea acida* Pouteria alnifolia Dialium guineense Lannea barteri/kerstingii Trichilia emetica* Synsepalum brevipes Lannea acida* Capparis tomentosa Manilkara obovata Hymenocardia acida Trichilia emetica* Drypetes gilgiana Mimusops kummel Uapaca guineensis Lannea humilis Monotes kerstingii Cola laurifolia Olax subscorpioides Canarium schweinfurthii Uapaca togoensis Mitragyna inermis

2.1454 1.4211 2.257 0.7 2.7656 2.1645 3.098 2.0187 2.2277 2.2663 2.0211 2.3718 2.0313 1.3301 2.2683 2.1408 1.004 2.8479 1.0239 2.0117 3.1854 2.5581 3.8225 2.1325 2.6474 3.1606 2.3237 1.9423 2.0948

0.0992 0.0454 0.0632 0.0185 0.0508 0.0349 0.0474 0.0285 0.0314 0.0286 0.0255 0.0282 0.0238 0.0141 0.0181 0.0162 0.0065 0.0154 0.0044 0.0083 0.0124 0.0099 0.0132 0.0073 0.0073 0.0061 0.0042 0.0029 0.0022

4.62 3.19 2.80 2.64 1.84 1.61 1.53 1.41 1.41 1.26 1.26 1.19 1.17 1.06 0.80 0.76 0.65 0.54 0.43 0.41 0.39 0.39 0.35 0.34 0.28 0.19 0.18 0.15 0.11 (continued on next page)

Please cite this article in press as: Collura, L.V., Neumann, K., Wood and bark phytoliths of West African woody plants, Quaternary International (2016), http://dx.doi.org/10.1016/j.quaint.2015.12.070

6

L.V. Collura, K. Neumann / Quaternary International xxx (2016) 1e18

Table 2 (continued ) PHV no.

Wood coll. no.

Family

Species

Sample weight

Silica weight in g

% Silica/g dry weight

886 920 828 830 358 834

920 1733 871.1 769 528.2 855

Malvaceae Boraginaceae Malvaceae Burseraceae Chrysobalanaceae Anacardiaceae

Sterculia setigera Cordia sinensis Bombax costatum Boswellia papyrifera Parinari curatellifolia Sclerocarya birrea

1.3636 2.212 2.51 1.2 0.91 1.65

0.0013 0.0015 no data no data no data no data

0.10 0.07 no data no data no data no data

Table 3 Absolute silica content in bark samples. * ¼ species with more than one processed sample. PHV no.

Wood coll. no.

Family

Species

Sample weight

Silica weight in g

% Silica/g dry weight

Specific phytoliths present

911 908 982 972 841 853 859 895 851 974 891 930 978 904 985 850 981 931 932 948 887 849 983 892 905 896 848 879 929 986 973 868 874 871 910 987 975 837 943 883 876 862 979 861 894 909 890 860 844 984 903 858 835 928 856 877 897 980 838

1007.1 866.1 1634 1649 866.2 35 482.1 577 1700 1638 1010.2 249.2 1617 1192 881.1 734 1624 1686 1157.1 1554 92 817.1 1058.0 253.1 1637 838 1583 1196.2 1577 1583 1633 1034 925.1 1485 799.1 1680 1589 751 694 221 178 1066 895.0 1051 914.1 919.1 908.2 1054 909 867.2 1176 639.2 1593 1009 1177.1 1295.2 155 1678 1743

Bignoniaceae Bignoniaceae Irvingiaceae Phyllanthaceae Bignoniaceae Euphorbiaceae Leguminosae Caesalpinioideae Capparaceae Chrysobalanaceae Euphorbiaceae Capparaceae Capparaceae Cannabaceae Combretaceae Moraceae Capparaceae Annonaceae Burseraceae Sapotaceae Malvaceae Capparaceae Capparaceae Malvaceae Moraceae Urticaceae Capparaceae Chrysobalanaceae Rubiaceae Phyllanthaceae Chrysobalanaceae Euphorbiaceae Moraceae Opiliaceae Olacaceae Meliaceae Cannabaceae Phyllanthaceae Combretaceae Capparaceae Sapotaceae Sapotaceae Leguminosae Mimosoideae Cannabaceae Leguminosae Caesalpinioideae Ochnaceae Lamiaceae Meliaceae Leguminosae Caesalpinioideae Phyllanthaceae Salicaceae Capparaceae Leguminosae Caesalpinioideae Dipterocarpaceae Burseraceae Leguminosae Papilionoideae Salvadoraceae Malvaceae Euphorbiaceae Combretaceae

Stereospermum kunthianum Kigelia africana* Klainedoxa cf. gabonensis Uapaca guineensis Kigelia africana* Sapium shirakiopsis Dialium guineense Boscia salicifolia Maranthes polyandra Macaranga heudelotii Boscia angustifolia Capparis tomentosa* Celtis adolfi-fridericii Guiera senegalensis Morus mesozygia Boscia senegalensis* Anonidium mannii Canarium schweinfurtii Manilkara obovata Cola laurifolia Boscia senegalensis* Maerua crassifolia Cola cordifolia Ficus sur Musanga cecropioides Capparis decidua Parinari curatellifolia* Mitragyna inermis Uapaca togoensis Parinari curatellifolia* Mallotus oppositifolius Ficus cf. ingens* Opilia celtidifolia Olax subscorpioides Khaya senegalensis Trema orientalis* Securinega virosa Combretum molle Cadaba glandulosa Synsepalum brevipes Pouteria alnifolia Acacia nilotica var. cf. nilotica* Celtis zenkeri Cassia sieberiana Ochna afzelii Vitex doniana Pseudocedrela kotschyi Detarium microcarpum Antidesma venosum Oncoba spinosa Crateva religiosa Tamarindus indica Monotes kerstingii Commiphora africana Lonchocarpus laxiflorus Salvadora persica Grewia mollis Alchornea cordifolia* Terminalia prunioides

2.0281 2.2383 2.1352 1.2918 1.32 0.1996 2.5488 1.6589 1.948 0.8122 2.0185 1.8629 2.1886 1.0593 3.3231 0.74 1.0007 2.1105 0.981 1.7173 4.0915 2.43 3.807 3.4095 0.7377 2.4194 2.23 1.3558 2.1335 2.089 1.09 2.2464 0.8765 0.173 2.2124 1.9037 0.4627 0.79 0.2186 1.2222 0.8055 2.5618 2.4472 2.106 4.6892 2.3594 0.3767 3.1214 0.96 0.438 0.3794 3.721 2.36 0.9712 2.7516 1.0355 2.0331 0.6067 0.3

0.3569 0.2467 0.2273 0.1265 0.1252 0.0175 0.2225 0.1396 0.1584 0.0583 0.1348 0.0907 0.0951 0.0308 0.0914 0.0194 0.0219 0.0246 0.0204 0.0332 0.0767 0.0442 0.0658 0.0567 0.0114 0.0354 0.0310 0.0179 0.0270 0.0262 0.0131 0.0262 0.0093 0.0016 0.0178 0.0153 0.0036 0.0060 0.0016 0.0089 0.0053 0.0160 0.0145 0.0012 0.0258 0.0125 0.0019 0.0015 0.0042 0.0019 0.0016 0.0146 0.0088 0.0035 0.0085 0.0032 0.0062 0.0017 0.0008

17.60 11.02 10.65 9.79 9.48 8.76 8.73 8.42 8.13 7.18 6.68 4.87 4.35 2.91 2.75 2.62 2.19 2.08 2.08 1.93 1.88 1.82 1.73 1.66 1.55 1.46 1.39 1.32 1.27 1.25 1.20 1.17 1.06 0.92 0.80 0.80 0.78 0.76 0.73 0.73 0.66 0.62 0.59 0.57 0.55 0.53 0.50 0.48 0.44 0.43 0.42 0.39 0.37 0.36 0.31 0.31 0.31 0.28 0.27

x x

x x x x x x x

x x x

x x x x x x

x

x

x

x

x

Please cite this article in press as: Collura, L.V., Neumann, K., Wood and bark phytoliths of West African woody plants, Quaternary International (2016), http://dx.doi.org/10.1016/j.quaint.2015.12.070

L.V. Collura, K. Neumann / Quaternary International xxx (2016) 1e18

7

Table 3 (continued ) PHV no.

Wood coll. no.

Family

Species

Sample weight

Silica weight in g

875 888 873 863 847 865 945 907 854 855 889 893 829 880 902 869 831 898 900 836 976 866 899 885 878 857 977 906 946 864 827 839 881 901 833 793 842 846 792 840 843 794 845 789

242.2 740 1187.1 1595 1288 1183.1 906 766 922.1 1083.1 1053 1375.2 769 1094 1184.1 433.2 851 819 729 1030 1582 117.1 810 920 1084 1197.2 882.0 650 216.1 1291 871.1 856 624 406 855 1228.1 1578 1014 1619 1733 883 889.2 1164.1 928.2

Sapotaceae Bignoniaceae Olacaceae Leguminosae Mimosoideae Zygophyllaceae Leguminosae Mimosoideae Bignoniaceae Moraceae Moraceae Leguminosae Papilionoideae Combretaceae Ochnaceae Burseraceae Rubiaceae Leguminosae Caesalpinioideae Meliaceae Anacardiaceae Leguminosae Mimosoideae Leguminosae Mimosoideae Combretaceae Annonaceae Leguminosae Mimosoideae Leguminosae Mimosoideae Malvaceae Polygalaceae Leguminosae Papilionoideae Ulmaceae Celastraceae Annonaceae Leguminosae Mimosoideae Bombacaceae Annonaceae Rhamnaceae Leguminosae Mimosoideae Anacardiaceae Euphorbiaceae Phyllanthaceae Capparidaceae Cannabaceae Boraginaceae Putranjivaceae Putranjivaceae Phyllanthaceae Cannabaceae

Mimusops kummel Kigelia africana* Ximenia americana Albizia zygia Balanites aegyptiaca Parkia biglobosa Markhamia tomentosa Ficus cf. ingens* Ficus sycomorus Pterocarpus erinaceus Combretum glutinosum Lophira lanceolata Boswellia papyrifera Crossopteryx febrifuga Isoberlinia doka Trichilia emetica Lannea acida Acacia tortilis var. raddiana* Faidherbia albida Terminalia avicennioides Annona senegalensis* Acacia seyal Acacia nilotica* Sterculia setigera Securidaca longepedunculata Pterocarpus lucens Holoptelea grandis Gymnosporia senegalensis Uvaria chamae Acacia tortilis* Bombax costatum Annona senegalensis* Ziziphus jujuba Acacia mellifera Sclerocarya birrea Alchornea cordifolia* Bridelia ferruginea Capparis tomentosa* Celtis mildbraedii Cordia sinensis Drypetes floribunda Drypetes gilgiana Hymenocardia acida Trema orientalis*

1.3453 2.054 2.4733 2.4115 2.15 2.1032 2.1777 2.0629 2.4168 3.308 1.015 2.348 2.02 2.3656 2.6541 2.2496 3.98 2.1083 1.0383 3.54 2.0897 2.2494 2.0606 2.3773 2.0188 2.0862 2.251 2.351 2.954 2.6082 2.65 2.4 2.0972 3.402 2.6 0.34 2.51 0.59 1.26 0.87 0.55 0.22 3.03 0.78

0.0034 0.0050 0.0059 0.0057 0.0046 0.0042 0.0043 0.0036 0.0276 0.0057 0.0017 0.0039 0.0032 0.0038 0.0040 0.0033 0.0057 0.0027 0.0013 0.0044 0.0023 0.0024 0.0020 0.0019 0.0016 0.0014 0.0015 0.0015 0.0018 0.0015 0.0015 0.0012 0.0010 0.0010 0.0007 no data no data no data no data no data no data no data no data no data

4.2. Morphotype classes in bark and wood In the bark, 34% of the studied species have specific phytoliths with consistent morphologies that can be attributed to either a tissue or a clearly definable morphotype (Table 4). Most of these species have two or more different specific morphotypes. Fibres (A.1) are represented in 4%, Sclereids (A.2) in 10%, Parenchyma/Cork (B) in 22%, Cork aerenchyma (C) in 9%, Elongates in 14%, and Globulars s.l. (F) in 11%. Class (G) Aggregates is only present in 2% of the studied species (Fig. 2). 57% have no specific phytoliths, but only Silica particles/Accumulations (H) and Blockies unspecific (E) (SOM Table 1). These two classes are almost universally present: Silica particles/accumulations in 80% of all studied species, and Blockies unspecific in 63%. In only eight species (9%) no silica was observed in the slides of the bark phytolith samples. In these cases the absolute silica production is so low that the few phytoliths are probably lost during extraction. In the wood, most morphotypes belong to the Class Globulars s.l. (F) and Aggregates (G) (Table 5). They can occur either in isolated or composite forms (Fig. 3f). The most common type is the Globular psilate, being present in 22 out of 26 studied species (84.6%), (types

% Silica/g dry weight 0.25 0.24 0.24 0.24 0.21 0.20 0.20 0.17 0.17 0.17 0.17 0.17 0.16 0.16 0.15 0.15 0.14 0.13 0.13 0.12 0.11 0.11 0.10 0.08 0.08 0.07 0.07 0.06 0.06 0.06 0.06 0.05 0.05 0.03 0.03

Specific phytoliths present x

x x

x

1 and 2 in Table 5, Fig. 3aed). These spherical phytoliths often consist of a thin shell enclosing a central cavity (Fig. 3b). The Globular decorated (type 6, Fig. 3eeg) is also well represented in 14 species (53.8%). Less common are Lobate psilate (type 5, Fig. 3h and i) or Ellipsoid/Cylindric (types 3 and 4, Fig. 3j). Of the Class Aggregates (G), the Nodular (type 7, Fig. 3mep) occurs in 10 species, the Granular (type 8, Fig. 3q and r) in 3 species. A unique Ellipsoid/ Cylindric echinate was found in the bark of three Boscia species (type 9, Fig. 3k and l) and a unique Globular with cylindric projections in the bark of Celtis adolfi-fridericii (type 10, Fig. 3s and t). Fibres and Sclereids (A), Cork/parenchyma (B), and Cork aerenchyma (E) are completely absent in wood. Elongates (D) are only present in the wood of Capparis tomentosa, Blockies unspecific (E) only in Vitellaria paradoxa. Silica particles/accumulations (H) have been observed in the wood of Canarium schweinfurthii, Commiphora africana, Bombax costatum, and Vitellaria paradoxa, rarely in Hymenocardia acida and Lannea acida. In the processed wood samples of Sclerocarya birrea, Cordia sinensis, Boswellia papyrifera, Bombax costatum, and Sterculia setigera no identifiable phytoliths were found. A few phytoliths had been observed in the anatomical slides of these species, but due to the low silica content in the wood, they were probably lost during extraction (Table 2).

Please cite this article in press as: Collura, L.V., Neumann, K., Wood and bark phytoliths of West African woody plants, Quaternary International (2016), http://dx.doi.org/10.1016/j.quaint.2015.12.070

8

Family

Species

A.1. Fibres A.2. Sclereids B. Parenchyma/Cork C. Cork aerenchyma D. Elongates E. Blockies unspecific F. Globulars G. Aggregates H. Slica Silica in wood particles/Accumulations of species

Annonaceae Bignoniaceae Bignoniaceae Burseraceae Burseraceae Cannabaceae Cannabaceae Cannabaceae Capparaceae Capparaceae Capparaceae Capparaceae Capparaceae Chrysobalanaceae Chrysobalanaceae Dipterocarpaceae Euphorbiaceae Euphorbiaceae Euphorbiaceae Irvingiaceae Leguminosae Caesalpinoideae Malvaceae Meliaceae Moraceae Moraceae Moraceae Moraceae Phyllanthaceae Phyllanthaceae Sapotaceae Urticaceae

Anonidium mannii Kigelia africana Stereospermum kunthianum Canarium schweinfurthii Commiphora africana Celtis adolfi-fridericii Celtis mildbraedii Trema orientalis Boscia angustifolia Boscia salicifolia Boscia senegalensis Capparis tomentosa Maerua crassifolia Maranthes polyandra Parinari curatellifolia Monotes kerstingii Macaranga heudelotii Mallotus oppositifolius Sapium shirakiopsis Klainedoxa cf. gabonensis Dialium guineense

a a e e e e e e e e e e e a e e a e e e e

r r e a e a e e e e e e e r e e r e e a e

Cola cordifolia Khaya senegalensis Ficus ingens Ficus sur Ficus sycomorus Morus mesozygia Uapaca guineensis Uapaca togoensis Synsepalum brevipes Musanga cecropioides

e e e e e e e e e e

e e e e c e a e e e

a a* a* a e a* a e e e e a* a e e c a a a a e a r c c a a e a

r e c a e e e e e e e e e e e e e e e e

c c e a e c e a e e e e e c e e c e e a a

e a c r r e e e e c

e e e a a e r c e e

c e a e e e a c c e r r e e c c

e e e e c a* e e a* a* a* e e e a a e e e e e

e e e e e e e e e e e e e e e e e e e e e

a a c a a a e a e e e a e a e a a a a a a

e e e x x e e e e e e x e x x x e e x e x

e r a r e e r r e e

a e e e e e a a e

c e e e e e e r e e

e a a a a e a c a e

x e e e e e x x x e

a a r

L.V. Collura, K. Neumann / Quaternary International xxx (2016) 1e18

Please cite this article in press as: Collura, L.V., Neumann, K., Wood and bark phytoliths of West African woody plants, Quaternary International (2016), http://dx.doi.org/10.1016/j.quaint.2015.12.070

Table 4 Distribution of morphotype classes in bark. Only species having specific phytoliths with consistent morphologies were considered. * ¼ diagnostic morphotypes present. Abundance: (a) abundant, one or more in each image field; (c) common, one or more in each slide transect; (r) rare, on the order of one to three in each slide.

L.V. Collura, K. Neumann / Quaternary International xxx (2016) 1e18

9

Table 5 Distribution of Globulars s.l., Aggregates and diagnostic Globulars s.l. in wood and bark. 1. Globular psilate; 2. Globular psilate composed; 3. Ellipsoid psilate; 4. Ellipsoid psilate composed; 5. Lobate psilate; 6. Globular decorated; 7. Nodular; 8. Granular; 9. Ellipsoid echinate; 10. Globular with cylindric projections. Abundance: (x) abundant; (r) rare. Family

Wood

Bark

Anacardiaceae Anacardiaceae Anacardiaceae Anacardiaceae Bursaceae Burseraceae Capparaceae Chrysobalanaceae Chrysobalanaceae Dipterocarpaceae Euphorbiaceae Euphorbiaceae Leg. Caesalpiniaceae Malvaceae Malvaceae Meliaceae Olacaceae Phyllantaceae Phyllantaceae Phyllantaceae Rubiaceae Sapotaceae Sapotaceae Sapotaceae Sapotaceae Sapotaceae Burseraceae Cannabaceae Capparaceae Capparaceae Capparaceae Chrysobalanaceae Dipterocarpaceae Malvaceae Phyllanthaceae Sapotaceae

Species

Lannea acida Lannea barteri/kerstingii Lannea fruticosa Lannea humilis Canarium schweinfurthii Commiphora africana Capparis tomentosa Maranthes polyandra Parinari curatellifolia Monotes kerstingii Drypetes gilgiana Sapium shirakiopsis Dialium guineense Cola cordifolia Cola laurifolia Trichilia emetica Olax subscorpioidea Hymenocardia acida Uapaca guineensis Uapaca togoensis Mitragyna inermis Manilkara obovata Mimusops kummel Pouteria alnifolia Synsepalum brevipes Vitellaria paradoxa Commiphora africana Celtis adolfi-fridericii Boscia angustifolia Boscia salicifolia Boscia senegalensis Parinari curatellifolia Monotes kerstingii Cola cordifolia Uapaca togoensis Synsepalum brevipes

Class F. Globular s.l. 1

2

3

4

x x x x r r

x x

r x x

r

5

x x x x x

r r r x r x x x x

x

r

x

r r

x x x x x x x r r

x x

Class G. Aggregates

Class F. Globular s.l. Diagnostic types

7

9

8

x

x r

x x x x

x x x x

x

r

x

x x r

x r r

r r

10

r x r

r r x x r x x x x x r x

6

x x

x

x x

r x x x x x x x x x

x x x x x

x x x r r

r

x

x x

5. Discussion 5.1. Phytoliths in wood Globular s.l. (F) is the most common class in wood (Table 5, Fig. 3aej, s, t). The shape of these morphotypes is independent from the shape of the cells in which they have been formed. The diagnostic significance of the psilate forms (Globular/Subglobular psilate, types 1 and 2, Fig. 3aed; Ellipsoid/Cylindric psilate, types 3 and 4, Fig. 3j; and Lobate (type 5, Fig. 3h and i) is low. They occur both in wood and bark and are also found in many different plant families and in various parts of the plant, e.g. in leaves, seeds and fruits (Kondo et al., 1994; Piperno, 1988, 2006; Runge, 1996, 1999; € mberg, 2003; Iriarte and Paz, 2009; Mercader et al., 2009). Stro Decorated bodies with a globular to ellipsoid shape (type 6, Fig. 3eeg) are well represented in the limited number of species that produce silica in their wood. They are the major constituents of phytolith assemblages in soils under the tropical forests in Central Africa (Runge, 1999). Summarized under the names ‘Circular rugose’ (Alexandre et al., 1997; Barboni et al., 1999), ‘Rugose/ver€ mberg, 2004), ‘Rough spherical’ (Bremond rucate spheres’ (Stro et al., 2005), ‘Globular granulate’ (ICPN Working Group, 2005; Barboni et al., 2007; Mercader et al., 2009), ‘Spherical rugose’ (Iriarte and Paz, 2009), or ‘Globular decorated’ (Neumann et al., 2009; Garnier et al., 2013), this morphotype is used as indicator of woody dicotyledons in the calculation of the index D:P (e.g. Alexandre et al., 1997; Barboni et al., 1999, 2007; Bremond et al.,

x x x r r x

x r

r

2005, 2008). Most of them are not specific for dicotyledon wood because they also occur in leaves, fruits and seeds of many monocots and some eudicots (Kondo and Peason, 1981; Piperno, 1988, 2006; Iriarte and Paz, 2009). Especially the decorated morphotypes with more or less regular globular outlines are also very common in monocotyledons, notably the Zingiberales (Piperno, 2006, p. 37; Chen and Smith, 2013). 50 45 40 35 30 25 20 15 10

5 0 > 5%

2-< 5%

1-< 2%

0,5-< 1%

wood (n = 35)

0,05-< 0,5%

< 0,05%

no data

bark (n = 103)

Fig. 1. Absolute silica contents in % of g dry weight in wood and bark.

Please cite this article in press as: Collura, L.V., Neumann, K., Wood and bark phytoliths of West African woody plants, Quaternary International (2016), http://dx.doi.org/10.1016/j.quaint.2015.12.070

10

L.V. Collura, K. Neumann / Quaternary International xxx (2016) 1e18

Scurfield et al. (1974) have described silification of vessels, fibres, and occasionally ray parenchyma, but in general this is a rare phenomenon in wood. In our study, silicified cells were only observed in the wood of two species: Elongates vesicular in Capparis tomentosa (Fig. 4e), and Blockies unspecific in Vitellaria paradoxa. The rarity of tissue silification in wood corresponds with the very limited occurrence of ‘Vitreous silica’ - a term introduced by Amos (1952) e in the InsideWood database (character 163). Of the 5833 wood descriptions in InsideWood, only 23 taxa have vitreous silica, among them one species from Africa, Streblus africanus (Moraceae). Therefore the silicified pitted ‘vessels’ and ‘tracheids’ often mentioned as specific for wood (e.g. Runge, 1999; Piperno, 2006, p. 42, Fig. 2.17e) are actually not; they must be attributed to other plant organs.

90 80 70 60 50 40 30 20 10 0

5.2. Bark phytoliths in their anatomical context

Fig. 2. Distribution of morphotype classes in bark, in % of studied species.

Much more specific for wood are aggregates of nodules or tiny spherical particles (Class G, types 7 and 8, Fig. 3mer). They occur in 46% of the studied woods, but only in 2% of the bark samples. In our classification, we use the names ‘Nodular’ and ‘Granular’ as first descriptor to indicate that these morphotypes are aggregates and not solid bodies with various surface decorations. In his study on the silicious inclusions in the wood of Lauraceae, Richter (1980) describes the Nodular type as ‘Cauliflower-like rough grains’ and the Granular type as ‘Aggregate grains’ (see also Kondo et al., 1994, p. 22, Plates 14, 15). The aggregate nature of the Nodulars and Granulars clearly separates them from the globular morphotypes with nodular surfaces that have been observed for example in Marantaceae (Piperno, 1988, p. 223, Fig. 3; Chen and Smith, 2013). The shape of the latter is more regularly spherical, and the nodular or rugulose structure represents a surface ornamentation. In the Nodulars and Granulars, by contrast, the single nodules and granules are mostly well distinguishable and the shape of the aggregates is much more irregular. When working with microscopic slides in wood identification, the presence of silica is a very valuable character because it only occurs in a limited number of families, genera and species (Metcalfe and Chalk, 1983; InsideWood, 2004 onwards, http://insidewood.lib. ncsu.edu; Wheeler et al., 2007). Among the 5833 coded taxa in InsideWood, only 534 (ca. 10%) have silica; among the African taxa the ratio is similar (146 out of 1170). The location of the silica bodies is also diagnostic. In his extensive study on silica in 1300 Neotropical woods, Ter Welle (1976a) found that in 85% of the species silica bodies are located in the ray cells, to a lesser extent (20%) in axial parenchyma, and very rarely in fibres and tyloses of the vessels. Isolated phytoliths cannot be attributed to a cell type in which they have been formed, but our data highlight the diagnostic potential of the various shapes, decorations and composition of the silica bodies. The psilate and decorated Globulars are present in all silica-bearing woods, whereas the occurrence of Nodulars and Granulars is much more limited and might serve as diagnostic feature for separating taxa with otherwise identical wood anatomical features. An example is the genus Lannea: Nodulars are common in L. fruticosa, rare in L. humilis and absent in L. barteri/ kerstingii.

The bark, in the most general sense defined as ‘all tissues outside of the vascular cambium’ (Trockenbrodt, 1990), has a very complex anatomical structure (Esau, 1965; Roth, 1981). Besides assimilate transport, its main function is protection. The secondary phloem in the inner bark originates from cambium activity and is responsible for assimilate transport. Its conducting phloem cells are short-lived and very thin-walled and will probably not silicify. The secondary phloem also contains fibre bundles and storage parenchyma that in principle can silicify. The outer part of the bark consists of one to several periderm layers (in the latter case called rhytidome, Trockenbrodt, 1990), each originating from a secondary cork cambium, the phellogen (Roth, 1981). The anatomical structure of the periderm and rhytidome is very diverse and includes several different cell types in variable arrangements: parenchyma, fibres, sclereids, secretory cells, and cork (Roth, 1981). The exact anatomical origin of phytoliths isolated from the bark cannot be determined, but most specific phytoliths with consistent morphologies reflect the anatomy of the cells and can thus be attributed to a tissue: fibres, sclereids, parenchyma/cork, and cork aerenchyma. For some Elongates (Fig. 4aej) the original cell function is not clearly identifiable but they also result from silification of one of the tissues mentioned above. This is a major difference to the wood phytoliths which develop their shapes independently from the cell in which they are formed. However, 10 species also show Globulars s.l. in the bark, mostly Globular psilate and rarely Globular decorated (Tables 4 and 5). They probably develop in living parenchyma cells, comparable to the Globulars in wood. Class A. Sclerenchyma includes both fibres and sclereids. Distinguishing between the two is often not easy because there are numerous transitional forms, the fibre-sclereids (Roth, 1981; Trockenbrodt, 1990). Fibres are long, slender cells which develop directly from meristem derivates (Trockenbrodt, 1990), with either straight or curved outline, elongate oblique pits or pointed ends. Bark fibres are between 300 and 4000 mm long (Roth, 1981, p. 35), a length rarely reached in isolated phytoliths. Isolated silicified fibres break into fragments of various length in soils and sediments, and also during processing of fresh plant material. Long Elongates with psilate surfaces, oblique pits and straight or irregular ends (Fig. 5a and b) are fragments of longer fibres. They correspond to morphotype A6 often occurring in rainforest soils of Central Africa (Runge, 1999). Some fibres have faceted surfaces resulting from the impressions of adjacent cells (Fig. 5c and d). In comparison with fibres, sclereids are more difficult to define. Shapes of sclereids vary between cylindric, irregularly elongate, sometimes with protrusions, to different types of Blockies. Their irregular shapes result from their special ontogeny in the bark. They develop and enlarge in already differentiated tissue, where other cell elements may become obstacles in their way so that the

Please cite this article in press as: Collura, L.V., Neumann, K., Wood and bark phytoliths of West African woody plants, Quaternary International (2016), http://dx.doi.org/10.1016/j.quaint.2015.12.070

L.V. Collura, K. Neumann / Quaternary International xxx (2016) 1e18

11

Fig. 3. Class F. Globulars s.l.: a. Globular psilate, Malacantha alnifolia (W); b. Globular psilate, Canarium schweinfurthii (W), c. Subglobular psilate, Pachystela brevipes (W); d. Globular psilate, with dark core, Cola cordifolia (W); e. Globular decorate, Olax subscorpioides (W); f. Globular decorated composed, Uapaca togonensis (W); g. Globular decorated, Lannea acida (W); h. Lobate psilate, Malacantha alnifolia (W); i. Lobate psilate, Lannea acida (W); j. Ellipsoid psilate, Lannea acida (W); k, l. Ellipsoid echinate, Boscia senegalensis (B); Class G. Aggregates: m. Nodular, Mimusops kummel (W); n, o. Nodular, Sapium ellipticum (W); p. Nodular, Cola cordifolia (W); q, r. Granular, Mitragyna inermis (W); s, t. Globular with cylindric projections, Celtis adolfi-fridericii (B); u. Class H. Silica particles/Accumulations, Kigelia africana (B). (B) ¼ bark, (W) ¼ wood. Frame solid line: diagnostic morphotypes in bark. Frame dashed line: diagnostic morphotypes in wood.

outlines of the sclereids are often bent or wavy (Roth, 1981, p. 41). The surface is either psilate or columellate, depending on the type of silicification. The columellate surface results from complete silification of the lumen and the pits in a stage when lignification of the secondary walls had already been advanced. After deterioration of the thickened lignified cell wall, a cast remains with the silicified pits as tiny protrusions in a dense, irregular arrangement on the surface. The pit fillings can be shorter or longer depending on the original thickness of the secondary wall (Fig. 5feo). The pitted sclereids correspond to type C1 that Runge (1999) attributes to ‘sclereids or tracheids’ of woody plants. However, in the bark they must be sclereids because, with the exception of very few species (Roth, 1981, p. 282), the bark does not contain any waterconducting elements. In the bark of Kigelia africana, Ficus sycomorus and Canarium schweinfurthii elongate phytoliths with a psilate or finely granulate (Fig. 5e and g) surface occur. Some show projections (Fig. 5s and t), some an irregular network on the surface (Fig. 5q), probably representing impressions of adjacent cells. In some cases it is clear that some sclereids result from silification of the wall; the non-silicified lumen is often still discernible, and fissures in the cell wall mark the non-silicified pits (Fig. 5p and r). In the parenchyma two different types of cells silicify: elongate with straight, rectangular outlines, and polyhedral ± isodiametric (Roth, 1981, pp. 21 f.). The silicified polyhedral parenchyma cells are thin-walled, sometimes with a foraminate wall representing the non-silicified pits; when isolated, they appear as thin tabular

Blockies (Fig. 7c and d). In their original anatomical context, the elongate parenchyma cells are arranged in strands of several cells (Fig. 6eeg). Isolated elements of parenchyma strands appear as parallelepipedal Elongates or Blockies (Fig. 7e and f). Cork is the major protective tissue in the outer bark (Esau, 1965, p. 338). During the ontogeny of the cork cells, polymeric suberin and waxes are deposited in their walls, making them waterimpermeable. Sometimes lignin (Roth, 1981, p. 92) and, as our study shows, silica can be added to the cells walls. Radial arrangement of ± parallelepipedal cells, indicating their origin from periclinal divisions of the phellogen meristematic cells (Esau, 1965, p. 678), was observed in Sapium shirakiopsis (Euphorbiaceae); the cork cells are thin-walled and arranged compactly in regular radial lines (Fig. 6b). The diagnostic Blockies with irregular projections of Stereospermum kunthianum and Kigelia africana (Bignoniaceae) are compact casts of the original cell lumen filled with silica (Fig. 6hek). They are also arranged in a regular parallel pattern, indicating their origin from cork cells. A very special silicified cork occurs in Celtis mildbraedii (Cannabaceae) (Fig. 7g-l). Its cells (¼ Blocky cavate fimbriate) have thick silicified walls with a central convex cavity, and the side opposite the cavity has fimbriate edges interconnecting the cells. We attribute this tissue to cork and not sclerenchyma because of its thick solid cell walls without any pit channels, and the absence of intercellular spaces which is typical for cork (Esau, 1965, p. 340). Cork tissues with similar cells e but non-silicified - have been observed in a few Neotropical tree species (Roth, 1981, pp. 88 ff.).

Please cite this article in press as: Collura, L.V., Neumann, K., Wood and bark phytoliths of West African woody plants, Quaternary International (2016), http://dx.doi.org/10.1016/j.quaint.2015.12.070

12

L.V. Collura, K. Neumann / Quaternary International xxx (2016) 1e18

Fig. 4. Classes D. Elongate and B. Blocky: a, b. Elongate scalariform, Trema guineensis (B); c, d. Elongate facetate, Kigelia africana (B); e. Elongate vesicular, Capparis tomentosa (W); f, g. Elongate with globular parts, Ficus sycomorus (B); hej. Elongate irregular, single and multicellular, Ficus sycomorus (B); Class E. Blockies unspecific: k. Blocky irregular, Lophira lanceolata (B); l. Blocky irregular, Pseudocedrela kotschyi (B); m. Blocky irregular, Antidesma venosum (B); n. Tabular irregular, Kigelia africana (B); o. Tabular irregular, side view, Lonchocarpus laxiflorus (B). (B) ¼ bark, (W) ¼ wood.

A very distinctive morphotype in the bark, occurring in 8.7% of the studied species, consists of thin platelets with rounded or irregular perforations. We attribute these platelets to the cork aerenchyma, an air-filled tissue which is usually an adaptation of trees in humid forests, but can also occur in species of drier habitats (Roth, 1981, p. 229). A similar tissue type can also develop in the filling tissue of the lenticels (Esau, 1965, pp. 348e350). Some of the larger perforations might have developed through breaking down of the cell walls (Fig. 7m). Others with a more regular rounded outline might represent primary intercellular spaces, or air-filled cork cells (Fig. 7n). Runge (1996) found similar perforated platelets in the leaves of Gilbertiodendron dewevrei. We also observed them in fruits of Sarcophrynium macrostachyum (Marantaceae) and leaves of Parinari congensis (Chrysobalanaceae) from the Frankfurt reference collection. 5.3. Taxonomic and ecological relevance of phytoliths in wood and bark More than 90% of the studied species from a representative number of West African plant families deposit silica in the bark. By contrast, only 10% of the woody species worldwide produce silica in the wood (InsideWood, 2004 onwards, http://insidewood.lib.ncsu. edu). Despite its wide occurrence, the greatest part of silica in the bark is deposited in the form of unspecific Blockies and Silica

particles/Accumulations. This type of deposition has no taxonomic relevance, but can be regarded as ‘waste disposal’ of excessive silica without the detour through silification of tissues. The correlation of high silica production with the presence of specific morphotypes indicates that silification in the bark is at least partly under genetic control. However, the high variability of silica content within different bark samples (and also in the wood samples) of one species suggests a strong influence of age and environmental factors in the deposition and accumulation of silica. One third of the studied species show specific morphotypes with consistent morphologies in the bark. Almost half of these species also have phytoliths in their wood (Table 4), indicating that wood and bark silification often go hand in hand and are genetically determined. However, in most plant families phytolith production is unevenly distributed. Chrysobalanaceae are the only family with consistent silica production in wood and bark (Ter Welle, 1976a,b; Metcalfe and Chalk, 1983; InsideWood, 2004 onwards). Most, but not all species of Burseraceae have phytoliths in wood and bark. These two families are generally known as good phytolith producers in which subfamily- and genus-specific forms occur (Group I in Piperno, 2006, p. 7). Moraceae, Urticaceae, and Cannabaceae (formerly p.p. Ulmaceae) also belong to Piperno's Group I, and eight species of these families with specific phytoliths in the bark support their status as

Please cite this article in press as: Collura, L.V., Neumann, K., Wood and bark phytoliths of West African woody plants, Quaternary International (2016), http://dx.doi.org/10.1016/j.quaint.2015.12.070

L.V. Collura, K. Neumann / Quaternary International xxx (2016) 1e18

13

Fig. 5. Class A. Sclerenchyma: a, b. Fibre pitted, Macaranga heudelotii (B); c. Fibre facetate, Kigelia africana (B); d. Fibre psilate, Ficus sycomorus (B); e. Sclereid pitted, Canarium schweinfurtii (B); f. Sclereid pitted, Uapaca guineensis (B); g, h. Sclereid pitted, Canarium schweinfurtii (B); i, j. Sclereid pitted, Uapaca guineensis (B); k. Sclereid pitted, Celtis adolfifridericii (B); l. Sclereid pitted, Klainedoxa gabonensis (B); m. Sclereid pitted, Maranthes polyandra (B); n. Sclereid pitted, Celtis adolfi-fridericii (B); o. Sclereid pitted, Annonidium mannii (B); per. Sclereid psilate, Ficus sycomorus (B); s, t. Sclereid psilate with dendritic projections, Ficus sycomorus. (B) ¼ bark.

good silica producers. Moraceae, Urticaceae, Cannabaceae, and Ulmaceae have recently been grouped together as Urticalean Rosids (Sytsma et al., 2002; Simpson, 2010, p. 336). In addition to the presence of specific phytoliths in their bark, the Urticalean Rosids are also characterized morphologically by the presence of unicellular trichomes on their abaxial leaf surfaces (Sytsma et al., 2002) which often silicify (Piperno, 1988, p. 253 f., authors' observation in the Frankfurt reference collection). Among the Cannabaceae, Celtis is an especially interesting genus. Celtis species produce diagnostic phytoliths in their endocarp and leaves (Piperno, 1989; Bozarth, 1992; Kealhofer and Piperno, 1998), their leaves are heavily

silicified with the typical unicellular trichomes, and some species have diagnostic morphotypes in their bark. The Globular with cylindric projections (Table 6, Fig. 3s and t) is diagnostic for C. adolfi-fridericii, and the conspicuous Blocky cavate (Table 6, Fig. 7gel) for C. mildbraedii. Woody species of the Urticalean Rosids are a very promising group for further phytolith studies. However, it is surprising that the studied Moraceae, Urticaceae and Cannabaceae, including Celtis, do not produce silica in their wood. Generally, silica is rare or absent in the wood of these families (Ter Welle, 1976a; Metcalfe and Chalk, 1983; InsideWood 2004 onwards, http://insidewood.lib.ncsu.edu).

Please cite this article in press as: Collura, L.V., Neumann, K., Wood and bark phytoliths of West African woody plants, Quaternary International (2016), http://dx.doi.org/10.1016/j.quaint.2015.12.070

14

L.V. Collura, K. Neumann / Quaternary International xxx (2016) 1e18

Table 6 Diagnostic morphotypes in bark. Family

Species

Morphotype

Figure

Cannabaceae

Celtis adolfi-fridericii Celtis mildbraedii Kigelia africana Stereospermum kunthianum Boscia angustifolia Boscia salicifolia Boscia senegalensis Maerua crassifolia

Globular with cylindric projections Blocky cavate Blocky with irregular projections

3set 7gel 6hek

Ellipsoid echinate

3kel

Blocky multilayered

7aeb

Bignoniaceae Capparaceae

The example of the Urticalean Rosid shows that correlated occurrence of specific phytoliths in different organs of one and the same plant species is not universal, and there can be considerable discrepancies on the different taxonomic levels. In many families, the pattern of silica production in wood and bark is highly variable between different genera and species. In the Malvaceae,

Euphorbiaceae, Phyllanthaceae, and Meliaceae, some species have specific phytoliths in wood and bark, others only in one organ, and some no specific phytoliths at all. Lannea (Anacardiaceae), Drypetes (Putranjivaceae), and Sapotaceae usually have silica in their wood (Ter Welle, 1976a; Metcalfe and Chalk, 1983; InsideWood, 2004 onwards, http://insidewood.lib.ncsu.edu), but rarely in the bark.

Fig. 6. Class B. Parenchyma/Cork: a. P/C Blocky polyhedral, Kigelia africana (W); b. P/C Blocky parallelepipedal, Sapium ellipticum (B); c. P/C Blocky polyhedral, Morus mesozygia (B); d. P/C Blocky polyhedral, Ficus sycomorus (B); e. Parenchyma strand, Anonidium mannii (B); f. Parenchyma strand, Maranthes polyandra (B); g. Parenchyma strand, Macaranga heudelotii (B); hek. P/C Blocky with irregular projections, Stereospermum kunthianum (B). (B) ¼ bark. Frame solid line: diagnostic morphotypes in bark.

Please cite this article in press as: Collura, L.V., Neumann, K., Wood and bark phytoliths of West African woody plants, Quaternary International (2016), http://dx.doi.org/10.1016/j.quaint.2015.12.070

L.V. Collura, K. Neumann / Quaternary International xxx (2016) 1e18

15

Fig. 7. Classes B. Parenchyma/Cork and C. Cork aerenchyma: a, b, Blocky multilayered, Maerua crassifolia (B); c. Blocky polyhedral, Kigelia africana (B); d. Blocky polyhedral, Morus mesozygia (B); e. Parenchyma strand, single elongate cell, Macaranga heudelotii (B); f. Parenchym strand, single blocky cell; Maranthes polyandra (B); gel, P/C Blocky cavate fimbriate, Celtis mildbraedii (B); Class C. Cork aerenchyma: m. Ficus capensis (B); n. Khaya senegalensis (B). (B) ¼ bark. Frame solid line: diagnostic morphotypes in bark.

Other families show the opposite pattern, i.e. specific phytoliths in the bark, but not in the wood. Conspicuous examples are the Bignoniaceae, Capparaceae, and Annonaceae. Kigelia africana and Stereospermum kunthianum (Bignoniaceae) are top producers of silica in the bark, with values up to 18%. Both species have a diagnostic Blocky with irregular projections (Table 6, Fig. 6hek). In the leaves of K. africana characteristic jigsaw-shaped epidermal cells with a central point-shaped elevation have been observed (Runge, 1996, p. 334). In contrast, Piperno (2006, p. 7) includes Bignoniaceae in the families with uncommon to rare phytolith production (Group V), and silica in wood is also uncommon (Metcalfe and Chalk, 1983, p. 219). The wood of Capparaceae (except for Capparis tomentosa with its Globulars and characteristic Elongate vesicular) is silica-free, but in the bark genus- and species-specific morphotypes occur. The Ellipsoid echinate is present in all Boscia species and can be regarded as diagnostic for the genus (Table 6, Fig. 3kel). The Blocky multilayered is diagnostic for Maerua crassifolia (Table 6, Fig. 7a and b), while the three other studied

Capparaceae species have only unspecific silica in the bark. Annonaceae, as representatives of basal angiosperms, are listed in Piperno's Group I. They are known as good phytolith producers, but they do not have silica in their wood. In our study, only Anonidium mannii shows specific phytoliths in the bark, whereas Annona senegalensis and Uvaria chamae have only Blockies unspecific and Silica particles/Accumulations. This confirms Runge's (1996) statement on the special status of A. mannii; she found the typical faceted phytoliths only in A. mannii, but not in the other two studied Annonaceae species Monodora tenuifolia and Uvaria chamae. Specific bark phytoliths are absent in the three large families Combretaceae, Rubiaceae, and Leguminosae, with the exception of Dialium guineense (Caesalpinioideae). This corresponds with the absence of silica in the wood of Combretaceae and its rare occurrence in the wood of Leguminosae and Rubiaceae (Ter Welle, 1976a; Metcalfe and Chalk, 1983; InsideWood, 2004 onwards, http:// insidewood.lib.ncsu.edu). All Leguminosae species, except for

Please cite this article in press as: Collura, L.V., Neumann, K., Wood and bark phytoliths of West African woody plants, Quaternary International (2016), http://dx.doi.org/10.1016/j.quaint.2015.12.070

16

L.V. Collura, K. Neumann / Quaternary International xxx (2016) 1e18

D. guineense, have very low silica contents in wood and bark, generally <0.5%. The rarity of phytoliths in the three families stands in contrast to their ecological significance. Leguminosae, Rubiaceae and Combretaceae are dominant in many African plant communities, but their wood and bark cannot be expected to leave a considerable phytolith signal in the soils underneath them. The same is true for some families which do neither have silica in wood nor specific phytoliths in bark, among them Celastraceae, Lamiaceae, Ochnaceae, Opiliaceae, Salicaceae, Salvadoraceae, and Rhamnaceae. 5.4. Wood and bark phytoliths in archaeological and palaeoecological contexts Which morphotypes can be regarded as diagnostic for wood and bark in assemblages of isolated phytoliths from modern and ancient sediments and archaeological contexts? Our study shows that wood and bark produce clearly different phytoliths, making the accumulative categories ‘wood/bark’ (Albert et al., 2000, 2006; Albert and Weiner, 2001) and ‘stem’ (Mercader et al., 2009) obsolete. Considering that only 10% of the woody species have silica in their wood and most of the wood phytoliths are Globulars that also occur in other plant organs, the value of phytoliths for detecting wood in unknown assemblages is limited. Nodulars and Granulars occurring in a few species are diagnostic for wood (Table 6, Fig. 3mer) and occur only rarely in other tissues (Baas et al., 1982). But less than half of the wood phytolith producers in tropical Africa, i.e. ca 5% of all woody species, have specific Nodulars or Granulars. We conclude that most woods do not leave any unambiguous phytolith signal and are thus mostly silent in the phytolith record. In temperate regions the phytolith evidence for wood is even weaker because temperate trees and shrubs produce much less silica in their wood than tropical species. Among the 922 wood descriptions for temperate Europe and Eurasia in InsideWood, only five have silica (InsideWood, 2004 onwards, http://insidewood.lib. ncsu.edu). Bark phytoliths have been reported for some temperate woody plants. The irregular blockies found in shrubs of European subalpine and alpine plant communities (Carnelli et al., 2004, polyhedral phytoliths ¼ type POLY) are equivalent to the Blockies unspecific of our study and should be attributed to bark. Carnelli et al. (2004) had processed twigs with their bark adhering, leading to the erroneous assumption that the irregular blockies occur in wood. While silica is deposited only in a limited number of woods, it is almost universally present in bark. However, the most common form of silica deposition in the bark consists of Blockies unspecific and Silica particles/Accumulations. They correspond to the ‘phytoliths with variable morphologies’ and the ‘siliceous aggregates’ found in ash deposits from Palaeolithic cave sites and in modern woody species of the Eastern Mediterranean (Schiegl et al., 1996; Albert and Weiner, 2001; Albert et al., 2000). In contrast to the interpretation of Schiegl et al. (1996), these aggregates are not identical with the Nodulars and Granulars in wood which we found in a limited number of species and which have been described in the wood anatomical studies of Amos (1952), Scurfield et al. (1974), Ter Welle (1976a,b), Richter (1980), and Metcalfe and Chalk (1983). Albert and Weiner, (2001) studied wood and bark of 13 Mediterranean woody species and concluded that the ratio of phytoliths with variable morphologies vs. phytoliths with consistent morphologies is much higher in wood than in bark. Our data, on the contrary, suggest that phytoliths with variable morphologies and unspecific silica accumulations are much more common in bark than in wood and can thus be used as indicators for the presence of bark in unknown assemblages. However, this is only applicable in special contexts where larger amounts of bark might have been

deposited in situ, for example in archaeological rock shelter sites. In sediments and soils with mixed assemblages, it might be difficult to identify bark. The same is valid for the specific bark phytoliths with consistent morphologies. In archaeological sites, accumulations of silicified fibres, sclereids, parenchyma, cork, or cork aerenchyma can indicate deposition of bark from those species that produce specific morphotypes. But in mixed assemblages of palaeoecological or archaeological sites, assignment of single morphotypes to bark is tricky due to their redundancy with those of other plant organs. Parenchyma and cork, fibres and sclereids are not exclusively produced in bark. Long fibres have not been observed in leaves of eudicots, but apart from the bark, they can also occur in the primary tissues of monocots and in the stems of herbaceous dicots which have not been studied comprehensively for their silicified elements. Long silicified fibres are probably not preserved in sediments because they easily break into fragments. Lignified sclereids are widely distributed in different plant tissues (Evert, 2006), and silicified sclereids, especially with irregular or blocky shapes, have also been found in fruits, for example of Uapaca guineensis (authors' observation in the Frankfurt collection). Although reproductive plant parts produce more diagnostic phytoliths than vegetative organs (Piperno, 2006, p. 41), redundancy with phytoliths of fruits cannot be excluded; comprehensive phytolith studies for the fruits of trees and shrubs are not available for tropical Africa and still rare for other parts of the world (e.g. Piperno, 1989; Kealhofer and Piperno, 1998). Despite of redundancy, some general differences can be stated between the morphotypes of dicot leaves and bark which are major producers of phytoliths in tropical forests. Leaves are characterized by the presence of epidermal cells, trichomes, trichome bases, €mberg, mesophyll and tracheary elements (Baas et al., 1982; Stro 2003) which are all absent in bark. The geniculate to stellate nonpitted sclereids which are typical for dicot leaves (Kondo and Peason, 1981) usually do not occur in bark. Assemblages from bark would be characterized by abundant blockies from parenchyma or cork, irregularly shaped sclereids, and fibres. Blockies from bark are also dominant in the ‘stem’ samples from Mozambique (Mercader et al., 2009). Bark sclereids with their dense columellate pitted surface and irregular shape can be distinguished from the terminal tracheids and vessels in leaves which are often elongate and show parallel arrangement of pits or ring- or spiral shaped surface structures due to their function as water-conducting cells (Kondo and Peason, 1981; Postek, 1981; Runge, 1996). Runge's pitted sclereids with very irregular shapes (type C1) which are dominant in soil profiles underneath evergreen rainforest in NE Congo can be tentatively attributed to bark; they are surely neither diagnostic for the fruit of Gilbertiodendron dewevrei as Runge (1999) supposed, nor for wood in general as Piperno (2006, p. 42) suggested. The cork aerenchyma, although only occurring in the bark of a limited number of species, can be regarded as diagnostic for bark because silicified aerenchyma is rare or morphologically different in other plant tissues. The bark aerenchmya can unambiguously be separated from silicified aerenchyma in the culms of Cyperaceae which is composed of stellate cells (Fernandez Honaine et al., 2009, see also; Esau, 1965, p. 184). The perforated platelets of the cork aerenchyma are also clearly different from the opaque platelets with ±regular perforations described by Bozarth (1992) for the inflorescence of Asteraceae. Although the cork aerenchyma only occurs in the bark of a limited number of species, it can be regarded as diagnostic for bark in general because silicified aerenchyma is rare or morphologically different in other plant tissues. In contrast to multicellular phytoliths, isolated Blockies and Elongates resulting from post-depositional fragmentation usually

Please cite this article in press as: Collura, L.V., Neumann, K., Wood and bark phytoliths of West African woody plants, Quaternary International (2016), http://dx.doi.org/10.1016/j.quaint.2015.12.070

L.V. Collura, K. Neumann / Quaternary International xxx (2016) 1e18

cannot be attributed to the tissues they originated from, such as parenchyma/cork and sclerenchyma. Even assignment to dicotyledons or monocotyledons is tricky, especially in assemblages where grasses with their much higher phytolith production are involved. Poaceae display a myriad of highly variable phytoliths, among them Elongates and Blockies which are mostly bulliform cells. Parallelepipedal Blockies from bark and Bulliforms from grasses can look very similar, and identification of these morphotypes is basically an interpretation and depends on the context. In an open landscape dominated by grasses, it is much more probable that Blockies parallelepipedal come from Poaceae and not from the bark of woody plants (Albert et al., 2006). In the case of single phytoliths from small assemblages, assignment of single Blockies, Globulars and Elongates to wood or bark is almost impossible. Because of strong redundancy, the illustrated blocky phytolith of Fig. 3b and the ‘wood/bark’ morphotypes listed in SOM Table 3 of Henry et al. (2012) are not specific for bark, and therefore the statement ‘Australopithecus sediba consumed wood and bark (Henry et al., 2012) is not justified. 6. Conclusions Wood and bark produce very different and clearly discernible phytoliths and should therefore never be regarded as a unit ‘wood/ bark’ or ‘stem’. With a few exceptions, the morphotypes in bark and wood clearly belong to different classes. The bark phytoliths generally reflect the anatomy of the cells. For the morphotype classes Sclerenchyma (with subclasses Fibres and Sclereids), Parenchyma/ Cork, and Cork aerenchyma the tissues in which they were formed can be identified. Isolated Elongates and Blockies cannot always be unambiguously assigned to a tissue, but most of them are probably single elements of sclerenchyma, parenchyma and cork. Wood, in contrast, produces Globulars s.l. and Aggregates that develop their own shape and do not reflect the original cell form. Many more species produce silica in the bark than in the wood and very high silica contents up to 18% are reached in bark. This, together with the ubiquitous presence of unspecific Blockies and Silica particles/Accumulations indicates that the function of silica production in bark is, at least partly, deposition of excess silica without the detour of cell silification. However, high silica production in bark is often correlated with the presence of specific phytoliths with consistent morphologies, i.e. silification of cells and tissues which seems to be under genetic control. These specific phytoliths with consistent morphologies are present in about one third of the studied species. The distribution of specific phytoliths with consistent morphologies is uneven in wood and bark, and there is considerable variation on different taxonomic levels. Species from Burseraceae and Chrysobalanaceae produce phytoliths in wood and bark, others only in wood (e.g. Sapotaceae, Drypetes and Lannea), or only in bark (e.g. Urticalean Rosids, Annonaceae). Urticalean Rosids, Capparaceae, and Bignoniaceae are especially interesting for future taxonomic studies because they were found to produce unique species-, genus- or family-diagnostic phytoliths in the bark. Only 10% of the woody dicotyledons produce phytoliths in their wood. The majority of the wood phytoliths are psilate and decorated Globulars which are not diagnostic for wood because they also occur in other plant parts of dicots and are abundant in monocots. Aggregates (Nodulars and Granulars) are unique and diagnostic for wood, but they are produced only in a few species. We conclude that wood cannot be expected to leave a significant record in soils and sediments, especially in mixed assemblages. In exceptional cases of in-situ decay wood might be identified by the abundant presence of Globulars and/or Aggregates.

17

One of the major objectives of this study was to find out which morphotypes can be regarded as diagnostic for bark. In pure assemblages resulting from in-situ decay of bark or twigs with adhering bark, high amounts of unspecific Blockies and Silica particles/Accumulations are indicators. If species with specific phytoliths with consistent morphologies are involved, concentrations of fibres, sclereids, parenchyma and cork provide additional evidence for bark. In mixed assemblages of soils and sediments identification of bark is much trickier because almost all morphotypes present in bark are fragments of vegetative tissues that also occur in other plant organs of dicot trees and shrubs, as well as in other taxonomic groups. Dicot bark and leaves produce different assemblages of phytoliths, but assigning a single phytolith, e.g. a sclereid, fibre or parenchyma cell to bark or leaves is difficult. Identification of single phytoliths in smaller assemblages, for example from tooth calculus (Henry et al., 2012), are even more problematic than in soils and sediments. Extreme scrutiny is indicated especially if far-reaching archaeological, anthropological or palaeoecological conclusions are inferred from phytolith identification. Several questions remain open. Is production of specific phytoliths with consistent morphologies a general feature of tropical woody plants? Are there differences in bark phytolith production between species from dry and humid habitats and from different plant formations, i.e. savannas and forests? Are patterns of bark silification similar in other tropical regions (SE Asia, Central and South America)? What about silica production in the bark of woody species from temperate areas which rarely have silica in their wood? Our data show that systematic studies of modern phytoliths in their anatomical context are very rewarding. They provide a better and safer basis for identifying phytoliths in unknown assemblages and they can additionally result in unexpected detection of diagnostic morphotypes. Acknowledgements The wood and bark samples of this study have been collected in the frame of several research projects funded by the German Research Foundation (DFG). We thank Jennifer Markwirth for managing the collections and for technical assistance, Monika Heckner for artwork, Manfred Ruppel for assistance with the SEM, € hn for valuable comRichard Byer for language editing, Alexa Ho ments and fruitful discussions, and two anonymous reviewers whose comments contributed greatly to the improvement of the paper. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.quaint.2015.12.070. References Albert, R.M., Weiner, S., 2001. Study of phytoliths in prehistoric ash layers from Kebara and Tabun caves using a quantitative approach. In: Meunier, J.D., Colin, F. (Eds.), Applications in Earth Sciences and Human History. Balkema Publishers, Lisse, Netherlans, pp. 66e251. Albert, R.M., Weiner, S., Bar-Yosef, O., Meignen, L., 2000. Phytoliths in the Middle Paleolithic deposit of Kebara cave, Mt Carmel, Israel: study of the plant materials used for fuel and other purpose. Journal of Archaeological Science 27, 47e931. Albert, R.M., Bamford, M.K., Cabanes, D., 2006. Taphonomy of phytoliths and macroplants in different soils from Olduvai Gorge (Tanzania) and the application to Plio-Pleistocene paleoanthropological samples. Quaternary International 148, 78e94. zine, A.M., Vincens, A., Schwartz, D., 1997. PhytoAlexandre, A., Meuniere, J.D., Le liths: indicators of grasslands dynamics during the late Holocene in intertropical Africa. Palaeogeography, Palaeoclimatology, Palaeoecology 136, 29e213. Amos, S.L., 1952. Silica in timbers. CSIRO Bulletin 267, 1e55.

Please cite this article in press as: Collura, L.V., Neumann, K., Wood and bark phytoliths of West African woody plants, Quaternary International (2016), http://dx.doi.org/10.1016/j.quaint.2015.12.070

18

L.V. Collura, K. Neumann / Quaternary International xxx (2016) 1e18

Baas, P., van Oosterhoud, E., Scholtes, C.J.L., 1982. Leaf anatomy and classification of the Olacaceae, Octoknema, and Erythropalum. Allertonia 3, 155e210. Barboni, D., Bonnefille, R., Alexandre, A., Meuniere, J.D., 1999. Phytoliths as paleoenvironmental indicators, west side Middle Awash Valley, Ethiopia. Palaeogeography, Palaeoclimatology, Palaeoecology 152, 87e100. Barboni, D., Bremond, L., Bonnefille, R., 2007. Comparative studies of modern phytolith assemblages from inter-tropical Africa. Paleogeography, Palaeoclimatology, Palaeoecology 246, 454e470. Bowdery, D., Hart, C., Lentfer, C., Wallis, L.A., 2001. A universal phytolith key. In: Meunier, J.D., Colin, F. (Eds.), Applications in Earth Sciences and Human History. Balkema Publishers, Lisse, pp. 267e278. Bozarth, S., 1992. Classification of opal phytoliths formed in selected dicotyledons native to the Great Plains. In: Rapp Jr., G., Mullholland, S.C. (Eds.), Phytolith Systematics: Emerging Issues. Plenum Press, New York, pp. 193e214. ly, C., Guiot, J., 2005. A phytolith index as a proxy of Bremond, L., Alexandre, A., He tree cover density in tropical areas: calibration with leaf area index along a forest-savanna transec in south-eastern Cameroon. Global and Planetary Change 45, 93e277. Bremond, L., Alexandre, A., Peyron, D., Guiot, J., 2008. Definition of grassland biomes from phytoliths in West Africa. Journal of Biogeography 35, 2039e2048. 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. Chen, S.T., Smith, S., 2013. Phytolith variability in Zingiberales: a tool for the reconstruction of past tropical vegetation. Palaeogeography, Palaeoclimatology, Palaeoecology 370, 1e12. Dominguez-Rodrigo, M., Serrallonga, J., Tresserras-Juan, J., Alcala, L., Luque, L., 2001. Woodworking activities by early humans: a plant residue analysis on Acheulien stone tools from Peninj (Tanzania). Journal of Human Evolution 40, 289e299. Esau, C., 1965. Plant Anatomy, second ed. John Wiley & Sons, New York. Evert, R.F., 2006. Esaus Pflanzenanatomie. Meristeme, Zellen und Gewebe der Pflanzen e Ihre Struktur, Funktion und Entwicklung. Walter de Gruyter Verlag, Berlin. Fernandez Honaine, M., Zucol, A., Osterrieth, M., Del Rio, L., 2009. Phytolith analysis of Cyperaceae from the Pampean region, Argentina. Australian Journal of Botany 57, 512e523. Garnier, A., Neumann, K., Eichhorn, B., 2013. Phytolith taphonomy in the middle to late Holocene fluvial sediments of Ounjougou (Mali, West Africa). The Holocene 23, 416e431. Henry, A.G., Ungar, P.S., Passey, B.H., Sponheimer, M., Rossouw, L., Bamford, M., Sandberg, P., de Ruiter, D.J., Berger, L., 2012. The diet of the Australopithecus sediba. Nature 487, 90e93. ICPN Working Group, Madella, M., Alexandre, A., Ball, T., 2005. International code for phytolith nomenclature 1.0. Annals of Botany 96, 253e260. http://insidewood.lib.ncsu.edu (17.08.15.). Iriarte, J., Paz, E.A., 2009. Phytolith analysis of selected native plants and modern soils from south-eastern Uruguay and its implications for paleoenvironmental and archaeological reconstruction. Quaternary International 193, 99e123. Kealhofer, L., Piperno, D.R., 1998. Opal Phytoliths in Southeast Asian Flora. Smithsonian Contributions to Botany, vol. 88. Smithsonian Institution Press, Washington D.C. Kondo, R., Peason, T., 1981. Opal phytoliths in trees leaves (part 2): opal phytoliths in dicotyledon angiosperm tree leaves. Research Bulletin of Obihiro University 12, 217e229. In Japanese, with English abstract. Kondo, R., Childs, C., Atkinson, I., 1994. Opal Phytoliths of New Zealand. Manaaki Press, Lincoln, New Zealand. 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., Chalk, L., 1983. Anatomy of the Dicotyledons, second ed. Oxford University Press, London. Neumann, K., Fahmy, A., Lespez, L., Ballouche, A., Huysecom, E., 2009. The Early Holocene palaeoenvironment of Ounjougou (Mali): phytoliths in a multiproxy context. Palaeogeography, Palaeoclimatology, Palaeoecology 276, 87e106.

€ Ostlund, L., Ahlberg, L., Zachrisson, O., Bergman, I., Arno, S., 2009. Bark-peeling, food stress and tree spirits e the use of pine inner bark in Scandinavia and North America. Journal of Ethnobiology 29, 94e112. Pettersen, R.C., 1984. The chemical composition of wood. In: Rowell, R.M. (Ed.), The Chemistry of Solid Wood. Advances in Chemistry Series, vol. 207. American Chemical Society, Washington, D.C, pp. 57e126. Piperno, D., 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 selected tropical angiosperms and their significance in tropical palaeoecology, palaeoethnobotany and systematics. Review of Palaeobotany and Palynology 61, 147e173. Piperno, D., 2006. Phytoliths. A Comprehensive Guide for Archaeologists and Paleoecologists. Altamira Press, Lanham. Postek, M.T., 1981. The occurrence of silica in the leaves of Magnolia grandiflora L. Botanical Gazette 142, 124e134. Richter, H.G., 1980. Occurrence, morphology and taxonomic implications of crystalline and siliceous inclusions in the secondary xylem of the Lauraceae and related families. Wood Science and Technologies 14, 34e44. Roth, I., 1981. Structural Patterns of Tropical Barks. Gebrüder Borntraeger Verlag, Würzburg. Runge, F., 1996. Opal Phytolithe in Pflanzen aus dem humiden und semi-ariden Osten Afrikas und ihre Bedeutung für die Klima- und Vegetationsgeschichte. Botanische Jahrbücher für Systematik 188, 303e363. Runge, F., 1999. The opal phytolith inventory of soils in Central Africa  Quantities, shapes, classification and spectra. Review of Paleobotany and Palynology 107, 23e53. Runge, F., 2000. Opal-Phytolithe in den Tropen Afrikas und ihre Verwendung bei € oo € kologischer Umweltverha €ltnisse. Books on Demand der Rekonstruktion pala Verlag, Norderstedt. Schiegl, S., Goldberg, P., Bar-Yosef, O., Weiner, S., 1996. Ash deposits in Hayonim and Kebara caves, Israel: macroscopic, microscopic and mineralogical observations, and their archaeological implications. Journal of Archaeological Science 23, 763e778. Scurfield, G., Anderson, C.A., Segnit, E.R., 1974. Silica in woody stems. Australian Journal of Botany 22, 29e211. Simpson, M.G., 2010. Plant Systematics, second ed. Elsevier Academic Press. Stearn, W.T., 2008. Botanical Latin, fourth ed. David & Charles Publishers, Davon. €mberg, K., 2003. The Origin and Spread of Grass-dominated Ecosystem during Stro the Tertiary of North America and How it Relates to the Evolution of Hypsodonty in Equids. PhD thesis. University of California, Berkeley, USA. €mberg, K., 2004. Using phytolith assemblages to reconstruct the origin and Stro spread of grass-dominated habitats in the Great Plains during the late Holocene to early Miocene. Palaeogeography, Palaeoclimatology, Palaeoecology 207, 75e239. Swetnam, T., 1984. Peeled ponderosa pine trees: a record of inner bark utilization by Native Americans. Journal of Ethnology 4, 177e190. Sytsma, K.J., Morawetz, J., Pires, J.C., Nepokroeff, M., Conti, E., Zjhira, M., Hall, J.C., Chase, M.W., 2002. Urticalean Rosids. Circumscription, rosid ancestry and phylogenetics based on rbcL, trnL-F and ndhF sequences. American Journal of Botany 89, 1531e1546. Ter Welle, B.J.H., 1976a. Silica grains in woody plants of the Neotropics, especially Surinam. Leiden Botanical Series 3, 42e107. Ter Welle, B.J.H., 1976b. On the occurrence of silica grains in the secondary xylem of the Chrysobalanaceae. International Association of Wood Anatomists Bulletin 2, 19e29. http://www.theplantlist.org/ (11.08.15.). Trockenbrodt, M., 1990. Survey and discussion of the terminology used in bark anatomy. International Association of Wood Anatomists Bulletin New Series 2, 141e166. Wheeler, E.A., 2011. Inside Wood - A web resource for hardwood anatomy. IAWA Journal 32, 199e211. Wheeler, E.A., Baas, P., Rodgers, S., 2007. Variations in dicot wood anatomy: a global analysis based on the InsideWood database. International Association of Wood Anatomists Journal 3, 229e258.

Please cite this article in press as: Collura, L.V., Neumann, K., Wood and bark phytoliths of West African woody plants, Quaternary International (2016), http://dx.doi.org/10.1016/j.quaint.2015.12.070