Phytoliths in dicotyledons occurring in Northwest Europe: establishing a baseline

Rosalie Hermans; Caroline A E Strömberg; Tessi Löffelmann; Luc Vrydaghs; Lien Speleers; Alexandre Chevalier; Karin Nys; Christophe Snoeck

doi:10.1093/aob/mcae217

. 2024 Dec 16;135(5):885–908. doi: 10.1093/aob/mcae217

Phytoliths in dicotyledons occurring in Northwest Europe: establishing a baseline

Rosalie Hermans ^1,^✉, Caroline A E Strömberg ², Tessi Löffelmann ³, Luc Vrydaghs ⁴, Lien Speleers ⁵, Alexandre Chevalier ⁶, Karin Nys ⁷, Christophe Snoeck ⁸

PMCID: PMC12064429 PMID: 39680404

Abstract

Background and Aims

The absence of a modern plant-based ‘dicotyledon’ phytolith reference baseline impedes the accurate interpretation of fossil phytolith records in archaeological and palaeoecological research within Northwest (NW) Europe. This study aims to fill this gap by documenting and analysing the phytolith record from modern dicotyledon taxa occurring in this region.

Methods

Phytoliths were extracted from several plant parts of 117 plant specimens representing 74 species (one or two specimens per species). The study employed light microscopy to examine phytolith production (non-producer, trace, common or abundant) and phytolith assemblage composition. The data were analysed statistically to (1) determine the influence of taxonomy and plant part on phytolith presence (absent/present) using a mixed model, (2) assess phytolith assemblage variation using a permutational multivariate analysis of variance (PerMANOVA) and (3) identify patterns among sample groups including segregation for plant part, life form (forbs vs. shrubs/trees) and order using linear discriminant analyses (LDA).

Key Results

Morphotype analysis revealed diagnostic morphotypes and features for specific plant families, genera and plant parts. LDA effectively segregated plant parts and life forms, though taxonomic groupings showed limited segregation. Phytolith presence (absent/present) was found to vary, influenced by both plant part and taxonomy. For species examined through two specimens, although phytolith production varied considerably, phytolith assemblage composition was consistent.

Conclusions

This study establishes a ‘dicotyledon’ phytolith baseline for NW Europe, showing that the phytolith record can be informative in terms of plant part and life form and that several phytolith morphotypes and/or features are taxonomically diagnostic below ‘dicotyledon’ level. The findings constitute a foundation upon which future research can build, refining and expanding our knowledge of the NW European region.

Keywords: Phytoliths, reference collection, dicotyledons, northwestern Europe

INTRODUCTION

Phytoliths are opal silica bodies that can accumulate in intra- and inter-cellular spaces or on cell walls of plant tissues. They take on a wide variety of shapes. They are formed when a plant absorbs monosilicic acid Si(OH)₄ from the soil and deposits opaline silica (SiO₂.nH₂O) within its different organs and tissues, such as roots, stems, leaves, inflorescence bracts, seed bracts and fruit rinds (Bozarth, 1992; Sangster and Hodson, 1992; Piperno, 2006). When the plant decays, its phytoliths are released into the environment, where they may be preserved for thousands or even millions of years depending on their morphology and chemical composition and the environmental conditions (Bartoli and Wilding, 1980; Meunier et al., 1999; Cabanes and Shahack-Gross, 2015). The resulting phytolith assemblages serve as unique indicators of past plant communities in archaeological and palaeoenvironmental settings that do not preserve organic material (Piperno, 2006).

In Northwest (NW) Europe, phytolith analysis has emerged as a prominent tool in Quaternary palaeoecology and archaeobotany only in the last few decades. Phytoliths are commonly studied from soils and sediments (e.g. Gebhardt et al., 2023; Trant et al., 2024), but they are also retrieved from ecofacts and artefacts such as sherd residues (e.g. Hart, 2011), stone artefacts (e.g. Risberg et al., 2002), dental samples (e.g. Prado and Noble, 2024) and faeces (e.g. Powers et al., 1989; Persaits et al., 2015). They are used to investigate the palaeoenvironment (e.g. Wade et al., 2021) and to address a wide range of archaeological topics, including anthropogenic plant use (Saul et al., 2013), spatial organization at archaeological sites (e.g. Borderie et al., 2020), fuel use (e.g. Braadbaart et al., 2017; Power, 2018; Dekoninck et al., 2024), crop cultivation (Devos et al., 2022), burial practices (Out et al., 2022) and urbanization processes (Wouters et al., 2019) in the region.

The surging interest in phytolith research has led to an augmented demand for comprehensive phytolith reference collections. These collections play a pivotal role in establishing contemporary baseline data, enhancing our understanding of modern phytolith assemblages, and aiding in the identification and quantification of fossil phytolith records (Bozarth, 1992; Tsartsidou et al., 2007; Morris et al., 2009; Pearce and Ball, 2020; Piperno and McMichael, 2020). In NW Europe, such a baseline is lacking. Although previous studies of phytolith assemblages from modern plants have documented taxa also occurring in NW Europe, there has been no systematic focus on the flora specific to this region (Kealhofer and Piperno, 1998; Albert, 2000; Tsartsidou et al., 2007; Morris et al., 2009; Pearce and Ball, 2020; Castilla‑Beltran et al., 2024).

In this study, we focus on non-monocotyledonous angiosperm (hereafter ‘dicotyledons’) that occur in NW Europe. We are interested in understanding the phytolith record of these taxa, including information on phytolith production level, taxonomically diagnostic morphotypes and phytolith assemblage composition. Our research specifically targets dicotyledons. While most phytolith reference studies concentrate on diagnostic phytoliths from monocotyledons such as grasses and palms, there is a lack of published research on dicotyledons, which tend to have fewer phytoliths that are ‘diagnostic’ at a low taxonomic level (Strömberg et al., 2018).

The plants included in this study were selected to support the expanding field of phytolith research in Belgium, which has helped identify past human activities such as pasturing, crop cultivation and manuring during medieval times, as well as the diachronous evolution of the urban landscape (Devos et al., 2009, 2013a, b, 2017; Vrydaghs et al., 2016). Taxa include a wide range of human-used plants – fruits, nuts, vegetables, herbs, oil and fibre plants, as well as various weedy and wild species (Table 1). While these taxa are found in multiple medieval or post-medieval contexts in Brussels (Speleers and van der Valk, 2017), their frequent occurrence in other regions of NW Europe across various time periods makes this study of general importance to many palaeo- and archaeobotanists (Bakels, 1999; Brombacher and Hecker, 2015; Moffett, 2018; Hald et al., 2020; Helweg, 2020; Hondelink and Schepers, 2020). At the most fundamental level, the dataset advances our general understanding of phytolith production and assemblage composition within dicotyledons, which has both ecological and evolutionary implications (see e.g. Katz, 2015; Strömberg et al., 2016).

Table 1.

Sample table including information on family, order, superorder, species, specimen, life form, phytolith production level per sample, accession number, herbarium location, country of origin and laboratory protocol (LP).

Family (Order/Superorder)	Specimen	Life form	Plant parts^{production level}	Accession number	Herbarium	Climate zone of specimen	Country of origin	LP
Amaranthaceae (Caryophyllales/Caryophyllanae)	Atriplex patula	Forb	R(I)^NP, HS^NP, LE^C	253557	WTU	Dfb	Finland	2
	Beta vulgaris	Forb	R(I)^NP, HS^NP, LE^NP	46366	WTU	Csb	Denmark	1/2
	Beta vulgaris subsp. vulgaris	Forb	R(I)^NP, HS^NP, LE^NP	NA	NY	Cfb	USA	2
	Chenopodium album (1)	Forb	R(I)^NP, HS^NP, LE^T	427105	WTU	Csb	USA	2
	Chenopodium album (2)	Forb	R(I)^T, HS^NP, LE^C	400469	WTU	Csb	USA	2
	Lipandra polysperma	Forb	R(I)^NP, HS^NP, LE^NP	263189	WTU	Csb	USA	2
Apiaceae (Apiales/Asteranae)	Aethusa cynapium (1)	Forb	R(I)^NP, HS^NP, LE^NP	NA	NY	Dfb	Sweden	2
	Aethusa cynapium (2)	Forb	R(I)^NP, HS^NP, LE^NP	279501	WTU	Cfb	Denmark	1
	Anethum graveolens	Forb	R(S)^NP, HS^NP, LE^T	03519865	NY	NA	Unknown	2
	Apium graveolens (1)	Forb	R(I)^NP, HS^NP, LE^T	63573	NY	Cfb	the Netherlands	2
	Apium graveolens (2)	Forb	HS^NP, LE^C	NA	NY	Cfb	Germany	2
	Coriandrum sativum (1)	Forb	R(S)^NP, HS^NP, LE^NP	03455721	NY	Cfa	USA	2
	Coriandrum sativum (2)	Forb	R(S)^NP, R(I)^NP, HS^NP, LE^NP	144964	WTU	Csb	USA	1
	Daucus carota	Forb	R(I)^T, HS^NP, LE^T	279733	WTU	Dfb	Poland	2
	Petroselinum crispum	Forb	LE^NP, HS^NP	422275	WTU	Csb	USA	2
Asteraceae (Asterales/Asteranae)	Anthemis cotula (1)	Forb	R(I)^NP, HS^NP, LE^C	158545	WTU	Cfb	the Netherlands	2
	Anthemis cotula (2)	Forb	R(I)^C, HS^NP, LE^C	411102	WTU	BSk	USA	1/2
	Bidens cernua (1)	Forb	R(I)^C, HS^NP, LE^A	250875	WTU	Cfb	Germany	2
	Bidens cernua (2)	Forb	R(I)^T, HS^NP, LE^C	413090	WTU	Dfb	USA	2
	Centaurea cyanus (1)	Forb	R(I)^T, HS^NP, LE^C	225582	WTU	Cfb	Denmark	2
	Centaurea cyanus (2)	Forb	R(I)^T, HS^NP, LE^C	279349	WTU	Cfb	Germany	2
	Cirsium arvense (1)	Forb	R(I)^C, HS^NP, LE^A	279782	WTU	Dfb	Poland	2
	Cirsium arvense (2)	Forb	R(I)^C, HS^NP, LE^A	NA	WTU	Cfb	Austria	2
	Lapsana communis (1)	Forb	R(I)^T, HS^NP, LE^T	181514	WTU	Cfb	the Netherlands	2
	Lapsana communis (2)	Forb	R(I)^NP, HS^T, LE^T	NA	WTU	Cfb	Austria	1/2
	Sonchus asper	Forb	R(I)^C, HS^NP, LE^A	158547	WTU	Cfb	the Netherlands	2
Betulaceae (Fagales/Rosanae)	Corylus avellana (1)	Shrub/tree	R(F)^NP, R(I)^C, WS^NP, LE^A	158416	WTU	Cfb	the Netherlands	2
Betulaceae (Fagales/Rosanae)	Corylus avellana (2)	Shrub/tree	R(F)^C, WS^NP, LE^A	17141	WTU	Cfb	(Northern) France	2
Boraginaceae (Boraginales/Asteranae)	Myosotis scorpioides (1)	Forb	HS^T, LE^A	NA	NY	Cfb	Switzerland	2
Boraginaceae (Boraginales/Asteranae)	Myosotis scorpioides (2)	Forb	HS^NP, LE^A	NA	NY	Cfb	Austria	2
Brassicaceae (Brassicales/Rosanae)	Brassica rapa (1)	Forb	R(I)^NP, HS^NP, LE^T	221723	WTU	Dfc	Finland	2
	Brassica rapa (2)	Forb	R(F)^NP, R(I)^NP, HS^NP, LE^C	3536656	WTU	Cfb	USA	1/2
	Capsella bursa-pastoris (1)	Forb	R(I)^NP, HS^NP, LE^NP	279525	WTU	Cfb	Germany	2
	Capsella bursa-pastoris (2)	Forb	R(I)^NP, HS^NP, LE^T	72615	WTU	Csb	USA	1/2
	Raphanus raphanistrum (1)	Forb	R(F)^C, HS^NP, LE^C	NA	NY	Cfb	Germany	2
	Raphanus raphanistrum (2)	Forb	R(F)^NP, R(I)^NP, HS^NP, LE^C	334142	WTU	Csb	USA	1/2
	Rhamphospermum nigrum	Forb	R(I)^T, HS^NP, LE^C	328295	WTU	Csb	USA	2
Cannabaceae (Rosales/Rosanae)	Cannabis sativa	Forb	R(S)^C, HS^C, LE^A	01839809	NY	Cfa	USA	2
Caryophyllaceae (Caryophyllales/Caryophyllanae)	Agrostemma githago	Forb	R(I)^T, HS^NP, LE^C	319228	WTU	ET	Switzerland	2
	Silene flos-cuculi	Forb	R(I)^NP, HS^NP, LE^T	NA	NY	Cfb	Sweden	2
	Stellaria media	Forb	HS^NP, LE^NP	404648	WTU	Csb	USA	2
Fabaceae (Fabales/Rosanae)	Lathyrus oleraceus	Forb	R(F)^NP, HS^NP, LE^NP	50076	WTU	Cfa	Italy	2
Fabaceae (Fabales/Rosanae)	Vicia tetrasperma	Forb	R(F)^NP, HS^NP, LE^NP	225374	WTU	Cfb	Germany	1
Juglandaceae (Fagales/Rosanae)	Juglans regia	Forb	R(F)^NP, WS^NP, LE^A	361409	WTU	Cfb	USA	2
Lamiaceae (Lamiales/Asteranae)	Lamium maculatum	Forb	R(I)^NP, HS^NP, LE^NP	NA	NY	Cfb	Czechia	2
	Lycopus europaeus	Forb	HS^NP, LE^C	250616	WTU	Dfb	Germany	2
	Prunella vulgaris	Forb	R(I)^NP, HS^NP, LE^T	321057	WTU	Cfb	USA	2
Linaceae (Malpighiales/Rosanae)	Linum usitatissimum	Forb	HS^NP, LE^C	325769	WTU	Dfb	(Northern) France	2
Moraceae (Rosales/Rosanae)	Ficus carica (1)	Shrub/tree	R(F)^C, WS^A, LE^A	01042437	NY	Cfa	Georgia	2
	Ficus carica (2)	Shrub/tree	R(F)^C, WS^C, LE^A	6308	WTU	Dfb	USA	1
	Morus nigra (1)	Shrub/tree	R(F)^C, WS^A, LE^C	NA	NY	Cfa	Georgia	2
	Morus nigra (2)	Shrub/tree	WS^C, LE^C	48660	WTU	Dfb	USA	2
Papaveraceae (Ranunculales/Ranunculanae)	Chelidonium majus (1)	Forb	HS^NP, LE^NP	241761	WTU	Cfb	Finland	2
	Chelidonium majus (2)	Forb	R(I)^NP, HS^NP, LE^NP	232317	WTU	Cfb	Switzerland	1
	Fumaria officinalis (1)	Forb	HS^NP, LE^NP	325411	WTU	Csb	Switzerland	2
	Fumaria officinalis (2)	Forb	R(I)^NP, HS^NP, LE^NP	337879	WTU	Cfb	USA	2
	Papaver argemone	Forb	R(F)^A, HS^C, LE^A	NA	NY	Dfb	Denmark	2
	Papaver somniferum (1)	Forb	R(F)^C, HS^T, LE^C	01414452	NY	Csb	Canada	2
	Papaver somniferum (2)	Forb	R(F)^C, R(I)^C, HS^NP, LE^A	408182	WTU	Cfb	USA	1/2
Plantaginaceae (Lamiales/Asteranae)	Plantago major	Forb	R(I)^T, HS^NP, LE^C	280004	WTU	Dfc	Denmark	2
Polygonaceae (Caryophyllales/Caryophyllanae)	Persicaria hydropiper	Forb	HS^NP, LE^T	221587	WTU	Cfb	Finland	2
	Persicaria lapathifolia (1)	Forb	R(I)^NP, HS^NP, LE^T	229689	WTU	Cfb	Denmark	2
	Persicaria lapathifolia (2)	Forb	R(I)^NP, HS^NP, LE^NP	NA	WTU	Dfc	Austria	1
	Persicaria maculosa (1)	Forb	HS^NP, LE^T	338194	WTU	Cfb	Russia	2
	Persicaria maculosa (2)	Forb	R(I)^NP, HS^NP, LE^T	NA	WTU	Dwb	Austria	1/2
	Polygonum aviculare	Forb	HS^NP, LE^NP	337014	WTU	Csb	Russia	2
	Rumex acetosella (1)	Forb	R(I)^NP, HS^NP, LE^T	403509	WTU	Csb	USA	2
	Rumex acetosella (2)	Forb	R(I)^NP, HS^NP, LE^NP	141701	WTU	Cfb	USA	1
	Rumex conglomeratus (1)	Forb	R(I)^NP, HS^NP, LE^NP	231045	WTU	Cfb	Switzerland	2
	Rumex conglomeratus (2)	Forb	R(I)^T, HS^NP, LE^C	116510	WTU	BSk	Sweden	2
	Rumex crispus (1)	Forb	R(I)^NP, HS^NP, LE^NP	211154	WTU	Csb	USA	2
	Rumex crispus (2)	Forb	R(I)^NP, HS^NP, LE^C	388837	WTU	Cfb	USA	1/2
	Rumex obtusifolius (1)	Forb	R(I)^NP, HS^NP, LE^NP	330010	WTU	Cfb	Switzerland	2
	Rumex obtusifolius (2)	Forb	R(I)^NP, HS^NP, LE^NP	419839	WTU	Cfb	USA	1
	Rumex sanguineus	Forb	R(I)^C, HS^NP, LE^C	NA	WTU	NA	Unknown	2
Primulaceae (Ericales/Asteranae)	Lysimachia arvensis (1)	Forb	HS^NP, LE^NP	329576	WTU	Cfb	Switzerland	2
Primulaceae (Ericales/Asteranae)	Lysimachia arvensis (2)	Forb	R(I)^NP, HS^NP, LE^NP	240778	WTU	Cfb	Denmark	2
Ranunculaceae (Ranunculales/Ranunculanae)	Ranunculus repens (1)	Forb	HS^NP, LE^C	NA	NY	Cfb	Poland	2
	Ranunculus repens (2)	Forb	HS^NP, LE^C	02629303	NY	Dfb	Armenia	2
	Ranunculus sardous (1)	Forb	HS^NP, LE^C	63659	NY	Cfb	the Netherlands	2
	Ranunculus sardous (2)	Forb	R(I)^C, HS^NP, LE^C	241998	WTU	Cfa	USA	2
	Ranunculus sceleratus (1)	Forb	R(I)^T, HS^NP, LE^C	329848	WTU	Dfc	USA	2
	Ranunculus sceleratus (2)	Forb	R(F)^NP, HS^NP, LE^T	230810	WTU	Cfb	Denmark	1/2
Resedaceae (Brassicales/Rosanae)	Reseda luteola	Forb	R(I)^T, HS^NP, LE^C	427418	WTU	Csb	USA	2
Rosaceae (Rosales/Rosanae)	Crataegus germanica (1)	Shrub/tree	R(I)^NP, WS^NP, LE^T	168185	WTU	Cfb	(Northern) France	2
	Crataegus germanica (2)	Shrub/tree	R(F)^NP, WS^NP, LE^T	NA	NY	Cfb	Spain	2
	Fragaria vesca (1)	Forb	R(I)^C, HS^NP, LE^C	103476	WTU	Cfb	(Northern) France	2
	Fragaria vesca (2)	Forb	R(F)^NP, HS^NP, LE^C	241612	WTU	Dfb	Finland	2
	Malus sylvestris	Forb	R(I)^T, WS^NP, LE^T	03263404	NY	Dfb	Finland	2
	Prunus avium (1)	Shrub/tree	R(I)^T, WS^NP, LE^NP	427971	WTU	Csb	USA	2
	Prunus avium (2)	Shrub/tree	R(I)^T, WS^NP, LE^C	424306	WTU	Csb	USA	2
	Prunus cerasus (1)	Shrub/tree	R(I)^NP, WS^NP, LE^NP	354723	WTU	Csb	USA	2
	Prunus cerasus (2)	Shrub/tree	R(S)^NP, R(I)^C, WS^NP, LE^NP	352647	WTU	Csb	USA	1/2
	Prunus domestica (1)	Shrub/tree	R(F)^NP, WS^NP, LE^C	404736	WTU	Csb	USA	2
	Prunus domestica (2)	Shrub/tree	R(F)^NP, WS^NP, LE^T	352650	WTU	Csb	USA	1/2
	Prunus spinosa (1)	Shrub/tree	R(F)^NP, WS^NP, LE^C	424265	WTU	Csb	USA	2
	Prunus spinosa (2)	Shrub/tree	R(F)^NP, WS^NP, LE^C	218377	WTU	Dfb	Russia	1/2
	Pyrus communis	Shrub/tree	R(F)^NP, WS^NP, LE^C	423989	WTU	Csb	USA	2
	Rubus caesius (1)	Shrub/tree	R(F)^NP, WS^NP, LE^NP	NA	WTU	Dfb	Sweden	2
	Rubus caesius (2)	Shrub/tree	WS^NP, LE^NP	251771	WTU	Cfb	Sweden	2
	Rubus fruticosus (1)	Shrub/tree	R(F)^NP, WS^NP, LE^NP	NA	NY	Cfb	Switzerland	2
	Rubus fruticosus (2)	Shrub/tree	R(I)^NP, WS^NP, LE^T	67378	WTU	Cfb	Sweden	2
	Rubus idaeus	Shrub/tree	R(I)^NP, WS^NP, LE^C	295675	WTU	Csb	USA	2
Solanaceae (Solanales/Asteranae)	Alkekengi officinarum (1)	Forb	R(F)^T, HS^NP, LE^C	328761	WTU	Dfc	Switzerland	2
	Alkekengi officinarum (2)	Forb	R(F)^NP, HS^NP, LE^C	NA	NY	Cfa	USA	2
	Hyoscyamus niger (1)	Forb	R(F)^T, HS^NP, LE^C	328772	WTU	ET	Switzerland	2
	Hyoscyamus niger (2)	Forb	R(F)^NP, HS^NP, LE^NP	269520	WTU	Dfb	USA	2
	Solanum dulcamara (1)	Forb	R(F)^NP, HS^NP, LE^C	NA	WTU	Dfb	Finland	2
	Solanum dulcamara (2)	Forb	R(I)^NP, HS^NP, LE^C	328711	WTU	ET	Switzerland	2
	Solanum nigrum (1)	Forb	R(F)^NP, HS^NP, LE^C	159150	WTU	Cfb	Switzerland	2
	Solanum nigrum (2)	Forb	R(F)^NP, HS^NP, LE^C	NA	WTU	Cfb	the Netherlands	2
Urticaceae (Rosales/Rosanae)	Urtica dioica subsp. dioica	Forb	R(I)^C, HS^C, LE^A	273553	WTU	Csb	USA	2
Urticaceae (Rosales/Rosanae)	Urtica urens	Forb	R(I)^C, HS^T, LE^C	8750	WTU	Cfb	USA	2
Viburnaceae (Dipsacales/Asteranae)	Sambucus ebulus	Forb	WS^NP, LE^C	01040914	NY	Dfb	USA	2
Viburnaceae (Dipsacales/Asteranae)	Sambucus nigra	Shrub/tree	R(I)^NP, WS^NP, LE^C	NA	NY	Dfb	Finland	2
Vitaceae (Vitales/Rosanae)	Vitis vinifera (1)	Shrub/tree	R(F)^NP, WS^NP, LE^A	03097763	NY	Cfb	Austria	2
Vitaceae (Vitales/Rosanae)	Vitis vinifera (2)	Shrub/tree	R(F)^NP, WS^NP, LE^A	369640	WTU	Csb	USA	1/2

Open in a new tab

NP = non-producer, T = trace producer, C = common producer, A = abundant producer, NA = not applicable.

R(I) = reproductive part (inflorescences), R(S) = reproductive part (seeds), R(F) = reproductive part (fruits), WS = woody stem, HS = herbaceous stem, LE = leaf.

BSk = cold semi-arid climate, Cfa = humid subtropical climate, no dry season, hot summer, Cfb = oceanic climate, no dry season, warm summer, Csb = Mediterranean climate, warm summer, Dfb = humid continental climate, no dry season, warm summer, Dfc = subarctic climate, no dry season, cool summer, Dwb = monsoon-influenced humid continental climate, dry-winter, warm summer, ET = tundra climate (following Köppen-Geiger climate classification; Beck et al., 2018).

Laboratory protocols are designated as ‘1’ for wet oxidation, ‘2’ for a combination of dry and wet oxidation, and ‘1/2’ indicating the use of both protocols for different samples within a specimen.

The study has three main goals. The first is to assess the taxonomically diagnostic value of phytoliths in the selected plant taxa. Since many phytoliths produced by dicotyledons exhibit redundant morphologies across different plant taxa (Rovner, 1971), we will focus on both individual morphotypes and the relative abundances of morphotypes within phytolith assemblages. By doing so, we seek to enhance the precision with which dicotyledons can be identified and differentiated. The second aim is to assess whether phytolith assemblages can be used effectively to segregate different groupings, namely taxonomic groups (e.g. families, genera), life forms (shrubs/trees and forbs) and plant parts (fruits, seeds, inflorescences, stems and leaves). Achieving separation within each of these groupings would vastly enhance the information that is provided by phytolith analysis. For example, identifying plant parts that serve distinct functions can offer valuable insights into past plant use and human activities, such as dietary practices or the specific functions of archaeological sites (e.g. areas used for food processing). This approach has been successfully applied to monocotyledons, as demonstrated by Harvey and Fuller (2005), who used phytolith assemblages to identify crop by-products such as straw and husks. The third and final aim is to investigate phytolith production levels in dicotyledonous taxa, addressing concerns that certain taxa or plant parts may produce fewer phytoliths, potentially biasing the fossil phytolith record (Albert and Weiner, 2001). To do so, we examine phytolith production levels within species (using two specimens per species for selected taxa) and across different plant parts. Taken together, these results provide a uniquely detailed view of the phytolith production and assemblage composition of important trees, shrubs and forbs in NW Europe.

MATERIALS AND METHODS

Selection of plant taxa and specimen

In this study, extant dicotyledonous taxa were selected based on their numerous occurrences (n ≥ 3) in the carpological record of archaeological deposits in Brussels, Belgium (Devos et al., 2017; Speleers and van der Valk, 2017). Taxonomic nomenclature follows Plants of the World Online (POWO, 2023). Dried plant specimens were sourced from the University of Washington Herbarium (WTU) and the William and Linda Steere Herbarium of the New York Botanical Garden (NY). Although these herbaria are located in the USA, they house a large number of species originating from NW Europe, making them well-suited sources for this study. Of the total 117 specimens analysed, 58 were procured from NW Europe (the Netherlands, northern France, Germany, Denmark, Sweden, Switzerland, Finland, Iceland, Austria). The remaining 60 specimens were from other regions, as detailed in Table 1. This selection strategy was employed to reveal broader patterns of phytolith production and assemblage formation across a range of environmental conditions. Our hypothesis is that, consistent with previous studies, the diverse environmental origins of the specimens will impact phytolith production but will have a limited influence on the morphotypes and their broad patterns of relative abundance (Amos, 1952; Lanning, 1960; Iriarte and Paz, 2009; Liu et al., 2016; Brightly et al., 2020; Ge et al., 2020).

A total of 74 species from 24 plant families, which included 60 forbs and 14 shrubs/trees (Table 1), were studied. For a subset of the taxa, specifically 43 species, we were able to analyse two specimens to address intra-specific variation (Mulholland et al., 1990). ‘Duplicated species’ refer to two specimens of the same species, each independently collected, such as two Anthemis cotula specimens from separate herbarium sheets. ‘Duplicate samples’ are matching plant parts from duplicate species, such as two leaf samples coming from two individual Anthemis cotula specimens. Samples were typically gathered in spring or summer (Supplementary Data File 1 – Methods S1).

The study comprises 338 samples in total: 117 leaf samples (including leaf base and petiole), 117 stem samples (89 herbaceous and 28 woody samples) and 104 samples from reproductive structures (66 inflorescence, 33 fruit and five seed samples). Roots were excluded from this study due to their susceptibility to contamination from surrounding soil, which makes them challenging to analyse accurately (see discussion in Tromp and Dudgeon, 2015). Given that roots from dicotyledons remain generally understudied, we recommend dedicating a separate study – using a tailored roots cleaning protocol – to better address root phytolith production and composition. To account for the variation in phytolith assemblages within plant organs, entire organs were processed, with stems being standardized to sampling ~4–5 cm of biomass (Blackman, 1971; Mulholland et al., 1988, 1990). All plant tissues were documented visually before processing; images are available in the online repository (see data availability).

Plant tissue laboratory processing

Of the 338 samples analysed in this study, 46 underwent processing using wet oxidation (protocol 1), while the majority, 292 samples, were treated with a combination of wet and dry oxidation (protocol 2). Both protocols are detailed in Supplementary Data File 1 – Methods S2. Initially, wet oxidation was the primary method employed; however, as the research progressed, a strategic shift was made towards integrating wet and dry oxidation. Specifically, 50 samples that were first processed with protocol 1 were subsequently reprocessed using protocol 2 (Supplementary Data File 1 – Table S1). This decision was influenced by the observation that samples processed with protocol 2 tended to yield phytoliths with less breakage in contrast to samples processed using protocol 1, thereby enhancing the study of their phytolith morphologies. Samples that did not produce phytoliths or where phytolith morphotypes had limited breakage, which were processed initially using protocol 1, were not subjected to reprocessing using protocol 2, resulting in the 46 samples that underwent processing using protocol 1. The differences in laboratory processing and their implications are discussed later in this work.

Data acquisition

The samples were examined and photographed at magnifications of ×200, ×400 and ×500 under plane-polarized light (PPL). Different microscopes were used: a Zeiss Axioscope 5 TL/RL polarizing microscope with a Zeiss Axiocam 208 digital camera, an AmScope T490B-5M Composite Microscope with an AmScope MU503 digital camera, and a Nikon i80 compound microscope with a DS-Fi1 5-Meg Color C Mount digital camera. Microphotographs (n = 964) are accessible in the online repository creating a visual phytolith atlas. Each microphotograph includes the morphotype names within its metadata (Table 2).

Table 2.

Morphotype groups, names, codes, diagnostic character (dia.) following the compound variables of Strömberg (2003), occurrence of the morphotype (on a morphotype and sample level), corresponding ICPN2.0 name (if applicable) and figure.

Morphotype group (Strömberg, 2004)	Previously used morphotype names (Strömberg, 2003)^‡	Code	Dia.	Occurrence (100 % = all morphotypes)	Occurrence (100 % = all samples with phytoliths)	ICPN2.0 (ICPT, 2019)	Fig.
‘Mesophyll’	‘Parenchyma/honeycomb aggregate’	M-1	Y	12.97	85.19
	‘Infilled parenchyma/honeycomb aggregate’	M-2	Y	0.41	11.85
	‘Concentric silica aggregate’	M-3	Y	1.01	19.26
	‘Solid parenchyma cell’	M-8	Y	0.21	11.11
	‘Transparent polyhedral sheet’	M-11	Y	3.37	2.96
	Amoeboid tuberous^*	AMO_TUB	Y	0.11	1.48		1A
‘Stomata’	‘Dicotyledon-type stomata’	St-1	Y	1.55	37.78
‘Dicotyledon epidermis’	‘Polyhedral epidermis’	Epi-1	Y	27.15	79.26
	‘Anticlinal epidermis’	Epi-2	Y	8.41	41.48
	Polygonal papillar^*	POL_PAP	Y	0.87	2.96		1B
‘Tracheary element’	‘Hollow and infilled helix (helical tracheary element)’	Tra-1	Y	12.96	72.59	Tracheary annulate/helical
	‘Worm/pupa-like, infilled helical tracheary element’	Tra-2	Y	1.05	8.89
	‘Ornamented (conifer) 3D polyhedrons (ridged)’	Tra-5 type B	Y	0.11	1.48
	‘Ornamented (conifer) 3D polyhedrons (stalked or elevated circular pits)’	Tra-5 type C	Y	0.06	1.48
‘Cystolith and non-grass trichome’	‘Stalked cystolith’	Cy-1 type A	Y	0.30	5.93
	‘Verrucate cystolith with trichome’	Cy-1 type B	Y	0.25	3.70
	‘Trichome bases infilled with small masses made up of verrucate or VI material’	Cy-1 type C	Y	0.18	5.19
	Sacciformis verrucate^*	SAC_VER	Y	0.65	1.48		1C
	‘Articulated trichome’	Tri-2	Y	0.59	11.11
	‘Non-segmented, armed trichome’	Tri-3	Y	2.74	9.63
	‘Trichome with laminar wall without laminated infillings’^†	Tri-4 type A	Y	1.76	20.00
	‘Trichome with laminar wall with laminated infillings’^†	Tri-4 type B	Y	1.33	13.33
	‘Laminar/simple trichome with attached polyhedral/anticlinal epidermis’	Tri-5	Y	0.68	14.07
	‘Simple hollow trichome’	Tri-6	N	6.81	35.56
	‘Simple solid trichome’	Tri-7	N	0.90	18.52
	Annuloid conical^*	ANN_CON	N	0.28	10.37		1D
	Acicular bulbous^*	ACI_BUL	Y	0.31	2.96		1E
	Brachiate acicular^*	BRA_ACI	Y	0.03	0.74		1H
	Tholus pappilar^*	THO_PAP	Y	0.21	2.96		1G
	Spheroid quadrifidius^*	SPH_QUA	Y	0.47	2.96		1G
‘Smooth/compound sphere’)	‘Homogeneous VI spheres and subspheres’	Cl-1	Y	1.38	35.56	Spheroids
	‘Attached verrucate silica’	Cl-2	Y	0.41	2.96
	‘Small, smooth, pink sphere’	Cl-4	Y	1.90	6.67
	‘Compound sphere’	Cl-5	Y	0.20	0.74
	‘Small rugulose sphere and subsphere’	Cl-7	Y	0.33	9.63
	‘Large nodular body’	Cl-8	Y	0.06	2.22
	‘Small grainy sphere’	Cl-11	N	0.18	4.44
	Spheroid verrucate^*	SPH_VER	Y	0.35	2.22		1I
‘Vesicular infilling (VI)’ (thick silica with laminar internal structure)	‘VI spheres’	VI-1	Y	2.29	35.56
	‘Non-spherical VI bodies’	VI-2	Y	0.68	14.07
	‘Sedge-like VI plate’	VI-3	Y	0.14	8.15
‘Elongate’	‘Elongate entire’	Elo-1	N	1.82	19.26	Elongate entire
‘Rod’	‘Thin, straight, flat band’	Elo-5	N	1.67	19.26
‘Rod’	‘Sinuous, irregular rod’	Elo-8	Y	0.28	10.37
‘Blocky polyhedron’	‘3D blocky polyhedron’	Blo-6	Y	0.10	7.41	Blocky
‘Blocky polyhedron’	‘Large, faceted/scalloped sphere’	Blo-11	Y	0.05	2.22	Blocky
‘Sheets’	‘Striped sheet’	She-2	N	0.44	5.93

Open in a new tab

^*Newly named and described morphotypes. More information can be found in Supplementary Data File 2.

^†In this study, ‘Trichome with laminar wall’ (Tri-4) from Strömberg (2003) was split up into ‘Trichome with laminar wall without laminated infillings’ and ‘Trichome with laminar wall with laminated infillings’.

^‡Work is underway to describe and name these morphotypes (R. Albert, K. Neumann, L. Scott Cummings, C. Strömberg, L. Vrydaghs, C. Yost, unpubl. data).

For each sample, phytoliths were counted until 200 ‘diagnostic’ phytoliths had been recorded. Concurrently, ‘non-diagnostic’ phytoliths were tallied and documented. The classification of ‘diagnostic’ and ‘non-diagnostic’ phytoliths adheres to the ‘compound variables’ (roughly comparable to plant functional types) of Strömberg, (2003) (see Table 2). If the number of diagnostic phytoliths was fewer than 200 for a sample, the phytoliths were still documented (see Supplementary Data File 3) but the sample was excluded from calculation of the Horn–Morisita index, permutational multivariate analysis of variance (PerMANOVA) and the linear discriminant analyses (LDAs) (see further below).

Phytolith production was estimated by observing microscope slides at 200× magnification, employing a modified version of the semi-quantitative descriptive index of Wallis (2003). We relied on a combination of counting on slides and measuring the amount of silica left in a 1.5-mL Eppendorf tube after slides were made to classify samples into:

Non-producer (NP): no phytoliths were observed across the entire slide(s). No residues were left in the Eppendorf tube.
Trace (T): few (non-)diagnostic phytoliths (1–100) were observed when scanning the entire microscopic slide(s). No residues were left in the Eppendorf tube.
Common (C): a moderate number (≥101) of (non-)diagnostic phytoliths were observed when scanning the entire microscopic slide(s). Residues were ≥2 mm high in the Eppendorf tube.
Abundant (A): a large number (≥201) of (non-)diagnostic phytoliths were observed when scanning the entire microscopic slide(s). Residues were <2 mm high in the Eppendorf tube.

Plant material for each sample was not standardized by weight, and not all extracted phytoliths were necessarily mounted on the slide(s), especially in cases of high yield. For several samples, multiple slides were prepared to observe a total of 200 diagnostic phytoliths. Details regarding the number of slides examined for each sample, along with information about any additional yield available per sample, are documented in the ‘phytolith_counts_table.xlsx’ file, which can be found in the online repository. Both the ‘common’ and ‘abundant’ categories indicate a relatively high presence of phytoliths. To reduce analytical noise, ‘common’ and ‘abundant’ categories were merged into a single ‘Common/Abundant (CA)’ category for analysis.

The phytolith nomenclature system in Strömberg (2003, 2004) was employed. Table 2 illustrates the correspondence of this nomenclature with the International Code for Phytolith Nomenclature (ICPN) 2.0 (ICPT, 2019). The ongoing efforts of the International Committee for Phytolith Taxonomy (ICPT) include the descriptions of many dicotyledon phytolith morphotypes in line with the ICPN 2.0 (R. Albert, K. Neumann, L. Scott Cummings, C. Strömberg, L. Vrydaghs, C. Yost, unpubl. data). We therefore retain the original morphotype names of Strömberg (2003, 2004). Nine phytolith morphotypes not previously described by Strömberg (2003) and ICPN 2.0 were named and described including: (1) Amoeboid tuberous (AMO_TUB), (2) Polygonal papillar (POL_PAP), (3) Sacciformis verrucate (SAC_VER), (4) Annuloid conical (ANN_CON), (5) Acicular bulbous (ACI_BUL), (6) Brachiate acicular (BRA_ACI), (7) Tholus papillar (THO_PAP), (8) Spheroid quadrifidius (SPH_QUA) and (9) Spheroid verrucate (SPH_VER) (Fig. 1; Supplementary Data File 2).

This is a figure plate of micrographs of newly described phytolith morphotypes in this article. It includes micrographs of nine morphotypes. Supplementary Data File 2 contains detailed descriptions of the morphologies of these morphotypes. — Micrographs of newly described morphotypes. (A) Amoeboid tuberous in *Sambucus ebulus* leaf sample. (B) Polygonal papillar in *Papaver argemone* fruit sample. (C) Sacciformis verrucate in *Urtica dioica* subsp. *dioica* leaf sample. (D) Annuloid conical in *Urtica urens* inflorescence sample (D1), *Raphanus raphanistrum* leaf sample (D2) and *Cannabis sativa* leaf sample (D3). (E) Acicular bulbous in *Urtica urens* inflorescence sample. (F) Tholus papillar in *Urtica dioica* subsp. *dioica leaves* (F1) and *U. urens* leaf sample (F2 and F3). (G) Spheroid quadrifidius in *Rumex crispus* leaf sample (G1), *R. conglomeratus* leaf sample (G2) and *R. sanguineus* leaf sample (G3). (H) Brachiateacicular in *Capsella bursa-pastoris* leaf sample. (I) Spheroid verrucate in *Urtica dioica* subsp. *dioica* leaf sample (I1) and *Urtica urens* leaf sample (I2 and I3).

Data analysis

A generalized linear mixed model (GLMM) assessed the effect of taxonomy and plant part on phytolith presence (present/absent) (Supplementary Data File 1 – Table S2). This type of modelling is suitable for our dataset because it accommodates the hierarchical structure of our data, allowing for both fixed effects (taxonomy and plant part) and random effects (specimen variation), thus providing a comprehensive analysis of factors affecting phytolith presence. The GLMM provides coefficients that indicate the impact of each predictor variable on the log-odds of the binary outcome (phytolith presence/absence). Positive coefficients indicate increased log-odds with an increase in the predictor, while negative coefficients imply decreased log-odds. Larger coefficients represent stronger predictor–outcome relationships (Stroup et al., 2024). Given the limited sample size across taxonomic levels – with 338 samples spanning four superorders, 15 orders, 24 families and 59 genera – the analysis focused on superorders to maintain the statistical robustness of the GLMM.

The Horn–Morisita index was utilized to calculate pairwise distances between phytolith assemblages having 200 diagnostic phytoliths. This index, previously used in other phytolith studies (Gallego and Distel, 2004; Fernández Honaine et al., 2006), quantifies the similarity or difference in composition between two assemblages. Values near zero indicate high similarity, while values closer to one suggest greater dissimilarity. This index not only considers whether phytolith morphotypes are present or absent but also their abundance, providing a nuanced measure of assemblage overlap (Horn, 1966). Subsequently, we used a PerMANOVA to analyse the variability among phytolith assemblages using the Horn–Morisita distance matrix. PerMANOVA is a non-parametric test that efficiently detects differences between multivariate sample groups (Anderson, 2017).

Lastly, LDA was employed to discern patterns among the samples with 200 diagnostic phytoliths, focusing on plant part, life form and taxonomy (order) as grouping variables. LDA is effective in distinguishing between groups by maximizing between-group variance and minimizing within-group variance, thereby highlighting unique phytolith assemblages related to the specified variables (Holden et al., 2011).

RESULTS

Phytolith production

Inter-duplicate sample variation.

In total, 114 duplicate samples (for 43 duplicate species, with —two or three plant part samples per species) were investigated. Duplicate samples were considered consistent when they were both labelled as either non-producer (NP), trace (T), or common/abundant producer (CA), which was true for 76 % (87 out of 114 duplicate samples). Based on the studied material, on a superorder level, for Asteranae, 76 % (29 out of 38 duplicate samples) has consistent phytolith production levels. For Caryophyllanae, Ranunculanae and Rosanae, this is 70 % (16 out of 23 duplicate samples), 85 % (11 out of 13 duplicate samples) and 78 % (31 out of 40 duplicate samples), respectively. Further, 67 % (29 out of 43 duplicate samples) of species exhibited consistent phytolith production levels in their leaves. Inflorescences and seeds/fruits showed similar patterns of consistency, with 65 % (11 out of 17 duplicate samples) and 64 % (7 out of 11 duplicated samples), respectively. For stems, 93 % (40 out of 43 duplicate samples, with many being consistent non-producers) of species showed consistent phytolith production.

Plant parts.

The phytolith production index showed distinct patterns across the different plant plants. Among all leaf samples analysed, 73 % (85 out of 117 samples) produced phytoliths, with 55 % being common/abundant producers and 18 % being trace producers. Leaf samples accounted for the majority of common/abundant production (Fig. 2). Inflorescence samples showed phytolith production in 42 % (28 out of 66 samples), with 23 % being trace producers and 20 % being common/abundant producers. In contrast, among fruit and seed samples, only 29 % (11 out of 38) showed production, with 5 % being trace producers and 24 % being common/abundant producers. Stem samples had the lowest phytolith production, 9 % (11 out of 117), with 3 % bring trace producers and 6 % being common/abundant producers. Phytoliths in reproductive parts or stems were generally present only in cases where the corresponding leaves also produced phytoliths, with Prunus as a notable exception to this pattern.

A bar plot showing the relative frequency of phytolith production index per plant part (n = 338). The x-axis represents the phytolith production index, divided into three categories: Non-producer, Trace, and Common/Abundant. The y-axis represents the absolute frequency. Plant parts are color-coded: red for reproductive part (fruits or seeds), green for reproductive part (inflorescences), blue for leaves, and purple for stem. Around 60% of the samples are non-producers, with stems accounting for approximately half of these. Roughly 10% of the samples are trace producers. About 30% are common/abundant producers, with leaves being responsible for the majority of these samples. — Relative frequency of phytolith production index per plant part (n = 338). Rep. part = reproductive part.

GLMM analysis showed that different plant parts significantly impact phytolith presence (present/absent). Using ‘Asteranae’ and ‘fruits and seeds’ as the reference group (value of −5.72, P < 0.001), the model found that leaves and inflorescences strongly increased the likelihood of phytolith presence, with coefficients of 13.02 and 6.50 (both P < 0.001). In contrast, stems had a significant negative effect, with a coefficient of −3.63 (P = 0.018) (Supplementary Data File 1 – Table S3).

Taxonomy.

GLMM analysis highlighted variation in phytolith presence (absent/present) among superorders. Compared to the reference group (Asteranae and seeds/fruits), Caryophyllanae had significantly low phytolith production, with a coefficient of −12.04 (P < 0.001). In contrast, Ranunculanae and Rosanae showed no significant effect on phytolith presence (present/absent), with coefficients of 0.39 (P = 0.887) and −0.07 (P = 0.970), respectively (Supplementary Data File 1 – Table S3). These results show that there are differences in phytolith production levels at taxonomic levels below superorder, for example the family level (Fig. 3). The most abundant producing families (with over 75 % of the samples being common/abundant phytolith producers) are the Cannabaceae, Moraceae and Urticaceae. Non-producing families or families with a limited number of phytolith producing samples include the Amaranthaceae, Apiaceae, Fabaceae, Juglandaceae, Lamiaceae, Polygonaceae and Primulaceae. For several plant families, there are also substantial intra-familial differences, such as within the Rosaceae or the Papaveraceae, where some genera are characterized by consistent non- or trace producing samples, while other genera include common/abundant producing samples (Table 1).

A horizontal bar plot showing the phytolith production index by family (n = 338). Each bar represents a family and is segmented into three production levels: Non-producer (red), Trace (green), and Common/abundant (blue). The proportion of each production level is displayed on the x-axis. Families such as Amaranthaceae, Apiaceae, Fabaceae, Lamiaceae, Polygonaceae, and Primulaceae show a high proportion of non-producing samples, while families like Asteraceae, Cannabaceae, Moraceae, and Urticaceae show a high proportion of common/abundant producing samples. — Phytolith production index presented at family level (n = 338).

Life form.

Phytolith production is widespread across both northern European forbs and shrubs/trees. In forb samples, 39 % (101 out of 260) contained phytoliths, with 13 % being trace producers and 26 % being common/abundant producers. Among forbs, leaf samples show higher phytolith production, at 70 % (64 out of 91), including 18 % trace producers and 53 % common or abundant producers. Conversely, stem samples had a markedly lower yield at 8 % (7 out of 91), with 4 % being trace producers and 3 % being common/abundant producers. Among inflorescence samples, 41 % (23 out of 56) were productive, with 21 % being trace producers and 20 % being common/abundant producers. Among seed and fruit samples, 32 % (7 out of 22) were productive, with 9 % being trace producers and 23 % being common/abundant producers.

In shrub and tree samples, 44 % (34 out of 78) contained phytoliths, with 10 % being trace producers and 33 % being common/abundant producers. Among shrubs and trees, leaf samples showed higher phytolith production, at 81 % (21 out of 26), including 19 % being trace producers and 62 % being common/abundant producers. Conversely, stem samples had a markedly lower yield at 15 % (4 out of 26), all categorized as common/abundant producers. Among inflorescence samples, 50 % (5 out of 10) were productive, with 30 % being trace producers and 20 % being common/abundant producers. Among seed and fruit samples, 33 % (4 out of 12) were productive, with solely common/abundant producers.

Phytolith morphotypes

Occurrence of morphotypes.

A total of 22 174 phytoliths were identified, including 19 508 diagnostic phytoliths and 2 666 non-diagnostic phytoliths. Phytoliths were observed in 135 out of the 338 samples. Table 2 shows the relative occurrence of each morphotype within all phytolith assemblages (=X/22 174), as well as the relative occurrence of each morphotype within all samples with phytoliths (=X/135). Morphotypes that are present in over 30 % of all samples, and can be considered highly redundant, include ‘parenchyma/honeycomb aggregate’ (M-1), ‘polyhedral epidermis’ (Epi-1), ‘hollow and infilled helix (helical tracheary element)’ (Tra-1), ‘anticlinal epidermis’ (Epi-2), ‘dicotyledon-type stomata’ (St-1), ‘simple hollow trichome’ (Tri-6), ‘vesicular infilling spheres’ (VI-1) and ‘homogeneous vesicular infilling spheres and subspheres’ (Cl-1). Collectively, these morphotypes account for 68.30 % of all phytoliths. Morphotypes that exhibit redundancy at lower frequencies (here defined as observed in 3–30 % of all samples) encompass various trichome and cystolith types, among other morphotypes (Table 2). The morphotypes in this group that show potential to discriminate below dicotyledon level are the cystoliths. ‘Stalked cystolith’ (Cy-1 type A), ‘verrucate cystolith with trichome’ (Cy-1 type B) and ‘trichome bases infilled with small masses made up of verrucate or vesicular infilling material’ (Cy-1 type C) were observed only in four families within this dataset: Cannabaceae, Moraceae, Urticaceae and Viburnaceae.

Non-redundant morphotypes (observed in less than 3 % of all samples) can be categorized into two groups. The first group comprises morphotypes known from a broad spectrum of plant taxa, including ‘ornamented (conifer) 3D polyhedrons (ridged)’ (Tra-5 type B), ‘ornamented (conifer) 3D polyhedrons (stalked or elevated circular pits)’ (Tra-5 type C), ‘attached verrucate silica’ (Cl-2), ‘compound sphere’ (Cl-5) and ‘large, faceted/scalloped sphere’ (Blo-11) (Strömberg, 2003). ‘Ornamented (conifer) 3D polyhedrons (ridged)’ (Tra-5 type B) and ‘ornamented (conifer) 3D polyhedrons (stalked or elevated circular pits)’ (Tra-5 type C), for instance, occur mainly in conifers (Strömberg, 2003), and have only a limited presence in dicotyledons. Their non-redundant character in this study arises from the absence of conifers in the dataset.

The second group includes morphotypes that either (1) are highly diagnostic, meaning that they can differentiate taxonomic groups below dicotyledon level or distinguish between plant parts, or (2) whose diagnostic status has yet to be established. Examples of highly diagnostic morphotypes in this study include ‘transparent polyhedral sheet’ (M-11), which occur in stems of woody dicots (Strömberg, 2003), and were observed in the woody stems of Ficus carica and Morus nigra in this study. Polygonal papillar (Fig. 1B) occurred in four Papaver samples and the similar ‘silicified papillae (-like)’ has previously been described from Leucadendron spissifolium (Proteaceae) (Novello et al., 2018). Metcalfe and Chalk (1950) reported ‘epidermis papillose’ across various plant families, but it is unclear whether these structures undergo silicification and would be included in Polygonal papillar. Spheroid verrucate (Fig. 1I) and Sacciformis verrucate (Fig. 1C) were observed for several Urtica samples herein and in previous work (Bozarth, 1992; Fernández Honaine et al., 2018), and Spheroid verrucate additionally sporadically (n = 2) for Cannabis sativa (Supplementary Data File 3). Tholus papillar (Fig. 1F) was observed for Ficus carica and several Urtica samples. Brachiate acicular (Fig. 1H) was exclusively identified for Capsella bursa-pastoris. Spheroid quadrifidius (Fig. 1G) was observed for Rumex crispus, R. conglomeratus, R. sanguineus and Solanum dulcamara. Both types of trichomes are distributed across various plant families according to Metcalfe and Chalk (1950). ‘Large nodular body’ (Cl-8) was observed for Ficus carica and Cannabis sativa. Strömberg (2003) documented the presence of this type for Eucommia (Eucommiaceae) and Chimonanthus (Calycanthaceae), two genera not occurring in NW Europe.

Lastly, three morphotypes with an unknown anatomical origin were observed but not systematically recorded because of their rare occurrence. First, a stomata-like morphotype having a fusiform shape was observed for a stem and leaf sample of Papaver argemone, classified as ‘dicotyledon-type stomata’ (St-1) (Fig. 5E). The stomatal nature of this morphotype is not definitively established but seems likely.

This figure displays a plate with micrographs of phytolith morphotypes, including those with diagnostic features and those with an unknown anatomical origin. The morphological descriptions of these phytoliths can be found in the 'phytolith morphotypes' and 'morphotypes with distinctive features' sections of the results. — Micrographs of phytolith morphotypes with diagnostic features (A–D) and phytolith morphotypes with an unknown anatomical origin (E–H). (A) Cone-shaped ‘simple hollow trichome’ (Tri-6) with a sphere inside and cone-shaped ‘simple hollow trichome’ (Tri-6) having a sphere-like hair base in *Morus nigra* fruit sample. (B) ‘Articulated trichome’ (Tri-2) with a striate surface in *Centaurea cyanus* leaf sample. (C) ‘Articulated trichome’ (Tri-2) with squarish segments and thick cell walls in *Centaurea cyanus* leaf sample. (D) ‘Laminar/simple trichome with attached polyhedral/anticlinal epidermis’ (Tri-5) with ‘vesicular infilling spheres’ (VI-1) in the adjacent cells in *Myosotis scorpioides* leaf sample. (E) Stomata-like morphotype having a fusiform shape in *Papaver argemone* stem sample. (F) Small (20–30 µm in diameter) silica skeletons forming a concentric pattern around an inner cell that is split in two in *Cannabis sativa* leaf sample. (G) Bilateral or trilateral phytoliths with lobed extensions radiating from the centre in *Rhamphospermum nigrum* leaf sample. (H) ‘Concentric silica aggregate’ (M-3) with a large central circular cell with ‘vesicular infilling spheres’ (VI-1) in the adjacent cells in *Rhamphospermum nigrum* leaf sample (H1 and H2) and *Rhamphospermum nigrum* inflorescence sample (H3).

For Cannabis sativa, small (20–30 µm in diameter) phytoliths (n < 5) forming a concentric pattern around an inner cell that is split in two were observed (Fig. 5F). The precise anatomical origin of these structures remains unidentified, but they may represent mesophyll, closed stomatal cells or hair bases.

For Rhamphospermum nigrum, bilateral or trilateral phytoliths with lobed extensions radiating from the centre were noted (Fig. 5G). The anatomical origin of the morphotype is unknown. Considering their highly unusual shape, they may be highly diagnostic. However, their low occurrence in the sample (n < 5) prevents firm conclusions regarding both their diagnostic character and anatomical origin.

Morphotypes with distinctive features.

This study observed morphotypes with distinct features – beyond those outlined in the general classification scheme (see Table 2) – that may serve as characteristic indicators for specific genera or species, thereby facilitating a taxonomic botanical attribution of these morphotypes below dicotyledon level. These features encompass various dicotyledon epidermis structures and certain trichomes. Their relative frequencies in the phytolith assemblages were not recorded; instead, we provide a descriptive report of their presence here and in Supplementary Data File 3.

For dicotyledon epidermal phytoliths, specific, potentially diagnostic features included hexa- or pentagonal with ‘vesicular infilling spheres’ (VI-1), ‘homogeneous vesicular infilling spheres and subspheres’ (Cl-1) or ‘attached verrucate silica’ (Cl-2), which were observed for Morus nigra and Ficus carica (Fig. 4A). Previous literature has documented the occurrence of such hexa- or pentagonal epidermal phytoliths for Moraceae (Geis, 1973; Kealhofer and Piperno, 1998). The pentagonal or hexagonal ‘polyhedral epidermis’ (Epi-1) is not exclusive to Moraceae species; for example, Vitis vinifera and Centaurea cyanus also exhibited these shapes within our dataset (Fig. 4A). Nevertheless, in combination with ‘vesicular infilling spheres’ (VI-1), ‘homogeneous vesicular infilling spheres and subspheres’ (Cl-1) or ‘attached verrucate silica’ (Cl-2), these epidermal cell phytoliths may be diagnostic for the Moraceae.

This figure displays a plate with micrographs of epidermal phytolith morphotypes with diagnostic features. The morphological descriptions of these phytoliths can be found in 'morphotypes with distinctive features' section of the results. — Micrographs of epidermal phytoliths with diagnostic features. (A) Hexagonal and pentagonal ‘polyhedral epidermis’ (Epi-1) with ‘vesicular infilling spheres’ (VI-1), ‘homogeneous vesicular infilling spheres and subspheres’ (Cl-1) or ‘attached verrucate silica’ (Cl-2) in *Ficus carica* leaf sample (A1) and *Morus nigra* leaf sample (A2); hexagonal and pentagonal ‘polyhedral epidermis’ (Epi-1) without ‘vesicular infilling spheres’ (VI-1), ‘homogeneous vesicular infilling spheres and subspheres’ (Cl-1) or ‘attached verrucate silica’ (Cl-2) in *Vitis vinifera* leaf sample (A3) and *Centaurea cyanus* leaf sample (A4). (B) Elongated ‘anticlinal epidermis’ ‘(Epi-2) in *Anthemis cotula* leaf sample (B1) and *Bidens cernua* inflorescence sample (B2). (C) ‘Polyhedral epidermis’ (Epi-1), ‘anticlinal epidermis’ (Epi-2) and ‘concentric silica aggregate’ (M-3) with a single spherical to ovoid filling in *Fragaria vesca* leaf sample. (D) ‘Polyhedral epidermis’ (Epi-1) with an angular outline and a verrucate surface pattern in *Cannabis sativa* seed sample. (E) ‘polyhedral epidermis’ (Epi-1) with small, infilled voids (possibly plasmodesmata or the space between them) in the cell walls and thick laminations [‘Non-spherical vesicular infilling bodies’ (VI-2)] in *Sambucus ebulus* leaf sample (E1: this study, E2: Vrije Universiteit Brussel unpublished reference collection). (F) Small voids in ‘anticlinal epidermis’ ‘(Epi-2) that may be (imprints of) small oxalate crystals in *Bidens cernua* inflorescence sample.

Elongated ‘anticlinal epidermis’ (Epi-2) was observed for Anthemis cotula and Bidens cernua (Fig. 4B). Elongated ‘anticlinal epidermis’ (Epi-2) has previously been described for Anthemis cotula by Hart (2007). This morphotype is considered diagnostic for Asteraceae.

Small (10–20 µm in diameter) ‘anticlinal epidermis’ (Epi-2) with an angular outline and a verrucate surface pattern was exclusively observed for the seeds of Cannabis sativa in this dataset (Fig. 4D). Seeds of species within the order Urticales are known to be diagnostic at the genus or species level, as observed in, for example, Celtis and Trema (Kealhofer and Piperno, 1998; Watling and Iriarte, 2013; Piperno and McMichael, 2020). This type of ‘anticlinal epidermis’ (Epi-2) may therefore be diagnostic of Cannabis seeds.

‘Polyhedral epidermis’ (Epi-1), ‘anticlinal epidermis’ (Epi-2), ‘concentric silica aggregate’ (M-3) and ‘laminar/simple trichome with attached polyhedral/anticlinal epidermis’ (Tri-5), all with single spherical to ovoid filled cells, were observed for Fragaria vesca (Fig. 4C) and have previously been reported from the leaves of Rosa banksiae by Strömberg (2003). These morphotypes might be diagnostic of the Rosaceae.

Small (<1 µm in diameter), infilled voids (possibly plasmodesmata or the space between them) were observed in the cell walls of the ‘polyhedral epidermis’ (Epi-1) in a leaf sample of Sambucus ebulus in this study (Fig. 4E). The cell walls are laminated with ‘non-spherical vesicular infilling bodies’ (VI-2). These two features are shared with other Sambucus specimens, as illustrated by a micrograph of Sambucus adnata in Ge et al. (2020) and another sample of Sambucus ebulus that is part of the (unpublished) modern plant-based phytolith reference collection at the Vrije Universiteit Brussel (VUB). The latter exhibited similar yet larger (∼2 µm in diameter), infilled holes (Fig. 4E).

Two specimens of Bidens cernua exhibited small (∼1 µm in diameter), rounded voids in both ‘polyhedral epidermis’ (Epi-1) and ‘anticlinal epidermis’ (Epi-2), and in both inflorescence and leaf samples (Fig. 4F). These voids may probably be (imprints of) small oxalate crystals. It is unclear if this feature would be preserved in the archaeological or fossil record, and hence if it is suitable for taxonomic differentiation.

For trichomes, the following features were observed: cone-shaped ‘simple hollow trichome’ (Tri-6) with a sphere inside and cone-shaped ‘simple hollow trichome’ (Tri-6) having a sphere-like hair base were identified for Morus nigra (Fig. 5A). This observation probably aligns with what Watling and Iriarte (2013) refer to as ‘hair base’ observed for Ficus americana. Geis (1973) reported that Morus rubra produced abundant hook-shaped trichomes with a spherical protuberance of solid silica on the upper surface. Kealhofer and Piperno (1998) observed two types of ‘unsegmented hairs with cystoliths’ for a fruit sample of Morus alba. It is plausible that these descriptions correspond to the observed trichome type, although the limited number of published images prevents definitive confirmation.

‘Articulated trichome’ (Tri-2) with a striate surface was observed for Centaurea cyanus (Fig. 5B). Striated trichomes have been reported from other Asteraceae species, potentially rendering this trichome type diagnostic for the family Asteraceae (Parry et al., 1984, 1986). It should be noted that images provided by Parry et al. (1984, 1986) depict trichomes with striation patterns parallel to the long axis of the trichomes, whereas in this study, the striation pattern is perpendicular to the long axis of the trichomes. Kealhofer and Piperno (1998) also noted striated hairs in a leaf sample of Ficus hispida, providing no accompanying image, thereby making it uncertain whether the striations provide morphological distinctions between the two families.

‘Articulated trichome’ (Tri-2) with squarish segments and thick cell walls was observed for Centaurea cyanus (Fig. 5C). This type is common for Asteraceae, particularly for Bidens, and is only shared by some of the Cucurbitaceae. This morphotype might be diagnostic of the Asteraceae in temperate regions, where few Cucurbitaceae taxa occur naturally (Piperno, 1988).

‘Laminar/simple trichome with attached polyhedral/anticlinal epidermis’ (Tri-5) with ‘vesicular infilling spheres’ (VI-1) in the adjacent cells were observed for Myosotis scorpioides (Fig. 5D). Similar trichomes were observed in Myosotis scorpioides by Fernández Honaine et al. (2023), but it is unclear whether they would be diagnostic for this taxon. ‘Concentric silica aggregate’ (M-3) with a large central circular cell with ‘vesicular infilling spheres’ (VI-1) in the adjacent cells, which probably also derives from a large trichome base (inner diameter 75 µm, outer diameter >100 µm), were observed exclusively for Rhamphospermum nigrum (Fig. 5H).

Phytolith assemblages

Inter-duplicate sample variation.

To examine the similarity between phytolith assemblages of duplicate samples, as well as their dissimilarity from assemblages of other samples, we analysed the mean pairwise distance among all samples with 200 diagnostic phytoliths (n = 83 samples). This analysis revealed a mean pairwise distance of 0.62, with a standard deviation of 0.32. Notably, the duplicate samples having at least 200 diagnostic phytoliths (17 pairs) demonstrated a remarkable degree of similarity, with pairwise distances of duplicate samples ranging from a minimum of 4.91 × 10⁻⁴ to 0.31, except for Myosotis scorpioides leaves, which stood out with a dissimilarity value of 0.72 (Supplementary Data File 1 – Table S4 and Fig. S1). This pronounced similarity in the phytolith assemblage composition between duplicate samples (intra-group) compared to among all samples (inter-group) is supported in the PerMANOVA. Here, inter-group labels explained 97.59 % (R²) of the phytolith assemblage variance, whereas intra-group residuals contribute only 2.41 % (R²) to phytolith assemblage variance (Supplementary Data File 1 – Table S5).

Discriminating between assemblages (LDA)

Plant part grouping:

The results of LDA conducted on samples containing 200 diagnostic phytoliths (n = 83) clearly distinguished phytolith assemblages of woody stem samples (Ficus carica and Morus nigra) from those of herbaceous stem samples (Urtica dioica subsp. dioica, Cannabis sativa and Papaver argemone) (Fig. 6; Supplementary Data File 1 – Tables S6 and S7). ‘Transparent polyhedral sheet’ (M-11) was exclusively observed in woody stem samples in this study, contributing to the high, positive values of woody stems on the LD1 axis. On LD2, higher values for herbaceous stems were driven primarily by high loadings associated with ‘trichome bases infilled with small masses made up of verrucate or vesicular infilling material’ (Cy-1 type C), Tholus papillar, Spheroid verrucate and ‘large, faceted/scalloped sphere’ (Blo-11). These morphotypes are not exclusive to herbaceous stems, but herbaceous stems have higher relative frequencies of them compared to other groups. The phytolith assemblages of leaves, inflorescences, and seeds and fruits overlapped strongly, clustering in the lower left quadrant of the LDA scatterplot (Fig. 6).

This figure displays two LDA scatterplots. The first plot shows the grouping of plant parts, with herbaceous and woody stems segregating from reproductive parts (seeds and fruits), reproductive parts (inflorescences), and leaves, which cluster together in the lower-left quadrant. The second plot displays the same LDA grouping of plant parts, but without stem samples. However, reproductive parts (seeds and fruits), reproductive parts (inflorescences), and leaves still overlap in the second scatterplot. — LDA scatterplots (LD1 and LD2) of samples with 200 diagnostic phytoliths using plant part as a grouping variable. One LDA including stem samples and another LDA excluding stem samples. HS = herbaceous stem sample, LE = leaf sample, R(F/S) = reproductive part (fruit or seed) sample, R(I) = reproductive part (inflorescence) sample, WS = woody stem sample.

To test whether leaves, inflorescences, seeds and fruits could be distinguished from each other – an outcome that was unclear in the initial LDA, due to the distinct assemblages of herbaceous and woody stems appearing as ‘outliers’ in the LDA scatterplot – we conducted another LDA excluding the two stem groups. However, overlap among leaves, inflorescences, seeds and fruits remained (Fig. 6).

Life form grouping:

The LDA clearly discriminated between life forms (Fig. 7; Supplementary Data File 1 – Tables S6 and S7). Many shrubs and trees had higher values along the LD1-axis, while forbs had lower values. Morphotypes with high, positive loadings on LD1 include ‘parenchyma/honeycomb aggregate’ (M-1), ‘transparent polyhedral sheet’ (M-11), ‘polyhedral epidermis’ (Epi-1), ‘anticlinal epidermis’ (Epi-2), Tra-1 and Sacciformis verrucate. Spheroid verrucate and Tholus papillar contributed to lower values for LD1.

The figure displays an LDA plot segregating life forms (forbs vs. shrubs/trees). The LDA effectively separates most of the samples along a single LD. — LDA scatterplot (LD1 and LD2) of samples with 200 diagnostic phytoliths using life form (forbs vs. shrubs/trees) as a grouping variable. LE = leaf sample, R(I) = reproductive part (inflorescence) sample.

Taxonomy grouping:

In general, most species clustered closely by order. Nevertheless, there was also substantial inter-order overlap preventing clear differentiation of individual orders (Fig. 8; Supplementary Data File 1 – Tables S6 and S7). The exceptions were the leaf samples from Sambucus nigra and S. ebulus (Dipsacales) and leaf samples of Myosotis scorpioides (Boraginales) exhibiting distinct grouping patterns. Dipsacales had high, negative values along both LD1 and LD2, while Boraginales showed high, positive values on LD1 and high negative on LD2. In the case of Dipsacales, the negative values on LD1 were influenced primarily by high relative abundances of Spheroid verrucate. The high positive scores along LD1 for Boraginales were due to high frequencies of ‘laminar/simple trichome with attached polyhedral/anticlinal epidermis’ (Tri-5).

The figure displays an LDA plot segregating samples based on the taxonomic level 'order'. However, the LDA does not effectively separate most of the orders. — LDA scatterplot (LD1 and LD2) of samples with 200 diagnostic phytoliths using the taxonomic level 'order' as a grouping variable.

DISCUSSION

Intra-species variation in production and assemblage composition

The main objective of this study was to lay the groundwork for a baseline for dicotyledon phytoliths of plant taxa occurring in NW Europe. We concentrated on taxa occurring in NW Europe, integrating samples from diverse environmental context for the purpose of uncovering broad patterns in phytolith production and assemblages across various environmental conditions. The study found that 76 % of duplicate samples (87 of 114) consistently matched in phytolith production labels: non-producer (NP), trace (T) or common/abundant (CA) (Table 2). The variation observed in phytolith production may stem from two main factors. First, the phytolith index we used is semi-quantitative in nature, providing only a coarse gauge of production levels, which in of itself may contribute to variability. Shifting to a binary classification (classifying samples simply as producing or non-producing) can resolve estimation errors inherent to this index. However, even after using a binary classification system, 18 % (21 out of 114 duplicate samples) still show inconsistency in phytolith production. Second, as hypothesized at the start of this study, the broad geographical range of the samples, encompassing various environmental contexts (e.g. several climate zones; Table 1), may have influenced production variation. Additional factors, such as tissue maturity, soil silicon levels, precipitation and temperature, may also have played a role (Lanning, 1960; Liu et al., 2016; Brightly et al., 2020).

Although phytolith production varied widely across duplicate samples in our dataset, phytolith assemblages showed low variation within duplicate samples. This consistency was confirmed by the PerMANOVA, which revealed minimal intra-group variance (between duplicate samples) but significant compositional inter-group variance (between sample groups) (Supplementary Data File 1 – Table S5). In addition, we note some similar patterns in subtle morphotype features for a number of duplicate samples coming from different environmental contexts. Some examples of these include: small (∼1–2 µm in diameter), infilled voids (possibly plasmodesmata or the space between them) in the cell walls of the ‘polyhedral epidermis’ (Epi-1) of Sambucus leaves (Fig. 4E); small (∼1 µm in diameter) voids that might be (imprints of) small oxalate crystals in both ‘polyhedral epidermis’ (Epi-1) and ‘anticlinal epidermis’ (Epi-2) of Bidens cernua inflorescence and leaf samples (Fig. 4F); and elongate ‘anticlinal epidermis’ (Epi-2) within Asteraceae species including Bidens cernua and Anthemis cotula (Fig. 4B). These results tend to support the notion that genetic factors play a role in phytolith assemblages (Supplementary Data File 1 – Table S5).

Phytolith production variation

Taxonomic variation.

Our analysis showed that phytolith production differs substantially among plant families, consistent with previous studies (Fig. 3) (Piperno, 2006; Iriarte and Paz, 2009; Katz, 2015; Strömberg et al., 2016). In this study, the constrained sample size limited our statistical examination of phytolith production to taxonomic levels below superorder, which proved to be too broad a category to yield overall significant results (Supplementary Data File 1 – Table S3). Statistically, to achieve more definitive conclusions on the influence of taxonomy on phytolith production, future research should strive to test patterns of phytolith production at taxonomic levels sich as family, genus or even species.

Here we compare our results with those of Hodson et al. (2005) who provide mean shoot silicon concentrations for various plant species using leaf and non-woody shoot tissues. Their study includes species from all the families included in our study except for Resedaceae and Viburnaceae. When we compare their family-level mean silicon percentages with the percentage of common/abundant phytolith-producing samples from our study, grouped by quartiles, we find limited alignment. Specifically, only six out of the 22 families fall into matching quartiles across both datasets (Supplementary Data File 1 – Table S8). This discrepancy can be attributed to the limited sample sizes and species selection for several families in both studies. In some cases, families are represented by only one or two species, which do not align between the studies. However, even at the species level, we also see conflicting results. For instance, Hodson et al. (2005) report a low mean silicon concentration of 0.09 % for Cannabis sativa, whereas in our study, the reproductive part, herbaceous stem and leaf samples of C. sativa were all common or abundant phytolith producers. In general, the results of both studies do agree that silica concentration and phytolith production vary notably below family level, at least on genus level, probably as a result of differences in environmental conditions or plant tissue maturity.

Plant part variation.

In terms of phytolith production across different plant parts, leaves appear to be primary contributors to the phytolith record, followed by inflorescences and seeds/fruits, whereas stems were largely non-producers (Fig. 2). In most cases, leaves in this study produce either the same level (non-producer, trace, common/abundant) of phytoliths as the reproductive parts or a higher level. In general, these findings align with previous observations reporting on phytolith production patterns (Piperno, 1988; Strömberg, 2003; An and Xie, 2022). Piperno (1988) noted that seed phytoliths were often produced by species whose leaves contributed significant amounts of phytoliths but were never present in species showing no or low phytolith production. Similarly, Schoelynck et al. (2023) found that dicotyledon leaves generally contain higher silicon concentrations than flowers, although some flowers can produce more biogenic silicon (BSi). For instance, Brassica carinata has a whole flower BSi of 0.21 % and a leaf BSi of 0.125 %. In contrast, our study found that two Brassica rapa specimens did not exhibit higher phytolith production in their reproductive parts. Another genus examined by both Schoelynck et al. (2023) and in our study is Morus. Schoelynck et al. (2023) reported a whole flower BSi of 0.3 % and a leaf BSi of 0.22 % for Morus alba. In our analysis, all Morus nigra samples (1× reproductive part, 2× woody stems, 2× leaves) were classified as common/abundant phytolith producers, aligning with the findings of Schoelynck et al. (2023) for this genus.

Taxonomic resolution of dicotyledon phytoliths

The implications of diagnostic morphotypes.

Consistent with prior research, our study revealed substantial overlap in phytolith assemblage composition among taxonomic groups due to redundancy in morphotype production (Fig. 8) (Rovner, 1971; Piperno, 1988; Bozarth, 1992; Wallis, 2003; Mercader et al., 2009; Novello et al., 2012). However, even redundant phytolith morphotypes (in case of not having distinctive features, see ‘morphotypes with distinctive features’ result section) such as ‘polyhedral epidermis’ (Epi-1), ‘anticlinal epidermis’ (Epi-2) and ‘dicotyledon-type stomata’ (St-1) are taxonomically diagnostic at the level of dicotyledons, which makes it possible to differentiate them from monocotyledons (Strömberg, 2003).

Certain dicotyledon phytoliths are, or might be, highly diagnostic. The ability to discriminate within dicotyledonous angiosperms relies on either non-redundant morphotypes or morphotypes with unique features (e.g. Bozarth, 1985; Piperno, 1988, 2006; Delhon et al., 2003; Strömberg, 2003; Bremond et al., 2004; Carnelli et al., 2004). This study has shown that even redundant morphotypes, in particular the highly redundant ‘polyhedral epidermis’ (Epi-1) and ‘anticlinal epidermis’ (Epi-2), may possess diagnostic characteristics, to differentiate plant taxa below dicotyledon level (Figs 4 and 5). In general, families that did not contain taxa with phytoliths that were diagnostic below dicotyledon level included the Amaranthaceae, Apiaceae, Betulaceae, Caryophyllaceae, Fabaceae, Juglandaceae, Lamiaceae, Linaceae, Plantaginaceae, Primulaceae, Ranunculaceae, Resedaceae and Vitaceae. Families that did contain taxa with phytoliths that were diagnostic below dicotyledon level were the Asteraceae, Boraginaceae, Brassicaceae, Cannabaceae, Papaveraceae, Polygonaceae, Rosaceae, Solanaceae, Urticaceae and Viburnaceae.

Our study suggests that phytoliths may be used as markers for the cultivation and use of some oil and fibre plants, some medicinal plants and some taxa that serve as markers for anthropogenic disturbance of natural lands. We identified diagnostic phytoliths for the fibre-producing Cannabis sativa based on its epidermal cells. Cannabis seeds have distinct epidermal cells that may aid in identifying them in archaeological contexts (Fig. 4C). Future research should also examine Humulus lupulus, another economically important member of the family Cannabaceae, to determine whether phytoliths can distinguish it from Cannabis. We also found phytoliths that are diagnostic for Urtica (or the Urticaceae), a genus known for textile fibres (Bergfjord et al., 2012). The textile is made from the nettle plant stem, which yielded diagnostic phytoliths in our study in the form of silicified microhairs (Acicular bulbous). Phytoliths may therefore be a suitable proxy to detect nettle fibres. Lastly, in our study, we did not find diagnostic phytoliths for Linum (producer of flaxseed oil), though previous research suggests they exist (Albert, 2000; Risberg et al., 2002).

Additionally, we observed diagnostic epidermal cells in Papaver, a genus containing species with medicinal qualities. Specifically, Polygonal papillar were found in both P. somniferum (cultivated) and P. argemone (wild), indicating it cannot differentiate between the species. This suggests that identifying cultivated fields with P. somniferum might be difficult, as P. argemone, a common weed, also grows in cultivated fields.

Lastly, plant taxa within the family Asteraceae are often indicative of open and disturbed habitats. Considering the highly diagnostic nature of several Asteraceae species, they may serve as markers for anthropogenic disturbance of natural lands (Piperno and Jones, 2003; Piperno, 2006). In this study, elongated ‘anticlinal epidermis’ (Epi-2) and ‘articulated trichome’ (Tri-2) with squarish segments and thick cell walls are considered diagnostic for the Asteraceae.

The implications of compositional patterns within phytolith assemblages.

Aside from the phytolith production and the taxonomic diagnostic character of morphotypes and morphotype features, we were interested in finding discriminative patterns in the overall phytolith assemblage compositions. The results demonstrated that different plant parts can, to a certain extent, be differentiated based on their phytolith assemblage composition. The LDAs effectively separated herbaceous stems and woody stems, whereas leaves, inflorescences, seeds and fruits clustered together (Fig. 6). Thus, edible parts of dicotyledons cannot easily be distinguished from non-edible parts based on phytolith assemblages, such that phytolith analysis may not be useful to directly answer research questions about dicot plant diet, subsistence strategies, food storage and crop by-products. Nevertheless, there are cases where identification of plant parts does not rely on phytolith assemblages as a whole. Unique morphotypes, such as those observed for Cannabis sativa seed phytoliths, can potentially aid in identification of specific seeds (Fig. 4C), and the morphotype ‘transparent polyhedral sheet’ (M-11) has proven to be diagnostic for woody stems. To test the broad applicability of our results, future studies should explore the phytolith production and potential morphotype overlap of plant parts such as wood, bark and roots. These plant organs remain understudied and relatively unknown in terms of phytolith production and assemblages (but see Albert, 2000; Collura and Neumann, 2017) but could contain diagnostic morphotypes, add important noise to statistical analyses or both.

Phytolith assemblage compositions were effective at distinguishing life forms (shrubs/trees vs. forbs) in this study (Fig. 7). An and Xie (2022) noted that, in most cases, the dicotyledon phytoliths described in Bozarth (1992) cannot be used to differentiate dicotyledonous herbs from broadleaved trees. Despite overlap in the morphotype distribution of forbs and shrubs/trees, we demonstrated that LDA could discriminate life forms for most samples based on differences in relative frequencies rather than unique morphotypes linked to life form. With this type of information, differentiation between forested areas and non-forested areas is possible, making phytoliths a suitable proxy for identification of land use, which aligns with findings from studies conducted on soil samples (e.g. Nogué et al., 2017; Witteveen et al., 2023). The forest vs. non-forest distinction can shed light on the process of anthropogenic alteration of the natural environment in the form of woodland clearing to establish cropland and pasture, and the exploitation of forest for fuel and construction materials (Kaplan et al., 2009). Similarly, analysis of fossil phytoliths may help detect ecological succession or spatial variation in vegetation structure not related to humans, on Quaternary timescales or potentially even further back.

Recommendations for application of phytolith baseline study results and creating new baselines

This study was a first step in establishing a robust and comprehensive survey of phytoliths produced by dicotyledons in NW Europe. However, more work documenting the silica levels and phytolith morphotype distribution among modern plant taxa occurring in this region is needed to evaluate the diagnostic potential of morphotypes below the level of dicotyledons.

For phytolith laboratory processing, we recommend thoroughly cleaning samples beforehand, particularly for tissues with a low phytolith production level. In our study, we soaked plant tissues overnight in pH-neutral detergent (Aquet) within 50-mL centrifuge tubes on a sample shaker to remove contaminants such as soil, diatoms and glue, followed by a tap water rinse. While this method was generally effective, some samples still contained residual contaminants. We suggest that researchers explore varied cleaning protocols, such as sonication of plant tissues, to enhance contaminant removal. However, as sonication may risk damaging phytoliths (e.g. causing fragmentation or breaking silica skeletons), further experiments are needed to assess its effects on phytoliths in modern plant tissues.

For chemical processing, we initially used wet oxidation but switched to a combined method: dry oxidation to remove most organic material, followed by a brief wet oxidation to clear any remaining residues. This change was made because some phytoliths subjected to wet oxidation exhibited visible physical breakage, which hindered morphotype identification. Our findings align with prior studies indicating that dry oxidation causes less breakage (Jenkins, 2009; Wang et al., 2014). However, the combined method also presented challenges, as some samples showed structural alterations – such as darkened or melted silica and crumpled, lightly silicified sheets – probably from dehydration during dry oxidation (500 °C for 6 h). We advise researchers working with modern and fossil phytoliths to consider the limitations of each method, as both impact phytolith preservation and identification.

For scholars looking to utilize or expand upon the current baseline, we recommend consulting Supplementary Data File 3, which includes detailed descriptions of the phytolith assemblages and comparisons to previous species- and genus-level studies. Additionally, we offer a visual atlas of 964 labelled microphotographs with key metadata, to help researchers familiarize themselves with the morphotypes present in this dataset and raw data tables and R code for further analysis. We suggest that they follow our lead in providing full, freely available documentation of observed morphotypes and assemblages, as well as the methods used to extract, study and analyse the phytolith data (Karoune, 2022; Kerfant et al., 2023). Doing so allows future researchers to more easily compare their studies to ours, and avoid the challenges that we encountered during our literature review due to unavailable raw counts or inadequate descriptions of results.

The goal of describing this reference collection of modern phytolith assemblages is to use it for interpreting fossil assemblages. However, when doing so, it is important to exercise caution. For one, our study could not account for taphonomic factors such as uneven transport and preservation among phytolith morphotypes (see e.g. Song et al., 2016; Strömberg et al., 2018). Preservation bias may affect the comparability of modern and fossil assemblages, and we invite scholars to investigate dicotyledon phytolith turnover in more detail before using the segregation patterns of the modern phytolith assemblages as a baseline for fossil datasets. That said, the broad production patterns of taxa should still be informative for anticipating major biases in representation, as was done in Crifò and Strömberg (2021) for plant- and soil-derived phytolith assemblages from modern Neotropical forests. For example, we predict that leaf-derived, well-silicified phytolith morphotypes—such as ‘dicotyledon epidermis’, ‘tracheary elements’, ‘cystoliths’, ‘spheroids’ and ‘blocky polyhedrons’ (see Table 2) – will be preferentially represented in fossil phytolith assemblages. Our findings also indicate considerable variation in phytolith production at both the taxonomic and intra-species levels, which could impact interpretation of fossil assemblages.

Future research should consider region-specific intra-species analyses to assess whether phytolith production patterns are consistent under similar environmental conditions as well as test for the influence of factors such as plant part, tissue maturity, soil silicon availability, precipitation and temperature that may influence production variability (Lanning, 1960; Liu et al., 2016; Brightly et al., 2020). Ideally, studies should include at least two specimens per taxon, and this paper offers a comprehensive framework for comparing intra-species variation in both phytolith production and composition. We recommend that researchers using our results focus on identifying general trends and patterns in phytolith production and morphotype representation rather than on precise quantifications. Unless there is independent evidence for factors that specifically influence phytolith production, prioritizing broader patterns will help prevent over-interpretation of minor quantitative variations (e.g. the ‘general approach’ of Strömberg, 2004).

CONCLUSIONS

This study focused on phytoliths extracted from a collection of modern dicotyledon plant taxa that occur in numerous archaeological deposits in NW Europe. We sought to characterize the level of production of phytoliths and the composition of the resulting phytolith assemblages. We also tested to what degree plant parts/organ, taxonomic group and life form influenced their production and composition.

We found that taxonomy influences both phytolith production and phytolith morphologies in dicotyledon taxa common to NW Europe. Overall, in our study, we observed a wide range of phytolith morphotypes, with some types abundantly appearing across samples and others occurring less frequently or even having rare occurrences. The potential for taxonomic differentiation below the dicotyledon level lies either in non-redundant morphotypes or in redundant types that exhibit distinct, diagnostic features. Several plant families produce diagnostic phytoliths below dicotyledon level, including Asteraceae, Boraginaceae, Brassicaceae, Cannabaceae, Papaveraceae, Polygonaceae, Rosaceae, Solanaceae, Urticaceae and Viburnaceae.

Our study also showed that differentiation between life form (forbs vs. shrubs/trees) is possible based on phytolith assemblage composition. However, other observed variation in phytolith production probably relates to the fact that we used samples from a wide range of environmental and geographical contexts. The study underscores the importance of further research to better understand the diagnostic value of phytolith morphotypes through additional modern plant-based phytolith investigations, and the complexities of phytolith preservation and turnover for the study region of interest. We therefore encourage other researchers to develop phytolith baselines on a local or regional level and test for differences in phytolith production when factors such as geographical region, season and soil type are held constant. When local or regional baselines do not exist for a region, it is advisable to employ an approach that uses broad phytolith production trends rather than focusing on precise quantification to avoid over-interpretation. Our study adds to a growing body of work that will promote understanding of such general patterns in phytolith production among dicotyledons.

SUPPLEMENTARY DATA

Supplementary data are available at Annals of Botany online and consist of the following.

Data File 1: Detailed methods and results. Data File 2: Detailed descriptions and images of newly described morphotypes. Data File 3: Phytolith catalogue: description of the assemblages of all samples, with a comparison to previous descriptions at species and genus level.

mcae217_suppl_Supplementary_File_S1

mcae217_suppl_supplementary_file_s1.docx^{(71KB, docx)}

mcae217_suppl_Supplementary_File_S2

mcae217_suppl_supplementary_file_s2.docx^{(91.4MB, docx)}

mcae217_suppl_Supplementary_File_S3

mcae217_suppl_supplementary_file_s3.docx^{(49.6MB, docx)}

ACKNOWLEDGEMENTS

We thank the University of Washington Herbarium (WTU) and the William and Linda Steere Herbarium of the New York Botanical Garden (NY) for providing plant material from their collections; Dr Lara Stas for her advice on the data analyses; and Brussels Capital Region.

Contributor Information

Rosalie Hermans, Archaeology, Environmental Changes and Geo-Chemistry Research Group, Vrije Universiteit Brussel, Pleinlaan 2, Elsene 1050, Belgium.

Caroline A E Strömberg, Department of Biology and Burke Museum of Natural History and Culture, University of Washington, Seattle, WA 98195, USA.

Tessi Löffelmann, Archaeology, Environmental Changes and Geo-Chemistry Research Group, Vrije Universiteit Brussel, Pleinlaan 2, Elsene 1050, Belgium.

Luc Vrydaghs, Archaeology, Environmental Changes and Geo-Chemistry Research Group, Vrije Universiteit Brussel, Pleinlaan 2, Elsene 1050, Belgium.

Lien Speleers, Royal Belgian Institute of Natural Sciences, Quaternary Environments and Humans, Rue Vautier 29, Brussels 1000, Belgium.

Alexandre Chevalier, Royal Belgian Institute of Natural Sciences, Quaternary Environments and Humans, Rue Vautier 29, Brussels 1000, Belgium.

Karin Nys, Archaeology, Environmental Changes and Geo-Chemistry Research Group, Vrije Universiteit Brussel, Pleinlaan 2, Elsene 1050, Belgium.

Christophe Snoeck, Archaeology, Environmental Changes and Geo-Chemistry Research Group, Vrije Universiteit Brussel, Pleinlaan 2, Elsene 1050, Belgium.

CONFLICT OF INTEREST

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

FUNDING

This work was supported by a PhD fellowship to R.H. by the Research Foundation – Flanders (Fonds voor Wetenschappelijk Onderzoek, FWO) (11A8922N), and a PhD fellowship to R.H. by the Belgian American Education Foundation (BAEF).

DATA AVAILABILITY

Raw data tables, the analysis script (R), plant tissue sample images and 964 labelled microphotographs are publicly available via the online repository (Zenodo) for this project: https://doi.org/10.5281/zenodo.14269045

LITERATURE CITED

Albert RM. 2000. Study of ash layers through phytolith analyses from the Middle Paleolithic levels of Kebara and Tabun caves. Thesis, University of Barcelona, 250.
Albert RM, Weiner S.. 2001. Study of phytoliths in prehistoric ash layers from Kebara and Tabun caves using a quantitative approach. In: Phytoliths: applications in earth sciences and human history. Lisse: A.A. Balkema, 251–266. [Google Scholar]
Amos G. 1952. Silica in timbers. Vol. 267. Australia: Commonwealth Scientific and Industrial Research Organization. [Google Scholar]
An X, Xie B.. 2022. Phytoliths from woody plants: a review. Diversity 14: 339. https://www.mdpi.com/1424-2818/14/5/339 [Google Scholar]
Anderson MJ. 2017. Permutational multivariate analysis of variance (PERMANOVA). In: Balakrishnan N, Colton T, Everitt B, Piegorsch W, Ruggeri F, Teugels JL, eds. Wiley StatsRef: statistics reference online. Wiley, 1–15. doi: https://doi.org/ 10.1002/9781118445112.stat07841 [DOI] [Google Scholar]
Bakels C. 1999. Archaeobotanical investigations in the Aisne valley, northern France, from the Neolithic up to the early Middle Ages. Vegetation History and Archaeobotany 8: 71–77. doi: https://doi.org/ 10.1007/BF02042844 [DOI] [Google Scholar]
Bartoli F, Wilding LP.. 1980. Dissolution of biogenic opal as a function of its physical and chemical properties. Soil Science Society of America Journal 44: 873–878. doi: https://doi.org/ 10.2136/sssaj1980.03615995004400040043x [DOI] [Google Scholar]
Beck HE, Zimmermann NE, McVicar TR, Vergopolan N, Berg A, Wood EF.. 2018. Present and future Köppen-Geiger climate classification maps at 1-km resolution. Scientific Data 5: 180214. doi: https://doi.org/ 10.1038/sdata.2018.214 [DOI] [PMC free article] [PubMed] [Google Scholar]
Bergfjord C, Mannering U, Frei KM, et al. 2012. Nettle as a distinct Bronze Age textile plant. Scientific Reports 2: 664. doi: https://doi.org/ 10.1038/srep00664 [DOI] [PMC free article] [PubMed] [Google Scholar]
Blackman E. 1971. Opaline silica bodies in the range grasses of southern Alberta. Canadian Journal of Botany 49: 769–781. [Google Scholar]
Borderie Q, Ball T, Banerjea R, et al. 2020. Early Middle Ages houses of Gien (France) from the inside: geoarchaeology and archaeobotany of 9th–11th c. floors. Environmental Archaeology 25: 151–169. [Google Scholar]
Bozarth SR. 1985. Distinctive phytoliths from various dicot species. Phytolitharien Newsletter 3: 7–8. [Google Scholar]
Bozarth, S. R. 1992. Classification of opal phytoliths formed in selected dicotyledons native to the Great Plains. In: Phytolith systematics: emerging issues. Berlin: Springer, 193–214. [Google Scholar]
Braadbaart F, van Brussel T, van Os B, Eijskoot Y.. 2017. Fuel remains in archaeological contexts: experimental and archaeological evidence for recognizing remains in hearths used by Iron Age farmers who lived in peatlands. The Holocene 27: 1682–1693. [DOI] [PMC free article] [PubMed] [Google Scholar]
Bremond L, Alexandre A, Véla E, Guiot J.. 2004. Advantages and disadvantages of phytolith analysis for the reconstruction of Mediterranean vegetation: an assessment based on modern phytolith, pollen and botanical data (Luberon, France). Review of Palaeobotany and Palynology 129: 213–228. [Google Scholar]
Brightly WH, Hartley SE, Osborne CP, Simpson KJ, Strömberg CA.. 2020. High silicon concentrations in grasses are linked to environmental conditions and not associated with C4 photosynthesis. Global Change Biology 26: 7128–7143. [DOI] [PubMed] [Google Scholar]
Brombacher C, Hecker D.. 2015. Agriculture, food and environment during Merovingian times: plant remains from three early medieval sites in northwestern Switzerland. Vegetation History and Archaeobotany 24: 331–342. doi: https://doi.org/ 10.1007/s00334-014-0460-4 [DOI] [Google Scholar]
Cabanes D, Shahack-Gross R.. 2015. Understanding fossil phytolith preservation: the role of partial dissolution in paleoecology and archaeology. PLoS One 10: e0125532. [DOI] [PMC free article] [PubMed] [Google Scholar]
Carnelli AL, Theurillat JP, Madella M.. 2004. Phytolith types and type-frequencies in subalpine–alpine plant species of the European Alps. Review of Palaeobotany and Palynology 129: 39–65. [Google Scholar]
Castilla‑Beltran A, Fernandez‑Palacios E, Vrydaghs L, Mallol C, Fernandez‑Palacios J-M, de Nascimento L.. 2024. Phytoliths from modern plants in the Canary Islands as a reference for the reconstruction of long‑term vegetation change and culture‑environment interactions. Vegetation History and Archaeobotany 33: 705–723. doi: https://doi.org/ 10.1007/s00334-024-00995-9 [DOI] [Google Scholar]
Collura LV, Neumann K.. 2017. Wood and bark phytoliths of West African woody plants. Quaternary International 434: 142–159. [Google Scholar]
Crifò C, Strömberg CA.. 2021. Spatial patterns of soil phytoliths in a wet vs. dry neotropical forest: Implications for paleoecology. Palaeogeography, Palaeoclimatology, Palaeoecology 562: 110100. [Google Scholar]
Dekoninck M, Deforce K, Kaal J, et al. 2024. Fuelling the Roman salt industry. Developing a new multiproxy approach to identify peat fuel from archaeological combustion residue. Journal of Archaeological Science 161: 105892. [Google Scholar]
Delhon C, Alexandre A, Berger JF, Thiebault S, Brochier JL, Meunier JD.. 2003. Phytolith assemblages as a promising tool for reconstructing Mediterranean Holocene vegetation. Quaternary Research 59: 48–60. [Google Scholar]
Devos Y, Vrydaghs L, Degraeve A, Fechner K.. 2009. An archaeopedological and phytolitarian study of the ‘Dark Earth’ on the site of Rue de Dinant (Brussels, Belgium). Catena 78: 270–284. [Google Scholar]
Devos Y, Nicosia C, Vrydaghs L, Modrie S.. 2013a. Studying urban stratigraphy: Dark Earth and a microstratified sequence on the site of the Court of Hoogstraeten (Brussels, Belgium). Integrating archaeopedology and phytolith analysis. Quaternary International 315: 147–166. [Google Scholar]
Devos Y, Nicosia C, Vrydaghs L, et al. 2017. An integrated study of Dark Earth from the alluvial valley of the Senne river (Brussels, Belgium). Quaternary International 460: 175–197. [Google Scholar]
Devos Y, Vrydaghs L, Collette O, Hermans R, Loicq S.. 2022. Understanding the formation of buried urban Anthrosols and Technosols: an integrated soil micromorphological and phytolith study of the Dark Earth on the Mundaneum site (Mons, Belgium). Catena 215: 106322. [Google Scholar]
Devos Y, Wouters B, Vrydaghs L, Tys D, Bellens T, Schryvers A.. 2013b. A soil micromorphological study on the origins of the early medieval trading centre of Antwerp (Belgium). Quaternary International 315: 167–183. [Google Scholar]
Fernández Honaine M, Borrelli NL, Martinez Tosto AC.. 2023. A review of anatomical and phytolith studies of cystoliths: silica-calcium phytoliths in dicotyledonous angiosperms. Botanical Journal of the Linnean Society 202: 149–165. [Google Scholar]
Fernández Honaine M, De Rito M, Osterrieth M.. 2018. Análisis de los silicofitolitos presentes en especies de las familias Cannabaceae, Moraceae y Urticaceae del SE bonaerense y estudio comparativo de los cistolitos. Boletín de la Sociedad Argentina de Botánica 53: 1–10. [Google Scholar]
Fernández Honaine M, Zucol AF, Osterrieth ML.. 2006. Phytolith assemblages and systematic associations in grassland species of the South-Eastern Pampean Plains, Argentina. Annals of Botany 98: 1155–1165. [DOI] [PMC free article] [PubMed] [Google Scholar]
Gallego L, Distel RA.. 2004. Phytolith assemblages in grasses native to central Argentina. Annals of Botany 94: 865–874. [DOI] [PMC free article] [PubMed] [Google Scholar]
Ge Y, Lu H, Wang C, Gao X.. 2020. Phytoliths in selected broad-leaved trees in China. Scientific Reports 10: 15577. [DOI] [PMC free article] [PubMed] [Google Scholar]
Gebhardt A, Mansuy-Huault L, Robin V, Vrydaghs L, Lorgeoux C, Poszwa A.. 2023. Paleoenvironmental study of modern charcoal making activity on forest soils in the Northern Vosges Mountains (Bitche, France): a multidisciplinary study of two remaining charcoal platforms and associated soils sequences. Geoarchaeology 39: 187–211. doi: https://doi.org/ 10.1002/gea.21986 [DOI] [Google Scholar]
Geis JW. 1973. Biogenic silica in selected species of deciduous angiosperms. Soil Science 116: 113–130. [Google Scholar]
Hald MM, Magnussen B, Appel L, et al. 2020. Fragments of meals in eastern Denmark from the Viking Age to the Renaissance: new evidence from organic remains in latrines. Journal of Archaeological Science: Reports 31: 102361. doi: https://doi.org/ 10.1016/j.jasrep.2020.102361 [DOI] [Google Scholar]
Hart TC. 2007. A stroll through the park: evaluating the usefulness of phytolith and starch remains found on medieval sherds from Wicken, Northamptonshire, England. Thesis, University of Missouri-Columbia. [Google Scholar]
Hart TC. 2011. Evaluating the usefulness of phytoliths and starch grains found on survey artifacts. Journal of Archaeological Science 38: 3244–3253. [Google Scholar]
Harvey EL, Fuller DQ.. 2005. Investigating crop processing using phytolith analysis: the example of rice and millets. Journal of Archaeological Science 32: 739–752. doi: https://doi.org/ 10.1016/j.jas.2004.12.010 [DOI] [Google Scholar]
Helweg KK. 2020. Gardening at medieval farmsteads: archaeobotanical indications of horticulture in Denmark and southern Sweden, AD 1000–500. Archaeobotanical Studies of Past Plant Cultivation in Northern Europe 5: 119–130. [Google Scholar]
Hodson MJ, White PJ, Mead A, Broadley MR.. 2005. Phylogenetic variation in the silicon composition of plants. Annals of Botany 96: 1027–1046. doi: https://doi.org/ 10.1093/aob/mci255 [DOI] [PMC free article] [PubMed] [Google Scholar]
Holden JE, Finch WH, Kelley K.. 2011. A comparison of two-group classification methods. Educational and Psychological Measurement 71: 870–901. [Google Scholar]
Hondelink MMA, Schepers M.. 2020. The common and the rare: a review of Early Modern Dutch plant food consumption based on archaeobotanical urban cesspit data. Vegetation History and Archaeobotany 29: 553–565. doi: https://doi.org/ 10.1007/s00334-019-00766-x [DOI] [Google Scholar]
Horn HS. 1966. Measurement of ‘Overlap’ in comparative ecological studies. The American Naturalist 100: 419–424. doi: https://doi.org/ 10.1086/282436 [DOI] [Google Scholar]
ICPT. 2019. International code for phytolith nomenclature (ICPN) 2.0. Annals of Botany 124: 189–199. doi: https://doi.org/ 10.1093/aob/mcz064 [DOI] [PMC free article] [PubMed] [Google Scholar]
Iriarte J, Paz EA.. 2009. Phytolith analysis of selected native plants and modern soils from southeastern Uruguay and its implications for paleoenvironmental and archeological reconstruction. Quaternary International 193: 99–123. [Google Scholar]
Jenkins E. 2009. Phytolith taphonomy: a comparison of dry ashing and acid extraction on the breakdown of conjoined phytoliths formed in Triticum durum. Journal of Archaeological Science 36: 2402–2407. doi: https://doi.org/ 10.1016/j.jas.2009.06.028 [DOI] [Google Scholar]
Kaplan JO, Krumhardt KM, Zimmermann N.. 2009. The prehistoric and preindustrial deforestation of Europe. Quaternary Science Reviews 28: 3016–3034. doi: https://doi.org/ 10.1016/j.quascirev.2009.09.028 [DOI] [Google Scholar]
Karoune E. 2022. Assessing open science practices in phytolith research. Open Quaternary 8: 20220310. [Google Scholar]
Katz O. 2015. Silica phytoliths in angiosperms: phylogeny and early evolutionary history. The New Phytologist 208: 642–646. Retrieved from https://www.jstor.org/stable/newphytologist.208.3.642 [DOI] [PubMed] [Google Scholar]
Kealhofer L, Piperno DR.. 1998. Opal phytoliths in Southeast Asian flora. Washington, DC: Smithsonian Institution Press, 1–39. [Google Scholar]
Kerfant C, Ruiz-Pérez J, García-Granero JJ, Lancelotti C, Madella M, Karoune E.. 2023. A dataset for assessing phytolith data for implementation of the FAIR data principles. Scientific Data 10: 479. [DOI] [PMC free article] [PubMed] [Google Scholar]
Lanning F. 1960. Nature and distribution of silica in strawberry plants. Paper presented at the Proceedings American Society of Horticultural Science 76: 349–358. [Google Scholar]
Liu L, Jie D, Liu H, et al. 2016. Assessing the importance of environmental factors to phytoliths of Phragmites communis in north-eastern China. Ecological Indicators 69: 500–507. doi: https://doi.org/ 10.1016/j.ecolind.2016.05.009 [DOI] [Google Scholar]
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: 91–113. [DOI] [PMC free article] [PubMed] [Google Scholar]
Metcalfe CR, Chalk L.. 1950. Anatomy of the Dicotyledons: leaves, stem, and wood, in relation to taxonomy, with notes on economic uses. Oxford: Clarendon Press. [Google Scholar]
Meunier JD, Colin F, Alarcon C.. 1999. Biogenic silica storage in soils. Geology 27: 835–838. doi: https://doi.org/ [DOI] [Google Scholar]
Moffett L. 2018. The archaeobotany of Late Medieval plant remains: the resource and the research. In: The Oxford Handbook of Later Medieval Archaeology in Britain, Oxford Handbooks. Oxford Academic. doi: https://doi.org/ 10.1093/oxfordhb/9780198744719.013.63 [DOI] [Google Scholar]
Morris LR, West NE, Baker FA, Van Miegroet H, Ryel RJ.. 2009. Developing an approach for using the soil phytolith record to infer vegetation and disturbance regime changes over the past 200 years. Quaternary International 193: 90–98. doi: https://doi.org/ 10.1016/j.quaint.2007.08.040 [DOI] [Google Scholar]
Mulholland SC, Rapp G Jr, Ollendorf AL.. 1988. Variation in phytoliths from corn leaves. Canadian Journal of Botany 66: 2001–2008. [Google Scholar]
Mulholland SC, Rapp G Jr, Ollendorf AL, Regal R.. 1990. Variation in phytolith assemblages within a population of corn (cv. Mandan Yellow Flour). Canadian Journal of Botany 68: 1638–1645. [Google Scholar]
Nogué S, Whicher K, Baker AG, Bhagwat SA, Willis KJ.. 2017. Phytolith analysis reveals the intensity of past land use change in the Western Ghats biodiversity hotspot. Quaternary International 437: 82–89. [Google Scholar]
Novello A, Bamford MK, van Wijk Y, Wurz S.. 2018. Phytoliths in modern plants and soils from Klasies River, Cape Region (South Africa). Quaternary International 464: 440–459. doi: https://doi.org/ 10.1016/j.quaint.2017.10.009 [DOI] [Google Scholar]
Novello A, Barboni D, Berti-Equille L, Mazur J-C, Poilecot P, Vignaud P.. 2012. Phytolith signal of aquatic plants and soils in Chad, Central Africa. Review of Palaeobotany and Palynology 178: 43–58. doi: https://doi.org/ 10.1016/j.revpalbo.2012.03.010 [DOI] [Google Scholar]
Out WA, Hasler M, Portillo M, Bagge MS.. 2022. The potential of phytolith analysis to reveal grave goods: the case study of the Viking-age equestrian burial of Fregerslev II. Vegetation History and Archaeobotany. doi: https://doi.org/ 10.1007/s00334-022-00881-2 [DOI] [Google Scholar]
Parry D, Hodson M, Sangster A.. 1984. Some recent advances in studies of silicon in higher plants. Philosophical Transactions of the Royal Society of London. B, Biological Sciences 304: 537–549. [Google Scholar]
Parry D, O’Neill C, Hodson M.. 1986. Opaline silica deposits in the leaves of Bidens pilosa L. and their possible significance in cancer. Annals of Botany 58: 641–647. [Google Scholar]
Pearce M, Ball TB.. 2020. A study of phytoliths produced by selected native plant taxa commonly used by Great Basin Native Americans. Vegetation History and Archaeobotany 29: 213–228. doi: https://doi.org/ 10.1007/s00334-019-00738-1 [DOI] [Google Scholar]
Persaits G, Gulyás S, Náfrádi K, Sümegi P, Szalontai C.. 2015. Phytolith aided paleoenvironmental studies from the Dutch Neolithic. Open Geosciences 7: 20150049. [Google Scholar]
Piperno DR. 1988. Phytolith analysis: an archaeological and geological perspective. San Diego, CA: Academic Press. [DOI] [PubMed] [Google Scholar]
Piperno DR. 2006. Phytoliths: a comprehensive guide for archaeologists and paleoecologists. Lanham, New York, Toronto, Oxford: AltaMira Press (Rowman & Littlefield), ix + 238. [Google Scholar]
Piperno DR, Jones JG.. 2003. Paleoecological and archaeological implications of a late Pleistocene/Early Holocene record of vegetation and climate from the Pacific coastal plain of Panama. Quaternary Research 59: 79–87. doi: https://doi.org/ 10.1016/S0033-5894(02)00021-2 [DOI] [Google Scholar]
Piperno DR, McMichael C.. 2020. Phytoliths in modern plants from Amazonia and the Neotropics at large: Implications for vegetation history reconstruction. Quaternary International 565: 54–74. [Google Scholar]
Power RC. 2018. 23. Mesolithic site organisation and fuel use at Derragh: insights from phytoliths. In: McGlynn G, Stuijts I, Stefanini B, eds. The Quaternary the Irish Midlands: 208–214. Dublin: Irish Quaternary Association. [Google Scholar]
Powers A, Padmore J, Gilbertson D.. 1989. Studies of late prehistoric and modern opal phytoliths from coastal sand dunes and machair in northwest Britain. Journal of Archaeological Science 16: 27–45. [Google Scholar]
POWO. 2023. Plants of the World Online. http://www.plantsoftheworldonline.org/(7 January 2025, date last accessed). [Google Scholar]
Prado S, Noble G.. 2024. Flavours of Pictish life: using starch grains and phytoliths to trace late Roman and early medieval culinary traditions. Journal of Archaeological Science: Reports 58: 104695. [Google Scholar]
Risberg J, Bengtsson L, Kihlstedt B, et al. 2002. Siliceous microfossils, especially phytoliths, as recorded in five prehistoric sites in Eastern Middle Sweden. Journal of Nordic Archaeological Science 13: 11–26. [Google Scholar]
Rovner I. 1971. Potential of opal phytoliths for use in paleoecological reconstruction. Quaternary Research 1: 343–359. [Google Scholar]
Sangster AG, Hodson MJ.. 1992. Silica deposition in subterranean organs. In: Phytolith systematics: emerging issues. Berlin: Springer, 239–251. [Google Scholar]
Saul H, Madella M, Fischer A, Glykou A, Hartz S, Craig OE.. 2013. Phytoliths in pottery reveal the use of spice in European prehistoric cuisine. PLoS One 8: e70583. doi: https://doi.org/ 10.1371/journal.pone.0070583 [DOI] [PMC free article] [PubMed] [Google Scholar]
Schoelynck J, De Block P, Van Dyck E, Cooke J.. 2023. Is there silicon in flowers and what does it tell us?. Ecology and Evolution 13: e10630. [DOI] [PMC free article] [PubMed] [Google Scholar]
Song Z, McGrouther K, Wang H.. 2016. Occurrence, turnover and carbon sequestration potential of phytoliths in terrestrial ecosystems. Earth-Science Reviews 158: 19–30. [Google Scholar]
Speleers L, van der Valk JM.. 2017. Economic plants from medieval and post-medieval Brussels (Belgium), an overview of the archaeobotanical records. Quaternary International 436: 96–109. [Google Scholar]
Strömberg CA. 2003. The origin and spread of grass-dominated ecosystems during the Tertiary of North American and how it relates to the evolution of hypsodonty in equids. Berkeley, CA: University of California. [Google Scholar]
Strömberg CAE. 2004. Using phytolith assemblages to reconstruct the origin and spread of grass-dominated habitats in the great plains of North America during the late Eocene to early Miocene. Palaeogeography, Palaeoclimatology, Palaeoecology 207: 239–275. doi: https://doi.org/ 10.1016/j.palaeo.2003.09.028 [DOI] [Google Scholar]
Strömberg CAE, Di Stilio VS, Song Z.. 2016. Functions of phytoliths in vascular plants: an evolutionary perspective. Functional Ecology 30: 1286–1297. doi: https://doi.org/ 10.1111/1365-2435.12692 [DOI] [Google Scholar]
Strömberg CA, Dunn RE, Crifò C, Harris EB.. 2018. Phytoliths in paleoecology: analytical considerations, current use, and future directions. In: Methods in paleoecology: reconstructing Cenozoic terrestrial environments and ecological communities. Cham: Springer, 235–287. [Google Scholar]
Stroup WW, Ptukhina M, Garai J.. 2024. Generalized linear mixed models: modern concepts, methods and applications. 2nd edn. Boca Raton, FL: CRC Press, 555. doi: https://doi.org/ 10.1201/9780429092060 [DOI] [Google Scholar]
Trant PLK, Wouters B, Croix S, Sindbæk SM, Deckers P, Kristiansen SM.. 2024. A multi-proxy geochemical and micromorphological study of the use of space and stratigraphy of a Viking-age house in Ribe, Denmark. Archaeological and Anthropological Sciences 16: 59. doi: https://doi.org/ 10.1007/s12520-024-01962-1 [DOI] [Google Scholar]
Tromp M, Dudgeon JV.. 2015. Differentiating dietary and non-dietary microfossils extracted from human dental calculus: the importance of sweet potato to ancient diet on Rapa Nui. Journal of Archaeological Science 54: 54–63. [Google Scholar]
Tsartsidou G, Lev-Yadun S, Albert R-M, Miller-Rosen A, Efstratiou N, Weiner S.. 2007. The phytolith archaeological record: strengths and weaknesses evaluated based on a quantitative modern reference collection from Greece. Journal of Archaeological Science 34: 1262–1275. [Google Scholar]
Vrydaghs L, Ball TB, Devos Y.. 2016. Beyond redundancy and multiplicity. Integrating phytolith analysis and micromorphology to the study of Brussels Dark Earth. Journal of Archaeological Science 68: 79–88. [Google Scholar]
Wade K, Shillito L-M, Marston JM, Bonsall C.. 2021. Assessing the potential of phytolith analysis to investigate local environment and prehistoric plant resource use in temperate regions: a case study from Williamson’s Moss, Cumbria, Britain. Environmental Archaeology 26: 295–308. [Google Scholar]
Wallis L. 2003. An overview of leaf phytolith production patterns in selected northwest Australian flora. Review of Palaeobotany and Palynology 125: 201–248. [Google Scholar]
Wang X, Jiang H, Shang X, et al. 2014. Comparison of dry ashing and wet oxidation methods for recovering articulated husk phytoliths of foxtail millet and common millet from archaeological soil. Journal of Archaeological Science 45: 234–239. doi: https://doi.org/ 10.1016/j.jas.2014.03.001 [DOI] [Google Scholar]
Watling J, Iriarte J.. 2013. Phytoliths from the coastal savannas of French Guiana. Quaternary International 287: 162–180. [Google Scholar]
Witteveen NH, White C, Sanchez Martinez BA, et al. 2023. Phytolith assemblages reflect variability in human land use and the modern environment. Vegetation History and Archaeobotany 33: 221. [DOI] [PMC free article] [PubMed] [Google Scholar]
Wouters B, Devos Y, Vrydaghs L, Ball T, De Winter N, Reygel P.. 2019. An integrated micromorphological and phytolith study of urban soils and sediments from the Gallo‐Roman town Atuatuca Tungrorum, Belgium. Geoarchaeology 34: 448–466. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

mcae217_suppl_Supplementary_File_S1

mcae217_suppl_supplementary_file_s1.docx^{(71KB, docx)}

mcae217_suppl_Supplementary_File_S2

mcae217_suppl_supplementary_file_s2.docx^{(91.4MB, docx)}

mcae217_suppl_Supplementary_File_S3

mcae217_suppl_supplementary_file_s3.docx^{(49.6MB, docx)}

Data Availability Statement

[CIT0001] Albert RM. 2000. Study of ash layers through phytolith analyses from the Middle Paleolithic levels of Kebara and Tabun caves. Thesis, University of Barcelona, 250.

[CIT0002] Albert RM, Weiner S.. 2001. Study of phytoliths in prehistoric ash layers from Kebara and Tabun caves using a quantitative approach. In: Phytoliths: applications in earth sciences and human history. Lisse: A.A. Balkema, 251–266. [Google Scholar]

[CIT0003] Amos G. 1952. Silica in timbers. Vol. 267. Australia: Commonwealth Scientific and Industrial Research Organization. [Google Scholar]

[CIT0004] An X, Xie B.. 2022. Phytoliths from woody plants: a review. Diversity 14: 339. https://www.mdpi.com/1424-2818/14/5/339 [Google Scholar]

[CIT0005] Anderson MJ. 2017. Permutational multivariate analysis of variance (PERMANOVA). In: Balakrishnan N, Colton T, Everitt B, Piegorsch W, Ruggeri F, Teugels JL, eds. Wiley StatsRef: statistics reference online. Wiley, 1–15. doi: https://doi.org/ 10.1002/9781118445112.stat07841 [DOI] [Google Scholar]

[CIT0006] Bakels C. 1999. Archaeobotanical investigations in the Aisne valley, northern France, from the Neolithic up to the early Middle Ages. Vegetation History and Archaeobotany 8: 71–77. doi: https://doi.org/ 10.1007/BF02042844 [DOI] [Google Scholar]

[CIT0007] Bartoli F, Wilding LP.. 1980. Dissolution of biogenic opal as a function of its physical and chemical properties. Soil Science Society of America Journal 44: 873–878. doi: https://doi.org/ 10.2136/sssaj1980.03615995004400040043x [DOI] [Google Scholar]

[CIT0008] Beck HE, Zimmermann NE, McVicar TR, Vergopolan N, Berg A, Wood EF.. 2018. Present and future Köppen-Geiger climate classification maps at 1-km resolution. Scientific Data 5: 180214. doi: https://doi.org/ 10.1038/sdata.2018.214 [DOI] [PMC free article] [PubMed] [Google Scholar]

[CIT0009] Bergfjord C, Mannering U, Frei KM, et al. 2012. Nettle as a distinct Bronze Age textile plant. Scientific Reports 2: 664. doi: https://doi.org/ 10.1038/srep00664 [DOI] [PMC free article] [PubMed] [Google Scholar]

[CIT0010] Blackman E. 1971. Opaline silica bodies in the range grasses of southern Alberta. Canadian Journal of Botany 49: 769–781. [Google Scholar]

[CIT0011] Borderie Q, Ball T, Banerjea R, et al. 2020. Early Middle Ages houses of Gien (France) from the inside: geoarchaeology and archaeobotany of 9th–11th c. floors. Environmental Archaeology 25: 151–169. [Google Scholar]

[CIT0012] Bozarth SR. 1985. Distinctive phytoliths from various dicot species. Phytolitharien Newsletter 3: 7–8. [Google Scholar]

[CIT0013] Bozarth, S. R. 1992. Classification of opal phytoliths formed in selected dicotyledons native to the Great Plains. In: Phytolith systematics: emerging issues. Berlin: Springer, 193–214. [Google Scholar]

[CIT0014] Braadbaart F, van Brussel T, van Os B, Eijskoot Y.. 2017. Fuel remains in archaeological contexts: experimental and archaeological evidence for recognizing remains in hearths used by Iron Age farmers who lived in peatlands. The Holocene 27: 1682–1693. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CIT0015] Bremond L, Alexandre A, Véla E, Guiot J.. 2004. Advantages and disadvantages of phytolith analysis for the reconstruction of Mediterranean vegetation: an assessment based on modern phytolith, pollen and botanical data (Luberon, France). Review of Palaeobotany and Palynology 129: 213–228. [Google Scholar]

[CIT0016] Brightly WH, Hartley SE, Osborne CP, Simpson KJ, Strömberg CA.. 2020. High silicon concentrations in grasses are linked to environmental conditions and not associated with C4 photosynthesis. Global Change Biology 26: 7128–7143. [DOI] [PubMed] [Google Scholar]

[CIT0017] Brombacher C, Hecker D.. 2015. Agriculture, food and environment during Merovingian times: plant remains from three early medieval sites in northwestern Switzerland. Vegetation History and Archaeobotany 24: 331–342. doi: https://doi.org/ 10.1007/s00334-014-0460-4 [DOI] [Google Scholar]

[CIT0018] Cabanes D, Shahack-Gross R.. 2015. Understanding fossil phytolith preservation: the role of partial dissolution in paleoecology and archaeology. PLoS One 10: e0125532. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CIT0019] Carnelli AL, Theurillat JP, Madella M.. 2004. Phytolith types and type-frequencies in subalpine–alpine plant species of the European Alps. Review of Palaeobotany and Palynology 129: 39–65. [Google Scholar]

[CIT0020] Castilla‑Beltran A, Fernandez‑Palacios E, Vrydaghs L, Mallol C, Fernandez‑Palacios J-M, de Nascimento L.. 2024. Phytoliths from modern plants in the Canary Islands as a reference for the reconstruction of long‑term vegetation change and culture‑environment interactions. Vegetation History and Archaeobotany 33: 705–723. doi: https://doi.org/ 10.1007/s00334-024-00995-9 [DOI] [Google Scholar]

[CIT0021] Collura LV, Neumann K.. 2017. Wood and bark phytoliths of West African woody plants. Quaternary International 434: 142–159. [Google Scholar]

[CIT0022] Crifò C, Strömberg CA.. 2021. Spatial patterns of soil phytoliths in a wet vs. dry neotropical forest: Implications for paleoecology. Palaeogeography, Palaeoclimatology, Palaeoecology 562: 110100. [Google Scholar]

[CIT0023] Dekoninck M, Deforce K, Kaal J, et al. 2024. Fuelling the Roman salt industry. Developing a new multiproxy approach to identify peat fuel from archaeological combustion residue. Journal of Archaeological Science 161: 105892. [Google Scholar]

[CIT0024] Delhon C, Alexandre A, Berger JF, Thiebault S, Brochier JL, Meunier JD.. 2003. Phytolith assemblages as a promising tool for reconstructing Mediterranean Holocene vegetation. Quaternary Research 59: 48–60. [Google Scholar]

[CIT0025] Devos Y, Vrydaghs L, Degraeve A, Fechner K.. 2009. An archaeopedological and phytolitarian study of the ‘Dark Earth’ on the site of Rue de Dinant (Brussels, Belgium). Catena 78: 270–284. [Google Scholar]

[CIT0026] Devos Y, Nicosia C, Vrydaghs L, Modrie S.. 2013a. Studying urban stratigraphy: Dark Earth and a microstratified sequence on the site of the Court of Hoogstraeten (Brussels, Belgium). Integrating archaeopedology and phytolith analysis. Quaternary International 315: 147–166. [Google Scholar]

[CIT0027] Devos Y, Nicosia C, Vrydaghs L, et al. 2017. An integrated study of Dark Earth from the alluvial valley of the Senne river (Brussels, Belgium). Quaternary International 460: 175–197. [Google Scholar]

[CIT0028] Devos Y, Vrydaghs L, Collette O, Hermans R, Loicq S.. 2022. Understanding the formation of buried urban Anthrosols and Technosols: an integrated soil micromorphological and phytolith study of the Dark Earth on the Mundaneum site (Mons, Belgium). Catena 215: 106322. [Google Scholar]

[CIT0029] Devos Y, Wouters B, Vrydaghs L, Tys D, Bellens T, Schryvers A.. 2013b. A soil micromorphological study on the origins of the early medieval trading centre of Antwerp (Belgium). Quaternary International 315: 167–183. [Google Scholar]

[CIT0030] Fernández Honaine M, Borrelli NL, Martinez Tosto AC.. 2023. A review of anatomical and phytolith studies of cystoliths: silica-calcium phytoliths in dicotyledonous angiosperms. Botanical Journal of the Linnean Society 202: 149–165. [Google Scholar]

[CIT0031] Fernández Honaine M, De Rito M, Osterrieth M.. 2018. Análisis de los silicofitolitos presentes en especies de las familias Cannabaceae, Moraceae y Urticaceae del SE bonaerense y estudio comparativo de los cistolitos. Boletín de la Sociedad Argentina de Botánica 53: 1–10. [Google Scholar]

[CIT0032] Fernández Honaine M, Zucol AF, Osterrieth ML.. 2006. Phytolith assemblages and systematic associations in grassland species of the South-Eastern Pampean Plains, Argentina. Annals of Botany 98: 1155–1165. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CIT0033] Gallego L, Distel RA.. 2004. Phytolith assemblages in grasses native to central Argentina. Annals of Botany 94: 865–874. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CIT0034] Ge Y, Lu H, Wang C, Gao X.. 2020. Phytoliths in selected broad-leaved trees in China. Scientific Reports 10: 15577. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CIT0035] Gebhardt A, Mansuy-Huault L, Robin V, Vrydaghs L, Lorgeoux C, Poszwa A.. 2023. Paleoenvironmental study of modern charcoal making activity on forest soils in the Northern Vosges Mountains (Bitche, France): a multidisciplinary study of two remaining charcoal platforms and associated soils sequences. Geoarchaeology 39: 187–211. doi: https://doi.org/ 10.1002/gea.21986 [DOI] [Google Scholar]

[CIT0036] Geis JW. 1973. Biogenic silica in selected species of deciduous angiosperms. Soil Science 116: 113–130. [Google Scholar]

[CIT0038] Hald MM, Magnussen B, Appel L, et al. 2020. Fragments of meals in eastern Denmark from the Viking Age to the Renaissance: new evidence from organic remains in latrines. Journal of Archaeological Science: Reports 31: 102361. doi: https://doi.org/ 10.1016/j.jasrep.2020.102361 [DOI] [Google Scholar]

[CIT0039] Hart TC. 2007. A stroll through the park: evaluating the usefulness of phytolith and starch remains found on medieval sherds from Wicken, Northamptonshire, England. Thesis, University of Missouri-Columbia. [Google Scholar]

[CIT0040] Hart TC. 2011. Evaluating the usefulness of phytoliths and starch grains found on survey artifacts. Journal of Archaeological Science 38: 3244–3253. [Google Scholar]

[CIT0041] Harvey EL, Fuller DQ.. 2005. Investigating crop processing using phytolith analysis: the example of rice and millets. Journal of Archaeological Science 32: 739–752. doi: https://doi.org/ 10.1016/j.jas.2004.12.010 [DOI] [Google Scholar]

[CIT0042] Helweg KK. 2020. Gardening at medieval farmsteads: archaeobotanical indications of horticulture in Denmark and southern Sweden, AD 1000–500. Archaeobotanical Studies of Past Plant Cultivation in Northern Europe 5: 119–130. [Google Scholar]

[CIT0043] Hodson MJ, White PJ, Mead A, Broadley MR.. 2005. Phylogenetic variation in the silicon composition of plants. Annals of Botany 96: 1027–1046. doi: https://doi.org/ 10.1093/aob/mci255 [DOI] [PMC free article] [PubMed] [Google Scholar]

[CIT0044] Holden JE, Finch WH, Kelley K.. 2011. A comparison of two-group classification methods. Educational and Psychological Measurement 71: 870–901. [Google Scholar]

[CIT0045] Hondelink MMA, Schepers M.. 2020. The common and the rare: a review of Early Modern Dutch plant food consumption based on archaeobotanical urban cesspit data. Vegetation History and Archaeobotany 29: 553–565. doi: https://doi.org/ 10.1007/s00334-019-00766-x [DOI] [Google Scholar]

[CIT0046] Horn HS. 1966. Measurement of ‘Overlap’ in comparative ecological studies. The American Naturalist 100: 419–424. doi: https://doi.org/ 10.1086/282436 [DOI] [Google Scholar]

[CIT0047] ICPT. 2019. International code for phytolith nomenclature (ICPN) 2.0. Annals of Botany 124: 189–199. doi: https://doi.org/ 10.1093/aob/mcz064 [DOI] [PMC free article] [PubMed] [Google Scholar]

[CIT0048] Iriarte J, Paz EA.. 2009. Phytolith analysis of selected native plants and modern soils from southeastern Uruguay and its implications for paleoenvironmental and archeological reconstruction. Quaternary International 193: 99–123. [Google Scholar]

[CIT0049] Jenkins E. 2009. Phytolith taphonomy: a comparison of dry ashing and acid extraction on the breakdown of conjoined phytoliths formed in Triticum durum. Journal of Archaeological Science 36: 2402–2407. doi: https://doi.org/ 10.1016/j.jas.2009.06.028 [DOI] [Google Scholar]

[CIT0050] Kaplan JO, Krumhardt KM, Zimmermann N.. 2009. The prehistoric and preindustrial deforestation of Europe. Quaternary Science Reviews 28: 3016–3034. doi: https://doi.org/ 10.1016/j.quascirev.2009.09.028 [DOI] [Google Scholar]

[CIT0051] Karoune E. 2022. Assessing open science practices in phytolith research. Open Quaternary 8: 20220310. [Google Scholar]

[CIT0052] Katz O. 2015. Silica phytoliths in angiosperms: phylogeny and early evolutionary history. The New Phytologist 208: 642–646. Retrieved from https://www.jstor.org/stable/newphytologist.208.3.642 [DOI] [PubMed] [Google Scholar]

[CIT0053] Kealhofer L, Piperno DR.. 1998. Opal phytoliths in Southeast Asian flora. Washington, DC: Smithsonian Institution Press, 1–39. [Google Scholar]

[CIT0054] Kerfant C, Ruiz-Pérez J, García-Granero JJ, Lancelotti C, Madella M, Karoune E.. 2023. A dataset for assessing phytolith data for implementation of the FAIR data principles. Scientific Data 10: 479. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CIT0055] Lanning F. 1960. Nature and distribution of silica in strawberry plants. Paper presented at the Proceedings American Society of Horticultural Science 76: 349–358. [Google Scholar]

[CIT0056] Liu L, Jie D, Liu H, et al. 2016. Assessing the importance of environmental factors to phytoliths of Phragmites communis in north-eastern China. Ecological Indicators 69: 500–507. doi: https://doi.org/ 10.1016/j.ecolind.2016.05.009 [DOI] [Google Scholar]

[CIT0058] 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: 91–113. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CIT0059] Metcalfe CR, Chalk L.. 1950. Anatomy of the Dicotyledons: leaves, stem, and wood, in relation to taxonomy, with notes on economic uses. Oxford: Clarendon Press. [Google Scholar]

[CIT0060] Meunier JD, Colin F, Alarcon C.. 1999. Biogenic silica storage in soils. Geology 27: 835–838. doi: https://doi.org/ [DOI] [Google Scholar]

[CIT0061] Moffett L. 2018. The archaeobotany of Late Medieval plant remains: the resource and the research. In: The Oxford Handbook of Later Medieval Archaeology in Britain, Oxford Handbooks. Oxford Academic. doi: https://doi.org/ 10.1093/oxfordhb/9780198744719.013.63 [DOI] [Google Scholar]

[CIT0062] Morris LR, West NE, Baker FA, Van Miegroet H, Ryel RJ.. 2009. Developing an approach for using the soil phytolith record to infer vegetation and disturbance regime changes over the past 200 years. Quaternary International 193: 90–98. doi: https://doi.org/ 10.1016/j.quaint.2007.08.040 [DOI] [Google Scholar]

[CIT0063] Mulholland SC, Rapp G Jr, Ollendorf AL.. 1988. Variation in phytoliths from corn leaves. Canadian Journal of Botany 66: 2001–2008. [Google Scholar]

[CIT0064] Mulholland SC, Rapp G Jr, Ollendorf AL, Regal R.. 1990. Variation in phytolith assemblages within a population of corn (cv. Mandan Yellow Flour). Canadian Journal of Botany 68: 1638–1645. [Google Scholar]

[CIT0065] Nogué S, Whicher K, Baker AG, Bhagwat SA, Willis KJ.. 2017. Phytolith analysis reveals the intensity of past land use change in the Western Ghats biodiversity hotspot. Quaternary International 437: 82–89. [Google Scholar]

[CIT0066] Novello A, Bamford MK, van Wijk Y, Wurz S.. 2018. Phytoliths in modern plants and soils from Klasies River, Cape Region (South Africa). Quaternary International 464: 440–459. doi: https://doi.org/ 10.1016/j.quaint.2017.10.009 [DOI] [Google Scholar]

[CIT0067] Novello A, Barboni D, Berti-Equille L, Mazur J-C, Poilecot P, Vignaud P.. 2012. Phytolith signal of aquatic plants and soils in Chad, Central Africa. Review of Palaeobotany and Palynology 178: 43–58. doi: https://doi.org/ 10.1016/j.revpalbo.2012.03.010 [DOI] [Google Scholar]

[CIT0068] Out WA, Hasler M, Portillo M, Bagge MS.. 2022. The potential of phytolith analysis to reveal grave goods: the case study of the Viking-age equestrian burial of Fregerslev II. Vegetation History and Archaeobotany. doi: https://doi.org/ 10.1007/s00334-022-00881-2 [DOI] [Google Scholar]

[CIT0069] Parry D, Hodson M, Sangster A.. 1984. Some recent advances in studies of silicon in higher plants. Philosophical Transactions of the Royal Society of London. B, Biological Sciences 304: 537–549. [Google Scholar]

[CIT0070] Parry D, O’Neill C, Hodson M.. 1986. Opaline silica deposits in the leaves of Bidens pilosa L. and their possible significance in cancer. Annals of Botany 58: 641–647. [Google Scholar]

[CIT0071] Pearce M, Ball TB.. 2020. A study of phytoliths produced by selected native plant taxa commonly used by Great Basin Native Americans. Vegetation History and Archaeobotany 29: 213–228. doi: https://doi.org/ 10.1007/s00334-019-00738-1 [DOI] [Google Scholar]

[CIT0072] Persaits G, Gulyás S, Náfrádi K, Sümegi P, Szalontai C.. 2015. Phytolith aided paleoenvironmental studies from the Dutch Neolithic. Open Geosciences 7: 20150049. [Google Scholar]

[CIT0073] Piperno DR. 1988. Phytolith analysis: an archaeological and geological perspective. San Diego, CA: Academic Press. [DOI] [PubMed] [Google Scholar]

[CIT0074] Piperno DR. 2006. Phytoliths: a comprehensive guide for archaeologists and paleoecologists. Lanham, New York, Toronto, Oxford: AltaMira Press (Rowman & Littlefield), ix + 238. [Google Scholar]

[CIT0075] Piperno DR, Jones JG.. 2003. Paleoecological and archaeological implications of a late Pleistocene/Early Holocene record of vegetation and climate from the Pacific coastal plain of Panama. Quaternary Research 59: 79–87. doi: https://doi.org/ 10.1016/S0033-5894(02)00021-2 [DOI] [Google Scholar]

[CIT0076] Piperno DR, McMichael C.. 2020. Phytoliths in modern plants from Amazonia and the Neotropics at large: Implications for vegetation history reconstruction. Quaternary International 565: 54–74. [Google Scholar]

[CIT0077] Power RC. 2018. 23. Mesolithic site organisation and fuel use at Derragh: insights from phytoliths. In: McGlynn G, Stuijts I, Stefanini B, eds. The Quaternary the Irish Midlands: 208–214. Dublin: Irish Quaternary Association. [Google Scholar]

[CIT0078] Powers A, Padmore J, Gilbertson D.. 1989. Studies of late prehistoric and modern opal phytoliths from coastal sand dunes and machair in northwest Britain. Journal of Archaeological Science 16: 27–45. [Google Scholar]

[CIT0079] POWO. 2023. Plants of the World Online. http://www.plantsoftheworldonline.org/(7 January 2025, date last accessed). [Google Scholar]

[CIT0080] Prado S, Noble G.. 2024. Flavours of Pictish life: using starch grains and phytoliths to trace late Roman and early medieval culinary traditions. Journal of Archaeological Science: Reports 58: 104695. [Google Scholar]

[CIT0081] Risberg J, Bengtsson L, Kihlstedt B, et al. 2002. Siliceous microfossils, especially phytoliths, as recorded in five prehistoric sites in Eastern Middle Sweden. Journal of Nordic Archaeological Science 13: 11–26. [Google Scholar]

[CIT0082] Rovner I. 1971. Potential of opal phytoliths for use in paleoecological reconstruction. Quaternary Research 1: 343–359. [Google Scholar]

[CIT0083] Sangster AG, Hodson MJ.. 1992. Silica deposition in subterranean organs. In: Phytolith systematics: emerging issues. Berlin: Springer, 239–251. [Google Scholar]

[CIT0084] Saul H, Madella M, Fischer A, Glykou A, Hartz S, Craig OE.. 2013. Phytoliths in pottery reveal the use of spice in European prehistoric cuisine. PLoS One 8: e70583. doi: https://doi.org/ 10.1371/journal.pone.0070583 [DOI] [PMC free article] [PubMed] [Google Scholar]

[CIT0085] Schoelynck J, De Block P, Van Dyck E, Cooke J.. 2023. Is there silicon in flowers and what does it tell us?. Ecology and Evolution 13: e10630. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CIT0086] Song Z, McGrouther K, Wang H.. 2016. Occurrence, turnover and carbon sequestration potential of phytoliths in terrestrial ecosystems. Earth-Science Reviews 158: 19–30. [Google Scholar]

[CIT0087] Speleers L, van der Valk JM.. 2017. Economic plants from medieval and post-medieval Brussels (Belgium), an overview of the archaeobotanical records. Quaternary International 436: 96–109. [Google Scholar]

[CIT0088] Strömberg CA. 2003. The origin and spread of grass-dominated ecosystems during the Tertiary of North American and how it relates to the evolution of hypsodonty in equids. Berkeley, CA: University of California. [Google Scholar]

[CIT0089] Strömberg CAE. 2004. Using phytolith assemblages to reconstruct the origin and spread of grass-dominated habitats in the great plains of North America during the late Eocene to early Miocene. Palaeogeography, Palaeoclimatology, Palaeoecology 207: 239–275. doi: https://doi.org/ 10.1016/j.palaeo.2003.09.028 [DOI] [Google Scholar]

[CIT0090] Strömberg CAE, Di Stilio VS, Song Z.. 2016. Functions of phytoliths in vascular plants: an evolutionary perspective. Functional Ecology 30: 1286–1297. doi: https://doi.org/ 10.1111/1365-2435.12692 [DOI] [Google Scholar]

[CIT0091] Strömberg CA, Dunn RE, Crifò C, Harris EB.. 2018. Phytoliths in paleoecology: analytical considerations, current use, and future directions. In: Methods in paleoecology: reconstructing Cenozoic terrestrial environments and ecological communities. Cham: Springer, 235–287. [Google Scholar]

[CIT0092] Stroup WW, Ptukhina M, Garai J.. 2024. Generalized linear mixed models: modern concepts, methods and applications. 2nd edn. Boca Raton, FL: CRC Press, 555. doi: https://doi.org/ 10.1201/9780429092060 [DOI] [Google Scholar]

[CIT0093] Trant PLK, Wouters B, Croix S, Sindbæk SM, Deckers P, Kristiansen SM.. 2024. A multi-proxy geochemical and micromorphological study of the use of space and stratigraphy of a Viking-age house in Ribe, Denmark. Archaeological and Anthropological Sciences 16: 59. doi: https://doi.org/ 10.1007/s12520-024-01962-1 [DOI] [Google Scholar]

[CIT0094] Tromp M, Dudgeon JV.. 2015. Differentiating dietary and non-dietary microfossils extracted from human dental calculus: the importance of sweet potato to ancient diet on Rapa Nui. Journal of Archaeological Science 54: 54–63. [Google Scholar]

[CIT0095] Tsartsidou G, Lev-Yadun S, Albert R-M, Miller-Rosen A, Efstratiou N, Weiner S.. 2007. The phytolith archaeological record: strengths and weaknesses evaluated based on a quantitative modern reference collection from Greece. Journal of Archaeological Science 34: 1262–1275. [Google Scholar]

[CIT0096] Vrydaghs L, Ball TB, Devos Y.. 2016. Beyond redundancy and multiplicity. Integrating phytolith analysis and micromorphology to the study of Brussels Dark Earth. Journal of Archaeological Science 68: 79–88. [Google Scholar]

[CIT0097] Wade K, Shillito L-M, Marston JM, Bonsall C.. 2021. Assessing the potential of phytolith analysis to investigate local environment and prehistoric plant resource use in temperate regions: a case study from Williamson’s Moss, Cumbria, Britain. Environmental Archaeology 26: 295–308. [Google Scholar]

[CIT0098] Wallis L. 2003. An overview of leaf phytolith production patterns in selected northwest Australian flora. Review of Palaeobotany and Palynology 125: 201–248. [Google Scholar]

[CIT0099] Wang X, Jiang H, Shang X, et al. 2014. Comparison of dry ashing and wet oxidation methods for recovering articulated husk phytoliths of foxtail millet and common millet from archaeological soil. Journal of Archaeological Science 45: 234–239. doi: https://doi.org/ 10.1016/j.jas.2014.03.001 [DOI] [Google Scholar]

[CIT0100] Watling J, Iriarte J.. 2013. Phytoliths from the coastal savannas of French Guiana. Quaternary International 287: 162–180. [Google Scholar]

[CIT0101] Witteveen NH, White C, Sanchez Martinez BA, et al. 2023. Phytolith assemblages reflect variability in human land use and the modern environment. Vegetation History and Archaeobotany 33: 221. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CIT0102] Wouters B, Devos Y, Vrydaghs L, Ball T, De Winter N, Reygel P.. 2019. An integrated micromorphological and phytolith study of urban soils and sediments from the Gallo‐Roman town Atuatuca Tungrorum, Belgium. Geoarchaeology 34: 448–466. [Google Scholar]

PERMALINK

Phytoliths in dicotyledons occurring in Northwest Europe: establishing a baseline

Rosalie Hermans

Caroline A E Strömberg

Tessi Löffelmann

Luc Vrydaghs

Lien Speleers

Alexandre Chevalier

Karin Nys

Christophe Snoeck

Abstract

Background and Aims

Methods

Key Results

Conclusions

INTRODUCTION

Table 1.

MATERIALS AND METHODS

Selection of plant taxa and specimen

Plant tissue laboratory processing

Data acquisition

Table 2.

Fig. 1.

Data analysis

RESULTS

Phytolith production

Inter-duplicate sample variation.

Plant parts.

Fig. 2.

Taxonomy.

Fig. 3.

Life form.

Phytolith morphotypes

Occurrence of morphotypes.

Fig. 5.

Morphotypes with distinctive features.

Fig. 4.

Phytolith assemblages

Inter-duplicate sample variation.

Discriminating between assemblages (LDA)

Plant part grouping:

Fig. 6.

Life form grouping:

Fig. 7.

Taxonomy grouping:

Fig. 8.

DISCUSSION

Intra-species variation in production and assemblage composition

Phytolith production variation

Taxonomic variation.

Plant part variation.

Taxonomic resolution of dicotyledon phytoliths

The implications of diagnostic morphotypes.

The implications of compositional patterns within phytolith assemblages.

Recommendations for application of phytolith baseline study results and creating new baselines

CONCLUSIONS

SUPPLEMENTARY DATA

ACKNOWLEDGEMENTS

Contributor Information

CONFLICT OF INTEREST

FUNDING

DATA AVAILABILITY

LITERATURE CITED

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases