Introduction

Phytoliths are micro silica bodies produced by plants from silica deposits made in and around the cells1. As phytoliths maintain the shape of the cells and tissue in which they are formed, phytoliths can be taxonomically significant1,2,3. Compared with other plant micro-remains, phytoliths can especially reveal information about Poaceae species, as Poaceae plants produce more phytoliths than most other taxa4,5, and phytoliths can be preserved in sediments where organic material (such as pollen or seed) is typically not well preserved, such as in fire pits (where materials were directly burnt) and highly oxidized soils4,5,6. Thus, phytolith analysis has been a valuable tool for researchers.

Phytoliths are considered to reflect local vegetation due to their in situ deposition7, reference collections of regional scale8,9,10,11,12,13,14,15,16 and certain taxa17,18,19,20,21,22,23 have been shown to be useful for geological and archaeological studies2,3,24,25,26,27,28,29,30,31,32. In recent years, phytolith analysis has helped researchers make much progress in understanding vegetation change in paleoecology33,34,35, the reconstruction of paleoclimate36,37 and the exploitation of plant resources in the early stages of agriculture38,39,40,41,42,43,44. However, as woody plants typically have shown a comparatively low degree of silicification4,5,45, phytoliths in broad-leaved trees have not been extensively studied. While a few phytolith studies involving species of broad-leaved trees from tropical areas and other regions have been conducted12,15,46,47,48,49,50,51,52,53,54, little phytolith research has been conducted on woody taxa from sub-tropical and temperate China4,55,56.

Previous studies commonly illustrated the morphology of phytoliths observed in the leaves of broad-leaved trees by SEM45,46,50,51 or light microscope12,15,48,49,52,55,56,57,58. Some studies also revealed that spherical and elongate types of phytoliths could be found in the stem59, wood and bark60 in some woody plants. However, although many studies provided the morphology of phytoliths observed in broad-leaved trees, there has not been a reliable identification criterion, especially in temperate China. The illustration of phytoliths in broad-leaved trees sometimes was used as identification criteria4, however, no systematic comparison has been made. Thus, the identification of phytoliths from broad-leaved trees in sediments was difficult in practice, which hindered the precise reconstruction of the paleoenvironment and the understanding of woody plant utilization by the ancestors.

To solve the issues on phytoliths in broad-leaved trees, we selected specimens that cover the taxa of common broad-leaved trees in temperate China to carry out the phytolith analysis. In this study, we provided the phytoliths morphology of common broad-leaved trees in China, and several morphotypes were proposed to be potentially diagnostic for the identification of broad-leaved trees in general. Our results could be a valuable tool for the identification of phytoliths both in natural and archaeological sediments, especially in temperate zones that covered by broad-leaved trees.

Material and methods

A total of 110 species, belonging to 33 families Table 1 were collected for analysis. These species were collected from four regions, Changbai Mountain (N 41°40′, E 125°45′), Gongga Mountain (N 30°02′, E 101°57′), Beijing Botanical Garden (N 40°10′, E 116°12′), and Xiamen Botanical Garden (N 24°27′, E 118°05′), during August to October, in the years 2001, 2004, 2015 and 2019, respectively. To investigate the phytolith types and frequencies in these species, the leaves, branchs and fruit were separately treated using a modified wet oxidation method61. Every part (leaf, twig and fruit) of each specimen was cleaned with distilled water in an ultrasonic water bath to remove adhering particles and then dried in an air drying box for 24 h, the dried materials (mostly 5 g, the species with large leaves were used one whole leaf), were cut into smaller parts and placed in separate tubes and the tubes filled with 20 ml (or enough to submerge the materials) saturated nitric acid and left for one night; the next day the tubes with materials were heated in a water bath (at 90 °C) for at least 2 h, then the solutions were centrifuged at 3000 rpm for 10 min. After removing the supernatant, 5 to 10 ml (or enough to submerge the materials) perchloric acid was added to each tube and then heated in the water bath until the solution became clear and transparent; then the solutions were centrifuged and rinsed with distilled water 3 times and then with ethyl alcohol for a last rinse. Then, 3 ml of ethyl alcohol was added into each tube, and mixed using a Vortex Mixers for 30 s to make the residues homogenous. One drop of the mixture from each tube was mounted on separate slides using Canada Balsam for further observation.

Table 1 Information of the studied specimens.
Full size table

Analyses of the phytoliths thus extracted were conducted under a Leica DM 750 microscope at 400 × magnification. For phytolith identification and counting a total of 100 fields (10 × 10, evenly distributed) under the microscope was analyzed on each slide. If no phytoliths were observed in a slide after scanning the whole slide, then another slide was prepared and scanned, and a replica using more dried materials was conducted for a final examination. After observing all the slides, representative phytolith types were chosen to provide photographic images. All morphotypes are described using the International Code for Phytolith Nomenclature 2.0 (ICPN 2.0)62.

The Principal Components Analysis (PCA analysis) was conducted in C2 program63 to study the relationship between phytoliths types and studied species. The Mann–Whitney U test was conducted in R software64 to find out the significance of differences between phytoliths type and production in species from southern and northern China.

Results

Phytoliths types in the studied species

A total of 23 different types of phytoliths were observed in the studied species. Typical phytoliths types are shown in Figs. 1 and 2, and more detailed illustrations of phytoliths produced by each specimen can be found in the Supplementary Figures 1–13. Phytoliths types are described in Table 2.

Figure 1
figure 1

Phytoliths types observed in this study: 1–2. Stomate stellate (Paulownia fargesii and Mahonia bealei, leaf); 3. Elongate brachiate geniculate (Quercus mongolica, leaf); 4. Irregular sinuate (Lespedeza bicolor, leaf); 5. Polygonal tabular (Paulownia fargesii, leaf); 6. Trichome irregular tubercule (Cornus schindleri sub poliophylla, leaf); 7. Trichome bulbous irregular (Smilax sp., leaf); 8. Elongate facetate (Pittosporum truncatum, leaf); 9. Tracheary annulate/facetate geniculate (Pittosporum truncatum, leaf); 10. Tracheary annulate/facetate claviform (Oyama sieboldii, leaf); 11. Tracheary annulate (Rhus potaninii, leaf); 12. Tracheary helical (Mahonia bealei, leaf); 13. Spheroid favose (Cornus controversa, leaf); 14. Elongate entire and Spheriod hollow (Acer oliverianum, leaf), they are often found articulate; 15. Irregular articulated granulate (Aleurites moluccana, fruit husk). Scale bars are 20 μm.

Full size image
Figure 2
figure 2

Phytoliths types observed in this study: 1. Acute bulbosus (Rosa helenae, leaf); 2. Acute uncinate (Smilax sp., leaf); 3. Acute (Leptopus chinensis, leaf); 4. Acute acicular (Morus australis, leaf); 5. Acute echinate (Ficus tikoua, leaf); 6. Hair base (Acer komarovii, leaf); 7. Trichome spheroid plicate/cavate (Euptelea pleiosperma, leaf); 8. Ellipsoidal nodulate ( Populus sp., leaf); 9. Trichome fusiform cavate ( Cornus controversa, leaf). Scale bars are 20 μm.

Full size image
Table 2 Phytoliths types observed in this study.
Full size table

Most phytoliths observed in this study were found in leaves, except for Elongate entire (Fig. 1-10) which were also observed in the vine of Ficus tikoua, the twig of Pittosporum truncatum and Tilia mandshurica, and Irregular articulated granulate (Fig. 1-15), which were only observed in the fruit husk of Aleurites moluccana. Because many phytolith types have the same anatomical origin, to simplify the further analysis, we further classify the phytoliths types into 4 categories or classes:

  • the stomata class, phytoliths that were formed in the stomata in the leaves, which includes the Stomate stellate;

  • the hair tissue class, phytoliths that were formed in the hair tissues in the leaves, which includes the Trichome irregular tubercule, Trichome bulbous irregular, Acute bulbosus, Acute uncinate, Acute, Acute acicular, Acute echinate, Hair base, Trichome spheroid plicate/cavate, Trichome fusiform cavate;

  • the tracheid/vascular tissue class, phytoliths that were formed in the tracheid/vascular tissues in the leaves, which included the Elongate facetate, Tracheary annulate/facetate geniculate, Tracheary annulate/facetate claviform, Tracheary annulate, Tracheary helical;

  • the silicified cell class, phytoliths that were formed in the cells of mesophyll or epidermis in leaves/branches/fruit, which includes the Elongate brachiate geniculate, Irregular sinuate, Polygonal tabular, Spheroid favose, Elongate entire, Spheriod hollow, Ellipsoidal nodulate.

The total count of phytoliths in each specimen and the percentage of phytoliths in each category are reported in Table 3. We carried out a PCA analysis using this set of data, to find out the relationship between the phytoliths types and species. The result is reported in Fig. 3. We note that the spheres (the red spheres) that represent the four categories of phytolith types form a tetrahedron in the coordinate system Fig. 3, with each sphere occupying an apex of the tetrahedron, indicating that the four categories can be clearly separated. We further note that the spheres that represent the species are scattered throughout the coordinate system with their positions reflecting their relationship with the four phytolith type categories. This PCA closest relationship paradigm between phytolith type categories and the species suggests that phytoliths of the stomata class could be more representative of Aceraceae and Ericaceae, phytoliths of the hair tissue class could be more representative of Moraceae, phytoliths of the tracheid/vascular tissue class could be more representative of Tiliaceae and Euphorbiaceae, phytoliths of the silicified cell class could be more representative of Fagaceae, Saxifragaceae, Liliaceae, Magnoliaceae, Cornaceae, Rosaceae and Lauraceae.

Table 3 Phytolith percentage and phytolith count in studied specimens.
Full size table
Figure 3
figure 3

Relationship among specimens and phytoliths using PCA analysis. Red spheres: indicates the types of phytolith; Green spheres: represents specimens collected in southern China; Yellow spheres: represents specimens collected in northern China. The size of the Green and Yellow spheres relates to the total count of phytoliths of the specimen, the larger the sphere the more phytoliths identified in each specimen. The red and black dots are the projection of spheres on different quadrant. Refer to the result part for more details.

Full size image

Phytolith production in the studied species

To evaluate phytolith production in each specimen, we adapted the production index (PI) used by Pearce and Ball (2019)15:

  • NP (non producer): no phytoliths observed

  • R (rare): one or two phytoliths observed

  • U (uncommon): 3–30 phytoliths observed

  • C (common): 30–100 phytoliths observed

  • A (abundant): more than 100 phytoliths observed

Of 110 species we analyzed, 58 produced phytoliths and 52 were non phytolith producers Table 1. The production index for 58 phytolith producers was mostly recognized as abundant (A) and common (C), except for Berberis poiretii (which was rare), and Leptopus chinensis and Zelkova schneideriana (which are uncommon).

Among the phytolith producers, 21 species were collected from Northern China (Changbai Mountain and Beijing) and 37 were from Southern China (Gongga Mountain). To compare phytolith production between the two regions, we applied an independent-samples Mann–Whitney U test using the data in Table 3. The results showed that phytolith production in the stomata class (Sig. = 0.147), the hair tissue class (Sig. = 0.792) and the silicified cell class (Sig. = 0.226) showed no significant differences between the two regions, however, phytolith production in the tracheid/vascular tissue class (Sig. = 0.028) was significantly different between Northern and Southern China. Also, despite some differences in the taxa, the total count of phytolith showed no significant differences between the two regions (Sig. = 0.601). Such results indicated that although the tracheid/vascular tissue class differed between the two regions, the production of most other phytoliths types might not be influenced by regional differences. The differences in the production of the tracheid/vascular tissue class might reflect the different hydrothermal conditions in the two regions.

Discussion and conclusions

It is widely known that in general, woody plants produce fewer phytoliths than grasses4,5. The results of our study are consistent with the previous studies. Only 58 out of the 110 species we analyzed were phytoliths producers. Most of the phytoliths we observed were extracted from leaves, the other plant parts, such as twigs and fruits typically showing a lack of silicification. Phytolith types belonging to the silicified cell class make up the largest portion of the phytoliths produced by the 58 phytolith producing taxa, followed by the stomata class, the tracheid/vascular class and the hair tissue class. Species belonging to the same genus usually produced the same types of phytoliths, and the phytolith production was typically similar. However, we found that phytolith types and production in species belonging to different genera of the same family can be very different. Such results suggest the possibility of identification of taxa on the genus level using phytolith analysis, which is in consist with the study of grasses22, however, studies that involve more species and more samples of species are needed to confirm such findings.

To date, no especially diagnostic types of phytoliths have been identified for broad-leaved trees in general or a certain family. After reviewing other phytolith studies of species belonging to the broad-leaved trees12,15,46,47,48,49,50,51,52,55,56 (also see Table 4, we here propose several phytolith types that have the potential to be diagnostic to broad-leaved trees: Elongate brachiate geniculate (Fig. 1-3), Polygonal tabular (Fig. 1-5), Elongate facetate (Fig. 1-8), Tracheary annulate/facetate geniculate (Fig. 1-9) and Tracheary annulate/facetate claviform (Fig. 1-10). Because these types of phytoliths are rarely seen in grasses and have been extracted from broad-leaved tree taxa in other studies, we suggest that they might have the potential to be diagnostic types for broad-leaved trees. Although some types of phytoliths have distinct morphological differences with other types (such as Trichome irregular tubercule (Fig. 1-6), Trichome spheroid plicate/cavate (Fig. 2-7), Ellipsoidal nodulate (Fig. 2-8) and Trichome fusiform cavate (Fig. 2-9), considering the lack of cross-examination of these types, further studies were needed to evaluate their potential in being diagnostic types. The Acute acicular (Fig. 2-4) and Acute echinate (Fig. 2-5) were only observed in Moraceae plants4,7,48, combined with our results, they might be the potential diagnostic types for Moraceae, while observation of more specimens from Moraceae and other plants was needed to confirm this finding. Although Irregular sinuate phytoliths were observed in many broad-leaved trees, they were also observed in many ferns4,45,54,65, thus they were not proposed as the potential diagnostic types for broad-leaved trees. The Irregular articulated granulate (Fig. 1-15) which we found in the fruit husk of Aleurites moluccana (which could be used as food or sauce in Malaysia and Indonesia), is also noteworthy as it has not been reported yet. Such silicification in fruit husks might be a protection strategy22,66, and the presence of this type may provide insight into ancient plant resource exploitation.

Table 4 Comparison of phytoliths nomenclature and evaluation of their potential in being diagnostic types for broad-leaved trees.
Full size table

In this study, we have provided an illustration of several distinct phytolith types we observed in the common broad-leaved trees in temperate China, and reported that there appears to be little difference in broad-leaved trees phytolith production between the northern and the southern regions. Although we have proposed several specific phytoliths types as potentially diagnostic (which we believe to be reliable), pending further confirming research involving more taxa and samples, researchers should not solely use our findings as identification criteria, but rather as a guidance and reference for the future studies.