Phytoliths in Archaeology: Chemical Aspects
Phytoliths have now become almost a routine aspect of many archaeological investigations. They are very widely used in many contexts to gain information about the plant species grown by or associated with humans in the past. Up until relatively recently most workers have used phytolith morphology to investigate archaeological contexts, and still the majority of papers published take that approach. However, there is increasing interest in using the chemical or even isotopic makeup of phytoliths to provide more information than is available merely from using morphology. The second key area where phytolith chemistry is important in archaeology is taphonomy. In particular, phytoliths with different chemistries might be expected to degrade in soils or sediments at different rates, and soil chemical conditions may also affect degradation. Both these factors could seriously impact on the interpretation of phytolith results from archaeological contexts.
Philippa Ryan wrote an excellent overall introduction to the use of phytoliths in archaeology for the Encyclopedia of Global Archaeology (Ryan 2014), but of necessity the coverage of the more chemical aspects of phytoliths and their potential for use in archaeology was brief. Although the number of archaeological investigations that have made direct use of phytolith chemistry is so far relatively small, the potential for the future is huge. Chemical aspects can also have potentially large effects on the taphonomy of phytoliths. In this entry I will outline the progress in these areas so far.
Ryan (2014) gave an overall definition of what phytoliths are in her paper and, like her, I will also only consider siliceous structures. From a purely chemical point of view phytoliths are deposits of amorphous silica that are formed within some plant cells. Other compounds and elements are often included in their structures, and these will be discussed below.
Phytoliths have been studied since the nineteenth century, but it is not easy to say when investigations of their chemistry began. However, it was undoubtedly Jones and Beavers (1963) who produced the first detailed chemical work. They carried out x-ray diffraction, infrared absorption, differential thermal analysis, specific gravity, and refractive index measurements on phytoliths. Another key figure in the early chemical work on phytoliths was François Bartoli whose work on crystallochemistry was seminal, and is still much cited (Bartoli 1985). In more recent times Carole Perry has been a leading figure in research into the chemistry of phytoliths (Currie and Perry 2007).
Key Issues/Current Debates
Phytoliths are found in leaves, flowers, seeds, stems, and roots in many different cell types. So, unlike the situation with pollen grains, a single plant species can produce many different morphological types of phytoliths. This has its challenges for phytolith analysts, but advances in image analysis and statistical techniques are steadily improving our ability to identify species from the mixture of phytolith types often found in archaeological samples. It is less well recognized that because phytoliths form in different environments in the plant, they are not just morphologically different but chemically heterogeneous. Kumar et al. (2017) assessed the variation between different cell types in grass species. The main deposition sites observed were in the endodermis of the roots, the epidermis of the leaves, and the outer epidermis of the inflorescence bracts. Lesser amounts have been detected in other locations such as the brush hairs at the ends of some cereal grains. There are two main locations where silica deposition leads to formation of phytoliths: the cell wall and the cell lumen. The authors postulated three deposition mechanisms: passive cell wall silicification, controlled cell wall silicification, and deposition in silica cells where the phytolith develops on the external side of a functional plasma membrane. As phytoliths are formed in many different locations in the plant and the mechanisms of their formation are thought to vary, it should not be too surprising that their chemistries differ.
The silica deposited in phytoliths is amorphous with no order down to the level of 10 Å (Currie and Perry 2007; Hodson 2016b), and has the general formula [SiO n/2(OH)4−n ] m , where n = 0 to 4 and m is a large number. Occasional reports of crystalline material are probably either due to contamination or the silica being laid down on ordered cellulose molecules. The silica within cell walls is likely to differ markedly from that laid down in the lumen. Where silica is deposited in cell walls, it is laid down onto a carbohydrate matrix. When it is deposited in the cell lumen, the silica would not be expected to contain as much carbohydrate but might well contain the breakdown products of the protoplast: proteins, lipids, and nucleic acids among them. So the silica structure itself will vary depending on the matrix onto which it is laid down. The carbon within phytoliths has received much recent attention. The two other elements in phytoliths that have gained some interest are aluminum and nitrogen. Stable isotopes of silicon, oxygen, and carbon have all been suggested as potential proxies for use in archaeological work, and radiocarbon dating of phytoliths with 14C has been widely used. All of these parameters could be useful in providing information beyond morphological analysis.
Jones and Beavers (1963) were the first to measure carbon in phytoliths, estimating that those isolated from a Cisne silt loam had 0.86% C. They suggested that this carbon might be occluded within the phytolith and protected from oxidation (so-called phytolith occluded carbon). Since then there have been many attempts to measure carbon in phytoliths. Acid digestion, heating in a muffle furnace at 450 °C or more, and microwave digestion are the three most usual methods. The results vary depending on the methodology employed from 0.1% to 6.0% C or more (Hodson 2016b). Perhaps surprisingly, this has caused one of the major current controversies in phytolith research. For quite some time, researchers around the world have been interested in whether phytoliths might have the potential to sequester substantial amounts of carbon, and thus contribute to the fight against global warming. The percentage of carbon in phytoliths determined in laboratories using different techniques then has a major influence on the amount of carbon that is thought to be sequestered in soil phytoliths at an ecosystem or even a global scale. Song et al. (2017) consider that very low carbon concentrations are due to extreme preparatory techniques causing the disintegration of the phytoliths. Others have suggested that high carbon concentrations are unrealistic as much of the carbon measured will be labile, easily degraded in the soil, and therefore not strongly sequestered. My concern (Hodson 2016a) is that none of the methodologies used necessarily give good estimates of the carbon present in soil phytoliths. This topic might be considered tangential to archaeological concerns, but the carbon concentration of phytoliths in soils or sediments may be a factor in how quickly they degrade (see “Taphonomy” below).
What form does the carbon take in phytoliths? In cell wall phytoliths, such as those considered by Currie and Perry (2007), carbohydrates are major components with smaller amounts of protein and amino acids. The type of carbohydrate present has a significant effect on the morphology of the silica produced. Rather less is known about lumen silica deposits, but it is likely that they are much lower in carbohydrates and potentially higher in lipids from the breakdown of membranes and possibly nucleic acids.
Other Elements in Phytoliths
Elements other than silicon, oxygen, and carbon have also frequently been found in phytoliths (Hodson 2016b), and there has been some interest in their use as potential markers in archaeological contexts, but so far I am unaware of any publications that have taken this idea forward.
Bartoli (1985) was among the first to work on aluminum in phytoliths. He found that phytoliths from conifers had a higher aluminum content than those from deciduous trees. Most of the aluminum seemed to be present on the surfaces of the phytoliths, and conifer phytoliths were significantly less soluble as a result. Unlike many other elements, it is possible for aluminum to substitute for silicon in the structure of a phytolith, isomorphous substitution. Carnelli et al. (2002) conducted a survey of 20 species growing in the Valaisan Swiss Alps. They confirmed that phytoliths isolated from woody taxa were higher in aluminum than those of monocotyledons and suggested that this could be used as a marker for woody taxa in paleoecological studies. However, aluminum concentration in phytoliths does not only depend on whether a species is woody but also whether the soil in which the parent plants are growing is acidic and has high available aluminum for uptake (Hodson 2016b).
Nitrogen was first reported in phytoliths at low concentrations by Jones and Beavers (1963). In wheat, Hodson et al. (2008) found between 0.01% and 0.06% N, depending on the organ that the phytoliths were isolated from. Findings like these suggest that amino acids, proteins, and nucleic acids could be among the nitrogenous compounds sequestered in phytoliths. Alexandre et al. (2015) used nanoSIMS to analyze individual wheat phytoliths and observed that both carbon and nitrogen appeared to be distributed throughout these structures and at an N/C ratio of 0.27. As the authors point out this suggests that amino acids are present. The one major attempt to isolate DNA from phytoliths was not successful, but glycoproteins were found (Elbaum et al. 2009). I suspect that as techniques for the detection of DNA get better and better there will be further attempts at isolating it from phytoliths. Success would revolutionize archaeobotany.
There has been considerable interest in using stable isotopes in biogeochemical investigations and some work on isotopes in plant silica and phytoliths. The elements that have been investigated so far in phytoliths are silicon, oxygen, and carbon.
There are three stable isotopes of silicon, 28Si, 29Si, and 30Si, with 28Si being the most common in the environment. The ratios of these isotopes to each other are relatively constant but change when Si is deposited as a solid from a solution, a so-called fractionation event. In some situations the heavier isotopes come out of solution quicker leaving the solution with a greater proportion of lighter isotopes. However, it appears that in flow through situations the lighter isotopes come out of solution faster, leaving the heavier isotopes at greater concentrations in the solution (Hodson 2016b). So in wheat (Hodson et al. 2008), and all other plant species so far investigated, the heavier isotopes are concentrated higher up the shoot and further along the transpiration stream. This is an interesting and highly consistent observation that could eventually have uses in archaeology, as it suggests a method for distinguishing between stem and leaf phytoliths or stem and inflorescence phytoliths. As yet, however, there have been no such applications. Hodson et al. (2008) concluded “At present, then, we feel that the fractionation of heavy Si isotopes into the upper parts of plants is a finding that is awaiting an application.” That remains the case.
There are three naturally occurring stable isotopes of oxygen, 16O, 17O, and 18O, and the former is by far the most common in the environment. Because it is lighter, 16O, when present in water molecules, is slightly more likely to evaporate than the heavier isotopes. This has led to many applications in geochemistry, and there has been some interest in attempting to relate oxygen isotopes in phytoliths to the environments in which plants are growing (Hodson 2016b). However, there are many factors that influence evaporation of water from plants, and these tend to complicate the oxygen signal derived from actively transpiring tissues such as leaves. This led Alexandre et al. (2012) to investigate wood phytoliths in tropical forest trees in Queensland, Australia, that were growing along altitudinal gradients. These phytoliths should not be influenced by evaporation, and Alexandre et al. found that wood phytolith oxygen isotope signatures were related to mean annual soil water isotopic values and to mean annual temperature.
There are two stable isotopes of carbon, 12C and 13C, with the former being more common. These isotopes are important in paleoecology and archaeology as they can be used to distinguish between plants showing C3 and C4 photosynthesis. C3 plants generally prefer more mesic conditions, while C4 plants are better at surviving dry conditions. So the balance between C3 and C4 plants can give an indication of the aridity of the environment. Plants all discriminate against the 13C isotope when it is taken up as carbon dioxide in photosynthesis, but C3 plants do so more than C4 plants. This means that C3 plants such as beans, rice, and wheat have markedly different isotopic signatures to C4 plants such as maize and sugar cane. This isotopic signal is then passed on to animals, including humans that have eaten either C3 or C4 plant products. Not surprisingly there has been some interest in measuring stable carbon isotopes in phytoliths. In general, this work has been successful and, for example, wheat phytoliths showed carbon isotopic signatures within the expected range for C3 plants (Hodson et al. 2008). However, there have been some problems with phytolith carbon signatures not matching those of the whole plant (Hodson 2016b), and this phenomenon has been even more pronounced with work on 14C.
The Controversy Over 14C
When they discovered that carbon was occluded within phytoliths, Jones and Beavers (1963) pointed out that it might be possible to use this in dating. Trace amounts of the radioactive isotope 14C occur naturally in the environment with a half-life of 5700 years, and this is used routinely in dating of all types of materials. From the late 1960s onwards the technique was quite widely used on phytoliths and seemed to give good results. However, in recent years there have been a number of anomalous results reported where phytolith 14C determinations were giving results that were hundreds or even thousands of years older than they were known to be (Piperno 2016). Phytoliths isolated from some modern plant material seem to give much older dates than the plants themselves. This finding has caused a great controversy, and sometimes a heated one, and as I write in 2017 there is still no definitive answer to the problem.
It is generally recognized that most of the carbon in phytoliths originates from photosynthetic carbon uptake. However, Santos et al. (2012) proposed that “old carbon” taken up by plant roots from the soil is transported through the plant and is incorporated into phytoliths, thereby making the phytoliths apparently older than the rest of the plant. They reviewed the evidence that a small amount of carbon is taken up by plants in this way and that it joins the much larger pool of photosynthetic carbon. My question concerning the “old carbon” hypothesis is the mechanism whereby transport of old carbon is focused towards the phytoliths. This would be required if they were to be apparently older than the plants from which they are derived (Hodson 2016b). I suggested that depletion of 14C in lipids within phytoliths might be a possible explanation for the results.
Piperno (2016) analyzed phytoliths isolated from herbarium specimens of neotropical plants collected between 1964 and 2013. She found that the dates obtained were post-bomb 14C dates, reflecting collection of the samples after 1955 after the beginning of thermonuclear testing. There was no suggestion of dates being older than expected, except for one sample which had originally been treated with organic chemicals made from fossil fuels that were radiocarbon dead. Piperno suspects that many of the problems arise from the methodologies used and that surface contamination with old carbon is a major difficulty.
The search for a reliable methodology was the focus of Asscher et al. (2017). They dissolved silica from phytoliths using mild conditions and released insoluble fractions which gave reliable radiocarbon dates. They compared their phytolith dating results with those for seeds from the same contexts at the archaeological site of Bet Shemesh in Israel. For insoluble material that was extracted from phytoliths where the percentage carbon was greater than 40%, the dating results were identical with the seeds. The authors also compared the percentage of modern carbon in phytoliths extracted from wheat plants (97.76%), from cellulose extracted from the plant (99.78%), and from the soil in which the plants were growing (72.53%). They suggested that the lowered modern carbon in the phytoliths was due to uptake of old carbon from the soil.
The importance of carbon dating of phytoliths is well illustrated by the work of Zuo et al. (2017). They used the technique to take the date of rice cultivation in China back to the beginning of the Holocene period between 9000 and 9400 cal year B.P., the earliest dates seen so far. The authors validated these dates by comparing them with other dated materials from the same level or context. The dates observed were apparently not affected by “old carbon.” Zuo et al. concluded, “We find no evidence to support arguments from a few investigators that old carbon absorbed from soils bias phytolith dates.”
Clearly, we have a major problem here with researchers split into those favoring the “old carbon” hypothesis, and those who think that the difficulties largely arise from methodology. Some look to fractionation events within the plant to explain the phenomenon, and yet others seem to be largely ignoring the issue. Carbon dating of phytoliths is a fundamental issue for many of their applications in archaeological science, and this disagreement needs resolution soon. It seems that the technical problems in obtaining clean samples have largely been overcome, but the problems arising from the “old carbon” hypothesis remain. Why do some researchers find carbon dating of phytoliths is unreliable, but others seem to find no problems? How does “old carbon” from the soil get concentrated into phytoliths? Are other processes within the plant responsible for the observed results? The answers to these questions require detailed investigations of the processes involved in phytolith formation.
Phytoliths with a variety of chemistries are produced by plants, and these find their way into different archaeological contexts. Many are incorporated into the soil after the plant dies. Some plant parts will be harvested and either eaten or used in other ways. Phytoliths can be found associated with human or animal teeth. Sometimes the plant material will be burnt, leaving the phytoliths in the ash. In all of these situations the chemistry of the phytoliths involved will interact with the chemistry of the environment. In some cases the phytoliths will be highly resistant to dissolution and degradation, while in others they will soon disappear from the archaeological record. So the phytoliths remaining in a sample for morphological analysis will be affected to a considerable extent by chemistry.
Some soil conditions are known to increase the dissolution and breakdown of phytoliths, for example, high soil pH and temperature. Prentice and Webb (2016) investigated a pH range from 4.0 to 10.0, and temperatures between 4.0 °C and 44.0 °C and found that dissolution increased with higher pH and temperature. The authors also took this work a stage further by looking at changes in the isotopic composition of phytoliths during dissolution. In the early stages of dissolution the lighter isotopes 28Si and 16O were preferentially removed from the phytoliths leaving them enriched in heavier isotopes. However, once the surrounding solution reached 30–40% saturation, then precipitation of new silica onto the phytoliths began, and lighter isotopes were redeposited. This study suggests that changes in the isotopic composition of phytoliths after they are deposited in soils or sediments need to be taken into account, and that the complexities in using isotopes resulting from processes within the plant are exacerbated in the post-depositional environment.
Bartoli (1985) determined that aluminum deposited within phytoliths decreased their solubility. As we noted above, conifer phytoliths are particularly high in aluminum provided that the trees producing them are growing in acidic soils with reasonably available aluminum. So we might expect that conifer phytoliths would dissolve more slowly than those from other species (e.g., grasses) and so might be overrepresented in the depositional environment.
Another factor that is likely to affect the solubility of phytoliths and how quickly they break down in soils or sediments is their organic matter content. When they are lost from the host plant, cell wall phytoliths will have a higher organic content (principally cellulose and other carbohydrates) than lumen phytoliths. I previously assessed the evidence that when this organic matter is lost from the phytolith, either through acid digestion, burning, or degradation in the soil, the remaining silica is fairly porous (Hodson 2016b). Thus, cell wall phytoliths may well be more susceptible to dissolution and breakdown within the soil. More work is undoubtedly needed in this area.
Osterrieth et al. (2009) investigated the taphonomical factors affecting phytoliths deposited in the loess sediments of the Argentinian Pampas. They considered that the moderately alkaline pH of the loess was a major factor in phytolith dissolution. However, they also recognized that the surface/volume ratio of the phytolith type, the amount of impurities within the silica, and the degree of silicification of the original plant cell were also important. Smaller phytoliths have less surface area when compared to their volume, and the authors noted that small rondel and trapezoid phytoliths have less surface pitting than elongates or bulliforms. The most detailed work so far on phytolith taphonomy was conducted by Cabanes and Shahack-Gross (2015). They investigated dissolution of phytoliths isolated from both four modern plant species and archaeological sediments. The phytolith morphotype assemblages from modern plants subjected to controlled partial dissolution were significantly changed from the original assemblages. So different types of phytolith dissolved at different rates and the principal factor involved seemed to be geometric surface area to bulk ratio. These results have major implications for those wishing to use phytolith analysis on archaeological samples. If some phytolith types disappear from the assemblages over time, this could markedly skew interpretation of the data obtained. In conclusion, it appears that even archaeological work focused on using phytolith morphology needs to take the chemistry of the environment and of the phytoliths into account.
The work outlined in the present entry is of a very interdisciplinary nature, and it is published in a very wide range of journals by scientists from many countries. Unlike many areas within archaeology there is less limitation set by location, and researchers can be anywhere provided that they have a reasonably well-equipped laboratory. There are a number of international meetings where work on the chemistry of phytoliths is discussed. At the latest International Meeting on Phytolith Research (IMPR) held in Aix-en-Provence, France, in 2016, the keynote speaker on phytolith chemistry was Carole Perry. The Integrated Microscopy Approaches in Archaeobotany (IMAA) meeting held in Reading, UK, in February 2017 had many presentations including phytolith work, and mine looked specifically at phytolith development and chemistry. The meetings of Isotopes in Biogenic Silica (IBiS) often feature relevant work, particularly on isotopes, and the most recent meeting was in Girona, Spain, in June 2017.
It has become very clear that phytoliths are deposited in plants in many different locations and that these provide very different chemical environments for deposition (Hodson 2016b; Kumar et al. 2017). Just as the morphology of phytoliths varies within the plant so does the chemistry. Plant scientists are only now beginning to understand this variation. At present our incomplete understanding of the processes involved in phytolith formation is holding up progress in using phytolith chemistry as a proxy in archaeological contexts. This has even recently affected confidence in dating with 14C. With all isotopes we need a better understanding of partitioning between the phytolith and the rest of the plant. Almost certainly phytoliths in soil and sediments will degrade at different rates depending on their chemistry, but again our knowledge in this area is far from complete. What is really needed is a resurgence in interest among plant scientists and chemists in working out the deposition process in plants and how that affects phytolith heterogeneity. As I write in 2017 it appears that this may be beginning to happen, but time will tell.
- Alexandre, A., J. Crespin, F. Sylvestre, C. Sonzogni, and D.W. Hilbert. 2012. The oxygen isotopic composition of phytolith assemblages from tropical rainforest soil tops (Queensland, Australia): Validation of a new paleoenvironmental tool. Climate of the Past 8: 307–324. https://doi.org/10.5194/cp-8-307-2012.CrossRefGoogle Scholar
- Alexandre, A., I. Basile-Doelsch, T. Delhaye, D. Borshneck, J.C. Mazur, P. Reyerson, and G.M. Santos. 2015. New highlights of phytolith structure and occluded carbon location: 3-D X-ray microscopy and NanoSIMS results. Biogeosciences 12: 863–873. https://doi.org/10.5194/bg-12-863-2015.CrossRefGoogle Scholar
- Asscher, Y., S. Weiner, and E. Boaretto. 2017. A new method for extracting the insoluble occluded carbon in archaeological and modern phytoliths: Detection of 14C depleted carbon fraction and implications for radiocarbon dating. Journal of Archaeological Science 78: 57–65. https://doi.org/10.1016/j.jas.2016.11.005.CrossRefGoogle Scholar
- Bartoli, F. 1985. Crystallochemistry and surface properties of biogenic opal. Journal of Soil Science 36: 335–350. https://doi.org/10.1111/j.1365-2389.1985.tb00340.x.CrossRefGoogle Scholar
- Hodson, M. 2016a. Carbon concentrations in phytoliths and carbon sequestration in soil. Journal of Brief Ideas. http://beta.briefideas.org/ideas/c2ee0d247f8fc516a87f2b0f385c6031.
- Piperno, D.R. 2016. Phytolith radiocarbon dating in archaeological and paleoecological research: A case study of phytoliths from modern neotropical plants and a review of the previous dating evidence. Journal of Archaeological Science 68: 54–61. https://doi.org/10.1016/j.jas.2015.06.002.CrossRefGoogle Scholar
- Prentice, A.J., and E.A. Webb. 2016. The effect of progressive dissolution on the oxygen and silicon isotope composition of opal-A phytoliths: Implications for palaeoenvironmental reconstruction. Palaeogeography, Palaeoclimatology, Palaeoecology 453: 42–51. https://doi.org/10.1016/j.palaeo.2016.03.031.CrossRefGoogle Scholar
- Zuo, X., L. Houyuan, J. Leping, J. Zhang, X. Yang, X. Huan, K. He, C. Wang, and N. Wu. 2017. Dating rice remains through phytolith carbon-14 study reveals domestication at the beginning of the Holocene. Proceedings of the National Academy of Science, USA 114: 6486–6491. https://doi.org/10.1073/pnas.1704304114.CrossRefGoogle Scholar