For each slide representing a given sample, researchers tend to randomly count out a certain amount of phytoliths that represent the phytolith assemblage. This is defined by Piperno (1988) as "... the tabulation and quantification in percentages, absolute numbers, or ratios of all morphological variants observed in a sample". The minimum number of phytoliths counted that are considered to be representative of a slide (and therefore of a sample) differs quite widely:
It is clear from this list that the amount of phytoliths counted in each slides differs depending on the goals of the research. There is a general agreement that a good number (a statistically meaningful number) is between 200 and 300 individuals and this is supported by accumulation curves that show that at around 200 individuals the curve starts to flatten (Zurro, in preparation). Researchers dealing with palaeoenvironmental reconstructions often count a greater amount of phytoliths to be able to discuss at least 200 meaningful morphotypes (e.g. grass short cells), like in the works of Barboni et al. (1999).
For archaeological samples, Piperno (1998) recommends that after counting a statistically meaningful number of phytoliths to be between 200 and 300, a quick scan is carried out to assess the presence of rare, but often very important, morphologies. These morphologies are mostly related to plants with low phytolith production or low incidence in the settlement.
According to the work done by van der Veen & Fieller (1982) on charred seeds, the required sample size depends on four variables (that are still valid when counting phytoliths):
The following table was created using the values from an example according to van der Veen & Fieller (1982, Table 4). Each phytolith type in a phytolith assemblage would probably account for around 20% of that assemblage.
These values are very similar to those routinely used in phytolith analysis (see above) and if a quick scan is also added most or all of the rare forms will be observed.
The calculation of phytolith concentration is an important step in the interpretation of archaeological phytolith assemblages. The total number of phytoliths present in a preparation can be estimated by counting all phytoliths in a subsample of known weight or volume, or by adding an exotic particle to the prepartion (e.g. Lycopodium spores, microscopic glass beads, etc.) and then counting the phytoliths to an arbitrary sum. However, when the mineralogy of the deposits is varied and/or diagenesis is an important taphonomical processes volumes or weights can not be compared directly and therefore there is the need of a better standardization. Albert & Weiner (2001) proposed to calculate the number of phytolith per gram of the sediment's Acid Insoluble Fraction (AIF). This is the mineralogical fraction that survives -during the extraction procedure- the digestion of organic matter and the attack by Hydrochloric and Nitric acids.