Dissolution does not affect grass phytolith assemblages

Dissolution is one among several taphonomical processes that may bias paleoenvironmental, paleoclimatic or taxonomic interpretation of phytolith assemblages. To improve our understanding of dissolution on grass phytoliths, we studied systematic changes of surface features, morphotype assemblages, and dissolution rates of phytoliths extracted from two grass species Hyparrhenia involucrata (Panicoideae), Nastus borbonicus (Bambusoideae), one soil from La Réunion Island (approximate mean age < 800 yr), and three paleosols from Ethiopia (approximate age of 4.4 million years). We used heavy-liquid to extract phytoliths, and 1% Na2CO3 to perform partial dissolution experiments. Physicochemical surface properties, morphotypes, and assemblages were analyzed using optical and scanning electron microscopy, laser diffraction, and X-ray diffractometry. Our results show that 1) phytoliths from different grass species may have different dissolution rates: phytoliths from the leaves of Hyparrhenia involucrata (Panicoideae) are more prone to dissolution than those from Nastus borbonicus (Bambusoideae). 2) Silicon (Si) released by phytolith assemblages (i.e., phytolith dissolution rate) decreases as follows: plant > soil > paleosol. 3) Dissolution leads to cavity formation on phytolith surfaces and disappearance of fragile silica particles. 4) Partial dissolution does not significantly change percentages of common grass phytolith morphotypes in a given assemblage. These results provide a benchmark for assessing the reliability of paleoenvironmental reconstructions using grass phytolith assemblages from buried soils and sediments.