Comment on the paper in Quaternary International: “Methodological concerns for analysis of phytolith assemblages: Does count size matter?” (C.A.E. Strömberg)

Abstract

For the last three decades, fossil phytolith assemblages from soils, buried soils and lake sediments have been used, in addition to other vegetation indicators, to reconstruct paleovegetation changes. However, in the objective of strengthening further accuracy for these reconstructions, and in addition to comparison with reference plant collections, recent effort has been made to provide quantitative relationships between modern phytolith assemblages from soils, vegetation and bio-climatic parameters (e.g., Alexandre et al., 1997; Fredlund and Tieszen, 1997; Bremond et al., 2004, 2005a, b, 2008; Blinnikov, 2005; Albert et al., 2006; Lu et al., 2006; Barboni et al., 2007). Along this line, Stromberg (Quaternary International, this volume) presented a paper entitled ‘‘Methodological concerns for analysis of phytolith assemblages: Does count size matter?’’, articulated on (1) calculation of confidence intervals related to count size and (2) investigation on how addition of these confidence intervals may affect vegetation interpretations. The first point is a relevant contribution for increasing accuracy of phytolith data. We comment here on the second point. For the purpose of her demonstration, the author used published modern phytolith data from West Africa (Bremond et al., 2005a, b) for which she misinterpreted the vegetation significance. The phytolith index D/P was first presented by Alexandre et al. (1997) and later further defined by Bremond et al. (2005a) as the ratio of ligneous dicotyledon phytolith type (globular granulate) characteristic of trees in tropical areas of low elevation, over the sum