The shape factors of phytoliths in selected plants from the Changbai Mountains and their implications

Phytoliths as a climate proxy are one of the most important and effective tools for reconstruction of paleoenvironment. In this paper, twelve Gramineae species of the subfamilies Panicoideae, Pooideae and Arundinoideae from the Changbai Mountains were selected, and 3698 phytoliths were counted. We measured six parameters of dumbell, elongate, lanceolate, crenate, saddle and silicified stomata. The seven shape factors of different types of phytoliths were deduced by the six parameters measured above and analyzed by nonparametric testing and discriminant analysis. The nonparametric testing indicated that the average value of the shape factor SF4 among the three subfamilies was obviously different. Discriminant analysis for the three subfamilies, also revealed that the discriminant accuracy was high (89.81%). Total discriminant accuracy using shape factors of dumbell phytoliths of Panicoideae was 73.12%. Discriminant accuracies using shape factors of elongate, lanceolate and crenate phytoliths of Pooideae were respectively 60.87, 59.21 and 87.07%. The discriminant accuracies using shape factors of saddle phytoliths and silicified stomata of Arundinoideae were 59.06 and 71.77%. At the same time, the shape factors of phytoliths from Setaria viridis and Leymus chinensis were used to test whether the discriminant functions established at the subfamily level were effective or not. Results showed discriminant functions precisely classify them, thus verified discriminant functions were useful. In conclusion, using the seven shape factors of phytoliths not only may improve the accuracy of classifying Gramineae, but also provides basis reference data for better understanding the presence and function of Gramineae plants at archaeological sites. (C) 2015 Elsevier B.V. All rights reserved.