TRACKING THE DEEP-TIME EVOLUTION OF GRASSES AND GRASSLANDS: BUILDING A KEY TO GRASS SHORT CELL PHYTOLITH MORPHOLOGY

ERRA GEORGINA; STRÖMBERG C. A. E

Salvador

Congreso; XI Congreso Latinoamericano de Botánica /LXV Congresso Nacional de Botânica; 2014

Sociedade Botânica do Brasil y la Universidade Estadual de Feira de Santana

Historically our knowledge about the Cenozoic spread of grassland vegetation has relied almost exclusively on indirect evidence (faunal remains, fossil soils), because of the sparse record of grass fossils in the form of pollen, leaves, and seeds. However, in the last decade, plant silica (phytoliths) has emerged as a new source of data for tracking the evolutionary history of grasses (Poaceae). It has long been known that phytoliths, in particular so-called grass silica short cells (GSSC) are taxonomically sensitive within Poaceae, but it remains unclear how precisely they can diagnose specific lineages (e.g., C4 panicoids vs. C3 panicoids), hence different ecologies and physiologies, in deep time. To solve this problem, we are conducting the first detailed study of 3-D morphology of GSSC phytoliths in a phylogenetic framework. Our initial study samples the grass phylogeny broadly to get a basic understanding of the patterns of GSSC morphology within the grass clade. As a first step, we describe the GSSC phytoliths from each modern grass sample and assign them to six basic morphotypes (rondel, pyramidal, saddle, crenate/polylobate, bilobate, cross). This initial coarse classification has so far shown, for example, that crenate/polylobates can dominate assemblages in both pooids and bambusoids and are common in many species of phylogenetically distant panicoids. Saddles dominate in many chloridoids and bambusoids, and bilobates are dominant or common in many panicoids, ehrhartoids, and pooids. These results are consistent with previous work and highlight the problem of redundancy in phytolith shape among grasses when only the coarsest, 2-D morphotypes classes are used to characterize GSSC assemblages. We tackle this issue by using geometric morphometrics to analyze the morphology within each morphotype class, focusing on morphotypes that make up >5% of the GSSCs. So far we have focused on bilobates morphotypes, for which we use twenty landmarks and semilandmarks to describe outline shape and several more to describe ornamentation. We are mapping the resulting shape data onto the grass phylogeny to identify instances of morphological change. Our goal is that the end result will be a detailed, robust and phylogenetically based key to grass phytolith shape that can be used to diagnose grass lineages in the fossil record (e.g., grass phytoliths from Las Violetas).