Phylogenetic and functional growth form diversification in the Cape grass genus Ehrharta Thunb.

Abstract This thesis uses phylogenetic and comparative data to test an hypothesis of adaptive radiation in the Cape grass genus Ehrharta Thunb. sensu stricto. Morphological data and sequence data from two noncoding regions of DNA (lTS1 and trnL-F) are used to produce a phylogenetic hypothesis for the tribe Ehrharteae. Combined analysis of these data sets resolves four principal clades that approximate the genera Ehrharta s. s., Micro/aena, Tetrarrhena and Zotovia and this result thus supports a four-genus classification. Poor resolution and a reduction in branch length at the base of a clade nested within Ehrharta s. s. suggests past radiation. Parsimony-based reconstruction of ancestral habitats and growth form attributes indicates that such radiation is associated with a historical transition to seasonallydrier but more fertile habitats, and the coincident or subsequent evolution of several growth form novelties (e.g. buried and swollen culm bases and annualness). These traits are interpreted to reflect divergent strategies for surviving seasonal drought (Le. via seed or storage). Much higher transpiration rates in summer-deciduous leaves than in perennating culms of two species suggest that the evolution of summer-deciduous foliage was important in the occupation of seasonally-arid habitats. Controlled growth experiments are used to test the hypothesis that divergence in persistence traits is associated with differences in seedling biomass allocation and relative growth rate (RGR). Ehrharta s. s. shows wide variation in seedling RGR and regressions based on phylogenetically independent contrasts suggest that differences are better explained by early biomass allocation than leaf area indices. Species with a high allocation to leaves grow faster and flower sooner, so these traits are typical of seeding species. Experimental data plus a comparison of species' soil-preferences suggest that high RGR's are sustainable only in comparatively fertile habitats. Thus resource-limited systems favour low RGR-species that invest more heavily in vegetative persistence (e.g. storage). In addition, experimental and field data show that species from resource-limited habitats capitalize on increased resource availability after fire. Data from a defoliation! fertilization experiment indicate that apparent fire-stimulated flowering in E. capensis is a response to ameliorated growth conditions rather than a strict fire cue. In conclusion, this study supports the hypothesis that Ehrharta s. s. has undergone adaptive radiation in summer-arid habitats at the Cape, presumably following the inception of a summer-arid climate in the late Tertiary. Parallel studies on other taxa are needed to determine whether such radiation is a general feature of the Cape flora.