Deserts are the most abundant terrestrial biome on Earth, and microbial processes assume a major role since environmental stress severely limits higher plant and animal life. A major hurdle to developing an understanding of microbial ecology in deserts has been the lack of knowledge at the fundamental biodiversity level. This is due to lack of research focus and also the inherent bias in ‘traditional’ microbial diversity estimates based upon cultivation. In this thesis an evaluation of culture-independent approaches employing both DNA and RNA from environmental samples was made with comparison to more traditional cultivation techniques. These were then applied to soils from hot and cold deserts, and along stress gradients from semi-arid to hyper-arid. A literature review was first conducted to assess the extent of current knowledge and identify critical knowledge gaps. The scientific study was then carried out as follows. The second chapter presents an evaluation of fungal taxa using cultivation, DNA and RNA based techniques. The findings indicated major taxa are revealed in all approaches, yet differences in less abundant taxa occur. The third chapter describes fungal assemblages in the soils of the McMurdo Dry Valleys, a cold polar desert. In this study, RNA based approaches tracked active fungal assemblages, whilst DNA and cultivation revealed additional taxa. Chapter four analyzed microbial communities in the Thar Desert, a hot monsoon desert in India. This study revealed a diverse community that comprised known desiccation-tolerant taxa but also a phylogenetically broad range of bacteria, archaea and eukarya. Chapter 5 focuses on the delineation of total versus active microorganisms in environmental samples from the hot deserts. As with the initial experiments, this revealed that total and active assemblages track each other broadly in desert soils. A synthesis of the study revealed that certain common microbial phyla are likely well-adapted to xeric stress, although distinct hot and cold desert assemblages also develop. For such low-diversity systems it is likely that DNA-based approaches are reasonable tools for diversity analysis, and will be especially useful in arid systems when long periods of inactivity may confound attempts to estimate active populations. Broader significance of the study includes an increased appreciation of eukaryotic microbial presence in arid soils, and how latent soil microbiota may act as a reservoir for development of future microbial macro-structures (e.g. soil crust) that function in soil stabilization, and should therefore be included in conservation planning for deserts. / published_or_final_version / Biological Sciences / Doctoral / Doctor of Philosophy