Ecology is facing a challenging time. The pressures from an expanding human population is resulting in a global decline in biodiversity. What remains is confined to fragmented patches of habitat broken by urbanisation, agriculture and industry. To understand biodiversity in our natural world, our ecological theories of fragmentation must be more complete and our models more accurate.

Recent technological advances have enabled a broad array of new methods which ecologists can employ. In this thesis, I will explore one branch of ecological modelling -- neutral theory -- to study patterns of biodiversity in fragmented landscapes. I will present my new packages for simulating spatially explicit neutral dynamics, a powerful tool for studying fragmentation. I will then demonstrate how my models can be applied in three broad areas: theoretical ecology, tropical ecology and palaeoecology.

My theoretical investigations will explore habitat fragmentation in hypothetical scenarios, revealing important links between habitat structure and biodiversity outcomes. These links will be demonstrated using an analytical approach for estimating species loss from habitat loss. Using these findings, I will provide a neutral perspective for two ongoing debates surrounding fragmentation and reserve design in ecology. In my final two studies, I will apply neutral models to two real and disparate systems. Firstly, I will test neutral theory with reference to tropical tree diversity in Panama. My simulations will highlight the current limitations of neutral theory and in our understanding of biodiversity generation. In my second scenario, I will use neutral models to test theories of early tetrapod diversity during a time of worldwide rainforest fragmentation. Beyond this single example, I will signal the broader potential neutral theory has for addressing palaeoecological questions.