Much research has been carried out in recent years on Pickering emulsions, but understanding of the underlying physics requires considerable strengthening. This thesis seeks to address several fundamental aspects by presenting the results of recent experimental work. This work has focused on a model oil-in-water emulsion system stabilised by fluorescent colloidal silica particles and using a mixture of dodecane and isopropyl myristate as the oil phase. The phase behaviour of the particle dispersions has been altered using sodium chloride and sodium iodide, whilst sodium hydroxide and hydrochloric acid have been used to adjust the pH of samples. Comparisons are also made to emulsions stabilised by commercially available fumed silica. Conventionally, it was assumed that a weakly flocculating particle dispersion is required in order to generate a stable Pickering emulsion. It is shown in this work, however, that in some circumstances a weakly flocculating dispersion leads to the least stable emulsion. It is therefore argued that a more nuanced view of Pickering stabilisation is required, taking into account the factors affecting whether particles will adsorb to the interface during emulsification. Very recently it has begun to be suspected that Pickering emulsions sometimes aggregate due to the sharing of particles between two droplets, an effect known as bridging. In this thesis it is also shown that particle bridges can form in Pickering emulsions at high shear, and that they can subsequently be broken by low shear or by modifying the particle wettability. For the first time, electron microscopy has been used to provide direct evidence of droplets sharing particles. A simple theoretical model is developed, based on collisions between partially coated droplets, which captures the trends observed experimentally. It is argued that particle bridging may have been overlooked in the literature, and that the shear history of emulsions is a crucial determinant of subsequent behaviour. The deaggregation of bridged emulsions has been studied using a novel method where two different colours of particles are used. By starting with two emulsions which are bridged, each stabilised by a different colour of particle, and then using confocal microscopy to study them as they are mixed together and deaggregate, the processes involved in deaggregation can be elucidated. These experiments have also shown, for the first time, the dynamic nature of particles in Pickering emulsions; particles transfer readily between droplets when the samples are placed on a roller bank. It is found that a period of unbridging and rebridging takes place prior to deaggregation of the emulsions, and the timescale of deaggregation can be tuned by varying the particle wettability. The two-colour method has also been applied to the study of Pickering emulsions which are repeatedly sheared. It is found that limited coalescence is not reestablished simply by re-applying the shear rate which was used in the initial emulsification. This behaviour is attributed to the presence of an elastic shell of particles at the interface, which inhibits droplet breakup, and is in contrast to that of surfactant-stabilised emulsions, where increasing the stabiliser concentration makes droplets more liable to deform and breakup. Finally, a short study has been carried out attempting to increase the scale of the experiments presented in this thesis to sample volumes of approximately one litre. This study has demonstrated the relevance of particle bridging to industrial emulsification processes. Overall, experiments with carefully controlled model Pickering emulsions, including those using two colours of particles, have revealed the fundamental workings of these arrested systems.