Bolted joints provide one of the most common means of joining two structural components together. The joints themselves use friction to transmit force, torque and motion across a common interface from one component to another. In many cases a pretensioned bolt, running through a common hole at the joint interface, provides the clamping force. The friction force at a joint interface is highly nonlinear. This makes the analysis of dynamic systelTIS with joints unrealistic with conventional linear techniques. It has also been shown that the contact pressure at a joint interface is not necessarily uniform. A variable contact pressure results in a variable limiting friction load. Where the contact pressure can be shown to be smallest on an interface, local microslip can take place whilst the joint maintains its sticking contact elsewhere. Microslip is responsible for the dissipation of energy from within bolted joints that otherwise maintain their integrity. The level of energy dissipation caused by microslip can be significantly larger than that provided by other dissipative mechanisms within a structure. This provides an incentive to be able to describe and predict the energy losses and overall joint behaviour accurately. Difficulties arise when considering 3-Dimensional contact, changing contact conditions during dynamic loading and the nonlinear nature of friction phenomena. To investigate microslip behaviour in bolted joints a detailed finite element model of an isolated lap joint interface was constructed. The joint interface was subjected to a variety of preloads and applied torque. Output from the joint is in the form of hysteresis loops that reveal information about the energy dissipated and overall joint stiffness during a loading cycle. Representative models are presented that reduce the complexity of the joint, yet still maintain the defining characteristics of the hysteretic behaviour. The first representative model uses Jenkins elements that match the physical response of the joint at a number of discrete points during the loading cycle. Good agreement between the finite element model and the Jenkins element model is illustrated. The Jenkins element model is also capable of predicting the response of the finite element model when different magnitudes of preload and applied torque are applied. The second representative model is the Bouc-Wen representation of hysteresis. This model offers significant gains in efficiency when approximating the smooth transition from a fully sticking interface to the onset of joint failure. All of the hysteresis can be described using just four parameters, and matching with the finite element model is demonstrated. To demonstrate microslip behaviour physically an individual joint was experimentally analysed. A cantilever beam with a single lap joint near the clamped end is resonated to generate the dynamic joint hysteresis. The joint behaviour is monitored by local time domain measurements at a number of different preloads and excitation amplitudes. Microslip is demonstrated in the joint when the preload is reduced from a maximum "rigid" clamping value. Notably at low preloads the spectral content of the response reveals a large contribution from the superharmonics of the excitation frequency. Both the Jenkins element model and the Bouc-Wen model are successfully matched to the hysteresis output of the experimental joint.