An example of non-linearity in neuroendocrine systems is the generation of action potentials in oxytocin cells which result from fluctuations in membrane potential induced by synaptic activity which summate to cross the threshold for action potential generation. These properties were described by an adaptation of the leaky integrate and fire model. I used this model to predict which aspects of action potential generation are important for producing the linear changes in oxytocin cell firing rate seen in response to intravenous infusions of hypertonic saline. In addition to the changes in membrane potential seen in oxytocin cells under these conditions, only with parallel increases in the excitatory and inhibitory synaptic input could the model accurately reproduce the linear increase in firing rate seen in vivo. Magnocellular neuroendocrine systems also show non-linearities during parturition. Due to changes in the steroid milieu at this time oxytocin receptor expression in the uterus and oxytocin release from the neural lobe are increased. In order to stimulate uterine contractions in late pregnant rats oxytocin pulses were administered to 22 day pregnant rats and this advanced the onset of parturition. Fos protein immunocytochemistry was performed on brain sections from these animals and showed that oxytocin pulses increased the Fos expression in the supraoptic nucleus and nucleus tractus solitarii. Administration of progesterone inhibited the effect of oxytocin pulses on both the timing of parturition and the induction of Fos protein. Fos expression indicates whether individual cells are activated but gives no information on the dynamics of the activation. This thesis describes the use of both theoretical and experimental approaches to investigate a number of non-linear behaviours in neuroendocrine systems.