Dendritic spines are fundamental units of information processing within the nervous system, responsible for independent modulation of synaptic input to neurons. Filopodia, often morphologically indistinguishable from spines, are involved in formation of synapses during neuronal development. Despite the importance of these structures for neuronal function, no detailed study of their presence on motoneurons has yet been made. Here, the presence of spines on hypoglossal motoneurons (HMs) is described at three developmental stages: at P0–2 and P9–11, spines are present at an average density of ~0.1 spines/micron, but at P19 spine density becomes negligible. In P0–2 and P9–11, spines are nonuniformly distributed, occuring in clusters, and at lower density in the most proximal and distal regions to the soma than at intermediate regions. HM spines coincide with a decrease in cell input resistance, which reduces excitability during development. Thus one may speculate that these spines are involved in the formation of new synapses required to maintain adequate excitatory drive. A major difficulty for the study of spines is their small size, which complicates measurement using optical methods. Here, I present a novel method for reconstructing spine morphology using geometric models based on a priori knowledge of spine structure. Tests of the technique using simulated data indicate that it has a resolving capability of up to 40 nm (limited by noise). The technique has been used to measure dendritic spines on HMs, showing that these structures have necks as small as 0.22 micron. For purely passive modulation of synaptic strength, spine necks need to be <~ 0.15 micron. These data suggest that if modulation of synaptic input occurs, biochemical and/or active electrical processes are needed. The methods developed in this Thesis, which have here been applied to HMs, are generally applicable to the study of spine morphology, and its effect on synaptic processing, in all classes of neurons.