We have used serial electron microscopy and 3-dimensional reconstructions of dendritic spines from Purkinje spiny branchlets of normal adult rats to evaluate 2 questions about the relationship of spine geometry to synaptic efficacy. First, do relationships between spine geometry and other anatomical indicators of synaptic activity suggest that spine size and shape might be associated with synaptic efficacy? Reconstructed spines were graphically edited into head and neck compartments; the area of the postsynaptic density (PSD) was measured; the volume of spine smooth endoplasmic reticulum (SER) was computed; and all of the vesicles in the axonal varicosities were counted. Spine head volume and the volume of SER contained in the head are well correlated with the area of the PSD and the number of vesicles in the presynaptic axonal varicosity. Spine neck diameter does not fluctuate with PSD area, head volume, or the vesicle number. These results suggest that the dimensions of the spine head, but not of the spine neck, are likely to reflect differences in synaptic efficacy. Second, does the geometry of cerebellar spine necks reduce the transfer of synaptic charge to the recipient dendrite from the theoretical maximum that could be transferred if the synapse were on a dendritic shaft? Comparison of volume to surface area showed that the spine heads are approximately spherical and the necks are approximately cylindrical. Application of results from a biophysical model that assumed these geometrical shapes for spines (Wilson, 1984) showed that the cerebellar spine necks are unlikely to reduce transfer of synaptic charge by more than 5-20% even if their SER were to completely block passage of current through the portion of the neck that it occupies. We suggest that the constricted spine neck diameter might serve to isolate metabolic events in the vicinity of activated synapses by reducing diffusion to neighboring synapses, without significantly influencing the transfer of synaptic charge to the postsynaptic dendrite.