Developing dendritic trees often acquire their mature form by selective pruning and reorientation relative to anatomical boundaries, such as cortical 'barrel' walls. Whether similar constraints are imposed on the developing dendrites of subcortical somatosensory neurons is not clear, although it is known that the cells in trigeminal nucleus principalis (PrV) of adult rats have polarized trees. In attempting to resolve this issue we adopted and subsequently optimized a strategy of intracellular injection of Lucifer Yellow into PrV neurons and cerebellar granule cells in slices of fixed brain. Postinjectionally, immunohistochemical staining produced a stabilized image of the Lucifer Yellow injected cells and created the opportunity to apply also an antiserum to J1-tenascin in order to detect whisker-related compartmental boundaries in the neonatal PrV. In 6-day-old rats, most PrV dendrites are polarized and they do not cross tenascin-stained, whisker-related, patch borders. Notable exceptions are dendrites from the minority of PrV cells that have large somata and are responsive to multiple whiskers.