The presence of nerves in human tooth pulp has been recognized for over a hundred years, and the innervation of dentine for about 40 years. These observations have been made in permanent teeth. Very few studies have reported on the innervation of the primary pulp and dentine. The purpose of this study was to describe the innervation of the primary tooth pulp–dentine complex. Ten mature primary teeth (one incisor, six canines and three molars) were used. Immediately following extraction they were divided into three sections using a diamond disc and saline coolant. They were then immersion fixed in a solution of formaldehyde and picric acid dissolved in a phosphate buffer pH 7·4). The teeth were then demineralized for 1–3 weeks in formic acid. Following complete demineralization, 30 μm sections were cut on a freezing microtome. Neural tissue was stained using a specific antibody to calcitonin gene related peptide (CGRP). Sections were mounted on glass slides and examined using light microscopy. No individual nerve fibres were seen in the control sections, suggesting that the method used was specific for CGRP-containing nerve fibres. The primary teeth appeared to be well innervated. Myelinated and unmyelinated nerves were seen. There was a dense but variable subodontoblastic plexus of nerves (plexus of Raschkow) and nerve fibres were seen to leave this to travel towards the odontoblast layer. Most terminated here, but a few penetrated the odontoblast layer to enter predentine and the dentine tubules. The maximum penetration was 125 μm but most terminated within 30 μm of the dentinopulpal junction. The coronal region was more densely innervated than the root. Within the crown the cervical third was the most densely innervated region, followed by the pulp horn and the middle third. In conclusion, this study has demonstrated that mature primary tooth contains a pulp which is well innervated and has many nerve endings terminating in or near the odontoblast layer, with a small number penetrating into dentine.