This study aimed to determine the diagnostic utility of mast cell densities in distinguishing neurotised ("neural") melanocytic naevi from neurofibromas. Three groups of lesions were studied: neurofibromas, neural naevi, and naevi showing no neural change (control naevi). A Giemsa stain was used to demonstrate mast cells. The median mast cell density in the neurotised naevus group was significantly higher (p < 0.005) than that of both the neurofibroma and control naevus groups, but the distributions of the individual density counts overlapped considerably. The sensitivity and specificity of the mast cell density as a potential discriminator between neurotised naevi and neurofibromas, determined in relation to the optimal discrimination value obtained using Bayes' minimum cost decision rule, were low. It is concluded that mast cell density on its own is of little use as a classification tool but could be of value in the context of a multivariate decision rule.