The variability in the external gyri pattern of human brains has made the identification of specific cortical areas difficult. Studies correlating cortical structure and function have not consistently controlled for this variability. The aim of the present study was to develop a reliable and reproducible regimen for sampling five speech-associated and one non-speech associated cortical region in the human brain. The gyri of interest were labelled using non-aqueous dye prior to coronal slicing of brains at 3 mm intervals. Using the labelled gyri, a set of internal brain landmarks was established to aid in sampling one block of each cortical region of interest. The position of each internal landmark was determined as a percentage of the total brain length and breadth. The variability in the position of each internal landmark was investigated using analysis of variance and found to be consistent in three dimensions in all cases. The correlation of the sampled cortical region to the internal landmark was consistent in different cases with point to point agreement of 100%. This contrasts with the variability between cases in external gyri features. The sampled region was tested to determine cytoarchitectural variability by measuring the depth of each cortical layer. This technique found that the same cytoarchitectural regions were sampled in each case. As expected, these regions were distinguishable by the significant difference in the depth of different cortical layers. Accurate identification of both the external gyri and internal landmarks occurred with interrater point to point agreement of 90-100%.