This study sought to demonstrate the relative effects of different techniques for treatment of the severe anomia of a 67 year-old woman. Both a semantic and a word form problem underlaid the anomia. A single subject approach was employed with an alternating treatment design to contrast four therapy conditions and two control conditions. Condition 1 involved semantic matching tasks and word form training (mentally visualising the written word, naming with the help of the first letter), condition 2 involved semantically unrelated matching tasks and word form training, condition 3 involved semantic matching tasks, condition 4 involved semantically unrelated matching tasks. In the two control conditions, the patient was only provided with the opportunity to try to name the pictures. With the first therapy condition, the criterion for the termination of therapy was reached. Moreover, there was overall significant difference between the different conditions. Some maintenance of the effects of therapy were observed two days later. No generalisation to untreated items was observed, although the patient seemed to be better at identifying the first syllables of words when naming. These results demonstrate that techniques which are designed for remediating particular anomic problems are a better means of therapy than techniques which are not.