This paper describes two patients, O. I. and B. Y., with confabulatory syndromes. O. I. was diagnosed with probable frontotemporal dementia, whereas B. Y. met the criteria for probable Alzheimer's disease. O. I. (but not B. Y.) was impaired on tests of frontal/executive functions, and performed better than B. Y. on clinical tests of memory. Both patients confabulated in episodic-autobiographical memory tasks and in personal future planning tasks. B. Y. confabulated also in a semantic memory task. The paper concludes that the pattern of confabulation and the cognitive profiles shown by these two patients are better explained by the hypothesis proposed by Dalla Barba and his co-workers than by current theories of confabulation.