To determine whether a new, simple, quick measure, the Repeat and Point test, reliably differentiates between semantic dementia (SD) and progressive non-fluent aphasia (PNFA). Fifteen patients with SD, six patients with PNFA and 18 healthy controls were administered the Repeat and Point test. Participants were required to repeat 10 multi-syllabic concrete nouns and, following each repetition, to point to the word's pictorial referent amongst an array of six semantically and perceptually similar foils. Patients with SD were consistently impaired relative to PNFA patients and controls on the comprehension (pointing) component of the task, whereas patients with PNFA showed no significant deficit on pointing but were impaired at the production (repeating) component. Discriminant function analysis confirmed perfect classification of the individual patients into their respective groups: criteria involving a ratio of the two scores are provided. The Repeat and Point test is particularly appropriate for routine use in a clinical context: it is quick and easy to administer and score; it reliably discriminated between the two patient groups, SD and PNFA; and it offers a simple rule of thumb, i.e., the Repeat-to-Point ratio, to aid in the diagnosis of these two language variants of frontotemporal dementia (FTD). © 2007 Elsevier Srl. All rights reserved.