Using the Holzman-Johnston Thought Disorder Index, thought disorder was examined in the first-degree relatives of schizophrenic, manic, and schizoaffective patients. In all three groups, there was a tendency for probands with higher thought disorder to have first-degree relatives with higher thought disorder. Furthermore, the quality of thought disorder in the groups of relatives was similar to that in the groups of probands, although it was clear that the relatives of schizoaffective-manic patients showed the highest amount of thought disorder, which was not found in the proband sample. Although based on a small sample, these findings suggest that amount and type of thought disorder differ not only among medicated patient groups but also among their unmedicated relatives.