Two experiments are reported. In the first experiment subjects recalled a trigram stimulus after copying a string of random consonants, using either the preferred or non-preferred hand. Retention was higher in the non-preferred hand condition in which fewer interpolated letters were written per unit time. The second experiment measured the latency of solving a simple mental arithmetic task under each of the two writing conditions. No difference was found, which, it is suggested, supports the view that the two tasks prevented rehearsal equally well in the memory experiment.