Express saccades: is bimodality a result of the order of stimulus presentation?

Subjects undertook a saccadic gap task, in which the fixation target is extinguished for a period before the appearance of the peripheral stimulus. The majority showed a population of short-latency express saccades in addition to the main, slower, distribution. However, closer analysis showed that nearly all of this bimodality was due to the order in which trials were performed: the faster responses came almost entirely from trials in which the target was on the opposite side from the preceding trial, slower ones when it was on the same side. Further experiments using a novel two-gap task demonstrated that this inter-trial effect is due to the return eye movement of one trial conditioning the first saccade of the next. Consequently, in a two-gap task the latency of the second saccade falls into the faster category if it is in the same direction as the immediately preceding one: this may be the result of the oculomotor system predicting target direction, saccades in the expected direction having a shorter latency. It seems therefore that the bimodality is not primarily the result of some kind of randomising process within the oculomotor system: rather, it is a consequence of the way in which saccadic experiments are normally conducted.