Number counted:
The instruction to count the number of longer-than-usual task-relevant shapes resulted in a more difficult task than counting the number of task-relevant shapes. Fifteen participants in Group Number reported the correct number (i.e., 48), whereas only three participants in Group Longer gave the correct number (i.e., 12). 

False alarms:
The frequency of false alarms committed in each block, collapsed across groups, is shown in Figure 3. As can be seen, more false alarms occurred during task-relevant shapes than during task-irrelevant shapes in each trial block. 

Secondary RT:
When group RT was compared, the results did not replicate those found by Siddle et al. (1996). Thus, there was no interaction between the group, probe position, and stimulus type factors.  Figure 4 shows secondary RT to probes presented during task-relevant and task-irrelevant shapes in Group Number (top panel) and Group Longer (lower panel).  As can be seen, secondary RT was slower during task-irrelevant shapes than during task-relevant shapes at all three probe positions in both groups. This difference represents the pattern in secondary RT that is characteristic of the dissociation effect. 

Two additional results are shown in Figure 4.  First, the difference between shapes tended to be the same across the three probe positions. Second, there was a downward trend in secondary RT across probe positions (i.e., secondary RT at the 50 ms probe positions was slower than at the 150 and 250 ms probe positions and RT at the 150 ms probe position was slower than at the 250 ms probe position). Thus, the data showed a downward parallel trend in secondary RT for the task-relevant and task-irrelevant shapes. 

Groups differed only in overall secondary RT across the experiment.  Secondary RT was slower in Group Longer than in Group Number in the first trial block. In all subsequent trial blocks, however, this pattern was reversed. 
 

Correspondence to: D.Neumann@ccrn.uwa.edu.au