Six pigeons were trained to discriminate between two intensities of white light in a symbolic matching‐to‐sample procedure. These stimuli were then used to signal which schedule was available on the main key in a switching‐key concurrent schedule. The concurrent schedules led to a symbolic matching‐to‐sample phase in which the subject identified the concurrent schedule to which it last responded before a reinforcer could be obtained. The concurrent schedules were varied across conditions. Discriminability, measured during the symbolic matching‐to‐sample performance, was high throughout and did not differ across the two procedures. Performance in the concurrent schedules was like that typically obtained using these schedules. Delays were then arranged between completion of the concurrent schedules and presentations of the symbolic matching‐to‐sample phase. A series of conditions with an intervening delay of 10 s showed that both concurrent‐schedule performance and symbolic matching‐to‐sample performance were affected by the delay in a similar way; that is, choice responding was