The present experiments examined two forms of savings after the extinction
of a conditioned response (CR) in the rabbit nictitating membrane preparati
on. Each experiment entailed three groups. First, a reacquisition (RAQ) gro
up received initial conditioned stimulus-unconditioned stimulus (CS-US) acq
uisition, followed by CS-alone extinction, and then CS-US reacquisition wit
h the same stimulus. Second, a cross-modal acquisition (CMA) group received
initial acquisition and extinction with one stimulus but received its fina
l stage of training with a stimulus from another modality. Third, a naive a
cquisition group (designated as REST) received only the final stage of CS-U
S training. In all three groups, tone and light were used as CSs in a count
erbalanced fashion. In addition, the experiments tested whether the manipul
ation of contextual factors-namely, trial spacing and the number of trials
per session-could influence the rates of acquisition in Groups RAQ and CMA
after extinction. The experiments demonstrated that (1) reacquisition of th
e CR to the original CS was very rapid relative to Group REST, (2) cross-mo
dal acquisition was also rapid relative to Group REST, and (3) in Group CMA
, tests of the original CS showed concurrent recovery, which is a new pheno
menon and is distinct from spontaneous recovery. However, the relative rate
s of acquisition after extinction appeared constant across the contextual m
anipulations. The results are discussed with respect to the relative stabil
ity of excitatory and inhibitory learning, as well as two alternative theor
ies of savings after extinction.