This research examined whether an expanding training series protracts
retention for infants as it does for children and adults. In three ses
sions spanning an 8-day period, 3-month-olds learned to move a crib mo
bile by kicking. Intersession intervals were Either constant (1 or 4 d
ays) or progressively expanding (average ISI = 4 days). The expanding-
series group exhibited significant retention on a delayed recognition
test 3 weeks after training was over, but the two constant-series grou
ps exhibited none. Although the I-day constant-series group remembered
after I week, the 4-day constant-series group did not. Surprisingly,
a reactivation treatment administered 4 weeks after training was over
was ineffective whether infants were trained, reminded, and tested in
a distinctive context or not. These results demonstrate that the reten
tion advantage afforded by programming training sessions in an expandi
ng series extends to infants and suggest that the upper limit on react
ivation is timed from initial encoding and not from the point of forge
tting. (C) 1998 John Wiley & Sons, Inc.