Surprisingly, Bartlett's (1932) famous repeated reproduction experiments, i
n which he found systematically increasing errors in recall from the same p
eople tested over time, have never been successfully replicated. Several st
udies have attempted partial replications, which were unsuccessful, and the
ir authors concluded that the original observations might not be replicable
. We conducted a study modeled closely after Bartlett's procedures: Subject
s studied "The War of the Ghosts," took an initial test 15 min later, and t
hen took a delayed test after 1 week. A follow-up test was conducted 6 mont
hs later on as many subjects as could be obtained. We did replicate Bartlet
t's results, in that (1) subjects forgot the story over delays but (2) intr
oduced rationalization and distortion into their accounts of the story, wit
h increases in the proportion of material distorted as retention interval i
ncreased. Subjects also imported new propositions at long delays, further c
onfirming Bartlett's empirical observations and conclusions. Bartlett's rep
eated reproduction results can be replicated.