In the first of three experiments, we found that 26% of our subjects r
eversed the mental image of Jastrow's (1900) duck/rabbit figure. We sp
eculated that the remaining 74% did not perform the reversal in imagin
ation because they were unable to maintain an accurate image of the fi
gure sufficiently long to allow for the reversal to occur. Therefore,
in a second experiment, subjects attempted to reverse the memory image
of the duck/rabbit, while simultaneously feeling an embossed version
of the figure. This resulted in a 66% reversal rate, suggesting that t
he haptic form allowed for a greater number of reversals by preventing
the mental image of the duck/rabbit from becoming distorted or decayi
ng over time. Because a good portion of these reversals may have been
tactual and not imaginal, a fragmented haptic form was used in a third
experiment. A 50% reversal rate was obtained, even though this form c
ould not be reversed by touch alone.