Saito and Shiraishi (2002) have reported that people bind live postures together into a series of movements (the binding). They used a cued reconstruction test in which participants were presented a list of seven asymmetrical front postures sequentially presented, after which the participants were asked to reconstruct the order of the presented postures from cued postures printed on a sheet. We conducted two experiments using the cued reconstruction test to investigate the effect of the lifelike quality of postures and the processing based on an egocentric frame of reference on the activation of the binding. We used symmetrical front and back postures with less lifelike quality depicted with three dimensional computer graphics. We manipulated the number of postures (eight postures in Experiment 1, nine and eleven postures in Experiment 2) in the lists of stimuli and the number of joints (1- and 4-step conditions) functioning in the changes in the neighboring postures to vary the difficulty of task both in Experiment 1 and 2. The still pictures of front and back postures were used in Experiment 1, and the still pictures and the animation of back postures were used in Experiment 2. The results of Experiment 1 indicated that the accuracy in the 1 step condition with back postures was higher than the 4 step condition, but not in the front condition. The results of Experiment 2 showed that the accuracy in the still condition with back postures decreased as the number of postures increased from nine to eleven, but not in the animation condition. These results suggest that the decrease in lifelike quality interrupts the activation of the binding with front postures when only observing them without imitation, whereas the observation of back postures elicits the binding due to the ease in using the processing based on an egocentric frame of reference. Furthermore, the observation of sequentially presented postures is involved in forming a motor representation of neighboring postures, whereas the resulting motor representation is different from such motion representation as that shown in the animation.