In the wood condition, five photos of wood were selected from the genuine wood shown in Fig. 1a. Using data obtained from the main experiment, cluster analysis was conducted using 14 genuine types of wood. We then selected five types of wood that belonged to different clusters in order to prevent distribution bias in terms of the selected woods. The selected types of wood were cedar, falcate, walnut, oak, and guibourtia. We did not conduct auditory nor tactile conditions for the subsidiary experiment, because stones were part of the showpieces of the museum and unable to be touched or tapped.