In the cluster analysis of the chemical composition, 216 datapoints corresponding to the contents of chemical componentsobserved in six samples of the essential oils were involved. Theresult was showed in Fig. 3. When the Euclidean distance wasselected as 23, like the previous case, the samples were clusteredinto two groups with an exactly same mode. The samples in Group Awere also clustered into a group (Group a), while the other two wasgrouped together into Group b. However in this case, the assign-ment of samples was decided by the similarity of the chemicalcomposition. Since objects in the same group are more similar toeach other than to those in other groups (Gomez-Polo et al., 2015),it was considered that the repellent activity characteristic of theessential oil samples might be related to their chemical composi-tions, comparing the results obtained in above two cases.