Study on the correlation of volatile between brown rice plant hoppers and rice stem based on electronic nose

Background: Brown rice plant hopper (BRPH) is one of the main insect pests for rice plant. To strengthen the acquisition of BRPH insect pest information has important significance for prevent and cure insect pests of rice plant. As one of the new tools of real-time acquisition of BRPH insect pest information, electronic nose has been preliminary studied in the detection of BRPH. But all of those studies are in the test phase and many questions still unexplored.

Methods: This paper used electronic nose (PEN3) sample the volatile of U31N (under the 3th-instar nymphs) and 031N (over the 3th-instar nymphs) and rice stem. Decrease the disturbance of electronic nose response values which caused by environment factors via remove benchmark value. Using principal component analysis (PCA), comparison of sensor response values, Euclidean distance (ED), fuzzy C- means clustering analysis (FCM) and loading analysis (Loadings) for analysis.

Results: The results of ED analysis prove the correctness of PCA and the comparison analysis of sensor response values. The classification accuracy of FCM for U31N, 031N and rice stem is 70%. The misclassifications are existed between rice stem and 031N, but not existed between rice stem and U31N.

Discussions: The classification effect of PCA is poor. Compare with 031N has two-part change trends of sensor value closed to rice stem, U31N only have one. We can preliminary infer that the correlations between 031N and rice stem are stronger than the correlation between U31N and rice stem.

Conclusion: According to Loadings results, we can speculate that in the young stage of BRPH's growth, BRPHs are less influenced by rice stem. But the influence will be enlarged with the increase of the age. This experiment indicates the correlations are exists between the volatiles of BRPH and rice stem, which supports scientific reference for further BRPH detection via electronic nose.