Abstract: To distinguish organic vegetables from non-organic ones, we analyzed the stable carbon and nitrogen isotope compositions (δ¹³C and δ¹⁵N) of 19 non-organic and 34 organic samples of 17 categories of vegetables collected from various sources (physical retailers, e-commerce platforms, takeaway platforms, farming base). The results showed that there was no significant difference of δ¹³C between organic and non-organic samples, while δ¹⁵N was significantly higher in the organic samples. The effectiveness of δ¹⁵N varied in different vegetable categories. The fussy C-means clustering based on dual isotope compositions (δ¹³C and δ¹⁵N) showed that 44.45% of non-organic samples and 15.14% of organic samples were consistent with the market labeling. 21.22% of market-labeled organic products were identified as clustered non-organic. No market-labeled non-organic samples were classified as clustered organic. The rest of the samples were classified as transitional products. The δ¹⁵N of both organic and non-organic samples in this study was higher than that in other studies, possibly due to different fertilizing habits, which suggests that regional difference should be fully considered when using δ¹⁵N methods.

Key words: organic vegetable, stable isotope, fussy C-means clustering