An Emerging High-Risk Pathway for Biological Invasions: Online Pet Trade of Non-Native Terrestrial Arthropods in Mainland China

Published: 26 June 2026| Version 1 | DOI: 10.17632/h5499d66jv.1
Contributor:

Description

The trade in non-native pets has grown steadily in mainland China over the past decade, with terrestrial arthropods representing a dominant group among traded species. If released or escaped, these species could become invasive, posing substantial threats to biodiversity, food security, and public health. However, no comprehensive analysis of this trade has been available for mainland China. Here, we systematically characterised the taxonomic makeup, geographic origins, and market supply structure of these species. From June 2023 to June 2024, a survey was conducted across three major e-commerce platforms: Taobao, JD.com, and Xianyu, focusing on five arthropod classes: Arachnida, Chilopoda, Diplopoda, Malacostraca, and Insecta. A total of 580 non-native terrestrial arthropod pet species were recorded, representing 14 orders, 52 families, and 248 genera. The proportion of non-native species among all terrestrial arthropod pets sold online was strikingly high in Araneae, Hymenoptera, Scorpiones, Mantodea, and Coleoptera. South America and Southeast Asia were identified as the primary source regions, with Brazil and Indonesia as the top supplier countries. Domestically, sellers were heavily concentrated in the eastern coastal region, particularly Guangdong, Zhejiang, Jiangsu, Beijing, and Shandong. Commercial success was consistently associated with larger-bodied, longer-lived, brightly coloured, venomous, and carnivorous species. Our findings provide an urgent snapshot of the emerging online pet trade in mainland China and underscore the need for stricter regulation of high-risk species and public education to avert future invasions.

Files

Steps to reproduce

This dataset derives from a 12-month systematic survey (June 2023–June 2024) of non-native terrestrial arthropod pets on three major Chinese e-commerce platforms: Taobao, JD.com, and Xianyu. These platforms cover both B2C sales from commercial breeders and C2C trades among hobbyists, capturing most public online non-native pet transactions. Concealed channels (WeChat/QQ groups, short-video platforms, forums, offline shows) were excluded, as they lack standardized public listings and cannot be reliably quantified via web scraping. Standardized Chinese keyword searches targeted five ornamental pet arthropod classes: Arachnida, Chilopoda, Diplopoda, Malacostraca, and Insecta, covering scorpions, spiders, whip spiders, centipedes, millipedes, woodlice, cockroaches, mantises, katydids, walking sticks, beetles, and ants. Only live ornamental pet arthropod listings were retained; feeder insects, breeding equipment, dried specimens and non-pet products were excluded. Extracted fields included scientific name, common name, seller’s province, vendor count, and retail price. Surveys ran biweekly. Unique sellers were identified by a composite key of platform, shop ID and shipping province to avoid double-counting; only new seller-species pairs were added incrementally. The dataset is a cumulative annual snapshot of species and seller presence, not a time series, as individual listing timestamps were not retained. All data came from publicly accessible listings. A strict multi-step taxonomic verification protocol eliminated errors from homonyms, synonyms and misidentification. Preliminary identifications were cross-validated against GBIF, Insect Identification Complete Database, Animal Diversity Web, Insecta.pro, AntWeb, AntWiki and peer-reviewed monographs. Ambiguous taxa were reviewed by taxonomic experts; species not reliably identified to species level were excluded. Non-native status was defined by biogeographic origin: species with no confirmed natural distribution in mainland China were classified as non-native, based on GBIF, AntMaps, AntWeb and peer-reviewed literature. Native ranges were grouped into 10 global regions and further refined by country. Domestic vendor distribution was mapped by province from listed shipping addresses. Six functional traits (body size, lifespan, native range size, colouration, venomousness, dietary habits) were compiled from the same authoritative databases. Unique seller count was used as a proxy for commercial success, since direct sales volume data are unavailable and displayed metrics are prone to manipulation. All statistical analyses were performed in R 4.5.0. Seller counts were ln(seller count + 1) transformed and prices log₁₀-transformed to correct skewness. Methods included Spearman’s rank correlation, linear regression, Mann–Whitney U, Kruskal–Wallis with Dunn’s post-hoc tests, Likelihood Ratio Tests and GLMs, with α = 0.05 as the significance threshold.

Categories

Arthropoda, Biological Invasion, Conservation Biology, Wildlife Management

Funders

Licence