Species identity
Alignment resulted in a consensus of 167bp sequences recovered from 86 tortoiseshell products. The length corresponded to the positions (Pos) 360-526bp of the published E. imbricata COI DNA sequence (GenBank accession code: JX454986; Duchene et al. 2012). A total of three distinct haplotypes were scored (GenBank accession #: OP473969, OP473970 and OP473971). Subsequent BLAST of the haplotypes again the GenBank found 34 Refs that had detailed collection information and shared similarity >95%. On the basis of taxonomic identity of all the Refs, the samples all belonged to E. imbricata and a confirmed hybrid progeny of female E. imbricata ´ male Chelonia mydas (KU254594; Kelez et al. 2016), basically in agreement with the initial morphological speculation (Table 1).
Stock origin
Two of the 3 haplotypes, P1 (n=25) and P2 (n =13) joined 11 Refs (P1, n=25 plus 8 Refs; P2, n=13 plus 3 Refs) that all shared the origin in the Indo-Pacific stock (PS), including Australia, Hawaii, Singapore, Piura and Peru; the other haplotype, A1 (n= 48) was combined with 23 Refs all originated in the Atlantic stock (AS), including the Caribbean Sea of Magdalena and Bolivar, Colombia, Tortuguero, Costa Rica and Puerto Rico. Except that 20 Refs joined A1 (n = 48 plus 20 Refs), the rest three formed A2 that included no sequences derived from our tortoiseshell samples. Detailed information was exhibited on Table 1.
In the dataset of all 120 of 167bp E. imbricata COI sequences (including 34 Refs), eight polymorphic sites were scored (4.8%). Overall, the nucleotide diversity (Pi) was 1.92%±0.08% and haplotype diversity (Hd) was 0.59±0.03. The minimum number of nucleotide substitutions between stocks was 6 (Pos 4, 10, 61, 95, 100 and 139) observed between P1 and A1, both being dominant haplotypes in PS and AS, respectively (Fig. 2). P1 and P2 within PS differed in a single nucleotide site (Pos 52), and so did A1 and A2 within AS (Pos 5). Despite being two sister clades in a monophyletic group independent of other species in Cheloniidae on the ML tree (Fig 2), PS and AS had a K-2-P distance of 3.97%±1.53%, which was significantly greater than the distances within the stocks (PS, 0.27%±0.28%; AS, 0.05%±0.05%). This level of inter-stock variation was noticeably higher than that between the two major lineages in C. mydas (1% between the Atlantic and the Pacific Ocean (Camacho-Sánchez et al. 2021), and even the two sibling species in Lepidochelys (1.68%±1.65% between L. kempii and L. Olivacea (NARO‐MACIEL et al. 2010). Rüppell (1835) once, proposed two subspecies for E. imbricata, i.e. E. i. imbricata representing the population nesting in the warmer parts of Western Atlantic Ocean (Ernst & Barbour 1989) and E. i. bissa (Linnaeus 1766) in the tropical areas of the Indian and Pacific Oceans, which could correspond to AS and PS in this study, respectively.
Implications for conservation
Our results provided concrete evidence that the ongoing illegal trade of E. imbricata tortoiseshells is global, instead of regional. It involved the turtles of multi-stock origins across the oceans. First, under the criterion of Moritz (1994) by acknowledging the significant differences in the three biological attributes, i.e. nesting sites for reproduction (stock origin), morphology for taxonomy (subspecies status) and mitochondrial DNA sequence for genetics, the two E. imbricata stocks should undoubtedly be merited as separate conservation units. Second, as the illegal poaching and trade remains the focal continue to threat to the species, time-efficient and cost-effective tools, like the non-invasive sampling-based DNA analysis for accurate determination of species identity and stock origin of E. imbricata and other wildlife products are needed. The establishment of the short COI DNA sequence using the primer set of Villanueva‐Zayas et al. (2021) meets this demand for tracing stock origins of sea turtle products across the oceans. In terms of sea turtle conservation, traditional approaches at the population level based on physical tags had limitations in the inference of stock origin, especially in the species with complex life-history and broad-scale distributional range (Avise & Bowen 1994); therefore, future efforts for conservation research of sea turtles should be focused on the aspects in behavior biology, population ecology and life history.