A matter of sight: Evaluating taxonomic fidelity and hidden hybridization in iNaturalist records of Callithrix marmosets (Primates, Platyrrhini)

Resumen

Abstract Citizen Science platforms, such as iNaturalist, are increasingly utilized to gather large-scale data on species distribution and occurrence, complementing traditional ecological methods. However, data quality concerns, particularly regarding taxonomic accuracy and the ability to detect morphologically intermediate individuals, limit their application for species complexes experiencing hybridization. The marmoset genus Callithrix in southeastern Brazil presents such a challenge: invasive species (C. jacchus and C. penicillata) hybridize extensively with the endangered endemic C. aurita, producing individuals with intermediate phenotypes that complicate identification. This study aimed to assess the taxonomic fidelity and hybrid detection reliability of Callithrix records from the iNaturalist platform within southeastern Brazil, providing validated data for subsequent ecological analyses. We applied a multi-stage validation protocol to “research grade” observations, including manual screening of all images to confirm species identity and classify hybrid phenotypes based on literature-derived morphological criteria. The analysis of the records indicated considerable variation in the phenotypic classification of the populations. While hybridization proved to be predominant in C. jacchus, a moderate incidence was observed in C. penicillata, and a reduced frequency of hybrid phenotypes in C. aurita. Inter-observer reliability assessment between the primary screener and a specialist yielded fair agreement, indicating that phenotypic hybrid classification, while not achieving high precision, provides a consistent and non-random assessment suitable for population-level analyses. Species misclassification by the iNaturalist community was rare, confirming high taxonomic accuracy of “research grade” records for this genus. Our findings demonstrate that iNaturalist data achieve high taxonomic accuracy for Callithrix species identification, but that reliable hybrid detection requires expert phenotypic screening. These validated data provide a foundation for spatial and ecological analyses of marmoset invasion and hybridization dynamics.