A call for the review of public biodiversity databases.

Big data and specimen digitization have profoundly revolutionized the use of natural history collections (NHCs) in ways that were unfathomable in previous decades (Hendrick et al. 2020). The NHC digitization revolution (Marshall et al. 2018) has led to a paradigm shift in the use of NHC data across scientific fields and has demonstrated the relevance of NHCs to the broader scientific community for addressing problems such as climate change, biodiversity loss, epidemiology, national security, and even diversity, equity, and inclusion (Eversole et al. 2019). Because specimen data are now used in largescale bioinformatic analyses, the accuracy of associated results, and the subsequent propagation of scientific knowledge, is determined by the quality of specimen input data (Araujo et al. 2019). Currently specimen data that are publicly available in databases such as Arctos, iDigBio, GBIF, and VertNet are fraught with error (e.g., mis-identifications, mis-tabulations, transcription and annotation error, mis-interpreted localities, nomenclatural confusion), and the inclusion of citizen science data in some databases has compounded this issue (Minelli 2017, 2019, Zizka et al. 2020, Moudry et al. 2020, Durso et al. 2021). Errors can be reported; however, this approach leaves this responsibility to users, rather than trained curators, who may not be attentive to data errors in what should be highly accurate and reliable datasets. This is inefficient, unsustainable, and even dangerous in the hands of unresponsible or uninformed users. The consequences for science are apparent and must be addressed. It is time for a prompt and thorough review of databased specimens to increase data quality and accuracy, secure a future of integrity, and improve the use of NHCs by the scientific community. Meanwhile, mandatory reporting of the use of non-reviewed specimens in studies published in the primary scientific literature will improve transparency and demonstrate where caution is warranted in the interpretation or implementation of results. Funding agencies also must address these short-falls by providing designated financial resources to NHC staff to review, correct, and enhance data quality.