Managing high-volume specimen movements in the entomological collection of Naturalis with a new type of record

What this was about

Myriam van Walsum explained a workload problem in Naturalis's pinned insect collection (about a third of 45 million specimens, only about 10% individually digitised). Records exist mainly at drawer level, with a table of identifications per storage unit. When taxa shifted between drawers, staff had to delete and manually re-enter identifications, so moves were often not recorded. Instead of disruptive independent taxon records, Naturalis introduced a 'lot' record: specimens of the same lowest determined taxon stored in one contiguous location. It is created by migrating existing storage-unit records, so moving a taxon only means updating the lot's link. A Catalogue of Life API lookup also reduced entry errors. She asked how lots should map to Darwin Core or DwC-DP and whether GBIF or DiSSCo want such records.

Why it matters. It is a practical data-model fix for a common curation bottleneck in drawer-level digitised collections. It also raises an open standards question about publishing lot and storage-unit records.

In the room

• Naturalis holds about 45 million specimens; over a third are pinned insects, only about 10% individually digitised (mostly bees).
• The zoology/geology CMS is from a Dutch vendor not designed for natural history, so Naturalis writes its own ABCD-aligned requirements.
• Taxonomically ordered storage creates many-to-many relationships between taxa and storage units.
• Manual delete/re-enter of identifications caused backlog, errors and lost locations, despite a legal obligation to know where specimens are.
• Lot definition: specimens of the same lowest determined taxon in one contiguous location; each specimen belongs to exactly one lot.
• Lots were created by data migration from existing storage-unit records; no new digitisation was needed.
• Catalogue of Life API integration in identifications reduced data entry errors.

Notable moments