...
University and Jepson Herbaria at UC Berkeley
During a site visit, Chris and Jess took some notes especially about the /wiki/spaces/collectionspace/pages/666274318November 19, 2008: Chris Hoffman and Jess Mitchell visited the University and Jepson Herbaria on the UC Berkeley campus. Dick Moe, Manager of Collections Data and Informatics, gave us a tour of the Herbaria collections and talked with us about various workflows and use cases. Captured here are some notes especially about object entry.
Object entry use cases
As we came in to the museum main entry, Dick pointed to an open folio of dried plant specimens (collected in the late 1800s according to the label on one of the folio pages), saying these were probably about to go out on loan. These lacked bar codes or they had not yet been entered into the collection management system. Only something like 20% of Herbaria collections are catalogued in the database. Of importance to us, the first time an object is entered into the collection management system might be when it is about to be loaned out! The Herbaria only needs to collect a minimum amount of information as the specimen goes out on loan.
In other cases, initial data entry occurs when data entry resources have been acquired (e.g., via a grant to enter a specific collection into the database in order to make it available to other researchers and the public). In that case, the existing folio pages are gathered together, data entry from the physical labels is performed, and bar codes are then assigned and attached to the folio pages. So at this point, the folio pages will have annotations such as re-identifications. I don't know how much of that enriched data is entered into the collection management system.
Entering the collections area itself, Dick took us through the rows of cabinets, pointing out some of the different kinds of specimens they house (cones, fruits, seeds, as well as dried plants). One of the activities that is quite common in the Herbaria is taking samples (e.g., leaf cuttings) to send to other researchers and collections.
Here is a fairly typical workflow for object entry in the Herbaria:
- A researcher brings in a set of specimens that have been partially prepared: They are dried, stored between sheets, and will have some varying amount of information on labels (where located, species included, date, collector).
- Museum staff ties these into larger bunches that are sandwiched between heavy cardboard sheets into bunches that are up to about 5 inches thick. One collector's submission might end up as one or many of these bunches. Each bunch is assigned a lot number, and information is written into a physical register (the lot log). Dick said they are thinking of creating a digital lot log so there is some initial computerized information.
- These lots are placed in storage cabinets (is the cabinet information then entered into the lot log?)
- Due to a scarcity of data entry resources, these lots could sit in the storage cabinet for decades.
- If there is justified reason to do further work on a lot (e.g., researcher interested in the species collected, is a specimen from a specific area), then it will be processed using a workflow that we didn't really discuss. I suspect that steps include further physical processing to produce proper folio pages, creation of a label that has the appropriate information, data entry into the collection management system, assignment of a barcode, and filing of the physical folio into the correct location. I imagine there is a lot of variability in timing and amount and quality of data, driven by research needs as much as anything.
Some general observations: This workflow reminded me of some of the archives. A group of materials comes in - perhaps a couple hundred specimens perhaps of different species from one collecting event or one donation event. That group is divided into an appropriate number of lots, basically to make them physically manageable. The core information (collector, date, location and such) is probably shared across the group. If there is funding or need, the group (a set of lots at this point) will be further divided, processed, identified and annotated. So the hierarchical splitting of an original grouping of objects seems to hold. Dick Moe and others talked about how nice it would be if the original information from the initial group would cascade into the data entry screens as they processed the objects into more granular groupings.