Royal Botanical Gardens Herbarium (HAM)

Study Extent

Label digitization and image capture of vascular plant specimens from the Royal Botanical Gardens Herbarium (HAM).

Sampling

The label digitization protocol follows two periods. First, two thirds of the specimens in this resource (21,961 specimens) were transcribed in full into a Microsoft Access database (2002 – 2009). In 2011, this data was batch imported into the current database, BG-Base. Since then, the full label text has been manually entered into BG-Base by botany interns (2011 – Present). Specimen images were captured and saved as .jpg files using a book scanner (2010 – 2012) or a HerbScan flatbed scanner (2015 – Present). Moving forward, the current focus for digitization is newly collected specimens. New digitization protocols to improve the rate of label transcription and to georeference the legacy specimens are to be determined.

Quality Control

In the past decade, label transcription has been carried out by botany interns trained to use BG-Base. However, on review of the dataset, there were still issues of unfilled fields, misuse of free text fields, etc. Therefore, the data was cleaned prior to transformation to fit the Darwin Core terms. Some fields such as coordinates, countries or abundances were populated by scraping free text fields for keywords. Other measures of control include setting range constraints for valid dates or coordinates. Future transcriptions will follow stricter controlled vocabularies and can be expected to contain fewer errors.

Method steps

1. Newly collected specimens are digitized systematically.
2. Specimens are identified by trained field botanists.
3. Specimen field notes are then transcribed in BG-Base and a label is created.
4. The specimen is mounted with the label affixed and frozen at -30°C for seven days.
5. Specimens are then scanned and filed into the collection.
6. Legacy specimens are digitized opportunistically, but with an emphasis on locally collected specimens.
7. Determinations are made or updated if the user is qualified and time permits.
8. Specimen and annotation labels are transcribed in BG-Base using the HAM number present on the specimen.
9. Annotation labels are printed and affixed if applicable.
10. Specimens are then frozen at -30°C for seven days.
11. Specimens are re-filed into the collection.

By class, the digitized collection at HAM contains 89% Magnoliopsida, 8% Polypodiopsida, 2% Pinopsida, and 2% Lycopodiopsida The families most represented are Cyperaceae (9%), Asteraceae (8%), Rosaceae (7%), Poaceae (7%) and Oleaceae (4%).

1. Cyperarceae
   rank: family
2. Asteraceae
   rank: family
3. Rosaceae
   rank: family
4. Poaceae
   rank: family
5. Oleaceae
   rank: family

51% of the digitized HAM collection is representative of Royal Botanical Gardens' Nature Sanctuaries and to a small degree its horticultural collections. An additional 42% of the specimens were collected in Ontario, Canada. Two percent of the collection displays representation from other continents with no specimens from Antarctica. Royal Botanical Gardens coverage is overrepresented in the currently digitized collection based on the priority to digitize locally collected specimens first. In the physical collection, it is still the dominant representation, but likely less than 51%.