IU Herbarium goes digital: BioNews: Fall 2017: Newsletters: News & Events: Department of Biology: Indiana University Bloomington

IU Herbarium goes digital

By: Eric Knox

Friday, October 20, 2017

The IU Herbarium was established in 1885 as a research museum that now holds over 150,000 dried, pressed plant specimens mounted on large sheets of acid-free paper with attached labels that record the original collection information, any subsequent taxonomic changes, and indicators of additional research done with particular specimens. Half of the specimens were collected by Charles Deam, who wrote the Flora of Indiana in 1940, and the rest were collected by IU faculty and graduate student researchers during the past 130 years.

In the past, specimens were sent on loan from one herbarium to another, or botanists would make extended visits to herbaria in order to study the specimens and record the collection information, which would then be synthesized in published floras or research articles. Technological advancements in computing and digital photography now allow us to unlock this vast storehouse of information and make it publicly available in ways that are useful to scientists, students, and the general public.

Associate Curator Paul Rothrock and Director Eric Knox of the IU Herbarium have already completed the first three years of a five-year project designed to complete the IU Herbarium Digitization Project in advance of IU’s Bicentennial Celebration.

Rapid advances in plant molecular systematics have resulted in many taxonomic changes as researchers determine precise relationships among species in the great ‘tree of life.’ During the previous centuries, botanists had developed a good classification scheme of species in genera, genera in families, families in orders, etc., as part of the taxonomic system developed by the great Swedish botanist Carl Linnaeus. This general understanding, however, was depicted more as the great ‘cactus of life’ because the fine details of relationships could not be determined.

The first steps of IU Herbarium Digitization Project were to inventory all the species and curate all the specimens to ensure that they were correctly identified with current names. Three years may seem like a lot of time to accomplish these tasks, but if someone worked nonstop for every minute of every hour, eight hours a day, five days a week, for 50 weeks during a year, that still only totals 120,000 minutes. Approximately 40 percent of the collection (60,000 specimens) needed new name slips printed and glued to the mounting sheets, and we completed this phase by June, which is an amazing accomplishment.

The next phase involves photographing the specimens next to a ruler and color gauge using a high-resolution digital camera. A barcode is attached to each mounting sheet and the camera is tethered to a computer, so that each photograph can be renamed using a barcode reader, and the barcode number is used to start a new entry in the database. The previous inventory and curation work provided a carefully checked electronic list of the species in the collection, so the species name, along with the state and county collection information, is selected from drop-down lists, thereby avoiding the need to type any information during the imaging phase. Over 125,000 specimens have already been photographed, so this phase will be completed by December.

Most of the IU Herbarium specimens have typed specimen labels, which allows us to use an optical character recognition (OCR) program to scan the digital photograph and recover any printed information in a text file (that is also named according to the barcode number). Handwritten labels must still be transcribed into the database, but by using the digital images, we can zoom in on the label, making it easier to transfer the information. The OCR files make the work go even faster because it is just a matter of checking for OCR mistakes (e.g., the letter combination of rn is easily misread as the letter m) and then using an almost magical parsing button to copy the information from the OCR file into the appropriate fields of the database. Our database management system includes a ‘machine learning’ component so that the computer gets better at parsing the label information with each new entry. Over 57,000 specimen labels have been transcribed, and we hope to finish the transcription by June 2018.

The final stage

When all the label information is entered into the database, we will turn our attention to the final stage, which involves estimating the geographic coordinates of the site where each specimen was collected so that the information can be used in map-based applications. Botanists today routinely carry a geographic positioning system (GPS) unit in the field so that they can record the GPS data on the specimen label. Previously, the only option was to provide a written description of the site (e.g., 10 miles north of Bloomington on State Road 37), and there is an online locator system (GEOLocate) that accepts written descriptions and provides estimated coordinates; however, this work needs to be done carefully with a clear sense of place and time. In Charlie Deam’s day, 10 miles north of Bloomington would have been on the two-lane road that is now called Old State Road 37, not the 4-lane divided highway that comes to mind today (and which will eventually be renamed Interstate 69).

The specimen information is now available ‘live’ on a data portal run by the Consortium of Midwest Herbaria (http://midwestherbaria.org). This data portal draws information from all herbaria that are part of a much larger institutional network and displays information for plants collected anywhere in the world. It is possible to find information on a particular species or to download a checklist of species for all of Indiana or just a particular county. An interactive key allows people to quickly identify plants using easily observed features and without needing to know the extensive technical vocabulary used in a traditional written key. The photographs of herbarium specimens and living plants provide quick confirmation of the species’ identity, but there are also written descriptions that provide all the technical details. The IU Herbarium will continue to add features and information to the data portal (including Deam’s observations from 1940), and the portal allows inventories to be made for individual state parks or nature preserves.

The collective effort of more than a century of Indiana botany will soon be available to everyone online. As technology advances, this information will become available through mobile applications so that people can identify plants in the field, and eventually there will be photo recognition programs for plant identification. As a child watching Star Trek on television, I yearned for my own tricorder (the futuristic Starfleet's multifunction handheld device used for sensor scanning, data analysis, and recording data) for identifying all the plants and animals in my backyard. Our technology has not advanced that far yet, but soon it will be easy for most people to identify most plants in Indiana.

The IU Herbarium Digitization Project is jointly funded by the Department of Biology, College of Arts and Sciences, Center for Biological Research Collections, Vice Provost for Research, Vice President for Research, University Information Technology Services, and IU Libraries.

Eric Knox is a senior scientist with the IU Department of Biology and director of the IU Herbarium. Knox is a Fellow of the Indiana Academy of Science and recipient of the George R. Cooley Award from the American Society of Plant Taxonomists.

IU Herbarium goes digital

The final stage

Department of Biology social media channels

The College of Arts & Sciences