The Ouellet-Robert Entomological Collection in the digital age

Natural history collections have long been primary sources of research data for systematists, and technological advances are progressively enhancing the contributions of collections to the systematics disciplines (Wen et al. Reference Wen, Ickert-Bond, Appelhans, Dorr and Funk2015). Collections are also increasingly relevant to comparatively new and emergent research fields such as global change (Kharouba et al. Reference Kharouba, Lewthwaite, Guralnick, Kerr and Vellend2019), biodiversity (Meineke et al. Reference Meineke, Davies, Daru and Davis2019), conservation (Drew Reference Drew2011), evolution (Holmes et al. Reference Holmes, Hammond, Wogan, Walsh, LaBarbera and Wommack2016), and genomics (Buerki and Baker Reference Buerki and Baker2015). Despite their documented value (Suarez and Tsutsui Reference Suarez and Tsutsui2004; Bradley et al. Reference Bradley, Bradley, Garner and Baker2014), or perhaps rather because of it, collections face new resource-limited challenges in providing the services and data requested by the world’s biological researchers; in addition to the traditional obligations of specimen curation and collection growth, natural history collections are called upon to make their specimen data available in electronic form (Council of Canadian Academies and Expert Panel on Biodiversity Science 2010; Kemp Reference Kemp2015; Schindel and Cook Reference Schindel and Cook2018).

The Ouellet-Robert Entomological Collection was late in adopting the informatics technology and modern curatorial practices that underpin much of the modern added value of a collection (Baird Reference Baird2010; Beaman and Cellinese Reference Beaman and Cellinese2012). In 2012, the collection moved across town to new climate-controlled facilities at the Université de Montréal’s Biodiversity Centre, located on the grounds of the Montreal Botanical Garden. Having also hired a new curator (Favret), it was an opportune moment for a thorough appraisal of the collection and for creating a vision of its future development. Up until recently, knowledge of the important holdings of the Ouellet-Robert Entomological Collection was either transferred from person to person or stored within the collection archives. Revitalising the collection and planning for its medium-term to long-term growth and improvement required an accurate understanding of its current status. In order to increase the profile and relevance of the Ouellet-Robert Entomological Collection to systematics and other disciplines, over the past 10 years, collection personnel have sought to acquire, assemble, and systematise electronically data useful both to potential users of the collection and for setting curatorial priorities. Specifically, we profiled the curatorial condition of the collection, inventoried its holdings, and digitised the specimen metadata of select taxa.

Collection profile

A collection profile is a quantitative assessment of the health of a natural history collection. The collection is evaluated in terms of the long-term conservation and stability of the specimens, their labels, and their containers, and the level of preparation, identification, and digitisation of the specimens. Profiling units are identified; in the case of entomological collections, these would typically be insect drawers, vial racks, and slide boxes. Once each profiling unit in the collection has been evaluated, curators have quantitative data with which to make informed decisions regarding the parts of the collection requiring attention. One of the first formal profiling systems, developed by the Department of Entomology at the United States National Museum of Natural History, graded each collection profiling unit on a single scale of 1–10 (McGinley Reference McGinley, Rose, Williams and Gisbert1993). Similar systems were developed specifically for vertebrate (Williams et al. Reference Williams, Monk and Arroyo-Cabrales1996), paleontological (Adrain et al. Reference Adrain, Lewis and Horton2006), slide (Neuhaus et al. Reference Neuhaus, Schmid and Riedel2017), and mammal collections (Rivera-León et al. Reference Rivera-León, Hortelano-Moncada and Cervantes2018). In order to establish curatorial priorities across the multiple biological natural history collections at a single institution, the Illinois Natural History Survey (Champaign, Illinois, United States of America) developed a profiling system that parsed collection health into eight criteria (Favret et al. Reference Favret, Cummings, McGinley, Heske, Johnson and Phillips2007): conservation status (the physical condition and long-term stability of the specimens), processing state (from unprepared bulk samples to fully curated specimens), container condition (e.g., the stability and fitness of unit trays or vial stoppers), condition of labels (i.e., their legibility and permanence), identification level (unidentified or identified to order, family, genus, or species), arrangement level (e.g., mixed taxa or crowded specimens), data quality (presence of all pertinent data on labels), and computerisation level (no computerisation to fully databased specimen data). Most of these criteria are graded on a scale of 1–4: 1 – problematic, generally requiring immediate attention; 2 – substandard, to be addressed when resources allow; 3 – acceptable, stable, and not requiring intervention; 4 – ideal, the best possible scenario. See Favret et al. (Reference Favret, Cummings, McGinley, Heske, Johnson and Phillips2007) for a full description of the profiling criteria and their respective scales.

We began by assigning unique identifiers to three kinds of profiling units – Cornell drawers, vial racks, and slide trays and boxes: identifiers were simple sequential numbers from 1–2962, irrespective of kind, taxonomy, or location. Secondly, in order in the future to be able to quickly find specific profiling units requiring attention, each was identified by kind and located in space by cabinet (1–200) and shelf number (1–24). For example, profiling unit 2285 is the 15th vial rack located on the 11th shelf in cabinet 193. Third, the profile was conducted using the aforementioned eight criteria, data being entered into the FileMaker Pro (Santa Clara, California, United States of America) database described by Favret et al. (Reference Favret, Cummings, McGinley, Heske, Johnson and Phillips2007). As we were examining the Cornell drawers, we took the opportunity to take an overhead photograph of each and upload it to the database. These photographs give us a quick view of approximately how many specimens and unit trays are in the drawer, how tightly packed the specimens are, and whether there are any anomalies (e.g., envelopes or microscope slides in what is otherwise a drawer of pinned specimens, empty or non-standard unit trays).

The collection profile provides quantitative data to help inform curatorial priorities. Unsurprisingly, given the attention accorded to them by relatively recent personnel, the aquatic orders Ephemeroptera, Plecoptera, and Trichoptera are in excellent condition. Along most of the eight assessment criteria, the profiling units scored acceptably. Of note regarding the container condition are substandard hard-bottom unit trays in much of the pinned collection. The single most important criterion is conservation status, judging the long-term safety of the specimens. Of serious concern is the condition of certain parts of the pinned collection, especially the Hemiptera (Fig. 1). The poor scores are largely due to specimens having become unglued from their points (Deans Reference Deans2018). Meanwhile, most of the Odonata are contained in paper envelopes and thus not easily examined, and many of the alcohol-preserved Diptera are low on preservative; these two parts of the collection, therefore, received many substandard scores. Thanks to the profiling, we located the exact drawers, scattered in the collection, that contain unprocessed material (e.g., specimens in pill boxes), unit trays with mixed taxa, and unsorted or too-densely-arranged specimens. We upgraded the computerisation level of the entire collection by performing the inventory and the specimen label data capture (see below). The collection profile thus has provided us with specific action items for the contents of specific drawers and vial racks that we can prioritise as collection improvement progresses.

Fig. 1.

Relative conservation status of parts of the pinned and enveloped (dry) and alcohol-preserved (wet) collection with more than 100 profiling units; in black, grey, and white, the proportion of profiling units scoring 1 (problematic), 2 (substandard), and 3 (acceptable), respectively.

Collection inventory

Rather than focussing immediately on specimen-level data capture, a critical first step in collection computerisation would be an inventory of the holdings of the collection, that is, a simple list of the taxa in the collection, perhaps including the number of specimens and some indication of their geographic provenance. Such a list would inform potential users whether the collection holds specimens of interest and whether to request additional information, submit a loan request, or schedule a visit. An inventory also quantifies the taxonomic and geographic strengths of a collection and therefore helps curatorial staff decide which taxa should be targeted for specimen-level data capture.

We did not have the resources to perform particularly labour-intensive tasks such as updating the nomenclature of all the taxa or recording the full range of geographic provenance of the specimens. We therefore limited our taxonomic data to the names as found with the specimens and the geographic provenance of taxa to three hierarchical categories: “Québec” if at least one specimen of a taxon was collected in the province, “North America” if at least one specimen was collected on the continent (but not in Québec), and “exotic” for all others. For each taxon within each profiling unit, we counted the number of collection objects, that is, the number of insect pins, envelopes, pill boxes, alcohol vials, or microscope slides. We did not record the number of specimens associated with each collection object, although the vast majority of pins held a single specimen. Inventory data were added to a relational table in the same database used for profiling. These data are available for search and download as a spreadsheet-compatible Darwin Core archive (Wieczorek et al. Reference Wieczorek, Bloom, Guralnick, Blum, Döring and Giovanni2012) at Canadensys.net: https://doi.org/10.5886/vehj9v.

The Ouellet-Robert Entomological Collection houses over 20 000 insect species, of which half include specimens from Québec (Table 1). This latter figure represents 62% of the 16 600 described species recorded from or probably residing in Québec (Loiselle and Francoeur Reference Loiselle and Francoeur1992). There are 542 752 dry specimens in 1735 Cornell drawers identified at least to order on pins or in envelopes and pill boxes; 72% are identified to species level. A rough subsampling of the vials provided a conservative estimate of 25 specimens per vial, although this number varies greatly by taxon and processing state (e.g., a thousand bulk-sampled larvae versus a single holotype). The 38 281 vials in 1198 vial racks in the collection give us a rough estimate of 957 025 alcohol-preserved specimens. Therefore, omitting backlogged material not identified at least to the ordinal level, the collection houses approximately 1.5 million specimens, of which roughly one-third consists of pinned material. The microscope slide collection is negligible, although the aphid collection is growing quickly per Favret’s taxonomic expertise.

Table 1.

Summary of inventory: number of identified species and amount of preserved material of select orders.

Specimen digitisation

Because collections are arranged taxonomically, they are predisposed to be used by their traditional clientele, taxonomists. However, rather than seeking the geographic or temporal provenance of specific taxa, non-traditional users in conservation biology, ecology, and other fields might prefer to see the list of taxa from a specific locale, time period, or host plant, for example. These specimen metadata, data normally found on specimen labels, can only be made available and analysed by computerising them (Shaffer et al. Reference Shaffer, Fisher and Davidson1998; Baird Reference Baird2010). Although the value of specimen metadata is well known (Favret and Dewalt Reference Favret and Dewalt2002; Scoble Reference Scoble2010; Smith and Blagoderov Reference Smith and Blagoderov2012; Meineke et al. Reference Meineke, Davies, Daru and Davis2019), due to their sheer size in terms of number of specimens and taxa, entomological collections lag far behind vertebrate collections and herbaria in their level of computerisation (Constable et al. Reference Constable, Guralnick, Wieczorek, Spencer and Peterson2010; Sikes et al. Reference Sikes, Copas, Hirsch, Longino and Schigel2016; Sweeney et al. Reference Sweeney, Starly, Morris, Xu, Jones and Radhakrishnan2018).

As part of the initial funding for Canadensys.net, a network of Canadian collections and collections-based researchers serving organismal occurrence data (namely museum specimen data) on the Internet, selected taxonomic strengths of the Ouellet-Robert Entomological Collection were targeted for specimen-level metadata capture. One exceptional strength made evident by our inventory is the Odonata collection, including 33 123 museum objects (envelopes, pins, and vials) representing some 60 000 specimens. The label data associated with the Odonata museum objects were electronically recorded in another FileMaker Pro relational database, each object being given a unique identifier label. The taxonomic nomenclature was updated using recent checklists (Garrison and von Ellenrieder Reference Garrison and von Ellenrieder2016; Paulson and Dunkle Reference Paulson and Dunkle2018; Savard Reference Savard2018). North American collection geoposition coordinates were found retrospectively using the Canada Geographical Names Database at Natural Resources Canada and the Geographic Names Information System at the United States Geological Survey. The other aquatic taxa targeted for specimen-level databasing were the Ephemeroptera (7062 records) and Trichoptera (8844 records). All Ouellet-Robert Entomological Collection specimen-level data are available at Canadensys.net: https://doi.org/10.5886/qwvt63fz.

Perspectives

The results of the data capture presented here provide us the information needed to map the way forward. Along with assuring healthy collection growth, the priority clearly is to stabilise and upgrade the problematic parts of the collection, for example, by replacing hard-bottom unit trays and securing specimens fallen from point mounts. As the value of a collection is entirely based on its active use, it is hoped that the inventory will inform interested parties of the collection holdings and stimulate inquiries, loan requests, and visits. The profile and online inventory will be updated as various sections of the collection are re-curated and as new material is added. Likewise, the Odonata data should prove useful to a number of researchers, and we hope their use will help us secure funding to extend specimen digitising to other parts of the collection. The Odonata specimen label data of six other insect collections in Québec have recently been digitised, complementing the Ouellet-Robert Entomological Collection dataset.

In addition to providing collection data to external parties, we are also stimulating local interest in the collection. Natural history collections are unique educational resources (Cook et al. Reference Cook, Edwards, Lacey, Guralnick, Soltis and Soltis2014) and venues for public engagement (Ballard et al. Reference Ballard, Robinson, Young, Pauly, Higgins, Johnson and Tweddle2017). Students from the entomology and insect systematics courses taught at the Université de Montréal contribute to the collection as part of their course activities; a student-led natural history collection club (Gerdes et al. Reference Gerdes, Harris, Beas-Moix and Marsico2017), “Club QMOR”, participates in collection curation on a biweekly basis; a Facebook page and blog (where entomology students publish species pages: http://qmor.umontreal.ca) maintain a social media presence for the collection. An upcoming guidebook to the insects of Québec includes 2300 species, almost all of which are presented with high-quality photographs of Ouellet-Robert Entomological Collection specimens. Having gone through three phases of development, collecting the initial bulk of specimens (1930–1964), the integration of the Ouellet, Robert, and other collections and the official inauguration of the Ouellet-Robert Entomological Collection (1965–1984), and the physical growth, curatorial improvement, and move to new facilities (1985–2012), the Ouellet-Robert Entomological Collection has now entered into a new phase of electronic data capture and dissemination and increased activity.