In the first part of this blog series, I explained a portion of the analyses I performed during my time as an intern for the Biodiversity Heritage Library (July-August 2021). These analyses revolved around metadata patterns in BHL’s collection that highlight shortcomings in terms of diversification in the Library’s catalogue. In that post, by focusing on comparative metadata and the case of BHL México, I argued that an outreach plan that included the establishment of global partnerships between BHL and institutions in the Global South was a solid strategy to diversify the Library’s collections. This same argument is sustained by the second portion of the analyses I performed during my internship and that I present here, which deal primarily with patterns of associations and representation in the subject lists of BHL’s materials and the specific case of Central America and Panama.

The goal of the second part of my internship was thus to identify semantic patterns in subject lists that highlight the diversification—or lack thereof—in materials about Latin American biodiversity contained in BHL. To this end, I began my analyses with the extraction of the subject lists of those materials. In order to perform this extraction, and following, once more, the methodology employed by Chris Freeland (Freeland, steps 1–3), I cross-referenced BHL’s data file on items and that on subjects to match each title ID to the subjects it includes.[i] Finally, from the resulting dataset, I extracted the records of materials that included subjects related to Latin America to build six subsets:[ii]

1. West Indies (WI): This subset contains all materials (228 title IDs) that include the subject/term West Indies in their subject list.
2. Central-Latin-South America (CLS): This subset contains all materials (710 title IDs) that include at least of one of these subjects/terms in their subject list: Central America, Latin America, South America. Out of these three, South America is considerably more frequent at 528 occurrences,[iii] more than three times the occurrences of the other subjects, with 168 occurrences of Central America and 104 of Latin America.
3. Central American Countries (CAC): This subset contains all materials (485 title IDs) that include at least one Central American country in their subject list: Belize, Guatemala, Honduras, Nicaragua, Costa Rica, Panama, El Salvador. Out of these countries, Panama is by far the most frequent subject (240 occurrences).
4. South American Countries (except Brazil) (SAC): This subset contains all materials (1221 title IDs) that include at least one South American country (except Brazil) in their subject list: Argentina, Bolivia, Colombia, Chile, Ecuador, Guyana, (French) Guiana, Paraguay, Peru, Surinam/Suriname, Uruguay, Venezuela. Out of these subjects, Peru, Chile, and Argentina are the most frequent, with over 230 occurrences each, followed by Colombia (193) and Ecuador (154). The remaining countries occur less than 100 times each, with Uruguay being the least frequent (25 occurrences).
5. Brazil (BR):[iv] This subset contains all materials (543 title IDs) that include the subject/term Brazil in their subject list.
6. Mexico (MEX): This subset contains all materials (917 title IDs) that include the subject/term Mexico in their subject list.

As I explained in my previous blog post, the geopolitics of publication and housing of materials in BHL’s catalogue reflect an unequal and still colonial relationship between the Global South and North. In addition to this, I argue that a lack of diversification in terms of knowledge production results in unequal representations and valorization of the Global South and North that perpetuate colonial dynamics in biodiversity-related knowledge production. To test this thesis through subject lists, I generated frequency lists, co-occurrence networks, and hierarchical clusters using KH Coder 3 for each of the previously mentioned subsets.[v] By taking a closer look at the patterns of subject lists in the subsets described above and revealed by these text analyses, it becomes clearer how a colonial niche of biodiversity-related publications can lead to biased and colonial representations of biodiversity and communities from the Global South. That is the case, for instance, of materials in BHL that focus on the biodiversity of Central America and, specifically, Panama.

As previously mentioned, in the CAC subset, Panama is the most frequent subject with 240 occurrences, followed by the compound field note with 188 and Costa Rica with 99 (Figure 1). These numbers suggest that Panama is the most studied Central American country in BHL’s materials. Furthermore, when looking at the subject lists in which it appears, it is possible to note that it is often associated with the Canal Zone and the United States, terms that are significantly frequent in this subset as well (Figure 2).

Figure 1 The 20 most frequent words in the CAC subset. Generated on KH Coder 3 in September 2021. Data from https://www.biodiversitylibrary.org/data as of July 1st, 2021.

Figure 2 Sample of subject lists that include the word Panama. Generated on KH Coder 3 in September 2021. Data from https://www.biodiversitylibrary.org/data as of July 1st, 2021.

These initial observations in terms of word-frequencies and associations in the CAC subset led to the hypothesis that the frequency of Panama in this subset is due to the biological (Smithsonian Tropical Research Institute, ‘Why Is the Smithsonian in Panama?’) but also economic importance of the country and the Central American region, especially given its strong association with US-related subjects. For instance, in the co-occurrence network and hierarchical clustering of the CAC subset, it is possible to identify a strong relationship between Panama and the United States. The most intricated subgraph in the co-occurrence network is built around Panama and includes nonhuman species and biodiversity-related topics (mammalogy, ethology, Animalbehavior, Ornithology) alongside US-related subjects and institutions (UnitedStates, NationalMuseum), as well as the compound Canal Zone (Figure 3). Given these observations and to further test this hypothesis, I repeated the extraction and selection process to create a subset specifically for this country, the PAN subset.[vi] The goal of the creation of this additional subset was to verify whether these US-centric interests and findings around Panama also manifest in the rest of the metadata of records containing the subject Panama.

Figure 3 Co-occurrence network of the CAC subset. The overlapping of subgraphs 01 and 06 shows the strong association between Panama and the United States. Generated on KH Coder 3 in September 2021. Data from https://www.biodiversitylibrary.org/data as of July 1st, 2021.

As I explained in the first part of this blog series, mapping the places of publication of subject-based subsets of materials in BHL can help visualize the relationships between the Global South and North as both objects and subjects of knowledge production. When mapping the PAN subset, however, I noticed that almost 70% of the materials (125 out of 185 title IDs) had no identified place of publication. When looking more closely at each of these materials, I found that almost all of them[vii] were handwritten field notes. Therefore, I looked at each of these titles to identify the origin of these field notes. While their place of publication cannot be identified because these materials are not publications per se, identifying the affiliation of the field work and expeditions that originated them helped more clearly illuminate the geopolitics of biodiversity-related knowledge production about Panama. For instance, out of the 125 materials with no identified place of publication, 76 (60.8%) are field notes resulting from US-based expeditions to Central America,[viii] meaning that such expeditions constitute the predominant point of origin of materials in the PAN subset[ix] (Figure 4). Looking at these details for each of the records with no identified place of publication in the PAN subset is particularly important as the hierarchical clustering for the CAC subset also shows this significantly strong connection between the subjects Panama and field note(s) (Figure 5). Given that the majority of field notes with no identified place of publication in the metadata are the product of US-based expeditions to Central America, this connection further emphasizes the determining role that US interests play in the presence and representation of Panama in BHL.

Figure 4 Number of titles per place of publication in the PAN subset (with cleaned data for missing fields). Generated on Tableau in September 2021. Data from https://www.biodiversitylibrary.org/data as of July 1st, 2021.

Figure 5 Hierarchical clusters for the CAC subset. Cluster 8 (pink) shows a strong association between Panama and field note(s). Generated on KH Coder 3 in September 2021. Data from https://www.biodiversitylibrary.org/data as of July 1st, 2021.

Of the remaining materials, seven (5.6%) are affiliated with the Smithsonian Environmental Research Center (with its headquarters in Chesapeake Bay) and the Smithsonian Migratory Bird Center (in Washington, D.C.). The latter is particularly interesting, as it could contribute, at least in part, to the importance of the subject ornithology in connecting the subject Panama with the subgraph built around the term United States in the CAC subset (Figure 3, subgraph 06). Finally, the remaining 41 materials with no identified place of publication are affiliated to the Smithsonian Tropical Research Institute (STRI). What is particularly notable about the STRI is that it is located in Ancón, Panama (Smithsonian Tropical Research Institute, ‘Why Is the Smithsonian in Panama?’). This means that even if the STRI is affiliated with the Smithsonian and, therefore, the United States, the knowledge production about Panama sponsored by the Institute could be seen as an epistemic collaboration between researchers from Panama, the US, and beyond (Smithsonian Tropical Research Institute, Biological and Cultural Diversity of the Tropics 4). Nevertheless, even when considering the place of publication of STRI affiliated documents as being located in Ancón, the knowledge production about Panama and her biodiversity as housed in BHL continues to be greatly dominated by the United States (Figure 6).

Figure 6 Map showing places of publication (density per number of records) in the PAN subset (with cleaned data for missing fields). Generated on Tableau in September 2021. Data from https://www.biodiversitylibrary.org/data as of July 1st, 2021.

Other than places of publication showing a predominance of the United States as the main producer of the knowledge about Panama available in BHL, the year of these publications[x] is particularly important to further evidence the impact of US politics and interests in the production of these materials, especially given the history of the US construction and administration of the Panama Canal throughout the 20th Century. After its beginnings as a French project in the late 19th Century, the construction of the Panama Canal was initiated by the US in 1904 and concluded in 1914 (Autoridad del Canal de Panamá). The concession of this region was a result of US participation in “the separatist movement in Colombia [after which] the separatists achieved in breaking the province of Panama from Colombia [and] the United States was awarded with what was to become the Panama Canal Zone” (‘Panama Canal Zone in World War II’). The US control of the Canal continued until 1979, when it was transferred to “the Panama Canal Commission, a joint agency of the United States and the Republic of Panama” (Padelford). It was not until 1999 that the Canal’s administration was left exclusively in the hands of Panama (Autoridad del Canal de Panamá).

Figure 7 Distribution per year of titles in the PAN subset (with cleaned data for missing fields). Generated on Tableau in September 2021. Data from https://www.biodiversitylibrary.org/data as of July 1st, 2021.

Mirroring this historical context, 161 out of the 185 records in the PAN subset (87%) were published between the years 1903 and 1996, that is, during the US administration of the Panama Canal, with the peak occurring in the late fifties (Figure 7). Furthermore, out of the 16 materials in the PAN subset published prior to the 20th Century, only six are affiliated with the US: four published in New York and two being the product of US-based expeditions. In contrast, the other 74 of the 76 materials that are the product of US-based expeditions (as previously explained) were published between 1910 and 1983, meaning that 97.4% of these expeditions occurred during US control of the Panama Canal (Figure 8). Moreover, the STRI itself, in spite of its global collaboration outlook, is “closely tied to the construction of the Panama Canal” as it resulted from “Smithsonian scientists and naturalists across the United States urg[ing] U.S. President Theodore Roosevelt to support a biological expedition to take an inventory of the future Canal Zone’s flora and fauna” (Smithsonian Tropical Research Institute, ‘Why Is the Smithsonian in Panama?’). Therefore, given these patterns of publication across time—alongside the importance of US expeditions and the STRI in the production of the knowledge contained in BHL and the semantic associations and word frequencies of subject lists in the PAN subset—it is possible to conclude that biodiversity-epistemologies about Panama in the Library are greatly subsumed to US history, politics, and economic interests.

Figure 8 Chronological distribution per place of publication of titles in the PAN subset (with cleaned data for missing fields). Generated on Tableau in September 2021. Data from https://www.biodiversitylibrary.org/data as of July 1st, 2021.

In addition to these categories of metadata showing the dominance of US-centric and Global-North-centric epistemologies and interests in BHL’s materials about Panama, it is essential to consider the language of the materials, another fundamental category in terms of the diversification not only of representation but also of access and audiences, as I have argued elsewhere.[xi] In this regard, out of the 185 records in the PAN subset, only seven (3.78%) are in languages other than English,[xii] of which only three (1.62%) are in Spanish, the official language of Panama (Zajícová 185)(Figure 10).

Figure 9 Number of titles per language in the PAN subset (with cleaned data for missing fields). Generated on Tableau in September 2021. Data from https://www.biodiversitylibrary.org/data as of July 1st, 2021.

The three materials in the PAN subset that are written in Spanish are James Zetek’s Los moluscos de la República de Panamá (1918), the Smithsonian Tropical Research Institute’s Panamá, puente biológico: Las Charlas Smithsonian del Mes (2001),[xiii] and Rafael Tobías Marquís Oropeza’s Algunas palmeras industriales de la flora istmeña (1908). The first two, albeit being published in Panama, are affiliated to the US through their authors and holding institutions. These two materials are held by the Smithsonian Libraries, with the STRI also being the author of the second one. Likewise, James Zetek (1886-1959) was a US entomologist who worked for the US Department of Agriculture and was in charge of research expeditions and international relations with Central America, especially the Canal Zone in Panama, during both world wars (Snyder et al. 1230–31). This record, thus, further evidences the importance of the history of the Canal and US interests in the materials in the PAN subset. On the contrary, the third record is a unique example that counteracts US-centric tendencies. Marquís Oropeza was a Venezuelan scientific philosopher and agronomist who acted as the first Director of the Museo Nacional de Panamá (Moreno 169–70). Therefore, and even though this work is held by the New York Botanical Garden, Marquís Oropeza’s book is the only record in the PAN subset (out of 185) that is written in Spanish and where the knowledge production took place in Panama and independently from the US and Europe. Moreover, this work highlights biodiversity-related collaboration within the Global South, further acting as an example of epistemic agency in non-hegemonic spaces.

Nevertheless, the evident shortcoming of the record by Marquís Oropeza is the rareness of its case, which, given the observations made through the analysis of the CAC and PAN subsets explained so far, turns it into a one-case exception. The goal of the diversification of BHL’s collection should then be to make these cases the rule and not the anomaly, an objective that could be achieved through a more thorough incorporation of non-hegemonic materials and the establishment of more BHL nodes across the Global South. In this sense, the findings explained throughout this blog post reveal the important need for BHL to strengthen bonds with institutions in the Global South, not only to achieve a truly global repository but to diversify representation, a particularly pressing matter in topics related to colonial and (neo)imperial issues, such as the case of Panama. The overwhelming predominance of US-centric narratives and perspectives in the knowledge production about Panama and her biodiversity in BHL’s catalogue reveals the urge to diversify, decentralize, and decolonize bio-diverse epistemologies. As I argued in my previous blog post, a strong collaboration with institutions in the Global South, as exemplified by BHL México, can lead to more diverse collections, not only in terms of language but also of semantic associations. In this regard, for example, when comparing the co-occurrence networks of the CAC and MEX subsets, it is possible to note that the MEX subset presents an arguably more diversified network of subjects (Figure 10). For instance, it shows a wider variety of both nonhuman species and groups and of geographical regions. Furthermore, while subjects and locations related to the United States are still present (subgraphs 01, 03, and 08), their connection to the subject Mexico itself is less strong than in the case of the CAC subset and the subject Panama (Figure 3). Thus, given that a large number of materials in the MEX subset are the product of BHL México, I continue to argue that a more solid, profound, meaningful, and critical strategy for the establishment of BHL partnerships throughout the Global South can lead to a truly global, open, decolonial, and bio-diverse BHL, deeply in line with the Library’s goals and mission.

Figure 10 Co-occurrence network of the MEX subset. Generated on KH Coder 3 in September 2021. Data from https://www.biodiversitylibrary.org/data as of July 1st, 2021.

[i] This process included all materials in BHL—regardless of subject—as of July 1^st, 2021. The full list of title IDs and subjects can be found in BHL’s GitHub repository.

[ii] Tables for these subsets can be found in BHL’s GitHub repository. Due to the constraints of the extension and nature of this blog post, not all subsets are discussed in detail. However, the files are publicly available, and the author can be contacted at any point with any further questions or inquiries.

[iii] Word frequency tables for each subset can be found in BHL’s GitHub repository.

[iv] Both Brazil and Mexico were treated as separate subsets given that a large portion of materials about Brazilian and Mexican biodiversity in BHL comes from the BHL SciELO and BHL México projects respectively. Thus, considering these materials as separate subsets can illuminate the impact of global collaboration in the diversification of the Library’s collection.

[v] All graphs for each subset can be found in BHL’s GitHub repository.

[vi] All files (subject lists, word frequency, co-occurrence network, hierarchical clustering) for this subset can be found in BHL’s GitHub repository.

[vii] The one exception is Dr. William Mark Whitten’s Ph.D. thesis (1985) at the University of Florida.

[viii] The affiliations of these expeditions vary but are mostly associated with academic and research institutions throughout the United States, such as the New York Botanical Garden and Harvard University, and government dependencies, such as the US Navy and the US Department of Agriculture.

[ix] Tables for the PAN subset with both raw data and data cleaned for missing fields (places of publication, year, and language) can be found in BHL’s GitHub repository. In the file with cleaned data, previously missing fields are highlighted in yellow.

[x] Missing fields of year of publication in the PAN subset (see previous note) were fixed by looking into other metadata categories (such as the start date for periodicals and the publication details) and verified with the material itself.

[xi] I discussed the importance of multilingual representation in the Library for the diversification of access in more detail in a previous blog post and in my presentation for BHL Day 2021.

[xii] Missing fields of language in the PAN subset (see note ix) were mostly for albums and field notes. These fields were fixed by considering the language of notes, comments, introductions, and/or descriptions/captions for images or photographs, which were all in English.

[xiii] This is the one material affiliated to the STRI in the PAN subset that is not written in English, meaning that almost all materials affiliated with the STRI are in English, which, again, hinders the collaborative nature of their research endeavours.

Works cited

Autoridad del Canal de Panamá. ‘Reseña Histórica del Canal de Panamá’. Canal de Panamá, https://micanaldepanama.com/historia-del-canal/resena-historica-del-canal-de-panama/. Accessed 5 Oct. 2021.

Freeland, Chris. ‘BHL Poster for AETFAT2010’. ChrisFreeland, 19 Apr. 2010, http://blog.chrisfreeland.com/2010/04/.

Moreno, Hiram A. ‘Tras las elusivas huellas de Rafael T. Marquís Oropeza. El primer Director del Museo Nacional de Panamá’. Canto Rodado, vol. 10, 2015, pp. 163–75.

Padelford, Norman J. ‘Panama Canal’. Encyclopedia Britannica, 2005, https://www.britannica.com/topic/Panama-Canal.

‘Panama Canal Zone in World War II’. WW2DB – World War II Database, https://ww2db.com/country/panama_canal_zone. Accessed 5 Oct. 2021.

Smithsonian Tropical Research Institute. Biological and Cultural Diversity of the Tropics: Extraordinary Opportunities and Challenges in the Midst of the Anthropocene. Strategic Plan 2019-2024. Smithsonian Tropical Research Institute, 2019, https://stri.si.edu/sites/default/files/stri_strategic_plan_aug_2019.pdf.

—. ‘Why Is the Smithsonian in Panama?’ Smithsonian Tropical Research Institute, Smithsonian Tropical Research Institute, 19 Dec. 2016, https://stri.si.edu/why-panama.

Snyder, Thomas, et al. ‘James Zetek. 1886-1959’. Journal of Economic Entomology, vol. 52, no. 6, 1959, pp. 1230–32, https://doi.org/10.1093/jee/52.6.1230.

Zajícová, Lenka. ‘Lenguas Indígenas En La Legislación de Los Países Hispanoamericanos’. Onomázein. Revista de Lingüística, Filología y Traducción, 2017, pp. 171–203, https://doi.org/DOI: 10.7764/onomazein.amerindias.10.

During my time as an intern for the Biodiversity Heritage Library (July-August 2021), I worked on a project, I hope, will help engender important and critical conversations around the Library’s work and responsibilities vis-á-vis the sometimes harmful and problematic origins of its materials, as well as around the possibilities for the decolonization of its collection and archival practices. By focusing on the case of Latin America and her biodiversity, the main goal of this project was to identify patterns in the metadata of BHL’s collection that can inform decolonial policies and strategies for the diversification of the Library’s catalogue.

To identify such patterns, I extracted and analyzed the metadata of materials that include a subject related to Latin America in their subject lists. These analyses shed important light on diversification issues, specifically in the case of this region. The first part of this project was to perform subject-based extractions from BHL’s collection. Following a similar methodology to that of Chris Freeland (Freeland, steps 1–3), I extracted all records from BHL’s online datasets[i] (as of July 1st, 2021[ii]) that included specific subjects[iii] to build five subsets:

1. Greater Regions (GR) subset. This subset incorporates all BHL records (4465 total[iv]) that include one of the following subjects—all of which refer to Latin America as a whole or in subcontinental terms—in their subject list: Latin America, Central America, South America, West Indies.
2. Latin American Countries (LAC) subset. This subset incorporates all BHL records (6801 total) that include the name of at least one Latin American country—except Mexico—in their subject list: Argentina, Belize, Bolivia, Brazil,[v]British Guiana, Chile, Colombia, CostaRica/Costa-Rica, Ecuador, El Salvador, French Guiana, Guatemala, Guiana/Guyana, Honduras, Nicaragua, Panama, Paraguay, Peru, Surinam/e, Uruguay, Venezuela.
3. Mexico (MEX) subset. This subset incorporates all BHL records (2995 total) that include the subjects Mexico; Mexico, North; and/or Mexico, Northern in their subject lists. Mexico was treated as a separate subset given that a large portion of materials about Mexican biodiversity in BHL comes from the BHL México project. Thus, considering these materials as a separate subset can illuminate the impact of global collaboration in the diversification of the Library’s collection.
4. Indigenous Peoples – General (IP-G) subset. This subset incorporates all BHL records (1052 total) that include one of the following subjects—all of which refer to Latin American Indigenous peoples in generalizing/homogenizing terms—in their subject list: Indians of Central America, Indians of Mexico, Indians of South America, Indians of the West Indies. Although of a different nature, the subjects Aztecs and Incas are also included in this subset, as they refer to the Indigenous empires that existed at the arrival of European colonizers in the Americas but that no longer exist and were conglomerates of several Indigenous groups of these regions.
5. Indigenous Peoples – Specific (IP-S) subset. This subset incorporates all BHL records (135 total) that include one of the following subjects—all of which refer to specific Latin American Indigenous peoples—in their subject list: Carib Indians, Choco Indians, Cuna Indians, Diaquita Indians, Goajiro Indians, Huichol Indians, Kickapoo Indians, Mapuche Indians, Mayas, Mayoruna Indians, Mojo Indians, Shipibo-Conibo Indians, Taino Indians, Tairona Indians, Tarahumara Indians, Yahgan Indians.

Once the subsets were created, from the different BHL datasets available online and using Microsoft Access, I extracted specific information for each of the records included in these subsets:[vi] title ID (used as primary key), full title, author, holding institution, year of publication (including start and end year of publication for periodicals), publication details, language, title URL, and the subject they include to be part of the subset.

The next step of this project was to prepare the data, especially for geolocation. In this regard, it is important to mention that BHL’s metadata do not have separate categories for place of publication and publisher. On the contrary, these two are included under the category Publication details, which sometimes includes the year of publication as well. Therefore, to understand the geopolitical affiliations of the records included in each subset, I cleaned the data[vii] to have a separate category for place of publication. In some cases, this required the modernization and/or translation of place names, given that BHL’s metadata often include publication information in the original language and format of the material. For example, the publication details of a volume of José de Acosta’s Historia natural y moral de las Indias (included in the LAC subset) read “Impresso en Seuilla :en casa de Iuan de Leon.,Año de 1590.” This record, then, required the isolation of the place of publication (Seuilla), its modernization from 16th-century Spanish to its modern form Sevilla, and its translation into English, Seville, for software readability. After the cleaning phase and running Microsoft Excel’s geographical data identification, I obtained coordinates (longitude and latitude) for each place of publication included in the subsets.[viii]

After preparing the datasets, the next step was to generate statistics and visualizations, mostly done in Tableau. For geographical data, I also employed Google Maps, which makes it easier to map the locations of holding institutions, as they are already recorded in Google. For Google Maps, given its restrictions in the number of nodes that can be mapped at a time, only unique IDs were included:[ix]GR, LAC, MEX, IP-G, IP-S.[x] In Tableau, however, I generated separate maps and graphs for two groups per subset, one for all title IDs and one for unique IDs only. For most subsets, the generated data visualizations[xi] include a density map for places of publication, graphs for number of titles (unique and non-unique) per place of publication, graphs for number of titles (unique and non-unique) per language of publication, graphs for number of titles (unique and non-unique) per year of publication, graphs for frequency of subject based on number of titles (unique and non-unique), graphs for frequency of holding institution based on number of titles (unique and non-unique), a graph for language versus year of publication, a graph for included subject versus place of publication, and a graph for included subject versus holding institution.[xii]

In almost all subsets,[xiii] mapping metadata categories evidences a discrepancy between the object of study, in this case, the biodiversity and peoples of Latin America, and the networks of knowledge production about said object. These colonial epistemic dynamics are particularly important for the IP-G and IP-S subsets, which not only include Indigenous peoples as objects of study but also show—arguably more than any other subsets—a clear distinction of the Global North as producer of knowledge about Indigenous peoples in Latin America (Figure 1).

Figure 1 Map showing places of publication (density per number of records) in the IP-G subset. Generated on Tableau in August 2021. Data from https://www.biodiversitylibrary.org/data as of July 1st, 2021.

Moreover, the IP-G subset is anchored in homogenizing terms such as “Indians of…” The very fact that the word Indians is being employed to refer to Indigenous peoples in Latin America perpetuates colonial homogenizing discourses that began with the first arrival of Europeans to the Americas (and not India) in the late 15th Century, thus flattening the cultural diversity of the Global South. Additionally, employing this formula alongside general geographical areas further blurs the cultural diversity of Indigenous peoples in the region. For instance, there are 70 identified Indigenous peoples in Mexico (Sistema de Información Cultural SIC México), at least four in Belize, 24 in Guatemala, seven in Honduras, three in El Salvador, nine in Nicaragua, eight in Costa Rica, eight in Panama, 30 in Argentina, 36 in Bolivia, 241 in Brazil, 83 in Colombia, nine in Chile, 12 in Ecuador, nine in Guyana, six in French Guiana, 20 in Paraguay, 43 in Peru, five in Suriname, and 37 in Venezuela (UNICEF España). However, these more than 660 Indigenous cultures are homogenized in the generalizing terms Indians of Mexico, Indians of Central America, and Indians of South America. In sum, the use of the formula Indians of… in subject lists requires a critical reformation that allows for a better representation of Indigenous peoples and topics in metadata practices.

In contrast to the IP-G subset, the IP-S subset is built around subjects that specify the Indigenous group they are referring to, which could potentially mean that the cultural specificity of these communities is being more accurately represented. Nevertheless, the IP-S subset is the only subset where all materials were published and are held in the Global North (Figure 2). Therefore, despite its apparent cultural diversity and specificity, this subset reveals that the Library’s materials about Indigenous peoples continue to be anchored in a stark contrast between the North as epistemic subject—those who know—and the South as a passive object—that to be known. Furthermore, the dichotomy subject-object perpetuates the historical colonial association between nonhuman animals and Indigenous peoples. The animal-human dichotomy is the primeval binary on which other power-based categories are constructed and that characterizes the relationship between colonizers and colonized (Wolfe xx; Rajamannar 5). In this sense, several critics have identified a profound connection between colonization in the Americas and the establishment of a racial system for the classification of human groups (Greer et al.). Therefore, a link between species and race is constructed in the world-making process of colonization, a link that is still palpable in the use of subjects such as Indians of… in BHL’s metadata. These colonial dynamics can only be counteracted by an acknowledgement of the colonial roots of such materials as well as a stronger diversification of the Library’s collection that increases the presence of Indigenous peoples as agents of knowledge production.

Figure 2 Map showing places of publication (density per number of records) in the IP-S subset. Generated on Tableau in August 2021. Data from https://www.biodiversitylibrary.org/data as of July 1st, 2021.

Even in subsets such as LAC, where there is a considerable diversity of places of publication,[xiv] the holding institutions of the materials are almost exclusively located in the Global North (Figure 3). In fact, there are only three works—out of 1982 unique IDs in this subset (LAC)—that are held in institutions outside of Europe and Anglo North America: Jean Baptiste Boussingalt and Francois Desire Raulin’s Viajes científicos a los Andes Ecuatoriales ó Colección de memorias sobre física, química e historia natural de la Nueva Granada, Ecuador y Venezuela (1849) translated by Joaquín Acosta and held in the Universidad Autónoma de Nuevo León in Mexico; the Anales del Museo de Historia Natural de Valparaíso, authored and held by the Museo de Historia Natural de Valparaíso in Chile; and Evangelina Schwindt, Nicolás Battini, Clara Giachetti, Karen Castro, and Alejandro Bortolus’s Especies exóticas marino-costeras: Argentina (2018), edited and contributed to BHL by authors Schwindt and Bortolus.

Figure 3 Map showing places of publication (density per number of records) in the LAC subset. Generated on Tableau in August 2021. Data from https://www.biodiversitylibrary.org/data as of July 1st, 2021.

Boussingalt and Raulin’s work and, especially, Acosta’s translation is, in fact, a considerable example of historical efforts to decentralize biodiversity-related knowledge. Acosta himself, in his prologue, highlights the goal of his translation as being that of the sharing of the knowledge produced by Boussingalt and Raulin in French, especially so that it can be accessed by “Granadinos, Venezolanos y Ecuatorianos” (Acosta 10) that is, the people from the places established as the object of study of this work, Colombia, Venezuela, and Ecuador. As the translator mentions, the people from these countries had, during that time, little to no access to these volumes due to their limited distribution and high costs. Furthermore, Acosta emphasizes that his translation is the product of a strong collaboration between him and the authors and that the edition was sponsored by a French editor (10-11). Thus, already in 1844, Acosta’s translation of this work was an example of collaboration between the Global South and North. Moreover, his goals echo in the access provided to these volumes by BHL, even more so given that they are contributed by a Mexican institution, meaning that a fundamental part of the collaborative network of bio-diverse knowledge production of this text, from the 19th century to today, is notably located across the Global South.

In turn, both the volumes of the Anales of the Museo de Historia Natural de Valparaíso in Chile and Schwindt et al’s book are powerful examples of local biodiversity-related knowledge production. The latter is also a more recent example of global collaboration and the role of BHL in promoting it, and it exemplifies the fruitful outcomes of a truly diverse and trans-geopolitical network of bio-diverse knowledge production. The result of an extensive local project to understand marine species in Argentina, Especies exóticas marino-costeras is also bilingual (English and Spanish) and includes the voices of researchers in different parts of the world. It incorporates, for instance, three prologues by researchers from the US, South Africa, and Argentina respectively. Similarly, this book includes a remarkably diverse list of acknowledgements, with individuals from Canada, Argentina, Uruguay, Brazil, France, Sweden, South Africa, the US, the UK, Spain, the Netherlands, Colombia, and Chile, who the authors thank for “provid[ing] valuable assistance and help during the entire creative process … by supplying administrative assistance, photographs and specimens, as well as commenting and improving the text” (Schwindt et al. 15–16). Finally, this work was also featured in a BHL blog post published in October 2020 and written by former Outreach and Communication Manager for BHL, Grace Costantino. At the moment of publication, the blog post was also shared on the Library’s accounts on Twitter and Facebook, thus promoting this bio-diverse trans-geopolitical collaborative work through other important online avenues, a decolonial and truly global representation of Latin America for which I myself have argued elsewhere.

Despite, however, the richness and diverse geopolitical affiliations of these materials, their small numbers blur their presence and continue to evidence an overwhelming predominance of the Global North as the housing site of biodiversity-related knowledge production, thus positing it as a sort of legitimizing epistemic centre for such production. While local and collaborative knowledge production is present in it, BHL still has a long way ahead to achieve a more equitable and diversified collection. In sum, what the statistics and maps of these subsets show is a colonial object-subject relationship between the Global South and North that requires a profound diversification of the networks of knowledge production that constitute BHL’s catalogue.

In the case of BHL, global partnerships are a strong strategy to achieve such diversification, both in terms of its collection and of access to materials. A notable example of this is the MEX subset. A significant portion of the materials about Mexico in BHL is contributed to its collection through the BHL México project, which is the result of a partnership between the Library and several Mexican institutions led by Mexico’s Comisión Nacional para el Conocimiento y Uso de la Biodiversidad (CONABIO). As a result, the MEX subset shows significant diversification in comparison to other subsets, especially regarding places of publication and language distribution. For instance, Mexico City is, by far, the most frequent place of publication in this subset, meaning that the MEX subset is the only one where the most frequent place of publication is not in the Global North and is located in the same country that is included as subject (Figure 4).

Figure 4 Frequencies of places of publication (non-unique IDs) in the MEX subset. Generated on Tableau in August 2021. Data from https://www.biodiversitylibrary.org/data as of July 1st, 2021.

Such a strong presence of Mexico in the geographical affiliations of these texts thus posits the country not only as an object of study but, especially, as an agent of knowledge production, particularly of local knowledge production. Furthermore, although in most[xv] subsets—including the MEX subset—English continues to be the most frequent language of publication, the MEX subset shows a considerable number of materials in Spanish, with the difference between the two languages being significantly smaller than in other subsets (Figure 5). Therefore, at least to a certain extent, the meaningful presence of Spanish in materials about Mexico—which are part of both the BHL México project and the collection Publicaciones en español—counteracts the English dominance found throughout BHL’s collections, as I have also shown before.

Figure 5 Frequency of English and Spanish in all subsets. Data from https://www.biodiversitylibrary.org/data as of July 1st, 2021.

Moreover, it is important to note that the substantial presence of Spanish as a fundamental language of publication in the MEX subset is only true for non-unique IDs. On the contrary, when considering only unique IDs in this subset, the gap between English and Spanish increases dramatically, showing, once more, an overwhelming predominance of the former (Figure 6). Given that many of the materials contributed through BHL México and CONABIO are also periodicals, the difference of distribution when considering unique and non-unique IDs further strengthens the hypothesis that linguistic diversification in the MEX subset is strongly related to this partnership. Thus, the diversification of BHL’s collection proves to be deeply related to the diversification of its global partnerships, the BHL México project being a case in point.

Figure 6 Language distribution for non-unique (left) and unique (right) ID titles in the MEX subset. Graph generated on Tableau in August 2021. Data from https://www.biodiversitylibrary.org/data as of July 1st, 2021.

The Biodiversity Heritage Library is a rich and invaluable resource for researchers across the world. Nevertheless, to maximize access to its collection and an equitable and decolonial representation of diverse human and nonhuman subjects, the Library must continue its efforts regarding decoloniality, diversity, multiculturalism, and multilingualism. Diversification must be a primary goal of BHL’s outreach strategies. As I argued in a previous post on BHL’s blog, increasing multicultural representation can lead to diversified and equitable access that can increase the participation of audiences from the Global South including, at the very least in the case of Latin America, Indigenous communities. Therefore, a critical understanding of metadata, representational paradigms, and collection practices can highlight the dimensions of the Library where more work needs to be done as well as the benefits of previous strategies undertaken towards diversification. For instance, in the case of the subsets I analyzed during my internship with BHL, a stronger collaboration between the Library and institutions in the Global South seems to have a positive impact on the diversification and multicultural representation of topics related to these regions of the world. Thus, the establishment of more and stronger BHL nodes in the Global South can potentially lead to a more decolonial, equitable, and truly global collection anchored in bio-diverse networks of knowledge production, so that the Library can continue to honour its mission of providing global and open access to biodiversity-related content in collaborative and ethical ways.

[i] The files used for the extraction and building of the subsets were creator.txt, item.txt, subject.txt, and title.txt.

[ii] As of this date, the total number of BHL records was 146,806.

[iii] All subjects are presented here and used in the analyses exactly as they appear in BHL’s subject dataset (as of July 1st, 2021).

[iv] Total numbers for these subsets include non-unique ID records, meaning that some title IDs are repeated when the work includes several numbers, volumes, or issues.

[v] Includes the subjects Brazil 0 and Brazil 7 as they appear in BHL’s subject dataset.

[vi] Tables for these subsets (including cleaned geographical data as explained later in this text) can be found in BHL’s GitHub repository.

[vii] Data were cleaned directly on Microsoft Excel, separating values by punctuation and replacing terms that needed modernization and/or translation.

[viii] See note vi.

[ix] This means that works in volumes, issues, and/or numbers were treated as a single record and comprised in their shared title ID.

[x] Additional details can be found in the description of each map on Google. For the IP-G and IP-S subsets, additional information about Indigenous groups is included in each node on the map.

[xi] All visualizations for all subsets can be found in BHL’s GitHub repository.

[xii] Given the one-subject nature of the subset, all subject-based visualizations are not included in the MEX subset.

[xiii] The exception is the MEX subset, which is discussed later in this text.

[xiv] Meaning that they include several places of publication throughout Latin America. It is important to note, however, that there is almost no knowledge production about Latin America from other regions in the Global South.

[xv] The most frequent language in the IP-G subset is French, followed by English. In the GR, LAC, and IP-S subsets, French is the second most frequent language, after English.

[xvi] It is notable that the IP-S subset, along with its dichotomy between the Global South and North in terms of places of publication, is also the only subset with no materials in Spanish.

Works Cited

Acosta, Joaquín, translator. ‘Advertencia Preliminar’. Viajes Cietíficos a Los Andes Ecuatoriales ó Colección de Memorias Sobre Física, Química e Historia Natural de La Nueva Granada, Ecuador y Venezuela Presentada a La Academia de Ciencias de Francia, by Jean Baptiste Boussingault and Francois Desire Raulin, Librería Castellana, 1849, pp. 10–11, https://www.biodiversitylibrary.org/item/235575#page/6/mode/1up.

Freeland, Chris. ‘BHL Poster for AETFAT2010’. ChrisFreeland, 19 Apr. 2010, http://blog.chrisfreeland.com/2010/04/.

Greer, Margaret, et al., editors. Rereading the Black Legend: The Discourses of Religious and Racial Difference in the Renaissance Empires. 2008.

Rajamannar, Shefail. Reading the Animal in the Literature of the British Raj. Palgrave Macmillan, 2012.

Schwindt, Evangelina, et al. Especies Exóticas Marino-Costeras: Argentina. Marine-Coastal Exotic Species: Argentina. Edited by Evangelina Schwindt and Alejandro Bortolus, Vázquez Mazzini Editores, 2018, https://www.biodiversitylibrary.org/bibliography/169323.

Sistema de Información Cultural SIC México. ‘Pueblos Indígenas’. Gobierno de México, 2021, https://sic.cultura.gob.mx/lista.php?table=grupo_etnico&disciplina=&estado_id=.

UNICEF España. UNICEF Presenta El Atlas Sociolingüístico de Pueblos Indígenas En América Latina | UNICEF. https://www.unicef.es/prensa/unicef-presenta-el-atlas-sociolinguistico-de-pueblos-indigenas-en-america-latina. Accessed 29 June 2020.

Wolfe, Cary. ‘Introduction’. Zoontologies: The Question of the Animal, edited by Cary Wolfe, University of Minnesota Press, 2003, pp. ix–xxiii.

The large-leaved kōwhai (Sophora tetraptera) is native to the North Island of Aotearoa New Zealand, where it is widespread and common. While it grows naturally only in Aotearoa, it has been cultivated extensively outside of this range, including as one of a number of New Zealand plants historically introduced into English gardens.[1]

Natural history literature provides a record of such introductions, with many authors remarking on the suitability of Sophora tetraptera—often referring to it under the synonym Edwardsia grandiflora—to the English climate. For example George Loddiges, within his The Botanical Cabinet (1826), remarked that it was “sufficiently hardy to bear our climate, planted against a wall; in very severe frost, a mat should be hung over it.”[2] Two decades later, British writer and botanical authority Jane Wells Webb Loudon included the plant in her The Ladies’ Flower-Garden of Ornamental Greenhouse Plants (1848), where she noted that the original plant could still be found growing in Chelsea Garden in 1848.[3]

Sophora tetraptera, under the synonym Edwardsia grandiflora. Plate 1162. The Botanical Cabinet. v. 12 (1826). Art by George Cooke. Contributed in BHL from the Missouri Botanical Garden, Peter H. Raven Library.

Loddiges’ and Loudon’s references are just two of many sources related to the history of Sophora tetraptera that Mark Laird (Professor, John H. Daniels Faculty of Architecture, Landscape, and Design, University of Toronto) identified thanks to the Biodiversity Heritage Library (BHL) whilst conducting research for a prospective book. One of the book’s chapters explores a variety of New Zealand plants introduced into English gardens from the 1770s to 1840s, in the context of both Kew’s colonial collecting and Māori heritage. BHL was an invaluable resource for this research—especially during the COVID-19 related lockdowns of 2020.

“I am a huge fan of BHL, and my admiration increases every year,” lauds Laird. “Last year, under Covid, I discovered I could research a large topic through online access to the BHL. The search engine allowed me to find all the collected references to many plant species, including illustrations of those species.”

Mark Laird (Professor, John H. Daniels Faculty of Architecture, Landscape, and Design, University of Toronto). Image courtesy of Mark Laird.

Laird’s expertise in historic landscape architecture is rooted in four decades of research. Over his career, he’s held fellowships at Chelsea Physic Garden, London, and Dumbarton Oaks, Washington, DC, and, prior to his appointment at the University of Toronto, was a Senior Lecturer in the Graduate School of Design at Harvard University for fifteen years. Laird has also authored many publications on historic landscape gardening.

Dust jacket for A Natural History of English Gardening (2015) by Mark Laird. Laird sourced many of the illustrations used within the publication from the Biodiversity Heritage Library (BHL). Image courtesy of Mark Laird.

It was whilst working on his most recent publication, A Natural History of English Gardening (2015), that Laird first discovered BHL. Within the work, which was awarded the 2019 Elisabeth Blair MacDougall Book Award from the Society of Architectural Historians, Laird draws on historic documents to help chronicle and illustrate the history of early modern gardening and its cultural and scientific intersections. Jacqueline Dearborn (née Ford), then a Digital Projects Librarian at Harvard Botany Libraries, introduced Laird to BHL, who made enthusiastic use of the Library as a resource for many of the illustrations included in his Natural History.

“When sourcing images for my publications twenty-five years ago, many institutions were charging exorbitant reproduction fees,” recalls Laird. “It was a game changer to find that a good number of those institutions have now made their collections accessible online with a simple cost-free download of a High Resolution Image. BHL, as part of that shift in ‘intellectual commons,’ has changed the game for me.”

The African Land-Tortoise (Testudo graeca) by George Edwards from his A Natural History of Uncommon Birds, v.4 (1751). Contributed in BHL from National Library Singapore. Edwards’ illustration appears within Laird’s A Natural History of English Gardening (2015), where he shares that naturalist George White likened his aunt’s tortoise, Timothy—of the same species as Edwards’ Testudo graeca, to “a snail with a shell on its back.” White’s work represents a major theme in Laird’s Natural History.

BHL is now Laird’s go-to resource for natural history literature, particularly when it comes to accessing high quality illustrations. While he appreciates the access BHL provides, he does hope to see the Library implement a more streamlined high resolution image download process in the future, as the current method involves multiple steps.

Laird isn’t alone in his love of BHL’s illustration collections, as evidenced by the popularity of the BHL Flickr, which provides access to over a quarter of a million images from the Library. As BHL further develops its tools and services, staff will continue to explore ways to improve access to and features around scientific artworks.

As for Laird, BHL has opened up a whole new world of possibilities when it comes to online research, and he intends to make the most of these opportunities. “This summer, I expect to continue ‘remote’ research thanks to BHL, which is a game-changer in how a scholar can work from home and at any hour of the day,” affirms Laird.

As researchers continue to adapt to and evolve within increasingly dispersed, asynchronous working environments, BHL will be there to support them wherever and whenever their work takes them.

References

The yellow fever mosquito (Aedes aegypti), under the synonym Stegomyia fasciata. Source: Howard, L. O. (Leland Ossian). Mosquitoes; how they live; how they carry disease; how they are classified; how they may be destroyed. 1901. Contributed in BHL from Smithsonian Libraries and Archives.

Aedes aegypti, dubbed the yellow fever mosquito, is a globally invasive, pervasive threat to human health. As the common name suggests, the species can carry a range of diseases, including not only yellow fever but also dengue, the Zika virus, and the chikungunya virus. It is responsible for an estimated 400 million infections each year.[1]

Originating in Africa, Aedes aegypti has spread across the globe, initially transported via ships carrying enslaved people to the Americas. Soon after, trade ships bearing goods like sugar from the Americas brought the species to Europe, and from there it eventually made its way to Asia by way of the Suez Canal. Aedes aegypti’s success is secured by its ability to survive in a range of tropical, subtropical and temperate regions and its threat is exacerbated by its ability to breed in clean water supplies and propensity to bite during the day, rendering protective bed nets ineffective. The effects of climate change are expected to increase A. aegypti’s potential range significantly, escalating the urgency of implementing effective species control methods.[1,2]

Notonecta indica perched on a piece of wood. Photo courtesy of: Gavin Campbell.

The aquatic insect Notonecta indica—a species in the family Notonectidae, commonly called backswimmers because they swim on their backs—preys on Aedes aegypti mosquitoes in the Americas. Gavin Campbell, a PhD candidate at the University of the West Indies in Jamaica, is researching this predator / prey relationship as part of his PhD. Specifically, Campbell hopes to quantify the control of Aedes aegypti by Notonecta indica by determining how many mosquitoes each adult predator can consume daily and throughout their lifetimes.

The Biodiversity Heritage Library (BHL) has been an invaluable resource for this research.

“Thanks to the articles from BHL, I was able to substantiate my data with the foundational data to prove that other researchers have found similar things,” shares Campbell. “Before I found the BHL resources, I was only able to find a single paper from 2020 that was able to offer anything close to my research. I am very grateful for these and other documents I have come across since then.”

Gavin Campbell, a PhD candidate at the University of the West Indies in Jamaica, collecting water samples from a pond as egrets fly around. Photo courtesy of Gavin Campbell.

Campbell has been studying the aquatic and terrestrial phases of temporary ponds since 2017, with the life history of Notonecta indica and its impact on mosquito populations being a major component of that research. As Campbell explains on his website, “temporary waters are bodies of water which recurrently dry”, ranging from the minute—”water inside a snail shell”—to the massive—”the sizes of lakes and rivers”. As these bodies may be “freshwater, brackish, saline and even hypersaline (saltier than the sea)”, a wide variety of species are adapted to live in these diverse ecosystems.[3]

Campbell regularly uses BHL to access historic documents related to his research, typically downloading selected pages as PDF files to read offline.

“BHL has been critical in providing me with baseline information for my research as the documents in question were difficult to find,” asserts Campbell.

A variety of Notonecta species, including Notonecta indica (fig. 2). Source: Hungerford, H. B. The Genus Notonecta of the World. 1933. Contributed in BHL from Harvard University, Museum of Comparative Zoology, Ernst Mayr Library.

Of all of the titles in BHL’s collection, The Genus Notonecta of the World (1933) by H. B. Hungerford has had the greatest impact on Campbell’s research. Published in The University of Kansas Science Bulletin, the work attempted to assemble all available information about the Notonecta genus, noting that more than thirty years had passed since the last comprehensive account of the genus was published. The publication represented twelve years of research, during which Hungerford consulted both the original descriptions and the types in museums throughout North America and Europe in an effort to “account for every named species”. The work was illustrated in color by Kathleen Doering.[4]

The Genus Notonecta of the World helped Campbell perform species identifications and capitalize on various behavioral aspects. For example, Hungerford notes that eggs are deposited on hard surfaces. Campbell applied this knowledge to his own work by placing a mesh within his rearing container, which facilitated easy removal of eggs each day. Hungerford also provides useful details about Notonecta behaviour in natural environments. For example, in a pond which N. indica frequents during the rainy season, a large number of branchiopods are present, which Hungerford notes are a food source for N. indica’s younger stages. This thus serves as a contributing factor in their dispersal and biology.

The illustrations in The Genus Notonecta of the World were also useful for Campbell’s research, helping him identify organs and structures within individuals and distinguish males and females for successful mating and to help determine differences in mosquito consumption between males and females.

“The BHL was key to providing access to the valuable information in this publication, as few papers at present go into such detail on the family,” affirms Campbell. “Without the BHL providing access to this foundational article, I would have been set back and hindered in my research.”

With Aedes aegypti expected to pose continued, significant threats to human health, we are proud to know that BHL is empowering important research that can help scientists better-understand the species and implement effective control methods. As Campbell shares, “With a better understanding of the role of N. indica in mosquito control, I aim to create and support more habitats for these natural predators to suppress mosquito populations, using nature-based solutions to address the extensive effects of climate change.”

As Campbell’s testimonial demonstrates, immediate, online access to biodiversity literature is ever-more important, allowing research to proceed more quickly and efficiently and improving our ability to respond to the many crises facing our planet and our species today. We look forward to continuing to provide researchers like Gavin Campbell with the essential literature they need to empower their work—anytime, anywhere.

Thank you @BioDivLibrary for this and so many other treasures pic.twitter.com/qDkfTH0kSc

— ?️‍?Gavin Campbell ?️‍? (@dragonecology) July 1, 2020

https://platform.twitter.com/widgets.js

References

The Joseph F. Cullman 3rd Library of Natural History at the Smithsonian Libraries and Archives in Washington, D.C. is home to many rare and special books. Amongst the approximately 20,000 volumes in the library’s collection is a particularly rare item—one of only six known copies in public and institutional libraries worldwide.

Title Page. Byron, Edmund. What we did in South Africa in 1873 (1874). Digitized copy from the Joseph F. Cullman 3rd Library of Natural History at the Smithsonian Libraries and Archives. Freely available in the Biodiversity Heritage Library.

The book is a privately printed journal by Edmund Byron entitled What we did in South Africa in 1873, which details Byron and his wife Charlotte’s 1873 exploratory and hunting expedition to South Africa. In 2015, at the request of a researcher, the Smithsonian digitized its copy of the journal and made it freely available online in the Biodiversity Heritage Library (BHL), where it was discovered by Dr. Nigel Elliott, who was conducting research on the Byron family.

“My research was prompted by a donation to the Museum of Croydon, UK, of an archive of family papers relating to the Byrons, who were Lords of the Manor and Squires of the village of Coulsdon from 1782 to 1921,” explains Elliott. “Since the archive’s acquisition in 2012, a team from the Bourne Society has been working on cataloguing the archive, with a view to online publication. The archive provides a rich resource for exploring the domestic, social, business and farming history of the family, their role and relationship as local squires within the wider village community of Coulsdon, and the extensive foreign travels undertaken over the years by members of the family.”

Dr. Nigel Elliott. Photo courtesy of Dr. Elliot.

Inspired by his work with the Bourne Society cataloging these archives—and the intensive research that it required—Elliott authored a book on the family, published by the Society: The Byrons of Coulsdon: Abroad and at Home (2020). The book itself covers three generations, with a particular focus on Edmund Byron.

Cover: The Byrons of Coulsdon: Abroad and at Home (2020) by Dr. Nigel Elliott. Image courtesy of Dr. Elliott.

One of Elliott’s chapters is devoted to Edmund and Charlotte’s South African expedition—as detailed in Edmund’s journal—during which the pair, accompanied by Edmund’s manservant Mr. Kemp, traveled via a combination of coaster (a shallow-hulled ship used for trade) and ox wagon from Cape Town to present-day Kruger National Park and Great LimpopoTransfrontier Park, where, faced with the remote and challenging terrain of that area in 1873, they travelled and hunted wholly on foot. The Byrons employed a team of local Black staff—on whose hunting skills the Byrons were particularly dependent—as well as a White employee, Mr. Dubois, to manage the expedition.

“In the chapter, I follow Edmund’s journal closely and in detail,” shares Elliott. “I have attempted to place their visit in the context of South Africa at the time. My background reading revealed to me the extent to which the 1870s were critical in the country’s history, and I have attempted to set this decade within the wider trajectory of South African history.”

Online access to Edmund’s journal was indispensable for Elliott’s research and the publication of his book.

“From Edmund’s journal, which I accessed online via a PDF download of the Smithsonian’s copy on BHL, I was able to construct a detailed narrative of the tour,” affirms Elliott. “This easy access to the journal was critical in providing a thorough history of the family and in meeting the research objectives set out in the book.”

According to Elliott, the journal provides valuable perspective on the dualities of the Byrons’ lives as both a gentry family deeply rooted in a small English community and intrepid world travelers eager to experience different places and cultures. The journal also reveals a duality within Edmund himself as he—a privileged White male—interacts with new cultures and people. When referring to the Black men and women they encountered, Edmund sometimes uses derogatory terms reflective of the White supremacist culture so pervasive during the European colonial period. At the same time, and unlike many of his compatriots, Edmund strove to act with integrity towards his Black staff and others and even, when faced with a direct appeal for help, enabled one woman to escape enslavement by providing her with protection and employment.

“The journal provides a fine opportunity to deepen the character study of Edmund, a truly patriarchal man in the context of his home life in Coulsdon, that comprises one of the central themes to my book,” shares Elliott.

The Smithsonian’s copy of Edmund’s journal was originally owned by Eliza Vans Agnew, who had connections to Edmund through his wife, Charlotte, whose mother was a Vans Agnew. Today, the volume is part of the Russell E. Train Africana Collection of the Smithsonian Libraries and Archives, which contains manuscript and printed materials, photographic prints and negatives, slides, audio tapes, film, original and reproduction artwork, maps, scrapbooks, and historical and natural artifacts related to the history of African exploration and natural history. The collection was originally assembled by the Honorable Russell E. Train, a former judge, top administrator of the Environmental Protection Agency, and a past president of the World Wildlife Fund. It was acquired by the Smithsonian in 2004.

The circumstances surrounding the digitization of Edmund’s journal was a fortuitous coincidence for Elliott. The Smithsonian Libraries and Archives digitized the volume—unbeknownst to Elliott—at the request of one of his colleagues, who was conducting tangential research with possible connections to the Byrons’ South African expedition.

“My colleague was impressed by the courtesy of the Smithsonian staff, who responded quickly to his expression of interest by digitizing Edmund’s journal,” shares Elliott. “As it turned out, the book was of no direct use to him but, as this story reveals, there was a serendipitous consequence of the Smithsonian’s helpful response and the digitized journal proved invaluable for my research.”

When Edmund Byron decided to “have printed a few copies of the Journal which we kept during our travels”, he did so with the expectation that it “would be of no interest to the general public.” Would he be surprised to know that, nearly 150 years later, researchers would view this journal as an informative historical document useful for modern-day research? Thanks to the Smithsonian Libraries and Archives and the Biodiversity Heritage Library, this volume—once available only to a few select readers—is now freely accessible to researchers everywhere.

A version of this post was originally published on the Cogapp blog on 30 March 2021.

AI generated illustrations of birds created with illustrations from the BHL Flickr using machine learning with a StyleGAN.

I work as a web developer for the agency Cogapp, which is based in Brighton, UK. We create websites and other digital services for museums, art galleries, archives and the like, but every couple of months we hold a “hack day.” A hack day involves spending a day working on projects which generally revolve around a particular theme and which ideally we can do in one day. This allows us to get the creative juices flowing and to further our agenda of innovation.

The theme this past hack day at Cogapp was “Museum APIs,” but the looser interpretation was that we were to use open data provided by museums in our projects. I was inspired by the Biodiversity Heritage Library’s Flickr, which is a massive collection of free-to-use scientific images. I immediately knew I wanted to utilise this resource as I love scientific illustrations of nature.

I’ve also had an interest in Machine Learning (ML) for a while, and I recently discovered Derrick Schultz and his YouTube channel Artificial Images. Here, he publishes videos of his Machine Learning courses which he runs for people who want to use ML for creative purposes.

I watched Derrick’s tutorials on training a StyleGAN Neural Network and the things he was saying made a degree of sense to me, plus he had published a handy Google Colab notebook with step-by-step code, so I decided it was something I might be able to have a go at.

What is a StyleGAN?

StyleGAN stands for Style Generative Adversarial Network and was originally developed by NVLabs. The purpose of a StyleGAN is to train and learn from input images and produce its own “fake” images back. You may have seen the site ThisPersonDoesNotExist.com, which features fake photos of people which a StyleGAN has produced. This network has been trained on thousands of images of people’s faces and has in turn learned to produce new images which (for the most part) look like real photographs of people’s faces.

Three images from ThisPersonDoesNotExist.com; two successful, one less so…

A StyleGAN does this by processing through the set of real images it is given numerous times during which it also produces its own images. The “adversarial” part of the name refers to the fact that a StyleGAN actually comprises two Neural Networks which work against each other. The first network produces the fake images and the second will look at those images and try to determine whether it thinks they are “real” images or “fake.” The longer each network trains, the better each gets at its job, and the more likely the images produced will trick the second network and ultimately also humans into thinking they’re real.

StyleGAN has been updated a few times and in 2020 StyleGAN2-ADA was released, which now allows us to train a network on very few images, actually as few as 500–1000 as opposed to 10,000s just a couple of years ago, and it will learn to produce good quality fakes in a very short period of time (a few hours of training). This advancement allowed me to feel like I could attempt to train my own, as 1000 images was something I could get and I felt like I could train the model during the course of a hack day and get some kind of result to show at the end.

Creating the dataset

The biggest (active) part of this project time-wise was collecting the images and making them look roughly uniform by cropping them to the same size and having the subject roughly in the same spot in each one. This was important as when training this model I wanted the network to focus on the subject of the image itself and not, for example, where in the image the subject was, as that would waste precious training time.

I looked through the BHL Flickr and decided to use scientific illustrations of birds. I love birds and the style of scientific illustrations, and I was excited at the idea of the new and weird birds my StyleGAN model might produce! I chose illustrations in which the bird was in the centre of the page and the rest of the space was ideally empty. Due to time constraints, I did allow some illustrations with more than one bird in them and some with background features such as grass. This wasn’t ideal for the above reasons, but I decided to give it a go and see what the results would be like anyway.

A contact sheet of “reals” or real images taken from my dataset from the BHL Flickr. This is produced by the StyleGAN at the beginning of training.

Training my model

Once I had my dataset ready to go, I used Derrick’s Google Colab notebook to train my model. In order to run StyleGAN2-ADA, you need “1–8 high-end NVIDIA GPUs with at least 12 GB of GPU memory,” which is more memory than most home computers have, so using something like Google Colab where you’re essentially borrowing their super powerful computers means that even an average joe like me can have a go at running memory-heavy code like this and it’s free for the basic version, which is enough for this purpose! The other benefit is that you can connect it to a Google Drive for storage. This was useful in my case because I needed to store 1000 images to train my model with, and the model outputs large images itself.

The code I was using was set to train a new StyleGAN model based off of one that had already been trained for many hours. A model in Machine Learning is essentially a file that has been trained to recognise patterns. Using a pre-trained model as a starting point like this is called transfer-learning, and it can drastically speed up the time it takes for your StyleGAN to learn from your images and start producing something which looks good. The model you’re using as a base for training a StyleGAN doesn’t have to be similar in subject matter to the images you will be training it with, and in this case I used the ffhq model by NVlabs, which is trained from a set of images of faces. This meant that when it started to train, it initially produced images which looked like a mixture of faces and my illustrations.

After each few iterations through the dataset, the model produced a 7 by 4 grid of sample images so I could see how it was getting on. The image below is the first sample image I got, which shows the original faces mixing with some features of my images!

Totally not terrifying image produced by the StyleGAN model after one of the first cycles of training.

As time went on, the shapes it was producing started looking more and more bird-like. By the end of the hack day, I had only managed to train my model for about 3–4 hours. But, by that time I was starting to see the faces disappear and bird features appear. Pretty impressive for just a few hours, but also still pretty abstract!

The fakes produced by the end of the hack day in around 3–4 hours of training.

I wanted to see whether the model could produce more recognisable bird images, so I continued to train it after the hack day ended and the images started to look better and better…

The fakes produced after about 3–4 hours more training.

I left the model training over the weekend and the fakes actually did start looking like they could be illustrations of birds in many cases! Below is the final set of sample images I produced with my model. I would guess I spent about 40 hours training to get to this point.

The final set of sample images generated by the StyleGAN after around 40 hours of training.

Okay, so there are a few images in there that maybe wouldn’t fool a human into thinking they were real illustrations of birds… ahem…

But the majority kind of do look like birds. I’m pretty sure if you didn’t know this was generated by a Neural Network you would have no trouble believing it was a real bird illustration.

It does maybe look overly happy for a bird though?

Here are a few more images produced which I think look very convincing.

AI generated illustrations of birds created with illustrations from the BHL Flickr using machine learning with a StyleGAN.

This is so impressive considering I only had 1000 images and it was around a couple of days worth of training!

Creating interpolation animations

While learning from the given dataset, a StyleGAN will look for patterns or “features” in the images and organise them into a multi-dimensional graph. This is called latent space.

When I talk about “features,” in the simplest terms I mean things like brown vs white feathers, for example. So, brown feathers might be point A in this space and white feathers is point B. If you move from point A to point B and produce images every so often in between, you would see the feathers turn from brown to white over time. This is called interpolation.

Using another of Derrick Schultz’s Google Colab notebooks, I generated a couple of videos of interpolations through my model. I used the “circular” interpolation which gives a nice “walk” through a few diverse points in the latent space and you see different birds morph into one another. Here is one of the outcomes:

As it “walks” through the latent space, different features morph from one to another.

Conclusion

The fact that with StyleGAN2-ADA you can use such a limited dataset and need so little time, and that you can use a free service like Google Colab to do the training and it still produces results like this is incredibly impressive and opens the word of Machine Learning and StyleGANs to many more people than previously. It is also very exciting to see and think about all of the creative, as well as practical, applications of this technology. Practical applications of GANs include: “image-to-image translation,” which might involve converting a photograph taken in the day time to night time for example, so as a sort of AI-led photoshop; being able to generate higher resolution versions of images; or even generating missing parts of an incomplete image.

I have to obviously shout out to Derrick Schultz for his Colab notebooks and tutorials which also helped lower the barrier. It meant that I was able to get training much quicker than I would have otherwise but the code is still there for me to learn from, amend and dissect. I encourage you to check out his YouTube channel if you’re interested in learning more about training StyleGANs. He also runs regular courses on Machine Learning for creative purposes and you can find out more about them on his website.

I also encourage you to have a look through the Biodiversity Heritage Library’s 250,000+ images on their Flickr. It’s an impressive and incredibly interesting collection of images from the natural world, organised into different categories. It’s the perfect place to find datasets to use in a StyleGAN model or just for inspiration!

I’m looking forward to using this technology with more datasets and seeing what new and weird flora, fauna and other images I might be able to create!

Leer en español

A view of the Panamanian tropical rainforest in Colón Province. Photo by Samanta Orellana.

In the last decades of the 19th century, a monumental publication on the biodiversity of Mexico and Central America began publication—Biologia Centrali-Americana. Published in 215 parts from 1879 to 1915 by the editors Frederick DuCane Godman and Osbert Salvin, the work describes over 50,000 species and is illustrated with over 1,600 lithographic plates depicting over 18,000 species. Remarkable for its time, the title is still vitally important for the study of Neotropical biodiversity today, as it contained virtually all known information at the time about Mexican and Central American flora and fauna.

Biologia Centrali-Americana is a particularly important resource for entomologist Samanta Orellana, a PhD student in evolutionary biology at the Dr. Nico Franz Lab of Arizona State University (ASU) and a research assistant in the ASU Biocollections of the Biodiversity Knowledge Integration Center. Orellana began studying insects and working with entomological collections more than a decade ago, during her undergraduate studies in her home country of Guatemala.

“For many insect groups in Guatemala and the rest of Central America, Biologia Centrali-Americana still represents the only source of information available for the region,” states Orellana.

Samanta Orellana in the Arizona State University Biocollections. Photo by Kevin Cortés.

Biologia Centrali-Americana is also meaningful to Orellana because it led her to the Biodiversity Heritage Library (BHL). In 2010, whilst identifying Guatemalan insects as part of her undergraduate studies, Orellana started using the “Electronic Biologia Centrali-Americana,” available from the Smithsonian Libraries and Archives, which directed her to BHL. She hasn’t stopped using BHL since.

“BHL is simply wonderful,” lauds Orellana. “I think it is an amazing tool for entomologists all around the world, especially for those of us who work or have worked in collections in developing countries, usually without access to specialized libraries or journals, but with a lot of material to identify.”

One of the plates of Central American anthribids, included in the chapter written by Karl Jordan in the Biologia Centrali-Americana (Insecta. Coleoptera. v.4, pt.6, 1895-1907). Contributed in BHL from Smithsonian Libraries and Archives.

Orellana specializes in the family Anthribidae (Coleoptera), commonly known as fungus weevils. Her research on these insects began during her studies at the Universidad de Panamá, where she completed an M.Sc. in entomology, and continues through her PhD work, with her dissertation focused on the family’s diversity and evolution. As part of this research, Orellana recently published a catalogue of the Anthribidae species from Panama.

BHL’s collections have been instrumental in enabling Orellana to pursue her work on the family.

“BHL was essential to complete my first published work on Anthribidae, a catalogue of the species from Panama,” shares Orellana. “This group of weevils has been poorly studied in the Neotropics, so identifying specimens is always challenging due to the lack of taxonomic keys. I had to rely mostly on original descriptions available through BHL to identify the specimens I examined, both in Panamanian and American collections. The catalogue ended up with 44 new records for the country, most of them identified thanks to the literature available in BHL.”

Eugonus robustus Jordan, 1904 (Coleoptera: Anthribidae), from Panama. Photo by Samanta Orellana.

BHL has now become a core part of Orellana’s research process. She uses the Library at least several times a week, reading content online or downloading material via BHL’s custom PDF generator to build her own local collection of Anthribidae literature. Her favorite feature is BHL’s full text search service, which allows her to easily find scientific names or specific localities within sources.

“I use BHL all the time,” affirms Orellana. “I continue using it for my PhD studies to identify collection specimens, as taxonomy is an important component of my research with this group of weevils. I also maintain a website in Spanish about insects of Guatemala and constantly consult BHL to obtain and share information about the diversity of this group in my country.”

When asked what improvements she would like to see in BHL, Orellana replies, “This is a difficult question, because I love BHL so much. There is nothing I can think of that I would change. I can always find what I am looking for with the features already included on the website.”

Orellana’s experience testifies to the importance of open access for scientific research, whether you are a researcher working with limited resources or even if you have the benefit of a large, institutional library. For researchers everywhere, BHL’s collections offer free, easy access to centuries of scientific literature and sources that may otherwise have been inaccessible.

“Even now that I am working at a larger collection, and I have access to specialized literature, there are older works that can only be obtained through BHL,” confirms Orellana.

We are proud to know that BHL is empowering biodiversity research globally. As we face unprecedented biodiversity declines—with more than a million species threatened with extinction—immediate, online access to essential literature is ever-more important, allowing scientists to conduct research more quickly and efficiently and improving our ability to respond to today’s biodiversity crisis. We look forward to continuing to work with our global community of partners to ensure that researchers like Samanta Orellana have the resources they need to study and conserve biodiversity, from the smallest weevil to the largest sequoia and everything in between.

———————

Biodiversidad Oculta: Explorando los Gorgojos Neotropicales con la ayuda de BHL

Una vista del bosque tropical de Panamá en la Provincia de Colón. Fotografía por Samanta Orellana.

En las últimas décadas del siglo XIX, una publicación monumental acerca de la biodiversidad de México y Centroamérica inició su difusión –la Biología Centrali-Americana. Publicada en 215 partes de 1879 a 1915, por los editores Frederick DuCane Godman y Osbert Salvin, el trabajo describe más de 50,000 especies y está ilustrada con más de 1,600 láminas litográficas que contienen más de 18,000 especies. Muy notoria en su tiempo, la obra aún es sumamente importante para el estudio de la biodiversidad neotropical en la actualidad, ya que contiene prácticamente toda la información conocida, hasta el momento, acerca de la flora y fauna de México y Centroamérica.

La Biologia Centrali-Americana es un recurso particularmente importante para la entomóloga Samanta Orellana, estudiante de Doctorado en Biología Evolutiva, en el laboratorio del Dr. Nico Franz, en Arizona State University (ASU) y asistente de investigación en las Colecciones Biológicas del Centro de Integración para el Conocimiento de la Biodiversidad en ASU. Orellana empezó a estudiar insectos y a trabajar en colecciones entomológicas hace más de una década, durante sus estudios de pregrado en su país natal, Guatemala.

“Para muchos grupos de insectos de Guatemala y el resto de Centroamérica, la Biología Centrali-Americana aún representa la única fuente de información disponible en la región”, asegura Orellana.

Samanta Orellana en las Colecciones Biológicas de Arizona State University. Fotografía por Kevin Cortés.

La Biologia Centrali-Americana también es importante para Orellana, ya que la llevó a descubrirBiodiversity Heritage Library (BHL), que en español podría denominarse como “Biblioteca del Patrimonio de Biodiversidad”. En 2010, mientras identificaba insectos de Guatemala, como parte de sus estudios de pregrado, Orellana empezó a utilizar la versión electrónica de la Biología Centrali-Americana,disponible en el sitio de las Bibliotecas y Archivos del Instituto Smithsonian, lo cual la dirigió hacia BHL, y no ha dejado de usarla desde entonces.

“BHL es simplemente maravillosa”, asegura Orellana. “Pienso que es una herramienta increíble para los entomólogos alrededor del mundo, especialmente para aquellos que trabajamos en colecciones de países en vías de desarrollo, usualmente sin acceso a bibliotecas o revistas especializadas, pero con mucho material por identificar”.

Una de las láminas de antríbidos de Centroamérica, incluidas en el capítulo escrito por Karl Jordan en la Biología Centrali-Americana (Insecta. Coleoptera. v.4, pt.6, 1895-1907). Compartido en BHL por las Bibliotecas y Archivos del Instituto Smithsonian.

Orellana se especializa en la familia Anthribidae (Coleoptera), comúnmente conocidos como “gorgojos de los hongos”. Sus investigaciones con estos insectos iniciaron durante sus estudios en la Universidad de Panamá, donde completó su maestría en entomología, y continúan durante su doctorado, con su tesis enfocada en la diversidad y evolución de la familia. Como parte de su investigación, Orellana recientemente publicó un catálogo de las especies de Anthribidae de Panamá.

Las colecciones de literatura de BHL han sido indispensables para permitir que Orellana trabaje con esta familia.

“BHL fue esencial para completar mi primer trabajo publicado acerca de Anthribidae, un catálogo de especies de Panamá”, comparte Orellana. “Este grupo de gorgojos ha sido pobremente estudiado en la región Neotropical, por lo que identificar las especies siempre es un reto debido a la falta de claves taxonómicas. Yo tuve que basarme mayormente en descripciones originales, disponibles en BHL, para identificar las especies que estudié, tanto en colecciones panameñas como norteamericanas. El catálogo finalizó 44 nuevos registros para el país, muchos de los cuales fue posible identificar gracias a la literatura disponible en BHL.”

Eugonus robustus Jordan, 1904 (Coleoptera: Anthribidae), de Panamá. Fotografía por Samanta Orellana.

BHL se ha vuelto una parte esencial en el proceso de investigación de Orellana, ya que utiliza este recurso al menos varias veces por semana, leyendo el contenido en línea o descargando material, por medio del generador de archivos PDF de BHL, para construir su propia colección de literatura de Anthribidae. Su característica favorita de BHL es el servicio de búsqueda de texto, que le permite encontrar nombres científicos y localidades de una manera fácil dentro de los archivos disponibles.

“Uso BHL todo el tiempo”, afirma Orellana. “Durante mis estudios de doctorado, uso BHL constantemente para identificar especímenes de colecciones, ya que la taxonomía es un componente importante de mi investigación con este grupo de gorgojos. Como una actividad adicional, administro un sitio web enfocado en insectos de Guatemala, y consulto BHL constantemente para obtener y compartir información acerca de la diversidad de este grupo en mi país.”

Al consultarle acerca de las mejoras que le gustaría ver en BHL, Orellana responde, “Esta es una pregunta difícil, porque BHL me gusta demasiado. No puedo pensar en nada que me gustaría cambiar. Siempre puedo encontrar lo que estoy buscando con las características ya incluidas en el sitio web”.

Las experiencias de Orellana son un testimonio de la importancia de libre acceso a las investigaciones científicas, así sea para científicos trabajando con recursos limitados, o para aquellos que cuentan con acceso a bibliotecas institucionales. Para investigadores de todas partes, la colección de BHL ofrece acceso gratuito y fácil, a siglos de literatura y recursos científicos que, de otra manera, serían inaccesibles.

“Aún ahora, que me encuentro trabajando en una colección grande y tengo acceso a literatura especializada, existen trabajos muy antiguos a los que puedo acceder únicamente a través de BHL,” confirma Orellana.

Estamos orgullosos de saber que BHL empodera la investigación de biodiversidad a nivel global. Mientras nos enfrentamos a la desaparición sin precedentes de la biodiversidad —con más de un millón de especies amenazadas con la extinción— el acceso inmediato a literatura en línea es más importante que nunca, permitiendo que los científicos conduzcan sus investigaciones de una manera más rápida y eficiente, mejorando la habilidad de responder a la crisis de biodiversidad actual. Esperamos continuar trabajando con nuestra comunidad de socios para asegurar que los investigadores como Samanta Orellana tengan a la mano los recursos que necesitan para estudiar y conservar la biodiversidad, desde el gorgojo más pequeño a la secoya más grande.