From species descriptions to diversity patterns: the validation of taxonomic data as a keystone for ant diversity studies reproducibility and accuracy

Abstract

Research findings in natural sciences need to be comparable and reproducible to effectively improve our understanding of ecological and behavioural patterns. In this sense, knowledge frontiers in biodiversity studies are directly tied to taxonomic research, especially in species-rich tropical regions. Here we analysed the taxonomic information available in 470 studies on Brazilian ant diversity published in the last 50 years. We aimed to quantify the proportion of studies that provide enough data to validate taxonomic identification, explore the frequency of studies that properly acknowledge their taxonomic background, and investigate the primary resources for ant identification in Brazil. We found that most studies on Brazilian ant diversity (73.6%) explicitly stated the methods used to identify their specimens. However, the proportion of papers that provide complete data for the repository institutions and vouchered specimens is vanishingly small (5.8%). Additionally, only 40.0% of the studies consistently presented taxon authorities and years of description, rarely referencing taxonomic publications correctly. In turn, the number of specialists and institutions consulted for ant identification in Brazil has increased in the last years, along with the number of studies that explicitly provide their taxonomic procedures for ant identification. Our findings highlight a shift between generations regarding the recognition of taxonomy as fundamental science, deepening our understanding of biodiversity.

Keywords: biological collections; community ecology; myrmecology; scientific reproducibility; scientometrics; vouchers.

Figure 1.

Number of ant diversity studies in Brazil published in the last 50 years recovered in our systematic literature review. The trend line is based on the arithmetic mean of the number of biodiversity studies over the past five years (n observations) using the function SMA of the R package TTR (Technical Trading Rules, [35]). The red arrow indicates 1994, when we had more than four ant diversity studies published yearly. From that year onwards, we investigated the temporal patterns of variation of the taxonomic data presented in figure 2. The periods highlighted in blue are the two moments in which we compared the frequency of mentions in the binomial proportion tests.

Figure 2.

Frequency of mentions of taxonomic data over the years obtained from ant diversity studies in Brazil. The trend lines are based on time series modelling and describe the frequency of mentions based on the arithmetic mean of the series over the past five observations using the function SMA of the R package TTR (Technical Trading Rules, [35]). From top to bottom: (a) frequency of studies providing ant species list and explicitly mentioning taxa authorities; (b) frequency of studies that provided the methods for taxonomic identification; (c) frequency of mentions for the six most cited taxonomic publications; (d) frequency of mentions for the six most cited reference collections; (e) frequency of mentions for the six main repository collections that received ant vouchers; (f) frequency of mentions of ant specialists consulted for taxonomic identification. CPDC = Centro de Pesquisas do Cacau, Ilhéus, BA; MZSP = Museu de Zoologia da Universidade de São Paulo, São Paulo, SP; INPA = Instituto Nacional de Pesquisas da Amazônia, Manaus, AM; DZUP = Coleção Entomológica Padre Jesus Santiago Moure da Universidade Federal do Paraná, Curitiba, PR; LABECOL = Laboratório de Ecologia de Comunidades da Universidade Federal de Viçosa, Viçosa, MG; LAMAT = Laboratório de Mirmecologia do Alto Tietê – Universidade de Mogi das Cruzes, Mogi das Cruzes, SP; LEIS/UFU = Laboratório de Ecologia de Insetos Sociais da Universidade Federal de Uberlândia, Uberlândia, MG; MPEG = Museu Paraense Emílio Goeldi, Belém, PA; UNISINOS = Laboratório de Genética de Insetos Sociais da Universidade do Vale do Rio dos Sinos, São Leopoldo, RS. The asterisks indicated differences detected by the binomial proportion test comparing the frequency of mentions in the two extreme periods of the time series modeling. In cases where only one asterisk was associated with the trend line, the main proportion was compared with a null proportion (zero). We added asterisks only in tests where the p-value was less than 0.05 of the significance level. The values of statistics are available in the main text.

Figure 3.

Collaboration maps in the myrmecology field based on ant diversity studies in Brazil, including international networks. The circles represent authors, identification specialists, and collections localities, while the lines represent collaborations. Collaboration directionality (i.e. the direction of collaboration from source locality to target locality) is represented by clockwise curved lines. Redder circles represent localities with high values of collaboration (defined by high in-degree values), while bluer circles represent localities with low values of collaboration (defined by low-in-degree values). (a), (c) and (e) collaborations between authors and reference collections on a global scale; (b), (d) and (f) collaborations between authors and reference collections in Brazil. The four Brazilian institutions with higher in-degree values are indicated in maps (b), (d) and (f). Abbreviations: CPDC, Coleção de Formicidae do Centro de Pesquisas do Cacau; DZUP, Coleção Entomológica Padre Jesus Santiago Moure da Universidade Federal do Paraná; INPA, Coleção de Insetos do Instituto Nacional de Pesquisas da Amazônia; MZSP, Acervo de Formicidae do Museu de Zoologia da Universidade de São Paulo.