Michelle van der Bank

Aim: Insects provide essential ecosystem services, but there is a concern that insect abundance is in rapid decline. Identifying trends in insect biodiversity is challenging because we lack baseline data for most biogeographic regions, and most insect species remain undescribed. Here, we aim to report the first nationwide inventory of insect biodiversity in South Africa. Location: Eleven National Botanical Gardens representing four major South African biomes: Savanna, Grassland, Fynbos and Albany Thicket. Time Period: Sampling from 1 February 2022 to 31 January 2023. Major Taxa Studied: Insects (Arthropods) and vascular plants. Methods: We deployed two Malaise traps per botanical garden and used DNA barcoding to identify insect specimens. We reconstructed phylogenies for insect molecular clusters (BINs) and 12,149 vascular plant species. Insect and plant community structure were compared using dissimilarity metrics and Mantel tests. Results: We generated DNA barcodes for 337,809 specimens, representing 29,919 unique molecular clusters, as species proxies. We estimated that the true richness of above ground insect biodiversity in South Africa may be closer to 100,000 species, yet 97% of barcoded taxa could not be assigned to species level. At current rates of taxonomic description, formally cataloguing this dark diversity would take more than 200 years. Estimated insect species richness varied strongly among biomes, with higher insect diversity in the Savanna biome than the floristically rich Fynbos. Turnover in plant lineages was associated with insect community turnover, particularly insect herbivores and predators. Main Conclusions: We show that plant species richness alone is a poor predictor of insect species richness, with biome‐specific factors playing an important role in shaping insect richness gradients. However, turnover in plant lineages was a strong predictor of insect compositional differences between gardens. Our study provides both critical baseline data on insect diversity in South Africa and new insights into the ecological and evolutionary drivers of insect biodiversity gradients.

Earlier phylogenetic analyses of the family Thymelaeaceae using combined plastid ( matK , rbcL , rps16, trnL‐F ) and nuclear (ITS) regions showed that the African‐Malagasy genus Gnidia was extensively polyphyletic in relation to the southern African genera Lachnaea , Passerina and Struthiola , as well as to other Southern Hemisphere genera such as Drapetes , Kelleria and Pimelea . We present here a robust and well‐sampled phylogenetic analysis of the subfamily Thymelaeoideae, with sampling in Gnidia increased to 76% of the species, to better resolve the polyphyly of the genus . Our analysis retrieved seven lineages of Gnidia nested in four independent evolutionary clades within the “Gnidia group” of the subfamily Thymelaeoideae. We analysed the options available for re‐circumscribing Gnidia as a monophyletic genus, with secondary considerations of nomenclatural stability, phylogenetic and biogeographic information, and ease of identification. The option that we present retains the traditionally recognized genera Drapetes , Kelleria , Lachnaea , Passerina , Pimelea and Struthiola through the recognition of six lineages of Gnidia at the generic level. We propose that Gnidia be retained as a monotypic genus, including only the type of the genus name G. pinifolia . We reinstate the four previously recognized genera Basutica , Craspedostoma , Gnidiopsis and Thymelina and recognize the two new genera Epichroxantha and Phidia for the remaining two lineages of continental African Gnidia for which no names were available, plus the monotypic Zacharia for an anomalous Malagasy species that was previously excluded from Lasiosiphon. A taxonomic key and descriptions are provided for the genera in the “Gnidia group”. We also provide the necessary new combinations for names of Gnidia species now transferred to other genera.

Pteronia is a large genus of shrubs or woody perennials with more than 70 species. The genus is largely restricted to southern Africa, with one species in Zambia. The last taxonomic treatment of Pteronia was published in 1917 by J. Hutchinson and E. P. Phillips. In South Africa, where most of the species occur, the genus has been identified as one of the top priority genera for taxonomic research. Its taxa are poorly known in terms of their distribution and morphological variation within and among species. Morphological and anatomical studies revealed variation in characters such as leaf size, shape and vestiture, capitula shape, involucre shape, pappus series, cypselae shape, size and vestiture, presence or absence and type of trichomes, number of vascular bundles, type of secretory structure in the mesophyll, epidermal glands, epidermal cell shape in the leaves and cypselae (in transverse section) as well as the number of ribs and type of cells associated with vascular bundles at the ribs in the cypselae (in transverse section). A detailed taxonomic treatment is presented with descriptions, nomenclature, diagnostic characters, geographical distribution and maps, ecological information, as well as an updated key to the species. A total of 76 accepted and one imperfectly known species are recognised within the genus. Several species are treated as synonyms, and all names are typified, with justification presented for the choice of lectotypes.

Spines are a major ecological innovation supporting plant defence and diversification. Spine anatomy is diverse, having arisen in multiple plant lineages from many different plant organs and parts, which may differ in relative advantages across environmental gradients. Systematic analyses of the correlates of spiny plant diversity are limited, but climate and soil properties may be important. We analysed spatial patterns of the proportional richness of spiny woody plant species (fraction of total woody species richness) and the proportional richness of species with particular spine types (fraction of richness of spiny plants) across three regions with high plant geolocational data density spanning three continents, China (Asia), South Africa (Africa), and Australia. Spiny plants accounted for 12% of woody species, but there are strong phylogenetic biases in the evolution of spiny lineages and lineages bearing different spine types. The proportion of spiny plants increased towards drier environments and higher soil clay contents, and decreased towards soils with greater total N. Species bearing different spine types appear to be distributed differently across climate and soil gradients, suggesting trade‐offs across productivity gradients, specialization for climate space, and constraints on environmental adaptability. The spatial richness of spiny plants was positively correlated with estimated historical richness of large herbivorous mammals (body mass >20 kg, diet >90% plant material), and species bearing different spine types also mostly show positive relationships with mammal richness. Plants with spines appear to be advantaged over non‐spiny species when exposed to high mammal browsing pressure in arid environments or over certain soil conditions, and species bearing different spine types are differentially advantaged across climate and soil gradients.

Societal Impact Statement: Recent shifts in flowering times are an index of, and a response to, human driven climate change. However, most information on these flowering changes is heavily skewed to the northern hemisphere. This imbalance limits our understanding of how climate change is affecting ecosystems, including the mismatches of flowering times between species, increased risks from frost or drought, and shifts in crop growing seasons. Here we show how this knowledge gap can be addressed by combining modern machine learning tools with large‐scale data collected by citizen scientists. We present a new map showing flowering patterns across South Africa, revealing complex regional differences and wide variation between plant species. Summary: Phenology—the timing of recurring life history events—is strongly linked to climate. Shifts in phenology have important implications for trophic interactions, ecosystem functioning, and community ecology. However, data on plant phenology can be time consuming to collect and current records are biased across space and taxonomy. Here, we evaluate the performance of convolutional neural networks (CNN), a machine learning tool, for classifying flowering phenology on a very large and taxonomically diverse dataset of citizen science images sourced from the iNaturalist database. Our analysis focuses on plants listed in the National Botanical Gardens (NBGs) of South Africa, a country famed for its floristic diversity (approximately 21,000 species), but poorly represented in phenological databases. We trained our CNN on a dataset of 10,000 images, representing several thousand species across >200 families, and applied the trained model to >1.8 million iNaturalist records. We were able to correctly classify images with >90% accuracy. Using metadata associated with each image, we then reconstructed the timing of peak flower production and length of the flowering season across South Africa and within each National Botanical Garden. Our analysis illustrates how machine learning tools can leverage the vast wealth of citizen science biodiversity data to describe large‐scale phenological dynamics. We suggest such approaches may be particularly valuable where data on plant phenology is currently lacking.

Insects are one of the most species-rich groups on Earth. They comprise much of animal diversity and play vital roles in ecosystems, including pollination, pest control, and decomposition. However, only a fraction of this diversity has been formally described. South Africa is recognised as one of the most biologically diverse countries globally, with an estimated 44 000 insect species. Many crops rely on insect pollinators, including canola, apples, oranges, and sunflowers. A shortage of wild pollinators currently threatens crop yields, yet our knowledge of insect diversity within South Africa is sparse. There are few taxonomic specialists relative to South Africa’s biodiversity, and the methods used for insect identification can be time-consuming and expensive. DNA barcoding provides an important research tool to accelerate insect biodiversity research. In this review, we queried the public DNA barcoding BOLD (Barcode of Life Data System) database for records of “Insecta” within South Africa, and 416 211 published records assigned to 28 239 unique BINs (Barcode Index Numbers) were returned. We identified five taxonomic orders with more BINs than known species in southern Africa (Hymenoptera, Diptera, Thysanoptera, Plecoptera, and Strepsiptera). Most of the barcoded records were derived from Malaise trap sampling in Gauteng, Mpumalanga and Limpopo, while the rest of South Africa remains poorly sampled. We suggest that there is a need for a comprehensive national sampling effort alongside increased investment in taxonomic expertise to generate critical baseline data on insect biodiversity before species are lost to extinction.

The Meliaceae are broadly distributed worldwide, with about 50 genera and over 1400 species. There are 11 genera in South Africa, with 13 indigenous and three naturalized species. Considering the diversity of the indigenous species of this family in South Africa and the lack of recent studies encompassing these species, a taxonomic revision of the South African indigenous species of Meliaceae is presented here. Phylogenetic analysis, anatomical data, herbarium collections, and online data sources were used in this study. The results confirm the monophyly of Melioideae and Swietenioideae. The incongruence of Turraea previously reported was resolved in this study. Most representative genera of South African Meliaceae were recovered monophyletic with strong support. However, multiple samplings of species and including more markers could provide a better understanding of the relationships among South African species of Meliaceae. The review of the taxonomy of the South African Meliaceae, and especially the study of diagnostic characters and hitherto recorded natural distributions, have value in providing an up-to-date inventory of the indigenous genera and species and an easy means of identifying the taxa. Anatomical characters may be of systematic value to explore higher-level relationships in the family. This study is a contribution to tropical botany and to a more comprehensive database for the Meliaceae.

Summary: Phenology — the timing of recurring life history events—is strongly linked to climate. Shifts in phenology have important implications for trophic interactions, ecosystem functioning and community ecology. However, data on plant phenology can be time consuming to collect and current records are biased across space and taxonomy. Here, we explore the performance of Convolutional Neural Networks (CNN) for classifying flowering phenology on a very large and taxonomically diverse dataset of citizen science images. We analyse >1.8 million iNaturalist records for plants listed in the National Botanical Gardens within South Africa, a country famed for its floristic diversity (∼21,000 species) but poorly represented in phenological databases. We were able to correctly classify images with >90% accuracy. Using metadata associated with each image, we then reconstructed the timing of peak flower production and length of the flowering season for the 6,986 species with >5 iNaturalist records. Our analysis illustrates how machine learning tools can leverage the vast wealth of citizen science biodiversity data to describe large-scale phenological dynamics. We suggest such approaches may be particularly valuable where data on plant phenology is currently lacking.