An image dataset of cleared, x-rayed, and fossil leaves vetted to plant family for human and machine learning

Osmoxylon-like fossils from early Eocene South America: West Gondwana-Malesia connections in Araliaceae

PREMISE

Araliaceae comprise a moderately diverse, predominantly tropical angiosperm family with a limited fossil record. Gondwanan history of Araliaceae is hypothesized in the literature, but no fossils have previously been reported from the former supercontinent.

METHODS

I describe large (to macrophyll size), palmately compound-lobed leaf fossils and an isolated umbellate infructescence from the early Eocene (52 Ma), late-Gondwanan paleorainforest flora at Laguna del Hunco in Argentine Patagonia.

RESULTS

The leaf fossils are assigned to Caffapanax canessae gen. et sp. nov. (Araliaceae). Comparable living species belong to five genera that are primarily distributed from Malesia to South China. The most similar genus is Osmoxylon, which is centered in east Malesia and includes numerous threatened species. The infructescence is assigned to Davidsaralia christophae gen. et sp. nov. (Araliaceae) and is also comparable to Osmoxylon.

CONCLUSIONS

The Caffapanax leaves and Davidsaralia infructescence, potentially representing the same source taxon, are the oldest araliaceous macrofossils and provide direct evidence of Gondwanan history in the family. The new fossils and their large leaves enrich the well-established biogeographic and climatic affinities of the fossil assemblage with imperiled Indo-Pacific, everwet tropical rainforests. The fossils most likely represent shrubs or small trees, adding to the rich record of understory vegetation recovered from Laguna del Hunco.

Fossils of an endangered, endemic, giant dipterocarp species open a historical portal into Borneo's vanishing rainforests

PREMISE

Asia's wet tropical forests face a severe biodiversity crisis, but few fossils record their evolutionary history. We recently discovered in situ cuticles on fossil leaves, attributed to the giant rainforest tree Dryobalanops of the iconic Dipterocarpaceae family, from the Plio-Pleistocene of Brunei Darussalam (northern Borneo). Studying these specimens allowed us to validate the generic identification and delineate affinities to living dipterocarp species.

METHODS

We compared the leaf cuticles and architecture of these fossil leaves with the seven living Dryobalanops species.

RESULTS

The cuticular features shared between the fossils and extant Dryobalanops, including the presence of giant stomata on veins, confirm their generic placement. The leaf characters are identical to those of D. rappa, an IUCN red-listed Endangered, northern Borneo endemic. The D. rappa monodominance at the fossil site, along with Dipterocarpus spp. leaf fossils, indicates a dipterocarp-dominated forest near the mangrove-swamp depocenter, most likely in an adjacent peatland.

CONCLUSIONS

The Dryobalanops rappa fossils are the first fossil evidence of a living endangered tropical tree species and show how analysis of in situ cuticles can help illuminate the poorly known floristic history of the Asian tropics. This discovery highlights new potential for fossils to inform heritage values and paleoconservation in Southeast Asia.

Opportunities and Challenges in Applying AI to Evolutionary Morphology

Artificial intelligence (AI) is poised to revolutionize many aspects of science, including the study of evolutionary morphology. While classical AI methods such as principal component analysis and cluster analysis have been commonplace in the study of evolutionary morphology for decades, recent years have seen increasing application of deep learning to ecology and evolutionary biology. As digitized specimen databases become increasingly prevalent and openly available, AI is offering vast new potential to circumvent long-standing barriers to rapid, big data analysis of phenotypes. Here, we review the current state of AI methods available for the study of evolutionary morphology, which are most developed in the area of data acquisition and processing. We introduce the main available AI techniques, categorizing them into 3 stages based on their order of appearance: (1) machine learning, (2) deep learning, and (3) the most recent advancements in large-scale models and multimodal learning. Next, we present case studies of existing approaches using AI for evolutionary morphology, including image capture and segmentation, feature recognition, morphometrics, and phylogenetics. We then discuss the prospectus for near-term advances in specific areas of inquiry within this field, including the potential of new AI methods that have not yet been applied to the study of morphological evolution. In particular, we note key areas where AI remains underutilized and could be used to enhance studies of evolutionary morphology. This combination of current methods and potential developments has the capacity to transform the evolutionary analysis of the organismal phenotype into evolutionary phenomics, leading to an era of "big data" that aligns the study of phenotypes with genomics and other areas of bioinformatics.

Network feature-based phenotyping of leaf venation robustly reconstructs the latent space

Despite substantial variation in leaf vein architectures among angiosperms, a typical hierarchical network pattern is shared within clades. Functional demands (e.g., hydraulic conductivity, transpiration efficiency, and tolerance to damage and blockage) constrain the network structure of leaf venation, generating a biased distribution in the morphospace. Although network structures and their diversity are crucial for understanding angiosperm venation, previous studies have relied on simple morphological measurements (e.g., length, diameter, branching angles, and areole area) and their derived statistics to quantify phenotypes. To better understand the morphological diversities and constraints on leaf vein networks, we developed a simple, high-throughput phenotyping workflow for the quantification of vein networks and identified leaf venation-specific morphospace patterns. The proposed method involves four processes: leaf image acquisition using a feasible system, leaf vein segmentation based on a deep neural network model, network extraction as an undirected graph, and network feature calculation. To demonstrate the proposed method, we applied it to images of non-chemically treated leaves of five species for classification based on network features alone, with an accuracy of 90.6%. By dimensionality reduction, a one-dimensional morphospace, along which venation shows variation in loopiness, was identified for both untreated and cleared leaf images. Because the one-dimensional distribution patterns align with the Pareto front that optimizes transport efficiency, construction cost, and robustness to damage, as predicted by the earlier theoretical study, our findings suggested that venation patterns are determined by a functional trade-off. The proposed network feature-based method is a useful morphological descriptor, providing a quantitative representation of the topological aspects of venation and enabling inverse mapping to leaf vein structures. Accordingly, our approach is promising for analyses of the functional and structural properties of veins.

The first Gondwanan Euphorbiaceae fossils reset the biogeographic history of the Macaranga-Mallotus clade

PREMISE

The spurge family Euphorbiaceae is prominent in tropical rainforests worldwide, particularly in Asia. There is little consensus on the biogeographic origins of the family or its principal lineages. No confirmed spurge macrofossils have come from Gondwana.

METHODS

We describe the first Gondwanan macrofossils of Euphorbiaceae, represented by two infructescences and associated peltate leaves from the early Eocene (52 Myr ago [Ma]) Laguna del Hunco site in Chubut, Argentina.

RESULTS

The infructescences are panicles bearing tiny, pedicellate, spineless capsular fruits with two locules, two axile lenticular seeds, and two unbranched, plumose stigmas. The fossils' character combination only occurs today in some species of the Macaranga-Mallotus clade (MMC; Euphorbiaceae), a widespread Old-World understory group often thought to have tropical Asian origins. The associated leaves are consistent with extant Macaranga.

CONCLUSIONS

The new fossils are the oldest known for the MMC, demonstrating its Gondwanan history and marking its divergence by at least 52 Ma. This discovery makes an Asian origin of the MMC unlikely because immense oceanic distances separated Asia and South America 52 Ma. The only other MMC reproductive fossils so far known are also from the southern hemisphere (early Miocene, southern New Zealand), far from the Asian tropics. The MMC, along with many other Gondwanan survivors, most likely entered Asia during the Neogene Sahul-Sunda collision. Our discovery adds to a substantial series of well-dated, well-preserved fossils from one undersampled region, Patagonia, that have changed our understanding of plant biogeographic history.

Decoding family-level features for modern and fossil leaves from computer-vision heat maps

PREMISE

Angiosperm leaves present a classic identification problem due to their morphological complexity. Computer-vision algorithms can identify diagnostic regions in images, and heat map outputs illustrate those regions for identification, providing novel insights through visual feedback. We investigate the potential of analyzing leaf heat maps to reveal novel, human-friendly botanical information with applications for extant- and fossil-leaf identification.

METHODS

We developed a manual scoring system for hotspot locations on published computer-vision heat maps of cleared leaves that showed diagnostic regions for family identification. Heat maps of 3114 cleared leaves of 930 genera in 14 angiosperm families were analyzed. The top-5 and top-1 hotspot regions of highest diagnostic value were scored for 21 leaf locations. The resulting data were viewed using box plots and analyzed using cluster and principal component analyses. We manually identified similar features in fossil leaves to informally demonstrate potential fossil applications.

RESULTS

The method successfully mapped machine strategy using standard botanical language, and distinctive patterns emerged for each family. Hotspots were concentrated on secondary veins (Salicaceae, Myrtaceae, Anacardiaceae), tooth apices (Betulaceae, Rosaceae), and on the little-studied margins of untoothed leaves (Rubiaceae, Annonaceae, Ericaceae). Similar features drove the results from multivariate analyses. The results echo many traditional observations, while also showing that most diagnostic leaf features remain undescribed.

CONCLUSIONS

Machine-derived heat maps that initially appear to be dominated by noise can be translated into human-interpretable knowledge, highlighting paths forward for botanists and paleobotanists to discover new diagnostic botanical characters.

First fossil-leaf floras from Brunei Darussalam show dipterocarp dominance in Borneo by the Pliocene

The Malay Archipelago is one of the most biodiverse regions on Earth, but it suffers high extinction risks due to severe anthropogenic pressures. Paleobotanical knowledge provides baselines for the conservation of living analogs and improved understanding of vegetation, biogeography, and paleoenvironments through time. The Malesian bioregion is well studied palynologically, but there have been very few investigations of Cenozoic paleobotany (plant macrofossils) in a century or more. We report the first paleobotanical survey of Brunei Darussalam, a sultanate on the north coast of Borneo that still preserves the majority of its extraordinarily diverse, old-growth tropical rainforests. We discovered abundant compression floras dominated by angiosperm leaves at two sites of probable Pliocene age: Berakas Beach, in the Liang Formation, and Kampong Lugu, in an undescribed stratigraphic unit. Both sites also yielded rich palynofloral assemblages from the macrofossil-bearing beds, indicating lowland fern-dominated swamp (Berakas Beach) and mangrove swamp (Kampong Lugu) depositional environments. Fern spores from at least nine families dominate both palynological assemblages, along with abundant fungal and freshwater algal remains, rare marine microplankton, at least four mangrove genera, and a diverse rainforest tree and liana contribution (at least 19 families) with scarce pollen of Dipterocarpaceae, today's dominant regional life form. Compressed leaves and rare reproductive material represent influx to the depocenters from the adjacent coastal rainforests. Although only about 40% of specimens preserve informative details, we can distinguish 23 leaf and two reproductive morphotypes among the two sites. Dipterocarps are by far the most abundant group in both compression assemblages, providing rare, localized evidence for dipterocarp-dominated lowland rainforests in the Malay Archipelago before the Pleistocene. The dipterocarp fossils include winged Shorea fruits, at least two species of plicate Dipterocarpus leaves, and very common Dryobalanops leaves. We attribute additional leaf taxa to Rhamnaceae (Ziziphus), Melastomataceae, and Araceae (Rhaphidophora), all rare or new fossil records for the region. The dipterocarp leaf dominance contrasts sharply with the family's <1% representation in the palynofloras from the same strata. This result directly demonstrates that dipterocarp pollen is prone to strong taphonomic filtering and underscores the importance of macrofossils for quantifying the timing of the dipterocarps' rise to dominance in the region. Our work shows that complex coastal rainforests dominated by dipterocarps, adjacent to swamps and mangroves and otherwise similar to modern ecosystems, have existed in Borneo for at least 4-5 million years. Our findings add historical impetus for the conservation of these gravely imperiled and extremely biodiverse ecosystems.