Search

Understanding the extremely variable, complex shape and venation characters of angiosperm leaves is one of the most challenging problems in botany. Machine learning offers opportunities to analyze large numbers of specimens, to discover novel leaf features of angiosperm clades that may have phylogenetic significance, and to use those characters to classify unknowns. Previous computer vision approaches have primarily focused on leaf identification at the species level. It remains an open question whether learning and classification are possible among major evolutionary groups such as families and orders, which usually contain hundreds to thousands of species each and exhibit many times the foliar variation of individual species. Here, we tested whether a computer vision algorithm could use a database of 7,597 leaf images from 2,001 genera to learn features of botanical families and orders, then classify novel images. The images are of cleared leaves, specimens that are chemically bleached, then stained to reveal venation. Machine learning was used to learn a codebook of visual elements representing leaf shape and venation patterns. The resulting automated system learned to classify images into families and orders with a success rate many times greater than chance. Of direct botanical interest, the responses of diagnostic features can be visualized on leaf images as heat maps, which are likely to prompt recognition and evolutionary interpretation of a wealth of novel morphological characters. With assistance from computer vision, leaves are poised to make numerous new contributions to systematic and paleobotanical studies. [ABSTRACT FROM AUTHOR]

The Paleocene-Eocene Thermal Maximum (PETM, 55.8 Ma), an abrupt global warming event linked to a transient increase in pCO2, was comparable in rate and magnitude to modern anthropogenic climate change. Here we use plant fossils from the Bighorn Basin of Wyoming to document the combined effects of temperature and pCO2 on insect herbivory. We examined 5,062 fossil leaves from five sites positioned before, during, and after the PETM (59-55.2 Ma). The amount and diversity of insect damage on angiosperm leaves, as well as the relative abundance of specialized damage, correlate with rising and falling temperature. All reach distinct maxima during the PETM, and every PETM plant species is extensively damaged and colonized by specialized herbivores. Our study suggests that increased insect herbivory is likely to be a net long-term effect of anthropogenic pCO2 increase and warming temperatures. [ABSTRACT FROM AUTHOR]

South America has some of the most diverse floras and insect faunas that are known, but its Cenozoic fossil record of insects and insect herbivory is sparse. We quantified insect feeding on 3,599 leaves from the speciose Laguna del Hunco flora (Chubut, Argentina), which dates to the early Eocene climatic optimum (52 million years ago) and compared the results with three well preserved, rich, and identically analyzed early- and middle-Eocene floras from the following sites in North America: Republic, WA; Green River, UT; and Sourdough, WV. We found significantly more damage diversity at Laguna del Hunco than in the North American floras, whether measured on bulk collections or on individual plant species, for both damage morphotypes and feeding groups. An ancient history of rich, specialized plant-insect associations on diverse plant lineages in warm climates may be a major factor contributing to the current biodiversity of South America. [ABSTRACT FROM AUTHOR]

Neotropical rainforests have a very poor fossil record, making hypotheses concerning their origins difficult to evaluate. Nevertheless, some of their most important characteristics can be preserved in the fossil record: high plant diversity, dominance by a distinctive combination of angiosperm families, a preponderance of plant species with large, smooth-margined leaves, and evidence for a high diversity of herbivorous insects. Here, we report on an ≈58-my-old flora from the Cerrejón Formation of Colombia (paleolatitude ≈5 °N) that is the earliest megafossil record of Neotropical rainforest. The flora has abundant, diverse palms and legumes and similar family composition to extant Neotropical rainforest. Three-quarters of the leaf types are large and entire- margined, indicating rainfall >2,500 mm/year and mean annual temperature >25 °C. Despite modern family composition and tropical paleoclimate, the diversity of fossil pollen and leaf samples is 60-80% that of comparable samples from extant and Quaternary Neotropical rainforest from similar climates. Insect feeding damage on Cerrejón fossil leaves, representing primary consumers, is abundant, but also of low diversity, and overwhelmingly made by generalist feeders rather than specialized herbivores. Cerrejón megafossils provide strong evidence that the same Neotropical rainforest families have characterized the biome since the Paleocene, maintaining their importance through climatic phases warmer and cooler than present. The low diversity of both plants and herbivorous insects in this Paleocene Neotropical rainforest may reflect an early stage in the diversification of the lineages that inhabit this biome, and/or a long recovery period from the terminal Cretaceous extinction. [ABSTRACT FROM AUTHOR]