Humans in the loop: Community science and machine learning synergies for overcoming herbarium digitization bottlenecks.

Premise: Among the slowest steps in the digitization of natural history collections is converting imaged labels into digital text. We present here a working solution to overcome this long-recognized efficiency bottleneck that leverages synergies between community science efforts and machine learning approaches. Methods: We present two new semi-automated services. The first detects and classifies typewritten, handwritten, or mixed labels from herbarium sheets. The second uses a workflow tuned for specimen labels to label text using optical character recognition (OCR). The label finder and classifier was built via humans-in-the-loop processes that utilize the community science Notes from Nature platform to develop training and validation data sets to feed into a machine learning pipeline. Results: Our results showcase a >93% success rate for finding and classifying main labels. The OCR pipeline optimizes pre-processing, multiple OCR engines, and post-processing steps, including an alignment approach borrowed from molecular systematics. This pipeline yields >4-fold reductions in errors compared to off-the-shelf open-source solutions. The OCR workflow also allows human validation using a custom Notes from Nature tool. Discussion: Our work showcases a usable set of tools for herbarium digitization including a custom-built web application that is freely accessible. Further work to better integrate these services into existing toolkits can support broad community use.

Urban forest biodiversity and cardiovascular disease: Potential health benefits from California's street trees.

Enhanced immune functioning in response to biodiversity may explain potential health benefits from exposure to green space. Using unique data on urban forest biodiversity at the zip code level for California measured from 2014 to 2019 we test whether greater diversity of street trees is associated with reduced death from cardiovascular disease. We find that urban forests with greater biodiversity measured via the Shannon Index at the genus level are associated with a lower mortality rate for heart disease and stroke. Our estimates imply that increasing the Shannon Index by one standard deviation (0.64) is associated with a decrease in the mortality rate of 21.4 per 100,000 individuals for heart disease or 13% and 7.7 per 100,000 individuals for stroke or 16%. Our estimates remain robust across several sensitivity checks. A policy simulation for tree planting in Los Angeles based on our estimates suggests that if these relationships were causal, investment in planting for a more biodiverse set of street trees would be a cost-effective way to reduce mortality related to cardiovascular disease in urban areas.

The socially parasitic ant Polyergus mexicanus has host-associated genetic population structure and related neighbouring colonies.

The genetic structure of populations can be both a cause and a consequence of ecological interactions. For parasites, genetic structure may be a consequence of preferences for host species or of mating behaviour. Conversely, genetic structure can influence where conspecific interactions among parasites lay on a spectrum from cooperation to conflict. We used microsatellite loci to characterize the genetic structure of a population of the socially parasitic dulotic (aka "slave-making") ant (Polyergus mexicanus), which is known for its host-specificity and conspecific aggression. First, we assessed whether the pattern of host species use by the parasite has influenced parasite population structure. We found that host species use was correlated with subpopulation structure, but this correlation was imperfect: some subpopulations used one host species nearly exclusively, while others used several. Second, we examined the viscosity of the parasite population by measuring the relatedness of pairs of neighbouring parasitic ant colonies at varying distances from each other. Although natural history observations of local dispersal by queens suggested the potential for viscosity, there was no strong correlation between relatedness and distance between colonies. However, 35% of colonies had a closely related neighbouring colony, indicating that kinship could potentially affect the nature of some interactions between colonies of this social parasite. Our findings confirm that ecological forces like host species selection can shape the genetic structure of parasite populations, and that such genetic structure has the potential to influence parasite-parasite interactions in social parasites via inclusive fitness.

Asymmetrical selection maintains heritable phenotypic variation between two subspecies of Monardella villosa.

PREMISE: Monardella villosa is an evolutionarily young species complex distributed across a large geographic range. Our goal was to determine whether the phenotypic difference between two subspecies of M. villosa was heritable and whether the alternative phenotypes were adaptive to their respective local habitats. METHODS: We collected seeds from 25 populations of M. villosa, 14 from subspecies franciscana, which grows closer to the coast, and 11 from subspecies villosa, which has a larger and more inland geographic distribution. We reciprocally transplanted the two subspecies into their respective habitats and compared plant germination, post-emergence survival, and growth. We used linear mixed models to quantify the effects of genotype and environment to determine whether subspecies were locally adapted and whether leaf traits that distinguish these subspecies were genetically based. RESULTS: Plants of both subspecies grown at the coastal site had significantly lower survival and biomass than the inland site. The subspecies were not locally adapted; however, the coastal subspecies franciscana did have a home site advantage. We also found that distinctive leaf morphological traits were genetically based, with high broad-sense heritability of traits. CONCLUSIONS: The two subspecies of Monardella villosa were not locally adapted to their respective habitat, but rather we found that selection for local genotypes may be stronger at the coastal site. Despite the lack of evidence for local adaptation in the strict sense, the subspecies had heritable variation in several leaf phenotypes, indicating that heterogeneous selection imposes an adaptive trade-off for leaf trichome production within this species.

Old Plants, New Tricks: Phenological Research Using Herbarium Specimens.

The timing of phenological events, such as leaf-out and flowering, strongly influence plant success and their study is vital to understanding how plants will respond to climate change. Phenological research, however, is often limited by the temporal, geographic, or phylogenetic scope of available data. Hundreds of millions of plant specimens in herbaria worldwide offer a potential solution to this problem, especially as digitization efforts drastically improve access to collections. Herbarium specimens represent snapshots of phenological events and have been reliably used to characterize phenological responses to climate. We review the current state of herbarium-based phenological research, identify potential biases and limitations in the collection, digitization, and interpretation of specimen data, and discuss future opportunities for phenological investigations using herbarium specimens.

Edaphic adaptation maintains the coexistence of two cryptic species on serpentine soils.

PREMISE OF THE STUDY: Divergent edaphic adaptation can contribute to reproductive isolation and coexistence between closely related species, yet we know little about how small-scale continuous edaphic gradients contribute to this phenomenon. We investigated edaphic adaptation between two cryptic species of California wildflower, Lasthenia californica and L. gracilis (Asteraceae), which grow in close parapatry on serpentine soil. METHODS: We reciprocally transplanted both species into the center of each species' habitat and the transition zone between species. We quantified multiple components of fitness and used aster models to predict fitness based on environmental variables. We sampled soil across the ridge throughout the growing season to document edaphic changes through time. We sampled naturally germinating seedlings to determine whether there was dispersal into the adjacent habitat and to help pinpoint the timing of any selection against migrants. KEY RESULTS: We documented within-serpentine adaptation contributing to habitat isolation between close relatives. Both species were adapted to the edaphic conditions in their native region and suffered fitness trade-offs when moved outside that region. However, observed fitness values did not perfectly match those predicted by edaphic variables alone, indicating that other factors, such as competition, also contributed to plant fitness. Soil water content and concentrations of calcium, magnesium, sodium, and potassium were likely drivers of differential fitness. Plants either had limited dispersal ability or migrants experienced early-season mortality outside their native region. CONCLUSIONS: Demonstrating that continuous habitats can support differently adapted, yet closely related, taxa is important to a broader understanding of how species are generated and maintained in nature.

The evolution of postpollination reproductive isolation in Costus.

Reproductive isolation is critical to the diversification of species. Postpollination barriers may be important in limiting gene flow between closely related species, but they are relatively cryptic and their evolution is poorly understood. Here, we review the role of postpollination reproductive isolation in plants, including the various stages at which it operates and the hypotheses for how it may evolve. We then review empirical studies in the plant genus Costus, evaluating documented postpollination barriers in light of these hypotheses. We summarize isolation due to parental style length differences and present evidence supporting the hypothesis that the differences are in part a by-product of selection on floral morphology. Additionally, we show that reduced pollen adhesion, germination, and tube growth contribute to reproductive isolation between two closely related sympatric species of Costus. Geographic variation in the strength of these crossing barriers supports the hypothesis that they evolved under reinforcement, or direct natural selection to strengthen isolation.