Search for: All records

Abstract—Podostemaceae are a clade of aquatic flowering plants that form important components of tropical river ecosystems. Species in the family exhibit highly derived growth forms and high vegetative phenotypic plasticity, both of which contribute to taxonomic confusion. The backbone phylogeny of the family remains poorly resolved, many species remain to be included in a molecular phylogenetic analysis, and the monophyly of many taxa remains to be tested. To address these issues, we assembled sequence data for 73 protein-coding plastid genes from 132 samples representing 68 species (∼23% of described species) that span the breadth of most major taxonomic, morphological, and biogeographic groups of Podostemaceae. With these data, we conducted the first plastid phylogenomic analysis of the family with broad taxon sampling. These analyses resolved most nodes with high support, including relationships not recovered in previous analyses. No evidence of widespread, well-supported conflict among individual plastid genes and the concatenated phylogeny was observed. We present new evidence that four genera (Apinagia,Marathrum,Oserya, andPodostemum), as well as four species, are not monophyletic. In particular, we show thatPodostemum flagelliformeshould not be included inPodostemumand is better recognized asDevillea flagelliformis,and thatMarathrum capillaceumis embedded withinLophogynes.l. and should be recognized asLophogyne capillacea. We also place a previously unsampled and undescribed species that likely represents a new genus. In contrast to previous studies, the neotropical generaDiamantina,Ceratolacis,Cipoia,andPodostemumare resolved as successive sister groups to a clade of all paleotropical Podostemoideae taxa sampled, suggesting a single dispersal event from the neotropics to the paleotropics in the history of the subfamily. These results provide a strong basis for improving the classification of Podostemaceae and a framework for future phylogenomic studies of the clade employing data from the nuclear genome. 

Abstract Contextualizing current increases in Northern Hemisphere temperatures is precluded by the short instrumental record of the pastca.120 years and the dearth of temperature-sensitive proxy records, particularly at lower latitudes south of <50 °N. We develop a network of 29 blue intensity chronologies derived from tree rings ofTsuga canadensis(L.) Carrière andPicea rubensSarg. trees distributed across the Mid-Atlantic and Northeast USA (MANE)—a region underrepresented by multi-centennial temperature records. We use this network to reconstruct mean March-September air temperatures back to 1461 CE based on a model that explains 62% of the instrumental temperature variance from 1901−1976 CE. Since 1998 CE, MANE summer temperatures are consistently the warmest within the context of the past 561 years exceeding the 1951−1980 mean of +1.3 °C. Cool summers across MANE were frequently volcanically forced, with significant (p<0.05) temperature departures associated with 80% of the largest tropical (n=13) and extratropical (n=15) eruptions since 1461 CE. Yet, we find that more of the identified cool events in the record were likely unforced by volcanism and either related to stochastic variability or atmospheric circulation via significant associations (p<0.05) to regional, coastal sea-surface temperatures, 500-hpa geopotential height, and 300-hpa meridional and zonal wind vectors. Expanding the MANE network to the west and south and combining it with existing temperature-sensitive proxies across North America is an important next step toward producing a gridded temperature reconstruction field for North America. 

Abstract Researchers have investigated whether machine learning (ML) may be able to resolve one of the most fundamental concerns in personnel selection, which is by helping reduce the subgroup differences (and resulting adverse impact) by race and gender in selection procedure scores. This article presents three such investigations. The findings show that the growing practice of making statistical adjustments to (nonlinear) ML algorithms to reduce subgroup differences must create predictive bias (differential prediction) as a mathematical certainty. This may reduce validity and inadvertently penalize high‐scoring racial minorities. Similarly, one approach that adjusts the ML input data only slightly reduces the subgroup differences but at the cost of slightly reduced model accuracy. Other emerging tactics involve weighting predictors to balance or find a compromise between the competing goals of reducing subgroup differences while maintaining validity, but they have been limited to two outcomes. The third investigation extends this to three outcomes (e.g., validity, subgroup differences, and cost) and presents an online tool. Collectively, the studies in this article illustrate that ML is unlikely to be able to resolve the issue of adverse impact, but it may assist in finding incremental improvements.