Phylogenetic Relationships of Orb-weaving Spiders from Species to Superfamily Open Access
Downloadable ContentDownload PDF
Spiders are a speciose group of arachnids that are a dominant predator of terrestrial arthropods. Among these, perhaps the most iconic are the orb-weavers, which spin webs from silk produced in abdominal glands. Despite steps forward in spider phylogenetics, there are still many undescribed species and a number of difficult problems in understanding their phylogeny that must be remedied before one can understand evolution of morphology and behavior that define these fascinating creatures. In this work, I focus on two families of orb-weavers: Tetragnathidae and Araneidae. First, I discuss the tetragnathid subfamily Metainae, which has not been previously revised, has low support in previous works, and questionable inclusion of certain genera. My target gene analyses on an increased taxon sampling of metaines found good support for the subfamily for the first time, and suggest inclusion of a new genus, Zhinu, from Taiwan. After synonymy of Prolochus and Menosira, I determined there are four genera in Metainae: Zhinu, Meta, Metellina, and Dolichognatha. I also describe a new species of Orsinome with exaggerated genitalic morphology. Second, I revise the Australasian leaf-curling araneid genera Phonognatha and Deliochus using molecular and morphological characters. I place these genera in the wider context of Araneidae, which lacks a family-level treatment. Based on resulting phylogenies, I found evidence for a third genus, Artifex, and monophyly of Araneidae following recent taxonomy changes. Furthermore, this phylogeny allowed me to explore two comparative questions relating to these genera: evolution of leaf retreats in orb-webs and biogeography. I found evolution and subsequent loss of an integrated leaf retreat in these genera, and colonization of New Caledonia consistent with geology of the island. Finally, I used next next-generation transcriptome-based techniques to explore the phylogeny of Araneidae, which has been limited by the resolution power of current markers and morphology. Using a diverse sampling of 19 araneids, I examined the effects of orthology assessment methods and gene occupancy/missing data on the resulting topologies. The results showed broad congruence across most analytical treatments regardless of orthology method, occupancy, or tree inference method, with more genes (and more missing data) to be the best predictor of high node supports. I also found little overlap between transcripts used by orthology programs, suggesting good signal in the data.