Phylogenetic Systematics of Ipomopsis (Polemoniaceae): Relationships and Divergence Times Estimated from Chloroplast and Nuclear DNA Sequences

The genus Ipomopsis (Polemoniaceae) encompasses about 29 species and 24 subspecies generally divided into three sections: sect. Ipomopsis, sect. Microgilia, and sect. Phloganthea. We employed maximum likelihood and Bayesian inference of DNA sequences from the nuclear ribosomal ITS region (ITS1, 5.8S ribosomal subunit, ITS2) and the chloroplast trnL–F region (trnL intron + trnL–trnF intergenic spacer) to estimate phylogenetic relationships within this genus and its placement among other genera of Polemoniaceae. The chloroplast and combined sequences provide support for the monophyly of Ipomopsis, but only if four species previously included in the genus are removed: Ipomopsis havardii, I. sonorae, Microgilia minutiflora (= I. minutiflora), and Loeseliastrum depressum (= I. depressa). Of the three sections, two are conditionally supported as being monophyletic. Section Microgilia (with 11 species and 11 infra-specific taxa) is supported as monophyletic if I. polycladon, I. sonorae, I. depressa, and I. minutiflora (the type of the section) are removed. This clade is treated here as section Elaphocera. Section Ipomopsis is inferred to be monophyletic with the inclusion of several members of sect. Phloganthea (I. multiflora, I. pinnata, and I. polyantha). There is no support for monophyly or paraphyly of sect. Phloganthea. The Giliopsis group (I. effusa, I. guttata, and I. tenuifolia) is supported as monophyletic by both data sets, and the cpDNA sequences place it as sister to the remainder of Ipomopsis. This clade is treated as a new section, Giliopsis. Nuclear data place Giliopsis in a clade with Ipomopsis havardii, I. sonorae, Microgilia minutiflora, Loeseliastrum depressum, Eriastrum spp., Langloisia, and Dayia grantii. Using the Eocene fossil Gilisenium hueberii to calibrate the most recent common ancestor of tribe Gilieae, we estimate that Ipomopsis has its origin 28.2 ± 0. 40–39.0 ± 1.14 MYA (trnL–F and ITS, respectively). Using this same relaxed clock, the node (or coalescent event) that defines the I. aggregata complex is dated at 16.2 ± 0.38 and 27.1 ± 0.83 MYA (trnL–F and ITS, respectively). The deep divergence of the I. aggregata complex suggests that reticulation, rather than lineage sorting, is the source of conflict among phylogenetic markers used to infer the placement of I. macrosiphon.