Abstract
The Celmisia group (i.e. an informal taxonomic proposal) belongs to the tribe Astereae of the Asteraceae family. Prior to the work presented in this dissertation, the delimitation of the Celmisia group, based on insights from morphology, chromosome number data, and geographic distributions, included the following genera: Achnophora (1 sp.), Celmisia (68 spp.), Damnamenia (1 sp.), Olearia (ca 180 spp.) in part, Pachystegia (3 spp.), and Pleurophyllum (3 spp.) with an Australasian distribution, and Pacifigeron (1 sp.), an endemic genus to Rapa Island in French Polynesia. However, only a few species of most of these genera have been included in DNA-based phylogenetic studies, and the morphology of most of the group has not been studied in-depth. This situation makes its delimitation and generic taxonomy unclear, warranting further research. The overarching aim of this dissertation was to study the taxonomy of the Celmisia group by using molecular phylogenetic and morphological analyses in order to test its delimitation and generic classification and to propose taxonomic changes when appropriate.
I performed phylogenetic analyses (Bayesian Inference and Maximum Parsimony) based on nuclear ribosomal DNA sequence data (ETS and ITS regions) of 107 species (69% of the Celmisia group). Likewise, I carried out morphological analyses (Non-metric multidimensional scaling and Ward’s clustering method) based on 83 characters scored from 587 herbarium specimens representing 148 species (95% of the Celmisia group). Based on these analyses, I proposed a new circumscription of the Celmisia group. Following the principles of monophyly (i.e. recognition of monophyletic groups), diagnosability (i.e. recognition of morphologically well-defined groups), and nomenclatural stability (e.g. disrupt present classifications as little as possible), I proposed a new generic and infrageneric taxonomic classification for the group. Further, based on the study of herbarium specimens and fieldwork, I undertook taxonomic studies at a specific level for some of its members.
I proposed that the Celmisia group should be redefined to exclude Achnophora, Pacifigeron, and part of the genus Olearia. I found Achnophora to be part of a large clade containing Australian, Hawaiian, and South American taxa, although without clear relationships. Pacifigeron, for which I also described a new species to science, is related to South American Andean genera. Regarding Olearia, as the type species, O. tomentosa, does not belong to the Celmisia group, only about half of the currently accepted Olearia species (ca 84 spp.) belong to the Celmisia group.
The Celmisia group, as redefined here, includes ca 156 species distributed in New Zealand, Australia, and New Guinea, belonging to Celmisia, Damnamenia, Olearia in part, Pachystegia, and Pleurophyllum. Overall, my results indicate that the Celmisia group is divided into nine main clades, with most of them well-defined by the overall morphological similarity of its members and/or diagnostic character states; however, these nine clades do not completely align with the aforementioned genera. Celmisia species fall into two non-sister clades, whereas smaller genera such as Pachystegia and Pleurophyllum are monophyletic. The monotypic Damnamenia is confirmed as an independent lineage not nested within another genus. The Olearia species belonging to the Celmisia group species do not form a monophyletic group and are instead divided into several clades not directly related to each other. The new taxonomic classification proposed here includes the recognition of eight genera. A new generic name, Lignosia Saldivia, is proposed for the species formerly placed in Celmisia subgenera Caespitosae, Glandulosae, and Lignosae since they form a monophyletic group not sister to the clade containing subgenera Celmisia and Pelliculatae (i.e. Celmisia sensu stricto). Damnamenia, Pachystegia, and Pleurophyllum are retained as distinctive genera. Olearia species belonging to the Celmisia group (ca 84) are included in the reinstated Shawia and Steiractis, and the new name Macrolearia Saldivia. Additionally, Steiractis is divided into six sections. For a clade including five Olearia species endemic to the North Island of New Zealand, here recognized as the Olearia furfuracea complex, I did not propose a new generic affiliation because the molecular analyses did not show clear phylogenetic relationships of this clade within the Celmisia group.
I undertook a taxonomic revision of Celmisia subgenera Caespitosae, Glandulosae, and Lignosae (i.e. Lignosia), for which I accepted 26 species, two subspecies, and two varieties. This revision was based on the study of 2701 herbarium specimens plus extensive fieldwork along the South Island of New Zealand, where 25 of the 26 accepted species are distributed.
Finally, I studied the taxonomy of the Pleurophyllum clade (i.e. Pleurophyllum, Damnamenia, and the six macrocephalous Olearia species), focusing on the taxonomic alternatives to excluding the macrocephalous Olearia species from Olearia. The molecular phylogenetic and morphological (431 herbarium specimens studied) analyses indicate the maintenance of Pleurophyllum and Damnamenia, and the recognition Macroleria as a new generic name for the macrocephalous Olearia species is the most reasonable taxonomic choice.
This dissertation made a significant and novel contribution to the understanding of Asterae in the Southern Hemisphere by demonstrating that previous hypotheses regarding the classification of both the Celmisia group and its constituent genera needed reevaluation. In the same way, the findings of this thesis indicate that taxonomy best benefits when it is approached from a broad perspective, in which the focus of a study should never be completely isolated or dissociated from the higher level lineage to which it belongs.