Olearia, Moench, 1802

Saldivia, Patricio, Wood, Kenneth R., Orlovich, David A. & Lord, Janice M., 2020, Pacifigeron indivisus (Asteraceae: Astereae), a new species endemic to Rapa, Austral Islands, and a new delimitation of the Celmisia group, Phytotaxa 442 (4), pp. 239-266 : 250-252

The polyphyletic Olearia : its infrageneric taxonomy and its representation within the Celmisia group

The classification of Olearia requires special attention in the context of both the definition of the Celmisia group and the classification of the Australasian Astereae . As currently circumscribed, Olearia is the largest Asteraceae genus in Australasia ( Lander 1994, Cross et al. 2002, Messina et al. 2014, Hind & Johns 2014). Olearia is composed of 174 accepted species distributed in Australia , New Zealand, and New Guinea with no species naturally shared between countries ( Koster 1966, Hind & Johns 2014, de Salas & Baker 2018, Schönberger et al. 2019, CHAH 2019). Since the taxonomy at the specific level has unevenly been studied in different countries or groups, and no complete taxonomic work has been completed for the genus so far, the exact number of species remains unclear.

Nesom (1994b) stated that one of the most significant remaining problems in Astereae classification is the infra-generic organization of Olearia and its relation to other Australian Hinterhuberinae . The same author ( Nesom 1992, 1993a, 1993b, 1994b) pointed out that Olearia is not monophyletic. Based mainly on chromosome number, he stressed the idea that the highly polyploid species should be removed from Olearia and rather be part of the Celmisia group. Subsequent studies based on DNA sequence data ( Cross et al. 2002, Brouillet et al. 2009) have demonstrated that Nesom’s finding was mainly correct, showing that although Olearia is highly polyphyletic (at least eight well-supported clades not directly related to each other, Cross et al. 2002), its constituent taxa fall within two main clades distantly related in the tribal phylogeny ( Brouillet et al. 2009). Clade I in Cross et al. (2002), which corresponds to the Australasian lineage in Brouillet et al. (2009), has a derived position in the tribal phylogeny and contains mostly diploid species (see Watanabe et al. 1996). Clade II in Cross et al. 2002, named as the New Zealand clade by Brouillet et al. (2009) and equivalent to the Celmisia group, shows a relatively basal position in the tribal phylogeny, and the Olearia species contained, along with the other associated taxa, possess high polyploidy levels ( Hair 1980, Beuzenberger & Hair 1984).

Archer (1860), based on the trichome morphology of the leaves, proposed an infra-generic taxonomy for the Australian and New Zealand species that at that time were placed in the genera Olearia and Eurybia ( Cassini 1818: 166) Gray (1821: 464) . He proposed five sections as follows: Dicetrotriche, centrally-attached or divaricately forked (T- or Y-shaped) trichomes; Asterotriche , stellate trichomes; Eriotriche, densely intricate woolly trichomes; Adenotriche, glabrous, usually glutinous; and Merismotriche, glabrous, glandular-pubescent, hirsute or glutinous. However, his proposal was not validly published because sectional epithets were not clearly associated with the name of a genus ( Messina et al. 2014; see ICN Art. 35.2, Turland et al. 2018). Bentham (1867) validated Archer’s sectional proposal when synonymized Eurybia with Olearia , and thus attributing the sections to a single genus. However, section Dicerotriche was not validly published because it includes the type of the genus, Olearia tomentosa , and therefore, an autonym was required (ICN Art. 22.1, 22.2; Turland et al. 2018). Concerning Eurybia, Nesom (1994d) regarded it as a genus with 28 species from the USA, Canada to Alaska, and north-eastern Asia ( Nesom & Robinson 2007) and commented: “[…] the long association of Eurybia with Olearia is unjustified”.

Most of the subsequent authors dealing with Olearia in Australia have followed Bentham’s (1867) infra-generic taxonomy (e.g. Domin 1929, Lander 1989b, 1990, 1991, 2008a, 2008b, Messina et al. 2014). In contrast, authors working on New Zealand flora (e.g. Hooker 1867, Kirk 1899, Cheeseman 1906, 1925, Allan 1961) have not followed it, although some modern authors such Heenan & Cameron (2002) and Heenan (2005) have suggested the placement of some New Zealand Olearia species in section Dicerotriche (despite not being a validly published section). With regard to the New Guinea species, most of the authors ( Mattfeld 1937, Koster 1966, van Royen 1983) have rejected Archer’s sectional taxonomy arguing that in one species, Olearia heterotricha Mattfeld (1937: 254) , the indumentum of the leaves shows a combination of three types of trichomes that define different sections.

Two other groups have been proposed within Olearia after the original classification by Archer (1860) and Bentham (1867): the macrocephalous olearias and section Divaricaster . Archer (1860) commented: “The best natural group, however, is that composed of O. colensoi and the three following species [ O. oporina , O. angustifolia , O. lyallii ], which would form a fair genus, and have large solitary or few heads, long villous achenia, subduplex rufescent pappus, and large, flat, denticulate, woolly-haired leaves”. Kirk (1891) coined the name “macrocephalous olearias” for this group adding the two Chatham Islands endemics Olearia chathamica Kirk (1891: 444) and O. semidentata , and treated them informally as a section (i.e. he mentioned it in the text but did not formally describe it). Drury (1968) proposed a closer relationship between the macrocephalous olearias and the subantarctic genus Pleurophyllum , rather than with Olearia .

Heads (1998) erected section Divaricaster , the sixth section of Olearia to accommodate a group of New Zealand species characterized by their divaricate architecture, that is, plants developing short shoots or brachyblasts bearing leaves and capitula, long shoots with apical abortion, and at least a portion of their branches growing plagiotropically or even geotropically. This remarkable architecture has recently been referred to as “Philipson’s model” ( Heads 2019). Heenan & Molloy (2004) added one species to section Divaricaster , and Heenan et al. (2008) described Olearia telmatica Heenan & de Lange in Heenan et al. (2008: 575) and expanded the section to include two species endemic to the Chatham Islands, O. traversiorum ( Mueller 1864: 19) Hooker (1867: 731) and O. telmatica . The latter two species have no divaricate architecture, which raises questions about the morphological definition of the section.

However, the question of what comprises the complete assemblage of Olearia species belonging to the Celmisia group is still unclear. Based on our results, on previous DNA-based phylogenetic works ( Cross et al. 2002, Brouillet et al. 2009, Wagstaff et al. 2011, Messina et al. 2014), and on extensive taxonomic literature referring to Olearia (e.g. Archer 1860, Bentham 1867, Kirk 1891, Koster 1966, Drury 1968, Lander 1975, 1989b, 1989c, 1990, 1991, 2008a, 2008 b, Heads 1998, Cross et al. 2002, Heenan & Cameron 2002, Heenan & Molly 2004, Heenan 2005, Heenan et al. 2008, Lang 2008, Hind & Johns 2014, Messina et al. 2014, Walsh 2014, Bean & Mathieson 2015, Bean & Jobson 2017, Messina & Walsh 2019), the following points are made:

– Messina et al. (2014) redefined section Asterotriche as endemic to Australia , part of the Australasian linages ( Brouillet et al. 2009), and sister to the type species Olearia tomentosa . However, they did not study the morphology of the New Zealand species formerly placed in Asterotriche ( Archer 1860, Cross et al. 2002) nor the New Guinea species with stellate or lepidote-like trichomes. Therefore, the question of how to differentiate Asterotriche sensu stricto from the New Zealand and New Guinea species with stellate-like trichomes has not been answered. Three out of eight New Zealand species with stellate trichomes have been analyzed with molecular phylogenetics and were shown to be monophyletic within the Celmisia group ( Cross et al. 2002).

– Sections Adenotriche, Eriotriche (excluding the macrocephalous olearias), and Merismotriche are endemic to mainland Australia , Tasmania, and the Lord Howe Island. Molecular evidence ( Cross et al. 2002, Brouillet et al. 2009) based on the analysis of 40, 25, and 30% of their respective species, indicates that these sections are artificial groups with members variously related to other Australasian taxa.

– We included two of the 19 Olearia species from New Guinea in our phylogenetic analyses ( Fig. 2 ). Both species form a sister clade to some of the New Zealand species with T- or Y-shaped trichomes. Nevertheless, more sampling is needed to confirm both the monophyly of this group and whether all its members belong to the Celmisia group.

– Species formerly placed in section Dicerotriche make up the largest group with 39 species, 25 from Australia (mainland, Tasmania, and Lord Howe Island), and 14 from New Zealand, of which 48 and 29% of the respective species have been studied using molecular data. All of the analyzed species belong to the Celmisia group except for Olearia tomentosa , which is the type species. The relationships among species of this group are not clear. Our phylogenetic analyses ( Fig. 2 ) show that they (excluding Olearia tomentosa ) do not form a monophyletic group.

– The macrocephalous olearias and section Divaricaster (including O. traversiorum and O. telmatica ; Heenan et al. 2008) are endemic to New Zealand and belong to the Celmisia group. However, while all the macrocephalous olearia species have been analyzed with molecular data and shown to have closer relationships with the subantarctic Pleurophyllum and Damnamenia rather than with the rest of the Olearia species ( Wagstaff et al. 2011), the sampling within section Divaricaster is still insufficient.

– In our analyses, Olearia species in the Celmisia group form several well-supported clades that mostly reflect previously recognized groups. However, more sampling and research is needed in order to establish a topology firmly supporting the exclusion of these taxa from Olearia .

Figures

FIGURE 2. Bayesian inference phylogram of concatenated internal transcribed spacer (ITS) and external transcribed spacer (ETS) regions sequences. Posterior probability values are indicated above branches and bootstrap (%) values from the maximum parsimony analysis are indicated below the branches. Distributions are shown in colors. Olearia lyallii is also present in Stewart Island, New Zealand. Laennecia sophiifolia is also distributed in Central America, Mexico, and the southwestern USA.