Transfer of Cotula alpina to the genus Leptinella (Asteraceae: Anthemideae)

Introduction

Tribe Anthemideae of the Asteraceae family comprises an estimated 1800 species (Oberprieler et al. 2007) with a predominantly Old World extra-tropical distribution (Oberprieler et al. 2009). They are herbs or shrubs with frequently dentate to deeply divided leaves, lack pappus bristles, and are often aromatic. Because of the latter trait, members of the tribe have variously been used as medicinal or tea plants (e.g. chamomile), flavour or spices (e.g. absinth, tarragon–estragon), and for scent (cotton lavender) (Simpson 2009). Other species are popular ornamentals (e.g. chrysanthemums, marguerites), and some are significant weeds such as the ox-eye daisy Leucanthemum vulgare Lam. (Stutz et al. 2021).

Despite their diversity at the global level, Anthemideae is poorly represented in the Australian flora. Only the following nine species from two genera are indigenous: Cotula alpina (Hook.f.) Hook.f., an alpine to montane species of south-eastern mainland Australia and Tasmania (Fig. 1a); C. australis (Sieber ex Spreng.) Hook.f., which is widespread (Fig. 1e); C. cotuloides (Steetz) Druce, occurring in saline areas and swamps of south-western Western Australia (Fig. 1f); C. vulgaris Levyns, growing in wet saline areas of South Australia, Victoria, and Tasmania; Leptinella drummondii (Benth.) D.G.Lloyd & C.J.Webb of south-western Western Australia; L. filicula (Hook.f.) Hook.f. found in wet forests of south-eastern Australia and Tasmania (Fig. 1b); L. longipes Hook.f., of wet saline areas ranging from South Australia to Queensland and Tasmania (Fig. 1d); L. plumosa Hook.f. of Macquarie Island; and L. reptans (Benth.) D.G.Lloyd & C.J.Webb of South Australia, New South Wales, Victoria, and Tasmania (Fig. 1c). Both these genera belong to the Cotulinae, a subtribe of 10 genera and 137 species with a southern hemisphere distribution (Oberprieler et al. 2009).

Fig. 1.

Six of the nine Australian indigenous representatives of Asteraceae subtribe Cotulinae and state or territory where the photograph was taken. (a) Cotula alpina (Hook.f.) Hook.f., New South Wales. (b) Leptinella filicula (Hook.f.) Hook.f., New South Wales. (c) Leptinella reptans (Benth.) D.G.Lloyd & C.J.Webb, Tasmania. (d) Leptinella longipes Hook.f., New South Wales. (e) Cotula australis (Sieber ex Spreng.) Hook.f., Australian Capital Territory. (f) Cotula cotuloides (Steetz) Druce, Western Australia. Note procumbent, stoloniferous, and rosette-and-scape ‘Leptinella’ habit in a–d and upright to ascending ‘Cotula’ habit with branched, leafy aerial stems in e, f. All photos were taken by the first author.

Over the past two decades, the generic placement of Cotula alpina has come under scrutiny. The species is unusual in Cotula L., whose members are generally erect to ascending herbs with branching, leafy aerial stems (Fig. 1e, f). Instead, C. alpina shares with members of Leptinella Cass. a prostrate, stoloniferous (or sometimes long-rhizomatous), rosette-and-scape growth form (Fig. 1a–d). In overall habit, C. alpina is very similar to L. filicula (Fig. 1a, b), generally distinguished most easily by plants of the former being entirely glabrous, and the two species are also geographically close. However, like other Cotula and unlike Leptinella, C. alpina does not produce corollas in female florets, a character that has traditionally been considered defining of Cotula (Lloyd and Webb 1987; Thompson 2007).

A comprehensive molecular phylogenetic study of Leptinella using the nuclear ribosomal internal transcribed spacer region (ITS), chloroplast psbA–trnH, and trnC–petN regions found Cotula alpina nested within the former genus (Himmelreich et al. 2012). However, the authors provided the caveat that its base chromosome number of x = 9 would match those known for Cotula better than the x = 13 of Leptinella (Oberprieler et al. 2009). Another potential caveat is that the ITS sequence of the species in the study (Genbank accession HE860701), which evolves faster than the chloroplast regions and thus provides the largest number of informative characters, comprises only the first half of the entire region (ITS1), in contrast to other ITS sequences from the same study. The taxonomic affiliation of C. alpina therefore remains unresolved.

In the present study, we used target enrichment to recover genetic data for hundreds of nuclear genes to provide confidence in the phylogenetic placement of Cotula alpina relative to other species of Australian Anthemideae and resolve this outstanding taxonomic question.

Materials and methods

Results

The phylogeny (Fig. 2) showed the ingroup divided into the following two large, strongly supported clades: one comprised the genera Achillea L., Anthemis L. (including Tripleurospermum maritimum (L.) W.D.J.Koch), Argyranthemum Webb ex Sch.Bip., Artemisia L., Matricaria L., and Tanacetum L.; the other comprised Cotula, Hippia L., Leptinella, Schistostephium Less., and Soliva Ruiz & Pav. Cotula and Leptinella were reciprocally monophyletic except for the placement of C. alpina in the latter genus. The monophyly of Leptinella including C. alpina was strongly supported (UFB = 100), but the Cotula clade without C. alpina received a UltraFast Bootstrap value of only 93. Conspecific samples of Anthemis arvensis L., C. alpina, L. filicula, L. reptans and Soliva sessilis Ruiz & Pav. were not placed as sister terminals.

Fig. 2.

Maximum-likelihood phylogeny of concatenated Angiosperms353 target-capture data of Australian native and introduced Anthemideae species, rooted on three representatives of related tribes. Numbers above branches are UltraFast Bootstrap values. Identifiers starting with ERR and SRR are Sequence Read Archive accession numbers of data published by PAFTOL; the five-digit identifier 80092 is a Genomics for Australian Plants sample number, with sequence reads published in the Bioplatforms Australia Data Portal (data.bioplatforms.com); identifiers starting with CANB or CBG are herbarium accession numbers of specimens sequenced specifically for this study. Red text indicates current circumscription of Cotula, blue text is Leptinella, and bold text indicates species indigenous to Australia, including Macquarie Island. Scale bar indicates estimated substitutions per site.

Discussion

Although neither Cotula nor Leptinella was comprehensively sampled across their global diversity, our results provided strong support to confirm the results of Himmelreich et al. (2012), namely that C. alpina is a sister lineage to or potentially nested within the genus Leptinella, as suggested by its growth habit. Within Leptinella, Cotula alpina was part of a small clade sister to the remainder of the genus, the ‘filicula-group’ of Himmelreich et al. (2012). In addition to the Australian C. alpina and L. filicula, the clade also included L. altilitoralis (P.Royen & D.G.Lloyd) D.G.Lloyd & C.J.Webb and Leptinella wilhelminensis (P.Royen) D.G.Lloyd & C.J.Webb, both of New Guinea, which we did not sequence.

This relationship is not replicated in our phylogeny because our two samples of Cotula alpina and our two samples of Leptinella filicula formed a grade under the remainder of Leptinella. A caveat of our dataset is that it includes only a single sample of most species, and that duplicate samples from the same species were rarely placed as sister terminals. Possible reasons are the use of protein-coding gene regions, which can be expected to evolve slowly, and the patchy nature of the sample × gene matrix, because the median sample had data for 289 of the 353 targeted loci (147 and 305 in the case of the two samples of C. alpina). Resolution at shallower levels is often improved by replicate sampling of five or more specimens per species (Schmidt-Lebuhn 2022), which is rarely feasible in higher-level phylogenetic studies. However, our data do not provide strong evidence against the monophyly of the filicula-group, given this limited sampling.

Our results imply that the character traditionally used to delineate Cotula, namely, the absence of corollas in female florets, is in this case homoplasious. However, this is not without precedent, because Thompson (2007, p. 46) noted in one inverse case that ‘the outer florets [of Cotula] are female and lack a corolla except for a weakly developed one in C. bipinnata’. Without more comprehensive sampling across the clade of Cotula, Leptinella and Soliva, it remains unclear whether absence or presence of corollas in female florets is the ancestral state, but either inference would require several gains or losses.

More generally, genetic analyses of the past two decades have repeatedly demonstrated that what might be called a total-evidence approach relying on a larger suite of characters including growth habit and branching pattern is more informative about evolutionary relationships than individual reproductive characters considered important a priori. Examples include the cases of Odixia Orchard and Ozothamnus R.Br., which differ only in the absence or presence of the pappus (Schmidt-Lebuhn and Constable 2013), and Leucochrysum fitzgibbonii (F.Muell.) Paul G.Wilson, which differs from members of Waitzia J.C.Wendl. only in the plesiomorphy of lacking a beak on the cypsela (Weber and Schmidt-Lebuhn 2015).

Because we confirmed the grouping of Cotula alpina with Leptinella, the question of a transfer to the latter genus arises. A counter-argument is the possible non-monophyly of Cotula in its current circumscription. The phylogeny of Himmelreich et al. (2012) showed Cotula as a clade comprising the type species C. coronopifolia L., C. australis, and C. turbinata L., but C. mexicana (DC.) Cabrera nested in Soliva, and C. abyssinica Sch.Bip. ex A.Rich. in an isolated position. Should these relationships be corroborated, sinking all three genera into a broadly circumscribed Cotula might be justified, rendering a name change for C. alpina unnecessary.

However, this is entirely speculative on current evidence and, given the monophyly of Leptinella and the placement of the type species of Cotula in a clade of at least several species (including also C. bipinnata Thunb. and C. cotuloides in our phylogeny), a less disruptive solution would be to transfer a few other species of Cotula as required to maintain the genus-level taxonomy mostly as it is now.

On the basis of the phylogenetic evidence presented in this study, the shared stoloniferous, scapose growth form of Cotula alpina and members of Leptinella, and the currently accepted circumscription of the two genera, we therefore transfer the species to the latter genus. We also take the opportunity to lectotypify the basionym Ctenosperma alpinum Hook.f.

Taxonomy

Data availability

Raw reads of data generated for this project are available on the Sequence Read Archive (https://www.ncbi.nlm.nih.gov/sra) as accessions SRR24286507–SRR24286534 and SRR26780972. The concatenated data matrix and phylogenetic tree are available on the CSIRO Data Access Portal at doi:10.25919/2z7m-5866.

Conflicts of interest

The authors declare that they have no conflicts of interest.

Declaration of funding

Laboratory work was funded by the CSIRO Future Science Platform Environomics. Sequencing was funded by Bioplatforms Australia. The Genomics for Australian Plants Framework Initiative consortium is supported by funding from Bioplatforms Australia (enabled by NCRIS), the Ian Potter Foundation, Royal Botanic Gardens Foundation (Victoria), Royal Botanic Gardens Victoria, the Royal Botanic Gardens and Domain Trust, the Council of Heads of Australasian Herbaria, CSIRO, Centre for Australian National Biodiversity Research and the Department of Biodiversity, Conservation and Attractions, Western Australia.