Comparison of the four species groups comprising the tribe Geissoieae (Lamanonia in South America, Geissois sensu stricto from islands in the south-western Pacific, and Pseudoweinmannia and Geissois in Australia) showed differences in the stipules, inflorescences and flowers. Molecular phylogenetic analyses suggested that Geissois sensu lato is paraphyletic, with Australian Geissois being closer to Pseudoweinmannia than to New Caledonian Geissois. The name Karrabina is published to accommodate the Australian species formerly placed in Geissois, namely K. benthamiana Rozefelds & H.C.Hopkins and K. biagiana Rozefelds & H.C.Hopkins. Distinguishing characters for Karrabina are: four lateral stipules per node that vary from free to largely fused between and sometimes across the petioles; inflorescences with a variable number of metamers and racemes, with both median and lateral modules that always lack a median raceme; and pale flowers with five or six calyx lobes and shortly exserted stamens. Geissois s.s. differs in having: two intrapetiolar stipules per node; inflorescences consisting of few metamers and few racemes, with the modules being always lateral and possessing a median raceme; flowers with four red calyx lobes and prominently exserted stamens. The generic limits of Lamanonia and Pseudoweinmannia remain unchanged.

Lomatia R.Br. is a genus of 12 species in South America and eastern Australia. Hybridisation is extensive in the Australian species and molecular work is required to understand phylogenetic relationships and examine potential gene flow among species. We developed a library of microsatellite markers for Lomatia silaifolia (Sm.) R.Br. These markers were tested across population samples of L. silaifolia and L. myricoides (C.F.Gaertn.) Domin, assessed for cross amplification across all 12 species of Lomatia, sequenced and inspected for variation in the microsatellite flanking region (MFR), and utilised for phylogeographic and phylogenetic analysis. Nineteen microsatellite markers were tested, 13 of which were polymorphic in size analysis. Four of the markers amplified consistently within Lomatia and the MFR had equivalent to or more sequence variation than the three universal markers (psbA–trnH intergenic spacer, PHYA, ITS). Variation within five individuals of L. silaifolia indicated that MFR could be phylogeographically informative. Combined phylogenetic analysis of Lomatia using universal markers and MFR resulted in a well supported tree; however, phylogenetic analysis of only MFR suggested non-monophyly of the species. Phylogenetic trees supported South American species as being diverged from the Australian species but paraphyletic with respect to the Australian lineage. Relationships among Australian species of Lomatia are correlated with geography rather than morphology. We found that microsatellite markers designed for Lomatia and the flanking regions can be informative at population, phylogenetic and phylogeographic levels.

Three new species of Gingidia (Apiaceae: Apioideae) segregated from the G. montana (J.R.Forst. & G.Forst.) J.W.Dawson complex are named, with G. montana now regarded as a New Zealand endemic. The new Australian endemic, G. rupicola I.Telford & J.J.Bruhl, is restricted to the eastern escarpment of the New England Tableland, New South Wales. With few populations and limited numbers of plants, the conservation assessment of G. rupicola is Endangered. G. haematitica Heenan is described as a new species from North-West Nelson, South Island, New Zealand, where it is restricted to base-rich substrates in the Burnett Range. Because of its restricted distribution and the continued mining for dolomite at the site of the largest population, G. haematitica is considered to have a conservation assessment of Nationally Critical. The second New Zealand endemic, G. amphistoma Heenan, is known from alpine habitats in the Southern Alps, South Island. Distributions of the three new species are mapped, habitats noted, and a table compares attributes of these species and G. montana.

Species delimitation is of critical importance in systematics and biological and conservation research. The general-lineage species concept, which defines species as separately evolving metapopulation lineages, considers multiple lines of evidence to identify lineages and delimit species boundaries. Here, we apply the general-lineage concept to the New Zealand endemic Myosotis petiolata Hook.f. (Boraginaceae) species complex, to test its usefulness in the New Zealand Myosotis L. species radiation. We aimed to determine whether the complex contains separately evolving lineages to assess the criteria of monophyly, distinct genotypic clusters and fixed morphological differences by using amplified fragment length polymorphism (AFLP) and morphological data. The use of multiple criteria to identify separately evolving lineages within the M. petiolata complex was effective, but the different criteria were satisfied to varying degrees. Species rank is recommended for each of the currently recognised varieties as Myosotis pottsiana (L.B.Moore) Meudt, Prebble, R.J.Stanley & Thorsen (comb. & stat. nov.), M. pansa (L.B.Moore) Meudt, Prebble, R.J.Stanley & Thorsen (comb. & stat. nov.) and M. petiolata Hook.f. (North Island individuals only). A new allopatric subspecies, M. pansa subsp. praeceps Meudt, Prebble, R.J.Stanley & Thorsen (subsp. nov.), is also described. The distinguishing morphological characters and conservation status of each species are discussed. In addition, the low genetic diversity revealed in our population genetic analyses, coupled with few, fluctuating, disjunct populations, underscores the conservation priority of these three rare endemic New Zealand species.

The reciprocal monophyly of Craspedia and Pycnosorus (Asteraceae, Gnaphalieae) is tested, with a phylogenetic analysis of ribosomal and chloroplast DNA. Although one species of the latter genus was not sampled, the results of the present study indicated that it is most likely monophyletic as opposed to paraphyletic with respect to Craspedia. Within Craspedia, deeper sampling results in the appearance of several species (C. aurantia, C. coolaminica, C. glabrata and C. variabilis) in both major Australian rDNA clades, indicating that the available molecular phylogenies have to be considered gene trees instead of species phylogenies. Additional studies using more independent loci and species-tree approaches are needed to resolve species relationships in the genus.

The first comprehensive quantitative study of morphological characters in all Australian species of the billy button genus Craspedia (Asteraceae, Gnaphalieae) is presented. Homogeneity analysis and pair-wise reclassification tests are used to test species circumscriptions under the genotypic-cluster concept. Although most species are supported, C. aurantia and C. jamesii do not form separate morphological clusters, and C. jamesii is reduced to a variety of C. aurantia. Although our results indicated that C. variabilis may be synonymous with C. glauca, we reserve judgment because of the small number of specimens of the latter species included in the present study. Other species were found to be less problematic; however, additional and more detailed studies will be needed to clarify the circumscription of various species with woolly leaf indumentum, in particular C. canens, C. coolaminica and C. macrocephala. A synopsis of all Australian species of Craspedia is presented, together with a preliminary key.