Peter B. Heenan

Aim To evaluate competing views on the origin and distribution of the New Zealand flora by testing the hypothesis that the geographical distribution of species is unrelated to ecological traits such as habitat requirements and dispersal capabilities. Location The New Zealand archipelago. Methods An analysis of the factors correlated with distribution and endemism for alpine plants within New Zealand, and for the New Zealand biota as a whole. Results Woody plants are highly endemic; nonendemic plants tend to be herbaceous and are concentrated among the highly dispersible ferns and fern allies, orchids and wetland plants. These groups make up 32% of the total flora but contribute 78% of nonendemics. Alpine plants with wide spatial distribution tend to have greater altitudinal ranges, a broader habitat preference and better dispersal ability. Main conclusions Most vascular plants reached New Zealand by long‐distance transoceanic dispersal, probably during the Late Miocene to early Pleistocene period. During the Miocene and Pliocene, similar climates and landscapes to those of Australia and northern island groups, and highly invasible terrain, permitted dispersal of woody plants. Cooling climates and formation of a more mountainous, more compact landscape after that time reduced dispersal of woody plants and favoured herbaceous, wetland and highly dispersible plant groups. The prominence of dispersal has led to intense selective immigration, and is responsible for many characteristic features of the flora. Species selection by glacial–interglacial cycles has restricted acquisition or retention of cool or arid climate adaptations, particularly in the lowland flora. Endemic and range disjunction patterns in the New Zealand mainland are not, in general, directly caused by Pliocene inundations or the faulting and associated horizontal displacement of terrain that has continued since the Miocene. They have arisen mainly through Pleistocene extinctions, speciation and dispersal, and some patterns are strongly linked to repeated glaciation. Endemic centres are associated with differentiated terrain and climates providing isolation, distinctive environments, and habitat continuity conducive to speciation.

The generic taxonomy of the Nothofagaceae is revised. We present a new phylogenetic analysis of morphological characters and map these characters onto a recently published phylogenetic tree obtained from DNA sequence data. Results of these and previous analyses strongly support the monophyly of four clades of Nothofagaceae that are currently treated as subgenera of Nothofagus. The four clades of Nothofagaceae are robust and well-supported, with deep stem divergences, have evolutionary equivalence with other genera of Fagales, and can be circumscribed with morphological characters. We argue that these morphological and molecular differences are sufficient for the four clades of Nothofagaceae to be recognised at the primary rank of genus, and that this classification will be more informative and efficient than the currently circumscribed Nothofagus with four subgenera. Nothofagus is recircumscribed to include five species from southern South America, Lophozonia and Trisyngyne are reinstated, and the new genus Fuscospora is described. Fuscospora and Lophozonia, with six and seven species respectively, occur in New Zealand, southern South America and Australia. Trisyngyne comprises 25 species from New Caledonia, Papua New Guinea and Indonesia. New combinations are provided where necessary in each of these genera.

The mustard family (Brassicaceae) is a scientifically and economically important family, containing the model plant Arabidopsis thaliana and numerous crop species that feed billions worldwide. Despite its relevance, most phylogenetic trees of the family are incompletely sampled and often contain poorly supported branches. Here, we present the most complete Brassicaceae genus-level family phylogenies to date (Brassicaceae Tree of Life or BrassiToL) based on nuclear (1,081 genes, 319 of the 349 genera; 57 of the 58 tribes) and plastome (60 genes, 265 genera; all tribes) data. We found cytonuclear discordance between the two, which is likely a result of rampant hybridization among closely and more distantly related lineages. To evaluate the impact of such hybridization on the nuclear phylogeny reconstruction, we performed five different gene sampling routines, which increasingly removed putatively paralog genes. Our cleaned subset of 297 genes revealed high support for the tribes, whereas support for the main lineages (supertribes) was moderate. Calibration based on the 20 most clock-like nuclear genes suggests a late Eocene to late Oligocene origin of the family. Finally, our results strongly support a recently published new family classification, dividing the family into two subfamilies (one with five supertribes), together representing 58 tribes. This includes five recently described or re-established tribes, including Arabidopsideae, a monogeneric tribe accommodating Arabidopsis without any close relatives. With a worldwide community of thousands of researchers working on Brassicaceae and its diverse members, our new genus-level family phylogeny will be an indispensable tool for studies on biodiversity and plant biology.