The breeding of new, high-quality citrus cultivars depends on dependable information about the relationships of taxa within the tribe Citreae; therefore, it is important to have a well-supported phylogeny of the relationships between species not only to advance breeding strategies, but also to advance conservation strategies for the wild taxa. The recent history of the systematics of Citrus (Rutaceae: Aurantioideae) and its allies, in the context of Rutaceae taxonomy as a whole, is reviewed. The most recent classification is tested using nine cpDNA sequence regions in representatives of all genera of the subfam. Aurantioideae (save Limnocitrus) and numerous species and hybrids referred to Citrus s.l. Aurantioideae are confirmed as monophyletic. Within Aurantioideae, tribe Clauseneae are not monophyletic unless Murraya s.s. and Merrillia are removed to Aurantieae. Within tribe Aurantieae, the three traditionally recognized subtribes are not monophyletic. Triphasiinae is not monophyletic unless Oxanthera is returned to Citrus (Citrinae). Balsamocitrinae is polyphyletic. Feroniella, traditionally considered allied closely to Limonia (=Feronia), is shown to be nested in Citrus. The proposed congenericity of Severinia and Atalantia is confirmed. The most recent circumscription of Citrus is strongly supported by this analysis, with hybrids appearing with their putative maternal parents. The genus was resolved into two clades, one comprising wild species from New Guinea, Australia, and New Caledonia (formerly Clymenia, Eremocitrus, Microcitrus, Oxanthera), but surprisingly also Citrus medica, traditionally believed to be native in India. The second clade is largely from the Asian mainland (including species formerly referred to Fortunella and Poncirus).

A long-standing hypothesis in evolutionary biology is that polyploid plants have a fitness advantage over diploids in climatically variable or extreme habitats. Here we provide the first empirical evidence that polyploid advantage in these environments is caused by two distinct processes: homeostatic maintenance of reproductive output under elevated abiotic stress, and fixed differences in seed development. In an outdoor climate manipulation experiment using coastal to inland Australian populations of the perennial grass Themeda triandra Forssk., we found that total output of viable seed in drought- and heat-stressed tetraploid plants was over four times higher than in diploids, despite being equal under more favourable growing conditions. Tetraploids also consistently produced heavier seeds with longer hygroscopic awns, traits which increase propagule fitness in extreme environments. These differences add to fitness benefits associated with broader-scale local adaptation of inland T. triandra populations to drought stress. Our study provides evidence that nucleotypic effects of genome size and increased reproductive flexibility can jointly underlie polyploid advantage in plants in stressful environments, and argue that ploidy can be an important criterion for selecting plant populations for use in genetic rescue, restoration and revegetation projects, including in habitats affected by climate change.

Controlled pollinations were used to examine the mating system of diploid and tetraploid individuals of the endangered grassland herb Rutidosis leptorrhynchoides F.Muell. Crosses among unrelated plants gave 1.5–2 times as many fruit as crosses between plants that were half-sibs, while selfed crosses generally resulted in no fruit. Three classes of compatibility reaction were observed within outcross treatments: (1) reciprocal compatibility, (2) one-way compatibility and (3) reciprocal incompatibility. This is diagnostic of sporophytic control of self-incompatibility, which is characteristic of the Asteraceae. This is supported by the occurrence of a dry stigma and trinucleate pollen. Analysis of the behaviour of self- and outcross pollen on the stigma by using fluorescence microscopy shows that rejection of self-pollen does not all occur at one point but at a number of stages, with cumulative reductions in the adherence of pollen to the stigma, pollen germination, pollen tube penetration of the stigma and fertilisation. On the basis of both fruit set and pollen behaviour data, for any level of relatedness, tetraploid plants are about 20% less likely to be compatible with each other than diploid plants. This presumably reflects the greater likelihood of matching S alleles given the greater potential for polymorphism at the individual level. The occurrence of two plants, one diploid and one tetraploid, that set large amounts of fruit on selfing shows that self-incompatibility can break down.