Aim  To infer phylogenetic relationships among Antirrhinum species and to reconstruct the historical distribution of observed sequence polymorphism through estimates of haplotype clades and lineage divergence. Location  Antirrhinum is distributed primarily throughout the western Mediterranean, with 22 of 25 species in the Iberian Peninsula. Methods  Plastid (83 trnS‐trnG and 83 trnK‐matK) and nuclear (87 ITS) sequences were obtained from 96 individuals representing 24 of the 25 Antirrhinum species. Sequences were analysed using maximum parsimony, Bayesian inference and statistical parsimony networking. Molecular clock estimates were obtained for plastid trnK‐matK sequences using the penalized likelihood approach. Results  Phylogenetic results gave limited support for monophyletic groups within Antirrhinum. Fifty‐one plastid haplotypes were detected and 27 missing haplotypes inferred, which were all connected in a single, star‐like network. A significant number of species shared both the same haplotypes and the same geographical areas, primarily in eastern Iberia. Furthermore, many species harboured populations with unrelated haplotypes from divergent haplotype clades. Plastid haplotype distribution, together with nucleotide additivity in 59 of the 86 nuclear ribosomal ITS sequences, is interpreted as evidence of extensive hybridization. Lineage divergence estimates indicated that differentiation within Antirrhinum post‐dates the Miocene, when the Mediterranean climate was established. Main conclusions  Incongruence between plastid sequences, nuclear sequences and taxonomic delimitation is interpreted as strong evidence of limited cladogenetic processes in Antirrhinum. Rather, extensive nucleotide additivities in ITS sequences in conjunction with haplotype and haplotype‐clade distributions related to geographical areas support both recent and ancient hybridization. This geographical pattern of Antirrhinum speciation, particularly in eastern Iberia, is congruent with isolation–contact–isolation processes in the Pleistocene.

No abstract

Several aspects of the reproductive biology of Fumana juniperina were analyzed to find out whether delayed selfing occurs within a representative natural population in southwest Sardinia. Similar fruit set was obtained after autonomous, within-flower, and between-flower self-pollination (0.35, 0.53, and 0.46, respectively), crosspollination (0.45), and control (0.39). Seed set and seed mass did not differ significantly between pollination treatments (P > 0:05). Autonomous self-pollination was favored at the end of floral anthesis, when the petals fell and the sepals closed and pushed the anthers onto the stigma, making contact. No pollinators were seen visiting the flowers during the 42-h observation period, and the stigmatic pollen load on open-pollinated flowers, collected after the petals had fallen and before anther and stigma contact had occurred, was not enough to fertilize all the ovules. Low levels of inbreeding depression were found for fruit set (d ¼ À0:167), seed set (d ¼ À0:086), seed mass (d ¼ 0:008), and seed germination (d ¼ À0:052). These results support the idea that F. juniperina is selfcompatible and capable of autonomous self-pollination through changes in the position of the stigma and anthers at the end of floral anthesis. This is a case of how delayed selfing can be an efficient mating strategy that ensures reproduction in this plant when pollinators are scarce. This is the first time that a mechanism of delayed selfing has been described within the family Cistaceae.

Reproductive barriers play an important role in the maintenance of species boundaries. However, to date, few studies have provided a detailed analysis of reproductive isolation barriers between species or examined their importance in maintaining species identity. This is the first detailed study into pre-and post-zygotic reproductive isolation barriers in Antirrhinum, based on a mixed population with two species that rarely co-occur. The study revealed that pollinator constancy and preference and poor hybrid seed viability were the most important reproductive isolating mechanisms. Reproductive isolation was practically complete by both pre-and post-zygotic barriers. Average pre-zygotic isolation was greater than post-zygotic isolation, in accordance with the trend observed in flowering plants in which reproductive isolation is principally caused by pre-zygotic mechanisms. However, average post-zygotic isolation was also high, in contrast to what was expected among Antirrhinum spp. This case highlights the importance of quantifying the reproductive isolation barriers thoroughly to understand how and why species boundaries are maintained.