No abstract

The genus Taraxacum is readily divided into “primitive” and “advanced” forms, on morphological and cytological criteria. It is thought that the genus arose in the west Himalayas during the Cretaceous and that apomixis arose at an early stage by means of polyploidy, precocious embryony and asynapsis in the female meiosis. Advances of sexual primitive types and “precursor” types west into Europe were paralleled by the spread of arctic‐alpine types into many regions of the world. During the Pleistocene the precursor types are thought to have generated the widespread advanced section Ceratophora, which gave rise to many of the advanced species after the last glacial period by hybridizing with primitive and precursor sexuals, thus “fixing” a hybrid swarm as apomicts.

Apomixis is a common feature of perennial plants, which occurs in ca. 60% of the British flora, but has been largely ignored by reproductive theoreticians. Successful individuals may cover huge areas, and live to great ages, favoured by 'symmetrical' selection. Apomixis is favoured by colonizing modes, for instance post-glacially. Despite its theoretical advantages, apomixis usually coexists with sexuality, suggesting 'hidden' disadvantages. Agamospermy (apomixis by seed) is relatively uncommon, but gains from the attributes of the seed. It pays agamospermy genes, which discourage recombination, to form co-adapted linkage groups, so that they become targets for disadvantageous recessive mutant accumulation. Consequently, agamospermy genes cannot succeed in diploids and agamosperms are hybrid and highly heterotic. Agamospermous endosperm may suffer from genomic imbalance, so that nutritious ovules, which can support embryos without endosperm, may be preadapted for agamospermy. When primary endosperm nucleus fertilization ('pseudogamy') continues as a requirement for many aposporous agamosperms, selfing sex becomes preadaptive and archesporial sex remains an option. Apomictic populations can be quite variable although apomictic families are much less variable than sexuals. Only in some diplosporous species does sex disappear completely, and in those species some release of variability may persist through somatic recombination. The search for an agamospermy gene suitable for genetic modification should target fertile sexuals with a single localized agamospermy (A) gene, which therefore lack a genetic load. The A gene should coexist alongside sexuality, so that it would be easy to select seedlings of sexual and asexual origins. Plants with sporophytic agamospermy provide all these attributes.

We sequenced the trnL and rpl16 introns of the chloroplast DNA from 95 of the ca. 425 species (30 of 37 sections, seven of eight subgenera) of Primula L. in order to reconstruct the phylogenetic history of the group. Among the 24 additional taxa sampled are representatives of all genera that are likely to be embedded in Primula, as well as outgroups from the Maesaceae, Theophrastaceae, and Myrsinaceae. In the strict consensus of the most parsimonious trees, Primula and the genera embedded in it (Dionysia Fenzl., Sredinskya [Stein] Fedorov, Dodecatheon L., and Cortusa L.) are sister to a clade of several genera previously suspected to be embedded in Primula (Hottonia L., Omphalogramma [Franchet] Franch., and Soldanella L.). In recognition of this, two new rankless names are defined for these clades (/Primula and /Soldanella). Close relationships are inferred between Dionysia and Primula subgenus Sphondylia (Duby) Rupr., Sredinskya and Primula subgenus Primula, Dodecatheon and Primula subgenus Auriculastrum Schott, and Cortusa and Primula subgenus Auganthus (Link) Wendelbo. The largest subgenus, Aleuritia (Duby) Wendelbo, is dispersed among three clades that are not each other's closest relatives. Primula sections Muscarioides Balf. f., Soldanelloides Pax, Denticulata Watt, Armerina Lindley, and Aleuritia Duby are resolved as para-or polyphyletic with moderate to strong support. Throughout, we consider the striking morphological and cytological variation seen in Primula within a phylogenetic context, particularly as it relates to the close relationship implied here between Dionysia and Primula subgenus Sphondylia. The homology of involute leaf vernation in Primula is reconsidered in light of its two independent origins, and we come to the conclusion that vernation in subgenus Sphondylia is better characterized as conduplicate.