Summary• DNA methylation can cause heritable phenotypic modifications in the absence of changes in DNA sequence. Environmental stresses can trigger methylation changes and this may have evolutionary consequences, even in the absence of sequence variation. However, it remains largely unknown to what extent environmentally induced methylation changes are transmitted to offspring, and whether observed methylation variation is truly independent or a downstream consequence of genetic variation between individuals.• Genetically identical apomictic dandelion (Taraxacum officinale) plants were exposed to different ecological stresses, and apomictic offspring were raised in a common unstressed environment. We used methylation-sensitive amplified fragment length polymorphism markers to screen genome-wide methylation alterations triggered by stress treatments and to assess the heritability of induced changes.• Various stresses, most notably chemical induction of herbivore and pathogen defenses, triggered considerable methylation variation throughout the genome. Many modifications were faithfully transmitted to offspring. Stresses caused some epigenetic divergence between treatment and controls, but also increased epigenetic variation among plants within treatments.• These results show the following. First, stress-induced methylation changes are common and are mostly heritable. Second, sequence-independent, autonomous methylation variation is readily generated. This highlights the potential of epigenetic inheritance to play an independent role in evolutionary processes, which is superimposed on the system of genetic inheritance.

Apomixis, asexual reproduction through seeds, has the potential to revolutionize agriculture if its genetic basis can be elucidated. However, the genetic control of natural apomixis has remained obscure until quite recently, owing to all the complications of Mendelian genetics, such as epistatic gene interactions, components that are expressed sporophytically and gametophytically, expression modifiers, polyploidy, aneuploidy, segregation distortion, suppressed recombination, etc., that seem to have accumulated during the evolution of apomixis. In this review we show how molecular markers and superior phenotypic methods have been used to clarify the genetics of apomixis in monocots as well as dicots during the past 15 years. Many of the complexities in the genetics of apomixis are likely secondary and have evolved as a consequence of the reproductive process. New mapping techniques, such as comparative mapping, linkage disequilibrium mapping, and deletion mapping, and new high-throughput sequencing methods, will help to penetrate the core of apomixis chromosomal regions. If the evolutionary genetic load can be exposed and removed, the apomixis genes may be used in agriculture as a tool to fix elite genotypes.

The ecological and evolutionary opportunities of apomixis in the short and the long term are considered, based on two closely related apomictic genera: Taraxacum (dandelion) and Chondrilla (skeleton weed). In both genera apomicts have a wider geographical distribution than sexuals, illustrating the short-term ecological success of apomixis. Allozymes and DNA markers indicate that apomictic populations are highly polyclonal. In Taraxacum, clonal diversity can be generated by rare hybridization between sexuals and apomicts, the latter acting as pollen donors. Less extensive clonal diversity is generated by mutations within clonal lineages. Clonal diversity may be maintained by frequency-dependent selection, caused by biological interactions (e.g. competitors and pathogens). Some clones are geographically widespread and probably represent phenotypically plastic 'general-purpose genotypes'. The long-term evolutionary success of apomictic clones may be limited by lack of adaptive potential and the accumulation of deleterious mutations. Although apomictic clones may be considered as 'evolutionary dead ends', the genes controlling apomixis can escape from degeneration and extinction via pollen in crosses between sexuals and apomicts. In this way, apomixis genes are transferred to a new genetic background, potentially adaptive and cleansed from linked deleterious mutations. Consequently, apomixis genes can be much older than the clones they are currently contained in. The close phylogenetic relationship between Taraxacum and Chondrilla and the similarity of their apomixis mechanisms suggest that apomixis in these two genera could be of common ancestry.

DNA methylation is an epigenetic mechanism that has the potential to affect plant phenotypes and that is responsive to environmental and genomic stresses such as hybridization and polyploidization. We explored de novo methylation variation that arises during the formation of triploid asexual dandelions from diploid sexual mother plants using methylation-sensitive amplified fragment length polymorphism (MS-AFLP) analysis. In dandelions, triploid apomictic asexuals are produced from diploid sexual mothers that are fertilized by polyploid pollen donors. We asked whether the ploidy level change that accompanies the formation of new asexual lineages triggers methylation changes that contribute to heritable epigenetic variation within novel asexual lineages. Comparison of MS-AFLP and AFLP fragment inheritance in a diploid x triploid cross revealed de novo methylation variation between triploid F(1) individuals. Genetically identical offspring of asexual F(1) plants showed modest levels of methylation variation, comparable to background levels as observed among sibs in a long-established asexual lineage. Thus, the cross between ploidy levels triggered de novo methylation variation between asexual lineages, whereas it did not seem to contribute directly to variation within new asexual lineages. The observed background level of methylation variation suggests that considerable autonomous methylation variation could build up within asexual lineages under natural conditions.

Some dandelions, Taraxacum, are diplosporous gametophytic apomicts. Crosses between closely related diploid sexuals and triploid apomicts were made to study the inheritance of apomixis. Seed-set was less than one-third of that in diploid´diploid crosses, probably because of the inviability of aneuploid pollen or zygotes. Almost 90% of the viable o spring were diploid and the result of sel®ng, as was shown by a discriminating allozyme marker. Aneuploid outcross pollen had a mentor e ect on self-pollen, causing a breakdown of the sporophytic self-incompatibility system. A similar phenomenon has been reported before in wide crosses. Of the 26 allozyme-con®rmed hybrids, four were diploids, 15 were triploids and seven were tetraploids. Diploid hybrids were signi®cantly less frequent than triploid hybrids, suggesting either low ®tness of haploid pollen or more numerous formation of diploid pollen. Emasculation and bagging of¯owers indicated apomictic seed-set in none of the diploid, in one-third of the triploid and in all of the tetraploid hybrids. All apomictic hybrids showed partial seed-set, but additional cross-pollination did not increase seed-set. Cytological analysis of the F 2 progeny con®rmed that partial apomixis was caused by semisterility and not by residual sexuality (facultative apomixis). The di erence in segregation for apomixis between triploid and tetraploid hybrids may be because the triploids originated from partially reduced diploid pollen grains, whereas the tetraploids originated from unreduced triploid pollen grains.