Nature 450, 1184-1189 (20 December 2007) | doi:10.1038/nature06393; Received 2 May 2007; Accepted 18 October 2007

After a dozen years of progress the origin of angiosperms is still a great mystery

Michael W. Frohlich1 & Mark W. Chase1


Here we discuss recent advances surrounding the origin of angiosperms. Putatively primitive characters are now much better understood because of a vastly improved understanding of angiosperm phylogenetics, and recent discoveries of fossil flowers have provided an increasingly detailed picture of early diversity in the angiosperms. The 'anthophyte theory', the dominant concept of the 1980s and 1990s, has been eclipsed; Gnetales, previously thought to be closest to the angiosperms, are related instead to other extant gymnosperms, probably most closely to conifers. Finally, new theories of flower origins have been proposed based on gene function, duplication and loss, as well as on morphology. Further studies of genetic mechanisms that control reproductive development in seed plants provide a most promising avenue for further research, including tests of these recent theories. Identification of fossils with morphologies that convincingly place them close to angiosperms could still revolutionize understanding of angiosperm origins.

After a dozen years of progress the origin of angiosperms is still a great mystery

Less than a dozen years ago even the most basic questions regarding the origin of angiosperms were still disputed, including the nature of primitive flowers, what sorts of gymnosperms might have given rise to angiosperms, and the broadest outlines of the evolutionary trajectory between them. (See ref. 1 for references not cited here.) Studies of fossil flowers2 showed that bisexual and unisexual flowers both occurred in the earliest fossil flower floras, so it was still possible that plants with unisexual flowers consisting of a single stamen or a single carpel (resembling extant Ceratophyllum or Hedyosmum) might reflect the ancestral angiosperm condition. Other analyses supported the directly opposite view: the overall organization of the bisexual angiosperm flower (flat structures surrounding male organs surrounding central female organs) had been inherited directly from gymnosperm ancestors. Subsequent advances, derived from new data and reinterpretations of older data, have narrowed the range of alternative explanations for origins of both flowers and angiosperms. New data have come especially from molecular phylogenetics, but also from studies of gene function, duplication and loss and from palaeobotany. The rise of evolutionary developmental biology ('evo–devo') has reinvigorated the study of plant anatomy and led to new, increasingly synthetic theories; they seek to fuse disparate fields to explain various aspects of flower origins. Formulation of detailed, testable theories combined with study of fossils and genes has the power to dispel the mystery surrounding the origin of both flowers and angiosperms.


Relationships between extant angiosperms

Relationships of relatively few groups of angiosperms are still in dispute. Furthermore, morphological and molecular phylogenetic results are now considerably more congruent than in the past. Analyses in the past five years have consistently pointed to the 'ANA' (formerly called ANITA) taxa— Amborellaceae, Nymphaeales and Austrobaileyales— as successive sister groups to the larger clades of magnoliids, eudicots and monocots (Fig. 1), although there has been some dispute over the relative positions of Amborellaceae and Nymphaeales3. A noteworthy new addition near the basal nodes of the angiosperm tree is the Hydatellaceae, a small family of minute aquatics with small simple flowers that were previously thought to be members of the monocot order Poales. These fall as sister to Nymphaeales (Fig. 1) and extend the range of morphologies among these clades4, 5.

The ANA taxa, including Hydatellaceae, are each individually highly specialized. For example, Amborella grows in wet, forest understorey habitats in New Caledonia and is dioecious (but with vestigial organs of the opposite sex), whereas Nymphaeales with perfect flowers are all adapted to aquatic habitats. Improvements in morphological reconstructions of primitive angiosperms can yet be expected (particularly with the application of likelihood methods that consider branch lengths in projecting character states down to the basal node1), but the range of hypotheses now considered relevant is considerably narrower than in the past. For example, the old view of the primitive carpel as conduplicate— folded lengthways and fused at the edge— was based on magnoliid taxa now known to be relatively derived. Carpels of most ANA taxa are bucket-shaped and sealed only by a secretion3. Most importantly, several previously popular ideas can now be discarded, such as the idea that angiosperms arose from more than one 'gymnosperm' ancestor.


Fossil flowers

The only direct evidence of early angiosperm flowers comes from fossils. Mesofossils, up to a few millimetres in size, often show exquisite three-dimensional preservation, including cell structure. Most mesozoic flowers fall within the mesofossil range. Diverse mesofossil assemblages span the late-Early to Late Cretaceous period (reviewed in ref. 6), and others extending back to the upper Jurassic are now known (E.-M. Friis, personal communication). Flowers referable to ANA angiosperms are found in the earliest mesofossil assemblages, along with flowers of Chloranthaceae, which are sister to the magnoliids in recent analyses of complete sets of genes from the plastid genome53. Fossil pollen provides yet earlier evidence of angiosperms at roughly 136 Myr ago (Hauterivian6; mid-Early Cretaceous), about 10 Myr before the earliest published mesofossil floras7. Fossil pollen shows that diverse magnoliids, monocots and early eudicots had appeared by the early Aptian, about 125 Myr ago, demonstrating an early, rapid major radiation.

Molecular results have sharpened evolutionary interpretations. For example, unisexual fossil flowers similar to the extant genus Hedyosmum (Chloranthaceae) are among the earliest flowers known6. However, our knowledge that Chloranthaceae insert well above the ANA taxa clearly indicates that unisexuality in Hedyosmum and most probably also these fossil taxa is due to secondary reduction. The addition of Hydatellaceae as sister to Nymphaeales does not change this inference.

There are no studied fossils clearly representing stem-group angiosperms, that is, of plants related to extant angiosperms but attached below the basal node of extant angiosperms in the tree. Such fossils might provide spectacular direct evidence of morphological change along this unknown stretch of evolutionary history. Archaefructus, originally thought to be a stem-group angiosperm of Jurassic age, is not; it has been re-dated as mid-Early Cretaceous, and its reproductive unit has been reinterpreted as an inflorescence, not a flower8, 9. Reinterpretation of Archaefructus is a good example of initial morphological interpretations leading to remarkably different ideas of relationship compared with subsequent analyses.


Evidence from phylogenetics and morphology

Living gymnosperms and angiosperms constitute the extant seed plants. The four groups of living gymnosperms are only a remnant of the substantial diversity of Palaeozoic and Mesozoic times10. In the 1980s, morphological cladistic analyses of living and fossil seed plants11, 12 placed cycads sister to the other living taxa and identified 'anthophytes' as consisting of angiosperms, Gnetales (Fig. 2) and two extinct groups, Bennettitales (Fig. 3e) and Pentoxylon. The first three include members with reproductive units that have sterile appendages surrounding male structures with female structures in the centre, suggesting that this overall organization might be homologous in the three groups, hence much antedating origin of angiosperms. Otherwise, their reproductive structures differ markedly, but, given these relationships, morphological transformations have been proposed12. Subsequent analyses placed Caytoniales (Fig. 3b, c), which lack this overall organization, within anthophytes as sister to angiosperms, undermining this supposed homology in overall organization13.

Figure 2: Welwitschia cones.
Figure 2 : Welwitschia cones. Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, or to obtain a text description, please contact

a, Female. b, Close-up of male cone, showing pollen organs and pollination droplet in between them. c, Male cones. pd, pollination droplet.

High resolution image and legend (207K)

Figure 3: Fossil gymnosperms.
Figure 3 : Fossil gymnosperms. Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, or to obtain a text description, please contact

a, Glossopteris showing cupules borne on stalk above a leaf (from ref. 23). b, Caytonia male (above) and female (below) reproductive units (from ref. 11). c, Caytonia cupule (from ref. 36). d, Corystosperm (Umkomasia) cupule containing one ovule (from ref. 52). Cupule wall almost surrounds ovule, except for a slit facing the stalk. e, Bennettitales (Williamsoniella) bisexual reproductive unit (from ref. 11); each oval pollen sac consists of several fused microsporangia. Ovules are borne among scales on the central stalk; in Vardekloftia each is enclosed by a cupule wall. Green, cupule wall; red, ovule; yellow, pollen organ.

High resolution image and legend (90K)

Molecular phylogenetic analyses of seed plants now indicate that living gymnosperms are monophyletic, with Gnetales related to conifers1, 14, although this remains controversial6, 15, 16. Palaeobotanists are increasingly willing to consider extant gymnosperm monophyly, but with varying levels of surprise and disquiet over the implications17. In the two most recent morphological analyses, placing Gnetales with conifers made trees one step longer17 or forcing extant gymnosperm monophyly cost four additional steps15, showing that the signal against extant gymnosperm monophyly is not especially strong.

Extant gymnosperm monophyly moves Gnetales and cycads over many nodes compared with the shortest morphology-based trees15, 17, but only rarely have phylogenetics studies of morphology and DNA data agreed in plant studies, even in well-studied groups. Early morphological cladistic analyses of angiosperms underwent a radical rethinking of character homologies in the light of DNA analyses, which generated results much more in line with the DNA trees18. Interpretation of morphological homologies can radically shift if evidence of alternative relationships triggers re-examination; however, no source of phylogenetic evidence is infallible. Future studies should show whether current molecular or morphological results are erroneous.

New types of data are promising, from large-scale sequencing of nuclear genes19 to molecular fossils. Oleanane is a diagenetic product of triterpenoids found in most angiosperms. Taylor et al.20 demonstrated oleanane associated with several Bennettitales and Permian gigantopterid fossils, but oleananes were not found in Gnetales, Palaeozoic medullosan pteridosperms or in the conifer relatives Cordaitales. Other chemical fossils are markers for Cordaitales21.

Monophyly of extant gymnosperms places them all equally distant from the angiosperms, which means that the lineage that eventually produced angiosperms diverged from the common ancestor with extant gymnosperms much earlier than previously thought, from among the 'pteridosperms' ('seed ferns'). Living gymnosperms show a great diversity of reproductive morphologies, and these must have resulted from numerous specializations. This makes comparison with angiosperms much more difficult.

Some extinct 'gymnosperm' groups must be closely related to angiosperms. If living and fossil 'gymnosperms' are considered together, then angiosperms arose from within them, making 'gymnosperms' paraphyletic (which we indicate with quotation marks). Both Caytoniales (Fig. 3b, c) and Bennettitales (Fig. 3e), fossil 'gymnosperms' with remarkably different morphologies, have long figured in theories of angiosperm origins11, 22 and appear as successive sister groups to angiosperms in recent studies15, 17. Caytoniales have cupules that could plausibly be transformed into angiosperm bitegmic ovules (see Box 1), but these are borne on slender stalks unlike carpels. The detached male structures also differ significantly from angiosperm stamens, and it is not known how either of these was borne on the plant. Some Bennettitales have bisexual reproductive units11, but there is no obvious carpel precursor, and except for Vardekloeftia17 the ovules are not borne inside a cupule, so the source of angiosperm-type bitegmic ovules is also uncertain. Microsporophylls are highly variable, but none so far known closely resemble angiosperm stamens.

Retallack and Dilcher23 suggested that the angiosperm carpel could be derived from structures resembling those of glossopterids (Fig. 3a), a group of Permian 'gymnosperms' that had a cupule or cupules borne on stalks above foliage leaves. Glossopterids (Fig. 3a) had one or more cupules borne on stalks above foliage leaves and have been suggested as angiosperm ancestors23, but their early (Permian) age is problematic. Doyle17 suggested that Caytoniales may be related to glossopterids and may also have had the reproductive stalk borne above a subtending leaf. If the stalk became fused to the leaf, the resulting structure would be an ideal carpel precursor.

Friedman and Floyd24 proposed a theory about the angiosperm female gametophyte that uses the idea of developmental modules to account for arrangements and fates of nuclei, including those that participate in double fertilization to make zygote and endosperm (food store). They suggested that the basic module consists of four nuclei, one of which moves to the centre of the initially coenocytic gametophyte to fuse with the second sperm forming the endosperm nucleus. In the module near the micropyle, where the pollen tube enters, the other three nuclei organize the two synergids and the egg cell that fuses with a sperm, making the zygote. Most angiosperms have a second module that also sends a nucleus to the centre of the gametophyte, so fusion generates the standard, triploid endosperm nucleus. There is much variation on this basic pattern, but most of this diversity is explicable by changing the numbers of modules. Nymphaeaceae have only one module, which could be the ancestral condition, especially because Amborella has a unique system.

Some unreasonable theories posit multiple origins of angiosperms from 'gymnosperm' ancestors25, 26. Angiosperms have many shared derived characters11, 12, and it is most unlikely that such complex features, arising independently, would fail to show differences that reveal their independent origins. All molecular and morphological analyses support angiosperm monophyly.


MADS genes

MADS-box transcription factors are important for flower origins because they specify the major floral organs and because their expression zones typically correspond to their zones of action, so expression studies are useful for inferring function. According to the classic 'ABC' model (in Arabidopsis terminology), sepals are specified by the 'A' gene Ap1 (and the non-MADS AP2), petals by the 'A' in combination with the two 'B' genes PI and AP3, stamens by the 'B' genes and the 'C' gene AG and carpels by AG alone. SEPALLATA (SEP) or 'E' genes are now known to be required for all four organ types, and the 'D' genes have been proposed as specifying ovules. Expression of the A, B, C and E MADS genes is upregulated by LEAFY, a non-MADS transcription factor.

Gene phylogenetics shows that each major MADS subgroup extends back to the base of extant angiosperms. There have been many duplication events within these clades, some probably reflecting whole-genome duplications, such as at the base of eudicots19.

Several pairs of major MADS clades result from duplications below extant angiosperms, such as the PI and AP3 clades, and also the AG clade and the putative 'D' gene clade. The most closely related gymnosperm genes are sister(s) to these clade pairs.

Classical 'A' function may be limited to relatives of Arabidopsis, whereas in other plants 'A' function may not be separable from the other major role of the 'A' genes in helping to specify apices as floral27. Lack of a unique sepal-specifying system is consistent with the suggestion that in the original flower the perianth may have been composed entirely of petals28.

Understanding the specification of ovules would be especially important but remains problematic. Overexpression of a Petunia 'D' gene in Petunia generates ectopic ovules on sepals and petals, but overexpression of the Arabidopsis orthologue in Arabidopsis does little. It is unclear whether ovule function versus stamen + carpel function characterized the 'D' versus 'C' clades from their initial divergence. Kramer et al.29 concluded that they do, but Zahn et al.30 produced contradictory evidence. Specification of ovules and their components is highly complex31, and there may be differences between taxa.

In ANA angiosperms, 'B' (and to some degree 'C') MADS genes show much broader messenger RNA expression than in eudicots, which has led to the 'fading borders' model of floral organ specification19, 32. This posits an activity gradient of floral genes that determine organ identity, resulting in a gradient in organ morphology from the outside to the centre, in contrast to flowers of most eudicots, which have sharply distinguished organs. Many ANA grade flowers have variable numbers of floral organs often arranged spirally (instead of in whorls), suggesting less developmental homeostasis than in eudicots or monocots1, 19, 32. Perhaps lower developmental homeostasis implies simpler systems for specifying floral organs in these plants that are more appropriate for comparison with gymnosperms than are systems of other angiosperms.

In gymnosperms, 'B' gene homologues are primarily expressed in developing male structures, resembling their role in angiosperm stamens. Homologues of the 'C' + 'D' clade are expressed in both male and female structures (including ovules), also suggesting broad conservation in their roles. These apparently conserved functions allow theories of flower origin based on these genes33, 34, 35.


Theories based on evo–devo analysis

The 'mostly male' theory1, 36, 37, 38 was triggered by studies of the LEAFY gene. It has two paralogues in gymnosperms but only one in angiosperms, in which it helps specify the flower. Data from pine suggested that the two gymnosperm paralogues may specify male versus female cones. Angiosperms have lost the latter copy, suggesting that the flower may be derived mostly from the male reproductive unit. At the extreme, the minimal female structure (for example a cupule) might have moved onto the male structure ectopically, creating the antecedent for the carpel bearing angiosperm-type ovules (Fig. 4).

Figure 4: Steps in the mostly male theory.
Figure 4 : Steps in the mostly male theory. Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, or to obtain a text description, please contact

a, Gymnosperm with separately borne microsporophylls (male; left) and cupules (female; right). b, Cupules have moved ectopically onto some microsporophylls. c, Microsporophylls bearing cupules are transforming into carpels and cupules into angiosperm-style ovules (from ref. 38).

High resolution image and legend (93K)

This is supported by other observations: within carpels, ovules have highly variable numbers and placements, some of which must represent ectopic movement and increase in numbers. Stamens, by contrast, are highly uniform. Arabidopsis null mutants of lfy make no stamens but still form carpels, showing that LFY is required for male specification, but LFY independent genes can specify carpels. Ectopic ovules can be generated in Petunia by the overexpression of a single gene and in Arabidopsis by a different mutation, and functional ectopic ovules occur naturally on leaves of some plants of Ginkgo (a gymnosperm) (Fig. 5), suggesting that ectopic ovules are relatively easily produced. Liquid exuded by sterile ovules in Gnetales (Fig. 2) attracts insects to male structures, and pollination droplets on the functional female ovules also attract insects, resulting in pollination. Ectopic ovule placement in male cones in angiosperm ancestors might have conferred an immediate selective advantage by encouraging insect pollination36.

Figure 5: Ginkgo leaves bearing ectopic ovules (and showing autumn colour).
Figure 5 : Ginkgo leaves bearing ectopic ovules (and showing autumn colour). Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, or to obtain a text description, please contact

a, Leaf bearing mature seed at the base of an indentation. b, Leaf with many indentations that have ectopic ovules. (From ref. 37.)

High resolution image and legend (103K)

Recent work on three conifers found complex patterns of expression of the two paralogues, with both being expressed at about equal levels in early female cone development, and both being expressed in early-developing male structures39. This argues against any role in specifying male versus female cones. This undermines the mostly male theory, but preliminary expression data from Welwitschia favours the theory (E. Moyroud and M.W.F., unpublished observations), and the supporting morphological evidence remains.

An intriguing observation is the large number of mutations that can homeotically transform the outer integument of an ovule into a carpelloid structure31, which is consistent with the homology of both of these to a leaf. It might also suggest some ancestral structure resembling cupules borne within a cupule, but no such structure is known among fossil gymnosperms. An alternative would be the spread of some elements of cupule-wall identity from ectopic cupules onto the microsporophyll that bears them, precipitating modification of the latter into a carpel wall.

The old 'gametoheterotopy' theory of Meyen40 remains relevant. It begins with the bisexual reproductive structures of Bennettitales and posits a homeotic conversion that partly imposes morphology of the pollen-bearing organs onto female structures. Some Bennettitalian microsporophylls were flattened with pollen organs on their upper (adaxial) surface. If homeotic conversion resulted in ovules borne on upper surfaces of flat microsporophyll-like structures, then the combination would serve as a carpel precursor. Bennettitales typically have a single whorl of male structures, so homeotic transformation rather than ectopic placement of ovules onto pre-existing structures would be required.

The 'out-of-male/out-of-female' theory of Theissen et al.33, 34 centred on the origin of flower bisexuality. They noted that modern conifers sometimes make bisexual cones. Downregulation of 'B' gene expression in the distal portion of a male gymnosperm cone could permit the tip to become female, or upregulation of 'B' gene expression in basal regions of a female cone could make that region male, generating bisexual reproductive units from either male or female ancestral structures. They suggest that insect pollination could confer an immediate selective advantage, as in the mostly male theory. However, in conifers the resulting cones show normal male and female morphology in both cone regions. There is no novel morphology beyond bisexuality, so the origin of the angiosperm carpel structure is not explained.

Baum and Hileman35 proposed a theory that adds mechanistic detail to the out-of-male theory. They suggested that greatly increased expression of protein encoded by a 'C' gene in the terminal region of a male cone could have been complexed with all the sepallata-encoded protein, preventing its interaction with the protein encoded by the 'B' gene, so switching its developmental fate to female. They also suggested that this 'C' protein might have repressed WUSCHEL, a gene required for maintenance of the apical meristem, resulting in floral determinacy.

The more explicit a theory is, the easier it is to test, so such explicit theories are especially valuable. These recent theories differ from earlier views in the crucial aspect of being testable, not only through the discovery of fossils but also by data from evo–devo studies.


The future

We certainly hope that spectacular palaeobotanical discoveries will clarify flower origins, but failing that it is evo–devo studies that will provide the most important new data, both by suggesting and testing theories of flower origin. Even simple gene-expression data may help in judging whether particular structures are homologous or not41, although such comparisons can be misleading42. Vestigial genes or gene expression patterns may indicate directions of evolutionary change43. Gene trees analysed within organismal trees offer special power for detecting neofunctionalization as opposed to retained (plesiomorphic) gene functions44. All of these results help to limit the range of possible theorizing, which we hope will converge on a historically accurate account of flower origins.

The comparative method applied to morphology and development fuelled the first great advances in evolutionary understanding, and similarly the comparative method applied to DNA sequences revolutionized our understanding of phylogenetic relationships between land plants. The comparative method applied to gene function and genetic controls that determine morphology will vastly increase the power of evo–devo to explain both evolutionary mechanisms and the history of evolutionary change.

Relatively inexpensive 454 Life Sciences45 and Solexa sequencing can detect virtually all mRNAs in a tissue, so expressed genes are known, and microarrays can measure their relative abundances. At present, in non-model organisms, gene function is often assumed to resemble that of closely related genes in model organisms, but improved reverse-genetics methods, such as VIGS (virus-induced gene silencing)46 and TILLING (targeting induced local lesions in genomes), can downregulate genes to demonstrate function directly47.

Phylogenetic footprinting between species identifies conserved non-coding DNA segments that probably have shared protein-binding sites that are important for regulating gene expression48. Segments shared by distant species with similar, homologous morphologies versus segments shared among taxa with differing morphologies should reveal similarities and differences in transcription-factor-binding sites, explaining inputs to gene expression that result in various morphologies. Surface plasmon resonance and other methods can measure equilibrium and kinetics constants for protein–protein49 and protein–DNA interactions50 on short DNA segments and perhaps on promoters of a few thousand base pairs in length; this potentially allows measurement of regulatory outputs of the proteome and inputs to gene expression. In combination, these methods should greatly facilitate elucidation and comparison of genetic control networks in non-model organisms, vastly increasing the power of evo–devo; however, before these studies are available, phylogenetic analyses of individual gene families and expressed sequence tag/microarray studies of whole flowers and floral organs will continue to provide the most useful data, such as those of the Floral Genome Project19.

The appearance in the past decade of theories of flower origin, stimulated by developmental genetic data from modern plants, marks a major shift in attempts to solve Darwin's "abominable mystery". By building a model of the common aspects of floral developmental controls and comparing these with common elements of gymnosperm systems, we can build a picture of the genetic architecture underpinning floral structure in primitive angiosperms and test theories of how floral systems could have arisen19. This could lead to the realization that the fossils we need for understanding angiosperm origins may already be known. Incremental fossil discoveries should allow increasingly complete reconstructions of currently poorly known extinct taxa, which may then be included in phylogenetic analyses, but a palaeobotanical deus ex machina is possible at any time if a fossil is discovered that illustrates intermediate steps in the evolution of critical angiosperm attributes, such as the carpel with its included ovules or the angiosperm stamen with its specialized structure.



  1. Frohlich, M. W. in Developmental Genetics of the Flower (eds Soltis, D. E., Leebens-Mack, J. H. & Soltis, P. S.) 63–127 (Academic, San Diego, CA, 2006)
  2. Crane, P. R., Friis, E. M. & Pedersen, K. R. The origin and early diversification of angiosperms. Nature 374, 27–33 (1995) | Article | ISI | ChemPort |
  3. Soltis, P. S., Endress, P. K., Chase, M. W. & Soltis, D. E. Angiosperm Phylogeny and Evolution (Sinauer, Sunderland, MA, 2005)
  4. Saarela, J. M. et al. Hydatellaceae identified as a new branch near the base of the angiosperm phylogenetic tree. Nature 446, 312–315 (2007) | Article | PubMed | ChemPort |
  5. Rudall, P. J. et al. Morphology of Hydatellaceae, an anomalous aquatic family recently recognized as an early-divergent angiosperm lineage. Am. J. Bot. 94, 1073–1092 (2007)
  6. Friis, E. M., Pedersen, K. R. & Crane, P. R. Cretaceous angiosperm flowers: Innovation and evolution in plant reproduction. Palaeogeogr. Palaeoclimatol. Palaeoecol. 232, 251–293 (2006) | Article |
  7. Rydin, C., Pedersen, K. R., Crane, P. R. & Friis, E.-M. Former diversity of Ephedra (Gnetales): Evidence from Early Cretaceous seeds from Portugal and North America. Ann. Bot. (Lond.) 98, 123–140 (2006) | Article | PubMed |
  8. Friis, E. M., Doyle, J. A., Endress, P. K. & Leng, Q. Archaefructus—angiosperm precursor or specialized early angiosperm? Trends Plant Sci. 8, 369–373 (2003) | Article | PubMed | ChemPort |
  9. Ji, Q. et al. Early Cretaceous Archaefructus eoflora sp. nov. with bisexual flowers from Beipiao, Western Liaoning, China. Acta Geol. Sin. Engl. Edn 78, 883–896 (2004)
  10. Anderson, J. M. & Anderson, H. M. Heyday of the Gymnosperms: Systematics and Biodiversity of the late Trassic Molteno Fructifications (National Botanical Institute, Pretoria, 2003)
  11. Crane, P. R. Phylogenetic analysis of seed plants and the origin of angiosperms. Ann. Mo. Bot. Gard. 72, 716–793 (1985) | Article |
  12. Doyle, J. A. & Donoghue, M. J. Seed plant phylogeny and the origin of angiosperms: an experimental cladistic approach. Bot. Rev. 52, 321–431 (1986) | ISI |
  13. Doyle, J. A. Seed plant phylogeny and the relationships of Gnetales. Int. J. Plant Sci. 157, S3–S39 (1996) | Article | ISI |
  14. Burleigh, J. G. & Mathews, S. Assessing among-locus variation in the inference of seed plant phylogeny. Int. J. Plant Sci. 168, 111–124 (2007) | Article | ChemPort |
  15. Hilton, J. & Bateman, R. M. Pteridosperms are the backbone of seed-plant phylogeny. J. Torrey Bot. Soc. 133, 119–168 (2006) | Article |
  16. Bateman, R. M., Hilton, J. & Rudall, P. J. Morphological and molecular phylogenetic context of the angiosperms: contrasting the 'top-down' and 'bottom-up' approaches used to infer the likely characteristics of the first flowers. J. Exp. Bot. 13, 3471–3503 (2006) | Article | ChemPort |
  17. Doyle, J. A. Seed ferns and the origin of angiosperms. J. Torrey Bot. Soc. 133, 169–209 (2006) | Article |
  18. Doyle, J. A. & Endress, P. K. Morphological phylogenetic analysis of basal angiosperms: Comparison and combination with molecular data. Int. J. Plant Sci. 161, S121–S153 (2000) | Article | ISI | ChemPort |
  19. Soltis, D. E. et al. The evolving floral genome: A history of genome-wide duplications and shifting patterns of gene expression. Trends Plant Sci. 12, 358–367 (2007) | Article | PubMed | ChemPort |
  20. Taylor, D. W. et al. Biogeochemical evidence for the presence of the angiosperm molecular fossil oleanane in Paleozoic and Mesozoic non-angiospermous fossils. Paleobiology 32, 179–190 (2006) | Article |
  21. Auras, S. et al. Aromatized arborane/fernane hydrocarbons as biomarkers for Cordaites. Naturwissenschaften 93, 616–621 (2006) | Article | PubMed | ChemPort |
  22. Stebbins, G. L. Flowering Plants Evolution above the Species Level (Harvard Univ. Press, Cambridge, MA, 1974)
  23. Retallack, G. & Dilcher, D. L. Arguments for a glossopterid ancestry of angiosperms. Paleobiology 71, 54–67 (1981)
  24. Friedman, W. E. & Williams, J. H. Modularity of the angiosperm female gametophyte and its bearing on the early evolution of endosperm in flowering plants. Evolution Int. J. Org. Evolution 57, 216–230 (2003) | PubMed |
  25. Hughes, N. F. 1994. The Enigma of Angiosperm Origins (Cambridge Univ. Press, Cambridge, 1994)
  26. Krassilov, V. A. Angiosperm Origins: Morphological and Ecological Aspects (Pensoft Publishers, Sofia, Bulgaria, 1997)
  27. Litt, A. Evaluation of A-function: evidence from the APETALA1 and APETALA2 gene lineages. Int. J. Plant Sci. 168, 73–91 (2007) | Article | ChemPort |
  28. Albert, V. A., Gustafsson, M. H. G. & Di Laurenzio, L. in Molecular Systematics of Plants II: DNA Sequencing (eds Soltis, D. E., Soltis, P. S. & Doyle, J. J.) 349–374 (Kluwer Academic, Dordrecht, 1998)
  29. Kramer, E. M., Jaramillo, M. A. & Di Stilio, V. S. Patterns of gene duplication and functional evolution during the diversification of the AGAMOUS subfamily of MADS box genes in angiosperms. Genetics 166, 1011–1023 (2004) | Article | PubMed | ChemPort |
  30. Zahn, L. M. et al. Conservation and divergence in the AGAMOUS subfamily of MADS-box genes: evidence of independent sub- and neofunctionalization events. Evol. Dev. 8, 30–45 (2006) | Article | PubMed | ChemPort |
  31. Skinner, D. J., Hill, T. A. & Gasser, C. S. Regulation of ovule development. Plant Cell 16, S32–S45 (2004) | Article | PubMed | ChemPort |
  32. Buzgo, M., Soltis, P. S. & Soltis, D. E. Floral developmental morphology of Amborella trichopoda (Amborellaceae). Int. J. Plant Sci. 165, 925–947 (2004) | Article |
  33. Theissen, G. et al. in Developmental Genetics and Plant Evolution (eds Cronk, Q. C. B., Bateman, R. M. & Hawkins, J. A.) 173–206 (Taylor & Francis, London, 2002)
  34. Theissen, G. & Becker, A. Gymnosperm orthologues of class B floral homeotic genes and their impact on understanding flower origin. Crit. Rev. Plant Sci. 23, 129–148 (2004) | Article | ChemPort |
  35. Baum, D. A. & Hileman, L. C. in Flowering and its Manipulation (ed. Ainsworth, C.) 3–27 (Blackwell, Oxford, 2006)
  36. Frohlich, M. W. & Parker, D. S. The Mostly Male theory of flower evolutionary origins: from genes to fossils. Syst. Bot. 25, 155–170 (2000) | Article | ISI |
  37. Frohlich, M. W. in Developmental Genetics and Plant Evolution (eds Cronk, Q. C. B. & Bateman, R. M. & Hawkins, J. A.) 85–108 (Taylor & Francis, London, 2002)
  38. Frohlich, M. W. An evolutionary scenario for the origin of flowers. Nature Rev. Genet. 4, 559–566 (2003) | Article |
  39. Vázquez-Lobo, A. et al. Characterization of the expression patterns of LEAFY/FLORICAULA and NEEDLY orthologs in female and male cones of the conifer genera Picea, Podocarpus and Taxus: implications for current evo-devo hypotheses for gymnosperms. Evol. Dev. 9, 446–459 (2007) | PubMed |
  40. Meyen, S. V. Origin of the angiosperm gynoecium by gametoheterotopy. Bot. J. Linn. Soc. 97, 171–178 (1988)
  41. Kim, S. et al. Sequence and expression studies of A-, B-, and C-class Mads-box homologues in Eupomatia (Eupomatiaceae): support for the bracteate origin of the calyptra. Int. J. Plant Sci. 166, 185–198 (2005) | Article | ChemPort |
  42. Jaramillo, M. A. & Kramer, E. M. The role of developmental genetics in understanding homology and morphological evolution in plants. Int. J. Plant Sci. 168, 61–72 (2007) | Article | ChemPort |
  43. Long, J. & Barton, M. K. Initiation of axillary and floral meristems in Arabidopsis. Dev. Biol. 218, 341–353 (2000) | Article | PubMed | ISI | ChemPort |
  44. Causier, B. et al. Evolution in action: Following function in duplicated floral homeotic genes. Curr. Biol. 15, 1508–1512 (2005) | Article | PubMed | ISI | ChemPort |
  45. Margulies, M. et al. Genome sequencing in microfabricated high-density picolitre reactors. Nature 437, 376–380 (2005) | Article | PubMed | ISI |
  46. Burch-Smith, T. M., Anderson, J. C., Martin, G. B. & Dinesh-Kumar, S. P. Applications and advantages of virus-induced gene silencing for gene function studies in plants. Plant J. 39, 734–746 (2004) | Article | PubMed | ISI | ChemPort |
  47. David-Schwartz, R. & Sinha, N. Evolution and development in plants: bridging the gap. Int. J. Plant Sci. 168, 49–59 (2007) | Article | ChemPort |
  48. De Bodt, S., Theissen, G. & Van de Peer, Y. Promoter analysis of MADS-box genes in eudicots through phylogenetic footprinting. Mol. Biol. Evol. 23, 1293–1303 (2006) | Article | PubMed | ChemPort |
  49. Boozer, C. et al. Looking towards label-free biomolecular interaction analysis in a high-throughput format: a review of new surface plasmon resonance technologies. Curr. Opin. Biotechnol. 17, 400–405 (2006) | Article | PubMed | ISI | ChemPort |
  50. Nalefski, E. A., Nebelitsky, E., Lloyd, J. A. & Gullans, S. R. Single-molecule detection of transcription factor binding to DNA in real time: Specificity, equilibrium, and kinetic parameters. Biochemistry 45, 13794–13806 (2006) | Article | PubMed | ChemPort |
  51. Pryer, K. M. et al. Horsetails and ferns are a monophyletic group and the closest living relatives to seed plants. Nature 409, 618–622 (2001) | Article | PubMed | ISI | ChemPort |
  52. Klavins, S. D., Taylor, T. N. & Taylor, E. L. Anatomy of Umkomasia (Corystospermales) from the Triassic of Antarctica. Am. J. Bot. 89, 664–676 (2002)
  53. Moore, M. J. et al. Using plastid-genome-scale data to resolve enigmatic relationships among basal angiosperms. Proc. Natl. Acad. Sci. USA 104, 19363–19368 (2007) | Article | PubMed | ChemPort |


M.W.F. thanks the National Science Foundation (USA) for supporting work in this area. We thank J. A. Doyle, E.-M. Friis, P. S. Soltis, R. M. Bateman, P. Kenrick, D. E. Soltis and J. Hilton for commenting on the manuscript, and J. Trager and Huntington Gardens for Welwitschia materials.

  1. Royal Botanic Gardens Kew, Richmond, Surrey TW9 3DS, UK

Correspondence to: Michael W. Frohlich1 Correspondence should be addressed to M.W.F. (Email:


These links to content published by NPG are automatically generated.


Botany The family tree flowers

Nature News and Views (25 Nov 1999)

Time for the angiosperms

Nature News and Views (16 Dec 1993)

See all 7 matches for News And Views