I also stated that there is growing consensus among some molecular systematists and paleobotanists on the existence of a 160 million year old angiosperm ghost lineage rooted at the angiosperm-gymnosperm split, 300 million years ago, prior to the end-Permian extinction.

Insights into the rapid radiation of crown group flowering plants from the angiophyte stem group based on contrasting phylogenetic research perspectives are available in several reviews (J. A. Doyle and Donoghue 1993, Magallón and Castillo 2009, Stephen A. Smith et al. 2010). While I disagree with some of the proposals in a couple of these reviews, some of the ideas are open to debate.

Pictured to the right is a flower of Protea compacta (Proteaceae, Proteales, Proteanae) photographed by the author.

The Cretaceous Yixian Formation of Asia yields potentially interesting fossilized inflorescences, flowers, seed, and pollen of crown group angiosperms including Asiatifolium, Jixia, Shenkuoa, and Xingxueina (G. Sun et al. 2001). Recent research findings on the Cretaceous Yixian Formation and other strata from northeastern Asia are reviewed by Ge Sun et al. (2008).

Archaefructus liaoningensis (Magnoliophyta, Magnoliopsida, Archaemagnoliidae), has been described (G. Sun and Dilcher 1997) and discussed within the context of a Jurassic aquatic origin of flowering plants (G. Sun et al. 1998). Radioisotope decay data suggest that the Yixian Formation is much younger than originally believed by Ge Sun and colleagues (see Friis et al. 2006).

Phylogenetic placement of Archaefructus as a stem group flowering plant is problematic (Friis et al. 2003, J. A. Doyle 2008). Several workers suggest that the angiosperm fossils from the Cretaceous Yixian Formation are better placed with the crown group of eudicot angiosperms.

Hyrcantha decussata (Sinocarpus decussatus, Leng and Friis 2006) is one of the controversial angiosperm-like fossils recovered from northeastern Asian Cretaceous sediments (Dilcher et al. 2007). However, it is increasingly unlikely that any of the aforementioned taxa belong in the angiosperm stem group.

Fossilized flowers are also known from exposed Cretaceous sediments of Antarctica (Eklund 2003), Asia (G. Sun et al. 1998, Poinar and Chambers 2005), North America (Crane and Herendeen 1996), South America (Endressinia brasiliana, Mohr and Bernardes-de-Oliveira 2004), and elsewhere (Friis et al. 2006).

Another interesting but incomplete Cretaceous North American angiosperm fossil is Archaeanthus linnenbergerii (Dilcher and Crane 1985). The morphology of pollen-bearing organs and stem anatomy of Archaeanthus is unknown, therefore critical comparison with Afropollis pollen and fossil wood of Cretaceous Winteraceae from Antarctica (Poole and Francis 2000) is impossible.

"It seems the perpetuation of the 'abominable mystery' is due more to a disagreement over the cladistic position of various fossil taxa in relationship to the angiosperm crown group, than the lack of data."

The preceding quotation is from page 127 of M. S. Zavada (2007), The identification of fossil angiosperm pollen and its bearing on the time and place of the origin of angiosperms, Plant Systematics and Evolution 263: 117-134.

The most recent flowering plant classification system (Chase and Reveal 2009) will require revision to accommodate results of ongoing molecular phylogenetic research. Some workers (D. E. Soltis and P. S. Soltis 2003) suggest abandoning the angiosperm subclass system enthusiastically followed by Arthur Cronquist (1981) and others, but this idea is too inflexible.

A temporary fix using the classification level of superorder is proposed by Chase and Reveal (2009). Some Cretaceous angiosperm and mesoangiosperm fossils do not fit in Chase and Reveal's superorders, therefore an informal approach is warranted here.

Tables 7-13 are organized along the lines adopted in T. N. Taylor et al. (2009) with consideration of updated concepts introduced by Chase and Reveal (2009).

Cretaceous explosive radiation of angiosperms. Rapid radiation of flowering plants during the Albian Age of the early Cretaceous Period is demonstrable based on paleontologic evidence gathered to date (Friis et al. 2006, T. N. Taylor et al. 2009, Friis et al. 2010), molecular evidence from duplication and neofunctionalization of MIKC-type MADS-box genes (Irish 2006), and phylogenetic analyses (D. E. Soltis et al. 2008, Endress and J. A. Doyle 2009).

Diversification of flowering plants might have been attributable to decline in atmospheric carbon dioxide levels during the Cretaceous Period of the Mesozoic Era instead of developing aridity according to McElwain et al. (2005).

Clues to angiosperm diversification may be gleaned from studies of xylem heterochrony (Carlquist 2009).

Was the Big Bang of angiosperm evolution during the Aptian Age (Gallic Epoch) attributable to the effects of the end-Barremian Age (Neocomian Epoch) biogeochemical perturbation (BaCCE) on coevolving angiosperm hosts and insect antagonists?

Mounting geophysical evidence points toward extensive episodes of Cretaceous (Aptian) volcanism in the southwest Pacific Basin near the undersea Ontong Java Plateau (Tejada et al. 2009). Methane clathrates released from undersea beds could explain significant spikes in atmospheric methane and carbon dioxide, triggering the BaCCE and later carbon cycle anomalies of the Cretaceous Period.

Heimhofer et al. (2005) suggest that the BaCCE might have accelerated the diversification of early magnoliid flowering plants and possibly monocots. Phylogenetic analysis supports the idea that an explosive radiation of the Malpighiales sensu APG II occurred during the Aptian Age of the Gallic Epoch, a few million years after the BaCCE (Davis et al. 2005, Figure 1).

In 1981, Friis and Skarby reported a remarkable find of abundant indeterminate and tiny angiosperm eudicot flowers from the Late Santonian-Early Campanian Age (Senonian Epoch, Late Cretaceous Period). Friis and Skarby (1981), Friis (1985), Crane et al. (1986), Knobloch and Mai (1986), Schönenberger and Friis (2001), Eklund (2003), Kvaček and Eklund (2003), and Friis et al. (2006) are key pieces of the scientific literature on early angiosperm floras and fossils.

Studies of Cretaceous permineralized woods (V. M. Page 1967) paint a different picture of Maastrichtian forests once thought to be dominated by gymnosperms. A recent study of wood permineralizations sampled from large, fallen in situ logs from southwestern North American Maastrichtian (Senonian) deposits of the Aguja and Javelina Formation suggests that "dicotyledonous" trees were more abundant than conifers. Petrified stumps of Javelinoxylon multiporosum (Malvales, Dilleniidae) were more than a meter in diameter with extrapolated tree axes up to 40 meters tall (Wheeler and Lehman 2000).

Existence of dilleniid Javelinoxylon trees belonging to derived crown group eudicots as an important floristic element of late Cretaceous stratified tropical forests of southwestern North America (Wheeler and Lehman 2000), and large trees of Paraphyllanthoxylon indigenous to European forests of that time (Oakley and Falcon-Lang 2009), detract from the idea that early flowering plants were paleoherbs of upland habitats.

Basal angiosperms and magnoliids. The floral biology and evolution of basal angiosperms is reviewed by Peter Endress (2010). The Angiosperm Phylogeny Group's (APG) recent update (APG III 2009) includes a "working classification" (pages 123 and 124, Table 1, Chase and Reveal 2009), which I have adopted in this third of three essays on the origin of flowering plants.

A benchmark phylogenetic study by Y.-L. Qiu (2000) established a group of extant flowering plant families as basal to all other living angiosperms. Chase and Reveal (2009) recognize four superorders of basal angiosperms and magnoliids: Amborellanae, Austrobaileyanae, Magnolianae, Nymphaeanae. Chloranthales are unplaced (Chase and Reveal 2009).

Certain ANITA-grade basal flowering plants first appear in the fossil record of the Cretaceous Period (Friis et al. 2000, Friis et al. 2001, Krassilov and Golovneva 2004, Takahashi et al. 2007, von Balthazar et al. 2008). The fossil history of basal angiosperms is reviewed by Friis et al. in three reviews (2006, 2009, 2010).

On the left is a picture of a flower of Degeneria roseiflora (Degeneriaceae, Magnoliales, Magnolianae), and several fragrant flower buds at different stages of maturity. Two of the largest flower buds shown on this kodachrome opened one-by-one on the next two successive nights, releasing a rose-like fragrance (photographed by the author). Degeneriaceae are related to Winteraceae and Magnoliaceae (A. C. Smith 1981, J. M. Miller 1988, A. C. Smith 1991).

D. E. Soltis et al. (2005) support the assignment of ANITA-grade angiosperms to a basal position in several calibrated, bootstrap-supported, molecular-based phylogenies of extant flowering plants. The acronym ANITA is composed of first letters from the taxa Amborella, Nymphaeales, Illiciaceae, Trimeniaceae, and Austrobaileyaceae.

Since 2005 phylogenetic analyses of extant basal flowering plants point to the Amborellaceae, Nymphaeales, and Austrobaileyales (ANA) as sister groups to all other basal magnoliids, monocots, eudicots, core eudicots, rosids, and asterids (Leebens-Mack et al. 2005, Y.-L. Qiu et al. 2006, Frohlich and Chase 2007, D. E. Soltis et al. 2008, Magallón and Castillo 2009). Reevaluation of gymnosperm molecular sequences when subjected to phylogenetic analysis (Y.-L. Qiu et al. 2001) support the basal position of ANITA grade angiosperms with respect to magnoliids, monocots, and eudicots.

Basal flowering plants sensu Cronquist (1981) belong to Amborellaceae and Trimeniaceae (Laurales), Austrobaileyaceae (Magnoliales), Hydatellales (monocots), Illiciales, and Nymphaeales.

Certain other Magnoliales once regarded by Cronquist (1981) as primitive angiosperms (e.g. Degeneriaceae, Magnoliaceae, and Winteraceae) are in a more derived position on phylogenetic trees based on molecular data (Wikström et al. 2001, Leebens-Mack et al. 2005, Y. L Qiu et al. 2006, D. E. Soltis et al. 2007) than ANITA-grade taxa.

Tremendous progress has been made on understanding the anatomy, basic biology, organelle genetics, developmental genetics, ecology, floral genetics, molecular evolution, morphology, natural history, and phylogenetic systematics of ANITA grade basal angiosperms (Parkinson et al. 1999, J. A. Doyle and Endress 2000, Endress and Igersheim 2000, Floyd and Friedman 2000, S. W. Graham and Olmstead 2000, Mathews and Donoghue 2000, Y. L Qiu et al. 2000, P. S. Soltis et al. 2000, and Thien et al. 2000).

Work on ANITA grade basal angiosperms within the last decade is published by Endress (2001), Furness and Rudall (2001), Friedman (2001), Yamada et al. (2001), Borsch et al. (2003), Feild et al. (2003), Ronse De Craene et al. (2003), E. L. Schneider et al. (2003), Zanis et al. (2003), Aoki et al. (2004), Endress (2004), Stefanovič et al. (2004), Sage et al. (2009), P. S. Soltis et al. (2009), and Thien et al. (2009).

Despite elucidation of MIKC-type MADS-box gene function in several groups of flowering plants (Irish 2006), the importance of A-function in basal angiosperms is unclear (D. E. Soltis et al. 2007). Recent molecular phylogenetic work has been published by Carpenter (2005), S. Kim et al. (2005), Leebens-Mack et al. (2005), Podoplelova and Ryzhakov (2005), Y. L Qiu et al. (2005), Zahn et al. (2005), Carpenter (2006), P. S. Soltis et al. (2006), Endress and J. A. Doyle (2007), Feild and Arens (2007), Moore et al. (2007), D. E. Soltis et al. (2007), Rudall et al. (2008), Endress (2008), and P. S. Soltis et al. (2009).

Zahn et al. (2005) report class E MADS-box gene homologs of SEPALLATA (SEP) from basal angiosperms. The molecular phylogenetic studies by Zahn et al. (page 2220, 2005), "strongly supports the hypothesis that the SEPALLATA subfamily has been present in the angiosperms since before the diversification of the extant basal-most angiosperms Amborella and the Nymphaeales."

Pollination ecology of primitive flowering plants has the subject of much field research and debate over the past 20 years, even before the term basal was widely used by plant systematists. Certain beetles (Coleoptera) are considered pollinators of basal angiosperms but the list also includes several groups of flies (Diptera), moths (Lepidoptera), and wasp-like bees (Hymenoptera). A review of some of these studies (Bernhardt et al. 2003) adds considerable controversy to the role of stigmatic secretions as pollinator attractants, and calls into question character polarity of wet-type stigmas determined by Endress and Igersheim (2000).

Liliales. Records of Mesozoic fossils referable to the order Liliales consist mainly of pollen casings (palynomorphs) preserved in sediments (J. Muller 1981).

A study by Ramírez et al. (2007) sheds light on the time of origin of a specific orchid and its pollinator with implications toward the phylogeny and origin of monocots. Lily flowers express APETALA1-like MIKC-type MADS-box genes as reported by M.-K. Chen et al. (2008).

Commelinid monocots. According to APG III (Chase and Reveal 2009), commelinid monocots consist of four orders (Arecales, Commelinales, Poales, Zingiberales) and one unplaced family (Dasypogonaceae).

The picture to the left is a flower of Strelitzia nicolai (Strelitziaceae, Zingiberales, Lilianae) from The University of the South Pacific Botanical Garden, Suva, Fiji (photographed by the author in 1988).

Arecales. The order Arecales contains several economically important palms used as a source of fiber (copra), food, and oil (Cocos nucifera, Elaeis guineensis, Copernicia, among others).

A review of palms of India is available (Ramanujam 2004). Additional details on the fossil history of certain palm groups may be gleaned from Uhl and Dransfield (1987). Fossil pollen belonging to this subclass may be found in sediments as old as the early Cretaceous Period.

Fabids were evidently common in Cretaceous floras including evolutionary lines leading to Paleogene and Neogene species of beeches (including Nothofagus, a favorite of biogeographers), birches, hazelnuts, oaks, and walnuts (Friis et al. 2006, Manos et al. 2007, Golovneva 2008, Narita et al. 2008, Oh and Manos 2008, Tschan et al. 2008, X. Wang 2008, Y.-H. Wang et al. 2009).

Several problematic morphotype (form) genera of detached leaves, and pollen from the Maastrichtian Age possibly belonging to fabids are reported in the literature (Manchester 1987).

Table 12 outlines the Mesozoic fossil history, at least of definitive reproductive remains of the fabid clade of rosids.

The photograph above is Castilleja cinerea (Plantaginaceae, Lamiales, Asteranae) from the San Bernardino Mountains of southwestern North America.

Extant Asteranae are being explored as research subjects to shed light on coevolutionary questions from the research perspective of population ecology, such as shifting pollinators (P. Wilson et al. 2007, S. D. Smith et al. 2008).

Table 13 is the Mesozoic fossil history of asterids.

There are great gaps in our understanding of the fossil history of flowering plants based on data recorded in Tables 7-13 and a more detailed review by T. N. Taylor et al. (2009). Paleontologic data reveal several general trends but due to insufficient sampling it is too soon to make any definitive statements on the origin, paleobiodiversity, and evolution of angiosperms.

Paleobotanical data in the preceding tables often consist of a single specimen from one isolated locality (sometimes only a single, tiny charcoalified flower or seed), and therefore, can no way support assertions of a "Big Bang," "explosive," or "first" radiation of angiosperms in early Cretaceous paleoenvironments. Considerably more field work is needed with possible focus on outcrops older in geologic age.

Coevolution between phytophagous insect antagonists and Carboniferous, Permian, and Triassic seed plant hosts at the level of their respective developmental tool kits and homeotic gene cis-regulatory modules was likely. I reject the widely-held belief of a late Jurassic-early Cretaceous origin of flowering plants. This view is supported by recent molecular phylogenetic analysis of nucleic acid data suggesting a late Triassic (Norian) age of the flowering plant crown group (Stephen A. Smith et al. 2010).

Further, our knowledge of carpel, floral, and ovular transcriptional regulators in extant angiosperm model organisms does not preclude derivation of evo-devo models that explain curling, inrolling, and fusion in 260- to 300 million-year-old spermopteroid Phasmatocycas bridwellii leaves to form carpels, ovaries, and pistils.

Therefore, 160 million years of neotenic evolution to include condensation of hypothetical gigantopteroid proanthostrobili is the most simple evo-devo process to explain the origin of reproductive organs in Mesozoic crown group angiosperms and extant basal Amborellanae, Austrobaileyanae, Nymphaeanae, and Magnolianae.

Adaptive radiation within the major clades of eudicots, rosids, and asterids during the Cretaceous Period is evident from paleontological data summarized by Crepet et al. (2004), Friis et al. (2006), D. E. Soltis et al. (2008), and T. N. Taylor et al. (2009).

Is the albeit sometimes asynchronous diversification and adaptive radiation of the angiosperm crown group and coevolving insect clades attributable to delayed climatic effects of the early Cretaceous end-Barremian biogeochemical perturbation (BaCCE) and other global carbon cycle anomalies triggered by bolide impacts, global warming, undersea volcanism, plate tectonics, and/or island arc orogenesis?

Clues from our redoubled efforts to excavate ("mining the rock record") and to painstakingly study coalified, compressed, permineralized, petrified, and preserved fossil plant material (page 249, E. L. Taylor and T. N. Taylor 2009), to better understand the anatomy, biology, and morphology of coevolving colonies of holometabolous insect antagonists, and to reconstruct whole protoflowers, might help us solve the riddle of angiosperm beginnings within an evolutionary framework.

Interestingly, pteridosperms represented by the Paleogene fossil Komlopteris cenozoicus, survived the K-T extinction radiating with angiosperms and modern conifers in the forests of Tasmania (McLoughlin et al. 2008). Discovery of a population of Cenozoic Corystospermales detracts from proposals such as MMT (Frohlich 2002) on the origin of flowering plants and the suggestion by Hilton and Bateman (2006), that "pteridosperms are the backbone of seed plant evolution."

Botanists should conduct field surveys and search the museum shelves for unidentified herbarium specimens recording a surviving remnant of a Komlopteris population. This endeavor may be equal in importance to redoubling our efforts to mine and describe early Mesozoic and Paleozoic seed plants, as it would be very important to gain knowledge of homeotic transcriptional regulation in a corystosperm seed fern to compare with other extant vascular plant model organisms.

Are there any extant corystosperms "lurking there"?

Further, insect-seed plant interactions affected by temperature extremes and global hypoxia may have led to diversification at the molecular level in seed plant and holometabolous insect lineages, formation of paleoflowers from bisexual cone axes, coevolutionary development of reproductive modules, and moulting novelties in insect larvae leading to evo-devo of the adult insect head, thorax, and abdomen.

The origin of angiosperms and certain clades of phytophagous insect antagonists is potentially a consequence of coevolution of animal and seed plant cis-regulatory modules and developmental tool kits. Further, I suggest that phytoecdysones secreted by Permo-Carboniferous and Permo-Triassic shrub lifeboats potentially affected body size and moulting time of phytophagous insect antagonists.