New ideas on a Paleozoic origin of the flower will require a deeper understanding of selective pressures on molecular evolution of the helix-turn-helix DNA-binding motif of homeotic enzymes, based on structural studies and x-ray crystallography of LFY protein cis-acting developmental switches (Hamès et al. 2008).

"Our theory may also be tested, without the vagaries of fossil discovery, in modern plants. Of the early-acting genes that control flower development, more should have close homologs (or orthologs, if gene trees are sufficiently resolved to demonstrate orthology) active in male gymnosperm reproductive structures rather than in female structures... In the coming years, topology-violating transformations that are of evolutionary importance may become recognizable using evidence from genes that control development."

The above quotation is from page 167 of M. W. Frohlich and D. S. Parker (2000), The mostly male theory of flower evolutionary origins: from genes to fossils (MMT), Systematic Botany 25(2): 155-170.

Frohlich suggests in two updates of MMT (2002, 2003) that early Mesozoic Corystospermales might be one of the possible ancestors of flowering plants. Frohlich (2003) placed an interesting idea on the table, namely that potential pollinators of gymnosperms were attracted by ectopic ovules displayed on upper (abaxial) leaf surfaces.

For example, none of the work on Permian gigantopterids published to date convincingly demonstrates physical connections between reproductive structures, stems, roots, and leaves. Whole plant morphologies are completely unknown for this group of Paleozoic seed plants.

With possible insect and seed plant co-cladogenesis in mind, I now present a coevolutionary/thigmomorphogenetic hypothesis on the origin of flowering plants:

Coevolutionary/Thigmomorphogenetic Hypothesis:

During the hypoxic late Devonian-early Carboniferous ice-house (DeCARB), invertebrate Hox proteins, hemocyanin respiratory enzymes, and certain nuclear receptors of plecopteran (stonefly) ancestors diverged, leading to novel multimeric proteins in several hemimetabolous and holometabolous clades of late Paleozoic insects.

Candidates for the early insect developmental tool kit may also include certain homeotic selector genes of the Hox complex such as homologs and paralogs of abdominal A (abd-A), Abdominal B (Abd-B), Hox3, proboscipedia (pb), sex combs reduced (Scr), and Ultrabithorax (Ubx), the compartment-selector gene engraled; also including the proteins these genes encode such as Ubx and engraled, the ecdysone hormone signaling machinery, hemocyanins, hexamerins, and nuclear receptor proteins.

At the same time that homeotic genes and transcription factors were diverging in early Carboniferous insect clades, key vascular plant developmental genes including AP2, ARF, Class I and II TCP, Class III HD-Zip, GRAS, KNOX, LEAFY, MIKC-type MADS-box, PIN, TIR, and YABBY, and their modular protein transcription factors, were evolving into molecular novelties in Paleozoic seed plant lineages.

Coevolution of insect helix-turn-helix homeodomain proteins, seed plant homeotic helix-turn-helix LEAFY (LFY) proteins, other phytophagous insect- and seed plant host developmental cis-regulatory modules, and their respective developmental tool kits, potentially occurred during the DeCARB and/or end-Guadalupian carbon cycle event (GuCCE), with local hypoxia and temperature extremes as underlying and potentially powerful selective forces.

Global oxygen levels declined to 13% during the course of millions of years of the DeCARB. Resident chewing, crawling, ovipositing, piercing, and sucking insects probably became dependent on host seed plant shrubs and trees for food, oxygen gas; and shelter from predators, cold temperatures, and ultraviolet radiation. Phytophagous insect associates of Permo-Carboniferous vegetation potentially coevolved with their oxygen-generating plant hosts.

The GuCCE and associated global hypercapnia reinforced the already close association between Permian plants and protocoleopterans, weevils, and other hemi- and holometabolous insects, ostensibly established after the DeCARB. Resident phytophagous insect associates of developmentally plastic Permian seed plant shrubs including populations and species of Gigantopteridales and Vojnovskyales, allowed stressed hemimetabolous insect mutualists to survive hot, hypoxic environments.

Permo-Triassic seed plants were probably confined to small isolated populations indigenous to fragmented biomes and fitness landscapes resulting from the deleterious effects of the end-Permian extinction (EPE). Surviving insects of the EPE and Triassic-Jurassic boundary carbon cycle event (TrCCE) fed, reproduced, and might have secreted biologically active molecules while trampling the surfaces of shoot apical meristems (SAMs) and accessory bisexual cone axes.

Signaling molecules such as jasmonates and volicitins originating in the bodies of phytophagous insect associates of Permo-Triassic shrubs and trees, when applied to growing points of the host plant with mechanical force might have affected biosynthesis of certain phytohormones including gibberellic acid (GA). Transduction of biologically active insect signaling molecules potentially stimulated biosynthesis of GA, activating, modifying, and upregulating homeotic selector and meristem identity genes affecting developmental switches in expanding bisexual cone axes.

Phytoecdysones originating in tissues of Carboniferous and Permian seed plants could have signalled the hemocyanin respiratory and hexamerin storage protein manufacturing machinery of eggs and instars of phytophagous insect antagonists profoundly affecting their development and body size.

Permo-Triassic shrubs and trees ostensibly developed anatomical and biochemical defenses against both invertebrate and vertebrate herbivores possibly ameliorating mechanical damage to SAMs and accessory fertile meristems. Dependence of insect colonies on seed plant shrubs and trees might have been reinforced by paleoclimatic changes triggered by global catastrophe.

Paleozoic megasporophylls described by Axsmith et al. (2003) as Phasmatocycas bridwellii were probably detached female pieces of much larger 300 million-year-old flower-like hermaphroditic organs. Further, no "assembly" (page 816, J. A. Doyle 2008) of perianth parts, microsporophylls, and megasporophylls was needed during early stages of angiosperm evolution prior to the end-Permian extinction.

Bisexual cone axes consisting of inner (terminal) Phasmatocycas bridwellii megasporophylls and an outer ring of microsporophylls subtended by a spirally-arranged perianth of tepals and colored andropetals, hereinafter termed "gigantopteroid paleoflowers," potentially existed in populations of poorly understood Paleozoic seed plants described as gigantopterids and Vojnovskyales, groups omitted from most phylogenetic analyses.

I hypothesize that insect-mediated natural hybridization between gigantopterids and certain Vojnovskyales might have been a method through which MIKC-type MADS-box genetic novelties were recruited into populations of ancestral angiosperms during the Permo-Carboniferous approximately 300 million years ago.

Enormous paleoflowers known only from fragments and detached megasporophylls were probably modifications of developmentally plastic bisexual cone axes representing divergent clades of several Permian gymnosperms that survived the end-Permian apocalypse.

Carpel, floral, and ovular transcriptional regulation 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 ovule-bearing gigantopteroid Phasmatocycas bridwellii leaves to form carpels, ovaries, and pistils.

Neoteny to include condensation of hypothetical gigantopteroid paleoflowers is the most simple evo-devo process to explain the origin of reproductive organs in Mesozoic crown group angiosperms and extant basal Amborellaceae, Nymphaeales, and Magnoliales from a putative 160 million year old ghost lineage.

Reproductive traits of certain developmentally plastic Paleozoic seed plants including early Mesozoic flowering plants are ostensibly intertwined with 200 to 300 million year-old gene duplications and subsequent molecular evolution of cis-acting floral protein quartets essential to meristem, organ, and reproductive module identity.

At the same time that gene duplications led to the evolution of floral quartets in Paleozoic shrub lifeboats, certain phytophagous insect associates underwent a second burst of molecular evolution of hexamerin storage proteins. Molecular evolution comprising the second burst of hexamerin diversification led to Coleoptera (bees), Diptera (flies), Hymenoptera (bees and wasps), and certain Lepidoptera (moths).

Moulting in developmentally plastic insect mutualists of Permian seed plant shrub- and tree-like lifeboats was probably affected by phytoecdysone hormones manufactured by the host seed plant. Signaling of respiratory protein genes of the insect genome by the plant, and mechanical and secretory activities of insect antagonists leading to signaling of host plant homeotic genes in SAMs, were likely hallmarks of Permo-Carboniferous and Permo-Triassic coevolutionary compartments.

Enveloped and surrounded by an acidified hypoxic desert, shrub lifeboats were venues where reciprocal, simultaneous, and specialized communication between invertebrate and plant took place at the molecular level. Insect hexamerins might have been hypoxia-inducible moulting storage proteins of the developmental tool kit of the larger-bodied insects of Carboniferous and Permian time. Insect bodies became smaller in plant-dwelling populations that survived the EPE and TrCCE.

Innovative traits of Permo-Triassic seed plants possibly included double fertilization, encapsulation of female gametophytes inside hardened ovules, folding of megasporophylls into carpels, and development of staminodes and stamens from laminar microsporophylls, but critical fossil evidence is needed. Populations of Permo-Triassic seed plants were ostensibly subjected to the vagaries of genetic drift, gene flow, meiotic drive, mutation, and natural selection of phenotypes resulting from developmental recombination, genetic accommodation, and gene duplications potentially leading to evolutionary change and possible angiospermization.

Selection pressures in populations of Carboniferous and Permo-Triassic seed plant populations were probably much greater than believed. An herbaceous origin of flowering plants cannot be explained by mutualism and coevolution of insects and seed-bearing shrub hosts alone. A Late Jurassic aquatic origin for flowering plants is incongruent with likely deleterious effects of sulfuric acid poisoning of lakes, shorelines, and wetlands, caused by basalt outpouring from the Central Atlantic Magmatic Province (CAMP), which is associated with the Triassic-Jurassic boundary carbon cycle event (TrCCE).

Assuming coevolution of phytophagous coleopterans with angiosperms, various proposals on rapid diversification of flowering plants during the Albian Age of the Gallic Epoch of the early Cretaceous Period are inconsistent with molecular-based phylogenies of Coleoptera that suggest a Triassic origin of certain non-chrysomelid beetle lineages. Concomitantly, existence of a 160 million-year-old angiosperm ghost lineage is likely. Therefore, extant basal flowering plants such as Amborella trichopoda are probably derived species.

Further, the origin of flowering plants and their coevolving insect hosts is potentially traceable to hybridization between unrelated gigantopterid and vojnovskyalean Carboniferous and Permian seed plant lineages about 300 million years ago. When supported by future phylogenetic analyses a monophyletic origin of flowering plants becomes less likely.

When taking into account the cyclic nature of angiospermization, flowering plants as traditionally defined, are a loose amalgam of parallel evolutionary lines traceable to surviving geographically disparate early Triassic remnants of already divergent Permian seed plant lineages.

Discussion of a Coevolutionary/Thigmomorphogenetic Hypothesis:

I now consider and discuss evidence drawn from studies of insect- and plant anatomy and development, ecology, evolution, physiology, genetics, and phylogenetic analyses, which is critical in understanding the origin of flowering plants. Throughout this essay I adopt the integrative biologic approach, which is subject to periodic revision. Adjustments and refinements are necessary to take into account ongoing review and study of the considerably large body of scientific evidence available to me from nearby research libraries.

One place and time to begin a discussion is a seminal 14th Pennsylvania State University Plant Physiology Symposium: Plant Reproduction: Insights into the Abominable Mystery.

Nancy Eckardt prepared a meeting report in 2002 that summarized key research findings of papers read to the assembled American Society of Plant Biologists (Eckardt 2002). Professor D. E. Soltis, one of the meeting's keynote speakers, explained how evolution of floral characters may be examined from a developmental (ontogenetic) perspective focusing on genetic studies of the four stem cell sectors of shoot apical meristems (SAMs):

merism (number and growth patterns of floral meristems)

phyllotaxis (spiral versus whorled organ positions)

perianth differentiation

ovary position

The 2002 conferees reported on three aspects of exciting research on the evolution of floral characters, which is underway is several different laboratories. These aspects involve the floral genome project, evolution and function of floral regulators, and important advances dealing with the anatomy of plant reproduction.

The present essay updates scientific evidence gleaned from the 2002 Pennsylvania State University Plant Physiology Symposium with 2009 interdisciplinary data. I consider evidence from research on insect and plant anatomy, development, and morphology in the next chapter.

Anatomic and Developmental Considerations:

The following chapter of the essay considers evidence drawn from studies of insect and plant anatomy and development that might help us understand a possible coevolutionary origin of angiosperms and certain clades of holometabolous phytophagous insect antagonists.

Cis-acting transcription factors and signaling cascades are hallmarks of the anatomic and developmental regulatory networks in insect antagonists and their seed plant hosts. Logically, understanding the structure of developmental regulatory networks, communication conduits, and signaling systems in extant phytophagous arthropods and seed plants might allow us to better interpret paleontologic transformational series.

This line of reasoning is necessary to decipher character homologies and to solve the mystery of angiosperm origins within a coevolutionary and phylogenetic context. However, homology assessment must take into account comparisons between animal and plant genetic, developmental, and transcriptional regulatory systems at the highest levels (Meyerowitz 2002, Niklas 2006).

According to a reviews by Meyerowitz (2002) and Becker and Theißen (Figure 1, page 468, 2003), the developmental systems of animals and plants were derived independently over the course of more than a billion years of organic evolution. The homeobox genes (e.g. HOM/Hox genes) essential for animal patterning are not the same as MADS-box homeotic genes from plants. Further animal MADS-box genes e.g. MEF2 encode certain enzymes involved in muscle differentiation of both invertebrates and vertebrates and wing vein and tracheal development in certain insects (Meyerowitz 2002).

The image to the right is the passive insect trapping flowering plant, Darlingtonia californica (Sarraceniaceae, Nepenthales, Dilleniidae), photographed by the author at a seep on Eight Dollar Mountain located in the Klamath Region of western North America.

Were plant-derived ecdysone signals sent to the homeotic genetic machinery in cells of developing herbivorous insects residing among the clasping leaves bases of massive shoot apical meristems (SAMs) and bisexual cone axes of Permo-Carboniferous and Permo-Triassic seed plant host shrubs?

Invertebrate body plans. Sean B. Carroll (1995), Rogers et al. (1997), Gellon and McGinnis (1998), and S. B. Carroll et al. (2005) review molecular evolution of homeobox genes and homeodomain transcription factors needed to understand the molecular basis of body ground plan development in Paleozoic Ecdysozoa including insects.

A key paper on the control of insect body size (Nijhout 2003) outlines the molecular mechanisms (involving cis-acting transcription factors and hormones) and environmental controls (nutrition and temperature) behind growth and cell division in hemimetabolous and holometabolous insects.

Molecular control over arthropod growth varies among the major clades of hemimetabolous and holometabolous insects (Grimaldi and Engel 2005). Several neurosecretory hormones play an important part in mechanisms that regulate cell division and growth including insulin-like peptides (Drosophila insulin-like proteins [DILPs] and bombyxins), chitenase-derived imaginal disk factor proteins, the steroid hormone ecdysone, local autocrine and paracrine transcription factors, and brain neurosecretory prothoracicotropic hormone (PTTH) (Nijhout 2003).

Cessation of growth in holometabolous insects leading to a new moulting cycle is triggered by PTTH that initiates the ecdysone growth regulatory cascade. Prothoracicotropic hormone and/or ecdysone secretion in Holometabola is negatively controlled by juvenile hormone (JH) (Truman and Riddiford 2002). Once JH circulating in the hemolymph is destroyed by juvenile hormone esterases, then PTTH secretion resumes under circadian (22-24 hour) photoperiodic control (Nijhout 2003). On the other hand, certain hemimetabolous bugs (Hemiptera) possess abdominal stretch receptors that activate secretion of PTTH (Nijhout 2003).

Insect developmental tool kit. The insect developmental tool kit is comprised of certain homeotic selector genes (including Hox genes), zygotic (gap- maternal-, and pair-rule-) genes, field-specific selector genes, compartment selector genes, cell-type-specific selector genes, and segment polarity genes; and the transcription factors they encode. In addition, the insect developmental tool kit is comprised of controlling factors behind the cessation of insect growth including bioactive PTTH, JH, juvenile hormone esterases, and ecdysone steroids (Truman and Riddiford 2002, Nijhout 2003, S. B. Carroll et al. 2005).

Thummel and Chory (2002) point to a possible coevolutionary connection between the 20E-ecdysone/cytochrome P450 biosynthetic machinery of insect antagonists and seed plant hosts. Structurally similar to bioactive plant brassinosteroids, 20E-ecdysone induces a cascade of transcription factor biosynthesis important in the regulation of insect development (Truman and Riddiford 2002, De Loof 2008).

Juvenile hormone (JH) biosynthesized in the corpora allata of the insect brain is a sesquiterpenoid epoxide methyl ester (Hartfelder 2000). Interestingly, many naturally-occurring plant sesquiterpene esters and lactones are bioactive and exhibit insecticidal properties. Juvenile hormone and its homologs are integral in vitellogenesis (Hartfelder 2000), regulation of moult cycles (Truman and Riddiford 2002), and caste development and behavior in social Hymenoptera (Guidugli et al. 2005).

Were bioactive brassinolides and sesquiterpenes manufactured by Paleozoic seed plants used as chemical warfare agents to affect growth, development, and behaviour of herbivorous insects?

Several developmental gene families, transcription factors, and enzymes involved in hormone signaling cascades are known in invertebrates based in part, on experimental studies of the Drosophila model arthropod (S. B. Carroll et al. 2005, Hittinger and S. B. Carroll 2008, among others). Key elements of the Drosophila genetic tool kit are:

Achaete-Scute [AS-C] complex (four-gene complexes encoding transcription factors critical for bristle patterning and neural cell development)

apterous [ap] (compartment-selector gene encoding Apterous protein needed for subdivision of imaginal discs into dorsal and ventral compartments)

bicoid [bcd] (maternal effect gene necessary for anterior embryo development)

caudal and nanos (maternal effect genes needed for posterior development of the embryo)

decapentaplegic [dpp], short-gastrulation [sog], snail [sna], twist [twi], and zerknüllt [zen] (maternal effect genes involved in subdivision of cells in the dorsiventral axis of larvae)

distal-less [dll] (field-specific selector gene necessary for limb [antennae, genitalia, mouthparts, legs] development)

engraled [en] (compartment-selector gene encoding distinct homeodomain transcription factors which determine the posterior identity of embryos and wings)

even-skipped, fushi-tarazu, hairy, and paired (pair-rule genes which act in the periodicity of double-segments)

eyeless [ey] (encodes Ey protein, a DNA transcription factor which regulates other field-specific genes)

giant, huckebein, hunchback, knirps, Krüppel, and tailless (gap genes that regulate larval segmentation)

hedgehog and wingless (segment polarity genes which organize the orientation and patterns of cells within the anteroposterior axis of each larval segment)

Hox genes (clusters of eight-linked homeotic selector genes absolutely required for embryo, larval, and adult development)

Runx genes (clusters of homeotic selector genes that regulate development)

scalloped [sd] and vestigial [vg] (field-specific selector genes required for haltere and wing development)

tinman (field-specific selector gene that controls development of the insect heart)

Evolution of the Hox complex. Molecular diversification of the Hox gene complex over the course of 600 million years of metazoan evolution is analogous to the 400 million-year old molecular evolution of MIKC-type MADS-box genes and related cis-acting transcription factors of land plants (Theißen et al. 2000, Theißen 2001, and Theißen and Becker 2004).

Evolution of the Hox complex probably involved small gene duplications (SGDs), whole genome duplications (WGDs), divergence of homeodomains, disintegration of the Hox cluster at breakpoints, and rapid changes in the nucleotide sequence of homeodomains (S. B. Carroll et al. 2005, Negre et al. 2005). Further, changes in the arthropod homeodomain and evolution of new protein motifs led to new Hox developmental functions in certain insect lineages (S. B. Carroll et al. 2005).

The homeotic selector genes (S. B. Carroll et al. 2005, among others) of the Drosophila Hox complex are:

abdominal A [abd-A] (selects the developmental fate of the front-middle parts of the abdomen)

Abdominal B [Abd-B] (selects the developmental fate of the rear-middle part of the abdomen)

antennapedia [ANTb] (affects development of the middle legs and middle part of the thorax)

deformed [Dfd] (affects development of the upper part of the head)

labial [lab] (selects the developmental fate of the labium of the head)

proboscipedia [pb] (affects development of the lower half of the head)

sex combs reduced [Scr] (selects the developmental fate of the most posterior region of the head, labial segment, front legs, prothorax, and front and anterior portions of the thorax) (Rogers et al. 1997)

ultrabithorax [Ubx] (selects the developmental fate of the rear legs and thorax and front of the abdomen)

Ancestral arthropods possess two additional homeotic selector genes of the Hox cluster that together comprise the HOM-C, ten-gene complex (see discussion in Negre et al. 2005). These additional genes are:

fushi tarazu (ftz), and

Hox3

Genomic analyses suggest that derived winged insects lost functional copies of ftz and Hox3 through disintegration of the HOM-C complex (Negre et al. 2005). Duplication of the Hox3 gene of ancestral Cyclorrhaphan flies gave rise to two maternal effect genes, bicoid (bcd) and zerknüllt (zen) (Stauber et al. 2002). Based upon this study it is important to include Hox3 as part of the ancestral diverging insect developmental tool kit.

Early divergent arthropod developmental tool kit. Possible candidates for the early divergent insect developmental tool kit might include certain homeotic selector genes of the Hox complex such as homologs and paralogs of abdominal A (abd-A), Abdominal B (Abd-B), Hox3, proboscipedia (pb), sex combs reduced (Scr) (Rogers et al. 1997), and Ultrabithorax (Ubx), and the field-specific selector gene necessary for limb development in Drosophila (Diptera) known as distal-less (dll) (S. B. Carroll et al. 2005).

The importance of Ubx protein encoded by the Ubx gene in the early divergent insect developmental tool kit cannot be neglected in the present analysis since significant changes in the carboxy-terminal (C-terminal) region (Galant and Carroll 2002) and serine/threonine phosphorylation sites (Ronshaugen et al. 2002) are probably behind many insect body plan novelties seen in the paleontologic record of the past 400 million years of arthropod and crustacean evolution (Pavlopoulos and Averof 2002).

I also add hexamerin moulting storage proteins which are related to hemocyanin respiratory enzymes (Burmester et al. 1998, Burmester 2001, Burmester et al. 2006, Burmester and Hankein 2007), juvenile hormone esterases, vitellogenin genes and yolk proteins (Isoe and Hagedorn 2007), pheromone chemoreceptors (Robertson and Wanner 2006), and certain nuclear receptor proteins (Bonneton et al. 2003, Bonneton et al. 2008) including ultraspiracle, and ecdysone-inducible transcription factors to the list of developmental tools in the ancestral, diverging insect kit.

Respiratory enzymes (hemocyanins) and moulting storage proteins (hexamerins) are key elements of the early divergent arthropod developmental tool kit that tie-in with the evolution of insect legs and wings from bilaterian gills. Ancient insect wings probably functioned as respiratory organs. Wings, halteres, arachnid spinnerets, and insect legs are all organs that develop from limb fields of cells where Ubx expression is prevalent (S. B. Carroll et al. 2005).

Several insect systematists studying beetle (Coleoptera) evolution are employing some of the aforementioned genes and proteins of the insect development tool kit in their phylogenetic analyses (Gómez-Zurita and Galián 2005). Gómez-Zurita and Galián (2005) discuss the utility of molecular phylogenetic characters appearing in the entomological literature in a review paper, which is organized along the lines of Floyd and Bowman (2007) for land plants (see section below).

Molecular model systems used as tools in beetle genomic research and phylogenetic studies include proteins central to development (JH esterases), diapause proteins, heat shock proteins, ultraspiracle (an ecdysone nuclear receptor protein), cuticle proteins, hexamerins, genes encoding vitellogenin, and apolipophorins, among others (see review by Gómez-Zurita and Galián 2005).

While the work published by coleopterists also involves study of digestive enzymes and proteinase inhibitors associated with herbivore control and the insect-plant arms race, molecular coevolution and the origin of angiosperms is apparently not a topic of focus. However, a molecular-based phylogeny of chrysomelids provides evidence of independent diversification of beetle clades and supposedly coevolving host plant lineages (Gómez-Zurita et al. 2007), which is at odds with paleontologic evidence of insect feeding traces on late Cretaceous and Paleogene (Eocene) fossilized hispine ginger leaves (Wilf et al. 2000).

Understanding the nature and timing of early molecular diversification of homeotic selector genes, developmental proteins, nuclear receptor proteins, and cis-acting transcription factors of both invertebrate antagonists and vascular plant hosts is the critical first step in understanding a coevolutionary origin of angiosperms.

The place and time to begin a molecular phylogenetic analysis is the late Devonian-early Carboniferous hypoxic icehouse (DeCARB)

Plant architecture, body plans, and mutualism. A theoretical model exists for evolution of animal or plant body size and allometry by natural selection, which is known as the WBE (West-Brown-Enquist) Model. Enquist et al. (2007) provide evidence that a single 3/4 allometric scaling formula applies to plants at the organismal level (anatomy, body size, and metabolism), and is highly conserved having "been under strong stabilizing selection for millions, if not hundreds of millions, of years" (page 745, Enquist et al. 2007).

Three-quarter scaling may also extend to the level of paleopopulations and communities according to Enquist and coworkers. Scaling in quarters may apply to all levels from molecules to organ (or body) size. Perhaps scaling functions and integrals could predict the structure of herbivore-plant compartments complete with hypothetical colonies of interacting insects and massive shoot apical meristems of seed plants indigenous to hypoxic or hypercapnic paleoenvironments.

The focal point in the present evolutionary and morphologic analysis is the shoot apical meristem (SAM) and accessory fertile meristems of extinct and extant seed plants. Accessory fertile meristems and cone axes probably function differently at the genetic level than SAMs (Glover 2007).

One or more of these questions might be answered by scientific studies:

Were certain seed plant morphologies exploited by respiring insect eggs, larvae, instars, and adults?

Can we identify plant adaptations that helped shield insect eggs and instars from hypoxia, temperature extremes, and ultraviolet (UV) radiation in paleoenvironments?

What are the kinds of miniature seed-plant fitness landscapes exploited by insects?

Did the architecture of some Paleozoic seed plants provide critical, oxygenated habitat for phytophagous insect and tetrapod associates?

Was the development and form of pollen and seed cones (strobili) affected by wind currents in Triassic fitness landscapes?

Are some seed plant morphologies, such as ovule placement (adaxial or abaxial) on the surfaces of large, leathery leaves a consequence of the mechanostimulatory effects of grazing tetrapods on expanding SAMs of seed plant monopodia?

Can we identify an extant seed plant species that offers an example of a monopodial shrub and resident insect colony? The living cycad Zamia furfuracea (Zamiaceae) forms a mutualistic relationship with curculionid weevils Rhopalotria mollis (Coleoptera) (Norstog and Fawcett 1989).

There is mounting evidence from detailed studies of coalified and permineralized fossil flowers that Upper Cretaceous flowering plants already had mutual relationships with phytophagous and pollinating insects (Gandolfo et al. 2004, Friis et al. 2006). Paleoecologic and anatomic studies may shed more light on coevolutionary relationships between seed plants and large herbivores (Tiffney 1992).

"Ontogeny in land plants can be viewed as a complex, partly hierarchical, series of developmental processes, which together with their underlying genetic controls, provide the raw material for morphological innovation. Ontogeny is thus the creative force behind botanical diversification, and small modifications at the genetic level may have a disproportionate effect on plant form as their consequences cascade and multiply through development. Plant evolution occurs as variation in genetic and epigenetic developmental processes is winnowed by ecology..."

The preceding quotation is from page 161 of P. R. Crane and P. Kenrick (1997), Diverted development of reproductive organs: a source of morphological innovation in land plants, Plant Systematics and Evolution 206: 161-174.

How can botanists better understand the origin and homology of angiosperm characters from biochemical and genetic studies of co-option, modularity, dissociation, and diverted development in cone axes (strobili and flowers)?

"The origin of a new direction of adaptive evolution starts with a population of variably responsive, developmentally plastic organisms. That is, before the advent of a novel trait, there is a population of individuals that are already variable, and differentially responsive, or capable of producing phenotypic variants under the influence of new inputs from the genome and the environment."

The preceding quotation is from page 6544 of M. J. West-Eberhard (2005), Developmental plasticity and the origin of species differences, Proceedings of the National Academy of Sciences 102(Supplement 1): 6543-6549.

Land plant developmental tool kit. Floyd and Bowman (2007) are the first workers to estimate the developmental tool kit of early land plants including Paleozoic seed plant homeotic genes potentially important in the later evolution and diversification of angiosperms and origin of the first flowers from bisexual cone axes sensu Baum and Hileman (2006).

The work by Floyd and Bowman (2007) focuses on a molecular phylogenetic analysis of Chara (a green alga), Physcomitrella (a moss), Selaginella (a lycophyte), Arabidopsis (angiosperm eudicot), Antirrhinum (angiosperm asterid), Oryza (angiosperm monocot), Populus (angiosperm eudicot), Picea (gymnosperm conifer), and Pinus (among others).

Many developmental gene families and transcription factors have been identified in land plants (Langdale 2008). These genes, and their transcription factors, together with phytohormone biosynthetic and regulatory factors, homeotic selector genes, and certain microRNAs, make-up the developmental tool kit of land plants including vascular seed plants and angiosperms. The main developmental and regulatory gene tools of focus are:

ARF (auxin response factor gene family)

ARP (KNOX regulating genes)

AS2 (LOB family genes that function in lateral organ formation in SAMs)

Class III HD-Zip (function in SAM and leaf development)

Class I KNOX (essential genes to meristem maintenance and SAM formation affected by auxins and gibberellins)

Class I TCP (leaf sculpting genes and cell division controllers)

Class II TCP (genes that determine leaf shape, floral symmetry, and flower patterns)

CUC (important genes in embryo and flower development)

euANT (regulating genes involved in lateral organ development)

euAP2 (SAM maintenance and timing of flowering including APETALA 1 [AP1]and APETALA 2 [AP2])

GRAS (a family of genes necessary for radial patterning of roots and shoots and axillary meristem development [HAIRY MERISTEM, LATERAL SUPRESSOR, SCARECROW], and in mediation of gibberellic acid responses during flowering and development of trichomes [DELLA genes])

KANADI (GARP family transcriptional regulators antagonistic to Class III HD-Zip genes important to leaf and integument development)

LEAFY [LFY] (controllers of the downstream expression of several genes including MIKC-type MADS-box genes)

MIKC-type MADS-box genes (developmental genes that control organogenesis in seed plants)

PIN (auxin related genes needed in land plant architecture)

TIR (auxin related genes)

WOX Class genes including WUSCHEL (essential genes for stem cell maintenance in SAMs)

YABBY (seed plant specific family of genes that promote differentiation of abaxial tissues and leaf growth)

Based on a molecular phylogenetic analysis by Floyd and Bowman (2007) the ancestral land plant homeotic gene tools of our focus to deduce the timing of the origin of the first flower should be AP2, ARF, Class I and II TCP, Class III HD-Zip, GRAS, KNOX, LEAFY, MIKC-type MADS-box, PIN, TIR, and YABBY genes.

Cell geometry of shoot apical meristems (SAMs). Much progress has been made on understanding developmental biochemistry and hormonal controls behind the anatomy and cell geometry of the shoot apical meristems (SAMs) and accessory fertile meristems of a number of diverse flowering plants and gymnosperms (Glover 2007, Busov et al. 2008). The best studied example is the diminutive mouse-ear cress (Arabidopsis thaliana [Brassicaceae, Capparales, Dilleniidae]), an extant, highly derived annual species of mustards (Kwiatkowska 2006).

Floral meristem identity and integrator genes of massive shoot apical meristems of the oil palm (Elaeis guineensis [Arecaceae, Arecales, Arecidae]) have also been studied (Adam et al. 2007).

Biomechanics is a fertile area of research in botany (Niklas 1992, Niklas 2000, Rowe and Speck 2005, Niklas et al. 2006, Read and Stokes 2006). Can the experiments drawn from principles of geometry and engineering answer some of the basic questions posed by Coen et al. (2004)?

"Whatever the detailed mechanism for aligning cell axes with gradients [of morphogen flow], these considerations indicate that morphogens play two distinct roles in the generation of shape. The first is a regionalizing role, involved in elaborating differences between regions [of meristems], such as levels of gene activity, growth rate, or anisotropy. The second is a polarizing role involved in specifying directions of cell axes. Morphogens with a regionalizing role often act by influencing transcription factors, which in turn can affect parameters such as growth rate and anisotropy. Cells respond only to the overall concentration of these morphogens, not to the direction of their gradient. Morphogens with a polarizing role, influence the orientation of cell axes and can determine the principal directions of growth. Cells respond to the highest slope of the morphogen gradient or to the direction of morphogen flow. Such responses [of plant cells] are unlikely to be mediated by transcription factors, as the directional information would be lost through conversion to a gene expression level."

The preceding quote is from E. Coen, A.-G. Rolland-Lagan, M. Mathews, A. Bangham, and P. Prusinkiewicz (2004), The genetics of geometry, Proceedings of the National Academy of Sciences 101(14): 4728-4735. The phrases in brackets [] are mine.

Certain anatomical and enzymatic tools of the land plant hormone signaling machinery align cell axes and control morphogen flow. Some of these affect SAM development and phyllotaxis.

Phyllotaxis. Phyllotactic patterns of extant flowering plants are of ongoing interest to paleobotanists and plant systematics who seek to understand the origin of the angiosperm flower from an evolutionary developmental (evo-devo) perspective (J. A. Doyle and Endress 2000, Endress and J. A. Doyle 2007). Floral phyllotaxis is distinct from vegetative phyllotaxis.

Extant basal angiosperms including the Magnoliales exhibit both spiral and whorled floral phyllotactic patterns (J. A. Doyle and Endress 2000, Ronse De Craene et al. 2003, Staedler et al. 2007). When floral phyllotactic characters in basal angiosperms are subjected to phylogenetic analyses, reversals from spiral to whorled and whorled to spiral become apparent (Ronse De Craene et al. 2003, Endress and J. A. Doyle 2007).

Phyllotaxis is reviewed by Kuhlemeier (2007). The most recent Arabidopsis model suggests that in addition to PIN proteins, the family of AUX1 LAX auxin influx carrier enzymes stabilize spiral phyllotaxis (Bainbridge et al. 2008). An earlier model of developmental phyllotaxis and initiation of procambial strands leading to vasculature in the SAMs of extant flowering plants implicates auxins, auxin efflux carrier protein (ATPIN1), mRNA transcripts such as ATHB8-GUS, and other auxin related developmental genes (R. S. Smith et al. 2006). Development of phyllotactic geometric patterns and vasculature in SAMs of some modern flowering plant species may be traced to a common genetic and physiologic mechanism (Dengler 2006).

Were the possibly conserved mechanisms of SAM differentiation that underlie geometric patterning in sporophytes operating the same way in long extinct seed plant species?

According to Floyd and Bowman (2007) the answer to the second question is "yes." Orthologues of PIN and ARF, auxin related genes and transcription factors needed in land plant architecture and phyllotaxis, were part of the ancestral land plant developmental tool kit according to Floyd and Bowman (2007).

Floyd and Bowman (2007) identified several "PIN-like" genes from expressed sequence tag (EST) studies of moss gametophytes belonging to Physcomitrella. Five PIN genes were identified in sporophytes of extant Selaginella, and eight additional PIN genes were found in Arabidopsis (Floyd and Bowman 2007).

"If the hypothesis that increasing complexity of auxin metabolism, transport, and perception was related to the increased sporophytic complexity accompanying land plant evolution, then we might expect these particular gene families to be less complex in the moss and lycophyte lineages than in angiosperms, as indeed appears to be the case."

The preceding quote is from page 22 of Sandra K. Floyd and John L. Bowman (2007), The ancestral developmental tool kit of land plants, International Journal of Plant Sciences 168(1): 1-35.

Did auxin exist in Paleozoic vascular plants? Yes, according to a study by Rothwell and Lev-Yadun (2005).

Macromolecular secretions (microRNAs, bioactive peptides, and other large [and small-sized] molecules) when applied to the surfaces of SAMs by chewing, crawling, ovipositing, piercing, and sucking resident, phytophagous insects, could potentially signal the host plant genome and affect phyllotaxis (among other developmentally plastic traits), but experimental studies are needed to support this hypothesis.

Leaf development. Ongoing research on the anatomy and cell biology of leaf development from meristematic cells of SAMs is incrementally reviewed by Sinha (1999), Roth-Nebelsick et al. (2001), Engström et al. (2004), Aida and Tasaka (2006), Brodribb et al. (2007), and Champagne et al. (2007). Development of leaves depends on the activity of Class III HD-ZIP genes (Floyd and Bowman 2007), and a KNOX/ARP genetically-based, modular switch mechanism (Beerling and Fleming 2007).

Whether the KNOX genetic switch is on or off depends on the activity of ARP repressors, and may lead to the development of different leaf forms e.g. pinnate, palmate, or compound (Beerling and Fleming 2007).

Several books and review papers place leaf development of vascular plants including angiosperms in a paleontologic context including Melville (1969), Stewart and Rothwell (1993), Trivett and Pigg (1996), Sinha (1999), Roth-Nebelsick et al. (2001), Boyce (2005), Feild and Arens (2005), J. A. Doyle (2006), Beerling and Fleming (2007), Brodribb et al. (2007), and Sanders et al. (2007).

Land plants possess two leaf hydraulic strategies as expressed by their anatomy and morphology (Brodribb et al. 2007). According to Brodribb et al. (pages 1894 and 1895, 2007), the first hydraulic method is to branch the system of veins to deliver water to the site of transpiration (this is the method employed by many flowering plants). The second method seen in some members of the gymnosperm clade is to modify hydraulics by lignifying certain cells of the leaf mesophyll thus channeling water to the stomates.

The Permian gigantopterid plant Delnortea abbottiae possesses leaves with four orders of veins, and certain cells of the leaf hypodermis are sclerotic (pages 1423 and 1424, Fig. 36 on page 1420, Mamay et al. 1988), which is anatomically transitional to the two means of water distribution in leaves described by Brodribb et al. (2007).

The image to the left is a piece of rock slab containing the largest known gigantopterid fossil leaf compression collected by the author and colleagues (USNM 387473, Fig. 10, page 1413, Mamay et al. 1988). The piece of fossil leaf was photographed by the writer on the day it was pulled from the rock layer in the Lower Permian (Leonardian) Cathedral Mountain Formation, Del Norte Mountains of southwestern North America in 1982.

Was the paleophysiology of gigantopterid plant Delnortea as expressed by the presence of two anatomical innovations to maximize leaf hydraulics somehow linked with the success of these plants in arid Permian paleoenvironments? Can these same leaf traits be found in later Mesozoic seed plants? And what aspects of transcriptional regulation were behind the evolutionary development of Delnortea and Evolsonia (Mamay 1989) leaves: morphologies not seen in any older Paleozoic fossils?

Abaxial-adaxial leaf organ development in angiosperms begins in the shoot apical meristem: the adaxial side of leaf primordia faces toward the SAM while the abaxial surface is orientated away from the SAM. The genetic basis of cell polarity in Arabidopsis SAMs is reviewed by Engström et al. (2004). Three classes of transcription factors establish adaxial-abaxial leaf organ identity in Arabidopsis: Class III HD-Zip, KANADI, and YABBY proteins (Nole-Wilson and Krizek 2006).

AINTEGUMENTA (ANT) gene expression in certain extant flowering plants determines which cells of the SAM "leave the meristem to form lateral organs" (page 983, Nole-Wilson and Krizek 2006). Further, ANT and APETALA2 are transcription factors of focus in phylogenetic analyses of seed plants (Floyd and Bowman 2007) and the origin of angiosperms (S. Kim et al. 2006).

Ovule development from leaves and axillary meristems is reviewed by Skinner et al. (2004) and Scutt et al. (2006) (see next section).

Development of reproductive modules. Modular development as it pertains to seed plant reproduction may be viewed at the cellular or organ level. Friedman and Ryerson (2009) in a recent review of megagametogenesis in angiospermous seed plants argue that the female gametophyte of flowering plants consists of a modular quartet of cells. However, from an evolutionary-developmental (evo-devo) perspective, aggregations of cells (tissue fields and compartments) and whole organs are also potentially modular. I adopt this philosophy in the present discussion and while assigning character polarities in a later chapter on phylogenetic considerations.

Paleontologic evidence of ovule development in bennettitalean seed plants suggests early innovation in the production of these reproductive modules (Rothwell and Stockey 2002). Cellular and anatomic details of sexual reproduction in extinct seed plant groups are sometimes "accessible" contrary to statements in the literature (page 145, Williams 2009) but require tedious preparation of rare permineralizations (e.g. Nishida et al. 2004, Rothwell et al. 2009).

Several key papers and reviews are devoted to the study of reproductive structures in seed plants as an approach to gain a better understanding of evolution of gymnosperms and the origin of flowering plants (Friedman 1992, Crepet and Nixon 1996, Crane and Kenrick 1997, Labandeira 2000, Lord and Russell 2002, Friedman et al. 2004, Friedman and Williams 2004, Williams and Friedman 2004, Friedman et al. 2008, Friedman and Ryerson 2009, Williams 2009).

The image to the right is the permineralized nucellus (the megasporangium wall of a fossilized ovule) of Eoantha zerikhinii (from a gnetalean "pre-flower") from the Baisian Assemblage, early Cretaceous Period, of Transbaikalia in central Asia. The fossilized ovule contains a female gametophyte with a pollen chamber filled with Ephedrites-type pollen (this picture is from an original image provided by Professor Valentin Krassilov, posted here with his permission).

Reproductive modules are specific organs, tissue fields, or cells derived from a SAM or cone axis. Evolutionary development and genetics of the female reproductive tract is reviewed by B. C. W. Crawford and Yanofsky (2008). The bulleted list below includes some key references:

archegonia (haploid tissue receptacles in certain gymnosperms that contain the egg and accessory cells)

carpel (an inrolled [conduplicately folded] ovule-bearing leaf) (Scutt et al. 2006)

central cell (female haploid cell of the embryo sac involved in double fertilization) (Bemer et al. 2008, Berger et al. 2008, R. J. Scott et al. 2008)

eggs or egg nuclei (female haploid cells or nuclei, respectively) (Berger et al. 2008, R. J. Scott et al. 2008)

embryo (an often tiny [usually dormant] diploid daughter plant [complete with a basal cell, suspensor, and growing point] that develops from the fertilized egg, usually housed within a mature ovule or ripened seed) (Tobe et al. 2000, Holloway and Friedman 2008)

endosperm (diploid, triploid, or polyploid nutritive cells or nuclei usually housed inside of ovules and embryo sacs) (Baroux et al. 2002, Raghavan 2003)

interseminal scale (modified female organs of bennettitaleans possibly used to house toxins needed as defense against potential herbivores)

megasporangia (multicellular sacs that produce megaspore mother cells and megaspores)

megaspore mother cells (female diploid cells where meiosis occurs)

megaspores (female haploid products of meiosis)

megagametophytes (cellular or acellular haploid tissues derived from cell- and/or nuclear divisions in megaspores [termed modular quartets of cells in flowering plants]) (Yadegari and Drews 2004, Tobe et al. 2007, Blanvillain et al. 2008, Friedman et al. 2008, Madrid and Friedman 2008, Friedman and Ryerson 2009, Madrid and Freidman 2009)

microgametophytes (Russell 1991, Rudall and Bateman 2007, Blanvillain et al. 2008, Borg et al. 2009)

microsporangia (diploid multicellular sacs that produce diploid microspore mother cells, haploid microspores, and haploid pollen) (Russell 1991, Furness et al. 2002)

microspore mother cells (male diploid meiotic cells) (Russell 1991, Furness et al. 2002)

microspores (male haploid products of meiosis) (Russell 1991, Furness et al. 2002, Nadot et al. 2008)

ovule (a complex organ of the mother seed plant that contains sterile diploid protective tissues termed integuments [often hardened by lignified cell walls], a vestigial megaspore wall [nucellus], megaspores; and at later stages of development: megagametophytes, other nutritive tissues in some seed plants [endosperm], archegonia [in certain seed plants], and eggs [to include accessory cells and synergids in some seed plants]) (Skinner et al. 2004, Scutt et al. 2006, Colombo et al. 2008, Skinner and Gasser 2009)

polar nuclei (haploid nuclei of the embryo sac) (Berger et al. 2008)

pollen grain (an endosporic microgametophyte that consists of a hardened microspore containing haploid male nuclei and/cells) (Borg et al. 2009, Lora et al. 2009)

pollen tube (a living tubular cell derived from the male [micro] gametophyte, which is lined with callose in extant flowering plants) (Russell 1991, Williams 2008, Cai and Cresti 2009, Williams 2009)

seed (a ripened ovule that contains an embryo)

sperm or sperm nuclei (male haploid cells [sometimes motile] or nuclei, respectively) (Berger et al. 2008)

stamen (a complex, highly modified male branch or grouping of fused male [micro] sporophylls that contains one or more microsporangia [and pollen when ripe]) (Alves-Ferreira et al. 2007)

stigma (the pollen receptive surface of a carpel)

staminode (a sterile, often leaf-like module derived from the male [micro] sporophyll or stamen) (Walker-Larsen and Harder 2000)

suspensor (an often linear array of diploid cells [often containing endoreduplicated nuclei] of the embryo)

synergids (specialized cells of the embryo sac that interact with the pollen tube during double fertilization) (Sandaklie-Nikolova et al. 2007)

Megagametophytes. Despite the importance of the megagametophyte in hypotheses of the origin of angiosperms almost nothing is known of the anatomy of female haploid stages in Paleozoic gymnosperms. Only recently have paleobotanists begun to understand the anatomy and morphology of female gametophytes of Mesozoic seed plants (Stockey and Rothwell 2003, Rothwell et al. 2009). Preserved megagametophytes of unequivocal stem group flowering plants are unknown.

Additional insight has been gained from studies of megagametogenesis in evolutionarily derived extant gymnosperms including Gnetum (Gnetales) (Friedman and Carmichael 1998). Now that molecular systematists converge on the root of phylogeny of extant flowering plants, plant anatomists have turned their research focus to studies on the reproductive anatomy of basal angiosperms including magnoliids.

Anatomical studies of megagametophytes of basal flowering plants have been conducted by Friedman and Williams (2003), Madrid and Friedman (2008), Friedman and Ryerson (2009), and Madrid and Freidman (2009). The angiosperm female gametophyte consists of a basic modular quartet of cells according to Friedman and Ryerson (2009).

In 2006 Friedman described an unusual eight-celled and nine-nucleate female gametophyte in the basal flowering plant Amborella trichopoda (Amborellaceae, Laurales, Magnoliidae), a shrubby species indigenous to New Caledonia in the southwest Pacific. While discussing the peculiar four-celled egg apparatus (consisting of three synergids and one egg cell) of Amborella, in comparison with the common seven-celled, eight-nucleate Polygonum-type female gametophyte of angiosperms, Friedman states:

"Ironically, it is now evident that none of the most ancient lineages of flowering plants produces a seven-celled, eight-nucleate female gametophyte, a stark reminder that much remains to be discovered or correctly circumscribed for the earliest angiosperms."

The preceding quote is from page 339 of W. E. Friedman (2006), Embryological evidence for developmental lability during early angiosperm evolution, Nature 441: 337-340.

Historically the seven-celled, eight-nucleate female gametophyte, also known as the Polygonum-type, was regarded as the plesiomorphic state for flowering plants. Tobe et al. (2000) reported this anatomical type in Amborella. However, in 2006 Friedman discovered a peculiar eight-celled, nine-nucleate Amborella-type female gametophyte in this same group of New Caledonian angiosperms.

Detailed anatomical studies of Austrobaileya (Austrobaileyales) by Tobe et al. (2007), Illicium (Illiciales) by Williams and Friedman (2004), Kadsura (Illiciales) by Friedman et al. (2003), and Nymphaeales (including Nuphar and Hydatellaceae [Hydatella and Trithuria]) (Williams and Friedman 2002, Friedman 2008, Rudall et al. 2008), reveal a four-celled, four-nucleate, Nuphar-type or Schizandra-type female gametophyte.

According to the review by Friedman and Ryerson (2009) the three types of female gametophyte development in basal clades of extant flowering plants are:

Amborella-type (known only from Amborella trichopoda (Amborellaceae)

Nuphar/Schizandra-type (common to all Austrobaileyales and Nymphaeales), and

Polygonum-type (“ancestral” to magnoliids, Ceratophyllaceae, Chloranthaceae, eudicots, monocots)

Further, Friedman and Ryerson propose that a four-celled egg apparatus consisting of an egg cell, two synergids, and a central cell is the most plesiomorphic character state for the angiosperm female gametophyte. Concomitantly, these authors suggest that either a one-module megagametophyte (i.e. in Austrobaileyales and Nymphaeales) or a two-module female gametophyte (i.e. among magnoliids, Ceratophyllaceae, Chloranthaceae, eudicots, monocots) could be the plesiomorphic condition in two different parsimony analyses (Friedman and Ryerson 2009).

In my opinion, all of the female gametophyte types described for extant crown group basal angiosperms including magnoliids are derived (apomorphic). The possibility that species of the long extinct stem group of flowering plants might have possessed a many-celled (i.e. more than 16-celled) megagametophyte compartmentalized into two or more modular octets of cells, should be entertained when computing future phylogenetic reconstructions of the flowering plant lineage.

A concerted effort by paleobotanists is needed to unearth, painstakingly prepare, and to describe the anatomy of female gametophytes of Carboniferous and Permo-Triassic seed plants, including long extinct stem group angiosperms (when discovered from the 160-million-year old ghost lineage) and their potential Paleozoic gigantopterid and vojnovskyalean ancestors.

Ovules. Considerable discussion by paleobotanists on the origin of flowering plants focuses on the origin of bitegmic (having a two-layered integument), anatropous (reflexed) ovules from gymnospermous ancestors (page 458, Stewart and Rothwell 1993, J. A. Doyle 2006).

Ovules of flowering plants are composed of the embryo sac (megagametophyte), a nucellus (old megasporangium wall), and sterile integuments. Integuments are generally regarded as modified leaves. Many but not all angiosperms possess two integuments. It is important therefore, to understand the evolutionary-developmental genetics of integuments and details of ovule ontogeny.

Frohlich (2003) offers additional discussion on the origin of ovule integumentation and placement on leaf surfaces from the standpoint of YABBY signaling from shoot apical meristems (SAMs). James A. Doyle discusses YABBY within the context of ovule tissue layers and attachment point (abaxial or adaxial) to carpels and leaves in Amborella, a basalmost angiosperm (J. A. Doyle 2006).

Pollen-containing male gametophytes. Considerable insight on the question of the origin of angiosperms has been gained from the recovery and study of palynomorphs (fossilized sporopollenin-filled pollen walls), whole pollen including preserved endosporic microgametophytes (Wing and Boucher 1998, Lupia et al. 1999, J. A. Doyle 2005, Zavada 2007), and preserved pollen inside fossilized insect guts (Krassilov 1997, Labandeira 2000). Ultrastructural details of fossilized pollen and spores are known from many vascular plants (T. N. Taylor et al. 1996). Further, angiospermous pollen of the perforate-reticulate and/or columellate type is known from Permian, Triassic, and Jurassic rocks (Cornet 1989, Hochuli and Feist-Burkhardt 2004, Zavada 2007).

Evolution and anatomical diversity of seed plant male gametophytes is reviewed by Friedman (1993) and T. N. Taylor et al. (2009).

Focused studies of the male gametophytes of seed plants have been underway for several years including investigations of Ginkgo (Ginkgoales) (Friedman 1987), Ephedra (Gnetales) (Friedman 1990), and Arabidopsis (eudicot Brassicales) (Friedman 1999).

Microgametophytes and sperm are often poorly preserved in the fossil record with the notable exception of Permian glossopterids (Nishida et al. 2003, Nishida et al. 2004). More often than not, the only evidence of male gametophytes in the fossil record consists of empty permineralized sporopollenin casings termed palynomorphs (T. N. Taylor et al. 2009).

An in-depth account of the stratigraphic distribution of seed plant palynomorphs (including charts and images) is available on the Cornet Web Site. The World Pollen Database is an important online resource.

The biology and diversity of pollen modules and pollination syndromes is discussed by several authors including Hughes (1976), J. Müller (1984), Hotton (1991), Cornet and Habib (1992), Krassilov (1997), Labandeira (2000), Ramanujam (2004), J. A. Doyle (2005), Zavada (2007), Friedman and S. C. H. Barrett (2008), and J. A. Doyle (2009).

Evolutionary-development of the seed plant progamic phase The progamic phase of seed plant reproduction is defined as, "the life history period between pollination and fertilization" (page 144, Williams 2009). The flowering plant diplophase and male and female haplophases coordinate development of pollen tubes through maternal tissues to ovules at the same time that female gametophytes develop from a megaspore mother cell. Biochemical interactions between cells of sporophytic and gametophytic reproductive modules that underlie orchestration of the progamic phase, vary widely among the angiosperms (Williams 2009).

Solitary carpels of the "primitive" magnoliid tree, Degeneria vitiensis (Degeneriaceae, Magnoliales), are shown on either side of the text. The left image is a scanning electron micrograph of a portion of the stigmatic secretion of a carpel. A tiny, thread-like pollen tube emanates from a single boat-shaped monosulcate pollen grain, which is just visible on the left-hand surface of the secretory plug. The pollen grain adheres to the secretory surface, ×15.

The pollen tube probably penetrates the flared stigmatic secretion and would eventually be guided to the egg of a modular female gametophyte inside one of several ovules within the carpel. The right image is a cutaway view revealing a row of ovules with cellular detail of the carpel, a thin placenta, and the stigmatic secretion, ×30.

Both samples were field-fixed in 1986; dissected from a flower collected in the canopy of a tagged and vouchered Degeneria tree; Naitaradamu Study Area, Viti Levu, Fiji. The scanning electron micrographs were prepared by Al Soeldner, Director, Oregon State University Electron Microscopy Laboratory. I thank the National Geographic Society for providing research funding for this study.

Extant crown group flowering plants including basal angiosperms exhibit remarkable diversity in pollen tube development (Williams 2008). Yet, nothing is known of microgametophyte development and the progamic phase in long extinct individuals and populations of the stem group.

Recent reviews of the angiosperm progamic phase, evo-devo of pollen tube transmitting tissues, and the biology of pollination of the basal crown groups of flowering plants are published by Sage et al. (2009), Thien et al. (2009), and Williams (2009).

Fertilization and embryology of seed plants. Friedman (1992, 1994) and Dumas and Rogowsky (2008), outline and discusses the large body of scientific literature on the reproductive biology of seed plants including fertilization, development of nutritive tissue (megagametophytes-discussed above), endosperm, and embryogenesis. Questions of reproductive success of flowering plants are discussed in several later reviews (Friedman 1995, 1999, 2001).

Double fertilization has historically been regarded as one of the hallmarks of angiospermy (Cronquist 1968, Takhtajan 1969, Stebbins 1974). Yet, later studies reveal that certain extant gymnosperms also possess this syndrome.

Specific studies of double fertilization in gymnosperms belonging to the anthophyte clade sensu J. A. Doyle and Donoghue (1986, 1987) have been carried out by Carmichael and Friedman (1995, 1996) and Friedman and Carmichael (1996) on Gnetum gnemon (Gnetales).

Many studies of double fertilization, endosperm formation, and embryology appear in the extensive recent literature on extant basal angiosperms including magnoliids. These include papers on endosperm development (Floyd and Friedman 2000) and embryogenesis (Arias and Williams 2008).

An important goal toward a better understanding of coevolution with insect antagonists and the early angiosperm reproduction might be to elucidate anatomical and evolutionary-developmental (evo-devo) interactions between insects and megasporophylls, pollen and megasporophylls, pollen and female pollen receptive surfaces, pollen tubes and ovaries (when present), pollen and carpels (including secretions), and pollen tubes and ovules.

Reviews by Lord and Russell (2002), Bernhardt et al. (2003), Bernasconi et al. (2004), Nasrallah (2005), Friedman et al. (2008), Cai and Cresti (2009), Sage et al. (2009), Thien et al. (2009), and Williams (2009) form a basis for opening new lines of research and later discussion from other perspectives.

Did the mechanical activities of animal antagonists and arthropod secretions affect evolutionary development of megasporophylls (and carpels), microsporophylls (and stamens), and the reproductive modules contained within these organs?

The next section of the essay discusses a possible coevolutionary origin of flowering plants from a thigmomorphogenetic research perspective.

Thigmomorphogenesis. Two recent reviews of thigmorphogenesis (Jaffe et al. 2002, Braam 2005) open a door to potentially novel approaches to better understand coevolution and mutualism. Surprisingly, discussion of the early work by Mordecai Jaffe (1973) on thigmomorphogenesis is neglected in most essays and books on coevolution.

Thigmomorphogenesis involves mechanostimulation, mechanoperception, signal transduction, and cellular differentiation. Mechanostimulation may be passive (caused by wind currents, flowing and freezing water, and scraping effects of soil particles and moving rocks) or active, triggered by touching animals, fungi, and neighboring plants (Braam 2005).

Mechanostimulation of the plant cell surfaces is perceived by the plant cytoskeleton (mechanoperception). Transduction of thigmo- (touch) stimuli by plant cells and tissues may signal certain touch-inducible (TCH) genes (Braam and Davis 1990, Lee et al. 2005), and potentially trigger developmental switches and differentiation, and change the shape and function of developmentally plastic organs (morphogenesis).

The diagram below illustrates how an instar of a Permo-Triassic insect might convey biochemical and mechanical signals to nuclei of dividing cells of the host plant shoot apical meristem (SAM). Mechanostimulation of cell walls, plasmalemmas, protoplasts, and the cytoskeleton of the plant SAM by the insect antagonist may involve a signal transduction cascade that potentially leads to de-repression of DELLA chromosomal proteins and switching on or off of meristem identity genes, differentiation, and thigmomorphogenesis.

Thigmomorphogenesis might be important toward better understanding evo-devo and the origin of flowering plants. There are many biochemical studies that link thigmo with the host plant and pathogen signaling cascade. Signal transduction to the nuclear envelope, chromatin loops, and labile protein switches in shoot apical meristems (SAMs) could be linked to mechanostimulation of cell walls, plasmalemmas, and the cytoskeleton by abiotic or biotic stimuli. Further, the host plant wounding response and biosynthesis of defense molecules is at the heart of the insect-plant coevolutionary arms race.

Interestingly, strobilus formation in club mosses (Selaginella) a modern evolutionary derivative of certain Paleozoic lycopods, is a thigmo response (Jaffe et al. 2002, Braam 2005, Read and Stokes 2006). Based on morphological and evo-devo studies of extant club mosses, were indeterminate membrane receptors, microtubule enzymes, and TCH genes and transcription factors tied-in with the developmental tool kit of early land plants?

An almost completely unexplored and potentially important phenomenon is whether insect bristles are capable of mechanically stimulating plant cell surfaces and the cytoskeleton, and inducing thigmorphogenesis in host plant organs. Arthropod bristles are innervated while hairs are not. Both types of insect organs are under developmental control and require the activity of the Achaete-Scute (AS-C) complex and shavenbaby/ovo (svb/ovo) gene, at least in Drosophila (S. B. Carroll et al. 2005).

Could the potential mechanostimulatory effects of insect bristles on seed plant cell surfaces of SAMs and floral meristems induce signaling of cis-regulatory modules on nuclear DNA of plant cells, triggering differentiation, and morphogenesis?

The engineering approach is used to study chewing in grasshoppers and locusts (acridids) and caterpillars of butterflies and moths (lepidopterans) (Clissold 2008), and to ascertain the magnitude of mechanical force potentially applied by ants (Paul and Gronenberg 1999), and other insects to artificial surfaces (Zhendong et al. 2002, Barbakadze et al. 2006). With the exception of in vivo x-ray analytical techniques there are few experimental studies in the literature to date that explore biomechanics of insect body parts and application of physical force to plant cell surfaces in a behavioral context (Clissold 2008).

At the cellular level, mechanostimulation of plant cells and protoplasts in the laboratory affects movements of their nuclei (Qu and M.-X. Sun 2007). Mechanostimulation of Arabidopsis plants by air currents or touch may also alter gene expression by upregulating touch-inducible (TCH) genes (Braam and Davis 1990). Lee et al. (2005) found that 589 genes are upregulated in Arabidopsis thaliana by touch including those of the cytochrome P450 family, genes encoding calmodulin-related proteins, APETALA2, and WRKY transcription factors. The APETALA2 gene encodes one of the class A cis-acting transcription factors in the ancestral developmental tool kit of land plants (Floyd and Bowman 2007).

Transduction of signals that originate from abiotic and biotic sources in environments external to the stationary plant body potentially involves cell walls, the plasmalemma, Hechtian strands, microtubules, calcium-binding proteins, and expression of genes of the CAM/CML and XTH gene families (Lee et al. 2004, Braam 2005). Microtubules in plant cells convey signals from the cell wall, plasmalemma, and cytoplasm to other parts of the cell, including the nuclear envelope and endoplasmic reticulum (Wasteneys 2004).

Did extinct plants have the anatomical machinery to convey mechanostimulatory and hormonal signals from insect bodies rubbing against cell walls, to adjoining membranes, microtubules, and the nucleus? Yes.

Intact protoplasm containing fossilized plastids and secondary plant substances has been identified in the cells of Paleogene "green-leaf" fossils of dicotyledonous flowering plants (Giannasi and Niklas 1981, Niklas and Brown 1981, Niklas 1982). Calcium carbonate permineralizations of pteridophyte (Tubicaulis fern) sieve elements yield fine cell wall structure (Smoot and Taylor 1984). Xin and Y. Wang (2006) report that 17-million year old fossil plants possess the machinery for transcytosis; a physio-anatomical syndrome needed for signal transduction (Jaffe et al. 2002).

Is there anatomical evidence preserved in the fossil record that cells have responded to touch? Possibly.

When fossilized leaves containing chew marks and punctures are permineralized, the preserved plant remains exhibit thicker cells surrounding the damaged tissues (Labandeira 1998).

Wind currents and moving parts of large herbivorous animals potentially had thigmomorphogenic effects on development of pollen dispersing reproductive modules (microsporophylls, microsporangia, synangia, stamens) and ovular points of attachment (abaxial or adaxial) on large, seed-bearing leaves, but paleontologic evidence in support of this idea is lacking.

In certain seed plant species constant exposure to wind currents results in endogenous biosynthesis of ethylene, signaling of primary and secondary meristems, leading to differentiation, morphogenesis, and wood formation (Rowe and Speck 2005).

Certain geometries of ovule-bearing shoots in extant gnetophytes (among other plant groups) affect the aerodynamics of pollen, possibly resulting in differential reproductive success (Niklas et al. 1986).

Thigmomorphogenesis in plants is incompletely understood but some progress has been made in understanding the effects of touch stimuli on plant differentiation and the derivation of plant growth forms (Niklas 1998, Pruyn et al. 2000, Pigliucci 2002, Lee et al. 2005, Rowe and Speck 2005, Telewski 2006, Fluch et al. 2008).

I now consider ecological, climatic, and geochemical aspects of global mass extinctions which had potentially profound effects on ancient animal and plant interactions. Biotic interactions probably took place at all levels from biomes to plant-herbivore compartments, to venues in tissues and cells of Paleozoic insects and seed plants. Coevolution of insect and plant developmental tool kits is proposed to explain the origin of angiosperms, flowers, and certain clades of Holometabola.

Further, I propose that the origin of flowers and a coevolutionary origin of seed plant lineages leading to angiosperms was tied-in with the rise and fall of atmospheric oxygen and greenhouse gases over more than 400 million years of geologic time.

Ecologic Considerations:

Was angiospermization a consequence of coevolution between seed plants and resident insect mutualists, reinforced by global ecologic catastrophe, temperature extremes, or hypoxia?

The image to the right is an artist's reconstruction of an early Triassic landscape a few thousand years following the end-Permian extinction (EPE). The Hercynian Mountains are in the distance and the dark regions to the left are decayed remains of coastal gigantopteroid mangrove at the edge of the Panthalassa Sea.  Lighter areas on land represent acidified ground probably unsuitable to mycorrhizal fungi and some vascular plants.

"To begin to connect ecology to the early radiation of angiosperms, we must understand the ecological starting point for angiosperm evolution."

The preceding statement is quoted from page 385 of T. S. Feild and N. C. Arens (2005), Form, function and environments of the early angiosperms: merging extant phylogeny and ecophysiology with fossils, New Phytologist 166: 383-408.

Little if any vegetation and few insects and vertebrates survived the end-Permian mass extinction. Early anthophytes may have survived and radiated in several biomes including those of low elevation coastal and desert regions of Pangaea.

Paleoecology of global catastrophe. Among the planet's untold episodes of climatic and ecologic catastrophe (including fire, flood, glaciation, drought, global warming, and volcanism) several events stand-out as "extinction-level" global cataclysms. These include the late Devonian-early Carboniferous ice-house (DeCARB) (McGhee 1996), the carbon cycle anomaly marking the end of the Guadalupian Epoch (GuCCE) (Lai et al. 2008); the end-Permian extinction (EPE), also known as the "Great-Dying" (Reichow et al. 2009); the Triassic-Jurassic boundary carbon cycle event (TrCCE) (Cleveland et al. 2008), the early Cretaceous end-Barremian biogeochemical perturbation (BaCCE) (Heimhofer et al. 2005), and the Kretaceous-Tertiary asteroid impact (K-T Event) (Nichols and Johnson 2008).

"Many paleontologists persist in using the number of species (or higher taxa) present as a metric of recovery. But the response to a mass extinction is above all an ecological one, and the number of taxa is only an indirect measure of evolutionary activity. Understanding these events requires an ecological approach. What are the dynamics and interactions between species? How well were ecosystems functioning? How complex were food webs and other ecological services during the survival and recovery intervals? Perhaps most importantly, how does the structure of the ecosystem facilitate the speciation involved in biotic recovery?"

The preceding quotation is from page 241 of D. H. Erwin (2006), Extinction: How Life on Earth Nearly Ended 250 Million Years Ago, Princeton, Princeton University Press, 306 pp.

Late Devonian-early Carboniferous ice-house (DeCARB). One of the greatest mass extinctions on Earth occurred during the Frasnian Age of the Devonian Period, 375 million years ago (McGhee 1996, Ward et al. 2006, Bishop et al. 2009). During the late Devonian Period island fragments and a chunk of Gondwana were clustered south of the equator and in the south polar region.

The late Devonian environmental crisis on land 375 to 345 million years ago (m.y.a.) resulted in an hypoxic ice-house with oxygen levels at 13% (Berner 1999, Berner and Kothavala 2001, Huey and Ward 2005, Ward et al. 2007) during the time interval known as Romer's Gap (Ward et al. 2006). The warm interval preceding the DeCARB was the first phase of the expansion of invertebrate arthropod herbivores in terrestrial environments (Labandeira 2006).

According to Ward et al. (page 16820, 2006) "no new clades" of arthropods occurred during the low oxygen interval demarcating Romer's Gap. Adaptive radiation of arthropods, land plants, and early limbed vertebrates (stegocephalians) occurred on land during the millions of years following Romer's Gap (Ward et al. 2006). As the ice-house Earth slowly warmed during the Carboniferous Period, global oxygen concentration increased to levels exceeding 31% (Berner and Kothavala 2001, Ward et al. 2006), coinciding with the second phase of arthropod herbivore expansion (Labandeira 2006).

"While mass extinction and recovery, accommodation to new habitats, and body-plan modernization have all been factors in the evolution of life, atmospheric O₂ level also has been a major, if underappreciated, driver in the major features of life's history. These features include origination, diversity levels, and probably absolute size increases through time."

The preceding quotation is from page 16821 of P. Ward, C. Labandeira, M. Laurin, and R. A. Berner (2006), Confirmation of Romer's Gap as a low oxygen interval constraining the timing of initial arthropod and vertebrate terrestrialization. Proceedings of the National Academy of Sciences 103(45): 16818-16822.

While certainly not applicable to events following Romer's Gap, I suggest that evolutionary biologists also consider decreases in insect body size and cone- (or floral-) axis length as likely adaptations emerging in coevolving species pairs resulting from the selective forces of the two Paleozoic mass extinctions and later Triassic-Jurassic carbon cycle event.

The Permian Period, despite evidence of episodic glaciation, saw a continuation of climatic warming, which was punctuated by at least two global mass extinctions (see next section). Evidence from the study of fossil soils (paleosols) supports stratigraphic division of the Permian Period into three epochs: Cisuralian (299-270 m.y.a.), Guadalupian (270-260 m.y.a.), and Lopingian (260-251 m.y.a.) (Retallack 2005).

End-Guadalupian carbon cycle event (GuCCE). Retallack et al. (2006) present evidence of a global carbon isotope anomaly coinciding with the close of the Guadalupian Epoch of the late Permian Period. Oxygen levels in Earth's atmosphere declined significantly in the final days of the Guadalupian triggered by ignition of coal seams with hot igneous intrusions and extruding lava (Lai et al. 2008), and by expulsion of carbon monoxide and carbon dioxide into the atmosphere (Retallack et al. 2006, Retallack and Jahren 2008).

Stratigraphic evidence from the Emeishan carbonates and volcanics of Asia (Wignall et al. 2009) and magnetostratigraphy of the Illawara Reversal (Isozaki 2009) suggests a trigger for the GuCCE. Methane clathrates stored in cold undersea sedimentary layers were disrupted by volcanic activity releasing gigatons of ozone-damaging CH₄ into the atmosphere (Knoll et al. 2007).

The Earth slowly became a hot house with far-reaching ecological effects on late Permian terrestrial ecosystems, as determined from study of the geochemistry of fossil soils and coals, and biostratigraphy of plant and vertebrate fossils (Retallack et al. 2006).

Invertebrate paleontologists, however, view the end-Guadalupian event as a prolonged biotic crisis that led to a gradual decrease in marine biodiversity from the Wordian to end of the Changhsingian, about 253 million years ago (Clapham et al. 2009). The chart pictured below illustrates the rise and fall of oxygen levels in Earth's atmosphere, timing of the DeCARB, and the five greatest mass extinctions (GuCCE, EPE, TrCCE, BaCCE, and K-T).

The graph below is based on several data sets (Berner 1999, Huey and Ward 2005, Ward et al. 2006) including GEOCARB III (Berner and Kothavala 2001) and GEOCARBSULF (Berner 2006). The graphic does not show the geologic interval of the Neoproterozoic snowball Earth, or periods of less severe global temperature swings (ice-house to hot house and vice-versa, except the DeCARB), and the Oligocene-Eocene cooling interval of the Paleogene Period.

At the Permian-Triassic boundary, global oxygen levels on the shores of the Panthalassa Sea further declined to a level of free atmospheric oxygen less than 12% (Ward et al. 2006). Carbon dioxide levels in the atmosphere were elevated with profound effects on the physiology of aquatic and terrestrial animals (Knoll et al. 2007).

When taking into account supposed adaptations of terrestrial animals to pre-Guadalupian high oxygen levels of 31%, and effects of altitudinal compression, animals might have experienced hypoxia comparable to the higher Himalaya or the summits of Kilimanjaro and Mount McKinley (Huey and Ward 2005). Hypoxia at sea level and in the Hercynian Mountains was probably exacerbated by carbon dioxide and sulfur dioxide outgassing linked to catastrophic volcanism that led to hotter climates within 600,000 years of the global apocalypse marking the close of the Paleozoic Era (Retallack et al. 2006).

End-Permian extinction (EPE). Global catastrophe caused extinction of most life at the close of the Permian Period, 251.3 million years ago (Erwin 2006, Knoll et al. 2007). Known as the "Great Dying," the apocalypse coincides with a 2,000,000 million year interval of volcanic unrest causing extrusion of epicontinental lava lakes known as the Siberian Traps dated precisely 250.3 million years ago (Reichow et al. 2009). Geochemical studies of at least one subvolcanic intrusive rock formation associated with the Siberian Traps reveal high levels of sulfur contamination of the magma which potentially exacerbated the effects of SO₂ outgassing on the atmosphere (C. Li et al. 2009).

Several other hypotheses have been proposed that might explain the greatest of all of Earth's mass extinctions. These include collision of the Earth with a bolide, comet, or icy dwarf; release of thousands of gigatons of methane gas from methane clathrate deposits in the undersea bed, methane gas derived from burning coal measures ignited by volcanic activity (Retallack and Jahren 2008), disruption of the Earth's magnetosphere by galactic dust and gas clouds, or catastrophic atmospheric disturbances attributable to solar flaring (Benton and Twitchett 2003).

However, no single event listed in the preceding paragraph explains the oceanic δ¹³ data gleaned from paleogeochemical and isotopic studies of carbonates in sediments bracketing the Permian-Triassic (P/Tr) boundary (Berner 2002). Lipid biomarkers suggest a prolonged period of ecological disruption in marine environments preceding and following the main global kill event (Cao et al. 2009).

Important insight on the effects of the GuCCE on late Permian animals and plants, and the EPE "trigger and kill" mechanisms (page 296, Knoll et al. 2007) on early Triassic aquatic and terrestrial ecosystems, has been gained from detailed biostratigraphic, paleoclimatic, and paleophysiologic studies (Z.-Q. Wang and Zhang 1998, Looy et al. 1999, Z.-Q. Wang 2000, Rees 2002, Rees et al. 2002, Berner 2005, Gastaldo et al. 2005).

Recovery of biota on land probably took a very long time relative to other global extinctions according to the review by Grauvogel-Stamm and Ash (2005). However, recovery of the insect fauna following the EPE is unclear, and later radiation of arthropod clades is even more nebulous due to a general lack of meaningful paleontologic data (Béthoux et al. 2005).

Zi-Qiang Wang and co-workers in a series of papers report that Cathaysian floras were reduced to isolated, fungal-enriched refugia dotted within arid, desolate landscapes by late Permian time. Some of the Upper Permian sedimentary red beds studied contained only charcoal fragments of vascular plants while some stratigraphic sequences up to 70-meters thick were devoid of plant fossils (Z.-Q. Wang 2000). A later paper by Peng and Shi (2009), which is based on detailed biostratigraphic data from continuous sections across the Permian-Triassic boundary (PTB), suggests uneven recovery of the Cathaysian Gigantopteris flora following the EPE.

When seed plant compressions of gigantopterids were found the fossilized leaves had thick cuticles suggestive of drought conditions in Cathaysia. Based on biostratigraphic and taphonomic data cycads, peltasperms, ferns, ginkgoes, and certain other conifers were represented in the fossil record on both sides of Permo-Triassic boundary (Z.-Q. Wang 2000).

Paleogeochemical data show a rise in the proportion of rate of burial of sulphur and iron-containing pyrite over organic carbon burial rate at the Permo-Triassic boundary, 251.3 m.y.a., which was probably attributable to forest decline at that time (Fig. 4, page 3214, Berner 2005). Some estimates based on geochemical data suggest a catastrophic drop in atmospheric oxygen over a 20 million-year period following the end-Permian apocalypse (Berner 2006).

The late Devonian-early Carboniferous ice-house, GuCCE, EPE, and Triassic-Jurassic boundary carbon cycle event (see below) had a common thread: significant fluctuation in global carbon dioxide and oxygen levels (Berner and Kothavala 2001, Huey and Ward 2005, Berner 2006, Ward et al. 2006).

Proportional changes in the relative amounts of oxygen, carbon dioxide, and methane gas at the ground surface, when accompanied by temperature swings, might have affected low-elevation terrestrial biomes with far-reaching global paleoclimatologic and paleoecologic consequences (Looy et al. 1999, Barrett and Willis 2001, Huey and Ward 2005, Retallack et al. 2006). Based on biostratigraphic studies of Late Permian and earliest Triassic (Scythian Age) rocks in East Greenland (Looy et al. 2001) ecosystem recovery may have varied spatially and temporally from place to place on the Earth.

Paleontologic evidence from climatic and geophysical phenomena listed above includes on fossil soils (paleosols) of Antarctica, southwestern North America, and southern Africa (Retallack 2005, Retallack et al. 2005), and from taphonomy of fossorial therapsids and paleosols preserved in Permian rocks (Retallack et al. 2003). Conditions on land following the EPE were inhospitable to insects, tetrapods, and plants (Z.-Q. Wang 2000, Retallack et al. 2006).

However, reinterpretation of channel-fill deposits in the Bethulie region of the Karoo Basin of South Africa suggest that the EPE was a non-event (Gastaldo et al. 2009). I disagree with the far-reaching conclusions of Gastaldo and company, which are based on stratigraphic studies of sediments deposited in a single region of Permo-Triassic Pangaea. More geochemical and stratigraphic work is clearly needed to understand the magnitude of disturbances of Siberian Traps volcanism, sulfur dioxide outgassing, acid rain, and effects of hypoxia on terrestrial animals, fungi, and plants at a global scale, including southern reaches of Permo-Triassic Pangaea.

Paleoclimatology is the subject of active research (Berner 1999, Barrett and Willis 2001, Berner and Kothavala 2001, Beerling and Berner 2002, Huey and Ward 2005, Ward et al. 2006, Lamarque et al. 2006, A. C. Scott and Glasspool 2006). Refinements in the global oxygen level curve from reanalysis taking into account the effects of weathering of sulfur-containing iron pyrites on atmospheric free oxygen "... may have an important, relatively unrecognized relation to animal evolution." (page 5663, Berner 2006).

Triassic-Jurassic boundary carbon cycle event (TrCCE). The end-Triassic mass extinction was preceded by a global increase in temperature followed by severe climatic disruption over many millions of years (Cleveland et al. 2008). Many studies suggest that the mass extinction was prolonged and connected with at least two intervals of higher carbon dioxide levels in the atmosphere during the Rhaetian Age preceding the TrCCE (Cleveland et al. 2008, Ruhl et al. 2009).

Relatively little is known of the effects of the Triassic-Jurassic boundary carbon cycle event on terrestrial vegetation (D. H. Erwin 2007). Studies on the classic Triassic floras of Greenland suggest that tropical forests of this boreal region began to change in composition and stratification at the onset of global warming well before the TrCCE (McElwain et al. 1999, McElwain et al. 2007).

Beerling and Berner (2002, abstract) state that "the end-Triassic mass extinctions represent one of the five most severe biotic crises in Earth history, yet remain one of the most enigmatic." Further, Beerling and Berner (abstract, 2002) propose that global warming, "due to a buildup of volcanically derived CO₂," triggered "destabilization of seafloor methane hydrates and the catastrophic release of CH₄ [Palfy et al. 2001]."

The two papers by Beerling and Berner (2002, 2004) present data and findings that disruption of the global carbon cycle at the end of the Triassic Period resulted in the release of up to 9,000 gigatons of carbon (as carbon dioxide gas) during the CAMP (Central Atlantic Magmatic Province) basalitic eruptions, and an additional 5,000 gigatons of methane released from methane hydrates imbedded in the seafloor.

Early Cretaceous end-Barremian biogeochemical perturbation (BaCCE). Biogeochemical evidence exists suggesting that a global atmospheric and climatic perturbation occurred in the Late Barremian (Neocomian, early Cretaceous). The early Cretaceous end-Barremian biogeochemical perturbation (BaCCE) might have accelerated the diversification of early magnoliid flowering plants and possibly monocots (Heimhofer et al. 2005).

Paul M. Barrett and Katherine J. Willis (page 437, 2001) report that elevated levels of carbon dioxide in Earth's atmosphere during the Cretaceous Period coincide with "... emergence of the major plant groups ..." and "... increasing rates of species turnover ..." Further, the initial radiation of flowering plants may be attributable to carbon-dioxide build-up in the atmosphere according to an interesting hypothesis by Barrett and Willis (2001).

Implications of the BaCCE and other thermal maximums toward the colonization of ice-free landscapes, and diversification and speciation of flowering plants and their invertebrate and vertebrate antagonists, is discussed by Barrett and Willis (2001) and Friis et al. (2006).

Ecology of Paleozoic seed plants. Paleozoic seed plants did not possess flowers and fruits. Rather ovules and seeds (ripened ovules) were probably phyllospermous being attached to fertile leaves termed megasporophylls.

The best studied example of a fossilized megasporophyll is probably Phasmatocycas bridwellii (Axsmith et al. 2003). However, paleobotanists do not know the whole plant morphology or taxonomic relationships of the gymnosperm to which Phasmatocycas leaves were attached.

Were ovulate Phasmatocycas bridwellii leaves (Axsmith et al. 2003) detached pieces of an enormous bisexual cone axis with the pollen-bearing leaves (microsporophylls) and strap-shaped bracteopetals missing?

To answer this question is one of many paleobotanical challenges discussed in the second essay that require a coordinated attack strategy.

"... Nonetheless, it is possible that co-expression of AP3- and PI-homologs [B-class MADS-box genes, see Chapter on Genetics Considerations] mediated the evolutionary innovation of animal-attractive, petal-like organs well before the appearance of flowers."

A coevolutionary origin of flowering plants based on co-radiations between specific groups of phytophagous insects and seed plant hosts is discussed in several papers (Thien et al. 1985, Farrell 1998, Grimaldi 1999, Pellmyr and Leebens-Mack 1999, Bernhardt 2000, Wilf et al. 2000, Gorelick 2001, Crepet and Niklas 2009). Some of these studies however, focus on possible coevolution of arthropods and the first angiosperms during the Jurassic or Cretaceous periods of the Mesozoic Era, not the Carboniferous or Permian periods of the Paleozoic Era.

The "right insect traits" (page 422, Bronstein et al. 2006) to study are helix-turn-helix proteins encoded by the compartment-selector gene engraled and certain homeotic selector genes of the Hox complex (e.g. homologs and paralogs of abd-A, Abd-B, Hox3, pb, Scr, and Ubx). Bronstein and company's traits of focus might also include enzymes belonging to the respiratory, moulting, and vitellogenin developmental tool kit.

Specific invertebrate respiratory protein phenotypes of potential evolutionary interest are hemocyanins (Hcs), hemoglobins (Hgs), and hexamerins (Hxs) (Van Holde et al. 2001, Burmester and Hankein 2007, Hagner-Holler et al. 2007).

The diagram below illustrates how signals originating in the bodies of phytophagous insects might potentially affect the biochemical machinery of differentiating cells in SAMs of certain developmentally plastic, late Paleozoic host seed plants. Further, host plants might have manufactured phytoecdysones (including brassinolide hormones) that potentially signaled the developmental, respiratory, and storage protein biosynthetic machinery of insect antagonists inhabiting Paleozoic and early Mesozoic vegetation compartments.

Clues on whether or not hypoxia is a selective force come from ecologic, physiologic, and rearing studies of extant insects (Gorr et al. 2006). The Holometabola exhibit complete metamorphosis of the fertilized egg to a wingless larval stage, a quiescent pupal stage, culminating with reproductively mature adults (page 331, Grimaldi and Engel 2005).

Hypoxia in extant terrestrial environments is defined as an oxygen concentration below ca. 20% (Hoback and Stanley 2001). Adaptations to hypoxia in holometabolous insects are specific to developmental stages. According to the review by Hoback and Stanley (2001), several kinds of extant insects are adapted to hypoxic environments including Coleoptera (beetles and weevils), Collembola (springtails), Diptera (flies), Isoptera (termites), Lepidoptera (butterflies and moths), and Orthoptera (grasshoppers and katydids).

Tenebrio molitor beetles exhibit 80% larval mortality, decreased instar body mass, and an increased number of moults when reared in an atmosphere of 10% oxygen in a nitrogen (Greenberg and Ar 1996). Aquatic insects are generally more susceptible to hypoxia than terrestrial insects (Hoback and Stanley 2001). Hoback and Stanley (page 540, 2001) conclude "... that the developmental plasticity of insects facilitates morphological changes in respiratory systems in chronically hypoxic situations."

Choice of respiratory enzymes (hemocyanins) and moulting storage proteins (hexamerins) as some of the right insect traits to study ties-in with the evolution of insect legs and wings from bilaterian gills. Ancient insect wings probably functioned as respiratory organs. Wings, halteres, arachnid spinnerets, and insect legs are all organs that develop from limb fields of cells where Ubx expression is prevalent (S. B. Carroll et al. 2005).

At the same time that hexamerin molecules were evolving in populations of hemi- and holometabolous insects and colonies of Carboniferous and Permian vascular plants, trampling and secreting insect antagonists might have sent signals to the developmental machinery including MIKC-type 2 MADS-box genes inside the nuclei of growing cells of shoot apical meristems (SAMs) and bisexual cone axes of host seed plants.

"Coevolution is a well characterized process that takes place at all biological levels, from ecosystems to molecules."

The preceding quotation is from page 934 of Juan et al. (2008), High-confidence prediction of global interactomes based on genome-wide coevolutionary networks, Proceedings of the National Academy of Sciences 105(3): 934-939.

Understanding the selective forces behind molecular coevolution of insect and seed plant developmental cis-acting regulatory modules in deep-time i.e. hundreds of millions of years, is an important first-step in unraveling the enigmatic origin of angiosperms, evolutionary development of the first flower, and adaptive radiation of certain clades of coevolving Holometabola.

Coevolution of seed plants and tetrapods. Weishampel and Jianu (2000), in an extensive review of dinosaurian herbivory, cast doubt on the coevolutionary-based idea that ornithischian tetrapods had an effect on the rapid radiation of flowering plants during the Cretaceous Period. Dinosaurians probably did not play a key role in the contrived Cretaceous Terrestrial Revolution (G. T. Lloyd et al. 2008).

According to Weishampel and Jianu (page 137, 2000), three geologic intervals of diversification of euornithopod transverse-grinders occurred during the Cretaceous Period "with the middle peak arguably associated with the origin of angiosperms." However, when ghost lineages are added to the cladistic analyses "the middle peak disappears" (page 137, Weishampel and Jianu 2000).

"Whatever coevolutionary interactions may have existed between flowering plants (angiosperms) and ornithischian dinosaurs, they are not apparent at the onset of the evolutionary radiation of angiosperms, as least as far as diversity patterns are concerned."

The preceding quotation is from page 139 of Weishampel and Jianu (2000), Chapter 5, Plant-eaters and ghost lineages: dinosaurian herbivory revisited. Pp. 123-143 In: H.-D. Sues (ed.), Evolution of Herbivory in Terrestrial Vertebrates: Perspectives from the Fossil Record, Cambridge: Cambridge University Press, 256 pp.

Tiffney (2004), in an extensive review of vertebrate dispersal of seed plants, concludes that tight coevolution between animal dispersers and plant is rare. Biotic dispersal of seed plant propagules probably arose through "transfer of function," according to Tiffney (page 6, 2004). Further, he suggests that "the Mesozoic dynamic was dominated by reptile-gymnosperm interactions to which the angiosperms were newcomers" (page 20, Tiffney 2004).

Fitness landscapes. How do we apply Sewall Wright's adaptive (fitness) landscape concept toward a better understanding of evolutionary ecology of biome fragments, paleohabitats, seed plant shrubs, trees, mycorrhizal fungi, vertebrate herbivores, and resident phytophagous insects of Pangaea? Niklas (1997) expands and modifies Wright's fitness landscape concept to better understand the evolution of early land plants.

Shrub lifeboats. Pockets of air trapped by sheathing, leathery photosynthetic leaves of a monopodium and other crevices in organs of arborescent and herbaceous seed plant species might have had higher amounts of oxygen gas than the surrounding atmosphere of Devonian, Carboniferous, Permian, and Triassic time.

I name the individual seed plant shrubs and small trees of hypoxic late Devonian-early Carboniferous, Permian, and Triassic fitness (adaptive) landscapes that house herbivorous insect antagonists in oxygenated air spaces and crevices, shrub lifeboats.

A higher partial oxygen pressure within the air spaces between host plant leaves could potentially protect resident insect mutualists from atmospheric disruption and hypoxic conditions attributable the late Devonian-early Carboniferous ice-house, GuCCE, EPE, TrCCE, and BaCCE. Trapped oxygen gas of air pockets in the host shrub might have been critical for the survival and metamorphosis of instars (fertilized eggs, larvae, pupae).

Optimal thermogenic microhabitats cast from food-bearing, sheathing leaves and woody tissues could have sustained a resident insect and tetrapod population during times of temperature extremes, hypoxia, and high ultraviolet radiation levels.

The diagram on the left side of the page illustrates a hypothetical seed plant shrub lifeboat. The monopodium with its shoot apical meristem (SAM) and large, leathery sheathing leaves is shown from an overhead view. The resident insect colony of eggs and instars (or larvae, pupae, and adults) shown as red-colored objects occupies the oxygenated air spaces that surround the growing tip and hollows formed by clasping leaves of the shrub.

Invoking coevolutionary nomenclature drawn from studies of extant biomes, a shrub lifeboat might also be viewed as a compartment housing antagonists (herbivores and predators) in a locally hot, hypercapnic, hypoxic, Permo-Triassic fitness landscape.

Did perennial seed plant shrubs furnish respiring insects with cavities and air spaces having a higher partial pressure of oxygen than the surrounding hypoxic oxygen desert?

What happened to complex food webs, plant communities and habitats of early Triassic time following the disruptive and extirpatory forces of the end-Permian extinction (EPE)?

Some early Paleozoic forests of conifers, ferns, lycophytes, and sphenopsids were stratified and contained complex food webs involving decomposers, primary producers, mycorrhizal fungi, and populations of invertebrates and amphibians (Z.-Q. Wang 2000).

Were elevated carbon dioxide levels (hypercapnia) in the atmosphere of late Permian time, when combined with low oxygen levels (hypoxia), a selective force in animal populations, including terrestrial arthropod invertebrates and vertebrate tetrapods?

Knoll et al. (2007) review the possible physiological effects of hypercapnia on animals and plants in connection with the end-Permian global apocalypse and biotic recovery of early Triassic aquatic and terrestrial landscapes. Geochemical studies suggest that late Permian and early Triassic air of Pangaean shorelines had oxygen levels comparable to the highest summits (>5000 meters) of most modern mountain ranges (Huey and Ward 2005).

Based on studies of fossils, paleosols, and geochemistry it was possible that during millions of years of Paleozoic global climate changes, complex networks of generalist and antagonist animal and plant species broke down and/or were disrupted (Retallack et al. 2003, Retallack et al. 2005).

Hypoxic ice-house late Devonian-early Carboniferous fitness landscapes. The ice-house terrestrial paleoenvironments of the late Devonian and early Carboniferous Periods (DeCARB) might have been challenging to the respiratory systems of invertebrates including ancient plecopteran insects (Labandeira 2006). The DeCARB was also a time of seed plant innovation (Rothwell et al. 1996).

Early arthropods exhibited an explosive radiation of molecular novelties of the hemocyanin class of gas-binding respiratory enzymes during the hypoxic DeCARB. At least seven clades of hemimetabolous and holometabolous insects diverged from the thawing ice-house as evidenced by molecular clock estimates of hexamerin moulting storage proteins (Hagner-Holler et al. 2007).

Was the burst of evolution of insect hexamerin proteins from hemocyanins somehow linked to low temperature and low oxygen levels in late Devonian-early Carboniferous terrestrial paleoenvironments?

Hypoxic and hypercapnic hot-house Permo-Triassic fitness landscapes. How did terrestrial invertebrates and vertebrates cope with high carbon dioxide content and low levels of oxygen in the global atmosphere (<12%) following the end-Guadalupian carbon cycle event (GuCCE) and end-Permian extinction (EPE)? Was hypoxia greater at higher altitudes in, for example, the Hercynian Mountains?

What was the effect of oxygen depletion (hypoxia) and elevated carbon dioxide (hypercapnia), methane, and hydrogen sulfide levels on animals and plants indigenous to aquatic and terrestrial environments?

High carbon dioxide levels in the atmosphere and water column preceding and immediately following the end-Permian extinction (EPE) might have had profound effects on the physiology of animals (Knoll et al. 2007). According to Knoll et al. (page 302, 2007), "... hypercapnia can reduce respiratory capacity, exacerbating the physiological effects of environmental oxygen limitation."

"It is possible, therefore, that the high CO₂ [Silurian] concentrations acted as an environmental 'barrier' that prevented the origination and spread of large megaphylls. Once that environmental barrier was removed, ecological processes, especially competition between species for light, soil and water nutrients, would have driven both the co-evolution of the root and shoot and the well-documented increase in plant height (Beerling and Berner 2004)."

The previous statement is quoted from page 10 of David J. Beerling and Andrew J. Fleming (2007), Zimmermann's telomb theory of megaphyll leaf evolution: a molecular and cellular critique, Current Opinion in Plant Biology 10: 4-12.

Did Paleozoic seed plant shrubs and trees (or even herbs) furnish respiring insect larvae with cavities and air spaces having a higher partial pressure of oxygen than the surrounding hypoxic landscape? Could plant organs, tissues, and waxy cuticles shield insect eggs and instars from harmful ultraviolet solar radiation?

The end-Permian extinction marking the beginning of the Triassic Period of the Mesozoic Era also had devastating effects on growing substrates including extirpation of mycorrhizal fungi and certain soil bacteria. Global terrestrial and aquatic environments became acidified by 600,000 years of global volcanism and more than a million years of precipitation containing dissolved sulfur dioxide (SO₂) outgassed from magma underlying the Siberian Traps, a name applied by geologists to early Triassic epicontinental lava lakes.

The art board below shows a community of potentially coevolving animal antagonists that inhabit shrubs and trees of a biome fragment of an hypoxic late Permian fitness landscape based on descriptions by DiMichele and Hook (1992). A biome fragment near a water source in an otherwise arid landscape is depicted with overstory conifers and understory scrub, which is punctuated by dicynodont burrows. Herbivorous dicynodonts were probably the most common vertebrate animals of such late Permian landscapes (DiMichele and Hook 1992).

The vertebrate tetrapod population resided with the oxygen generating vegetation consisting of seed plant shrubs and coniferous trees, and an understory of bryophytes, ferns, horsetails, lycopsids, and seed ferns. Seeds rained on to the floor of the biome forming a seed bank sensu Harper (1977), and bisaccate pollen was transported to other shrubs and trees. Insect colonies and free-living invertebrates (antagonistic generalists and predators) are shown in dark red.

Free-living insects and certain generalists were probably only able to crawl or fly short distances from the oxygenated crevices of one shrub lifeboat to another. The surrounding hypoxic desert might have been lethal to most respiring animals. The increasingly small body size of Triassic invertebrates and larger lung cavities of certain tetrapodial vertebrates might have been adaptations of these animals to withstand oxygen deprivation.

The preceding diagram emphasizes one possible fate (of many) of a biome fragment immediately following the EPE. In this case, the biome fragment is dominated by antagonistic networks of interacting plants and animals compartmentalized around oxygen-generating vegetation including shrub- and tree-like lifeboats.

Paleoecological processes in Paleozoic animals and plants were probably not homologous to living systems. Inextricably linked to paleosuccession in Permo-Triassic terrestrial ecosystems following disturbance was the feeding behavior of herbivorous and insectivorous amphibians and early reptiles including archosaurians, dicynodonts, and therapsids, among other tetrapods (Tiffney 1992).

What were adaptations of land plants available for exploitation by resident insect colonies?

Were these invertebrate colonies subject to predation by insectivorous amphibians and reptiles?

How did the activities of dicynodonts, therapsids, and archosaurian dinosaurs (among other vertebrates) affect the evolution, diversification, and dispersal of seed plants and their angiospermization?

Hypocapnic and hyperoxic Cretaceous fitness landscapes.  McElwain et al. (2005) in a review chapter on the effects of atmospheric carbon dioxide on angiosperm origin, radiation, and paleodiversity conclude:

"It is likely, therefore, that a long-term trend of CO₂ decline through the Cretaceous period may have played an extremely important role in the diversification of angiosperms ..."

The previous phrase is quoted from page 158 of J. C. McElwain, K. J. Willis, and R. Lupia, 2005, Chapter 7. Cretaceous CO₂ decline and the radiation and diversification of angiosperms. Pp. 133-165 In: J. R. Ehleringer, T. E. Cerling, and M.-D. Dearing, A History of Atmospheric CO₂ and its Effects on Plants, Animals, and Ecosystems. New York: Springer, 530 pp.

Cretaceous angiosperm diversification might have been triggered by abiotic effects of carbon dioxide decline (hypocapnia) on the physiology of flowering plants as McElwain et al. suggest. However, a concomitant rise in atmospheric oxygen content probably also had physiologic effects on insect antagonists of early Aptian angiosperms.

Rapid floral diversification at the population and species level might also be traced to the potential genetic release of insect-induced molecular effects on the host plant developmental machinery, attributable to shifts in coevolving pollinator populations away from host plant species to other oxygen-enriched habitats. The notion of genetic release in a host plant-insect antagonist coevolutionary relationship at the level of transcriptional regulators is unsupported by scientific data.

However, a coevolutionary arms race at the molecular level between populations of coevolving flowering plant hosts and insect antagonists is supported by real-time experimental studies. Phylogenetic analyses of species groups with divergent floral characters point toward recurring pollinator shifts in the evolutionary history of certain parallel lineages, attributable to natural selection.

Further, evolutionary-developmental studies using biochemically altered probes, genetically-engineered transcription factors, and microRNAs demonstrate the lability of flowers and insect bodies. Does the inherent lability of floral and insect organs inferred from laboratory studies have a molecular genetic basis applicable at the population level to coevolving plants and animals?

The next two chapters consider scientific evidence drawn from plant physiology and genetics critical in understanding the Paleozoic beginnings and Mesozoic adaptive radiation of the first angiosperms.

Physiologic Considerations:

This chapter considers evidence from the physiology of interacting plants and insects from a different research perspective. A discussion of phytohormones and defense substances and insect developmental and respiratory proteins, hormones, and neurotransmitters is necessary to understand possible molecular coevolution of insect and seed plant developmental tool kits, animal-plant signaling, and the ongoing coevolutionary arms race (Pauw et al. 2009).

Insect development is apparently controlled by cis-acting transcription factors (S. B. Carroll et al. 2005) and hormone signals acting in regulatory networks and cascades (Bonneton et al. 2008). Further, structural similarities exist in the trihelical bundle of certain helix-turn-helix cis-regulatory homeodomain proteins of the model insect Drosophila melanogaster and seed plant Arabidopsis thaliana.

The image to the right is a 20 × 20 cm polyamid thin layer chromatogram representing successive gel filtration column fractions from a mixture of floral flavonoids (x-axis), partitioned in an aqueous solvent mixture (y-axis) by TLC. Development of the thin layer chromatogram and treatment with Natural Products Reagent A reveals a separation of flavonol mono-, di-, and triglycosides (viewed under long-wave ultraviolet light and captured on film). The yellow-orange spot to the right is a standard sample of quercetin-3-O-glucoside, a flavonol monoglycoside.

Natural products, phytohormones, and signaling. This section briefly discusses the high points of a great body of literature in plant physiology, which is devoted to bioactive secondary plant products. Certain natural products are natural insecticides and phytohormones.

According to Floyd and Bowman's phylogenetic analyses (2007) the auxin, ethylene, and gibberellin metabolic pathways are highly conserved and involve several ancient land plant enzyme systems including APETALA2/auxin response proteins (AP2/ARFs), DELLAs (transcription factors of the GRAS family), and ethylene response factors (ERFs).

Did Paleozoic seed plants manufacture bioactive plant natural products and phytohormones and signal moulting insect instars residing in oxygenated crevices and leaf bases of massive SAMs?

Likewise, did phytophagous insect antagonists of Permo-Carboniferous and Permo-Triassic seed plants signal the homeotic genetic machinery of host plants affecting SAM development and morphology?

An important clue comes from naturally occurring but fossilized triterpenoids known as oleanone triterpanes (oleananes). Oleananes, together with ursenes, lupenes, and taraxerenes are important taxon-specific biomarkers (TSBs) that belong to a class of Β-amirin triterpenoids (Moldowan and Jacobson 2002).

Certain flavonoids are antioxidants and inhibit auxin efflux carriers. Many flavonoids are implicated in hormone signaling, plant defense, and regulation of gene expression (Koes et al. 2005, Owens et al. 2008).

Biosynthesis of chalcone synthetase is affected by the hormone gibberellic acid (GA) in some flowering plants. Chalcone synthetase manufactures many classes of flavonoids involved as antioxidants, bee-violet attractants in flowers such as Rudbeckia hirta, and as ultraviolet radiation shields.

Steroids are well-known natural products manufactured by plants for use in defense and signaling. Insects are unable to manufacture cholesterol ring systems required for the biosynthesis of 20E-ecdysone, an important steroid in endocrine signaling, but must obtain precursor molecules from digested products of animal and plant tissues, which are then assembled by cytochrome P450 enzymes (Truman and Riddiford 2002).

Auxin down-regulates Class I KNOX genes in developing SAMs of extant flowering plant experimental systems permitting GA activity in leaf primordia and promoting growth (Weiss and Ori 2007). Membrane proteins known as efflux and influx carriers control auxin movement from cell to cell in the SAM.

Further, auxin response factor (ARF) genes encode cis-acting ARF proteins that bind to auxin responsive genes (Busov et al. 2008, Delker et al. 2008, R. S. Smith 2008). Auxin response factors figure prominently in the regulation of leaf morphology and belong to the ancestral developmental tool kit of land plants (Floyd and Bowman 2007). Interestingly, the MIKC type 2 MADS-box transcription factor SEPALLATA3 interacts with ARF and other elements of several hormone signaling cascades in Arabidopsis (Kaufmann et al. 2009).

Clues on the possible existence of polar auxin flow in woody stems have been uncovered in comparative anatomical studies of a 375 million-year-old fossilized progymnosperm known as Archaeopteris and modern conifers (Rothwell and Lev-Yadun 2005). The fossil record reveals that auxin regulation of secondary growth was a conserved physiologic process (Rothwell et al. 2008).

Based on the phylogenetic analysis of Class I KNOX genes (Floyd and Bowman 2007), paleobotanical evidence of polar auxin flow in Devonian plant fossils (Rothwell and Lev-Yadun 2005), and numerous examples of cross-talk between auxins and GA in extant vascular plants, it is not unreasonable to view auxins and gibberellins as components of tool kits of early diverging land plants.

Brassinosteroids (BRs) also known as brassinolides, are structurally similar to 20S-ecdysone, a potent insect steroid moulting hormone. Brassinolides are also implicated in endogenous ethylene production (J. Li and Chory 1999, Thummel and Chory 2002, Zullo and Adam 2002).

Ethylene is a gaseous hormone with profound effects on plant growth, development, and responses to biotic stress including herbivory. A family of five membrane-bound receptor proteins perceives ethylene molecules from the external environment. The ethylene signaling cascade crosstalks with the biosynthetic machinery of most other phytohormones including abscisic acid (ABA), auxins, brassinosteroids (BAs), and gibberellins (Y. F. Chen et al. 2005).

For example, ethylene adversely affects gibberellic acid (GA) by activating DELLA repressing proteins. In turn, DELLAs repress the floral integrator genes LEAFY (LFY) and SOC1 resulting in delayed flowering in Arabidopsis (Achard et al. 2007). Ethylene and gibberellic acid (GA) biosynthesis in plants is enhanced by insect neurotransmitters such as catecholamines (Schultz 2002).

Gibberellins are key diterpenoid hormones of the developmental tool kit of plants (Weiss and Ori 2007). The effects of nuclear regulatory growth repressing DELLA proteins that belong to the GRAS family of transcriptional regulators are relieved by gibberellic acid (GA) via the ubiquitin 26S proteasome biosynthetic pathway (Yu et al. 2004, Axtell and Bartel 2005, Lu et al. 2005, Jiang and Fu 2007, Weiss and Ori 2007, Gao et al. 2008, Hartweck 2008).

Studies of Arabidopsis provide evidence that GA controls the development and morphology of trichomes on leaves, flowers, and shoots by de-repressing DELLAs (Gan et al. 2007). Whether the mechanism of GA action on the negative effects of nuclear regulatory growth repressing DELLA proteins on glandular hair and trichome development in other seed plants is the same as Arabidopsis is not known. Trichomes and glandular hairs play an important role in defense against herbivores.

DELLAs restrain plant growth and development under adverse environmental conditions in extant flowering plants, and repress the accumulation of reactive oxygen species (Achard et al. 2008). Did these nuclear repressing proteins play a role in the diversification and adaptive radiation of flowering plants in fluctuating paleoclimates?

The ubiquitin 26S proteasome pathway is employed by animals and plants to control levels of- and degrade regulatory proteins such as gibberellin sensitive plant DELLAs and certain ethylene regulatory enzymes (Y. F. Chen et al. 2005).

Several other plant hormones including abscisic acid, auxin, cytokinins, and ethylene, interact with GA; some of these interactions are negative and reciprocal. For example, when rice is grown under hypoxic conditions, culm and internode elongation requires both ethylene and GA. KNOX1 proteins induce the production of cytokinins in SAMs that in turn inhibit the biosynthesis of GA (Weiss and Ori 2007).

Jasmonic acid (JA) and related esters belong to the oxylipin family of molecules (Wasternack 2007, Chico et al. 2008). Jasmonates are hormones implicated in both specific and conserved plant morphogenic effects (I. T. Baldwin 1998, Biondi et al. 2001, Balbi and Devoto 2008, Katsir et al. 2008). Specific effects include the ability of oxylipins to reprogram expression of jasmonate-inducible genes (Pauwels et al. 2009). Jasmonates, biosynthesized as a response to mechanostimulation (wounding) of tissues by herbivores (I. T. Baldwin 1998), affect development of stamens and pollen (Devoto and Turner 2003).

Nitric oxide (NO) is a gas with profound mediating effects on phytohormone cross-talk, development of flowers and pollen tubes, defense, and signaling (Perazzolli et al. 2006, Besson-Bard et al. 2008, Besson-Bard et al. 2008, Palavan-Unsal and Arisan 2009). Nitric oxide activates the enzyme protein kinase G (PKG) in mammals and both molecules are important to survival of insects in artificial hypoxic environments (Wingrove and O'Farrell 1999, Hoback and Stanley 2001).

Interestingly, NO also binds to hemoglobins, a class of oxygen-binding enzymes found in plants (Perazzolli et al. 2006), and vertebrates and invertebrates including certain species of the insect orders Coleoptera, Diptera, Hemiptera, Hymenoptera, and Lepidoptera (Van Holde et al. 2001, Burmester and Hankein 2007). Further, both NO and PKG enhance the survival of Drosophila larvae when deprived of oxygen in the laboratory (Wingrove and O'Farrell 1999).

Were Paleozoic insect antagonists attracted to potential emissions of nitric oxide of host plants where NO possibly played a role in ameliorating the detrimental effects of local hypoxia on egg and larval development?

Carbon dioxide is also a gaseous signaling molecule, which is perceived by the olfactory receptors of blood-sucking, meat-eating, and herbivorous insects. At certain concentrations carbon dioxide (CO₂) becomes an odor to Manduca moths (Goyret et al. 2008).

How did hypercapnic paleoenvironments following the end-Guadalupian extinction (GuCCE), end-Permian apocalypse (EPE), Triassic-Jurassic carbon cycle event (TrCCE), and end-Barremian biogeochemical perturbation (BaCCE) affect molecular evolution of insect olfactory receptor proteins and coevolution of stem group angiosperms and their pollinators?

Salicylic acid (SA) is a well-known natural signaling molecule that a play key role in activating genes for defense and integrating WRKY transcription factors during plant responses to herbivory and wounding (Wasternack 2007, Balbi and Devoto 2008, Blanco et al. 2009).

Signaling by herbivory and wounding. Molecular studies on the herbivorous sphingid moth, Manduca sexta by Ian Baldwin and students, demonstrate several molecular interactions between the insect and the wild tobacco plant host Nicotiana attenuata (Asteridae, Solanales) that potentially affect the ecology and development of both organisms. The activities of the specialized insect herbivore induces "... a dramatic ethylene release, a jasmonate burst, and a suppression of nicotine accumulation ..." in simulations of herbivory on the host plant (abstract, Winz and I. T. Baldwin 2001). Further, plant mRNAs accumulate "... in response to insect-derived cues" (title, Schittko et al. 2001).

The mechanism of action of hormone signals in SAMs is an active area of research in plant physiology. Through a process termed signal transduction biochemical signals that originate in the insect exoskeleton manipulate and control the genetic machinery needed for orderly (or disorderly) growth and expansion of shoots and other organs of the plant (Pandey et al. 2008).

Gall-forming insects inject plant cells with modified transfer RNAs (tRNAs), which are in fact, cytokinins. Injected cytokinins biosynthesized by phloem-feeding insects cause rapid differentiation and disorderly growth in plant tissues resulting in gall formation (Schultz 2002).

Do hormones secreted by parasitic invertebrates or insect instars and larvae upregulate or repress homeotic genes in host seed plant SAMs? Small CSE polypeptides biosynthesized in the esophageal glands of parasitic nematodes when injected through feeding stylet into plant cells, affect development of floral meristems (Germain et al. 2006). Interestingly, the elusive flowering hormone florigen is a small protein that is transported through the phloem of angiosperms (Shalit et al. 2009).

Plant light receptors. Phytochrome, an enzymatic light receptor implicated in several physiological processes in plants, including flowering, affects regulation of MIKC-type MADS-box genes. Issue Number 12 of Volume 58 (2007) of the Journal of Experimental Botany is devoted to photomorphogenesis.

Insect hexamerins: molecular evolution of hemocyanins. As the Neoproterozoic snowball Earth slowly warmed during the Cambrian and Ordovician periods global oxygen concentration increased. Photosynthesis in the chloroplasts of aquatic algae, protozooans, and land plants was the source of oxygen gas (Berner 1999, Berner and Kothavala 2001).

Hemocyanins (Hc) of invertebrate circulatory systems evolved during the early Paleozoic as a class of gas-binding, copper-containing enzymes needed in respiration (Burmester and Hankein 2007, Pick et al. 2009). Iron-containing hemoglobins (Hg) of animals also evolved and diversified during this time (Burmester et al. 1998, Van Holde et al. 2001, Burmester 2004).

Molecular evolution of copperless insect storage hexamerins (Hx) from the respiratory hemocyanin class (Hc) of invertebrate copper-containing enzymes may be linked with the late Devonian/early Carboniferous ice-house cooling of the Earth, accompanied by a drop in oxygen levels to 13% (Berner and Kothavala 2001, Ward et al. 2006). Protein storage macromolecules of the hemocyanin family of enzymes are termed hexamerins (Burmester 2001, Moreira et al. 2004).

Thorsten Burmester and coworkers (1998) describe the molecular evolution of hexamerin hemolymph protein subunits from the hemocyanin family of enzymes in two important reviews (Burmester 2001, Burmester et al. 2006). Hexamerins are moulting storage proteins composed of several polypeptide chains (multimers). Each enzyme subunit has a molecular weight of about 80,000.

Hemocyanins, including hexamerins are critical to developing and respiring entognathous hexapods and true (ectognathous) insects, being necessary for instar development and successive moults (Hagner-Holler et al. 2004, Pick et al. 2009). Larvae digest hexamerin storage proteins to release certain amino acids required for development (Moreira et al. 2004).

Moulting food storage proteins, homeotic selector genes including Hox genes, zygotic genes, field-specific selector genes, compartment selector genes, cell-type-specific selector genes, segment polarity genes, and controlling factors behind the cessation of insect growth (brain neurosecretory prothoracicotropic hormone [PTTH], juvenile hormone [JH], and ecdysone steroids) comprise key elements of the developmental tool kit of insects (Truman and Riddiford 2002, S. B. Carroll et al. 2005).

Trans-acting chromatin proteins and cis-regulatory transcription factors of eukaryotic genomes respond to oxygen levels in controlled environments. For example, lowering oxygen concentration in controlled environments affects levels of nonubiquitinated proteins and activates hypoxia-responsive genomic DNA elements (HREs) (Flück et al. 2007). Further, "hypoxia-regulated gene expression very likely reflects an ancient adaptive trait that was established in primitive eukaryotes in response to alterations in levels of oxygen" (page 529, Flück et al. 2007).

The below graphic is a crude plot of changing oxygen levels in Earth's atmosphere over geologic time based on GEOCARB III (Berner and Kothavala 2001), and other work (Beerling and Berner 2002, Huey and Ward 2005, Ward et al. 2006). I superimposed an evolutionary tree showing cladogenesis of insect hexamerins from invertebrate hemocyanins, which is based on the most recent molecular clock estimates by Hagner-Holler et al. (2007) and morphological data. In addition, I plotted three of the four insect herbivore expansion phases outlined by Labandeira (2006) on the graph.

The preceding plot is redrawn from Figure 6 on page 1072 of Hagner-Holler et al. (2007), Diversity of stonefly hexamerins and implication for the evolution of insect storage proteins, Insect Biochemistry and Molecular Biology 37: 1064-1074, with additional data added from GEOCARB III (Berner and Kothavala 2001), GEOCARBSULF (Berner 2006), and Labandeira (2006).

Three bursts of molecular evolution of insect hexamerins are apparent from the Hagner-Holler et al. data (2007). The first of the hexamerin molecular evolutionary radiations corresponds with the late Devonian/early Carboniferous hypoxic ice-house (DeCARB) Earth from 360 to 320 M.Y.A. (Hagner-Holler et al. 2007), and the initiation of Insect Herbivore Expansion Phase 2 (Labandeira 2006). During the hypoxic interval of the DeCARB seven clades of hexamerin storage proteins diverge from primitive Plecoptera (stoneflies) leading to many derived hemimetabolous and holometabolous insect orders (Hagner-Holler et al. 2007).

A second burst of insect hexamerin evolution occurs prior to the end-Guadalupian extinction and a decline in atmospheric oxygen levels from 31% to 12% (Berner and Kothavala 2001, Ward et al. 2006). During this interval, hexamerins diverged into clades leading to the Coleoptera (beetles and weevils), Hymenoptera (bees and wasps), and Diptera (flies), and certain Lepidoptera (butterflies and moths). This interval of hexamerin diversification corresponds with Insect Herbivore Expansion Phase 2 (Labandeira 2006).

The point of divergence between the hexamerin class belonging to beetles and those Hxs belonging to bees and wasps coincides with the second burst of hexamerin evolution.

Was the second explosion of molecular evolution of insect hexamerin protein subunits driven by developing hypoxia following the end-Guadalupian extinction (GuCCE) of the late Permian, or did diversification occur before the GuCCE as the data suggest?

Dipteran hexamerins diverged about the same time as oxygen levels in Earth's atmosphere declined to 12%. Interestingly, the third burst of hexamerin evolution leading to blattarians (cockroaches) and Isoptera (termites) occurs at a point in geologic time coincident with the Triassic-Jurassic carbon cycle event (TrCCE) (Beerling and Berner 2002).

Despite the rough estimate of molecular divergence times of insect hexamerins by Hagner-Holler et al. (2007) plotted against GEOCARB III, future phylogenetic inference could employ other cladistic algorithms and be fine-tuned with more data on insect hexamerins, possibly including a detailed study of hemoglobin evolution of beetles and wasps (Hgs).

A more detailed synthesis of invertebrate gene duplications and later molecular evolution of hexamerins and vitellogenins in populations of coevolving insect antagonists and Permo-Triassic seed plants is possible. Finely-tuned molecular phylogenetic and isotope studies are needed to implicate global hypoxia as an underlying selective force in the evolution of insect hexamerins. Insect hemoglobins (Hgs) remain relatively unstudied in this context.

Were Paleozoic changes in atmospheric oxygen concentration the selective force driving the molecular evolution of gas-binding hemocyanin respiratory enzymes and moulting storage proteins of Permian insects?

Are hexamerins hypoxia-inducible storage proteins of insects?

What was the role of nitric oxide (NO) in mediating developmental and physiologic effects of hypoxia on Paleozoic insects living with plants?

Did certain steroid hormones manufactured by Permo-Triassic seed plant host shrubs act like ecdysones while sending signals to the genetic machinery of developing eggs, instars of phytophagous insects inhabiting shrub lifeboats?

At the same time that hexamerin molecules were evolving in insect populations and colonies of shrub lifeboats, trampling and secreting insect antagonists might have sent signals to the genetic machinery of SAMs of host plants. Signals originating in the bodies of phytophagous insects had potential far-reaching effects on the biochemical machinery of differentiating cells in SAMs of certain developmentally plastic Permo-Triassic seed plants.

One example of a far-reaching effect might be signaling of the homeotic gene biosynthetic machinery in the nuclei of cells of bisexual cone axes of seed plants sensu Becker and Theißen (2003), Baum and Hileman (2006), and Theißen and Melzer (2007).

Insect sensing and hormone signaling. Arthropods manufacture several classes of hormones that regulate development and, together with anatomically modified appendages they facilitate communication among individuals, within castes of social insects, or possibly with plant hosts.

Some of the hormones produced by insects which are active in signaling and sensing include ecdysone steroids, epoxide methyl esters of sesquiterpenoids, peptides and polypeptides, and pheromones (Robertson and Wanner 2006).

Insect olfactory and gustatory chemoreceptors (Abdel-latief 2007) and innervated bristles are of potential interest with respect to plant to insect, insect to insect, and insect to plant communication by biomolecules and thigmo.

Interestingly, from the research perspective of possible neotenous coevolution of insect antagonists and seed plant hosts, Nicholas Strausfeld (2009) reviews the deeply conserved evo-devo of arthropod mushroom bodies in the biochemical and phylogenetic context. Mushroom bodies are paedomorphic neuropils of insects and firebrats that function in control of behavior, and in olfactory perception (Strausfeld et al. 1998).

Olfactory organs and chemoreceptors (Ray et al. 2008) allow certain insects to sense minute concentrations of pheromones, nitric oxide (NO), carbon dioxide, and other volatiles (VOCs) in the air. Beetles exhibit many adaptations to smell and taste and this is reflected in considerable anatomical, biochemical, and morphological diversification in sensillae of antennae, legs, mouthparts, and wings (Abdel-latief 2007).

The preceding paragraph is quoted from page 2635 of C. Hamès, D. Ptchelkine, C. Grimm, E. Thevenon, E. Moyroud, F. Gérard, J.-L. Martiel, R. Benlloch, F. Parcy, and C. W. Müller (2008), Structural basis for LEAFY floral switch function and similarity with helix-turn-helix proteins. The EMBO Journal 27: 2628-2637.

The focus of the exercise should now be further development of hypotheses on the origin of the angiosperms, the origin of the flower, and MIKC-type MADS-box and associated homeodomain protein evolution proposed by Theißen and Saedler (2001), Albert et al. (2002), Kaufmann et al. (2005), Maizel et al. (2005), Baum and Hileman (2006), and Theißen and Melzer (2007).

Further, phylogenetic analysis suggests that a critical MIKC-type MADS-box gene duplication or whole genome duplication (WGD) occurred in the common seed plant ancestor of angiosperms and gymnosperms at least 300 million years ago (Becker et al. 2000, Zahn et al. 2005).

Based upon a brief paleontologic survey of Paleozoic seed plants the prime gymnosperm candidate population for the 300-million-year-old gene duplication event was among the species and genera of Vojnovskyales, a Paleozoic gymnosperm with bisexual cone axes; and the oleanane-containing gigantopterids, and their potential intergeneric hybrids.

The leaves and fertile leaf fragments (microsporophylls and megasporophylls) of Paleozoic gigantopterids and Phasmatocycas might have been attached to bisexual cone axes and shoots (and the plants were probably shrub-like or vine-like), but published paleobotanical evidence is lacking.

Evolution in MADS-box proteins is reviewed by Rijpkema et al. (2007), Veron et al. (2007), and Leseberg et al. (2008), among others.

The next chapter of the essay considers evidence from evolutionary-developmental genetics and genomics needed to better understand the origin of the first flower and the origin of angiosperms.

Genetic Considerations:

In this chapter I discuss molecular evolution of transcription factors and cis-regulatory modules, developmental recombination, genetic accommodation, homeotic selector genes, horizontal transfer of genes (HGT), chloroplast capture, gene duplications, and duplication of whole genomes (WGDs) including polyploidy.

In addition, I summarize key research advances in our understanding of the possible effects of signals on whole genomes, homeotic selector genes, gene transcripts such as microRNAs (miRNAs and siRNAs), and cis-regulatory sequences that bind transcription factors.

The image on the left is a stained preparation of the chromosomes of Joinvillea plicata (Commeliniidae, Restionales, Joinvilleaceae), a species of grass-like monocots with plesiomorphic characters. The chromosomes in this image were fixed at the moment of metaphase I of meiosis in a microspore mother cell, × 2000. In the living Joinvillea plant, meiotic chromosomes would be identifiable in two places: inside the immature megasporangium of the ovule and within young microsporangia of very young stamens.

Could the application of mechanical force and secretion of phytohormone signals by chewing, crawling, feeding, ovipositing, and sucking insect mutualists on elongated (or shortened) cone axis SAMs affect developmental switches that trigger reproductive organ identity and formation?

"I argue that the origin of species differences can be explained, and the synthesis of Darwinism with genetics can be improved, by invoking two concepts: developmental recombination and genetic accommodation. Developmental recombination, or developmental reorganization of the ancestral phenotype, explains where new variants come from: they come from the preexisting phenotype, which is developmentally plastic and therefore subject to reorganization to produce novel variants when stimulated to do so by new inputs from the genome or the environment. Genetic accommodation, or genetic change in the regulation or form of a novel trait, is the process by which new developmental variants become established within populations and species because of genetic evolution by selection on phenotypic variation when it has a genetic component."

The preceding passage is quoted from page 6544 of M. J. West-Eberhard (2005), Developmental plasticity and the origin of species differences, Proceedings of the National Academy of Sciences 102(Supplement 1): 6543-6549.

These and many other questions are in need of answers:

What is the role of developmental recombination and genetic accommodation as underlying factors that lead to co-option, modularity, dissociation, and diverted development of seed plant characters involved in angiospermization?

Were the genes, morphogens, and regulatory factors having to do with mechanoperception and shoot development linked or the same in extinct Permo-Triassic seed plants?

Are phytophagous insects and their secretions a source of signals perceived by the developing plant SAM (including accessory reproductive meristems) and sent to nuclei of cells where they upregulate (or repress) floral integrator and meristem identity genes?

Do certain insect nuclear receptor proteins and HSLs bind to plant GA?

Are microRNAs passed on to male and female reproductive modules (pollen and ovules)?

Can we identify potentially interfertile sources of reproductive modules, and model the flow of pollen and rain of seeds in pre-angiospermous seed plant populations following the end-Permian extinction (EPE) and Triassic-Jurassic boundary carbon cycle event (TrCCE)?

What were the effects of epigenesis, including input of foreign genetic material, on developmental plasticity and phenotypic variation in populations of Permo-Triassic seed plants and the first angiosperms?

How has the horizontal transfer of genes in paleopopulations affected the origin and evolution of flowering plants?

Do phytophagous insects send biochemical signals via the host plant cytoskeleton and microtubules to the nucleus thereby upregulating SAM maintenance- and floral integrator and meristem identity genes?

Homeotic selector genes. Homeotic selector genes are known from metazoans, fungi, and plants. These genes are key regulators in insect and land plant developmental tool kits being important in plant organ or body segment identity (S. B. Carroll et al. 2005, Floyd and Bowman 2007). Homeotic selector genes of insects include Hox genes, zygotic (gap- maternal-, and pair-rule-) genes, field-specific selector genes, compartment selector genes, cell-type-specific selector genes, and segment polarity genes (S. B. Carroll et al. 2005).

Homeotic selector genes of plants include shoot apical meristem (SAM) maintenance genes and floral meristem organ identity and integrator genes that encode MIKC-type MADS-box transcription factors.

Homeotic selector genes of plants and the developmental proteins they determine have been the subject of active research by biochemists, plant cell biologists, and molecular systematists for nearly two decades (Coen and Meyerowitz 1991, Theißen et al. 1996, Kramer et al. 1998, Becker et al. 2000, Cubas et al. 2001, Stuurman et al. 2002, Becker and Theißen 2003, Irish 2003, Litt and Irish 2003, Tucker 2003, Yamada et al. 2003, Kaufmann et al. 2009,, and Melzer et al. 2009).

More recent studies having to do with the origin of the flower, perianth parts, bisexual cone axes, ovules and carpels include Buzgo et al. (2004), P. Zhang et al. (2004), Adam et al. (2005), Buzgo et al. (2005), Irish and Litt (2005), Kaufmann et al. (2005), S. Kim et al. (2005), S. Kim et al. (2005), Zahn et al. (2005), Baum and Hileman (2006), Cronk (2006), De Bodt et al. (2006), Doerner (2006), Endress (2006), Feng et al. (2006), Irish (2006), S. Kim et al. (2006), Jack (2004), S. Kim et al. (2004), Scutt et al. (2006), P. S. Soltis et al. (2006), Kramer et al. (2007), Rijpkema et al. (2007), and Theißen and Melzer (2007).

Developmental cis-regulatory modules. Clusters of DNA and nucleoprotein spaced at irregular intervals on the chromosomes of diverse eukaryotes sometimes contain promoter cis-regulatory sequences that bind transcription factors. Promoters are cis-regulatory sequences that affect transcription. Some promoters work together with additional factors termed enhancers. Transcription factors are proteins that bind to cis-regulatory sequences (Wray et al. 2003).

The clustered transcriptional-regulating protein-DNA binding sites and the structural genes being regulated may be viewed as cis-regulatory modules (S. B. Carroll 2005). De Bodt et al. (page 1293, 2006) define these modules as clusters of cis-regulatory binding sites that generate "discreet aspects of the total transcription profile."

At the level of populations and species, transfer of genetic identity and function from ancestral structural modules to derived ones is termed homeoheterotopy (Baum and Donoghue 2002).

It is increasingly clear that cis-regulatory MADS-box domain (see section below) transcription factors interact even with their own promoters. Further, "MADS-box proteins seem to function mostly as subunits of larger protein complexes" (page 34, Rijpkema et al. 2007).

Understanding the selective forces behind molecular coevolution of insect and seed plant developmental cis-acting regulatory modules in deep-time i.e. hundreds of millions of years, is an important first-step in unraveling the enigmatic origin of angiosperms, evolutionary development of the first flower, and adaptive radiation of certain clades of coevolving Holometabola.

Evolution of cis-regulatory modules in relation to animal and plant body plan patterning, development of organs of insects and seed plants, and natural selection is reviewed by Rodríques-Trelles et al. (2003), Wray (2003), Wray et al. (2003), Theißen and Becker (2004), S. B. Carroll (2005), De Bodt et al. (2006), Wittkopp (2006), Hoekstra and Coyne (2007), Wray (2007), and S. B. Carroll (2008).

The diagram below, which is redrawn from Carroll (2005), shows how the cis-regulatory molecular evolution and duplication of genes in insects or plants may potentially increase gene function while minimizing pleiotropy. The blue lines are segments of coding and non-coding DNA. Intron pairs are gray-colored rectangles. Orange circles are cis-regulatory promoters. Finally, asterisks denote mutations.

The first line of "A" illustrates a progenitor gene with its introns and promoter. The second two lines of "A" shows a gene duplication. Potential mutations of the cis-regulatory promoter and gene intron are denoted by asterisks. "B" illustrates expansion of cis-regulatory elements. "C" depicts the genesis of a novel intron and potential splicing sites that may lead to biosynthesis of alternative protein forms without the deleterious effects of mutations.

The above graphic is redrawn from Figure 1 of Sean B. Carroll, 2005, Evolution at two levels: on genes and form, PLoS Biology 3(7): 1159-1166.

A typical promoter consists of fewer than 50 binding sites for less than 15 different transcription factors. Transcription factor binding sites are about a dozen nucleotide base pairs wide, spaced irregularly along a segment of chromatin. Some promoters are highly conserved while others are not.

Shoot apical meristem (SAM) maintenance genes. Function and patterning in the shoot apical meristem (SAM) requires several gene families and their cis-acting transcription factors. The main shoot apical meristem identity genes are Class III HD-Zip, KANADI, CLASS 1 KNOX (KNOTTED1-like), and ARP genes (Ori et al. 2000, Hake et al. 2004, Di Giacomo et al. 2008).

Extant vascular plants including the lycophyte Selaginella kraussiana contain KNOX/ARP gene modules. The SAMs of extant flowering plants typically express KNOX but not ARP genes. Conversely expression of ARP genes in leaf primordium cells of the SAM leads to suppression of KNOX function. Whether the KNOX genetic switch is on or off depends on the activity of ARP repressors, and may lead to the development of different leaf forms e.g. pinnate, palmate, or compound (Beerling and Fleming 2007).

Meristem maintenance- and organ patterning gene families are deeply conserved in land plants (Floyd and Bowman 2007). Cytokinins and gibberellins (see previous chapter) act downstream of Class I KNOX genes in the rice and mouse-eared cress subjects studied (Di Giacomo et al. 2008). Class II TCP genes encoding TCP proteins are involved in the shaping of leaves and flowers (Cubas 2004, Damerval et al. 2007).

Further, Class I KNOX genes are important when considering the genetic basis of land plants and the origin of angiosperms as they figure prominently in discussions of classic theories on telomb transformational series within the evo-devo context (Beerling and Fleming 2007, Sanders et al. 2007).

Floral meristem organ identity and integrator genes and transcription factors. The development of perianth parts of the flower involves cis-acting transcription factors (Glover 2007). Pamela S. Soltis et al. (2009) provide an excellent overview of floral developmental genetics and evolution.

The main floral meristem organ identity and integrator genes in Arabidopsis include LEAFY (LFY), AGAMOUS (AG), APETALA 1 (AP1), APETALA 2 (AP2), CAULIFLOWER (CAL), SUPRESSOR OF OVEREXPRESSION CONSTANS 1 (SOC1), TERMINAL FLOWER 1 (TFL1), and UNUSUAL FLORAL ORGANS (UFO)..

Several transcription factors are implicated in changing the shape and size of flowers (Glover 2007). These include AINTEGUMENTA (ANT), ARGOS, BIG BROTHER (BB), CYC, DICH, and DIVARICATA (Glover 2007). AINTEGUMENTA (ANT) is a principal controller of both flower- and leaf size. Further ANT negatively regulates the expression of the Class C MIKC-type MADS-box gene AGAMOUS (AG) (see section below). Interestingly, ANT is a component of the auxin-regulated signaling cascade (Busov et al. 2008) (see previous chapter).

AINTEGUMENTA and APETALA2 (and a number of other genes known from Arabidopsis) comprise the AP2 subfamily, which is part the larger AP2/ERF (ethylene response factor) family of transcription factors (S. Kim et al. 2006). All of these genes including their paralogs, orthologues, and cis-acting transcription factors, figure prominently in discussions of the ancestral developmental tool kit of land plants (Floyd and Bowman 2007), the origin of the angiosperm flower, and floral diversification (S. Kim et al. 2006, P. S. Soltis et al. 2009).

LEAFY (LFY) protein encoded by the LFY gene is a key developmental switch in floral meristems of many angiosperms including Arabidopsis (Glover 2007, Hamès et al. 2008). A gymnosperm ortholog of LFY termed NEEDLY (NLY) has been isolated from Pinus (Mouradov et al. 1998). The LFY family of transcription factors controls the downstream expression of several genes including MIKC-type 2 MADS-box genes (Theißen and Melzer 2007).

In 2006 Baum and Hileman proposed a modification of Theißen and Saedler's (2001) protein quartet model involving genetic interactions of B and C homeotic selector genes, SEP genes, and small-sized growth promoting molecules (phytohormones and other factors) aligned along gradients within a developing seed plant cone axis.

The protein quartet model of floral molecular evolution is now widely accepted among plant biologists (Leseberg et al. 2008). Plant systematists view the expanded ABC model as "the default program" behind the identity of organs of the flower (page 18, D. E. Soltis et al. 2008; page 117, Figure 2, P. S. Soltis et al. 2009). In fact, E-class MIKC-type MADS-box proteins such as SEPALLATA3 bind to DNA in the test tube as quartet-like complexes (Melzer et al. 2009).

Theißen and Melzer's (2007) proposed "Out-of-Male" and "Out-of-Female" hypotheses on the origin of the angiosperm flower, and the molecular evolution of MIKC-type MADS-box protein quartets (Theißen 2001) are examples of how cis-acting protein modules "might exert their function as transcription factors by binding to the promoters of target genes" (page 132, Theißen and Becker 2004).

While discussing evidence that seed plant class B homeotic selector genes diverged into the two lineages leading to extant gymnosperms and angiosperms 300 million years ago during the late Carboniferous Period following a gene- or whole-genome duplication, Theißen and Becker (2004) state:

"Changes in cis-regulatory elements, typically located in the promoter region of a respective gene, in enhancers or silencers up- or downstream of the coding region, or even in introns, might have rendered B gene expression more susceptible to a hypothetical apical-basal gradient already present within the cone. Such a gradient could be based on a low molecular weight compound such as a phytohormone or on a protein (presumable another transcription factor). The underlying molecular changes within the B gene could range from a simple single nucleotide change to major rearrangements at the B gene locus."

The preceding statement is quoted from page 142 of Günter Theißen and Annette Becker (2004), Gymnosperm orthologues of class B floral homeotic genes and their impact on understanding flower origin, Critical Reviews in Plant Sciences 23(2): 129-148.

Is Theißen and Becker's hypothetical cone concept (2004) applicable to bisexual axes of Carboniferous Vojnovskyales?

MIKC-type MADS-box genes and transcription factors. Certain MADS-box floral meristem organ identity and integrator genes encode MIKC-type MADS-box transcription factors that determine development of cones and flowers (Ma and dePamphilis 2000, Theißen et al. 2000, Theißen 2001, Becker and Theißen 2003, Theißen and Melzer 2007).

The acronym "MADS" was coined from names of homeotic genes isolated from yeast (MCM1), snapdragon (AG and DEFICIENS), and humans (SRF) (Ma and dePamphilis 2000). Protein subunits encoded by the MIKC-type MADS-box genes possess a modular organization that includes a MADS (M) intervening (I), keratin-like (K), and a C-terminal (C) domain (Theißen et al. 2000).

The extreme age of MADS-box genes as controllers of development is supported by phylogenetic analyses (Nam et al. 2003). Gymnosperms studied thus far, also possess orthologs of certain MADS-box genes. It is not unreasonable to accept the proposal that at least some MADS-box genes operated in more general developmental patterning in Carboniferous seed plants before the angiosperm-gymnosperm split 300 million years ago (Nam et al. 2003).

Further, based on studies of extant Chara (an algal charophyte), Physcomitrella (a moss), and Selaginella (a lycophyte), early land plants probably had two MIKC-type MADS-box genes and that the MADS-box gene family diverged widely after the seed plant/fern split (Floyd and Bowman 2007).

A compilation of review papers on genetic regulation of floral development is published as Volume 44 of Advances in Botanical Research, Developmental Genetics of the Flower (Soltis et al., eds. 2006). Glover (2007), Kramer (2007), and P. S. Soltis et al. (2009) review advances in our understanding of the genetics and regulation of flowering since the 2006 compilation by the Douglas Soltis lab.

There are three fundamental classes of floral organ identity genes, "A," "B," and "C" that comprise the first unifying principle of fertile meristem development (Jack 2004). In extant flowering plants "A" function is responsible for sepal identity (M.-K Chen et al. 2008, among others), "A" + "B" expression determines the identity of petals (Su et al. 2008), "B" + "C" expression affects stamen morphology (Kramer et al. 1998, Piwarzyk et al. 2007), and "C" gene expression controls the identity of carpels. Variations in "B" gene expression determine the identity, shape, and form of unusual perianth parts in columbines (Kramer et al. 2007), grasses (Whipple et al. 2007), and papaya (Ackerman et al. 2008).

Gymnosperm orthologues of the B and C genes including AG are known: their conserved evolution is demonstrable (P. Zhang et al. 2004). Apparently gymnosperm "B" genes are only expressed in male cones (Theißen and Becker 2004).

Additional classes "D," "E," and "F" been identified in certain extant seed plants. "D"-class genes apparently operate to determine organ identity in the flowering plant genus Petunia (Solanaceae, Solanales, Asteridae) (Ferrario et al. 2006).

"... the most parsimonious explanation [i.e. differential expression of class B, C, and D genes as the 'primary sex-determination mechanism in all seed plants'] would be that the mechanism of female/male reproductive organ specification/distinction in angiosperms was recruited from a similar system that was already established in the lineage leading to all extant seed plants about 300 MYA (Theißen et al. 2000). At the molecular level, therefore, flower origin seems less mysterious than the morphological difference between flowers and gymnosperm cones or strobili might suggest."

The preceding statement is from page 138 of Günther Theißen and Annette Becker (2004), Gymnosperm orthologues of class B floral homeotic genes and their impact on understanding flower origin, Critical Reviews in Plant Sciences 23(2): 129-148. The phrase and within-context but out of place quote from the original article denoted by brackets [] is mine.

Further, I propose that insect-mediated natural hybridization between gigantopterids and certain Vojnovskyales was the method through which MIKC-type MADS-box genetic novelties were recruited into populations of ancestral angiosperms during the Permo-Carboniferous approximately 300 million years ago prior to both the end-Guadalupian and end-Permian mass extinctions.

In the derived eudicot angiosperm species Arabidopsis thaliana (Brassicaceae, Capparales, Dilleniidae), commonly known as mouse-eared cress, a favorite research subject of plant developmental biologists, there are apparently two of A-class floral meristem identity genes, (AP1 and AP2). The B-class genes of Arabidopsis, including AP3 and PISTILLATA are specifiers of petaloid organs in whorl 2, and staminate organs in whorl 3 of the SAM (Hileman and Irish 2009). The C-class gene AGAMOUS integrates the activity of whorl 3 stamens and carpels of the fourth whorl. C-class genes evidently repress the activity of A-class genes that operate in whorls 3 and 4 (Jack 2004).

A fourth class of MIKC-type MADS-box genes known as SEPALLATA is implicated in determination of floral identity (Becker and Theißen 2003, Kaufmann et al. 2009, Melzer et al. 2009). Zahn et al. (2005) report SEP homologues in basal angiosperms including Amborella trichopoda.

"The coincidence between the origin and diversification of the SEP subfamily and the emergence of angiosperms suggests that the origin of the SEP subfamily may be part of a hypothesized 'molecular innovation' that made possible the morphological invention of the flower."

The previous quotation is from page 2219 of L. M. Zahn, H. Kong, J. H. Leebens-Mack, S. Kim, P. S. Soltis, L. L. Landherr, D. E. Soltis, C. W. dePamphilis, and H. Ma (2005), The evolution of the SEPALLATA subfamily of MIKC type MADS-box genes: a preangiosperm origin with multiple duplications throughout angiosperm history, Genetics 169: 2209-2223.

By knowing the interval in geologic deep time when molecular systematists predict divergence of reproductive patterning genes and regulatory proteins, paleobotanists might be able to better focus on candidate seed plant groups for detailed anatomical study of permineralized fossils (Crepet 2008). Pamela S. Soltis et al. (2009) offer a nice summary of the role of gene duplications in the diversification of angiosperm seed plant flowers.

Phylogenetic analyses suggest that duplications in floral meristem identity genes took place prior to the origin of angiosperms more than 260 million years ago (Bowe et al. 2000, Bowers et al. 2003, S. Kim et al. 2004). Further, mathematical modeling studies of Arabidopsis, the highly researched extant model eudicot plant, reveal that diversification of genes involved in signal transduction, SAM development, and the regulation of transcription are attributable to three WGDs during the last 350 million years (Maere et al. 2005).

Paleopolyploidy. A great body of data has been assembled on paleopolyploidy in plants (Bennett and Leitch 2005, J. J. Doyle et al. 2008). Evidence from an exhaustive genomic study of the cultivated grape overwhelmingly supports the existence of paleohexaploidy (Jaillon et al. 2007).

Nancy Eckardt (2004) in one of many journal article summaries for The Plant Cell reports on a paper by Blanc and Wolfe (2004) that estimates the timing of gene duplications in several monocots, eudicots, and rosids over the course of the last few million years as demonstrated by mathematical modeling studies (Blanc and Wolfe 2004), and later refinements of the analysis (Cui et al. 2006). Detailed phylogenetic analyses of this sort permit accurate estimates of the divergence times of major eudicot clades (D. E. Soltis et al. 2008), with precision enhanced by including fossil minimum age data mapped on to branch points (Crepet et al. 2004).

Based upon studies of gene and genome duplications in rice (Oryza sativa, Poaceae, Poales, Commelinidae) and mouse-eared cress, Chapman et al. (2006) provide insight on the functional divergence and buffering models for genome duplication. Interestingly, Chapman et al. (page 2734, 2006) suggest that as ancient duplicated genes diverge "... reciprocal buffering presumably erodes with time ..." leading to cyclic genome duplication.

De Bodt et al. (2005) propose that paleopolyploidy created much of the genetic material in modern flowering plants. For example, genes retained in duplicate from WGDs in Arabidopsis over the course of million of years are involved in developmental gene regulation and often do not retain redundant functions (Blanc and Wolfe 2004). In fact, gene duplications result in either neofunctionalization, loss of function thus forming pseudogenes, or provide raw material for divergence in animals (Carroll et al. 2005) and plants (Irish and Litt 2005).

"Given that such genes [developmental, regulatory, and signaling genes] are considered important for introducing phenotypic variation and increase in biological complexity, linking ancient polyploidy events with decisive moments in evolution becomes less speculative and the origin and evolution of angiosperms perhaps less of a mystery."

I reject the assertion by Wing and Boucher (page 380, 1998) that angiosperms are "not in a genealogical sense one of the major branches of land plants and did not originate with other major land plant clades." Feild and Arens (2007) are also incorrect in stating that flowering plants originated in the Cretaceous Period.

The absence of paleobotanical data is not a substitute for fact when dealing with a probable ghost lineage due to insufficient sampling, especially in view of more than a dozen molecular phylogenetic studies pointing to ancient gene duplications (see D. E. Soltis et al. 2007), and geochemical evidence of a Paleozoic oleanane signal (D. W. Taylor et al. 2006).

Perplexingly, the few paleobotanists who support a Cretaceous origin of angiosperms neglect to discuss the potential importance of two widespread Paleozoic seed plant lineages belonging to Permian gigantopterids (X.-G. Li and Z.-Q. Yao 1983, Mamay et al. 1988, Mamay 1989) and Carboniferous and Permian Vojnovskyales (Mamay 1976, Rothwell et al. 1996, Naugolnykh 2001) where scorable morphological character data are published and readily available for phylogenetic analysis.

Evolution of the Hox gene cluster and MADS-box genes. Gene duplications were an important part of the assembly of the early bilaterian animal tool kit that led to tetraploidization and the creation of several gene families including the HOM-Hox and Parahox clusters of homeobox genes (Carroll et al. 2005). Homeobox genes are not the same as homeotic MADS-box genes (Meyerowitz 2002).

An ancestral MADS-box gene duplication took place before the divergence of animals, fungi, and plants hundreds of millions of years ago (Alvarez-Buylla et al. 2000), probably before or during the Neoproterozoic snowball Earth or later warming leading to the Cambrian explosion of life. This earliest gene duplication gave rise to the Type 1 and Type II lineages of MADS-box genes (see Figure 1, page 468, Becker and Theißen 2003).

Animal genes such as MEF2 encode certain enzymes involved in muscle differentiation of both invertebrates and vertebrates and wing vein and tracheal development in certain insects. Duplicated MADS-box genes in plants have completely different functions (Meyerowitz 2002).

Some of the better understood duplicated MADS-box floral meristem organ identity and integrator genes encode MIKC-type MADS-box transcription factors that determine development of cones and flowers (Ma and dePamphilis 2000, Theißen et al. 2000, Theißen 2001, Becker and Theißen 2003, Theißen and Melzer 2007).

An evolutionary history of A-class MADS-box genes, including gene duplications and functional divergence is reviewed by S. Kim et al. (2006) and Shan et al. (2007). Recent functional analyses of A-class genes in monocots is published by M.-K. Chen et al. (2008).

Irish and Litt (2005) and Irish (2006) review the evolutionary history of B-, C-, and E-class MADS-box gene duplications in plants. The molecular evolution of B-class MADS-box genes has been extensively researched by Kramer et al. (1998), Kramer et al. (2003), Aoki et al. (2004), S. Kim et al. (2004), Zahn et al. (2005), and Hernández-Hernández et al. (2007).

Molecular evolution of C-class MADS-box genes over the course of 300 million years of seed plant evolution is reviewed by Kramer et al. (2004), P. Zhang et al. (2004), and Irish and Litt (2005).

Finally, the importance of MADS-box gene duplications and the interactions among the proteins they encode toward the evolution and radiation of flowering plants is reviewed by Rijpkema et al. (2007) and Veron et al. (2007).

MicroRNAs. MicroRNAs (miRNAs) are tiny ribonucleic acid fragments comprised of about a dozen nucleotides found in animal and plant cells. Their precursors possess a characteristic hairpin loop structure. RNAi interference probably existed in metazoan and land plant developmental systems as early as the Silurian Period (B. Zhang et al. 2006).

While many microRNA's are deeply conserved in land plants as determined from molecular phylogenetic analyses (S. Kim et al. 2006, Barakat et al. 2007), the birth and death of certain miRNAs may occur within a short evolutionary time span (Axtell et al. 2007).

Are plant microRNAs involved in repressing homeotic genes in growing cells and tissues? Yes, according to Axtell and Bartel (2005) and Axtell et al. (2007). Most microRNAs are involved in negative regulation of gene expression, specifically in control of transcriptional regulators that affect plant development and morphology (Axtell et al. 2007). They are powerful but negative repressors of key transcriptional regulators in the land plant developmental tool kit (B. Zhang et al. 2006, Floyd and Bowman 2007, Axtell and Bowman 2008).

Some key examples of miRNA regulation of transcription factor genes include repression of Class III HD-Zip genes necessary for leaf development by miR165 and miR166, targeting of APETALA2 floral development Class A genes by miR172 (S. Kim et al. 2006), repression of scarecrow-like (SCL) GRAS family genes by miR172, cleavage of GAMYB proteins (positive regulators of LEAFY gene expression) by miR159, loss-of-function of CUC genes (important in embryo and flower development) caused by miR164, and negative effects on ARF genes by miR160 and miR167 (B. Zhang et al. 2006).

Interestingly miRNAs, specifically small-interfering RNAs (siRNAs) are implicated in the regulation of plant defense responses to insect herbivore attack. These small RNAs may move through living cells and tissues via plasmadesmata from plant to plant, or from plant to insect and vice-versa (Pandey et al. 2008).

Chloroplast capture and horizontal transfer of genes. Phylogenetic systematists who study the origin and evolution of angiosperms have identified genes that originated outside of the host plant (Bergthorsson et al. 2004). Foreign chloroplasts and mitochondria (including their genetic material) may be incorporated into new symbionts (seed plants) (Tsitrone et al. 2003, Cho et al. 2004).

Horizontal transfer of genes (HGT) occurs in both prokaryotes and eukaryotes including animals, fungi, and plants. According to a recent survey by Richardson and J. D. Palmer (2007), HGT involves housekeeping respiratory- or ribosomal protein encoding- mitochondrial genes.

Transfer of genes from donor epiphytic or saprophytic (or other) species to certain organelles of seed plant hosts is demonstrable. At least 26 mitochondrial genes in the basal angiosperm Amborella trichopoda originated from other plants (Bergthorsson et al. 2004) including nine genes from probably epiphytic mosses, and another 19 from sympatric eudicots (Richardson and J. D. Palmer 2007). This is in contrast to the chloroplast genome of Amborella where no horizontally transferred genes are apparent (Goremykin et al. 2003).

"The horizontal gene transfer concept is potentially of great importance for leading evolutionary thinking beyond the ossified tenets of 'synthetic' theory, first of all, by introducing a long sought - and vigorously denied by traditionalists - mechanism by which macromutations can spread ..."

The passage above is quoted from page 381 of Krassilov (2002), Chapter 29, Character parallelism and reticulation in the origin of angiosperms, Pp. 373-382 In: M. Syvanen and C. I. Kado (eds.), Horizontal Gene Transfer (second edition), San Diego: Academic Press, 445 pp.

The prevalence of HGT in seed plants and its phylogenetic significance in the evolution of angiosperms is a matter of continuing debate (Goremykin et al. 2009).