LIGNOPHYTA

True roots +; lateral meristems: cork cambium producing cork abaxially, vascular cambium producing phloem abaxially and xylem adaxially.

EXTANT SEED PLANTS/SPERMATOPHYTA

Plant woody, evergreen; nicotinic acid metabolised to trigonelline, (cyanogenesis via tyrosine pathway); primary cell walls rich in xyloglucans and/or glucomannans, 25-30% pectin [Type I walls]; lignins derived from (some) sinapyl and particularly coniferyl alcohols, thus containing p-hydroxyphenyl and guaiacyl lignin units, (lignins derived from p-coumaryl alcohol, i.e. S [syringyl] lignin units); true roots present, apex multicellular, xylem exarch, and branching endogenous; arbuscular mycorrhizae +; shoot apical meristem multicellular, interface specific plasmodesmatal network; stem with ectophloic eustele, endodermis 0, xylem endarch, branching exogenous; vascular tissue in t.s. discontinuous by interfascicular regions; vascular cambium + [xylem ("wood") differentiating internally, phloem externally]; wood homoxylous, tracheids and rays alone, tracheid/tracheid pits circular, bordered; mature sieve tube/cell lacking functioning nucleus, plastids with starch grains; phloem fibres +; stem cork cambium superficial, root cork cambium deep seated; leaves with single trace from sympodium ["nodes 1:1"]; stomata ?; leaf vascular bundles collateral; leaves megaphyllous [determinancy evolved first, then ad/abaxial symmetry], spiral, simple, lamina with vein density up to 5 mm/mm² [mean for all non-angiosperms 1.8]; axillary buds associated with at most some leaves; prophylls [including bracteoles] two, lateral; plant heterosporous, sporangia eusporangiate, on sporophylls, sporophylls aggregated in indeterminate cones/strobili; true pollen [microspores, i.e. no distal pore for release of gametes] +, grains mono[ana]sulcate, exine and intine homogeneous; ovules unitegmic, crassinucellate, megaspore tetrad tetrahedral, only one megaspore develops, megasporangium indehiscent; male gametophyte development first endo- then exosporic, tube developing from distal end of grain, to ca 2 mm from receptive surface to egg, gametes two, developing after pollination, with cell walls, with many flagellae; female gametophyte endosporic, initially syncytial, walls then surrounding individual nuclei; seeds "large", first cell wall of zygote transverse, embryo straight, endoscopic [suspensor +], short-minute, with morphological dormancy, white, cotyledons 2; plastid transmission maternal; two copies of LEAFY gene, PHY gene duplications [three - [BP [A/N + C/O]] - copies], nrDNA with 5.8S and 5S rDNA in separate clusters; mitochondrial nad1 intron 2 and coxIIi3 intron and trans-spliced introns present.

MAGNOLIIDAE Takhtajan  Back to Main Tree

Lignans, O-methyl flavonols, dihydroflavonols, triterpenoid oleanane, non-hydrolysable tannins, quercetin and/or kaempferol +, apigenin and/or luteolin scattered, [cyanogenesis in ANITA grade?], S [syringyl] lignin units common, positive Maüle reaction [syringyl:guaiacyl ratio more than 2-2.5:1], and hemicelluloses as xyloglucans; root apical meristem intermediate-open; root vascular tissue oligarch [di- to pentarch], lateral roots arise opposite or immediately to the side of [when diarch] xylem poles; origin of epidermis with no clear pattern [probably from inner layer of root cap], trichoblasts [differentiated root hair-forming cells] 0; shoot apex with tunica-corpus construction, tunica 2-layered; reaction wood ?, with gelatinous fibres; starch grains simple; primary cell wall mostly with pectic polysaccharides, poor in mannans; tracheid:tracheid [end wall] plates with scalariform pitting, wood parenchyma +; sieve tubes enucleate, sieve plate with pores (0.1-)0.5-10< µm across, cytoplasm with P-proteins, cytoplasm not occluding pores of sieve plate, companion cells from same mother cell that gave rise to the sieve tube; sugar transport in phloem passive; nodes unilacunar [1:?]; stomata with ends of guard cells level with pore, paracytic, outer stomatal ledges producing vestibule; leaves petiolate, lamina [formed from the primordial leaf apex], development of venation acropetal, 2ndary veins pinnate, fine venation reticulate, veins (1.7-)4.1(-5.7) mm/mm², endings free; most/all leaves with axillary buds; flowers perfect, pedicellate, polysymmetric, parts spiral [esp. the A], free, numbers unstable, development in general centripetal; P not sharply differentiated, with a single trace, outer members not enclosing the rest of the bud, often smaller than inner members; A many, filament not sharply distinguished from anther, stout, broad, with a single trace, anther introrse, tetrasporangiate, sporangia in two groups of two [dithecal], ± embedded in the filament, with at least outer secondary parietal cells dividing, each theca dehiscing longitudinally by action of hypodermal endothecium, endothecial cells elongated at right angles to long axis of anther; tapetum glandular, binucleate; microspore mother cells in a block, microsporogenesis successive, walls developing by centripetal furrowing; pollen subspherical, tectum continuous or microperforate, ektexine columellar, endexine thin, compact, lamellate only in the apertural regions; nectary 0; G free, several, ascidiate, with postgenital occlusion by secretion, stylulus short, hollow, cavity not lined by distinct epidermal layer, stigma ± decurrent, dry [not secretory]; ovules few [?1]/carpel, marginal, anatropous, bitegmic, micropyle endostomal, outer integument 2-3 cells across, often largely subdermal in origin, inner integument 2-3 cells across, often dermal in origin, parietal tissue 1-3 cells across [crassinucellate], nucellar cap?; megasporocyte single, hypodermal, megaspore tetrad linear, functional megaspore chalazal, lacking sporopollenin and cuticle; female gametophyte four-celled [one module, nucleus of egg cell sister to one of the polar nuclei]; P deciduous in fruit; seed exotestal; pollen binucleate at dispersal, trinucleate eventually, germinating in less than 3 hours, pollination siphonogamous, tube elongated, growing at 80-600 µm/hour, with pectic outer wall, callose inner wall and callose plugs, growing between cells, penetration of ovules via micropyle [porogamous] within ca 18 hours, distance to first ovule 1.1.-2.1 mm, tube moves between nucellar cells; double fertilisation +, endosperm diploid, cellular [micropylar and chalazal domains develop diffently, first division oblique, micropylar end initially with a single large cell, divisions uniseriate, chalazal cell smaller, divisions in several planes], copious, oily and/or proteinaceous, embryo cellular ab initio, minute; germination hypogeal, seedlings/young plants sympodial; Arabidopsis-type telomeres [(TTTAGGG)_n]; whole genome duplication, ndhB gene 21 codons enlarged at the 5' end, single copy of LEAFY and RPB2 gene, knox genes extensively duplicated [A1-A4], AP1/FUL gene, paleo AP3 and PI genes [paralogous B-class genes] +, with "DEAER" motif, SEP3/LOFSEP and three copies of the PHY gene, [PHYB [PHYA + PHYC]].

Evolution. See Below.

Possible apomorphies for flowering plants are in bold. Note that the actual level to which many of these features, particularly the more cryptic ones, should be assigned is unclear. This is because some taxa basal to the [magnoliid + monocot + eudicot] group have been surprisingly little studied, there is considerable homoplasy as well as variation within and between families of the ANITA grade in particular for several of these characters, and also because details of relationships among gymnosperms will affect the level at which some of these characters are pegged. For example, if reticulate-perforate pollen is optimized to the next node on the tree (see Friis et al. 2009 for a discussion), it effectively makes the pollen morphology of the common ancestor of all angiosperms ambiguous... For other features such a nucellus only one (Nymphaeales) to three cells thick above the embryo sac and a stylar canal lacking an epidermal layer, although these are plesiomorphous for basal grade angiosperms (Williams 2009), where on the tree a thicker nucellus and a stylar epidermal layer are acquired has not yet been indicated.

Chemistry, Morphology, etc. For angiosperm and perhaps related gymnosperm fossils, see Krassilov (1997) and Taylor et al. (2009). Root morphology, cork development, etc., are unknown in Amborellaceae, an absence of knowledge complicated by the very distinctive "aquatic" morphology and anatomy of the next clade up, Nymphaeales. As to other characters of possible phylogenetic interest, the triperpenoid oleanane is widely distributed in angiosperms, but in no other extant seed plants (Taylor et al. 2006). An arsenite transporter controlled by the single-copy ACR3 gene and promoting arsenic tolerance is found throughout land plants other than flowering plants, as found when searching whole-genome and and EST databases (Indriolo et al. 2010), so another apomorphy.... The mycorrhizal condition of the ANITA grade (Amborellales, Nymphaeales and Austrobaileyales here) is largely unknown, as is that of Canellales, Piperales, and most of Laurales, although mycorrhizae are absent in Nymphaeales (and Ceratophyllales), as might be expected for aquatic groups (Landis et al. 2002; Wang & Qiu 2006). Magnoliophyta as a whole are likely to have vesicular-arbuscular mycorrhizae, indeed, this may be a much more "basal" character, and it is placed here as a commonality of all extant seed plants. Boyce at al. (2004) discuss interesting variation in the lignification of the primary cell wall; it is slight in "basal" angiosperms (and also in Drimys, a magnoliid), less in eudicots, and this may have functional implications. However, the sampling is very preliminary. p-hydroxybenzaldehyde, a component of many lignins, is apparently absent from broad-leaved angiosperms - at least from magnoliids and eudicots (Towers & Gibbs 1953), but is present in the monocots sampled and in some living gymnosperms (and also some Myrtaceae, etc.). Axial parenchyma is notably slight to absent in Amborellales, Austrobaileyales, Laurales, and Chloranthales (Herendeen et al. 1999). Leaf traces make connections only with xylem produced during the first year (Tomlinson et al. 2006); cf. Pinales. Stomatal morphology in many members of the ANITA grade is notably variable (Upchurch 1984). Leaf teeth of the chloranthoid type, with a central vein joined by branches from above and below and then proceeding to a thickened apex, may be plesiomorphic within angiosperms and synapomorphic for them (Doyle 2007).

For mannans, etc., see Popper and Fry (2004), note that Austrobaileya has mannans, although two other members of the order sampled lack them, Nymphaea also has mannans, but other Nymphaeales and Amborella were not sampled. Pribat et al. (2010) discuss the distribution of a folate-dependent phenyanaline hydrolase, which may be correlated with major plant groups when sampling is improved. For floral morphology and evolution, see Specht and Bartlett (2009 and references). Endress (2001a) describes a possible plesiomorphic flower and fruit morphology for angiosperms in some detail; he notes i.a. that the carpels may have uniseriate hairs - in Trimeniaceae and Nymphaeales the apical cell of these hairs is elongated and tanniniferous. The perianth members may each have a single trace, but I haven't followed this character in the basal pectinations carefully enough. For pollen characters of angiosperms of the ANITA grade, and also magnoliids, see Doyle (2007); note that pollen morphology of Amborellaceae is still not well understood. Indeed, some of the pollen characters may be incorrectly placed on the tree. Routley et al. (2004) find protogyny to be very common in "basal" angiosperms (Amborellaceae inapplicable, variable in eudicots and above Alismatales in monocots - see also Endress 1994b). In Amborellaceae and some other ANITA-grade angiosperms - note, including Hydatellaceae - the stigma has multicellular papillae. Whether or not an 8-nucleate embryo sac and triploid endosperm are synapomorphies for all angiosperms or only for those angiosperms in the [magnoliid + monocot + eudicot + Chloranthaceae] group has been unclear (Friedman 2001a, b, 2006; Baroux et al. 2002), although Friedman et al. (2003a, esp. b) and Friedman and Williams (2003, 2003) incline towards the latter hypothesis - see especially Friedman and Ryerson (2009). Similarly, there is much variation in microsporogenesis and pollen morphology in Nymphaeales, Amborellales, etc. (e.g. Furness et al. 2002). Sage et al. (2009) suggest that the basic condition of the angiosperm stigma may have been dry, while Staedler et al. (2009) note that the presence of an extragynoecial compitum could be an apomorphy of angiosperms - and then it would have to be lost at least twice (see also Williams 2009 for pollen tube growth, etc.). Lauraceae, Degeneriaceae and Magnoliaceae, at least, develop a massive, multiseriate suspensor during embryogenesis (Wardlaw 1955). Seeds of many angiosperms are notably smaller than those of extant gymnosperms (Moles et al. 2005a). For suggested patterns of evolution in endosperm development, see Floyd and Friedman (2000 [comprehensive treatment], 2001). Seedlings/young plants with decumbent lignotubers and sympodial growth are common in the ANITA grade and in Chloranthaceae, although they are not known in Nymphaeaceae, and early angiosperms may have been smallish trees (Feild et al. 2003, 2004; Feild & Arens 2005). Most eudicots have seedlings/young plants that are at least initially erect. Genome size in many angiosperms is small, less than 1.4 picograms in size, although there are some notable exceptions (e.g. a few Liliales and Asparagales); it is smaller than that of extant gymnosperms, although genomes in Gnetales are smaller than those of the others (Leitch et al. 2005); 1-1.4 picograms is the estimated ancestral genome size for angiosperms (Masterson 1994; Leitch et al. 2005). Note that at this level genome size and basic degree of ploidy seem not to be connected; for genome duplication in stem-group angiosperms, see e.g. Karlgren et al. (2011). Evans and Rees (1971) discuss variation in the length of the mitotic cycle, with that in eudicots being ca 4 hours longer than that in monocots (interphase, G1, is involved, 16 species sampled). For B-function genes, etc., see S. Kim et al. (2004b) - synonymization: AP3 and PI with DEF and GLO.

Phylogeny. On both the Cycadales (see under seed plant evolution) and especially the Students (see seed plants again!) pages there is further discussion about relationships between the major clades of seed plants; for some characters, whether or not they are apomorphies of angiosperms will depend on relationships between extant gymnosperms. The balance of evidence seems to be tilting towards the hypothesis that extant gymnosperms are sister to angiosperms, however, given the uncertainty in our knowledge of the relationships between the five major seed-plant clades, direct links are provided to the four other clades from here: Cycadales, Ginkgoales, Gnetales, and Pinales.

Relationships between members of the basal angiosperm pectinations are being clarified. Donoghue and Mathews (1998) listed 16 different hypotheses of relationships among basal angiosperms that involved the the first three nodes, but it seems that Amborellaceae are most likely to be sister to other angiosperms (not an hypothesis that Donoghue and Mathews included!), Nymphaeaceae sister to the rest, then Austrobaileyales - the ANITA grade (e.g. Mathews & Donoghue 1999; Qiu et al. 1999, 2000, 2005, 2006a, 2006b [some analyses], 2007; P. Soltis et al. 1999, 2000; Parkinson et al. 1999; Zanis et al. 2002; Magallón & Sanderson 2002; Kim et al. 2003; Borsch et al. 2003, 2005; Hilu et al. 2003; Nickerson & Drouin 2004; Aoki et al. 2004; P. Soltis & D. Soltis 2004; Müller et al. 2006a; Hansen et al. 2007; Duarte et al. 2008; McCoy et al. 2008). This is the topology followed here. There are also suggestions that the basal clade consists of Amborellales + Nymphaeales, perhaps favoured by analyses of mitochondrial genes (Qiu et al. 2006b: however, "it is sufficiently clear that the first diverging lineage of extant angiosperms consists of Amborella + Nymphaeales" [p. 845] seems something of an overstatement, and note that there are several unexpected if poorly supported relationships elsewhere in their prefered tree; Qiu et al. 2010) - see also Qiu et al. (2000, 2005, 2006a), Mathews & Donoghue (2000), Graham et al. (2000), Stefanovic et al. (2004), Leebens-Mack et al. (2005), Cai et al. (2006), some trees in Jansen et al. (2006b), Bausher et al. (2006), Chang et al. (2006), Wu et al. (2007), Huang et al. (2010: cf. rooting), Finet et al. (2010: support weak) and Moore et al. (2011: position not as stable as one might like...). Soltis et al. (2007; data from D. Soltis et al. 2000) found that the relationships obtained depended on the method of analysis; Bayesian analysis favoured [Amborellaceae + Nymphaeaceae], while parsimony yielded Amborellaceae as sister to the rest. What kinds of characters are analysed may also be important; Goremykin et al. (2009b) found an [Amborella + Nymphaea] clade after removing a relatively few (500) highly variable positions from the analysis. (Remember, too, that different seed plant topologies were obtained from analyses using single genes or the same number of sites chosen from twelve separate loci [Burleigh & Mathews 2007a], and maximum likelihood and maximum parsimony were susceptible to systematic error in an analysis of a twelve locus data set [Burleigh & Mathews 2007b].) On balance, however, Amborella alone being sister to all other extant angiosperms seems most likely. For other studies, see Ruhlman et al. (2007), Jansen et al. (2007) and Moore et al. (2007). Note that whichever position is occupies, optimisation of morphological characters onto the tree using strict parsimony will be little affected.

There are other possibilities. In a fairly recent study (Goremykin et al. 2003a) using complete chloroplast sequences, but for only 10 angiosperms, suggested the relationships [[[Amborellaceae + Calycanthaceae] [eudicots]] [monocots (= Poaceae only!)]], but poor taxonomic sampling with resultant long-branch attraction may be responsible for these results (D. Soltis & P. Soltis 2004; Jansen et al. 2004; Stefanovic et al. 2004; T. Degtjareva et al. 2004; D. Soltis et al. 2004). Goremykin et al. (2004) found the same general result when adding Nymphaea to their analysis; it linked with Amborella - which was still not sister to all other angiosperms. However, this study, too, suffers from the same basic sampling problem; grasses are highly derived monocots (see Kuhl et al. 2004 for the very distinct genome of Poaceae). Indeed, even when looking at complete chloroplast sequences of just a few flowering plants, the inclusion of Acorus, breaking up the long branch leading to Poaceae, had a major effect (Stefanovic et al. 2004), although there are also questions about the models used in the analyses (Lockhart & Penny 2005; Goremykin et al. 2005) and the jury may still be out (Goremykin et al. 2005: The monocots included in this analysis did not always form a monophyletic group...). Indeed, the non-monophyly of the monocots is a rather unlikely result that has been obtained in some studies where the nuclear gene 18S has been included (Troitsky et al. 1991: see Duvall et al. 2006 for references). Morton (2011: nuclear Xdh gene) found weak support for Ceratophyllaceae as sister to all other angiosperms, although otherwise much of the structure in her tree was similar to that of other studies mentioned here.

Finally, a further complication has been introduced by Lee et al. (2011). Like others, they found that Amboella and Nymphaeales weer successively sister to other angiosperms, but then monocots were sister to all angiosperms. Although this may be a sampling problem (101 taxa across seed plants as a whole: Austrobaileyales, Chloranthales or Ceratophyllales not included), Lee et al. (2011), there were massive amounts of data (almost 23,000 sets of orthologues from nuclear genomes).

Sampling strategies may well be critical, and this general issue has become particularly important in analyses using relatively few taxa each of which has relatively massive amounts of data. In some cases large amounts of data may indeed be the solution, in others, perhaps quite surprisingly little data per taxon but improved sampling will do the trick (e.g. Rokas et al. 2005; Hedtke et al. 2006), of course, exactly where sampling should be improved is important (Geuten et al. 2007), and each situation will have to be evaluated independently. The recent discovery of an association of Hydatellaceae with Nymphaeales (Saarela et al. 2006) unexpectly allows sampling in this area of the tree to be improved; however, that family consistently links with Nymphaeaceae and Cabombaceae and does not affect the overall topology of the tree. Finally, some kinds of DNA data may be positively misleading when it comes to understanding relationships (Duvall & Ervin 2004; Qiu et al. 2005; Duvall et al. 2006, 2008b; G. Petersen et al. 2006b), with horizontal transfer being notably common in mitochondrial genomes (Sanchez-Puerta et al. 2008; Hao et al. 2010; cf. Cusimano et al. 2008).

Within angiosperms, there are convenient summaries of the copious literature on relationships between the major clades in e.g. P. Soltis & D. Soltis (2004), D. Soltis et al. (2005b) and Qiu et al. (2005). There is additional literature cited at individual nodes, see especially the notes immediately preceding the Magnoliales, i.e. the magnoliid clade, and Acorales (both the [monocot + eudicot] clade and monocots themselves), Ranunculales (eudicots), Berberidopsidales (core eudicots), and Cornales (asterids). The discussion below is based on the usually rather conservative topologies of the trees in this site.

Classification. For bibliographic information on familial and ordinal names, etc., see Reveal and Chase (2011).

Synonymy: Alismatidae Takhtajan, Arecidae Takhtajan, Aridae Takhtajan, Asteridae Takhtajan, Bromeliidae C. Y. Wu et al., Burmaniidae Heintzw, Caycanthidae C. Y. Wu et al., Catyophyllidae Takhtajan, Ceratophyllidae Doweld, Chloranthidae C. Y. Wu et al., Commelinidae Takhtajan, Cornidae Reveal, Dillenidae Reveal & Takhtajan, Ericidae C. Y. Wu et al., Hamamelididae Takhtajan, Iliciidae C. Y. Wu et al., Juncidae Doweld, Lamiidae Reveal, Lauridae C. Y. Wu et al., Liliidae J. H. Schaffner, loranthidae Tieghem, Malvidae C. Y. Wu et al., Myrtidae J. H. Schaffner, Nelumbonidae Takhtajan, Nymphaeidae Takhtajan, Orchididae Heintze, Piperidae Reveal, Plumbaginidae C. Y. Wu et al., Polygonidae C. Y. Wu et al., Ranunculidae Takhtajan, Rosidae Takhtajan, Rutidae Doweld, Theidae Doweld, Triuridae Doweld, Winteridae Doweld, Zingiberidae Cronquist - Magnoliophytina Reveal - Magnoliophyta Reveal

SEED PLANT EVOLUTION (still being developed)

Lignophytes are sister to the monilophytes, ferns and their relatives. The split between the two clades is old, occurring mid- to later Devonian, some 375-400 million years before present (Pryer et al. 1995, 2000, 2001a, 2004; Schneider et al. 2002). Lignophytes are characterized by having distinctive secondary thickening: There is a bifacial vascular cambium producing secondary phoem externally and secondary xylem internally (e.g. see Donoghue 2005). Thus lignophytes often have stems with large amounts of wood, and there is also a relatively thin bark produced by a separate cork cambium; Robinson (1990) noted that the ratio of periderm to xylem was 8-20:1 in lycopsids and <1:4 in extant seed plants. The origins of seed plants, the focus of this site, are to be sought in mid-Devonian lignophytes, the progymnosperms, often homosporous plants with complex leaves and well-developed secondary thickening with much parenchyma mixed in with the tracheids (see below). The leaves proper may have been small, although the branch systems as a whole may have been flattened; there are no obvious leaf traces. The origin of a stele more like that of extant spermatophytes can be seen in taxa in which the solid central vascular tissue became ridged and dissected into vertical columns, pith developing (Stewart & Rothwell 1993 for a good summary). The leaves of extant seed plants are sometimes called megaphylls, and they may represent modified branch systems. Thus Floyd and Bowman (2010) compared gene expression patterns in shoots and leaves of seed plants and found that the marginal blastozones of leaves and the shoot apical meristem were similar in some respects, consistent with this hypothesis.

Some plants with the kind of vegetative morphology described in the previous paragraph may have been heterosporous, stems having been found in the same fossil beds as fossilized seeds (see e.g. Beck 1962, 1981; Carluccio et al. 1966; Namboodiri and Beck 1968). The progymnosperm Archaeopteridales and Aneurophytales plants of this kind; they have a relatively very rich fossil record, and may have morphologies very unlike those of any living plant, whether angiosperm or gymnosperm. Seed plants or spermatophytes in general are characterised by heterospory, i.e. the plant produces a number of microspores ("pollen") per microsporangium and a single megaspore per megasporangium (Kenrick & Crane 1997). Heterospory has evolved several times in land plants, although in seed plants it is quite distinctive because the dividing megaspore receives its nutrition from the parental sporophyte, whereas in other heterosporous land plants megaspore development is independent of the sporophyte. The ovule is the megasporangium and enveloping integument(s), and the seed is an ovule when it drops from the plant; ovules are known from the Devonian (Stewart & Rothwell 1993). In the Carboniferous in particular there was a considerable variety of plants with fern-like leaves and ovules, the seed ferns or pteridosperms (the Carboniferous has sometimes been called the age of ferns, rather, it is the age of tree ferns). Recent studies are helping to clarify their morphology (Taylor et al. 2006 and references, also other papers in J. Torrey Bot. Soc. 133(1). 2006), and this will help us to understand the phylogeny of seed plants as a whole. By the lower Carboniferous there were to be found the rather conifer-like Cordiatales, which had compound pollen-bearing structures and saccate pollen, and slightly later, the still more conifer-like ("ancestral") Voltziales. Seeds of Mesozoic seed plants are very diverse in their morphology (e.g. Anderson & Anderson 2004), and Bennettitales, Corystospermales (pteridosperms), which seem to have survived the end-Cretaceous mass extinction in Tasmania (McLoughlin et al. 2008), and conifers were all radiating in the Triassic.

Note that a major distinction has been drawn between manoxylic and pycnoxylic taxa. Secondary xylem of the former has much parenchyma mixed in with the tracheids, while in the secondary xylem of the latter there is much less parenchyma. The cycadophytes, which include seed ferns, cycads, and the immediately unrelated cycadeoids (e.g. Bennettitales), have manoxylic wood, while the coniferophytes, which include all other extant gymnosperms and several fossil groups (Chamberlain 1935; see also Gifford & Foster 1988), have pycnoxylic wood. Within gymosperms as a whole, manoxylic wood, megaphyllous leaves, and radiospermic (polysymmetric) seeds seem to be associated, as do pycnoxylic wood, microphyllous leaves (perhaps not strictly microphyllous - see Crown Euphyllophytes for the distinction between microphylls and megaphylls), and platyspermic (disymmetrical) seeds (Sporne 1965). It is perhaps unlikely that these represent completely independent lines of evolution, especially if Bennettitales are close to Gnetales (see below, but c.f Crepet & Stevenson 2009), and manoxyly versus pycnoxyly and radiospermy versus platyspermy seem not to represent fundamental distinctions.

Some early conifers and Cordiatales had microspores of a kind often called prepollen. There is no sulcus, but trilete or monolete ridges where the pollen grains were originally attached (these ridges are called haptotypic marks, i.e. they are proximal markings on the mature grains denoting where they were attached in the tetrad before it broke up). The development of the male gametophyte probably took place inside the spore, and germination occurred via these ridges (e.g. Friedman 1993; Friedman & Gifford 1997). All extant seed plants have true pollen in which germination is plesiomorphically distal, that is, the pollen tube grows out through the part of the pollen grain that was not adjacent to the the other members of the tetrad, and there are no haptotypic marks. However, the relationships between fossil plants with prepollen, other fossils with true pollen, and extant gymnosperms, which also have true pollen, are not well understood. A distinction is sometimes made between coniferophytes and conifers, although what the two contain and their relationships are unclear (Rothwell & Mapes 2001); phylogenetic studies are certainly not suggesting a single answer (e.g. Crane 1985b; Doyle & Donoghue 1986a, 1992; Rothwell & Serbet 1994; Doyle 1996; etc.).

Relationships among extant seed plants.

In the 1980s and 90s morphological phylogenetic studies suggested that extant seed plants were probably to be placed in five groups: cycads, Ginkgo, conifers, Gnetales (Gnetum, Ephedra and Welwitschia), and angiosperms. Gnetales and a larger or smaller group of fossil gymnosperms/pteridosperms were together sister to angiosperms (e.g. Crane 1985a, b; Doyle & Donoghue 1986a, b; Nixon et al. 1994; Doyle 1998a, b); Doyle (in Sanderson et al. 2000) noted that this position was well supported in bootstrap analyses that were carried out subsequently. Extant "gymnosperms" were thus thought to be paraphyletic, the botanical equivalent of "reptiles". Plants with a heterosporangiate strobilus, the so-called anthophytes, included flowering plants, Gnetales, and also fossil taxa like Bennettitales, while the glossophytes, also fairly close cladistically to the anthophyte clade, also included the glossopterid seed ferns (these seem to have had multiflagellate male gametes - Nishida et al. 2004). Conifers, cycads, etc., were more distantly related to flowering plants. Analyses of morphological data, especially those that include fossil taxa, continue to suggest that gymnosperms are paraphyletic, the four main groups being independently derived from a pteridosperm grade, with Gnetales close to angiosperms and often associated with Bennettitales, etc., thus they support some kind of anthophyte hypothesis (Doyle 2006; Hilton & Bateman 2006; Rothwell et al. 2009; Schneider et al. 2009; Crepet & Stevenson 2010). However, bootstrap support for these relationships is very low. A recent study (Wang 2010) even suggested that possible relationships among seed plants included a paraphyletic Gnetales, with angiosperms sister to [Gnetum + Welwitschia]; a morphological analysis of relationships among extant seed plants by the same author had placed [Archaefructus + Ceratophyllum] as sister to all other angiosperms.

Such relationships are strongly questioned in most analyses of molecular data, and extant gymnosperms appear to be monophyletic (e.g. Goremykin et al. 1996, Frohlich & Parker 2000; Antonov et al. 2000; Aris-Brosou 2003; Magallón & Sanderson 2002, including a summary of the literature; Qiu et al. 2006: support weak). Within extant gymnosperms, many studies suggest that Cycadales may be sister to all others (see Hasebe 1997 for the early literature). On the other hand, an association of Cycadales and Ginkgoales has also been recovered, especially in maximum parsimony analyses (e.g. Raubeson et al. 2006; Wu et al. 2007; Chumley et al. 2008; Finet et al. 2010). But yet other relationships appear in at least some molecular analyses. Thus Mathews et al. (2010) suggest that cycads are sister to angiosperms (no support values given), this clade in turn being sister to other gymnosperms; note that morphological data optimised with this topology as a constraint tree also yielded little bootstrap support, and posterior probabilities from unconstrained analyses were also very low (Mathews et al. 2010).

Long-branch attraction involving the branch leading to angiosperms (Rydin & Källersjö 2002; Stefanovic et al. 2004) may affect the results of these molecular studies, especially the position of Gnetales, but this is very hard to deal with given the paucity of extant gymnosperm taxa (see also Hilton & Bateman 2006 for sampling in the context of morphological versus molecular phylogenies, molecular results will necessarily be flawed because the sampling cannot be improved; Bateman et al. 2006b for much else besides). If there seems to be no extant taxon that could be used to break up the branch (see Geuten et al. 2007 for discussion, albeit in rather easier - although still difficult - examples), knowledge of fossils needs to be much improved for a phylogeny whose topology is determined by fossils to be convincing. Indeed, in some "basal" chordates it has been found that as the organism decayed later-derived characters tended to become unrecognisable before earlier-derived characters, hence fossils would tend to be assigned a more "basal" position in the tree than they should, and the age of the clade represnted by these fossils possibly underestimated... (Sansom et al. 2010, 2011).

Even if Gnetales are not particularly close to angiosperms, their position is still a matter of considerable interest. Some phylogenies suggest that Gnetales are sister to Pinales (e.g. Antonov et al. 2000; Sanderson et al. 2000; Chaw et al. 2000; Gugerli et al. 2001, rather strong support; de la Torre et al. 2006; Wu et al. 2007; the preferred topology in Englund et al. 2011 and Groth et al. 2011), or are even to be placed within Pinales, in particular being associated with Pinaceae (e.g. Burleigh & Mathews 2004; Hajibabaei et al. 2006: sampling within Pinales poor; Qiu et al. 2007; Finet et al. 2010: quite strong support; Zhong et al. 2010). Note that there may be problems in such analyses depending on exactly what sequences are analysed and how they are analysed (e.g. Burleigh & Mathews 2007a, b; Zhong et al. 2010). All ndh genes in the chloroplast of Pinus thunbergii are absent - or are present, but as pseudogenes (Wakasugi et al. 1994); recent work suggests that these genes are absent in all Gnetales and Pinales alone - additional support for the gnepine hypothesis (Braukmann et al. 2009). The rps16 gene in Gnetales and Pinaceae is sommonly lost (Wu et al. 2007, 2009). All Pinales sampled have but a single copy of the chloroplast inverted repeat (Strauss et al. 1988); all other seed plants have two copies (Raubeson & Jansen 1992), and so the incorporation of Gnetales within Pinales will imply parallel evolution of this loss (Lackey & Raubeson 2008), and this may be marked by micromorphological changes in the genome. Interestingly, one end of the inverted repeat of Welwitschia has expanded (Welwitschia is derived within Gnetales) with duplication of trnI-CAU and partial duplication of pscbA gene region at the end of the Large Single Copy region, and these match those of the remnant inverted repeat known from Pinus and other Pinaceae, but not other members of Pinales (Margheim et al. 2006; McCoy et al. 2006, 2008: note details of relationship depend on methods of analysis; see also Braukmann et al. 2009; Hirao et al. 2009). An analysis of variation in 83 plastid genes strongly suggested a set of relationships [Pinaceae [Gnetales + other Pinales]], although other relationships cannot be entirely rejected (Chumley et al. 2008). In an analysis of an amino acid matrix derived from chloroplast genomes, depending on whether or not quickly-evolving proteins and proteins in which there appeared to be much parallel evolution were removed, Gnetales were sister to Cupressaceae or to Pinaceae, the latter being the prefered position (Zhong et al. 2010). Clearly, Gnetales were on a long branch (Zhong et al. 2010).

Background.

Evolution in Stem Group Angiosperms.

Angiosperms and Insects.

Phytophagy.

Pollinators.

Angiosperms and Fungi.

Tertiary Diversification

The physiological-ecological context of angiosperm evolution.

Some patterns of diversity in extant angiosperms.

In Conclusion.

Background

When thinking about evolution in general, a well-supported phylogeny is of course a sine qua non. Beyond this, there are issues of dating, understanding fossils, working out diversification rates, optimising characters on trees, etc., that need to be taken into consideration. I discuss these issues briefly below, but please consult the primary literature for details; see also Freckleton (2009).

1. Dating is critical, but this whole issue remains a subject of intense discussion (e.g. Magallón and Sanderson 2001; Graur & Martin 2003; Pirie et al. 2005; Renner 2005b; Bell & Donoghue 2005; Magallón & Sanderson 2005; Rutschmann et al. 2007; Sanderson et al. 2004; H. Wang et al. 2009; Smith et al. 2010: molecular dating; Crepet et al. 2004: palaeontological dating; Magallón 2009; Milne 2009: sampling). Many clade ages are not very reliable at this stage of our knowledge, and in several cases there are substantially different estimates for the same event (e.g. compare Wikström et al. 2001, 2004; Davies et al. 2004: diversification rates in the context of a dated supertree; Soltis et al. 2008). The relaxed ages given by Magallón and Castillo (2009) are often substantially older than the constrained ages - for example, the relaxed crown group age for angiosperms is about 242 million years, and the constrained age about 130 million years. There are other older ages, e.g., see Smith et al. (2010), while Clarke et al. (2011) also reject post Jurassic ages for the origin of crown angiosperms, their estimates for the age of this clade being 240-175 million years. Molecular and paleontological dating can seem to be in conflict, and the former may give substantially older ages than the latter. Of course, fosils can yield only a minimum age (Donoghue & Benton 2007) - and note that fragmentary fossils in particular may be assignable to more than one node. Nevertheless, in some groups like platanoids (Crepet et al. 2004) the fossil record is quite substantial, and that of early angiosperms has been summarized by Friis et al. (2011). However, the fossil record is sure to have surprises, and the recent discovery of Leefructus from early Cretaceous deposits 125.8-122.6 million years old and assigned to stem Ranunculaceae (Sun et al. 2011) will, if confirmed, challenge many of our current ideas of angiosperm evolution. Many stem- and crown-group age estimates for orders, families, etc., are given here, but, hardly surprisingly, they should all be treated with extreme caution.

Be this as it may, it seems likely that the distribution patterns of a number of clades that were thought to reflect vicariance caused by plate tectonic events are quite often better explained by much more recent dispersal/migration events (e.g. Renner 2005b and de Queiroz 2005 for summaries; Higgins et al. 2003 and Nathan 2006 for mechanisms, about which we know little; Gillespie et al. 2012 and references; also Yoder & Nowak 2996; Wen & Ickert-Bond 2009; Carpenter et al. 2010; cf. in part Ladiges & Cantrill 2007; Heads 2008, etc.). Even the distributions of Lars Brundin's iconic chironimid midges may need reinterpretation (Krosch et al. 2011).

2. Understanding fossils is of course central to understanding angiosperm evolution (see also 5 below). Weins et al. (2010 and references) is a good example of the integration of morphological and molecular data, and fossils and extant organisms. Both amount and quality of data are important, less so proportion, i.e. the ratio of molecular to morphological characters, and of course it is not simply data that matter, but how they are analysed (Morlon et al. 2011).

However, as with dating, one has a sense of unease, and some studies question what had previously thought to be fairly well established fossil records (e.g. Cook & Crisp 2005 - Nothofagus; Biffin et al. 2010b: Araucariaceae). In groups like Poaceae (Poinar 2004: see below, also under Poaceae) there are amber fossils 110-100 million years old from the Early Cretaceous of Myanmar/Burma that would seem to suggest a substantially greater age for the clade and its diversification than is given by other dating methods, indeed, they would change many dates suggested for other flowering plants below and cause a rethinking of angiosperm evolution. Although there may be questions about the age of this fossil (Caroline Strömberg, pers. comm.), Poinar (2011) recently confirmed its probable identity as Poaceae-Pooideae. Moreover, other angiosperm fossils have been found in these amber deposits. These include a possible core eudicot with sepals, petals, an inferior ovary, and apparently a single style (?somewhere around Cornales - Poinar 2011), and a possible rosid with a floral formula of K 5, C ?, A 10, G [5], styles diverging (Poinar et al. 2007, 2008), and it is difficult to know what to make of them. Such issues aside, the identification of fossils and their selection for calibration of molecular trees should be treated very carefully (e.g. Gandolfo et al. 2004; Graham 2010; Clarke et al. 2011; etc).

3. Although "diversification" is mentioned frequently below, both it and the related term, "adaptive radiation", are very imprecise and difficult to estimate (e.g. Sanderson 1998; Davies et al. 2004; Ricklefs 2007; Olson & Arroyo-Santos 2009; Ackerly 2009 for measurements of aspects of radiation). However, interpretation of curves showing diversity in clades over time is not simple. In particular, what can seem like an abrupt radiation, with rapid diversification after a period when there was little apparent diversity - the "broom and handle" and "stemmy" patterns evident in many clades - may rather be the result of extinction, diversification after the extinction event resuming at a rate similar to that before the event and giving the appearance of a radiation (Crisp & Cook 2009). Quite extensive sampling (>80%) may be needed if accurate estimates of slowdowns in diversification are to be made (Cusimano & Renner 2010). Simple experiments estimating future extinctions showed that these might affect estimates of imbalances of clade size at nodes of some 50 million years age (Clarke et al. 2011). In general, estimating clade imbalance is a remarkably tricky operation, especially in the near absence of fossils, the usual situation (Tarver & Donoghue 2011; see also e.g. Rabowsky 2010a, b). Even when there is an excellent fossil record, strategies like removing recently-radiating clades may be necessary if one is to detect periods of diversity loss in other clades (Morlon et al. 2011, the example is cetaceans [whales, etc.]).

4. The apparently simple issue of numbers of species is in fact not that straightforward. We have to be very careful even when discussing the size of extant clades. There are two aspects to this - what is really the clade of interest?, and, how many species does it contain?

A. Orchidaceae are often considered to be a highly diverse clade, at least in terms of numbers of species, but since they are sister to the rest of the Asparagales, the disparity in species number, although considerable, is only three-fold (ca 22,000 vs. 7,100), and by some measure the Asparagales minus Orchidaceae could be considered vegetatively and even florally more diverse than Orchidaceae - although it is very hard to provide measures of morphological diversity. Furthermore, Asparagales, with ca 29,000 species, are sister to commelinids, with some 23,500 or more species, while within Orchidaceae much of the diversity is concentrated in the largely epiphytic Epidendroideae. So the related questions, "Are orchids really diverse, and if so, what do we mean?", are not easy to answer. Perhaps we should consider Epidendroideae, or a clade within it, to be the hyperdiverse group; it is clear that there are many similar examples of extreme clade size imbalance throughout the tree. However, recent work is beginning to move beyond simplistic "major clade"-type comparisons (e.g. Smith et al. 2011); again, positions of rate shifts there should be taken with a grain or two of salt and the "small backbone tree" is an extreme way of looking at trees.

B. Estimates of the number of extant species of flowering plants vary by a factor of about two - 422,127 (Govaerts 2001) to 223,300 (Scotland & Wortley 2003) - and perhaps add 20% (Joppa et al. 2010).

5. Exactly how one goes about optimising characters on trees is critically important. Thus using either parsimony or maximum likelihood, making apparently reasonable assumptions about weighting gains over losses (or vice versa), or just using the rather simple models of evolution explicit in ACCTRAN or DELTRAN to place the character on the tree (e.g. Donoghue & Ackerley 1996; Cunningham et al. 1998; Omland 1997, 1999; Ree & Donoghue 1999; Polly 2001; Webster & Purvis 2001; Ronquist 2004; Crisp & Cook 2005), may greatly affect the position of synapomorphies on trees, and hence our ideas of evolution. Sannier et al. (2007) give a good example concerning where on a tree one might peg changes in microsporogenesis in palms (see also Sannier et al. 2011).

There are related issues here. Although sampling in molecular studies is usually more or less incomplete, we often forget that the same is true of the morphological and chemical characters in whose evolution we are interested. Indeed, molecular sampling is now often better than morphological/chemical sampling. For many anatomical, chemical and embryological characters that are confidently said to characterise families and other groups, we may well have no idea if those characters are applicable to the whole clade, or just to a subgroup within that clade. Thus Albach et al. (2001a, see also D. Soltis et al. 2005b) assign possession of iridoids to the base of the asterid I + II clades. However, this feature is placed higher up the tree here, partly because of topological uncertainties, but partly because in Lamiales (for example), the first four clades that are successively sister to the remaining Lamiales either lack iridoids or (most Oleaceae) have iridoids different from those found in the other members of the clade. Similar problems arise when thinking of the evolution of ellagic acid in Ericales (Stevens 2006b), and here we run into the sampling problem, too. In nearly all studies of the evolution of characters (D. Soltis et al. 2005b is a good example), the distributions of those characters are optimised on more or less fully resolved trees; of course, some nodes may have little support, and optimisations based on such trees carry correspondingly little conviction...

6. The relationships of angiosperms to other seed plants still remain an abominable mystery - and as a result so do the whens, whys and hows of their initial diversification (see Davies et al. 2004b; Friis et al. 2005; Frohlich & Chase 2007; Pennisi 2009; Lee et al. 2011). Here a distinction needs to be made between the origin of the clade of which angiosperms are the only extant representative, i.e. stem angiosperms ("origin 1"), the origin of plants with carpels, tepals, and a heterosporangiate strobilus, i.e. the evolution of plants with flowers ("origin 2"), and finally, the origin of crown angiosperms, i.e. the clade represented by the immediate common ancestor of flowering plants as they occur today ("origin 3"). Stem angiosperms presumably are of early Carboniferous age or even older, 350±35-305-275±35 million years old, if the angiosperm clade is sister to the clade including all living gymnosperms (e.g. Savard et al. 1994; Crane et al. 1995; Crane 1999; Clarke et al. 2011), or even just to Pinales, to a younger bound of Permian in age (Doyle 1998a). For the bulk of their some 100 million years plus history stem angiosperms will probably have lacked flowers and even carpels and will have had naked seeds and other features of the extant gymnosperms (see above, cf. mammal-like reptiles and mammals). Even if crown angiosperms are somewhere between 270 and 182 million years old in age (Smith et al. 2010), there is still a substantial stem history that is largely unknown. Indeed, although current evidence (see below) now suggests that extant gymnosperms are monophyletic, when considering both extant and fossil taxa gymnosperms are paraphyletic with respect to angiosperms; angiosperms are derived from a gymnospermous ancestor. Little progress has been made in idenfiying "origin 2" angiosperms over the last twenty years or more.

7. Understanding the ecological and environmental contexts of the evolution of angiosperms is obviously made very difficult because of the unresolved problems just discussed above. Thus questions such as, Why are angiosperms so diverse? may have several answers. There is a series of factors - seed plant morphology, and its interactions with the environment s.l. - that have shaped stem- and crown-angiosperm diversification through the Cretaceous and back to the Permian. For instance, is the apparent Jurassic origin of bee-flies connected with the evolution of angiosperms then (see Wiegmann et al. 2011)? The ecological factors that are connected with the early Tertiary diversification of angiosperms are likely to be quite different from those in which this initial evolution of angiosperms occured, whenever that might be. The Tertiary factors are perhaps initially connected with or shaped by the bolide impact at the K/T boundary, although much of angiosperm diversity as we now appreciate it seems to be a phenomenon of the later Tertiary (e.g. Tiffney 1985a, b). For instance, Magallón et al. (1999) noted that major core eudicot clades like Fabaceae and (most of) Lamiales that together represent about 45% of core eudicot diversity appear only in the upper Cretaceous (Maastrichtian) and Tertiary. Even in far older clades like Myristicaceae, crown group diversification may also be largely a Tertiary phenomenon (J. A. Doyle et al. 2004, 2008; Scharaschkin & Doyle 2005; Richardson et al. 2004; but cf. Couvreur et al. 2011a, c).

Although we think of the evolution of angiosperms as being intimately connected with and dependent on the evolution of their pollinating insects, most angiosperms are symbiotic systems at a variety of levels. Features ascribed to plants may well be the result of interactions of plants and microbes (Friesen et al. 2011), and this goes far beyond the obvious endosymbiotic events that resulted in chloroplasts and mitochondria. Importantly, basic angiosperm physiology is mediated by fungal mycorrhizae and bacteria in the soil and by fungal endophytes in the plants, and this shaped and continues to shape both the local and the global environments.

8. All in all, it is a challenge to think about the evolution of novelties, morphological or otherwise. It is not simply that the gene/genome duplications that appear to be so pervasive in angiosperm evolution (see Soltis et al. 2009 for a summary) allow of a myriad possibilies for subfunctionalisation and neofunctionalisation, as well as the evolution of novel regulatory gene pathways (e.g. de Martino et al. 2006). Indeed, such duplications are widespread in land plant evolution. Jiao et al. (2011) provide evidence for a genome duplication in the lineage basal to all extant seed plants, and date the peak of the age curve of duplicated genes to (352-)349, 347(-343) million years ago in the early Carboniferous (Mississippian) - the overall age spread is ca 400 to just over 250 million years. Lang et al. (2010) discuss the evolution of transcription-associated proteins, perhaps associated with genome duplications; three new families evolved somewhere between the lycophytes and flowering plants. As more becomes known about details of molecular evolution, it is clear that there has been widespread homoplasy in the cooption of particular genes in vegetative development (e.g. Rosin & Kraemer 2009; Blein et al. 2010 for the development of compound leaves), while in floral development the CYC gene, for example, has been coopted numerous times in the evolution of core eudicots (Rosin & Kramer 2009). The CYC gene is expressed in the dorsal/adaxial part of the floral meristem and has often become involved in the development of monosymmetric flowers (e.g. Damerval & Manuel 2003; Preston et al. 2011b). Parallelisms occur at the amino acid level as in C₄ photosynthesis (e.g. Brown et al. 2011 for parallelisms between grasses and Capparaceae). As Preston et al. (2011b) put it as they summarized aspects of the developmental evolution of angiosperm flowers, "reduce, reuse, and recycle" has been the order of the day, and it seems that old dogs can indeed be taught new tricks (Rosin & Kraemer 2009). Add heterochrony (the male gametophyte of flowering plants is a good case in point - e.g. see Takhtajan 1976) and heterotopy to the mix (e.g. Baum & Donoghue 2002), and development becomes a very flexible operation.

It has been suggested that for some secondary metabolites in particular the important evolutionary step is the acquisition of the capability to synthesise a particular metabolite; this may then be switched off easily, but not lost, and so the metabolite can sometimes be "reacquired" again (e.g. Wink 2003; Liscombe et al. 2005). Furthermore, the sugar donor specificity of the enzymes which conjugate flavonoids with sugars are able to change quite readily (Noguchi et al. 2009), while individual terpene synthases may have many products, thus gamma-humulene synthase of Abies grandis can generate 52 different sesquiterpenes (Degenhardt et al. 2009). Hence the rather spotty distribution of some of these metabolites when considered in the context of phylogenies; note also that endophytic fungi may be the organisms that are actually producing the metabolite normally ascribed to the plant partner (see below).

In other cases particular phenotypes may be the result of parallel mutations that effect change only because of some previous but as yet undetected change in the larger clade (see Shubin et al. 2009 on deep homology) - hence "evolutionary tendencies" (for which, see Cantino 1985; Sanderson 1991). The ability of a plant to form an association with nitrogen-fixing bacteria is an example, and there is some discussion about this on the Fabales page. Similarly, the clustered origins of C₄ photosynthesis probably have a aimilar explanation (see e.g. McKown et al. 2004; Christin et al. 2011a; Grass Phylogeny Working Group II 2011). At a more molecular level, Irish (2009) suggests that features such as petals may evolve several times because of the independent cooption of underlying gene regulatory networks. Whatever the reasons, some characters - and not only those of secondary plant chemistry - seem to come and go on the tree almost willy-nilly, which, as just mentioned, makes their optimization a distinctly hazardous profession.

Evolution in stem group angiosperms. (This section in particular needs lots of work.)

Although it is unclear which seed plant fossils are stem-group angiosperms, such plants will, of course, be largely gymnospermous in their morphologies, regardless of whether extant gymnosperms are monophyletic or paraphyletic. The most promising candidates for fossil relatives of angiosperms include Corystospermales (Pteruchus, Ktalenia, etc. - see the "mostly male" theory of flower evolution, Frohlich & Parker 2000), Bennettitales, and Caytoniales. Cycadeoids or Bennettitales - "fossil beehives" - have long been associated with angiosperms (see also Doyle 2006 and Hilton & Bateman 2006 for cladistic analyses and entry into the older literature).

However, recent detailed work on seed morphology and anatomy in particular, but also pollen morphology, have suggested that Bennettitales should be placed in the BEG (Bennettitales, Erdmanithecales and Gnetales) group, along with Gnetales and some other wholly fossil assemblages like Erdmanithecales that persisted into the Late Cretaceous (Friis et al. 2007, 2009a, the latter describing four new genera in this complex; Mendes et al. 2010: cf. Rothwell et al. 2009). This group has radiospermic ovules that lack a cupule, they have a vascularized nucellus but not a vascularized integument, and their seeds have an outer sarcotesta, a sclerotesta, and a layer inside that (Rothwell & Stockey 2002: for further discussion, see seed plant evolution). Crane and Herendeen (2009) provide a careful interpretation of the reproductive structures of Bennettitales.

Whether plants with such seeds will shed direct light on angiosperm evolution is still unclear, but it is unlikely; in any event, some of the early (Upper Triassic) bennettitalean reproductive morphologies are rather different from those of later fossils (e.g. Pott et al. 2010). Gnetales themselves are far from having a flower, and the interpretation of the complex reproductive structures of Bennettitales is not easy (see also Crane & Herendeen 2009). Indeed, the consensus of molecular studies is that Gnetales are best placed inside Pinales, this position being supported by a growing amount of data (see above). If this position is confirmed, it means that the immediate relatives of Gnetales have little to do with angiosperm origins. Some morphological work questions any particular similarity between the "flower" of Bennettitales and those of angiosperms (Rothwell et al. 2008a, 2009; Crepet & Stevenson 2009, esp. 2010 - although in the last-mentioned paper it is difficult to interpret a morphological analysis in which relationships among the extant angiosperms included are largely different from those accepted here and the extant gymnosperms are not monophyletic in some analyses, and when changing a single character state in one taxon results in the loss of eleven nodes in the strict consensus tree...). Other pteridosperm groups that have been linked to angiosperms are Caytoniales, Pentoxylon and glossopterids (e.g. Soltis et al. 2008 for references), but in a recent comprehensive review on the bearing of fossil data on the origin of the flower, the overall conclusion was that our understanding of the fossil record was insufficient to help much in answering questions of angiosperm origins (Doyle 2008b).

Overall, it seems unlikely that Gnetales have anything immediately to do with flowering plants, although this hypothesis (the anthophyte hypothesis) was popular back in the 1990s, as several papers in Taylor and Hickey (1995) suggest, and the idea still persists (see Wang 2010, Fig. 8.10; Crepet & Stevenson 2010). This will be true of those seed ferns like Bennettitales that have been associated with Gnetales if current ideas of relationships hold, although Rothwell et al. (2009) strongly question the idea of a close relationship between Bennettitales and Gnetales. Indeed, some find no particular similarity between the "flower" of Bennettitales (Rothwell et al. 2008a, 2009) and that of angiosperms, and there are suggestions that the envelopment of the seed to produce a fruit-like structure has happened both within Bennettitales (Rothwell & Stockey 2010) as well as in angiosperms. But even in some morphological analyses where Bennettitales do not group with the anthophytes and are associated with cycadofilicalean plants (Crepet & Stevenson 2009, esp. 2010), gymnosperms are also not monophyletic and Gnetales are sister to angiosperms!

Despite the strong suggestion from analyses of molecular data that Gnetales have no immediate or even particularly close relationships with flowering plants, the issue is not yet completely closed (Rydin et al. 2002: Friis et al. 2007). In particular, analyses of morphological data that include fossil taxa continue to suggest that gymnosperms are paraphyletic, the four main groups being independently derived from a pteridosperm grade, with Gnetales close to angiosperms and often associated with Bennettitales (see also detailed studies of seeds: Friis et al. 2007; for pollen, with homogeneous or granular walls, see Zavialova et al. 2009), etc. However, bootstrap support for these relationships is very low (Doyle 2006; Hilton & Bateman 2006), as it was for the original anthophyte clade. Some recent work questions any particular similarity between the "flower" of Bennettitales (Rothwell et al. 2008) and that of angiosperms, and even some morphological analyses remove Bennettitales from the anthophytes and associate them with plants of the Cycadofilicales (Crepet & Stevenson 2009). Interestingly, rather than carrying out an independent morphological study, Doyle (2006) himself proceeded to think of seed plant evolution in the context of a morphological analysis that was constrained by a molecular topology in which Gnetales were nested within gymnosperms (he noted that this is in fact almost as parsimonious as if Gnetales were linked with angiosperms). Hilton and Bateman (2006), on the other hand, allow only a slight possibility that their morphology-based tree could be superseded (see also Farjon 2007). This argument, morphology with/without better/worse than molecules, is independent of group being studied (see e.g. Springer et al. 2007 for mammals).

Detailed morphological work suggests that angiosperm characters apparently found also in Gnetales fail Remane's criteria and are for the most part parallelisms. Thus the sieve areas in the phloem cells of Gnetales are very like those of other gymnosperms and are unlike those of the sieve tubes of angiosperms (Behnke 1990a), the vessels in the two develop differently (e.g. Carlquist 1996), and the tunica has only a single layer, not two or three as in angiosperms, etc. (e.g. Donoghue & Doyle 2000a; Doyle 2006). Similarly, the adaxial tension (reaction) wood in Gnetum, produced as the branches react against gravity to maintain their orientation, consists of gelatinous extra-xylary (reaction) fibres in the adaxial position - i.e., it is unique among seed plants, and is unlike the tension wood of angiosperms (Tomlinson 2001b, 2003; see also Höster & Liese 1966). The loss of sperm flagellae and the associated development of a pollen tube growing towards the ovule thus becomes a parallelism between [Pinales + Gnetales] and angiosperms, and venation density of Gnetum (Boyce et al. 2009) and details of the timing of events in the pollination-fertilization process (Williams 2008) suggest that there are yet more parallelisms between the two. In any event, wherever Gnetales are placed, they are clealy a very derived group.

This is probably another case where relationships that are supported by morphology alone deceive. Although long-branch attraction involving the branch leading to angiosperms (Rydin & Källersjö 2002; Stefanovic et al. 2004) may affect the results of these molecular studies, especially the position of Gnetales, it is very hard to deal with given the paucity of extant gymnosperm taxa (see also Hilton & Bateman 2006 for sampling in the context of morphological versus molecular phylogenies; Bateman et al. 2006b for much else besides). It is unclear which extant taxon might break up the branch, and our knowledge of fossils will need to be much improved for a phylogeny whose topology is determined by fossils to be convincing (e.g. see above).

All this being said, based on the patterns of duplication of PHY genes - and the assumption that they are not lost - Schmidt and Schneider-Poetsch (2002) had suggested that within the gymnosperms, Gnetales were sister to the rest since they had fewer duplicated genes (but see above); an early maximum parsimony analysis of Samigullin et al. (1999) had found a similar position. Indeed, a recent analysis of a large amounts of nuclear gene data from 101 genera of seed plants also suggests that Gnetales are sister to all other gymnosperms, and that Ginkgo and Cycadales are also sister groups (Lee et al. 2011: see also Cibrián-Jaramillo et al. 2010: most data from ESTs, much missing). Lee et al. (2011) discuss character evolution in the context of the distinctive topology they found, suggesting that "motile male gametes would be independently and uniquely evolved (apomorphic) in cycads plus Ginkgo, and loss of motile male gametes in Gnetales would would be ancestral in the gymnosperms (plesiomorphic)." (ibid., p. 2). From this point of view, and disregarding fossils, such a loss would be an apomorphy for [angiosperms + gymnosperms], and there would need to be a mechanism that would allow the restoration of motility to male gametes of cycads, e.g. regaining flagellae, etc. At the same time other character evolution interpreted in the context of this tree still need not be the same as in the Anthophyte hypothesis (cf. Lee et al. 2011). Even if this topology holds up, simple parsimony may be a rather blunt instrument to use in character evolution.

Hence ideas of relationships between seed plants remain somewhat in limbo (Taylor et al. 2009). In particular, it is unclear which seed plant fossils are stem-group angiosperms. These plants will, of course, be largely gymnospermous in their morphologies, some kind of seed fern, regardless of whether extant gymnosperms are monophyletic or paraphyletic of where Ginkgo, etc., go on the tree. The most promising candidates for angiosperm relatives may include Corystospermales (Pteruchus, Ktalenia, etc. - see the mostly male theory of flower evolution [Frohlich & Parker 2000]), Bennettitales (but recent work is suggesting that these are perhaps close to Gnetales), and Caytoniales (poorly known in the younger Mesozoic). I have set up the synapomorphy scheme below in a way that I hope allows one to understand possible synapomorphies whatever the real relationships are, however, I think it unlikely that Gnetales have anything immediately to do with flowering plants - and this will be true of seed ferns like Bennettitales that have been associated with them if the relationships discussed above hold. If extant gymnosperms are monophyletic, as is seeming more likely, then loss of sperm flagellae and the associated development of a pollen tube, etc., will represent parallelisms between extant Pinales and the clade ancestral to angiosperms, although a synapomiorphy for a clade that includes all extant seed plants could still be the presence of pollen (see Poort et al. 1996 for definitions, etc.).

Somewhere around Pinales seems the most likely position for Gnetales. Although nearly all Pinales have megasporangiate strobili with spirally-arranged ovuliferous scales or modifications of them, Gnetales are distinctive in having strobili with decussating bracts (Magallón & Sanderson 2002). Nevertheless, there are some morphological similarities between the Pinales and Gnetales, and within the former, perhaps particularly with Pinaceae. The binucleate sperm cells, basic proembryo structure, development of polyembryony, etc., of Ephedra agree with Pinales in general and perhaps Pinaceae in particular. Some Pinus species have stomata in which the subsidiary cells are produced from the same initial that gives rise to the guard cells (Gifford & Foster 1989; see also Mundry & Stützel 2004: the stomata are mesogenous), as in Gnetales. Interestingly, perhaps, strobili with both micro- and megasporangia are common as abnormalities in Pinales, and the megasporangia are borne above the microsporangia (Chamberlain 1935) - perhaps a similarity with the angiosperm flower? (Rudall et al. 2011a) - or perhaps just an analogy.

The challenge is to think of how the heterosporangiate strobilus with short internodes that is the angiosperm flower might have evolved from the separate male and female strobili of most gymnosperms. Baum and Hileman (2006) also proposed a developmental genetic model for the evolution of the flower, which may help in the interpretation in the significance of particular fossils, and this and other models are summarized by Bateman et al. (2011b). In the "mostly male" theory of Frohlich and Parker (2000) it was suggested that the heterosporangiate strobilus evolved from what was a male strobilus, but with the development of ectopic ovules. LFY/FLO genes were likely to be associated with male reproductive structures, they suggested, and NLY genes with female. However, recent work on the expression of LEAFY/FLORICAULA and NEEDLY orthologs suggest that both genes are expressed in early-stage primordia, but LFY/FLO genes are then expressed in ovules and microsporangia while NLY genes are expressed in the ovuliferous scale, aril, microsporophylls, etc. Thus both genes are expressed in both male and female cones, which is not consistent with the "mostly male" theory (Vásquez-Lobo et al. 2007 and references; Moyroud et al. 2010; Tavares et al. 2010; see also Bateman et al. 2011b).

Chanderbali et al. (2010) found that genes involved in microsporangium, etc., production in at least some gymnosperms are also expressed in the perianth of angiosperms; only a few genes involved in ovular expression are also expressed in the perianth. S. Kim et al. (2004b) estimate the split that gave rise to the paleo AP3 and PI genes as somewhere between (297-)290-230(-213) million years before present, well before the origin of crown angiosperms. Jiao et al. (2011) have recently suggested there was a whole genome duplication in these stem group plants - estimates of peak ages are (245-)236, 234(-225) million years ago, although the spread of ages is 275-150 million year; the former set of ages are in the first half of the Triassic.

Rudall and Bateman (2010) had earlier suggested that the morphology of crown group conifers, being highly derived, may be of little help in thinking about that of the ancestors of angiosperms. This turns the problem over to interpreting fossil remains, and here it seems to me that there has been little progress over the last fifty years. It has been suggested that there are similarities between the ovules of some Magnoliaceae and the cupules of Caytonia (e.g. Umeda et al. 1994), but these are probably superficial. Features like the lobing of the integuments which induced this comparison seem to have little systematic significance, certainly there is no suggestion that the two integuments are of fundamentally different nature (Endress & Igersheim 2000; Endress 2005c). Indeed, the ovule-bearing structures of Caytonia could be linked with the carpels of extant angiosperms by invoking appropriate morphological gymnastics (Doyle 2006 for literature), and Doyle (e.g. 2008b; Doyle & Donoghue 1986a, b, 1992; Doyle & Endress 2010) has long tried to ascertain what fossils might be close to angiosperms. It has also been suggested (Wang et al. 2007) that the early Jurassic Schmeissneria, previously placed in Ginkgoales, is angiospermous, having closed carpels, it is difficult to see this in the fossils; of course, angiospermy itself may have arisen more than once. In general, no pre-Cretaceous fossils that purport to be angiospermous are convincing (but see Wang 2010 for detailed descriptions of a number of possible pre-Cretaceous angiosperms). If columellate pollen is ancestral in angiosperms (but see above), there may be connections with the Triassic reticular-columellar Crinopolles pollen (Doyle 2001; Zavada 2007). However, Taylor and Taylor (2009) suggest there are simply not enough good data to determine the putative relatives of angiosperms.

So just which gymnosperms are close to angiosperms is unknown; no fossil gymnosperms are convincing stem group angiosperms. Furthermore, although there have been dramatic changes in gene expression during the course of evolution of land plants (e.g. Banks et al. 2011; Szövényi et al. 2010), sampling is currently too poor to know if those changes can be pegged to a particular part of the tree. Thus Szövényi et al. (2010, q.v. for many more details, ca 30% of the genome mapped) noted that a total of only ca 5% of the genes in Funaria hygrometrica were expressed uniquely in the sporophytic and gametophytic generations, but in Arabidopsis ca 5% of the genes were differentially expressed in the gametophyte alone and ca 25% in the sporophyte alone. Making the problem more interesting is the fact that gene expression in neither bryophyte generation was like that in the Arabidopsis gametophyte. Where in the tree between Funaria and Arabidopsis this shift took place is not known.

Angiosperms and Insects.

Associations between extant plants and insects may be particularly close, whether the insects are herbivores, detritivores, or pollinators. The diversification of angiosperms appears to be broadly contemporaneous with the massive diversification of many insect groups that are now more or less dependent on them, although there is some argument as to just how closely linked in time the two were. There have been suggestions that it is not so much increased diversification but reduced extinction that has characterised the evolution of insects (Labandeira & Sepkoski 1994), although this is unlikely.

Phytophagy. In general, clades of phytophagous insects are more speciose that their non-phytophagous sister groups (Mitter et al. 1988). Plant-feeding insects make up at least one quarter of all described species, and over half the beetles (Janz et al. 2006; Farrell 1998); there are well over 100,000 species of extant phytophagous beetles in some five clades that eat angiosperms. These beetles may have diversified since the early Cretaceous (Farrell 1998; see also Mayhew 2007), perhaps first on monocots and then moving on to broad-leaved angiosperms (Reid 2000). Of phytophagous beetles, about two thirds eat only one or a few species of angiosperms, i.e. they are are mono- or oligophagous. The phytophagous beetle sister taxa weevils (Curculionoidea) and leaf beetles (Chrysomeloidea) include about half of all herbivorous insects and seem to diversify in parallel with angiosperms (Farrell 1998). However, a number of not very speciose but old clades of these insect groups are found on gymnosperms, including cycads, an association that dates back to the Jurassic or earlier, and initial diversification seems to have been on those plants in the Jurassic (e.g. Labandeira et al. 1994; Farrell 1998; Mckenna et al. 2009). Weevils are a particularly diverse group with some 62,000 species described, 220,000+ species altogether; McKenna et al. (2009) suggest that there was a "massive diversification" of Curculionidae - ca 90% of all weevils - as angiosperms became floristically common some 112-93.5 million years ago. Bark beetles are weevils that are decomposers; they are less speciose in clades that returned to conifers. A major subset of these (Scolytinae, Platypodinae) includes the ambrosia beetles; parental care is almost universal here, and the weevils make tunnels in the wood and are associated with the ambrosia fungi (Ophiostoma, Ceratocystis: Ophiostomatales, ascomycetes). This association has evolved about seven times in these and related weevils and is unreversed (e.g. Farrell et al. 2001; Jordal et al. 2011). Of course, there are also very species-rich beetle clades that are neither herbivores nor decomposers (e.g. Barraclough et al. 1998).

There are at least 150,000 species of butterflies and moths (Lepidoptera) (Roe et al. 2009 for a summary), and larvae of about two thirds of these are herbivores, and most of these are mono- or oligophagous (Bernays & Chapman 1994). It seems that butterflies and moths are not sister groups, and the buttefly clade also includes a small clade made up of the mostly night-flying Hedylidae, the American moth-butterflies (Cho et al. 2011 and literature), but relationships among major groups of ditrysian insects is unclear; ditrysia also include the majority of moths, both macro- and microlepidoptera. Heikkilä et al. (2011) recently suggested the unsuspected set of relationships [Papilionidae [[Hedylidae + Hesperidiidae] other butterflies]]. Much lepidopteran diversification occured in the Tertiary (see Wahlberg et al. 2009 for references), thus although the main clades [families] with an expanded Papilionoidae diverged quickly in the early Cretaceous, the main diversification occured after the K/T boundary (Heikkilä et al. 2011).

Crown-group diversification of major angiosperm-associated weevil clades seems to have been underway by the Aptian 125-112 million years ago with diversification continuing through the Cretaceous into the Tertiary (McKenna et al. 2009). "Basal" Curculionidae show strong associations with monocots, but the significance of this is unclear; there is currently little evidence that monocots were either particularly abundant or ecologically successful early angiosperms (Crane et al. 1995; Friis et al. 2004; J. A. Doyle et al. 2008; cf. McKenna et al. 2009). Crown group diversification of termites and ants had also begun (Grimaldi & Agosti), and honeydew and scale insects, important in the later evolution of ants, were also abundant then, even if ants seem to be much less so (Grimaldi & Agosti 2000). The angiosperm-feeding Nymphalidae (butterflies) began diversifying in the Late Cretaceous some 90 million years ago, although much diversity is in clades that have originated in the Tertiary (Wahlberg et al. 2009); diversification seems to have occurred as butterflies moved on to new groups of food plants and specialized on them (Janz et al. 2006; Fordyce 2010).

Another set of associations between plant and insect result in the distinctive outgrowths of the plant evident as galls (see Redfern 2011 for a good introduction). Associations here may be close. Estimates of the numbers of gall-forming insects depend on estimates of the numbers of flowering plants (for which, see Joppa et al. 2010 for literature) because of this specificity of many gall insect/plant association - and so estimates range from (21,000-)132,930(-211,000) species. Cecidomyiidae (dipterans, gall midges), the largest group of galling insects and comprising perhaps 25% of galling insects in North America (Abrahamson & Weiss 1997), are worldwide in distribution but show no particular patterns of host associations (Yukawa & Rohfritsch 2005: but see below for geography), while Cynipidae (hymenopterans, gall wasps) may comprise as many as 50% of gallers and are north temperate. Smaller groups include psyllids (jumping plant lice, hemipterans) which are particularly common in Australia (Fernandes & Price 1991; Crespi et al. 2004; Espiritó-Santo & Fernandes 2007; Raman et al. 2005), while some aphids and other insects are also gallers. In general, gall-inducing insects are commonest on sclerophyllous plants growing on poor soils in warm climates between 25 and 45^o N and S, or perhaps more generally in species-rich communities, whether dry or wet, but not necessarily in tropical climates (Price et al. 1998; Yukawa & Rohfritsch 2005; see Price et al. 1987 for galling in an adaptive context). Other organisms may be directly involved, too, such as fungi in cecidomyid ambrosia galls (Rohritsch 2009 and references), and a whole network of parasites, hyperparasites and predators is all more or less dependent on the gall larvae (Redfern 2011).

Plants have evolved mechanical and especially chemical defences against herbivory, and some insects have evolved ways of tolerating these defences - or even eat only plants with particular defences that they then coopt for their own defence (e.g. Termonia et al. 2001 for chrysomelid leaf beetles). It is well known that what attracts an egg-depositing insect to one plant and prevents it laying eggs on another is often some aspect of plant chemistry (see Bernays & Chapman 1995 and Fernandez & Hilker 2007 [Chrysomelidae] for host plant selection). Some plant and insect groups do seem to be rather closely associated, showing loose coevolution (Ehrlich & Raven 1964 for an early statement of the concept; see/cf. e.g. Janzen 1980; Kato et al. 2010; Fordyce 2010); these are discussed further after individual orders and families. Futuyma and Agrawal (2009) and Janz (2011) for an evaluation of the concept - wjich includes everything from cospeciation to key innovations and adaptive radiation - and literature. In general, more related plants do have more similar animals eating them (Weiblen et al. 2006; see Futuyma & 2009 for literature), simply because they will tend to taste similarly. In general, host plant relationships may be conserved, even if there is little evidence for strictly parallel diversification (Winkler & Mitter 2008). In connection with Ehrlich and Raven's hypothesis, host switching might be expected to be associated with radiation of the insect on to a new host (Fordyce 2010). At the same time, the similarity between e.g. the monoterpenes produced by both plant and insect may be the result of the plant tapping in to pre-existing insect physiology and communication channels (Schiestl 2010).

Herbivorous insects may sequester secondary metabolites from the plant in the larva and/or adult stages, ensuring some measure of protection by so doing and often having a warning colouration (i.e., they are aposematic), or they may use plant metabolites for pheromones to attract mates, or these metabolites may simply act as oviposition cues, not being otherwise utilised by the insect (Brower & Brower 1964 on butterflies; Nishida 2002 for a review). Protective metabolites may be found in laticiferous structures, or they may be translocated via the vascular tissue, or there may be other specialised tissues involved; herbivorous insects that eat plants with such defences may show distinctive veing-cutting behaviours which stops the supply of the distateful metabolite to plant tissue and enables the insects to eat it (see e.g. Dussourd & Eisner 1987; McCloud et al. 1995; Becerra et al. 2001; Dussourd 2009). In any one local area, related plants may shows greater than expected diversity of traits involved in herbivore defence (e.g. Becerra 2007; Becerra et al. 2009; Kursar et al. 2009).

Within herbivores, there is a general decrease in host specificity both in temperate and tropical regions and following the same general sequence, granivores > leaf miners > fructivores > leaf chewers = sap suckers > wood eaters > root feeders (Novotny & Basset 2005), while specialization in weevil-plant associations is similar - fruit and seed > wood > root and stem eaters (McKenna et al. 2009). Indeed, how insect larvae feed, i.e., whether they are internal feeders like stem borers and whether they can tolerate raphides, or latex, etc., may be more conserved than associations between larvae and particular groups of plants or other types of feeding behaviours (e.g. Powell 1980; Peigler 1986; Powell et al. 1999 and references for associations with latex-containing plants and Konno 2011 for some chemistry; Farrell & Sequiera 2001; Lopez-Vaamonde et al. 2003, 2006). Furthermore, phylogenetic conservatism may be greater in groups in which the adults tend to remain close to plants in which they grew up, as with beetles, compared to the situation where the adult may fly away, as in many lepidoptera (Berenbaum & Passoa 1999). In general, ectophagous insect groups are more diverse than their endophagous sister taxa, phytophagous taxa more diverse than non-phytophagous, and taxa that eat angiosperms are more speciose compared to those that eat other plants (Winkler & Mitter 2008). In general, herbivory is greater in tropical than in temperate forests (Adams et al. 2011 and references).

Since some herbivorous insects effectively track plant secondary metabolites, they are found on whatever plant has a particular metabolite, and this may be independent of the phylogeny of the plant groups concerned (e.g. Winkler et al. 2009). Glucosinolates and some alkaloids are examples; glucosinolates are found in both Putranjivaceae and Brassicales, as are the pierid butterflies that are attracted to glucosinolates, while swallowtail butterflies are found on Rutaceae and Lauraceae, the two plant groups having similar alkaloids. Secondary metabolites involved in plant-insect relationships - and secondary metabolites in general - seem to have a very scattered distributions and show considerable homoplasy. It has been suggested that the genes involved in their production are only sporadically expressed, but are retained in the genome (e.g. Grayer et al. 1999; Albach et al. 2005c); the fact that endophytes on occasion synthesize some of these compounds (see below) further complicates the issue.

To sumarize: the evolution of insect/plant associations on plant diversification is in general not well understood (Futuyma and Agrawal (2009: also other papers in Proc. National Acad. Sci. U.S.A. 106()). In some cases, diversification of plants can be linked to the development opf particular defences, but this does not happen in any simple fashion; the mechanism by which insect diversification increases when feeding on angiosperms is unclear (Janz 2011. We need to know more about both the timing of diversification and patterns of phylogenetic relationships in both groups, but there is uncertainty in both areas. Was the radiation of ants the more or less immediate result of the radiation of angiosperms (Moreau et al. 2006), or did ant-plant relationships like myrmechory develop only when ants became abundant (e.g. Grimaldi & Agosti 2000; Dunn et al. 2007; Pie & Tschá 2009; Lengyel et al. 2010)? Indeed, the answer may be "both" (see below). Is monocot feeding largely restricted to a single clade of beetles, or are those beetles in two immediately unrelated clades (cf. Wilf et al. 2000 and Gómez-Zurita et al. 2007)? Furthermore, the sheer complexity of the matrix of defensive compounds inside the plant make simple explanations of the evolutionary dynamics of plant-insect relationships difficult, idiosyncracy being utterly central to the nature of chemical coevolution (Berenbaum & Zangerl 2008: p. 806); this problem can only be exacerbated by the difficulty of understanding ecological relationships over time. Finally, the whole issue can be almost turned on its head if one thinks of the plant "reacting" to pre-existing signalling and physiological circuitry of the insect (Schiestl 2010).

Pollinators. Insect pollination is not restricted to angiosperms. It probably occurred in Mesozoic gymnosperms as well, the most likely groups of pollinators being beetles, neuroptera, mecopterids (scorpion flies, mecoptera - perhaps) and true flies (the evolution of bee flies may be early Jurassic - see Wiegmann et al. 2011), and other groups may also have been involved (Labandeira 1998, 2010; Labandeira et al. 2007; Ren et al. 2009). Beetle, fly and moth pollination is known in extant gymnosperms (Kato & Inoue 1994; Schneider et al. 2002; Oberprieler 2004; Labandeira 2005: note that dioecy is common) and of course also in angiosperms. There pollination by such insects predominates in taxa in the ANITA grade, Annonaceae (magnoliids), Araceae, etc. - and these appeared quite early in the Cretaceous (but see above). Bees are not prominent pollinators in extant members of these clades. Beetles, etc., are attracted to flowers lacking definite symmetry signals (Leppik 1957); of course, other factors such as scent are also involved (see Barth 1985 for a very readable summary of the interrelationships between insects and flowers).

The common ancestor of all angiosperms may well have had a small, rather generalised flower (for suggestions as to what this flower might look like, see e.g. Crane et al. 1995; Doyle & Donoghue `1986a; Endress 2001a; Weberling 2007; Doyle 2008b; Doyle & Endress 200, 2010, 2011), and was likely to have been pollinated by insects (Hu et al. 2008). Dry stigmas and protogyny were probably the common condition (e.g. Sage et al. 2009; Endress 2010a). Many Cretaceous angiosperms had small flowers, and a remarkable number of these had inferior ovaries (e.g. Crane et al. 1995; Friis et al. 1999); it is likely that they were aggregated into inflorescences to attract pollinators (Friis et al. 2006b). Pollen seems initially to have been produced in rather low quantities, yet at first, however, it is probable that it was the stamens in particular that were a source of food, and pollen has been found in coprolites. Initially, nectar from specialized nectaries was unlikely to have been a common reward, although there appear to be "food bodies" in flowers from 115-100 million years ago of the Lower Cretaceous Burmese amber (Santiago-Blay et al. 2005). Finally, carpels of early angiosperms are likely to have had few ovules and stamens with few pollen grains (e.g. Crepet et al. 1991; Dilcher 2000; Friis et al. 2006b), as in extant members of the ANITA grade, and these often have diploid endosperm (see below). Specht and Bartlett (2009), Endress (2010a), and others have surveyed the floral morphology and biology of basal angiosperms.

By the mid-Cretaceous pollen became more abundant and is more often found in clumps, suggesting that the pollinators, perhaps bees, were more specialized (Hu et al. 2008), although there is no obvious signal of an increase in protein content of the pollen of extant representatives of angiosperm clades that had evolved by then that might suggest a shift in pollinating agent to bees (see Roulston et al. 2000). However, diversification of eudicots in the Cenomanian 110-90 million years ago has been linked with the evolution of bees that also occurred around that time; certainly, angiosperm flowers from this period showed a variety of quite specialized zoophilous morphologies, and nectar secretion became common (Hu et al 2008). The initial divergence within butterflies (Papilionoidea) may have been around 110-100 million years ago (Heikkilä et al. 2011). The triaperturate pollen of these eudicots may germinate faster yet at the same time be less viable than uniaperturate pollen (e.g. Dajoz et al. 1991; Furness & Rudall 2004), but overall rather little is known about many aspects of the functional evolution of pollen and how it relates to morphology (see e.g. Roulston et al. 2000; Fernández et al. 2009). Nectaries are known from plant fossils of Cenomanian age (mid Cretaceous), while sympetaly, inferior ovaries, and monosymmetry (evidence for the latter is indirect - seeds assignable to Zingiberales - Rodríguez-de la Rosa & Cevallos-Ferriz 1994) appear in the Late Cretaceous (Friis 1985; van Bergen & Collinson 1999; Friis et al. 2003a). With the advent of the core eudicots in particular nectar produced by receptacular nectaries may have become a major reward for pollinators (Friis et al. 2006b), but Proteaceae and Sabiaceae (both Proteales) may also have receptacular nectaries. Note that some fossils of Fagales and Platanaceae (also Proteales), groups that are now predominantly wind-pollinated, have nectaries. Septal nectaries may be an apomorphy for monocots, being scattered through that clade, and are certainly found in some Alismatales, so that their evolution is also probably old. Nectaries of various other types are scattered through angiosperms other than monocots and core eudicots.

There is considerable debate over the existence of pollination syndromes and what exactly pollinators might see and respond to (Fenster et al. 2004; Waser & Ollerton 2006; Raguso 2008; Ollerton et al. 2009a); Rodríguez et al. (2004) and Horridge (2009) discuss the honey bee's point of view, the former, thinking about mionosymmetry in particular, the latter more generally. Turning the question around, authors like Schaefer and Ruxton (2009, 2010), Schiestl (2010) and Schiestl et al. (2010) think about pollination as the exploitation of pre-existing perceptual/sensoty biases of the pollinator by the plant.

The evolution of bees is of particular importance, given the close involvement of many of them with angiosperm pollination (for bees and pollen, see Westerkamp 1996; for an account of all bee groups, see Michener 2007). The basic phylogenetic structure of the bees is [Dasypodaidae [[Meganomiidae + Melittidae s. str.] [Andrenidae [Halictidae [Stenotritidae [Colletidae [Apidae + Megachilidae]]]]]]], i.e. the mellitids s.l. (Dasypodaidae, Meganomiidae and Melittidae) are paraphyletic. The origin of Apidae has been dated to some 135-120 million years ago (Grimaldi & Engel 2005), and initial diversifification in the early to mid Cretaceous seems to have occured in association with the evolution of angiosperms (see also Engel 2000; Michez et al. 2009; Almeida & Danforth 2009; cf. Renner & Schaefer 2010 - [[Apidae + Megachilidae] [Andrenidae [Halictidae [Stenotritidae + Colletidae]]]]). Within Apidae, the corbiculate bees, relationships are [[Euglossini + Apini] [Meliponini + Bombini]] (Cameron 2004: trees based on morphology and behaviour conflict with those based on molecular data, the latter providing the relationships discussed here; see also Cardinal et al. 2010). For the phylogeny of Colletidae, see Almeida and Danforth (2009).

These general relationships are consistent with the appearance of bees in the fossil record. The earliest fossil bee, perhaps sister to other Apoidea, was found in amber of Upper Albian age (ca 100 million years old) from Burma; it is interesting that it is also quite small, being ca 5 mm long (Poinar & Danforth 2006, but are there problems with dating?), in line with the often rather small size of Cretaceous flowers. A younger fossil from the New Jersey amber of the Late Cretaceous (96-74 million years ago) was even assigned to the extant genus Trigona, a highly derived eusocial stingless bee (Apidae - Meliponini: Michener & Grimaldi 1988); both its age (now estimated at 70-65 million years) and its relationships (it is now placed in Cretotrigona) have been re-evaluated (Engel 2000). Apidae and Megachilidae are derived, long-tongued bees, and both are known from Baltic amber of Eocene age (Danforth et al. 2006 and references). However, these are simply minimum ages, and with the revision of ages for the origin of angiosperms (e.g. S. A. Smith et al. 2010), it has recently been suggested that the age of stem-group bees is some (182-)149(-119) million years ago, with crown-group Megachilidae, a major clade including the leaf-cutting bees, starting to diversify (154-)126(-100) million years ago (HPD: Litman et al. 2011).

Angiosperms and Fungi.

Close relationships between seed plants and fungi, whether as mycorrhizae or endophytes, are ubiquitous; Brundrett (2009, see also 2008 for updated online resource) provides a comprehensive survey against which information on mycorrhizal associations mentioned here has been checked. The evolution and ecological significance of mycorrhizae in particular have been widely discussed (see Malloch et al. 1980; papers in Allen 1992; Read et al. 2000; Landis et al. 2002; Egger & Hibbett 2004; Taylor et al. 2009, etc.), as has the morphology of the plant/fungus interface (e.g. Peterson & Massicotte 2004) and how the fungus uses the 10% or more of photosynthesate that it gets from the plant (Leake et al. 2004). Aquatic plants, hardly surprisingly, often lack mycorrhizae (see Safir 1987 and Radhika & Rodrigues 2007 and references for records; de Marins et al. 2009), but the frequent absence of mycorrhizae in Caryophyllales, Proteales, etc., is somewhat surprising; furthermore, epiphytic taxa are not often mycorrhizal (Janos 1993; see other papers in Mycorrhiza 4(1). 1994). Overall some 18% of flowering plants may lack mycorrhizae, and a further 12% are facultatively mycorrhizal (Molina et al. 1992). Note that the development of mycorrhizae is intimately but variously linked with root architecture - root thicknesss, root branching, the development of root hairs (Baylis 1975; Schweiger et al. 1995 and references) - as well as the nutrient status of the soil, and benefits accruing to the partners may be similarly labile (Newsham et al. 1995). Thus the development of root hairs or of a vesicular arbuscular mycorrhizal association may be alternative ways for the plant to obtain phosphorous in situations when it is not readily available (Schweiger et al. 1995 and references). Being mycorrhizal is not a simple either/or matter, and the one species of plant may have a variety of associations with fungi, as well as having distinctive root morphologies like dauciform roots which are also supposed to be responses of the plant to nutient-poor conditions (e.g. Gao & Yang 2010; Schweiger et al. 1995). Finally, a single individual of a mycorrhizal fungus may form associations with two or more species of plants simultaneously, so forming complex mycorrhizal networks (Simard & Durall 2004 for literature).

Vesicular-arbuscular mycorrhizae (endomycorrhizae) are very widespread. In such mycorrhizae the aseptate hyphae are intracellular, often forming vesicles or branching structures (the arbuscules) within the cells. There is substantial variation both in the morphological details of the fungus-plant association (e.g. Smith & Smith 1997; Peterson & Massicotte 2004) and in the proportion of fungal biomass inside and outside the plant (Maherali & Klironomos 2007). The fungi involved are Glomeromycota (Schüßler et al. 2001), and about 70% of seed plants are endomycorrhizal (80% plant species, 92% plant families - Blackwell 2011). This association is probably of very long standing indeed, and may be a feature of the common ancestor of all land plants (see also Baylis 1975; Redecker et al. 2000b; Kottke & Nebel 2005; Duckett et al. 2006b; Ligrone et al. 2007; Wang et al. 2010). There may be bacteria obligately associated with the fungus (Bonfante & Genre 2010). Sexual reproduction in the fungus is at most exceedingly uncommon, but the spores are multinucleate, the nuclei not having any onviousd immediate common ancestor, so the unit of selection may be the individual nucleus (Jany & Pawlowska 2010). Overall, Glomeromycota are not very speciose, containing perhaps some 300 species, and a number of these species have relatively limited distributions (Öpik et al. 2010).

In such associations, nutrient uptake by the plant - especially of phosphorous, although recent work suggests that nitrogen is also involved - is increased, and water uptake is improved (Allen 1992; Govindarajulu et al. 2005; Leigh et al. 2009 and references; Tian et al. 2010; Bonfante & Genre 2010), and details of how the fungus gets carbohydrate from the plant are becoming clearer (Helber et al. 2011). It is fair to say that we still know rather little of the details of the plant-mycorrhizal interactions (Whitfield 2007), however, a number of genes involved in the establishment of vesicular-arbuscular mycorrhizal associations are the same as those involved in nodulation in the nitrogen-fixing clade (Markmann et al. 2008; Yano et al. 2008) and the invasion of plant tissue may be similar to what goes on in parasitism (Bonfante & Genre 2010). Initial attraction of the fungus to the plant, and also hyphal branching, may be mediated by strigolactones secreted by the root (Akiyama 2010, and references).

There are many fewer species of ectomycorrhizal plants, but they often dominate the communities in which they occur. They are particularly prominent in (cool) temperate areas (see below); interestingly, a number of ectomycorrhizal plants are mast-fruiters (Newberry et al. 2006). Ectomycorrhizae form a Hartig net of hyphae investing rootlets and penetrating between the cortical cells; the hyphae are septate and are not intracellular - with the exceptions of Ericaceae and Orchidaceae (q.v.). There are other variants, such as tuberculate mycorrhizae, clusters of roots surrounded by hyphae (Smith & Pfister 2009). Basidiomycetes are frequent in such associations, but Pezizales (ascomycetes) are also quite common (Tedersoo et al. 2006, ascomycetes with a hypogeous life style are derived from them), as in Ericaceae.

Ectomycorrhizae are commonly found in plants growing on rather extreme soils, either poor in nutrients and/or rich in organic materials, especially in tropical montane and (cool) temperate habitats, but also in the tropics (e.g. Malloch et al. 1980; Sanford & Cuevas 1996 and references). In Boletales, at least, ectomycorrhizae may have evolved in the ecological context of soils low in nitrogen (and high in organic material) that are the result of the activities of brown rot fungi, a number of which are also found in Boletales (and elsewhere: Eastwood et al. 2011). Ectomycorrhizae may also obtain N and P from material as diverse as pollen and dead nematodes, and these are transferred to the plant as glutamine and ammonium, but the fungi may have lost their ability to break down plant cell walls, etc. (Read & Perez-Moreno 2003; Martin & Nehls 2010; Bonfante & Genre 2010). Ectomycorrhizae secrete low molecular weight organic compounds, including oxalate (complexing with toxic aluminium ions) and siderophores (they chelate iron); siderophores are also produced by the bacterial associates of ectomycorrhizae (Frey-Klett et al. 2007; Taylor et al. 2009), indeed, bacteria are integral to ectomycorrhizal associations, whether facilitating the establishment of the association (mycorrhiza helpers) or as integral to the functioning of the established association (Frey-Klett et al. 2007 and references). Associated bacteria have also been implicated in fixing nitrogen (Paul et al. 2007). When thinking of the biogeochemical effects of ectomycorrhizal plants, the fungi and bacteria are just as important - if not more so - as the plant itself (e.g. Landeweert et al. 2001; van Schö et al. 2008; Taylor et al 2009; Bonfante & Anca 2009).

Endophytes, fungi growing inside plants, have been broadly categorized. Clavicipitaceous endophytes, class one endophytes, occur in grasses with which they form very close associations (Schardl 2010). Other endophyte are nono-clavicipitaceous: Class two endophytes pervade all the tissues of the plant, although the fungi involved are not particularly speciose; class three endophytes are restricted to shoots and are very diverse; while class four endophytes are restricted to roots - the dark septate endophytes (DSE) (Rodriguez et al. 2009). Transmission may occur via the seed (vertical transmission), as is common in endophytic Poaceae-Poöideae in particular, or via fungal spores (horizontal transmission), more common in other endophytic associations (Arnold 2008). Endophytic fungi are probably to be found in all seed plants (Rodriguez et al. 2009; Hoffman & Arnold 2010). Indeed, it is becoming clear that the numbers of species of fungi that take part in endophytic associations with plants is very large indeed. For example, Arnold et al. (2001) found 418 morphospecies of class three endophytes in only 83 leaves of two species of tropical trees, Ouratea (Ochnaceae) and Heisteria (Erythropalaceae) (see also Bills & Polishook 1994; Frohlich & Hyde 1999; Arnold & Lutzoni 2007 and other articles in Ecology 88[3]. 2007); the whole phyllosphere, the above-ground surface of the plant, may contain several hundred species of fungi, although what they are doing is largely unclear (Jumpponen & Jones 2009).

Most endophytes are Ascomycota, and show various relationships within that clade. Indeed, fungi living inside lichens, endolichenic fungi - these are different from the mycobionts that form an association with algae and which are traditionally thought to constitute lichens - may give rise to many of these endophytic clades (Arnold et al. 2009b). However, other fungus groups are involved, and the basidiomycete Sebacinales are particularly diverse ecologically, being common endophytes (Weiß et al. 2009) as well as being involved in the distinctive mycorrhizae of Ericaceae and Orchidaceae (q.v.). Grass endophytes are derived from fungi that are insect pathogens, and some fungi are both pathogen and endophyte (Spatafora et al. 2007; Sasan & Bidochka 2012).

In the tundra habitat several plants take up substantial amounts of nitrogen as amino acids directly from the soil, but the ability to do this is not obviously correlated either with mycorrhizal status or taxonomy. Thus some Cyperaceae, which are generally thought to lack mycorrhizae (but cf. e.g. Muthukumar et al. 2004), take up nitrogen predominantly in an organic form, but other Cyperaceae take it up in an inorganic form (Raab et al. 1999). Largely ascomycetous fine endophytes are commonly found in plants from such habitats (Higgins et al. 2007), indeed, they may be more prevalent than arbuscular mycorrhizal fungi, since the prevalence of vesicular-arbuscular associations decreases with latitude (Olsson et al. 2004). However, in the widespread Kobresia pygmaea ecosystem of the Tibetan plateau, dauciform roots may be produced by Kobresia spp., and these may also be ectomycorrhizal, but some of the fungi found on the plants may also be dark septate hyphae and/or endomycorrhizae - indeeed, clear distinctions between these associations may be hard to draw (Gao & Yand 2010).

Finally, there are connections at the molecular level between the establishment of mycorrhizal associations and the development of symbioses with actinomycetes and Rhizobium that are involved in nitrogen fixation, a number of the genes involved being the same. This is discussed in more detail in the Fabales page - see also Markmann & Parniske (2009) for a recent review, Maillet et al. (2010), etc.

Indeed, the complexity of ectomycorrhizal and other fungal associations is considerable. The gametophytes of liverworts may form associations with the same fungi that form ectomycorrhizae with the flowering plant on which the liverwort is found. Examples include Marchantia/mycorrhizal fungus/Podocarpus and the mycoheterotrophic chlorophyll-less Cryptothallus/the basidiomycete Tulasnella/Pinus-Betula (Read et al. 2000; Bidartondo et al. 2003; Kottke & Nebel 2005 and references). The ascomycete Rhizoscyphus ericae is very commonly an associate of the hair roots of North Temperate Ericaceae, it can also be ectomycorrhizal on Pinus growing together with these Ericaceae (Grelet et al. 2010; see also Villarreal-Ruiz et al. 2004), and it also forms mycorrhizal associations with Jungermanniales-Schistochilaceae, leafy liverworts (Pressel et al. 2008). And, as just mentioned, mycorrhiza-plant associations often include bacteria as additional partners, whether growing on the surface of the mycelium and synthesising crucial metabolites or living within the hyphae; bacteria are involved both in the establishment and the functioning of ectomycorrhizae (Frey-Klett et al. 2007).

Complications also develop as one looks at endosymbionts of fungi associated with plants. Thus the bacterium Candidatus Glomeribacter gigasporarum (near Burkholderia) is found in the endomycorrhizal fungus Glomus (Castillo & Pawlowska 2009, 2010 and references). Such bacteria may affect the growth of the fungi, and they may be vertically transmitted like the fungi themselves (Bianciotto et al. 2003; Hoffman & Arnold 2009). Numerous bacteria (mostly Proteobacteria) are involved, and a diversity of fungi have been shown to be infected, although the relationship seems sometimes be rather casual (Hoffman & Arnold 2010). Viruses of endophytes may affect the ability of the host plant of the endophyte to grow in particular conditions (Márquez et al. 2007).

In several cases distinctive "plant" metabolites that function as animal toxins such as indolizidine (swainsonine) and ergoline alkaloids are in fact synthesized by fungal or bacterial associates of the plant (e.g. Popay & Rowan 1994; Findlay et al. 2003 and Sumarah et al. 2010 [spruce endophytes producing a variety of metabolites toxic to the easter spruce budworm]; Gunatilaka 2006; Markert et al. 2008 [Convolvulaceae]; Pryor et al. 2009 [Fabaceae]; Wink 2008); Celastraceae and especially Poaceae are also distinctive in this regard. Indeed, such substances function as if they were endogenous plant metabolites (Zhang et al. 2009; Friesen et al. 2011).

Sporadic associations between plant and fungus/microbe in both mycorrhizal and endophytic associations, and/or lateral transfer of genes, may also go some way towards understanding the rather unpredictable pattern of distribution of many secondary metabolites (Wink 2008; Lamit et al. 2009). Interestingly, secondary metabolites like terpenoids and quinolizidine alkaloids are produced more or less exclusively in mitochondria and/ot chloroplasts - i.e. in bacteria whose association with plants is of very long standing (Wink 2008). Wink (2008) also noted that enzymes apparently important in the synthesis of distinctive secondary metabolites in flowering plants were not infrequently much more widely distributed in flowering plants than the distribution of those metabolites would lead one to suspect.

Cretaceous History

Age estimates of crown angiosperms vary considerably, but are mostly in the range (130-)140-180(-210) million years before present (e.g. Doyle 2001; Sanderson & Doyle 2001; Wikström et al. 2001; Aoki et al. 2004; Sanderson et al. 2004; Bell et al. 2005; Leebens-Mack et al. 2005; Moore et al. 2007; Soltis et al. 2008: a variety of estimates; Moore et al. 2010: 95% highest posterior density); Bell et al. (2010) suggest ages of (199-)183(-167) or (154-)147(-141) million years depending on the method used. Estimates based on molecular data tend to be substantially older than others, Magallón (2008 and references) noting times of 182-158 million years before present for the basal split within angiosperms, i.e. Lower to Middle Jurassic, or an even older age (Magallón & Castillo 2009), with eudicots appearing in the uppermost Jurassic. Recently Smith et al. (2010) have suggested that crown angiosperms are (257-)217(-182) million years (with eudicot calibration) to (270-)228(-193) million years old (without: also note divergence time estimates in Table S3 of Smith et al. 2010), and similarly great ages of 275-215.6 million years are also suggested by Magallón (2010) and of (240-)205(-175) million by Davies et al. (2011: fossil calibrations). On the other hand, when we look at the plant fossil record, we find that unambiguous angiosperm fossils from before the Lower Cretaceous are at best few, indeed, the oldest generally-agreed remains of the clade are pollen from the Cretaceous Valanginian-Hauterivian only 141-132 million years before present. However, if columellate pollen is ancestral in angiosperms, there may be connections with the Triassic reticular-columellar Crinopolles pollen type (Doyle 2001; Zavada 2007).

If an earlier age for crown angiosperm origin - some 270 to 175 million years ago, so certainly Jurassic, if not earlier - of Smith et al. (2010, see also Sun et al. 2011; Davies et al. 2011) are correct, we are faced with a series of questions. We now have to rethink the ecological context of the evolution of angiosperms and of the insects associated with them - and then understand how angiosperms persisted as a presumably not very diverse clade for 50 million years or more. We are not immediately any closer to understanding which pteridosperms are close to angiosperms, but may need to think of a different set of gymnosperm reproductive structures to link with the angiosperm flower (e.g. Doyle 2008b; Specht & Bartlett 2009; Doyle & Endress 2010 for literature).

Dating of angiosperm evolution and diversification is in general very much up in the air. Even leaving aside the pre-Cretaceous dates for the origin of crown-group angiosperms, there are conflicting narratives for the evolution of angiosperms through the Cretaceous and into the Tertiary. During the Cretaceous, some suggest that angiosperm diversification was extensive, with angiosperms achieving ecological dominance by the end of the period; others suggest that tropical rainforest as we know it had not yet developed at the end of the Cretaceous, and that angiosperms were ecologically and physiologically rather unlike many of their Tertiary successors. As to events at the Cretaceous-Tertiary (K/T) boundary, some suggest that the bolide impact had a major effect only on North American plants, with only a muted more global effect; others suggest there were major changes in vegetation structure and composition in the early Tertiary. The same tensions are evident when in the literature on the evolution of mammals, birds, insects, etc.

The most comprehensive review of Cretaceous angiosperm history is that of Friis et al. (2006a; see also 2011), on which this section draws heavily (see also Dilcher 2010; Taylor 2010 [focus on genes possibly involved]; also Friis et al. 2010a for Early Cretaceous history). Diversification of angiosperms was well under way by 137 million years before present as judged by pollen remains, the only angiosperm fossils from that period, but there seems to be a gap of 10-30 million years or more before crown group diversification started (e.g. Feild & Arens 2005). There is a wealth of fossil material from slightly later in the Early Cretaceous, for instance, there are some 140-150 taxa from the Barremian-Aptian ca 125 million years before present in Portugal alone (e.g. Friis et al. 1999, 2000a, 2010b). All in all, a remarkably diverse flora, although recent work suggests that a somewhat younger age for at least some of this material, perhaps Albian and ca 112 million years ago, is more likely (Heimhofer et al. 2005, 2007). Although practically none of these fossils can be assigned to extant families, 85% of them represent plants of the magnoliid type or are somewhat monocot-like (Friis et al. 1997a, 1999, 2001; Heimhofer et al. 2007 - see also Doyle et al. 2008 for an evaluation of early fossils putatively of monocot origin; Friis et al. 2010a). Doyle (2001) noted the abundance of families with ascidiate carpels and exotestal seeds in these floras - and in extant members of the ANITA grade (Amborellales, Nymphaeales, Austrobaileyales) and Chloranthaceae. Stamens are often wedge-shaped and with a massive apex, and they open by laterally-hinged valves (e.g. Crepet & Nixon 1996; Endress 2008c and references). Quite "derived" features may be evident in these early fossils. Thus in Sinocarpus, from the Barremian-Aptian 139-122 million years ago, the carpels were apparently connate at the base (Leng & Friis 2003).

Moore et al. (2007) suggest rather older dates for a rapid separation of the Chloranthales, magnoliid, monocot, eudicot and Ceratophyllales clades, i.e. some time between 148.6-135.5 million years ago (see also Sun et al. 2011). Other dates are somewhat younger. Fossils assignable to Chloranthaceae are known from the late Barremian ca 130 million years ago onwards, with some fossils being very like the extant Hedyosmum (e.g. Crepet & Nixon 1996; Friis et al. 2006b for references). Magnoliids also diversify early, although somewhat later for the most part (Friis et al. 1997a, 2006b for reviews); Lauraceae are prominent, and include the somewhat younger Early Cenomanian Mauldinia (Drinnan et al. 1990). This has distinctive condensed monosymmetric inflorescences quite unlike those of other Lauraceae, yet its flowers are very like those of extant members of the family. Perhaps more remarkably, fossils ascribed to Sarraceniaceae (asterids, Ericales) have been described from deposits about the same age as those in which Archaefructus is found (Li 2005); this does seem something of a stretch, whiel Wang (2010) inclines to the idea that Caryophyllales represent a very ancient clade. Be all that as it may, pollen data suggests that monocot/magnoliids diversified in the early Aptian-mid Albian 125-105 million years ago (Heimhofer et al. 2005; Hochuli et al. 2006). Doyle and Endress (2010) should be consulted for the phylogenetic placement of a number of mostly magnoliid and Anita-grade Cretaceous fossils, albeit the constraint tree that they used has a rather different topology that that of the main tree here.

There was latitudinal spread of angiosperms - and a similar spread of increasing density and abundance - over a period of about 49 my from their initial appearance in more tropical environments (Axelrod 1959; see Wing & Boucher 1998; Hemihofer et al. 2005 for further references), spreading polewards in the early Late Cretaceous. The appearance of a significant number of eudicots and their replacement of free-sporing plants, i.e., not including conifers (see e.g. Wing & Boucher 1998: cycads may have declined, but see below) occurred in North America in the Albian-Turonian, ca 100 million years ago, although again slightly earlier at lower latitudes, that is, palaeolatitudes S of 30 N (e.g. Crane & Lidgard 1989, 2000; Lupia et al. 1999). Yet even areas where conifers remained common seem to have become more restricted, and ecological factors such as slow seedling growth and details of leaf construction, etc., can be adduced to explain this change (e.g. Bond 1989); broad-leaved podocarps seem to have begun diversifying in the Southern Hemisphere in the early Tertiary, but narrow-leaved podocarps have declined subsequently (Biffin & Lowe 2011; Biffin et al. 2011). The decline of cycads and Bennettitales (cycadophytes, an ecological grouping) may be linked with the contemporaneous decline in herbivorous stegosaurian dinosaurs, but there is no support for any even loose co-evolutionary relationships between early angiosperms and dinosaurs (Butler et al. 2009 and references). Indeed, the recent suggestion that diversification of extant Cycadalean clades is a mere ca 12 million years ago (Nagalingum et al. 2011; see also Crisp & Weston 2011), means that we have to rethink the reasons for their decline and subsequent diversification. Diversification of polypod ferns, perhaps associated with the evolution of a distinctive new photosystem, began in the Cretaceous (Kawai et al. 2003; Schuettpelz & Pryer 2009). A long-term warming trend from the early Aptian culminated in the Cenomanian-Turonian thermal maximum ca 99 million years ago (Heimhofer et al. 2005).

Early angiosperms may have been rather small (Friis et al. 2010b) tropical trees that tolerated shady, humid and disturbed conditions, or other environments, but also with a measure of disturbance (e.g. Wing & Boucher 1998; Feild 2005; Feild & Arens 2005; Berendse & Scheffer 2009 for a summary). Seed size of early angiosperms is likely to have been substantially smaller than that of most gymnosperms, extant or fossil, perhaps because the evolution of the carpel allowed control both over fertilization and allocation of resources to the seed (Lord & Westoby 2012 and references). The climate in the late Barremian to Aptian seems to have been notably unstable, which perhaps favoured plants like the angiosperms with their relatively short reproductive cycles (Williams 2008, 2009). Flowers, too, seem to have been rather small, although most seem to have been insect-pollinated (e.g. Crane et al. 1995; Friis & Crepet 1987; Friis & Endress 1990; Friis et al. 2000, 2006b, 2010b; Hu et al. 2008 for literature).

Labandeira (2010 and references) suggested that many pollinating clades of Hymenoptera, Diptera, and Lepidoptera originated around Late Barremian - end Albian some 125-100 million years ago - however, if revisions of the age of orgin of crown-group angiosperms and of bees hold (Smith et al. 2010; Litman et al. 2011), such scenarios will have to be revisited. Thermogenic (beetle) pollination mechanisms occur in some extant members of many "basal" lineages, including Araceae, although not Laurales, Amborellales, and Acorales (Thien et al. 2000; see also Seymour et al. 2003 for functional explanation of thermogenesis). Beetles are common pollinators in other groups, including in Calycanthaceae, but bees are not conspicuous. Although nectaries are uncommon, bee larvae (for example) obtain their fat from pollenkitt (Renner 2010) which is produced by the degeneration of the tapetum and is rich in plastid-derived lipids. Pollen was probably the main reward for those early pollinators that were bees; of course, few Lepidoptera or Diptera are pollen-eaters... Interestingly, evidence suggests that early bees were oligolectic, not polylectic (Litman et al. 2011 and references), so they would be restricted to one or a few species of flower. Disseminules were also small, the volume of the seed contents - probably mostly endosperm reserve - being a mere 2-3 mm³, and they were also quite often fleshy (Eriksson et al. 2000a, esp. b).

Even up to the later Cretaceous (Cenomanian), the diversity of fossils that probably belong to the ANITA-magnoliid grade is quite considerable (e.g. Kvacek & Friis 2010; Friis & Pedersen 2011). Many of these older plant fossils have very distinctive character combinations. For example, Archaefructus was probably an aquatic herb and lived in the Barremian-Aptian at least 124 million years ago (Sun et al. 2002). It has been interpreted as having perfect flowers that are unlike those of extant angiosperms - there is no perianth, the receptacle is very elongated, the stamens are paired, and the carpels are conduplicate. However, its flowers have also been interpreted as being inflorescences, the paired stamens then representing staminate flowers that have two stamens but no other structures (see Zhou et al. 2003; Friis et al. 2003b; Ji et al. 2004; Doyle & Endress 2007; also Crepet et al. 2004 for a critical analysis of this and other early fossil angiosperms). In any event, Archaefructus has a very distinctive floral morphology; it is unlikely to be sister to all extant angiosperms (cf. Sun et al. 2001; Crepet et al. 2004), and recent morphological work suggests that it could be a member of Nymphaeales (Doyle & Endress 2007, 2010a; Doyle 2008b). Some fossils - and perhaps Archaefructus itself - may represent quite distinct but now extinct clades (von Balthazar et al. 2008). Thus Hyrcantha, also more or less aquatic, has been described from Barremian-Aptian deposits in China (Dilcher et al. 2007); it has leaves with a sheathing stipule (i.e. the stipules are ochreate) and partly connate carpels with apparent resin bodies at their apices. Interpretation of fossil woods in the context of palaeoclimatic indiciators is also difficult (Wheeler & Baas 1993). Quite a diversity of strange-looking putative angiosperms have been discovered in northeastern China, and although the identities of a number are disputed (Sun et al. 2006), new fossil finds from this area continue to challenge our understanding of angiosperm evolution (Sun et al. 2011 - see eudicots below).

Even looking at extant magnoliid and Anita-grade angiosperms, distinctions between different kinds of floral parts can be hard to make, including the boundary between floral and other structures, such as the distinction between perianth and prophylls; the numbers of parts and their arrangement also vary here and in other members of the ANITA grade and magnoliids in particular (e.g. Upchurch 1984 [stomata]; Buzgo et al. 2004; Taylor et al. 2008; Endress 2008a; Doyle & Endress 2011 and references). Indeed, Chanderbali et al. (2009, esp. 2010 and references), promote a fading borders/sliding boundaries model of floral evolution (see also Thiessen & Melzer 2007), and find that the expression of genes that are quite tightly linked to particular floral whorls in eudicots show much less specificity in expression in more basal angiosperms (they studied Lauraceae and Nymphaeaceae). As gene expression is canalized, distinctions between different kinds of floral organs become sharper.

Tricolpate pollen, the signal of eudicots, has been reported from the Late Barremian-Early Aptian some 125-120 million years before present (e.g. Magallón et al. 1999; Sanderson & Doyle 2001). However, recent work suggests that in west Portugal and elsewhere tricolpates are only in low numbers in the early Aptian, but later in the Early Albian ca 112 million years ago there is evidence of rapid diversification of angiosperms including these tricolpates (Heimhofer et al. 2005). Tricolpate pollen may germinate faster, even if the grains themselves remain viable for a shorter time than monoaperturate pollen (e.g. Furness & Rudall 2004). The age of the first eudicot pollen is similar to that of the oldest monocot fossils; monocots and eudicots are sister taxa. Thus a distinctive monocot pollen type of a comparable age (120-110 million years before present) has been fairly safely identified as Araceae-Pothoideae (Friis et al. 2004; see also Doyle et al. 2008); however, it is not surprising that overall monocot fossils are not very common since monocots are a predominantly herbaceous group. This was a time of climatic and environmental instability and change, which may have favoured angiosperms, adapted as they may well have been to disturbed habitats (Heimhofer et al. 2005).

Whenever they originated, eudicots diversified rapidly only in the later Aptian and through the Albian (Friis et al. 2006b for references). Thus fossils of Nelumbonaceae - as Nelumbites, the leaves with rather different venation but the flowers with the distinctive expanded floral receptacle of extant Nelumbo - are reported from the from the mid to late Albian 118-112 million years ago (Upchurch & Wolfe 2005), while a "tubular gynoecium" (?connate carpels) is reported in a flower from Burmese amber 115-100 million years old (Santiago-Blay et al. 2005). Fossils of this age continue to have odd assemblages of characters (see also Friis et al. 1995), and the anthers often had stout filaments and connectives and dehisced by valves (Endress 2011a), as in all but the earliest angiosperm fossils.

There has been considerable work on the evolution of seed and fruit size and dispersal mechanisms (see also below). Tiffney (1986a) suggested that seeds of early angiosperms were mostly rather small (compared with extant gymnosperms - see also Haig & Westoby 1991; Linkies et al. 2010) and abiotically dispersed. However, Eriksson et al. (2000b) based on a sample of some 100 taxa from the Barremian-Aptian (132-112 million years ago) suggested that even then ca 25% were animal dispersed, although in size they were very like the abiotically-dispersed propagules (but cf. Eriksson 2008; Dilcher 2010). In any event, the embryos of early angiosperms are likely to have been small to minute (measured as the embryo:seed ratio), resulting in a period of physiological dormancy before germination could occur. Relative embryo size has increased during seed plant evolution and the need for dormancy has decreased (Forbis et al. 2002; Linkies et al. 2010).

Flowers unequivocally identified as those of core eudicots are first known from the Cenomanian only some 96-94 million years ago (Basinger & Dilcher 1984). The flower they described, the Rose Creek fossil, was not assigned to any extant family. It is relatively large compared to the tiny flowers so common in other Cretaceous angiosperms and the five stamens are somewhat unexpectedly opposite the petals; it is the earliest fossil flower with a nectary (Friis et al. 2011). However, the major groups of asterids, and rosids - and of monocots - were all probably diverging by the earlier part of the Cretaceous (Sanderson et al. 2004). The pollen record suggests core eudicots may have been around since the Albian 125-112 million years ago, and they diversified rapidly, with flowers assignable to a variety of asterid groups and also to Saxifragales (ovary inferior, crowned by a nectary, styles more or less separate, i.e. they look very like the old woody Saxifragaceae!) being especially well represented, as are Ericales (Friis et al. 2006b). Crane and Herendeen (1996) note that taxa referable to extant angiosperm families appear in the fossil record in east North America around 115-90 million years ago, and by some 85 million years ago their diversity had increased considerably (see also Lidgard & Crane 1988; Friis & Crepet 1987; Friis & Endress 1990; Crepet et al. 2004, etc.). The diversity of floral form in the Turonian of east North Americas is very considerable, magnoliids, rosids and asterid-Ericales all being represented (e.g. Crepet & Nixon 1996). Fossil wood of angiosperms is known from deposits of up to about 120 million years ago, although its assignment to extant clades is not easy (Oakley et al. 2009).

Dipsacales, Saxifragales, major clades of both rosids and monocots, etc., may all have radiated rather rapidly in this general period (Jian et al. 2008 and references; Wang et al. 2009), although modern representatives of the clades that diverged then belong to much more recent radiations (see below). Rosids in particular seem to have been common in the Late Cretaceous (Friis et al. 2010b), and diversification of the major rosid clades may have occurred (114-)108-91(-85) million years ago, and those of Fabidae and Malvidae very soon after, (113-)107-83(-76) million years ago (Wang et al. 2009). Saxifragales, although now not very speciose, may also represent an ancient and rapid radiation (Fishbein et al. 2001; Fishbein & Soltis 2004), early divergence of the main clades perhaps occurring over a period as short as 3-6 million years (Jian et al. 2008). Within Malpighiales, separation of several clades was estimated as occurring some time in the late Aptian, perhaps (119.4-)113.8(-110.7)/(105.9-)101.6(-101.1) million years before present (Davis et al. 2005a: high and low estimates). Initial diversification there seems to have been rapid, and relationships within the order could for some time be represented only by a major polytomy (Wurdack & Davis 2009; but cf.). Plants with distinctive pollen assignable to the Normapolles complex (Fagales - Fabidae) were both diverse and ecologically very prominent in the area from east North America to western Asia starting in the Late Cenomanian and Early Turonian ca 99 million years ago (there were other pollen provinces, the southermost of which was also characterized by pollen from Fagales - see e.g. Pacltová 1981 for a review; Kedves & Diniz 1983; Friis et al. 2006b, 2010b). However, odd character combinations - explained as the result of reticulate evolution (hybridization) - are found in pollen of late Campanian/early Maastrichtian age of mixed Aquilapollenites/Normapolles from Siberia (Hofmann et al. 2011).

The origins of several clades in Malpighiales and Ericales whose representatives now are major components of tropical rainforest are also to be pegged to the Mid Cretaceous or slightly later (e.g. Davis et al. 2005a). Interestingly, stem-group Rafflesiaceae are estimated to have diverged from other Malpighiales some 95 million years ago, with Sapria, representing the first clade to split off from the rest, diverging some (95.9-)81.7(-69.5) million years ago (Bendiksby et al. 2010); this might suggest the presence of humid forest - where all Rafflesiaceae are now found - by then. Similarly, mycoheterotrophic clades of Dioscoreales may have diverged from the rest (118-)109-79(-68) million years ago, with the mycoheterotrophic habit being established by some time before the beginning of the Palaeocene ca 65 million years ago (Merckz et al. 2010). Epiphytic ferns are commonly found growing on angiosperms and prefer humid conditions; they began to diversify in the Cretaceous (Schuettpelz & Pryer 2009; Watkins et al. 2010, but see below). An exception is Trichomanes and relatives (but not Hymenophyllum and its relatives) which were diversifying in the early Cretaceous - but then Trichomanes and relatives are commonly epiphytic on tree ferns, which themselves had begun diversifying in the Jurassic (Schuettpelz 2007; see also Schuettpelz & Pryer 2009; Rothwell & Stockey 2008 for early radiations of leptosporangiate ferns).

Coiffard et al. (2006, 2007) suggest that up to the Cenomanian ca 98 million years ago many angiosperms grew in aquatic, shady, or disturbed flood plain-type habitats, while Bond and Scott (2010) suggest that until the Mid or even Late Cretaceous angiosperms were mostly small herbs to small trees of the understory growing in dryish conditions, perhaps rather weedy plants (Feild et al. 2011b), and at least some early leaf floras from Portugal have leaves that are small in size and xeromorphic in appearance (Friis et al. 2010b). Taken together, this suggests a variety of water relationships for these early plants...

The period from 110-80 million years ago is that of the so-called Cretaceous Terrestrial Revolution (CTR: Meredith et al. 2011). This period has been characterised as one in which major changes in the terrestrial flora occured; there was also diversification of mammals, the net diversification rate showing a major increase around 80 million years ago at the end of the CTR, immediately followed by a comparable decrease (Meredith et al. 2011), and perhaps also bees (Engel 2000). Indeed, later in the Cretaceous not only did plant diversity increase, being high even close to the Cretaceous Arctic Circle (Hofmann et al. 2011), but angiosperms may have achieved at least some measure of ecological dominance (Friis et al. 2006b). Large trees first appear in the fossil record in the Late Albian to Cenomanian about at the time that angiosperm leaf venation density first becomes markedly greater that of non-angiosperms (Feild et al. 2011b; cf. Bond & Scott 2010 in part). Thus Friis et al. (2006a) note a dramatic increase of phylogenetic diversity and ecological abundance of angiosperms in the Late Albian-Cenomanian, about the time of a small increase in the net diversification rate of mammals (Meredith et al. 2011), while Crepet (2008) recorded the first appearance in the fossil record of a number of eudicot characters in the Cenomanian, and in particular during the Turonian some 90-88 million years ago. In Australia, angiosperm pollen increased from a low level in the middle Albian ca 105 million years ago to about 35% of the total spores at the end of the Cretaceous, while free sporing plants dropped from 80% to 45% of the total during the same period; interestingly, individual fern families did not all behave the same, and there were differences between Australia and North America (Nagalingum et al. (2002).

Wang et al. (2008) support this general idea, noting that there was contemporaneous diversification of ants, mosses, beetles and hemipterans at about this time. Trends like increasing seed size were already evident (Eriksson et al. 2000a and references), and the sugar-rich fruits of angiosperms may have provided a habitat for budding yeasts such as Saccharomyces cerevisiae in which there was a genome duplication ca 100 million years ago that was perhaps connected with their ability to exploit this habitat (Wolfe & Shields 1997: molecular clock for 18S ribosomal RNA assumed; Conant & Wolfe 2007). Locally, herbivory was quite prominent (Labandeira et al. 2002b).

Nevertheless, just how ecologically dominant angiosperms had become in the later Cretaceous remains unclear. In their discussion of Cretaceous angiosperm ecology, Wing and Boucher (1998, p. 379) concluded that at the end of the Cretaceous, diversification of flowering plants represented "the evolution of a highly speciose clade of weeds but not necessarily a major change in global vegetation", while Eriksson et al. (2000) suggest that Late Cretaceous vegetation was open, rather dry (leaf size was relatively small - Upchurch & Wolf 1987), and disturbed by herbivores (see also Schönenberger 2005). Fires were relatively common throughout the Cretaceous,and they may have encouraged a rather shrbby, low stature vegetation (Bond and Scott 2010). Some of these features are linked to nutrient cycling, and will be discussed below. It may be that through much of the Cretaceous, dominance of angiosperms tended to be restricted to fluvial or disturbed environments, although diversity did sometimes become quite high towards the end of this period, with angiosperms forming a canopy at least locally by the end-Cretaceous (e.g. Upchurch & Wolfe 1987; Crane & Lidgard 1990; Boyce et al. 2010). Note that there was a second major increase in angiosperm venation density at the end of the Cretaceous and beginning of the Tertiary some 109-60 million years ago, suggesting that only then did the forests assume a more "modern" physiology, the trees then having a venation density similar to that of plants in the most productive forests today (Brodribb & Feild 2010; Feild et al. 2011b).

Furthermore, at least parts of this whole story are challenged by the discovery of Leefructus, assigned to stem group Ranunculaceae, in deposits of at least 122.6 million years old (Sun et al. 2011); the discovery of tricolpate pollen to confirm the general assignment of theis fossil would be comforting - most extant Ranunculaceae have tricolpate pollen. The fossil, if correctly identified, will also suggest a notably greater age for Ranunculales - and hence of core eudicots; ca 152-140 million years or so would be an estimate, based on the ages given by Anderson et al. (2005). And of course somne other ages for angiosperm evolution (see above) would suggest that this scaanario needs to be revisited.

To conclude. The "museum" hypothesis of Tertiary diversification suggests that many clades persisted across the K/T boundary (Stebbins ; Couvreur et al. 2011a, b, esp. c), and Couvreur et al. (2011c) suggests that the diversification rate of palms held more or less constant from their origin ca 100 million years ago for the subsequent 65 million years - right across the K/T boundary. Other suggestions are that there was initial rapid diversification which slowed down as global cooling occured ("ancient cradle"), or that there was a progressive increase in diversification rate towards the present, the "recent cradle" theory (see Couvreur et al. 2011c for references). These theories are not mutually exclusive, and certainly many groups have diversified well after the palms, and mostly in the Tertiary (see below). On balance, it seems that if palms are the iconic plant of tropical rainforests today, either the rainforest in which they initially evolved some 100 million years ago was rather different from that of today (see e.g. Feild et al. 2011b), tropical rain forest of "modern" aspect was very restricted then and remained so for millions of years, or there are methodological problems with in analyses that yield such results (see e.g. Quental & Marshall 2010). Or perhaps we should read less from the present into the past; rainforests may have have changed over time, and to pigeon-hole them may not be very productive.

Tertiary Diversification

Details of the effect of the end-Cretaceous bolide impact that occurred ca 65.5 million years ago had on angiosperm evolution are unclear, although the end of the Cretaceous and beginning of the Tertiary clearly mark an important change in angiosperm ecology. The impact caused up to 80% loss of plant species, at least in some places in North America (Upchurch & Wolf 1987), and there was concomitant extinction of diet-specific herbivorous insects as evidences by a survey of the types of damage caused by herbivore feeding on the leaves (Labandeira et al. 2002a, b; Wilf 2008). Indeed, in North America at least locally the vegetation seems to have suffered "sudden ecosystem collapse", even common plants not transgressing the Cretaceous/Tertiary boundary (Wilf & Johnson 2004), although the severity of the impact seems to have depended in part on physical location (Johnson & Ellis 2002). In New Zealand the iridium anomaly associated with this impact was followed by a thin layer high in fungal remains, even if the original vegetation seems to have returned quite quickly, while in both hemispheres there are fern spikes (and, in the Netherlands, a bryophyte peak) after the impact (Vajda & McLoughlin 2007 and references; Nichols & Johnson 2008), although evidence from Australia is unclear (e.g. Macphail et al. 1994; Hill & Brodribb 2006). In seed plants, both insect-pollinated and/or evergreen taxa suffered more than wind-pollinated and/or deciduous taxa (Collinson 1990; McElwain & Punyasena 2007). Groups other than plants and dinosaurs suffered; there are estimates of about 60% loss of butterfly diversity at the K/T boundary (Wahlberg et al. 2009) and a mass extinction of birds, at least in western North America (Longrich et al. 2011), although they are inclined to think extinction may have been more widespread. Consistent with a substantial affect on the vegetation, plant and insect diversity was decoupled in the early Tertiary in North America for a couple of million years after the bolide impact (Wilf et al. 2006).

Estimates of the time that the vegetation took to recover from the impact range from anything from only a few thousand years (Vajda & McLoughlin 2007) to over a million years (McElwain & Punyasena 2007), depending on where on the globe the forests were. Similarly, recovery of algal primary productivity in marine ecosystems may have taken as little in the order of century or perhaps even less (Sepúlveda et al. 2009: Denmark), although most estimates of marine recovery are longer (literature in Wilf & Johnson 2004). In North America there may have been initial recolonization by swamp- and mire-loving plants, which survived the impact better (Johnson 2002; Labandeira et al. 2002b), and deciduous vegetation was also prominent early in the Palaeocene (Collinson 1990)

However, other evidence suggests that any effects of the bolide impact were not so severe. After evaluating all the evidence, Nichols and Johnson (2008) suggested that no major plant group disappeared at the end of the Cretaceous, even if species may have, although we shall see that Eocene and later vegetation is very different from that of the Late Cretaceous. Understanding exactly what clades persisted throught the K/T boundary is important. By and large the main pollen genera persist across the K/T boundary, even if species do not (Tschudy & Tschudy 1986). Recent work suggests that for Annonaceae there was a constant diversification rate for the first 80 million years of its existence, a period spanning the K/T boundary (Couvreur et al. 2011a), and this may be true for another iconic tropical family, Arecaceae (Couvreur et al. 2011c), although some clades show a relatively high diversification rate soon after their origin (literature summarized by Couvreur et al. 2011a). Seed ferns survived until well into the Palaeocene in Tasmania (McLoughlin et al. 2008). Such findings support the "museum" hypothesis of Tertiary diversity. Rather surprisingly, given what the text-books say, mammal diversification also seems to have been little affected by the end-Cretaceous extinction event (Bininda-Emonds et al. 2007; Meredith et al. 2011), also supporting the "museum" hypothesis of Tertiary diversity.

Although in general angiosperm diversity in the tropics and warm temperate areas was rather low during the Palaeocene (Wilf 2008), middle Palaeocene (ca 61 million years old) vegetation in France was diverse and also supported a diverse assemblage of herbivores, and this seems to be true of a number of sites far distant from the Mexican point of impact of the bolide (Wappler 2009 and references). Thus this event in Colombia is reflected more by changes in ecological structure, less in extinction (De la Parra et al. 2007; see Graham 2010 for a summary of the vegetational history of Latin America). Fossils from the Late Palaeocene in Colombia imply that although the basic familial composition of the forest was similar to that of current neotropical rainforest, although both plant and herbivore diversity were rather low (Jaramillo et al. 2006). This may reflect a rather belated recovery from the bolide impact and/or that the tropical rainforest ecosystem was just developing (Wing et al. 2009) - indeed, this Colombian flora studied by Wing et al. (2009) has a very high venation density and is the first evidence of functional equatorial neotropical megathermal rainforst (Feild et al. 2011b). Interestingly, Palaeocene and Eocene Patagonian vegetation was more diverse than its North American counterparts (Iglesias et al. 2007; Wilf 2008).

Angiosperms with their often dense veinlet reticulum and high leaf specific conductivity (Watkins et al. 2010), in turn associated with high rates of photosynthesis and transpiration, may have helped to drive the spread of widespread tropical forest with reliably high rainfall, numerous epiphytes, and overall high diversity (Boyce et al. 2008, 2009, 2010; Boyce & Lee 2010). Fast decomposition of their litter may also have speeded nutrient cycling in this forest (Cornwell et al. 2008; Berendse & Scheffer 2009), and this is discussed in more detail below. There was major increase in angiosperm venation density at the end of the Cretaceous and beginning of the Tertiary, suggesting that only then did the forests assume a more "modern" physiology (Feild et al. 2011b). Interestingly, wood with scalariform perforation plates was particularly common in the Cretaceous (Wheeler & Baas 1993), and there seem to have been some rather abrupt xylary changes across the K/T boundary in features like vessel length and perforation plate morphology - scalariform perforation plates became less common - that would favour xylem conductance; there were also some abrupt changes in wood parenchyma, but not in ray morphology, etc. (Wheeler & Baas 1991). Indeed, it has been suggested that only in the early Tertiary, and in the Eocene rather than the Palaeocene, did vegetation take on a more modern appearance with the development of a closed, multi-layered forest (e.g. Upchurch & Wolfe 1987; Wing 1987; Eriksson et al. 2000a; Burnham & Johnson 2004; Pennington et al. 2006; Crane & Carvell 2007 discuss the early Tertiary fossil record; Morley 2000 provided a good general account of the evolution of rainforests). Tropical forests as we think of them, the "modern archetypal tropical rain forest" of Burnham and Johnson (2004, see their Table 1), with lianes, epiphytic ferns, bromeliads and orchids, and relatively large leaf blades with entire margins, thus seem to be a Tertiary phenomenon (Upchurch & Wolf 1987; Schuettpelz 2006; Boyce et al. 2009: Schuettpelz & Pryer 2009; Watkins et al. 2010; Bond & Scott 2010). Looking at individual features of fruit and seed, there are substantial differences between Tertiary, especially Eocene and younger, and Cretaceous fossils, for instance, there was a wide variety of winged disseminules in the former, but not the latter, floras, some fruits were fleshy (but rather few in the Palaeocene) and far larger than early Cretaceous fleshy fruits (Collinson & van Bergen 2004).

Climate changes were also occurring at about this time. Herbivore intensity and diversity increased with the warming tend of late Paleocene-Eocene (Labandeira et al. 2002b), diversification rates of modern mammals also increasing (Bininda-Emonds et al. 2007). Tropical diversity peaked in the Eocene as it rebounded from the bolide impact, perhaps even topping today's levels (Wilf 2008 - check); by the early Eocene South American fossil floras were notably diverse, including lianes, declining only at the end of the Eocene (Jaramillo et al. 2006, 2010; Wilf et al. 2011, Patagonian floras; Herrera et al. 2011), and the same is true of floras in today's North Temperate zone, woody plants often with tropical affinities being prominent in the flora there (e.g. Tiffney 1985a, b). Upchurch et al. (2007; see also Zachos et al. 2008) suggest that the climate in the early Tertiary around 56.3 million years ago was much warmer and wetter, while during the post-Eocene thermal maximum (PETM), temperatures went up 3-5^o (or 5-8^o or more - estimates vary) to 31-34^o C (Willis & MacDonald 2011); note that today's tropical rain forest has a mean annual temperature of ca 27.5^o C). Over 2,000 gigatons of carbon were released in ca 10,000 years, the whole event lasting a mere 1-200,000 years (McInerney & Wing 2011). Interestingly, in South America plant diversity and origination rates increased at about this time, and there is no evidence of thermal damage to the leaves (Jaramillo et al. 2010) while in west-central North America plant diversity and herbivore diversity (and activity) on those plants increased (Currano et al. 2008); overall there is little evidence of extinction (Willis & MacDonald 2011). Fires seem to have decreased in frequency acter the Paleaecene, and Bond and Scott (2010) suggest this may have encouraged treees of larger stature. [Link rest of para.] In parts of Europe there is evidence for episodic fires in a vegetation dominated by ferns and perhaps Fagales (Collinson et al. 2007). A unique mixed broad-leaved deciduous and conifer forests grew north of 65-70^o N (Collinson 1990 and references).

After the PETM, diversity decreased strongly from the late Eocene-Oligocene through the Miocene, but then rebounded; tropical floras became less widespread, and a number of herbaceous groups diversified (e.g. Tiffney 1985a, b). There was cooling and drying ca 26 and 16 million years ago (Crisp & Cook 2011), and concentrations of carbon dioxide in the atmosphere were falling (e.g. Arakaki et al. 2011). Although ecological conditions then may still have been rather different from those of today, it is noteworthy that many fossils from this period are assignable to extant genera and any ecological differences are surely less than when comparing Cretaceous and extant angiosperms (Jaramillo et al. 2006; Mittelbach et al. 2007). Note that quite early in the Tertiary the distributions of a number of temperate and tropical taxa that are today rather restricted were much wider (e.g. Plaziat et al 2001: Nypa; Smith et al. 2008: Cyclanthaceae; Herrera et al. 2011, Stephania); a number of taxa now restricted to Southeast Asia also occurred in Europe and North America at various times from the Palaeocene to the Miocene (e.g. Ferguson et al. 1997; Manchester et al. 2009: East Asian endemics).

The increasing prevalence of angiosperms was not the result of a simple replacement of gymnosperms. Indeed, divergence within gymnosperm clades (i.e. genera), both Cycadales and Pinales, is for the most part quite recent, in the mid to later part of the Tertiary (Nagalingum et al. 2011; Crisp & Cook 2011). Extinction may have been higher in gymnosperms in general than in angiosperms, hence resulting in lower gymnosperm diversification, certainly, gymnosperm clades have longer stems and shallower crowns (Crisp & Cook 2011). Within podocarps, diversification is largely a Tertiary phenomenon, and that of members with flattened foliage of one sort or another is both later and notably greater than those with imbricate leaves (Biffin et al. 2011). Podocarps with flattened foliage units are often shade tolerant and diversified slightly after the venation density of angiosperm leaves increased - (94-)64(-38) versus 109-60 million years ago (cf. Biffin et al. 2011; Brodribb & Feild 2009; Biffin & Lowe 2011).

In the mid-Pliocene some 3.6-6 million years ago temperatures were 2-3^o warmer than they are now, again, novel vegetation assemblages develop and there was increased diversity (Willis & MacDonald 2011)Evidence of marked seasonality in fossil woods is a Neogene (Pliocene and since - the last ca 23 million years) phenomenon (Wheeler & Baas 1993). The Neogene might be called the age of grasses (with apologies to the Palaeos website). The rise to ecological prominence of grasses, including the clades that carry out C₄ photosynthesis, is a Neogene phenomenon; although the origin of this trait may go back 20 million years or more, major diversification in the clades concerned happened independently for the most part only 10-3 million years ago - and the same is true of many clades with succulent plants, whether terrestrial or epiphytic, and there CAM photosynthesis is common (Arakaki et al. 2011).

Turning now to individual clades, a number of common Cretaceous plants failed to survive into the Tertiary, and the familial composition of Early Tertiary forests differed from that of their Late Cretaceous counterparts (e.g. Johnson 2002; Wilf & Johnson 2004). For clades that did cross the boundary, features like seed size, notably larger in Tertiary floras, also increased across the K/T boundary (Collinson & van Bergen 2004). Magallón et al. (1999) noted that major core eudicot clades like Fabaceae and (most of) Lamiales that together represent about 45% of core eudicot diversity appear only in the upper Cretaceous (Maastrichtian) and Tertiary. Orchid diversification (ca 22,000 species) seems to have been a largely Tertiary phenomenon (Ramírez et al. 2007; Gustafsson et al. 2010), as was that of Asteraceae (23,000+ species: K.-J. Kim et al. 2005; Funk et al. 2009c for a summary) and Fabaceae (19,000+ species: e.g. Bruneau et al. 2008b; Bello et al. 2009), etc., while even in much older clades like Myristicaceae and Annonaceae, much diversification may also be largely a Tertiary phenomenon (J. A. Doyle et al. 2008 and references; Richardson et al. 2004; cf. in part Couvreur et al. 2011a). Ca 24,440 species of angiosperms are epiphytic (Schuettpelz & Pryer 2009), about half being members of Orchidaceae-Epidendroideae (Ramírez et al. 2007; Gustafsson et al. 2010; Conran et al. 2009) and Bromeliaceae (Givnish et al. 2008a) alone, and both of these groups diversified in the Tertiary.

It is perhaps hardly surprising that when we look at the diversification of other organisms now associated with angiosperms, we find a similar story. Non-angiosperm epiphytes - most of which are ferns, especially the polypod ferns which make up at least 80% of the living fern species - may have evolved in the Late Cretaceous after the initial diversification of the angiosperms. One third (ca 3,000 species) of all leptosporangiate ferns are epiphytic, about 10% of all epiphytes (Schuettpelz & Pryer 2009 - see Brodribb & Holbrook 2004 for the comparative leaf physiology of ferns and angiosperms), and their diversification occurred during the early Tertiary and was perhaps linked with the Palaeocene-Eocene thermal maximum, which occurred some 10 million years after the bolide impact (Schneider et al. 2004a, b; Schuettpelz 2007; esp. Schuettpelz & Pryer 2009: Supplemental Tables 2, 3; Watkins et al. 2010); an exception is Trichomanes and relatives, but not Hymenophyllum and its relatives (see above). About half - 190/380 species - of clubmosses, Lycopodium s.l. (Wikström & Kenrick 1997; Wikström 2001; Schuettpelz & Pryer 2009), while in liverworts, although the largely leaf-epiphytic Porellales diverged from the terrestrial Jungermanniales in the Jurassic, there may have been enhanced diversification of Porellales in the Cretaceous and early Tertiary (Heinrichs et al. 2007; see also Ahonen et al. 2003; Forrest & Crandall-Stotler 2004). Within the speciose pleurocarpous mosses - about 40% of all mosses of which many are epiphytic, especially in Hypnales - diversification seems to have been early-Cretaceous and rapid, with subsequent semi-stasis (Shaw et al. 2003b; Newton et al. 2006, 2007; see also Kürschner & Parolly 1999), and although there may also have been more recent ([post-]Cretaceous) diversification as well, the initial radiation seems to be at about the same time as the early rise of the angiosperms.

Even diversification of Cycadales (Oberprieler 2004; esp. et al. 2011; Nagalingum et al. 2011), and Equisetum (Des Marais et al. 2003, but cf. Stanich et al. 2009) seems be a Tertiary phenomenon. Nevertheless, the clade that contains Equisetum itself has probably been separate from other monilophytes since the Permian, ca 250+ million years before present, and taxa with some of the apomorphies of crown group Equisetum are known from Lower Cretaceous deposits some 136 million years or more old (Stanich et al. 2009).

Animal groups diversify notably in the early Tertiary, often somewhat later than do plants (Winkler & Mitter 2008; Janz 2011). Although initial diversification of ants seems to have started some 125 million years ago as angiosperms diversified, ants seem to have diversified more in the late Cretaceous-early Eocene 75-50 million years before present, well after the initial diversification of the angiosperms. Ecological dominance of ants may have occurred only in the later Eocene, ants becoming common in the fossil amber record only then (Grimaldi & Agosti 2000; Moreau et al. 2006; Dunn et al. 2007; Pie & Tschá 2009); there is some argument over this, but general agreement over the main timing of diversification (see also Brady et al. 2007; Wilson & Holldöbler 2005). Sugar obtained either directly or indirectly from plants is an important food/energy source for many ants, while plant material in general presents a great variety of resources for them (cf. Gorelick 2000). Ants make up to 86% of the total arthropod biomass in tropical rainforests, and although some are carnivores, most eat oplant materials (Rico-Gray & Oliveira 2007). Evidence of ant-plant associations can be obtained from the fossil record, e.g. in the extra-floral nectaries of Populus fossils from the Oligocene (Pemberton 1992). The commonly encountered associations between ants, plants and sap-sucking homopterans (Ueda et al. 2008) are mid-Tertiary or later.

Elaiosomes attractive to carnivorous ants - the fatty acids the elaiosomes contain may mimic those in the animal prey of these ants, the elaiosomes being "dead insect analogue[s]" (Carroll & Janzen 1973: p. 235; Hughes et al. 1994) - are found on small seeds or fruits and occur in many plants (Beattie 1985; Rico-Gray & Oliveira 2007). Their appearance in clades such as Polygalaceae-Polygaleae seems to be associated with the diversification of these clades and is a mid-Tertiary phenomenon (Forest et al. 2007b; Lengyel et al. 2009, 2010, see also Fokuhl 2008). Elaiosomes provide food for the ants, which will not eat the seeds themselves (cf. granivorous ants) and they aid in the dispersal of plant disseminules and perhaps in the establishment of the seedling. Elaiosomes that vary considerably in their morphological nature, their persistence, and content (e.g. Kubitzki et al. 2011) and myrmecochory are particularly common in the ground flora of the east North American and European forests (Sernander 1906; Berg 1975; Orians & Milewski 2007; Lengyel et al. 2009, and references: Türke et al. 2011 suggest gastropods may also be involved in the distribution of such seeds), the Australian flora, as well as in the South African flora (Milewski & Bond 1982; Bond et al. 1991). All told, myrmecochory may involve some 11,500 species (Lengyel et al. 2009, 2010 - estimate conservative), and myrmecochorous clades have about twice as many species as their non-myrmecochorous sister clades (Lengyel et al. 2009). In addition, perhaps half the species of stick insects (Phasmatodea) lay eggs that mimic those of the myrmecochorous plants discussed (Hughes & Westoby 1992). In general, plant-ant associations like myrmecochory may have evolved in the late Eocene and afterwards, but probably not earlier (Dunn et al. 2007).

Did leaf beetles, Chrysomelidae, and angiosperms diversify more or less together, or was the diversification of the insects later (cf. Farrell 1998 and Gómez-Zurita et al. 2007)? Again, the timing of such relationships is perhaps that animals diversify rather later than plants (Winkler & Mitter 2008), and diversification is particularly obvious in the early Tertiary. Even if initial diversification of beetles and angiosperms was associated, subsequent bouts of diversification have occurred well after the appropriate angiosperm host clades originated (implicit in Futuyma 1983; see Funk et al. 1995; Percy et al. 2004; Lopez-Vaamonde 2006; Wheat et al. 2007; Kölsch & Pedersen 2008; Janz 2011). This may not always be so, thus Kergoat et al. (2005) suggest that diversification of bruchids and Fabaceae may have occurred more or less together. In general, close co-evolution seems to be the exception rather than the rule, and is most evident in shallow rather than deep clades (Berenbaum & Passoa 1999 for references; cf. Farrell & Mitter 1998); looser "co-evolution", with host shifts associated with taxonomy, may be more common (see Futuyma & Mitter 1996).

Diversification within clades representing extant subfamilies of butterflies (Papilionoidea) seems to have happened after the K/T boundary (Heikkilä et al. 2011). Within Nymphalidae, clades that represent butterfly tribes today diverged only after the K/T boundary even if family clades are largely of late Cretaceous origin (Wahlberg et al. 2009); the butterfly (and other herbivore) clades that survived the K/T boundary may have grown on several alternative food plants and this gave them an oportunity to diversify - perhaps on a more restricted set of host plants - subsequently (Nylin & Wahlberg 2008; Fordyce 2010). However, there are rather different estimates for divergene within Papilioninae - all tribes date to before the K/T boundary (Michel et al. 2008) to divgence more or less at the boundary (Simonsen et al. 2011), but estimates of the position and hence age of Papilionidae as a whole differ (e.g. Heikkilä et al. 2011).

A general increase in fruit size occurred some 85-75 million years ago in the Late Cretaceous-Early Tertiary (e.g. Eriksson et al. 2000a; Dunn et al. 2007). Similarly, the average seed mass of angiosperms, initially rather low, increased markedly towards the end of the Cretaceous and the beginning of the Tertiary 85-60 million years before present (e.g. Tiffney 1986b; Sims 2010), indeed, Tiffney (1968a) suggested that increases in seed/fruit size had occurred within Juglandaceae and Fagaceae by the early Tertiary. These large, nutritious seeds (or large dispersal units in general) and fleshy fruits are likely to have been dispersed by animals, and by mammals, birds, and bats in particular. This increase in size may be linked primarily to a change in forest type, now closed and made up of large trees, and/or to the evolution of animals that dispersed the seeds (e.g. Eriksson et al. 2000a; Moles et al. 2005a, b; Eriksson 2008; cf. in part Tiffney 1984, 2004). Large seeds are common in plants that at least initially grow in shaded habitats, providing reserves for the initial growth of the seedling (Leishman et al. 2000), although they may also be favoured by dry conditions, or soils with low mineral nutrients, etc. (Leishman et al. 2000; Bolmgren & Eriksson 2005). Seed mass of extant angiosperms drops quite abruptly (seven-fold) at the edge of the tropics (Moles et al. 2007: sampling in the tropics not very good), and although the reasons for this are unclear, the efficacy of wind dispersal in the open habitats that are more common there may also be involved (Lorts et al. 2008). In general, features associated with the needs of plants living in tall, closed forests interacted with the evolution of frugivorous vertebrates (Eriksson 2008) as angiosperms came to dominate the forests, and overall great diversity in propagule size developed within the angiosperms (Tiffney 1986a).

Phyllostomid bats can be important frugivores, perhaps especially in the Neotropics, where they tend to specialize on fruits of particular genera or families of plants (Lobova et al. 2009; Mello et al. 2011). They can be particularly important in dispersing seeds of understorey plants.

After the end-Cretaceous disappearance of the dinosaurs, diversification of seed-dispersing animals, including mammals and birds, and of the plants they dispersed may have proceeded roughly in parallel, although again with something of a lag for the animals (e.g. Tiffney 1984; Wing & Tiffney 1987; Collinson & Hooker 1991; Dilcher 2000; Tiffney 2004). Mammals have a substantial fossil history before the Cretacous, but their diversification in the early Tertiary is particularly notable (Bininda-Emonds et al. 2007), and herbivory (including frugivory) is common. Phyllostomid and vespertilionid bat diversification and that of angiosperms is also associated (Jones et al. 2005); insect-eating bats (the second group) may have diversified because there were more insects because of the diversity of plants, and fruit-, pollen- and nectar-eating bats (the first group) because there were a greater diversity of fruit types and an abundance of flowers (Jones et al. 2005; Teeling et al. 2005). Crown-group diversification of phyllostomid (leaf-nosed) bats, a New World clade, began some time between 43.1 and 33.4 million years ago (probably in the late Eocene), with much diversification 26-16 million years ago in the late Oligocene to mid-Miocene (Datzmann et al. 2010; Rojas et al. 2011), and the ecological importance of these bats, which can be very abundant, in terms of the services they provide plants in the New World is considerable (Freeman 2000). Both frugivory and nectarivory arose several times in parallel, and some combination of insectivory with these modes of nutrition is common (Datzmann et al. 2010; Rojas et al. 2011); adoption of a vegetarian diet seems to have accelerated diversification rates (Datzmann et al. 2010). Radiation of important seed-dispersing birds such as Columbiformes (pigeons) occurred some (63.6-)54.4(-46.1) million years ago (95% CI), also in the earlier Tertiary (e.g. Tiffney 1986b; Pereira et al. 2007). Fleshy fruits, in addition to providing food for the dispersers, also support other organisms. This food "niche" was exploited by particular groups of flies, particularly by Drosophilinae, some of the relationships between particular fruits and flies is very close (Ashburner 1998 [on alcohol dehdrogenase in flies]; Harry et al. 1996, 1998 [fig-breeding Lissocephala]).

Currently, animals - including fish - allow for the wide dispersal of rather large propagules and many also for the (cross) pollination of the rather widely dispersed individuals that produce them (e.g. Regal 1977). Indeed, it has been suggested that trees have a distinctive evolutionary rhythm, speciating rather slowly. In any one species the number of individuals may be quite large, and although they may be rather dispersed they are long-lived, the species themselves also being rather long-lived (Petit & Hampe 2006). Woody plants with fleshy, animal-dispersed seeds tend to speciate more than plants with other dispersal mechanisms (Eriksson & Bremer 1991).

However, the causal links between plant disseminule size and vertebrate disperser are unclear (Eriksson et al. 2000a; Eriksson 2008; see also Mack 2000; Dilcher 2010). Note that although there is a connection between large seeds, fleshy fruits, and the arboreal habit, exactly what drives this connection is unclear, and the acquisition of fleshy fruits in particular is not linked with notable increases in diversification of clades with them (Bolmgren & Eriksson 2010 and literature). Indeed different ecosystem dynamics may have prevailed in the Cretaceous and Tertiary (Tiffney 2004), and it is in the context of these that changes in seed size (and of other aspects of the plant) occured.

In extant angiosperms, there is a correlation between woodiness and tannin frequency and a negative correlation between tannins (generalized defence) frequency and alkaloid and other secondary metabolites (specific defence) frequencies (e.g. Feeny 1976; Silvertown & Dodd 1996; see also Levin 1976; Mole 1993). In general, one can distinguish between plants in which defence is "qualitative", the defensive compounds being highly toxic and butterfly groups like Nymphalidae being specialized herbivores, and plants with "quantitative" defence, in which the defences are more generalised - polyphenolics and the like - and where groups like Lycaenidae are the herbivores (Fielder 1996). The nature of the defensive compounds produced by plants can also be linked with resource availability, qualitative defences being linked with low concentrations of available nutrients (Coley et al. 1985).

So along with a shift in ecology, there may have been a shift in metabolites involved in defence, and thus perhaps also in the pattern of herbivory. Major diversification of herbivorous beetles in particular and insects in general occurred in the thermal maximum of the later Paleocene-Eocene ca 56.8 million years ago (the Paleocene-Eocene Thermal Maximum, PETM: Farrell 1998; Wilf & Labandeira 1999; Wilf et al. 2001; Lopez-Vaaamonde et al. 2006), an event that also caused some marine extinction and shifts in the distributions of both plants and animals, althoughn perhaps little extinction (Wing et al 2005; Willis & MacDonald 2011).

Although Novotny et al. (2006) suggested that individual species of temperate and tropical plants (controlled for phylogenetic relationships) support a similar number of insect species, since there are many more species of plants in the tropics, there will be many more species of insects there. However, recent work has also suggested differing patterns of association (cf. Novotny et al. 2007; Dyer et al. 2007).

Other major shifts in seed mass are also rather strongly correlated with changes in life form/plant habit (Eriksson et al. 2000a; Moles et al. 2006a, b). Much angiosperm diversity is concentrated in groups that are annuals or herbaceous or shrubby perennials, with animal pollinated flowers and small disseminules that are not often dispersed by animals (Eriksson & Bremer 1991, 1992). Such herbs, especially annuals, have smaller seeds and rarely have fleshy fruits; several of the large groups with monosymmetric flowers mentioned below (Lamiales, Asteraceae, Fabaceae) also include many members that are more or less short-lived herbaceous or shrubby plants. However, in the fire-prone Mediterranean ecosystem a study suggested that neither diversification nor molecular evolution differed between seeders and resprouters, two "strategies" allowing plants to survive fires, although seeders under some scenarios should have shown more diversification (Verdú et al. 2007). Overall diversification rates are certainly highest in the asterid I and II clades, particularly in Asterales and Lamiales (Magallón & Sanderson 2001; Magallón & Castillo 2009). Dodd et al. (1999) also found that the evolution of herbs from trees was correlated with a rise in the speciation rate of the former, although Verdú (2002) suggested that is was not so much the tree habit per se that was important, but the associated condition, length of time to maturity.

Molecular studies have long suggested that there is correlation between the rate of molecular evolution and plant habit: molecular evolution is faster in herbs/annuals (e.g. Wilson et al. 1990; Bousquet et al. 1992: esp. chloroplast genes; Gaut et al. 1992: chloroplast rbcL gene, grasses evolve notably faster even than other monocot herbs, 1996; Andreasen & Baldwin 2001; Rydin et al. 2009b; Barker et al. 2009; Korall et al. 2010 [ferns]; Müller & Albach 2010; Yue et al. 2010; Frajman & Schönswetter 2011; cf. Whittle & Johnson 2003 [comparisons of branch lengths of species pairs, ?sampling]: Gaut et al. 2011 for a good summary). In a series of extensive analyses of both monocots and eudicots, Smith and Donoghue (2008) confirmed that there are usually substantial increases in the rate of molecular evolution in herbs as compared to trees, shrubs, or simply to plants with long life cycles. For instance, within commelinids the clades Arecaceae, Bromeliaceae and Rapateaceae, all with long life cycles, showed a low rate of molecular evolution. Both chloroplast and nuclear genes show an increased rate of molecular evolution (Gaut et al. 2011), but not all genes are equally affected (Yue et al. 2010).

Herbs also show an increased rate of climatic niche evolution (Smith & Beaulieu 2009). Rates of molecular evolution (substitution rates) may be correlated with the rate of speciation (Barraclough & Savolainen 2001), but this was not found to be the case in Veronica (Plantaginaceae: Müller & Albach 2010). Indeed, the cause of the correlation between molecular evolution and life history is unclear, perhaps having something to do with mutation rate or on population size and hence on speciation (Gaut et al. 2011 and references).

Of course, the success of angiosperms is often attributed in part to the pollination of their flowers by animals, of which insects predominate (Eriksson & Bremer 1992). Floral rewards and how they are offered vary: Some plants - probably including most early angiosperms - offer only pollen, while various kinds of nectaries are found in flowers of other angiosperms, e.g. septal nectaries in many monocots and receptacular nectaries of one sort or another in core eudicots (and perhaps also Proteales). Successful pollination entails the pollinator following a more or less complex and specific set of cues. Colouring of the corolla in particular, in terms of pigment type, amount, and pattern of deposition, seems to be under the control of a small family of regulatory genes in a diverse set of angiosperms (Schwinn et al. 2006). Self pollination is hindered by sporophytic and gametophytic incompatibility (see & Chase 2004), and less effectively by protandry or protygyny (the latter is commonest in members of the ANITA grade and the magnoliids). Pollination, especially by insects, but also bats and other mammals as well as birds, and seed dispersal, especially by mammals and birds, may interact, in that both may increase outcrossing and gene flow in general, and hence speciation (Kay et al. 2006; Kay & Sargent 2009 for a summary). For example, Schemske and Bradshaw (1999) in a classic paper discuss possible links between pollinator behaviour and pollination preferences of hummingbirds and bumble bees as drivers of speciation (see also Gegear & Burns 2007, floral features considered more or less separately), while Pauw et al. (2009) describe rather diffuse co-evolution between flies with long probosces and and a group of species with long-tubed flowers (see also Bascompte & Jordano 2007). But for general plant-pollinator relationships, it may be useful to think of plants tapping into pre-existing sensory biases of their pollinators and manipulating their behaviour (Schiest 2010).

However, when thinking of plant-pollinator relationships in a phylogenetic context, there is something of a paradox - we think of early angiosperms as having rather generalized flowers, yet the plesiomorphic condition for pollination specificity in bees seems to be oligolecty. Polylectic or generalist bees like bumble bees - definitely not basal in bee phylogenies - visit flowers with complex often monosymmetric corollas which the animal has to learn to work before visits are effective, while specialists prefer to visit shallow often polysymmetric flowers with easily accessible rewards (Wcislo & Cane 1996). Indeed, although a number of extant bees, including Melittidae s.l., show considerable host plant specificity, being oligolectic, this may have facilitated early angiosperm evolution (Danforth et al. 2006; Sipes et al. 2006; Michez et al. 2008). Polylectic behaviour - so no host specificity - may be derived (e.g. Müller 1996; Michener 2007; Sedivy et al. 2008; Litman et al. 2011). Thus oligolectic basal megachilids pollinate radially symmetrical and rather large flowers, while more polylectic derived members commonly pollinate monosymmetric Fabaceae and Lamiaceae (Litman et al. 2011). These preferences are by no means absolute; oligolectic bees may also pollinate monosymmetric flowers (e.g. Sedivy et al. 2008), and although a species or even an individual bee may visit many species of plants, on any one trip the bee is likely to be much more selective (Heard 1999), visiting only one or a few species, so being functionally mono- or oligolectic.

Particular bee groups may show general preferences. Thus of the Apidae, bumblebees (Bombini) in particular appear to have an innate preference for monosymmetric flowers (Leppik 1957; Kalisz et al. 2006: although Rodríguez et al. 2004 found bumblebees preferred monosymmetric flowers, the test was only against asymmetric flowers); Westerkamp (1997) described how bees pollinate flowers in which the pollen is enclosed in keels, a common monosymmetric flower type. Honey bees (Apini) frequent polysymmetric (radiate) flowers with relatively accessible nectar (for a review of the cues used by honey bees, see Horridge 2009). Bumble bess are effective buzz pollinators, honey bees are not (Goulson 2010). Halictidae-Rophitinae (sister to all other halictids) prefer to visit flowers of asterids, especially those of members of the asterid I clade, although other plants are also visited (Patiny et al. 2008). Similarly, some flower types are pollinated in a particular way. The morphologically distinctive buzz pollination floral syndrome has evolved many times, and perhaps some 4,000 species of angiosperms - mostly core eudicots - are pollinated in this way (Buchmann & Hurley 1978; Buchmann 1983). The evolution of flowers which have oils as their primary reward probably began in the Eocene. There are now some 1,500-1,800 species of oil flowers in some 11 families that are pollinated by 360-370 species of bees, and oil flowers may have evolved some 28 times even if the syndrome has been lost even more often (Renner & Schaefer 2010; see Neff & Simpson 1981 for the bees).

Although the initial diversification of lepidoptera may date to before or about the same time as angiosperm evolution, butterfly radiation - divergence of and within subfamilies - began in the Late Cretaceous some 90 million years ago (e.g. Pohl et al. 2009: 113-84 million years ago, gene duplications) but most happened after the K/T boundary (see Vane-Wright 2004; Wheat et al. 2007; Wahlberg et al. 2009 for references; Heikkilä et al. 2011), occurring in the last 75 million years. Of course butterflies are important both as insect pollinators when adult and as herbivores as larvae. Adult Lepidoptera prefer to visit radiate flowers, that is, polysymmetric flowers with a definite number signal (e.g. three-merous, five-merous flowers) and that have enclosed nectar (e.g. Leppik 1957). Evolutionary radiation of butterfly clades sometimes occurs when butterfly larvae shift their food preferences (Fordyce 2010). Turning to individual butterfly groups, diversification of of Nymphalidae-Nymphalinae seems to be a post K/T boundary phenomenon, occurring 65-33 million years before present (Wahlberg 2006), and the same is true of Nymphalidae-Papilioninae (Zakharov et al. 2004). Diversification may have begun before, as in Pieridae which began diversifying the the Late Cretaceous ([112-]95[-82] million years before present: Braby et al. 2006), but again, much speciation seems to have been a Tertiary phenomenon (see Simonsen et al. 2011 for the range of divergence times that have been suggested). Caterpillars of these groups tend to show rather high food-plant specificity.

The evolution of animals other than insects was also affected by angiosperm evolution; sometimes the connection between animal and plant is direct, sometimes, as with insect-eating bats, it is in part indirect. That the diversification of orb-weaving spiders, etc., was contemporaneous with that of angiosperms, or somewhat later is interesting - they were eating insects, at least some of which were eating plants - but it was probably of little effect on seed plant evolution. Bat pollination more directly involves plants, and this is sporadically distributed in angiosperms; bat pollinated flowers probably began evolving in the Miocene (Fleming et al. 2009), and there was much diversification in the major New World group of nectarivorous bats, the phyllostomids, 26-16 million years ago in the late Oligocene to mid-Miocene (Datzmann et al. 2010; Rojas et al. 2011). Bird pollination is also likely to be a Tertiary phenomenon, with three different groups of birds (Trochilidae, Nectariniidae, and Meliphagidae) predominantly being involved (Cronk & Ojeda 2008). C₄ photosynthesis in grasses seems to have originated in the middle Miocene, some 18-12.5 million years before present (Jacobs et al. 1999; Strömberg & McInerny 2011; McInerny et al. 2011), and it has been suggested that the origin of this pathway affected herbivory. C₄ plants tend to be less attractive to herbivorous animals because of their lower nitrogen concentration and greater amount of fibrous tissue (Caswell et al. 1973), and the rise and spread of C₄ grasses with their silica-rich tissues in the early Miocene was followed by the radiation of grazing mammals such as horses with hypsodont teeth that could deal with such refactory plant material (Thomasson & Voorhies 1990; Retallack 2001; Keeley & Rundel 2003) - and larvae of Satyrinae butterflies are found largely on Poaceae. However, Sanson and Heraud (2010) question the relation of tooth morphology and silica content of grasses, and although prairie grasses expanded in Nebraska in the Early Miocene ca 23 million years before present, hypsodont ungulates were already around by then (Strömberg 2004) - the connection between the radiation of grasses and the evolution of grazing animals needs clarification.

Overall, our understanding of the ecological-evolutionary connections between animals, in particular insects, and plants remains unclear (Janz 2011; there is no simple underlying theory to explain the variety of the interactions (Futuyma 1983). As Grimaldi and Engel (2005: p. 625) note, "Despite the fact that the mechanism is obscure as to how insects diversified with angiosperms, the overall patterns are extremely clear that the angiosperm radiations had a profound impact on insects, and vice versa." Although insects may only rarely act as selective agents on their hosts (surely not true????: Strong et al. 1983), in fact we know little about the details of long-term evolutionary-ecological interactions of plants and the organisms associated with them (see above; Fine et al. 2004 for habitat specialization and herbivore activity in the Amazon, also Janzen 1974a).

The physiological-ecological context of angiosperm evolution.

We are inclined to think of the evolution of angiosperms as being largely the consequence of the evolution of flowers (and fruits), diverse floral and fruit form representing adaptations to pollination and fruit dispersal, but there have been a number of more physiological changes that have profoundly affected the climate of the earth and the ecology of angiosperms. Some of these changes may have ensured the spread of the tropical rainforest habitat in which so much angiosperm diversity is now to be found. Others may be implicated in the long-term decline in atmospheric CO₂ concentration that characterises the Tertiary, although it started in the Jurassic. Although these changes affected the rate of photosynthesis, nutrient cycling and acquisition, and the like, our understanding of the eco-physiological dimension of angiosperm evolution is for the most part rather poor. Nevertheless, is clear that there are a number of feedback loops, many positive, implicated in changing climates and CO₂ concentration (e.g. Berner 1999; Beerling 2005a; Beerling & Berner 2005); interactions have been mutual. It is in the context of these physiological-ecological changes that angiosperm diversification in a more conventional sense has occurred. One way of thinking about these eco-physiological changes is that they allowed angiosperms to grow in a great diversity of environments, including large areas of ever-wet forests, and diversification occurred in these different environments; they provided the basic ecological space for features associated with pollination and seed dispersal in particular to interact with aspects of the biotic and abiotic environment (see also Boyce et al. 2010; e.g. Marazzi & Sanderson 2010 and diversification of a clade of American Senna with extrafloral nectaries).

Develop para.: About 30% of the organic carbon in the biosphere is currently locked up in lignin (Boerjan et al. 2003). There is no evidence of lignin-destroying fungi in the Paleozoic, and this is reponsible in part for the accumulation of carbon deposits in rocks of Carboniferous age (). In addition, there are suggestions that lycophytes had proportionally a large amount of lignin (Robinson 1990). Litter and wood decay of conifers in general is slower than that of angiosperms (e.g. Weedon et al. 2009), and brown rot fungi, which can barely degrade lignin, are commoner in conifer forests that the lignin-decaying white rot fungi (Boddy & Watkinson 1995). Indeed, there may have been an evolutionary sequence white rot -> brown rot -> ectomycorrhizae; fungi involved in the first two associations are widely scattered through Agaricomycetes (Eastwood et al. 2011).

The Permo-Triassic boundary is marked by major extinctions, and the global climate abruptly became warmer, with a mean annual temperature increase of ca 10^o C in rocks from both tropical and more temperate environments before gradually cooling again; atmospheric CO₂ concentrations increased from about 300 to 1,500 ppm (Retallack & Krull 1999 for a summary). There was a world-wide "coal gap", sediments with coal deposits being absent, and throughout the southern hemisphere cool forest area there was increased weathering in an environment that had become more unstable, the soils were more infertile (oligotrophic), and there was the replacement of deciduous (Glossopteris) forest by evergreen forest that had a lower albedo, etc. (Retallack & Krull 1999).

Carbon dioxide concentrations then decreased, but there was another major extinction event at the end of the Triassic. Again this was accompanied by an increase of atmospheric CO₂ (about four-fold, from ca 600 to 2,100-3,000 ppm) and an increase in temperature of ca 4^o C (McElwain et al. 1999). This may have increased leaf temperatures above the limit of lethality (45-52^o C for non-succulent leaves); species persisting across the Triassic-Jurassic boundary or first occurring in the early Jurassic had notably divided and/or narrow leaf blades compared to those of the late Triassic flora, and this would result in lower leaf temperatures than if the leaves were undivided (McElwain et al. 1999).

This next paragraph needs work - There are several hypotheses about the ecological preferences of early angiosperms. Most include the element that conditions were disturbed (e.g. Heimhofer et al. 2005; Berendse & Scheffer 2009 for a summary; Bond & Scott 2010; Boyce et al. 2010). Other than that, suggestions as to the conditions these plants faced range all the way from semi-arid (e.g. Stebbins 1965) or at least seasonally arid (Bond & Scott 2010) to aquatic or marsh-like, the latter being inhabited by nymphaealean-type plants (e.g. G. Sun et al. 2008), although it seems to me likely that both very xeromorphic plants and plants with an aquatic habit are likely to be derived. Conditions may be well lit (e.g. Royer et al. 2010) and/or drier (Bond & Scott 2010). Quite recently it has been suggested that earliest angiosperms were smallish, sympodial tropical trees that were rather tolerant of shade and disturbed conditions and that grew in humid conditions (Feild & Arens 2005). Since early angiosperms were probably woody, climatic niche evolution that occurred may have been slow, i.e., habitat evolution was slow (Smith & Beaulieu 2009). Ancestral angiosperms may have had chloranthoid teeth and guttated, but are likely to have had low drought tolerance (Feild et al. 2011a, c).

Thus the evolution of angiosperms has occured against a back-frop of changing - mostly decreasing - Co₂ concentrations. Plants with more sclerophyllous leaves may be favoured in conditions of high Co₂ concentrations because increasing concentrations cause a disproportiately large reduction on resistance to diffusion through the mesophyll compared to plants with thinner leaves (Niinemets et al. 2011).

Venation density, photosynthesis, and climate. There was a 3-4^o temperature increase at the Triassic-Jurassic boundary associated with a major faunal extinction. The dissection of plant leaves also increased then, and this allowed their heat load to be reduced (McElwain et al. 1999; Beerling & Berner 2005). The climate in the Late Jurassic-Early Cretaceous was dry - certainly Pangea had a notably dry interior - but continents were drifting apart, carbon dioxide concentrations were still high, and sea levels were rising. There were some areas with ever-wet tropical humid climates, although they were initially rather restricted (see also Boyce et al. 2009, 2010; Boyce & Lee 2010; Feild et al. 2009a). CO₂ concentrations were perhaps 3,500-5,000 ppm in the early Cretaceous 135-100 million years ago, since when they have declined - with the odd hiccup (Beerling & Franks 2010) - perhaps with a particularly abrupt decrease in the middle of the Cretacous (Feild et al. 2011b; Barcley et al. 2010). Thus concentrations briefly spiked at a high of over 1,200 ppm at the PETM (Paleocene/Eocene thermal maximum), but by the recent glacial periods they had dropped to 180-190 ppm, about as low as they have ever been during the whole period of land-plant evolution (Zachos et al. 2008; Gerhart & Ward 2010; Boyce et al. 2010). Interestingly, leaves of at least some species of angiosperms seem to be adapted to life in a low CO₂ environment, stomatal closing occuring at about 600 ppm, whereas in other land plants the stomata remained open (Brodribb et al. 2009), and a low CO₂ environment would have favoured increased venation density (Feild et al. 2011a and references).

During the 200+ million year existence of land plants with a dominant sporophytic generation prior to the evolution of flowering plants, venation density held largely constant despite considerable fluctuations in atmospheric carbon dioxide concentrations (e.g. Boyce et al. 2010; Lee & Boyce 2010), although stomatal density fluctuated inversely to changes in CO₂ concentrations. Venation density in non-flowering plants continued to hold steady through the Cretaceous (Feild et al. 2011b). However, the venation density of angiosperms has increased very considerably over that of other vascular plants. This increase is less evident in members of the ANITA grade (see above), which also show lower CO₂ exchange than most magnoliids and basal eudicots (Feild et al. 2011a), and also Chloranthaceae. The vention density increase dramatically reduced the the main element in the resistance to water flow through the plant, passage through the mesophyll (Sack & Holbrook 2006). This increased vein density came at a cost of increased carbon allocation, although this was partially mitigated by vein tapering and the high density of minor veins (McKown et al. 2010; Beerling & Franks 2010). However, it allowed an increase in stomatal conductance and this can be linked to a higher maximum photosynthetic capacity - a some 174% increase in maximum photosynthetic CO₂ uptake over other land plants (e.g. Brodribb et al. 2007; Brodribb & Feild 2010). Interestingly, there can be a 50% decrease in the water potential before there is stomatal closure, so allowing maximum leaf hydraulic conductivity to persist, whereas in ferns stomatal closure occurs before there is any decrease (Brodribb & Holbrook 2004). Overall, dense veinlet reticulum is correlated with higher rates of photosynthesis and transpiration, more tardy stomatal closure because hydraulic failure as water potential decreased was less easy, etc. (Feild et al. 2011a; for vein architecture, see Roth-Nebelsick et al. 2001; for leaf hydraulics, see McKown et al. 2010). The increased transpiration would also promote evaporative leaf cooling (Hetherington & Woodward 2003; Boyce & Lee 2010), perhaps important at times like the PETM. Could there be a link between details of xylem element evolution and increased transpiration?

Increased transpiration in turn quite possibly helped to drive the spread of widespread tropical forest with reliably high rainfall (e.g. Boyce et al. 2008, 2009, 2010). Simulations in which the Amazon rain forest is replaced with non-angiosperm vegetation decreased the extent of ever-wet rainforest there by about 80% (Boyce & Lee 2010; Lee & Boyce 2010). As Boyce et al. (2010) note, the extent of rain forest in other parts of the tropics shows less change in such simulations, but this might have been different under conditions earlier in the Tertiary - for instance, the elevation of large areas of continental Africa had not yet occurred, and continents were in different positions. Much of woody angiosperm diversification has been in such rainforestss, and epiphytes, another appreciable component of angiosperm diversity, are abundant there (Feild et al. 2009a; Boyce et al. 2009, 2010; Boyce & Lee 2010). Note that epiphytes do not often form endomycorrhizal associations, although epiphytic Orchidaceae and Ericaceae have the mycorrhizae typical of their respective families (see Mycorrhiza 4(1). 1993 for papers). The spread of perhumid conditions in which large angiosperms dominate may have selected against drier, rather small stature angiosperm forest subject to frequent fires; fires were widespread in the Cretaceous, but decreased notably from the mid Palaeocene to mid Eocene (Bond & Scott 2010; Belcher 2010).

This raises the issue of when tropical rainforest as we know it appeared (see above). The venation density of ANITA-grade angiosperms is rather similar to that of non-angiosperms, but this density showed two marked increases in the Cretaceous, one around the Middle to Late Albian ca 196-100 million years ago - large trees first appear in the fossil record then - and one around the K/T boundary, when values reached 12-16 mm mm^-2 (Feild et al. 2011b).

Vascular evolution. Amborella of course lacks vessels, and it has been suggested that the acquisition of vessels in Nymphaeales may be independent of that in other angiosperms (e.g. Schneider & Carlquist 2009). The earliest angiosperms may have been smallish, sympodial tropical trees that were rather tolerant of shade and disturbed conditions and that grew in humid conditions (Feild & Arens 2005), and vessels may have evolved in such plants (Feild 2005). These conditions may be reflected in the ecological proclivities of living members of the ANITA grade (other than the aquatic Nymphaeales). Interestingly, indiividual vessels in both magnoliids and ANITA-grade angiosperms may be more effective in transmitting water than tracheids, but on a xylem cross-sectional area basis such vessel-bearing plants may have a hydraulic efficiency similar to that of their tracheid-bearing relatives (Sperry et al. 2006; Feild & Holbrook 2001; Hudson et al. 2010). The overall hydraulic efficiency of the tracheids in conifers is higher than might be expected despite their short length and their end walls because of the very low resistance to water flow of the bordered pits - the central torus may block the pit if needed, yet the fibrils in the margo are relatively widely spaced - compared with the pits in angiosperms (Pittermann et al. 2005; Sperry et al. 2007; Hacke et al. 2007; Hudson et al. 2010). Furthermore, the resistance to flow of the scalariform perforation plates found in early-evolving xylem is higher than had been earlier estimated (e.g. Christman & Sperry 2010). (Some palaeozoic medullosan seed ferns had long, wide tracheids that presumably had high water conductivities like those of some angiosperms with vessels [Wilson & Knoll 2010], but ferns in general have wide and long tracheids with surprisingly high transport rates [Pittermann et al. 2011].) Indeed, the vessels in these magnoliids and ANITA-grade angiosperms are rather different from those prevalent in core eudicots ("basal vessels", see Sperry et al. 2007), being short, not very dense, with scalariform perforations, etc.

Hence there are similarities in the pattern of evolution of xylem and that of venation density. In both, ANITA grade angiosperms are in some physiological respects more similar to gymnosperms and extinct seed plants than to other amgiosperms. Thus homoxyxlic vasculature (no vessels) was constant in lignophytes until the evolution of angiosperms (Taylor et al. 2009), and the same is true for venation that is not dense. In both cases, it is not the evolution of angiosperms that marks the main changes, but changes that occured later. Details of what drove these changes are unclear, indeed, as Wheeler and Baas (1993) noted, there is conflict in the Cretaceous between features of fossil woods and paleoclimatic indicators - but these indicators are based on our understanding of how wood of extant plants functions. Infering the past from the present may present difficulties.

Overall, the evolution of vessel elements with the water-conducting capacity of eudicots may have been a rather protracted process, and the initial acquisition of vessels may even have been of immediate value to the plants with them more because heteroxylic wood allows the specialization of cells in the vascular tissue for support, storage, etc., the heteroxylly [sic] hypothesis (Hudson et al. 2010 and references; see also Carlquist 2009 for extensive discussion on xylem heterochrony and its implications for angiosperm evolution). Vessel conductance increases substantially when the peforation plates become simple (see e.g. Christman & Sperry 2010; Hudson et al. 2010; Feild et al. 2011c). There seem to have been some rather abrupt changes in the general conductive efficiency of angiosperm woods across the K/T boundary, at least based on changes in vessel length and decrease in the percentage of woods with scalariform perforation plates; there were also some changes in wood parenchyma, fewer in rays (Wheeler & Baas 1991, 1993) - could this have been necessitated by the increased transpiration made possible by the dense veinlet reticulum unique to angiosperms? Veinlet density, too, increased subsequent to the initial appearance of angiosperms, with two particularly marked increases occuring in the Cretaceous - the second at the K/T boundary (Feild et al. 2011b).

Although comparisons of water flow in the xylem of angiosperms and that of gymnosperms are quite frequent, less seems to be known about the ecological significance of any functional differences in their phloem. Even if some differences between sieve tubes (angiosperms) and sieve cells (gymnosperms) may be somewhat over-emphasised - the nucleus in both may be non-functional, even if they differ in how they become non-functional - the fact remains that they have different sieve plate morphologies and occlusion mechanisms, and their ontogenetic/functional associations with neighbouring cells differ (e.g. see Behnke 1986; Schulz 1992).

Wood and litter decay and dissolution. Another element to consider is litter and wood breakdown, as well as their loss by burning (Cornwell et al. 2009). Factors like leaf mass per area (M_A, the inverse of specific leaf area [SLA], leaf area to dry mass, cm² g–1)and primary and secondary venation type have vatriously been linked with rates of photosynthesis, plant growth, litter decay, nitorgen content, and nutrient cycling have all been linked, although there is much within-community variation in such features within angiosperms and the phylogenetic signal in such correlations needs to be better understood (e.g. Cornwell et al. 2008; Wieder et al. 2009; Walls 2011). There is likely to be much local heterogeneity in decay rates depending on the decay organisms present, the tree species present, and even the age of the tree (Weedon et al. 2009). Interestingly, there is a negative correlation between litter longevity and silcon concentration in tissues; Poales, and especially Poaceae, have a high silcon concentration, and this is especially high in the annual species (Cooke & Leishman 2011b).

Low leaf mass/area and high amounts of nutrients in litter are both implicated in speedy litter breakdown. Both Cretaceous (Potomac, 110-105 million years ago: Royer et al. 2007) and Tertiary (Eocene, 49-47 million years ago: Royer et al. 2010) angiosperm floras had a low leaf M_A, under 100g/m²; note that the three gymnosperm plants in the former flora had a mean of 291 g/m². Even contemporary tropical non-riparian lowland rainforest may have only a moderate M_A, e.g. ca 198 g/m², as on Barro Colorado Island (Royer et al. 2010), and extant gymnosperms, like the extinct gymnosperms just mentioned, have a lower SLA than do angiosperms (Berendse & Scheffer 2009 and references). However, M_A per se is unlikely to have been a major factor, since angiosperm woods are on the whole denser than gymnosperm woods, yet they, too, decompose faster (Weedon et al. 2009).

Decay is affected by the composition of the structural elements of plants. The lignin content of angiosperms may be 20% lower than that of gymnosperms (Robinson 1990), indeed, denser gymnosperm woods have proportionally still more lignin and proportionally less nitrogen, while the reverse relationships hold in angiosperms (Weedon et al. 2009). In general, both lignin and polysaccharide content are negatively correlated with the rate of litter breakdown (Cornwell et al. 2008); for the decomposition of lignocellulosic compounds, see Martínz et al. (2005). Note that the syringyl-rich lignins that characterise many angiosperms are more easily decomposed by fungi than the guaiacyl-rich lignins of other seed plants (Ziegler et al. 1985). Angiosperm leaves, litter and wood have more nitrogen and phosphorous (on a %age basis) than do those of gymnosperms (Cornwell et al. 2008; Weedon et al. 2009). Although the overall patterns of wood decomposition may be clear, they vary in detail, and the particular factors reponsible for the patterns are not understood (Weedon et al. 2009); the chemistry of woods is complex (e.g. Kögel-Knaber 2002).

In any event, Litter from extant ferns and fern allies and bryophytes is slow to decompose compared to that of gymnosperms and especially angiosperms (Cornwell et al. 2008). The litter of deciduous angiosperm trees decomposes faster than that of evergreens, angiosperm wood faster than gymnosperm wood, and the litter of angiosperm forbs in particular decomposes faster than that of any other group of land plants (Cornwell et al. 2008; Weedon et al. 2009) - note that these more easily decomposable leaves are likely to be more palatable to insects, too (see below). Interesting, graminoid litter - presumably mostly Poaceae and Cyperaceae are involved - decomposes more slowly than the litter of forbs (Cornwell et al. 2008). It is well known that graminoid lignin is somewhat different in composition from other lignins, having an appreciable amount of p-hydroxyphenyl units and being low in nitrogen (e.g. Wedin 1995). UV light also decomposes lignin (Austin & Ballaré 2010). Finally, mean annual precipitation - which, as we have seen, can be affected by angiosperm transpiration - and temperature are also positively correlated with litter and wood turnover and so with the release of the nutrients they contain (Yin 1999; Weedon et al. 2009; Wieder et al. 2009).

Low rates of litter decay (accompanied by high M_A values) are features of those plants, gymnosperms and some angiosperms, that are able to grow in stressful, nutrient-poor environments (Berendse & Scheffer 2009). Extant vascular plants other than angiosperms tolerate nutrient-poor but mineral-rich conditions, perhaps the ancestral condition for many of them (Page 2004). Angiosperms grow faster, and the nutrients that they need will be released by the decay of their litter, and the disturbed habitats that angiosperms may initially have favoured are also likely to be associated with elevated levels of nutrients (Berendse & Scheffer 2009). Since the high photosynthetic rates of angiosperms allow higher growth rates, these may allow angiosperms to immediately utilize any flushes of nutrients produced by litter and wood breakdown, effectively scavenging available nutrients (Berendse & Scheffer 2009). Litter decay of course results in the release of large amounts of CO₂, angiosperm growth removes CO₂. Jaramillo et al. (2010 and references) note that high levels of CO₂ and high levels of soil moisture improve the performance of plants when temperatures are high, as earlier in the Tertiary, so perhaps together being an element in the success of angiosperms.

The importance of fire should not be overlooked. Brener (2003) notes that rocks rich in organic matter derived from plants, of which substantial amounts may be inertinite, charcoal from fires (Scott & Glasspool 2006), are particularly prominent in the Mid Cretaceous 120-90 million years ago to the Palaecene, again in the Oligocene ca 30 million years ago, and fires increased again in the last 10 million years (Bond & Scott 2010; Belcher et al. 2010). Fires cause the release of CO₂ into the atmosphere, but on the other hand the inertinite produced is highly resistant to decay, so causing carbon sequestration.

Interestingly, pioneer plants - perhaps including early angiosperms - may be able to tolerate high herbivory because they have metabolically cheap, rapidly expanding leaves with a low amount of fibres and low concentrations of secondary metabolites like terpenoids, phenols, and tannins; their high quality habitat allows rapid growth and low defence (see e.g. Bond & Scott 2010). Indeed, deciduous plants in general, with their rather thin and short-lived leaves, tend to be eaten by insects more than plants with long-lived xeromorphic leaves (e.g. Coley & Barone 1996; Arnold et al. 2001; Wilf et al. 2001; Lewinsohn et al. 2005). But even tropical, largely evergreen, non-riparian lowland rainforest, as on Barro Colorado Island, is subject to insect damage (Royer et al. 2007). Floras from immediately after the bolide impact in the western interior United States, some ca 64 million years before present (early Palaeocene) had low plant diversity; the species were deciduous, with thin leaves and presumably low defences, and there were several kinds of insect damage, while in others, more diverse and with the facies of tropical rainforest, the species having tough, thick, and presumably tanniniferous leaves, and there were fewer kinds of herbivore damage (Wilf et al. 2006). There are also correlations with temperature, with more herbivory occurring at higher temperatures, as is evident at least locally in the fossil record (Wilf & Labandeira 1999).

All this is compatible with the general thesis that angiosperms seem to be responsible in part for the diversity, extent and perhaps even the very existence of tropical rainforests in the Tertiary. Indeed, Hoorn et al. (2010) recently suggested a role for climate in sustaining, and perhaps also driving, diversity - they gave as an example the wet, less seasonal western Amazonian rainforest - and the increased plant productivity and diversity allowed animals that ate, pollinated or dispersed angiosperms to diversify (see also Boyce et al. 2010).

Mycorrhizae and their associates. But this is just the beginning. Add to this the interaction of plants with their fungal associates, both ecto- and endomycorrhizal, the bacteria associated with these mycorrhizae, the evolution of lignin-decomposing fungi, and the effect of all these on mineral weathering in rocks, soil structure and (again) on carbon and nutrient cycling, and one can set up another distinctive physiological-ecological context in which to consider angiosperm evolution (e.g. Taylor et al. 2009; see Beerling 2005a for vascular plants in general).

Ectomycorhizal associations may be of major importance for angiosperm evolution not only because of their direct interactions with plants in facilitating nutrient and water supply (see above), but also more indirectly because of their effect on the soil, weathering, and hence on the earth's climate. To a certain extent their activities seem to work counter to the increasing nutrient availability enabled by the ready decomposition of angiosperm wood and litter just discussed. Some plant groups are particularly well known as being ectomycorrhizal, and these groups, which include Pinaceae, Dipterocarpaceae and relatives, some Fabaceae, members of Fagales, etc., can all dominate the communities in which they grow (see Hart et al. 1989 for some monodominant communities in the tropics; most of these are ectomycorrhizal), and many of them are also mast fruiters (e.g. Richards 1996). Ectomycorrhizae are commonly found in plants growing on rather extreme soils, including serpentines (Branco & Ree 2010), and in general, the soils that support communities dominated by ectomycorrhizal species are either poor in nutrients and/or rich in organic materials and/or without much other vegetation. In Boletales, at least, ectomycorrhizal associations may have evolved in the ecological context of soils low in nitrogen (and high in organic material) as the result of activities of brown rot fungi, a number of which are also found in Boletales (and other fungi: Eastwood et al. 2011). Ectomycorrhizal plants are often pioneers, and ectomycorrhiza-dominated plant communities are especially prominent in (cool) temperate and tropical montane habitats, but also in some tropical lowland forests (e.g. Malloch et al. 1980; Safer 1987; Hart et al. 1989; Sanford & Cuevas 1996). In such communities there is often a decrease in the pH of the soil. CO₂ sequestration may occur because the development of sometimes massive amounts of mor humus is common, especially in cooler climates, but even in the wet tropical West Malesian forests there are huge peat lenses on which ectomycorrhizal dipterocarps may dominate (Richards 1996); podzolization may also occur (van Schöll et al. 2008), and in general the acid, nutrient-poor conditions that develop are not conducive to the activity of potential decomposers of humus - a self-reinforcing cycle.

There is also an increase in mineral weathering faciltated in part by siderophores, chelating agents, and organic acids produced by the ectomycorrhizae and their bacterial associates and by the substantial amounts of CO₂ produced by the respiration of the association that is then used up in rock weathering (Beerling 2005a; Taylor et al. 2008). The basic equation for weathering is 2CO₂[atmosphere, in soil from respiration] + H₂O[from precipitation] + Ca/MgSiO₃[continent] -> SiO₂[continent + ocean] + 2HCO₃^- -> Ca⁺⁺ + 2HCO₃^- -> CaCO₃[ocean sediments] + CO₂ + H₂O, or overall, this can be represented as CO₂ + Ca/MgSiO₃[continent] -> SiO₂ Ca/MgCO₃[ocean] (e.g. Berner 1999; Beerling 2005a; Taylor et al. 2009). Low molecular weight organic acids produced by ectomycorhizae can mobilize cations such as Ca⁺⁺ and Mg⁺⁺, increase phosphorous availability, and oxalate, etc.; they may also form complexes with aluminium ions, detoxifying the aluminium but also increasing the weathering of aluminium-containing minerals in rocks (Landeweert et al. 2001; van Schöll et al. 2008). It is not for nothing that ectomycorrhizal fungi have been dubbed "rock-eating fungi" (Jongmans et al. 1997).

So answering the questions, "how many times did ectomycorrhizal associations develop, and when did they become common?", is important (see also Eastwood et al. 2011). Brundrett (2009) lists seed plants known to have ectomycorrhizal associations (see also Smith & Read 2008): Only perhaps 2,500-3,000 species of vascular plants are ectomycorrhizal, but ericoid mycorrhizae account for another 4,000 species and orchids over 22,000 more. In angiosperms, ectomycorrhizal associations have formed perhaps some 35 times, and there are over 40 families of angiosperms involved (Bruns & Schefferson 2004, Wang & Qiu 2006; Smith & Read 2008); most associations are with rosids. In fungi, the ability to form ectomycorrhizae has evolved at least 6 to 8 times (Hibbett & Matheny 2009). Compared to their hosts, many species of fungi form ectomycorrhizal associations, estimates being ca 5,000-6000 described speciesbut probably many more (Molina et al. 1992; Blackwell 2011), and impressive single-site diversity of fungi has been documented (Horton & Bruns 2001 - most examples here are from Pinaceae-dominated forests), with individual species of plants forming associations with several species of fungi (Molina et al. 1992). Forests where ectomycorrhizal associations are common are, however, not very diverse when it comes to angiosperms, perhaps because of the dearth of readily available nutrients (Taylor et al. 2009 for literature; see also above).

The basidiomycete clades involved in these associations are older than the plant clades (Hibbett & Matheny 2009). However, suggestions that ectomycorrhizal associations in Dipterocarpaceae and Fabaceae-Amherstieae developed before the break-up of Gondwana over 130 million years ago (Henkel et al. 2002; Moyersoen 2006) may be overestimates, although there are massive amounts of dipterocarp resin in India much later in the Early Eocene, some 52-50 million years ago (Rust et al. 2010). However, estimates of the age of Fagales, in which the ability to form actomycorrhizae may well be an apomorphy, are in the region of perhaps a little more than 100 million years, Albian in age (e.g. Cook & Crisp 2005; Friis et al. 2006a; Wang et al. 2009; Magallón & Castillo 2009), while Pinaceae, also commonly ectomycorrhizal, may be some 200-350 million years old (see Eckert & Hall 2006). There are suggestions that ectomycorrhizal fungi themselves first diversified in the Cretaceous, but perhaps especially in the Tertiary 60-25 million years ago (Bruns et al. 1998; Horton & Bruns 2001). Interestingly, Normapolles-type pollen and macrofossils associated with it have been linked uniquely to Fagales, but only to families other than Fagaceae and Nothofagaceae. This pollen was abundant and diverse in the Late Cretaceous fossil record in the area from eastern North America to west central Asia (e.g. Friis et al. 2006a, 2010b and references), elsewhere in the Northern hemisphere pollen named Aquilapollenites and Wodehouseia - affinities uncertain - predominated, in tropical Gondwanan areas pollen of Palmae was common, while in the south Nothofagites pollen was to be found (Nichols & Johnson 2008 for a summary). If the present and past are connected, Normapolles plants were probably ectomycorrhizal, Nothofagus, also Fagales, is ectomycorrhizal; if Normapolles plants were indeed abundant in the Late Cretaceous, they may have had a transformative effect on the environment.

It is not simply that ectomycorrhizal plants incease mineral weathering, but more rainfall in general also allows more silicate weathering, and this is a principal sink for atmospheric carbon dioxide (Boyce et al. 2010; Berner, 1999). in drier years, there may even be competition between ectomycorrhizae and lignin-decomposing fungi for water, leading to a reduction in the rate of wood decomposition (Koide & Wu 2003). Thus the sequestration of carbon in non-decomposing biomass (and in the plants themselves) that may become buried in sediments, and an increase in the amount of atmospheric CO₂ removed by the weathering of rock, may be connected with the decrease in atmospheric CO₂ concentration during the Tertiary (Pagani et al. 2009; Taylor et al. 2009).

In forests with endomycorrhizal associations fewer species of fungi are involved and the habit evolved perhaps once, the 200-250 species of Glomeromycota (Blackwell 2011) are all involved in mycorrhizal associations. Extant forests made up of endomycorrhizal trees tend to be more diverse than forests with ectomycorrhizal trees (Malloch et al. 1980; McGuire 2007). How many times the endomycorrhizal association evolved in land plants is unclear. Although it is likely embryophytes and fungi established associations, initially with the gametophytes of the former, very early in the Silurian/Devonian (Selosse & Tacon 1998; Redecker et al. 2000b; Nebel et al. 2004; Köttke & Nebel 2005), mosses in particular usually lack mycorrhizal associations (Read et al. 2000; Kottke & Nebel 2005; Duckett et al. 2006b; Ligrone et al. 2007; Wickett & Goffinet 2008; Stenroos et al. 2010; Pressel et al. 2008, 2010 for literature). Relatively few plants form both ecto- and endomycorrhizal associations (Molina et al. 1992 for some examples). The majority of angiosperms are endomycorrhizal, and the mycorrhizae have substantial beneficial effects on soil structure (Taylor et al. 2009) - improving drainage, and hence weathering, and also playing an important role in phosphorous uptake, effectively by scavenging for it more efficiently (S. E. Smith et al. 2011).

There is evidence that fungi that become associated with seeds affect seed germination and survival (U'Ren et al. 2009). Details of the roles these fungi play are unclear, and they seem to be more related to pathogens and endophytes than to saprotrophic fungi.

Finally, many species of fungi are involved in endophytic associations with plants. Endophytic fungi are probably to be found in all seed plants (Rodriguez et al. 2009; Hoffman & Arnold 2010; Friesen et al. 2011), perhaps particularly in Poaceae (q.v.) and Ericaceae (e.g. Petrini 1986 and other references in this volume; Saikkonen et al. 2004 and references). However, details of any advantages accruing the parties involved are for the most part unclear (e.g. see Jumpponen 2001: dark septate endophytes; Rodriguez et al. 2009: summary), although a number of Poaceae have been shown to benefit from the association. Endophytic associations are probably at least intermittently mutualistic (Carroll 1988, 1995); they may often be involved in facilitating stress tolerance in plants (Rodriguez & Redman 2008), and they also can affect various aspects of root growth (Sasan & Bidochka 2012). Van Bael et al. (2009) recently found that leaf-cutting ants seem to dislike plants that have numerous endophytes, and at the same time there are nitogen-fixing bacteria (esp. Klebsiella that are integrally involved in this system (Pinto-Tómas et al. 2009). In general, the line between mutualism - or at least prolonged symbiosis - and parasitism is a fine one (Eaton et al. 2010 and references).

C₄ photosynthesis, grass, and grasslands. The different ways in which angiosperms fix carbon are quite well understood (Keeley & Rundel 2003). C₄ photosynthesis is especiallly common in Poaceae (Sage et al. 2011); ca 50% of the 5,250+ species of the PACMAD clade of Poaceae alone have C₄ photosynthesis. However, the driver that promotes/promoted the evolution of this distinctive photosynthetic pathway is unclear (see Retallack 2001; Westhoff & Gowik 2010 for general literature). In addition to the evolution of C₄ photosynthesis several times in Poaceae, it has evolved more than sixty times in other clades, too (Sage et al. 1999, 2011; Ludwig 2011b), as in many species of Cyperaceae, some Amaranthaceae (ca 800 species are involved) and other core Caryophyllales, Euphorbiaceae (Euphorbia subg. Chamaesyce), etc. (Arakaki et al. 2011). As Brown et al. (2011: p. 1438) note, "functionally equivalent mechanisms that control the accumulation of proteins important for C₄ photosynthesis" have evolved in parallel in Cleome gynandra and in maize, and origins in monocots and eudicots are roughly contemporaneous, all occuring within the last 30 million years or so (Christin et al. 2011b, q.v. for a number of dates); ca 30 million years is the time of a possible rapid reduction in atmospheric CO₂ concentration (Pagani et al. 2005: there are some suggestions that C₄ photosynthesis persisted through the Mesozoic - Keeley & Rundel 2003 for literature). Although overall only somewhat over 2% of angiosperms are C₄ plants - recent estimates suggest that a total of only some 6,000-6,500 species are involved (R. F. Sage, pers. comm.) - yet they account for about 18-21% of terrestrial gross primary productivity, not to mention 14/18 of the world's worst weeds (Lloyd & Farquhar 1994; Ehleringer et al. 1997; Retallack 2001) - which also seem to be notably facultatively mycorrhizal (Safer 1987). C₄ photosynthesis is very efficient, especially in monocots (Braütigam et al. 2008); C₄ monocots do better in warmer environments, while C₄ eudicots are found in some combination of arid, ephemeral, disturbed and/or saline conditions (Ehleringer et al. 1997). C₄ photosynthesis is thus efficient, highly polyphyletic, showing parallelisms even at the molecular level (e.g. Bläsing et al. 2000; Christin & Besnard 2009), and also recent, having evolved within the last 30 million years (Christin et al. 2011b).

C₄ photosynthesis in grasses seems to have originated in the Oligocene about 25 million years before present, perhaps - it was initially thought - in response to declining CO₂ concentration in the atmosphere, and these were perhaps the result of the carbon-sequestering activities of ectomycorrhizal angiosperms (Taylor et al. 2009). High temperatures may also have help spurred the change, and more recent work emphasizes the importance of heat, drought, fire and water stress (Edwards & Still 2008; Edwards 2009; Strömberg & McInerney 2011; Christin et al. 2011b; see also Retallack 2001). The great expansion of C₄ grassland beginning in the Miocene a mere nine million years ago and completed only 3-2 million years ago (e.g. Strömberg & McInerney 2011; McInerney et al. 2011 for North America) may be due to some of these and/or associated environmental changes like accelerated fire cycles, etc. (Retallack 2001; Sage & Kubien 2003; Tipple & Pagani 2007; Christin et al. 2008; Vicentini et al. 2008; Bond & Scott 2010; Arakaki et al. 2011).

Grasses in particular - and often only a few species - dominate large areas of prairie and savannah habitats, an estimated one fifth of the earth's surface being covered by grasslands (Hall et al. 2000). In the later Tertiary the spread of grasslands interacted with animal evolution and soil and environment change, and the result was a net draw-down of CO2

Retallack (2001; most of the details below come from this paper) suggested that short sod grassland with shallow soils appeared in the early Miocene in dry regions with 400 mm annual precipitation ca 20 million years ago. The PACMAD-BEP clade, i.e., the majority of grasses, are mostly plants of well lit environments, in contrast to more basal clades in Poales (Givnish et al. 2010b, see also other ecological correlates of this clade). Tall sod grasslands with deeper soils appeared in the late Miocene ca 7-5 milliion years ago in areas that had up to 750 mm annual precipitation - and these grasslands in particular would be mostly made up of C₄ grasses. These grasslands largely replaced forest/woodland vegetation, although McInerny et al (2011) suggest that the late Neogene expansion of C₄ grasses in North America was at the expense of C₃ grasses rather than woody vegetation. In general, erosion of crumb peds from grasslands leads to a loss of organic carbon in sediment that is an order of magnitude larger than the corresponding loss of carbon from forests - root systems in mature grasslands are dense and the soils are deep. Nutrients are also rapidly mobilized and ultimately lost in run-off and then support ocean productivity. The total C sequestration in soil and above-ground biomass in grasslands is greater than that of the forests they replaced - there is a shift in the proportion of the biomass sequestered from above-ground parts to the soil. Much C in consumed by weathering, grassland soils being notably moister than corresponding woodland soils even if woodlands transpire more and have a higher albedo - so grasslands support a drier climate, even if weathering is increased (Retallack 2001). Fires have increased in the last 10 million years (Bond & Scott 2010), perhaps because of the spread of easily flammable grasslands (Retallack 2001). Overall, grasslands can be considered a long-term carbon and water vapour sink, one one of the consequences of their activities is long-term global cooling (Retallack 2001), and they may be particularly sensitive to changing climates (e.g. Hall et al. 2000; Knapp & Smith 2001).

Furthermore, Poaceae, apparently alone in flowering plants, acquire iron through chelation of ferric ions with siderophores which are then taken up by the roots (Schmidt 2003; Kraemer et al. 2006). Yet again, mycorrhizae may also be involved: C₄ grasses have roots with long root hairs, yet may respond positively in terms of phosphorous uptake when forming endomycorrhizal associations - long root hairs and endomycorrhizal associations tend to be thought of as alternative ways of securing phosphorous supply, etc. (Schweiger et al. 1995 and references). There is more discussion on the diversification of Poaceae below.

Crasssulacean acid metabolism (CAM) photosynthesis is a taxonomically more widespread photosynthetic variant particularly prevalent in clades like Crassulaceae, Bromeliaceae, Cactaceae and Orchidaceae-Epidendroideae that either grow in arid terrestrial environments or are epiphytes. Some 17,000 or more species may have CAM or its variants (Winter & Smith 1996b).

Next three paragraphs not yet integrated: The adoption of fructans as storage polysaccharide in Poaceae-Poöideae, many Asparagales and Asterales, etc., may also be linked with the ability of the plants involved to grow in seasonally dry or frost-prone climates outside the tropics (Hendry 1993; Hendry & Wallace 1993; Vijn & Smeekens 1999).Other variation in plumbing includes that in water supply to the flower. The large flowers of at least some of these taxa may get their water supply through the xylem, whereas smaller flowers, as in the core eudicots, may be hydrated primarily via the phloem (Feild et al. 2009a, b, but sampling).

There is also substantial variation in phloem pressure and phloem loading in angiosperms, and active phloem loading seems to be most common in the predominantly herbaceous asterid I + II clade; a variety of selective advantages for active loading can be suggested (Turgeon 2010b; Fu et al. 2011 - see Garryales page). Passive loading, with associated high sugar concentrations in leaf cells, may be correlated with the woody habit, and one kind of active loading involving the synthesis of raffinose family oligosaccharides is not connected with plant habit, but perhaps rather with climate (warmer: see Davidson et al. 2010). Interestingly, woody plants that have active phloem loading are usually asterids (e.g. Buddleja, Catalpa, Ilex, Syringa), while herbs with passive transport are members of the predominantly woody rosids (e.g. Rosaceae, Peaonia, Lythrum) - although Saxifraga, an herbaceous rosid, has active phloem transport, while the woody Cercis and Styrax also have active loading (Rennie & Turgeon 2009; Fu et al. 2011). However, the situation is a little confusing since the classifications of phloem transport types in Davison et al. (2011) and Fu et al. (2011) are somewhat different; I have not yet integrated them.

Both phytochrome genes duplicated before the origin of crown group angiosperms, and it has been suggested that PHYA in particular may have evolved and become sensitive both to pulses of light of very short duration (the very low fluence response - VLFR) such as sunflecks and to far-red light, generally enriched in low-light conditions. Thus PHYA is intimately involved in germination and in etiolation responses of the seedling, especially in shady conditions, and in germination and in etiolation responses of the seedling in low-light conditions such as occur on the forest floor (Mathews et al. 2003). Finally, genome size affects cell size and stomatal density - plants with larger genomes have larger and less dense stomata, although woody plants in general have small, dense stomata and low genome size (Beaulieu et al. 2008).

Some patterns of diversity in extant angiosperms.

A number of large clades (2,000+ species each) can be characterised by features that seem likely to affect diversification/speciation, although the lack of firm age estimates for branching points within these clades remains an obstacle in understanding their evolution. Five of these major clades have a preponderance of members with monosymmetric flowers - these are Orchidaceae, Zingiberales, Lamiales (at the lower of the nodes where Calceolariaceae or Gesneriaceae join), Fabaceae, and Asteraceae. Of course, reversion to polysymmetric flowers has occurred within these clades, perhaps most notably in Fabaceae-Mimosoideae. Euglossine bees and bumblebees (see above) in particular are attracted to monosymmetric flowers; Westerkamp and Claßen-Bockhoff (2007) consider such flowers to be the "ultimate response" to bees. Birds also visit monosymmetric flowers, and in the gullet-type ornithophilous syndrome stamens in the upper part of the mouth of the flowers deposit pollen on the head of the pollinator. Kalisz et al. (2006) suggest that the development of monosymmetric flowers may be linked to the evolution of dichogamy (separation of the time of pollen dispersal and stigma receptivity), in particular, to that of protandry. In fact, some kind of dichogamy is widespread, and poly- or asymmetrical ANITA grade angiosperms have protogynous flowers. Even if there are elements of common developmental mechanisms involved in independent acquisitions of monosymmetry (e.g. Feng et al. 2006, comparing Fabaceae and Plantaginaceae; Zhang et al. 2010, Malpighiaceae), duplication of CYC genes being involved, the plants in the major clades just mentioned are ecologically/functionally very different. Together with some rather smaller clades, e.g. Campanulaceae-Lobelioideae, Caprifoliaceae s.l., Lecythidaceae-Lecythidoideae and Iridaceae, these few clades comprise almost 1/3 of all angiosperm diversity.

Indeed, diversification in many clades with monosymmetric flowers seems to be greater than that in their sister taxa with polysymmetric flowers (Sargent 2004 [some comparisons need to be reworked], cf. Key et al. 2006, but see Kay & Sargent 2009), perhaps because pollinator fidelity is increased. However, a word of caution is in order. From the simply structural point of view, very many flowers are monosymmetric at some stage of their development (see Endress 1999, 201b, 2008a for discussion) and is unclear exactly what such crude categorisations as "flowers monosymmetric" mean functionally. Not only are many apparently polysymmetric flowers slightly monosymmetric, but how the pollinator approaches such flowers may lead to quite precise deposition of pollen on it, e.g. hummingbirds pollinating Aquilegia (Kay et al. 2006; see also Iridaceae below). Many highly reduced flowers are also monosymmetric, not only in Poaceae (see below), but also in the speciose Piperaceae, etc. Asteraceae are treated separately below because their small and quite often monosymmetric flowers are aggregated into functionally polysymmetric inflorescences.

There are five large clades with monosymmetrical flowers. 1. Orchidaceae (ca 20,000 species) are ground-dwelling or epiphytic and in part mycoheterotrophic herbs of small to moderate size producing as many as millions of tiny seeds per flower; the tepals are more or less free and the flowers, with their distinctive and complex morphologies centred on the gynostemium (stamens and stigma-style all congenitally fused) and labellum (median tepal of the inner whorl), are inverted. Insect behaviours involved in effective pollination are for the most part different from the pollen- and nectar-collecting behaviors of visitors to the other monosymmetric clades, deceit pollination being particularly common. Although the diversity of floral form in Orchidaceae is great, it is attained by variation on a rather limited basic floral theme. Diversification in the "higher epidendroids", the speciose, epiphytic Epidendroideae, may have occurred in the mid Eocene to Oligocene, some (64-)59-42(-36)/(49-)39-34(-22) million years before present, although diversification in the Epidendroideae clade itself (ca 18,000 species) began (72-)68-51(-44)/(62-)49-44(-29) million years ago (Ramírez et al. 2007; Gustafsson et al. 2010: calibration by fossils, so dates are minimums); there are more than twice as many epiphytic species in Orchidaceae than in all other families combined (Gentry & Dodson 1987). 2. Zingiberales (2,100 species) are quite large plants, mostly herbs, of the tropics with large flowers. Their fruits usually have only a moderate number of seeds that are mostly animal dispersed. Flowers in Zingiberales vary considerably in orientation, the parts that are petaloid, number of stamens, etc., and from this point of view are more variable than Orchidaceae. (Note that many taxa in Commelinales, sister to Zingiberales, also have monosymmetric flowers, and it is possible the the common ancestor of the two orders had such flowers.) 3. Lamiales (21,000 species) are more or less herbaceous plants perhaps particularly abundant oustide the tropics, although there are many tropical members which are often more or less woody. Individual flowers are moderate in size to quite large, each usually producing rather many and small seeds. Although clades like Lamiaceae produce only four seeds per flower, they are still quite small; in general dispersal is by wind. Most of the monosymmetric 4. Fabaceae (19,400 species, of which 3,300 are in the polysymmetric Mimosoideae) produce rather few and relatively large seeds in the single carpel of each flower. The flowers are inverted and often papilionaceous, i.e. they are keel flowers (see Stirton 1981; Westerkamp 1997 for keel flowers), the petals are more or less free, and the plants are either trees of more or less tropical forests, especially those of the neotropics, or herbs, found considerably more widely. Dispersal is either autochorous (ballistic) or animal-mediated. Many Fabaceae are nitrogen-fixers, and the family is noted for the diversity of secondary metabolites that it produces, sometimes in association with endophytic fungi.

Of the smaller but still quite large clades in which monosymmetry predominates, 5. Campanulaceae-Lobelioideae comprise some 1,200 species of laticiferous herbs or shrubs with slit-monosymmetric flowers and dehiscent fruits with many small seeds, while 6. Caprifoliaceae s.l., comprising some 850 species, has often rather weakly monosymmetric flowers and indehiscent fruits with at most few seeds. Ericales include 7. Lecythidaceae-Lecythidoideae, with some 200 species of trees that can be a prominent element in neotropical forests. There the polystaminate androecium alone is monosymmetric, the fruit is large, and the seeds are few and large. In 8. Iridaceae, one generally thinks of monosymmetry in connection with flowers of the speciose Gladiolus, for example. However, from the point of view of the pollinator the flowers of Iris, Moraea, etc., are also monosymmetric, even if they appear to us to be polysymmetric. Thus a single Iris flower consists of three strongly monosymmetric meranthia or part-flowers (see also Westerkamp & Claßen-Bockhoff 2007), and all told over half the family, 750+ species, may have functionally monosymmetric flowers. the fruits have moderately numerous seeds. 9. Polygalaceae, with some 1,050 species the majority of which are monosymmetric, are phylogenetically close to Fabaceae, although the exact relationships of the two remain unclear (see Bello et al. 2009). Their flowers are also papilionoid keel flowers and superficially like those of many Fabaceae, although different parts are involved (Westerkamp & Weber 1999). The fruits usually have few seeds.

The final clade with monosymmetric flowers to be mentioned is the insect-pollinated 10. Asteraceae (23,600 species). These are mostly herbaceous to shrubby plants with small, usually monosymmetric flowers that are aggregated into capitulae; only a single seed of moderate size is produced per flower. Dispersal is often by wind. Each capitulum is functionally a single polysymmetric flower which produces quite numerous seeds. Asteraceae is another clade that is noted for the diversity of secondary metabolites it contains. Note that exactly where monosymmetry is an apomorphy in Asterales is unclear; it may not be an apomorphy of Asteraceae.

Diversification as a possible result of the aquisition of monosymmetry in particular can be studied on much finer evolutionary scales. Stebbins (1974) suggested that zygomorphy/monosymmetry had evolved more than 25 times within angiosperm families and Westerkamp and Claßen-Bockhoff (2007) noted that it was found in 38 families. In fact it has evolved perhaps two hundred times or more in angiosperms (see also Endress & Matthews 2006a; Endress 2008), although understanding just how many independent origins of monosymmetry there have been depends critically on the evolutionary assumptions we make (e.g. Endress 1997a; Donoghue et al. 1998; Reeves et al. 2003; Cubas 2004; Jabbour et al. 2008). Indeed, as we come to know more about floral morphology and development, a clear definition of monosymmetry itself has become elusive since it includes a variety of quite disparate morphologies (Endress 2008; see also the "Characters" page). Whole inflorescences may be the attractive unit from the point of view of the pollinator, and these, too, can be functionally equivalent to monosymmetric flowers, as in Euphorbia-Pedilanthus, some Proteaceae, etc., although such inflorescences usually have no particular symmetry signal or are functionally polysymmetric. Indeed, the large peripheral monosymmetric flowers in Asteraceae, some Brassicaceae (e.g. Iberis amara: Busch & Zachgo 2007) and Apiaceae accentuate the similarity of the whole inflorescence to a single, polysymmetric flower; the strongly monosymmetric peripheral flowers are the visual equivalent of petals.

Dioecious plants tend to be woody, wind pollinated and to have small flowers. If insect pollinated, the floral displays tend to be dimorphic, those of the staminate plants being showier and so more visited; extinction is thus perhaps more likely (but cf. Amborella!). Dioecious clades in general are not notably speciose (Heilbuth 2000; Vamosi & Otto 2002; Kay et al. 2006, etc.). Clades in which wind pollination predominates are also not notably speciose, even if they locally dominate ecologically, as with Fagales (see above and ectomycorrhizae); the adoption of abiotic pollination elsewhere is often associated with a decrease in speciation rate, e.g. Dodd et al. (1999); Friedman and Barrett (2008, 2011) provide useful surveys of the evolution of wind pollination in angiosperms, suggesting i.a. that there has been selection for numerous small carpellate flowers, and this to a certain extent incidentally also involves a reduction in ovule number per flower - most wind-pollinated taxa have few to a single ovule per flower. A clear exception is the very speciose Poaceae, 10,050 or so species of largely herbaceous wind-pollinated plants with a single-seeded fruit (note that in most species the flowers can be categorized as being monosymmetric!). Finally, Cyperaceae-Juncaceae, another clade of herbaceous, wind-pollinated plants also often with single-seeded fruits, contains about 4,800 species. Kay and Sargent (2009) noted that both these groups were about seven times more diverse than their animal-pollinated sister clades.

Even in the clades just mentioned, simply listing clades and characters may not be very helpful. Over four out of five Orchidaceae are Epidendroideae, in which epiphytes are common, almost three quarters of Asteraceae are members of the chemically very distinct Asteroideae, while within Poaceae the first three clades that are successively sister to the remainder of the family contain some 26 species out of the 11,000+ of the whole family, and these three "basal" clades are forest plants (see e.g. also Givnish et al. 2010b; S. A. Smith et al. 2011, but treat examples with care). One needs to know exactly when diversification occurred in the clade, thus in clades like Halenia (Gentianaceae) with a "key innovation" of five nectar spurs (von Hagen & Kadereit 2003), diversification and acquisition of these spurs are not simply linked (see also above, Lupinus, Valerianoideae, etc.).

For Poaceae enough is known to show how complex understanding diversification can be. For example, Poaceae-Poöideae are noted for their association with endophytes, an association that could be ca 40 million years old (Schardl et al. 2004). The presence of some of these endophytes affects the palatability of grasses to herbivorous mammals and of their seeds to granivorous birds because of the metabolites produced by the fungi, animals eating the infected material sometimes not thriving at all. The level of aphid infestation and that of their parasites and parasitoids, and even the pattern and rate of decomposition of dead grass, are also affected (e.g. Madej & Clay 1991 - birds; Omacini et al. 2001 - aphids; Lemmons et al. 2005 - decomposition). A variety of alkaloids, including loliine (pyrrolizidine) and ergot alkaloids, are produced by the fungus; the distinctive loliine alkaloid is primarily active against insects (Schardl et al. 2007). Furthermore, the larvae of Phorbia (or Botanophila) flies eat the stroma (a mass of spore-bearing fungal tissue) of the endophyte Epichloë, and the adults transmit the spermatia in a fashion analogous to insect pollination of flowers, since the spermatia are associated with a sugary secretion of the stroma (Bultman 1995). Various aspects of root growth may also be affected (Sasan & Bidochka 2012). Large numbers of other apparently symptomless endophyte species may grow together on Poaceae, but little is known about their real effect on the plant, and the line between a balanced association and parasitism may be easily crossed (Eaton et al. 2010). For fungal records - very numerous and diverse - on grasses, see Tang et al. (2007); there are at least 1933 species of endophytic fungi known from bamboos alone. Relationships between the grass and the endophyte can be complex. Thus Márquez et al. (2007) found that only when the endophytic fungus (Curvularia) was infected with a virus was Dicanthelium lanuginosum, the host of the fungus, able to grow in volcanically-heated soils. Grasses are well known for the diversity of silica bodies in their leaves, and these play a role in protection against herbivory, while silicon concentration itself is correlated with the rate of breakdown of plant tissues, and so nutient cycling (see silica).

A number of features are clearly not associated with individual clades, but are common in flowering plants. Thus succulence of some form, whether of root, stem or leaf, occurs in some 690 genera and 12,500 species (Nyffeler & Eggli 2010b; see also von Willert et al. 1990; Eggli & Nyffeler 2009), and include species which either avoid drought - they are rarely found in the driest conditions - or are salt tolerant, and these are usually mutually exclusive strategies (Ogburn & Edwards 2010). CAM-type photosynthesis is particularly prevalent in clades that either grow in arid terrestrial environments or are epiphytes (succulent epiphytes are quite often CAM-type plants); some 17,000 or more species may have CAM or its variants (Winter & Smith 1996b; Sayed 2001).

At another level, it has been suggested that species-rich clades and genome duplication are associated, Soltis et al. (2009: p. 336) linking "a dramatic increase in species richness" in Poaceae, Fabaceae, Brassicaceae and Solanaceae with genome duplications there, although as van de Peer et al. (2009a) point out, duplication and diversification may not be particularly closely linked in time. Duarte et al. (2010) and others also emphasize the amount of genome/gene duplication (and gene loss) there has been in angiosperm evolution. Fawcett et al. (2009) have dated a series of genome duplications within angiosperms to about 70-57 million years ago, around about the time of the bolide impact, suggesting that these palaeopolyploids were at a selective advantage because of their hybrid vigour, also having extra genes/alleles available for selection (see also Crow & Wagner 2006 and references; van de Peer 2009a; Franzke et al. 2011). Such genome duplications may reduce the probablity of extinction by increasing environmental tolerance and genetic variation, while individual mutations would be less likely to have an immediate effect (note that Wood et al. 2009 suggest that polyploidy affects primarily cladogenesis, less the diversification rate of polyploid clades). Indeed, immediate connections between gene and/or genome duplication and diversification of angiosperms are unclear, and it has been suggested that polyploidy is most often an evolutionary dead end, with recently-formed polyploid plants speciating less and in particular showing higher extinction rates than diploids (Mayrose et al. 2011).

However, species number is only one way of thinking about "success" in evolution. I finish by thinking about aymmmetries in ecological interactions. These include 1, the role of particular ecologically important but small (in number of species numbers, not neceassarily in biomass) groups of pollinators and seed dispersers for the community as a whole. These pollinators and dispersers - "generalists" from one point of view - all service a disproportionally large number of plants - "specialists". I then turn to 2, asymmetries in more physiological-ecological relationships, where relatively few species have a major effect on their communities/the environment as a whole.

• 1A. In the New World, the approximately 200 species of the notably long-tongued (the tongue is 15-42 mm long) euglossine bees, the orchid bees (Euglossini), are very important pollinators (Roubik & Hanson 2004). They are vigorous fliers and trap-line the plants they visit; these are often steady-state flowerers of the understorey (Janzen 1971; Ackerman 1985; Borrell 2005). Plant groups that have many species pollinated by euglossines include Araceae (e.g. the very speciose Anthurium and Spathiphyllum), Bignoniaceae, Gesneriaceae-Gesnerioideae, Lecythidaceae-Lecythidoideae, perhaps 2,000 species of Orchidaceae-Epidendroideae (here the males largely pollinate, looking for fragrances: Cameron 2004; Roubik & Hanson 2004; Ramírez et al. 2011; Schiestl 2012 - and references), and Zingiberales (especially Costaceae, Marantaceae), also Apocynaceae, Convolvulaceae, Fabaceae (including Swartzia, the flowers have dimorphic stamens and often only a single petal), and Rubiaceae. Attractants are various: nectar, pollen, and for the males, fragrances. Overall, these few euglossine bees are the major pollinators of well over 4,000 species of Neotropical plants (Wiehler 1976; Williams 1982; Ramírez et al. 2002 for a summary of the literature; Ramírez pers. comm.), and in any one community there may be up to 50 species of bees (Roubik & Hanson 2004; Zimmermann et al. 2009). It is estimated that crown group euglossine diversification occurred only 42-27 million years ago, montane clades diverging only in the last 8-4 million years (Ramírez et al. 2010).

• 1B. Bumble bees are more widespread. They seem to have diversified over a similar time frame, i.e. about 40-25 million years ago. The Eocene-Oligocene boundary of ca 34 million years ago was a time of sharp cooling, and bumblebees flourish in cooler climates, being facultatively endothermic (Hines 2008 and references). They are prominent pollinators in alpine environments, many species in large genera like Gentiana, Rhododendron and Pedicularis depending on them for pollination; they are also major pollinators in temperate and arctic regions, but are absent from Africa (the Sahara and southwards) and the Papua-Australia-New Zealand area (Goulson 2010). There are about 200 species of bumblebees that pollinate, and of course they can buzz pollinate, too (Goulson 2010). Bumblebees first moved into South America some 8-6 million years ago, perhaps along with plant genera like Rubus, Scutellaria, Lathyrus and Lupinus, all bee-pollinated and which have subsequently diversified substantially there (Asmussen & Liston 1998; Hines 2008).

• 1C. There are about 330 species of humming birds, all from the New World. The centre of diversity (ca 1/2 the species) for humming birds is the Colombia-Ecuador region. However, some are migratory, and these may pollinate ca 130 species of plants in western North America alone (Grant 1994). An additional ca 1,000 species of Ericaceae-Vaccinioideae and Gesneriaceae-Gesnerioideae may be pollinated by humming birds, and both these groups have their centres of diversity in the Colombian-Ecuadorean region along with their pollinating birds (Luteyn 2002; Weber 2010). Some 250 species of Salvia (Wester & Claßen-Bockhoff 2011), 500-600 species of Acanthaceae (E. A. Tripp and L. McDade, pers. comm.: also Tripp & Manos 2006), etc., are also bird pollinated, and the overall imbalance of pollinator species number:numbers of species of plants pollinated is probably similar to that of the euglossine bees just mentioned. The earliest humming birds known are from Europe in deposits ca 30 million years old (Mayr 2004). In the New World diversification seems to have begun in lowland South America, with much speciation about 13-12 million years ago along with the uplift of the Andes (McGuire et al. 2007 and references).

• 1D. Three small New World genera of fruit-eating phyllostomid bats are known to be particularly involved in the seed dispersal of ecologically important groups of plants. Carollia, with perhaps 8 species, prefers species of Piper, Artibeus, with perhaps 21 species, particularly Ficus and Cecropia, while Stumira, with about 13 species, is particularly fond of Solanum (see Lobova et al. 2009 for records). Major diversification of phyllostomid bats, which also include the major groups of nectarivorous bats in the New World, began 26-16 million years ago in the late Oligocene to mid-Miocene (Datzmann et al. 2010; Rojas et al. 2011).

The proper functioning of many plant-animal relationships of these kinds may be more important to the individual plant species than to the individual animal species. However, this by no means implies that that the morphologies of the organisms involved are "generalized". Individual interactions may be exquisitely precise, witness the deposition of pollinaria on a visiting euglossine bee by an orchid, and the morphology of the orchid flower. Indeed, there are paradoxes. Floral specialization has increased over evolutionary time; "specialized" monosymmetrical flowers in which precise interactions between plant and pollinator are needed for effective pollination have become more common. Yet over this same time bees, for example, have increasingly tended to pollinate a greater variety of hosts, but such polylectic bees have specialized morphologies. Plants with generalist-type flowers are visited more by oligolectic pollinators, but also by polylectic pollinators, plants with specialized flowers more by polylectic pollinators (see in part Stang et al. 2007). Overall, pollination efficacy in monosymmetric flowers is less than that of generalized flowers (Ramirez 2003). Such asymmetries might suggest that for polylectic pollinators the effect of the extinction of a single species of plant is likely to be slight, but the extinction of a pollinator may affect the plants it pollinates more seriously (Stang et al. 2007; Vamosi & Vamosi 2012; see also Colles et al. 2009). Rezende et al. (2007) suggested that extinctions may have some phylogenetic signal - but cf. Ramírez et al. (2011); however, this may depend on the extent of strict co-evolution.

2.

• 2A. C₄ photosynthesis, grass, and grasslands together make up a major component of global ecology. C₄ photosynthesis is especiallly common in Poaceae (Sage et al. 2011). Ca 50% of the 5,250+ species of the PACMAD clade of Poaceae alone have C₄ photosynthesis, but it has also evolved more than sixty times in other clades, too (Sage et al. 1999, 2011; Ludwig 2011b), particularly in many species of Cyperaceae and some Amaranthaceae (ca 800 species are involved), Euphorbiaceae (Euphorbia subg. Chamaesyce), etc. (Arakaki et al. 2011). Although overall only somewhat over 2% of angiosperms are C₄ plants - recent estimates suggest that a total of only some 6,000-6,500 species are involved (R. F. Sage, pers. comm.) - yet they account for about 18-21% of terrestrial gross primary productivity, not to mention 14/18 of the world's worst weeds (Lloyd & Farquhar 1994; Ehleringer et al. 1997; Retallack 2001). Grasslands in gneeral - i.e., including both C₃ and C₄ species - currently account for 11-19% of net primary productivity on land and 10-30% of soil C storage (Hall et al. 2000). To add further to this asymmetry, relatively few species of grasses are ecologically dominant in any one place, and most of these seem to be C₄ photosynthesizers (Edwards et al. 2010).

• 2B. To add - ectomycorrhizae ....

In Conclusion.

Gorelick (2001) summarizes some twenty hypotheses that have been advanced to explain diversification/success of the angiosperms (see also Crepet & Niklas 2009), and some 120 hypotheses have been advanced to explain the patterns of species richness that are such a distinctive feature of the environment (Palmer 1994). Work has tended to focus on understanding patterns of speciation within individual very speciose clades (e.g. Davies et al. 2004c), and much of the literature has emphasized the aquisition of "key innovations", an apomorphic feature of supposed functional and ecological advantage whose development allowed a subsequent increase in the overall speciation rate of the clade in which it arose (e.g. Marazzi & Sanderson 2010 for useful discussion). Clades in which there has been the acquisition of latex (Farrell et al. 1991; see also Powell et al. 1999; Agrawal & Konno 2009 for a survey of laticiferous plants and latex; Konno 2011 for some chemistry of latex), nectar spurs (Hodges & Arnold 1995; Hodges 1997; Kay et al. 2006), monosymmetric flowers (Sargent 2004, see also Kay & Sargent 2009; cf. in part Kay et al. 2006), hummingbird pollination (Schmidt-Lebuhn et al. 2007), or, more broadly, animal pollination (Eriksson & Bremer 1992; Kay et al. 2006b) or floral characters in general (Endress 2011a), or in which the climbing habit (Gianoli 2004) has evolved, are often more diverse in terms of extant species than their sister clades lacking these distinctive features. On the other hand, Malpighiales and Ericales appear to be disproportionately common among the small trees of the understorey of tropical rain forests (Davis et al. 2005a). They include taxa with many kinds of flowers and fruits, and monosymmetric flowers of a variety of morphologies are scattered in both clades. Neither clade can be well characterised either morphologically or chemically, and key innovations of any kind for these clades are hard to identify.

The study of key innovations is far more than simply linking a feature to a node (e.g. Donoghue 2005; Marazzi & Sanderson 2010). For instance, increase in speciation that results from the acquisition of a key innovation has to be distinguished from simple radiation of a clade when it moves into in a new area, even if allowing the plant with an innovation to move into new ecological space may be part of the definition of a key innovation (Marazzi & Sanderson 2010). Key innovations are also rarely simple features, rather, they may involve a complex suite of changes, however, such problems are just the tip of the iceberg (e.g. see Cracraft 1990), and determining that an innovation might be a key innovation is a difficult process (e.g. Sanderson 1998; Ree 2005b; Maddison et al. 2007). Even if a feature such as evolution of extra-floral nectaries in some Senna may be a key innovation, this does not mean that it always is; the loss of such nectaries may equally be a key innovation in related taxa (Marazzi & Sanderson 2010). As suggested above, in many very speciose clades, including angiosperms as a whole and appareently characterizable by "key innovations", patterns of clade numbers do not suggest any immediate diversification after acquisition of the putative key innovation. Thus in Poaceae the first three pectinations have a mere 4, 14 and 11 species respectively, compared with the over 11,000 species of the PACMAD clade, and similar patterns are found in Ericaceae, Orchidaceae, and even Asteraceae, while many other families have the basic phylogenetic structure of a very small clade being sister to a very much larger clade. Key innovations that cause the more or less immediate diversification of the clade with them may be individually less important than we might like to think (e.g. Davies et al. 2004a; Erkens 2007; Crepet & Niklas 2009). When thinking of diversification, it is best to focus on the shape of the phylogenetic tree, the timing of branching events, the numbers of species in clades, and the like - not names.

But species number is simply one estimate of success - there are other and perhaps ecologically more important ways to think about this. If one thinks about the evolution of general plant-environment relationships over the last two humndred milliion years, flowering plants look like ecosystem engineers. We also need to consider the implications of the asymmetry of many relationships between animals and plants. Species numbers then are not so easy to interpret. A conclusion that both sets of examples just discussed might suggest is that by focusing on the construction and maintenance of the ecological scaffolding of community structure, albeit in a phylogenetic context, one can think of diversity more as the paintings in spandrels (cf. Gould & Lewontin 1979). These paintings are forever changing as specialized plants go extinct due to the breakdown of relationships - but others evolve. Angiosperms with dense venation, C₄ grasses and ectomycorrhizal trees construct the pillars, bumblebees, fruit bats and the like, the arches and spandrels, while the bulk of the tens of thousands of species of the [asterid I + II] clade make up the paintings in the spandrels.

At least some characters that might seem to facilitate diversification apparently evolve well before the diversification they are supposed to facilitate, so may be best considered as exaptions, and evolution is contingent (de Queiroz 2002). Along similar lines Edwards and Donoghue (2006) suggest that several key elements of the cactus ecological niche were established before the evolution of the cactus life form - and subsequent diversification of the cactus clade. Similarly, Donoghue (2005) noted that the resolution of paraphyletic groups (as in Cactaceae, see Ogburn 2007; Ogburn & Edwards 2009; Nyffeler & Eggli 2010 for information) has helped spread what appeared to be phylogenetically more or less linked characters through the tree. In fact, the importance of some changes may be less in the changes themselves, but subsequent changes that they make possible and/or their importance in ecological conditions developing long after their origin. Trait lability may also facilitate assemblage of character syndromes; another way of putting it is that it is the combination of traits that is important, rather than any one trait itself (Ogburn & Edwards 2008; see also Stebbins 1951). But there is lability of sorts at the level of individual traits themselves. Thus in basal angiosperms in particular distinctions between the nature and arrangement of floral parts that are obvious in say, core eudicots are less evident (e.g. Buzgo et al. 2004; Endress 2005c; Taylor et al. 2008; Friedman 2008b).

Although Howarth and Donoghue (2004, esp. 2005) note possible connections between changes in CYC-like genes and changes in floral form in Dipsacales, direct links remain to be established - and linking these changes with diversification is yet another issue. Thus crown-group Valerianaceae may be 60-55 million years old (Bell & Donoghue 2005a), but diversification in the Andean paramo - resulting in ca 1/7 of the species current recognized in the family - happened less than 5 million years ago on the arrival of Valerianaceae in that area (Bell & Donoghue 2005b; Moore & Donoghue 2007, see also Adoxaceae, Viburnum in particular) and is not associated with the evolution of particular floral (or other) "key innovations" (e.g. Richardson et al. 2001; von Hagen & Kadereit 2003: see above). The same seems to be true of Andean Lupinus species, rapid diversification there starting only some 1.76-1.19 million years before present and probably being driven by the ecological opportunities available in the high altitude habitats where they are now mostly found (Hughes & Eastwood 2006; Drummond 2008). Species of Lupinus there (and in montane North America) are largely perennials (Drummond 2008), the annual habit being plesiomorphous, and show much variation in habit, etc.; the presence of bumble bees may have been another important factor in their diversification (Hines 2008). Gentianella seems to show a broadly similar pattern of diversification (von Hagen & Kadereit 2001). Diversification in Ericaceae-Vaccinioideae - some 500 species in the tropical Andes above 1,000 m - may also have been encouraged by the Andean orogeny (Luteyn 2002), and the Colombian-Ecuadorian region is also the centre of diversity of hummingbirds, and these pollinate many of the Ericaceae (also Generiaceae) of the same region (Weber 2011 and references). Guatteria (Annonaceae), Inga (Fabaceae-Mimosoideae) and Ocotea (Lauraceae) between them contain over 1,000 species; species of the first two genera in particular tend to occur at lower altitudes. These genera are Neotropical and most species have evolved fairly recently; in Guatteria much speciation may have occurred only subsequent to its entry into South America (Erkens et al. 2007).

One also has to take into account both intrinsic and extrinsic traits of plants; the area a clade inhabits, especially if it is non-contiguous, may affect diversification (Vamosi & Vamosi 2010, 2011; see also Marazzi & Sanderson 2010 above). Along the same lines, Lavin et al. (2004) and Schrire et al. (2005) suggest that it is more profitable to think of diversification and distribution of Fabaceae in terms of vicariance of biomes rather than of the classic geographical areas.

The rise to dominance of the angiosperms and the diversification of particular angiosperm clades also involves other organisms - plants, animals, fungi, bacteria - as well as changes in the environment itself, and it is a thoroughly ecological process (e.g. Thompson 1998; Harmon et al. 2009). Indeed, it has been suggested species richness depends primarily on geographic area inhabited by a clade, diversity being limited by ecological factors and lineages having a carrying capacity set by these limits (e.g. Vamosi & Vamosi 2010). However, if the assumption that clades increase steadily in diversity with time is unreasonable (Rabosky 2009; Vamosi & Vamosi 2010), so is the implicit assumption that the environment does not change, since there is abundant evidence that climate has been changing dramatically throughout the history of crown-group angiosperms, spurred by angiosperms themselves and their associated fungi (e.g. Boyce et al. 2010; see the physiological-ecological context of angiosperm evolution)...

The morphology of clades immediately related to extant angiosperms remains conjectural and the early-diverging Nymphaeales are in many respects highly autapomorphic aquatics, so the polarity of many angiosperm characters is unclear (see also Friedman & Floyd 2001; Ronse De Craene et al. 2003). Furthermore, just about all of the distinctive features of angiosperms have evolved in parallel and/or have been lost. One can, however, develop plausible adaptive scenarios for the evolution of many angiosperm features.

The evolution of the distinctive angiosperm flower is considered a critical evolutionary event (e.g. Dilcher 2000), and the flower can be considered a key innovation, or, better, a group of innovations. Thus the development of a style allows competition among gametophytes (Mulcahy 1979) and may also be associated by an increase in size of pollen grains (Roulston 2000 for references). Frame (2003) the emphasized flexibility in construction of the flowers (there is abundant evidence that the development of flowers of ANITA grade angiosperms and magnoliids in particular is not at all highly canalized [add refs]), the speed of the reproductive cycle, the closure of the carpels, and the fact that flowers are edible hence attracting some kinds of potential pollinators. Closed carpels both protect the ovules and may become much elaborated when the seeds are mature, so promoting dispersal. Endosperm, tissue with both maternal and paternal genomes and usually with a diploid maternal and a haploid paternal contribution, that is involved in the nutrition of the embryo, is a tissue unique to angiosperms, although there is controversy over its origin (cf. e.g. Friedman & Williams 2004 and Nowack et al. 2007). Wny there should be variation in embryo sac development at the base of the angiosperm tree (and sporadically elsewhere, too) that affects the balance of maternal and paternal genes in the parent sporophyte-endosperm-seedling system, is still not well understood. However, a higher ratio of paternal genes in the endosperm may lead to more "selfish" behaviour of individidual endosperm tissues as they scavenge nutrients at the expense of other ovules in the carpel (e.g. Friedman et al. 2008; esp. Friedman & Ryerson 2009) - hence perhaps the rather low ovule number (per carpel) of many ANITA-grade angiosperms. But, as Olsen (2004: p S215) noted, "In spite of recent progress in understanding angiosperm phylogeny, all of the main questions regarding the evolutionary history of the nuclear endosperm remain unresolved." Acquisition of syncarpy would seem to be an important event (Friis et al. 2006b), at least in the current mythology of floral evolution. In fact, it may have evolved seventeen or more times independently, while a compitum, allowing pollen tubes from one stigma to pollinate the ovules in more than one carpel, also evolved in three quarters of these cases (Armbruster et al. 2002). Members of the ANITA grade - perhaps not Nymphaeales - quite frequently have an extragynoecial compitum in which pollination of ovules in more than one carpel from pollen landing on a single stigma is possible (e.g. Williams et al. 1993; Lyew et al. 2007; Williams 2009 - see also Igersheim & Endress 1997; Endress & Igersheim 2000). Angiospermy seems to be associated with a number of changes in the gametophyte phase of the life cycle, which is usually notably shorter than that of extant gymnosperms; reproduction was speeded up (Stebbins 1965). Thus the rate of pollen tube growth in angiosperms shows a great increase over that of extant gymnosperms - 80-600 µm/hour in ANITA-grade angiosperms (overall 60[Fagaceae]-20,000 µm/hour) versus 10-20(Gnetum) µm/hour in gymnosperms (Hoekstra 1983; esp. Williams 2008, 2009 - see also Rudall & Bateman 2008 for microgametophyte evolution). Fertilization occurs within about 24 hours in these angiosperms as compared to seven days - often far more - in most extant gymnosperms (Williams 2008). Such changes may be associated with the evolution of callose plugs in the tube and a wall made up largely of callose, other apomorphies of angiosperms; gymnosperms lack these plugs and have a cellulose-based pollen tube wall (Williams 2008; see Parre & Geitmann 2005 for the mechanical properties of callose). Aside from details of the life cycle, it has also been suggested that change in the hydration of flowers from the xylem to the phloem could represents a major evolutionary transition within angiosperms (Feild et al. 2009a, see also below). Angiosperms tend to become mature at a younger age than do gymnosperms, and that this may aid diversification by speeding up evolution (Verdú 2002).

However, none of these features may have been of immediate importance, at least if judged in terms of numbers of extant taxa in early-branching clades. The initial branches of the angiosperm tree are highly asymmetrical in terms of species number in extant clades (see also Sanderson & Donoghue 1994 in the context of a diferent topology), and as Friis et al. (2006b) emphasized, many of the extant clades that were evident early in the fossil record now have few representatives: thus Amborellales, Nymphaeales, Austrobaileyales, Chloranthaceae, and even most magnoliid families include only a few species. Such families may also differ from other angiosperms in ecophysiological features. Thus ANITA grade angiosperms tend to have rather low veinlet densities, not much above those of other gymnosperms and ferns, so transpiration rates and hence photosynthetic capacities (P_c) are rather low (Brodribb et al. 2007; Boyce et al. 2009; Feild et al. 2009a; Brodribb & Feild 2010). Seeds of plants that make up the basal pectinations of angiosperms are notably smaller than those of extant gymnosperms (Moles et al. 2005a; cf. Stebbins 1981); indeed, if early angiosperms were relatively small plants, this is to be expected, since plant and seed size are quite strongly linked (e.g. Eriksson et al. 2000; Moles et al. 2005b).

The interpretation of the immediate importance of particular features of the flower, such as closed carpels and the development of a style (this was initially short), should be placed in the context of the eco-physiological evolution of early angiosperms and the environmental changes that resulted (see above). This evolution may have enabled the later development of the diversity of floral form that characterises angiosperms, so promoting the further evolution of prezygotic reproductive barriers and an overall shorter generation time (Williams 2008), factors important in subsequent angiosperm diversification. Indeed, even if angiosperms did diversify considerably quite soon after their first appearance early in the Cretaceous, this diversification is likely to have been under ecological conditions rather different from those under which they prospered later; then continents were drifting apart, there was increasing carbon dioxide concentration and rising sea levels, and ever-wet tropical humid climates, initially rather restricted, were becoming more extensive. If "basal" clades that are now species poor were much more diverse in their early history (Magallón & Castillo 2009 and references), they may have speciated in response to conditions that are not found today.

The distinctive modern angiosperm-dominated vegetation is largely of Tertiary age (see above), and it is the early Tertiary environment that may provide an illuminating context for thinking about this. Indeed, venation density in Angiosperms further increased in the later Cretaceous 40-60 million years after the initial evolution of crown angiosperms, an increase associated with increasing maximum photosynthetic rates and perhaps linked to falling CO₂ concentrations and lower atmospheric humidity (Brodribb & Feild 2010). Subsequent Tertiary diversification occurred even in old but now somewhat more speciose clades such as Annonaceae, Lauraceae, and Myristicaceae (particularly striking in the last-named - J. A. Doyle et al. 2004, ?2008). Indeed, it seems likely that the distribution patterns of a number of clades that were thought to reflect vicariance caused by plate tectonic events are quite often better explained by much more recent dispersal/migration events (e.g. Renner 2005b and de Queiroz 2005 for summaries; Higgins et al. 2003 and Nathan 2006 for mechanisms). [Anderson et al. (2011) emphasize the importance of fish as seed dispersers in the Amazon region - out of place.]

Diversification of angiosperms is associated with that of insects, indeed, with diversification of just about all elements of the biota. Associations between species of extant plants and many insects are close, whether the insects are herbivores, detritivores, or pollinators. Although there have been suggestions that it is not so much increased diversification but reduced extinction that has characterised the evolution of insects (Labandeira & Sepkoski 1994), this is unlikely, and diversification of angiosperms appears to be overall contemporaneous with the massive diversification of many insect groups that are now more or less dependent on them. Thus insects feeding on flowering plants make up at least one quarter of all described species (Janz et al. 2006). However, understanding details of the patterns of associations between plant and insect groups is not easy (see Janz 2011 for the current status of ideas on coevolution). What attracts an egg-depositing insect to one plant and prevents it laying eggs on another may be some aspect of plant chemistry (see Bernays & Chapman 1995 for host plant selection). A number of organisms sequester secondary metabolites in the larva and/or adult stages, ensuring some measure of protection by so doing, or they use the metabolites for pheromones, or these metabolites simply act as oviposition cues, appearing not to be otherwise utilised by the insect (Nishida 2002 for a review). Overall, most plant secondary metabolites show considerable homoplasy, and to a certain extent, plant-insect relationships reflect this homoplasy (see alkaloids, glucosinolates, etc.).

In general, current plant and insect diversity are positively correlated, although Hawkins and Porter (2003, see also references) note that the correlation may well not be causal. There are also correlations with temperature, with more herbivory occurring at higher temperatures, as is evident at least locally in the fossil record (Wilf & Labandeira 1999). Novotny et al. (2006) suggest that individual species of temperate and tropical plants (controlled for phylogenetic relationships) support a similar number of insect species, however, there are of course many more species of plants, and hence of insects, in the tropics. South America in particular is very diverse in both plants and insects. This link may be evident as far back as the early Eocene, the diversity of herbivore damage in a fossil flora from Argentina being appreciably greater than in comparable North American floras (Wilf et al. 2005).

Even if initially diversification of insects and angiosperms was associated, some subsequent bouts of insect diversification may have occurred well after the appropriate angiosperm host clades originated (implicit in Futuyma 1983; see Funk et al. 1995; Percy et al. 2004; Lopez-Vaamonde et al. 2006), with diversification of insect groups occurring - and overall diversity increasing - after they adopt new hosts (Janz et al. 2006). Subsequently the relationship may be reversed. Although Kergoat et al. (2005) suggest that diversification of bruchids and Fabaceae may have occurred more or less contemporaneously, in the New World euglossine-pollinated Orchidaceae, we find that the orchids diversify considerably after the likely time of initial diversification of the bees that pollinate them (Ramírez et al. 2011). In general, close co-evolution seems to be the exception rather than the rule, and is most evident in shallow rather than deep clades (Berenbaum & Passoa 1999 for references; cf. Farrell & Mitter 1998); looser "co-evolution", with host shifts associated with taxonomy, may be more common (see Futuyma & Mitter 1996; Janz 2011). In any event, insects perhaps rarely act as selective agents on their hosts (Strong et al. 1983; see e.g. Ramírez et al. 2011).

Diversification of seed-dispersing animals, including birds, and of the plants they dispersed may have proceeded roughly in parallel, although with something of a lag for the animals (e.g. Tiffney 1984; Wing & Tiffney 1987; Dilcher 2000; Collinson & Hooker 1991).

Along with a shift in ecology, there may have been a shift in defence (see above), and perhaps also in associated herbivores. Herbivorous beetles in particular and insects in general diversified in the later Paleocene-Eocene (Farrell 1998; Wilf & Labandeira 1999; Wilf et al. 2001; Lopez-Vaamonde et al. 2006). However, not all post-bolide impact floras behaved the same; there are some from ca 64 million years before present (early Palaeocene) in the western interior United States with low plant diversity (species deciduous, with thin leaves and low defences) and high diversity of insect damage, and others more diverse and with the facies of tropical fainforest (tough, thick, probably tanniniferous leaves) but with low diversity of herbivore damage (Wilf et al. 2006). Diversity in the tropics may have peaked in the Eocene, perhaps even topping today's levels, but the plants and ecological conditions involved may have been rather different from those of today (Jaramillo et al. 2006); Morley (200) provides a good general account of the evolution of rainforests.

Details of the relationships between groups diversifying in temperate regions and their tropical relatives have been a matter of speculation for some time (e.g. Bews 1927). Thus Judd et al. (1994) found that the temperate family of temperate-tropical family pairs often arose from within the tropical family, making the latter paraphyletic. But other literature focuses not so much on details of morphological evolution, but on establishing global - especially latitudinal - patterns of diversity over time (see Mittelbach et al. 2007 for a critical summary), patterns that then can be explained in terms of higher rates of speciation or extinction. The question is, what global patterns of biodiversity can be discerned, and what causes them? (e.g. Kier et al. 2005). In general, plant diversity is broadly correlated with climate (Francis & Currie 2003: families!, but cf. Qian & Ricklefs 2004 for problems with the distribution maps used there). Lack of seasonality in the tropics may be a contributing factor (Janzen 1967; see also Ghalambor et al. 2006), as may evapo-transpiration, topographical diversity, and related factors (Kreft & Jetz 2007). Related to this, Linder (2008) linked the timing of diversification in particular areas to the local environment, whether or not it had been climatically and geologically stable during the Tertiary. There is a connection between diversity and environmental energy variously estimated (and this links with latitude), species richness and the rate of molecular evolution, but the connections are independent (Davies et al. 2004b, see also Allen et al. 2002, more individuals can be supported in productive environments, therefore ceteris paribus more mutations, etc.; Moser et al. 2005; Jaramillo et al. 2006). Diversification patterns in individual families may also be distinctive. For example, Davies et al. (2005) note that in Iridaceae diversification is greater in areas like southern Africa than in the northern hemisphere, although the Cape area seems to be notably diverse from a global point of view (Kreft & Jetz 2007). However, teasing apart historical and ecological signals in patterns of plant diversity is not at all straightforward (Ricklefs 2005).

As Feild and Arens (2005: p. 402) observed, diversification may well depend "on the fortuitous combinations of a large repertoire of traits" rather than on any particular key innovation (see also Crepet & Niklas 2009 and references; Magallón & Castillo 2009), and will be much affected by the ecological opportunities available; angiosperms are characterised by showing recurrent bursts on diversification in separate clades (Sims & McConway 2003; Crepet & Niklas 2009). The role of genome duplication (polyploidy), which seems to have ocurred many times and at all levels of the angiosperm tree (see Soltis et al. 2009 for a summary), in faciltating diversification by allowing the subfunctionalisation and neofunctionalisation of genes, is unclear (e.g. Mayrose et al. 2011); certainly, duplication occurs in gymnosperms, too, even if the chromosome numbers there are low, and also in other land plants. To understand the pattern of diversification we have to factor in the major climatic changes in the Tertiary; as late as ca 30 million years ago there were humming birds in Europe (Mayr 2004), and Cyclanthaceae are also known from earlier in the Tertiary of Europe (Smith et al. 2008). Both groups are now entirely and iconically New World, and humming birds are involved in the pollination of perhaps ca 2,000 species of flowering plants there. Overall angiosperm success seems to be in considerable part the result of diversification of individual angiosperm clades with apparently almost fortuitous combinations of characters and ecological/environmental contexts, and establishing an immediate connection between acquisition of an apomorphy or group of apomorphies and diversification is difficult. Phylogenetic niche conservatism of adaptations to "major ecological niches" may mean that groups will follow these niches when there is an opportunity (Donoghue 2008; especially Lavin et al. 2004; Schrire et al 2005; Marazzi & Sanderson 2010); adaptation to such niches may not occur very frequently. This may be particularly important for some eco-physiological traits.

For further discussions on evolution and diversification, see e.g. Mulcahy (1979: pollen tube competition and angiospermy), Bachelier and Friedman (2011: female gametophyte competition within a single ovule and angiosperm evolution), Taylor and Kirchner (1995: carpel evolution), Hickey and Taylor (1995: evolution of flower), Wing and Boucher (1998: ecology), Armbruster et al. (2002: syncarpy and increase in seed set, offspring quantity, and pollen tube competition, i.e. offspring quality), Donoghue (2004: general), Gianoli (2004: climbing habit), Whitney (2009: stronger selection for divergent flower than fruit morphology). For woodiness, see S. Kim et al. (2004a), on the evolution of the perianth, etc., see e.g. Hasebe (1999), D. Soltis et al. (2005a, b: note that rather than thinking how often petals/a corolla has evolved, it may be more helpful to think about the evolution of a more or less sharp distinction between calyx and corolla form and function in the flower, from a condition in which there is some sort of continuum between the two), for the evolution of features of wood anatomy, see Herendeen et al. (1999: a useful table), for changes in phyllotaxy, Ronse De Craene et al. (2003), for the perianth, see Endress (2008), for the evolution of the flower and fruit, see Dilcher (2000), for variation in the embryo sac, see Friedman (2006 and references) and Friedman and Ryerson (2009), endosperm development, see Williams and Friedman (2004 and references), stigma type, Thien et al. (2009), and for general floral evolution in the "basal" angiosperms in particular, see Endress (2004a, b), Endress and Doyle (2009: note topologies of trees on which characters are optimized), and many others. For microsporogenesis evolution, see Doyle and Endress (2000) and especially Furness et al. (2002b) and Taylor and Osborn (2006), for pollen micromorphology, see Sampson (2000), Doyle (2005: topologies of the trees on which the pollen characters are optimized are now often questionable, 2008) and Taylor and Osborn (2006), for variation in embryo size, see Forbis et al. (2002: very useful) and Verdú (2006), for pollen tube/male gametophyte development, nucellus, etc., see Williams (2008, 2009). For phytochrome evolution, see Mathews et al. (2003), and for oleanane, Taylor et al. (2010).

[Out of place.] This fruit food "niche" was exploited by particular groups of flies, particularly by Drosophilinae, some of the relationships between particular fruits and flies is very close (Ashburner 1998 [on alcohol dehdrogenase in flies]; Harry et al. 1996, 1998 [fig-breeding Lissocephala]) Rates of molecular evolution and speciation are linked, with more molecular evolution occurring in speciose clades (Webster et al. 2003, see also Barraclough & Savolainen 2001). The rate of rbcL gene evolution within families and the number of species those families contain is positively correlated (Barraclough et al. 1996).

AMBORELLALES Melikian, A. V. Bobrov & Zaytzeva  Main Tree, Synapomorphies.

Nodes 1:1; lamina margin toothed; plant dioecious; hypanthium +; A sessile, vascular bundle branched towards apex, pollen anaulcerate [operculum endexinous, margin poorly defined], ektexine cupulate [distinctive undulate, columella-less exine]; ovule 1/carpel, outer integument annular [cap-shaped], nucellar cap 0; embryo sac 9-nucleate [three synergids], bipolar, antipodal cells die very early, polar nuclei in chalazal region [two-modular]; fruit a drupelet; exotesta and exo- and endotegmen thick-walled, lignified; endosperm triploid. - 1 family, 1 genus, 1 species.

Includes Amborellaceae.

Synonymy: Amborellineae Shipunov

AMBORELLACEAE Pichon, nom. cons. Back to Amborellales

Shrub or small tree; alkaloids?, plant accumulates aluminium; cork?; axial parenchyma diffuse; pericycle with hippocrepiform sclereids; mucilage cells 0; petiole bundles arcuate; (stomata anomocytic); leaves spiral; inflorescence cymose; flowers small; P spiral, basally connate, 5-8, single trace; staminate flowers: A 10-25, outer adnate to the base of T, vascular bundle branched near thecae; carpellate flowers: staminodes 1-2; G 5-6, whorled; ovule ± median, pendulous, hemianatropous, sessile, micropyle endostomal; P persistent, stone surface sculpted; seed coat tanniniferous; n = 13; horizontal transfer of atp1 gene.

1[list]/1: Amborella trichopoda. New Caledonia. [Photo - Leaves, Flower.]

• Amborellaceae are sprawling shrubs with apparently two-ranked, exstipulate leaves; the margins are both coarsely serrate and rather undulate. The plant is dioecious and the flowers are small, with an undifferentiated perianth. The staminate flowers have sessile anthers and the carpellate flowers a few carpels that develop into drupelets with pock-marked stones and pockets of almost resinous substances.

Evolution. Floral Biology. Stigmatic exudate may cover all the stigmas of a single flower together, and pollination of ovules in more than one carpel from pollen landing on a single stigma is possible, i.e. there is an extragynoecial compitum (Williams 2009).

Genes & Genomes. It has recently been shown that the mitochondrial genome of Amborella contains genes from a number of land plants, including at least three different mosses; such "foreign" genes may migrate to the nucleus (Bergthorsson et al. 2004, but cf. Goremykin et al. 2009, the evidence for these tranfers perhaps questionable on methodological grounds). Although mitochondrial genomes like that of Amborella are as yet unknown from other angiosperms, sampling is very poor.

Chemistry, Morphology, etc. There is no reaction wood, the stems of Amborella tending to sprawl, especially when young; in terms of architectural models (Hallé et al. 1978) the plant conforms to Troll's model. Stomatal morphology is quite variable, although the brchyparacytic configuration is common (Carlquist & Scnier 2001). The leaves are described as being spiral at first (Cronquist 1981; Takhtajan 1997, but cf. Posluszny & Tomlinson 2003).

The perianth is spiral and undifferentiated. There seems to be no agreement on the pollen morphology of Amborella; Sampson (2000) and Hesse (2001) suggest that the pollen is not really tectate (see also Doyle 2000, 2001; Doyle and Endress 2000), and the aperture type is difficult to categorise, as well as not always being present. Williams (2008, 2009) describes pollen tube development and fertilization. The ovule has been described as being orthotropous, anatropous, or intermediate (Tobe et al. 2000). Bobrov et al. (2005) show that the drupe of Amborella differs from a typical drupe in that the bulk of the woody layer is mesocarpial in origin, unlike the drupes of Laurales, etc. Tha nature of the resinous cavities in the mesocarp is unclear; although not observed by Bobrov et al. (2005), they are unlikely to be an artefact caused by re-expansion of dried fruits prior to study. The seed coat appears to have thin, unlignified walls, as might be expected in such a fruit, although some lignification has been reported (Tobe et al. 2000).

Ovule "type" has also been variously interpreted. Friedman (2006; cf. Tobe et al. 2000) described a very distinctive embryo sac for Amborella; the third synergid cell arises from a cell division that also produces the female gamete. In other angiosperms the polar nuclei are sister to the egg nucleus (at one end) and the central chalazal nucleus (at the other), and the egg is produced by a nuclear division; however, the overall pattern is not necessarily fundamentally different since Amborella has a 9-nucleate embryo sac (cf. Friedman 2006); Friedman and Ryerson (2009) discuss the evolution of the angiosperm embryo sac in detail. But how widespread the "other angiosperm" pattern really is, is unclear, and in part it may rest on the belief that the micropylar and chalazal ends of the embryo sac are identical, each representing a single, highly reduced archegonium. Thus Porsch (1907) took this view, but note that he had an "Englerian" concept of seed plant evolution, with Amentiferae being primitive. Porsch and others at that time (e.g. Nawaschin 1895) saw chalazogamy in Amentiferae (see Fagales here, also Ulmaceae) as being in some way intermediate between porogamy and non-angiospermy, where the female gametophyte has more than a single archegonium.

Additional information is taken from Bailey and Swamy (1948: general), Metcalfe (1987: anatomy), Philipson (1993), Sampson (1993: pollen), Floyd and Friedman (2001: endosperm development), Carlquist and Schneider (2001: leaf and stem anatomy), Yamada et al. (2001a: ovules), Posluszny and Tomlinson (2003: floral morphology), Field et al. (2003: ecophysiology) and especially Tobe et al. (2000: embryology). Chemistry?

Previous Relationships. Amborellaceae were included in Laurales by Takhtajan (1981) and Takhtajan (1997).