Student 1

Roots with exarch protoxylem, lateral roots endogenous; leaf traces leaving a gap in the central stele; euphylls or megaphylls spirally arranged, with apical/marginal growth, venation develops basipetally, growth determinate; sporangia borne in pairs and grouped in terminal trusses, meiosis polyplastidic; sperm multiflagellate, monoplastidic, basal bodies staggered, centrosomes bicentriolar; 30kb chloroplast inversion in the large single-copy region of the chloroplast genome.

Evolution. For information on possible apomorphies for crown euphyllophytes, see Raubeson and Jansen (1992b), Kenrick and Crane (1997), Imaichi et al. (2008: position of some characters difficult to ascertain), and Schneider et al. (2009).

Crown euphyllophytes appear to date from 401-380 million years ago (Leebens-Mack et al. 2005) and are in turn made up of two clades, ferns and their relatives, the monilophytes or Moniliformopses, lacking true roots, and lignophytes, made up largely of seed plants or spermatophytes.

Chemistry, Morphology, etc. Note that details of the evolution of megaphylls - indeed, a satisfactory definition for them seems to be lacking - are unclear (see e.g. Sporne 1965, but cf. Boyce 2005 summarizes earlier literature; Tomescu 2008, 2009; Sanders et al. 2009). At the level of development, there is considerable similarity between magaphylls and microphylls (Harrison et al. 2005b). The leaf supply to megaphylls in monilophytes seem to have evolved by dissection of an amphiphloic siphonostele, while leaf gaps in seed plants are associated with a stele that consists of a series of sympodia of collateral vascular strands (see also below), so from this point of view megaphylls in the two groups may represent parallelisms rather than a synapomorphy and leaf gaps are not equivalent, being used in a descriptive sense only (Namboodiri & Beck 1968c; Beck et al. 1982, but cf. Schneider et al. 2009). However, it has been found that the vascular construction of the rhizome in some true ferns is also made up of sympodia (Karafit et al. 2005). Floyd and Bowman (2007) suggest that megaphylls have evolved independently in the angiosperms and ferns and relatives (see also Boyce & Knoll 2002; Gensel & Kenrick 2007; Tomescu 2009; Sanders et al. 2009), while Floyd and Bowman (2010) comparae gene expression patterns in shoots and leaves, suggesting that the marginal blastozones of leaves and the shoot apical meristem may be similar in some respects. Osborne et al. (2004) provide an ecological explanation for the origin of megaphylls based on falling CO₂ levels, although the developmental mechanisms involved may have evolved long before then (Beerling 2005a and references). Note that Schneider et al. (2009: p. 461 and references) suggest that euphylls did arise once, and can be characterized by apical/marginal growth, apical origin of the venation, determinate growth, etc.

FERNS s.l./MONILOPHYTA

Shoot apical meristem of a single cell, plasmodesmatal network cell lineage specific; amphiphloic siphonostele + [discontinuities in stele in t.s. +, caused by leaf gaps]; protoxylem restricted to lobes of central xylem strand [xylem development mesarch], primary xylem with circular bordered pits; phloem fibres rare; stem endodermis and pericycle +; leaves megaphyllous [ad/abaxial symmetry evolved first, then determinancy], frond veins not anastomosing; sporangia in sori, sporangium stalk 6< cells across, walls two cells thick, lacking an annulus, spores/sporangium 1000<, white, spores globose-tetrahedral, trilete, tapetum plasmodial, spore wall development centrifugal, exospore 3-layered, pseudoendospore +; gametophytes exosporic, green, photosynthetic, antheridium embedded, wall ³5 cells thick; first division of the zygote horizontal; nine-nucleotide insertion in the plastid rps4 gene.

Evolution. Some possible apomorphies (see e.g. Schneider et al. 2009) are in bold.

Schuettpelz and Pryer (2009, esp. Tables 2, 3 in the Supplement) provide extensive dating of divergence times in monilophytes, and also list a number of fossil records of the group (for the fossil record, see also Rothwell & Stockey 2008).

There have been several radiations of homosporous leptosporangiate ferns, the first in the Palaeozoic, giving rise to lineages that have since become extinct, in the Jurassic and again in the Cretaceous (Rothwell & Stockey 2008). General fern diversity decreased (along with that of the cycads) through the Cretaceous (Wing & Boucher 1998), and the diversification that gave rise to most living ferns, especially to the polypod ferns, which make up some 80% of living fern species, may have occurred subsequent to the diversification of the angiosperms (Lovis 1977; Rothwell & Stockey 2008; Schuettpelz & Pryer 2009). Indeed, ferns appear to have temporarily dominated at least locally after the end-Cretaceous bolide impact (Schneider et al. 2004a). Quite a number of the polygrammoid ferns (Polypodiaceae + Grammitidaceae) are epiphytic, and the Grammitidaceae in particular have green spores and accelerated plastid genome evolution, a correlation also found elsewhere in ferns, but not 100% (Schneider et al. 2004b). On the other hand there has been an abrupt reduction in the rate of molecular evolution in the largely arborescent Cyatheales (Korall et al. 2010: Marattiales, Osmundales, etc., not included). Indeed, the eusporangiate Marrattia and Angiopteris, and also the leptosporangiate tree ferns, may be something of living fossils showing little molecular and even morphological evolution (P. Soltis et al. 2002).

Ferns are noted for the high incidence of polyploidy within the group, and it is estimated that almost 1/3 (31%) of all speciation events there are accompanied by polyploidy (Wood et al. 2009).

Chemistry, Morpholgy, etc. Monilophytes or ferns s.l. are characterised by having a siphonostele, the protoxylem being restricted to lobes of the central xylem strand, hence bringing to mind a necklace (development of the xylem is mesarch, although notably variable in the Ophioglossum/Psilotum clade); spore wall development that is exclusively centrifugal; similarity in details of spermatozoid morphology and movement; a nine-nucleotide insertion in the plastid rps4 gene, etc. (e.g. Renzaglia et al. 2002; Schneider et al. 2002).

For general comparative anatomy, see Ogura (1972), and for details of stelar morphology and evolution, see Beck et al. (1982), for megaphylls, see Tomescu (2009).

Phylogeny. The circumscription of this clade has only recently become clear, and as just mentioned they include Psilotum (Tmesipteris is close) sister to Ophioglossum (support strong) in a clade sister to all other ferns. Equisetum, perhaps sister to Angiopteris, etc. (although support currently only moderate), may be in turn sister to remaining ferns (e.g. Pryer et el. 2001a, 2004a; Wikström & Pryer 2005; Qiu et al. 2007; cf. in part Wolf et al. 1998). However, recent work places Equisetaceae alone sister to all other ferns; some support came from a rps4 analysis, and also 4- and 5-gene analyses, the latter two with strong support (Schuettpelz et al. 2006). Wikström and Pryer (2005) note that Equisteum has no mitochondrial atp1 intron, and this is either a secondary (and parallel) loss or plesiomorphic absence, depending on the topology of the whole group (see the character hierarchy below). Spore wall ultrastructure of Calamites, an extinct member of Equisetaceae, is not so different from that of Ophioglossaceae and other ferns (Lugardon & Brousmiche-Delcambre 1994; Grauvogel-Stamm & Lugardon 2009). The inclusion of morphology alone or in combination also affects relationships (Wikström & Pryer 2005 and references).

Within the remaining ferns is a large clade made up of leptosporangiate ferns (with very strong support - see also Hasebe et al. 1994, 1995, Pryer et al. 1995; Wolf et al. 1998; Quandt et al. 2004; Schuettpelz et al. 2006) that originated perhaps 350 million years before present (e.g. Schneider et al. 2004a). Within this leptosporangiate clade, Osmunda and relatives, the sporangia of which have some eusporangiate features, are strongly supported as being sister to the rest. There is further substantial resolution of relationhips within leptosporangiate ferns (e.g. Pryer et al. 2004a, b and references). Davalliaceae and related taxa are sister to the polygrammoid ferns, and they, too, include a number of epiphytes (for their evolution, see Tsutsumi & Kato 2006).

Classification. Smith et al. (2006, 2008) propose a phylogeny-based reclassification of the ferns, and they also include literature, ordinal and familial synonymy, and a list of accepted genera and some major synonyms. However, it is likely that adjustments to this classification will be needed as details of the phylogeny become better understood (Schuettpelz & Pryer 2007, 2008). A provisional hierarchy of characters obtained from Smith et al. (2006, 2008) and also from Pryer et al. (1996), is given below. For pteridophytes in general (these often include lycophytes), see also Kato (2005) and Ranker and Haufler (2008).

Previous Relationships. Psilotum and relatives used to be considered the most primitive living vascular plants.

EXTANT SEED PLANTS/SPERMATOPHYTA

Plant woody, evergreen; true roots present, endomycorrhizal; vascular cambium + [xylem differentiating internally, phloem externally]; wood homoxylous, tracheids and rays alone, tracheid/tracheid pits circular, bordered; sieve tube/cell plastids with starch grains; stem cork cambium superficial, root cork cambium deep seated; nodes ?; stomata?; leaves megaphyllous [determinancy evolved first, then ad/abaxial symmetry], spiral, simple, axillary buds associated with at most some leaves, prophylls two, lateral; plant heterosporous [with microspores (pollen) in stamens and megaspores in ovules]; sporangia eusporangiate [i.e. developing from a group of superficial cells and with walls two or more cells across], on sporophylls, aggregated in indeterminate cones/strobili; pollen grains mono[ana]sulcate, germinating, or gamates exiting, distally and lacking proximal preformed suture or marks where they were attached to other members of the tetrad, exine and intine homogeneous; ovules unitegmic, crassinucellate, megasporangium indehiscent, monosporic [only one megaspore develops from meiotic tetrad and germinates to form female gametophyte]; male gametophyte development endosporic [with first divisions inside the pollen grain], gametes two, developing after pollination, with cell walls, with many flagellae; female gametophyte endosporic, initially syncytial [nuclei only, no cell walls], walls produced later surrounding individual nuclei; seeds "large", first cell wall of zygote transverse, embryo straight, short-minute, with morphological dormancy, cotyledons 2; plastid transmission maternal. [Back to Index]

Estimates of the age for crown-group seed plants, i.e., extant seed plants, range from 348-285 million years before present (e.g. Becker et al. 2000; Leebens-Mack et al. 2005).

Because of the probable sister group relationship between extant gymnosperms and angiosperms, many life cycle characteristics cannot be polarised. However, it is very likely that many of the features of gametophyte and young sporophyte that characterize gymnosperms are likely to be features of the extant seed plants as a whole.

EXTANT GYMNOSPERMS [Back to Index]

Phloem with sieve cells and associated Strasburger cells, the sieve area with small pores generally less than 0.8 µm across that have cytoplasm and E.R., joining to form a median cavity in the region of the middle lamella; transfusion tissue +; stomata in which subsidiary and guard cells come from the same initial; cataphylls [modified protective leaves] +; pollen endexine lamellate; ovule unitegmic, with pollen chamber developing by breakdown of nucellar cells, pollination droplet +, fertilisation 7 days to 4-6 months or more after pollination; pollen tube breaks down sporophytic cells, grows away from ovule, male gametophyte with 4 cells [two prothallial cells, tube cell, stalk/sterile cell] and two multiflagellate gametes released through pollen grain wall; female gametophyte with radially-elongated cells [alveoli]; testa mainly of coloured sarcotesta and sclerotesta [fleshy and hard testa layers], ± vascularized; proembryo syncytial, with many nuclear divisions before wall formation, massive suspensor formed; gametophyte persists in seed as major reserve.

Much of this characterisation may apply to all extant seed plants (see above), with more detail to be added if extant gymnosperms are paraphyletic, although this now seems increasingly unlikely.

For general information see Gifford and Foster (1988), Hill (2005) and Anderson et al. (2007: including fossils), for leaf anatomy, see Napp-Zinn (1966), for the shoot apex, see Johnson (1951), for the binding of ferulic acid to the primary cell wall, see Carnachan and Harris (2000), for pit membranes, see Bauch et al. (1972), for stelar evolution, see Beck et al. (1982), for nodal anatomy, see Kumari (1963), for differences between the roots of lycophytes and lignophytes, see Gensel and Berry (2001) and Gensel et al. (2001), for sieve element plastids, see Behnke (1974) and Behnke and Paliwal (1973), for pollen, especially the alveolate infratectal layer of the ectexine and lamellate endexine, see van Campo (1971), van Campo and Lugardon (1973), Thomas and Spicer (1986), Xi and Wang (1989), Faegri and Iversen (1989), Page (1990), Osborn and Taylor (1994) and Kurmann and Zavada (1994), for variation in the life cycle and embryology, see Singh (1978), for the integument and its evolution, see Andrews (1963), for double fertilisation, see Friedman (1992), for plastid transmission, understood fairly well in only Encephalartos and Ephedra outside Pinales, see Moussel (1978 - Ephedra), Wilson and Owens (2006: Pinales), Mogensen (1996 - summary), Cafasso et al. (2001 - Encephalartos), for the female gametophyte, see Maheshwari and Singh (1967), for genome size, see Leitch et al. (2001, 2005), for embryo size, see Forbis et al. (2002), for pollination, see Stützel and Röwekamp (1999b), Labandeira et al. (2007: insect pollination in Mesozoic gymnosperms), and Ren et al. (2009: scorpion flies mid-Mesozoic pollinators?), for duplication of the phytochrome gene, see Schmidt and Schneider-Poetsch (2002), and for morphological apomorphies, see Doyle (1998a, 2006).

CYCADALES Dumortier [Back to Index]

Roots and stems with contractile tissue; toxins + [ß-methylamino-L-alanine and compounds producing methylazoxymethanol], mucilage copious; association with Nostoc or Anabaena in apogeotropic coralloid roots [much-branched roots growing ± upwards]; primary thickening meristem +; wood manoxylic [not very dense, numerous rays, etc.], large amounts of secondary phloem persisting; foliage leaf nodes multilacunar, with traces girdling the stem; axillary buds 0; leaves large, pinnate, bases persisting; plants dioecious; strobili ?terminal; many abaxial microsporangia/sporophyll, dehiscing by the action of the epidermis [exothecium]; megasporophylls with terminal sterile portion, not aggregated into a cone; pollen tube usually unbranched, one prothallial cell; seeds with sarcotesta and inner fleshy layer, both vascularized; germination hypogeal, cryptocotylar. - 2 families, 10 genera, 305 species.

Cycads are known as fossils in the Upper Palaeozoic 290-265 million years before present, probably being derived from Palaeozoic pteridosperm-like plants (Mamay 1969; Gao & Thomas 1989), and the Cycas lineage may already have diverged from Zamiaceae by the Permian at least 250 million years before present (Hermsen et al. 2006a; see also Bogler & Francisco-Ortega 2004: there are also much younger estimates, e.g. Treutlein & Wink 2002).

The nitrogen-fixing cyanobacteria Nostoc and Anabaena have been found in several cycads growing either inter- or intracellularly in the apogeotropic coralloid roots near the surface of the soil (e.g. Lindblad et al. 1985).

There are widespread, close and specific associations between Zamiaceae and their beetle (weevil) and thrip pollinators (e.g. Schneider et al. 2002; Terry et al. 2007), although it seems likely that in their present form these are relatively recent (Downie et al. 2007); wind pollination may also occur (Kono & Tobe 2007). Indeed, the development of such associations with insect pollinators may even have contributed to the relatively recent diversification of cycad clades like Encephalartos, Macrozamia etc., in the (late) Tertiary (Oberprieler 2004; see also Downie & Donaldson 2005). Thermogenesis has been detected in the strobili of some Cycadales (Seymour 2001).

A few lepidopteran larvae eat cycads, in particular the lycaenid Eumaeus (Schneider at al. 2002 for references), although in S.E. Africa a group of brightly-coloured diptychine geometrids (loopers) is more or less restricted to cycads (Cooper & Goode 2004); it would be interesting to know if they are distasteful to potential predators.

Cycads are noted for having some rather potent toxins that, it has been suggested, may have contributed to the persistence of the clade. ß-methylamino-L-alanine (BMAA) is widespread, as is methylazoxymethanol (MAM), produced by the hydrolysis of glycosides such as cycasin (a monosaccharide) and macrozamin (with a disaccharide). BMAA, interestingly, is also probably produced by the cyanobacterial associates of cycads (Cox et al. 2005). These compounds are toxic: BMAA is a possible neurotoxin, and MAM can cause severe digestive upsets, cancers, etc. (Brenner et al. 2003 and references).

Hermsen et al. (2006a) suggest a number of additional synapomorphies, including the presence of pith cell packets and three unique biflavones. Although girdling traces are conspicuous in the vascular supply to the expanded foliage leaves, the vascular traces to the reduced leaves (cataphylls) and sporophylls take a direct course through the cortex (some traces may also proceed directly to the foliage leaves as well). Nodal anatomy appears to be complex (Pant 1973); the cotyledons have split lateral vascular traces (Coulter & Chamberlain 1917). There are up to some 40,000 flagellae per male gamete.

Ryberg et al. (2007) emphasise the large amounts of secondary phloem that persist in a cycad stem, presumably because the cork cambium is not very active; some larger cycads have fibres in tangential bands in this phloem. Both characters are common in the Bennettitales, which with the Cycadales make up the grade group, the cycadophytes, although in the Bennettitales the fibre bands alternate with tissue that is very largely made up of sieve cells.

For relationships within Cycadales as a whole, see e.g. K. D. Hill et al. (2003, 2004), Bogler and Francisco-Ortega (2004), Wink (2006). Cycas is sister to other cycads, but other details of relationships are rather unclear (support mostly low) and conflict in part, at least, with those suggested by previous morphological studies. The positions of Bowenia, Stangeria (the two have sometimes been associated, as by Stevenson 1992 in a morphological analysis [and placed in Stangeriaceae], see also Brenner et al. 2003), and Dioon are particularly uncertain; all three are distinctive genera. There are, however, good morphological characters supporting the basal division between Cycadaceae and Zamiaceae (K. D. Hill et al. 2003), so a conservative (broad) approach to family limits has been taken here - the whole order is not very big. A recent study (Chaw et al. 2005, see also Rai et al. 2003, 2008; Zgurski et al. 2008) suggests the following quite well supported relationships within Zamiaceae: [Dioon [Bowenia (not always here) [the rest - including Stangeria which is never close to the first two]]]. Such relationships are consistent with variation in the micromorphology of the cuticular waxes (Wilhelmi & Bartlott 1997), Dioon and Cycas having a plesiomorphic morphology.

For embryology, see Singh (1978), for branching, see Stevenson (1988), and for fossils, see Pant (1987), for general references, see Gifford and Foster (1988), Johnson and Wilson (1990), Stevenson (1990), Norstog and Nicholls (1997), and Schneider et al. (2002: the biology and evolution of the group), Jones (2002: account of all taxa), Walters and Osborne (2004: problems of species delimitation, etc.); see also Artabe and Stevenson (1999: especially anatomical variation), Hill et al. (2004: list of all taxa), The Cycad Pages and the Gymnosperm Database.

Includes: Cycadaceae, Zamiaceae.

CYCADACEAE Persoon - Hairs transparent; stem with more than one complete vascular cylinder [polyxylic]; leaflets circinate, midrib +; megasporophylls not forming a cone, margins lobed or toothed, (1-)3-6 erect ovules/sporophyll; seeds platyspermic. - 1/100. E. Africa and Madagascar, South East Asia to New Caledonia and Tonga.

ZAMIACEAE Horaninow - Hairs coloured; leaflets flat, no midrib, secondary veins regular, subparallel; megasporophylls peltate, 2(-3) inverted ovules/sporophyll; seeds radiospermic. - 9-10/200: Encephalartos (65), Zamia (55), Macrozamia (40). Scattered throughout the tropics and subtropics.

Stangeria is perhaps particularly distinctive: it develops buds from its roots, it lacks cataphylls, and its leaflets have dichotomously-branching pinnate venation. Stomata are to be found at the apex of the nucellus in Zamia, perhaps reflecting a time when the nucellus - really the wall of the megasporangium - was exposed. Each male gametophyte of Microcycas produces multiple spermatozoids and each female gametophyte produces several - even hundreds - of ovules.

Chaw et al. (2005) suggest apomorphies for Zamiaceae and some clades within it, as well as a realignment of generic limits. For relationships within the speciose Encephalartos which perhaps split from Lepidozamia 5-20 million years before present - which, despite the age spread, is quite recent - see Treutlein et al. (2005).

GINKGOALES + PINALES: wood pycnoxylic [dense, rays not so prevalent]; vascular pits bordered, with margo-torus construction; axillary buds +; strobili axillary.

GINKGOALES Gorozh. [Back to Index]

Nodes 1:2; leaf venation dichotomising, open; plants dioecious; 2 pendulous microsporangia/microsporophyll, dehiscing by the action of the hypodermis [endothecium]; megasporophyllar strobilus 0; 2(-4) terminal erect ovules together on megasporophyll, each with basal collar, on a long stalk; pollen tube branched, sperm cells binucleate, both nuclei fuse with female gametes; inner fleshy layer of seed alone vascularized; germination hypogeal, cryptocotylar. - 1 family, 1 genus, 1 species.

Includes: Ginkgoaceae.

GINKGOACEAE Engler - Plant resinous, mucilage +; short shoots +; leaves deciduous. - 1/1: Ginkgo biloba. E. China, perhaps extinct in the wild.

Ginkgoaceae are deciduous trees with stout and conspicuous short shoots; the leaves have dichotomous venation.

Ginkgoales were almost world-wide in distribution and included several genera in the Mesozoic, and possibly originated from Palaezoic pteridosperms, perhaps in the Upper Carboniferous (Thomas & Spicer 1987; Zhou 1997). The morphology of these early Ginkgo-like plants is uncertain, but the ovules may have been more numerous, very differently arranged and some at least were inverted rather than erect (and/or platyspermic). Ginkgo has very distinctive leaves, and Ginkgo-like leaves are known from the Permian onwards (see Zhou & Zhang 2003).

It is reported that dioecy is associated with chromosomal differentiation (female xx, male xy), but cf. Hizume (1997).

Note that the nuclei of the female gametophyte have the DNA content of diploid cells (Friedman & Gifford 1997). For more information, see Friedman (1987: male gametophyte), Soma (1997: female gametophyte, embryogeny), Hori et al. (1997: general), and Douglas et al. (2007: ovule).

PINALES: pollen tube unbranched, growing towards the ovule, gametes non-motile, released from the distal end of the tube [siphonogamy]; germination epigeal, phanerocotylar.

PINALES Dumortier [Back to Index]

Resin ducts/cells in phloem in vascular tissue [and elsewhere]; stem cork cambium ± deep seated; nodes 1:1; leaves with single vein; exine thick [³2 µm across], granular; microsporangium dehiscing by the action of the hypodermis [endothecium]; ovulate strobilus compound, with ± united flattened ovuliferous and bract scales; ovule lacking pollen chamber; seed coat dry, not vascularized; proembryo with 4 or 5 nuclear divisions, with upper tier or tiers of cells from which secondary suspensor develops, elongated primary suspensor cells and basal embryonal cells [or some variant]; plastid transmission paternal. - 7 families, 68 genera, 545 species.

Many rusts (Uredinales), including those on ferns, have their aecial stage (these produce binucleate aeciospores) on Pinales, especially on Pinaceae and Cupressaceae (Savile 1979b). Ambrosia and bark beetles (Platypodinae, Scolytinae, both weevils) seem to have been associated ancestrally with conifers, then shifted on to angiosperms and finally back to conifers several times - their current diversity in Pinales is lower (Farrell et al. 2001; see also Powell et al. 1999 for other insect-conifer associations). Bark beetles (Scolytinae) make their gallery systems in phloem, ambrosia beetles (Platypodinae, Scolytinae) in the wood, and they live mostly in dead or dying wood. The few bark beetles that are noxious pests invade living pines. Ambrosia beetles may also carry blue stain fungi (species from a few unrelated ascomycete genera) that can quickly invade the sapwood, rendering it non-functional, and the result is that the plant can die surprisingly quickly. Other fungi are involved in similar close associations, both with ambrosia and bark beetles. Some of the weevils cultivate and eat the fungus; the evolution of cultivation is unreversed. The weevils have highly modified cuticular structures that allow the transport of fungal spores (Beaver 1989; Jordal et al. 2008 and references). Not only pine beetles and fungi, but yeasts, bacteria (some nitrogen-fixing), parasitoids of the beetles and fungus-eating nematodes all form part of a very complex association.

Hudgins et al. (2003) examined the diversity of bark beetles in the context of various plant structures that might be defenses against such beasts; Francheschi et al. (2005) elaborate on the pine-beetle story. Conifers in general have layers of polyphenol-containing parenchyma cells in the phloem. Many bark beetles are found growing on Pinaceae despite the constitutive presence of resin ducts in both phloem and xylem (i.e. the ducts do not develop in response to some trauma, etc., but are always to be found there), there are intracellular crystals, etc. Other families of Pinales have such ducts only in the phloem, but they also have large numbers of small, extracellular, calcium oxalate crystals and also stratified phloem (Pinaceae have scattered sclereid cells or sometimes groups of such cells), both possibly protective structures - and a lower diversity of these beetles. Keeling and Bohlmann (2006a [detailed discussion], b) discuss terpenoids and conifer defence mechanisms, a complex subject; it is unclear just what drives the diversity of terpenoids in conifers.

Pitterman et al. (2005), Hacke et al. (2005) and Sperry et al. (2006) compare water transport in tracheids that have the torus:margo pits found in many conifers (and Ginkgo) with that in other kinds of tracheids and in vessels. Pore size in the thin, peripheral margo is relatively large, while the torus provides a valuable satefy feature guarding against embolism. Overall, hydraulic conductance in tracheids with such torus:margo pits is somewhat greater than in vessels of similar diameter when expressed on a sapwood area basis. The issue is complex, since hydraulic conductance and gas exchange per unit leaf area may be higher in angiosperms (Lusk et al. 2007, study carried out in Chile; see also Boyce et al. 2009).

A number of Pinales have pollen grains with paired sacci or wings, and there is a correlation between the presence of these pollen sacci and exine thickness and structure, whether (no wings) or not (wings) the pollen is wettable, etc. (Tomlinson 1994). In general, the sacci help orient the pollen grains in the pollination droplet (Doyle & O'Leary 1935; Salter et al. 2002 and references), or, more particularly, when the ovules are inverted, the pollen grains are wetted and float up to the micropyle where the sacci separate and expose the sulcus through which the pollen tube germinates (Salter et al. 2002). Previously it had been thought that sacci facilitated wind dispersal of the pollen, and indeed they may also increase the distance the pollen grain can travel before it falls to the ground, so facilitating wind pollination (Schwendemann et al. 2007). In Phyllocladus and many taxa with erect ovules the pollination droplet is resorbed through the micropyle, and again the pollen grains are brought close to the nucellus; in Juniperus communis and other taxa resorbtion of the ovule droplet may be an active process happening quite soon after the pollen grain lands (Mugnaini et al. 2007). There are further variants of these pollination mechanisms in Coniferales (Owens et al. 1998; Salter et al. 2002 for references). The character, "pollen exine shed during germination", is likely to have evolved more than once (?three times) in Pinaceae (see also Rydin & Friis 2005).

More than some other groups, we cannot understand the evolution of conifers just by looking at their extant representatives, and it is clear that the apomorphies of extant conifers depend critically on fossil outgroups (e.g. Hart 1987). There are no known synapomorphies for a clade containing living and extinct conifers (e.g. Rothwell & Serbet 1994). The morphology of extinct conifers and coniferophytes is currently being reevaluated as the morphologies of entire organisms are pieced together from what used to be separate form genera; the result is that many of the conventional taxonomic groupings are being radically overhauled (e.g. Rothwell et al. 2005; see also below). Furthermore, the extent of the diversity of fossil conifers is becoming clear. Not only are forked leaves common, but stomatal distribution, etc., may differ dramatically on leaves from the one plant, compound microsporangiate strobili are known (see also Gnetum, etc.), as are megasporagiate strobili which do not terminate vegetative growth of the axis on which they occur (e.g. Hernandez-Castillo et al. 2001; Rothwell & Mapes 2001). The current distributions of many extant conifer groups is much smaller than and/or very different from their past distributions - which in many cases are surprisingly well known - that go back to the Cretaceous (e.g. Manchester 1999: N. temperate distributions); for example, the African Widdringtonia (Cupressaceae) has been found in Cretaceous rocks from Alabama (McIver 2001).

Within conifers, relationships are being substantially clarified. Pinaceae (Pinus, Cedrus, etc.) are sister to the rest, as a morphological cladistic analysis by Hart (1987) suggested some time ago (but cf. Nixon et al. 1994; Doyle 1996b), other details of Hart's phylogeny have not been confirmed. Molecular data and additional morphological work largely confirm the relationships in the tree below, which is based on the work of Quinn et al. (2002: successive approximations weighting), see also Kelch and Cranfill (2000), Gugerli et al. (2001: e.g. the mitochondrial nadI gene), and Rai et al. (2002, and especially 2008a and references). There has been some uncertainty in the Cephalotaxaceae-Taxaceae area, the two being tentatively bring combined here since there is some evidence that the exclusion of Cephalotaxus would make Taxaceae paraphyletic (Quinn et al. 2002; Price 2003; Rai et al. 2008a). Taxaceae themselves have often been considered rather different from other conifers (e.g. Florin 1948, 1954; Miller 1999), but a reinterpretation of the nature of their reproductive structures (Stützel & Röwekamp 1999a) suggest that Taxus in particular can be linked with Torreya and thence to other conifers.

For a classic study of both fossil and extant conifers, see Florin (e.g. 1951), see also Doyle and Brennan (1972: cleavage polyembryony), Doyle (1945), Tomlinson (1994, 2000), and Tomlinson and Takaso (1998, 2002), all pollination, Page (1990: general), Kumari (1963: nodal anatomy), Butts and Buchholz (1940: cotyledon number), Herrmann (1951: intergeneric grafting), Napp-Zinn (1966: leaf anatomy), Den Outer (1967) and Schulz (1990), both phloem anatomy, Zhou and Jiang (1992: wood anatomy), Raubesen and Jansen (1992a: loss of a copy of the inverted repeat), Hill and Brodribb (1998: southern conifers), Owens et al. (1995b: cytoplasmic inheritance), Mundry (2000: cone/strobilus development, with an emphasis on Taxaceae and friends), Trapp and Croteau (2001: resin biosynthesis), Sklonnaya and Ruguzova (2003: spermatogenesis), Farjon (2001: checklist, 2005b: bibliography).

Includes: Araucariaceae, Cupressaceae, Pinaceae, Podocarpaceae, Phyllocladaceae, Sciadopityaceae, Taxaceae.

PINACEAE F. Rudolphi - Plant ectomycorrhizal; resin ducts in wood and phloem; sieve cells with nacreous [mother of pearl-like] walls, plastids with protein fibres; phloem fibres 0; 2 microsporangia/microsporophyll, pollen saccate, exine thin [2³ µm] except distally, alveolate, ovules 2/scale, inverted, pollination droplet 0 (+); sperm cells binucleate; "embryo tetrad" present [the free-nuclear stage has only four nuclei]; seeds winged, wing develops from adaxial side of scale; cotyledons (2-)4-11(-20); only one copy of the chloroplast inverted repeat. - 11/210: Pinus (105), Abies (46), Picea (33). North Temperate.

Although Pinaceae are known fossil only from the Early Cretaceous onwards (Miller 1999), the age of the lineage must be well over 200 million years (200-350 million years before present - see Eckert & Hall 2006; cf. also Gernandt et al. 2008) since they are sister to all other extant conifers. Within Pinaceae, estimates of the divergence of Pinus from the rest range from 190-102 million years before present (Wang et al. 2000; Eckert & Hall 2006; Willyard et al. 2007; Gernandt et al. 2008). Be that as it may, their restriction to the northern hemisphere is remarkable, and for the most part they are unable to compete in tropical broad-leaved rain forests (but see Pinus krempfii: Brodribb & Field 2008). There may also have been a bout of diversification in the family in the Palaeocene and more recently (Le Page 2003; Willyard et al. 2007).

A number of rusts, including those on ferns, have their aecial stages on Pinales, especially Pinaceae (Savile 1979b; Durrieu 1980). These include the white pine blister rust, Cronartium ribicola (alternate host Ribes, Grossulariaceae). For a general discussion of resins and defence, see the introduction to Pinales (above), while Mumm and Hilker (2006) discuss the chemical defence of pines against foliovores in particular. Adelgidae (aphids) are restricted to Pinaceae, and include Adelges piceae and A. tsugae, serious introduced pests in North America (Havill et al. 2007). Cecidomyiid gall midges are quite common on the family in North America (Gagné 1989).

The seed coat of Cedrus is vascularized. Cleavage polyembryony is common, as is true polyembryony (more than one archegonium is formed), but the seed generally contains only a single embryo.

Relationships within Pinaceae are unclear, details depending on data analysed (morphology, molecules) and methods of analysis (parsimony, Bayesian: Tsumura et al. 1995; Wang et al. 2000; Rydin & Källersjö 2002; Liston et al. 2006b; Gernandt et al. 2008). For the phylogeny of Pinus, see Syring et al. (2005) and Eckert and Hall (2006); Gernandt et al. (2005) also discuss infrageneric classification. Pinus has two subgenera; leaves of subgenus Pinus, the hard pines, have two vascular bundles, plesiomorphic, while those of subgenus Strobus, the soft pines, have but a single bundle. Analysis of nuclear ITS variation was largely uninformative in suggesting relationships between sections in Abies, but at lower levels was more useful (Xiang et al. 2009).

Additional information: for Pinus, see Farjon (2005a: monograph); for other Pinaceae, see Farjon (1990: general), for ovuliferous cone morphology and anatomy, see Hu et al. (1989) and Napp-Zinn and Hu (1989), and for the embryo, see Buchholz (1931).

[Araucariaceae + Podocarpaceae] [Sciadopityaceae [Cupressaceae + Taxaceae]]: resin ducts in phloem only; phloem with bands of fibres [stratified].

For other possible synapomorphies of this group, see Hart (1987). Isoflavonoids are known from Cupressaceae, Podocarpaceae and Araucariaceae (Reynaud et al. 2005).

Araucariaceae + Podocarpaceae: roots with nodules; one ovule/scale; proembryo with 5 or 6 free-nuclear divisions.

ARAUCARIACEAE Henkel & W. Hochst. - Branches whorled, plagiotropic; stem apex with tunica/corpus construction; wood lacks resin canals or cells, only resin plugs present; stomata tetracytic; branchlets, not leaves, shed [cladoptosis]; leaves multiveined; ovules inverted; pollination droplet 0, pollen germinates on ovuliferous scale [?Araucaria], prothallial cells numerous; seed winged [the wing is the entire bract scale] or not; free nuclear stage in proembryo multinucleate, central. - 3/33. Southern South America, Malesia to E. Australia and New Zealand).

Araucariaceae are well known as fossils from the Mid Jurassic onwards, Araucaria in particular having been found in Triassic deposits in many parts of the world in both hemispheres (Florin 1963; Stockey 1982, 1994; Hill & Brodribb 1989). Sequeira and Farrell (2001) suggest that the association between Araucaria and the scolytine Tomicini bark beetles that feed on them is probably Cretaceous in age; the beetles seem to have moved on to Araucaria from angiosperms, and from thence moved on to Pinaceae. Caterpillars of Agathiphagidae, a small group of lepidoptera the adults of which have jaws, are found on the family (Shields 1988).

The recent discovery very close to Sydney of a few trees of the remarkable Wollemia, very similar to some fossil Araucariaceae ca 90 million years old (see e.g. Chambers et al. 1998; Pastoriza-Piñol 2007 for a general account), has occasioned some excitement. It has variously been placed sister to Agathis or sister to the rest of the family (Gilmore & Hill 1997; Setoguchi et al. 1998).

The single leaf trace divides into three or more as it proceeds into the leaf. The stomata have a wax plug. The pollen grains lack sacci and do not rupture when placed in water (Tomlinson 1994).

For general information, see Stockey (1982), for comparative anatomy, see Thompson (1913), for details of reproductive biology compared with those of other Pinales, see Owens et al. (1995a, b, c), for pollen morphology, see Dettmann and Jarzen (2000), for vegetative construction, see Tomlinson (2008), and for possible apomorphies, perhaps including "dehiscent" seeds (i.e. separating from the cone-scale), see Cantrill and Raine (2006).

PODOCARPACEAE Endlicher - Roots with nodules [modified lateral roots]; sclereids numerous, with large lumina; transfusion tissue in patches lateral to the vascular bundle in the leaf; microsporophylls with two sporangia; pollen saccate; male gametophytes with 3-6(-8) prothallial cells, male gamete binucleate, whether or not one nucleus is extruded; ovules ± inverted (erect); proembryo cells [of the E tier] binucleate; seed with epimatium. - 18/130: Podocarpus (100), Dacrydium (20). Largely southern Hemisphere, scattered, N. to Japan, Central America and the Caribbean.

Podocarpaceae are known as fossils from as early as the early Middle Triassic (Axsmith et al. 1998); although quite common, they are largely restricted to the southern hemisphere, including Antarctica. For biogeographyic relationships in the family, see Mill (2006). The New Caledonian Parasitaxus ustus is parasitic on the roots of Falcatifolium taxoides, another podocarp.

For nodulation, see Becking (1965) and Russell et al. (2002); the fungus Glomus is involved, and nitrogen does not seem to be fixed in the nodules.

RbcL analyses (Conran et al. 2000; Wagstaff 2004b) tend to result in Phyllocladus being embedded in Podocarpaceae, other analyses, whether (Quinn et al. 2002) or not (Sinclair et al. 2002) also including rbcL sequences, have the two as sister groups. Peery et al. (2008) using the nuclear XDH gene also found Phyllocladus to be embedded in Podocarpaceae, and it is looking as if that is where it will have to go.

The morphological nature of the epimatium is controversial; Chamberlain (1935) interprets it as possibly being equivalent to the ovuliferous scale. For cone and ovule development, see Tomlinson et al. (1989), for pollination, see Tomlinson (1994, 1997: useful comparative table), Tomlinson et al. (1991) and Rydin and Friis (2005: correlation between absence of wings and the pollen exine being shed on germination), for nucleus number in the E-tier cells, see Quinn (1986), and for phylogeny, see Kelch (1998: comparison of morphology and molecules).

Phyllocladus - Phylloclades +, leaves reduced to scales; pollination droplet actively resorbed; ovule erect; seed arillate. - 1/ca 5. The Philippines (N. Luzon) to Australia (Tasmania) and New Zealand.

Phyllocladus has phylloclades, flattened, photosynthetic stems; these bear highly reduced, scale-like leaves, and it is the axils of these leaves that the reproductive structures are found.

Phyllocladus, with its phylloclades, highly reduced leaves that may lack any associated leaf gaps, active pollen capture, etc., has long been considered very distinctive, sometimes being separated from all other conifers (e.g. Keng 1974, 1979; cf. Quinn 1986). Its pollen has often been described as having a wing (e.g. Singh 1978), but this seems to be absent. The seedling has needle leaves.

Sciadopityaceae [Cupressaceae + Taxaceae]: pollen without sacci, exine shed on germination [microgametophyte naked], prothallial cells 0.

The pollen grains expand and rupture when placed in water (Tomlinson 1994), and the intine-clad pollen may deform more easily and so be tranferred along the narrow micropylar canal (Takaso & Owens 2008). Whether or not all taxa have male gametes each surrounded by cell walls needs to be confirmed (see Singh 1978).

SCIADOPITYACEAE Luersson - Leaves reduced to brown scales, short shoots with photosynthetic cladodes [modified stems]; microsporophyll with flattened apical expansion; 7-9 inverted ovules/ovuliferous scale. - 1/1: Sciadopitys verticillata. C. and S. Japan.

Sciadopityaceae have rather stout, needle-like leaves borne in loose whorls at the ends of the innovations.

Fossils of Sciadopitys are known from the Upper Cretaceous and are common in the European Tertiary.

There has been much debate over whether the photosynthesizing structures of Sciadopitys are phylloclades - perhaps formed by the connation of two leaves - or cladodes, basically stem structures. They are borne axillary to scales, as are the clusters of needles in pines. The "leaves" have two vascular bundles, each with its own endodermis and with abaxial xylem and adaxial phloem, a rather odd arrangement, and Sporne (1965) notes that on occasion branches develop from them. In general, that the "leaves" are cladodes is the favoured hypothesis (see also Farjon 2005c).

For pollen, see Page (1990), for a monograph, see Farjon (2005c).

Cupressaceae + Taxaceae: cone scales opposite.

CUPRESSACEAE Bartling - Ovules erect to inverted. - 29/140: Juniperus (52 [ca 50-70 - Adams 2004]), Callitropsis (18), Callitris (14), Cupressus (12). Northern hemisphere, more scattered in south temperate region, also NE Africa.

The telial stage of the basidiomycete Gymnosporangium rust in which thick-walled two-celled binucleate spores are produced is common on some Cupressaceae; the aecial stage (production of thinner-walled binucleate spores) characterises Rosaceae-Maloideae (Savile 1979b). In Cupressus dupreziana paternal apomixis, a phenomenon unknown from any other seed plant, occurs; here the embryo develops from unreduced male gametes (Pichot et al. 2001).

Characters of wood anatomy may yield phylogenetically interesting variation, but state delimitation is difficult. Proliferation of the ovuliferous cones is common, and the distribution of this feature may also be of phylogenetic interest (Schulz & Stützel 2007). Scales on the ovuliferous cones are wedge-shaped to peltate. A number, perhaps a majority, of Cupressaceae lack ovuliferous scales, having only bract scales (Zhang et al. 2004; see also Farjon 2005c), while Cryptomeria has several "teeth" on the ovuliferous scale - perhaps a reversion to a plesiomorphic morphology (see also Schulz & Stützel 2007).

Cupressaceae s. str. are embedded in a paraphyletic Taxodiaceae which form a basal grade (Quinn et al. 2002); phenetic analyses had earlier suggested the combination of the two (Eckenwalder 1976). For phylogenetic relationships within Cupressaceae, see Gadek et al. (2000), Farjon et al. (2002), Brunsfeld et al. (2003), and Little et al. (2004). For generic limits around Cupressus, which has turned out to be polyphyletic and is now restricted to the Old World, see Xiang and Li (2005) and especially Little (2006).

For cone morphology, see Farjon and Garcia (2003), and for a monograph (and far more) see Farjon (2005c).

TAXACEAE Berchtold & J. Presl - Plant dioecious; pollen inaperturate; ovule solitary, erect, on shoot in axils of vegetative leaves; male gametes unequal in size. - 6/30. Scattered in the Northern Hemisphere, esp. South East Asia, also New Caledonia.

For morphology, see Hart and Price (1990), for megasporangiate shoots, see Liang and Wang (1989), and for a general account, see Cope (1998).

GNETALES Luersson [Back to Index]

Stem apex with tunica/corpus construction; vessels + [perforations derived from circular bordered pits]; mucilage cells +; stomata mesogenous [subsidiary cells produced by the same cell that gives rise to the guard cell initials]; leaves opposite, joined at the base, with collateral buds; plant dioecious, strobili compound [micro- and megasporangium-bearing structures closely associated], bracts opposite; synangia present, surrounded by a tubular "bract", dehiscing apically by the action of the epidermis [exothecium], pollen striate, with granular layer under the tectum; ovules terminal, erect, inner integument with much-elongated beak, not vascularized, surrounded by vascularized connate structure ["outer integument"]; male gametes binucleate, both nuclei fuse with female gametes; seed with outer fleshy and inner sclerenchymatous layer derived from the outer integument; secondary suspensor developing from upper embryonal tier, no primary suspensor. - 3 families, 3 genera, 96 species.

Crane (1996) summarized the fossil history of Gnetales. Both Ephedra and Welwitschia have distinctive striate or polyplicate pollen of a kind that has a fossil record of ³250 million years. This Gnetalean-like pollen was common in both the Northern and Southern Hemispheres; in the former, records are from the Upper Triassic onwards, in the latter, it is especially common in the early Cretaceous from the northern half of South America (Dilcher et al. 2005). Ephedra and Welwitschia themselves may have diverged by 110 million years before present or more, given the South American welwitschioid seedling, Cratonia, that is of this vintage (Rydin et al. 2003). Detailed studies of small Early Cretaceous seeds suggests that Erdtmanithecales and Bennettitales have seeds very similar to those of Gnetum and Welwitschia in particular, the latter order agreeing in details of micropylar closure, and all have paracytic stomata (Friis et al. 2007, 2009; Mendes et al. 2008; cf. Rothwell et al. 2009). A further link with Ephedra is in the granular infratectum of the pollen that all share (Friis et al. 2007). Gnetales s.l., i.e., stem-group Gnetales and including these two wholly fossil groups, show a considerable amount of variation.

Ovules of all three extant genera are visited by diptera, which are often pollinators, and moths may also visit; sweetish droplets exude from the micropyle (see Labandeira 2005 for references).

Note that there are substantially different interpretations of the parts of both the microsporangium- and megasporangium-bearing structures (e.g. Gifford & Foster 1989; Hufford 1997a; Mundry & Stützel 2004). In the microsporangiate plants in all three genera both stamens and non-functional ovules (pollination droplets may still be produced) are closely associated, although this perhaps least marked in Ephedra, and the microsporangiate cones can be interpreted as being compound (Mundry & Stützel 2004), rather like the megasporangiate cones of Pinales. The plants themselves are functionally dioecious.

Within Gnetales relationships are clearly [Ephedra [Gnetum + Welwitschia]] (e.g. Price 1996).

For the morphology of Gnetales in the context of that of fossil gymnosperms, see Doyle (2006, and references), see also Osborn (2000) and Rydin and Friis (2005), both pollen, Kato and Inoue (1994: pollination), Martens (1971: detailed treatment), Gifford and Foster (1989: summary), Carmichael and Friedman (1996) and Friedman and Carmichael (1997), both double fertilisation, Carlquist (1997: wood anatomy), Endress (1997: megasporangium), and Hufford (1997a: microsporangium arrangement).

EPHEDRACEAE Dumortier - Xeromorphic; nodes 1:2; leaves reduced, or at least without a lamina; microsporangiophores with 2-8 synangia, each with 2(-4) sporangia, dehiscence porose, pollen lacking a colpus, exine shed on germination; archegonia exposed at base of deep pollen chamber; each nucleus of free-nuclear stage forms an embryo; seed with papillae on the inner side of the outer covering. - 1/65. North (warm) temperate, W. South America; drier habitats.

Ephedraceae are rather small, shrubby, much-branched, xeromorphic plants that may be readily recognised by their green, photosynthetic stems, often reduced opposite leaves and small cones borne along the shoots.

The distinctive pollen of Ephedra has been found inside fossil seeds that are morphologically also Ephedra in deposits that date from the late Aptian to Early Albian (early Cretaceous) from Portugal, suggesting that diversification in the genus, previously thought to be recent, 32-8 million years before present, may be much older, i.e. 127-110 million years before present (Rydin et al. 2004, but cf. Huang & Price 2003). Indeed, fossils of Ephedra with "modern" morphology from the early Cretaceous seem to be widespread, E. paleorhytidosperma having distinctive seeds very like those of the extant E. rhytidosperma (Yang et al. 2005). Other fossils apparently assignable to Ephedraceae are known from perhaps a little earlier in the lower Cretaceous in China (Zhou et al. 2003).

Because the pollen exine of Ephedra is shed on germination, the male gametophyte is naked. Fertilisation occurs only 10-15 hours after pollination largely because (). The "outer integument" surrounding the ovule may become fleshy and brightly coloured, or it may be dry and form a wing, or be faintly nondescript, the seeds being dispersed by scatter-hoarding rodents (Hollander & Vander Wall 2009).

Species of Ephedra are pharmacologically very active and contain a number of distinctive secondary metabolites (Caveney et al. 2001). For nodal anatomy, see Marsden and Steeves (1955).

Gnetaceae + Welwitschiaceae: vascular pits lacking central torus; nodes multilacunar; branched sclereids +; stomata paracytic; male gametophyte with one prothallial cell and no sterile cell; ovule with additional pair of bracts; megagametophyte development from all four megaspores [tetrasporic], no archegonia per se, female gametophyte lacking radial arrangement of cells [alveolation does not occur]; some cells of embryonal mass elongate, embryo all cellular, with lateral "feeder" [protrusion of the hypocotylar axis].

GNETACEAE Lindley - Vessel elements with vestured pits; sieve tubes with companion cells [derived from different mother cells]; laticifers +; leaves petiolate, blade broad, with more than two orders of reticulate venation; ovules and microsporangiophores at same node in staminate plant; microsporangiophore with (1-)2(-4) sporangia, pollen not striate, surface spinose; additional pair of bracts connate [i.e. there is a third integument]; proembryo initially with elongated suspensor tubes, nucleus at end divides forming an embryonal mass. - 1/30. Tropical, rather disjunct.

Gnetaceae are woody, evergreen plants with opposite, angiosperm-like leaves whose bases surround the stem; axillary to the leaves are a series of strobili.

Entomophily has been reported from Malesian species of Gnetum, moths visiting the pollination droplets (Kato & Inoue 1994). For biogeographical relationships in the genus, a story of post-Eocene diversification and dispersal, see Renner (2005b) and Won and Renner (2006).

Not surprisingly, the wood of the lianoid taxa is distinctive, with serial cambia being formed. See Martens (1971) for the vascularization of the leaves; pairs of vascular bundles leave the central stele in close proximity. There is vascular tissue in the outer two coverings of the ovule, but vascular bundles barely enter the base of the inner integument.

WELWITSCHIACEAE Caruel - Successive cambia + [in root - derived from phelloderm, the innermost tissue coming from the cork cambium]; three pairs of leaves only, the second pair long-ligulate, persisting for the life of the plant and elongating from the base, with stomata on both sides, venation parallel; ovules and microsporangiophores in intimate association; microsporangiophores 6, basally connate, with synangia of three sporangia, dehiscence radial; additional pair of bracts free; megagametophyte with multinucleate cells, some grow upwards through nucellus forming female gametophytic tubes, fertilisation in apical bulge [both gametes involved?], proembryo pushed back down tube by elongating embryonal suspensor. - 1/1: Welwitschia mirabilis. S.W. Africa.

Welwitschia becomes a rather massive plant that can immediately be recognised by the two, long leaves that ensheath the stem; these become curled and are dead and frayed at the apices. The plant has no trunk as such, but the stem becomes quite thick.

Welwitschia mirabilis grows in the Namib desert close to the ocean where precipitation comes in the form of condensed fog. The plants may be some hundreds of years old, the two persistent leaves growing at the base and fraying at the apex. Pollination appears to be by diptera (Wetschnig & Depisch 1999).

Cratonia cotyledon is a fossil seedling with distinctive cotyledon vasculature very like that of the leaves of Welwitschia, the secondary veins leaving from the primary veins fuse to form an inverted "Y" (Rydin et al. 2003). Cratonia was found in N.E. Brazil and is late Aptian or early Albian in age, perhaps 114-112 million years before present, and other fossils of welwitschiaceous affinity have been found in the same area (Dilcher et al. 2005).

Because of the abundant, branched sclereids, "One might as well try to cut sections of a thick Scotch plaid blanket as to try and cut a stem of Welwitschia without imbedding." (Chamberlain 1935, pp. 388-389).

MAGNOLIOPHYTA/angiosperms [Back to Index]

Plant woody, evergreen; non-hydrolysable tannins, quercetin and/or kaempferol +; lateral roots arise opposite or immediately to the side of [when diarch] xylem poles; cork cambium in root deep-seated; stem with 2-layered tunica + corpus construction; stem cork cambium superficial; circular bordered pits lacking margo and torus; sieve tubes eunucleate, with sieve plate, companion cells from same mother cell that gave rise to the sieve tube; sugar transport in phloem passive; nodes unilacunar [1:?]; stomata paracytic; leaves petiolate, lamina [formed from the primordial leaf apex], development of venation acropetal, 2ndary veins pinnate, fine venation reticulate, dense, with some free vein endings; axillary buds present; 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, with a single trace, introrse, filaments stout, anther ± embedded in the filament, tetrasporangiate, dithecal [sporangia in two groups of two], each theca dehiscing longitudinally by action of hypodermal endothecium, pollen subspherical, binucleate at dispersal, trinucleate eventually, germinating in less than 3 hours, siphonogamy, tube elongated, growing at 80-600 µm/hour, with pectic outer wall, callose inner wall and callose plugs, growing between cells, male gametes immobile, siphonogamy [gamete delivery via the tube], penetration of ovules within ca 18 hours, nectary 0, G free, several, ascidiate [forming an urn-shaped to subtubular structure], with postgenital occlusion by secretion, ovules anatropous, bitegmic, micropyle endostomal, inner integument 2-3 cells thick, nucellus at apex of ovule 1-3 cells thick, megasporocyte single, megaspore lacking sporopollenin, female gametophyte four-celled [one-modular, nucleus of egg cell sister to one of the polar nuclei], stylulus short, hollow, cavity not lined by distinct epidermal layer, stigma ± decurrent; P deciduous in fruit; seed exotestal; double fertilisation +, endosperm cellular, copious, oily and/or proteinaceous, embryo cellular; germination hypogeal.

Possible angiosperm synapomorphies are in bold. Note that the actual level to which many of these features should be assigned is unclear, partly because many taxa basal to the [[magnoliid + Chloranthales] [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, 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 the immediate outgroup to angiosperms is also uncertain. In particular, for details of variation in embryo sac morphology and endosperm ploidy in members of the ANITA grade, see below.

Within angiosperms, relationships between members of the basal pectinations have recently been clarified. Donoghue and Mathews (1998) listed 16 different hypotheses of relationships among basal angiosperms that involved the the first three nodes. However, Amborellaceae are most likely to be sister to other angiosperms (not one of the hypotheses that Donoghue and Mathews included!), Nymphaeales sister to the rest, then Austrobaileyales - the three making up the ANITA grade (e.g. Mathews & Donoghue 1999; Qiu et al. 1999, 2000; P. Soltis et al. 1999, 2000; Parkinson et al. 1999; Müller et al. 2006; Hansen et al. 2007; Jansen et al. 2007; Moore et al. 2007). There perhaps remains an outside possibility that Amborellaceae and Nymphaeales are sister taxa, the clade they form being sister to the rest of the angiosperms (e.g. Qiu et al. 2000; Leebens-Mack et al. 2005; Cai et al. 2006; Soltis et al. 2007a; Graham & Iles 2009; Goloboff et al. 2009: huge parsimony analysis).

Indeed, other relationships may be obtained when few taxa, each with massive amounts of data, are analyzed (e.g. Goremykin et al. 2004). Sampling is critical in such analyses. 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), so knowing where sampling should be improved is important (Geuten et al. 2007). For example, 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; see also Lockhart & Penny 2005). If Poaceae are included as the sole representative of the monocots, the results may be rather strange; its genome is very derived. Although the recent discovery of an association of Hydatellaceae with Nymphaeales (Saarela et al. 2007) unexpectly allows sampling in this area of the tree to be improved, inclusion of this taxon has little affected the topology and support values of the basal part of the angiosperm tree (Graham & Iles 2009). What kinds of characters are analysed may also be important; Goremykin et al. (2009b) found an [Amborella + Nymphaea] clade after removing a relatively few highly variable positions from the analysis. How data are analysed may also affect the results; parsimony is more susceptible than maximimum likelihood or Bayesian methods to long branch attraction. Finally, some kinds of DNA data such as the 18S nuclear gene, some mitochondrial genes, etc., may be positively misleading when it comes to understanding relationships (e.g. Duvall & Ervin 2004; Qiu et al. 2005; Duvall et al. 2006, 2008b; G. Petersen et al. 2006b), and horizontal transfer seems to be notably common in mitochondrial genomes (Sanchez-Puerta et al. 2008).

There are convenient summaries of the copious literature on relationships between the major angiosperm 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 discussion immediately preceding the Magnoliales, i.e. the magnoliid clade, and Acorales (both the [monocot + eudicot] clade and monocots themselves), Ranunculales (eudicots), Gunnerales (core eudicots), Dilleniales (more about core eudicots), Saxifragales (especially rosids), and Cornales (asterids).

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 bark produced by the separate cork cambium; Robinson (1990) noted that the ratio of periderm to wood 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).

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 groups. Secondary xylem of the former has much parenchyma mixed in with the tracheids, while in 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 probably do not represent fundamental distinctions. Extant gymnosperms appear to be monophyletic, but when considering both extant and fossil taxa gymnosperms are paraphyletic; angiosperms are derived from a gymnospermous ancestor.

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.).

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 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). However, bootstrap support for these relationships is very low.

These relationships are now 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). But other relationships appear in 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 thatmorphological data optimised with this topology as a constraint tree also had 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 more derived characters tended to become unrecognisable before less 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)...

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), 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). 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). 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 ro be much parallel evolution were removed, Gnetales were sister to Cupressaceae or to Pinaceae, the latter being the prefered position (Zhong et al. 2010).

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. Furthermore, strobili with both micro- and megasporangia are common as abnormalities in Pinales (Chamberlain 1935), while 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.

Detailed morphological work suggests that angiosperm characters apparently found also in Gnetales are for the most part clearly 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), the tunica has only a single layer, etc. (e.g. Donoghue & Doyle 2000a; Doyle 2006), and the adaxial tension (reaction) wood in Gnetum, produced as the branches react against gravity to maintain their orientation, is unlike that of angiosperms (Tomlinson 2003; see also Höster & Liese 1966). The loss of sperm flagellae and the associated development of a pollen tube growing towards the ovule will also represent 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 yet more parallelisms are involved. In any event, wherever Gnetales are placed, they will be a very derived group.

Thus 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 relatives of angiosperms include Corystospermales (Pteruchus, Ktalenia, etc. - see the mostly male theory of flower evolution, Frohlich & Parker 2000), Bennettitales, and Caytoniales. Indeed, it seems that genes from the male genetic programme of the male cone of gymosperms are active in the outer whorls of the basal angiosperm flower (Chanderbali et al. 2010). However, recent detailed work on seed morphology and anatomy in particular, but also pollen morphology, suggests that Bennettitales should be placed in the BEG group with Gnetales and some other fossil assemblages like Erdmanithecales, which persisted into the Late Cretaceous (Friis et al. 2007, 2009a, the latter describing four new genera in this complex; Mendes et al. 2010). Overall, it seems unlikely that Gnetales have anything immediately to do with flowering plants - and this will be true of those seed ferns like Bennettitales that have been associated with them if current ideas of relationships hold (see Crane & Herendeen 2009 for a careful interpretation of the reproductive structures of Bennettitales) - note, however, that Rothwell et al. (2009) strongly question the idea of a close relationship between Bennettitales and Gnetales. Other work questions any particular similarity between the "flower" of Bennettitales (Rothwell et al. 2008a, 2009) and that of angiosperms, with some morphological analyses removing Bennettitales from the anthophytes and associating them with cycadofilicales plants (Crepet & Stevenson 2009). Chanderbali et al. (2010) find 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.

Baum and Hileman (2006) advance a developmental genetic model for the evolution of the flower, which may help in the interpretation in the significance of particular fossils, while Rudall and Bateman (2010) suggest that the morphology of crown group conifers, being highly derived, may be of little help in thinking about that of the ancestors of angiosperms. The ovule-bearing structures of Caytonia can indeed be linked with the carpels of extant angiosperms by invoking appropriate morphological gymnastics (Doyle 2006 for literature). Thus 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).

Although it has recently been suggested (Wang et al. 2007) that the early Jurassic Schmeissneria, previously placed in Ginkgoales, is angiospermous, having closed carpels (see origin 2 above), it is difficult to see this in the fossils; of course, angiospermy itself may have arisen more than once. In general, pre-Cretaceous angiosperm fossils do not currently convince. 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 determines the putative relatives of angiosperms. Just which gymnosperms are close to angiosperms is unknown.

EVOLUTION AND DIVERSIFICATION OF THE ANGIOSPERMS (This section is very much out of date - but see the Amborellales page.)

Background.

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.

The evolution of crown-group angiosperms has taken place under several different ecological and environmental regimes during the course of their some 140 million years' (or 200+? see below) existence. Thus questions such as, Why are angiosperms so diverse? may have several answers. There are factors that shaped angiosperm diversification through the Cretaceous, or maybe even back to the Permian, and perhaps a rather different set of factors that are connected with early Tertiary diversification, factors that may in turn be in some way connected with or shaped by the bolide impact then. Indeed, much of angiosperm diversity as we appreciate it now seems to be a Tertiary phenomenon. 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 and Annonaceae, crown group diversification may also be largely a Tertiary phenomenon (J. A. Doyle et al. 2004, 2008; Scharaschkin & Doyle 2005; Richardson et al. 2004).

Background

1. Although "diversification" is mentioned frequently below, it, like the related "adaptive radiation", is a very imprecise term and is difficult to estimate (e.g. Sanderson 1998; Davies et al. 2004; Ricklefs 2007; Olson & Arroyo-Santos 2009; Ackerly 2009 for measurement of aspects of radiation).

2. When thinking about evolution in general, a well-supported phylogeny is a sine qua non, as is an understanding of how to interpret trees for morphological features of potential interest - see below. But beyond this, dating is critical. The whole issue of dating 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). 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 (but see Smith et al. 2010). Molecular and paleontological dating can seem to be in conflict, and the former sometimes give substantially older ages than the latter - which after all must only yield a minimum age (Donoghue & Benton 2007), even if in some groups like platanoids (Crepet et al. 2004) the fossil record is more substantial. Many age estimates for orders, families, etc., are given here, but they should all be treated with extreme caution.

3. Understanding fossils is critical to understanding angiosperm evolution (see also 7 below). Weins et al. (2010 and references) is a good example (albeit on reptiles) 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.

However, as with dating, one has a sense of unease, and some studies question what one had thought were 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 of the dates suggested below and cause a rethinking of angiosperm evolution - but there are questions about the age of this fossil (Caroline St�mberg, pers. comm.). 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, 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 plants 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, etc).

4. 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).

5. It is a challenge to think of how novelties, morphological or otherwise, evolve. It is not simply that the genome duplications that appear to be so pervasive in angiosperm - and land plant in general - evolution allow of a myriad possibilies for subfunctionalisation and neofunctionalisation (e.g. de Martino et al. 2006; Soltis et al. 2009 summaries of duplications). 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 is sometimes "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 - the ability of a plant to form an association with nitrogen-fixing bacteria is a case in point (see also Shubin et al. 2009 on deep homology), hence "evolutionary tendencies"! (for which, see Cantino 1985; Sanderson 1991). Whatever the reasons, some characters - and not only those of secondary plant chemistry - seem to come and go on the tree almost willy-nilly.

6. Estimates of the number of 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). Inded, we have to be very careful when discussing clade size. 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. 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 to be the hyperdiverse group. However, recent work is beginning to move far beyond such simplistic comparisons (see 1 above).

7. The relationships of angiosperms to other seed plants and the whens, whys and hows of their initial diversification still remain an abominable mystery (see Davies et al. 2004b; Friis et al. 2005; Frohlich & Chase 2007; Pennisi 2009). 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 diversification of crown angiosperms, i.e. 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), 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 (Smith et al. 2010), there is still a substantial stem history that is largely unknown.

Gymnospermous "origin 1" will be the case whether or not extant gymnosperms are monophyletic or paraphyletic. 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). Individual ovules are radiospermic and lack a cupule, they have a vascularized nucellus but not a vascularized integument, and seeds with an outer sarcotesta, a sclerotestsa, and a layer inside that (Rothwell & Stockey 2002). Interestingly, the triterpenoid oleanane is found pretty much throughout angiosperms, although not in all, and in Bennettitales, in the Permian Giganopteridales (remember, there are vessels there, too!), but not in any extant gymnosperms. This is consistent with a divergence of the angiosperm stem group from other seed plants by the late Paleozoic (Moldowan et al. 1994; Taylor et al. 2006). (For further discussion, see seed plant evolution.) Note, however, that cycadophytes include Bennettitales and Cycadales, and there are similarities in wood anatomy characterizing this group (see Cycadales page; Rydberg et al. 2007).

Recent work aligns bennettitalean pollen and seed closely with those of Gnetales (Friis et al. 2007, 2009a; cf. Rothwell et al. 2009); their ovules are radiospermic and 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). 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 (Pott et al. 2010). Gnetales 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), which, if confirmed, 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). Other pteridosperm groups that have been linked to angiosperms are Caytoniales, Pentoxylon and glossopterids (e.g. Soltis et al. 2008), but in a recent comprehensive review of the bearing of fossil data on the origin of the flower, the overall conclusion was that our understanding of the fossil record was currently insufficient to help much in answering questions of origins (Doyle 2008b).

8. Although we allow that the evolution of angiosperms as intimately connected with and dependent on the evolution of their pollinating insects, most angiosperms are symbiotic systems at additional levels. Importantly, basic angiosperm physiology is mediated by mycorrhizae and bacteria in the soil and endophytes in the plants, and this shapes the global environment.

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 these diversifications 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), this is unlikely, and diversification of angiosperms appears to be contemporaneous with the massive diversification of many insect groups that are now more or less dependent on them; plant-feeding insects make up at least one quarter of all described species (Janz et al. 2006).

Phytophagy. It was early suggested that 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), and 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 these species, 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 alll weevils - as angiosperms became floriticially common some 112-93.5 million years ago, and this was first on monocots, then on core eudicots. Bark weevils (Scolytinae) are less speciose in clades that returned to conifers, and also the weevils that make tunnels in the wood and are associated with the ambrosia fungi (Ophiostoma, Ceratocystis: Ophiostomatales, ascomycetes), an association that has evolved about seven times but is unreversed (Farrell et al. 2001). Of course, there are also very species-rich beetle clades that are not associated with plants (e.g. Barraclough et al. 1998).

Slightly under 100,000 species of butterflies and moths (Lepidoptera) are herbivores, of which some three quarters are mono- or oligophagous (Bernays & Chapman 1994). There are over 4,000 species of aphids (Aphididae - hemipterans) feeding on plant sap; again, diversification was Late Cretaceous/early Tertiary (von Dohlen & Moran 2000). Ants, almost 12,000 species, also seem to have diversified in the late Cretaceous-early Eocene 75-50 million years before present well after the evolution of the angiosperms, although the main clades may have diverged somewhat before; however, ants began to be ecologically dominant only in the Eocene (see below). 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 the specificity of the gall insect/plant association - and so estimates range from (21,000-)132,930(-211,000) species. Cecidomyiidae (dipterans), the most speciose (Yukawa & Rohfritsch 2005), are worldwide but show no particular patterns of host associations, Cynipidae (hymenopterans) are north temperate, while psyllids (jumping plant lice, hemipterans) are particularly common in Australia (Fernandes & Price 1991; Crespi et al. 2004; Espiritó-Santo & Fernandes 2007; Raman et al. 2005). In general, gall-inducing insects are commonest on sclerophyllous plants growing on poor soils in warm climates or perhaps more generally in species-rich communities, whether dry or wet (Yukawa & Rohfritsch 2005; see Price et al. 1987 for galling in an adaptive context).

Plants have evolved mechanical and especially chemical defences against herbivory, and some insects have evolved ways of tolerating these defeences - or even eat only plants with particular defences that they then coopt for their own defence (see Termonia et al. 2001 for chrysomelid leaf beetles). Some plant and herbivorous insect groups seem to be rather closely associated, showing loose coevolution (Ehrlich & Raven 1964 for an early statement of the concept; cf. e.g. Janzen 1980; Kato et al. 2010; see Futuyma & Agrawal 2009 for an evaluation of the concept and literature); these are discussed further after individual orders and families. In general, more related plants do have more similar animnals eating them (Weiblen et al. 2006; see Futuyma & 2009 for literature). 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). Most plant secondary metabolites show considerable homoplasy. Since some herbivorous insects effectively track these 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. Herbivorous insects may sequester secondary metabolites 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; 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, however, 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). Host plant relationships may be conserved, even if there is little evidence for strictly parallel diversification (Winkler & Mitter 2008).

Although the evolution of insect/plant associations is in general not well understood, Futuyma and Agrawal (2009: also other papers in Proc. National Acad. Sci. U.S.A. 106()) summarize much of the literature. Obviously, we need to know both the timing of diversification and patterns of phylogenetic relationships in both groups, and all seem to be in flux. 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 plant-insect relationships difficult, idiosyncracy being utterly central to the nature of chemical coevolution (Berenbaum & Zangerl 2008: p. 806); this whole problem can only be exacerbated by adding the difficulty of understanding ecological relationships over time. Secondary metabolites involved in these plant-insect relationships - and secondary metabolites in general - seem to have a very scattered distributions, and its 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.

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 seem to have occurred as butterflies moved on to new groups of food plants and specialized on then (Janz et al. 2006).

Pollinators. Insect pollination is not restricted to angiosperms. It probably occurred in Mesozoic gymnosperms as well, the most likely groups of pollinators being beetles, mecopterids (scorpion flies, mecoptera - perhaps) and true flies, although other groups may also have been involved (Labandeira 1998; Labandeira et al. 2007; Ren et al. 2009). Beetle, fly and moth pollination is known in both extant gymnosperms (Kato & Inoue 1994; Schneider et al. 2002; Oberprieler 2004; Labandeira 2005: note that dioecy is common) and angiosperms, where pollination by such insects predominates in taxa in the ANITA grade, Annonaceae (magnoliids), Araceae, etc. - and these appeared quite early in the Cretaceous; bees are also not prominent as pollinators in extant members of these clades. Beetles, etc., like 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; Endress 2001a; Weberling 2007), 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 a remarkable number of which had inferior ovaries (e.g. Crane et al. 1995; Friis et al. 1999) and it is likely that they were aggregated into inflorescences to attract pollinators (Friis et al. 2006b). 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 this may be connected with the fact that the endosperm is diploid (see below). See Specht and Bartlett (2009), Endress (2010a), etc., for a survey of floral morphology and biology in basal angiosperms.

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). 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. Pollen seems initially to have been produced in rather low quantities, but by the mid-Cretaceous it became more abundant and is more often found in clumps, suggesting that the pollinators, perhaps bees, were more specialized (Hu et al. 2008). However, there is no obvious signal of an increase in protein content of the pollen of representatives of extant angiosperm clades that might suggest a shift in pollinating agent to bees (see Roulston et al. 2000). Triaperturate - or multiaperturate - pollen may germinate faster and 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 morphology (see e.g. Roulston et al. 2000; Fernández et al. 2009). Nectar from specialized nectaries was unlikely to have been a common reward initially, 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). 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); even some fossil Fagales and Platanales had nectaries (note that Proteales and Sabiaceae may also have receptacular nectaries). Septal nectaries may be an apomorphy for monocots, being scattered through that clade, and are certainly found in some Alismatales. Nectaries of various other types are scattered through angiosperms other than monocots and core eudicots.

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. Bees initially diversified in the early to mid Cretaceous (see also 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). 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 [sic], 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 stingless bee (Apidae - Meliponini: Michener & Grimaldi 1988). Apidae and Megachilidae are long-tongued bees, and both these groups are known from Baltic amber of Eocene age (Danforth et al. 2006 and references). Many of the bees in these clades, including the Melittidae s.l., show considerable host plant specificity, being oligolectic. This feature may even be an apomorphy for all bees, and 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 (see also Michener 2007). Of course, bats, birds, mammals, as well as other insects are effective pollinators, and 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); Schaefer and Ruxton (2009, 2010) and Schiestl et al. (2010) discuss the exploitation of perceptual biases of the pollinator by the plant, while Rodríguez et al. (2004) discuss monosymmetry from the point of view of the bee.

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 has been checked. There are discussions on the evolution and ecological significance of mycorrhizae has been widely discussed (see Malloch et al. 1980; Read et al. 2000; Landis et al. 2002; Taylor et al. 2009; etc.), the morphology of the plant/fungus interface (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 Radhika & Rodrigues 2007 and references for records, also de Marins et al. 2009), but the frequent absence of mycorrhizae in Caryophyllales, Proteales, etc., is interesting.

Vesicular-arbuscular mycorrhizae (endomycorrhizae) are very widespread. In vesicular-arbuscular mycorrhizae the aseptate hyphae are intracellular, often forming vesicles or branching structures called arbuscules within the cells. Note that there is substantial variation both in the morphological details of the fungus-plant association (e.g. Smith & Smith 1997) 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 they are found in about 70% of seed plants. 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). Sexual reproduction in the fungus is at most exceedingly uncommon, but the spores are multinucleate, the unit of selection perhaps being the individual nucleus (Jany & Pawlowska 2010). In such mycorrhizal associations, nutrient uptake by the plant - especially of phosphorous, although recent work suggests that nitrogen is also involved - is increased, as well as water uptake is improved (Govindarajulu et al. 2005; Leigh et al. 2009 and references; Tian et al. 2010), although it is fair to say that we still know little of the complexity of the plant-mycorrhizal interactions (Whitfield 2007). It is becoming clear that 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). Overall, Glomeromycota are not very speciose, perhaps some 300 species, and a number of these species have relatively limited distributions (Öpik et al. 2010).

Ectomycorrhizal plants often dominate the communities in which they occur and are particularly prominent in (cool)temperate areas (see below); 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. 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, and some fungi both form associations with Ericaceae and also ordinary ectomycorrhizal associations, as with Pinus (Villarreal-Ruiz et al. 2004). Ectomycorrhizae are commonly found in plants growing on rather extreme soils, either poor in nutrient and/or rich in organic materials, especially in (cool) temperate and tropical montane habitats (e.g. Malloch et al. 1980). Ectomycorrhizae may obtain N and P from things like pollen and dead nematodes, and these are transferred to the plant (Read & Perez-Moreno 2003). Ectomycorrhizae also 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 these associations, whether facilitating the establishment of the association (mycorrhiza helpers) or as necessary for the functioning of the established association (Frey-Klett et al. 2007 and references). Thus 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).

Te complexity of ectomycorrhizal and other fungal associations is considerable. Thus liverworts may form associations with the same fungi that form ectomycorrhizae with the flowering plant on which the liverwort is found (Marchantia/mycorrhizal fungus/Podocarpus, the mycoheterotrophic chlorophyll-less Cryptothallus/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; this fungus can also be ectomycorrhizal on Pinus growing together with 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).

Endophytes, fungi growing inside plants, have been broadly categorized. Clavicipitaceous endophytes, class one endophytes, occur in grasses. 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 in endophytic Poaceae in particular, or it may be by 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 ("Olacaceae") (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 again, what they are doing is 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 traditionally have been 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.).

Things get still more complicated 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 and references). Such bacteria may affect the growth of endophytic fungi, and they may be vertically transmitted like the fungi (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). Various kinds of bacteria are involved both in the establishment and the functioning of ectomycorrhizae (Frey-Klett et al. 2007). 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 have been shown to be synthesised by fungal or bacterial associates of the plant (e.g. 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, Convolvulaceae and especially Poaceae are also distinctive in this regard. Indeed, such substances function as if they were endogenous plant metabolites (Zhang et al. 2009). Some distinctive metabolites produced by the host may be a response of the plant to the presence of a particular endophyte (Popay & Rowan 1994). Bacteria may also be involved in similar associations. 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 mitochondia 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.

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 lack mycorrhizae, 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).

There are connections at the molecular level between the establishment of mycorrhizal associations and the development of actinomycete and Rhizobium symbioses 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.

Cretaceous History

Age estimates of crown angiosperms vary, and 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 are also suggested by Magallón (2010: 275-215.6 million years). 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 the early age for crown angiosperm origin - some 270 to 182 million years ago - of Smith et al. (2010) are correct, we are faced with a series of questions. We are not immediately any closer to understanding which pteridosperms are close to angiosperms, and 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 may also need to think of a different set of gymnosperm reproductive structures to link with the angiosperm flower (see Doyle 2008b; Specht & Bartlett 2009; Doyle & Endress 2010 for literature).

The most comprehensive review of Cretaceous angiosperm history is that of Friis et al. (2006a), on which this section draws heavily (see 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 (Endress 2008c and references). Quite "derived" features may be evident in very 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).

Early angiosperms may have been smallish tropical trees that tolerated shady, humid and disturbed conditions (Feild 2005; Feild & Arens 2005). 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 of early angiosperms seem to have been rather small, although most seem to have been insect-pollinated (e.g. Crane et al. 1995; Friis et al. 2000, 2006b, 2010b; Hu et al. 2008 for literature). Nectaries are uncommon, but it should be noted that bee larvae obtain their fat from pollenkitt (Renner 2010) which is produced by the degeneration of the tapetum and is rich in plastid-derived lipids, so pollen was probably the reward. Early angiosperms may have been rather small plants (Friis et al. 2010b).

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. Quite a diversity of strange-looking putative angiosperms have been discovered in northeastern Chinea, although the identities of a number are disputed (Sun et al. 2006).

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). Indeed, Chanderbali et al. (2009, esp. 2010 and references), promote a fading borders/sliding boundaries model of floral evolution, 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.

Fossils assignable to Chloranthaceae are known from the late Barremian ca 120 million years ago onwards, with some fossils being very like the extant Hedyosmum (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 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. Doyle and Endress (2010) should be consulted for their 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.

Tricolpate pollen, the signal of eudicots, is known from this Late Barremian-Early Aptian age some 125-120 million years before present (e.g. Magallón et al. 1999; Sanderson & Doyle 2001 - in W. Portugal from Early Albian remains ca 112 million years old, Heimhofer et al. 2005). Such pollen may germinate faster, even if the pollen itself remains 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 perhaps not surprising that early fossils of the monocots, a predominantly herbaceous group, are not very common. Note that based on a recent re-evaluation of relationships between extant members of early-diverging angiosperm clades, Moore et al. (2007) suggest that there was rapid separation of the Chloranthales, magnoliid, monocot, eudicot and Ceratophyllales clades some time between 148.6-135.5 million years ago (but remember the problematic Poaceae fossils). Eudicots diversified rapidly 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). Although fossils of this age continue to have odd assemblages of characters (see also Friis et al. 1995), the major groups of monocots, asterids, and rosids were all probably diverging by the earlier part of the Cretaceous (Sanderson et al. 2004).

There has been considerable work on the evolution of seed and fruit size and dispersal (see also below). Tiffney (1986a) suggested that seeds of early angiosperms were mostly small 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.

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 did decline) 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 conifers do seem to have become more restricted in terms of areas where they remained common, 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). 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). At the same time, 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 ealy Aptian culminated in the Cenomanian-Turonian thermal maximum ca 99 million years ago (Heimhofer et al. 2005).

Flowers of core eudicots are first known unequivocally from the Cenomanian some 96-94 million years ago (Basinger & Dilcher 1984). The flower they described, called the Rose Creek fossil and not assigned to any extant family, is relatively large compared to the tiny flowers so common in other Cretaceous angiosperms and the five stamens are somewhat unexpectedly opposite the petals. However, the pollen record suggests core eudicots may have been around since the Albian 125-112 million years ago. Core eudicots 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). Rosids sem to have been common in the Let Cretaceous (Friis et al. 2010b), and within the rosids, plants with distinctive pollen assignable to the Normapolles complex (Fagales) were both diverse and ecologically very prominent in north temperate regions starting in the Late Cenomanian and Early Turonian ca 99 million years ago (there were different pollen provinces to the north and south: e.g. Pacltová 1981 for a review; Kedves & Diniz 1983; Friis et al. 2006b, 2010b). Major clades within Malpighiales that currently make up important elements of lowland tropical rain forests today seem to have appeared in the later Cretaceous some 114 million years before present (119.4-101.1 million years before present: Davis et al. 2005a); both Ericales and Malpighiales are now particularly notable as small trees in tropical rainforests. 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). Note, however, that modern representatives of the clades that diverged then represent much more recent radiations (see below). Similarly, 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; Crepet et al. 2004, etc.). Fossil wood of angiosperms is known from deposits of up to about 120 million years ago, although its assignment to clades is not easy (Oakley et al. 2009).

Overall, 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 for further references). 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. However, later in the Cretaceous not only did diversity increase, but angiosperms may have achieved at least some measure of ecological dominance (Friis et al. 2006b).

However, just how ecologically dominant angiosperms were in the later Cretaceous is unclear. Bond and Scott (2010) suggest that until the Mid Cretaceous angiosperms were mostly small herbs to small trees of the understory growing in dryinsh conditions, and at least some early leaf floras from Portugal have leaves that are small in size and xeromorphic in appearance (Friis et al. 2010b). Friis et al. (2006a) note a dramatic increase of phylogenetic diversity and ecological abundance of angiosperms in the Late Albian-Cenomanian, i.e., towards the end of the Cretaceous, while Crepet (2008) records 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. 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. 2000), 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). 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). Nevertheless, in their discussion of Cretaceous angiosperm ecology, Wing and Boucher (1998, p. 379) concluded that by 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). 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 (Upchurch & Wolfe 1987; Crane & Lidgard 1990).

Diversification of the major clades of rosids may have occurred (114-)108-91(-85) million years ago, and that of Fabidae and Malvidae very soon after, (113-)107-83(-76) million years ago (Wang et al. 2009). Saxifragales, although not very speciose, may 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). Similarly, separation of several clades within Malpighiales was estimated as occurring some time in the Cretaceous-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 in Malpighiales seems to have been rapid, and relationships within the order were for some time represented by a major polytomy (Wurdack & Davis 2009; cf.). Major divergence dates within monocots and asterids are also in this general time zone. Rosids in particular currently include a preponderance of denizens of the lowland tropical rainforest, and stem-group Rafflesiaceae are estimated to have diverged from other Malpighiales some 95 million years ago, with Sapria splitting off (86.6-)73.2(-60.8) million years ago, suggesting the presence of rainforest by then (Bendiksby et al. 2010). Clades like the ectomycorrhizal Fagales are a little more than 100 million years old. Epiphytic ferns are commonly found growing on angiosperms and prefer humid conditions; they started 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).

Tertiary Diversification

Details of the effect of the end-Cretaceous bolide impact 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 bolide impact ca 65.5 million years ago caused up to 80% loss of plant species, at least locally (Upchurch & Wolf 1987), and there was concomitant extinction of diet-specific herbivorous insects (Labandeira et al. 2002; Wilf 2008). The impact of the bolide seems to have had a considerable effect on the vegetation in North America and led at least locally to "sudden ecosystem collapse", even common plants not transgressing the boundary (Wilf & Johnson 2004), although the severity of the impact seems to have depended in part of physical location (Johnson & Ellis 2002). Events in Colombia are reflected more by changes in ecological structure, less in extinction (De la Parra et al. 2007; see Graham 2010 for a study of the vetational history of Latin America). In New Zealand the iridium anomaly associated with this impact was followed by a thin layer high in fungal remains (although the original vegtation seems to have returned quite quickly), and 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).

It has been suggested that recovery of the vegetation may have taken anything from only a few thousand years (Vajda & McLoughlin 2007) to over a million years (McElwain & Punyasena 2007); 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 seed plants, both insect-pollinated and/or evergreen taxa suffered more than wind-pollinated and/or deciduous taxa (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). Although after evaluating all the evidence, Nichols and Johnson (2008) suggested that no major plant group disappeared, even if species may have, we shall see that Eocene and later vegetation is very different from that of the Late Cretaceous.

Fawcett et al. (2009) have dated a series of genome duplications in angiosperms to the general time of 70-57 million years ago, 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); genome duplication may reduce the probablity of extinction by increasing environmental tolerance and genetic variation, while mutations would be less likely to have an immediate effect (note that Wood et al. 2009 suggest that polyploidy affects cladogenesis, less the diversification rate of polyploid clades).

Climatic changes also occurred at about this time. Upchurch et al. (2007; see also Zachos et al. 2008) suggest that the climate in the early Tertiary around 55 million years ago was much warmer and wetter (the post-Eocene thermal maximum - PETM); temperatures went up ca 5^oC, over 2,000 gigatons of carbon were released in ca 10,000 years and the concentration of carbon dioxide was at a high then, so conditions favourable for plant growth in part would in part counter any lingering effects of the bolide impact. 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; Boyce & Lee 2010). Fast decomposition of their litter may also have speeded nutrient cycling, again, to their own advantage (Cornwell et al. 2008; Berendse & Scheffer 2009).

In North America there may have been initial recolonization by swamp-loving plants, which survived the impact better (Johnson 2002). 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, both plant and herbivore diversity were rather low. 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). Palaecene Patagonian vegetation was more diverse than its North American counterpart (Iglesias et al. 2007), and other American early fossil floras seem notably diverse (Jaramillo et al. 2006). Angiosperm diversity in the tropics and warm temperate areas was rather low during the Palaeocene (Wilf 2008) but peaked in the Eocene as it rebounded from the bolide impact, perhaps even topping today's levels. 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 (e.g. Upchurch & Wolf 1987; Schuettpelz 2006; Boyce et al. 2009: Schuettpelz & Pryer 2009; Watkins et al. 2010; Bond & Scott 2010).

After the PETM, diversity then decreased strongly through the Miocene, but rebounded; of course, the plants and ecological conditions involved may have been rather different from those of today, although many fossils from this period seem to be assignable to extant genera and any difference is 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. Manchester et al. 2009: East Asian endemics; Plaziat et al 2001: Nypa; Smith et al. 2008: Cyclanthaceae), indeed, 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).

Turning now to individual clades, it seems that common Cretaceous plants often failed to survive into the Tertiary, and that the familial composition of Early Tertiary forests differed from that of their Late Cretaceous counterparts (e.g. Johnson 2002; Wilf & Johnson 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. Indeed, orchid diversification (ca 22,000 species) seems to have been a largely Tertiary phenomenon (Ramírez et al. 2007; Gustafsson et al. 2010), as is that of Asteraceae (23,000+ species: K.-J. Kim et al. 2005; Funk et al. 2009c for a summary), Fabaceae (19,000+ species: e.g. Bruneau et al. 2008b; Bello et al. 2009), etc., while in much older clades like Myristicaceae and Annonaceae, crown group diversification may also be largely a Tertiary phenomenon (J. A. Doyle et al. 2008 and references; Richardson et al. 2004).

It is perhaps hardly surprising that when we look at the diversification of other organisms now associated with angiosperms, we find a similar story. Most living ferns, especially to the polypod ferns which make up at least 80% of the living fern species, may have evolved subsequent to the initial diversification of the angiosperms in the Late Cretaceous. Diversification of epiphytic ferns in particular 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], in Hymenophyllaceae). One third of all leptosporangiate ferns are epiphytic (ca 3,000 species), making up about 10% of all epiphytes (Schuettpelz & Pryer 2009 - see Brodribb & Holbrook 2004 for the comparative leaf physiology of ferns and angiosperms), and similar radiations of epiphytes occurred in diverse groups including Bromeliaceae (Givnish et al. 2008a), Lycopodium (Wikström & Kenrick 1997; Wikström 2001 - about half - 190/380 species - of clubmosses are epiphytic [Schuettpelz & Pryer 2009]), Orchidaceae-Epidendroideae (Ramírez et al. 2007; Gustafsson et al. 2010; Conran et al. 2009). Ca 24,440 species of angiosperms are epiphytic (Schuettpelz & Pryer 2009), about half being members of Orchidaceae and Bromeliaceae. In liverworts the largely leaf-epiphytic Porellales diverged from the terrestrial Jungermanniales in the Jurassic, although 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), and Equisetum (Des Marais et al. 2003, but cf. Stanich et al. 2009) seems be a Tertiary phenomenon. Equisetum itself has been probably 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, and in general the timing of such relationships is perhaps that animals diversify rather later than plants (Winkler & Mitter 2008). Although initial diversification of ants seems to have started some 125 million years before present as angiosperms diversified, ants seem to have diversified more in the late Cretaceous-early Eocene 75-50 million years before present, again 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). Elaiosomes attractive to carnivorous ants - the fatty acids they contain may mimic those in the animal prey of these ants, the elaiosomes being "dead insect analogue[s]" (Carroll & Janzen 1973: 235; Hughes et al. 1994) - and found on small seeds or fruits occur in many plants (Beattie 1985). Their appearance in clades such as Polygalaceae-Polygaleae seems to be associated with the diversification of these clades and is a mid-Tertiary phenomen (Forest et al. 2007b; Lengyel et al. 2009, 2010, see also Fokuhl 2008); this is later than the beginning of diversification of other fleshy fruits and general increase in fruit size, which occurred some 85-75 million years ago in the Late Cretaceous-Early Tertiary (e.g. Eriksson et al. 2000a; Dunn et al. 2007). 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 of various kinds and myrmecochory are particularly common in the Australian flora, the ground flora of the east North American and European forests (Sernander 1906; Berg 1975; Orians & Milewski 2007; Lengyel et al. 2009, and references), 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 general, plant-ant associations like myrmecochory may have evolved in the late Eocene and afterwards, not earlier (Dunn et al. 2007), and the commonly encountered associations between ants, plants and sap-sucking homopterans (Ueda et al. 2008) are considerably younger.

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. (Along similar lines, it has been noted that diversification of C₄ grasses occurred some time after their initial evolution [see e.g. Christin et al. 2008]). Even if initial diversification of insects and angiosperms was associated, some subsequent bouts of 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 2006; Wheat et al. 2007; Kölsch & Pedersen 2008). 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). 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).

The large, nutritious seeds (or large dispersal units in general) and fleshy fruits that become particularly notable in early Tertiary angiosperms are likely to have been dispersed by animals, and by mammals, birds, and bats in particular. 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 occurred within Juglandaceae and Fagaceae by the early Tertiary. This change may be linked primarily to a change in forest type, now closed and made up of large trees 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 (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). Indeed, 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 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).

However, the causal links between plant disseminule size and vertebrate disperser are unclear (Eriksson et al. 2000a; Eriksson 2008; see also Mack 2000). 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 is not linked with notable increases in diversification (Bolmgren & Eriksson 2010 and literature). Indeed different ecosystem dynamics may have prevailed in the Cretaceous and Tertiary (Tiffney 2004). As another factor to consider, there also seems to have been a change in the general conductive efficiency of angiosperm woods around this period (Wheeler & Baas 1991) - could this have been necessitated by the increased transpiration made possible by the dense veinlet reticulum?

After the end-Cretaceous disappearance of the dinosaurs, diversification of seed-dispersing animals, including 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). 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-eating bats (the first group) because there were a greater diversity of fruit types (Jones et al. 2005; Teeling et al. 2005). 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 in mammals. 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).

Currently, animals allow for the wide dispersal of rather large propagules and 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, 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).

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).

There have also been major shifts in seed mass that are rather strongly correlated with changes in life form/plant habit (Eriksson et al. 2000a; Moles et al. 2006a, b). Thus 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.

Large seeds and fruits are dispersed predominantly by birds and mammals. 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]).

Molecular studies have suggested that there is correlation between the rate of molecular evolution and plant habit: molecular evolution is faster in herbs (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; Korall et al. 2010 [ferns]). In a series of extensive analyses of both monocots and eudicots, Smith and Donoghue (2008) showed that there are usually substantial increases in the rate of molecular evolution in herbs as compared to trees, shrubs, or simply 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. 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).

So along with a shift in ecology, there may have been a shift in metabolites involved in defence, and thus perhaps also in herbivores. 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 (Wing et al 2005). Furthermore, Novotny et al. (2006) suggested that individual species of temperate and tropical plants (controlled for phylogenetic relationships) supported a similar number of insect species, but 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).

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 varies: some plants - probably including most early angiosperms - offer only pollen, and of plants that produce nectar, monocots mostly have septal nectaries, core eudicots (and perhaps some clades immediately basal to them) receptacular nectaries of one sort or another, while nectaries of different types are sporadic in other angiosperms. 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). As mentioned below, a number of bees are quite specific pollinators, at least on any one trip, and can move pollen from plant to plant of the one species. Self pollination is hindered by sporophytic and gametophytic incompatibility, and less effectively by protandry or protygyny (the latter is commonest in members of the ANITA grade and the magnoliids). Thus the activities of pollinating animals can be linked with diversification, whether in generally affecting gene flow and hence speciation or being involved in barriers between closely related species (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).

When thinking of pollinator relationships, there is something of a paradox. Bees may be plesiomorphically oligolectic, even if generalist bees like bumble bees visit flowers with complex corollas which the animal has to learn to work before visits are effective, while specialists often visit shallow flowers with easily accessible rewards (Wcislo & Cane 1996). Within Apidae, bees such as many of the several hundred species of Meliponini (stingless bees) may each visit many species of plants, although on any one trip a bee is likely to be much more selective (Heard 1999), visiting only one or a few species, so being functionally mono- or oligolectic. 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 radiate flowers with relatively accessible nectar. Halictidae-Rophitinae (sister to all other halictids) prefer to visit flowers of asterids, especially members of the asterid 1 clade, although other plants are also visited (Patiny et al. 2008). 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; oil flowers may have evolved some 28 times, but the syndrome has been lost even more often (Renner & Schaefer 2010).

In the New World, the approximately 200 species of the notably long-tongued (the tongue is 15-42 mm long) euglossine bees (orchid bees - Euglossini) are 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, Bignoniaceae, Gesneriaceae-Gesnerioideae, Lecythidaceae-Lecythidoideae, perhaps 2,000 species of Orchidaceae-Epidendroideae (here the males largely pollinate, looking for fragrances: Cameron 2004 for 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. These few euglossine bees are the major pollinators of well over 4,000 species of plants in the Neotropics (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 euglossine bees (Roubik & Hanson 2004; Zimmermann et al. 2009). 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). Bumble bees seem to have diversified 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). They are prominent pollinators in alpine environments, large genera like Gentiana, Rhododendron and Pedicularis depending on them for pollination. Bumblebees first moved into South America some 8-6 million years ago - perhaps along with plant genera like Rubus, Scutellaria and Lupinus that have subsequently diversified substantially there (Hines 2008).

Although initial diversification of lepidoptera may date to before angiosperm evolution, butterfly radiation in particular beginning in the Late Cretaceous some 90 million years ago, much is a Tertiary phenomenon, occurring in the last 75 million years (see Vane-Wright 2004; Wheat et al. 2007; Wahlberg et al. 2009 for references); 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). 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, the bulk seems to have been Tertiary. Caterpillars of these groups tend to show rather high food-plant specificity. Bat pollination is sporadically distributed in angiosperms, and bat pollinated flowers probably began evolving in the Miocene (Fleming et al. 2009).

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. Bird pollination is likely to be a Tertiary phenomenon, with three different groups of birds (Trochilidae, Nectariniidae, and Meliphagidae) predominantly being involved (Cronk & Ojeda 2008).

The evolution of both plants other than angiosperms and animals other than insects was also affected by angiosperm evolution. 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. C₄ photosynthesis in grasses seems to have originated in the middle Miocene, some 12.5 million years before present (Jacobs et al. 1999), 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; Keeley & Rundell 2003) and butterfly Satyrinae larvae 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).

Overall, it is difficult to see any direct or even indirect effect of some of these associations, whether of fungi, insects, or other organisms, on the diversification of the plants involved. Our understanding of the ecological-evolutionary connections between insects and plants remains unclear; 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 it has been suggested that insects only rarely act as selective agents on their hosts (Strong et al. 1983), we know little about the details of long-term 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 consequences of the evolution of flowers (and fruits), but there have been a number of more physiological changes that have profoundly affected the physiology and 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 are implicated in the long-term decline in atmospheric CO₂ concentration that characterises the Tertiary, affected the rate of photosynthesis and nutrient acquisition, and the like. But although there are also these more purely physiological changes that may be linked to diversification, our understanding of the eco-physiological dimension of evolution is for the most part very poor - although this is changing. Indeed, it is against the backdrop caused by these changes that the evolution of angiosperm flowers and fruits has played out and angiosperm diversification occurred; interactions have been mutual. One way of thinking about these changes is that they provided angiosperms a suitable ever-wet environment in which diversification could occur; they provided the space for features associated with pollination and seed dispersal in particular to interact with other aspects of the biotic and abiotic environment (see also Boyce et al. 2010; also Marazzi & Sanderson 2010 and diversification of a clade of American Senna with extrafloral nectaries).

integrate this para.: About 30% of the organic carbon in the biosphere is currently locked up in lignin (Boerjan et al. 2003); when did lignin-decaying fungi evolve? 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 (). For the decomposition of lignocellulosic compounds, see Martínz et al. (2005). In addition, there are suggestions that lycophytes had proportionally a large amount of lignin (Robinson 1990), while 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). A connection needs to be established between the dark and disturbed/humid conditions hypothesis of angiosperm origins (Feild et al. 2004) and ideas where disturbance is still an important factor but conditions are well lit (e.g. Royer et al. 2010) and/or drier (Bond & Scott 2010).

The climate in the Late Jurassic-Early Cretaceous was dry - certainly Pangaea had a notably dry interior - but continents were drifting apart, there was high carbon dioxide concentration, rising sea levels, and 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).

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. 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); 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.

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). Vessels in magnoliids and ANITA-grade angiosperms may indeed be more effective in transmitting water than tracheids, but on a xylem cross-sectional area basis such vessel-bearing plants may not be any more hydraulically efficient than their tracheid-bearing relatives (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). It may also be noted that 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). Finally, 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. 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). There seems 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 perforation plates; there were also some changes in wood parenchyma, fewer in rays (Wheeler & Baas 1991) - could this have been necessitated by the increased transpiration made possible by the dense veinlet reticulum unique to angiosperms? Since early angiosperms were probably woody, climatic niche evolution that occurred may have been slow, i.e., habitat evolution was slow (Smith & Beaulieu 2009).

During the 200+ million year existence of land plants with a dominant sporophytic generation prior to the evolution of flowering plants, venation density was largely constant despite considerable fluctuations in atmospheric carbon dioxide concentrations (e.g. Boyce et al. 2010). The venation density of angiosperms has increased very considerably, although this is less evident in members of the ANITA grade (see above), so dramatically reducing the the main element in the resistance to water flow through the plant (Sack & Holbrook 2006). This increased vein density allowed an increase in stomatal conductance and 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). 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. (for vein architecture, see Roth-Nebelsick et al. 2001; for stomatal closure [ferns versus angiosperms], see Brodribb & Holbrook 2004; for leaf hydraulics, see McKown et al. 2010). Could there be a link between xylem evolution and increased transpiration?

Increased transpiration in turn quite possibly helped to drive the spread of widespread tropical forest with reliably high rainfall (Boyce et al. 2008, 2009, 2010). Indeed, 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). 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. Woody angiosperms diversified in such rainforests, and epiphytes, another appreciable component of angiosperm diversity, are abundant there (Feild et al. 2009a; Boyce et al. 2009, 2010; Boyce & Lee 2010). The spread of perhumid conditions in which large angiosperms dominate may have selected against drier, rather small stature angiosperm forest subject to frequent fires, and fires did decrease notably from the mid Palaeocene to mid Eocene (Bond & Scott 2010).

Another element to consider is litter breakdown. Factors like leaf mass per area (M_A, the inverse of specific leaf area (SLA), leaf area to dry mass, cm² g�1), and rates of photosynthesis, plant growth, litter decay, and nutrient cycling are linked, although there is much within-community variation in such features within angiosperms (e.g. Cornwell et al. 2008; Wieder et al. 2009). Low leaf mass/area and high amounts of nutrients in litter are both implcated. 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).

Decay is also affected by the composition of the structural elements of plants. It has been suggested that lignin content may be 20% lower in angiosperms that in gymnosperms (Robinson 1990), although this needs to be confirmed. Indeed, both lignin and polysaccharide content are negatively correlated with the rate of litter breakdown (Cornwell et al. 2008). 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). The litter of deciduous angiosperm trees decomposes faster than that of evergreens, and that of angiosperm forbs decomposes faster than that of any other group of land plants (Cornwell et al. 2008). Interesting, litter breakdown in forbs is faster than that of graminoids (Cornwell et al. 2008), so nutrients will be recycled fastest in forb-dominated communities. Graminoid lignin is somewhat different in composition from other lignins, having an appreciable amount of p-hydroxyphenyl units. UV light also decomposes lignin (Austin & Ballaré 2010).

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). Angiosperm leaves and litter have relatively high amounts of nitrogen and phosphorous (Cornwell et al. 2008), and so the relatively less bulky litter decomposes readily and speeds nutrient cycling - note that these more easily decomposable leaves are likely to be more palatable to insects, too (see below). Exactly those nutrients that the plants themselves need are released, and the disturbed habitats that angiosperms may initially have favoured are also likely to be associated with elevated levels of nutrients (Berendse & Scheffer 2009). Litter from forbs and deciduous trees decomposes more quickly than that of graminoids (mostly Poaceae and Cyperaceae) and evergreen trees (Cornwell et al. 2008). Mean annual precipitation is also positively correlated with litter turnover and the release of the nutrients it contains (Wieder et al. 2009). Hoorn et al. (2010) suggest 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 diversiify (see also Boyce et al. 2010). Finally, the high rates of photosynthesis of angiosperms allow a higher growth rate, and this might allow angiosperms to take immediate advantage of any increases in nutrient supply, effectively scavenging any available nutrients (Berendse & Scheffer 2009). 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 such rainforests in the Tertiary.

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). But even tropical, largely evergreen, non-riparian lowland rainforest, as on Barro Colorado Island, is be subject to insect damage (Royer et al. 2007). 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). In immediately post-bolide impact floras, some from ca 64 million years before present (early Palaeocene) in the western interior United States, plant diversity was low; the species were deciduous, with thin leaves and presumably low defences, 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, 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).

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).

Brundrett (2009) lists seed plants known to have ectomycorrhizal associations (see also Smith & Read 2008). Such associations are relatively uncommon; only ca 2% of vascular plants are ectomycorrhizal, but ericoid mycorrhizae account for another 1.3% and orchids some 8.1% more. However, ecytomycorrhizal associations have formed numerous times, perhaps some 35 origins or more, since there are over 40 families of angiosperms involved (Bruns & Schefferson 2004, Wang & Qiu 2006; Smith & Read 2008). The ability to form ectomycorrhizae in fungi has evolved at least 6 to 8 times in fungi, less frequently than in plants, and the basidiomycete clades involved are older than the plants with which they are associated (Hibbett & Matheny 2009). A wide variety of fungi form ectomycorrhizal associations - one estimate is ca 5,000 species (Molina et al. 1992), impressive single-site diversity of fungi has been documented (Horton & Bruns 2001 - most examples are from Pinaceae-dominated forests), and individual species of plants may form 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).

Nevertheless, ectomycorhizal associations may be of major importance for angiosperm evolution because of their effect on the soil, weathering, and hence on the earth's climate - and to a certain extent their activities seem to work counter to the increasing nutrient availability enabled by the decomposition of angiosperm litter just discussed. Ectomycorrhizal Pinales, Dipterocarpaceae and relatives, some Fabaceae, members of Fagales, etc., can all dominate the communities in which they grow. Ectomycorrhizae are commonly found in plants growing on rather extreme soils, either poor in nutrient and/or rich in organic materials, and especially in (cool) temperate and tropical montane habitats (e.g. Malloch et al. 1980). In such communities there is often a decrease in the pH of the soil; CO₂ sequestration may increase because the development of sometimes massive amounts of mor humus is common, especially in cooler climates, podzolization may occur (van Schöll et al. 2008), and the acid, nutrient-poor conditions in these communities are not conducive to the activity of potential decomposers of humus - a vicious cycle. Brener (2003) notes that rocks rich in organic matter (substantial amounts of this may be inertinite, charcoal from fires - Scott and Glasspool 2006) derived from plants are particularly prominent in the Mid Cretaceous 120-90 million years ago and the mid-Tertiary, 50-30 million years ago, although the incidence of fire seems to to have dropped at the end of the Palaecene until increasing again in the last 10 million years (Bond & Scott 2010).

There is also an increase in mineral weathering faciltated in part by the siderophore chelating agents produced by the ectomycorrhizae and their bacterial associates and by the respiration of the association that produces substantial amounts of CO₂ that is then used up in weathering (Taylor et al. 2008). The basic equation for weathering is CO₂[atmosphere, in soil from respiration] + CaSiO₃[continent] -> SiO₂[continent + ocean] + CaCO₃[ocean] (from 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., may form complexes with aluminium ions, detoxifying the aluminium but also increasing the weathering of aluminium-containing minerals (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 question, "When did ectomycorrhizal associations become common?", is important. Suggestions that such 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 (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 of plants producing this pollen, both pollen and macrofossils being linked uniquely to Fagales (but not Fagaceae or Nothofagaceae) - were 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 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). Normapolles plants were probably ectomycorrhizal, Nothofagus, also Fagales, is ectomycorrhizal, and if they were abundant in the Late Cretaceous, they may have had a transformatory 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). 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 linked to the decrease in atmospheric CO₂ concentration during the Tertiary (Pagani et al. 2009; Taylor et al. 2009). CO₂ concentration was at a high of over 3,000 ppm at the PETM, and it was perhaps ca 4,000 ppm in the Jurassic, but by the recent glacial periods it had dropped to 180-190 ppm, about as low as it has ever been during the whole period of land-plant evolution (Zachos et al. 2008; Gerhart & Ward 2010; Boyce et al. 2010).

In forests with endomycorrhizal associations fewer species of fungi are involved and the habit evolved perhaps once, the fewer than 200 species of Glomeromycota all being 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; Hart et al. 1989; McGuire 2007). How many times the 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.

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), 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), and they may often be involved in facilitating stress tolerance in plants (Rodriguez & Redman 2008). Van Bael et al. (2009) recently found that leaf-cutting ants seem to dislike plants that have numerous endophytes. However, the line between mutualism - or at least prolonged symbiosis - and parasitism is a fine one (Eaton et al. 2010 and references).

There is also substantial variation in phloem pressure in angiosperms, sugar-rich phloem exudate and phloem loading perhaps being most common in the asterid I + II clade, and a variety of selective advantages for this loading can be suggested (Turgeon 2010b). 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).

The different ways in which angiosperms fix carbon are quite well understood (Keeley & Rundell 2003). C₄ photosynthesis is especiallly common in Poaceae; ca 50% of the 5,250+ species of the PACCMAD clade of Poaceae alone have C₄ photosynthesis. However, the driver that promotes/promoted the evolution of this distinctive photosynthetic pathway is unclear (see Westhoff & Gowik 2010 for general literature). C₄ photosynthesis in grasses seems to have originated in the Oligocene about 25 million years before present, perhaps in response to declining CO₂ concentration - itself perhaps the result of the carbon-sequestering activities of ectomycorrhizal angiosperms (Taylor et al. 2009) - in the atmosphere, or of high temperatures; more recent work emphasizes the importance of drought and water stress (Edwards & Still 2008; Edwards 2009). The great expansion of C₄ grassland beginning in the Miocene may be due to other perhaps associated environmental changes like accelerated fire cycles, etc. (Sage & Kubien 2003; Tipple & Pagani 2007; Christin et al. 2008; Vicentini et al. 2008).

In addition to the evolution of C₄ photosynthesis several times in Poaceae, it has evolved in several other clades, too (Sage et al. 1999), thus it occurs in many species of Cyperaceae, some Amaranthaceae (ca 800 species are involved) and other core Caryophyllales, and Euphorbiaceae (Euphorbia subg. Chamaesyce). Indeed, 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). 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, and recent, having evolved within the last 25 million years or less. It is also an interesting case of parallelism at the molecular level, a conserved serine residue in the carboxy-terminal part of the enzyme being integral to its functioning and this amino acid is found in Poaceae, Asteraceae, etc. (e.g. Bläsing et al. 2000; Christin & Besnard 2009). Grasses in particular dominate large areas of pairie and savannah habitats, and interesting they, like ectomycorrhizae, secrete siderophores ().

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).

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). Sugar-rich phloem exudate and phloem loading is perhaps most common in the asterid I + II clade, and a variety of selective advantages for this loading can be suggested (Turgeon 2010b).

Not yet integrated: 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; polysymmetrical 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. It 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; seee 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. (Exactly where monosymmetry is an apomorphy in Asterales is unclear.)

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 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 10,000+ of the whole family, and the three "basal" clades are forest plants (see also above). Furthermore, one needs to know when diversification occurred in the clade, and when any supposed key innovation appeared. 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.).

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 1997b; 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. Although monosymmetry is clearly very widespread in angiosperms 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 is dominant are also not notably speciose, even if they locally dominate ecologically, as with Fagales; 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) provide a useful survey of the evolution of wind pollination in angiosperms. 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.

For Poaceae enough is known to show how complex understanding diversification can be. For example, Poaceae-Pooideae 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). 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 reasily 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. Indeed, mutualistic associations between microbes and insects are likely to be important in understanding plant-insect interactions (Janson et al. 2008), as well as associations between endophytic fungi and plants.

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, Brassiccaeae and Solanaceae with genome duplications there, although as van de Peer et al. (2009a) point out, the two may not be particularly closely linked in time. However, detailed connections between gene and/or genome duplication and diversification of angiosperms are unclear.

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 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). 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), 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 in general (Eriksson & Bremer 1992; Kay et al. 2006b), 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 are hard to identify here.

However, the study of key innovations is far more than simply linking a feature to a node (e.g. Donoghue 2005; Marazzi & Sanderson 2010). Thus key innovations have 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 arely 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). Determining that an innovation might be a key innovation is a dificult process (e.g. Sanderson 1998; Ree 2005b; Maddison et al. 2007). Even if a feature such as evolution of extra-floral nectaries may be a key innovation in one instance, this does not mean that is always is (Marazzi & Sanderson 2010). In many clades, including angiosperms as a whole, characterized by "key innovations", patterns of clade numbers do not suggest an immediate diversification. Thus in Poaceae the first three pectinations have a mere 4, 14 and 11 species respectively, compared with the over 10,000 of the PACCMAD clade, and similar patterns are found in Ericaceae, Orchidaceae, and even Asteraceae. 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).

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 not so evident there (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). Guatteria (Annonaceae), Inga (Fabaceae-Mimosoideae) and Ocotoea (Lauraceae) between them contain over 1,000 species; all are Neotropical and the majority have evolved fairly recently, and in the first genus much speciation may have occurred only subsequent to its entry into South America (Erkens et al. 2007).

It is clear that 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 capaity 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 angiospemrs, and beyond...

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 a group of key 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) emphasized flexibility in construction of the flowers (there is abundant evidence that the development of flowers of ANITA grade angiosperms and magnoliids 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 - usually 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). Furthermore, the reason for the 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. 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 occurs 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). Consistent with this, the realisation that the distribution patterns of clades that were thought to reflect vicariance caused by plate tectonic events are quite often better explained by more recent dispersal/migration events (e.g. Renner 2005b and de Queiroz 2005 for summaries; Yoder & Nowak 2996; Wen & Ickert-Bond 2009; Carpenter et al. 2010; cf. in part Ladiges & Cantrill 2007; Heads 2008, etc.) also emphasizes that biotas may be much younger than we had thought.

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; certainly, duplication occurs in gymnosperms, too, even if the chromosome numbers there are low, and also in other land plants. Overall angiosperm success seems to be in considerable part the result of diversification of individual angiosperm clades with such fortuitous combinations, and establishing an immediate connection between acquisition of an apomorphy or group of apomorphies and diversification is difficult. However, plants with monosymmetric flowers and the herbaceous habit are notably common. 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); adaptation to such niches by members of the local flora may not occur very frequently. This may be particularly important for some eco-physiological traits.

[Out of place.] 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. Thus it has been suggested that tropical taxa show more pronounced differentiation with respect to elevation/climate than do temperate taxa (Kozak & Wiens 2007: sister taxon comparisons, animals). The rate of rbcL gene evolution within families and the number of species those families contain is positively correlated (Barraclough et al. 1996).

For further discussions on evolution and diversification, see e.g. Mulcahy (1979: pollen tube competition and angiospermy), 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).

AMBORELLALES Melikian, A. V. Bobrov & Zaytzeva [Back to Index]

Nodes 1:1; plant dioecious; hypanthium +, stamens sessile, vascular bundle branched towards apex, microsporogenesis successive, walls developing by centripetal furrowing, pollen ektexine granulate, 1 ovule/carpel, embryo sac 9-nucleate [three synergids], bipolar; fruit of drupelets, mesocarp thickened and lignified; endosperm triploid. - 1 family, 1 genus, 1 species.

AMBORELLACEAE Pichon - 1/1: Amborella trichopoda. New Caledonia.

Amborellaceae are sprawling shrubs with apparently two-ranked, exstipulate leaves; the margins are both 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.

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 (Williams 2009).

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; cf. Goremykin et al. 2009a, the existence of these transfers is questionable). Although mitochondrial genomes like that of Amborella are as yet unknown from other angiosperms, sampling is as yet very poor.

There seems to be no agreement on the pollen morphology of Amborella, and ovule "type" has also been variously interpreted. The secondary chemistry of Amborella is almost unknown.

For information, see Metcalfe (1987: anatomy), Philipson (1993), Sampson (1993: pollen), Carlquist and Schneider (2001: wood anatomy), Yamada et al. (2001a: ovules), Posluszny and Tomlinson (2003: floral morphology), Feild et al. (2003: ecophysiology), Tobe et al. (2000: embryology), Bobrov et al. (2005: fruit), Friedman (2006: embryo sac).

NYMPHAEALES [AUSTROBAILEYALES [[CHLORANTHALES + MAGNOLIIDS] [MONOCOTS [CERATOPHYLLALES + EUDICOTS]]]]: vessels +, elements with scalariform perforation plates.

Presence of vesels is optimised here on the tree. Feild (2005) suggests that vessels may have evolved in plants growing in humid conditions. This is also true of Gnetales and fossil groups with vessels, too, although they are unlikely to have evolved in aquatics. For discussion as to where characters of pollen morphology and development are to be placed on the tree, see Taylor and Osborn (2006); it partly depends on how the characters are defined. For the possibility of a genome duplication at about this position, see Cui et al. (2006).

NYMPHAEALES Dumortier [Back to Index]

Aquatic herbs; mycorrhizae 0; primary root soon aborts; primary stem with ± scattered bundles; vascular cambium 0; aerenchyma common; ?nodes; stomata anomocytic; 4-celled uniseriate secretory trichomes with a large terminal cell, mucilage ?always produced; leaf base broad; seeds exotestal, operculate; endosperm scanty, perisperm copious [starchy], suspensor 0, embryo broad. - 3 families, 6 genera, 74 species.

The curious fossil Archaefructus, probably an aquatic plant and 124 million years old, has been linked with Hydatellaceae (Doyle & Endress 2007: Doyle 2008b). Cretaceous fossils assignable to Nymphaeaceae are quite common, and it has been suggested that Nymphaeales were "the first globally diverse clade" (Borsch " Soltis 2008: p. 1051).

Hydatellaceae have only recently been included here. They are superficially like Centrolepidaceae, and have long been considered to be monocots. However, it was realized that their combination of characters was unique within Poales, which also includes Centrolepidaceae, indeed, the combination was very distinctive within monocots as a whole (e.g. Hamann et al. 1979; Dahlgren et al. 1985). Trithuria and Xyris had been sister taxa (weak support) and in turn sister to Mayaca (still weaker support) of the Poales in some studies (e.g. Michelangeli et al. 2003); Janssen and Bremer (2004) suggested that the association of Hydatellaceae with Mayacaceae was probably an artefact (see also Chase et al. 2006). Recent studies (Saarela et al. 2007, several genes from two compartments, morphology) place Hydatellaceae firmly with Nymphaeales, and sister to [Cabombaceae + Nymphaeaceae]. Many of the morphological features of Hydatellaceae that made it so different from other monocots are consistent with this new position. Hamann (1998) even noted that the antipodal cells were absent or degenerated early, and absence of these cells would almost be expected if Hydatellaceae were placed here; indeed, it was subsequently confirmed that Hydatellaceae has a 4-celled embryo sac like other Nymphaeales and Austrobaileyales (Friedman 2008; Rudall et al. 2008).

Nelumbonaceae, also "water lily"-type plants, have been associated with Nymphaeaceae in the past, but they are here placed in Proteales.

HYDATELLACEAE U. Hamann - Plant rosette-forming; sieve tube plastids with proteinaceous inclusions; leaves linear, with a single vein; plant monoecious; inflorescence scapose, capitate, with involucral bracts; flowers imperfect, ebracteate; P 0; staminate flowers: A 1, filament long, slender; carpellate flowers: G with a single pendulous ovule, stigma penicillate. - 1/10. India, New Zealand and Australia.

Hydatellaceae are rather small more or less caespitose and often annual aquatic (submerged or not) herbs with linear leaves that have only a single vein, and capitate and usually scapose inflorescences bearing minute and very reduced flowers; the beaded, penicillate stigmas are distinctive.

Hairs with possible apical secretory cells are known only from the inflorescences. The inflorescence is described above as being capitate, with highly reduced flowers, i.e., it is a sort of pseudanthium, although alternative interpretations are possible (Rudall et al. 2007a). Early work suggested that the carpels might be initiated outside the stamens, and this has been confirmed (Rudall et al. 2007a); staminate flowers are the first to be initiated in the cymose inflorescence (see also Begoniaceae).

There is some disagreement over the interpretation of the morphology of the embryo. Tillich et al. (2007) compare it with that of a monocot, describing collar rhizoids, a coleoptile, two cotyledonary sheath lobes, and a haustorium. Sokoloff et al. (2008a) suggest that the sheathing structure with its bilobed apex that is found in some species can be interpreted as more or less completely connate cotyledons. The seed attaches to the sheathing structure, although details of exactly how are unclear. In some Hydatellaceae there is no sheathing structure at all and only some kind of lateral outgrowth from the seedling that goes into the seed. Both Tillich et al. (2007) and Sokoloff et al. (2008a) examined largely surface morphology, neither looked in any detail at anatomy.

For other information, see Cutler (1969: vegetative anatomy), Hamann (1975, 1998 - embryology and general respectively), and Hamann et al. (1979: seed anatomy).

Cabombaceae + Nymphaeaceae: hydrolyzable [ellagi]tannins +; leaves involute, peltate, 2ndary veins palmate; flowers single along stem, pedicel long; cortical vascular system in flower; P whorled (outer [inner] whorls in 3's), outer members enclosing the rest of the bud; A whorled, placentation ± laminar, carpel margins with postgenital fusion; exotesta palisade, anticlinal walls sinuous.

Fossils of Nymphaeaceae, Cabombaceae and/or stem-group Nymphaeales are known from the Lower Cretaceous from several parts of the world (e.g. D. W. Taylor et al. 2001, 2008; Mohr et al. 2008). Although other fossils possibly of this group - to a certain extent they show characters of the both families - are known from the Barremian-Aptian 125-113 million years before present in Portugal (Friis et al. 2001; see also von Balthazar et al. 2008 for another fossil perhaps assignable to this general [Nymphaeales-Austrobaileyales] area), they may also be from a member of Austrobaileyales (Gandolfo et al. 2004). There have been considerable arguments over the timing of diversification within Nympahaeales and Nymphaeaceae (e.g. Nixon 2008). Löhne et al. (2008) suggests that divergence is only Palaeocene in age, (75-)56.4(-38) million years ago, although Wikström et al. (2001) had found that divergence of the two appeared to occur 144-111 million years before present (in the latter study details of relationships within the clade differ from those given here, and of course Hydatellaceae were not included). Yoo et al. (2005) give a still younger date, pegging divergence within the crown group to only 44.6 ± 7.9 million years before present; the fossil Microvictoria was perhaps stem group Nymphaeales (cf. Gandolfo et al. 2004). D. W. Taylor et al. (2008) discuss the vegetative evolution of the group, noting how inclusion of diverse fossils affects relationships as suggested by analysis of morphological variation and hence evolutionary interpretations.

Taylor (2008) outlined the vegetative morphology of this clade. Warner et al. (2008) discuss perianth evolution; they provide a useful summary on the literature on perianth morphology.

For additional information, see Gwynne-Vaughan (1897: anatomy), Collinson (1980: seed anatomy), Osborn et al. (1991: pollen), Yamada et al. (2001b: ovules), Chen et al. (2004: seed anatomy), Les et al. (1999: general) and Schneider et al. (2003: general).

CABOMBACEAE A. Richard - Floating and rhizomatous; stem vascular tissue with two pairs of bundles; flowers 3-merous, perianth members with one trace, stamens with slender filaments; endosperm helobial [first division transverse, free-nuclear divisions in micropylar cell, chalazal cell ± enlarged]. - 2/6. World-wide.

Cabombaceae are rather small-flowered waterlilies with few ovules or seeds in each carpel; they have floating stems and all the flower parts are free.

Brasenia is wind pollinated, while Cabomba has paired nectaries on its inner tepals and is pollinated by insects; Taylor and Williams (2009) describe details of reproduction from pollination to fertlization in considerable detail..

The stems of Brasenia are encased in a very thick layer of mucilage; there are paired, glandular patches at the nodes. Peltate leaves in Cabombaceae are spirally arranged and floating, although in some taxa they are uncommon; the more or less dichotomously-divided submerged leaves are opposite.

For information, see Richardson (1969: development of Brasenia flowers), Schneider and Jeter (1982: pollination of Cabomba), Moseley et al. (1984: general, Cabomba), Williamson and Schneider (1993: general), Taylor et al. (2001: fossils), M. L. Taylor & Osborn (2006: pollen) and M. L. Taylor et al. (2008: pollen).

NYMPHAEACEAE Salisbury - Nodes 3:3; (flowers/branches replacing leaves in the genetic spiral); individual perianth members with separate sepal-like and petal-like areas. - 3/58. World-wide.

1. Nupharoideae Ito - Rhizome stout; perianth 5-14, inner ?perianth members tiny, with abaxial nectaries, pollen spiny. - 1/ca 10. North Temperate.

2. Nymphaeoideae - Sepals 4-5, staminodes showy, pollen with a ring-like sulcus [zonasulcate], ovary more or less inferior, stigmatic surfaces of carpels continuous; fruit maturation underwater; seeds arillate. - 2/48: Nymphaea (46). World-wide.

Nymphaeaceae are the quintessential waterlilies, with large flowers, usually many free perianth parts and stamens, and many ovules in each carpel, the placentation usually being clearly laminar. There is often no obvious stem, Nymphaeaceae being aquatic rosette plants.

The family was previously much more diverse, and its first known fossil is from the Turonian (ca 90 million years before present: New Jersey, U.S.A.). Victoria (probably to be included in Nymphaea, see below), has "paracarpels", stigmatic areas adnate to the receptacular cup immediately surrounding the gynoecium, and these are also found in the Turonian Microvictoria; indeed, this latter is like flowers of Victoria in almost all respects, except being less than 1/10th the size (Gandolfo et al. 2004, see also above). Although the family is widespread, individual clades within it are relatively localized, and diversification probably occurred in the northern hemisphere in the early Tertiary (Löhne et al. 2008).

There is thermogenesis in the flowers of some Nymphaeaceae (Seymour 2001); beetles, flies and bees may be pollinators (e.g. Padgett 2007). In some taxa the stigma produces a copious rather sweet secretion.

The vasculature of the stem is exceedingly complex, especially at the node, however, the basic stem structure is unlike that of monocotyledons (Weidlich 1980 and references). Stipules may be single, adaxial and with two keels, or paired and lateral. In both Nuphar and Nymphaea flowers and even branches may occur where leaves would be expected to be found (e.g. Groß et al. 2006). Weberling (1989) suggested that in at least some Nymphaeaceae the individual carpels were free laterally, although adnate to the central axis inside and to "hypanthial" tissue outside.

For a phylogeny of Nymphaea, see Borsch et al. (2007); it definitely includes the wind-pollinated and usually apetalous Ondinea, but its monophyly lacked much support. However, in a study with more complete sampling it seems very likely that the spiny Victoria and Euryale should also be included in Nymphaea (the spiny genera are sister taxa - see Löhne et al. 2007; Borsch et al. 2008).

For information, see Schneider and Williamson (1993: general), and Takhtajan (1988: ovules and seeds).

AUSTROBAILEYALES [[CHLORANTHALES + MAGNOLIIDS] [MONOCOTS [CERATOPHYLLALES + EUDICOTS]]]: ethereal oils in scattered spherical cells [lamina and perianth ± pellucid-punctate].

AUSTROBAILEYALES Reveal [Back to Index]

Nodes 1:2; perianth parts not obviously in 3's, mucilaginous extragynoecial compitum +, outer integument 5-7 cells thick; fruit a berrylet; mesotestal cells ± sclerotic. - 3 families, 5 genera, 100 species.

Pollen tubes grow through mucilage on the style, etc., and reach carpels other than that on which they initially landed, the distinctive "mucilaginous extragynoecial compitum" of the characterisation above.

For the circumscription of the order, see Soltis et al. (1997), Källersjö et al. (1999), and Qiu et al. (1999).

For vegetative anatomy, see Metcalfe (1987), for developmental morphology of ovules and seeds, see Yamada et al. (2003a), and for embryo sac morphology, see Tobe et al. (2007).

AUSTROBAILEYACEAE Croizat - Flowers with internal staminodes. - 1/2. Eastern Australia.

Austrobaileyaceae are lianes with opposite and entire leaves and large, pendulous flowers. These have spreading petals, many stamens with broad filaments, and internal staminodes. The fruit consists of more or less ellipsoid berrylets.

For information, see Endress (1980: floral morphology, 1993: general), and Carlquist (2001: wood anatomy).

SCHISANDRACEAE Blume, nom. cons. - Tetracyclic triterpenes [cycloartanes] +; astrosclereids +; mucilage cells +; leaf epidermis silicified; pollen tricolpate, colpi joining at distal pole of grain; exotesta palisade. - 3/92: Illicium (42). Sri Lanka and South East Asia to W. Malesia, S.E. U.S.A., E. Mexico, Greater Antilles.

Schisandra glabra (Schisandra s. str.) and Kadsura longipedunculata may have thermogenic flowers (Seymour 2001; Liu et al. 2006; Yuan et al. 2008); pollen may be a floral reward, and/or deceit may be involved. For the growth of the pollem tube over the surface of the epidermis, rather than between cells, see Lyew et al. (2007).

Kadsura is paraphyletic, based on the analysis of both trnL-F and ITS sequences, although this is not confirmed by morphological studies (Hao et al. 2001; Denk & Oh 2006; Liu et al. 2006). Liu et al. (2006) discuss character evolution in this group. Molecular and morphological studies also suggest rather different relationships within Illicium (Hao et al. 2000; Oh et al. 2003).

In the past, Illiciaceae and Schisandraceae have often been recognised as separate families.

For general information, see from Keng (1993), and Saunders (1997, 2000), for the female gametophyte, Friedman et al. (2003a, esp. b) and Friedman and Williams (2004), and for wood anatomy, see Yang and Lin (2007).

TRIMENIACEAE Gibbs - Wood rays 6-9-seriate; flowers small, carpels 1-2, 1 ovule/carpel; testa vascularized, almost all walls thick, lignified; perisperm +. - 1-2/6. New Guinea and S.E. Australia to Fiji.

Trimeniaceae are trees or lianes with opposite, serrate to entire leaves that often have pellucid glands and widely spreading and quite close secondary veins. The wind-pollinated flowers are small, there is usually only a single carpel with a single ovule, and the fruit is a berrylet.

Distinctive trimeniaceous seeds, albeit without a vascularized testa, have been found in Late Albian deposits some 118 million years old in Hokkaido, Japan (Yamada et al. 2008).

Some plants of Trimenia papuana have inaperturate pollen, while others have polyporate pollen (Sampson 2007).

For more information, see Endress and Sampson (1983: floral morphology) and Philipson (1993: general).

[[CHLORANTHALES + MAGNOLIIDS] [MONOCOTS [CERATOPHYLLALES + EUDICOTS]]] / MESANGIOSPERMAE: benzylisoquinoline alkaloids +; outer epidermal walls of root elongation zone with cellulose fibrils oriented transverse to root axis; P more or less whorled, 3-merous [possible position]; A whorled, carpels plicate; embryo sac bipolar, 8 nucleate; endosperm triploid.

Sesquiterpenes may be acquired here and lost at least three times, or acquired in the [Chloranthales + magnoliid] clade (and lost at least once there) and in Acorales. For the distribution of benzylisoquinoline alkaloids, alternatively, 1-benzyltetrahydroisoquinoline alkaloids, 1-btiq alkaloids, see Waterman (1999, 2007); the betalains of core Caryophyllales are biosynthetically similar to these alkaloids. This is perhaps the best place to put triploid endosperm on the tree; the other would be as a synapomorphy for all angiosperms, but in that case it would subsequently be lost twice or lost once and then regained.

Relationships between the lineages immediately above the basal pectinations in the main tree, the ANITA grade (Amborellales, Nymphaeales and Austrobaileyales here), are finally being cleared up. The positions of Ceratophyllales and Chloranthaceae have been particularly labile, the former having been placed e.g. as sister to the eudicot lineage, or sister to all flowering plants. Indeed, Graham et al. (2005) found that the inclusion of these two isolated groups in analyses could destabilise relationships among early-branching angiosperm clades, e.g., the position of the monocots became labile. Qiu et al. (2005; see also Löhne & Borsch 2005) found initial rather strong bootstrap support for an association between monocots and Ceratophyllaceae in a 9-gene analysis was vitiated by the failure to obtain much support for this relationship in any of the subanalyses and by details of the topology obtained in the 9-gene analysis itself that were rather improbable. Soltis et al. (2005b) very reasonably summarized their discussion on relationhips in this area by showing a pentachotomy.

Graham et al. (2005) found a rather weakly supported (73% bootstrap) [monocot + eudicot] grouping, but this was still weaker when Chloranthaceae and Ceratophyllaceae were included. Nevertheless, evolution of some floral developmental genes, e.g. in the C and D lineages, are also consistent with such a relationship (Kramer et al. 2004), as are the positional relationships between members of the androecium and the perianth (the stamens are individually opposite perianth members, although note that this relationship may be found in some magnoliids, e.g. Lauraceae), and perhaps some other characters (trimery of some floral whorls is fairly widespread within Ranunculales). Recently, Jansen et al. (2006b: 37 whole chloroplast genomes), and Duvall et al. (2006: overall, a relationships between magnoliids and monocots was preferred), Mathews (2006a: three PHY genes, 105 taxa), and Hansen et al. (2007: 61 protein-coding chloroplast genes) have found some support for this [monocot + eudicot] grouping. Furthermore, Jansen et al. (2006b) and Hansen et al. (2007) found support for a sister-group relationship between Chloranthales and the magnoliids. When sequences of complete chloroplast genomes are analysed, an association between monocots and eudicots is more strongly suggested. Thus Jansen et al. (2007) found strong support for this grouping, a number of alternative topologies being excluded, and although support for this grouping in Moore et al. (2007) was somewhat less strong, Ceratophyllum, with quite a long branch, was usually sister to eudicots. Similarly, Chloranthus was found to be sister to the magnoliids with moderate to strong support (Jansen et al. 2007), a position also found by Saarela et al. (2007) and many analyses in Moore et al. (2007). However, Goremykin et al. (2009b) recently found a [Ceratophyllum + magnoliid] clade (Chloranthaceae were not included) after removing 2,500 highly variable positions from the analysis; with the removal of 1,000 positions Ceratophyllum was sister to a [monocot + eudicot] clade.

Pending further studies, the set of relationships [[Chloranthales + magnoliids] [monocot [Ceratophyllales + eudicot]]] is recognised here, rather different from the relationships suggested by the tree in A.P.G. II (2003) and Soltis et al. (2005b).

CHLORANTHALES + MAGNOLIIDS: sesquiterpenes +; seed endotestal.

CHLORANTHALES J. F. Leroy [Back to Index]

Rays 6-10 cells wide; nodes often swollen, ?type; leaves opposite, toothed; stipules +; flowers very small, monosymmetric, parts whorled; P 3, basally connate, or 0; A 1; G 1, ± inferior, ascidiate, 1 apical pendulous straight ovule/carpel; fruit fleshy; endotesta palisade, lignified, crystalliferous. - 1 family, 4 genera, 75 species.

CHLORANTHACEAE Sims - 4/75. Tropics and subtropics, not Africa (Madagascar - Chloranthus only).

Chloranthaceae are aromatic plants that may be recognised by their opposite, serrate leaves with more or less sheathing stipules and often strongly swollen nodes that may collapse on drying. The flowers are very small and are often borne on branched inflorescences. The wood is soft.

Chloranthaceous fossils are common, diverse, and world-wide in distribution in the early fossil record. Fossil pollen grains, Asteropollis, are first known from the Barremian-Aptian of the early Cretaceous some 125 million years before present; these grains are assignable to Hedyosmum (Friis et al. 2005 for references; see Doyle & Endress 2007 for other palynomorphs assigned to the family). Estimates of the time of divergence of Chloranthaceae are 168-131(-126) million years before present (Soltis et al. 2008), however, estimates of the age of crown group diversification are within the last 60-29 million years (Zhang & Renner 2003b).

Endress (2001) has emphasized what he considered to be the plesiomorphic floral morphology of Chloranthaceae, although it is not a member of the basal ANITA grade; a number of aspects of its floral development, including the loss of the perianth, are clearly derived (e.g. G. S. Li et al. 2005). Eklund et al. (1997) and Eklund (1999) discuss the nature of the androecium in the family. In Chloranthus it has been suggested that the androecium is lobed, with 2 or 4 dithecal (i.e. four-sporangiate) stamens, and that staminate flowers of Hedyosmum have hundreds of anthers. However, staminate flowers in the family seem always to have but a single dithecal stamen, although this can be much modified, being flattened, lobed, and with the sporangia on the margins (see Kong et al. 2002, and references); the flowers with hundreds of stamens are closer to pseudanthia, that is, they are highly reduced and modified inflorescences that look like single flowers.

For more information, see Todzia (1988, 1993: general), Eklund (1999: general), Metcalfe (1987: vegetative anatomy), Carlquist (1992a: wood anatomy), Zhang and Renner (2003b: phylogeny), Doyle et al. (2003: morphology), and Eklund et al. (2004: evolution, inc. fossils).

MAGNOLIIDS [[MAGNOLIALES + LAURALES] [CANELLALES + PIPERALES]]: lamina margins entire; A many, spiral [possible position here], extrorse, raphal bundle branches at the chalaza.

4 orders, 20 families, 9900 species.

Caterpillars of Papilionidae-Papilioninae, swallowtail butterflies, are notably common (almost 33% of the records as of 2004) on members of this group (and Rutaceae! - similar alkaloids), although they are apparently so far unrecorded on Myristicaceae, in Laurales they predominate on Lauraceae, and in Piperales on Aristolochiaceae (see Scriber et al. 1995 for references; Zakharov et al. 2004).

Although the sister group relationship of Piperales with Canellales in particular is at first sight unexpected, the magnoliid clade as a whole and the relationships within the group are turning out to be robust. There was not - and still really is not - much if any morphological support for this grouping (Doyle & Endress 2000; see also the characterisation above). However, molecular support for the clade has been increasing in successive studies (e.g. Qiu et al. 2006b; Jansen et al. 2006b; Zhengqiu et al. 2006; Cai et al. 2006; Müller et al. 2006, support for [Canellales + Piperales] rather poor), and it includes the possession of unique indels (Löhne & Borsch 2005).

MAGNOLIALES + LAURALES: pollen with lamellate endexine.

It wouldn't take much for flowers with inner staminodes to be an apomorphy at this level.

MAGNOLIALES Bromhead [Back to Index]

Secondary phloem with bands of fibres [stratified]; pith with sclerenchymatous diaphragms [septate]; nodes 3:3; leaves two-ranked; outer integument 5-10 cells across; seeds medium-sized, testa vascularized, multiplicative; endosperm ?type, irregularly ruminate by inpushings of the seed coat. - 5 families, 154 genera, 2929 species.

Intra- and interfamilial variation of morphological characters that have often been used to reconstruct phylogenies, including vessel perforation type (simple versus scalariform), is considerable. There is much discussion on character evolution in the order in Sauquet et al. (2003), and much of the character hierarchy here is based on this. Note, however, that where characters like extrorse/introrse anther dehiscence, fruit dehiscence, and ruminate/non-ruminate testa are placed on the tree depends on how the characters are optimised, or even defined (the ruminations in the seeds of Myristacaceae and, say, Himantandraceae, look very different). Furthermore, there is some conflict with the positions of characters as they are optimised on a more extensive tree for "basal angiosperms", although less detailed for Magnoliales (cf. Ronse De Craene et al. 2003, also Judd et al. 2003). Doyle and Endress (2000) and Soltis et al. (2005) suggest additional characters for the clade.

Molecular data suggested that Myristicaceae are sister to the rest of the order, but support was only moderate (D. Soltis et al. 2000); the addition of morphological data strengthened that position, and also placed Magnoliaceae as sister to the remaining taxa (Doyle & Endress 2000; see also Sauquet et al. 2001). The family pairs [Annonaceae + Eupomatiaceae] and [Degeneriaceae + Himantandraceae] are both well supported (Sauquet et al. 2003; Müller et al. 2006 and references).

For more information, see Sugiyama (1976a, b, 1979: nodal anatomy), Endress (1977b, 1986a, 1994a: floral morphology), Metcalfe (1987: general anatomy), Taylor and Hickey (1995: general), Ronse Decraene and Smets (1996a: androecium), and Kimoto and Tobe (2001: embryology).

MYRISTICACEAE R. Brown - Tannin-containing tubes in the xylem; hairs branched or stellate (T-shaped); flowers imperfect; P 3, connate; staminate flowers: A whorled, connate, unithecate; carpellate flower: G 1, one ovule/carpel; fruit a follicle also dehiscing abaxially; seed large, arillate, endotesta palisade, lignified, crystalliferous, tegmen vascularized, massive, exotegmen sclerotic or tracheidal; endosperm nuclear; hypocotyl not developed in embryo. - 20/475: Myristica (175), Horsfieldia (100), Knema (95), Virola (60). Pantropical.

Myristicaceae are easily recognised: the plant has red sap, the terminal bud is long, the leaves are usually two-ranked, the petioles are often short and the blades are often glaucous below. The blades are often much damaged by insects, they are brittle when dry, and the venation is inconspicuous. Growth is fairly stereotyped: the main axis is orthotropic and bears pseudoverticillate, radially-departing branches at the end of each flush (this is called Massart's model). The plants are dioecious and the rather small flowers have a connate, 3-merous perianth, connate stamens, and a single carpel. The single-seeded, dehiscent fruits are often quite large and the seed is ruminate and usually arillate.

Diversification within the family may be recent, within 21-15 million years before present, although this seems very recent indeed given the age of the family, it distribution throughout the humid tropics, and its apparently low dispersability (J. A. Doyle et al. 2004), unfortunately, the recent discovery of fossil seeds apparently of Myristicaceae from the Eocene (London Clay) does not solve the problem because they cannot be placed accurately on the phylogenetic tree of Myristicaceae (J. A. Doyle et al. 2008). Note that although there is a long branch leading to the family, there is little molecular divergence between its extant members (Sauquet et al. 2003); this might suggest recent diversification. In this respect Myristicaceae can be compared with Annonaceae, also pantropical, but rather younger; there diversification may have occurred 84-57 million years before present (Doyle et al. 2004; Scharaschkin & Doyle 2005).

Relationships within the family are unclear. The African and Madagascan taxa may form a clade, possibly sister to Compsoneura (but perhaps long branch attraction), overall, geography and relationships may be summarized as [[Asia] + [America] [Madagascar and Africa]]. Within the Asian/Malesian representatives, Knema and Myristica may be sister taxa. Sauquet et al. (2001, 2003), Sauquet (2003) and Sauquet and Le Thomas (2003) recently suggested that the free stamens (in some species they are numerous and apparently spirally inserted) and small aril of Mauloutchia, apparently plesiomorphic features, are more likely to be derived.

For more information, see Kerster and Baas (1981: anatomy), Kühn and Kubitzki (1993: general), Jiménez-Rojas et al. (2002: growth patterns), Sauquet (2003: androecium), and Sauquet and Le Thomas (2003: pollen).

Magnoliaceae [[Himantandraceae + Degeneriaceae] [Eupomatiaceae + Annonaceae]]: wood with broad rays; flowers solitary, large [possible position], receptacle well-developed, cortical vascular system +; P = K + C; A many, spiral [another position!], filaments with three veins, anther thecae separate, embedded in the broad filaments, the connective prolonged.

MAGNOLIACEAE Jussieu - Nodes 6(+):6(+); stipules surrounding stem; flowers terminal; receptacle elongated, pollen boat-shaped. - 2/227: Magnolia (ca 225). The Americas (but not W. North America), and South East Asia to Malesia.

Magnoliaceae may be recognised by their entire or lobed leaf blades that are often more or less glaucous below and have large stipules that entirely surround the stem but are open on the side opposite the petiole. The flowers are rather large, usually single, and have an elongated receptacle bearing many petals, stamens and carpels that are obviously spirally arranged.

Beetles are common pollinators, and the flowers may be thermogenic; floral scents of the family have been extensively studied (Azuma et al. 1999; see also Seymour 2001 for thermogenesis). Nectar is secreted from the exposed surfaces of the carpels in some species of Magnolia. When the fruits open abaxially, the brightly-coloured seeds dangle from the fruit remaining attached by extended annular thickenings of the protoxylem.

Generic limits around Magnolia s. str. are unclear, the genus being wildly paraphyletic (e.g. Azuma et al. 2001, Kim et al. 2001a, b; Wang et al. 2006); it seems best to expanded the genus to encompass the whole of the current Magnolieae. The family thus includes only two genera. In Liriodendron the leaves are distinctively lobed, the fruit dry, single-seeded, and winged (it is a samara), the seeds lack a sarcotesta, and the endosperm is slight and not ruminate. In Magnolia the leaves are entire, the fruit may dehisce abaxially, there is a sarcotesta and a scleroendotesta, and there is much endosperm.

For information, see Nooteboom (1993: general), Yamada et al. (2003b: ovules), Pan et al. (2003: embryology) and Xu and Rudall et al. (2006: floral development), Xu and Kirchoff (2008: pollen morphology), Charlton (1994) and Liao and Xia (2007: phyllotaxis and vernation), and Figlar (2000: branching patterns). For a checklist and bibliography (under old generic names), see Frodin and Govaerts (1996).

[Himantandraceae + Degeneriaceae] [Eupomatiaceae + Annonaceae]: anthers valvate, staminodes internal, pollen without tectum, smooth; fruit indehiscent.

For the inner staminodes that are often so conspicuous in the flowers, and their function in pollination - food for pollinators, attractants - see Endress (1984).

Degeneriaceae + Himantandraceae: flowers axillary; x = 12.

DEGENERIACEAE I. W. Bailey & A. C. Smith - Leaves spiral; pollen boat-shaped; G 1; seeds with sarcotesta. - 1/2. Fiji.

Degeneriaceae may be recognised by their spiral, entire, estipulate leaves and large, axillary flowers with many tepals and a single carpel.

For some information, see Kubitzki (1993b: general).

HIMANTANDRACEAE Diels - Plant with peltate scales; P or bract + bracteoles enclosing the flower in 2 series, caducous, (1) 2 apical ovules/carpel; fruit ± syncarpous, a drupe with several stones; seeds flattened. - 1/2. The Celebes, New Guinea and N.E. Australia.

The two-ranked, estipulate leaves, the peltate scales that cover the plant, the flowers with many petals, and the more or less syncarpous and drupaceous fruits with laterally compressed stones/seeds allow one to recognise the family. The crushed leaves are very aromatic.

The perianth may be staminodial in origin.

For some information, see Endress (1993: general) and Doweld and Shevyryova (1997: seed).

Eupomatiaceae + Annonaceae: sieve tube plastids with protein crystalloids and starch; rays 8-15 cells wide; trunk leaves spiral; inflorescence present; fruit ± berry-like; mesotesta fibrous.

EUPOMATIACEAE Endlicher - Inflorescence fasciculate; receptacle concave, floral calyptra thick, deciduous, with sclereids; P 0; A introrse, basally connate, pollen with encircling equatorial sulcus. - 1/3. New Guinea and E. Australia.

Eupomatiaceae may be recognised by their two-ranked, exstipulate, aromatic leaves and their fairly large, calyptrate flowers with the fertile stamens surrounding many flat, gland-bearing staminodes that arch over the gynoecium. The receptacle is concave and bears many carpels tightly packed together with sessile but basically decurrent stigmas. The stamens and staminodes fall off the flower together, littering the ground beneath the plant with what look rather like sea anemones.

The main axes are mixed, initially being orthotropic and bearing spirally-arranged leaves, later becoming more or less plagiotropic and with distichous leaves. There are about three two-ranked bracts on the pedicel; the calyptra itself is often interpreted as being a modified, amplexicaul bract (e.g. Endress 2003b; esp. Kim et al. 2005b), although such a dramatic change in leaf insertion might seem unlikely.

For information, see Endress (1993: general) and Rix and Endress (2007 - an easy account).

ANNONACEAE Jussieu - Epidermal cells with crystals [?level]; flowers with open development; P 3 [smaller] + 3 + 3, very thick, valvate; A whorled, filaments with a single vein; endotestal plug +; tegmen crushed, ruminations irregular. - 129/2220. Largely tropical.

1. Anaxagorea - Trunk leaves distichous; uniseriate hairs terminate in a rounded cell; fruit a follicle; endotesta aerenchymatous, tegmen alone involved in ruminations, with oil globules. - 1/21. Tropical America, also Sri Lanka to West Malesia.

Ambavia clade + The Rest: internal staminodes 0.

2. The Ambavia clade - Anther connective ± tongue-like, ovules (1)2/carpel, integuments 3. - 6/54. Tropical, inc. Madagascar.

3. The Rest - Anthers with prominent ± truncate connective. - 128/2200.

3a. The Short Branch Clade - Polyalthia (150, see Mols et al. 2008a), Pseuduvaria (50: see Su & Saunders 2006; Su et al. 2008), Unoniopsis (45). Lowland tropics.

3b. The Long Branch Clade - Guatteria (280), Annona (120-175, soursop, sweetsop: inc. Rollinia), Xylopia (100-160), Uvaria (110-150: polyphyletic, see Mols et al. 2004), Artabotrys (100), Goniothalamus (50-120), Duguetia (70-95), Fissistigma (60), Monanthotaxis (55), Friesodielsia (50-60: ?monophyletic). Predominantly lowland tropics, rarely temperate.

Annonaceae are easily recognised, even when sterile. The twigs often dry blackish, the bark is conspicuously fibrous (when cut tangentially, there is a reticulum), the rays are broad, the young stems are often zig-zag, and the ± aromatic leaves are entire and two-ranked. The obliquely-pointing bud has clearly conduplicate leaves with the two halves of the blades pressed together. The flowers are nearly always pendulous and their development is more or less open, i.e., they are never obviously in bud, the tepals in particular continuing to enlarge even after the flower seems to have opened. The tepals are usually in three whorls of three members each, the outer whorl being notably smaller than the others, and there are numerous short stamens with a tough, truncate connective extending over the apex of the anthers. The seeds are quite large and most have a regularly ruminate endosperm with inpushings of the seed coat extending far into the endosperm.

Stem Annonaceae may be about 91-82 ± 4 million years old (Wikström et al. 2001), while diversification may have occurred 82-57 million years before present (Doyle et al. 2004; Scharaschkin & Doyle 2005). J. A. Doyle et al. (2004) and Richardson et al. (2004) discuss the historical biogeography of the family; the stem node of the ["long branch" + "short branch"] clade may date to 70-65 million years before present and the crown node to 66.7-56.6 ± 2.3 million years before present - in any event, although diversification is much earlier than in Myristicaceae, it is largely post continental drift in age. Divergence within the small genus Anaxagorea may have begun ca 44 million years before present (Scharaschkin & Doyle 2005), also comparatively early. Guatteria, with its ca 265 species, seems to have diversified abruptly in (Amazonian) South America perhaps 8.8-4.9 million years before present; the genus may have moved to South America from Central America (Erkens 2007; Erkens et al. 2007a; for its classification, see Erkens et al. 2007b).

In Neotropical Annonaceae pollination of the more or less odorous flowers is predominantly by a variety of beetles, although pollination by flies (e.g. Su et al. 2008) and thrips is also known; beetle pollination is common throughout the range of the family. It has been suggested that the tough, expanded connective that predominates in group 3 above may potect the pollen from the depradations of unwanted visitors (Gottsberger 1999; Silberbauer-Gottsberger et al. 2003). The flowers of some Annonaceae show thermogenesis (Seymour 2001). Dispersal of fruit is predominantly by mammals and birds.

The main axes of Annonaceae are often mixed (?Troll's model), initially being orthotropic and with spirally-arranged leaves, later being more or less plagiotropic and distichous (Johnson 2003), as also occurs in Eupomatia (?and other Magnoliales). Tetrameranthus sometimes seems to have opposite leaves; according to George Schatz (pers. comm.) it lacks plagiotropic branches and has all branches with spirally-arranged leaves. In some Xylopia, etc., the plagiotropic branch systems are frondose and very beautiful. Some Old World taxa like Artabotrys are lianes; there the ultimate branches have curved hooks which represent modified inflorescences while the orthtropic stems may bear stout, paired and rather vicious thorns.

Xylopia, like Anaxagorea also with staminodes and follicles, has micropylar-arillate seeds (Corner 1976), and other genera may also have small arils (Svoma 1997). Monodora has connate carpels, and in taxa like Annona the carpels become very tightly adpressed in fruit which is functionally a berry.

"The rest" is made up of two major clades, the Malmea-Piptostigma-Miliusa (MPM) clade, including a probably polyphyletic Polyalthia. The MPM clade shows relatively little molecular divergence (it is the "short branch" clade of Richardson et al. 2004) and is not very speciose, mostly lacking large genera other than Polyalthia; within this group, New World members form a monophyletic group (Pirie et al. 2006). Mols et al. (2008b) explore character evolution within the miliusoid clade. The other main clade is the inaperturate pollen clade, in which there is also more molecular divergence (the "long branch" clade); it includes Artobotrys, Guatteria, Xylopia, Annona, etc. Annona is in a clade members of which have pollen tetrads or polyads (see also J. A. Doyle et al. 2004; Pirie et al. 2005).

For information, see Jovet-Ast (1942: indumentum and anatomy), van der Wyck and Canright (1956: anatomy), Morawaetz (1986, 1988: esp. cytology), Christmann (1986: seed anatomy), van Heusden (1992: floral morphology), Kessler (1993: general), Doyle and le Thomas (1994, 1996: morphological phylogeny), Rudall and Furness (1997: tapetum), Svoma (1998a: seeds, 1998b: ovules), Deroin (1999a: receptacular vascular system), Doyle et al. (2000: pollen evolution), Bakker (2000a, b: general), Couvreur et al. (2008: evolution of syncarpy, etc.), and Sun et al. (2008: epidermal anatomy).

LAURALES Perleb [Back to Index]

Sesquiterpenes 0; sieve tube plastids with protein crystalloids; nodes 1:?; leaves opposite; inflorescence ± cymose; hypanthium +, inner staminodia +, 1(-2) basal ovules carpel, stylulus quite long; fruitlets single-seeded, indehiscent, hypanthium persistent; endotesta tracheidal [seed endotestal]; embryo long. - 7 families, 91 genera, 2858 species.

It is somewhat unlikely that the plesiomorphous ovule number is 1(-2) ovules per carpel; the latter number is found only in some Calycanthaceae, and the fruits of Laurales are always single seeded. The integuments do not always cover the apex of the ovule, which is thus exposed or naked, in at least some Atherospermataceae, Siparunaceae and Calycanthaceae (Endress & Igersheim 1997).

Relationships between Monimiaceae, Lauraceae and Hernandiaceae are difficult to work out, this being one of the few cases where there seems to be persistent disagreement between morphology and molecules (Renner & Chanderbali 2000). Here I follow morphology (see Renner 2005a for the most recent study).

See also Metcalfe (1987) for anatomy, Endress and Igersheim (1997) and Eklund (1999) for general information, Renner et al. (1997), Renner (1998, 1999) for relationships, Kimoto and Tobe (2001) for embryology, and Renner (2005a) for diversification.

CALYCANTHACEAE Lindley - Stem with four inverted cortical bundles; P many, spiral, no double fertilisation; endosperm diploid, developing autonomously. - 5/11. China, North America, N.E. Australia.

1. Idiospermum - G 1-2(-3), outer integument 12-15 cells thick, stigma massive; cotyledons (3-)4, massive, peltate. - 1/1: Idiospermum australiense. N.E. Australia, rain forest.

2. The Rest - Pollen disulcate; G 5-35, outer integument 5-6(-8) cells thick; cotyledons spirally twisted. - 4/10. China, North America.

Calycanthaceae can be recognised by their opposite, entire, exstipulate leaves that smell aromatic. The cortical bundles can be seen with a lens when the stem is cut, and they may also be visible externally as four, raised lines along the stem. The terminal bud often aborts. The large flowers have a bract-covered hypanthium; the inner members of the perianth secrete nectar or bear food bodies. The achenial fruits are attached to the inside of the persistent hypanthium.

Staedler et al. (2007) note that the numbers of floral parts, tepals, stamens, staminodes, etc., are more or less those of the Fibonacci series (3, 5, 8, 13,....). There is no triple fusion during fertilisation, and the endosperm develops autonomously. The seeds are poisonous and have characteristic alkaloids; the embryo of Idiospermum is the largest known in angiosperms.

Fossils identified as Calycanthaceae have interesting morphologies. Araipa florifera, from the Lower Cretaceous of Brasil, has flowers that externally are very like those of Calycanthaceae, but its leaves are lobed (Mohr & Ecklund 2003); unfortunately, nothing is known of the internal structure of the flower. The late Cretaceous Virginianthus calycanthoides (98-113 million years before present) has been placed in Calycanthaceae. It has small flowers, anthers dehiscing by lateral hinges (see also Sinocalycanthus which has laterally-attached flaps - independently evolved?), and reticulate pollen with a single sulcus (Friis et al. 1994). Its inclusion may change one or two ordinal/family group characters. Its anasulcate pollen is like that of Idiospermum, so mono(ana-)sulcate pollen may well be plesiomorphic for the order, disulcate pollen being an apomorphy for the rest of Calycanthaceae and inaperturate pollen an apomorphy for the rest of Laurales. Assuming the lateral hinges on the anthers of Virginianthus are equivalent to the rather differently oriented hinges found in most other Laurales, hinged anthers may be a synapomorphy for the order, and perhaps anthers with slits a synapomorphy for crown-group Calycanthaceae. However, Virginianthus may be sister to all other Laurales, or even may not belong to Laurales at all... (Eklund 1999; Zhou et al. 2006; Doyle & Endress 2007). The rather younger (Turonian, ca 90 million years before present) Jerseyanthus may be sister to Calycanthus; it has the distinctive disulcate pollen common in Calycanthaceae (Crepet et al. 2005). However, it is remarkable in having an arrangement of flower parts quite unlike that of any other angiosperm. From the outside, the sequence is petaloid tepal - introrse staminode - extrorse stamen - abaxially curved "petaloid staminode" - pistillode - but note that Staedler et al. (2007a) interpret the outer staminode series as being inner members of the tepals.

For general information, see Nicely (1965) and Kubitzki (1993b), for phylogeny and morphology, see Li et al. (2004), and for gynoecial development, see Staedler et al. (2007b).

[Siparunaceae [Gomortegaceae + Atherospermataceae]] [Monimiaceae [Hernandiaceae + Lauraceae]]: vessel elements with scalariform perforation plates; hippocrepiform [horseshoe-shaped] sclereids in pericycle; lamina with rather distant teeth; flowers rather small; A whorled, stamens with paired nectaries/glands at base, anthers bisporangiate/unithecal, with apical flaps [valvate], tapetum ?, pollen inaperturate, ± spinulose, 1 ovule/carpel; fruit proper a drupe, endocarp lignified.

There is little information on tapetal development, or of most other embryological details and of seed anatomy; this is especially true of the first three families (see Kimoto & Tobe 2001 for a summary; Bello et al. 2002a for Siparuna). For fruit anatomy, see Romanov et al. (2007); they note that even the dry fruits of Atherospermataceae have a lignified endocarp. What about Calycanthaceae?

Siparunaceae [Gomortegaceae + Atherospermataceae]: acicular crystals +; hypanthium closed by roof; embryo very small.

SIPARUNACEAE Schodde - Anthers bisporangiate/unithecal, with a single flap, integument 1, thick. - 2/75: Siparuna (74). Tropical America (Siparuna), W. Africa (Glossocalyx).

Siparunaceae may be recognised by their opposite, serrate, estipulate leaves with rather prominent venation, the spreading veins looping and joining towards the margin [brochidodromous]; the wood lacks broad rays and the hairs are often stellate or fasciculate. The anthers lack basal glands.

Pollination is by cecidomyid gall midges, which pollinate as they oviposit in the flowers, however, larvae have been found only in staminate flowers. Dicoey is common in Siparuna, and it seems to have evolved several times from monoecy (Renner et al. 19999; Renner & Won 2001). The individual drupes may have a fleshy appendage that often differs in color from the rest of the fruit (the "aril" of Renner & Hausner 1997); it occurs only in dioecious taxa. In most taxa the hypanthium splits to expose the drupelets inside, whether or not they have fleshy appendages, and the open hypanthium presents a striking color contrast with the drupes to the would-be frugivore.

For information, see Philipson (1993: general) and Kimoto and Tobe (2003: embryology).

Gomortegaceae + Atherospermataceae: bud scales +; sieve tube plastids also with proteinaceous fibrils; outer stamens alone staminodial.

GOMORTEGACEAE Reiche - Lamina entire; G [2-3(-5)], inferior, ovule apical, stigmatic branches erect; 1 drupe/flower, endocarp wall thick; embryo large, endosperm slight. - 1/1: Gomortega keule. C. Chile, rare.

Gomortegaceae are aromatic plants with opposite, estipulate, entire leaves. The flowers are small, with staminodes outside the stamens, the ovary is inferior and syncarpous, and the single-seeded fruit is drupaceous.

For information, see Kubitzki (1993b: general), Doweld (2001: seed) and Heo et al. (2004: embryology).

ATHEROSPERMATACEAE R. Brown - Stomata anomocytic; fruit achenial, plumose, hypanthium woody. - 6-7/16. New Guinea to New Zealand and New Caledonia, Chile.

Atherospermataceae have opposite, coarsely serrate leaves in which the prominent lateral veins join to form a looping vein inside the margin. The flowers are quite small, the hypanthium is dry, and the individual fruitlets are plumose and achenial.

The family is known from forests on the Antarctic Peninsula of the late Cretaceous/early Tertiary, with wood recorded from the Upper Eocene of Germany; the oldest fossils are from ca 88 million years ago. An age for the clade as a whole of ca 140 million years has been suggested (Renner et al. 2000).

For information, see Philipson (1993: general), and Poole and Gottwald (2001: wood anatomy).

Monimiaceae [Hernandiaceae + Lauraceae]: A whorled, pollen lacking both columellae, foot layer and endexine, ovule apical.

MONIMIACEAE Jussieu - Paired nectaries/glands at base of A absent, anthers dehiscing longitudinally, with 4 sporangia, style short, stigma stout. - 22/200: Mollinedia (90 [?20 - S. Renner, pers. comm.]), Tambourissa (45), Kibara (45). Tropical, but esp. Australasia.

Monimiaceae are recognisable by their opposite, rather distantly serrate leaves which have prominent venation on the lower surface with the veins looping and joining near the margin, and their twigs, which have broad rays and are often flattened below the somewhat swollen nodes; some taxa have "laticiferous threads" (Keller 1996). In fruit, the well-developed fleshy hypanthium and/or receptacle is conspicuous; this may enclose the drupelets and it then splits irregularly.

Decaryodendron has flowers with up to 1000 carpels, those of Tambourissa have up to 2000 carpels. A hyperstigma may be developed. When the hypanthium splits, as in Palmeria, the colour of the inside of the hypanthium forms a striking contrast with that that of the drupelets. Kibaropsis has four cotyledons.

Palmeria, Peumus, and Monimia form a clade that is sister to the rest of the family (Renner 2002 and references). Fruit anatomy correlates quite well with phylogeny, thus the first three genera have a massively thick endocarp, and in this they are alone in the family (Romanov et al. 2007).

For general information on Monimiaceae and other families previously included in it, i.e. Atherospermataceae and Siparunaceae, see Schodde (1970) and Philipson (1993), also Sampson (1993, 1997: pollen), Renner (1998: phylogeny), Poole and Gottwald (2001: wood anatomy), Romanov et al. (2007: fruit anatomy) and Kimoto and Tobe (2008b: embryology).

Hernandiaceae + Lauraceae: vessel elements with simple perforation plates; mucilage cells +; sieve tube plastids also with starch; leaves spiral (lobed); all flower parts whorled, tapetum amoeboid, exine of pollen thin, intine very thick; G 1, inferior [one optimisation: see below], outer integument ³4 cells across, embryo sac more or less linear; testa multiplicative, thick, tegmen not persisting; endosperm 0.

Another scenario for ovary evolution is that it became inferior independently in Hernandiaceae and in Hyphodaphnis - both require two steps (gain, loss versus gain, gain).

HERNANDIACEAE Berchtold & J. Presl - Fruit dry. - 5/55. Pantropical.

1. Hernandioideae - Anther valves laterally hinged, single layer of microspore mother cells, pollen grains 90-160 µm across, outer integument 10-23 cells thick, nucellus massive, 6-8 layers of parietal cells, nucellar beak +; bracteoles accrescent in fruit; testa vascularized, spongy, mesotesta massive, 7-17 cells across. - 3/44. Tropical, esp. Madagascar and Indo-Malesia.

2. Gyrocarpoideae Pax - Leaves with strong higher-order vein areolation; inflorescence ebracteate; pollen grains 19-45 µm across, apical part of embryo sac protruding; cotyledons contortuplicate [much folded]. - 2/10. Pantropical, esp. America.

Given the variation in life form and basic vegetative morphology in Hernandiaceae, this is not an easy family to recognise; Illigera can even be mistaken for the monocot Dioscorea.

Illigera (Hernandioideae) is a climber with palmately compound leaves and sensitive petioles, while Sparattanthelium (Gyrocarpoideae) is a climber with recurved stem hooks.

See Heo and Tobe (1995) for embryology, etc., Kubitzki (1969, 1993b) for general information, and Kimoto and Tobe (2008a) for more embryology, including a nice summary.

LAURACEAE Jussieu - Leaves entire. - Ca 50/2500: Ocotea (350), Cryptocarya (350), Litsea (?400), Beilschmiedia (250), Persea (200: P. americana, avocado - dianthesis, flowers in two phases, plants have flowers in only one phase), Cinnamomum (350: cinnamon), Lindera (100). Pantropical (temperate), lowland to montane. Some of the distributional area of the family, e.g. in most of West Australia, is attributable to Cassytha alone.

1. Hypodaphnis zenkeri - Anthers tetrasporangiate, staminodes 0. - 1/1: Hypodaphnis. Tropical West Africa.

[Beilschmeidia, Cryptocarya, Endiandra, etc.] [Cassytha [[Caryodaphnopsis + Neocinnamomum] + The Rest]]: subsidiary cells of paracytic stomata envelop the guard cell above and below, the latter having outer and inner cuticular ledges; staminodes +, glandular tapetum +, ovary superior, embryo sacs protruding.

2. Beilschmeidia, Cryptocarya, Endiandra, etc. - Stamens in two whorls, ovary superior [reversal, the first optimisation]. - 6/710. Pantropical, some subtropical, to New Zealand.

Cassytha [[Caryodaphnopsis + Neocinnamomum] + The Rest]: ?

3. Cassytha - Parasitic herb; ?tapetum, micropyle bistomal, nucellar cap 0; endosperm cellular. - 1/16. Old World tropics.

[Caryodaphnopsis + Neocinnamomum] + The Rest: ?

Caryodaphnopsis + Neocinnamomum - Anthers tetrasporangiate. - 2/13. South East Asia to the Philippines and Borneo.

5. The Rest - Tapetum amoeboid, embryo sac not protruding. - Ca 40/1730. Pantropical (temperate).

The family may be recognised by its aromatic smell, nearly always entire spiral or opposite (very rarely indeed two-ranked) leaves the blades of which are often quite thick (they are coriaceous when dry) and with a glaucous undersurface and rather steeply ascending secondary veins. The twigs are often more or less ridged, and branching often occurs as the current flush grows out, i.e., it is sylleptic - even in temperate taxa there is often a mixture of proleptic and sylleptic branching on the same plant. The flowers are small, trimerous, and have valvate anthers; the fruit has a single, large seed that lacks endosperm and the pedicel - and sometimes also tepals and hypanthium - is often ± swollen and conspicuous.

Some Lauraceae have a growth pattern similar to that of Myristicaceae, with whorled branches and periodic growth, but the leaves on the branches of Lauraceae are usually spiral, rarely opposite, while in Myristicaceae (and Annonaceae) the leaves are distichous.

Lauraceae are prominent in the Mid to Late Cretaceous flora, and include the Early Cenomanian Mauldinia ca 100 million years old (Drinnan et al. 1990; Viehofen et al. 2008). This has distinctive inflorescences in which the lateral units are flattened and bilobed, bearing sessile flowers on their adaxial surfaces that are very like flowers of extant members of the family (see also Crepet et al. 2004; Friis et al. 2006b for references). Recently a flower provisionally assigned to Lauraceae, although with a unusual combination of floral characters - there are only two whorls of stamens and no nectaries/glands - has been described from deposits in Virginia some 112-105 million years old (von Balthazar et al. 2007). Chanderbali et al. (2001) suggest an date of 174 ± 32 million years before present for the origin of the family (stem group age).

The family is particularly prominent in the lowland tropical rainforest of Africa and America, and in tropical montane forests in the latter it may be the most speciose family (Gentry 1988), but it is common throughout the tropics. The fruit proper is commonly blue and dispersal is predominantly by birds.

The genera are very difficult and rather unsatisfactory, often being based on single character differences in the androecium (Rohwer 1993a, 1994a; Werff & Richter 1997). However, both extrorse and introrse anthers can occur in the same flower, although the extrorse condition seems to be developmentally derived (Buzgo et al. 2007), there may be 2 or 4 sporangia, their arrangement varies, etc.

It has been suggested, largely on the basis of gene expression, that the perianth in some Lauraceae - Persea, at least - may represent modified stamens (Chanderbali et al. 2004, 2006), however, both the tepals and the stamens of Persea have three traces (Reece 1939). Yet other reports suggest that lauraceous stamens have single traces, even if both whorls of tepals have three traces (Laurus) or often one trace in the inner whorl (Umbellularia - see Kasapligil 1951). Trace number by itself is telling us little. Multistaminate Lauraceae attain this condition by the increase in number of the stamen whorls. The fruit is often a drupe, only sometimes a single-seeded berry.

Hypodaphnis, with an inferior ovary, is sister to the rest of the family, so if Hernandiaceae and Lauraceae are sister taxa, change from superior to inferior will be put at different places on the tree depending on if ACCTRAN or DELTRAN is used. A number of clades have long branches, and complex analyses by Rohwer and Rudolph (2005) suggest the relationships: Hypodaphnis [[the Cryptocarya group] [Cassytha (this has a long branch) [[Caryodaphnopsis + Neocinnamomum] [[the Mezilaurus group] [core Lauraceae]]]]] - most of these clades have about 1.00 posterior probabilities. If this topology is confirmed, it will have considerable implications for character evolution. In the topology suggested by Rohwer and Rudolph (2005: but cf. Kimoto et al. 2006) a glandular tapetum is to be found in all species that have been examined that occur on the first three successive branches of the tree that have woody taxa; embryo sacs protruding through the micropyle are found in most of them. However, these features are not, or only very rarely, found in the Mezilaurus group and in core Lauraceae, which suggests that there may have been reversal/loss of these characters there. I have optimised these and some other characters (especially von Balthazar et al. 2007) in the context of the provisional phylogeny of Lauraceae noted above, but it is clear that where most characters might properly be placed on the tree is very uncertain.

Distinctive paracytic stomata in which the subsidiary cell envelops the guard cell above and below, the latter having outer and inner cuticular ledges, may be an apomorphy for all Lauraceae except Hyphodaphnis (Carpenter et al. 2007; but cf. Nishida & van der Werff (007).

General information is taken from Rohwer (1993a), some embryological details from Heo et al. (1998) and Endress and Sampson (1983), cuticle from Christophel et al. (1996), cytology from Oginuma et al. (1998), general wood anatomy from Richter (1981) and van Rijckevorsel (2002), and androecial development from Buzgo et al. (2007).

CANELLALES + PIPERALES: aporphine alkaloids; nodes 3:3; G whorled.

CANELLALES Cronquist [Back to Index]

Indumentum 0; sieve tube plastids with starch and protein crystalloids and/or fibres; foliar sclereids +, branched; calyx and corolla distinct; fruit a berry. - 2 families, 9-13 genera, 75-105 species.

For nodal anatomy, see Sugiyama (1979), for general vegetative anatomy, see Metcalfe (1987).

CANELLACEAE Martius - Cortical vascular system +; G [2-6], placentation parietal, ovules campylotropous; fruit a berry, K persistent. - 5/13. Tropics; U.S.A. (S. Florida), Antilles, South America, Africa, Madagascar.

Canellaceae are aromatic evergreen shrubs or trees often with rather pale bark that may be recognised by their rather thick, exstipulate leaves with entire margins and flowers with three outer sepals, petals, connate stamens with extorse anthers, and ovary with parietal placentation. The fruit is a berry.

For a phylogeny of Canellaceae, see Salazar and Nixon (2008); Cinnamomodendron is polyphyletic.

Some information is taken from Kubitzki (1993b: general).

WINTERACEAE Lindley - Vessels 0; rays 10+-seriate; K calyptrate, pollen in tetrads, monoporate; exotesta palisade. - 4-7/60-90. Montane tropics, not mainland Africa.

Taktajanioideae Leroy - Stomata anomocytic; G [2], placentation parietal. - 1/1. Takhtajania perrieri. Madagascar.

Winteroideae - 3-6/60-90. New Guinea to New Zealand and New Caledonia, few Borneo and the Philippines and South America.

Winteraceae are often rather small evergreen trees or shrubs that can be recognised by their aromatic and/or peppery taste and their spiral, entire leaves that are often glaucous below. There is a more or less obviously calyptrate calyx more or less covering the flower although the other flower parts are nearly always free. The fruit is often a berry.

Fossils assigned to Winteraceae are have a much wider distribution than that of the family today (Vink 1993, for map), and fossil pollen is known from the Cretaceous Albian-Aptian some 125 million years before present (Doyle 1999; see also Sampson 2007 for fossil pollen). Fossil wood has been reported somewhat later from the Maastrichtian of California (see Vink 1993, for literature). However, divergence of Takhtajania from other Winteraceae may have occurred only some 46-41 million years ago (Wikström et al. 2001).

The occlusion of the stomata that is common in Winteroideae prevents the stomatal apertures from being wetted; Winteraceae in general grow in moist environments. The flowers of some Winteraceae show thermogenesis (Seymour 2001), and pollination is by a diversity of small insects, including thrips. In New Caledonia the primitive jawed moths, Sabatinca (Micropterigidae, sister to all other Lepidoptera), eat the pollen of Zygogynum; the tetrads are covered by much oily pollen-kitt which makes them stick to the moths. Sabatinca also uses the flowers as a place of assembly prior to mating (Thien et al. 1985: cf. Nothofagaceae).

Is Drimys piperita the only species in Malesia, with Drimys also in S. America, or should there be two genera - Drimys (flowers perfect: South America) + Tasmannia (plant dioecious: 30+ species in New Guinea alone and a few more at least in Borneo)? A generic distinction between the Old and New World taxa is strongly supported in molecular analyses by Karol et al. (2000), Doust (2003) and Doust and Drinnan (2004), while in the field in Papua New Guinea at least there are morphologically and ecologically very distinctive species that grow together yet show at most uncommon hybridisation.

See also Praglowski (1979: pollen), Vink (1985, 1993: general), Svoma (1998b: ovules), Tobe and Sampson (2000: embryology), Ehrendorfer and Lambrou (2000: cytology) and Keating (2000a: anatomy).

PIPERALES Dumortier [Back to Index]

Herbs; vessel elements with simple perforation plates; wood with broad rays; nodes often swollen; stomata ?; leaves two-ranked, 2ndary veins palmate; flower parts whorled; P and A and (±) G in 3's; G occlusion?; seed coat ± tegmic. - 4 families, 17 genera, 4090 species.

Carlquist et al. (1995) suggest a number of wood anatomical characters that may be common to this clade, e.g. wood both in some Aristolochiaceae and Piperaceae is storied (Carlquist 1992). There is extensive variation in the differentiation of the embryo in Piperaceae, and the polarity of evolution of this feature is unclear, as is that of micropylar morphology, etc. Variation in embryo sac morphology in the whole clade is also very considerable, and there are now attempts to put this in a phylogenetic context (e.g. Madrid & Friedman 2008a, 2008b, 2009).

Hilu et al. (2003) suggested that Aristolochiaceae are paraphyletic and include the rest of the order (Hydnoraceae were not sampled), Wanke et al. (2007: Hydnoraceae again not included) found strong support for inclusion of Lactoridaceae in Aristolochiaceae (as sister to Aristolochioideae), while Nickrent and Blarer (2005) found moderate support for the clade [Hydnoraceae + Aristolochioideae]. In some floral details, Saururaceae are very like Acoraceae (Buzgo & Endress 2000), e.g. they both have monosymmetric flowers. Similarly, the three-merous perianth and adaxial prophylls that seem to suggest a relationship between Piperales and monocots (and Nymphaeales) are also parallelisms.

For floral development, see Tucker and Douglas (1996).

Hydnoraceae + Aristolochiaceae: P connate, valvate, anthers extrorse, embryo undifferentiated.

HYDNORACEAE C. Agardh - Echlorophyllous root parasites; leaves 0; flowers arising endogenously from "roots", large; P uniseriate, thick and fleshy, A adnate to and opposite P, sessile, connate, polythecate, ovary inferior, placentation of parietal or apical lamellae, ovules straight, unitegmic, tenuinucellate, stigma broad, cushion-shaped. - 2/7. Arabian Peninsula, Africa, Madagascar; Costa Rica and S. South America.

Hydnoraceae are leafless and echlorophyllous parasites that can be recognised by their large flowers with a single, thick, basally connate, valvate perianth whorl that is brownish and cracked outside and coloured inside, stamens opposite the perianth lobes, and inferior ovary; the fruit contains many small seeds.

In Hydnora triceps both flower and fruit are underground. Pollination of the foetid flowers of Hydnora is by flies and beetles, as in Aristolochiaceae (Bolin et al. 2006b), and thermogenesis occurs in the flowers of Prosopanche (Cocucci & Cocucci 1996). Dispersal is by mammals. Each flower has up to 35,000 ovules.

Tennakoon et al. (2007) suggest that the so-called pilot roots, with their scattered vascular bundles, are in fact better interpreted as modified stems. Cork is continuous over this stem, even over the apical meristem.

For additional information, see Cocucci (1976: embryology) and Meijer (1993: general), and Bolin et al. (2006a: germination).

ARISTOLOCHIACEAE Jussieu - Stomata anomocytic; leaves conduplicate, heart-shaped, 2ndary venation palmate, prophyll adaxial; inflorescence cymose: P 3, odd member adaxial; A in 3's, ± sessile, connective extended apically, carpels basically free; fruit a follicle; endotesta palisade, usu. crystalliferous, exotegmen and layer underneath crossing fibres, endotegmen with reticulate thickenings. - 5-8/480. World-wide, not Arctic - three groups below.

1. Asaroideae O. C. Schmidt - Growth sympodial; sieve tube plastids with cuneate protein crystalloids and a large polygonal crystal. - 2/75: Asarum (70). N. Temperate, esp. East Asia.

Lactoris + Aristolochioideae: Growth monopodial; bracts distinct.

2. Lactoris - Sieve tube plastids with starch grains; nodes 1:2; leaves elliptic, 2ndary veins subpinnate, stipule sheathing-intrapetiolar, adnate to the petiole; bracteoles 0; flowers small; A 6, pollen in tetrads, saccate; G 3, ovules ?tenuinucellate, funicle long, endothelium +. - 1/1: Lactoris fernandeziana. Chile, the Juan Fernandez Islands.

See Carlquist (1964: general, 1990: wood anatomy), Crawford et al. (1986: chemistry), Metcalfe (1987: vegetative anatomy), Kubitzki (1993: general), Tobe et al. (1993) and González and Rudall (2001: general, the stipule is initially paired) for details.

3. Aristolochioideae - Benzylisoquinoline alkaloids +; sieve tube plastids with polygonal protein crystalloids plus starch grains or protein fibres (starch grains alone); axillary buds serially arranged; hairs hooked; inflorescences axillary; floral primordia monosymmetric; G [4-6], inferior, apically constricted; fruit opening laterally along the septal radii, or a schizocarp or dry-baccate. - 2-5/405. Tropics (temperate), relatively less diverse in Africa (inc. Madagascar), few in N. Australia.

Aristolochiaceae are more or less herbaceous plants (they are quite often vines) that may be recognised by their adaxial prophylls, exstipulate leaves with palmate venation and entire margins, and 3-merous flowers with extrorse stamens. The ovary is usually inferior and the perianth uniseriate and connate. Only Aristolochia has monosymmetric flowers, only Saruma a perianth differentiated into sepals and petals and a largely superior gynoecium of free carpels.

Lactoris can be recognised by its small, two-ranked leaves with intrapetiolar stipules; the prophylls are adaxial. The flowers are small and have three imbricate perianth members and three free carpels.

Pollen like that of Lactoris has been found in Late Cretaceous deposits from S.W. Africa (Turonian-Campanian - 93-76 million years ago) to Oligocene deposits in Australia, etc. (Zavada & Benson 1987; Gamerro & Barreda 2008 and references).

Aristolochia is eaten by caterpillars of the magnificent birdwing butterflies of the Papilionidae-Troidini, and the association between caterpillars of these butterflies and Aristolochiaceae (they are apparently absent from Saruma, although larvae of the related Luehdorfia [Zerynthiini] have been reported from this plant) has been studied in some detail (e.g. Weintraub 1995). However, there seems to be no particular association between the phylogeny or chemistry of the plant and the butterflies (Silva-Brandão & Solferini 2007). The flowers of some Aristolochiaceae show thermogenesis (Seymour 2001); pollination by flies and beetles, which become trapped in the flowers, is common. In Aristolochia grandiflora the abaxial tepal is prolonged into a dangling appendage to 1.5 m long (Bello et al. 2006). The shrubby habit is derived, and the woody Aristolochia arborea is cauliflorus; its flowers looks as if they have a small mushroom growing inside!

There has been much discussion about the perianth in Aristolochiaceae. González and Stevenson (2000) suggest that the uniseriate perianth is derived from the outer whorl of a biseriate perianth; in any inner whorl, "petal" bases are narrow, although the bases of members of the outer whorl encircle the axis. it has also been suggested that "petals" may be derived from stamens (cf. Ronse De Craene et al. 2003), although since these petals are found in the angles of the outer whorl and the stamens are more or less adnate to the style, this is unlikely. Jaramillo and Kramer (2004) describe the basic perianth condition for the family as being unipartite (= uniseriate), with its ancestors having "multiple" whorls. See Leins and Erbar (1995) for the flowers of Saruma, which seem very different from those of the rest of the family. The sepals and petals are quite distinct, the pollen is sulcate, and the nine carpels are adnate to hypanthium, but are otherwise free.

Some splitting of Aristolochia, perhaps into four genera, seems to be favoured (Neinhuis et al. 2005); Wanke et al. (2006b) note that the four main clades in the genus all have synapomorphies. González and Stevenson (2002) provide a phylogenetic analysis of Aristolochioideae and suggest that it be split into four genera; Ohi-Toma et al. (2006) propose dividing it into two. Lactoris has until very recently been placed in its own family, Lactoridaceae, and it was not considered to be immediately related to Aristolochiaceae, but molecular data suggest that it should be placed here (see above). Kelly (1998) discusses relationships in Asarum.

For information, see Huber (1985: seed characters, 1993: general), González (1999: inflorescence morphology), González et al. (2001a: microsporogenesis, 2001b: gynoecium), Behnke (2003: sieve tube plastids), Kelley and González (2003: a morphological phylogenetic analysis), Leins et al. (1988: floral development), and González and Rudall (2001: ovule and seed development).

Piperaceae + Saururaceae: stomata tetracytic; leaf base broad, ± sheathing; inflorescence spicate, terminal; flowers small; P 0, filaments rather slender, microsporogenesis simultaneous, pollen grains <20 µm, ovules straight; seed coat exo- and endotegmic; perisperm +, endosperm scanty, embryo broad.

Some information is taken from Murty (1960) and Tucker et al. (1993). See Jaramillo et al. (2004) for the complexities of floral evolution in this group; it is possible that having a four-carpellate gynoecium is the basic condition (but see also Wanke et al. 2007b).

PIPERACEAE Martynov - Vascular bundles not in a single ring; flowers very small, bracts peltate; embryo sac tetrasporic; fruit a drupe. - 5/3615. Pantropical.

1. Verhuellioideae Samain & Wanke - 1/3. Cuba and Hispaniola.

2. Zippelioideae Samain & Wanke - 2/6. China to Malesia, Central and South America.

3. Piperoideae Arnott - 2/3600: Piper (2000), Peperomia (1600). Pantropical.

Piperaceae may be recognised vegetatively by their herbaceous habit and/or their separate vascular bundles in the stem, their swollen nodes, and their soft and/or fleshy leaf blades the bases of which are often cordate. The minute flowers lack a perianth and are often borne in dense spikes; individual flowers are subtended by peltate to clavate bracts.

Piper and Peperomia diverged in the late Cretaceous, but species diversification is mid-Tertiary and later (Smith et al. 2008). Peperomia is a notable component of the epiphytic flora, particularly in the neotropics; the epiphytic habit is derived, as is the geophytic habit - several times (Symmank et al. 2008). Carollia bats (Phyllostomidae) are abundant New World fruit-eating bats that specialize on eating Piper living in the understorey; Old World Piper are bird-dispersed (Fleming 2004). Some New World species of Piper look remarkably like Araceae!

Peperomia shows considerable variation in the nature (druses, raphides) and pattern of oxalatae deposition in the leaf (Horner et al. 2009 - spectacular under polarizing light), but with little obvious correlation with phylogeny. Syncarpy is weak; Piper has separate carpel primordia. The endosperm of Peperomia may be up to 15n, while in taxa like Manekia it is triploid. Indeed, there is considerable variation - some infraspecific - in the particlar kind of tetrasporic embryo sac development in the family (Arias & Williams 2008: Verhuellia not yet studied). In Zippelia the zygote remains as such up to the maturity of the seed.

Relationships may be [Verhuellia [[Zippelia + Manekia] [Piper + Peperomia]]] (Jaramillo & Callejas 2004; Wanke et al. 2006a, 2007a, b); this entails redrawing the old subfamilial boundaries. The recent discovery that Verhuellia is sister to the rest of the family (Wanke et al. 2007b) changes hypotheses as to the plesiomorphous characters of the family; of the three genera in the two small clades that are successively sister to Piper and Peperomia, we know little about two. Jaramillo and Callejas (2004) and Smith et al. (2008) found that Piper s. str. was divided into New and Old World clades, the latter, Piper s. str., being divided into a mainland Asian clade, containing both the two endemic African species and a species from Australia, and also a Pacific islands Macropiper clade including the economically very important Piper methysticum (Jaramillo & Callejas 2004 found that one African species they examined grouped with their Pacific clade - see also Jaramillo et al. 2008; Smith et al. 2008). This Pacific clade, the Macropiper clade, is either sister to the rest of the genus or sister to the Asian clade (Jaramillo et al. 2008). Interestingly, in a trnK/matK analysis, Wanke et al. (2007a) found much less resolution within Piper than Peperomia.

Although Peperomia is a very distinctive genus with its bisporangiate anthers, inaperturate pollen, gynoecium of a single carpel with a penicillate stigma, cellular endosperm, etc., its recognition as a separate family would make Piperaceae paraphyletic; for the phylogeny of Peperomia, see Wanke et al. (2006a, 2007a). (Peperomia also has the dubious distinction of having the most herbarium names, about 1,530. These are names known primarily from herbarium sheets and were coined mostly by Trelease, and are mostly synonyms [Mathieu 2007]!) For the classification of Piperaceae followed here, see Samain et al. (2008); the subfamilies are not easily characterisable.

For information, see Weberling (1970: stipules), Bornstein (1991: general), Johri et al. (1992: embryology), Tebbs (1993: general), Lei et al. (2002: embryology of Zippelia), and for floral development, see Lei and Liang (1998: Piper; 1999: Peperomia and Tucker et al. (1993: Zippelia).

SAURURACEAE Martynov - Inflorescence bracts petaloid; A introrse, carpels connate, placentation parietal. - 4/6. North Temperate.

Saururaceae may be recognised by their often aromatic and rather fleshy leaves with broad bases and palmate venation and by their spicate inflorescences that usually have a large, basal inflorescence bract. The flowers are small, without a perianth, but there are clearly three to four carpels.

According to Murty (1960) the single intrapetiolar stipule represents two, connate stipules. It is possible that petaloid inflorescence bracts evolved twice in the family and are not an apomorphy for it. Saururus lacks these petaloid bracts (and its carpels are connate only at the base) and only one species of Gymnotheca has them. Houttuynia has tenuinucellate ovules

Houttuynia and Anemopsis are sister taxa in a matR analysis (Meng et al. 2002, 2003); this pair is also found in a three-gene analysis, but with poor support, although [Saururus + Gymnotheca] are a better-supported pair (Jaramillo et al. 2002; see also Neinhuis et al. 2005). These relationships are not recovered in morphological analyses.

For information, see Wu and Kubitzki (1993: general) and Carlquist et al. (1995: wood anatomy).

[MONOCOTS [CERATOPHYLLALES + EUDICOTS]]: (veins in lamina often 7-17mm/mm² or more [mean for eudicots 8.0]); (stamens opposite [two whorls of] P); (pollen tube growth fast).

Details of the exact position and magnitude of changes in venation density and pollen tube growth are still provisional (see Boyce et al. 2008, 2009; Williams 2008 for more details).

MONOCOTYLEDONS [Back to Index]

Herbaceous, plant sympodial; non-hydrolyzable tannins +, benzylisoquinoline alkaloids, hydrolysable [ellagi]tannins 0; root cork cambium [uncommon] superficial; root medullated, lateral roots arise opposite phloem poles; primary thickening meristem +; vascular bundles in stem scattered, closed [no interfascicular cambium developing]; sieve tube plastids with cuneate protein crystals alone; epidermis with bulliform cells [large, thin-walled cells: ?this level], leaves not differentiated into petiole plus lamina, main venation parallel, developing acropetally and basipetally from the base and converging towards the apex, intermediate [and other] veins basipetal from apex, endings not free, (margins with spiny teeth), base sheathing, sheath open [margins not fused]; inflorescence racemose; flowers 3-merous [6-merous to the pollinator?], polysymmetric, pentacyclic; T in two whorls, median member of outer whorl abaxial, members of whorls alternating, similar, [pseudomonocyclic, each providing a sector for the T tube when present]; stamens = and opposite each T member, anther and filament sharply distinguished, anthers subbasifixed; G [3], placentation axile, antipodal cells persistent and proliferating; fruit a loculicidal capsule; seed testal; embryo long, cylindrical, cotyledon 1, terminal, plumule lateral; primary root unbranched, not very well developed, "adventitious" roots numerous, cotyledon with a closed sheath, unifacial, assimilating and haustorial.

11 orders, families, 60,100 species.

Some features that are likely to be synapomorphies are in bold. There are non-secreting slits in the ovary septae of Acorus (Buzgo 2001); if these are considered to be septal nectaries, this feature may be a synapomorphy (lost many, many times) of monocots as a whole.

The oldest unequivocally monocotyledonous fossils (as pollen) are from the Late Barremian-Early Aptian of the Cretaceous some 120-110 million years ago and are assignable to Araceae-Pothoideae-Monstereae; Araceae are sister to other Alismatales, which are in turn sister to all monocots except Acorales (Friis et al. 2004: for fossil monocots, cf. Gandolfo et al. 2000 and Friis et al. 2006b). Bremer (2000b) had earlier suggested that the split between Acorales and other monocots could be dated to (147-)134(-121) million years before present, and this date was also used in a more recent and comprehensive analysis that formed the basis for dating the age of monocot groups in general (Janssen & Bremer 2004).

Caterpillars of skipper butterflies of the Castniidae are found on a variety of monocots (Forbes 1956; see Powell et al. 1999 for some other groups that prefer monocots). Larvae of the chrysomelid beetle group Galerucinae (ca 10,000 species) subribe Diabroticites are apparently quite common on monocots, where they feed on roots (Eben 1999), indeed, the sister group of Galerucinae, the ca 6,000 species of Hispinae-Cassidinae, are the major group of beetles feeding on monocots (Jolivet & Hawkeswood 1995; Wilf et al. 2000). Wilf et al. (2000) thought that the initial monocot food of these beetles was aquatic members of Acorales and Alismatales, the Hispinae+Cassidinae-commelinid association being derived, but Gómez-Zurita et al. (2007) suggest that the two main clades of monocot-eating chrysomelid beetles are unrelated, and also that the chrysomelids diversified 86-63 million years ago, well after the origin of monocots. The idea has been floated that monocots experience less herbivory in tropical lowland rainforests than do other angiosperms, in part because they are tough and in part because the leaves remain rolled up for a relatively long time (Grubb et al. 2008); many monocots also have raphides as their main crystalline form of calcium oxalate, and these may be involved in herbivore defence (see Araceae below; Franceschi & Nakata 2005). Monocots are practically never ectomycorrhizal.

It has long been noted that many of the distinctive features of monocots might suggest that they originated from aquatic or hydrophilous ancestors (e.g. Henslow 1893 and references: the style of comparison and suggested mechanisms are interesting!); many members of the first two pectinations in the monocot tree, Acorales and Alismatales, are water or marsh plants or at least prefer to grow in damp conditions. The scattered vascular bundles in the stem, long linear leaves, absence of secondary thickening (cf. biomechanics of living in water), clusters of roots, rather than a single, branched tapwoot (nature of substrate), even the sympodial habit (for which, see Holttum 1955), etc., are all compatible with such an origin. However, even if monocot origin can be linked to life in some kind of aquatic habitat, it does not help much in our understanding of the details of how the distinctive monocot anatomical features, etc., evolved. Indeed, monocots are so different in many respects from other angiosperms that relating their morphology, anatomy and development to that of broad-leaved angiosperms (BLAs) is difficult (e.g. Zimmermann & Tomlinson 1972; Tomlinson 1995). Nymphaeales and Ceratophyllales are scarcely less remarkable in this respect, but the common ancestors of all these clades with broad-leaved angiosperms must have been plants with broad, petiolate leaves and a woody stem with lateral thickening meristems (cork and vascular cambiums).

In the past taxa that have petiolate, net veined leaves have been considered "primitive" in the monocots, linking monocots to "dicots" with similar leaves. Recent work suggests that such taxa are scattered through the monocots, as are taxa with fleshy fruits (excluding taxa with arillate, ant-dispersed seeds), both features being adaptations to shady conditions that have evolved together but independently (Dahlgren & Clifford 1982; Patterson & Givnish 2002; Givnish et al. 2005, 2006b). The taxa involved include Smilax, Trillium (Liliales), Dioscorea (Dioscoreales), etc. A number of these plants are vines that tend to live in shady habits for at least parts of their lives, and there may also be an association with unoriented stomata (see Cameron & Dickison 1998 for references for the latter feature). Indeed, net venation seems to have evolved at least 26 times in monocots, fleshy fruits 21 times (they have sometimes been lost subsequently); the two features, although independent, showed very strong signs of tending to be gained or lost in tandem, a phenomenon that Givnish et al. (2005, 2006b) describe as "concerted convergence".

For the sympodial growth habit of many monocots, see Holttum (1955). Most monocots form tufts of leaves in part of each growth cycle, and/or are geophytes; internode elongation in such cases is very slight. Given that secondary thickening in monocots is uncommon, yet many are plants of quite considerable stature, even tree-like, there must be considerable changes in the plant in the period between germination and the mature (flower-producing) stage, particularly in the size of the apical meristem. This period is often designated as establishment growth (e.g. Tomlinson & Esler 1972; Bell & Bryan 2008).

It is interesting that monosymmetric flowers in monocots are very frequently presented in the BLA position with the median sepal adaxial; the main exception are most Zingiberales. It is unclear why this should be so, although perhaps the abaxial tepal that acts as a landing platform is then partly supported by the two adjacent tepals of the outermost perianth whorl - in those Commelinaceae where the abaxial tepal is very small, perhaps the well developed inflorescence bract serves the same purpose - whereas if it were a member of the outer whorl, there would not be the same support. Indeed, floral orientation in monocots is quite variable, and i.a. depends on the presence and position of the prophyll/bracteole, and also on the existence of other structures on the pedicel (see e.g. Eichler 1875; Engler 1888; Remizova et al. 2006b). Stuetzel and Marx (2005) note the variability of the position of monocot bracteoles, which they suggest may be because what appear to be single axillary flowers in fact represent reduced racemes.

Monocots and "dicots" were often distinguished in the past by the 3-merous flowers of the former and the predominantly 5-merous flowers of the latter, even as it was realised that some of the "primitive dicots" might have more or less 3-merous flowers. With our current knowledge of phylogeny and floral development, it seems that a 3-merous perianth in particular is widespread and may even be a synapomorphy for a clade [[Chloranthaceae + magnoliids] [monocots [Ceratophyllaceae + eudicots]]] (Soltis et al. 2005b and literature cited). The two perianth whorls in monocots are similar and petaloid, being called tepals; however, there is usually a slight different between the members of the two whorls. The stamens are individually opposite members of the two whorls, indeed, even the outer whorl may not encircle the floral apex. All in all, the 3-merous flowers of the monocots are rather highly stereotyped, usually being pentacyclic: pentacyclic 3-merous flowers are at best extremely uncommon in broad-leaved angiosperms and are here considered to be an apomorphy for monocots (cf. Soltis et al. 2005b; Bateman et al. 2006b).

The development of monocot leaves needs much more study. A Vorläuferspitze, an often small abaxial conical or cylindrical protrusion at the apex of the mature leaf, is quite common in monocots. It represents the upper/distal part of the leaf, and in BLAs it develops into the blade. The petiole and blade of some monocot leaves is thought to develop from the lower/proximal part of the leaf primordium and to be equivalent to the leaf base of BLAs. However, in Acorales and Alismatales in particular the blade may develop from the upper part of the leaf primordium, i.e., they are similar in this to broad-leaved angiosperms. However, Scindapsus, but not Arisaema, Orontium, Zamioculcas, and even Acorus itself, seems to develop in a "typical" monocot fashion (Troll & Meyer 1955; Bharathan 1996; Doyle 1998b). Given the little that we know, the "typical" monocot leaf development might even be a synapomorphy of a subgroup of monocots, i.e. of monocots minus Acorales and Alismatales. Although Acorus has a "typical" linear monocot leaf, many Araceae do not, and scattered through the group are taxa with petioles and net-veined leaf blades of a variety of morphologies; these include Smilax, Trillium (Liliales), Dioscorea (Dioscoreales), Lowia (Zingiberales - such leaves are very common here), Stemona (Pandanales), etc. Smilax has paired tendrils near the base of the petiole, and sometimes paired ligules born either at the base (e.g. Potamogetonaceae) or top (e.g. Poaceae) of the petiole or sheath are scattered through the group. Truly compound leaves are rare (Zamioculcas is an example), but cell death may result in the leaves appearing to be compound (Arecaceae, a few Araceae) or having distinctive perforations (some Araceae and Aponogetonaceae).

For monocots, in addition to references in the notes on the Characters page and under individual orders and families, there is much interesting information in Arber (1920, 1925), Dahlgren et al. (1985) and Tillich (1998); Tomlinson (1970) outlined the morphology and anatomy of monocots, emphasizing the woody groups. Volumes III and IV of Families and Genera of Vascular Plants, edited and with useful outline classications by Kubitzki (especially 1998a, c), also contain a great deal of information (see also references to individual articles below). For ovule morphology, see Rudall (1997), dimorphism in the cells of the root epidermis and hypodermis, see Kauff et al. (2000), seedlings, see Tillich (2003b and references, 2007), the distribution of operculate pollen, see Furness and Rudall (2006b), nuclear DNA content, see Bharathan et al. (1994), and for discussion on the evolution of the berry, see Rasmussen et al. (2006).

ACORALES Reveal [Back to Index]

Inflorescence a spadix with spathe; flowers ebracteate; ovules straight; endosperm copious, perisperm [derived from nucellar epidermis] +, not starchy. - 1 family, 1 genus, 2-4 species.

ACORACEAE Martinov - Leaves equitant and isobifacial [oriented edge on to the stem]; peduncle with two separate vascular systems; flowers weakly monosymmetric; anther thecae with confluent dehiscence; fruit a berry. - 1/2-4. E. America to East and South East Asia; perhaps naturalised in Europe and America (Mayo et al. 1997).

Acoraceae are easily recognisable by their sweetly-smelling, two-ranked, isobifacial leaves and their densely spicate inflorescence overtopped by and appearing lateral to the leaf-like spathe. The flowers are small, perfect and pentacyclic.

The abaxial tepal is large, bract-like, and completely encloses the young flower; it looks as if it has "merged" with the bract (Buzgo 2001), perhaps being an organ of "hybrid" nature (Bateman et al. 2006b). The ovules are encased in mucilage secreted by the intra-ovarian trichomes.

For information, see Kaplan (1970: leaf development), Tillich (1985: seedling), Grayum (1987: general), Carlquist and Schneider (1997: anatomy), Bogner and Mayo (1998: general), Buzgo and Endress (2000) and Buzgo (2001: both floral morphology), Keating (2003a: anatomy), and Stockey (2006: evaluation of fossil remains). For a checklist and bibliography, see Govaerts and Frodin (2002) and World Checklist of Monocots.

[ALISMATALES [PETROSAVIALES [[DIOSCOREALES + PANDANALES] [LILIALES [ASPARAGALES + COMMELINIDS]]]]]: ethereal oils 0; calcium oxalate raphides + (druses 0); pollen boat-shaped, tectum reticulate with finer sculpture at the ends of the grain, endexine 0, (septal [epithelial] nectaries + [intercarpellary fusion postgenital]).

Although raphides occurring in bundles and largely filling the cells containing them are common in this clade, druses may at least sometimes be found along with them (e.g. Prychid et al. 2008).

ALISMATALES Dumortier [Back to Index]

Anthers extrorse, tapetum amoeboid, carpels with completely unfused canals, styluli present; endosperm helobial; embryo large; seedling with hypocotyl and primary root relatively well developed. - 14 families, 166 genera, 4490 species.

Stem-group Alismatales are dated to ca 131 million years before present, crown group Alismatales to ca 128 million years before present (Janssen & Bremer 2004; ca 133 and 103 million years before present respectively in Bremer 2000b); the oldest fossils assignable to this clade are recently discovered and from the Early Cretaceous, some 110-120 million years old (Araceae-Pothoideae-Monstereae: Friis et al. 2004, see Stockey 2006 for a review of the fossils that have been placed in Alismatales).

Alismatales include the only marine angiosperms and the only monocots with green embryos. Cox (1988: review), Cox et al. (1991: computer simulation of underwater pollination), Cox and Humphries (1993: Cymodoceaceae) and Les et al. (1997: phylogeny and hydrophily) discuss the remarkable pollination devices of water-pollinated Alismatales, both marine and freshwater. These include staminate flowers that detach from the parent plant and rise to the surface, the flowers themselves then floating on the water and transporting pollen to the stigma, pollen variously aggregated and forming masses especially on the water surface, and underwater pollination where the pollen is sometimes very much elongated or forms elongated strands, so increasing the chanes of pollination. These adaptations associated are so striking that the flowers and inflorescences in particular, but also the vegetative bodies, of the plants appear very different both from one another and from Alismatales with more conventional life styles. The result is that many small families are recognised below, although even more families have been recognised in the past. Sulphated phenolic compounds are common in seagrasses, including members of Hydrocharitaceae (McMillan et al. 1980), and probably arose in parallel; their function is unclear, although they are probably involved in adaptation to life in the marine habitat. The distribution ranges of a number of the hydrophytic taxa are rather young (Les et al. 2003).

The apparent absence of mycorrhizae in Alismatales may be connected with the prevalence of the aquatic habitat that many of their members members prefer; mycorrhizae are usually absent in such situations. Caterpillars of Pyralidae-Schoenobiinae are found on aquatic monocots, as are larvae of Chrysomelidae-Donaciinae (especially Kölsch & Pedersen 2008); members of the latter group, at least, are found on aquatic plants in general and so are found on Nymphaeaceae and Haloragaceae, etc.

For seedling morphology, see Tillich (1985), for pollen, see Grayum (1992), for ovules, see Igersheim et al. (2001), and for carpel evolution, see Chen et al. (2004a); Den Hartog (1970) and Green and Short (2003) provide comprehensive taxonomic and ecological accounts of marine angiosperms, along with distribution maps, etc.

Includes Alismataceae, Aponogetonaceae, Araceae, Butomaceae, Cymodoceaceae, Hydrocharitaceae, Juncaginaceae, Posidoniaceae, Ruppiaceae, Potamogetonaceae, Scheuchzeriaceae, Tofieldiaceae, Zosteraceae.

ARACEAE Jussieu - Inflorescence terminal, densely spicate [spadix], subtended by an inflorescence bract [spathe] +, floral bracts 0; pollen endexine spongy; fruit a berry. - 106/4025. Mostly tropical.

Gymnostachydoideae + Orontioideae: ovules straight.

1. Gymnostachydoideae Bogner & Nicolson - Leaves two-ranked, margins minutely toothed; inflorescence complex, branched, spathe 0; nucellus not fully covered by integument. - 1/1: Gymnostachys anceps. E. Australia.

2. Orontioideae Mayo, Bogner & Boyce - 3/6. Temperate East Asia, W. and E. North America.

Lemnoideae [[Pothoideae + Monsteroideae] [Lasioideae + Zamioculcadoideae + Aroideae]]: ?

3. Lemnoideae Engler - Thallus-like, stemless, floating aquatic herbs; collenchyma and bundle fibres 0; vessels 0; P 0; A 1 or 2, pollen ulcerate, spiny; G 1, embryo sac bisporic, 8-celled, stigma funneliform; fruit an achene [sort of]; seed operculate; endosperm cellular, starchy, embryo undifferentiated. - 5/37. World-wide.

For cytology, see Urbanska-Worytkiewicz (1980), for morphology, see Landolt (1986, 1998), for chemistry, etc., see Landolt and Kandeler (1987), for phylogeny in particular see Les et al. (2002) and Rothwell et al. (2004), and for speciation, see Crawford et al. (2006).

[Pothoideae + Monsteroideae] [Lasioideae + Zamioculcadoideae + Aroideae]: petiole clearly differentiated from sheathing base.

Pothoideae + Monsteroideae: stem usu. aerial; styloids +; pseudopetiole apically geniculate; crystals often surrounding the embryo.

4. Pothoideae Engler - Secondary and tertiary veins reticulate; spathe not enclosing spadix; anther thecae often forming tip above slit; spathe persistent in fruit; endosperm with starch. - 4/900: Anthurium (825, notably variable in leaf morphology), Pothos (70). Tropical America, Madagascar to South and Southeast Asia, Malesia and N.E. Australia.

5. Monsteroideae - P often 0, pollen inaperturate, extended monosulcate or zonate - 12/360: Rhaphidophora (100: paraphyletic), Rhodospatha (75). Tropical South and Southeast Asia to the Pacific, South America, few in Africa.

Lasioideae + Zamioculcadoideae + Aroideae: ?

6. Lasioideae Engler - Inflorescence flowering basipetally; anthers with oblique pore-like slits, pollen grains lacking starch, sulcus with ectexine lamella and thick bilayered endexine [outer: flakes or lamellae; inner: spongy]. - 10/58. India and Southeast Asia to the Pacific, also Africa (Lasimorpha).

Zamioculcadoideae + Aroideae: plants monoecious (dioecious); spathe differentiated into tube plus blade, pollen extruded in strands.

7. Zamioculcadoideae Bogner & Hesse - Biforine raphides 0; (leaves compound, petiole geniculate); P +; staminate flowers: (pollen zonasulcate, columellae winding, forming a sort of internal tectum as well as the external tectum, endexine lamellate, intine thin). - 3/21. Africa.

8. Aroideae - Laticifers +; P 0; A connate, pollen inaperturate, ektexine thin, often lacking sporopollenin, endexine thick, spongy, intine massive. - 68/2650: Philodendron (500), Arisaema (170), Amorphophallus (150), Schismatoglottis (120), Homalomena (110), Alocasia (70), Xanthosoma (60), Cryptocoryne (50). Tropical and warm temperate (the latter - Arum and relatives), Calla alone in more northern Eurasia and northern North Ammerica.

Araceae are herbs with leaves that are usually divided into a petiole and expanded blade. They have a distinctive inflorescence of a more or less petaloid spathe plus a spadix made up of densely-packed, sessile, ebracteate flowers - in many taxa the spadix has a large, terminal, sterile part. The fruit is a berry.

Distinctive pollen that can be assigned to Pothooideae-Monstereae has been found in Early Cretaceous deposits of the late Barremian-early Aptian of some 110-120 million years old in Portugal (Friis et al. 2004); other pollen types that may also be Araceae were found at the same place (see also Hesse & Zetter 2007). The site is now, alas, developed. Stem-group Araceae have been dated to ca 131 million years before present, crown group Araceae to ca 128 million years before present (Janssen & Bremer 2004).

Many Araceae are plants of shaded conditions, and net veined leaves and fleshy fruits are associated with this habitat here and in other monocots (Givnish et al. 2005). Climbers and epiphytes are notably common in Pothoideae and Monsteroideae, particularly in Pothoideae (Pothos and relatives have monopodial shoots, distichous leaves, and flattened petioles), and seedlings of some species grow towards the dark, i.e., towards a tree up which they then climb (Strong & Ray 1975). Although many Araceae prefer damp or even marshy habitats, free-floating aquatics are represented in extant Araceae by two clades only: Lemnoideae, some of which have arguably the most reduced and modified plant body of any free-living angiosperm, and Pistia stratioides, the water lettuce, an unrelated clade in Aroideae. The vegetative body of Lemnoideae is variously interpreted as being some combination of leaf and shoot, while the reproductive parts either represent a reduced but perfect flower or a very highly reduced inflorescence. Wolffia and Wolffiella lack both roots and veins in the thallus, and the thallus of the former may be less than 2.5 mm across, the smallest flowering plant known (see Lemon & Posluszny 2000b for shoot development in Lemnoideae). Pistia has a much less unconventional plant body, and its inflorescence, although reduced, is basically similar to that of other Aroideae; vegetative shoots are monopodial (Lemon & Posluszny 2000a). Both Lemnoideae and Pistia have supernumerary axillary buds which increases the complexity of their branching patterns. Although the early Tertiary fossil Limnobiophyllum seems "intermediate" between these two groups, it is to be assigned to Aroideae (Stockey et al. 1997); there are also other unrelated fossil floating aquatics in Araceae (Stockey et al. 2007).

Araceae are not much liked by butterfly caterpillars (Ehrlich & Raven 1964). A number of species of galerucine beetles (Aplosonyx) have been found feeding on laticiferous Aroideae from South East Asia where they make circular tranches in the leaves to interrup the latex flow and then eat out the portion of the leaf so isolated - it looks as if there are paper punch holes in the blade (Darling 2007); galerucines are known from other monocots and beetle herbivory in Araceae may be geographically more widespread.

Thermogenesis has been detected in the flowers of a number of Araceae, and this is produced both by uncoupling proteins and the mediation of an alternative oxidase, the net result being that heat rather than energy in the form of adenosine triphosphate is produced by glycolysis (Watling et al. 2006; Onda et al. 2008 and references). The heat may volatilize compounds that attact pollinators, and/or provide a warm roost for them inside the spathe. A nectar-like but sometimes foul-tasting exudate may be produced by stigmatic hairs, etc., as in Anthurium (Daumann 1931); it attracts pollinators. Croat (1980) discussed pollination in this speciose neotropical genus. The spathe of Aroideae is usually differentiated into tubular and blade-like portions. The fertile flowers are restricted to the basal part of the spadix, hence being more or less enclosed by the sometimes inflated tubular portion of the spathe. Pollinators, attracted by the color of the blade, or the smell, or even the dangling apical portion of the spadix (e.g. some Arisaema) may be temporarily trapped inside the tubular portion by hairs, etc.; they are released when the staminate flowers open and they get covered with pollen. More or less unpleasant odours (to us), are common in Araceae, as is evident from common names like the marvelously-named dead horse arum (Helicodiceros [Dracunculus] muscivorus), and pollination by flies and beetles is common. Some neotropical Araceae are pollinated by euglossine bees (orchid bees; see Roubik & Hanson 2004). In a number of Aroideae the pollen is extruded from the anthers in almost toothpaste-like threads. The arborescent South American Montrichardia Aroideae) has "explosive" pollen; the massively thick intine swells to an elongated structure ca 400 µm long within a few seconds, perhaps aiding its attachment to the pollinator, a hairless dynastid beetle (Weber & Halbritter 2007). For a discussion on the evolution of the distinctive pollen that characterises most Aroideae, see Hesse (2006b). In Monsteroideae there are trichosclereids in the stylar tissue and the spathe is deciduous; the trichosclereids may protect the exposed ovary against insects.

Raphides in those taxa that have been studied are twinned calcium oxalate crystals, H-shaped in transverse section, and often with lateral barbs (Sakai & Hanson 1974); Lemna, etc., also have such raphides. Raphides develop earlier than druses, at least in Amorphophallus (Aroideae), and may help protect young tissue, as well as aiding in regulating calcium (Prychid et al. 2008). Distinctive biforine raphides are found in a number of Aroideae. Here the cell wall is thick, except for papillae at the two ends, and lignified, and the cell contents are mucilaginous. When the papillae break, the raphides are ejected with some force, and can perforate epithelial tissue of the throat if leaves, etc., are eaten by humans (as with dumb cane, Dieffenbachia, although a poison may also be involved).

Vegetative variation in Araceae is considerable. In addition to Gymnostachys, I have seen one taxon (unnamed, from Thailand) with softly dentate/spinulate lamina margins. A number of taxa have fenestrate or apparently compound leaves produced by localised cell death. The leaves of Monstera, the swiss cheese plant, are fenestrate (see Melville and Wrigley 1969), while compound-leaved taxa include Zamioculcas and Dracontium. Leaves of Zamioculcas appear to be truly compound, with localised development of the blastozone, the marginal leaf meristem, being responsible for the growth of the individual leaflets, while in Dracontium localised cell death results in what is clearly an initially simple leaf blade with entire margins separating into a complex structure with numerous "leaflets". Such leaves may be huge, thus in Dracontium gigas the dissected foliar part is up to 4m in diameter and is born on a petiole ca 5 m tall (Bown 2000). Pulvini occur on the petiole of taxa like Dracontium. Other details of leaf development seem to be quite variable in Araceae. Scindapsus, but not Arisaema, Orontium, and Zamioculcas, develops in a "typical" monocot fashion (Troll & Meyer 1955; Bharathan 1996; Doyle 1998b); in the latter genera the blade develops from the upper part of the leaf primordium, i.e., they are similar in this to broad-leaved angiosperms. The leaves of Anthurium are notably variable, being entire to deeply lobed or even compound. The leaves of climbers are often strongly heteroblastic, the leaves of a plant in the climbing phase being simpler than when it becomes reproductive.

The highly reduced vegetative body of Lemnoideae is variously interpreted as being some combination of leaf and shoot, while the reproductive parts either represent a reduced but perfect flower or a very highly reduced inflorescence; given the phylogenetic position of Lemnoideae, the former is perhaps more likely. Wolffia and Wolffiella lack both roots and veins in the thallus, and the thallus of the former may be less than 2.5 mm across; it is the smallest flowering plant known (see Lemon & Posluszny 2000b for shoot development in Lemnoideae). The aquatic Pistia has a much less unconventional plant body, and its inflorescence, although reduced, is basically similar to that of other Aroideae; vegetative shoots are monopodial (Lemon & Posluszny 2000a). Both Lemnoideae and Pistia have supernumerary axillary buds which increases the complexity of their branching patterns. Although the early Tertiary fossil Limnobiophyllum seems "intermediate" between Lemnoideae and Pistia (Stockey et al. 1997), those two groups are not at all close in molecular phylogenies, and the fossil is to be assigned to Aroideae like Pistia itself.

The uninucleate chalazal endosperm haustorium of Arum maculatum is reported to be 24,576 n (Werker 1997).

Early hypotheses of phylogeny based on restriction site analysis (French et al. 1995) suggested rather pectinate relationships in the family, but a consensus tree of morphological characters (Mayo et al. 1997) and a more recent tree based on the analysis of five plastid genes (Mayo et al. 2003) show somewhat less resolution. However, the clade [Gymostachydoideae + Orontioideae] remains as sister to all the rest of the family, and Lemnoideae are strongly supported as sister to the remainder; the groupings are largely those also evident in French et al. (1995). Tam et al. (2004: trnL-F sequences, Calla not examined) again suggest the phylogeny is rather pectinate, as do Cabrera et al. (2008: five chloroplast genes). The topology the latter present, quite well supported, is that given here; the exact position of Zamioculcadoideae is still uncertain, but it can reasonably be excluded from Aroideae. Gonçalves et al. (2007) discuss the phylogeny of the Andean Spathicarpeae, a clade in which the spathe is adnate to the spadix, many of which grow in very dry and/or high conditions, and Cabrera et al. (2008) offer a number of suggestions of tribal relationships in Aroideae. For a phylogeny of the large neotropical genus Philodendron, see Gauthier et al. 2008; Homalomena may be part of this clade. Calla palustris, in early (prior to Ed. 8) versions of this site placed in a separate subfamily, seems best included in Aroideae, with which it agrees in lacking a perianth but differs in having biporate pollen and perfect flowers. It is placed in a clade of Aroideae along with other rooted aquatics/marsh plants in molecular studies (see Cabrera et al. 2008), and it is possible its perfect flowers represent a reversal. It has a notably more northern distribution that all other members of the clade (see map on main page).

There is much information in Mayo et al. (1997, 1998); see also Ertl (1932: venation and petiole anatomy, more normal monocot venation may be common in the basal subfamilial pectinations), Ray (1988: shoot organisation), Tillich (1985, 2003b: seedlings, great variation), Grayum (1991, 1992: pollen morphology), Seubert (1993: starch grains, seeds and seedlings), Dring et al. (1995: chemistry), Behnke (1995a: sieve tube plastids), French (1998: stem anatomy), Buzgo and Endress (1999: Gymnostachys), Weber et al. (1999: pollen), Bown (2000: general, an excellent account), Keating (2000b: collenchyma, 2003a: general anatomy, b: leaf anatomy, 2004 a: classification, b: raphides), Gonçalves et al. (2004: collenchyma), Hesse (2006a, b: summary of pollen variation), and Bogner and Petersen (2007: chromosome numbers), Barabé and Lacroix (2008: development of Anthurium), and Tobe and Kadokawa (2008: good summary of embryological variation). For a checklist and bibliography, see Govaerts and Frodin (2002) and the World Checklist of Monocots, and for several keys and lots more, see CATE-Araceae.

Tofieldiaceae [Hydrocharitaceae + Scheuchzeriaceae groups]: carpels free.

TOFIELDIACEAE Takhtajan - Stomata anomocytic; leaves equitant and isobifacial [oriented edge on to the stem]; flowers subtended by a "calyculus"; A introrse to latrorse, microsporogenesis simultaneous, pollen disulcate, nectaries +, ovules unitegmic; seeds with appendage(s). - 3-5/27. S.E. U.S.A., N.W. South America, N. temperate.

Tofieldiaceae can be recognised by their equitant, isobifacial leaves; racemose inflorescences in which there is usually a 1-3-membered "calyculus" below the flower; and undistinguished monocot flowers with the tepals usually free and the tips of the carpels each with a clearly separate stylulus. The fruit is usually a septicidal capsule.

Branching in Tofieldiaceae needs study. Thus Remizowa et al. (2005) suggest that the first two leaves of axillary shoots in Tofieldia (the prophyll and also the next leaf) are adaxial; this would be a very unusual arrangement, if true. Generalised comparisons between the calyculus of Tofieldiaceae, made up of two or three connate "scales", with the spathe of Hydrocharitaceae and pseudowhorls of bracts in Alismataceae have been made (Remizowa and Sokoloff 2003; Remizowa et al. 2