LIGNOPHYTA

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

EXTANT SEED PLANTS/SPERMATOPHYTA

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

MAGNOLIOPHYTA

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

Evolution. Possible apomorphies for flowering plants are in bold. Note that the actual level to which many of these features, particularly the more cryptic ones, should be assigned is unclear. This is because some taxa basal to the [magnoliid + monocot + eudicot] group have been surprisingly little studied, there is considerable homoplasy as well as variation within and between families of the ANITA grade in particular for several of these characters, and also because details of relationships among gymnosperms will affect the level at which some of these characters are pegged. For example, if reticulate-perforate pollen is optimized to the next node on the tree (see Friis et al. 2009 for a discussion), it effectively makes the pollen morphology of the common ancestor of all angiosperms ambiguous... For other features such as details of sugar transport in the phloem, their placement on the tree is frankly speculative. Finally, for features such as parietal tissue/a nucellus only one (Nymphaeales) to three cells thick above the embryo sac and a stylar canal lacking an epidermal layer, although plesiomorphous for basal grade angiosperms (Williams 2009), I am unsure where on the tree a thicker nucellus and a stylar epidermal layer are acquired.

NYMPHAEALES [AUSTROBAILEYALES [[CHLORANTHALES + MAGNOLIIDS] [MONOCOTS [CERATOPHYLLALES + EUDICOTS]]]]: vessels +, elements with elongated scalariform perforation plates; wood fibres +; axial parenchyma diffuse or diffuse-in-aggregates; tectum reticulate-perforate [here?]; ?genome duplication; "DEAER" motif in AP3 and PI genes lost, gaps in these genes.

[ROSALES [CUCURBITALES + FAGALES]]: placentation apical; ovules 1-2/carpel.

ROSALES Perleb  Main Tree, Synapomorphies.

(Frankia infection via intercellular penetration); (isoflavonoids, dihydroflavonols +); mucilage cells +; roots diarch [lateral roots 4-ranked]; prismatic crystals in ray cells [not Barbeyaceae, Elaeagnaceae]; (sieve element with non-dispersive protein bodies; sieve element plastids lacking starch [Rhamnaceae, Dirachmaceae?]); lamina margins with teeth; inflorescence cymose; hypanthium +, nectariferous, K valvate, C clawed; styles +, stigma dry; ovule 1/carpel, apotropous, micropyle endostomal; K and/or hypanthium persistent in fruit; (polyembryony +). - 9 families, 261 genera, 7725 species.

Evolution. Divergence & Distribution. Wikström et al. (2001: relationships are [Fabales [Rosales [Cucurbitales + Fagales]]]) date the origin of Rosales to (90-)88(-86) million years before present, diversification beginning (79-)76(-73) million years before present. Most other ages are rather older. The age of crown group Rosales was estimated as (96-)93, 88(-85) million years (two penalized likelihood dates), the stem group age being (110-)105(-100) or (94-)89(-84) million years; Bayesian relaxed clock estimates were slightly older, to 103 or 111 million years respectively (Wang et al. 2009: note that relationships are [Rosales [Fabales [Cucurbitales + Fagales]]]), while Magallón and Castillo (2009: relationships are [Fabales + Rosales] [Cucurbitales + Fagales]) estimated ages of ca 101 and 101.3 million years for relaxed and constrained penalized likelihood datings for the age of stem Rosales, and ages of 93.9 and 94.1 million years for the crown group (relaxed and constrained again). Fossils are known from the Middle Eocene, ca 44 million years before present.

Rosales contain ca 1.9% of eudicot diversity (Magallón et al. 1999).

Plant-Animal Interactions. Quite a number of butterfly larvae - especially caterpillars of "basal" groups and Lycaeninae - feed here (Fiedler 1995; Janz & Nylin 1998).

Bacterial/Fungal Associations. Some taxa in at least Rosaceae, Rhamnaceae, Elaeagnaceae and Ulmaceae are ectomycorrhizal (see Malloch et al. 1980; Smith and Read 1997).

Ronse De Craene (2003, see also 2010) suggested that loss of petals may characterise Rosales, with apparent "petals" occupying the position of stamens and their evolution allowing e.g. Rosaceae to diversify. Comparing the vasculature of petals and stamens may bear on this idea, but whether or not it has anything to do with diversification is a separate issue. Indeed, if Rosales are sister to Fabales, they would not seem to be a notably diverse group in terms of species numbers, the more so since almost 4,000 species of Rosales are in the Ulmaceae-Urticaceae group, which lack petals of any sort. These taxa form a single clade, and the wind pollination that is also so common here cannot be considered basic to the order (cf. Ronse de Craene 2010).

Chemistry, Morphology, etc. Roots are commonly diarch in Rosaceae, but are also tetrarch, etc.; sampling elsewhere is poor, although less so in Ulmaceae and relatives, and diarch roots seem to be found throughout the order. Tracheary members in Rosaceae commonly have pseudotori (thickenings in pit membranes associated with plasmodesmata), while true tori occur in Rosaceae (Cercocarpus) and also Cannabaceae and Ulmaceae (Jansen et al. 2007), although note that these tori are formed in two different ways (Dute et al. 2010a). Sieve tube plastids lacking both starch and protein inclusions are rare outside Rosales, although they occur in some parasites as well as Crassulaceae and Malpighiaceae (Behnke 1991a). For wood anatomy, esp. of Elaeagnaceae, see Jansen et al. (2000b) and Baas et al. (2001: fibre pits not bordered). A granular layer below the tectum may be a synapomorphy for the clade. Kubitzki (2004) provides a summary of the order; there is much useful information in Thulin et al. (1998).

Phylogeny. Relationships within the order were for some time unclear, although Rosaceae appeared to be sister to the rest of the order (strong support: Savolainen et al. 2000a; Wang et al. 2009), and Ulmaceae and relatives (the old Urticales) and Rhamnaceae and relatives may form two more clades (also Thulin et al. 1998; Savolainen et al. 2000b; Richardson et al. 2000b; Sytsma et al. 2002 [position of Rosaceae, etc. uncertain]; Wang et al. 2009; Soltis et al. 2011), as in the tree below. However, Richardson et al. (2000b, successive approximation weighting) found that the position of Elaeagnaceae was rather labile, even being embedded in Rhamnaceae in a rbcL analysis, and there was quite strong support for a clade [Barbeyaceae + Dirachmaceae]. The tree found by S.-D. Zhang et al. (2011) is based on an analysis of twelve genes from 25 taxa of Rosales, and the major relationships within the order are very well supported, although somewhat less so in the subclade that includes Rhamnaceae (see below); their relationships are followed here.

There is also the issue of the placement of Cynomoriaceae. Z.-H. Zhang et al. (2009; see also Moore et al. 2011) placed the holoparasitic Cynomoriaceae in Rosales sister to Rosaceae based on analysis of chloroplast inverted repeat sequences (Moraceae were the only other family in the order examined), and with strong support; Cynomoriaceae certainly were to be excluded from Saxifragales (good sampling) where they had sometimes been included, albeit with modest support (see Saxifragales page for further details). S.-D. Zhang et al. (2011) did not include Cynomoriaceae in their study; 10/12 genes they sequenced were plastid genes. Cynomoriaceae are placed both here and in Saxifragales for now. Although there is little in their morphology to suggest relationships with Rosales, the floral morphology and development of Cynomoriaceae would repay careful examination.

Previous Relationships. In the past, Urticales (Urticaceae, Moraceae, etc.) were kept well separate from Rosaceae, largely because the very reduced and usually wind-pollinated flowers of Urticales suggested relationships to the old Amentiferae, and the other families now included in Rosales were usually placed elsewhere yet again.

Includes Barbeyaceae, Cannabaceae, Cynomoriaceae, Dirachmaceae, Elaeagnaceae, Moraceae, Rhamnaceae, Rosaceae, Ulmaceae, Urticaceae.

Synonymy: Rhamnineae Shipunov - Amygdalales Link, Artocarpales Martius, Barbeyales Takhtajan & Reveal, Cannabales Döll, Dryadales Link, Elaeagnales Berchtold & J. Presl, Ficales Dumortier, Frangulales Wirtgen, Morales Martius, Rhamnales Link, Sanguisorbales Link, Spiraeales Link, Ulmales Link, Urticales Berchtold & J. Presl

ROSACEAE Jussieu, nom. cons.   Back to Rosales

Triterpenes +, alkaloids 0; cork deep seated; (vessel elements with scalariform perforation plates); (true) and fibre tracheids +; sieve tubes with non-dispersive protein bodies; petiole vasculature of arcuate or annular bundles, or annular; leaves spiral (opposite), lamina vernation usu. conduplicate, (2ndary veins palmate), stipules also often petiolar (0); inflorescences racemose; (C 0); A (1-)15-many [ca 20 common - 10 + 5 + 5, centripetal, in groups], (latrorse); (pollen porate); G 1-5, free, stigmas punctate to expanded or down stylulus; ovule 1/carpel, epitropous, parietal tissue 2-4 cells across, nucellar cap ca 4 cells across; megaspore mother cells several; fruit aggregate of achenes; exotestal cells periclinally elongated, radial walls thickened, or palisade or tabular, walls with spiral or reticulate thickenings, outer wall often becoming mucilaginous, (mesotesta sclerotic), endotegmic cells slightly thickened, or seed coat undistinguished; chalazal endosperm haustorium +; x = 9; duplication of GBSSI [granule bound starch synthase I] gene.

90[list]/2520. World-wide, but esp. N. hemisphere (map: from Vester 1940; Hultén 1971). [Photos - [Collection, Collection.]

1. Rosoideae Arnott

Herbs to shrubs; 2-pyrone-4,6dicarboxylic acid, ellagic acid +; rays often narrow; cuticle waxes as narrow ribbons and triangular rodlets; leaves compound; (epicalyx +), carpels usu. many; ovule (straight), unitegmic; fruits achene or drupelet; x = 7; plant with phragmidiaceous rusts.

Especially temperate (to Arctic) areas.

1A. Filipendula - Plant herbaceous; receptacle enlarged; ovules 2/carpel. - 1/10. Eurasia.

Synonymy: Ulmariaceae Gray

Rosodeae T. Eriksson, Smedmark, & M. S. Kerr (= all other Rosoideae)

1B. Rubus - Prickly scrambling shrub; receptacle enlarged; ovules 2/carpel [Rubus], integument ca 6 cells across; fruit an aggregate of drupelets. - 1/± 250. ± Worldwide, esp. N. temperate.

Synonymy: Chamaemoraceae Lilja

1C. Colurieae Rydberg

Ovule apotropous [Geum].

3/42: Geum (40: Kajewski 1957 for classic cytological work; Smedmark & Eriksson 2006 for development of the stylar hook). Temperate, inc. montane tropics, Chile.

[Sanguisorbeae [Rosa + Potentilleae]]: ?

1D. Sanguisorbeae Candolle

G 1-5; integument 6-8 cells across; phragmidiaceous rusts 0.

Within Sanguisorbeae are two subtribes. Agrimoniinae J. Presl - 5/20: Agrimonia (15). N. Temperate, Africa. Sanguisorbineae Torrey & A. Gray - 7/360: Cliffortia (115), Acaena (100). ± Worldwide, few Indo-Malesia, tropical America.

Synonymy: Agrimoniaceae Gray[?], Fragariaceae Nestler, Poteriaceae Rafinesque, Sanguisorbaceae Berchtold & J. Presl

[Rosa + Potentilleae]: ?

1E. Rosa - Prickly arching shrub; hypanthium fleshy, urn-shaped; integument ca 8 cells across. - 1/100-150: see Bruneau et al. (2007), Wissemann and Cox (2007) and Koopman et al. (2008) for phylogenies, relationships not easy to disentangle. N. temperate; 1/3^rd spp. in Europe.

1F. Potentilleae Sweet

(Epicalyx +); receptacle enlarged; integument ca 4 cells across. Potentillineae: style often lateral/gynobasic. 5-6/540: Potentilla + Argentina et al. N. temperate to Arctic (montane tropics to S. temperate). Fragariinae + Alchemillinae: anther thecae more or less confluent; parietal tissue ca 2 cells across, nucellar cap ca 7 cells across. Fragariinae Torrey & A. Gray: (leaves simple); (G 1); phragmidiaceous rusts 0 (Fragaria +). 10/60. Alchemillinae 3/960-1100: Alchemilla (870[-1000+]), Lachemilla (80). N. temperate, esp. Europe, tropical mountains (S. temperate).

Synonymy: Alchemillaceae Martinov, Potentillaceae Berchtold & J. Presl, Tormentillaceae Martynov

[Dryadoideae + Spiraeoideae]: sugar alcohol sorbitol as transport carbohydrate, cyanogenic glycosides +.

2. Dryadoideae Juel

Association with N-fixing Frankia; G 1-many; ovules straight (anatropous, apotropous - Dryas); fruits achenes with hairy styles.

4/19: Cercocarpus (8). W. North America, Dryas circumboreal.

Synonymy: Cercocarpaceae J. Agardh, Dryadaceae Gray

3. Spiraeoideae C. Agardh

Plant woody; flavones +, ellagic acid 0; cuticle waxes as tubules or platelets; G <5, opposite petals, 2< ovules/carpel, papillate funicular obturator +, stigma usu. wet; fruit a follicle.

3A. Lyonothamnus - Cyanogenic glycosides 0; leaves opposite, compound, stipules deciduous; G seminferior, placentation apical; ovules 4-6/carpel. - 1/1: Lyonothamnus floribundus. California Islands, off S. California.

3B. Niellieae Maximowicz

Cyanogenic glycosides?; ovule 1/carpel, apical, apotropous (-5, pleurotropous); fruitlets hard, shiny.

2/24: Niellia (14). E. and W. North America.

Synonymy: Neilliaceae Miquel

3C. Amygdaleae Jussieu

Plant ectomycorrhizal; cork superficial; true tracheids 0; lamina vernation laterally or vertically conduplicate, extrafloral nectaries on petiole or abaxial lamina; G 1; outer integument 6-8 cells across, inner integument 3-6 cells across, obturator from ovary wall; fruit a drupelet; seed coat mostly pachychalazal; n = 8.

1/200. Temperate and tropical montane.

Synonymy: Amygdalaceae Marquand, Prunaceae Martinov

Kerriodae D. Potter, S. H. Oh, & K. R. Robertson [= Osmaronieae + Kerrieae]: phragmidiaceous rusts 0.

3D. Osmaronieae Rydberg

Cork superficial; pith chambered; stipules deciduous; styles lateral; obturator from ovary wall; fruit a drupe, or septicidal capsule, the carpels also opening adaxially [Exochorda]; n = 8.

3/9: Exochorda (4), Prinsepia (4). Central to East Asia, W. North America.

3E. Kerrieae Focke

Wart-like projections on lamina; G 1-5; ovule ?obturator; fruit an aggregate, nut-like units, (achenes: Neviusia); n = ?

4/4. East Asia, W. North America, Alabama.

Synonymy: Coleogynaceae J. Agardh, Rhodotypaceae J. Agardh

3F. Sorbarieae Rydberg

Leaves compound (simple: Adenostoma); (G 1, Adenostoma), placentation apical; (fruit an achene: Adenostoma); phragmidiaceous rusts 0.

4/8: Spiraeanthus (4). Central to East Asia, W. North America

3G. Spiraeeae Candolle

Vestured pits +; nodes 1:1 [?all]; stipules 0; ovules 6-8/carpel, unitegmic; (fruit an achene - Holodiscus).

8/106: Spiraea (80-100). N. temperate, to Columbia, (S. and) E. Africa, West Malesia.

Synonymy: Spiraeaceae Bertuch

Pyrodeae C. S. Campbell, R. C. Evans, D. R. Morgan, & T. A. Dickinson

Plant ectomycorrhizal; flavone C-glycosides +; cork superficial [?Gillenia]; rays often narrow; colleters + [probably elsewhere]; G ± connate, adnate to base of hypanthium, opposite sepals or odd member abaxial, gynoecial ring primordium +; ovules two/carpel, ± apotropous, (micropyle bistomal); exotesta ± thickened, often mucilaginous, mesotesta thick, sclerotic; Gymnosporangium rust common.

3H. Gillenia - Leaves compound. - 1/2. E. North America.

3I. Pyreae Baillon

N = 17; four copies of GBSSI [granule bound starch synthase I].

33/ca 1000.

[Kageneckia + Lindleya]

(Plant dioecious); 4-many pleurotropous ovules/carpel; Gymnosporangium rust 0.

2/5. Mexico, Peru, Chile. [Kageneckia Flower, Fruit.]

Synonymy: Lindleyaceae J. Agardh

Vauquelinia - Tannin-containing cells pervasive; fruit septicidal, carpels opening adaxially (and partially abaxially as well). - 1/3. S.W. North America.

Pyrinae Dumortier

(Leaves compound), stipules deciduous; G at least half inferior, (carpels laterally free); outer integument 5-14 cells across, inner integument 3-6 cells across; hypanthium fleshy in fruit, (endocarp +).

30/1000: Sorbus (260), Crataegus (260, inc. Mespilus), Cotoneaster (260), Pyrus (75), Malus (55). Largely north Temperate.[Photo - Flower]

Synonymy: Cydoniaceae Schnizlein, Malaceae Small, nom. cons., Mespilaceae Schultz-Schultzenstein, Pyraceae Vest, Sorbaceae Brenner

• Rosaceae are a heterogeneous family, but they may be recognised by their often compound, stipulate and serrate leaves that are never(?) two-ranked and are only rarely opposite. Their flowers have a valvate calyx, always free and often clearly clawed petals, usually many stamens, a well-developed nectary on the hypanthium or at the base of the stamens, and often free carpels. The fruits are often achenes or drupelets, but the fruit in the broad sense is fleshy in a variety of ways, the hypanthium or receptacle also being involved; follicles are also quite common. Cyanogenesis is common (crush the leaves of Prunus, and smell), but there is more than one pathway involved.

Evolution. Divergence & Distribution. Turonian fossils from some 90 million years before present are assignable to Rosaceae (Crepet et al. 2004 for references); other estimates tend to give younger ages (stem group ca 76 million years before present, crown group divergence [Rosoideae not included] 47-46 million years before present - Wikström et al. 2001). Prunus endocarps have been found in the early Eocene of China ca 55 million years ago, and considerable diversification of the family occured in the Eocene (Wehr & Hopkins 1994; Benedict et al. 2011; Li et al. 2011 and references). The inferior-ovaried clade of Pyreae seems to represent a rapid but ancient radiation (Campbell et al. 2007).

Divergence & Distribution. N-fixing is known from Dryadoideae, but not all members of the subfamily can do this, thus the widespread Dryas integrifolia appears to be unable to fix nitrogen (Markham 2009); nodules and association with Frankia have also been reported from elsewhere in the family, e.g. from Rubus ellipticus (Markham 2009). Species of different genera of Rosaceae, one N-fixing (Cowania - Dryadoideae) and the other (Fallugia - Rosoideae-Colurieae) not, can form successful grafts, but when the non-N-fixing genus is the stock it will still not fix nitrogen (Kyle et al. 1986).

Floral Biology. There may be a period of about 19 days between pollination and fertilization in Prunus persica (Herrera & Arbeloa 1989).

Plant-Animal Interactions. Galls caused by cecidomyid midges are quite common in North American Rosaceae (Gagné 1989). Cynipid gall wasps of the tribe Diplolepidini are notably common on Rosa, Diplolepis rosae forming the well-known robin's pincushion or bedeguar gall (Csoka et al. 2005; Redfern 2011).

Bacterial/Fungal Associations. At least some woody Rosaceae are ectomycorrhizal. Geopora (Pezizales, an ascomycete) has recently been identified as the fungus involved in the mycorrhizae of Cercocarpus (Dryadoideae), and it also forms associations with Quercus and Arctostaphylos, the result being a complex mycorrhizal network; ascomycete fungi are associated with other ectomycorrhizal members of the family (McDonald et al. 2010).

Savile (1979b; see also Jackson 2004b for possible codivergence) discusses the distribution of phragmidiaceous rusts within Rosaceae-Rosoideae; they occcur on no other Rosaceae, and only rarely on plants from other families. Most of these rusts are autoecious, that is, their entire life cycle occurs on the same species of rosaceous host. In Gymnosporangium rusts the telial stage (the teliospore is a thick-walled resting spore that germinates to produce basidisospores) is common on some Cupressaceae, the aecial stage, which produces thinner-walled binucleate aeciospores, is found on Spiraeaoideae-Pyrodeae. Interestingly, Rosaceae rarely produce phytoalexins, protective compounds induced by e.g. fungal infection (Harborne 1999).

Genes & Genomes. The basic chromosome number of the family may be x = 7 (Shulaev et al. 2010). It has long been suspected that Pyreae were of wide hybrid origin, as their chromosome number might suggest (n = 9 [Rosoideae] x n = 7 [Spiraeaoideae] = n = 16 [some Maloideae]), see Evans et al. (1998), however, Evans and Campbell (2002) suggest that this is unlikely. Polyploidisation with subsequent aneuploidy (9 x 9 = 18, 18 - 1 = 17) of the diploid Gillenia (herbaceous, with compound leaves) or something similar is more likely; Gillenia is sister to Pyreae (Potter et al. 2002; Evans & Dickinson 2002). Note that Gillenia is host to the same rusts that are found on other Pyrodeae, but it has only two copies of GBSSI, as is the condition in the rest of the family.

For the relationship between polyploidy and diversification in Rosaceae - perhaps direct - see Vamosi and Dickinson (2006). Allopolyploidy in Fragariinae and its taxonomic implications were studied by Lundberg et al. (2009), while within Fragaria itself, the only extant taxon with one of the parental genomes of F. magellanica is known only from eastern Japan (Rousseau-Gueutin et al. 2009). Apomixis is quite common, as in Rubus and Alchemilla (Rosoideae) and Amelanchia and Crataegus (Spiraeaoideae-Pyrodeae). Hybridisation occurs in Crataegus, and the hybrids are triploid and apomictic; within Crataegus clades are linked with geography, and as in Fragaria, parents of genomes may have very disparate geographic origins (Lo et al. 2009). In general, apomixis seems to have preceded hybridisation (Dickinson et al. 2007).

Chemistry, Morphology, etc. A polyderm is common, although perhaps not occurring in Pyreae (Mylius 1913). 1:1 nodes occur in Spiraea, a genus that also lacks stipules, although normally Rosaceae have stipulate leaves and 3:3 nodes. This correlation between stipule presence/absence and nodal vasculature is fairly general in eudicots, and so it is interesting that it is evident even within Rosaceae, as was early noted by Sinnott and Bailey (1914). There are extensive data on cuticle waxes in the family (Fehrenbach & Barthlott 1988), but they are recorded only as summaries in the context of conventional subfamilies; Barthlott (pers. comm.) kindly provided a more detailed breakdown.

For hypanthium development, see Rauh and Reznik (1951). The epicalyx that occurs in some Rosaceae seems to represent stipules associated with the calyx members. Are petals staminodial in origin here (Ronse de Craene 2007)? Even in taxa with inferior ovaries, there is great variation in whether or not the carpels are connate, or what parts are connate, and in whether or not the carpels are adnate to the axial tissue enveloping the carpels/hypanthium. Thus Cotoneaster has an inferior ovary yet carpels that are more or less separate although adnate to the "hypanthium", cf. also Pyracantha. Indeed, the odd genus Dichotomanthes, also included in Pyreae, has a single carpel that is superior in position (Rohrer et al. 1994); its distinctive gynoecial morphology must represent a reversal. There are often five traces to each carpel. Ovules of Rhodotypos have a protruding nucellus (cf. Rhamnaceae). Chamaebatia (Dryadoideae) has a single carpel with a single, basal ovule that lacks an obturator (Evans & Dickinson 2002a). The lignified exotesta common in the family can be found even in the drupe of Prunus.

Some general information is taken from Robertson (1974), Kalkman (2004), Judd et al. (2002), and especially Potter et al. (2007), while there is information on ovule and seed morphology in Péchoutre (1902, quite comprehensive), petiole vasculature in Morvillez (1917), general chemistry in Hegnauer (1973, 1990), 2-pyrone-4,6dicarboxylic acid distribution in Wilkes and Glasl (2001), and tannins in Okuda et al. (1992), on cork initiation and bark anatomy in Lotova and Timonin (e.g. 1998, 1999, 2002) and also Weiss (1890), on wood anatomy in Zhang (1992), on carpel orientation and general morphology in Focke (1888), Schäppi (1951: peltate and plicate carpels), Schäppi and Steindl (1950), Sterling (1969 and references), Kania (1973: also androecium) and Weberling (1989), information on floral axis morphology is to be found in Rauh and Reznik (1951), androecial diversity is discussed in Murbeck (1941: also vasculature, Neurada etc. "basal") and Lindenhofer and Weber (2000 and references), pollen morphology in Sanguisorbeae in Chung et al. (2010), carpel development in van Heel (1981, 1983), ovule morphology in Rosoideae in Schaeppi and Steindl (1950), rust hosts in Savile (1979), exotesta in Frohne and Jensen (1992), and floral development in Evans and Dickinson (1999a, 1999b, 2002, 2005); for information on Vauquelinia, see Hess and Henrickson (1987).

Phylogeny. Although Rosoideae and a number of clades within it, Spiraeoideae, Spiraeoideae minus Lyonothamnus, Pyrodeae + Sorbarieae and Pyrodeae are all well-supported clades, little can yet be said of larger patterns of relationship in the rest of the family (e.g. Morgan et al. 1994; Potter et al. 2002; Potter 2003; Potter et al. 2007). The position of Dryadoideae is uncertain, other than being a rather "basal" branch in the tree (Potter et al. 2002: Evans et al. 2002), hence perhaps the lack of obturators. Potter (2003) found that Dryadoideae were fairly well supported as sister to remaining Rosaceae in a analysis using several genes, but their position was still not secure in Potter et al. (2007), although a sister group relationships with Spiraeaoideae is indeed perhaps most likely.

Eriksson et al. (2003) provide a phylogeny of Rosoideae; Alchelmilla, Fragaria, and other genera form a well-supported clade outside Potentilla and relatives (see also Dobes & Paule 2010). Filipendula is sister to other Rosoideae. "Intergeneric" hybridisation seems to occur (Smedmark et al. 2003). Within Potentilleae, Potentilla is sister to the rest (see Eriksson et al. 2003), or, more precisely, Potentilla (including a few genera) is sister to a small clade including Argentina (Potentilla sect. Anserina), the two forming a clade sister to the other Potentilleae (all relationships with strong support: Dobes & Paule 2010). Fragariinae are sister to Alchemillinae; for the relationships of Alchemilla, to include Aphanes, see Gehrke et al. (2008). The current infrageneric classification of Rubus does not reflect relationships there (Alice & Campbell 1999).

Aldasoro et al. (2005) suggest morphological and biogeographic relationships in the inferior-ovaried Spiraeaoideae-Pyreae. Exochorda forms a small clade along with Oemleria and Prinsepia (Evans & Dickinson 1999a for information); they have been placed near Prunus (Potter et al. 2002, see also Lee & Wen 2001) to which they do show some morphological similarity. Amygdaleae are circumscribed narrowly here (one genus, Prunus, from which there have previously been segregates), following Potter et al. (2002, 2006). For phylogenies of Prunus, see Lee and Wen (2001) and Wen et al. (2008: some conflict between ITS and ndhF); the distribution and nature of calcium oxalate crystals correlate quite well with relationships (see Lersten & Horner 2000). In the recent past Maddenia (dioecious, apetalous) has also been placed near Prunus. Generic limits in Pyreae-Pyrineae are difficult, there being little molecular divergence between many of them (but considerable divergence within them (Dickinson et al. 2007; Lo et al. 2007, esp. Crataegus s.l.). Generic limits are difficult, with divisions perhaps reflecting the European origin of taxonomy and the fact that Pyreae are common in Europe (e.g. Walters 1961). Furthermore, apomixis, common also elsewhere in Rosaceae, occurs in Sorbus, Crataegus, and Cotoneaster, at least, indeed, there were ca 17 North American species of Crataegus in 1896, while 30 years later there were over 1000 species - at least some were triploid apomictic hybrids; C. S. Sargent described many of these. Cotoneaster forms grafts with Crataegus.

As is common, optimisation of characters on the tree presents problems. Potter et al. (2007) used DELTRAN, and as a result being host to Gymnosporangium rusts is not an apomorphy of their Pyrodeae; using ACCTRAN (as here) it is. Along the same lines, they reasonably divided up the presence of sorbitol into two states; it might be present in only small amounts (Dryadoideae), or it was more abundant (Spiraeaoideae); presence of at least some sorbitol characterizes the larger clade.

Classification. Fruit types are certainly not as good indicators of relationships within Rosaceae as was for a long time thought, but chemistry, chromosomes, and fungi all support the realignments suggested by molecular data (especially Morgan et al. 1994), as does developmental work by Evans and Dickinson (1999a, b, 2002). Thus the old Spiraeoideae, characterised by follicular fruit, include a considerable amount of variation and are strongly paraphyletic, now including the old Prunoideae/Amygdaloideae and Maloideae, and this group shows similarities in floral morphology (Kania 1973). In the past Spiraeoideae had been considered a very natural group (e.g. Kalkman 1988), however, tribes in Spiraeaoideae may represent clades (especially Evans et al. 2002; Potter et al. 2006). Prunoideae were characterized by drupaceous fruits, and Maloideae by their pomes (a pretty useless term). However, other sets of relationships have been suggested, thus looking at carpel morphology Schäppi (1951) found Prunoideae to be closer to Spiraeoideae (and Pomoideae) than to Rosoideae.

Generic limits in Potentilleae are becoming clearer (Dobes & Paule 2010, but see Soják 2008 for an alternative).

Previous Relationships. Although Rosaceae as described above are holding together very well despite their morphological heterogeneity, there have been departures. Chrysobalanaceae, often associated with Rosaceae in the past, are part of a distinct clade within Malpighiales, Quillaja rather unexpectedly is an isolated monotypic clade within Fabales, while recently (Oh & Potter 2006) Guametala has been found to be a correspondingly isolated clade in Crossosomatales.

Botanical Trivia. Trees of Polylepis tarapacana grow at the highest known altitude for all trees, some 5,100 m in Bolivia (Hoch & Körner 2005).

[[Barbeyaceae [Elaeagnaceae [Barbeyaceae + Dirachmaceae]]] [Ulmaceae [Cannabaceae [Moraceae + Urticaceae]]]]: trans-spliced intron in nad1 gene [cis-spicing elsewhere].

Evolution. Diversification in this clade (Rhamnaceae sister to the rest) may begin 64-62 million years before present (Wikström et al. 2001).

Chemistry, Morphology, etc. For nad1 intron splicing, see Qiu et al. (1998). For polyembryony, at least sporadic in the clade, see G. Dahlgren (1991).

[Rhamnaceae [Elaeagnaceae [Barbeyaceae + Dirachmaceae]]]: petiole bundle arcuate; stamens = and opposite C/alternate with P; capsule septicidal; ovule basal, erect, parietal tissue 5-6 cells across; coat multiplicative, exotesta palisade, thick-walled; cotyledons large.

Evolution. Divergence & Distribution. Note that most of the apomorphies mentioned above are those of the [Dirachmaceae + Rhamnaceae + Elaeagnaceae] clade for the pre vi.2011 version of the site; they will now reverse - or be uncertain - in the poorly-known Barbeyaceae...

Chemistry, Morphology, etc. The wood anatomy of Dirachmaceae is particularly similar to that of Rhamnaceae (Baas et al. 2001) and the flattened seed with an antiraphal vascular bundle of the former also finds similarities with seeds in the latter (Boesewinkel & Bouman 1997), and the distribution of testal and tegmic epidermal cells with sinuous anticlinal walls could be interesting. Dense, curly hairs on the abaxial surface of the leaf blade could be a synapomorphy for this group (Sytsma et al. 2002); Qiu et al. (1998) put this feature at a higher node. A granular layer below the tectum is more or less developed in both Rhamnaceae and Dirachmaceae - I am unsure of its distribution elsewhere in the clade. Optimising the position of changes in infratectum morphology on the tree is not easy (see also Doyle 2009).

Phylogeny. This grouping is suggested by Campbell (pers. comm. 2003), Sytsma et al. (2002: support weak) and especially Zhang et al. (2011). Rhamnaceae, Barbeyaceae and Dirachmaceae may form a clade (Richardson et al. 2000a), although this was not confirmed by S.-D. Zhang et al. (2011) - note that maximum parsimony bootstrap values found by the latter group for the relationships here are not strong, as is the maximum likelihood support for the clade [Elaeagnaceae [Barbeyaceae + Dirachmaceae]], although the posterior probabilities are all 1.0.

RHAMNACEAE Jussieu, nom. cons.   Back to Rosales

Woody (lianes); leaves much reduced and whole plant thorny); chelidonic acid +; saponins, biflavonyls, benzylisoquinoline alkaloids +, myricetin, ellagic acid 0; (vessel elements with scalariform perforation plates); libriform fibres +; lysigenous mucilage cavities +; leaves opposite or spiral, lamina vernation conduplicate(-plicate) or involute, (margins entire), (2ndary veins palmate), (stipules 0), colleters +; (plant dioecious); flowers small, 4-5(-6)-merous, hypanthium +, to long and tubular; K (connate), with a longitudinal median ridge adaxially, C cucullate, (0), enfolding A; infratectum granulate-intermediate, (nectary as disc; on ovary); G [2-3(-5)], ± inferior, opposite sepals or odd member adaxial, style + or styles separate, stigma papillate; ovules (2; median), epitropous or apotropous, (bistomal), outer integument 4-10 cells across, inner integument 3-4 cells across, parietal tissue 4-7 cells across, nucellar cap to 6 cells across, nucellus ± protruding through micropyle, hypostase +, funicular obturator +; often several megaspore mother cells, antipodals degenerate; K often deciduous; seeds often laterally flattened, (arillate); testa with median integumentary antiraphe bundle +/0, (mesotesta with a few sclerotic cells), endotegmen of cuboid cells, with scalariform thickenings (slightly lignified); endosperm +/0, (starchy), (perisperm +?), polyembryony common, embryo green; n = (6, 8-)12 (13).

52[list]/925 - three groups below. World-wide, especially tropics and warm temperate regions (map: see van Steenis & van Balgooy 1966; Meusel et al. 1978). [Photo - Flower, Dry fruit, Fleshy fruit.]

1. Rhamnoids

G inferior; fruit a drupe or samara; n = 6, 10, 12.

Rhamnus (125). Tropical, north temperate.

Synonymy: Frangulaceae de Candolle

2. Ampeloziziphoids

Venation palmate; fruit a drupe; n = ?

3/4. Cuba, north South America, Madagascar.

3. Ziziphoids

(Herbs; roots with N-fixing Frankia); (indumentum stellate); stipules also petiolar [Colletia]; G semi-inferior to inferior (superior); (embryo sac bisporic, eight celled [Allium type]); fruit a drupe, (winged), also partly loculicidal capsule; (seed arillate); n = 6, 8, 11.

Phylica (150), Zizyphus (100-170), Ceanothus (55), Gouania (50). Pantropical, warm temperate.

Synonymy: Gouaniaceae Rafinesque, Phylicaceae J. Agardh, Ziziphaceae Adanson

• Rhamnaceae are recognisable by their toothed, stipulate leaves that often have strong, parallel secondary veins and scalariform tertiary venation. Their flowers have stamens opposite the often clawed petals, a hypanthium with a nectary inside, and a valvate calyx with sepals that are ridged down the middle on their inner surfaces. The capsules are dehiscent, the fruit wall often separating into two layers with the inner woody layer twisting as dehiscence occurs. Since the seeds are basal, there is usually no columella, while on the ouside of the fruit towards the base there is often a conspicuous rim marking the place where the hypanthium was attached. In taxa such as Gouania with largely inferior ovaries the calyx, etc., is persistent. In lianes the shoots often develop somewhat laterally in the leaf axils and may have a bud at the very base.

  The capsules are rather similar to those of Euphorbiaceae, Phyllanthaceae and Putranjivaceae where the capsule also falls to bits as dehiscence, which can be explosive, occurs; there, too, the fruit wall often separates into two layers, the inner, woody, twisting as dehiscence occurs so helping to eject the seed. However, since the seeds are basal in Rhamnaceae, there is usually no columella, as is so typical of these families, but they lack a hypanthium; Phyllanthaceae and Putranjivaceae have two seeds per carpel.

Evolution. Divergence & Distribution. Fossils from the Cretaceous-Cenomanian, some 94 million years before present, have been identified as members of this family (Crepet et al. 2004 for references). Calvillo-Canadell and Cevallos-Ferriz (2007) describe Mexican fossils from the late Campanian onwards; those frome the late Campanian have tiny spathulate petals ca 0.7 mm long... However, Wikström et al. (2001), followed by Richardson et al. (2004), date stem Rhamnaceae to some 64-62 million years before present. Crown Rhamneae are some (31.2-)28.5, 27.6(-24.9) million years old and Paliureae (34.7-)31.6, 30.6(-27.5) million years (Richardson et al. 2004). Apparent rhamnaceous fossils with puzzling fruits and leaves from the Cretaceous-Maastrichtian some 68 million years ago are in serious conflict with such dates if assigned to these tribes, but not if they are placed incertae sedis (Correa et al. 2010).

Richardson et al. (2004, see also Richardson et al. 2001a) discuss the evolution and historical biogeography of the family in detail, noting i.a. the rapid diversification of the speciose Phylica within the last ca 8 million years. There seems to have been three origins of the climbing habit in the family (Richardson et al. 2000a), and in two of these cases the plants have winged fruits.

Bacterial/Fungal Associations. The North American Ceanothus has N-fixing actinomycetes, as do many Colletieae; they may form a monophyletic group, although there is no evidence of this yet (Richardson et al. 2000b). Within Ceanothus most diversification has been within the last ca 6 million year, beginning well before the origin of the Mediterranean-type vegetation species of the genus favour (Burge et al. 2011).

For the complex ansmycin maytasinoids found in Colubrina and - the precursors, at least - probably synthesized by a bacterium, see references in Cassady et al. (2004). Ectomycorrhizae have been reported from Rhamnus and Pomaderris (Malloch et al. 1980). Quite a few species are xeromorphic.

Seed Dispersal. Perhaps a third of the family - all members of the ziziphoid group - has myrmecochorous seeds (Lengyel et al. 2010).

Chemistry, Morphology, etc. New World species of Gouania have glands at the base of the lamina; Karwinskia has pellucid dots in the leaves. Is there some kind of chalazal haustorium (see Srinivasachar 1940)? The outer epidermis of the outer integument of the ovule can have rather large cells (Juel 1929).

General information is taken from Brizicky (1964) and Medan and Schirarend (2004); Vikhireva (1952: not read) described fruit anatomy, Hegnauer (1973, 1990) summarized chemistry, Medan (1985, 1988) discussed gynoecial development, Medan and Aagesen (1995) and Medan and Hilger (1992) comparative floral and fruit morphology, and Juel (1929) and Arora (1953) ovules, etc.

Phylogeny. There are three main clades in the family, the rhamnoids, which include three tribes, the ziziphoids (five tribes, plus assorted genera, inc. Cenaothus), which include most of the rest of the family (for the Pomaderreae, one of these tribes, see Kellermann et al. 2005 and Kellermann & Udovicic 2008), and the ampeloziziphoids (three tribes: three genera: four species) - see Richardosn et al. (2000b). Relationships between the major clades in the zizphoid group in particular are poorly understood, and there is weak support for a [zizphoid + ampeloziziphoid] clade (Richardson et al. 2000b). Zizyphus is paraphyletic (Islam & Simmons 2006).

Classification. Richardson et al. present a phylogeny (2000b) and a classification based on it (2000a).

[Elaeagnaceae [Barbeyaceae + Dirachmaceae]]: ?

ELAEAGNACEAE Jussieu, nom. cons.   Back to Rosales

Trees or shrubs; roots with N-fixing Frankia; dihydroflavonols?, 0-methyl flavonoids, ellagic acid +, myricetin 0; no mucilage cells in leaves; hairs lepidote or stellate; cambium storied; phloem stratified; pits vestured; true and fibre tracheids +; fibre pits bordered; wood with broad rays; sieve tubes with non-dispersive protein bodies, plastids lacking starch grains; nodes 1:1; petiole bundles arcuate or annular; mucilage cells?; leaves spiral or opposite, lamina vernation conduplicate-flat, margins entire, stipules 0; (plant dioecious - Hippophae); inflorescence a raceme, or flowers axillary; flowers (2-)4(-6)-merous; hypanthium long, C 0; A also 2 x P, borne in throat of tube; pollen 3-nucleate; G 1, stylulus long, stigma decurrent or capitate; ovule micropyle?, outer integument 5-16 cells across, inner integument 3-4 cells across, funicular obturator +; megaspore mother cells several; hypanthium accrescent and fleshy and closely investing fruit; pericarp thin [Hippophae]; testa very thick, exotesta with sinuous anticlinal walls at least in part, (not palisade), mesotesta ± thick-walled; endosperm with chalazal haustorium, (starchy), cotyledons usu. unequal; n = 6, 10, 11, 13, 14.

3[list]/45: Elaeagnus (20-45). North Temperate, warm tropical; Malesia and Australia - also quite widely cultivated and/or escaped (map: from Meusel et al. 1978; Hultén & Fries 1986). [Photos - Collection] [Photo - Shepherdia Fruit © R. Kowal]

• Elaeagnaceae are often thorny and mostly deciduous shrubs that are recognisable by their exstipulate leaves with entire margins; the whole plant is densely covered with scaly-stellate indumentum, which often gives it a greyish or silvery appearance. The hypanthium is relatively long and there is no corolla. The dry fruit proper is surrounded by the fleshy hypanthium (cf. some Nyctaginaceae), and so the result is very like a berry or drupe - which is all a hungry bird cares about.

Evolution. Bacterial/Fungal Associations. All genera are associated with N-fixing Frankia, and cluster roots have been reported from Hippophae (Shane & Lambers 2005).

Chemistry, Morphology, etc. The androecium is obdiplostemonous according to Huber (1963). Seed anatomy is rather like that of Rhamnaceae (Corner 1976); Harrison and Beveridge (2002) have clarified fruit and seed anatomy of Hippophae. For rust host preferences, see Savile (1979), for general information, see Bartish and Swenson (2004).

Previous Relationships. Elaeagnaceae have been difficult to place. They were included in Proteales by Cronquist (1981) because of superficial floral similarities, and in Elaeagnales - Rhamnanae, next to Proteanae, in Rosidae, by Takhtajan (1997).

Synonymy: Hippophaeaceae G. Meyer

[Barbeyaceae + Dirachmaceae]: connective produced; ovule with outer integument 3-5 cells across, parietal tissue 5-6 cells across.

Chemistry, Morphology, etc. In both families the embryo sac is quite deep seated, and in Dirachma socotrana in particular, the nucellar tissue at the chalazal end of the seed is very well developed.

BARBEYACEAE Rendle, nom. cons.   Back to Rosales

Trees; ellagic acid +; libriform fibres +; nodes 1:1; no mucilage cells in leaves; stomata laterocytic; hairs unicellular, spirally twisted; leaves opposite, lamina vernation supervolute-curved, margins entire, stipules 0; plant dioecious; inflorescence fasciculate, bracts and bracteoles 0; hypanthium 0; nectary 0; P 3-4; A (6-)9-12; pollen infratectum granulate-intermediate; G 1-2(-3), ± separate, placentation subapical, styluli long, stigma long-clavate, decurrent, ?type; ovule epitropous, micropyle long-endostomal, inner integument 5-6 cells across, nucellar cap ca 5 cells across, hypostase +; fruit a nutlet; P accrescent; seed coat undistinguished, exotesta perforated, not palisade, anticlinal walls sinuous, endotegmen tanniniferous, anticlinal walls sinuous; n = ?

1[list]/1: Barbeya oleoides. N.E. Africa, Arabia (map: from Aubréville 1974).

• Barbeyaceae may be recognised by their opposite, entire, exstipulate leaves with dense white indumentum on the undersurface and their shortly pedunculate, axillary, cymose inflorescences. The flowers are small, wind-pollinated and with a uniseriate perianth and the nutlets are enveloped by the accrescent perianth.

Chemistry, Morphology, etc. The sieve tubes have compound perforations, unlike Ulmaceae and its immediate relatives and other Rosales. The perforated seed coat is rather like that common in the Ulmaceae group (Bouman & Boesewinkel 1997).

Additional information is taken from Dickison and Sweitzer (1970: morphology), Tobe and Takahashi (1990: hairs and pollen), Friis (1993: general), and Bouman and Boesewinkel (1997: ovule and seed); Hegnauer (1990) has a little information on chemistry.

Previous Relationships. The monotypic Barbeyales were placed in Hamamelididae-Urticales by Cronquist (1981) and in Hamamelididae-Barbeyanae by Takhtajan (1997).

DIRACHMACEAE Hutchinson   Back to Rosales

Shrub; stalked glands +/0; chemistry?; phloem stratified; nodes ?lacunar; leaves spiral, stipules subulate, persistent; flowers single, terminal; flowers 5-8-merous; epicalyx of 4-8 lobes (in middle of pedicel); K basally connate, C contorted, vasculature fan-shaped; nectaries on base or on subbasal appendages, lacking stomata; anthers extrorse, long, opening from apex; G [8], lobed, opposite the K, style +, stigma clavate or punctate, elongated; micropyle zig-zag, inner integument ca 2 cells across, nucellar cap 0, hypostase ?+; fruit beaked, segments opening adaxially and from the base, wooly inside, with columella, K deciduous above the "hypanthium"; seeds laterally flattened, tegmen multiplicative?, median integumentary antiraphe bundle +, exotesta with anticlinal walls thickened, endotegmen tanniniferous; endosperm slight, embryo color?; n = ?

1/2. Socotra, Somalia (map: from Link 1991b).

• Dirachmaceae are shrubs that may be recognised by their small, strongly-toothed leaves which have subulate stipules; the latter persist along the stem and particularly prominent on short shoots. The flowers have an epicalyx and are relatively large, and the stamens are opposite the petals and have extrorse anthers that clearly open from the top downwards. The capsules open and separate from the base and so the dehiscing fruit looks like a parasol; the loculi are densely wooly inside.

Chemistry, Morphology, etc. The single flowers may represent a reduced, cymose inflorescence (Ronse De Craene & Miller 2004). A "hypanthium" is described as involving either the sepals and petals and also the petals and stamens (Ronse De Craene & Miller 2004); there does not seem to be a conventional hypanthium. Petal initiation is later than that of the stamens, as is common is rosids, and until quite late in development they are very much shorter than the stamens. There is a lot of starch in the pollen grains (Boesewinkel & Bouman 1997). The presence of an antiraphal vascular bundle means that the outer integument is very thick (6-9 cells or more) when viewed in the abaxial median plane; integument measurements are taken from the sides of the ovule (see Boesewinkel & Bouman 1997). There is a structure on the seed described as a small, funicular aril, perhaps similar to that found in some Rhamnaceae (Ronse De Craene & Miller 2004).

For additional information, see Link (1991b [cf. 1994a], 1994b), Boesewinkel and Bouman (1997: ovule and seed), Baas et al. (2001), Ronse De Craene and Miller (2004: floral morphology), and Bayer (2004: general).

Previous Relationships. The exotestal seeds with straight embryos suggest that Dirachma is not close to Geraniaceae (Geraniales), with which Dirachma had been linked, as by Cronquist (1981) - see e.g. Boesewinkel (1985) and Boesewinkel and Bouman (1997). Takhtajan (1997) included Dirachmaceae in his Malvales mainly because of its similarity in gross morphology.

[Ulmaceae [Cannabaceae [Moraceae + Urticaceae]]]: flavonols and their glycosides, myricetin [some Ulmaceae, Cannabaceae] +, ellagic acid 0; plant with ± watery exudate; hairs unicellular and multicellular-glandular; cambium ± storied; libriform fibres +; phloem stratified; sieve tubes with non-dispersive protein bodies (some Cannabaceae - 0), (plastids lacking starch grains); cystoliths [globose; usu. CaCO₃] and epidermal and hair cell wall silicification and calcification common; at least one prominent prophyllar bud; lamina with 2ndary veins proceeding straight to non-glandular teeth and higher-order veins convergent on those teeth [urticoid], stipules cauline; flowers small; hypanthium?, P +, stamens equal and opposite P members; pollen porate, infratectum granular; nectary 0; G [2], abaxial only fertile, placentation apical, stigmas sessile, spreading, receptive area extending down adaxial surface and ± confluent; ovule epitropous; fruit a drupe; testa perforated [rare in Ulmaceae]; endosperm scanty, polyembryony common; x = 14, centromeres both median and subterminal.

Evolution. Divergence & Distribution. This clade may be some 67-65 million years old, Ulmaceae diverging 57-55 million years before present, the rest ca 48-42 million years before present (Wikström et al. 2001).

Plant-Animal Interactions. Some Nymphalidae-Nymphalini butterflies and their immediate outgroups have larvae on members of these families (see also under Urticaceae) - but also on the immediately unrelated Euphorbiaceae (Malpighiales: see Ehrlich & Raven 1964). However, the ancestor of Nymphalinae may have fed on Urticaceae and relatives (Nylin & Wahlberg 2008). Thus, caterpillars of Acraea (Acraeinae) are found quite commonly on Urticaceae (including Cecropia), but also on Moraceae, etc.; this particular genus is also particularly common on Passifloraceae and their relatives.

Chemistry, Morphology, etc. Raffinose and stachyose are common oligosaccharides in phloem exudate in Ulmaceae, Moraceae and Cannabaceae sampled (Zimmermann & Ziegler 1975). The group has homogeneous wood anatomy: Rays are relatively broad, pits are simple, intervessel pitting is alternate, fibres are septate, and parenchyma is paratracheal (Baas et al. 2000). Furthermore, at least some members (Celtis, Ulmus) have a torus-bearing, pit membrane (Coleman et al. 2004) that is only weakly lignified. Two-ranked leaves may be an additional synapomorphy for the group (or pegged at a still higher level), as well as urticoid teeth. Because of the well-developed prophyllar bud(s), the inflorescences are often paired, with a bud between them, and/or the branches may have a bud on one or both sides at the base; Ulmus does not show this arrangement. The "stipular buds" of Cannabis (Miller 1970 and references) are really prophyllar buds.

It is not clear which taxa have a hypanthium; at least some species of Ulmus and Pilea do, but other species of Ulmus, Zelkova and Trema show no obvious signs of one (see also Bechtel 1921). Staedler (1923) discusses the absence of an anther epidermis in the group (but not in Ficus; situation in other Rosales?); there is no obvious link with dehiscence mechanism. Starchy pollen is common, but apparently not in Urticaceae. Bechtel (1921) and Eckardt (1937) described gynoecial morphology in considerable detail. Taxa with a perforated testa are quite common in the group, although this feature may have arisen more than once (Kravtsova & Oskolski 2007).

For further information, see Satake (1931: spodograms), Sweitzer (1971: anatomy), Giannasi (1978, 1986: chemistry), Behnke and Barthlott (1983: hairs), Takaso and Tobe (1990: testa), Omoron and Terabayashi (1991: phylogeny), Terabayashi (1991: vernation), Hennig et al. (1994: cuticle waxes), and Tobe and Takaso (1996: hairs).

Phylogeny. The phylogenies suggested by Sytsma et al. (2000) and Song et al. (2001) place Cannabaceae within Celtidaceae (see also e.g. Ueda et al. 1997b; note that Cannabaceae is the earliest name for the combined group) and Cecropiaceae within Urticaceae, and this set of relationships has been strongly supported by a recent, more comprehensive analysis (Sytsma 2002). The group as a whole is very well characterised, and it still may make sense to expand Urticaceae, as, for instance, Corner (1952) did when he placed Moraceae in Urticaceae. See Judd et al. (1994) for a morphological phylogeny and Zavada and Kim (1996) for a molecular phylogeny focussed on the old paraphyletic Ulmaceae s.l.

The copious information on the four families awaits synthesis, although this process has been begun by Sytsma et al. (2002); this paper should be consulted for details of character evolution. Sytsma et al. (2002) note that inflexed stamens and their dehiscence, fruit type, and laticifers need further detailed study within this clade; hypanthium presence and other characters can be added to this list! Mohan Ram and Nath (1964) noted that the chalazal megaspore was the functional one, but I do not know how important this is.

ULMACEAE Mirbel, nom. cons.   Back to Rosales

Trees, (growth symopodial, apex of innovation aborts); lignans +; (wood fluoresces); unicellular hairs smooth; sieve tube plastids often with protein crystalloids; cystoliths usu. pegless; leaves two-ranked, lamina vernation laterally (vertically) conduplicate-plicate, 2ndary veins going into teeth, stipules extrapetiolar; flowers perfect and mixed; P spiral, (connate); A extrorse, (2 x P); pollen 4-7-porate, exine rugulose; at least one stigma with 3(-5) vascular bundles; ovule (with bistomal micropyle), outer integument 4(-6 - Holoptelea) cells across, inner integument ca 4 cells across; fruit also a samara; seeds flattened, coat undistinguished, exotestal cells elongated, unthickened; (embryo curved - Zelkova); often terminal/subterminal diffuse-complex centromeres; 69bp ndhF deletion.

6[list]/35: Ulmus (20-30, species limits uncertain). Mostly N. temperate, esp. Asian, but scattered elsewhere except Australia and the Pacific (map: from Soepadmo 1977; Hultén & Fries 1986; Fl. N. Am. III 1997, incomplete). [Photos - Collection]

• Ulmaceae are usually trees and may be recognised by their usually two-ranked, strongly serrate, stipulate leaves with asymmetrical bases; the pinnate secondary veins proceed to the apex of the teeth (not in Ampelocera). At least some of the small flowers are perfect and the fruits are flattened.

Evolution. Divergence & Distribution. Ulmus is known as leaves and fruits from Early Eocene deposits of northeastern China some 50 million years old (Wang et al. 2010).

Chemistry, Morphology, etc. Distinctive fatty acids were founds in the seeds of some Ulmaceae, but not in those of Cannabaceae or Moraceae surveyed (Badami & Patil 1981: sampling). Ulmus lacks well-developed prophyllar buds and has one of the two stipules intrapetiolar, they are both intrapetiolar in seedlings of some species, and the leaves may also be opposite, as in seedlings. Hemiptelea has pegged cystoliths. The breeding system in the family is variable, but at least some flowers are perfect. Nawaschin (1895) suggested that chalazogamy occurred in Ulmus. Holoptelea has a thick-walled exotesta.

For gynoecial morphology, see Okamoto et al. (1992).

Phylogeny. The poorly-known Ampelocera is to be included here (see Ueda et al. 1997b; also Wiegrefe et al. 1998); although its hairs are smooth, its leaves have ascending veins.

[Cannabaceae [Moraceae + Urticaceae]]: C-glycoflavones also +; (sieve tube plastids with starch grains); unicellular hairs usu. micropapillate; 2ndary veins palmate; stipules cauline-intrapetiolar; flowers imperfect; embryo curved.

Evolution. Seed Dispersal. Members of all threee families are quite often dispersed by bats in the New World (Lobova et al. 2009).

CANNABACEAE Martynov, nom. cons.   Back to Rosales

Trees or ± herbaceous, erect or twining; sesquiterpene lactones +; true tracheids +; cystoliths usu. with pegs [distribution?]; leaves (spiral; opposite), lamina vernation (laterally) conduplicate-plicate (conduplicate; supervolute; (strongly palmately lobed or compound, veins proceeding to the apex of lobes - Cannabis, Humulus), 2ndary veins ascending, stipules connate or not; pollen 2-3(-5) porate, exine scabrate to verrucate; stigmas with single vascular bundle; micropyle bistomal, outer integument 2-4(-8) cells across, inner integument 2-3 cells across, parietal tissue ca 6 cells across, nucellar cap ca 2 cells across; embryo sac haustorium +; (fruit an achene); seeds globose, exotestal cells tangentially elongated, with arms, unthickened; endosperm +, embryo also coiled; (n = 8-11, 13, 15), centromeres medial/submedial, simple.

11[list]/170: Celtis (ca 100). Worldwide, but not Arctic, distribution of Humulus lupulus in Aa sia is unclear (map: from Wickens 1976; Soepadmo 1977; Hultén & Fries 1986; George 1989; Fl. N. Am. III 1997). [Photos - Collection] [Photo - Celtis Flower]

• Cannabaceae may be recognised by their two-ranked, serrate leaves with ascending veins that do not proceed straight into the teeth, more or less intrapetiolar stipules, and prominent prophyllar buds. The fruit is a rounded drupe. However, Humulus and Cannabis have opposite and spiral leaves respectively, secondary veins running into the teeth, and Humulus lacks buds at the base of the staminate inflorescence...

Evolution. Divergence & Distribution. For the early Tertiary fossil history of Cannabaceae that are now East Asian endemics, see Manchester et al. (2009).

Plant-Animal Interactions. Caterpillars of most Apaturinae (Nymphalidae) are found on Cannabaceae, although those of Hestinalis eat Urticaceae, Sephisa, Fagaceae, and Apatura, the spectacular purple emperor, has moved on to salicin-containing Salicaceae (Salix and Populus: Ohshima et al. 2010).

Bacterial/Fungal Associations. Parasponia is the only non-legume nitrogen fixer that is associated with other than actinomycetes, and its rhizobia remain in infection threads like those of some Fabaceae, especially "Caesalpinioideae". Infection of the plant is by entry through cracks in the epidermis. The nodules are modified lateral roots and have central vascular tissue, the bacteria being in the cortex; this is unlike the nodules of Fabaceae but like yhode found elsewhere in the N-fixing clade that are Frankia-induced (Streng et al. 2011 for some literature). Gironniera is reported to be ectomycorrhizal (Smits 1994).

Genes & Genomes. Cannabis and Humulus have an X-autosome balance system determining the 'sex' of the plant.

Chemistry, Morphology, etc. Humulus, with its opposite leaves, has split laterals/a commissural vascular bundle (Colomb 1887). The "distinctive" camptodromous (to semicraspedodromus) venation of Celtidaceae s. str. is disturbed both by the inclusion of Cannabis, etc., in this clade, but also by the occurrence of strictly craspedodromous venation in Palaeocene Celtis (Manchester et al. 2002).

Whether the laticifers of Cannabis etc., are really similar to those of Urticaceae and Moraceae must be confirmed; there is no milky exudate and they occur throughout the plant. Only the "basal" Aphananthe and Gironniera have flavonols. Lozanella, with its opposite leaves and boxy venation, looks rather like Urticaceae, but it has a bilobed stigma. The anticlinal walls of the testa of Humulus are sinuous and its embryo is green. n = 10 is common in Cannabaceae.

For general information, see Grudzinskaya (1967), Mohan Ram and Nath (1964: Cannabis ovules and seeds), Leins and Orth (1979: Cannabis flowers), Todzia (1993: general, as Ulmaceae), and Kubitzki (1993b: general, as Cannabaceae), for chemistry, see Hegnauer (1973, 1990, as Ulmaceae; also 1964, 1989).

Phylogeny. Pteroceltis, Humulus, and Cannabis are close, and they and some other members of this clade have sieve tube plastids with starch grains (Behnke 1989). Lozanella is sister to Aphananthe or sister to the rest of the family (Wiegrefe et al. 1998; Soltis et al. 2002).

Synonymy: Celtidaceae Link, Humulaceae Berchtold & J. Presl, Lupulaceae Link

[Moraceae + Urticaceae]: latex system +; (calcium carbonate crystals); (lamina with pinnate 2ndary veins); staminate flowers: stamens incurved in bud; pistillode +; polyembryony 0.

Evolution. Genes & Genomes. Rates of molecular evolution are likely to have increased at least twice in this clade, e.g. in most Urticaceae and in Dorstenia as compared with other Moraceae, i.e. the increase is associated with the herbaceous habit (Smith & Donoghue 2008).

There is a group of genera of Moraceae with explosive pollen dispersal, and inflexed stamens may be a synapomorphy for [Moraceae + Urticaceae] (Datwyler & Weiblen 2004), although where they change on the tree of the former family depends very much of how they are optimized (Clement & Weiblen 2009). The moraceous genera with such stamens are in the paraphyletic Moreae that was part of a basal polytomy within Moraceae in early reconstructions. Indeed, the straightening stamens and reflexing tepals of Morus alba are reported to show the fastest movement of any plant parts known, over half the speed of sound (Taylor et al. 2006).

Chemistry, Morphology, etc. Scattered in the group are taxa in which the tepals are persistent (e.g. Pilea) or accrescent (e.g. Morus), i.e. the fruits are anthocarps in the strict sense.

Phylogeny. For a morphological phylogeny focussing on Urticaceae s.l., i.e. Moraceae and Urticaceae here, see Kravtsova and Oskolski (2007).

MORACEAE Gaudichaud, nom. cons.   Back to Rosales

Largely woody; (isoflavonoids +); (cork in outer cortex); laticifers throughout the plant, latex milky; (stomata aniso- and cyclocytic); leaves spiral, lamina vernation variable, stipule also ensheathing stem [open in leaf axil - Ficus], (0 - Fatoua); dioecious [plesiomorphously so], inflorescence congested, spicate [staminate] or ± globose [carpellate); flowers 4-merous, (P 0-10); staminate flower: bracts peltate; P free; carpellate flower: P connate; (G inferior), styles 1 or 2, often unequal; ovule (subapical; campylotropous), outer integument 3-4 cells across, inner integument ca 3 cells across, nucelar cap ca 5 cells across [nucellar beak]; fruits a drupe or achene, receptacle often accrescent; seed coat undistinguished, exotesta ± tanniniferous, (several thickened layers - Prainea); (endosperm +); (n = 12 upwards, esp. 13, 14, variation great in Dorstenia), both terminal and median centromeres.

39[list]/1125 - six groups below. Mostly tropical to warm temperate (map: from Jalas & Suominen 1976; George 1989; Fl. N. Am. III 1997).

[Artocarpeae + Moreae]: carpellate flowers: bracts peltate.

1. Artocarpeae Lamarck & de Candolle

(P connate, that of adjacent flowers adnate in the middle portion [completely]); staminate flowers: 2 merous; A 1(-3), filaments straight.

3/67: Artocarpus (61). Largely tropical, Indo-Malesia (Artocarpus) and America. Photo: Fruit].

Synonymy: Dorsteniaceae Chevallier

2. Moreae Dumortier

Staminate flower: (bracts not peltate; filaments straight).

7/56. Tropical, some temperate.

[Maclureae [Dorstenieae [Ficeae + Castilleae]]]: ?

3. Maclureae Clement & Weiblen

Habit various, axillary thorns +; inflorescence with golden dye; staminate flowers: bracts not peltate, (filaments straight); pistillate flower: (bracts peltate).

1/11. Tropical to Temperate. [Photo - Inflorescence.

[Dorstenieae [Ficeae + Castilleae]]: plant monooecious, inflorescence bisexual; radial latex tubes +; pistillode conical.

4. Dorstenieae Dumortier

(Herbs); staminate flowers: (3-merous; P connate; filaments straight); carpellate flowers: (2-merous; embedded in receptacle.

13/145: Dorstenia (105). Pantropical. Photo: Fruit.

Synonymy: Dorsteniaceae Chevallier

[Ficeae + Castilleae]: inflorescence axis expanded, insect pollination; staminate flowers: filaments straight; carpellate flowers: bracts not peltate; P free.

5. Ficeae Dumortier

Habit various; (latex 0; leaves opposite); inflorescence a syconium, pollination by wasps, flowers 3-merous: staminate flowers: bracts not peltate; A 2; carpellate flowers: style branches unequal or not.

1/800. Pantropical. Photo: Fruit.

Synonymy: Ficaceae Berchtold & J. Presl

6. Castilleae C. C. Berg

(Branches spreading, cladoptosis); (cystoliths 0); infloresence involucrate, unisexual, discoid to urceolate; staminate flower: bracts 0, pistillode 0; carpellate flowers: (bracts 0).

11/62. Pantropical, esp. America.

• Moraceae are laticiferous plants with stipulate leaves and often boxy fine venation; the flowers are small and densely aggregated; the inflorescence axis is often large, becoming much accrescent and ± fleshy in fruit.

Evolution. Divergence & Distribution. Berg (2005) suggests that diversification in Moraceae occurred on a still physically coherent tropical supercontinent. Zerega et al. (2005) advance a more complex hypothesis to explain the distribution and diversification of the family; they date crown group diversification to at least 79 million years before present, and the divergence of the stem group to at least 89 million years before present.

Ecology. Moraceae, and in particular Ficus, includes a number of (hemi) epiphytes, stranglers and lianes. The family is often the second most speciose family in lowland tropical rainforest in America and Africa (Gentry 1988). Figs in particular are a very important and dependable source of food for frugivores, both birds and mammals, throughout the tropics, and the diversity of their growth forms means they are encountered throughout the forest. Indeed, figs are perhaps surprisingly nutritious (Herre et al. 2008 for references), and individuals of many species will be in fruit throughout the year. It has been found that the species richness of Ficus is correlated with that of their avian frugivores, particularly specialised frugivores, and the frugivores also select on particular aspects of the morphology of the fig species (Shanahan et al. 2001; Kissling et al. 2007; Herre et al. 2008). Some bats also eat figs, and, depending on whether the bats are New or Old World, they search for food in different ways and have selected figs with different qualities - thus in the New World figs tend to be greenish and odoriferous (Compton 1996 [a whole series of papers]; Korine et al. 2000; Shanahan et al. 2001; Harrison 2005).

Plant-Animal Interactions. Bombyx mori (the silkworm) caterpillars can eat quite widely within Moraceae, but not on most other members of the Ulmaceae group - although they will grow on Ulmus itself (Fraenkel 1959). Caterpillars of danaine butterflies quite commonly use Moraceae as food plants; both they and their usually preferred Apocynaceae are rich in latex, although Moraceae do not often have cardenolides (Ackery & Vane-Wright 1984).

Floral Biology & Seed Dispersal. Basic details of the association between figs and their agaonid fig-wasp pollinators (Agaonidae) may be summarized as follows, although the life cycle in dioecious figs is more complex. Female wasps are fertilized by flightless males inside the fig, and they pick up pollen from the male flowers - at least sometimes aggregated around the ostiole - before leaving the fig via exits prepared by the males. They enter a receptive fig via the ostiole, losing their wings, etc., in the process, and then lay eggs in the female flowers inside, pollinating the flowers at the same time. They cannot lay eggs in ovaries of flowers with long styles, their ovipositors being too short, and these flowers develop into fruits; they can lay eggs in shorter-styled flowers, and these develop into the next generation of wasps.

Note that despite fig pollen being dispersed by tiny agaonid wasps, these wasps can sometimes be transported up to 14 km in the rainforest, the result being that breeding units of figs may be an order of magnitude larger - covering some 100 square kilometres or more - than those of other plants in the rainforest (Herre 1996; Nason et al. 1998). Furthermore, transfer of pollen for distances up to 160 km has been found in riparian popularions of Ficus sycomorus in the Namib desert; in more open environments the female wasps can be passively transported very long distances (Ahmed et al. 2009).

The fig/fig wasp association has been touted as a classic example of an obligate one-on-one association of plant and wasp pollinator (e.g. Herre 1996 and references). The beginning of the association/co-divergence dates to 100-60 million years before present. Lopez-Vaamonde et al. (2009) discuss dating and biogeography of the wasps, suggesting that the ancestor of the wasps lived in Asia, dispersal rather than Gondwanan break-up being important for their subsequent distribution. Interestingly, Ficus, included in Ficeae, is sister to Castilleae (which in turn include some errant Artocarpeae), and members of both tribes have urceolate inflorescence axes and the insect pollinators (these are thrips in Castilleae) breed among the flowers (Datwyler et al. 2003; Datwyler & Weiblen 2004; Clement & Weiblen 2009).

However, recent work suggests that there is surprisingly often rather little specificity in the association between fig and wasp (Machado et al. 2005; Jackson et al. 2008), so questioning both the classic idea of an obligate association and also the frequency of occurrence of strict co-speciation. Nevertheless, there is at least a general association between figs and wasps, and in groups like section Ficus sect. Galoglychia cospeciation does seem to occur, even in the non-pollinating wasps that are part of the whole system (Jousselin et al. 2008). Pollinator sharing seems to be more common in monoecious neotropical figs than in palaeotropical dioecious figs (Moe et al. 2011).

The parasites of the agaonid wasps and and also galling wasps are part of the fig-fig wasp system (e.g. Lopez-Vaamonde et al. 2001; Jousselin et al. 2003; Jackson 2004a; Kjellberg et al. 2005; Rønsted et al. 2005b, 2008a for references). The parasitoid wasps that parasitize the fig wasps may be of considerable importance in preserving the relationship between the agaonid wasp and fig (Dunn et al. 2008; for more information on figs and wasps, as well as gallers and parasitioids, see also Kjellberg et al. 2005; the papers in Symbiosis 45, nos 1-3, 2008; Silvieus et al. 2008; especially Herre et al. 2008). Galling chalcid wasps belong to five groups of chalcids (Hymenoptera-Chalcidoidea) and gall individual flowers; they may have have moved on to figs some 50-40 million years ago, with subsequent dispersal to other groups (Cruaud et al. 2011). As many as 36 species of chalcids may be found on a single species of fig (Cruaud et al. 2011). Also involved are drosophilid flies that in Africa, at least, have a very close association with figs and oviposit either on the stomium or the exit holes made by the male fig wasps (Harry et al. 1996, 1998).

Seed dispersal is often by animals, but just what part of the plant contributes to the fleshiness of the infructescence varies. In Moraceae such as Morus the fruits - drupes - are closely packed on all sides of an erect inflorescence axis. In Artocarpus - it includes the bread and jack fruits - the more or less elongated infructesences can weigh up to 50 kg, and the fleshy part comes from perianth members amnd the inflorescence axis (see Zerega et al. 2010 for a phylogeny and morphology). Castilleae have both these kinds of inflorescences; in Antiaropsis the inflorescence axis is concave at first, but spreading ["dehiscing"] when ripe, when the fruits proper contrast in colour with the bracts, etc. Finally, in Ficus the flowers are borne on the inside of an invaginated inflorescence axis. The axis is fleshy and may be dull coloured and the whole infructescence may be scented; in the New World Artibeus bats (phyllostomids) commonly disperse the seeds of such fruits, although they also like fruits of Cecropia (Lobova et al. 2009) Bat-dispersed species in Malesia may lack odour; pteropodids are the dispersers there (Hodgkinson et al. 2003). The fruits of some species of Ficus are dispersed by birds, but fruits of such species are generally more or less brightly coloured (see also Hodgkinson et al. 2003; Lomáacolo et al. 2008, 2010). Dorstenia, etc., have "dehiscent" drupes; here the stone is shot out of the turgid mesocarp, the separation following a line of weakness in the tissue. The flowers are borne on the upper side of a flat inflorescence axis.

Chemistry, Morphology, etc. Laticiferous cells elongate and branch and intrude between other cells, the nuclei divide, but cell walls are not formed. Some species of Dorstenia have small, cauline stipules that do not overlap the petiole. For fig/syconium development, see Rauh and Reznik (1951). Chromosome numbers in Dorstenia are very variable.

Some information is taken from Rohwer (1993a); for embryology, see Johri and Konar (1956: Ficus), for chemistry, see Hegnauer (1969, 1990); for pollen, see Burn and Mayle (2008); chromosomes, see Oginuma and Tobe (1995).

Phylogeny. For phylogeny, see Datwyler et al. (2003), Datwyler and Weiblen (2004) and especially Clement and Weiblen (2009). The paraphyletic Moreae with their incurved stamens include Morus (fleshy perianth + drupe, a syncarp), Maclura (ditto), and Broussonetia (drupe - tapa cloth). Clement and Weiblen (2009) confirmed that Castilleae are sister to Ficeae, and the relationships between the tribesd are on the whole well supported. Analyses of morphological characters alone provided little resolution, only Ficeae (one genus!) being well supported, although most Castilleae formed a clade, albeit with very little support. Zerega et al. (2010) provide a detailed phylogeny of Artocarpeae, which they recircumscribe.

Classification. See Clement and Weiblen (2009) for tribal delimitation (note likely position of Treculia); they also discuss one or two problems with generic limits.

URTICACEAE Jussieu, nom. cons.   Back to Rosales

Herbs to shrubs (trees; lianes); dihydroflavonols?, (furanocoumarins) +; cork cortical [Urtica]); (wood fluoresces); (wood paremchyma not lignified); laticifers in bark only, latex 0 (milky); petiole bundle(s) annular or arcuate; bundle sheath extensions 0; cystoliths often elongated (0); stomata often anisocytic (paracytic, etc.); leaves two-ranked or spiral; lamina base often asymmetrical, vernation laterally or vertically conduplicate, stipule also intrapetiolar or sheathing [open on stem opposite leaf] (0 - Soleirolia); (plant dioecious); P (1-)4, 5(-6), valvate, (connate), staminate flowers: filaments explosively straightening, (not explosive - Poikilospermum), endothecium 0, pistillode +; carpellate flower: staminodes +; G "pseudomonomerous", style 1, or stigma sessile, penicillate; ovule basal, ± straight, both integuments often 2 cells across, (protruding into the stylar cala; inner integument obturator, of several 1-celled thick projections in t.s.), parietal tissue 4-6 cells across, nucellar cap 2-4 cells across; fruit also often a nut or achene; seed coat perforated, ± crushed, but various testal/tegmic layers persisting; chalazal endosperm haustorium +, (endosperm +; starchy), embryo straight, (cotyledons thick, radicle short - Pouruma, Myrianthus); n = 7-14 [x = 14 is unlikely to be basal], protein bodies in nuclei.

54[list]/2625: Pilea (500-600), Elatostema (300), Urtica (80). World-wide, but mainly tropical (map: from Hultén & Fries 1986; George 1989; Fl. N. Am. III 1997). [Photo - Shoot, Flower, Fruit.]

• Urticaceae are variable in habit but have stipulate leaves with boxy venation. The small flowers are borne in branched, usually clearly cymose inflorescences - carpellate inflorescences may be capitate - and the anthers are often explosively dehiscent. Some genera related to Cecropia are trees with stout stems, deeply lobed or compound leaves, densely spicate inflorescences, and stamens that are not explosive.

Evolution. It has been suggested that caterpillars of Nymphalini butterflies have a plesiomorphic association with Urticaceae as food plants (Janz et al. 2001); caterpillars in a clade of Nymphalidae-Heliconiinae-Acraeini utilise members of this family, probably switching from host plants in the Passifloraceae area (Silva-Brandão et al. 2008; see also Nylin & Wahlberg 2008).

Cecropia is a fast-growing pioneer tree that is associated with Azteca and some other ants that live in the stems and eat glycogen-rich food bodies (Müllerian bodies) produced by the plant at the abaxial base of the petiole (for a phylogeny, see Treiber & Weiblen 2009); beetles may also eat these food bodies, as well as ant eggs, etc. (Jolivet 1991; Longino 1991; Yu & Davidson 1997). The association between plant and ant may break down, especially on islands and at high altitudes (Janzen 1973); Musanga, e.g. M. cecropioides, from Africa, also lacks ants but is otherwise very similar to Cecropia. Note that species of Macaranga (Euphorbiaceae, q.v.) are ecological analogues of Cecropia in Malesian forests.

Explosive dispersal of the pollen as the filaments abruptly straighten is common in Urticaceae. In Pilea the achenes are similarly dispersed by the abrupt straightening of the fleshy, inflexed filaments of the staminodes.

Chemistry, Morphology, etc. Groups of cells in the vascular tissue may be unlignified and the pericyclic sheath may also be late in lignifying; the presence of unlignified apotracheal parenchyma was used to divide the family (minus Cecropiaceae) into two main groups (Bonson & ter Welle 1984); Cecropia et al. have lignified paremnchya (Bonson & ter Welle 1983). Boehmeria has a fleshy perianth. For the absence of an endothecium, see Staedtler (1923); this character needs more extensive sampling in the context of recent phylogenies. The gynoecium is basically bicarpellate, but one carpel is highly reduced (Eckardt 1937). Shamrov (2004) shows the inner integument of Leucosyke becoming much elaborated and functioning as an obturator. The seeds of Dendrocnide may lack holes in the exotesta, and the whole seed coat is relatively well developed, while Kravtsova (2006) found ingrowths in cell walls of both the pericarp and seed coat, observations that should be extended.

See also Staedler (1923: anther morphology, and in other families of this group), Bigalke (1933: cystoliths and hairs), Fagerlind (1944a: Elatostema - embryology, apomixis, quite a lot of variation in ovule), Miller (1971: general), Hegnauer (1973, 1990: chemistry), Berg (1978), Kubitzki (1993b: general), Friis (1993: general) and Kravtsova (2003: seed coat anatomy) for additional information.

Phylogeny. Urticaceae minus Cecropiaceae are paraphyletic (Sytsma et al. 2000, 2002 - three genes; Monro 2006 - two genes); the latter are polyphyletic and heterogeneous. Hadiah et al. (2008) found similar relationships, Poikilospermum being in a well-supported clade with Urtica, etc. However, Datwyler & Weiblen (2004 - one gene) and Zerega et al. (2005) find the reverse relationship, and strong support for Poikilospermum as sister to the rest of the family - [Poikilospermum [Cecropiaceae s. str. [rest of Urticaceae]]]. Either way, Cecropiaceae are best included in Urticaceae; I characterise them briefly below. Pellionia should probably be included in Procris (Hadiah et al. 2008). See Monro (2006) for suggestions as to how to proceed with the phylogenetic analysis of the large genus Pilea, and for a preliminary phylogenetic study of Elatostema, see Hadiah et al. (2003).

Ceropieae

Trees (hemiepiphytes); cystoliths spherical/0; bundle sheath extensions +; stomata anomocytic; hairs arachnoid, unicellular; leaves deeply lobed or compound; filaments straight.

5/147: Cecropia (61), Coussapoa (50). African and esp. New World tropics (Map: red, see Bonsen & ter Welle 1963).

Synonymy: Cecropiaceae C. C. Berg

CYNOMORIACEAE Lindley, nom. cons.   Back to Saxifragales

Echlorophyllous root parasite; vessel elements?; cork?; hairs 0; leaves spiral; plant monoecious, inflorescence capitate; flowers minute; P (1)4-5(-8), basally connate or not; staminate flowers: A 1, adnate to P, pollen colporate, ?nectary-stylodium +; carpellate flowers: staminodia 0; G 1, inferior, 1 pendulous straight unitegmic ovule, integument 5-7 cells thick, nucellar cap +?, style long, channeled; fruit an achene; testa ca 7 cells across, persistent, cells little thickened; endosperm cellular, copious, thick-walled, embryo undifferentiated; n = 12, size strongly bimodal.

1[list]/2. Mediterranean to C. Asia (map: from Jalas & Suominen 1976; Jäger et al. 1985; Hansen 1986). [Photo - Habit © D.L. Nickrent]

Evolution. In the Mediterranean area the host is often a member of Cistaceae or Amaranthaceae, elsewhere Cynomorium parasitizes Amaranthaceae, Tamaricaceae, Nitrariaceae, etc. (Jäger at al. 1985).

Chemistry, Morphology, etc. The root has root hairs. Perfect flowers are also known. The perianth is less well developed in pistillate than in staminate flowers, and there is debate as to its morphological nature. The pistillode in staminate flowers may be superior or inferior, according to Hooker (1856), clearly a matter that should be cleared up. The channeled style has two vascular bundles, together a rather odd combination.

For details of seed anatomy, see Takhtajan (2000), for morphology, see Weddell (1860), for ovule, etc., see Juel (1902) and Teryokhin et al. (1975), for some chemistry, see references in Zhang et al. (2009), and for general information, see the Parasitic Plants website (Nickrent 1998 onwards) and also Heide-Jørgensen (2008).

Previous relationships. Cynomoriaceae have usually been included in Balanophoraceae or Balanophoranae (e.g. Cronquist 1968; Takhtajan 1997), in Saxifragaceae, but with weak support, or their position has seemed to be completely uncertain (see above).