Plant woody, evergreen; nicotinic acid metabolised to trigonelline; primary cell walls rich in xyloglucans and/or glucomannans, 25-30% pectin [Type I walls]; lignins rich in guaiacyl units; true roots present, xylem exarch, branching endogenous; arbuscular mycorrhizae +; shoot apical meristem complex; 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 +; tracheid/tracheid pits circular, bordered; sieve tube/cell plastids with starch grains; phloem fibers +; stem cork cambium superficial, root cork cambium deep seated; nodes ?; stomata ?; leaf vascular bundles collateral; leaves spiral, simple, axillary buds?, prophylls [including bracteoles] two, lateral, veins -5(-8) mm/mm2; plant heterosporous, sporangia eusporangiate, on sporophylls, sporophylls aggregated in indeterminate cones/strobili; true pollen [microspores] +, 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, 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 duplication [N/O//A/C and P//BE lines], mitochondrial nad1 intron 2 and coxIIi3 intron present.


Plant woody, evergreen; lignans, O-methyl flavonols, dihydroflavonols, triterpenoid oleanane, non-hydrolysable tannins, quercetin and/or kaempferol +, apigenin and/or luteolin scattered, cyanogenesis via tyrosine pathway [ANITA grade?], lignins derived from both coniferyl and sinapyl alcohols, containing syringaldehyde [in positive Maüle reaction, syringyl:guaiacyl ratio less 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; stem with 2-layered tunica-corpus construction; wood fibers and wood parenchyma +; reaction wood ?, with gelatinous fibres; starch grains simple; primary cell wall mostly with pectic polysaccharides; tracheids +; sieve tubes eunucleate, with sieve plate, companion cells from same mother cell that gave rise to the tube, the sieve tube with P-proteins; nodes unilacunar; stomata with ends of guard cells level with aperture, paracytic; leaves with petiole and lamina [the latter formed from the primordial leaf apex], development of venation acropetal, 2ndary veins pinnate, fine venation reticulate, vein endings free; flowers perfect, polysymmetric, parts spiral [esp. the A], free, development in general centripetal, numbers unstable, P not differentiated, outer members not enclosing the rest of the bud, smaller than inner members, A many, with a single trace, introrse, filaments stout, anther ± embedded in the filament, tetrasporangiate, dithecal, 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, binucleate at dispersal, trinucleate eventually, tectum continuous or microperforate, exine columellar, endexine thin, compact, lamellate only in the apertural regions, pollen germinating in less than 3 hours, tube elongated, growing at 80-600 µm/hour, with callose plugs and callose-based walls, penetrating between cells, siphonogamy, penetration of ovules within ca 18 hours, distance to first ovule 1.1.-2.1 mm, nectary 0, G free, several, ascidiate, with postgenital occlusion by secretion, few [?1] ovules/carpel, ovules marginal, anatropous, bitegmic, [outer integument often largely subdermal in origin, inner integument dermal], micropyle endostomal, integuments 2-3 cells thick, megasporocyte single, megaspore lacking sporopollenin and cuticle, chalazal, female gametophyte four-celled [one-modular, nucleus of egg cell sister to one of the polar nuclei], stylulus short, hollow, stigma ± decurrent, dry [not secretory]; P deciduous in fruit; seed exotestal; double fertilisation +, endosperm diploid, cellular [first division oblique, micropylar end initially with a single large cell, chalazal end more actively dividing], copious, oily and/or proteinaceous, embryo cellular ab initio; germination hypogeal, seedlings/young plants sympodial; Arabidopsis-type telomeres [(TTTAGGG)n]; whole genome duplication, 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 PHYA + C/PHYB + E gene pairs.

Possible apomorphies are in bold. Note that the actual level to which many of these features, particularly the more cryptic ones, should be assigned is unclear, because some taxa basal to the [magnoliid + monocot + eudicot] group have been surprisingly little studied, there is considerable variation between families 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.

NYMPHAEALES [AUSTROBAILEYALES [[CHLORANTHALES + MAGNOLIIDS] [MONOCOTS [CERATOPHYLLALES + EUDICOTS]]]]: vessels +, elements with scalariform perforation plates; ?genome duplication; "DEAER" motif in AP3 and PI genes lost, gaps in these genes.

AUSTROBAILEYALES [[CHLORANTHALES + MAGNOLIIDS] [MONOCOTS [CERATOPHYLLALES + EUDICOTS]]]: ethereal oils in spherical idioblasts [lamina and P ± pellucid-punctate]; tension wood 0; tectum reticulate-perforate, nucellar cap + [character lost where in eudicots?]; 12BP [4 amino acids] deletion in P1 gene.

[CHLORANTHALES + MAGNOLIIDS] [MONOCOTS [CERATOPHYLLALES + EUDICOTS]] : benzylisoquinoline alkaloids +; P more or less whorled, 3-merous [possible position], carpels plicate; embryo sac bipolar, 8 nucleate, antipodal cells persisting; endosperm triploid.

MONOCOTS [CERATOPHYLLALES + EUDICOTS]: (A opposite [2 whorls of] P).

[CERATOPHYLLALES + EUDICOTS]: ethereal oils 0.

EUDICOTS: Myricetin, delphinidin scattered, asarone 0 [unknown in some groups, + in some asterids]; root epidermis derived from root cap [?Buxaceae, etc.]; nodes 3:3; stomata anomocytic; flowers (dimerous), cyclic, K/outer P members with three traces, "C" with a single trace, few, (polyandry widespread), filaments fairly slender, anthers basifixed, microsporogenesis simultaneous, walls developing by centripetal furrowing, pollen with endexine, tricolpate, G with complete postgenital fusion, style solid [?here]; seed coat?


TROCHODENDRALES [BUXALES + CORE EUDICOTS]: benzylisoquinoline alkaloids 0; euAP3 + TM6 genes [duplication of paleoAP3 gene: B class], mitochondrial rps2 gene lost.


CORE EUDICOTS: Ellagic and gallic acids common; micropyle?; PI-dB motif +, small deletion in the 18S ribosomal DNA common.

ROSIDS ET AL. + ASTERIDS ET AL.: root apical meristem closed; (cyanogenesis also via [iso]leucine, valine and phenylalanine pathways); flowers rather stereotyped: 5-merous, parts whorled, calyx and corolla distinct, stamens = 2x K/C, developing internal to the corolla whorl, (numerous, but then often fasciculate and/or centrifugal), pollen tricolporate, (nectary disc +), [G 5], [3] also common, compitum +, placentation axile, stigma not decurrent; endosperm nuclear; fruit dry, dehiscent, loculicidal [when a capsule]; euAP1 + euFUL + AGL79 genes [duplication of AP1/FUL or FUL-like gene], PLE + euAG [duplication of AG-like gene: C class], SEP1 + FBP6 genes [duplication of AGL2/3/4 gene].



CARYOPHYLLALES + ASTERIDS: seed exotestal; embryo long.

ASTERIDS - Sympetalae redux?: Nicotinic acid metabolised to its arabinosides; (iridoids +); tension wood decidedly uncommon; C sympetalous, if evident only early in development, petals free, anthers dorsifixed?, (nectary gynoecial), ovules unitegmic, integument thick, endothelium +, nucellar epidermis does not persist, style +, long; endosperm cellular, embryo long.

ERICALES [ASTERID I + II]: ovules tenuinucellate.

ASTERID I + II: Ellagic acid 0, proanthocyanidins not common; inflorescence cymose; C forming a distinct tube, A epipetalous, = and opposite sepals or P [polyandry (secondary) very uncommon indeed].

ASTERID II: myricetin 0; vessel elements with scalariform perforation plates; style short; flowers rather small, style short; endosperm copious, embryo short/very short.

ASTERALES [ESCALLONIALES [BRUNIALES [APIALES [PARACRYPHIALES + DIPSACALES]]]]: iridoids +; C tube initiation early, G [2-3], inferior.   Back to Main Tree

Various families have been placed here or elsewhere in the asterid II clade, but with uncertain support; recent work is clarifying relationships (Winkworth et al. 2008a). For Polyosmaceae and Escalloniaceae, see below, Paracryphiaceae, Quintiniaceae, and Sphenostemonaceae form a clade sister to Dipsacales, while Columelliaceae and Desfontainiaceae are perhaps close to Asterales, more particularly Bruniaceae.

ESCALLONIACEAE Dumortier, nom. cons.   Back to Unplaced

Plants Al accumulators; petals free, anthers basifixed, nectary ["disc"] +, integument 5-10 cells across.

9/130. Réunion, E. Himalayas and S. China to N.E. Australia and New Zealand, Central and South America.

The group is very poorly known; there is no information on endosperm development, etc. For general anatomy, see Gornall et al. (1998), for some information on ovules, see Mauritzon (1933).

A relationship between Escalloniaceae s. str. and Eremosynaceae has strong 3-gene support (Soltis et al. 2000; see also Hibsch-Jetter et al. 1997); an association of Tribelaceae with them is only weakly supported (rbcL alone analysed - Savolainen et al. 2000b). However, a three-gene Bayesian analysis suggested that Escalloniaceae were paraphyletic if Eremosynaceae and Tribelaceae were excluded (or Anopterus would have to be in a separate family), furthermore, the monogeneric Polyosmaceae were sister to the expanded Escalloniaceae (Lundberg 2001e). Lundberg (pers. comm.) suggests combining all into a single family, and so they are.


Trees; iridoids +; storying +; pericycle sclereidal; nodes 1:1; petiole bundle?; stomata ?; hairs unicellular; leaves opposite, margins toothed or not; inflorescence a raceme; flowers 4-merous; K connate, C valvate, free, pollen triporate, G [2], inferior, placentation intruding parietal, many ovules/carpel, integument to 10 cells across, ?endothelium, style +, stigma bilobed; fruit a 1-seeded drupe; seed coat to 10 cells thick; endosperm thick-walled, starchy, embryo undifferentiated; n = ?

1/60. E. Himalayas and S. China to N.E. Australia and New Caledonia.

Much information is taken from Lundberg (2001c); for information on embryology and seed, see Danilova (1996).

Escallonia, etc.

(Acaulescent) shrubs to trees (annual herbs); route I seco- and route II decarboxylated iridoids, flavonols +; (cork cambium deep-seated - Escallonia); nodes 3:3 (1:1, 5:5); petiole bundle arcuate; cells in heads of gland hairs radially arranged; stomata anomocytic; leaves spiral, supervolute (conduplicate), margins with broad glandular teeth and two accessory veins (entire); inflorescence a raceme (cymose - Eremosyne; flowers single, terminal - Tribeles); flowers 5-9-merous, anthers longer than connective, placentoid +, G [2-4], (largely superior), transverse, or median member adaxial, disc +, placentation also basal, parietal, 1 ascending-many (bitegmic - Tribeles) ovules/carpel, endothelium weak, integument 5-8 cells across, embryo sac protruding at micropyle, style long, ± divided or not, stigma capitate to clavate-lobed, wet; fruit a septicidal capsule, often splitting down the sides (indehiscent); exotestal cells with inner walls thickened and lignified (not), elongated or not (palisade - Tribeles), meso- and/or endotesta usu. persist; micropylar haustoria +, (embryo long); n = 12.


8[list]/68: Escallonia (40: stipular prickles occasional). Scattered, but largely southern: Central and South America, SE and SW Australia (Eremosyne pectinata), New Zealand, Réunion (Map: from Sleumer 1968). [Photos - Escallonia Flower, Valdivia Flower]

This is a very heterogeneous group. Eremosyne is an annual herb with leaf margins ± lobed and palmate-acrodromous venation, cymose inflorescence, small flowers with a valvate calyx, pollen with incomplete tectum and complex endaperture, two largely superior carpels with 1 ascending ovule/carpel. Its fruit is a capsule. Tribeles australis is a small, prostrate shrub with spirally-arranged leaves that have broad bases and three minute teeth at the apex. The small, terminal flowers have a contorted corolla and extrorse anthers. The shiny seeds long remain attached to the columella of the capsule. The pollen has interrupted muri.

Forgesia has sclereids alone in the pericyclic sheath. Anopterus has parietal placentation and winged seeds; the exotestal cells are elongated, with thickened inner walls. Escallonia lacks the mitochondrial coxII.i3 intron; other taxa have not been sampled.

As Bensel and Palser (1975d, p. 693, see also 1975c) noted of the Escallonioideae (Escalloniaceae here, also inc. Quintiniaceae), "The only conclusion that can be drawn about Escallonioideae is that it has been too little investigated in all aspects" - a conclusion that so far has unfortunately stood the test of time. Krach (1976, 1977) suggested that Escalloniaceae, including Abrophyllaceae and Argophyllum, were close to Hydrangeaceae, and should be placed in an Escalloniales (Escalloniaceae were placed in Hydrangeales by Takhtajan 1997); he wasn't sure about Anopterus. Eremosyne was previously frequently included in Saxifragaceae (e.g. Cronquist 1981) or Saxifragales (Takhtajan 1997). Tribeles was placed in Hydrangeales by Takhtajan (1997), who described the capsule as being loculicidal.

For anatomy of Forgesia (young stem with a complete vascular cylinder), see Ramamonjiarisoa (1980), and of Escallonia, see Stern (1974), for ovules, see Mauritzon (1933), for nodal anatomy, see Swamy (1954), for indumentum, Al-Shammary and Gornal (1994), for seed anatomy, Krach (1976) and Nemirovich-Danchenko and Lobova (1998), for embryology and testa of Eremosyne, Danilova (1996), for flowers, Ronse Decraene et al. (2000), for floral orientation, Schnizlein (1843-1870: fam. 170), for much general information, see Lundberg (2001).

Synonymy: Anopteraceae Doweld, Eremosynaceae Dandy, Tribelaceae Airy Shaw - Escalloniales Doweld, Tribelales Doweld - Escallonianae Doweld

A relationship between Apiales and Dipsacales has been suggested by Nandi et al. (1998), Lundberg (2001c), Winkworth et al. (2008), etc.

APIALES Nakai  Main Tree, Synapomorphies.

Woody; route II decarboxylated iridoids +; vessels solitary; pits [both vessels and fibres] bordered; ?stomata; hairs both uni- and multicellular; leaves pinnatinerved (margins toothed or lobed); inflorescence terminal, paniculate; plants dioecious; pedicels articulated; flowers small, C apparently free, A free from C, G [3], abaxial carpel fertile [?Pennantia], 1-2 apical apotropous ovules/carpel, nucellus type?, funicular obturator +; fruit a single-seeded drupe; endosperm nuclear; x = 6?; mitochondrial rpl2 gene lost. - 7 families, 494 genera, 5489 species.

Apiales contain ca 2.4% eudicot diversity (Magallón et al. 1999). Wikström et al. (2001) suggest a Santonian-Turonian age for the clade of some 90-85 million years before present, with divergence in the Apiaceae-Araliaceae area occuring 48-46 million years before present, although details of relationships within the clade differ from those given here.

The basal pectinations are genera which used to be in Cornales/Cornanae - and they have transseptal bundles in common, as well as clustered vessels, septate fibers with simple pits, some rays over 10 cells wide and with square or upright cells (Noshiro & Baas 1998; Baas et al. 2000) - as well as the even more different Pennantia, late of Icacinaceae. Thus Melanophylla was included in Hydrangeales and Torricellia, Griselinia and Aralidium in separate monogeneric orders, all in Cornidae-Cornanae, by Takhtajan (1997); all have been placed in Cornaceae s.l. Unfortunately, corolla initiation, pollen nucleus number, and many other characters are unknown for these taxa, although some information for the ex-Cornalean genera can be found in Patel (1973), Philipson (1967), and Philipson and Stone (1980). Griseliniaceae and Torricelliaceae both have griselinoside, transseptal bundles in the ovary, and the abaxial carpel alone is fertile, the single ovule being apotropous. Pennantia, sister to all other Apiales, has a superior ovary (and sometimes a thick, disciform, sessile stigma)... In particular, the interpretation of the gynoecium of Pennantia (summarised in Kårehed 2003a) has many implications for character evolution. Here it is interpreted as being tricarpellate, although appearing to consist of a single carpel, and it is assumed(!) that the abaxial carpel is fertile (see also Chandler & Plunkett 2004; Plunkett et al. 2004c for carpel number). Add the problem of the unclear outgroup to Apiales, and character evolution in the basal part of Apiales in particular becomes very uncertain. Thus of taxa close to Apiales, Sphenostemonaceae, of the Quintiniaceae group, has a superior ovary, and Adoxaceae, sister to other Dipsacales, the ovary is only semi-inferior (and in Tetradoxa the ovary was described as being superior - Ying et al. 1993).

There are similarities between [Pittosporaceae [Araliaceae [Myodocarpaceae + Apiaceae]]] and some Asterales (esp. Campanulaceae s.l. and Asteraceae) (e.g. Erbar & Leins 2004; Leins & Erbar 2004b). However, given the rather strongly supported set of relationships at the base of Apiales, the other families that are in this part of the phylogeny, and relationships within Asterales (Campanulaceae s.l. and Asteraceae are not immediately related), most of these similarities will probably be plesiomorphies or more likely parallelisms (see also below).

Within Apiales, the grouping (Griselinia (Aralidium + Torricellia)) is rather weakly supported, mainly in earlier studies (Backlund & Bremer 1997; see also Chandler & Plunkett 2002). The ex-Icacinaceae Pennantia is sister to other Apiales (Kårehed 2003; Lens et al. 2008a). Plunkett (2001) and Lundberg (2001c) both suggest that Torricelliaceae and Griseliniaceae are successive clades near the base of Apiales, and this has strong Bayesian support (Kårehed 2002a: four genes, all genera sampled), and this topology is followed here (note that the relationships is reversed in, but with a p.p. of 0.93). Chandler and Plunkett (2004) and especially Tank et al. () have begun to resolve relationships both within Apiaceae and between Apiaceae and their immediate relatives (see also tree). Some Bayesian analyses yield strong posterior probabilities for a relationship between Myodocarpaceae and Pittosporaceae (Chandler & Plunkett 2004), but on balance the relationships [Pittosporaceae [Araliaceae [Myodocarpaceae + Apiaceae]]] seem most likely.

Kårehed (2003, for which see for further details, also Chandler & Plunkett 2004) provides a good summary of what is known of the main clades in Apiales, although as just mentioned optimising character state change on the tree is not simple, even given its current constitution - see, for example, perforation plate variation. For wood anatomy, see Lens et al. (2008a). Endress (2003c) summarises the literature on nucellus development in the clade; there are a number of glaring gaps in our knowledge. Yi et al. (2004) suggest that the basic chromosome number in Apiales is x = 6, although the order then would have all polyploid/dysploid members, with several independent origins of polyploidy.

Includes Apiaceae, Araliaceae, Griseliniaceae, Myodocarpaceae, Pennantiaceae, Pittosporaceae, Torricelliaceae.

Synonymy: Apiineae Plunkett & Lowry, Aralidiinae Thorne & Reveal - Ammiales Small, Araliales Reveal, Aralidiales Reveal, Griseliniales (J. R. Forster & G. Forster) Reveal & Doweld, Pennantiales Doweld, Pittosporales Lindley, Torricelliales Reveal & Doweld - Aralianae Takhtajan

PENNANTIACEAE J. Agardh   Back to Apiales

Trees or shrubs; iridoids ?; diffuse apotracheal parenchyma +/0; nodes 3:3; petiole bundle annular (with medullary bundle) and with solid wing bundles; stomata paracytic; hairs uniseriate; leaf margins also entire; K minute, free, C connate, valvate, apex inflexed, disc/nectary 0: staminate flowers: A dorsifixed (epipetalous), pistillode +; carpellate flowers: staminodes +/0, G also [2], nucellus?, styles short, stigmas punctate, or stigma sessile, broad; drupe with inner cells transversely elongated; seed coat thin; endosperm development?, embryo short/minute [ca. 1/3 the seed length]; n = 25.


1/4. N.E. Australia, Norfolk Island, New Zealand (Map: from van Balgooy 1966; George 1984). [Photo - Habit.]

Collenchyma is only poorly developed, and a pericyclic sheath is present; pits in general are bordered. For the morphology of the stigma, see Kårehed (2002b).

For other information, both on Pennantiaceae and other members of the old Icacinaceae (here in Garryales and Aquifoliales), see Miers (1852: ovule orientation), Bailey and Howard (1941: anatomy); see also Fagerlind (1945: embryology), Heintzelmann and Howard (1948), Padmanabhan (1961: embryology), Sleumer (1971a: general), van Staveren and Baas (1973: epidermis), Baas (1973: epidermis, 1974: stomata), Lobreau-Callen (1980: pollen), Kaplan et al. (1991: chemistry), Teo and Haron (1999: anatomy), Kårehed (2001, 2002b, 2003: the taxa in their current circumscription), and Lens et al. (2008a: wood anatomy). Gardner and de Lange (2002) monographed the genus, providing a number of interesting morphological and anatomical details.

Torricelliaceae [Griseliniaceae [Pittosporaceae [Araliaceae [Myodocarpaceae + Apiaceae]]]]: young stems with peripheral collechyma; vessels mainly in groups; vessel element perforations simple; pits in general with at most narrow borders; paratracheal parenchyma +, scanty; pericyclic fibers 0 or few; nodes 5(+):5(+); leaf bases broad; C apparently free, imbricate, styles/stigmas recurved.

For the wood anatomy of this clade, see Lens et al. (2008a).

TORRICELLIACEAE H. H. Hu   Back to Apiales

Trees; griselinoside, ellagic acid +, tanniniferous [Torricellia]; (vessel elements with scalariform perforation plates - Aralidium); petiole bundles scattered [Aralidium]; crystal sand +; mucilage cells +; stomata anomocytic (anisocytic); glandular hairs +; leaf margins lobed and toothed or toothed, (2ndary veins ± palmate), leaf bases with petiole margin (+ small adaxial) flange, encircling stem or not; K imbricate or largely connate, staminate flowers: (C induplicate-valvate - Torricellia), pollen tectum reticulate; carpellate flowers: (C 0 - Torricellia), G [(2-4)], (disc vascularised - Aralidium), transseptal bundles +, integument massive [Aralidium], (stigmas ± bifid - Torricellia); endocarp with sclereids; seed ± curved, (ruminate, coat vascularised [Aralidium]), exotesta scalariform-thickened, unlignified [Melanophylla], testa slightly sclerified [Torricellia]; (embryo long, thin - Torricellia); n = 12, 20 ± 2.


3[list]/10. Madagascar, South East Asia and W. Malesia. [Photo: Melanophylla Habit, Flower.]

For fruit and seed of Melanophylla, see Trifonova (1998), for the circumscription of this clade, see also Plunkett et al. (2004). Other information is taken from from Philipson (1977), Philipson and Stone (1980), and Takhtajan (2000).

Synonymy: Aralidiaceae Philipson & B. C. Stone, Melanophyllaceae Airy Shaw

Griseliniaceae [Pittosporaceae [Araliaceae [Myodocarpaceae + Apiaceae]]]: petroselenic acid + [in seed: CH3-[CH2]10-CH=CH-[CH2]4-COOH]; hairs rarely glandular; petals 1-veined, nectary of sorts +, endothelium +.

GRISELINIACEAE A. Cunningham   Back to Apiales

Trees or shrubs (climbing); griselinoside +, tannins 0; vessels single, vessel element perforations scalariform; pits in general with distinct borders; diffuse-in-aggregates axial parenchyma +, scanty; petiole bundles (incurved) arcuate; mucilage cells +; stomata cyclocytic; hairs unicellular; leaves two-ranked, conduplicate, margins toothed or entire, bases with petiole margin + adaxial flange, encircling stem, or not; K minute, open; staminate flowers: A dorsifixed, pollen tectum striate, carpellate flowers: (C 0), G [(4)], disc +, transseptal bundles +, (2 carpels fertile), ovule crassinucellate [weakly so]; fruit baccate; testa ?many layered, outer two (and esp, third) layers with thickened walls; embryo long; n = 18.


1[list]/6. New Zealand and S. South America (Map: from Dillon & Muñoz-Schick 1993). [Photo: Inflorescence, Leaves, Habit, Flower (scroll to end).]

Eyde (1964) suggested a relationship with Garrya and Cornaceae. The wood has solitary vessels and apotracheal parenchyma (Baas et al. 2000).

Takhtajan (2000) provides details of embryo and testa. For leaf insertion, cf. Philipson (1967), and for leaf base, cf. Takhtajan (1997). For Griselinia flowers, see Eyde (1964). Further information is taken from Philipson (1977) and Dillon and Muñoz-Schick (1993).

Pittosporaceae [Araliaceae [Myodocarpaceae + Apiaceae]]: plant often aromatic, ethereal oils, polyacetylenes [mainly aliphatic, the C17 acetylenes, falcarinone, etc.], acetate-derived anthroquinones, coumarins +, iridoids, flavonols, tannins 0; lateral roots originating from either side of the xylem poles; schizogenous secretory canals in pericycle; (nodes 3:3); petiole bundles arcuate or annular; leaves conduplicate; flowers perfect and imperfect [plant andromonoecious or dioecious]; C tube formation early, pollen trinucleate, G [2], both fertile, stigma wet; hemicellulosic seed reserves common [?right place]; x = 12.

Lateral roots originate from either side of the xylem poles because a resin canal runs down the stem at the apex of the pole (Van Tieghem & Douliot 1888). Triterpenoid ethereal oils produce the distinctive odour characteristic of many of these plants; see Jay (1969) for suggestions based on plant chemistry that they are related. Which groups are food plants for butterfly larvae may be of interest; Ehrlich and Raven (1964) note that some do not seem to like Araliaceae, including Hydrocotyle, but they do like Apiaceae (?including Saniculoideae); the Papilio machaon group is not found on Araliaceae because that family lacks the furanocoumarins that the butterfly likes (Berenbaum 1983).

There are three sets of related issues when thinking about this group. One is exactly what its synapomorphies might be, another is the position of the Pittosporaceae, and the third concerns the splitting of the old Apiaceae/Araliaceae.

One. The following characters may also be synapomorphies: sesquiterpene lactones +, benzylisoquinoline alkaloids scattered; (inverted vascular bundles in the pith and/or cortex); (crystal sand +); palisade mesophyll often with arm cells; exotestal cells tangentially elongated, tanniniferous. As to other characters, the vessel:ray pits are bordered (Baas et al. 2000). Members of both Araliaceae s. str. and Apiaceae s. str. in clades that are sister to the rest of these families have simple leaves, as have Myodocarpaceae (Plunkett 2001) and of course Pittosporaceae, so compound leaves may have evolved independently in Apiaceae and Araliaceae. Leaf teeth have a broad glandular apex with a main and two accessory veins, or one vein proceeds above the tooth; a survey of tooth morphology might be interesting. For information on ventral carpel bundles, see Philipson (1970) and Eyde and Tseng (1971); when the united ventral bundles are opposite the carpels, the two bundles come from the same carpel, and when they are in the septal radii, the two bundles are derived from adjacent carpels. For ovule morphology, see Jurica (1922) and van Tieghem (1898). According to the latter, Apiaceae have a thick integument (e.g. ca 7 cells - Gupta 1970), that of Hedera is very thick, although Philipson (1977) described the integument of Araliaceae as being "thin". A synthesis of what is known about ovule number and type would be useful (but see Philipson 1970 for ovule number). Note that there is not only variation in what kinds of calcium oxalate crystals are found in the fruit wall, i.e., single rhomboidal crystals or druses, but also where they occur, e.g. in the endocarp or mesocarp, all around the fruit or only in commissural area (Liu et al. 2006); I have only just begun to work out the phylogenetic significance of the later variation (see also Rompel 1895; Burtt 1991a). Kleiman and Spencer (1982) surveyed Apiaceae and Araliaceae for the occurence of petroselenic acid.

Two. Pittosporaceae may be sister taxon of the whole group (e.g. Kårehed 2002c; Andersson et al. 2006, strong support, but Myodocarpaceae not included), or they may be embedded in the group (see also Chandler & Plunkett 2004). Whatever its position, it is likely that many of the apparently plesiomorphous characters of Pittosporaceae, e.g. 'normal'-looking entire leaf blade and petiole with rather narrow insertion, superior ovary, etc., are derived. Note that the leaves of seedlings of Pittosporum may be pinnatifid. Some Pittosporaceae and a few Araliaceae have basally connate petals (Plunkett 2001). The nectary "disc" of Apiaceae + Araliaceae is a carpellary flank nectary displaced by intercalary growth, and the axile placentation of these two families is easily related to the parietal placentation of Pittosporaceae; in the latter, there may be a short basal zone with separate loculi, the ovules being borne above this zone - essentially the same position in which the ovules of Apiaceae + Araliaceae are found (Erbar & Leins 1988, 1996a, 2004). However, gynoecial development, etc., in the whole Apiales needs reinvestigation, given the inclusion of the superior-ovaried Pennantiaceae in it as sister to the rest of the order (and taxa with superior ovaries are scattered throughout the [Asterales [Apiales + Dipsacales]] clade). The floral morphology of Pittosporaceae is quite probably derived, rather than representing the plesiomorphic condition from which the distinctive flowers of [Apiaceae + Araliaceae] evolved.

Three. Finally, there is the issue of the demise of the old woody Araliaceae/herbaceous Apiaceae distinction. Myodocarpaceae and Apiaceae-Mackinlayoideae as recognised here both include genera that used to be in Araliaceae, but the precise relationships of the former in particular is uncertain - see Plunkett and Lowry (2001), Lowry et al. (2001), Plunkett (2001) and Chandler and Plunkett (2004) for more information. Since there are a number of morphological similarities between the Mackinlaya clade and Apiaceae s. str. (Chandler & Plunkett 2002, 2004), the former is included in the latter. Hydrocotyloideae (herbaceous and with simple leaves) were previously included in Apiaceae, but they are polyphyletic - Hydrocotyle is here placed in Araliaceae (and this position makes morphological sense), Arctopus is in Apiaceae-Saniculoideae, while Centella, Micropleura, Actinotus, etc., are probably in Apiaceae-Mackinlayoideae (e.g. Downie et al. 1998, Downie et al. 2000; Chandler & Plunkett 2003, 2004; Andersson et al. 2006); only a few genera remain to be sequenced, although sampling must be improved in Hydrocotyle and Trachymene.

There are a number of similarities between this group of families and the Asterales, perhaps especially both/either Campanulaceae s.l. and Asteraceae (e.g. Erbar & Leins 2004; Leins & Erbar 2004b). However, given the rather strongly supported set of relationships at the base of Apiales and the other families that are somewhere in this part of the phylogeny, families which for the most part have a very different set of morphological/chemical features, most will probably be plesiomorphies or parallelisms.

For variation in the sequence of initiation of parts of the flower, see Erbar and Leins (1985: mostly Apiaceae, also Hydrocotyle, the latter rather different - and lacking a calyx!), for chromosome numbers, see Yi et al. (2004), and for fruit wings and fruit anatomy, see Liu et al. (2006). For other work on phylogenetic relationships in the Apiaceae-Araliaceae area, see e.g. Judd et al. (1994: Pittosporaceae not included), Oskolski et al. (1997), Oskolski & Lowry (2000), Plunkett (1998), Plunkett et al. (1996, 1997a, 1997b), and Kårehed (2002c).

PITTOSPORACEAE R. Brown, nom. cons.   Back to Apiales

Trees or twining vines; furanocoumarins +, (hydroxycoumarins), proanthocyanins +, petroselenic acid 0, C20 and C22 fatty acids abundant; young stem with (± interrupted - Pittosporum) vascular cylinder; nodes 1:3, 3:3; petiole bundles arcuate; stomata paracytic; hairs uniseriate, terminal cells vertically or transversely [T-shaped hair] elongated or glandular; leaf ptyxis supervolute-curved, margins entire, 2ndary veins pinnate, base narrow to sheathing, stipules 0; flowers medium-sized; K large, free (± connate), C often slightly basally connate, 3-5-veined, anthers ± basifixed, placentoid +, G [(3-5)], flank nectaries +, many (incompletely tenuinucellate; apotropous) ovules/carpel, endothelium 0, funicular obturator 0?, style +, straight, stigma capitate (lobed); fruit a loculicidal (+ septicidal) capsule or berry, K deciduous; seeds often with sticky pulp, testa many layered, exotestal cells thickened, unlignified.


6-9[list]/200: Pittosporum (140). Old World, especially Australia, tropical and warm temperate, outside the immediate Australian region, mainly Pittosporum (Map: from van Steenis & van Balgooy 1966; Good 1974; Palgrave 2002). [Photo - Flower]

Glucosinolates are reported from Bursaria spinosa, but this is probably a mistake (Fahey et al. 2001). In Pittosporum, at least, the flowers are often functionally unisexual. There is a tendency, especially evident in taxa like Cheiranthera, for the flowers to be obliquely monosymmetrical; asymmetry is largely because of the position of the stemans and gynoecium. Seedlings of Pittosporum have up to five cotyledons.

Cayzer (references in Chandler et al. 2007) has carried out a number of morphological revisions and phylogenetic analyses of parts of Pittosporaceae; changes in taxon limits that she suggests are largely confirmed by a preliminary molecular study (Chandler et al., in Plunkett et al. 2004c). However, generic limits in the Billardiera-Sollya area are still unclear (Chandler et al. 2007). For relationships between species of Pittosporum found in the Pacific, see Gemmill et al. (2002).

Pittosporaceae are included in Rosales by Cronquist (1981) and Mabberley (1997), however, there has been mounting evidence for over 100 years that they were best associated with Apiaceae/Araliaceae (Hegnauer 1969b for a summary of some literature).

For endothelium, see Batygina et al. (1985), for anatomy, Wilkinson (1992, 1998), for testa anatomy, Takhtajan (2000: the ovules may be apotropous).

Araliaceae [Myodocarpaceae + Apiaceae]: polyacetylene C18 tariric acid +; leaves (compound), margins toothed or otherwise incised; ultimate units of inflorescences umbels; K small, nectariferous stylopodium +, two epitropous ovules/carpel, one much reduced; fruits ± laterally compressed; 92 bp deletion in rpl16 gene.

Chemistry, Morphology, etc. For wood anatomy, see Oskolski (2001), and for the rpl16 deletion, see Downie et al. (2000a).

ARALIACEAE Jussieu, nom. cons.   Back to Apiales

Hydroxycoumarins +, furanocoumarins 0; stomata para- or aniso-(anomo-)cytic; hairs often stellate or dendroid; stipules cauline or petiolar, or 1 intrapetiolar, or 0; C valvate (imbricate - Aralia, Panax, etc.; A 3, many), ventral carpel bundles are fused bundles of adjacent placentaa (the same - e.g. Harmsiopanax), ovules crassinucellate to tenuinucellate, (nucellar cap +), integument "thin" [?level], (endothelium +), funicle with short hairs [cf. obturator], stigma punctate (dry); fruit drupaceous or dry, mesocarp thick-walled, lignified, (berry), with single rhomboidal crystals in cell layer immediately outside endocarp; (seeds ruminate), exotestal cell walls a little thickened; (n = 9, 11, 18).

43/1450. Largely tropical, few temperate. Two groups below.

1. Hydrocotyloideae Link

± herbaceous; nodes 3:3; leaves rounded-peltate, margin crenate, stipules cauline; (K 0), funicle "short"; (undivided carpophore +).

4/175: Hydrocotyle (130), Trachymene (45). Tropical, including montane, also warm temperate.

Synonymy: Hydrocotylaceae N. Hylander

2. Aralioideae Eaton

Woody (herbaceous); leaves also pinnately to palmately compound; G [1-5(-200+)], when 5, opposite petals, when 3, median abaxial.


41[list - but see Plunkett et al. 2004a for genera]/1275: Schefflera (650: wildly polyphyletic), Polyscias (150: paraphyletic), Oreopanax (80), Dendropanax (70), Aralia (68), Osmoxylon (50). Largely tropical, few temperate (Map: see Meusel et al. 1978; Hultén & Fries 1986; FloraBase 2006). [Photo - Flowers, Fruits.]

Synonymy: Hederaceae Giseke, Botryodendraceae J. Agardh

Evolution. For shifts in the rate of molecular evolution within Araliaceae that are correlated with changes in habit, see Smith and Donoghue (2008).

Chemistry, Morphology, etc. For the anatomy of flower and fruit of Hydrocotyle and genera previously asssociated with it, see Tseng (1967); Hydrocotyle lacks a calyx, and its integument is ca 5 cells across. Hedera may have an interrupted fibrous pericyclic sheath; creeping forms have two-ranked leaves. As to stipules - Fatsia (no stipules) x Hedera (no stipules) = X Fatshedera (stipules) - but note that some species of Hedera have a hollowed leaf base with a stipule-like margin, the whole enclosing the bud... Both Aralioideae and Hydrocotyle show early corolla tube initiation (Leins & Erbar 1997; Erbar & Leins 2004). Osmoxylon has basally connate petals. Plerandra (= Schefflera) may have up to 25 or more carpels and to 500 stamens; flowers in Aralioideae may be up to 12-merous. When stamens are numerous they are often in a single whorl, thus Tupidanthus calyptratus has up to 172 stamens and 132 carpels; the carpels are initiated in a single elongated and sometimes contorted whorl looking rather like a brain cactus (fasciated!: Oskolski et al. 2005; see also Eyde & Tseng 1971). In Plerandra there are about as many stamens as carpels, separate members of the calyx cannot be distinguished, and although the symplicate zone of the gynoecium develops first, the carpels are largely synacidiate; there ia a large, flat remnant of the floral axis within the carpel whorl (Sokoloff et al. 2007b). In Plerandra there may be up to four series of stamens (Philipson 1970). Tetraplasandra gymnocarpa and T. kavaiensis have secondarily more or less completely superior ovaries (Costello & Motley 2000, 2001, 2004 - see photograph on the cover of American J. Bot. 91(6) 2004).

For general information, see Philipson (1970), for wood anatomy, see Oskolski (1996).

Phylogeny. Basal Araliaceae may well be bicarpellate (see also Wen et al. 2001) and have simple leaves. Both these are features of the herbaceous Hydrocotyloideae (ex Apiaceae), which are sister to the rest of the family (Chandler & Plunkett 2004; Plunkett et al. 2004a). The S.W. Australian Neosciadium and probably also Homalosciadium belong to Hydrocotyloideae (Andersson et al. 2006). Hydrocotyle has laterally flattened fruits with a sclerified (= woody) endocarp and stipules that are either cauline or are borne on the leaf base. Trachymene may also be placed near or in Hydrocotyloideae, although only one species has been sampled so far; morphologically it is rather similar to them, and although there is a carpophore, it is undivided. Astrotricha and Osmoxylon may be part of a polychotomy at the node immediately above.

For more details on the phylogeny of Aralioideae in particular, see Henwood and Hart (2001) and especially Wen et al. (2001), Plunkett et al. (2004a, c), and Lowry et al. (2004); it is clear that generic limits badly need re-evaluation. Although Schefflera is apparently polyphyletic, there is no strong support for particular positions for each of the groupings into which it breaks. These are circumscribed geographically, and in large part represent earlier infrageneric groupings (Plunkett et al. 2005).

For a checklist, see Frodin and Govaerts (2003).

Myodocarpaceae + Apiaceae: furanocoumarins +.

MYODOCARPACEAE Doweld   Back to Apiales

Plants woody; ?coumarins; apotracheal (and paratracheal) parenchyma +; leaves pinnately compound (simple), margin entire (serrate), venation brochidodromous; inflorescence panicles or racemes, often of umbels, pedicels articulated; K valvate, C imbricate, A inflexed in bud, nucellus?, ventral carpel bundles fused bundles of adjacent placentae; fruits terete, fleshy, (dry, winged, seeds laterally flattened - Myodocarpus), endocarp with large oil ducts [different from vittae].


2 (Myodocarpus, Delarbrea)/19. New Caledonia, E. Malesia, and Queensland, Australia (Map: from van Balgooy 1993). [Photo - Habit]

Chemistry, Morphology, etc. Myodocarpus has a number of distinctive floral and in particular fruit features (compound umbel; schizocarp), the latter of which may be associated with wind dispersal. Indeed, the mericarps are beautiful little samaras that at first sight are similar to those of Serjania! In wood anatomy Myodocarpus is perhaps more like Cornaceae than any other members of the Apiaceae-Araliaceae complex, but in other features it is more like Apiaceae (Rodrigues C. 1957).

Some information is taken from Lowry (1986), Raquet (2004), and Plunkett et al. (2004b).

Previous Relationships. This group used to be included in Araliaceae - Myodocarpeae.

APIACEAE Lindley, nom. cons.//UMBELLIFERAE Jussieu, nom. cons. et nom. alt.   Back to Apiales

Plant usu. herbaceous; pyranocoumarins, myricetin, mannitol +, umbelliferose [raffinose (trisaccharide) isomer] the storage carbohydrate; hydroxycoumarins, flavones 0; vessel elements aggregated; axial parenchyma scanty, paratracheal; stomata various; leaf also supervolute, base ± encircling stem; umbels often compound; (outer flowers of umbel monosymmetric), K reduced to ring of teeth (obsolete), C free, clawed, valvate, usu. with inflexed tips, vein single, unbranched [?all], A inflexed in bud, ventral carpel bundles are fused bundles of the same placenta, funicle "short", stigma usu. capitate; fruit dry (fleshy), schizocarpic, mesocarp lignified, crystals single, rhombic, in small-celled inner layer, endocarp woody, 2(+) cell layers thick, sclereidal-fibrous, carpophore +, not free; exotestal cells thin-walled.

434[approx. list]/3780 - 4 groups below. World-wide, esp. N. temperate.

1. Mackinlayoideae Plunkett & Lowry

Plants woody (herbaceous); (centellose [oligosaccharide] - Centella); apotracheal (+ paratracheal) parenchyma; (leaf irregularly palmately compound - Mackinlaya); inflorescence panicles or racemes, often of umbels; (pedicel not articulated); nucellus?; (carpophore +).

6/67: Centella (40), Actinotus (18). Most S. Pacific Rim, Centella esp. S. Africa, C. asiatica pantropical.

Some genera are ex Araliaceae - Mackinlayeae, others ex Apiaceae - Hydrocotyloideae - other genera are Apiopetalum, Mackinlaya, Micropleura, Xanthosia. See Melikian and Konstantinova (2006) for discussion on the gynoecial structure of Actinotus which they consider to be so different from that of other Apiaceae that it merits the genus being placed in its own family.

Synonymy: Actinotaceae Konstantinova & Melikian, Mackinlayaceae Doweld

Azorelloideae [Saniculoideae + Apioideae]: fruits dorsally flattened; irregular anastomosing and/or branching mesocarp vittae +.

2. Azorelloideae Plunkett & Lowry

(Cork cambium deep-seated - Mulinum); (leaves compound), stipules +; nucellus "large", 2-4 megasporocytes, embryo sac tetrasporic, 16-nucleate [more than one embryo sac "type"], nucellus relatively persistent; (wings + [made up of the entire fruit wall, including a vascular bundle]), (druses in outer mesocarp - Azorella), companion cells [oil canals associated with the vascular bundles] +; x = ?8.

21/155: Azorella (70: ?inc. Laretia, Mulinum). South American-Australian, Antarctic islands; Drusa glandulosa from the Canary Islands and Somalia! (Map: from Bramwell 1972). [Photo - Habit.]

For the composition of this clade, see Downie et al. (1998, 2001), Mitchell et al. (1999), Plunkett & Lowry (2001), Henwood and Hart (2001: the Bowlesia clade), Chandler and Plunkett (2004) and particularly Andersson et al. (2006), which see for details. It includes Asteriscium, Azorella, Bolax, Bowlesia, Dichosciadium, Dickinsia, Diplaspis, Diposis, Domeykoa, Drusa, Eremocharis, Gymnophyton, Homalocarpus, Huanaca, ?Laretia, ?Mulinum, Oschatzia, Pozoa, Schizeilema, Spananthe, Stilbocarpa (this used to be in Araliaceae; it is probably sister to the Andean Huanaca. Bowlesia lacks petroselenic acid - it was apparently the only genus of this group studied (Kleiman & Spencer 1982).

Saniculoideae + Apioideae: basal leaf with pinnate venation [?level], stipules 0; ovules tenuinucellate [incompletely so], nucellar cap +, funicle "long"; secondary [i.e. lateral] ribs +, mesocarp cells lignified, endocarp single cell layer thick, parenchymatous, (walls lignified), calcium oxalate as druses dispersed in mesocarp and around commissure, intrajugal vittae [oil ducts in the ribs] + [?level].

3. Saniculoideae Burnett

Kaurene-type terpenoids +; inflorescence apparently a simple umbel (a capitulum - Eryngium; pseudoracemose - Sanicula) [probably reduced compound umbels], with showy inflorescence bracts; style separated from disc by a narrow groove; intrajugal secretory ducts/cavities +; cotyledons rounded.


10/335. World-wide (Map: see Meusel et al. 1978; Hultén & Fries 1986).

3a. Steganotaenieae C. I. Calviño & S. R. Downie

Plant woody; fruits 2-3 winged [wings exo- and mesocarp alone], intrajugal secretory ducts/cavities associated with the wings much expanded, carpophore +, (dispersed mesocarp druses 0); n = ?

2/2-3. Tropical Africa, S.W. Cape.

3b. Saniculeae Burnett

(R)-3'-O-ß-D-glucopranosylrosmarinic acid +; leaves blades often broad, (compound), with hairy or spiny tips to the teeth; (flowers blue), carpellate flowers sessile; fruit scaly or spiny, carpophore 0, (intrajugal secretory ducts/cavities 0), (dispersed crystals throughout mesocarp 0), endocarp not lignified; n = 8 (9, 11, 12).

8/333: Eryngium (250). World-wide (Map: see Meusel et al. 1978; Hultén & Fries 1986).

Synonymy: Eryngiaceae Rafinesque, Saniculaceae (Burnett) A. & D. Löve

4. Apioideae Seemann

Flavones, methylated flavonoids, furanocoumarins, phenylpropenes +; leaves usu. compound; umbels compound; hypostase +, stylopodium lacking groove; carpophore free, bifid [mericarps attached at apex], intrajugal oil ducts 0, vallecular vittae +, (+, 0 - Lichtensteinia), [crystals on commissural side of mericarp only, or none [although these characters are common, it is unclear which, if any, is an apomorphy]), endocarp not lignified; x = 11; cotyledons various.

400/3200: Bupleurum (190), Ferula (175), Pimpinella (150), Angelica (110), Seseli (110), Peucedanum s.l. (110), Lomatium (74), Heracleum (65), Chaerophyllum (60), Arracacia (55), Ligusticum (50). Worldwide, esp. N. temperate (Map: see Meusel et al. 1978; Hultén & Fries 1986).

Synonymy: Ammiaceae Barnhart, Angelicaceae Martynov, Bupleuraceae Martynov, Coriandraceae Burnett, Daucaceae Martynov, Ferulaceae Saccardo, Imperatoriaceae Martynov, Pastinacaceae Martynov, Smyrniaceae Burnett

Evolution. Caterpillars of Papilionideae butterflies are notably common (ca 13% of all records) on Apiaceae, but not on either Pittosporaceae or Araliaceae; thus they will not eat Hydrocotyle (here Araliaceae), many of the larvae of Papilio ajax tested absolutely refusing to eat it (see also Dethier 1941; Ehrlich & Raven 1967). Microlepidopteran larvae of the Elachistidae - Depressariinae are common on Apiaceae, although thay may have initially been associated with rosids and have also colonized some Asteraceae (Fetz 1994 for host plants; Berenbaum & Zangerl 1998 for the chemistry of the [co]evolution of resistance; Berenbaum & Passoa 1999 for a phylogeny). For general insect-umbellifer relationships, see Berenbaum (1990) and Sperling and Feeny (1995).

Apioideae are likely to have evolved in Southern Africa, but although a number of taxa from there have a woody habit, the ancestral condition for this subfamily, at least, is likely to be herbaceous (Calviño et al. 2006).

Variation in leaf morphology, even within Apioideae, is considerable. Thus the small circum-Pacific genus Oreomyrrhis includes species with ordinary-looking highly dissected leaves, leaves with a series of almost tooth-like lealets on either side of the rachis, linear leaves, sometimes lobed at the apex, although the lobes are not articulated, and small, undivided leaves, in the last case the plant being tussock-forming and almost moss-like (plants have also been identified as Centrolepidaceae, a monocot!), and all are found on the mountains of New Guinea. The genus is probably monophyletic, but it is well embedded in Chaerophyllum, a genus hitherto thought to be fairly well understood (Chung et al. 2005; Chung 2007). Lilaeopsis occidentalis and Oxyopolis greenmanii have linear, terete leaves that are marked by articulations at intervals. These leaves are comparable with the rhachis of compound leaves, hydathodes borne at the articulations being much reduced and modified pinnae (Kaplan 1970b, cf. esp. Figs 3E and 6A). Other North American taxa have similar leaves, and they appear to have evolved several times (Feist & Downie 2008). Some species of Bupleurum have almost orbicular, entire, perfoliate leaves, hence the common name, thorow wax.

Chemistry, Morphology, etc. Gums and resins are scattered in the family. For the distinctive rosmarinic acid glucoside found in Saniculoideae-Saniculeae, see Olivier et al. (2008). Peripheral collenchyma in the stem is often especially well developed, and petiole anatomy is very variable and complex (e.g. Metcalfe 1950). Taxa such as Foeniculum have stipules of sorts. Centella has a few branches from the petal bundle (Gustafsson 1995); petal vasculature may repay attention. Spichiger et al. (2002) show the two carpels as being collateral. According to Eyde and Tseng (1971), whether the ventral carpel bundles are fused bundles of adjacent or the same placentae varies within Apiaceae without any particular systematic significance. Van Tieghem (1898) noted that in Apiaceae the ascending ovule aborts and the pendulous ovule persists, however, other reports suggest that both ovules are pendulous (Philipson 1970).

For chemistry, see Hegnauer (1971) and Berenbaum (2001), for wood anatomy in Apioideae-Heteromorpheae, see Oskolski and van Wyk (2008), for the distribution of the hypostase, see Gupta (1970), for embryo sac morphology and fruit anatomy, esp. of genera in the old Hydrocotyloideae, see Tseng (1967), for a useful discussion of characters mainly in the context of Southern African taxa, see Burtt (1991a), for genera, old style, see Pimenov and Leonov (1993), for chromosome number and morphology, see Pimenov et al. (2003), for floral development, see Leins and Erbar (2004b), and for vittae and druses, see Liu et al. (2007). There is much useful information in Chandler and Plunkett (2003, 2004), and the subfamilial classification of Plunkett et al. (2004c) is followed here. I am particularly indebted to Mark Watson for comments on Apiaceae s.l.; mistakes remain mine.

Phylogeny. Platysace is not to be included in Mackinlayoideae, where it had been placed (e.g. Chandler & Plunkett 2004), although its exact position is currently uncertain. The positions of Klotzschia ("distinctive fruits") and Hermas (endocarp fibrous, single crystals in mesocarp cells, congested umbels, n = 7 - for the latter two characters, cf. Saniculoideae), both of which used to be in Hydrocotyloideae, are also unclear (Andersson et al. 2006; Calviño et al. 2006, 2008).

Monophyly of the distinctive Saniculoideae (except Lagoecia, now in Apioideae) is upheld in all molecular analyses, and details of relationships within it are given by Valiejo-Roman et al. (2002). In some analyses the African ex-hydrocotyloid Arctopus may be sister to other Saniculoideae, and the woody Steganotaenia and Polemanniopsis, and perhaps Lichtensteinia (for this genus see below in Apioideae), are probably also part of the same clade (Downie et al. 2001; van Wyck 2001; M. Liu et al. 2003; Plunkett et al. 2004c); at least some of the latter genera have a slightly lignified endocarp, apparently alone in both Saniculoideae and Apioideae (M. Liu et al. 2004). Calviño et al. (2006, esp. 2007 for a good general discussion of variation), however, express reservations about this expanded - and characterless - Eryngioideae, while Calviño and Downie (2007) found that the clade could be circumscribed satisfactorily so long as Lichtensteinia was excluded and placed in Apioideae; they recognised two tribes, both well supported and with indels. Carpophores, chromosome numbers other than 8, and inflorescences that are not condensed (all in African taxa) could be plesiomorphic in the clade (see also Calviño et al. 2008a; Kadereit et al. 2008). The usually compact inflorescence units of Saniculoideae seem best interpreted as a group of reduced umbellules (Froebe 1964, 1971). The large genus Eryngium has separate New and Old World clades (Calvinño et al. 2008b)

Within Apioideae, some taxa are woody, e.g. the Southern African Bupleurum fruticosum and a number of other Southern African taxa (Calviño et al. 2006); Myrrhidendron, from Central America and Colombia, is also woody. Relationships within the subfamily are something like [Heteromorpheae (mostly woody, African; vessel element walls with helical thickenings) [Bupleurum + The Rest]] (Downie & Katz-Downie 1999; Plunkett et al. 2004). Includes Lagoecia.

Generic and tribal limits are something of a disaster area. The traditional tribal and generic classification was based primarily on fruit morphology and is problematic; convergence of fruit architecture as adaptations to modes of fruit dispersal has resulted in many non-monophyletic groupings. Recent studies using molecular techniques are demonstrating the problems with the traditional classifications and are beginning to resolve generic and tribal affinities. For instance, 13/18 genera for which two or more sequences were included in in a study by Downie et al. (2000b) were found not to be monophyletic. The problem is further compounded by the disproportionately large number of mono- or di-typic genera (see also Spalik et al. 2001; Valiejo-Roman et al. 2006; Spalik & Downie 2007; Zhou et al. 2008 [China]; Ajani et al. 2008 [Iran]; Winter et al. 2008 [Africa]; Downie et al. 2000b and Kurzyna-Mlynik et al. 2008 [Ferula, somewhat paraphyletic, and relatives]).

For other relationships within Apioideae, see e.g. Downie et al. (2000a, b, 2001, 2002), Hardway et al. (2004), Plunkett and Downie (2000 - variation in size of chloroplast inverted repeat characterises a large clade in Apioideae), Spalik and Downie (2001a, b), Sun and Downie (2004), Sun et al. (2004: Cymopterus and Lomatium about as polyphletic as you can get), and Downie et al. (2008: Oenantheae). Among "basal" Apioideae are clades made up largely of Southern African genera, these include Lichtensteinia, with rounded, toothed cotyledons (Burtt 1991a) and strongly supported as sister to the rest of the subfamily (as Calviño et al. 2006 note, this has considerable implications for synapomorphy positions), the Annesorhiza clade, perhaps allied to Heteromorpheae (including Pseudocarum, which has branching/anastomosing vittae), and finally Bupleurum, also with branching vittae, strongly supported as sister to the remainder of the subfamily (Calviño et al. 2005, 2006).