Botanical Journal of the Linnean Society, 2009, 161, 105–121. With 1 figure An update of the Angiosperm Phylogeny Group classification for the orders and families of flowering plants: APG III boj_996105..121 THE ANGIOSPERM PHYLOGENY GROUP*1 1Recommended citation: APG III (2009). This paper was compiled by Birgitta Bremer, Kåre Bremer, Mark W. Chase, Michael F. Fay, James L. Reveal, Douglas E. Soltis, Pamela S. Soltis and Peter F. Stevens, who were equally responsible and listed here in alphabetical order only, with contributions from Arne A. Anderberg, Michael J. Moore, Richard G. Olmstead, Paula J. Rudall, Kenneth J. Sytsma, David C. Tank, Kenneth Wurdack, Jenny Q.-Y. Xiang and Sue Zmarzty (in alphabetical order). Addresses: B. Bremer, The Bergius Foundation at the Royal Swedish Academy of Sciences, PO Box 50017, SE-104 05 Stockholm, Sweden; K. Bremer, Vice Chancellor, Stockholm University, SE-106 91 Stockholm, Sweden; M. W. Chase, M. F. Fay, Jodrell Laboratory, Royal Botanic Gardens, Kew, Richmond, Surrey, TW9 3DS, UK; J. L. Reveal, L.H. Bailey Hortorium, Department of Plant Biology, 412 Mann Building, Cornell University, Ithaca, NY 14853-4301, USA; D. E. Soltis, Department of Biology, University of Florida, Gainesville, Florida 32611–8525, USA; P. S. Soltis, Florida Museum of Natural History, University of Florida, Gainesville, Florida, 32611–7800, USA; and P. F. Stevens, Department of Biology, University of Missouri-St. Louis and Missouri Botanical Garden, PO Box 299, St. Louis, Missouri 63166–0299, USA Received12August2009;acceptedforpublication18August2009 A revised and updated classification for the families of flowering plants is provided. Many recent studies have yielded increasingly detailed evidence for the positions of formerly unplaced families, resulting in a number of newlyadoptedorders,includingAmborellales,Berberidopsidales,Bruniales,Buxales,Chloranthales,Escalloniales, Huerteales, Nymphaeales, Paracryphiales, Petrosaviales, Picramniales, Trochodendrales, Vitales and Zygophylla- les.Anumberofpreviouslyunplacedgeneraandfamiliesareincludedhereinorders,greatlyreducingthenumber of unplaced taxa; these include Hydatellaceae (Nymphaeales), Haptanthaceae (Buxales), Peridiscaceae (Saxifragales), Huaceae (Oxalidales), Centroplacaceae and Rafflesiaceae (both Malpighiales),Aphloiaceae, Geisso- lomataceae and Strasburgeriaceae (all Crossosomatales), Picramniaceae (Picramniales), Dipentodontaceae and Gerrardinaceae (both Huerteales), Cytinaceae (Malvales), Balanophoraceae (Santalales), Mitrastemonaceae (Ericales) and Boraginaceae (now at least known to be a member of lamiid clade). Newly segregated families for genera previously understood to be in other APG-recognized families include Petermanniaceae (Liliales), Calophyllaceae (Malpighiales), Capparaceae and Cleomaceae (both Brassicales), Schoepfiaceae (Santalales), Anacampserotaceae, Limeaceae, Lophiocarpaceae, Montiaceae and Talinaceae (all Caryophyllales) and Linder- niaceae and Thomandersiaceae (both Lamiales). Use of bracketed families is abandoned because of its unpopu- larity, and in most cases the broader circumscriptions are retained; these includeAmaryllidaceae,Asparagaceace and Xanthorrheaceae (all Asparagales), Passifloraceae (Malpighiales), Primulaceae (Ericales) and several other smaller families. Separate papers in this same volume deal with a new linear order forAPG, subfamilial names thatcanbeusedformoreaccuratecommunicationinAmaryllidaceaes.l.,Asparagaceaces.l.andXanthorrheaceae s.l.(allAsparagales)andaformalsupraordinalclassificationforthefloweringplants. ©2009TheLinneanSociety of London, Botanical Journal of the Linnean Society, 2009, 161, 105–121. ADDITIONAL KEYWORDS: angiosperm classification – phylogenetic classification – DNA phylogenetics – classification system – APG system. *Corresponding author: Mark W. Chase. E-mail: [email protected] © 2009 The Linnean Society of London, Botanical Journal of the Linnean Society, 2009, 161, 105–121 105 106 THE ANGIOSPERM PHYLOGENY GROUP INTRODUCTION along APG lines. They have all agreed to adopt APG III as their standard and the linear order of Since the previousAPG classification (APG II, 2003), Haston etal. (2009). In general, the broader cir- a great deal more information about flowering plant cumscriptionshavebeenfavouredandareadopted relationships has emerged, and a more refined and here. better-resolvedclassificationisnowpossible.Ifawell- 2. Papers over the last few years have clarified the supported hypothesis of monophyly is a necessary positions of isolated families including Cerato- prerequisite for a group to be named, it is not suffi- phyllaceae, Chloranthaceae and Picramniaceae cient. Not all clades need be named and, indeed, it (Jansen etal., 2007; Moore etal., 2007; Wang would barely be practicable to do this, so to decide etal., 2009), and this has necessitated addition of whichcladesshouldbenamed,additionalcriteriacan orders not previously recognized by APG. be invoked. Backlund & Bremer (1998) provided a 3. A few genera/families, members of which had usefuldiscussionontheprinciplesofrank-basedphy- either not been sequenced before or for which logenetic classification that is applicable at all levels chimaeric sequences were produced, were wrongly apartfromspeciesandimmediatelyabove(seeAlbach placed. Thus, families like Guamatelaceae have etal., 2004; Entwisle & Weston, 2005; Pfeil & Crisp, been added; Guamatela used to be placed in 2005, etc., for examples). We follow these principles Rosaceae, but molecular data places it in here. Backlund and Bremer’s main principle is that Crossosomatales (Oh & Potter, 2006). Hydatel- taxa that are recognized formally should be mono- laceae have been moved from Poales to Nymphae- phyletic. However, this does not indicate which par- ales (Saarela etal., 2007). ticularcladesshouldbenamedasfamilies,ordersetc. 4. Thereareafewcaseswherethegeneralpatternof It is helpful if (1) taxa formally recognized are relationships has not changed much sinceAPG II, easily recognizable, (2) groups that are well estab- but our appreciation of the pattern of variation lished in the literature are preserved, (3) the size of has. For example, this helps justify inclusion of groups is taken into account (particularly small ones, Ixerbaceae in Strasburgeriaceae. which should be combined with others whenever pos- 5. Finally,afewfamilycircumscriptionssuggestedby sible) and (4) nomenclatural changes are minimized APG II did not reflect general usage, so we have (Backlund & Bremer, 1998). Thus, on the one hand, modified these, an example being the broadly numerous small groups have little to recommend circumscribed Brassicaceae, which are here split them, as individually they summarize only a small into three families. amount of information and tend to clog the memory, whereas groups that are too large may have few In general, we have tried not to change the status obvious shared traits that can be used to recognize of families if they have long been recognized, unless them.Asparagales and Lamiales in particular have a there are other good reasons for combining them. number of problems in this regard and include fami- However, we have taken the opportunity to combine lies that are still in a state of flux in this respect. mono- or oligogeneric families. Most of the family- Changes are being made to the classification level changes mentioned below have resulted in the adopted by APG II for several reasons, but bearing families now recognized being relatively well charac- the Backlund and Bremer guidelines in mind. terized. However, we realize that it is almost impos- 1. In the earlier versions of the Angiosperm Phylog- sible to achieve universal agreement for change – of eny Group classification, alternative circumscrip- any sort. Note that we do not see theAPG classifica- tions were permitted for some families. However, tion as continuing to mutate for the indefinite future. this seems unnecessary and more likely to cause Given the amount of phylogenetic work that has confusion than clarity. Here we indicate one of the taken place in the last five years, the changes pro- alternatives, a choice based in part on the circum- posed here are modest. We hope the classification scriptions adopted by textbooks (e.g. Judd etal., below will be found to be reasonable and, hence, will 2007), dictionaries (e.g. Mabberley, 2008) and the not need much further change, although we do note general literature. Note that, in preparation for those few areas where there is particular phylo- the third edition of The Plant Book, Mabberley genetic uncertainty that may necessitate further (2008: xi, 927, 929) consulted widely among tax- revision of familial or ordinal limits. For further dis- onomists about which alternative they preferred, cussion on the variation in the groups discussed, and, more recently, the issue of these alternative potential apomorphies, etc., see the literature cited circumscriptions was discussed by researchers and Stevens (2001); particularly important recent representingseveralEuropeanherbaria(e.g.K,E, work includes Wang etal. (2009: rosids), Tank & BM, P, G and the Dutch herbaria collectively) that Donoghue(inpress),Mooreetal.(2008,inpress:core are in the process of reorganizing their collections eudicots), Wurdack & Davis (2009: Malpighiales) and © 2009 The Linnean Society of London, Botanical Journal of the Linnean Society, 2009, 161, 105–121 APG III 107 Refulio-Rodriguez & Olmstead (2008, pers. comm.: Austrobaileyales Takht. ex Reveal (1992) Lamiales). Austrobaileyaceae Croizat (1943), nom. cons. Forthoseresearcherswhowishtoemployaformal, $Schisandraceae Blume (1830), nom. cons. (includ- higher-level classification for the land plants, Chase ing Illiciaceae A.C.Sm., nom. cons.) & Reveal (2009) have proposed a system. It is based Trimeniaceae L.S.Gibbs (1917), nom. cons. on the recognition that most previous systems Schisandraceae s.l. are well characterized. have inflated the ranks of the angiosperms and the †Chloranthales R.Br. (1835) other major groups of land plants. In this system, Chloranthaceae R.Br. ex Sims (1820), nom. cons. all embryophytes are a class, and, therefore, the angiosperms are recognized collectively as a single Chloranthaceae are probably sister to magnoliids subclass, Magnoliidae, which necessitated the use of (Moore etal., 2007). Separate ordinal status is superorders for the major angiosperm clades recog- warranted because of their phylogenetic position; nized here (rosids, asterids etc.). they are also morphologically distinctive. The general sequence of orders follows the left to right sequence of the largely ladderized tree in MAGNOLIIDS Haston etal. (2009). Within orders, the sequence of families is alphabetical. Symbols: *new family place- Canellales Cronquist (1957) ment; †newly recognized order for the APG system; Canellaceae Mart. (1832), nom. cons. §new family circumscription described in the text; Winteraceae R.Br. ex Lindl. (1830), nom. cons. $families that represent the broader circumscription of options available in APG II and favoured here, Piperales Bercht. & J.Presl (1820) $$families that were in square brackets in APG II, Aristolochiaceae Juss. (1789), nom. cons. thenarrowercircumscriptionsfavouredhere.Thelist Hydnoraceae C.Agardh (1821), nom. cons. reflects a starting date for all flowering plant family Lactoridaceae Engl. (1888), nom. cons. namesas4August1789(Jussieu,Generaplantarum). Piperaceae Giseke (1792), nom. cons. Fullcitationsareavailableelsewhere(Reveal,2008– Saururaceae F.Voigt (1811), nom. cons. onward). A summarized phylogenetic tree represent- The relationships of Hydnoraceae are unclear ing the relationships among the major groups within Piperales. recognized here is presented in Figure1. Laurales Juss. ex Bercht. & J.Presl (1820) Atherospermataceae R.Br. (1814) CLASSIFICATION OF FLOWERING PLANTS Calycanthaceae Lindl. (1819), nom. cons. Gomortegaceae Reiche (1896), nom. cons. †Amborellales Melikyan, A.V.Bobrov & Zaytzeva Hernandiaceae Blume (1826), nom. cons. (1999) Lauraceae Juss. (1789), nom. cons. Amborellaceae Pichon (1948), nom. cons. Monimiaceae Juss. (1809), nom. cons. The evidence that Amborellaceae are sister to all Siparunaceae Schodde (1970) other angiosperms is clear (e.g. Hansen etal., 2007; Magnoliales Juss. ex Bercht. & J.Presl (1820) Jansenetal.,2007;Mooreetal.,2007).However,even Annonaceae Juss. (1789), nom. cons. if they were sister to Nymphaeales (e.g. Goremykin, Degeneriaceae I.W.Bailey & A.C.Sm. (1942), nom. Viola & Hellwig, 2009), they should be kept separate cons. as their inclusion in Nymphaeales would result in a Eupomatiaceae Orb. (1845), nom. cons. taxon without characters. Himantandraceae Diels (1917), nom. cons. †Nymphaeales Salisb. ex Bercht. & J.Presl (1820) Magnoliaceae Juss. (1789), nom. cons. $$Cabombaceae Rich. exA.Rich. (1822), nom. cons. Myristicaceae R.Br. (1810), nom. cons. *Hydatellaceae U.Hamann (1976) $$Nymphaeaceae Salisb. (1805), nom. cons. MONOCOTS There seems to be a general preference for keeping Acorales Link (1835) Cabombaceae and Nymphaeaceae separate, although Acoraceae Martinov (1820) both are small in terms of species. The two families are easily characterizable. The inclusion of Hydatel- Alismatales R.Br. ex Bercht. & J.Presl (1820) laceae (previously included in Poales) here was unex- §Alismataceae Vent. (1799), nom. cons. (including pected, but it is well supported morphologically and Limnocharitaceae Takht. ex Cronquist) embryologically(Saarelaetal.,2007;Friedman,2008; Aponogetonaceae Planch. (1856), nom. cons. Rudall etal., 2008). Araceae Juss. (1789), nom. cons. © 2009 The Linnean Society of London, Botanical Journal of the Linnean Society, 2009, 161, 105–121 108 THE ANGIOSPERM PHYLOGENY GROUP Amborellales† Nymphaeales† Austrobaileyales a Piperales n Canellales g magnoliids io Magnoliales s Laurales p Chloranthales† e r Commelinales m Zingiberales s Poales commelinids Arecales Dasypogonaceae m Asparagales o n Liliales o Pandanales c o Dioscoreales ts Petrosaviales† Alismatales Acorales Ceratophyllales† Ranunculales Sabiaceae Proteales Buxales† e u Trochodendrales† d Gunnerales ico Cucurbitales t Fagales s Rosales Fabales Celastrales fabids Oxalidales Malpighiales Zygophyllales† Malvales c r Brassicales o o r s Huerteales† e id Sapindales malvids e s Picramniales† u d Crossosomatales ic Myrtales o Geraniales t s Vitales† Saxifragales Dilleniaceae Berberidopsidales† Santalales Caryophyllales Cornales Ericales Garryales a s Gentianales te Lamiales lamiids rid Solanales s Boraginaceae Aquifoliales Escalloniales† Asterales Dipsacales campanulids Paracryphiales† Apiales Bruniales† Figure1. Interrelationships of the APG III orders and some families supported by jackknife/bootstrap percentages greaterthan50orBayesianposteriorprobabilitiesgreaterthan0.95inlarge-scaleanalysesofangiosperms.Seetextfor literaturesupportingtheserelationships.Newly-recognized-for-APGordersaredenoted(†).Someeudicotfamiliesnotyet classified to order are not shown. © 2009 The Linnean Society of London, Botanical Journal of the Linnean Society, 2009, 161, 105–121 APG III 109 Butomaceae Mirb. (1804), nom. cons. Liliales Perleb (1826) Cymodoceaceae Vines (1895), nom. cons. §Alstroemeriaceae Dumort. (1829), nom. cons. Hydrocharitaceae Juss. (1789), nom. cons. (including Luzuriagaceae Lotsy) Juncaginaceae Rich. (1808), nom. cons. Campynemataceae Dumort. (1829) Posidoniaceae Vines (1895), nom. cons. Colchicaceae DC. (1804), nom. cons. Potamogetonaceae Bercht. & J.Presl (1823), Corsiaceae Becc. (1878), nom. cons. nom. cons. Liliaceae Juss. (1789), nom. cons. Ruppiaceae Horan. (1834), nom. cons. Melanthiaceae Batsch ex Borkh. (1797), nom. cons. Scheuchzeriaceae F.Rudolphi (1830), nom. cons. *Petermanniaceae Hutch. (1934), nom. cons. Tofieldiaceae Takht. (1995) Philesiaceae Dumort. (1829), nom. cons. Zosteraceae Dumort. (1829), nom. cons. Ripogonaceae Conran & Clifford (1985) Smilacaceae Vent. (1799), nom. cons. Convincing evidence for the monophyly of Alismataceaes.s.islacking(e.g.Chenetal.,2004a,b) Petermanniaceae are morphologically and phylo- andthefamilydoesnothaveanyapomorphies.When genetically distinct. Luzuriagaceae, consisting of two combined with Limnocharitaceae, a family that was small genera with generalized lily floral morphology, only relatively recently described (Cronquist, 1981), are sister to Alstroemeriaceae and have the same the enlarged family has several distinctive charac- distinctivetwistedpetioles,socombinationisinorder ters. It may be necessary to split off Maundia from (see also Mabberley, 2008). Juncaginaceae(Ilesetal.,2009;S.vonMering&J.W. Asparagales Link (1829) Kadereit, pers. comm.), and Maundiaceae Nakai is $Amaryllidaceae J.St.-Hil., nom. cons. (including available. However, it might be better in this case to Agapanthaceae F.Voigt, Alliaceae Borkh., nom. createalargersinglefamilyforthelargerclade.More cons.) study is needed before another monogeneric family is $Asparagaceae Juss. (1789), nom. cons. (including recognized in Alismatales. Agavaceae Dumort., nom. cons.,Aphyllanthaceae Burnett, Hesperocallidaceae Traub, Hyacintha- †Petrosaviales Takht. (1997) ceae Batsch ex Borkh., Laxmanniaceae Bubani, Petrosaviaceae Hutch. (1934), nom. cons. Ruscaceae M.Roem., nom. cons., Themidaceae The isolated position of Petrosaviaceae here is well Salisb.) supported (e.g. Tamura etal., 2004; Chase etal., Asteliaceae Dumort. (1829) 2006), hence its ordinal status. Blandfordiaceae R.Dahlgren & Clifford (1985) Boryaceae M.W.Chase, Rudall & Conran (1997) Dioscoreales R.Br. (1835) Doryanthaceae R.Dahlgren & Clifford (1985) Burmanniaceae Blume (1827), nom. cons. Hypoxidaceae R.Br. (1814), nom. cons. Dioscoreaceae R.Br. (1810), nom. cons. Iridaceae Juss. (1789), nom. cons. Nartheciaceae Fr. ex Bjurzon (1846) Ixioliriaceae Nakai (1943) Lanariaceae R.Dahlgren & A.E.van Wyk (1988) ThismiaceaeJ.Agardh(1858)mayturnouttobein Orchidaceae Juss. (1789), nom. cons. a clade separate from Burmanniaceae and, similarly, Tecophilaeaceae Leyb. (1862), nom. cons. the morphologically distinctive Taccaceae Dumort. $Xanthorrhoeaceae Dumort. (1829), nom. cons. (1829), nom. cons., from Dioscoreaceae. Phylogenetic (includingAsphodelaceae Juss. and Hemerocalli- relationships in Dioscoreales that support such daceae R.Br.) changes have been found by Merckx etal. (2006) Xeronemataceae M.W.Chase, Rudall & M.F.Fay and Merckx & Bidartondo (2008), and Merckx etal. (2000) (2009) even suggest that Thismiaceae s.s. may be paraphyletic. Given the problems in understanding An expanded Amaryllidaceae, including Agapan- relationships of mycoheterotrophic groups, we have thaceae and Alliaceae, are recognized here (Amaryl- been conservative and not adopted any altered cir- lidaceae was recently conserved over Alliaceae, cumscriptions at this stage. despiteAlliaceaebeingtheoldername).Severalchar- acterssupportthecombinedgroup.Agapanthaceae,if Pandanales R.Br. ex Bercht. & J.Presl (1820) keptseparate,areweaklycharacterized;thefamilyis Cyclanthaceae Poit. ex A.Rich. (1824), nom. cons. monogeneric. Pandanaceae R.Br. (1810), nom. cons. The area aroundAsparagaceae is difficult from the Stemonaceae Caruel (1878), nom. cons. standpointofcircumscription.AlthoughAsparagaceae Triuridaceae Gardner (1843), nom. cons. s.l. are heterogeneous and poorly characterized, Velloziaceae J.Agardh (1858), nom. cons. Asparagaceae s.s., Agavaceae, Laxmanniaceae, Rus- © 2009 The Linnean Society of London, Botanical Journal of the Linnean Society, 2009, 161, 105–121 110 THE ANGIOSPERM PHYLOGENY GROUP caceae and even Hyacinthaceae have few if any dis- Restionaceae R.Br. (1810), nom. cons. tinctivefeatures.AgavaceaeandRuscaceaesensuAPG Thurniaceae Engl. (1907), nom. cons. II(2003)areheterogeneousandhavebeendividedinto §Typhaceae Juss. (1789), nom. cons. (including several families in the past, further confusing the Sparganiaceae Hanin, nom. cons.) issue. This solution at least keeps the number of Xyridaceae C.Agardh (1823), nom. cons. difficult-to-recognize families to a minimum;Amaryl- Sparganiaceae are included in Typhaceae; the two lidaceae s.l. andAsparagaceae s.l. are easily differen- familiesaremonogeneric,occupysimilarhabitatsand tiated by umbels with a pair of enclosing bracts vs. share a number of features. That they were treated racemes or rarely umbels, but if umbels are present separately in APG II was a mistake (M. W. Chase, theycontainthreeormorebracts(e.g.Brodiaea,Milla pers. comm.). They have in the past been combined; and relatives, Themidaceae) and lack the enclosing Mabberley (2008) suggested that their combination pair of bracts. Xanthorrhoeaceae s.l. have been would be in order. maintained for similar reasons to Asparagaceae s.l. Xanthorrhoeaceae s.s. are monogeneric, and Zingiberales Griseb. (1854) Asphodelaceae are impossible to distinguish from Cannaceae Juss. (1789), nom. cons. genera such as Anthericum inAsparagaceae s.l. Costaceae Nakai (1941) For convenience and better communication, Heliconiaceae Vines (1895) a subfamilial classification of Amaryllidaceae, Lowiaceae Ridl. (1924), nom. cons. AsparagaceaeandXanthorrhoeaceaesensuAPGIIIis Marantaceae R.Br. (1814), nom. cons. proposed in Chase, Reveal & Fay (2009). This will Musaceae Juss. (1789), nom. cons. allow researchers to use a subfamily name where Strelitziaceae Hutch. (1934), nom. cons. previously they would have used one of the APG II Zingiberaceae Martinov (1820), nom. cons. bracketed family names. PROBABLE SISTER OF EUDICOTS COMMELINIDS Ceratophyllales Link (1829) Dasypogonaceae Dumort. (1829) Ceratophyllaceae Gray (1822), nom. cons. Relationships of Dasypogonaceae remain unclear. The molecular evidence that Ceratophyllaceae are They are not particularly distinctive morphologically sister to eudicots is becoming clearer (Jansen etal., and so do not warrant immediate ordinal status, 2007; Moore etal., 2007, but cf. Goremykin etal., although Dasypogonales Doweld is available. They 2009). In this and all other relationships that have could probably be combined with one of the other been suggested for Ceratophyllaceae, including sister commelinid orders should they fall as sister groups. tothemonocotsorChloranthaceae(Endress&Doyle, 2009), they are morphologically divergent from their Arecales Bromhead (1840) putative closest relatives. Arecaceae Bercht. & J.Presl (1820), nom. cons. Commelinales Mirb. ex Bercht. & J.Presl (1820) EUDICOTS Commelinaceae Mirb. (1804), nom. cons. Haemodoraceae R.Br. (1810), nom. cons. Ranunculales Juss. ex Bercht. & J.Presl (1820) Hanguanaceae Airy Shaw (1965) Berberidaceae Juss. (1789), nom. cons. Philydraceae Link (1821), nom. cons. $Circaeasteraceae Hutch. (1926), nom. cons. Pontederiaceae Kunth (1816), nom. cons. (including Kingdoniaceae Airy Shaw) Eupteleaceae K.Wilh. (1910), nom. cons. Poales Small (1903) Lardizabalaceae R.Br. (1821), nom. cons. Anarthriaceae D.F.Cutler & Airy Shaw (1965) Menispermaceae Juss. (1789), nom. cons. Bromeliaceae Juss. (1789), nom. cons. $Papaveraceae Juss. (1789), nom. cons. (including Centrolepidaceae Endl. (1836), nom. cons. Fumariaceae Marquis, nom. cons., Pteridophyl- Cyperaceae Juss. (1789), nom. cons. laceae Nakai ex Reveal & Hoogland) Ecdeiocoleaceae D.F.Cutler & Airy Shaw (1965) Ranunculaceae Juss. (1789), nom. cons. Eriocaulaceae Martinov (1820), nom. cons. Flagellariaceae Dumort. (1829), nom. cons. We adopt broad limits for Circaeasteraceae and Joinvilleaceae Toml. & A.C.Sm. (1970) Papaveraceae, as this is commonly done (Judd etal., Juncaceae Juss. (1789), nom. cons. 2007; Mabberley, 2008), and the two families are Mayacaceae Kunth (1842), nom. cons. well characterized in their broader circumscriptions. Poaceae Barnhart (1895), nom. cons. The two families into which Circaeasteraceae have Rapateaceae Dumort. (1829), nom. cons. been divided (Circaeasteraceae and Kingdoniaceae) © 2009 The Linnean Society of London, Botanical Journal of the Linnean Society, 2009, 161, 105–121 APG III 111 are both monogeneric; they are herbaceous and This family has no stable position as yet (Moore their leaves have the same distinctive dichotomous etal., in press). The ordinal name, Dilleniales DC. ex venation. Bercht. & J.Presl, is available. Sabiaceae Blume (1851), nom. cons. Saxifragales Bercht. & J.Presl (1820) Altingiaceae Horan. (1841), nom. cons. Although Moore etal. (2008) placed Sabiaceae as Aphanopetalaceae Doweld (2001) sister to Proteales, support is only moderate. Cercidiphyllaceae Engl. (1907), nom. cons. However, if further work confirms this position, Crassulaceae J.St.-Hil. (1805), nom. cons. Sabiaceae will be included in a broadened circum- Daphniphyllaceae Müll.-Arg. (1869), nom. cons. scription of Proteales; the two have features in Grossulariaceae DC. (1805), nom. cons. common. Sabiaceae remain poorly known. $$Haloragaceae R.Br. (1814), nom. cons. Proteales Juss. ex Bercht. & J.Presl (1820) Hamamelidaceae R.Br. (1818), nom. cons. Nelumbonaceae A.Rich. (1827), nom. cons. $Iteaceae J.Agardh (1858), nom. cons. (including $$Platanaceae T.Lestib. (1826), nom. cons. Pterostemonaceae Small, nom. cons.) $$Proteaceae Juss. (1789), nom. cons. Paeoniaceae Raf. (1815), nom. cons. $$Penthoraceae Rydb. ex Britt. (1901), nom. cons. Platanaceae, although monogeneric, are morpho- *§Peridiscaceae Kuhlm. (1950), nom. cons. logically distinct from Proteaceae, and the two have (including Medusandraceae Brenan, nom. cons., never been combined previously; members of the Soyauxia Oliver) broader family would have few features in common. Saxifragaceae Juss. (1789), nom. cons. †Trochodendrales Takht. ex Cronquist (1981) $$Tetracarpaeaceae Nakai (1943) $Trochodendraceae Eichler (1865), nom. cons. The limits of Iteaceae are broadened because the (including Tetracentraceae A.C.Sm., nom. combined clade is well characterized and Pteroste- cons.). monaceae are monogeneric. The limits of Halor- A separate order for this morphologically distinct agaceae are drawn narrowly as the inclusion of clade is warranted; the two monospecific genera in Penthoraceae and Tetracarpaeaceae would result in Trochodendraceae s.l., Tetracentron and Trocho- a family with no obvious characters and totally dendron, have much in common. novel limits. The three families are individually tol- erably well characterized. Recent molecular analyses †Buxales Takht. ex Reveal (1996) strongly support a placement of Peridiscaceae $Buxaceae Dumort. (1822), nom. cons. (including within Saxifragales, as sister to all other members Didymelaceae Leandri) of this clade (Soltis etal., 2007a; Jian etal., 2008), *Haptanthaceae C.Nelson (2002) rather than in Malpighiales as previously proposed. The limits of Buxaceae are expanded. The monoge- Peridiscaceae continue to be expanded (Davis & neric Didymelaceae have the same distinctive pollen Chase, 2004; Wurdack & Davis, 2009), but all and chemistry as at least part of Buxaceae, although members of the expanded family have similar dis- thereiscurrentlynoevidencefortheparaphylyofthe tinctive seeds etc. latter. Some morphological features suggest that Cynomoriaceae are another family of holoparasitic Haptanthaceae are best placed here, but they are angiosperms that have been difficult to place. Some distinctfromallotherangiosperms(Doust&Stevens, molecular analyses had placed them in Santalales 2005). An order for the two families is warranted. (Jian etal., 2008), although with little support. Note that relationships of Trochodendrales and However, Barkman etal. (2007) found no support for Buxales remain unclear, although they are certainly a position in that order or anywhere else. Neverthe- to be placed in this part of the tree. less, Nickrent (2002) and Nickrent etal. (2005) suggested that Cynomoriaceae should be placed in Saxifragales, but the evidence for placing them here versus in Santalales is not strong. Confounding the CORE EUDICOTS placement of this taxon is evidence for horizontal Gunnerales Takht. ex Reveal (1992) gene transfer involving its host for some mitochon- $$Gunneraceae Meisn. (1842), nom. cons. drial genes (Barkman etal., 2007). Zhang, Li & Li $$Myrothamnaceae Nied. (1891), nom. cons. (2009) analyzed sequences from the plastid inverted repeat and found that Cynomorium fell as sister The two families share no important features and to Rosaceae (Rosales) with high bootstrap support so are kept separate, although both are monogeneric. (99%). Due to these discordant results, we do not Dilleniaceae Salisb. (1807), nom. cons. assign Cynomoriaceae to an order here. © 2009 The Linnean Society of London, Botanical Journal of the Linnean Society, 2009, 161, 105–121 112 THE ANGIOSPERM PHYLOGENY GROUP ROSIDS Balanopaceae Benth. & Hook.f. (1880), nom. cons. Bonnetiaceae L.Beauvis. ex Nakai (1948) †Vitales Juss. ex Bercht. & J.Presl (1820) *Calophyllaceae J.Agardh Vitaceae Juss. (1789), nom. cons. Caryocaraceae Voigt (1845), nom. cons. Vitaceaeremainisolatedandordinalstatusisappro- *Centroplacaceae Doweld & Reveal (2005) priate. They are sister to the fabids+malvids (rosid $$Chrysobalanaceae R.Br. (1818), nom. cons. I+II) clade in most recent analyses, albeit without §Clusiaceae Lindl. (1836), nom. cons. strong support (reviewed in Wang etal., 2009). Ctenolophonaceae Exell & Mendonça (1951) $$Dichapetalaceae Baill. (1886), nom. cons. FABIDS Elatinaceae Dumort. (1829), nom. cons. $$§Erythroxylaceae Kunth (1822), nom. cons. †Zygophyllales Link (1829) (including Aneulophus Benth.) $$Krameriaceae Dumort. (1829), nom. cons. Euphorbiaceae Juss. (1789), nom. cons. $$Zygophyllaceae R.Br. (1814), nom. cons. $$Euphroniaceae Marc.-Berti (1989) Although Krameriaceae are monogeneric, they are Goupiaceae Miers (1862) readily distinguished from the heterogeneous Humiriaceae A.Juss. (1829), nom. cons. Zygophyllaceae; the two are sister taxa. Combin- Hypericaceae Juss. (1789), nom. cons. ing the two would simply make a heterogeneous Irvingiaceae Exell & Mendonça (1951), nom. cons. Zygophyllaceae still more so. An order is needed for Ixonanthaceae Planch. ex Miq. (1858), nom. cons. this family pair as it is placed with strong support as Lacistemataceae Mart. (1826), nom. cons. sister to a clade containing more than two fabid Linaceae DC. ex Perleb (1818), nom. cons. orders in the analysis by Wang etal. (2009). Lophopyxidaceae H.Pfeiff. (1951) Malpighiaceae Juss. (1789), nom. cons. Celastrales Link (1829) $Ochnaceae DC. (1811), nom. cons. (including $Celastraceae R.Br. (1814), nom. cons. (including Medusagynaceae Engl. & Gilg, nom. cons., Lepuropetalaceae Nakai, Parnassiaceae Marti- Quiinaceae Choisy, nom. cons.) nov, nom. cons., Pottingeriaceae Takht.) Pandaceae Engl. & Gilg (1912–1913), nom. cons. Lepidobotryaceae J.Léonard (1950), nom. cons. $Passifloraceae Juss. ex Roussel (1806), nom. cons. [including Malesherbiaceae D.Don, nom. cons., The limits of Celastraceae are broadened because Turneraceae Kunth ex DC. (1828), nom. cons.] the three small families included show every sign Phyllanthaceae Martinov (1820), nom. cons. of making Celastraceae paraphyletic if excluded Picrodendraceae Small (1917), nom. cons. (Zhang & Simmons, 2006). The dismemberment of Podostemaceae Rich. ex Kunth (1816), nom. cons. Celastraceae needed to maintain the families, of Putranjivaceae Meisn. (1842) which only Parnassiaceae are well known, would be *Rafflesiaceae Dumort. (1829), nom. cons. extensiveandyieldpoorlycharacterizedfamilies,and $$Rhizophoraceae Pers. (1807), nom. cons. Celastraceae s.l. are better characterized than Celas- Salicaceae Mirb. (1815), nom. cons. traceae,excludingParnassiaceae(formorphology,see $$Trigoniaceae A.Juss. (1849), nom. cons. Matthews & Endress, 2005a). Violaceae Batsch (1802), nom. cons. Oxalidales Bercht. & J.Presl (1820) Brunelliaceae Engl. (1897), nom. cons. The holoparasitic Rafflesiaceae are best assigned Cephalotaceae Dumort. (1829), nom. cons. to Malpighiales, perhaps making Euphorbiaceae s.s. Connaraceae R.Br. (1818), nom. cons. paraphyletic (e.g. Davis & Wurdack, 2004; Davis Cunoniaceae R.Br. (1814), nom. cons. etal., 2007); the recognition of Peraceae Klotzsch Elaeocarpaceae Juss. ex DC. (1816), nom. cons. (1859) would be needed to maintain monophyly of *Huaceae A.Chev. (1947) Euphorbiaceae. However, pending further studies, Oxalidaceae R.Br. (1818), nom. cons. Peraceae are not recognized here. Limits of clades in the Bonnetiaceae–Podostemaceae area are becoming Huaceae are tentatively included in Oxalidales clearer (Wurdack & Davis, 2009), and this necessi- because a number of recent studies (e.g. Wurdack & tates the removal of Calophyllaceae from Clusiaceae. Davis, 2009) have indicated that they are sister to Thealternativeswouldbeafamilythatincludedboth OxalidalesasrecognizedinpreviousversionsofAPG. ofthesefamiliesandBonnetiaceae,Hypericaceaeand This is not a well-characterized clade, and it remains Podostemaceae or one that included the last two poorly understood. families plus Calophyllaceae; in both cases Hyperi- Malpighiales Juss. ex Bercht. & J.Presl (1820) caceae would be the correct name. The four families Achariaceae Harms (1897), nom. cons. in the area of Chrysobalanaceae, Dichapetalaceae, © 2009 The Linnean Society of London, Botanical Journal of the Linnean Society, 2009, 161, 105–121 APG III 113 Euphroniaceae and Trigoniaceae are kept separate similar, even although Rhoipteleaceae have superior as,althoughclearlyrelated,abroadlydrawnChryso- and Juglandaceae inferior ovaries; ovary position is balanaceae would be heterogeneous (see Matthews & variable in many families, and in other such cases, Endress, 2008, for the morphology of this group). e.g. Ebenaceae (including Lissocarpaceae), we have Bhesa (formerly of Celastraceae) and Centroplacus recognized the larger unit. (formerly of Euphorbiaceae) form an isolated clade Rosales Bercht. & J.Presl (1820) (Davis etal., 2005; Wurdack & Davis, 2009) that has Barbeyaceae Rendle (1916), nom. cons. distinguishing features; recognition of a bigeneric Cannabaceae Martinov (1820), nom. cons. Centroplacaceae is reasonable. Salicaceae are Dirachmaceae Hutch. (1959) broadly drawn, i.e. including Samydaceae Vent. and Elaeagnaceae Juss. (1789), nom. cons. Scyphostegiaceae Hutch. Although the combined Moraceae Gaudich. (1835), nom. cons. clade is only moderately distinct morphologically and Rhamnaceae Juss. (1789), nom. cons. less so phylogenetically (Chase etal., 2002), recogni- Rosaceae Juss. (1789), nom. cons. tion of more families in this area is premature Ulmaceae Mirb. (1815), nom. cons. pending a more detailed sampling of the genera. Urticaceae Juss. (1789), nom. cons. Trichostephanus Gilg, unplaced previously, shares wood anatomy, disc lobing and seed structure with SamydeaeVent.(=Salicaceae),andunpublishedDNA MALVIDS datasupportthisplacement(M.Alford,pers.comm.). Geraniales Juss. ex Bercht. & J.Presl (1820) Rhizophoraceae are kept separate from their $Geraniaceae Juss. (1789), nom. cons. (including sister taxon, Erythroxylaceae, although Aneulophus, Hypseocharitaceae Wedd.) of Erythroxylaceae, is to a certain extent morphologi- $Melianthaceae Horan. (1834), nom. cons. (includ- callyintermediate;thetwofamilieshavehithertonot ing Francoaceae A.Juss., nom. cons.) been combined. Passifloraceae and Ochnaceae are §Vivianiaceae Klotzsch, nom. cons. prop. (including broadly delimited here yet remain readily character- Ledocarpaceae Meyen) izable; relationships between the component clades within the two families are uncertain, and both Geraniales are a heterogeneous and poorly known Medusagynaceae (Ochnaceae s.l.) and Malesherbi- order. The inclusion of the monogeneric Hypseochari- aceae (Passifloraceae s.l.) are monogeneric. taceae in Geraniaceae, monogeneric Francoaceae in Melianthaceae and bigeneric Ledocarpaceae in Cucurbitales Juss. ex Bercht. & J.Presl (1820) Vivianiaceae leaves these expanded families with a Anisophylleaceae Ridl. (1922) number of characters. Begoniaceae C.Agardh (1824), nom. cons. Coriariaceae DC. (1824), nom. cons. Myrtales Juss. ex Bercht. & J.Presl (1820) Corynocarpaceae Engl. (1897), nom. cons. Alzateaceae S.A.Graham (1985) Cucurbitaceae Juss. (1789), nom. cons. Combretaceae R.Br. (1810), nom. cons. Datiscaceae Dumort. (1829), nom. cons. Crypteroniaceae A.DC. (1868), nom. cons. Tetramelaceae Airy Shaw (1965) Lythraceae J.St.-Hil. (1805), nom. cons. $Melastomataceae Juss. (1789), nom. cons. (includ- Fabales Bromhead (1838) ing Memecylaceae DC., nom. cons.) Fabaceae Lindl. (1836), nom. cons. §Myrtaceae Juss. (1789), nom. cons. (including PolygalaceaeHoffmanns.&Link(1809),nom.cons. Heteropyxidaceae Engl. & Gilg, nom. cons., Quillajaceae D.Don (1831) Psiloxylaceae Croizat) Surianaceae Arn. (1834), nom. cons. Onagraceae Juss. (1789), nom. cons. Fagales Engl. (1892) §Penaeaceae Sweet ex Guill. (1828), nom. cons. Betulaceae Gray (1822), nom. cons. (including Oliniaceae Arn., nom. cons., Rhyn- Casuarinaceae R.Br. (1814), nom. cons. chocalycaceae L.A.S.Johnson & B.G.Briggs) Fagaceae Dumort. (1829), nom. cons. Vochysiaceae A.St.-Hil. (1820), nom. cons. §Juglandaceae DC. ex Perleb (1818), nom. cons. In Melastomataceae, and still more in Myrtaceae, [including Rhoipteleaceae Hand.-Mazz. (1932), common usage is for broadened family circumscrip- nom. cons.] tions. Both Heteropyxidaceae and Psiloxylaceae are Myricaceae A.Rich. ex Kunth (1817), nom. cons. small families and when included in Myrtaceae s.l. Nothofagaceae Kuprian (1962) that family remains characterized by possession of Ticodendraceae Gómez-Laur. & L.D.Gómez (1991) pellucid glands containing ethereal oils. Juglandaceae are expanded to include their sister A close relationship between Crypteroniaceae, clade, the monogeneric Rhoipteleaceae. The two look Penaeaceae, Oliniaceae, Alzataeaceae and Rhyn- © 2009 The Linnean Society of London, Botanical Journal of the Linnean Society, 2009, 161, 105–121 114 THE ANGIOSPERM PHYLOGENY GROUP chocalycaceae is clear; Van Beusekom-Osinga & van Brassicales Bromhead (1838) Beusekom(1975)includedthelasttwofamiliesinthe $Akaniaceae Stapf (1912), nom. cons. (including first. All families are morphologically similar, Bretschneideraceae Engl. & Gilg, nom. cons.) although they show variation in floral morphology Bataceae Mart. ex Perleb (1838), nom. cons. andembryosac,inparticular.Somecombinationisin §Brassicaceae Burnett (1835), nom. cons. order,andPenaeaceaehavebeenexpandedtoinclude *Capparaceae Juss. (1789), nom. cons. Rhynchocalycaceae and Oliniaceae; Penaeaceae s.l. Caricaceae Dumort. (1829), nom. cons. can be characterized. *Cleomaceae Bercht. & J.Presl (1825) Emblingiaceae J.Agardh (1958) Crossosomatales Takht. ex Reveal (1993) Gyrostemonaceae A.Juss. (1845), nom. cons. *Aphloiaceae Takht. (1985) Koeberliniaceae Engl. (1895), nom. cons. Crossosomataceae Engl. (1897), nom. cons. Limnanthaceae R.Br. (1833), nom. cons. *Geissolomataceae A.DC. (1856) Moringaceae Martinov (1820), nom. cons. *Guamatelaceae S.Oh & D.Potter (2006) Pentadiplandraceae Hutch. & Dalziel (1928) Stachyuraceae J.Agardh (1858), nom. cons. Resedaceae Martinov (1820), nom. cons. Staphyleaceae Martinov (1820), nom. cons. Salvadoraceae Lindl. (1836), nom. cons. *§Strasburgeriaceae Soler. (1908), nom. cons. Setchellanthaceae Iltis (1999) (includingIxerbaceaeGriseb.exDoweld&Reveal) Tovariaceae Pax (1891), nom. cons. TheadditionofseveralfamiliestoCrossosomatales Tropaeolaceae Juss. ex DC. (1824), nom. cons. iswelljustified(e.g.Sosa&Chase,2003;Soltisetal., 2007b; Wang etal., 2009), although monogeneric Inclusion of monogeneric Bretschneideraceae into Guamatelaceae are a somewhat surprising addition, the monogeneric Akaniaceae is justified by the Guamatela having previously been included in morphological similarities of the two, which are Rosaceae (Oh & Potter, 2006). For the most part, sister taxa. Although a broad circumscription of relationships among families included in Crossoso- Brassicaceae was recognized inAPG (1998) andAPG matales have not been suggested before. The order is II (2003), the consensus prefers the recognition of heterogeneous, although the families are all small. three families, all of which can be characterized, Nevertheless, Matthews & Endress (2005b, 2006) albeit Capparaceae only rather poorly so. The final found a number of floral features to be at least phylogenetic positions, and hence taxonomic disposi- common in the order. The sister taxa Strasburgeri- tion, of some genera, particularly those previously aceae and Ixerbaceae are two monogeneric families included in Capparaceae–Stixeae, remain uncertain that agree in several characters, including base chro- (Hall, Sytsma & Iltis, 2002; Hall, Iltis & Sytsma, mosome number and stamen and gynoecial morphol- 2004). Nonetheless, the name Stixaceae Doweld ogy etc.; combination is in order.As the order is now (2008) is available if it is required. defined, Staphyleales Mart. (1835) is an older name Malvales Juss. ex Bercht. & J.Presl (1820) for Crossosomatales Takht. ex Reveal (1993). It need $Bixaceae Kunth (1822), nom. cons. (including not be adopted, however. CochlospermaceaePlanch.,nom.cons.,Diegoden- †Picramniales Doweld (2001) draceae Capuron,) *Picramniaceae Fernando & Quinn (1995) Cistaceae Juss. (1789), nom. cons. *Cytinaceae A.Rich. (1824) Picramniaceaewerepreviouslyunplacedrosids,but Dipterocarpaceae Blume (1825), nom. cons. thereisnowstrongsupportforapositionintherosid Malvaceae Juss. (1789), nom. cons. II/malvid clade (Wang etal., 2009). Muntingiaceae C.Bayer, M.W.Chase & M.F.Fay †Huerteales Doweld (2001) (1998) *Dipentodontaceae Merr. (1941), nom. cons. Neuradaceae Kostel. (1835), nom. cons. *Gerrardinaceae Alford (2006) Sarcolaenaceae Caruel (1881), nom. cons. Tapisciaceae Takht. (1987) Sphaerosepalaceae Tiegh. ex Bullock (1959) Thymelaeaceae Juss. (1789), nom. cons. This assemblage of three small families is well supported, and recognition of Huerteales is appro- AbroadcircumscriptionforBixaceaeisadopted;the priate given their position. Within Huerteales, the threefamiliesincludedareallsmall,andthecombined recently described Gerrardinaceae (Gerrardina was familycanbecharacterizedmorphologically.Thepara- previously placed in Flacourtiaceae) are sister to the sitic Cytinaceae (including Bdallophyton Eichl.) find rest, and Dipentodontaceae, although monogeneric their resting place here (Nickrent, 2007). The novel like Gerrardinaceae, are distinctive (see Worberg dismemberment of Malvaceae by Cheek (2006), see etal., 2009). also Cheek in Heywood etal., 2007) is not followed © 2009 The Linnean Society of London, Botanical Journal of the Linnean Society, 2009, 161, 105–121