Apg IV (1) (2)

Apg IV (1) (2)

(Parte 1 de 7)

An update of the Angiosperm Phylogeny Group classification for the orders and families of flowering plants: APG IV


1Recommended citation: APG IV (2106). This paper was compiled by James W. Byng, Mark W. Chase, Maarten J. M. Christenhusz, Michael F. Fay, Walter S. Judd, David J. Mabberley, Alexander N. Sennikov, Douglas E. Soltis, Pamela S. Soltis and Peter F. Stevens, who were equally responsible and listed here in alphabetical order only, with contributions from Barbara Briggs, Samuel Brockington, Alain Chautems, John C. Clark, John Conran, Elspeth Haston, Michael M€oller, Michael Moore, Richard Olmstead, Mathieu Perret, Laurence Skog, James Smith, David Tank, Maria Vorontsova and Anton Weber. Addresses: M. W. Chase, M. J. M. Christenhusz, M. F. Fay, Royal Botanic Gardens, Kew, Richmond, Surrey TW9 3DS, UK; J. W. Byng, M. J. M. Christenhusz, Plant Gateway, 5 Talbot Street, Hertford, Hertfordshire SG13 7BX, UK; J. W. Byng, School of Biological Sciences, University of Aberdeen, Aberdeen AB24 3UU, UK; M. W. Chase, University of Western Australia, 35 Stirling Highway, Crawley, Western Australia 6009, Australia; W. S. Judd, D. E. Soltis, Department of Biology, University of Florida, Gainesville, FL 32611-8525, USA; D. J. Mabberley, Wadham College, University of Oxford, UK; Universiteit Leiden and Naturalis Biodiversity Center, Leiden, the Netherlands; Macquarie University and National Herbarium of New South Wales, Sydney, Australia; A. N. Sennikov, Botanical Museum, Finnish Museum of Natural History, PO Box 7, FI-00014, Helsinki, Finland and Komarov Botanical Institute, Prof. Popov 2, RU-197376, St. Petersburg, Russia; D. E. Soltis, P. S. Soltis, Florida Museum of Natural History, University of Florida, Gainesville, FL 32611-7800, USA; P. F. Stevens, Department of Biology, University of Missouri-St. Louis and Missouri Botanical Garden, PO Box 299, St. Louis, MO 63166-0299, USA.

Received 10 January 2016; revised 17 January 2016; accepted for publication 17 January 2016

An update of the Angiosperm Phylogeny Group (APG) classification of the orders and families of angiosperms is presented. Several new orders are recognized: Boraginales, Dilleniales, Icacinales, Metteniusiales and Vahliales. This brings the total number of orders and families recognized in the APG system to 63 and 416, respectively. We propose two additional informal major clades, superrosids and superasterids, that each comprise the additional orders that are included in the larger clades dominated by the rosids and asterids. Families that made up potentially monofamilial orders, Dasypogonaceae and Sabiaceae, are instead referred to Arecales and Proteales, respectively. Two parasitic families formerly of uncertain positions are now placed: Cynomoriaceae in Saxifragales and Apodanthaceae in Cucurbitales. Although there is evidence that some families recognized in APG I are not monophyletic, we make no changes in Dioscoreales and Santalales relative to APG I and leave some genera in Lamiales unplaced (e.g. Peltanthera). These changes in familial circumscription and recognition have all resulted from new results published since APG I, except for some changes simply due to nomenclatural issues, which include substituting Asphodelaceae for Xanthorrhoeaceae (Asparagales) and Francoaceae for Melianthaceae (Geraniales); however, in Francoaceae we also include Bersamaceae, Ledocarpaceae, Rhynchothecaceae and Vivianiaceae. Other changes to family limits are not drastic or numerous and are mostly focused on some members of the lamiids, especially the former Icacinaceae that have long been problematic with several genera moved to the formerly monogeneric Metteniusaceae, but minor changes in circumscription include Aristolochiaceae (now including Lactoridaceae and Hydnoraceae; Aristolochiales), Maundiaceae (removed from Juncaginaceae; Alismatales), Restionaceae (now re-including Anarthriaceae and Centrolepidaceae; Poales), Buxaceae (now including Haptanthaceae; Buxales), Peraceae (split from Euphorbiaceae; Malpighiales), recognition of Petenaeaceae (Huerteales), Kewaceae, Limeaceae, Macarthuriaceae and Microteaceae (all Caryophyllales), Petiveriaceae split from Phytolaccaceae (Caryophyllales), changes to the generic composition of Ixonanthaceae and Irvingiaceae (with transfer of Allantospermum from the former to the latter; Malpighiales), transfer of Pakaraimaea (formerly Dipterocarpaceae) to Cistaceae (Malvales), transfer of Borthwickia, Forchhammeria, Stixis and Tirania (formerly all Capparaceae) to Resedaceae (Brassicales), Nyssaceae split from Cornaceae (Cornales), Pteleocarpa moved to Gelsemiaceae (Gentianales), changes to the generic composition of Gesneriaceae (Sanango moved from Loganiaceae) and Orobanchaceae (now including Lindenbergiaceae and

*E-mail: m.chase@kew.org 1© 2016 The Linnean Society of London, Botanical Journal of the Linnean Society, 2016

Botanical Journal of the Linnean Society, 2016. With 1 figure.

Rehmanniaceae) and recognition of Mazaceae distinct from Phrymaceae (all Lamiales). © 2016 The Linnean Society of London, Botanical Journal of the Linnean Society, 2016

ADDITIONAL KEYWORDS: Apodanthaceae – Aristolochiaceae – Boraginales – Cistaceae – Cynomoriaceae – Dasypogonaceae – Dilleniales – Francoaceae – Gesneriaceae – Icacinales – Metteniusales – Orobanchaceae – Phrymaceae – Phytolaccaceae – Resedaceae – Restionaceae – Sabiaceae – Santalales – Vahliales.

In 1998, the first Angiosperm Phylogeny Group (APG) classification of the orders and families of flowering plants (which we will term APG I; APG, 1998) was published, and this classification initiated a new approach to this long tradition. APG I was not written by one or two authoritative individuals; rather the APG process tried to produce a consensus classification that reflected results and opinions of experts in many groups of flowering plants. The initial focus was to produce a classification of families in orders, without too much emphasis on the issue of family delimitation; in 1998, few families had been appropriately studied, and so such issues had limited consideration in APG I. Exceptions were families, such as Saxifragacaeae (Morgan & Soltis, 1993), Geraniaceae (Price & Palmer, 1993), Liliaceae (Chase et al., 1995), Onagraceae (Conti, Fischbach & Sytsma, 1993) and Ericaceae (Judd & Kron, 1993; Kron & Chase, 1993), that had been the focus of early molecular studies, some of them due to their suspected polyphyletic nature (e.g. Saxifragaceae sensu Cronquist, 1981). Because the rule of priority does not apply at the level of order, the biggest issue in APG I was standardization of names being applied to orders so that researchers (many of them using molecular techniques) studying similar sets of families were not using different names. Prevention of chaos was the objective, and consensus was relatively easily reached. The summary consensus tree (fig. 1, p. 535) provided in APG I was highly unresolved, an indication of the preliminary nature of what was known at that time about higher-level (interordinal) relationships, even though the composition of what were then considered orders was reasonably clear.

The general scheme of the arrangement of major groups was also clear: a grade of isolated taxa (the ANA grade, or ANITA grade as it was then called) leading to the major radiation of angiosperms, a clade of all monocots, a clade of magnoliid families and a large eudicot (tricolpate) clade composed of several small clades and two major groups, rosids and asterids, each composed of two major subclades.

As the general framework of angiosperm relationships became clearer, the focus started to shift toward issues of family delimitation, with an emphasis on those that most angiosperm taxonomists had a sense might be problematic, such as Dioscoreaceae (Caddick et al., 2002), Flacourtiaceae (Chase et al., 2002), Lamiaceae/Verbenaceae (Wagstaff & Olmstead, 1997), Loganiaceae (Backlund, Oxelman & Bremer, 2000), Malvaceae (Judd & Manchester, 1997; Bayer et al., 1999), Rutaceae (Chase, Morton & Kallunki, 1999) and others. These early studies of putatively problematic families resulted in mostly clearcut solutions, especially if one followed the principles of Backlund & Bremer (1998; Vences et al., 2013 presented a zoological perspective on these same issues). Again, gaining a consensus was relatively straightforward, and APG I (2003) tried to make this easier by offering APG users optional circumscriptions, narrower and broader (a ‘bracketed’ system), permitting any permutation thereof and still allowing authors to claim that they were ‘following APG’. Delimitation of families was clearly becoming a major issue, and the use of the bracketed system was, in addition to an attempt to maintain a broad consensus of support, focused much more on family delimitation and the issue of lumping versus splitting.

In response to negative reactions received by the compilers on the use of the bracketed system, this usage was abandoned in APG I (2009). In most, but not all, cases the broader circumscriptions implied by the bracketing were accepted. As reviewed by Wearn et al. (2013) and Christenhusz et al. (2015), this lumping approach was made in an effort to simplify the parts of a classification that users emphasize, principally orders and families, an approach that has generally received support. An additional effort to assess support from both taxonomists and users of classification for broader versus narrower circumscriptions was made by conducting an online survey in August 2014 (Christenhusz et al., 2015), with the realization that any survey may have biases due to the way questions are phrased.

Higher-level classification of angiosperms has received continuing attention since APG I (2009) and enough progress has been made that an update to the APG classification is warranted. Several important studies have been published since 2009

© 2016 The Linnean Society of London, Botanical Journal of the Linnean Society, 2016


(APG I), particularly those of Soltis et al. (2011), Ruhfel et al. (2014) and Stull et al. (2015). Soltis et al. (2011) used 17 genes from all three genomes for 640 angiosperm taxa, whereas Ruhfel et al. (2014) used 78 protein-coding plastid genes for 360 green plant taxa (including green algae). Both analyses reached similar general conclusions for the angiosperms. Stull et al. (2015) concentrated on the lamiids, but this was the clade in which the greatest uncertainty existed, particularly with the former Icacinaceae, which had been known to be polyphyletic (Savolainen et al., 2000).

Researchers have speculated about what analyses of low-copy nuclear genes would reveal about plant relationships and whether these relationships would be different from those portrayed so far by plastid, mitochondrial and nuclear ribosomal genes. Nuclear data, particularly low-copy genes, have so far been poorly represented in broader phylogenetic studies of the angiosperms. Morton (2011) surveyed xanthine dehydrogenase (Xdh) for 247 genera of seed plants and obtained results generally congruent with those of previous studies, although the branching order within some larger clades was different from other

Magnoliids Commelinids

Fabids Malvids

Campanulids Lamiids


Eudicots Rosids



Superrosids Superasterids

Amborellales Nymphaeales Austrobaileyales Magnoliales Laurales Piperales Canellales Chloranthales Arecales Poales Commelinales Zingiberales Asparagales Liliales Dioscoreales Pandanales Petrosaviales Alismatales Acorales Ceratophyllales Ranunculales Proteales Trochodendrales Buxales Gunnerales Fabales Rosales Fagales Cucurbitales Oxalidales Malpighiales Celastrales Zygophyllales Geraniales Myrtales Crossosomatales Picramniales Malvales Brassicales Huerteales Sapindales Vitales Saxifragales †Dilleniales Berberidopsidales Santalales Caryophyllales Cornales Ericales Aquifoliales Asterales Escalloniales Bruniales Apiales Dipsacales Paracryphiales Solanales Lamiales †Vahliales Gentianales †Boraginales Garryales †Metteniusales


Figure 1. Interrelationships of the APG IV orders and some families supported by jackknife/bootstrap percentages >50 or Bayesian posterior probabilities >0.95 in large-scale analyses of angiosperms. See text for literature supporting these relationships. The alternative placements representing incongruence between nuclear/ mitochondrial and plastid results for the Celastrales/Oxalidales/ Malpighiales (COM) clade are indicated by slash marks (\\). †Orders newly recognized in APG.

© 2016 The Linnean Society of London, Botanical Journal of the Linnean Society, 2016

APG IV 3 studies. Zeng et al. (2014) and Wickett et al. (2014) both analysedlow-copynuclear genes (59and 852 genes, respectively),but relativelyfew angiosperms (60 and 37, respectively, the latter focused on all green plants), and reached similar conclusions about relationships to those found in the majority of earlier studies.

Although the results using low-copy nuclear genes may not substantially alter our ideas of the major framework of relationships within the angiosperms, there is at least one consistent and significant difference. Celastrales, Oxalidales and Malpighiales (the COM clade), in general found in the fabid clade of rosids (rosid I) based on the mostly plastid DNA results published up to 2011, are instead members of the malvid clade (rosid I) in trees inferred from lowcopy nuclear and mitochondrial genes. This is consistent with the nuclear results of Morton (2011), Zeng et al. (2014) and Wickett et al. (2014) and mitochondrial results of Zhu et al. (2007) and Qiu et al. (2010). Sun et al. (2015) reviewed the history of these incongruent results and added additional studies of mitochondrial and nuclear genes. It is possible that some sort of horizontal transfer of plastid DNA, perhaps via ancient hybridization, produced this incongruence (Sun et al., 2015). We have indicated this incongruence in Figure 1. It is not yet clear if this incongruence extends to Zygophyllaceae, which fell as sister to the rest of the fabid clade (including the COM clade) in plastid analyses in Sun et al. (2015).

In this update of APG, there are some changes from

APG I as a result of placements of some genera that required erection of new families, and we recognize several new orders as a result of studies incorporating many genes/whole plastid genomes (Soltis et al., 2011; Ruhfel et al., 2014; Stull et al., 2015), for example Boraginales, Dilleniales, Icacinales and Metteniusales (see below). We deviate here from previous APG papers in placing the families in the linear order of Haston et al. (2009; LAPG) and provide comments on changes and other issues in the text below, thus keeping the linear sequence of orders and families intact. For a formal, higher-level classification of plants, see Cantino et al. (2007) and Chase & Reveal (2009), which can still be applied to this version of APG. Recently, linear orders and revised classifications have been published for ferns and lycopods (Smith et al., 2006; Christenhusz & Chase, 2014) and gymnosperms (Christenhusz et al., 2011), which provide companion classifications for the remainder of the vascular plant flora.

(Parte 1 de 7)