The Families and Genera of Vascular Plants

THE FAMILIES AND GENERA OF VASCULAR PLANTS

Edited by K. Kubitzki Volumes published in this series

Volume I Pteridophytes and Gymnosperms Edited by K.U. Kramer and P.S. Green (1990) Date of publication: 28.9.1990

Volume II Flowering Plants. Dicotyledons. Magnoliid, Hamamelid and Caryophyllid Families Edited by K. Kubitzki, J.G. Rohwer, and V. Bittrich (1993) Date of publication: 28.7.1993

Volume III Flowering Plants. Monocotyledons: Lilianae (except Orchidaceae) Edited by K. Kubitzki (1998) Date of publication: 27.8.1998

Volume IV Flowering Plants. Monocotyledons: Alismatanae and Commelinanae (except Gramineae) Edited by K. Kubitzki (1998) Date of publication: 27.8.1998

Volume V Flowering Plants. Dicotyledons: Malvales, Capparales and Non-betalain Caryophyllales Edited by K. Kubitzki and C. Bayer (2003) Date of publication: 12.9.2002

Volume VI Flowering Plants. Dicotyledons: Celastrales, Oxalidales, Rosales, Cornales, Ericales Edited by K. Kubitzki (2004) Date of publication: 21.1.2004

Volume VII Flowering Plants. Dicotyledons: Lamiales (except Acanthaceae including Avicenniaceae) Edited by J.W. Kadereit (2004) Date of publication: 13.4.2004

Volume VIII Flowering Plants. Eudicots: Asterales Edited by J.W. Kadereit and C. Jeffrey (2007)

Volume IX Flowering Plants. Eudicots: Berberidopsidales, Buxales, Crossosomatales, Fabales p.p., Geraniales, Gunnerales, Myrtales p.p., Proteales, Saxifragales, Vitales, Zygophyllales, Clusiaceae Alliance, Passifloraceae Alliance, Dilleniaceae, Huaceae, Picramniaceae, Sabiaceae Edited by K. Kubitzki (2007) The Families and Genera of Vascular Plants

Edited by K. Kubitzki

Flowering Plants · Eudicots IX Berberidopsidales, Buxales, Crossosomatales, Fabales p.p., Geraniales, Gunnerales, Myrtales p.p., Proteales, Saxifragales, Vitales, Zygophyllales, Clusiaceae Alliance, Passifloraceae Alliance, Dilleniaceae, Huaceae, Picramniaceae, Sabiaceae Volume Editor : K. Kubitzki in Collaboration with C. Bayer and P.F. Stevens

With 174 Figures

123 Professor Dr. Klaus Kubitzki Universität Hamburg Biozentrum Klein-Flottbek und Botanischer Garten Ohnhorststraße 18 22609 Hamburg Germany

Library of Congress Control Number: 2006928744

ISBN-10 3-540-32214-0 Springer Berlin Heidelberg New York ISBN-13 978-3-540-32214-6 Springer Berlin Heidelberg New York

This work is subject to copyright. All rights are reserved, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilm or in any other way, and storage in data banks. Duplication of this publication or parts thereof is permitted only under the provisions of the German Copyright Law of September 9, 1965, in its current version, and permissions for use must always be obtained from Springer-Verlag. Violations are liable for prosecution under the German Copyright Law.

Springer is a part of Springer Science+Business Media springer.com

© Springer-Verlag Berlin Heidelberg 2007

The use of general descriptive names, registered names, trademarks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use.

Cover design: WMXDesign, Heidelberg, Germany Typesetting and production: LE-TEX Jelonek, Schmidt & Vöckler GbR, Leipzig, Germany

Printed on acid-free paper 31/3150/YL – 5 4 3 2 1 0 Preface

The present volume contains treatments of various eudicot orders which all are strongly supported in molecular analyses. A first group comprises Proteales, Buxales and the enigmatic Sabiaceae which, together with Ranunculales and Trochodendrales, treated earlier in Vol. II of this series, represent the basal grade of early-diverging eudicots. Al- though all of them have clearly tri-orate pollen, making them eudicots, they otherwise lack the strict eudicot floral organisation, particularly with regard to flower merosity, phyllotaxis and perianth structure. The same is true for the order Gunnerales which, however, according to findings of molecular systematics, forms part of the strongly sup- ported core eudicots. All these orders to some extent bridge the morphological gap be- tween basal angiosperms and typical core eudicots. Plesiomorphic floral traits, although less pronounced, are also found among the isolated but clearly core eudicotyledonous Saxifragales, treated in this volume, and particularly in the early-diverging woody fam- ilies of this order. Some of these latter families were included in Vol. II, which followed a classification used prior to the discovery of the modern concept of Saxifragales. The exact interrelationships among the early-diverging eudicot orders still remain largely unresolved. This is even more true for many of the core eudicot orders included in the present volume, i.e. Vitales, Crossosomatales, Geraniales, Zygophyllales and Myrtales, and also for two of the families not assigned to order, i.e. Huaceae and Picramniaceae. A possible relationship between Dilleniaceae and woody Caryophyllales, as suggested by recent studies, opens an interesting new perspective on the evolution of this family. Included in this volume are two subclades of the vast Malpighiales. These are Passi- floraceae with two satellite families, and Clusiaceae/Hypericaceae with Podostemaceae, which recently have been identified as very close relatives. My deep thanks go to all authors of this volume, who have provided such highly in- teresting and scholarly contributions, and to all those who have freely shared additional information and/or have commented on earlier drafts of the contributions. These include B.G. Briggs, T. Clifford, G. Jordan, B. Makinson, P. Olde, R. Barker, J.A. Doyle, P.J. Rudall and C.A. Furness (Proteaceae); H. Manitz (Aphanopetalaceae); A.E. Orchard (Halor- agaceae); P.H. Linder and M. Weigend (Geraniales); A. Bernhard and W.J.J.O. de Wilde (Passifloraceae); the late M. Ricardi S. (Malesherbiaceae); I. Jäger-Zürn and T. Philbrick (Podostemaceae); and S. Renner, P.G. Wilson and J. Schönenberger (Myrtales). I am also grateful to M.L. Matthews and P.K. Endress for showing me their papers prior to publication elsewhere. Mark C. Chase is thanked for always making available the newest results of his pathbreaking studies. The copyright holders of the illustrations included in this volume are thanked for their generous permission to use their valuable material. As always, it is a pleasure to acknowledge the agreeable collaboration with the staff of Springer-Verlag, who kindly responded to all requests I had in connection with the production of this volume, and to thank Monique T. Delafontaine for her meticulous copy editing of the manuscript.

Hamburg, September 2006 K. Kubitzki Contents

Introduction to the Groups Treated in this Volume K. Kubitzki ...... 1 IntroductiontoBerberidopsidales ...... 1 IntroductiontoBuxales ...... 2 IntroductiontotheClusiaceaeAlliance(Malpighiales) ...... 3 IntroductiontoCrossosomatales ...... 3 IntroductiontoFabales ...... 5 IntroductiontoGeraniales ...... 5 IntroductiontoGunnerales ...... 7 IntroductiontoMyrtales ...... 7 Introduction to the Passifloraceae Alliance (“Passiflorales” = Malpighiales) ...... 12 IntroductiontoProteales ...... 12 IntroductiontoSaxifragales ...... 15 IntroductiontoVitales ...... 18 IntroductiontoZygophyllales ...... 19 FamiliesUnassignedtoOrder ...... 20 General References ...... 21 Aextoxicaceae K. Kubitzki ...... 23 Alzateaceae S.A. Graham ...... 26 Aphanopetalaceae K. Kubitzki ...... 29 Aphloiaceae K. Kubitzki ...... 31 Berberidopsidaceae K. Kubitzki ...... 33 Bonnetiaceae A.L. Weitzman, K. Kubitzki and P.F.Stevens ..... 36 Buxaceae E. Köhler ...... 40 Clusiaceae-Guttiferae P.F. Stevens ...... 48 Combretaceae C.A. Stace ...... 67 Crassulaceae J. Thiede and U. Eggli ...... 83 Crossosomataceae V. S o s a ...... 119 Crypteroniaceae S.S. Renner ...... 123 Daphniphyllaceae K. Kubitzki ...... 127 VIII Contents

Didymelaceae E. Köhler ...... 129 Dilleniaceae J.W. Horn ...... 132 Geissolomataceae F. Forest ...... 155 Geraniaceae F. Albers and J.J.A. Van der Walt ...... 157 Grossulariaceae M. Weigend ...... 168 Gunneraceae H.P. Wilkinson and L. Wanntorp ...... 177 Haloragaceae K. Kubitzki ...... 184 Huaceae C. Bayer ...... 191 Hypericaceae P.F. Stevens ...... 194 Iteaceae K. Kubitzki ...... 202 Ixerbaceae J.V. Schneider ...... 205 Krameriaceae B.B. Simpson ...... 208 Ledocarpaceae M. Weigend ...... 213 Leeaceae J. Wen ...... 221 Lythraceae S.A. Graham ...... 226 Malesherbiaceae K. Kubitzki ...... 247 Melianthaceae H.P. Linder ...... 250 Oliniaceae M. von Balthazar and J. Schönenberger ...... 260 Paeoniaceae M. Tamura ...... 265 Passifloraceae C. Feuillet and J.M. MacDougal ...... 270 Penaeaceae J. Schönenberger, E. Conti and F. Rutschmann . 282 Penthoraceae J. Thiede ...... 292 Peridiscaceae C. Bayer ...... 297 Picramniaceae K. Kubitzki ...... 301 Podostemaceae C.D.K. Cook and R. Rutishauser ...... 304 Polygalaceae B. Eriksen and C. Persson ...... 345 Proteaceae P.H. Weston ...... 364 Pterostemonaceae K. Kubitzki ...... 405 Quillajaceae K. Kubitzki ...... 407 Rhynchocalycaceae J. Schönenberger ...... 409 Sabiaceae K. Kubitzki ...... 413 Saxifragaceae D.E. Soltis ...... 418 Stachyuraceae J.V. Schneider ...... 436 Staphyleaceae S.L. Simmons ...... 440 Strasburgeriaceae W.C. Dickison ...... 446 Contents IX

Surianaceae J.V. Schneider ...... 449 Tetracarpaeaceae K. Kubitzki ...... 456 Turneraceae M.M. Arbo ...... 458 Vitaceae J. Wen ...... 467 Vo chysia ceae M.L. Kawasaki ...... 480 Zygophyllaceae M.C. Sheahan ...... 488 AdditionsandCorrectionstoVolumesII–VI ...... 501 IndextoScientificNames ...... 503 List of Contributors

Albers, Focke Botanischer Garten, Universität Münster, Schlossgarten 3, 48149 Münster, Germany Arbo, María M. Instituto de Botánica del Nordeste, Casilla de Correo 209, 3400 Corrientes, Rep. Argentina Balthazar, Maria Swedish Museum of Natural History, Division of von Palaeobotany, P.O. Box 50007, 10405 Stockholm, Sweden Bayer, Clemens Palmengarten der Stadt Frankfurt, Siesmayerstr. 61, 60323 Frankfurt/Main, Germany Conti, Elena Institut für Systematische Botanik und Botanischer Garten, Universität Zürich, Zollikerstr. 107, 8008 Zürich, Switzerland Cook, Christopher Institut für Systematische Botanik und Botanischer Garten, D.K. Universität Zürich, Zollikerstr. 107, 8008 Zürich, Switzerland Dickison, W.C. (deceased) Eggli, Urs Sukkulenten-Sammlung, Mythenquai 88, 8002 Zürich, Switzerland Eriksen, Bente Department of Botany, University of Göteborg, P.O. Box 461, 40530 Göteborg, Sweden Feuillet, Christian Department of Botany, Smithsonian Institution, P.O. Box 37012, Washington, DC 20013-7012, USA Forest, Felix L. School of Biological Sciences, Plant Sciences Laboratories, The University of Reading, Reading RG6 6AS, UK Graham, Shirley A. Missouri Botanical Garden, P.O. Box 299, St. Louis, MO 63166-0299, USA Horn, James W. Department of Biology, Duke University, P.O. Box 90338, Durham, NC 27708-0338, USA. Present address: Fairchild Tropical Botanic Garden, 11935 Old Cutler Road, Miami, FL 33156-4242, USA Kawasaki, M. Lúcia Department of Botany, Field Museum, 1400 S. Lake Shore Drive, Chicago, IL 60605-2496, USA Köhler, Egon Spezielle Botanik und Arboretum, Humboldt-Universität, Späthstr. 80/81, 12437 Berlin, Germany Kubitzki, Klaus Biozentrum Klein-Flottbek und Botanischer Garten, Universität Hamburg, 22609 Hamburg, Germany XII List of Contributors

Linder, H. Peter Institut für Systematische Botanik und Botanischer Garten, Universität Zürich, Zollikerstr. 107, 8008 Zürich, Switzerland MacDougal, John M. Missouri Botanical Garden, P.O. Box 299, St. Louis, MO 63166-0299, USA Persson, Claes Department of Botany, University of Göteborg, P.O. Box 461, 40530 Göteborg, Sweden Renner, Susanne S. Botanische Staatssammlung, Menzinger Str. 67, 80638 München, Germany Rutishauser, Rolf Institut für Systematische Botanik und Botanischer Garten, Universität Zürich, Zollikerstr. 107, 8008 Zürich, Switzerland Rutschmann, Frank Institut für Systematische Botanik und Botanischer Garten, Universität Zürich, Zollikerstr. 107, 8008 Zürich, Switzerland Schneider, Julio V. Spezielle Botanik, Universität Leipzig, Johannisallee 21–23, 04103 Leipzig, Germany Schönenberger, Jürg Department of Botany, University of Stockholm, Lilla Frescativägen 5, 10691 Stockholm, Sweden Sheahan, Mary C. Jodrell Laboratory, Royal Botanic Gardens Kew, Richmond SRY TW9 3DS, UK Simmons, Sara L. Department of Integrative Biology, University of Texas, Austin, TX 78712, USA Simpson, Beryl B. University of Texas, Section of Integrative Biology, 1 University Station A 6700, Austin, TX 78712, USA Soltis, Douglas E. Department of Botany, University of Florida, Gainesville, FL 32611-7800, USA Sosa, Victoria Instituto de Ecología, A.C., Apartado Postal 63, 91000 Xalapa, Veracruz, México Stace, Clive A. Department of Botany, University of Leicester, University Road, Leicester LE1 7RH, UK Stevens, Peter F. Missouri Botanical Garden, P.O. Box 299, St. Louis, MO 63166-0299, USA Tamura, Michio 4-25-7 Ao-gein, Mino, Osaka 562-0025, Japan Thiede, Joachim Schenefelder Holt 3, 22589 Hamburg, Germany Van der Walt, J.J.A. (deceased) Wanntorp, Livia Department of Botany, University of Stockholm, Lilla Frescativägen 5, 10691 Stockholm, Sweden Weigend, Institut für Biologie/Systematische Botanik, Maximilian Freie Universität Berlin, Altensteinstr. 6, 14195 Berlin, Germany Weitzman, Anna L. Department of Botany, Smithsonian Institution, P.O. Box 37012, Washington, DC 20013-7012, USA List of Contributors XIII

Wen, Jun Department of Botany, Smithsonian Institution, P.O. Box 37012, Washington, DC 20013-7012, USA Weston, Peter Royal Botanic Gardens, Mrs Macquaries Road, Sydney, NSW 2000, Australia Wilkinson, Hazel P. Jodrell Laboratory, Royal Botanic Gardens Kew, Richmond SRY TW9 3DS, UK Introduction to the Groups Treated in this Volume

K. Kubitzki

Introduction to Berberidopsidales gynoecia. The spiral sequence of initiation of floral organs in Berberidopsis,withatendencyof 1. Dioecious tree, young parts covered with ferrugineous arrangement in alternating groups of five, may scales; leaves conduplicate, entire; flowers 5(6)- represent an incipient case of pentamery (Ronse merous, enveloped in bud by firm calyptrate bract; DeCraene 2004) but this is problematic, in view of stamens 5, alternating with nectary glands; gynoecium 1-carpellate; ovules 2, pendulous from apex of locule; the firmly established pentamerous floral structure style apically bifid; fruit dry, indehiscent; seed with characteristic for core eudicots which exists in ruminate endosperm and embryo of about half the parts of Berberidopsidaceae and in the closely length of seed. 1/1, S Chile and adjacent Argentina related Aextoxicum (see Berberidopsidaceae and Aextoxicaceae – Scandent shrubs, largely glabrous; leaves involute Aextoxicaceae, this volume). (Berberidopsis), spiny-toothed or entire; flowers hermaphrodite, acyclic and with disk, or cyclic, pen- tamerous and without disk; gynoecium 3–5-carpellate; ovules several to many, each on 3–5 placentas; style not bifid; fruit berry-like; embryo small. 2/3, S Chile and SE Australia Berberidopsidaceae

A close relationship between Berberidopsidaceae and Aextoxicaceae has never been considered un- til gene sequence studies provided strong support for a relationship between them (see family treat- ments). In the four-gene analysis of eudicots (Soltis et al. 2003), Gunnerales and subsequently Berberi- dopsidales are sister to all other core eudicots, the latter being strongly supported by molecular data and isolated from all other clades (Fig. 1). Aextox- icum has long been known for its peculiar wood anatomy, particularly the high number of bars of the vessel element perforations. A recent study by Carlquist (2003) has revealed many important sim- ilarities in the wood anatomy of the two families, although these are plesiomorphic. Pollen grains are relatively small and tricolpate to indistinctly colpo- rate. The two families share encyclocytic stomata (Soltis et al. 2005), a rare character in angiosperms, stout filaments, and a ring of vascular bundles in the petiole (Judd and Olmstead 2004). Unfortunately, many important characters are not known for both taxa but available information shows that Berberidopsidales are very plastic in their floral structure, combining (even within the same family, Berberidospidaceae) both spiral Fig. 1. A phylogenetic hypothesis of eudicot relationships, and whorled patterns, and 1-, 3- and 5-merous based on a four-gene dataset (Soltis et al. 2003) 2K.Kubitzki

Morphologically, basal eudicots exhibit con- peculiar steroidal pregnan alkaloids. The most siderable structural disjunctions, which underlines obvious trait of Buxales is the plasticity and sim- their relict nature. This is also corroborated by plicity of perianth organisation. In some of their the remarkable angiospermous fossil from the members (Didymeles,maleStyloceras), a perianth Early Cretaceous, Teixeira lusitanica, which shows is completely lacking and, in Buxaceae, the tepals affinities to members of Ranunculales, and to hardly differ from vegetative bracts below the Berberidopsidaceae, Hamamelidaceae and Daph- flower (von Balthazar and Endress 2002a) and in niphyllaceae (von Balthazar et al. 2005). Characters female flowers they are spirally arranged, making such as the dimerous floral structure, known from the delimitation of flowers difficult. The stamens Gunnera, and presumably plesiomorphic traits are always antesepalous and the stamen-sepalum (decurrent stigmas, antepetalous stamens, etc.), complex of Buxaceae is similar to that of Pro- known from other basal eudicot families such teaceae, also in the supply of the sepals by a single as Proteaceae and Sabiaceae, are not found in trace. Stamens, when occurring in low number, Berberidopsidales. are arranged in dimerous whorls but, for higher numbers (in Notobuxus 6, 8, and up to more than 40), less regular arrangements prevail. References Palynologically, Buxales are highly diverse (Bessedik 1983; Doyle 1999). An early fossil at- Balthazar, M. von, Pedersen, K.R., Friis, M.E. 2005. Teixeira tributable to Buxales (Doyle 1999) is a pollen from lusitanica,anewfossilflowerfromtheEarlyCreta- the Aptian/Albian of northern Gondwana, which ceous of Portugal with affinities to Ranunculales. Pl. has simple colpate apertures and a striate(-re- Syst. Evol. 255:55–75. Carlquist, S. 2003. Wood anatomy of Aextoxicaceae and ticulate) sculpture and has been related to the Berberidopsidaceae is compatible with their inclusion buxaceous megafossil Spanomera (Drinnan et al. in Berberidopsidales. Syst. Bot. 28:317–325. 1991). In the late Albian of Gabon and Brazil, Judd, W.S., Olmstead, R.G. 2004. A survey of tricolpate the tricolpodiorate pollen Hexaporotricolpites (eudicot) phylogenetic relationships. Amer. J. Bot. 91:1627–1644. (Boltenhagen 1967) appears. This pollen type may Ronse DeCraene, L.P. 2004. Floral development of Berberi- be related to extant Didymeles from Madagascar dopsis corallina: a crucial link in the evolution of flow- (cf. Fig. 36), which has left a fossil record in the ers in the core eudicots. Ann. Bot. 94:741–751. southern Indian Ocean, Australia, New Zealand Soltis, D.E. et al. 2003. See general references. and New Caledonia. Similar pollen grains with an Soltis, D.E. et al. 2005. See general references. increasing number of pores and meridional colpi, later in pantocolporate and eventually pantoporate Introduction to Buxales configuration, the latter combined with a crotonoid exine pattern (cf. Fig. 11D), appear both in the fossil record and in extant Buxus (Köhler 1981; 1. Dioecious trees; flowers apetalous, male with one stamen pair, female often paired, a single carpel; Köhler and Brückner 1982; Bessedik 1983). pollen grains tricolpo-di-orate; seeds exalbuminous. Buxalesformpartofthegradeofearly- 1/2, Madagascar Didymelaceae diverging tricolpate(-derived) dicots or eudicots, – Monoecious, rarely dioecious shrubs or herbs; flowers which also comprises Ranunculales, Sabiaceae, with weakly differentiated perianth, male with decus- sate tepals and 4, 6 or more stamens, female with spiral Proteales and Trochodendraceae (cf. Fig. 1). With tepals and a 2–4-carpellate, syncarpous gynoecium; several early-diverging eudicots, and partly also pollen grains 3–7-colporate with 3–6 pores per colpus, with some basal core eudicots (Gunneraceae, or pantoporate; seeds albuminous. 5/c. 100, all conti- Myrothamnaceae and some basal families of nents, except Australia Buxaceae Saxifragales), Buxales share characters which are Buxales comprise Buxaceae and Didymelaceae, known also from the eumagnoliids. Particularly grouped together by traits such as cyclocytic remarkable are the dimerous flowers, the supply stomata, leaf venation pattern, wood anatomical of the sepals by a single trace, and the stamen- peculiarities including many sclereids, racemose sepalum complex, in which Buxaceae agree with inflorescences, small, imperfect, often dimerous Proteaceae. Conspicuous connective protrusions flowers with decurrent stigmas extending the are known from other early-diverging eudicots and entire length of the stylodia, stamens with more or some basal core eudicots, including Proteaceae, less basifixed anthers and conspicuous connective Platanaceae, Trochodendraceae, Myrothamnaceae; anther protrusions, and the occurrence of very basifixed anthers are widespread in early-diverging Introduction to the Groups Treated in this Volume 3 eudicots. Elongate stigmas decurrent in two crests now well established (Soltis et al. 2000; Savolainen, are shared with Platanaceae, Myrothamnaceae Fay et al. 2000), and Podostemaceae appear sister and Trochodendraceae but are also found in to Hypericaceae or perhaps nested inside this fam- some Saxifragales. Nectary disks are rare in ily (Gustafsson et al. 2002). As the sinking of Po- early-diverging eudicots and, apart from the in- dostemaceae in the broadly delimited Clusiaceae trastylodial nectariferous structures in Buxaceae, would lead to a highly heterogeneous unit, it ap- are known only from Proteaceae and Sabiaceae. pears preferable to “save” an independent family Podostemaceae by segregating Hypericaceae from Clusiaceae s.l., following the approach of many ear- References lier authors such as Takhtajan (1997), although the separation in terms of contrasting characters be- Doyle, J.A. 1999. The rise of angiosperms as seen in the tween the latter two is not very strong. Charac- AfricanCretaceouspollenrecord.In:Heine,K.(ed.) ters common to the three families include resin Palaeoecology of Africa and the surrounding islands. cells and secretory ducts, containing xanthones, Rotterdam: Balkema, pp. 3–29. For other references, see the selected bibliographies of Bux- and bitegmic and tenuinucellate ovules. aceae and Didymelaceae, and the General References (this volume). References

Introduction to the Clusiaceae Alliance For references, see the Selected Bibliography of Clusiaceae- (Malpighiales) Guttiferae and the General References (this volume).

1. Annual cataract-dwellers with unclear differentiation of stems, roots and leaves (roots often crustaceous, ribbon-like; leaves sometimes terminal and double- Introduction to Crossosomatales sheathed); fertile pollen and fertilisable embryo sacs developed underwater; [pollination autogamous or 1. Perianth biseriate 2 cleistogamous, rarely allogamous; female gameto- – Perianth uniseriate; [flowers solitary] 6 phyte reduced Allium type; no double fertilisation and 2. Leaves opposite, pinnately compound, rarely unifolio- no endosperm; seed set high]. 49/c. 280, worldwide, late; embryo green; [ovary syncarpous or apocarpous; tropical and warm-temperate regions Podostemaceae ovules anatropous]. 2/40–50, temperate to tropical re- – Woodyorherbaceouslandplants 2 gions, mainly of the northern hemisphere 2. Leaves alternate, serrulate, initially setulose, convolute; Staphyleaceae latex 0; stamen connective glands 0; capsule with per- – Leaves alternate, simple (if opposite and simple to sistent column. 3/40, northern South America, West slightly trilobate and gynoecium apocarpous, see Indies, and SE Asia to New Guinea Bonnetiaceae Apacheria in Crossosomataceae); embryo achlo- – Leaves opposite or alternate, entire, not setulose, not rophyllous 3 convolute; latex often present in glands or secretory 3. Flowers solitary 4 canals; stamen connectives often with glands produc- – Flowers in panicles, racemes or spikes 5 ing oil or resin; fruit, if capsular, then rarely with per- 4. Sepals 4–10; stamens5+5;anthersdorsifixed;pistil sistent column 3 4–7-carpellate; style simple; ovules 1 per locule, ana- 3. Stylodia free, at least distally; flowers perfect; sepals tropous; fruit indehiscent, fibrous; seed with rudimen- and petals 3–5; aril 0; trichomes, if multicellular, then tary aril; embryo straight; vessel element perforation stellate; woody or herbaceous. 9/540, worldwide scalariform; T-shaped unicellular trichomes present. Hypericaceae 1/1, New Caledonia Strasburgeriaceae – Stylodia free or fused to form a simple style; flowers – Sepals (3)4–5(6); stamens 4–50 (flower haploste- perfect or unisexual; sepals 2–20, petals 0–8; aril some- monous, diplostemonous or polystemonous); anthers times present; stellate hairs very rare (Caraipa, Marila); basifixed; gynoecium apocarpous, 1–5(–9)-carpellate; woody. 27/1090, pantropical Clusiaceae ovules 1–many per carpel, campylotropous; fruit Althoughthesefamilieshavebeenintenselystudied follicular; seed arillate; embryo curved; vessel element perforation mostly simple; T-shaped trichomes 0. bygenerationsofbotanists,recentworkhasconsid- 4/10, North America, with Mexico erably modified our understanding of their phylo- Crossosomataceae genetic relationships and details of their family and 5. Flowers strictly 5-merous; ovules 2 per locule; pollen tribal delimitation. Molecular studies have revealed 4(5)-colporate; fruit capsular; aril rudimentary; vessel element perforation scalariform; T-shaped unicellular one enigma of long standing – the systematic po- trichomes present. 1/1, New Zealand Ixerbaceae sition of Podostemaceae. Their close relationship – Flowers strictly 4-merous; ovules many per locule; with Clusiaceae s.l. (i.e. including Hypericaceae) is pollen 3-colporate; fruit berry-like; seed with soft 4K.Kubitzki

funicular aril; vessel element perforation simple; ously neglected floral characters, which are shared T-shaped trichomes 0. 1/c. 16, E Asia Stachyuraceae in different constellations by groups of two, three 6. Leaves decussate, entire; tepals 4; stamens4+4;an- or more families of the whole alliance. thers dorsifixed; ovary 4-locular, with 4 twisted sty- lodia; ovules anatropous; fruit capsular; seeds with The group as a whole is only weakly char- swollen funicle; embryo straight; T-shaped unicellu- acterised. Stomata are usually anomocytic. Leaf lar trichomes present. 1/1, South Africa margins are usually toothed. Stipules are lacking Geissolomataceae only in Ixerbaxceae and some Crossosomataceae. – Leaves alternate, serrate; tepals 4–5(6); stamens many; anthers basifixed; ovary unilocular; style simple; Vessel elements have scalariform perforation, ovules campylotropous; fruit a berry; seeds arillate, Crossosomataceae and Stachyuraceae excepted. incurved with hippocrepiform embryo. 1/1, E and Sepal and petal aestivation is imbricate throughout, southern Africa, islands of Indian Ocean Aphloiaceae and stamens are always incurved in bud; anthers Until very recently, these apparently disparate fam- are tetrasporangiate; nectary disks are present. ilies had been placed in different rosid orders and Ovules are bitegmic and crassinucellar, mostly some had been “dumped” in larger families such anatropous; Aphloiaceae and Crossosomataceae as the broadly construed Saxifragaceae (Ixerba) have campylotropous ovules. Pollen grains are col- or Flacourtiaceae (Aphloia). The taxonomic his- porate and usually have lalongate endoapertures; tory of the individual families is briefly described the gynoecium is often stalked; the carpel tips are in the family treatments, and has been treated in often postgenitally united to form a compitum. more depth by Matthews and Endress (2005). Al- The seed coat is testal. Sieve element plastids are S though Takhtajan (1987) for the first time used the type throughout. Ellagitannins and gallotannins, name of the order Crossosomatales which, in his but no proanthocyanidins, are known from Cros- approach, comprised only the name-giving family, sosomataceae. More restricted are the following a broader concept of the order was not achievable traits. Ixerbaceae and Strasburgeriaceae have in the pre-molecular era largely because the char- largeflowerswithpetalsformingatight,pointed acters traditionally used in higher-level classifica- cone in bud, stamens with sagittate anthers, and tion are very variable in these seven families (see a rudimentary aril. These families share with Geis- Conspectus). solomataceae T-shaped unicellular trichomes and During recent years, several molecular stud- a punctiform stigma on postgenitally united and ies have contributed to the recognition of the rela- twisted carpel tips, and only one or two ovules per tionships in the entourage of Crossosomataceae. carpel. Aphloiaceae, Geissolomataceae, Ixerbaceae In their rbcL and combined morphological and and Strasburgeriaceae share pollen grains with rbcL studies, Nandi et al. (1998) found a clade pronounced protruding endoapertures (“pollen of {[(Crossosoma + Stachyurus) Staphylea] Geis- buds”). Crossosomataceae, Stachyuraceae and soloma}, albeit without significant support. Strong Staphyleaceae have polygamous or functionally support for [(Crossosoma + Stachyurus) Staphylea] unisexual flowers, and Crossosomataceae and was adduced by further rbcL studies (Savolainen, Aphloiaceae (although not resolved as sisters in Fay et al. 2000; Sosa and Chase 2003) and multi- molecular studies) share polyandrous flowers, gene analyses (Soltis et al. 2000; Cameron 2003), basifixed anthers, a stigma with two or more and for Ixerba + Strasburgeria by rbcL(Savolainen, decurrent crests, campylotropous ovules and Fay et al. 2000; Sosa and Chase 2003) and multi- reniform seeds (data from Matthews and Endress gene studies (Cameron 2003). When included in 2005). the analysis, Ixerba + Strasburgeria, Aphloia and Crossosomatales are core eudicots but other- Geissoloma usually appeared in the same clade as wise their relationships are still unclear: they ap- Crossosoma, Stachyurus and Staphylea,although pear at the base of eurosids II (Savolainen, Fay et al. statistical support for this was low. The concept of 2000; Soltis et al. 2000) or eurosids I (Hilu et al. Crossosomatales proposed by Savolainen, Fay et al. 2003), or in a polytomy with Geraniales, Myrtales, (2000) and Soltis et al. (2000), comprising Crosso- eurosids I and eurosids II (APG II 2003), but always somataceae, Stachyuraceae and Staphyleaceae, has with low statistical support. later been extended to include all seven aforemen- tioned families (see also Stevens 2005). This con- cept is now confirmed by the broad-based compar- References ative study of Matthews and Endress (2005), which has revealed structural traits, particularly previ- For references, see the General References (this volume). Introduction to the Groups Treated in this Volume 5

Introduction to Fabales Soltis, P.S.,Soltis, D.E., Chase, M.W. 1999. Angiosperm phy- logeny inferred from multiple genes as a tool for com- parative biology. Nature 402:402–404. 1. Stylodia gynobasic; [woody; flowers regular; gynoe- Soltis, D.E. et al. 2000. See general references. cium apocarpous, 1–5-carpellate; ovule unitegmic Stevens, P.F. 2005. See general references. (only Suriana known); endosperm 0 or sparse; nectary only rarely present; vestured pits in Recchia]. 5/8, in warm-temperate and tropical regions, widely distributed Surianaceae – Style or stylodia not gynobasic 2 Introduction to Geraniales 2. Gynoecium syncarpous, 2–8-carpellate (sometimes 1- locular); pollen grains 7–28-colporate; seeds mainly 1. Embryo small, straight, achlorophyllous; endosperm endotestal; leaves estipulate; [woody or herbaceous; copious; secondary xylem always with rays; [either nodes unilacunar with a single trace; vessel element (sub)shrubs or small trees with mostly 5-merous flow- perforations usually simple; vestured pits sometimes ers and simple leaves and regular flowers (Greyia), present; flowers actinomorphic to zygomorphic; nec- = or with pinnate leaves and ± zygomorphic flowers tary a disk, a gland, or 0; seeds often arillate]. n 6–23. (Melianthus, Bersama), or herbs with mostly 4-merous 21/800–1,000, widely distributed in tropical, subtropi- flowers and commissural stigmas (Francoa, Tetilla)]. cal and temperate regions Polygalaceae 5/19, subsaharan Africa, southern South America – Gynoecium (nearly) apocarpous; pollen grains mostly Melianthaceae 3-aperturate; seeds exotestal; leaves stipulate; [nectar- – Embryo large, circinate, twisted, or cochlear, rarely iferous disk usually present] 3 (Rhynchotheca) straight, chlorophyllous; endosperm 3. Flowers strictly actinomorphic; carpels 5, only basally usually scant; secondary xylem often rayless 2 connate; pollen grains in monads; seeds exarillate; en- 2. Pollen grains tricolp(or)ate; style elongate, with 5 style dosperm thin; cotyledons convolute; stipules small; branches or (Hypseocharis) unbranched with capitate nodes unilacunar; vessel elements with simple and stigma; fruits schizocarpic with 1-seeded awned meri- scalariform perforation; vestured pits 0; [woody; bark = carps or (Hypseocharis) loculicidal capsules; seed coat strongly saponiferous]. n 14. 1/2, warm-temperate with crystalliferous endotesta and thickened but not southern South America Quillajaceae lignified exotegmen. 5/c. 835, nearly worldwide – Flowers actinomorphic to zygomorphic; carpel usu- Geraniaceae ally 1 or very rarely more (and then each carpel with – Pollen grains pantoporate or inaperturate; style very a terminal stylodium); pollen grains in monads, tetrads short, with 5–3 elongate stigmatic branches; fruits sep- or polyads; seeds arillate or not; endosperm usually 0, ticidal or septifragous capsules with 1–many-seeded rarely sparse or even copious; stipules sometimes mod- locules; seed coat usually lacking mechanical layers ified into prickles or spines; nodes tri-(penta-)lacunar; (Viviania, exotegmic); Balbisia with mucilaginous exo- vessel elements with simple perforations, the lateral testa. 4/c. 18, South America, mostly Andean pits often vestured; [roots very often with N-fixing root Ledocarpaceae nodules]. 640/1800, widely distributed throughout the world The former, broadly construed concept of the or- Leguminosae–Fabaceaes.l.(nottreatedinthisvolume) der Geraniales (e.g. Engler 1892) comprised 15–20 A clade comprising these four families was resolved families including disparate groups such as Oxal- as belonging to eurosids I by early molecular stud- idaceae, Tropaeolaceae, Zygophyllaceae, Rutaceae ies (Chase et al. 1993; Fernando et al. 1993; Morgan and Euphorbiaceae. Based on the work of many etal.1994)andisstronglysupportedinseveralmul- authors, notably Takhtajan (e.g. 1959, 1987) and tigene analyses (e.g. Soltis et al. 1999, 2000). Mor- Dahlgren (e.g. 1980), Geraniales were stepwise re- phologically, the four families have little in com- stricted by the exclusion of orders such as Rutales, mon, apart from the basically core eudicot floral Polygalales and Malpighiales. Yet,in a still more re- organisation. Stevens (2005) notes green embryos cent survey of dicotyledon families, Thorne (2001) and often fluorescent wood, and absence of ellagi- merged Geraniales with Linales (= Malpighiales), tannins (which are, however, present in Legumi- mainly on account of these sharing a tendency nosae) as common traits. for obdiplostemonous flowers with 10–15 stamens and a 5-partite gynoecium. It is difficult to un- derstand why Oxalidaceae, for instance, were for so long considered to belong to Geraniales, al- References though the former differ in possessing traits such asfreestylodia,abundantendosperm,andcapsular Chase, M.W. et al. 1993. See general references. fruits (see treatment of Oxalidales in Vol. VI of this Fernando, E.S. et al. 1993. See selected bibliography of Suri- anaceae. series). Morgan, D.R. et al. 1994. See selected bibliography of Quil- In the pioneering molecular studies of Price lajaceae. and Palmer (1993), Morgan and Soltis (1993) and 6K.Kubitzki

Chase et al. (1993), the five genera of Geraniaceae Geraniaceae agree, however, in numerous traits, s.str. grouped with Hypseocharis and, in some anal- such as rayless wood, acuminate to awned sepals, yses, also with Viviania, Greyia and Francoa, and broadened filaments and sometimes basal nectarif- often also with Crossosoma, whereas other families erous appendages, tanniniferous seed coats, green of the erstwhile Geraniales were clearly excluded. embryos and (Rhynchotheca excepted) curved or More recent molecular work (Soltis et al. 2000; cochlear embryos. Differences between the two Savolainen, Fay et al. 2000; Sosa and Chase 2003) families exist in growth habit, pollen morphology has resolved Geraniales and Crossosomatales (see and seed coat anatomy (Boesewinkel 1997). treatment in this volume) as sister groups placed Anthecologically, Geraniales families are di- at the base of eurosid II orders, but the statistical versified. Nectaries are present in all families, and support is tenuous and the Angiosperm Phylogeny Geraniaceae depend on a broad variety of insect Group (APG II 2003) lists both orders among the groups as pollinators and only occasionally on unresolved rosids. Evidence for a sister group bird, whereas Melianthaceae rely strongly on bird relationship of Melianthaceae and Ledocarpaceae pollination. In Ledocarpaceae, Viviania produces and, in turn, of both of these with Geraniaceae is copious nectar as reward for insect and other provided by the rbcL analyses of Savolainen, Fay pollinators whereas the remaining genera, Balbisia et al. (2000) and Sosa and Chase (2003). This is and Rhynchotheca, lack nectaries. Balbisia has notable, as morphologically Ledocarpaceae and pollen flowers, and its showy corolla indicates that Geraniaceaeseemtohavemoreincommonthan it is also zoophilous; one species, B. gracilis,may any of them have with Melianthaceae (see below). be anemophilous. Rhynchotheca has apetalous Support is also strong for recognising Melanthi- flowers with large, pendulous anthers and shows aceae and Geraniaceae in the circumscription synchronous mass-flowering, all indicative of adopted in this volume but it is less strong for anomophily (Weigend 2005). Thus, the morpho- the morphologically more diverse Ledocarpaceae. logical disparity of Ledocarpaceae appears to be Biebersteinia, a Eurasian monotype which often related to their range of pollination syndromes, had been related to Geraniaceae (e.g. Knuth which may explain the difficulties morphological 1931), agreeing with this family in some details of workers had in recognising the circumscription fruit and seed morphology (Boesewinkel 1988), and affinities of the family. in molecular studies is resolved as sapindalean Phytochemically, Geraniales are characterised (APG II 2003). by the typical presence of ellagitannins; ellagic acid A morphological characterisation of Gerani- has been recorded from all genera (Tetilla not stud- ales is difficult because the genera of Melianthaceae ied); gallotannins are also present in Geranicaeae, and, less so, of Ledocarpaceae are diverse. It is and geraniin, an ellagitannin based on dehydroxy- true that all Geraniales have anomocytic stomata, hexahydroxydiphenic acid, is a prominent com- vessels with simple perforation, 5- or 4-merous pound in Geranium. Proanthocyanidins are uni- hypogynous flowers with a persistent calyx, and formly lacking from aerial parts but occur in seed either haplostemonous or more often obdiploste- coats and are recorded also from the rootstocks monous androecia with paired antepetalous of Geraniaceae. Other biodynamic compounds in- stamens, 5(–3)-carpellate syncarpous gynoecia clude the bufodienolides and pentacyclic triter- with simple styles (extremely short in Balbisia)ter- penoids in Melianthaceae (Hegnauer 1969, 1989). minating in 3–5 stigmatic lobes or branches, axile or rarely basal placentation, anatropous to campy- lotropous bitegmic and crassinucellar ovules and, where known, a Polygonum type embryo sac References and Nuclear endosperm. All Melianthaceae have copious endosperm, and Melianthus, Bersama and APG II 2003. See general references. Boesewinkel, F.D. 1988. The seed structure and taxonomic Greyia share multilacunar nodes. Francoa and relationships of Hypseocharis Remy. Acta Bot. Neerl. Tetilla, formerly included in Saxifragaceae, differ 37:111–120. from this family in the commissural stigma, the Boesewinkel, F.D. 1997. Seed structure and phylogenetic re- 4-merous flowers, Nuclear endosperm and the lack lationships of the Geraniales. Bot. Jahrb. Syst. 119:277– 291. of myricetin. Otherwise, multilacunar nodes are Chase, M.W. et al. 1993. See general references. not known in Geraniales, and the endosperm is ab- Dahlgren, R. 1980. A revised system of classification of the sent or scanty in Geraniaceae. Ledocarpaceae and angiosperms. Bot. J. Linn. Soc. 80:91–124. Introduction to the Groups Treated in this Volume 7

Engler, A. 1892. Syllabus der Vorlesungen über spezielle und many Ranunculaceae but usually also co-occurs medizinisch-pharmazeutische Botanik. Berlin: Gebr. with trimery in basal angiosperms (Kubitzki Borntraeger. 1987) or even with pentamery in early-diverging Hegnauer, R. 1969, 1989. See general references. Knuth, R. 1931. Geraniaceae. In: Engler, A., Harms, H. (eds) eudicots (Soltis et al. 2003). This is not seen in Die natürlichen Pflanzenfamilien, ed. 2, 19a. Leipzig: character reconstructions if only a single exemplar W. Engelmann. per family is included in a tree, as in the recon- Kubitzki, K. (ed.) 2004. Flowering plants. Dicotyledons. struction of perianth merosity in Fig. 3 of Soltis Celastrales, Oxalidales, Rosales, Cornales, Ericales. The Families and Genera of Flowering Plants, VI. et al. (2003). Indeed, both in basal angiosperms Berlin Heidelberg New York: Springer. and early-diverging eudicots, there is pronounced Morgan, D.R., Soltis, D.E. 1993. See general references. variation in floral merosity, and the stereotyped Price, R.A., Palmer, J.D. 1993. See general references. pentamerous floral structure with diplostemony Savolainen, V., Fay, M.F. et al. 2000. See general references. or haplostemony occurs only above the node to Soltis, D.E. et al. 2000. See general references. Sosa, V., Chase, M.W. 2003. See general references. Gunnerales. Thus, in Ranunculaceae, pentamery, Takhtajan, A.L. 1959. Die Evolution der Angiospermen. where it occurs, is restricted to the perianth and, in Stuttgart: G. Fischer. pentamerous Sabia (Sabiaceae), the sepals, petals Takhtajan, A. 1987. See general references. and stamens stand on the same radius, a clear Thorne, R.F. 2001. See general references. Weigend, M. 2005. Notes on the floral morphology in Vivia- violation of Hofmeister’s rule but quite common niaceae (Geraniales). Pl. Syst. Evol. 253:125–131. in “basal” eudicots, as in Gunnera itself (see also Doust and Stevens 2005). Yet, following the node to Gunnerales, the typical pentamerous eudicot pattern is strictly conserved, and further variation Introduction to Gunnerales is limited to processes such as fusion, reduction and multiplication of stamens and/or carpels and 1. Poikilohydric shrubs; nodes trilacunar; axial of perianth parts. Gunnerales, although forming parenchyma 0; rays uniseriate; leaves opposite; flowers unisexual, hypogynous; perianth of up to 4 scales; part of the core eudicot clade, have not achieved stamens 3–8; ovary 3–4-locular; stylodia 3–4, broad, (or have lost?) the pentamerous pattern, but agree recurved, with ventrally decurrent stigma; pollen in with many core eudicots in possessing potent tetrads; embryo sac Allium type (bisporic, 8-celled); allelochemicals based on ellagic acid. fruit a septicidal capsule; sieve-element plastids S type. 1/2, E and South Africa, Madagascar Myrothamnaceae (see Vol. II) – Perennial herbs, often giant and nearly acaulescent, References with endosymbiontic Nostoc cells; nodes multilacunar; vascular system nearly always polystelic; leaves alter- nate; flowers epigynous, 2-merous; ovary 1-locular; Doust, A.W., Stevens, P.F. 2005. A reinterpretation of the stylodia subulate; pollen in monads; embryo sac Peper- staminate flowers of Haptanthus. Syst. Bot. 30:779–785. omia type (tetrasporic, 16-celled); stylodia 2; fruit Hilu, K.W. et al. 2003. See general references. a drupe; ellagitannins present; sieve-element plastids Kubitzki, K. 1987. Origin and significance of trimerous flow- Pcf type. 1/c. 60, mainly in southern hemisphere ers. Taxon 36:21–28. Gunneraceae Kubitzki, K., Rohwer, J.G., Bittrich, V. (eds) Flowering plants. Dicotyledons. Magnoliid, Hamamelid and Traditionally, Gunneraceae were included in Halor- Caryophyllid families. The Families and Genera of agaceae; Takhtajan (1997) placed them in Saxifra- Vascular Plants, II. Berlin Heidelberg New York: Springer. ganae. Numerous molecular studies recovered the Soltis, P.S., Soltis, D.E. 2004. The origin and diversification family in close association with the desert shrub of angiosperms. Amer. J. Bot. 91:1614–1626. Myrothamnus atthebaseofcoreeudicots,where Soltis, D.E. et al. 2000. See general references. they form a strongly supported clade. There is also Soltis, D.E. et al. 2003. See general references. evidence for the position of this clade as sister to all Takhtajan, A. 1997. See general references. remaining core eudicots (cf. Fig. 1), a grouping in- dicated in various analyses and strongly supported particularly by Soltis et al. (2000, 2003) and Hilu Introduction to Myrtales et al. (2003). Gunnera very probably has dimerous flowers 1. Ovary unilocular with apical placentation, inferior (half-inferior in Strephonema); indumentum almost and the same may apply to Myrothamnus (see always of slender, unicellular, thick-walled, pointed Conspectus). Dimery is not only widespread in hairs with a distinctive basal compartment; inflo- early-diverging eudicots such as Proteaceae and rescences indeterminate; [stamen whorls 1 or 2, the 8K.Kubitzki

outersometimeswith2or3timesthenormalnumber and mangrove trees with pneumatophores. x = 12. of stamens; pseudocolpi + (0 in Strephonema); 1/5, coastal Africa to Pacific islands intrastaminal disk often +; fruit 1-seeded, a drupe, Lythraceae p.p. (Sonneratia) usually flattened, ridged and/or winged; cotyledons 8. Annual aquatic with floating leaves and submerged twisted (massive and hemispheric in Strephonema)]. filiform-dissected stipules; flowers emergent, 4- 14/500, pantropical Combretaceae merous; sepals basally connate into a tube, 2 or 4 – Ovary usually multilocular with axile placentation, of them accrescent in fruit as hornlike or spine-like more or less inferior; fruit 1–many-seeded, usually projections; stamens 4, alternipetalous; gynoecium acapsuleorberry;indumentumvariousbutnotas 2-carpellate, ovary partly inferior; fruits 1-seeded; above 2 endosperm 0 but one cotyledon starchy, very large, 2. Leaves with secretory cavities usually containing essen- retained within the fruit. 1/3 (or 15?), temperate to tial oils (0 in Psiloxylon), spirally arranged or opposite; tropical regions of Old World, except Australia pollen oblate, brevi- to longicolpate, lacking pseudo- Lythraceae p.p. (Trapa) colpi; style base often sunken in apex of gynoecium 3 – Terrestrial 9 – Leaves lacking secretory cavities, not aromatic; pollen 9. Endothelium present (at least, in Axinandra); often with pseudocolpi; style base generally not [glabrous trees, often with quadrangular twigs; sunken (sunken in some Vochysiaceae) 4 inflorescences indeterminate; flowers hypogynous to 3. Dioecious; leaves spiral; stamens ≥ 10, erect in perigynous, 4–5-merous, obhaplostemonous or rarely bud; anthers tetralocular at anthesis, dehiscing by diplostemonous; anther endothecium ephemeral; slits; embryo sac bisporic. 2/4, south-eastern Africa, pollen tricolporate-pseudocolpate or (Crypteronia) Mascarenes Myrtaceae–Psiloxyloideae∗1 bisyncolporate; petals 0 or (Axinandra)smalland – Flowers bisexual, rarely andromonoecious; leaves connate apically, falling off as a cup when the flower spiral or opposite; stamens (few–)usually many, in- opens; ovary ± inferior, 2–6-locular; capsule woody flexed in bud; anthers bilocular at anthesis, dehiscing or chartaceous]. 3/10, South and Southeast Asia, with longitudinal or apical slits or pores; embryo sac Malaysia Crypteroniaceae monosporic; [ovary (nearly superior–)inferior, (1)2– – Endothelium 0 10 4(–18)-locular; fruits capsular, indehiscent, or fleshy]. 10. Flowers strictly obhaplostemonous; [plants woody; c. 140/over 5,500, tropical to warm-temperate regions anther endothecium ephemeral] 11 mainly of southern hemisphere, with greatest diversity – Flowers usually diplostemonous or multistami- in the Australasian region Myrtaceae–Myrtoideae∗ nate; Melastomataceae and Lythraceae rarely 4. Flowers strongly mono- or asymmetric; fertile stamen (ob)haplostemonous 14 1, staminodes 0–several; pollen without pseudocolpi; 11. Hypanthium rim ending with some blunt teeth [petals (0)1–3(5); style base ± sunken in apex of (“epicalyx”); ovary inferior, 3–5-locular; pollen gynoecium; fruit a loculicidal capsule or samaroid]. heteropolar with unequal colpi and “half pseudocolpi” 8/200, most neotropical, 2/5 of them in West Africa restricted to one polar face; sepals conspicuous, white Vochysiaceae or pinkish, inserted on margin of hypanthium; petals – Flowers usually not strongly zygomorphic; stamens scale-like, minute, closing the hypanthium in bud; > than 1; pollen often with peudocolpi 5 [stamens inserted on inner rim of tube below petals; 5. Pollen with viscin threads on proximal surface and fruit drupaceous]. x = 10. 1/c. 8, southern and eastern unique paracrystalline beaded exine, the central Africa, from Ethiopia to South Africa Oliniaceae body of the grain circular to triangular with (2)3(–6) – Epicalyx 0; ovary superior; pollen isopolar; petals, if protruding apertures, pseudocolpi 0; embryo sac present, not closing the hypanthium 12 monosporic, 4-nucleate; endosperm diploid; [flowers 12. Flowers (5)6(7)-merous; hypanthium stellate; petals (2)4(5, 7)-merous; stamens usually arising from rim of minute, hood-like, lobate and unguiculate, arising the well-developed hypanthium; fruit a capsule, or dry from hypanthial rim; septum between the two and indehiscent; interxylary phloem often present; microsporangia of each theca persistent; pollen vegetative parts rich in oxalate raphides; exotegmen 3-colporate with pseudocolpi; gynoecium 2(3)- fibrous]. 17–24/675, widely distributed from tropics to carpellate, ovary 1-locular/partly bilocular; embryo arctic-alpine regions Onagraceae∗ sac monosporic/8-nuclear; fruit capsular. n = 10. 1/1 – Viscin threads 0, exine different; pollen apertures Rhynchocalycaceae usually not protruding but pseudocolpi often present; – Flowers 4–5(6)-merous; hypanthium tubular; petals embryo sac 8-nucleate, endosperm triploid 6 strongly reduced or 0; septum between the two 6. Pollen grains 3-porate, lacking pseudocolpi; [ovary su- microsporangia not persistent; embryo sac bisporic or perior to partly inferior; branched foliar sclereids +] 7 tetrasporic 13 – Pollen grains (2)3-colporate; pseudocolpi present or 13. Flowers 4-merous; nodes unilacunar; foliar sclereids not 8 0; hypanthium large, often conspicuously coloured; 7. Gynoecium 4–8-carpellate and -locular; petals shortly ovary 4-locular; pollen with pseudocolpi isomerous clawed; stamens 12 or numerous; tall trees with droop- with apertures; embryo sac tetrasporic, 16-nucleate. ing, 4-angled ultimate branches. x = 10. 1/2, Southeast n = 10. 7/23, Cape Province of South Africa Asia, New Guinea Lythraceae p.p. (Duabanga) Penaeaceae – Gynoecium 10–20-carpellate and -locular; petals – Flowers 5(6)-merous; nodes trilacunar; branched foliar linear(-lanceolate) or 0; stamens numerous; swamp sclereids +; pollen without pseudocolpi; hypanthium green to yellow, 4–6 mm long; ovary 2-locular; embryo sac bisporic. 1/1, Costa Rica to Bolivia Alzateaceae 1 The asterisk denotes taxa not treated in this volume. Introduction to the Groups Treated in this Volume 9

14. Stamen connectives dorsally enlarged and often and Johnson (1979) and Dahlgren and Thorne massive; anther dehiscence often ± poricidal; [pollen (1984), heterogeneous elements were subsequently with pseudocolpi; leaves opposite; crystal druses excluded. Since then, the circumscription of the and/or styloids +] 15 – Stamen connectives dorsally not enlarged; anther order has remained unchanged, the only exception dehiscence by longitudinal slits 16 being the reinsertion of Vochysiaceae which, 15. Leaf venation pinnate or brochidodromous (very based on their unique floral morphology, had not rarely, acrodromous); plants woody; flowers strictly been included by most authors. Rather, they had epigynous, diplostemonous; stamen connectives generallyprovidedwithdepressedellipticterpenoid- associated it with families such as Polygalaceae and producing dorsal glands; anthers with fibrous Euphroniaceae. However, phylogenetic analyses of endothecium, dehiscing by slits (sometimes short molecular data as well as morphological evidence and functioning as pores); terminal leaf sclereids strongly supports the inclusion of Vochysiaceae +; stomata paracytic; secondary xylem with axially into the order (Conti et al. 1996). included phloem islands; fruit baccate; seed coat with fibrous exotegmen; seeds 1–few, generally with well- In the present circumscription, the order developed storage cotyledons. x =? 6/440, pantropical comprises12families(seeFig.2foraphylogenetic Memecylaceae∗ hypothesis) and more than 9,000 species, repre- – Leaf venation acrodromous (very rarely, pinnate); senting about 6% of core eudicot diversity. Myrtales plants woody or herbaceous; flowers actino- to zygo- morphic, wholly or partly epigynous, diplostemonous or (ob)haplostemonous; stamen connectives without dorsal glands; anthers mostly poricidal, endothecium 0; terminal leaf sclereids 0; stomata anemocytic, polycytic or encyclocytic; secondary xylem generally without included phloem islands; fruit capsular or baccate; seed coat without fibrous exotegmen; seeds many, with small cotyledons; [indumentum very diverse, trichomes multicellular, scale-like]. x = 17. 185/4,500, mainly in tropical and subtropical regions oftheworld,withgreatestdiversityinSouthAmerica Melastomataceae∗ 16. Herbaceous or woody; ovary superior (to infe- rior); flowers usually diplostemonous or flowers (ob)haplostemonous; petals crumpled in bud; stamens inserted at the base of floral tube or above, whorls of unequal length; heterostyly widespread; pollen grains tricolporate, pseudocolpi 0 or isomerous with or double the number of apertures; fruit capsular or baccate. 30/c. 600, worldwide, mainly in subtropical and tropical regions Lythraceae p.m.p. – Woody; ovary inferior; stamens many, covering the inner floral tube surface from the rim to the ovary; homostylous; pollen tricolporate, with indistinct pseudocolpi; ovary 7–9(–15)-loculate, carpels in 1 whorl or in 2–3 superposed layers; fruit a leathery berry; seeds many, with translucent sarcotesta. 1/2, from Balkan Peninsula to Himalayas and on Soqotra Lythraceae p.p. (Punica) Earlier hypotheses on the composition of the order Myrtales are partly congruent with modern concepts. A.P. de Candolle, for instance, in 1828 in the third volume of his Prodromus, combined all major myrtalean families, such as Combretaceae, Onagraceae, Memecylaceae, Melastomataceae, Myrtaceae and, surprisingly, also Vochysiaceae into Myrtales. In addition, he included Alan- giaceae, Rhizophoraceae and Lecythidaceae. This and similar circumscriptions of Myrtales persisted Fig. 2. A phylogenetic hypothesis of relationships of Myr- in most classifications up to the second half of the tales families, mainly based on Clausing and Renner (2001), 20thcentury.MainlyduetotheworkofBriggs Sytsma et al. (2004) and Wilson et al. (2005) 10 K. Kubitzki are characterised by the combination of vestured Modern phylogenetic studies in Myrtales pits and bicollateral vascular bundles in the pri- started with the seminal work of Johnson and mary xylem, resulting in the appearance of phloem Briggs (1984), and many of their findings have included within the secondary xylem (van Vliet later been confirmed by molecular studies (Conti and Baas 1984), and also by several embryological et al. 1996, 1997, 2002; Clausing and Renner 2001; features (Tobe and Raven 1983). Additional char- Sytsma et al. 2004; Wilson et al. 2005). Molecular acters found in part throughout the order include analyses generally provided strong support for the opposite leaves with undivided laminas, even in the monophyly of individual families, and also inter- aquatic members; small or rudimentary stipules; familial relationships have been greatly clarified. short to elongate hypanthia; stamens incurved The morphological circumscription of families in bud; vessel elements with simple perforations, and larger clades turned out to be more difficult paratracheal axial parenchyma and usually non- (see Conspectus of families). The exact position septate fibres; secondary phloem stratified in of the order within the eudicots is not clear and, youngtwigs;unilacunarnodes;simplestyles; together with Crossosomatales and Geraniales, pollen with subsidiary “colpi” (“pseudocolpi”, i.e. Myrtales are left unplaced within the rosids (APG meridional invaginations in the intercolpial re- II 2003; see also the angiosperm-wide analysis of gions, apparently with a harmomegathic function); matK sequences by Hilu et al. 2003). 2-celled pollen; a crystalliferous endotesta; scarce Combretaceae often appear to be sister to the or no endosperm; and copious amounts of gallo- rest of the order but statistical support for this is and ellagitannins, the latter often methylated. Mor- still tenuous. Onagraceae and Lythraceae are sister phological studies significantly contributed to our taxa, the former possessing raphides, the latter knowledge of Myrtales in the 1970s and 1980s, and alkaloids. Onagraceae are highly autapomorphic many of these appear in the Myrtales symposium (see Conspectus). Among the remaining families, volume published in the Annals of the Missouri Melastomataceae, Memecylaceae and their sister Botanic Garden (vol. 71, 1984). Following the group, the CAROP families (Crypteroniaceae, comparative analysis of inflorescence structure by Alzateaceae, Rhynchocalycaceae, Oliniaceae and Briggs and Johnson (1979), Weberling (1988) ana- Peneaeaceae), are characterised by dorsally lysed the inflorescences from a typological point of massive anther connectives. Renner (1993) and view. Monotelic thyrsopaniculate inflorescences, Clausing and Renner (2001) determined the postulated to be basic in the order, predominate in acrodromous leaf venation and lack of a fibrous Myrtaceae, Melastomataceae, Oliniaceae and other anther endothecium as being synapomorphic for smaller families whereas Combretaceae and Ona- Melastomataceae, and the terpenoid-producing graceae are polytelic throughout. Nevertheless, the connective glands as synapomorphic for Memecy- aspect of character polarity of inflorescences is yet laceae. The few genera of Melastomataceae which not completely settled, and recent work shows that lack the peculiar, arching leaf venation are nested within Lythraceae alone the monotelic condition within the family and thus are clearly derived, and is derived at least four times from the polytelic theoccurrenceofanantherendotheciuminthe (Graham et al. 2005). basal melastom genus Pternandra is interpreted The so-called pseudocolpi are a peculiar char- as a plesiomorphy (Clausing and Renner 2001). acter of the pollen grains of many Myrtales, and Interestingly, Pternandra also has the interxylary the distribution and different expressions of these phloem islands (included phloem) which generally structures are problematic. The absence of pseu- occur in Memecylaceae. This trait is interpreted docolpi from part of Lythraceae is striking, as is as parallelism, and not as plesiomorphy, because theoccurrenceof“double”pseudocolpiinanother Pternandra agrees with Melastomataceae in most part(Pateletal.1984).Theircompleteabsencefrom other wood characters. Onagraceae and the Myrtaceae clade have led ear- The CAROP families share the loss of an anther lier authors (Dahlgren and Thorne 1984; see also endothecium (probably evolved independently Johnson and Briggs 1984) to postulate the origin of from Melastomataceae) with obhaplostemonous the pseudocolpi after the branching off of these two flowers (Schönenberger and Conti 2003) and the families.Inviewofrecentphylogenetichypotheses presence of stipules (Johnson and Briggs 1984). (Fig. 2), however, it seems more parsimonious to Apart from these features, the four families are consider pseudocolpi as ancestral for the order as strongly diversified. Particularly Crypteroniaceae, awhole. with their variable androecium and gynoecium Introduction to the Groups Treated in this Volume 11 structure, defy any attempt for a sound morpholog- dispersal in the Neogene, when the Atlantic ical family characterisation; the circumscription had already rifted c. 80 million years ago in the of this family follows largely molecular findings. equatorial region (Sytsma et al. 2004). The initial The melastom/CAROP clade is sister to the Myr- radiation of Melastomataceae is hypothesised taceae/Vochysiaceae clade. Myrtaceae represent to have occurred during the Palaeocene/Eocene the largest, most diverse family of the order, for along the northern shore of the Sea of Tethys which a detailed classification has recently been (Renner et al. 2001). From there, the family may established (Wilson et al. 2005). The inclusion have dispersed to North America and throughout of the somewhat aberrant genera Psiloxylon and Eurasia, later also to South America and from Heteropyxis in Myrtaceae increases the support for there with repeated long-distance dispersal events the monophyly of the family, compared to a sepa- to Africa, Madagascar, India and Indochina. rate treatment of these two taxa as monogeneric families. The great size of Myrtaceae encompasses much variation in inflorescence, floral and fruit References structure, which has been explored by numerous studies subsequent to the seminal contributions APG II 2003. See general references. by Briggs and Johnson (1979) and Johnson and Briggs, B.G., Johnson, L.A.S. 1979. Evolution in the Myr- Briggs (1984). Until the advent of molecular taceae – evidence from inflorescence structure. Proc. systematics, the close relationship of Vochysiaceae Linn. Soc. New South Wales 102:157–256. Candolle, A.P. de 1828. Prodromus systematis naturalis and Myrtaceae had been camouflaged by the regni vegetabilis. Pars III. Paris: Treuttel & Würtz. distinct floral organisation of the former family Clausing, G., Renner, S.S. 2001. Molecular phylogenetics of but Vochysiaceae have many myrtalean traits, Melastomataceae and Memcylaceae: implications for and share the plesiomorphic (see the hypothetic character evolution. Amer. J. Bot. 88:486–498. Conti, E. et al. 1996. See general references. “Protomyrtalis” of Johnson and Briggs 1984) Conti, E. et al. 1997. See general references. sunken styles and 1–2-celled hairs with many Conti, E. et al. 2002. See general references. Myrtaceae (Stevens 2005). Dahlgren, R., Thorne, R.F. 1984. The order Myrtales: cir- Much work has been dedicated to the eluci- cumscription, variation, and relationships. Ann. Mis- dation of the biogeographic history of Myrtales. souri Bot. Gard. 71:633–699. Graham, S.A. et al. 2005. See general references. This task is complicated due to the ancient Hilu, K.W. et al. 2003. See general references. origin of the order and its poor fossil record (see Johnson, L.A.S., Briggs, B.G. 1984. Myrtales and Myrtaceae Sytsma et al. 2004). Recent dating analyses have – a phylogenetic analysis. Ann. Missouri Bot. Gard. estimated the crown group of Myrtales to be 107 71:700–756. Patel, V.C., Skvarla, J.J., Raven, P.H. 1984. Pollen charac- (Wikström et al. 2001) or 110 million years old ters in relation to the delimitation of Myrtales. Ann. (Sytsma et al. 2004), corresponding to the Albian Missouri Bot. Gard. 71:858–969. of the Lower Cretaceous. The split between the Renner, S.S. 1993. Phylogeny and classification of the Melas- Myrtaceae/Vochysiaceae clade and the Melas- tomataceae and Memecylaceae. Nordic J. Bot. 13:519– tom/CAROP clade may also have occurred in the 540. Renner, S.S., Clausing, G., Meyer, K. 2001. Historical bio- Albian. This implies that Gondwanan vicariance geography of Melatomataceae: the roles of Tertiary was an important factor in the biogeographic migration and long-distance dispersal. Amer. J. Bot. history of this lineage, as seems reflected by, 88:1290–1300. for instance, the disjunct distribution of the Rutschmann, F., Eriksson, T., Schönenberger, J., Conti, E. 2004. Did Crypteroniaceae disperse out of India? CAROP families and the out-of-India dispersal Molecular dating evidence from rbcL, ndhF, and rpl16 of Crypteroniaceae (Rutschmann et al. 2004). intron sequences. Intl J. Pl. Sci. 165, suppl. 4:S69–S83. As for Myrtaceae, Sytsma et al. (2004) could not Schmid, R. 1980. Comparative anatomy and morphology unambiguously determine the place of their initial of Psiloxylon and Heteropyxis, and the subfamilial and diversification, although the extant members of tribal classification of Myrtaceae. Taxon 29:559–595. Schönenberger, J., Conti, E. 2003. Molecular phylogeny and the family are clearly Australasian in origin and floral evolution of Penaeaceae, Oliniaceae, Rhynchoca- amorerecentmovetoSouthAmericaoccurred lycaceae, and Alzateaceae (Myrtales). Amer. J. Bot. in the early Eocene, possibly using the temperate 90:293–309. Antarctic land bridge. Vochysiaceae are clearly Stevens, P.F. 2005. See general references. Sytsma, K.J. et al. 2004. See general references. neotropical; the African representatives of the Tobe, H., Raven, P.H. 1983. An embryological analysis of family are nested within a South American clade the Myrtales: its definition and characteristics. Ann. and may have reached Africa by long-distance Missouri Bot. Gard. 70:71–94. 12 K. Kubitzki

Vliet, G.J.C.M. van, Baas, P. 1984. Wood anatomy and clas- members of Salicaceae and Achariaceae (de Wilde sification of the Myrtales. Ann. Missouri Bot. Gard. 1971b, at that time included in Flacourtiaceae) 71:783–800. where, however, no cyanogens are present; still, Weberling, F. 1988. The architecture of inflorescences in the Myrtales. Ann. Missouri Bot. Gard. 75:226–310. acoronaisknownfromAbatia, which led earlier Wikström, N. et al. 2001. See general references. researchers to include the genus in Passifloraceae. Wilson, P.G., O’Brien, M.M., Heslewood, M.M., Quinn, C.J. Its placement in Salicaceae on the basis of molec- 2005. Relationships within Myrtaceae sensu lato based ular evidence (Chase et al. 2002) is corroborated on matK phylogeny. Pl. Syst. Evol. 251:3–19. by morphological evidence. Thus, in contrast to Passifloraceae, Abatia has opposite leaves, valvate calyx aestivation and extrorse anthers, and both Introduction to the Passifloraceae the coronal threads and stamens are irregularly Alliance (“Passiflorales” = Malpighiales) arranged whereas in Passifloraceae the corona elements are in distinct whorls and the stamens in 1. (Andro)gynophore 0; petal aestivation contorted; the polystemononous genera of this family are in corona rarely present and then weakly developed; a single whorl (Bernhard 1999); the corona may calyx and corolla separating from developing fruit and not be homologous in both groups. falling together; seeds arillate, pitted. 10/+200, Africa, America Turneraceae – (Andro)gynophore usually present; petal aestivation cochlear; corona often present and strikingly coloured References 2 2. Stylodia inserted beneath the top of ovary; stamens 5; pollen grains 3-colporate; seeds exarillate; calyx per- For references, see the Selected Bibliography of Passiflo- sistent in fruit; tendrils 0. 1/24, Chile, Peru raceae and the General References (this volume). Malesherbiaceae – Stylodiainsertedontopofovary;stamens4,5,ormany; pollen grains 3–12-colporate or -foraminate; seeds ar- illate; tendrils often present. 17/700–750, pantropical Introduction to Proteales Passifloraceae 1. Herbaceous, cambial activity 0; sepals 2, stamens nu- These three families are closely related and also merous, spirally arranged; carpel closure by secretion; couldbemergedintoone,assuggestedasan ovule 1 per carpel, anatropous; pollen grains vari- option by the Angiosperm Phylogeny Group ouslyfurrowedorrarelysulcateortricolpate;sim- (APG II 2003); here, they are treated separately ple benzylisoquinoline alkaloids (aporphines) present; because their authors prefer the traditional family myricetin and condensed tannins 0; carpel closure by secretion; flowers thermogenic]. 1/1 or 2, North Amer- circumscription. Whereas the molecular data of ica, Asia, Australia Nelumbonaceae (see Vol. II) Chase et al. (2002) confirm that these families form – Woody; sepals or tepals > 2, stamens whorled; a clade, at the time of writing of these accounts carpels postgenitally fused; ovules 1 or 2, rarely and this introduction (Sept. 2005), it still remains more per carpel, usually orthotropous; pollen grains triaperturate(-derived); benzylisoquinoline alkaloids unclear whether the separation of these families 0; myricetin and condensed tannins present but gallate involves paraphyly. They share important char- 0; [seeds lacking starch but containing fat oil and acters such as an extrastaminal corona, exotestal protein] 2 seeds, cyclopentenoid cyanogenic glucosides 2. Stipules 2, often fused; flowers small, unisexual, in and/or cyclopentenyl fatty acids, and biparental globular heads; perianth 3–4(–7)-merous, the petals vestigial; carpels 3–8, distinct, ovules 1(2); pollen or paternal transmission of plastids (the latter grains tricolpate; triterpenes +. 1/c. 7, North America not observed in Malesherbiaceae). Whereas seed and Asia Minor to East Asia Platanaceae (see Vol. II) structure and chemical make-up appear quite – Stipules 0; flowers usually bisexual, in racemes, pan- constant in the group, a corona is not always icles, or condensed, often paired; perianth of 4 (very rarely 3 or 5) valvate tepals; stamens antetepalous, present and its expression is quite diverse. It is often adnate to tepals, alternating with hypogynous developed in its full-fledged form in Passifloraceae, nectar secreting glands; carpel 1; ovules 1–2(–many); mainly Passiflora, but is only weakly in the pos- pollen grains triporate, rarely tricolporoidate or sibly basal Adenias and also in Turneraceae and diporate; triterpenes 0. 80/1,700, mainly southern Malesherbiaceae; it is difficult to decide whether hemisphere, best developed in Australia Proteaceae this represents an anagenetic or reductional The three families united in this order form an un- transformation. Accessory or superposed buds, expected alliance, which was discovered by early as in Passifloraceae, are found also in various molecular work and since has received support in Introduction to the Groups Treated in this Volume 13 various multigene analyses. Sabiaceae, here left un- which is inconsistent with their interpretation as placed as to ordinal allocation (see family treat- erstwhile petals. Both extant and fossil Platanaceae ment), are also often found together with Proteales show considerable variability in the number (cf. Fig. 1). As is evident from the characters given of floral parts. Some of their mid-Cretaceous in the Conspectus, Nelumbonaceae appear quite and early Tertiary representatives had strictly out of place in this alliance. Their ranunculalean pentamerous perianths with five stamens and five chemistry (see Gottlieb et al. 1993) is accompanied carpels respectively (Friis and Crane 1989), but by completely ascidiate carpels without any post- clear tetramery existed, for instance, in the Late genital fusion, which Nelumbo shares only with Cretaceous Quadriplatanus (Magallón et al. 1997), Berberidaceae; the ovules are anatropous (Igers- in which the female flowers had two perianth heim and Endress 1998). Nelumbo has a dimerous whorls and eight carpels. calyx (Hayes et al. 2000; in contrast to information Proteaceae pollen is known for differing from given erroneously by Kubitzki 1993), but dimer- the widespread developmental pattern in eudicots, ous whorls are widespread in basal eudicots (Drin- in which apertures are formed in pairs at six points nan et al. 1994; Doyle and Endress 2000; Soltis et in the developing tetrad, following Fischer’s Rule. al. 2003) and are by no means exclusive for early- In Proteaceae, the apertures are formed in groups diverging eudicots. of three at four points in the tetrad (Garside 1946). Nelumbonaceae are remarkable with regard Furness and Rudall (2004), who quoted Illiciales, to pollen development. Whereas in basal an- Proteaceae and Olacaceae as the only examples giosperms (monosulcates) there is much variation in angiosperms for pore orientation according to between the simultaneous and successive type of Garside’s Rule, argued for origins of this develop- microsporogenesis, almost all eudicots (triapertu- mental mode of triaperturate pollen independent rates) have simultaneous microsporogenesis, with from the developmental pattern following Fis- the notable exception of Nelumbonaceae (Kreunen cher’s Rule which characterises the majority of the and Osborn 1999) and Proteaceae (Furness et al. eudicots. Illiciales are, however, irrelevant in this 2002; including the diporate proteaceous pollen: context because their “tricolpate” condition is an Blackmore and Barnes 1995). The co-occurrence extension of the trichotomosulcate arrangement, of putatively monosulcate and triaperturate pollen which is often found among monosulcates in which in Nelumbo (Kuprianova 1979; Blackmore et al. Illiciales are embedded (Huynh 1976; Doyle et al. 1995) raised great phylogenetic interest but later 1990). The pollen grains of some Olacaceae, which (Borsch and Wilde 2000) was found to be part of are formed according to Garside’s Rule, appear an extensive, regular variation pattern influenced autapomorphic because this family is deeply by factors such as a delay in aperture ontogeny embedded within the triaperturate group. Also for (Kreunen and Osborn 1999). Proteaceae, it is difficult to envisage a completely Platanaceae and Proteaceae, which are re- independent origin of the triaperturate condition solved as sister taxa in most molecular analyses, from a monosulcate/trichotomosulcate ancestor: have much in common morphologically, including even if Proteales were basal in eudicots, Proteaceae the presence of five carpellary bundles (rather are nested within Proteales, in which Platanaceae than three in most Ranunculales and in the and Nelumbonaceae produce “normal” triaper- early-branching Proteacea Bellendena), ample turates. Any hypothesis of an independent origin tanniniferous tissue in the carpels, mostly one within this order would then require one or two or two large, orthotropous ovules of which the additional origins of the normal Fischer’s Rule upper is pendent, and floral organs which may tricolpate type, a non-parsimonious assumption. be arranged in dimerous whorls. Traditionally, Rather, these considerations all favour an origin Proteaceae had been interpreted as tetramerous of the proteaceous condition from normal tricol- but the ontogenetic work of Douglas and Tucker pates, much as concluded for Olacaceae. This view (1996a) supports the interpretation of their is shared by Blackmore and Crane (1998) who perianth and androecium as dimerous. Possibly, tend to view the Garside’s Rule arrangement in Proteaceae are primarily apetalous; the nectarial Proteaceae as derived. hypogynous scales, which alternate with the A triporate fossil pollen from the Cenoma- tepals, are positioned inside the stamen whorl nian (mid-Cretaceous) of the Northern Gondwana and their initiation takes places after that of all Province, Triorites africaensis,hasoftenbeenre- other floral organs (Douglas and Tucker 1996), lated to Proteaceae. The ultrastructural analysis of 14 K. Kubitzki

Triorites by Ward and Doyle (1994) suggests that Cladistic analysis and implications. Amer. J. Bot. Triorites pollen is not tricolporate-derived, as usu- 77:1558–1568. ally is the case with triporates, but perhaps directly Drinnan, A.N., Crane, P.R., Hoot, S.B. 1994. Patterns of floral evolution in the early diversification of non-magnoliid from tricolpate. Ward and Doyle (1994) consid- dicotyledons (eudicots). Pl. Syst. Evol. suppl. 8:93–122. ered this as an additional piece of evidence against Endress, P.K., Igersheim, A. 1999. Gynoecium diversity and aderivationofthefamilyfromarosidancestor systematics of the basal eudicots. Bot. J. Linn. Soc. – of course, amply confirmed by molecular data. 130:305–393. A number of Late Cretaceous (Late Santonian-Early Friis, E.M., Crane, P.R. 1989. Reproductive structures of Cre- taceous Hamamelidae. In: Crane, P.R., Blackmore, S. Campanian) follicular fruits from southern Sweden (eds) Evolution, systematics and fossil history of the (Leng et al. 2005) exhibit several similarities with Hamamelidae, 1. Oxford: Clarendon Press, pp. 155– Proteaceae, particularly with the first branching 174. lineages in the family. These include a plicate carpel Furness, C.A., Rudall, P.J. 2004. Pollen aperture evolution – a crucial factor for eudicot success? Trends Pl. Sci. structure with a vascular system of three bundles, 9:154–158. several anatropous, probably bitegmic ovules, and Furness, C.A., Rudall, P.J., Sampson, F.B. 2002. Evolution a more or less sessile stigmatic area which is lo- of microsporogenesis in angiosperms. Intl J. Pl. Sci. cated at the distal-most part of the ventral slit and 163:235–260. extends over the topological apex to the abaxial side Garside, S. 1946. The developmental morphology of the pollen of Proteaceae. J. S. African Bot. 12:27–34. of the follicle. Although these fossils differ from ex- Gottlieb, O.R., Kaplan, M.A.C., Zocher, D.H.T. 1993. tant Proteaceae in having unisexual and obviously A chemosystematic overview of Magnoliidae, Ra- perianth-free flowers and several ovules, they rep- nunculidae, Caryophyllidae and Hamamelidae. In: resent an extinct lineage of basal eudicots which Kubitzki, K. (ed.) The Families and Genera of Vascular probably was close to modern Proteaceae. Plants, 2. Berlin Heidelberg New York: Springer, pp. 20–31. Johnson and Briggs (1975), with admirable in- Hayes, V., Schneider, E.L., Carlquist, S. 2000. Floral devel- tuition, anticipated that Proteaceae are not only opment of Nelumbo nucifera (Nelumbonaceae). Intl J. an “isolated” but also a fairly basal family, rather Pl. Sci. 161, suppl. 6:S183–S191. than belonging somewhere in the rosids, this hav- Hoot, S.B., Magallón, S., Crane, P.R. 1999. Phylogeny of basal eudicots based on three molecular data sets: atpB, rbcL, ing been fully confirmed by the evidence available and 18S nuclear ribosomal DNA sequences. Ann. Mis- 30 years later. souri Bot. Gard. 86:1–32. Huynh, K.-L. 1976. L’arrangement du pollen du genre Schisandra (Schisandraceae) et sa signification phylogénique chez les Angiospermes. Beitr. Biol. References Pflanzen 52:227–253. Igersheim, A., Endress, P.K. 1998. Gynoecium diversity Blackmore, S., Barnes, S.H. 1995. Garside’s rule and the and systematics of the paleoherbs. Bot. J. Linn. Soc. microspore tetrads of Grevillea rosmarinifolia A. Cun- 127:289–370. ningham and Dryandra polycephala Bentham (Pro- Johnson, L.A.S., Briggs, B. 1975. See general references. teaceae). Rev. Palaeobot. Palynol. 85:111–121. Kreunen, S.S., Osborn, J.M. 1999. Pollen and anther devel- Blackmore, S., Crane, P.R. 1998. The evolution of apertures opment in Nelumbo (Nelumbonaceae). Amer. J. Bot. in the spores and pollen grains of embryophytes. In: 86:1662–1676. Owens, S.J., Rudall, P.J. (eds) Reproductive biology. Kubitzki, K. 1993. Platanaceae. In: Kubitzki, K., Rohwer, J.G., Royal Botanic Gardens, Kew, pp. 159–182. Bittrich, V. (eds) Flowering plants. Dicotyledons. Blackmore, S., Stafford, P., Persson, V.1995. Palynology and Magnoliid, Hamamelid and Caryophyllid families. systematics of Ranunculiflorae. Pl. Syst. Evol. suppl. The Families and Genera of Vascular Plants, II. Berlin 9:71–82. Heidelberg New York: Springer, pp. 521–522. Borsch, T., Wilde, V. 2000. Pollen variability within spe- Kubitzki, K., Rohwer, J.G., Bittrich, V. (eds) Flowering cies, populations, and individuals, with particular ref- plants. Dicotyledons. Magnoliid, Hamamelid and erence to Nelumbo. In: Harley, M.M., Morton, C.M., Caryophyllid families. The Families and Genera of Blackmore, S. (eds) Pollen and spores: morphology Vascular Plants, II. Berlin Heidelberg New York: and biology. Royal Botanic Gardens, Kew, pp. 285–299. Springer. Douglas, A.W., Tucker, S.C. 1996. Comparative floral on- Kuprianova, L.A. 1979. On the possibility of the develop- togenies among Persoonioideae including Bellendena ment of tricolpate pollen from monosulcate. Grana (Proteaceae). Amer. J. Bot. 83:1528–1555. 18:1–4. Doyle, J.A., Endress, P.K. 2000. Morphological phylogenetic Leng,Q.,Schönenberger,J.,Friis,E.M.2005.Late Cretaceous analysis of basal angiosperms: comparisons and com- follicular fruits from southern Sweden with systematic bination with molecular data. Intl J. Pl. Sci. 161, suppl. affinities to early diverging dicots. Bot. J. Linn. Soc. 6:S121–S153. 148:377–407. Doyle, J.A., Hotton, C.L., Ward, J.V. 1990. Early Cretaceous Magallón, S., Herendeen, P.S., Crane, P.R. 1997. Quadripla- tetrads, zonasulculate pollen, and Winteraceae. II. tanus georgianus gen. et sp. nov.: staminate and pis- Introduction to the Groups Treated in this Volume 15

tillateplatanaceousflowersfromtheLateCretaceous (Coniacian-Santonian) of Georgia, U.S.A. Intl J. Pl. Sci. licular; seeds very small, winged; low glabrous shrub]. 158:373–394. 1/1, Tasmania Tetracarpaeaceae Ressayre, A., Dreyer, L., Triki-Teurtroy, S., Forchioni, A., – Ovary inferior or semi-inferior, the carpels at least Nadot, S. 2005. Post-meiotic cytokinesis and pollen basally connate; anther wall with fibrous endothecium aperture pattern ontogeny: comparison of develop- 9 ment in four species differing in aperture pattern. 9. Leaves alternate, opposite, or verticillate; ovary infe- Amer. J. Bot. 92:576–583. rior,4(–2)-carpellate;stylodiafreewithpenicillatestig- Soltis, D.E. et al. 2003. See general references. mas; vessel elements with simple perforation; pollen 4– Ward, J.V., Doyle, J.A. 1994. Ultrastructure and relation- 6(–20)-colpate or -porate, often aspidiate; [tanninifer- ships of mid-Cretaceous polyforate and triporate ous terrestrial or aquatic herbs, shrubs or small trees]. pollen from northern Gondwana. In: Kurmann, M.H., n = 7 (6, 9, 21, 29). 8/150, worldwide but mainly Aus- Doyle, J.A. (eds) Ultrastructure of fossil spores and tralia Haloragaceae pollen. Royal Botanic Gardens, Kew, pp. 161–172. – Leaves opposite; ovary semi-inferior, 4-carpellate; style shortly 4-lobed; stigmas papillate; vessel elements with scalariform perforation; pollen tricolporate; [petals small or 0; climbing shrubs]. 1/2, Australia Introduction to Saxifragales Aphanopetalaceae 10. Shrubs; [leaves alternate; vessel element perforation 1. Trees; stigmas decurrent; pollen colpate or pantopo- mainly scalariform; ovary syncarpous] 11 rate; [anthers with protruding connectives] 2 – Herbs; [seeds exotestal] (but see Crassulaceae) 13 – Trees or herbs; stigmas subulate, capitate or spatulate; 11. Gynoecium 5-carpellate; style shortly 5-lobed; [ovary pollen colpor(oid)ate, rarely porate 5 largely inferior with 4–6 ascending ovules/locule; stig- 2. Flowers mostly hermaphroditic; anthers mostly mas radiate; stipules minute; pollen 3-colporate; vessel dehiscing with valves; [trichomes mostly stellate elements also with simple perforations]. 1/3, Mexico or tufted; flowers (2–)4–5(–7)-merous, calyx rarely Pterostemonaceae 0, petals often adaxially circinate; ovary inferior to – Gynoecium 2-carpellate; style cleft or not 12 superior, 2-carpellate with straight stylodia; iridoids 12. Ovary inferior, 1-locular; fruit a berry; seeds usually 0]. n = 8, 12, 18. 27/82, tropical to temperate, C and E numerous, small, mucilaginous; embryo small; pollen NorthAmerica,SEEuropethroughS,EandSEAsiato 8-zonocolporate, pentacolpo-di-orate, or pantoporate; New Guinea and NE Australia erect, arching, trailing or prostrate shrubs often with Hamamelidaceae (see Vol. II) 3- or 2-forked or simple nodal spines and smaller in- – Dioecious; anthers dehiscing with slits or rudimentary ternodal bristles, and long-petiolate, basally 2-veined valves 3 leaves; [seed coat with exotestal mucilaginous palisade, 3. Ovary unicarpellate with abaxial suture [but the soli- endotesta crystalliferous]. n = 8. 1/150–200 tary carpels (= female flowers) united into pseudan- Grossulariaceae thia]; iridoids 0; [fruit a samara; seed with large em- – Ovary nearly superior to 3/4 inferior, 2-locular; fruit bryo]. n = 19. 1/2, China and Japan acapsule;seedsfewtomany,dry;embryolarge,curved; Cercidiphyllaceae (see Vol. II) pollen bilateral, 2-porate; unarmed shrubs or small – Ovary bicarpellate; no pronounced shoot dimorphism; trees with short-petiolate, pinnately veined leaves; [an- iridoids present 4 thers with globular protrusion of the connective; sty- 4. Female flowers in globose heads, male in terminal lodia separate to fused but apically coherent with glob- globose racemes; stipules present; pollen pantoporate; ular stigmas]. n = 11. 1/c. 27, E and SE Asia, one sp. in embryo > half the length of the seed; secretory ducts in Africa, one in North America Iteaceae all vegetative tissues. n = 8. 1/13, C America, E Mediter- 13. Fruit a 5–7-carpellate and -beaked stellate capsule, the ranean, E Asia to Malesia beak of each carpel circumscissile above the syncar- Altingiaceae (see Vol. II under Hamamelidaceae) pous region; nodes unilacunar; vessel element perfo- – Flowers in elongate racemes; stipules 0; pollen tricol- ration scalariform; [rhizomatous perennials; petals 0 pate; embryo minute; secretory ducts 0. n = 8. 1/10, E or very small]. n = 8, 9. 1/2, E North America, E and Asia, Malesia Daphniphyllaceae SE Asia Penthoraceae 5. Flowers polyandrous 6 – Fruiting carpels not dehiscing along a circumscissile – Flowers haplostemonous or diplostemonous. suture; nodes uni-, tri- or multilacunar; vessel element Core Saxifragales 7 perforation simple 14 6. Apocarpous; stamens in 5 fascicles; seeds with shin- 14. Succulent herbs, subshrubs or rarely shrubs; stipules ing sarcotesta; perennial herbs or (half)shrubs. 1/40, 0; nodes trilacunar or unilacunar; leaves usually sim- northern hemisphere Paeoniaceae ple and entire; gynoecium isomerous with perianth; – Syncarpous, ovary unilocular; stamens not distinctly nectariferous scale near base of each carpel (petaloid fasciculate; seeds with black crustaceous coat; trees. in Monanthes and some Sedum). n = 4–22+. 33/1,410, 3/11, South America, Africa Peridiscaceae widely distributed mostly in arid temperate or warm 7. Flowers essentially 4-merous; [leaves estipulate; nodes regions, and centred in Mexico and South Africa unilacunar] 8 Crassulaceae – Flowers essentially 5 >-merous 10 – Not succulent, perennial and annual herbs; stipules 8. Ovary superior, nearly apocarpous; anther wall with- present or leaf basis sheathing; nodes trilacu- out fibrous endothecium; [pollen tricolporate; fruit fol- nar or multilacunar; leaves simple or pinnately or 16 K. Kubitzki

palmately compound or decompound; gynoecium 2(– et al. 1993; Morgan and Soltis 1993; Soltis and Soltis 5)-carpellate; nectariferous disk mostly present. n = 1997; Fishbein et al. 2001; APG II 2003; Davis and (5, 6)7(11, 12, 15, 17, 18). 33/1,410, nearly cosmopolitan Chase 2004; Fishbein and Soltis 2004, among oth- but mainly in northern temperate zone and centred in North America Saxifragaceae ers). The monophyly of this clade is strongly sup- ported. Moreover, there is a 1 bp insertion com- Saxifragales, in the circumscription followed in this mon to all members of the order (Soltis and Soltis volume, are the result of a series of molecular ana- 1997). In addition to the Core Saxifragales tradi- lyses carried out over the past 15 years or so (Chase tionally considered to belong to this order (see Fig. 3), it comprises also Haloragaceae, the contro- versial Paeoniaceae, some woody “hamamelidid” families (Cercidiphyllaceae, Hamamelidaceae, Al- tingiaceae, Daphniphyllaceae), and the enigmatic Peridiscaceae.OutgrouprelationshipsofSaxifraga- les are weakly supported (Savolainen, Chase et al. 2000; Soltis et al. 2000) but the group is often found together with Vitaceae at the base of the large eu- rosid clade. A group comprising the families now con- stituting the order Saxifragales has never before been recognised in traditional systematic studies. In comparison with older concepts such as of those of Bentham (1865), Engler (1891, 1930) and Cronquist (1981), and partly also with the more recent but essentially morphologically based classifications by Huber (1991), Takhtajan (1997) and Thorne (2001), the present circumscription differs in three major points, first, in the exclusion of the lineages having tenuinucellate ovules and containing iridoids; second, in the inclusion of Haloragaceae, Peridiscaceae and Paeoniaceae (in- cluded in Saxifragales by Huber 1991); and third, in the addition of several woody families showing presumably plesiomorphic characters such as valvate anther dehiscence, apiculate connective protrusions and tricolpate pollen, some of which persist here and there in the Core Saxifragales. ad 1. The first to recognise the systematic significance of ovules was Warming (1878), and the bitegmic ovules of Itea led van Tieghem (1898) to propose the transfer of Itea from Escalloniaceae to Saxifragaceae. He later (van Tieghem 1901) used the distinction between bitegmic and unitegmic ovules as a rigorous criterion in the classification of the whole plant kingdom, which resulted in a system containing inconsistent and unnatural groupings, discrediting the use of this character. Consequently, his views were wholly rejected by Fig. 3. A phylogenetic hypothesis of relationships of Saxi- other botanists and Engler (1930), when com- fragales families, based on Fishbein et al. (2001), Fishbein menting on Hydrangioideae and Escalloniodeae and Soltis (2004) and, for Peridiscaceae, on Davis and Chase which he included in his Saxifragaceae, argued that (2004). Note that resolution among the basal woody fam- ilies is weekly supported. Hamamelidaceae, Altingiaceae the number of integuments had little systematic and Cercidiphyllaceae were treated in Vol.II (Kubitzki et al. significance because, among otherwise clearly 1993) of this series related genera of Ranunculaceae, their number Introduction to the Groups Treated in this Volume 17 can be variable. However, Mauritzon (1933), in his rate compound apertures (with well-differentiated embryological studies of Saxifragales, found ovule exo- and endoapertures) are the rule but some- characters to be useful, and Philipson (1974) sug- times (Saxifragaceae) they are not fully developed. gested that a distinction be made between families On the whole, the woody basal families of (“non- in which the ovular characters are constant, as Core”) Saxifragales appear as isolated remnants opposed to those in which some variation in this of formerly more richly developed, archaic lin- respect occurs sporadically. Taxa such as Escalloni- eages, as is particularly well documented for Cer- aceae, Hydrangeaceae, Phyllonoma, Montinia and cidiphyllaceae. Resolution within the Core Saxifra- Eremosyne, transferred to the asterids by Soltis and gales is better supported, mainly due to the efforts Soltis (1997) on the basis of molecular evidence, of Fishbein et al. (2001), and even a cursory glance are all unitegmic and iridoid-positive. Some at the topology reproduced in Fig. 3 reveals that unitegmic genera (Darmera, Micranthes) persist in a broader concept of Saxifragaceae (with the inclu- Saxifragaceae but these are embedded in broader sion of Tetracar paea and Penthorum)isuntenable, bitegmic lineages and lack iridoids, whereas the unless Crassulaceae and Haloragaceae were to be few iridoid-positive Saxifragales are bitegmic. incorporated, too. The topology of Fig. 3 is also ad 2. Haloragaceae traditionally have been re- useful for a comparison with character transfor- lated with Myrtales but Takhtajan (1997) demon- mations which can be recognised in the Core Saxi- strated that they have more characters in common fragales. The transition from woody to herbaceous with Saxifragales. Peridiscaceae have often been growth, usually accompanied by the loss of scalari- related to Flacourtiaceae but the three-gene anal- form perforation plates of the vessel elements, has ysis of Davis and Chase (2004) adds the family, taken place some five times within Saxifragales – in with the inclusion of Soyauxia, to the Saxifraga- Paeoniaceae, Saxifragaceae, Crassulaceae (the few les where they come out with Daphniphyllaceae woody members of which are definitely secondari- and the other woody groups at the base of the ly woody; see Crassulaceae, this volume), Pentho- Saxifragales clade, albeit with low support (M.W. raceae and the woody/herbaceous Haloragaceae. Chase, pers. comm. Nov. 2003). A close relation- Penthoraceae are remarkable for having “retained” ship between these three taxa is not reflected by scalariform perforation in spite of being herba- morphological traits, although the anther flaps of ceous. Grossulariaceae are strictly woody whereas Soyauxia are found in some basal Saxifragales fam- their sister group Saxifragaceae is largely herba- ilies as well. Paeoniaceae, large-flowered, apocar- ceous – a remarkable difference, although both pous, with striking seed-presentation and strongly agree in details of shoot morphology and growth autapomorphic2, in the analysis of Fishbein and dynamics,asiswelldescribedbyWeigendunder Soltis (2004) are basal to Core Saxifragales. “Affinities” of Grossulariaceae (this volume). ad 3. Within the Saxifragales, but outside the Anthers in Saxifragales are remarkably uni- Core Saxifragales, iridoids occur in two families, formly basifixed but gynoecium structure, par- Altingiaceae and Daphniphyllaceae, where they are ticularly in Core Saxifragales, is labile. Pteroste- poorly diversified chemically (Kaplan and Gottlieb monaceae stand out with an isomerous and apo- 1982), perhaps due to the small size of these fami- carpous gynoecium within an otherwise 2–3-car- lies. These are the only reports of iridoids outside pellate Saxifragaceae alliance; the gynoecia of Cras- the asterids. sulaceae and Tetracarpaeaceae are also (nearly) ThetricolpatepollenpredominatinginCer- apocarpous. Free stylodia are widespread, and cidiphyllaceae, Hamamelidaceae and Daphniphyl- most groups with this character seem to lack laceae is likely to be a plesiomorphic trait; in fact, a compitum. The functionally advantageous fusion the apertures of Cercidiphyllum appear quite ar- of stylodia into a common style is uncommon chaic, and are intermediate between the poroidate (Grossulariaceae, Iteaceae, Aphanopetalaceae). and colpoidate condition (Praglowski 1975); how- Although there is no indication that apocarpy here, ever, there are transitions to compound (colporoi- or in the (other) eurosids where it also occurs, date or colporate) apertures known from within is due to a reversal, it is difficult to imagine that Daphniphyllum. In the Core Saxifragales, elabo- this character expression should be plesiomorphic in these groups. Minute embryos characterise 2 The record of iridoids sometimes indicated for Paeonia (e.g. in Peridiscaceae, Daphniphyllaceae, Paeoniaceae, Stevens 2005) may be based on Nekratova et al. (1988, in Rast. Resur. 24:392–399), who may have mistaken monoterpene glucosides of Grossulariaceae and Tetracarpaeaceae; all other the Paeoniflorin type for iridoids (Hegnauer 1990). groups have medium-sized or large embryos. 18 K. Kubitzki

References Tieghem, Ph. van 1898. Structures de quelques ovules et parti qu’on en peut tirer pour améliorer la classifica- tion. J. Bot. (Paris) 12:197–220. APG II 2003. See general references. Tieghem, Ph. van 1901. L’œuf des plantes considéré comme Bentham, G. 1865. Ordo LIX. Saxifrageae. In: Bentham, G., base de leur classification. Ann. Sci. Nat., Bot. VIII, Hooker, J.D., Genera Plantarum, I, ii. London: Reeve, 14:213–390. pp. 629–655. Walker, J.W. 1974. Aperture evolution in the pollen Chase, M.W. et al. 1993. See general references. grains of primitive angiosperms. Amer. J. Bot. Cronquist, A. 1981. See general references. 61:1112–1136. Cutler, D.F., Gregory, M. (eds) 2000. Anatomy of the Warming, E. 1878. De l’ovule. Ann. Sci. Nat. VI, 5:177–266. Dicotyledons, 2nd edn. Vol. 4, Saxifragales. Oxford: Clarendon Press. Davis, C.C., Chase, M.W. 2004. Elatinaceae are sister to Malpighiaceae; Peridiscaceae belong to Saxifragales. Introduction to Vitales Amer. J. Bot. 91:262–273. Engler, A. 1891. Saxifragaceae. In: Engler, A., Prantl, K., Die natürlichen Pflanzenfamilien III, 2a. Leipzig: W. En- 1. Small trees, shrubs, or herbs; tendrils 0; stipular wings gelmann, pp. 41–93. conspicuous, sheathing; inflorescences terminal; Engler, A. 1930. Saxifragaceae. In: Engler, A., Prantl, K., floral disk tubular, not producing nectar; due to Die natürlichen Pflanzenfamilien, ed. 2, 18a. Leipzig, secondary septation, ovary 4–6(–8)-locular, ovule 1 W. Engelmannn, pp. 74–226. per locule. 1/34, mainly southern Asia, extending to Fishbein, M., Soltis, D.E. 2004. Further resolution of the Africa/Madagascar and Australia Leeaceae rapid radiation of Saxifragales (Angiospetrms, Eudi- – Woody lianas usually with leaf-opposite tendrils, cots) supported by mixed-model Bayesian analysis. rarely succulent small trees or erect herbs; stipules not Syst. Bot. 29:883–891. sheathing the petiole margins; inflorescences often Fishbein, M. et al. 2001. See general references. leaf-opposed; floral disk intrastaminal, ring-shaped, Hegnauer, R. 1990. See general references. cupular, or gland-shaped, usually nectariferous; ovary 2-locular, ovules 2 per locule. 14/c. 750, pantropical Huber, H. 1991. Angiospermen. Leitfaden durch die Ord- Vitaceae nungen und Familien der Bedecktsamer. Stuttgart: G. Fischer. Traditionally, Vitales were included in Rhamnales Kaplan, M.A.C., Gottlieb, O.R. 1982. Iridoids as systematic – both have antepetalous stamens – but Takhta- markers in dicotyledons. Biochem. Syst. Ecol. 10:329– 347. jan (1997) dismembered this association because Kubitzki, K., Rohwer, J.G., Bittrich, V. (eds) Flowering Vitaceae and Leeaceae differ from Rhamnaceae in plants. Dicotyledons. Magnoliid, Hamamelid and their berry-like fruits and seed structure, and in Caryophyllid families. The Families and Genera of having raphide sacs in the parenchymatous tissue; Vascular Plants, II. Berlin Heidelberg New York: he placed them as “Vitalanae” close to his Pro- Springer. Mauritzon, J. 1933. Studien über die Embryologie der Fam- teanae at the end of his Rosidae. Corner (1976) was ilien Crassulaceae und Saxifragaceae. Ph.D. Thesis, much impressed by the thick, lignified endotesta Lund University. and small embryo of the seeds of Vitaceae, which Morgan, D.R., Soltis, D.E. 1993. Phylogenetic relationships he found “scarcely improved on that of Magno- among members of Saxifragaceae sensu lato based on rbcL sequence data. Amer. J. Bot. 80:631–660. lia and. . . even more primitive”. Vitales also have Nemirovich-Danchenko, E.N. 1994. Morphology and a tracheidal exotegmen, which is a rare feature in anatomy of the seeds of Iteaceae (in Russian). Bot. angiosperm seeds – apart from Dilleniaceae; it is Zhurn. (Moscow & Leningrad) 79:83–87. listed only for Cunoniaceae by Nandi et al. (1998). Philipson, W.R. 1974. Ovular morphology and the major Vitaceae, Leeaceae and Dilleniaceae are the only classification of the dicotyledons. Bot. J. Linn. Soc. 68:89–108. angiosperm families which share the lignified en- Praglowski, J. 1975. Pollen morphology of the Trochoden- dotesta and tracheidal exotegmen. draceae, Tetracentraceae, Cercidiphyllaceae and Eu- The first to find an association between pteleaceae with reference to taxonomy. Pollen Spores Vitaceae and Dilleniaceae were Nandi et al. (1998) 16:449–467. in a combined rbcL/morphological analysis. In Savolainen, V., Chase, M.W. et al. 2000. See general refer- ences. an rbcL analysis (Savolainen, Fay et al. 2000), Soltis, D.E., Soltis, P.S. 1997. Phylogenetic relationships in Vitales and Dilleniaceae appear in a sister position Saxifragaceae sensu lato: a comparison of topologies to Caryophyllales. The two-gene analysis of based on 18S rDNA and rbcL sequences. Amer. J. Bot. Savolainen, Chase et al. (2000) and the three-gene 84:504–522. analysis of Soltis et al. (2000) place Vitales at the Soltis, D.E. et al. 2000. See general references. Stevens, P.F. 2005. See general references. base of the rosid clade. In the matK analysis of Takhtajan, A. 1997. See general references. Hilu et al. (2003), the rosids are sister to Vitis and Thorne, R.F. 2001. See general references. Tetracera (in turn, sister taxa) and to other taxa Introduction to the Groups Treated in this Volume 19 including Berberopsidales, Santalales (relative 2 times as many as petals; ovary (2–)5(–12)-locular; positions were uncertain), and Caryophyllales vessels with vestured pits; axial parenchyma usually plus asterids. In the four-gene study of Soltis with 2–4 cells per strand; storying present in axial parenchyma, sometimes in rays; crystals one per cell or et al. (2003), Vitales occupy a position at the base septate portion of cell in wood or secondary phloem; of a Caryophyllales/Saxifragales clade. None of x = 6–15. 22/230–240, in hot dry regions all over the theseassociationsisstronglysupported.Thus, world Zygophyllaceae a closer relationship between Vitaceae/Leeaceae Previously, Krameriaceae and Zygophyllaceae and Dilleniaceae can not be ruled out. Perhaps were placed in different orders, and no close both families branched off at the base of the core relationship between them had been recognised. eudicot tree and both, but Dilleniaceae more Molecular studies, particularly the multigene anal- probably than Vitaceae/Leeaceae, may be related yses of Soltis et al. (2000) and Savolainen, Chase to Caryophyllales. Evidence for placing Vitales et al. (2000), revealed a strongly supported clade in rosids is tenuous because the molecular data consisting of the two families within eurosids I. are not convincing. Note, however, that Stevens Ordinal status for this clade, which appears not to (2005), citing Oxelman et al. (2004), mentions fit in any other rosid order, was suggested by Soltis that the RPB2genemaynotbeduplicatedin et al. (2000). Zygophyllaceae and Krameriaceae Vitales, perhaps suggesting a position outside core are quite diverse but have more or less pentamer- eudicots. ous and (ob-)diplostemonous(-derived) flowers, Summarising, one may agree with Stevens bitegmic/crassinucellate ovules, and simple styles, (2005, on Vitales) that Vitales have no firm posi- and thus conform to a generalised rosid pattern. tion as yet, although a more strongly supported They agree in various wood characters such as sim- association with Dillenaceae and Caryophyllales ple perforation plates in vessels, and the presence would not come as a surprise. of tracheids (vasicentric, in the case of Zygophyl- laceae), which are considered as plesiomorphous within eurosids whereas other characters, listed References by Carlquist (2005) and partly included in the Conspectus above, are autapomorphous; the Corner, E.J.H. 1976. See general references. paedomorphic rays of Krameria are probably Hilu, K.W. 2003. See general references. related to its hemiparasitism. The presence of Nandi, O.I. et al. 1998. See general references. anthraquinones may indicate their relationship to Oxelman, B., Yoshikawa, N., McConaughy, B.L., Luo, J., Den- ton, A.L., Hill, B.D. 2004. RPB2genephylogenyinflow- the nitrogen-fixing clade (Cucurbitales, Fagales, ering plants, with particular emphasis on asterids. Mol. Fabales, Rosales) where these compounds are Phylog. Evol. 32:462–479. more often found (Savolainen, Chase et al. 2000), Savolainen, V., Chase, M.W. et al. 2000. See general refer- and to which they appear close in some analyses, ences. Savolainen, V., Fay, M.F. et al. 2000. See general references. although with low support. Soltis, D.E. et al. 2000. See general references. Apart from the presence of harman alkaloids in Soltis, D.E. et al. 2003. See general references. both families, the remarkable diversification of lig- Stevens, P.F. 2005. See general references. nans and neolignans is a strong link between them Takhtajan, A. 1997. See general references. (see “Phytochemistry” in family treatments) al- though, according to our present knowledge, these compounds in Krameriaceae are localised in the Introduction to Zygophyllales roots but in Zygophyllaceae on the leaf surface and in the wood. 1. Hemiparasitic; stipules 0; flowers solitary or in botryoid panicles, zygomorphic; the two abaxial petals lipid-secreting, the three adaxial ones forming a flag; stamens ± as many as petals; ovary 1-locular; References pollen 3-porate; vessels with non-vestured pits; axial parenchyma usually with one cell per strand; storying Carlquist, S. 2005. Wood anatomy of Krameriaceae with absent or nearly so; crystals many per cell, mostly in comparisons with Zygophyllaceae: phylesis, ecology axial phloem parenchyma; n = 6. 1/18, New World and systematics. Bot. J. Linn. Soc. 149:257–270. Krameriaceae Savolainen, V., Chase, M.W. et al. 2000. See general refer- – Autotrophic; stipules +; flowers solitary, paired or in ences. few-flowered cymes, regular or rarely slightly zygo- Soltis, D.E. et al. 2000. See general references. morphic; nectar-secreting disk often +; stamens 1 or Stevens, P.F. 2005. See general references.