Generic relationships of selected African genera of Apiaceae
by
Anthony Richard Magee
Thesis submitted in fulfilment of the requirements for the degree
PHILOSOPHIAE DOCTOR
in
BOTANY
in the
FACULTY OF SCIENCE
at the
UNIVERSITY OF JOHANNESBURG
SUPERVISOR: PROF. B.-E. VAN WYK CO-SUPERVISOR: PROF. P. M. TILNEY CO-SUPERVISOR: PROF. S. R. DOWNIE
September 2009
AFFIDAVIT: MASTER'S AND DOCTORAL STUDENTS TO WHOM IT MAY CONCERN
This serves to confirm that I Anthony Richard Magee Full Name(s) and Surname
ID Number 8109065065089 Student number 920001887 enrolled for the
Qualification PhD (Botany)
Faculty Science Herewith declare that my academic work is in line with the Plagiarism Policy of the University of Johannesburg which I am familiar.
I further declare that the work presented in the Generic relationships of selected African genera of Apiaceae (thesis) is authentic and original unless clearly indicated otherwise and in such instances full reference to the source is acknowledged and I do not pretend to receive any credit for such acknowledged quotations, and that there is no copyright infringement in my work. I declare that no unethical research practices were used or material gained through dishonesty. I understand that plagiarism is a serious offence and that should I contravene the Plagiarism Policy notwithstanding signing this affidavit, I may be found guilty of a serious criminal offence (perjury) that would amongst other consequences compel the UJ to inform all other tertiary institutions of the offence and to issue a corresponding certificate of reprehensible academic conduct to whomever request such a certificate from the institution.
Signed at Johannesburg on this the 2 S' September 2009
Signature Print name A f-) Ti-V7 H RqC-C.
STAMP COMMISSIONER OF OATHS Affidavit certified by a Commissioner of Oaths This affidavit conforms with the requirements of the JUSTICES OF THE PEACE AND COMMISSIONERS OF OATHS ACT 16 OF 1963 and the applicable Regulations published in the GG GNR 1258 of 21 July 1972; GN 903 of 10 July 1998; GN 109 of 2 February 2001 as amended. "There is grandeur in this view of life, with its several powers, having been originally breathed into a few forms or into one; and that, whilst this planet has gone cycling on according to the fixed law of gravity, from so simple a beginning endless forms most beautiful and most wonderful have been, and are being, evolved."
— CHARLES R. DARWIN (The Origin of Species) TABLE OF CONTENTS
SUMMARY i
ACKNOWLEDGEMENTS VII
CHAPTER 1: GENERAL INTRODUCTION AND OBJECTIVES 1
1 .1 GENERAL INTRODUCTION 1
1.2 OBJECTIVES 3
CHAPTER 2: GENERAL MATERIALS AND METHODS 6
2.1 MORPHOLOGY 6
2.2 FRUIT ANATOMY AND VITTAE STRUCTURE 7
2.3 DATA CAPTURING 8
2.4 PHYLOGENETIC ANALYSES 8
CHAPTER 3: SYSTEMATICS OF THE CAPNOPHYLLUM GROUP AND PUTATIVE
RELATIVES 12
3.1 INTRODUCTION 12
3.2 MATERIAL AND METHODS 14
3.3 RESULTS 18
3.4 DISCUSSION 30
3.5 KEY TO GENERA OF THE CAPNOPHYLLUM GROUP 35 CHAPTER 4: TAXONOMIC REVISION OF THE GENUS CAPNOPHYLLUM 36
4.1 INTRODUCTION 36
4.2 VEGETATIVE AND REPRODUCTIVE MORPHOLOGY 37
4.3 TAXONOMY OF THE GENUS CAPNOPHYLLUM.. 44
CHAPTER 5: TAXONOMIC REVISION OF THE GENUS DASISPERMUM 58
5.1 INTRODUCTION 58
5.2 VEGETATIVE AND REPRODUCTIVE MORPHOLOGY 59
5.3 TAXONOMY OF THE GENUS DASISPERMUM 68
CHAPTER 6: TAXONOMIC REVISION OF THE GENUS SCARABOIDES 97
6.1 INTRODUCTION 97
6.2 VEGETATIVE AND REPRODUCTIVE MORPHOLOGY 98
6.3 TAXONOMY OF THE GENUS SCARABOIDES.. 100
CHAPTER 7: SYSTEMATICS OF AFRICAN AND MALAGASY PIMPINELLA AND RELATED
GENERA 104
7.1 INTRODUCTION 104
7.2 MATERIAL AND METHODS 105
7.3 RESULTS AND DISCUSSION 106 CHAPTER 8: SYSTEMATICS OF MALAGASY PEUCEDANUM 117
8.1 INTRODUCTION 117
8.2 MATERIAL AND METHODS 118
8.3 RESULTS AND DISCUSSION 119
8.4 TAXONOMY OF THE GENUS BILLBURTTIA 125
CHAPTER 9: SYSTEMATICS OF THE GENUS EZOSCIADIUM 128
9.1 INTRODUCTION 128
9.2 MATERIAL AND METHODS 129
9.3 RESULTS AND DISCUSSION 130
9.4 TAXONOMY OF THE GENUS EZOSCIADIUM 139
CHAPTER 10: NEW TRIBAL DELIMITATIONS OF THE EARLY DIVERGING LINEAGES OF
APIACEAE SUBFAMILY APIOIDEAE 143
10.1 INTRODUCTION 143
10.2 MATERIAL AND METHODS 144
10.3 RESULTS 148
10.4 DISCUSSION 155
10.5 TAXONOMY OF THE PROTOAPIOID TRIBES 165
CHAPTER 11: GENERAL CONCLUSIONS 171
LITERATURE CITED 177 APPENDIX A: VOUCHER SPECIMENS OF FRUIT MATERIAL STUDIED 198
APPENDIX B: DNA ACCESSIONS USED FOR THE PHYLOGENETIC ANALYSES 202
APPENDIX C: PUBLICATIONS RESULTING FROM THIS STUDY 216 SUMMARY
Recent anatomical and molecular studies have highlighted the importance of the African and Malagasy Apiaceae, many of which have been found to occupy early diverging positions within the subfamilies Apioideae and Saniculoideae. Despite the recent interest in the African contingent however, there remain several anomalous and poorly known African and Malagasy taxa in which generic boundaries remain unclear and which have yet to be incorporated within the emerging tribal classification for the family. Generic circumscriptions and affinities amongst hitherto poorly known African and Malagasy genera are here explored using anatomical, cytological, morphological and molecular sequence data. Substantial rearrangements at almost all infrafamilial levels are formalized in order to incorporate the unique African and Malagasy members for the first time.
Generic circumscriptions and phylogenetic relationships of the Cape genera
Capnophyllum, Dasispermum, and Sonderina are explored through parsimony and
Bayesian inference analyses of nrDNA ITS and cpDNA rps16 intron sequences, morphology, and combined molecular and morphological data. The relationship of these genera with the North African genera Krubera and Stoibrax is also assessed.
Analyses of both molecular data sets place Capnophyllum, Dasispermum,
Sonderina, and the only southern African species of Stoibrax (S. capense) within the newly recognized Lefebvrea Glade of tribe Tordylieae. Capnophyllum is strongly supported as monophyletic and is distantly related to Krubera. The monotypic genus
Dasispermum and Stoibrax capense are embedded within a paraphyletic Sonderina.
This complex is distantly related to the North African species of Stoibrax in tribe
Apieae, in which the type species, Stoibrax dichotomum, occurs. Consequently,
i SUMMARY
Dasispermum is expanded to include both Sonderina and Stoibrax capense. A taxonomic revision of Dasispermum s.l. is presented which includes keys to the species, complete nomenclature, typifications, descriptions as well as geographical distributions. New combinations are formalized for Dasispermum capense, 0 hispidum, D. humile, and D. tenue. In addition two new species, namely D. grandicarpum and D. perrenans, are described. As a result seven species of
Dasispermum s.l. are recognised and can be distinguished from one another by their habit (life history and growth form), leaf morphology (leaf texture, leaf colour and breadth of the ultimate leaflet segments), inflorescence structure (length of the peduncle, presence or absence and division of involucre and involucel bracts), fruit morphology (relative length of the styles, fruit size, rib prominence and relative orientation) and fruit anatomy (shape of the cells external to the vittae).
The Cape endemic genus Capnophyllum is revised. As a result of valuable recent collections and extensive fieldwork, this hitherto neglected genus was found to comprise four annual species, two of which are newly described, namely
Capnophyllum lutzeyeri and C. macrocarpum. The four species are distinguished from one another by their fruit morphology (relative length of the styles, the shape and position of the stylopodium, fruit size, surface sculpturing, and the presence or absence of a sterile apical portion) and fruit anatomy (marginal wings slightly or prominently involute and secondary ribs present or absent). A comprehensive key to the species, their complete nomenclature and typification, together with complete descriptions and known geographical distributions for all the species are presented and illustrated.
The genus Scaraboides is described herein to accommodate a new species,
S. manningii, from the Tanqua Karoo in South Africa. This monotypic genus shares
ii SUMMARY the dorsally compressed fruit and involute marginal wings with Capnophyllum, but is easily distinguished by its erect branching habit, green leaves, scabrous umbels, and fruit with indistinct median and lateral ribs, additional solitary vittae in each marginal wing, and parallel, closely spaced commissural vittae. Despite the marked fruit similarities with Capnophyllum, analyses of DNA sequence data place
Scaraboides closer to Dasispermum, with which it shares the erect habit, green
(non-glaucous) leaves, and scabrous umbels.
The phylogenetic position of the African and Malagasy species of Pimpinella are assessed using nrITS sequence data. These species are found to ally with their
Eurasian counterparts within the tribe Pimpinelleae. The genus Pimpinella is rendered paraphyletic by the inclusion of African Cryptotaenia and the small African and Malagasy endemic genera Frommia and Phellolophium. Within the paraphyletic
Pimpinella three major clades are recovered, with the African species widely separated into two of the three clades. Based on the results of the molecular analyses it is clear that the current sectional classification for the genus, based largely on fruit vestiture, is largely artificial. Chromosome base number, however, was found to be consistent with the groupings recovered in the molecular trees.
Optimisation of this cytological data supports the separation of the African species into two clades. Those African and Malagasy Pimpinella species with a chromosome base count of n=11 form the most early diverging Glade together with Frommia which also has a base count of n=11. The remaining African species ally with several
Eurasian species and share a chromosome base count x= 9. Fruit anatomical data is also explored for both African and Malagasy Pimpinella as well as Cryptotaenia,
Frommia and Phellolophium.
iii SUMMARY
The genus Billburttia is described herein to include two new Madagascan endemic species, B. capensoides and B. vaginoides. Although the species appear superficially similar to those of the southern African endemic genus Notobubon, they are easily distinguished by fruit anatomical characters, such as the narrower commissure, the six commissural vittae, the position of the vascular tissue in the tip of the ribs, and sphaerocrystals distributed in and around the epidermis. The two last mentioned characters are proposed as generic apomorphies for Billburttia. The phylogenetic position of the genus is assessed using ITS and rps16 intron sequence data. Both parsimony and Bayesian analyses place Billburttia within the tribe
Apieae, and not closely related to either Peucedanum (Selineae) or the African peucedanoid genera (Lefebvrea Glade of Tordylieae).
The hitherto poorly known Cape endemic genus Ezosciadium (Apiaceae) is revised. This genus is highly distinctive and can be distinguished from other annual genera of the region by its pilose vegetative and reproductive organs, the sessile compound umbels with conspicuously unequal rays, the non-inflexed petal tips, the relatively small, highly-inflexed stamens which appear almost sessile, and the prominent carpophores which persist on the plant. The fruit are unusual in the presence of druse crystals around the carpophore and tanniniferous substances in the epidermal cells of the ribs. The phylogenetic position of the genus within the subfamily Apioideae is assessed using rbcL, trnQ-trnK and nrITS sequence data.
Ezosciadium capense is found to form part of an early diverging lineage within the subfamily, sister group to the Annesorhiza Glade and the genera Molopospermum and Astydamia, here described as the tribe Annesorhizeae. A comprehensive taxonomic revision, including typification, detailed descriptions, geographical range and illustrations of the genus Ezosciadium, is presented.
iv SUMMARY
Phylogenetic analyses of the cpDNA trnQ-trnK 5'exon region for 27 genera and 42 species of Saniculoideae and early diverging lineages of Apioideae were carried out to assess or confirm the tribal placements of the following anomalous genera: Annesorhiza, Astydamia, Chamarea, Choritaenia, Ezosciadium, ltasina,
Lichtensteinia, Marlothiella, Molopospermum and Phlyctidocarpa. To accommodate these unique early diverging members of the Apiaceae and to reflect their relationships, a new classification system has become necessary. Many of the early diverging genera (herein referred to the as protoapioids) can readily be distinguished from the euapioids (the remaining apioids) by the presence of scattered druse crystals in the mesocarp. The major discontinuity within the family, however, lies between the combined protoapioids and euapioids (representing an expanded Apioideae s.l., including the Saniculoideae) and the subfamilies
Azorelloideae and Mackinlayoideae. The broadened subfamily Apioideae s.l. is diagnostically different from the other subfamilies in the absence of rhomboidal crystals, the presence of druse crystals scattered throughout the mesocarp
(subsequently lost in the euapioids), and the non-woody endocarp. No such diagnostic characters are available to support the traditional or recently expanded concept of Saniculoideae. The broadened concept of Apioideae is also supported by the sporadic presence of true wings. This character can be variously interpreted from a phylogenetic point of view, but nevertheless has considerable diagnostic value. A new tribal classification system for the protoapioids is proposed on the basis of molecular, morphological and anatomical evidence. It makes provision for hitherto poorly known African taxa and comprises the following eight tribes, five of which are newly described: Annesorhizeae, Choritaenieae, Heteromorpheae,
v SUMMARY
Lichtensteinieae, Marlothielleae, Phlyctidocarpeae, Saniculeae and
Steganotaenieae.
vi ACKNOWLEDGEMENTS
The following people/organisations are sincerely thanked for their contribution(s) to this study:
My supervisors Prof. Ben-Erik van Wyk, Prof. Patricia M. Tilney and Prof.
Stephen R. Downie for their enthusiasm, patient supervision, valuable input
and advice, financial support, and the many exciting opportunities made
possible. It was truly an honour working with each of them.
Mrs Deborah Katz-Downie for her exceptionally kind hospitality and logistic
support during my visit to Illinois.
Dr Samuel N. Beshers and Prof. Lynn Wiley for generously hosting me in
their home in Illinois.
Mr Heiner and Mrs Eva Lutzeyer for their enthusiasm and hospitality while
performing field studies on their reserve in Stanford and for subsequent
collections of mature fruit.
Prof. Abraham E. van Wyk for useful discussion on the identity of the crystals
in Billburtia and valuable information and material of Choritaenia.
Dr John Manning for assistance with field work and alerting us to unusual
Apiaceae collections.
Mr Jean-Pierre Reduron for supplying mature fruit of Molopospermum.
Prof. Peter Goldblatt, and Mrs Edwina Marinus for there assistance in
obtaining recent herbarium material of Ezosciadium.
Dr Antoine Nicholas and Prof. Gregory Plunkett for providing DNA aliquots of
Choritaenia.
vii ACKNOWLEDGEMENTS
Dr. Jean-Noel Labat for providing a digital image of the type of Caucalis
capense.
Dr Hugh F. Glen, Dr Ian Hedge and Dr Gerrit Koorsen for translating the
diagnoses into Latin.
The curators and staff of the cited herbaria for their kind hospitality and
assistance during visits and for making specimens available on loan.
Dr Annah N. Moteetee for the management of loans to the University of
Johannesburg Herbarium.
Dr Rebecca Liu for keen interest in and assistance with fruit anatomy of
Apiaceae.
Ms Mary Ann E. Feist, Ms Jenny M. Cordes and Mr Clark Danderson for
logistical support, advice and welcome friendship during my work in the
molecular laboratory at the University of Illinois. Mr Clark Danderson is also
thanked for providing sequences of Astydamia and Bupleurum.
Dr Carolina I. Calvin° for supplying the published trnQ-trnK dataset of the
Saniculoideae with which I could build upon, as well as one of the trnQ-trnK
sequences for Phylictidocarpa.
Dr Alexei Oskolskii for hosting us during our visit to St. Peterburg.
Dr Eleena Y. Yembaturova for receiving and guiding us in Moscow.
Mr Pieter J. D. Winter for his enthusiasm and many useful discussions on
Apiaceae systematics.
Mr Andre Marais and his team from De Hoop Nature Reserve for their kind
hospitality during field studies.
viii ACKNOWLEDGEMENTS
The National Research Foundation and the South African Biosystematics
Initiative for funding and travel grants to visit Russia (Apiales IV) and the USA
(University of Illinois).
Department of Botany and Plant Biotechnology for the use of their facilities.
The Faculty of Science at the University of Johannesburg for financial support to visit Prof. Stephen Downie's laboratory at the University of Illinois.
The Lesley Hill Laboratory, the Jodrell Laboratory and Kew Herbarium for the
DNA aliquots.
The Molecular Systematics Laboratory at the University of Johannesburg for the use of their facilities.
Fellow students in the Department of Botany and Plant Biotechnology at
University of Johannesburg for their friendship.
Ms Marianne le Roux for much treasured friendship and companionship during field studies.
Dr James S. Boatwright for his unwavering support, companionship, critical proofreading and assistance with almost all facets of this study.
My family for their love, unlimited support and understanding.
0
ix CHAPTER 1
GENERAL INTRODUCTION AND OBJECTIVES
1.1 GENERAL INTRODUCTION
The Apiaceae are a large and notoriously complex family comprising ca. 455
genera and 3600-3751 species (Pimenov and Leonov 1993) or ca. 463 genera and
3,500 species (Plunkett et al. in press). Although the family has a near cosmopolitan
distribution, the highest diversity is found within the temperate regions of the
Northern Hemisphere. Because of the distinctive umbellate inflorescence and highly
specialised bicarpellate fruit, the Umbelliferae (=Apiaceae) were the first major
natural group of flowering plants to be recognised (Constance 1971). Despite being
the focus of intensive research (particularly over the last four to five decades) from
diverse fields of study (including anatomical, chemical, cytological, morphological,
and most recently molecular approaches), there remain many problems at
practically all infrafamilial levels of classification. Burtt (1991) attributed this in part to
"the very uneven state of our knowledge of the family in different parts of the word",
and highlighted the importance of the hitherto neglected southern Africa taxa in
gaining a better understanding of the family as a whole.
Currently, the family is undergoing a renewed global research effort in which
the results of molecular sequence data is allowing for a re-exploration and re-
interpretation of anatomical, morphological and cytological data with the aim to
produce a more natural classification system to replace the outdated and largely
Eurocentric treatment of Drude (1897-98). In this classical and monumental treatment, Drude (1897-98) proposed that the family be divided into three well-
1 CHAPTER 1: GENERAL INTRODUCTION AND OBJECTIVES defined subfamilies (viz. Apioideae, Hydrocotyloideae and Saniculoideae) based primarily on fruit anatomical data. The recent treatment by Plunkett (2004), based mainly on molecular sequence data, retained Apioideae and Saniculoideae as largely natural groups (although with some rearrangements required). The members of the polyphyletic Hydrocotyloideae were accommodated within two additional subfamilies, the Azorelloideae and Mackinlayoideae. The type genus Hydrocotyle, along with a few related genera were transferred to the closely related sister family
Araliaceae. At the tribal level, however, the system of Drude (1897-98) has been shown to be highly unnatural as a result of large-scale convergence (Downie and
Katz-Downie 1996; Downie et al. 1996, 1998, 2000a, b, c, 2001; Kondo et al. 1996;
Plunkett et al. 1996a, b, 1997; Valiejo-Roman et al. 1998; Katz-Downie et al. 1999;
Plunkett and Downie 1999; Spalik et al. 2004; Calvin° and Downie 2007). As a result the current tribal classification for the family is based exclusively on molecular sequence data (Downie et al. in press) and therefore many genera have not yet been allocated to tribes due to sampling limitations.
The sub-Saharan African and Malagasy contingent of Apiaceae is comparatively small with only 76 genera and 368 species (Van Wyk and Tilney
2004) and exhibits high levels of endemism (40 and 321, respectively). The high incidence of woodiness and other unusual leaf and particularly fruit anatomical characters (Van Wyk 2001; Liu 2004), as well as the many isolated and anomalous taxa, indicate that the African genera are critical to an understanding of higher order relationships within the family (Van Wyk and Tilney 2004). Recent molecular systematic studies (Downie and Katz-Downie 1996, 1999; Plunkett et al. 1996;
Chandler and Plunkett 2004; Calvin° et al. 2006; Calvin° and Downie 2007) have shown that the African and Malagasy genera are often sister to or basally divergent
2 CHAPTER 1: GENERAL INTRODUCTION AND OBJECTIVES within the major lineages of the family (discussed in more detail in Chapters 9, 10 and 11), suggesting an African or likely southern African origin for both subfamilies
Apioideae and Saniculoideae (Calvino et al. 2006; Calvin() and Downie 2007).
Despite the increased sampling of African and particularly southern African taxa in recent phylogenetic studies (Calvino et al. 2006; Calvin() and Downie 2007;
Magee et al. 2008d; Winter et al. 2008; Nicolas and Plunkett 2009), there remain several small African genera with uncertain circumscription and affinity (Burtt 1991;
Lebrun and Stork 1992; Van Wyk and Tilney 2004). While early diverging positions within the largest subfamily Apioideae and the sister subfamily Saniculoideae have been suggested for some of these genera based on recent fruit anatomical data (Liu et al. 2003a, b, 2007b, c; Liu 2004), the majority remain either poorly known or unstudied. Due to the improved molecular sampling of African and particularly the early diverging lineages of the Apioideae-Saniculoideae Glade by Calvirio (Calvino et al. 2006, 2008a, b; Calvino and Downie 2007) and the family-wide survey of fruit anatomical characters, particularly the early diverging lineages of Apioideae and
Saniculoideae, by Liu (Liu et al. 2003a, b, 2006, 2007a, b, c, 2009; Liu 2004), comparisons between these poorly known and previously unstudied African genera and other genera throughout the family are now possible.
1.2 OBJECTIVES
This project is aimed at providing more clarity on generic circumscriptions and affinities amongst hitherto poorly known African and Malagasy genera using anatomical, morphological, and molecular sequence data. The aim was also to consider the best way in which the existing tribal classification system be formally
3 CHAPTER 1: GENERAL INTRODUCTION AND OBJECTIVES modified to allow for the incorporation of anomalous genera. The specific objectives of the study are to:
Explore the putative relationships between the Cape endemic genera,
Sonderina and Capnophyllum, and the North African genera, Stoibrax and
Krubera, as well as the placement of the only southern African species of
Stoibrax, viz. S. capense.
Clarify the generic circumscriptions of the South African genera
Capnophyllum, Stenosemis, Sonderina and Dasispermum.
Revise the species delimitations, nomenclature, typifications, diagnostic
features, and geographical distributions of the taxonomically complex genera
Sonderina, Capnophyllum and the monotypic genus Dasispermum.
Determine the systematic placement of the African and Malagasy species of
Pimpinella, as well as their putative relatives Cryptotaenia, Frommia and
Phellolophium.
Determine the correct systematic placement of the Malagasy species of
Peucedanum.
Confirm the proposed systematic placement of the Malagasy genera
Andriana, Anisopoda, Cannaboides, Pseudocannaboides and Tana within the
early diverging lineages of the Apioideae and particularly the tribe
Heteromorpheae.
Examine the generic circumscription of the poorly known South African genus
Ezosciadium and assess its systematic affinities.
Determine the systematic placement of the anomalous genera Choritaenia,
Phlyctidocarpa and Marlothiella and explore the tribal delimitations of the
4 CHAPTER 1: GENERAL INTRODUCTION AND OBJECTIVES early diverging lineages of the subfamily Apioideae and their affinities to subfamily Saniculoideae.
5 CHAPTER 2
MATERIALS AND METHODS
Authorities for scientific plant names (according to Brummitt and Powell 1992)
are given in Appendices A and B or in Table 11.1. The list of literature references
includes only those cited in the text and not the taxonomic citations included in the
relevant synonymies (abbreviated as in Stafleu and Cowan 1976).
2.1 MORPHOLOGY
During the last five years extensive field work was undertaken in which many of the South African Apiaceae were studied in situ. During these excursions fresh as well as silica- (Chase and Hill 1991) and FAA-preserved materials (formaldehyde: acetic acid : alcohol : water; Sass 1958) for molecular and anatomical studies were collected. Photographs were taken to record habit and certain features of gross morphology. All specimens collected are kept at the University of Johannesburg
Herbarium (JRAU). In addition relevant collections of African and Madagascan
Apiaceae from the following herbaria were studied: BM, BOL, JRAU, K, LE, MO,
NBG, PRE, S, SAM and THUNB-UPS (herbarium acronyms as in Holmgren et al.
1990). Information concerning habitat and phenology was obtained from these sources as well as field notes. From this material, together with information from
Leistner and Morris (1976), the recorded distribution of all the species was ascertained and mapped. The distribution data was recorded using the quarter degree reference system (Edwards and Leistner 1971). Line drawings were made
6 CHAPTER 2: MATERIALS AND METHODS
with the aid of a camera lucida attachment on a Zeiss compound microscope or a
Wild M3Z stereomicroscope.
2.2 FRUIT ANATOMY AND VITTAE STRUCTURE
Preserved (FAA) and herbarium materials were used to study fruit anatomy.
Herbarium material was first re-hydrated and then placed in FAA for a minimum of
24 h. This material was subsequently treated according to a modification of the
method of Feder and O'Brien (1968) for embedding in glycol methacrylate (GMA).
This modification involves a final infiltration in GMA for five days. Transverse
sections, about 3 1.1rn thick, were cut using a Porter-Blum ultramicrotome. The
sections were examined for the presence of crystals using a light microscope, after
which they were stained according to the periodic acid Schiff/toluidine blue (PAS/TB)
method of Feder and O'Brien (1968). A list of voucher specimens for the fruit
anatomical study is given in Appendix A1. The terminology used to describe the fruit
anatomical features follows that proposed by Kljuykov et al. (2004) and is indicated
in the relevant labelled figures.
To study the three-dimensional structure of the vittae, mature fruit were
softened by soaking in boiling water for 24 h. The exocarp was then peeled off while
keeping the fruit submerged in water to prevent desiccation. A list of voucher
specimens used to study the external vittae structure of the mature fruit is given in
Appendix A2.
7 CHAPTER 2: MATERIALS AND METHODS
2.3 DATA CAPTURING
Whole fruit and anatomical sections were photographed using a JVC KY-
F1030 digital camera mounted on either a Zeiss compound microscope or a Wild
M3Z stereomicroscope. These images were then deep etched using the computer
program Adobe Photoshop CS version 8.0. Drawings were made with the aid of a
camera lucida attachment mounted on either the before-mentioned compound
microscope or the stereomicroscope.
2.4 PHYLOGENETIC ANALYSES
2.4.1 DNA extraction, amplification and sequencing— Total DNA was
extracted from herbarium or silica material using either the 2x CTAB method of
Doyle and Doyle (1987), or one of the following commercial extraction kits: DNeasy
Plant Mini Kit (Qiagen) or PureLink TM Plant Total DNA Purification Kit (Invitrogen).
The voucher specimen information and GenBank accession numbers for the
material used in the analyses are given in Appendix B.
The internal transcribed spacers (ITS) of nuclear ribosomal DNA was
amplified using the primer combinations of Sun et al. (1994). For amplification of the
chloroplast DNA rps16 intron and the flanking exon regions (tmQ-rps16 5' exon and
rps16 3' exon-trnK 5' exon) the primers of Downie and Katz-Downie (1996) and Lee
and Downie (2006) were used, respectively. Successfully amplified PCR products were purified using either a QlAquick PCR purification kit (Qiagen Inc.) according to the manufacturer's instructions, or according to the ExoSAP protocol of Werle et al.
(1994) using 5 units of Exonuclease I (New England Biolabs, Ipswich,
8 CHAPTER 2: MATERIALS AND METHODS
Massachusetts, USA) and 0.5 units of Shrimp Alkaline Phosphatase (Promega,
Madison, Wisconsin, USA). Sequencing reactions were carried out using the BigDye
Terminator version 3.1 Cycle Sequencing Kit (Applied Biosystems Inc.) and
sequenced using either an ABI (Applied Biosystems) 3130 XL or 3730 XL
sequencer.
2.4.2 Sequence alignment and phylogenetic analyses— Complementary
strands were assembled and edited using Sequencher version 3.1.2 (Gene Codes
Corporation) and manually aligned in PAUP* version 4.0b10 (Swofford 2002), with
gaps positioned so as to minimise nucleotide mismatches. For the large trnQ - trnK
matrix of Chapter 11, the sequences were initially aligned using the default pairwise
and multiple alignment parameters in the computer program Clustal X (gap opening
cost =15.00, gap extension cost = 6.66, DNA transition weight = 0.50; Jeanmougin
et al. 1998). This alignment was then checked and adjusted manually as necessary,
with gaps positioned so as to minimise nucleotide mismatches. In the latter matrix
unambiguous gaps were scored as presence/absence characters using the simple
indel coding method of Simmons and Ochoterena (2000).
Phylogenetic analyses of all data sets were conducted initially using the
parsimony (MP) algorithm of PAUP*. Character transformations were treated as
unordered and equally weighted (Fitch parsimony; Fitch 1971). As tree search
strategies differed between the datasets they are specified in the relevant Chapters.
Branch support for the MP analyses was determined using bootstrap percentage values (BP; Felsenstein 1985). Only values greater than or equal to 50% are reported and the following scale was applied for support percentages: 74%, weak;
75%-84%, moderate; and 85%-100%, strong.
9 CHAPTER 2: MATERIALS AND METHODS
After model selection with Modeltest version 3.1.2 under the corrected Akaike
information criterion (Akaike 1974; Posada and Crandall 1998), Bayesian inference
(BI; Yang and Rannala 1997) was implemented using MRBAYES version 3.1.2
(Huelsenbeck and Ronquist 2001; Ronquist and Huelsenbeck 2003). We employed
the 'standard' model (using default parameters) for both the indel data as well as the
coded gaps (Lewis 2001). The search strategies implemented for the different
datasets are specified in the relevant Chapters. The analyses were judged to have
reached stationarity when the standard deviation between the split frequencies
stabilised below 0.009. The initial one-fourth of trees where discarded as the 'burn-
in' phase. A majority rule consensus tree was produced from the remaining trees in
order to show the posterior probabilities (PP). The following scale was used to
evaluate the PPs: 0.5-0.84, weak; 0.85-0.94, moderate; 0.95-1.0, strong.
To assess congruency of relationships within the Capnophyllum group
(Chapter 3), as inferred by separate MP analyses of the ITS, rps16 intron and
morphological data sets, the bootstrap consensus trees from each analysis were
compared. These trees were considered incongruent only if they displayed 'hard'
(i.e., incongruencies with strong bootstrap values) rather than 'soft' (i.e.,
incongruencies with weak bootstrap values) incongruence (Seelanan et al. 1997;
Wiens 1998). In addition, a partition homogeneity test (incongruence length difference test, ILD; Farris et al. 1995) was performed in PAUP*. This test was implemented with 1,000 replicate analyses, using the heuristic search option with simple addition of taxa, TBR and the MULTREES option selected. To evaluate the significance of differing topologies, we used the Shimodaira-Hasegawa test (SH;
Shimodaira and Hasegawa 1999), as implemented in PAUP* (applying the RELL re- sampling method with 1,000 bootstrap replicates).
10 CHAPTER 2: MATERIALS AND METHODS
2.4.3 Evolution of morphological characters— Selected morphological characters were reconstructed onto the respective MP trees using parsimony with
Mesquite version 2.5 (Maddison and Maddison 2008).
1 1 CHAPTER 3
SYSTEMATICS OF THE CAPNOPHYLLUM GROUP AND PUTATIVE RELATIVES
3.1 INTRODUCTION
A more natural classification of the large and taxonomically complex
cosmopolitan family Apiaceae is currently emerging as a result of molecular
systematic studies, together with rigorous comparisons of morphological and
anatomical data. Several small genera of uncertain circumscription and affinity are
evident in recent checklists of African Apiaceae (Built 1991; Lebrun and Stork 1992;
Van Wyk and Tilney 2004). The majority of these genera are either poorly known or have not been studied in recent years but may be extremely important in the understanding of relationships within the family as a whole. The South African endemic genera Capnophyllum, Dasispermum and Sonderina were identified as three such taxa. Of these, only two species (Dasispermum suffruticosum and
Sonderina humilis) have previously been included in molecular systematic studies
(Calvin() et al. 2006; Winter et al. 2008). In the phylogenetic analysis by Winter et al.
(2008) using nuclear ribosomal DNA internal transcribed spacer (ITS) sequences, both of these species were shown to be closely related to a group of recently circumscribed African peucedanoid genera, here referred to as the Lefebvrea Glade
(viz. Afroligusticum, Afrosciadium, Cynorhiza, Lefebvrea A.Rich., Nanobubon and
Notobubon) within tribe Tordylieae.
The monophyly of Sonderina has of yet not been assessed, nor has its putative relationship with Stoibrax been confirmed. The genus Sonderina was described by Wolff (1927) to accommodate four of five South African species
12 CHAPTER 3: SYSTEMATICS OF THE CAPNOPHYLLUM GROUP AND PUTATIVE RELATIVES
previously included in Ptychotis W.D.J.Koch by Sonder (1862). Wolff (1927) transferred the fifth species, Ptychotis didyma Sond., to the genus Tragiopsis Pomel
(now Stoibrax), which already included four North African species. Adamson (1939) considered this geographically disjunct treatment to be unnatural and transferred the
South African species, Tragiopsis didyma (Sond.) H.Wolff, to Sonderina. Burtt
(1989), however, argued that such a Cape and North African disjunction was not uncommon and transferred Sonderina didyma (Sond.) Adamson, along with the
North African species, back to the genus Stoibrax, as Stoibrax capense. Burtt (1989, pg 145), furthermore, expressed his doubts about the generic concept of Sonderina, stating that the genus was "probably too close to Stoibrax for it to be maintained."
Burtt (1991), in his checklist of southern African Umbelliferae, treated five species within Sonderina. One of these, the Namibian endemic Sonderina streyi Merxm., has subsequently been transferred to the early diverging African genus Anginon
(Allison and Van Wyk 1997). As a result, only four closely related species are now recognised within the taxonomically difficult genus Sonderina.
A similar disjunction has also been proposed for the genus Capnophyllum, with some authors (e.g., Tutin et al. 1968; Dyer 1975) expanding the genus to include the Mediterranean Capnophyllum peregrinum (L.) Lange. Meikle (1977), however, treated the Mediterranean species as distinct under the monotypic genus
Krubera, a decision maintained by Burtt (1991). A recent taxonomic revision of
Capnophyllum (Chapter 4; Magee et al. 2009c) recognized four species, two of which were newly described and excluded Krubera peregrina on the basis of important differences in fruit anatomy.
A thorough taxonomic study of the genera Capnophyllum, Dasispermum and
Sonderina along with extensive field work has revealed one new monotypic genus
13 CHAPTER 3: SYSTEMATICS OF THE CAPNOPHYLLUM GROUP AND PUTATIVE RELATIVES
(herein described as "Scaraboides manningii") and four new species (Magee et al.
2009c; Magee et al. unpublished). The present study is aimed at resolving generic
circumscriptions and relationships of these previously neglected South African
endemic genera. As the phylogenetic relationships of African Apiaceae genera are
often hard to predict on the basis of morphological characters alone, analyses of
both morphology and anatomy in combination with molecular data (specifically, ITS
and rps16 intron sequences) are here presented and explored.
3.2 MATERIAL AND METHODS
3.2.1 Taxon sampling— In order to assess the generic delimitations and
phylogenetic relationships of the Cape endemic genera Capnophyllum (12 new accessions), Dasispermum (two new accessions), Sonderina (13 new accessions), the undescribed monotypic genus "Scaraboides"(two new accessions) and the largely North African genus Stoibrax (five new accessions). Additional accessions of the rps16 intron region for the closely related African peucedanoid genera Cynorhiza
(two new accessions), Nanobubon (2 new accessions) and Notobubon (five new accessions) were also included. The 45 new accessions for which ITS (18 accessions) and rps16 intron (27 accessions) sequences were obtained and the previously published rps16 intron accessions are listed in Appendix B1 and previously published ITS accessions are available in Winter et al. (2008). The newly obtained ITS sequences were added to the 125 taxon ITS matrix of Winter et al.
(2008). This matrix represents all tribes and major clades of the apioid superclade plus outgroups from tribes Smyrnieae and Oenantheae (Downie et al. 2001), with those species of the latter used to root the trees. The newly obtained rps16 intron
14 CHAPTER 3: SYSTEMATICS OF THE CAPNOPHYLLUM GROUP AND PUTATIVE RELATIVES
sequences were analysed with 27 additional rps16 intron sequences from GenBank
(Appendix B1), the latter also representing several relevant major clades of the
apioid superclade. The rps16 intron trees were rooted with Sium latifolium and
Berula erecta of tribe Oenantheae. To further explore relationships within the
Capnophyllum group, combined data sets (ITS/rps16 intron, ITS/morphology and
ITS/rps16 intron/morphology) for 31 taxa of the Lefebvrea Glade were analysed, with
Lefebvrea abyssinica A.Rich. used as the outgroup. A matrix of 23 morphological
and anatomical characters was prepared based on examination of herbarium
specimens and literature (Tables 3.1 and 3.2; Magee et al. 2008a, b, 2009a, c;
Winter et al. 2008).
3.2.2 Phylogenetic analyses— Phylogenetic analyses of all data sets were
conducted initially using maximum parsimony, as described in Chapter 2, with gaps treated as missing data. Tree searches were performed using a heuristic search with
500 random sequence additions, TBR branch swapping and the MULPARS option in effect, but saving no more than five of the shortest trees from each search. These equally parsimonious trees were then used as starting trees for TBR branch swapping (MULPARS and STEEPEST DESCENT in effect) with the maximum number of trees saved set at 12,000; these trees were permitted to swap to completion (Downie et al. 1998). Bootstrap percentage values for the separate ITS and rps16 intron data sets were determined from 500,000 replicate analyses using fast stepwise addition of taxa, while BP values for the morphological and combined data sets of the Lefebvrea Glade were determined from 1,000 bootstrap replicates, holding 10 trees per replicate and with TBR and MULPARS selected. All data sets
15 CHAPTER 3: SYSTEMATICS OF THE CAPNOPHYLLUM GROUP AND PUTATIVE RELATIVES
TABLE 3.1 Morphological and anatomical characters and states used in the phylogenetic analysis of the Lefebvrea Glade.
1. Life history' (monocarpic = 0; short-lived perennial = 1; perennial = 2); 2. Habit (herbs =
0; rhizomatous = 1; suffrutices = 2; shrubs or shrublets =3); 3. Growth pattern (monopodial = 0; sympodial = 1); 4. Leaf persistence (one-seasoned or deciduous = 0; permanent, evergreen = 1); 5. Leaf arrangement (radical or if somewhat cauline then borne on deciduous branches = 0; cauline, borne on permanent branches =1); 6. Leaf texture (coriaceous = 0; flimsy = 1; sclerophyllous =2); 7. Leaf colour (concolourously green or green above = 0; glaucous =1); 8. Inflorescence vestiture (glabrous = 0; scabrous = 1); 9. Ratio of functionally male flowers (equal ratio of male to female flowers in all raylets of the umbellule = 0; inner raylets of umbellules functionally male =1); 10. Involucre and involucel bracts2 (present = 0; absent or very much reduced = 1); 11. Involucre and involucel bracts type 3 (absent or all simple = 0; at least some compound, resembling the leaves = 1); 12. Petal vestiture (leathery = 0; papillose = 1); 13. Fruit length (more than 9 mm = 0; less than 9 mm = 1); 14. Fruit compression (platyspermous = 0; isodiametric = 1); 15. Fruit in lateral view (narrowly elliptic = 0; very broadly elliptic to rotund = 1); 16. Ribs (median and lateral ribs markedly less developed than the marginal ribs = 0; median and/or lateral ribs as well developed as the marginal ribs = 1); 17. Ribs (obtusely tipped = 0; almost trifid with prominent tapering tips = 1); 18. Secondary ribs (absent = 0; usually present = 1); 19. Marginal wings (absent or flat = 0; involute = 1); 20. Commissural surface (flat = 0; concave = 1); 21. Commissure (100% from rib tip to rib tip = 0; from near rib tip to near rib tip = 1; from at most rib base to rib base = 2); 22. Rib vittae (absent = 0; present at base of all ribs = 1; present in marginal wings = 2); 23. Cells external to vittae (indistinct =0; square = 1; enlarged, upright = 2).
1 Field observations of Dasispermum suffruticosum and Sonderina sp. 1 indicate that these species are not monocarpic annuals but rather short-lived perennials lasting for only a few seasons depending on rainfall, possibly an adaptation to the dune habitat in which they both occur. Perennials include all shrubs and also species with permanent fleshy roots.
2 Some species of Sonderina are distinct in that their involucre and involucel bracts are usually absent or at best strongly reduced and rudimentary.
3 Sonderina hispida, Sonderina sp. 1 and Stoibrax capense are unusual in that at least some of the involucre and involucel bracts are pinnately divided, thus resembling the leaves.
16
C=)COIDCO N DO 0 OCOQC, 0 CD 000 0000 0 N CNN-
N 000000000000 N 00.-0T-N00000000
N 00000000N00000000000000NNNNNT-
O N 000T- T- T- T- 000000000000000T-00000000
oaaaaaaaaaaaa,--aaaoaaaa
de. 03 000.-.-?-aapacpaacpaaaaaaaaaaaaaacpa0 la G I•••• oaaaa aaaaaaaaaaaaaaaa a coo aaa,
bvrea CO CD 000000" 0000 OD QC) CI C■ QC) fe
Le oaa , ,-,-,a,-aaaaaaaaaaaaa