View/Open - ResearchSpace - University of KwaZulu-Natal

Pollination of Plectranthus L’Hér. 

(Lamiaceae) along the 

Eastern seaboard of southern Africa 

by 

Christina Johanna Potgieter 

Submitted in fulfilment of the academic requirements 

for the degree of Doctor of Philosophy 

in the School of Biological & Conservation Sciences, 

University of KwaZulu-Natal, Pietermaritzburg. 

December 2009

ABSTRACT 

Pollination data is provided for a third of the Plectranthus species in southern Africa. In 

the largest genus of Lamiaceae in the region (53 species), 18 were studied, plus two 

species of allied genera (Pycnostachys urticifolia and Aeollanthus parvifolius). Study of 

these 20 species aimed to describe the groups of pollinators that have driven pollinator 

specialisation. Case histories are provided upon which future studies of Lamiaceae 

pollination, breeding systems and speciation may be based. Bees (Apidae) and flies 

(Nemestrinidae, Acroceridae and Tabanidae) are the main pollinating insect groups. 

Seven straight-tubed Plectranthus species show a match between corolla tube- and 

proboscis length of nectar-feeding pollinators. Long-proboscid nemestrinid flies are 

specialised on long-tubed Plectranthus species (P. ambiguus, P. hilliardiae, P. reflexus 

and P. saccatus), while shorter-proboscid flies of all three families are important 

pollinators of straight-tubed species with medium- and short corolla tubes. Seven 

species with sigmoid corolla tubes are bee-pollinated, with fly-pollination prevalent in 

some. Bent corolla tubes, coupled with length, act as barriers to illegitimate nectarfeeders 

and ensures alignment of pollinators for effective pollen placement and carryover. 

It is suggested that straight-tubed species may have evolved from sigmoid-tubed 

species. Long-tubed species with straight corollas in other Lamiaceae may show 

convergent pollination by long-proboscid flies, with the guild being dependent on 

habitat and distribution of plants and flies. Formal establishment of the 

Stenobasipteron wiedemanni pollination guild extends the study from Plectranthus to 

selected Acanthaceae, Orchidaceae, Balsaminaceae, Gesneriaceae and Iridaceae, 

occurring in forested habitat along the Eastern seaboard. Micro-distillation of essential 

oils confirmed parentage of a putative natural hybrid; once established, hybrid data 

allows studies of the importance of natural hybridisation events in explaining pollinator 

fidelity. Nectar sugar studies in Plectranthus mostly showed sucrose dominance; cases 

of hexose dominance are noted and discussed. Nectar volume and concentration 

proved variable and do not fit any trends. Pollination by medium-proboscid acrocerid 

flies has importance for ‘medium-tubed’ plants, since six of the Plectranthus species 

are solely or partially reliant on Acroceridae for pollination. An appendix with 

consolidated data describes the 20 study species i.t.o. morphology, habitat, study sites, 

field work, pollinator observations and insect vouchers. 

ii

STUDENT DECLARATION 1 

Pollination of Plectranthus L’Hér. (Lamiaceae) along the Eastern seaboard 

of southern Africa 

(Thesis Title) 

declare that : 

I, Christina Johanna Potgieter 

902409960 

iii 

(Full name) 

(Student Number) 

(i) The research reported in this dissertation, except where otherwise indicated, 

is the result of my own endeavours in the School of Biological and 

Conservation Sciences, University of KwaZulu-Natal, Pietermaritzburg; 

(ii) This dissertation has not been submitted for any degrees or examination at 

any other University; 

(iii) This thesis does not contain data, figures or writing, unless specifically 

acknowledged, copied from other researchers; and 

(iv) Where I have reproduced a publication of which I am an author or co-author, I 

have indicated which part of the publication was contributed by me. 

Signed at Pietermaritzburg on the ………..…. day of ………………………………, 2009. 

__________________________ 

SIGNATURE

STUDENT DECLARATION 2: PUBLICATIONS 

Publication 1: 

Potgieter, C.J., Edwards, T.J., Miller, R.M., Van Staden, J., 1999. Pollination of 

seven Plectranthus spp. (Lamiaceae) in southern Natal, South Africa. Plant 

Systematics and Evolution 218, 99–112. 

All field work and write-up was done by C. Potgieter, with assistance from Prof. T. 

Edwards (discussions and editing), Dr R. Miller (Diptera identification) and Prof. J. van 

Staden (discussions and editing). 


Potgieter, C.J., Edwards, T.J., 2001. The occurrence of long, narrow corolla tubes 

in southern African Lamiaceae. Systematics and Geography of Plants 71, 493– 

502. 


Edwards (discussions and editing). 


Potgieter, C.J., Edwards, T.J., 2005. The Stenobasipteron wiedemanni (Diptera, 

Nemestrinidae) pollination guild in eastern southern Africa. Annals of the Missouri 

Botanical Garden 92: 254–267. 


Edwards (discussions and editing). 


Viljoen, A.M., Demirci, B., Baser, K.H.C., Potgieter, C.J., Edwards, T.J., 2006. 

Microdistillation and essential oil chemistry - a useful tool for detecting 

hybridisation in Plectranthus (Lamiaceae). South African Journal of Botany 72, 

99–104. 

All field work and sample collection in South Africa was done by C. Potgieter; the idea 

of analyzing hybrids was put forward by C. Potgieter and Prof. T. Edwards. The 

laboratory-based work and analysis was arranged and executed by Prof. A Viljoen, in 

collaboration with Dr B. Demirci and Dr K. Baser (in Turkey). 


Potgieter, C.J., Edwards, T.J., Van Staden, J., 2009. Pollination of Plectranthus 

spp. (Lamiaceae) with sigmoid flowers in southern Africa. South African Journal 

of Botany 75: 646–659. 


Edwards (discussions and editing) and Prof. J. van Staden (discussions and editing). 

Signed at Pietermaritzburg on the ………..…. day of ………………………………, 2009. 

__________________________ 

SIGNATURE 

iv

DECLARATION BY SUPERVISOR 

We hereby declare that we acted as Supervisors for this MSc / PhD student: 

Student’s Full Name: Christina Johanna Potgieter 

Student Number: 902409960 

Thesis Title: Pollination of Plectranthus L’Hér. (Lamiaceae) 

along the Eastern seaboard of southern Africa 

Regular consultation took place between the student and ourselves throughout the 

investigation. We advised the student to the best of our ability and approved the final 

document for submission to the Faculty of Science and Agriculture Higher Degrees 

Office for examination by the University appointed Examiners. 

__________________________ 

SUPERVISOR: PROFESSOR T.J. EDWARDS 

__________________________ 

CO-SUPERVISOR: PROFESSOR J. VAN STADEN 

v

DECLARATION BY CO-SUPERVISOR 

We hereby declare that we acted as Supervisors for this MSc / PhD student: 

Student’s Full Name: Christina Johanna Potgieter 

Student Number: 902409960 

Thesis Title: Pollination of Plectranthus L’Hér. (Lamiaceae) 

along the Eastern seaboard of southern Africa 

Regular consultation took place between the student and ourselves throughout the 

investigation. We advised the student to the best of our ability and approved the final 

document for submission to the Faculty of Science and Agriculture Higher Degrees 

Office for examination by the University appointed Examiners. 

__________________________ 

SUPERVISOR: PROFESSOR T.J. EDWARDS 

__________________________ 

CO-SUPERVISOR: PROFESSOR J. VAN STADEN 

vi

ACKNOWLEDGEMENTS 

Financial support was received from the Foundation for Research Development (FRD), 

now National Research Foundation (NRF); and the University of Natal Research Fund 

(URF), now University of KwaZulu-Natal (UKZN) Research Office. 

All-important access to and accommodation at field sites were provided by the Natal 

Parks Board, now Ezemvelo KZN Wildlife (Oribi Gorge and Umtamvuna Nature 

Reserves); many thanks to Rod Potter and Rob Wolter in particular. Miles Hunt, previous 

owner of Leopard’s Bush NR, Karkloof, and Vernon Green and the Booysens (of the 

Dargle area) are also thanked for access to field sites. 

The National Botanical Institute, now South African National Biodiversity Institute, 

provided data from the National Herbarium, Pretoria (PRE) Computerised Information 

System (PRECIS). The National Herbarium, Pretoria (PRE), KwaZulu-Natal Herbarium, 

Durban (NH) and Bews Herbarium, UKZN (NU) are thanked for access to distribution 

data and plant collections. 

Many persons contributed their time and expertise towards this project and I 

wish to thank and list them: 

Connal Eardley, Denis Brothers, David Barraclough, Ray Miller and Brian Stuckenberg 

(now late) gave assistance with insect identifications. Fred Gess (Entomology 

Department, Albany Museum), Margie Cochrane (Entomology Collections Manager, 

South African Museum) and Brian Stuckenberg and David Barraclough (Natal Museum) 

assisted with insect distribution and collectors’ data; their respective institutions are 

also thanked for access to this data. 

Access to a scanning electron microscope and photographic and imaging equipment 

was courtesy of the Centre for Electron Microscopy at UKZN, Pietermaritzburg, and the 

friendly staff at this facility assisted generously with their time. Riyad Ismail, Andrew 

Simpson, Dave Thompson, Mark Todd and Toni Boddington (variously of the 

Cartographic Unit, UKZN), gave assistance with mapping. 

Rogan Roth gave useful photographic help and Trevor Edwards, Steve Johnson, Neil 

Crouch, Guy Upfold and Geoff Nichols generously shared their photographs for posters 

and publications. 

vii

Specific pollinator field observations were shared by John Manning, Peter Goldblatt, 

Craig Symes, Tracy McLellan, Dino Martins, Dirk Bellstedt, Neil Crouch and Geoff 

Nichols. 

Jeff Finnie and Mike Smith kindly advised me on GC techniques; Ben-Erik van Wyk, 

Clinton Carbutt and Tracy Odendaal generously double-checked a few nectar sugar 

samples. 

Much appreciated field assistance and company in the field were provided by Trevor 

Edwards, Edna Meter, Joslyn Taylor, Dave Thompson, Clinton Carbutt, Carol-Ann 

Rolando, Barbara Bleher, Richard Beckett, Tony Abbott, Neil Crouch, Cameron and 

Rhoda McMaster, Bella and Danie du Toit, Isabel Johnson and Pev Curry. 

Various discussions and advice on manuscripts came from Esmé Hennessy, Kathleen 

Gordon-Gray, Steve Johnson, Brian Stuckenberg (now late), Steven Piper (now late), 

Denis Brothers, David Barraclough, Mervyn Lotter, Neil Crouch, Shelah Morita and 

Alan Paton. 

The technical and administrative staff of the School of Biological & Conservation 

Sciences (and previously of the Department of Botany), were always prepared to 

assist; and several of the students that I have known through the years have helped 

me in many ways. Special thanks go to Angela Beaumont for her constant 

encouragement. 

I thank my supervisors, Trevor Edwards and Hannes van Staden, for their patience and 

continued support of the project. In particular, I wish to thank Trevor Edwards, who 

suggested the project and acted as primary supervisor, for maintaining a high level of 

interest and enthusiasm over many years, which is rare to find. 

My husband, Pev Curry, was very patient throughout and I thank him for supporting me 

in many practical ways, which allowed me the time to put the thesis together. His 

presence in the field also added some fortunate pollinator observations, even though 

he disagrees with killing voucher insects. 

This thesis is dedicated to my late parents, Carel and Marina Potgieter, who both 

passed away during the course of the study. They instilled in me the love of natural 

history that set me on this path, for which I am grateful. 

viii

CONTENTS 

Abstract ……………………………………………………………………… ii 

Student Declaration 1 …………………………………………………….………. iii 

Student Declaration 2 ……..……………………………………………………. ... iv 

Declaration by Supervisor ……………………………………………………. v 

Declaration by Co-supervisor ……………………………………………………. vi 

Acknowledgements ……………………………………………………………... vii 

Contents ………………………………………………………………………. ix 

Chapter 1: Introduction …………………………………………………….. 1 

Chapter 2: Pollination of Straight-tubed species …………………………… 14 

Chapter 3: Pollination of Sigmoid-tubed species …………………………… 29 

Chapter 4: Convergent pollination in southern African Lamiaceae …………. 44 

Chapter 5: A new Pollination Guild …………………………………………….. 55 

Chapter 6: Natural Hybrids ……………………………………………………… 70 

Chapter 7: Nectar studies ……………………………………………………… 77 

Chapter 8: Discussion and Conclusions ……………………………………. 102 

Appendix: Descriptive and pollinator accounts for twenty study species ……... 113 

ix

CHAPTER 1: INTRODUCTION 

Plectranthus L’Hér. (Lamiaceae) is a horticulturally important genus of predominantly 

herbaceous plants that is increasingly popular in indigenous landscaping in South 

Africa (Van Jaarsveld 2006), and internationally in the potted plant trade (Brits et al. 

2001, Brits & Ling Li 2008). Current horticultural research is focused on aspects of 

flowering in the genus (Ascough & Van Staden 2007, Ascough et al. 2008), while 

chemical research (Abdel-Mogib et al. 2002, Stavri et al. 2009) and ethno-botanical 

research (Rabe & Van Staden 1998, Lukhoba et al. 2006, Van Zyl et al. 2008) point to 

the biomedical potential of this ornamental genus. 

The genus Plectranthus is diverse in terms of floral morphology, especially on the 

sandstone islands of southern KwaZulu-Natal (KZN) and the northern parts of the 

Eastern Cape (EC), i.e. the Pondoland Centre of Endemism sensu Van Wyk & Smith 

(2001) where a total of 29 described species of Plectranthus occur. Eleven of these 

species (or sub-species) are endemic or near-endemic to the region: Plectranthus 

aliciae (Codd) Van Jaars. & T.J.Edwards, P. brevimentum T.J.Edwards, P. ernstii 

Codd, P. hilliardiae Codd, P. malvinus Van Jaars. & T.J.Edwards, P. oertendahlii 

T.C.E.Fr., P. oribiensis Codd, P. praetermissus Codd, P. reflexus Van Jaars. & 

T.J.Edwards, P. saccatus Benth. subsp. pondoensis Van Jaarsv. & Milstein, P. stylesii 

T.J.Edwards (Van Jaarsveld & Edwards 1997; Van Wyk & Smith 2001). These endemic 

species probably form part of a natural group of species, found in Eastern and Southern 

Africa and in Madagascar. The details of relationships within this group are not clear. 

The Pondoland Centre is described as a “major centre of diversity and endemism for 

the genus” [Plectranthus] (Van Wyk & Smith 2001). 

The general vegetation type of the Pondoland Centre is grassland plateaux, dissected 

by deep, narrow river gorges lined by patches of forest (Van Wyk & Smith 2001). It is in 

and around these forested gorges that most of the endemic Plectranthus species 

occur. The sandstone regions of KZN and Pondoland have been described as 

remarkable centres of endemism with several species which are uncommon or absent 

on surrounding substrates (Van Wyk 1990). Many of these species are palaeoendemic 

forest and forest margin elements, and studies on reproductive biology and 

population dynamics have been encouraged for those species that exhibit poor seedset 

and that are vulnerable to extinction (Van Wyk 1990). The area is also rich in

apparent neo-endemics with strong affinities to the Cape and Afro-montane areas (Van 

Wyk & Smith 2001). Plectranthus does not fit well within either of these broad phytogeographic 

groups. The genus is Tropical in origin (Paton et al. 2004) with relatively 

few species occurring in the Cape Floral Region and along montane corridors. 

1 

1 

1 

1 

1 

1 1 

2 

1 1 3 

2 

1 

1 1 3 4 1 1 3 

1 1 2 1 4 10 8 8 1 

1 1 1 5 13 5 7 

1 5 10 9 5 

5 6 7 

1 1 7 8 18 2 

1 1 5 

4 1 2 3 9 11 

9 

1 1 2 2 1 6 3 

1 1 2 1 7 2 4 1 

1 2 3 4 22 4 2 

2 1 7 1 7 2 

10 

1 1 

6 8 2 

1 5 3 10 3 

2 11 9 1 2 1 

1 1 6 2 4 4 4 12 5 1 

1 1 3 2 4 12 1 

2 1 2 4 3 7 11 8 2 

1 2 5 4 2 3 1 6 11 15 2 1 

1 

2 1 2 

1 1 

1 1 4 3 3 

3 4 4 

4 7 1 

2 7 9 8 11 1 

1 5 5 

Figure 1: Map of southern Africa showing number of Plectranthus species per half 

degree square, indicating areas of high diversity; after Codd (1985a) and specimens 

from PRE and NU Herbaria. Areas with ten or more species are shown in red, with the 

three main areas of species diversity indicated with red boxes. 

In southern Africa, Plectranthus exhibits its highest level of diversity in the Pondoland 

Centre of Endemism, where the main study areas for this research were located. Oribi 

Gorge (3030C) is home to 18 species and Umtamvuna Gorge (3030C and 3130A) has 

12 species (Fig. 1). Several other regions of high Plectranthus diversity are indicated, 

with a prominent area being 2531C with 15 species (the area surrounding Barberton in 

Mpumalanga Province). The area north-west of Pietermaritzburg, in the KZN Midlands, 

Chapter 1/ 2

also boasts a high species number, with 13 species of Plectranthus present in the half 

degree square that includes the Karkloof (2930A). 

Taxonomy and Phylogeny 

Plectranthus belongs to the subfamily Nepetoideae, tribe Ocimeae, subtribe 

Plectranthinae. The Ocimeae originated in Asia (Paton et al. 2004) and gave rise to 

the African Ociminae and Plectranthinae. With about 300 species world-wide, the 

genus covers a wide distribution in the tropical and warm regions of the Old World 

(Retief 2000). Its generic boundary is controversial, with Paton et al. (2004) showing 

that the current circumscription is paraphyletic and requires expansion to include allied 

genera such as Pycnostachys Hook., Solenostemon Thonn., Aeollanthus Mart. ex 

K.Spreng., Thorncroftia N.E.Br. and Tetradenia Benth. Even without these inclusions it 

is the largest genus of Lamiaceae in southern Africa, with 53 species described to date 

(Codd 1975, 1985a; Van Jaarsveld & Edwards 1991, 1997; Van Jaarsveld & Hanky 

1997; Edwards et al. 2000; Van Jaarsveld & Van Wyk 2004; Edwards 2005; Winter & 

Van Jaarsveld 2005). 

The taxonomy of Plectranthus is fairly well established (Codd 1975 & 1985a, b), but 

what has not been adequately resolved is a phylogeny for the genus. A phylogenetic 

study of the tribe Ocimeae, to which Plectranthus belongs, by Paton et al. (2004) 

included only five Plectranthus and one Pycnostachys species relevant to the current 

pollination study, with no long-tubed species included. In addition, it was found that 

Plectranthus is paraphyletic and that further sampling is needed to recognise 

monophyletic groups within the subtribe Plectranthinae (Paton et al. 2004). These 

samples will need to include more of the endemic species of the Pondoland Centre 

before conclusions can be drawn with respect to phylogeny and the evolution of their 

pollination systems. 

Studies based on phylogenies are able to polarise characters and interpret pollinator 

shifts within a genus, for example Johnson et al. (1998) - Disa P.J.Bergius 

(Orchidaceae), Goldblatt et al. (2000) - Sparaxis Ker Gawl (Iridaceae), Goldblatt et al. 

(2001) - Gladiolus L. (Iridaceae), Johnson et al. (2002) - Zaluzianskya F.W.Schmidt 

(Scrophulariaceae), Beardsley et al. (2003) - Mimulus L. (Scrophulariaceae), Wilson et 

al. (2006) - Penstemon Mitch. and Keckiella Straw (Scrophulariaceae). Phylogenies 

Chapter 1/ 3

enhance the understanding of floral character evolution (Kay et al. 2006); hence the 

lack of such a complete phylogeny in Plectranthus and its allies is unfortunate. 

General importance of pollination studies 

Pollination Biology is a growing field of study, complementing the natural history of 

“anthecology” by using rigorous analytical techniques, the quantification of 

observations and experimentation (Baker & Baker 1986). Many long-held theories 

regarding co-evolution, mimicry and plant-animal interactions are being tested 

experimentally, with some researchers suggesting that descriptive pollination biology 

studies, combined with theoretical mating system studies, will lead to a “new plant 

reproductive biology” (Morgan & Schoen 1997). The “traditional descriptive, naturalhistory 

approach” [to understanding floral evolution] is being broadened to incorporate 

various forms of experimentation, genetics, mathematical theory and pollinator 

behaviour, thus integrating plant reproductive biology into evolutionary ecology (Harder 

& Barrett 2006). 

There is, however, still a great deal of base-line observational data that needs to be 

collected to elucidate the pollinators or pollination syndromes for most plant genera, 

which restricts the theoretical conclusions that can be made globally. In developing 

countries that are species rich, little is known about pollination systems (Johnson & 

Steiner 2000), which means that theories from northern hemisphere studies may not be 

appropriate for generalisations about pollination syndromes. In a review of angiosperm 

speciation, as driven by pollinators and as evidenced by floral diversity, Johnson (2006) 

concludes that the environmental factors behind this evolution need more attention, 

and that “a diversity of approaches from natural history to molecular biology” is 

necessary. In the editorial of a recent volume of the South African Journal of Botany, 

which is dedicated to pollination studies, Johnson et al. (2009) reiterate that “solid 

documentation of pollination systems” is the basis upon which the ecology and 

evolution of plants may be understood w.r.t. pollination systems. 

Apart from adding to basic natural history information, pollination studies are critical in 

conservation management and agriculture. Conservation management plans need to 

incorporate studies on floral reproductive biology in their syntheses. The pollination 

component of any ecosystem is important for: (1) management of ecosystems; (2) 

servicing agricultural needs and (3) revealing the co- and contra-evolutionary 

Chapter 1/ 4

development of plants and animals (Stirton 1981). Kevan (1975) discussed pollination 

with respect to the use of insecticides and concluded that the elucidation of critical 

pollination phenomena, which have far-reaching ramifications on whole ecosystems, is 

clearly needed. Johnson & Steiner (2000) mentioned the urgent need for pollination 

studies to assess the viability of plant populations i.t.o. conservation. 

Baker & Hurd (1968) stressed that detailed studies of plant families should be 

complemented by intensive studies of a single species so that the mechanics of the 

pollination systems may be discovered, their operations placed on a meaningful 

quantitative basis and their modes of evolution inferred with some measure of certainty. 

For example, restrictions may be placed on the geographical distribution of plants by 

the lack of suitable pollinators and, conversely, the opportunity for anthophilous 

animals to live in an area may be limited by a lack of suitable flowers for them to visit 

(Baker & Hurd 1968). In an overview of anthecology in the Labiatae, Meeuse (1992) 

refers to the “limited number of case histories” [of pollination syndromes] in the 

Labiatae, pointing out that more case histories are required. 

Previous studies on Plectranthus pollination 

Scott Elliot (1891) reported the honeybee Apis mellifera L., 1758, a bombyliid fly and 

two lepidopterans as visitors to P. ecklonii Benth. in South Africa. Marloth (1932) did 

not record any insect visitors to two Plectranthus species from the Cape (South Africa), 

but noted that self-pollination would be unlikely as the stigma matures after the last 

anther has withered. Van der Pijl (1972) further mentioned butterflies and species of 

Bombus Latreille, 1802 and Apis L., 1758 bees as visitors to Plectranthus species in 

Nepal, Australia and Java. Gupta et al. (1984) studied the foraging activity of two Apis 

species on Plectranthus rugosus Wall. [=Isodon rugosus (Wall. ex Benth.) Codd] in India, 

and found that bumble bees and lepidopterans also visit the flowers. In Madagascar 

Pachymelus limbatus Saussure, 1890 bees (Anthophoridae) and a Stylogaster Macquart, 

1835 fly species (Conopidae) are visitors to Plectranthus vestitus Benth., with the former 

species being the principal pollinator (Nilsson et al. 1985). Pachymelus limbatus was also 

shown to exhibit male patrolling and territoriality associated with plants of Plectranthus aff. 

vestitus Benth. and P. madagascariensis (Pers.) Benth. in Madagascar (Nilsson & 

Rabakondrianina 1988). 

Stirton (1977) listed the following South African insect visitors to cultivated plants of 

Plectranthus neochilus Schltr.: Hymenoptera - five species of Megachile Latreille, 1802, 

Chapter 1/ 5

three species of Xylocopa Latreille, 1802, one species of Anthophora Latreille, 1803, 

Apis mellifera (all Apidae); Diptera - unidentified bombyliids, Asarkina Macquart, 1842 

(Syrphidae); Lepidoptera - Macroglossum trochilus Hubner, 1823 (Sphingidae). Two 

species of Xylocopa and Macroglossum trochilus also visited Plectranthus barbatus 

Andr. Only the bees were seen to consistently and effectively work the pollination 

mechanism (Stirton 1977). 

Huck (1992) reviewed pollination in the Lamiaceae and added Bombus diversus Smith, 

1869 bees (Apidae) and Gurelca himachala Butler, 1875 moths (Sphingidae) as 

pollination vectors of the Japanese species, Plectranthus inflexus Vahl ex Benth. 

In summary the documented insect visitors to Plectranthus belong to the hymenopteran 

families Apidae, including the Anthophoridae and Megachilidae – now Anthophorinae 

and Megachilinae (Brothers 1999); the dipteran families Syrphidae, Bombyliidae and 

Conopidae; and Sphingidae and other Lepidoptera. This mirrors groups of pollinators in 

the Ocimeae in general, as Paton et al. (2004) noted that these include bees, 

butterflies and flies. 

A recent study, also centred on southern African species of a genus of Lamiaceae, was 

conducted on five species of Syncolostemon E.Mey. (Ford & Johnson 2008) and overlapped 

with one of the general study sites of the current study (grassland habitat at 

Umtamvuna Gorge). It showed that a variety of pollinator groups, ranging from sunbirds 

to long-proboscid flies, day-flying hawkmoths and bees, were active on species that 

have a range of corolla tube sizes. One major difference between the genus 

Syncolostemon and Plectranthus is that the former has more or less trumpet-shaped 

flowers with more open corolla mouths, while Plectranthus tends to have narrow 

entrances to the corolla or smaller, laterally compressed entrances than 

Syncolostemon, which restricts the possible suite of pollinators, especially sunbirds. 

Rationale of the project 

This study was initiated during 1994 in KZN, for various reasons. There was little 

baseline data available on pollination in Plectranthus (Nilsson et al. 1985), especially 

from South Africa. The main (Pondoland) Centre of endemism for the genus occurs in 

this province, making it an ideal study area. There has been increased interest in the 

cultivation of this genus, with new species discoveries and artificial hybrids becoming 

Chapter 1/ 6

available in recent years. As the largest genus of Lamiaceae in southern Africa, with 

remarkable floral diversification, it makes the ideal subject for broad-based pollination 

studies. Species of Plectranthus associated with the Pondoland Centre in particular, 

show considerable diversification with respect to corolla tube length and shape, which 

suggests that pollinator syndromes are involved in the speciation process. Both 

endemic and non-endemic species were studied in the main study area, and the study 

was extended to provide a comparison of Plectranthus pollination in other sandstone 

areas further south and non-sandstone areas further north in KZN, thus extending the 

study to the greater Eastern seaboard. 

In light of discussions surrounding the ‘generalisation’ versus ‘specialisation’ debate in 

pollination systems (Waser & Price 1993, Waser et al. 1996, Johnson & Steiner 2000), 

the study aimed to describe the syndromes or groups of pollinators that may have 

driven pollinator specialisation in Plectranthus, while adding base-line observational 

information to provide case histories for eighteen species of Plectranthus and two 

species of allied genera (Pycnostachys urticifolia Hook. and Aeollanthus parvifolius 

Benth.). 

Thesis structure 

Following this introductory chapter, five chapters are in the form of published papers, 

with an additional chapter on Nectar studies followed by a Discussion & Conclusions 

chapter. The study of Plectranthus has lead to further discussions on the Lamiaceae in 

general, as well as a newly described pollination syndrome that extends beyond 

Plectranthus to other plant families and genera. 

Chapter 1: Introduction 

This current chapter provides an introduction to the genus Plectranthus and the study 

of pollination in the family Lamiaceae, and introduces the project. The introductory 

sections of the five papers presented in Chapters 2 – 6 provide further general 

background information to the study and are not repeated here. 

Chapter 1/ 7

Chapter 2: Pollination of Straight-tubed species 



Systematics and Evolution 218: 99–112. 

The pollination of seven straight-tubed Plectranthus species of varying tube length is 

described, showing that both bees (Hymenoptera) and flies (Diptera) with varying 

proboscid lengths are the main pollinator groups. 

Chapter 3: Pollination of Sigmoid-tubed species 

Potgieter, C.J., Edwards, T.J., Van Staden, J., 2009. Pollination of Plectranthus 

spp. (Lamiaceae) with sigmoid flowers in southern Africa. South African Journal 


The pollination of five sigmoid-tubed Plectranthus (and two allied species with sigmoid 

corollas) is described. Bees are the main pollinators, but fly pollination is also prevalent 

in some species. The potential origin of the sigmoid corolla shape is discussed. 

Chapter 4: Convergent pollination in southern African Lamiaceae 

Potgieter, C.J., Edwards, T.J., 2001. The occurrence of long, narrow corolla 

tubes in southern African Lamiaceae. Systematics and Geography of Plants 71: 

493–502. 

This paper uses known Plectranthus pollination data to speculate on the possible 

pollinators of other Lamiaceae in the region that have long, straight corolla tubes. 

Chapter 5: A new Pollination Guild 


Nemestrinidae) pollination guild in eastern southern Africa. Annals of the 

Missouri Botanical Garden 92: 254–267. 

The discovery of this exciting new pollination guild led to a discussion on Plectranthus 

and species of other genera and plant families that have representatives that conform 

to this long-proboscid fly guild. 

Chapter 1/ 8

Chapter 6: Natural Hybrids 



hybridisation in Plectranthus (Lamiaceae). South African Journal of Botany 72: 

99–104. 

This paper tests a microdistillation technique to detect hybridization, using a putative 

Plectranthus hybrid for the study. The discussion relates the occurrence of natural 

hybrids to pollinator fidelity in long-tubed Plectranthus. 

Chapter 7: Nectar studies 

This is a general chapter on nectar, outlining data on Plectranthus nectar sugar 

composition and comparing it with trends in other Lamiaceae and long-proboscid flypollinated 

plants. Nectar concentration and volume were studied in selected 

Plectranthus species. The relevance of nectar studies is discussed. 

Chapter 8: Discussion and Conclusions 

The final chapter brings the various chapters and published papers together. 

Appendix 

This final section provides consolidated data on the 20 studied plant species, with brief 

plant and habitat descriptions, study site and field work information, and pollinator 

observations and vouchers. The appendix should be read in conjunction with the 

published papers, since it provides the most recent data for all studied species. It also 

contains unpublished records that have not been, or are in the process of being, 

compiled for publication. 

Chapter 1/ 9

References 

Abdel-Mogib, M., Albar, H.A., Batterjee, S.M., 2002. Chemistry of the genus 

Plectranthus. Molecules 7: 271-301. 

Ascough, G.D., Rice, L.J., Van Staden, J., 2008. Considerations for evaluating flower 

abscission in potted plants with multiple inflorescences - Plectranthus as a case 

study. South African Journal of Botany 74: 753–756. 

Ascough, G.D., Van Staden, J., 2007. Preventing flower abscission in potted 

Plectranthus plants and discovery of differential regulation of dark-induced 

abscission. South African Journal of Botany 73: 277–278. 

Baker, H.G., Baker, I., 1986. The occurrence and significance of amino acids in floral 

nectar. Plant Systematics and Evolution 151: 175–186. 

Baker, H.G., Hurd, P.D. Jr., 1968. Intrafloral Ecology. Annual Review of Entomology 

13: 385–414. 

Beardsley, P.M., Yen, A., Olmstead, R.G., 2003. AFLP phylogeny of Mimulus section 

Erythranthe and the evolution of hummingbird pollination. Evolution 57: 1397– 

1410. 

Brits, G.J., Selchau, J., Van Deuren, G., 2001. Indigenous Plectranthus (Lamiaceae) 

from South Africa as new flowering pot plants. Acta Horticulturae 552: 165–170. 

Brits, G.J., Ling Li, 2008. Polyploid breeding of wild South African Plectranthus 

(Spurflowers) as new flowering pot plants. Acta Horticulturae 774: 437–442. 

Brothers, D.J., 1999. Phylogeny and evolution of wasps, ants and bees (Hymenoptera, 

Chrysidoidea, Vespoidea and Apoidea). Zoologica Scripta 28: 233–249. 

Codd, L.E., 1975. Plectranthus (Labiatae) and allied genera in Southern Africa. 

Bothalia 11: 371–442. 

Codd, L.E., 1985a. Plectranthus (Lamiaceae). Flora of Southern Africa 28(4): 137–172. 

Codd, L.E., 1985b. Plectranthus hilliardiae. The Flowering Plants of Africa 48: Plate 

1904. 

Edwards, T.J., 2005. Two new Plectranthus species (Lamiaceae) and new distribution 

records from the Pondoland Centre of Plant Endemism, South Africa. Bothalia 

35: 149–52. 

Edwards, T.J., Paton, A., Crouch, N.R., 2000. A new species of Plectranthus from 

Zimbabwe. Kew Bulletin 55: 459–464. 

Ford, C.M., Johnson, S.D., 2008. Floral traits, pollinators and breeding systems in 

Syncolostemon (Lamiaceae). Plant Systematics and Evolution 275: 257–264. 

Goldblatt, P., Manning, J.C., Bernhardt, P., 2000. Adaptive radiation of pollination 

mechanisms in Sparaxis (Iridaceae: Ixioideae). Adansonia 22: 57–70. 

Chapter 1/ 10

Goldblatt, P., Manning, J.C., Bernhardt, P., 2001. Radiation of pollination systems in 

Gladiolus (Iridaceae: Crocoideae) in southern Africa. Annals of the Missouri 


Gupta, J.K., Mishra, R.C., Kumar, J., 1984. Plectranthus as forage for Apis cerana 

indica F. and Apis mellifera L. Apidologie 15: 75–82. 

Harder, L.D., Barrett, S.C.H., 2006. Preface. In: Harder, L.D., Barrett, S.C.H (Eds.), 

Ecology and Evolution of Flowers, pp. vii-viii. Oxford University Press, Oxford. 

ISBN9780198570851. 

Huck, R., 1992. Overview of pollination biology in the Lamiaceae. In: Harley, R.M., 

Reynolds, T. (Eds.), Advances in labiate science, pp. 167–181. Royal Botanic 

Gardens, Kew. ISBN0947643524. 

Johnson, S.D., 2006. Pollinator-driven speciation in plants. In: Harder, L.D., Barrett, 

S.C.H (Eds.), Ecology and Evolution of Flowers, pp. 295–310. Oxford University 

Press, Oxford. ISBN9780198570851. 

Johnson, S.D., Edwards, T.J., Carbutt, C., Potgieter, C.J., 2002. Specialization for 

hawkmoth and long-proboscid fly pollination in Zaluzianskya section Nycterinia 

(Scrophulariaceae). Botanical Journal of the Linnean Society 138: 17–27. 

Johnson, S.D., Linder, H.P., Steiner, K.E., 1998. Phylogeny and radiation of pollination 

systems in Disa (Orchidaceae). American Journal of Botany 85: 402–411. 

Johnson, S.D., Manning, J.C., Pauw, A., 2009. Advances in the pollination biology of 

South African plants (Editorial). South African Journal of Botany 75: 625–629. 

Johnson, S.D., Steiner, K.E., 2000. Generalization versus specialization in plant 

pollination systems. Trends in Ecology and Evolution 15: 140–143. 

Kay, K.M., Voelckel, C., Yang, J.Y., Hufford, K.M., Kaska, D.D., Hodges, S.A., 2006. 

Floral characters and species diversification. In: Harder, L.D., Barrett, S.C.H 

(Eds.), Ecology and Evolution of Flowers, pp 311–325. Oxford University Press, 

Oxford. ISBN9780198570851. 

Kevan, P.G., 1975. Pollination and environmental conservation. Environmental 

Conservation 2: 293–298. 

Lukhoba, C.W., Simmonds, M.S.J., Paton, A.J., 2006. Plectranthus: a review of 

ethnobotanical uses. Journal of Ethnopharmacology 103, 1–24. 

Marloth, R., 1916–1932. The Flora of South Africa. Cape Town, Darter Brothers. 

Meeuse, A.D.J., 1992. Anthecology of the Labiatae: and armchair approach. In: Harley, 

R.M., Reynolds, T. (Eds.), Advances in labiate science, pp. 183–191. Royal 

Botanic Gardens, Kew. ISBN0947643524. 

Morgan, M.T., Schoen, D.J., 1997. The role of theory in an emerging new plant 

reproductive biology. Trends in Ecology and Evolution 12: 231–234. 

Nilsson, L.A., Jonsson, L., Rason, L., Randrianjohany, E., 1985. Pollination of 

Chapter 1/ 11

Plectranthus vestitus (Lamiaceae) by trap-lining hovering bees in Madagascar. 

Plant Systematics and Evolution 150: 223–236. 

Nilsson, L.A., Rabakonandrianina, E., 1988. Chemical signalling and monopolization of 

nectar resources by territorial Pachymelus limbatus (Hymenoptera, 

Anthophoridae) male bees in Madagascar. Journal of Zoology 215: 475–489. 

Paton, A.J., Springate, D., Suddee, S., Otieno, D., Grayer, R.J., Harley, M.M., Willis, F., 

Simmonds, M.S.J., Powell, M.P., Savolainen, V., 2004. Phylogeny and Evolution 

of Basils and Allies (Ocimeae, Labiatae) based on three Plastid DNA Regions. 

Molecular Phylogeny and Evolution 31: 277–299. 

Rabe, T., Van Staden, J., 1998. Screening of Plectranthus species for antibacterial 

activity. South African Journal of Botany 64, 62–65. 

Retief, E., 2000. Lamiaceae (Labiatae). In: Leistner, O.A. (Ed.), Seed Plants of 

Southern Africa: Strelitzia, vol. 10, pp. 323–334. 

Scott Elliot, G. F., 1891. Notes on the fertilisation of South African and Madagascar 

flowering plants. Annals of Botany 5: 330–344. 

Stavri, M., Paton, A., Skelton, B.W., Gibbons, S., 2009. Antibacterial diterpenes from 

Plectranthus ernstii. Journal of Natural Products 72: 1191–1194. 

Stirton, C.H., 1977. Broad-spectrum pollination of Plectranthus neochilus. Bothalia 12: 

229–230. 

Stirton, C.H., 1981. Petal sculpturing in papilionoid legumes. In: Polhill, R.M. & Raven, 

P.H. (Eds.), Advances in Legume Systematics. Royal Botanic Gardens, Kew. 

ISBN1842460544. 

Van der Pijl, L., 1972. Functional considerations and observations on the flowers of 

some Labiatae. Blumea 20: 93–104. 

Van Jaarsveld, E., 2006. The South African Plectranthus and the art of turning shade 

into glade. Fernwood Press, Cape Town. ISBN9781874950806. 

Van Jaarsveld, E.J., Edwards, T.J., 1991. Plectranthus reflexus. The flowering plants of 

Africa 51: Plate 2034. 

Van Jaarsveld, E.J., Edwards, T.J., 1997. Notes on Plectranthus (Lamiaceae) from 

southern Africa. Bothalia 27: 1–6. 

Van Jaarsveld, E.J., Hanky, A., 1997. Plectranthus venteri Van Jaarsveld and Hankey 

spec. nov. (Lamiaceae), a new species from the Northern Province, South Africa. 

Aloe 34: 40–41. 

Van Jaarsveld, E.J., Van Wyk, A.E., 2004. Plectranthus mzimvubuensis, a new species 

from Eastern Cape, South Africa. Bothalia 34: 30–32. 

Van Wyk, A.E., 1990. The Sandstone Regions of Natal and Pondoland: Remarkable 

Centres of Endemism. In: Heine, K. (Ed.), Palaeoecology of Africa 21: 243–257. 

Chapter 1/ 12

Van Wyk, A. E., Smith, G. F., 2001. Regions of Floristic Endemism in Southern Africa: 

A review with Emphasis on Succulents. Umdaus Press, Pretoria (Hatfield). 

ISBN1919766189. 

Van Zyl, R.L., Khan, F., Edwards, T.J., Drewes, S.E., 2008. Antiplasmodial activities of 

some abietane diterpenes from the leaves of five Plectranthus species. South 

African Journal of Science 104: 62–64. 

Waser, N.M., Price, M.V., 1993. Specialisation and generalisation in pollination 

systems. In: Veeresh, G.K., Shaanker, R.U., Ganeshaiah, K.N. (Eds.), Pollination 

in Tropics: Proceedings of the International Symposium of Pollination in tropics. 

International Union for the Study of Social Insects, Indian Chapter, Bangalore, pp 

165–169. 

Waser, N.M., Chittka, L., Price, M.V., Williams, N.M., Ollerton, J., 1996. Generalization 

in pollination systems, and why it matters. Ecology: 77: 1043–1060. 

Wilson, P., Castellanos, M.C., Wolfe, A.D., Thomson, J.D., 2006. Shifts between bee 

and bird pollination in Penstemons. In: Waser, N.M., Ollerton, J. (Eds.), Plantpollinator 

interactions: from specialization to generalization, pp 47–68. The 

University of Chicago Press, Chicago. ISBN0226874001. 

Winter, P.J.D., Van Jaarsveld, E.J., 2005. Plectranthus porcatus, a new species 

endemic to the Sekhukuneland centre of plant endemism, Limpopo Province, 

South Africa. Bothalia 35, 169–173. 

Chapter 1/ 13

CHAPTER 2: 

POLLINATION OF STRAIGHT-TUBED SPECIES 

Potgieter, C.J., Edwards, T.J., Miller, R.M., Van Staden, J., 1999. 

Pollination of seven Plectranthus spp. (Lamiaceae) in southern Natal, 

South Africa. 

Plant Systematics and Evolution 218: 99–112.

Plant Syst. Evol. 218:99-112 (1999) Plant Systematies 

and Evolution 

Pollination of seven Plectranthus spp. 

(Lamiaceae) in southern Natal, South Africa 

C. J. Potgieter j, T. J. Edwards 1, R. M. Miller 2, and J. Van Staden 1 

Department of Botany, University of Natal Pietermaritzburg, Scottsville, South Africa 

2 Department of Zoology and Entomology, University of Natal Pietermaritzburg, Scottsville, South Africa 

Received February 13, 1998 

Accepted August 28, 1998 

Abstract: The genus Plectranthus (Lamiaceae) shows 

remarkable radiation on the sandstones of southern Natal 

and northern Transkei in South Africa, where six endemic 

species occur. Two of these endemic species, P. hilliardiae 

and P. oribiensis, are included in this study, as well as P. 

reflexus, for which only limited data are available. The other 

species that were studied are P. ambiguus, P. ciliatus, P. 

ecklonii, P. madagascariensis and P. zuluensis. Four of these 

taxa, P. ambiguus, P. hilliardiae, P. reflexus and P. saccatus 

var. longitubus, have uniquely long corolla-tubes (20-30 mm) 

and this is related to pollination by nemestrinid flies of the 

genus Stenobasipteron that have proboscides of similar 

length. Other nemestrinid species of the genus Prosoeca 

have shorter proboscides and pollinate two species of 

Plectranthus with shorter corolla-tube lengths (6-15ram). 

Acrocerid flies, tabanid flies and anthophorid bees are also 

important visitors to these species. This study on the 

pollination of seven species of varying corolla-tube lengths 

shows a correlation between floral tube length and proboscis 

length of insect visitors, many of which are recorded for the 

first time as pollinators of Plectranthus. 

Key words; Lamiaceae, Plectranthus, Nemestrinidae, 

Stenobasipteron, Prosoeca, Acroceridae, Psilodera, Antho- 

phoridae, Amegilla. Pollination, long-tubed flowers, long- 

proboscid flies, Natal Group sandstone, adaptive radiation, 

speciation. 

The genus Plectranthus (Lamiaceae) is represented by 

45 species in southern Africa and shows a remarkable 

diversity in terms of floral morphology. This diversity 

is most prominent on the sandstones of southern Natal 

and northern Transkei where a total of 25 species 

Chapter 2/ 15 

© Springer-Verlag 1999 

Printed in Austria 

occur, including six endemic species-Plectranthus 

ernstii Codd, P. hilliardiae Codd, P. oertendahlii 

Th. Fries jun., P. oribiensis Codd, P. praetermissus 

Codd and P. reflexus E. J. van Jaarsveld and T. J. 

Edwards. The endemics are distinctive species with no 

obvious relationships to each other or more wide- 

spread species (Codd 1985a). The sandstones of 

Pondoland form forest refugia in which allopatric 

speciation of Plectranthus has occurred. These 

sandstone regions have been described as remarkable 

center of endemism with several species which are 

uncommon or absent from surrounding substrates 

(Van Wyk 1990). 

Plectranthus species are predominantly herba- 

ceous varying from soft herbs (P. ciliatus E. Mey. ex 

Benth.) to erect soft shrubs [P. oribiensis, P. zuluensis 

T. Cooke and P. ambiguus (H. Bol.) Codd], that are 

occasionally woody below (P. ecklonii Benth.). 

Plectranthus hilliardiae is a short, erect semi-succu- 

lent herb and P. madagascariensis (Pers.) Benth. is a 

semi-succulent herb with trailing or erect stems (Codd 

1985b). The latter is the only grassland species in the 

study, but it also grows along forest margins. The 

other six species grow in forest or forest margins, with 

P. ciIiatus restricted to moist areas. Plectranthus 

oribiensis, P. ecklonii and P. ambiguus may also be 

found in semi-shade. 

The inflorescences of Lamiaceae are typically 

decussate cymes that form verticillasters. Superim- 

posed upon this basic pattern are structural elabora- 

tions to form branched indeterminate thyrses, e.g.

100 C.J. Potgieter et al.: Pollination in PIectranthus 

P. ecklonii and P. ambiguus. Inflorescence size varies 

from 40-80 mm in P. zuluensis to 120-250 mm tall in 

P. ecklonii (Codd 1985b). 

Flower colour of the studied species varies from 

white (P. ciliatus and P. madagascariensis), to pale 

mauve (P. hilliardiae), mauve (P. oribiensis), pinkish 

purple (P. ambiguus), pale blue (P. zuluensis) and 

bluish purple (P. ecklonii). Nectar guides are present 

as purple speckles on the upper and lower corolla 

limbs of P. ciliatus and P. hilliardiae, as rows of 

mauve speckles on the upper lip of P. zuluensis and as 

a few speckles on the upper lip of P. ecklonii. Faint 

purple vertical lines occur on the upper lip of P. 

ambiguus and pale blue lines in P. madagascariensis. 

An interesting phenomenon is the occurrence of 

species of Plectranthus with long corolla-tubes in the 

general study area. Plectranthus hilliardiae and 

P. ambiguus are included in the current study, but 

P. reflexus occurs south of the study area at Port 

St. Johns, and only limited time was available for 

observations on this species. Plectranthus reflexus has 

a pale mauve corolla-tube of 28-30mm (Van 

Jaarsveld and Edwards 1991). Another species, 

P. saccatus Benth., has very variable corolla-tube lengths 

and one of the varieties that occur at Umtamvuna 

(P. saccatus var. longitubus Codd) has a tube length of 

20-26 mm (Codd 1985b). These long-tubed species of 

Plectranthus are unique within the African members 

of the genus. It is hypothesised that these species have 

coevolved with a pollinator with a similar proboscis 

length. 

The possibility of pollination by long-tongued flies 

exists for the long-tubed species of Plectranthus as 

some of their floral traits correspond to those listed by 

authors reporting this syndrome (Rebelo et al. 1985; 

Whitehead et al. 1987; Johnson and Steiner 1995, 

1997; Manning and Goldblatt 1995, 1996, 1997). The 

floral tubes of P. hilliardiae, R ambiguus and P. 

reflexus, as well as that of the long-tubed variety of P. 

saccatus, are long relative to other species in the 

genus, narrowly flattened laterally and straight. 

Flowers are zygomorphic with exserted stigmas and 

anthers and flowers are oriented horizontally. Nectar is 

hidden at the base of saccate corolla-tubes, except in 

P. ambiguus with its narrow base that raises the nectar 

level slightly. Flowers are not sweetly scented but 

inflorescences emit terpenoid-like scents from a 

variety of glandular trichomes. 

Records of pollination by long-proboscid flies of 

the families Nemestrinidae and Tabanidae are scat- 

tered through older texts such as those of Marloth 

(1916-1932) and Vogel (1954) and in recent years the 

topic has received increasing attention. Whitehead 

et al. (1987) observed that general texts on pollination 

Chapter 2/ 16 

had not given much attention to bee flies and long- 

proboscid flies and that the importance of pollination 

by flies, particularly in South Africa, was not 

recognised. Reviews of insect pollination in the Cape 

flora of South Africa, done by Whitehead et al. (1987) 

and Johnson (1992), outline the syndrome of pollina- 

tion by long-proboscid flies. Accounts of individual 

species or groups of species that rely on this syndrome 

are given by Rebelo et al. (1985) for Ericaceae; 

Goldblatt (1991) and Goldblatt et al. (1995) for 

Iridaceae; and Johnson and Johnson (1993) and 

Johnson and Steiner (1995, 1997) for Orchidaceae. 

Manning and Goldblatt (1995) mention a floral guild 

adapted to pollination by long-proboscid flies in the 

southern Cape (South Africa) and elaborate on two 

distinct guilds of long-tubed flowers specialised for 

this pollination syndrome that occur in winter rainfall 

areas along the west coast and near interior of southern 

Africa (Manning and Goldblatt 1996, 1997). 

The genus Plectranthus contains around 350 

species (Codd 1985b) that are distributed throughout 

the Old World tropics, with the majority of species 

occurring in eastern and southern Africa and Mada- 

gascar. Codd subdivides species on the subcontinent 

into six subgenera, with all of the species mentioned in 

this paper restricted to subgenus Plectranthus. Plec- 

tranthus madagascariensis falls within the section 

Coleoides, while the other species under study fall 

within the typical section (Codd 1975). 

Although the species delimitation of the South 

African species is fairly well established (Codd 1975, 

1985b), little is known about their pollination, 

especially in natural habitats (Nilsson et al. 1985). 

Marloth (1932) did not record any insect visitors to 

two Plectranthus species from the Cape (South 

Africa), but noted that self-pollination would be 

unlikely as the stigma matures after the last anther 

has withered. 

Scott Elliot (1891) reported Apis melIifera, a 

bombyliid fly and two lepidopterans as visitors to P. 

ecklonii in South Africa. Van der Pijl (1972) further 

mentions a number of butterflies and species of 

Bombus and Apis as visitors to Plectranthus species 

in Nepal, Australia and Java. Gupta et al. (1984) 

studied the foraging activity of two Apis species on R 

rugosus Wall. in India, and found that bumble bees 

and lepidopterans also visit the flowers. In Madagascar 

Pachymelus limbatus (Hymenoptera, Anthophoridae) 

and a Stylogaster species (Diptera, Conopidae) are 

visitors to P. vestitus Benth., with the former species 

being the principal pollinator (Nilsson et al. 1985). 

Pachymelus limbatus was also shown to exhibit male 

patrolling and territoriality associated with plants of 

Plectranthus aff. vest#us Benth. and P. madagascar-

C. J. Potgieter et al.: Pollination in Plectranthus 101 

iensis in Madagascar (Nilsson and Rabakondrianina 

1988). 

Stirton (1977) listed the following South African 

insect visitors to cultivated plants of /3. neochilus 

Schltr.: Hymenoptera - five species of Megaehile, three 

species of Xylocopa, one species of Anthophora, Apis 

mellifera (Apidae); Diptera - unidentified bombyliids, 

Asarkina (Syrphidae); Lepidoptera -Macroglossum 

trochilus (Sphingidae). Two species of Xylocopa 

and Macroglossum trochilus also visited Plectranthus 

barbatus Andr.. Only the bees were seen to work the 

pollination mechanism effectively every time. 

Huck (1992) reviewed pollination in the Lamia- 

ceae and added Bombus diversus (Apidae, Hymenop- 

tera) and Gurelca himachala (Sphingidae, 

Lepidoptera) as pollinators of Plectranthus inflexus 

Vahl ex Benth.. 

In summary the documented insect visitors to 

Plectranthus belong to the families Anthophoridae, 

Apidae and Megachilidae (Hymenoptera); Syrphidae, 

Bombyliidae and Conopidae (Diptera) and Sphingidae 

and other Lepidoptera. 

This study considers variation in corolla-tube 

length within seven species of Plectranthus and 

correlates these data to various insect pollinators that 

are recorded for the first time. 

Materials and methods 

Field observations and collections were made during the 

flowering seasons (December-May) of 1995, 1996 and 

1997. Voucher specimens of insects are lodged at the Natal 

Museum Pietermaritzburg and names are listed in Appendix 

1. Plant vouchers are lodged at the University of Natal 

herbarium (NU) and are listed in Appendix 2. 

Study sites. Field work was conducted at Umtamvuna 

and Oribi Gorge Nature Reserves in southern Natal, South 

Africa (Fig. 1). These sandstone gorges are separated by 

about 35 kin and four of the endemic species occur in the 

two reserves. Umtamvuna is closest to the coast with study 

sites ranging from 3 to 8 kin inland, while Oribi Gorge is 

situated about 15 km inland. Additional observations were 

done at World's View and Ferncliffe Nature Reserve in 

Pietermaritzburg, 75 kin inland and separated by 125 kin and 

160kin from Oribi Gorge and Umtamvuna respectively. 

Limited observations were made on P reflexus at Port St. 

Johns (Fig. lb). 

Species studied. The following species were studied: 

P ambiguus and P. hilliardiae at Umtamvuna, P oribiensis 

and P. zuluensis at Oribi Gorge, and P ecklonii, P ciliatus 

and P maclagascariensis at all three study sites. Plec- 

tranthus oribiensis and P hilliardiae are endemic to 

southern Natal. Plectranthus reflexus is endemic to forest 

along the Bulolwe River at Port St. Johns. 

Observations. Populations of flowering Plectranthus 

species were observed during the daytime and notes were 

made of the types of insect visitors, type of floral reward 

utilised and insect behaviour on the flowers. Voucher insects 

were netted and killed in separate ethyl acetate-containing 

vials to prevent pollen contamination. Each specimen was 

set on a pin with its proboscis extended forward. 

Length measurements. Measurements of proboscis 

length were done from the tip up to the point of attachment 

of the proboscis to the face of the insect. Corolla-tube 

lengths of the relevant Plectranthus species were measured 

from the base (at the junction to the calyx) to the mouth of 

the corolla (at the point where the upper and lower lips 

diverge). Style and filament lengths were also recorded; 

where the filaments are partially attached to the corolla the 

measurement included the length of the corolla. Both 

posterior (shorter) and anterior (longer) filaments were 

measured. These values were averaged and compared to 

proboscis length of insect visitors. 

Visitation frequency. To give an indication of the 

importance of various insect visitors, an estimate of 

visitation frequency was made by calculating the proportion 

of observed visits made by each insect species. This 

estimation was only done for species with proboscis lengths 

that fall within a range that may promote outcrossing, i.e. 

nectar and pollen robbing species were excluded. 

Pollen loads. Pollen loads of insects were examined 

under a Hitachi $570 scanning electron microscope to 

establish whether insects carried mixed pollen loads. Insects 

were examined under a dessecting microscope to establish 

where pollen grains were deposited on the insect body. 

Small pieces of double-sided tape were used to pick pollen 

off various parts of the insect body and these were placed on 

a stub, coated with gold-palladium and examined under the 

scanning electron microscope (SEM). The percentages of 

Plectranthus and foreign pollen were estimated. 

Results 

Chapter 2/ 17 

Distribution. Figures 1 and 2 show the distribution of 

the studied species, with longer-tubed species in Fig. 1 

and shorter-tubed species in Fig. 2. 

Phenology. Flowering times are indicated in Table 

1. December to April comprises the main flowering 

season, with intermittent flowering during the rest of 

the year for some species. Plectranthus reflexus 

flowers from January to March. The flowering of 

Plectranthus within the gorges is strongly seasonal, 

with marked overlaps occurring between species. 

Time of visit. Flowering populations are visited by 

insects between 8.00 and 17.00 with continuous visits 

throughout the day. 

Insect behaviour during visits. All the studied 

Plectranthus species are herkogamous and dichoga- 

tutus. Pollen is presented upon elongate filaments, and 

after a few days these curl downward and the style 

elongates and becomes receptive. While autogamy is 

avoided in this manner, geitonogamy can occur. 

Foraging behaviour of both dipteran and hymenop-

102 

2 14 ° I 

.j. 

/ 2 

.e- 

¢..,~ 

}e.-~ 

C. J. Potgieter et al.: Pollination in Plectranthus 

Fig. 1. Location of study sites and distribution of four longer-tubed species of Plectranthus. a P. hilliardiae, OG 3030C (Oribi 

Gorge), P 2930C (Pietermaritzburg), U 3130A (Umtamvuna). b P. reflexus (3129D, Port St. Johns). e P. ambiguus, tl P. 

ecklonii. Bar: 200 km 

reran visitors follows a typical pattern with insects 

moving up an inflorescence with basipetally maturing 

flowers, first depositing pollen on stigmas of female 

stage flowers and picking up pollen from male stage 

flowers before leaving the inflorescence. Thus geito- 

nogamy within the inflorescences is minimised. 

Dipteran and hymenopteran visits to individual 

flowers last between one and four seconds and result 

in sternotrobic pollen deposition. Anthophorid bees 

Chapter 2/ 18 

rest on the lower lip of the corolla while probing the 

flower for nectar, while nemestrinid and acrocerid flies 

hover in front of the flower while probing the tube 

horizontally to reach nectar at the base of the flower. 

Nemestrinid flies sometimes grasp the filaments and 

style while probing the flower. All the insects 

represented in Figs. 3 and 4 pick up pollen from 

Plectranthus ventrally on the abdomen, thorax, bases 

of the legs and wings or on the head and broad base

C. J. Potgieter et al.: Pollination in Plectranthus t03 

24 o ! 26 ° ~ 280 t ~0 ° i ~32 ° 

2 I 6 : ° : j /o 

"ti.t ~d¢ i I ) ~/I ~-" 

Fig. 2. Distribution of four shorter-tubed species of Plectranthus. a P. zuluensis, b P. madagascariensis, c P. oribiensis, d P. 

ciliatus. Bar: 200kin 

Table 1. Phenology of the seven studied species of Plectranthus with months arranged from July to June. F flowering 

July Aug Sept Oct Nov Dec Jan Feb March April May June 

P. ambiguus F F F 

P. hilIiardiae F F F F 

P. ecklonii F F F F 

P. zu/uensis F F F F F F 

P. ciliatus F F F F F 

P. oribiensis F F 

P. madagascariensis F F F F 

F 

F F 

Chapter 2/ 19

104 

a 

C 

d 

Stenobasipteron sp. (N, D) U 

Amegilla mimadvena (An, H) U 

Xylocopa hottentotta (Ap, H) U 

AIIodape pemix (Ap, H) U 

Apis mellifera (Ap, H) U 

Stenobasip~eron sp. (N, D) U 

Psilodera sp. A (Ac, D) U 

AIIodape pemix (Ap, H) U 

Stenobasipteron sp. (N, D)OG 

Psilodera sp. A (Ac, D) OG 

Stenobasipteron sp. (N, D) U 

Prosoeca sp. A (N, D) U 

Philoliche sp. (T, D) U, P 


Xylocopa hottentotta (Ap, H) U 

AIIodape pernix (Ap, H) U 

Fig. 3. Comparisons of floral tube length with proboscis 

lengths of insect visitors to the longer-tubed species of 

Plectranthus. Lines below flower indicate insect proboscis 

length in relation to floral tube length, a P. ambiguus, b P. 

hiltiardiae, c P zuluensis, d R ecklonii, N Nemestrinidae, D 

Diptera, An Anthophoridae, H Hymenoptera, Ap Apidae, 

Ac Acroceridae, T Tabanidae; U at Umtamvuna, OG at 

Oribi Gorge, P at Pietermaritzburg. Bar above flower: 5 mm 

a 

b 

C 

Chapter 2/ 20 


Psilodera sp. A (Ac, D) P, O(3 

Amegilla bothai (An, H) U 

Amegilla caelestina (An, H) OG 

Amegilla aspergina (An, H) P 

Prosoeca sp. B (N, D) P 

Prosoeca sp. C (N, D) P 


Amegilla caelestina (An, H) OG, P 

Prosoeca sp. D (N, D) OG 

Amegilla spilostoma (An, 14) U 

AIIodape pernix (Ap, H) OO 

Bombylius sp. A (B, D) OG 

Bombyiius sp. B (B, D) U 

Amegilla bothai (An, H) OG 

Amegflta caelestina (An, H) OG 

Amegilla spilostoma (An, H) OG 

Fig. 4. Comparisons of floral tube length with proboscis 

lengths of insect visitors to the short-tubed species of 

Plectranthus. Lines below flower indicate insect proboscis 

length in relation to floral tube length, a P. ciliatus, b P. 

madagascariensis, c P oribiensis. An Anthophoridae, H 

Hymenoptera, N Nemestrinidae, D Diptera, Ap Apidae, B 

Bombyliidae, Ac Acroceridae; U at Umtamvuna, OG at 

Oribi Gorge, P at Pietermaritzburg. Bar above flower: 5 mm


Table 2. Measurement values (in ram) for floral tube, filament and style lengths of seven species ofPlectranthus (n = 22, except 

P. ciliatus and P. zuluensis where n = 20 and P. oribiensis where n = 8). SD standard deviation 

Species Corolla tube Upper filament Lower filament Style 

Mean SD Range Mean SD Range Mean SD Range Mean SD Range 

P. ambiguus 28.1 4.3 20-33 32.2 5.5 22-39 35.1 6.0 23-41 33.2 3.9 26-39 

P. hilliardiae 25.7 2.7 21-29 31.2 4.1 25-38 32.7 4.2 27-39 31.9 3.6 25-39 

P. ecklonii 12.5 1.5 10-15 20.8 5.4 14-36 23.7 5.8 16-38 27.3 4.3 17-34 

P. zuluensis 12.5 0.6 12-13 13.3 0.6 12-15 18.2 0.9 16-20 16.6 1.8 15-20 

P. oribiensis 7.4 0.9 7-9 9.4 1.0 8-11 10.7 1.1 9-12 9.6 0.6 9-10 

P. ciliatus 7.1 0.7 6-8 10.5 1.2 10-13 11.9 1.1 8-12 11.0 2.3 7-14 

P. madagascariensis 5.8 0.5 4-6 9.2 1.3 6-10 10.0 1.2 8-12 9.9 1.2 8-11 

of the proboscis. This places pollen in an optimal 

position to be transferred to the receptive styles of 

flowers visited subsequently. 

Floral tube, filament and style lengths. Values 

indicating the average and range of floral tube, 

filament and style lengths are represented in Table 2. 

Plectranthus ambiguus and P. hilliardiae have corolla- 

tubes longer than 25mm and filaments and styles 

longer than 30ram. Plectranthus eeklonii and P. 

zuluensis have corolla-tubes of 12.5mm, but P 

zuluensis has shorter filaments and styles than P. 

ecklonii where the style length approaches that of the 

previous two species. In P zuluensis the upper pair of 

filaments is reduced to staminodes that protrude 

slightly from the mouth of the corolla. Plectranthus 

oribiensis and P. ciliatus have corolla-tubes longer 

than 7 mm and filaments and styles longer than 9 ram. 

Plectranthus madagascariensis has the shortest cor- 

olla-tube of the studied species, with correspondingly 

shorter filaments and styles. 

Insect proboscis lengths. Table 3 presents mea- 

surements of proboscis lengths of insect visitors, some 

of which are represented in Figs. 3 and 4. 

Insect identifications. Positive identifications 

were obtained for most of the Hymenoptera, but the 

Diptera (in particular the Nemestrinidae) proved 

problematic. The revision of South African Nemestri- 

nidae (Bezzi 1924) appears to have excluded many 

Natal specimens, making identification below generic 

level unreliable (Stuckenberg, pers, comm.). The four 

species of Prosoeca (Nemestrinidae) will be referred 

to as species A-D; those of Bombylius (Bombyliidae), 

Allobaccha (Syrphidae) and Psilodera (Acroceridae) 

as species A and B respectively for each genus. 

Insect visitors. Figures 3-7 show the range of 

dipteran and hymenopteran visitors and the similarity 

between corolla-tube and insect proboscis length for 

the seven species of Plectranthus. Observations on P. 

reflexus showed that a Stenobasipteron sp. visits this 

species, but more extensive observations may show 

that other insect species are also visitors. 

Frequency. Nemestrinids (Diptera) and anthophor- 

ids (Hymenoptera) were the principle insect visitors of 

the studied species of Plectranthus. The proportion of 

visits made by each insect species, excluding pollen and 

nectar robbers, is represented in Table 4. 

Table 3. Length measurements (in mm) of proboscis lengths of insect visitors to Plectranthus. n number of measurements, SD 

standard deviation 

Diptera Hymenoptera 

Chapter 2/ 21 

Mean proboscis SD Range Mean proboscis SD Range 

Taxa length (n) Taxa length (n) 

Psilodera sp. A 10.6 (7) 1.2 9-12 Amegilla bothai 8.6 (5) 0.4 8-9 

Bombylius sp. A 3.0 (1) 0 3 A. caelestina 8.0 (8) 0.7 7-9 

Bombylius sp. B 3.0 (1) 0 3 A. mimadvena 9.0 (4) 0 9 

Prosoeca sp. A 15.5 (2) 0.7 15-16 A. aspergina 9.5 (1) 0 9.5 

Prosoeca sp. B 9.3 (3) 1.2 8-10 A. spilostoma 7.5 (2) 0.7 7-8 

Prosoeca sp. C 9.3 (3) 1.2 8-10 Allodape pernix 3.2 (8) 0.5 2.5-4 

Prosoeca sp. D 7.0 (1) 0 7 Apis mellifera 2.7 (3) 0.3 2.5-3 

Stenobasipteron sp. 25.1 (6) 6 22-29 Xylocopa hottentotta 8.3 (2) 0.6 8-9 

Philoliche sp. 

\ 

8.7 (3) 3 8-9 Megachilidae 3.0 (1) 0 3

106 C.J. Potgieter et al.: Pollination in Plectranthus 

J 

El e 

b 

d 

J 

Chapter 2/ 22 

Fig. 5. Eight species of Diptera showing the range in body size and proboscis length of flies that visit Plectranthus spp, a-e 

Nemestrinidae: a Stenobasipteron sp., b Prosoeca sp. A, c Prosoeca sp. B, d Prosoeca sp. C, e Prosoeca sp. D; f Tabanidae: 

Philoliche sp.; g-h Acroceridae: Psilodera sp. A, note variability in proboscis length/body size ratio. Bars: 5 mm


a d 

b 

C 

I 

Pollen loads. Plectranthus pollen grains are 

radially symmetrical and 6-colpate with double- 

reticulate exine patterns. Pollen grain diameters vary 

from 19 x 29 pm (P. ciliatus) and 23 x 24 gm (P. 

madagascariensis), to 27 x 37 gm (P. hilliardiae). The 

genus is stenopalynous, which makes it difficult to 

distinguish between pollen of different species, but it 

is possible to distinguish Plectranthus pollen from 

grains of other plant species. 

Specimens of anthophorid bees and nemestrinid 

files that were caught visiting P. ambiguus, P. ecklonii, 

m 

[] 

f 

m 

Chapter 2/ 23 

107 

Fig. 6. Six species of Hymenoptera showing 

the range in body size and proboscis length 

of bees that visit Plectranthus spp. a-d 

Anthophoridae: a Amegilla caelestina, b A. 

mimadvena, e A. bothai, d A. spilostoma; 

e-f Apidae: e Apis mellifera, f Allodape 

pemix. Bars: 5 mm 

P. ciliatus P. oribiensis and P. madagascariensis were 

found to contain more than 90% Plectranthus pollen 

on their bodies. In cases where mixed pollen loads 

were found the majority of foreign pollen was 

restricted to the scopae (of female bees) or dorsally 

on the insect body (of flies). Three species of Pieridae 

and one of Lycaenidae were the observed butterfly 

visitors to P. madagascariensis, but no pollen was 

found on these specimens. The pierids were only 

abundant towards the end of the main flowering season 

(May) in 1996.


Table 4. Proportions (frequency) of visits made by different insect species to seven species of Plectranthus, excluding nectar- 

and pollen-robbing species with very short mouthparts 

Plant species Insect visitor % of visits Plant species Insect visitor % of visits 

P. ambiguus 

P. hilliardiae 

P. zuluensis 

P. ecklonii 

Stenobasipteron sp. 72 P. ciliatus Amegilla caelestina 50 

Amegilla mimadyena 14 Psilodera sp. A 33 

Xylocopa hottentotta 14 Amegilla bothai 17 

Stenobasipteron sp. 67 P. madagascariensis Amegilla caelestina 38 

Psilodera sp. A 33 Amegilla mimadvena 23 

Psilodera sp. A 67 Amegilla aspergina 8 

Stenobasipteron sp. 33 Amegilla spilostoma 8 

Stenobasipteron sp. 33 Prosoeca sp. B 8 

Amegilla mimadvena 27 Prosoeca sp. C 8 

Philoliche sp. 20 Prosoeca sp. D 8 

Xylocopa hottentotta 13 R oribiensis Amegilla caelestina 50 

Prosoeca sp. A 7 Amegilla bothai 33 

Amegilla spilostoma 17 

Table 5. Pollen collecting insect visitors with short or non- 

sucking mouthparts that are not presented in Figs. 3 and 4. U 

at Umtamvuna, OG at Oribi Gorge, P at Pietermaritzburg 

Insect species Plant species 

Diptera 

Syrphidae 

Allobaccha sp. A 

Allobaccha sp. B 

Episyrphus sp. 

Rhingia sp. 

Acroceridae 

Psilodera sp. B 

Hymenoptera 

Megachilidae (1 sp.) 

P. ambiguus (U) 

P. ecklonii (U) 

P. ambiguus (U) 

P. madagascariensis (OG) 

P. ciliatus (OG) 

P. madagascariensis (P) 

P. madagascariensis (OG) 

Nectar robbing. In three of the longer-tubed 

species (P. ambiguus, P. hilliardiae and P. ecklonii) a 

small bee, Allodape pernix, was observed to make a 

hole in the base of the corolla through which nectar is 

robbed. The proboscis of this bee is 3.2mm long. 

These bees are fairly frequent visitors to the flowers 

and the holes left by the robbers are commonly found. 

In addition, individuals of A. pernix collect pollen 

from the dehisced anthers. 

Pollen-collecting insects. Table 5 lists pollen- 

collecting Diptera and Hymenoptera with short 

mouthparts that are not represented in Figs. 3 and 4. 

The only specimen that yielded sufficient pollen to 

allow any assumptions about pollination was that of 

Bombylius sp. B, collected on P. madagascariensis 

form Umtamvuna (Fig. 4b). This specimen had 90% 

Plectranthus and 10% foreign pollen ventrally on the 

junction between the thorax and the abdomen. 

Discussion 

Chapter 2/ 24 

Our results indicate that the Nemestrinidae are 

significant pollinators of Plectranthus in Natal and 

are the only observed insects capable of exploiting the 

nectar in the long-tubed species. Proboscis lengths 

correspond well to corolla-tube lengths of the 

Plectranthus species visited. In the long-tubed species 

(P. ambiguus, P. hilliardiae and P. reflexus) the 

nemestrinid fly, Stenobasipteron sp. appears to be 

the prominent pollinator, but anthophorid bees also 

contribute to pollen transfer in P. ambiguus. No pollen 

was found on the specimen of acrocerid fly (Psilodera 

sp. A) seen on P. hilliardiae, and the medium-length 

probscis of this species would not have reached 

nectar at the base of the flower. The proboscis of 

Stenobasipteron sp. is longer than the corolla tube of 

P. ecklonii and P. zuluensis, but the extended style and 

filaments are grasped by the insect during feeding, 

enhancing sternotrobic pollen transfer. The shorter 

corolla tubes in the latter two species allow insects 

with shorter proboscides to effect pollination as well. 

Thus Psilodera sp. A accounts for the majority of 

visits to P. zuluensis (Figs. 3c, 7b) and Prosoeca sp. A 

frequently visits P. ecklonii (Fig. 3d). The tabanid fly 

and anthophorid bee species that visit P. ecklonii also 

have proboscis lengths comparable to that of the 

corolla-tube (Fig. 3d). 

The shorter-tubed species rely predominantly on 

anthophorid bees and nemestrinid flies of the genus 

Prosoeca for pollination. In P. ciliatus, P. oribiensis 

and P. madagascariensis the proboscis lengths of


I / 

Fig. 7. Comparisons of floral tube length with proboscis length of visiting dipterans for two species of Plectranthus. a P. 

hilliardiae above with Stenobasipteron sp. below, b P. zuluensis above with Psilodera sp. A below; note the close correlation 

between proboscis and floral tube length. Bars: 5 mm 

visiting bees and flies in most cases correspond well 

with that of the corolla-tube, style and filaments of the 

Plectranthus species in question (Fig. 4). Insects with 

shorter mouthparts tend to concentrate on pollen 

collection. 

Pollen loads. Voucher specimens of anthophorid 

bees and nemestrinid and other flies that yielded a 

high percentage of Plectranthus pollen show that 

although pollen transfer is achieved for Plectranthus, 

these vectors are not exclusive to the genus. A few 

observations of vectors moving between Plectranthus 

flowers and those of other genera, e.g. Lobelia spp. 

(Campanulaceae) and IsogIossa spp. (Acanthaceae), 

have been made and this is supported by the 

placement of foreign pollen in patches inaccessible 

to Plectranthus stigmas on some insect specimens. It 

is difficult to distinguish between pollen grains from 

different species of Plectranthus, but it is seldom that 

more than two species grow in close proximity. 

Differences in filament and style lengths place pollen 

in different areas on the insect body, thus isolating 

pollen from different Plectranthus species. The lack of 

pollen on the voucher specimens of Lepidoptera 

suggests that butterflies are not important polinators 

of P. madagascariensis. The Pieridae were present of 

a limited period of the flowering season and were 

seldom encountered. 

Nectar robbing. The proboscis of Allodape pernix 

is too short to reach nectar legitimately, hence the 

behaviour of robbing nectar through a hole pierced at 

the base of the corolla tube. This corresponds to 

Chapter 2/ 25 

primary nectar robbing according to the system 

suggested by Inouye (1980). Glandular trichomes are 

abundant on the calyces of Plectranthus and in other 

members of the Lamiaceae have been shown to 

secrete essential oils that have insecticidal properties 

(see Levin 1973 and Konstantopoulou et al. 1992). 

Individuals of A. pernix do not appear to be deterred 

by these oils, as it is the corolla tissue, containing 

fewer terpenoid-containing trichomes, that is pierced 

above the calyx and nectary. 

The generally small body size, short proboscis 

length and pollen-collecting habit of the insects listed 

in Table 5 make them unlikely primary pollinators of 

Plectranthus. In his account on floral larceny Inouye 

(1980) refers to pollen theft which is the collection of 

pollen directly from the anthers by small bees, without 

contacting the stigma. This is probably the case for 

these species and that of the nectar robber A. pernix, 

but the possiblility that some contribution is made to 

intra-specific pollen transfer cannot be discounted. 

Comparison of tube length with proboscis 

length. This study records several new groups of 

visitors to Plectranthus: flies of the families Nemes- 

trinidae, Acroceridae and Tabanidae, and lepidopter- 

ans of the families Lycaenidae and Pieridae. The 

occurrence of pollination by long-tongued flies 

(primarly of the family Nemestrinidae, but also one 

species of Tabanidae) is confirmed for both longer- 

and shorter-tubed species of Plectranthus. Species 

with intermediate tube length (P. ecklonii and P. 

zuluensis) allow for visits by shorter-tongued insects -


anthophorid bees, acrocerid flies or Prosoeca app. 

(Nemestrinidae) - as well as long-tongued nemestri- 

nids (Stenobasipteron sp.). Style length (27.3 mm) and 

that of the anterior pair of filaments (23.7 mm) in P. 

ecklonii (tube length 12.5 mm) approximate that of the 

average proboscis length of Stenobasipteron sp. 

(25.1 ram). 

In P. zuluensis the style (15-20mm) and anterior 

filament (16-20mm) lengths are somewhat shorter, 

but the lengths of Stenobasipteron sp. proboscides 

range from 22 mm to 29 ram, thus allowing for some 

pollen carryover via the base of the proboscis. At 

Oribi Gorge both Stenobasipteron sp. and Psilodera 

sp. A (Acroceridae) visit P. zuluensis, but the mean 

proboscis length of Psilodera sp. A (9-12 mm) fits the 

dimensions of P. zuluensis flowers more closely. This 

species visits populations of P. ciliatus at Oribi Gorge 

as well as in Pietermaritzburg and observations of its 

behaviour suggest it to be an important pollinator of a 

number of other plant species. Psilodera sp. A tends to 

be restricted to forest and forest margins, while 

Prosoeca spp. prefer open grassland and patches 

bordering forest margins. 

Evolution of long corollas. No specimens of 

Stenobasipteron sp. have been seen visiting Plec- 

tranthus in Pietermaritzburg, suggesting its replace- 

ment by shorter-tongued Prosoeca species in this 

region. One could speculate that corolla-tube length in 

the Natal species of Plectranthus has evolved in 

response to the proboscis length of visiting insects. A 

similar case was suggested by Johnson and Steiner 

(1997) for the adaptation of spur length in the Disa 

draconis (L. E) Sw. species complex, except that the 

flowers of this species are non-rewarding, thus 

discounting evolution of fly proboscis length in 

response to corolla-tube length. For Plectranthus it 

is possible that there has been a coevolutionary 

process between flower and fly. 

Through experimental testing of Darwin's hypoth- 

esis regarding the evolution of flower depth, Nilsson 

(1988) confirmed that long-tongued pollinating vec- 

tors can drive the evolution of longer floral tubes. This 

may be true in the cases of P. ambiguus (average tube 

length 28.1 mm) and P. hilliardiae (average tube 

length 25.7 mm) that are visited by individuals of 

Stenobasipteron sp. (average proboscis length 

24.1 mm), as well as for P. reflexus (tube length 28- 

30 mm). Frequency of insect visits also shows that the 

long-tubed P. hilliardiae and P. ambiguus are visited 

most often by this long-tongued fly species, while this 

species only accounts for a third of insect visits to the 

'medium-tubed' P. zuluensis and P. ecklonii. Slightly 

longer corolla tubes will encourage visiting insects to 

throust their heads deeper into the flower, thus ensuring 

Chapter 2/ 26 

pollen deposition on the insect head or body, while 

slightly shorter tubes still allow for pollen-carryover 

due to the exserted style and filaments (longer than 

30 mm for three of these species). 

Pollen placement. In contrast to the orchid 

pollinaria found attached to the proboscides of 

nemestrinids and tabanids in studies by Johnson and 

Steiner (1995, 1997), very few pollen grains were 

found on the thin parts of proboscides of the dipteran 

specimens examined in the current study. Members of 

the Lamiaceae do not shed pollen in units, thus 

individual pollen grains adhere more readily to hairy 

areas at the base of the proboscis, bases of the wings 

and legs and ventrally over the thorax and abdomen. It 

appears that, for Plectranthus, pollen placement on the 

actual body of the insect optimises the chances of it 

being transferred to the style of a subsequent species. 

Similar areas of pollen placement were found on 

the long-proboscid flies that visit Pelargonium (Ger- 

aniaceae), Geissorh&a, Hesperantha and Ixia (all 

Iridaceae) in studies by Manning and Goldblatt (1996, 

1997) and Johnson and Steiner (1997). 

Lamiaeeae. The current study provides the first 

evidence of long-tongued fly pollination in the 

Lamiaceae. The floral features of the long-tubed 

species of Plectranthus correspond to many of those 

reported for this pollination syndrome, with some 

variations on the theme. Flower colours range from 

white to purple and contrasting nectar guides are 

positioned on the upper or both corolla limbs. The 

nectar guides on P. ambiguus are only feint lines and 

in P. reflexus and the long-tubed variety of P. saccatus 

there are no contrasting nectar guides visible to the 

human eye. Nevertheless the latter two species would 

fit in with the 'Stenobasipteron' syndrome on account 

of flower colour and corolla-tube length. 

In addition, the importance of flies with 'medium 

tongue lengths' should be emphasised. The observed 

species of Prosoeca (Nemestrinidae), Philoliche 

(Tabanidae) and Psilodera (Acroceridae) do not have 

exceptionally long proboscis lengths but, together with 

anthophorid bees, play a significant role in the 

pollination of shorter-tubed Plectranthus species. It 

is hypothesised that these insects, especially the 

acrocerid flies at Oribi Gorge, are the pollinators of 

shorter-tubed species of Plectranthus that occur in the 

same area, such as P. oertendahlii (tube length 8- 

13 mm), P. ernstii (tube length 4-8 mm), P. fruticosus 

L'H~rit. (tube length 5-13 mm) and the shorter-tubed 

variety of P. saccatus Benth. (tube length 8-16mm) 

for which no insect visitors have yet been observed 

(tube length values from Codd 1985b). 

Implications. This study and other documenting 

pollination by long-tongued flies highlight the impor-


tance of conserving sufficiently large areas of natural 

vegetation in order to support pollination systems. 

Little detail is known about the life cycles of these 

flies, except that their larvalstages are parasitic on egg 

pods of orthoperans (in Nemestrinidae) and arachnids 

(in Acroceridae). Those flower-feeding species with 

adults that have proboscides that are highly modified, 

elongate and specialised for feeding on nectar from 

long-tubed flowers, such as the nemestrinids docu- 

mented in this study, require 'established complex 

biocoenoses' in order to survive (Bowden 1978). 

Bowden (1978) noted that the relations between 

flowers and Diptera merit more attention than they 

have received and that the distribution of these flies 

may be driven by adult biology where the adults are 

flower visitors. The documentation of these specia- 

lised flower-fly relationships in the Cape and Natal 

flora is addressing the need for these studies, but more 

information is needed on the life histories of the flies 

and the distribution of this syndrome. 

The authors would like to thank C. Eardley, D. Brothers, 

D. Barraclough and B. Stuckenberg for insect identifica- 

tions, B. Stuckenberg for interesting discussions, S. Johnson 

for reading the manuscript and offering helpful advise, the 

Centre for Electron Microscopy at the University of Natal 

Pietermaritzburg and R. Roth for photographic help, the 

Natal Parks Board for allowing access to and accomodation 

in their Reserves and the Foundation for Research 

Development (FRD) for funding. 

Appendix 1. List of insect visitors collected during the 

study (lodged at Natal Museum Pietermaritzburg). 

Diptera 

Acroceridae 

Psilodera-2 spp. 

Bombyliidae 

Bombylius-2 spp. 

Nemestrinidae 

Prosoeca-4 spp. 

Stenobasipteron sp. 

Syrphidae 

Allobaccha-2 spp. 

Episyrphus sp. 

Rhingia sp. 

Tabanidae 

Philoliche sp. 

Hymenoptera 

Anthophoridae 

Amegilla (Aframegilla) bothai 

A. (Aframegilla) caelestina 

A. (Aframegilla) mimadvena 

A. (Zebramegilla) aspergina 

A. (Zebramegilla) spilostoma 

Apidae 

Allodape pernix 

Apis mellifera 

Xylocopa hottentotta 

Megachilidae-1 sp. 

Lepidoptera 

Lycaenidae-1 sp. 

Pieridae-3 spp. 

Appendix 2. List of plant voucher specimens (lodged 

at NU). 

Plectranthus ambiguus Potgiester 86 

P. ciliatus Potgieter 67 


P.. ciliatus Potgieter 69 



P. ecklonii Potgieter 65 

P. eckIonii Potgieter 66 

P. ecklonii Potgieter 70 

P. eckIonii Potgieter 114 

P. hilliardiae Potgieter 110 

P. hiIliardiae Potgieter 111 

P. hilliardiae Potgieter 112 

P. madagascariensis Potgieter 89 




P. oribiensis Potgieter 102 

P. saccatus var. longitubus Potgieter 107 

P. saccatus var. longitubus Potgieter 120 

P. saccatus var. saccatus Potgieter 58 

P. saccatus vat. saccatus Potgieter 103 

P. saccatus var. saccatus Potgieter 106 



P. zuluensis Potgieter 64 

P. zuluensis Potgieter 118 

References 

Chapter 2/ 27 

Bessi M. (1924) The South African Nemestrinidae (Diptera) 

as represented in the South African Museum. Ann. S. 

African Mus. 19: 164-190. 

Bowden J. (1978) Diptera. In: Werger M. J. A. (ed.) 

Biogeography and ecology of southern Africa. Junk, The 

Hague, pp. 775-796. 

Codd L. E. (1975) Plectranthus (Labiatae) and allied genera 

in southern Africa. Bothalia 11: 371-442. 

Codd L. E. (1985a) Plectranthus hilliardiae. F1. P1. Africa 

48: Plate 1904. 

Codd L. E. (1985b) Plectranthus (Lamiaceae). F1. S. Africa 

28 (4): 137-172. 

Goldblatt P. (1991) An overview of the systematics, 

phylogeny and biology of the African Iridaceae. In: 

Linder H. P., Hall A. V. (eds.) Systematics, biology and 

evolution of some South African taxa, Contr. Bolus Herb. 

13, pp. 1-74.


Goldblatt R Manning J. C., Bernhardt R (1995) Pollination 

biology of Lapeirousia subgenus Lapeirousia (Iridaceae) 

in southern Africa: Floral divergence and adaptation for 

long-tongued fly pollination. Ann. Missouri Bot. Gard. 

82: 517-534. 

Gupta J. K., Mishra R. C., Kumar J. (1984) Plectranthus as 

forage for Apis cerana indica E and Apis mellifera L. 

Apidologie 15: 75-82. 

Huck R. (1992) Overview of pollination biology in the 

Lamiaceae. In: Harley R. M., Reynolds T. (eds.) 

Advances in labiate science. Royal Botanic Gardens, 

Kew, Richmond, pp. 167-181. 

Inouye D. W. (1980) The terminology of floral larceny. 

Ecology 61: 1251-1253. 

Johnson S. D. (1992) Plant-animal relationships. In: 

Cowling R. (ed.) The ecology of fynbos: Nutrients, fire 

and diversity. Oxford University Press, Cape Town, pp. 

175-205. 

Johnson S. D., Johnson K. (1993) Beauty and the beast: A 

Cape orchid pollinated by horseflies. Veld and Flora 

(1975+) 79: 38-39. 

Johnson S. D., Steiner K. E. (1995) Long-proboscid fly 

pollination of two orchids in the Cape Drakensberg 

mountains, South Africa. Plant Syst. Evol. 195: 169-175. 

Johnson S. D., Steiner K. E. (1997) Long-tongued fly 

pollination and evolution of floral spur length in the Disa 

draconis complex (Orchidaceae). Evolution 51: 45-53. 

Konstantopoulou I., Vassilopoulou L., Mavragani-Tsipidou 

R, Scouras Z. G. (1991) Insecticidal effects of essential 

oils. A study of the effects of essential oils extracted from 

eleven Greek aromatic plants on Drosophila auraria. 

Experientia 48: 616-619. 

Levin D. A. (1973) The role of trichomes in plant defense. 

Quart. Rev. Biol. 48: 3-15. 

Manning J. C., Goldblatt E (1995) Cupid comes in many 

guises. Veld and Flora (1975+) 81: 50-52. 

Manning J. C., Goldblatt R (1996) The Prosoeca peringueyi 

(Diptera: Nemestrinidae) pollination guild in southern 

Africa: Long-tongued flies and their tubular flowers. Ann. 

Missouri Bot. Gard. 83: 67-86. 

Manning J. C., Goldblatt P. (1997) The Moegistorhynchus 

longirostris (Diptera: Nemestrinidae) pollination guild: 

Long-tubed flowers and a specialized long-proboscid fly 

Chapter 2/ 28 

pollination system in southern Africa. Plant Syst. Evol. 

206: 51-69. 

Marloth R. (1916-1932) The flora of South Africa. Darter 

Brothers, Cape Town. 

Nilsson L. A (1988) The evolution of flowers with deep 

corolla tubes. Nature 334: 147-149. 

Nilsson L. A., Rebakondrianina E. (1988) Chemical sig- 

nalling and monopolization of nectar resources by territorial 

Pachymelus limbatus (Hymenoptera, Anthophoridae) male 

bees in Madagascar. J. Zool. 215: 475-489. 

Nilsson L. A., Jonsson L., Randrianjohany E. (1985) 

Pollination of Plectranthus vestitus (Lamiaceae) by 

trap-lining hovering bees in Madagascar. Plant Syst. 

Evol. 150: 223-236. 

Rebelo A. G., Siegfried W. R., Olivier E. G. H. (1985) 

Pollination syndromes of Erica species in the south- 

western Cape. S. African J. Bot. 51: 270-280. 

Scott Elliot G. E (1891) Notes on the fertilisation of South 

African and Madagascar flowering plants. Ann. Bot. 5: 

330-344. 

Stirton C. H. (1977) Broad-spectrum pollination of 

Plectranthus neochilus. Bothalia 12: 229-230. 

Van der Pijl L. (1972) Functional considerations and 

observations on the flowers of some Labiatae. Blumea 

20: 93-104. 

Van Jaarsveld E. J., Edwards T. J. (1991) Plectranthus 

reflexus. F1. P1. Africa 51: Plate 2034. 

Van Wyk A. E. (1990) The sandstone regions of Natal and 

Pondoland: Remarkable centres of endemism. Palaeoecol. 

Africa 21: 243-257. 

Vogel S. (1954) Bltitenbiologische Typen als Elemente der 

Sippengliederung. Bot. Stud. 1. 

Whitehead V. B., Giliomee J. H., Rebelo A. G. (1987) Insect 

pollination in the Cape flora. In: Rebelo A. G. (ed.) A 

Preliminary synthesis of pollination biology in the Cape 

flora. CSIR, Pretoria, pp. 52-82. 

Addresses of the authors: Christina J. Potgieter, T. J. 

Edwards, J. Van Staden, Department of Botany, University 

of Natal Pietermaritzburg, P/Bag X01, Scottsville, 3209, 

South Africa. R. M. Miller, Department of Zoology and 

Entomology, University of Natal, Pietermaritzburg, P/Bag 

X01, Scottsville, 3209, South Africa.

CHAPTER 3: 

POLLINATION OF SIGMOID-TUBED SPECIES 

Potgieter, C.J., Edwards, T.J., Van Staden, J., 2009. 

Pollination of Plectranthus spp. (Lamiaceae) with sigmoid flowers in 

southern Africa. 

South African Journal of Botany 75: 646–659.

Pollination of Plectranthus spp. (Lamiaceae) with sigmoid flowers in 

southern Africa 

C.J. Potgieter a,⁎ , T.J. Edwards a , J. Van Staden b 

a School of Biological and Conservation Sciences, University of KwaZulu-Natal Pietermaritzburg, Private Bag X01, Scottsville 3209, South Africa 

b Research Centre for Plant Growth and Development, School of Biological and Conservation Sciences, University of KwaZulu-Natal Pietermaritzburg, 

Private Bag X01, Scottsville 3209, South Africa 

Abstract 

Received 31 March 2009; received in revised form 6 July 2009; accepted 14 July 2009 

Within the South African Plectranthus species two specialized lines of corolla adaptations have evolved. Long-proboscid flies (Nemestrinidae) 

appear to have driven the development of Plectranthus species with long corolla tubes that are limited to the sub-continent. Plectranthus s.l. 

(including Coleus) species with sigmoid corollas are far more widespread and evidence presented here supports the hypothesis that this floral type 

has evolved as a response to melittophily. Thirty percent of southern African Plectranthus species have corolla tubes that are bent to some degree. 

Pollination of the following four labiate species with sigmoid corollas was studied in detail: Plectranthus petiolaris, P. laxiflorus, P. calycinus 

and Pycnostachys urticifolia. The pollination of three other species was investigated to a lesser degree: Plectranthus spicatus, P. rehmannii and 

Aeollanthus parvifolius. Bee pollination is confirmed for P. laxiflorus and Py. urticifolia and is recorded here for the first time in P. petiolaris and 

P. calycinus. A new group of floral visitors comprising nemestrinid flies of the genus Prosoeca with moderately long proboscids is recorded for 

P. laxiflorus and P. calycinus, where the corolla tube shape allows visits by medium-proboscid floral visitors. The sigmoid corolla shape limits the 

type and size of insects that can access nectar and act as pollinators. Explanations for the existence and function of the sigmoid corolla shape are 

suggested. 

© 2009 SAAB. Published by Elsevier B.V. All rights reserved. 

Keywords: Apidae; Anthophorinae; Lamiaceae; Nemestrinidae; Plectranthus; Pollination; Sigmoid corolla; Southern Africa 

1. Introduction 

Plectranthus (Lamiaceae), a member of subfamily Nepetoideae, 

tribe Ocimeae, subtribe Plectranthinae (Paton et al., 2004), 

comprises ±300 species that occur in the tropical and warm 

regions of the Old World (Retief, 2000). It is the largest genus of 

the Lamiaceae in southern Africa, represented by ca. 53 species 

(Codd, 1975, 1985; Van Jaarsveld and Edwards, 1991, 1997; 

Van Jaarsveld and Hankey, 1997; Edwards et al., 2000; Van 

Jaarsveld and Van Wyk, 2004; Edwards, 2005; Winter and Van 

Jaarsveld, 2005). Recent evidence (Paton et al., 2004) suggests 

that the genus is paraphyletic as currently circumscribed. 

⁎ Corresponding author. 

E-mail address: potgietercj@ukzn.ac.za (C.J. Potgieter). 

Available online at www.sciencedirect.com 

South African Journal of Botany 75 (2009) 646–659 

0254-6299/$ - see front matter © 2009 SAAB. Published by Elsevier B.V. All rights reserved. 

doi:10.1016/j.sajb.2009.07.009 

Chapter 3/ 30 

www.elsevier.com/locate/sajb 

In an overview of pollination biology in the Lamiaceae, 

Huck (1992) commented on the large gaps that exist in our 

knowledge and recommended that pollination studies focus on 

species in situ. In view of the large size of the genus we 

identified it as an ideal group for pollination studies. The 

pollination of long-tubed Plectranthus species by longproboscid 

flies was first reported by Potgieter et al. (1999), 

along with the pollination of a number of medium- and shorttubed 

species of the genus by bees and flies with medium- to 

short proboscids. The discussion surrounding long-tubed 

species of Plectranthus was extended to the rest of the 

Lamiaceae in southern Africa by relating the distribution of 

long-tubed Lamiaceae to the biogeography of long-proboscid 

flies (Potgieter and Edwards, 2001). Pollination of Lamiaceae 

with corollas of intermediate length, by pollinators with intermediate 

length proboscids, may have predisposed the group to

pollination by long-proboscid flies, with subsequent extension 

of corolla tubes. Elongation of corolla tubes leads to increased 

protection of nectar resources, which leads to increased 

pollinator fidelity (Potgieter and Edwards, 2001). Publication 

of the Stenobasipteron wiedemanni (Diptera, Nemestrinidae) 

Pollination Guild (Potgieter and Edwards, 2005) established the 

multi-family plant guild to which long-tubed species of Plectranthus 

(with tube lengths of 20–33 mm) belong. 

Both the specialized long-proboscid fly pollinated and more 

generalized shorter-proboscid fly and bee pollination guilds that 

have been studied within southern African Plectranthus 

(Potgieter et al., 1999; Potgieter and Edwards, 2001, 2005) 

involve species with relatively straight corolla tubes. A new 

pollination guild is described here for species with sigmoidshaped 

corollas. 

Initial phylogenetic work by Paton et al. (2004),supplemented 

by Lukhoba et al. (2006), creates a framework for phylogenetic 

interpretation of Plectranthus and its allies. Subsequent work 

shows that there are three main clades in the genus Plectranthus, 

but the relationships between them are unresolved (A. Paton, pers. 

comm.). They are: 1) a sigmoid ‘Coleus’ clade, including P. 

rehmannii Gürke and P. calycinus Benth., possibly P. spicatus E. 

Mey. ex Benth,. and Pycnostachys urticifolia Hook.; 2) a sigmoid 

Plectranthus clade, including P. petiolaris E.Mey. ex Benth. and 

P. laxiflorus Benth.; and 3) a straight Plectranthus clade. The 

genus Aeollanthus Mart. ex K.Spreng is placed near the base of 

the Plectranthinae (A. Paton, pers. comm.). The definition of 

sigmoid corollas by Codd (1985) included corollas with varying 

degrees of geniculation of the tube, and for the purpose of this 

paper the term ‘sigmoid’ is used in a loose sense. Within southern 

African Plectranthus 30% of species have evolved short to 

medium sigmoid corollas similar to those in Solenostemon 

Thonn., Aeollanthus and Pycnostachys Hook. These three genera 

all fall within the broader generic circumscription of Plectranthus 

proposed by Paton et al. (2004). 

The bulk of this paper deals with four species that were studied 

in detail: P. petiolaris and P. laxiflorus, with sigmoid tubes bent 

to a similar degree; P. calycinus Benth. [=Rabdosiella calycina 

(Benth.) Codd], with weakly bent corollas, and Pycnostachys 

urticifolia Hook., with distinctly bent corollas. Three other species 

were studied in less detail, but are recorded and discussed since 

they fall within the same syndrome: P. spicatus (with distinctly 

bent corollas), P. rehmannii (with weakly bent corollas), and 

Aeollanthus parvifolius Benth. (with curved corolla tubes). 

During the course of the study it was noted that a site in 

Pietermaritzburg with extensive stands of P. laxiflorus showed 

a temporal separation of pollinator classes during each flowering 

season. At first only bee pollinators were present, but 

nemestrinid flies emerged en masse at the end of March/early 

April each year, with bees and flies actively visiting flowers for 

the rest of the season. An investigation on the effectiveness of 

the two pollinator types was included in the study. 

2. Literature on sigmoid Lamiaceae 

In a discussion on corolla adaptations in the Lamiaceae, 

Meeuse (1992) noted two important aspects: the size of specialized 

C.J. Potgieter et al. / South African Journal of Botany 75 (2009) 646–659 

Chapter 3/ 31 

corollas and the resistance offered to insect visitors attempting to 

access nectar. The upper corolla lip of Plectranthus is erect and 

does not hamper access, whereas the lower lip allows for easy 

landing, thus some kind of barrier in the tube would function as a 

selective device (Meeuse, 1992). Van der Pijl (1972) considered 

why some Plectranthus species, as a transition to Coleus, should 

have ‘geniculate’ corolla tubes. One suggestion was that the bend 

may be a mechanical necessity for a horizontal flag-type blossom 

(with an upright upper limb) to combine with the long vertical tube 

that already exists in this group (Van der Pijl, 1972). Another is 

that sigmoid corollas provide an effective shift from butterfly to 

bee pollination without major changes in tube dimensions (Van 

der Pijl, 1972), possibly since the sigmoid shape may allow bees 

better access to nectar than in a long-tubed, straight corolla. 

There are no published records on the pollination of P. 

petiolaris and only one for P. laxiflorus (Scott Elliot, 1891). In 

P. calycinus the sides of the lower lip are bent upwards, giving it 

a boat-like appearance, enclosing the stamens and style. As an 

insect visitor forces the lower lip down the stamens are exposed, 

dusting pollen onto the insect. In both P. calycinus and P. 

laxiflorus this action requires considerable force from the insect 

(Scott Elliot, 1891) and the corolla shape of P. petiolaris suggests 

that a similar system operates in this species. Van der Pijl (1972) 

mentioned the bee-blossoms of P. laxiflorus and tried to explain 

their sigmoid corolla shape. It was first suggested that the bend in 

the corolla tube may have functioned in excluding bee visitors 

while fitting a lepidopteran proboscis, yet the fused filaments and 

hinged carina suggests bee visitors, or a regression to melittophily 

(bee pollination). Vogel (1954) listed Plectranthus under 

melittophily and sphingophily (moth pollination), while Pycnostachys 

was placed under melittophily. Percival (1965) discussed 

how bees with abdominal brushes (the Dasygastrae) exploit 

sternotrobic flowers, such as Py. urticifolia, by collecting pollen 

ventrally on the abdomen. She mentioned that the open corolla 

mouth of the labiates did not hinder bees from probing, but that 

visitors were stratified according to tube length. 

Stirton (1977) described the insect visitors to cultivated 

plants of the South African species Plectranthus neochilus 

Schltr. The corolla of this species has a narrow tube (15–18 mm 

long) that ascends, then bends knee-like and expands about the 

middle (Codd, 1985). Although the corolla tube is not bent to 

the same extent as found in P. petiolaris and P. laxiflorus, P. 

neochilus makes for an interesting comparison. Stirton (1977) 

found five species of Megachile, three Xylocopa species, one 

species of Anthophora (now genus Amegilla) and Apis mellifera 

(all Hymenoptera, Apidae) to be effective pollinators. The bees 

landed on the boat-shaped lower lip, depressed it and exposed 

the stigma and stamens which transferred pollen ventrally onto 

the insects (Stirton, 1977). This study also listed unidentified 

bombyliids (Diptera: Bombyliidae), one syrphid species 

(Diptera: Syrphidae) and a sphingid moth, Macroglossum 

trochilus (Lepidoptera: Sphingidae), as ineffective visitors. 

Paton et al. (2004) noted that the sigmoid tubes of the Coleus 

clade and the sigmoid Plectranthus clade (P. laxiflorus and P. 

petiolaris) always combine with a horizontal lower (anterior) 

corolla lobe, a combination which they believed would favour 

landing insects with flexible proboscids. 

647

648 C.J. Potgieter et al. / South African Journal of Botany 75 (2009) 646–659 

3. Materials and methods 

All seven plant species were studied in KwaZulu-Natal 

(KZN), with some work done in the Eastern Cape Province of 

South Africa (Fig. 1a). Field work was conducted in various 

localities from 1995 to 2009 (see Appendix 1 for study site, year 

Chapter 3/ 32 

of study and voucher details). Plant species were identified using 

Codd (1985), as well as the texts listed in the introduction, for 

species described subsequent to Codd's (1985) revision. 

Plant distributions were compiled from flora accounts (Codd, 

1985), herbarium records (NU, NH and PRE) and field 

observations. Pollinator observations were made across a number 

Fig. 1. Distribution maps for study species and field study sites: (a) Study sites in eastern South Africa; (b) distribution of P. petiolaris (solid circles) and P. calycinus 

(open circles); (c) distribution of P. laxiflorus (solid circles) and Py. urticifolia (open circles); (d) distribution of P. spicatus (closed circles), A. parvifolius (open 

circles) and P. rehmanii (open squares). S = Stutterheim–Kologha Forest, U = Umtamvuna N.R., O = Oribi Gorge N.R., D = Dargle, P = Pietermaritzburg, K = 

Karkloof–Leopards' Bush N.R., N = Ngoye Forest. Bar: 200 km.

of flowering seasons, between 7.00am and 6.00pm. Insects that 

visited study flowers were netted and asphyxiated in separate 

glass pill vials to prevent pollen transfer between specimens. Vials 

were prepared by compressing absorbent paper into the base of 

each vial and saturating it with ethyl-acetate. Bees and flies were 

pinned with proboscids extended forward. 

Insect voucher specimens are lodged with the Natal Museum 

in Pietermaritzburg (Diptera) and with the Biosystematics 

Division, Plant Protection Research Institute, Pretoria (Hymenoptera). 

Hymenoptera were classified according to Brothers 

(1999). Plant vouchers are housed at the Bews Herbarium, 

University of KwaZulu-Natal (NU) (Appendix 1). 

Areas of pollen deposition on insects were determined using 

a dissecting microscope. Initially pollen samples were removed 

with double-sided tape for Scanning Electron Microscopy 

(SEM). Stubs were coated with gold-palladium and examined 

under a Hitachi S570 scanning electron microscope at an 

accelerating voltage of 10 kV. Pollen grains of Plectranthus 

and Pycnostachys species are morphologically similar, with 6colpate 

grains and reticulate exine patterns, and can be distinguished 

from pollen grains of other plant families. 

Percentages of Plectranthus to total pollen loads were 

estimated. For Pycnostachys, representative voucher insects 

were killed and examined, but subsequent specimens were 

caught and cooled to allow handling. Tiny cubes of Fuchsin gel 

were used to remove pollen from different areas on the insect 

body, placed on separate slides and gently heated to make semipermanent 

mounts. Insects were marked with tiny drops of 

white correction fluid on the thorax to avoid re-examination if 

captured subsequently. Slides were examined with a compound 

light microscope to determine the percentage Pycnostachys 

pollen present. The Fuchsin gel technique was followed in later 

years. 

Proboscis length measurements were made from the tip of 

the proboscis to the point where the proboscis attaches to the 

head of the insect. In the case of bees this measurement was 

divided into the length of the solid base of the proboscis (galea) 

including the clypeus, and the protruding flexible part (glossa). 

Corolla tube lengths were measured using a fine piece of 

wire bent in the shape of the sigmoid corolla, extending from 

the corolla base to the point where the upper and lower corolla 

lips diverge. The wire was then straightened and measured. 

Filament and style measurements include the entire, functional 

length of the corolla, although filaments attach to the corolla for 

some distance. Measurements were made from fresh, preserved 

and pressed plant specimens. 

Nectar levels were recorded in P. petiolaris by comparing 

the height of the nectar column from the base of the corolla, 

with the distance from the base to the bend in the corolla. This 

was measured in unvisited flowers, early in the morning (at 

9.00am) and again at 2.00pm (after many bee visits on a sunny 

day). The ratio of nectar column to ‘base-to-bend’ distance was 

compared between the morning and the afternoon. 

To gauge the relative efficiency of apinid bees and 

nemestrinid flies in pollinating P. laxiflorus at Ferncliff Nature 

Reserve (NR), Pietermaritzburg, fruit set was recorded for two 

intervals: before (bees only) and after fly emergence (bees and 


flies). As soon as flies emerged in 2003, the position of the most 

recently opened verticil of flowers was marked on each of 16 

inflorescences. At the end of the flowering season the resulting 

infructescences were collected. To ensure a suitable time lapse 

between observed and actual fly emergence, fruit set was only 

recorded from verticils three rows below the mark on each 

inflorescence (to represent ‘bee only’ visitation) and then from 

the verticil above the mark (to represent ‘bee and fly’ visitation). 

Fertilisation rate was calculated by comparing actual fruit set 

(no. of swollen calyces) with potential fruit set (no. of pedicels 

with and without calyces, representing no. of flowers). 

By accessing unpublished data on other straight-tubed species 

of Plectranthus, combined with data on sigmoid species, twenty 

species were first grouped by corolla shape and length and then 

graded by reliance on shaded (forest) through to sunny habitat. 

The proportion of visits received by effective fly pollinators versus 

bee pollinators was estimated for each species, using pollinator 

observation and voucher data. Flies included the families 

Nemestrinidae, Tabanidae and Acroceridae; bees included the 

sub-families Apinae and Megachilinae of the family Apidae. 

4. Results 

Chapter 3/ 33 

The study species are more or less widely distributed along 

the eastern seaboard of southern Africa, with more than one 

species often co-occurring and one species (P. rehmannii) 

endemic to the KZN Midlands (Fig. 1b–d; Appendix 1). The 

habit, floral characteristics and habitat varies between species, 

but all flower in late summer to autumn (see Table 1). 

Two floral subtypes were represented (see Table 2): flowers up 

to 11 mm long, with narrow corolla bases: P. laxiflorus (Figs. 2, 

7c), P. petiolaris (Figs. 3, 7a) and Py. urticifolia (Figs. 5, 7b); and 

smaller flowers less than 7 mm long, with saccate bases: P. 

calycinus (Fig. 4). In Py. urticifolia the filaments are fused for a 

few millimetres beyond the corolla tube, enclosing the style in a 

rigid sheath that provides mechanical support once the bent 

filaments and style elongate and straighten after anthesis (Fig. 7b). 

Sigmoid Plectranthus flowers have tubular corollas and are 

zygomorphic, with the stigma and four anthers enclosed by the 

lower lip (Fig. 7a–d). Flowers are protandrous, with anther 

dehiscence in the male phase followed by an extension of the 

style, bringing the stigma upwards into a position previously 

occupied by the anthers. The style is smooth and generally pinlike 

in appearance, with closely appressed bifid stigmatic lobes 

which open to reveal inner receptive surfaces during the female 

phase (Fig. 7c). The flowers are herkogamous, with the 

filaments and anthers dropping down into the boat-shaped 

lower lip of the flower during the female phase. In some of the 

non-sigmoid Plectranthus species with long filaments (e.g. P. 

ecklonii Benth.) the anthers and filaments curl away sideways 

after dehiscence (Potgieter et al., 1999). 

The upright upper limbs of the bilabiate flowers are held 

vertically in all studied species and function in advertising 

nectar, which is secreted by a nectariferous disc around the 

ovary at the base of the corolla; nectar guides, when present, are 

most often located on the inner or adaxial surface of the upper 

corolla limb. 

649


Chapter 3/ 34 

Fig. 3. Comparisons of floral tube shape and length with proboscids of insect 

visitors to Plectranthus petiolaris. Single lines show flexible parts of proboscis, 

double lines show solid parts. H = Hymenoptera, A = Apidae, Ap = Apinae. Bar: 

2.5 mm. 


visitors to Plectranthus laxiflorus. In bees the single lines show the proximal 

flexible part of a proboscis (glossa) that extends along the whole length; double 

lines show the rigid basal part (clypeus and galea). D = Diptera, N = Nemestrinidae, 

H = Hymenoptera, A = Apidae, Ap = Apinae, Me = Megachilinae. Bar: 2.5 mm.


visitors to Plectranthus calycinus. Single lines show flexible parts of proboscis, 

double lines show solid parts. D = Diptera, N = Nemestrinidae, H = Hymenoptera, 

A = Apidae, Ap = Apinae. Bar: 2.5 mm. 

Nectar levels are generally confined to below the bend of the 

corolla. The base of the corolla tube is held at a near vertical or 

oblique angle upwards before the tube bends downwards; this 

presumably acts gravitationally to retain nectar. The nectar level 

fluctuation that was measured in P. petiolaris (by comparing 

the height of the nectar column from the base of the flower with 

the ‘base-to-bend’ corolla tube length), showed that unvisited 

flowers had an average nectar to tube bend ratio of 0.88 (SD 

1.00, n=27) and visited flowers had an average ratio of 0.41 

(SD 1.26, n=31). Thus in the morning newly secreted nectar 

was pushed up closer to the bend in the corolla tube, but later in 

the day it became more difficult for insects with shorter 

proboscids to extract nectar, as the nectar level dropped due to 

repeated insect visitation. 

Flower colour varies from predominantly white (P. laxiflorus, 

P. calycinus) and creamy-white (P. rehmannii), to pale 

pink (Aeollanthus parvifolius), pink (P. petiolaris Oribi Gorge), 

deep purple (P. petiolaris Umtamvuna, P. spicatus) and deep 

blue (Py. urticifolia) (Table 1). In sigmoid species the darker 

coloured species tend to be pollinated by bees only, while the 

paler, whitish species appear to favour bee and fly pollinators. 

This pattern does not, however, extend to the rest of the genus 

when data from 20 species with a range of corolla shape is 

compared (C. Potgieter, unpubl. data). 

Insect identifications were obtained for most bee species, 

excepting the megachilinid and halictinid bees, where generic 


Table 1 

Floral, plant and habitat characteristics of the four main study species. 

Species Habit Habitat Flowering 

time 

Flower colour 

P. petiolaris Branched herb Scree below 

cliffs 

covered by 

scarp forest; 

forest 

margins 

P. laxiflorus Freely branching 

soft shrub or 

herb, up to 1.5 m 

tall 

P. calycinus Erect, branched 

shrub with 

slightly woody 

annual stems 

from woody 

rootstocks, stems 

up to 1.5 m tall 

Py. urticifolia Erect herb or 

shrub with a 

woody base, 

1–2.5 m tall 

Forest 

margins; 

damp open 

vegetation 

December to 

April 

Mid- 

February to 

April 

(sporadic in 

October and 

Deep purple in 

northern KZN 

and 

Umtamvuna 

NR; pink at 

Oribi Gorge NR 

White, with 4–5 

thin purple 

linear nectar 

guides on flaglike 

upper lip 

Grasslands 

November) 

January to Creamy-white, 

May tinged with 

(sporadically mauve on edges 

to July) of the corolla 

limbs 

Moist areas, April to June Deep blue 

such as 

forest 

margins and 

grassy stream 

banks 

identities are provided (Table 3; Appendix 2). Few KwaZulu- 

Natal specimens were included in the revision of nemestrinid 

flies done by Bezzi (1924), but subsequent work done by 

Barraclough (2006) allowed him to identify most Prosoeca 

(Diptera, Nemestrinidae) specimens to species-level for this 

study. Some species are new collections awaiting description 

(D. Barraclough, pers. com.). 

The main daily period of activity of insect visitors was 

between 9.00 am and 4.00pm. Apinid bees of the genera Amegilla 

and Xylocopa are the main pollinators of Plectranthus 

petiolaris (Fig. 3; Table 4). Bees of the genera Amegilla and 

Xylocopa, and pollen-collecting megachilinid bees of the genera 

Megachile and Chalicodoma, pollinate Py. urticifolia (Fig. 5; 

Table 4). Plectranthus laxiflorus is pollinated by species of 

Chalicodoma (Megachilinae), Amegilla and Xylocopa (both 

Apinae), as well as nemestrinid flies of the genus Prosoeca, of 

Table 2 

Mean floral measurements (in mm) for the four main studied sigmoid species: 

tube length measured from base of corolla to junction with lower lip; filament 

and style measurements include the full length of tube; SD standard deviation 

(in brackets after mean); n sample size. 

Species (n) Corolla tube 

(SD) 

Style 

(SD) 

Chapter 3/ 35 

Upper filament 

(SD) 

Lower filament 

(SD) 

P. petiolaris (22) 10.9 (0.7) 15.1 (4.2) 14.3 (0.6) 15.3 (2.3) 

P. laxiflorus (18) 10.5 (0.9) 16.9 (1.8) 15.9 (1.3) 17.4 (1.4) 

P. calycinus (29) 6.6 (0.6) 11.7 (1.1) 10.5 (0.8) 11.6 (0.9) 

Py. urticifolia (26) 11.0 (0.7) 19.3 (1.0) 17.2 (1.1) 18.6 (1.0) 

651


Table 3 

Proboscis (and components of proboscis) length measurements of bee and fly 

visitors to all of the seven studied species. 

Visitor Average proboscis length (mm) N 

Solid base 

(SD) 

Flexible tip 

(SD) 

Total length 

(SD) 

Prosoeca umbrosa 10.5 (0.8) 10 

Prosoeca circumdata 9 1 

Prosoeca sp. nov. 5 10 1 

Amegilla mimadvena 5.1 (0.4) 3.9 (0.4) 9 (0.3) 6 

Amegilla caelestina 4.3 (0.3) 4.2 (0.5) 8.5 (0.5) 12 

Amegilla bothai 4.8 (0.3) 3.7 (0.3) 8.6 (0.4) 11 

Amegilla fallax 4 3 7 1 

Xylocopa hottentotta 4.2 (0.2) 3.3 (0.6) 7.5 (0.5) 3 

Xylocopa scioensis 3.5 3.0 6.5 1 

Xylocopa flavorufa 3.5 3.0 6.5 1 

Xylocopa flavicollis 3.3 (0.4) 3.3 (0.4) 6.5 (0.7) 2 

Thyreus sp. 4.3 (0.4) 2.0 (0) 6.3 (0.4) 2 

Chalicodoma sp. A 3.0 (0) 2.8 (0.4) 5.8 (0.4) 2 

Chalicodoma sp. B 2.0 (0.4) 2.8 (0.3) 4.8 (0.5) 4 

Apis mellifera 3.3 (0.3) 5 

Allodape pernix 3.2 (0.5) 8 

Megachile sp. A 3 1 

Megachile sp. B 3 1 

Pseudoanthidium truncatum 2.8 (0.8) 2 

SD standard deviation (in brackets after measurement); n sample size. 

which P. umbrosa is the most abundant (Fig. 2; Table 4). 

Prosoeca umbrosa also pollinates P. calycinus at the Dargle, 

but is replaced by the apinid bee Xylocopa scioensis at 

Umtamvuna NR (Fig. 4; Table 4). 

In all cases these insects were the most abundant visitors to the 

plants under study and all populations under study showed high 

levels of fruit set, as evidenced by the retention of swollen calyces. 

Under greenhouse conditions, where bees and flies are excluded, 

fruit set in the study species was found to be negligible, confirming 

that insects are necessary for pollination. Hand-pollination 

experiments done in a few non-sigmoid species of Plectranthus 

indicate that most species can set fruit from geitonogamous pollen 

transfer, but a few species do not (C. Potgieter, unpubl. data). 

Observations made on a single clone of the sigmoid P. petiolaris, 

grown in a garden, show that geitonogamous fruit set is possible. 

In the four species that were studied in greater detail the edges 

of the lower corolla lip are folded inwards to partially conceal the 

anthers of unvisited flowers until a suitable insect visits the 

flower. Bees and flies with flexible proboscids of appropriate 

length can access nectar at the base of the corolla and, in the case 

of bees, the lengths of the galea and glossa in relation to the 

positioning of the corolla bend within the tube, also determine 

whether nectar can be reached (Table 3; Figs. 2–5). Before 

landing, bees swing their proboscids forward (the galea hinge 

does not bend beyond linear alignment with the bee's body). 

After landing on the lower lip the insect must angle its proboscis 

upwards into the corolla tube to access nectar at the base of the 

declined flower and since the proboscis is locked in linear 

alignment, the insect is forced to lower its body to raise the head 

and proboscis upwards. This action results in forced contact 

between the ventral surface of the insect's body and the anthers 

Table 4 

Pollen placement and pollen loads (% Plectranthus/Pycnostachys pollen) on 

effective insect visitors of the four main study species. 

Labiate species: with 

insect visitor 

Chapter 3/ 36 

Pollen placement (ventrally) Pollen 

load (%) 

P. petiolaris 

Amegilla mimadvena Thorax, abdomen. 25–95– 

100 

Scopae. 50 

Amegilla caelestina Thorax, abdomen. 100 

Amegilla bothai Thorax: between leg bases. 50–75 

Xylocopa hottentotta Thorax: between leg bases. 50 

P. laxiflorus 

Amegilla mimadvena Base of proboscis, head/thorax junction, 

abdomen, hind legs, scopae. 

100 

Thorax: between leg bases. 95 

Amegilla caelestina Thorax, abdomen. 100 

Amegilla bothai Thorax, abdomen, scopae. 75–90 

Prosoeca circumdata Base of proboscis, head/thorax junction, 

thorax: leg bases, abdomen. 

100 

Prosoeca umbrosa Head/thorax junction, thorax: leg bases 100 

Prosoeca sp. nov. 5 Head/thorax junction, thorax, abdomen 100 

Chalicodoma sp. A Thorax, abdomen. 60–10 

Xylocopa flavicollis Head/thorax junction, thorax: between leg 

bases. 

90 

Scopae. 75 

P. calycinus 

Prosoeca umbrosa Base of head, head/thorax junction, thorax, 

abdomen. 

Py. urticifolia 

Amegilla mimadvena Thorax, abdomen. 90 

Apis mellifera Thorax, abdomen, scopae. 100 

Xylocopa scioensis Head/thorax junction, base of proboscis. 100 

Thorax, abdomen. 95 

Xylocopa flavorufa Proboscis, head. 100 

Thorax, abdomen, hind legs. 75–100 

Thorax: bases of hind legs. 90 

Megachile sp. A Abdomen, hind legs 95–100 

Megachile sp. B Proboscis, base of head, thorax, thorax: leg 

bases, abdomen, hind legs 

100 

Chalicodoma sp. B Abdomen 75 (few 

grains) 

and style that are concealed in the lower lip, which facilitates 

pollen transfer. The lower lip does not return to its original 

position, and the anthers or stigma thus remain exposed for 

future visits. 

Bees are covered with setae (hairs), especially along the 

groove between the legs on the ventral surface; these hairs 

generally point towards the posterior of the insect (Fig. 7g). As 

an insect arrives at the flower and lands on the lower lip (or 

contacts the sexual organs) the bifid stigma of a female phase 

flower dislodges pollen from between the hairs of the sternum 

onto the stigmatic surface. Pollen can only be picked up by the 

stigma as the insect moves into the flower and not as it retreats. 

Upon retreat from a male phase flower, pollen is passively 

loaded onto the hairs of the sternum, since the dehisced anthers 

brush close to the insect and force pollen to lodge between the 

hairs. 

100


visitors to Pycnostachys urticifolia. Single lines show flexible parts of proboscis, 

double lines show solid parts. H = Hymenoptera, A = Apidae, Ap = Apinae, Me = 

Megachilinae. Bar: 2.5 mm. 

In all cases the pollinators of sigmoid species picked up and 

transported pollen on the ventral parts of their bodies (Table 4), 

with the thoracic area between the leg bases and the hairy area 

below the head being good sites for pollen carryover (Fig. 7g–h). 


In the case of P. calycinus studied at Umtamvuna NR, no 

vouchers of Xylocopa caffra were caught, hence areas of pollen 

deposition could not be checked. Since bees crawl over the 

anthers and filaments to access nectar at the base of the floral tube, 

pollen placement is not always localised in discrete areas on the 

insect body. All visitors carried substantial amounts of Plectranthus 

pollen on their bodies, but the percentages of foreign 

pollen varied from 0 to 75% (Table 4). 

Observations during the course of this study show that 

nemestrinid flies visit flowers for nectar; apinid bees visit 

flowers for nectar and sporadically for pollen collection, while 

megachilinid bees utilise nectar and/or load pollen onto the 

ventral abdominal scopae. With thoracic lengths of 13–18 mm 

and widths of 6–7 mm, apinid bees and nemestrinid flies 

comprise a class of large-bodied, nectar-feeding pollinators 

sufficiently powerful to depress the lower lip of a Plectranthus 

flower and pick up pollen from the anthers, but too broad to 

fully enter the mouth of a corolla tube (Table 5). Likewise, the 

smaller pollen-collecting megachilinid bees are also too large to 

fully enter into the laterally compressed corolla tubes (Table 5). 

Results from the 2003 study on bee and fly visitor effectiveness 

on P. laxiflorus did not show a definitive increase or 

decrease as a result of fly emergence: average fruit set was 

57.6% (SD =19.5) before fly emergence (i.e. bees only), and 

49.2% (SD =22.8) after fly emergence (i.e. bees and flies). In 

seven inflorescences the fruit set was higher after fly emergence 

while in nine cases fruit set was lower. At a site in the Dargle it 

was noted that on an overcast morning flies were the only 

visitors to P. laxiflorus flowers, until about 11.30am, when the 

sun came out and a few bees emerged. In this instance, P. 

umbrosa was the main floral visitor for the morning. 

A few butterfly and one day-flying hawkmoth species feed on 

nectar of sigmoid Plectranthus species (Appendix 2: Lepidoptera); 

no pollen was found on the examined voucher specimens. 

The pollination of three species with variously sigmoid corolla 

shapes was studied in less detail and is summarized as follows. 

Plectranthus spicatus is a savanna species (Fig. 1d) with a 

succulent perennial habit, producing many decumbent stems 

from the base, with sub-spicate inflorescences that ascend up to 

60 cm (Codd, 1985). The small flowers are blueish-purple in 

colour, with a sigmoid tube that is basally narrow, ascending at 

first and then curving sharply downwards, expanding at the 

throat (Figs. 6, 7d). Floral shape is similar to that of Py. 

urticifolia, but the corolla tube is shorter. Xylocopa caffra bees 

(proboscis length 6.5 mm) are the main pollinators of P. 

spicatus flowers at Oribi Gorge NR, with sporadic visits made 

by Amegilla mimadvena bees (proboscis length 9 mm). 

Table 5 

Range of body (thorax) size of insects that form the major pollinator groups of 

all seven studied plant species. 

Insect group Length of thorax 

(mm) 

Chapter 3/ 37 

Amegilla spp. (Apinid bee) 14–15 6–6.5 

Xylocopa spp. (Apinid bee) 15–18 6–8 

Prosoeca spp. (Nemestrinid fly) 13 7 

Chalicodoma spp. (Megachilinid bee) 13–15 5–6 

653 

Width of thorax 

(mm)


Fig. 6. Corollas of three additional Plectranthus and allied species studied: 

Plectranthus spicatus, P. rehmanii, Aeollanthus parvifolius. Bar: 2.5 mm. 

Plectranthus rehmannii is endemic to the KZN Midlands 

(Fig. 1d), where it grows along or near forest margins. It is an 

erect, perennial sub-shrub reaching 1.2 m, with paniculate 

inflorescences up to 350 mm tall (Codd, 1985). The small 

flowers are creamy-white, with saccate bases and deflexed tubes 

(Fig. 6), similar to that of P. calycinus. At Leopard's Bush NR in 

the Karkloof, where P. rehmannii and P. laxiflorus grow in 

close proximity, the pollinator is a megachilinid bee, Chalicodoma 

sp. A (proboscis length 6 mm), which visits both species 

of Plectranthus. In a forest in the Dargle these two species also 

grow together, but none of the abundant Prosoeca flies at that 

site were seen to visit P. rehmannii. Honey bees (Apis mellifera) 

and Allodape pernix visited flowers of P. rehmannii, but only 

collected pollen from the anthers. 

Aeollanthus parvifolius is a semi-succulent, perennial, herbaceous 

species that grows amongst rocks in grassland. The white to 

pinkish red flowers are borne on relatively lax, branched inflorescences. 

Cylindrical corolla tubes are 7–10 mm long, narrow at 

the base, expanding slightly towards the mouth (Codd, 1985), with 

a midway curve approaching the sigmoid shape. On the granite 

outcrops at Ongoye Forest (Fig. 1d), Aeollanthus parvifolius was 

pollinated by the apinid bees Amegilla bothai, Amegilla mimadvena 

and Amegilla fallax, all with proboscids ranging from 7 to 

9 mm long. An acrocerid fly species (Psilodera sp.), with a flexible 

proboscis of similar length to those of the visiting bees, also probed 

flowers, but was not caught. 

Estimations of proportional visitation by fly and bee 

pollinators to straight-tubed Plectranthus species showed that 

species that grow in or near shaded forest habitat tend to have 

more fly pollination visits, while species growing out in the 

open or in more sunny areas tend to have more bee visits 

(Table 6). This pattern is not clear in the sigmoid species, where 

most species are associated with sunlit patches in or near forest, 

or with sunny areas away from forest altogether. 

5. Discussion 

Chapter 3/ 38 

The observed modes of visitation by bees to sigmoid labiates 

generally correspond to that described by Scott Elliot (1891) 

and Stirton (1977). The nemestrinid flies (Pr. umbrosa, Pr. 

circumdata and Prosoeca sp. nov. 5, with proboscids 9– 

10.5 mm long) that visit P. laxiflorus and P. calycinus, are a 

new group of pollinators of sigmoid Plectranthus species. 

Nemestrinid flies with proboscids ranging from 8 to 30 mm also 

pollinate other, straight-tubed species of Plectranthus (Potgieter 

et al., 1999; Potgieter and Edwards, 2005). 

There is an apparent close fit between the length of corolla tubes 

and the length of the proboscids of most apinid bees and 

nemestrinid flies that pollinate P. laxiflorus and P. petiolaris (as 

evidenced by visual comparisons in Figs. 2, 3).Theflexibletipsof 

bee proboscids accommodate the bend of the corolla tube, as does 

the flexibility of the nemestrinid proboscis along its whole length 

(Figs.2,3). These insects pick up pollen ventrally on the head, 

thorax and abdomen as the insect moves over the anthers. Species 

such as Apis mellifera (Fig. 2, P. laxiflorus) andAllodape pernix 

(Fig. 3, P. petiolaris) cannot reach nectar and rather collect pollen. 

In the case of P. calycinus this fit by visual comparison only 

seems to hold for the apinid bee, Xylocopa scioensis (Fig. 4). 

Yet, despite its longer proboscis, the nemestrinid fly, Prosoeca 

umbrosa, picks up P. calycinus pollen on the hairy junction at 

the base of the head, as well as on the ventral surface of the 

thorax and abdomen (Table 4; Fig. 7h), when it appears as 

though only the base of the head should remove pollen from the 

anthers. The saccate base and inflated corolla of P. calycinus 

may be responsible for better contact between the anthers (and 

stigma) and the fly body, since the flexible proboscis may follow 

the curve of the upper part of the dorso-laterally flattened tube as 

it probes, which bends the proboscis and brings the fly body 

closer. Despite the weakly sigmoid shape of the corolla (it only 

bends close to the base) this design may function in excluding 

insects that do not have flexible proboscids in a way similar to 

other sigmoid corollas, while allowing for shorter-proboscid 

bees with flexible proboscids to probe directly towards the 

nectary by aligning with the lower part of the corolla (Fig. 4). 

In Py. urticifolia the lengths and point of flexibility of 

proboscids of Amegilla mimadvena, Xylocopa scioensis, X. 

flavorufa and Thyreus sp. fit the corolla bend perfectly for 

nectar extraction and pollen removal (following visual comparison 

in Fig. 5). The recorded megachilinid bees cannot reach the 

corolla base for nectar. These bees actively collect pollen 

ventrally into abdominal scopae, in the same way as described 

by Percival (1965).

In general, the knee-like bend in the corolla tubes of P. 

laxiflorus, P. petiolaris and Py. urticifolia acts as a physical 

barrier in the corolla tube that only allows insect groups with 

flexible proboscids of sufficient length access to the nectar; the 

boat-shaped lower lip also limits the type of insect that can pick 

up pollen on its body. In P. calycinus this bend is less restrictive, 

sitting quite close to the corolla base, yet the two recorded 


Chapter 3/ 39 

Fig. 7. A comparison of corollas of four sigmoid species (a–d), two insect pollinators (e–f) and undersides of an apinid bee and a nemestrinid fly (g–h). 

(a) Plectranthus petiolaris; (b) Pycnostachys urticifolia; (c) Plectranthus laxiflorus; (d) Plectranthus spicatus; (e) Amegilla mimadvena (Apinae, Hymenoptera); 

(f) Prosoeca umbrosa (Nemestrinidae, Diptera); (g) ventral surface of a typical apinid bee, Amegilla mimadvena, showing hairs on thoracic area and below the head; 

(h) ventral surface of a typical nemestrinid fly, Stenobasipteron wiedemanni, showing hairs below the head and groove between leg bases. 

visitors belong to bee and fly genera that also visit other sigmoid 

species. The same situation exists in P. rehmanii, while A. 

parvifolius has more gently curved corollas (Fig. 6). These act in 

much the same way as the knee-like bends discussed earlier, and 

that of P. spicatus, since a flexible proboscis is required to access 

nectar resources. The declined angle of the corolla tube forces 

insects into close contact with the reproductive parts of the 

655


flower, but this means that an ascending tube may be necessary 

at the base of the corolla as it may prevent the gravitational bleed 

of nectar—the combination of these two evolutionary forces are 

most likely responsible for the sigmoid shape of the corollas. In a 

study on the pollination of Disa versicolor (Orchidaceae) it was 

found that there was a good fit between orchid spur length and 

that of bee mouthparts, while the sharply decurved spur of this 

species functions to accommodate the ‘long curved tongue of 

Amegilla or similar large long-tongued bees' (Johnson, 1995). 

The effect of the sigmoid corolla shape on forcing contact 

between insect and floral reproductive parts (i.e. anthers and 

stigma) is more pronounced in bees than in flies. The flight of 

the approaching, probing bee is interrupted by the need to settle 

on the lower lip and/or reproductive parts of the flower, since 

the proboscis locks in linear alignment with the insect's body; to 

accommodate the angle in the sigmoid corolla the insect body is 

forced below the horizontal. By settling there is close contact 

between the ventral surface of the bee body and the anthers/ 

stigma and any movement by the bee to probe into or retreat out 

of the corolla tube leads to further chances of pollen deposition 

onto the bee and removal from the anthers. In nemestrinid flies 

the proboscis is more flexible and can bend to some extent to 

accommodate the sigmoid shape of the corolla tube, hence flies 

can hover while feeding and often bypass the anthers/style in 

the process. This happens during some visits, but observations 

also showed that the flies may grip the floral reproductive parts 

as they settle on the lower lip during feeding. 

The short proboscis of Allodape pernix cannot legitimately 

reach the nectar in P. petiolaris (Fig. 3), other than by observed 

instances of robbing, where the bee pierces holes into the 

corolla base. This is akin to the robbing mentioned by Scott 

Elliot (1891) for P. calycinus, where a small ‘fly’ [sic.] was 

implicated. Species of Plectranthus with straight corolla tubes 

are robbed in a similar way by A. pernix (Potgieter et al., 1999), 

where the bees act as “primary nectar robbers” and possibly as 

secondary robbers as well, where nectar is accessed through 

holes made by primary robbers (Inouye, 1980). 

Honey bees, Apis mellifera, are only able to reach the nectar 

in P. laxiflorus and Py. urticifolia (Figs. 2, 5) if the bees force 

their way into the corolla and if nectar levels are very high (e.g. 

at the start of the day). In most cases honey bees were observed 

rather to collect pollen from the anthers of flowers, as in the case 

of Allodape pernix. Observations in P. petiolaris show that 

while nectar may be more accessible to shorter-proboscid 

insects early in the morning, residual nectar is available later in 

the day only to longer-proboscid insects, when the bend in the 

corolla tube functions as an effective exclusion mechanism. 

The short proboscids of megachilinid bees prohibit nectar 

feeding altogether, although these species collect copious 

amounts of pollen on the modified hairs (abdominal scopae) 

on the ventral surfaces of their abdomens. Unlike the situation 

in other bee families where the insects groom pollen off the 

body into scopae located on the hind legs (and hence out of the 

reach of stigmatic surfaces), the abdominal scopae of megachilinid 

bees make pollen constantly available for subsequent 

deposition on stigmatic surfaces. While this ensures effective 

pollination, it may lead to increased levels of geitonogamous 

fruit set, compared to other bees where pollen is sporadically 

groomed out of the reach of stigmas. 

Although Van der Pijl (1972) suggested that sigmoid corollas 

may provide a shift from butterfly to bee pollination, the current 

study records both bee and a few butterfly visitors, thus the corolla 

shape allows access to both groups. Lepidoptera are, however, not 

effective pollinators of Plectranthus since they tend to bypass the 

reproductive parts of the flowers by hovering or briefly settling on 

the lower lip of the flower without contacting stigma or anthers. 

The sigmoid species that have both bee and fly pollinators, such 

as P. laxiflorus and P. calycinus, pose interesting questions with 

respect to which pollinator group is more efficient at pollen 

carryover. The experiment conducted at Ferncliff NR to test this in 

P. laxiflorus does not, however, answer these questions fully. The 

results show little or no increase in fruit set later in the season, 

despite a considerable increase in pollinator activity when the 

nemestrinid flies emerge and forage actively alongside apinid bees. 

Table 6 

Twenty species of Plectranthus with straight or sigmoid, long or shorter corolla 

tubes listed according to habitat, showing proportions of bee and fly visits (C. 

Potgieter, unpubl. data). 

Species of Plectranthus 

(or allied genus) 

Chapter 3/ 40 

Habitat % % 

Flies Bees 

Long-tubed, straight (20–32 mm) 

P. saccatus Benth. (long-tubed 

variety) 

FOR: deep shade 100 

P. reflexus E.J. Van Jaarsveld 

and T.J.Edwards 

FOR: deep shade 100 

P. hilliardiae Codd FOR: deep shade 100 

P. ambiguus (H.Bol.) Codd 

Shorter-tubed, straight (4–18 mm) 

FOR: sunlit patches 80 20 

P. oertendahlii Th.Fries jun. FOR: deep shade 100 

P. ciliatus E.Mey ex Benth. FOR: deep shade and sunlit 60 40 

P. zuluensis T. Cooke FOR: shade 100 

P. praetermissus Codd FOR: shade and sunlit patches 100 

P. fruticosus L'Hérit. FOR: shade and sunlit patches/ 100 

margins 

P. ecklonii Benth. FOR: sunlit patches/margins 70 30 

P. oribiensis Codd FOR/SUN: forest margins and 

wooded areas 

100 

P. madagascariensis (Pers.) SUN: grassland, woodland, 50 50 

Benth. 

forest margin 

P. ernstii Codd 

Sigmoid-tubed (5–11 mm) 

SUN: sunny cliffs 100 

P. petiolaris E.Mey. ex Benth. FOR: sunlit patches/margins 1 99 

P. laxiflorus Benth. FOR: sunlit margins (late 

season) 

50 50 

P. rehmannii Gürke FOR: shade and sunlit patches/ 

margins 

100 

Py. urticifolia Hook. SUN: stream banks, sunlit 

forest margin 

100 

P. spicatus E.Mey. ex Benth. SUN: grassland and dry 

woodland 

100 

P. calycinus Benth. SUN: grassland 100 

A. parvifolius Benth. SUN: sunlit rocky areas 25 75 

FOR: species associated with forest habitat (sometimes on sunny forest margins); 

SUN: species in sunny habitats.% Flies represents an estimate of the number of 

visits received by effective fly pollinators in the families Nemestrinidae, 

Tabanidae and Acroceridae; % Bees represents an estimate of the number of 

visits received by effective bee pollinators in the apid sub-families Apinae and 

Megachilinae.

One explanation for this could be the genetic constraint on 

ovule number in the Lamiaceae: flowers of this family have just 

four ovules per flower, setting four nutlets per fruit. The 

sequential maturation of fruit did not allow for individual nutlets 

to be counted at the end of the season, since some of the smooth 

seeds were shaken from the calyces before the last fruits matured, 

hence the number of nutlets per fruit could not be considered a 

reliable count and only fruit set was represented here. The limited 

number of ovules in Plectranthus may explain why bee visitors 

alone effected as much fruit set as bees combined with flies, since 

only four pollen grains are needed to give full fertilisation of a 

fruit; any subsequent visits would not increase fruit set. Bees 

alone may then be adequate to provide maximum fruit set, or, at 

least, as much fruit set as when flies and bees are visiting together. 

While the maximal fruit set was not established with the aid of 

hand pollinations, it would seem likely that heavy visitation by 

two pollinator groups later in the season would account for most 

of the potential fruit set in the population. 

The distribution of P. laxiflorus extends beyond that of the most 

abundant fly visitor, Pr. umbrosa, but where the plants do co-occur 

with the flies it is heavily visited by these flies (in addition to bees); 

in overcast conditions where bee activity may be limited, the flies 

may be the dominant visitor (as was seen at the Dargle site). This 

suggests that flies could be more important as pollinators in years 

with predominantly overcast weather. Most of the sigmoid species 

discussed here are limited to sunny or grassland habitats (P. 

calycinus, P. spicatus, Py. urticifolia, A. parvifolius), sunlit forest 

margins (P. laxiflorus, P. rehmannii) or open patches in forests and 

along forest margins (P. petiolaris) (Table 6). In other, straighttubed 

species of Plectranthus there is a general pattern of species in 

shaded forest habitat being associated with a greater proportion of 

fly pollinators, while species in sunny areas (away from forest and 

forest margins) favour more bee pollinators (Table 6). This is 

confirmed by a study on pollination in a South African grassland 

community, where it was found that long-proboscid solitary bees 

were the most important floral visitors (Johnson et al., 2009-this 

issue). This pattern is not as clear in the sigmoid species, none of 

which occur in deep forest shade, but it could point to a trend in 

increasing reliance on fly pollinators adjacent to forests (e.g. P. 

laxiflorus with a 50:50 split in the latter part of the season—after fly 

emergence). The possibility exists that, as sigmoid species became 

more shade-tolerant and moved closer to forest habitat, they 

became pollen-limited as a result of bees not being active in denser 

forest shade; nemestrinid flies would then be of greater importance 

as pollinators. Nemestrinid flies are, however, also active in some 

grassland habitats (e.g. P. calycinus). 

It is possible that nemestrinid flies are more efficient at pollen 

carryover than bees since the latter group of insects groom pollen 

off their bodies more frequently than flies do (S. Morita, pers. 

com.) At one stage it was thought that nemestrinid flies do not 

groom, but observations during this study confirmed that limited 

grooming happens during hovering and whilst resting. This 

happens especially to clean the eyes (S. Morita, pers. com.). 

The observation that a species of acrocerid fly, Psilodera sp., 

visited A. parvifolius at Ongoye, where Amegilla bees were the 

more abundant floral visitors, is another case of a fly species 

occupying the same niche as a bee species. In this case the fly looks 


and sounds like one of the resident anthophorine bees, A. fallax.A 

similar observation on Iridaceae prompted Goldblattetal.(1997)to 

describe a related acrocerid fly as a bee mimic. Acrocerid flies are 

pollinators of a number of straight-tubed Plectranthus species that 

are also pollinated by various species of Amegilla bees with similar 

proboscis lengths (Potgieter et al., 1999; Viljoen et al., 2006). 

The phylogenetic work by Paton et al. (2004 and pers. com.) 

does not adequately explain the relationships between the different 

floral types of Plectranthus, hence the question of whether 

sigmoid tubes evolved from straight corollas, or vice versa, is 

difficult to answer. We suggest the possibility that straight corollas 

have evolved from sigmoid ones, for the following reasons. 

Our data from P. laxiflorus shows that a shift from sigmoid to 

straight corolla tubes could occur since nemestrinid flies can 

accommodate the bend in sigmoid tubes.Whythendidcorollas 

straighten? When nemestrinid flies visit sigmoid corollas they do 

not always contact the anthers, since direct observations show that 

their flexible proboscids can bypass the anthers and stigma if the fly 

is hovering. This reduced efficiency is not significant where visiting 

insects are plentiful. However data collected from forest species of 

Plectranthus show widespread pollen limitation and field observations 

confirm that insect visits are rare (C. Potgieter, unpubl. data). 

Under these circumstances increased efficiency of pollen transfer 

would be strongly selected. By straightening the corolla tube the 

anthers and stigma are shifted into the flight path of flies which 

makes it more difficult to bypass. Those grassland labiate species 

(e.g. Orthosiphon, Thorncroftia) that are pollinated by nemestrinid 

flies all have straight corolla tubes (Potgieter and Edwards, 2001). 

A shift from sigmoid to straight corolla tubes could explain the 

saccate corolla bases found in many straight-tubed forest Plectranthus 

species (e.g. P. saccatus, P. reflexus), since the saccate 

base may represent the upper remnants of a sigmoid corolla. In this 

study the corolla shapes of P. calycinus and P. rehmannii 

represent a possible intermediate situation of a saccate corolla base 

combined with a bend near the base. Only two forest Plectranthus 

species, P. ambiguus and P. ecklonii, have straight corolla bases. 

These questions will only be answered fully once a comprehensive 

phylogeny of Plectranthus and its allies, including all the longtubed 

species from southern Africa, has been constructed. 

Acknowledgements 

Chapter 3/ 41 

The following persons and institutions are thanked for 

assistance: University of KwaZulu-Natal (UKZN) Research Office 

and National Research Foundation for funding; Prof. D. Brothers, 

Dr D. Barraclough, Dr C. Eardley and the late Dr B. Stuckenberg 

for insect identification; Centre for Electron Microscopy (UKZN) 

for access to photographic equipment and a scanning electron 

microscope; National Herbarium, Pretoria (PRE), KwaZulu-Natal 

Herbarium, Durban (NH) and Bews Herbarium, UKZN (NU) for 

access to distribution data; Craig Symes, Dave Thompson, Carol 

Rolando, Clinton Carbutt, Neil Crouch, Pev Curry andCameron 

McMaster for assistance with field observations; Miles Hunt 

(Leopards Bush NR), Vernon Green and the Booysens (Dargle) 

and Ezemvelo KZN Wildlife (Oribi Gorge NR, Umtamvuna NR) 

for access to field sites; Alan Paton for providing ideas and updates 

on phylogenetic work. 

657


Appendix 1 

Study site details and plant vouchers for species studied at each site. 

KZN = KwaZulu-Natal Province; EC = Eastern Cape Province; NR = Nature Reserve. 

Study site Province Quarter deg. grid Plant species studied Year studied Voucher 

Umtamvuna NR KZN 3030CC P. petiolaris 1995–1998 C. Potgieter 115 

P. calycinus 2000 

Oribi Gorge NR KZN 3030CB P. petiolaris 1996–1998 C. Potgieter 100 

P. laxiflorus 1998 

P. spicatus 1998 C. Potgieter 142 

Ferncliff NR: Pietermaritzburg KZN 2930CB P. laxiflorus 1996–2003 C. Potgieter 135 

Leopard's Bush NR: Karkloof KZN 2930CB P. laxiflorus 1999 C. Potgieter 145 

P. rehmannii 1999 C. Potgieter 150 

Dargle: KZN Midlands KZN 2930AC P. laxiflorus 2009 

P. calycinus 2000 C. Potgieter 154 

P. rehmannii 2009 T. Edwards 3518 

Ngeli Forest: Weza KZN 3029DA P. laxiflorus 1998, 2001 

Garden, Pietermaritzburg KZN 2930CB Py. urticifolia 1998 C. Potgieter 1064 

Ongoye Forest: Empangeni KZN 2831DC A. parvifolius 1998 

Kologha Forest: Stutterheim EC 3227CB P. laxiflorus 1998, 2000 

Appendix 2 

Chapter 3/ 42 

All observed insect visitors to flowers of the four species studied in detail: Plectranthus petiolaris, P. laxiflorus, P. calycinus and 

Pycnostachys urticifolia. 

Localities indicated; U, Umtamvuna Nature Reserve; O, Oribi Gorge Nature Reserve; P, Pietermaritzburg (Ferncliff Nature 

Reserve); K, Karkloof (Leopards Bush Nature Reserve); D, Dargle area; N, Ngeli area; S, Stutterheim (Kologha forest). 

P. petiolaris P. laxiflorus Py. urticifolia P. calycinus 

Hymenoptera Hymenoptera Hymenoptera Hymenoptera 

Apidae Apidae Apidae Apidae 

Apinae Apinae Apinae Apinae 

Amegilla mimadvena U, O Amegilla mimadvena P, S Amegilla mimadvena P 

Amegilla bothai O Amegilla bothai P, K 

Amegilla caelestina U, O, P Amegilla caelestina O Thyreus sp. P 

Apis mellifera P Apis mellifera P Apis mellifera D 

Xylocopa hottentotta U Xylocopa flavicollis S Xylocopa scioensis P Xylocopa scioensis U 

Allodape pernix U Allodape ceratinoides P Xylocopa flavorufa P 

Halictinae Halictinae 

Lasioglossum sp. O Zonalictus sp. P 

Megachilinae Megachilinae 

Chalicodoma sp. A K Chalicodoma sp. B P 

Pseudoanthidium truncatum P 

Megachile sp. A P 

Megachile sp. B P 

Diptera Diptera Diptera 

Nemestrinidae Nemestrinidae 

Prosoeca umbrosa P Prosoeca umbrosa D 

Prosoeca circumdata P, N, S 

Prosoeca sp. nov. 5 P 

Syrphidae Syrphidae 

Asarkina sp. O Asarkina sp. A N 

Asarkina sp. B K 

Episyrphus sp. U Oniromyia sp. P 

Voria sp. P 

Bombyliidae P 

Lepidoptera Lepidoptera Lepidoptera 

Pieridae U Pieridae S 

Lycaenidae U, O, P Hesperidae K, P Lycaenidae P 

Papilionidae 

Papilio nireus lyaeus S 

Sphingidae 

Macroglossum trochilus P, S

References 

Barraclough, D.A., 2006. An overview of the South African tangle-vein flies 

(Diptera: Nemestrinidae), with an annotated key to the genera and a checklist 

of species. Zootaxa 1277, 39–63. 

Bezzi, M., 1924. The South African Nemestrinidae (Diptera) as represented in the 

South African Museum. Annals of the South African Museum 19, 164–190. 

Brothers, D.J., 1999. Phylogeny and evolution of wasps, ants and bees 

(Hymenoptera, Chrysidoidea, Vespoidea and Apoidea). Zoologica Scripta 

28, 233–249. 

Codd, L.E., 1975. Plectranthus (Labiatae) and allied genera in southern Africa. 

Bothalia 11, 371–442. 

Codd, L.E., 1985. Plectranthus (Lamiaceae). Flora of Southern Africa 28, 

137–172. 

Edwards, T.J., 2005. Two new Plectranthus species (Lamiaceae) and new 

distribution records from the Pondoland Centre of Plant Endemism, South 

Africa. Bothalia 35, 149–152. 

Edwards, T.J., Paton, A., Crouch, N.R., 2000. A new species of Plectranthus 

from Zimbabwe. Kew Bulletin 55, 459–464. 

Goldblatt, P., Manning, J.C., Bernhardt, P., 1997. Notes on the pollination of 

Gladiolus brevifolius (Iridaceae) by bees (Anthophoridae) and bee mimicking 

flies (Psilodera: Acroceridae). Journal of the Kansas Entomological Society 

70, 297–304. 

Huck, R., 1992. Overview of pollination biology in the Lamiaceae. In: Harley, 

R.M., Reynolds, T. (Eds.), Advances in Labiate Science. Royal Botanic 

Gardens Kew, Kew, pp. 167–181. 

Inouye, D.W., 1980. The terminology of floral larceny. Ecology 61, 1251–1253. 

Johnson, S.D., 1995. The pollination of Disa versicolor (Orchidaceae) by 

anthophorid bees in South Africa. Lindleyana 9, 209–212. 

Johnson, S.D., Harris, L.F., Proches, S., 2009. Pollination and breeding systems 

of wildflowers in a southern African grassland community. South African 

Journal of Botany 75, 630–645 (this issue). 

Lukhoba, C.W., Simmonds, M.S.J., Paton, A.J., 2006. Plectranthus: a review of 

ethnobotanical uses. Journal of Ethnopharmacology 103, 1–24. 

Meeuse, A.D.J., 1992. Anthecology of the Labiatae: an armchair approach. In: 

Harley, R.M., Reynolds, T. (Eds.), Advances in Labiate Science. Royal 

Botanic Gardens Kew, Kew, pp. 167–181. 

Paton, A.J., Springate, D., Suddee, S., Otieno, D., Grayer, R.J., Harley, M.M., 

Willis, F., Simmonds, M.S.J., Powel, M.P., Savolainen, V., 2004. 

Edited by JC Manning 


Chapter 3/ 43 

Phylogeny and Evolution of Basils and Allies (Ocimeae, Labiatae) based 

on three Plastid DNA Regions. Molecular Phylogenetics and Evolution 31, 

277–299. 

Percival, M.S., 1965. Floral Biology. Pergamon Press, Oxford. 

Potgieter, C.J., Edwards, T.J., 2001. The occurrence of long, narrow corolla 

tubes in southern African Lamiaceae. Systematics and Geography of Plants 

71, 493–502. 



Missouri Botanical Garden 92, 254–267. 




Retief, E., 2000. Lamiaceae (Labiatae). In: Leistner, O.A. (Ed.), Seed Plants of 

Southern Africa: Strelitzia, vol. 10, pp. 323–334. 

Scott Elliot, G.F., 1891. Notes on the fertilisation of South African and 

Madagascan flowering plants. Annals of Botany 5, 330–344. 

Stirton, C.H., 1977. Broad-spectrum pollination of Plectranthus neochilus. 

Bothalia 12, 229–230. 

Van der Pijl, L., 1972. Functional considerations and observations on the 

flowers of some Labiatae. Blumea 20, 93–104. 

Van Jaarsveld, E.J., Edwards, T.J., 1991. Plectranthus reflexus. Flowering 

Plants of Africa 51, Plate 2034. 

Van Jaarsveld, E.J., Edwards, T.J., 1997. Notes on Plectranthus (Lamiaceae) 

from southern Africa. Bothalia 27, 1–6. 

Van Jaarsveld, E.J., Hankey, A., 1997. Plectranthus venteri Van Jaarsveld and 

Hankey spec. nov. (Lamiaceae), a new species from the Northern Province, 

South Africa. Aloe 34, 40–41. 

Van Jaarsveld, E.J., Van Wyk, A.E., 2004. Plectranthus mzimvubuensis, a new 

species from Eastern Cape, South Africa. Bothalia 34, 30–32. 


Microdistillation and essential oil chemistry — a useful tool for detecting 

hybridisation in Plectranthus (Lamiaceae). South African Journal of Botany 

72, 99–104. 

Vogel, S., 1954. Blütenbiologische Typen als Elemente der Sippengliederung. 

Botanischer Studien 1. 

Winter, P.J.D., Van Jaarsveld, E.J., 2005. Plectranthus porcatus, a new species 

endemic to the Sekhukuneland centre of plant endemism, Limpopo Province, 

South Africa. Bothalia 35, 169–173. 

659

CHAPTER 4: 

CONVERGENT POLLINATION IN SOUTHERN AFRICAN 

LAMIACEAE 

Potgieter, C.J., Edwards, T.J., 2001. 

The occurrence of long, narrow corolla tubes in southern African 

Lamiaceae. 

Systematics and Geography of Plants 71: 493–502.

Syst. Geogr. Pl. 71: 493-502 (2001) 

Chapter 4/ 45 

The occurrence of long, narrow corolla tubes 

in southern African Lamiaceae 

Christina J. Potgieter* & Trevor J. Edwards 

School of Botany & Zoology, University of Natal Pietermaritzburg, P/Bag XO 1, Scottsville, 3209, South Africa 

* author for correspondence [potgietercj @nu.ac.za] 

Abstract. - Long, narrow corolla tubes have evolved several times in genera of southern African 

Lamiaceae, such as Plectranthus, Thorncroftia, Orthosiphon, Hemizygia, Stachys and Salvia. A study 

of the pollination of Plectranthus in KwaZulu-Natal and the Eastern Cape of South Africa showed a 

nemestrinid fly, Stenobasipteron wiedemanni, to be the pollinator of the four species that have long 

corolla tubes. The proboscis length of thisfly (19 - 30 mm) corresponds to the corolla tube lengths of 

theflowers (20 - 33 mm). This fly is restricted to forest patches and woodlands along the eastern parts 

of southern Africa, while long-proboscid species of the nemestrinid genus Prosoeca occur mostly in 

grassland habitats. It is postulated that S. wiedemanni is the pollinator of Stachys tubulosa, Salvia 

scabra, Salvia repens var. keiensis and Hemizygia ramosa in forest and woodland habitats. Longproboscid 

species of the nemestrinid genus Prosoeca are suggested as pollinators of Orthosiphon 

tubiformis, Thorncroftia longiflora, T. succulenta, Hemizygia rugosifolia, H. gerrardii and Salvia repens 

var. keiensis, in grassland habitats. Novel distribution maps for four long-proboscid fly species are 

provided. This paper aims to discuss the distribution of long-tubed members of the Lamiaceae in 

relation to that of long-proboscid flies by relating the biogeography of the flies to that of the plant 

species. 

Key words: long corolla tubes, Lamiaceae, Plectranthus, Orthosiphon, Thorncroftia, Salvia, Stachys, 

Hemizygia, long-proboscidflies, Nemestrinidae, Stenobasipteron, Prosoeca. 

Abbreviations: SEM, scanning electron microscopy; LM, light microscopy; PRECIS, Pretoria (PRE) 

Computerised Information System. 

Resume. - Presence de tubes corollins longs et etroits chez des Lamiaceae sud-africaines. Des 

tubes corollins longs et etroits se sont developpes dans plusieurs genres sud-africains de Lamiaceae, 

comme Plectranthus, Thorncroftia, Orthosiphon, Hemizygia, Stachys et Salvia. Une etude de la 

pollinisation de Plectranthus au KwaZulu-Natal et dans I'Eastern Cape en Afrique du Sud ont montre 

qu'une mouche nemestrinide, Stenobasipteron wiedemanni, pollinisait quatre especes a- longs tubes 

corollins. La longueur du proboscis de cette mouche (19-30 mm) correspond a celle du tube corollin 

des fleurs (20-33 mm). La presence de cette mouche est limitee aux tlots forestiers et aux savanes 

arborees des regions situees a 1est de IlAfrique australe, alors que les especes de nemestrides ac long 

proboscis appartenant au genre Prosoeca se rencontrent principalement dans des milieux herbaces. 

II est postule que S. wiedemanni est le pollinisateur de Stachys tubulosa, Salvia scabra, Salvia repens 

var. keiensis et Hemizygia ramosa dans les habitats forestiers et en foret claire. Les especes de 

nemestrides ai long proboscis du genre Prosoeca sont supposees etre les pollinisatrices de Orthosiphon 

E. Robbrecht, J. Degreef & I. Friis (eds.) Plant systematics and phytogeography for the understanding of African biodiversity. 

Proceedings of the XVIth AETFAT Congress, held in 2000 at the National Botanic Garden of Belgium. 

Subject to copyright. All rights reserved. ? 2002 National Botanic Garden of Belgium 

Permission for use must always be obtained from the National Botanic Garden of Belgium. 

ISSN 1374-7886

Syst. Geogr. PI. 71 XVIth Ss.Gg.l.1Xt AETFA AETFAT Congress 

tubiformis, Thorncroftia longiflora, T. succulenta, Hemizygia rugosifolia, H. gerrardii et Salvia repens 

var. keiensis dans les milieux herbaces. De nouvelles cartes de distribution pour quatre especes de 

mouches a long proboscis sont etablies. Cet article a pour objectif de discuter la distribution de 

representants a long tube de la famille des Lamiaceae en relation avec celle de mouches a long 

proboscis en comparant la biogeographie des mouches a celle des especes vegetales. Traduit par le 

journal. 

Introduction 

Long, narrow corolla tubes have undoubtedly evolved several times in insect-pollinated genera of 

southern African Lamiaceae. Of the 49 species of Plectranthus in southern Africa, four have (or include 

forms with) corolla tubes that are substantially longer than those in the rest of the genus. These longtubed 

species are pollinated by the long-proboscid fly, Stenobasipteron wiedemanni (Nemestrinidae, 

Diptera); cited as Stenobasipteron sp. in Potgieter & al. (1999). 

Stenobasipteron wiedemanni has been recorded or inferred as a pollinator in few other studies 

(Goldblatt & Manning 1998, 2000; Manning & al. 1999) and is a relatively unknown pollinator that is 

restricted to forested or wooded areas along the eastern parts of southern Africa. 

Other nemestrinid fly species are well-known as pollinators in southern Africa, particularly in the 

Cape flora and montane grasslands of the eastern seaboard. A number of studies in recent years have 

highlighted this long-proboscid fly pollination syndrome, which appears to be almost unique and highly 

evolved in southern Africa. These studies were conducted in the families Ericaceae (Rebelo & al. 1985), 

Iridaceae (Goldblatt & al. 1995; Goldblatt & Manning 1998, 1999), Orchidaceae (Johnson & Johnson 

1993; Johnson & Steiner 1995, 1997) and Geraniaceae (Struck 1997), and a number of long-proboscid 

fly pollination guilds were identified by Manning & Goldblatt (1995, 1996, 1997) in the Cape flora. 

Goldblatt & Manning (2000) recently published a revision of this pollination syndrome in southern 

Africa. 

During the course of our study on pollination in Plectranthus in KwaZulu-Natal and the Eastern 

Cape of southern Africa, it was shown that S. wiedemanni is the primary pollinator of the four species 

with long corolla tubes (P. hilliardiae Codd, Codd P. ambiguus a (.Bol.) dd, P. reflexus E.J.Van Jaarsveld 

& T.J.Edwards and P. saccatus Benth. - long-tubed forms). No other long-proboscid insect was seen to 

visit a long-tubed Plectranthus, except for one instance where a swallowtail butterfly (Papilio sp.) 

visited a flower of P. reflexus at Port St Johns (Potgieter & Edwards, unpubl. data). Other visits to longtubed 

Plectranthus were either by S. wiedemanni or pollen-collecting bees. 

This pattern of long corolla tubes is repeated in other genera of the Lamiaceae, such as Thorncroftia, 

Orthosiphon, Stachys, Salvia and Hemizygia (table 1). 

494 

Table 1. Occurrence of long and medium corolla tubes in southern African Lamiaceae. 

Numbers are numbers of species. Long tube lengths fall in the range of 18 - 35 mm; 

medium tube lengths fall in the range of 10 - 17 mm. From Codd 1985; Van Jaarsveld & Edwards 1991, 1997. 

Genus Total long tubes medium tubes 

Plectranthus 49 4 7 

Thorncroftia 3 1 1 

Orthosiphon 9 1 2 

Hemizygia 28 3 21 

Stachys 40 2 4 

Salvia 22 3 | 15 

Chapter 4/ 46

Family F i systematics sr 

Stenobasipteron wiedemanni visits Gladiolus macneilii Oberm. (Iridaceae) in woodland and 

grassland habitats in northern Mpumalanga and was also seen to visit Orthosiphon tubiformis R.Good 

in this area (M. Lotter pers cor.; Goldblatt & Manning 1998, 1999, 2000). 

In the same area another long-proboscid nemestrinid fly species, Prosoeca robusta, was seen to visit 

Gladiolus calcaratus G.J.Lewis, and a couple of the examined insect specimens carried pollen of a 

Hemizygia species (Goldblatt & Manning 1998, 1999). Apart from the above records, no observational 

data on insect pollination of long-tubed Lamiaceae in southern Africa is available. This excludes the 

genus Leonotis, which is primarily bird pollinated (Vos & al. 1994). The genus Syncolostemon is insect 

pollinated, but has flowers with a wide corolla throat which are more suited to bee-pollination. 

Stachys tubulosa MacOwan, Salvia scabra L.f. and Hemizygia ramosa Codd occur in forests in 

KwaZulu-Natal and the Eastern Cape and have similar mauve, lilac or purple flowers with long, narrow 

corolla tubes. Other long-tubed labiate species that occur in grassland and woodland habitats in cool 

mist-belt areas include Orthosiphon tubiformis, Thorncroftia longiflora N.E.Br., Thorncroftia 

succulenta (Dyer & Bruce) Codd, Hemizygia rugosifolia Ashby and Hemizygia gerrardii (N.E.Br.) 

Ashby. Salvia repens Burch. ex Benth. var. keiensis Hedge occurs both in open woodland and grassland 

habitats (Codd 1985). In this paper we propose that these long-tubed labiates are also pollinated by longproboscid 

flies such as S. wiedemanni (in forest and woodland) or Prosoeca spp. (in grassland). 

Material and methods 

Pollinator observations on the long-tubed members of Plectranthus were carried out during the flowering season (December - May) 

from 1995 - 2000. Field work was conducted at Stutterheim, Port St Johns, Umtamvuna Nature Reserve, Oribi Gorge Nature Reserve 

and Karkloof (fig. 1). Insect visitors to flowers were captured after observation, killed in separate ethyl-acetate containing vials, 

pinned and inspected under a dissecting microscope. Areas of pollen deposition on the insects were noted. Pollen grains were removed 

for SEM using small strips of double-sided tape, or LM using tiny blocks of fuchsin jelly. 

Data on the occurrence and distribution of long-tubed species in Plectranthus and other genera of the Lamiaceae was obtained 

using Codd (1985), Van Jaarsveld & Edwards (1991,1997) and PRECIS records from 1984 until present. The latter records 

contributed to extended distributions for endemic species such as P. hilliardiae, P. reflexus and T. longiflora. 

Proboscis length measurements of flies were done from vouchers collected during the study and specimens from the Natal 

Museum. Insect distributions were augmented using specimen data from the Natal Museum in Pietermaritzburg, Albany Museum in 

Grahamstown and South African Museum in Cape Town. 

Additional distributional and proboscis length information was obtained from Goldblatt & Manning (1998, 1999). 

SOUTH AFRICA 

BOTSWANA 

Chapter 4/ 47 

C.J. Potgieter & T.J. Edwards, Long, narrow corollas in Lamiaceae 

Northern Province 

IMBAB 

LI /l. North-West n(gpumalarga' 7 

Figure 1. NAMIBIA 

Field study sites (0) 

.fr Plectranthus 

vN 

Free State Nat 

KwaZulual 

pollination observations 

in KwaZulu-Natal and 

the Eastern Cape. 

Northern Cape 

K 

K, Karkloof (2930AC); 

OG, Oribi Gorge Nature Reserve \ / u 

(3030CA); PSJF 

U, Umtamvuna Nature Reserve 

(3030CC & 3130AA); 

PSJ, Port St Johns (3129 DA); 

S, Stutterheim (3227CB). 

\ 

estern Cape 

_. 

_ 

Easteape 

F 

/ 

* 

A 

I 

M 

495

Syst. Geogr. Pl. 71 Syst. Geogr. 

P1.71XVIhAETFATCongres-s XVIth AETFAT Congress 

Results and discussion 

The long-tubed species of Lamiaceae that (potentially) belong to the long-proboscid fly pollination guild 

are listed in table 2. Tube length, habitat, flower colour and distribution are shown for each species in 

table 2. The long-tubed forms of Plectranthus saccatus have arisen convergently in both subspecies, P. 

saccatus subsp. saccatus and P. saccatus subsp. pondoensis E.J.Van Jaarsveld & S.Milstein (sensu Van 

Jaarsveld & Edwards 1997), thus only the long-tubed forms are listed and discussed. 

Table 2. Floral tube length, habitat, general distribution and flower colour of fourteen long-tubed 

South African species of Lamiaceae. 

Tube lengths in mm. See figs. 2 & 3 A for exact distributions. After Codd 1985; Van Jaarsveld & Edwards 1991, 1997; PRECIS; 

Potgieter unpubl. data. KZN, KwaZulu-Natal; EC, Eastern Cape; WC, Western Cape; Mp, Mpumalanga; NP, Northern Province. 

Plant species Tube length Habitat Distribution Flower colour 

Plectranthus 

P. ambiguus 

P. reflexus 


P. saccatus (long-tubed) 

Stachys 

St. tubulosa 

St. thunbergii 

Hemizygia 

H. ramosa 

H. rugosifolia 

H. gerrardii 

Thorncroftia 

T. longiflora 

T. succulenta 

Orthosiphon 

0. tubiformis 

Salvia 

Sa. scabra 

Sa. repens 

var. keiensis 

496 

20- 33 

24- 30 

21 - 32 

20- 30 

(12) 18-23 

16- 20 

20- 22 

ca. 18 

17- 20 

30- 38 

15- 20 

20- 36 

20- 35 

15-19 

Forest 

Forest 

Forest 

Forest 

Forest 

Forest margins 

Woodland 

Rocky 

grassland 

Rocky 

grassland 

Rocky 

grassland 

Open woodland 

Open woodland 

KZN & EC 

EC (Pt St Johns) 

S. KZN & EC 

KZN 

KZN & Swaziland 

WC & EC 

N. KZN 

Mp & NP 

Mp, Swaziland 

Mp & Swaziland 

NP & Mp 

Mp 

Forest margins KZN & Swaziland 

Grassland & EC 

woodland 

Chapter 4/ 48 

violet - purple 

pale blue 

pale bluish 

mauve 

pinkish white 

red/ purple 

mauve 

pale pink 

mauve- pink 

mauve pink 

bluish mauve 

whitish- mauve 

mauve/ lilac 

mauve - purple

Family F systematics 

- 

Table 3. Proboscis length, habitat and general distribution of four long-proboscid fly species. 

Proboscis lengths in mm. See fig. 3 B for exact distributions. 

Flies are (potential) pollinators of long-tubed Lamiaceae in the eastern parts of South Africa. 

After literature cited and measurements of Natal Museum (NM) and voucher specimens from this study. N, Nemestrinidae: Diptera; 

KZN, KwaZulu-Natal; EC, Eastern Cape; Mp, Mpumalanga; G & M, Goldblatt & Manning; J & S, Johnson & Steiner. 

Name of dipteran Proboscis Distribution Habitat Reference 

length 

Stenobasipteron 19- 30 EC & KZN Forest Potgieter & al. '99, 

wiedemanni (N) NM, our vouchers. 

(23 - 29) Mp (Abel Erasmus Pass) Woodland G & M '98, '99. 

Stenobasipteron 14 - 24 Mp (Barberton & Forest NM 

cf. gracile (N) Mariepskop) 

Chapter 4/ 49 


Prosoeca 20- 23 Mp Grassland G & M '98, '99. 

robusta (N) 

Prosoeca 17 - 42 Eastern areas Grassland 

ganglbaueri (N) (29- 35) Mp G & M '98, '99; NM. 

(19- 30) KZN Drakensberg G & M '98, '99; NM. 

(17 - 29) EC (Naude's Nek) J & S '95; NM. 

(25 - 42) EC G & M '98, '99; NM. 

Distribution maps of long-tubed labiate species and subspecies are given in figs. 2 A, B & 3 A. 

Four nectar-feeding nemestrinid fly species from the summer rainfall region have proboscides that 

fall within the range of tube lengths of long-tubed labiate species. Table 3 lists each species with 

proboscis length, habitat and general distribution. Fig. 3 B shows a distribution map of all four fly 

species. 

While measuring proboscis lengths of S. wiedemanni specimens at the Natal Museum it became 

apparent that large collections from two localities in Mpumalanga looked different to the rest, i.e. from 

Barberton, De Kaap (in 1920) and Mariepskop (in 1932). These specimens are similar to Stenobasipteron 

gracile and will be referred to as Stenobasipteron cf. gracile, as the proboscis lengths place 

them in a functionally different pollinator category. It is likely that these flies pollinate flowers that occur 

in Mpumalanga and have corolla tubes slightly shorter than those visited by S. wiedemanni. 

The nemestrinid fly, Pr. ganglbaueri, is restricted to grassland habitats at high altitudes and is the 

likely pollinator of a few grassland Lamiaceae with very long tubes (table 3 shows variation in proboscis 

length across its distribution). 

Fig. 4 A, B, C & E shows the similarities in floral tube length for Plectranthus; the similarly long, 

narrow corolla tubes of 0. tubiformis are shown in fig. 4 F, and fig. 4 D & G shows the long-proboscid 

nemestrinid flies S. wiedemanni and Pr. ganglbaueri. 

Most of the labiate species listed in table 2 have flower colours in the purple/mauve/pale blue to 

white spectrum, coupled with long, constricted corolla tubes and odourless flowers, which conforms to 

the requirements of long-proboscid fly pollinators. In the case of P. saccatus the corolla tube is vertically 

broad, but the entrance is laterally compressed, thus restricting foraging to pollinators with long 

proboscides. 

One species, Stachys thunbergii Benth., deviates from this colour range and it is the only long-tubed 

labiate that grows in the winter rainfall area of the Western and Eastern Cape. It has red, magenta or 

purple flowers and is probably pollinated by sympatric long-proboscid flies, e.g. Philoliche rostrata 

(Tabanidae) or Prosoeca nitidula (Nemestrinidae). 

497

Syst. Geogr. Pl. 71 

Figure 2. 

Distribution of long-tubed Lamiaceae. 

A: *, Thorncroftia longiflora (tube 30 - 38 mm); 

A, T. succulenta (tube 15 - 20 mm); 

O, Orthosiphon tubiformis (tube 20 - 36 mm); 

E, Plectranthus ambiguus (tube 20 - 33 mm); 

0, P. reflexus (tube 24 - 30 mm). 

B: O, Stachys tubulosa (tube 18 - 23 mm); 

O, St. thunbergii (tube 16- 20 mm); 

*, Salvia scabra (tube 20 - 35 mm); 

A, Sa. repens var. keiensis (tube 15 - 19 mm). 

498 

Chapter 4/ 50 

XVIth AETFAT Congress 

I -1 

Figure 3. Distribution of long-tubed Lamiaceae 

and long-proboscid flies. 

A, Lamiaceae. 

A, Hemizygia rugosifolia (tube ca. 18 mm); 

E, H. gerrardii (tube 17- 20 mm); 

*, H. ramosa (tube 20- 22 mm); 

0, Plectranthus hilliardiae (tube 21 - 32 mm), 

*, P. saccatus (long-tubed forms, 20 - 30 mm). 

B, Flies: 

O, Stenobasipteron wiedemanni (proboscis 19 - 30 mm); 

0, Stenobasipteron cf: gracile (proboscis 14 - 24 mm); 

*, Prosoeca ganglbaueri (proboscis 17- 42 mm); 

A, Prosoeca robusta (proboscis 20 - 23 mm).

Family ? ,g systematics 

D 

Chapter 4/ 51 


Figure 4. Long-tubed Lamiaceae and their pollinators. 

A, Plectranthus hilliardiae; B, Plectranthus reflexus; C, Plectranthus ambiguus; D, Stenobasipteron wiedemanni; E, Plectranthus 

saccatus (long-tubed form, Umtamvuna); F, Orthosiphon tubiformis,; G, Prosoeca ganglbaueri (with orchid pollinaria, 

photo: S. Johnson); scale bars A, B, C, E & F = 2 cm, scale bars D & G = 1 cm. 

499

Syst. Svst. Geogr. Pl. P1. 71 XVIth AETFAT AETFAT Congress Congress 

Stachys tubulosa is the only other long-tubed Stachys, but it has pinkish white flowers, a tube that is 

nearly straight and a deflexed lower lip; it also grows in moist forests and forest margins, thus S. 

wiedemanni is proposed as its pollinator. 

Two other species may be pollinated by the fly, S. wiedemanni. Salvia scabra has a straight tube (in 

a genus where the corolla tube is often curved) and grows in forest margins and bush clumps at the lower 

end of the distribution of S. wiedemanni. Salvia repens var. keiensis is a variable species with a straight 

upper lip (often hooded in the genus) and occurs in grassland or open woodland in the Eastern Cape. At 

the coastal sites S. wiedemanni is proposed as the pollinator in woodland areas, but at Naude's Nek Pr. 

ganglbaueri may be the grassland pollinator. 

Thorncroftia longiflora grows in rocky grassland and has a very narrow, long tube (30 - 38 mm) 

which closely matches the long proboscis of Pr. ganglbaueri in Mpumalanga (29 - 35 mm). Thorncroftia 

succulenta also occurs on rock outcrops in grassland, but the shorter tube length (15 - 20 mm) 

suggests that Pr. robusta may be its pollinator. 

No nemestrinid flies are recorded in the area where Hemizygia gerrardii (tube 17 - 20 mm) grows 

in grass among rocks, but the habitat suggests that a species of Prosoeca (such as Pr. robusta or Pr. 

ganglbaueri) may be its pollinator. The narrow endemic H. rugosifolia with a tube ca. 18 mm long 

(Codd 1985), is probably pollinated by Pr. robusta (proboscis length 20 - 23 mm). Hemizygia ramosa 

(tube 20 - 22 mm) grows among rocks in open woodland towards the coast and may be pollinated by S. 

wiedemanni. 

A number of the labiate genera that have long-tubed members, also have species that have 'mediumtubed' 

corollas in the range of 10 - 17 mm (table 1). 

In northern KwaZulu-Natal we observed visits of a horsefly, Philoliche aethiopica (family Tabanidae, 

proboscis length 8 - 9 mm), to flowers of Hemizygia pretoriae (Giirke) Ashby. The corolla tubes 

are of medium length (10 - 12 mm; Codd 1985), but other floral features are similar to those visited by 

long-proboscid flies. In the same area we also observed repeated visits by Philoliche aethiopica to 

flowers of Orthosiphon serratus Schltr., which has a medium corolla tube length of 9 - 16 mm (Codd 

1985). 

A similar situation occurs in Plectranthus where medium-tubed species are pollinated by a corresponding 

suite of medium-proboscid flies and bees (Potgieter & al. 1999). Plectranthus ecklonii Benth., 

for example, is visited by S. wiedemanni, shorter-proboscid flies (such as Philoliche aethiopica) and 

bees (Potgieter & al. 1999). In P. ecklonii and a number of other cases the filament and style lengths of 

these species approach that of the long-tubed species, but the shortened corolla tube allows for nectar 

exploitation by shorter-proboscid insects as well. 

The occurrence of a guild of medium-tubed Lamiaceae adapted for pollination by medium-proboscid 

flies may explain how long tubes could have evolved simply by elongation of the corolla tube, as the 

other floral features of medium-tubed species are already adapted to fly pollinators with extended 

proboscides. 

Since four species of long-proboscid flies with varying proboscis lengths occur in Mpumalanga, with 

Prosoeca spp. limited to grasslands and Stenobasipteron spp. occurring in forest or woodland, we will 

concentrate future field studies in this area to confirm which flies actually pollinate the species for which 

observations are not available. Studies are also needed to ascertain how closely pollinator proboscis 

lengths and floral tube lengths are correlated in specific localities, as there is a large amount of variation 

in these measurements (see Prosoeca ganglbaueri, table 3). 

It is, however, clear that long-tubed Lamiaceae are adapted for pollination by long-proboscid flies in 

the eastern parts of southern Africa. Furthermore, the habitat and distribution of the fly species determine 

which Lamiaceae are able to exploit this pollination syndrome. 

The production of long corolla tubes is energetically expensive and results in elevated water loss 

(Potgieter & Edwards unpubl. data). However, the advantages of this syndrome lie in the protection of 

500 

Chapter 4/ 52

Family Famil systematics sti C.J. Potgieter Potier & T.J. Edwards, Long, narrow crla corollas in Lamiaceae Lamia 

Table 4. Recorded natural hybrids of Plectranthus from Oribi Gorge. 

Natural hybrids are only derived from species with short or medium corolla tube lengths. Measurements are in mm. 

Standard deviation (SD) given after mean value. Sample size (n) = 20 for each species. Data from Potgieter & al. (2000). 

Hybrid Parent Plectranthus spp Tube length: Tube length: 

range 

mean (SD) 

P. zuluensis x P. ciliatus P. zuluensis T.Cooke 10 - 12 11.5 (0.6) 

P. ciliatus E.Mey. ex Benth. 6- 8 6.8 (0.7) 

P. oribiensis x P. ernstii P. oribiensis Codd 6.5 - 8 7.2 (0.5) 

P. ernstiiCodd 6 - 10.5 7.6 (1.2) 

P. ciliatus x P. oertendahlii P. ciliatus E.Mey. ex Benth. 6 - 8 6.8 (0.7) 

P. oertendahliiTh.Fries jun. 8 - 11.5 9.6 (1.2) 

nectar, which leads to heightened pollinator fidelity. Potgieter & al. (1999) outline the ability of 

pollinators to access nectar in seven species of Plectranthus. These results indicated that only S. 

wiedemanni is capable of reaching the nectar of long-tubed Plectranthus species. Extensive fieldwork 

a t Oribi Gorge N.R. has revealed natural hybrids between medium - and short-tubed Plectranthus 

species (table 4), but no natural hybrids have been recorded from species with long corolla tubes. 

This system would break down if there were areas where many long-tubed species co-occur, but the 

long-tubed Plectranthus species are mostly endemics that seldom co-occur. Where the more widespread 

P. ambiguus grows in Umtamvuna N. R. with P. hilliardiae and long-tubed forms of P. saccatus, the 

populations are all ecologically separated along vertical gradients along the slopes of the gorge, or occur 

either along the Umtamvuna River or one of its tributaries. 

Acknowledgments. - The authors would like to thank the following: National Research Foundation 

(NRF) for financial support; Mervyn Lotter of the Mpumalanga Parks Board for discussion; the 

KwaZulu-Natal Nature Conservation Services for access to and accommodation in their reserves; Centre 

for Electron Microscopy, University of Natal Pietermaritzburg for assistance with electron microscopy 

and photographs; Dr Fred Gess (Entomology Department, Albany Museum), Margie Cochrane (Entomology 

Collections Manager, South African Museum) and Drs Brian Stuckenberg & Dave Barraclough 

(Natal Museum) for insect identifications and insect distribution data; The National Botanical Institute 

for the use of data from the National Herbarium, Pretoria (PRE) Computerised Information System 

(PRECIS); Dr Steve Johnson for the use of his slide of Prosoeca ganglbaueri and access to a slide 

scanner; Dave Thompson, Mark Todd and Toni Boddington of the Cartographic Unit for assistance with 

maps and graphics. 

References 

Chapter 4/ 53 

Codd L.E. (1985) Lamiaceae. Flora S. Africa 28(4). 

Johnson S.D. & Johnson K. (1993) Beauty and the beast: a Cape orchid pollinated by horseflies. Veld and Flora (1975+) 79: 38-39. 

Johnson S.D. & Steiner K.E. (1995) Long-proboscid fly pollination of two orchids in the Cape Drakensberg mountains, South 

Africa. Plant Syst. Evol. 195: 169-175. 

Johnson S.D. & Steiner K.E. (1997) Long-tongued fly pollination and evolution of floral spur length in the Disa draconis complex 

(Orchidaceae). Evolution 51(1): 455-53. 

Goldblatt P. & Manning J.C. (1998) Gladiolus in Southern Africa. Vlaeberg, Fernwood Press. 

501

Syst. Geogr. Pl. 71 P1.71 XVIth AETFAT 

Syst.- Geogr.~ 

XVIth AETFAT Congress~- Congress 

Goldblatt P. & Manning J.C. (1999) The long-proboscid fly pollination system in Gladiolus (Iridaceae). Ann. Missouri Bot. Gard. 

86: 758-774. 

Goldblatt P. & Manning J.C. (2000) The long-proboscid fly pollination system in southern Africa. Ann. Missouri Bot. Gardl. 87: 

146-170. 

Goldblatt P., Manning J.C. & Bernhardt P. (1995) Pollination biology of Lapeirousia subgenus Lapeirousia (Iridaceae) in 

southern Africa; floral divergence and adaptations for long-tongued fly pollination. Ann. Missouri Bot. Gard. 82: 517-534. 

Manning J.C. & Goldblatt P. (1995) Cupid comes in many guises: the not-so -humble fly and pollination guild in the Overberg. Veld 

and Flora (1975+) 81: 50-52. 

Manning J.C. & Goldblatt P. (1996) The Prosoeca peringueyi (Diptera: Nemestrinidae) pollination guild in southern Afiica: longtongued 

flies and their tubular flowers. Ann. Missouri Bot. Gard. 83: 67-86. 

Manning J.C. & Goldblatt P. (1997) The Moegistorhynchus longirostris (Diptera: Nemestrinidae) pollination guild: long-tubed 

flowers and a specialized long-proboscid fly pollination system in southern Africa. Plant Syst. Evol. 206: 51-69. 

Manning J.C., Goldblatt P. & Winter P.J.D. (1999) Two new species of Gladiolus (Iridaceae: Ixioideae) from South Africa and 

notes on long-proboscid fly pollination in the genus. Bothalia 29: 217-223. 

Potgieter CJ., Edwards TJ., Miller R.M. & Van Staden J. (1999) Pollination of seven Plectranthus spp. (Lamiaceae) in southern 

Natal, South Africa. Plant Syst. Evol. 218: 99-112. 

Potgieter CJ., Edwards TJ. & Viljoen A. (2000) The significance of hybrids in South African species of Plectranthus (Lamiaceae). 

Poster presentation at XVIth AETFAT Congress August 28 - September 2, 2000. 

Rebelo A.G., Siegfried W.R. & Olivier E.G.H. (1985) Pollination syndromes of Erica species in the south-western Cape. S.A. Jnl. 

Bot. 51: 270-280. 

Struck M. (1997) Floral divergence and convergence in the genus Pelargonium (Geraniaceae) in southern Africa: ecological and 

evolutionary considerations. Plant Syst. Evol. 208: 71-97. 

Van Jaarsveld E.J. & Edwards T.J. (1991) Plectranthus reflexus. Fl. P1. Africa 51: Plate 2034. 

Van Jaarsveld E.J. & Edwards T.J. (1997) Notes on Plectranthus (Lamiaceae) from southern Africa. Bothalia 27: 1-6. 

Vos W.T., Edwards T.J. & Van Staden J. (1994) Pollination biology of annual and perennial Leonotis species (Lamiaceae). Plantzr 

Syst. Evol. 192: 1-9. 

Manuscript received November 2000; accepted in revised version February 2001. 

502 

Chapter 4/ 54

CHAPTER 5: 

A NEW POLLINATION GUILD 

Potgieter, C.J., Edwards, T.J., 2005. 

The Stenobasipteron wiedemanni (Diptera, Nemestrinidae) pollination 

guild in eastern southern Africa. 

Annals of the Missouri Botanical Garden 92: 254–267.

THE STENOBASIPTERON 

WIEDEMANNI (DIPTERA, 

NEMESTRINIDAE) 

POLLINATION GUILD IN 

EASTERN SOUTHERN 

AFRICA 1 

ABSTRACT 

ANN. MISSOURI BOT. GARD. 92: 254–267. 2005. 

C. J. Potgieter 2 and T. J. Edwards 2 

Stenobasipteron wiedemanni, a long-proboscid nemestrinid fly, services a Guild of flowers distinct from the Prosoeca 

ganglbaueri pollination Guild in which it has previously been placed. The former fly is the recorded pollinator of 19 

plant species in six families: the Acanthaceae, Balsaminaceae, Gesneriaceae, Iridaceae, Lamiaceae, and Orchidaceae. 

Stenobasipteron wiedemanni is a mainly forest-dwelling fly with a proboscis length of 19–30 mm. The plant species 

that are pollinated by this fly are also restricted to forest (or woodland) habitat along the eastern parts of southern 

Africa. The flowers of plants in this pollination Guild tend to have long, narrow nectaries and corollas in shades of 

purple, pink, mauve, pale blue, and white. Nectar is present in all species, pollination is diurnal, and most species 

have zygomorphic flowers with nectar guides. This pollination Guild is separated on the basis of the limitation of the 

fly, S. wiedemanni—and hence the plant species—to forest or closed-canopy habitat. Biogeographically, the S. wiedemanni 

Guild also occurs at lower altitudes in subtropical regions of southern Africa. 

Key words: Acanthaceae, Asystasia, Balsaminaceae, Barleria, Brownleea, forest, Gesneriaceae, Hesperantha, Hypoestes, 

Impatiens, Iridaceae, Isoglossa, Lamiaceae, long-proboscid flies, Nemestrinidae, Orchidaceae, Plectranthus, 

pollination, Stenobasipteron wiedemanni, Streptocarpus. 

The study of long-proboscid fly pollination is a 

subject of increasing interest in southern Africa. 

This pollination syndrome is well represented on 

the sub-continent and has been reported for a number 

of plant families: Ericaceae (Rebelo et al., 

1985), Iridaceae (Goldblatt et al., 1995; Goldblatt 

& Manning, 1999; Manning et al., 1999), Orchidaceae 

(Johnson & Steiner, 1995, 1997), Geraniaceae 

(Struck, 1997), and Lamiaceae (Potgieter et 

al., 1999; Potgieter & Edwards, 2001). This pollination 

syndrome was reviewed by Goldblatt and 

Manning (2000), and three discrete Guilds were 

identified. These are the Prosoeca peringueyi Guild, 

the Moegistorhynchus–Philoliche Guild, and the 

Prosoeca ganglbaueri Guild. 

The Prosoeca peringueyi Guild is restricted to the 

western half of the winter-rainfall area in southern 

Africa and comprises two fly species that pollinate 

flowers with intense shades of violet, deep purple, 

and magenta, often with cream to yellow markings 

and areas of darker pigmentation (Goldblatt & 

Manning, 2000). The Moegistorhynchus–Philoliche 

Guild comprises six to seven species of nemestrinid 

and tabanid flies, the tabanids extending from 

southern Namibia to southeastern parts of the Western 

Cape (see Fig. 1 for location of provinces in 

South Africa), with the nemestrinids being restricted 

to narrow ranges in the southwestern parts of the 

Western Cape. Flowers in this second Guild are 

white to cream with pink undertones, or pale to 

deep pink; nectar guides are red or deep pink 

(Goldblatt & Manning, 2000). These two Guilds occur 

in the winter-rainfall areas of southern Africa, 

while the Prosoeca ganglbaueri Guild (in which 

Stenobasipteron wiedemanni has been placed) occurs 

mainly in summer-rainfall areas. 

1 Financial support for this research was received from the NRF (National Research Foundation) and the URF 

(University of Natal Research Fund). The authors thank the following people and institutions: Riyad Ismail and Andrew 

Simpson of the Cartographic Unit, University of KwaZulu-Natal, for assistance with mapping; Cameron and Rhoda 

McMaster, Dave Thompson, Clinton Carbutt, and Carol Rolando for assistance with fieldwork; Brian Stuckenberg for 

identifying the fly species; Tracy McLellan, John Manning, Peter Goldblatt, Guy Upfold, and Geoff Nicholls for pollinator 

observations and information on S. wiedemanni; S. Piper for commenting on the manuscript; The National Botanical 

Institute for the use of data from PRECIS (National Herbarium, Pretoria (PRE) Computerised Information System); staff 

of the South African Museum (Cape Town), Albany Museum (Grahamstown), and Natal Museum (Pietermaritzburg) for 

making collector’s notes from fly specimens available; and two anonymous reviewers for helpful comments. 

2 School of Botany & Zoology, University of KwaZulu-Natal, Pietermaritzburg, P/Bag X01, Scottsville, 3209, KwaZulu- 

Natal Province, South Africa. potgietercj@ukzn.ac.za. 

Chapter 5/ 56

Volume 92, Number 2 

2005 

Potgieter & Edwards 

255 

Stenobasipteron wiedemanni Pollination Guild 

Figure 1. Map of South Africa showing provinces and study sites. Note that Northern Province is now called 

Limpopo Province. —Nk. Nkandla Forest. —Ng. Ongoye Forest. —K. Karkloof (Leopards Bush Nature Reserve). 

—H. Hlabeni Forest (Creighton). —O. Oribi Gorge Nature Reserve. —U. Umtamvuna Nature Reserve. —P. Port St. 

Johns. —S. Stutterheim (Kologha Forest). 

The Prosoeca ganglbaueri Guild sensu Goldblatt 

and Manning (2000) included four nemestrinid fly 

species: Prosoeca ganglbaueri Lichtwardt, Prosoeca 

longipennis Loew, Prosoeca robusta Bezzi, and Stenobasipteron 

wiedemanni Lichtwardt. Of these species, 

the floral Guilds pollinated by the first three 

show considerable overlap, but no overlap has been 

shown with the Guild pollinated by S. wiedemanni. 

Flowers pollinated by the first three fly species are 

pink with dark pink markings, but some species 

are cream or white or deep blue. The fourth fly 

species, S. wiedemanni, visits flowers in shades of 

pink, pale blue, or mauve (Goldblatt & Manning, 

2000). Goldblatt and Manning (2000) indicated that 

further research may show this latter species to 

constitute a separate Guild of pollinating flies, and 

in this paper we confirm the existence of this separate 

S. wiedemanni pollination Guild. 

Stenobasipteron wiedemanni (cited as Stenobasipteron 

sp. in Potgieter et al., 1999) is a brown ne- 

mestrinid fly species with a proboscis length of 19– 

30 mm. The species is largely limited to subtropical 

and temperate forests along the eastern seaboard of 

southern Africa (Potgieter & Edwards, 2001), and 

adults have been collected from December to June, 

with activity in each locality restricted to a few 

months during summer (see Appendix 1). By contrast, 

the nemestrinid genera Moegistorhynchus and 

Prosoeca occur in temperate fynbos, montane grasslands, 

and other habitats without a closed canopy. 

We proposed (Potgieter & Edwards, 2001) that 

there are a number of long-tubed species in genera 

of Lamiaceae in South Africa that are pollinated by 

either Stenobasipteron or Prosoeca, depending on 

habitat. Comparative plant and long-proboscid fly 

pollinator distributions are presented in Potgieter 

and Edwards (2001). These species are distributed 

over the eastern part of the country and include 

long-tubed members of Plectranthus (Potgieter et 

al., 1999), Hemizygia, Salvia, Stachys, Orthosiphon 

Chapter 5/ 57

256 Annals of the 

Missouri Botanical Garden 

(Goldblatt & Manning, 1999, 2000), and Thorncroftia. 

Goldblatt and Manning (2000) included 

species of Thorncroftia as inferred members of their 

Prosoeca ganglbaueri Guild. 

In forests of the eastern seaboard of southern Africa 

a number of plant families have evolved long, 

narrow corolla tubes, spurs, or hypanthia (ca. 20– 

35 mm), and in our study area we suggest that Stenobasipteron 

wiedemanni may be the primary pollinator 

of such species. Most of these plant species 

are unscented and have flowers of mauve, pink, 

purple, pale blue, or white. 

MATERIALS AND METHODS 

Fieldwork was conducted during the main flowering 

season of Plectranthus, between December 

and May, from 1995 to 2002. Study sites for pollinator 

observations are shown in Figure 1. We 

avoided over-collecting fly specimens, since the 

species occurs at low densities and we did not want 

to impact on the pollination service it renders. 

Where appropriate, observations of insect visits 

were followed by capture of voucher specimens that 

were killed in separate ethyl-acetate-containing vials 

and pinned with proboscides extending forward 

for easy measurement. Areas of pollen deposition 

were identified under a dissecting microscope, and 

samples of pollen were collected using blocks of 

Fuchsin Jelly and mounted. Slides were studied 

with a compound light microscope, and pollen samples 

were identified by comparison with reference 

slides of pollen from plants flowering in the area. 

Stenobasipteron wiedemanni is not easily confused 

with other insects in the study area, and thus 

only one (or rarely a few) vouchers were collected 

per plant species to ascertain areas of pollen deposition 

on the insect. After the first voucher per 

plant species was captured, careful observations of 

flies during subsequent flower visits allowed the observer 

to judge whether pollen would be deposited 

on certain areas of the insect. 

In cases where observations were not made by 

the authors of this paper, we relied on personal 

communication with other field workers that obtained 

photographic records or insect vouchers 

where possible. 

Insect vouchers for this study are lodged with the 

Natal Museum in Pietermaritzburg, South Africa, 

and plant vouchers are housed at NU, Natal University 

Herbarium (see Appendices 1 & 2). 

Distributions of the fly species were compiled 

from field observation and by using data from the 

Natal Museum (Pietermaritzburg), Albany Museum 

(Grahamstown), and the South African Museum 

(Cape Town). 

Distributions of Guild constituents were compiled 

from NU and PRECIS (National Herbarium, 

Pretoria (PRE) Computerised Information System), 

as well as published information (Codd, 1985; Pooley, 

1998; Linder, 1981; Hilliard & Burtt, 1986; Edwards, 

1988) and observations made during the 

course of this study. Distribution maps for shortertubed 

species of Plectranthus are available in Potgieter 

et al. (1999). 

Distribution maps were generated using ArcGIS, 

combining data from MS Access. Data were collected 

at quarter-degree-grid level, and hence the 

forest land cover was re-sampled to quarter-degreegrid 

resolution. 

Floral and other characters of plants shown to be 

effectively visited by Stenobasipteron wiedemanni 

were compared, and these data were used to suggest 

other possible members of the Guild. 

Nectar volume measurements were made during 

mornings from cultivated plants in a greenhouse, 

and these represent the maximum amount of nectar 

available to insect visitors. Nectar concentrations 

(as percentage of sucrose equivalents) were measured 

using a Bellingham & Stanley Eclipse (0– 

50%) hand-held refractometer. Nectar samples 

were dried on Whatmans no. 1 filter paper and analyzed 

using a modified Gas Chromatographic nectar 

sugar analysis method based on Tanowitz and 

Smith (1984). Samples of 10 to 30 flowers were 

pooled to obtain sufficient nectar for analysis. 

RESULTS 

THE POLLINATOR 

Stenobasipteron wiedemanni is a day-flying dipteran 

that visits flowers from about 9 A.M. to4P.M., 

extracting nectar as a reward. Specimens have been 

collected between December and July, but we 

found that in the Eastern Cape and KwaZulu-Natal 

activity continued to about May (depending on the 

year). 

The behavior of these flies makes them ideal pollinators 

of long-tubed flower species. During flight 

the proboscis is folded underneath the body, but as 

a flower is approached the proboscis is brought forward 

to extend in front of the body (Fig. 2A). Flies 

hover in front of flowers while probing for nectar, 

but occasionally grasp protruding filaments or 

styles (in the case of Lamiaceae) or rest on landing 

platforms (e.g., in Isoglossa hypoestiflora Lindau, 

Acanthaceae). Visits to individual flowers last between 

one and four seconds, and a number of flowers 

on the same inflorescence or plant may be 

Chapter 5/ 58


2005 


257 


Figure 2. —A. Stenobasipteron wiedemanni visiting a flower of Hypoestes aristata. The proboscis is extended forward 

and upward on approach of the flower. —B. S. wiedemanni, with pollinaria of Brownleea coerulea attached to the base 

of the proboscis. This specimen also carried pollen from Isoglossa hypoestiflora on the dorsal surface, from Hesperantha 

huttonii on the face and from Plectranthus on the ventral surface. Scale bars 10 mm. 

probed sequentially. The flies are highly mobile, 

strong fliers that can cover large distances in the 

forest understory. This is evident from the fact that 

we observed low frequency visitation at individual 

patches of studied plants. The relatively large body 

size (up to 20 mm long, excluding proboscis) offers 

many sites for pollen deposition (Fig. 2B). 

THE FLORAL GUILD 

Nineteen species from six plant families are currently 

included in the Stenobasipteron wiedemanni 

pollination Guild (Table 1). Members of this Guild 

have functional floral tube lengths that range from 

16 to 39 mm (Table 1). This length coincides with 

that of the proboscis of Stenobasipteron wiedemanni 

(19–30 mm). Shorter-tubed species that also belong 

to the Guild have tube lengths of 6–21 mm, but 

often have filaments that approach the length of the 

long-tubed species (indicated by 1 in Table 1). Plectranthus 

saccatus Benth. contains forms with variable 

corolla tube length, hence the emphasis on the 

long-tubed forms relevant to this study. 

Flowers of species that belong to this Guild are 

zygomorphic with flowers held horizontally, and 

where they tend to be more actinomorphic (e.g., 

Hesperantha), the floral tubes are held horizontally. 

Pollinators are thus directed to visit in a certain 

way. Nectar guides are present in many of the species. 

DISTRIBUTION OF THE GUILD 

The distribution of Stenobasipteron wiedemanni 

coincides with the Forest Biome along the eastern 

seaboard of South Africa (Fig. 3A). Similarly, mem- 

bers of the Guild coincide with forest and fly distribution 

(Figs. 3, 4). The presence of S. wiedemanni 

at Ongoye Forest, Oribi Gorge Nature 

Reserve (N.R.), Umtamvuna N.R., Ngeli Forest, 

Hlabeni Forest (Creighton), Hlatikhulu Forest, and 

Port St. Johns is based on new distribution records 

resulting from this study. 

The maps only include South Africa (excluding 

Swaziland and Mozambique), but it is likely that 

the fly distributions, and that of some of the plant 

species involved in the Guild, extend into these 

neighboring countries. 

FLOWER-POLLINATOR INTERACTION 

A comparison of areas of pollen deposition on 

Stenobasipteron wiedemanni shows that few of the 

long-tubed plant species share the same site on the 

insect (Fig. 5 and listed in Table 1). In cases where 

this does happen, the plants do not co-occur (except 

in a case of two Plectranthus species at one of 

the study sites). Even though we did not study Orthosiphon 

tubiformis R. D. Good, we included it in 

Figure 5 to show where pollen would be deposited 

on the insect. It shares the same areas of deposition 

on the insect as long-tubed Plectranthus species, 

but is allopatric. 

Nectar volumes range from 0.1 to 8.7 l per 

flower with high maximum volumes found in the 

longer-tubed species (Table 2). Nectar sugar concentrations 

range from 24% to 33% sucrose equivalents 

(Table 2), which falls within the range reported 

by Goldblatt and Manning (2000). Nectar 

tends to be sucrose dominant, i.e., sucrose:hexose 

ratio 1 (Table 2), with two species having sucrose 

Chapter 5/ 59



Table 1. Floral and other characters of plants pollinated exclusively or partially by Stenobasipteron wiedemanni (the 

‘‘fly’’). In five Plectranthus species with short and medium tubes, as well as in Hypoestes aristata (indicated with 1 ), the 

style and stamen filament lengths approach that of the tube lengths of long-tubed species. F forest, W highaltitude 

open woodland/wooded savanna. 2 narrow portion of floral tube, where tube length is longer, but sufficiently 

broad distally to allow access to a fly or bee body. See Appendix 2 for references for floral tube lengths and pollinator 

observations. 

Plant species (arranged by family) Habitat Flower color 

Length of 

floral tube 

spur (mm) 

Area of 

pollen 

deposition 

on fly 

Lamiaceae 

Plectranthus ambiguus F pinkish 20–33 ventral thorax 

purple 

& abdomen 

P. hilliardiae F pale 21–32 ventral thorax 

mauve 

& abdomen 

P. reflexus F pale blue 24–30 ventral thorax 

P. saccatus (long-tubed) F mauve to 

pale 

blue 

1P. ecklonii F bluish pur- 

& abdomen 

20–30 ventral thorax 

& abdomen 

Is the fly the 

only observed 

visitor to reach 

nectar? 

10–15 ventral head & no 

ple 

thorax 

1P. zuluensis F pale/dark 12–13 ventral head & no 

blue 

proboscis 

base 

1P. ciliatus F white 6–8 proboscis no 

1P. fruticosus F bluish 6–9 ventral head & no 

mauve 

thorax 

1P. praetermissus F blue to 

purple 

13–15 proboscis no 

Orthosiphon tubiformis 

Acanthaceae 

W pale pink 20–36 ventral thorax 

& abdomen 

yes 

Isoglossa hypoestiflora F mauve 24–27 dorsal thorax yes 

1Hypoestes aristata F bright pink 11–21 ventral thorax 

& abdomen 

no 

Barleria obtusa F edge blue to 212–18 ventral & lat- no 

Orchidaceae 

purple 

eral thorax 

Brownleea coerulea 

Balsaminaceae 

F mauve 20–24 base of proboscis 

yes 

Impatiens hochstetteri subsp. hochstetteri 

Gesneriaceae 

F pink (10–)16–24 

(–27) 

proboscis no 

Streptocarpus formosus 

Iridaceae 

F pale violet 231–36 dorsal body yes 

Hesperantha brevicaulis — pink 25–37 — — 

H. huttonii F purple 22–27; 33–39 ventral & lateral 

head & 

thorax 

no 

Gladiolus macneilii W pink 232–37 dorsal thorax yes 

yes 

yes 

no 

yes 

Chapter 5/ 60


2005 


259 


Figure 3. Distribution of forest biome, the fly (Stenobasipteron wiedemanni) and members of the S. wiedemanni 

pollination Guild in eastern South Africa. The forest biome is mapped at quarter-degree level for maps B–D (shaded 

squares). Stenobasipteron wiedemanni is shown in each map (). Plant species are shown on different maps. —A. 

Actual extent of the forest biome (black spots). —B. Plectranthus ambiguus (). —C. Brownleea coerulea (). —D. 

Impatiens hochstetteri subsp. hochstetteri (). 

Chapter 5/ 61



Figure 4. Distribution of forest biome, the fly (Stenobasipteron wiedemanni), members and inferred members of the 

S. wiedemanni pollination Guild in eastern South Africa. The forest biome is mapped at quarter-degree level for each 

map (shaded squares). Stenobasipteron wiedemanni is shown in each map (). Plant species are shown on different 

maps. —A. Isoglossa hypoestiflora (), Isoglossa cooperi (). —B. Plectranthus hilliardiae (), P. reflexus (), Hesperantha 

huttonii (). —C. Orthosiphon tubiformis (), P. saccatus—long-tubed forms (). —D. Asystasia varia (). 

Chapter 5/ 62


2005 


261 


Figure 5. Areas of pollen deposition (indicated by black shading on insect) on Stenobasipteron wiedemanni for 

floral forms in the Guild. —A. Isoglossa hypoestiflora (Acanthaceae). —B. Pollen placed dorsally on thorax. —C. 

Brownleea coerulea (Orchidaceae). —D. Pollinaria attached at base of head and proboscis. —E. Orthosiphon tubiformis 

(Lamiaceae). —F. Pollen placed ventrally on thorax. —G. Plectranthus hilliardiae (Lamiaceae). —H. Pollen placed 

ventrally on thorax. —I. Hesperantha huttonii (Iridaceae). —J. Pollen placed ventrally and on lower sides of thorax 

and head. —K. Impatiens hochstetteri subsp. hochstetteri (Balsaminaceae). —L. Plectranthus zuluensis (Lamiaceae). — 

M. Impatiens (K) pollen placed on proboscis; Plectranthus (L) pollen placed ventrally on head and thorax. Scale bar 

20 mm. Drawn by T. Edwards. 

Chapter 5/ 63



Table 2. Nectar sugars in species belonging to the Stenobasipteron wiedemanni pollination guild. G & M data 

from Goldblatt and Manning (2000); otherwise data are from the present study. The four Plectranthus species with short 

or medium tubes are indicated by 1 . 

Species 

Lamiaceae 

Orthosiphon tubiformis 

Plectranthus reflexus 


P. ambiguus 

P. saccatus (long tube) 

1 P. ecklonii 

1 P. ciliatus 

1 P. fruticosus 

1 P. zuluensis 

Nectar 

Volume l (n) Conc. % 

2.7–4.1 (10) 

0.9–3.2 (8) 

0.2–2.4 (40) 

— 

1.4–8.7 (10) 

— 

0.2–2.8 (25) 

— 

0.4–1.4 (11) 

24.5 

— 

31 

— 

4–18 

— 

29–33 

24 

— 

Range of sugars % 

Fru Glu Suc 

0–1 

1 

5 

24 

30 

9 

13 

19 

17–19 

0–5 

3 

4 

22 

29 

16 

15 

25 

21–23 

94–100 

96 

91 

54 

41 

75 

72 

56 

57–62 

Sucrose/ 

Fru Glu Reference 

25.2 

24.0 

10.11 

1.17 

0.70 

3.0 

2.57 

1.27 

1.5 

G&M 

Acanthaceae 

Isoglossa hypoestiflora 

Orchidaceae 

0.1–5.3 (27) — 18 33 49 0.96 


Iridaceae 

1.3–1.8 (2) 25–27 10 0–11 79–90 5.45 G & M 

Gladiolus macneilii 4.5–5.8 (5) 26 — — — — G & M 

rich nectar (sucrose:hexose ratio 0.5–0.99), using 

the categories of Baker and Baker (1990). 

NOTES ON POLLINATION IN EACH FAMILY 

LAMIACEAE 

In Plectranthus the deposition of pollen is sternotribic, 

since the bilabiate flowers have filaments 

and stamens that are ventrally aligned within the 

floral tube and lower lip. The long-tubed species 

force the fly, Stenobasipteron wiedemanni, to probe 

the floral tube fully, which maximizes pollen deposition 

and subsequent removal from the insect’s 

body. Shorter-tubed species of Plectranthus are also 

visited by the fly, but it is not the only pollinator 

(Potgieter et al., 1999). 

ACANTHACEAE 

The flowers of Isoglossa hypoestiflora have long, 

narrow tubes (up to 27 mm, Table 1) that widen to 

form a raised palate with petal sculpturing on the 

lower lip, and a hooded upper lip where the anthers 

and stigma are positioned. Pollen transfer is nototribic 

on the thorax and abdomen of the fly, with a 

few grains placed on the proboscis. Isoglossa cooperi 

C. B. Cl. may be conspecific with I. hypoestiflora. 

The species were mapped separately in accordance 

with current taxonomy (Clarke, 1912). 

Hypoestes aristata (Vahl) Sol. ex Roem. & Schult. 

has a distribution greater than that of the fly (ex- 

tending into coastal parts of the Western Province 

and into Tropical Africa), and it is not exclusively 

pollinated by Stenobasipteron wiedemanni; xylocopid 

bees, tabanid flies, and papilionoid butterflies 

have been observed as floral visitors (G. Nicholls 

and G. Upfold, pers. comm.). Insects access nectar 

from the lilac, bilabiate, medium-tubed flowers 

(tubes up to 21 mm, Table 1), with filament and 

style lengths that reach 34 mm (measured at NU) 

that are crowded into verticillate inflorescences 

(Balkwill & Getliffe Norris, 1985). Barleria obtusa 

Nees was seen to be visited by S. wiedemanni at 

Oribi Gorge (F. Field, pers. comm.), but bees also 

visit this species. 

ORCHIDACEAE 

Brownleea coerulea Harv. ex Lindl. is the only 

orchid seen to be pollinated by Stenobasipteron wiedemanni. 

We studied it at Stutterheim, where the 

fly accesses nectar from the slightly recurved spur 

(20–24 mm long, Table 1) and picks up pollinaria 

at the base of the proboscis. 

BALSAMINACEAE 

Impatiens hochstetteri Warb. subsp. hochstetteri is 

a widespread African species ranging from northern 

Sudan to South Africa. It grows in shaded moist 

places—often in forests—and flowers throughout 

the year, with seasonal flowering in some areas 

Chapter 5/ 64


2005 

(Grey-Wilson, 1980). Corollas are pink with a 

curved calyx spur (10–)16–24 mm long (Grey-Wilson, 

1980), but specimens from our study areas in 

KZN (measured at NU) have spur lengths 16–27 

mm long. We have seen a number of visits by the 

fly, but never managed to collect a voucher. One 

was caught by M. Byrne at Nkandla Forest after 

observing visits to I. hochstetteri subsp. hochstetteri 

(McLellan, pers. comm.), but very little pollen remained 

on the voucher. Impatiens L. pollen adheres 

dorsally to the proboscis of the fly, since the short 

spur and sessile anthers prevent deposition elsewhere 

on the insect’s body. The throat of the spur 

is exceedingly narrow, and the sessile anthers are 

situated directly above this aperture. 

GESNERIACEAE 

Streptocarpus formosus (Hilliard & B. L. Burtt) T. 

J. Edwards is pollinated by Stenobasipteron wiedemanni 

(J. Manning, pers. comm.). Flies enter the 

flower fully, since the 40–55 mm long corolla (Weigend 

& Edwards, 1994) is only narrow in the lower 

part (31–36 mm, Table 1). Pollen deposition is nototribic 

on the fly thorax. The pale violet corollas 

have dark stippling in the throat, which acts as a 

nectar guide, and yellow pigmentation on the floor 

of the corolla tube. It is a narrow endemic in forest 

habitats from Umtamvuna to the area just north of 

Oribi Gorge (NU records), and observations were 

made at Umtamvuna in January 2002. 

IRIDACEAE 

Hesperantha huttonii (Baker) Hilliard & Burtt is 

a day-flowering species with purple flowers and 

narrow, straight perianth tubes. Flowers have exserted 

anthers and perianth tube lengths that vary 

from 22 to 39 mm, which fall into two size classes 

(Table 1); those with longer tubes occur at the Katberg, 

but specimens collected outside of the Katberg 

(in the Amatole Mountains) have shorter tubes 

(Hilliard & Burtt, 1986), such as the plants studied 

at Stutterheim in this study (tubes 24–27 mm long). 

We found Hesperantha pollen on Stenobasipteron 

wiedemanni individuals caught in two different 

years at this site, but we never saw the visits. Our 

voucher plant specimen of H. huttonii had grains 

of Plectranthus pollen on the styles and anthers, 

which shows that pollinators that visited a nearby 

Plectranthus also visited H. huttonii. Subsequent to 

this observation, P. Goldblatt (pers. comm.) recorded 

S. wiedemanni visiting flowers of H. huttonii, 

confirming our indirect observations. 


263 


DISCUSSION 

The suite of characteristic floral attributes associated 

with pollination by Stenobasipteron wiedemanni 

includes long, narrow floral tubes or spurs 

(Table 1); flower colors in shades of purple, mauve, 

pale blue, pink, and white; both subtropical and 

montane forest habitats; and the presence of nectar 

guides as darker spots or lines in certain species 

(e.g., Brownleea coerulea, Hypoestes aristata, Plectranthus 

hilliardiae Codd, P. ecklonii Benth., P. zuluensis 

T. Cooke, P. ciliatus E. Mey. ex Benth., P. 

praeternissus Codd, P. fruticosus L’Hér., P. ambiguus 

(Bolus) Codd, Orthosiphon tubiformis, and Streptocarpus 

formosus). 

Nectar rewards are offered in all studied species 

and nectar is sucrose dominant or rich. Nectar concentrations 

are similar to those previously recorded 

for long-proboscid fly pollinated plants (Goldblatt 

& Manning, 2000). 

Natural forest habitat in southern Africa is largely 

restricted to the eastern parts of the sub-continent, 

extending as far west as the eastern section 

of the Western Cape Province (not fully shown in 

Figs. 3, 4) (Low & Rebelo, 1996). Observations and 

museum records show that members of the Stenobasipteron 

wiedemanni Guild are largely limited to 

forested areas. 

In Mpumalanga Province this Guild does not operate 

in the same kind of closed-canopy, moist forest. 

Orthosiphon tubiformis is visited by S. wiedemanni 

in rocky grassland that appears to have been 

recently transformed from more wooded to more 

open vegetation (Goldblatt, pers. comm.), and Gladiolus 

macneilii Oberm. occurs in more wooded savanna 

or forest margins (Manning et al., 1999). 

Nevertheless, there is still some tree cover and a 

similarity in temperature regimes. The KwaZulu- 

Natal and Eastern Cape forests where S. wiedemanni 

is found are warm coastal or cool mist-belt 

forests, while the Mpumalanga observations are 

from cooler, higher altitudes. 

At least eight plant species within this pollination 

Guild appear to rely solely on Stenobasipteron 

wiedemanni for pollination (indicated in Table 1 as 

having the fly as the only observed visitor to reach 

nectar). In Plectranthus reflexus Van Jaarsv. & T. J. 

Edwards, an opportunist butterfly visitor was seen 

(Potgieter & Edwards, 2001), but it is likely that 

the corolla design is geared toward long-proboscid 

fly pollination. In P. ambiguus visits by two species 

of long-proboscid bees were seen (Potgieter et al., 

1999), but they are unable to reach nectar in most 

flowers. Some specimens have shorter corolla tubes 

(20 mm is the lower end of the range, Table 1) and 

Chapter 5/ 65



these may allow exploitation by bees. Other species 

in the Guild (with shorter corolla tubes) are not 

solely visited by S. wiedemanni, but still need to 

be included since the fly provides an efficient pollination 

service (Potgieter et al., 1999). 

DISCUSSION BY FAMILY 

LAMIACEAE 

Pollination of Plectranthus hilliardiae, P. reflexus, 

P. ambiguus, and long-tubed forms of P. saccatus 

has previously been discussed in Potgieter et 

al. (1999) and Potgieter and Edwards (2001). In 

most areas where Stenobasipteron wiedemanni has 

been observed, Plectranthus species (long- or shorttubed) 

tend to form a major component of the forest 

understory. Orthosiphon tubiformis occurs in different 

habitat to forest understory Plectranthus species 

and has been discussed by Goldblatt and Manning 

(2000). 

ACANTHACEAE 

Honeybees collect pollen from Isoglossa hypoestiflora, 

but the long, narrow corolla tube prohibits 

nectar access. Honeybees would not contribute to 

significant pollen carryover over large distances, 

since pollen is groomed into the scopae before the 

worker bees return to their hives. Isoglossa cooperi 

closely resembles I. hypoestiflora in floral shape 

and size, differing only in the presence of glandular 

hairs on the calyx and bracts. The species are often 

sympatric (Fig. 4A). From a pollination perspective, 

these two species are very similar and fit the Stenobasipteron 

wiedemanni pollination Guild. They occur 

in afromontane and scarp forest and are generically 

anomalous with respect to flower color and 

corolla morphology. Most Isoglossa Oerst. species 

in South Africa are hymenophilous (pers. obs.), 

with creamy white flowers and short corolla tubes 

(mostly less than 10 mm in length). These creamy 

white species are common in savanna and tropical 

and subtropical forests. 

The observation of visits to Hypoestes aristata 

shows that Stenobasipteron wiedemanni may form 

part of a more generalized suite of pollinators in 

species with shorter floral tubes. As in shortertubed 

Plectranthus species, the fly still contributes 

to pollen carryover, since the elongate anther filaments 

deposit pollen on the body of the fly and the 

elongate style is able to remove pollen from this 

position. A similar situation exists in Barleria obtusa, 

where long-proboscid bees can also access 

nectar by crawling into the widened distal part of 

the corolla tube. 

ORCHIDACEAE 

The record for long-proboscid fly pollination in 

Brownleea Harv. ex Lindl. (Goldblatt & Manning, 

2000) highlights the disjunction between grasslandand 

forest-growing species. Brownleea macroceras 

Sond. is pollinated by Prosoeca ganglbaueri (Johnson 

& Steiner, 1995), which occurs in montane 

grassland at high altitudes. By contrast, B. coerulea 

is a forest margin species of lower altitudes and is 

pollinated by Stenobasipteron wiedemanni. Brownleea 

coerulea is a widespread species (Fig. 3C) that 

also occurs in Madagascar, where its pollination has 

not been studied. 

BALSAMINACEAE 

The floral visits by Stenobasipteron wiedemanni 

to Impatiens hochstetteri subsp. hochstetteri are the 

first published records of nemestrinid fly pollination 

in the Balsaminaceae. This plant species is 

widespread across Africa, but the distribution of S. 

wiedemanni elsewhere in Africa is not known. 

Grey-Wilson (1980) recorded butterflies as pollinators 

of the ‘‘flat type’’ flowers in the section to 

which I. hochstetteri subsp. hochstetteri belongs, yet 

little information is available on the pollination of 

specific Impatiens L. (Grey-Wilson, 1980). We have 

observed papilionoid butterfly visits to I. hochstetteri 

subsp. hochstetteri at Ferncliffe Nature Reserve 

in Pietermaritzburg (KZN) confirming the above. 

The distribution of Impatiens hochstetteri subsp. 

hochstetteri in southern Africa (Fig. 3D) shows a 

number of plots outside forested areas. This species 

relies on moist forest habitats and may occur in 

scrub forest patches that are below the resolution 

of our Geographic Information System that was 

used for mapping. 

GESNERIACEAE 

Very little is known about pollination in Streptocarpus 

(Hilliard & Burtt, 1971), and it is likely 

that Stenobasipteron wiedemanni is the pollinator of 

a number of Streptocarpus species that conform to 

the floral morphology of S. formosus. 

IRIDACEAE 

The genus Hesperantha comprises many scented, 

pale species with crepuscular anthesis that are 

thought to be pollinated by moths (Goldblatt, 1984). 

Contrary to this pattern, H. huttonii is odorless, has 

diurnal anthesis and colored flowers—attributes 

suited to nemestrinid fly pollination. Long-proboscid 

fly pollination of day-flowering species of Hesperantha 

is not uncommon. Hesperantha latifolia 

Chapter 5/ 66


2005 

(Klatt) M. P. de Vos is pollinated by Prosoeca peringueyi 

in Namaqualand, while H. grandiflora G. J. 

Lewis, H. scopulosa Hilliard & B. L. Burtt, and Hesperantha 

cf. woodii Baker are pollinated by P. ganglbaueri 

in grasslands in the eastern parts of the 

country (Goldblatt & Manning, 2000). In Mpumalanga 

Hesperantha brevicaulis (Baker) G. J. Lewis is 

pollinated by Stenobasipteron wiedemanni (Goldblatt 

& Manning, 2000). The switch between moth and 

nemestrinid pollination was recently also recorded 

in Zaluzianskya F. W. Schmidt (Scrophulariaceae) by 

Johnson et al. (2002). 

INFERRED MEMBERS OF THE STENOBASIPTERON 

WIEDEMANNI GUILD 

There are a number of other plant species that 

may belong to the Stenobasipteron wiedemanni pollination 

Guild. Labiate species such as Stachys tubulosa 

MacOwan, Salvia repens Burch. ex Benth. var. 

keiensis Hedge, Salvia scabra L.f., and Hemizygia 

ramosa Codd have been suggested as candidates 

(Potgieter & Edwards, 2001). 

Manning et al. (1999) suggested that Gladiolus 

sekukuniensis P. Winter and G. saxatalis Goldblatt 

& J. C. Manning (both Iridaceae) are pollinated by 

Stenobasipteron wiedemanni in the highlands of 

Mpumalanga and Northern (Limpopo) Province, 

since they have similar morphology and phenology 

to G. macneilii and occur in forest margins and 

wooded savannas. 

Streptocarpus species (Gesneriaceae) that may 

conform to this Guild include S. rexii (Bowie ex 

Hook.) Lindl. (forests of the Western and Eastern 

Cape Provinces, from Knysna to Kokstad), S. primulifolius 

Gand. (forests of the Eastern Cape Province 

and KwaZulu-Natal), and S. cyaneus S. Moore 

(forests of Mpumalanga Province). 

Pelargonium transvaalensis (Geraniaceae) is a 

forest species of the eastern seaboard. It is likely 

that Stenobasipteron wiedemanni is the pollinator, 

since the hypanthial tubes are ca. 20 mm long and 

pink in color. 

Rhinacanthus gracilis Klotzsch and Asystasia 

varia N. E. Br. (both Acanthaceae) are also likely 

to belong to this Guild; R. gracilis was inferred as 

a member of the Prosoeca ganglbaueri pollination 

system of Goldblatt and Manning (2000) that included 

Stenobasipteron wiedemanni. Rhinacanthus 

gracilis is a subtropical species of forest and woodland; 

it has mauve to white narrow floral tubes ca. 

20 mm long. Similarly, A. varia has mauve flowers 

with tubes 35–40 mm long of which the basal half 

to two-thirds is narrowed, and the lower floral lobe 

has a raised palate with violet nectar guides (Ed- 


265 


wards, 1988). It occurs in forest patches ranging 

from northern KwaZulu-Natal to the Eastern Cape, 

a distribution coinciding with that of Stenobasipteron 

wiedemanni (Fig. 4D). Asystasia gangetica T. 

Anderson and A. pinguifolia T. J. Edwards are beepollinated 

species of grasslands (Edwards, 1988) 

with shorter tubes and white flowers. 

Asystasia varia may indicate a shift from bee to 

long-proboscid fly pollination associated with elongation 

of the floral tube and a flower color shift from 

creamy white to mauve. A similar shift may have 

happened in Isoglossa (also Acanthaceae), where 

most species have creamy white flowers with short 

corolla tubes (adapted to bee pollination), but I. 

hypoestiflora and I. cooperi have mauve flowers with 

longer tubes. Without cladistic analyses we can 

only speculate about such shifts. In the Iridaceae 

it has been shown that long-proboscid fly flowers 

are often closely related to ancestors pollinated by 

long-tongued bees (Goldblatt & Manning, 2000). 

CONCLUSIONS 

The Stenobasipteron wiedemanni pollination 

Guild should be recognized separately from the 

three Guilds suggested by Goldblatt and Manning 

(2000). There is no overlap in the plant species 

pollinated by this fly compared to the other three 

Guilds. On the basis of habitat there is no or very 

little overlap in the distribution of the fly species. 

Unlike the mostly pink flowers with dark pink to 

red markings pollinated by the Prosoeca ganglbaueri 

Guild (sensu Goldblatt & Manning, 2000), 

the S. wiedemanni Guild consists of flowers with 

shades of purple, mauve, pale blue, pink, and 

white. In cases where flower color is similar, e.g., 

the blue-flowered Nivenia stenosiphon Goldblatt 

and pale pink-flowered Nerine Herb. and Brunsvigia 

Heist. species that were included in the Prosoeca 

ganglbaueri Guild of Goldblatt and Manning 

(2000), there is no overlap in plant species at the 

microgeographic scale. The S. wiedemanni pollination 

Guild is thus limited by the (forest) distribution 

of the fly, rather than by floral color or shape. 

We have confirmed the existence of long-proboscid 

fly pollination in the Acanthaceae, as suggested 

by Goldblatt and Manning (2000) and report it here 

for the first time in the families Balsaminaceae and 

Gesneriaceae. 

Literature Cited 

Baker, H. G. & I. Baker. 1990. The predictive value of 

nectar chemistry to the recognition of pollinator types. 

Israel J. Bot. 39: 157–166. 

Balkwill, K. & F. Getliffe Norris. 1985. Taxonomic studies 

Chapter 5/ 67



in the Acanthaceae; The genus Hypoestes in southern 

Africa. S. African J. Bot. 51: 13–144. 

Clarke, C. B. 1912. Isoglossa (Acanthaceae). In: W. T. Thiselton-Dyer 

(editor), Fl. Capensis 5: 79–84. L. Reeve, 

Kent. 

Codd, L. E. 1985. Plectranthus (Lamiaceae). Fl. S. Africa 

28(4): 137–172. 

Edwards, T. 1988. Asystasia varia. Fl. Pl. Africa 50: Plate 

1976. 

Goldblatt, P. 1984. A revision of Hesperantha (Iridaceae) 

in the winter rainfall area of southern Africa. S. African 

J. Bot. 50: 15–141. 

& J. C. Manning. 1999. The long-proboscid fly 

pollination system in Gladiolus (Iridaceae). Ann. Missouri 

Bot. Gard. 86: 758–774. 

& . 2000. The long-proboscid fly pollination 

system in southern Africa. Ann. Missouri Bot. 

Gard. 87: 146–170. 

, & P. Bernhardt. 1995. Pollination biology 

of Lapeirousia subgenus Lapeirousia (Iridaceae) in 

southern Africa; Floral divergence and adaptation for 

long-tongued fly pollination. Ann. Missouri Bot. Gard. 

82: 517–534. 

Grey-Wilson, C. 1980. Impatiens of Africa. A. A. Balkema, 

Rotterdam. 

Hilliard, O. M. & B. L. Burtt. 1971. Streptocarpus: An 

African Plant Study. Univ. Natal Press, Pietermaritzburg. 

& . 1986. Hesperantha (Iridaceae) in Natal 

and nearby. Notes Roy. Bot. Gard. Edinburgh 43: 407– 

438. 

Johnson, S. D. & K. E. Steiner.1995. Long-proboscid fly 

pollination of two orchids in the Cape Drakensberg 

Mountains, South Africa. Pl. Syst. Evol. 195: 169–175. 

& . 1997. Long-tongued fly pollination 

and evolution of floral spur length in the Disa draconis 

complex (Orchidaceae). Evolution 51: 455–53. 

, T. J. Edwards, C. Carbutt & C. J. Potgieter. 2002. 

Appendix 1. Vouchers of Stenobasipteron wiedemanni (Diptera, Nemestrinidae) lodged at Natal Museum, Pietermaritzburg. 

KZN KwaZulu-Natal Province, E Cape Eastern Cape Province. 

Collector’s no. Locality (all in South Africa) Plant Date 

C. Potgieter 1 















F. Field s.n. 

M. Byrne s.n. 

KZN: Umtamvuna 



KZN: Oribi Gorge 


KZN: Karkloof, Leopards Bush 

KZN: Karkloof, Leopards Bush 

KZN: Ongoye Forest 

E Cape: Port St Johns 

E Cape: Stutterheim, Kologha Forest 







KZN: Nkandla Forest 

Specialization for hawkmoth and long-proboscid fly pollination 

in Zaluzianskya section Nycterinia (Scrophulariaceae). 

Bot. J. Linn. Soc. 138: 17–27. 

Linder, H. P. 1981. Taxonomic studies on the Disinae: 1. 

A revision of the genus Brownleea Lindl. S. African J. 

Bot. 47: 13–48. 

Low, A. B. & A. G. Rebelo (Editors). 1996. Vegetation of 

South Africa, Lesotho and Swaziland: A companion to 

the vegetation map of South Africa, Lesotho and Swaziland. 

Department of Environmental Affairs & Tourism, 

Pretoria. 

Manning, J. C., P. Goldblatt & P. J. D. Winter. 1999. Two 

new species of Gladiolus (Iridaceae: Ixioideae) from 

South Africa and notes on long-proboscid fly pollination 

in the genus. Bothalia 29: 217–223. 

Pooley, E. 1998. A Field Guide to Wild Flowers: Kwa- 

Zulu-Natal and the Eastern Region. Natal Flora Publications 

Trust, Durban. 

Potgieter, C. J. & T. J. Edwards. 2001. The occurrence of 

long, narrow corolla tubes in southern African Lamiaceae. 

Syst. & Geogr. Pl. 71: 493–502. 

, , R. M. Miller & J. Van Staden. 1999. 

Pollination of seven Plectranthus spp. (Lamiaceae) in 

southern Natal, South Africa. Pl. Syst. Evol. 218: 99– 

112. 

Rebelo, A. G., W. R. Siegfried & E. G. H. Olivier. 1985. 

Pollination syndromes of Erica species in the southwestern 

Cape. S. African J. Bot. 51: 270–280. 

Struck, M. 1997. Floral divergence and convergence in 

the genus Pelargonium (Geraniaceae) in southern Africa: 

Ecological and evolutionary considerations. Pl. 

Syst. Evol. 208: 71–97. 

Tanowitz, B. D. & D. M. Smith. 1984. A rapid method for 

qualitative and quantitative analysis of simple carbohydrates 

in nectars. Ann. Bot. 53: 453–456 

Weigend, M. & T. J. Edwards. 1994. Notes on Streptocarpus 

primulifolius (Gesneriaceae). S. African J. Bot. 60: 

168–169. 

Plectranthus ambiguus 

P. ambiguus 

P. ambiguus 

P. ciliatus 

P. zuluensis 

P. fruticosus 

in spider web 

on ground (with I. hypoestiflora pollen) 

P. praetermissus 

P. ciliatus 

P. ciliatus 

P. ecklonii 

P. ecklonii 

P. ecklonii 


Barleria obtusa 

Impatiens hochstetteri 

17 Mar. 1995 

18 Mar. 1996 

15 Mar. 1996 

25 Mar. 1997 

25 Mar. 1997 

3 Mar. 1999 

3 Mar. 1999 

22 Apr. 1999 

9 Mar. 1998 

5 Mar. 1998 

5 Mar. 1998 

5 Mar. 1998 

5 Mar. 1998 

29 Feb. 2000 

29 Feb. 2000 

5 Apr. 1997 

31 Jan. 2001 

Chapter 5/ 68


2005 


267 


Appendix 2. Vouchers for studied plant species lodged at NU Herbarium, Pietermaritzburg. References for pollinator 

observations and/or floral tube measurements are given where this has been recorded in addition to observations in the current 

study. P & E Potgieter & Edwards, P et al. Potgieter et al., G&M Goldblatt & Manning, G-W Grey-Wilson, W 

&E Weigend & Edwards. KZN KwaZulu-Natal Province, E Cape Eastern Cape Province, PMB Pietermaritzburg. 

Plant species (arranged by family) Collector’s no. 

Locality (all in 

South Africa) 

Reference of 

previous work 

Lamiaceae 

Plectranthus ambiguus (Bolus) 

Codd 

C. Potgieter 86 KZN: Umtamvuna P & E (2001) 

P. hilliardiae Codd C. Potgieter 110, 111, 

112 

KZN: Umtamvuna P & E (2001) 

P. reflexus Van Jaarsv. & T. J. Ed- T. Edwards 1554 E Cape: Port St P & E (2001) 

wards 

P. saccatus Benth. (long-tubed) C. Potgieter 107, 120 KZN: Umtamvuna P & E (2001) 

P. ecklonii Benth. C. Potgieter 65, 114 KZN: Oribi Gorge P et al. (1999) 

C. Potgieter 70 KZN: Umtamvuna 

P. praetermissus Codd T. Edwards 1556 E Cape: Port St 

Johns 

P. zuluensis T. Cooke C. Potgieter 64, 118 KZN: Oribi Gorge P et al. (1999) 

P. ciliatus E. Mey. ex Benth. C. Potgieter 67 KZN: Oribi Gorge P et al. (1999) 

C. Potgieter 68, 69, 116 KZN: Umtamvuna 

C. Potgieter 136 KZN: PMB, Ferncliffe 

P. fruticosus L’Hér. C. Potgieter 66 KZN: Oribi Gorge 

C. Potgieter 134 KZN: PMB, Ferncliffe 

C. Potgieter 144 KZN: Karkloof, 

Leopards Bush 

Orthosiphon tubiformis R. D. Good 

Acanthaceae 

G & M (2000); P 

& E (2001) 

Isoglossa hypoestiflora Lindau C. Potgieter s.n. KZN: Oribi Gorge G. Nicholls (pers. 

C. Potgieter & D. Thompson 

735 

Johns 

KZN: Creighton, 

Hlabeni Forest 

comm.) 

Hypoestes aristata (Vahl) Sol. ex 

G. Nicholls (pers. 

Roem. & Schult. 

comm.) 

Barleria obtusa Nees 

Orchidaceae 

F. Field (pers. 

comm.) 

Brownleea coerulea Harv. ex Lindl. 

Balsaminaceae 

Edwards & Potgieter 1841 E Cape: Stutterheim, 

Kologha Forest 

G & M (2000) 

Impatiens hochstetteri Warb. subsp. C. Potgieter & D. Thomp- KZN: Creighton, McLellan (pers. 

hochstetteri 

Gesneriaceae 

son 744 

Hlabeni Forest comm.); G-W 

(1980) 

Streptocarpus formosus (Hilliard & 

W & E (1994); 

B. L. Burtt) T. J. Edwards 

Iridaceae 

Manning (pers. 

comm.) 

Hesperantha brevicaulis (Baker) G. 

J. Lewis 

G & M (2000) 

H. huttonii (Baker) Hilliard & B. L. Edwards & Potgieter E Cape: Stutterheim, Goldblatt (pers. 

Burtt 

1841a 

Kologha Forest comm.) 

Gladiolus macneilii Oberm. G & M (1999) 

Chapter 5/ 69

CHAPTER 6: 

NATURAL HYBRIDS 

Viljoen, A.M., Demirci, B., Baser, K.H.C., Potgieter, C.J., Edwards, T.J., 

2006. 


hybridisation in Plectranthus (Lamiaceae). 

South African Journal of Botany 72: 99–104.

Microdistillation and essential oil chemistry—a useful tool for detecting 

hybridisation in Plectranthus (Lamiaceae) 

Abstract 

A.M. Viljoen a, *, B. Demirci b , K.H.C. Bas¸er b , C.J. Potgieter c , T.J. Edwards c 

a School of Pharmacy, Tshwane University of Technology, Private Bag X680, Pretoria, 0001, South Africa 

b Department of Pharmacognosy, Faculty of Pharmacy, Anadolu University, 26470 Eskis¸ehir, Turkey 

c School of Botany and Zoology, University of KwaZulu-Natal Pietermaritzburg, Private Bag X01, Scottsville 3209, South Africa 

Received 26 November 2004; accepted 27 May 2005 

The essential oil composition is reported for Plectranthus ciliatus, Plectranthus zuluensis and their putative hybrid. The essential oil chemistry 

is in support of morphological data and pollination studies, which have indicated a natural hybrid between P. ciliatus and P. zuluensis. The hybrid 

plant contains terpenoids from both putative parents together with Fhybrid compounds,_ which are not present in any of the two parents. The 

composition of the essential oil obtained through microdistillation is virtually identical to the analysis of the hydrodistilled essential oil. 

D 2005 SAAB. Published by Elsevier B.V. All rights reserved. 

Keywords: Plectranthus; Essential oil; Chemotaxonomy; Hybridisation 

1. Introduction 

In southern Africa, the Lamiaceae are most abundantly 

represented by the genus Plectranthus with 48 indigenous 

species (Hankey, 1999; Van Jaarsveld and Edwards, 1997; 

Edwards et al., 2000). This group of aromatic Labiates have 

been the subject of various taxonomic studies (Codd, 1975, 

1985; Van Jaarsveld and Edwards, 1991, 1997; Edwards et al., 

2000). The species in some complexes are notoriously difficult 

to identify. Pollination studies in the genus have revealed a 

number of natural hybrids amongst the medium-tubed species 

of Plectranthus, particularly at Oribi Gorge Nature Reserve 

(3030CA) in southern KwaZulu-Natal, South Africa. Potgieter 

et al. (2000) illustrated that the leaf and flower morphology of 

the putative Plectranthus ciliatus Plectranthus zuluensis 

hybrid appear to be intermediate to that of the sympatric 

putative parents. P. ciliatus E. Mey. ex Benth is the only 

species in the study area that has nectar guides on both upper 

and lower corolla lips; thus, the presence of this character in the 

hybrid plant indicates putative parentage in P. ciliatus P. 

zuluensis. Hybrids of P. ciliatus P. zuluensis show a general 

* Corresponding author. 

E-mail address: viljoenam@tut.ac.za (A.M. Viljoen). 

South African Journal of Botany 72 (2006) 99 – 104 

0254-6299/$ - see front matter D 2005 SAAB. Published by Elsevier B.V. All rights reserved. 

doi:10.1016/j.sajb.2005.05.003 

Chapter 6/ 71 

www.elsevier.com/locate/sajb 

habit and purple colouration on the lower leaf surface that is 

intermediate to that of the putative parent species. Corolla 

shape, size and colouration are also intermediate to that of the 

putative parents. Furthermore, two different dipteran species 

were seen to visit and move between the inflorescences of 

sympatric populations of P. zuluensis T. Cooke and P. ciliatus 

at Oribi Gorge; these are Stenobasipteron wiedemanni, a longproboscid 

fly (family Nemestrinidae), and Psilodera confusa, a 

medium-proboscid fly (family Acroceridae) illustrated in 

Potgieter et al. (1999). 

Hybrids are not unknown in this genus of horticulturally 

interesting plants and many crosses have been made by plant 

breeders (E. van Jaarsveld, personal communication). Natural 

hybridisation has a pronounced impact on the understanding of 

taxonomic relationships (reticulate vs. divergent evolution). As 

hybrids could often defy detection by morphology alone, 

additional information is required to detect and/or confirm 

possible hybridisation events. The aromatic character of 

Plectranthus leaves allows for the extraction and analysis of 

the essential oils. The aim of this investigation was to 

determine if essential oil chemistry could be used to confirm 

hybridisation in Plectranthus. As hybrid plants are often scarce 

and the essential oil yields for Plectranthus species is generally 

very low, microdistillation was investigated as an alternative 

method to extract the essential oils and compare it to the

100 

A.M. Viljoen et al. / South African Journal of Botany 72 (2006) 99–104 

Chapter 6/ 72 

Table 1 

Essential oil composition for (1) P. ciliatus (ex hort Witwatersrand Botanical Garden), (2) P. ciliatus (Ferncliff), (3A) P. ciliatus P. zuluensis (ex Oribi Gorge), 

hydrodistilled oil, (3B) P. ciliatus 

Oribi Gorge) 

RRI : Relative Retention Indices 

P. zuluensis (ex Oribi Gorge), microdistilled oil, (4) P. zuluensis (ex hort Witwatersrand Botanical Garden), (5) P. zuluensis (ex 

RRI Compound 1 2 3A 3B 4 5 

1032 a-Pinene 0.26 – 0.08 1.06 0.02 0.02 

1035 a-Thujene – – – 0.08 – – 

1076 Camphene – – – 0.04 – – 

1118 h-Pinene 0.01 0.01 2.44 18.01 0.01 0.01 

1132 Sabinene 0.01 0.03 0.08 0.45 0.01 – 

1159 y-3-Carene 0.12 – 0.47 2.49 – – 

1174 Myrcene 0.09 – – – 0.01 – 

1176 a-Phellandrene – – 0.08 0.62 – – 

1187 o-Cymene – – 0.01 0.13 – – 

1188 a-Terpinene – – 0.01 0.48 – – 

1195 Dehydro-1,8-cineole – – – 0.03 – – 

1203 Limonene 3.80 0.04 0.11 0.56 0.01 0.01 

1205 Sylvestrene 0.02 – – – – – 

1213 1,8-Cineole – 0.01 – – 0.01 0.01 

1218 h-Phellandrene – – 0.07 0.66 – – 

1224 o-Mentha-1(7),5,8-triene – – – 0.05 – – 

1225 (Z)-3-Hexenal – – – 0.04 – – 

1232 (E)-2-Hexenal – – – – – 0.02 

1244 Amyl furan (=2-Pentyl furan) – – – 0.02 – – 

1246 (Z)-h-Ocimene 1.27 – 0.02 0.09 – – 

1255 g-Terpinene 0.02 – 0.13 3.58 – – 

1266 (E)-h-Ocimene 0.42 – 0.01 0.05 – – 

1278 m-Cymene 0.01 – 0.10 0.31 – – 

1280 p-Cymene 0.12 0.02 1.18 2.01 – – 

1286 Isoterpinolene – – 0.02 0.13 – – 

1290 Terpinolene 0.07 – 0.02 0.28 – – 

1319 (E)-2,6-Dimethyl-1,3,7-nonatriene 0.01 – – 0.01 0.01 – 

1345 3-Octyl acetate – 0.07 0.02 – – 0.04 

1360 Hexanol – – – – – 0.02 

1382 cis-Alloocimene 0.02 – – – – – 

1386 Octenyl acetate 0.76 4.25 0.06 0.09 – – 

1391 (Z)-3-Hexenol – – – – – 0.02 

1393 3-Octanol – 0.03 0.02 0.01 0.05 0.05 

1400 Nonanal 0.01 – – 

1406 a-Fenchone – – – – 0.01 – 

1412 (E)-2-Hexenol 0.01 – – – – – 

1415 Rose furan 0.02 – – – – – 

1452 a,p-Dimethylstyrene 0.01 – 0.03 0.14 – – 

1452 1-Octen-3-ol 1.11 1.83 0.71 0.13 0.07 0.13 

1466 a-Cubebene 0.15 0.10 0.05 0.04 0.04 0.11 

1468 trans-1,2-Limonene epoxide 0.07 – – – – – 

1474 trans-Sabinene hydrate 0.16 0.02 0.02 0.03 – – 

1479 y-Elemene – 0.09 – – – – 

1476 (Z)-h-Ocimene epoxide 0.01 – – – – – 

1478 cis-Linalool oxide (Furanoid) – – – 0.01 – – 

1495 Bicycloelemene 0.07 0.43 0.04 0.03 0.21 0.03 

1492 Cyclosativene 0.90 – 0.08 0.04 – – 

1493 a-Ylangene – 1.23 – – – – 

1497 a-Copaene 2.25 0.48 1.53 0.76 1.16 2.97 

1528 a-Bourbonene 0.01 0.22 0.03 – 0.29 0.22 

1532 Camphor – – – 0.02 – – 

1535 h-Bourbonene 0.21 3.08 0.36 0.05 3.85 2.65 

1544 a-Gurjunene 1.00 0.08 0.07 0.04 0.03 – 

1545 cis-a-Bergamotene – 0.08 0.06 – – – 

1548 (E)-2-Nonenal – – – – 0.02 – 

1549 h-Cubebene 0.26 0.28 0.44 0.18 0.60 1.25 

1553 Linalool 0.21 0.30 0.98 2.14 1.24 6.21 

1565 Linalyl acetate – – – – – 0.06 

1568 1-Methyl-4-acetylcyclohex-1-ene – 0.04 0.03 – – – 

1571 trans-p-Menth-2-en-1-ol 0.03 0.01 0.01 0.01 – –

Table 1 (continued) 

Chapter 6/ 73 

A.M. Viljoen et al. / South African Journal of Botany 72 (2006) 99–104 101 



1572 h-Ylangene 0.06 0.51 0.09 0.03 – 0.39 

1586 Pinocarvone – – 0.03 0.09 – – 

1594 trans-h-Bergamotene – 0.34 0.21 0.09 – – 

1597 Bornyl acetate 0.10 5.48 0.02 0.09 – – 

1598 Thymol methyl ether (=Methyl thymol) – – – 0.04 – – 

1599 (E,Z)-2,6-Nonadienal – – – – 0.01 – 

1600 h-Elemene 0.81 1.16 0.56 0.25 0.38 0.36 

1611 Terpinen-4-ol – – – 0.01 – – 

1612 h-Caryophyllene 9.29 2.64 3.35 3.90 6.19 7.45 

1614 Carvacrol methyl ether (=Methyl carvacrol) – – – 0.06 – – 

1617 6,9-Guaiadiene 0.12 0.61 – – – – 

1638 cis-p-Menth-2-en-1-ol – – 0.02 0.04 – – 

1648 Myrtenal – – 0.04 0.11 – – 

1628 Aromadendrene 0.05 0.20 – – – – 

1639 trans-p-Mentha-2,8-dien-1-ol 0.06 – – – – – 

1650 g-Elemene 0.02 0.48 – – – – 

1658 Sabinyl acetate – – 1.44 2.51 – – 

1661 Alloaromadendrene 3.82 1.12 – – 0.40 – 

1664 trans-Pinocarveol – – 0.08 0.11 – – 

1668 (Z)-h-Farnesene – 0.03 0.12 0.04 0.02 0.02 

1674 g-Gurjunene – 0.20 – – 0.54 0.23 

1677 epi-Zonarene 0.16 – – – – – 

1684 h-Guaiene – 0.20 – – – – 

1687 a-Humulene 0.86 0.35 8.19 6.99 21.40 9.24 

1698 Myrtenyl acetate – – 0.53 0.88 – – 

1700 p-Mentha-1,8-dien-4-ol (=Limonen-4-ol) – – – 0.06 – – 

1704 g-Muurolene 0.26 0.40 – – – – 

1706 a-Terpineol 0.17 – 0.08 0.51 – 0.11 

1708 Ledene 2.09 – – – – – 

1709 a-Terpinyl acetate – – – 0.06 – – 

1726 Germacrene D 3.10 9.37 6.41 5.52 7.11 6.53 

1740 a-Muurolene 0.38 – – – – – 

1741 h-Bisabolene – 1.32 0.74 0.29 – – 

1743 Eremophilene – – – – 0.06 – 

1747 trans-Carvyl acetate 2.54 – 0.10 0.13 – – 

1755 Bicyclogermacrene 3.54 16.82 1.73 1.30 7.28 1.41 

1773 y-Cadinene 11.77 0.33 1.28 0.88 0.17 0.28 

1776 g-Cadinene – 0.25 0.06 – 0.02 0.02 

1782 cis-Carvyl acetate 0.34 – – 2.00 – – 

1784 (E)-a-Bisabolene – 5.97 8.10 2.56 – – 

1786 Kessane – – – 0.13 – – 

1798 Methyl salicylate – – – 0.16 – – 

1799 Cadina-1,4-diene (=Cubenene) 0.15 – – – – – 

1804 Myrtenol – – 0.24 0.07 – – 

1808 Nerol – – – – – 0.03 

1810 3,7-Guaiadiene 0.08 – – – – 

1830 2,6-Dimethyl-3(E),5(E),7-octatriene-2-ol 0.05 – – – – – 

1838 h-Damascenone – – – 0.03 – – 

1845 trans-Carveol 0.22 – 0.18 0.09 – – 

1853 cis-Calamenene 0.39 – – – – – 

1853 Dehydrocostuslactone – – 10.61 6.86 – – 

1854 Germacrene-B – 2.60 – – – – 

1857 Geraniol – – – – 0.01 0.08 

1864 p-Cymen-8-ol 0.10 – 0.91 0.95 – – 

1871 p-Mentha-1,8-dien-10-yl acetate 0.01 – – – – – 

1882 cis-Carveol 0.04 – – 0.02 – – 

1900 epi-Cubebol 6.19 0.10 0.05 – 0.02 0.05 

1941 a-Calacorene 0.26 0.03 – – – – 

1945 1,5-Epoxy-salvial-4-14-ene – 0.16 – – 0.06 0.02 

1953 Palustrol 1.07 0.04 – 0.01 – – 

1957 Cubebol 1.95 0.27 0.07 – – 0.05 

(continued on next page)

102 

Table 1 (continued) 



1958 h-Ionone – – – 0.02 – – 

1981 (Z)-Methyl cinnamate – – – – 0.01 – 

1984 g-Calacorene 0.27 – – – – – 

2001 Isocaryophyllene oxide – 0.14 0.10 – 0.05 0.03 

2008 Caryophyllene oxide 1.50 1.13 0.55 0.37 0.19 0.17 

2030 Methyl eugenol – – 0.13 – 8.56 2.93 

2033 Epiglobulol – 0.80 – – – – 

2037 Salvial-4(14)-en-1-one – 0.12 0.04 – – – 

2045 Humulene epoxide-I – – – – 0.10 0.02 

2050 (E)-Nerolidol – – 0.32 0.14 0.16 0.06 

2057 Ledol 7.42 0.11 0.21 0.12 – – 

2069 1,6-Germacradien-5h-ol (=Germacrene D-4h-ol), 

(=1(10),5-Germacradien-4h-ol) 

2.24 0.46 – – – – 

2071 Humulene epoxide-II – – 0.79 0.42 0.71 0.16 

2080 Cubenol 0.93 – – – – 0.02 

2081 Humulene epoxide-III – – – – 0.05 0.01 

2088 1-epi-Cubenol 1.18 – – – – – 

2096 Elemol – 1.52 2.37 1.23 – – 

2096 (E)-Methyl cinnamate – – – – 0.02 – 

2103 Guaiol – – 0.23 0.17 – – 

2104 Viridiflorol 0.55 0.27 – – 0.03 – 

2109 cis-Methylisoeugenol – – – – 0.51 0.78 

2127 10-epi-g-Eudesmol – – 9.00 6.50 – – 

2144 Spathulenol 2.23 15.52 1.11 0.36 0.60 0.09 

2185 g-Eudesmol – – 0.15 0.51 – – 

2187 T-Cadinol 3.38 – – – – – 

2200 trans-Methylisoeugenol – – – – 0.62 2.79 

2202 1,6-Germacradien-5a-ol (=Germacrene D-4a-ol), 

(=1(10),5-Germacradien-4a-ol) 

0.15 – – – – – 

2219 y-Cadinol 0.68 – – – – – 

2209 T-Muurolol 1.26 0.22 – – – – 

2232 a-Bisabolol – 1.33 1.33 0.46 – – 

2245 Elemicine – – 4.30 2.16 10.56 0.63 

2247 trans-a-Bergamotol 0.16 0.86 – – – – 

2255 a-Cadinol 2.99 0.66 – – – 0.04 

2257 h-Eudesmol – – 0.40 0.38 – – 

2282 g-Asarone – – 0.74 0.20 14.93 35.76 

2361 (Z)-Asarone – – – – – 2.48 

2403 trans-Isoelemicine – – – 0.02 2.79 1.54 

2478 (E)-Asarone – – – – 0.15 3.66 

2324 Caryophylla-2(12),6(13)-dien-5a-ol 

(=Caryophylladienol-II) 

0.06 – – – – – 

2438 Kaur-16-ene 0.07 – – – – – 

2607 14-Hydroxy-y-cadinene 0.02 – – – – – 

2676 epi-13-Manool 0.37 – 1.32 0.03 – – 

Total (%) 89.00 86.85 77.70 85.72 91.36 91.27 

composition of hydrodistilled oils. Although the essential oil 

composition has been reported for some species of Plectranthus 

(Ascensão et al., 1998; Buchbauer et al., 1993; Smith 

et al., 1996), research on the South African representatives of 

this genus has been neglected. This paper forms part of a 

broader investigation on the composition, chemotaxonomy and 

biological activity of the essential oil of South African 

Plectranthus species. 

2. Materials and methods 

Leaf material of P. ciliatus, P. zuluensis and their putative 

hybrid were collected from various localities (Table 1). Only 


Chapter 6/ 74 

one clone of the putative hybrid was present; hence, only one 

collection could be made. Voucher specimens of P. ciliatus 

(Ferncliff) C. Potgieter 745, P. ciliatus P. zuluensis (Oribi 

Gorge) C. Potgieter 556 and P. zuluensis (Oribi Gorge) C. 

Potgieter 64 are deposited at the University of KwaZulu–Natal 

Herbarium (UN). Vouchers of P. ciliatus (WBG) AV 78 and P. 

zuluensis (WBG) AV 23 are housed in the Department of 

Pharmacy and Pharmacology, University of the Witwatersrand. 

The essential oils were obtained through hydrodistillation (3 h) 

using a Clevenger apparatus. In addition, leaves collected from 

the hybrid plant (P. ciliatus P. zuluensis) were air dried and 

subjected to microdistillation using the following method; 

crushed leaves (¨500 mg) were placed in a sample vial

together with 10 ml of water. NaCl (2.5 g) and water (0.5 ml) 

were placed in the collecting vial. n-Hexane (300 Al) was 

added into the collecting vial to trap volatile components. 

Sample vials were heated to 108 -C at a rate of 20 -C/min and 

then kept at 108 -C for 90 min, heated to 112 -C at a rate of 20 

-C/min and kept at this temperature for 30 min. Finally the 

samples were subjected to post-run for 6 min under the same 

conditions. Collecting vials were cooled to 1 -C during 

distillation. After distillation was completed, the organic layer 

in the collection vial was analyzed by GC/MS. 

The essential oils were analysed using a Hewlett-Packard 

G1800A GCD system. Innowax FSC column (60 m 0.25 mm 

Ø, with 0.25 Am film thickness). Helium (0.8 ml/min) was used 

as carrier gas. GC oven temperature was kept at 60 -C for 10 min 

and programmed to 220 -C at a rate of 4 -C/min and then kept 

constant at 220 -C for 10 min to 240 -C at a rate of 1 -C/min. 

Mass range was recorded from m/z 35 to 425. Split ratio was 

50:1 and the splitless mode was used for essential oil samples 

obtained from micro-distillation. Injection port temperature was 

at 250 -C. MS were taken at 70 eV. Relative percentage amounts 

of the separated compounds were calculated automatically from 

peak areas of the total ion chromatogram. Library searches were 

carried out using the Wiley GC/MS Library and the Bas¸er 

Library of Essential Oil Constituents. Cluster analysis was 

carried out with the NTSYS-PC package version 2.00 (Rohlf, 

1997). The entire data set as presented in Table 1 was converted 

to qualitative data (presence/absence) and a hierarchical cluster 

analysis was performed using the UPGMA clustering algorithm. 

3. Results and discussion 

The compounds identified in the essential oil for all the taxa 

analyzed are summarized in Table 1. Eighty-three compounds 

were identified in the essential oil of P. ciliatus (89%) from the 

Witwatersrand Botanical Garden and sixty-two compounds 

(86.9%) were identified in the essential oil obtained from P. 

ciliatus from Ferncliff. Although quantitative variations are 

apparent, the essential oil composition from the two plants 

growing some ca. 600 km apart is similar as indicated in the 

dendrogram presented (Fig. 1). Both samples contained 6,9guaiadiene, 

aromadendrene, g-elemene, g-muurolene, a-calacorene, 

1,6-germacradien-5h-ol, T-muurolol, trans-a-bergamotol 

and a-cadinol. These compounds are absent in both P. 

zuluensis and the putative P. ciliatus P. zuluensis hybrid. The 

high correlation coefficient (Fig. 1) suggests that the oil 

composition of the two samples obtained for P. zuluensis are 

congruent. In both instances, a-humulene, germacrene D and 

g-asarone are the major compounds. 

The morphological data and field observations presented in 

the Introduction prompted us to investigate the essential oil as 

an independent test to confirm hybridisation events in 

Plectranthus. Demarne and Van der Walt (1989) and others 

(Emboden and Lewis, 1967; Kokkini, 1992; Viljoen and Van 

Wyk, 2001) have illustrated the value of using phytochemical 

data to confirm the origin of natural hybrids. Certain 

compounds found only in P. ciliatus are present in the hybrid 

(e.g., ledol) while some terpenoids found only in P. zuluensis 

0.45 0.63 0.81 

similarity coefficient 

are present in the hybrid (e.g., g-asarone). A chemotaxonomic 

survey of the essential oils of South African species of 

Plectranthus shows that g-asarone is restricted to P. zuluensis 

(Maistry, 2001). Compounds found in both parents are present 

in the hybrid (e.g., linalool). FHybrid compounds_ are absent in 

both parents and only found in the hybrid, e.g., 10-epi-geudesmol. 

It could be hypothesised that new compounds form 

when enzymes which were previously mutually exclusive in 

the two parents combine in the hybrid forming Fnew hybrid 

phytochemicals_. Recently Viljoen and Van Wyk (2001) and 

Viljoen (1999) demonstrated very similar chemical patterns for 

phenolic compounds in artificial and natural hybrids of the 

genus Aloe. The cluster analysis (Fig. 1) shows the hybrid plant 

nested between the putative parents which is indicative of 

chemical Fcharacters_ shared with both P. ciliatus and P. 

zuluensis. Fig. 1 also clearly indicates that the composition of 

the hydrodistilled oils and the oil obtained through microdistillation 

are virtually identical. Microdistillation is a very 

quick and accurate method to extract the oils from small 

amounts of low-yielding leaf material such as Plectranthus 

hybrids. 

The importance of hybridisation was noticed by Linnaeus 

who proposed a model of speciation through hybridisation. 

Lotsy (1916, 1931) identified hybridisation as the most 

important factor in evolutionary change. Hybrids do, however, 

pose a problem to systematics as it is often believed that Fgood 

species_ do not hybridise. Many species concepts consider the 

process of natural hybridisation at best to be nonexistent and 

that they make taxonomic treatments difficult (Arnold, 1997). 

Since the cladistic method is based on the assumption of 

divergence, it is imperative to determine the role of hybridisation 

in any taxonomic study as this could lead to reticulate 

phylogenies. The example presented here provides evidence 

that natural hybridisation occurs in Plectranthus, which is 

instrumental in unraveling the evolutionary history and 

taxonomy of this genus. 

References 

P. ciliatus (WBG) 

P. ciliatus (Ferncliff) 

P. ciliatus X P. zuluensis (OG, hydrodistilled) 

P. ciliatus X P. zuluensis (OG, microdistilled) 

P. zuluensis (OG) 

P. zuluensis (WBG) 

Chapter 6/ 75 

A.M. Viljoen et al. / South African Journal of Botany 72 (2006) 99–104 103 

Fig. 1. Dendrogram constructed on qualitative data (absence/presence) using all 

compounds in Table 1. WBG=Witwatersrand Botanical Garden, Ferncliff=- 

Ferncliff Nature Reserve, Pietermaritzburg, OG=Oribi Gorge Nature Reserve. 

Arnold, L., 1997. Natural Hybridization and Evolution. Oxford University 

Press, New York. 

Ascensão, L., Figuiredo, A.C., Barroso, J.G., Pedro, L.G., Schripsema, J., 

Deans, S.G., Scheffer, J.C., 1998. Plectranthus madagascariensis: morphology 

of the glandular trichomes, essential oil composition, and its 

biological activity. International Journal of Plant Sciences 159, 31–38.

104 

Buchbauer, G., Jirovetz, L., Wasicky, M., Nikiforov, A., 1993. Volatile 

constituents of the headspace and essential oil of Plectranthus coleoides 

Marginatus (Labiatae). Journal of Essential Oil Research 5, 311–313. 

Codd, L.E., 1975. Plectranthus (Labiatae) and allied genera in southern Africa. 

Bothalia 11, 371–442. 

Codd, L.E., 1985. Plectranthus. Flora of Southern Africa 28, 137–172. 

Demarne, F.E., Van der Walt, J.J.A., 1989. Origin of the rose-scented 

Pelargonium cultivar grown on Réunion Island. South African Journal of 

Botany 55, 184–191. 

Edwards, T.J., Paton, A., Crouch, N.R., 2000. A new species of Plectranthus 

(Lamiaceae) from Zimbabwe. Kew Bulletin 55, 459–464. 

Emboden, W.A., Lewis, H., 1967. Terpenes as taxonomic characters in Salvia 

section Audibertia. Brittonia 19, 152–160. 

Hankey, A., 1999. The genus Plectranthus (Lamiaceae) in South Africa: 

diagnostic characters and simple field keys. Plant Life 21, 5–15. 

Kokkini, S., 1992. Essential oils as taxonomic markers in Mentha. In: Harley, 

R.M., Reynolds, T. (Eds.), Advances in Labiate Science. Royal Botanic 

Gardens, Kew. 

Lotsy, J.P., 1916. Evolution by Means of Hybridization. M. Nijhoff, The 

Hague. 

Lotsy, J.P., 1931. On the species of the taxonomist in its relation to evolution. 

Genetica 13, 1 – 6. 

Maistry, K., 2001. The antimicrobial properties and chemical composition of 

leaf essential oils of indigenous Plectranthus (Lamiaceae) species. MSc 

research report, University of the Witwatersrand, South Africa. 


Chapter 6/ 76 

Potgieter, C.J., Edwards, T.J., Viljoen, A.M., 2000. The significance of hybrids 

in South African species of Plectranthus. Poster presentation: XVIth 

AETFAT Congress, National Botanic Garden of Belgium, Meise, 28 

August–2 September. 




Rohlf, F.J., 1997. NTSYSpc-2.00. Department of Ecology and Evolution, State 

University of New York. 

Smith, R.M., Bahaffi, S.O., Albar, H.A., 1996. Chemical composition of the 

essential oil of Plectranthus tenuiflorus from Saudi Arabia. Journal of 

Essential Oil Research 8, 447–448. 

Van Jaarsveld, E.J., Edwards, T.J., 1991. Plectranthus reflexus. Flowering 

Plants of Africa 51 (Plate 2034). 

Van Jaarsveld, E.J., Edwards, T.J., 1997. Notes on Plectranthus (Lamiaceae) 

from southern Africa. Bothalia 27, 1 – 6. 

Viljoen, A.M., 1999. A chemotaxonomic study of the phenolic leaf 

compounds in the genus Aloe. PhD thesis, Rand Africans University, 

South Africa. 

Viljoen, A.M., Van Wyk, B.-E., 2001. A chemotaxonomic and morphological 

appraisal of Aloe series Purpurascentes, Aloe section Anguialoe and 

their hybrid, Aloe broomii. Biochemical Systematics and Ecology 29, 

621–631.

CHAPTER 7: NECTAR STUDIES 

Nectar is primarily composed of water and sugars, containing lesser amounts of a 

variety of possible compounds, e.g. amino acids, lipids, antioxidants, alkaloids, 

proteins, vitamins, inorganic ions, organic acids, phenolics and terpenoids (Nicolson & 

Thornburg 2007). It constitutes the main calorific reward to almost all known pollinators 

(Dafni 1992). Nectar is a reward that is not part of the sexual system of the plant, thus 

a tight correlation between pollinator demands and behaviour, and the composition, 

amount and rhythm of secretion of nectar, is not surprising. It is widely accepted that 

nectar plays an important role in plant-pollinator interactions and this reflects the coevolution 

between plants and pollinators (Dafni 1992). Since nectar characteristics tend 

to be similar for plants that are visited by the same groups of animals, nectar may show 

differences between related plants that have different pollinators (Kearns & Inouye 

1993). 

The nectary of Plectranthus, as in the rest of the Lamiaceae, is gynobasic, with an 

asymmetrical disc at the base of the four-lobed ovary (Dafni et al. 1988, Petanidou et 

al. 1999). Nectar is secreted at the base of a tubular corolla and visiting insects require 

proboscides of sufficient length to access it. 

In the current study, (A) nectar sugar composition was analysed, and (B) nectar sugar 

concentration and nectar volume were measured, for a selection of species. 

A. Nectar Sugar Composition 

Nectar sugars are derived from sucrose that is translocated in phloem sap. The 

enzyme invertase (located in nectaries) determines the final composition of nectar by 

hydrolysing sucrose to glucose and fructose; the extent of this reaction determines the 

nectar sugar ratio (Nicolson & Fleming 2003). Sucrose and the hexose sugars fructose 

and glucose are the most common nectar sugars, the occurrence and relative 

proportions of which tend to remain constant in any one species (Percival 1961), but 

which may show wide inter-specific differences (Baker & Baker 1983b). The ratios in 

which these sugars occur have traditionally been correlated with the type of pollinator 

involved (Baker & Baker 1983a, 1990; Percival 1961). Percival (1961, 1965) showed 

that sucrose-dominated nectars are common in long-tubed flowers pollinated by 

bumble- and honey bees (i.e. long-tongued bees), as well as butterflies and moths. In

contrast, shallow flowers with relatively unprotected nectar tend to produce hexosedominated 

nectars. 

The following nectar classes were identified by Baker & Baker (1983a) on the basis of 

sucrose to hexose ratios, with corresponding % sucrose indicated as per Nicolson & 

Thornburg (2007): 

‘Sucrose-dominant’: sucrose:hexose ratio > 1; sucrose 51 – 100 % 

‘Sucrose-rich’: sucrose:hexose ratio 0.5 – 1; sucrose 34 – 50 % 

‘Hexose-rich’: sucrose:hexose ratio 0.1 – 0.5; sucrose 10 – 33 % 

‘Hexose-dominant’: sucrose:hexose ratio < 0.1; sucrose 0 – 9 % 

Baker & Baker (1983a) tabulated insect pollinator preference for the four nectar classes 

and showed that hawkmoths prefer ‘sucrose-dominant’ and ‘sucrose-rich nectars’, 

while other Lepidoptera such as settling moths, butterflies and skippers tolerate a 

greater range from ‘sucrose-dominant and -rich’, to ‘hexose-rich’ nectar. The latter is 

also true for long-tongued bees, with ‘sucrose-dominant’ nectar being most popular. 

Short-tongued bees and butterflies, on the other hand, show a strong preference for 

‘hexose-dominant and -rich’ nectar. Wasps and beetles show a more or less equal 

preference for ‘sucrose- or hexose-dominant’ nectar (observation based on a low 

sample size), while flies seem to prefer ‘hexose-dominant and -rich nectar’, with a few 

preferring higher sucrose ratios (Baker & Baker 1983a). 

Percival (1961) warned that the observed pattern of certain pollinator classes preferring 

certain types of nectar may be overridden by the plant family to which the species 

belongs, hence a phylogenetic effect may operate. This is echoed by Dafni (1992), Van 

Wyk (1993) and Nicolson & Van Wyk (1998) who caution that, in addition to 

considering pollinator demands, one should also consider plant phylogeny when 

studying the components of nectar, since genetic constraints limit the base on which 

selection can act. 

The recent review on nectar chemistry by Nicolson & Thornburg (2007) cautions 

against using the terminology (outlined above) suggested by Baker & Baker (1983a), 

since it over-emphasises sucrose; describing nectar sugar ratios as the percentage 

sugar composition is preferable (Nicolson & Thornburg 2007). Both practises are 

followed in this chapter, with new results presented as % sucrose and interpreted w.r.t. 

Chapter 7/ 78

Baker & Baker’s (1983a) categories, while discussions involving past studies will utilise 

the nectar sugar categories in inverted commas, for ease of comparison. 

B. Nectar Volume and Concentration 

Knowledge of nectar volume, concentration, composition and spatial distribution is 

considered important for the interpretation of pollinator behaviour, as well as 

understanding pollinator energetics and nutrient requirements (Kearns & Inouye 1993). 

A study of honeybee attraction in nine species of Lamiaceae was conducted by Dafni 

et al. (1988) with the intention of correlating a number of floral traits (including nectar 

volume and concentration) and the rate of attraction of honeybees, but the results 

failed to establish any clear relationships. Experimental manipulations of irrigation and 

fertilisation showed some surprising results in a study on three species of Lamiaceae in 

the phrygana, with volume and sugar content varying considerably between species 

and treatments (Petanidou et al. 1999). Nectar production was studied under natural 

and experimental conditions in three species of Lamiaceae by Macukanovic-Jocic et al. 

(2004), and results showed variation in volume and concentration in response to 

microclimate (habitat) and other factors. 

The techniques used to collect nectar, and the limitations imposed by small flowers 

with small nectar volumes in particular, introduce possible sources of variation in nectar 

volume and concentration estimates. It is for this reason that Nicolson & Thornburg 

(2007) caution against the practise of attributing ecological significance to nectar 

concentration, especially in cases where averages are used. 

Materials and Methods 

A. Nectar Sugar Analysis 

The nectar sugar ratios of 14 species of Plectranthus were analysed, as well as an 

additional seven varieties or forms of species. Two species that fall within the broader 

concept of Plectranthus (Paton et al. 2004) were included: one Pycnostachys (Py. 

urticifolia Hook.) and one Thorncroftia N.E.Br. (T. longiflora N.E.Br), as well as five 

other species of Lamiaceae for comparison: three Orthosiphon Benth. (O. tubiformis 

R.D.Good [=Ocimum tubiforme (R.D.Good) A.J.Paton], O. labiatus N.E.Br. [=Ocimum 

labiatum (N.E.Br.) A.J.Paton], O. serratus Schltr. [=Ocimum serratum (Schltr.) 

A.J.Paton]), and two Hemizygia (Benth.) Briq. (H. albiflora (N.E.Br.) M.Ashby 

Chapter 7/ 79

[=Syncolostemon albiflorus (N.E.Br.) D.F.Otieno], H. incana Codd [=Syncolostemon 

incanus (Codd) D.F.Otieno]). One member of the Acanthaceae, Isoglossa hypoestiflora 

Lindau, was also analysed since it falls within the nemestrinid fly (Stenobasipteron 

wiedemanni Lichtwardt, 1910) pollination syndrome. 

Sample preparation and analysis followed that outlined by Tanowitz & Smith (1984), 

with some modifications. Nectar was sampled by pulling the corolla tubes from the 

calyces of freshly harvested flowers and squeezing the tubes gently to force nectar 

from their bases. These droplets were spotted onto Whatman’s No. 1 filter paper and a 

circle was drawn in pencil to mark the spots before they dried. Nectar was allowed to 

air-dry and filter paper was then stored in separate stamp collector’s envelopes in a 

container with silica gel to keep it dry until analysis. 

Nectar spots were carefully cut from filter paper discs and collected in pill vials. In 

species with very little nectar per flower, a number of nectar spots had to be pooled to 

obtain a sufficient sample for Gas Chromatographic (GC) analysis. This amount was 

determined by trial and error and was recorded for each species (it varied from 4 to 30 

flowers per sample, according to nectar volume). Usually several flowers were sampled 

from the same plant, to reduce potential variation. 

Nectar was dissolved by adding 1 ml of 80% re-distilled ethanol to each vial with the 

nectar sample disc, closing it and sonicating for 15 min., then heating it for 10 min. at 

40 o C. Samples were dried down under a nitrogen stream at 40 o C. Sugars were 

converted to their oxime forms by adding 0.5 ml of a solution containing 25 mg.ml -1 of 

hydroxylamine hydrochloride and 6 mg.ml -1 of phenyl-β-D-glucoside in dry, silylation 

grade pyridine and heating at 40 o C for 20 min. Samples were cooled to room 

temperature and 0.1 ml of each was removed and placed in another pill vial and 

reduced to dryness under a nitrogen stream at 40 o C. Trimethylsilyl derivatisation of 

sugars was achieved by adding 0.2 ml Sylan BTZ to each sample and leaving it to 

react at room temperature for 15 min. Samples were stored in Eppendorf tubes and 

refrigerated until GC analysis. Care was taken to analyse samples within two days of 

preparation. 

Sugar standards of sucrose, glucose and fructose were prepared in the same way by 

dissolving 5 mg of sugar standard in 5 ml of 80% re-distilled ethanol in a pill vial, 

heating it at 40 o C until completely dissolved (and sonicating if necessary). The 

Chapter 7/ 80

technique outlined above was followed to convert sugars to their oxime forms and for 

trimethylsilyl derivatisation. 

Nectar sugars were analysed by injecting a 1 µl sample into a Varian 3700 Gas 

Chromatograph equipped with a 1.8 m x 6 mm O.D. glass column packed with OV-17 

on Chromosorb HP 80/100, with a flame ionising detector. Helium gas was used as a 

carrier and chromatographic conditions were: injector temperature 200 o C, ion detector 

temperature 300 o C, with program conditions: 3 min. at 125 o C, raising temperature at 

4 o C.min. -1 to 270 o C and holding at 270 o C for 10 min. Each run took about one hour to 

complete and allow the machine to cool down to the starting temperature. 

Peak integration was performed using an HP 3800A integrator and sample peaks were 

compared to those of standards. In cases of very low sugar concentration in samples 

(i.e. low nectar yield), attenuation threshold was dropped to its lowest limit for 

detection. The phenyl-β-D-glucoside that was included with the solvent during sample 

preparation acted as an internal standard to provide consistency for each run. In 

samples with very low sugar concentration (and hence small peaks) this internal 

standard confirmed that the correct amount of sample had been injected and that the 

run was successful. 

Nectar sugar percentages were calculated from integrated areas for sucrose, glucose 

and fructose, which were the major sugar peaks detected. 

Variability in nectar sugar composition was tested using a number of samples of P. 

oertendahlii T.C.E.Fr., a species that is endemic to the Oribi Gorge area and that 

showed initial results contrary to expectation. Samples of pooled flowers, analysed by 

GC, were compared to pooled and single-flower samples sent away for High 

Performance Liquid Chromatography (HPLC) analysis by B.-E. van Wyk at the 

University of Johannesburg. A number of duplicate samples of P. ecklonii Benth. 

were also analysed, using GC and compared with HPLC performed by Tracy 

Odendaal at the University of KwaZulu-Natal (UKZN), Pietermaritzburg campus. 

Duplicate samples of P. oribiensis Codd and T. longiflora were also analysed. 


Nectar volume ranges were recorded for 19 species or varieties of Plectranthus, and 

Py. urticifolia. 

Chapter 7/ 81

Nectar volumes were measured by using calibrated micro-capillaries as well as the 

spot-size technique recorded in Dafni (1992). In the latter, nectar was collected from 

detached corollas by squeezing the corolla gently and spotting the nectar onto 

Whatman No. 1 filter paper; the diameter of the spot was measured and then a circle 

was drawn around the spot so that this nectar could be used for later sugar analysis. 

The diameter of each spot was converted to a volume measurement using known 

nectar volumes from Dafni (1992: Table 3, Chapter 5, p 140). In the first method 

corollas were detached and the micro-capillary was held at the base of the corolla to 

take up nectar for a direct reading. 

In most cases nectar volume and concentration were measured mid-morning, using 

cultivated plants in greenhouses where insects do not enter, or garden plants that were 

covered by netting at dawn, which meant that no nectar had yet been taken by insects. 

Volume results reflect the maximum amount of nectar available to insects. 

Nectar concentrations were initially recorded for ten species of Plectranthus, using an 

Atago hand-held refractometer (type N1). Some species of Plectranthus have small 

nectar volumes, thus a number of flowers were used to give pooled samples since the 

refractometer was not capable of measuring small samples. Another refractometer 

(Bellingham and Stanley Eclipse hand-held refractometer), capable of measuring 

smaller samples, was subsequently used (in 2005) to measure nectar concentration. 

Consequently, individual flower nectar concentrations are known for 8 species of 

Plectranthus. 

Results 

A. Nectar Sugar Analysis 

Examples of GC traces are shown in Fig. 1. The hexose peaks appear first (fructose, 

then glucose), followed by the internal standard (phenyl-β-D-glucoside), with sucrose 

appearing last. 

The majority of studied Plectranthus species have (truly) sucrose-dominant nectar, 

ranging from 50 – 96% sucrose (Table 1). All three species of Orthosiphon (71 – 81% 

sucrose) and Py. urticifolia (87% sucrose) had sucrose-dominant nectar, while the two 

species of Hemizygia (both 0 % sucrose) as well as T. longiflora (6% sucrose) had 

hexose-dominant nectar. 

Chapter 7/ 82

A 

B 

Fr 

Fr 

Gl 

Gl 

Figure 1: Examples of gas chromatographic (GC) traces of nectar sugars for two species of 

Plectranthus. A: P. ecklonii (sucrose-dominant), B: P. oertendahlii (hexose-dominant sample). 

Fr: Fructose, Gl: Glucose, Su: Sucrose, Std: Internal standard – phenyl-β-D-glucoside. 

Std 

Std 

Su 

Su 

Chapter 7/ 83

Table 1: Nectar sugar properties of Plectranthus and other species analysed in this study, using 

Gas Chromatography (GC), with an indication of main pollinator class. Plant vouchers are 

lodged at Bews Herbarium (NU). 

CP: C. Potgieter, NV: no voucher; Fr: Fructose, Gl: Glucose, Su: Sucrose; S/H nectar: ‘True’ nectar 

sugar class dominance; S: more than 50% sucrose, H: more than 50% hexose; B&B nectar: Nectar 

sugars classified according to Baker & Baker (1983a); HR: ‘hexose-rich’, HD: ‘hexose-dominant’, 

SR: ‘sucrose-rich’, SD: ‘sucrose-dominant’; Apin: Bee (Apidae, Apinae), Mega: Bee (Apidae, 

Megachilinae), Nem: Fly (Nemestrinidae), Acro: Fly (Acroceridae), Tab: Fly (Tabanidae). Species 

marked with an asterisk* are visited for nectar by the nemestrinid fly, S. wiedemanni, or expected to 

be visited by a Stenobasipteron species. 

Species % % % S/H B&B Voucher Pollinator class 

Plectranthus spp. 

Fr Gl Su nectar nectar 

P. reflexus * 1 3 96 S SD CP 95 Nem 

P. hilliardiae * 5 4 91 S SD CP 112 Nem 

P. petiolaris (purple) 3 7 90 S SD CP 115 Apin 

P. ernstii 5 5 90 S SD CP 85 Apin 

P. oribiensis 2 13 85 S SD CP 102 Apin 

P. hadiensis 4 11 85 S SD CP 92 Apin, Nem 

P. ecklonii * 9 16 75 S SD CP 114 Nem, Apin, Tab 

P. ciliatus * 13 15 72 S SD CP 116 Acro, Apin 

P. zuluensis (pale blue) * 17 21 62 S SD CP 118 Acro 

P. zuluensis (dark blue) * 20 23 57 S SD CP N12 Acro 

P. fruticosus (long spur) * 19 25 56 S SD CP 126 Nem 

P. ambiguus * 24 22 54 S SD CP 86 Nem 

P. laxiflorus 17 33 50 50:50 SR CP 927 Apin, Nem 

P. oertendahlii 53 36 11 H HR CP N13 Acro 

Varieties of P. saccatus 

short, white 22 21 57 S SD CP 131 Acro?/Apin? 

long tube * 30 29 41 H SR CP 120 Nem 

medium, fine spots 30 28 42 H SR NV Nem? 

medium, speckled * 33 31 36 H SR CP 109 Nem 

medium, no spots 41 37 22 H HR NV Nem? 

medium, blotched 45 37 18 H HR CP 108 Nem? 

medium, pale spots 50 43 7 H HD NV Nem? 

Other Lamiaceae 

Pycnostachys urticifolia 5 8 87 S SD CP 1064 Meg 

Orthosiphon labiatus 4 16 80 S SD CP N7 ? 

Orthosiphon serratus 7 15 78 S SD CP N8 ? 

Orthosiphon tubiformis * 7 22 71 S SD CP N6 Nem 

Stachys tubulosa * 19 23 58 S SD CP N4 ? 

Stachys grandifolia 29 34 37 H SR CP N5 ? 

Thorncroftia longiflora 12 60 28 H HR CP N3 Nem? 

Hemizygia albiflora 69 31 0 H HD NV ? 

Hemizygia incana 60 40 0 H HD NV ? 

Acanthaceae 

Isoglossa hypoestiflora * 16 34 50 50:50 SR CP N10 Nem 

Chapter 7/ 84

Of the seven forms of P. saccatus that were analysed, two medium-tubed forms and 

one long-tubed form had ‘sucrose-rich’ nectar (36 – 42% sucrose), three medium-tubed 

forms had ‘hexose-rich or -dominant’ nectar (7 – 22 % sucrose) and the short-tubed 

succulent form from Umtamvuna was ‘sucrose-dominant’ (57% sucrose). In reality, six 

of the forms had truly hexose-rich nectars, while only the short-tubed form had nectar 

with more than 50% sucrose. 

Initial GC results, analysed in Pietermaritzburg in 2003, showed P. oertendahlii to have 

‘hexose-dominant’ to ‘hexose-rich’ nectar (0–1% sucrose), with the small percentage of 

sucrose in the latter case only detected when integrator attenuation was set at the 

lowest setting (Table 2). Subsequent samples, analysed by HPLC at the University of 

Johannesburg, showed varying results; pooled flower samples had sucrose 

percentages of 54 – 59%, while individual flowers varied from 38 – 60% sucrose. 

_______________________________________________________________________________________________ 

Table 2: Variability in nectar sugar properties of samples of Plectranthus oertendahlii, a species 

endemic to Oribi Gorge (OG). All samples were collected in the field, at OG, except for sample 

1, which was from material cultivated in a greenhouse. Some samples represent pooled nectar 

from different flowers; some are from individual flowers. Plant vouchers are lodged at Bews 

Herbarium (NU). 

CP: C. Potgieter, NV: no voucher; GC: Gas Chromatography, done by C. Potgieter; HPLC: High 

Performance Liquid Chromatography, done by B.-E. van Wyk; Fr: Fructose, Gl: Glucose, Su: 

Sucrose; S/H nectar: ‘True’ nectar sugar class dominance, S: more than 50% sucrose, H: more 

than 50% hexose; B&B nectar: Nectar sugars classified according to Baker & Baker (1983a), 

HR: ‘hexose-rich’, HD: ‘hexose-dominant’, SR: ‘sucrose-rich’, SD: ‘sucrose-dominant’. 

Plant 

no. 

Sample 

details 

Sample 

collection 

date 

1 Pooled Feb ‘99 

Method 

& date 

GC, 

June ‘03 

Voucher % 

Fr 

% 

Gl 

% 

Su 

S/H 

nectar 

B&B 

nectar 

CP 96 58 42 0 H HD 

2a Pooled March ‘03 “ CP N13 59 41 0 H HD 

2 b 

Above: Rerun, 

low 

attenuation 

3a Single flower Feb ‘04 

“ “ CP N13 53 36 11 H HR 

HPLC, 

June ‘04 

CP N11 36 26 38 H SR 

3b Single flower “ “ CP N11 20 20 60 S SD 

3c Single flower “ “ CP N11 22 23 55 S SD 

3d… 

Pooled; same 

plant as above 

“ “ CP N11 20 22 58 S SD 

4 

Pooled; plants 

of same patch 

Pooled; plants 

“ “ NV 22 24 54 S SD 

5 of different 

patches 

“ “ NV 24 17 59 S SD 

Chapter 7/ 85

To test whether these results could be attributed to the GC technique that was followed 

for the majority of species shown in Table 1, samples of P. ecklonii were re-analysed 

using HPLC in Pietermaritzburg. Samples from the same plant, collected on the same 

day, showed 65% sucrose with GC and 77 – 89% sucrose with HPLC (Table 3). 

Repeat analyses of the same sample by HPLC showed, in three cases, the same or 

very similar results (Table 3). A repeat sample of P. oribiensis, analysed with GC in 

both cases, showed the same result (85% sucrose), while repeat samples of T. 

longiflora, using GC, showed slightly different results (6% and 28% sucrose), but both 

were truly hexose-dominant (Table 3). 

_______________________________________________________________________________________________ 

Table 3: Duplicate samples and nectar sugar analysis technique checks (GC / HPLC), for three 

species of Lamiaceae. Plant vouchers are lodged at Bews Herbarium (NU). Sub-samples a & b 

of samples 2 – 4 & 7 are duplicate analyses of the same nectar sample. 

GC: Gas Chromatography, HPLC: High Performance Liquid Chromatography; Fr: Fructose, Gl: 

Glucose, Su: Sucrose; S/H nectar: ‘True’ nectar sugar class dominance, S: more than 50% 

sucrose, H: more than 50% hexose; B&B nectar: Nectar sugars classified according to Baker & 

Baker (1983a); HR: ‘hexose-rich’, HD: ‘hexose-dominant’, SD: ‘sucrose-dominant’. 

Species & sample 

collection date 

Sample 

no. 

Analysis method 

& date 

% 

Fr 

% 

Gl 

% 

Su 

S/H 

nectar 

B&B 

nectar 

P. ecklonii C. Potgieter N9 (all from same plant, cultivated in UKZN botanical garden) 

April 2008 1 

GC: C. Potgieter 

April 2008 

14 21 65 S SD 

June 2008 2a 

HPLC: T. Odendaal 

June 2008 

6 14 80 S SD 

“ 2b “ 6 14 80 S SD 

“ 3a “ 3 8 89 S SD 

“ 3b “ 4 10 86 S SD 

“ 4a “ 8 16 76 S SD 

“ 4b “ 8 15 77 S SD 

P. oribiensis C. Potgieter 102 (from same cultivated cutting, at UKZN botanical garden) 

March 1999 5 

February 2005 6 


July 2003 

GC: C. Carbutt 

February 2005 

Thorncroftia longiflora C. Potgieter N3 (same sample, re-run) 

2 13 85 S SD 

2 13 85 S SD 

March 1999 7a 


July 2003 

12 60 28 H HR 

March 1999 7b “ 16 78 6 H HD 

Chapter 7/ 86


Nectar volume ranges for 19 studied taxa are shown in Table 4, arranged according 

to corolla tube length and with main nectar-feeding pollinator class indicated. The 

highest volume of nectar was found in a flower of the long-tubed form of P. saccatus 

(8.7 μl), with the other three long-tubed species of Plectranthus (pollinated by S. 

wiedemanni) showing moderate maximum nectar levels (2.4 – 3.2 μl). The beepollinated 

P. ernstii, which has a relatively short corolla tube, showed a maximum of 

6.8 μl, while the bee-pollinated P. petiolaris E.Mey. ex Benth. had a maximum of 3.2 μl. 

______________________________________________________________________________________________ 

Table 4: Nectar volume range and corolla tube length of 19 species or forms of Plectranthus, 

and Py. urticifolia. Pollinator class represents the main nectar-feeding class of pollinators; Apin: 

Bee (Apidae, Apinae), Nem: Fly (Nemestrinidae), Acro: Fly (Acroceridae), Tab: Fly (Tabanidae); 

Shape of corolla base and presence of spur indicated; Nar: Narrow corolla base, Sac: Saccate 

corolla base. Species marked with an asterisk* are visited for nectar by the nemestrinid fly, S. 

wiedemanni. 

Species Volume 

range 

(µl) 

Floral visitors excluded 

n Corolla 

tube 

length 

(mm) 

Corolla 

base 

shape 

Pollinator 

class 

Voucher 

P. ambiguus * 0 – 2.7 32 20 – 33 Nar Nem CP 86 

P. hilliardiae * 0.2 – 2.4 40 21 – 32 Sac Nem CP 112 

P. reflexus * 0.9 – 3.2 8 24 – 30 Sac Nem CP 95 

P. saccatus (long) * 1.4 – 8.7 10 20 – 30 Sac Nem CP 120 

P. zuluensis (dark)* 0.4 – 1.4 11 10 – 16 Sac Acro CP N12 

P. zuluensis (pale)* 0 – 1.1 33 10 – 16 Sac Acro CP 118 

P. ecklonii * 0.1 – 1.4 28 10 – 15 Nar Nem, Apin, Tab CP 114 

P. saccatus (med.)* 0.1 – 1.8 17 9 – 11 Sac Nem CP 108 

P. petiolaris 0 – 3.2 61 7 – 11 Nar Apin CP 115 

P. oribiensis 0.2 – 0.6 7 6 – 12 

Sac, 

Spur 

Apin CP 102 

P. hadiensis 0 – 2.4 82 6 – 8 Sac Apin, Nem CP 92 

P. ciliatus 0.2 – 2.8 25 6 – 8 Sac Acro, Apin CP 116 

P. fruticosus * 0.1 – 0.9 23 5 – 8 Sac Nem CP 126 

P. saccatus (short) 0.2 – 1.1 24 5 – 7 Sac Acro? / Apin? CP 131 

P. ernstii 0.3 – 6.8 25 4 – 8 Sac Apin CP 85 

Floral visitors not excluded 

Py. urticifolia 0 – 1.4 74 11 Nar Apin CP 1064 

P. laxiflorus 0.2 – 1.7 40 10.5 Nar Apin, Nem CP 135 

P. oertendahlii 0 – 0.4 33 8 – 13 Sac Acro CP 97 

P. petiolaris (pink) 0 – 1.8 31 7 – 11 Nar Apin CP 100 

Chapter 7/ 87

Nectar sugar concentration results are presented (as % sucrose equivalents) in Tables 

5 & 6. Table 5 shows measurements for individual flowers, taken with a more recent 

model of refractometer, while Table 6 shows measurements where flowers were 

pooled to obtain sufficient nectar for a reading with a less modern Atago handheld 

refractometer. In the latter case the ranges of pooled readings are indicated, while 

range and average are presented for individual readings in Table 5. Both tables show 

corolla tube length and major class of pollinator. 

_______________________________________________________________________________________________ 

Table 5: Mean and range of nectar concentration (as % sucrose equivalents) for individual 

flowers, taken in 2005 with a sensitive Bellingham and Stanley Eclipse hand-held refractometer. 

Readings were taken from cultivated plants from which floral visitors were excluded, except for 

P. hilliardiae. Pollinator class represents the main nectar-feeding class of pollinators; Apin: Bee 

(Apidae, Apinae), Nem: Fly (Nemestrinidae), Acro: Fly (Acroceridae), Tab: Fly (Tabanidae). 

Plectranthus 

species 

Mean 

nectar 

concen- 

tration 

Floral visitors not excluded 

Concen- 

tration 

range 

SD 

n 

Corolla 

tube 

length 

(mm) 

Pollinator 

class 

Voucher 

P. hilliardiae 17 % 13–20 % 2.6 5 21–32 Nem CP 112 

Floral visitors excluded 

P. ambiguus 17 % 13–22 % 2.0 23 20–33 Nem CP 927 

P. zuluensis 29 % 15–36 % 5.6 11 10–16 Acro CP 118 

P. ecklonii 24 % 13–30 % 3.8 22 10–15 Nem, Apin, Tab CP 114 

P. saccatus 

(medium) 

34 % 4– 49 % 11.3 13 9–11 Nem CP 108 

P. petiolaris 30 % 25–36 % 2.6 21 7–11 Apin CP 100 

P. hadiensis 28 % 10–49 % 11.8 17 6–8 Apin, Nem CP 92 

P. fruticosus 

(short) 

30 % 22– 7 % 3.3 20 5–8 Nem CP 131 

Chapter 7/ 88

Plectranthus nectar concentrations ranged from 4 – 55% and results were highly 

variable (Tables 5 & 6). The short-tubed forms of P. saccatus showed highly variable 

concentrations, from 7 – 55% (Table 6), with individual flower measurements of a 

medium-tubed species of P. saccatus also being highly variable, from 4 – 49% (Table 

5). 

_______________________________________________________________________________________________ 

Table 6: Mean and range of nectar concentration (as % sucrose equivalents) of pooled flower 

samples, measured using an Atago hand-held refractometer (type N1), prior to 2005. 

All were cultivated plants from which floral visitors were excluded. 

Pollinator class represents the main nectar-feeding class of pollinators; Apin: Bee (Apidae, 

Apinae), Mega: Bee (Apidae, Megachilinae), Nem: Fly (Nemestrinidae), Acro: Fly (Acroceridae), 

Tab: Fly (Tabanidae). 

Species Nectar 

concentration 

(%) 

No. 

pooled 

flowers 

Corolla 

tube length 

(mm) 

Conc. 

range 

(%) 

Pollina 

tor 

class 

Voucher 

P. hilliardiae 7 20 21–32 

8 2 21–32 7–8 Nem CP 112 

8 15 21–32 

P. saccatus (long) 18 5 20–23 

4 

7 

2 

3 

23–25 

23–25 

4–18 Nem CP 120 

4 2 23–25 

P. saccatus (med.) 26 18 13–15 

36 

33 

21 

17 

13–15 

10–11 

26–51 Nem CP 108 

51 13 9–11 

P. zuluensis 49 

52 

15 

20 

10–16 

10–16 

49–52 Acro CP 118 

P. saccatus (short) 7 25 7–9 

6 

55 

11 

10 

6–8 

6–8 

7–55 

Acro? / 

Apin? 

CP 131 

54 10 6–8 

P. ciliatus 33 

29 

41 

28 

6–8 

6–8 

29– 33 

Acro, 

Apin 

CP 116 

P. fruticosus 24 

34 

35 

30 

5–8 

5–8 

4–34 Nem CP 131 

P. hadiensis 

P. madagascariensis 

44 

30 

20 

18 

6–8 

5–6 

30– 4 

Apin, 

Nem 

CP 92 

CP 90 

P. ernstii 21 

23 

5 

3 

4–8 

4–8 

21–23 Apin CP 85 

Chapter 7/ 89

Discussion 

A. Nectar Sugars 

Very few studies have reported on nectar sugar ratios in the genus Plectranthus, but 

surveys have revealed the general condition for the Lamiaceae (Percival 1961, Baker & 

Baker 1983a). 

A semi-quantitative paper chromatographic technique was used by Percival (1961) to 

describe the nectar sugars of about 900 species from wide-ranging plant families. 

Descriptions of the abundance of sugars were made subjectively, based on the size 

and depth of spot colour on the chromatogram (Percival 1961). Plectranthus 

oertendahlii nectar was found to fall in class SFG, the capitals indicating plenty of each 

sugar and the bold S indicating “a strong preponderance of” sucrose (Percival 1961). 

This result was based on one sample of 20 flowers from 1 locality, probably a cultivated 

plant. Pycnostachys urticifolia, a relative of Plectranthus, showed a very similar nectar 

sugar profile (Percival 1961). 

The Lamiaceae in general have long-tubed flowers with nectar in which sucrose 

dominates (Percival 1961). This observation was confirmed by Baker & Baker (1983a) 

by studying 765 species with more modern methods. Certain families show 

conservatism in the proportions of major nectar sugars and in the Lamiaceae and 

Ranunculaceae, for example, nectars are characteristically ‘sucrose-dominant’ or 

‘sucrose-rich’. In contrast, nectars from the Asteraceae and Brassicaceae were found 

to be dominated by hexose sugars (Baker & Baker 1983b). This pattern breaks down in 

the alpine zone of the Rocky Mountains (Colorado), where Baker & Baker (1983b) 

found that most of the plant species have ‘hexose-rich’ nectar; this energy-economical 

nectar type is thought to be in response to the “less than optimum environment for 

sugar production by photosynthesis” (Baker & Baker 1983b). 

The data presented in Table 1 confirms that most of the Lamiaceae analysed in this 

study have truly sucrose-dominant nectar that also fall in the ‘sucrose-rich’ and 

‘sucrose-dominant’ categories of Baker & Baker (1983a). In a few cases the categories 

of Baker & Baker (1983a) were found to be misleading, for example three of the 

varieties of P. saccatus have truly hexose-dominant nectar that would previously have 

classified as ‘sucrose-rich’ (Table 1). This general trend of sucrose dominance in a 

lamiaceous genus, as well as the re-analysis of samples of P. ecklonii with HPLC 

Chapter 7/ 90

(Table 3), indicate that the GC technique followed for the majority of samples is 

acceptable. 

The nectar data for Py. urticifolia (see Table 1) corresponds to the ‘sucrose dominance’ 

noted by Percival (1961), but some of the samples did not correspond to her result for 

P. oertendahlii. Two sets of independent nectar samples of P. oertendahlii were initially 

analysed in this study: one collected from the field in the same year as analysis, and 

one collected from cultivated material four years prior to analysis. Initial results showed 

no sucrose, only fructose and glucose, with the former slightly dominant. One sample 

was re-run with the lowest possible integrator threshold and then a very small peak of 

sucrose was detected (values in Table 2). A possible explanation is that the cultivated 

specimen that Percival would have sampled could have been misidentified (possibly 

confused with P. ciliatus E.Mey. ex Benth., another common container plant). Another 

is that Percival’s technique depended on visual estimation of sugar abundance, which 

by the author’s admittance was subjective. This is a concern also raised by Van Wyk 

(1993). Subsequent analyses of P. oertendahlii nectar showed variability in samples 

from individual flowers taken from the same plant, with values varying from hexosedominant 

to sucrose-dominant (Table 2). 

The nutrient-poor Natal Group Sandstones on which the endemic P. oertendahlii 

grows, linked with its dim understorey habitat, could lead to low levels of 

photosynthesis which may favour energy-economical hexose-dominant nectar. 

However, sampling of cultivated material under good light and nutrient regimes showed 

that the nectar composition was similar (hexose-dominant) in three samples collected 

from the greenhouse and the field, but different from sucrose-dominant samples 

collected in the field at another time. A number of other species in this study, such as 

P. hilliardiae and P. reflexus, are also sandstone endemics that grow on the forest 

floor, yet these species show highly dominant levels of sucrose in their nectar. It is 

possible that some species simply display large variance in nectar sugar characteristics 

and since most studies do not analyse multiple samples of the same species, such 

variability most probably goes unnoticed. 

The pooling of nectar from flowers collected from the same plant, and even from the 

same population, for sugar analysis, was shown to be an acceptable practise by Lanza 

et al. (1995). In their study they investigated various nectar characteristics of Impatiens 

capensis Meerb. at the level of the individual flowers, plant and population. Significant 

Chapter 7/ 91

differences in sucrose concentration were detected only in between-population 

comparisons (Lanza et al. 1995). Since pollinating insects visit more than one flower in 

a foraging bout, pooling of nectar from various flowers in a population was considered 

to be acceptable for the determination of average nectar sugar composition in the 

current study. 

The methods of Freeman et al. (1991) could be used as a guideline whereby a species 

needs to show two or more samples with consistent nectar sugar composition, before 

inclusion into the results, with exemptions being single samples of species where 

conspecific or congeneric taxa are included. Interestingly, Freeman et al. (1991) 

included two Lamiaceae in their survey of South and South East Asian taxa. Leucas 

zeylanica R.Br. was shown to have nectar with 86% sucrose, but Orthosiphon aristatus 

(Blume) Miq. had only 53% sucrose. In the present study O. tubiformis [= Ocimum 

tubiforme (R.D.Good) A.J.Paton] was found to have 71% sucrose (based on one 

sample), and Goldblatt & Manning (2000) found the latter species to have nectar with 

94 – 100% sucrose, based on three samples. In light of such variation, even the 

inclusion of two or more samples could be questioned, let alone congeneric 

comparisons. 

Since one of the most interesting findings during the course of this project has been the 

discovery and description of the S. wiedemanni (Nemestrinidae) pollination guild, it 

makes sense to establish whether long-proboscid flies have a particular preference for 

a certain class of nectar. In a pollination study of Lapeirousia Pourr. subgenus 

Lapeirousia (Iridaceae), Goldblatt et al. (1995) found that most species have ‘sucrosedominant 

or -rich’ nectar and their pollinators include long-proboscid nemestrinid and 

tabanid flies with proboscis lengths of 30 – 60 mm. The authors noted that this adds a 

new dimension to the categorisation of nectar and pollinators by Baker & Baker (1983a, 

1990), since the hexose-loving flies in their studies were short-tongued species from 

families such as the Syrphidae, Muscidae and Phoridae. Goldblatt et al. (1995) pointed 

out that in the same way that long-tongued bees prefer sucrose in nectar, the flowers 

visited by long-proboscid flies also produce sucrose. They suggested that large-bodied 

insects that are physically active and that maintain wing movement during feeding, may 

require more sucrose rather than hexose. This holds for the nemestrinids, tabanids 

such as Philoliche Wiedemann, 1820, hawkmoths and some anthophorine bees 

(Goldblatt et al. 1995). 

Chapter 7/ 92

A study on the pollination of Sparaxis (also Iridaceae) by Goldblatt et al. (2000) shows 

that a species of Mesomyia Macquart, 1850 (Tabanidae) and the nemestrinid fly 

Prosoeca peringueyi Lichtwardt, 1920 both visit flowers with ‘sucrose-dominant’ nectar. 

Subsequent work showed that the nectar sugar characteristics in long-proboscid flypollinated 

species of Iridaceae, Orchidaceae and Lamiaceae are ‘sucrose-rich or - 

dominant’ (Goldblatt & Manning 2000). However, surveyed Pelargonium L’Hér. ex 

Aiton species (Geraniaceae), with one exception, have ‘hexose-rich or -dominant’ 

nectar (Goldblatt & Manning 2000). From such divergent nectar properties, Manning & 

Goldblatt (1996) concluded that nectar sugar composition is not a significant factor in 

the Prosoeca peringueyi (Nemestrinidae) pollination guild. This observation was 

confirmed for other long-proboscid species by Goldblatt & Manning (2000) in their 

review on long-proboscid fly pollination in southern Africa. 

Data from the current study show that samples of long-tubed Lamiaceae such as P. 

reflexus, P. hilliardiae, P. ambiguus, O. tubiformis and S. tubulosa all had truly sucrosedominant 

nectar (54 – 96% sucrose), while long-tubed species such as P. saccatus 

(long-tubed form) and T. longiflora had truly hexose-dominant nectar (Table 1). 

Plectranthus species that are not long-tubed, but that are visited for nectar by the longproboscid 

S. wiedemanni (e.g. P. ecklonii, P. ciliatus, P. zuluensis and P. fruticosus) all 

have truly sucrose-dominant nectar, as do O. tubiformis and S. tubulosa, that have 

been inferred to be visited by this nemestrinid fly species (Potgieter & Edwards 2001). 

The sample from the long-tubed I. hypoestiflora (Acanthaceae), which also belongs to 

the S. wiedemanni pollination guild, showed equal amounts of sucrose and hexose 

(Table 1). While most of the species correspond to a sucrose-dominant pattern, there 

are exceptions. 

The question then, is whether Plectranthus species that are visited by different types of 

pollinator have different nectar classes as a result. Most Plectranthus species recorded 

in Table 1 have sucrose-dominant nectar, with P. saccatus and some samples of P. 

oertendahlii being the exceptions with hexose-dominant nectar. Plectranthus 

oertendahlii is pollinated by medium-proboscid acrocerid flies (see Appendix), but then, 

so is P. ciliatus and P. zuluensis T.Cooke, which have sucrose-dominant nectar. 

Plectranthus saccatus happens to be the one variable species for which limited 

pollinator data was evident during this study: the long-tubed form of P. saccatus is 

long-proboscid fly-pollinated and, from garden-based observation, at least one 

medium-tubed form is pollinated by S. wiedemanni (see Appendix). A further exception 

Chapter 7/ 93

is that the shortest-tubed P. saccatus, which possibly has an acrocerid, tabanid or 

apinid pollinator (see Appendix), has sucrose-dominant nectar, unlike the other longtubed 

and medium-tubed forms of this species. 

Species of Orthosiphon and Pycnostachys (recorded in Table 1) showed sucrose 

dominance, with Thorncroftia and Hemizygia having hexose-dominant nectar. In the 

case of Stachys one species was hexose-dominant and another was sucrosedominant. 

Nectar from P. laxiflorus showed an equal split between hexose and 

sucrose, which is interesting in a species where both bees and medium-proboscid 

nemestrinid flies are frequent floral visitors (Potgieter et al. 2009). No definite pattern 

can at present be seen in the comparison of nectar sugars and pollinator type in a 

range of Plectranthus species and their relatives. 

In a study of bees (Anthophoridae) and bee-mimicking flies (Acroceridae) that pollinate 

Gladiolus brevifolius Jacq. (Iridaceae), Goldblatt et al. (1997) noted that ‘sucrosedominant’ 

nectar is characteristic of most of the insect-pollinated Gladiolus species 

(and of the subfamily Ixioideae). This nectar class for Iridaceae is expected, but what is 

of interest is the range of pollinators that visit these flowers for ‘sucrose-dominant’ 

nectar. Long-tongued anthophorine bees (referred to as apinid bees in the current 

study), short-tongued bees (species of Allodape Lepeletier & Serville, 1825) and 

interestingly, a species of Psilodera Gray, 1832 (a medium-proboscid acrocerid fly 

similar to the species that visits P. oertendahlii) all pollinate this Gladiolus. 

Does this mean that acrocerid flies prefer ‘sucrose-dominant’ nectar? The data 

presented here for three medium-tubed Plectranthus species that are, amongst others, 

pollinated by acrocerid flies of the genus Psilodera (Table 1), show sucrose-dominant 

nectar (regardless of nectar classification system), but P. ciliatus is also visited by longtongued 

bees that ‘prefer’ sucrose (according to Baker & Baker 1983a). Plectranthus 

zuluensis is only visited by acrocerid and nemestrinid flies and has sucrose-rich nectar, 

while P. oertendahlii only has acrocerid visitors and may have either sucrose- or 

hexose-rich nectar (Table 2). 

The Lapeirousia study by Goldblatt et al. (1995) shows a similar range of pollinators in 

plants with mostly ‘sucrose-dominant’ nectar. Apart from the long-proboscid 

nemestrinids and tabanids, these included hawkmoths (22 mm long proboscides), 

Chapter 7/ 94

ombyliid flies and honey bees (6 – 8 mm long proboscides) and noctuid moths and 

anthophorine bees (7 – 8 mm long proboscides) (Goldblatt et al. 1995). 

Some of the conclusions drawn by Baker & Baker (1983a) summarise the debate 

around phylogenetic constraints versus the adaptation of nectar sugar ratios (nectar 

classes) to pollinator preference. They point out that there does seem to be a tendency 

for intra-familial resemblances in the ratio between sucrose and hexoses in nectar. 

They mention that some nectar sugar ratio modification can take place when new 

flower-pollinator partnerships are established, and they suggest a tendency for nectar 

chemistry to predispose members of a particular plant family to pollination by certain 

pollinator classes on the basis of nectar chemistry; however they caution that this could 

be outweighed by morphological and phenological adaptations in the features of 

flowers and inflorescences (Baker & Baker 1983a). 

The recent review on nectar chemistry by Nicolson & Thornburg (2007) concludes that 

the phylogenetic history of a plant group is likely to be the primary determinant of 

nectar sugar composition, but that pollinator class may have a secondary effect. 

For the Plectranthus and other species of Lamiaceae studied here, with a few 

exceptions, the occurrence of sucrose- or hexose-dominance appears to relate more to 

phylogeny, rather than the preference of a certain class of pollinator. A more diverse 

range of plant families will have to be sampled locally for nectar sugars and studied 

with respect to pollination to tease this issue apart. It is clear that some of the 

previously held ideas regarding the nectar preferences of long-tongued bees and other 

pollinator classes (such as those held by Baker & Baker 1983a) may be a case of 

inferring patterns from circumstantial evidence, rather than stringent testing for 

preferences. 


Nectar volume and concentration was highly variable in the studied species of 

Plectranthus. This is not uncommon, since volume measurements for a range of longproboscid 

fly-pollinated families in Goldblatt & Manning (2000: Table 3, pp 163 – 167) 

often show a doubling or tripling of the lowest measured volume (and occasionally 

even more) to reach the upper range. 

Chapter 7/ 95

Nectar volume and concentration data is sparsely recorded for other members of the 

Lamiaceae. One record is given in Goldblatt & Manning (2000) for the long-tubed 

O. tubiformis which had nectar volumes of 2.7 – 4.1 μl and concentration of 24.5% (SD 

2.2). Nectar volumes in three bird-pollinated species of Leonotis (Pers.) R.Br. were 

measured by Vos et al. (1994), and ranged from 5.3 – 11.3 μl per flower with 

concentrations of 19 – 28%. These volumes reflect the nectar-feeding capacity of 

sunbirds, but the nectar concentration falls within the range of insect-pollinated 

species. Nine species of the Lamiaceae were studied by Dafni et al. (1988) and nectar 

volume per flower was found to be small, ranging from 0.14 – 6.3 μl, with most of the 

species having less than 1 μl per flower. Nectar sugar concentration was generally 

high, from 25 – 52%, and no correlation was found between honeybee preference for 

different plants and nectar volumes or sugar concentrations (Dafni et al. 1998). A 

subsequent study, measuring floral parameters in thirteen species of Labiatae in Israel 

(Dafni 1991), presents nectar concentration values of 17 – 60% sucrose equivalents, 

with nectar volumes of 0.4 – 18.2μl per flower over 24 hours. Larger flowers were found 

to offer more calorific reward per flower, and large flowers lasted longer. Nectar 

rewards in large flowers, with longer floral tubes, were also better protected from shorttongued 

insects and general climatic conditions (Dafni 1991). The study by Ford & 

Johnson (2008) showed four species of Syncolostemon (Lamiaceae) with tube lengths 

varying from 9.6 – 28.0 mm to have corresponding nectar volumes of 0.7 – 3.2 μl and 

variable nectar concentrations of 21.6 – 29.3 %. 

In the current study there does not appear to be a relationship between corolla tube 

length and maximum volume of nectar (Table 4), since short-tubed species such as P. 

hadiensis (Forssk.) Schweinf. ex Spreng. and P. ciliatus had nectar volumes similar to 

those of the long-tubed P. ambiguus and P. hilliardiae. This is contrary to results in 

Syncolostemon, where Ford & Johnson (2008) found a positive correlation between 

nectar volume and corolla tube length. The occurrence of saccate corolla bases in 

Plectranthus shows no obvious relationship to nectar volume either (Table 4). 

There is, however, a general trend for lower concentrations of nectar to be found in 

longer-tubed flowers of Plectranthus that are pollinated by S. wiedemanni, but no other 

trend is evident with respect to pollinator class (Tables 5 & 6). The nectar concentration 

values for Plectranthus species pollinated by S. wiedemanni vary from 4 – 55%, which 

extends the 20 – 32 % range suggested by Goldblatt & Manning (2000) for longproboscid 

flies. The limited nectar concentration data presented in Potgieter et al. 

Chapter 7/ 96

(2005) was based on preliminary results and more recent data are contained here, in 

Table 5. In Syncolostemon nectar concentration was shown to vary among species, but 

was not correlated to corolla tube length (Ford & Johnson 2008). 

The way in which nectar volume is presented has been extended in the present study 

to include the range, since the practice of only giving averages does not reflect what is 

experienced at individual flowers by a visiting insect (Table 4). For the most recent 

volume measurements in this study zero readings for nectar (i.e. no nectar in the 

flower) were included, but in other cases not, which would make a comparison of 

averages meaningless. 

A survey of pollination literature shows that most often only averages and/or ranges, 

without zero measurements, are given, which makes it difficult to gauge whether the 

researcher(s) included data for flowers containing no nectar. Southwick et al. (1981) 

recorded volume averages and ranges of floral nectar production, but none with zero at 

the lower limit; likewise, the long-proboscid fly pollination review by Goldblatt & 

Manning (2000) showed ranges, but none with zero values. Many studies (e.g. Spira 

1980, Whitten 1981, Johnson 2000, Luyt & Johnson 2001, Potts et al. 2001, Johnson 

et al. 2002, Johnson et al. 2003, Macukanovic-Jocic 2004, Goldblatt et al. 2005, 

Manning & Goldblatt 2005, Larsen et al. 2008) only reflect averages and standard 

deviation/error, but not range. However, data on diurnal nectar secretion dynamics, by 

Macukanovic-Jocic (2004), show cases where nectar volume is zero at certain times of 

day; these authors also ascribe high values of standard error to absence, or very low 

levels, of nectar in some flowers. 

In the current study most measurements were made on unvisited plants, hence the 

zero readings (see Table 4) indicate that the variability in nectar volume extends to 

some flowers not producing any nectar at all at any one time. This is important with 

respect to pollinator movement, since there is unpredictability in whether an insect will 

find nectar in any one flower, which forces movement between flowers, inflorescences 

and plants. In a study on bumble-bee foraging movements in Monarda fistulosa L. 

(Lamiaceae), Cresswell (1990) suggests that bees initiate movements between 

inflorescences on the basis of the amount of nectar found at the most recently probed 

flower, with the number of flowers already probed at an inflorescence only exerting a 

weak influence on such choices. 

Chapter 7/ 97

Complete rewardlessness is rarely recorded in plant families other than orchids, 

although many plant populations are suggested to contain a proportion of rewardless 

flowers because individual plants can produce both nectar-bearing and empty flowers. 

Some species are known to be polymorphic for the presence or absence of nectar 

production (Smithson & Gigord 2003). A few authors draw attention to the phenomenon 

of occasional nectarless flowers in their results. Petanidou et al. (1999), whilst 

presenting average values for nectar volume in three species of Lamiaceae in an 

experimental array, recorded some nectarless flowers in 10 out of 15 cases in pot 

treatments. Petanidou et al. (1999) also found that increasing the nutrient status of 

three species of Lamiaceae, increased the number of empty flowers in experimental 

treatments. A study by Real & Rathke (1991) showed that two out of 32 studied flowers 

of Kalmia latifolia L. (Ericaceae) did not accumulate any nectar over a 24 hour period, 

over the life span of the flower. 

This chapter provides new records of nectar sugar composition and measurements of 

nectar volume and concentration for species of Plectranthus and other Lamiaceae that 

have not had such data recorded previously. While general trends may be seen in the 

data, it is clear that exceptions occur, that results can be variable and that the types of 

conclusions that may be drawn from nectar studies may have limited value, other than 

showing that variation may be inherent in the interactions between plants and insect 

pollinators. 

Chapter 7/ 98

References 

Baker, H.G., Baker, I., 1983a. Floral nectar constituents in relation to pollinator type. In: 

Jones, C.E., Little, R.J. (Eds.), Handbook of experimental pollination biology. Van 

Nostrand Reinhold Company Inc., New York. ISBN 0442246765. 

Baker, H.G., Baker, I., 1983b. A brief historical review of the chemistry of floral nectar. 

In: Bentley, B., Elias, T. (Eds.), The biology of nectaries. Columbia University 

Press, New York. ISBN 0231044461. 

Baker, H.G., Baker, I., 1990. The predictive value of nectar chemistry to the recognition 

of pollinator types. Israel Journal of Botany 39: 157–166. 

Cresswell, J.E., 1990. How and why do nectar-foraging bumblebees initiate 

movements between inflorescences of wild bergamot Monarda fistulosa 

(Lamiaceae)? Oecologia 82: 450–460. 

Dafni, A., 1991. Advertisement, flower longevity, reward and nectar protection in 

Labiatae. Acta Horticulturae 288: 340–346. 

Dafni, A., 1992. Pollination Ecology: a Practical Approach. IRL Press, Oxford. ISBN 

0199632987. 

Dafni, A., Lensky, Y., Fahn, A., 1988. Flower and nectar characteristics of nine species 

of Labiatae and their influence on honeybee visits. Journal of Apicultural 

Research 27: 103–114. 



Freeman, C.E., Worthington, R.D., Jackson, M.S., 1991. Floral nectar sugar 

compositions of some South and Southeast Asian species. Biotropica 23: 568– 

574. 

Goldblatt P., Manning J.C., 1999. The long-proboscid fly pollination system in Gladiolus 

(Iridaceae). Annals of the Missouri Botanical Garden 86: 758–774. 

Goldblatt, P., Manning, J.C., 2000. The long-proboscid fly pollination system in 

southern Africa. Annals of the Missouri Botanical Garden 87:146–170. 

Goldblatt, P., Manning, J.C., Bernhardt, P., 1995. Pollination biology of Lapeirousia 

subgenus Lapeirousia (Iridaceae) in southern Africa; floral divergence and 

adaptations for long-tongued fly pollination. Annals of the Missouri Botanical 

Garden 82: 517–534. 

Goldblatt, P., Manning, J.C., Bernhardt, P., 1997. Notes on the pollination of Gladiolus 

brevifolius (Iridaceae) by bees (Anthophoridae) and bee mimicking flies 

(Psilodera: Acroceridae). Journal of the Kansas Entomological Society 70: 297– 

304. 

Goldblatt, P., Manning, J.C., Bernhardt, P., 2000. Adaptive radiation of pollination 

mechanisms in Sparaxis (Iridaceae: Ixiodeae). Adansonia 22: 57–70. 

Chapter 7/ 99

Goldblatt, P., Bernhardt, P., Manning, J.C., 2005. Pollination mechanisms in the African 

genus Moraea (Iridaceae: Iridoideae): floral divergence and adaptation for 

pollinators. Adansonia 27: 21–49. 

Johnson, S.D., 2000. Batesian mimicry in the non-rewarding orchid Disa pulchra, and 

its consequences for pollinator behaviour. Biological Journal of the Linnean 

Society 71: 119–132. 

Johnson, S.D., Alexandersson, R., Linder, H.P., 2003. Experimental and phylogenetic 

evidence for floral mimicry in a guild of fly-pollinated plants. Biological Journal of 

the Linnean Society 80: 289–304. 

Johnson, S.D., Edwards, T.J., Carbutt, C., Potgieter, C., 2002. Specialization for 

hawkmoth and long-proboscid fly pollination in Zaluzianskya section Nycterinia 

(Scrophulariaceae). Biological Journal of the Linnean Society 138: 17–27. 

Kearns, C.A., Inouye, D.W., 1993. Techniques for pollination biologists. University 

Press of Colorado, Colorado, pp 153–216. ISBN 0870812793. 

Lanza, J., Smith, G.C., Sack, S., Cash, A., 1995. Variation in nectar volume and 

composition of Impatiens capensis at the individual, plant, and population levels. 

Oecologia 102: 113–119. 

Larsen, M.W., Peter, C., Johnson, S.D., Olesen, J.M., 2008. Comparative biology of 

pollination systems in the African-Malagasy genus Brownleea (Brownleeinae: 

Orchidaceae). Botanical Journal of the Linnean Society 156: 65–78. 

Luyt, R., Johnson, S.D., 2001. Hawkmoth pollination of the African epiphytic orchid 

Mystacidium venosum, with special reference to flower and pollen longevity. Plant 


Macukanovic-Jocic, M., Duletic-Lausevic, S., Jocic, G., 2004. Nectar production in 

three melliferous species of Lamiaceae in natural and experimental conditions. 

Acta Veterinaria (Beograd) 54: 475–487. 

Manning J.C., Goldblatt P., 1996. The Prosoeca peringueyi (Diptera: Nemestrinidae) 

pollination guild in southern Africa: long-tongued flies and their tubular flowers. 

Annals of the Missouri Botanical Garden 83: 67–86. 

Manning J.C., Goldblatt P., 2005. Radiation of pollination systems in the Cape genus 

Tritoniopsis (Iridaceae: Crocoideae) and the development of bimodal pollination 

strategies. International Journal of Plant Science 166: 459–474. 

Nicolson, S.W., Fleming, P.A., 2003. Nectar as food for birds: the physiological 

consequences of drinking dilute nectar solutions. Plant Systematics and Evolution 

238: 139–153. 

Nicolson, S.W., Thornburg, R.W., 2007. Nectar Chemistry. In: Nicolson, S.W., Nepi, M., 

Pacini, E. (Eds.), Nectaries and Nectar. Springer, Dordrecht, pp 215–263. 

ISBN9781402059377. 

Nicolson, S.W., Van Wyk, B.-E., 1998. Nectar sugars in Proteaceae: Patterns and 

process. Australian Journal of Botany 46: 489–504. 

Chapter 7/ 100




Molecular Phylogeny and Evolution 31: 277-299. 

Percival, M.S., 1961. Types of nectar in angiosperms. New Phytologist 60: 235–281. 

Percival, M.S., 1965. Floral Biology. Pergamon Press, Oxford. pp 80–99. 

Petanidou, T., Goethals, V., Smets, E., 1999. The effect of nutrient and water 

availability on nectar secretion and nectary structure of the dominant Labiatae 

species of phrygana. Systematics and Geography of Plants 68: 233–244. 

Potgieter, C.J., Edwards, T.J., 2001. The occurrence of long, narrow corolla tubes in 

southern African Lamiaceae. Systematics and Geography of Plants 71: 493– 

502. 


Nemestrinidae) pollination guild in eastern southern Africa. Annals of the Missouri 


Potgieter, C.J., Edwards, T.J., Van Staden, J., 2009. Pollination of Plectranthus spp. 

(Lamiaceae) with sigmoid flowers in southern Africa. South African Journal of 

Botany 75: 646–659. 

Potts, S.G., Dafni, A., Ne’eman, G., 2001. Pollination of core flowering shrub species in 

Mediterranean phrygana: variation in pollinator diversity, abundance and 

effectiveness in response to fire. Oikos 92: 71–80. 

Real, L.A., Rathke, B.J., 1991. Individual variation in nectar production and its effect on 

fitness in Kalmia latifolia. Ecology 72: 149–155. 

Smithson A., Gigord L. D., 2003. The evolution of empty flowers revisited. American 

Naturalist 161: 537–552. 

Southwick, E.E., Loper, G.M., Sadwick, S.E., 1981. Nectar production, composition, 

energetics and pollinator attractiveness in spring flowers of western New York. 

American Journal of Botany 68: 994–1002. 

Spira, T.P., 1980. Floral parameters, breeding system and pollinator type in 

Trichostema (Labiatae). American Journal of Botany 67: 278–284. 

Tanowitz, B.D., Smith, D.M., 1984. A rapid method for qualitative and quantitative 

analysis of simple carbohydrates in nectars. Annals of Botany 53: 453–456. 

Van Wyk, B.-E., 1993. Nectar sugar composition in southern African Papilionoideae 

(Fabaceae). Biochemical Systematics and Ecology 21: 271–277. 

Vos, W.T., Edwards, T.J., Van Staden, J., 1994. Pollination biology of annual and 

perennial Leonotis species. Plant Systematics and Evolution 192: 1–9. 

Whitten, W.M., 1981. Pollination ecology of Monarda didyma, M. clinopodia, and 

hybrids (Lamiaceae) in the Southern Appalachian Mountains. American Journal 


Chapter 7/ 101

CHAPTER 8: DISCUSSION AND CONCLUSIONS 

This thesis contributes a number of new findings to the field of Lamiaceae pollination 

and pollination studies in southern Africa in general. 

Chapter 2 represents a selection of the initial data gathered during the course of this 

study, showing that bees and flies are the main pollinators of seven straight-tubed 

species of Plectranthus with varying tube lengths (Potgieter et al. 1999). It goes on to 

show a correlation of corolla tube length with the proboscis lengths of the main 

pollinator groups. This paper presents the first record of long-proboscid (nemestrinid) 

fly pollination in the Lamiaceae family, while pointing to the importance of flies with 

medium proboscid lengths (nemestrinid, tabanid and acrocerid flies) in the pollination 

ecology of Plectranthus with medium- and short corolla tubes. 

Since the publication of this paper, a number of specific identifications for the dipteran 

vouchers have been made. The nemestrinid fly, Stenobasipteron sp., is now known to 

be S. wiedemanni, and the four species of Prosoeca (Nemestrinidae) are actually 

representatives of two species of variable size and appearance (D. Barraclough pers. 

comm.): Prosoeca umbrosa is represented by Prosoeca spp. A and B, while Pr. 

circumdata incorporates Prosoeca spp. C and D. The tabanid fly Philoliche sp. is Ph. 

aethiopica. Acrocerid flies have been identified as Psilodera confusa and Psilodera aff. 

confusa; hence the photographs of ‘Psilodera sp. A’ in Fig. 5 g & h (p 106, Potgieter et 

al. 1999) are in fact two species, with Fig. 5g representing Ps. confusa and Fig. 5h 

Psilodera aff. confusa. 

These positive identifications are largely due to the efforts of Dr David Barraclough 

who, during the course of this study, undertook a revision of type material of the 

southern African genus Stenobasipteron (Barraclough 2005), followed by the 

publication of an overview of the South African Nemestrinidae (Barraclough 2006). This 

was done in response to the need created by various pollination studies, including the 

current Plectranthus study, and the overview was specifically intended for interdisciplinary 

use by botanists, entomologists and pollination biologists, with the aim of 

stimulating interest in the conservation and taxonomic research on South African 

Nemestrinidae (Barraclough 2006). One of the reasons stated for not sinking the genus 

Stenobasipteron as a synonym of Prosoeca is the use of this generic name in the

pollination biology literature (Barraclough 2005), which points to the synergy that is 

developing between insect taxonomists and pollination biologists in South Africa. 

Dr Barraclough identified the vouchers of Acroceridae collected during the study and 

pointed out that the species collected at Umtamvuna NR represents a new taxon. 

Likewise, five of the species of Prosoeca collected during the study as a whole, are 

new taxa that need further study by insect taxonomists. A number of the bee species 

collected during the study are new range extensions when compared to distributions 

given in revisions of the genera Amegilla (Eardley 1994) and Xylocopa (Eardley 1983). 

In addition to taxonomic clarifications of insect vouchers cited in Potgieter et al. (1999), 

some new pollinator observations were also made since publication of this paper. Most 

of these observations expand the class of pollinator to other species of the same genus 

already recorded, but in P. ambiguus the nemestrinid fly Prosoeca umbrosa was a new 

observation and in P. madagascariensis the acrocerid fly Psilodera valida was added. 

In P. ciliatus the original observation of one species of acrocerid fly, ‘Psilodera sp. A’, 

was expanded to include four species of Psilodera. Three additional apinid bee species 

were added to the list of valid floral visitors in P. oribiensis, with the genera Xylocopa 

and Thyreus now represented. The Appendix to the thesis should be consulted for 

updated pollinator data for the seven species that are the main subjects of Potgieter et 

al. (1999): P. ambiguus, P. hilliardiae, P. ecklonii, P. zuluensis, P. ciliatus, P. 

madagascariensis and P. oribiensis. 

Chapter 3 is the most recently published paper (Potgieter et al. 2009), outlining the 

pollination of five species of Plectranthus (plus two allied species) with sigmoid-shaped 

corolla tubes. In addition to showing that corolla tube length is related to the proboscis 

length of pollinating groups (as is the case with the straight-tubed species), this paper 

also shows that the bend in the corolla tube acts as a barrier that protects nectar 

resources from illegitimate visitations, as well as ensuring that the bodies of bee 

visitors are aligned in a way that enhances effective pollen placement and carry-over. 

Chapter 8/ 103 

Sigmoid species with corollas bent or declined to some degree, account for a third of 

southern Africa Plectranthus species and this functional group is shown to be geared 

towards melitophily. Two of the three clades in Plectranthus, the ‘Coleus’ clade and the 

‘sigmoid Plectranthus’ clade (Paton et al. 2004, Paton pers. comm.) show this syndrome 

and are bee-pollinated, with a few species also receiving nectar-feeding visits from

medium-proboscid nemestrinid flies (Potgieter et al. 2009). This is a new discovery, 

not previously recorded or suggested for sigmoid Plectranthus and allied species. 

While straight corolla tubes may be considered pleisiomorphic due to its presence in the 

Ociminae (sister group to the Plectranthinae), the discussion in Chapter 3 suggests the 

possibility that straight corolla tubes may have evolved from the sigmoid condition. 

As a result of its recent publication, basic data contained in Chapter 3, such as the 

number of species of Plectranthus that occur in southern Africa (ca. 53), is updated 

compared to the 45 Plectranthus species originally noted in Chapter 2. This highlights 

new species discoveries in the intervening decade, systematic work in the genus, and 

a generally increased interest in Plectranthus. As our knowledge of insect taxonomy 

and systematics improves, we can also draw better conclusions from data linking floral 

features to insect morphology. 

The paper on convergent pollination syndromes in southern African Lamiaceae 

(Potgieter & Edwards 2001) is a case in point. Chapter 4 aims to describe what is 

known about pollination in long-tubed Lamiaceae and speculates that most of these 

long, straight-tubed species are adapted to pollination by long-proboscid flies. With the 

overview of nemestrinid flies (Barraclough 2006) and the review of Stenobasipteron 

(Barraclough 2005) not yet available in 2001, some of the species concepts of longproboscid 

flies in the summer rainfall region of the eastern part of southern Africa were 

not clear. 

It is now known that the ‘Abel Erasmus Pass form’ of S. wiedemanni (cited by Goldblatt 

& Manning 1998, 1999) is not S. wiedemanni, but rather an undescribed species, which 

accounts for its different habitat requirements to the S. wiedemanni guild (described in 

Potgieter & Edwards 2005). True S. wiedemanni only occurs in forest habitat in the 

Eastern Cape and KwaZulu-Natal (Barraclough 2005), while a number of undescribed 

species of Stenobasipteron are recorded from savanna and grassland habitats in the 

Limpopo and Mpumalanga Provinces (Barraclough 2006). 

Similarly, the various forms of Prosoeca ganglbaueri (with varying proboscid lengths) 

listed in Table 3 of Potgieter & Edwards (2001, Chapter 4) may represent a species 

complex, with the Drakensberg populations comprising at least two species 

(Barraclough 2006), while the identification of Pr. robusta from Mpumalanga Province 

Chapter 8/ 104

(by Goldblatt & Manning 1998, 1999) is incorrect and probably represents a different 

species (Barraclough 2006). 

Such details do not change the notion that these plant species are adapted for longproboscid 

fly pollination, but the details of the possible interactions may be slightly 

different. It is, however, clear that there are guilds based on the habitat preferences of 

the flies, with S. wiedemanni being limited to forest or adjacent areas, while Pr. 

ganglbaueri tends to occur in open habitats such as high-altitude grassland. The 

mechanism by which floral extension was possible in long-tubed species may be 

explained by the presence of shorter-tubed species with straight corollas in the same 

genera, that are also pollinated by medium-proboscid flies and/or long-tongued bees 

(see Discussion in Potgieter & Edwards 2001). 

The genus Syncolostemon was not considered by Potgieter & Edwards (2001) as 

potentially being pollinated by long-proboscid flies, but rather by bees, on account of 

the wide corolla entrance. The subsequent study by Ford & Johnson (2008) showed 

that the long-tubed (22 – 28 mm) species, S. macranthus, S. rotundifolius and S. 

densiflorus, are pollinated, at least in part, by medium-proboscid nemestrinid (Prosoeca 

sp.) and tabanid (Philoliche aethiopica) flies. 

One of the most exciting aspects of this project was the discovery (in Plectranthus) and 

description of the Stenobasipteron wiedemanni pollination guild, which constitutes 

Chapter 5 (Potgieter & Edwards 2005). Having tentatively been included in the 

Prosoeca ganglbaueri–Pr. robusta pollination system by Goldblatt & Manning (2000), it 

was evident that this nemestrinid fly species serviced a completely different guild of 

plants, largely in forest habitat, which is quite distinct from Pr. ganglbaueri. Based on 

the work done by Barraclough (2005) it is now clear that the Mpumalanga species that 

was identified as S. wiedemanni by Goldblatt & Manning (2000), is a different, 

unidentified species of Stenobasipteron that was found to pollinate Gladiolus macneilii 

(Iridaceae) and Orthosiphon tubiformis (Lamiaceae). This anomaly in habitat type in 

Mpumalanga was pointed out in the discussion of Potgieter & Edwards (2005), since it 

did not fit the pattern seen along the rest of the distribution of S. wiedemanni. 

During the current study S. wiedemanni was first observed pollinating the long-tubed 

Plectranthus species, P. hilliardiae and P. ambiguus, at Umtamvuna NR in 1995, with 

subsequent observations extending the species list to seven more species of 

Chapter 8/ 105

Plectranthus (see Appendix). Anecdotal evidence, publications by Goldblatt & Manning 

(1999, 2000) and confirmation by pollen collected from voucher specimens of S. 

wiedemanni, extended the Guild to members of other plant families (Acanthaceae, 

Orchidaceae, Balsaminaceae, Gesneriaceae and Iridaceae) that occur in forested 

habitat along the Eastern seaboard of southern Africa. 

A study on pollination systems in Brownleea (Larsen et al. 2008) confirmed that in a 

Grahamstown population (EC, South Africa), B. coerulea is pollinated by S. 

wiedemanni, with flies at this site also visiting stands of Hypoestes aristata (Vahl) Sol. 

ex. Roem & Schult. (Acanthaceae) which may act as a magnet species; this species 

was included in the S. wiedemanni guild by Potgieter & Edwards (2005). Interestingly, 

B. coerulea at Umtamvuna was rather found to be pollinated by a tabanid fly, Ph. 

aethiopica, with P. ciliatus creating a possible magnet effect at the study site which was 

a forest patch situated next to grassland (Larsen et al. 2008). 

This type of habitat is similar to that occupied by succulent forms of P. saccatus with 

short, white flowers that have blue speckling on the corolla limbs, which superficially 

resemble the speckled nectar guides of P. ciliatus (and that of B. coerulea). A few 

plants of B. coerulea were seen amongst the population of P. saccatus studied at 

Beacon Hill, Umtamvuna NR, but for which no pollinator data was recorded. It is 

possible that Ph. aethiopica pollinates this form of P. saccatus at Umtamvuna NR, 

since the average proboscis length (8.9 mm) listed in Table 2 of Larsen et al. (2008) 

corresponds favourably to the 5 – 7 mm tube length of P. saccatus at this site. It is, 

however, also possible for acrocerid flies or anthophorine bees to be the pollinators of 

this Plectranthus species (see Appendix, P. saccatus). The month of observation of B. 

coerulea at Umtamvuna NR is not given by Larsen et al. (2008), but it is likely that S. 

wiedemanni had not yet emerged at that site at the time of the study, since the fly 

species has been recorded from forests at that site (C. Potgieter unpubl. data). At the 

Kologha Forest study site where P. ciliatus was observed for pollinators in the current 

Plectranthus study (near Stutterheim, EC), B. coerulea was seen to be pollinated by S. 

wiedemanni, with a voucher carrying pollinaria of the orchid (Potgieter & Edwards 

2005). 

Stenobasipteron wiedemanni is responsible for the evolution of specialised long-tubed 

corollas in four species (or forms) of Plectranthus, three of which are endemic to the 

Pondoland Centre of Endemism. The long-tubed form of P. saccatus that occurs on 

Chapter 8/ 106

steep forest slopes of the Umtamvuna Gorge does not have the exact floral 

morphology as the long-tubed P. saccatus endemic to the Hlatikulu Reserve area 

(Kaliweni forest) in far northern KZN; the latter form has a vertically narrower corolla, 

but with similar or shorter tube length, and was presumed to be pollinated by S. 

wiedemanni as well. A search for the pollinator in early February 2008 at Hlatikulu 

Reserve did not reveal any Stenobasipteron activity, but visits by Philoliche aethiopica 

(with shorter proboscis than S. wiedemanni) were noted to long-tubed P. saccatus, as 

well as some visits by apinid bees that tended to be brief. It is possible that 

Stenobasipteron had not yet emerged so early in the flowering season, since 

specimens of a fly species of that genus have been reported from that site 

(Barraclough pers. comm.). This species appears to be distinct from S. wiedemanni 

and more collections are needed to confirm its identity. 

As is the case for all the varieties of P. saccatus (see Appendix), more field work is 

needed to tease out the diversity of floral forms in this species, some of which are 

pollinated by S. wiedemanni and possibly other long-proboscid flies. 

The topic of Chapter 6 is not centred on pollination, but has relevance to understanding 

the importance of natural hybrids and hybridisation events in the group. The paper by 

Viljoen et al. (2006) tests microdistillation as a method for analysing small samples of 

essential oil-containing material and succeeds in showing that this technique, which 

may be used to analyse very small samples, compares well with hydro-distillation. It 

also shows that for a case study of a putative hybrid, P. zuluensis x P. ciliatus at Oribi 

Gorge, the essential oil profile of the confirmed hybrid plant is intermediate to those of 

the parent species. It is worth noting that flowers of the putative hybrid had four 

functional stamens, despite the fact that P. zuluensis only has two functional stamens. 

A number of cases of suspected natural hybridisation were found during the course of 

the study and in most cases it was possible to speculate on the parentage of the 

hybrids, based on plant morphology, with nectar guide and other floral characteristics 

proving particularly useful. All the cases of suspected hybridisation were between 

species with medium- and/or short corolla tube lengths, with no hybridisation events 

recorded between long- and shorter-tubed species (as discussed in Potgieter & 

Edwards 2001, Chapter 4). This shows that long corolla tubes offer greater pollinator 

fidelity while shorter tubes tend to be visited by a larger array of insect pollinators that 

also visit other related species (see Appendix, and Potgieter et al. 1999, Chapter 2). 

Most hybrids observed during the study were not fertile (no fruits or remaining calyces 

Chapter 8/ 107

were seen on older inflorescences), even in the case of P. ciliatus x P. zuluensis where 

a large, unidentified species of Prosoeca was seen to visit the hybrid inflorescence 

(see Appendix, P. zuluensis). 

In at least one case the pollinator of a species could be inferred from the presence of a 

hybrid plant, even when pollinators were never observed for the species, as was the 

case in P. ernstii (see Appendix, P. ernstii). 

In some Plectranthus populations (e.g. at Ngeli and Oribi Gorge) extensive hybrid 

stands were present due to vegetative propagation, since the genus grows very easily 

from cuttings and even small bits of broken-off stem will grow under favourable 

conditions. In some of these populations hybrid swarms were evident, with one or both 

species apparently crossing back onto parent clones, e.g. stands of P. ciliatus x P. 

fruticosus at Magwa (see Appendix, P. ciliatus); in such cases some hybrids would 

have had to be fertile in order to cross back. 

More studies involving essential oil composition will allow the identity of more hybrids to 

be revealed, especially in contentious cases such as Plectranthus brevimentum 

T.J.Edwards from Lupatana River Gorge near Port St Johns, EC (Edwards 2005). The 

author of the species presents it as a distinct species, while Van Jaarsveld (2006) 

suggests it to be of hybrid origin. In his predominantly horticultural review of 

Plectranthus in southern Africa, Van Jaarsveld (2006, p 104) maintains that P. 

brevimentum (incorrectly referred to as “P. brevilabrum Edwards”) represents a hybrid 

of P. hilliardiae and P. strigosus Benth. (a very short-tubed species that occurs in the 

current study area, but was not studied). If this is the case it would be the first recorded 

natural hybrid between a long-tubed and short-tubed species. A number of very 

attractive horticultural hybrids appear to involve P. hilliardiae and shorter-tubed species 

(e.g. Plectranthus Mona Lavender), hence crosses between long-and short-tubed 

species are possible. 

The nectar studies of Chapter 7 represent the most complete set of nectar data for the 

genus. As is the case in most genera of the Lamiaceae, Plectranthus nectar tends to 

be dominated by sucrose, with a few notable exceptions. 

One aspect of this study that has not yet been fully published is pollination of a number 

of Plectranthus species by acrocerid flies of the genus Psilodera. The pollinator data 

contained in the Appendix (see P. ciliatus, P. zuluensis, P. oertendahlii, P. 

Chapter 8/ 108

praetermissus, P. madagascariensis and Aeollanthus parvifolius) is being compiled into 

a paper on this topic (Potgieter et al. in prep.), since limited mention has been made of 

this pollinator group in other papers published from this thesis. There is a growing 

appreciation of the importance of this group of medium-proboscid flies in the pollination 

literature, with Borkent & Schlinger (2008a, 2008b) describing aspects of the floral 

visitation behaviour and pollen-carrying ability of Eulonchus spp. (Acroceridae) in North 

America. They point out that this family of flies has potential to form an important part 

of the pollinator fauna, but that more research is needed on the topic (Borkent & 

Schlinger 2008a). This mirrors the opinion expressed in this thesis, with Potgieter et al. 

(1999) suggesting that the behaviour of acrocerid flies would make them important 

pollinators of ‘medium-tubed’ plant species, including several Plectranthus species, 

while Potgieter et al. (2009) mentions that these flies fill a similar niche, w.r.t. proboscis 

length, to apinid bee pollinators. 

A total of five different species of Acroceridae, representing two sub-genera of 

Psilodera (C. Conway & D. Barraclough pers. comm.) were observed as potential 

pollinators of Plectranthus. Some species show wide distribution ranges, e.g. Ps. valida 

collected from Karkloof to Stutterheim (see Appendix, P. ciliatus), while others are 

potentially new species with apparent restricted distribution (Psilodera aff. confusa from 

Umtamvuna NR). Most acrocerid species were nectar-feeding floral visitors to straighttubed 

species of Plectranthus that matched the proboscis length of the pollinator. In 

species such as P. oertendahlii and P. praetermissus, both of which are narrow 

Pondoland endemics occurring in forest habitat, the corolla shape is basally saccate 

with a distinct narrowing at the mouth of the corolla, which restricts the range of floral 

visitors to insect species with thin proboscides, capable of hovering, such as Psilodera 

spp. (see Appendix, P. oertendahlii & P. praetermissus). 

This study provides a basis for future studies of Lamiaceae pollination, breeding 

systems and speciation. In addition to conducting more field studies to elucidate the 

pollinators for Plectranthus species that have not yet been recorded, a number of 

questions beg to be answered. 

Field-based experimental studies could investigate whether Plectranthus species are 

pollen-limited, as was found in insect-pollinated Syncolostemon (Ford & Johnson 

2008), and whether apinid bees or nemestrinid flies are more efficient at pollen carryover. 

Following on from the study on P. laxiflorus, conducted at Ferncliff in 2003 

Chapter 8/ 109

(Potgieter et al. 2009), developing inflorescences could be bagged and, once matured, 

opened to allow one flower to receive one visit from either pollinating group, and closed 

up again to record seed set. 

Studies that track gene flow in Plectranthus, such as the one using microsatellite and 

RFLP markers in two species of Streptocarpus (Gesneriaceae), by Hughes et al. 

(2007), would be useful in understanding the Pondoland Plectranthus distribution 

patterns and diversity. The forest-dwelling species of Plectranthus are trapped in deep 

gorges by aspects of their life history, such as the inability to disperse, and intolerance 

of desiccation and fire (on surrounding grassland plateaux), which favours allopatric 

speciation (Edwards 2005). Streptocarpus presents a similar situation and the forestdwelling 

species, S. primulifolius, which is pollinated by the nemestrinid fly S. 

wiedemanni, shows patterns of infrequent seed-dispersal coupled with limited 

movement of pollen between populations (Hughes et al. 2007). 

The inclusion of Tetradenia in future pollination studies will broaden the scope for 

analyses of floral diversification in the Plectranthus clade. Tetradenia is embedded within 

the Plectranthus clade and is the likely sister-group to Thorncroftia (Edwards 2006). Unlike 

Plectranthus and most of its allies, Tetradenia has small, unisexual flowers that appear 

almost actinomorphic. A study of its pollination syndrome would highlight the ability of 

clades to diverge i.t.o. floral morphology, and may explain its anomalous floral design. 

Finally, by building on the work done by Paton et al. (2004), an extended phylogeny 

that incorporates all of the southern African endemic and specialised species of 

Plectranthus, would provide the ideal platform for testing theories w.r.t. pollinator shifts 

and convergence in this large genus. 

Chapter 8/ 110

REFERENCES 

Barraclough, D.A., 2005. A review of the type material of the Southern African genus 

Stenobasipteron (Diptera: Nemestrinidae), with transfer of two species to 

Prosoeca Schiner, 1867. Zootaxa 1094: 41–51. 

Barraclough, D.A., 2006. An overview of the South African tangle-vein flies (Diptera: 

Nemestrinidae), with an annotated key to the genera and a checklist of species. 

Zootaxa 1277: 39–63. 

Borkent, C.J., Schlinger, E.L., 2008a. Flower-visiting and mating behaviour of 

Eulonchus sapphirinus (Diptera: Acroceridae). Canadian Entomologist 140: 

250–256. 

Borkent, C.J., Schlinger, E.L., 2008b. Pollen loads and pollen diversity on bodies of 

Eulonchus tristis (Diptera: Acroceridae): implications for pollination and flower 

visitation. Canadian Entomologist 140: 257–264. 

Eardley, C.D., 1983. A taxonomic revision of the genus Xylocopa Latreille 

(Hymenoptera: Anthophoridae) in southern Africa. Entomology Memoir, 

Department of Agriculture, Republic of South Africa 58: 1–67. 

Eardley, C.D., 1994. The genus Amegilla Friese (Hymenoptera: Anthophoridae) in 

southern Africa. Entomology Memoir, Department of Agriculture, Republic of 

South Africa 91: 1–68. 

Edwards, T.J., 2005. Two new Plectranthus species (Lamiaceae) and new distribution 

records from the Pondoland Centre of Plant Endemism, South Africa. Bothalia 

35: 149–52. 

Edwards, T.J., 2006. Notes on the Lamiaceae: A new Tetradenia and a new Thorncroftia 

from South Africa. South African Journal of Botany 72: 202–204. 



Goldblatt, P., Manning, J.C., 1998. Gladiolus in Southern Africa. Vlaeberg, Fernwood 

Press, pp 88–89. ISBN1874950326. 


Gladiolus (Iridaceae). Annals of the Missouri Botanical Garden 86: 758–774. 


southern Africa. Annals of the Missouri Botanical Garden 87:146–170. 

Hughes, M., Möller, M., Edwards, T.J., Bellstedt, D., De Viliers, M., 2007. The impact of 

pollination syndrome and habitat on gene flow: a comparative study of two 

Streptocarpus (Gesneriaceae) species. American Journal of Botany 94:1688– 

1695. 

Larsen, M.W., Peter, C., Johnson, S.D., Olesen, J.M., 2008. Comparative biology of 

pollination systems in the African-Malagasy genus Brownleea (Brownleeinae: 

Orchidaceae). Botanical Journal of the Linnean Society 156: 65–78. 

Chapter 8/ 111




Molecular Phylogeny and Evolution 31: 277-299. 

Potgieter, C.J., Edwards, T.J., 2001. The occurrence of long, narrow corolla tubes in 

southern African Lamiaceae. Systematics and Geography of Plants 71: 493– 

502. 




Potgieter, C.J., Edwards, T.J., Barraclough, D., Van Staden, J., In prep. Pollination of 

Plectranthus by medium-proboscid Psilodera spp. (Diptera: Acroceridae) in 

southern Africa. 

Potgieter, C.J., Edwards, T.J., Miller, R.M., Van Staden, J., 1999. Pollination of seven 

Plectranthus spp. (Lamiaceae) in southern Natal, South Africa. Plant 




Botany 75: 646–659. 

Van Jaarsveld, E., 2006. The South African Plectranthus and the art of turning shade 

into glade. Fernwood Press, Cape Town. ISBN9781874950806. 


Microdistillation and essential oil chemistry – a useful tool for detecting 

hybridisation in Plectranthus (Lamiaceae). South African Journal of Botany 72: 

99–104. 

Chapter 8/ 112

APPENDIX: 

Descriptive and pollinator accounts for twenty study species 

This appendix provides a summary of the main results from pollinator field observations 

for the twenty studied plant species, as a complementary reference to the published 

and nectar chapters. Each account includes a short description of the habitat, habit, 

distribution and phenology of the species, with a description of inflorescence design 

and floral morphology. The basic inflorescence structure in Plectranthus is cymose. 

Each cyme is subtended by a bract and these are arranged in a decussate 

synflorescence, resembling a raceme or a panicle. The species accounts also include 

study sites and observation details, with corresponding pollinator and insect voucher 

information. 

Descriptions are based on observations and measurements made during the study, 

with additions from the revision by Codd (1985a), as well as the following publications: 

Dyer (1934), Verdoorn (1949), Dyer & Bruce (1951a), Dyer & Bruce (1951b), Lewis 

(1951), Codd (1957), Codd (1970), Codd (1975), Codd (1977), Codd (1979), Codd 

(1980), Codd (1982a), Codd (1982b), Codd (1985b), Codd (1985c), Van Jaarsveld & 

Edwards (1991), Ryding (1993), Codd (1994), Van Jaarsveld & Edwards (1997), Van 

Jaarsveld & Van Wyk (2001). 

Insect vouchers are lodged at the Natal Museum, Pietermaritzburg (Diptera), or with 

the Biosystematics Division, Plant Protection Research Institute, Pretoria 

(Hymenoptera). Family classification for Hymenoptera follows Brothers (1999). 

Voucher details are followed by date of collection (or observation where no voucher 

was collected), and locality. 

Plant vouchers collected during the study, cited in the various publications and 

chapters, are lodged at the Bews Herbarium (NU).

The following abbreviations are used: 

NV: No voucher collected; 

NR: Nature Reserve; 

Localities 

DV: Dargle Valley, KwaZulu-Natal Province (KZN) Midlands; 

FC: Ferncliff Nature Reserve, Pietermaritzburg, KZN Midlands; 

HF: Hlatikulu NR, Kaliweni forest, northern KZN; 

KF: Kologha forest, Stutterheim, Eastern Cape Province (EC); 

KK: Karkloof, Leopard’s Bush Nature Reserve, KZN Midlands; 

MF: Magwa Forest, EC Coast; 

NG: Ngeli Forest, Weza, southern KZN; 

OG: Oribi Gorge Nature Reserve, KZN South Coast; 

ON: Ongoye Nature Reserve, near Empangeni, KZN North Coast; 

PMB Gdn: Pietermaritzburg, in a garden, KZN Midlands; 

PSJ: Port St Johns, Bulolwe River, EC Coast; 

UNR: Umtamvuna Nature Reserve, KZN South Coast; 

UV: Umgeni Valley Nature Reserve, Howick, KZN Midlands; 

WV: World’s View, near Ferncliff, Pietermaritzburg, KZN Midlands. 

Appendix/ 114

Study species, grouped according to corolla tube shape and length 

A. Long-tubed species with straight corollas; corolla tube lengths 20 – 33 mm 

Species from this group are included in Potgieter et al. (1999) and Potgieter & 

Edwards (2005). 

1. Plectranthus ambiguus (Bolus) Codd Tube 20 – 33 mm 

2. Plectranthus hilliardiae Codd Tube 21 – 32 mm 

3. Plectranthus reflexus E.J. van Jaarsv. & T.J.Edwards Tube 24 – 30 mm 

4. Plectranthus saccatus Benth. Tube 20 – 28 mm 

B. Shorter-tubed species with straight corolla; corolla tube lengths 4 – 18 mm 

Five species from this group, indicated with *, are included in Potgieter et al. 

(1999). 

5. Plectranthus ecklonii Benth. * Tube 10 – 15 mm 

6. Plectranthus zuluensis T.Cooke * Tube 10 – 16 mm 

7. Plectranthus ciliatus E.Mey ex Benth. * Tube 6 – 8 mm 

8. Plectranthus ernstii Codd Tube 4 – 8 mm 

9. Plectranthus oribiensis Codd * Tube 6 – 12 mm 

10. Plectranthus fruticosus L’Hér. Tube 5 – 13 mm 

11. Plectranthus oertendahlii T.C.E.Fr. Tube 8 – 13 mm 

12. Plectranthus praetermissus Codd Tube 12 – 15 mm 

13a. Plectranthus madagascariensis (Pers.) Benth. * Tube 5 – 6 mm 

13b. Plectranthus hadiensis (Forssk.) Schweinf. ex Spreng. Tube 7 –18 mm 

C. Sigmoid-tubed species; with corolla tube lengths 5 – 11.5 mm 

Species from this group are included in Potgieter et al. (2009). 

14. Plectranthus petiolaris E.Mey. ex Benth. Tube 7– 11 mm 

15. Plectranthus laxiflorus Benth. Tube 10 – 11 mm 

16. Plectranthus calycinus Benth. Tube 6 – 7 mm 

17. Plectranthus rehmannii Gürke Tube 5 mm 

18. Plectranthus spicatus E.Mey. ex Benth. Tube 5 mm 

19. Pycnostachys urticifolia Hook. Tube 10.5 – 11.5 mm 

20. Aeollanthus parvifolius Benth. Tube 7 – 10 mm 

Appendix/ 115

1. Plectranthus ambiguus (H.Bol.) Codd 

Habit, habitat, distribution and phenology 

Plectranthus ambiguus is a soft herb with erect or decumbent stems, growing 0.4 – 1.2 

m in height, in clearings in forest, on forest margins and on shaded rocky slopes. It 

occurs along semi-coastal and coastal areas of the Eastern Cape towards Ongoye 

Forest in KZN. It flowers from January to April and produces a showy mass of purple to 

violet flowers in the forest understorey. 

Inflorescence and floral morphology 

Synflorescences are simple (or sparingly branched), congested panicles, 40 – 170 mm 

in height. Flowers are purple to violet, with narrow, straight, long corolla tubes (20 – 33 

mm) that expand slightly towards the mouth to 2 mm deep. The upper lip is 4 – 5 mm 

long with fine vertical dark lines as nectar guides; the concave lower lip is 3 – 5 mm 

long. Four free stamens and the style extend 6 – 7 mm and 6mm respectively from the 

corolla mouth. 

Study sites and observations 

Most pollinator observations for this species were done at Umtamvuna NR where large 

patches occurred along the forested slopes of the Bulolo River Gorge. One large 

population was mixed with Plectranthus ecklonii and insects frequently moved from the 

one species to the other. Less dense populations were studied at Oribi Gorge NR. 

Observation time on P. ambiguus was over a number of days in different years for at 

least 20 hours, during the main period of insect activity from 9.00 am to 4.00 pm. 

Pollinators 

The long-proboscid nemestrinid fly, Stenobasipteron wiedemanni, is the main pollinator 

for this species; few other insect visitors can reach nectar at the base of the long 

corolla tube. Visits by S. wiedemanni lasted from 1 – 4 seconds per flower, with the fly 

hovering into position, then settling on the exserted stamens and style while probing for 

nectar. Pollen was deposited ventrally on the thorax, especially in the hairy area 

between the bases of the legs, and on the abdomen. The length of each visit appeared 

to be dictated by the amount of nectar at the corolla base where the nectary is 

positioned. Flies tended to forage from the base to the top of the inflorescence, often 

visiting a few flowers per inflorescence before moving on to the next. 

Appendix/ 116 

P. ambiguus

Other visitors included the apinid bees Amegilla mimadvena and Xylocopa hottentotta, 

both with proboscis lengths too short to extract nectar (mean length 9 and 8 mm 

respectively), unless flowers on the short end of the range were visited while nectar 

levels were very high. Syrphid flies (two species, one being Allobaccha sp.), apinid 

honeybees (Apis mellifera) and Allodape pernix collected pollen from the anthers; the 

latter also robbed nectar from holes made in the base of the corolla tissue near the 

nectary. 

Insect vouchers 

Stenobasipteron wiedemanni (Nemestrinidae) 

Potgieter 1 UNR, 17-3-95 



Prosoeca umbrosa (Nemestrinidae) 


Allobaccha sp. 2 (Syrphidae) 

Potgieter 8 

Syrphidae sp. 1 

UNR, 17-3-95 


Apis mellifera (Apidae) 


Allodape pernix (Apidae) 






Amegilla mimadvena (Apidae) 

NV 

Xylocopa hottentotta (Apidae) 

NV 

Appendix/ 117 

P. ambiguus

2. Plectranthus hilliardiae Codd 


Plectranthus hilliardiae is an erect, short herb growing 300 – 650 mm tall. It is semisucculent 

with a branched habit. It grows among rocks – often in sandy soil near rivers 

– in forest areas. The species is endemic to coastal forests from the Umtamvuna area 

in KZN, southwards to the Mbotyi/Magwa area in the Eastern Cape. It is locally sparse 

to abundant in a narrow niche, but in some years formed large displays, especially 

along paths in the Umtamvuna NR. It flowers from December to April. 


Synflorescences are simple racemes / panicles, 80 – 150 mm in height, with one or two 

branch pairs at the base. The corolla tube is slightly deflexed at the base, straight and 

long (21 – 32 mm) with a saccate base; the tube is parallel-sided or slightly narrowed 

towards the mouth (from 5 – 3 mm deep). Flowers are pale blue with nectar guides as 

dark purple flecks on the inside of the upper and lower lips. The upper lip is 5 – 6 mm 

long and the concave lower lip is 4 mm; the latter may sit horizontally or be deflexed, 

depending on floral age. The four, free stamens are didynamous, with the lower pair 

extending up to 8 mm from the corolla mouth. The style extends 8 – 10 mm from the 

corolla mouth. 


Pollinator observations were mainly conducted at Umtamvuna NR, along the Bulolo 

River, which is a tributary to the Umtamvuna River where Codd (1985a) recorded the 

species as being endemic. Subsequent collections (Van Jaarsveld & Van Wyk 2001) 

extended the distribution south to the northern Eastern Cape (Transkei) where 

populations were found at Mbotyi Forest, Magwa Falls, Fraser Falls, Myokane Gorge 

and Noyokaan Gorge (subsp. australis sensu Van Jaarsveld & Van Wyk), as well as 

Lupatana Gorge and Mkambati Gorge (which, with Umtamvuna, represents the 

distribution of subsp. hilliardiae sensu Van Jaarsveld & Van Wyk). We conducted 

shorter periods of pollinator observations at Lupatana Gorge and Magwa Gorge. Total 

observation time was at least 20 hours over a number of days in different years, 

between 9.00 am and 4.00 pm. 

Appendix/ 118 

P. hilliardiae

Our observations suggest that the glossy-leafed population at Lupatana Gorge is closer 

to subspecies australis than hilliardiae. This subspecies or form is much easier to 

cultivate than the typical subspecies. 

Pollinators 

The main pollinator of P. hilliardiae is the long-proboscid nemestrinid fly, 

Stenobasipteron wiedemanni. The fly hovered in front of the flower, briefly settling on 

the exserted stamens and style, while probing for nectar. Pollen is deposited ventrally 

on the thorax and abdomen. Pollinator sightings for this plant species were much less 

frequent than observations on other species that occur in the same locality, such as P. 

ambiguus. During 15 hours of observation on P. hilliardiae, on sunny days, only three 

observations of visits by S. wiedemanni were made, compared to 25 visits to P. 

ambiguus over the same amount of time. 

The only other recorded visit to P. hilliardiae was by an acrocerid fly (Psilodera aff. 

confusa) which probed a flower briefly. This fly could not reach nectar with its mediumlength 

proboscis (16 mm), even if it extended the proboscis slightly, but may have 

attempted a visit since other medium-tubed species of Plectranthus look superficially 

similar. 

Nectar robbing visits by the apinid bee, Allodape pernix, were observed on P. hilliardiae 

flowers, with about 30% of flowers showing robbing holes on the dorsal side of the 

saccate corolla base, near the nectary 




Psilodera aff. confusa (Acroceridae) 

Potgieter s.n. UNR, 17-3-95 

Patellapsis sp. (Apidae, Halictinae) 



NV 

Appendix/ 119 

P. hilliardiae

3. Plectranthus reflexus Van Jaarsv. & T.J.Edwards 


Plectranthus reflexus is a semi-succulent, soft herb with an erect habit. It grows 1 – 1.5 

m tall and has few, ascending branches. It is a narrow endemic of densely shaded 

coastal forest along the Bulolwe River at Port St Johns in the Eastern Cape (Transkei), 

and further south at Mkambati (D. Styles pers. comm.). Flowering occurs between 

January and March. 


Synflorescences are lax, terminal racemes (150 – 200 mm long), with occasional 

subtending lateral branches. Flowers are pale blue without any markings. Floral tubes 

are long (24 – 30 mm) with a slightly saccate base 3 mm deep, tapering to 1 mm in the 

corolla mouth. The upper lip is reflexed to the extent that it may touch the corolla tube, 

as is the boat-shaped lower lip. Stamens are four, free and didynamous, with the lower 

pair exserted by 12 mm and the upper pair by 6 mm from the corolla mouth. Stamens 

coil away in older flowers and the style extends 8 – 9 mm from the corolla mouth. 


The population of P. reflexus was visited once in March 1998 at Port St Johns. About 

five hours of observation time was spent with the population, from mid-morning to midafternoon. 

Pollinators 

The nemestrinid fly, Stenobasipteron wiedemanni, is the main pollinator, but a brief visit 

to one inflorescence by a Citrus Swallowtail butterfly (Papilionidae) was also seen. Due 

to the long proboscis of this butterfly it is unlikely that pollen was deposited on the 

insect’s body. The reflexed upper and lower corolla lips make it difficult for insects to 

settle on the flowers, hence hovering insects with long proboscides are at an advantage 

for reaching nectar. Nemestrinid flies grasped the extended stamens and style during 

visitation, thus ensuring pollen carry-over on the fly body. Pollen was deposited 

ventrally on the thorax and abdomen of S. wiedemanni. 

No bees were seen to visit this species, other than a pollen-collecting apinid Allodape 

pernix. A syrphid fly was observed eating pollen. 

Appendix/ 120 

P. reflexus



NV on P. reflexus (but voucher caught soon after on P. praetermissus, at 

PSJ: Potgieter 191) 


Potgieter 119 PSJ, 9-3-98 



Appendix/ 121 

P. reflexus

4. Plectranthus saccatus Benth. 


Plectranthus saccatus is a freely branched, erect to spreading soft shrub with semisucculent 

stems. It grows 0,5 – 1,2 m tall in coastal or semi-coastal forests or areas of 

semi-shade. It is distributed from the southern Transkei area of the Eastern Cape, 

northwards to northern KZN. It flowers from December to April (and sporadically in 

June). 


Synflorescences are simple racemes (occasionally branched near the base), 50 – 120 

mm in height, with relatively few, deflexed flowers per inflorescence, of which only a 

couple are open at any time. The large mauve to pale blue or white flowers make a 

striking display in the forest understorey. Floral tubes are markedly saccate at the base 

(up to 6 mm), with laterally compressed parallel sides that may narrow slightly towards 

the mouth. The upper lip is upright and the boat-shaped lower lip starts off horizontal 

and deflexes with floral age. The upper lip may be unmarked or heavily blotched with 

purple around the edges; forms occur with fine purple speckling across upper and 

lower lips. The four stamens are free and exserted from the corolla mouth by 8 – 10 

mm; the style is exserted by ca. 10 mm. 

This is a highly variable species and two varieties were recognized by Codd (1985a). 

P. saccatus var. saccatus has a shorter corolla tube (8 – 16 mm long) with broader 

vertical sides (5 – 6 mm deep at the base), while P. saccatus var. longitubus has a 

longer corolla tube (20 – 28 mm), but narrower vertical sides (4 – 5 mm deep at the 

base). The latter variety occurs at the northernmost range in the Kaliweni Forest near 

Ingwavuma, with some long-tubed populations further south at Umtamvuna NR. 

On the basis of succulence two subspecies were recognized by Van Jaarsveld & 

Edwards (1997). Plectranthus saccatus Benth. subsp. pondoensis Van Jaarsv. & 

Milstein has succulent leaves, a procumbent to decumbent habit and flexible stems that 

may reach 4 m in length. This subspecies occurs in scrub near forest along the lips of 

gorges, while P. saccatus subsp. saccatus occurs in forest understorey. The variation 

in corolla tube length in this species presents an apparent continuum, which questions 

the validity of recognizing the variety P. saccatus var. longitubus Codd (Van Jaarsveld 

Appendix/ 122 

P. saccatus

& Edwards 1997). For the purposes of this study we refer to populations only, referring 

to long-tubed, medium-tubed and short-tubed forms for convenience. 


Pollinator observation was attempted at four populations of this species. At 

Umtamvuna NR the mauve, long-tubed population (with no nectar guides and tubes 20 

– 30 mm long) was studied in the forest understorey along the steep slopes of the 

Umtamvuna Gorge. One study visit was made to Hlatikulu NR at Kaliweni forest in 

northern KZN in early February 2008, with ca. eight hours spent at two different 

populations of the long-tubed form of P. saccatus. Another, short-tubed (5 – 7 mm), 

white- to pale mauve-flowered variety (with dark blue speckles as nectar guides) with 

succulent stems, was studied on the forest edge above the Bulolo River Gorge. At Oribi 

Gorge NR a medium-tubed population (9 – 11 mm) with pale mauve flowers, speckled 

upper lips and succulent stems was studied. At least 23 hours of observation time 

included observations from early morning to dusk over a number of days, in different 

years. 

Pollinators 

Pollinators of this species proved to be the most elusive at the Pondoland study sites, 

despite many hours of observation time. A long-proboscid nemestrinid fly (S. 

wiedemanni) was recorded on the long-tubed population at Umtamvuna NR. This 

population also showed evidence of nectar-robbing at the bases of most of the open 

flowers on each plant, as well as on fallen corollas. The apinid bee, Allodape pernix, 

was even seen making holes in the bases of unopened flowers in search of nectar. 

No observations were made on the short-tubed variety at Umtamvuna, but based on 

corolla tube dimensions it is highly likely that an acrocerid fly of the genus Psilodera, or 

an apinid bee of the genus Amegilla, is the pollinator. 

At Oribi Gorge one brief visit to a flower of the medium-tubed population was seen at 

dusk (6.15 pm in April 1998), but the identity of the visitor remains unclear. It looked 

like a brown moth, which would make sense considering the time of day (and 

nemestrinids had not been seen out at that time before), but Plectranthus has no scent 

typical of that which would attract moths at night. It is likely that S. wiedemanni, or one 

of the shorter-proboscid Prosoeca species, is the pollinator at this site. Repeated 

visitations of the long-proboscid S. wiedemanni were seen to a medium-tubed form of 

Appendix/ 123 

P. saccatus

P. saccatus (12 – 14 mm) in cultivation in a garden in New Germany, Durban area (A. 

Beaumont pers. comm.). 

At the northern KZN site, where P. saccatus with long corolla tubes were prominent in 

certain parts of the forest understorey, growing alongside a species of Justicia L. 

(Acanthaceae), no nemestrinid flies were active at the start of the flowering season. A 

few tabanid flies (Ph. aethiopica), were seen to visit the flowers, and appeared to reach 

nectar. Apinid bee species such as Amegilla mimadvena were frequent visitors to P. 

saccatus flowers, but spent more time on the adjacent Justicia species. Fewer visits 

were noted by A. bothai bees which attempted visits to P. saccatus but then moved on 

the more accessible Justicia flowers. Individuals of Xylocopa hottentotta appeared to 

occasionally reach nectar in P. saccatus flowers, but often relocated to the base of the 

flower to pierce holes for robbing nectar. 

The paucity of pollinator observations for P. saccatus in the Pondoland area is 

unfortunate, since the range in floral tube length is the most extensive in the genus. 

This aspect of the project will be extended in future years to ascertain why there are so 

many different floral varieties in this species. 



NV UNR, 18-3-96 

Philoliche aethiopica (Tabanidae) 

Potgieter 300 HF, 5-2-08 






Potgieter 304 

Amegilla bothai (Apidae) 

HF, 5-2-08 


Appendix/ 124 

P. saccatus

5. Plectranthus ecklonii Benth. 


Plectranthus ecklonii is a tall, erect, robust sub-shrub that grows 0.7 – 2.5 m tall; stems 

are succulent below with ascending branches. It is locally common forming populations 

along forest margins, clearings and stream banks. It has a coastal, semi-coastal and 

midlands distribution from the Eastern Cape, through KZN into Mpumalanga Province. 

Flowering time extends from January to April (and sporadically in August). 

Inflorescence design and floral morphology 

Plants produce a showy display of dark blue-purple or mauve flowers on large, 

terminal, dense panicles (120 – 250 mm in height). Populations of pale blue, white and 

pink-flowered individuals have been recorded, but the populations studied in this 

project were of the commonest colour morph (purple). The whole inflorescence, 

including calyces, pedicels, peduncles and bracts, are similar in colour to the flowers. 

Corolla tubes are straight, laterally compressed and expanding from a narrow base (1 

mm deep) towards the mouth (3 mm deep) to form a trumpet shape. The tube is of 

medium length (10 – 15 mm); the upper lip is 5 – 6 mm and the concave lower lip is 4 – 

5 mm long. Nectar guides of darker purple irregularly shaped blotches occur on the 

upper lip. The four stamens are free and didynamous, extending up to 15 mm from the 

corolla mouth. Stamens coil away from the style as the flower ages; the style extends 

12 – 16 mm from the corolla mouth. 


Populations of P. ecklonii were studied at Oribi Gorge NR, Umtamvuna NR, Kologha 

Forest at Stutterheim and in cultivation at Pietermaritzburg. Observation time was at 

least 25 hours between 9.00 am and 4.00 pm over a number of days in different years, 

with similar amounts of time spent at Oribi Gorge and Umtamvuna, and one visit to 

Stutterheim. Mixed populations with P. ambiguus occurred at Umtamvuna NR. 

Pollinators 

The long-proboscid nemestrinid fly, Stenobasipteron wiedemanni, was a common 

visitor at the Stutterheim population, and was also seen at Oribi Gorge and 

Umtamvuna NR. Large-bodied, medium-long proboscid nemestrinid flies of the genus 

Prosoeca (P. umbrosa) were frequent visitors at Oribi Gorge and Umtamvuna NR; 

these flies ‘fit’ the corolla tube very well with a proboscis length of 15 – 16 mm. 

Appendix/ 125 

P. ecklonii

Nemestrinid flies hovered while probing the tubes, briefly settling on the stamens and 

style while extracting nectar. Visits lasted from 1 – 5 seconds. The proboscis was 

folded beneath the fly body during flight and was raised into a horizontal position on 

approach of a flower. Pollen was deposited ventrally on the head and thorax in the 

case of S. wiedemanni, and on the thorax and abdomen in P. umbrosa. 

Medium-proboscid tabanid flies (Philoliche aethiopica) and apinid bees of the genus 

Amegilla (A. caelestina and A. mimadvena), as well as Xylocopa hottentotta, were 

frequent visitors. These four species had proboscides 8 – 9 mm long and were able to 

access nectar if the body was forced into the trumpet-shaped corolla. This action 

brushed the ventral surface of the insect body over the exserted anthers and deposited 

pollen on the abdomen of the insect. In the female phase, when the anthers were 

coiled away, the pin-like style picked up pollen off the insect in the same way. 

Hummingbird hawkmoths (Macroglossus trochilus) were occasional nectar feeders, 

while syrphid flies and the apinid bee, Allodape pernix, collected pollen. The latter also 

robbed nectar by making holes in the corolla bases and buds. 

The flowers of P. ecklonii straddle the long- and medium-tubed guilds in that the corolla 

is of medium length and dilated at the mouth, which allows visits by medium-proboscid 

insects, as well as long-proboscid insects that pick up pollen on the body from the long 

anthers that approximate the length of long-tubed species. As a result more insect 

species visit P. ecklonii effectively than the four preceding long-tubed species. 



Potgieter 194 KF, 5-3-98 





Potgieter 5 UNR, 18-3-96 (= Prosoeca sp. A, Potgieter et al. 1999) 

Prosoeca sp. nov. 1 (Nemestrinidae) – close to Potgieter 171 

Potgieter 190 OG, 7-4-98 

Appendix/ 126 

P. ecklonii



Potgieter 39 PMB Gdn, 7-4-94 

Amegilla caelestina (Apidae) 



NV 


NV 

Macroglossus trochilus (Sphingidae) 

NV 





Appendix/ 127 

P. ecklonii

6. Plectranthus zuluensis T. Cooke 


Plectranthus zuluensis is an erect, soft sub-shrub that grows 1 – 2 m tall, with 

ascending branches, along streams and margins of the forest. It has a coastal and 

semi-coastal distribution from southern KZN northwards to central KZN, occurring 

again in Swaziland. There is an extended flowering season for this species: September 

to November and January to May. 


It is a showy species when in flower, with tightly packed, cylindrical blue 

synflorescences. These are racemose (occasionally branched near the base), 40 – 80 

mm in height. Flower colour in the study area was pale blue (sometimes appearing 

white), but dark blue forms also exist. Six rows of darker blue dots act as nectar guides 

on the upright upper lip. The corolla tube is of medium length (10 – 16 mm), slightly 

deflexed and laterally compressed, with a large saccate base that narrows towards the 

corolla mouth. Both the upper lip and concave lower lip are 5 – 6 mm long; the latter 

reflexes as the lower stamens fold away when the style elongates. The stamens are 

free with only the lower pair fertile, extending 5 – 7 mm from the corolla mouth. The 

upper pair is reduced to small staminodes that only extend 1 – 2 mm from the mouth. 

This is the only species with only two functional anthers. The style exerts 5 mm from 

the corolla mouth. 


Pollinator observations on P. zuluensis were only made at Oribi Gorge, where an easily 

accessible, large population occurred along the upper end of the gorge. At least 15 

hours were spent on observations over a number of days in different years (mostly 

1997 – 1998), from early morning (8 am) to late afternoon (5.30 pm). 

Pollinators 

The acrocerid fly, Psilodera confusa, was the commonest visitor. These flies have 

proboscis lengths (9.5 – 11 mm) matching or slightly shorter than that of the corolla 

tubes, which forced close contact between the fly bodies and the anthers or styles. The 

flies never occurred in large numbers, but one or two were seen visiting the population 

at any one time. The flies hovered in front of the flower and briefly settled on the sexual 

organs while probing for nectar. Pollen was deposited ventrally on the head and hairy 

Appendix/ 128 

P. zuluensis

thorax. These flies moved between P. zuluensis and an adjacent population of P. 

ciliatus and a hybrid between these two species was found close by. 

A less frequent visitor was the nemestrinid fly, Stenobasipteron wiedemanni. Despite 

the long proboscis of this fly, it could contribute to pollen carryover by brushing against 

the anthers with its proboscis, or in individuals at the short end of the proboscis range 

pollen may be carried on the hairy patch below the head. These flies also moved 

between P. zuluensis and the shorter-tubed P. ciliatus. One observation was made of a 

very large-bodied black fly (probably Prosoeca) visiting P. zuluensis and subsequently 

visiting the hybrid plant nearby, but this specimen could not be captured to verify 

identity. This fly was not observed again, but a similar-looking species was seen at 

Ongoye Forest. 


Psilodera confusa (Acroceridae) 







Appendix/ 129 

P. zuluensis

7. Plectranthus ciliatus E.Mey ex Benth. 


Plectranthus ciliatus is a low-growing soft, branching herb with decumbent to 

ascending stems, growing up to 0.6 m tall. It occurs in moist areas in forests and other 

shady places, with a coastal to midlands distribution from Knysna in the Western Cape, 

through the Eastern Cape and KZN and into Mpumalanga Province. It flowers from 

November to April. 


Synflorescences are fairly lax, simple racemes (sometimes with a pair of branches near 

the base), 50 – 200 mm in height. Flowers are white with dense purple speckles over 

the inside surfaces of the upper and lower lips. The white flowers contrast well with the 

dark green foliage (with purple undersides) in shaded, dark places. The corolla tube is 

medium to short (6 – 8 mm), slightly deflexed, with a saccate base, narrowing slightly 

towards the corolla mouth. The upper lip is 5 – 8 mm and the boat-shaped lower lip is 3 

– 7 mm long; the latter may be held horizontally or deflex as the flower ages. The four 

anthers are free and extend beyond the corolla mouth by 5 mm (upper pair) and 7 mm 

(lower pair). 


Pollinator observations were made at Oribi Gorge and Umtamvuna NR in southern 

KZN, Kologha Forest near Stutterheim and Magwa area in the Eastern Cape, 

Leopard’s Bush in the Karkloof and Ferncliff in Pietermaritzburg. At least 25 hours of 

observations were made over a number of days in different years, between early 

morning (8.00 am) and late afternoon (5.30 pm). 

Pollinators 

At Oribi Gorge the medium-proboscid acrocerid fly, Psilodera confusa (proboscis 9.5 – 

11 mm), was a common visitor to P. ciliatus. In Pietermaritzburg the acrocerid flies, 

Psilodera hessei (proboscis 8.5 – 9.5 mm) and Psilodera nhluzane (proboscis 12 mm), 

were common at Ferncliff, with Psilodera valida visiting P. ciliatus at a nearby 

population (World’s View). In the Karkloof and at Umgeni Valley, and further south at 

Stutterheim, this species (a yellow fly of a different sub-genus to Ps. confusa, with 

proboscis 8 mm), was observed visiting P. ciliatus. A similar yellow acrocerid species 

was seen visiting P. ciliatus and P. fruticosus at Magwa in an area where putative P. 

Appendix/ 130 

P. ciliatus

ciliatus x P. fruticosus hybrids occurred. It is clear that acrocerid flies are important 

pollinators of P. ciliatus over a large extent of its range. The different acrocerid species 

that visited P. ciliatus have proboscides in a common range (8 – 12 mm), which matched 

or were longer than the corolla tubes of P. ciliatus, but corresponded to the length of 

the stamens and style, which facilitated pollen carry-over on the ventral head surface 

and base of the proboscis of the fly. 

The long-proboscid nemestrinid fly, Stenobasipteron wiedemanni, also visited P. 

ciliatus populations in Stutterheim and at Oribi Gorge. The fly reached nectar without 

contacting the sexual organs of the plant with its body. While some pollen may be 

carried over on the proboscis, this is more a case of the fly exploiting the flowers for 

nectar. 

The apinid bee species Amegilla caelestina and A. bothai visited P. ciliatus at 

Umtamvuna NR. No acrocerid flies were observed on P. ciliatus at this a site, but 

Psilodera aff. confusa did occur at this site and may also pollinate the species. 

Likewise, species of Amegilla bees occur at the other study sites and may also visit P. 

ciliatus there. 

At Magwa both the apinid bees and acrocerid flies were seen visiting flowers of the 

same population of the tentative hybrid P. ciliatus x P. fruticosus. The bees were 

Amegilla mimadvena and Zebramegilla sp. These bees may function as effectively as 

pollinators as acrocerid flies, since they are similar in body shape and size and the 

proboscis lengths match that of the P. ciliatus corolla tube. In most of the cases where 

Amegilla bees and Psilodera flies visited the same species, they fulfilled a similar role 

in terms of body sizes and proboscis lengths. 

Other visitors to this tentative hybrid included a medium-proboscid nemestrinid fly 

(Prosoeca sp.) and three lepidopterans: the hummingbird hawkmoth Macroglossus 

trochilus, the papilionoid butterfly Papilio dardanus and a species of hesperid butterfly. 

These Lepidoptera were unlikely to be major pollinators, since the proboscides were 

much longer than the floral parts, which prevented bodily contact. Syrphid flies visited 

flowers at Oribi Gorge and Pietermaritzburg, and the short-proboscid halictinid bee, 

Lasioglossum sp., was observed at Pietermaritzburg. These species collected pollen, 

but are unlikely to reach nectar from the corolla base. 

Appendix/ 131 

P. ciliatus

Plectranthus ciliatus is a relatively common and widespread species of Plectranthus; 

this, coupled with its shorter corolla tube, allows the species to have many insect 

visitors, of which the various species of Acroceridae and Apinae are effective 

pollinators in different localities. 





Psilodera hessei (Acroceridae) 

Potgieter 108 FC, 7-3-97 




Psilodera nhluzane (Acroceridae) 


Psilodera valida (Acroceridae) 

Potgieter 31 WV, 11-4-96 


Potgieter 156 KK, 3-3-99 

Potgieter 221 MF, 19-4-02 

D. Martins s.n. UV, 18-3-05 







Potgieter 215 


MF, 19-4-02 (on P. ciliatus hybrid) 


Potgieter 94 

Amegilla bothai 

NV 

UNR, 2-3-97 

Appendix/ 132 

P. ciliatus

Amegilla mimadvena 

NV 



Papilio dardanus (Lepidoptera) 

NV 

Hesperidae (Lepidoptera) 


Asarkina sp. (Syrphidae) 

Potgieter 100 

Rhingia sp. (Syrphidae) 

FC, 7-3-97 


Zonalictus sp. (Apidae, Halictinae) 


Appendix/ 133 

P. ciliatus

8. Plectranthus ernstii Codd 


Plectranthus ernstii is a semi-succulent herb that branches from thickened stem bases; 

it reaches 250 mm in height and survives dry conditions with the potato-like stem 

bases. The species escapes fire by growing in pockets of humus-rich soil between 

rocks and on south-facing cliffs. It is an endemic species confined to the area from 

Oribi Gorge to Umtamvuna and Mkambati further south. Flowering time extends from 

February to April, but in cultivation the species flowers over a longer period of time. 


The pale bluish-mauve flowers are produced in relatively lax, often secund 

synflorescences, 30 – 120 mm in height. Corolla tubes are straight and of medium 

length (4 – 8 mm) with a saccate base 4 – 5 mm wide, narrowing to 2 mm wide at the 

corolla mouth. The upper lip is 4 – 5 mm, with fine vertical dark blue lines as nectar 

guides; the boat-shaped lower is lip 3 – 4 mm long. Four free, didynamous stamens 

extend by 1.5 mm (upper pair) and 3 mm (lower pair) from the corolla mouth. The style 

exerts by 2 mm. 


A population at Oribi Gorge was observed for a whole morning, but the rocky cliff face 

where the plants were growing was inaccessible and plants were observed from some 

distance. No pollinators were seen. A subsequent visit to another, slightly more 

accessible, population on a neighbouring farm was disappointing, since it was not in 

flower when expected. Observations made in Pietermaritzburg on cultivated plants in a 

garden showed the apinid bee Amegilla caelestina to be a constant visitor. This bee 

species occurs at Oribi Gorge and may prove to be the natural pollinator. It has the 

correct proboscis length to match the corolla tube of P. ernstii, as do other Amegilla 

species recorded at the site. The bees have also been observed in the kind of exposed 

areas where this species grows. 

A putative hybrid between P. ernstii and the endemic P. oribiensis was found amongst 

the P. ernstii population. In cultivation the hybrid showed vegetative features of P. 

ernstii and floral features of P. oribiensis. The latter is pollinated by four species of 

Amegilla at Oribi Gorge – evidence that at least one of these bee species must have 

visited the rocky P. ernstii site to transfer pollen. 

Appendix/ 134 

P. ernstii

9. Plectranthus oribiensis Codd 


Plectranthus oribiensis is a tall, erect, branched soft shrub with ascending stems, 

reaching up to 1.5 m in height. It grows along forest margins and in wooded kloofs on 

steep slopes, especially south-facing ones. According to Codd (1985a) this species is 

endemic to Oribi Gorge and the Umtamvuna River, but during the course of this study it 

was only found at Oribi Gorge. Flowering time extends from February to May. 


Mauve flowers are produced in lax racemose synflorescences with one or two pairs of 

branches near the base, up to 200 mm tall. The floral display is not striking, with 

inflorescences scattered at low density throughout the population. The upper and lower 

corolla lips have mauve hairs and white gland-dots on the outer surface. The 

compressed corolla tube is deflexed, straight and of medium length (6 – 12 mm). The 

base is saccate and extends into a short spur; the tube narrows slightly towards the 

corolla mouth. The upper lip is 5 – 6 mm and the boat-shaped lower lip 5 – 7 mm long; 

the latter deflexes as the flower ages. The four stamens are free, extending 2 – 3 mm 

from the corolla mouth, as does the style. 


Pollinators were observed at Oribi Gorge, mostly along the road that passes through 

the mid- to upper part of the gorge, where plants were abundant. Twelve hours of 

observations were made from February to April, during 1995, 1997 and 1998. 

Pollinators 

Mostly bees visit this species, with apinid bees of the genus Amegilla as the most 

abundant pollinator. Four species of Amegilla were recorded: A. caelestina, A. 

mimadvena, A. bothai and A. spilostoma. The proboscis lengths of these species range 

from 7 – 9 mm, which fits the range of P. oribiensis corolla tubes (7 – 9 mm). A bee 

would inserted the whole length of its proboscis into the corolla, and pollen would be 

deposited ventrally on the thorax. An apinid Thyreus sp. was a frequent visitor and 

Xylocopa flavorufa (proboscis 9 mm long) was also recorded. One visitation of a small 

lycaenid butterfly was observed. 

Appendix/ 135 

P. oribiensis

As discussed for P. ernstii, a tentative hybrid of P. oribiensis x P. ernstii was found near 

a population of P. ernstii. The range of corolla tube length of both overlap, and even 

though no bee visits were observed for P. ernstii at Oribi Gorge, at least one of the 

bees that pollinate P. oribiensis must visit P. ernstii as well. 







Potgieter 180 


OG, 7-4-98 


Amegilla fallax (Apidae) 


Thyreus vachali (Apidae) 


Xylocopa flavorufa (Apidae) 

NV 

Lycaenidae (Lepidoptera) 


Appendix/ 136 

P. oribiensis

10. Plectranthus fruticosus L’Hér. 


Plectranthus fruticosus is a variable soft suffrutex, growing 0.6 – 2 m tall. It branches 

freely, with branches ascending or rarely decumbent. The different varieties differ in 

terms of growth forms and/or floral shape, colour and size. The species occurs in forest 

and scrub forest, or in shaded places between rocks. It has the widest distribution of all 

the studied species, extending from the southern parts of the Western Cape, along the 

eastern seaboard through the Eastern Cape, KZN, Swaziland, Mpumalanga and the 

Northern Province. Flowering time extends from December to June. 


The branched paniculate synflorescences, 80 – 300 mm tall, are relatively dense, with 

large flowers, tapering towards the apex. Plants produce showy displays of bluish 

mauve flowers. Pink and pale blue forms also occur, but not in the study sites of this 

project. Corolla tube lengths vary from 5 – 13 mm, with the broad upper lip 2.5 – 8 mm 

and the boat-shaped lower lip 2 – 8 mm long. The latter deflexes as the flower ages; 

the upper lip reflexes. Purple blotches or speckles mark the upper lip. The corolla tube 

is deflexed and saccate at the base, extending upwards and backwards into a spur. 

The tube narrows slightly towards the mouth, giving a 2 mm vertical aperture. Four 

free, didynamous stamens extend 8 – 14 (upper pair) and 10 – 17 mm (lower pair) from 

the corolla mouth. These coil away as the flower ages and the style elongates to a 

similar length as the lower stamens. Some populations have reduced floral spurs, while 

others have spurs up to 5 mm long (Ongoye Forest). 


Populations were studied at Oribi Gorge, Leopard’s Bush in the Karkloof, Ferncliff in 

Pietermaritzburg, Ngeli Forest, and Magwa area. At least 18 hours were spent 

observing this species for pollinators, on various occasions over a number of years, 

between 9.00 am and 4.00 pm. At both the Magwa and Ngeli sites hybrid swarms 

involving P. fruticosus were observed. 

Pollinators 

Nemestrinid flies were the most commonly observed pollinators of P. fruticosus. These 

included the long-proboscid Stenobasipteron wiedemanni, and medium and shorterproboscid 

species of Prosoeca with proboscides from 7 – 11mm long. At Magwa 

Appendix/ 137 

P. fruticosus

acrocerid flies of the genus Psilodera (P. valida) visited a mixed stand of P. fruticosus, 

P. ciliatus and putative P. fruticosus x P. ciliatus hybrids. Medium-proboscid species of 

Prosoeca also visited these hybrids. The proboscis of S. wiedemanni exceeds the 

length of the P. fruticosus corolla tube, but some varieties have stamen and style 

lengths that approach the length of the S. wiedemanni proboscis and facilitate pollen 

deposition at the base of the proboscis. Medium-proboscid Prosoeca species carry 

pollen ventrally on the thorax and abdomen. 

No bee visits were seen to P. fruticosus, except for Allodape pernix (Apidae, Apinae) 

that collected pollen form the anthers on the exserted stamens. One visit of a pierid 

butterfly was seen at Oribi Gorge. 




Prosoeca sp. nov. 4 (Nemestrinidae) – near Potgieter 34 


Prosoeca circumdata (Nemestrinidae) 




Potgieter 223 

Episyrphus sp. (Syrphidae) 

MF, 16-4-02 





NV 

Pieridae (Lepidoptera) 

NV 

Appendix/ 138 

P. fruticosus

11. Plectranthus oertendahlii T.C.E.Fr. 


Plectranthus oertendahlii is a short semi-succulent herb with decumbent branches, 

reaching 200 mm in height. It grows in open forest understorey near rivers and is 

endemic to the coastal areas of Oribi Gorge and the Uvongo River near Port 

Shepstone. Flowering time extends from March to April. 


White flowers are borne on 70 – 200 mm tall synflorescences of relatively lax, simple or 

branched racemes. The white flowers are sometimes suffused with pale mauve, and 

nectar guides of four thin, purple, vertical lines occur on the upper lip. Straight, laterally 

compressed corolla tubes of medium length (8 – 13 mm) with saccate bases 4 mm 

wide, narrow gradually to 1.5 mm wide at the corolla mouth. The upper lip is 5 mm and 

the boat-shaped lower lip 4 – 5 mm long; the latter may deflex slightly from a horizontal 

position. Four free stamens extend slightly (2 – 3 mm) beyond the corolla mouth, with 

the style extending by 3 mm. 


Observations were conducted at Oribi Gorge only, since no populations were found at 

the Uvongo River. After initial attempts yielded no observations of floral visits at the 

scattered plants at the base of the gorge, a population was studied at dusk, but with no 

success. At least 20 hours of observations were made on various days in different 

years, between 9.00 am and 6.30 pm. Finally, in March 2004, pollinator visits were 

recorded to a large plant in a patch situated at a more elevated level along the gorge. 

Pollinators 

The acrocerid fly, Psilodera confusa, pollinates the species. Even on a rainy day the fly 

visited 30 – 40 flowers out of the ca. 60 on the plant, before moving on. A year later the 

same patch was observed and Psilodera confusa was once again seen pollinating the 

flowers. The proboscis length of this fly species (9.5 – 11 mm) matches the length of 

the corolla tube and sexual organs of P. oertendahlii closely. The floral tube is 

constricted at the mouth to the extent that only insects with suitably long, thin 

proboscides can probe for nectar. Pollen is deposited on the base of the proboscis and 

ventrally on the head and hairy thorax of the fly, between the bases of the legs. 

Appendix/ 139 

P. oertendahlii

This fly species also pollinates white P. ciliatus and pale blue forms of P. zuluensis at 

Oribi Gorge. It appears to seek out white or pale coloured flowers, since visits to a 

white Lobelia sp. were also seen in the same area. 



NV (but same species as Potgieter 105, 106, 134, 136, collected at OG) 



Appendix/ 140 

P. oertendahlii

12. Plectranthus praetermissus Codd 


Plectranthus praetermissus is a freely branched herb with decumbent stems growing 

20 – 50 cm tall. It occurs in open areas in coastal forest and is endemic to the Port St 

Johns area of the Eastern Cape (Transkei). It was studied along the Bulolwe River, not 

far from populations of another Pondoland endemic, P. reflexus. This species flowers 

from January to March. 


Racemose synflorescences are 120 – 200 mm tall, occasionally branching near the 

base. Flowers are violet to mauve with darker blotches as nectar guides on the inside 

of the upper and lower lips. Corolla tubes are straight and of medium length (12 – 15 

mm) with saccate base 4 mm deep, narrowing to 1.5 – 2 mm at the corolla mouth. The 

upper lip is 5 mm and the concave lower lip is 4 mm long, deflexing with age. The four 

free, didynamous stamens extend beyond the corolla mouth by 1.5 – 2 mm (upper pair) 

and 4 – 5 mm (lower pair). The style exerts from the corolla mouth by 4 mm. 


Two populations of P. praetermissus were studied in March 1998 at Port St Johns. 

Observations were made in the morning and the afternoon, for four hours in total. 

Pollinators 

The main pollinator was the acrocerid fly, Psilodera nhluzane, with a proboscis length 

(at this site) of 9 mm, which, together with the small head size of the fly, allows the fly 

to reach into the constricted floral tube for nectar. This facilitates pollen deposition 

ventrally on the head and thorax of the fly. The basic floral shape and size of P. 

praetermissus is very similar to that of P. oertendahlii and both species are pollinated 

by acrocerid flies. 

One visit of the nemestrinid fly, Stenobasipteron wiedemanni, was observed at a 

population near the other endemic, P. reflexus. This fly species may not pick up pollen 

on its body, since its proboscis is longer than the corolla tube, but the narrow opening 

of the corolla mouth and the short extension of the upper pair of anthers facilitated 

some pollen deposition on the fly proboscis. Other floral visitors included a species of 

Appendix/ 141 

P. praetermissus

syrphid fly, a hesperid butterfly and the apinid bee, Allodape pernix, which collected 

pollen. 


Psilodera nhluzane (Acroceridae) 







NV 


NV 

Appendix/ 142 

P. praetermissus

13. Plectranthus madagascariensis (Pers.) Benth. 

(and Plectranthus hadiensis (Forssk.) Schweinf. ex Sprenger.) 

Plectranthus madagascariensis and P. hadiensis will both be treated under P. 

madagascariensis, since these widespread, variable species are not always easy to 

distinguish in KZN (Codd 1985a) – the three varieties of each tend to inter-grade. The 

floral shape is functionally very similar and there is an overlap in pollinators. Most 

observations were made on P. madagascariensis, with a few P. hadiensis records 

included. 


Plectranthus madagascariensis is a relatively low-growing herb with erect to 

procumbent or decumbent stems, which may be semi-succulent. It is tolerant of a 

variety of habitats, growing along forest margins, in rocky grasslands or dry woodlands. 

The distribution of the varieties of P. madagascariensis and P. hadiensis overlap to a 

large extent, being widespread from the eastern part of the Western Cape, through the 

Eastern Cape, KZN, Swaziland, Mpumalanga, Northwest and the Northern Province, 

extending further into Mozambique and the Mascarenes (P. madagascariensis) and 

Tropical East Africa to Somalia and beyond (P. hadiensis). Flowering time extends 

from February to May, with sporadic flowering in July and November. 


Synflorescences are produced terminally on the main stem and on side branches, as 

simple racemes or with 1 – 2 pairs of branches at the base (90 – 250 mm in height). In 

P. hadiensis the synflorescences may be taller (up to 500 mm) with flowering verticils 

further apart. Flowers are white or shades of mauve or purple, often with red glands on 

the lips. Corolla length is relatively short, varying from 5 – 6 to 7 – 18 mm long; the 

variety studied at Oribi Gorge had corolla tubes 4 – 6 mm long. The tube expands 

gradually from the base and is bent (downwards) at about a third of the distance from 

the base, after which the sides are more or less parallel. The upper lip is relatively short 

and the concave lower lip may be longer than the corolla tube. The four stamens are 

free to the base and extend more or less as long as the lower lip. The variety at Oribi 

Gorge has didynamous stamens, extending 2 – 4 mm (upper pair) and 4 – 6 mm (lower 

pair) from the corolla mouth. 

Appendix/ 143 

P. madagascariensis


Observations on P. madagascariensis were made at Oribi Gorge, Umtamvuna and 

World’s View (near Ferncliff in Pietermaritzburg); a population of P. hadiensis was 

studied once at Ongoye Forest. A total of at least 20 hours was spent observing these 

species, on various days over a number of years, from 9.00 am to 4.00 pm. 

Pollinators 

The population at World’s View occurred in rocky grassland and was pollinated by 

medium-proboscid nemestrinid fly species (Prosoeca circumdata & P. umbrosa) and 

Psilodera valida (Acroceridae). The apinid bee, Amegilla aspergina, also visited 

flowers. These observations were made in April after the mass emergence of Prosoeca 

spp. Both the fly and the bee species have proboscides that reach nectar easily in the 

short corolla tubes, while still permitting pollen carryover on the ventral head and 

thoracic surfaces. 

More time was spent making observations at Oribi Gorge, which resulted in more floral 

visitors being recorded. Populations at this study site were most often in rocky areas 

near the river. The apinid bees Amegilla caelestina, A. bothai, A. mimadvena and A. 

aspergina, Xylocopa caffra and X. hottentotta were common visitors to P. 

madagascariensis; the proboscis lengths of this group are slightly longer than that of 

the floral tube, but as the bee settles on the boat-shaped lower lip while probing for 

nectar, the head rubs over the sexual organs contained in the lip. The bend in the 

corolla tube angles the mouth of the tube downwards, which forces large-bodied floral 

visitors to depress the lower lip that contains the stamens and style, as it angles 

upwards to access nectar which sits at a slightly higher level at the base of the tube. 

Medium- and short-proboscid nemestrinid flies of the genus Prosoeca were frequent 

visitors at this study site, with the medium-proboscid flies functioning in the same way 

as the apinid bees discussed earlier, while the short-proboscid species match the 

corolla tube lengths so as to pick up pollen ventrally on the thorax and abdomen. The 

tabanid fly, Philoliche aethiopica, visited occasionally; it had similar proboscis lengths 

to that of the apinid bees and medium-proboscid Prosoeca species. 

Less frequent visits by shorter-proboscid apinid bees at this site included Thyreus sp. 

and Apis mellifera, both of which probed for nectar, while a megachilinid bee species 

collected pollen ventrally onto the abdomen. The apinid bee, Allodape pernix, reached 

Appendix/ 144 

P. madagascariensis

nectar by crawling into the corolla tube, but also collected pollen from the anthers of P. 

madagascariensis. The halictinid bee Lasioglossum sp. accessed nectar legitimately by 

crawling into the corolla tube, as well as by robbing it from a hole pierced at the base of 

the corolla. Four different species of syrphid fly were also occasionally seen probing for 

nectar with their relatively short proboscides; these flies are also known to collect pollen. 

Small lycaenid butterflies were occasional visitors to flowers and larvae of this species 

were also found feeding on the inflorescences. 

At Umtamvuna similar bee species were visitors in open patches in the forest, as well 

as a number of lycaenid butterflies, a pierid butterfly species, a bombyliid fly with a 

short proboscis and a syrphid fly species. 

At Ongoye Forest one set of observations at midday yielded a number of visitors to a 

form of P. hadiensis with blue corollas that occurred in an open, rocky area beyond the 

forest margin. Two species of the nemestrinid genus Prosoeca were very active 

visitors; P. umbrosa specimens had medium proboscides and the other, unidentified, 

species had shorter proboscides. The apinid bee Amegilla bothai was also observed, as 

well as a small halictinid bee and a bombyliid fly. All these species were capable of 

pollinating P. hadiensis and were also found on P. madagascariensis in other localities. 

Plectranthus madagascariensis and P. hadiensis had the shortest corolla tubes of the 

ca. 20 species studied in this project and they showed the greatest number of visiting 

insects that were capable of facilitating pollination. 




Potgieter 30 WV, 11-4-96 (= Prosoeca sp. D, Potgieter et al. 1999) 



Potgieter 114 OG, 20-4-97 (= Prosoeca sp. C, Potgieter et al. 1999) 





Prosoeca sp. nov. 3 (Nemestrinidae) – near Potgieter 34 

Appendix/ 145 

P. madagascariensis



Potgieter 29 WV, 11-4-96 (= Prosoeca sp. B, Potgieter et al. 1999) 


Potgieter 166 ON, 22-4-99 (on P. hadiensis) 

Potgieter 167 ON, 22-4-99 (on P. hadiensis) 



Amegilla aspergina (Apidae) 




Potgieter 175 


OG, 3-4-98 





Potgieter 182 


OG, 25-4-98 


Xylocopa caffra (Apidae) 


Potgieter 184a OG, 25-4-98 

Potgieter 184b OG, 25-4-98 





Potgieter 137 

Thyreus vachali (Apidae) 

OG, 4-4-98 

Potgieter 118 


OG, 25-4-98 


Pseudoanthidium truncatum (Apidae, Megachilinae) 

Appendix/ 146 

P. madagascariensis


Chalicodoma sp. B (Apidae, Megachilinae) 



Potgieter 122 


OG, 25-4-98 






Allodape ceratinoides (Apidae) 



Potgieter 150 

Braunsapis (Apidae) 

OG, 3-4-98 


Lasioglossum sp. (Apidae, Halictinae) 



Potgieter 164 

Bombylius (Bombyliidae) 

ON, 22-4-99 



Potgieter 165 

Lycaenidae (Lepidoptera) 

ON, 22-4-99 (on P. hadiensis) 



Potgieter 208 


ON, 22-4-99 (on P. hadiensis) 




Appendix/ 147 

P. madagascariensis

14. Plectranthus petiolaris E.Mey. ex Benth. 


Plectranthus petiolaris is a sprawling branched herb with ascending and descending 

stems, reaching 1 m in height. It grows on forest margins and in scree below cliffs 

covered by scarp forest. It occurs in the Eastern Cape (Transkei) northwards along the 

coast through KZN and inland to Mpumalanga. It flowers from December to May. 


Flowers are produced in lax, racemose synflorescences (that occasionally branch at 

the base), 100 – 250 mm in height. At Umtamvuna the flowers were deep purple 

(violet), with the corolla lips often tinged with blue; at Oribi Gorge the flowers were pale 

pink. The corolla tube is laterally compressed, 7 – 11 mm long and sigmoid in shape; 

from the narrow base it ascends for 3 mm, then deflexes and expands to 3 mm wide at 

the corolla mouth. The upper lip is 6 – 8 mm and the shallowly boat-shaped lower lip is 

7 – 9 mm long. The four free stamens extend 4.5 mm (upper pair) and 5.5 mm (lower 



Pollinator observations were made at Umtamvuna and Oribi Gorge, where at least 18 

hours were spent with different populations on various days over a number of years, 

between 9.00 am and 5.30 pm. Some observations were also made on cultivated 

plants in Pietermaritzburg. 

Pollinators 

The main pollinators of P. petiolaris were apinid bees of the genera Amegilla and 

Xylocopa. At Umtamvuna Amegilla caelestina, A. mimadvena and Xylocopa hottentotta 

were common visitors. At Oribi Gorge Amegilla bothai and A. caelestina were 

commonly seen to visit the pink populations of P. petiolaris. At Pietermaritzburg 

Amegilla caelestina was a frequent floral visitor to cultivated plants. These bees picked 

up pollen ventrally on the thorax and abdomen. The flexible tip of the bee proboscis in 

most cases corresponds to the bend in the corolla of P. petiolaris, and the full length of 

the proboscis corresponds to the full tube length. 

At Oribi Gorge one visit by the long-proboscid fly, Stenobasipteron wiedemanni, was 

seen on P. petiolaris. The hovering fly bent its proboscis around the corolla bend to 

Appendix/ 148 

P. petiolaris

each nectar. It is unlikely that pollen was deposited on the body, but some may have 

rubbed off onto the proboscis; no voucher was collected. 

Two syrphid fly species were seen collecting pollen at Oribi Gorge and another syrphid 

species did the same at Umtamvuna. The apinid bee Allodape pernix was regularly 

seen robbing nectar at the base of the corolla of P. petiolaris at Umtamvuna, with about 

25% of corollas pierced in this way. These bees crawled onto the lower lip and around 

the outside to the base of the flower. The small halictinid bee Lasioglossum sp. crawled 

into the corolla tube to access nectar at Oribi Gorge. A few individuals of a pierid 

butterfly species were seen probing for nectar at Umtamvuna, but these made no 

contact with the style or anthers of the flowers. 







Potgieter 174 

A. mimadvena (Apidae) 

OG, 2-4-98 










Potgieter 82 

Amegilla caelestina 

NV 

OG, 10-1-97 

Stenobasipteron wiedemanni 

NV 

Asarkina sp. (Syrphidae) 


Appendix/ 149 

P. petiolaris


Potgieter 9 

Syrphidae 

UNR, 11-5-96 






Lasioglossum sp. (Apidae) 




Appendix/ 150 

P. petiolaris

15. Plectranthus laxiflorus Benth. 


Plectranthus laxiflorus is a soft suffrutex or freely-branched herb with spreading or 

ascending stems, growing 0.7 – 1.5 m tall. Populations occur on forest margins and 

shaded stream banks – seldom under forest canopy – and may form extensive stands 

that emit a characteristic citronella-like scent. The species is widely distributed, 

extending from the eastern part of the Western Cape, along the coast and midlands 

northwards through the Eastern Cape, KZN, Swaziland, Mpumalanga and the Northern 

Province, continuing into tropical Africa. This species flowers from mid-February to May 

and sporadically in October/November. 


Many synflorescences are produced, creating a striking show. These are lax racemes 

or panicles, 100 – 300 mm in height. The white flowers may be tinged with mauve, and 

4 – 5 vertical, thin purple lines are positioned on the upright upper lip. The corolla tube 

is sigmoid in shape (mean length 10.5 mm long), narrow at the base and ascending for 

2.5 mm, then bending downwards, expanding to 2.5 mm wide and laterally compressed 

at the corolla mouth. The upper lip is 6 – 7 mm and the boat-shaped lower lip 5 – 7 mm 

long. The four free, didynamous stamens extend 6 mm (upper pair) and 7.5 mm (lower 



Pollinator observations were done at a small population in Oribi Gorge and extensive 

populations at Ferncliff in Pietermaritzburg, Leopard’s Bush in the Karkloof, the Dargle 

Valley in the KZN Midlands, Ngeli Forest in southern KZN, and Kologha Forest near 

Stutterheim. At least 22 hours were spent observing at the various sites, over a number 

of days in different years, from 9.00 am to 4.00 pm. Additional observations were made 

at Long Tom Pass, Mpumalanga, where nectar collection was done for nectar studies. 

Pollinators 

Apinid bees of the genus Amegilla (A. bothai and A. mimadvena) were common 

pollinators throughout the flowering time at Ferncliff. However, a seasonal change in 

pollinators was noted late in each season (late March/early April) when mediumproboscid 

flies of the genus Prosoeca emerged in large numbers and joined the bees. 

These flies were also common at Kologha Forest in early March and Ngeli forest and 

Appendix/ 151 

P. laxiflorus

Dargle Valley in late March. On a overcast day at Dargle Valley, mostly nemestrinid 

flies (Prosoeca umbrosa) were seen to visit flowers during mid-morning, with only 

occasional bee visits later in the morning. 

At Kologha Forest the apinid bees Amegilla mimadvena and Xylocopa flavicollis were 

also abundant pollinators. At Leopard’s Bush the apinid bee Amegilla bothai was a 

common pollinator in early March, while the megachilinid bee Chalicodoma sp. A 

collected pollen in abdominal scopae. At Oribi Gorge the apinid bee Amegilla 

caelestina was a frequent pollinator in early April. 

The apinid bees have proboscides that match the length of the corolla tube, with flexible 

tips that allow the bees to reach nectar beyond the bend near the base of the tube. 

Likewise, the flexible proboscides of the medium-proboscid nemestrinid flies also match 

the corolla tube length. These bees and flies pick up pollen ventrally on their thorax, 

abdomen and head. 

Other floral visitors included Apis mellifera and two syrphid fly species that collected 

pollen at Ferncliff; two syrphid species collected pollen at Ngeli and one at Leopard’s 

Bush. A few lepidopterans visited flowers: a hesperid butterfly at Leopard’s Bush, a 

pierid butterfly and papilionoid butterfly (Papilio nireus lyaens) at Kologha Forest and 

the diurnal sphingid moth species, Macroglossus trochilus, at the latter site and at 

Ferncliff. 












Appendix/ 152 

P. laxiflorus


Potgieter 181 


FC, 4-3-98 

Potgieter 173 

Xylocopa flavicollis (Apidae) 

OG, 2-4-98 

Potgieter 202 

Chalicodoma sp. A (Apidae) 

KF, 29-2-00 


Potgieter 163 


KK, 3-3-99 












Prosoeca sp. nov. 5 (Nemestrinidae) 








Potgieter 144 NG, 20-3-98 

Potgieter 145 NG, 20-3-98 



Potgieter 220 NG, 03-01 

NV DV, 03-09 

Appendix/ 153 

P. laxiflorus

Bombyliidae (Diptera) 


Asarkina sp. A (Syrphidae) 

Potgieter 140 

Asarkina sp. B (Syrphidae) 

NG, 3-5-98 

Potgieter 160 

Syrphidae (Diptera) 

KK, 3-3-99 

Potgieter 139 


NG, 3-5-98 


Potgieter 101 


NV 

KK, 3-3-99 

Papilio nireus lyaens (Papilionidae) 




Appendix/ 154 

P. laxiflorus

16. Plectranthus calycinus Benth. [= Rabdosiella calycina (Benth.) Codd] 

This species was placed in a new genus, Rabdosiella, by Codd (1984), which meant 

that it was known as Rabdosiella calycina (Benth.) Codd until 1993 when it was placed 

back into Plectranthus (Ryding, 1993). The phylogeny by Paton et al. (2004) shows it to 

cluster with other species of Plectranthus, Pycnostachys and Holostylon within the 

Coleus clade of the Plectranthinae. 


Plectranthus calycinus is a tall, erect, rigid sub-shrub with branches arising annually 

from a perennial woody rootstock, growing 0.6 – 1.5 m tall. It occurs in grasslands from 

the Eastern Cape northwards to KZN, the eastern Free State, Swaziland, Mpumalanga 

and the Northern Province. It flowers from January to May. 


The dense synflorescences are terminal panicles of scorpioid cymes, 100 – 300 mm in 

height. The flowers are creamy-white with mauve edging on the upper and lower 

corolla lips. The short compressed corolla tube (6.5 mm long) is saccate at the base (4 

mm wide), narrowing to 3 mm wide at the mouth. The erect upper lip is 2 mm and the 

spreading, shallowly boat-shaped lower lip 4 – 5 mm long. The four stamens are 

declinate with free filaments, enclosed in the lower lip, extending 4 mm (upper pair) and 

5 mm (lower pair) from the corolla mouth. 


Pollinator observations were made at a grassland site in the Dargle area (KZN 

Midlands), as well as on a grassland plateau at Umtamvuna. A total of ten hours was 

spent making observations at the three sites. 

Pollinators 

The Dargle population was pollinated exclusively by a medium-proboscid nemestrinid 

fly, Prosoeca umbrosa, with pollen loads deposited ventrally on the head, thorax and 

abdomen. At Umtamvuna the population was visited by the apinid bee Xylocopa 

scioensis, but no voucher was collected. This bee species has a short proboscis length 

(6.5 mm) that fits the corolla tube length well. 

Appendix/ 155 

P. calycinus



Potgieter 218 DV, 1-5-00 

Potgieter 219 DV, 1-5-00 

Xylocopa scioensis (Apidae) 

NV UNR, 05-00 

Appendix/ 156 

P. calycinus

17. Plectranthus rehmannii Gürke 


Plectranthus rehmannii is an erect subshrub or branched herb with ascending stems, 

growing 0.6 – 1.2 m tall, in or near forest margins. It is endemic to the KZN Midlands 

and may be locally abundant. It flowers from January to April. 


Synflorescences are paniculate, 250 – 350 mm in height, bearing small, creamy-white 

flowers that are densely tomentose. The deflexed, laterally compressed, short corolla 

tube (5 mm long) is saccate at the base and narrows slightly towards the mouth. The 

upper lip is very short (2 mm) and the boat-shaped lower lip is 4 mm long, curving 

upwards. The four free stamens extend about 2.5 mm from the corolla mouth. 


Pollinator observations were made at Leopard’s Bush in the Karkloof in 1999, where at 

least four hours were spent with the population, as well as at a forested site in the 

Dargle, in 2009. At both sites P. rehmannii grows alongside P. laxiflorus. 

Pollinators 

The megachilinid bee Chalicodoma sp. A visited both P. rehmannii and P. laxiflorus in 

the Karkloof. This bee has a short proboscis (ca. 6 mm) which could not reach nectar in 

P. laxiflorus (it only collected pollen), but it accessed nectar from the short corolla tubes 

of P. rehmannii. This was the main pollinator of P. rehmannii at the Karkloof site, and 

actively collected pollen in its abdominal scopae. 

At the Dargle site there were no megachilinid bees on an overcast day, when general 

bee activity was limited. Interestingly, none of the Prosoeca umbrosa flies that visited 

flowers of adjacent P. laxiflorus plants, were seen on P. rehmannii. 

At the Karkloof site Apis mellifera was another, less frequent, floral visitor that 

accessed nectar and collected pollen, while many Allodape pernix bees visited flowers 

to collect pollen. 

Appendix/ 157 

P. rehmannii


Chalicodoma sp. A (Apidae) 


Potgieter 163 


KK, 3-3-99 




Appendix/ 158 

P. rehmannii

18. Plectranthus spicatus E.Mey. ex Benth. 


Plectranthus spicatus is a succulent herb producing several annual stems from a 

perennial rootstock. The stems are decumbent, reaching 0.6 m in height. It grows in dry 

woodland or rocky grassland (where it was studied), extending coastally from the 

Eastern Cape to KZN, and inland to Swaziland and Mpumalanga. The flowering time 

extends from March to July. 


Purple flowers are produced in compact, simple, subspicate synflorescences 

(occasionally with a pair of branches near the base), 90 – 300 mm in height. Corolla 

tubes are sigmoid in shape, short (5 mm) and narrow at the base, ascending at first, 

then curving sharply downwards, expanding towards the mouth. The upper lip is 2.5 

mm and the boat-shaped lower lip 2.5 – 3 mm long. Four free stamens extend 2.5 – 3 

mm from the corolla mouth. 


This species was studied at Oribi Gorge, in a dry, steep, rocky grassland, where four 

hours were spent on one occasion on a hot afternoon. 

Pollinators 

The main pollinator (about 80% of nectar-feeding floral visitors) was the apinid bee 

Xylocopa caffra, with these large-bodied bees picking up pollen ventrally on the thorax. 

The relatively short proboscis (6.5 mm) fitted the short tube of the flower, and the 

flexible tip to the proboscis allowed it access to nectar at the base of the sigmoid tube. 

About 20% of nectar-feeding floral visits were by the apinid bee Amegilla mimadvena. 

Allodape pernix was an infrequent apinid bee visitor, collecting pollen from the anthers 

in the lower lip. 


Xylocopa caffra (Apidae) 


Potgieter 183A OG, 25-4-98 

Amegilla mimadvena (Apidae) NV 

Allodape pernix (Apidae) NV 

Appendix/ 159 

P. spicatus

19. Pycnostachys urticifolia Hook. 


Pycnostachys urticifolia is an erect soft shrub or herb that branches from a woody 

base, growing 1 – 2.5 m in height. It occurs in moist areas such as grassy stream 

banks and forest margins and is widely spread in South Africa, Zimbabwe, 

Mozambique, Malawi and Tanzania. In South Africa it was originally restricted to the 

Northern Province and Mpumalanga, but records from Bews Herbarium (NU) show it to 

have spread into KZN from the 1970’s. It is now common in roadside vegetation in 

subtropical areas. It has flowers of similar dimensions to Pycnostachys reticulata 

(E.Mey.) Benth. which occurs naturally in KZN, hence similar pollinators should be 

found on both. It flowers from April to June. 


Synflorescences are dense, terminal, spicate panicles of helicoid cymes on the ends of 

upright branches, with the largest, central one 70 – 100 mm in height. Unlike those in 

Plectranthus, the sepal lobes of the calyces are elongate, rigid and spinescent, 

protecting the basal nectaries. Flowers are deep, bright blue (occasionally whitish-blue) 

with no visible nectar guides. The deflexed corolla tube is sigmoid in shape and of 

medium length (11 mm), with a narrow base. It ascends for 8 mm, then bends 

downwards, enlarging near the corolla mouth. The upper lip is short (3 mm) and the 

boat-shaped lower lip is 9 mm long. The four didynamous stamens are fused for a few 

mm outside the corolla mouth, providing mechanical support for the style that is 

enclosed in this rigid sheath. The upper pair extends 4 – 6 mm and the lower pair 6 – 

7.5 mm from the corolla mouth; stamens are housed in the lower lip and are separated 

laterally by an apical fold of the lower lip. The style extends slightly further than the 

stamens. 


Pollination observations were conducted on cultivated material in Pietermaritzburg, for 

five hours on a sunny day, from late morning onwards. 

Pollinators 

The main pollinators of Py. urticifolia were nectar-seeking apinid bees of the genera 

Amegilla (A. mimadvena) and Xylocopa (X. caffra and X. scioensis), as well as pollen- 

Appendix/ 160 

Py. urticifolia

collecting megachilinid bees of the genera Chalicodoma (Chalicodoma sp. B) and 

Megachile (Megachile spp. A & B). 

The apinid bees probed corollas for nectar and the flexible tips of their proboscides 

allowed them to negotiate the sigmoid corolla bend; the longer proboscis of A. 

mimadvena allowed pollen deposition ventrally on the bee thorax and abdomen, while 

the shorter proboscis of Xylocopa resulted in pollen placement on the base of the head 

as well. The downwards angle of the distal limb of the corolla tube forced the bees to 

probe upwards, and ensured body contact with the anthers and stigma in the lower lip. 

The megachilinid bees had large ventral pollen loads. These bees have abdominal 

scopae into which pollen is gathered, which places it in a perfect position for stigma 

deposition. 

Other floral visitors were the apinid bee Apis mellifera, collecting pollen, a lycaenid 

butterfly, a megachilinid bee, Pseudoanthidium truncatum, on which no pollen was 

found, and the apinid bee Thyreus sp., on which no pollen was found. 



Xylocopa scioensis (Apidae) 

NV 

Potgieter 128 

Xylocopa flavorufa (Apidae) 

PMB Gdn, 28-4-98 

Potgieter 129 

Chalicodoma sp. B (Apidae) 

PMB Gdn, 28-4-98 

Potgieter 126 

Megachile sp. A (Apidae) 

PMB Gdn, 28-4-98 

Potgieter 123 

Megachile sp. B (Apidae) 

PMB Gdn, 28-4-98 

Potgieter 127 


PMB Gdn, 28-4-98 


Pseudoanthidium truncatum (Apidae) 


Thyreus sp. (Apidae) NV 

Lycaenidae (Lepidoptera) NV 

Appendix/ 161 

Py. urticifolia

20. Aeollanthus parvifolius Benth. 


Aeollanthus parvifolius is a semi-succulent herb or sub-shrub with ascending branches 

that spread out from a woody base, growing 0.2 – 0.5 m tall. It usually occurs among 

rocks, coastally in the Eastern Cape and KZN, to Swaziland and at higher altitudes in 

Mpumalanga, North West and the Northern Province. It flowers from December to 

June. 


Relatively lax panicoid synflorescences that are often branched, reach 50 – 200 mm in 

height. Flowers are white, sometimes with a pinkish tinge, and the upper lip has nectar 

guides in the form of reddish purple spots. The corolla tube is 7 – 10 mm long, curving 

downwards from a cylindrical, narrow base, expanding towards the mouth. The upper 

lip is 5 mm and the concave lower lip 8 mm long. The four didynamous stamens are 

attached at the mouth of the corolla, with free filaments lying inside and exserted 

slightly beyond the lower lip, bearing one-celled anthers. 


Pollinator observations were made at Umtamvuna and Ongoye Forest, with about 

equal amount of time spent at each locality – a total of six hours from early to late 

afternoons on sunny days. 

Pollinators 

At Umtamvuna Apis mellifera and a few syrphid flies were seen collecting pollen, but 

the apinid bee, Amegilla fallax, visited flowers for nectar and appeared to be a 

pollinator. At Ongoye Forest three apinid bee species (Amegilla mimadvena, A. bothai 

and A. fallax) were pollinators. All these bees could reach nectar at the base of the 

corolla and picked up pollen passively on their ventral abdominal surfaces. An 

acrocerid fly, Psilodera sp., visited flowers in a similar way as the apinid bees, with its 

black and white striped abdomen looking superficially like that of A. fallax (subgenus 

Zebramegilla). 

Appendix/ 162 

A. parvifolius



NV 


NV 



Potgieter 169 

Psilodera sp. (Acroceridae) 

ON, 22-4-98 

NV ON, 22-4-98 

Appendix/ 163 

A. parvifolius

References 

Brothers, D.J., 1999. Phylogeny and evolution of wasps, ants and bees (Hymenoptera, 

Chrysidoidea, Vespoidea and Apoidea). Zoologica Scripta 28: 233–249. 

Codd, L.E., 1957. Plectranthus dregei. Flowering Plants of Africa 32: Plate 1244. 

Codd, L.E., 1970. Plectranthus saccatus. Flowering Plants of Africa 43: Plate 1601. 

Codd, L.E., 1975. Plectranthus (Labiatae) and allied genera in southern Africa. Bothalia 

11: 371–442. 

Codd, L.E., 1977. Plectranthus oertendahlii. Flowering Plants of Africa 44: Plate 1729. 

Codd, L.E., 1979. Plectranthus praetermissus. Flowering Plants of Africa 45: Plate 

1791. 

Codd, L.E., 1980. Plectranthus oribiensis. Flowering Plants of Africa 46: Plate 1809. 

Codd, L.E., 1982a. Plectranthus ecklonii. Flowering Plants of Africa 47: Plate 1854. 

Codd, L.E., 1982b. Plectranthus ernstii. Flowering Plants of Africa 47: Plate 1855. 

Codd, L.E., 1985a. Plectranthus (Lamiaceae). Flora of Southern Africa 28: 137–172. 

Codd, L.E., 1985b. Plectranthus hilliardiae. Flowering Plants of Africa 48: Plate 1904. 

Codd, L.E., 1985c. Plectranthus petiolaris. Flowering Plants of Africa 48: Plate 1905. 

Codd, L.E., 1994. The genus Isodon (Schrad. ex Benth.) Spach in Africa and a new 

genus Rabdosiella Codd (Lamiaceae). Bothalia 15: 7–10. 

Dyer, R.A., 1934. Pycnostachys urticifolia. Flowering Plants of Africa 14: Plate 560. 

Dyer, R.A., Bruce, E.A., 1951a. Plectranthus fruticosus. Flowering Plants of Africa 28: 

Plate 1101. 

Dyer, R.A., Bruce, E.A., 1951b. Plectranthus zuluensis. Flowering Plants of Africa 28: 

Plate 1110. 

Lewis, G.J., 1951. Plectranthus behrii. Flowering Plants of Africa 28: Plate 1109. 




Potgieter, C.J., Edwards, T.J., Miller, R.M., Van Staden, J., 1999. Pollination of seven 

Plectranthus spp. (Lamiaceae) in southern Natal, South Africa. Plant Systematics 

and Evolution 218: 99–112. 

Appendix/ 164



Botany 75: 646-659. 

Ryding, O., 1993. A reconsideration of the genus Rabdosiella (Lamiaceae, Nepetoidea, 

Ocimeae). Plant Systematics and Evolution 185: 91–97. 

Van Jaarsveld E.J., Edwards T.J., 1991. Plectranthus reflexus. Flowering Plants of 

Africa 51: Plate 2034. 

Van Jaarsveld E.J., Edwards T.J., 1997. Notes on Plectranthus (Lamiaceae) from 

southern Africa. Bothalia 27: 1–6. 

Van Jaarsveld, E.J., Van Wyk, A.E., 2001. Plectranthus hilliardiae subsp. australis, a 

new taxon from Eastern Cape, South Africa. Bothalia 31: 44–45. 

Verdoorn, I.C., 1949. Plectranthus ciliatus. Flowering Plants of Africa 27: Plate 1051. 

Appendix/ 165

View/Open - ResearchSpace - University of KwaZulu-Natal

You also want an ePaper? Increase the reach of your titles

Delete template?

Save as template?