Evolution of Host-Plant use

in Cymothoe (Nymphalidae) Feeding on

Rinorea (Violaceae)

Robin van Velzen (791105 866 070)

Supervisors: Dr. Freek Bakker

Biosystematics Group Dr. Joop van Loon Laboratory of Entomology

November 2006 Wageningen University

2

3

TITLEPAGE.

1. Harma theobene male; 2. Cymothoe beckeri female;

3. Cymothoe indamora female; 4. Cymothoe herminia female RVVB226;

5. Cymothoe caenis male RVV B364; 6. Cymothoe coccinata male RVV B353;

7. Cymothoe oemilius female. 8. Rinorea longisepala young fruit;

9. Rinorea dentata young fruit; 10. Rinorea caudata fruit RVV49;

11. Rinorea subintegrifolia capsule; 12. Cymothoe fumana caterpillar RVV C009

on Rinorea oblongifolia leaf.

4

Preface

This is the report of the second Msc thesis for my biology studies at the Wageningen University, supervised by Freek Bakker from the Nationaal Herbarium Nederland – Wageningen Branch & Biosystematics Group and Joop van Loon from the Laboratory of Entomology. This thesis was meant to be a preparative and orientating pilot for the planned PhD project “Evolution of host-specificity in the tropical African butterfly genus Cymothoe (Nymphalidae) feeding on Rinorea (Violaceae)” for which an application has been submitted to the Nederlandse Organisatie voor Wetenschappelijk onderzoek, Aard- en Levenswetenschappen (NWO-ALW).

The pilot study aimed at evaluating the feasibility of studying Cymothoe butterflies and their associated host plants in the field and in the laboratory, and my work consisted consequently of field work as well as analytical work. The field work was carried out in the tropical forests of Cameroon, in collaboration with the Herbier National in Yaoundé. The analyticalwork could be carried out thanks to data that was kindly provided by laboratories in Guelph (Canada), Wageningen (Netherlands) and Davis (USA).

This report is divided into two chapters, with different objectives. The first chapter is the scientific thesis report, where an introduction to the context of the subject is given and the analytical questions will be adressed. The second chapter is the field report and can be found at page 45. It deals with the work that has been done in Cameroon and gives details about the expedition. The results, in terms of collections and observations are discussed.

I thank the reader for the interest and hope that this report will give a satisfactory overview of my work on this subject and a boost to the study of this highly fascinating subject.

Robin van Velzen.

Wageningen, November 2006

5

CHAPTER 1

Thesis Report

6

Index Abstract............................................................................................................................ 7

1 Introduction ........................................................................................................ 8 1.1 Host plant use in phytophagous insects...................................................................... 8 Phytochemistry............................................................................................................................. 8 Evolution of host plant use............................................................................................................. 8

1.2 Host specificity ........................................................................................................ 9 Evolution of host specificity ......................................................................................................... 10 Causes of host specificity............................................................................................................ 10

1.3 The case of Cymothoe (Nymphalidae) feeding on Rinorea (Violaceae)...........................11 Rinorea ...................................................................................................................................... 11 Cymothoe and Harma ................................................................................................................. 12 Host plant relationships............................................................................................................... 14 Host plant chemistry ................................................................................................................... 15

1.4 DNA barcoding........................................................................................................16 1.5 Research questions .................................................................................................18 Phylogeny estimates ................................................................................................................... 18 Host plant associations ............................................................................................................... 19 DNA barcoding ........................................................................................................................... 19

2 Materials & Methods ......................................................................................... 20 2.1 Taxon sampling.......................................................................................................20 2.2 DNA extraction, amplification and sequencing............................................................22 2.3 DNA barcoding........................................................................................................23 2.4 Phylogeny reconstruction.........................................................................................23

3 Results............................................................................................................. 24 3.1 DNA barcoding........................................................................................................24 3.2 Phylogeny reconstruction.........................................................................................25 Cymothoe and Harma ................................................................................................................. 25 The C. egesta complex ............................................................................................................... 27 Rinorea ...................................................................................................................................... 28

4 Discussion........................................................................................................ 30 4.1 DNA barcoding........................................................................................................30 4.2 Phylogenetic estimation ...........................................................................................30 Rinorea ...................................................................................................................................... 30 Cymothoe and Harma ................................................................................................................. 31

4.3 Cryptic sibling species in Cymothoe..........................................................................32 4.4 Synthesis................................................................................................................34

5 Conclusions...................................................................................................... 37

6 Recommendations ............................................................................................ 37

7 Acknowledgements ........................................................................................... 38

8 References....................................................................................................... 39

7

Abstract

Together, living plants and the herbivorous insects that eat them make up half of all extant organisms on earth. Here, I focus on two aspects of host plant association patterns. First, the majority herbivorous insect groups do not feed indiscriminative on any plant, but are usually associated with plants that are somehow related, an observation that has led authors to hypothesise a process of coevolution where plants and herbivorous insects have influenced each others evolution through reciprocal selection pressures, mediated through plant secondary compounds and the butterflies’ adaptations to overcome this anti-herbivore phytochemistry. However, phylogenetic evidence for chemical coevolution is limited and evidence must be sought for at other levels of resolution than most previous studies have achieved. Second, most phytophagous insects are highly host specific, with many species being monophagous. But the proximate and ultimate factors underlying this specificity remain largely unclear and phylogenetic studies of highly specific insect-plant associations are very few.

Here, I present a study on the evolution of host plant use and host specificity in the highly specific interaction between Cymothoe butterflies and their Rinorea host plants in the Afrotropics; the first detailed study of a system where the specificity is at the species level and involves many congeneric species of both butterflies and plants and the first such published study in the tropics. The research questions consider the scale and extent of this herbivore plant interaction, the phylogeny of both butterflies and host plants, and the mechanisms involved in host plant recognition, but given the limitations for an Msc thesis, a few number of questions have been selected as a basis of a pilot study aiming at evaluating the feasibility of studying Cymothoe butterflies and their associated host plants in the field and in the laboratory. This chapter evaluates the preliminary phylogenetic patterns in the light of host plant association patterns as well as the use of DNA barcodes as a tool in species identification.

The phylogenetic tree of Rinorea based on nrDNA ITS sequences is largely concordant with the taxonomic classification by Achoundong and the overall pattern is congruent with the previous Rinorea phylogeny estimation based on cpDNA trnL-F sequences. The phylogeny estimation of Cymothoe based on mtDNA cox1 sequences shows that Harma can be maintained as a monotypic genus and sister to Cymothoe, and that within Cymothoe, Rinorea-colonising species have probably evolved from Achariaceae-feeding ones. In at least three species, intraspecific DNA sequence variation indicates that they may contain cryptic sibling species, two of which were previously predicted based on morphological studies. There seems to be an overall pattern of phylogenetic conservation in the phytophagous relationship between Cymothoe and Rinorea, where related butterflies feed on related hosts. There is also an indication that the phylogenetic order of divergence among the Rinorea host plants corresponds to that among their associated Cymothoe herbivores. But there are also a few cases where the occurrence of host jumping is most likely. DNA barcoding is a very powerful tool for the study of Cymothoe species. It can be successfully used to reliably identify eggs and caterpillars. Additionally it can also be of value in disclosing hidden diversity in cryptic sibling species. The associations between Rinorea and Cymothoe are likely to be explained by a combination of historical and adaptive factors which makes it an ideal subject for a study of the mechanisms and relative importance of these factors in the evolution of host plant use and host specificity.

8

1 Introduction

It is widely known that plants constitute the largest part of Earths biomass and play an important role in shaping the evolutionary arena for other biota. Insects are the most diverse group of organisms on land, with an estimated 4 to 7 million species (1996; Janz and Nylin, 1998; Novotny et al., 2002), a diversification that has been associated with the adoption of phytophagy (Mitter, Farrell, and Wiegmann, 1988). Together, living plants and the herbivorous insects that eat them make up half of all extant species on earth (Coley, 1999; Schoonhoven, van Loon, and Dicke, 2005).

1.1 Host plant use in phytophagous insects.

An important observation is that there is large scale conservation in the patterns of association between insects and their host plants. The majority of insects groups do not feed indiscriminative on any plant, but are usually associated with plants that are somehow related (Ehrlich and Raven, 1964; Futuyma and Moreno, 1988; Janz and Nylin, 1998; Novotny et al., 2006).

Phytochemistry

Of all insect plant interactions, the most well-known cases are those of butterflies and their host plants. This is mainly due to the seminal paper by Ehrlich and Raven (1964), in which they explain several butterfly host-plant relationships by phytochemical similarities in host plants. Plants exhibit a daunting variety of secondary compounds that play a role in resistance against herbivory but may also act as chemical cues for host-plant recognition for feeding and oviposition in insects (Fraenkel, 1959; Ehrlich and Raven, 1964; Schoonhoven, van Loon, and Dicke, 2005).

Evolution of host plant use

The tendency of groups of related butterflies to colonise plants that are chemically similar (Jaenike, 1990) has led Ehrlich and Raven (1964) to hypothesise that butterflies and their dicotyledonous host plants have diversified by a process of chemical coevolution. Coevolution can be defined as ‘an evolutionary change in a trait of the individuals of one population in response to a trait of the individuals of a second population, followed by an evolutionary response by the second population to the change in the first’ (Janzen, 1980). Ehrlich and Raven hypothesised that the butterfly host-plant relationships are the result of a coevolutionary process that acts through plant secondary compounds and the butterflies’ adaptations to overcome this anti-herbivore phytochemistry (Fraenkel, 1959; Ehrlich and Raven, 1964). Others proposed that chemical coevolution was not involving species pairs, but signified rather a diffuse process, where the evolution of plant lineages occurs in response to suites of herbivore species and vice versa (Janzen, 1980; Thompson, 1989; Farrell and Mitter, 1998; Cornell and Hawkins, 2003).

An important preposition for coevolution is the reciprocal selection pressure of plants on their herbivores and vice versa. It is generally accepted that plants exert selective pressure on their herbivores, but the opposite case is still controversial. Many authors consider selection on plants by herbivorous insects to be too weak and variable to drive plant evolution (Jermy, 1984; Farrell and Mitter, 1990; Cornell and Hawkins, 2003). But conversely other authors state that insect herbivores are actually able to shift the competitive balance among plant species (Marquis, 2004) and drive the evolution of some plant defences (Simms and Rausher, 1989; Thompson and Pellmyr, 1991), even

9

when the amount of damage is relatively low (Crawley, 1985). Another complicating factor is that selection is likely to vary spatially, creating a geographic mosaic of coevolution (Thompson and Cunningham, 2002; Schoonhoven, van Loon, and Dicke, 2005; Thompson, 2005).

If insect and plant lineages have indeed diversified in association, it is expected that the phylogenetic order of divergence among host plants should correspond in some way to that among their associated herbivores (Farrell and Mitter, 1998). Additionally, the timing of the divergences should be contemporary on a geographical scale (Jermy, 1984; Becerra, 2003; Percy, Page, and Cronk, 2004; Lopez-Vaamonde et al., 2006).

By producing dated phylogenies, evidence for phytochemical coevolution for insect herbivores and their host plants has been found in beetles (Coleoptera; Farrell and Mitter, 1990; Farrell and Mitter, 1998; Wilf et al., 2000; Becerra, 2003) and a recent diversification of Psyllids (Hemiptera) on the Canary Islands (Percy, Page, and Cronk, 2004). Examination of the responses of herbivores to plant secondary compounds also complies to the predictions made by chemical coevolutionary theory. Studies of the diversification of the main groups of Lepidoptera, however, revealed that the host plant clades are much older than their associated butterflies; apparently butterflies colonised already diversified plants (Janz and Nylin, 1998; Braby and Trueman, 2006; Lopez-Vaamonde et al., 2006). In fact, most insect groups seem to be evolving in such a “sequential” way; the insects following the evolution of plants (Jermy, 1984; Menken, Herrebout, and Wiebes, 1992; Bernays, 2001; Percy, Page, and Cronk, 2004). At a more detailed level, host plant associations of insect species seem to be even more flexible (Farrell and Mitter, 1998; Janz and Nylin, 1998), with colonisation of distantly related plants (Farrell and Mitter, 1990; Pratt, 1994; Braby and Trueman, 2006), unstable host use (Bernays and Graham, 1988) and intraspecific variation (Fox and Morrow, 1981; Radtkey and Singer, 1995). Overall, the relative contributions of the two factors phylogenetic history and adaptation to host plant use by insects remain largely unclear (Moran, 1988). Although direct support for chemical coevolution in herbivorous insects is limited, it may still be an important process (Bernays, 2001). Evidence must be sought for at other levels of resolution than most previous studies have achieved (Janz and Nylin, 1998), because above the species level coevolutionary patterns may become obscured by subsequent processes such as host jumping.

1.2 Host specificity

Most phytophagous insects, approximately 80%, are highly host specific, (Bernays and Graham, 1988; Jaenike, 1990; Schoonhoven, van Loon, and Dicke, 2005). For example, more than half of all Lepidoptera species feed on plants of only one genus (Bernays and Chapman, 1994). Questions about the evolution and causes of host specificity are therefore very important in the field of insect plant relationships. The most commonly recognised terms used for designating diet breadth are monophagous, oligophagous and polyphagous (Bernays and Chapman, 1994; Schoonhoven, van Loon, and Dicke, 2005), but different definitions of these terms appear in the literature (Symons and Beccaloni, 1999), with monophagous signifying feeding exclusively on one species, species of one genus, or even species of one family. Simply counting the number of host taxa at one particular taxonomic level can also lead to contradictory results, because higher taxa are not necessarily equivalent unless they are sisters (Symons and Beccaloni, 1999). The current view is that analyses of host specificity need to consider the phylogenetic relationships of the hosts (Novotny and Basset, 2005), resulting in new specificity indices such as the Non Specificity Index (NSI; Desdevises, Morand, and Legendre, 2002), the Root Phylogenetic Diversity index (Root PD) or the Clade Dispersion index (CD; Symons and Beccaloni, 1999).

10

Evolution of host specificity

The creation of phylogenetic trees for the herbivore allows inferences about whether a particular ancestral herbivore was generalised or specialised (Coley, 1999). Studies mapping host specificity on herbivore phylogenies indicate that specialism is derived (Kelley and Farrell, 1998; Nosil, 2002; Janz, Nylin, and Wahlberg, 2006). However, specialisation does not necessarily represent an evolutionary dead-end (Nosil, 2002) and some highly specialized plant-herbivore interactions can indeed be very old (Becerra, 2003).

In their recent study of the evolution of host specificity in Nymphalidae butterflies, Janz and co workers (2006) have found recurring oscillations between host expansions and specialisation. They hypothesise that these oscillations are an important driving force behind the diversification of plant feeding insects, with host expansions allowing the species to increase its geographical distribution and thereby setting the stage for subsequent population fragmentation by secondary specialisation on different hosts in the expanded repertoire (Janz, Nylin, and Wahlberg, 2006).

Causes of host specificity

Causes for host specificity can be divided into proximate and ultimate factors, the first providing mechanistic explanations for host specificity and the second focussing on ecological factors.

a. Proximate factors The proximate causes for host specificity will be discussed for butterflies, because of its relevance for the case studied here (see below) and the fact that there is much information specific to Lepidoptera. Although the mechanism behind exploitation of a particular plant as a source of food by butterflies depends on metabolic adjustments (Ehrlich and Raven, 1964; Weingartner, Wahlberg, and Nylin, 2006) and host plant recognition by the larvae (Wiklund, 1975), the initial distribution of young larvae on food plants is determined by the oviposition behaviour of the adult female (Singer, 1971). Female preference should therefore match larval performance (Wiklund, 1975; Singer, Ng, and Thomas, 1988). The precise behavioural sequence used by females when choosing plants for oviposition varies among species (Thompson and Pellmyr, 1991) but in general the behavioural sequence leading to oviposition by a butterfly follows a sequence of 1. searching, orientation, encounter and 2. landing, surface evaluation, and acceptance (Renwick and Chew, 1994). The predominant sensory cues at the searching phase in host finding are visual (Rausher, 1978; Prokopy and Owens, 1983; Kelber, 1999). Landing may be triggered by visual and/or chemical cues, (Renwick and Chew, 1994; Schoonhoven, van Loon, and Dicke, 2005). But Wiklund (1977) found that Leptidea females land on plant species in proportion to the plants’ density, indicating that females land on plants at random and that plant recognition occurs only after the female has landed (Thompson and Pellmyr, 1991). After an insect lands on a plant, receptors present on the tarsi, antennae, proboscis and ovipositor (Calvert, 1974; Qiu, van Loon, and Roessingh, 1998) are involved in contact perception of both physical and chemical characteristics of the leaf, to determine the suitability for oviposition (Renwick and Chew, 1994; Schoonhoven, van Loon, and Dicke, 2005). All species of day-active butterflies probably show drumming behaviour at this stage, where the forelegs move rapidly so that the terminal tarsi drum against the leaf surface (Renwick and Chew, 1994).

Overall, non-volatile plant chemistry is probably the most important source of information contributing to the final decision by a butterfly to oviposit or not (Fraenkel, 1959; Ehrlich and Raven, 1964; Jermy, 1984; Renwick and Chew, 1994). Specific chemicals can act as stimulants (Carter, Feeny, and Haribal, 1999; van Loon et al., 2002) or deterrents (Fraenkel, 1959; Qiu, van Loon, and Roessingh, 1998), either individually or in combination (Fraenkel, 1959; Thompson and Pellmyr, 1991; Schoonhoven, van Loon, and Dicke, 2005).

11

b. Ultimate factors Ecologically, the much larger numbers of relative specialists imply that specialization has greater overall advantages than polyphagy (Bernays and Graham, 1988). Many hypotheses about advantages of specialism over generalism exist, including predator avoidance (Bernays and Graham, 1988), food predictability (Jaenike, 1978; Futuyma and Moreno, 1988; Novotny, 1994), feeding efficiency (Ehrlich and Raven, 1964; Bernays and Graham, 1988; Singer, Ng, and Thomas, 1988; Thompson and Pellmyr, 1991; Cornell and Hawkins, 2003), location of mates (Bernays and Graham, 1988; Jaenike, 1990) and information processing (Bernays, 2001; Janz, 2003). But host associations may also be highly evolutionary constrained, preventing optimal responses to changing selective forces (Moran, 1988). All in all it is clear that host plant specialisation is an important process, but the mechanisms underlying specialization are not fully understood in ecology (Desdevises, Morand, and Legendre, 2002).

Empirical study of host specificity patterns is the first step towards the ultimate goal of explaining host plant use by evolutionary and ecological factors (Novotny and Basset, 2005), but phylogenetic studies of highly specific insect-plant associations are very few (Farrell and Mitter, 1990; Schoonhoven, van Loon, and Dicke, 2005). Here, I want to present a study on the evolution of host plant use and host specificity in the highly specific interaction between Cymothoe butterflies and their Rinorea host plants in the Afrotropics. It is the first detailed study of a system where the specificity is at the species level and involves many congeneric species of both butterflies and plants. Additionally, it will be one of the first such published study in the tropics. Tropical regions are known to have a much higher biodiversity, and although tropical herbivores do not seem to be more specialised than their temperate counterparts (Novotny et al., 2002; Novotny and Basset, 2005), high investment in chemical defences by tropical plants indicate that they do exert more selective pressure on their food plants (Coley, 1999).

1.3 The case of Cymothoe (Nymphalidae) feeding on Rinorea (Violaceae)

Rinorea

Rinorea Aublet. (1775) (Violaceae, Violoideae) comprises approximately 200 species of woody shrubs distributed in Old and New World tropics (Hekking, 1988). This makes it the most speciose genus in the tribe Rinoreeae, the other genera (i.e. Allexis, Decorsella, Gloeospermum, Hekkingia, and Rinoreocarpus) being mostly monotypic to comprising <10 species. Of the 200 Rinorea species, 110-120 are known from Africa where most species occur in Central Africa, with the centre of diversity in the coastal plain and the plateau areas of Cameroon and Gabon (Achoundong, 1996). The abundance of Rinorea species in Cameroon has been regarded a ‘forest radiation’. Local endemic Rinorea species have usually extremely small distribution areas (Achoundong, 1996), indicating poor dispersal capacity. Generally, species of Rinorea are considered to be important and sensible bio indicators for forest typification as they are usually locally common to abundant and show close association with different forest types (Achoundong, 1996). A revision of all African Rinorea has been conducted by Achoundong (1997), who has subdivided the species into 8 subgroups, mainly based on flower symmetry, inflorescence architecture and characteristics of the androecium and the stamens.

Bakker et al. (2003) have published a haplotype tree for Rinorea based on cpDNA trnL-F sequences. Their only Neotropical representative, R. crenata appeared as sister to all other included Rinorea species and despite a lack of resolution, they concluded that African Rinorea was paraphyletic with respect to the included Asian species. An optimisation of androecium symmetry showed that actinomorphy was plesiomorphic for African Rinorea. Non-monophyletic grouping of conspecifics

12

made them suspect relative young species complexes (Bakker, van Gemerden, and Achoundong, 2003). Because their evidence was based on cpDNA alone, it remains a question whether their haplotype tree is a good estimate of the organismal evolution. Corroboration with signals from other genotypes such as mtDNA or nrDNA is needed (Miyamoto and Fitch, 1995).

Cymothoe and Harma

Cymothoe Hübner (1819) (Nymphalidae, Limenitidinae, Neptini) constitutes a large genus of 72 species, all occurring in tropical Africa (Ackery, Smith, and Vane-Wright, 1995; Larsen, 2005). Around 50 of these so called Gliders are known from Cameroon. The species are mainly forest butterflies (Fontaine, 1982; Amiet and Achoundong, 1996; Larsen, 2005) though a few can survive considerable habitat degradation (Larsen, 2005, pers. obs.). They feed on fermenting fruit and nectar (Fontaine, 1982; Larsen, 2005, pers. obs.; see field report). Nearly all species show considerable sexual dimorphism. In C. oemilius, the largest species in the genus, sexual dimorphism is only slight; the male costa being white, that of the female not. But for all other species the females are generally larger and have a different form and coloration pattern than the males (Fontaine, 1982; Amiet and Achoundong, 1996; Larsen, 2005; see figure 1). The species of this genus are mainly distinguished by the ground colour of the males, which are usually some shade of white, yellow or red (see figure 1). But male similarity is often accompanied by female dissimilarities, complicating identification and classification (van de Weghe, pers. comm.). The larval stages of Cymothoe in Cameroon were studied by Dr J.-L. Amiet, with the majority of the species covered (Amiet and Achoundong, 1996). He concludes that larval characters can be important for the classification (Amiet, 2000) as well as reliable indications for the existence of some cryptic species, as in the yellow C. egesta and the red C. sangaris (Amiet, 1997). The Angular Glider, Harma Doubleday (1848) is a genus with a single widespread afrotropical species H. theobene. It is generally regarded as closely related to Cymothoe (Fontaine, 1982; Kielland, 1990), but differs strongly in wing shape (Larsen, 2005). However, Amiet (2000) argues against maintaining Harma because it is more similar to Cymothoe species in both adult and larval characters, than C. oemilius is to its congenerics. Chermock (1950), Overlaet (1955) and van Son (1979) also share the opinion that Harma and Cymothoe should be merged.

FIGURE 1 (opposite page). Morphological diversity in Harma and Cymothoe adults. 1A Harma theobene; 1B Cymothoe oemilius; 2A C. beckeri. Yellow species: 2B C. fumana; 3A C. egesta; 3B C. lucasii 4A C. reinholdii; 4B C. haynae. Red species: 5A C. ogova; 6A C. sangaris; 7A C. coccinata; 8A C. haimodia; 9A C. reginaeelisabethae. White species: 5B C. jodutta; 6B C. caenis; 7B C. indamora; 8B C. alticola; 9B C. althaea. Images are ½ x life size and taken from d’Abrera (2004) except 5A taken from Larsen (2005).

13

14

Host plant relationships

Van Someren (1939), Chermock (1950) and Fontaine (1982) already described the feeding of Cymothoe and Harma on either Achariaceae (ex. Flacourtiaceae; Chase et al., 2002) or on Rinorea (Violaceae). But Amiet & Achoundong (1996) were the first to note the high specificity of the relationships between species of Cymothoe and their Rinorea host plants in the Cameroonese forests. They discovered that 18 Cymothoe species were strictly monophagous, feeding on only one Rinorea host species and 6 more species fed on just two or three related Rinorea hosts. The three remaining Rinorea-feeding species had five or six hosts (table 1).

Rinorea ↓ ↓ C

ymo

tho

e a

ram

is

arc

ua

ta

co

ccin

ata

co

lma

nti

cro

cea

cyc

lad

es

dis

tinct

a

eg

est

a c

on

fusa

eg

est

a "o

ran

ge"

exc

els

a

fu

ma

na

ha

imo

dia

ha

rmill

a h

arm

illa

he

liad

a h

elia

da

he

sio

do

tus

hyp

ath

a

fo

nta

ne

i

luca

sii

luri

da

nig

eri

en

sis

og

ova

orp

hn

ina

pre

uss

i

re

gin

ae

elis

ab

eth

ae

sa

ng

ari

s I 1

sa

ng

ari

s I 2

sa

ng

ari

s II 1

sa

ng

ari

s II 2

amietii ���� ���� 2

angustifolia ���� ���� 2

batesii ���� ���� 2

campoensis ���� 1

caudata ���� 1

convallarioides ���� 1

dentata ���� ���� ���� 3

dewildei ���� 1

dewitii ���� 1

dimakoensis ���� 1

gabunensis ���� 1

ilicifolia ���� 1

keayi ���� ���� 2

ledermannii ���� ���� 2

lepidobotrys ���� 1

letouzeyi ���� 1

longicuspis ���� ���� ���� 3

longisepala ���� ���� 2

mezilii ���� 1

oblongifolia ���� ���� 2

preussii ���� 1

rubrotincta ���� ���� 2

sinuata ���� 1

spongiocarpa ���� 1

subsessilis ���� ���� ���� 3

verrucosa ���� 1

vivienii ���� ���� 2

welwitschii ���� ���� ���� 3

yaundensis ���� ���� 2

zenkeri ���� 1

1 1 3 1 1 1 2 6 1 1 5 2 1 1 3 1 1 1 1 2 1 2 1 5 1 1 1 TABLE 1. Host plant relationships of Cymothoe feeding on Rinorea in Cameroon. Redrawn from Amiet and Achoundong (1996).

Host plant use in Cymothoe is in general congruent with the main groupings based on male ground colour. All yellow and red Cymothoe species appear to be exclusively colonising Rinorea (Amiet and Achoundong, 1996; Amiet, 2000). Species of Achariaceae are colonised exclusively by the white species of Cymothoe, C. oemilius, C. beckeri and Harma. The only exception is C. harmilla that feeds on Rinorea. It has a white male, but a female that strongly resembles the females of species with red males (Amiet, 2000).

15

For some species feeding on Achariaceae, additional host plants from other families have been reported (table 2). The majority of these records do belong to the same order as Achariaceae and Rinorea (Malpighiales): Salicaceae, Euphorbiaceae and Hypericaceae (Ackery, 1988). But there are also records of host plants from Bignoniaceae belonging to the Lamiales (van Someren, 1974; Ackery, 1988). Together with Amiet (1996) and Larsen (2005), I consider most of these host plant records improbable, for several reasons: First, the reliability of the identifications of the insects but especially of the plants is unknown. Achariaceae are notoriously difficult to identify (Spencer and Seigler, 1985) and easily confounded with members of other families such as the Euphorbiaceae (Breteler pers. comm.). The recorded Salicaceae hosts have even long been considered as members of the same family (Flacourtiaceae) as the Achariaceae due to their morphological similarities (Chase et al., 2002). Combined with the fact that host plants are usually discovered by entomologists and not botanists, the chances of error are substantial. Second, for many data it is unclear whether they signify host plants discovered in the field, or plants on which caterpillars have successfully reared in captivity. The latter is no proof for the role as host plant for natural populations. Third, authors of review articles and butterfly guides tend to copy each others data, so that any incorrect host plant records can remain in the literature for a long time. Fourth, there is no chemical evidence to support these host plant records (see next paragraph). Because there are no recent, verifiable records of Cymothoe host plants other than Achariaceae and Rinorea, it is safe to assume that there are no other plant families playing a major role as host plant for these butterflies.

Host Plant Family ↓Genus ↓ ↓ B

utte

rfly

H

arm

a th

eo

be

ne

C. a

lcim

ed

a

C. a

ltico

la

C. a

ltisi

do

ra

C. a

ma

nie

nsi

s

C. a

uri

villi

i

C. b

eck

eri

C. ca

en

is

C. ca

pri

na

C. co

ran

us

C. h

erm

inia

C. h

yarb

ita

C. jo

du

tta

C. lu

rid

a

C. o

em

iliu

s

C. rh

ein

old

i

C. te

ita

Buchnerodendron � 1

Caloncoba � � � � � � � � 8

Heterophragma � 1

Kiggelaria 0

Lindackeria � � 2

Poggea � 1

ACHARIACEAE

Rawsonia � � � � � � � 7

Fernandoa � 1 BIGNONIACEAE

Kigelia � � 2 DICHAPETALACEAE Dichapetalum � 1

EUPHORBIACEAE Macaranga � 1 HYPERICACEAE Vismia � 1

Casearia � 1 SALICACEAE

Dovyalis � � � 3 4 1 1 1 1 1 2 3 1 5 2 2 1 2 1 1 1

TABLE 2. Reported host plant relationships of Cymothoe feeding on Achariaceae and other host plant families (van Someren, 1974; van Son, 1979; Fontaine, 1982; Kielland, 1990; Bampton, Collins, and Dowsett, 1991; Larsen, 1991; Amiet and Achoundong, 1996).

Host plant chemistry

An obvious chemical characteristic of the Achariaceae is cyanogenesis (Spencer and Seigler, 1985; Clausen et al., 2002). Cyanogenesis refers to the ability to release hydrogen cyanide (HCN), usually following damage (Jones, 1988; Thomsen and Brimer, 1997). As HCN is a universal toxicant due to its inhibition of the respiratory chain, it is generally regarded as a mechanism to inhibit herbivory and although the effectiveness of cyanogenesis as a herbivore defence system is not always clear (Hruska, 1988) because cyanogenesis is not effective against all herbivores, and not all cyanogenic plants release enough cyanide to be toxic (Gleadow and Woodrow, 2002), many cyanogenic plants are indeed avoided by herbivores (Nahrstedt, 1988; Jones, 1998).

16

The most common cyanogens found in plants occur in a glycosidic form (Thomsen and Brimer, 1997; Gleadow and Woodrow, 2002). These cyanogenic glycosides are cleaved in a catalysed process resulting in a highly unstable cyanohydrin that degrades either enzymatically or spontaneously to produce HCN (Gleadow and Woodrow, 2002).

In Achariaceae, the cyanogens are a group of cyclopentanoid glycosides that is found exclusively in the passifloraceous group containing the Malpighiales plant families Passifloraceae, Turneraceae, Malesherbiaceae and Flacourtiaceae (Spencer and Seigler, 1985; Clausen et al., 2002). These cyclopentanoid glycosides have been found in all the Achariaceae genera that are reported as Cymothoe host plants that were examined biochemically: Caloncoba (Cramer, Rehfeldt, and Spener, 1980), Kiggelaria (Raubenheimer and Elsworth, 1988), Lindackeria (Jaroszewski, Ekpe, and Witt, 2004), and Rawsonia (Andersen et al., 2001).

Although Rinorea is not cyanogenic, there are a number of butterfly genera other than Cymothoe that feed on members of the passifloraceous group and Rinorea. The passifloraceous group together with Rinorea was regarded by Ehrlich and Raven (1964) as one of the especially important butterfly host plant groups. When I performed a search on an online butterfly host plant database HOSTS (http://www.nhm.ac.uk/research-curation/projects/hostplants/) all four butterfly genera listed as feeding on Rinorea were also feeding on Achariaceae (Cymothoe, Phalantha and Terinos) and Passifloraceae and Turneraceae (Acraea; see table 3). These are strong indications of biochemical similarities among Rinorea and the cyanogenic passifloraceous group (Ehrlich and Raven, 1964; Clausen et al., 2002).

Host Plant ↓ Subfamily ↓ Genus ↓ Rinorea FLACOURTIACEAE PASSIFLORACEAE TURNERACEAE

Acraea � � � �

Phalantha � � Heliconiinae

Terinos � �

Cymothoe � � Limenitidinae Harma �

TABLE 3. Host plant relationships of Harma and butterflies feeding on Rinorea with families belonging to the passifloraceous group, according to the online HOSTS database (http://www.nhm.ac.uk/research-curation/projects/hostplants/).

In their search for these biochemical similarities, Clausen et al. (2002), discovered a novel cyclopentanoid in Rinorea ilicifolia that is similar to those of passifloraceous group. This led them to hypothesise that cyclopentanoid glycosides may be the chemical basis for the host plant associations of members of Acraea and Cymothoe butterflies (Clausen et al., 2002).

1.4 DNA barcoding

An assessment of host plant use and host specificity depends on a sound taxonomy and reliable identifications of herbivores and plants (Novotny and Basset, 2005). This is an important complication in the study of tropical systems because taxonomic treatments are usually inexistent or outdated. Fortunately, for Rinorea the taxonomy is up to date thanks to the work of Achoundong (1997). Nevertheless, leaf characters can be highly variable making it very difficult to identify sterile plants (Hawthorne and Jongkind, 2006), especially in the herbarium (Amiet and Achoundong, 1996).

For Cymothoe, a recent taxonomic treatment is lacking. Although some species are easy to recognise, there are a number of problems for others: Some species can only be reliably identified by either the male or the female and matching sexes is difficult without behavioural observations (Janzen et al., 2005). For a few Cymothoe, cryptic sibling species have been reported (Amiet, 1997). But the main problem lies in the identification of eggs and caterpillars. Although the last caterpillar stages of many species have been described by Amiet (Amiet and Achoundong, 1996; Amiet, 1997, 2000), early stages

17

are lacking discriminative characters. Because the discovery of eggs or caterpillars on plants is one of the main observations underlying host plant data, an alternative to the time consuming rearing of eggs or larvae to adulthood for identification would be a boon (Moritz and Cicero, 2004; Janzen et al., 2005).

Recently, the use of standardised short DNA sequences as a tool for species identification (DNA barcoding; Hebert et al., 2004b; Hebert and Gregory, 2005) is receiving increasing attention in the scientific world (Matz and Nielsen, 2005) and an international Consortium for the Barcode of Life (CBOL) has been installed. CBOL has more than 130 member organisations from 40 countries and is devoted to developing DNA barcoding as a global standard in taxonomy (http://barcoding.si.edu). The main scientific benefits of the DNA barcoding approach lie in the fast and digital species identification of any life stage or fragment, and the facilitation of species discoveries (Savolainen et al., 2005).

Plant DNA barcoding projects are lacking behind, which is mainly due to problems with finding a universally amplifiable DNA sequence region that provides enough variation for species identification (Cowan et al., 2006; Newmaster, Fazekas, and Ragupathy, 2006). The mitochondrial genes in plants exhibit typically much lower nucleotide substitution rates than those in animals (Newmaster, Fazekas, and Ragupathy, 2006), but see Bakker et al. (2000) for an exception. The substitution rates in plastid genes that are commonly used in phylogenetic studies of plants are even lower (Chase et al., 2005), although relatively variable regions or codon position do exist (Newmaster, Fazekas, and Ragupathy, 2006). It is clear that for plants a multi gene approach is necessary (Chase et al., 2005; Newmaster, Fazekas, and Ragupathy, 2006). In a collaborate effort to find the best barcode to apply to all land plants, a selected set of cpDNA gene regions is currently being tested for their use for species level identifications (Cowan et al., 2006).

In animals the use of the mitochondrial region cox1, that codes for a part of cytochrome c oxidase I (Slonimski and Tzagoloff, 1976), has successfully been applied as a DNA barcode in different groups such as birds (Hebert et al., 2004a), fishes (Ward et al., 2005), nematodes (Bhadury et al., 2006), sponges (Gómez et al., 2006) and many arthropods (e.g. Monaghan et al., 2005; Caterino and Tishechkin, 2006; Memon et al., 2006; Witt, Threloff, and Hebert, 2006). While applying DNA barcoding methods, Hajibabaei and co workers (2006) could discriminate among Lepidoptera species in Costa Rica. They were also able to associate sexes and reinforce the identification of butterfly species that are difficult to distinguish morphologically (Janzen et al., 2005; Hajibabaei et al., 2006). In addition DNA sequence variation has revealed a substantial number of cryptic species, differing mainly in their food plants and caterpillar morphology (Hebert et al., 2004b; Janzen et al., 2005).

An online database for the storage of DNA sequence data together with voucher information and photographs has been set up. The Barcoding of Life Database (BoLD; http://www.boldsystems.org) now contains thousands of vouchered and species-level identified mitochondrial cox1 sequences from over 6000 animal species and more than 700 species of Lepidoptera.

The main prerequisite for the effectiveness of DNA barcodes is that the maximum sequence divergence among individuals of a species does not exceed the minimum sequence divergence from another species (Matz and Nielsen, 2005; Cowan et al., 2006; Hajibabaei et al., 2006). A true test of the precision of DNA barcodes to assign individuals to species would therefore need to include comparisons with sister species (Moritz and Cicero, 2004). Because most DNA barcoding efforts have been region based (e.g. butterflies of Costa Rica, Birds of North America) instead of clade based, sequence divergences between sister species have not yet received sufficient attention (but see Kress et al., 2005 for a good example of a study regarding sister species pairs). In this study, the application of DNA barcoding for the identification of Cymothoe species is being

assessed. This clade based approach ensures that there is an efficient sampling of sister species or even sibling species, providing an adequate assessment of the performance of the DNA barcoding approach.

18

1.5 Research questions

The ultimate goals of the project presented here is to study the evolution of host plant use and host specificity of Cymothoe butterflies feeding on Rinorea. A PhD proposal titled “Evolution of host-specificity in the tropical African butterfly genus Cymothoe (Nymphalidae) feeding on Rinorea (Violaceae)” has been submitted to the Nederlandse Organisatie voor Wetenschappelijk onderzoek, Aard- en Levenswetenschappen (NWO-ALW). The research questions consider the scale and extent of this herbivore plant interaction, the phylogeny of both butterflies and host plants, and the mechanisms involved in host plant recognition, and are formulated as follows:

• What is the scale and extent of herbivore/plant interactions, i.e. o Can similar associations be found in different Cymothoe / Rinorea species? o Are specific associations constant throughout tropical Africa? o How many plant taxa have we still missed?

• Based on the butterfly and host plant molecular phylogenetic trees becoming available during the project:

o Can we infer whether Rinorea proliferation preceded that of Cymothoe, i.e. can we reject co-evolution in favour of sequential evolution, or have both clades proliferated simultaneously and are we looking at a rare occasion of true co-evolution?

o Did closely related Cymothoe species colonise closely related Rinorea species, or do we see broader host-jumping?

o Is strict monospecific host dependence phylogenetically derived? o Have Rinorea-colonising species of Cymothoe evolved from those feeding on

Achariaceae or vice versa, or, are they monophyletic within Cymothoe? • What is the nature of the interactions among plants, butterflies or caterpillars, and what are

the mechanisms involved? o Are cyclopentanoid glycosides involved in chemosensory host-plant recognition by

Cymothoe butterflies? o How do Cymothoe butterflies recognise their host, and how could this mechanism

have driven Cymothoe speciation? Could, for instance, host plant species-specific chemicals have gained physiological or ecological significance in Cymothoe species through pheromone production or through providing chemical defences against natural enemies.

o In case of ‘vacant’ Rinorea species (i.e. absence of colonising Cymothoe species) is a deterrent mechanism involved or are chemical recognition cues, viz. the cyclopentanoids, modified through derivatisation and no longer active or even absent?

Given the plethora of questions raised above, and the limitations in both time and money for an Msc thesis, a few number of questions have been selected as a basis of a pilot study aiming at evaluating the feasibility of studying Cymothoe butterflies and their associated host plants in the field (see chapter 2 for the field report) and in the laboratory. In this report I evaluate the preliminary phylogenetic patterns in the light of host plant association patterns as well as the use of DNA barcodes as a tool in species identification. The specific questions addressed are the following:

Phylogeny estimates

• Are the taxonomic groups as defined by Achoundong congruent with the molecular phylogenetic tree based on cpDNA ITS sequences?

19

• What is the phylogenetic status of Harma in relation with Cymothoe, based on a mtDNA phylogenic tree?

• Does the molecular phylogenic tree of Cymothoe support the division of this genus according to male ground colour and host plant use?

• Have Rinorea-colonising species of Cymothoe evolved from those feeding on Achariaceae or vice versa?

Host plant associations

• Is host plant use correlated with the classification of Cymothoe and Rinorea?

• Are there cases where host jumping might have occured?

DNA barcoding

• What is the level of divergence of mtDNA cox1 sequences within and between species of Cymothoe?

• Can we use DNA barcodes to identify caterpillars and eggs of Cymothoe species?

20

2 Materials & Methods

2.1 Taxon sampling

The genus Rinorea is represented by 18 species and a total number of 30 accessions, coming from West Africa (Ivory Coast, Liberia), Central Africa (Cameroon, Gabon) and Madagascar (table 4). These species comprise six of the eight taxonomic groups recognized by Achoundong (1997). Three other members of the subfamily Violoideae; Allexis Pierre, Decorsella A. Chev. (both Rinoreeae) and Hybanthus Jacq. (Violeae), were used as outgroups. Hybanthus was used to root the trees, because it belongs to a different tribe.

TABLE 4. Sampled taxa of Rinorea and Violaceae outgroups.

I have sampled 59 Cymothoe and Harma butterflies, comprising 12 and 1 species respectively (table 5). Although the number of species is low, eight of the ten main taxonomic groups as recognized by Amiet (2000) are represented. An additional 12 eggs and 23 caterpillars were sampled to assess the applicability of DNA barcoding for their identification. All of these samples were collected in Cameroon between April 10 and June 8, 2006 (see field report). For the phylogeny reconstruction, 10 of the 13 collected species could be analysed, with an additional 2 Cymothoe species, of which Carloyn McBride MSc (UC Davis, USA) has kindly provided the sequences. I have also added GenBank sequences for three Limenitidinae outgroup taxa Adelpha (Limenitidini), Euphaedra (Adoliadini) and Parthenos (Parthenini). The tribe Parthenini is considered as sister to all other tribes in the Limenitidinae (Wahlberg, Weingartner, and Nylin, 2003), so the latter was used to root the trees.

Genus Species Author SampleID Collection Locality Country Allexis cauliflora Pierre allex003 Andel 3244 Kribi, Manale Cameroon Decorsella paradoxa A.Chev. decor116 Jongkind 6655 Grand Gedeh Liberia Hybanthus enneaspermus (L.) F.Muell. hyban020 Jongkind 2845 Volta Region Ghana Rinorea subintegrifolia (P.Beauv.) Kuntze subin021 Achoundong 2108 Edea, 5km on Edéa–Douala road Cameroon Rinorea subintegrifolia (P.Beauv.) Kuntze subin002 Gideon Shu 327 Bipindi–Lolodorf Cameroon Rinorea subintegrifolia (P.Beauv.) Kuntze subin113 Jongkind 6054 Grand Cape Mount, north of Lake Piso Liberia Rinorea sp. nov ined. sp.110 Achoundong 2337 Sangmelina, Meyomessala Cameroon Rinorea simonae Achoundong simon067 Achoundong 2110 Bella SW of Edéa Cameroon Rinorea rubrotincta Chipp rubro066 Achoundong 2124 Right band of Sanaga River Cameroon Rinorea ovata Chipp ovata026 Achoundong 2122 Njabilobe, Kribi–Ebolowa road Cameroon Rinorea ovata Chipp ovata001 Gideon Shu 7751 Bipindi–Lolodorf Cameroon Rinorea mezilii Achoundong mezil064 Achoundong ? Kribi, Boaben Cameroon Rinorea mezilii Achoundong mezil016 Achoundong 2119 W of Song Mbong Cameroon Rinorea longicuspis Engl. longc008 Achoundong 2112 Elogbantindi, Bella SW of Edéa Cameroon Rinorea liberica Engl. liber114 Jongkind 6148 Grand Cape Mount, Grand Cape Mountains Liberia Rinorea letouzeyii Achoundong letou063 Achoundong 2180 Route Bella Cameroon Rinorea leiophylla M.Brandt leiop062 Achoundong 2115 Elogbantindi, Bella SW of Edéa Cameroon Rinorea kamerunensis Engl. kamer006 Achoundong 2111 Elogbantindi, Bella SW of Edéa Cameroon Rinorea ilicifolia (Welw. ex Oliv.) Kuntze ilici028 Jongkind 3314 Mahajanga (Majunga) Madagascar Rinorea ilicifolia (Welw. ex Oliv.) Kuntze ilici115 Jongkind 6241 Grand Gedeh Liberia Rinorea ilicifolia (Welw. ex Oliv.) Kuntze ilici112 Wieringa 5388 Agboville, Yapo forest Ivory Coast Rinorea gabunensis Engl. gabun024 Achoundong 2118 Douala–Edéa road Cameroon Rinorea gabunensis Engl. gabun042 Wieringa 4419 Ngounié, 5km on the road Ikobey–Bakongue Gabon Rinorea gabunensis Engl. gabun060 Wieringa 4451 Ngounié, 10km on the road Ikobey–Bakongue Gabon Rinorea gabunensis Engl. gabun072 Wieringa 4566 Ngounié, 27km road Mimongo/Koulamoutou Gabon Rinorea exappendiculata Engl. ex M.Brandt exapp059 Wieringa 4382 Ngounié, Sindara Gabon Rinorea dimakoensis Achoundong dimak108 Achoundong 2167 Campo, Nkoe Elon Cameroon Rinorea dimakoensis Achoundong dimak057 Sonke 2591 10 km E of Kika village Cameroon Rinorea campoensis M.Brandt campo030 Gideon Shu 8294 Bipindi–Lolodorf Cameroon Rinorea campoensis M.Brandt campo055 Wieringa 4640 Ngounié, 33 km on road Mbigou/Malinga Gabon Rinorea breviracemosa Chipp brevi109 Achoundong 2208 Akak Cameroon Rinorea breviracemosa Chipp brevi111 Valkenburg 2549 Nyanga, Doudou Mountains Gabon Rinorea angustifolia (Thouars) Baill. angus054 Achoundong 2117 Douala–Edéa road Cameroon

21

Genus Species Author Stage Sample ID Collection Locality Country Adelpha bredowii Geyer adult AY788591 unknown Oregon, Benton co. USA Cymothoe beckeri Herrich-Schaeffer adult RVV B290 Velzen B290 Edea, Ducam-Duclair Cameroon Cymothoe beckeri Herrich-Schaeffer adult RVV B298 Velzen B298 Edea, Ducam-Duclair Cameroon Cymothoe beckeri Herrich-Schaeffer adult RVV B365 Velzen B365 Kribi, Londji 2 Cameroon Cymothoe beckeri Herrich-Schaeffer adult RVV B374 Velzen B374 Kribi, Lobe 1 Cameroon Cymothoe beckeri Herrich-Schaeffer adult RVV B376 Velzen B376 Kribi, Bissiang Cameroon Cymothoe beckeri Herrich-Schaeffer adult RVV B386 Velzen B386 Kribi, Nkolo Cameroon Cymothoe beckeri Herrich-Schaeffer adult RVV B389 Velzen B389 Kribi, Bissiang Cameroon Cymothoe caenis Drury adult RVV B249 Velzen B249 Edea, Malimba Cameroon Cymothoe caenis Drury adult RVV B364 Velzen B364 Kribi, Londji 2 Cameroon Cymothoe caenis Drury adult RVV B383 Velzen B383 Kribi, Londji 2 Cameroon Cymothoe caenis Drury adult RVV B384 Velzen B384 Kribi, Djabilobe Cameroon Cymothoe caenis Drury adult RVV B385 Velzen B385 Kribi, Djabilobe Cameroon Cymothoe egesta Cramer adult RVV B081 Velzen B081 Yaoundé, Kala Cameroon Cymothoe egesta Cramer adult RVV B082 Velzen B082 Yaoundé, Kala Cameroon Cymothoe egesta Cramer adult RVV B266 Velzen B266 Edea, Eding Cameroon Cymothoe egesta Cramer adult RVV B270 Velzen B270 Edea, Eding Cameroon Cymothoe egesta Cramer adult RVV B300 Velzen B300 Edea, Ducam-Duclair Cameroon Cymothoe egesta Cramer adult RVV B332 Velzen B332 Edea, Ducam-Duclair Cameroon Cymothoe egesta Cramer adult RVV B372 Velzen B372 Kribi, Londji 2 Cameroon Cymothoe egesta Cramer caterpillar RVV C025 Velzen C025 Yaoundé, Elounden Cameroon Cymothoe egesta Cramer egg RVV E007 Velzen E007 Edea, Ducam-Duclair Cameroon Cymothoe fontainei Overlaet adult CSM379 McBride 379 unknown Cameroon Cymothoe fumana Westwood adult RVV B080 Velzen B080 Yaoundé, Kala Cameroon Cymothoe fumana Westwood adult RVV B107 Velzen B107 Yaoundé, Kala Cameroon Cymothoe fumana Westwood adult RVV B321 Velzen B321 Edea, Ducam-Duclair Cameroon Cymothoe fumana Westwood adult RVV B378 Velzen B378 Kribi, Londji 2 Cameroon Cymothoe fumana Westwood caterpillar RVV C001 Velzen C001 Yaoundé, Kala Cameroon Cymothoe fumana Westwood caterpillar RVV C002 Velzen C002 Yaoundé, Kala Cameroon Cymothoe fumana Westwood caterpillar RVV C009 Velzen C009 Yaoundé, Kala Cameroon Cymothoe fumana Westwood caterpillar RVV C013 Velzen C013 Yaoundé, Kala Cameroon Cymothoe herminia Grose-Smith adult RVV B226 Velzen B226 Yaoundé, Kala Cameroon Cymothoe indamora Hewitson adult RVV B319 Velzen B319 Edea, Ducam-Duclair Cameroon Cymothoe jodutta Westwood adult RVV B060 Velzen B060 Yaoundé, Kala Cameroon Cymothoe jodutta Westwood adult RVV B169 Velzen B169 Yaoundé, Kala Cameroon Cymothoe oemilius Doumet adult RVV B322 Velzen B322 Edea, Ducam-Duclair Cameroon Cymothoe oemilius Doumet adult RVV B361 Velzen B361 Kribi, Londji 2 Cameroon Cymothoe oemilius Doumet adult RVV B377 Velzen B377 Kribi, Londji 2 Cameroon Cymothoe ogova Plötz adult CSM337 McBride 337 unknown Cameroon Cymothoe sangaris Godart adult RVV B061 Velzen B061 Yaoundé, Kala Cameroon Cymothoe sangaris Godart adult RVV B199 Velzen B199 Yaoundé, Kala Cameroon Cymothoe sangaris Godart adult RVV B240 Velzen B240 Yaoundé, Elounden Cameroon Cymothoe sangaris Godart adult RVV B344 Velzen B344 Yaoundé, Elounden Cameroon Cymothoe sangaris Godart adult RVV B346 Velzen B346 Yaoundé, Elounden Cameroon Cymothoe sangaris Godart adult RVV B354 Velzen B354 Yaoundé, Elounden Cameroon Cymothoe sangaris Godart egg RVV E010 Velzen E010 Yaoundé, Elounden Cameroon Cymothoe sangaris Godart egg RVV E011 Velzen E011 Yaoundé, Elounden Cameroon Euphaedra herbertii Sharpe adult AY218241 unknown Lesombo River Zambia Harma theobene Doubleday adult RVV B224 Velzen B224 Yaoundé, Kala Cameroon Harma theobene Doubleday adult RVV B225 Velzen B225 Yaoundé, Kala Cameroon Harma theobene Doubleday adult RVV B235 Velzen B235 Yaoundé, Elounden Cameroon Harma theobene Doubleday adult RVV B241 Velzen B241 Yaoundé, Elounden Cameroon Harma theobene Doubleday adult RVV B265 Velzen B265 Edea, Eding Cameroon Harma theobene Doubleday adult RVV B275 Velzen B275 Edea, Eding Cameroon Harma theobene Doubleday adult RVV B287 Velzen B287 Edea, Ducam-Duclair Cameroon Harma theobene Doubleday adult RVV B299 Velzen B299 Edea, Ducam-Duclair Cameroon Harma theobene Doubleday adult RVV B351 Velzen B351 Yaoundé, Elounden Cameroon Harma theobene Doubleday adult RVV B356 Velzen B356 Yaoundé, Elounden Cameroon Harma theobene Doubleday adult RVV B379 Velzen B379 Kribi, Londji 2 Cameroon Harma theobene Doubleday egg RVV E001 Velzen E001 Yaoundé, Elounden Cameroon Harma theobene Doubleday egg RVV E009 Velzen E009 Yaoundé, Elounden Cameroon Parthenos sylvia Cramer adult AY090218 unknown Stratford Butterfly Farm UK

TABLE 5. Sampled taxa of Cymothoe, Harma and Limenitidinae outgroups.

To have a better view of the C. egesta species complex, DNA sequences of another 15 Cameroonian specimens collected by Amiet and McBride were added: eight C. egesta “orange” and seven C. egesta confusa (see table 6 for the taxa analysed). C. fontainei, C. fumana, C. ogova and C. fumana were used as outgroup taxa, Harma theobene was used to root the tree.

22

Genus Species Form Stage Sample ID Collection Locality Country Host Plant Cymothoe egesta "orange" adult CSM345 McBride 345 Campo Cameroon R. breviracemosa Cymothoe egesta "orange" adult CSM349 McBride 349 Campo Cameroon R. breviracemosa Cymothoe egesta "orange" adult CSM350 McBride 350 Campo Cameroon R. breviracemosa Cymothoe egesta "orange" adult CSM355 McBride 355 Campo Cameroon R. breviracemosa Cymothoe egesta "orange" adult CSM356 McBride 356 Campo Cameroon R. breviracemosa Cymothoe egesta "orange" adult RVV B081 Velzen B081 Yaoundé, Kala Cameroon unknown Cymothoe egesta "orange" adult RVV B082 Velzen B082 Yaoundé, Kala Cameroon unknown Cymothoe egesta "orange" adult JLA5 Amiet. Yaoundé, Zamakoe Cameroon R. lepidobotrys Cymothoe egesta "orange" adult JLA6 Amiet. Yaoundé, Zamakoe Cameroon R. lepidobotrys Cymothoe egesta "orange" adult JLA11 Amiet Yaoundé, Zamakoe Cameroon R. lepidobotrys Cymothoe egesta confusa adult JLA2 Amiet Mbalmayo, Nkolngock Cameroon R. keayii Cymothoe egesta confusa adult JLA7 Amiet East Province, Dimako Cameroon R. dimakoensis Cymothoe egesta confusa adult CSM333 McBride 333 Edea Cameroon unknown Cymothoe egesta confusa adult RVV B300 Velzen B300 Edea, Ducam-Duclair Cameroon unknown Cymothoe egesta confusa adult RVV B332 Velzen B332 Edea, Ducam-Duclair Cameroon unknown Cymothoe egesta confusa egg RVV E007 Velzen E007 Edea, Ducam-Duclair Cameroon R. mezilii Cymothoe egesta confusa adult RVV B266 Velzen B266 Edea, Eding Cameroon unknown Cymothoe egesta confusa adult RVV B270 Velzen B270 Edea, Eding Cameroon unknown Cymothoe egesta confusa adult JLA9 Amiet, J.-L. Edea, Ndoupe Cameroon R. mezilii Cymothoe egesta confusa adult RVV B372 Velzen B372 Kribi, Londji 2 Cameroon R. ilicifolia Cymothoe egesta confusa adult JLA10 Amiet, J.-L. Mbalmayo, Memian Cameroon R. ilicifolia Cymothoe egesta confusa caterpillar RVV C025 Velzen C025 Yaoundé, Elounden Cameroon R. ilicifolia Cymothoe egesta confusa adult JLA3 Amiet, J.-L. Yaoundé, Kala Cameroon R. ilicifolia Cymothoe egesta confusa adult JLA8 Amiet, J.-L. Yaoundé, Kala Cameroon R. keayii Cymothoe fontainei adult CSM379 McBride 379 unknown Cameroon Cymothoe fumana adult RVV B378 Velzen B378 Kribi, Londji 2 Cameroon Cymothoe ogova adult CSM337 McBride 337 unknown Cameroon Harma theobene adult RVV B356 Velzen B356 Yaoundé, Elounden Cameroon

TABLE 6. Sampled taxa of Cymothoe egesta and Limenitidini outgroups.

2.2 DNA extraction, amplification and sequencing

The Violaceae sequences were produced by Lestrade, Bakker and Vrielink at the molecular lab of the Biosystematics Group in Wageningen. DNA was extracted from either silica dried leaves or herbarium collections, using the modified CTAB extraction protocol of Doyle & Doyle (1987), following the DNA extraction protocol of the Biosystematics Group.

The nuclear internal transcribed spacer (ITS) regions were amplified using the primer pair ITS1 (5’-TCCGTAGGTGAACCTGCGG-3’) and ITS4 (5’-TCCTCCGCTTATTGATATGC-3’). When these primers did not produce a PCR product, two internal primers were added, using primer pairs ITS1 with ITS2 (5’-GCTACGTTCTTCATCGATGC-3’) and ITS3 (5’-GCATCGATGAAGAACGTAGC-3’) with ITS4. Amplification was performed in a volume of 50 μL containing genomic DNA, 0.2 μM of each primer and each dNTP, 3 mM MgCl2, 75 mM Tris-HCL, 20 mM (NH4)2SO4, 0.01% Tween 20, and 1.5 unit Taq Polymerase. The following temperature profile was used: 30 cycles of 0:30 min. denaturisation at 94°C, 0:30 min. annealing at 55°C, 1:30 min. extension at 72°C, followed by one cycle of 10 min. at 72°C.

After visual inspection on a 1%-agarose gel, fragments were cleaned using Qiaquick purification columns (Qiagen) following the manufacturer’s protocol and eluted in MQ water. Cleaned fragments were sequenced using a fluorescent dye-labelled sequencing reaction (DYEnamic™ ET Terminator Cycle Sequencing Kit; Amersham Biosciences), and the high-throughput ABI sequencing facilities at Greenomics™, Wageningen. All generated sequences will be submitted to GenBank (www.ncbi.nlm.nih.gov/Genbank). The Lepidoptera samples were processed by the Canadian centre for DNA barcoding in Guelph, run by dr. Paul Hebert’s group who has kindly offered to sequence DNA barcodes for the butterflies, catepillars and eggs.. DNA was extracted from single legs, eggs or caterpillar prolegs using an automated Glass Fiber plate extraction method on a Beckman-Coulter Biomek NX Span-8 liquid handling system (Ivanova, deWaard, and Hebert, 2006 in press).

23

The target 658-bp fragment of cox1 was amplified using primer pair LepF (5’-ATTCAACCAATCATAAAGATATTGG-3’) and LepR (5’-TAAACTTCTGGATGTCCAAAAAATCA-3’). In the cases where these primers did not produce a PCR product, two internal mini-primers were added, using primer pairs LepF with mLR (5’- cctgttccagctccattttc-3’) and mLF (5’-gctttcccacgaataaataata-3’) with LepR. Amplification was performed in a volume of 12.5 μL containing 2 μL H2O, 1.25 μL 10x buffer, 2 μL DNA template, 0.125 μL of each primer, 0.0625 μL dNTP’s, 0.625 μL MgCl2, 0.0625 μL Platinum Taq Polymerase (5U / μL). The following temperature profile was used: 5 cycles of 0:40 min. denaturisation at 94°C, 0:40 min. annealing at 45°C and 1:00 min. extension at 72°C, followed by 35 cycles of 0:40 min. denaturisation at 94°C, 0:40 min. annealing at 51°C and 1:00 min. extension at 72°C and one cycle of 5 min. at 720:40 min. denaturisation at 94°C, 0:40 min. annealing at 45°C and 1:00 min. extension at 72°C. All generated sequences, together with photographs and collection details, will be deposited at the

Barcoding of Life Database for DNA barcodes (BoLD; www.boldsystems.org) under the project code CYMO.

2.3 DNA barcoding

DNA barcoding analyses were done through the online interface of the BoLD website. The taxon identification tree was based on the Kimura 2 parameter distance model, with the Specimen sample ID as terminal branch label, the filter set to sequences with length >100 basepairs, and all codon positions included. The resulting tree was transformed into a radial tree in TreeIllustrator 0.52 beta (Trooskens et al., 2005), and coloured according to the species. A nearest neighbour summary was provided through the online interface of BoLD, with only sequences longer than 425 base pairs included in order to ensure reliable sequence distance estimates (Hanner & al. in press).

2.4 Phylogeny reconstruction

Violaceae ITS sequences were assembled and edited using the Staden software version 1.5.3 (Staden, 1996), and aligned using the ClustalW multiple alignment option (Thompson, Higgins, and Gibson, 1994) in the BioEdit sequence alignment editor 7.0.4.1 (Hall, 1999). Further refinement of the alignment was done by eye using the same software, and MacClade version 4.07 (PPC) (Maddison and Maddison, 1989). Insertions and deletions (indels) were coded following the ‘simple indel coding’ protocol as described by Simmons & Ochoterena (2000). Sequences of the cox1 marker for Cymothoe and Harma generated for DNA barcoding and for the outgroup taxa coming from GenBank were aligned by eye using BioEdit.

Phylogenies were inferred using Bayesian methods in the program MrBayes 3.1.2 (Ronquist and Huelsenbeck, 2003). For the ITS and cox1 sequence data the general time-reversal model (GTR; Lanave et al., 1984) with gamma-distributed rate heterogeneity (Yang, 1993) and invariant sites was used (nst=6 rates=invgamma). Because only informative indels of the ITS alignment were coded, the model for the indel partition was set accordingly (coding=informative).

All analyses consisted of four Metropolis-Coupled Markov Chain Monte Carlo (MCMCMC; Huelsenbeck et al., 2002) runs, with a sample frequency of 100. Each MCMCMC run consisted of one cold chain and two heated chains with a temperature of 0.1 to improve mixing (temp=0.1; Beiko et al., 2006). In order to assess convergence, the standard deviation of the split frequencies in the four runs was calculated every 5000 generations. The analyses were automatically stopped when this standard deviation dropped below one percent (diagnfreq=5000 stoprule=yes stopval=0.01).

24

3 Results

3.1 DNA barcoding

A preliminary taxon ID tree showed that four Cymothoe sequences were incorrectly clustering with an unknown Lepidoptera caterpillar. This can be explained by the fact that they all had very short sequences lengths (230-279 basepairs), due to low quality tracer files. Therefore I have removed these sequences from the analyses.

The radial taxon identification tree shows that all sequences cluster nicely according to the species (figure 2). It appears that 9 unknown caterpillars are not Cymothoe or Harma species. This applies for the caterpillars (C018-020) found on Scottellia klaineana (Achariaceae), on R. oblongifolia (C012, C016), and on R. longicuspis (C023-024) that are clustering with a clearly non Liminitidinae found on R. longisepala (C017). Another eight eggs and caterpillars could be confidently identified using DNA barcodes: two eggs of Harma theobene found on Lindackeria schweinfurthii (Achariaceae; E001, E009), four caterpillars of C. fumana found on R. oblongifolia (C001-002, C009, C009), one egg of C. sangaris found on R. batesii (E010) and one egg and one caterpillar of C. egesta found on R. mezilii and R. ilicifolia, respectively (E007, C025). Maximum intraspecific sequence divergence and distances to the nearest neighbour species for

sequences above 500 basepairs are given in table 7. Distances were always higher between species than within species. Intraspecific variation was absent in C. beckeri, C. fumana and C. oemilius, low for Harma theobene (0.49%) and high for C. sangaris, C. egesta and C. caenis: 1.18%, 1.95% and 2.01%, respectively. The two nearest neighbours with the least divergence were C. sangaris and C. oemilius (2.82%), the two ingroup species having the highest distance between them were Harma theobene and C. sangaris (7.14%).

Species n Mean Intra-sp Max Intra-Sp Nearest Neighbour NN sample ID Distance to NN Cymothoe beckeri 5 0 0 Cymothoe oemilius RVV B377 3.51

Cymothoe caenis 5 1.17 2.01 Cymothoe sangaris RVV E011 4.65

Cymothoe egesta 5 1.09 1.95 Cymothoe sangaris RVV B199 3.33

Cymothoe fumana 4 0 0 Cymothoe oemilius RVV B377 4.1

Cymothoe indamora 1 - - Cymothoe beckeri RVV B365 4.68

Cymothoe jodutta 1 - - Cymothoe egesta RVV B372 4.11

Cymothoe oemilius 3 0 0 Cymothoe sangaris RVV B199 2.82

Cymothoe sangaris 4 0.58 1.18 Cymothoe oemilius RVV B377 2.82

Harma theobene 8 0.16 0.49 Cymothoe sangaris RVV B199 7.14 TABLE 7. Distance summaries of the within species (intraspecific) cox1 sequence diversity and the minimum distance to the nearest neighbour species (NN). Only sequences > 425 basepairs were included.

25

FIGURE 2. Radial taxon identification tree of collected Cymothoe and Harma butterflies (coloured), caterpillars (RVV C; black) and eggs (RVV E; black) . Phenogram based on cox1 sequence kimura 2 parameter distances.

3.2 Phylogeny reconstruction

Cymothoe and Harma

The Bayesian analysis was stopped after 4270000 generations of MCMCMC because the convergence diagnostic dropped below 0.01, indicating that the runs had converged on the same posterior probability distribution. The rough plots of generation versus log likelihood showed that the four runs reached stationarity after 2700 samples, which were discarded as burnin. Calculations were thus based on 160000 (4*40000 samples taken from 4000000 generations.

The proportions of successful exchanges between the chains ranged from 0.17 – 0.34; values which are indicative for good mixing. The acceptance rates of all parameter proposals were between 17.3% and 70.2%.

The resulting phylogenetic tree (figure 3) shows 13 clades in the ingroup, each representing a species. Species clades all have a posterior probability (p.p.) of over 0.95. Interspecies relationships are

26

not supported, however, except one. The clade of Harma theobene is sister to all Cymothoe in this analysis (0.98 p.p.), and the branch length between Harma theobene and Cymothoe is longer (0.89 expected changes per site) than the branch length between any Cymothoe species (C. caenis and C. jodutta 0.63 expected changes per site). Of the outgroup taxa, Adelpha is most closely related to Harma and Cymothoe (figure 3).

FIGURE 3. Phylogenetic tree of Cymothoe, Harma and outgroups based on mtDNA cox1 sequences. Node labels indicate Bayesian posterior probability values.

27

The C. egesta complex

The analysis of the C. egesta complex reached convergence (convergence diagnostic < 0.01) after 885000 generations. Rough plots of generation versus log likelihood showed that three of the four runs reached stationarity after 600 samples, which were discarded as burnin. The fourth run had an error in the probability file and had to be excluded from the stationarity assessment, but I have no reason to suspect that this run should have another burnin value. All four runs were used in the phylogeny reconstruction and calculations were thus based on 32000 (4*8000) samples taken from 800000 generations.

The proportions of successful exchanges between the chains ranged from 0.13 – 0.27; values indicating good mixing. The acceptance rates of all parameters were between 24.0% and 76.1%.

The resulting phylogenetic tree (figure 4) showed a monophyletic C. egesta (0.95 p.p.) containing two sister clades (0.75 p.p.), each containing exclusively specimens of C. egesta confusa or C. egesta “orange”, respectively.

FIGURE 4. Phylogenetic tree of Cymothoe egesta confusa, C. egesta “orange” and outgroups based on mtDNA cox1 sequences. Node labels indicate Bayesian posterior probability values.

28

Rinorea

The Bayesian analysis was stopped already after 660000 generations because the convergence diagnostic dropped below 0.01, indicating that the runs had converged earlier on the same posterior probability distribution than the runs of the Cymothoe and Harma analysis. The rough plots of generation versus log likelihood showed that the four runs reached stationarity after 600 samples, which were discarded as burnin. Calculations were thus based on 24000 (4*6000) samples taken from 600000 generations.

The proportions of successful exchanges between the chains ranged from 0.58 – 0.62; values which are indicative for good mixing. The acceptance rates of all parameters were between 16.7% and 91.8%. The resulting haplotype tree (figure 5) has R. angustifolia and R. exappendiculata as sister to all other Rinorea (0.89 p.p.), which show a polytomy with four highly supported (1.00 p.p.) clades: The first clade contains two accessions of R. breviracemosa. The second clade contains the species R. campoensis and R. subintegrifolia, both member of subgroup IIB. The third clade consists of six species of the tribe ilicifoliae (IIA1): R. dimakoensis, R. ilicifolia, R. letouzeyii, R. mezilii, R. simonei and an undescribed species. The fourth clade is a cluster of six species from subgroups IIA2 and IIA3: R. gabunensis, R. kamerunensis, R. leiophylla, R. longicuspis, R. ovata and R. rubrotincta. Rinorea liberica (IIA2) seems to be sister to this clade, but this relationship has a very low posterior probability value (0.54). Within each clade, species relationships are unclear with different accessions of the same species not clustering together (figure 5).

FIGURE 5. Phylogenetic tree of Rinorea and outgroups based on nrDNA ITS sequences. Species coloured according to taxonomic subgroup (Achoundong, 1997). Node labels indicate Bayesian posterior probability values.

29

FIGURE 6. Phylogenetic tree of Rinorea and outgroups based on cpDNA trnL-F sequences, redrawn from Bakker et al. (2003). Species coloured according to taxonomic subgroup (Achoundong, 1997). Node labels indicate Bayesian posterior probability values.

30

4 Discussion

4.1 DNA barcoding

The use of cox1 sequences as DNA barcodes seems to be a good tool for the identification of Cymothoe species. There was no intraspecific sequence variation for the species C. fumana, C. oemilius and Harma theobene, even though the sequences were from specimens collected in different regions in Cameroon. Sequence variation was always lower within than between species. I was able to positively match samples of eggs and caterpillars with those of identified adult butterflies, which can facilitate a quick and easy collection of host plant relationship data.

A number of flaws in this study need to be discussed, however. First, the number of samples was very low, both in number of species and in the number of sampled individuals per species. An assessment on the reliability of DNA barcoding for species identification depends greatly on the number of samples (Moritz and Cicero, 2004; Matz and Nielsen, 2005), and there need to be at least 10 samples per species from what seems to be one site (Janzen et al., 2005). In this study the only adequately sampled species is Harma theobene, for the other species the number of samples ranged from only 1 to 9. Moreover, ideally all species within a genus should be included to be sure that the between species distance always exceeds the intraspecific variation (Moritz and Cicero, 2004). Finally, taxa should be included from more than one geographic region to adequately reveal the extent of intraspecific variation (Moritz and Cicero, 2004; Janzen et al., 2005). Although we did include samples from three different regions in Cameroon, this geographical range is incomparable to the distribution of some of the widespread species occurring throughout tropical Africa. All the same, I feel that there are not much problems to be expected, for several reasons. First, the

cox1 sequences can discriminate between cryptic sibling species that have not been discovered until recently due to their high morphological similarities (see the next paragraphs). Therefore, it is likely that other, less closely related species can also be discriminated based on DNA barcodes. Second, a first analysis including a sequence of C. egesta confusa from Zambia shows that it falls nicely within the variation already found in the Cameroonian specimens (McBride, pers. comm.), indicating that intraspecific divergence is low even at a wide geographical scale.

4.2 Phylogenetic estimation

Rinorea

The main rDNA ITS clades as supported in the Bayesian phylogeny estimation represent subgroups that were described by Achoundong (1997). The overall pattern is congruent with the previous Rinorea phylogeny estimation based on cpDNA trnL-F sequences (Bakker, van Gemerden, and Achoundong, 2003), where the subgroups IIA1, IIA2+3 and IIB were also apparent (figure 6). At the subgroup level there were no apparent incongruencies between the haplotype trees based on cpDNA trnL-F and nrDNA ITS sequences, indicating phylogenetic congruence (Miyamoto and Fitch, 1995). The overall resolution, however, is higher. For both subgroups IIA1 and IIB the posterior probabilities have increased from 0.75 to 1.00 and 0.95 to 1.00, respectively. R. longicuspis (IIA3) has moved from an unresolved position into the IIA2+3 clade (1.00 p.p.). While the backbone of African Rinorea was completely unresolved in the trnL-F phylogeny, there now seems support for R. angustifolia (subgroup IA1) being sister to all other African Rinorea. This could be expected because this species is regarded as

31

having many ancestral characters (Achoundong, 1997; Amiet, 2000). Probably subgroup IB2 (not included here) shares this position, because it appeared as sister to R. angustifolia in the trnL-F haplotype tree (Bakker, van Gemerden, and Achoundong, 2003; figure 6). This would indicate that the supergroups I and II are sisters. The status of R. exappendiculata remains unclear as it still does not cluster with its proposed subgroup IIB, but remains unresolved next to R. angustifolia.

It appears that the taxonomic groups as defined by Achoundong (Achoundong, 1997) are natural, with a basic division of the supergroups I and II, with support for subgroups IA1, IB2, IIA1, IIB (except for R. exappendiculata) and combined subgroups IIA2 and IIA3 (figure 5). For the latter two subgroups it appears that IIA3 renders IIA2 paraphyletic, but this pattern may very well change when more taxa and more markers are added. In any case, they also have a similar androecium and inflorescence structure (Achoundong, 1997).

Cymothoe and Harma

The mtDNA haplotype tree of Cymothoe and Harma based on cox1 sequences (figure 3) shows the same overall patterns as the phenetic tree used for DNA barcoding (figure 2), with all species being monophyletic with high probability (0.96-1.00 p.p.). Within the species C. egesta, C. sangaris and C. caenis intraspecific variation indicates phylogenetic patterns below the species level, which will be further discussed in the next paragraph.

Virtually all nodes above species level, however, are collapsed. Apparently interspecies DNA sequence synapomorphies had been lost due to saturation. This is not surprising as cox1 is a relatively fast evolving marker, and shows deep divergences among species (Hebert and Gregory, 2005). Many authors have already warned that DNA barcodes are not to be used for phylogeny reconstruction (e.g. Moritz and Cicero, 2004; Will and Rubinoff, 2004; Hebert and Gregory, 2005; Janzen et al., 2005), a message that our results underline. Nevertheless, there are some interesting patterns to be seen.

The molecular phylogenetic tree presented here shows that Harma can be maintained as a monotypic genus. Although it is closely related to Cymothoe, there is enough divergence, either measured in distance or branch lengths, between the two genera to support its generic status. It is clear, however, that Harma and Cymothoe are sisters, as can be seen by their morphological similarities and the fact that all higher level molecular phylogeny estimates to date have shown these two genera as a clade (e.g. Wahlberg, Weingartner, and Nylin, 2003; Janz, Nylin, and Wahlberg, 2006).

This enables us to hypothesise about the evolution of host plant use in Cymothoe. Because its sister genus Harma shows the same relation with members of the Achariaceae as the species with white males, it is most parsimonious to assume that it is the ancestral host plant use. This means that the Rinorea feeding species have probably evolved from those feeding on Achariaceae. In lack of knowledge about the interspecies relationships within Cymothoe it is not possible to make any conclusions, but the congruence of host plant use with morphological characters such as male ground colour (white in Acharaiaceae feeding species, yellow or red in Rinorea feeding ones) indicates that the shift from Achariaceae- to Rinorea-feeding has occurred only once, and that there have been no reversals. Question remains whether the groups based on male ground colour are monophyletic, or that there

is a sequence with one group evolving from within another. The phylogenetic position of species that deviate from these groups such as C. oemilius and C. beckeri (figure 1) is also still unknown. In order to resolve the interspecies relationships in Cymothoe, addition of both species and (nrDNA) markers is needed.

32

4.3 Cryptic sibling species in Cymothoe

In at least three species, intraspecific DNA sequence variation indicates that they may contain two or more cryptic sibling species: C. egesta, C. sangaris and C. caenis. This is concordant with the finding of Amiet who claims that C. egesta and C. sangaris each consist of several cryptic sibling species (Amiet, 1997).

The case of C. egesta has been looked at in more detail (figure 4), thanks to the study of McBride from UC Davis (USA) who has collected and sequenced specimens from two different forms of this species. In Cameroon Amiet (1997) has discovered that, apart from the widespread C. egesta confusa, another form exists which is smaller, more orange in tone, a darker female and a different larva (C. egesta “orange”, possibly C. megaesta Staudinger 1890). While C. egesta confusa feeds on Rinorea species of the Ilicifoliae (subgroup IIA1), C. egesta “orange” is feeding on and R. lepidobotrys or R. breviracemosa (subgroup IA2). He concluded that these forms are distinct species (Amiet, 1997).

The DNA barcode sequence diversities within C. egesta samples show high levels of divergence between two groups (table 8). The phylogenetic tree of my C. egesta collections combined with the sequences from C. McBride confirms that the two forms represent two distinct, reproductively isolated species. Even sympatric specimens group according to their morphological differences and food plants (figure 4). The sequence of a recently added C. egesta confusa from Zambia falls nicely within the C. egesta confusa clade (McBride, pers. comm.), demonstrating that the sequence variation within this clade covers the entire geographical range of the species. Also, sequences of the nominate subspecies C. egesta egesta from West Africa (Ghana and Sierra Leone) are clustering in the C. egesta “orange” clade, indicating that these are sister species. Maybe C. egesta “orange” has formed when C. egesta egesta made secondary contact with C. egesta confusa within Cameroon (McBride, pers. comm.).

Distance Sample ID Region Locality RVV B270 RVV B300 RVV B372 RVV B081 RVV B082

RVV B270 Edéa Eding -

RVV B300 Edéa Ducam-Duclair 0 -

RVV B372 Kribi Londji 2 0 0 -

RVV B081 Yaoundé Kala 1.54 1.73 1.54 -

RVV B082 Yaoundé Kala 1.95 1.91 1.86 0.17 - TABLE 8. Sequence diversity within the C. egesta complex, based on Kimura 2 parameter distances of mtDNA cox1 sequences >425 base pairs long.

The studies of Amiet (1997) have also revealed cryptic species in the complex around C. sangaris. This complex has posed a problem for taxonomists for a long time already. Overlaet has worked on it for several years, but commented shortly before his death that “everything needs to be redone” (Larsen, 2005). The male is characterised by its blood red ground colour, but in contrast to the morphological uniformity of the males the females exhibit much variation in wing patterns and –colour (Amiet, 1997). Based on larval and pupal characters in combination with female morphology, Amiet (1997) concluded that there are at least four distinct species hiding in the C. sangaris complex, each having their own Rinorea host plant(s). We have found two populations of C. sangaris in the Yaoundé area: one at Mont Kala, feeding on R. preussi and another at Mont Eloundem feeding on R. batesii (both subgroup IIA2; Achoundong, 1997). These would be respectively the species Cymothoe sgI.1 and C. sgII.2 as described by Amiet (1997). Although the localities are only 20 km. apart, the sequence divergence is 1.15% (table 9) and the species appear as sisters in the phylogeny (figure 3). Indeed, the two C. sangaris specimens that appear as sister to all other C. sangaris correspond to Amiet’s type II.2, feeding on R. preussi.

33

Distance Sample ID Region Locality RVV B346 RVV B354 RVV E011 RVV B199

RVV B346 Yaoundé Elounden -

RVV B354 Yaoundé Elounden 0 -

RVV E011 Yaoundé Elounden 0 0 -

RVV B199 Yaoundé Kala 1.15 1.15 1.18 - TABLE 9. Sequence diversity within the C. sangaris complex, based on Kimura 2 parameter distances of mtDNA cox1 sequences >425 base pairs long.

Of the species C. caenis, specimens have been found at three localities in the littoral plain of Cameroon; two near Kribi and one near Edéa. Although the first two localities are both in the same region (Kribi), I have found a sequence divergence of 1.7% (table 10). Contrastingly, the sequences coming from one of these localities had only 0.3% difference with the sequence coming from Edéa, more than 80 kilometres north, so the sequence divergences are not related to geographical distance. The Achariaceae feeding C. caenis is less well documented by Amiet, but different forms and subspecies have been described in the past (Birket-Smith, 1960; Fontaine, 1982). When examining the four collected male butterflies, I found that the two individuals with deviating sequences share a morphological difference in that on the hindwing the blackening of the discocellular vein and veins 6 and 7 is more extended, with vein 7 almost entirely black and even an extension into vein 5 (figure 7). When taking these cryptic species into account, the minimum distance between two species decreases to 1.15% (C. sgI.1 and C. sgII.2). But the maximum within species divergence also decreases to a mere 0.49% (Harma theobene). Although DNA divergences should not be a primary criterion for recognising species boundaries (Moritz and Cicero, 2004; Will and Rubinoff, 2004), and the organism should always be taken into account (Will and Rubinoff, 2004), it seems that a sequence divergence of 1% could be used to detect cryptic sibling species in Cymothoe. The three examples above show that analysis of morphology, biogeography and DNA sequence data, through reciprocal illumination, can be very valuable in revealing cryptic species.

Distance Sample ID Region Locality RVV B249 RVV B384 RVV B385 RVV B364 RVV B383

RVV B249 Edéa Malimba -

RVV B384 Kribi Djabilobe 0.15 -

RVV B385 Kribi Djabilobe 0.3 0.15 -

RVV B364 Kribi Londji 2 2.01 1.86 1.7 -

RVV B383 Kribi Londji 2 2.01 1.86 1.7 0 - TABLE 10. Sequence diversity within C. caenis, based on Kimura 2 parameter distances of mtDNA cox1 sequences >425 base pairs long.

34

FIGURE 7. Morphological variation in the four male C. caenis specimens (life size). Top: numbering system of the veins of the hindwing; A. RVV B385; B. RVV B383; C. RVV B249; D. RVV B364.

4.4 Synthesis

Now, what can we conclude about the evolution of host plant use in Cymothoe based on the currently available data? It appears that the use of Rinorea as host plant is derived from Achariaceae feeding, but an evolutionary view on the host plant use within the Rinorea-colonising Cymothoe species has not yet been given.

In order to shed light on this matter, I have rearranged the host plant relationship data as presented by Amiet and Achoundong (1996). There are two assumptions underlying this rearrangement. First, in absence of sufficiently resolved phylogenies of Cymothoe and Rinorea, I have based the rearrangement on the most recent taxonomic classifications as evolutionary hypotheses by Achoundong (1997) and Amiet (2000). In the case of Rinorea, this seems to be a safe assumption, as the phylogenetic tree based on ITS sequences seems to indicate that its classification is natural. For Cymothoe it is not yet possible to make any conclusions, but because the classification is based on adult as well as larval characters there is no reason to suspect that it does not reflect the evolutionary relationships. Second, I have made the assumption that Cymothoe butterflies feeding on the same Rinorea host are more closely related to each other than to any other species in the same taxonomic group. Rinorea that are colonised by the same Cymothoe species are also regarded as related. This is also a safe assumption because, as outlined in the introduction, there is an evolutionary conservation in both host use and plant chemistry.

35

The host plant rearrangement is shown in table 11. It illustrates that the host plant relationships are congruent with the classification of both Cymothoe and Rinorea. The majority of Rinorea species are colonised by Cymothoe belonging to the same lineage, and there are no plants that are colonised by species from more than two lineages. Likewise, nearly all Cymothoe species feed on Rinorea coming from one group only. There were no associations that caused conflict in the rearrangement.

In general, there is a main division with the red Cymothoe species exclusively feeding on Rinorea from subgroup IIB and the yellow species feeding on all other represented Rinorea subgroups. But there are two exceptions: Even though it has a yellow ground colour Cymothoe fumana, it is feeding on the same Rinorea species as a number of red Cymothoe related to C. coccinata (COC). Similarly, species from the blood red C. sangaris complex (SAN) colonise the same Rinorea as two yellow lineages (LUR and LUC). Although in both cases there are enough characters indicating that these lineages are distinct from all other lineages with the same ground colour (Amiet, 2000), male ground colour probably is a stable trait. This means that the incongruence should be explained by other factors than phylogeny. In at least one of these two cases a host-jumping event should be assumed.

COC

FUM

OGO

EGE

LUR

SAN

LUC

HAR

HEL

CYC

Rinorea↓ ↓ C

ymo

tho

e p

reu

ssii

ara

mis

co

ccin

ata

cro

cea

re

gin

ae

elis

ab

eth

ae

arc

ua

ta

dis

tinct

a

exc

els

a

ha

imo

dia

fu

ma

na

og

ova

orp

hn

ina

eg

est

a c

on

fusa

eg

est

a "o

ran

ge"

he

sio

do

tus

co

lma

nti

hyp

ath

a

luri

da

nig

eri

en

sis

sa

ng

ari

s I 1

sa

ng

ari

s II 1

sa

ng

ari

s I 2

sa

ng

ari

s II 2

luca

sii

ha

rmill

a h

arm

illa

he

liad

a h

elia

da

fo

nta

Ine

i

cyc

lad

es

verrucosa A � 1 dewildei � 1 ledermannii B � � 2 longisepala � � 2 oblongifolia � � 2 subsessilis � � � 3 amietii � � 2 sinuata � C 1 campoensis � 1 dentata � � � D 3 yaundensis � � 2

IIB

zenkeri � 1 longicuspis E � � � 3

IIA3 welwitschii � � � 3 rubrotincta � � 2 vivienii � � 2 batesii � � 2 spongiocarpa � 1 preussii � 1

IIA2

gabunensis � 1 dewitii F � 1 dimakoensis � 1 mezilii � 1 letouzeyi � 1 ilicifolia � 1

IIA1

keayi � � 2

IA2 lepidobotrys � 1

IA1 angustifolia G � � 2

IB2 caudata � 1

IB1 convallarioides � 1 2 1 3 1 1 1 2 1 2 5 2 1 6 1 3 1 1 1 5 1 1 1 1 1 1 1 1

TABLE 11. Host plant relationships of Cymothoe and Rinorea, modified from Amiet and Achoundong (1996). Rinorea species arranged according to subgroups from Achoundong (1997), Cymothoe species arranged according to lineages from Amiet (2000). Cymothoe species with red males are in red, species with yellow or white (C. harmilla) males are in black. Shadings indicate relational groups.

36

Based on host plant and herbivore similarities, seven relational clusters can be recognised (table 11), which are discussed below.

A. Within the Rinorea subgroup IIB, R. verrucosa and R. dewildei are hosts for C. ogova (OGO) only. The other plants are colonised by species related to C. coccinata (COC) and C. fumana (FUM; see above).

B. This cluster contains all the hosts of C. fumana. These are also colonised by four Cymothoe from the COC lineage. C. hamoidia feeds on the very similar R. longisepala and R. ledermannii. The closely related R. amietii and R. subsessilis are hosts for C. distincta.

C. The two hosts of C. preussii (R. sinuata and R. campoensis) are considered to be closely related by Achoundong (1997).

D. This group is characterised by the hosts plants of C. coccinata (R. dentata, R. yaundensis and R.

zenkeri), two of which are also colonised by other Cymothoe. E. This is the relational cluster with species from the lineage of the yellow species C. lurida

(LUR), feeding on five Rinorea species from subgroups IIA2 and IIA3. All of these host plants are also fed on by two forms of the blood red C. sangaris (SAN). All other Rinorea from subgroup IIA2 are also colonised by C. sangaris or by the yellow C. lucasii (LUC). As mentioned earlier, the Rinorea subgroups IIA2 and IIA3 are morphologically similar and form a clade in the ITS phylogeny (figure 5).

F. This cluster is defined by C. egesta confusa and its host plant relationship with species exclusively coming from the Ilicifoliae (subgroup IIA1). The closely related C. orphnina is also feeding on a member of this subgroup, indicating that this is the normal preference for species of this lineage (EGE). Only the cryptic sibling species C. egesta “orange” feeds on species from the different subgroup IA2, which may indicate a host-jumping event.

G. Rinorea angustifolia is regarded as a species with many ancestral characters (Amiet, 2000), and is appearing as sister to almost all other Rinorea in the ITS phylogeny (figure 5). It is being colonised by C. harmilla (HAR) and C. heliada (HEL). The male morphology of C. harmilla is very similar to that of species feeding on Achariaceae, an indication that this species may be sister to all other Rinorea feeding species. When taking into account the possible “basal” (but see Crisp and Cook, 2005) position of both the butterfly as its host plant, this is a signal of phylogenetic congruence in Cymothoe and Rinorea. The fact that the yellow lineages of C.

heliada (HEL) and C. Cyclades (CYC) are the only ones also feeding on Rinorea from supergroup I, may point to a position as sister to the other yellow lineages LUR and LUC.

Concluding, there seems to be an overall pattern of phylogenetic conservation in the phytophagous relationship between Cymothoe and Rinorea, where related butterflies feed on related hosts. There is also an indication that the phylogenetic order of divergence among the Rinorea host plants corresponds to that among their associated Cymothoe herbivores. But there are also a few cases where the occurrence of host jumping is most likely.

37

5 Conclusions

This study has made considerable progress for the Rinorea Cymothoe project outlined in the introduction. First, I have shown that DNA barcoding is a very powerful tool for the study of Cymothoe species. It can be successfully used to reliably identify eggs and caterpillars. Additionally it can also be of value in disclosing hidden diversity in cryptic sibling species such as in C. egesta, the C.

sangaris complex and possibly C. caenis. Phylogenetic studies have shown that Harma can be maintained as monotypic genus. Within

Cymothoe, Rinorea-colonising species have probably evolved from Achariaceae-feeding ones. Rinorea phylogeny is largely concordant with the classification by Achoundong (1997).

Host plant relationship patterns suggest a phylogenetic signal and it seems that the phylogenetic order of divergence is congruent in Rinorea and their colonising Cymothoe species. Host-jumping has occurred at least two times; once in C. egesta “orange” and once in C. fumana or C. sangaris.

The fact that the associations between Rinorea and Cymothoe are likely to be explained by a combination of historical and adaptive factors makes it an ideal subject for a study of the mechanisms and relative importance of these factors in the evolution of host plant use. Better resolved species level phylogenies will certainly provide valuable data, including hypotheses about the evolution of host specificity. Timing of molecular sequence divergences can reveal whether the congruent phylogenetic patterns in Cymothoe and Rinorea have been contemporary or sequential.

6 Recommendations

In addition to the promising lines of investigation mentioned above, I have some recommendations for analyses and studies in the direct future.

Many collected Cymothoe butterflies were absent or represented by only a short cox1 sequence due to extraction problems caused by methanol contamination. A second extraction and sequencing of these specimens is needed to increase the number of included species and enhance the sequence distance estimates. Addition of specimens coming from other regions in Africa would ensure that the entire intraspecific sequence variation is taken into account. Cymothoe butterflies collected by Gael van de Weghe in Gabon and Torben Larsen in Sierra Leone and Ghana will soon be available.

In order to efficiently enhance the information content of the Rinorea phylogeny I recommend inclusion of taxa from subgroups IA2 and IB1 that have been absent from all phylogenetic analyses to date. Species of subgroup IA2 have recently been collected (see field report), so material is available. Collection of representatives of subgroup IB1 should have priority. Furthermore, more taxa from subgroup IIB should be added. This subgroup is the largest containing 23 species but it has been represented here by only 2 species. Coverage of all host species is recommended. The use of both markers ITS and trnL-F would enable assessment of conflict between nuclear and plastid signals, and the two markers combined probably contain enough sequence variation to produce a well resolved phylogeny on the level of Rinorea subgroups.

38

7 Acknowledgements

First I wish to thank Freek Bakker and Joop van Loon for their supervision during this project. Their expertise has greatly benefited the quality of my work. They have always made time when I needed advice or support. Second, I thank Jean-Louis Amiet and Gaston Achoundong. Without their excellent previous work, this study would not have been possible. Both have also been very kind in answering my numerous questions about Cymothoe and Rinorea. I am indebted to the people from staff and laboratory of the Canadian Centre for DNA Barcoding in Guelph who generously sequenced many of my butterfly collections; particularly Paul Hebert, Gregory Downs and Jeremy deWaard. Their team has also been responsible for setting up the very user friendly online BoLD database used for storing the DNA sequence and voucher data. The Rinorea sequences were available thanks to the work of Bram Lestrade and Ria Vrielink in the molecular lab of the Biosystematics Group in Wageningen. The results about Cymothoe egesta are almost entirely thanks to work done by Carolyn McBride of UC Davis, California who I have been collaborating with on this subject. Torben Larsen and Michel Beaurain have been of great help in correctly identifying butterfly specimens. Rienk de Jong and Yde Jongema have instructed me how to mount the butterflies. I thank Marleen Botermans for her support and the comments on the manuscript. This project has received financial support from the following organisations:

• Hugo de Vries Foundation

• Stichting der Korinthiërs

• Wageningen University Travel Fund

• Alberta Mennega Foundation

• Biosystematics Fund

39

8 References

ACHOUNDONG, G. 1996. Les Rinorea comme indicateurs des grands types forestiers du Cameroun. In J. v. d. Maesen [ed.], The Biodiversity of African Plants, 536-544. Kluwer Academic Publishers.

______. 1997. Rinorea du Cameroun, systématique, biologie, écologie, phytogéography., Université de Yaoundé I, Yaoundé.

ACKERY, P. R. 1988. Hostplants and Classification - a Review of Nymphalid Butterflies. Biological

Journal of the Linnean Society 33: 95-203. ACKERY, P. R., C. R. SMITH, AND R. I. VANE-WRIGHT. 1995. Carcasson’s African Butterflies: A

Catalogue of the Rhopalocera of the Afrotropical Region. CSIRO, East Melbourne. AMIET, J.-L. 1997. Trophic specialization and early stages in the Cymothoe: Taxonomic involvement

(Lepidoptera, Nymphalidae). Bulletin de la Societe Entomologique de France 102: 15-29. ______. 2000. Les premiers etats des Cymothoe: Morphologie et interet phylogenique (Lepidoptera,

Nymphalidae). Bulletin de la Societe Entomologique de France 106: 349-390. AMIET, J.-L., AND G. ACHOUNDONG. 1996. An example of trophic specialization within the

Lepidoptera: The Cameroonian Cymothoe feeding on Rinorea (Violaceae) (Lepidoptera, Nymphalidae). Bulletin de la Societe Entomologique de France 101: 449-466.

ANDERSEN, L., V. CLAUSEN, H. A. OKETCH-RABAH, M. LECHTENBERG, A. ADSERSEN, A. NAHRSTEDT, AND J. W. JAROSZEWSKI. 2001. Gynocardin and cyclopentenylglycine in Rawsonia lucida. Biochemical Systematics and Ecology 29: 219-222.

AUBLET, J. B. C. F. 1775. Histoire des plantes de la Guiane Françoise. Didot, Paris. BAKKER, F. T., B. S. VAN GEMERDEN, AND G. ACHOUNDONG. 2003. Molecular systematics of African

Rinorea Aublet. (Violaceae). Taxonomy and ecology of African plants, their conservation and sustainable use : proceedings of the 17th AETFAT congress, Addis Abeba.

BAKKER, F. T., A. CULHAM, C. E. PANKHURST, AND M. GIBB. 2000. Mitochondrial and cloroplast DNA-based phylogeny of Pelargonium (Geraniaceae) (vol 87, pg 727, 2000). American Journal of Botany 87: 933-933.

BAMPTON, I., S. C. COLLINS, AND R. DOWSETT. 1991. Butterflies (Lepidoptera) of the Mayombe and lower Kouilou (with a checklist for Congo). Tauraco Research Report 4: 57-108.

BASSET, Y., G. A. SAMUELSON, A. ALLISON, AND S. E. MILLER. 1996. How many species of host-specific insects feed on a species of tropical tree? Biological Journal of the Linnean Society 59: 201-216.

BECERRA, J. X. 2003. Synchronous coadaptation in an ancient case of herbivory. Proceedings of the

National Academy of Sciences of the United States of America 100: 12804-12807. BEIKO, R. G., J. M. KEITH, T. J. HARLOW, AND M. A. RAGAN. 2006. Searching for convergence in

phylogenetic Markov chain Monte Carlo. Systematic Biology 55: 553-565. BERNAYS, E., AND M. GRAHAM. 1988. On the Evolution of Host Specificity in Phytophagous

Arthropods. Ecology 69: 886-892. BERNAYS, E. A. 2001. Neural limitations in phytophagous insects: Implications for diet breadth and

evolution of host affiliation. Annual Review of Entomology 46: 703-727. BERNAYS, E. A., AND R. F. CHAPMAN. 1994. Host - plant selection by phytophagous insects. Chapman

& Hall, New York. BHADURY, P., M. C. AUSTEN, D. T. BILTON, P. J. D. LAMBSHEAD, A. D. ROGERS, AND G. R. SMERDON.

2006. Development and evaluation of a DNA-barcoding approach for the rapid identification of nematodes. Marine Ecology-Progress Series 320: 1-9.

BIRKET-SMITH, J. 1960. Results from the Danish Expedition to French Cameroons (1948-1950). XXVII - Lepidoptera (part II). Bulletin de l'institut Français d'Afrique Noire 22: 924-983.

BRABY, M. F., AND J. W. H. TRUEMAN. 2006. Evolution of larval host plant associations and adaptive radiation in pierid butterflies. Journal of Evolutionary Biology 19: 1677-1690.

40

CALVERT, W. H. 1974. External Morphology of Foretarsal Receptors Involved with Host Discrimination by Nymphalid Butterfly, Chlosyne-Lacinia. Annals of the Entomological Society of America 67: 853-856.

CARTER, M., P. FEENY, AND M. HARIBAL. 1999. An oviposition stimulant for spicebush swallowtail butterfly, Papilio troilus, from leaves of Sassafras albidum. Journal of Chemical Ecology 25: 1233-1245.

CATERINO, M. S., AND A. K. TISHECHKIN. 2006. DNA identification and morphological description of the first confirmed larvae of Hetaeriinae (Coleoptera : Histeridae). Systematic Entomology 31: 405-418.

CHASE, M. W., S. ZMARTZY, M. D. LLEDO, K. J. WURDACK, S. M. SWENSEN, AND M. F. FAY. 2002. When in doubt, put in Flacourtiaceae: a molecular phylogenetic analysis based on plastid rbcL DNA sequences. Kew Bulletin 57: 141-181.

CHASE, M. W., N. SALAMIN, M. WILKINSON, J. M. DUNWELL, R. P. KESANAKURTHI, N. HAIDAR, AND V. SAVOLAINEN. 2005. Land plants and DNA barcodes: short-term and long-term goals. Philosophical

Transactions of the Royal Society B-Biological Sciences 360: 1889-1895. CHERMOCK, R. L. 1950. A Generic Revision of the Limenitini of the World. American Midland Naturalist

43: 513-564. CLAUSEN, V., K. FRYDENVANG, R. KOOPMANN, L. B. JORGENSEN, D. K. ABBIW, P. EKPE, AND J. W.

JAROSZEWSKI. 2002. Plant analysis by butterflies: Occurrence of cyclopentenylglycines in Passifloraceae, Flacourtiaceae, and Turneraceae and discovery of the novel nonproteinogenic amino acid 2-(3 '-cyclopentenyl)glycine in Rinorea. Journal of Natural Products 65: 542-547.

COLEY, P. D. 1999. Paleoecology - Hungry herbivores seek a warmer world. Science 284: 2098-2099. CORNELL, H. V., AND B. A. HAWKINS. 2003. Herbivore responses to plant secondary compounds: A

test of phytochemical coevolution theory. American Naturalist 161: 507-522. COWAN, R. S., M. W. CHASE, W. J. KRESS, AND V. SAVOLAINEN. 2006. 300,000 species to identify:

problems, progress, and prospects in DNA barcoding of land plants. Taxon 55: 611-616. CRAMER, U., A. G. REHFELDT, AND F. SPENER. 1980. Isolation and Biosynthesis of

Cyclopentenylglycine, a Novel Non-Proteinogenic Amino-Acid in Flacourtiaceae. Biochemistry 19: 3074-3080.

CRAWLEY, M. J. 1985. Reduction of Oak Fecundity by Low-Density Herbivore Populations. Nature 314: 163-164.

CRISP, M. D., AND L. G. COOK. 2005. Do early branching lineages signify ancestral traits? Trends in

Ecology & Evolution 20: 122-128. D'ABRERA, B. 2004. Butterflies of the Afrotropical Region part II. Nymphalidae, Libytheidae. Hill

House Publishers, Melbourne. DESDEVISES, Y., S. MORAND, AND P. LEGENDRE. 2002. Evolution and determinants of host specificity in

the genus Lamellodiscus (Monogenea). Biological Journal of the Linnean Society 77: 431-443. DOUBLEDAY, E. 1848. The Genera of Diurnal Lepidoptera: Comprising Their Generic Characters, a

Notice of Their Habits and Transformations, and a Catalogue of the Species of Each Genus. Longman, Brown, Green & Longmans, London.

DOYLE, J. J., AND J. L. DOYLE. 1987. A rapid DNA isolation procedure for small quantities of fresh leaf tissue. Phytochemical Bulletin 19: 11-15. EHRLICH, P. R., AND P. H. RAVEN. 1964. Butterflies and Plants - a Study in Coevolution. Evolution 18:

586-608. FARRELL, B., AND C. MITTER. 1990. Phylogenesis of Insect Plant Interactions - Have Phyllobrotica Leaf

Beetles (Chrysomelidae) and the Lamiales Diversified in Parallel. Evolution 44: 1389-1403. FARRELL, B. D., AND C. MITTER. 1998. The timing of insect/plant diversification: might Tetraopes

(Coleoptera : Cerambycidae) and Asclepias (Asclepiadaceae) have co-evolved? Biological Journal of

the Linnean Society 63: 553-577. FONTAINE, M. 1982. Genre Cymothoe Hübner. Lep. Nymphalidae - S. fam. Nymphalinae. Note sur les

premiers états. Lambillionea 82: 63-64, 67-72, 95-98.

41

FOX, L. R., AND P. A. MORROW. 1981. Specialization - Species Property or Local Phenomenon. Science 211: 887-893.

FRAENKEL, G. S. 1959. Raison Detre of Secondary Plant Substances. Science 129: 1466-1470. FUTUYMA, D. J., AND G. MORENO. 1988. The Evolution of Ecological Specialization. Annual Review of

Ecology and Systematics 19: 207-233. GLEADOW, R. M., AND I. E. WOODROW. 2002. Constraints on effectiveness of cyanogenic glycosides in

herbivore defense. Journal of Chemical Ecology 28: 1301-1313. GÓMEZ, A., P. WRIGHT, D. LUNT, J. CANCINO, G. CARVALHO, AND R. HUGHES. 2006. Mating trials

validate the use of DNA barcoding to reveal cryptic speciation of a marine bryozoan taxon. Proceedings of the Royal Society B: Biological Sciences: FirstCite-FirstCite.

HAJIBABAEI, M., D. H. JANZEN, J. M. BURNS, W. HALLWACHS, AND P. D. N. HEBERT. 2006. DNA barcodes distinguish species of tropical Lepidoptera. Proceedings of the National Academy of Sciences

of the United States of America 103: 968-971. HALL, T. A. 1999. BioEdit: a user-friendly biological sequence alignment editor and analysis program

for Windows 95/98/NT. Nucleic Acids Symposium 41: 95-98. HAWTHORNE, W. D., AND C. C. H. JONGKIND. 2006. Woody Plants of Western African Forests. A guide

to the forest trees, shrubs and lianes from Senegal to Ghana Royal Botanical Gardens, Kew, Richmond.

HEBERT, P. D. N., AND T. R. GREGORY. 2005. The promise of DNA barcoding for taxonomy. Systematic

Biology 54. HEBERT, P. D. N., M. Y. STOECKLE, T. S. ZEMLAK, AND C. M. FRANCIS. 2004a. Identification of birds

through DNA barcodes. Plos Biology 2: 1657-1663. HEBERT, P. D. N., E. H. PENTON, J. M. BURNS, D. H. JANZEN, AND W. HALLWACHS. 2004b. Ten species in

one: DNA barcoding reveals cryptic species in the neotropical skipper butterfly Astraptes fulgerator. Proceedings of the National Academy of Sciences of the United States of America 101: 14812-14817.

HEKKING, W. H. A. 1988. Violaceae Part I-Rinorea and Rinoreocarpus. . The New York Botanical Garden, New York.

HRUSKA, A. J. 1988. Cyanogenic Glucosides as Defense Compounds - a Review of the Evidence. Journal

of Chemical Ecology 14: 2213-2217. HÜBNER, J. 1819. Verzeichniss bekannter Schmetterlinge. Jacob Hübner, Augsburg. HUELSENBECK, J. P., B. LARGET, R. E. MILLER, AND F. RONQUIST. 2002. Potential applications and

pitfalls of Bayesian inference of phylogeny. Systematic Biology 51: 673-688. IVANOVA, N. V., J. R. DEWAARD, AND P. D. N. HEBERT. 2006. An inexpensive, automation-friendly

protocol for recovering high-quality DNA. Molecular Ecology Notes. JAENIKE, J. 1978. Resource Predictability and Niche Breadth in Drosophila-Quinaria Species Group.

Evolution 32: 676-678. ______. 1990. Host Specialization in Phytophagous Insects. Annual Review of Ecology and Systematics 21:

243-273. JANZ, N. 2003. The cost of polyphagy: oviposition decision time vs error rate in a butterfly. Oikos 100:

493-496. JANZ, N., AND S. NYLIN. 1998. Butterflies and plants: A phylogenetic study. Evolution 52: 486-502. JANZ, N., S. NYLIN, AND N. WAHLBERG. 2006. Diversity begets diversity: host expansions and the

diversification of plant-feeding insects. Bmc Evolutionary Biology 6. JANZEN, D. H. 1980. When Is It Coevolution. Evolution 34: 611-612. JANZEN, D. H., M. HAJIBABAEI, J. M. BURNS, W. HALLWACHS, E. REMIGIO, AND P. D. N. HEBERT. 2005.

Wedding biodiversity inventory of a large and complex Lepidoptera fauna with DNA barcoding. Philosophical Transactions of the Royal Society B-Biological Sciences 360: 1835-1845.

42

JAROSZEWSKI, J. W., P. EKPE, AND M. WITT. 2004. Cyclopentanoid cyanohydrin glucosides and amides of a Lindackeria dentata. Planta Medica 70: 1001-1003.

JERMY, T. 1984. Evolution of Insect Host Plant Relationships. American Naturalist 124: 609-630. JONES, D. A. 1988. Cyanogenesis in animal-plant interactions. In D. Evered and S. Harnett [eds.],

Cyanide compounds in biology, 151-170. Wiley, Chichester. ______. 1998. Why are so many food plants cyanogenic? Phytochemistry 47: 155-162. KELBER, A. 1999. Why 'false' colours are seen by butterflies. Nature 402: 251-251. KELLEY, S. T., AND B. D. FARRELL. 1998. Is specialization a dead end? The phylogeny of host use in

Dendroctonus bark beetles (Scolytidae). Evolution 52: 1731-1743. KIELLAND, J. 1990. Butterflies of Tanzania. Hill House, Melbourne London. KRESS, W. J., K. J. WURDACK, E. A. ZIMMER, L. A. WEIGT, AND D. H. JANZEN. 2005. Use of DNA

barcodes to identify flowering plants. Proceedings of the National Academy of Sciences of the United

States of America 102: 8369-8374. LANAVE, C., G. PREPARATA, C. SACCONE, AND G. SERIO. 1984. A New Method for Calculating

Evolutionary Substitution Rates. Journal of Molecular Evolution 20: 86-93. LARSEN, T. B. 1991. The Buterflies of Kenya and their natural history. Oxford University Press, Oxford. ______. 2005. Butterflies Of West Africa. Apollo Books, Stenstrup. LETOUZEY, R. 1968. Etude phytogéographique du Cameroun. Lechavelier, Paris. LOPEZ-VAAMONDE, C., N. WIKSTRÖM, C. LABANDEIRA, H. C. J. GODFRAY, S. J. GOODMAN, AND J. M.

COOK. 2006. Fossil-calibrated molecular phylogenies reveal that leaf-mining moths radiated millions of years after their host plants. Journal of Evolutionary Biology 19: 1314-1326.

MADDISON, W. P., AND D. R. MADDISON. 1989. Interactive Analysis of Phylogeny and Character Evolution Using the Computer-Program Macclade. Folia Primatologica 53: 190-202.

MARQUIS, R. J. 2004. Herbivores rule. Science 305: 619-621. MATZ, M. V., AND R. NIELSEN. 2005. A likelihood ratio test for species membership based on DNA

sequence data. Philosophical Transactions of the Royal Society B-Biological Sciences 360: 1969-1974. MEMON, N., R. MEIER, A. MANAN, AND K. F. Y. SU. 2006. On the use of DNA sequences for

determining the species limits of a polymorphic new species in the stink bug genus Halys (Heteroptera : Pentatomidae) from Pakistan. Systematic Entomology 31: 703-710.

MENKEN, S. B. J., W. M. HERREBOUT, AND J. T. WIEBES. 1992. Small Ermine Moths (Yponomeuta) - Their Host Relations and Evolution. Annual Review of Entomology 37: 41-66.

MITTER, C., B. FARRELL, AND B. WIEGMANN. 1988. The Phylogenetic Study of Adaptive Zones - Has Phytophagy Promoted Insect Diversification. American Naturalist 132: 107-128.

MIYAMOTO, M. M., AND W. M. FITCH. 1995. Testing Species Phylogenies and Phylogenetic Methods with Congruence. Systematic Biology 44: 64-76.

MONAGHAN, M. T., M. BALKE, T. R. GREGORY, AND A. P. VOGLER. 2005. DNA-based species delineation in tropical beetles using mitochondrial and nuclear markers. Philosophical Transactions of

the Royal Society B-Biological Sciences 360: 1925-1933. MORAN, N. A. 1988. The Evolution of Host-Plant Alternation in Aphids - Evidence for Specialization as

a Dead End. American Naturalist 132: 681-706. MORITZ, C., AND C. CICERO. 2004. DNA barcoding: Promise and pitfalls. Plos Biology 2: 1529-1531. NAHRSTEDT, A. 1988. Cyanogenesis and the role of cyanogenic compounds in insects. In D. Evered and

S. Harnett [eds.], Cyanide compounds in biology. Wiley, Chichester. NEWMASTER, S. G., A. J. FAZEKAS, AND S. RAGUPATHY. 2006. DNA barcoding in land plants: evaluation

of rbcL in a multigene tiered approach. Canadian Journal of Botany-Revue Canadienne De Botanique 84: 335-341.

NOSIL, P. 2002. Transition rates between specialization and generalization in phytophagous insects. Evolution 56: 1701-1706.

43

NOVOTNY, V. 1994. Association of Polyphagy in Leafhoppers (Auchenorrhyncha, Hemiptera) with Unpredictable Environments. Oikos 70: 223-232.

NOVOTNY, V., AND Y. BASSET. 2005. Review - Host specificity of insect herbivores in tropical forests. Proceedings of the Royal Society B-Biological Sciences 272: 1083-1090.

NOVOTNY, V., Y. BASSET, S. E. MILLER, G. D. WEIBLEN, B. BREMER, L. CIZEK, AND P. DROZD. 2002. Low host specificity of herbivorous insects in a tropical forest. Nature 416: 841-844.

NOVOTNY, V., P. DROZD, S. E. MILLER, M. KULFAN, M. JANDA, Y. BASSET, AND G. D. WEIBLEN. 2006. Why are there so many species of herbivorous insects in tropical rainforests? Science 313: 1115-1118.

OVERLAET, F. G. 1955. Danaidae, Satyridae, Nymphalidae, Acracidae. Exploration du Parc National de

l’Upemba I. mission G.F. de Witte 27: 1-106. PERCY, D. M., R. D. M. PAGE, AND Q. C. B. CRONK. 2004. Plant-insect interactions: Double-dating

associated insect and plant lineages reveals asynchronous radiations. Systematic Biology 53: 120-127. PRATT, G. F. 1994. Evolution of Euphilotes (Lepidoptera, Lycaenidae) by Seasonal and Host Shifts.

Biological Journal of the Linnean Society 51: 387-416. PROKOPY, R. J., AND E. D. OWENS. 1983. Visual Detection of Plants by Herbivorous Insects. Annual

Review of Entomology 28: 337-364. QIU, Y. T., J. J. A. VAN LOON, AND P. ROESSINGH. 1998. Chemoreception of oviposition inhibiting

terpenoids in the diamondback moth Plutella xylostella. Entomologia Experimentalis Et Applicata 87: 143-155.

RADTKEY, R. R., AND M. C. SINGER. 1995. Repeated Reversals of Host-Preference Evolution in a Specialist Insect Herbivore. Evolution 49: 351-359.

RAUBENHEIMER, D., AND J. F. ELSWORTH. 1988. Gynocardin from the Leaves of Kiggelaria-Africana. Journal of Natural Products 51: 779-779.

RAUSHER, M. D. 1978. Search Image for Leaf Shape in a Butterfly. Science 200: 1071-1073. RENWICK, J. A. A., AND F. S. CHEW. 1994. Oviposition Behavior in Lepidoptera. Annual Review of

Entomology 39: 377-400. RONQUIST, F., AND J. P. HUELSENBECK. 2003. MrBayes 3: Bayesian phylogenetic inference under mixed

models. Bioinformatics 19: 1572-1574. SAVOLAINEN, V., R. S. COWAN, A. P. VOGLER, G. K. RODERICK, AND R. LANE. 2005. Towards writing the

encyclopaedia of life: an introduction to DNA barcoding. Philosophical Transactions of the Royal

Society B-Biological Sciences 360: 1805-1811. SCHOONHOVEN, L. M., J. J. A. VAN LOON, AND M. DICKE. 2005. Insect-Plant Biology. Oxford University

Press, Oxford. SIMMONS, M. P., AND H. OCHOTERENA. 2000. Gaps as characters in sequence-based phylogenetic

analyses. Systematic Biology 49: 369-381. SIMMS, E. L., AND M. D. RAUSHER. 1989. The Evolution of Resistance to Herbivory in Ipomoea

Purpurea .2. Natural-Selection by Insects and Costs of Resistance. Evolution 43: 573-585. SINGER, M. C. 1971. Evolution of Food-Plant Preference in Butterfly Euphydryas-Editha. Evolution 25:

383-&. SINGER, M. C., D. NG, AND C. D. THOMAS. 1988. Heritability of Oviposition Preference and Its

Relationship to Offspring Performance within a Single Insect Population. Evolution 42: 977-985. SLONIMSKI, P. P., AND A. TZAGOLOFF. 1976. Localization in Yeast Mitochondrial-DNA of Mutations

Expressed in a Deficiency of Cytochrome-Oxidase and-or Coenzyme Qh2-Cytochrome-C Reductase. European Journal of Biochemistry 61: 27-41.

SPENCER, K. C., AND D. S. SEIGLER. 1985. Cyanogenic Glycosides and the Systematics of the Flacourtiaceae. Biochemical Systematics and Ecology 13: 421-431.

STADEN, R. 1996. The Staden sequence analysis package. Molecular Biotechnology 5: 233-241. SYMONS, F. B., AND G. W. BECCALONI. 1999. Phylogenetic indices for measuring the diet breadths of

phytophagous insects. Oecologia 119: 427-434.

44

THOMPSON, J. D., D. G. HIGGINS, AND T. J. GIBSON. 1994. Clustal-W - Improving the Sensitivity of Progressive Multiple Sequence Alignment through Sequence Weighting, Position-Specific Gap Penalties and Weight Matrix Choice. Nucleic Acids Research 22: 4673-4680.

THOMPSON, J. N. 1989. Concepts of Coevolution. Trends in Ecology & Evolution 4: 179-183. ______. 2005. Coevolution: The geographic mosaic of coevolutionary arms races. Current Biology 15:

R992-R994. THOMPSON, J. N., AND O. PELLMYR. 1991. Evolution of Oviposition Behavior and Host Preference in

Lepidoptera. Annual Review of Entomology 36: 65-89. THOMPSON, J. N., AND B. M. CUNNINGHAM. 2002. Geographic structure and dynamics of

coevolutionary selection. Nature 417: 735-738. THOMSEN, K., AND L. BRIMER. 1997. Cyanogenic constituents in woody plants in natural lowland rain

forest in Costa Rica. Botanical Journal of the Linnean Society 124: 273-294. TROOSKENS, G., D. DE BEULE, F. DECOUTTERE, AND W. VAN CRIEKINGE. 2005. Phylogenetic trees:

visualizing, customizing and detecting incongruence. Bioinformatics 21: 3801-3802. VAN LOON, J. J. A., C. Z. WANG, J. K. NIELSEN, R. GOLS, AND Y. T. QIU. 2002. Flavonoids from cabbage

are feeding stimulants for diamondback moth larvae additional to glucosinolates: Chemoreception and behaviour. Entomologia Experimentalis Et Applicata 104: 27-34.

VAN SOMEREN, V. G. L. 1939. The butterflies of Kenya and Uganda, II, 2, family Nymphalidae, subfamily Nymphalinae. Journal of East Africa and Uganda Natural history Society 14: 15-100.

______. 1974. List of foodplants of some east african rhopalocera, with notes on the early stages of some Lycaenidae. Journal of the Lepidopterists' Society 28: 315-333.

VAN SON, G. 1979. The Butterflies of Southern Africa, Part 4, Nymphalidae: Nymphalinae. Transvaal Museum, Pretoria.

WAHLBERG, N., E. WEINGARTNER, AND S. NYLIN. 2003. Towards a better understanding of the higher systematics of Nymphalidae (Lepidoptera : Papilionoidea). Molecular Phylogenetics and Evolution 28: 473-484.

WARD, R. D., T. S. ZEMLAK, B. H. INNES, P. R. LAST, AND P. D. N. HEBERT. 2005. DNA barcoding Australia's fish species. Philosophical Transactions of the Royal Society B-Biological Sciences 360: 1847-1857.

WEINGARTNER, E., N. WAHLBERG, AND S. NYLIN. 2006. Dynamics of host plant use and species diversity in Polygonia butterflies (Nymphalidae). Journal of Evolutionary Biology 19: 483-491.

WIKLUND, C. 1975. Evolutionary Relationship between Adult Oviposition Preferences and Larval Host Plant Range in Papilio-Machaon L. Oecologia 18: 185-197.

______. 1977. Oviposition, Feeding and Spatial Separation of Breeding and Foraging Habitats in a Population of Leptidea-Sinapis (Lepidoptera). Oikos 28: 56-68.

WILF, P., C. C. LABANDEIRA, W. J. KRESS, C. L. STAINES, D. M. WINDSOR, A. L. ALLEN, AND K. R. JOHNSON. 2000. Timing the radiations of leaf beetles: Hispines on gingers from latest Cretaceous to recent. Science 289: 291-294.

WILL, K. W., AND D. RUBINOFF. 2004. Myth of the molecule: DNA barcodes for species cannot replace morphology for identification and classification. Cladistics 20: 47-55.

WITT, J. D. S., D. L. THRELOFF, AND P. D. N. HEBERT. 2006. DNA barcoding reveals extraordinary cryptic diversity in an amphipod genus: implications for desert spring conservation. Molecular

Ecology 15: 3073-3082. YANG, Z. H. 1993. Maximum-Likelihood-Estimation of Phylogeny from DNA-Sequences When

Substitution Rates Differ over Sites. Molecular Biology and Evolution 10: 1396-1401.

45

CHAPTER 2

Field Report

46

Index

Abstract...........................................................................................................................47

1 Introduction ...................................................................................................... 48 1.1 Goals .....................................................................................................................48 1.2 Collaboration ..........................................................................................................48 1.3 Study areas ............................................................................................................48 1.4 Timeplan ................................................................................................................50

2 Materials and methods ...................................................................................... 51 2.1 Plant collecting .......................................................................................................51

a. Collections....................................................................................................................................... 51 b. DNA ................................................................................................................................................ 51 c. Samples for chemical analysis........................................................................................................... 51

2.2 Butterfly collecting ..................................................................................................51 a. Collections....................................................................................................................................... 51 b. DNA. ............................................................................................................................................... 51

3 Results............................................................................................................. 52 3.1 Rinorea species found .............................................................................................52 Biogeography............................................................................................................................. 52 Taxonomic groups ...................................................................................................................... 52

3.2 Cymothoe species found..........................................................................................55 Taxonomic groups ...................................................................................................................... 55 Cymothoe behaviour ................................................................................................................... 55 a. Adult feeding.................................................................................................................................... 55 b. Larval feeding .................................................................................................................................. 55 c. Host=plant selection and oviposition ................................................................................................... 57 d. Mating & territoriality ........................................................................................................................ 58

3.3 Relationships found .................................................................................................59

4 Discussion........................................................................................................ 61

5 Conclusion ....................................................................................................... 62

6 Acknowledgements ........................................................................................... 63

7 References....................................................................................................... 64

47

Abstract

This field report describes the expedition to Cameroon in April and May 2006, organised by the NATIONAAL HERBARIUM NEDERLAND – WAGENINGEN BRANCH & BIOSYSTEMATICS GROUP, in close collaboration with the HERBIER NATIONAL in Yaoundé, Cameroon. The expedition aimed at evaluating the feasibility of studying Cymothoe butterflies and their associated host plants in the field. Its main goals were to find previously described Rinorea and Cymothoe populations, study the host plant selection and oviposition behaviour of Cymothoe females, examine which Cymothoe species colonise which Rinorea species and sample chemicals of the leaf surface that might be used as recognition template by Cymothoe females ovipositing on Rinorea.

In the limited time span of two months we were able to find populations of Rinorea and Cymothoe as described by Amiet in 1996 and have collected 32 species of Rinorea and 11 species of Cymothoe. With this we have made a good start for a reference collection, in which the main taxonomic groups of both genera are represented. All the observed host plant relationships were consistent with those previously described, indicating that these data are reliable. Host plant relationships have been found for all six Rinorea-feeding Cymothoe species that we have collected. Chemical samples have been made for 29 Rinorea species. In the light of these results, the field expedition has been a successful one.

An expansion of the fieldwork to other regions in Africa, and a study of the biogeography of both Rinorea and Cymothoe based on previous collections are recommended.

48

9 Introduction

This is the field report of an expedition to Cameroon that took place from April 7 to June 9 2006. This expedition was conducted for my thesis project about the relationships between Rinorea (Violaceae) and Cymothoe (Nymphalidae, Limenitidinae) species in the African tropical forests. This thesis project was part of my biology studies at the Wageningen University, and organised by the NATIONAAL HERBARIUM NEDERLAND – WAGENINGEN BRANCH & BIOSYSTEMATICS GROUP, in close collaboration with the HERBIER NATIONAL in Yaoundé, Cameroon. This thesis was meant to be a preparative and orientating pilot for the planned PhD project “Evolution of host-specificity in the tropical African butterfly genus Cymothoe (Nymphalidae) feeding on Rinorea (Violaceae)” for which an application has been submitted to NWO-ALW in august 2006. For a general introduction to the subject, I would like to refer to the thesis report in chapter 1.

9.1 Goals

The main goals for the expedition were to:

� Find (a part of) the Rinorea and Cymothoe populations as described by Amiet (1996; 1997; 2000) � Study the host plant selection and oviposition behaviour of Cymothoe females; how quickly are

host-plants recognised? � Examine which Cymothoe species colonise which Rinorea species. � Sample chemicals of the leaf surface that might be used as recognition template by ovipositing

Cymothoe females, of as many Rinorea species as possible.

9.2 Collaboration

The expedition was carried out in collaboration with the HERBIER NATIONAL in Yaoundé, Cameroon; under the guidance of Dr Jean Michel Onana and part of the INSTITUTE DE RECHERCHE AGRICOLE POUR LE DÉVELOPPEMENT (IRAD). Nearly all fieldwork was carried out together with Olivier Valery Sene Belinga MSc, under supervision of Dr Gaston Achoundong, who is expert on African Rinorea and has also participated in a part of the fieldwork.

The INSTITUTE DE RECHERCHE POUR LE DÉVELOPPEMENT (IRD) in Cameroun has provided me with a room in Yaoundé. The IRAD research station in Kribi has kindly lend us a car for a few days of collecting.

9.3 Study areas

Cameroon has been chosen as a destination for several reasons. First, the only previous studies on the host plant relationships of Cymothoe were also conducted in Cameroon (Amiet 1996). Thus, we had a fairly good impression of which species and associations we could expect. Second, also for Rinorea the taxonomy is in good condition only for Cameroon, thanks to the work of Dr. Achoundong with whom the Nationaal Herbarium Nederland in Wageningen has a very good relation.

Within Cameroon we have worked in three main areas (figure 1 and 4): Yaoundé (Central province), Edéa (Littoral province), and Kribi (South province). The Yaoundé area can be considered as part of the forest plateau, which has a red ferralitic soil on acidic eruptive rock, and a congolean and semi-deciduous forest type. The Edéa and Kribi areas lay within the littoral plain, having yellow ferralitic soils on sedimentary rock, and an Atlantic forest type (Letouzey, 1968; Achoundong, 1996).

49

FIGURE 1. The three main study areas in Cameroon with towns in red and

field localities in blue. For a detailed view see figure 4.

All three areas lay within the equatorial climate regime, which has four distinct seasons: two rainy and two dry (Letouzey, 1968). The main monthly temperature has its maximum in March-April and its minimum in July-August (figure 2). Yaoundé is relatively the cooler, drier area, while Edéa is the hottest and Kribi the wettest area (figures 2and 3), mainly influenced by the proximity of the ocean (Letouzey, 1968).

FIGURE 2. Annual climatic changes in the three main study areas. Average monthly rainfall (left) and temperature (right), data taken from Letouzey (1968) is concordant with data available on www.worldclimate.com.

3046

2601

1529

0

500

1000

1500

2000

2500

3000

3500

Yaoundé Edéa Kribi

Me

an

An

nu

al R

ain

fall

(mm

)

25.8

26.5

23.4

20

22

24

26

28

Yaoundé Edéa Kribi

Me

an

An

nu

al T

em

pe

ratu

re (

°C)

FIGURE 3. Mean annual rainfall (left) and temperature (right) for the three main study areas, data taken from Letouzey (1968) is concordant with data available on www.worldclimate.com.

20

22

24

26

28

30

jan feb mar ap r may jun jul aug sep oct nov dec

Ave

rag

e T

em

pe

ratu

re (

°C)

Yaoundé

Edéa

Kribi

0

100

200

300

400

500

600

jan feb mar apr may jun jul aug sep oct nov dec

Ave

rag

e R

ain

fal (

mm

) Yaoundé

Edéa

Kribi

50

FIGURE 4. The studied localities in Cameroon. Red circles are large towns, red houses are the towns of the three main study areas. White circles are field localities, with roads along which we have collected marked with red outlines. Maps taken from the Google Earth website http://maps.google.com.

9.4 Timeplan

The expedition took place from April to early June 2006, which is during the start of the first rainy season and the temperature maximum (see figure 2). There were a total of 9 weeks (63 days) in Cameroon, of which 43 days (70%) were spent working in the field. During 2 of these 43 days I was unable to be in the field due to illness, but Olivier was able to continue the work. During three other days I have accompanied Dr. Achoundong during the fieldwork for an environmental impact assessment of the Alucam aluminium factory in Edéa. In the end, 18 field days were dedicated to the Yaoundé area, and the Edéa and Kribi areas each received 11 days of field work (table 1). Most of the remaining days were spent on preparative work such as shopping, consultation, drying plant material etc. Other days were lost due to travelling or heavy rain. Province Department Area Locality lat long Field Days Total Days

Mont Kala N 03°50.23' E11°20.86' 11 Central province

Mfoundi Yaoundé

Mont Elounden N 03°49.83' E11°26.84' 7 18

Malimba N 03°49.73' E10°06.73' 2

Ile Eding N 03°48.96' E10°08.16' 3

Littoral province

Sanaga Maritime Edéa

Ducam Duclair N 03°55.10' E10°04.29' 6

11

Londji 2 N 03°04.96' E09°59.26' Kribi-Edéa road

Bebambwe 2 N 03°03.51' E09°59.26'

6

Kribi-Lolodorf road Bissiang N 02°59.00' E09°59.60' 2

Kribi-Ebolowa road Djabilobe N 02°48.50' E10°20.67' 1

Mbongo N 03°24.97' E10°08.74'

Bella N 03°15.70' E10°11.75'

Elokbatindi-Bipindi road

Nkolo N 03°14.49' E10°13.35'

1

South Province

Océan Kribi

Kribi-Campo road Lobe 1 N 02°52.6' E09°52.6' 1

11

TABLE 1. Localities in the three main study areas with coordinates and the number of days in the field.

51

10 Materials and methods

10.1 Plant collecting

c. Collections Branches with leaves and preferably flowers and/or fruits were pressed between papers in a plant press (figure 1.13). Later they were heat-dried above a stove. Some flowering branches were collected and stored in 90% alcohol to preserve their three-dimensional shape.

d. DNA For all collections one or two leaves were stored in plastic zip bags with silica gel to facilitate future DNA extraction.

e. Samples for chemical analysis For some representative individual plants, leaves for chemical analysis were also pressed and dried: approximately 30-60 mg dry weight per individual. In order to sample the surface chemicals of the leaves, 8 alcohol swabs were taken from 4 leaves of different age, by firmly rubbing the upper surface with a cotton swab (Sarstedt AG & Co, Germany 80.625) saturated with 96% ethanol (figure 1.1).

10.2 Butterfly collecting

a. Collections Butterflies were caught using hand nets (40 cm diameter, black, Vermandel, Netherlands 60.332; figure 1.10) or sometimes using a butterfly fruit trap (Vermandel, Netherlands 60.702) with fermenting bananas. After capture, they were pinch-killed or killed in a jar with plaster with saturated ethyl acetate, and preserved in 52x77 mm envelopes on silica gel.

b. DNA. Of all individuals of Cymothoe collected, two legs were preserved in 90% ethanol for future DNA extraction. Eggs and caterpillars were also preserved in 90% ethanol, the caterpillars only after being photographed. Unfortunately, after the Cymothoe material (eggs, caterpillars and legs) had been used for DNA extraction and amplification at the Canadian Centre for DNA barcoding in Guelph, we discovered that the alcohol used for the preservation of this material was contaminated with methanol. The alcohol was bought in a store in Yaoundé without sufficiently checking its exact contents. This had caused serious deterioration of the DNA and as a result many eggs and caterpillars were lost for DNA analysis. For the butterflies, legs dried on silica gel are still available.

52

11 Results

11.1 Rinorea species found

We have collected 66 specimens of Rinorea, belonging to 32 different species (table 2), which is more than half of all the 55 species occurring in Cameroon (Achoundong, 1996).

Biogeography

In the Yaoundé area, we have found 12 species (table 2). At Mont Kala, these were mainly forest species occurring in the littoral plain as well as the congolean forest (e.g. R. longicuspis) or species with a very wide afrotropial distribution such as R. angustifolia, R. dentata and R. oblongifolia (figure 1.14; Achoundong, 1996). The species composition of Mont Elounden was very different: we have not found any Rinorea species that also occurred at Mont Kala. This locality was characterised by species that are associated with semi-deciduous forests: R. batesii, R. ilicifolia, R. yaundensis and R. zenkeri

(Achoundong, 1996). The first two are species with very thick, leathery leaves (sclerophyllous, figure 1.11), the latter two have smaller more papery leaves and are strongly related to each other. Other, more widespread species that we found were R. angustifolia and R. welwitschii. We collected a total number of 7 species.

We found 11 species in the area around Edéa. Two of these are in common with Mont Kala (R.

longicuspis and R. oblongifolia). Three other species we found are restricted to the littoral plain, of which Edéa is the northernmost part (Achoundong, 1996), such as R. amietii (figure 1.7-8), R. caudata and R.

gabunensis (figure 1.18). The Kribi area has a species composition which is similar but more diverse, with 16 species in total.

There are again two species in common with Mont Kala (R. dentata and R. longicuspis). The number of littoral plain species is higher, comprising R. amietii, R. breviracemosa, R. kamerunensis and R. mezilii

(figure 1.15), with additional species that are restricted to the southern part of the littoral plain (R.

albidiflora, R. ovata, R. umbricola and R. verrucosa). This high diversity is consistent with the reports by Achoundong (1996) who reported a total diversity of 16 Rinorea species in Edéa, and 33 in the Campo-Bipindi region southeast of Kribi.

Taxonomic groups

The Cameroonian Rinorea can be divided into two main supergroups (I and II), based on androecium symmetry and inflorescence type (Achoundong, 1997). These supergroups each contain two groups A and B with differences in the staminal tube, which are further subdivided into a total of eight subgroups (table 3). Our collections contain species of seven of those eight subgroups, with coverage of at least 50% of the species within each of those seven subgroups. The only group missing in our collections is subgroup IB1 with 3 species. (table 3) This means that, assuming that this classification is based on characters identifying natural groups, nearly all main evolutionary lines are represented.

53

Genus↓ species ↓ ↓ A

rea

Ya

ou

nd

é

Ed

éa

Krib

i

albidiflora � 1

amietii � � 2

angustifolia � 1

batesii � 1

breviracemosa � 1

campoensis � 1

caudata � 1

dentata � � 2

dewildei � 1

dewitii � 1

gabunensis � 1

ilicifolia � 1

kamerunensis � 1

ledermannii � 1

leiophylla � � 2

lepidobotrys � 1

letouzeyi � 1

longicuspis � � � 3

longisepala � 1

mezilii � � 2

oblongifolia � � 2

ovata � 1

preussii � 1

rubrotincta � 1

simonae � 1

subintegrifolia � 1

umbricola � 1

verrucosa � 1

welwitschii � 1

woermanniana � 1

yaundensis � 1

Rin

ore

a

zenkeri � 1

32 12 11 16 TABLE 2. Rinorea species found in the three main study areas

Supergroup Group Subgroup species IA1 R. angustifolia*

R. lepidobotrys*

R. breviracemosa*

R. sp. 1

IA IA2

R. sp. 2

R. convallarioides

R. microglossa IB1

R. ebolowensis

R. caudata*

I

IB

IB2 R. albidiflora*

R. ilicifolia*

R. dimakoensis

R. dewitii*

R. simonei

R. letouzeyi*

R. mezilii*

R. keayi

IIA1

R. crassifolia

R. kamerunensis*

R. gabunensis*

R. spongiocarpa

R. batesii*

R. ovata*

R. rubrotincta*

R. vivienii

IIA2

R. preussii*

R. welwitschii*

R. longicuspis*

IIA

IIA3

R. leiophylla*

R. subintegrifolia*

R. bosii

R. faurei

R. fausteana

R. campoensis*

R. sinuata

R. verrucosa

R. dentata*

R. yaundensis*

R. zenkeri*

R. longisepala*

R. ledermannii*

R. gilletii

R. villiersii

R. umbricola*

R. exappendiculata

R. dewildei*

R. amietii*

R. subsessilis*

R. woermanniana

R. dichroa

II

IIB IIB1

R. oblongifolia* TABLE 3. Rinorea classification according to Achoundong (1997)

54

FIGURE 5. 1. making swabs of surface chemicals; 2. Mont Elounden; 3. Olivier Sene Belinga; 4. Mballa Joseph climbing in R. oblongifolia; 5. inflorescence R.longisepala; 6. Marcel Melingui Engama and Olivier Sene Belinga; 7. R. amietii fruits; 8. R. amietii inflorescence RVV087; 9. children at mont Kala; 10. with Gaston Achoundong; 11. R. dewitii leaves; 12. Olivier Sene Belinga and Ndtongo Etienne; 13. Olivier Sene Belinga pressing plants; 14. R. oblongifolia fruits and inflorescence RVV040; 15. R. mezilii inflorescence; 16. with Olivier Sene Belinga; 17. Gaston Achoundong; 18. R. gabunensis inflorescence; 19. Olivier Sene Belinga pressing plant collections at the Herbier National in Yaoundé; 20. R. subintegrifolia fruits; 21.Menguele Roger.

55

11.2 Cymothoe species found

We have collected 53 specimens of Cymothoe butterflies, belonging to 11 species, and twelve specimens of its monospecific sister genus Harma (figure 2.7). An additional Cymothoe colmanti was only found as a caterpillar, adding up to a total of 13 species (table 4).

Taxonomic groups

Amiet (2000) has divided the species of Cymothoe in 10 groups containing 18 lineages, based on both adult and larval morphology. The lineages can be grouped by the host plants used; being either exclusively members of the Achariaceae or Rinorea (Violaceae), or by the ground colour of the male, being white, yellow or red. Harma, the monospecific sister genus of Cymothoe, is also included in this classification. For an overview of the lineages, see table 5. Our collections of 13 species represent only 31% of all 42 species in Cameroon, but they cover 80% of the groups, and 66% of the lineages. This is because we have collected 1 species per lineage; only the group of C. coccinata (COC) is represented by 2 species.

Cymothoe behaviour

Cymothoe species are real forest butterflies; during our expedition they were always found in or on the border of forests. Nevertheless many species seem to withstand forest degradation and remain numerous even in newly cultivated understorey, as long as their host plants are still available, as we have observed on Mont Elounden where Musa (bananas) and Xanthosoma sagittigolium (Cocoyam or Macabo) were cultivated.

Their flight is rather directional, and they interrupt their wing beats regularly for a moment of gliding - comparable to the flight of the red admiral Vanessa atalanta (Nymphalidae, Nymphalinae) - paying tribute to their English vernacular name Gliders. In general, as with most forest butterflies, they can be found at sunlit clearings within the forest.

a. Adult feeding In the Eding locality near Edéa I observed a male Harma theobene (figure 2.7) feeding on the yellow flowers of Rinorea oblongifolia (figure 1.14). This is remarkable because this species oviposits exclusively on several Achariaceae species, never on Rinorea. This indicates that adult feeding and oviposition behaviour may have different selection mechanisms. Rinorea oblongifolia was one of the main flowering trees at that particular locality and time, and it seems plausible that adult feeding is much less discriminative. In the Ducam-Duclair locality, we observed C. indamora females three times feeding on an abundant flowering cauliflorous Maesobotrys klaineana (Euphorbiaceae) tree (figure 2.8-8). Another Cymothoe species was found feeding on the yellow flowers with large white bracts of a Mussaenda species (Rubiaceae; figure 2.12).

b. Larval feeding Cymothoe caterpillars were always found on the upper side of the leaves, or along the stems of their host plant. However, even though we did check the under side of the leaves of (potential) host plants, it cannot be ruled out that we missed some caterpillars there, as the under sides are not readily visible. But in any case, it seems that the caterpillars are not taking much effort to hide from predators, but rely on their dorsal spines to act as a deterrent. Indeed, when the first caterpillar was found, our

56

young guide did not dare to touch it because he believed it would have a poisonous sting. When they are being disturbed, some caterpillars raise their head and thorax and make hostile movements. I have never noticed any actual stinging or irritation due to these spines.

It appears that the caterpillars eat the leaves starting from the edges, moving inward (figure 2.5). In the end leaves can be consumed almost entirely: only the midrib remains. In adult trees or shrubs, herbivory damage is not very significant, but for young trees (up to around 0.5 m) defoliation can be very severe: up to 90% of the leaves, possibly more.

Genus↓ species ↓ ↓ A

rea

Ya

ou

nd

é

Ed

éa

Krib

i

beckeri � � 2

caenis � � 2

coccinata � 1

colmanti* � 1

egesta � � � 3

fumana � � � 3

herminia � 1

indamora � 1

jodutta � � 2

oemilius � � 2

reginaeelisabethae � 1

Cym

oth

oe

sangaris � 1

Harma theobene � � � 3

13 9 7 7 Table 4. Cymothoe and Harma species found in the three main study areas.

Host Plant Use Colour Group Lineage species - 1 THE H. theobene*

- 2 OEM C. oemilius*

C. alticola

C. caenis*

C. alticola

C. caprina

C. amenides

C. consanguis

CAE

C. capella

C. indamora* IND

C. zenkeri

C. herminia*

3

HER C. weymeri

C. jodutta* 4 JOD

C. altisidora 5 BEC C. beckeri*

C. hyarbita

Achariaceae

6 HYA C. reinholdi

white

7 HAR C. harmilla 8 FUM C. fumana*

C. heliada HEL

C. fontainei CYC C. cyclades

C. egesta* EGE

C. orphnina LUC C. lucasii

C. lurida

C. hesiodotus

C. hypatha

yellow 9

LUR

C. colmanti*

C. "sangaris"* SAN

C. euthalioides OGO C. ogova

C. coccinata*

C. crocea

C. reginaeelisabethae*

C. excelsa

C. arcuata

C. haimodia

C. preussi

C. aramis

Rinorea

red 10

COC

C. distincta TABLE 5. Cymothoe classification according to Amiet (2000)

57

c. Host-plant selection and oviposition Actual host plant selection is done by the female when she chooses the plants on which to lay her eggs (Amiet 1996, personal observation). When looking for a host plant, the C. sangaris and C. egesta females we observed were flying very low: at around 30 cm above the ground. This is in contrast to their normal flight range from the forest floor to the canopy, over 15 meters high. They landed preferentially on the upper side of leaves of small Rinorea species less than 0.30 m tall, but other plant species would also be visited. After landing, a very brief contact appears to be sufficient to identify the plant, as non-host species are mostly rejected in an instance: less than 0.5 seconds after landing.

When a host plant species was encountered, the female would curl her abdomen to reach the under side of the leaf, where a single egg would be laid. This behavioural sequence takes up to about 2 seconds, after which it would take off and continue its flight. Not all individuals of their respective host plant species (Rinorea batesii for C. sangaris and R. mezilii for C. egesta) that were visited were accepted; in general the females would choose plants with relatively young leaves of small plants for oviposition.

For the Achariaceae feeding species Harma theobene, we observed oviposition behaviour several times at the Elounden locality near Yaoundé (figure 1.2). Here it would oviposit on the locally abundant Lindackeria schweinfurthii (Achariaceae; figure 2.3). Searching behaviour was about the same; although the flight was generally higher (up till 80 cm), it would remain relatively low, and prefer to visit plants that were relatively small and unexposed. But its oviposition behaviour was slightly different from that of the two Rinorea feeding species described above. When it would land on a L.

schweinfurthii leaf, it would briefly drum the leaf surface with its legs. This could be noticed because this particular Lindackeria species has very thick, tough leaves (sclerophyllous), and the drumming was clearly audible. It can therefore not be excluded that Cymothoe species also have this drumming behaviour, but it might be less easy to observe.

The main difference lies in the behaviour after recognition of the host plant. Where the two observed Cymothoe species would lay an egg on the under side of the recognised leaf itself, Harma

theobene would always take off, and fly -with its abdomen curled to the front- to oviposit one egg on the under side of a neighbouring leaf. Thus, the egg is not laid on, but in the vicinity of the recognised host plant leaf. An important observation is that this neighbouring leaf can also belong to an entirely different species. I even observed oviposition on the stem of a herbaceous climber next to a L.

schweinfurthii plant. This behaviour can be seen as disadvantageous for the offspring, because the newly hatched larvae will first have to find the appropriate host-plant, but as their host will be in close proximity, this may not be a big problem. Previous experiments by Amiet (1996) have shown that larvae can indeed discriminate their host plants, and move directionally to it. Another explanation is that less specific oviposition may prevent parasitisation of the eggs. It is known from several parasites find their prey by locating its host plant (e.g. Kaiser and Carde, 1992; Hilker and Meiners, 2006).

For C. fumana it seems that females have an oviposition preference for sunlit, young leaves. At a small stream in the forest at the foot of Mont Kala near Yaoundé, there was one particular old Rinorea

oblongifolia tree that had fallen and produced new shoots with large, light green leaves that were in the sunshine for most of the midday. We found many small caterpillars on these leaves and at different days during our study, indicating that females had a strong oviposition preference for these leaves. We observed an ovipositing C. fumana once, at this same location and have been able to film a part of the host plant searching behaviour. She systematically visited many leaves, briefly touching them as all the other observed species did. These leaves were up to 2 meters above the ground so she did not seem to have the preference for leaves that were just above ground like the females of C. egesta and C.

sangaris. Maybe this is also in relation to the habit of their host species, as R. oblongifolia is in general larger than R. dewittii or R. batesii (figure 1.4).

58

d. Mating & territoriality Cymothoe males are very territorial. They can usually be found basking on a sunlit leaf, around 5-10 m above the ground. As soon as a conspecific male (or any other butterfly remotely resembling one) appears, they will approach it at full speed, fight and chase it away before returning to the same spot. I have seen them successfully chase off butterflies twice their own size. As a result of this aggressiveness, many collected Cymothoe males show severe damage to their wings.

Males choose their territory such that it contains host plants. Thus, they can wait for a female approaching, in search for a suitable oviposition site. I observed mating behaviour once, at the stream at Mont Kala mentioned above. I was observing the C. fumana female visiting host plant leaves, when a male that had been perched 3 m higher flew down to her at full speed. Immediately after their encounter they flew together, spiralling fast around each other up to the forest canopy. This behavioural sequence lasted less than one second.

FIGURE 6. 1. C. beckeri male RVV B283; 2. Mballa Joseph and Olivier Sene Belinga; 3. Lindackeria schweinfürthii; 4. unidentified caterpillar RVV C017 on R. longisepala; 5. Cymothoe caterpillar RVV C022 on R. yaundensis RVV055; 6. C. egesa confusa egg RVV E012 on R. dewitii; 7. Harma theobene male; 8. Maesobotrys klaineana; 9. C. indamora RVV B319; 10. on bridge to Eding Island; 11. C. caenis female RVV B384; 12. Mussaenda sp.; 13. C. egesta "orange"RVV C03 on R. lepidobotrys; 14. C. caenis male RVV B364.

59

11.3 Relationships found

We have found seven distinct host plant relationships between Cymothoe species and Rinorea, that all confirmed the data as presented by Amiet in 1996 (see table 6). Almost all relationships were observed in the Yaoundé area.

Rinorea ↓ ↓ C

ymo

tho

e a

ram

is

arc

ua

ta

co

ccin

ata

co

lma

nti

cro

cea

cyc

lad

es

dis

tinct

a

eg

est

a c

on

fusa

eg

est

a "o

ran

ge

"

exc

els

a

fu

ma

na

ha

imo

dia

ha

rmill

a h

arm

illa

he

liad

a h

elia

da

he

sio

do

tus

hyp

ath

a

fo

nta

ne

i

luca

sii

luri

da

nig

eri

en

sis

og

ova

orp

hn

ina

pre

uss

i

re

gin

ae

elis

ab

eth

ae

sa

ng

ari

s I 1

sa

ng

ari

s I 2

sa

ng

ari

s II 1

sa

ng

ari

s II 2

amietii ���� ���� 2

angustifolia ���� ���� 2

batesii ���� ���� 2

campoensis ���� 1

caudata ���� 1

convallarioides ���� 1

dentata ���� ���� ���� 3

dewildei ���� 1

dewitii ���� 1

dimakoensis ���� 1

gabunensis ���� 1

ilicifolia ���� 1

keayi ���� ���� 2

ledermannii ���� ���� 2

lepidobotrys ���� 1

letouzeyi ���� 1

longicuspis ���� ���� ���� 3

longisepala ���� ���� 2

mezilii ���� 1

oblongifolia ���� ���� 2

preussii ���� 1

rubrotincta ���� ���� 2

sinuata ���� 1

spongiocarpa ���� 1

subsessilis ���� ���� ���� 3

verrucosa ���� 1

vivienii ���� ���� 2

welwitschii ���� ���� ���� 3

yaundensis ���� ���� 2

zenkeri ���� 1

1 1 3 1 1 1 2 6 1 1 5 2 1 1 3 1 1 1 1 2 1 2 1 5 1 1 1 TABLE 6. Host plant relationships of Cymothoe feeding on Rinorea in Cameroon, redrawn from Amiet and Achoundong (1996). Observed relationships during this expedition are marked with orange; relationships deduced from syntopy are marked in green.

At the foot of Mont Kala, caterpillars in various stages of C. fumana have been found on different individuals of R. oblongifolia. We also observed host plant searching behaviour and male territoriality of this species involving a stand of this host plant species.

In that same area we found two 5th instar caterpillars of C. egesta on two different trees of R.

lepidobotrys belonging to one dense population (figure 2.13). On the summit of Mont Kala, we have seen C. sangaris butterflies that were clearly associated with

the population of R. preussi occurring there. A pupa was found on one of these trees, which gave rise to a blood red C. sangaris male 10 days later.

60

At Mont Elounden I observed one C. sangaris female ovipositing on two very small R. batesii individuals. There was also a small caterpillar found on R. yaundensis, but we have not been able to identify it (figure 2.5).

Another R. batesii besides the path, hosted two 5th instar caterpillars: belonging to C. sangaris and C.

colmanti. The plant was almost completely consumed, having only two intact leaves. At another locality, more on the south side of Mont Elounden, a 5th instar caterpillar of C. egesta

was found on R. ilicifolia. For this same species we also observed oviposition behaviour and found eggs at the other two

areas Edéa and Kribi, but this time in relation to R. mezilii and R. dewitii, other members of the Ilicifoliae (subgroup IIA1 in table 3; see also figure 2.6).

All our host plant relationship data for Harma theobene come from the Mont Elounden locality (figure 1.2). We have observed oviposition and found eggs on and directly next to individuals of Lindackeria schweinfurthii (figure 2.3), member of the Achariaceae which is the host plant family for this species. At Mont Kala we also found a pupa of H. theobene on a small treelet of a species belonging to the Olacaceae. Members of this family are not reported as host plants for this species, and we are not confident as to whether the caterpillar also had been feeding on this treelet. Unfortunately, we were unable to verify if there were any Achariaceae standing near because that forest patch was cut for cultivation before our next visit. When we also take into account the syntopy (co-occurrence at one site) of certain butterflies and their host species, we can conclude that the host plant relationship between C. coccinata and C.

reginaeelisabethae with R. yaundensis, as described by Amiet (1996) is also reliable. Indeed, both species were seen perched on the cultivated Musa, and caught flying at a site where R. yaundensis was relatively abundant. However, this was the same site where oviposition of C. sangaris on R. batesii was observed, so we cannot exclude that C. coccinata and C. reginaeelisabethae are ovipositing on R. batesii. But at least the syntopy can be explained by the above mentioned relationship with R. yaundensis.

61

12 Discussion

Although the relationships between species of Rinorea and species of Cymothoe are very strict, they are not interdependent. Cymothoe butterflies need their host plants for survival, but the Rinorea species do not need the butterflies, because they do not play a key role in their pollination. Therefore, locating a population of Rinorea does not guarantee the discovery of Cymothoe butterflies. Because we collaborated mainly with the expert of Rinorea, Dr. G. Achoundong, we visited localities with a known high diversity of Rinorea. The idea was that at those localities we would have a higher chance of finding associated Cymothoe. It appears, however, that we have maximised only the chance of finding many Rinorea species, as can be seen in the number of collected species for Rinorea (32) and Cymothoe (11, of which 6 feeding on Achariaceae).

As a preparation for a next field trip, I would recommend studying Cymothoe collections in order to know where populations of the Cymothoe species of interest occur(red). There are extensive collections of Cymothoe accessible at Tervuren, the British museum of Natural History and the museum of the African Butterflies Research Institute in Nairobi. With that information it will be possible to search the host plant(s) for that butterfly population. The fact that we have been able to find many of the populetions studied by Amiet a decade ago (1996; 1997) indicates that the populations are fairly stable. This may increase both efficiency for finding new species and the chance to find new associations.

Another point of discussion is the fact that we have visited mainly the same regions as Dr. J-L. Amiet. Although his studies were conducted more than a decade ago, it is imaginable that we found the same associations because we sampled in the same genetic pool. In order to reliably test the general validity of the host plant association data of Amiet (1996), it is necessary to visit other areas, both on a regional and a more pan-African scale. On the regional scale, it would be good to search for Cymothoe-Rinorea associations in other areas in Cameroon, or even in Gabon. Because the species composition of both Rinorea and Cymothoe is likely to be comparable throughout Central Africa, this will enable us to test whether the associations are constant in allopatric populations. The high species diversity of both genera in this region will maximise the success in the number of species found. On the pan-African scale, fieldwork in e.g. west and/or east Africa can provide even more convincing support for the general validity of the host plant associations. For the widespread Rinorea-dependent Cymothoe species, such as C. egesta, C. fumana, C. lurida, and the C. sangaris species complex, it will be highly interesting to see whether their host plant relationships in these parts of the continent are the same as for Cameroon. Additionally, we may be able to add relationship data of (endemic) species that were not studied before, filling possible gaps in our knowledge of Cymothoe and its host plant relationships.

In relation to both recommendations mentioned above, a study of the biogeography through the examination of collections of both Rinorea and Cymothoe is of prime importance. Additionally, such a study can shed light on a number of questions, such as

� Do the areas of distribution of Cymothoe species coincide with those of their proposed host plant(s), or do we need to postulate additional host plant relationships to explain their biogeography?

� Do strictly monophagous Cymothoe have a narrower distribution as compared to their oligophagous congenerics?

� Have Cymothoe species colonised the entire distribution area of their host plant(s), or are there (historical or ecological) factors limiting their distribution?

Knowledge of the biogeography can thus contribute to both the efficiency of future fieldwork and the understanding of the extent of the already known host plant relationships, as well as generate new questions and answers about this highly interesting insect-plant system.

62

13 Conclusion

It can be concluded that this field expedition has been a successful one. We were able to find populations of Rinorea and Cymothoe as described by Amiet in 1996. In the limited time span of two months we have collected 32 species of Rinorea and 11 species of Cymothoe. With this we have made a good start for a reference collection, in which the main taxonomic groups of both genera are represented. This also provides a good basis for the reconstruction of their respective phylogenies. But almost more importantly we have acquired hands-on experience with the identification as well as localisation of these species in the field. Our knowledge of the habitats and behaviour (in the case of Cymothoe) of the species ensures that future fieldwork can be even more efficient and effective.

Our observations of oviposition behaviour were still limited. This is in part due to the fact that our priorities were to collect many species for reference. Consequently, many female butterflies were caught before they could perform any oviposition behaviour to observe. Now that we can identify species better in the field, it will become possible to dedicate more effort to the observation of behaviour in the future. Even though the observations are limited, they were all consistent with the host plant relationships as described by Amiet (1996; 1997; 2000), indicating that these data are reliable. Host plant relationships have been found for all six Rinorea-feeding Cymothoe species that we have collected. In the case of C. egesta we have even found four different host plant relationships.

It is clear that oviposition and host plant searching behaviour is not identical in all species. We have observed deviations from the common pattern in both search flight and the actual oviposition behaviour of three species of Harma and Cymothoe. But it can be concluded that host plants are very quickly recognised and that drumming behaviour plays a role in the recognition process for at least one of the species.

Of all Rinorea species that we have found, chemical samples have been made: entire leaves as well as alcohol swabs of the upper leaf surface. These samples will be used in future studies to assess the chemical composition of Rinorea host plants, and characterise the compounds that are potentially used in host plant recognition (such as cyclopentenyl glycines; Clausen et al., 2002). This may lead to a better understanding of the mechanisms of host plant recognition by Cymothoe, and their role in the evolution of these plants and butterflies.

I recommend an expansion of the fieldwork to other regions in central Africa, and a study of the biogeography of both Rinorea and Cymothoe based on previous collections.

63

14 Acknowledgements

In addition to the acknowledgements in the thesis report, I want to thank the following people and institutes for their contribution to the success of this field expedition. First, I wish to thank Gaston Achoundong for his supervision during the expedition. He has been involved in the planning of the expedition, and has instructed me in Rinorea recognition in the field. Together with his wife Atiago Nzombeng Achoundong Fauste he has ensured that my stay in Cameroon was very pleasant and I thank them both for their kind hospitality. I also thank Jean Michel Onana, director of the HERBIER NATIONAL in Yaoundé, for his advice and support in terms of time and materials. I especially thank Olivier Sene Belinga who has been involved in nearly all the field work. He has proved to be a valuable plant taxonomist and biologist, and was excellent company. I owe much of the success of this expedition to him. I thank François Rivière and Mohammed Elomo Molo of the IRD for the wonderful room at their guesthouse in Yaoundé. The atmosphere in the guesthouse was very friendly, thanks to the international students and researchers staying there and the night watch Mohamed Saoud. Jean Folack, director of the IRAD field station in Kribi, has been so kind to lend us a car for a few days of collecting. For all the other days I thank the many taxi drivers of Kribi and Edéa. In the Youndé area, Adeng William has organised our transport to the localities under study. The local field guides have been of great help in finding suitable study localities. I want to mention Aboubakar (Edéa, Ducam-Duclair), Etienne Ndtongo (Kribi, Londji 2 and Bebambwe 2) and especially Mballa Joseph (Yaoundé, Kala) who was a passionate butterfly hunter and has collected many Cymothoe. During the preparation of the expedition, I have received good advice from many people, especially Jean-Louis Amiet, Rienk de Jong, Xander van de Burgt, Frans Breteler, Hans de Wilde, Freerk Molleman, Maaike de Jong, Carolyn McBride and my supervisors Freek Bakker and Joop van Loon. This field expedition could not have been carried out without the financial support by the following organisations, which I thank for their generosity:

• Hugo de Vries Foundation

• Stichting der Korinthiërs

• Wageningen University Travel Fund

• Alberta Mennega Foundation

• Biosystematics Fund

64

15 References

ACHOUNDONG, G. 1996. Les Rinorea comme indicateurs des grands types forestiers du Cameroun. In J.

v. d. Maesen [ed.], The Biodiversity of African Plants, 536-544. Kluwer Academic Publishers. ______. 1997. Rinorea du Cameroun, systématique, biologie, écologie, phytogéography., Université de

Yaoundé I, Yaoundé. AMIET, J.-L. 1997. Trophic specialization and early stages in the Cymothoe: Taxonomic involvement

(Lepidoptera, Nymphalidae). Bulletin de la Societe Entomologique de France 102: 15-29. ______. 2000. Les premiers etats des Cymothoe: Morphologie et interet phylogenique (Lepidoptera,

Nymphalidae). Bulletin de la Societe Entomologique de France 106: 349-390. AMIET, J.-L., AND G. ACHOUNDONG. 1996. An example of trophic specialization within the

Lepidoptera: The Cameroonian Cymothoe feeding on Rinorea (Violaceae) (Lepidoptera, Nymphalidae). Bulletin de la Societe Entomologique de France 101: 449-466.

CLAUSEN, V., K. FRYDENVANG, R. KOOPMANN, L. B. JORGENSEN, D. K. ABBIW, P. EKPE, AND J. W. JAROSZEWSKI. 2002. Plant analysis by butterflies: Occurrence of cyclopentenylglycines in Passifloraceae, Flacourtiaceae, and Turneraceae and discovery of the novel nonproteinogenic amino acid 2-(3 '-cyclopentenyl)glycine in Rinorea. Journal of Natural Products 65: 542-547.

HILKER, M., AND T. MEINERS. 2006. Early herbivore alert: Insect eggs induce plant defense. Journal of

Chemical Ecology 32: 1379-1397. KAISER, L., AND R. T. CARDE. 1992. In-Flight Orientation to Volatiles from the Plant-Host Complex in

Cotesia Rubecula (Hym, Braconidae) - Increased Sensitivity through Olfactory Experience. Physiological Entomology 17: 62-67.

LETOUZEY, R. 1968. Etude phytogéographique du Cameroun. Lechavelier, Paris.

65

Appendix: Network of contacts During the project, an informal, international network has been set up. A list of contacts related to the Cymothoe Rinorea research, in alphabetical order, can be found below. Of each contact the name, institute, email and postal address and telephone number –if available- is given, together with a short description of his / her specialisation relevant to the project.

Botany GASTON ACHOUNDONG Herbier National du Cameroun gachoundong@yahoo.fr Herbier National BP 1601 Yaoundé CAMEROUN 7660107 2314416 Specialist on African Rinorea. Identifier of all Rinorea specimens used. GREG WAHLERT Department of Environmental and Plant Biology Ohio University gw807105@ohio.edu 315 Porter Hall Ohio University Athens, OH 45701 USA 740 593.1126 PhD student working on Rinorea phylogenetics. Fieldwork in Madagascar. JEAN MICHEL ONANA Herbier National du Cameroun jmonana@yahoo.fr Herbier National BP 1601 Yaoundé CAMEROUN 2314416 Director of the National Herbarium in Yaoundé.

JEAN-PAUL GOGUE Herbier National du Cameroun jpghogue1062@yahoo.fr s/c Herbier National BP 1601 Yaoundé CAMEROUN PhD student on phytogeography, affiliated with the Institut for Systematic Botany in Switzerland. JERZY JAROSZEWSKI Danish university of Pharmaceutical Sciences Department of Medicinal Chemistry Natural products Research jj@dfuni.dk DENMARK (+45) 35 30 63 72 Professor. Specialist of cyanogenic glycosides in Flacourtiaceae and Rinorea. MAC ALFORD Department of Biological Sciences University of Southern Mississippi mac.alford@usm.edu Johnson Science Tower 308 118 College Drive #5018 Hattiesburg, MS 39406-0001 USA 601-266-6531 Assistant Professor and Curator of the Herbarium, specialist on Malpighiales phylogeny and chemistry.

66

MENGUELE MENGUELE ROGER s/c Eya Ebengue Gustave Ministere des Forets et de la Faune (MINFOF) direction des Forets BP 12489 Yaoundé CAMEROUN 9606656 7846298 Forest technician. Field assistant of Achoundong.

OLIVIER VALERY SENE BELINGA Herbier National du Cameroun olivier_sene@yahoo.fr s/c Herbier National BP 1601 Yaoundé CAMEROUN Plant taxonomy student. Field experience of Cymothoe and Rinorea in Cameroon.

Butterflies CAROLYN MCBRIDE Population Biology Grad. Group University of California Davis csmcbride@ucdavis.edu University of California Davis One Shields Avenue Davis, CA 95616 USA PhD student working on the molecular evolution of Cymothoe egesta. Field experience in Cameroon. FREERK MOLLEMAN Department of Entomology University of California Davis fmolleman@ucdavis.edu freerkmolleman@hotmail.com Office: 69 Briggs Hall One Shields Avenue Davis, CA 95616-8584 USA Makerere University Biological Field Station P.O.Box 409 Fort Portal UGANDA (530) 754-4872 Working on butterflies in Kibale National Park, Uganda. Also field experience in Cameroon.

GAËL VANDE WEGHE gvw19@hotmail.com GABON Working on the Butterfly Guide of Gabon. Field experience in Gabon, Guinée Equatorial and Congo Brazzaville. Is interested in sending Cymothoe legs for DNA analysis. JEAN-LOUIS AMIET 48, Rue des Souchères F – 26110 Nyons FRANCE Specialist on the host plant relationships of Cymothoe and other Limenitidinae. Field experience in Cameroon. MAAIKE DE JONG Evolutionary Biology dejongm@rulsfb.leidenuniv.nl Leiden Universiteit NETHERLANDS 071-5274882 PhD student working on Bicyclus anyana, field experience in Uganda and South Africa. MICHEL BEAURAIN direction@topduck.fr 5 rue Abel Gody ZI de la Boitardière 37400 Amboise FRANCE (0)247576363 Specialist on Cymothoe, field experience in Cameroon. Has identified photographs of Cymothoe specimens.

67

NIKLAS WAHBERG Laboratory of genetics University of Turku niklas.wahlberg@utu.fi Laboratory of Genetics Department of Biology University of Turku 20014 Turku FINLAND +46 8 164 024 Specialist on molecular systematics of Nymphalidae. RIENK DE JONG Naturalis Jong@naturalis.nnm.nl Naturalis Postbus 9517 2300 RA Leiden NETHERLANDS 071-5687652 Taxonomist of Hesperiidae with emphasis on Southeast Asia. Field experience in the tropics.

STEVE COLLINS African Butterfly Research Institute (ABRI) scollins@iconnect.co.ke No: 256 Dagoretti Road Karen, Nairobi KENYA P.O. Box 14308 884972 – 884973 0722 363 288 Director of the ABRI. Specialist on African Butterflies. TORBEN LARSEN torbenlarsen@compuserve.com 358 Coldharbour Lane London SW9 8PL UK (0)20 7737 1689 Author of the “Butterfly Guide of West Africa”. Field experience in Cameroon and West Africa. Has identified photographs of Cymothoe specimens and is interested in sending Cymothoe legs for DNA analysis.

DNA barcoding

GREGORY DOWNS Canadian Centre for DNA Barcoding gdowns@uoguelph.ca CANADA Network Manager, Programmer. JEREMY DEWAARD Canadian Centre for DNA Barcoding jdewaard@uoguelph.ca CANADA Laboratory Manager.

PAUL HEBERT Canadian Centre for DNA Barcoding phebert@uoguelph.ca Department of Integrative Biology University of Guelph 50 Stone Rd Guelph, ON N1G 2W1 CANADA +1 5198244120 Director. SUJEEVAN RATNASINGHAM Canadian Centre for DNA Barcoding sratnasi@uoguelph.ca CANADA Informatics Lead.

68

Cameroon JEAN FOLACK Institute de la Recherche Agricole pour le Développement (IRAD) folack@yahoo.fr j.folack@odinafrica.net IRAD Kribi BP 219 Kribi CAMEROUN 3461646 7761480 Director of the IRAD field station in Kribi. MARCEL MELINGUI ENGAMA s/c Dr. Edoo Edoo BP 55 Edéa CAMEROUN Chef of the village at the Ducam-Duclair locality near Edéa. MOHAMMED ELOMO MOLO Institute de la Recherche du Développement (IRD) elomo_molo@yahoo.fr IRD BP 1857 Yaoundé CAMEROUN 2201508 Staff member of the IRD office in Yaoundé.

NDTONGO ETIENNE (“Dodo”) S/c Nzié Nzié Daniel Infirmier à l’hopital d’Ebomé BP 637 Kribi CAMEROUN 6139673 Fieldguide at Londji 2 and Bebambwe 2 near Kribi. VINCENT EDJAMBA TOUGOULOU Ministère des Forets et de la Faune (MINFOF) direction des Forets BP 12489 Yaoundé CAMEROUN 7723249 2239231 Staff member of the MINFOF ministry in Yaoundé.