Thesis (MSc (Genetics))--University of Pretoria, 2005. / Includes summary. Includes bibliographical references. Available on the Internet via the World Wide Web.

Thesis (MSc)--Stellenbosch University, 2015. / ENGLISH ABSTRACT: Cuticular hydrocarbon (CHC) profiles of insects play roles in behavioural interactions within and between species, encompassing species-, colony- and mate-recognition. CHCs are largely genetically determined and are thus unique to each species, making them useful in chemotaxonomy. However, species exhibit intra-species variation in their CHC profile which can be the result of both intra-species genetic variation as well as environmental influences such as habitat effects, colony effects, diet, host switching, as well as adsorption of CHCs from other insects. Studies have found that the CHC profiles of a specific insect species will often exhibit variations between regions as well as the species of host the insect is associated with. Therefore, an ideal system to investigate the effects of genetic population structure and environment on the CHC profiles of insects is within the fig – fig wasp mutualism. Fig species occur in a wide variety of habitats and host a diverse complement of fig wasp species. We were therefore offered the opportunity to investigate a wide range of potential influences on fig wasp CHC profiles ranging from environmental to genetic effects. Firstly, through GC-MS we found that the CHC profiles of the fig wasps investigated are both species-specific and species-group-specific, with the species Elisabethiella glumosae, Elisabethiella stuckenbergi and Ceratosolen capensis, and two Otitesella species-groups (the Uluzi and Sesqui species-groups) separating out significantly. Consensus phylogenies (based on COI, Cytb and EF-1α) showed that within the galling fig wasp genus Otitesella there were multiple genetic lineages within a species-group which corresponds to species-level genetic variation, and that each genetic lineage was confined to a single host fig species. The CHC profiles reflected the genetic relationships between the two species-groups, and the CHC profiles within a species group could be differentiated by genetic lineage/host species. This indicated that although genetic lineage was mostly responsible for the observed variation in CHC profiles, factors associated with different host species also had an effect. Strong regional variation overriding both the influence of genetic lineage and factors associated with host species were observed in the CHC profiles of the fig wasps within a species-group. This regional variation in CHC profiles was also observed within two pollinating fig wasp species, Elisabethiella stuckenbergi and Ceratosolen capensis, which was not supported by population genetic data (COI and Cytb). In fact, very little genetic population structure was found within the pollinating species, even though the pollinators were collected across South Africa. The lack of genetic structure in pollinating fig wasps can be the result of high gene flow caused by the large dispersal capability of pollinating fig wasps. Our results indicated that fig wasp CHC profiles have the potential to be used in chemotaxonomy and are possibly used as species and mate-recognition cues by the fig wasps. Furthermore, we found both a regional and associated host species effect on the CHC profile. We suggest that the observed regional effect in this study could be attributed to habitat differences and differences in fig wasp community between regions. Moreover, the effect host species had on the CHC profiles may be as a result of dietary differences between galls in different host species. A possible consequence of the observed regional/host speciesassociated effect on fig wasp CHC profiles is that it could lead to pre-mating isolation within fig wasp species, which could ultimately result in speciation. In addition, our results indicated that the interpretation of the variation in the fig wasp CHC profile was dependent on the scale of the analysis: on a broad, inter-species-level scale, fig wasp CHC profiles were species-specific; on a finer intra-species scale, variation in CHC profiles occurred between fig wasps collected from different regions; and on a within-region scale, variation in CHC profiles within species-groups occurred between genetic lineages/host species. Future studies should look at the application of CHCs in chemotaxonomic studies on the fig wasp phylogeny, as well as the effect of fig wasp community composition on fig wasp CHCs. / AFRIKAANSE OPSOMMING: Kutikulêre koolwaterstof (KK) profiele van insekte speel rolle in die gedragsinteraksies binne sowel as tussen spesies, en behels die herkenning van spesieof kolonielidmaatskap asook potensiële maats. Kutikulêre koolwaterstowwe word meestal deur gene bepaal en is dus uniek vir elke spesie, wat dit handig maak vir chemotaksonomie. Spesies vertoon egter soms intraspesie variasie in hul KK profiele wat die gevolg kan wees van beide intraspesie genetiese variasie sowel as omgewingsinvloede soos habitat effekte, kolonie effekte, dieet, tussen-gasheer skuiwings, asook die adsorpsie van ander insekte se kutikulêre koolwaterstowwe. Studies het gevind dat die kutikulêre koolwaterstof profiele van ʼn spesifieke insek spesie op ʼn gereelde basis verskille vertoon tussen streke asook tussen die verskillende gasheer spesies waarmee die insek geassosieer is. Om hierdie redes is die vy – vy-wesp mutualisme ʼn ideale sisteem om die uitwerking van genetiese populasie struktuur en omgewing op die KK profiele van insekte te ondersoek. Vy spesies kom in ʼn wye verskeidenheid van habitatte voor en ondersteun ʼn diverse groep vy-wesp spesies. Dit het ons die geleentheid gebied om ʼn wye reeks moontlike invloede van vy-wesp KK profiele te ondersoek, van omgewings- tot genetiese invloede. Eerstens, deur die gebruik van GC-MS het ons gevind dat die KK profiele van die vy-wespe wat ondersoek was beide spesie-spesifiek en spesie-groep-spesifiek is, met die spesies Elisabethiella glumosae, Elisabethiella stuckenbergi en Ceratosolen capensis, asook twee Otitesella spesie-groepe (die Uluzi en Sesqui spesie-groepe) wat betekenisvol onderskei kon word. Konsensus filogenieë (gegrond op COI, Cytb en EF1-1α) het getoon dat daar in die gal-induserende vy-wesp genus Otitesella veelvuldige genetiese lyne binne die spesie-groepe voorgekom het ooreenstemmend met tussen-spesie genetiese variasie, en dat elke genetiese lyn beperk was tot ʼn enkele gasheer vy spesie. Die KK profiele het die genetiese verhoudings tussen die twee spesie-groepe weerspieël, en die KK profiele binne ʼn spesie-groep kon onderskei word op grond van hul genetiese lyn/gasheer spesie. Hierdie het getoon dat, alhoewel genetiese lyn meestal verantwoordelik was vir die waargeneemde variasie in KK profiele, faktore wat met verskille in gasheer spesies gepaard gaan ook ʼn effek gehad het. Sterk streeks-verbonde variasie wat beide die invloed van genetiese lyn, én faktore wat met verskille in gasheer spesie gepaard gaan, oortref het, was waargeneem in die KK profiele van die vy-wespe binne ʼn spesie-groep. Hierdie streeks-verbonde variasie in KK profiele was ook waargeneem in twee bestuiwende vy-wespe, Elisabethiella stuckenbergi en Ceratosolen capensis, ʼn resultaat wat nie ondersteun was deur die genetiese bevolkingsdata nie (COI en Cytb). In werklikheid was baie min genetiese bevolkings-struktuur opgespoor binne die bestuiwer spesies, selfs as was die bestuiwer spesies regoor Suid-Afrika ingesamel. Die tekort aan genetiese struktuur in die vy-wesp bestuiwers kan die gevolg wees van hoë geenvloei wat veroorsaak word deur die hoë verspreidingskapasiteit van bestuiwende vy-wespe. Die resultate toon aan dat vy-wesp KK profiele die potensiaal besit om in chemotaksonomie gebruik te word, en word moontlik deur vy-wespe gebruik as kenmerke vir die herkenning van spesie en potensiële maats. Verder was daar gevind dat daar beide ʼn streekseffek en ʼn effek geassosieer met gasheer spesie op KK profiele was. Ons stel voor dat die waargeneemde streekseffek in hierdie studie toegeskryf kan word aan verskille tussen habitatte asook streeksverbonde verskille tussen vy-wesp gemeenskappe. Boonop kan die effek wat gasheer spesie op die KK profiele gehad het ʼn gevolg wees van dieetverskille tussen die galle in verskillende gasheer spesies. ʼn Moontlike gevolg van die waargeneemde streeks/gasheer-spesie-geassosieerde effek op vy-wesp KK profiele is dat dit moontlik kon lei tot voor-paring-isolasie binne vy-wesp spesies, wat uiteindelik spesiasie kon veroorsaak het. Daarbenewens wys ons resultate dat die interpretasie van die variasie in die vy-wesp KK profiel was afhanklik van die skaal van die analise: op ʼn breë interspesie vlak was die vy-wesp KK profiele spesiespesifiek; op ʼn fyner intra-spesie vlak het variasie in KK profiele voorgekom tussen vy-wespe wat in verskillende streke ingesamel was; en op streeksvlak het variasie in die KK profiele binne spesie-groepe voorgekom tussen genetiese lyne/gasheer spesies. Toekomstige studies behoort te kyk na die toepassing van kutikulêre koolwaterstowwe in chemotaksonomiese studies van die vy-wesp filogenie, asook die effek wat vy-wesp gemeenskap samestelling het op vy-wesp kutikulêre koolwaterstowwe.

Male fig wasps (Hymenoptera: Chalcidoidea) exhibit a fascinating range of morphology and behaviour. A cluster analysis, based on descriptions of the males of several hundred species, distinguished six major morphological groups. Behaviourial observations suggest that male morphology is related to the levels of inter-male aggression. Three behaviourial groupings were identified. Fighting species generally mated in the fig cavity, pacifist species mated in the females' galls or outside the figs. Mating sites are thus the primary determinants of male morphology and behaviour. In fighting species males were larger than their females, whereas pacifists and aggressors were equal in size or smaller than conspecific females. The large males in fighting fig wasps appear to be a consequence of sexual selection because larger males tended to win fights. Within a species there were no differences in the size of the galls that produced males and females, even in species where sexual size differences were present, suggesting that there is a heritable component to wasp size. No alternative advantages for smaller males were detected. Although fights were sometimes fatal, damage was not always a consequence of fighting behaviour and was recorded in both fighting and pacifist species. Sex ratios in several species were more female biased at higher population densities. Sex ratios of species with 'internally' ovipositing species were heavily female biased, but approached 1:1 in more outbred species with 'externally' ovipositing females. Levels of matedness, among females ranged from 73% to 99%. No evidence for sperm exhaustion was obtained. Species of Philotrypesis with both winged and flightless males were present only in southern African Ficus species from subsections Platyphyllae and Chlamydodorae. No species had only winged males. The flightless males of some Philotrypesis species were themselves polymorphic. In one polymorphic Philotrypesis species, winged males were found to be rare at high densities, but common at low densities. Digitata and religiosa males of Otitesella differed in coloration, size and behaviour. Digitata males were aggressors while religiosa males were fighters. Digitata males escaped from the figs whereas religiosa males remained inside the figs, perhaps because only digitata males were attracted to Light. Proportionally more digitata than religiosa males were present in low density populations and females were found to respond differently to the two morphs.

The Sycoecinae are a distinct and well-defined subfamily of old world fig wasps (Agaonidae, Chalcidoidea 1 Hymenoptera) , exclusively associated with the figs of Ficus species (Moraceae). The most likely sister group of the Sycoecinae was determined to be the Sycoryctini (Sycoryctinae) based largely on synapomorphies of the underside of the head. 67 sycoecine species and 3 subspecies were recognised and included in a phylogenetic analysis of the subfamily. This analysis clearly delimited six genera (four African and two extra-African), although the phylogenetic relationships between the genera were not strongly supported and remain flexible. Comparisons of the phylogeny of the Sycoecinae with the classifications of the Agaoninae and their host fig trees (Ficus, Moraceae) suggest a degree of cospeciation sensu lato. Numerous homoplasies were detected within the Sycoecinae, some of which were shared with another group of fig wasps that also enter the fig to oviposit, the Agaoninae. The anatomy of the figs apparently provides strong selection pressures that have resulted in both parallelisms and convergences within and between the two subfamilies. Among the 67 species and 3 subspecies that were recognised, 43 species and 2 subspecies are described as new. The males of three previously recognised species are also described for the first time. One generic and two specific synonyms are established together with five new combinations. Keys are provided to the genera and species, for both sexes.

The fig–fig wasp interaction is a classic example of obligate mutualism and coevolution. It is also a nursery pollination mutualism and supports a diversity of exploiter/parasite/non-pollinating fig wasp species. Mutualists and exploiters comprise the fig wasp community. All the wasp community members are obligately dependent on the fig syconium (a globular closed structure comprising of hundreds to thousands of uniovulate florets) for completing their life cycle. The fig florets can be sessile (without a stalk) or pedicellate (stalked) and can support a community comprising 3–30 wasp species. Fig wasps can access the floral resources for oviposition directly by entering into the syconium (internal oviposition) or by penetration of the syconium surface (external oviposition). Most studies on the fig–fig wasp interaction have investigated the stability of the interaction, pollination biology, pollen dispersal, co-evolution or the effect of exploiters on this mutualism. However, studies dealing with community ecology and species coexistence mechanisms in these communities are rare. Factors contributing to coexistence of mutualists and exploiters in a fig wasp community were studied using a reasonably speciose fig wasp community associated with Ficus racemosa in south India. The wasp community of Ficus racemosa comprises a single species of pollinator and six species of exploiters; together they represent three genera of fig wasp species. The community members show differences in their feeding habit; they could be 1) gallers (feed on floral tissue after pollination and/or after inducing abnormal tissue development of the floret that is also called the gall), 2) inquilines (feed on gall tissue but cannot induce galling; survive by feeding on gall tissue and starving the host larva to death), or 3) parasitoids (lay eggs in or on developing offspring of a galler or inquiline species; develop by feeding on host tissue). Resource partitioning across temporal and spatial axes on this fig wasp community have been quantified. Ovipositor traits of each community member were also investigated since variation in ovipositor traits might facilitate resource partitioning. Finally, the role of life-history traits in species coexistence in this community was also explored. Temporal resource partitioning among members of the fig wasp community was studied (1) across the resource phenology, i.e. over the development phases of the fig syconium, and (2) on a diel scale. The seven members of the wasp community were found to partition their oviposition periods across fig syconium development phenology; some species used very young syconia (soft and smaller in size) for oviposition whereas others used mature (hard and bigger in size) syconia for oviposition. The first species to colonise the syconia were gallers and these were followed by parasitoids in a definite oviposition sequence. Pollinators arrived concurrently with an exploiter galler species and had the shortest oviposition window in terms of days. Although fig wasps are known to be largely diurnal, night oviposition in several fig wasp species was documented for the first time. Wasp species showed a peak in their activity period across the diel cycle and phenology. This is probably the first study to simultaneously investigate temporal partitioning across the syconium phenology as well as the diel scale in a fig wasp community. Partitioning of syconium space was investigated by quantifying the quality (type of floret—sessile or pedicellate) of floral resources. The number of individuals of each species developing in a syconium was quantified along with host accessibility during oviposition by each wasp species. The association between community members developing within a syconium was also tested. The differential occupancy of florets by each species based on their distance from the base of the syconium was evaluated. For the first time the relative distribution of males and females of the entire fig wasp community was quantified. The wasp community members used similar types of florets for oviposition. Seeds were found mostly in sessile florets and wasps were present in large numbers in pedicellate florets. Except for one wasp species, all others occurred uniformly within the syconium with respect to the distance from the base of the syconium. Species distribution models revealed higher prediction ability for the location of mutualists (seed and pollinator) within the syconium compared to exploiters. Within a syconium, all species pairs exhibited positive associations indicating either an absence of or low competitive exclusion. Some florets were modified by their gall occupants such that they were longer in length indicating the possibility of creation of an enemy-free zone by the gall occupant. Yet, most florets were accessible to ovipositing wasps based on ovipositor lengths and flexibility. The probability of finding a male decreased with increase in floret length when all wasp species were grouped together; however, this trend did not hold true when males and females of species were tested individually. Based on these results, the fig wasps of F. racemosa could be grouped into—(1) Early-arriving galler species which used immature florets, inducing large galls that protruded into the cavity, and with fewer individuals per syconium, (2) Galler species arriving concurrently with the pollinator, inducing galls that were morphologically indistinguishable from those of the pollinator, and with many individuals developing per fig syconium, and (3) Parasitoids and/or inquilines of the galler species, with variable abundance per syconium. Thus, these results show that the wasp species do not clearly partition floral resources between syconia and within syconium but they can modify their oviposition sites and also differ in the proportion of florets within a syconium used for oviposition. Oviposition sites of the fig wasps can be reached only by using their ovipositor. The resources for oviposition are hidden and hence might require tools for resource location and utilisation. The frequency and diversity of sensilla on the ovipositor, as well as ovipositor structure (morphology and sclerotisation of the tip) was documented for the entire wasp community. The internally-ovipositing pollinator had the simplest ovipositor, negligible sclerotisation and only one type of sensillum on its ovipositor; the externally ovipositing exploiter species had teeth on their ovipositors, sclerotisation and various types of ovipositor sensilla. Ovipositor sclerotisation and lateness of arrival for oviposition in syconium development were positively correlated. Ovipositor teeth height increased from gallers to parasitoids. Presence of different types of sensilla was noted which included mechano- and chemosensilla, as well as combined mechano-chemosensilla. Chemosensilla were most concentrated at ovipositor tips while mechanoreceptors were more widely distributed. Ovipositor traits of one putative parasitoid/inquiline species differed from those of its syntopic galler congeners and clustered with those of parasitoids within a different wasp subfamily. Thus ovipositor tools show lability based on adaptive necessity, and are not constrained by phylogeny. Life-history traits such as fecundity, pre-adult and adult lifespan were studied for each wasp member of the community. Trade-offs in life-history traits were also investigated. Interspecific variation in life-history traits was observed. Gallers were pro-ovigenic (all eggs were mature upon adult emergence) whereas parasitoids were synovigenic (eggs matured progressively during adult lifespan). Initial egg load was correlated with body size for some species, and there was a trade-off between egg number and egg size across all species. Although all species completed their development and left the syconium concurrently, they differed in their adult and preadult lifespans. Providing sucrose solutions increased parasitoid lifespan but had no effect on the longevity of some galler species. While feeding regimes and body size affected longevity in most species, an interaction effect between these variables was detected for only one species. Life-history traits of wasp species exhibited a continuum in relation to their arrival sequence at syconia for oviposition during syconium development, and therefore reflected their ecology. The largest number of eggs, smallest egg sizes, and shortest longevities were characteristic of the earliest-arriving galling wasps at the smallest, immature syconia; the converse characterised the later-arriving parasitoids at the larger, already exploited syconia. Thus life-history is an important correlate of community resource partitioning and can be used to understand community structure. The comparative approach revealed constraints and flexibility in trait evolution. This is probably the first comprehensive study of life-history traits in a fig wasp community.

Mating system, mating behavior and the evolution thereof is the foundation of this study. More specifically the effect of inbreeding on the evolution of mating behavior is investigated. To this end the pollinating fig wasp, Platyscapa awekei, lends itself to inquiry about inbreeding and the effect on its behavior. A pollinating fig wasp female will lay her eggs inside a syconium, and all offspring will mate with each other. Interestingly the abovementioned pollinating wasp exhibits male dispersal, not commonly expected to occur in a haplodiploid species observed to inbreed frequently. Several theories attempt to explain the evolution of male dispersal in this case, but very little work has been done on the effect of inbreeding on the choice to disperse. In order to study the effects of inbreeding it was necessary to be able to measure the inbredness of individuals. For this reason I developed micro satellite markers both to determine the inbredness of individuals but also to derive parentage from offspring genotypes. With the inbreeding status in hand I had to correlate this with fitness measures in order to derive the effect of inbreeding on this species. Interestingly I found both inbreeding and outbreeding depression, with optimal fitness at some point between fully inbred and fully outbred status. I give some explanations for the occurrence of dispersal in this species but come to the conclusion that dispersal is merely part of a mixed mating system and that more detailed work need to be done to derive what the specific effect of dispersal is on fitness. / Dissertation (MSc(Agric))--University of Pretoria, 2009. / Genetics / unrestricted

The behaviour of sex allocation has been extensively studied in hymenopterans (ants, bees and wasps) as an adaptive trait with respect to intra-specific competition within the framework of kin selection theory. Mating in these organisms often takes place in patchy populations established by the offspring of a few foundresses. Typically, there is a bias in favour of female dispersal from these patches. Theory predicts that foundresses that oviposit alone will do best to produce just enough sons to mate all of their daughters so as to maximize the number of dispersing daughters, under conditions of what is referred to as Local Mate Competition (LMC) between brothers to mate their sisters. If foundresses co-found a patch with other foundresses, they are expected to invest more resources in sons insofar as opportunity to sire offspring with the daughters of the other foundresses presents itself. Among organisms with such a life histories are fig wasps, the insects that pollinate and lay their eggs in the flowers that grow inside young figs. There is thought to be strong selective pressure for foundresses to use information about clutch size differences in species where clutch sizes are small and low foundress numbers are frequently encountered. However, less rigorous modes of sex allocation are thought to suffice in species encountering intermediate foundress numbers. Theory thus predicts a positive relationship between the degree of structure within mating populations and the information utilized by foundresses with respect to intra-specific competition for resources and mating opportunities. This is being extensively tested across the diverse species range of fig wasps and their hosts with the larger objective in mind of contributing to a better understanding of the role of natural selection in accounting for variation observed of intra-specific behaviour. This dissertation reports on a study of the sex allocation behaviour of the pollinating fig wasp Platyscapa awekei, a species characterized by low foundress numbers and clutch size differences brought about by foundress competition over oviposition sites. Offspring collected from experimentally controlled twofoundress broods were fingerprinted using microsatellite genetic markers to assign maternity and work out clutch size differences. These data are used to test what information foundresses use when allocating sex. It is reported that foundresses appear to use information of clutch size differences in two foundress broods. This observation provides evidence of advanced information utilization in fig wasps. More generally, the findings add support to the hypothesis that natural selection can bring about subtle adaptive behaviour at the individual level, but simultaneously highlights the importance of accounting for the selective regime of the organism being studied when attempting to understand the role of natural selection in the evolution of fine scale adaptive traits. / Dissertation (MSc)--University of Pretoria, 2011. / Genetics / unrestricted

Obligate nursery pollination mutualisms such as the fig–fig wasp system, with their central plant–pollinator mutualism associated with non-pollinating satellite wasp species, can function as closed system microcosms representative of tritrophic communities. In this system, enclosed inflorescences (syconia) function as sites of seed production, as well as brood-sites for the progeny of herbivorous mutualistic pollinators, non-pollinating gallers and parasitoids of the two. Plant reproductive traits such as inflorescence size (syconium volume) and within-plant phenology (within-tree asynchrony) as can affect inter-species relationships among the three trophic levels in such plant–herbivore–parasitoid systems. Induced or natural variations in such plant traits could influence various direct and indirect effects among the organisms in the community and could even lead to the formation of feedback cycles. Furthermore, changes in the abiotic environment could have major impacts on the biotic associations in the system either by affecting the community members directly, or through their effects on plant reproductive traits. Ficus racemosa with its fig wasp community comprised of a single herbivorous pollinator mutualist, three non-pollinating parasitic gallers and three non-pollinating parasitoids was used as a model to investigate: (1) the role of mutualistic and parasitic fig wasps in affecting within-tree phenology; (2) direct and indirect biotic associations between various groups of fig wasps (pollinators, gallers and parasitoids) and the influence of inflorescence size and within-tree phenology on them; and (3) variations in the reproduction of and the biotic associations between the organisms of the community under variable abiotic climatic conditions. Patterns of plant reproductive phenology are usually considered evolved responses directed at optimizing resource use, pollen receipt/donation schedules and seed dispersal for plant individuals. Within-plant reproductive synchrony or asynchrony can arise due to variation in floral initiation patterns, as well as from localised proximate mechanistic responses to interactants such as pollinators, parasites and herbivores which could affect floral longevity or fruit development time. The investigation was begun by exploring the role of a mutualistic pollinator, and for the first time in a brood-site mutualism, that of parasitic herbivores (gallers) and parasitoids in influencing within-plant reproductive phenology. Since a syconium functions as an inflorescence which develops into a fruit after pollination, investigations were carried out on the impact on syconium synchrony of fig wasps that began their development within the brood site syconium at pre-pollination, pollination, and post-pollination stages via their effects on the development time of individual syconia in Ficus racemosa. We found that syconium initiation patterns were not the only proximate mechanism for within-tree reproductive asynchrony, and that individual syconia (even within a tree) had highly plastic development times dependent on their sizes, pollination time and the species of wasp progeny developing within them. Syconium volume, pollination early in the pollen-receptive phase and presence of early-ovipositing galler progeny reduced syconium development time, whereas the presence of late-ovipositing parasitoid progeny or pollination late in the pollen-receptive phase increased syconium development time. These results suggest an ongoing tug-of-war between syconium inhabitants to modify syconium development times. Parasitic fig wasps pull in different directions to suit their own needs, such that final syconium development times are likely to be a compromise between conflicting demands from developing seeds and from different wasp species. Inter-species relationships among the three trophic levels in plant–herbivore–parasitoid systems can potentially include various direct and indirect effects possibly mediated by induced or natural variations in plant traits. Analysing the seed and fig wasp compositions of microcosm replicates, i.e. individual syconia, shows that besides direct competition for resources and predator–prey interactions, the F. racemosa community also displays exploitative or apparent competition and trait-mediated indirect interactions. Syconium volume and within-tree asynchrony were reproductive plant traits that not only affected plant–herbivore and plant–parasitoid associations, but also possibly modified herbivore–herbivore and herbivore–parasitoid interactions. Our results also indicated that the reciprocal effects of higher trophic level fauna on plant traits (and vice versa) within this system drive a positive feedback cycle between syconium inhabitants and within-tree reproductive asynchrony. In the F. racemosa system, abiotic environmental factors could affect the reproduction of mutualistic pollinators, non-mutualistic parasites and seed production via seasonal changes in plant reproductive traits such as syconium volume within-tree asynchrony. Temperature, relative humidity and rainfall defined four seasons: winter; hot days, cold nights; summer and wet seasons. Syconium volumes were highest in winter and lowest in summer, and affected syconium contents positively across all seasons. Greater transpiration from the nurseries was possibly responsible for smaller syconia in summer. The 3–5°C increase in mean temperatures between the cooler seasons and summer reduced fig wasp reproduction and increased seed production nearly two-fold. Yet, seed and pollinator progeny production were never negatively related in any season confirming the mutualistic fig–pollinator association across seasons. Parasites affected seed production negatively in some seasons, but had a surprisingly positive relationship with pollinators in most seasons. While within-tree reproductive phenology did not vary across seasons, its effect on syconium inhabitants varied with season. In all seasons, within-tree reproductive asynchrony affected parasite reproduction negatively, whereas it had a positive effect on pollinator reproduction in winter and a negative effect in summer. Seasonally variable syconium volumes probably caused the differential effect of within-tree reproductive phenology on pollinator reproduction. Within-tree reproductive asynchrony itself was positively affected by intra-tree variation in syconium contents and volume, creating a unique feedback loop which varied across seasons. Therefore, nursery size affected fig wasp reproduction, seed production and within-tree reproductive phenology via the feedback cycle in this system. Climatic factors affecting plant reproductive traits can cause biotic relationships between plants, mutualists and parasites to vary seasonally and must be accorded greater attention, especially in the context of climate change.

The inadequacy of arbitrarily classifying interactions between species as antagonistic, neutral or mutualistic has become clear in recent years. Both direct and indirect interactions between species can vary between mutualism and antagonism, depending on the intrinsic and extrinsic contexts of the interaction. This study investigated the characteristics of an ant-plant-homopteran interaction in southern Africa. The polyphagous homopteran Hilda patruelis (Tettigometridae) feeds primarily on the trunk-borne fruiting branches and figs of the Cape fig tree, Ficus sur, and produces honeydew which attracts tending ants. Ten of the sixteen ant species/species groups present on F. sur tended H. patruelis, with Pheidole megacephala the most frequent attendant. Ants attracted to F. sur by H. patruelis honeydew or other liquid food sources also preyed on insects on the tree, including adults of the small agaonid fig wasps whose larvae feed on the ovules in the developing figs. One fig wasp species (Ceratosolen capensis) is also the tree's only pollinator. No benefits to H. patruelis from being tended by ants were detected, either in terms of reduced parasitism, or predation by a lycaenid caterpillar. A P. megacephala colony foraging on a F. sur tree was found to receive a high proportion of its likely energy requirements from the tree, mainly in the form of H. patruelis honeydew, during periods when it was bearing fruit. It is probable that the H. patruelis-P. megacephala interaction constitutes a direct mutualism at times, but that benefits to the homopteran are intermittent or weak. Both H. patruelis and ants benefitted from F. sur, directly or indirectly, through the provision of food (and for some ants, nesting sites). The removal of phloem sap by H. patruelis did not detectably reduce the trees' reproductive output, either in terms of pollinator or viable seed production. The indirect effects of ant and H. patruelis presence on the F. sur trees were on average positive, as ants preyed disproportionately heavily on fig wasp species parasitic on or competing with the pollinator, thus increasing pollinator production. Effects of ant presence on seed production were not investigated, but have been demonstrated as beneficial elsewhere. However, there is great varatlon both in the composition of the wasp fauna arriving to oviposit at different crops, and in ant densities per fig, on several temporal and spatial scales. This results in high variability in the effects of ants on the pollinator and, through it, the tree, from positive to zero and potentially even negative. Despite this conditionality of beneficial outcomes for the tree, the mean effect of ants on the F. sur population studied was to increase pollinator production by up to nearly 20%. This study is among the few to have demonstrated an overall benefit to a plant of having homopteran-tending ants present on it.

Fig trees (Ficus spp., Moraceae) and fig wasps (Chalcidoidea, Agaonidae) are uniquely associated. In one fig wasp group, the pollinators (Agaoninae), each species is generally host species-specific. The relationship is one of obligate mutualism where the wasps provide pollination services and in return utilises some of the ovules for larval development. Non-pollinating fig wasps (generally belonging to subfamilies other than the Agaoninae) may be gallers or parasitoids, and can also be host species-specific. In the accompanying studies we examined the factors governing the interactions between fig wasps and their host trees. Surveys of fig trees and their associated pollinating fig wasps conducted in southern Africa, Madagascar and The Comores generally confirmed their specific relationships. An examination of F. sycomorlls in Madagascar resulted in the reclassification of F. sakalavarum as a distinct species with its own specific pollinator species. Biological and chemical evidence is presented demonstrating that the pollinators were able to distinguish their hosts through volatiles which emanated from the figs when they were ready to be pollinated. Environmental factors were found to influence wasp behaviour. Ambient temperature governed the timing of wasp emergence from their natal figs. When dispersing from their natal figs, the fig wasps flew upwards and then were blown downwind. Once nearing trees bearing figs ready to be pollinated, the wasps lost height and flew upwind towards the trees. E. baijnathi females apparently avoided figs which already contained a conspecific foundress. Scanning electron microscope studies of pollinating female fig wasp antennae showed that while all the species possessed multiporous plate sensilla, in only a few species were these sensilla elongated. Multiporous plate sensilla elongation is rare or absent among other female chalcids and may have evolved within the Agaoninae in order to facilitate their location on receptive host figs. Pollinator choice specificity appears to break down in a number of cases. In the first case examined, two pollinator species were recorded from the figs of African F. sycomorus. One. C. arabicus, pollinates the figs while the other, C. galili, acts as a 'cuckoo' by utilising some of ovules for oviposition without providing pollen. In the second case three pollinating fig wasp species were recorded from the rigs of F. lutea. Two were found to be incidental visitors and were not specifically attracted to the tree. The hybrid seeds from these crosses were successfully germinated but the seedlings did not grow passed the cotyledon stage of their development. In the concluding study the consequences of Ficus phenology and the structure of the fig's unusual inflorescence on the nonpollinating fig wasp community were examined. Various factors affecting the population levels and species richness were also examined. Future possible research directions were discussed.