Ecology and degree of specialization of South African milkweeds with diverse pollination systems

Like orchids, the complexity of flowers found in asclepiads (Asclepiadoideae, Apocynaceae) and the fact that pollen is presented as pollinaria, offers excellent opportunities to study various aspects of plant-pollinator interactions. In this thesis I investigated two broad themes: ecological aspects of the pollination biology of hymenopteran and fly-pollinated asclepiads as well as the degree of specialization to certain pollinators in these species. Colonizing plants often reproduce through self-pollination, or have highly generalized pollination systems, or both. These characteristics facilitate establishment in small founding populations and generates the prediction that reproductive success should be independent of population size in these species. Chapter one examines the pollination biology of Gomphocarpus physocarpus, an indigenous, weedy species and investigates the relationship between reproductive success and population size. In this species, there is no evidence of an Allee effect and reproductive success is not correlated with population size. In addition G. physocarpus is not capable of self-pollination, suggesting it is completely reliant on pollinators for seed set. The lack of a relationship between pollination success and population size is therefore likely explained by the generalized wasp pollination system of this species. Several milkweeds are invasive outside of their native ranges. Invasive species either need to co-opt native pollinators in order to reproduce or reduce their reliance on pollinators through having the ability to self-pollinate. Co-opting native pollinators is expected to be easier in species that have generalized pollination systems, alternatively species with specialized flower morphologies need to rely on similar functional groups of pollinators to be present within the invaded range. Chapter two investigates the pollination biology and pollination success of the invasive milkweed, Araujia sericifera, and finds that in South Africa, this species is visited mainly by native honeybees and nocturnal moths. Moths however contribute little to pollen removal, and deposition. Based on the apparent morphological mismatch between the flower of A. sericifera and native honeybees, I propose that the native pollinators of this species are likely to be larger Hymenoptera (e.g. Bumblebees). Data from a breeding system study, indicated that this species is not capable of automatic self-pollination, but could set fruit from geitonogamous self-pollinations pointing to the importance of native pollinators for successful reproduction. The pollinaria of milkweeds can accumulate on pollinators to form pollen masses large enough to physically interfere with the foraging behaviour of pollinating insects. In chapter three I describe the pollination biology of Cynanchum ellipticum and find that this species is mainly pollinated by honeybees although this species is visited by several other members of Hymenoptera, Lepidoptera and Diptera. Due to the structure of the pollinaria, these chain together relatively efficiently and frequently form large pollinarium loads on the mouthparts of honeybees. However there is little evidence that these pollinarium loads influence the foraging times of pollinators and only a few individual honeybees exhibited longer foraging times and most honeybees were unaffected by the presence of large pollinarium loads. Within the genus Cynanchum there is large variation in the gynostegium structure that may influence the pattern of pollinarium loading on pollinators as well as pollen reception as shown in chapter three. In Chapter four, the pollination biology of Cynanchum obtusifolium is examined, and like that of C. ellipticum, this species is visited by a wide diversity of pollinators but honeybees appear to be the primary pollinators. More importantly this species is shown to be andromonoecious and produces two morphologically different flower types, that may be distinguished based on differences in the gynostegium structure. These two types of flower could mainly be distinguished by the length of the anther wings. I found that flowers with short anther wings function as male flowers by only exporting- and rarely receiving pollinia. Flowers with longer anther wings function as hermaphrodite flowers and can both export and receive pollinia. The ratio of male to hermaphrodite flowers varied at different times during the flowering season, but preliminary data suggested that this was not related to levels of pollination success. The genera Stapelia and Ceropegia are well known for their intricate floral adaptations that mimic the brood and feeding substrates of pollinating flies. Despite several studies that have documented the various adaptations to fly pollination in different species, there is a lack of natural history studies documenting different flower visitors, pollen loads and long term levels of pollination success in these species. In Chapter six I document the pollination biology of Ceropegia ampliata by documenting different pollinators and quantifying average levels of pollination success and the nectar reward. I also experimentally manipulated the trapping hairs of this species to determine whether trapping hairs influence average levels of pollen export and receipt. I show that Ceropegia ampliata is pollinated by a generalist guild of flies (mainly Tachinidae, Sarcophagidae, Muscidae and Lauxaniidae) and produces minute quantities of relatively dilute nectar as a reward. Pollination success was generally low in this species and increases periodically suggesting that the abundance of pollinators is patchy. I found that flowers with trapping hairs that had already wilted had higher levels of pollinarium removal than flowers with erect hairs, however experimentally removing the hairs had no significant effect on pollen export and receipt. In Chapter seven, I document the pollinators, pollen loads and long term levels of pollination success in Stapelia hirsuta var. bayllissi, a rare sapromyiophilous stapeliad. I find that, in contrast to C. ampliata, this species was specialized to pollination by small flies of the family Anthomyiidae. Similar to the results from Chapter seven, I find that long term levels of pollination success were typically low but could increase periodically, although such increases were generally unpredictable. There are currently very few records documenting pollinator interactions in the Periplocoideae. Many species within this subfamily exhibit open-access flowers suggestive of pollination by short-tongued insects. I investigated the pollination biology of Chlorocyathus lobulata, a rare species with a highly localized distribution. I aimed to determine the pollinators, average levels of pollination success and demography of this species in order to determine whether this rare species is suffering from the collapse of a highly specialized pollinator mutualism. I also quantified the high incidence of flower herbivory caused by larvae of the moth, Bocchoris onychinalis. I find that C. lobulata has a highly generalized fly pollination system and average levels of pollination success suggested that a large proportion of flowers had pollen removed and deposited suggesting that this species is not experiencing pollination failure. The large numbers of juveniles present also indicated that recruitment is taking place