Degree Name

Doctor of Philosophy


Department of Biological Sciences


European honeybees were introduced to Australia in the 1820s and are now widespread as feral and domestic colonies in the temperate region, foraging for nectar and pollen from plants in at least two hundred genera. Many of the plant species visited by honeybees are among the nearly 1,000 species that have floral structures that facilitate pollination by birds or mammals. Numerous ecologists have expressed concerns that the pollination systems of such plants, which evolved in the absence of winged social insects, are at risk of disruption by honeybees, particularly because anthers, nectaries and/or stigmas are often separated by several centimetres. At the same time, honeybee industry advocates continue to argue for greater access to natural areas, citing a lack of evidence of negative impacts. To date, experimental studies of potential honeybee disruption of Australian pollination systems have produced a range of outcomes apparently demonstrating that honeybees can enhance, decrease or have little effect on the quantity of fruit or seed produced in different vertebrate-adapted species.

Less apparently, honeybees m a y be altering plant fitness by reducing seed quality. Such an effect would be expected if short-distance honeybee movement facilitate geitonogamy and inbreeding in contrast to longer-distance movements by native pollinators. As a result, honeybees could be expected to be a significant evolutionary pressure for some plant species.

A comparative approach is needed to determine what types of pollination systems are most at risk of negative honeybee effects such as severe reproductive decline or inbreeding depression. In this study, I compared the roles of honeybees and native animal visitors in the pollination ecology of bird- and insect-adapted plant species using confamilial pairs of plant species in three plant families: Callistemon citrinus and Baeckea imbricata (Myrtaceae); Styphelia tubiflora and Epacris microphylla (Epacridaceae); and Grevillea acanthifolia and G. sphacelata (Proteaceae). In all six species, I identified the suites of probable pollinators and observed their foraging behaviour at flowers and among plants. I compared the proportion of flowers that produced fruit among selective pollination exposure treatments - exclusion of birds, exclusion of all pollinators, and open pollination - to deduce the likely role of honeybees. In the Epacridaceae and Proteaceae, I determined if plants were self-compatible and compared fruit set of manually pollinated flowers with that of open flowers to determine if plants were pollinator limited. Finally, in order to determine whether honeybees may be reducing seed quality, I compared seed produced from selective pollinator exposures for viability and genriination rate (in Callistemon citrinus) and outcrossing rate (in Grevillea acanthifolia).

Overall, the results of this study did not support the hypothesis that pollination of bird-adapted plants by honeybees alone would produce less fruit than open pollination. Rather, fruit set in such species following exposure to honeybees alone was equivalent (Callistemon citrinus), marginally lower (Grevillea acanthifolia) or higher (Styphelia tubiflora) than fruit set following exposure to both honeybees and birds. The last outcome appeared to be an artefact of the exclusion cages, which excluded native bird nectar robbers; such birds appeared to be responsible for damaging more than 60% of flowers on 60% of exposed plants.

This study showed that honeybees might interfere with the pollination of insect-adapted plant species, even if they brush pollen onto stigmas. In two populations of Grevillea sphacelata, each flower was likely to receive more than 20 honeybee visits per day, whereas native insect visitors were so infrequent that a comparable measure could not be reliably calculated. Individual honeybees moved infrequently among plants and pollen was nearly always removed from open flowers by previous visitors; of 870 flower visits, only 5 resulted in visible amounts of pollen deposited on the honeybee. Honeybees appeared to promote geitonogamous pollination in the self-incompatible G. sphacelata, pre-empting potentially more effective pollination by native insects and thus limiting fruit set. Hand-pollination studies revealed that cross-pollination produced 4 to 13 times more fruit than open pollination. In contrast, honeybees appeared to contribute to fruit set along with numerous native insect species in two self-compatible species (Epacris microphylla and Baeckea imbricata).

Honeybees may not provide pollinations of equivalent quality to native pollinators in some bird-adapted species. In the self-compatible Callistemon citrinus seed germination experiments suggested that seed produced following pollination by honeyeaters and honeybees germinated more rapidly (80 % cumulative germination at 4.6 days) than seed produced by honeybee pollination alone (6.5 days). Seed produced by autogamy germinated slowest (9.4 days). In two Grevillea acanthifolia populations, the outcrossing rates of seed produced by open pollination were not significantly different from those of seeds produced by exposure to insects alone, suggesting that honeybees do not cause inbreeding. Nevertheless, behavioural differences between honeybees and birds suggest that the social insect may be decreasing genetic neighbourhood size. Taken together, the results of studies in these two species suggest that honeybees may be facilitating changes in plant populations which have not yet been detected, including inbreeding, character shifts (e.g. shift toward self-compatibility) and decreasing genetic neighbourhood size.

Pollination "syndromes" are not a good predictor of honeybee effects. It is also evident that honeybees may have little effect on the pollination systems of some species. However, this study demonstrates that honeybees may be disrupting reproductive success in self-incompatible species, even if they are adapted to insect pollination, and may be causing genetic shifts in species with mixed mating systems, even if they do not limit seed production. Based upon the evidence in this study, the conservation consequences for many plant species might include severe pollinator limitation of fruit set, a shift toward self-compatibility in plants with mixed mating systems and a decline in population viability because of loss of genetic variability.



Unless otherwise indicated, the views expressed in this thesis are those of the author and do not necessarily represent the views of the University of Wollongong.