Degree Name

Doctor of Philosophy


Department of Biology


Mating systems are of fundamental importance in any consideration of the dynamics of plant populations, since patterns of mating determine the transmission of genes in a population, and thus have a major influence on the structure and dynamics of populations. Moreover, plant reproduction is influenced by events which occur throughout the mating process. Factors such as flowering patterns, pollinator effectiveness and behaviour, population structure, and incompatibility mechanisms are important determinants of mating patterns. Any quantitative and definitive assessment of plant mating systems therefore requires an integrated approach, which encompasses these factors. This study used such an approach to investigate the mating system of one species of Banksia. The reproductive biology of a population of Banksia spinulosa was studied in detail at a site within Barren Grounds Nature Reserve near Wollongong. Several important aspects of the reproductive biology of this species were investigated, including; (i) flowering and fruiting patterns, (ii) pollinator visitation and behaviour, (iii) the mating system and genetic structure of the population and (iv) the potential for choice in the system.

Flowering and fruiting patterns exhibited a great deal of variability, both amongst plants in a given season and over seasons. Total numbers of inflorescence produced in the population were greatest in 1987, due partly to greater numbers of inflorescences per plant, and also to increases in the number of plants flowering. In contrast, the total number of infructescences produced were similar for the three years. However, seed output was greatest in 1988, due to increased numbers of follicles per infructescence in this year. Some plants in the population consistently had a higher reproductive output than others. These plants produced seed each year, and had a greater number of mature infructescences each year and overall. The total reproductive output of these plants (numbers of follicles produced) was approximately double that of other plants. Another group of plants never produced seed during the study, even though they flowered each year. These results illustrate the importance of considering individual variability in the population, rather than the more commonly measured, total, or mean reproductive success.

The principal pollinators at this site were mammals and birds. These included the sugar glider, the brown antechinus, the eastern pygmy possum and the eastern spinebill. Although several insect species (including moths and honeybees) also visited inflorescences, these were found to be less effective at pollinating than the vertebrate visitors. Moths carried very little pollen, and the foraging behaviour of other insects was unlikely to promote pollination. Examination of pollen tube growth indicated that pollination success was similar for nocturnal and diurnal visitors. However, nocturnal visitors were more effective at removing pollen from newly opened flowers. Furthermore, the foraging behaviour of mammals was such that they would transfer copious amounts of pollen to inflorescences, and probably contact receptive stigmas more often. Overall, mammals were considered to be slightly more effective pollinators than the more obvious daytime visitors, eastern spinebills.

Mammalian visitors showed considerable variability in numbers of inflorescences visited per plant and distances moved between plants visited. Some of this variability was associated with changes in flower abundance. Fewer flowers were available later in the season, and at this time, animals tended to visit fewer inflorescences on a plant and travelled greater distances between visited plants. As a consequence, pollen dispersal was also likely to exhibit within-season variability, with dispersal distances increasing as the abundance of flowers declined. Nevertheless, at any time during the flowering season, the foraging behaviour of animals was such that pollen loads transferred to flowers would normally contain mixtures of self- and cross-pollen.

In contrast to apparent patterns of pollen receipt, outcrossing rates estimated for this population were found to be very high overall. However, several assumptions of the mixed mating model used in estimations were violated. For example, alleles of Adhi apparently did not conform to Mendelian inheritance, and showed evidence of a null allele. Apparent non-Mendelian inheritance was also found in Pgii. Some slight but significant linkage was also detected between loci. Analysis of the genetic structure of the population revealed further violations associated with mating system analysis. Some genetic differentiation was detected, and the population as a whole deviated from a panmictic unit. However, these deviations were small compared to species which are predominantly selffertilizing, and they were primarily due to the behaviour of one locus, Pgii. This locus exhibited a large deficit of heterozygotes, possibly in association with a Wahlund effect. Despite the abovementioned problems associated with the mating system analysis, estimates of outcrossing were considered to be a reasonable approximation for this population.

Experimental pollinations of inflorescences using applications of self- and cross-pollen, and mixtures of the two, indicated that, at the early pre-zygotic stage, all treatments had an equal potential to produce seed. In contrast, the autogamous treatment did not produce any pollen tubes. However, few seeds resulted from the selfpollination treatment, indicating that self-pollen may have been, inhibited at some later stage in the mating process. The genetic composition of seed from mixed pollinations gave further weight to this hypothesis. All progeny genotyped were apparently derived from outcrossing events. It was suggested that this preferential outcrossing may occur via some sort of recessive lethal system, and that it may also account for the observed deviations from Mendelian inheritance. These results also explain the apparent discrepancy between apparent pollen dispersal by pollinators (i.e. the likelihood of a large amount of self-pollination) and the high outcrossing rates obtained.

The results obtained from this study highlight the value of integrating ecological and genetical data in investigations of plant mating systems. Patterns of mating in B. spinulosa were influenced by many variables, and selective processes were considered to be of major importance in this system. This study has also posed some new directions for future research. Firm conclusions were often frustrated by a lack of understanding of factors which might limit seed production in this and other species. Studies specifically designed to examine this, and many other questions, are urgently required for further advancement in this field of research.