Degree Name

Doctor of Philosophy


Department of Biological Sciences


The species is one of the most important concepts in organismic biology. There is, however, a long history of disagreement about how to define species and how they arise. All current species concepts are logically and operationally flawed because they rely on prospective narration, or equivalently, they depend upon the future. The solution is to use a species concept that is appropriate to a specific objective. The objective in this thesis was to study the processes producing divergence and ultimately perhaps speciation. It was most appropriate then to define species here as the most inclusive monophyletic group of organisms having the potential for genetic and/or demographic exchangeability. Persoonia mollis R. Br. sens, lat (Proteaceae) is a complex species that offers great opportunity to effectively address unresolved concepts relating to both systematic pattern and evolutionary process at the specific and infra-specific level.

Persoonia mollis R.Br. sens, lat was circumscribed and formally revised, and, as here defined, is clearly separated from all other species and constitutes a monophyletic group. A key and description were provided for the nine taxa recognised as subspecies. Five new taxa were described in P. mollis: subsp. maxima, subsp. nectens, subsp. leptophylla, subsp. livens and subsp. budawangensis. Three new combinations were made: P. mollis subsp. caleyi, P. mollis subsp. revoluta and P. mollis subsp. ledifolia.

Support for the recognition of these nine infraspecific taxa within P. mollis was found following a multivariate phenetic analysis of morphological variation. Eighty-seven flowering herbarium specimens, collected from throughout the range of P. mollis, were measured for 24 characters incorporating size, shape and pubescence on vegetative and floral parts. These data were analyzed by two ordination procedures, multidimensional scaling (MDS) and canonical variate analysis (CVA). Plots of canonical variate scores against latitude or longitude across subspecies boundaries revealed phenetic homogeneity within subspecies with sharp transitions between subspecies. The geographic variation in morphology within P. mollis can be best described as a mosaic of nine recognisably distinct allopatric or parapatric taxa, where narrow zones of morphological transition separate neighbouring taxa.

The breeding system of P. mollis was characterised to firstly assess its importance as a mechanism promoting genotypic diversity and secondly to investigate the modee of control over selective fruit abortion. Fruit quantity and quality was assessed following self- and outcross-pollination manipulations. Twenty percent of outcrossed flowers set fruit, compared to only 1% of flowers fertilized with self-pollen. Fruits produced by self-fertilization were 72% the weight of cross-fertilized fruits. Fruits produced by self-fertilization were significantly fewer in number and lighter than fruits following natural pollination of unmanipulated flowers (17% fruit set), but outcrossed and naturally pollinated fruits were equivalent. Flower to fruit demography suggested that a post-zygotic mechanism may be preferentially selecting the most vigorous genotypes, as ovary abscission occurs mostly between 4 and 30 weeks after pollination regardless of pollen source. Self-pollen tube growth was found to be inhibited within the style, while pollen tubes were found in the ovary for 50% of all outcrossed flowers. These data suggest that a pre-zygotic "pseudo" self-incompatibility mechanism is the cause of low fruit set following self-pollination. The breeding system of P. mollis was found to promote outbreeding, with an emphasis on flexibility and post-zygotic choice following pre-zygotic "pseudo" self-incompatibility.

Severely restricted gene flow may be a factor contributing to the remarkable amount of morphological variation within P. mollis. The mating system and realized pollen dispersal were studied to assess their effect on gene flow. Mating system parameters were estimated in seven natural populations over two seasons using allozyme electrophoresis. Realized pollen dispersal was measured in two natural populations over two seasons by monitoring the dispersion of a rare allozyme from a known source plant in each population. Single- and multi-locus estimates of outcrossing rate (t) were consistently equal to or greater than one (i.e. complete outcrossing). Realized pollen dispersal distances showed that 99% of the pollen received by given females was donated by males on average within 33m. However, 70% of all pollen dispersal was on average to the paternal plant's immediate neighbour. Genetic neighbourhood sizes due to pollen dispersal alone ranged from 1 to 5 plants, and paternity pool sizes ranged from 4 to 22 plants. These population sizes are small enough to allow genetic differentiation in the absence of selection. However, in contrast to the expectation that small population size leads to biparental inbreeding and reduced heterozygosity compared to Hardy-Weinberg expectations, the mean fixation index (F) of -0.035 indicated a slight excess of heterozygotes in the seed cohort. This apparent paradox could be the result of selection for heterozygous seeds, disassortative mating, or more likely because gene flow through seed dispersal substantially increases the neighbourhood sizes estimated here through pollen dispersal alone.

Hierarchical patterns of genetic diversity and gene flow were estimated within P. mollis from allozyme frequency data. The total gene diversity (HT) within P. mollis was extremely low (0.139) compared to an average of 0.310 for 406 plant species. This low gene diversity m a y be typical of the Proteaceae. However, P. mollis was typical in the way its gene diversity was distributed, with 78.3% of the total gene diversity found within populations. Of the 21.7% found among populations, 17.9% was attributed to differences among subspecies and only 3.8% attributed to differences among populations within subspecies. Indirect estimates of gene flow (Nm) among all 18 populations of P. mollis were approximately 1 individual per generation. Estimates of gene flow (Nm) between populations of neighbouring subspecies were generally well in excess of 1. A more detailed comparison of gene flow within and between some neighbouring subspecies revealed only 1 of 4 cases where Nm within subspecies was significantly in excess of Nm between subspecies. However, even for this exception, Nm was still in excess of 1. These results show that P. mollis is morphologically differentiated despite the homogenizing effects of gene flow. Consequently, natural selection rather than genetic drift was inferred to be primarily responsible for the patterns of morphological differentiation within P. mollis.

The phylogenetic analysis of conspecific populations is essential for an understanding of historical processes within species such as range expansion, divergence and ultimately speciation. The phylogeny of 18 populations representing all nine subspecies within P. mollis was estimated from allozyme frequency data. Trees were constructed under different models and assumptions. These procedures were maximum likelihood (CONTML), maximum parsimony (FREQPARS), UPGMA , distance Wagner and neighbor joining. Major differences in topology between trees constructed under an assumption of an evolutionary clock (UPGMA) and trees that do not assume equal rates of divergence indicated that evolutionary rates are not equal in different lineages in P. mollis. Of the non-UPGMA trees, the maximum likelihood and maximum parsimony trees produced similar topologies and were the most optimal under the criteria of likelihood (given the model of all change due to drift) and tree length. The major patterns produced included an extremely close congruence between geographic distance between populations and the position of each population on the tree for most populations, the early differentiation of subsp. maxima, the well supported clade of all other P. mollis populations and, within this clade, the split into two clades that although distinct, were weakly differentiated at their base. These trees were consistent with a scenario of range expansion along two distinct lineages in a southern direction. These lineages currently terminate in populations that share a hybrid zone of apparently secondary origin west of the Budawang Range.

Hybrid zones provide a unique opportunity to study the processes involved in the differentiation of populations and ultimately speciation. Narrow hybrid zones between P. mollis subsps. revoluta, livens and ledfolia were studied. These zones were associated with ecotones. Gene flow was assessed indirectly from allozyme frequency data. The fitness of hybrids was assessed by fruit set and weight following controlled pollination manipulations across the hybrid zones. The relationship between fitness and the spatial distance between mates was assessed to test for the presence of an optimal outcrossing distance. This was tested for initially at distances of up to 15 k m across the hybrid zones, and in a second experiment involving plants separated naturally by distances of up to 150 km. Unique allozyme markers were tracked from pollen to seed in a manipulated population to determine whether pollinators (bees) discriminate between plants from different subspecies. There was no evidence for restricted gene flow across these zones, as indirect estimates of gene flow were extremely high (average Nm over 8 loci was 19), and pollinators did not discriminate between subspecies. There was no detectable fitness effect following pollination across the hybrid zones and no evidence for optimal outcrossing at these distances as fruit set and weight were not significantly different for different pollen sources. These results indicate that hybrid seeds are neither advantageous nor disadvantageous compared to parental seeds. There was, however, evidence for asymmetric outbreeding depression at distances of 150 km.

Conclusions about the role of natural selection in affecting phenotypic variation among populations must first distinguish the variation due to phenotypic plasticity, as only the former is heritable. The extent of phenotypic plasticity within P. mollis was assessed by reciprocal transplant experiments of vegetatively propagated seedlings between 3 pairs of populations of markedly different phenotypes. After 15 months in "home" or "away" environments, 10 leaf characters were measured on recently produced leaves from transplants and adults, the adults being the source plants of cuttings. The change in phenotype was assessed by the ordination procedure non-metric multidimensional scaling and by ANOVA . Although there was evidence for plasticity, the diagnostic differences between reciprocal populations were not removed from leaves produced under new conditions. Therefore, the diagnostic differences between these populations appear not to be due to plasticity, but are under heritable control. However, similar experiments with seeds are required to assess the extent of genetic canalization in developing seeds and seedlings.



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.