The proportions of male and female plants In natural populations were investigated in twenty-five dioecious species in the Brisbane area, with detailed studies being confined to species of three genera, namely Lomandra, Casuarina and Baccharis.
Observations of tagged individuals in populations of four Lomandra species were made at frequent intervals throughout their flowering seasons. These revealed that reliable estimates of the sex ratio required careful attention to the following factors: the constancy of sex expression, the accurate delimitation of individuals, the proportion of sexually indeterminate individuals in the populations, differences between the sexes in the time and duration of flowering, and the degree of conspicuousness and persistence of inflorescences, both during and after flowering. The relative importance of these factors was found to vary with the species studied. Results revealed that when all the above factors were taken into consideration, no marked deviation from the expected unity value of the sex ratio was detected for any of the Lomandra species studied.
Similar conclusions were reached for Baccharis halimifolia. Because in this species male inflorescences tend to be lost from the plant soon after flowering, it was found necessary to score populations at the height of the flowering season in order to ensure nonflowering males were not overlooked.
Variation in the sex ratio with plant size detected in Lomandra confertifolia led to the consideration of the possibility that sex ratios might be age dependent. Investigations were undertaken with Casuarina where, in the absence of the formation of annual growth rings in the local species, girth at breast height (gbh) was adopted as an index of age. Variation in the sex ratio with gbh was detected both within and between populations of C. littoralis at Stradbroke Island and Caboolture. An excess of females was observed amongst individuals of the smallest class interval in both populations, but only in the Stradbroke Island population was there a marked tendency for the sex ratio to increase with gbh.
As a means of elucidating the fluctuations in the sex ratio of C. littoralis, two sample areas at Caboolture were mapped so that spatial relationships between the sexes might be studied. Results revealed marked differences between the two areas; an apparently pronounced aggregation of males around females in one area was absent in the second. The origin and significance of variations in the sex ratio, and in the spatial distribution of plants in the Stradbroke Island and Caboolture populations were considered with special reference to cyclone and fire damage, mistletoe infestation, differential longevity and differential germination and establishment of the sexes. Of the environmental factors considered, fire undoubtedly had the most pronounced effect. The clustering of progeny in the vicinity of female plants appears to result in an increased accumulation of flammable material around the parent tree, which is then more prone to fire damages. Thus, fire may well contribute to an increase in the frequency of males in populations of this species.
Preliminary estimates of sex ratios were made for a further eleven species from eight genera. An additional thirteen species were rejected for study because of an insufficient number of plants being available, unresolved difficulties in delimiting individuals, or a high proportion of sexually indeterminate plants in populations.
The problem of sexually indeterminate plants was found to be important in all species studied, particularly with respect to seedling populations. In an endeavour to overcome this problem, various methods were employed to find some means by which nonflowering plants might be sexed. These involved a study of the morphology, physiology, biochemistry or cytology of plants, but none of these techniques was successful.
Sex ratio data collected for local species were compared with those published for taxa growing elsewhere in Australia and overseas. The origin and significance of ratios differing from the expected unity value were discussed.
The relationships between dioecism and selected attributes associated with reproduction were investigated for taxa of the Queensland flora, the analyses being undertaken at both the family and generic levels.
Analyses of both the Monocots and Dicots revealed there to be a marked tendency for dioecious taxa to have many-flowered inflorescences and few-seeded fruits. The biological significance of the association of each of these characters with dioecism was considered.