I investigated the patterns of male and female flower production in two populations of a monoecious, self-compatible, prostrate vine, Apodanthera undulata Gray (Cucurbitaceae). Small, and probably young, plants produce no flowers. Larger and older plants produce only male flowers, while a somewhat greater threshold size is necessary for female flower production. Beyond these threshold effects, femaleness, a proportional measure of allocation to female function, did not increase with plant size. Thus, allocation to both male and female functions increased with size. In contrast, femaleness decreased with increasing flower production; plants with many flowers opened relatively more male flowers than plants with fewer flowers. This trend, which seemed stronger in the low density population, can have important effects on pollinator behavior and thus on pollen dispersal and receipt. Femaleness was positively correlated between years for all plants and for all cosexes, plants that produced both male and female flowers. Thus, plants that opened only male flowers one year were likely to open only male flowers the next year. Similarly, cosexes were likely to be cosexes again in the following year, with similar femaleness values. Approximately 10% of all flowering plants changed gender group. These patterns suggested that all plants were male until they reached a certain size and that plants had an intrinsic femaleness value due to either genotype, microsite or environmental effects. Finally, some of the reproductive consequences of these gender allocation patterns were investigated. The factors determining success, such as predation and rainfall, were unpredictable between populations and years. Unless success was especially low (i.e., few fruits matured in the population), success through male and female functions were positively correlated, as could be predicted for a monoecious species. The gains for male success were greater with increasing allocation to that function in the high density as opposed to the low density population. This suggested that plants in the high density population should allocate more to male function, as was observed. Though preliminary, these data support the predictions of sex allocation theory for cosexual species.

Improvements in soft fruit cultivars have been achieved by the selection of superior genotypes through plant breeding. Limitations imposed by plant breeding, led to work being initiated into the development of a gene vector technique for soft fruit, to permit improvement without altering their overall genetic makeup. Little research has been carried out in this area with perennial crops, cultivars of which are highly heterozygous, their genetic structure being maintained by vegetative propagation. Probably the most successful method of transformation available involves the use of the soil bacterium Agrobacterium. This bacterium is manipulated to contain the gene of interest, and used to infect small explants of the genotype which, can result in gene transfer to some cells of the explant. These transformed cells are induced to regenerate whole plants, which are assessed to select those containing the gene of interest. A transformation system suitable for use in three soft fruit species (Rubus. Ribes and Fraaaria spp.) was developed. Initially these species and Vaccinium spp. were inoculated with a range of wild type Agrobacterium isolates to demonstrate the ability of the isolates to infect. All but one isolate was capable of effecting gene transfer (to some degree) into the plants, which was demonstrated by the production of disease symptoms. Whole plant inoculation was not useful for Agrobacterium transformation, as only the cells at the site of infection contained the foreign DNA. Regeneration techniques were developed for each species, so that the infected cell(s) would regenerate to form whole plants containing the foreign DNA in every cell. With Rubus spp., both leaf disc and internodal segment explants were induced to regenerate whole plants. Generally a larger number of internodal segments than leaf discs regenerated, though the productivity of plantlets from a leaf disc was greater. Cultivars differed in their ability to regenerate. For blackcurrant, an internodal segment system was developed, and for strawberry, leaf disc regeneration was achieved. In Vaccinium spp. regeneration occurred using leaf material. Using these regeneration techniques, transformation experiments were initiated using both the Neomycin phosphotransferase II (NPT II) and the Beta- Glucuronidase (GUS) marker genes. Once putative transformants were selected on kanamycin medium, dot blot, fluorometric, histochemical, and Southern hybridisation assays were carried out. The GUS gene proved especially useful in the identification of transformants, providing a reliable, simple, quick and inexpensive marker gene. The development of transformation systems in soft fruit species has led to experiments being initiated for plant improvement, by the insertion of specific genes, notably genes conferring virus and disease resistance on Rubus and strawberry plants by Agrobacterium.

No description available.

Thesis--Paris-Sud, 1978. / Includes bibliographical references (p. 177-188).