Thesis (M. Sc.)--University of Hong Kong, 2005.

Thesis (M. Journ.)--University of Hong Kong, 2001. / Includes bibliographical references (leaves i-vii).

Thesis (M.Journ.)--University of Hong Kong, 2001. / Includes bibliographical references (leaves i-vii). Also available in print.

Thesis (Ph.D.)--Aberdeen University, 2009. / Title from web page (viewed on Feb. 23, 2010). Includes bibliographical references.

Acacia cyclops A. Cunn. ex G. Don is an invasive alien plant that invades fynbos and coastal dunes. Several acacia species are grown commercially in South Africa and this has limited biocontrol agents to those that reduce only reproductive capacity. Dasineura dielsii was released in 2002 as a biocontrol agent for A. cyclops. This gall-forming midge destroys inflorescences and prevents seedpods from forming, but allows continued harvesting. Insects overwinter as larvae within their galls. This study examined the levels of parasitism experienced by dormant D. dielsii larvae, as well as the trigger that causes them to break dormancy. Gall clusters were sampled over autumn and early winter, and were dissected to determine occupancy. The effects of temperature and light on dormant larvae were also examined. Observations showed that as the season progressed, more D. dielsii larvae entered dormancy. At the same time, a greater proportion of dormant larvae were parasitized. Parasitism was highest, at 18.5%, at the end of the study period, but this level of parasitism is not enough to reduce the biocontrol power of D. dielsii. A positive relationship between mass of gall clusters and the number of galls they contain was established. Larval dormancy could not be artificially broken by either light or temperature, and it appears that neither factor alone can trigger a break in dormancy.

Game farming is becoming more popular in southern Africa and the introduction of large indigenous ungulates into confined enclosed areas could alter plant communities and ecosystem processes. This is of particular concern in semi-arid rangelands of the Succulent Karoo where the evolutionary history of grazing is not clear and the compatibility of large herbivores in confined areas remains to be demonstrated. The establishment of Sanbona Wildlife Reserve, a 54 000 hectare private game reserve in the Little Karoo, which converted from livestock farming to game farming, allows an opportunity to study the vegetation dynamics in a confined plant-herbivore system. This study investigates the current community structure and the changes in the floral composition and vegetation structure of enclosed and comparable sites subjected to grazing by large herbivores after twelve years (2004-2015). It also determines the relative effect of grazing and rainfall on the observed patterns. Finally, the implications of these findings for management are discussed. Data from drop-point surveys in fenced (exclosure) and unfenced (grazed) plots in the dominant vegetation types as well as annual and seasonal rainfall totals, stocking rates of herbivores and annual game census information, were analysed. These were used in multivariate ordination techniques, regressions and linear mixed-effects models to determine the communities and their relationship with herbivory and rainfall over time and identify a set of indicator species. The annual game census information was used to determine areas of herbivore preference or 'hotspots' and for the identification of highly-utilised areas. Cluster analysis, using the flexible beta method in PC-Ord, was used to determine the current plant communities. Non-metric multidimensional scaling ordination (NMS) was used to determine the relationship of these communities with the environmental variables and illustrate the trajectories in floristic data. Species were also assigned to plant growth forms and examined as communities and growth form types. The Bray-Curtis distance measures were used to investigate the difference between each treatment over time, within each vegetation community and between treatments. Finally, the effects of rainfall and herbivory were examined using linear mixed-effects models of change over time vs the various potential determinants of change using lmer functions in R. Four communities were identified. These communities corresponded well with to the vegetation type descriptions for Western Little Karoo, Little Karoo Quartz Vygieveld and Renosterveld as described in the National Vegetation Map of Mucina and Rutherford (2006). However, the Western Little Karoo was too broad and two communities were recognised within this vegetation type. The finer scale mapping by Vlok et al. (2005) corresponded relatively well to these communities. Results showed an increase in species richness, abundance and cover over time, with the ungrazed plots experiencing more change than the plots exposed to grazing. Most growth forms exhibited an increase in cover, although low leaf succulents declined in both grazed and ungrazed plots. Medium evergreen shrubs declined in the exclosures and stem succulents declined in the grazed plots. The effects were found in both grazed and ungrazed treatments. In addition, many species which declined in abundance were unpalatable or toxic to herbivores. Because of this, the decline in cover of such species was not attributed to grazing, but was instead interpreted as being a response to other disturbance mechanisms, to competitive displacement and to rainfall events. The low stocking rates in the first five years of the study resulted in there being very little difference evident between the treatments. However, once stocking rates increased from 2008, both species richness and cover increased more rapidly in the ungrazed plots, compared to the grazed plots. An increase in palatable and unpalatable species was observed within both ungrazed and grazed plots indicating that grazing did not change the proportion of palatability classes. However, specific plots in the areas of high animal utilisation were more affected as indicated by the response of cover, species richness and palatable species in these specific plots. This suggests that the grazing pressure may be too high within those areas. The linear mixed-effect model supports the argument that grazing pressure is the dominant driver of the community change within grazed plots. Similarly, the results show that rainfall is the primary driver of the vegetation community in the absence of grazing. Timing, amount and intensity of rainfall can mask these impacts. Thus, the contribution of grazing to vegetation change can probably only be detected by tracking the trends over decades or longer. The use of indicators as a management tool is well documented. In order to identify indicators, a theoretical framework for determining indicators species in the different vegetation communities was created. This was based on the correlation between species abundance and sampling period in the different treatments, which identified species that have significantly increased or decreased over time as a result of the change in land use. Species identified as potential indicators were selected on the basis based on their abundance and ranged in lifespans and palatability. The indicators chosen need to be monitored into the future to confirm their utility as indicators. A small but significant difference between grazed and ungrazed plots suggest that herbivore impact is apparent. Identifying this trend indicates that the monitoring programme is providing a useful tool for assessing the impact of herbivores on an ongoing basis. The recovery process following the withdrawal of domestic livestock from Sanbona was much slower in the grazed plots than in the protected plots. Therefore, for the continued recovery of the vegetation to occur and for there to be a sustained increase in cover, active management of animal numbers needs to take place. The results from this study can contribute to future management decisions on the reserve and form a basis for future analyses.

The pollinator spectrum, diversity of flower colour, natural pollination and fruit set levels were examined in Scottish <i>L. borealis </i>patches.  Flowers were visited by a diverse spectrum of insect pollinators but, despite high natural pollination levels, fruit set was very low in the majority of patches examined.  The breeding system of <i>L. borealis </i>was confirmed as highly self-incompatible and limited compatible mate availability was identified as the cause of reproductive failure.  Due to the limited pollen dispersal capabilities of flies which dominate the pollinator spectrum, isolation from pollen exchange between compatible mates began at very short distances.  A diversity of compatible mates situated within close proximity (&lt;6 m) was shown to be the key requirement for high natural fruiting success in <i>L. borealis.  </i>Cross-pollination between neighbouring <i>L. borealis </i>patches resulted in high fruiting success suggesting that transplantation of compatible mates into isolated patches could restore their reproductive success. Natural pollination levels were generally high on Colonsay and a very low, but consistent, level of capsule and seed set was recorded in most years.  Capsules produced by Colonsay plants contained very low quantities of seed (&lt;100/capsule) and high numbers of empty testa, and seed set was extremely low (&lt;1%).  Severe reproductive failure was evident and the cause was attributed to self-incompatibility and/or inbreeding depression.  Seed set was also very low for all of the crosses made within Scotland, even between the two distinct genetic groups.  The high levels of seed set achieved following crossing of Scottish and North American plants suggest that restoration of reproductive success in Scotland may require the introduction of some North American genes.

published_or_final_version / Environmental Management / Master / Master of Science in Environmental Management

<i>Melampyrum sylvaticum</i> (small cow-wheat) is a hemiparasitic annual of boreal-montane regions of Europe.  The Species Action Plan recommended that in addition to protecting extant populations, by 2010 there should be an additional five populations that have been created with the aim of enhancing greater genetic diversity of the species.  Consequently this project was set up in order to provide the ecological knowledge required to meet such targets. There are various management options available to conservationists looking to prevent <i>Melampyrum sylvaticum’s</i> extinction from the UK but from the results of this project it is clear that some methods have drawbacks that should preclude their use.  Population augmentation with seeds from other populations is not advised due to the risk of genetic ‘swamping’ or outbreeding depression.  Seed amplification would avoid these problems but may introduce different complications by artificially promoting certain genotypes within a population. Population expansion by mimicking ant dispersal is recommended as a way of minimizing density dependent mortality in larger populations but is not suitable in smaller populations. Seed translocation to unoccupied sites is therefore, the best option but the exact details of seed-sourcing and sowing should be guided by the results of the Species Recovery Project in order to avoid predicted limitations. The long-term outlook for <i>M. sylvaticum</i> will depend entirely on whether populations can be created that operate as part of a functioning ecosystem (including pollinating and seed-dispersing insects) with enough demographic and genetic stability to survive predicted climate change.

Little is known about the demographics of the genus Acacia in Africa, despite its prominence and the economic and environmental importance of this group. The demographics and species composition of stands of four different Acacia species was investigated in the Hluhluwe-Umfolozi Game Reserve in northern Kwazulu-Natal in order to determine whether stands were self-replacing, or whether other Acacia species were invading them. Soil and stand density as well as the density and composition of the grass layer were investigated in order to determine whether any of these affected Acacia demography and species composition. In three out of four cases, the species whose large size class dominated the stand (known as the "stand species") was found to have a strongly bimodal size class distribution, skewed primarily towards the large size class and secondarily towards the small size class. Other Acacia species in the stands, with few or no large individuals present ("nonstand species") tended to have unimodal size class distributions, skewed primarily towards the small size class and secondarily towards the medium size class. Based on the proportion of small to large individuals, non-stand species are more likely to increase in overall numbers in future, suggesting that the species composition of the stands may be in a state of flux. Few relationships were found between stand density, grass density, grass composition and numbers of small and medium acacias. This could be as a result of differentiation between Acacia species, meaning that they cannot be analysed collectively. Furthermore, sample sizes of individual species may have been too small to analyse individually. Alternatively, it could mean that none of these factors have a significant effect on one another and that other explanations need to be found for the demographics of this genus.