Thesis (Ed.D.)--Boston University.

Thesis (Ed.D.)--Boston University. Note: cover says "3, 6, 17 insert at page 342"

Includes bibliographical references. / The roAp stars are high overtone pulsators characterised by strong global magnetic fields with effective field strength of between a few hundred to a few thousand Gauss. They also show overabundances in rare-earths and iron peak element. Multi-colour photometry of roAp stars shows that their pulsation amplitudes decrease more rapidly with increasing wavelength compared to other classically pulsating stars (e.g. Medupe & Kurtz 1998). In addition, this thesis shows that the pulsation phase can vary by up to 80° from Johnson U to V filters. In an attempt to explain the amplitude vs wavelength Kurtz & Medupe (1996) derived an analytic formula that indicated that the steep variation of the temperature eigenfunction with depth in the roAp atmospheres is the possible reason the rapid decrease of amplitude with wavelength. Even though Medupe & Kurtz (1998) showed that the analytical formula they derived in 1996 was a special form of a more general formula presented in Watson (1988), they found limitations in both formulae (see also Medupe, Christensen-Dalsgaard & Kurtz 1998 and chapter 3 of this thesis). This motivated me to attempt a numerical solution of the pulsation equations with consistent treatment of radiative transfer to investigate the multi-colour photometric data. In this thesis, we report on the multi-colour pulsation amplitudes and phases for HD 128898, HD 101065, HD 134214 and HD 137949. We also report on the campaign undertaken in 1996 and 1998 to determine whether the HR 3831 amplitude and phase vs wavelength changes with rotation. No variations with rotation are found. We find that the pulsation phase varies by up to about 80° from Johnson U to V in HD 101065 and HR 3831.

Bibliography: leaves 102-117 / An investigation was undertaken into the biology and genetics of Witches' Broom disease on Protea cynaroides. The investigation was approached in two ways: firstly, from a physiological and pathological angle and secondly at the genetic level. As very little is known about the causes of Witches' Broom disease on P. cynaroides, an attempt was made to identify a pathogen which could be held responsible for the disease. A number of plants were studied in the field and from these samples were taken and cultured on culture medium. Healthy P. cynaroides tissue was not established in tissue culture, while more success was obtained with teratoma tissue. Attempts were made to transmit the disease but these were unsuccessful. Four strains of Agrobacterium tumefaciens were unable to induce tumours on P. cynaroides seedlings. Sections of vascular tissue from teratoma and healthy tissue were viewed under the electron microscope but revealed no pathogen.

Includes bibliographic references. / Detailed variation studies of 23 species of the genus Eulopbia R. Br. (Orchidaceae) indicate that six taxa may be upheld as distinct species, together with two or possibly three intraspecific taxa, whose status can only be assessed on the basis of further field work. It is considered highly likely that the remaining 17 taxa should be relegated to the synonymy of the six species upheld, possibly together with a further eight doubtfully distinct species of which only descriptions were available.

The variation in abundance of the species G. bulloides and G. pachyderma with depth in a core of diatomaceous sediment off Walvis Bay, South West Africa, seems to correlate with the Little Ice Age that occurred during the late Holocene. Although similar research, using species and oxygen isotope variations has been completed in similar sediment off northern Venezuela and in the Santa Barbara Basin (Kipp and Towner 1975, Dunbar 1983), these are the first downcore foraminiferal species records of the South West African deposit. The abundances of some of the minor species, most notably G. ruber, show very distinctive peaks at apparently random intervals. It is thought that these sudden and short-lived increases could be linked to major El-Nino-type events. In order to measure accurately all the down-core foraminiferal changes, we had to develop a chemical solution which dissolves diatoms and organic material without damaging the foraminiferal skeletons. It consists of K2CO3, CaCl2, NaOH and H202 in particular concentrations. By using this technique, it is possible to prevent biassing of the foraminiferal signal during extraction of foraminifera from diatomaceous mud.

The stable isotope (SI) approach is widely used in ecological research to tackle problems such as delineating food web structure or tracing the migratory origins of various organisms. This thesis first tested the widely accepted assumption that SI ratios are fixed in an inert tissue, and then used the SI approach to infer the food web structure, from a marine top predator point of view, of a profoundly impacted marine ecosystem off southern Namibia. In bird research, it is assumed that SI ratios are fixed in feathers once they have completed their growth during moult. This assumption is crucial in determining where birds moult, and has been used to infer changes in the environment over time, as well as changes in the trophic levels of individuals. Recent comparisons of feathers collected from several penguin species during their annual moult have shown systematic differences between newly moulted and old feather SI ratios. I thus tested whether a change in SI ratios occurs as feathers age by comparing the carbon and nitrogen SI ratios of black and white feathers collected from captive, individually known African (Spheniscus demersus) and northern rockhopper (Eudyptes moseleyi) penguins at three occasions over a year. I found a clear trend for the rockhopper penguin feathers with new and old black feathers differing in their δ13C and δ15N values; this trend was not as clear for the African penguins. I then tested factors related to feather wear as a possible mechanism for differences in SI ratios between new and old feathers; these factors were feather reflectance and microstructure. In both penguin species, old black feathers reflected more light, and had a larger proportion of their barbs without barbules near their tips compared to new feathers. Feather wear may result in melanin leakage, which may explain the observed trends in the SI ratios between new and old pigmented penguin feathers. Differences in SI values were observed between species and may be a result of facility at which the penguins were housed, where one facility was exposed to more sunlight than the other, rather than the differences being a result of species. Although the differences observed were subtle, the state of feather wear (i.e. timing of feather collection within the moult cycle) should be considered in order to make accurate ecological inferences based on their SI ratios. Further research is needed to fully understand the phenomenon and to test whether the same process affects pigmented feathers of flying birds. I then used SI ratios to update our knowledge of resource partitioning among a marine top predator community in southern Africa, and to infer the marine food web structure in a Marine Protected Area off the southern Namibian coast. The Namibian Islands Marine Protected Area supports the most important breeding population of bank cormorants (Phalacrocorax neglectus; Endangered), and historically was important for two other Endangered seabirds: African penguins and Cape gannets (Morus capensis). Non-threatened marine top predators studied in the system were: greater crested terns (Thalasseus bergii) and Cape fur seals (Arctocephalus pusillus). In the last 50 years, shelf waters off southern Namibia have been overfished, resulting in an altered marine ecosystem; shoaling fish have been replaced by less nutrient-rich species such as hake (Merluccius spp.), jellyfish, and salps. Previous studies of the region’s food webs were based on traditional dietary analyses such as stomach content and scat analysis which provide short-term diet information. Here, I measured carbon and nitrogen SI ratios in several tissues of the marine top predators, and those of their potential prey species, to infer the marine food web for this region, and to complement short-term data obtained from traditional dietary analyses. Among the predator guild, Cape gannets had the lowest δ13C and δ15N values, indicating that they fed the farthest offshore and at the lowest trophic level both during the breeding season (from whole blood) and moulting period (from feathers). As expected, bank cormorant δ13C values indicated that they fed more benthically than the other predators, and Cape fur seals fed at the highest trophic level. African penguin tissue SI ratios were intermediate between those of Cape gannets and Cape fur seals. Greater crested terns exhibited the highest δ13C values, suggesting that they fed closest to shore. Bayesian mixing models used with species-specific discrimination factors (when available) revealed some resource partitioning among the marine top predators in this region but many made use of similar resources; sardine (Sardinops sagax), squid (Loligo reynaudii), and rock lobster (Jasus lalandii). This work highlights the most likely prey items used by marine top predators outside of the breeding period, and provides new insights into the food web of this region. Prior to the start of industrial fishing, marine top predators in this region mainly ate sardine and anchovy. Despite the collapse of these species’ populations in the 1970s, marine top predators currently still make use of these prey resources, which concurs with the results found from stomach content and scat analyses. In this thesis, I have shown experimentally that SI ratios vary slightly as feathers age in penguins. I have also used the SI technique in an ecological context to add to the knowledge on the diet of marine top predators of an overfished ecosystem. Overall I have shown how the SI approach can add to our understanding of trophic ecology, and also how the method is dependent on accurate SI inputs in order to make accurate dietary inferences.

This thesis reports on an exploratory investigation into the immobilization of homogeneous olefin polymerisation cataiysts on dendritic supports...; The dendrimers . employed were essentially allylaryl-ether type systems and consisted of both dendritic wedges .and compltte dendrimers.. These materials were used as supports for homogeneous irconocene catalysts. The zirconium moiety was attached to the surface of the dendrimer through the ally! functionality. This was dope via hydrozirconation of the pendant C=C of the ally! groups on the periphery of the dendrimers. A number of model compounds such as Cp2Zr{(CH2)4iPh}(Cl) and Cp2Zr{(CH2)30Ph}(Cl) were also prepared and their pronerties compared with the dendritic catalyst precursors .. The supported catalyst species together with. analogous mononuclear model compounds were activated using a range of activating agents. Included amongst these activating agents are, silver salts of the type AgX, (where X = OS02CF3, Cl04 and BP); the perfluoroaryl borane, B(C6Fs)3 and MAO. The activated complexes were assessed for their. polymerisation ability towards ethylene. The dendrimers were characterised using techniques such as FTIR spectroscopy, 1 H NMR and 13C NMR spectroscopy, GC mass spectrometry and microanalysis. All new organometallic compounds produced were characterised using techniques such as 1 H NMR and 13C NMR spectroscopy. The polymers prepared were characterised by Gel Permeation Chromatography (GPC), Fourier Transform InfraRed (FTIR), 1 H and 13C NMR spectroscopy, Differential Scanning Calorimetry (DSC) as well as Scanning Electron Microscopy (SEM).

This study investigated the nature and effects of a number of human disturbances' on the spatial distribution and dynamics of sand forest in and around Tembe Elephant Park (TEP). Maputaland, South Africa The first disturbance investigated was that of fire and its role in determining the spatial patterning of sand forest at the landscape level. Secondly human disturbance was investigated i11 the fom1 of utilisation of forest products and the effects of this on the d}11amics of sand forest as a ·whole as well as on individual species. Sand forest in this area was restricted to deep \ilite sands and occurred as patchy strips mlh a north-south onen.latio11 within a mosaic of grassland and wooded grassland. An abrupt forest grassland boundary was evident in these patches. Two alternatives were investigated as causal factors of the observed patterning: firstly, that sand forest boundaries could be set by underlying soil conditions and thus mosaic components were static; secondly that patches could be remnants of larger tracts of forest, such were fragmented or maintained by past disturbances such as fires. In the absence of these disturbances, patcl1es should be able to expand. II was found that mosaic components did not differ significantly between altin1de (used as a surrogate for s1il moisture) which could imply that distribution patlems are not solely governed by Sil determinants. Prevailing and direction and intensity over tl1e dry (fire) season were fol.llld lo be strongly southerly and n1,rtherly and this is considered 10 be the direction for fratllentation and maintooance of north-south orientated strips. Evidence of bush clumping in ti1e grassland occurred around Dum .whletteri nuclei. 'Ibis process appears lo be slow as. although evidence of thickening in lire-protected areas occurred in aerial photographs over time, it is gradual. The removal of lire in the grassland should lead to the development of sand forest clumps. llie dynamics of sand forest in terms of composition and size-class distributio11s was compared between a marginally disturbed site used by people south of the reserve and an unharvested site within the reserve from which elephants had been excluded. Overall dynamics were found be similar between the sites although certain species showed significant higher densities of seedlings in disturbed sites than in protected sites. Canopy dominants were found to have low levels of seedling recruitment and an analysis of the 'grain· of sand forest showed it to ha'e a course grain relative to other forest types in KwaZulu-Natal. TI1ese finding imply tliat relatively large gaps may be needed within sand forest in order to allow for regeneration of canopy species. four types of human use of sand forest products were investigated in the context of their sustainability medicinal, cmfl and fuehvood sales and hut building. A number of plant products were harvested for the medicinal trade. Of the species investigated, the bark of Cassipourea mossambicensis appeared 10 be harvested unsustainably and this species has been seriously impacted outside the reserve. Crafters targeted Clei.rtanlhus schlecteri for the manufacture of bowls for sale along the roadside. A model of a population of these trees adjacem to the main crafi stalls revealed that reproductively viable adult troos would be depleted within 25 years. This indicated lhat era.ft manufacture was unsustainable. Newtonia hifdebrandtii was the most commonly sold fuelwood along lhe roadside. Sellers mainly used deadwood and turnover rates of bundles sold was .1remely slow. TI1us al the time of the study (1997) fuehvood sale appeared to be sustainable. The main framework of huts built was composed of thin understorey and subcanopy sand forest woody species. Strong evidence of resprouting occurred in these species and species composition of huts was similar to an earlier study underrnken in J 982. This suggested lhat impact 011 forest dynamics was not substantial for this type of use. In conclusion, management and utilisation of sand forest needs to be within the dynamics and regeneration capacity of this forest and its constituent species. t'or example l predict that an absence of fire will lead to tltickcning of woody 'egetation and a decline of grasslands. Canopy species may require large gaps for regeneration to take place. vse of plant products for commercial purposes (e.g. C. mossambicensis for bark and C. schfecferi for crafts) is unsustainable given the slow growth rates of the forest. Allhough sale of tirewood appeared to be sustain.able at the time of the study, with increasing demand rrom tourists. live trees may increasingly be used and this industry could soon be problematic. U1ilisation of trees for domestic purposes appeared lo be sustainable as sand forest species were resilient resprouters. However, commercial use of this forest type was felt to be unsustainable. Species found to be most at risk were C mossambicensis, C. schlecteri, and N. hildebrandtii.

Cataclysmic Variable stars (CVs) are excellent examples of multi-wavelength objects. They have been detected at all frequencies, from radio to γ-ray. However, while large sets of multi-wavelength data exist, few studies have been performed on the broad spectral energy distributions (SEDs) of these objects. I perform such a study here, and present the largest collection of broad CV SEDs yet constructed. I define a sample of well-studied, nearby cataclysmic variables of several different CV classes, and gather existing multi-wavelength data over a broad frequency domain. I identify the contributions to the spectrum and compare the spectra between classes. I then use the SEDs to constrain the mass accretion rates by modelling the accretion disc as a series of concentric annuli, each of which radiates as a black body with effective temperature appropriate to the radius of that annulus. The resulting bestfit accretion rates are compared to published values of mass accretion rates for the systems. My method of modelling the emission from a CV accretion disc as a sum of blackbodies produces accretion rate estimates that are reliable to within a factor of 5 for those systems that have steady state discs. However, the data are significantly underfitted by these simple models in all cases. While this method is a useful tool for modelling spectra and finding accretion rates, it cannot be applied to a small amount of data quickly and simply. It must be performed on as large a dataset as possible, and other contributions to the spectrum must be accounted for. As a result, an attempt to obtain accretion rate estimates for a larger, volume-limited sample of non-magnetic CVs was unsuccessful. I also investigate the ratios of flux densities in different wavebands. My data reveal two previously reported correlations, namely that shorter period CVs have higher X-ray to optical flux ratios and that magnetic CVs are X-ray bright compared to non-magnetic CVs. In addition, I find that polars are radio bright compared to IPs and non-magnetic CVs, and illustrate this, for the first time.