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

Thesis (M.Sc.)--York University, 2004. / Typescript. Includes bibliographical references (leaves 81-88). Also available on the Internet. MODE OF ACCESS via web browser by entering the following URL: http://wwwlib.umi.com/cr/yorku/fullcit?pMQ99335

Thesis (M.A.)--Ohio University, June, 2002. / Title from PDF t.p.

Thesis (Ph. D.)--Oregon State University, 2008. / Printout. Includes bibliographical references (leaves 173-178). Also available on the World Wide Web.

Thesis (M.S.) University of Missouri-Columbia, 2006. / The entire dissertation/thesis text is included in the research.pdf file; the official abstract appears in the short.pdf file (which also appears in the research.pdf); a non-technical general description, or public abstract, appears in the public.pdf file. Title from title screen of research.pdf file (viewed on August 27, 2007) Includes bibliographical references.

Thesis (Ph.D.)--Murdoch University, 2008. / Thesis submitted to the Faculty of Health Sciences. Includes bibliographical references (leaves 307-371)

abstract: The southwestern willow flycatcher (Empidonax traillii extimus) is listed as an endangered species throughout its range in the southwestern United States. Little is known about its sub-population spatial structure and how this impacts its population viability. In conjunction with being listed as endangered, a recovery plan was produced by the US Fish and Wildlife Service, with recovery units (sub-populations) roughly based on major river drainages. In the interest of examining this configuration of sub-populations and their impact on the measured population viability, I applied a multivariate auto-regressive state-space model to a spatially extensive time series of abundance data for the southwestern willow flycatcher over the period spanning 1995-2010 estimating critical growth parameters, correlation in environmental stochasticity or "synchronicity" between sub-populations (recovery units) and extinction risk of the sub-populations and the whole. The model estimates two parameters, the mean and variance of annual growth rate. Of the models I tested, I found the strongest support for a population model in which three of the recovery units were grouped (the Lower Colorado, Gila Basin, and Rio Grande recovery units) while keeping all others separate. This configuration has 6.6 times more support for the observed data than a configuration assigning each recovery unit to a separate sub-population, which is how they are circumscribed in the recovery plan. Given the best model, the mean growth rate is -0.0234 (CI95 -0.0939, 0.0412) with a variance of 0.0597 (CI95 0.0115, 0.1134). This growth rate is not significantly different from zero and this is reflected in the low potential for quasi-extinction. The cumulative probability of the population experiencing at least an 80% decline from current levels within 15 years for some sub-populations were much higher (range: 0.129-0.396 for an 80% decline). These results suggest that the rangewide population has a low risk of extinction in the next 15 years and that the formal recovery units specified by the original recovery plan do not correspond to proper sub-population units as defined by population synchrony. / Dissertation/Thesis / M.S. Biology 2012

The general aim of this thesis was to examine the merits of ex-situ vs. in-situ strategies for the conservation of the endangered freshwater pearl mussel, Margaritifera margaritifera, and to investigate the relationship of the larval parasitic stages of the mussel (glochidia) with the salmonid hosts. To this end, I critically reviewed the literature on conservation of freshwater mussels, developed methods for quantifying the behaviour and activity patterns of adult mussels in captivity, experimentally studied host specificity, and quantified the physiological and behavioural effects of glochidia upon salmonid hosts. The results indicate that the conservation of the freshwater pearl mussel is probably best addressed at the watershed scale, and will benefit from a combination of ex-situ and in-situ techniques, as well as from a more critical assessment of findings, many of which are only reported in the grey literature. Empirical, peer-reviewed data are badly needed to inform current conservation efforts. Novel Hall-effect magnetic sensors were used to quantify and characterise discrete mussel behaviours without adversely affecting the welfare or survival of adult mussels, and these hold considerable potential for determining optimal rearing conditions for ex-situ conservation. Arctic charr was shown to be a potentially suitable host for M. margaritifera, and occupied an intermediate position in host suitability between brown trout and Atlantic salmon. Physiological impacts of glochidia upon brown trout included swelling of secondary lamellae and spleen enlargement, but the latter tended to be slight and was restricted to 1 month post-exposure. Glochidia encystment had no significant effect on blood haematocrit, respiratory performance, or cryptic colouration of brown trout hosts. The behavioural effects were more subtle and glochidiosis made brown trout more risk-averse and less willing to explore a novel habitat, without affecting the host's ability to chemically recognise and avoid cues from a predator. Overall, the results of this thesis indicate that the impacts of glochidia upon salmonid hosts are probably slight and temporally variable, and may perhaps lead to increased host survival, which would support the symbiosis-protocooperation theory of glochidia-salmonid interaction.

With the increase in the number of fast-moving vehicles and the simultaneous development of road building technology, roadside bird mortality has become an increasingly important environmental issue that has the potential to do serious damage to already vulnerable bird populations. This project was therefore initiated after an alarming number of owls were found dead along the N17 and R550 roads in the rural areas between Springs and Devon in the East Rand highveld of Gauteng Province. Five hundred and fifty four owls of four species, namely Marsh Owls, the Red Data listed Grass Owl, Barn Owl and the Spotted Eagle Owl were collected on the stretches of the R550 and N17 during the period between October 2001 and September 2003. They accounted for 53.6%, 27.4%, 17.5% and 1.3% of the mortalities respectively. Unidentified species of owls accounted for the remaining 0.2%. It was found that the monthly mortality rates of the birds varied throughout the year, with the greatest losses being suffered during July, as opposed to relatively lower mortalities occurring during the warmer months. All carcasses were collected, their GPS locations plotted on a map, and ‘hotspot areas’ identified as places of highest incidences of mortalities of the owls. Vehicle-induced mortalities are then discussed in relation to these hotspots, in terms of vegetation and habitat descriptions, daily vehicle counts along the route, as well as fixtures found along the route. It was found that traffic density was indirectly proportional to owl mortalities, with higher traffic speeds definitely having an increasingly detrimental effect on the owl mortalities. Weather conditions also play a role in mortality counts, with the mortalities being significantly negatively correlated to rainfall. Moon phases were also related to the times of highest mortalities, however, this factor did not play a significant role in influencing road mortalities. Another factor that was studied was the influence of differing tarmac road surface temperatures as opposed to gravel road verge surface temperatures, and how these temperatures differed from the ambient temperature. It was found that there was not a big enough difference in the temperatures that would warrant (the previously thought notion) that the owls were attracted to the roads at night to gain heat. Gravel roads had very low incidences of owl mortalities with the highest mortalities recorded along tarmac roads that are bordered by open grasslands or cattle grazing paddocks. It was shown that tarmac roads, bordered by croplands, had a lesser effect on the owl mortalities. Another factor influencing the road mortalities of the owls is grain that is spilled on the road during transport. This initially seemed to be the major factor in attracting granivorous rodents to the roads, and in turn, attracting the owls to prey on them. Rodents identified from recovered pellets and the stomach contents of dead owls confirmed the fact that the majority of rodent prey items were indeed granivorous species, namely Mastomys natalensis and Rhabdomys pumilio. This is in disagreement with previous studies that indicated that a large proportion of the prey species of the Grass Owl was Otomys irroratus, a grass-eating species, even though a rodent trapping study to determine prey abundance within the area indicated a healthy population of O. irroratus. Further studies into prey items of the owls that were dissected indicated that the majority of the prey items recovered were not caught directly on the road as it was already partially digested, suggesting that the prey was caught prior to the owl being killed. From the pellet analyses, other prey items were also found to form part of the owls’ diets. It was found that insects formed largely the diet of the Marsh Owl during the spring, summer and autumn months, with them resorting to smaller rodents during the winter months. Spotted Eagle Owls also preyed almost exclusively on insects. Grass Owls, on the other hand, preyed on small mammals exclusively, with the very rare exception of some insects also being taken. Of all of the dead owls recovered on the roads, post mortems were carried out on only 78 of the carcasses. All of the dead owls examined were in good health prior to death. Various morphometrics of the examined owls were noted. Comparisons of body mass showed that females were larger than males for most species. This was also found for most other measurements as well. It was also found that, according to body mass comparisons, Barn Owls and Marsh Owls were significantly similar. Conducting ANOVA analysis on other morphometrics to determine gender differences, it was found that Grass Owl males were significantly different to females in terms of body mass and length. Marsh Owl males were significantly different to females in terms of body mass and tail length; with Barn Owl males being significantly different in terms of tarsus length to females. Except for Spotted Eagle Owl tarsus lengths all other measurements were in favour of females being significantly larger. These findings were also confirmed when applying the Dimorphism Index to all morphometrics measurements, especially body mass. The degree of parasite infestation was also studied during post mortem examinations. Very few cestodes and nematodes were found, with too few to have an effect on the overall health of the birds prior to death. The vegetation type was studied at transects that coincided with hotspot and non-hotspot sites. Using the PRIMER statistical software package, hotspot sites were found to have highest plant cover and diversity, whereas nonhotspot sites showed lowest plant cover and diversity, generally dominated by Hyparrhenia hirta. During these studies, the degree of available nesting habitat was determined and nesting sites were identified, using the ropedragging technique to flush out roosting and nesting owls that would otherwise have been impossible to find in the thick cover. It was found that Grass Owls preferred a habitat rich in thick grass cover that was relatively high (0.75 m–1 m). It was found that the grass species preferred by these owls were Eragrostis curvula, Paspalum sp., Setaria sp., Sporobolis sp., with few other small herbaceous plants. Marsh Owls, on the other hand, seemed not to be too partial regarding roosting and nesting sites, with them roosting and breeding in more mixed vegetation grasslands that had sparser cover, not reaching the height of the grass cover typical of the habitat preferred by the Grass Owls. Opposed to this were the non-hotspot vegetation sites. These sites were found to have vegetation cover unsuited to both the Marsh and Grass owls, with mixed H. hirta grassland not forming the dense cover, or the height, needed by those two species of owls. Foraging owls were also observed, with the vegetation type in the immediate vicinity noted. Vegetation types similar to breeding areas were noted in these foraging areas. Habitat preferences as well as breeding performance were noted for both grassland species of owls, and found to be directly related to land usage in terms of varying agricultural practices and regimes. Fallow, undisturbed lands were found to be highly productive for the owls. Lands planted with Eragrostis sp. were also found to be very productive, but only if left undisturbed for a period of time sufficient to allow the grassland owls to colonise it. Maize-planted fields were found to be utilised only as foraging fields and no breeding of owls was found to take place close to these fields. After extensive nest searching, it was found that both Marsh and Grass owls were breeding from late March to early June, with the Barn Owls breeding in October and again in March. Grass Owls occur in the study area because of the presence of a natural corridor of suitable habitat that runs parallel with the Blesbokspruit. This favourable habitat of the study area is thus conducive to high population density of grassland owls utilising these uncultivated patches of dense and tall vegetation. The high incidence of mortalities on the road in the study area is due to the concomitant high population densities. This healthy population seems to be sustaining the losses occurring on roads. Owls also seem to be gathering in larger numbers in hotspot zones because of the easy available prey, which are attracted to these high productive areas. Agricultural practices in the area lead to the spillage of grain on the road during transportation. Potential prey species foraging on the roads expose themselves to the nocturnal hunters offering an easy dinner. This process leaves these owls vulnerable to vehicle collisions. The overall population size may be larger than previously thought, not with standing the high mortalities already recorded. The small patches of viable habitat in the study area remains suitable for the breeding of the two grassland owl species allowing for such high densities to occur in the area. The Grass Owl, nonetheless, remains severely threatened as it already occurs as a high priority species for conservation concern in the Gauteng Province. This study provides the first assessment of this owl species of this scale in South Africa and this will ultimately promote the long-term survival of these owls. / Dr. V. Wepener

Coryphantha scheeri var. robustispina (Schott) L. Benson, the Pima pineapple cactus, herein referred to as C. scheeri, was declared an endangered species in 1993 and only occurs in a limited range in Arizona, USA and northern Sonora, Mexico between 2,300'-4,500' asl. Development within the range of C. scheeri threatens individuals, but transplanting to conserve them while allowing for development has been considered to be ineffective for conservation due to low post-transplant survival rates in past studies. The construction of a natural gas pipeline provided the opportunity to conduct a transplant experiment on 82 individual C. scheeri transplanted in July and August 2014. The plants were randomized into one of four transplant methods: bare-root with supplemental water, bare-root without supplemental water, soil-and-plant moved with supplemental water, and soil-and-plant moved without supplemental water. Higher than average precipitation occurred during the 2014 monsoon season including after transplanting. A subset of the transplanted C. scheeri (n=17) were transplanted back onto the pipeline after pipeline construction was completed. Survival rates were monitored through December 2016 and compared to undisturbed C. scheeri near the pipeline ROW and those on other sites. For the plants transplanted once, no significant effect of moving the plants with soil compared to no soil (X2 = 2.9, p = 0.09), no significant effect of adding water at the time of transplant compared to not adding water (X2 = 1.2, p = 0.26), and no significant interaction among treatments (X2 = 0.06, p = 0.81) was observed. For plants transplanted twice, a significant effect of moving the plants with the soil compared to no soil (X2 = 5.0, p = 0.02) was found, while due to the random selection of plants to be transplanted twice there was too little data to adequately test other comparisons. There was no significant difference in mortality between the transplanted once (27% mortality) and the transplanted twice (31% mortality) treatments (p = 0.78), but there was a significant difference between transplanted and non-transplanted plants (2% mortality in non-transplanted plants; p < 0.05). Soil series did not appear correlated with mortality. Plants in good condition (scored 4 or 5 on scale of 0-5) at the time of transplanting had low mortality rates (16%) while plants scored 3 or lower had high mortality rates (60%) but deaths did not occur immediately after transplanting: 5 died after 8 or 9 months, 4 after 13-16 months, and 9 after 23 or 24 months. The majority of the deaths occurred after numerous months of declining in condition but six plants died suddenly. Good condition plants were more likely to flower than those in poor condition. Transplanting appears to conserve some of the C. scheeri population which would have otherwise been lost to development.