The six chapters of this thesis contain the results of four years spent in the investigation of old open clusters at the SAAO. Each chapter is a preprint of a paper which has been submitted or accepted for publication and is therefore an independent entity with its own list of acknowledgements and references. Since each chapter is intelligible without continual reference to its fellow chapters, considerable overlap has sometimes occurred. As a thesis, the end product is of necessity a little disjointed. This approach has been adopted because 'the publication of the results has higher priority than the presentation of a thesis. The style is as compact as seems consistent with reasonable clarity and I hope that reading it will not present too great a problem of digestion to those interested in the evolution of solar-type stars.

Includes bibliographical references. / This thesis is based on the Edinburgh-Gape Blue Object Survey and includes a general review of blue star sky surveys, a detailed discussion of the selection techniques, the present status of the survey, and follow-up photometry and spectroscopy of the cataclysmic variable stars (GVs) found by the survey. Because of the north-south imbalance in the number of blue objects, the Edinburgh-Gape Blue Object survey was designed to discover previously unknown blue objects with B≤~18 at high galactic latitude in the southern hemisphere. At present, all the planned fields (61 fields) in the North Galactic Pole region and 26 fields in the South Galactic Pole regions are complete, covering 2488 square degrees. The preliminary results show that the survey has an internal completeness of ~94% for objects with U-B bluer than the cutoff --0.3, and selects objects with an accuracy in photographic U-B of 0.16 mag. UBV measurements and spectra have been obtained for ~2000 objects. Half of these objects are hot subdwarfs which comprise the sdO, sdB, and sdOB stars. The white dwarfs account for about 17% of the total. About 18% of the objects are horizontal branch B and normal B stars. Cataclysmic variable stars and quasars comprise only a small fraction, 2 and 6 percent, respectively. Less than one percent are galaxies. The remaining ~5% have unknown spectral classes at present. So far, 25 cataclysmic variable stars have been found by the survey. The space density was estimated to be 1.6xl0-⁶ pc-³. Among these cataclysmic variables, 6 dwarf novae and 8 nova-like variables are new discoveries; detailed study of their photometric and spectroscopic characteristics are discussed in Chapters 6-10. One of the stars discovered, a close binary with large reflection effect, is a predecessor of the cataclysmic variables. Napier's (1968) method for the reflection effect in close binaries was used to model the light curve in order to solve for the physical properties of the system. Photometry, spectroscopy, and model fits for this system are discussed in Chapter 5.

Context: Stellar population modelling is a popular technique that has been extensively applied to main sequence galaxies. Yet starburst galaxies and Luminous InfraRed Galaxies (LIRGs) have, so far, not been studied as much using the method. LIRGs in the local universe are known to be highly interacting galaxies with strong star formation in obscured environments. Still, LIRGs also have diversity in terms of morphology and mode and location of star formation. Aim: This thesis investigates the stellar population properties of a group of 52 starbursts and luminous infrared galaxies (LIRGs) in the local universe that forms part of the SUperNovae and starBurst in the InfraReD (SUNBIRD) survey. The galaxies in a distance range of 3.5 < Dl < 280 Mpc and infrared luminosity of 10.30 < LIR < 11.91 L were observed with the Southern African Large Telescope in long-slit spectroscopy mode. Method: The stellar populations of the galaxies are derived by fitting Bruzual &amp; Charlot (2003) templates to the reduced spectra using STARLIGHT software with a Monte Carlo method implemented to recover uncertainties on age, metallicity and extinction. The derived stellar population models are then subtracted from the observed spectra to produce emission spectra from which emission line fluxes are measured. Both integrated spectra and spatially resolved apertures are extracted to be analysed in this work. Results: The light-weighted and mass weighted age of the sample is found to be 160 Myr and 7.2 Gyr respectively. The star formation history of the sample shows a rise of activity in the past ∼ 50 Myr and with a jump of an order of magnitude in the past 3 Myr. Analysis of the stellar metallicity hints at inflow of pristine gas, which decreases the observed metallicity content as well as ignites SF-activity. Analysis of the oxygen abundances shows that while LIRGs and SF galaxies are under abundant, their under-abundance may have previously been over-estimated as compared to main sequence galaxies. The radial age profile of the sample is flat, similar to that of late-type Sd galaxies. Interaction is found to cause a drop in the age of apertures although the post-merging stages shows continued star forming activity in the nuclear region. The stellar metallicity gradient is found to be −0.029 ± 0.018 dex/kpc, comparable to Sb or Sbc galaxies favouring an inside-out formation scenario for the galaxies. As interaction stage increases, both age and metallicity gradients are seen to get flatter, eventually getting slightly positive. The more active interaction stages are HII driven, while isolated and post merging stages shows higher AGN activity. The current work offers an update on the abundances of IR dominated galaxies from the previous work done by Rupke et al. (2008). The formation scenario of our LIRGs in the local universe is shown to be in line with the scenario put forward by Hopkins et al. (2008). Future works with medium resolution spectra acquired during the course of this thesis should allow for detection of gas inflows and better constrain the different ionising mechanisms involved at different interaction stages.

Includes bibliographical references (p. 233-242). / This thesis presents the results of a detailed multi-wavelength study of the nearby blue compact dwarf galaxies NGC 2915 and NGC 1705. The primary data set (nearly 100 hours of on-source data) for each galaxy consists of new observations of the neutral hydrogen (Hi) line obtained with the Australia Telescope Compact Array, October 2006 - May 2007. The stellar disk of NGC 1705 is host to an intense star-bursting core which is rapidly depleting the galaxy's central Hi reservoir. This galaxy can be used to rigorously test theories of star formation. Detailed studies of the distribution and kinematics of the neutral inter- stellar medium (ISM) within each galaxy are carried out. A suite of star formation recipes and models are examined for each galaxy to quantify the relationship between the observed star formation activity and the distribution and kinematics of the ISM.

Bibliography: p. 185-292. / This is primarily an observational thesis concerned firstly, with improving our knowlegde of the 12 SC stars, defined as a group by Catchpole and Feast (1971) and secondly, with determining the chemical composition of UY Cen, the brightest member of the group. An atlas, based on 13 A mm-1 spectra is presented for the S star π Gru and the C star 19 Psc between 5430 and 6850 A. In order to overcome the problems caused by the high density of absorption lines, we use a method of line identification that relies on the statistical interpretation of intensity against wavelength-agreement and Intensity against Excitation diagrams. The objective being to determine which elements are present in the spectrum before assigning identifications to individual lines. The spectrum is found to be rich in 's' process elements and shows weak molecules of both carbides, oxides and hydrides. U to L (0.36 to 3.4μ) photometry is listed for all the stars and shows them to be a very homogeneous group. A set of wavelengths derived from the line list is used to find the radial velocities of all the stars. A solar model atmosphere, with pure absorption as the method of line formation, is used to derive astrophysical log gf values from solar equivalent widths and meteor abundances. The model atmosphere also shows, that for our set of CN equivalent widths, both the CN abundance and the derived isotope ratio depend on the adopted C/O ratio.

This thesis, entitled "On evolution of star forming galaxies: the metallicity of dwarfs and the effect of environment on local luminous infrared galaxies" is divided into two main areas. First a detailed study of the environment and star formation relationship of local Luminous Infrared Galaxies (LIRGs), with infrared (IR) luminosity (LIR) between 1011L and 1011L, where L is solar luminosity was carried out. Secondly, a chemical abundance analysis was done on three Local Volume dwarf irregular galaxies. In the first part, various redshift surveys were used to quantify the environment around LIRGs. It was found that LIR 1011L is a remarkable luminosity point among IR galaxies.

This thesis makes use of sensitive Hi observations to map the distribution and kinematics of the low column density neutral hydrogen in late-type nearby mostly isolated galaxies and in the nearby M81 group. Using the KAT-7, GBT and MeerKAT AR1 telescopes, we present in the second chapter a survey of the Hi in a sample of twenty nearby, mostly isolated galaxies down to low column density levels. This provided a new Hi view of some of these galaxies, which allowed to derive their kinematics out to unprecedented extents. Despite the short spacings of KAT-7 and MeerKAT AR1, and the large size of the single-dish GBT that make these telescopes ideal for detecting faint structures, the observations revealed no clear detection of low column density Hi clouds down to a typical sensitivity of ∼2.2 × 1018 cm−2 that could be associated to gas accretion in the observed galaxies. However, we do not discard the existence of such structures that, we note, could be in the form of discrete clouds smaller than the beam size of the telescopes. In the third chapter, we use the DRAO telescope to perform a sensitive survey of the Hi in a 5 ◦ × 5 ◦ area of the M81 group. Similarly to previous observations, we find that the three major and interacting galaxies of the group – M81, M82 and NGC 3077 – are connected through Hi bridges and intergalactic Hi clouds. One of the major findings of the survey is the more complete map of the western Hi arm connecting the three galaxies to the dwarf galaxy NGC 2976. These observations offer enough resolution to map the structure of the arm, and reveal a complex of small clouds filling the space between the arm and the Hi forming “main body” of the interacting galaxies. Using a tilted-ring model, we also construct a large-scale rotation curve of the system formed by the interacting galaxies. Consistently with the large-scale velocity field, we observe a flat trend for the rotation velocity of the system from 20 kpc out to 80 kpc, well beyond the outskirts of the M81 disk, although with asymmetries like a wiggle at the vicinity of M82. The fourth chapter focuses on a subset of the M81 survey containing the dwarf galaxy IC 2574 and the Hi complex HIJASS J1021+68. In this chapter we perform a thorough analysis of the distribution and kinematics of the Hi in the two systems, and thanks to the high sensitivity we are able to detect a substantially large amount of low column density Hi around IC 2574, in the form of an Hi envelope, and in two large concentrations around the galaxy. We find evidence that HIJASS J1021+68 – which is found to be connected to IC 2574 through a filament of discrete clouds – is not a dark galaxy as previously suggested, but is instead a complex of clouds either stripped from, or falling onto the primordial Hi envelope of IC 2574. The kinematical analysis of IC 2574 using a 3D tilted-ring model brings us to derive its rotation curve out to a larger extent than previous works and allows us to constrain its Dark Matter halo parameters, which we find consistent with the literature. Overall, the different results presented in this work prove that the Hi content of local galaxies is higher than what current observations reveal, and new sensitive telescopes such as MeerKAT and the upcoming SKA will unveil a new Hi view of galaxies.

Super star clusters (SSCs) represent the youngest and most massive form of known gravitationally bound star clusters in the Universe. They are born abundantly in environments that trigger strong and violent star formation (SF) such as in galaxy mergers and interacting systems. SSCs are thus used as fundamental tools to understand the context of massive SF and galaxy evolution in general. This thesis investigates properties of these young, massive and dense star clusters in a sample of 42 nearby starbursts and luminous infrared galaxies (LIRGs) ...

Accurate stellar magnitudes on a uniform system are necessary for astronomical research. The North Polar Sequence, which defines the International System, is not accessible to southern observers so it has been necessary to establish magnitude sequences in the Southern Hemisphere. A number of observing programmes to establish such sequences in the nine Harvard Standard E Regions have been carried out using modern methods. Two of these programmes employing the Fabry method and a photoelectric photometer, respectively, are described in some detail. A special series of observations was made to ensure that the nine E regions have the same zero point. Accurate photographic and photovisual magnitudes were obtained for about 270 of the brighter stars. In a concluding section, modern material is collected and combined to give accurate magnitudes, the majority in two colours, for 640 stars, and approximate data for many more. The data have been examined and found to be free from systematic errors between magnitudes 3.0 and 10.0. The reductions to the adopted colour systems are also satisfactory. The most urgent need at present is for more observations of the fainter stars.

Word processed copy. / Includes bibliographical references (leaves 219-227). / Fitting binary lens models to gravitational microlensing event is currently a time consuming and labour-intensive process. Example-based algorithms more commonly used in the field of data mining were applied to simulatded and observed light curves in order to facilitate the fitting of a simple-seven-parameter binary lens model with minimal human intervention.