The lines of sight to QSOs are powerful probes of large-scale structure from redshifts corresponding to the most distant QSOs to the local universe. In this thesis, spectroscopy of QSO pairs and groups are used to study superclustering at high redshift and to estimate the tranverse dimensions of the Lyα absorbers. We present high resolution (∼30km s⁻¹) echelle spectra obtained with the CTIO 4-m telescope of the wide QSO pair Tol 1037—2704 (z(em) = 2.193) and Tol 1038—2712 (z(em) = 2.331), as well as two neighboring quasars. The quasars exhibit a large number of apparently correlated C IV absorption systems over a narrow redshift range 1.48 ≤ 2 ≤ 2.15 which is thought to be produced by an intervening supercluster. The velocity correlation function of C IV absorbers distributed among the four lines of sight show significant clustering signal on comoving scales out to ∼30 h⁻¹ Mpc at redshift z ∼ 2 (h ≡ H₀/100 km s⁻¹ Mpc⁻¹; q₀ = 0.5). The spatial correlation function shows a marginally significant peak on scales of < 18h⁻¹ Mpc. The clustering amplitude on these scales is larger than that predicted by current theories of the formation of large scale structure. We present spectroscopy of three close pairs of quasars with angular separations 10" to 2' in order to measure the sizes of the Lyα forest absorbers from scales of a few tens of kpc out to hundreds of kpc. Ground-based estimates of the pair Q1343+2640.A (z(em) = 2.029) and B (z(em) = 2.031) imply a characteristic radius of the Lyα absorbers of ∼100 h⁻¹ kpc at z ≃ 2. Ultraviolet FOS spectra of the pair Q0107—025A (z(em) = 0.956) and B (z(em) = 0.952) in the redshift range 0.5 < 2 < 0.9 show a number of Lyα absorption features common to both spectra as well as several features which are not in common, and imply characteristic radii of 400 h⁻¹ kpc to bigger than 1 h⁻¹ Mpc. Furthermore, the rms velocity difference between the common systems between the two lines of sight is only about 100 km s⁻¹ These measurements lead to a picture of absorbing clouds that are larger in extent than previously thought and surprisingly quiescent. Using a new statistical technique, we tested the relative likelihood of three geometric models, namely, spherical absorbers, with and without a distribution in size, as well as filamentary and disk-like absorbers. Spherical absorbers with uniform radius cannot represent the observations and are ruled out. Randomly-inclined disks and filaments match the data comparably, with disks being slightly favored over filaments. Our results are in remarkable agreement with hydrodynamical simulations in which the Lyα absorption is found to arise in diverse structures with coherence lengths as great as 1 Mpc. Finally, we present FOS observations of a second pair of quasars, LB 9605 (z(em) = 1.834) and LB 9612 (z(em) = 1.898), over the redshift range 1.1 < z < 1.7 from which we placed an 95% confidence upper limit on the radii of the Lyα absorbers of 280 h⁻¹ kpc. The estimates span the redshift range 0.5 < z < 2, corresponding to roughly a third of the age of the universe, and provide tantalizing evidence for evolutionary growth in the size of the Lyα absorbers

The star formation history of galaxies is the primary influence on galaxy evolution, and hence the evolution of almost all the visible matter in the universe. In this dissertation, I present studies of the star formation history of galaxies which have come from two distinct perspectives: the study of galaxies that have unusual star formation histories, and the search within the general galaxy population for galaxies with unusual star formation histories. A spectrophotometric atlas of 40 merging and strongly interacting galaxies is obtained and analyzed in order to examine their stellar populations and star formation histories. Within the sample, the subsample of 10 ultraluminous IRAS systems is compared with the optically selected subsample. The population of objects in the sample with anomalously strong Balmer absorption lines, a spectral signature indicative of post-starburst evolution, is examined and compared with distant "E+A galaxies" which have similar spectrophotometric properties. Spectrophotometry across the entire optical wavelength range is obtained and analyzed for a sample of 8 E+A galaxies, ranging in redshift from 0.09 ≤ z ≤ 0.54. The method of stellar population modeling, widely used with only minor variations in the astronomical community, is examined and its strengths and limitations are discussed.

Many groups are trying to find faint stellar companions using a variety of techniques. The most obvious is to take long exposure pictures. Here adaptive optics (AO) is useful in correcting the blurring effects of the atmosphere. Starlight is concentrated into a tighter peak and less energy is put into the surrounding halo. Another method is to look for periodic decreases in a star's irradiance due to a planet blocking some of the light as it crosses in front of the star or transits. The most productive technique so far has been to look for periodic doppler shifts in the light coming from a candidate star, i.e. radial velocity searches. Several large planets have been found this way. This dissertation explores aspects of these three methods. The ability to detect a companion amid noise depends on the properties of the noise. While theoretical expectations exist for the properties of the halo produced by an AO system, experimental studies are few. The temporal and spatial characteristics of the halo produced by the AO system at the Starfire Optical Range are explored from the viewpoint of searching for faint stellar companions. We set limiting companion magnitudes for stars imaged while searching for brown dwarfs. The primary limitation to ground based transit searches is scintillation, irradiance fluctuations due to atmospheric turbulence. We explore reduction of this noise via differencing signals from binary stars. Theory is extended to include temporal and angular separation effects simultaneously. Evaluation of the derived equations and simulations demonstrate the benefits and limitations. Finally the design of an echelle spectrograph for use on the 6.5 m MMT or the Magellan telescope that is well suited to radial velocity searches for planets is presented. It has throughput between 10% and 18%, resolution of 200,000 per pixel and can sample the entire 0.31 to 1.1 μm range at once. This is accomplished by avoiding metallic reflections whenever possible and by using a mosaic of CCDs fit to the curved focal surface of a Schmidt camera. Efficient simultaneous observation of many spectral lines makes this a powerful instrument for radial velocity companion searches.

Beam profiles from the Infrared Astronomical Satellite (IRAS) at 60 and 100μm and for both hot and cold sources are analyzed using Fourier techniques to check the efficiency of the short wavelength blockage in the 60 and 100μm filters. Comparison of these power spectra with the expected passbands and an estimate of the noise supports the hypothesis that the 100μm filters have a short wavelength leak of 14.3 ± 3.6%, but that there is no detectable leak in the 60μm filters. Following decrement of the 100μm flux by 14.3%, stellar atmosphere models are anchored at 12μm to the IRAS flux measurements for 33 chosen stars and predicted flux measurements at 25, 60 and 100μm are computed. Comparison of these predictions with those measured by IRAS shows that the IRAS measurements are overestimated at 25 and 60μm and underestimated at 100μm with regard to atmospheric models.

Impact crater scaling laws are used to predict the diameter and amount of melt produced in impacts on quartz sand of projectiles with mass from 100 gm to 100 kg and velocity from 20-40 km/sec. A one dimensional cooling model incorporating conduction, change of phase, and radiation is used to predict the cooling history of the crater. Several possible initial distributions (exposed to surface, shallow or moderate burial by cooler material) of the impact melt are considered. Infrared spectra are calculated for the modeled surface temperature distribution at several times during the cooling. The impact IR signature is prominent in the wavelength range 1.5-6 μ against the lunar nightside background. The optimum wavelength for detecting the smallest accessible impact is between 3 and 4 μ. It is found that the maximum signal strength is dependent on the initial distribution of melt as well as impact energy. The duration of the signal above a minimum detectability threshold is proportional to impact energy with only modest dependence on the initial melt distribution. Basic design requirements and capabilities for sensors to detect the impact signature from lunar orbit and earth orbit are considered. The lunar orbiting sensor can detect impacts as small as ∼50 gm. With a field of view covering ∼640000 km² a rate of approximately 2 events per week might be expected. An earth orbiting sensor could detect impacts of ∼100 gm at the sub earth point. Larger impacts could be detected closer to the lunar limb. Monitoring a large fraction of the nighttime hemisphere visible from earth orbit the observable event rate is similar to that from the lunar orbiter. Ground based observation at wavelengths between 2 and 2.4 μ could detect ∼2 kg impacts with an event rate estimated at 1 per 400 hours observing time.

The Infrared Astronomical Satellite (IRAS) Serendipitous Survey Catalog (SSC) was constructed from pointed observations made by the satellite when it was in Additional Observations (AO), i.e., non-survey, mode. Analysis of the SSC shows that it contains a higher percentage of sources which have been detected only at 60mum than are found in the IRAS Point Source Catalog (PSC). This could reflect the existence of a large population of faint extragalactic (galaxian) objects due to the increased sensitivity of the SSC relative to the PSC, especially in the 60 mum band, or simply be a result of spurious sources in the SSC. Inspection of cirrus contamination over each AO showed that it had little or no effect on the high 60 mum-only source count. An automated optical identification program presented here indicated that 60 mum-only sources were as likely to have optical counterparts as all other sources, and more likely than randomly-placed artificial sources. The SSC sources studied had, on average, one more optical source found nearby than did artificial data: probably the optical counterpart to the IRAS source. These results support the validity of the 60 mum-only sources and the SSC in general.

Using Magellan data, I have investigated two crater characteristics that have been cited as diagnostic of oblique impacts: an uprange offset of the central peak in complex craters, and an increasing central peak diameter relative to crater diameter with decreasing impact angle. I find that the offset distribution is random and very similar to that for high-angle impacts, and that there is no correlation between central peak diameter and impact angle. Accordingly, these two crater characteristics cannot be used to infer the impact angle or direction.

The heat source for partial melting of the mare basalt source region is believed to be the urKREEP layer at the base of the crust. An initially spherically symmetric Moon model, hot or cold, allowed to evolve by axisymmetric convection inside a spherical shell for 300--600 Myr, is impacted by a large thermal anomaly, representing South Pole-Aitken basin, or a small thermal anomaly, representing Imbrium basin, and is then observed for 100 Myr of evolution. The energy of the large impact (South Pole-Aitken) is enough to remove the urKREEP layer from beneath the ejecta blanket by mantle circulation and is interpreted as the removal of the heat source for partial melting of a shallow mare basalt source region and explains the absence of fill in the basin while a small impact (Imbrium) cannot produce enough mantle circulation in the time allotted to remove the layer of heat producing elements from the basin surroundings.

The Solar Tower Atmospheric Cerenkov Effect Experiment (STACEE) is a ground-based atmospheric Cerenkov telescope constructed to conduct gamma-ray astronomy. The STACEE telescope employs the mirrors of the National Solar Thermal Test Facility (NSTTF) to observe gamma-rays at a lower energy threshold than the current generation of imaging atmospheric Cerenkov telescopes. A prototype of the telescope was constructed in 1998 which used 32 heliostat mirrors of the NSTTF with a total collection area of ~1200 m 2. This prototype called STACEE-32, was used to search for high energy gamma-ray emissions from the active galactic nucleus (AGN) ON + 231 (also known as W Coma and 1219 + 285). The STACEE-32 response as a function of energy was calculated for this source and we report a spectral energy threshold of E th = 130 +/- 50GeV. The analysis of data collected in the spring of 1999 resulted in a non-detection of the AGN at this energy. We derived a 95% confidence level upper limit on the integrated gamma-ray flux for this source of phiint(E > 130 +/- 50GeV) ≤ 2.4 x 10-10 cm-2sec -1. This upper limit further constrains the high energy flux extrapolations of this AGN.

In this thesis we study a population of gravitationally lensed, extremely distant, dusty, star-forming galaxies (DSFGs) discovered in large area millimetre surveys. This work is a collection of theoretical and observational studies which shed light on different aspects of this population, and pave the way for future use of these sources to study galaxy formation at high-redshift, dark matter, and dark energy. Two chapters of this work study the statistical properties of this population, demonstrating that i) the existence of extremely bright, high-redshift galaxies in millimetre and sub-millimetre surveys can be understood in the context of gravitational lensing of high-redshift sources by intervening structures along the line-of-sight, and ii) statistical lensing models allow us to understand the selection effects and statistical biases of this sample, which are crucial for a correct interpretation of the underlying source population. This thesis also studies how the lensing of high redshift sources modifies the statistics of the background radiation, resulting in a biased measurement of dark energy parameters using Sunyaev-Zel'dovich cluster detection experiments, such as the South Pole Telescope. The simulations presented in this work allow us to correct for this effect. We then present high resolution Atacama Large Millimeter Array (ALMA) observations of four of the sources in this sample, unambiguously confirming that these sources are gravitationally lensed by intervening galaxies. We develop a lens modeling technique appropriate for interferometric data and apply it to the ALMA observations to derive the lensing magnification, allowing the calculation of the intrinsic (unlensed) properties of the DSFGs. Finally, we study the prospects for detecting low-mass dark matter subhalos embedded in the halos of the lensing galaxies through their gravitational effects on the images of the background DSFGs. We show that by taking advantage of the spectral resolution of ALMA, it may be possible to detect dark matter halos with masses as low as 10^8 solar mass. / Cette thèse a pour objet l'étude d'une population de distantes et poussiéreuses galaxies à sursauts d'étoiles (DSFGs) ayant subit l'effet de lentille gravitationnelle et ayant été découvertes lors de grands sondages des ondes millimétriques. Le présent travail est constitué d'une collection d'études à la fois théoriques et observationnelles qui ont fait la lumière sur différents aspects de cette population de galaxies, et qui ont pavé la voie pour l'utilisation future de ces sources pour l'étude de la formation des galaxies à décalage vers le rouge élevé, de la matière noire et de l'énergie sombre. Deux chapitres du présent travail présentent l'étude des propriétés statistiques de cette population. On y démontre, premièrement, que l'existence d'une telle population de galaxies, extrêmement brillantes et à décalage vers le rouge élevé, au sein de sondages des longueurs d'ondes millimétriques et sub-millimétriques peut être comprise dans le contexte du phénomène de lentille gravitationnelle avec pour source ces distantes galaxies et pour lentille des structures massives alignées dans la même ligne de visée. Deuxièmement, on y démontre que la modélisation statistique des lentilles permet la compréhension des effets de sélection ainsi que les biais statistiques qui entrent en jeux dans l'échantillon de galaxies composant cette population; effets qui sont cruciaux pour une interprétation juste des propriétés réelles de la population. Cette thèse étudie également la manière par laquelle l'effet de lentille gravitationnelle modifie les propriétés statistiques du rayonnement de fond cosmologique diffus, ce qui a pour conséquence de biaiser les mesures des paramètres de l'énergie sombre via les expériences de détection d'amas par l'effet Sunyaev-Zel'dovich, tel le Télescope du Pôle Sud (SPT). Les simulations présentées dans le présent travail permettent la correction de ce biais. Nous présentons ensuite les observations de quatre sources tirées de l'échantillon de cette population en haute résolution, prises à l'aide du réseau ALMA (Atacama Large Millimeter/submillimeter Array). Ces observation confirment sans ambiguité que ces sources ont bel et bien subi l'effet de lentille gravitationnelle. Nous développons une technique de modélisation des lentille appropriée pour et adapté aux données interférométriques, puis nous appliquons cette dernière aux observations d'ALMA afin de dériver les facteurs d'agrandissement spécifiques à ces quatre sources, nous permettant de calculer les paramètres intrinsèques des-dites sources. Finalement, nous explorons la possibilité de détecter des sous-halo à faible masse de matière noire qui se trouvent enclavés dans le halo des galaxies lentilles, par le biais de leur influence gravitationnelle sur l'image des DSFGs d'arrière-plan. Nous démontrons qu'en tirant avantage de la résolution spectrale d'ALMA, il pourrait être possible de détecter des halos de matière noire d'une masse aussi faible que 10^8 Masse solaire.