The work in this disseration has arisen from the Cambridge Optical Aperture Synthesis Telescope (COAST) project. COAST is a novel stellar interferometer under construction in Cambridge, designed to obtain stellar images with an angular resolution equivalent to that of a 100m telescope. The success of this design requires innovation in a number of areas including photo-detector design. The first intention was to specify the best detector and the optimal operating wavelength for COAST. This was done by considering the signal-to-noise ratio in fringe visibility modulus measurements with various detectors. The result of these calculations and other ancillary considerations was to choose the Silicon Avalanche Photodiode (Si APD) operated in a photon counting mode. Although Si APDs have been available for several years, no commercial photon counting unit was available at the outset of this research, so having specified the parameters for the COAST detector system, research and development of a photon counting APD detector was undertaken. The result is a small self-contained unit employing an RCA APD. This unit offers exceptionally high efficiency ( 40% in the near infrared), broad spectral response and low noise whilst being rugged, reliable and reasonably cheap. These COAST detectors have a much wider range of applications than just stellar interferometry as replacements for the commonly used photomultiplier tubes, e.g. in telescope autoguiding and photometry. Atmospheric seeing is of crucial importance to the performance of stellar interferometers and large telescopes. Specifically, stellar interferometry is sensitive to a very wide range of scales in the seeing - both temporal and spatial. The well-established theoretical models of seeing are currently not well constrained experimentally, especially in relation to the outer scale of turbulence. The size of this latter parameter is the subject of debate, but is vital in determining the maximum fringe motion in an interferometer. To resolve some of these problems, a new seeing monitor (WFTI) was proposed. WFTI is a novel interferometer designed to make measurements of both the temporal and spatial scales of the seeing. It uses photon counting APD detectors for the first time in an astronomical application and was operated for two nights on the William Herschel Telescope at the La Palma Observatory (LPO ). Results verified seeing models over scales of three orders of magnitude and have shown that the underlying seeing at the LPO is exceptionally good. A direct optical detection of the outer scale was made for the first time at the LPO and may be as short as 2m. These are very promising results for future large telescopes and interferometers at the site and indicate that care must be taken with telescope design to exploit this quality of seeing to the full. Unfortunately the outer scale detection was a little more tenuous than might be wished, due to an excess of low frequency power in the data from WFTI. Possible causes (both instrumental and atmospheric) for this excess were suggested and a new instrument (WFTI II) proposed. This improved interferometer will separate instrumental from atmospheric effects by taking multiple observations concurrently. COAST path stability measurements were made using WFTI for beam combining. The results show that it will acceptable for the beams returning from the individual array elements to the optics laboratory to propagate in free air, at least with short baselines. Finally, a new method for locating the COAST "white light" fringe was described.