Vision begins in the retina, where photoreceptors have the task of discriminating incomingphotons from intrinsic cellular noise. Vertebrates have evolved a duplex retina, in which rod and cone photoreceptors respond over a broad range of light intensities from single photons to bright light. Rods evolved from less sensitive cone-like photoreceptors and are responsible for mediating the lowest 8 orders of magnitude of light intensity. At the lowest threshold of vision, rods detect single photons and transmit signals to downstream neurons in a specialized circuit. In the rod visual pathway, tens of rods converge onto a single rod bipolar cell, where noise from the rods is filtered, and rod signals are processed for further downstream transmission. It is unclear which evolutionary changes imparted such sensitivity to rods, and how these advantages allow for rods to discriminate single photons from intrinsic noise. In downstream bipolar cells, the mechanism for setting the filtering threshold is unknown. Further, it is unclear what implications of the thresholding of rod bipolar cells have on vision in scotopic light levels. In this thesis, I investigated the mechanisms that work together to define the threshold of vision. To gain understanding of the nature of vision at the absolute threshold, I worked with three main goals in mind: • Goal 1: Determine the set of changes that make feasible the incredible sensitivity of rods. • Goal 2: Assess contributions of transduction proteins to the detection of single photons versus intrinsic noise. • Goal 3: Measure the ability of rod bipolar cells to shape and parse rod signals over several orders of magnitude of ambient light levels. To achieve these goals, I collected single-cell electrophysiological from mouse rods and rod bipolar cells. I analyzed the data with statistical models to determine parameters that described the mechanisms of visual threshold and adaptation. I also used statistical modeling to help validate results of mathematical modeling from our collaborators in the Reingruber Lab (Paris, France), which show that the important differences between rods and cones are reflected in the cone as a reduced outer segment volume, a decreased amplification through the transduction cascade, a faster turnover rate of the cyclic guanosine monophosphate second messenger, and an accelerated rate of decay of the phosphodiesterase (PDE) enzyme (goal 1). Through analysis of intrinsic noise in rods, I showed that PDE is the critical transduction element both setting the noise level and limiting the detection threshold and variability for single-photon absorptions (goal 2). Further, measurements in rod bipolar cells revealed that rod bipolar cells have a calcium-dependent feedback mechanism that allows rod bipolar cells to respond robustly to single-photon absorptions in the minority of the rods, while continuing to signal in brighter backgrounds that also cause rod adaptation (goal 3). Taken together, my work has helped delineate the fundamental mechanisms of signal detection and processing at the threshold of vision and over the dynamic range of rod vision.