The easiest and most effective way to influence the separation process in gas chromatography (GC) is achieved by controlling the temperature of the chromatographic column. In conventional GC, the temperature along the length of the column is constant at any given time, T(t). In my research, I investigated the effects of temperature gradients on GC separations as a function of time and position, T(t,x), along the column. This separation mode is called thermal gradient GC (TGGC). The research reported in this dissertation highlights the fundamental principles of axial temperature gradients and the separation potential of the TGGC technique. These goals were achieved through the development of mathematical models and instrumentation that allowed study of the effects of axial temperature gradients. The use of mathematical models and computer simulation facilitated evaluation of different gradient profiles and separation strategies prior to development of the instrumentation, providing theoretical proof of concept. Three instruments capable of generating axial temperature gradients, based on convective cooling and resistive heating, were developed and evaluated. Unique axial temperature gradients, such as nonlinear and moving sawtooth temperature gradients with custom profiles were generated and evaluated. The results showed that moving sawtooth temperature gradients allowed continuous analysis and were well-suited for comprehensive GCxGC separations. The use of custom temperature profiles allowed unique control over the separation power of the system, improving separations, as well as selectively increasing the peak capacity and signal-to-noise. A direct comparison of TGGC with conventional GC methods showed that TGGC produces equivalent separations to temperature programmed GC. This technology holds great promise for performing smart separations in which the column volume is most efficiently utilized and optimum separations can be quickly achieved. Moreover, precise control of the elution of compounds can be used to greatly reduce method development time in GC. This feature can be automated using feedback to develop efficient separations with minimum user intervention. This technology is of special interest in micro-GC systems, which allows relatively easy incorporation of resistive heating elements in the micro-column design.