Within the realm of research toward the emulation of human intelligence, the problem of how to perform qualitative reasoning with computer programs has received considerable attention. The research field of qualitative physics focuses on the special problems of identifying the basic concepts of nature and the issues related to mechanizing inference about these concepts, so that we might build qualitative models of the physical world. These models are indispensable for providing machines with "common sense" inference capabilities to reason about and describe observations of their environment. This dissertation contributes a qualitative theory of the gas dynamics domain. The theory comprises a qualitative representation of the dynamic behavior of gases, constructed upon a collection of fundamental process elements. These process elements affect the description of a dynamic scenario by manipulating the qualitative values of physical parameters that change in response to certain physical phenomena. The phenomena covered are those occurring in one- and two-dimensional flows of an ideal gas. Deployed in conjunction with an automated inference mechanism, the models built from this theory constrain results to those situations validated by natural physical laws. These models provide qualitative reasoning and simulation capabilities to intelligent computer-aided design systems. The presentation constructs a foundation within the computational framework of Qualitative Process Theory, and shows how an algebra of Qualitative Ratios, introduced herein, provides a convenient representation for describing physical change. The presentation and internal hierarchy of the theory parallels the engineering perspective of this domain, and encourages composition of the fundamental elements into larger, more complex, reasoning components. Several examples demonstrate the utility of the theory during the course of the presentation.