This paper examines the discovery of the carbon dioxide theory of climate change, its eclipse during the first five decades of the twentieth century, and its reemergence in the work of G.S. Callendar. It provides historical perspectives on the study of climate dynamics (�C/�t) from the perspective of science dynamics (�S/�t). In the nineteenth century, the work of John Tyndall, Svante Arrhenius, T.C. Chamberlin and others drew scientific attention to the role of carbon dioxide as a possible mechanism of climate change. Throughout the first half of the twentieth century, however, most scientists did not think that increased carbon dioxide levels would result in climate warming. Spectroscopic work by Knut �ngstr�m, Clemens Schaefer, and others led meteorologists to believe that water vapor was controlling the infra-red heat budget and that doubling or halving the amount of CO2 in the atmosphere would not appreciably affect the amount of radiation actually absorbed by it. Most meteorologists gave other mechanisms of climatic change more credence. Beginning in 1938 Guy Steward Callendar began a revival of the carbon dioxide theory of climate change and placed it on a more solid scientific basis. Callendar based his theory on new detailed measurements of the infrared spectrum, rising fossil fuel emissions, and the warming trend recorded in the Northern Hemisphere from about 1900 to 1950. Hans Seuss and Roger Revelle later named the rising levels of atmospheric carbon dioxide caused by industrial fuel combustion the "Callendar effect." The CO2 theory of climate change has changed dramatically on timescales of decades to centuries (eg. 1850 to 2000). Since climate ideas (in relation to changes in technology, and social organization of science) can change faster than the climate itself, they are worthy of serious historical study. Clearly, a student of climate dynamics must also be a student of science dynamics.