The dissertation is an investigation of the Swedish engineering company AGA’s inventive activity during the years 1904-1959. Inventive activity denotes the company’s efforts in rying to develop innovations. Operations such as patents, experiments, business methods, business co-operations, technical development and other related activities have been studied. Through its lighting system for beacons, which emanated from the international gas industry, AGA had a strong economic base, and could thus go through periods of strong diversification. This led to the growth of more branches on the company’s product tree, and the aim of this dissertation has been to map out and understand how this AGA tree developed. One point of departure for this study has been to investigate how the inventive activity at AGA reacted to changes in the demand side of the economy. According to the American economist Jacob Schmookler the demand determines the development of innovative activity. The American economic historian Nathan Rosenberg has criticised Schmookler, however, arguing that it is the resources of knowledge which dictate the innovative course, since technology transfers are costly to put into economic practice. A third perspective, partly bridging these differences of opinions, is the discussion on the influence of technology procurement, which, among others, the Swedish innovation researcher Charles Edquist has presented. These perspectives frame my study, which maps AGA through two major changes in technology during the first half of the twentieth century. The AGA product tree consists of path dependent shifts in technology; possibilities to develop new technology opened up in the interaction between the company and the market. Among other things, the main innovation, the AGA flasher, originally developed for the lighting in beacons, proved to be functional for railway signalling devices and respirators. Through general market changes outside the company, similar opportunities arose for AGA to develop already existing technique for new markets. During the period of research the inventive activity was characterized by a constant experimenting, where the company’s success to a large extent rested on the engineers’ ingenuity. AGA, being product diversified, had little room to act independently on the market. Thus, to a high degree the company had to adjust its inventive activity to market demand. By cooperating with initiated customers, above all public ones, AGA had the opportunity to continuously develop products in demand, despite limited resources.