IN THIS paper, one attempts to develop a method for investigating counterfactual histories of science, that is, possibles histories that did not occur. We develop a theory of science based on the notion of "advance" (ideas, data etc.), which are units passed from scientist to scientist and which would be conserved in passing from one possible history to another. Advances are connected to each other by nets of causal influence, and we distinguish strong and weak influences. Around sixty types of advances are grouped in ten classes. As our case study, we examine the beginning of the Old Quantum Theory (1850-1915), using computation to store and process historical information. Based only on secondary literature, we describe four great possible histories and six small paths to the appearance of this theory. As continuation of the research, we will have to examine in greater detail the history of the period, especially of experimental science, and improve the strategies for postulation of counterfactual histories.
