Abstract

My dissertation philosophically analyzes experiments in evolutionary biology, an area of science where experimental approaches have tended to supplement, rather than supercede more traditional approaches, such as field observations. I conduct the analysis on the basis of three case studies of famous episodes in the history of selection experiments: H. B. D. Kettlewell's investigations of industrial melanism in the Peppered Moth, Biston betularia; two of Th. Dobzhansky's studies of adaptive radiation in the fruit fly, Drosophila pseudoobscura; and M. Wade's studies of group selection in the flour beetle, Tribolium castaneum. The case studies analyze the arguments and evidence these investigators used to identify the respective roles of experiments and other forms of inquiry in their investigations. I discuss three philosophical issues. ;First, the analysis considers whether these selection experiments fit models of experimentation developed in the context of micro-and high energy physics by Allan Franklin and Peter Galison . My analysis documents that the methods used in the case studies can be accommodated on both Franklin and Galison's views. I conclude the case studies do not support claims regarding the relative autonomy of biology. ;Second, the analysis documents a number of important roles for life history data acquired by strictly observational means in the process of experimentation, from identification of research problems and development of experimental designs to interpretation of results. Divorced from this context experiments in biology make no sense. Thus, in principle, experimental approaches cannot replace more traditional methods. ;Third, the analysis examines a superficial tension between the use of experiments, which I characterize by the presence of artificial intervention, and the stated goal of most investigations in evolutionary biology, that of understanding how systems behave in the absence of intervention. Experiments involve trade-offs between the control one has over the circumstances of the study and how informative the study is with regard to questions of interest to biologists regarding specific, actual systems in nature. Experimental simulations of natural phenomena in other historical sciences involve similar trade-offs, but there are reasons for believing this tension is more prominent in biology