solution of differential equations involves arbitrary constants of integration, the laws do not specify absolute heat, energy, or entropy, but merely changes.

Thermodynamics could have been developed differently, e.g., by following Rankine’s suggestions rather than Clausius’s. Prigone (1989) has argued that non-equilibrium thermodynamics supplies a better basis than the thermodynamics of quasi-equilibrium states. In the late nineteenth century, however, thermodynamics seemed to set a new standard for scientific theories. It was based on two laws of unrestricted validity. It could apply to the cosmos as a whole, leading to concerns with heat death. It prescinded from properties of any particular substance. Planck's dissertation defended the traditional concept of entropy against Boltzmann's reduction to statistical mechanics. The aged Einstein, reflecting on his youthful struggles with quantum theory claimed: "By and by I despaired of the possibility of discovering the true laws by means of constructive efforts based on known facts. The longer and more despairingly I tried, the more I came to the conviction that only the discovery of a universal formal principle could led us to assured results. The example I saw before me was thermodynamics." (citation from Schilpp, 1949, p. 53)

1.32. Electrodynamics. The development of electrodynamics had a Baconian phase culminating in quantitative concepts and laws (See Heilbron, 1979); new experimental breakthroughs consequent upon the discovery of electrical currents; and two competing theoretical traditions, Continental action at a distance and British field theory. It also featured a creative theorist who manifested an unparalleled critical concern with the status of basic concepts and the experimental support for theoretical hypotheses. For this reason we start with the problematic situation as seen by Maxwell in 1865, when he resigned his chair at King's College, retired to the family estate of Glenlair, Scotland, and began work on the systematic account that was later published as his Treatise (Maxwell 1954, original 1871; henceforth cited by pages). He later asserted that the basic purpose of the Treatise was to educate himself by presenting a view of the stage he had reached. (See Everitt, p. 80)

The underlying problematic was that the two approaches to electrodynamics seemed to be mutually incompatible but empirically equivalent. I use 'approach' , rather than 'theory' because Maxwell insisted he did not yet have a theory. He thought of a theory as an explanation in terms of causes (or sources). In his brutally honest appraisal of his own work he did not even have the beginnings of a theory because he did not know what electricity is or how fast it moves (p. 574), nor whether it is a substance, a form of energy, or belongs to any known category of physical quantities (p. 35). He did not have the roughest idea of the velocity of electrical currents or even their direction (p. 570). His account involved stress in the medium and an account of light in terms of undulations in the medium. Yet, he did not know how this stress originates (p. 644) or what light is (p. 821).

Maxwell's electrodynamics emerged from a protracted attempt to systematize experimental results on the basis of geometrical reasoning (See Wise, 1979). When this proved an insufficient basis for a coherent treatment he adapted Thomson's method of analogies, of adapting a mathematical formalism developed for a different field (See Buchwald, 1977). Thomson developed electricity by adapting the mathematics of heat flow. In his first paper on electrodynamics (1855) Maxwell interpreted Faraday's lines of force as tubes of variable cross section carrying an incompressible fluid. This allowed