This is the story of a radical change in humans' concept of light. In 1896 most physicists were convinced that light consisted of wave disturbances in a medium, the electromagnetic aether. The energy transported in radiation was thought to propagate spherically outward from its source and to spread out over successively larger volumes in space. Thirty years later, a remarkably different concept of light prevailed. Physicists then took seriously the evidence collected over the preceding two decades showing that the energy of radiation does not spread in space. Under certain conditions, light behaves like a stream of particles.

Our subject is more than the story of a shift from one theoretical explanation of radiation to another. This reconsideration of the nature of light was a significant event in our scientific understanding of the world. To resolve the paradox that faced them, physicists rejected the venerable Platonic dictum that the microscopic realm recapitulates the macroscopic; that laws generalized from the behavior of objects in the perceivable world may be applied to the imperceivable one. By 1927 physicists had assigned to all forms of radiation a curious amalgam of wave and particle behavior. Waves spread energy over larger and larger volumes of space; particles do not. Reconciliation of these conflicting properties was possible only through appeal to an ontology that transcended mechanical incompatibility.

Nature was declared to be only imperfectly rationalizable in terms of human experience with macroscopic interactions. The programmatic goal formulated in the seventeenth century to reduce all physical phenomena to consistent mechanical representations was here recognized to be unattainable.

German-speaking physicists arrived at an explanation for the problems facing radiation theory different from that of their British colleagues. Moreover, their attempt to discover a consistent theory formed a complete contrast to the British attempt. The problems themselves were perceived differently. It would be an exaggeration to say that the problems ever reached the same status as paradoxes in Germany as they did in Britain. Physics as practiced in Germany was not as dependent on conceptual pictures, and physicists there were much more willing to adopt formal principles to solve the difficulties without demanding a consistent physical interpretation.

By 1905 the very word mechanics meant to many Germans something essentially different from its meaning in Britain. Influential voices had proclaimed that matter itself is only a construct of the mind, fashioned out of the more ontologically significant electromagnetic forces that give rise to human perceptions of mass and extension. The more influential Germans were not closely tied to the logical requirements of mechanistic thought. When the electromagnetic impulse hypothesis of x-rays was accepted in Germany, it was interpreted by many as a further example of the versatility of an electromechanical ontology. Electromagnetic impulses were considered by some, most notably those who first tried to make sense of x-ray behavior, as a form of wave with extended properties in space. Consequently, as we shall see in this chapter, the sharp distinction between x-rays and periodic light was never made as strongly in Germany as it was in Britain.

In the spring of 1912 two research assistants in Munich directed an x-ray beam through a crystal and found that the beam was reformed into a well-defined interference pattern. The property most characteristic of periodic waves – their ability to interfere – is shared by the x-rays. Max Laue, the man chiefly responsible for the discovery, thought he had found proof that characteristic secondary x-rays from the crystal are periodic waves. But H. A. Lorentz quickly pointed out that impulses should interfere too. He showed, in a tour de force argument, that the accepted square pulse is an impossible representation of x-rays. William Henry Bragg and his son concluded that the interference maxima could, indeed, be due to irregular x-ray pulses. But, as such, x-ray impulses were not different from ordinary white light. They supported this claim with a new technique of crystal analysis fully analogous to ordinary optical spectroscopy.

Crystal diffraction provided a new tool for the analysis of x-rays. Pushed furthest by Henry Moseley and Charles Darwin, the technique soon showed that some x-rays comprise periodic wave trains of great length. The extremely sharp angular resolution of observed x-ray interference maxima indicated beyond doubt that x-rays are no different, except in frequency, from ordinary light. Rutherford soon extended the technique to the y-rays. Not only could one isolate characteristic γ-rays, he believed, one could show, with some effort, that they interfere too.

The successful integration of the new spectroscopy with the Bohr atom came, as had x-ray diffraction, from Sommerfeld's Munich.

The same x-ray spectroscopic techniques that led to the identification of the atomic origin of x-ray spectra also offered means to test the localization of energy in the radiation. In the second decade of this century, x-ray spectrometers used as monochromatizers provided more purely defined x-rays than had been available from natural characteristic radiation. This, coupled with techniques to determine the kinetic energy of secondary electrons, provided opportunities for precise testing of the quantum transformation relation, E = hv. These experiments form the subject of this chapter.

We are primarily concerned here with the absorption of radiation, not with its emission. To be sure, the quantum regulation of the emission of radiant energy has no classical explanation; yet there is no electromechanical inconsistency implied in the creation of a spherical wave containing a definite amount of energy. It is the inverse case that causes real difficulty. How can that quantum of spherically radiating energy concentrate its full power on a single electron? For this reason, verification of the quantum relationship for emitted radiation lies outside our direct concerns. The Franck-Hertz experiments beginning in 1912, for example, demonstrated the quantum nature of energy transfer, but they did little to encourage acceptance or even consideration of the lightquantum. On the other hand, the experiments detailed here that verified the quantum nature of the absorption of light and x-rays gave substance to the lightquantum hypothesis because they verified the particlelike transfer of radiant energy to matter.

Between 1898 and 1912, a majority of physicists thought that x-rays were impulses propagating through the electromagnetic field. Only the extremely large number of pulses gave the x-ray beam its seeming continuity. Although this hypothesis was compatible with the wave theory of light, it was a special case of that theory. Impulses are not ordinary waves. Although they propagate spherically outward from their source, pulses are not periodic oscillations. The energy in an impulse is temporally but not spatially localized. It is contained within an ever-expanding shell, but the shell's radial thickness remains constant and small. Along the circumference, energy is distributed uniformly. But radially, from front to back so to speak, electromagnetic energy rises quickly from zero and drops back just as rapidly. When it passes a point in space, an impulse exerts only a single push or a single push – pull. A pulse collides, rather than resonates, with an atom.

In their temporal discontinuity, impulses differ decisively from their periodic-wave cousins. Light has an intrinsic oscillatory character that allows it to interfere; the superposition of two beams of coherent monochromatic light produces alternate regions of constructive and destructive interaction, the well-known interference fringes. A pulse has no oscillatory structure. Its interference properties are qualitatively different from those of light. A truly monochromatic light wave must extend infinitely in time; if it does not, an intrinsic ambiguity arises in the definition of its frequency. A pulse is restricted in temporal extent, and the very concept of frequency cannot readily be applied.

When x-rays or γ-rays strike atoms, electrons and secondary x-rays are emitted. Experiments to sort out the differing properties of the two secondary components produced new and perplexing observations in the first decade of this century. The most perplexing problems concerned the effect of the rays in producing secondary electrons. First, x-rays and γ-rays ionize only a small fraction of the total number of gas molecules through which they pass. Spherically expanding pulses should affect all molecules equally; manifestly, they do not. Second, the velocity that x-rays impart to electrons is many orders of magnitude higher than one would expect to come from a spreading wave. The energy in the new radiations seemed to be bound up in spatially localized packages, available to an electron in toto.

British physicists responded to these paradoxes with attempts to revise, rather than replace, classical electromechanics. J. J. Thomson suggested that old ideas about the microscopic structure of the aether might have to be reformulated. William Henry Bragg concluded that x-rays and γ-rays are not impulses at all but rather neutral material particles. Bragg and Thomson tried at first to find explanations using models based on human experience with machines. Each sought a resolution within the context of classical mechanics; each ultimately failed.