The need for solar electricity is clear. It is safe, ecologically sound, efficient, continuously available, and is has no moving parts. The basic problem with the use of solar photovoltaic devices is economics, but until recently very little progress has been made toward the development of low-cost photovoltaic devices. The larger part of research funding has been devoted to study of single-crystal silicon solar cells, despite the evidence, including that of the leading manufacturers of crystalline silicon, that the technique holds little promise. The reason for this pattern is understandable and historical. Crystalline silicon is the active element in the Very successful semiconductor industry, and virtually all of the solid state devices contain silicon transistors and diodes. Crystalline silicon, however, is particularly unsuitable to terrestrial solar cells. Crystalline silicon solar cells work well and are successfully used in the space program, where cost is not an issue. While single-crystal silicon has been proven in extraterrestrial use with efficiencies as high as 18 percent, and other more expensive and scarce materials such as gallium arsenide can have even higher efficiencies, costs must be reduced by a factor of more than 100 to make them practical for commercial use. Beside the fact that the starting crystalline silicon is expensive, 95 percent of it is wasted and does not appear in the final device. Recently, there have been some imaginative attempts to make polycrystalline and ribbon silicon, which are lower in cost than high-quality single crystals. But to date the efficiencies of these apparently lower-cost arrays have been unacceptably small. Moreover, these materials are cheaper only because of the introduction of disordering in crystalline semiconductors, and disorder degrades the efficiency of crystalline solar cells. This dilemma can be avoided by preparing completely disordered or amorphous materials. Amorphous materials have disordered atomic structure as compared to crystalline materials. That is, they have only short-range order rather than the long-range periodicity of crystals. The advantages of amorphous solar cells are impressive. Whereas crystals can be grown as wafers about four inches in diameter, amorphous materials can be grown over large areas in a single process. Whereas crystalline silicon must be made 200 microns thick to absorb a sufficient, amount of sunlight for efficient energy conversion, only 1 micron of the proper amorphous materials is necessary. Crystalline silicon solar cells cost in excess of $100 per square foot, but amorphous films can be created at a cost of about 50 per square foot. Although many scientists were aware of the very low cost of amorphous solar cells, they felt that they could never be manufactured with the efficiencies necessary to contribute significantly to the demand for electric power. This was based on a misconception about the feature which determines efficiency. For example, it is not the conductivity of the material in the dark which is relevant, but only the photoconductivity, that is the conductivity in the presence of sunlight. Already, solar cells with efficiencies well above 6 percent have been developed using amorphous materials, and further research will doubtless find even less costly amorphous materials with higher efficiencies. The author is primarily concerned with ______.