Page:Popular Science Monthly Volume 10.djvu/752

732 Here there is nothing speculative or doubtful. Here are certain things to be done, which can be done only by men of the most thorough training, equipped with the best appliances. In such a laboratory, chance would have but little place. The work would run to hard routine, to the solid establishment of accurate scientific data, to the systematic determination of substantial facts. Precise physical measurements would precede generalization, just as the labor of the quarryman goes before that of the builder. Startling and brilliant discoveries might possibly be made, but incidentally rather than as the result of special effort. The real value of the institution would be independent of anything sensational, and would rest upon considerations of the most severely practical kind.

Examples of the sort of work appropriate to an endowed laboratory may easily be found. For instance, one of the greatest of all scientific problems is that of the connection between the composition of a substance and its physical properties. Suppose this question were to be taken up systematically by a well-organized body of investigators. The first step in the research would manifestly be to determine, carefully and with the utmost rigor, the physical properties of the so-called chemical elements. At the outset each one of these substances would have to be isolated in quantity, and in a chemically pure condition—a labor which of itself would involve a great amount of research. Some of the elements have never yet been seen in a state of absolute purity, or have been obtained only in very minute portions, and accordingly new methods of treatment would need to be devised. Then would come the measurement of physical relations, thermal, electrical, optical, magnetic, mechanical, and so on. For each element, as far as possible, should be determined the melting-point, the boiling-point, the density, the coefficient of expansion, the specific and latent heat, thermal and electrical conductivity, the thermo-chemical constants, and many other data of much importance and value. Furthermore, these constants should be determined under widely-varied conditions, notably of pressure and temperature. For example, it would be necessary to ascertain the coefficient of expansion, and also the specific heat of a body at every degree, through a wide range of temperatures, and in not merely one, but in several series of observations. Thus, and thus only, could we attain to the exactness which science rigidly demands. Besides the actual measurements, this great labor would in many cases involve the comparative testing of various methods of research, and in some instances the invention of new experimental processes.

Years could be spent upon the metals alone, and the work done would add not only to our knowledge of their properties, but also much to science as regards variety and precision of methods. In connection with these researches would naturally arise an investigation of metallic alloys—a subject of which true science knows as yet