another, each attacking that corner of the unknown which most attracts his fancy. All are ambitious to accomplish great results, each one hopes to make some discovery of signal importance; and so the drier and less attractive details of investigation are oftentimes neglected. The field is cut up into many fields, between which the ground is uncultivated, and there no harvest is gathered. To systematize research, to bring about co-operation, to put the art of discovery itself more truly upon a scientific basis, is a problem for the future. In the final solution of this problem the practical inventor may help. The wealth created by invention should serve as the organizer. The law of mechanics that action and reaction are equal and opposite, applies to human affairs as well as to physical forces. Hence, since scientific discovery makes invention possible, it is clear that the inventor owes something to science in return. That some of the harvest should go back to its source as seed is not an unreasonable expectation. Indeed, it is justified by history; and if we trace back to their origin the endowments of our universities, we shall find that the successful inventors have done their fair share. What more is needed, and on what new lines?
In the science of astronomy this question is partly answered already. Every endowed observatory is an institution for research, and outside of that the observers have little else to do. They are employed primarily to gather and discuss data, the raw material of science, and all other duties are secondary. In the solution of large problems several observatories may co-operate, each taking a definite and prescribed portion of the field; and so the science grows symmetrically, with fewer gaps than exist in other departments of knowledge. Perfection of work, completeness in the absolute sense of the term, is of course unattainable, but to that ideal within the limits of its province astronomy approaches most nearly. By its example the other sciences may profit.
Now, for chemistry and physics institutions should be organized resembling in policy the astronomical observatories. I mean, of course, endowed laboratories for research, in which the greater problems could be effectively handled, and important data determined with the highest accuracy. The more precise and at the same time the most difficultly measurable physical constants are of direct value to industrial science, and their determination should not be left to the caprice or convenience of individuals. They represent routine work of the most tedious kind; their measurement involves the highest degree of skill and the most elaborate resources, and they are the foundation-stones of exact theory. They are needed by pure and applied science alike; and yet, under existing conditions, their determination is but scantily
