The actinic balance

THE writer bas been, during some time, making expel'iments on tbe device and construction of an instrument more delicate and more prompt than the tbermopile; an advance whicb bis recent researches into the distribution of radiant energy in the spectrum have proved to be indispensable. These researches have involved expenses for special apparatus which have been in part met by a grant from the American Academy of Arts and Sciences as trustees of Count Rumford, and the writer bas with gratitude to acknowledge his past in· debtedness to this aid. A communication of some of the principal results obtained was made to that society in the early part of December, and will appear with illustrations of the apparatus in a forthcoming volume of their Proceedings, to which the reader who desires fuller details is referred. The following independent description of the newly devised apparatus is rendered necessary here, as an introduction to a future account of researches in the true distribution of radiant energy in the solar spectrum. We see within the past few years a greatly increased atten· tion to this subject and an attempt by many skilled observers to measure the distribution of heat for each individual ray with the minimum of error which the vicious method of tbe prism admits. Even the use of the prism, however, demands most delicate means of measurement. Tyndall, employing every instrumental aid science commanded in his experiments on the electric light, was obliged to operate on a spectrum only an inch and a half in length, and it is from this that the wellknown heat curves of our text-books are derived. When we form a much longer spectrum, we must either make the face of our thermopile larger, or expect to find the radiation so weakened that we cannot measure it. Now the use of the prismatic spectrum involves two prominent causes of error. One of these (well-known and partly guarded against) concerns the selective absorption of the mate· rial employed, the other, far more important and pernicious in its results, has been almost completely neglected. It concerns the fact that th~ prism acts a part analogous to that of a cylindrical lens, concentrating tbe rays in the lower part of the spectrum as compared with the upper and entirely falsifying in its specious results the true distribution of the beat. Even if there were no theoretical difficulty the measure of the refracted heat as distributed by the prism is far from easy on account of its feebleness, but there has seemed to be no choice,