packed in cotton in a light glass cylinder, one and three quarter inches diameter by six and a half inches high, and closed the mouth of the cylinder by a flat cork carefully paraffined so as to be air-tight. The apparatus was then placed upright on the balance and counterpoised carefully. The balance beam was then lowered and the apparatus removed from the pan, held inverted for a second or two, and replaced on the pan. On raising the beam and releasing the pans a distinct loss of weight made itself immediately apparent, some seven to nine scale divisions. (The apparatus weighed from one hundred and ten to one hundred and twenty grains.) After some three quarters of an hour the normal weight returned.
"This loss of weight could not be due to the fact that I was weighing a vessel filled with air partially rarefied by the heat produced, for both vessels were closed air-tight. Nor could it be due to currents of hot air rising (the usual explanation in text-books) for two reasons: first, because the loss of weight was immediate, and, owing to the cotton packing, the outside of the apparatus did not become perceptibly warm to the touch for three quarters of a minute, and never became more than barely warm; secondly, because the effect noticed was too great to have been produced by convection. . . . It would seem, then, that convection currents have been greatly overestimated, and that the loss of weight noticed in weighing hot bodies is in large part a true effect. Either mass is a function of temperature, or else (which is more probable) heat weakens gravitation just as it weakens magnetic attraction, only that the magnetic attraction does not come back when the steel cools."
Furthermore, the authors show that the combining weights of the elements vary with the temperature, and they record a series of very interesting experiments. When applied to celestial gravitation, to the motion of the earth and other planets, and to their apparent irregularities, the new theory leads to surprising conclusions. If it stand the test of a more widespread examination, it will entirely change our conception of astronomical physics, and make necessary a radically different cosmology. The speculations in this department, however, are put forward very tentatively, and more by way of suggestion than as settled convictions. It will be noticed here, as elsewhere throughout the book, that the observed phenomena of Nature are not called in question, but only our conceptions and interpretation of them.
In conclusion—and it is quite time that we should stop we—are disposed to believe that Mr. Singer's and Mr. Berens's book is destined to attract wide attention, perhaps to provoke warm controversy, and certainly to stimulate wholesome doubt and inquiry. It has a high value quite aside from whether one accepts the
