the four sides of the pit, and a kilometre or so cut off the lower pointed end to allow space for its descent. The mass of this weight is 326 × 106 tons. Its heaviness, three quarters of the heaviness of an equal mass at the sun's surface, is 244 × 106 tons solar surface-heaviness. Now a horse-power is 270 metre-tons, terrestrial surface heaviness, per hour; or ten metre-tons, solar surface-heaviness, per hour. To do 78,000 horse-power, or 780,000 metre-tons, solar surface-heaviness, per hour, our weight must therefore descend at the rate of one metre in 313 hours, or about twenty-eight metres per year.
To advance another step, still through impracticable mechanism, toward the practical method by which the sun's heat is produced, let the thread of the screw be of uniformly decreasing steepness from the surface downward, so that the velocity of the weight, as it is allowed to descend by the turning of the screw, shall be in simple proportion to distance from the sun's center. This will involve a uniform condensation of the material of the weight; but a condensation so exceedingly small in the course even of tens of thousands of years, that, whatever be the supposed character, metal or stone, of the weight, the elastic reaction against the condensation will be utterly imperceptible in comparison with the gravitational forces with which we are concerned. The work done per metre of descent of the top end of the weight will be just four fifths of what it was when the thread of the screw was uniform. Thus, to do the 78,000 horse-power of work, the top end of the weight must descend at the rate of thirty-five metres per year: or seventy kilometres, which is one one hundredth per cent (110000) of the sun's radius, per two thousand years.
Now let the whole surface of our cool solid sun be divided into squares, for example, as nearly as may be, of one square metre area each, and let the whole mass of the sun be divided into long, inverted pyramids or pointed rods, each 700,000 kilometres long, with their points meeting at the center. Let each be mounted on a screw, as already described for the long tapering weight which we first considered; and let the paddle at the top end of each screw-shaft revolve in a fluid, not now confined to a vat, but covering the whole surface of the sun to a depth of a few metres or kilometres. Arrange the viscosity of the fluid and the size of each paddle so as to let the paddle turn just so fast as to allow the top end of each pointed rod to descend at the rate of thirty-five metres per year. The whole fluid will, by the work which the paddles do in it, be made incandescent, and it will give out heat and light to just about the same amount as is actually done by the sun. If the fluid be a few thousand kilometres deep over the paddles, it would be impossible, by any of the appliances of solar physics, to see the difference between our model mechanical sun and the true sun.
Now, to do away with the last vestige of impracticable mechanism, in which the heavinesses of all parts of each long rod are supported
