In an ordinary furnace the draft moves upward, obeying the impulse due to the lightness of its heated gases. This direction is reversed in down-draft furnaces which were originally devised by Lord Dundonald more than a century ago. In their modern types a fan blast forces the draft downward through the fuel, with the effect that the gases are so intensely heated as to be thoroughly burned. The grate-bars are of water-tube, connected to the boiler as part and parcel of its heating surface. In the Loomis gas-producer a like method is adopted: the fuel is charged through an open door in the top of the generator and the gas is exhausted from the bottom of the fire. Thus all tarry and volatile matter in bituminous coal or wood is converted into a fixed gas. Thirty years ago one would have supposed the wheels of ordinary carts and carriages to be safe from change, to be among the heirlooms secure of transmission to posterity. Not so. Observe the wheel of a bicycle and note that instead of stout spokes upholding the hub, there are thin steel wires from which the hub is suspended. Thus strength is gained while the wheel is lightened and material economized. Wheels of like model are now used in many other vehicles where lightness is particularly desired. This plan of using spokes in tension instead of in compression is credited to Leonardo da Vinci who flourished four centuries ago. [Illustration: Bicycle wheel suspended from axle by wires.] Judgment in Theorizing: Rules Have Limits. While the men who add to known truth, whether in the realm of matter or of mind, must build on acquired knowledge, they do so with common sense, by exercise of the supreme faculty of judgment. To begin with, they perceive that every force acts within limits, acts concurrently with other forces which modify its effects. Speaking of gravity Professor William James says:--“A pendulum may be deflected by a single blow and swing back. Will it swing back the more often, the more we multiply the blows? No. For if they rain upon the pendulum too fast it will not swing at all, but remain deflected in a sensibly stationary state. Increasing the cause numerically need not increase numerically the effect. Blow through a tube; you get a certain musical note; and increasing the blowing increases for a certain time the loudness of the note. Will this be true indefinitely? No; for when a certain force is reached, the note, instead of growing louder, suddenly disappears and is replaced by its higher octave. Turn on the gas slightly and light it; you get a tiny flame. Turn on more gas and the flame increases. Will this relation increase indefinitely? No, again; for at a certain moment up shoots the flame into a ragged streamer and begins to hiss.”