The front strips O^{8} and rear strips O^{9} and the front sheets O^{10} and the rear sheets O^{11} are connected to the levers O^{6}, so as to form, in effect, single expansion strips and sheets of relatively great length. They are fastened, however, at their front upper ends to the cross-bars O^{4}, so that the expansion cannot extend beyond that point and takes place in a direction towards the opposite end, and, of course, the contraction takes place in the opposite direction. Assuming now that at a temperature of say 75 degrees Fahr. these heavy strips and sheets lie in the position shown in Figs. 4 and 9 (the heavy strips O^{8}, O^{9} being shown in Fig. 9, and the heavy wide sheets O^{10}, O^{11} in Fig. 4), on a decrease in temperature of say five degrees Fahr., the heavy strips O^{8}, O^{9} and sheets O^{11}, O^{12} will contract in the direction of the arrows, depressing the rear ends of the levers O¹, O^{6}, and thereby through the levers O¹ lifting the knife-bars E, and the balance-levers suspended thereon, with the result that the force normally exerted at the ends of each expansion coil is increased to the extent of the lifting power of the contraction of the metal strips and sheets. I have found by experiment as well as observation that the average daily change of temperature in residence and office buildings is about five degrees. Sometimes the changes will be much greater, and sometimes less. On even a low average of temperature change, my apparatus will be able to generate force in larger amounts than required, and the surplus will be stored in a power-storage device such as above described, or by means hereinafter referred to, which surplus will [Illustration: _Fig. 4._ _Fig. 5._ BANGERTER’S PERPETUAL TIME CLOCK] be drawn upon when it should happen that the average temperature is approximately uniform.