movement ot the pulley was very much reduced. A string, passing over the circumference of the pulley, was wrapped round r, so that, as the weight descended, the pulley moved round, and the string of the pulley caused r to rotate very rapidly. Now, the motion of the axis r was conducted within the covered box B, where there was attached to r a system of paddles, of which a sketch is given in figure; and therefore, as r moved, these paddles moved also. There were, altogether, eight sets of these paddles revolving between four stationary vanes. If, therefore, the box were full of liquid, the paddles and the vanes together would churn it about, for these stationary vanes would prevent the liquid being carried along by the paddles in the direction of rotation.
Now, in this experiment, the weight was made to descend through a certain fixed distance, which was accurately measured. As it descended, the paddles were set in motion, and the energy of the descending weight was thus made to churn, and hence to heat some water contained in the box B. When the weight had descended a certain distance, by undoing a small peg p, it could be wound up again without moving the paddles in B, and thus the heating effect of several falls of the weight could be accumulated until this became so great as to be capable of being accurately measured by a thermometer. It ought to be mentioned that great care was taken in these experiments, not only to reduce the friction of the axles of the pulley as much as possible, but also to
