suffices to join the body, by means of a rigid connection, with a writing-lever, which touches a revolving cylinder, covered with smoked paper. The tracer on the writing-lever, moved with variable velocities, and in a direction parallel to the axis of the cylinder, will draw sinuous curves, whose parts will indicate by their inclination the velocities of the motions which produced them.
But the motions in walking are too extended to be traced on the revolving cylinder in their real magnitudes; in order to reduce them, while at the same time I preserved their characteristics unaltered, I had recourse to a train of wheel-work. In this apparatus, each wheel working into another, whose teeth are ten times more numerous than those of the former, it follows that the motion communicated to the first axis will be reproduced by the second with a reduction of 110; the third axis will reproduce the motion reduced to 1100; and the fourth axis will reduce it to 11000 etc.
If we attach to the foot of a walker a thread, which is wrapped around the wheel on the first axle of the wheel-work, and if to the third axle we connect the writing-lever, we can obtain traces on the revolving cylinder which will have only 1100th of the extent of the paths gone over by the foot of the walker.
Fig. 1 shows five traces obtained from the foot when walking with various velocities. A has been produced by the slowest walk; "B" is the ordinary gait; while C is the most rapid: the remaining traces have been obtained from gaits less rapid than that of C.
Fig. 1.
(The figure represents the smoked paper, unrolled from the revolving cylinder after the experiment. The paper revolved with the cylinder in the direction from O to Y. Therefore, the axis of the cylinder was parallel to OX, and the tracer on the writing-lever moved parallel to OX. It follows that, if the foot had remained stationary while the cylinder revolved, the tracer would have described a straight line parallel to OY. CH is the trace of the vibrating tuning-fork; each bend of its sinuous line is equal to 110th of a second of time. This chronographic trace gives us the means of estimating accurately the
