DESCRIPTION OF APPARATUS. 41 



of a cord passing over a groove in the wheel to rotate it in one direction 

 as the body was lifted, a pawl p preventing any backward motion. 

 When the body returned to the normal position, the cord slipped over 

 the pulley without producing any movement of the wheel. Each 

 upward movement of the body accordingly resulted in a forward 

 motion of the wheel. The work adder-wheel was directly connected 

 to the shaft of a revolution-counter, and a record of the total move- 

 ment of the wheel was thus obtained. (See C, fig. 6.) The total dis- 

 tance the body was raised would theoretically be that found by multi- 

 plying the number of revolutions of the wheel by its circumference. 



At the very rapid rate of walking used in some of these experiments, 

 great difficulty was experienced in the earlier development of the appa- 

 ratus in securing an accurate record. The cord would frequently 

 slip and there was considerable reverse movement, for it was not easy 

 to find a pawl arrangement which would work perfectly. This reverse 

 movement or "back-lash" was not overcome until practically the end of 

 the series of observations; recently, however, a very thin laminated 

 spring has been used as a pawl, which has practically eliminated this lost 

 motion. The apparatus as described represents the completed form. 

 As a matter of fact, in many of our experiments we were unable to use 

 the records obtained from this wheel. The height to which the body 

 was raised was therefore determined by the method previously referred 

 to, i. e., measuring on the kymograph curve the general average dis- 

 tance to which the body was raised and multiplying this by the record 

 of the number of steps on the step-counter. 



In order to have these measurements of absolute value, the distance 

 traveled by the pointer on the kymograph or that traveled by the cord 

 passing over the work adder-wheel must represent exactly the distance 

 over which the body is elevated or lowered. This assumes that there 

 is no disturbance in the attachment of the cord to the body and that 

 the tension of the spring is such as to prevent any diag or inequality 

 in the movement of the pointer over the drum. As may be seen from 

 the kymograph curve reproduced in figure 7, the constancy in the 

 movement of the pointer was very satisfactory. On the other hand, 

 the method of attaching the cord to the subject was certainly open to 

 serious criticism. Wire was used for the most part, violin string being 

 substituted for such portions as passed over the flexible parts, thus 

 minimizing the tension and the danger of alterations in the length. 

 There was, however, opportunity for considerable flexibility in the play 

 of this cord at the point where it was attached to the belt holding up the 

 trousers of the subject; accordingly, the records obtained with the kymo- 

 graph and with the work adder- wheel are without doubt invariably some- 

 what too small and the body was in all instances raised to a higher point 

 than that indicated in table 2. The expediency of attaching a wire 

 direct to the body by means of surgeon's plaster was not resorted to, 

 although the omission was undoubtedly an error in our observations. 



