EXPERIMENTS WITH HORIZONTAL WALKING. 155 



The average number of steps taken by E. D. B. at different speeds 

 is shown in table 42. (See p. 156.) The increase in the number of 

 steps with greater speed is not regular, but shows a diminishing rate 

 as the speed increases. Thus, the number of steps was 8.4 greater for 

 55 meters per minute than for 45 meters per minute, or an increase of 

 0.84 step for each meter per minute increase in speed, and 0.82 step 

 when the speed became 65 meters per minute. The increase to 72 and 

 77 meters per minute was accompanied by an increase of practically 

 0.62 per meter increase in speed in each case. With E. D. B., there- 

 fore, the increase in the speed of walking appears to have been more 

 nearly taken care of at the lower speeds by an increase in the number 

 of steps, but with the higher speeds this became a lessening factor. 



STEP-LIFT DURING HORIZONTAL WALKING. 



In considering the data recorded in tables 29 to 33 for the elevation 

 of the body, or what we have termed the step-lift, it is recognized that 

 there are considerable variations in this factor for the same subject at 

 the same speed on different days, and also that substantial differences 

 appear at times for the same subject in the periods for the same day. 

 This would naturally lead to a questioning of the technique by which 

 the measurements were made. 



POSSIBLE CAUSES FOR VARIATION IN STEP-LIFT. 



-The most apparent fault in the technique which would lead to these 

 differences in the records would be a failure to have the fork of the 

 recording device (see fig. 1, p. 19) held firmly against the shoulder of 

 the subject, thereby not giving the full effect of the step to the counter. 

 It is, of coarse, possible that an error may have occurred in this way in 

 a few instances, but as it was recognized that such difficulty might occur 

 we were especially careful to be on the watch for it. Another source of 

 error would be in the slipping of the cord on the periphery of the wheel 

 which operated the counter, This would naturally produce too low 

 a registration. This, we know, did occur in a few instances, and in 

 such cases we have made use of the kymograph record in estimating 

 the lift. After such conditions were discovered, it was the practice to 

 rub a little powdered rosin on the cord at the beginning of each period. 

 Even with this precaution and when no slipping was apparent, dif- 

 ferences were still found in the elevation. It seems probable, there- 

 fore, that this difference was due to the gait of the subject, produced 

 either by more shoulder-motion or by a lateral swaying of the body 

 or by a difference in the absolute lift. It should be recalled that if 

 a subject, while taking 100 steps a minute, lifted his body 1 cm. a step, 

 it would require a displacement of only 2 mm. from any cause to produce 

 a variation of 20 per cent in the final reading. That the subject 

 changed his gait by changing the number and length of his steps has 

 been shown in the preceding section, but how much this change in the 



