174 



PROCEEDINGS OF THE ACADEMY OF 



[1894. 



a end cannot go down on account of the string h d and the b end can- 

 not go down for the string a c. If rubber bands be stretched on 

 the same pins, then the upper horizontal piece will, if released, spring 

 forward very far and very fast compared with the actual shortening 

 ot the rubbers (fig. 2). 



Fig. 2. 



The model of the entire limb had all the merit of a first approxi- 

 mation. It imitated pretty closely the movement of the leg itself. 

 A second and closer approximation can now be had by considering 

 in terms of my first elements, the variations from Nature made in 

 them to gain simplicity of construction. In the first place, articular 

 surfaces are not circular, so that a pivot does not accurately represent 

 their motion. Neither do bones moving in constraint, one against 

 the other, keep in the same plane, but their axes describe w^arped 

 surfaces from the helical character of the articulations. 



There is a constant change in the relative rates of motion of the 

 joints, involving likewise a constant change in the relative leverage. 



The problem becomes instantly one of extreme difficulty, yet a fair 

 appreciation may be obtained. 



The leg should be studied in several positions not consecutive, but 

 so far apart as to cause decided changes in the relative rates of the 

 ditterent levers, and so small a part of the path or trajectory should 

 be considered, that a simple curve or straight line may, without 

 error, be substituted as in the manner of studying evolutes. 



