454 Journal of Comparative Neurology and Psychology. 



ently only a few of the cilia at the anterior tip of the left side 

 change their direction of stroke ; with a stronger stimulus the 

 number is greater. It is possible further that the amount of 

 change in the stroke of individual cilia is variable ; from some 

 of my observations I believe this probable. But whatever the 

 nature of the variation, the following results are produced. If 

 the effective beat of these left cilia is only slightly changed, the 

 anterior end then describes but a small circle, as in Fig. 8, a. 

 As the effective beat of these cilia changes farther, the swerv- 

 ing becomes stronger and the revolution slower, so that the an- 

 terior end swings in a larger circle (Fig. 8, d). Finally all the 

 cilia beat toward the oral side ; the revolution on the long axis 

 has then entirely ceased, while the swerving toward the aboral 

 side is very rapid. As a result the anterior end describes the 

 circumference of a circle, in the radii of which lies the long 

 axis of the body (Fig. 9). Thus the swerving toward the 

 aboral side varies inversely as the rate of revolution on the long 

 axis. In the unstimulated swimming the revolution is rapid 

 and the swerving slight ; in the strongest reaction the revolution 

 is zero and the swerving is strong, while between these two ex- 

 tremes an indefinite number of gradations exist. The change 

 in the forward stroke of the cilia seems more nearly independ- 

 ent of the two interconnected sets of changes just described. 

 The rapid forward swimming may be combined with the mini- 

 mum of swerving and the maximum of rotation ; the animal 

 then shoots rapidly forward. On the other hand, the forward 

 swimming may either entirely cease, or be converted into a 

 backward movement, in combination with the same minimum 

 of swerving and maximum of rotation. In the former case the 

 animal merely rotates rapidly on its long axis, neither advanc- 

 ing nor retrograding ; in the latter case it shoots rapidly back- 

 ward. But whenever the swerving toward the aboral side be- 

 comes largely increased, the longitudinal motion seems to de- 

 crease ; this is probably a necessary consequence of the fact 

 that the effective stroke of many of the cilia is in this case lat- 

 eral, so that only a comparatively weak component is left for 

 movement along the long axis. While swerving strongly, how- 



