144 MOVEMENT OF CILIA AND FLAGELLA 



passing up the cilium were to work against this and straighten 

 out the ciUum, then the elasticity of the ciHum might be respon- 

 sible for the return of the cilium to the resting position. It is 

 interesting in this respect that an abfrontal cilium in a hypertonic 

 medium (3g NaCl in 100 ml seawater) showed normal amplitude 

 and normal duration of the preparatory stroke, but the duration 

 of the *' effective " stroke was nearly twice the normal value 

 (Kinosita and Kamada, 1939). In addition, the resting periods, 

 which occur from time to time, always occur between effective 

 and recovery phases, i.e. when the cilium is in the contracted 

 position according to the usual theories. Most cilia rest in the 

 relaxed position if they rest at all. 



The calculations of Harris (1961) (quoted at length on pp. 

 146 ff.) have a bearing on the interpretation of the beating of 

 these cilia. Harris deduced in this work that, for a single cihum, 

 the time taken for the effective stroke should be proportional to 

 the cube of the cilium length, while the time for the recovery 

 stroke is related to the time taken for the wave of flexure to pass 

 up the length of the cilium, and is thus proportional to the cilium 

 length. In a long cilium, therefore, the effective phase may 

 occupy a much larger fraction of the whole beating cycle than in 

 a short cilium. These abfrontal cilia are compound, which 

 modifies this picture since more energy is available for moving 

 the cilium against water resistance; the resistance to bending is 

 assumed to be always proportional to the bending force, i.e. the 

 same ratio for a compound cilium as for a single one. These 

 arguments may be part of the answer, but cannot provide the 

 whole answer, for abfrontal cilia showing the same form of beat 

 may vary in length from some 20 /x to 100/x, and other cilia within 

 this size range may occur on the gill and show a quite different 

 form of beat (compare series (a) with series (b) in PI. XVII), 



From this extensive discussion it must be obvious that the 

 beat of these cilia can not be explained by any conventional 

 theory, and it would be wise if further experimental work, as 

 well as an electron microscopic study of structure, were to precede 

 any further speculation. 



Some of the long cilia that occur among the frontal cilia on 

 Mytiliis gill filaments show a form of beat that appears identical 

 with that of the normal frontal cilia, and is easier to see. The 



