258 RALPH S. LILLIE AND EARL N. JOHNSTON. 



precipitate, although, under certain conditions the suiface 

 filaments show a regular cross striation, a feature probably 

 indicating periodic precipitation. 



Under some conditions not yet clearly defined small burr-like 

 structures are formed on the surface of an iron wire. Their 

 formation seems to depend more on the condition of the surface 

 of the metal than on the concentration of the solution. Some 

 of these structures are shown in Fig. 2. They are always very 

 small, and in the presence of a large number of filaments may 

 easily be overlooked. During their formation they show a 

 regular rhythmic motion causing the fine filaments on their 

 surface to wave to and fro with a cilium-like movement. In 

 some cases the whole structure pulsates, but in others only the 

 filaments show any motion. This movement is comparatively 

 slow, the rhythm being at first only about once every second, 

 and later slowing down to once every five to eight seconds; 

 sometimes this rhythmical motion continues for ten minutes. 

 In the formation of zinc ferricyanide filaments very interesting 

 rhythmical phenomena are exhibited. A small strip of zinc, in 

 contact with a nobler metal (Cu, Fe), when placed in a ferri- 

 cyanide solution becomes entirely covered with a fungus-like 

 coat of filaments in a few seconds. At the extremity of each 

 filament there is visible a regular rhythmic motion or pulsation, 

 very similar to that described above as accompanying the 

 formation of sac-like structures from zinc filaments at the surface 

 of the solution. An almost transparent membrane swells out 

 from the end of the filament, remains an instant, and then falls 

 back quickly; then again the membrane slowly pushes out and 

 the process is repeated. There is no visible evidence that the 

 membrane is ruptured at any stage of the motion; neither can 

 any suspended particles of precipitate be seen to flow from the 

 end of the filament. The filament slowly increases in length, 

 and the rate of increase seems to be proportional to the rate of 

 rhythmical motion. It seems probable that each time the 

 pulsating membrane falls back, it has just been ruptured at its 

 edges, some precipitate being then formed and deposited. The 

 rate of the motion varies from about once per second in the 

 larger filaments to five or six per second in the small ones. No 



