PRECIPITATION STRUCTURES SIMULATING ORGANIC GROWTH. 255 



croscope. Masses of amorphous precipitate were formed on 

 Mn and Al, but Ag never showed any reaction, although left in 

 the ferricyanide solution for hours at a time. 



IV. Conditions Modifying Form and Rate of Growths. 



i. Electric Current. (a) The suggestion had been made that 

 rate of growth of filaments could be accelerated or inhibited by 

 electric influence. 1 Iron wires were made the electrodes in a 

 ferricyanide solution by connecting to a battery through a 

 rheocord, pole-changer and key. The strength of current re- 

 quired to affect the rate of growth is very small. By use of the 

 rheocord the potential of the current could be altered at will. 

 The general results were that both the rate of growth and the 

 number of structures formed were accelerated or increased at 

 the anode (Fig. 22) and retarded or decreased at the cathode 

 (Fig. 23). By use of the pole-changer the current could be sent 

 through the solution in the opposite direction, and the pro- 

 moting and inhibiting effects would then show on the opposite 

 wires. This experiment indicates clearly the electrical nature 

 of the conditions controlling precipitation-formation from 

 metals. 



(b) A galvanic current from a storage battery was sent through 

 a fine iron wire immersed in the solution. The result was both 

 to increase the rate of growth and to change the shape of the 

 structure. Upon making the current the rate of growth is 

 increased and the growing filaments bulge to almost double their 

 original diameter. By alternately making and breaking the 

 current the filaments could be made to resemble a string of 

 beads (Figs. 24 and 25). These effects are most probably due 

 chiefly to the heating eflects of the current. 



2. Sudden Changes of Concentration of the Solution. By re- 

 moving the solution from around a filament-forming metal 

 and then immediately replacing it with water the growing ends 

 of filaments are caused to swell out in a manner shown in Fig. 

 26. After such an experiment growth usually ceases, and the 

 walls of the filaments soon break down and fall apart in a peculiar 

 manner as shown in Fig. 27. These are osmotic effects, and 

 show that the walls of these structures are semipermeable. 



1 Lillie, loc. cit., p. 157. 



