STRUCTURES RESEMBLING ORGANIC GROWTHS. 153 



striking if a metal of higher solution-tension than zinc is used, 

 such as aluminium, manganese, or magnesium. Magnesium is 

 especially effective, since the ferricyanide of this metal is soluble 

 and hence does not accumulate at the anodal regions and arrest 

 the inhibitory action by polarization ; such an iron wire remains 

 bright until the magnesium is completely dissolved; then fila- 

 ments grow out as usual. Apparently when zinc is used the 

 accumulation of zinc ferricyanide may interrupt contact between 

 the metals or render the resistance of the local circuit too high 

 (i. e., have a polarizing effect), and by degrees filaments form 

 from the iron. In addition to this effect, contact of the zinc ferri- 

 cyanide precipitate with the iron appears to form local couples in 

 which the iron acts as anode, so that iron in contact locally with 

 this precipitate usually exhibits in course of time a development 

 of blue filaments. This, however, is an incidental effect due to 

 the formation of another type of circuit. Pieces of aluminium or 

 manganese are almost as effective as magnesium in preventing 

 the formation of iron filaments in the above solution. A large 

 variety of experiments of this type may be performed with iron 

 wire. In a similar manner zinc may be used as the filament- 

 forming metal, and the process inhibited by magnesium, alumin- 

 ium, or manganese. 



As already mentioned, this inhibitory influence may be effec- 

 tive for a distance of several centimeters from the region of 

 contact of the two metals, and a gradient in the intensity of the 

 influence is apparent. After the current of the circuit has 

 decreased, as the result of the accumulation of precipitate or 

 other polarization, an iron wire with zinc connection often begins 

 to send out filaments; these tend to be first formed at the region 

 farthest removed from the zinc. Irregularities, however, are 

 frequently observed in such experiments, due no doubt to local 

 irregularities in the composition and physical condition of the 

 iron. The biological analogies to these effects will be considered 

 later. 



An interesting type of effect is seen when an iron wire 3 or 4 cm. 

 long is in contact at one end with a piece of zinc and at the other 

 with copper or platinum. With this arrangement gradients in 

 the formation of iron filaments are very distinct; at first these 



