may be hud on their side ami cai red :il.<mi without break- 

 ing the shoots. 



The solution is always alkaline, and as most of the salts 

 which grow well are also precipitated by caustic soda it 

 might be thought that perhaps we were merely dealing 

 with ordinary precipitates. It is found, however, that 

 adding solutions of salts of the metals to a solution of water- 

 glass gives altogether different results to those obtained 

 on dropping a crystal of the same salt into the solution. 

 In the former case a gelatinous precipitate usually forms, 

 but this disappears on shaking, or it may be is suspended 

 as a, possibly, colloidal solution. Id one or two cases the 

 precipitate remains, but these are not the salts which give 

 the best growths. Thus, uranium nitrate gives dense yellow 

 flocks on adding it to a solution of silicate, and these appear 

 to be similar to what is produced on dropping in crystals, 

 but there is scarcely any growth in the latter case. A fair 

 number of the salts while giving precipitates with caustic 

 soda are soluble in excess of the reagent. Yet such salts, 

 e.g., zinc chloride or sulphate, aluminic chloride, and cobalt 

 nitrate grow well and quickly. 



A more probable suggestion is that growth is due to the 

 formation of a semi-permeable membrane. We may con- 

 sider some phenomena which at first sight bear considerable 

 analogy to those with which we are dealing. If sulphate 

 of copper, powdered, be mixed with a little sugar moistened 

 to a paste, and then allowed to dry, a hard mass is formed. 

 A piece of this carefully dropped on to a solution of ferro- 

 cyanide of potassium often floats. In a few minutes a 

 brown shoot protrudes, and in the course of an hour or two 

 may grow to a length of three or four inches. The shoots 

 are, however, very tender, and a slight shake will generally 

 cause them to break off and fall to the bottom of the glass. 

 If the piece of sugar mixture sinks, it may grow up from 



