Experiments for ohtaining and preserving Potassium. 207 



are made of malleable iron, of a cylindrical form, containing about a 

 gallon and having an aperture at one end, closed by an iron plug by 

 means of a screw one inch in diameter. In order to prepare it for 

 use, the aperture must be bored out to receive a tube of at least an 

 inch and a half internal diameter, and connected by a screw which 

 accurately fits the bottle. If one much smaller than this be used, it 

 is soon choked and the process is impeded or completely stopped. A 

 convenient length for our furnace is sixteen inches. A more definite 

 idea of the furnace and apparatus may be obtained from the figure 

 on p. 212, which see. The lute for coating the retort, which after 

 many trials, was found to answer the best purpose, is composed of 

 equal parts of sand, pure clay and finely pulverized soapstone, mixed 

 up with water into a paste and this, with a little cut tow to render it 

 more adhesive, is applied in thin layers and each layer is dried in the 

 sun ; or if necessary it may be dried by a charcoal fire. This ce- 

 ment sustains the heat better than any I have tried. It adheres bet- 

 ter when the retort is previously bound with iron wire. 



Charging the retort, ^c. 



Although there is litde doubt that Brunner's process of adding car- 

 bon to the iron turnings is preferable to any which preceded it, yet, 

 not having had an opportunity of examining his original paper, I have 

 been obliged to ascertain the best proportion of the ingredients by re- 

 peated experiments. It is found, when a large proportion of char- 

 coal is used, that notwithstanding the metal sublimes at a lower tem- 

 perature, yet so much charcoal in the state of fine powder is driven 

 over with it, that the tube, unless much larger than the one described, 

 would soon be choked. If too little charcoal is used, the high tem- 

 perature required to decompose the potassa, would endanger the fu- 

 sing of the retort, either of which would cause the experiment to 

 fail. Hence the more charcoal is used in the process, the more ca- 

 pacity is required in the tube for condensing the vapors. 



The result of my experiments is, that for a tube and retort of the size 

 above mentioned, two ounces of pulverized charcoal, twelve ounces of 

 potassa, and the retort filled up with clean iron turnings, or what I think 

 still better, clean card teeth,* form the best proportion for the materi- 



* Damn2;cd cr.n] Icrt'i, which ;ins\vcr very well, can lie obtained at a reasonable 

 ritp of (he card niamifartuiers. 



