1903.] KOENIG—ARTIFICIAL PRODUCTION OF CRYSTALS. 221 
come to life, and one would make many more failures by either too 
much or too little heat. The wear and tear of the wire seems 
trifling when held against this loss. Being thus prepared the tube 
is ready to be charged. 
At A (Fig. 2) I place from five to ten grams of resublimed 
arsenic, on top of this a loose plug of asbestos (P). In the first 
experiments I thought copper filings would be the best material to 
act upon. These filings I poured on top of P, forming a column 
about one inch high, and secured this column by means of a second 
asbestos plug (P’). Such an arrangement of parts promised to 
restrain the arsenic vapors from passing by the copper without 
action. It proved an unnecessary precaution, as the copper acts 
toward that vapor as asponge toward water. Coarse turnings were 
tried instead of filings, and later solid copper bars with even better 
results than the filings had given. Similarly the close proximity of 
the copper to the plug P was found objectionable and, therefore, 
in all the later experiments the tube was placed, after charging the 
arsenic and inserting P, in a horizontal position by means of a 
clamp at the open end. Then the metal pieces to be acted upon 
were shoved into the desired distance from the plug, a loose asbes- 
tos plug next to the metal to avoid air currents, and finally a stop- 
per holding a narrow glass tube, bent at right angles, was inserted 
into the open end. The glass tube was then made to dip under 
mercury and thus expansion of the air made possible, without 
danger of air entering. Whatever oxygen was in the tube made 
As,O,, which was always found as a ring sublimate behind the 
metal. Fig. 1 shows the outside of the tube, clamped to the stand - 
S. The stout contact wires, w, w’, were found very serviceable. 
By their use the field of high temperature may be enlarged or 
restricted as well as shifted. These wires are simply laid upon the 
coil wire. Their position must not be shifted or altered without 
switching off the current; the thin wire will melt in the moment 
when the contact is broken. I spoiled considerable wire in this 
way, besides the time consumed in rewinding is quite an item. A 
suitable, easily changeable resistance to modify the tone of the 
