40 BULLETIN 1179, U. S. DEPARTMENT OF AGRICULTURE. 
rundum arch of the furnace crucible. A water curtain was caused 
to flow over this shaft by means of a perforated pipe coiled around 
the upper part of the steel shell, this water being caught in the 
bronze basin mentioned above, from which it flowed by gravity to a 
steel tank holding 7,000 gallons and was returned to the furnace by 
a circulating pump. Individual steel plates (without any water-cooling 
devices) were bolted to the sides and ends of the crucible, it having 
been learned from previous experience that only the center of the 
crucible arch was appreciably affected by the temperatures attained. 
The first of the tests described below was made with this modified 
furnace, but while the water-cooled shaft and bronze basin answered 
the purpose for which they were intended admirably, leakage of the 
flame and gases through the joints of the crucible caused overheat- 
ing and buckling of the steel plates, a condition which grew worse as 
the run continued. Considerable difficulty was also caused by the 
breaking through of the slag at the tap holes, which made it apparent 
that it was not only necessary to have a gas-tight crucible but water- 
cooled cinder notches as well. While this test was only continued 
efficiently for five and a half hours after the furnace was first charged, 
a brief summary of the operation is given below. 
During the period between 2.40 and 5.15 p. m. July 26, 1922. 1,175 
pounds of briquets having a phosphoric acid content of 15.65 per 
cent were charged to the furnace, this being largely tapped off at 
5.40 p. m. The oil pressure during this period was maintained at 45 
pounds, equivalent to a consumption of 17 gallons per hour, and the 
temperature of the preheated air delivered by the stoves to the oil 
burners was 135° C. The average phosphoric acid content of this slag 
was 0.72 per cent, which is equivalent to a volatilization of 96.7 
per cent, or 3.5 pounds per gallon of fuel oil consumed. This 
evolution was surprisingly high in view of the fact that the temper- 
ature recorded on the hearth at the time of tapping was only 1,470° C. 
While this slag was being run off, 200 pounds more of briquets were 
added, but unfortunately almost immediately after the furnace was 
tapped an explosion occurred which blew off the top of one of the 
stoves, thus entailing a shutdown for a considerable period and a 
cooling of the furnace. Between 6 and 9.30 p. m. the furnace was 
only operated spasmodically, but during this time 450 pounds of 
briquetted charge were added. Much of this froze in the shaft, 
causing a stoppage of the draft and rendering it impossible to melt 
this charge out of the shaft with the oil flame, although attempts were 
made to dislodge it until 2.30 the following morning,, at which time 
the burners were cut off and the furnace closed down. 
Before the last test described below was undertaken a steel shell 
varying in thickness from one-fourth to three-eighths inch was fitted 
around the furnace crucible, the joints being acetylene welded and 
the junction between the shell and the bronze basin braized. This 
change rendered the crucible and shaft of furnace gas tight. Water- 
cooled cinder notches of cast iron were inserted in the slag holes and 
bolted to the steel jacket. A system of perforated pipes made it 
possible to run a water curtain over the entire furnace shell. This 
water ran into a 6-inch steel trough riveted to the base of the fur- 
nace jacket and from here flowed into the large tank previously 
mentioned, from which it was returned to the system. From the 
standpoint of durability this furnace has proven eminently satis- 
factory, as after the last experimental run, lasting 70 hours, during 
