J. M. Silliman—Evamination of the Bessemer Flame. 303 
te oer 
of this in the statements of others, and it may account for some 
of the first period at the beginning of a more decided flame. 
Tt comes flash 
of the other lines make their pppestanes in vivid flashes as 
1 ithi ree OF 
ized sodium to produce its spectrum, and thus occasion the 
flashes so characteristic in the first appearance of that line. 
Lithium exists in a much smaller quantity and requires y55}535 
