492 
PEOFESSOR W. N. HARTLEY AND ME. H. RAMAGE ON THE SPECTRA 
experiments were made in which, first, a very small quantity of the potassium 
compound was introduced into the oxyhydrogen flame; secondly, the potassium salt 
was volatilised in the oxygen and carhon monoxide flame. The line was of greater 
relative intensity in the latter spectrum, but it.still did not equal that observed in the 
Bessemer flame. 
There are two possible explanations of the reason for the disappearance of this line : 
first, its emissive power and its chemical action may be greater at lower temperatures, 
and, therefore, its jihotographic intensity greater ; and, secondly, it may be that when 
the flame is less dense, so that the vapour pressure of the metal is decreased, the 
molecules are endowed with greater freedom of motion. 
The following photographs were taken with the object of ascertaining Avhether 
the density of the vapour in the flame really affected the intensity of the line 4642 :— 
Plate 103, Sj^ectritin (2.)—Paper moistened with a solution containing OT gramme 
of potassium chloride. Made into a coil, and burnt in the oxyhydrogen flame. 
Spectrum (3).—The same as (2), but the paper was burnt in thin strips. 
Spectrum (4).—Similar to (3), but 0'2 gramme of potassium chloride was present. 
Plate 104, Spectrum (3).—Potassium chloride as in 103, Spectrum 3, but about one- 
half the quantity taken. 
Spectrum (4).—Potassium chloride, about one-quarter of the quantity used for 
photographing Spectrum (3) on Plate 103. The vapour of potassium in the flame is 
known to be proportionally increased with the larger c[uantity of substance volatilised, 
and the intensity of the chemical action or emissive power of the continuous rays in 
these spectra is observed to be in the inverse proportion to the vapour pressure of the 
substance. With reduction of vapour pressure there is a diminution of chemical 
action exerted l)y continuous rays over a wide range of differences in wave-length, 
and this is accompanied by a greater intensity of chemical action or emissive power of 
the molecules for the ray X 4642. 
The question of the temperature was next examined, as up to this time the line had 
not been observed in a Bessemer flame. 
Plate 373, Spectrum (1).—Spark spectra of metallic lines of known wave-lengths 
for measurements. Exposure, 30 seconds. 
Spectra (2), (3), and (4).—Potassium chloride heated on a support of cyanite. 
(2.) In the flame of hydrogen. Exposure, 1|- hours. 
(3.) In air and hydrogen, using the same burner as that for the oxyhydrogen flame. 
Exposure, 15 minutes. 
(4.) In the oxyhydrogen flame. Exposure, 1^ minutes. 
(5.) In the oxyhydrogen flame. Exposure, 5 minutes. 
Two red lines were seen l^y using a small direct-vision sj^ectroscope. These are 
doubtless the doublets and which have been observed in the arc 
spectrum of potassium. 
