( 275 ) 
belonging to the nearest vicinity of the absorption lines, that the 
above regular dispersion vanishes if compared with this anomalous 
dispersion. Waves lying on the red -facing side of absorption lines, 
and very near to them, we shall denote by R-light in this paper; 
neighbouring waves lying on their vio/tf-facing side, we shall call 
V-light. With the R-light the refraction constant has rapidly increasing 
positive values as we approach the line; with the V-light it has 
rapidly increasing negative values as we approach the line. 
The refracting power n — 1 of a gas is equal to the product If A 
of its refraction-constant into its density. 
Arago and Biot found that the refracting power of a mixture ot 
gases equals the sum of the refracting powers of the constituents: 
R m h in — RAx -f + etc ’ 
As with most kinds of light all gases have a positive R, the 
refracting power R m A m of the mixture is in general positive. But 
with V-light R m L m may be negative, because that term of the sum, 
which bears upon the constituent producing the absorption line in 
question, may with V-light have a negative value so great, as to 
surpass the sum of the remaining, positive, terms. It is presupposed, 
of course, that the A of the said constituent is not too small, or in 
other words: the vibrating system, producing the line which we 
consider, should not be too scantilj represented in the mixture. 
Near lines of principal constituents the anomalous refraction pre¬ 
dominates to such a degree, that the negative refracting power of 
the mixture with V-light is, in absolute magnitude, relatively little 
inferior to its positive refracting power with R-light. 
What are we to conclude from these facts with regard to the 
spot-spectrum? 
Let us first consider the case, that a region with a minimum of 
density is found just in front of the middle of the disk. Waves 
with which R m A m is positive, especially those belonging to the 
much refracted R-light, will in that region deviate conformably to 
the scheme of our drops of salt solution; the V-light, on the other 
hand, with which R m A m has great negative values, will in the same 
system of gradients deviate in the opposite way, that is, corresponding 
to the scheme of the glycerine drops. Both the R-light and the 
V-light are more weakened by dispersion than the rest of the spec¬ 
trum: the Fraunhofer lines must in general be widened or winged 
in the spot-spectrum. 
As the wings are caused by refraction, not by absorption, the 
