226 CARNEGIE INSTITUTION OF WASHINGTON. 



Two more adapters to permit the use of ordinary measuring machines 

 for observing rectangular coordinates have been constructed. These 

 are especially useful in the reduction of interferometer plates. 



A small 200-line half-tone screen has been provided for use in the 

 study of intensity distribution in spectral lines by the method of 

 Nicholson and Merton, and a double logarithmic sector for obtaining 

 the necessaiy steep intensity gradient has been constructed. 



A simple glass mercury still and a new vacuum pump have been 

 added to the laboratory equipment. 



ANOMALOUS DISPERSION WITH THE ELECTRIC FURNACE. 



The preUminary experiments on anomalous dispersion by Mr. King, 

 mentioned in last year's report, showed the effectiveness of the electric 

 furnace in producing this phenomenon, especially with vapors of the 

 more refractory substances. A more extended investigation has been 

 carried out during the past year, in the course of which 254 photo- 

 graphs were made in a study of the effect for 12 elements. In the 

 spectra of iron, chromium, and titanium, the relative degree of anoma- 

 lous dispersion was recorded for those lines which showed the effect, 

 namely, 73 lines of iron, 47 of chromium, and 102 of titanium. This 

 permitted a comparison of other lines with these lines of known behav- 

 ior in the furnace under various conditions. Lines of special interest 

 in the other spectra were studied in the same way. 



The general conclusion was that the anomalous dispersion shown bj' 

 a line is proportional to its strength in absorption, provided a proper 

 prismatic distribution is present in the absorbing vapor. The greater 

 prevalence of anomalous dispersion in the region of shorter wave- 

 length appears to result from the greater absorptive power generally 

 found for lines in this region. A point of special interest is the fre- 

 quent inversion of the anomalous refractive effect for a line when the 

 temperature is raised, indicating that the vapor prism absorbing such 

 a line has been inverted, the stronger absorption now taking place in 

 the cooler region above. At the same time, other lines, requiring higher 

 temperature for their production, show anomalous dispersion of regular 

 type. This simultaneous refraction in opposite directions for Hnes 

 of different character occurred many times when two or more elements 

 were vaporized together, and, in the case of calcium, even for lines of 

 the same element, the low-temperature hne X 4227 shov/ing an effect 

 opposite to that simultaneously given by the H and K lines. The 

 experiments have thus shown a dependence on the distribution of the 

 radiating particles rather than on the density of the vapor as a whole. 

 This conclusion was borne out by the observation of distinct anoma- 

 lous dispersion for the band at X 3883. On the assumption that this 

 band is due to nitrogen alone, the results showed that the strongest 

 absorption was given in the highly heated region, where the density of 

 the gas was at a minimum. 



