276 CARNEGIE INSTITUTION OF WASHINGTON. 



Institution of Washington and deposited by him in the Pasadena labor- 

 atory. A Babinet compensator has also been purchased for use in test- 

 ing mica preparations and for the study of polarized Ught in general. 



A new compound X/4 plate has been constructed and tested with 

 the Babinet compensator. It has about 100 mica sections, each 1 mm. 

 wide, adjusted for a relative retardation of X/4 for X6350. As far as 

 X4925 its retardation is practically a Unear function of the wave- 

 length. Retardations have also been measured for several of our other 

 mica preparations. 



A measurement has been made of the angular error between the axis 

 of the screw of the ruling-engine and the plane of the ruby which is 

 used as an end-thrust bearing. By means of a concave mirror of 36 

 feet radius of curvature, a high-power micrometer eyepiece, and a 

 method involving twice the angle under consideration, it was found 

 that the angle differs from 90° by about 0':05. 



ELECTRIC-FURNACE SPECTRA. 



The investigation of electric-furnace spectra by Mr. King has been 

 continued for elements of special interest in the Observatory work, a 

 detailed study of the cobalt and nickel spectra having been made 

 according to the plan followed for other elements, together with a set 

 of plates for the comparison of the effect of different vapor densities 

 upon certain hues in the iron spectrum. 



The photographs of the furnace spectra of cobalt and nickel have 

 covered the range from X3000 to X7100 for temperatures ranging from 

 1850° to 2600°. Three complete sets of spectrograms, taken at approxi- 

 mately 2000°, 2300°, and 2600°, respectively, have been used to classify 

 the spectrum lines according to the temperature at which they first 

 appear and their rate of increase in intensity as the temperature rises. 

 The Unes classified on this basis number 840 for cobalt and 423 for 

 nickel. The leading features of these spectra are similar to those 

 previously noted in the study of other elements. For each of them, the 

 furnace spectrum in the region of shorter wave-length is rich as com- 

 pared with the arc spectrum, while toward the red the differences be- 

 tween arc and furnace spectra, especially for nickel, are very striking. 

 As indicators of the temperature conditions in other sources, par- 

 ticularly various solar regions, the variations of the cobalt and nickel 

 Unes at different furnace temperatures promise to be of much service. 



In the detailed pubUcation, consideration is given to lines showing 

 unusual types in the furnace, to the behavior of certain cobalt Unes 

 which have been classified as enhanced, and to Unes relatively stronger 

 in the furnace than in the arc. 



THE TUBE-ARC SPECTRUM OF IRON. 



A survey of the iron spectrum as given by the "tube-arc" was 

 carried out by Mr. King following the installation of the 100-kw. 

 transformer. The tube-arc is obtained by causing the resistor tube of 



