MOUNT WILSON OBSERVATORY. 259 



source should be measured, rather than the emission Hnes of the carbon 

 arc. The experiments required the use of a plugged tube in the furnace 

 at atmospheric pressure and at very high temperatures in order to 

 give sufficient continuous ground in this region. The long exposures 

 needed for the high dispersion and the slow plates imposed a severe 

 strain on the furnace, and after some partially satisfactory results were 

 obtained, the special furnace described above was constructed. Strong 

 photographs were then obtained without difficulty, temperatures of 

 3,200° C. and higher being employed. A set of these plates with iron- 

 arc comparison is now ready for measurement. 



An examination under high dispersion of the X3883 band in emission 

 for the furnace and arc spectra showed clearly an effect which had been 

 noticed on spectrograms of lower dispersion, namely, that the com- 

 ponent hnes of the band do not behave alike when the excitation is 

 altered. Lines belonging to certain series are relatively stronger in 

 the furnace than in the arc, and reverse in the furnace more easily than 

 lines of other series which are strong in the arc. Spectrograms of 

 moderate dispersion indicate that the same condition prevails in the 

 band at X5165, and it is probably general in the carbon spectrum. 

 This phenomenon has direct apphcation to questions involving the 

 wave-length of these band-lines in different sources, since, if there is 

 an analogy with the line-spectra of metals, the lines relatively strong 

 in the furnace are less subject to variation of wave-length under condi- 

 tions which produce displacements. An effort will be made to deter- 

 mine how far the effect is a result of temperature difference. 



STANDARDS OF WAVE-LENGTH. 



The interferometer program for determination of standards of wave- 

 length in the spectrum of iron has been completed by Mr. Babcock for 

 the region X3370-X6750 and the results are now being prepared for 

 publication. About 610 iron hnes have been observed in this interval, 

 many of them on a large number of photographs. Various orders of 

 interference have been used and much care has been given to the elim- 

 ination of mistakes. The discussion of this observational material 

 is not complete, but it is apparent that for good lines the accordance 

 with other observers is very satisfactory, while in the case of sensitive 

 hnes our interferometer measures show clearly a systematic difference 

 when compared with observations which are influenced by pole effect. 



The extension of our wave-length standards toward both shorter and 

 longer wave-lengths is in progress. Fifty-two photographs of the 

 iron-arc spectrum have been taken with the interferometer in con- 

 junction with the large quartz spectrograph, covering the region X2270- 

 X4500. The final reduction of these plates is not yet complete, but 

 the best of them have been measiu-ed. In the other du-ection five 

 photographs have been made of the interference spectrum of iron from 

 X5500 to about X8000. These have been measured and reduced, and 



