RECENT ADVANCES IN SCIENCE 357 



extend the ultra-violet spectrum down to X 1,200, a limit set 

 by the absorption of fluorite. 



Schumann's apparatus was not suitable for the accurate 

 measurement of the short waves, because of the complex dis- 

 persion curve of fluorite in this region, although admirably 

 adapted for general exploration purposes. This difficulty was 

 removed by T. Lyman, who, by employing a concave grating 

 of speculum metal, paved the way for a further advance. The 

 grating spectroscope was contained in a vacuum chamber con- 

 taining hydrogen at a low pressure, the light from the discharge 

 tube being admitted through a fluorite window. Special plates, 

 prepared according to Schumann's methods, were used. With 

 a disruptive discharge, an extension of the spectrum down to 

 about A, 900 was secured, and fairly accurate measures of the 

 principal lines of many substances in the new region were made. 



The work of Laue, W. H. and W. L. Bragg, and others on the 

 diffraction of X-rays by crystals has shown conclusively that 

 such rays are a form of ether vibration with the extremely short 

 wave-length of about one angstrom. Lyman's work, therefore, 

 left a gap of about nine hundred units between the ultra-violet 

 extension of the visible spectrum and the X-ray region. 



A remarkable advance towards bridging this gap has recently 

 been secured by R. A, Millikan in the Ryerson Laboratory of 

 the University of Chicago {Physical Review, 12, 168, 191 8 ; 

 Science, 19, 138, 191 9 ; Astroph. Journ., 52, 47, 1920). The 

 essentially new features in Millikan 's work consist in (i) the 

 employment of a very perfect vacuum, which eliminates all 

 absorption by residual gases ; (2) the use of a source con- 

 sisting of very high potential sparks beween metal electrodes 

 placed very close together, so insuring the production of the 

 highest frequencies ; (3) the use of special gratings, so devised 

 as to throw as much light of short wave-length as possible into 

 the first-order spectrum, and with sufficient regularity of ruling 

 to produce good images when the ratio of grating-interval to 

 wave-lengh is as much as 70, instead of having a value of 3 to 

 6, as employed in common practice, or of about 20, as in 

 Lyman's work. The gratings used were ruled in the Ryerson 

 Laboratory, and are the outcome of long experience in grating 

 ruling ; it is probable that the limitations imposed by the use 

 of a grating have been reduced to a minimum. It is considered 

 by Prof. Millikan that all possible question as to the validity 

 of the results has now been removed : a detailed account of the 

 various difficulties encountered and the methods of over- 

 coming them is left for future publication, but the announce- 

 ment is made that the ultra-violet spectrum has been carried 

 down to a wave-length of 202 angstroms. 



The spectra were obtained by intermittent sparking between 



