426 L. Bell — Ultra-violet Spectrum of Cadmium. 



become luminous, the forms of the bodies become rounded in 

 front, and the paths described are straight lines. 



9. The meteoroids encountered by the earth on the 27th of 

 November, in 1872, and in 1885, did not leave the immediate 

 neighborhood of the Biela comet earlier than 18-11'5, and may 

 be treated as having at that time orbits osculating that of the 

 comet. The determination of the paths of these meteoroids 

 through their five and seven last revolutions about the sun 

 seems to be a problem capable of complete solution. 



Art. XL. — Tlxe Ultra-violet Spectrum of Cadmium; by Louis 

 Bell, Fellow in Physics of the Johns Hopkins University. 



As the ultra-violet spectrum of cadmium has long served as 

 a standard of reference in the measuring of other spectra it lias 

 seemed desirable to determine its principal wave-lengths with 

 an accuracy greater than has been heretofore reached. Thus 

 far all determinations of ultra-violet wave-lengths have de- 

 pended, in general, upon the method devised by Cornu, that of 

 photographing upon the same plate the lines in question and a 

 series of direct reflections of the slit corresponding to known 

 angles of deviation. 



While this process is an ingenious one and has done good 

 service in preliminary work, it is clear that the accuracy of 

 which it is susceptible is decidedly limited. For in the first 

 place it is liable to the errors that always attend the determina- 

 tion of an angle by a small number of measurements and in 

 addition to the numberless minor difficulties that must attend 

 the taking and measuring of small photographs when the dis- 

 persion used is not large and the focussing is mainly a matter 

 of experiment. A casual examination of the wave-lengths 

 given for the same lines by various experimenters will give a 

 clear idea of the difficulties and faults of the method. Even 

 the numbers given b}^ such careful workers as Cornu and Hartley 

 not infrequently differ by as much as one part in five thousand. 



A radically different method has been used in making the 

 determinations reported in the present paper. The apparatus 

 used has been the great spectrometer in the physical laboratory 

 of this university, armed with a concave grating of twenty-one 

 feet focal length and containing, in the space of six inches, 

 nearly 80,000 lines. An idea of the dispersion produced by 

 this instrument may be formed from the fact that photographs 

 taken directly in the focus of the third order are on a scale 

 somewhat larger than that of Angstrom's map. An eyepiece 

 micrometer with a very accurate screw two decimeters in length 

 answers for the exactness of all measurements made in the vis- 



