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



based on a measure of absolute wave-length made by Mr. C. 

 S. Pierce, which gives for D, 5896*22 against 5895*13 given by 

 Angstrom. Although the details of Pierce's work have never 

 been published, it is without doubt far more accurate than 

 Angstrom's. Pierce's value, however, is somewhat too large. 



If this difference of scale be taken into account the figures 

 are brought into somewhat better accord, though there will 

 still remain discrepancies. Prom the fact that these occur in 

 parts of the spectrum where the author's measurements were 

 direct and particularly free from sources of error, it seems prob- 

 able that many of them may be due to the unavoidable errors 

 in working with small apparatus by a method which several 

 sources of uncertainty tend to vitiate. , 



Most of the wave-lengths determined in the present research 

 are probably correct to within one fifty-thousandth part of their 

 respective values, and probably in no case does the error reach 

 0*1 tenth meter. A greater degree of accuracy can be attained 

 only by the use of a spark apparatus powerful enough to allow 

 one to photograph freely in the higher orders. 



The photographs were all taken on Stanley instantaneous dry 

 plates, developed by the ordinary ferrous sulphate process. 

 The average exposure was about thirty minutes. Thanks are 

 due to Mr. C. T. Child for efficient aid in the photographic 

 part of the experiments. 



A number of interesting points were noted during the pro- 

 gress of the work. The relative intensities of the lines, while 

 they agree upon all the negatives, often do not agree with 

 Hartley's estimates. This is probably due to the individual 

 peculiarities of the gratings, and shows how fallacious all such 

 estimates are liable to be unless made on the same apparatus. 

 It was curious to note how little the sensibility of the plates 

 decreased, even in the extreme ultra-violet. The curves ordi- 

 narily given for the sensibility of silver salts apparently depend 

 on sunlight, and hence really represent not so much the sensi- 

 tiveness of the salts employed as the intensity of the solar 

 spectrum after absorption by the earth's atmosphere. While in 

 photographing the solar spectrum the time of exposure in- 

 creases very rapidly in the ultra-violet portion, it must have 

 been noticed by every one who has photographed metallic 

 spectra that there is little trouble in reaching w. 1. 2000. This 

 fact is brought out with peculiar force in the author's negatives 

 of the superimposed spectra of the first and second orders. 

 Taking into account the fact that the first spectrum is some- 

 what the brighter there is really little difference in intensity 

 between w. 1. 4800 and w. 1. 2200. The maximum intensity 

 seems to . be between w. 1. 4000 and vv. 1. 3600 and at w. 1, 

 2200 it appears to be about half the maximum. These are of 



