Z. Bell — Rainband Spectroscopy, 



349 



to have the selenite plate cover the entire slit, as the field is 

 then uniform and the obtrusive black line produced by the 

 edge of the plate is absent. 



In passing it must be remarked that the nomenclature of the 

 atmosphere lines is in a state of confusion thrice confounded. 

 Each investigator seems to have numbered or lettered them to 

 suit his taste. Only two telluric groups of lines have designa- 

 tions which are generally received, the band at w. 1. 628 (a) 

 and that at w. 1. 578 (o). In this state of affairs the writer 

 would suggest that for sake of brevity it would be well to 

 apply the letter y to the well known rainband near D, reserving 



A a B 



C a. yDcP 



E h 



ft for the faint band given in Angstrom's chart between a and 

 D, and using Brewster's lettering for the more refrangible part 

 of the spectrum. Then in referring to the individual lines 

 that go to make up y or any other of the bands it would be 

 well to simply number them in order of their wave length 

 —yl, y% etc. And a careful map of y is badly needed. 



After being arranged as above, the spectroscope was used 

 constantly for six months with complete success. The place 

 of observation was Baltimore, Md., a location well calculated 

 to give rainband observations a severe test, by reason of a cli- 

 mate more than usually variable. Having arranged a satisfac- 

 tory scale, the point to be determined was, how far rainband 

 observations alone are to be trusted as prophetic of the weather, 

 and how they should be taken with a small instrument like the 

 one used in order to secure maximum efficiency. 



As to the first point the annexed chart answers the question. 

 An observation was taken at 8 A. M. and another at 2 p. 1, 

 usually in the western sky. No other instrument was used to 

 confirm the indications of the spectroscope. From day to day 

 the instrument was now and then readjusted as regards slit, 

 focus, etc., and there was no difficulty in securing identical 

 scale readings after such readjustment. The instrument polar- 

 ized too well, so that the interference band could be made 



