L. Bell — Rainband Spectroscopy. 353 



in conjunction with the usual meteorological observations, can 

 hardly fail to give indications of real value. Alone it appears 

 to be of about the same order of usefulness as the barometer 

 alone and unaided, but in connection with other instruments it 

 certainly deserves to be generally used, and ought to be added 

 to the equipment of meteorological stations. Its cost is quite 

 small — an instrument like the writer's, fitted with a similar 

 scale, costing not more than $20. In practice a pile of glass 

 plates is cheaper and better than a Nicol prism, since it polar- 

 izes less perfectly and thus allows the use of a more extended 

 scale. 



If an untrained observer begins rainband observations he is 

 almost sure to be, at first, disappointed with the results. He 

 procures a small pocket instrument of insufficient dispersion 

 and goes to work under the impression that a daily observation 

 will enable him to predict the weather for the next twenty- 

 four hours or more with almost unerring certainty. Naturally 

 he fails, and the more thoroughly he has believed the claims 

 made in various popular articles the more bitterly he abuses 

 the method. And very absurd claims have been made for it. 

 There may be some gifted persons to whose eyes the rainband 

 spectroscope becomes an oracle to reveal the meteorological 

 future, discriminate between hail, snow and rain, designate 

 the amount of each, predict the electric potential of an ap- 

 proaching cloud, indicate the per cent of ozone (!) in the atmo- 

 sphere, and in general answer the scientific prayers of the 

 astute suppliants. But the writer is not one of this fortunate 

 number. If there is any method of predicting the kind of 

 precipitation which is to occur, or of discovering the afore- 

 mentioned potential, he would like to see it demonstrated in 

 black and white, so that it will not require the eye of faith to 

 comprehend it. 



The rainband spectroscope, like most other scientific instru- 

 ments, is capable of giving good results in competent hands, 

 but there is no jugglery about it, neither is it the universal 

 solvent of meteorological difficulties, and success in its use can 

 be obtained only by frequent and careful observations, accom- 

 panied if possible by observations with other instruments. 



For practical work one of the larger pocket spectroscopes 

 should be chosen, giving a bright well defined spectrum with a 

 moderate amount of dispersion. For use with the scale de- 

 scribed the dispersion should not be great enough to resolve 

 the rainband into distinct lines, but should show it simply as a 

 dull band, so that it will be comparable to the interference 

 bands. If the individual lines are to be studied a much larger 

 dispersion is needed — enough to divide D without difficulty. 

 The instrument should be used with a rather narrow slit and 



