352 L. Bell — Rainband Spectroscopy. 



amount of aqueous vapor in the entire thickness of air in the 

 direction of observation, it tells nothing of its distribution 

 along that direction. Thus without any increase in the rain- 

 band there may be a precipitation of rain due to a descending 

 mass of vapor. So also clouds may or may not affect the rain- 

 band, though as a rule they do. The amount of vapor present 

 is often quite independent of the amount and kind of visible 

 clouds. Now and then the spectroscope will show dense masses 

 of vapor where the eye can detect at most only the lightest 

 cirri. Again, toward the end of a storm the sky may fre- 

 quently be overcast with threatening clouds when the rainband 

 clearly shows that the storm is over. Indeed the close of a 

 storm is indicated quite as certainly as its beginning, and in 

 this fact lies one of the great advantages of the spectroscopic 

 method. A typical case is shown in the storm of May 1st, 

 when the hygrometer reached its maximum some twelve hours 

 after the rainband had fallen far enough to indicate the cessa- 

 tion of rain. 



It is difficult, if not impossible, to designate the intensity of 

 rainband that may be regarded as a sure precursor of a storm. 

 Certainly no general rule can be laid down, for at present no 

 two scales are alike and each locality probably has its own 

 peculiarities. For the instrument used by the writer in Balti- 

 more, a rainband corresponding to 4 of the scale generally 

 meant rain within twenty-four hours. With the limited scale 

 possessed by the instrument it was best to register the faint 

 traces of a rainband as 0— too small to measure easily, al- 

 though in a larger spectroscope one might venture on an esti- 

 mate. It is very desirable that the readings of intensity should 

 be reduced to an absolute scale, in which case observations 

 taken at all points would be comparable and might lead to the 

 discovery of the laws which govern the distribution of aqueous 

 vapor. A knowledge of these could hardly fail to be of great 

 value to the science of meteorology. 



With the polarization scale herein described it is quite possi- 

 ble to determine the value of each scale division in per cent of 

 light absorbed, by a careful comparison by means of the 

 spectro-photometer, since the interference bands are wide 

 enough to be easily measured. The coefficients of absorption 

 once determined it would be possible to reduce them to terms 

 of standard column of aqueous vapor, and then the problem 

 would be well in hand and rainband observations would mean 

 something more than mere qualitative estimation. Until these 

 quantitative measurements are generally made, the spectro- 

 scope can be of comparatively small service except as an aid 

 to the prevision of local storms. 



For this latter purpose it is certainly effective and when used 



