354 L. Bell — Rainband Spectroscopy. 



should be focused for sharp definition of D ; y will then ap- 

 pear as a narrow shading along the red edge of D if the 

 weather is fine, or a dark stripe if it is rain} 7 , and can be very 

 easily compared with the scale. 



The observations should be taken not less than twice a day, 

 one of them being as late in the afternoon as practicable, since 

 changes are so apt to appear about sunset. If convenient, ob- 

 servations should be taken at various altitudes. For a single 

 altitude 20° is perhaps as good as any. As to the question of 

 direction, it is a matter to be guided by local conditions. The 

 illumination should be as good as can be attained without 

 looking directly at or near the sun, and a quarter of the sky 

 should be chosen where this condition is fulfilled. Observa- 

 tions in all directions are to be desired ; but if only one is 

 chosen it should preferably be the one from which storms 

 usually come. An additional observation in the direction of 

 the wind is likely to be useful. Observe systematically the 

 same directions and altitudes. Avoid cumulus clouds when 

 practicable, as they are likely to increase the rainband without 

 a corresponding probability of rain. Employ clear skylight, if 

 possible without much variation of direction or altitude. 



More trust is to be put in a sharp rise or fall of rainband 

 intensity than upon absolute intensity, since the latter is likely 

 to average differently in various locations and be subject to 

 various modifjnng factors, while the former in nearly every case 

 means a definite change. With these and other precautions 

 which will suggest themselves to individual observers, the 

 rainband spectroscope is capable of giving ver} 7 " good results, 

 and the field of investigation is a decidedly promising one. 



For the convenience of those interested in the subject a brief 

 bibliography is appended comprising the more important pub- 

 lications which have appeared upon the subject. A few of 

 them refer to telluric lines in general. 



Xature, xii, 231, 252 ; xiv, 9; xvi, 389 ; xxii, 194. Piazzi Smyth. 

 Edinburgh Astronomical Observations, xiv. Smyth. 

 Scottish Meteorological Soc. Jour., N. S., Nos. 51, 52. Smyth. 

 Madeira Meteorologic. Knowledge, ii, No. 48. Smyth. 

 Proc. Roy. Soc. Edinburgh, xii. Mill. 

 The Rainband (Hilger. 1S83). Mill. 

 The Use of the Spectroscope in Meteorology TJ. S. 

 Signal Service Notes. No. iv, 1883. Upton. 

 Comptes Reudus, Ixiii. 293, xcv. 885. Janssen. 

 A Plea for the Rainband (Browning, 1881). Capron. 

 Science, ii, 488. Cook. 



Journal de Physique (2 me ser.), ii, 58, hi, 109. Comu. iii. 5. Thollon. 

 How to Foretell the Weather with the Pocket Spectroscope. (Chatto and 

 Windus, 1884.) Cory. 



