248 AUROEA BOREALIS. 



eirable from a series of stations stretcliing along the entire northern frontier of 

 the United States, and the observations shoukl be continuous from year to year. 

 They should embrace a record of all auroras, even the faintest, and should state 

 concisely their degree of brightness. Such records Avill, in time, enable us to 

 decide the questions which may arise respecting the periodical character of these 

 phenomena. 



2. Whenever well-defined auroral arches are observed, it is important to locate 

 accurately their east and west extremities with reference to the magnetic me- 

 ridian. We may thus find that the vertex of the ai'ch does not ordinarily lie 

 cxacth/ in the magnetic meridian. 



3. Whenever auroral streamers are noticed, which are so peculiar that they 

 may be easily identified, locate them accurately with reference to neighboring 

 stars, and determine in what direction they move, and with what angular velocity. 



4. Record carefully the precise time of every auroral observation. 



