TIME OF OBSERVATIONS. 35 



evening; the particular color of the sun and of the moon at their 

 rising, especially in fair days. 



Heat lightnings without thunder, and sometimes without clouds ; 

 indicate their direction and the aspect of the clouds in their neigh- 

 borhood. 



The Aurora Borealis, or northern light, for the observation of 

 which the special instructions, page 48, must be followed. 



Shooting-stars. The observer must be particularly attentive 

 to their frequency, during the periods near the 10th and llth of 

 August, and the 10th and 15th November, in which it is sup- 

 posed that they are more numerous than at any other time. He 

 will designate the quarter of the heavens from which they seem 

 to issue, and their direction. 



Fireballs. Describe their aspect, their size, their course in the 

 heavens, and note the exact hour of their appearance. 



All the other luminous phenomena, which present any extra- 

 ordinary appearance, should be noted down. 



These descriptions should be made in simple and well-defined 

 terms. The observer will take great care to enter scrupulously 

 what he sees without drawing any conclusion, or attempting any 

 explanation of the phenomenon. He ought to reflect that, in 

 order to make a good observation, he must keep his mind in a 

 state of perfect freedom in respect of any preconceived theory, 

 and to consider the phenomenon before him as being one of the 

 data for the foundation of the science, and that the knowledge of 

 the truth will depend upon the fidelity of his observation. 



TIME OP OBSERVATIONS. 



The time of observations will be the mean time at each station 

 The observations will be made three times daily, viz: 

 At 7 o'clock a. m. 

 2 " p. m. 

 9 " p. m. 



The mean of these three hours will be very nearly the true 

 mean, as it would be obtained by observation made every hour 

 of the day and night. 



The rain gage will be observed only once a day, unless very 

 abundant rains should make a second measurement necessary. 



