36 TIME OF OBSERVATIONS. 



The best time will be 2 o'clock p. m., the observation being made 

 daily ; if another hour is selected, it should, when once fixed, 

 remain the same. 



The maxima and minima thermometers will be read once a day, 

 always at the same hour. The most suitable hour will be 9 o'clock 

 in the evening. 



If an observer desires to examine Ihe daily oscillations of the 

 barometer, he will also observe at 10 a. m. and 4 p. m., which 

 give the daily maximum and minimum. It will be well to note 

 also, at the same time, the state of the hygrometer. 



If he desires to complete the data upon the diurnal course of 

 the temperature, he will add observations of the thermometer at 

 10 a. m. and 6 p. m. In all cases it is desirable that, if an ob- 

 server has leisure to increase the number of the hours of obser- 

 vations, he should fix them at equal intervals between the prin- 

 cipal hours indicated above. 



Besides these observations at regular hours, additional obser- 

 vations ought to be made during remarkable storms, as has been 

 remarked above. 



It is very important that the observations should be made at 

 the exact hour, fixed by a well regulated watch. All the instru- 

 ments should be read rapidly, so that the observations may be as 

 simultaneous as possible. 



Tlie order in which they are to be observed will be as follows : 



A few minutes before the hour, observe the thermometer before 

 opening the window ; then wet the psychrometer. While it is 

 taking the temperature of evaporation, n6te the height of the 

 barometer, observe the wind, the course of the clouds, their 

 quantity, the aspect of the sky, &c. ; then read the temperature 

 of the psjchrometer. 



The observations must be recorded for each instrument at the 

 moment when they are made, without trusting anything to the 

 memory. A strict rule should be laid down for one's self, to note 

 exactly the indications of the instruments, without subjecting 

 them mentally to any corrections or any reductions ; these should 

 not be applied until all the elements are at hand. 



If the observer has been unavoidably hindered from making 

 the observations at the exact hour, he will note in the column of 

 hours the number of minutes of the delay. If he is obliged to 

 procure a substitute, he must choose one accustomed to this kind 



