Magnitude of the Diurnal Motion of the Magnetic Needle. 221 



The divergence, according to this, is least in the month of 

 January and greatest in the month of May. 



It is generally believed that one principal cause of this diurnal 

 change is to be traced to the warming of the earth. I will not 

 venture to decide what part is played by the temperatm-e in the 

 production of these regidar alterations ; true it is that a minimum 

 of the declination occurs in the morning nearly at the time when 

 the temperature is lowest, and that a maximum occurs in the 

 afternoon about the time when the temperature is highest. But 

 the magnetic declination attains a second minimum between 10 

 and 12 o^clock at night, and a second maximum between 2 and 

 4 o'clock in the morning ; that is to say, two regular maxima 

 and two minima in the day, while in the temperature only one 

 minimum and one maximum occurs. If the heat were the only 

 or even the chief cause of the magnetic variation, the magnitude 

 of the daily variation of the declination must proceed side by side 

 with the increase and decrease of temperature during the year, 

 which is not proved by the observations which have been thus 

 far made. True it is that the said magnitude is least at the time 

 of lowest temperature ; but at the time of highest temperatui-e 

 (July) it has already decreased, having attained its maximum in 

 the month of May. 



But another remark involuntarily suggests itself when the 

 foregoing results are reflected upon : — 



The magnitude of the daily variation of the declination through- 

 out the months of the year runs parallel with the changes in the 

 humidity of the air ; is smallest at the time of greatest humidity, 

 and greatest at the time of maximum dryness. 



For the proof of this proposition I give here — 



a. The mean diurnal deviation of the declination in the single 

 months. 



b. The mean relative atmospheric humidity in parts of 100 in 

 the single months. 



c. The mean temperature in the single months from the year 

 1842 to 1850. 



