MANNER OP OBSERVATION. 133 



than of the ground. The cause of this lies in 

 the slight density of the air, and in the increased 

 effect winch tins gives to any heat-influences act- 

 ing simultaneously with the sun upon the thermo- 

 meter. 



A thermometer of extreme accuracy, but rather 

 sluggish in its movements, should be suspended in 

 the shade, if possible on the north side of a build- 

 ing, at a distance of one foot from the wall ; and 

 its position should always be such that it is neither 

 exposed directly or indirectly to the sun's rays, 

 that it should not be wetted by rain, and that it 

 may be cooled as little as possible by nocturnal 

 radiation. The wind, however hard it may blow, 

 should have no effect at the point where the in- 

 strument is placed. If the index of a thermo- 

 meter exposed to the wind stands at any given 

 time lower than that of another, which is pro- 

 tected from its influence, it is evident that there 

 are certain foreign heating agencies which are not 

 perfectly shut out from the latter. 



To determine with accuracy the course of the 

 temperature in the course of the day, observa- 

 tions should be made at least every hour, i. e. 

 twenty-four times a day. 



The numbers of the hours are usually pointed 

 off on a horizontal straight line (for the abscissae), 

 dividing it into equal parts, and the correspond- 



