OCT. — MAR. 1858-59.] Report of the Magnetic Survey, 



349 



divided part. Besides the T ^ of the degree being read with per- 

 fect accuracy, it showed very rapidly even the minutest changes in 

 the atmospheric conditions. 



7. The surface of the ground. The thermometer was laying in 

 the reddish sand forming the general deposit in the plains of Hin- 

 doostan. It was during the first experiments covered with a 

 stratum of the sandy soil about 1 centimetre thick, but in the ex- 

 periments at Umballa, a little hole was filled with mercury and 

 this covered with the stratum of the soil. This arrangement made, 

 that the thermometer was, through the medium of mercury, in a 

 much more intimate contact with the particles of the soil, the tem- 

 perature of which it is to indicate. 



8 and 9, Metals exposed to a tropical sun are felt remarkably 

 hot when touched, though the conducting power has in this evi- 

 dently a great part. Their real temperature seemed an interest- 

 ing object for direct determination. I therefore exposed two metal 

 vessels, the one was an iron bottle, well screwed and filled with 

 mercury to two-thirds of its volume ; the other was a flat square 

 copper basin (used generally as an artificial horizon,) it was black- 

 ened, and about two cubic centimetres of mercury were put upon 

 it, forming isolated drops. Both were put upon a soft and thick 

 layer of cotton. For the observation of the copper basin, the 

 cotton was slightly pressed down at one place, so as to incline the 

 basin without touching this itself, and to cause the mercury to col- 

 lect itself in one corner. Then the thermometer was dipped in the 

 mercury, which it could be expected had assumed the tempera- 

 ture of the metallic surface with which it was in contact. 



The following Table contains the reading of the different ther- 

 mometers at Umballah ; the correction for index errors are appli- 

 ed to the reading. The instruments with black bulb, and on black 

 wool, were read with a telescope from 4 to 5 feet distance, since 

 any approaching for reading with a magnifier altered very sud- 

 denly the indications of the thermometer. 



Vol. xx. o. s. Vol. v. n. s. 



