190 A. Pedhv—VoIafiUfij of some of the Ooiiipnuuds of Men-ury. [No. 2, 



reflected glare from the ground, it is protected from the actual direct 

 rays of tlic sun. It was not receiving any direct rays of the sun at the 

 time the following facts were observed, and its position is such that, 

 exeejjt for perhaps a very brief period during the winter season, it is 

 never reached by the sun's rays. This observatory was inspected on 

 October 23rd, 1888, and in the Torricellian vacuum of the barometer, at 

 some little distance above the surface of the mercury in the tube and 

 immediately behind the usual position of the vernier, I found a deposit 

 of from 100 to 120 small globules of mercury ranged all round the 

 inside of the tube, the size of which averaged about half that of an 

 ordinary pin's head. I could not actually determine the weight of 

 mercury which had thus distilled upwards, but I estimated it roughly to 

 have been more than half a gramme. The Superintendent of the obser- 

 vatory has informed me that the deposit of mercury was first noticed on 

 September 20th, 188S, and that after having been removed and tlio 

 instrument put right by mo on October 2.3rd, the deposit had again 

 commenced to form on the 15th of November. In this case then it is 

 proved that mercury can at the temperature of Buxar for the period 

 named give olf sufficient vajDour in vacuo to effect an actual slow dis- 

 tillation. The meteorological elements for Buxar in October 1888, 

 were : — 



mean maximum temperature 90'1° F = 32'3° C 

 mean minimum temperature 70'0° F= 21'1'^ 0 

 and the mean temperatiire of the solar radiation thermometer in Behar 

 in October was about 139'1° P — 59'5° C. 



The cause of the mercury collecting in the portion of the vacuous 

 part of the barometer tube shaded by the vernier is easily understood, 

 for it is this part which is protected from the glare from the ground 

 heated by the sun, and will therefore be at a slightly lower tomperaturo 

 than the rest of the tube. The only other cause which could be acting 

 in this case would be tho difference between the highly illuminated 

 and shaded parts of the tube. If this were the cause it would show 

 that the distillation was caused by the astiou of light, but this alter- 

 native can probably bo neglected. 



The fact of this slow distillation of mercury in a barometer tube 

 should be carefully noted by meteorologists, particularly in tropical 

 countries. In the case in question the barometer was one of Forfin's 

 principle, and the presence of the globules of distilled mercury did not 

 introduce any error into the barometric readings, but if tho instrument 

 had been constructed on tho Kew principle, as is the case very largely 

 now with barometers, the presence of this sublimed mercury would have 

 entirely vitiated the accuracy of the readings of the baromclor. 



