370 S. A. Hill — Psychrometer and Condensing Hygrometer. [No. 4, 



vations, and to place behind it a dark coloured screen when it is in use. 

 For the purpose of the second series of observations, given below, the 

 vessel belonging to one of Regnaulfc's instraments, made about 15 years 

 ago by Casella, was freshly electro-plated, burnished, and polished by my 

 own hands. 



An objection to the use of all condensing hygrometers which at 

 first sight appears a serious one has been put forward in Symons' 

 Monthly Meteorological Magazine for June 1885 by Mr. R. Strachan, 

 who says, " A condensing hygrometer, whether Daniell's, Regnault's, 

 Dines's, or Alluard's, has the thermometer's bulb immersed in a cooling 

 medium and one surface of the dew plate is also in contact with the 

 cooling medium, but the surface upon which the dew is formed is cooled 

 by conduction, and is exposed to the air, which may be many degrees, 

 50 or 60, or more, warmer. In these circumstances when dew appears 

 the thermometer must be colder than the outside of the plate. When 

 the dew disappears the thermometer cannot have received the same 

 addition of heat as the outer surface of the plate." Had there been 

 any real weight in this objection, it would have doubtless been an- 

 ticipated by Regnault, who, however, merely says that the surface 

 on which the dew is deposited has the same temperature as the 

 liquid, because the metal is very thin and is in immediate contact 

 with the liquid, which must have sensibly the same temperature 

 throughout, since it is constantly stirred by the bubbles of air. A 

 little calculation will prove that, though the outside of the vessel is no 

 doubt warmer than the inside, as Mr. Strachan suggests, the difference 

 is so small as to be of no consequence whatever. 



Suppose the vessel to be made of copper and to be -^^ of an inch 

 thick. It is usually made of brass coated with silver, the combination 

 having probably about half the conducting power of copper, and being 

 therefore equivalent to one of copper twice as thick. Now at page 216 

 of Professor Tait's book on Heat are given several experimental values 

 for the thermal conductivity of copper, ranging from 4-11 to 2"04, on the 

 pound, foot, and second system of units, the mean of all the values being 

 3"3. This is the number of thermal units which would be transmitted 

 per second through a square foot of a plate one foot thick, if the two 

 surfaces were kept at temperatures differing by 1°. Through a super- 

 ficial area of 1 square foot and a thickness of -^^ inch the flow of heat 

 would be 3*3 x 12 X 50 = 1980 units per second. Under the as- 

 sumption made about the metal actually employed, the heat trans- 

 mitted would be half this or 990 units per second. Now suppose this 

 heat is brought to the plate by air blowing at the rate of 20 miles per 

 hour, — a somewhat extreme assumption, at any rate in India. In one 



