FROM MOOD AND COAT,. 1 <) 



the room, must therefore, from its situation, and the nature ol 

 its walls, change very little ; if. however, during an experiment, 

 any indication of an increase in its temperature is observed, 

 the upper sash in the window, which is suspended with 

 weight^, is lowered the required distance to correct it: but if 

 decreasing, a lire of wood can be immediately kindled in the 

 stove, a lamp being kept burning in this room for the purpose, 

 although never required hut in two instances during my ex- 

 periments. 



The required difference of temperature between the two 

 rooms being adjusted as described, it is maintained for aboul 

 half an hour by burning dry charcoal. The article to be sub- 

 jected to experiment is then accurately weighed, and if it 

 is wood, the unconsumed charcoal is wholly removed from the 

 stove by a small pair of tongs, and deposited in another room. 

 md the wood which is used in pieces two inches long, and 

 half to one quarter of an inch thick, is ignited by applying it 

 to the flame of a lamp; hut if it is any of the species of coal 

 which cannot be ignited per se, the burning charcoal is taken 

 from the stove and weighed, and its quantity either increased 

 or diminished so as to make half an ounce, which is, quicklj 

 returned to the stove, and on my notes, the name of the arti- 

 cle, its quantity, and the time, by an accurate watch, are then 

 sel down, together with the state of the thermometers, the 

 barometer, and hygrometer. The heights of the thermometers 

 are noted im-vy ten minutes during the experiment, thai in the 

 exterior room being always known by comparing the mercu- 

 rial and differential thermometers of the interior room. 



The last ten minutes of time which is entered to finish an 

 experiment, is that to which it approaches the nearest ; the dif- 

 ference therefore from the proper time; cannot be more, hut 

 will generally be less than five minutes, which is, in manj 

 s, as near perhaps as it can be determined, and the gn iteeri 

 difference stated will not affed the mean of the results on< 

 per cent. 



The anthracite coal cannot he whollj consumed, even in 

 the improved stat< of tbestove, the upper cbamln r hai ins bi 



