308 



cox. 



^/=/u 



l irrmm z 



Flue 



wall 



^•To aspir 



Fig. 1. 



The apparatus consisted of an iron gas pipe of 1.5 centimeters internal diam- 

 eter, passing through a suitable fixture attached to the shell of the chimney, 

 long enough to extend across the flue and leave a few centimeters projecting. 

 The inner end was capped and four holes 1.5 millimeters in diameter were 

 bored, one 7.5 centimeters from each wall of the flue and the other two dividing 

 the intervening distance into thirds. The two end holes were slightly enlarged 

 (about 0.2 millimeter) to counter balance the increased draft in the middle of 

 the chimney and the increased suction in the middle of the sampler when the 

 gases were exhausted. A piece of glass tube was fitted into the open end of 

 the iron pipe, by means of a tightly fitting plug, so that the end would rea<;h 

 to the middle of the perforated pipe. The apparatus was tested and proved to 

 have tight joints. The sampler was inserted into the flue and the gases drawn 

 off through the glass tube.' The holes were placed away from the current to 

 prevent their being filled with soot. An aspirator was constructed of a large 

 bottle fitted with the necessary siphon tubes. 



A concentrated salt solution was used in the aspirator. It is realized 

 that since the gases are somewhat soluble, this is not as accurate as their 

 collection over mercury, but is probably as accurate as the sample itself. 



'Attention has been called to the fact that samples taken with an apparatus 

 similar to this compared very favorably with those taken with the sampler 

 recommended by the American Society of Mechanical Engineers. V. S. G. S., 

 P. P. 48 (1906), 2, 311. 



