94 BULLETIN 1036, U. S. DEPARTMENT OF AGRICULTURE. 
expect the flasks to give more concordant results than the retorts, 
and this, in general, is true. 
The Forest Products Laboratory has shown (50) that, although 
there is no practical difference between the average numerical 
results obtained by the use of the flask and retort, the flask gives a 
somewhat sharper separation than the retort and is therefore to be 
preferred to the retort. 
The argument advanced in favor of the retort is that it has been 
used up to this time and should be continued. The argument 
against it is that it does not give as concordant results as the flask, 
and that the flask gives practically the same numerical value. Two 
associations — the National Electric Light Association (51) and the 
American Railway Engineering Association (52) — have made tests 
on the comparative checking value of the flask and the retort. The 
data obtained by them are published in their proceedings. A careful 
survey of these data shows that, after the elimination of what are 
evidently experimental errors, the ordinary flask is superior to the 
retort in checking value. 
OFFICIAL SPECIFICATIONS OF FOUR SOCIETIES. 
The specifications for the distillation tests adopted as standard 
by the American Wood Preservers' Association (1917), the American 
Railway Engineering Association (1917), the American Society for 
Testing Materials (1918), and the National Electric Light Asso- 
ciation (1921), are as follows: 
Retort. — This shall be a tubulated glass retort of the form and approximate dimen- 
sions shown in figure 45 with a capacity of 250 to 290 c. c. The capacity shall be 
measured by placing the retort with the bottom of the bulb and the end of the off- 
take in the same horizontal plane, and pouring water into the bulb through the tubu- 
lature until it overflows the offtake. The amount remaining in the bulb shall be 
considered its capacity. 
Condenser tube. — The condenser tube shall be a suitable form of tapered glass 
tubing of the following dimensions: 
Diameter of small end 12.5 mm. ; permissible variation 1.5 mm. 
Diameter of large end 28.5 mm.; permissible variation 3.0 mm. 
Length 360.0 mm. ; permissible variation 4.0 mm. 
Shield. — An asbestos shield of the form shown in figure 45 shall be used to protect 
the retort from air currents and to prevent radiation. This may be covered with 
galvanized iron, as such an arrangement is more convenient and more permanent. 
Receivers. — Erlenmeyer flasks of 50 to 100 c. c. capacity are the most convenient 
form. 
Thermometer. — The thermometer shall conform to the following requirements: 
The thermometer shall be made of thermometric glass of a quality equivalent to 
suitable grades of Jena or Corning make. It shall be thoroughly annealed. It shall be 
filled above the mercury with inert gas which will not act chemically on or contami- 
nate the mercury. The pressure of the gas shall be sufficient to prevent separation 
of the mercury column at all temperatures of the scale. There shall be a reservoir 
